the journal of engineering research (tjer), vol. 14, no. 2 (2017) 137-144 doi: 10.24200/tjer.vol.14iss2pp137-144 development of bioclimatic design tool for oman using dry bulb and dew point temperatures open database n.a. al-azri*,a, y. zurigatb and n. al-rawahia * a department of mechanical and industrial engineering, college of engineering sultan qaboos university, muscat, oman. b department of mechanical engineering, american university of sharjah, sharjah, uae. received 14 march 2017; accepted 28 september 2017 abstract: bioclimatic charts are used by engineers and architects in implementing passive cooling systems and architectural optimization with respect to natural air conditioning. conventionally, the development of these charts is based on the availability of typical meteorological year which requires a record of meteorological data that are rarely available in sufficient amounts. bioclimatic charts in oman were developed earlier by the authors for limited locations based on the available typical meteorological years. using dry bulb and dew point temperatures only, bioclimatic charts are developed for adam, buraimi, ibra, muscat, nizwa, rustaq, saiq, salalah, suhar and sur. these charts are better representative of bioclimatic trends since their development is mainly based on the relevant parameters, namely dry bulb temperature and dew point. keywords: bioclimatic charts; passive cooling; tmy. تطوير تصميم أداة املناخ احليوي لعمان باستخدام قاعدة بيانات درجات حرارة املصباح اجلاف و قطر الندى أو نبيل الرواحي ب، يوسف زريقات*،أناصر العزري مت استخدام جداول املناخ احليوي من قبل املهندسني واملهندسني املعماريني يف تنفيذ أنظمة التربيد السليب والوصول :امللخص اىل امثل نظام معماري فيما يتعلق بتكييف اهلواء الطبيعي. ومن الناحية التقليدية ، يستند تطوير هذه اجلداول إىل توافر سنة ي يتطلب سجال لبيانات األرصاد اجلوية واليت نادرا ما تكون متوفرة بكميات كافية. وقد أرصاد جوية منوذجية , األمر الذ مت تطوير جداول املناخ احليوي يف عمان يف وقت سابق من قبل املؤلفني ملواقع حمدودة استنادا إىل سنوات األرصاد اجلوية ندى فقط ، مت وضع جداول مناخية حيوية ملناطق كل النموذجية املتاحة. وباستخدام درجات حرارة املصباح اجلاف وقطر ال من آدم و الربميي وإبرا ومسقط ونزوى والروستاق والسيق وصاللة و صحاروصور. و متثل هذه اجلداول أفضل االجتاهات احلرارة درجةتطويرها مبين أساسا على معامالت التغري ذات الصلة ، وهذه املتغريات هي نظرا الن لقياس املناخ احليوي .والرطبة اجلافة : جداول المناخ الحيوي ، التبريد السلبي ، السنة النموذجية لألرصاد الجوية الكلمات المفتاحية * corresponding author’s e-mail: nalazri@squ.edu.om mailto:nalazri@squ.edu.om n.a. al-azri, y. zurigat and n. al-rawahi 138 1. introduction bioclimatic charts are tools used at the design phase in building architecture. these charts help engineers and architects decide suitable passive air conditioning tools and also indicate the effectiveness of each tool. the development of bioclimatic charts usually requires the typical meteorological data at a given location. for the development of typical meteorological years (tmys), a full record of weather data including dry bulb temperature, dew point, wind speed and solar radiation must be available. typical metrological years (tmys) are usually intended for a wide spectrum of applications including energy simulation, renewable energy applications, agricultural and architectural applications. for such versatility of uses, all of the above meteorological parameters are required. however, a full record of all data is rarely available and if it is, the likelihood of having missing gaps, unrealistic measurements and inaccuracies is very high. in the development of typical metrological years, a historical record of hourly readings of the meteorological parameters is analyzed statistically. the historical record extends to up to 12 years with each parameter analyzed separately using frankenstein-statistics. for each of the 12 months in the year, a representative month is selected from the record based on a weighed sum of the different parameters. parameters that have greater impact will take a higher weight. since solar radiation impacts all other meteorological parameters, it is always assigned the highest weight in the development of a general-purpose typical metrological year. this statistical method was first devised by sandia national laboratories (nsrdb 1992) for us locations with historical records between 1961 and 1990. two other versions of sandia method, namely tmy2 and tmy3, came later and they were different in the weights assigned to each parameter. the same method was used in different works in the literature but by introducing or excluding some of the parameters and redistributing the assigned weights (al-azri 2016). al-azri (2016) used a tailored typical meteorological year for the development of bioclimatic charts which only considers the relevant bioclimatic parameters; namely, dry bulb temperature and dew point. in the development of the tailored typical meteorological year by al-azri (2016), the method developed by sandia national laboratories (hall et al. 1978, nsrdb 1992) was used; however, both solar radiation and wind speed were assigned zero weight. the other relevant parameters were assigned the weights shown in table 1. currently, bioclimatic charts for oman are only developed for eight locations where tmys already exist (al-azri et al. 2013). these locations are: buraimi, majis, seeb, sur, fahud, masirah, marmul and salalah using the typical meteorological year developed by sawaqed et al. (2005) based on the then available records of meteorological data, mainly between 1995 and 2004. in the presented work, dry bulb temperature and dew point are used to develop bioclimatic charts for adam, buraimi, ibra, muscat, nizwa, rustaq, saiq, salalah, suhar and sur. these locations, shown in fig. 1, feature diverse weather behavior. the northern costal locations are intensively affected by high humidity. the interior locations are usually hot and dry except saiq. two locations maintain special cases which are saiq due to its high elevation (1,755 m above sea level) and salalah which is affected by the monsoon. table 1. the weights used in the development of the typical meteorological year by al-azri (2016). parameter dry bulb temp. dew point maximum daily value 1/10 1/10 minimum daily value 1/10 1/10 mean daily value 2/10 2/10 daily range 1/10 1/10 figure 1. map of oman with the ten studied locations. development of bioclimatic design tool for oman using dry bulb and dew point temperatures open database 139 2. bioclimatic charts the bioclimatic chart model used in this work is developed by givoni (1992, 1994). it features five zones projected on a psychrometric chart with each zone representing a passive cooling technique. the typical meterological year for each location was developed. the average of the daily maximum and minimum dry bulb temperature and absolute humidity are calculated for each month. for each month, the average of the daily maximum temperature is matched with the average of the minimum daily absolute humidity represented by the pair (�̅�𝑀𝑎𝑥, �̅�𝑀𝑖𝑛). likewise, the average of the daily minimum temperature is matched with the average daily minimum absolute humidity to form (�̅�𝑀𝑖𝑛, �̅�𝑀𝑎𝑥). the two points are connected to form a line segment representing the distribution of the points in that month. several variants of bioclimatic charts can be found in the literature. all charts are overlaid on psychrometric charts and hence the same parameters, dry bulb temperature and absolute humidity, are used. with the knowledge of the average atmospheric pressure, absolute humidity can be calculated as a function of the dew point and the dry bulb temperature. relative humidity was calculated as a function of dry bulb temperature, dew point and atmospheric preassure; the complete calculations associated with the chart development can be found in (al-azri et al. 2013). 3. methodology due to the scarcity of solar radiation data, development of a typical meteorological year based on the conventional sandia method or any of its variants is not possible. typical meteorological years based on dry bulb temperature and dew point is followed as described by al-azri (2016). the method is advantageous in that it requires records of dry bulb and dew point temperatures only which are more commonly available than other meterological parameters. moreover, restricting the development of this typical year to the relevent parameters should be more accurate. having irrelevant parameters such as solar radiation, which is assigned greater weight in the conventional tmy development, can drive the typical values to lean towards the trend of the irrelevant parameters and hence compromise the accuracy of the representation of the more relevant parameters in the developed typical year. figure 2 shows the bioclimatic chart of muscat using the third generation sandia method (tmy3) along with the one based on the relevant parameters: dry bulb temperature and dew point. the comparison features identical distribution of the line segments on the bioclimatic charts but with little misalignment between the two approaches. the misalignment is due to the exclusion of the irrelevant parameters. 4. results records of hourly weather data were collected for ten locations for 12 years (2005-2016). the ten locations are adam, buraimi, ibra, muscat, nizwa, rustaq, saiq, salalah, suhar and sur. the data was acquired from the national climatic data center (ncdc) of the american national oceanic and atmospheric administration (noaa). information on the ten locations is given in table 2. for each of the locations, the typical meteorological years for dry bulb temperature and dew point were developed using records of hourly readings of the 12 years; from 2005 to 2016. the acquired data has many missing points. nevertheless, they do not have a continuous period of missing data and hence missing data were properly treated by interpolation. short hourly gaps were treated using linear interpolation while extended gaps were interpolated for using readings of the same hours in the adjacent days. figures 3 to 12 show the developed bioclimatic charts for adam, buraimi, ibra, muscat, nizwa, rustaq, saiq, salalah, suhar and sur, respectively. the ten charts represent different behaviour depending on the location. for most locations, natural thermal comfort mainly takes place in the months from november to february. the months from june to september are typically hot months where mechanical air conditioning is necessary. n.a. al-azri, y. zurigat and n. al-rawahi 140 figure 2. givoni bioclimatic chart for muscat, oman using sandia third version of meterological typical years (tmy3) approach (solid lines) and the approach based on the dry bulb and dew point temperatures (dashed lines) (al-azri 2016). table 2. the ten locations for which bioclimatic charts are developed. location station id latitude longitude elevation (m) adam 412640 22.5 57.367 328 buraimi 412440 24.233 55.783 299 ibra 412650 22.733 58.5 469 muscat 412560 23.593 58.284 14.6 nizwa 412550 22.867 57.55 462 rustaq 412530 23.417 57.433 322 saiq 412540 23.067 57.633 1,755 salalah 413160 17.039 54.091 22.3 suhar 412460 24.467 56.633 4.0 sur 412680 22.533 59.483 14.0 figure 3. bioclimatic chart for adam. development of bioclimatic design tool for oman using dry bulb and dew point temperatures open database 141 figure 4. bioclimatic chart for buraimi. figure 5. bioclimatic chart for ibra. figure 6. bioclimatic chart for muscat. n.a. al-azri, y. zurigat and n. al-rawahi 142 figure 7. bioclimatic chart for nizwa. figure 8. bioclimatic chart for rustaq. figure 9. bioclimatic chart for saiq. development of bioclimatic design tool for oman using dry bulb and dew point temperatures open database 141 figure 10. bioclimatic chart for salalah. figure 11. bioclimatic chart for suhar. figure 12. bioclimatic chart for sur. 143 n.a. al-azri, y. zurigat and n. al-rawahi 144 the ten locations can be divided into four groups:  interiors at normal elevation: adam, buraimi, ibra, nizwa and rustaq  interiors at high elevation: saiq  north coastal: muscat, suhar and sur  south coastal: salalah the interior locations at normal elevation are dry and hot. natural comfort is mainly maintened from part of november to february. the only interior location at high altitude which is saiq, has a very dry climate with thermal comfort restored from april to october while heating is necessary from november to march. the north coastal locations are apperantly very humid with natural comfort attained mostly from december to february. salalah in the south presents a special case due to the impact of the monsson and so dry bulb temperature is relatively lower than those of northern coastal regions. 5. conclusion due to the lack of complete record of weather and solar radiation data, developing bioclimatic charts based on dry bulb temperature and dew point is very useful to the majority of locations in oman. bioclimatic charts were developed for ten locations and they showed consistent trends with their geographical features. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references al-azri na (2016), development of a typical meteorological year based on temperature and humidity for passive cooling applications. energy for sustainable development 33: 61-74. al-azri n, zurigat y, al-rawahi n (2013), development of bioclimatic chart for passive building design. international journal of sustainable energy 32(6): 713-723. givoni b (1992), comfort, climate analysis and building design guidelines. energy and buildings 1: 11-23. givoni b (1994), passive and low energy cooling of buildings. van nostrand reinhold, new york. hall i, prairie r, anderson h, bose e (1978), generation of typical meteorological years for 26 solmet stations. sand78-1601. albuquerque, nm, sandia national laboratories. nsrdb (1992), user’s manual national solar radiation data base (1961-1990). version 1.0. golden, co: national renewable energy laboratory and asheville, nc: national climatic data center. sawaqed nm, zurigat yh, al-hinai ha (2005), step-by-step application of sandia method in developing typical meteorological years for different locations. international journal of energy research 29(8): 723–737. copy (2) of elec140704.qxd vol. 3, no. 1 (2006) 63-68ring researchrnal of engineethe jou 1. introduction oil wells, in general, produce oil, water, and gas. knowing the percentage of water in oil is very important for oil production. oil from nearby wells is carried via pipelines into a collection station. basic sediment and water (bs & w) devices are usually used to determine the percentage of the water in the produced oil. conventional bs & w devices are based on measuring the capacitance or impedance characteristics of the oil (smit, et al. 1998; lucas, 1994). however, such devices cannot be used in a full range of water percentage (0-100%). other devices such as radioactive and microwave (shaofan, 1994; beckwith, et al. 1981) are harmful and require a special operation procedure. fiber optic sensors (betta, et al, 1993; giallorenzi, et al. 1986; grattan, 1987) are also available. however, these devices are expensive. in this paper we are reporting experimental results for a new, user friendly, and accurate off-line device that can determine the water constitute in oil. the experimental results show that the device can attain a very high resolution up to +/0.4% and it can be used to measure a full range of water percentage levels (0-100%). in addition, the experimental results were compared with the theoretical ones and showed a good agreement. _______________________________________ *corresponding author’s e-mail: mhabli@squ.edu.om 2. proposed water-cut measurement method the proposed measurement method of the water-cut in a mixture of oil and water is based on the pressure of a sample from the mixture. consider the system in fig. 1. the system consists of an apparatus where the pressure of a liquid sample can be measured. the needed sample from a liquid is of a fixed quantity of height hs. in principle, for a water and oil mixture, the mixture sample will consist of heights ho and hw, respectively: (1) where, ho = height/level of water hw = height/level of oil the pressures of the oil and water are: (2) (3) where, p = pressure, kpa g = gravitational acceleration, m/s2 inexpensive and accurate measuring device for water constitute in oil m. habli*1, m. meribout1, a. al-naamany1, k. al busaidi2 1 electrical & computer engineering department, sultan qaboos university, p.o. box 33, muscat 123, oman 2 petroleum development oman, p.o. box 81, muscat 113, oman received 14 july 2004; accepted 24 october 2005 abstract:this paper presents an inexpensive and accurate measuring device for water constitute in oil. the new device is based on the relationship between the water constitute in oil and the pressure of a sample from the oil. experimental results show that the device can attain a very high resolution that can reach up +/0.4% and it can be used to measure a full range of water percentage levels (0-100%). experimental results showed good agreement with theory. keywords: water percentage, oil, pressure, inexpensive measuring device swo hhh =+ ghp ooo ρ= ghp www ρ= âjõdg »a aéªdg ᫪c ¢sé«≤d ≥«bo h …oé°üàbg ré¡l »∏ñm .ω1@•ƒhôe .ω ^1»f骩ædg .cg ,1…~«©°sƒñdg .∑ ,2 áá°°uuóóîîddggøe ¬æ«y øe §¨° dg h âjõdg »a aéªdg ᫪c ø«h ábó©dg ≈∏y ~ªà©j ~j~édg ré¡édg gòg .âjõdg »a aéªdg ᫪c áaô©ªd ≥«boh …oé°üàbg ¢sé«b ré¡l ω~≤j ábqƒdg √òg : ≈dg áaé°v’éh .aée (%100-0) øe ä’éëdg ™«ªéd πª©à°ùj ¿’ é° jg ¬æµªj h +/-%0^4 ≈dg õ««ªàdg ≈∏y ájq~b π°üj ¿g øµªj ré¡édg ¿g äô¡xg á«∏ª©ªdg èféàædg .âjõdg .√~«l ¬≤agƒe äô¡xg h ájô¶ædg èféàædéh âfqƒb á«∏ª©ªdg èféàædg ¿éa gòg áá««mmééààøøªªddgg ääggooôôøøªªddgg.…oé°üàbg ,§¨°v ,âjr ,aéªdg áñ°ùf : 64 vol. 3, no. 1 (2006) 68-68ring researchrnal of engineethe jou h = height/level, m (pressure head); ρ = density, kg/m3 the subscripts (o, w and s) stand for oil, water and sample, respectively. the total pressure pt at the bottom of the container is equal to: (4) where pa is the atmospheric pressure. subtracting pa from (4) yields: (5) ps is the pressure of the mixture sample which reflects the heights or the fractions of both oil and water in the sample. this system can be used to indicate the fractions of water and oil for different mixtures. substituting eqs. (2) and (3) in eq. (5) yields: (6) equations (1) and (6) are two independent equations with two unknowns, ho and hw. the sample quantity hs in eq. (1) is known and fixed. on the other hand, the sample pressure ps in eq. (6) is the measured pressure. for known ρo and ρw , it is possible to solve eqs. (1) and (6) for ho and hw as: (7) and (8) this device can also be used to find the densities of the oil and water samples ρo and ρw , respectively, from the pressures of two samples of known heights. for example, eq. (6) can be written for the first sample as: (9) similarly, for the second sample, eq. (6) can be written as: (10) equations (9) and (10) can be solved for: (11) and (12) it should be noted that this device is an off-line measuring instrument. in order to determine the water-cut in a flow, a sample from the flow is needed. this can be taken automatically from a side pipe through a valve that can close and open automatically by a switch or from a base station computer command. from the pressure of a known amount of the flow, we can determine the percentage of water in oil. 3. experimental setup the experimental setup consisted of a piezo-resistive pressure sensor. the sensor was attached to the bottom of an apparatus to measure the pressure of a sample. the output of the sensor was connected to an amplification circuit as shown in fig. 2. the sample height, hs , for all experiments was set to 25 cm. the pressure sensor that was used in the experiments was a differential type, i.e. the output of the sensor indicates ps as in eq. (5). figure 3 shows a photo for the experimental setup. the experimental samples were prepared using tap water and motor oil. 4. experiments a number of experiments were conducted to test the accuracy and the capability of the device. in the first experiment, a full range test was conducted starting from a 25 cm water level and no oil and ending with no water glass tube hs = ho + hw ho hw piezo-resistive pressure sensor ps=po+pw = ρohog + ρwhwg figure 1. a pressure measuring apparatus awot pppp ++= wos ppp += ghghp wwoos ρ+ρ= o ww s o h g p h ρ ρ− = osw hhh −= ghghp 1ww1oo1s ρ+ρ= ghghp 2ww2oo2s ρ+ρ= 2w 2oo 2s w h h g p ρ− =ρ 2w 2o1w 1o 2s 2w 1w1s o h hh h g p h h g p − − =ρ 65 8 63-6vol. 3, no. 1 (2006)ring researchrnal of engineethe jou and a 25 cm oil level with 1 cm increments. theoretical results were deduced from the equations presented in section 2.0 and compared with the experimental results. for the theoretical analysis, an oil density of ρo=0.87 gm/cm3 and a water density of ρw=1.104 gm/cm3 were used. figure 4 shows both the theoretical and experimental results. the data indicate that since oil has a less density, the pressure dropped as the oil level increased. the results show almost a linear relation between the oil level and the voltage. the second experiment was designed to test the accuracy of the device and if it can detect small changes in the mixture. this experiment began with a 25 cm water level and no oil. the oil in the mixture was then increased from zero to 0.8 cm using 0.1 cm increments. the results of measurements are shown fig. 5. from these results, it is clear that the device can detect the 0.1 cm change in the 25 cm mixture. the 0.1 cm step is equivalent to +/0.4% change in the oil concentration in the mixture. therefore, it is possible to deduce that the device has a resolution of at least +/0.4%. a third experiment was conducted and aimed also to test the accuracy of the device within another range of oil and water mixture. the range taken was from 20 cm water and 5 cm oil mixture to 19 cm water and 6 cm oil mixture in increments of 0.1 cm. the results are shown in fig. 6. in the fourth experiment, another mixture range was taken figure 2. the piezo-resistive pressure sensor and its amplification circuit figure 3. experimental set up 66 vol. 3, no. 1 (2006) 63-68ring researchrnal of engineethe jou from 5 cm water and 20 cm oil stand to 3 cm water and 22 cm oil stand in increments of 0.1 cm. the obtained results are shown in fig. 7. the experiment was conducted using four trials in order to determine the standard error and precision of the system. as an example, table 1 displays the measurement results of the first experiment alongwith the average and standard error. the overall average precision on the pressure measurement was determined to be approximately 0.083 matm. 5. analysis of the experimental results based on the experiments conducted, it was concluded that the device can be used and trained to detect water-cut in an oil and water mixture. the results of the first experiment showed a clear change in the output voltage as the percentage of the oil increased in the mixture. the other three experiments showed that the device can detect small changes at different oil level ranges. in all the ranges considered, the device was able to clearly detect 0.1 cm pr es su re ( at m ) oil level (cm) 0.028 0.027 0.026 0.025 0.024 0.023 0.022 0.021 0.020 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0.0256 0.02555 0.0255 0.02545 0.0254 0.02535 0.0253 5 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5 oil level (cm) pr es su re ( at m ) 67 vol. 3, no. 1 (2006) 63-68ring researchrnal of engineethe jou change of oil in the mixture. the 0.1 cm change in a 25 cm level is equivalent to a measurement resolution of +/0.4%. one other very important observation is that the device can operate over a full range of water-cut from 0100%. 6. conclusions a number of water-cut measurement devices were evaluated, which included capacitance and conductance devices. other devices such as microwave, radioactive and fiber optics sensor were either unsafe or expensive cannot be used in the whole water-cut range. the proposed technique, which is based on the relationship between the water-cut in an oil and water mixture and the pressure of a sample from the mixture, is to be very cheap compared to the existing devices. this device is robust, user friendly and can operate in the whole water-cut range. pr es su re ( at m ) 0.02685 0.0268 0.02675 0.0267 0.02665 0.0266 0.02655 0.0265 0.02645 0.0264 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0222 0.02215 0.0221 0.02205 0.022 0.02195 0.0219 0.02185 0.0218 0.02175 0.0217 0.02165 pr es su re ( at m ) 20 20.2 20.4 20.6 20.8 21 21.2 21.4 21.6 21.8 22 86 vol. 3, no. 1 (2006) 63-68ring researchrnal of engineethe jou experimental results showed that the new, inexpensive, and simple measuring device for water-cut in oil can attain a very high resolution that can reach up +/0.4% and it can be used to measure a full range of water-cut levels (0100%). acknowledgements this work was supported by petroleum development oman (pdo). references beckwith, t.g., 1981, mechanical measurements, addison wesley publishing co. pp. 445-477. betta, g. and d'apuzzoand, l.m., 1993, "an intrinsic optical fiber sensor", imeko tc4, pp. 113-119. grattan, k.t.v., 1987, "recent advances in fiber optic sensors," measurement, vol. 5, pp. 122-134. giallorenzi, t.g., 1986, "optical fiber sensors challenge the competition", ieee spectrum. lucas, g.p., 1994, "flow rate measurement in vertical oil-water flows using conductivity sensors and a void fraction wave model," advances in sensors for fluid flow measurement, pp. 13/1-13/3. smit, q., mortimer, b.j.p. and tapson, j., 1998, "general purpose self-tuning capacitance sensor [for oil recycling and soil moisture measurement application]," instrumentation and measurement technology conference, imtc/98, vol. 2, pp. 107478. shaofan, w.q.d., 1994, "a high-accuracy microwave sensor and calibration for measuring three-phase saturations in cores," instrumentation and measurement technology conference, imtc/94, vol. 3, pp. 1273-76. shaofan, w.q.d., 1994, "a microwave technique for measuring three-phase saturations in cylindrical cores," precision electromagnetic measurements, pp. 71-71. pressure (matm) oil (cm) trial 1 trial 2 trial 3 trial 4 average std error 0 25.258 25.156 25.342 25.529 25.321 0.079 1 25.202 25.034 25.286 25.408 25.233 0.079 2 25.137 24.932 25.184 25.305 25.140 0.078 3 25.034 24.857 25.109 25.230 25.058 0.078 4 24.904 24.773 25.025 25.146 24.962 0.080 5 24.773 24.717 24.969 25.090 24.887 0.087 6 24.661 24.577 24.829 24.950 24.754 0.084 7 24.559 24.493 24.745 24.866 24.666 0.085 8 24.428 24.428 24.680 24.708 24.561 0.077 9 24.372 24.344 24.596 24.642 24.489 0.076 10 24.204 24.288 24.540 24.540 24.393 0.087 11 24.139 24.148 24.400 24.428 24.279 0.078 12 23.989 24.064 24.316 24.288 24.164 0.081 13 23.859 23.905 24.157 24.111 24.008 0.074 14 23.775 23.840 24.092 23.989 23.924 0.072 15 23.672 23.737 23.989 23.887 23.821 0.072 16 23.532 23.625 23.877 23.812 23.712 0.080 17 23.430 23.467 23.719 23.728 23.586 0.080 18 23.327 23.402 23.653 23.672 23.514 0.087 19 23.196 23.299 23.551 23.532 23.395 0.088 20 23.075 23.187 23.439 23.448 23.287 0.093 21 22.972 23.047 23.299 23.290 23.152 0.084 22 22.851 22.870 23.122 23.224 23.017 0.093 23 22.730 22.748 23.000 23.122 22.900 0.096 24 22.608 22.646 22.898 23.010 22.791 0.097 25 22.487 22.571 22.823 22.870 22.688 0.094 average 0.083 table 1. measurement results of the first experiment along with the average and standard error values the journal of engineering research (tjer), vol. 15, no. 2 (2018) 124-128 doi: 10.24200/tjer.vol15iss2p124-128 tuning the resonance frequency and miniaturization of a novel microstrip bandpass filter a. rezaeia,* and l. noorib a department of electrical engineering, kermanshah university of technology, kermanshah, iran b young researchers and elite club, kermanshah branch, islamic azad university, kermanshah, iran received 8 june 2016; accepted 24 december 2017 abstract: in this paper, a compact microstrip bandpass filter is designed using two open loop resonators. in order to obtain a tunable bandpass response with low insertion loss, two stubs are loaded inside them. the design process is based on obtaining the input admittance. then, using the input admittance, a method is presented to control the resonance frequency and miniaturization simultaneously. the obtained insertion loss and the return loss at the resonance frequency are 0.1 db and 19.7 db respectively. to verify the design method, the proposed filter is fabricated and measured. the measured results are in good agreement with the simulated results. keywords: microstrip filter; tunable; miniaturization; insertion loss. نطاق ترددي مضغوط باستخدام شرحية جديدة متناهية الصغري بوليلى نوري * ، أ عباس رضائي ذواتا حلقة مفتوحة . و من أجل ني باستخدام رناناتمرشح نطاق ترددي مضغوط ، مت تصميم البحثية : يف هذه الورقة امللخص وتعتمد عملية .احلصول على استجابة نطاق قابل للضباط ذو خبسارة منخفضة عند اإلدراج ، يتم حتميل شرحيتني بداخله ا بتقديم طريقة للتحكم يف تردد الرنني والتصغري يف الستخدام مدخل االدراج ، قمنو .التصميم على معرفة حاصل اإلدخال دي 19,7. دي بي و ,اوتبلغ نسبة اخلسارة يف اإلدراج اليت مت احلصول عليها واخلسارة عند رجوع تردد الرنني: .نفس الوقت قياس توافقا بشكل جيد و اظهرت نتائج ال .وللتحقق من طريقة التصميم ، يتم تصنيع الفلرت املقرتح وقياسه .على التواليبي .مع نتائج احملاكة اخلسارة يف اإلدراج ،التصغري ،قابل لالنضباط ،رغشرحية متناهية الص ،مرشح : الكلمات املفتاحية * corresponding author’s e-mail: unrezaei@yahoo.com mailto:unrezaei@yahoo.com a. rezaei and l. noori 125 1. introduction high performance and compact microstrip filters play an important role in developing modern microwave/rf communication circuits and systems. to design microstrip filters several types of resonators such as: open loops (hayati et. al. 2012; salehi et. al. 2014; hayati et. al. 2014; zhang et. al. 2012), parallel-coupled lines (moradian et. al. 2009; othman et. al. 2013; fathelbab et. al. 2005; kuan et. al. 2010) and quarter/half wavelength resonators (li et. al. 2010; deng et. al. 2010) have been used. also, several approaches on suppressing harmonics have been proposed (kang et. al. 2010; wang et. al. 2008; chen et. al. 2009; cheng et. al. 2014; wu et. al. 2017). in zhang et. al. (2012), three open loop resonators are used to design a bandpass filter with a high selectivity. however, it has an undesired fractional bandwidth. the common weakness of these reported filters is their large sizes and large insertion losses. in moradian et. al. (2009) a third-order bandpass filter is proposed to attenuate the harmonics. in kuan et. al. (2010), the parallel coupled lines and step impedance resonators are used to design a microstrip bandpass filter. a microstrip bandpass filter operating at 1.78 ghz with suppressed harmonics up to 6.2 ghz is presented in li et. al. (2010). but this filter has a large return loss. in kang et. al. (2010), a ringbalun bandpass filter has been proposed to attenuate the harmonics up to 12 ghz. in wang et. al. (2008), coupled ring resonators have been utilized to design a bandpass filter with a complex structure, while using a complex structure leads to hard fabrication. in chen et. al. (2009), various types of bandpass filters have been introduced utilizing cross coupled resonators to suppress the second harmonic. nevertheless, their stop bands are narrow. in cheng et. al. (2014), the inductive-coupled stepped-impedance quarter-wavelength resonators are used to design a microstrip bandpass filter with an undesired insertion loss. a dualmode microstrip bandpass filter operating at 1.67 ghz has been introduced in lu et. al. (2017), which has a narrow fractional bandwidth and large area. a microstrip bandpass filter with high return loss has been designed in guan et. al. (2017). in this filter, a triangular cell has been coupled to the step impedance feed structures that results in improving the bandwidth. in this paper, first a simple resonator (open loop) is chosen similar to hayati et. al. (2012); salehi et. al. (2014); and hayati et. al. (2014). then, a method is proposed for miniaturizing and tuning the resonance frequency simultaneously. using this method, the shapes and dimensions of the internal stubs can be determined as well as the resonance frequency. next, a simple tunable bandpass filter is presented using two proposed resonators to solve the problems of previous works in terms of large implementation area and large insertion loss. in addition, the harmonics are attenuated reasonably and fractional bandwidth and selectivity are improved. finally, the effect of changing the dimensions on the frequency response is investigated. 2. filter design an open loop resonator is shown in fig. 1a. a large size simple step impedance cell is loaded inside the proposed resonator to control the resonance frequency. the electrical lengths θi are used to obtain the input admittance, where i=1, 2, 3, s. an lc equivalent circuit of the proposed resonator is shown in fig. 1b, where c and l are the capacitors and inductors of the bends respectively. the parameters cg and cp present the gap capacitors. the parameters of la, lb, lc, le, lf and lg are the inductors of the stubs with the physical lengths la, lb, lc, le, lf, and lg respectively. co is the capacitor of the open end, which its position is subsequent the loaded step. cs, ls1 and ls2 are the capacitance and inductance of the step impedance cell. the input admittance from the open end of θ1 can be written as follows: )](tan)(tan)()[tan(tan1 )(tan)(tan)(tan)(tan jyy s321 s321 in      (1) where θs is the total electrical length of the step impedance cell, y is the admittance of the step impedance cell and open loop resonator. for the even mode, when, yin=0, the resonance condition is obtained from: 0)(tan)(tan)(tan)(tan s321   (2) so the resonance condition can be tuned by adjusting θ1, θ2, and θ3 while θs is fixed. the electrical length has a direct relation with the physical length. therefore, the resonance frequency can be controlled by adjusting the loop and step impedance open stub dimensions. tuning the resonance frequency and miniaturization of a novel microstrip bandpass filter 126 when the electrical length θs is maximum, then from θs= ls β (where β is propagation constant) ls must be maximum. therefore, by choosing a maximum value for ls, (from equation 2 and also from θs= ls β) tan(θs) is increased and the total of tan(θ1)+ tan(θ2)+ tan(θ3) is decreased (tan(θ1)+ tan(θ2)+ tan(θ3) is minimum). under this condition θ1, θ2, and θ3 are small values. therefore, (from θ3=(la+ lg+ lf) β, θ2= lcβ and θ1= leβ) the open loop dimensions consisting of la, lb, lc, le, lf, and lg are decreased. this is a method to decrease the resonator size and adjust the resonance frequency simultaneously. as a total result of above discussion, a method to control the resonance frequency and miniaturization is obtained by the following steps: in the first step, a resonator is selected which some stubs are loaded inside it. in the second step, the main resonator size is decreased and the dimensions of stubs are increased, so that the desired resonance frequency can be obtained. in the proposed resonator (fig. 1a), the physical lengths la, lb, lc, le, lg and lf must be smaller while the internal stub length must be larger. therefore, the inductors la, lb, lc, le, lg and lf can be smaller and ls2 or/and ls1 can be larger. according to equation (2), in some cases, there is a degree of freedom to control the resonance frequency so that we have to use the optimization method. therefore, in order to obtain a compact size at the target resonance frequency, the additional optimization is performed. in order to design a bandpass filter as shown in fig. 2a, two open loop resonators consisting of different step impedance cells are used. the loops are connected together using mix coupling. the coupling structure consists of three coupled lines with different widths, which are used to attenuate the harmonics. the feed structures with the step impedance forms are added to the input and output ports to decrease the insertion loss without size increment. the simulated and measured frequency responses of the propose filter are shown in fig. 2b. according to the above discussion, the resonance frequency can be controlled by adjusting the physical lengths of open loops while the internal stubs have a maximum size. therefore, a method to control the resonance frequency is changing of the lengths (l3, l9) or/and (l4 , l11). the frequency response as a function of l3, l9, l4 and l11 are shown in fig. 3a and fig. 3b. figures 3a and 3b depict the loop size changing effect on the resonance frequency. when the loops are large, the resonance frequency is shifted to the left. when the loops are small, the resonance frequency moves to the right. a photograph of the fabricated filter is shown in fig. 3c. figure 1. proposed resonator (a) layout, (b) lc equivalent circuit. figure 2. proposed filter (a) layout, (b) frequency response. a. rezaei and l. noori 127 3. results the proposed filter is simulated by advanced design system (ads) full wave em simulator. the proposed filter is designed and fabricated on a rt_duroid_5880 substrate with 15 mildielectric thickness, εr=2.22 and loss tangent 0.0009. the dimensions of the proposed structure shown in fig. 2a are presented in table 1. this filter has the cutoff frequencies at 1.68ghz, 1.93ghz and the resonance frequency is 1.8ghz. the harmonics are attenuated from 2.16 up to 7.1ghz with a maximum level of 19.5db. therefore, the harmonics are attenuated up to 3.94fo where fo is the resonance frequency. the obtained insertion loss at 1.8ghz is better than 0.1db, while the return loss is better than 19.7db. the filter size is 0.23λg × 0.1λg (26.7×12mm²). in comparison to the previous works, the filter size is small and the best insertion loss is obtained. the achieved fractional bandwidth (fbw) is 14%. the insertion loss, return loss, fractional bandwidth, and filter size are compared to the previous works in table 2. in the table 2, il, rl, and fbw are the insertion loss, return loss, and fractional bandwidth respectively. 4. conclusion a simple compact low loss bandpass filter is designed, fabricated and measured for rf systems. in this structure, two open loop resonators are loaded by two large stubs. by using the additional large size microstrip cells inside the open loops, the resonance frequency is decreased without size increment. the overall size of the proposed filter is 320mm². in comparison to all of the references, the lowest table 1. the dimensions of proposed structure (in mm). parameters l1 l2 l3 l4 l5 l6 l7 l8 value 0.3 1 6.8 3.5 3 3.5 9.7 4.5 parameters l9 l10 l11 l12 l13 w1 w2 w3 value 1 4.5 4.8 0.2 1 3 1.2 1 parameters w4 w5 w6 w7 s g value 5 0.6 0.5 1 0.1 0.4 table 2. comparison between the proposed filter and previous works. reference il (db) rl (db) fbw size (mm2) fo(ghz) this work 0.1 19.7 14% 320 1.8 (hayati et. al. 2012) 0.53, 0.59 10, 13.4 --320 2.4, 5.2 (salehi et. al. 2014) 0.35 , 0.25 13.6, 18.2 17%, 13% 363 2.4, 5.7 (hayati et. al. 2014) 0.2, 0.4 15, 12.7 13%, 11% 233 2.4, 5.2 (zhang et. al. 2012) 2.2 --6% 11595 0.95 (moradian et.al. 2009) 3.85 --13% 9000 1 (kuan et.al. 2010) 2.5 15 12% 2034 2 (li et.al. 2010) 1. 8 12 16.5% 488 1.78 (deng et.al. 2010) 1.66 --9.5% 606 2.41 (kang et.al. 2010) 1.34 21.6 4.08% 1564 2.45 (chen et.al. 2009) 0.83 --11.2 2300 2 (cheng et.al. 2014) 0.85 --11.5% 220 2.35 tuning the resonance frequency and miniaturization of a novel microstrip bandpass filter figure 3. (a) frequency response as a function of l3 and l9 , (b) frequency response as a function of l4 and l11 , (c) a photograph of the fabricated filter. insertion loss is obtained, while the harmonics are attenuated from 2.16 to 7.1ghz with a maximum attenuation level of -19.5db. in addition, the resonance frequency is tuned by calculation of the input admittance conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references chen wl, wang gm (2009), effective design of novel compact fractal-shaped microstrip coupled-line bandpass filters for suppression of the second harmonic. ieee microwave and wireless components letters 19(25): 74-76. deng ph, chiu pt (2010), new bandpass filters using half-wavelength and branch-line resonators. progress in electromagnetics research, 16: 241-249. fathelbab wm, steer mb (2005), parallelcoupled line filters with enhanced stopband performances. ieee transactions on microwave theory and techniques 53(12): 3774-3781. guan x, gui p, xiong t, ren b, zhu l (2017), hybrid microstrip/slotline ultra-wideband bandpass filter with a controllable notch band. international journal of antennas and propagation 2017: 1-7. li yc, zhang xy, xue q (2010), bandpass filter using discriminating coupling for extended out-of-band suppression. ieee microwave and wireless components letters 20 (7): 369-371. moradian m, tayarani m (2009), improving the stopband of microstrip parallel-coupled line bandpass filters using partially coupled resonators. iet microwaves, antennas and propagation 4(7): 906-916. othman ma, sinnappa m, hussain mn, abd mza, ismail mm (2013), development of 5.8 ghz microstrip parallel coupled line bandpass filter for wireless communication system. international journal of engineering and technology 5(4): 3227-3235. salehi mr, abiri e, noori l (2014), design of a microstrip dual-band bandpass filter with compact size and tunable resonance frequency for wlan application. international journal of electronics communication and computer engineering 6(3): 248-251. wang y, wang bz, wang j (2008), the design of coupled resonator bandpass filter with wide stop-band. ieee microwave and wireless components letters 18(4): 251-253. wu xh, wei xb, lv hg, shi y (2017), a compact trisection microstrip bandpass filter with frequency-dependent cross-coupling. electromagnetics 37(7): 454-461. zhang xy, xue q (2009), harmonic-suppressed bandpass filter based on discriminating coupling. ieee microwave and wireless components letters 19(11): 695-697. 128 http://ieeexplore.ieee.org/search/searchresult.jsp?searchwithin=%22authors%22:.qt.w.%20m.%20fathelbab.qt.&newsearch=true http://ieeexplore.ieee.org/search/searchresult.jsp?searchwithin=%22authors%22:.qt.m.%20b.%20steer.qt.&newsearch=true microsoft word paper 9.docx the journal of engineering research (tjer), vol. 14, no. 1 (2017) 94-104 mathematical model describes treatment of waste water using modified activated carbon a.s. ibrahim*, a.s. al buloshi, s.s. al zaabi and l.a. al yafai chemical engineering department, dhofar university, salalah, sultanate of oman received 7 march 2016; accepted 2 november 2016 abstract: the proposed mathematical model covered in this paper includes the most important parameters associated with the rates of adsorption and desorption. also, partial pressure is included since it is an important factor that affects rates of adsorption and desorption. the study focuses on the effects of the constant rates on adsorption of pollutant concentrations for benzene, nickel, cadmium, and copper using modified active carbon. when the rate constant of adsorption decreases, the pollutant concentration will also decrease, yielding high acceptable evidence of the logic of the proposed mathematical model. also, the proposed model is compared with experimental data and other models to give good outcomes with high accuracy. keywords: mathematical model, waste water, adsorption, desorption, activate carbon. אא אאאא kאk*אkkאkk،אkk،    wא אאא    א   א א אא  א אאkאאאאkאא א א ،א ،א   א א    א א kאאאאא،אא      א א    kא א   א kא  אאw،אא،א،אאאא،א.  * corresponding author’s e-mail: ahmadsaadi47@yahoo.com a.s. ibrahim, a.s. al buloshi, s.s. al zaabi and l.a. al yafai   95 nomenclature symbol definition (unit) a waste water. , rate constant for adsorption of forward direction. ca.s concentration of pollution of adsorption for surf modified activated carbon. ca concentration of adsorption. cai initial concentration of adsorption. cv v modified activated carbon concentration. , rate for constant adsorption of a reverse direction. , rate for constant desorption of a reverse direction. , rate for constant desorption of a forward direction. pa partial pressure. r rate of adsorption. rd rate of desorption. ra rate of adsorption. s modified activated carbonative carbon. vi volumetric flow rate. v volume. t time (s). kfr adsorption cap modified activated carbonity of the sorbent mg/g (l/mg)1/n. n freundlich’s constants. qe amount of adsorbate adsorbed at equilibrium (mg/g). qmax maximum monolayer adsorption cap modified activate carbonity of the adsorbent (mg/g). ce equilibrium concentration of adsorbate (mg/l). kln langmuir’s adsorption constant related to free energy adsorption (l/mg). ctm temkin’s constant related to the heat of sorption (j/mol). ctmi temkin’s isotherm constant (l/g). r gas constant (8.314 j/mol k). t absolute temperature (k). k1 first-order rate constant (min−1). qt amount of adsorbate adsorbed at any time (mg/g). mathematical model describes treatment of waste water using modified activated carbon   96   1. introduction waste water is one of the biggest pollution problems in the world, and suggestions for its treatment have been drawn from experimental and theoretical works. in experimental work, a simplified method for ion exchange has been used to evaluate kinetic data. many techniques have been explored using modified activated carbontivated carbon to remove carbon dioxide. one of the most important techniques that has been used is semi-batch remodified activated carbonator fixed inside modified activated carbontivated carbon to adsorb methane and carbon dioxide (prasetyo and dod 1998). another technique used biomass of sargassum fluitant to make bio sorption for heavy metals and tannin gel has been used to adsorb chromium (nakano et al. 2001). some special materials have been used to remove cadmium and mercury ions from aqueous solutions using sorption on treated pinus pinaster bark (vazquez et al. 2002). numerous experimental studies have used rice husk as an adsorbent for waste water treatment (ajmal et al. 2003), and chitosan-cellulose hydrogel beads have been used to adsorb copper (li and bai 2005). modified activated carbontivated carbon has been represented as one of the most effective modified activated carbontivated adsorption materials to be used to absorb carbon dioxide (bog et al. 2006). many researchers have used modified activated carbontivated carbon for waste water treatment (afsaneh et al. 2008; mohamed et al. 2008; muhammad et al. 2008). a hybrid technique was used to produce a t-shirt model for waste water treatment. other exploratory works have explored waste water treatment options through a combination of experimental and theoretical work. the first adsorption mathematical model which was proposed in 1906 by freundlich has been used in hetrogenous surf modified activated carbon adsorbent systems where the binding sites are not equivalent. a form of the freundlich’s model can be represented as follows: / (1) (2) the constants kfr and n can be evaluated from the intercept and the slope of the linear plot of experimental data of ln qe versus ln .ec a second important mathematical model for adsorption is langmuir’s isotherm model (okieimen and ogbeide 2009), which depends on an isothermal state when all the sites are homogenous compare to freundlish’s model, and all these sites are filled by molecules to be adsorbed. the linear form of the langmuir isotherm can be represented by the following equation: (3) the values of the constants kl and qmax can be evaluated from the intercept and the slope of the linear plot of experimental data of ce/qe versus ce. temkin and pyzhev (lalhruaitluanga et al. 2010) studied extensively the heat of adsorption and the adsorbate-adsorbent intermodified activated carbonation on adsorption isotherms. temkin and pyzhev’s mathematical model can be represented as follows: ln ln (4) the constants ctmi and ctm can be determined from the intercept and the slope of the linear plot of the experimental data of qe versus ln ce. the values of the constants ctmi and ctm are listed in table 1. the lagergren mathematical model is represented as proportional to the first power of sorption cap modified activated carbonity of the adsorbent and can be expressed as follows (khaled et al. 2009). table 1. langmuir, freundlich and temkin models’ constants and correlation coefficients for sorption of methylene blue (mb) into modified activated carbon. isotherm parameters values langmuir qo (mg/g) kln (l/mg) 8.75 0.23 fruendlich kfr n 4.21 4.13 temkin ctmi (l/g) ctm 11.3 1.5 a.s. ibrahim, a.s. al buloshi, s.s. al zaabi and l.a. al yafai   97 (5) integrating eq. (5) for the initial and end conditions t = 0 to t = t and qt = 0 to qt = qt, and, after some rearrangement, a linear plot is obtained: . (6) values of k1 and qe were obtained from the slope and intercept, respectively. table 2 lists lagergren’s mathematical model constants. . (6) plots of log (qe − qt) versus t for the lagergren mathematical model where the values of k1 and qe were obtained from the slope and intercept, respectively. table 2 lists lagergren’s mathematical model constants. 2. methodology deriving a mathematical model requires the provision of assumptions and a comparison with other models to create new ideas for a proposed model (tables 3 and 4). adsorption and desorption states represent active mechanisms for the system (fig. 1) and can be derived as follows: 2.1. adsorption state this state depends on the properties of surface of adsorbent, partial pressure of fluid and rates constant of adsorption as seen in equations below: ⇌ , , . (7) , , . (8) , . , , (9) substitute (9) in (8) ∗ , , (10) substitute (10) in (9) table 2. correlation coefficients for adsorption of benzene on modified activated carbontivated carbon. initial concentration 10 g/l qe exp. (mg/g) pseudo-first-order kinetic model qe cal. (mg/g) k1 (1/min) r2 benzene 4.14 2.3 0.003 0.915 nickel 6.31 3.8 0.003 0.898 copper 6.53 4.21 0.0037 0.927 cadmium 7.31 5.47 0.0028 0.934 table 3. the list of proposed mathematical model assumptions. 1. the fixed bed comprises liquid and solid phases. 2. no chemical remodified activate carbonation occurs inside a fixed bed. 3. negligible radial temperature and concentration gradients exist in the fixed bed. 4. adsorption occurs to adsorbent solid particles inside holes of modified activated carbon. 5. desorption occurs to transfer solid particles from inside holes to the surf modified activated carbon of modified activated carbon. 6. mass transfer occurs from surf modified activated carbon of modified activated carbon to the bulk flow. 7. the dynamics study is represented by the rates of adsorption and desorption at the surf modified activated carbon of modified activated carbon. 8. heat transfer is considered in the new mathematical model. mathematical model describes treatment of waste water using modified activated carbon   98   table 4. differences between the proposed mathematical model and the other models. no functions new mathematical model langmuir model freundlich isotherm model pseudo-firstorder model 1 phases liquid, gas and solid phase liquid phase liquid phase liquid phase 2 mass transfer calculated without chemical remodified activate carbonation not calculated not calculated calculated without chemical remodified activate carbonation 3 adsorption mass transfer from surf modified activate carbon to the entrails holes of modified activate carbon without chemical considered considered considered 4 desorption mass transfer from entrails holes of modified activate carbon to the surface activate carbon without chemical not considered not considered not considered 5 energy transfer considered not considered not considered not considered r ate of de sorption to the bulk flow r ate of adsorption   figure 1. steps of adsorption process. 2.2 desorption state this state depends on the same properties of adsorption but the molecules of adsorbent left to the bulk flow in opposite direction of adsorption’s flowrate as seen in equations below: , . ∗ (11) a.s. ibrahim, a.s. al buloshi, s.s. al zaabi and l.a. al yafai   10 . ⇌ , , (12) , . , (13) , . , , (14) ∗ , , (15) let , . ∗ (16) the mechanism of the rate of adsorption can be controlled in all states of the system. thus, this is the most important assumption. 0. from equation (16) c . ∗ (17) substitute (17) in (11) , ∗ ∗ (18) . ∗ (19) ∗ (20) substitute (20) in (18) , ∗ ∗ , ∗ (21) representing the system as a continuous stirred-tank reactor (cstr) process: (22) the volumetric flow rate compared to the volume of the system is very limited, so 0 (23) equation (22) will be (24) the rate of remodified activate carbonation is represented as (25) substitute equation (21) in (25) , ∗ (26) solve equation (26) , ∗ t constant (27) boundary conduction at t=0 , c c , (28) substitute in (21) , ∗ t c (29) mt , ∗ 3. results and discussion the discussion that follows focuses on the experimental results associated with adsorbent benzene. the proposed mathematical model yielded good behavior for the experimental data from adsorbent benzene against the first order lagergren model (fig. 2). figure 3 shows results for adsorbent nickel using the proposed model and lagergren model. also, the proposed mathematical model was highly accurate, with results close to experimental data as compared to other models for copper and cadmium (figs. 4 and 5, respectively). rates of adsorption and desorption have a big effect on pollutant concentrations. the rate of adsorption decreased the change of pollutant concentrations and also decreased the benzene, nickel, cadmium and copper (figs. 6–9, respectively). 99 mathematical model describes treatment of waste water using modified activated carbon   98     figure 2. comparison between proposed mathematical model and lagergren model for benzene. figure 3. comparison between proposed mathematical model and lagergren model for nickel. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.8 1 1.2 1.4 1.6 1.8 2 time (h) c o n c e n tr a ti o n g /l actual data propose model lagergren model 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 time (h) c o n c e n tr a ti o n g /l actual data propose model lagergren model 100 a.s. ibrahim, a.s. al buloshi, s.s. al zaabi and l.a. al yafai   101 figure 4. comparison between proposed mathematical model and lagergren model for copper.   figure 5. comparison between proposed mathematical model and lagergren model for cadmium. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.8 1 1.2 1.4 1.6 1.8 2 time (h) c o n c e n tr a ti o n g /l actual data propose model lagergren model 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.8 1 1.2 1.4 1.6 1.8 2 time (h) c o n c e n tr a ti o n g /l actual data propose model lagergren model mathematical model describes treatment of waste water using modified activated carbon   102   figure 6. effect of different rates of adsorption and desorption on pollutant concentration of benzene.   figure 7. effect of different rates of adsorption and desorption on pollutant concentration of nickel. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 8.5 9 9.5 10 time (h) c o n c e n tr a ti o n g /l k(+ad)k(s)-k(-ad)=0.24 k(+ad)k(s)-k(-ad)=0.15 k(+ad)k(s)-k(-ad)=0.1 k(+ad)k(s)-k(-ad)=0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 time (h) c o n c e n tr a ti o n g /l k(+ad)k(s)-k(-ad)=0.18 k(+ad)k(s)-k(-ad)=0.15 k(+ad)k(s)-k(-ad)=0.1 k(+ad)k(s)-k(-ad)=0.05 a.s. ibrahim, a.s. al buloshi, s.s. al zaabi and l.a. al yafai   101   figure 8. effect of different rates of adsorption and desorption on pollutant concentration of cadmium.   figure 9. effect of rates of adsorption and desorption on pollutant concentration of copper. 4. conclusion this proposed mathematical model has enough ability to evaluate dynamic adsorption and desorption for modified activate carbon to give clear view about mechanism of the system and very acceptable results due to inclusion of all 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 time (h) c o n c e n tr a ti o n g /l k(+ad)k(s)-k(-ad)=0.18 k(+ad)k(s)-k(-ad)=0.15 k(+ad)k(s)-k(-ad)=0.1 k(+ad)k(s)-k(-ad)=0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.8 1 1.2 1.4 1.6 1.8 2 time (h) c o n c e n tr a ti o n g /l k(+ad)k(s)-k(-ad)=0.18 k(+ad)k(s)-k(-ad)=0.15 k(+ad)k(s)-k(-ad)=0.1 k(+ad)k(s)-k(-ad)=0.05 103 mathematical model describes treatment of waste water using modified activated carbon   101   the rates types for adsorption and desorption compare to the other models. references afsaneh s, mohammad r, hosein d, amin g, mohammad h (2008), adsorption of carbon dioxide using impregnated modified activate carbontivated carbon promoted by zinc. international journal of greenhouse gas control 3: 249-254. ajmal m, rak r, jamal a, ahmad r (2003), adsorption studies on rice husk: removal and recovery of cd (ii) from wastewater. bioresource technology 86: 147–149. bog g, liping c, kechang x (2006), adsorption of carbon dioxide on modified activate carbontivated carbon. journal of natural gas chemistry 15: 223-229. lin n. bai r (2005), copper absorption on chitosan-cellulose hydrogel beads: hydro gelbeads: behaviors and mechanisms. separation and purification technology 42: 237–247. lalhruaitluanga h, jayaram k, prasad m, kumar k (2010), lead (ii) adsorption from aqueous solutions by raw and modified activate carbontivated charcoals of melocanna modified activate carboncifera roxburgh (bamboo)-a comparative study. journal of hazardous materials 175: 311-318. khaled a, nemr a, sikaily a, abdelwahab o (2009), treatment of artificial textile dye effluent containing direct yellow 12 by orange peel carbon. desalination 238: 210232. mohamed k, wan m, chun y, donni a (2008), adsorption cap modified activate carbonites of carbon dioxide, oxygen, nitrogen and methane on carbon molecular basket derived from polyethyleneimine impregnation on microporous palm shell modified activate carbontivated carbon. separation and purification technology 62: 609–613. muhammad r, raziya n, muhammad a, tariq m, khalil r (2008), pb (ii) biosorption from hazardous aqueous streams using gossypium hirsutum (cotton) waste biomass. journal of hazardous materials 161: 88-94. nakano y, takeshita k, tsutsumi t (2001), adsorption mechanim of hevalent chromium by redox within condensedtannin gel, water research 35 :496-500. okieimen c, ogbeide s (2009), the dependence on temperature of carbonization and chemical modified activate carbontivation on charmodified activate carbonteristics of granular modified activate carbontivated carbon charmodified activate carbonteristics. advanced materials research 62-64 : 398-403. prasetyo i, dod d (1998), adsorption rate of methane and carbon dioxide on modified activate carbontivated carbon by the semibatch constant molar flow rate method. chemical engineering science 53(19): 3459-3467. vazquez g, gonzalez a, freire s, lopez, antorrena g (2002), removal of cadmium and mercury ions from aqueous solution by sorption on treated pinus pinaster bark: kinetics and isotherms. bioresource technology 822: 47-51. 104 ece050919.qxd the journal of engineering research vol. 4, no.1 (2007) 64-68 1. introduction the world prevalence of dm is projected to increase from 135 million in 1995 to 300 million in 2025 (king, h., aubert, r.e. and herman, w.h., 1988). cardiovascular complications are the most common causes of morbidity and mortality in diabetic patients (amos, a.f., mccarty, d.j. and zimmet, p., 1997). diabetic cardiomyopathy is associated with a variety of myocardial dysfunctions including impairment of contractile force generation, abnormal ventricular filling and delayed relaxation of cardiac muscle and these defects can occur in the absence of vascular complications (dhalla, n.s., pierce, g.n., innes, i.r. and beamish, r.e., 1985; howarth, f.c., qureshi, a., lawrence, p. and adeghate, e., 2000; howarth, f.c., qureshi, m.a., bracken, n.k., winlow, w. and singh, j., 2001; howarth, f.c. qureshi, m.a. and white, e., 2002). _______________________________________ *corresponding author’s e-mail: mjacob@ieee.org early detection of dm, prior to clinical events, coupled with implementation of preventive management strategies, can delay the progression of the disease. in fact, proper diet, physical exercise, cessation of smoking, control of blood pressure, and reduced cholesterol lower a patient's odds of myocardial infarction by 19 times (wood, d., 1998; vinik, a. and flammer, m., 2002). the hr is controlled by the sympathetic and parasympathetic branches of the autonomic nervous system (ans). the parasympathetic vagal nerves slow the hr through the release of acetylcholine while the sympathetic nerves accelerate the hr through the release of noradrenaline. the variability of the hr is the result of a balance between the sympathetic and parasympathetic nerves (bootsma, m., swenne, c.a., van bolhuis, h.h., chang, p.c., cats, v.m. and bruschke, a.v., 1994). fluctuations in hr are known to be quasi-periodical, repeating with a period, which is not strictly constant, showing changes in morphology, amplitude and phase from one beat to the next. in steady state conditions, these heart rate variability time analyses of the streptozotocindiabetic rat m. jacobson*1, f.c. howarth2, e. adeghate2, k. fatima-shad3 *1 college of engineering, uae university, al-ain, p.o. box 17666, u.a.e. 2 faculty of medicine, uae university, al-ain, p.o. box 17666, u.a.e. 3 faculty of science, uae university, al-ain, p.o. box: 17666, u.a.e. received 19 september 2005; accepted 10 may 2006 abstract: as the world prevalence of diabetes mellitus (dm) increases, animal models of the disease's progression are required for researching effective treatment. the streptozotocin (stz) treated rat is known to cause hyperglycaemia. this study confirms that this animal model also displays dm physiological effects in the animal heart rate (hr) and heart rate variability (hrv). in particular, 5 minutes of rat (n=13) electrocardiogram (ecg) is acquired hourly for 30 days. at day 10, the animal (n=7) is dosed with stz and the ecg is analyzed in order to determine the hr and hrv. the hrv is indexed using two time-based analyses, based on long-term (24hr) and short-term (5min) analyses. all analyses are compared to control non-stz dosed animals (n=6) and display significant dm effects. keywords: heart tate variability, time-domain analysis, streptozotocin-diabetic …ôµ°ùddg úcƒjhrƒàjî°s áyôl i飩ÿg ¿gôäø∏d ö∏≤dg äé° ñf ∫~©e ‘ ägò¨à∏d »æeõdg π«∏ëàdg øø°°ùhhƒƒccééll ..ωω1@çqqggƒƒgg,,ìì..±±,,2ââ««ll~~jjgg ,,…… ,,2ooéé°°tt á᪪wwééaa ,,∑∑ ,,3 áá°°uuóóÿÿggáyôl √飩ÿg ¿gôäødg ¿g ±hô©ÿg øe .∫é©ødg êó©dg øy éãëh ¢vôÿg qƒ£j á©héàÿ á«fgƒ«m êpé‰ ¤g álé◊g äogorg éª∏c ⁄é©dg ‘ …ôµ°ùdg ¢vôe qé°ûàfg ogorg éª∏c : .∫~©ÿg gòg ò¨jh ö∏≤dg äé° ñf ∫~©e ‘ á«lƒ∏«°ùa ägò¨j é¡«∏y ô¡¶à°s ¿gôäødg √òg ¿g é° jg âñãj á°sgq~dg √òg .ω~dg ‘ ôµ°ùdg ioéjr •ôa øe êé©j úchrƒàjî°ùdg øe øe áyôl aé£yg ” ,ô°té©dg ωƒ«dg ‘ .éeƒj úkók i~ÿh áyé°s πµd ≥fébo ¢ùªn i~y é¡d ö∏≤dg §£fl iqé°tg π«é°ùj ” å«m ,(13) o~y ¿gôäa ≈∏y áhôéàdg agôlg ” ≈∏y äé° ñædg ∫g~©e ‘ ägò¨àdg á°sgqo â“h .ö∏≤dg äé° ñf ò¨j ∫~©eh äé° ñædg ∫~©e ≈∏y ∫ƒ°üë∏d é¡d ö∏≤dg §£fl iqé°tg π«∏– ”h ¿gôäødg øe (7) ∫ úchrƒàjî°ùdg òkéj á¶móe ”h áyô÷g √ò¡d √飩ÿg òz ¿gôäø∏d áà°ùdg ä’é◊g ™e âfqƒb âjôlg »àdg π«déëàdg πc .(≥fébo ¢ùªn πc) ò°ü≤dg i~ÿg ≈∏yh (áyé°s24 πc) πjƒ£dg i~ÿg . ∂dp gáá««mmééààøøÿÿgg ääggooôôøøÿÿ .…ôµ°ùdg ,úchrƒàjî°ùdg ,øeõdg ∫é› ‘ π«∏ëàdg ,ö∏≤dg äé° ñf ∫~©e ò¨j : 65 the journal of engineering research vol. 4, no.1 (2007) 64-68 oscillations are maintained around a certain mean value. during periods of rest, the parasympathetic response dominates and the hr lowers. during activity, the sympathetic response dominates and the hr increases (ori, z., et al. 1992). change in the hr is termed the hrv. since the hr is controlled primarily by the ans, assessment of the hrv provides important information about sympathetic and parasympathetic cardiovascular control (cerutti, s., bianchi, a.m. and mainnardi, l.t., 1995). in case of dysfunction, compensatory changes occur in the ans control system in order to keep the entire circulatory system within normal operating limits. in particular, the presence of dm is known to reduce hrv and disrupt the sympathetic and parasympathetic nervous systems in humans (bootsma, m., swenne, c.a., van bolhuis, h.h., chang, p.c., cats, v.m. and bruschke, a.v., 1994; karkuszewski, l., bissinger, a., rosaik, m., kuberska-kedzierska, m. and rogulski, b., 2005; al-hazimi, a., al-ama, n., syiamic, a., qosti, r. and abdul-galil, k., 2002). in order to develop clinical screening procedures, appropriate animal models are required to develop the methodology and research possible treatment. the stz dosed rat displays hyperglycaemia and this study confirms dm physiological effects on hrv comparable to humans. 2. method 2.1 animal preparations the ecg was monitored in 13 young, age-matched, adult male wistar rats using a telemetry system (data sciences int., usa). the system was comprised of transmitter devices, associated receivers, a data exchange matrix and personal computer (pc) for system configuration, control, acquisition, and storage (fig 1). the transmitters were surgically implanted under general anesthesia and inserted into the peritoneal cavity with the electrodes arranged in the einthoven bipolar lead ii configuration (right foreleg and left hind leg). one week after surgery, transmitters in the animals were activated via magnetic switches and baseline data was recorded for 10 days before induction of dm by administration of the drug streptozotocin in 7 of the 13 rats. streptozotocin causes selective pancreatic-cell necrosis, which results in hypoinsulinaemia, hyperglycaemia and reduced body weight gain (howarth, f.c., qureshi, a., lawrence, p. and adeghate, e., 2000; howarth, f.c., qureshi, m.a., bracken, n.k., winlow, w. and singh, j., 2001; howarth, f.c. qureshi, m.a. and white, e., 2002; choi, k.m., zhong, y., hoit, b.d., grupp, i.l., hahn, h., dilly, k.w., guatimosin, s., lederer, w.j. and matlib, m.a., 2002; okayama, h., hamada, m. and hiwada, k., 1994; ren, j. and davidoff, a.j., 1997). ethical approval for the project was obtained from the faculty of medicine & health sciences ethics committee for animal research of the united arab emirates university. 2.2 physiological data acquisition per the guidelines of the european society of cardiology and the north american society of pacing and electrophysiology (task force of european society of cardiology and north american society of pacing and electrophysiology, 1996), the ecg was sampled at 1khz, for 5 minutes each hour over one month. since the average rat hr is about 300 beats per minute (bpm), a 5minute sample ensures measurement of approximately 1500 contractions. from the ecg signal, the r-waves of the qrs complexes were determined using a secant detector and the beat-to-beat hr was computed. ectopic beats, such as pre-ventricular contractions (pvc) and skipped beats, were visually removed. the resulting signal was then used for computation of the 5 minute average hr and hrv analyses. 2.3 hrv analysis (1) specifically, the sdann is computed by measuring the ecg and determining the 5-minute average hr each hour for 24 hours. then, the standard deviation of 24 successive samples is determined. thus, the measurement is spans circadian cycle and indexes the range of hr over maximum and minimum activity. the short-term hrv analysis determined only from the 5-minute ecg. specifically, the standard deviation of the 5-minute, normal-to-normal (sdnn) hr is computer per (2), where hri is the beat-to-beat hr, n is the number of beats, and is the 5-minute average heart rate (kleiger, r.e., stein, p.k., bosner, m.s. and jeffrey, n.r., 1995). (2) figure 1. telemetry system to collect diabetic and control ecg the 24-hour, long-term hrv analysis was computed as the standard deviation of the mean 5-minute norma lto-normal hr, which is termed the sdann per (1), where hr is the 5 -minute average heart rate and 24hr is the 24-hour average heart rate (kleiger, r.e., stein, p.k., bosner, m.s. and jeffrey, n. r., 1995). this analysis highlights changes in the hr range o ver one circadian cycle of 24 hours. ∑ = −= 24 1i 2 24 2 i hrhr 24 1 sdann ∑ = −= n 1i 22 i hrhrn 1 sdnn dsi 66 the journal of engineering research vol. 4, no.1 (2007) 64-68 thus, the sdnn indexes changes in the hr over a 5minute period. the activity and circadian cycle of the animal are not considered. thus, though both the sdann and sdnn both give measurements of the hrv, one indexes 24-hourr hr changes while the other focuses on beat-to-beat changes. in other words, though both are derived from the ecg, each measures a different type of hrv: 24-hour long-term and 5-minute short-term hrv. 3. results the blood glucose of the 13 rats was measured at days 3 and 20, per fig. 2. the stz-treated rats displayed elevated blood glucose confirming the presence of dm while the control rats displayed normal blood glucose levels. the 5-minute average hr is plotted in fig. 3 for the stz and control groups. stz treatment occurred on day 0 and the stz group (n=6) displayed a drop in hr from day 0 to day 5, until a new steady-state hr is achieved. at day 20, there is an approximate 80bpm significant difference between the stz and control groups. significant testing is accomplished utilizing the t-test with p < 0.001 (considered significant). the standard deviation of 24 successive, 5-minute hr averages is used to determine the sdann, per the previous discussion. figure 4 displays the results of the stz and control groups. after treatment, the stz group sdann results are immediately unclear but are obviously declining at day 5, while the control group sdann increased. the difference between the groups continues increasing and is significant at day 20. per the previous discussion, the standard deviation of the 5-minute, normal-to-normal beats in the ecg is used to computed the sdnn. this short-term hrv analysis is displayed in fig. 5 for the stz and control groups. after treatment, the sdnn of the stz group immediately became decreasing to a new steady state value at about day 8. this new value is about 10 bpm less than pre-stz treatment. the control group sdnn values, though varied, continue much the same throughout the 30 day experiment. in order to better compare between the two time-based hrv algorithms, a 24-hour moving average is applied to the sdnn. therefore, the time basis of each hrv index spans 24 hours. furthermore, in order to compare between the relative changes due to stz treatment, each is normalized and plotted in fig. 6. observably, each hrv index gives a similar deflection due to stz treatment. the specific amplitude changes over the 30-day experiment are recorded in table 1 along with the associated standard error. note that each hrv measurement gives a similar normalized magnitude and error over the experiment period. statistically, the sdann and sdnn have a 0.96 cor 0 50 100 150 200 250 300 350 400 3 days post stz 20 days post stz b lo o d g lu co se ( m g d l1 ) control stz figure 2. stz and control group blood glucose levels 200 220 240 260 280 300 320 340 360 380 400 -10 -5 0 5 10 15 20 days h ea rt r at e (b ea ts p er m in ) control stz figure 3. comparison in average 5 minute hr between stz and control groups 10 15 20 25 30 35 40 45 -10 -5 0 5 10 15 20 day h ea rt ra te v ar ia bi lit y (b ea ts p er m in ut e) control stz figure 4. comparison in hrv sdann between stz and control groups 0 10 20 30 40 50 60 -10 -5 0 5 10 15 20 day h ea rt ra te v ar ia bi lit y (b ea ts p er m in ut e) control stz figure 5. comparison in sdnn hrv between stz and control groups 67 the journal of engineering research vol. 4, no.1 (2007) 64-68 relation coefficient indicating that the two hrv measurements are highly correlated and give a similar index in response to stz treatment. 4. discussion the sdnn and sdann analyses of the hrv both display similar changes in response to the stz treatment. specifically, after stz treatment, the hrv begins decreasing to a new steady-state value over a 10 day period. reduction in hrv is indicative of autonomic neuropathy, specifically the control of the hr. autonomic neuropathy in known to occur in the presence of dm, and so measurement of the hrv may be used in the screening of dm. sdann and sdnn are two types of time-based hrv analyses. though the sdann is based on long-term (24 hours) measurement and the sdnn is based on shortterm (5-minutes) measurement, each analysis, when normalized and adjusted to the same time scale is found to give a similar hrv deflection in the presence of dm with a similar measurement error. considering this result in clinical setting, it is likely that short-term hrv measurement, utilizing the sdnn, is sufficient for the screening of dm. 5. conclusions the sdnn and sdann analyses of the hrv both display changes in the animal's physiology due to stz administration, and thus, dm, as has been shown in humans. both hrv indices are highly correlated and both display a similar reduction in the hrv in comparison to control groups, thus confirming the dm physiological effects of the stz treated rat model. acknowledgments this project was supported by interdisciplinary grant (01-05-8-12/03) from the united arab emirates university. references al-hazimi, a., al-ama, n., syiamic, a., qosti, r. and abdel-galil, k., 2002, “time-domain analysis of heart tate variability in diabetic patients with and without autonomic neuropathy, annals of saudi medicine,” vol. 22, pp. 5-6. amos, a.f., mccarty, d.j. and zimmet, p., 1997, “the rising global burden of diabetes and its complications: estimates and projections to the year 2010,” diabetic medicine, 14: s1-s85. bootsma, m., swenne, c. a., van bolhuis, h. h., chang, p. c., cats, v. m. and bruschke, a. v., 1994, “heart rate and heart rate variability as indexes of sympathovagal balance,” am. j. physiol 266, h1565-h1571. cerutti, s., bianchi, a. m. and mainardi, l. t., 1995, “spectral analysis of the heart rate variability signal,” heart rate variability, futura publishing company. choi, k.m., zhong, y., hoit, b.d., grupp, i.l., hahn, h., dilly, k.w., guatimosin, s., lederer, w.j. and matlib, m.a., 2002, “defective intracellular ca2+ signaling contributes to cardiomyopathy in type 1 diabetic rats,” am j physiol, 283: h1398-h1408. dhalla, n.s., pierce, g.n., innes, i.r. and beamish, r.e., 1985, “pathogenesis of cardiac dysfunction in diabetes mellitus,” can j cardiol, vol. 1, pp. 263281. howarth, f.c., qureshi a, lawrence p. and adeghate e., 2000, “chronic effects of streptozotocin-induced diabetes on the ultrastructure of rat ventricular and papillary muscle,” acta diabetol, vol. 37, pp. 119124. howarth, f.c., qureshi, m.a. and white, e., 2002, “effects of hyperosmotic shrinking on ventricular myocyte contraction and intracellular ca2+ in streptozotocin-induced diabetic rats,” pflügers archiv, vol. 444, pp. 446-451. howarth, f.c., qureshi, m.a., bracken, n.k., winlow, w. and singh, j., 2001, “time-dependent effects of streptozotocin-induced diabetes on contraction of ventricular myocytes from rat heart,” emirates medical journal, vol. 19, pp. 35-41. king, h., aubert, r.e. and herman, w.h., 1998, “global burden of diabetes, 1995-2025: prevalence, numerical estimates, and projections,” diabetes care, vol. 21, pp. 1414-1431. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% -10 -5 0 5 10 15 20 day n or m al is ed sdnn sdann figure 6. comparison between the 24-hour average sdnn and sdann for the stz group day sdnn sdann -5 85 ±2% 76 ± 3% 0 87 ± 2% 85 ± 3% 5 33 ± 3% 25 ± 3% 10 11 ± 4% 10 ± 5% 15 14 ± 2% 22 ± 2% 20 11 ± 2% 9 ± 5% table 1. normalized 24-hour average values and error for the sdnn and sdann hrv indices 68 the journal of engineering research vol. 4, no.1 (2007) 64-68 kleiger, r. e., stein, p. k., bosner, m. s. and jeffrey, n. r., 1995, “timed-doman measurements of heart rate variability,” heart rate variability, futura publishing company. markuszewski, l., bissinger, a., rosiak, m., kuberskakedzierska, m. and rogulski, b., 2005, “heart rate turbulence and heart rate variability in patients with diabetes mellitus,” folia cardiologica, 12, supl. c; p. 181. okayama, h., hamada, m. and hiwada, k., 1994, “contractile dysfunction in the diabetic-rat heart is an intrinsic abnormality of the cardiac myocyte,” clinical science, vol. 86, pp. 257-262. ori, z., et al. 1992, “heart rate variability,” frequency domain analysis, cardiology clinics, vol. 10(3), pp. 499-537. ren, j. and davidoff, a.j., 1997. “diabetes rapidly induces contractile dysfunctions in isolated ventricular myocytes,” am j physiol, vol. 272, pp. h148-h158. task force of the european society of cardiology and the north american society of pacing and electrophysiology., 1996, “heart rate variability, standards of measurement, physiological interpretations, and clinical use,” european heart journal, vol. 17, p. 354. vinik, a. and flemmer, m., 2002, “diabetes and macrovascular disease,” j. of diabetes and its complications, vol. 16, pp. 235-245. wood, d., 1998, “prevention of coronary heart disease in clinical practice,” recommendations of the second joint task force of european and other societies on coronary prevention, arteriosclerosis, vol. 140, pp. 199-270. microsoft word paper 1.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 103-114 the simulation study of a restructured residential low-voltage distribution system microgrid   l. ramesha,* and umamageswarib   a,* dr. m.g.r educational and research institute university, chennai, india. b sustainable power and energy research centre, dr. m.g.r educational and research institute chennai, india.    received 19 february 2015; accepted 11 october 2015 abstract: a primary and necessary focus in creating a greener environment is the conversion of existing power-generation sources to renewable power sources in the near future. another important focus is to develop sustainable household power generation to a low-voltage electricity grid with a power purchase and selling facility. to help with achieving the above vision, the objective of this work is to critically analyze the existing low-voltage distribution system and make suggestions for restructuring it to the low-voltage interconnected microgrid (mg). the test was carried out in the tamil nadu electricity board (tneb) 100kva transformer feeder which was connected to supply around 100 houses with electricity. the performance analysis of the proposed system was examined through different case studies, represented as a normal operating condition of the existing distribution system and a reconstructed and interconnected mg to the tneb grid. the project was designed and analyzed using pscad software. the results discussed in the project are helpful in examining the effects of multiple distributed energy resources on distributed generation. in future, knowledge of these effects may be helpful for rural area electrification. keywords: distribution systems, renewable energy, wind, solar, microgrid. אאא אk،،   wאkאאאאאאא אאאאאאאk   א א א       א א  א  ، א  אאאkאאאא ١٠٠אאאkאאאאא אא،אאא،،אא  א   k   א  א pscad k א   א א אאאאkאkאא אא  :אאא،א،אא،k * corresponding author’s e-mail: lramesh@theiet.org l. ramesh and umamageswari   104 1. introduction nowadays conventional power systems face numerous problems including depletion of fuel resources, poor energy efficiency, and pollution. thus, a new trend in generating power is introduced by using renewable, nonconventional energy sources including biogas, photovoltaic (pv) cells, combined heat and power (chp) systems, micro turbines, and wind turbines. this type of power generation is known as distributed generation (dg) and the energy sources are distributed energy sources. microgrids (mg) are generally attractive to consumers, and a great number of them will be installed at consumers’ sites. the cluster of loads and microsources consider mgs a viable source of both power and heat for local communities. this concept also provides a new model for defining the operation of dg (brabandere et al. 2007; carrasco et al. 2006; cheng et al. 2009). a mg operates on small units of less than 100 kw with static electronic interfaces and low costs and low voltage but high reliability in a system that is easily placed on customer sites. the effectively designed power electronics and controllers insure that the mg can meet the needs of its customers and utilities. the best backup power for consumer sites is the installation of a mg and the interconnection of dg with an electricity grid that consists of electricity generation, energy storage and power management systems. the mg contains a combination of technologies, which provide supply-and-demand-style management and can be located near the energy center. mgs also enable a more efficient use of waste heat in chp applications, which boosts efficiency and lowers emissions. the chp systems provide electricity, heat for industrial processes, hot water, refrigeration, space heating and cooling and humidity control to improve indoor air quality and comfort. over the last few years, however, a number of influences have combined to lead to an increased interest in mgs’ schemes. a mg, when operated as an autonomous systems, can cause several technical problems in its operation and control. such technical issues presented by the present researchers are discussed here. the load sharing is possible in multiple dg units when operated in islanded mode. the frequency and voltage can be controlled through droop control to achieve power sharing in a decentralized manner (corradini et al. 2010; fornari et al. 2013; guerrero et al. 2005; guerrero et al. 2011; gungor et al. 2011; he and li 2011; mohamed et al. 2009; ramanarayanan et al. 2010). a number of methods have been proposed to solve mgs’ power control issues. the new framework for the frequency-voltage and real and reactive power concept were proposed (li et al. 2011; mehrizi-sani and iravani 2010) to improve the stability of a mg system. these methods cannot suppress the reactive power of sharing errors that can be experienced when small synchronous generators are incorporated into the mg. the proper power sharing between inverter-based dg units and electric machine-based dg units are more challenging problems to be addressed. harmonic current injection methods (melicio et al. 2008) were used to reduce reactive and harmonic power sharing errors. ji et al. (2014) presented a case study of a mg arranged between two commercial/residential buildings in order to overcome regulatory barriers. in the study, a load management strategy aimed at controlling power withdrawal at the point of delivery was developed. pogaku et al. (2007) proposed a mg smart grid and presented a review of the optimal distribution models of mg in a smart grid environment, pointing out issues in the existing models. the optimal load distribution model of mg both in objective functions and constraints was established and discussed with effective output. zhou et al. (2014) presented an optimization problem for residential mg energy management with combined cooling, heating and electricity technology. it was executed in two-stages to find the optimal installed capacity investment and operation control of chp and cooling. the proposed approach was tested with residential area energy management in northern china. computational results were gathered under different robustness levels and analyzed in order to provide investment decisions and operations strategies. the aim of this work is to critically analyze part of an existing low-voltage distribution system and its restructured low-voltage interconnected mg. the study was carried out with the tamil nadu electricity board (tneb) 100kva transformer feeder which delivers power to 100 houses. in the following sections, this work models renewable sources, presents distribution feeders, the restructuring of residential mgs and interconnected mgs. performance analyses were carried out for the the simulation study of a restructured residential low-voltage distribution system microgrid   105 figure 1. the pscad-generated layout of the tamil nadu electricity board (tneb) distribution feeder. proposed mg and are ready for implementation in chennai. 2. modeling of test system the work developed in this project was used to understand and model the electrical effects of multiple distribution resources on a distribution system. the analyzed test system was taken from the tneb distribution system. the existing system is connected with a 100 kva transformer rating, which supplies power to the houses connected to this feeder. section 2 will discuss modeling the test distribution feeder, house distribution system, and wind and solar sources. 2.1 modeling the tneb low-voltage distribution system initial studies on this feeder were done in july 2013 at rampoornanagar, which is in athipet village. the test system consists of three feeders running between the 100kva transformer and 40 houses, including single, double, triple and bungalow type houses. the layout, generated with pscad software (electrotek concepts, inc., beverly, massachusetts, usa) and data drawn from the tneb power engineers’ handbook (2010), is modeled in fig. 1. 2.2 modeling the home distribution system the distribution layout of the houses was reconstructed according to a limited number of nodes in order to reduce complexity without affecting the respective loads. the breaker box was considered the bus and all loads were chosen based on resistive, inductive, or capacitive capabilities. the lumped load component in the pscad was considered in this analysis so as to change the r, l, and c components for the load. the sample model of the pscad layout is shown in figs. 2–3. the selected houses were also supplied with renewable resources, in this case wind and solar sources which are ideal in chennai’s climate. modeling wind and solar sources with pscad is described in the section below. 2.3 modeling renewable sources wind turbines convert wind energy to electrical energy for distribution. aerodynamic modeling helps determine the optimum tower height, control systems, number of blades, and blade shape. wind turbines are generally divided depending on their axis of rotation (ie. horizontal or vertical axis), with the horizontal axis type always providing better performance. of the wind turbine costs, 20% includes the blade for converting wind energy to low speed l. ramesh and umamageswari 106 figure 2. single-bedroom house layout. figure 3. double bedroom house layout. figure 4. pscad wind turbine model. the simulation study of a restructured residential low-voltage distribution system microgrid   107   figure 5. a simulated solar model where anti-islanding schemes were not implemented.   figure 6. layout of reconstructed mg system into four sub-systems. rotational energy, 45% includes the electrical generator and 12% includes the control electronics and gearbox. the remaining 23% of the wind turbine cost includes the tower, land, foundation, road and installation. the pscad models of the wind turbine generator connections are shown in fig. 4. a photovoltaic (pv) cell is an array that carries solar voltaic material that changes solar energy into electrical energy. it includes multiple mechanical and electrical engineering components. the materials used in pvs include mono-, poly-, and micro-crystalline silicon, cadmium telluride, and copper indium selenide. the dc power output at the pv device terminals can be directly used to switch on small loads like lighting systems or small dc motors. maximum power from a pv array is possible through maximum power point tracking technology. to integrate the pv cell into an alternating current (ac) system, the inverter model is connected at the output of the cell, and the dc source for the currentcontrolled inverter is fed from the pv source. figure 5 shows a simulation model with the inverter model when anti-islanding schemes were not implemented. 2.4 modeling mgs the restructuring of the present system was formulated based on a performance analysis of the system and the present customers’ willingness to embrace renewable power. house numbers 2, 4, 8, 10, 14, 20, 24, 26, 28, 32, and 34 used renewable power through solar and wind. after the survey study with existing customers, it was decided to convert the present system into four sub-systems. assumptions were made that 50% of the total maximum loads would go for transition to the renewable power. in reference to above, the optimal planning and location of renewable sources were identified and placed for simulation. figure 6 shows pscad mg layout. l. ramesh and umamageswari   108   figure 7. voltage characteristics of the existing distribution system showing that the tail end customer receives lower voltage than the customers at the beginning of the system. figure 8. house seven’s power characteristics. 3. simulated results after designing the model, it was simulated with various parameters and the grid capability was assessed. in the designed model, there were different types of houses which were also being supplied with power from renewable resources like wind and solar. the impact of the designed the simulation study of a restructured residential low-voltage distribution system microgrid 109   model with renewable resources is investigated in the following case studies. three different case studies were carried out to analyze the performance of the designed model. the cases are as below:  normal operating conditions of the existing distribution system,  the present system reconstructed into a mg system, and,  an interconnected mg sub-system. 3.1. normal operating condition of the existing distribution system. to evaluate the normal operating condition of the existing distribution system, the present system with a normal operating load was considered for analysis. in this condition, the houses are simply supplied with power from a grid supply. the existing distribution system was designed and evaluated to verify the performance of the system. the representation of bus voltage with respect to the bus number (fig. 7) shows that the tail end customer receives lower voltage than the initial customers. from fig. 7, it can be concluded that there is a gradual decrease in voltage at the tail end of the system, implying that the system performance also gradually reduces. this implementation for renewable resources is that such a system will improve performance. the graphic results for houses seven and 38 show a gradual decrease in voltage which will impact the quality of power (figs. 8 and 9). the above graph shows the real reactive power characteristics of house seven. from the graph, it is clear that the performance is reduced when compared with the transformer characteristics, and especially the power curves. the above graph shows the real reactive power characteristics of house 38. from the graph, it is clear that the performance is reduced when compared with the transformer characteristics. the vi-characteristics and the power curves become more oscillatory as a result of the reduction in the voltage profile to 25%. 3.2. the present system restructured into a mg system the present system restructured into four sub-systems connected through renewable power generation is called a residential mg. each subsystem operates with its own renewable power sources connected through the different houses. the voltage output of the subsystems with respect to bus number is represented in figs. 10–13. figure 9. house 38’s characteristics. l. ramesh and umamageswari   110     figure 10. voltage characteristics of subsystem one. figure 11. voltage characteristics of sub-system two. figure 12. voltage characteristics of sub-system three. figure 13. voltage characteristics of sub-system four. the simulation study of a restructured residential low-voltage distribution system microgrid   111 from the above graphs, it can be observed that the proposed system will be able to deliver power under normal load conditions. each graph indicates an increase and decrease in voltage between 390 and 440 volts, indicating that houses with renewable power have good voltage profiles. the pscad output of the sample house is expressed in figs. 14 and 15. from the above results, it can be observed that the voltage performance is maintained at 440 under a sudden change in load. the limitation of the above system is that only 20% of the houses had voltage profiles of 400 volts. to improve the voltage performance to 440 volts at all the buses, the sub-systems should be interconnected with the main grid, as is discussed in next section. 3.3 an interconnected mg sub-system in the case of an interconnected mg subsystem, the interconnections of all of the mgs are connected with the help of a breaker. by making this connection, the power among each sub-system mg can be managed; thus, in case of power failure, they can receive power through another dg. the interconnected mg is shown in fig. 16. figure 14. house three’s voltage characteristics. figure 15. house 36’s voltage characteristics. time, s v o lt ag e, v time, s v o lt ag e, v l. ramesh and umamageswari   112   figure 16. interconnected mg layout. the above restructured mg layout was executed in pscad in different cases, including in the case of opening/closing the breaker, thereby checking the performance of the system. by doing this, the supply can be shared throughout the network. in such a case, even if there is a failure in any part of the network, the entire system will be supplied. the comparison of the existing and proposed system is shown in fig. 17. it can be observed that an interconnected mg will provide better performance when compared with independent systems. the mg will provide stable, good quality power throughout the system, all the way to the tail end. figure 17. comparison of the existing and proposed mg. the simulation study of a restructured residential low-voltage distribution system microgrid   113 3.4 discussion the existing and proposed feeders were simulated and analyses of bus voltage in all the cases are shown in previous figures. the existing bus voltage varies from 370–420 v, with a pscad voltage regulation of 16%. the practical voltage regulation was measured at 23%. the bus voltage of the restructured mgs varies from 400–440 v, with sustainable power throughout the system. the voltage in bus numbers 2, 4, 8, 10, 14, 20, 24, 26, 28, 32 and 34 was maintained at 440 v, clearly indicating that the dgs are connected at these buses in the houses. the case study analysis and the tail end 5% voltage regulation show that the proposed system is of a high accuracy. due to space limitations, results of the case study on a sudden drop or increase in load and various fault conditions are not reported. however, the case study shows that the system would be able to handle such a situation, resulting in better voltage on the bus. the proposed system was recommended to the customers for implementation in april 2015. 4. conclusion the present necessity of energy saving through a restructured mg is discussed in this paper. the local generation of renewable power and the interconnection between mgs may be a green solution for the issue of an energy shortage. the team, after performing a local survey with the customers, identified a distribution feeder for conversion to four local grids (mg1, mg2, mg3 and mg4). the proposed isolated and interconnected mgs were tested through case studies. the output predicted that the proposed system will work well even through interrupted faults. the customers agreed to implement the restructured mg feeders for effective operation and energy savings. the cost analysis for implementation will be reported in future work. references brabandere b, bolsens jvd, keybus a, woyte j, drisen, belmans r (2007), a voltage and frequency droop control method for parallel inverters. ieee transactions on power electronics 22(4): 1107–1115. carrasco lg, franquelo it, bialasiewicz e, galvan rcp, guisado am, prats, leon j, moreno-alfonso n (2006), power-electronic systems for the grid integration of renewable energy sources: a survey. ieee transactions on industrial electronics 53: 1002–1016. cheng pt, chen ca, lee tl, kuo sy (2009), a cooperative imbalance compensation method for distributed generation interface converters. ieee transactions on industrial applications 45(2): 805–815. corradini l, mattavelli p, corradin m, polo f (2010), analysis of parallel operation of uninterruptible power supplies though long wiring cables. ieee transactions on power electronics 25(4): 2806–2816. fornari s, di giorgio a, liberati f (2013), a case study of a commercial/residential mg integrating cogeneration and electrical local users. 12th international conference environment and electrical engineering (eeeic) 363–368. guerrero jm, de vicuña lg, matas j, castilla m, miret j (2005), output impedance design of parallel-connected ups inverters with wireless load sharing control. ieee transactions on industrial electronics 52(4): 1126–1135. guerrero jm, vasque jc, matas j, de vicuña lg, castilla m (2011), hierarchical control of droop-controlled ac and dc mgs—a general approach toward standardization. ieee transactions on industrial electronics 55(1): 158–172. gungor vc, sahin d, kocak t, ergut s, buccella c, cecati c, hancke gp (2011), smart grid technologies: communications technologies and standards. ieee trans. industrial information 7(4): 529–539. he j, li yw (2011), an accurate reactive power sharing control strategy for dg units in a mg. in proceedings from the 8th international conference on power electronics and ecce asia, jeju, korea 551–556. ji l, niu dx, huang gh (2014), an inexact two-stage stochastic robust programming for residential micro-grid management-based on random demand. elsevier energy 67(1): 186– 199. li w, li yw (2011), power management of inverter interfaced autonomous mg based on virtual frequency-voltage frame. ieee transactions on smart grid, 2(1): 30–40. mehrizi-sani a, iravani r (2010), potentialfunction based control of a mg in islanded and grid-connected modes. ieee transactions on power systems 25(4): 1883–1891. l. ramesh and umamageswari   114   melicio r, mendes vmf, catalão jps (2008), wind energy systems and power quality: matrix versus two level converters. in proceedings of the international conference on renewable energy and power quality, santander, spain. mohamed f (2006), mg modeling and simulation. thesis report helsinki university of technology control engineering laboratory. pogaku n, prodanovic m, green t (2007), modeling, analysis and testing of autonomous operation of an inverter-based mg. ieee trans. on power electronics 22(2): 613–625. ramanarayanan v (2010), decentralized parallel operation of inverters sharing unbalanced and nonlinear loads. ieee transactions on power electronics 25(12): 3015–3025. tneb engineers’ association (2008), power engineer’s handbook, 6th ed., chennai, india: allied publishers. zhou kl, yang sl, chen zq, ding sa (2014), optimal load distribution model of microgrid in the smart grid environment. elsevier renewable and sustainable energy reviews 35: 304–310. microsoft word paper 5.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 149-159 realization of a photovoltaic fed sparse alternating current (ac)-link inverter r. ramaprabha*, g. ramya, u. ashwini and a.h. fathima humaira department of electrical and electronics engineering, ssn college of engineering, rajiv gandhi sala, kalavakkam-603110, tamilnadu, india received 9 september 2015; accepted 4 january 2016 abstract: in this paper, a soft-switched alternating current (ac)-link buck-boost inverter with a reduced number of switches, referred to as a sparse ac-link buck-boost inverter, was designed and implemented for a photovoltaic (pv) interface. important features of the sparse configuration included a lower number of switches, lower failure rates, compactness, and cost-effectiveness. the link was composed of a low reactive rating series inductor/capacitor pair. significant merits of the ac-link buck-boost inverter are a zero voltage turn on and a soft turn off of the switches, resulting in minimum voltage stress on the switches and negligible switching losses. in this paper, 10 switches were used instead of 20 switches as are used in existing buck-boost inverter topology. the reduction in the number of switches did not change the principle of operation of the sparse configuration; hence, it remains the same as that of the original configuration. the pulse width modulation (pwm) technique was used for gating the switches. the inverter operation was validated and implemented for pv interface using a microcontroller. keywords: sparse inverter, photovoltaic systems, zero voltage switching, matlab, microcontroller. אאאא אאk*،אk،،kאk    א:א،א אf–אאkאאאאאאאeא אאאאkאא  א א k  א א      l    א  אאאאאאא א   אא  א   k١٠   ٢٠א         א    א  א     א   kא  kא  א    אא  kא א    א kאאא אאw ،אא،،  * corresponding author’s e-mail: ramaprabhar@ssn.edu.in r. ramaprabha, g. ramya, u. ashwini and a.h. fathima humaira   150 1. introduction renewable energy is defined as energy from a source that is not depleted when used (carlson, 1989), and photovoltaic (pv) energy is the most important renewable energy source available. (shen 2008). pv systems are the fastest growing energy source worldwide because of an increase in energy demand, and the fact that fossil energy sources are finite and expensive. pv energy is renewable, inexhaustible, and nonpolluting, and can be used for various applications (zahedi 2002). unlike other nonrenewable energy sources, solar energy sources produce few or no waste products such as carbon dioxide and other harmful chemical substances (subudhi and pradhan 2013). the major challenge in pv conversion is to use the proper power converter circuits in order to convert produced direct current (dc) to practical dc-alternating current (ac) (chakraborty et al. 2009). in a standalone or grid-connected system, the conversion of dc to ac with the proper power quality is important (atcitty et al. 2011). a number of pv inverter topologies have been presented in the literature (kerekes et al. 2011). the major challenges are a reduction of switches to reduce losses and the generation of nearly sinusoidal voltage with reduced total harmonic distortion (kjaer et al. 2005). these converters, which can be configured in many ways, including dc-dc, dc-ac, ac-dc, and ac-ac and have many advantages over other types of converters. specifically, these converters are more compact and reliable, and offer longer lift-time (qin et al. 2002). however, one complication is that these converters employ more switches than any other power processors (sood et al. 1988). their remarkable merits have paved the way for a soft-switching ac-link universal power converter which has received noticeable attention during the last few years (amirabadi et al. 2009). in spite of advantages over the ac-link buckboost inverter, a soft-switching ac-link universal power converter requires many switches when compared to a more conventional inverter (amirabadi 2013); hence, the control method is complicated. in a conventional ac-link buck-boost inverter, the complementary switches are added to provide the pathway for an ac-link current (balakrishnan et al. 2008). this ac-link current may reduce the system efficiency with an increased harmonic content. due to this potential issue, the sparse ac-link inverter is considered, and the configuration with the reduced number of switches is called a sparse ac-link buck-boost inverter. the operating principle of a sparse ac-link inverter is similar to that of a conventional one. the proposed inverter has high reliability and low switching loss (toliyat et al. 2008). the most important feature of a sparse inverter is that it has unidirectional switches which can be fabricated by switch modules. the sparse configuration is considered to be less expensive, less complicated, and more compact compared to the ac-link buck-boost inverter (amirabadi 2013). the sparse ac-link inverter supports bidirectional power flow. in this inverter, the complimentary switches are not essential, although they provide ac (amirabadi 2014). the converter transfer power through the dclink inductor is charged in the input phase and discharged at the output phase. the link frequency is greater than the output line frequency. during the resonance mode, no switches conduct, initiating the soft switching principle (amirabadi et al. 2009). however, due to the number of switches in the sparse ac-link inverter, the switching loss can be increased, affecting the system’s efficiency. hence, in this paper, the sparse ac-link inverter is proposed with a reduced number of switches. due to the reduced number of switches, the control method is very simple compared to a conventional ac-link inverter. in the proposed method, a reduced number of switches is used which makes the control process easier. a softswitching ac-link buck-boost inverter traditionally uses 20 switches, but the sparse configuration proposed by amirabadi et al. (2014) used 18 switches instead of 20. in this paper, a sparse configuration with 10 switches was implemented for pv interface. the operation was checked with matlab/simulink software (mathworks, natick, massachusetts, usa) and validated through practical implementation. the paper is organized as follows: section 2 describes the design and operation of sparse configuration. section 3 presents the simulation results, and section 4 presents the hardware implementation. the conclusion is presented in section 5. realization of a photovoltaic fed sparse alternating current (ac)-link inverter 151   2. design and operation of pv connected sparse ac–link inverter the basic circuit configuration of pv connected sparse ac link inverter is shown in fig 1. the input side and the output side are formed by four unidirectional switches and six bidirectional switches respectively. this inverter is most similar to a dc-ac buck-boost converter in which the link is charged through pv array and discharged into the output phases. the charging and discharging of the link in reverse is facilitated by the complementary switches in each leg. this style of discharging leads to an ac in the link, which leads to better utilization of the link inductor and capacitor. a single-phase high-frequency transformer can be added to the link if isolation is required. the pv panel was modelled using a one-diode model which consists of a current source in parallel with a diode (villava et al. 2009), a shunt resistance, and a series resistance (fig. 2). key modelling is shown in eqns. 1–4. in this case, the link inductance of the transformer is playing the role of magnetizing inductance (amirabadi et al. 2014). 1 tav dtvkocnvexp dtikscnioi                (1) ng g pvnidtikpvi       (2) q aktsn tav  (3)                   1vi n n r n v exp ninii pp ss s s ta ppopppvm                  (4)  figure 1. photovoltaic (pv) interfaced sparse alternating current (ac)-link inverter. figure 2. circuit model used for modelling a photovoltaic (pv) panel. r. ramaprabha, g. ramya, u. ashwini and a.h. fathima humaira   152 the pv array is sized accordingly (considering maximum power point voltage, vmpp and current, impp) and the system parameters are listed in table 1. table 1. system parameters. parameters standards vmpp of single panel 16.54 v impp of single panel 2.55 a pv array size/arm 15×2 (15 panels in series) pv = photovoltaic; v = volts; a = amps. the design of an ac link pv inverter is explained below (amirabadi et al. 2012). for the purpose of simplicity, a resonating time which is less than the power transfer time at full power is neglected. the phase carrying maximum current is involved in de-energizing of the link at modes three and five. the equivalent output voltage and current is given by cosθ 2 opeakπv oeqv  (5) π opeak3i oeqi  (6) where, iopeak is the output peak phase current, vopeak is the output peak phase voltage, and cos (θ) is the output power factor. the link current during energizing and de-energizing is given by l chargetpvv peaklink,i   (7) l dischargetoeqv peaklink,i   (8) where, tcharge and tdischarge are the total charging times during mode one and total discharging time during modes three and five. the average pv current based on the peak of the link current is given as follows:          chargetpeaklink,it 1 pvi (9) where, t is the period of the link current. based on the above equations, the peak link current can be written as:        oeqipvi2peaklink,i (10) the frequency of the link current is determined based on the power rating of the system and the characteristics of available switches. the link inductance is calculated by:                         oeqv pvv12f 1 peaklink,i pvvl (11) link capacitance is calculated such that the resonating periods within 5% of the link cycle at full power. there are 12 operating modes in 10 switch sparse ac link inverter. the link is energized from the input during modes 1 and 7 and is de-energized to the outputs during modes 3, 5, 9 and 11. modes 2, 4, 6, 8, 10 and 12 are devoted to resonance. the link current during mode 1 is given by as below: dt tlinkdilpvv       (12)               0linkil tpvvdt t 0 pvvl 1tlinki (13) the operation in different modes is depicted in table 2. the operating principle is shown in fig. 3 for each mode (amirabadi 2013). realization of a photovoltaic fed sparse alternating current (ac)-link inverter 153   figure 3. operating principle of sparse ac link inverter. table 2. modes of operation of 10 switch sparse ac-link inverter. mode link operation switches in on state switches in conduction 1 energizing s0 and s3 s0 and s3 2 partial resonance s10, s14 and s15 none 3 de-energizing s10, s14 and s15 s10 and s14 4 partial resonance s10 and s15 none 5 de-energizing s10 and s15 s10 and s15 6 partial resonance s1 and s2 none 7-12 the operations are similar to modes 1 to 6. but the link charges and discharges in the reverse direction through complimentary switch at each leg. table 3. simulation parameters parameter values nominal pv voltage (dc link) 240 v output voltage 230 v link inductance 5.5 μh link capacitor 1000 μf pv side filter 0.001 μh, 1000 μf ac side filter 0.200 μh, 200 μf pv = photovoltaic; dc = direct current; v = volt; mh = megahertz; mf = megaxxx; ac = alternating current; μh = millihenry; μf = millifarad. 3. simulation results the circuit shown in fig. 1 was simulated for a 1kw pv system using matlab/simulink (fig. 4), and the simulation parameters are presented in table 3. the filter components were designed using the equations given by amirabadi et al. (2014). pulse width modulation (pwm) pulses were generated in proper sequence to turn on the switches. r. ramaprabha, g. ramya, u. ashwini and a.h. fathima humaira   154   figure 4. matlab model of the 10-switch sparse alternating current (ac)-link inverter. figure 5. control circuit for pulse width modulation (pwm) generation. figure 6. voltage and current at the direct current (dc) link. the nominal voltage of 240 v forming the pv array is given as an input-to-input side bridge, while the output is observed from the three legs of the output side bridge. the input side bridge consists of unidirectional switches while the output side consists of bidirectional switches. the pwm pulses are generated through the control circuit whereas the output current is compared with the reference current (fig. 5). the simulation was carried out using matlab, and the results are presented below. figure 6 shows the dc voltage and current from the pv array. the maximum power from the pv array was tracked by the maximum power point tracking algorithm. the maximum power was tracked from the pv array only when the source impedance matched the load impedance. figure 7 shows the pwm pulses for switches s0–s3. the gating sequence to switches s4–s15 is given by comparing the output current from the inverter (fig. 8). figure 9 shows the input from the dc link voltage and current. in such a set-up, when the capacitor starts charging, the current flowing through the system decreases; when it starts discharging, the current flowing through the system increases. the frequency of the link current is much higher than the frequency of the line. realization of a photovoltaic fed sparse alternating current (ac)-link inverter 155   figure 7. pulses for the switches in the input side bridge. figure 8. pulses for the switches in the output side bridge. soft switching is carried out in the sparse ac link inverter. accordingly, there is zero voltage switching when the voltage across the switch is zero. then the capacitor starts charging. in case of zero current switching, when the voltage across the switch is zero, the current flowing through the inductor is also zero. the soft switching principle is shown in fig. 10 r. ramaprabha, g. ramya, u. ashwini and a.h. fathima humaira   156   figure 9. input dc link current and voltage. figure 10. zero voltage and zero current switching. figure 11. voltage and current at the alternating current (ac) side. figure 11 shows the output voltage and current after filter. the output voltage and current are 3ϕ ac. the peak value of the line to line voltage of each phase is 230 v. the peak value of output current is 5 a. the output voltage is nearly sinusoidal but has a few ripples. the output voltage and current are lower than the input voltage and current. hence, the output power is lower than the input power, which accounts for the losses. realization of a photovoltaic fed sparse alternating current (ac)-link inverter 157   4. hardware implementation the experimental setup of the sparse ac link inverter is shown in fig. 12. the gating pulses were created using a field programmable gate array (fpga) board. the generated pulses are given to the switches of the multi mini capacitor (mmc) via the mct2e opto-coupler. the power supply circuit for the optocoupler and processor was designed using a step-down transformer, bridge rectifier, and regulator. a link inductor and link capacitor are present between the input and output sides to perform partial resonance and to provide galvanic isolation. in order to protect the microcontroller and the control circuit from power circuit an isolation circuit is provided. the experimental results are observed using digital storage oscilloscope (dso). figure 13 shows the dc link output voltage and current. the three-phase output voltage and current waveform from the inverter is shown in fig. 14, where the voltage and current are in phase with each other. the synchronised current is shown for phase a. figure 12. hardware setup of the sparse ac-link inverter. figure 13. direct current (dc) side voltage (200 v/division) and current (5 a/division). figure 14. alternating current (ac) side voltage (200 v/division) and current (5 a/division). r. ramaprabha, g. ramya, u. ashwini and a.h. fathima humaira   158   figure 15. link current (20 a/division). the output ac link current was sinusoidal (fig. 15). the output ac current obtained was approximately 5 a. from the results, it can be observed that a nearly sinusoidal voltage and current were obtained using a sparse ac-link inverter with reduced switches. 5. conclusion this paper presents a soft-switched ac-link sparse inverter with reduced switches. the proposed inverter has some advantages such as zero voltage upon turn on and soft turn off of switches. although the proposed inverter has reduced switches compared to the existing sparse inverter, the efficiency of the system is not affected. the inverter has an lc (inductorcapacitor) pair which acts as a partial resonant circuit. this inverter provides isolation between the input and output without a transformer. therefore, the inverter is more reliable and compact for a grid-connected pv system. the performance of the inverter was evaluated through both simulated and experimental results. acknowledgment the authors wish to thank the management of ssn college of engineering, chennai, india, for providing all the facilities to carry out this work. references amirabadi m, toliyat ha, alexander w (2009), battery-utility interface using soft switched ac link supporting low voltage ride through. ieee energy conversion congress expo and conference. san jose, ca, usa. 2606–2613. amirabadi m (2013), soft-switching highfrequency ac-link universal power converters with galvanic isolation. ph.d. dissertation, texas a&m university, college station, tx, usa. amirabadi m, baek j, toliyat ha (2014), sparse ac-link buck–boost inverter. ieee transactions on power electronics 29(8): 3942– 3953. amirabadi m, balakrishnan a, toliyat h, alexander w (2014), high frequency ac-link pv inverter. ieee transactions on industrial electronics 61(1): 281–291. amirabadi m, toliyat ha, alexander w (2009), battery-utility interface using soft switched ac link buck boost converter. electric machines and drives conference 1299-1304. atcitty s, granata je, quinta ma, tasca ca (2011), utility-scale grid-tied pv inverter reliability workshop summary report. sandia national labs, sandia rep. sand2011-4778. balakrishnan a, toliyat ha, alexander wc (2008), soft switched ac link buck boost converter. twenty-third annual ieee appl. power electronics conference exp, austin, texas: ieee 1334–1339. realization of a photovoltaic fed sparse alternating current (ac)-link inverter 159   carlson de (1989), fossil fuels. the greenhouse effect and photovoltaics. ieee aerospace and electronics systems magazine, 4(12): 3-7. chakraborty s, kramer b, kroposki b (2009), a review of power electronics interfaces for distributed energy systems towards achieving low-cost modular design. renewable and sustainable energy reviews 13(9): 2323-2335. kjaer sb, pedersen jk, blaabjerg f (2005), a review of single-phase grid-connected inverters for photovoltaic modules. ieee transactions on industry applications 41(5): 1292-1306. kerekes t, teodorescu r, rodriguez p, vazquez g, aldabas e (2011), a new high-efficiency single-phase transformerless pv inverter topology. ieee transactions on industrial electronics 58(1): 184-191. qin yc, mohan n, west r, bonn r (2002), status and needs of power electronics for photovoltaic inverters. sandia national labs, sandia rep.: sand2002-1535. shen wx (2008), design of standalone photovoltaic system at minimum cost in malaysia. ieee conference on industrial electronics and applications 702 -707. sood pk, lipo ta, hansen ig (1988), a versatile power converter for high frequency link systems. ieee transactions on power electronics 3(4): 383–390. subudhi b, pradhan ra (2013), comparative study on maximum power point tracking techniques for photovoltaic power systems. ieee transactions on sustainable energy 4(1): 89–98. toliyat h a, balakrishnan a, amirabadi m, alexander w (2008), soft switched ac-link ac/ac and ac/dc buck-boost converter. in: ieee power electronics specialist conference; 15-19 june 2008; rhodes: ieee: 4168–4176. villalva mg, gazoli jr, filho er (2009), comprehensive approach to modeling and simulation of photovoltaic arrays. ieee transactions on power electronics 24(5): 1198– 1208. zahedi a (2002), development of a numerical model for evaluating the performance of renewable generation systems. proceedings of ieee tencon, 1950-1953. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 16, no. 2 (2019) 96-102 in achieving the most of limited research budget for engineering research projects: a case study at sultan qaboos university m. zarog * department of mechanical and industrial engineering, college of engineering, sultan qaboos university, muscat, oman. abstract: motivating engineering research in universities with limited or tight research budgets is a challenging task. there is always a conflict between researchers attempting to secure enough funds to ease their own research, and funding bodies and universities trying to achieve high quality research while lowering research expenses. this conflict leads to a situation where universities and researchers cannot achieve their goals altogether. this study attempts to partially resolve this issue by optimizing research expenditure for the benefit of both universities and researchers and suggests some policies regarding what type of research (modelling based or experimental based) universities should support. other factors related to the components of research expenditure (effect of local and international collaboration, hiring research assistants (ras), establishing research laboratories, and buying professional engineering software) are also considered. _________________________________________________________________________________________ keywords: research expenses; research budget; research fund; research collaboration; research types. دراسة حالة بجامعة السلطان قابوسلمشاريع البحثية الهندسية في ظل ميزانية محدودة: من اتحقيق األفضل *مصعب عبدهللا زروق الدائم صراع نوع من ال. هناك للجامعات مهمة صعبة ةصغيرأو ذات ميزانية بحث محدودة يعد تحفيز البحوث الهندسية في الجامعات :الملخص التمويل والجامعات التي تحاول ورجهات هيئاتبين ، ورجرا بحوههم العلمية بههولة ويهرإلالمادي الدعمبين الباحثين الذين يحاولون الحصول على في ال تهتطيع فيه الجامعات والباحثين تحقيق أهدافهم البحثية األبحاث. يؤدي هذا الصراع إلى وضعهذه تحقيق أبحاث عالية الجودة مع خفض نفقات الجامعات التورجيه األفضل لميزانية البحث العلمي بحيث تخدم كال الطرفين: تحاول هذه الدراسة حل هذه المشكلة رجزئيًا من خالل ورجميعًا. آن واحد و باختيار البحوث المناسبة للدعم المادي في المتعلقة بعض المقترحات اإلرجرائية من رجهة أخرى. كما تقدم الدراسةوالباحثين والجهات الممولة من رجهة اعتماد البحث على التجارب المعملية أو على النمذرجة البحث ) يةنوعظل ميزانية دعم محدودة ، وهذه المقترحات تتعلق بعدة رجوانب منها : اختيار هندسية البرامج ال، وشرا شرا أرجهزة اختبار علمية توظيف مهاعدي البحوث ، لي، والمحاكاة والتصميم الرياضي( ، التعاون المحلي والدو تخصصة. مال _____________________________________________________________________________ .أنواع البحوث العلمية ؛التعاون في البحوث العلمية ؛الدعم المالي للبحوث ؛ميزانية البحث العلمي ؛منصرفات البحوث العلمية :الكلمات المفتاحية *corresponding author’s e-mail: musaabh@squ.edu.om doi:10.24200 /tjer.vol.16iss2pp96-102 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 96-102 1. background and objectives universities in general rely heavily on public funds to support their research expenditure. the federal government of the united states, for example, spends billions of dollars each year to support r&d at universities and that huge amount of money still does not cover what these universities actually spend on r&d (evan 2017 and daridah 2014). as a result, universities still have to spend on researchers from their own resources. in many cases, universities all over the world are facing more restrictions on getting public fund for their research (faridah 2014). regardless of the amount that universities contribute to the research from their own resources, that fund should be spent wisely and efficiently (paul and beth 2014). there is an important fact that is usually overlooked which is, universities and researchers do not share the same view on research and its expenditure. universities attempt to spend less on research while seeking to have a large revenue from it. conducting a large amount of high quality research makes universities more reputable and can attract more postgraduate students, and funds from both public and private sectors. on the other hand, researchers seem to care less about the research expenses as long as they can secure funds for their research. the tendency of most researchers is to publish in highly reputed journals (bennis 2005, elisha 2018) and there is evidence that professors and researchers mainly publish to obtain different academic rewards such as promotion (rajaeian 2018) and joining a better institutions and universities. these two different objectives of conducting research, make it hard to balance research funding against research expenses. research spending in engineering is related to the cost of carrying out the research itself (such as equipment, experiments, software, etc.) and the cost of hiring researchers to conduct it. this paper attempts to investigate some factors that can lead to research satisfaction for both researchers and universities. these factors include whether universities with limited research budgets should focus on modeling or experimentally based research, and how they would distribute available funds among research projects. questions include: whether it is better for universities, with limited research budgets, to fund a large number of less costly research projects or whether they should focus more on the quality of the research regardless of its cost? is it advisable for universities to focus more on local or international collaboration? how can universities optimize their spending between hiring researchers, buying professional engineering software, and establishing or improving labs for experiments? the objective of this paper is to address the aforementioned points. sultan qaboos university (squ) is considered to be one of the highly reputable universities in the region and it is very much concerned with its research progress. the questionnaire on which this study is based was carried at squ department of mechanical and industrial engineering. 2. survey method and results the questionnaire used in this study was developed by the author in english and it was distributed to the entire academic staff at the department of mechanical and industrial engineering of sultan qaboos university (squ). the questionnaire was handed in person to the target group. two questionnaires were prepared for each staff member, one to measure the impact of a conference research and the other is to evaluate the impact of a peer-reviewed journals published research. 28 of the 29 academic staff responded successfully to the questionnaire – each staff member completed the two questionnaires-with an almost 97% response rate, and 54 questionnaires were collected. some follow-up attempts were made to ensure this high response percentage, and all the staff members who received the questionnaire showed a willingness to participate. every staff member was given two questionnaires, one of which is concerned with research that has led to a recent publication in a peer-reviewed journal, and the second is about recent research that has led to a conference publication. the author referred to the recent publications lists (2016 and 2017) of the department of mechanical and industrial engineering and picked one journal paper and one conference paper for almost each staff member. since the questionnaire requires knowledge about the research details (such as cost and funding, collaboration, people involved, etc.), the author selected only those papers in which the researcher was the first author. the questions were clear and direct as shown in table 1 (see appendix). 3. results and discussion the overall summary of the results is presented in fig.1 and each factor is discussed separately in the following subsections 3.1 modeling-based vs experimental based research modeling-based research is the research that was carried out only at the modelling, mathematical, and/or simulation level, without experimental verifications of the results. mathematical based modelling seems to be underestimated in many cases and almost the whole weight is given to experimental results which are not always true. a model might be needed, however, to interpret data or to confirm or reject a hypothesis via generating predictions and they are essential to the experimental analysis (mazur 2006, and mobius and laan 2015). they can also be used to predict or control responses (mazur 2006). when there is not enough budget to carry out experimental work, researchers have no other choice than focusing on modelling-based research. although research that is only modellingbased is usually cheaper to conduct, on the other hand researchers are usually unable to publish their results in a highly reputed journal and that might also affect their promotion. most reviewers in reputed journal ask for experimental verification of the modelled results. it can be noticed from the survey results (fig.1.a) that 97 m. zarog a large portion of the research was only based on modelling (almost 37%). on the other hand, the low cost of modelling based research can be proved by checking the cost of this type of research. 36.8% of this research type cost less than omr 100 (compared to only 15% in the experimental based research) and almost 63% of the modelling based research cost below omr 2000 (compared to only 40% in the purely experimental based research). modelling based research usually requires design or analysis via professional engineering software. it is found that 36.8% of modelling based papers used software that was not provided by squ. this indicates that squ should increase college budget in purchasing engineering software which will increase the number of this type of research. another point that was noticed from the survey results is that 63% of the modelling based research involved research assistants and postgraduate students which will definitely, and unnecessary, increase the cost of this type of research. common practice in universities is to involve research assistants and postgraduates in practical research and/or in creating new research kits that could further lead to a more experimental-based research. staff can conduct their modelling research on their own provided that enough training is provided for the staff about the particular engineering software that they are interested in. 3.2 level of collaboration findings from a dataset on all the researchers from the engineering departments in the uk in the past 20 year, showed that research collaboration with industry not only benefits knowledge transfer, attracts funds for research, and increases economic impacts, but it also positively affects the rate of publication for researchers (bonet et al. 2008). on the other hand, international collaboration with other universities and research groups may indicate negative issues like resource dependence, less research experience for needing help, and high concern of researchers to advance their own career. it may also indicate that collaborators are working in a narrow or very specialized fields (iglič et al. 2017). there is evidence that the rate of publication increases within a certain threshold of collaboration between researchers and industry, but too much collaboration will have a negative impact on the rate of publication. this section investigates issues related to conducting research such as, availability of research equipment and research assistants and postgraduates. the survey shows that a large portion of the research conducted (almost 63%) was experimentally based, and almost only half of these experiments (53%) were conducted inside squ labs (fig.1.b). this is a good sign of research collaboration in terms of conducting research outside squ labs (inside and outside oman). the results also show that most of these experiments that involved external collaboration were carried out with international institutions (87%) and only 13% were carried out inside oman, but not in squ. these results obviously show dependence on external international resources and low collaboration within oman. even though international collaboration is recommended sometimes to attract more research funds, squ should have their own laboratories to attract local and regional companies. in terms of research human resources, surprisingly the survey showed that 35.3% of the experiments were conducted by the professors themselves without any help of research assistants and postgraduate students. this will inevitably affect teaching quality and other duties of the professors and put more burdens on them. 3.3 costs versus funds one of the main factors to consider is the research cost, so we should not treat projects that cost hundreds of thousands with those which cost only a few, or tens of thousands. since research funds are limited in most cases, is it advisable to distribute the available fund to a larger number of research projects or spending it all in a lower number of research projects. the results in fig.1.c show that almost 28% of the research is conducted at almost no cost (less than omr 100) and nearly 60% of the research is conducted at a low cost (less than omr 2000) which also indicates that a large portion of the research can be carried out at a very low cost. hence, squ can accommodate more research with the same budget. the availability of research funds greatly affects both the quality and quantity of research produced. universities always strategically plan to attract more external funds for research than using their own budget to fund research. with this in mind, the survey shows that more than half of the experimental research (53%) was funded by squ, compared to 33% funded by local bodies in oman (fig.1.d). international bodies and institutions funded only 14% of the research. on the other hand, when comparing the modeling-based research and experimentally-based research, one expects that experimental research costs more but the results indicate that both experimental and modeling based research receives the same number of funds from squ. 3.4 level of involvement of ras and postgraduate students additional research cost comes from employing research assistants (ras) and offering postgraduate scholarships, but it is obvious that the availability of research assistants and postgraduate students improves the quality as well as the quantity of research conducted. even though, it is expected to see more ras in experimental-based research, the result showed little difference between the two types of research (43% of the modeling-based research employs ras as opposed to 57% in the experimentally based). when looking at both the ras and post graduates involved in the research, it can be noticed from the results 98 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 96-102 (a) (b) (c) (d) figure 1. summary of the results (a): modeling-based vs experimental based research, (b): conducting research (c): research cost, (d): research fund. that 63.2% of the modelling based research involved ras and postgraduates as opposed to 64.5% of the experimental based research. again, it is expected that the former should have much less percentage. 4. conclusion conducting good quality research is always a challenge that faces most universities, subsequently it is of major concern for researchers and faculties at universities. the main barrier from conducting high quality research is the limit in budget. below are some recommendations based on observations from this study. universities with tight research budgets should give more attention to modeling-based research, which is normally cheaper to conduct than experimental research. the results show that this can reduce research expenses by 60% to 37%. to achieve the aim of the above point, universities have to invest more in purchasing necessary engineering modeling software and to provide full training for the staff and researchers so that they can carry their research independently. this should reduce the cost associated with hiring ras or giving scholarships for postgraduates. to encourage researchers to publish modeling based papers, universities should avoid giving high value for publishing in a very highly reputable journals specially when promoting staff. the results also recommend reducing unnecessary this appointments of ras in modeling type research as it should be more limited to experimental based research. the results also indicate clearly the high level of collaboration between squ and international institutions compared to local collaboration. this scenario should be completely reversed through investing in research that is more relevant to local industry in oman. a great portion of the university research budget can be allocated to establish new research infrastructure relevant to local industry. international collaboration is still needed to continue but it could be more limited to modeling based research as opposed to experimental research. the results also show the high engagement of academic staff in conducting research experiments themselves. this issue can be tackled by directing ras towards experimental research type. the survey also shows that many research projects can be conducted by the staff at low cost. hence, it is recommended for certain allocated research budgets to increase the number of funded research projects while reducing the fund share for each project. this can also be achieved by distributing the fund of costly research projects into a larger number of less costly projects. in other words, universities should fund lower cost research rather than giving large funds to fewer 99 m. zarog projects. the results also suggest that the number of researchers and research projects can be increased by investing more in buying professional research software while providing sufficient training for the staff. following the above recommendations should definitely lead to involving more staff in research and/or involving staff in more research. conflict of interest the author declares no conflicts of interest. funding no funding was received for this research. acknowledgment the author would like to thank all teaching staff at the department of mechanical and industrial engineering for their participation in the survey that led to this paper. also special thanks to dr. nabeel al-rawahi for revising the questionnaire draft and giving some comments. references bennis wg, o’toole j (2005), how business schools lost their way. harv. bus rev. 83 (5): 96–104. bonet a., estanol b. meissner m (2008) the impact of industry collaboration on academic research output: a dynamic panel data analysis. working papers 08/14, department of economics, city university london. cater s.a. lane m, rajaeian mm (2018), determinants of effective knowledge transfer from academic researchers to industry practitioners. journal of engineering and technology management 47: 37-52. deborah ab, elisha rt (2018), research trends and collaborations by applied science researchers in south african universities of technology: 2007– 2017. the journal of academic librarianship 44: 468–476 evan c, michael bs, samuel s, thomas cf (2017), universities getting the most money from the federal government 24/7 wall st.march 22, https://247wallst.com/special-report/2017/03/22/ uni-versities-getting-the-most-money-from-thefederal-government/5. faridah ha, ibrahim ka, kalsom s, noor hh, syed ns (2014), funding trends for research universities in malaysia. international conference on accounting studies. icas 2014, 18-19 august 2014, kuala lumpur. iglič, h., doreian, p., kronegger, l. and anuska ferligo (2017), with whom do researchers collaborate and why? scientometrics 112: 153. jensen p, webster b (2014), let’s spend more wisely on research in australia, http:// the conversation.com/lets-spend-more-wisely-on-research-inaustralia-29053, retrieved february 15, 2019. mobius w, laan l.(2015), physical and mathematical modeling in experimental papers. cell 163(7):157783. mazur j.e. (2006). mathematical models and the experimental analysis of behavior. journal of the experimental analysis of behavior 85(2), 275-91. 100 https://247wallst.com/special-report/2017/03/22/%20uni-versities-getting-the-most-money-from-the-federal-government/5 https://247wallst.com/special-report/2017/03/22/%20uni-versities-getting-the-most-money-from-the-federal-government/5 https://247wallst.com/special-report/2017/03/22/%20uni-versities-getting-the-most-money-from-the-federal-government/5 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 96-102 appendix a table a.1 survey categories and questions. category no. question type of research and publication 1 what is the type of research? 1.1 modelling based 1.2 experimental based 1.3 both 2 where the results were published? 2.1 peer-reviewed journal 2.2 conference proceeding availability of resources 3 if experimental data are used, what is the source? 3.1 literature or no data was used 3.2 experiment carried at squ 3.3 experiment carried outside squ 4 if any software used in the research, is it provided by squ or not? 4.1 yes 4.2 no availability of research assistants and postgraduates 5 what is the number of research assistants and postgraduates (if any) involved in the research with squ? 5.1 msc students (no.= ?) 5.2 phd students (no.= ?) 5.3 research assistants (no.= ?) level of collaboration (locally and internationally) 6 was there any collaboration with other universities or individuals? 6.1 none 6.2 within department 6.3 within squ 6.4 regional 6.5 international research cost and funding 7 what is cost for the research (rough estimation)? 8.1 0 to 100 in omr 8.2 100 to 1000 8.3 1000 to 2000 8.4 2000 to 5000 8.5 5000 to 10,000 8.6 more than 10000 9 have you received any fund for the research from? 9.1 squ, percentage of fund to cost = ? 9.2 local bodies in oman, percentage of fund to cost= ? 9.1 bodies in gulf region, percentage of fund to cost =? 9.1 international bodies, percentage of fund to cost =? 101 m. zarog appendix b table b.1 questionnaire response table. table keys: q = question number in questionnaire x(y)= number of participants who chose answer (y) (p,q) = multiple answers mix= participants selected multiple choices (0) = no or participants didn’t answer q.5 = sum (postgraduates, ras) q.8 = sum (funding body, percentage of fund) journal papers q.1 q.2 q.3 q.4 q.5 q.6 q.7 q.8 q.9 q.10 11(1) 11(1) 7(1) 13(no) 4(1,1) 6(1) 4(1) 8(1,100) 14(1) 11(1) 3(2) 6(2) 6(2) 13(yes) 3(2,1) 3(2) 3(2) 4(2,100) 2(2) 3(2) 2(3) 7(3) 5(3) 8(3,1) 1(3) 3(3) 3(1&2, 100) 5(3) 5(3) 4(4) 2(mix) 2(mix) 1(3,2) 6(4), 8(5) 5(4), 7(5) 3(6) 2(4) 4(4) 0(4) 6(mix) 6(no) 10(no) 2(mix) 1(no) 1(1,4) 1(mix) 7(mix) 8(0) conference papers 9(1) 8(1) 5(1) 14(no) 1(1,1) 7(1) 7(1) 9(1,100) 18(1) 17(1) 7(2) 12(2) 9(2) 14(yes) 2(2,1) 1(2,2) 11(2) 8(2) 7(2,100) 2(2) 4(2) 1(3) 8(3) 3(3) 10(3,1) 2(3) 3(3) 1(3, 100) 4(3) 3(3) 6(4) 3(mix) 1(3,3) 3(4), 2(5) 2(4) 3(5) 2(6) 1(4) 3(4) 1(4) 5(mix) 8(0) 1(mix) 1(0) 3(mix) 3(0) 10(no) 1(mix) 3(mix) 102 copy of copy of comp050508.qxd the journal of engineering research vol. 3, no. 1 (2006) 69-74 1. introduction model predictive control or a receding horizon control has become an attractive feedback strategy during recent years. generalized predictive control (gpc) of linear systems has found successful applications, especially in the process industries, due to its robustness to parameter uncertainties (boucher and dumar, 1996). however, many systems are, in general, inherently nonlinear. thus, linear models are often inadequate to describe the process dynamics and nonlinear models have to be used. much effort has been made to extend gpc to nonlinear systems (michalska and mayne, 1993). however, in model predictive control, a nonlinear optimization problem must be solved online with computational complexity (henson and seborg, 1997). therefore, this kind of control law can be applied only for systems with slow dynamics (for example chemical process). to avoid this, several nonlinear predictive laws have been developed (ping, 1995; singh, et al. 1995; souroukh and kravaris, 1996) where the one-step ahead predictive error is obtained by expanding the output signal and reference signal in a rith order taylor series, and rith is the relative degree of the ith output. then, the continuous minimization of the predicted _________________________________________ *corresponding author’s e-mail: hedjar@ccis.ksu.edu.sa tracking errors is used to derive the control law. this paper examines the nonlinear continuous-time generalized predictive control approach based on a finite horizon dynamic minimization of predicted tracking errors with end point constraints to achieve tracking objectives. it is noticed that, the proposed approach increases the tracking performance with regards to ping's lu method (ping, 1995) and chen's method (chen, et al. 2003). indeed, the two performance indexes used by ping (1995) and chen, et al. (2003) are combined (control integration) to increase the dynamic of the tracking error. moreover, the recent chen et al. method given in chen, et al. ( 2003) assumes that decoupling matrix is not singular, which is not practical. with the proposed combination control, the assumption above will be weakened. it will be shown that some advantages of this control scheme include good tracking performance, clear physical meaning of maximum and minimum control values when saturation occurs with regard to input-output linearization method. minimizing a predictive cost function derives the proposed nonlinear predictive controller. a mobile robot is used as an illustrative example to show the tracking performance achieved by this nonlinear predictive controller. note that taylor approximation is used to overcome the computation burden. the rest of the paper is organized as follows: in section 2, the problem statement is given; a control law is develnonlinear predictive control with end point constraints r. hedjar* computer engineering department, college of computer and information sciences king saud university, p.o.box 51178riyadh 11543, saudi arabia received 8 may 2005 ; accepted 16 november 2005 abstract: the optimal nonlinear predictive control structure with end point constraints is presented, which provides asymptotic tracking of smooth reference trajectories. the controller is based on a finite horizon continuous time minimization of nonlinear predicted tracking errors. a key feature of the control law is that its implementation does not need to perform an online optimization, and asymptotic tracking of smooth reference signal is guaranteed. the proposed control scheme is applied to planning motions problem of a mobile robot. simulations results are performed to validate the tracking performance of the proposed controller. keywords: nonlinear predictive control, stability, mobile robot and tracking trajectory á«fé¡ædg á£≤ædg ~j~– ™e ájdƒñæàdg á«£ÿg á«æñdg ::áá°°uuóónnggá«æer ᪫b ~ªà©e »ªµëàdg èdé©ÿg .á∏°ù∏°s á«©lôe ägqé°ùÿ io~fi á≤móe øedƒj »àdgh ,á«fé¡ædg á£≤ædg ~j~– ™e ájdƒñæàdg á«£ÿg á«æñdg ¤g ádé≤ÿg √òg ¥ô£àj á«©lôÿg iqé°tód ioh~ùg á≤móeh ,ô°téñe πµ°ûh ≈∏ãe’g ™°vƒdg ré‚’ êéàëj ö«cîdg ¿g »g ºµëàdg ¿ƒfé≤d á«méàøÿg ᪰ùdg .á«£ÿg òz ájdƒñæàdg á≤móÿg aé£n’ ájô¨°ug .ìî≤ÿg »ªµëàdg èd驪∏d á≤móÿg agocg â«ñãj πlg øe iécéùg ré‚g ” éªc .∑ôëàe äƒhôd ᣣfl äécôm ≈∏y ámî≤ÿg ºµëàdg iqgo ≥«ñ£j ” ~bh .áfƒª° e á°ù∏°ùdg gáá««mmééààøøÿÿgg ääggooôôøøÿÿ.á≤móe qé°ùeh ∑ôëàe äƒhhq ,á«jéñãdg ,»£n òz …dƒñæj ºµ– : 70 the journal of engineering research vol. 3, no. 1 (2006) 69-74 oped to minimize the difference between predicted and desired response. the properties of the control law are discussed, including stability. in section 3, the proposed controller is applied to the planning motion problem of the mobile robot. the results are summarized in section 4 where some directions for related research are also provided. 2. model formulation 2.1 optimal nonlinear predictive control in the receding horizon control strategy, the following control problem is solved at each t > 0 and x(t) : (1) subject to the state equation: (2) on one hand, the objective function used by ping is the one step predicted tracking error given by: (3) on the other hand, the cost function used by chen et al. (2003) is the predicted tracking error over a fixed horizon without control penalty, and is given by: (4) it is noticed that the derived optimal control of the above cost function increases the dynamics of the tracking error with regard to ping's method, but the decoupling matrix should be nonsingular. in this paper, to maximize the tracking performance of previous methods and overcome the assumption of the decoupling matrix, we propose the integration of these two control structures (control integration), and the cost function to be minimized is rewritten as: (5) where q1 and q2 are positive definite matrices respectively and r positive semi-definite matrix. (6) a simple and effective way to predict the influence of u(t) on x(t+h) is to use rith order taylor series expansion, in such a way as to obtain for each component of the vector: (7) the predicted state variable in compact form is given by: (8) where we also expand each component of xref (t+h) in rith order taylor series to have: xref (t+h) = xref (t) + d(t,h) (9) where d(t,h) = [d1 d2 ... dn]t with [ ]τττ+ττ = ∫ + d)(ur)(u)(xq)(x 2 1 min)t),t(u),t(x(jmin ht t tt )t(u )t(u)x(g)x(f)t(x +=& and the terminal end point constraint x(t+h) = 0 for some h > 0, where x(t) ⊂ x∈ ℜn is the state vector, u(t) ⊂ u∈ ℜm represents the control vector. the functions f(x): ℜn → ℜn and g(x) : ℜ n → ℜ n×m are sufficiently differentiable, q is positive definite matrix and r positive semi -definite matrix. we denote the optimal control vector to the above problem by: u*( τ), τ∈ [t, t+h]. the currently applied control u(t) is set equal to u*(τ). this process is repeated for every next t for stabilization of the system at the origin. however, solving a nonlinear dynamic optimization problem with equality constraints is highly computationally intensive, and in many cases it is impossible to be performed within a reasonable time limit. furthermore, the global optimization so lution cannot be guaranteed in each optimization procedure since, in general, it is a non convex, constrained nonlinear optimization problem (henson and seborg, 1997). to overcome these kinds of problems, ping (1995) and chen, et al. (2003) proposed to use taylor approximation. the desired trajectory is specified by a smooth vector function x ref(t) ∈ ℜ n for t∈ [t0, tf]. the problem consists in elaborating a control law u(x, t) that improves tracking accuracy along the interval [t, t+h], where h > 0 is a prediction horizon, such that x( t+h) tracks x ref(t+h). that is, the tracking error is defined by: )ht(x)ht(x)ht(e ref +−+=+ )t(u)f(ll !r h )f(l !r h )f(l 2 h )f(lh)t(x)ht(x i 2r fg i r i 1r f i r if 2 i 0 fii i i i i −− ++ +++=+ l with fi(x) being the i th component of f(x) and )f(l i k f denotes the k th order lie derivative of fi with respect to f(x). [ ]tn21 )h,x(z)h,x(z)h,x(z)h,x(z l= ; ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = !r h !r h diag)h(ë n r 1 1r n l ; w(x) = [w 1 w2 … wn] t with i 1ri f i ri if 2 ii fl!r h .......fl 2 h fh)h,x(z −+++= and [ ]i2rifgmi2rif1gi fllfllw −−= l for i=1,…,n. 2 r 2 q1 )t(u 2 1 )ht(e 2 1 j ++= ττ+= ∫ d)t(e2 1 j h 0 2 q2 2 r h 0 2 2q 2 q1 )t(u 2 1 d)t(e 2 1 )ht(e 2 1 j +ττ+++= ∫ )t(u)x(w)h(ë)h,x(z)t(x)ht(x ++=+ )r( iref i ri iref 2 irefi i x !r h .........x 2 h xhd +++= &&& 71 the journal of engineering research vol. 3, no. 1 (2006) 69-74 the tracking error at the next instant (t+h) is then predicted as a function of u(t) by: (10) (11) where 2.2 tracking performances we assume that the matrix w(x) is of full rank. this assumption is needed for the stability analysis, but not necessary for the control law to be applicable, since one can always choose r>0, then the inverse matrix in eq. (11) will still exist. let r=0, then eq. (11) becomes: (12) let q1=q1in and q2=q2in, then the dynamic of the closed loop equation of the ith component of tracking error vector e(t) is given in compact form by: (13) where for most mechanical systems with actuator dynamics neglected, the relative degree is ri=1 or ri=2. in the case where ri=1, the eigenvalue of the characteristic equation of the error dynamic is: we conclude that the proposed controller achieves faster tracking error dynamic with regards to ping, (1995) or chen et al. (2003) method. 3. simulation example in this section, the reference trajectory-tracking problem of mobile robot is simulated to show the validity and achieved performance of the proposed method. a kinematics model of a wheeled mobile robot with two degrees of freedom is given by kim, et al. (2003): (14) where the forward velocity v and the angular velocity ω are considered as the inputs, (x,y) is the center of the rear axis of vehicle, and θ is the angle between heading direction and x-axis. d is the distance from the origin coordinate of the mobile robot to the axis of the driving wheel. the nonholonomic constraint is written as the following equation: (15) the nonlinear model of the mobile robot can be rewritten as: z = g(θ) u (16) where note that the above model matches the multi-variable affine nonlinear system given by eq. (2) with f(x) = 0. consider the problem of tracking a reference trajectory given by the equations: )t(u)x(w)h(ë)h,t(d )h,x(z)t(e)ht(x)ht(x)ht(e ref +− +=+−+=+ in order to find the current control u( t) that improves tracking er ror along a fixed interval and to avoid the computational burden, the expression of the above predicted tracking error is used in the objective function eq. (5). therefore, the unique control signal u op that minimizes j, obtained by setting 0 u j = ∂ ∂ , is (see appendix): ( ))h,x,x(v)t(e)h(k)h(ã)x(wu ref11op +−= −− ping’s method (ping,1995) : h 1 s 0 −= . chen et al. (2003): h2 3 s 0 −= proposed method: h 3 hq q 2 hq q s 2 1 2 1 0 ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ + −= . ⎪⎩ ⎪ ⎨ ⎧ ω=θ θ= θ= refref refrefref refrefref )sin(vy )cos(vx & & & ( ) ( )dt)t,t(d)t,x(zq)t(e )h,t(d)h,x(zq)h(e)h,x,x(v h 0 2 t 1 t ref −+ −= ∫ && && )t(x)t(x)t(e ref−= ; ∫+= h 0 21 t dtq)t(ëq)h(ë)h(k ; ∫+= h 0 2 tt 1 t dt)t(ëq)t(ë)h(ëq)h(ë)h(ã ; ( ) ( ))h,x,x(v)t(e)h(k)x(w r)x(w)h(a ~ )x(wu ref t 1t op + +−= − ∑ = =⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ++ + ir 0j i )j(j i 2 1 0e!j h 1rj hq q iref )j( i 1j fi )j( xfle −= − for 0 < j < r i and irefiii )0( xxee −== for j=0. the error dynamics eq. (13) are linear and time invariant. thus, the proposed controller that minimizes the predicted tracking error naturally leads to a special case of input/state linearization. the advantage of this controller with regards to the linearization method is a clear physical meaning of maximum and minimum when saturation occurs. note that, by using routh criterion, we can show that the tracking error dynamics eq. (13) are stable only for systems with r i ≤ 4. θ=θ−θ d)(sinx)(cosy && [ ]tyxz θ= ; 10 )cos(d)sin( )sin(d)cos( )(g θθ θ−θ =θ and [ ]tvu ω= . 72 the journal of engineering research vol. 3, no. 1 (2006) 69-74 (17) the optimal control that minimizes the objective function, eq. (5), subject to constraints, eq. (1) is: (18) in the simulation, the control parameters are chosen as: h = 0.002, q1 = q2 =500 i3, r=10-3 i2. the reference model and initial conditions are: (19) the limitations on control signals are: (20) figures 1 and 2 show the tracking trajectory in the xy plane. notice how gracefully the mobile robot tracks the reference trajectory in less than one revolution. figure 3 depicts the position tracking error given by: (21) we can see that the mobile robot tracks the reference trajectory successfully. it was stated in the introduction that the advantage of the proposed approach with regard to input-output linearization method is a clear physical meaning of maximum and minimum control values when saturation occurs. figure 4 depicts the control histories v(t) and w(t) with saturation. note the initial saturation for both controls due to the large tracking error at the beginning. the control parameters q1, q2 and r can be chosen as in classical optimal control theory. the choice of the control parameter h, on the other hand, affects the rate of convergence and the amplitude of the control signal. indeed, to bring the control signals inside the saturation limits, one has to increase the value of h (see fig. 5) and this will increase time response of the system slightly. 4. conclusions in this paper, optimal nonlinear predictive controller with end point constraints using taylor approximation is presented and is applied to multi-variable affine nonlinear system. minimizing a quadratic cost function of the predicted tracking error and the control input derived control laws. one of the main advantages of this control scheme is that it does not require an online optimization and asymptotic tracking of the smooth reference signal is guaranteed. moreover, the dynamic of the closed loop system is better than the dynamic obtained by both ping's and chen's methods. the proposed controller is applied to planning motion problem of mobile robot under nonholonomic constraints. finally, we expect that the results presented here can be explored and extended to discrete implementation of this continuous-time predictive controller either through computers or special purpose chips that can run at higher speeds. ⎪⎩ ⎪ ⎨ ⎧ ω=θ θ= θ= refref refrefref refrefref )sin(vy )cos(vx & & & ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +−⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +θ ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ +θ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +θ−= − − )t(zq 3 h qh)t(eq 2 h q)(g rh)(gq 3 h q)(g h 1 u ref2121 t 1 2 21 t & where )t(z)t(z)t(e ref−= figure 1. tracking performance for t ∈ [0, 0.3s] position of the mobile robot and the reference trajectory y( t) x(t) figure 2. tracking performance for t ∈ [0, 2s] position of the mobile robot and the reference trajactory for t = [0.03] y( t) x(t) ωref = 4rad/s, d=1m, vref =30 m/s, x(0)=0, y(0)= 4 m and 2 )0( π −=θ rad. smv /500≤ and sradw /200≤ 73 the journal of engineering research vol. 3, no. 1 (2006) 69-74 boucher, p. and dumur, d., 1966, la commande prédictive , collection méthode et pratique de l'ingénieur, edition technip. paris. chen, w.h., balance, d.j. and gawthrop, p.j., 2003, “optimal c ontrol of nonlinear systems: a predictive c ontrol approach,” automatica, vol. 39, pp. 633 -641. demircioglu , h. and gawthrop, p.j., 1991, “continuous -time generalized predictive control (gpc),” automatica , vol. 27 (1), pp. 55-74. henson, m.a. and seborg, d.e., 1997, “nonlinear process control,” prentice hall . kim, m.s., shin, j.h., hong, s.g. and lee, j.j., 2003, “designing a robust adaptive dynamic controller for nonholonomic mobile robots under modeling uncertainty and disturbances, ” mechatronics , vol. 13, pp. 507 -519. michalska , h. and mayne, d.q., 1993, “robust receding horizon control of constrained nonlinear s ystems,” ieee transactions on automatic control , vol. 38 (11), pp. 1623 -1633. ping, l., 1995, “optimal predictive control of continuous nonlinear systems,” int. j. of control , vol. 62(2), pp. 633 -649. singh, s.n., steinberg, m. and digirolamo , r.d., 1995, “nonlinear predictive control of feedback linearizable systems and flight control system design,” j. of guidance, control and dynamics , vol. 18(5), pp. 1023 -1028. souroukh, m. and kravaris, c., 1996, “ a continuoustime formulation of nonlinear model predictive control,” int. j. of control , vol. 63 (1), pp. 121 146. references figure 3. tracking error figure 4. control signals figure 5. control signals with h = 0.01 without saturation 74 the journal of engineering research vol. 3, no. 1 (2006) 69-74 appendix by using the predicted tracking error given in eq. (8), the cost function can be rewritten: )t(ur)t(u 2 1 d)t(eq)t(e 2 1 )ht(eq)ht(e 2 1 j t h 0 2 t 1 t + ττ+τ++++= ∫ [ ]t)t(u)x(w)h(ë)h,t(d)h,x(ze 2 1 j +−+= [ ])t(u)x(w)h(ë)h,t(d)h,x(zeq1 +−+ [ ]∫ τ+τ−τ++ h 0 t)t(u)x(w)(ë),t(d),x(ze 2 1 [ ] ττ+τ−τ+ d)t(u)x(w)(ë),t(d),x(zeq 2 (t)ur(t)u 2 1 t+ ( ) ( ) .ur)x(w d)(eq)(e)h(eq)h(e)x(u 2 1 d),t(d),x(zq)(e )h,t(d)h,x(zq )h(eedq)(eq)h(e )x(wu)h,x,e(mj h 0 2 t 1 ttt h 0 1 t 1 th 0 2 t 1 t tt ⎥⎦ ⎤+ ⎢ ⎣ ⎡ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ τττ++ ⎥ ⎦ ⎤ ττ−ττ+ − ⎢ ⎣ ⎡ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ττ+ += ∫ ∫ ∫ &&&&&&&& && &&&&&& ( ) ( ))h,x,x(v)t(e)h(k)x(w r)x(w)h(a ~ )x(wu ref t 1t op + +−= − ( ) ( )dt)t,t(d)t,x(zq)t(e )h,t(d)h,x(zq )h(e)h,x,x(v h 0 2 t 1 t ref −+ − = ∫ && && where m(e,x,g) represents terms that are independent of u(t). in compact form: ( ) ( )ur)x(w)h(ã)x(wu 2 1 h,x,x(ve)h(k)x(wu)h,x,e(mj tt ref tt ++ ++= where )t(x)t(x)t(e ref−= ; ∫+= h 0 21 t dtq)t(ëq)h(ë)h(k ; ∫+= h 0 2 t 1 t dt)t(ëq)t(ë)h(ëq)h(ë)h(ã ; click here, type the title in |times new roman(14) capitalized first letter the journal of engineering research (tjer) vo. 14, no. 2 (2017) 115-123 doi: 10.24200/tjer.vol.14iss2pp115-123 characterization of mechanical properties of aligned date palm frond fiber-reinforced low density polyethylene k.i. alzebdeh*,a, m. m. nassarb, m.a. al-hadhramia, o. al-aamria, s. al-defaaia and s. al-shuailya a department of mechanical and industrial engineering, college of engineering, sultan qaboos university, oman. b palestine polytechnic university, hebron, state of palestine. received 10 march 2017; accepted 15 august 2017 abstract: in recent decades, natural fibers have received attention of scientists and researchers due to their ecofriendly characteristics that qualify them as potential reinforcement in polymer composites in place of synthetic fibers. in this study, an experimental investigation has been conducted to evaluate the effect of orientation of fibers on mechanical properties of a newly developed bio-composite in which date palm fronds (dpf) are embedded as fibers in low-density polyethylene (ldpe) matrix. three bio-composite sheets with orientations of 0°, 45° and 90°, respectively have been fabricated after the date palm fronds were chemically treated. the fabricated composite specimens are tested under tensile load using universal testing machine (utm) in accordance with the astm d-638 standard. then, a comparison of the experimental results against analytical results is made to examine the accuracy and agreement between the two. an inconsistency in moduli, as was discovered, is attributed to the adhesion quality between the fibers and surrounding matrix. output results help to assess the applicability of such class of bio-composites in real-life applications. the results of tensile strength, young’s modulus, and elongation at break revealed that date palm fronds can be used as reinforcement material in polymer-based composites for low strength applications. keywords: bio-composites; date palm fronds (dpf); low density polyethylene (ldpe); tensile strength. البولي إيثيلني منخفض نخيل املصفوف و املقوى اللياف سعف التوصيف اخلواص امليكانيكية ل الكثافة أي، سعود الشعيلأرفاعي، شذى الأامري، عثمان العأ، حممد أ. احلضرميبنصار ماجد، حممود *ه,أالزبدابراهيم خالد الصديقة للبيئة واليت تؤهلها ها: القت األلياف الطبيعية يف العقود األخرية اهتمام العلماء والباحثني بسبب خصائصامللخص لتكون مقوي حمتمل يف مركبات البوليمر بدال من األلياف االصطناعية. قامت هذه الدراسة بإجراء حبث جترييب لتقييم حشو سعف النخيل فيه من خالل ،تأثري مدى تكيف األلياف مع اخلواص امليكانيكية للمركب احليوي املطور حديثا °90و °45, ° .من املركب احليوي بقياسات الواحإثيلني منخفض الكثافة. مت تركيب ثالث كألياف يف نسيج البولي على التوالي بعد معاجلة سعف النخيل كيميائيا. ومن اختبار عينات املركب املصنعة حتت محل الشد يف آلة االختبار . بعد ذلك مت مقارنة نتائج التجربة بالنتائج التحليلية من أجل دراسة مدى الدقة astm d-638( وفقا ملعايري utmالعاملية ) إىل جودة االتصاق بني األلياف واألنسجة التضارب يرجع معامل القيمة والتطابق بني االثنني. و تبني أن عدم التطابق يف النوع من املركبات احليوية يف واقع احلياة. كما احمليطة بها . وتساعد هذه النتائج يف تقييم مدى امكانية تطبيق هذا أظهرت نتائج قوة محل الشد ومعامل يونغ واستطالة الكسر أنه ميكن استخدام سعف النخيل كمواد حشو يف مركبات تعتمد على البوليمر يف تطبيقات منخفضة القوة. ؛ محل الشد.(ldpe)الكثافة ثيلني منخفض ؛ البولي إي(dpf): املركبات احليوية. سعف النخيلالكلمات املفتاحية * corresponding author’s e-mail: alzebdah@squ.edu.om mailto:alzebdah@squ.edu.om k.i. alzebdeh, m. m. nasser, m.a. al-hadhrami, o. al-aamri, s. al-defaai and s. al-shuaily 116 1. introduction the development of advanced renewable materials to replace the non-renewable and petroleum-based products has aroused the interest of researchers in the last few decades. although natural fiber reinforced composites have been successfully used in various noncritical products, there is still a need to develop similar materials to broaden their industrial applications. two constituents are processed and mixed together; natural fiber and hosting matrix, which can be either non-biodegradable or biodegradable based resin. the mechanical properties of composite materials are dependent on the properties of the two constituent phases. fibers typically have low density, moderate specific strength and modulus, but are often very brittle. generally, fiber distribution, concentration, and orientation are the main factors that influence the overall properties of the fabricated fiber reinforced composites (askeland et al. 2009; callister et al. 2007). besides, the quality of the interface between the dispersed phases and surrounding matrix is a highly critical factor in determining the local and nonlocal behavior of a composite. composite materials can be classified into three main types: particle-reinforced, fiber-reinforced, and structural composites. fibers used as reinforcement make improvements in strength, stiffness, or high-temperature performance in case of metals and polymers material, and improve toughness to ceramics matrix. currently, about 40,000 composite products are in use for an array of applications in diverse sectors of the industry around the world (gohil and shaikh 2010). date palm (phoenix dactylifera l) is one of the oldest cultivated and most valuable fruit trees due to its ritual significance in human societies, health benefits, productive capacity in harsh semi-arid and arid environments. it is also known for the range of subsistence produced from its fruits, fronds, date seeds and other parts of the tree. date palm plantations worldwide are estimated to have over 150 million trees (al-khayri et al. 2015). from each individual tree, 10 to 15 branches are cut down. thus, on average, 35 kg of palm residues are obtained per tree. however, the bulk of the material is discarded as waste. therefore, efficient utilization of this natural resource in fabricating natural fiber composite would have a positive impact on environment and could improve the economic status of rural areas (kocak and mistik 2015; mohanty et al. 2014; wazzan 2006). overall, application of the dpf as reinforcement or filler in polymers is feasible, especially in case of annual pruning wastes, noting an abundance of available biomass of this type (sbiai et al. 2010). natural fibers need to be chemically treated before mixing with the matrix. chemical treatment of fiber helps in removing sticking dirt and improving its physical and chemical characteristics. almaadeed et al. (2013) and shalwan et al. (2016) reported that the use of high naoh concentration can cause deterioration in the fiber strength. also, alawar et al. (2009) concluded that hydrochloric acid treatment resulted in deterioration in its mechanical properties. in contrast, some results revealed that 6% concentration of naoh is the optimum solution for treating date palm fiber to maintain high interfacial adhesion and strength with epoxy matrix (alsaeed et al. 2013). scanning electron microscopy of the composite specimens fracture surfaces indicates that the maleic anhydride grafted polypropylene (mapp) and the treated fibers improved the interfacial interaction between the fiber and the matrix (eslami-farsani 2014). based on the findings, it can be concluded that its physical and mechanical properties are not directly related to increased treatment conditions (taha et al. 2007). the potential use of extracted date palm tree fiber has been investigated based on different aspects and methods (al-kaabi et al. 2005a, 2005b; alrawahi et al. 2015; kocak and mistik 2015; mohanty et al. 2014; taha et al. 2007; wazzan 2006). in general, date palm fibers can be extracted from the stem mesh, midribs, bark and leaves or fronds. also, the particulate or powder can be produced by milling the date palm fruit seeds. studies have shown that date palm fibers provide a viable alternative for exploitation in composite material fabrication and they can serve as a replacement for glass fiber, thus solving associated environmental problems. polymer matrix in natural fiber composites is divided into two main categories; bio-based and characterization of mechanical properties of aligned date palm frond fiber-reinforced low density polyethylene 117 petroleum based. ldpe is a petroleum thermoplastic matrix widely used in packaging applications due to its chemical and corrosion resistance, light weight, good impact resistance, high stiffness and good process ability. however, the usage of ldpe, as a polymer matrix for reinforced composites, contributes to a serious environmental problem due to ldpe non-biodegradable properties. also, ldpe is soft, flexible and inert, thus resists reacting with any other elements. additionally, it possesses a low static charge, so it does not attract dust and dirt. hence, it has a huge potential in fabrication of polymer based composites (al-nasir 2013; fahim et al. 2012; rahman et al. 2012; sarifuddin and ismail 2013). an enhancement in the mechanical properties of the developed biocomposites compared to the virgin ldpe has been observed (fahim et al. 2012; nur et al. 2010; rahman et al. 2012). sarifuddin and ismail (2013) stated that the optimum tensile strength was obtained when 10 wt.% of kenaf fiber loaded into the ldpe. in the case of cannabis/ldpe, it has been found that the best fibers ratio is 5 wt.% to improve the mechanical properties of the developed bio-composites (alnasir 2013). moreover, lower elongation of break compared to pure ldpe occurs. in another study conducted by prasad et al. (2016), the potential use of banana fiber in ldpe matrix was explored. it showed that a composite with composition of 25 wt.% banana fiber is the optimal rate on the basis of biodegradability and mechanical properties. in this study, alkali treated aligned date palm fronds are used as reinforcing fibers in lowdensity polyethylene bio-composites. mechanical properties (modulus of elasticity, tensile strength, and strain at break) were determined via experimental testing at different fiber orientations (0o, 45o, and 90o) at ambient temperature. in addition, analytical calculations of mechanical properties based on micromechanics are presented. 2. methodology 2.1 composite fabrication and physical testing the date palm fronds were manually extracted from khusab palm tree of a local farm. chemical analysis showed the detailed composition of all constituents (table 1). firstly, fronds, as shown in fig. 1 (a), were dried under the sun for a month, and cleaned to remove any contaminants. the extracted fronds were cut into 1250 mm in length. a secondhand treatment was applied using tap water as a preliminary step of the cleaning process. in order to improve morphology structure of fibers, alkali treatment using 1 wt.% of naoh for 2 hours (alawar et al. 2009; taha et al. 2007) was applied. then, 20 wt.% volume of fibers was aligned based on the specified direction in the mold (200mm×200mm×5 mm) using a hand lay-up method. care was taken to achieve a uniform distribution of fibers while being layered in the ldpe matrix. secondly, an electrical oven was used to mix the composite constituents at 300˚c (almaadeed et al. 2013) until the matrix fully encapsulated the fibers as shown in fig.1(b). sheets were manually pressed under 25 kg load for 15 minutes to eliminate any porosity or void formation and maintain a uniform thickness as specified. lastly, the composite sheets were pressed manually again for a day prior to cutting. then, specimens were cut according to the astm d 638 using cnc machine in order to characterize their mechanical properties as shown in fig. 2. the prepared specimens were tested using universal testing machine (utm) at a fixed cross head speed of 5 mm/min, at room temperature 23˚c and humidity 50%. for each direction, three samples were tested at a fixed gage length of 57 mm and average response value was recorded. 2.2. analytical modeling a connection with the classical micromechanics may be established here. our experimental testing table 1. chemical composition of dpl fiber (mohanty et al. 2013). constitutes percentage (%) cellulose 54.75 lignin 15.3 hemi-cellulose 20.00 pectin 1.2 moisture 6.5 ash 1.75 wax 0.50 k.i. alzebdeh, m. m. nasser, m.a. al-hadhrami, o. al-aamri, s. al-defaai and s. al-shuaily 118 (a) (b) figure 1. manually extracted date palm fronds fibers; (a) raw, and (b) encapsulated. (a) (b) figure 2. tensile testing procedure (θ = 90˚); (a) specimen, and (b) testing setup. results can be compared with voigt (rule of mixture) and reuss bounds. for instance, voigt bound as given by eqn. (1), assumes a constant strain and therefore gives upper bound on effective elastic modulus of a composite (e c ). m emv f efv v c ue       (1) while reuss bound assumes a constant stress and gives lower bound on effective elastic modulus as given in eqn. (2). me mv fe fv r c l e       (2) where ef, em are the young’s moduli of fiber and matrix, respectively and vf and vm are the volume fractions of fiber and matrix, respectively. however, it is well known that the hashin-strikman bounds (h-s) (hashim 1963) give more accurate and rigorous values as compared to those of voigt and reuss bounds. characterization of mechanical properties of aligned date palm frond fiber-reinforced low density polyethylene 119 such bounds are given in eqns. (3) & (4) for isotropic elastic modulus of elasticity. (𝐸𝑢 𝑐 )𝐻−𝑆 = 𝐸𝑓 + 𝑉𝑚 1 (𝐸𝑚−𝐸𝑓) + 𝑉𝑓 2𝐸𝑓 (3) (𝐸𝑙 𝑐 )𝐻−𝑆 = 𝐸𝑚 + 𝑉𝑓 1 (𝐸𝑓−𝐸𝑚) + 𝑉𝑚 2𝐸𝑚 (4) where (eu c )h−s and (el c)h−sare the upper and lower h-s bounds on effective elastic modulus of a composite. it is important to mention that all these bounds assume prefect bonding between fibers and surrounding matrix. 3. results and discussion the mechanical properties of both fiber and matrix used in this study are given in table 2. one observation on the tensile properties of both fiber (frond) and matrix (ldpe) is a noticeable mismatch between the two ingredi table 2. mechanical properties of fibers and matrix used. tensile strength (mpa) tensile strain (%) young's modulus (gpa) specific gravity khusab frond 19 4.7 5.4 1.17 neat ldpe 9.6 100 0.4 1.08 figure 3. tensile strength (left) and tensile modulus (right) of the tested bio-composites sheets. figure 4. strain at break of the tested bio-composites sheets. neat ldpe 0 45 90 k.i. alzebdeh, m. m. nasser, m.a. al-hadhrami, o. al-aamri, s. al-defaai and s. al-shuaily 120 -ents can be detected. this may affect the interfacial properties of the interface region and can lead to a degradation in tensile properties of the composites. for all the tested composite specimens, mechanical properties were degraded in relation to those of neat ldpe. for tensile strength, around 14% drop was observed in case of θ = 0o (fig. 3), which is considered to be reasonable as there were several limitations in the fabrication process that contributed to a poor adhesion between fiber and surrounding matrix. at θ = 45o, a slight stronger composite was obtained. as expected, the weakest composite occurred at θ = 90o. young’s modulus follows a similar behavior (fig. 3). a drop of 15% in strain at failure was obtained for composite at θ = 0o. it is almost an equal strain at failure for both specimens at θ = 0o and θ = 45o (fig. 4). recently, alzebdeh et al. (2016) studied the mechanical properties of chopped date palm fronds reinforced ldpe. at 20% volume of date palm fronds, a tensile strength in the range of 5.81 to 7.2 mpa was obtained. therefore, continuous date palm frond reinforcement composites showed slightly higher strength. the recorded stress-strain curves under uniaxial tensile loading are shown in fig. 5. as reflected in the approximate linear region of these curves, young’s moduli are similar in both cases θ = 0o and θ = 45o with the weakest value at θ = 90o. yield for the first two cases occurred at strain equal to 0.3 while for the third case, it occurred at strain around 0.6. ultimate tensile strength of 8.4 mpa was achieved at θ = 45o with higher toughness. analytical modeling results in the whole range of volume fraction of fibers are depicted in fig. 6. again, it is worthwhile to emphasize that these bounds assume perfect bonding between fiber and hosting matrix. with admitted limitations on hand-lay-up method of fabrication, it is clear that this assumption is not well satisfied here. figure 5. stress-strain curves in tension of tested specimens. characterization of mechanical properties of aligned date palm frond fiber-reinforced low density polyethylene 121 figure 6. bounds on effective young’s modulus. table 3. young’s modulus (gpa) as obtained experimentally and analytically. experiment voigt reuss h-s (upper) h-s (lower) 0.25 1.4 0.49 0.99 0.56 at 20% volume fraction and 0o orientation, young’s modulus of 0.25 gpa is obtained, which is less than the corresponding analytical values (table 3). a better controlled fabrication process that ensures a strong adhesion between fibers and hosting matrix will result in a closer matching of experimental values of young’s modulus to those estimated analytically. 4. conclusion the increasing awareness toward the biodegradable natural fibers, as an alternative to synthetic fibers, has generated an interest to develop eco-friendly polymer composite materials. in this study, three sheets of ldpe were fabricated using hand lay-up method in which continuous palm tree fronds were aligned as fibers. based on experimental testing for mechanical properties, a few observations can be highlighted:  in general, mechanical properties have been degraded with a drop of 16% in tensile strength, 37% in tensile modulus and 15% in strain at failure.  this degradation in mechanical properties is attributed to the poor adhesion between fibers and matrix.  45˚ orientation of fiber resulted in higher tensile strength.  a mismatch between the experimental and analytical results occurred due to lack of perfect bond between fiber and matrix. regardless of such drop in mechanical properties of the new developed dpf/ldpe composites, preliminary results indicate that using date palm fronds as fillers in plastics and other types of polymers is promising as these biodegradable fronds are priceless, have abundant source, which is considered as waste. finally, it can be concluded that date palm fronds fibers can be used as “reinforcing” material for low strength applications with cost effectiveness and environmental awareness. conflict of interest the authors declare no conflicts of interest. 0 1 2 3 4 5 6 0 0.2 0.4 0.6 0.8 1 e c ( m p a ) vf voigt reuss hs-u hs-l k.i. alzebdeh, m. m. nasser, m.a. al-hadhrami, o. al-aamri, s. al-defaai and s. al-shuaily 122 funding no funding was received for this research. references al-kaabi k, al-khanbashi a, hammami a (2005a), date palm fibers as polymeric matrix reinforcement: dpf/polyester composite properties. polymer composites, 26(5): 604–613. al-kaabi k, al-khanbashi a, hammami a (2005b), natural fiber reinforced composites from date fibers. the european society for composite materials: 1–8. al-khayri jm, jain sm, johnson dv (2015), date palm genetic resources and utilization. date palm genetic resources and utilization: volume 1: africa and the americas (vol. 1): 318. al-nasir ea (2013), study of the mechanical properties of low density polyethylene composites with cannabis ( hemp fibers ). journal of basrah researches (sciences), 39(3):15–22. alawar a, hamed a, al-kaabi k (2009), characterization of treated date palm tree fiber as composite reinforcement. composites part b: engineering, 40(7): 601–606. almaadeed ma, kahraman r, noorunnisa kp, al-maadeed s (2013), characterization of untreated and treated male and female date palm leaves. materials and design 43: 526–531. alrawahi hh, alzebdeh ki, nassar mma (2015), statistical modeling of mechanical properties of chopped date palm frond reinforced polypropylene composites. in proceeding from the 8th international conference on materials science and engineering , al-sharjah , uae 74–77. alsaeed t, yousif bf, ku h (2013), the potential of using date palm fibers as reinforcement for polymeric composites. materials and design, 43:177–184. alzebdeh k, nassar, m, alrawahi n , al–hinai h (2016), characterization of mechanical properties of date palm fronds reinforced low density polyethylene (ldpe). in proceedings of the asme 2016 international mechanical engineering congress and exposition, phoenix, arizona, usa 1-10. askeland dr, fulay pp, bhattacharya dk (2013), essentials of materials science and engineering, 3nd ed. cl engineering. callister w, rethwisch d, wiley j (2009), materials science and engineering: an introduction. 8th ed., john wiley and sons eslami-farsani r (2015), effect of fiber treatment on the mechanical properties of date palm fiber reinforced pp/epdm composites. advanced composite materials, 24: 27-40. fahim is, elhaggar sm, elayat h (2012), experimental investigation of natural fiber reinforced polymers. materials sciences and applications 3: 59–66. gohil pp, shaikh aa (2010), experimental investigation and micro mechanics assessment for longitudinal elastic modulus in unidirectional cotton-polyester composites. international journal of engineering and technology 2(2): 111–117. hashim z (1963), approach to the theory of the elastic behaviour of multiphase materials. journal of the mechanics and physics of solids 11(42): 127–140. kocak d, mistik si (2015), the use of palm leaf fibers as reinforcements in composites. in book: biofiber reinforcements in composite materials 273-281 mohanty j, das s, das h, mahanta t, ghadei s (2014), solid particle erosion of date palm leaf fiber reinforced polyvinyl alcohol composites. advances in tribology, 2014:1-8. mohanty jr, das sn, das hc, swain sk (2013), effective mechanical properties of polyvinylalcohol biocomposites with reinforcement of date palm leaf fibers. polymer composites 34(6): 1548–569. nur hp, hossain ma, mollah mm (2010), preparation of polymer composites using natural fiber and their physicomechanical properties. bangladesh journal of scientific and industrial research 45(2): 117–122. prasad n, agarwal vk, sinha s (2016), banana fiber reinforced low-density polyethylene composites: effect of chemical treatment and compatibilizer addition. iranian polymer journal 25: 229–241. rahman wa, sudin sna, din sn (2012), physical and mechanical properties of pandanus amaryllifolius fiber reinforced low density polyethylene composite for packaging application. shuser 2012 2012 ieee symposium on humanities, science and engineering research (9002): 345–349. sarifuddin n, ismail h (2013), comparative study on the effect of bentonite or feldspar filled low-density polyethylene / thermoplastic sago starch / kenaf core fiber characterization of mechanical properties of aligned date palm frond fiber-reinforced low density polyethylene 123 composites. journal of physical science 8(2): 4238–4257. sbiai a, maazouz a, fleury e, souterneau h, kaddami h (2010), short date palm tree fibers / polyepoxy composites prepared using rtm process: effect of tempo mediated oxydation of the fibers. bioresources 5(2): 672–689. shalwan a, oraby s, alaskari a (2016), mathematical modelling of the interfacial adhesion of date palm/epoxy. journal of materials science research 5(3): 29. taha i, steuernagel l, ziegmann g (2007), optimization of the alkali treatment process of date palm fibers for polymeric composites. composite interfaces 14(7–9): 669–684. wazzan a a (2006), the effect of surface treatment on the strength and adhesion characteristics of phoenix dactylifera-l(date palm) fibers. international journal of polymeric materials 55(7): 485–499. the journal of engineering research (tjer), vol. 15, no. 2 (2018) 114-123 doi: 10.24200/tjer.vol15iss2p114-123 evaluation of actual creep length under heading up structures aprons m.a. el mollaa, n.y. saadb,*, and g.s. ezizahb a civil engineering department, higher technological institute, cairo, egypt. b irrigation and hydraulics department, faculty of engineering, ain shams university, cairo, egypt. received 13 january 2016; accepted 26 november 2017 abstract : in this paper a 2d finite element model (seep2d) is implemented to study the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure. different scenarios for the thickness of the pervious layer under the apron, the length of the apron and the depths of the upstream and downstream sheet piles are studied. results show that assuming the outer and inner faces of the sheet piles have the same weight for estimating the creep length while designing aprons of hydraulic structures is weak. design equations for the actual head loss along the outer and inner faces for both the upstream and downstream sheet piles are driven. these equations can be used as a tool in the practical design for aprons of heading up structures formed on pervious soil and provided with upstream and downstream sheet piles at its ends. keywords: seep2d; seepage; heading-up structures; sheet piles; creep length. اهلياكل فوق مئزرتقييم طول الزحف الفعلي ب، غادة س عزيزة ب،*يوسف سعد .نيفني ،أحممد علي املوىل لدراسة الفقدان الفعلي للقمة (seep2d) احملدود الطبقة التحتيةيتم تطبيق منوذج لعنصر البحثية،يف هذه الورقة :امللخص سيناريوهات خمتلفة لسمك الطبقة دراسة ويتمالعلوية اهلياكلاملتواجدة على طول أكوام االطبقات املثبتة يف نهايات مئزر النتائج أن االفرتاض بأن روتظه .والسفليةوطول املئزر وكذلك أعماق أكوام الطبقات العلوية املئزر،القابلة لالخرتاق حتت طول الزحف املقدر عندما يكون تصميم مآزر اهلياكل نفس وزناألوجه اخلارجية والداخلية ألكوام الطبقات هلا للقمة على طول األوجه اخلارجية والداخلية للخسارة الفعليةويتم عمل معادالت التصميم .ضعيفهو إفرتاض اهليدروليكية هذه املعادالت كأداة يف التصميم العملي ملآزر ترتكز على ستخداما والسفلية. وميكنلكل من أكوام الطبقات العلوية .نهاياتها العلوية والسفلية يفومزودة بأكوام من الطبقات لالخرتاق،هياكل متكونة على تربة قابلة .طول الزحف ،مرتاكمةطبقات ،هياكل علوية ،تسرب :الكلمات املفتاحية * corresponding author’s email: neveen_yousif@hotmail.com mailto:neveen_yousif@hotmail.com m.a. el molla, n.y. saad, and g.s. ezizah 115 notations the following symbols are used in this paper. d1 depth of upstream sheet pile (l); d2 depth of downstream sheet pile (l); ho1 head lost along the outer face of upstream sheet pile (l); ho2 head lost along the outer face of downstream sheet pile (l); hi1 head lost along the inner face of upstream sheet pile (l); hi2 head lost along the inner face of downstream sheet pile (l); g gravitational acceleration (l/t2); h head difference between upstream and downstream the apron (l); k hydraulic conductivity of the homogeneous pervious stratum (l/t); l horizontal distance between the upstream and downstream sheet piles (l); p1 head at point (1) (l); p2 head at point (2) (l); p3 head at point (3) (l); p4 head at point (4) (l); p5 head at point (5) (l); p6 head at point (6) (l); t thickness of pervious stratum under the apron (l); ρ density of seeping water (m/l3); evaluation of actual creep length under heading up structures aprons 116 1. introduction seepage is one of the most important factors that affects the stability of a heading-up structure. using sheet piles under the aprons of heading-up structures increases the percolation length, decreases the hydraulic gradient, and hence provides more safety against piping and uplift under the apron of a structure. figure (1) represents the concept of creep length and hydraulic gradient diagram. numerical methods (finite element, finite difference, boundary element and total variation diminishing method ) can be used to solve seepage problems with different degrees of complexities and accuracies harr (1962); serge leliavsky (1965); u.s. army corps of engineers (1986). ahmed and ellboudy (2010) carried out a number of numerical analysis to study the influence of different sheet pile configurations on the seepage losses, the uplift force on the apron, and the exit gradient at the end of the apron. they concluded that driving a sheet pile under a hydraulic structure that surrounds the downstream floor apron from all sides, has greatly reduced the exit gradient at the end toe of the floor. however, this was accompanied with some increase in the uplift force. ahmed (2011) studied different sheet piles configurations and flow through canal banks. kamble et al. (2014) described the potential of different approaches such as geological and geotechnical methods, dam instrumentation, geophysical methods, tracer techniques, nuclear logging and mathematical modeling for monitoring, detecting and analyzing seepage in hydraulic structures. three dimensional modeling techniques (3d) have recently been attempted to study seepage beneath and around hydraulic structures for different purposes. eftekhar and barani (2013) used seep3d to pinpoint the best locations of cutoff walls. they revealed that a cutoff wall installed at the lower toe of the structure eliminates piping and the one located at the upper toe reduces the uplift pressure. guerra et al. (2012) suggested that complex junctions in levee systems (i.e. 90 degrees bend in the levee alignment as well as a box culvert through the levee) should be studied using 3d models. koltuk and iyisan (2013) compared between the 2d and 3d modeling techniques in studying a rectangular shaped cofferdam taking into account the vertical anisotropy of the permea bility of soil layers. he concluded that the difference between the values obtained from 2d and 3d analysis decreased with increasing the length to width ratio of the cofferdam. gad et al. (2016) conclude that 3d simulations are essential in studying seepage under large dams especially in complicated 3d configurations and lateral heterogeneity. el tahan and elmolla (2013) used the electric analogue method to investigate models of hydraulic structures' aprons provided with two equal cutoffs. they investigated the efficiency of both faces (outer and inner) of these cutoffs. they concluded that the estimation of the percolation length under the hydraulic structure must consider a weighted length factor for various faces of the cut-offs. seep2d is a finite element program that has been applied in many researches to study seepage and has proved to be an efficient tool for seepage analysis (ozkan 2003; noori and ismaeel 2011; anas 2012; aboulatta et al. 2014). elmolla (2001) used seep2d to investigate seepage under the aprons of heading up structures provided with a single cut-off. in this paper, seep2d is used to investigate the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure formed on a pervious soil. equations of the relation between head loss along both faces (outer and inner) of upstream (u.s.) and downstream (d.s.) sheet piles and the involved variables will be driven to be used in creep length designing purpose which, to the best of the authors’ knowledge, has not been covered in the literature yet. figure 1. line of creep and hydraulic gradient diagram (after serge leliavsky, 1965). m.a. el molla, n.y. saad, and g.s. ezizah 117 in this study sensitivity analysis for the variables involved in the problem as well as different scenarios for the thickness of pervious layer under the apron, the length of the apron, the depths of the upstream and downstream sheet piles, and the hydraulic conductivity are studied. 2. dimensional analysis in the present study, all the variables involved in the problem can be expressed as (refer to fig. (2)): f (h, d1,d2,l,t,k,r,g,p1,p2,p3,p4, p5,p6 = 0 (1) where: h = head difference between upstream and downstream the apron; d1 & d2 = depth of upstream and downstream sheet pile respectively; l = horizontal distance between the upstream and downstream sheet piles; t = thickness of pervious stratum under the apron; k= hydraulic conductivity of the homogeneous pervious stratum under the apron; ρ= density of seeping water; g = gravitational acceleration; p1, p2, p3, p4, p5, p6 = head at points(1,2,3,4,5,6). applying buckingham's π theorem, the relationship between the above different variables can be presented as follow: 0) h h , h h , h*g k , d t , t l , d d ( 2i 2o 1i 1o 22 1  (2) where 32 21 1 1 pp pp h h i o    (3) and 54 65 2 2 pp pp h h i o    (4) ho1/ hi1 = the ratio between the head loss along the outer face and the inner face of the u.s sheet pile, and ho2/hi2 = the ratio between the head loss along the outer face and the inner face of the d.s sheet pile. 3. description of the model the seep2d software was developed by the united states army engineer waterways experiment station to model a variety of problems involving seepage. the governing equation used in this model is the laplace equation. laplace equation is the governing differential equation for two dimensional, steady state incompressible, isotropic flow in the xy plane which is 0 y h x h 2 2 2 2       (5) where h is the head at any point in the flow domain. transient or time varying problems and unconfined plan view (aerial ) models cannot be modelled using seep2d. seep2d allows for different hydraulic conductivities along the major and minor axes (anisoropic conditions) to be defined. heterogeneous models can be created by specifying different values of hydraulic conductivity for the elements representing the different layers or regions. post-processing includes contouring of the total head (equipotential lines), drawing flow vectors, and computing flow potential values at the nodes. these values can be used to plot flow lines together with the equipotential lines (i.e. flow nets). the phreatic surface can also be displayed (seep2d primer 1998). figure 2. the variables involved in the problem. evaluation of actual creep length under heading up structures aprons 118 in a typical modelling problem involving the seep2d software, a series of tasks are performed in a specific sequence as follows: 1. mesh generation 2. setting boundary conditions; 3. seep2d execution; and 4. postprocessing the output. 4. model application the dimensions for different variables as recommended from a previous study ezizah et al. (2000) are as follows: d1 = (0.2 to 0.6)*l, d2/d1 = (0.7 to 0.2) , h/l = (1/2, 1/3, 1/6, 1/8) , d1/t = (0.1 to 0.5) or t= 1.5l. the model consists of a heading up structure formed on a pervious homogeneous isotropic soil layer with thickness t (t=15, 22.5,30, 60 m). two values of head difference (h) between u.s and d.s the structure are considered (h=3, 7 m). the apron of the structure is of length l (l=10, 15, 20 m) and provided with two sheet piles at its ends driven to depths d1 and d2, where d1 is the depth of the u.s sheet pile (d1= 3,4,5,6 m) and d2 is the depth of the d.s sheet pile (d2=1.5, 2.5 m). figure (3) shows a finite element mesh used for one of the simulations. figure (4) shows a sample of flow net obtained from seep2d (seep2d output). 4.1 boundary conditions all external edges of the problem were modelled as impermeable boundaries including the upstream and downstream edges and the bottom of the problem. the apron of the hydraulic structure was also modelled as impermeable boundary. all upstream nodes of the structure in the canal bed were allocated a prescribed head (h) (h=3 or 7 m). nodes in the canal bed downstream of the structure were assigned a prescribed head value of zero. 4.2 calibration and verification of the model experimental readings of a previous study el tahan and elmolla (2013) were used to calibrate the model to choose a reasonable cell size and check that the input data were entering correctly. three different mesh cell sizes were figure 3. sample of seep2d mesh. figure 4. sample of flow net obtained from seep2d (seep2d output). m.a. el molla, n.y. saad, and g.s. ezizah 119 used in order to calibrate the model in an attempt to match the experimental data. then mesh would be generated as smaller mesh cell sizes were chosen around and between the two sheet piles in order to accurately show the changes in head when using sheet piles. the experimental study includes two equal sheet piles under the apron of the heading up structure. two values of h were used h= 3, 6 m, l was taken 15 and 20m while d1 and d2 were taken 5m. figure (5) illustrates a sample of the model calibration using the chosen best cell size, which represents the head readings at points p1 to p6. it is obvious that there is a good matching between the experimental readings and the numerical model. using the calibrated mesh size, the model is verified by another set of electric analogue experimental readings (aboulatta et al. 2010). the length of apron l= 19m, the depth of the u.s and d.s sheet pile d1= 5m, d2= 2.5m. two values for h (3, 7.5m) were used. figure (6) illustrates a sample of the model verification, which represents the head readings at points p1 to p6. it is obvious that there is a good matching between the experimental readings and the numerical model. 4.3 sensitivity analysis sensitivity analysis was carried out in order to find the most effective parameters for different variables. from the sensitivity analysis it is noticed that the variables which have the biggest effect on the head loss ratio along the upstream sheet pile are the thickness of pervious soil layer under the apron (t) and the depth of u.s sheet pile (d1). for the d.s sheet pile, it is noticed that the thickness of pervious soil layer under the apron (t) and the depth of downstream sheet pile (d2) have the biggest effect on the results. moreover, the head difference h and the hydraulic conductivity k have negligible effect on the ratio of the head lost along the faces for the both sheet piles. figure 5. sample of model calibration h= 6m, l= 20m, d1 & d2= 5m. figure 6. sample of model verification h= 7.5 m, l= 19 m, d1 = 5m, d2= 2.5m. evaluation of actual creep length under heading up structures aprons 120 figures (7) and (8) show the effect of each variable on the ratio between the head loss along the outer face and the inner face of u.s and d.s sheet piles respectively (ho/ hi). 5. results analysis the readings obtained from seep 2d are investigated and analyzed in order to determine the effect of changing the length of apron, the thickness of soil layer and the depths of upstream and downstream sheet piles on the ratio between the head loss along the outer and the inner faces for both the upstream and downstream sheet piles. plotting the head loss ratio along the outer face and the inner face for the u.s sheet pile ((ho1/ hi1) against the depths ratio of u.s and the d.s sheet pile (d1/d2) for different values of (t/d2) is shown in fig. (9). it is noticed that (ho1/ hi1) decreases with the increase of the ratio between (d1/d2) for all (t/d2) values. so, the efficiency of the inner face of the u.s sheet pile improves by 8-16% with increasing the ratio d1/d2. also, it is obvious that the ratio of (t/d2) is directly proportional to (ho1/ hi1). that means the efficiency of the inner face of u.s sheet pile improves by decreasing the values of t/d2. moreover, changing t/d2 affects ho1/ hi1 values for bigger values of d1/d2 than smaller ones. from fig. 9, it was found that the best relation between (ho1/hi1) and (d1/d2) is given by linear equation with correlation equal r (r measures the strength of the relationship between variables): figure 7. effect of each variable on the ratio (ho1/hi1). figure 8. effect of each variable on the ratio (ho2/hi2). m.a. el molla, n.y. saad, and g.s. ezizah 121 2 1 1 1 d d ba h h i o  (6) where a, b are coefficients depending on t/d2. the value of a and b are given in the table 1. so, the general empirical head loss equation along the u.s. sheet pile could be presented as follows: 2 1 2 2 1 21i 1o d d 184.0 d t 006.0 d d d t 005.018.2 h h   (7) plotting the ratio of the head loss along the outer and the inner face for the d.s. sheet pile ((ho2/ hi2) against the depths ratio of u.s. and the d.s. sheet pile (d1/d2) for different values of (l/t) is shown in (fig. 10). it is noticed that ((ho2/ hi2) decreases with the decrease of (d1/d2) for all values of (l/t). that means the efficiency of the inner face of the d.s. sheet pile improves by 1-3% with decreasing the ratio (d1/d2). moreover, comparing the ho2/ hi2 values at the same depth ratio values, we can conclude that as (l/t) values increase the ((ho2/ hi2) values decrease by 2-4%. from fig. 10, it was found that the best relation between (ho2/hi2) and (d1/d2) is given by linear equation with correlation equal r. figure 9. design chart for the ratio between the head lost along the outer and the inner faces of u.s sheet pile. figure 10. design chart for the ratio between the head lost along the outer and the inner faces of d.s sheet pile. table 1. the value of a and b constants. t/d2 6 9 10 12 15 20 a -0.16 -0.14 -0.13 -0.1 -0.11 -0.09 b 2.2 2.26 2.24 2.24 2.3 2.28 r 0.92 0.96 0.97 0.99 0.96 0.98 evaluation of actual creep length under heading up structures aprons 122 2 1 2 2 d d ec h h i o  (8) where c, e are coefficients depending on l/t. the value of c and e are given in the table (2). so, the general empirical head loss equation along the d.s sheet pile could be presented as follows: 2 1 2 1 2i 2o d d 016.0 t l 181.0 d d t l 027.0251.2 h h   (9) for the studied variables range, the ratio (ho1/ hi1) for the upstream sheet pile ranges from 1.55 to 2.15, and the ratio (ho2/hi2) for the downstream sheet pile ranges from 2.08 to 2.3. 6. conclusion seep2d is implemented to study the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure. the following main conclusions may be drawn:  for the u.s. sheet pile, the variables, which have the biggest effect on the head loss ratio along both of its faces, are the thickness of pervious soil layer (t) and the depth of u.s. sheet pile (d1).  for the d.s. sheet pile, the thickness of pervious soil layer (t), the apron length (l) and the depth of d.s. sheet pile (d2) have the biggest effect on the head loss ratio.  the outer and inner faces of the sheet piles do not have the same weight for estimating the creep length. actually, the creep length along the inner face of the u.s. pile is about 55% of its supposed value and for the d.s. pile is about 45% of its supposed value for most of the tests ranges.  increasing the sheet pile depth ratio (d1/d2=4) improves the efficiency of the inner face of the u.s. sheet pile by 816%.  the improvement of the d.s. sheet pile efficiency by decreasing d1/d2 or increasing l/t is small. so, it is better to increase the sheet pile depth ratio to improve the efficiency of the u.s. sheet pile, although it undermines the downstream sheet pile efficiency by small amount.  design equations for the actual head loss along the outer and inner faces for both the upstream and downstream sheet piles are driven. these equations may be used as a tool in practical design for aprons of heading up structures formed on pervious soil and provided with upstream and downstream sheet piles at the ends of the apron. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references aboulatta n, ezizah g, yousif n, anas d (2010), study of the efficiency of downstream blanket in heading-up structures. nile basin water science and engineering 3(1): 79-89. aboulatta n, ezizah g, el molla d (2014), modelling seepage effects in heterogeneous soil under heading-up structures using an experimental and numerical methodology. ph.d. thesis, faculty of engineering, ain shams university, cairo, egypt. ahmed aa, ellboudy am (2010), effect of sheet pile configuration on seepage beneath hydraulic structures. asce, international conference on scour and erosion (icse-5) san francisco, california, united states. ahmed aa (2011), design of hydraulic structures considering different sheet pile configurations and flow through canal banks. computers and geotechnics 38(4): 559-565. anas d, gad ma, samy g, aboulatta n (2012), a comparative analysis between electric and numerical seepage modeling techniques. the seventh conference on environmental hydrology, american society of civil engineers (asce), cairo, egypt. eftekhar, afzali s, barani ga (2013), application of the seep 3d software in m.a. el molla, n.y. saad , and g.s. ezizah 123 modeling the effects of cutoff walls on the reduction of the seepage pressure and exit hydraulic gradient under hydraulic structure. water engineering journal 5(15): 79-90. el molla am (2001), new trend for evaluating the percolation length under aprons of hydraulic structures provided with cut-off and founded on isotropic soil. cerm, faculty of engineering, al -azhar university 23(1). el tahan am, el-molla am (2013), effects of cut-offs depth ratio on uplift and efficiency of front and back faces of upstream cut-off. 21st csce canadian hydrotechnical conference, alberta, canada. ezizah gms, el-molla am, abdellateef m, and el-niazy hammad m (2000), seepage under hydraulic structures founded on complex formations. ph.d. thesis, faculty of engineering, ain shams university, cairo, egypt (this for ratios). gad ma, anas d, atta na, samy g (2016), worth of the 3d simulation of seepage in the vicinity of heading-up structures. ksce journal of civil engineering 20 (2): 679-686. guerra p, priestley d, fredlund m (2012), 3d groundwater seepage analysis of a levee intersection. canadian dam association (cda) annual conference. harr me (1962), groundwater and seepage. mcgrawhill, new york. kamble rk, muralidhar b, hanumanthappa ms, patil av, edlabadkar js (2014), multiple approaches to analyse and control seepage in hydraulic structures", ish journal of hydraulic engineering, 20(1): 7–13. koltuk s, iyisan r (2013), numerical analysis of groundwater flow through a rectangular cofferdam. electronic journal of geotechnical engineering (ejge), 18. noori bma, ismaeel khs (2011), evaluation of seepage and stability of duhok dam. alrafidain engineering journal 19(1). ozkan s (2003), analytical study on flood induced seepage under river levees. ph.d dissertation, louisiana state university, the department of civil and environmental engineering. seep2d primer (1998), brigham young university, engineering computer graphics laboratory. serge leliavsky (1965), design of dams for percolation and erosion. volume iii, oxford and ibh publishing co. u.s. army corps of engineers, washington dc (1986), seepage analysis and control for dams. engineer manual 1110-1901. chen050427.qxd the journal of engineering research vol. 3, no. 1 (2006) 38-42 1. introduction research interest in catalytic hydrodesulfurization (hds) of petroleum feedstock has been increasing recently. this is partly due to the necessity of processing feedstocks that have large amounts of sulfur-containing compounds and also as a result of more stringent environmental regulations with respect to emission of sulfur-containing gases to the atmosphere. sulfur removal has economic benefits of prolonging catalyst life and decreasing corrosion of the process equipment downstream. perhaps one of the most important factors that spurred interest in hds is a recent review of the allowable sulfur content of fuels, flue gases and other petroleum products, in the u.s. ______________________________________ *corresponding author's e-mail: baba@squ.edu.om the environmental protection agency (usepa) guidelines, for instance, restrict the amount of sulfur allowed in diesel fuels to 1.3 x 10-2 g/dl3 by the year 2010 (usa-epa, 2001). this requires reduction of about 97% from the current level. achieving this limit will require a major improvement or redesign of the catalysts employed for hds. therefore, it is necessary to gain further understanding of the chemical reactions involved. this could be achieved by studies of kinetics and networks of the reactions. dibenzothiophene (dbt), as a typical sulfur-containing compound in the petroleum feedstock, has been studied extensively. a recent study has shown that at certain conditions, the dbt kinetic closely resembles that of an overall behavior of dbts in a light oil fraction (steiner, et al. 2002). aspects of the hds of dbt kinetics have kinetics of hydrodesulfurization of dibenzothiophene on sulfided commercial co-mo/γ-al2o3 catalyst y.s. al-zeghayer1 and b.y. jibril*2 1chemical engineering department, king saud university, p.o. box 800, riyadh 11421, saudi arabia 2petroleum and chemical engineering department, sultan qaboos university, p.o. box 33, al-khoud, pc 123, muscat, oman received 27 april 2005; accepted 17 september 2005 abstract: kinetics of hydrodesulfurization of dibenzothiophene (dbt) has been studied on a commercial como/γ-al2o3 catalyst at 633 683 k and 10 atm. a low dbt concentration typically obtained in hydrodesulfurization operations was used. pseudo-first-order model was found to fit the experimental data for the consumption of dbt. the activation energy for the conversion of dbt was found to be 51.7 kcal/mol. biphenyl (bp) and cyclohexylbenzene (chb) were obtained as dominant products. for the reaction network, both parallel and parallel-sequential routes were explored. the latter was found to give a better description of the bp and chb distributions. the ratio of bp to chb depended on the reaction temperature. the values of activation energies of dbt hydrogenolysis to bp (ebp), dbt hydrogenation to chb (echb1) and hydrogenation of bp to chb (echb2) were found to be in a decreasing order of echb2 > ebp > echb1. the result suggests the presence of different catalytic sites leading to the two products on the catalysts. keywords: hydrodesulfurization, dibenzothiophene, como/al2o3, biphenyl, cyclohexylbenzene, kinetics ioéÿ á«æ«lh~«¡dg âjèµdg ádgpg äóyéøj á«cômdibenzothiophene…qéœ »àjèc …õøm πeéy ≈∏y ògõdg .¢s.…1πjèl .….ü h2@ á°uóÿg ioéÿ á«æ«lhq~«¡dg âjèµdg ádgpeg äóyéøj áñ«côj :dibenzothiopheneúh ée ájqgôm äélqo ~æy …qéœ »àjèc …õøm πeéy ≈∏y é«∏ª©e â°sqho633 ¤g683á≤∏£e álqo ~æy h10 ioéÿ áø«©°v ägõ«côj .…ƒl §¨°vdbtá«∏ª©ÿg èféàædg §ñæà°ùj ¿g øµá áhpéµdg ¤h’g álq~dg øe »°véjq êpƒ‰ ¿g âñkg .âe~îà°sg á«æ«lhq~«¡dg âjèµdg ádgpcg á«∏ªy øe é¡«∏y π°ué◊g io颢ÿdbt ioée π˘jƒ˘ë˘à˘d á˘£˘°ûæÿg á˘bé˘£˘dg .ᢵ˘∏˘¡˘à˘°ùÿgdbt…h颰ùj é˘¡˘fg ~˘˘lhkcal/mol51.7,,~l é˘ª˘g é˘¡˘«˘∏˘y ∫ƒ˘°üùg ᢫˘˘°ù«˘˘fô˘˘dg äé˘˘é˘˘à˘˘æÿg ¿g(chb) biphenyl (bp) and cyclohexylbenzene»£©j á©héààÿg ájrgƒàÿg äóyéøàdg ¿g ~lhh .âhôl á©héààÿg ájrgƒàÿgh …rgƒàÿg äóyéøàdg éªgh äóyéøàdg áµñ°t øe úækg ¿gh . újoée ™jrƒàd π° ag ∞°uhchbbp and ioée áñ°ùf ¿gh .bpioée ¤gchb ioée πjƒëàd ᣰûæÿg ábé£dg º«b ¿g .iqgô◊g álqo ≈∏y ~ªà©jdbt ioée ¤g) bpbp(e ¤g ioéÿg ¢ùøf πjƒ–hchb)chb1(e ioée πjƒ–hbp ¤g)chb2chb (e ‹réæj πµ°t ‘ 顪«b i~lhchb1< ebp< echb2e.èféàædg ¿gh õø◊g πª©dg ≈∏y újoée êéàf’ äoéb iõø◊g πeé©dg ≈∏y áø∏àfl ™bgƒe oƒlh ìî≤j.i áá««mmééààøøÿÿgg ääggooôôøøÿÿgg úlhq~«¡dg ᣰsgƒh âjèµdg ádgreg äóyéøjáñ«côj :,3o2 como/a1,cyclohexylbenzen, biphenyl,.…qéœ »àjèc …õøm πeéy 39 the journal of engineering research vol. 3, no. 1 (2006) 38-42 been reported. the reaction was shown to follow a pseudo-first-order kinetics and both parallel and consecutive reaction networks were identified (girgis, et al. 1991; farag, et al. 1997; wang, et al. 2004; pille, et al. 1994; broderick, et al. 1981). when the reaction of dbt was tested on como/mcm-41, a pseudo-first-order kinetic model was found to represent the experimental data better than a langmiur-hinshelwood model. this is a typical observation for low dbt concentrations. it has also been shown that the relative significance of hydrogenolysis and hydrogenation routes depended on co/mo ratios (wang, et al. 2004). furthermore, the two routes could also be substantially affected by the presence of naphthalene or hydrogen sulfide for the hds of dbt or 4,6dimethyldibenzothiophene (farag, et al. 1999). sulfided promoted como/al2o3 and other transition metal catalysts have been widely researched and shown to be effective for the reaction (topsoe, et al. 1984; prins, et al. 1989; yang, et al. 2002; venezia, et al. 2002; damyanova, et al. 2003; papadopoulou, et al. 2003; alzeghayer, et al. 2005). further understanding of the nature of interactions between the reactants and catalysts and effects of reaction conditions thereupon is important. therefore, the objective of this work was to describe the kinetics of dbt hydrodesulfurization on a sulfided commercial como/γ-al2o3 catalyst. effects of variation of reaction temperature and feed flow on products were explored. 2. experimental 2.1 materials the materials used for the catalyst testing were dibenzothiophene (dbt), tetralin, decalin and dimethyldisulfide (ch3)2s2 (dmds). all chemicals were of analytical grade quality supplied by aldrich chemical company. a commercial catalyst was obtained from a supplier. its composition and physical properties as given by the manufacturer are shown in table 1. the catalyst was sulfided to a catalytically active form using a solution of dimethyldisulfide (2 wt% sulfur equivalent) in decalin solvent at 623 k until h2s breakthrough was observed. 2.2 catalyst testing the catalysts evaluation was carried out by passing solutions containing dibenzothiophene at a concentration of one or two weight percent in the hydrogen donor solvent (tetralin) through a fixed bed containing 1 g of the catalyst particles. the catalyst particle sizes and reaction conditions were chosen to minimize mass transfer limitations, based on separate preliminary experiments. the use of a hydrogen donor solvent ensured an excess supply of hydrogen readily available for the reaction without a need for adding gas phase hydrogen. the reaction was conducted at temperature range of 633 683 k, pressure of 10 mpa, and flowrates required to give measurable conversions were between 0.5 and 2 g/min. details on catalyst testing and gc analysis were reported earlier (alzeghayer, 2005). 3. results and discussion the major products identified were biphenyl (bp) and cyclohexylbenzene (chb). negligible amounts of bicyclohexyl were detected at high conversions. the conversion and products distribution over the catalyst are shown in fig. 1. the figure shows variation in dbt conversion and bp and chb selectivities with inverse of weight hourly space velocity (whsv) from 0 250 g catalyst h/g feed. the effect of hydrogen concentration on the reaction was neglected as it was considered to be in excess. the dbt degree of conversion increased rapidly from 0 to 92% in the inverse whsv range. at the initial conversion, bp appeared to be an exclusive product indicating that it is a primary product. at the highest conversion, its selectivity decreased to 80% with a corresponding increase in the selectivity to chb from 0 to 20%. at certain conditions, the kinetics of hds of components wt% dry basis molybdenum oxide (moo 3) 15.40 cobalt oxide (coo 3.20 sodium oxide (na 2o) 0.03 iron (fe) 0.03 sulphate (so 4) 0.30 silicon dioxide (sio 2) 0.10 physical properties poured bulk density, lb/ft 3 33.0 compacted bulk density, lb/ft 3 36.0 crush strength, lb/m m 3.4 surface area, m 2/gm 310.0 pore volume, cm 3/gm 0.8 table 1. chemical and physical properties of the commercial catalyst c on ve rs io n (% ) an d se le ct iv it y (% ) (1/whsv)*1000 (g cat. h/g feed) figure 1. variation of dbt conversion and products distribution with 1/whsv 40 the journal of engineering research vol. 3, no. 1 (2006) 38-42 dibenzothiophene (dbt) was shown to represent that of an aggregate of dbts in a petroleum feedstock (steiner, 2002). in an earlier study of kinetics of hds of dbt in liquid feed, the data were well fitted to a langmuirhinshelwood (l-h) model in a plug-flow reactor (singhal, et al. 1981). however, at a lower dbt concentration, as expected, a first order relation between the rate of consumption of dbt and concentration was obtained. it has been shown that for an analysis of the reaction network, even simple power law could give satisfactory results (levenspiel, 1972). therefore, for the catalyst under study, the kinetic data for the reaction were evaluated assuming integral plug-flow reactor. a logarithmic plot of mole of dbt reacted and the inverse of whsv showed a linear relation (fig. 2). this indicates a pseudofirst order kinetic similar to earlier reports (wang, et al. 2004; farag et al. 1999; singhal, et al. 1982). however, at a low temperature, dbt adsorption may be important despite the low concentration that was employed. the deviation of the data from the pseudo-first order for a low feed flow or high contact time may be due to adsorption of dbt. at a low temperature, such adsorption and species diffusion may be important. therefore, the data may deviate from a first-order rate expression. the distribution of the products and reaction network were explored by changing the reaction temperature and feed flow rate (fig. 3). both show a similar trend of change in chb/bp ratio with dbt conversions. but for all conversion levels, bp was the dominant product; contrary to an earlier work on a similar catalyst (singhal, et al. 1981; broderick, 1980). many workers have proposed a reaction network similar to scheme i in fig. 4. the products observed were biphenyl and cyclohexylbenzene. in addition, the presence of tetrahydrodibenzothiophene and hexahydrodibenzothiophene were also reported (broderick, 1980). other reports suggested their presence as undetected, highly reactive intermediates (rollmann, 1977). biphenyl was observed to react to give mainly cyclohexylbenzene (singhal, et al. 1982; broderick, 1980). it was concluded that the rate of reaction of biphenyl was slow in comparison with the disappearance of dibenzothiophene. hydrogenation of cyclohexylbenzene to give bicyclohexyl was too slow to be measurable under typical hydrodesulfurization conditions (broderick, 1980). the obvious path for further reaction of biphenyl is a series of hydrogenations, giving first cyclohexylbenzene and then bicyclohexyl. by comparing the relation between product ratios and conversions at constant temperature and also with increasing temperature, we consider that both parallel and sequential paths may be involved in the reaction network as suggested in scheme i. perhaps partially hydrogenated dibenzothiophene intermediates were formed. if the rate of hydrodesulfurization of the intermediates (path 3') were very fast, the selectivity profiles would follow both parallel and sequential (complex) path. if the rate were slower, that would have hindered the appearance of cyclohexylbenzene in the early stages of the reaction; but contributed to its appearance by sequential reactions. the reaction network in scheme i was used to further lo g( m ol e of d b t re ac te d/ g fe ed ) (1/whsv)*1000 g cat.h/g feed figure 2. logarithmic variation of mole of dbt converted versus 1/whsv effect of temperature effect of flow rate p ro du ct r at io ( c h b /b p ) dbt conversion (%) figure 3. variation of chb/bp ratio with dbt conversion 0.28 0.24 0.20 0.16 0 20 40 60 80 100 0.00 0.01 0.02 0.03 0.04 0 -1 -2 -3 figure 4. scheme i 41 the journal of engineering research vol. 3, no. 1 (2006) 38-42 study the kinetics of the reaction. the overall first-order rate constants and the activation energies were determined from the experimental data. to measure these parameters, it was assumed that at low conversion, the reaction of biphenyl could be neglected compared to its rate of appearance and that cyclohexylbenzene formed directly from dibenzothiophene (sapre, et al. 1980). arrhenius plot for the steady state consumption of dbt is shown in fig. 5. the activation energy for the disappearance of dbt over the catalyst was calculated to be 51.7 kcal/mole as shown in table 2. the same table compares the activation energies with those obtained by others with the exception of the (5.3 kcal/mol) value obtained by bartsch and tanielian (1974); the energies of activation are in good agreement. the low values obtained by bartasch and tanielian indicate a diffusion-controlled reaction. based on the relation between the rate constants for consumption of dbt and that of production of bp and chb, the activation energies (el and e2) have been evaluated from the arrhenius plots. the relation between the product ratio and the conversions obtained assuming the parallel-sequential path was fitted to the experimental data to determine the pseudo first-order rate constant for the reaction of biphenyl to cyclohexylbenzene (k3). the activation energy (e3) was also calculated from the arrhenius relation. the energies have the same order of magnitude as shown in the table 3. the activation energy for the hydrogenolysis of dibenzothiophene to biphenyl is 50.3 kcal/mole. this is insignificantly greater than the activation energy for hydrogenation of dibenzothiophene to cyclohexylbenzene by about 6.15 kcal/mole. this is contrary to an earlier report that showed the rate of bp transformation to chb to be two orders of magnitude slower than that of dbt hydrogenolysis, when the reaction was tested on alumina supported molybdenum carbide (bartsh, et al. 1974). it was reported that at dbt conversions degrees higher than 40%, the selectivity to chb decreased with an increasing temperature (rollmann, 1977). in the low conversion range, the results are in agreement with those of (broderick, 1980). he used the ratio of rate equations to correlate the differential concentration data for the hydrogenation and hydrogenolysis reaction of dibenzothiophene. however, this observation was limited because the study took account of the parallel reaction scheme without further reaction of biphenyl due to the low conversions (less than 15%). the results obtained by singhal et al. favored the parallel and sequential reaction, which became increasingly important at high conversions (singhal, et al. 1981). accordingly, the change in selectivity with temperature was said to reflect the differences in the activation energies. in this work, the results covered the full range of conversions. it shows that the activation energy for hydrogenolysis reaction of dibenzothiophene is greater than that for hydrogenation of dibenzothiophene to cyclohexylbenzene. the values of the activation energies for hydrogenolysis (e1) and hydrogenation (e2) of dibenzothiophene and the activation energy for hydrogenation reaction of biphenyl to cyclohexylbenzene (e3) are in a decreasing order of e3 > el > e2. thus, the product distribution is partly determined by the temperature employed for the reaction. 4. conclusions the pseudo-first-order kinetic model was found to fit the experimental data for the dbt consumption in 5.0 4.0 3.0 2.0 1.0 0.0 lo g e k ( g fe ed /g c at .h ) 1.45 1.50 1.55 1.60 1/t x 103 , k-1 figure 5. arrhenius plot for hds of dbt on the catalyst temp., k chb/bp g feed / g catalyst h k, kcal/mol k k1 k2 k3 e e1 e2 e3 633 653 683 0.000 0.145 0.152 4.03 18.04 83.73 4.03 15.76 72.69 0.00 2.28 11.04 0.00 0.53 5.00 53.1 50.3 44.1 63.8 k=k1+k2 and cchb/cbp = k2/k1 table 3. the pseudo-first-order rate constants and activation energies for reactions in scheme i temperature range, k ea, kcal/mol.k solvent 633-683 558-623 573-723 not stated 548-598 53.1 39.1 36.0 5.3 28.0-32.8 tetralin [this work] tetralin [16] aromatics [19] not stated [21] n-hexane [18] table 2. comparison of apparent activation energies for dbt consumption 42 the journal of engineering research vol. 3, no. 1 (2006) 38-42 hydrodesulfurization on a commercial como/γ-al2o3 catalyst. biphenyl and cyclohexylbenzene were the major products observed. their distributions suggest a reaction network that involves both parallel and sequential routes. the differences in activation energies of the respective reaction leading to biphenyl and cyclohexylbenzene indicate the importance of reaction temperature in determining the products distribution. further, there appeared to be separate sites responsible for production of biphenyl and cyclohexylbenzene. references al-zeghayer, y.s., sunderland, y.s., al-masry, y.w., almubaddel, f., ibrahim, a.a., bhartiya, b.k. and jibril, b.y., 2005, “activity of como/γ-al2o3 as a catalyst in hydro-desulfurization: effects of co/mo ratio and drying condition," appl. catal. a, vol. 282, pp. 1-10. bartsh, r. and tanielian, c., 1974, "hydrodesulfurization: i. hydrogenolysis of benzothiophene and dibenzothiophene over coo-moo3-al2o3 catalyst," j. catal, vol. 35, p. 353. broderick, d.h., 1980, ph.d dissertation, university of delaware, newark, delaware. broderick, d.h. and gates, b.c., 1981, "hydro-desulfurization and hydrogenation of dibenzothiophene catalyzed by sulfided co-mo/γ-al2o3: the reaction kinetics," aichej, vol. 27, p. 663. da costa, p., potvin, c., manoli, j.m., lemberton, j.l., perot, g. and djega-mariadassou, g., 2002, "new catalysts for deep hydrotretment of diesel fuel: kinetics of 4,6-dimethyldibenzothiophene hydrodesulfurization over alumina-supported molybdenum carbide," j. mol. catal. a, vol. 184 p. 323. damyanova, s., petrov, l. and grange, p., 2003, "xps characterization of zirconium-promoted como hydrodesulfurization catalysts," appl. catal. a, vol. 239, p. 241. farag, h., sakanishi, k., mochida, i. and whitehurst, d.d., 1999, "kinetic analyses and inhibition by naphthalene and h2s in hydrodesulfurization of 4,6dimethyldibenzothiophene (4,6-dmdbt) over como-based carbon catalyst," energy and fuel, vol. 13, p. 449. farag, h., whitehurst, d.d., sakanishi, k. and mochida, i., 1997, "catalysis in fuel processing and environmental protection," proc. 214th national meeting of acs sep 7-11, las vegas, usa, am. chem. soc, vol. 42 p. 569. federal register 65 (2000) 35430. girgis, m.j. and gates, b.c., 1991, "reactivities, reaction networks, and kinetics in high-pressure catalytic hydroprocessing," ind. eng. chem. res, vol. 30, p. 2021. levenspiel, o., 1972, chemical reaction engineering, 2nd edn, john wiley & sons, new york. papadopoulou, c., vakros, j., matralis, h.k., kordulis, c. and lycourghiotis, a., 2003, “on the relationship between the preparation method and the physicochemical and catalytic properties of the como/γ-al2o3 hydrodesulfurization catalysts," j. coll. interf. sci, vol. 261, pp. 146. pille, r.c., yu, c. and froment, g.f., 1994, "kinetic study of the hydrogen sulfide effect in the conversion of thiophene on supported co-mo catalysts," j. mol. catal, vol. 94, p. 369. prins, r., de beer, v.h.j. and somorjai, g.a., 1989, "structure and function of the catalyst and the promoter in co-mo hydrodesulfurization catalysts," catal. rev. sci. eng, vol. 31 p. 1. rollmann, l.d., 1977, "catalytic hydrogenation of model nitrogen, sulfur, and oxygen compounds," j. catal, vol. 46, p. 243. sapre, a.v. and gates, b.c., 1980, "prepr., div. fuel chem., hydrogenation of aromatic hydrocarbons catalyzed by sulfided coo-mo/γ-al2o3: reactivity, reaction network, and kinetics," am. chem. soc, vol. 21, p. 66. singhal, g.h., espino, r.l., sobel, j.e. and huff, g.a., 1981, hydrodesulfurization of sulfur heterocyclic compounds: kinetics of dibenzothiophene," j. catal, vol. 67, p. 457. steiner, p. and blekkan, e.a., 2002, “catalytic hydrodesulfurization of a light oil over a nimo catalyst: kinetics of selected sulfur components,” fuel process. tech, vol. 79 p. 1. topsoe, h. and clausen, b.s., 1984, "importance of como-s type structures in hydrodesulfurization," catal. rev. sci. eng, vol. 26, p. 395. venezia, a.m., la parola, v., deganello, g., cauzzi, d., leonardi, g. and predieri, g., 2002, "influence of the preparation method on the thiophene hds activity of silica supported como catalysts,” appl. catal. a, vol. 229, p. 261. wang, y., sun, z., wang, a., ruan, a.l., lifeng, l., ren, j., li, x., li, c., hu, y. and yao, p., 2004, "kinetics of hydrodesulfurization of dibenzothiophene catalyzed by sulfided co-mo/mcm-41," ind. eng. chem. res, vol. 43, p. 2324. yang, p., yan, f. and liao, k., 2002, "a study on the effect of preparation parameters on the catalytic performance and active components of a new type hydrodesulfurization catalyst," petr. sci. tech, vol. 20, p. 763. the journal of engineering research (tjer), vol. 14, no. 2 (2017) 182-190 doi: 10.24200/tjer.vol.14iss2pp182-190 investigation of effective parameters of drop-on-demand droplet generator m.s. sarebandia, m. ghodsi*,b, y. hojjata, h. sadeghiana, h. ziaiefarb, m. mohammadzaherib and a. al-yahmedib a department of mechanical engineering, tarbiat modares university ,tehran, iran. b mechanical and industrial engineering, college of engineering, sultan qaboos university, muscat, sultanate of oman. received 12 march 2017; accepted 8 october 2017 abstract: this article presents a design and development of a drop-on-demand (dod) droplets generator. this generator uses molten metal as a liquid and can be used in fabrication, prototyping and any kind of printing with solder droplets. this setup consists of a vibrator solenoid with tunable frequency to produce a semi-spherical shape of molten metal, close to the surface of fabrication. this design also has a nozzle with micro-size orifice, a rod for transmitting force and a heater to melt the metal and keep it in superheat temperature. this dod can produce droplets in different sizes (less than 550 µm) by controlling the vibration frequency of solenoid. this ability together with the accuracy of the droplets in positioning (the error is less than ±20 µm for 1.5 mm amplitude) can be used in different applications. moreover, in this paper, the impact of initial position of the head and temperature on the average diameter of droplets and the impact of the frequency on the shape of the droplets have been tested and discussed keywords: drop-on-demand; metal droplets; rapid prototyping fabrication; solenoid transduce. مالت التغري الفعالة ملولد القطرة املسمى قطرة عند الطلباإختبار مع عامر،بظاهرى، مرتضى حممد بضيالي رضا ، محيدأيانقاديصسام ح، أحجت، يوسف *ب،قدسي، جمتبى أبنديرسمشسي مهدي باليحمدي املعدن يقوم هذا املولد باستخدام(. dodاملسمى قطرة عند الطلب ) تقدم هذه املقالة تصميما وتطويرا ملولد القطرة :امللخص تصنيع والقولبة وأي نوع من أنواع الطباعة بقطريات سبائك اللحام. ويتكون هذا الاملنصهر كسائل وميكن استخدامه يف كثريا السطح املصنع. و يوجد اجلهاز من ملف لوليب هزاز ذو تردد موقوت إلنتاج شكل شبه دائري من املعدن املنصهر يشبه بهذا التصميم أيضا فوهة ذات فتحة صغرية احلجم وقضيب لنقل الطاقة وسخان إلذابة املعدن واحلفاظ عليه يف درجة حرارة ميكرون( عن طريق التحكم يف 550( انتاج قطريات ذات أحجام خمتلفة )أقل من dodمرتفعة. وميكن هلذا املولد ) اللوليب. و ميكن هلذه االمكانية جنبا إىل جنب مع دقة القطريات حتديد املوقع املراد حلامه )و نسبة ترددات اهتزاز امللف مت فقدملم( كما ميكن استخدامه يف تطبيقات خمتلفة. وعالوة على ذلك 1.5 كلميكرون ل 20 ±أ فيه هي أقل من اخلط طر القطريات ومدى تأثري ُقللرأس ودرجة احلرارة على متوسط يف هذه الورقة البحثية اختبار ومناقشة تأثري املوقع األولي الرتدد على شكل القطريات. : قطرة بناء على الطلب قطرات معدنية؛ تصنيع النماذج األولية السريعة؛ حمول امللف اللوليبالكلمات املفتاحية * corresponding author’s e-mail: ghodsi@squ.edu.om mailto:ghodsi@squ.edu.om m.s. sarebandi, m. ghodsi, y. hojjat, h. sadeghian, h. ziaiefar, m. mohammadzaheri and a. al-yahmedi 183 1. introduction metal dod generators have become a useful tool in prototyping devices, layer-by-layer fabrication, solid fabrication, solder droplet printing and surface treatment. recently, due to their cheap price, droplets have attracted more attention. droplets can be tuned for different applications, from their old version which has been used for printing (waldvogel et al. 1996) to newer metal dods that can be used as 3d printers for prototyping (lawrence et al. 2017). at present, dods have complex technology and therefore they are not very common. producing a dod needs several manufacturing processes (etching, sputtering and so on) that are time consuming (liu and orme 1999). recently, researchers conducted many studies to overcome the aforementioned drawbacks and develop a cheaper and simpler dod. these generators mainly include a droplet producer unit and a droplet deposition unit. in the former unit, a heating furnace melts metal to a specific temperature; and a vibrator ejects droplets from a nozzle using pulses passing through the molten metal (sohn and yang 2005). the nozzle orifice and the latter unit deposit the droplets. in addition to dods, continuous droplet generators are another type of metal droplets generators. although these two methods of drop generating have some differences, both use periodic pulses to eject the molten metal drops through the nuzzle (sohn and yang 2005). for soldering droplets printing, continuous rows of droplets are ejected to produce a line or stream on the board. this uniform metal droplets can be produced by exciting a laminar molten metal jet. formation by using continuous drops of liquid from a steam was examined by savarat (1833), and many years later, rayleigh developed its mathematical equation (rayleigh 1878; rayleigh 1879). working frequency of dod depends on many factors like the speed and the size of drops. the frequency has a wide range between 5 to 44 khz, and lower frequency means larger drops (liu and orme 1999). on the other hand, for printing using drop-ondemand, the system can generate a discrete row of metal droplets to make the desired form. as an example, yamaguchi uses a chamber attached to a diaphragm to eject metal for fabrication. yamaguchi has used piezoelectric as a vibrator (yamaguchi et al. 2000). moreover, jun luo showed that it is possible to make a drop-on-demand droplet generator by controlling the pneumatic force of gas as a pressure source (jun luo et al. 2012). piezoelectric buzzers can be implemented in dods for a room temperature liquid (kuangchao fan 2008). using a piezoelectric vibrator (sadeghian et al. 2013; mohammadzaheri et al. 2017) or magnetostrictive one (ghodsi et al. 2017; karafi et al. 2015; ghodsi et al. 2010; ghodsi et al. 2008), dod can reach high frequency of vibration; on the other hand, the harsh environment (high temperature) can limit their use in dods. the main aim of this research is to design and fabricate a dod generator with the ability of controlling the size and position of drops in 2-d surface. moreover, in this article, the impact of different factors such as temperature of superheating metal, position of transmitting rod at the beginning, frequency of generator and amplitude of impulse-transmitting rod will be discussed. figure 1 shows the schematic view of the dod. 2. experimental setup as shown in fig. 1, the experimental setup consists of a vibrator as an impulse generator, a rod to transmit the impulses, a micro-nozzle, a tubular shape container for molten metal and a heating part for heating the metal to a desired temperature. the source of vibration is a precise solenoid at the top of the generator. figure 1. schematic of the fabricated generator. investigation of effective parameters of drop-on-demand droplet generator 186 after that, an impulse-transmitting rod transmits the impulses to the metal chamber. the rod is placed near the orifice of the nozzle at the beginning. the vibrator generates the impulses and accordingly, the molten superheated metal is ejected from the nozzle. four springs are attached to bring the rod back to the initial place for another ejection. sticking of the molten metal to the surface of the droplet generator is very common. the main reason is the small size of the orifice. to prevent the nozzle getting stuck with materials, non-wetting liquids should be used for injection process, like ceramic materials. however, by choosing a proper material for nozzle, this problem can be reduced. for instance, in previous research, stainless steel was reported as a nozzle material for tin (sohn and yang 2005). hence, some stainless steel parts are in contact with tin (sus303). several factors influence the size of drops, one of them is the size of the nozzle orifice. to produce spherical drops, the shape of nozzle orifice has to be perfectly round. moreover, the aspect ratio of the nozzle is important for ejection of droplets. ln/dn must be equal to or smaller than one to break the drops more easily (sohn and yang 2005). the diameter of orifice is 198 µm, and the length is 180 µm. hence, the aspect ratio of micro-drilled orifice is 0.90 and less than one. for melting the metal and keeping it in superheat temperature, a tubular heating element was used. this heater can produce temperatures up to 700oc. for sensing the temperature, a k type thermocouple is placed into the cylinder hole (drilled hole) to be close enough to the nozzle. to control the temperature of the heater at the specific desired temperature, proportional-integral-derivative method was employed. in this study, droplets of sn63/pb37 are produced by the presented setup. table 1 summarizes the physical properties of sn63/pb37. 3. solenoid design in order to evaluate the required force of transducer to eject the molten metal out of the table 1. physical properties of sn63/pb37. melting temperature (oc) surface tension ( n.m-1) density ( kg.m-3) viscosity (10-3 pa.s) 183 0.49 8420 1.33 orifice, bernoulli's equation in fluid dynamics was used. the flow rate is q=ua, where u is the flow speed of the molten metal and a is the cross-sectional area. from the continuity equation and according to the law of conservation of mass, flow speed of the liquid and droplet, a force of magnitude 0.09 n is required to be exerted on the plunger to make molten metal squirt from the orifice. in this setup, a solenoid is used as a force generator. in a solenoid, number of turns, air gap, size of solenoid, the characteristic of the steel of solenoid are factors that have impact on the produced force. to have the desired force for transducer, a coil with 200 turns of wire with 0.45 mm diameter was used. using this solenoid, the amplitude of displacement is 3 mm, and maximum payload is 9.05 n. force equations and solenoid equations can be found in appendix a. during the experiment, the payload can be adjusted by controlling the current. for adjusting the initial position of the plunger, a number of washers were used. 4. effective parameters on size of droplet’s to have an acceptable dod for droplet manufacturing, several experiments had been conducted to study the impact of different parameters on the characteristic of metal droplets. 4.1 temperature of molten metal and position of head at the beginning (lh) the melting temperature of sn63/pb37 is 183˚c. at this temperature, the viscosity of metal is not low enough to be used in the dod (manko 1979). to reach the desired viscosity, the temperature was set at 260, 290, and 320˚c. in this temperature, the molten metal is in a superheated condition. another factor for optimization is the gap between the head end of transmitting rod and the chamber. this factor has an impact on the efficiency of material supply for dropping and impulse transfer. in this design, this clearance is set at 125 µm, the vibration frequency (fsol) is set at 3 hz, displacement (a) is set at 0.5 mm and 40 ml of the melted metal is inserted in the tubular vessel. figure 2 shows the relationship between the position of the head end at the beginning and temperature of the molten metal and their effect on the diameter of spherical droplets. to have better estimation, the average diameter of 10 droplets was calculated. as it is shown in fig 2, by increasing the temperature of the super184 m.s. sarebandi, m. ghodsi, y. hojjat, h. sadeghian, h. ziaiefar, m. mohammadzaheri and a. al-yahmedi 187 heated metal, the diameter of droplets decreased. this happens because higher temperature decreases the viscosity and surface tension of the metal (beer 1972), therefore, the superheated metal can flow through the orifice more easily. in the test where lh (initial position) was smaller than 1.5 mm, an ejected droplet lost its unity and became two or more separated droplets. the reason is due to the excessive force of the solenoid. in positions higher than 4.5 mm, droplets fell down continuously because of gravity. 4.2 amplitude of displacement and the position of head at the beginning the effect of the position of the head at the beginning and amplitude on the size of drop lets are shown in fig 3. figure 2. impact of temperature and the position of the head at the beginning on the size of droplets. figure 3. effect of the position of the head at the beginning and amplitude on the size of droplets. to investigate the effect of amplitude on the size of droplets, using amplitude limitation washers, which are located on the top of the actuator, displacement amplitude of solenoid (a), were set at 0.5, 1 and 1.5 mm. in this experiment, the temperature is set at 290oc and other parameters such as the gaps between the slot of the nozzle and the head of impulsetransmitting rod, the frequency of the solenoid and melted metal inserted in the tubular vessel are set as in the previous experiment. figure 3 shows that by increasing the gap size of the head end position, the size of droplets increases. with increasing amplitude of the transducer, the average of droplet diameter increases too. when the amplitude increases, the impulse-transmitting force in the nozzle chamber increases too, thus, more melted metal is ejected from the nozzle orifice, and larger droplets are produced. the corresponding range of droplet diameter is between 542 µm and 554 µm. 4.3 solenoid vibrator frequency to study the effect of frequency, droplets are produced and analyzed at different frequencies, and in each case, more than 10 measurements on droplets size are carried out. in these experiments, displacement amplitude is set at 0.5 mm, and other parameters are set as in the previous experiment. figure 4 shows photographs of the products of molten metal ejected from nozzle orifice at different frequencies. figure 4. ((a)-(c)) visualized the top view of deposited droplets and their diameter. in fig. 4 ((a)-(c)), the droplets have a number of edges caused by the solidification of the solder at very low frequencies (goldin 1999). as it can be seen in fig. 4, at lower frequencies, the shape of the droplets is better and closer to spherical shape. the main reason is that when the frequency is low the molten metal has enough time to eject from the nozzle and has time to become solid. at low frequencies, time is enough for a droplet not to have a negative impact on the previous or next drops. on the other hand, at higher frequencies, because of surface tension between the droplets and viscosity, there is no enough time for the material to break and make spherical separated droplets. at last, the molten metal will become solid in the intended line of fiber. figure 4 ((d)-(f)) shows the solid droplets of metal after dropping. 185 investigation of effective parameters of drop-on-demand droplet generator 186 5. positioning of drop-on-demand droplets the accuracy of prototyping and fabrication by a dod generator depends on the droplets precise location and the frequency of the drop. to test the drop-on-demand capability, the temperature of the device was set at t= 290oc, and the frequency of dropping was set at 3 hz, and the gap between the head of the nozzle and surface was maintained at 5 mm. experimentally, the velocity of the droplets was calculated around v=10mm/s. the setup of the experiment and the results of the tests are shown in fig 5. figure 5. shows a row of droplets in a specific location (according to frequency). the result of the tests shows that the setup works correctly, and the droplets are in the intended positions. to calculate the droplets velocity, a high-speed camera captured a series of photos. figure 4. photographs ((a)-(f)) of products of molten metal ejected from orifice under different frequencies. figure 6. shows the photos of the high speed camera respectively. the flight distance of the droplets was measured through image processing, and the velocity of droplets was measured 0.7 m/s. 6. accuracy of positioning the distance of the nozzle and target point on the surface can reduce positioning accuracy. this may happen due to insufficient accuracy in machining of the rod or imperfect alignment of the rod and vibrator. figure 7 shows the schematic of procedure for measuring the error and accuracy of the droplets positioning (fig. 8). for this test, a straight line of droplets was dispositioned, where the value of hn of each droplet, (position of the head of the nozzle at the beginning) was different. when hn is 3 mm, the initial position of the head end is 0.5 mm, and the amplitude is 1.5 mm, the value of error is less than ±20 µm which is an acceptable error for rapid prototyping according to the size of droplets. figure 5. experimental results of the drop-ondemand deposition of droplets. figure 6. photograph of metal droplet ejected from nozzle. m.s. sarebandi, m. ghodsi, y. hojjat, h. sadeghian, h. ziaiefar, m. mohammadzaheri and a. al-yahmedi 187 figure 7. schematic of tests for measuring error. figure 8. deposition error measurement. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. acknowledgment the authors express their thanks to the advanced metrology lab, tarbiat modares university, iran, for their appreciated technical supports. references beer sz (1972), liquid metal – chemistry and physics. marcel dekke, 474-488, new york. chun jh, passow ch (1993), droplet based manufacturing. journal of annals of the cirp 42: 235-238. fan kc, chen jy, wang ch, pan wc (2008), development of a drop-on-demand droplet generator for one-drop-fill technology. sensors and actuators a: physical 147(2): 649-655. ghodsi m, ueno t, teshima h, hirano h, higuchi t, summers e (2007), zero-power positioning actuator for cryogenic environments by combining magnetostrictive bimetal and hts. sensors and actuators a 135: 787-791. ghodsi m, ueno t, teshima h, hirano h, higuchi t (2007), numerical modeling of iron yoke levitation using the pinning effect of high temperature superconductor. ieee transactions on magnetics 43(5): 2001-2008. ghodsi m, ueno t, higuchi t (2008), novel magnetostrictive bimetal actuator using permendur. advanced materials research 47-50: 262-265. ghodsi m, ueno t, modabberifar m (2010), quality factor, static and dynamic responses of miniature galfenol actuator at wide range of temperature. international journal of physical sciences 6(36): 8143-8150. ghodsi m (2015), optimization of mover acceleration in dc tubular linear direct-drive machine using response surface method. international review of electrical engineering 10(4): 492-500. ghodsi m, hosseinzadeh n, ozer a, rajabzadeh h, garjasi vn, hojjat y, talebian sl, sheykholeslami m, al-yahmadi a (2017), development of gasoline direct injector using giant magnetostrictive materials. ieee transactions on industry applications 53(1): 521529. goldin r (1999), flip chip bumping with high speed metal jet technology. proceedings of nepcon west'99, los anglos. hayes dj, wallace db, boldman mt (1993), method and apparatus for dispensing spherical shaped quantities of a liquid solder. us patent: 5: 229 016. karafi mr, ghodsi m, hojjat y (2015), development of magnetostrictive resonant torsional vibrator. ieee transactions on magnetics 51(9): 1-8. lawrence l, james l, jennifer r (2017), examine a 3d-printed carbon fiber part created using a direct-ink writing process developed at llnl. lawrence livermore national laboratory, usa. 1mm investigation of effective parameters of drop-on-demand droplet generator 186 liu q, orme m (1999), high precision solder droplet printing technology and the state-ofthe-art. journal of materials processing technology 115: 271-283. luo j, qi lh, zhou jm, hou xh, li hj (2012), modeling and characterization of metal droplets generation by using a pneumatic drop-on-demand generator. journal of materials processing technology 212(3): 718-726. manko hh (1979), solders and soldering: materials, design, production, and analysis for reliable bonding. mcgraw-hill, new york. mohammadzaheri m, qallaf a (2017), nanopositioning systems with piezoelectric actuators, current state and future perspective. science of advanced materials 9(7): 1071–1080. murr le, johnson wl (2017), 3d metal droplet printing development and advanced materials additive manufacturing. journal of materials research and technology. özer a, ghodsi m, sekiguchi a, saleem a, alsabari n (2015), design and experimental implementation of a beam-type twin dynamic vibration absorber for a cantilevered flexible structure carrying an unbalanced rotor: numerical and experimental observations. shock and vibration, article id 154892. rayleigh l (1878), on the instability of jets. proceedings of the london mathematical society 10: 4-13. rayleigh l (1879), on the capillary phenomena of jets. proceedings of the london mathematical society 29: 71-97. savarat f (1833), memoire sur la constitution des veines liquids lances par des orifices circulaires en mince. annales de chimie et de physique 53: 337-386. sadeghian h, hojjat y, ghodsi m, sheykholeslami mr (2013), an approach to design and fabrication of a piezo-actuated microdroplet generator. the international journal of advanced manufacturing technology 70(5-8): 1091-1099. sohn h, yang ky (2005), drop-on-demand deposition of superheated metal droplets for selective infiltration manufacturing. journal of materials science and engineering 392: 415–421. waldvogel jm, poulikakos d, wallace db, marusak rm (1996), transport phenomena in pico liter size solder droplet dispensing on a composite substrate. asme journal of heat transfer 118: 148–156. yingxue y, shengdong g, and chengsong c (2004), rapid prototyping based on uniform droplet spraying. journal of materials processing technology 146: 389-395. yamaguchi k, sakai k, yamanaka t, and hirayama t (2000), generation of threedimensional micro structure using metal jet. journal of precision engineering 24(1): 2–8. zaugg fg, wagner p (2003), drop-on-demand printing of protein biochips arrays. mrs bulletin 28(11): 837-842. appendix a exit velocity of a molten metal from a small nozzle is around 1 m/s. according to the dimensions of nozzle in figure a-1 for calculating the speed of the guidance rod, we will have 𝑢 𝑈 = ( 𝐷 𝑑 )2 (a1) 1 𝑈 = ( 0.006 0.0002 )2 ⇒ 𝑈 = 0.9 𝑚/𝑠 (a2) where u is the speed of the piston. we can write the bernoulli equation ( 𝑝 𝜌𝑔 + 𝑣2 2𝑔 + 𝑧)1 = ( 𝑝 𝜌𝑔 + 𝑣2 2𝑔 + 𝑧)2 (a3) where p is the pressure, v is the velocity, 𝜌 is the density of liquid and z is the height of the fluid. the z is equal, therefore we can write ( 𝑝 𝜌𝑔 + 𝑣2 2𝑔 )1 = ( 𝑝 𝜌𝑔 + 𝑣2 2𝑔 )2 (a4) 188 m.s. sarebandi, m. ghodsi, y. hojjat, h. sadeghian, h. ziaiefar, m. mohammadzaheri and a. al-yahmedi 185 189 in addition, we can calculate the pressure differentiation as 𝑝1 = 𝑝0 + 𝜌 𝑢2−𝑈2 2 (a5) ⇒ 𝑝1 – 𝑝0 ≅ 𝜌 𝑈2 2 (a6) the density of the liquid in this experiment in test temperature is 7830𝑘𝑔. 𝑚−3. by substituting this value in last equation we will have ⇒ 𝑝1 – 𝑝0 ≅ 3170 (a7) moreover, the force of the piston can be written as ∑ 𝐹 = 0 ⇒ ∑ 𝐹 = 𝑚𝑎 = 0 (a8) ⇒ 𝐹 = (𝑝1 – 𝑝0 )𝐴 ⇒ 𝐹 = 3170 ∗ 𝜋 ∗ (0.003) 2 (a9) ⇒ 𝐹 = 0.09 𝑁 (a10) to simplify the equations we made some certain assumptions. in the real experiment, the piston has friction, there is heat transfer in the system and the liquid velocity is not continuous. hence, this value of force for the piston is less than the real value. figure a-1. dimension and velocity direction of the nozzle. for calculating the force of solenoid, thanks to maxwell equation we can write 𝐹 = 𝐵2𝐴 (2𝜇0)⁄ (a11) where f is the force of solenoid, b is the magnetic flux density and magnetic permeability of nozzle. magnetic flux density can be written as 𝐵 = 𝜇𝐻 𝐵 = 𝜇0𝜇𝑟 ( 𝑁𝐼 𝐿 ) (a12) where n is the number of turns in the coil, i is the current, l is the length of the coil, 𝜇0is permeability in vacuum and 𝜇𝑟 is the relative permeability. by substituting the values into equation we will have investigation of effective parameters of drop-on-demand droplet generator 190 𝐵 = 4𝜋 × 10−7 × 1000 ( 200×0.1 0.042 ) = 0.598 𝑇 (a13) 𝐹 = 𝐵2𝐴 (2𝜇0) = 9.05 𝑁⁄ (a14) microsoft word paper 8.docx the journal of engineering research (tjer) vol. 14, no. 1 (2017) 85-93 a comparative study on energy derived from biomass a.m. algarny, i.m. al-naimi and m.a.m alhefnawy* * college of architecture and planning, university of dammam, saudi arabia. received 7 april 2016; accepted 17 october 2016 abstract: the paper promotes sustainable community through empowering the production and utilization of biomass renewable energy. the aim of this paper is to urge societies to adopt sustainable energy practices and resources; the objective is to appraise the possibilities of biomass energy produced through a neighborhood in eastern province, saudi arabia. the system incorporates an evaluation of the measure of biomass created, then utilizes two ascertaining techniques to gauge whether the measure of energy can be delivered. the computation strategies are hypothetical, with one drawn from past works and the other from a biomass calculation template performed as part of the evaluation of biomass resources for municipalities study (ebimun) by the waterford county council. the outcomes demonstrate that the aggregate potential biogas generation of the study area is around 43,200 m3/year, the methane mass is around 18,000 m3/year, and the energy production amount is around 250 mwh/year. contrasting the capability of biogas creation from both techniques, the figure assessed by ebimun is around 7,000 m3/year less than the hypothetically computed amount. the figures suggest that biogas is worthy of consideration as a renewable source of energy. keywords: biomass, biowaste, renewable energy, sustainable energy. אאאא  kאkkk،אkk،אk*    אkאאאאאאאאאאw   א א   א k   א    א     ؛אאאאאאא אא؛אאאאאkאאאא  אא k א  א א   א   אא؛  א אאאאא(ebimun) kא    א  אא   א א א    43,200א٣ ،א  18,000א٣אאא٢٥٠אkאא אאאk(ebimun)٧٠٠٠٣ ؛אאאאאאאkאא kאא  אאאאא،אא،אא،אאw.  * corresponding author’s email: malhefnawi@uod.edu.sa a.m. algarny, i.m. al-naimi and m.a.m alhefnawy   86 1. introduction the word biomass denotes the plants and other substances derived from land. it is the oldest and most important source of renewable energy for many reasons. it is renewable, spread around the globe, and is easily collected by environmentally friendly technologies (us department of energy 2000). biomass energies can be derived from a variety of resources, including timber, cultivated lands, agroindustrial wastes, animal waste, sewage, municipal solid waste (msw), and food handling wastes (cleantech1 2016). 2. waste in the middle east middle eastern and north african (mena) countries have rich unexplored biomass energy resources that hold the potential to generate about 400twh annually. in the middle east, 150 million tons of urban waste are generated per year, mainly by egypt, yemen, iraq, syria, and jordan. on the other hand, bahrain, saudi arabia, the united arab emirates (uae), qatar, and kuwait are among the top ten countries worldwide that produce solid waste per capita (bioenergy1 2016). furthermore, saudi arabia experiences pollution due to fast development and population growth, and more than 15 million tons of solid waste are annually generated. solid waste generation has been assessed at 1.5 to 1.8 kg per capita/day. having a population of around 29 million, solid waste control is a great challenge for the kingdom. more than six million tons of solid waste are annually generated in each of the major cities–riyadh, jeddah, and dammam. recycling, reuse, and energy recovery are getting increased attention, but they are still at an early stage. the recycling rate varies from 10–15% for metals, paper, and plastics. composting has received increased attention in saudi arabia because of the high organic content of the msw, which has been observed to reach 40% (bioenergy2 2016). 3. conversion of biomass to energy in 2000, only 1.5% of the world’s electricity was bioenergy-created electricity. on the other hand, the world’s biofuel production expanded from 16 billion liters in 2000, to more than 100 billion liters in 2010, constituting 3% of the world’s fuel for transport (cleantech2 2016). the yearly prospective biogas generation in europe has been surveyed at more than 200 billion m3 (appels et al. 2008). biomass-based energy constituted around 10% of the total main energy source in 2009 worldwide (cleantech2 2016). biochemical and thermal technologies are usually used to convert biomass into energy and fuels in the form of electrical power, heat, or both. electricity production efficiency from biomass is around 30%, and can reach up to 85% if heat gain is considered. biochemical strategies, such as anaerobic digestion, harvest energy in the form of biogas that is transformed to power and heat by means of a gas machine (cleantech1 2016). 3.1 the microbial process of anaerobic digestion anaerobic digestion is a methanogenic decomposer of biomass. it is a reactor, which can be blended or non-blended and warmed or non-warmed. the methodology utilizes a blend of organisms to complete the decomposition. there is no need for feedstock sterilization or contamination concerns. the methane separates and then moves in pipes to a storage tank for later use. present means of methane production result in a 50–97% conversion of substrate to methane depending on the feedstock. carbohydrates substrate yield 50% methane and 50% carbon dioxide (co2), while lipids yield higher proportion of methane [table 1] (wilkie 2008). 3.2 digestion tanks anaerobic tanks are made of shielded concrete or carbon steel. air is not necessary, and mixing provides bacteria an avenue for interaction with waste, provides a constant temperature, and prevents solid buildup. the container is equipped with warming and blending tools, a floating top for gas collection, or a detached gas vessel. in the usa and germany, egg-formed tanks are frequently used as they provide for full blending of the table 1. amount of methane produced from different types of biomass (wilkie 2008). biomass yield (liter/g of volatile solids) carbohydrate 0.350 protein 0.570 fat 0.950 a comparative study on energy derived from biomass   87   waste due to their declining impact. the tank should be totally blended to let raw influent waste have contact with actively processing waste, thereby speeding up the process, ensuring an even temperature in the tank, and preventing buildup of stones in the tank and the formation of pockets of non-degraded sludge (esru 2016). there are two commonly used blending systems—mechanical and gas. gas recirculation systems are used in most modern plants where gas is redirected through the waste. a mechanical system is cheaper to install but more costlier to maintain. a third blending method, draft-tube, is energy effective and recommended for egg-formed tanks (scottish envirotech 1995). 3.2 energy recovery biogas is 60–70% methane, 30–40% co2, and traces of different gases such as nitrogen, hydrogen, hydrogen sulphide and water vapor. the caloric value of 70% methane content biogas is around 21–23.5 mj/m³, meaning that one cubic meter of biogas matches 0.5–0.6 liters of diesel, or around six kwh [table 2] (rohstoffe 2009). to lend perspective, two kwh is sufficient to run a 100 w light bulb for 20 hours (electrigaz 2016). the produced gas may be turned to energy or run the digestion process, saving electricity. the surplus biogas may be utilized in a variety of means, including for combined heat and power generation (chp), gas distribution, sludge dehydration, and fueling waste incinerators (frost et al. 1990). some studies estimate that 150–200 m3 of biogas might be produced from one ton of food waste depending on tank, treatment circumstances, and wastes (cleantech2 2016). electricity generation from biogas is infrequent in developing nations. on the contrary, power generation is the chief target of biogas production in germany and other industrialized nations. theoretically, in a fuel cell, biogas may be turned into electricity. however, expensive fuel cells and precise amounts of pure gas are required for this operation. in most cases, biogas fuels a combustion engine, which turns it to mechanical energy that powers an electric generator for electricity production. gas turbines are used as biogas engines in the usa. they are very costly, small, and their outputs range from 30–75 kw. external combustion engines accept different fuel qualities, but they are rather costly and minimally effective (giz 2013). table 2. typical details of biogas (hanjie 2010). composition 55–70% methane, 30–45% carbon dioxide, traces of other gases energy content 6.0–6.5 kwm-3 fuel equivalent 0.6–0.65 l oil/m3 biogas explosion limits 6–12% biogas in air ignition temperature 650–750 ºc critical pressure 75–89 bar critical temperature -82.5 ºc normal density 1.2 kg/m3 odor sulphur     4. practical work the aim of the practical work is to estimate the potential of the biomass energy product in the university of dammam housing compound in petromin district, eastern province, saudi arabia. the objectives of the study follow: 1. to estimate the amount of biomass produced in the housing compound, 2. to estimate the amount of energy delivered utilizing a hypothetical computation technique generated from past works, 3. to estimate the amount of energy produced using the biomass calculation template as part of the evaluation of biomass resources for municipalities (ebimun) study created by the waterford county council (wcc). 4.1 the study area the university of dammam housing compound in petromin district, dammam city, is located on the main road to the university of dammam alongside king saud and king faisal roads (fig. 1). the compound is allocated to the staff of the university, and has both saudi and expatriate residents. the population is about 820 persons and, when maintenance workers, staff, drivers, and administrative staff are included, the population is 1,200 persons. a.m. algarny, i.m. al-naimi and m.a.m alhefnawy   88 figure 1. university of dammam residential compound in petromin. table 3. hypothetical calculation steps. population size: 1200 persons the per capita msw in saudi arabia: 1.5–1.8 kg/day, and 40% of it is biomass (zafer 2015) the area can provide: 1.65 * 1200= 1,980 kg/day of msw containing 792 kg/day of biomass each ton of food waste produces: 150–200 m3 of biogas (gp, 2013) potential of biogas: 0.792*175 = 138.6 m3 daily, (50,589 m3 yearly) the average calorific value of biogas: 21–23.5 mj/m³ (rohstoffe, 2009) 1 m³ of biogas: corresponds to 0.5–0.6 liter of diesel fuel, or about 6 kwh (rohstoffe, 2009) potential of biomass energy: 138.6 * 0.55 = 76 liter of diesel fuel daily equivalent to: 76 * 6 = 456 kwh/day, (166,440 kwh/year) 4.2 potential of biomass energy this study calculated the energy potential of biowaste in the study area by two different methods. the first method depended on the hypothetical figures regarding generation from biomass, and the second method was based on an excel sheet (microsoft, corp., redmond, washington, usa) titled biomass calculation template (ebimun) made by the waterford county council (wcc) (2013). 5. results 5.1 hypothetical calculations household biomass includes all domestic organic waste such as food, paper, tree leaves, etc. the potential energy production derived from biowaste in the study area was calculated through the following steps (table 3). 5.2 biomass calculation template of the evaluation of biomass resources for municipalities (ebimun) study the aim of the ebimun project was to increase the capability of the authorities in estimating the local organic waste potential and implement the most effective energy production method (universidad de leon 2016). the calculation template is used to calculate the potential of biomass energy of different waste materials from households, industry, sewage etc. from fig. 2, the total potential of biogas production of the study area is around 43,200 m3/year, and the methane mass is 18,000/year, the energy production is 900,000 mj/year, and energy production is 250 mwh/year. comparing the potential of biogas production from both methods (table 4 and fig. 3), the potentials a comparative study on energy derived from biomass   89 figure 2: energy potential of the waste by the ebimun template. table 4. biogas and energy potentials in the study area. theoretical calculations ebimun template biogas production potential 50589 m3/year 43200 m3/year biomass energy potential 166440 kwh/year 143138 kwh/year figure 3. biogas and energy production potentials in the study area. estimated by ebimun is around 7,000 m3/year less than the theoretical calculation. the figures suggest that biomass-produced gas is worthy of consideration as a renewable energy source for the study area. 5.3 the proposed plan an anaerobic digester is proposed to be placed in the residential compound in order to divert the biomass waste from the municipal waste stream and to produce renewable energy "biogas" that can be used as a local source and distributed to the housing units. in order to increase the outcomes of the operation, it is planned to include the waste produced from the whole neighborhood. the system is proposed as follows: o establishing a garbage collection system supervised by a board from the compound residents. it includes small trucks, recycling bins, and a collection and classification space. o building an anaerobic digester with all required facilities and connections. it should be heated in order to maximize the production rate. o extending a network of biogas to supply the residential units. a study based upon an analysis of the actual monthly electricity consumption for 115 dwellings in dhahran for the year 2012 (included 62 apartments, 28 villas, and 25 traditional houses), showed that the annual average electricity consumption was 176.5 0 20000 40000 60000 80000 100000 120000 140000 160000 180000 hypothetical method ebimun method biogas cubic m/year energy kwh/year a.m. algarny, i.m. al-naimi and m.a.m alhefnawy   90   kwh/m², which was a higher value than international energy-efficiency benchmark (alrashed and asif 2014). saudi arabia’s population has grown rapidly in recent decades, and electricity consumption (kw) per capita was 7,527 kwh in 2013. considering this figure, the electricity produced from biomass in the study area can fulfill the full needs of around 22 of the 1200 persons in the study area, which is almost 1.8% of the population. the energy outcome is obviously very limited due to the limited size of the residential compound, and the high consumption rate of electricity in ksa in general (alrashid and asif 2014). as such, recycling biomass for energy conversion should be extended to include the surrounding sites adjacent to the housing compound in order to increase the efficiency of the operation. therefore, the study proposes that household biowaste should be collected from larger areas, the overall context of al raka district, petromin district, and alsafa district, in order to get more biowaste and therefore more energy produced. the majority of electricity consumption in saudi arabia is for air conditioning loads; the intent of the proposal, therefore, is mainly to reduce energy needed from the public electricity network. biomass collection systems should be placed throughout the district to collect biowaste from houses, shops, markets, restaurants, etc. local recycling stations will be placed on streets at reasonable intervals with sign boards and usage instructions. this initiative would also benefit from information media (fig. 4). the collection devices will all be collected and directed to the anaerobic digester, which will be located in the vacant area southeast of the site, where the biomass will be processed for biogas production. a portion of the gas produced will be recirculated within a network of pipes back to houses for cooking use as a replacement to electric stoves. the other portion of produced gas is to be directed into an engine cell to get converted into electricity. this process is always accompanied by heat production, so the heat produced will be used to warm the reactor to the needed temperature in a thermophilic mode to get the maximum production rate of biogas. figure 4. proposed plan at the university housing compound. recycling stations anaerobic digester a comparative study on energy derived from biomass   93   6. conclusion the current study sought to contrast the potential energy created by the two distinct strategies. the hypothetical technique delivered around 51,000 m3 of biogas, which surpassed about 18% of the biogas delivered by the ebimun strategy, which was evaluated at 43,000 m3. the contrast between the two strategies may be clarified when the accompanying elements that firmly influenced the beneficial possibilities of energy from biomass are considered:  the assorted qualities in the sort and synthesis of organic waste, whether provincial squanders, household, agricultural, sludge, energy crops, catering waste, industrial waste, or harvested biomassnoticeably affected the outcomes. these differences can be seen from city to city, and from nation to nation and prompts the varieties of carbon substance in the waste, and along these lines the varieties in the measure of energy delivered.  the variety of environmental conditions under which the organic waste was produced and delivered affected the resultant energy. the variation in the quality of fertilizers and animal feed utilized as well as the climate conditions such as the amount of rain, moisture content and temperature ranges, and other atmospheric conditions influenced the synthesis of the waste and its humidity content.  the assortment of the accessible technologies in the form of digestion tanks from the european union where ebimun is embraced and in other nations where the hypothetical estimation strategy was involved logically, affect energy creation outcomes.  including just the waste created by the university of dammam housing compound would not result in enough biomass. along these lines, energy generation would be significantly more advantageous if the three locales around the compound—petromin, al raka, and alsafa—were to be included. saudi arabia has a wide range of assets in the area of biomass waste. it has the potential to generate renewable energy and decrease reliance on fossil fuels. it is also among the top ten waste-generating nations in the world. however, these opportunities are not being utilized yet. in the small community presented in the study, about 2% of the energy required for family utilization can be obtained by domestic biomass anaerobic digestion. if wastewater treatment, including anaerobic digestion of both domestic and industrial sewage streams, were to be added, more biogas could be produced and a greater reduction in dependency on electricity networks could be accomplished. the proposed framework might play an imperative part in energy saving and generation. recommendations efforts should be directed to increase the amount of household organic waste that can be collected from the residents of the housing complex. a team of volunteers from the residents of the housing complex should be recruited in order to oversee the implementation and development of this proposal collection and classification bags for organic waste should be supplied to every residential unit in the compound, and the time and location of organic waste collection should be conveyed to the population wedding halls, specialized kitchens in the neighborhoods, and restaurants should be contacted and coordinated with in order to incorporate the remnants of their events into the anaerobic digester. the population of the housing compound and of saudi arabia should be made aware of the importance organic waste and renewable energy use. a simulation model for the area should be developed that depicts the amount of energy consumed before and after implementation of the proposal and reflects the benefits from the proposed energy saving. the model should detail the economic impact of the proposal on the local population. a board should be assembled to approach the bodies concerned with the treatment of municipal waste in order to coordinate the 91 a.m. algarny, i.m. al-naimi and m.a.m alhefnawy   92   possibility of establishing an organic waste sewage plant linked to the anaerobic digester to increase production efficiency and rates of recycling. it would be more advantageous to include the three surrounding districts—petromin, al raka, and alsafa—in the framework in order to reap enough usable waste and, consequently, energy generation. nearby neighborhoods, districts, towns, and cities should encourage their inhabitants to embrace recycling mechanisms and elevate biomass to increase energy conversion. biomass energy should be utilized as a vital source of renewable energy that can be applied to save the earth from pollution and add to the economy by saving conventional energy production costs. biomass conversion innovation should be embraced in all nations. more investigation should be urged to develop more potential sources for energy protection and lessen the amount of solid waste being disposed of after just a single use. acknowledgment this project (no. 2012168) was funded by the deanship of scientific research at the university of dammam. references alrashed f, and asif m (2014), trends in residential energy consumption in saudi arabia with particular reference to the eastern province. journal of sustainable development energy, water and environmental systems 2(4): 376–387. appels l, baeyens j, degrève j and dewil r (2008), principles and potential of the anaerobic digestion of waste-activated sludge. energy and combustion science 34:755–781. bioenergy1 (2016), biomass feedstock in the middle east and north africa. from bioenergyconsult.wordpress.com/2011/09 /30/biomass-feedstock-in-middle-eastand-north-africa-mena. accessed apr 2016. bioenergy2 (2016), municipal wastes in saudi arabia. from bioenergyconsult.com / municipal-wastes-in-saudi-arabia. accessed april 2016. cleantech1 (2016), introduction to biomass energy. from www.cleantechloops.com / an-introduction-to-biomass-energy. accessed april 2016. cleantech2 (2016), a glance at biomass energy. from www.cleantechloops.com/a-glanceat-biomass-energy. accessed apr 2016. electrigaz (2016), biogas faq. from www.electrigaz.com/faq_en.htm. accessed apr 2016. esru (2016), energy from waste. from www.esru.strath.ac.uk/eande/web_sites/ 97-8/energy_from_waste/ page3.htm. accessed jan 2016. frost r, powlesland c, hall je, nixon sc, and young c.p (1990), review of sludge treatment and disposal techniques. wrc report prd 2306-m/1. giz (2013), small scale electricity generation from biomass: part 2. from www.giz.de/themen/de/dokumente/gtz 2010-en-small-scale-electricity-generationfrom-biomass-part-2.pdf. accessed may 2013. google maps (2016), petromin area, dammam city, ksa. from www.google.com/maps/search/petromin +dammm+ksa/@26.4067682,50.1638835,16z accessed jun 2016. green power (gp) (2013), green power. from www.greenpower.cn. accessed jul 2013. hanjie z (2010), sludge treatment to increase biogas production, trita-lwr degree project 10-20, kth land and water resource engineering, sweden. jun h (2013), saudi arabia’s domestic energy situation and policy: focusing on its power sector. kyoto bulletin of islamic area studies 107–135. myriad power (2016), an anaerobic digester image. from myriadpower.com / anaerobic.html. accessed sep 2016. rohstoffe fn, (2009), biogas basisdaten deutschland – stand. 7p. very short but comprehensive overview of the biogas situation in germany. from www.fnrserver.de/ftp/pdf/literatur / pdf_185basisdaten _ biogas _ 2009.pdf. accessed apr 2016. scottish envirotech (1995), scottish envirotec. 3(3): 24–25. u.s. department of energy (doe) (2000), biofuels for sustainable transportation, the national renewable energy laboratory, a doe national laboratory. doe/go102000-0812. from portal.ncdenr.org / c / document_library/get_file?uuid= 41c5f270 a comparative study on energy derived from biomass   93   58d7 4930-93d3 -172e1b0d2cd & groupid=38322. accessed mar 2016. universidad de leon (2016), evaluation of biomass resources for municipalities (ebimun-bio-en-area). from www.unileon.es / investigadores/ proyectos-internacionales-de-idi/ebium. accessed jun 2016. waterford county council (wcc) (2013), ebimun: evaluation of biomass resources for municipalities. from www.bioenarea.eu / index.php?q =node/217. accessed mar 2014. wilkie ac (2008), biomethane from biomass, biowaste, and biofuels. bioenergy, asm press, washington. zafer s (2015), solid waste management in saudi arabia. from www.ecomena.org /solid-waste-manage-ment-in-saudiarabia. accessed jun 2016. civl19303_ii.qxd 1. introduction the saudi arabian standards organization, saso, is the sole standardization body in the country. the ministry of commerce, in association with saso, has established a comprehensive certification program termed the international conformity certification program. it is a combined assessment program, which deals with a preshipment inspection and a certification scheme. based on these inspection guidelines, various consignments are allowed into the country. it applies to approximately 66 categories of products, and includes new tires for all kinds of imported automobiles (ministry of commerce, 2001). _____________________________________________ *corresponding author e-mail: nratrout@kfupm.edu.sa saso standard no.448/1986 deals with the methods of testing that are performed on tires in order to obtain a conformity certificate (saudi arabian standard organization, 1986a). saso standard no. 447/1986, on the other hand, documents the values of these tests that have to be satisfied (saudi arabian standard organization, 1986b). these standards have been in effect since 1987 and are similar to the u.s. federal motor vehicle safety standard no.109 (national highway traffic, 2001). this federal standard specifies the tire dimension and laboratory test requirements for bead unseating resistance, strength, endurance, and high speed performance that are applicable to all passenger car pneumatic tires manufactured after 1948. the objective of this paper was to examine the conformity of imported tires with saso standards and to assess their physical characteristics based on these standards. ring research 1 (2004) 01-06rnal of engineethe jou conformity of physical characteristics of imported tires with government standards: a case study n.t. ratrout*1, s. anwar basha1, h. i. al-abdul wahhab1 and k. a. al-ofi1,2 1 department of civil engineering, king fahd university of petroleum & minerals, dhahran 31261, saudi arabia 2 research institute, king fahd university of petroleum & minerals, dhahran 31261, saudi arabia. received 19 march 2003; accepted 7 july 2003 abstract: the saudi arabian standards organization, saso, has established a set of tests that all passenger car tires have to pass in order to be allowed into the country. these tests are: visual inspection, dimension measurement, bead unseating, tensile strength and elongation, aging, endurance, and high speed performance tests. this paper describes the application of the first five tests on eight different brands of tires that are taken to represent the categories of “frequently failed” and “infrequently failed” tires in the country. the results showed that the dimensions of almost all the tested tires were somewhat different from those documented in the saso standards. on the other hand, all the tested tires passed the bead unseating and strength tests. all tires in the “frequently failed” category failed the tensile strength test. in contrast, all tires from the “infrequently failed” category passed this test. the average values of the test for these two tire categories were statistically different at a 95% confidence level. key words: saso tests, tire tests, bead unseating test, tires in saudi arabia. :¢ü∏îà°ùÿÿg¤g é¡dƒn~h íª°ùj »àm iò¨°üdg ägqé«°ùdg ägqéwg égréàœ ¿g öéj »àdg ägqéñàn’g øe áyƒª› ájoƒ©°ùdg ¢ù«jé≤ÿgh äéø°ugƒÿg áä«g â©°vh äéyô°ùdg ‘ πªëàdg qéñàngh ,òª©àdg ,ádé£à°s’g h ~°ûdg áehé≤e qéñàng ,áféàÿg qéñàng ,qéw’g oé©hg ¢sé«b ,…ô¶ædg ¢üëødg :ägqéñàn’g √òg πª°ûjh .áµ∏ªÿg áyƒª›h qééøf’g iòãc ägqéw’g áyƒª› éªgh ägqéw’g øe úàyƒª› á°sgq~d ’hg iqƒcòÿg ägqéñàng ¢ùªÿg ≥«ñ£jh ωg~îà°sg ” ábqƒdg √òg ‘ .á«dé©dg ¢ù«jé≤ÿgh äéø°ugƒÿg áä«g ~æy á∏é°ùÿg á«°sé«≤dg oé©h’g øy ée éyƒf ∞∏àîj ägqéw’g ™«ªl oé©hg ¿g qéw’g oé©hg ¢sé«b èféàf äô¡xg .qééøf’g á∏«∏b ägqéw’g ägqéw’g áyƒª› ‘ ägqéw’g ™«ªl â∏°ûa ~≤a iông á«méf øe .áféàÿg qéñàng äréàlg ~b úàyƒªûg ‘ ägqéw’g ™«ªl ¿g èféàædg äô¡xg é° jg .ájoƒ©°ùdg »fé°üm’g π«∏ëàdg èféàf äô¡xg é° jg .qéñàn’g gòg qééøf’g á∏«∏b ägqéw’g áyƒª› ‘ ägqéw’g ™«ªl äréàlg éªæ«h ~°ûdg áehé≤e qéñàng ré«àlg ‘ qééøf’g iòãc .áäÿéh 95 á≤k álq~h »fé°ümg ¥ôa ∑éæg ¿g ~°ûdg qéñàng ‘ úàyƒªéª∏d :á«méàøÿg ägoôøÿg.ájoƒ©°ùdg á«hô©dg áµ∏ªÿéh ägqéw’g ,áféàÿg qéñàng ,ägqéw’g ägqéñàng ,ƒ°sé°s saso standard no. 448/1986 describes eight tests for passenger car tires, namely visual inspection, dimension measurement, bead unseating, strength, tensile strength and elongation, aging, endurance, and high speed performance. this paper summarizes the results of the first five tests that were conducted on abroad sample of imported tire brands. the remaining three tests (aging, endurance, and high speed performance) are still in progress. 2. experimental design four infrequently failed tire brands and four frequently failed tire brands were selected (ratrout et al., 2002). this categorization was based on observing the brand of failed tires (blow out or tread separation) along four major roads in saudi arabia during three months in the summer of 2001. the number of failed tires for each brand together with its market share along the four roads was used as rates to rank the tire brands. the highest and lowest four ranks were selected for this research. it should be emphasized that because of the limited tire sample size in the previous project, the term of frequently and infrequently failed tires should only be taken as a rough measure for segmenting the tire population into two main categories. consequently, sampling from both categories, simultaneously, should increase the probability of covering a wide spectrum of tire properties and characteristics. the four infrequently failed tire brands were termed as brands a, b, c, and d, while the other four frequently failed tire brands were referred to as brands e, f, g and h. three tires from each brand were subjected to the testing program. all tires were radial, new, and less than six months old at the time of testing. with the exception of brands e and f, all tires were of the size 225/70 r15. this size is currently the most frequently used size in the country. however, brands e and f of this size are not imported (ratrout et al., 2002). consequently, the next most frequently used size, namely 205/65 r15, was selected for these two brands (ratrout et al., 2002). 3. test procedures the visual inspection test recommended by saso standard no. 448/1986 (saudi arabian standard organization, 1986a), requires the tire to be visually inspected by the naked eye, for any cracks or cuts in the tread or sidewalls. information and markings molded on the tire that are required by saso standard no.445/1986 (saudi arabian standard organization, 1986c), were checked. the dimension measurement aimed to compare the actual overall width and diameter with the standard values given in saso standard no.445/1986 (saudi arabian standard organization, 1986c). each tested tire was mounted on a measuring rim and to the pressure specified in the standard, and allowed to stand for 24 hours at room temperature (23 ± 2°c). before conducting the measurements, the pressure was adjusted to the original values and the tire width was callipered at six different points spaced equally around the circumference. this was done using a vernier caliper having an accuracy of 0.05 mm. the average value of these readings was reported as the section width. the maximum outer circumference was measured with a typical steel tape and then the value was divided by π to obtain the overall diameter. these values were then compared with the saso standard (saudi arabian standard organization, 1986c). the bead unseating test evaluates the tire resistance to unseating the tire bead from its rim when subjected to lateral forces such as during severe turning movements. for the tire sizes tested in this paper (section widths 205 mm and 225 mm), this force should not be less than 11,350 n (saudi arabian standard organization, 1986b). the tires were washed and dried, mounted on a clean rim and inflated according to the proper pressure specified in the standard at ambient temperature. a material testing machine (instron) was modified to host the tire-wheel assembly horizontally. the testing machine had a force measuring accuracy of 0.5%. the load was applied through the unseating block (shoe) to the tire sidewall at a rate of 50 mm/min, figure 1. the load was increased until the bead unseated or until the load of 11,350 n was reached, whichever happened first (saudi arabian standard organization, 1986a). for each tire, a sample of six readings was taken at six equally spaced points around the circumference. the strength test evaluates the strength of the reinforcing material in the tire. according to the previous saso standard 447/1986, the bearing energy for the tires tested should not be less than 294 n.m. each tested tire was mounted on an appropriate rim and inflated to the proper pressure as specified in the standard. the material testing machine (instron), which was used in the bead unseating test, was modified to support a cylindrical steel plunger of 19 mm diameter with a hemispherical end, figure 2. the plunger was forced into the tread as near to the centerline as possible at a rate of 50 mm/min. the distance the plunger moved was reported by an external linear transducer capable of providing an accuracy better than 0.1 mm. five measurements of force and penetration were taken at points equally spaced around the circumference of the tire. the breaking energy was calculated according to the following formula: [1] where: w = energy at break, n.m f =force at break, n p = perforation at break, mm 2 ring research 1 (2004) 01-06rnal of engineethe jourout et al. /rat 310 2 −××= pfw the tire was considered to pass the test when the breaking energy was not less than 294 n.m. if the plunger reached the face of the rim without any break occurring (plunger bottoms out), the tire was also considered to pass this test. the tensile strength and elongation test was conducted using a test specimen of the shape and dimensions shown in figure 3a and was taken from each tested tire using a press mold, figure 3b, in the direction of the circumference. the specimen was clamped between the two jaws of a material testing machine (lloyd lr300k), figure 4, and stretched at a rate of 500 ± 25 mm/min. the testing machine had a force measuring accuracy of 0.5%. the stretch in the specimen was reported by an internal extensometer resolution having an accuracy better than 0.005 mm. the tensile strength and elongation were calculated according to the following formula: [2] and [3] where: t = tensile strength, kg/mm2 l = maximum load, kg 3 ring research 1 (2004) 01-06rnal of engineethe jourout et al. /rat figure 1. bead unseating test figure 2. strength test a lt = elongation = 100 1 12 × − i ii average diameter (mm) brand size tire 1 tire 2 tire 3 standard•• (mm) a 697 697 697 696 b 689 689 689 696 c 693 693 698 696 d 225/70 r15 682 682 689 696 e 632 632 634 ** f 205/65 r15 631 631 634 ** g 679 687 679 696 h 225/70 r15 679 679 695 696 table 2. tire diameter * saso standard no. 445/1986. no tolerance limit given in standard, **no standard exists average tire width (mm) (% coefficient of variation) brand size tire 1 tire 2 tire 3 standard• (mm) a 232 (0.2) 232 (0.2) 232 (0.2) 220 b 229 (0.4) 229 (0.4) 229 (0.3) 220 c 220 (0.4) 220 (0.4) 221 (0.3) 220 d 225/70 r15 229 (0.2) 229 (0.2) 229 (0.2) 220 e 214 (0.3) 214 (0.3) 214 (0.3) ** f 205/65 r15 211 (0.3) 210 (0.5) 210 (0.3) ** g 224 (0.2) 231 (0.2) 231 (0.2) 220 h 225/70 r15 235 (0.2) 235 (0.2) 235 (0.2) 220 table 1. tire width a lt = elongation = 100 1 12 × − i ii * saso standard no. 445/1986. no tolerance limit given in standard, **no standard exists a = cross section area of test specimen, mm2 i1 = standard length i2 = length at breaking to pass the saso standard, the tensile strength of the rubber should not be less than 1.2 kg/mm2 and the elongation should not be less than 300% (saudi arabian standard organization, 1986b). it is interesting to note that the unit used in this test was kilogram rather than newton. this is inconsistent with the si system and other saso tests. 4. results visual inspection of the tires revealed that all of them were free from cracks or cuts. all information required by the saso standards, namely designation of dimensions and construction, designation of tire rim diameter, designation of load and speed, country of origin, date of manufacture and batch number, were present (saudi arabian standard organization, 1986c). the width measurements of the tested tires, table 1, showed that the coefficient of variation within each tire width measurement never exceeded 0.5%. this indicates that the width of each tire was uniform around its circumference. the diameter measurements in table 2 show that almost all tire brands had some degree of discrepancy with the saso standard. the standard does not specify the values for tires of width greater than 195 mm under the 65 series. consequently, 4 ring research 1 (2004) 01-06rnal of engineethe jourout et al. /rat 40mm20mm40mm 5 m m 2 5 m m 100mm 1 0 m m 1 0 m m r 25 m m r 11 m m 5 m m figure 3a. dimension of the tensile strength and elongation specimens average tire unseating values ( n) (% coefficient of variation) brand tire 1 tire 2 tire 3 a 11,900 (0.41) 11,870 (0.93) 11,530 (0.41) b 11,540 (0.74) 11,410 (1.52) 11,660 (1.63) c 11,570 (1.52) 11,760 (1.02) 11,880 (1.27) d 11,480 (1.34) 11,460 (1.38) 11,760 (1.66) e 11,750 (1.61) 11,890 (1.41) 11,590 (1.40) f 11,680 (1.65) 11,530 (1.72) 11,930 (1.39) g 11,670 (1.62) 11,900 (0.85) 11,810 (1.69) h 11,930 (2.07) 11,480 (1.53) 11,460 (1.57) table 3. tire unseating brand i-iv: frequently failed brands figure 3b. shape and mold of the tensile strength and elongation specimens elongation (%) tire brand tensile strength kg/mm2 (% coefficient of variation) tread samples** (% coefficient of variation) sidewall samples (% coefficient of variation) a 3.21 3.25 3.28 (1.1) 331.9 291.3 56.7 (65.5) 702.0 911.4 911.4 (14.4) b 2.80 2.82 2.84 (0.7) 584.2 593.8 562.3 (2.8) 498.4 464.3 475.0 (3.6) c 2.34 2.36 2.37 (0.6) 392.2 272.3 390.8 (19.6) 416.1 416.0 462.5 (6.2) d 1.37 1.33 1.31 (2.3) 113.8 113.5 50.1 (39.7) 400.4 402.7 325.2 (11.7) e 0.81 0.81 0.86 (3.5) 161.3 6.4 241.0 (87.6) 532.7 506.8 506.3 (2.9) f 1.19 0.90 0.77 (22.6) 296.4 240.0 313.9 (13.6) 405.9 316.4 375.7 (12.4) g 1.02 0.98 0.93 (4.6) 341.7 295.1 295.0 (8.7) 513.8 521.2 521.2 (0.8) h 0.51 0.53 0.53 (2.2) 41.0 88.4 66.0 (36.5) 511.1 582.8 578.2 (7.2) table 4. tire tensile strength and elongation ** the large variation in the elongation is due to the existance of groove on samples there were no standard values with which brands e and f could be compared. however, the average discrepancy for the remaining tires was 7.5 mm and 9.0 mm for the diameter and width measurements, respectively. surprisingly, the saso dimension standards do not mention any tolerance values for tire dimensions, hence almost all the tested tires did not comply with these standards. clearly, this is impractical from an engineering point of view. the standard should have some tolerance to account for the accuracy of measuring instruments and other random errors. the bead unseating test was easily passed by all tire brands, table 3. it is also interesting to notice that the results in table 3 show little variation within, and between, different brands. the coefficient of variation within brands was less than 2.1%. the average value for all brands (i.e. 24 tests) was 11,685 n with a coefficient of variation equal to 1.5%. the u.s. national highway safety administration reported that the bead unseating test and strength test seem to be ineffective in differentiating between radial tires and that these tests are more useful for bias ply tires (national highway traffic, 2002 ). in the strength test, the results for all tires were so good that the plunger reached the bottom (rim surface) without breaking the reinforcement material in the tread. this seemed to be consistent with what is being reported in the u.s. the national highway traffic safety administration has recently reported that this strength test, which was originally designed for evaluating the strength of bias ply tires, was not so effective in testing radial tires (national highway traffic, 2002). they also reported that “because of the belt package, radial tires far exceed the strength requirements of the test and many times the plunger bottoms out on the rim instead of breaking reinforcement material in the radial tire” (national highway traffic, 2002). the u.s. rubber manufacturer association members also reported conducting about 19,000 strength tests on radial tires, none of which failed (national highway traffic, 2002). the results of the tensile strength test for our study, table 4, show that the variation within each brand is small. the coefficient of variation was less than 5% for all brands except brand f. the latter had a value of 22.6%, table 4. brands a–d (infrequently failed) passed the strength requirement of 1.2 kg/mm2. on the other hand, none of the frequently failed tires (brands e-h) passed this requirement. the average strength was 2.44 kg/mm2 for the infrequently failed tires and 0.82 kg/mm2 for the frequently failed tires. these two averages are statistically unequal at a significance level of 0.05%, table 5. the elongation test results for the tread samples was inconsistent and the variation within brands was high. the coefficient of variation ranged from 88% for brand e and 5 ring research 1 (2004) 01-06rnal of engineethe jourout et al. /rat 100mm tire groove a a b b section a-a section b-b 5mm5mm varying depth based on the groove figure 4a. longitudinal cross-section of the tensile strength and elongation test infrequently failed tires frequently failed tires mean (kg/mm 2) 2.44 0.82 variance 0.551 0.04 observations 12 12 pooled variance 0.298 hypothesized mean difference 0 df 22 t stat 7.27 p(t ≤ t) one -tail 1.4 x 10−7 reject h o t critical one -tail 1.72 p(t ≤ t) two -tail 2.8 x 10−7 reject h o t critical two -tail 2.07 table 5. t-test analysis for tensile strength test figure 4b. tensile strength and elongation test 2.8% for brand b, table 4. a possible reason for this variation was the fact that the cross-sections of the samples were not uniform due to the existence of the groove in the tire tread, figures 4a and 4b. this resulted in a non-uniform strain along the sample length. with the exception of brand b, none of the brands passed the required 300% elongation value in all of the three tire samples tested. only 33% of the tested tires passed this test. more specifically, 50% of the tires in the infrequently failed brands category, and 17% from the frequently failed brands category, passed the elongation requirements. 5. conclusions the dimensions of almost all the tested tires were at variance with the saso standards. the average difference was 7.5 mm and 9.0 mm for the diameter and section width, respectively. however, it is not known whether these differences are acceptable from a safety and operational point of view. it might be beneficial for the standard to include some tolerance value for the dimensions, to overcome this shortcoming. all the tested tires easily passed the bead unseating and the strength tests. both tests seem not to be stringent enough for radial tires. the infrequently failed tires passed the tensile strength test with an overall average value of 2.44 kg/mm2. whereas, none of the frequently failed tires was able to reach the 1.2 kg/mm2 strength required to pass this test. the overall average tensile strength of these frequently failed tires was 0.82 kg/mm2. this was statistically different from the average value of the infrequently failed tires at a 95% confidence level. consequently, this tensile strength test seemed to be useful in the screening of imported radial tires. the elongation test results were inconsistent for the tread samples because of the existence of grooves in the tire tread. due to the erratic elongation results, another set of specimens were taken from the sidewall of the tire to avoid the effects of the groove. the results of these specimens appeared to be more consistent and to have less variation within brands. the coefficient of variation was less than 15% for all the tested brands, table 4. the use of sidewall specimens to measure the elongation gave better results. in this case, some adjustment to the specifications may be needed to determine the allowable minimum value since all results passed the minimum required elongation of 300%. 6. recommendations our research study indicated that the saso standards need to be critically reviewed and updated. dimension standards should be completed for all tire sizes and reasonable tolerance in the dimension standard should be considered. all units in the saso standards should be consistent and should follow the si system. the bead unseating test and the strength test are not stringent enough for radial tires. the merits of these two tests for radial tires should be studied and appraised. in the tensile strength and elongation test, it is recommended to use the sidewall specimens in measuring elongation and to adjust the passing minimum value accordingly. acknowledgements the authors would like to acknowledge the financial support of king abdulaziz city for science and technology and king fahd university of petroleum & minerals. references ministry of commerce & saudi arabian standards organization, 2002, "international conformity certification program for exporting countries: comprehensive procedures and guidelines," august 2001 (www.iccp.com). national highway traffic safety administration, department of transportation, 2001, "federal motor vehicle safety standard no. 109; new pneumatic tires," code of federal regulations, title 49, volume 5, revised as of october 1, 2001 from the u.s. government printing office via gpo access (http/frwebgate.access.gpo.gov). national highway traffic safety administration, 2002 "notice of proposed rulemaking: federal safety standard; tires," docket no. nhtsa-00-8011, federal register: march 5, 2002 (www.nhtsa.dot.gov/cars/rules/rulings/tread/mileston es/ 66frpg10049 .html). ratrout, n.t. et al., 2002, "tire failure reasons and consequences," 1st interim report, king abdulaziz city for science and technology, 1423h. saudi arabian standards organization, 1986a, "passenger car tires, part 3: methods of test," saso standard no. 448/1986, riyadh, saudi arabia. saudi arabian standards organization, 1986b, "passenger car tires, part 2: general requirements," saso standard no. 447/1986, riyadh, saudi arabia. saudi arabian standards organization, 1986c, "passenger car tires, part 1: nomenclature, designation, marking, dimensions, load capacity and inflation pressure," saso standard no. 445/1986, riyadh, saudi arabia. 6 6ring research 1 (2004) 01-0rnal of engineethe jourout et al. /rat the journal of engineering research (tjer) vol. 14, no. 2(2017) 156-165 doi: 10.24200/tjer.vol.14iss2pp156-165 an enhanced smith predictor based control system using feedback-feedforward structure for time-delay processes m. mohammadzaheria and r. tafreshi*b a department of mechanical & industrial engineering, college of engineering, sultan qaboos university, muscat, oman. b reza tafreshi, mechanical engineering, texas a and m university at qatar, po box 2387, doha, qatar. received 17 march 2017; accepted 1 october 2017 abstract: this paper presents enhanced smith predictor based control systems (spbcss) for processes with a time-varying or fixed time delay. this work focuses on improving the arrangement and asynchrony of spbcs components rather than the design of the predictor and the feedback controller, which have been well discussed in the literature. the proposed control system advances spbcs through implementation of two design recommendations: (i) replacing the classical feedback controller by a feedback-feedforward control system, and (ii) synchronizing the reference signal and the predicted output. as a result, common shortcomings of spbcss or control systems based on pade approximation, i.e. the trade-off between performance and steady-state error, and instability associated with non-minimum-phase systems do not exist in the proposed spbcs. the superior performance of the proposed control system is demonstrated with two examples: temperature control of an infrared dryer (a system with fixed time-delay) and air-fuel ratio of a lean-burn spark-ignition engine (a system with time-varying delay and lag). the proposed control system is shown to clearly outperform the conventional spbcs and internal model control (imc) pid based on pade approximation for aforementioned examples and performs satisfactorily in the presence of noises, actuator saturations, and severe model inaccuracies. keywords: smith predictor; time delay system; dryer; engine. التغذية -نظام املراقبة باستخدام جهاز التغذية الراجعة احملسن للتنبؤ القائم على نظام مسيث املرتدة لعمليات تأخري الوقت ب،*تفرشي، رضا أظاهريمرتضى حممد للعمليات اليت بها (spbcss)نظام املراقبة احملسنة للتنبؤ القائم على : تعرض هذه الورقة البحثية أنظمة مسيثامللخص بدال من تصميم نظام spbcss تأخري زمين متغري أو ثابت. ويركز هذا العمل على حتسني الرتتيب وعدم التزامن ملكونات التنبأ ومراقبة التغذية الراجعة اليت مت مناقشتها بشكل جيد يف األدبيات. ويقدم نظام املراقبة املقرتح نظام سبكس من ( االستعاضة عن نظام مراقبة التغذية الراجعة الكالسيكية بنظام مراقبة التغذية 1ني للتصميم هما : )خالل تنفيذ توصيت أو (spbcss)( مزامنة اإلشارة املرجعية والنتائج املتوقعة. ونتيجة لذلك ، فان أوجه القصور املشرتكة يف نظام 2و ) سبقةامل بية لنظام بايد, أي املقايضة بني األداء وخطأ احلالة الثابتة، وعدم االستقرار أنظمة املراقبة القائمة على التقديرات التقري املقرتح. ويتجلى األداء املتفوق لنظام املراقبة املقرتح يف مثالني هما : (spbcss)املقرتن بالنظم غري الدنيا , ال توجد يف نظام خري زمين ثابت( ونسبة وقود اهلواء الناتج عن اشتعال شرارة مراقبة درجة احلرارة يف جمفف األشعة حتت احلمراء )نظام تأ (. وقد تبني أن نظام التحكم املقرتح يتفوق بشكل التباطؤ االحرتاق الضعيفة للمحرك )نظام ذو متغريات زمنية يف التأخري و التقريبية لألمثلة القائم على التقديرات (imc-pid) التحكم الداخليالتقليدي و نظام (spbcss)واضح على نظام املذكورة آنفا، كما أنه يقوم باداءه بشكل مرض يف وجود الضوضاء والتشبع يف احملركات والقصور احلاد يف دقة النموذج. ، احملرك ،اجملفف ، نظام تأخري الوقت : تنبؤ مسيثالكلمات املفتاحية * corresponding author’s e-mail: reza.tafreshi@qatar.tamu.edu m. mohammadzaheri and r. tafreshi 157 1. introduction a wide variety of dynamic systems witness dead-time or delay. so called time-delay systems (zhong 2006; wu et al. 2010) include chemical processes, engines, manufacturing, micro-aerial vehicles and telecommunication systems, etc. (liu 2014; sanz et al. 2017). timedelay is generally regarded as a main source of instability and poor performance in control systems (li and desouza 1997). in order to effectively address this control issue, the difference between two categories of time-delay systems needs to be recognised. in the first category, states of a system influence their time derivative with a delay. this type of delay is mainly network-induced, i.e. it is resulted by the time needed to exchange data between devices in electrical/communication networks. the control of this type of time-delay is addressed using lyapunov-krasovskii and razumikhin theorems (wu et al. 2010) and is beyond the scope of process control and this paper. in the second category, the control inputs influence the time derivative of the system states with a delay, i.e. a noticeable time is needed for the actuator(s) to affect the states and the output (zhong 2006). this type of time-delay is a prevalent concern in process control (bequette 2008) and is studied in this paper. two main approaches have been employed to control the second category of time-delay systems: pade approximation and predictive methods. in the first approach, the term presenting the delay is approximated using pade technique (bequette 2008). then, the resultant model is used in control system design. pade approximation results in model inaccuracy and transforms a stable minimumphase system to a non-minimum-phase one with a higher order (ebrahimi et al. 2012). on the other hand, a predictive method, such as a smith predictor (smith 1957), does not add the aforementioned restrictions; hence, it is expected to present higher performance. smith predictors were originally used for stable siso systems, but further research extended their application to unstable systems (kwak et al. 1999, mataušek and ribić 2012) and systems with long delays (astrom et al. 1994). as a result, smith predictors were suggested to be a part of control solution for any stable or unstable time-delay system (depoar 1985; zhong 2006). despite all the advantages and advances of spbcss, there are reports showing the superiority of feedback control systems based on pade approximation over spbcss. for example, in the control of linear systems with identifiable time-varying delays (ebrahimi et al. 2014a; ebrahimi et al. 2014b; mohammadzaheri et al. 2015a and 2015b). this research investigates and identifies two reasons behind relatively poor performance of spbcss and proposes two design recommendations to resolve the identified issues. unlike most of the literature on spbcss, the focus of this research is on the structure of spbcss rather than the design of the predictor or the feedback controller. the proposed control system, an enhanced spbcs, is assessed on systems with fixed or time-varying delays. a thorough assessment shows the superior performance of the proposed control system. 2. shortcomings of spbcs figure 1 shows a typical smith-predictor-based control system (spbcs), where g(z) and ˆ ( )g z are the plant original and approximated transfer functions without the delay, respectively, r is the delay, and c(z) is a linear classical controller designed for g(z). regardless of the methods used to design the predictor and the classical controller, this structure has two main shortcomings causing a relatively low performance: (i) design of a feedback classical controller (a vital part of spbcs) is always a trade-off between performance and steady-state error. moreover, windup phenomenon (the influence of actuator’s saturation on integrals (aryan et al. 2010)) needs to be compensated. these limits influence the performance of the spbcs; where other methods, e.g. the ones based on pade approximation, may use non-classical controllers (e.g. sliding mode (ebrahimi et al. 2014b)) with no trade-off or windup issues. (ii) figure 2, a re-arrangement of fig. 1, demonstrates that the reference yd (k) is actually compared with the estimated output at a different time, ˆ( )y k r , i.e. there is an asynchrony in conventional spbcs. an enhanced smith predictor based control system using feedback-feedforward structure for time-delay processes 158 figure 1. a conventional smith-predictor-based control system. figure 2. another representation of the common smithpredictor-based control system shown in fig. 1, assuming the discrepancy of g(z) and ˆ ( )g z is negligible. both aforementioned shortcomings are resolved in the proposed structure of spbcs. 3. methodology the following design elements are recommended to address the shortcomings listed in the previous section. recommendation 1: a feedback-feedforward control law replaces the feedback controller of c(z) in fig. 2 (which is designed for the plant without delay or g(z)). the recommended control system is proposed in the following corollary. this removes trade-off and windup issues. corollary: a process modelled by (1) or its discrete equivalent, is asymptotically stabilized by a control law of (2) if condition (3) is satisfied: ( ) , 1 b g s ts   (1) * ( ) ( ) ( ),u t ke t u t  (2) 1 0, bk t   (3) where ( ) ( ) ( ) d e t y t y t  . subscript d refers to desired status ( , d d y y y y  ) and u*(t) is derived from (4): * ( ) ( ) ( ). d d ty t y t bu t  (4) proof is in the appendix. accordingly, if, in (4), b>0 and t>0, with any positive k, stability is maintained, and steadystate error converges to zero. as a result, there is no trade-off between stability and steady-state error. besides, there is no integral in (2); hence, windup does not exist. the sole source of performance restriction is the incapability to achieve high values of k due to actuator limits. however, the feedforward controller, which generates u*(t), is model-based. hence, 158 m. mohammadzaheri and r. tafreshi 159 inaccuracy in model identification may lead to loss of performance as a disadvantage of the employed feedback-feedforward structure; this issue has been investigated for the proposed design in example 2. it should be noted that the control systems with feedforward-feedback structure have been investigated and used for different purposes (de luca 2000; bresch-pietri and krstic 2009; mohammadzaheri et al. 2009). in this paper, this structure is adopted in spbcss. recommendation 2: in tracking problems, i.e. where the reference, yd, is subject to change, the value of reference signal at r instants ahead of current time, yd (k+r), is used both to generate feedforward control commands and to be compared with the output of the smith predictor, as shown in fig. 3. the value of reference signal is normally available upfront in process plants (bresch-pietri and krstic 2009). this recommendation can be disregarded for regulation problems, where the reference is constant. 3.1 example1: an infrared dryer, a system with fixed delay figure 4 shows an infrared dryer with a pair of lamps/thermocouples. such dryers can be used to dry food and other bio products (reed et al. 2000; chuanzhu et al. 2005; sharma et al. 2005; roknul et al. 2014) or in a variety of other engineering applications (mirsepahi et al. 2017). using the experimental data and the least square method, the system transfer function has been approximated by (mohammadzaheri et al. 2015a): 1.3 52.7108 ( ) 1 152.812 s dryer h s e s    (5)                    65.0 8844.1 s 1 65.0 0190.0 65.0 9114.2 )s(dryerimccg   (6) figure 3. the proposed smith-predictor-based control system. figure 4. the infrared dryer. an enhanced smith predictor based control system using feedback-feedforward structure for time-delay processes 160 the input is the control command (in the range of 0 and 10v) which excites the lamp power amplifier, and the output is the measured temperature in ºc. equation (5) evidently meets the condition of the proposed corollary, i.e. (3), with any k > 0. three control systems were examined on the dryer model: (i) a feedback pid controller, (6), designed through pade approximation and internal model control (imc), with the method detailed in (bequette 2008): (ii) a conventional spbcs (fig. 1) with an imc pi controller of (7) (a continuous equivalent of c(z) in fig. 1): 2.8991 0.0190 1 ( ) dryer c s s    , (6) where λ=1 was chosen empirically for both (6) and (7). (iii) the proposed enhanced spbcs (fig. 3) with feedforward component of u*=0.0190 yd, derived for the time-delay process presented by (5) at the steady state desired situation of 0 d y  . with such a feedforward component, the condition presented by (3) is satisfied for any positive feedback gain of k. a feedback gain of k = 1 was chosen empirically to maintain reasonable trade-off between fast response and sensitivity to noise. the mean of absolute controller error (mace) was used as the control performance criterion: ( ) ( ) mace f i t t d t t f i y i y i dt t t       . (7) where tf and ti are final and start times of operation. figure 5 and table 1 compare the performance of three controllers with and without a random noise in the range of [-3 +3]ºc and with the fixed sampling frequency of 100 hz. feedback pid imc fails to control the system in presence of noise. the proposed spbcs outperforms both conventional spbcs and pid imc feedback with 9% to 22% less mace. figure 5. performance of three control systems on an infrared dryer model without noise, where conventional imc spbcs and feedback imc pid almost match. table 1. mean of absolute control error for different controllers. mace [ºc] noise-less with noise imc pid 2.0068 failure conventional imc spbcs 1.9911 2.2875 enhanced spbcs 1.5686 2.0898 m. mohammadzaheri and r. tafreshi 161 3.2 example 2 – air to fuel ratio of a leanburn engine, a system with timevarying delay and lag the second example is control of normalized air-fuel ratio (afr) of a lean-burn spark-ignition (si) engine. the importance of this control problem is on rise due to increasing concerns and new regulations about emissions (pace and zhu 2014; lin et al. 2017) and fuel consumption (ghodsi et al. 2017). normalized afr is defined as following (zope et al. 2011). ,a s f m1 afr = afr m (8) where a m and fm are air and fuel mass flow rates, respectively, and afrs is the stoichiometric air-fuel ratio which is commonly 14.7 for gasoline engines (ebrahimi et al. 2012). input afr to the engine can be calculated using measured fuel and air mass flow rates. the output afr is measured by the universal exhaust gas oxygen (uego) sensor (fig. 6). the input and output afr relationship is approximated by a first-order transfer function (zope et al. 2011). 1 ( ) . 1 s fuel path h s e ts    (9) where lag (t) and time delay (θ) can be estimated as following: , 120 , a t n n m            (10) where n is engine rotational speed, and  and are coefficients which are identified empirically with the values of  = 997.4 and β = 0.0305 lb for the investigated engine (ebrahimi et al. 2012). figure 7 shows experimental data of engine speed and air mass flow as well as the time delay and lag time collected in a typical federal test procedure (ftp) for an f-150 ford truck with a v8 4.6l lean-burn si engine (ebrahimi et al. 2012). figure 6. a schematic of air-fuel path in a leanburn si engine (ebrahimi et al. 2014). figure 7. engine speed and intake air flow mass rate according to a typical federal test procedure (ftp) and time delay (θ) and lag time (t) calculated based on (11).  161 an enhanced smith predictor based control system using feedback-feedforward structure for time-delay processes 162 for the aforementioned engine and operating conditions, depicted in fig. 7, a feedback imc pid was designed for (10). furthermore, an imc pi controller was designed for (10) excluding the exponential term to be used in the conventional smith arrangement depicted in fig. 1. for these controllers, the values of λ were empirically found as 1.2 and 0.1, respectively. equation (10) evidently meets the condition of the proposed corollary, i.e. (3), with any k > 0. therefore, the enhanced spbcs (depicted in fig. 3) is applicable on this system. the feedforward control command is u*=yd, based on (4 and 10). a feedback gain of k=1 was also empirically chosen. a random noise in the range of ±0.5% of afrs (fiengo et al. 2005), i.e. ±0.0735, was considered to assess the noise rejection of different control systems. a fixed sampling frequency of 12.5 hz, which is achievable by commercially available actuators (woodward 2014) was also chosen. figure 8 and table 2 compare the performance of three investigated control systems with the control input saturation limits of 0 and 6. in this example, the coefficients of  and β in (11) may be inaccurate due to imperfection of their identification process. in order to evaluate the effect of the inaccuracies, 20%, 30% and 50% of identification error were considered. then, the control systems were designed based on inaccurate lag (t) and delay (θ) obtained from the erroneous  and β (fig. 9). as demonstrated in fig. 10 and table 3, the proposed control system exhibits satisfactory performance with inaccurate and β, where two other control systems witness either a significant drop in performance or total failure in most cases. table 2. mean of absolute control error for different controllers on a lean-burn engine. mace noise-less with noise feedback imc pid 0.138 0.295 conventional imc spbcs 0.122 0.286 enhanced spbcs 0.072 0.082 figure 8. performance of three control systems on an engine model without noise. m. mohammadzaheri and r. tafreshi 163 figure 9. delay and lag variations in the case of error in identification of  and β. figure 10. performance of the enhanced spbcs for afr control in the case of error in identification of  and β. table 3. mean of absolute control error with inaccurate parameters. control system mace for various errors in identification of  and β +30%, -30% -30%,+30% -30%,-30% +30%,+30% feedback imc pid 0.0147 0.0245 failure failure conventional imc spbcs failure 0.0411 0.0582 0.1454 enhanced spbcs 0.0090 0.0076 0.0096 0.0076 4. conclusion in this paper, two shortcomings of the structure of smith predictor based control systems were identified: (i) inherited limits of the classical feedback controllers and (ii) asynchrony between the reference and the predictor output. on this basis, two design recommendations were suggested to enhance spbcss: (i) replacing the feedback controller by a feedbackfeedforward structure and (ii) replacing reference signal at the current time by the reference signal at the target time of prediction. within a corollary, the proposed control system an enhanced smith predictor based control system using feedback-feedforward structure for time-delay processes 164 was proved to produce a high performance with guaranteed stability for a wide class of timedelay processes. the enhanced spbcs does not face issues of common control techniques for time-delay processes, i.e. windup, trade-off between performance and steady-state error, and non-minimum-phase systems (the latter is a result of pade approximation). the proposed control system was assessed for processes with fixed/time-varying delays, in the presence of parameter identification errors and/or severe noises and/or actuator limits. the proposed control system exhibited superior performance in all cases and evidently outperformed a number of well-designed controllers based on pade approximation or smith predictor. conflict of interest the authors declare no conflicts of interest. funding this work was supported by nprp grant from the qatar national research fund (a member of qatar foundation), grant number is 7-829-2-308. references aryan p, mohammadzaheri m, chen l, ghanbari m, mirsepahi a (2010), ga-imc based pid control design for an infrared dryer, chemeca adelaide, south australia. astrom kj, hang cc, lim b (1994), a new smith predictor for controlling a process with an integrator and long dead-time. ieee transactions on automatic control 39(2): 343345. bequette bw (2008), process control, modeling, design and simulation. new delhi, prentice hall of india. bresch-pietri d, krstic m (2009), adaptive trajectory tracking despite unknown input delay and plant parameters. automatica 45(9): 2074-2081. chuanzhu s, xiangyou w, chao g, lifang p (2005), development of far infrared dryer of 5hy model for vegetables [j]. transactions of the chinese society of agricultural engineering 5: 020. de luca a (2000), feedforward/feedback laws for the control of flexible robots, ieee international conference on robotics and automation san francisco, usa. depoar a (1985), a modified smith predictor and controller for unstable processes with time delay. international journal of control 41(4): 1025-1036. ebrahimi b, tafreshi r, franchek m, grigoriadis km mohammadpour j (2014a), a dynamic feedback control strategy for control loops with time-varying delay. international journal of control 87(5): 887-897. ebrahimi b, tafreshi r, masudi h, franchek m, mohammadpour j, grigoriadis k (2012), a parameter-varying filtered pid strategy for air-fuel ratio control of spark ignition engines. control engineering practice 20(8): 805-815. ebrahimi br, tafreshi j, mohammadpour m, franchek k, grigoriadis, masudi h (2014b), second-order sliding mode strategy for air-fuel ratio control of lean-burn si engines. ieee transactions on control system technology 22(4): 1374-1384. fiengo g, grizzle j, cook ja, karnik ay (2005), dual-uego active catalyst control for emissions reduction: design and experimental validation. ieee transactions on control systems technology 13(5): 722-736. ghodsi m, hosseinzadeh n, özer a, dizaj hr, hojjat y, varzeghani ng, sheykholeslami mr, talebian s, ghodsi mh, al-yahmadi a (2017), development of gasoline direct injector using giant magnetostrictive materials. ieee transactions on industry applications 53(1): 521-529. kwak hj, sung sw, lee ib, park jy (1999), a modified smith predictor with a new structure for unstable processes. industrial and engineering chemistry research 38(2): 405-411. li x, desouza ce (1997), delay-dependent robust stability and stabilization of uncertain linear delay systems: a linear matrix inequality approach. ieee transactions on automatic control 42(8): 11441148. lin jh, chen p, wang (2017), afr control of a gasoline engine using triple-step method, 29th chinese control and decision conference, chongqing, china. liu pl (2014), further results on delay-rangedependent stability with additive timevarying delay systems. isa transactions 53(2): 258-266. mataušek mr, ribić ai (2012), control of stable, integrating and unstable processes by the modified smith predictor. journal of process control 22(1): 338-343. m. mohammadzaheri and r. tafreshi 165 mirsepahi a, mehdizadeh a, chen l, o’neill b, mohammadzaheri m (2017), comparison of inverse modelling and optimizationbased methods in the heat flux estimation problem of an irradiative dryer/furnace. journal of computational science 19: 77-85. mohammadzaheri m, chen l, behnia-willison f, aryan p (2009), a design approach for feedback-feedforward control systems. ieee international conference on control and automation, christchurch, new zealand. mohammadzaheri m, chen l, mirsepahi a, ghanbari m, tafreshi r (2015), neuro‐ predictive control of an infrared dryer with a feedforward‐feedback approach. asian journal of control 17(5): 1972-1977. mohammadzaheri m, tafreshi r, ebrahimi b, grigoriadis (2015), a predictive strategy to control time-varying delay systems: leanburn engines, asme dynamic systems and control conference, columbus, ohio, usa. pace s, zhu gg (2014), transient air-to-fuel ratio control of an spark ignited engine using linear quadratic tracking. journal of dynamic systems, measurement, and control 136(2): 021008. reed tr, meyer wj, van em, hermann d (2000), infrared dryer system for printing presses, google patents. roknul as, zhang m, mujumdar as, wang y (2014), a comparative study of four drying methods on drying time and quality characteristics of stem lettuce slices (lactuca sativa l.). drying technology 32(6): 657-666. sanz rp, garcía qc, zhong, albertos (2017), predictor-based control of a class of timedelay systems and its application to quadrotors. ieee transactions on industrial electronics 64(1): 459-469. sharma g, verma r, pathare p (2005), mathematical modeling of infrared radiation thin layer drying of onion slices. journal of food engineering, 71(3): 282-286. smith oj (1957), closer control of loops with dead time. chemical engineering progress 53(5): 217-219. woodward (2014), fuel controls for engines. from http:// www.woodward.com / applications-engines.aspx. wu m, he y, she jh (2010), stability analysis and robust control of time-delay systems, springer. zhong qc (2006), robust control of time-delay systems, springer. zope r, mohammadpour j, grigoriadis k, franchek m, tafreshi r, masoudi h (2011), identification of air-fuel ratio dynamics in si engines using linear parameter varying techniques control and applications, vancouver, canada appendix: proof of the corollary eq. (1) results in ( ) ( ) ( )ty t y t bu t  (a1) let us consider * ( ) ( ) ( ).u t u t u t   subtracting (4) from (a.1) results in the error dynamics: ( ) ( ) ( )te t e t b u t   . (a2) with the control law of (2), ( ) ( ),u t ke t   (a3) then ( ) ( ) ( )te t e t bke t   . consequently,  0 0 1 1 ( ) ( ) ( ) ( ) exp bk bk e t e t e t e t t t t t                       as a result, if 1 0 bk t   , the error consistently decreases and the closed loop system is asymptotically stable. the journal of engineering research (tjer, vol. 14, no. 2 (2017) 105-114 doi: 10.24200/tjer.vol.14iss2pp105-114 design of a testing facility for investigation of drill pipes fatigue failure j. abdoa*, e. hassana, a. al-shabibia and j. kwakb a department of mechanical and industrial engineering, college of engineering, sultan qaboos university, muscat, oman. b qatar university, doha, qatar. received 9 march 2017; accepted 6 june 2017 abstract: drillstring and down-hole tool failure usually results from failing to control one or more of the vibration mechanisms. the solution starts with the ability to measure different modes of vibration, hence identifying different vibration mechanisms. lateral, torsion and axial are vibration modes that take place when drill pipes run into problems downhole. due to the three modes of vibration mechanisms such as bit bounce, stick-slip, lateral shocks, bit and bottom hole assembly (bha) whirl, parametric and torsional resonance occur. understanding the causes of the destructive loads is the main step towards developing approaches to prevent or reduce their effects, hence improving drilling performance. vibration modes and mechanisms lead to failure of the drill pipes, bha and drill bits. drill pipes fatigue failure is very common due to capability of producing all vibration modes and mechanisms. drill pipe and downhole tool assembly failure usually result from failing to have power over one or more of these vibration mechanisms. a novel in house experimental setup has been developed to mimic downhole axial, lateral and torsional vibration modes and mechanisms in drilling operations. in this paper, we focus on the design and construction of the testing facility. a number of tests were conducted to validate the capability and performance of the test setup. drill pipe fatigue failure due to lateral cyclic stresses induced in the drill pipe has also been investigated and presented in this paper. the results show that operating on a rotation speed higher than 90% of the drillstring critical speed leads to yielding in the drillstring. keywords: drillstring experiment setup; fatigue failure; vibration modes. أنابيب احلفر اجهاداالخفاق الناتج عن من تصميم مرفق اختبار للتحقق بكواك ان, ج أالشبييب عبداهلل, أادريس حسن, *وأعبدمجيل أسفل البئر عن االخفاق يف السيطرة على واحد أو أكثر من آليات حفر يف جهاز احلفر و أداة خفاق : عادة ما ينتج االامللخص تعد االهتزازات االهتزازا املختلفة.حتديد آليات يتم خمتلفة وبالتالي امناط اهتزازاتاالهتزاز.و يبدأ احلل بالقدرة على قياس من االهتزازات اليت حتدث عندما تصادف أنابيب احلفر مشاكل أسفل البئر. و حيدث ااجلانبية و االلتوائية واحملورية أمناط جلانبية و انزالق القضيب و الصدمات ا املثقاب وهي : هتزاز ارتدادا الياتالرنني االلتوائي احلدودي نتيجة ثالثة أمناط من هو اخلطوة الرئيسية حنو تطوير اساليب ملنع أو احلد من آثارها اهلدمفهم أسباب أمحال ان جتمع االجزاء أسفل احلفرة. االخفاقعد يأنابيب ومثقاب احلفر. و أداء يف اخفاقوبالتالي حتسني أداء احلفر. كما أن اشكال وآليات االهتزاز تؤدي إىل عادة ما ينتج ف. ب احلفر من االمور الشائعة جدا بسبب امكانية قيامه جبميع انواع واآلليات االهتزازأنابي اجهادالناتج عن القدرة على التحكم يف واحد أو أكثر من آليات االهتزاز هذه. وقد االخفاق يفاخفاق أنبوب احلفر و ادوات أجزاء البئر عن انواع االهتزازات احملورية و األفقية وااللتوائية واآلليات املستخدمة يف حملاكاة (يف املختربحمليا ) مت تطوير جهاز جترييب قمنا بإجراء عدد من االختبارات للتحقق من مدى ويف هذه الورقة على تصميم وبناء مرفق احلفر. حنن نركز عمليات احلفر. الدوري اجلانيب الناجم عن جهادإلقدرة وكفاءة أداء اجلهاز. وقمنا أيضا بالتحقق من سبب اخفاق أنبوب احلفر بسبب ا و احلفر منطقة يف ٪90أنبوب احلفر وقدمناه يف هذه الورقة البحثية. وتظهر النتائج أن التشغيل على سرعة دوران أعلى من يف جهاز احلفر. مرونة إىل تؤدي احلرجة السرعة اجهاد جهاز احلفر: جهاز احلفر, اجلهاز التجرييب. االخفاق الناتج عن كلمات املفتاحيةال * corresponding author’s e-mail: jdabdo@squ.edu.om mailto:jdabdo@squ.edu.om j. abdo, e. hassan, a. al-shabibi and j. kwak 106 1. introduction drillstring consists of a drill pipe and bottom hole assembly (bha). the drill pipe is a long hollow shaft that drives a bit at the bottom of the wellbore. drillstring failure occurs in approximately one out of seven drilling rigs in low to moderate well depth and three out of seven in deep drilling wells. the cost of each failure ranges from $100,000 to $200,000 depending on well depth. due to the frequent and high cost of drillstring failure, investigations have focused on raising awareness of the factors causing drillstring failure, and eliminating or reducing the occurrence of such failure incidents. during drilling operations, drillstring interacts continuously with the rock formation, which results in severe shock and vibrations. drillstring shock and vibrations are identified as the principal sources of performance deterioration in oil and gas drilling operations (macdonald and bjune 2007; moradi and ranjbar 2009; reid and rabja 1995). drill pipe, bha and drill bit are affected by the vibration induced as a result of high input impact energy. this energy introduces different states of stresses, which translate into excess vibrations that may lead to failure. the severity of shock and vibrations depends on three parameters: shock magnitude, duration or length of time and frequency or number of shocks. the main causes of these vibrations include: contact and friction at the borehole/drillstring and bit/rock formation interfaces, eccentricity, imbalance, initial curvature in the drill collar sections, and various linear or nonlinear resonances. severe shock and vibration may subsequently produce fatigue and abrasive wear in drill pipes, drill bit damage and reduction of penetration rate rop (jardine et al. 1994; spanos et al. 2003; khulief and al-naser 2005). as a consequence, the drilling process becomes inefficient and costly. in a drilling operation, drillstring exhibits one or more of three common modes of vibrations: axial, torsion and lateral. preventing drill pipe failure calls for addressing all possible factors that can cause failure. sometimes this is as simple as managing loads and load capacities, and having a good inception of the drilling operation. however, with complex causes, much more is involved in order to prevent failure (jardine et al. 1994). regardless of vibration mechanisms, drill pipes happen to deteriorate as a result of flaws in one or more of the following areas: design, inspection, operation and characteristics of drilling operation. the mechanical and chemical environments surrounding the drillstring can have a major effect on whether or not failure may take place. vibration, mud type, dissolved gas content, salinity, and other factors play key roles in many drill pipes failure. lateral vibrations are recognized as a leading cause of drillstring and bha failures. lateral vibration also damages the well when bha large shocks impact the wellbore. several attempts have been made to theoretically study drill pipe lateral vibration in order to overcome the difficulties encountered during drilling process (berlioz et al. 1996; yigit and christofourou 1996). most of the mathematical models report slight progress in order to determine causes of drill pipe failure. however, not much has been reported about the procedures to accurately account for the effects of induced lateral vibration in drillstrings during drilling operations. very few modeling attempts have been undertaken in order to address the effects of lateral vibration on drill pipes, drill bit, and bha. it suffices to say that nearly all models are very much simplified and far away from addressing the real problem. one should note that modeling of such a long rotating element is non-linear and complex due to the fact that drillstring dynamics involve broad vibration profile, which includes axial, lateral and torsional modes (baryshnikov et al. 1997; wang at al. 2016). in addition, drilling process is very complex as many influencing parameters are not under drilling engineers' control (abdo and al-sharji 2015). some attempts have been made to study the non-linear dynamics of the drilling string. non-linear treatments were proposed in (al-hiddabi et al. 2003, abdo 2006, abdo et al. 2015, farhang and lim 2007). the authors undertook such a procedure to suppress torsional vibrations and reduce lateral vibrations of drilling sting non-linear modes. this paper aims at gaining a deeper understanding of drill pipe complex behavior under vibrations and its detrimental effect on drilling operation. the present work focuses on the design and development of an in-house testing facility. the developed facility is proposed to be capable of generating various vibration modes that a drillstring experiences when it runs into hard rock formation during drilling operations. the facility intends to provide further expansion for more comprehensive drillstring testing. it can readily accommodate other relevant effects such as wellbore and drillstring contact, drillstring interaction and stick-slip phenomenon. the controlled testing facility also aims to offer oil and gas industries with satisfactory answers and justifications on the drillstring performance under various loading and operating conditions. the investigations are expected to lead to better understanding of drill pipe behaviors. it also accounts for the number of cycles to fatigue design of a testing facility for investigating of fatigue failure of drill pipes 107 failure due to induced lateral vibration in the drill pipe under various loading conditions. 2. drillstring testing facility drilling processes are complex as many of their influencing parameters are not in control of drilling engineers. in many situations, mathematical formulations do not necessary lead to accurate explanations of failure causes. drill pipe fatigue failure is one of the most frequent failures that occur due to rapid and continuous cyclic stresses induced by vibration and high shock loads. mathematical and computational models have important roles in the investigation of vibrations in drillstrings. such models enable configuration of drillstring behaviour under range of loading, physical and boundary conditions. however, a combination of the drillstring geometry (long thin-walled rotating object) as well as forces and torques acting on it during drilling operations lead to complications in drillstring dynamic response and broad vibration profile that includes axial, lateral and torsional patterns. hence, there is a need to develop an inhouse testing facility that is capable of imitating drilling process and has the potential of generating the three different types of vibrations. the testing facility is also utilized to investigate drillstring connections behavior under all modes of vibrations. in the current setup, a braking mechanism was developed and assembled to the lower side of the facility in order to introduce torsional vibration in the drill pipe. the braking mechanism is employed to induce torsional vibration of various amplitudes in the drill pipe. lateral vibration is created due to the eccentric rotation of the drill pipe. the frequency and amplitude of the lateral vibration are controlled by varying the drillstring rotational speed. the following sections briefly discuss the testing facility design and construction. in subsequent sections, the effect of inducing lateral vibration of the drill pipe on its lifetime fatigue failure is experimentally investigated utilizing the developed testing facility. 2.1 experimental setup the experimental setup shown in fig.1 consists of the following: a testing pipe that represents the drillstring; support system that includes supporting plates fixed to the ground to provide stability for the facility, c-channels beams to connect the supporting tables; loading and rotation mechanism which includes motor, pulleys, belts, chucks, pillow bearings, connections, flanges and housing pipes; mild steel cylinders representing the wellbore and also used as stabilizers for the drillstring; a control unit and speed controller; sensing mechanism, that includes load cells and sensors for measuring top and bottom torques, rotation speed, number of cycles, and vibration amplitude; braking mechanism and pressure mechanism utilized for inducing torsional vibration in the drill pipe at various frequencies and amplitudes; and a data acquisition system which includes twelve channels data logger and computer. figure 1a shows an image of the facility while fig. 1b shows a schematic of the facility and its individual components and systems. the representations for the numbers indicated on fig. 1b are listed in table 1. the facility structural support consists of two 16 mm thick mild steel tables, two ring shape stabilizers and four c-channel beams connecting the tables to the stabilizers and providing stability to the facility. the c-channels are fixed to a concrete floor. the structural support is responsible for providing safe working environment and holding other mechanisms such as rotation and braking mechanisms. the rotation mechanism is responsible for rotating drill pipe and generating lateral vibration. three phase 12 hp electric motor rotates the pulley whereas the controller regulates the drill pipe rotational speeds. the v-belts rotate the housing pipe at a specified rotational speed hence the drill pipe rotates at the same rotation speed as it is connected to a chuck as shown in fig. 1. the three table 1. main components of the drilling facility. component/system #. nomenclature 1 computer 2 data acquisition system 3 3-phase electrical motor 4 motor speed controller 5 rotary shaft 6 pulleys and belts 7 chuck and flange assembly 8 stabilizer 9 specimen (drill pipe) 10 pillow block bearing 11 air compressor 12 pressure controller 13 hall effect sensor 14 brake on/off switch 15 ultrasonic sensor 16 linear pneumatic actuator 17 brake pedal 18 master cylinder 19 caliper and pads 20 brake disc j. abdo, e. hassan, a. al-shabibi and j. kwak 108 jaw chuck is connected to the housing pipe. a similar housing pipe and a chuck are fixed on the other end of the testing facility. the housing pipe guides the drillstring through these for about 0.5 m on each end of the facility to provide stability and rigidity to the drill pipe. the housing pipes are also guided through two pillow block bearings. such configuration allows for smooth rotation. a number of test specimens (mild steel pipes) representing drillstrings of various external diameters (1”, 2”, and 3”) were used during testing. the drill pipe is 6 m long; however, the length between the two chucks is 5 m. a total of 1 m of the drill pipe is inserted inside the two sides of housing pipes using three-jaw chuck and pillow block bearing. fig. 1a illustrates a zoomed-in image of the drill pipe, chuck and flange to housing pipes connection. 2.2 braking mechanism a brief description of the braking mechanism is discussed in this section. the braking mechanism is used to provide time varying resistance to simulate wellbore rock formation resistance and prompt torsional stress in the drill pipe. such engagement should also simulate stick-slip phenomenon in drilling operations. table 2 includes images of the braking system along with other attached subsystems and deliverables. (a) (b) figure 1. drillstring testing facility (a: image of the testing facility and loading mechanism, b: schematics shows individual components). design of a testing facility for investigating of fatigue failure of drill pipes 109 table 2. braking mechanism along with deliverables of the sub-mechanism. part/subsystem picture deliverables braking mechanism  less than 10 cm stroke length  pressure supply of 8 bar  time required to brake and frequent duration are possible  repeatability of mimicking stick-slip condition braking pads  easy assembly to a system.  parts are available.  potential to mimic stick-slip motion. control unit  brake actuator has a control switch  can be controlled automatically or manually  for automatic control, the delay time varies from 1second up to 1 hour measuring sensors  ultrasonic sensor and halleffect sensors have been installed  rpm at both ends of the string and deflection in the middle of the string are measured. 109 j. abdo, e. hassan, a. al-shabibi and j. kwak 110 #. nomenclature 1 brake disk 2 brake hub 3 booster stand 4 brake booster 5 pneumatic actuator 6 caliper stand 7 brake caliper 8 safety box figure 2. configuration of parts of the braking mechanism. the braking mechanism consists of a steel hosting structure and a linear pneumatic actuator that is connected to the brake pedal to initiate braking action. booster and master cylinder are both bolted into a c-channel, which is welded to the testing facility base. pneumatic double acting actuators with controlling valves are used in the system. required pressure needed to actuate the braking mechanism is provided by 12 bar compressor. a control unit, as shown in table 2, has been utilized to control stroke pressure, the braking time and the time of engagement. figure 2 configures exploded view of the brake mechanism parts using 3d cad modeling. table 2 also illustrates parts of breaking mechanism. 3. results and discussion the experiments were carried out under controlled loading and unloading conditions to examine the facility performance and investigate the effects of inducing lateral vibration on the drill pipe fatigue failure. one of the useful characteristics is to rotate the drill pipes near their resonances. this allows for quick testing using constructed facility despite the high strength of the component. the typical drill pipes diameters used in the tests are 1”, 2” and 3”. two steel rings are placed to constrain the drilling pipe in case of catastrophic failure for safety reasons. ultrasonic displacement sensors are used to measure the displacement at the middle of the drill pipe. two hall effect sensors are placed; one near the first chuck and the other near the second chuck. the sensors are utilized to record the pipe's rotational speed near the motor and the braking mechanism. displacement, deformation and rotational speed have offered a complete description of the drill pipe's dynamic vibration profile. the drill pipe rotational speed is constrained up to a certain critical rotation close to the drill pipe resonance. sets of 1”, 2”, and 3” diameters and 6 m long drill pipes were selected and utilized in the testing procedure. the first few experiments were performed by rotating the 1” diameter drill pipes at various rotational speeds. the lateral vibration amplitudes (deflection from the center axis of the drill pipe) were then measured using ultrasonic deflection sensors. the experiments were repeated using drill pipes of 2” and 3” diameters. the drill pipes' lateral deflections versus the rotational speeds for different pipe sizes is shown in fig. 3 the maximum lateral vibration amplitude (donated as amp) takes place at the middle of the drill pipe. it can be observed from fig. 3 that the deflection is lower when a larger drill pipe diameter is used and is higher when a smaller diameter pipe is used for corresponding rotational speed. the lateral vibration frequency increases almost linearly as the drill pipe rotational speed increases. figure 4 illustrates a 1” diameter drill pipe that has been subjected to a rotational speed near its resonance or critical speed. the deflection suggests that any further increase in the rotational speed might lead to unacceptable deformation of the drill pipe. theoretical resonance rotational speeds of the three drill pipes used in the experimental work were calculated using eq. (1) and presented in table 3. 𝑓 = λ 2π √gei μl4 (1) j. abdo, e. hassan, a. al-shabibi and j. kwak 111 table 3. theoretical resonance rotational speeds of the drillstrings. drillstring diameters in (inches) 1” 2” 3” resonance rotational speed (rpm) 660 1089 1518 90% of resonance rotational speed (rpm) 594 980 1366 0 4 8 12 16 20 0 100 200 300 400 500 600 a m p ( cm ) rpm 3" pipe 2" pipe 1" pipe figure 3. drillstrings behavior at various rotational speeds. figure 4. large deflection of the 1” diameter drill pipe at rotational speed close to its resonance. f: natural frequency g: gravitational constant e: modulus of elasticity i: moment of inertia 𝑙: span length λ: frequency factor, dimensionless μ: weight per unit length the second set of tests was performed to examine the fatigue failure due to induced lateral vibration in the drill pipes. drill pipes of three sizes were rotated at various rotational speeds and the number of cycles to fatigue failure were recorded (crakes were also visually observed). figure 5 shows number of cycles to fatigue failure for each drill pipe against various rotational speeds. it can be noticed from the figure that the number of cycles to failure is decreased for all drill pipes as the rotational speed is increased. meanwhile, the number of cycles to failure are increased as the diameters of drill pipes are increased. figure 5 also demonstrates that operating at a rotational speed near the resonance of j. abdo, e. hassan, a. al-shabibi and j. kwak 112 each drill pipe, which is approximately at 10% less than the resonance rotational speed as it is given in table 3, has significantly reduced the number of cycles to failure. the reduction in the number of cycles to failure in the vicinity of resonance is due to the high lateral vibration induced in each drill pipe as the rotational speed is about 90% of the drill pipes resonance rotational speed. the 90% resonance rotational speeds for the three-size drillstrings are presented in table 3. it is worth mentioning here that rotating drillstrings at a speed higher than 90% of their critical speeds leads to yielding of drillstrings. figure 6 shows a 1” diameter drillstring that has been yielded due to its operation at about 94% of its critical speed (620 rpm). the drillstring yielded after rotating the drillstring for a few thousands of cycles. measured lateral deflections (amplitude, amp) at 620 rpm were around 11.5 cm. figure 7 shows fatigue failure of a 1” diameter drill pipe when it was rotated near its resonance rotationed speed. the pipe failed at 53110 cycles due to operation at a rotational speed of 594 rpm which is corresponded to 90% of the drill pipe resonance rotation speed. it is apparent that operating near resonance rotational speed is a useful characteristic in testing as it allows quick test results and provides important information about the number of cycles to failure if operated near resonance rotation speed. 4.0e+04 4.0e+05 4.0e+06 4.0e+07 0 200 400 600 800 1000 1200 1400 n o . o f c yc le s rpm 3" pipe 2" pipe 1" pipe figure 5. drillstrings rotational speeds versus number of cycles to fatigue failure. figure 6. yielding of 1” diameter drillstring due to operating at 94% critical speed. design of a testing facility for investigating of fatigue failure of drill pipes 113 figure 7. fatigue failure of 1” diameter drillstring. 4. conclusion an experimental setup was developed to imitate the vibration modes induced in the drillstring when it runs downhole in oil or gas wells. the drilling operation setup is capable of investigating fatigue failure of drillstrings due to individual and coupled modes of vibrations. design and construction of the testing facility as well as its testability were the focus of this paper. drill pipe fatigue failures due to lateral cyclic stresses induced in the drill pipe have also been investigated. the performance of the testing facility was first validated. then, tests were executed to investigate the effects of rotational speeds and vibration amplitudes on drillstrings fatigue life for various drill pipe sizes. results have shown that the drill pipe size, rotational speed and lateral vibration amplitude have significant effects on the drillstring fatigue life. investigations have also illustrated that operating near resonance of drill pipe reduced the number of cycles to fatigue failure. results also showed that operating on a rotational speed higher than 90% of the pipes resonance rotational speeds prompts yielding in the drill pipes. it has also been established that drillstring is yielded due to its operation at 94% of its critical speed (620 rpm). the drillstring yielded after rotating the drillstring for a few hundreds additional cycles beyond its critical speed. lateral deflections (amplitude) at 620 rpm were almost 11.5 cm. the design and construction of such test facility as well as the conducted tests and obtained results from this study should offer appropriate technical indicators for oil and gas drilling operation. the work is currently under way to test and present full vibration profile including torsional, axial and stick-slip phenomenon. this can certainly lead to better control of drilling parameters and reduce instances of failure in drillstrings. conflict of interest the authors declare no conflicts of interest. funding this research was made possible by nprp grant # [8-690-2-298] from qatar national research fund (a member of qatar foundation). references abdo j, al-sharji h (2015), investigation of vibration effects on friction and axial force transfer of buckled rod constrained in a horizontal cylinder. tribology international 92: 317-327. abdo j (2006), modelling of frictional contact parameters of a mechanical systems. international journal of applied mechanics and engineering 11(3): 449-465. abdo j, abualsoud a (2011), analytical approach to estimate amplitude of stick-slip oscillations. journal of theoretical and applied mechanics 49(4): 971-986. al-hiddabi s, samanta b, seibi a (2003), non-linear control of torsional and bending vibrations of oil well drillstrings. journal of sound and vibration 265(2): 401-415. baryshnikoy a, calderoni a, ligrone a, ferrara p (1997), a new approach to the analysis of drillstring fatigue behaviour. spe drilling and completion 12(2): 77-84. berlioz a, der hagopian j, dufour r, draoui e (1996), dynamic behavior of drill-string: experimental investigation of lateral instabilities. transaction of the american society of mechanical engineers, journal of vibration and acoustics 118(3): 292-298. http://www.ptmts.org.pl/article.xsl?vol=49&no=4&page=971 http://www.ptmts.org.pl/article.xsl?vol=49&no=4&page=971 j. abdo, e. hassan, a. al-shabibi and j. kwak 114 farhang k, lim a (2007), a kinetic friction model for viscoelastic contact of nominally flat rough surfaces. journal of tribology 129 (3): 684-688. jardine s, malone d, sheppard m (1994), putting a damper on drilling’s bad vibrations. oilfield review 6(1): 15-20. khulief a, al-naser h (2005), finite element dynamic analysis of drillstrings. finite element analysis and design 41: 1270-1288. macdonald a, bjune v (2007), failure analysis of drillstrings. engineering failure analysis 14: 16411666. moradi s, ranjbar k (2009), experimental and computational failure analysis of drillstrings. engineering failure analysis 16(3): 923-933. nouby m, abdo j, mathivanan d, srinivasan k (2011), evaluation of disc brake materials for squeal reduction. tribology transaction 54(4): 644-656. reid d, rabja h (1995), analysis of drillstring failure. presented in spe drilling conference, amsterdam. spanos d, chevallier m, politis p, payne l (2003), oil well drilling: a vibrations perspective. the shock and vibration digest 35(2): 85-103. wang r, zang y, zhang r, bu y, li h (2016), drillstring failure analysis and its prevention in northeast sichuan, china. engineering failure analysis 1 18(4): 1233-1241. yigit as, christofourou ap (1996), coupled axial and transverse vibrations of oil well drillstrings. journal of sound and vibration 195(4): 617-627. http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&result=true&prevsearch=%2bauthorsfield%3a(nouby%2c+m.) http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&result=true&prevsearch=%2bauthorsfield%3a(abdo%2c+j.) http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&result=true&prevsearch=%2bauthorsfield%3a(mathivanan%2c+d.) http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&result=true&prevsearch=%2bauthorsfield%3a(srinivasan%2c+k.) http://www.tandfonline.com/loi/utrb20?open=54 microsoft word paper 3.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 124-136 an investigation of excessive seepage from the al-fulaij recharge dam, oman y.e.a. mohamedzein* , s.m. al-hashmi and a.a. al-rawas department of civil and architectural engineering, college of engineering, sultan qaboos university, muscat, oman received 22 july 2013; accepted 8 december 2015 abstract: the al-fulaij recharge dam is located on the al batinah coast in oman and was constructed in 1992. the dam is about 3.3 km long and 7.7 m high with a storage capacity of 3.7 million cubic meters of water. it is an earthfill dam with silty, sandy gravel fill in the embankment. excessive seepage of between 5,000–12,500 m3/day was observed during floods in 1993, and several sinkholes were noticed close to the upstream toe. remedial work consisting of an upstream blanket and a cutoff trench wall was performed in 2000. however, these remedial measures failed and almost the same seepage was noticed again during the impoundment. this paper investigates possible causes of the seepage using a finite element model. the input data for the model were collected from site investigations and field records during the construction and monitoring of the dam. the study reveals that the most probable cause of the excessive seepage is the presence of a permeable soil layer underneath the dam due to the dissolution of the gypsum material. keywords: cut-off walls, earthfill dam, recharge dam, remedial measures, seepage. אאאא ،אא  ∗אא،א، אאאאw١٩٩٢אאk ٣}٣א٧}٧٧}٣kאא    א א    k٥٠٠٠א ١٢٥٠٠    א   ١٩٩٣  ، kא א٢٠٠٠k אאאא، kאאאאאאkא אאאאkאאאאא kאאאאאאאאא   אא،א،א،א،wk،א * corresponding author’s email: yahiaz@squ.edu.om y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 125 1. introduction all earth and rock-fill dams are subject to seepage through the embankment, foundation, and abutments. many case studies of dams with seepage through their foundations have been reported in the literature (leonards 1987; piqueras et. al. 2012; richards and reddy 2007; unal et. al. 2008). seepage through the foundation is the main cause of failure of many dams (leonards 1987; richard and william 1985; richards and reddy 2007; zhang and chen 2006). one of the factors contributing to seepage through the foundation is the presence of soluble materials such as gypsum (piquerase et. al. 2012), because gypsum has a tendency to dissolve as a result of water impoundment and create seepage paths. a well-known example is the failure of the st. francis dam in 1928 due to the dissolution of the conglomerate with gypsum in the left abutment (yilmaz 2001). this paper presents an investigation of the possible causes of excessive seepage in the al-fulaij dam and its remedial works. the computer program seep/w (geo-slope, 2009) was used in the analysis. 2. background 2.1 the al-fulaij dam oman has an arid climate, and the main source of fresh water is rain as there are no rivers or lakes in oman. however, there are many seasonal streams where rain water flows in huge quantities for a short period of time. these streams discharge their water either in the sea or inland. to reserve these enormous quantities of water, the government started building recharge dams to retain water and store it for a certain period of time (usually for a maximum of two weeks) in order to recharge the underground water aquifers. more than 30 recharge dams have been constructed in oman (ministry of regional municipalities and water resources (mrmwr files, 1998-2010), and most are earth and rock fill dams. the al-fulaij dam is located in the barkarumays area of the al batinah coast, about 40 km west of muscat, oman (fig. 1). figure 1. location of al-fulaij dam (from geological map of ne oman, glennie et al. 1974 and shackleton et al. 1990). dam site an investigation of excessive seepage from the al-fulaij recharge dam, oman 126 table 1. main features of the dam. main dam total length 3346 m crest elevation 65.55 m above sea level (asl) max. height 7.70 m upstream slope (v:h) 1:2 downstream slope (v:h) 1:2.7 spillway type gabion mattress length 1150 m crest elevation 64.00 m above sea level (asl) max. height 6.15 m upstream slope (v:h) 1:2 downstream slope (v:h) 1:4 outlet conduits type ductile iron pipe, no. 2, dia. 80 cm reservoir capacity 3.7 million cubic meters catchment 117 km2 the dam aims to enhance the groundwater level in the barka-rumays area by retaining the water from seasonal streams that otherwise would be lost to the sea. construction of the dam was completed in september 1992 (mrmwr files, 1998-2010). during floods in 1993, significant seepage was observed downstream of the dam, and several sinkholes were noticed close to the upstream toe. the amount of seepage was between 5,000–12,500 m3/day (binnie partners overseas ltd. 2001) when the reservoir was full. remedial work was performed in 2000 consisting of an upstream blanket 10 meters wide and 0.6 meters thick of silty sandy gravel with 35% fines and a 2.5 m wide cut-off trench that was deep enough to reach the hard strata (limestone). the cut-off trench was filled with the same material as the blanket. however, these remedial measures failed for many reasons, and almost the same seepage was noticed again during the impoundment (binnie partners overseas ltd. 2001). the main features of the dam are listed in table 1. the dam is about 3.3 km long and 7.7 m high. the upstream slope of the dam is protected by rip rap whereas the downstream slope is protected with a 20 cm thick-layer of gravel. the total storage capacity of the reservoir is 3.7 million cubic meters. the dam has a spillway to discharge water freely during floods when the water quantities are beyond the dam storage capacity. the spillway is 1.15 km long with a maximum height of 6.15 meters. the upstream slope of the spillway is protected with rip rap whereas the downstream slope of the spillway is protected by a 30 cm layer of gabion mattresses. the earthfill dam’s embankment is made of a silty, sandy gravel fill with a maximum particle size of about 100 mm. the maximum dry density and optimum moisture content of the embankment fill as determined by the modified proctor compaction test varied from 2.15–2.20 g/cm3 and 6.1–6.9%, respectively. in situ density tests by the sand cone method gave degrees of compaction of 100%. one test carried out on the embankment fill material in a fixed-wall permeameter gave a coefficient of hydraulic conductivity (k) equal to 3.24 x 10-7 m/sec (electowatt-ekono 2006). 2.2 geology and subsurface conditions two site investigations were carried out at the dam site: one during the feasibility study in 1989 and the second in 2006 (electrowatt-ekono 2006). the two site investigations consisted of 11 boreholes and 19 test pits. according to the investigations in 1989 and 2006, the subsurface profile at the dam site consists of tertiary rocks, belonging to the jufnayn formation and is overlain by shallow quaternary deposits. the jufnayn formation comprises marl and limestone with a regional dip of 15–30° to the northwest. the quaternary deposits consist of recent alluvial deposits and conglomerate. the alluvium may reach a thickness of up to about three meters. it is composed of evaporate-rich y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 127 (gypsum and halite), loose to medium dense, silty fine to coarse gravel with cobbles and boulders. the conglomerate extends below the alluvial deposits. the depth to the conglomerate layer in trial pits varied generally between 0.5–3 meters. the gypsum material was encountered in many trial pits (fig. 2). table 2 shows, by weight, the gypsum content of some samples taken from different depths of the trial pits. the gypsum content varies by up to 28% and generally increases with depth. depth (m) visual description symbol 0.0 to 0.3 soft light brown silty clay 0.3 to 1.5 medium dense brown silty clayey sand with gravel and cobbles 1.0 to 1.7 weak grey conglomerate with gypsum, highly weathered figure 2. gypsum concentration in trial pit no. 8 (swissboring, 2006). table 2. gypsum content in some sample taken from the dam site (electrowatt-ekono, 2006). trial pits no. chainage (m) depth (m) gypsum content % 2 0+500 0.20-0.50 8.12 1.0-1.50 3.76 1 0+750 0.20-0.50 0.58 3 1+900 0.20-0.50 0.17 0.50-1.0 0.11 1.0-2.0 0.09 4 1+900 1.0-1.20 0.05 5 2+230 0.50-1.0 2.54 2.0 6.58 6 2+500 0.20-0.50 0.28 0.50-1.0 0.26 8 2+900 0.20-0.50 26.25 1.0 22.64 7 2+900 0.20-0.50 1.05 0.50-1.0 5.76 1.90 13.33 trial trench 1 2+230 0.80-1.0 2.0 1.20-1.70 27.76 1.30-1.70 23.93 trial trench 2 2+500 0.0-0.50 0.58 1.0-1.20 5.16 1.0-1.50 17.05 2.3 hydraulic conductivity the site investigation in 2006 (swissboring, 2006) included 14 packer tests (astm d4630, 2008) (table 3). the hydraulic conductivity was found to generally be low and variable. in the top four meters (i.e. the layer of alluvial and conglomerate materials), the hydraulic conductivity was 10-5 m/sec. below this depth (i.e. the limestone layer), the hydraulic conductivity varied from as low as 10-9 to 10-5 m/sec. an investigation of excessive seepage from the al-fulaij recharge dam, oman 128 2.4 the seepage problem during floods, significant seepage has been observed at many points downstream of the dam (fig. 3). as a result of the gypsum occurrence, numerous sinkholes, probably a few hundred within the reservoir basin just upstream of the dam, were formed. they are mostly circular but sometimes elongated, sinuous and channel–like (fig. 4). typically the sinkholes are 0.4–0.6 m in diameter with a depression of about 0.2 m deep, with the occasional sinkhole measuring 0.3–0.4 m deep (binnie partners overseas ltd. 2001). table 3. measured hydraulic conductivity using packer test (electrowatt-ekono, 2006). test no. borehole no. chainage (m) depth of section tested (m) hydraulic conductivity (m/s) 1 1 0+750 10.00-13.50 1.78x10-8 2 2 2+230 9.00-13.00 1.83x10-5 3 3 2+500 10.00-14.00 2.53x10-8 4 4 0+750 1.50-2.80 8.79x10-5 5 5 0+750 7.00-10.00 2.67x10-9 6 6 2+230 3.80-5.00 9.23x10-8 7 7 2+500 3.80-5.00 <1x10-9 8 8 2+500 3.00-8.00 3.13x10-6 9 9 2+230 4.00-8.00 <1x10-9 10 6 2+230 1.00-2.30 3.32x10-5 11 7 2+500 1.00-2.00 1.96x10-5 12 2 2+230 7.00-8.00 9.81x10-6 13 3 2+500 8.50-9.50 3.30x10-7 14 tp-5 2+230 0.80-1.80 1.24x10-5 figure 3. seepage from the dam. y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 129 figure 4. a sinkhole in reservoir area about 40 m from the dam heel. figure 5. crack in the crest of the dam at chainage 2+700. a crack in the crest of dam embankment about 5–7 m long and 80–100 cm deep running parallel to the dam axis was also noticed at about chainage 2 + 700 (fig. 5). the formation of this crack may have been due to some settlement in the foundation caused by gypsum leaching from the foundation. the dam has over spilled at least three times; the latest overspill was due to cyclone gonu in june 2007. 3. investigation of the seepage in the dam the most probable causes of excessive seepage in the al-fulaij recharge dam were evaluated based on the available information and data from previous site investigations and the inspection of the dam during the impoundment. the possible causes of the seepage in this dam were thus attributed to the site geology, dam design, and material used for dam construction and are listed here (fig. 6). an investigation of excessive seepage from the al-fulaij recharge dam, oman 130 • the presence of a permeable soil layer beneath the dam as a result of gypsum material dissolution and the insufficient depth of the key trench. • seepage through the dam body because of the segregated layers of fill. • seepage through the limestone or the hard strata as a result of cavities, faults, or open joints. figure 6. possible causes of seepage. table 4. the gap between the cutoff wall and the hard strata (electrowatt-ekono, 2006). cross section chainage (m) (start – end) average gap between foundation and hard strata (m) 1 400 450 0.18 2 525 575 0.8 3 675 700 0.8 4 725 800 0.88 5 925 975 0.34 6 1100 1325 1.55 7 1375 1500 1.31 8 1550 1600 1.05 9 1750 1850 1.02 10 1950 2075 1.50 11 2250 2300 0.03 12 2325 2375 0.39 13 2425 2475 0.46 14 2500 2700 0.94 15 2725 2900 1.54 1) permeable layer and insufficient depth of key trench 2) segregated layer 3) open joints, faults or cavities in conglomerate or limestone y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 131 figure 7. design case of the dam as modeled in the program (seep/w 2007). 0 1 2 3 4 5 0 2 4 6 8 10 12 14 16 s ee p ag e (m 3 /d ay /m ) cross section no. figure 8. the seepage modeling results of the design case of al-fulaij dam. the first cause is valid since all of the site investigations showed the presence of gypsum in the foundation of the dam. furthermore, according to the design, the foundation of the dam was not extended to the hard formation (i.e. the limestone) in some places. a second cause may be related to the dam’s construction since the silty sandy gravel used, if not mixed properly, will lead to seepage problems. however, if such a case had existed, then the seepage would have been noticed directly after the water impoundment, but it was not noticed. additionally, the water which seeps from the dam reservoir takes about two to three days to appear downstream. the third cause is not valid because many permeability tests were carried out during the site investigation in 2006 in the hard strata or the bedrock underneath the dam foundation. the coefficient of permeability obtained from the packer tests in these materials was found to vary from less than 10-9 m/s to 10-5 m/s (table 3). since the first potential cause, the presence of a permeable soil layer beneath the dam as a distance 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 5 e le va tio n 0 1 2 3 4 5 6 7 8 9 10 11 12 13 subsurface foundation (hard formation) surface foundation dam body stone filter key trench cross section no. s ee p ag e (m 3 / d ay /m ) an investigation of excessive seepage from the al-fulaij recharge dam, oman 132 result of gypsum material dissolution and insufficient depth of the key trench, is the only plausible solution, it is considered for further study in the following sections. 3.1 modeling to determine the most possible causes of excessive seepage the dam was modeled using the 2009 seepage modeling program seep/w2007 (geoslope international ltd. calgary, alberta, canada). seep/w2007 is a finite element software used for analyzing groundwater seepage and excess pore-water pressure dissipation problems. as a first step to model the dam, the locations along the dam where the key trench was not extended to the hard strata were identified. the effect of the gypsum material was introduced later in the analysis. table 4 shows the gap between the key trench and the hard strata at 15 cross sections along the dam. to show the effect of the presence of the permeable layer underneath the dam on the seepage quantity, the problem was modeled according to the design case and the actual case. the design case represents the dam without the causes of the seepage problem. this means that the presence of the gypsum material at the site was not considered since it was not evaluated properly during the design stage; hence, the values of the hydraulic conductivity (k) of the different material are as shown in table 5, and the case is shown schematically in fig. 7. table 5. materials properties obtained during feasibility study (mrmwr files (1998-2010)). no. material k (m/s) 1 dam body (silty sand with gravel) 3.24x10-7 2 surface foundation (silty sandy gravel) 3.40x10-5 3 subsurface foundation (limestone) 9.10x10-6 4 stone filter (fine) 1.00x10-3 the seepage modeling results of this case (fig. 8) show that the seepage quantities of the design case range from 0.50–2.450 m3/day/m (total seepage 1250 to 6125 m3/day). during the impoundment, 5,000–12,500 m3/day of seepage was noticed per day. based on this, trials were carried between 5,000 and 12,500 m3/day out using the modeling program to find out the corresponding k-value of the foundation material. the trials were conducted using the average cross section of the dam where the water head is 4.05 meters and the thickness of the permeable layer underneath the dam is 0.85 meters. a total seepage figure of 8,750 m3/day was used in the analysis. this equates to a seepage amount of 3.5 m3/day/meter for a 2,500 meter long dam. after conducting the trials, the average kvalue was estimated at 1.00 x 10-4 m/s (fig. 9). it is clear that the average foundation permeability was higher than the one obtained during the feasibility study. this is because the leaching process of gypsum material as a result of water flowing underneath the dam created voids within the dam’s soil structure. in this section, the actual case was simulated using the revised permeability value of the foundation material. the results that were obtained from the modeling of this case along with the design case are presented in fig. 10. as can be seen, the seepage quantities of the actual case ranged from 0.85–5 m3/day/meter with a total amount of seepage between 2,125–12,500 m3/day), which is about double the seepage for the design case. 4. remedial measures as mentioned before, the previous remedial measures did not work and almost the same amount of seepage was noticed. two methods of remedial measures are considered in the current study: the installation of a grouting curtain at the foundation of the dam and the use of a geomembrane at the upstream face of the dam anchored to the upstream cutoff. grouting is one of the most popular techniques to reduce the seepage amounts from storage dams. grouts such as clay, cement, bentonite or bentonite-cement can be successfully grouted in a formation having a permeability of more than 10-3 cm/second (shroff and shan 1993). one of the remedial measures suggested is to carry out grouting under the dam foundation (fig. 11). a permeability value of less than 10-6 m/second can be expected at the foundation after applying the grouting (shroff and shan 1993). a sixmeter deep hole is recommended with 1.5-meter spacing between the holes on the first round. more holes between them can be added as a second round depending on the results achieved during the grouting execution. the grouting was modelled in the program as a y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 133 0 2 4 6 8 10 12 14 0 1 2 3 4 5 6 s ee p ag e (m 3 / d ay ) hydraulic conductivity, k ( 10-4 m/sec) figure 9. trails to obtain the revised k-value of the foundation material. 0 1 2 3 4 5 6 0 2 4 6 8 10 12 14 16 s ee p ag e (m 3 / d ay /m ) cross section no. design actual (current) figure 10. the seepage modeling results of the design and actual cases. hydraulic conductivity, k (10-4 m/sec) s ee p ag e (m 3 / d ay ) cross section no. s ee p ag e (m 3 / d ay /m ) design actual (current) an investigation of excessive seepage from the al-fulaij recharge dam, oman 134 figure 11. grouting at the dam foundation as modeled in the program. cutoff wall as shown in fig. 11. the results of the dam modeling with the grouting curtain along with the current case are shown in fig. 13, which demonstrates that the seepage quantities after applying the grouting range from 0.5–2.6 m3/day/meter (500–2,600 liters/day/m) whereas the seepage quantities of the current case range from 0.85–5 m3/day/meter (850– 5,000 liters/day/meter). this shows 41–48% expected decrease in the seepage quantities. it is very important to mention here that grouting also has the advantage of improving the stability of the dam which is achieved by making the foundation denser since it will replace the materials which were washed away and prevent any depression or settlement of the dam foundation. a geomembrane can also be used to control seepage by inserting sheets of this material on the upstream face of the dam and extending it down to be anchored in the limestone (fig. 12). the figure shows the seepage profile of the dam when using the geomembrane. the value of hydraulic conductivity (k) used in the program for the geomembrane was 10-9 m/second. as can be seen from fig. 13, the seepage quantities after installing the geomembrane range from 0.4–1.8 m3/day/m (400–1,800 liters/day/meter) whereas the seepage quantities of the current case range from 0.85–5 m3/day/m (850–5,000 liters/day/meter). this shows an expected decrease in the seepage quantities between 53– 64%. this means that technically using the geomembrane is the best way to solve the seepage problem. however, from a financial point of view, this solution would be very costly because the installation process is very expensive and requires special techniques and care. the need to remove all the riprap and provide a smooth layer as well as supports will increase the total cost. al-hashmi (2010) presented a cost estimate for the two remedial options. figure 12. dam with geomembrane. distance 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 e le va tio n 0 1 2 3 4 5 6 7 8 9 10 grouting geomembrane y.e.a. mohamedzein, s.m. al-hashmi and a.a. al-rawas 135 0 1 2 3 4 5 6 0 2 4 6 8 10 12 14 16 s e e p a g e (m 3 /d a y /m ) cross section no. actaul (current) remedial (failed) grouting design geomembrane figure 13. seepage quantities profile of all cases. 5. conclusion the objective of the current study was to investigate possible causes for excessive seepage at the al-fulaij recharge dam and to suggest remedial measures. a finite element program (seep w 2007) was used to simulate the dam’s seepage, considering the most probable causes and suggested remedial measures. the most possible causes of seepage at al-fulaij dam are the presence of permeable alluvium deposits and conglomerates with cavities that formed due to the dissolution of the gypsum material and the insufficient depth of the dam key, which constitutes the dam’s foundation. as a result of alluvium deposits and the dam key’s insufficient depth, the seepage quantities increased to about double the expected amounts. there was also a sign of settlement that had taken place because of gypsum leaching from the dam’s foundation. two methods were proposed to control the seepage and were modelled by the program to validate their efficiency. the first method is the installation of a geomembrane at the upstream face of the dam. the second method is the installation of grout at the foundation of the dam. it is expected that using a geomembrane would reduce the current seepage by about 53– 64% while installing grout at the dam foundation would reduce the seepage by about 41–48%. grouting is the best option although a geomembrane would give better results as the total process of installation of the geomembrane is costly and needs special care as well a protection from sunlight and vandalism. grouting is more suitable and also has the advantage of enhancing the dam stability by filling the voids and cavities which were created by gypsum leaching. references al-hashmi smk (2010), seepage investigation of al-fulaij recharge dam using a numerical model. m.sc. thesis, department of civil and architectural engineering, sultan qaboos university, sultanate of oman. astm d4630 (2008), determining transmissivity and storage coefficient of low-permeability rocks by in situ measurements using the constant head injection test. annual book of astm standards, 04.08. binnie and partners overseas ltd. (2001), wadi fulayj (halban) dam, safety improvement works, report on causes of seepage. ministry of water resources, sultanate of oman. electrowatt-ekono (2006), safety evaluation of existing recharge dams in oman. final report, ministry of regional municipalities and water resources. geo-slope international ltd. (2009), seepage modelling with seep/w 2007. an cross section no. s ee p ag e (m 3 / d ay /m ) an investigation of excessive seepage from the al-fulaij recharge dam, oman 136 engineering methodology (4th ed.). alberta, canada. glennie kw, boeuf mga, hughes cmw, moody-stuart m, pillar wfh, reinhardt bm (1974), geology of the oman mountains. verhandelingen van het koninklijk nederlands geologisch mijnbouwkundig genootschap, part 1, 31: 182. leonards ga (ed.) (1987), special issue on dam failures, engineering geology. 24, nos 1-4. (also: leonards, g.a. (ed), (1985). proceedings international workshop on dam failures. purdue university, west lafayette, indiana, usa.) mrmwr files (1998-2010). dams in oman. ministry of regional municipalities and water resources. piqueras jam, sanz pe, menendez-pidal i (2012), water seepage beneath dams on soluble evaporite deposits: a laboratory and field study (caspe dam, spain). bulletin of engineering geology and environment 71: 201– 213. richard lv, william ek (eds.) (1985), seepage and leakage from dams and impoundments. proceedings of geotechnical engineering division, asce national convention, denver, colorado, usa. richards ks, reddy kr (2007), critical appraisal of piping phenomena in earth dams. bulletin of engineering geology and environment 66 : 381–402. shackleton rm, ries ac, bird pr, filbrand jb, lee cw, cunningham gc (1990), the batain melange of ne oman. in, the geology and tectonics of the oman region. robertson ahf, searle mp, ries ac (eds). geological society special publication 49: 673-696. shroff av, shan dl (1993), grouting technology in tunnelling and dam construction, 2nd ed. a. a. balkema publishers, rotterdam, brookfield. swissboring (2006). geotechnical investigations at al-fulaij dam. final report, ministry of regional municipalities and water resources. unal b, celik m, yildirim t (2008), evaluation of seepage problems of the armagan dam, kirklareli, northwestern turkey. bulletin of engineering geology and environment 67: 335– 344. yilmaz i (2001), gypsum/anhydrite: some engineering problems. bulletin of engineering geology and the environment 60(3): 227-230. zhang lm, chen q (2006), seepage failure mechanism of the gouhou rockfill dam during reservoir water infiltration. soil and foundation, japanese geotechnical society 46(5): 557-568. civl50917.qxd the journal of engineering research vol. 4, no.1 (2007) 17-22 1. introduction high performance concrete (hpc) represents a rather recent development in concrete materials technology. hpc is a concrete with properties which satisfy the performance criteria. generally, concrete with higher strength and attributes superior to conventional concrete is desirable in the construction industry. usually hpc is defined in terms of strength and durability (fidjestol, 1993). foster (1994) defined hpc as “a concrete made with appropriate materials combined according to a _______________________________________ *corresponding author’s e-mail: aloraimi@squ.edu.om selected mix design and properly mixed, transportedplaced, consolidated, and cured so that the resulting concrete will give excellent performance in the structure in which it will be exposed, and with the loads to which it will be subjected for its design life”. the strategy to produce hpc is to use mineral admixtures such as silica fume, as a partial cement replacement. silica fume reacts to form hydration products such that when introduced in the portland cement paste it can partially replace portland cement. silica fume has been used more and more in concrete due to its benefits in terms of strength and durability. there is a wealth of information compressive strength and surface absorption of high strength silica fume concrete under different curing conditions s.k. al-oraimi*, a.w. hago, h.f. hassan and r.taha department of civil and architectural engineering, sultan qaboos university, p.o. box 33, al-khoud 123, muscat, oman received 17 september 2005; accepted 15 april 2006 abstract: the effect of curing conditions and silica fume replacement on the compressive strength and the initial surface absorption of high performance concrete is reported. the silica fume contents were 5, 10, 15 and 20%, by weight of cement. four different curing conditions were used: air curing, control curing and two other curing conditions recommended by bs8110 and aci308-81. the cementitious material (binder) content was constant (400 kg/m3); the water/cement (w/c) ratio was also maintained at a constant value of 0.35; while the water/binder (w/b) ratio ranged from 0.35 to 0.28. the addition of silica fume enhanced the compressive strength significantly up to 30%. the 28-day compressive strength was found to be 69.9 mpa without silica fume and it was determined to be 89.9 mpa with silica fume under the standard curing condition. the 28-day compressive strength results under the control curing condition were found to be higher than the compressive strength for specimens cured under other curing conditions. the surface absorption (ml/m2.s) was found to decrease as the percentage replacement of silica fume was increased. control curing also decreases the surface absorption of water compared with air curing. concrete with silica fume was less sensitive to drying than that without silica fume. keywords: curing; high strength concrete; compressive strength; flexural strength; splitting strength; initial surface absorption; silica fume êéé° f’g ´gƒfg ∞∏àı á°vô©ÿgh 鵫∏«°ùddg qéîh ≈∏y …ƒà– »àdg á«dé©dg iƒ≤dg ägp áfé°sôî∏d »ë£°ùddg ¢ué°üàe’g h •é¨° f’g iƒb égéw .q h ø°ùm .±.ì ,ƒlég .h.cg ,»áô©dg .ñ .¢s ::áá°°uuóóÿÿggä髪c âe~îà°sg ~bh .ago’g á«déy áfé°sôî∏d »f~ñÿg »ë£°ùdg ¢ué°üàe’g h •é¨° fg iƒb ≈∏y 鵫∏«°ùdg ωg~îà°sgh êé° f’g ´gƒfg òkéj á°sgqo åëñdg gòg ‘ …ôœ ωƒµùg êé° f’gh »fgƒ¡dg êé° f’g :â檰 j êé° f’g øe áø∏àfl ´gƒfcg ™hqcg âe~îà°sg éªc ,ω~îà°ùÿg â檰s’g ¿rh øe áféÿg ‘ 20h 15 h 5 áñ°ùæh 鵫∏«°ùdg qéîh øe (ö©µÿg îª∏d ωgôlƒ∏«c 400) á«à檰s’g ioéÿg øe áàhék áñ°ùf âe~îà°sg ~bh . 81-308 … ¢scg »µjôe’g h 8110 ¢s ü êé£jèdg oƒµdg √qôb éªñ°ùm øjông úyƒfh áñ°ùæh •é¨° fe’g iƒb äogr 鵫∏«°ùdg qéîh áaé°veéhh .áféÿg ‘ 28 ¤g 35 øe á£hgôdg ioéÿg ¤g aéÿg áñ°ùf äòz éªæ«h áféÿg ‘ 35 ~æy âàñk â檰s’g ¤g aéÿg áñ°ùfh ≈∏y …ƒà– »àdg áfé°sôî∏d ∫éµ°séh騫e 89^9 ¤g 鵫∏«°ùdg qéîh ≈∏y …ƒà– ’ »àdg áfé°sôî∏d ∫éµ°séh騫e 96^9 øe •é¨° f’g iƒb äogr ~≤a ,áféÿg ‘ 30 ¤g π°üj iòñc øe iông ´gƒféh áé° æÿg äéæ«©∏d •é¨° f’g iƒb ≈∏y ~jõj ωƒµùg êé° f’g øe éeƒj 28 ~©h •é¨° f’g iƒb ¿g ~lh ~bh .êé° f’g øe á«°sé«b ´gƒfg ωg~îà°séh 鵫∏«°ùdg qéîh ωg~îà°séh é° jg ¢ü≤æj ¬fg ¿g éªc ,ω~îà°ùÿg 鵫∏°ùdg qéîh áñ°ùf äogr éª∏c ¢ü≤æj ¬fg ~lh ~≤a (™hôe îe πµd á«féãdg ‘ î«∏«∏e) »f~ñÿg »ë£°ùdg ¢ué°üàe’g éeg .êé° f’g .鵫∏«°ùdg qéîh ≈∏y …ƒà– »àdg ∂∏j øe ±éøé∏d á«°sé°ùm πbg 鵫∏«°ùdg qéîh ≈∏y …ƒà– »àdg áfé°sôÿg ¿gh ,ωƒµùg êé° f’g áá««mmééààøøÿÿgg ääooggôôøøÿÿgg.鵫∏«°s qéîh ,»f~ñÿg »ë£°ùdg ¢ué°üàe’g ,q飰ûf’g iƒb ,aéæëf’g iƒb ,•é¨° f’g iƒb ,iƒ≤dg á«déy áfé°sôÿg ,êé° f’g : 18 the journal of engineering research vol. 4, no.1 (2007) 17-22 in the literature related to the effect of different silica fume replacement on concrete properties. silica fume continues to be a popular element of high performance concrete. not only does it provide an extremely rapid pozzolanic reaction, but also investigators including detwiler and mehta ((1989), and goldman and bentur (1998) found that its very fine size also appears to provide a beneficial contribution to concrete. those researchers examined the effect of silica fume on mechanical behaviour of concrete. luther (1989) in his review examined durability effects, while fidjestol (1993) and khayat and aitcin (1992) have provided general reviews of the effects. all those investigators confirm findings that silica fume tend to improve both mechanical properties and durability. bentur et al. (1998) reported that the strength of silica fume concrete is greater than that of silica fume paste. they attributed it to the change in the role of aggregate as an inert filler, but, due to the presence of the weak interfacial zone, composite concrete is weaker than cement paste. a similar conclusion made by toutanji and elkorchi (1995) was that the increase in compressive strength of mortar containing silica fume as a partial replacement for cement, greatly contributes to strengthening the bond between the cement paste and aggregate. they reported that partial replacement of cement by silica fume and the addition of superplasticizers increase the strength of mortar, but have no influence on the strength of cement paste. on the other hand, some investigators have reported that the increase in concrete strength is due, in great part, to a higher quality of the cement paste matrix. cong et al. (1992) found that the replacement of cement by silica fume (up to 18%) and the addition of superplasticizers increase the strength of cement paste. darwin et al. (1989) found that the strength of both cement paste and mortar increases when 15% of the cement is replaced by silica fume. meeks and carino (1999) concluded that current practice and standards are based on studies related primarily to strength development characteristics of conventional concrete. most hpc, however, are fundamentally different from conventional concrete, because they typically have a low water/binder ratio (w/b) and one or more admixtures. since the composition of a high performance concrete differs from conventional mixtures, early age characteristics of the hydrating paste will also differ. therefore, existing curing practices may not be optimal for high performance concrete. a better understanding is needed of the role of an external supply of moisture and of the adequacy of membrane-forming compounds when a low w/b is used. one of the most controversial topics that emerged from the literature review by meeks and carino (1999) concerns the sensitivity of various properties of high performance concrete to different curing conditions. some researchers have reported that high performance concrete is more sensitive to the details of curing than normal concrete; whereas others have found the opposite to be true, at least for some properties. these differences may be attributed to the different experimental procedures that have been used. for example, hasni et al. (1994) reported that the use of silica fume makes high performance concrete more sensitive to different curing methods when considering both strength and durability properties. in addition, they hasni et al. (1994) reported that high performance concrete with silica fume is more sensitive to different curing methods than normal concrete for characteristics such as compressive strength and flexural strength, depth of carbonation, and microcracking. with respect to resistance to penetration of chloride ions, results showed that high performance concrete with and without silica fume, as well as normal concrete, were not sensitive to the curing method. research on the effect of silica fume on the mechanical properties of concrete is still evolving. curing effects on normal concrete have been studied extensively, and results have been established (meeks et al. (1999), however, the effect of different curing procedures on silica fume concrete still needs more research. the purpose of the research reported here is to improve our knowledge about the role of silica fume on enhancing the mechanical properties of concrete under different curing conditions. initial surface absorption tests (isat) were also carried out on concrete cubes to study the effect of silica fume on the permeability of concrete. 2. materials 2.1 cement the cement used in this study was ordinary portland cement (opc) obtained from a single batch conforming to the omanian standard specification (os 7:2001 opc) [15]. physical and chemical properties of this cement as supplied by the manufacturer are given in tables 1 and 2, respectively. 2.2 aggregate the fine aggregate was crushed gravel corresponding to zone 2 grading according to os 2:1982 [16]. it was purchased from a nearby crusher in al-khodh area which is from the same batch. the coarse aggregate was a crushed limestone supplied locally and has a maximum size of 10 mm. 2.3 silica fume silica fume consists primarily of amorphous (non-crystalline) silicon dioxide (sio2). the individual particles are extremely small, approximately 1/100th the size of an average cement particle. the silica fume (emsac 500s) was used in these experimental studies and it was added to the mix in a powder form. the quality of silica fume is specified by astm c1240-03a. physical and chemical properties of silica fume as supplied by the manufacturer are given in tables 1 and 2, respectively. 2.4 superplasticizer the superplasticizer (conplast sp430) conforms to bs 5075: part 3 and astm c494 as type a and type f, 19 the journal of engineering research vol. 4, no.1 (2007) 17-22 depending on dosage used. conplast sp430 is a chloride free, based on selected sulphonated naphthalene polymers, which had 40% active solids in solution. it was supplied as a brown solution, which instantly disperses in water. 3. experimental program 3.1 mix design the mix proportions chosen for this study were selected from a range of trial mixes and are given in table 3. the dry constituents were mixed in the rotating pan mixer with a speed of 40 rpm. the standardized mixing procedure was as follows: cement, sand and silica fume were mixed first; then water and coarse aggregate were added; and finally superplasticizers were introduced in the mix to achieve the desired workability. the overall mixing time was about 4 minutes. from each mix, thirty 100 mm cubes, three 100x100x500 mm prisms and three 150x300mm cylinders were cast. all specimens were compacted by means of a vibrating table. the frequency of the vibrating table was 2280 cycles/minute and all specimens were vibrated for 60 seconds. the specimens were covered with a wet burlap and plastic covers thirty minutes after casting, and later demoulded after 24 hours. the specimens were then cured as described below. 3.2 curing conditions four different types of curing regimes, table 4 were chosen to study the effect of different curing conditions on the compressive strength and surface absorption of high performance concrete. the curing conditions of c2 and c3 conformed to that recommended by bs 8110 and aci 308-81, respectively for ordinary portland cement (type 1), while c4 is the control curing which is a standard practice in our concrete laboratory. the room temperature was 25oc and the relative humidity was 60% for c4 curing. for all curing methods, curing in water was done prior to curing in air. the temperature in the laboratory was 25oc while the relative humidity varied from 35 to 60% during the time of curing and testing. all specimens were oven dried before testing for water surface absorption. the oven temperature was 100oc and the specimens were oven dried for 24 hours. 3.3 testing procedure the compressive strength was obtained by testing 100 mm cubes according to the bs 1881: part 116: 1983 (17). the cubes were tested in a 3000 kn testing machine at a rate of 2.5 kn/s. the splitting tensile strength was tested using 150x300 mm cylinders, according to bs 1881: part 117: 1983 (18). the flexural strength was tested using 100x100x500 mm prisms according to the bs 1881: part 118: 1983 (19). all these tests were conducted at the age of 28 days. the initial surface absorption tests (isat) were conducted in accordance with bs 1881: part 208: 1996 (20) on all specimens after 2 months from the date of casting. the initial surface absorption test was performed as an indicator for the durability of the mixes. the concrete cubes were initially oven-dried. a cap was attached to the concrete surface. this cap was connected to a reservoir with a water pressure head of 200 mm. the cap was also connected to a horizontal capillary tube. water was allowed to penetrate the concrete surface from the reservoir into the cap. at the end of 0, 30, 60 and 120 minutes, the reservoir flow was closed and the water was allowed material/test type ordinary portland cement silica fume specific gravity 3.05 2.2 fineness (cm2/g) 3,357 200,000 initial setting (minutes) 110 table 1. physical properties of ordinary portland cement and silica fume component ordinary portland cement (%) silica fume (%) sio2 21.94 93.2 al2o3 4.95 0.44 fe2o3 3.74 0.20 cao 62.33 na mgo 2.08 0.81 so3 2.22 0.45 k2o 0.56 1.28 na2o 0.32 0.59 tio2 0.17 na mn2o3 0.05 na cl 0.01 na loi 1.78 na ir 0.24 na c na 0.55 c (isolated) na 0.41 ph na 7.9 si (isolated) na 0.29 sic na 0.50 table 2. chemical composition of ordinary portland cement and silica fume na: not applicable 20 the journal of engineering research vol. 4, no.1 (2007) 17-22 to penetrate the concrete surface from the capillary tube. the test gives the water flow (ml/m2.sec) into the surface of the dry cube specimen using the calibration of the capillary tube. 4. results and discussion 4.1 strength 4.1.1 compressive strength the compressive strength results for the five mixes are presented graphically in fig. 1. the compressive strength under the control curing condition (c4) increased as the water/binder (w/b) ratio decreased. similar trends were also observed for the other three types of curing conditions. the use of silica fume as a replacement of cement increased the compressive strength of concrete up to 30%. the increase in compressive strength is due to the high fineness of silica fume, which allows the material to act as a filler and to provide an additional strength by reducing the void content in the cement paste matrix. increasing the silica fume replacement beyond 15% will not affect the compressive strength too much. considering the cost of silica fume and the effect of its percentage replacement on the compressive strength, one would expect that a 1015% of silica fume replacement is optimum. the effect of moist curing condition on the compressive strength of concrete is also presented in fig. 1. the concrete cubes with curing conditions of c1, c2, and c3 had a lower strength value compared to those with the control moist curing condition c4, for concretes both with and without silica fume. this is expected, since concrete with a prolonged moist curing should have a higher strength. it is also observed from fig. 1 that concrete with silica fume seems to be less sensitive to drying than that without silica fume. this is because of the rapid reaction happening between the silica fume and the calcium hydroxide released by c3s and c2s hydration. 4.1.2 flexural indirect tensile strength the 28-day flexural and splitting tensile strength results are presented graphically in fig. 2. these specimens were cured under the curing condition c4 (control moit curing) quantities kg/m3 constituents mix. 1 mix. 2 mix. 3 mix. 4 mix. 5 cement 400 380 360 340 320 fine aggregate 710 710 710 710 710 coarse aggregate 1190 1190 1190 1190 1190 water 140 133 126 119 112 silica fume -20 40 60 80 superplastisizer 10 ltr 11 ltr 12 ltr 13 ltr 15 ltr w/c 0.35 0.35 0.35 0.35 0.35 w/b 0.35 0.33 0.32 0.30 0.28 table 3. details of mix proportions days in air condition days in mold days in water compressive surface absorption c1 1 0 27 58-64 c2 1 2 25 56-62 c3 1 6 21 52-58 c4 1 27 0 31-37 table 4. curing conditions 30 40 50 60 70 80 90 100 c1 c2 c3 c4 curing condition c om p. s tre ng th , m pa . mix 1 mix 2 mix 3 mix 4 mix 5 figure 1. relationship between 28-days compressive strength and curing conditions 0 2 4 6 8 10 12 14 16 mix 1 mix 2 mix 3 mix 4 mix 5 mixture t en si le s tre ng th , m pa . flexural strength splitting strength figure 2. tensile strength at 28-days for different mixes 21 the journal of engineering research vol. 4, no.1 (2007) 17-22 to investigate the effect of silica fume on them. the flexural and splitting tensile strength varied from 8.6 to 13.9 mpa and from 3.9 to 7.8 mpa, respectively. the tensile/compressive strength ratio was found to increase as the percentage replacement of silica fume was increased. from fig. 2, it is also clear that the tensile strength increases as the compressive strength increases. 4.2 water absorption the water absorption of concrete was investigated by the initial surface absorption test (isat). the test results are presented in figs. 3, 4, 5 and 6 for the four curing conditions c1, c2, c3 and c4, respectively. the results obtained for condition 1 (c1), which is air curing, is shown graphically in fig. 3. it is clear that the flow of water decreased as the percentage replacement of silica fume was increased. this is expected since the silica fume act as a microfiller and its extreme fineness allows it to fill the microscopic voids between cement particles, which greatly reduces the permeability. the results obtained for condition 4 (c4), which is control moist curing, is shown graphically in fig. 6. the conclusions drawn for condition 1 (c1) apply also to condition 4 (c4). surface permeability is reduced more with moist curing rather than air curing. almost the same effect was produced by the addition of silica fume. the results obtained for specimens under curing condition 2 (c2) and curing condition 3 (c3) are presented graphically in figs. 4 and 5, respectively. here it is also clear that the flow of water decreased as the percentage replacement of silica fume was increased. the flow values obtained for curing condition 3 (c3) were lower than those obtained for curing condition 2 (c2). this was due to the fact that pores in specimens cured under curing condition 3 (c3) became smaller due to the higher bond between the cement paste and the aggregate. 5. conclusions from the experimental work conducted in this study, it may be concluded that: 1. the compressive strength of concrete increases as the water/binder (w/b) ratio decreases. 2. the replacement of cement by silica fume increases the compressive strength significantly up to 30%. 3. examining the compressive strength results, it was clear that concrete prepared with silica fume was less sensitive to air curing compared to the concrete prepared without silica fume. 4. it is optimum to use silica fume in the range of 1015% as a replacement for cement. 5. concrete with silica fume has a lower surface absorptivity than concrete without silica fume. 6. concrete cured under the moist curing condition had lower values of surface absorptivity compared to other curing conditions. concrete cured in air had the highest surface absorptivity. 0 0.1 0.2 0.3 0.4 10 30 60 120 t ime (min) fl ow ( m l / m 2 . s) mix. 1 mix. 2 mix. 3 mix. 4 mix. 5 figure 3. flow versus time for different mixes (condition 1 c1) 0 0.1 0.2 0.3 0.4 10 30 60 120 time (min.) fl ow (m l / m 2 .s ) m ix. 1 m ix. 2 m ix. 3 m ix. 4 m ix. 5 figure 4. flow versus time for different mixes (condition 2 c2) 0 0.1 0.2 0.3 10 30 60 120 time (min.) fl ow ( m l / m 2 .s ) m ix. 1 m ix. 2 m ix. 3 m ix. 4 m ix. 5 figure 5. flow versus time for different mixes (condition 3 c3) 0 0.05 0.1 0.15 0.2 0.25 10 30 60 120 time (min.) fl ow (m l / m 2 .s ) m ix. 1 m ix. 2 m ix. 3 m ix. 4 m ix. 5 figure 6. flow versus time for different mixes (condition 4 c4) 22 the journal of engineering research vol. 4, no.1 (2007) 17-22 7. further investigation using x-ray or electron scanning microscopy is recommended to provide visual assessment of the effect of silica fume on the microstructure of concrete. acknowledgments the authors would like to express their thanks to sultan qaboos university for providing financial support to this project under grant no. ig/eng/civl/03/02. references bs 1881: part 116: 1983, "method for determination of compressive strength of concrete cubes," british standard institution. bs 1881: part 117: 1983, "method for determination of tensile splitting strength," british standard institution. bs 1881: part 118: 1983, "method for determination of flexural strength," british standard institution. bs 1881: part 208: 1996, "recommendations for the determination of the initial surface absorption of concrete," british standard institution. bentur, a., goldman, a. and cohen, m.d., 1998, "the contribution of the transition zone to the strength of high quality silica fume concrete," proc mater res soc. pp. 114: 97-102. cong, x., gong, s., darwin, d. and mccable, s., 1992, "role of silica fume in compressive strength of cement paste, mortar, and concrete," aci mater j, vol. 89(4), pp. 375-387. darwin, d., shen, z. and harch, s., 1989, "silica fume, bond strength, and the compressive strength of mortar," proc mater res soc. vol. 114, pp. 105-110. detwiler, r.j. and mehta, p.k., 1989, "chemical and physical effects of silica fume on the mechanical behaviour of concrete," aci mater j, vol. 86(6), pp. 609-614. fidjestol, p., 1993, " applied silica fume concrete," concrete international, vol. 15(11), pp. 33-36. foster, s.w., 1994, "high performance concrete stretching the paradigm," concrete international, vol. 16(10), pp. 33-34. goldman, a. and bentur, a., 1993, "influence of microfillers on enhancement of concrete strength," cem con res, vol. 23(4), pp. 962-972. hasni, l., gallias, j.l. and saloman, m., 1994, "influence of the curing method on the durability of concrete," third intr conf, nice, france, sp-145, vm malhotra, ed., aci, farmington hills, mi, pp. 131-155. khayat, k.h. and aitcin, p.c., 1992, "silica fume in concrete an overview," in: proceeding of the 4th international conference on the use of fly ash, silica fume, slag and natural pozzolans in concrete, istanbul, turkey, pp. 835-872. luther, m.d., 1989, "microsilica (silica fume) concrete durability in severe environment, structural materials," proceeding of the sessions at asce structures congress 1989, san francisco, ca, usa, pp. 95-105. meeks, k.w. and carino, n.j., 1999, "curing of highperformance concrete: report of the state-of-the art," nistir 6295, nat. inst. of stds. and tech., gaithersburg, md. mindess, s., 1994, "materials selection, proportioning and quality control,” in: shah sp, ahmad sh, editors. high performance concrete and applications. london: edward arnold, pp. 1-26. nevile, a.m. and brooks, j.j., 2001, "concrete technology," prentice hall, england, revised edition. os7:2001 : ordinary portland cement; specification. ministry of commerce & industry, directorate general for specifications & measurements. oman. os2:1982 : natural aggregate; specification. ministry of commerce & industry, directorate general for specifications & measurements, oman. toutanji, h. and el-korchi, t., 1995, "the influence of silica fume on the compressive strength of cement paste and mortar," cement and concrete research. vol. 25(7), pp. 1591-1602. full page photo the journal of engineering research (tjer),vol. 13, no. 2 (2016) 187-196 vibrations of circular plates with elastically restrained edge against translation and resting on elastic foundation l.b. rao*, a and c.k. raob a school of mechanical and building sciences, vit university, chennai-600127, tamil nadu, india. b department of mechanical engineering, school of engineering, hyderabd-500 088, a.p. india. received 17 june 2014; accepted 23 may 2016 abstract: the present paper deals with exact solutions for the free vibration characteristics of thin circular plates elastically restrained against translation and resting on winkler-type elastic foundation based on the classical plate theory. parametric investigations are carried out for estimating the influence of edge restraint against translation and stiffness of the elastic foundation on the natural frequencies of circular plates. the elastic edge restraint against translation and the presence of elastic foundation has been found to have a profound influence on vibration characteristics of the circular plate undergoing free transverse vibrations. computations are carried out for natural frequencies of vibrations for varying values of translational stiffness ratio and stiffness parameter of winkler-type foundation. results are presented for twelve modes of vibration both in tabular and graphical form for use in design. extensive data is tabulated so that pertinent conclusions can be arrived at on the influence of translational edge restraint and the foundation stiffness ratio of the winkler foundation on the natural frequencies of uniform isotropic circular plates. keywords: plate, frequency, elastic edge, translational stiffness, elastic foundation. اعدة اھتزازات األلواح الدائریة ذات الحافة المقیدة المرنة مقابل االنتقال والسكون على ق مرنة بوشیالبیال كامیسواره راو، ∗أ،وكافرابو بھاسكارا راول تتناول ھذه الورقة الحلول الدقیقة لخصائص االھتزاز الحر أللواح دائریة رقیقة مرنة ومقیدة مقابل االنتقال :الملخص والسكون على قاعدة مرنة من نوع وینكلر باالعتماد على نظریة اللوح التقلیدیة. وتم إجراء استقصاء بارامتري من أجل لقاعدة مرنة على الترددات الطبیعیة أللواح دائریة. وتبین أن حافة المرونة تقدیر تأثیر حافة مقیدة مقابل االنتقال والصالبة المقیدة مقابل االنتقال وبوجود القاعدة المرنة لھا تأثیر شدید على خصائص اھتزاز اللوح الدائري الخاضع الى اھتزازات تنوعة من نسبة الصالبة االنتقالیة ومتغیر عرضیة حرة. تم إجراء العملیات الحسابیة للترددات الطبیعیة لالھتزازات لقیم م الصالبة لقاعدة نوع وینكلر. وقد تم عرض النتائج ألثنى عشر طریقة لالھتزاز سویًا على شكل جداول ورسوم بیانیة وذلك الستخدامھا في التصمیم. وتم جدولة بیانات واسعة النطاق بحیث أن االستنتاجات ذات الصلة منھا یمكن التوصل الیھا معرفة تأثیر الحافة المقیدة االنتقالیة ونسبة الصالبة لقاعدة نوع وینكلر على الترددات الطبیعیة لأللواح الدائریة ذات ل الخواص الموحدة. صالبة انتقالیة، قاعدة مرنة، تردد، حافة مرنة، لوح: الكلمات المفتاحیة * corresponding author’s e-mail: bhaskarbabu_20@yahoo.com l.b. rao and c.k. rao 188 nomenclature h thickness of a plate a radius of a plate ν poisson’s ratio e young’s modulus ρ density of a material ),( θrw transverse deflection of the plate d flexural rigidity of a plate tk translational spring stiffness wk stiffness of winkler foundation t translational spring stiffness ratio ξ foundation stiffness ratio mnω is the natural frequency of vibrations mnλ eigenvalue without foundation * mnλ eigenvalue with winkler foundation nm, positive integers corresponding to the number of concentric circles and nodal diameters in each flexural mode vibrations of circular plates with elastically restrained edge against translation and resting on elastic foundation 189 1. introduction circular plates resting on elastic foundation have wide range of application in the static and dynamic design of linear/nonlinear vibration absorbers, dynamic exciters, telephone receiver diaphragms, computer discs, printed circuit boards etc. (leissa 1969). due to the essential use of vibration data in the computation of stresses in such structures, reliable prediction of vibration data is of great importance. in view of its importance in engineering design, the problem of vibration of circular plates on elastic foundation has attracted the focus and attention of many researchers. some of the recent studies have reestablished the efficiency of the classical approach in analyzing the vibrations of variety of structures. circular plate problems allow for significant simplification in view of their symmetry but still many difficulties arise when the boundary conditions of the plate become complex involving linear and rotational restraints. a recent survey of literature shows that very few studies exist on the study of circular plates resting on elastic foundation. wang and wang (2003), who observed the switching between axisymmetric and asymmetric vibration modes, recently investigated the effect of internal elastic translational supports. the vibration characteristics of plates resting on an elastic medium are different from those of the plates supported only on the boundary. leissa (1993) discussed the vibration of a plate supported laterally by an elastic foundation. leissa deduced that the effect of winkler foundation merely increases the square of the natural frequency of the plate by a constant. salari et al. (1987) speculated the same conclusion. ascione and grimaldi (1984) studied unilateral frictionless contact between a circular plate and a winkler foundation using a vibrational formulation. leissa (1969), who tabulated a frequency parameter for four vibration modes of simply supported circular plate with varying rotational stiffness, presented one of the earliest formulations of this problem. kang and kim (1996) presented an extensive review of the modal properties of the elastically restrained beams and plates. zheng and zhou (1988) studied the large deflection of a circular plate resting on winkler foundation. ghosh (1997) studied the free and forced vibration of circular plates on winkler foundation by exact analytical method. chang and wickert (2001), kim et al. (2000) and tseng and wickert (1994) studied the dynamic characteristics of bolted flange connections involving circular plates displaying beating type of repeat frequencies and typical mode shapes of vibration. bolted flange connections are practically the best examples for the elastically restrained boundary conditions of circular plates on partial or continuous winkler type elastic foundation. the most general soil model used in practical applications is the winkler (1867) model in which the elastic medium below a structure is represented by a system of identical but mutually independent elastic linear springs. recent investigations have reiterated the efficiency of the classical approach (soedel 1993) in analyzing the behavior of structures under vibrations. there are other papers (weisman 1970; dempsey et al. 1984; celep 1988) dealing with the study of plates on a winkler foundation. in general, papers dealing with vibrating plates, shells and beams are concerned with the determination of eigenvalues and mode shapes (leissa 1969). a good number of studies are made by investigators (wang and lin 1996; kim et al. 2001; yayli et al. 2014) using the method of fourier series for estimating the frequencies of beams with generally restrained end conditions including the effect of elastic soil foundation. the method includes use of stoke’s transformation in suitably modifying the complex boundary conditions. very much similar to the dynamic stiffness matrix approach, the elements of the matrix involving infinite fourier series are explicitly obtained in these studies. the determinant of this matrix for each case considered leads to the frequency equation and the same can be solved using well known numerical methods. the results obtained for various elastically restrained beam cases in these studies tallied well with those available in the literature establishing the efficiency of this method. in view of the necessity of using complex combinations of rotational and translational springs at the boundary of the circular plate to suitably simulate the practical non-classical boundary connections being adopted in a wide range of industrial applications (bhaskara and kameswara 2009; bhaskara and kameswara 2010; lokavarapu and chellapilla 2013), the use of exact method of solution becomes imperative and hence the same is adopted in this paper. even though the method adopted here is l.b. rao and c.k. rao 190 classical, the particular case of vibration of elastically restrained circular plate resting on elastic foundation considered here is not dealt with in the available literature. utilizing the classical plate theory, this paper deals with exact method of solution for the analysis of free transverse vibrations of thin circular plate that is elastically restrained against translation and resting on winkler-type elastic foundation. for estimating the influence of edge restraint against translation and stiffness of the elastic foundation on the natural frequencies of circular plates, parametric investigations are carried out varying the values of elastic edge restraint stiffness against translation and the stiffness of the elastic foundation. the results obtained on natural frequencies of vibration clearly show that the vibration characteristics of the circular plate undergoing free transverse vibrations are found to have been profoundly influenced by these variations. computations are carried out for natural frequencies of vibrations for varying values of translational stiffness ratio and stiffness parameter of winkler-type foundation. results presented for twelve modes of vibration both in tabular and graphical form are believed to be quite useful for designers in this area. 2. mathematical formulation of the system the considered elastic thin circular plate is supported on a winkler foundation as shown in fig. 1. in the classical plate theory (leissa 1969), the following fourth order differential equation describes free flexural vibrations of a thin circular uniform plate. 0/),,(),,(. 224 =∂∂+∇ ttrwhtrwd θρθ (1) figure 1. a thin circular plate with translational elastic edge restraint and supported on elastic foundation. where )1(12/ 23 ν−= ehd is the flexural rigidity of a plate and νρ ,,,, eha are the plate’s radiuses, thickness, density, young’s modulus and poisson ratio respectively. the homogeneous equation for kirchhoff’s plate on one parameter elastic foundation is given by the following equation. ++∇ ),,(),,(. 4 trwktrwd w θθ (2) 22 /),,( ttrwh ∂∂ θρ =0 displacement in equation (2) can be presented as a combination of spatial and time dependent components as follows; let tierwtrw ωθθ ),(),,( = (3) now substitute the eq. (3) in eq. (2) 0),().(),(. 24 =−+∇ θωρθ rwhkrwd w (4) the solution of the equation takes the following form where3, 2, 1, 0,= ;…3 2, 1, 0,= cos. ),( θ λλ θ n a r ic a r jarw mnnmn mn nmnmn             +      = (5) where amn and cmn are constants, nj is bessel function of the first kind of first order and ni is modified bessel function of the first kind of first order. considering an elastically supported plate as shown in fig. 1, boundary conditions can be formulated at ar = , in terms of translational stiffness ( tk ) as follows: 0),( =θam r (6) ),(.),( θθ awkav tr −= (7) where the kelvin-kirchhoff and bending moment are defined as follows                         ∂ ∂ + ∂ ∂ + ∂ ∂ −= 2 2 2 2 2 ),(1 ),(1 ),( .),( θ θ θ ν θ θ aw r r aw r r aw dam r (8)                       ∂ ∂ − ∂∂ ∂ ∂ ∂ −+∇ ∂ ∂ −= θ θ θ θ θ νθ θ ),(1),(1 1 )1(),( .),( 2 2 2 aw rr aw r r aw r davr (9) vibrations of circular plates with elastically restrained edge against translation and resting on elastic foundation 191 by applying eqs. (6) and (8), we obtain the following equation 0 ),(1 ),(1 ),( 2 2 2 2 2 =                         ∂ ∂ + ∂ ∂ + ∂ ∂ θ θ θ ν θ aw r r aw r r aw (10) from eqs. (5) and (10), we derive the following equation )( 4 2 2 )( 4 2 2 2 2 2 2 mnn mn mn mn mn mnn mn mn mn mn mn i n st j n pq c λ λ ν λ ν λ λ ν λ ν         −++         +++− = (11) where mnp = )()( 11 mnnmnn jj λλ −+ − ; )()( 22 mnnmnnmn jjq λλ −+ += ; )()( 11 mnnmnnmn iis λλ −+ += ; )()( 22 mnnmnnmn iit λλ −+ += ; from eqs. (7) and (9), we get the following ),(. ),(1 ),(1 1 )1(),(. 2 2 2 θ θ θ θ θ θ νθ awk aw r r aw r r aw r d t−=                         ∂ ∂ − ∂∂ ∂ ∂ ∂ −+∇ ∂ ∂ − (12) from eqs. (5) and (12), we derived the following equation )( 84)3(8)2(44 3 2 )( 84)3(8)2(44 3 2 33 2 2 2 33 2 2 2 mnn mnmnmn mn mnmn mn mn mnn mnmn mn mnmn mn mn mn it n s nt u jt n p nq r c λ λλλ ν λ ν λ λ λλλ ν λ ν λ         −+ − +         −− −++         −− − −         −+ +−− = (13) where, d ka t t 3 = )()( 11 mnnmnnmn jjp λλ −+ −= ; )()( 22 mnnmnnmn jjq λλ −+ += ; )()( 33 mnnmnnmn jjr λλ −+ −= ; )()( 11 mnnmnnmn iis λλ −+ += ; )()( 22 mnnmnnmn iit λλ −+ += ; )()( 33 mnnmnnmn iiu λλ −+ += ; if ∞→tk then this case becomes simply supported boundary condition as shown in fig. 2. the frequency equation can be calculated from eqs. (11) and (13), which allows determining eigenvalues mnλ . the mode shape parameters mnc can be determined corresponding to these eigenvalues by using either eq. (11) or eq. (13). the amplitude of each vibration mode in eq. (5) is set by the normalization constant mna determined from the following condition. nqmpmnpq a mn mrdrdrwrw δδθθθ π =∫ ∫ ),().,( 2 0 0 (14) where, mnm is a mass of the plate, 1== nqmp δδ if qnpm == , and nqmpδδ = 0 if m ≠ p or n ≠ .q the normalization constant mna can be derived using eqs. (5) and (14) as given below: 1 2 0 0 2 2 cos. . . 1 −                 ∫ ∫                                   +       = π θθ λ λ π a mn nmn mn n mn rdrdn a r ic a r j a a (15) in eq (4), mnω is the natural frequency of vibrations               = h d a mn ρ λ 2 2 (16) it is clear from the eq. (16) the natural frequency of vibrations is dependent on the plate radius and eigenvalues. from eq. (16) we can express d ha mn mn 24 4 ωρλ = (17) l.b. rao and c.k. rao 192 244* ξλλ += mnmn (18) where d ak w 4 2 =ξ (19) [ ]4 1 24* ξλλ += mnmn (20) where mnλ is eigenvalue without foundation and *mnλ is eigenvalue with winkler’s foundation. 3. solution using matlab programming, computer software with symbolic capabilities, solves the above set of equations. the program determines eigenvalues ( *mnλ ), for a given range of boundary conditions. the boundary linear translational non-dimensional restraint parameter can be defined as follows: d ak t t 3 = (21) d ak w 4 2 =ξ (22) the following are the input parameters to the program; (i) translational stiffness ratio ( t ) (ii) foundation ratio ( ξ ) (iii) poisson ratio (ν ) (iv) upper bound for eigenvalues (n) (v) suggested for eigenvalues (d) (vi) number of mode shape parameters (n). the program finds eigenvalues * mnλ by using matlab root finding function. 4. results and discussion the code developed is used to determine eigenvalues of any set or range of translational and foundation constraints. this code also implanted for various plate materials by adjusting poisson ratio. such a wide range of results is not available in the literature yet. the eigenvalues for the plate edge, which is elastically restrained against translation and fully resting on the elastic foundation, at various values of the translational stiffness ratios, are computed and the results are given in table 1. the effects of the translational stiffness ratios are plotted in fig. 3. as seen from fig. 3, eigenvalues increases with an increment in the translational stiffness ratio, and the plates become unstable in the region when the translational stiffness ratio exceeds a certain value. twelve vibration modes are presented in fig. 3. the smoothened stepped variation is observed in fig. 3. the stepped region increases with increase in translational stiffness ratio and vibration modes. the location of the stepped region with respect to t changed gradually from the range of 0.01526 † [9.9997] ‡ – 5587.5316 [10.] to 16.62296 [14.6739] – 611824.96917 [16.75055]. the simply supported boundary conditions (fig. 2) could be accounted for by setting ( ∞→tk ) shown in fig.1. the translational edge supports becomes simply supported (or hinged) for very high values (close to infinity) of translational stiffness parameter i.e. ∞→tk . the frequency in this case is 2.23175 and this is in good agreement with the results published by wang (2005). another result considered for comparison is from rao and rao (2009) on study of the case of vibrations of elastically restrained circular plates supported on partial winkler foundation. when the support position is full span which means that when b = 1, the case becomes a circular plate having full foundation support with elastically restrained edge against translation. for this case, the frequency is 2.1834 which is in good agreement with the frequency of 2.18341 obtained from the present study. here †.represents translational stiffness ratio and ‡ represents eigen values throughout the text. the eigenvalues at various values of the foundation stiffness ratios [ ]33.0&100 == νt are computed and the results are given in table 2. the effects of the foundation stiffness ratio on eigenvalues are plotted in fig. 4. as seen from fig. 4, the eigenvalue increases with increase in the foundation stiffness ratio, and the plate becomes stiffer and stronger as the value of foundation stiffness becomes greater than 102. as seen from the tables 1 and 2, the influence of foundation stiffness ratio on eigenvalue is relatively greater than the translation stiffness ratio in increasing the overall natural frequencies of the plate support system. as seen from fig. 4, for all the modes considered here, up to a value of 10 the eigenvalues stay constant and beyond this value all the curves tend to converge to a constant eigenvalue as the foundation stiffness ratio increases up to 103. the convergence starts from 1.07897 [2.0325779] and continues up to a constant value of 9.63274 [13.84796]. vibrations of circular plates with elastically restrained edge against translation and resting on elastic foundation 193 0 2 4 6 8 10 12 14 16 18 -5 0 5 10 15 e ig en va lu e, λ m n transverse stiffness ratio, log10(t) λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 table 1. eigenvalues for different translational stiffness ratio for ξ =100 and ν =0.33. table 2. eigenvalues for different foundation stiffness ratio for t =100 and ν =0.33. figure 2. a simply supported thin circular plate resting on elastic foundation. figure 3. effect of translational stiffness ratio ξ on eigenvalues, λmn. log10(t) λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 -3 10 10.0205 10.3517 10.10358 10.79568 12.46774 10.29701 11.43857 13.52984 10.63672 12.2505 14.67332 -2 10 10.02051 10.35171 10.10359 10.79568 12.46775 10.29702 11.43857 13.52985 10.63673 12.25051 14.67332 -1 10.00005 10.02061 10.3518 10.10368 10.79576 12.46779 10.29711 11.43864 13.52989 10.63682 12.25056 14.67335 0 10.00047 10.02162 10.35263 10.10462 10.79649 12.46827 10.29803 11.43927 13.53026 10.63769 12.25109 14.67365 1 10.003 10.03287 10.36125 10.11468 10.804 12.47302 10.30745 11.44568 13.53403 10.64649 12.25645 14.67667 2 10.00568 10.1199 10.46887 10.2319 10.89095 12.52351 10.42003 11.5759 13.57333 10.74585 12.31326 14.70772 3 10.00614 10.203 11.00421 10.49407 11.61576 13.1793 10.91134 12.27823 14.08459 11.41839 12.99719 15.0966 4 10.00619 10.21322 11.14468 10.53794 11.94459 14.15323 11.03916 12.88272 15.33948 11.70347 13.90786 16.53545 5 10.00619 10.21423 11.15725 10.54212 11.97273 14.2438 11.05093 12.93527 15.47979 11.72934 13.99516 16.74092 6 10.00619 10.21433 11.15849 10.54254 11.97546 14.25215 11.05209 12.94027 15.49231 11.73186 14.00329 16.75866 12 10.00619 10.21434 11.15862 10.54258 11.97576 14.25307 11.05222 12.94082 15.49369 11.73214 14.00418 16.7606 log10(ξ) λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 -3 2.18341 4.70075 6.69703 5.56683 7.98678 10.99197 6.50356 9.33292 12.43922 7.59879 10.67534 13.84974 -2 2.18341 4.70075 6.69703 5.56683 7.98678 10.99197 6.50356 9.33292 12.43922 7.59879 10.67534 13.84973 -1 2.18365 4.70077 6.69704 5.56684 7.98679 10.99197 6.50357 9.33292 12.43922 7.59879 10.67535 13.84974 0 2.20704 4.70315 6.69786 5.56827 7.98727 10.99216 6.50447 9.33323 12.43935 7.59936 10.67555 13.84983 1 3.3284 4.92488 6.77875 5.7064 8.03541 11.01074 6.5926 9.36352 12.45218 7.65514 10.69583 13.84914 2 10.00568 10.1199 10.46887 10.2319 10.89095 12.52351 10.42003 11.5159 13.57333 10.74585 12.31326 14.70772 3 31.62296 31.62664 31.63867 31.63037 31.6549 31.73756 31.63691 31.68259 31.81038 31.6491 31.72496 31.90972 l.b. rao and c.k. rao 194 figure 4. effect of foundation stiffness ratio, ξ on eigenvalues, λmn. figure 5. effect of translational, t and foundation, ξ stiffness ratio son eigenvalues, λmn. the eigenvalues at various values of the translational stiffness ratios and foundation stiffness ratios are computed and the results are given in table 3. the effects of the translation and foundation stiffness ratios on eigenvalues are clearly observed in fig. 5, eigenvalues increases with an increment in both the translational and foundation stiffness ratios. as observed from the table 1 and 3, the influence of foundation stiffness ratio on eigenvalue is more predominant than that of translation stiffness ratio alone. as observed from table 1, 2 and 3, in table 3, lower eigenvalues are recorded for lower values of foundation and translation stiffness ratios together. as seen from fig. 5, all the curves are stable up to certain region and beyond this all the curves tend to converge as the value of translation and foundation stiffness ratios increases. the eigenvalues for different plate materials and various values of translational, foundation stiffness ratios are computed, and the results are given in table 4. it was observed that for high ξ , eigenvalues are independent of poisson ratio, as shown in fig. 6. in addition, it was observed that for any value of t, eigenvalues are independent on poisson ratio. 5. conclusion this paper deals with a method of computation of eigenvalues of axi-symmetric flexural vibrations of a circular plate with translational edge supports and resting on winkler foundation using a specifically written matlab code. in this paper, the computed numerical results are presented in a tabular format to enable an estimating the accuracy of 0 5 10 15 20 25 30 35 -4 -3 -2 -1 0 1 2 3 4 ei ge n va lu e( λm n) foundation parameter, log10( ξ) λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 0 5 10 15 20 25 30 35 -4 -3 -2 -1 0 1 2 3 4 e ig en va lu e( λm n) transverse stiffness, log10 (t) and foundation ratos, log10(ξ) λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 vibrations of circular plates with elastically restrained edge against translation and resting on elastic foundation 195 table 4. eigenvalues for different poisson ratios. ν 1000== ξt 10,100 == ξt 1000,10 == ξt 10,1 == ξt 50,50 == ξt 0 31.62925 4.92456 10.03027 3.62553 7.28924 0.1 31.62929 4.92466 10.03111 3.64904 7.2896 0.2 31.62934 4.92476 10.0319 3.6708 7.28993 0.3 31.62938 4.92485 10.03265 3.69097 7.29025 0.4 31.62942 4.92494 10.03336 3.70974 7.29056 0.5 31.62946 4.92503 10.03404 3.72725 7.29085 figure 6. effect of poisson ratio, ν on eigenvalues, λmn. approximate methods being used by other researchers for solving such problems. twodimensional plots of eigenvalues are drawn for a wide range of translational and foundation stiffness ratios facilitating their use in design. it has been observed that the eigenvalues remain constant without change only in a limited range of constraints (0 to 10) specific to each vibration mode and then steeply increase with increasing values of foundation ultimately converging to a constant value. it is also observed that the influence of foundation stiffness ratio on eigenvalues is more predominant than that of translational stiffness ratio. the effects of various parameters such as translational stiffness, foundation stiffness and poisson ratio parameters on natural frequencies of the plate with elastic edge and resting on elastic foundation are studied in detail. references ascione l, grimaldi a (1984), unilateral contact between a plate and an elastic foundation. mechanica 19(3): 223-233. bhaskara rl, kameswara rc (2009), buckling of circular plates with a ring support and elastically restrained edge. advances in vibration engineering 8(1): 61-69. bhaskara rl, kameswara rc (2010), buckling analysis of circular plates with elastically restrained edges and resting on internal elastic ring support. mechanics based design of structures and machines 38(4): 440-452. 0 5 10 15 20 25 30 35 0 0.1 0.2 0.3 0.4 0.5 e ig en va lu e( λ1 0) poisson ratio(ν) t=1000, ξ=1000 t=100, ξ=10 t=10, ξ=1000 t=1, ξ=10 t=50, ξ=50 l.b. rao and c.k. rao 196 celep z (1988), circular plate on tensionless winkler foundation. journal of engineering mechanics 114(10): 1723-1739. chang jy, wickert ja (2001), response of modulated doublet modes to traveling wave excitation. journal of sound and vibration 242(1): 69-83. dempsey jp, keer lm, patel nb, glasser ml (1984), contact between plates and unilateral supports. journal of applied mechanics 51(2): 324-328 ghosh ak (1997), axisymmetric dynamic response of a circular plate on an elastic foundation. journal of sound and vibration 205(1): 112-120. kang kh, kim kj (1996), modal properties of beams and plates on resilient supports with rotational and translational complex stiffness. journal of sound and vibration 190(2): 207-220. kim m, moon j, wickert ja (2000), spatial modulation of repeated vibration modes in rotationally periodic structures. journal of vibration and acoustics, transactions of asme 122(1): 62-68. kim km, kim ms (2001), vibration of beams with generally restrained boundary conditions using fourier series. journal of sound and vibration 245(5): 771-784. leissa aw (1969), vibration of plates (nasa sp-160). office of technology utilization, washington, dc. leissa aw (1993), vibration of plates. acoustical society of america, sewickley, pa. lokavarapu brl chellapilla kr (2013), fundamental buckling of circular plates with elastically restrained edges and resting on concentric rigid ring support. frontiers of mechanical engineering 8(3): 291-297. rao ck, rao lb (2009), vibrations of elastically restrained circular plates resting on partial winkler foundation. the open acoustics journal 2: 68-74. salari m, bert cm, striz ag (1987), free vibration of a solid circular plate free at its edge and attached to a winkler foundation. journal of sound and vibration 118(1): 188-191. soedel w (1993), vibrations of shells and plates. mareel dekker, inc. tseng jg, wickert ja (1994), on the vibration of bolted plate and flange assemblies. journal of vibration and acoustics, transactions of asme 116(4): 468-473. wang jts, lin cc (1996), dynamic analysis of generally supported beams using fourier series. journal of sound and vibration 196(3): 285-293. wang cy, wang cm (2003), fundamental frequencies of circular plates with internal elastic ring support. journal of sound and vibration 263: 1071-1078. wang cy (2005), fundamental frequency of a circular plate supported by a partial elastic foundation. journal of sound and vibration 285(4-5): 1203-1209. weisman y (1970), on foundations that react in compression only. journal of applied mechanics 37: 1019-1030. winkler e (1867), die lehre von der elasticitaet and festigkeit, prag. dominicus. yayli mö, aras m, aksoy s (2014), an efficient analytical method for vibration analysis of a beam on elastic foundation with elastically restrained ends. shock and vibration 2014: 1-7. zheng xj, zhou yh (1988), exact solution of nonlinear circular plate on elasticfoundation. journal of engineering mechanicsasce 114: 1303-1316. microsoft word paper 4.docx the journal of engineering research (tjer) vol. 14, no. 1 (2017) 39-48 quantitative analysis of force and torque in bone drilling   k. alam*, a, r. muhammadb, a. shamsuzzohaa, a. alyahmadia and n. ahmedb a department of mechanical and industrial engineering, college of engineering, sultan qaboos university, oman. b department of mechanical engineering, cecos university of it and emerging sciences, peshawar, kpk, pakistan. received 3 august 2015; accepted 5 september 2016 abstract: bone drilling is an important and the most frequent operation in orthopaedics and other bone surgical procedures. prediction and control of drilling force and torque are critical to safe and efficient surgeries. this paper studies the drilling force and torque arising from bone drilling process. drilling parameters such as drilling speed, feed rate, drill size and drill condition (sharp and worn) were changed to measure the force and torque in the direction of the drill penetration. experimental results demonstrated lower drilling force using a sharp drill compared to a worn drill for similar drilling conditions. contrary to the drilling force, lower torque was measured using a worn drill compared to a sharp drill. the drilling force was found to decrease with increase in drill speed and increased with rise in the feed rate using both types of drills. a linear drop in drilling torque was measured with increase in drilling speed. this study provided scientific information to orthopaedic surgeons and technicians to use appropriate surgical drill and cutting parameters to avoid overstressing of the bone tissue and drill breakage during drilling operations. keywords: orthopaedic, bone drilling, drilling force, drilling torque, drill wear. אאאאא k،*k،k،אk،k،    א א      אא   א  א wא    kא  א  א אא א א  א  א  אא אאאאאאkאאאאkא  אא  ،א  ،א   א f      א  e   א א אאא אא     א  א  kא   אאkkאאא، kא    א א  kא    א א    א א  אאאאאאאאאאא kאאאא   אא،אאw،א،א،אאאאk  * corresponding author’s e-mail: kalam@squ.edu.om k. alam, r. muhammad, a. shamsuzzoha, a. alyahmadi and n. ahmed   40 1. introduction cutting bone for fixation is one of the oldest surgical procedures in the history of medicine. it is believed that an operation to remove a portion of the skull bone was performed as early as 10,000 b.c. (krause 2003). in modern times, bone fracture is a feature of everyday life due to accidents or aging, and fusing bone fragments together or replacing human joints usually requires cutting bone. knee and hip implant surgeries are frequently performed around the world, with a total of 300,000 knee arthroplasties performed each year in the united states alone (harrysson et al. 2007). popular methods of bone cutting include scraping, grooving, sawing, and drilling. among these methods, drilling is a surgical technique which is frequently and widely discussed in the literature (alam 2014; alam et al. 2011a; alam et al. 2014b; augustin et al. 2012; pandey and panda 2013). measurements of forces and temperatures in bone drilling operation have been discussed widely in the literature (abouzgia and symington 1996; alam et al. 2011b; alam et al. 2015; alam and silberschmidt 2014; bachus et al. 2000; brisman 1996). like drilling of other materials (muhammad et al. 2010; muhammad et al. 2011; muhammad et al. 2012), bone drilling requires force to push the tool into the bone. lower cutting forces in bone surgical procedures are required to avoid unnecessary damage to the bone tissue. elevated temperature is harmful for bone tissue since it causes death in living cells. several empirical studies have found that the bone temperatures in drilling operation are strongly correlated with drilling force (abouzgia and symington 1996; bachus et al. 2000; brisman 1996; matthews and hirsch 1972). the application of heavy force on the drill can reduce both the maximum cortical temperature and its duration above 50oc (bachus et al. 2000). an increase in bone temperature was observed with an increase in drilling force from 1.5–4 n (brisman 1996). some empirical studies have shown that cortical bone temperatures are inversely related to drilling force. matthews and hirsch (1972) found an inverse relationship between drilling force and bone temperature using force levels ranging from 20–118 n while drilling human cortical bone. in that study, the decrease in bone temperature was attributed to the reduced time necessary to penetrate the drill into the cortex of the bone. abouzgia and symington (1996) found that both maximum temperatures and the duration of application decreased when drilling force was increased from 1.5 n to 9.0 n. allan et al. (2005) presented a relationship between the level of drill bit wear and the temperature of the bone being drilled. bone tissue is sensitive to many osteotomes. bone cracking under sharp tool penetration is a common phenomenon and has been extensively studied previously (sugita et al. 2009; alam et al. 2012; alam et al. 2014a). in a bone drilling operation, an orthopedic surgeon pushes the drill into the bone with a significant amount of force. harvesting long bone grafts may also cause pain due to stress fractures in the bone around the donor site (tanishima et al. 1995). heavy drilling force may induce microcracks and impart damage to the surrounding tissue (brett et al. 2004; kendoff et al. 2007). in addition, heavy drilling force and torque are the prime reasons for drill breakage during surgical procedures (jantunen 2002). lower cutting force was also measured in plane cutting of cortical bone in the presence of imposed vibrations on the cutting tool (alam et al. 2013). perhaps in response to these findings, some recent studies have proposed ultrasonically-assisted drilling (uad) of bone to minimize drilling force, torque, temperature, and micro cracks (alam et al. 2011b; alam et al. 2014a; alam and silberschmidt 2014; wang et al. 2013). some studies have also attributed lower force to appropriate drill geometry and drilling parameters (wiggins and malkin 1976; saha et al. 1982). heavy drilling force could damage bone tissue and adversely affect the success of a surgical procedure. recent technological advancements have been concerned with efforts to decrease the force required by a surgeon when cutting bone, thereby expediting surgical procedures without compromising safety. conventional surgical tools are continually replaced by robot-assisted tools for achieving safety and performance goals, but bone drilling is mostly performed using handheld surgical drills with no detection system to measure force and torque. despite technological enhancement in bone surgical procedures, metallic drills are still the primary choice of orthopedic surgeons. the cutting performance of surgical drills dramatically drops when sharp cutting edges of the drill wear out. as a result, surgical clinics spend quantitative analysis of force and torque in bone drilling   41   heavily to replace surgical tools when they wear out. no study performed has provided a comparison between the level of force and torque produced by a sharp and worn drill. this study is a step forward; therefore, to measure and compare the drilling thrust force and torque generated using a sharp versus worn drill. the effect of drilling parameters such as drilling speed and feed rate on the drilling force for both sharp and worn drills is studied using a series of experiments discussed below. 2. materials and methods 2.1. specimen preparation drilling tests were performed on compact bone excised from a bovine femoral shaft. fresh femurs of a young cow with no visible disease were obtained from a local slaughter house. the middle portions of the femurs were cut using a mechanical hacksaw. water irrigation was used while cutting the specimen from the femoral shaft in order to avoid damage to the bone tissue. specimens of cylindrical shape (fig. 1a) were obtained for the drilling operation. specimens were kept refrigerated at -10°c for two days before using them. the thin, elastic film (periosteum) was removed from the top surface of the bone to prevent clogging of the drill flutes. each cylindrical specimen was cut into two halves for attaching them to a metallic plate with glue. this was done to secure the bone specimen firmly to a fixture during drilling tests. test specimen attached to metallic plate is shown in fig. 1b. 2.2. experimental setup and procedure the experimental setup consisted of a vertical cnc drilling machine, a stationary dynamometer (type 9271a) and surgical drill bits (kistler group ac, winterthur, switzerland). the dynamometer measures both the thrust force and torque as a function of time figure 1. bone specimen used in drilling experiments, (a) specimen excised from femur, and (b) specimen attached to metallic plate. and was placed directly below the sample for drilling. piezoelectric materials were used in the sensors, generating a charge proportional to the applied load. the torque was not measured directly but instead was calculated from the various measurement signals of the force sensors. charge amplifiers were used to charge the difference arising when sensors were loaded. the charge signals were then converted into equivalent voltage signals. two sizes of orthopedic drills (2.5 mm and 4.8 mm in diameter) (zimmer biomet, warsaw, indiana, usa) donated by a local hospital were used in the experiments. the sizes of drills used in the current study were within the range reported in the literature related to bone drilling and are frequently used in surgical clinics (alam 2014; alam et al. 2014b; augustin et al. 2012; hillery and shuaib 1999). drill bits’ geometrical angles included a 90-degree point angle, 20-degree helix angle, and a 116-degree chisel edge angle. both sharp drills and worn drills of the prescribed sizes were used. the bone surgical drill bits used in the experiments are shown in fig. 2. according to hospital sources, drills are considered worn after drilling approximately 500–600 holes. the rigid structure and high power of the drilling machine enable precise identification of drilling speed and feed rate. the picoscope series 2000 oscilloscope with a maximum frequency of 10 mhz was used to acquire the data for force and torque in a digital format (pico technologies, cambridgeshire, uk). drilling tests were conducted at room temperature (25°c). drill penetration was performed perpendicular to the major figure 2. orthopaedic surgical drills used in drilling experiments. k. alam, r. muhammad, a. shamsuzzoha, a. alyahmadi and n. ahmed   44   anatomical direction of the bone, along the longitudinal axis of the femur bone. room temperature tap water was continually supplied to the drilling region to mimic in vitro drilling conditions. each drill was replaced after boring 40 holes in order to get reliable force data for comparison. a maximum drilling speed and feed rate of 3,000 rpm and 50 mm/min, respectively, were used in experiments. ten samples were prepared for drilling operation. each sample accommodated twenty holes on average. each experiment was repeated three times for a particular set of drilling parameters to observe repeatability in the measurements. the experimental setup for bone drilling is shown in fig. 3. 3. results and discussion force evolution during drill penetration was studied. the force was observed to rise rapidly when the drill bit started penetrating the bone. the force attained maximum value when the cutting lips of the drill were fully engaged with the bone and remained constant until the drill crossed the cortex of the bone. the force evolution with respect to the drilling time using a worn and sharp drill is shown in fig. 4. similar force profiles were observed in all drilling tests. all the data points in the subsequent plots represent the maximum value of the force obtained during drill penetration. the drilling thrust force (the force component in the direction of drilling) and torque (the moment about the axis of rotation of the drill) were measured. each data point in the subsequent plots showed an average of the three consecutive drilling tests. the standard error was plotted for each set of drilling tests to show variation from the mean value. the effect of drilling speed, feed rate on thrust force, and drilling speed on torque is discussed below. figure 3. experimental set up for bone drilling. 3.1. effect of drilling speed on force the effect of drilling speed on force was studied. drilling was performed for a set of parameters using a sharp drill followed by a worn drill. the drilling speed was changed from 600 rpm to 3,000 rpm, and the feed rate was kept constant at 30 mm/min. figures 5 and 6 show the plots of drilling thrust force for different drilling speeds using both types and sizes of drills. it was observed that the thrust force decreased with the progressive increase in drill speed. the thrust force was observed to decrease by 32% (from an average maximum force of 67 n to 45 n) when the drilling speed was increased from 600 rpm to 3,000 rpm using a 4.8-mm worn drill (fig. 5). similarly, a decrease of 33% (from an average maximum force of 51 n to 34 n) was noted for a similar change in drilling speed when a sharp drill was used. the force dropped from 67 n to 51 n when the drill was changed from worn to sharp using 600 rpm and a 4.8-mm drill. at the higher drilling speed used in this study (3,000 rpm), for the same size of drill the force was dropped from 45 n to 34 n. a comparison of a sharp and worn drill showed a similar trend using a 2.5-mm diameter drill. a decrease of 23% was measured when a 2.5-mm worn drill was replaced with a sharp drill of the same size using 600 rpm (fig. 6). approximately the same drop was measured when 3,000 rpm were used. although not shown here, the drop in the force with an increase in the drill speed was observed for the range of feed rates used in this study. figure 4. evolution of thrust force in bone drilling using wd and sd (drilling speed – 1800 rpm, feed rate – 30 mm/min, drill size – 4.8 mm). 42 time (s)  quantitative analysis of force and torque in bone drilling   41   figure 5. effect of drill speed on thrust force (drill diameter – 4.8 mm, feed rate – 30 mm/min). figure 6. effect of drill speed on thrust force (drill diameter – 2.5 mm, feed rate – 30 mm/min) 3.2. effect of feed rate on force the effect of the feed rate on the drilling force using both types and size of drills was also studied. the drilling speed was kept constant at 1,800 rpm to find the effect of feed rate alone. the trend with which the force rose with an increase in feed rate was similar for both types of drills. the variation of force with feed rate for both types and sizes of drills is shown in figs. 7 and 8. a linear increase in force was found with an increase in feed rate. the force was observed to increase by 39% (from an average maximum force of 41–57 n) when the feed rate was increased from 10 mm/min to 50 mm/min using a worn drill 4.8 mm in diameter (fig. 7). similarly, an increase of 62% (from an average maximum force of 29– 47 n) was noted for a similar change in the feed rate when a sharp drill was used. the force dropped from 41 n to 29 n when the drill was changed from worn to sharp using a feed rate of 10 mm/min and a 4.8-mm drill (fig. 7). at a higher feed rate of 50 mm/min for the same figure 7. effect of feed rate on thrust force (drill diameter – 4.8 mm, drill speed – 1800 rpm). figure 8. effect of feed rate on thrust force (drill diameter – 2.5 mm, drill speed – 1800 rpm). size drill, the force dropped by 10 n. a comparison of sharp and worn drills showed a similar trend when using a 2.5-mm-diameter drill (fig. 8). a higher feed rate increased the chip cross-sectional area that in turn raised the cutting force. thicker chips increased the resistance, which caused an increase in the cutting energy. 3.3. effect of drill speed on torque the effect of drill quality and size on drilling torque was also studied using different spindle speeds. figures 9 and 10 show the variation in torque with drill speed when a sharp and worn drill penetrated into bone. similar to drilling force, torque dropped when the drill speed was increased. the drop in torque was similar for both types of drills. however, small variation between the torque measurements was observed between the sharp and worn drills for the same drilling speed. the slightly lower torque values obtained when using a worn drill may be due to the slip of the blunt edges of the drill on the bone in a tangential direction. such 43 k. alam, r. muhammad, a. shamsuzzoha, a. alyahmadi and n. ahmed   44   figure 9. variation of torque with drill speed (drill diameter – 4.8 mm, feed rate – 30 mm/min). figure 10. variation of torque with drill speed (drill diameter – 2.5 mm, feed rate – 30 mm/min). conditions did not generate enough tangential force compared with those produced in a sharp drill. 3.4. statistical analysis a simple statistical analysis was performed to determine the effect of drill speed and feed rate on the level of drilling force and torque for both types and sizes of drills. the average drilling force measured with a worn drill was found to be larger than that measured with a sharp drill. the average drilling torque measured with sharp drill was found to be larger than that measured with a worn drill. figures 11 and 12 show a comparison of drilling force and torque for both types and sizes of drills for different drilling conditions. 4. discussion the wear in drilling metals may significantly increase with an increase in the drilling thrust force and torque (el-hofy 2013). as discussed earlier, in the present study, the thrust force gradually increased when the drill bit penetrated into the bone and dropped during the exit stage. the rise and drop in force was due to the engagement of the drill cutting edges with the bone during the penetration and exit stages. the decrease during the exit stage reduced the average thrust force over the drilling time. large drilling force and torque may cause drill breakage during surgical incision (jantunen 2002). this study demonstrated that using a worn drill would result in a greater chance of breakage due to the large drilling force inflicted as compared to a sharp drill. different readings of force and torque obtained using the same drill size and type may be attributed to the anisotropic nature of bone, its mechanical properties, and differing bone porosity at different locations. the velocity with which material is removed in the vicinity of a drill bit axis is small. in drilling, the chisel edge experiences an indentation process while the cutting lips perform orthogonal cutting (galloway 1957). modeling the forces experienced by the chisel edge is important since it significantly contributes to the overall thrust force experienced by the drill (stephenson and agapiou 1992). several factors such as stiffness of the material being cut, feed rate, rotational speed, cutting fluids, drill size, drill geometry, and the friction between the drill and the workpiece material contribute to varying thrust force. the findings of this study are comparable to those reported in other empirical works (alam et al. 2011b; hillery and shuaib 1999). the level of force was significantly reduced with an increase in the number of revolutions of the drill. a higher drilling speed produced a lower force due to the quick chip formation mechanism and the decrease in mean friction between the drills and the bone (alam et al. 2011b). the decrease in drilling force with an increase in drilling speed was also observed in other studies performed on bovine cortical bone (hillery and shuaib 1999). in a recent study performed on porcine bone (xu et al. 2014), the drilling force was found to decrease both with and without saline cooling. the obvious reason for an increase in drilling force with an increased feed rate was due to the increase in the material removal rate per revolution of the drill. higher feed rates caused the average friction between the drill and the bone to rise, which increased pressure on the drill. quantitative analysis of force and torque in bone drilling   45   figure 11. variation of drilling force with feed rate for a constant speed of 18000 rpm. (a) wd, and (b) sd for different diameters of drills.   figure 12. variation of drilling torque with drilling speed for a constant feed rate of 30 mm/min. (a) wd, and (b) sd for different diameters of drills. the results presented in the current study overlapped with another recent study (lughmani et al. 2015). the drilling force was also found to be strongly influenced by varying feed rates while simulating dry drilling using finite element analysis (lughmani et al. 2013; lughmani et al. 2015). studies have found that maximum bone temperatures were strongly correlated to the drilling force (abouzgia and symington 1996; bachus et al. 2000). the majority of previous studies found an increase in bone temperature that correlated with an increase in drilling speed (abouzgia and symington 1996; bachus et al. 2000), with some exceptions which reported an inverse relationship between the drilling force and bone temperature (abouzgia and symington 1996). the conflicting results reported in the literature arise from the variety of test specimens and drilling conditions. whether or not drilling force affects bone temperature, it is important to minimize pressure on the drill to avoid overstressing the bone tissue. surgeons and technicians in orthopedic surgical clinics have to discard drills when they become blunt. they discard the drills based on their experience and sense of penetration force. however, they do not have any experimental data on the level of force and torque caused by sharp and worn drills. experimental data from the current study will provide scientific information to the medical community based on quantitative analysis of drilling force and torque. the results of this study should also contribute to the development of new surgical drills which will have force and torque sensation mechanisms. such systems will signal the surgeon to discard the drill if the force and torque values reach a threshold level, therefore making bone drilling safer.   k. alam, r. muhammad, a. shamsuzzoha, a. alyahmadi and n. ahmed   46   5. limitations of this study this study did not measure the force and torque as a function of drilling depth. in the drilling process, the chips compact as the depth increases inside the flutes, thereby increasing the torque (mellinger et al. 2002). the increase in the drilling depth generates improper conditions for the cooling fluid to access the cutting zone. under these conditions, the drilling tools’ wear increases, imparting more pressure on the drill in the cutting direction. the maximum speed used in this study was 3,000 rpm. further studies are required to explore the effect of drilling speeds higher than those used in this study. further research is required to relate the amount of the drill wear with the drilling force and torque. geometries of drills other than those used in this study are required to be tested to obtain more generalized data on bone drilling. in this study, drilling tests are performed on hard cortical bone only. depending on the fracture site, the drill may be used to produce holes in spongy bone (trabecular bone). further research is suggested to investigate the effect of drill quality on thrust force in drilling spongy bone under various drilling conditions. further studies are required to check the quality of the hole surface using both types of drills and its effect on bone regeneration. new techniques are required to check real time wear of the tool based on cutting force and torque measurements. 6. conclusion in this article, the drilling thrust force and torque produced during the drilling process was measured and compared using sharp and worn surgical drills. the effect of drilling speed and feed rate on the thrust force was investigated in the presence of cooling with water. the main aim of this research study was to provide engineering-based information to surgeons and technicians about the force required by them to push drill into bone. experimental results demonstrated a lower drilling force using a sharp drill compared to a worn drill for similar drilling conditions. the thrust force may be reduced using a combination of low speed of penetration and higher drill rotation speed. the torque may be lowered by using a higher drill rotational speed. the thrust force must be kept at minimum to avoid mechanical and thermal damage to the bone tissue. it is hoped that the results of this research will be useful for surgeons to understand the effect of drill wear on the outcomes of the drilling process. the data obtained from this study may be useful for the improvement of automatic bone surgical drilling machines and the optimization of bone cutting procedures. references abouzgia mb, symington j (1996), effect of drill speed on bone temperature. international journal of oral and maxillofacial surgery 25(5): 394-399. alam k, ghafoor a, silberschmidt vv (2011a), analysis of forces and temperatures in conventional and ultrasonically-assisted cutting of bone. advanced materials research, trans tech publ. alam k, mitrofanov av, silberschmidt vv (2011b), experimental investigations of forces and torque in conventional and ultrasonically-assisted drilling of cortical bone. medical engineering and physics 33(2): 234-239. alam k, kerckhofs g, mitrofanov av, lomov s, wevers m, silberschmidt vv (2012), online analysis of cracking in cortical bone under wedge penetration. proceedings of the institution of mechanical engineers, part h: journal of engineering in medicine 0954411912451824. 226(9): 709-719. alam k, khan m, silberschmidt vv (2013), analysis of forces in conventional and ultrasonically assisted plane cutting of cortical bone. proceedings of the institution of mechanical engineers, part h: journal of engineering in medicine 227(6): 636-642. alam k (2014), experimental and numerical investigation of cracking behavior of cortical bone in cutting. technology and health care: 22(5): 741-750. alam k, silberschmidt vv (2014), analysis of temperature in conventional and ultrasonically-assisted drilling of cortical bone with infrared thermography. technology and health care: 22(2): 243-252. alam k, ahmed n, silberschmidt vv (2014a), comparative study of conventional and quantitative analysis of force and torque in bone drilling   47   ultrasonically-assisted bone drilling. technology and health care: 22(2): 253-262. alam k, khan m, silberschmidt vv (2014b), 3d finite-element modelling of drilling cortical bone: temperature analysis. journal of medical and biological engineering 34(6): 618-623. alam k, hassan e, bahadur i (2015), experimental measurements of temperatures in ultrasonically assisted drilling of cortical bone. biotechnology and biotechnological equipment 29(4): 753-757. allan w, williams e, kerawala c (2005), effects of repeated drill use on temperature of bone during preparation for osteosynthesis self-tapping screws. british journal of oral and maxillofacial surgery 43(4): 314-319. augustin g, zigman t, davila s, udilljak t, staroveski t, brezak d and babic s (2012), cortical bone drilling and thermal osteonecrosis. clinical biomechanics 27(4): 313-325. bachus kn, rondina mt, hutchinson dt (2000), the effects of drilling force on cortical temperatures and their duration: an in vitro study. medical engineering and physics 22(10): 685-691. brett p, baker d, taylor r and griffiths m (2004), controlling the penetration of flexible bone tissue using the stapedotomy microdrill. proceedings of the institution of mechanical engineers, part i: journal of systems and control engineering 218(5): 343351. brisman dl (1996), the effect of speed, pressure, and time on bone temperature during the drilling of implant sites. the international journal of oral and maxillofacial implants 11(1): 35-37. el-hofy hag (2013), fundamentals of machining processes: conventional and nonconventional processes, crc press. galloway d (1957), some experiments on the influence of various factors on drill performance. transaction of asme journal engineering for industry 79(2): 191-231. harrysson ol, hosni ya, nayfeh jf (2007), custom-designed orthopedic implants evaluated using finite element analysis of patient-specific computed tomography data: femoral-component case study. bmc musculoskeletal disorders 8(1): 91. hillery m and shuaib i (1999), temperature effects in the drilling of human and bovine bone. journal of materials processing technology 92: 302-308. jantunen e (2002), a summary of methods applied to tool condition monitoring in drilling. international journal of machine tools and manufacture 42(9): 997-1010. kendoff dm, citak m, gardner mj, stübig t, krettek c and hüfner t(2007), improved accuracy of navigated drilling using a drill alignment device. journal of orthopaedic research 25(7): 951-957. krause w (2003), orthopaedic cutting procedures. bonezone 2: 4-10. lughmani wa, bouazza-marouf k, ashcroft i (2013), finite element modeling and experimentation of bone drilling forces. journal of physics: conference series, iop publishing. lughmani wa, bouazza-marouf k, ashcroft i (2015), drilling in cortical bone: a finite element model and experimental investigations. journal of the mechanical behavior of biomedical materials 42: 32-42. matthews ls, hirsch c (1972), temperatures measured in human cortical bone when drilling. the journal of bone and joint surgery 54(2): 297-308. mellinger jc, ozdoganlar ob, devor re, kapoor sg (2002), modeling chipevacuation forces and prediction of chipclogging in drilling. journal of manufacturing science and engineering 124(3): 605-614. muhammad r, ahmed n, abid m and silberschmidt vv (2010), 3d modeling of drilling process of aisi 1010 steel. journal of machining and forming technologies issn 1947: 4369. muhammad r, ahmed n, shariff ym, silberschmidt vv (2011), effect of cutting conditions on temperature generated in drilling process: a fea approach. advanced materials research, trans tech publ. muhammad r, ahmed n, shariff ym, silberschmidt vv (2012), finite-element analysis of forces in drilling of ti-alloys at elevated temperature. solid state phenomena, trans tech publ. pandey rk, panda s (2013), drilling of bone: a comprehensive review. journal of clinical orthopaedics and trauma 4(1): 15-30. saha s, pal s, albright j (1982), surgical drilling: design and performance of an improved drill. journal of biomechanical engineering 104(3): 245-252. stephenson d, agapiou j (1992), calculation of main cutting edge forces and torque for k. alam, r. muhammad, a. shamsuzzoha, a. alyahmadi and n. ahmed   46   drills with arbitrary point geometries. international journal of machine tools and manufacture 32(4): 521-538. sugita n, osa t, aoki r, mitsuishi m (2009), a new cutting method for bone based on its crack propagation characteristics. cirp annals-manufacturing technology 58(1): 113118. tanishima t, yoshimasu n, ogai m (1995), a technique for prevention of donor site pain associated with harvesting iliac bone grafts. surgical neurology 44(2): 131-132. wang y, cao m, zhao y, zhou g, liu w, li d (2013), experimental investigations on microcracks in vibrational and conventional drilling of cortical bone. journal of nanomaterials 2013, articl3e idf 845205: 5. wiggins k, malkin s (1976), drilling of bone. journal of biomechanics 9(9): 553-559. xu l, wang c, jiang m, he h, song y, chen h, shen j, zhang j (2014), drilling force and temperature of bone under dry and physiological drilling conditions. chinese journal of mechanical engineering 27(6): 1240-1248. 48 microsoft word paper 6.docx the journal of engineering research (tjer) vol. 14, no. 1 (2017) 64-73 storage of nitrous oxide (nox) in diesel engine exhaust gas using alumina-based catalysts: preparation, characterization, and testing a. alsobaai* chemical engineering department, faculty of engineering and petroleum, hadhramout university, mukalla, yemen. received 28 january 2016; accepted 22 september 2016 abstract: this work investigated the nitrous oxide (nox) storage process using alumina-based catalysts (k2o/al2o3, cao/al2o3, and bao/al2o3). the feed was a synthetic exhaust gas containing 1,000 ppm of nitrogen monoxide (no), 1,000 ppm i-c4h10, and an 8% o2 and n2 balance. the catalyst was carried out at temperatures between 250–450°c and a contact time of 20 minutes. it was found that nox was effectively adsorbed in the presence of oxygen. the nox storage capacity of k2o/al2o3 was higher than that of bao/al2o3. the nox storage capacity for k2o/al2o3 decreased with increasing temperature and achieved a maximum at 250°c. potassium loading higher than 15% in the catalyst negatively affected the morphological properties. the combination of ba and k loading in the catalyst led to an improvement in the catalytic activity compared to its single metal catalysts. as a conclusion, mixed metal oxide was a potential catalyst for de-nox process in meeting the stringent diesel engine exhaust emissions regulations. the catalysts were characterized by a number of techniques and measurements, such as x-ray diffraction (xrd), electron affinity (ea), a scanning electron microscope (sem), brunner-emmett-teller (bet) to measure surface area, and pore volume and pore size distribution assessments. keywords: nox storage, lean de-nox, exhaust catalyst, diesel exhaust, mixed metal oxide.  nox ،אwאאאאא אא  kא*  wא    א noxא א  א fk2o ،al2o3 ،cao/al2o3 bao/al2o3eאאk١٠٠٠אno،١٠٠٠אic4h10 ،٨  ٪o2  אn2 א   א  א  k٢٥٠   ٤٥٠   ٢٠אknoxאאknoxאk2o/al2o3   א bao/al2o3>   א א  אא    ٢٥٠  k ١٥אאkאאא٪bak אא؛kאאאאאאא de-noxאאאkאאאxrd،ea،sem، אbetאא،k   wאאnox،de-noxkאא،א،א،  * corresponding author’s email: alsobaai@yahoo.com a. alsobaai 65 1. introduction diesel engines are caught in an area of conflict between a wide variety of requirements ranging from maximum customer benefit to minimum fuel consumption and emissions (michael et al. 2014). however, they have many adverse environmental effects, particularly due to the emission of excessive nitrogen oxide (nox) in its emissions. due to the lean burn combustion in the engine, the exhaust gas also contains an excess of oxygen that complicates conventional approaches to chemically reduce nox to environmentally benign nitrogen gas. therefore, there is an urgent need to develop diesel emission control technology to take full advantage of the fuel efficiency and durability of diesel vehicles (athanasios et al. 2015; kabin et al. 2004). nox storage and reduction (nsr) catalysts, which are also sometimes referred to as nox adsorption catalysts and lean nox traps, have been developed as a promising alternative method to remove nox from diesel engine emission (athanasios et al. 2015; centi et al. 2003; epling et al. 2004; michael et al. 2014). these catalysts operate in a cyclic manner where during the lean reductant (conventionally ammonia [nh3], carbon monoxide [co], or hydrocarbons) period of operation, the catalyst stores or traps nox as a nitrate species. a periodic and short rich pulse of reductant is then introduced so that the trapped nox is released and reduced to n2 and the catalyst is regenerated (epling et al. 2004). the nox trapping materials can be found among the alkali (potassium [k], magnesium [mg], calcium [ca]) and alkaline earth metal (barium [ba], lihium [li]) oxides, with barium oxide (bao) having been most extensively studied experimentally (bethke et al. 1995; fridell et al. 1999; josh et al. 2013; michael et al. 2014; milt et al. 2003a; milt et al. 2003b; sedlmair et al. 2003; su and amiridis 2004; takahashi et al 1996; westerberg and fridell 2001). in general, these metals show different efficiencies in storing nox as nitrates that eventually have different stabilities in the catalysts. therefore, efforts to determine the right types and content of oxides, or possibly a combination of oxides in the catalyst, should be explored. the main aim of this work was to investigate the performance of oxides of k and ba supported on an aluminum oxide (al2o3) catalyst for storage of nox in diesel engine exhaust gas. the efficiency of the process was studied with single and combined oxide systems. the behavior of the process was characterized and elucidated with the final aims of identifying the active and efficient mixed metal oxides to store nox from diesel engine exhaust gas. 2. materials and methods 2.1 catalyst preparation oxides of ba and k supported on al2o3 were tested for their effectiveness in the storage of nox during the lean reductant stage of the denox process. the introduction of active metals was carried out through a wet impregnation method using respective metal nitrates as the precursors. in a typical procedure, for the preparation of bao/al2o3 (10 wt. %), 3.5 ml of deionized water was added while stirring to 4.5 g of the supporting oxide. after that, 10 ml of an aqueous solution containing 0.96 g of barium nitrate [ba(no3)2] was mixed with the paste and stirred for two hours at room temperature. after evaporating the solvent at ≈ 100°c, the sample was dried at 110°c overnight and finally calcined in the air at 500°c for two hours. a fraction with particle sizes between 425–600 µm was obtained by pressing, crushing, and then sieving. the catalysts synthesized are denoted as mo(x)/al2o3, where x represent the weight composition of the metal oxide in the catalyst. 2.2 experimental setup the feed used in this study consisted of nitrogen (n2), oxygen (o2), 5,000 ppm of no in n2 and 5,000 ppm of isobutene (i-c4h10) in n2 gases. the flow rate was controlled by means of mass flow controllers. the reactor was made of 20 mm i.d. stainless steel tubing and was designed as a cross sectional detachable type with a center joint, where the catalyst packing could be mounted and removed easily from the reactor. the pressure drop across the reactor was negligible under the normal operating conditions. a thermocouple type k 88500-10 (cole-parmer, vernon hills, illinois, usa) connected to a multi-channel digital temperature scanner (cole-palmer-92000-05) was used to measure the temperature inside the reactor. the reaction temperatures (200–450 °c) were achieved by means of a horizontal tubular furnace (lindberg-tf-55035c, thermo fisher scientific, inc., waltham, massachusetts, usa). the feed gas components comprised of storage of nitrous oxide (nox) in diesel engine exhaust gas using alumina-based catalysts: preparation, characterization, and testing 66 1,000 ppm of no, 1,000 ppm of butane (c4h10), 8% o2, and balanced n2 were mixed to give the feed gas. the no storage activity of the catalysts was evaluated at atmospheric pressure with a total feed flow rate of 50 ml/minute. lean no that presented in feed was reacted and chemically stored in the catalyst storage component for 20 minutes. the packing of 200 mg of catalyst was mounted at the center of the catalytic reactor. the furnace was allowed to cool down and the tubing system was flushed with nitrogen for about 30 minutes to remove excess oxygen. the feed gas mixture was fed into the system at 50 ml/minute, and the reactor was heated to the desired reaction temperature before measuring the concentration of gases. the feed gas mixture was allowed to run for about 10 minutes to ensure a steady state and a uniform mixture before measuring the outlet concentration of the gases. an in-line gas chromatography (gc 8a) (shimadzu corp., kyoto, japan) was used for the analysis of the feed and product gases. the gc unit was operated in an isothermal condition (80°c) with helium as the carrier gas. the gc used was equipped with a thermal conductivity detector, two separating columns, and an integrator. a supelco molecular sieve 5a column (sigma aldrich, st. louis, missouri, usa) was used for separating n2, o2, and ic4h10 while a porapak q column (supelco) was used for the analysis of co2 and no. 2.3. characterization of catalysts the surface area pore volume and average pore diameter of the synthesized catalysts were measured using the accelerated surface area and porosimetry system (asap 2000) supplied by micromeritics instruments corporation, norcross, georgia, usa. the samples were dried overnight at 105°c and degassed for 12 hours under vacuum at a temperature of 300°c using the asap2000 instrument before measurements were performed (alsobaai et al. 2007a). powder x-ray diffraction (xrd) patterns of prepared materials were obtained using a d5000 x-ray diffractometer (siemens, munich, germany) with cukα radiation (λ = 1.54056 å) at 40 kv and 30 ma and a scanning speed of 2 degrees per minute (alsobaai et al., 2007b). scanning electron microscope (sem) studies were performed using a cambridge stereo scan 360 (wetzlar, germany) and polaron for the sputter coating. before scanning, the powder samples were spread onto the double-side carbon tape and coated with gold (20–30 nm thickness) in order to increase the conductivity and therefore the quality of the results (alsobaai et al. 2007c). an elemental analyzer (ea) was also used to analyze the composition of fresh and aged catalyst samples. 3. results and discussion 3.1 characterization of catalysts the porosity and surface characteristics in terms of brunner-emmett-teller (bet) surface area, pore volume, and average pore diameter of the catalysts are given in table 1. the catalysts loaded with k2o and k2obao showed lower surface area and pore volume compared to the unloaded al2o3 support. this finding is reasonable as the impregnated metals both fill up and plug some pores, making less area available for nitrogen adsorption and contributing to the weight of the catalyst, lowering the surface area and pore volume measured on a weight basis. figure 1(a) shows nitrogen adsorption-desorption isotherms of alumina and alumina-based catalysts. the isotherms were type iv, corresponding to mesoporous solids (halachev et al. 1996). as the relative pressure increased (p/po >0.2), the isotherms exhibited sharp inflections characteristic of capillary condensation within mesopores. in any case, the sharpness of this step suggested a uniform size pore system and provided evidence of the high quality of the catalysts. the pore size distributions of catalysts are illustrated in fig. 1(b). as shown in this figure, all catalysts exhibit mesopores with size peaks centered at 5–10 nm. figure 2 presents xrd traces of the catalysts. for al2o3, only diffraction peaks due to the alumina support material were observed, and no k or ba-related phases were evident. for k2o/al2o3 and k2obao/al2o3, new peaks at 2θ values of 22, 36, and 42 in addition of the al2o3 pattern which did not change from the original pattern. the presence of these new peaks confirmed that the k and ba species were present in the prepared materials. these results are in good agreement with those reported by kim et al. (2007). the structure and morphology of the fresh and spent catalysts were also investigated. topological information such as crystal structure and morphology of the fresh and a. alsobaai 67 table 1. bet surface area, pore volume, average pore diameter of alumina-based catalysts. sample surface area, m2/g mesoporous area, % microporous area, % pore volume, cm3/g average pore size (aps), nm al2o3 77.285 100 0 0.208 10.773 k2obao/al2o3 39.516 89.239 10.761 0.138 13.99 k2o/al2o3 7.457 98.155 1.845 0.0403 21.598 figure. 1 porosity measurements of alumina-based catalysts (a) adsorption (solid line) and desorption (dotted line) isotherm, and (b) pore size distribution. spent k2o(10)/al2o3 and k2o(5)bao(5)/al2o3 catalysts were studied using an sem. figure 3(a) shows a sem micrograph of fresh k2o(10)/al2o3 catalyst while fig. 3(b) shows the sem micrograph of the fresh k2o(5)bao(5)/al2o3 catalyst. as shown in these figures, the sample k2o(5)bao(5)/al2o3 had a clear crystalline shape compared to k2o(10)/al2o3. the micrograph of both samples suggested that the surface of alumina in the sample of k2o(5)bao(5)/al2o3 was homogenously covered by metal as compared to the sample of k2o(10)/al2o3. analytical techniques were also carried out to characterize the structure of the aged catalyst. the information about the structure 0 20 40 60 80 100 120 140 160 0 0.2 0.4 0.6 0.8 1 p o re v o lu m e, c c/ g al2o3 k2o/al2o3 k2obao/al2o3 0 50 100 150 200 250 300 350 0 20 40 60 80 100 120 140 k2obao/al2o3 al2o3 k2o/al2o3 (a) (p/po) pore diameter, nm (b) d if fe re n ti al p oo r vo lu m e (c c/ g x n m ) x e -0 5 storage of nitrous oxide (nox) in diesel engine exhaust gas using alumina-based catalysts: preparation, characterization, and testing 68 . figure 3. sem micrograph of fresh and aged catalysts: (a) fresh k2o (10)/al2o3, (b) fresh k2o(5)ba(5)/al2o3 (c) aged k2o(10)/al2o3, (d) aged k2o(5)bao(5)/al2o3. and morphology of aged catalyst was investigated. the topological information of the aged k2o(10)/al2o3 and (k2o(5)bao(5)/al2o3 catalysts were studied. the used sample was the catalyst that was previously subjected to the no storage condition for a duration of 1600 0 10 20 30 40 50 60 70 80 90 100 20 40 60 80 100 120 140 in te n si ty , c p s 2 theta figure 2. xrd pattern for alumina-based catalysts. (a) (b) (c) (d) a. alsobaai 69 seconds under a reaction temperature of 350°c. figures 3(c) and (d) show the sem micrograph of the k2o(10)/al2o3 and k2o(5)/al2o3 aged samples, respectively. as shown in the micrograph, there was no structured change to either sample after being subjected to the nox storage process if compared to the micrographs shown in figs. 3(a) and (b). as compared the fresh and aged samples, it could be concluded that as the coverage of metal in the surface increased, the capacity of storage also increased. similar observation were reported by fanson et al. (2003). however, the researchers were unable to unambiguously identify the species of the surface compound that resulted and were unsure of their exact origin. elemental analysis was also conducted on the spent catalysts. the composition of c, hydrogen (h), and n content is shown in table 2. the k2o(10)/al2o3 catalyst with notation (a) was tested under a reaction temperature of 250°c, while the catalyst with notation (b) was tested under a reaction temperature of 350°c. the percentages of c, h, and n were found to increase from those of the fresh catalyst. the existence of c and h was caused by the oxidation of the hydrocarbon (c4h10) in the feed gas by high temperatures (>250°c) (despres et al. 2003). the clear evidence of n chemisorption could be observed when n content varied between 0.35–0.45 in the fresh sample. the decomposition process of no to n2 and o2 inside the catalyst contributed to the growing n content (bethke et al. 1995). on the other hand, it was found that bao/al2o3 was not efficient to catalyze the oxidation of no to no2 and then to n2 if compared to k2o/al2o3 and cao/al2o3. it has been reported that bao/al2o3 did not store no to a great extent. a similar observation was reported by lietti et al. (2001). table 2. composition of carbon, hydrogen, nitrogen and sulfur content in aged catalyst. element k2o(10) /al2o3 a k2o(10) /al2o3 b bao(10) /al2o3 carbon (wt %) 0.54 0.50 0.78 hydrogen (wt %) 0.66 0.57 0.50 nitrogen (wt %) 1.83 2.08 0.79 atested at 250°c; btested at 350°c 3.2 process studies the experimental results obtained from a series of catalytic tests on no storage by single and mixed metal oxide catalysts were presented and discussed. experiments were carried out to determine the effect of the important independent process variables (i.e. storage compounds, reaction temperature, percentage of metal loading, and metal composition) at fixed amounts of catalyst, and a fixed reaction duration and feed composition. the experiments were performed at a temperature of 350°c with a reaction duration of 20 minutes. in order to examine the performance of different storage components, 20 mg of different catalysts (i.e. bao(10)/al2o3 and k2o(10)/al2o3) were investigated. table 3 compares the activities of these catalysts in nox storage at 350°c. for each experiment, the concentration of no in the outlet gas measured was lower than that of the inlet gas. these results indicate that no was stored on the catalyst under the lean conditions. in the presence of 8% o2, k2o/al2o3 was more active for the storage of no than bao/al2o3. the sequence of basicity strength is k>ba. therefore, the result suggested that the stronger the basicity of the nox storage compound, the larger the quantity of the no stored. a similar observation was reported by takahashi et al. (1996). it was attributed to the higher stability of nitrates formed by oxides of stronger basicity (takahashi et al. 1996). potassium also favored surface mobility, and had a higher surface basicity (milt et al. 2003a). on the other hand, it was found that bao/al2o3 was not efficient in catalyzing the oxidation of no to no2; therefore, it could not efficiently store no to a significant extent, even in the presence of oxygen (lietti et al. 2001). also the ionic size and charges of k+ is less than ba+, and this may be considered in explaining ion-supported interactions and a loss of porosity in these samples. the larger ionic size can cause more loss in porosity and less activity. it is generally reported in the literature that the formation of no2 from no and o2 is a necessary initial step before nox storage can take place during lean conditions (bethke et al. 1995; fridell et al. 1999). however, this simple step already includes several uncertainties and, further, may be the sum of several elementary reaction steps (fridell et al. 1999). the nox storage process for lean conditions can be storage of nitrous oxide (nox) in diesel engine exhaust gas using alumina-based catalysts: preparation, characterization, and testing 70 table 3. amount of no storage by different catalysts at a temperature of 350°c and reaction time of 20 min catalyst no concentration (ppm) amount of no stored (%) (cno)in (cno)out bao(10)/al2o3 1014 638.3 39.0 k2o(10)/al2o3 1009 450.4 55.4 assumed to comprise the following steps with m as a storage component in the catalyst: ag nono  (1) ag 2 o2o  (2) 2 aa noono  (3) 23)no(mmoo3no2  (4) 232 )no(mmoono2  (5) the superscript (g) refers to gas form while superscript (a) refers to the absorbed form of the chemical compound. this is a somewhat simplified reaction scheme. for example, no2 adsorption and desorption are not included (fridell et al. 1999). a key role of the catalyst surface was found to be the oxidation of no to no2 (milt et al. 2003b). in order to determine the parameters for the nox storage process, the temperature dependence of the reaction at constant feed gas composition was investigated. for this purpose, catalysts bao(10) / al2o3 and figure 4. percentage of no storage onto bao(10)/al2o3 and k2o(10)/al2o3 as a function of reaction temperature. k2o(10)/al2o3 were evaluated on a catalytic reactor rig to determine their properties as a function of temperature at 200–450°c. both metal oxide catalysts were tested under the same reaction conditions. figure 4 shows the result of the temperature dependence of the catalytic activity under lean conditions. the difference between inlet and outlet no concentrations indicated the nox storage. the maximum no storage for k2o(10)/al2o3 of around 58% occurred at 250 oc. after this, the no storage reaction decreased with increasing temperature. as reported by fridell et al. (1999), the decrease in no storage at higher temperatures was associated with the stability of metal nitrate. in contrast to the k2o(10)/al2o3 catalyst, the bao(10)/al2o3 catalyst showed higher conversions at higher temperatures. for bao(10)/al2o3, the lean feed shows a peak steady-state nox storage of approximately 47% at 450°c. the bao(10)/al2o3 catalyst had higher no storage activity than the k2o(10)/al2o3 catalyst at higher temperatures, particularly above 400°c. this suggested that the ionic bonding character of ba nitrates increased with increasing temperatures (sedlmair et al. 2003). the no storage for both catalysts exceeded 40% at 200°c. it was because of the al2o3 catalysts that they were capable of playing an important role as the storage site at temperatures below 300°c (westerberg and fridell 2001). contrary to the literature, maximum nox storage was not seen at around 380oc for either case. similar behavior was also reported by fridell et al. (1999). a series of k2o/al2o3 catalysts containing different k loadings (in the range 0–25 wt. %) was prepared and tested for no storage. the objective was to investigate the effect of metal loading on the storage process. accordingly, the role of the k loading on the storage of nox could be analyzed. figure 5 shows the no storage performance of the catalyst as a function of k2o loading. supporting 5–10 wt % 20 30 40 50 60 70 150 200 250 300 350 400 450 500 temperature ( o c) n o r em o v ed ( % ) . bao(10)/al2o3 k2o(10)/al2o3 a. alsobaai 71 figure 5. amount of no storage for different k2o loadings in k2o/al2o3 catalyst. of k2o on al2o3 was found to improve the lean no storage activity of al2o3, but the incorporation of 15 wt % k showed no significant improvement in the storage activity. the further addition of k2o proved to be detrimental to the performance of the catalyst. westerberg and fridell (2001) proposed that k2o could play an important role as a storage site at temperatures below 300°c. as the loading of the catalysts increased, the activity increased as indicated by the percentage of no storage. the highest no storage over k2o/al2o3 occurred when k contents were between 10–15 wt %. however, beyond that, the percentage of no storage started to decreased, probably due to the diffusion limitations at higher loading. these data were in agreement with some of the literature (bethke et al. 1995; castoldi et al. 2004; milt et al. 2003a). the decrease in activity for the higher loading samples was also attributed to the dissolution of some amount of k(no3) during the impregnation process and caused higher k coverage of the al2o3 surface (castoldi et al. 2004; dawody et al. 2004). both k and al2o3 were able to react with no2-forming nitrates species. both k2o(5)/al2o3 and k2o (20)/al2o3 produced no retentions lower than, or similar to, pure al2o3. in this way, k2o (10)/al2o3 showed the highest stability for nitrate species among the k(x)/al2o3 formulations studied. the effect of the addition of bao to the k2o(x)/al2o3 system was studied with the objective of improving the interaction with no molecules and improving the catalytic activity for the no storage reaction by increasing the formation of surface no intermediates. activity studies were performed on k2o(3)bao(7)/al2o3, k2o(5)bao(5)/al2o3 and 1k2o(7)bao(3)/al2o3. the amounts of no stored are shown in table 4. first, the presence of ba led to lower no storage activity for k2o(3)bao(7)/al2o3. by increasing the loading of potassium into these catalysts, an increase in the activity could be observed. the highest no storage over ba,k/al2o3 occurred when the loading of both metals was equal (i.e. k2o (5)bao(5)/al2o3). the k2o(7)bao(3)/al2o3 catalyst presented an activity similar to k2o (10)/al2o3, which was in agreement with results reported by matsumoto, 2004. the combination of the ba and k loading resulted in a catalyst capable of showing improvement in the catalytic activity compared to single metal loading, and may be attributed to some synergistical effects between these two metal oxides. the catalyst behavior was influenced by three factors: ion size, alkalinity of the element, and preservation of the original surface area and porosity of the al2o3. the heavy element ba produces fewer ba++ ions than the lighter element k+. the presence of more k+ ions can give k/al2o3 better reactivity for nox storage. 4. conclusion the no storage amount in a lean exhaust stream was dependent on the basicity of the storage compounds. k2o/al2o3 catalyst was found to be the more active for the no storage process than bao/al2o3. the no storage capacity for k2o(10)/al2o3 also decreased with increasing temperature, especially at above 300°c. the no storage process achieved its maximum at an intermediate temperature (250°c) for k2o(10)/al2o3. as for bao(10)/al2o3, the no storage capacity increased with increase in operating temperature. the effect of k loading 0-25 % (w/w) in the storage over k2o/al2o3 samples was investigated and the results indicated that the addition of k2o negatively affected the morphological properties of the catalyst system at loading higher than 15 %. nevertheless, the increase in the k2o loading resulted in a strong increase of the no adsorption at breakthrough. under these conditions, a maximum value of no storage of around 60% was observed for the k(15)/al2o3 sample. significant amounts of no were found to be stored in the catalysts 20 30 40 50 60 70 0 5 10 15 20 25 30 ko loading (wt. %) n o r em o ve d ( % ) . k2o loading (wt. %) storage of nitrous oxide (nox) in diesel engine exhaust gas using alumina-based catalysts: preparation, characterization, and testing 72 table 4. amount of no storage over different storage composition in lean phase at a temperature of 350°c and reaction time of 20 min. catalyst no concentration (ppm) amount of no stored (%) (cno)in (cno)out k2o(3)bao(7)/al2o3 988 514.1 48.0 k2o(5)bao(5)/al2o3 996 339.2 65.9 k2o(7)bao(3)/al2o3 1012 422.6 58.2 containing k(5)-ba(5)/al2o3. the combination of ba and k loading in catalyst led to an improvement in the catalytic activity compared to single metal catalyst. references alsobaai am, zakaria r, hameed hb (2007a), hydrocracking of petroleum gas oil over niw/mcm-48-usy composite catalyst. journal of fuel processing technology 88: 921928. alsobaai am, zakaria r, hameed hb, (2007b), characterization and hydrocracking of gas oil on sulphided niw-mcm-48 catalysts. chemical engineering journal 132(1-3): 173181. alsobaai am, zakaria r, hameed hb (2007c), gas oil hydrocracking on niw/usy catalyst: effect of tungsten and nickel loading. chemical engineering journal 132(13): 77-83. athanasios gk, margaritis k, carlo b, gabriele db, abdurrahman i, ingemar d (2015), impact of combination of egr, scr, and dpf technologies for the low-emission rail diesel engines. emission control science and technology 1(3): 213-225. bethke ka, kung mc, yang b, shah m, alt d, li c, kung hh (1995), metal oxide catalysts for lean nox reduction. catalysis today 26: 169-183. castoldi l, nova i, lietti l, forzatti p (2004), study of the effect loading for catalytic activity of pt-ba/al2o3 model catalysts. catalysis today 96: 43-52. centi g arena ge, perathoner s (2003), nanostructured catalyst for nox storagereduction and n2o decomposition. journal of catalysts 216: 443-454. dawody j, skonglundh m, fridell e (2004), the effect of metal oxide additives (wo3, moo3, v2o5, ga2o3) on the oxidation of no and so2 over pt/al2o3 and pt/bao/al2o3 catalysts. journal of molecular catalysis a: chemical 209: 215-225. despres j, koebel m, kröcher o, elsener, wokaun a (2003), storage of no2 on bao/tio2 and the influence of no. applied catalysis b: environmental 43: 389-395. epling ws, parks je, campbell gc, yezerets a, currier nw, campbell le (2004), further evidence of multiple nox sorption sites on nox storage/reduction catalysts. catalysis today 96: 21-30. fanson pt, horton mr, delglass wm, lauterbach j (2003), ftir analysis of storage behavior and sulfur tolerance in barium-based nox storage and reduction (nsr) catalysts. applied catalysis b: environmental 46: 393-413. fridell e, skoglundh m, westerberg b, johansson s, smedler g (1999), nox storage in barium-containing catalysts. journal of catalysis 183: 196-209. halachev t, antanasova p, agudo al, arias mg, ramirez j (1996), activity of p-niw/al2o3 catalysts with varying phosphorus content in the hydrogenation of naphthalene. catalysis a: general 136: 161-175. josh ap, jennifer al, todd jt, james ep (2013), lean nox trap chemistry under lean-gasoline exhaust conditions: impact of high nox concentrations and high temperature. topics in catalysis 56(1): 8993. kabin ks, muncrief rl, harold mp (2004), nox storage and reduction on a a. alsobaai 73 pt/bao/alumina monolithic storage catalyst. catalysis today 96: 79-89. kim dh, kwal jh, szanyi j, burton sd, peden ch (2007), water-induced bulk ba(no3)2 formation from no2 exposed thermally aged bao/al2o3. applied catalysis b: environmental 72: 234-240. lietti l, forzatti p, nova i, tronconi e (2001), nox storage reduction over pt-ba/γ-al2o3 catalyst. journal of catalysis 204: 175-191. matsumoto s (2004), recent advances in automobile exhaust catalyst. catalysis today 90: 183-190. michael f, andreas w, bastian h, sebastian k (2014), particulate emissions from diesel engines: correlation between engine technology and emissions. journal of occupational medicine and toxicology 9:6. milt vg, pisarello ml, miró ee, querini ca (2003a), abatement of diesel-exhaust pollutants: nox storage and soot combustion k/la2o3 catalysts. applied catalysis b: environmental 41: 397-414. milt vg, querini ca, miró ee, ulla ma (2003b), abatement of diesel-exhaust pollutants: nox adsorption on co, ba,k/ceo2 catalyst. journal of catalysis 220: 424-432. su y, amiridis md (2004), in situ ftir studies of the mechanism of nox storage and reduction on pt/ba/al2o3 catalysts. catalysis today 96: 31-41. sedlmair c, seshan k, jentys, lercher jj (2003), elementary steps of nox adsorption and surface reaction on a commercial storagereduction catalyst. journal of catalysis 214: 308-306. takahashi n, shinjoh h, ijima t, suzuku t, yamazaki k, yokota k, suzuki h, miyoshi n, matsumoto s, tanizawa t, tanaka t, tateishi s, kasahara k (1996), the new concept 3-way catalyst for automotive leanburn engine: nox storage and reduction catalyst. catalysis today 27: 63. westerberg b, fridell e (2001), a transient ftir study of species formed during nox storage in the pt/bao/al2o3 system. journal of molecular catalysis a: chemical 165: 249-263. copy (2) of comp140404.qxd the journal of engineering research vol.3, no.1 (2006) 1-9 1. introduction sleep apnea is complete or partial cessation of breathing during sleep. obstructive sleep apnea (osa) is the common form of apnea that occurs when the upper airways are partially or completely obstructed for 10 seconds or more due to the relaxation of the dilating muscles. the dilating muscles contract during inspiration to prevent the collapse of the upper airways caused by inward air suction (rodenstein, et al. 1990). hypoapnea is a form of apnea where the airways partially collapse and _________________________________________ *corresponding author’s e-mail: abhossen@squ.edu.om cause 50% reduction of air accompanied by drops in oxygen desaturation of at least 4% followed by compensating hyperventilation (rodenstein, et al. 1990; tsi 1999; aasm task force report, 1999; penzel, 2000). the severity of the apnea is measured by apnea hypoapnea index (ahi), which is the number of apnea and hypoapnea episodes per hour. osa is said to be mild if ahi is between 5 and 15 per hour, moderate if ahi is between 15 and 30 per hour, and severe if ahi is greater than 30 per hour. if ahi is less than 5 per hour, the case is considered normal (aasm task force report, 1999). the most commonly used technique for sleep analysis is an overnight polysomnographic recording, which is an statistical signal characterization of spectral analysis of heart rate variability for screening of patients with obstructive sleep apnea a. hossen*1, b. al ghunaimi1 and m.o. hassan2 1department of electrical and computer engineering, college of engineering , sultan qaboos university , p.o. box 33, al-khoud, 123, muscat, 2department of physiology, college of medicine, sultan qaboos university , p.o. box 35, al-khoud, 123, muscat, sultanate of oman received 14 april 2004; accepted 9 march 2005 abstract: a new screening technique for obstructive sleep apnea (osa) is implemented. this technique is based on finding the statistical signal characterization (ssc) parameters of the spectrum of pre-processed r-r-interval (rri) data. a single classification factor (cr) is selected to classify between patients with obstructive sleep apnea and normal controls. both nonparametric spectral analysis techniques (with welch method as an example) and parametric techniques (with burg method as an example) are used to estimate the power spectral density. the data tested in this work are drawn from mit database. both trial and test (challenge) groups of data are used. each of these two groups contains 20 osa and 10 normal records. the trial data is used to set the threshold value of the classification factor, which is then used in identifying the challenge data. the total accuracy of the screening is about 93% and 90% using burg and welch algorithms respectively. keywords: sleep apnea, statistical signal characterization, parametric and non-parametric spectral analysis .ωƒædg aéækcg ¢ùøæàdg ´é£≤fg ≈°vôe á∏hô¨d ö∏≤dg äé° ñf ∫~©e ò¨àÿ …ooîdg ∞«£dg π«∏ëàd »fé°üm’g ∞°uƒdg á≤jôw ägò¨àe äéfé«ñd …ooîdg ∞«£dg π«∏ëàd »fé°üëm’g ∞°uƒdg πegƒy oééjg ≈∏y á≤jô£dg √òg ~ªà©j .ωƒædg aéækcg ¢ùøæàdg ´é£≤fg ≈°vôe á∏hô¨d √~j~l á«æ≤j åëñdg øª° àj :á°uón …ooîdg ∞«£dg oééj’h .aéë°u’g ¢uéî°t’g h ωƒædg aéækcg ¢ùøæàdg ´é£≤fg ¢vôã ≈°vôÿg úh õ««ªà∏d ~mgh πeé©e á≤jô£dg √òg ω~îà°ùj .é≤ñ°ùe á÷é©ÿg ö∏≤dg äé° ñf ∫~©e äéeƒ∏©ÿg áµñ°t øe äéfé«ñdg äòncg .(∑qƒh á≤jô£c) »ª∏©ÿg ∞«£dg á≤jôwh (¢û∏jh á≤jô£c) …~«∏≤àdg »ª∏©eódg ∞«£dg á≤jôw éªg úà«°sé°sg úà≤jôw åëñdg ‘ ω~îà°ùf iôn’gh ¬e~îà°ùÿg á«erqgƒÿg öjq~àd éªg~mcg äéfé«ñdg øe ¿éàyƒª› âe~îà°sg .i~ëàÿg äéj’ƒdg ‘ »lƒ∏æµàdg õà«°s ƒ°t é°sée ~¡©e äéfé«h i~yéb øe ~j~ëàdéhh á«ÿé©dg ᪫≤dg oééj’ öjq~àdg áyƒª› âe~îà°sg .aéë°ucg ¢séf’ é¡æe (10) h ,ωƒædg aéækg ¢ùøæàdg ´é£≤fg ≈°vôÿ é¡æe (20) ,óé°s (30) º° j áyƒª› πc .á«erqgƒÿg iaéøc ¢üëød áñ°ùæh äéfé«ñdg úh õ««“h á∏hôz ábo á≤jô£dg â£ygh .¢üëødg áyƒª› äéféñ«h ∞«æ°üàd álô◊g ᪫≤dg √òg âe~îà°sg ºk øeh ,º«∏°sh ¢ jôe úh äéfé«ñdg ∞«æ°üàd álô◊g .‹gƒàdg≈∏y ¢û∏jh h ∑qƒh éà≤jôw ωg~îà°sg ~æy %90 h %93 áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.á«∏ª©eódg h á«∏ª©ÿg ¥ô£déh ∞«£dg π«∏– ,iqé°tód »fé°üm’g ∞°uƒdg ,ωƒædg aéækg ¢ùøæàdg ´é£≤fcg : 2 the journal of engineering research vol.3, no.1 (2006) 1-9 overnight monitoring of sleep-related body functions, such as brain activity, eye movements, muscle activity, leg and arm movements, heart rate, air snoring, respiration effort, and blood oxygen level. these activities are measured by several special electrodes and sensors attached to the body. although the polysomnography provides reliable results in the diagnosis of sleep disorders, it is expensive, time consuming and inconvenient to the patient (boyer and kapur, 2002). usually, patients undergo a primary diagnosis, by primary screening techniques, to evaluate their condition, after which a decision is made if further in depth diagnosis is required through polysomnography (stevenson, 2003; pack, 1993). these techniques are also helpful for follow up of treatment. an example of primary diagnosis techniques is the pulse oximetry, which is based on measuring oxygen saturation in the blood via an oximeter attached to a fingertip or ear. the presence of apnea is evaluated by measuring the number of oxygen desaturations per hour, which is also called oxygen desaturation index (odi). however, the screening results by oximetry methods are inconsistent as reported by pack (1993) and netzer (2001). the development in biomedical signal processing technology has led several researchers to propose other methods for apnea detection and screening by processing and manipulating one or more of the body activities such as the electrocardiogram (ecg) signal, electroencephalogram (eeg) signal and the heart rate variability (hrv) or the r-r intervals (rri). this study utilized the rri data, which is extracted from ecg data records, to accomplish the screening process. the pattern of ecg signal is composed of several waves that are repeated on every beat. these waves were named arbitrarily as p, q, r, s and t as shown in fig.1. the beat is recognized by the qrs waves, which are also called qrs complex. the beat is said to be normal (n) if it is originated from the sino-atrial node of the heart, and ectopic if it is originated elsewhere. the time between peaks of two consecutive r waves of the ecg signal is referred to as r-r interval (rri), which is used to measure the heart rate. heart rate variability (hrv) implies the variation of heart rate that has been proven to accompany the variation of several physiological activities such as breathing, thermoregulation and blood pressure changes (task force of the european society of cardiology and the north american society of pacing and electrophysiology, 1996). theoretically, in osa the cessation of breathing will cause the respiration center in the brain to activate its autonomic components (sympathetic and parasympathetic) that send feedback impulses to the heart to compensate for the lack of oxygen and low blood pressure. this interaction between the heart and the brain is reflected into the beat-to-beat variation of the heart rate. therefore, the analysis of hrv should somehow reveal the variations in breathing. many algorithms manipulate hrv or rri data either in frequency-domain or in timedomain, to extract screening parameters. several studies have examined hrv using spectral analysis techniques under different conditions (task force of the european society of cardiology and the north american society of pacing and electrophysiology, 1996) and in osa (khoo, et al. 1999). in (drinnan, et al. 2000), the fft is applied directly on the rri signal, and the area under the spectral curve between 0.01 and 0.05 hz is divided by the area between 0.005 and 0.01 hz. the algorithm used led to the classification accuracy of 93.3 % when applied to the mit challenge data. another technique employed the hilbert transformation of the rri time series to derive the instantaneous amplitudes and frequencies of the series (mietus, et al. 2000). some basic statistics such as the mean and the standard deviation were calculated for every 5-minute sliding window. a comparison is made between the probability distribution of the measured statistics in order to set up detection threshold limits. the algorithm used obtained accuracy of 93.3% on mit challenge data. a comparison of several time domain and frequency domain algorithms for apnea detection is found in the work reported by penzel et al. (2002). in this paper a new technique is presented. this technique computes the statistical signal characterization parameters for the power spectral density (psd), which is computed using different spectral estimation techniques, of the pre-processed rri data. so instead of finding the power at different frequency bands in the psd, statistical parameters are calculated for the estimated psd. these parameters depend on the morphology (number of maxima and minima) of the psd. the paper is organized as follows: section 2 deals with the pre-processing steps of the data used in this work. in section 3, welch and burg spectral analysis techniques are introduced. the statistical signal characterization method is discussed in section 4. data implementation steps are explained in section 5. different results of burg and welch methods are listed and compared in section 6. this section also includes the results of parametric methods other than burg. moreover, a comparison with other techniques (drinnan, et al. 2000; mietus, et al. 2000) are also provided in this section. conclusions are given in section 7. figure 1. ecg pattern 3 the journal of engineering research vol.3, no.1 (2006) 1-9 2. data: sources and pre-processing 2.1 the data the ecg records are drawn from the database of massachusetts institute of technology (mit) (physionet: an nih/ncrr research resource for complex physiologic signals). the ecg signals (single channel) were extracted from polysomnographics recordings with sampling rate of 100 hz and 12-bit resolution for an average duration of 8 hours. the database contains 35 trial records and 35 test records. each set of the records contains 20 osa records, 10 normal records, and 5 borderline records. apnea records are those having 100 minutes or more of apnea during the recordings and at least one hour with apnea/hypoapnea index of 10 or more. the normal records are those having less than 5 minutes of apnea during the recording. the borderline records are those having 5 to 99 minutes of apnea during the recording, and at least one hour with apnea/hypoapnea index of 5 or more (physionet: an nih/ncrr research resource for complex physiologic signals). in this study we only used the apnea records and the normal records; borderline records are excluded. the trial data is used to set up the classification factor. this factor is then used to classify the test data. 2.2 pre-processing of data rri series could be exposed to different types of errors throughout its generation process. the original ecg may be exposed to different type of physiological and environmental interferences that could overwhelm the ecg features and lead to detection of false qrs peaks or misdetection of normal qrs peaks (friesen, et al. 1990). moreover, the variation of heartbeats is non-periodic and would result in non-periodic variation of the instantaneous rri and irregular inter-sample spacing. furthermore, the qrs detection algorithm, regardless of its accuracy, could miss normal peaks or detect false or ectopic beats. large errors induced by missing values, outliers, non-stationarity and irregular inter-sample spacing could influence the accuracy of any analysis technique. therefore the rri data usually undergo several processing and filtering stages before they can be used in an analysis. figure 2 shows the pre-processing steps needed before applying the new technique. the qrs detection, which is the primary process for every ecg signal analysis technique, is accomplished by the “ecgpuwave” software (physionet: an nih/ncrr research resource for complex physiological signals). the rri data are the normal-to-normal (n-n) intervals obtained directly from qrs detector without any smoothing and filtering steps; therefore it could contain false intervals, missed and/or ectopic intervals. removing of outliers is achieved by using a 41-points moving average filter (maf). re-sampling at 1 hz and substituting for the missed peaks are then accomplished by simple linear interpolation (mietus, et al. 2000). the re-sampling and the estimation of missed values are intended to have an equally spaced rri data and preserve the temporal sequence that is necessary for the frequency domain analysis (bigger, 2000). the next processing steps are intended to emphasize the oscillations of rri data that are commonly found in apnea patients but less common in normal subjects. these oscillations lie in the vlf band of rri spectrum, especially between 0.01 and 0.04 hz. a high pass filtering (hpf) is used to remove the oscillations below 0.01 hz. hpf is implemented by a local de-trending over an 81point moving window, which gives a highpass 3db cutoff at 0.01 hz. in other words, a simple linear regression is used to estimate the trend line of the rri series over a sliding 81point window. the estimated value at the center of each window forms the high pass filtered rri series (rri-hf) (mietus, et al. 2000). this set of data is now ready for the analysis. 3. spectral analysis spectral analysis is used to find the psd, which is the frequency content of a signal or a system based on a finite set of data. the various techniques of psd estimation can be classified into non-parametric and parametric methods. 3.1 non-parametric methods nonparametric methods estimate the psd directly from the signal itself by fft. the simplest method of such methods is the periodogram, which is the magnitude squared of the fft of the signal or a section of the signal. examples of improved versions of the periodogram are welch's method, which is one of the most popular nonparametric techniques. it is employed in this study as an example of such methods (kay, 1988). this method uses the following steps to compute the psd of a signal: 1. apply a non-rectangular data window (hamming) to different sections of the signal. 2. find the periodogram, that is the magnitude squared of the fft of the windowed section. 3. compute the average of the periodograms of all sections. qrs detection remove outliers ecg rri resample 1 hz hpf rri-hf figure 2. block-diagram of pre-processing steps 4 the journal of engineering research vol.3, no.1 (2006) 1-9 3.2 parametric methods parametric methods are based on modeling the signal, to estimate its psd, as the output of a linear system driven by white noise. these methods firstly estimate the parameters (coefficients) of the linear system that hypothetically construct the signal. the most commonly used linear system model is the all-pole model, a filter with all of its zeroes at the origin in the z-plane. the output of data of such a filter for white noise input is called an autoregressive (ar) process of order p, which represents the number of poles. examples of parametric methods are the yule-walker autoregressive (ar) method, covariance method (pcov), modified covariance method (pmcov) and the burg method (proakis and manolakis, 2000). the burg method, which is used in this study as an example of parametric techniques, is based on minimizing the forward and backward prediction errors while satisfying the levinson-durbin recursion (marple, 1987). this method avoids the calculation of the autocorrelation function; and instead computes the reflection coefficients. this method has a good accuracy in estimating the ar model for short data length and ensures stable ar model. simply, the burg method has the following steps: 1. select an order to the all-pole filter model. 2. estimate the lattic reflection coefficients from the signal. 3. fit an ar model to the estimated reflection coefficients. some parametric methods generate frequency component estimates for a signal based on an eigenanalysis or eigendecomposition of the correlation matrix. examples are the multiple signal classification (music) method and the eigenvector method, peig (proakis and manolakis, 2000). in fig. 3, the spectrums of both normal and osa data segments are estimated using welch (with n=128) and burg (with n =128, p=10) algorithms. 4. statistical signal characterization the statistical signal characterization (ssc) is a method that characterizes a waveform not only as a function of the frequency component amplitudes (like fft) but also as a function of the relative phases of its frequency components (hirsch, 1992). the input waveform to the ssc process is basically divided into segments where each segment is bounded by two extrema: maxima and minima as shown in fig. 4. the segment amplitude, an, and the segment period, tn, are defined as: (1) n=1,2,3, (2) there are four ssc parameters that can be computed from the time-domain signal. the parameters are the amplitude mean ma, the period mean mt, the amplitude mean deviation da, and the period mean deviation, dt: (3) (4) 0 10 20 30 40 50 60 70 -60 -50 -40 -30 -20 -10 spectral analysis of normal subject a m p lit u d e burg welch 0 10 20 30 40 50 60 70 -70 -60 -50 -40 -30 -20 -10 spectral analysis of osa subject frequency index a m p lit u d e burg welch figure 3. spectrums of normal and osa data segment estimated by welch (n=128) and burg (n=128, p=14) methods 1nnn ttt −−= 1nnn aaa −−= ( )∑ = = sn 1i sia n/am ( )∑ = = sn 1i sit n/tm 5 the journal of engineering research vol.3, no.1 (2006) 1-9 (5) (6) where ai is the amplitude of the ith segment, and ti is the period of the ith segment and ns is the total number of segments. the ssc process is usually applied to a time domain signal. in this work, the ssc process is modified by applying it to the spectrum (psd) of a time domain signal. for example, a similar figure to fig. 4 can be plotted by considering the maxima and the minima in the spectral signal as in fig. 5. in fig. 4 the x-axis represents the time whereas in fig. 5 it represents the frequency. hence, corresponding changes have to be applied to eqs. (4) and (6) by replacing any time point with a frequency point. to initially set up the screening algorithm, the ssc were applied to the trial data. this is in order to find the best ssc parameter(s) that screen the osa patients from normal subjects. only two parameters (ma and da), which are both related to the amplitude of the psd, are found to be useful in our work. figure 6 shows the steps of determining the two ssc parameters from the estimated specrum of the (rri-hf) data. 5. data implementation 5.1. performance measures the performance of the technique is evaluated by three main metrics: specificity, sensitivity and accuracy as defined below (rangayyan, 2001). (7) time a m pl itu de a1 a2 a3 a4 a5 a6 a0 t0 t1 t3 t4 t5 t6 t2 figure 4. segment amplitude and period characterization ( )∑ = −= sn 1i saia n/mad ( )∑ = −= sn 1i stit n/mtd f0 f1 f2 f3 f4 f5 f6 a0 a1 a2 a4 a5 a6 a3 a m pl itu de frequency figure 5. segment amplitude and frequency characterization 100 fntp tp (%) ysensitivit + = 6 the journal of engineering research vol.3, no.1 (2006) 1-9 (8) (9) where tp, tn, fn, fp are defined in table 1 (rangayyan, 2001), and t is the total number of data under test. sensitivity represents the ability of a classifier to detect the positive cases, e.g. osa. specificity indicates the ability of a classifier to detect negative cases, e.g. normal subjects. accuracy represents the overall performance, which indicates the percentage of correctly classified positives and negative cases from the total cases. 5.2. implementation procedure the following procedure is implemented in this work: 1. estimate the psd of all successive segments of the rri-hf data using welch or burg method or any other method. 2. compute the ssc parameters (ma and da) of the psd of all successive segments. 3. find the maximum values, max (ma) and max (da) among the results of all segments. 4. calculate (cr) as the sum of max (ma) and max (da). 5. find (manually) the threshold value of (cr) to identify the trial data with best performances. this can be done by sketching a line that can separate better between the data results of both groups (normal and apnea). 6. find the performance of the algorithm on test data using the same threshold. 7. compute the overall performance of the screening process. 6. results and discussion in order to study the effect of the transform length n and the filter order p in the performance of the algorithm, the procedure is repeated for n = 64, 128 and 256 for both welch and burg methods, and for p = 10, 14 and 20 for burg method. other filter orders were also attempted but the results obtained were low and, therefore, been excluded. 6.1. results tables 2-4 show the performance results of welch method at different transform lengths n. tables 5-9 show the results for burg method at different transform length n and filter order p. the best accuracy obtained for trial data by welch method is 93.3% with n=128 and 256. for test data the maximum accuracy is 86.7%, obtained with n=256. the overall performance shows that the accuracy increases from 81.7% to 90% with n increasing from 64 to 256. moreover, the overall specificity is constant at 85%; whereas sensitivity (detection of more osa cases) improves with increasing n causing the improvement in accuracy. the maximum accuracy obtained by burg method for trial and test data are 96.7% and 90% respectively. the best overall accuracy is 93.3%. these results are obtained with n=128 and p = 14. moreover, by observing the burg accuracy results in tables (5,7,9) for p=14, which are 80%, 93.3% and 81.7%, it can be noticed that the increase in n had not always lead to better results. in addition, the results of burg method with n=128 at different filter order show that the accuracy is not improving consequently with increasing the filter order. therefore, the results of burg method depend on selecting an optimum filter order and transform length n. the parametric burg method results in better screening capability (accuracy=93.3%) than the non-parametric welch method (accuracy=90.0%). other parametric methods have been tested with n=128 100. fptn tn (%) yspecificit + = normal osa normal tn fp osa fn tp table 1. confusion matrix efficiency trial test overall sensitivity 85% 75% 80% specificity 70% 100% 85% accuracy 80% 83.3% 81. 7% table 2. results of welch method (n=64, cr=0.517) efficiency trial test overall sensitivity 90% 85% 87.5% specificity 100% 70% 85% accuracy 93.3% 80% 86.7% table 3. results of welch method (n=128, cr=0.378) efficiency trial test overall sensitivity 100% 85% 92.5% specificity 80% 90% 85% accuracy 93.3% 86.7% 90% table 4. results of welch method (n=256, cr=0.27) 100. fntnfptp tntp (%) accuracy +++ + = figure 6. statistical signal characterization parameters 7 the journal of engineering research vol.3, no.1 (2006) 1-9 and p=14 for performance comparison with burg. the methods and their results are shown in table 10. the results of welch and burg methods are also shown in this table for comparison. the parametric methods resulted in different accuracies above 80%. the second best accuracy after burg is 90%, which is obtained by peig method. figures 7 and 8 show the results (computed values of cr) of all data using both welch and burg algorithms respectively at fft length n=128 and filter order p =14 (for burg only). in these figures an optimum threshold value (cr) is selected manually (see the line sketched between the two groups) from the trial data, with a value of 0.378 for welch method and 0.63 for burg method. any subject with (cr) above the threshold is classified as normal and any subject with (cr) less than the threshold is classified as osa. 6.2. comparison with other methods the new technique is compared with other methods by drinnan, et al. (2000) and mietus, et al. (2000). the results of screening of both trial data and test data are shown in table 11. by applying the new technique with burg spectral analysis to the trial set, we could correctly classify 29 out of 30 of combined osa and normal subjects. when this technique is applied to the test set, it correctly classified 27 out of 30 of the subjects. the results show that this method have overall better performances compared to other methods used in the comparison. 7. conclusions a new method for osa screening is investigated in this paper. it is based on estimating the ssc parameters of the spectral analysis of the pre-processed and filtered rri data. two basic algorithms are used for the spectral estimation (welch and burg) as examples to non-parametric and parametric spectral analysis techniques respectively. the work shows significant results in screening of subjects with osa from those of normal controls. this study has shown that the morphology of the spectrum could be used to screen osa and normal subjects by deriving relevant statistical parameters. common spectral methods usually find and compare spectral power between two different frequency bands (drinnan et al. 2000). the study also revealed that both parametric and non parametric methods have obtained acceptable results (accuracy above 85%). however, parametric method could over-perform non-parametric method if optimum filter order and transform length are selected. the best sensitivity result of implementing this technique with burg method (n = 128, p = 14) and welch method (n = 256) is 92.5% for both. the specificity of efficiency trial test overall sensitivity 80% 75% 77.5% specificity 80% 90% 85% accuracy 80% 80% 80% table 5. results of burg method (n=64, p=14, cr=0.648) efficiency trial test overall sensitivity 80% 80% 80% specificity 80% 90% 85% accuracy 80% 83.3% 81.7% table 6. results of burg method (n=128, p=10, cr=0.72) efficiency trial test overall sensitivity 100% 85% 92.5% specificity 90% 100% 95% accuracy 96.7% 90% 93.3% table 7. results of different burg method (n=128, p=14, cr=0.63) efficiency trial test overall sensitivity 90% 85% 87.5% specificity 80% 100% 90% accuracy 86.6% 90% 88.3% table 8. results of burg method (n=128, p=20, cr=0.532) efficiency trial test overall sensitivity 95% 65% 77.5% specificity 90% 90% 90% accuracy 90% 73.3% 81.7% table 9. results of burg method (n=256, p=14, cr=0.6) parameter yuler pcov pmusic pmcov peig welch burg sensitivity 80% 85% 85% 86.7% 90% 87.5% 92.5 specificity 90% 90% 80% 86.7% 90% 85% 95 accuracy 83.3% 86.7% 83.3% 86.7% 90% 86.7 93.3 cr 0.6 0.62 0.468 0.632 0.588 0.378 0.63 table 10. results of parametric and non-parametric methods, n=128, p=14 method trial test overall accuracy new-welch-128 28/30 24/30 83.3% new-welch-256 28/30 26/30 90% new-burg-128 29/30 27/30 93.3% mietus 26/30 28/30 90% drinnan 27/30 27/30 90% table 11. screening results of different methods 8 the journal of engineering research vol.3, no.1 (2006) 1-9 both algorithms are 95% and 85% respectively, while the total accuracy is about 93.33% with burg algorithm and about 90% with welch spectral estimation method. the most important advantage of this technique is that a single-parameter is used for screening. we would like to pointout that this is a screening technique and is not purely diagnostic. it helps to short-identify patients who need full polysomnography especially in a very busy center like the one we have at sultan qaboos university hospital in oman. figure 7. results of all data using welch method (n=128) figure 8. results of all data using burg method (n=128) 9 the journal of engineering research vol.3, no.1 (2006) 1-9 references aasm task force report, 1999, "sleep -related breathing disorders in adults, recommendations for syndrome definition and m easurement techniques in clinical research," sleep, vol. 22, pp. 667-689. bigger, t., 2000, "overview of rr variability, heart rhythm instruments," inc, 2002. available: http://www.nervexpress.co m. boyer, s. and kapur, v., 2002, "obstructive sleep apnea: its r elevance in t he care of diabetic patients, " clinical diabetes , vol. 20( 3), pp. 126132. drinnan, m., aller, j., langley, p. and murray, a., 2000, "detection of sleep apnea from frequency analysis of heart rate variability," computers in cardiology , vol. 27, pp. 259 -262. friesen, g.m., jannett, t.c., jadallah, m.a., yales, s.l., quint, s. r. and nagle, h.t., 1990, "a comparison of the n oise sensitivity of n ine qrs detection algorithms,” ieee trans. biomed. eng. , vol. 37(1), pp. 85 -97. hirsch, h.l.,1992, "statistical signal charact erization," artech house, norwood, ma, usa. kay, s.m., 1988, "modern spectral estimation," englewood clis, nj : prentice hall. khoo, m., kim, t. and berry, r., 1999, "spectra l indices of cardiac autonomic function in obstructive sleep apnea," sleep, vol. 22(4), pp. 443-451. marple, s.l., 1987, "digital spectral analysis," englewood clis, nj : prentice hall. mietus, j., peng, c., ivanov, p. and goldberger, a.i., 2000, "detection of o bstructive sleep apnea from cardiac interbeat interval time series," computers in cardiology , vol. 27, pp.753 -756. netzer, n., "overnight p ulse oximetry for sleepdisordered breathing in adults," chest, aug., 2001, pp. 625 -633. pack, a.i., 1993, "simplifying the diagnosis of obstructive sleep apnea, " annals of internal medicine, vol. 119( 6), pp. 528 -529. penzel, t., 2000, "the apnea ecg d atabase," computers in cardiology , vol. 27, pp. 255 -258. penzel, t., mcnames, j., de chazal, p., raymond, b., murray, a. and moody, g., 2002, "systematic c omparison of different algorithms for apnoea detection based on electrocardiogram r ecordings," med. biol. eng. comput., vol. 40, pp. 402 -407. physionet: an nih/ncrr research resource for complex physiologi c signals, apnea ecg data files online. available: physionet.org/physiobank/database/apnea -ecg physionet: an nih/ncrr research resource for complex physiological signals, qrs detection and waveform boundary recognition using ecgpuwave, computer program. avail able: http//www.physionet.org/physiotools /ecgpuwave proakis, j.g. and manolakis, d.g., 2000, "digital signal processing: principles , algorithms, and applications, " 3rd ed, prentice -hall of india . rangayyan, r.m., 2000, "biomedical signal analysis: a case -study approach," ieee press. rodenstein, d.o., doom, g., thomas, y., luestro, g., stanescu, d.c., culee, c. and aubert tulkens, g., 1990, "pharyngeal s hape and dimensions in h ealthy subjects, snores, and patients with o bstructive sleep apnea," thorax, vol. 45, pp. 722-727. stevenson, j.e., (2003), "diagnosis of sleep apnea," wisconsin medical j . vol. 102( 1), pp. 25-27. task force of the european society of cardiology and the north american society of pacing and electrophysiology, 1996, "heart r ate variability; standard of measurements, physiologic al interpretation, and clinical use," circulation, vol. 93, pp. 1043-1065. tsi, w.h., flemons, w.w., whitelaw, w.a. and remmers, j.e., 1999, "a c omparison of apneahypoapnea indices derived from different definitions of hypoapnea, " am j. respir. crit. care med., vol. 159, pp. 43 -48. the journal of engineering research (tjer) vol. 15, no. 1 (2018) 26-41 doi: 10.24200/tjer.vol15iss1pp26-41 evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria a.w. salamia, r.b. mudashiru*, b and s.o. bilewuc adepartment of water resources and environmental engineering, university of ilorin, ilorin, nigeria. bdepartment of civil engineering, federal polytechnic offa, offa, nigeria. cdepartment of water resources and environmental engineering, university of ilorin, ilorin, nigeria. received 28 january 2016; accepted 2 february 2017 abstract: this paper presents the evaluation of methods of determination of peak runoff for an ungauged catchment. the present study made use of geographical information system (gis) for the morphometric analysis of four sub-catchments within jere sub-basin in gurara basin, nigeria. various morphometric parameters were computed and analyzed as sub-basin characteristics are important factors in determining the runoff. three methods of peak runoff estimation were used to determine peak runoff of the study area for four return periods. the results of the peak runoff estimation based on rational method for the four sub-catchments for 25-year, 50-year, 75-year and 100-year return periods varied between 224m3/s and 1036m3/s, while the results for the scs method varied between 105.3m3/s and 162.05m3/s and the results from the cypress creek method varied between 22.8m3/s and 86.55m3/s. the results obtained from the three methods showed that there are significant differences in the peak runoff results and also there are significant differences in the peak runoff results for the different return periods. keywords: digital elevation model; geographic information system; morphometric parameters; return periods; runoff; sub-basin; sub-catchment. يف تقييم أساليب حتديد ذروة اجلريان السطحي للمياه باستخدام خصائص مستجمعات املياه حوض جريي الفرعي وحوض نهر غورارا، مشال نيجرييا الوسطى ت, س. او. بيليووب, ر.ب. موداشريو اأ. و.سالمي قامت مستجمعات املياه غري املقفلة. يف طرق حتديد ذروة اجلريان السطحي للمياهلتقييم البحثية تقدم هذه الورقة :امللخص (مورفومرتيقياسات الشكل اخلارجي و التضاريس )تحليل ل جي اي اس()هذه الدراسة باستخدام نظام املعلومات اجلغرايف نيجرييا. ومت احتساب وحتليل العديد من العوامل بألربعة مستجمعات فرعية داخل حوض جريي الفرعي يف حوض غورارا استخدام ثالث ب قام البحث لحوض الفرعي باعتبارها عوامل مهمة يف حتديد اجلريان السطحي.لاملورفومرتية كخصائص اجلريان السطحي ذروة نتائج تقدير تراوحت و ألربع فرتات عودة. نطقة الدراسة مبطرق لتحديد ذروة اجلريان السطحي / ث 3م 224بني سنة 100سنة و 75سنة و 50سنة و 25فرتات ل جمعات الفرعية األربعستالقائم على الطريقة املنطقية للم / ث ونتائج طريقة 3 م 162,05/ ث و 3م 105,3اس سي اس تراوحت بني يقةيف حني أن نتائج طر , / ث 3م 1036و على الثالث طرقالنتائج اليت مت احلصول عليها من ال ودلت / ث. 3م 86,55/ ث و 3م 22,8السيربس كريك تراوحت بني وأيضا نطقة الدراسة مباجلريان السطحي للطرق املستخدمة يف املستجمعات الفرعية األربع ذروة يف نتائج ةق جوهريووجود فر اجلريان السطحي لفرتات العودة املختلفة.ذروة فروق ذات داللة إحصائية يف نتائج على وجود معامالت الشكل اخلارجي و التضاريس ،نظام املعلومات اجلغرايف ،: منوذج االرتقاء الرقمي الكلمات املفتاحية .املستجمع الفرعي ،احلوض الفرعي ،اجلريان السطحي فرتات العودة، ،)املورفومرتية( * corresponding author’s e-mail: bunmbello@gmail.com mailto:bunmbello@gmail.com a.w. salami, r.b. mudashiru and s.o. bilewu 27 1. introduction in nigeria and the world generally, peak flows determine the design specifications of structures such as embankments, channels, bridges and dams. peak flows also determine the capacity of the control sections of flow-through measurement systems and the collection pipes and transfer conduits of volumetric collection vessels. some estimate of peak flows must be made before the design of these systems can be completed (miller 1994). design peak flows are linked to particular return periods, such as the maximum flow in 5, 10, 25, etc., years and design specifications are a balance between economic cost and the prevention of failure of the structure (miller 1994). jere is a sub-basin of gurara river basin where the federal ministry of agriculture and water resources (fma and wr) in nigeria is constructing a dam for the purpose of inter basin water transfer to supplement the lower usama and shiroro reservoirs for municipal water supply and hydroelectric power generation, respectively (fmawr 1986) ibrahim and isiguzo 2009). the relevance of determining the peak runoff for the basin cannot be overemphasized as past events that showed gurara river had flooded at one time or another. in 1998 and 1999, gurara river flooded above the capacity of its channel taking over farm lands and houses in kutamgba, a village, at the same time when heavy rains pressured the kainji, jebba and shiroro dams located in the same middle belt region of nigeria with gurara river. catchment characteristics add a lot to runoff volume. generally, the basin size is the most important basin characteristic in determining the quantity and timing of surface runoff at the outlet. the larger the basin size, the greater potential amount of rainfall that can be captured and routed to the basin outlet. basin size primarily controls the volume of runoff past the outlet. basin shape and topography are key basin characteristics controlling the routing of runoff to the basin outlet, and primarily control the timing of the peak, and to a lesser extent, the magnitude of the peak flow. soil properties determine, to a large degree, the infiltration rate, storage, and release of the precipitation from the overburden. soils affect the amount and type of vegetation, which also influence the infiltration rate. land use and modifications to the natural surface by practices such as deforestation, mining, and farming, as well as structures such as dams, levees, bridges, channels, and pavement also can have a significant effect (carluer et al. 2004). in the present study, the basin characteristics were determined using arc gis, historical rainfall data, rainfall intensity and the curve number used was based on soil characteristics and land use details of watershed. many hydrologic methods are available for estimating peak flows from a basin, and no single method is applicable to all basins (donald and richard, 2005). estimating the volume of runoff or the peak runoff rate is expected to occur from a specific watershed in response to a specific rainfall event is a very difficult challenge. hydrologists have struggled to define the many critical relationships that define the processes and the interrelationships between these processes for years. one of the first successful attempts to quantify a watershed’s runoff response was developed by the usdascs (1972) when they published what we know today as the scs or soil-cover-complex method. a second early method was known as the rational method. the soil-cover-complex method was originally developed to calculate the volume of runoff only. in recent years, this method has been further developed to yield the peak rate of runoff as well. the rational method is a simple equation that is only able to estimate the peak runoff rate. the rational method which estimates peak flows is a simplified representation of the complicated process whereby rainfall amount and intensity, catchment conditions and size as well as human activity, determine runoff amount, but it is suitable where the consequences of the failure of structures are limited. a great deal of research has been undertaken to develop hydrological models that can predict runoff peak flows and volumes. the majority, however, are not suited to general use. sometimes they are too complex, but most frequently they are limited by the geographical localities and hydrological conditions within which the data were collected. many models are regression models and their value is difficult to assess outside their own particular circumstances (miller 1994). presented here are three methods that can be used to predict peak flows and three models that are suitable to estimate runoff volumes. they are suitable for use with a wide range of catchment sizes and conditions. these methods of estimation have certain drawbacks: they can be relatively inaccurate because they make simplifying assumptions; and they require the evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 28 availability of some primary data such as catchment physical characteristics and rainfall. however, they have been used for some time in a variety of environments with success and are based on measurements from a great number of catchments, with a wide range of physical characteristics. 2. materials and methods 2.1 description of the study area the jere is a sub-basin in the gurara basin, which lies between 9°28’9°44’n and longitudes 7°18’-7°32’e covering a total area of about 483km2 , with basin length of 354.74km and average slope of 0.0004 (as determined in this study using gis); it consists of four subcatchments and has three stream orders as shown in table 1. figure 2 shows the jere subbasin and the stream order in the sub-basin. jere sub-basin falls in kargarko local government area of kaduna state in nigeria. it has an area of 1864km2 and a population of 240,943 at 2006 census. it comprises about eleven districts which are shadalafiya, katugal, kushe, jere, dogon kurmi, kenyi, kurmin dangana, kubacha, aribi and kagarko districts. the people of kagarko are predominantly koro by tribe. the people in this region are mostly farmers. about 70% of ginger produced in kaduna state comes from this local government area. figure 1 shows the location of the study area in gurara basin and in nigeria. 2.2 data collection and analysis 2.2.1 delineating of sub-basin the drainage network of jere sub-basin was delineated using esri software arc gis 10.3. the steps followed include; setting up work environment, creating a depressionless digital elevation model, creating a flow direction grid, creating a flow accumulation grid, creating outlet point, creating pour points and then delineating the sub-basin. the parameters required for morphometric analysis are computed using arcgis software. reflection radiometer (aster) images were loaded onto geographic information systems (gis) environment. the remotely sensed data are geometrically rectified to digitize the pattern of sub-basin with the help of gis software. the drainage patterns for the jere sub-basin along with basin boundaries are digitized as a line feature giving a unique identity for each order of stream which was engendered from the dem. the arcgis approach was used for drainage generation which is more coherent and reliable when compared to a manual approach. based on the cumulative number of the upstream cells draining to each cell, flow accumulation, flow direction was defined. also, a stream network in the watershed was defined. the watershed area was evaluated by calculating the geometry of the derived watershed polygons, and length of the watershed was calculated by summing the length of the main stream channel and the distance from the top of the main channel to the watershed boundary. by summing the lengths of all stream segments in each sub-watershed, the total stream length was calculated, by using strahler’s ordering system, a stream segment with no tributaries that flows from the stream source is denoted as a first-order segment. a second order segment is created by joining two first-order segments, a third-order segment by joining two second order segments, and so on. jere sub-basin consists of four sub-catchments and has three stream orders as shown in table 1. table 2 shows the stream lengths of the subcatchments, table 3 shows the linear properties of the sub-catchments and table 4 shows the relief properties of the sub-catchments which were obtained from delineation and morphometric analysis of the sub-basin. figure 2 shows the jere sub-basin, the sub-catchments and the stream order in the sub-basin. also, the map showing the jere sub-basin slope, soil distribution and land use are presented in figs. 3, 4, and 5 respectively. 2.3 estimating of peak runoff 2.3.1 rational method the rational method of predicting a design peak runoff rate is expressed by the equation (schwab et al. 1981). 𝑄𝑝 = 0.0028𝐶𝑖𝐴 (1) where c= the coefficient of runoff (0-1). the soil in the sub-basin belongs mostly to the a.w. salami, r.b. mudashiru and s.o. bilewu 29 figure 1: location map of nigeria showing gurara basin and jere sub-basin. table 1. stream order of the sub-catchments within jere sub-basin. river subbasin orders sub catchments 1 2 3 4 jere 1st 7 4 13 7 2nd 2 1 4 1 3rd 0 0 1 0 total stream 9 5 18 8 table 2. stream length of the sub-catchments within jere sub-basin (km). sub-catchment id 1 2 3 4 total stream length in order (km) mean stream length(km) stream length ratio bifurcation ratio order 1st 16.36 10.17 22.58 16.8 66.31 2.3933 0.3864 3.875 2nd 5.18 3.04 13.2 4.20 25.62 2.6420 0.0722 8.000 3rd 1.85 1.85 2.1128 total stream length in subcatchment 21.54 13.21 38.03 21 93.78 2.6250 evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 30 figure 2. the jere sub-basin showing the stream orders and the sub-catchments. table 3. linear properties of the sub-catchments within jere sub-basin. table 4. relief properties of the sub-catchments within jere sub-basin. sub-catchment id 1 2 3 4 total basin area (km2) 124.06 50.73 197.37 111.25 483.423 basin perimeter (km) 92.07 46.46 140.07 85.88 364.479 basin length (km) 92.00 46.34 130.12 86.11 354.738 sub-catchment id 1 2 3 4 basin relief (km) 0.31 0.31 0.31 0.31 relief ratio (slope) 0.003367 0.006690 0.002382 0.003600 ruggedness number 0.05382 0.08072 0.059732 0.058517 a.w. salami, r.b. mudashiru and s.o. bilewu 31 figure 3. slope of the jere sub-basin. figure 4. soil distribution of the jere sub-basin. evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 32 dystric regosols as shown in figure 3 belonging to base of soil resources which are weakly developed from mineral soil in unconsolidated material. some regosols are used for capital intensive irrigated farming, but most common land use is low volume grazing. the soil type is particularly common in arid areas, in dry tropics and in mountain regions. some parts of the jere sub-basin contain plinthic ferric lixisols which are strongly weathered soils from unconsolidated, strongly leached, finely textured materials from region of tropical warm temperate with pronounced dry season. they are often used for perennial crops or low volume grazing and these groups of soil are prone to erosion because of their unstable surface soil structure. for the purpose of this study, the coefficient of runoff of 0.20 would be used as the soil type and land use of the study area falls within ( ruralaverage infiltration ratescultivated pasture from the table of runoff coefficient for rational formula). a = the catchment area (ha) i = rainfall intensity (mm/h) qp = peak runoff rate (m3/s) note: 1km2= 100ha, the area for each subcatchment in jere sub-basin were obtained from the dem. the rainfall intensities for return periods of 25-years, 50-years, 75-years and 100 years were obtained in (mudashiru 2015) as 78.89mm/h, 86.35mm/h, 90.68mm/h and 93.75mm/h respectively. the peak flow for various rainfall intensities were determined by adopting equation (1). the results are summarized in tables 5 for the four sub-catchments within the jere sub-basin. figure 5. land use map of the jere sub-basin. a.w. salami, r.b. mudashiru and s.o. bilewu 33 2.3.2 soil conservation service (scs) method the peak discharge, the time to peak and the lag time can be determined in accordance to (scs 1972; viessman et al. 1989; and ogunlela and kasali 2002). peak discharge: the peak discharge can be obtained through the eq. (2), (ramirez 2002). qp = 2.08a tp (2) where, qp= peak discharge (m3/s/cm) a = watershed area (km2) tp= time to peak(h) time to peak and lag time: tp = tc + 0.133tc 1.7 (3) where; tc= time of concentration (h) tr= storm duration (h) tl = lag time (h) tc = time of concentration (hr) (kirpich’s equation). 𝑡𝑐 = 0.06628 { 𝐿0.77 𝑆 0.385 } (4) where; l = length of channel (stream) in km s = slope of channel (m/m) equation 4 was adopted in the estimation of the time of concentration for each of the subcatchments. this was used to determine the values for both the peak discharge (qp) and time to peak (tp) which was applied to the dimensionless hydrograph ratios to obtain points for the unit hydrograph using equations 2 and 3. the peak discharge, time to peak and lag time were determined in accordance to scs (1972), viessman et al. (1989) and ogunlela and kasali (2002). the parameters used for the analysis are as presented in tables 3 and 4. however, the parameters obtained such as catchment length (l), time of concentration (tc), time to peak (tp), catchment slope (s), peak discharge (qp) and the catchment area (a) are presented in table 6. the estimated values for both the peak discharge qp and time to peak tp are applied to the dimensionless hydrograph ratios to obtain points for the unit hydrograph as presented in figs. 6-9. table 5. summarized peak runoff values using rational method. recurrence interval peak runoff values (m3/s) sub-catchment 1 sub-catchment 2 sub-catchment 3 sub-catchment 4 25-year 548.08 224.12 871.95 491.48 50-year 599.90 245.31 954.40 537.96 75-year 629.99 257.61 1002.26 564.94 100-year 651.32 266.33 1036.19 584.06 table 6. parameters for the generation of unit hydrograph. subcatchment l(km) a(km2) sc(slope) tc(h) tp(h) qp(m3/s/cm) 1 92.00 124.06 0.0034 19.25 2.6 20.48 2 46.34 50.73 0.0067 8.74 5.72 18.44 3 130.12 197.37 0.0024 28.74 18.81 21.82 4 86.11 111.25 0.0030 19.19 12.56 18.42 total 354.57 483.41 0.0155 75.92 39.69 79.16 evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 34 figure 6. unit hydrograph flows versus time for sub-catchment 1. figure 7. unit hydrograph flows versus time for sub-catchment 2. figure 8. unit hydrograph flows versus time for sub-catchment 3. 0.00 5.00 10.00 15.00 20.00 25.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 u n it h y d r o g r a p h f lo w s (m 3 /s /c m ) time (h) 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 u n it h y d r o g r a p h f lo w s (m 3 /s /c m ) time (h) 0.00 5.00 10.00 15.00 20.00 25.00 0.00 20.00 40.00 60.00 80.00 100.00 u n it h y d r o g r a p h f lo w s (m 3 /s /c m ) time (h) a.w. salami, r.b. mudashiru and s.o. bilewu 35 a) estimation of design storm (runoff) hydrograph the estimated synthetic unit hydrograph from scs method was used to develop the runoff hydrographs due to annual peak daily rainfall event over the sub-basin. the design runoff hydrographs for the selected rainfall of recurrence intervals of 25-years, 50-years, 75 years and 100-years were developed through hydrograph convolution. hydrograph convolution involves multiplying the unit hydrograph ordinates by incremental rainfall excess, adding and lagging in a sequence. b) determination of rainfall depth of different return periods. the rainfall depth of 25-years, 50-years, 75years and 100-years were 157.79mm, 172.71mm, 181.38mm and 187.51mm respectively were determined according to (mudashiru 2015). c) estimation of rainfall excess in the scs method, the excess rain volume, q, depends on the volume of precipitation, p, and the volume of the total storage, s, which includes both the initial abstraction, ia, and the total infiltration f. the relation between rainfall excess and total rainfall (on twenty-four hour basis) is then (mccuen and bondelid 1983). 𝑄 = (𝑃−0.2𝑆)2 (𝑃+0.8𝑆) (5) 𝑃 = 𝑃∗ 24 ∗ 𝑃𝑇 (6) where, p = accumulated rainfall (mm). pt = rainfall recurrence interval of the sub-basin (mm). p* = precipitation ratio. p*/24 = precipitation ratio dimensionless value in scs type ii for estimation of rainfall excess. s = volume of total storage (mm). note that equation (15) contains only one unknown, the storage parameter s. this parameter in (mm) can be obtained from 𝑆 = 25400 𝐶𝑁 − 254 (7) where; cn is runoff curve number cn value of 75 was adopted based on the soil type which is particularly common in arid areas, in dry tropics and in mountain regions. some parts of the jere sub-basin contain plinthic ferric lixisols which are strongly weathered soils from unconsolidated, strongly leached, finely textured materials from region of tropical warm temperate with pronounced dry season. they are often used for perennial crops or low volume grazing and these groups of soil are prone to erosion because of their unstable surface soil structure. figure 9. unit hydrograph flows versus time for sub-catchment 4. 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 u n it h y d r o g r a p h f lo w s (m 3 /s /c m ) time (h) evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 36 equations (5) – (7) were used to estimate the rainfall excess, other parameters used include the rainfall depth for different return period and curve number (cn) was selected based on soil distribution and land use of the study area described in figures 4 and 5. tables 8 11 presents the results of excess rainfall for different return period. the following are the estimated rainfall excess (r) for 25-years, 24-hr storm was 4.4599cm, 2.3946cm, 0.8815cm, 0.5881cm and 0.4725cm, 50-years; 24-hr storm was 5.2233cm, 2.6957cm, 0.9872cm, 0.6574cm and 0.5276cm, 75yr, 24-hr storm was 5.6778cm, 2.8708cm, 1.0485cm, 0.6977cm and 0.5596cm and 100-years, 24-hr storm was 6.0035cm, 2.9946cm, 1.0919cm, 0.7261cm and 0.5821cm respectively. d) hydrograph convolution (runoff hydrograph development) the discrete convolution equation allows the computation of direct runoff qn. let r=incremental rainfall excess (cm) u= unit hydrograph ordinate (m3/s/cm) the equations of the ordinates are given in the form of the equation below, generally: 𝑄𝑛 = 𝑅1𝑈𝑛 + 𝑅2𝑈𝑛−1 + 𝑅3𝑈𝑛−2 (8) the results from peak runoff hydrograph development are summarized in table 7. table 7. summarized peak runoff values using scs method. recurrence interval peak runoff values (m3/s) sub-catchment 1 sub-catchment 2 sub-catchment 3 sub-catchment 4 25-year 117.1 105.42 124.75 105.3 50-year 134.52 121.11 143.31 120.96 75-year 144.79 130.35 154.25 130.2 100-year 152.11 136.94 162.05 136.78 table 8. summarized peak runoff values using cypress creek method. recurrence interval peak runoff values (m3/s) sub-catchment 1 sub-catchment 2 sub-catchment 3 sub-catchment 4 25-year 48.03 22.80 70.72 43.86 50-year 53.38 25.34 78.60 48.74 75-year 56.53 26.83 83.24 51.62 100-year 58.78 27.90 86.55 53.68 table 9. values of peak runoff using rational, scs & cypress creek formula for sub-catchment 1. recurrence interval rational method(m3/s) scs method (m3/s) cypress creek(m3/s) 25-year 548.08 117.10 48.03 50-year 599.90 134.52 53.38 75-year 629.99 144.79 56.53 100-year 651.32 152.11 58.78 a.w. salami, r.b. mudashiru and s.o. bilewu 37 2.3.3 cypress creek formula the nature of flood flows has been found to be dependent on the physiographic and climatological characteristics of the stream drainage basin. this relationship can be expressed by an equation such as the following: (hydrology training series, module 206d). 𝑄 = 𝐴𝑥 + 𝐵𝑦 + 𝐶 𝑥 + 𝑁 𝑁 (9) where, q = peak flow for a given return frequency. a, b, c, and n = measurable characteristics of the basin, such as size, main channel gradient, land slope, etc. x, y, z, n = power functions, obtained by regression analysis. studies show that the dominant, and often the only obtainable, basin characteristic of flatland watersheds is size. in this case, the complex general equation may be reduced to the relationship, 𝑄 = 𝐶𝐴𝑥 (10) the cypress creek formula is: 𝑄 = 𝐶𝑀 5 6⁄ (11) q = average runoff rate, for the 24-hour period of the greatest runoff for a particular storm event (cfs). c = a coefficient, is mainly dependent upon the degree of protection desired. m = drainage area, square miles. i. development of c the value of c varies with runoff, which can be related to rainfall frequency for a 24-hour period, and was computed based on maximum runoff rates. c reflects frequency, rainfall excess, cover, intensity, etc. for the experimental watersheds, and it is based on the 24-hr average runoff. next, the coefficient c is evaluated in terms of rainfall excess, and this is related to storm recurrence interval. 𝑄 = 𝐶𝑀 5 6⁄ (12) based on numerous runoff events computed by the method of least squares, the regression equation for flatland watersheds is: 𝑦 = 16.39 + 14.75𝑅𝑒 (13) 𝑅𝑒 = (𝑃 − 0.2𝑆)2 (𝑃 + 0.8𝑆) (14) where, y = predicted value of the coefficient c . re = rainfall excess for an individual storm, in. s= volume of total storage (in). note: equation (14) is the same as equation (7). the value of curve number (cn) used was selected based on soil distribution and land use of the study area as described in figs. 4 and 5. 𝐶𝑁 = 75 when analyzing the peak runoff for cypress creek method, the following conversion factors were applied: 1mile = 1.609344km; 1km=0.6214mile; 1km2 = (0.6214*0.6214) mi2; 1km2 = 0.3861mi2; 1in = 25.4mm ; 1cfs = 0.02831684m3/s 2.4 graphical comparison of the peak runoff results based on the three methods the results obtained from the three methods used in estimating peak runoff for 25-year, 50year, 75-year and 100-year return period for the sub-catchments within jere sub-basin are summarized and presented in tables 9-12 while figures 10-13 are bar charts, which are plotted to show the graphical representation of the results from the three methods of peak runoff of the different return periods for the subcatchments. evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 38 the comparison is presented graphically as shown in figs. 10 to 13. table 10. values of peak runoff using rational, scs & cypress creek formula for sub-catchment 2. recurrence interval rational method(m3/s) scs method (m3/s) cypress creek(m3/s) 25-year 224.12 105.42 22.80 50-year 245.31 121.11 25.34 75-year 257.61 130.35 26.83 100-year 266.33 136.94 27.90 table 11. values of peak runoff using rational, scs & cypress creek formula for sub-catchment 3. recurrence interval rational method(m3/s) scs method (m3/s) cypress creek(m3/s) 25-year 871.95 124.75 70.72 50-year 954.40 143.31 78.60 75-year 1002.46 154.25 83.24 100-year 1036.19 162.05 86.55 table 12. values of peak runoff using rational, scs & cypress creek formula for sub-catchment 4. recurrence interval rational method(m3/s) scs method (m3/s) cypress creek(m3/s) 25-year 491.48 105.30 43.86 50-year 537.96 120.96 48.74 75-year 564.94 130.20 51.62 100-year 584.06 136.78 53.68 figure 10. comparison of rational, scs and cypress creek methods for sub-catchment 1. 0 100 200 300 400 500 600 700 25-yr, 24hr 50-yr, 24hr 75-yr,24hr 100-yr,24hr p e a k f lo w (m 3 /s ) return period rational method scs method cypress creek a.w. salami, r.b. mudashiru and s.o. bilewu 39 figure 11. comparison of rational, scs and cypress creek methods for sub-catchment 2. figure 12. comparison of rational, scs and cypress creek methods for sub-catchment 3. 0 50 100 150 200 250 300 25-yr, 24hr 50-yr, 24hr 75-yr,24hr 100-yr,24hr p e a k f lo w (m 3 /s ) return period rational method scs method cypress creek 0 200 400 600 800 1000 1200 25-yr, 24hr 50-yr, 24hr 75-yr,24hr 100-yr,24hr p e a k f lo w (m 3 /s ) return period rational method scs method cypress creek evaluation of methods of peak runoff determination using catchment characteristics for jere sub-basin, gurara river basin, north central nigeria 40 figure 13. comparison of rational, scs and cypress creek methods for sub-catchment 4. 3. results and discussions the area considered is a sub-basin within gurara basin in the north central part of nigeria with a catchment area of 483km2 and length of 355km. three methods of peak runoff estimation were considered, namely the rational, scs and cypress creek methods. from the comparison of peak runoff results obtained from the three methods, it shows that there is a significant difference in the peak runoff results obtained by the three methods. also, there are significant differences in the peak runoff results obtained for the return periods of 25-years, 50-years, 75-years and 100years. furthermore, the results show that the positive contribution to the runoff value obtained is based on the values of the morphometric parameters of the basin. 4. conclusion in this study, physiography and morphometric parameters of catchment area such basin length, basin area, basin slope, soil distribution, land use were analyzed. the parameters were used in estimating/ predicting peak runoff for return periods of 25-years, 50-years, 75-years and 100-years using the rational, scs and cypress creek methods. it can be established that catchment characteristics like area, length, slope etc. have a positive effect on the peak runoff of the sub basin. the results of the peak runoff obtained showed that the values obtained using the rational method was higher compared to the scs and cypress creek methods for the four sub-catchments. it can therefore be concluded that scs and cypress creek method can be adopted in the estimation of peak runoff in the sub-basin since the values obtained using the two methods are not far apart and are reasonable. both methods considered the excess rainfall within the sub-catchment and the curve number, which is a function of the properties of the soil and vegetation, cover the sub-catchment in the estimation of peak runoff. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references carluer n, de m, ghislain d (2004), assessment and modeling of the influence of man-made networks on the hydrology of a small watershed–implications for fast flow components, water quality and landscape management. journal of hydrology 285(1): 76–95. 0 100 200 300 400 500 600 25-yr, 24hr 50-yr, 24hr 75-yr,24hr 100-yr,24hr p e a k f lo w (m 3 /s ) return period rational method scs method cypress creek a.w. salami, r.b. mudashiru and s.o. bilewu 41 donald c, richard l (2005), comparison of peak discharge and runoff characteristic estimates from the rational method to field observations for small basins in central virginia. u.s. department of the interior and u.s. geological survey. ibrahim h, isiguzo e (2009), flood frequency analysis of gurara river catchment at jere, kaduna state, nigeria. science research essay 4(6): 636–646. mccuen r, bondelid t (1983), estimating unit hydrograph peak rate factors. journal irrigation drainage engineering asce 110(7): 887-904. miller s (1994), handbook for agrohydrology, natural resources institute, overseas development administration. mudashiru r (2015), hydrologic characterization and peak runoff prediction for jere sub-basin, gurara river basin, nigeria. master of engineering thesis submitted to the school of postgraduate studies, university of ilorin, nigeria. ogunlela a, kasali m (2002), evaluation of four methods of storm hydrograph development for an ungauged watershed. nigerian journal technology development 2: 25-34. ramirez j (2002), prediction and modeling of flood hydrology and hydraulics. chapter 11 of inland flood hazards: human, riparian and aquatic communities. edited by ellen wohl. cambridge university press, usa. schwab g, frevert r, edminister t, barnes k (1981), soil and water conservation engineering. john wiley and sons, new york 525. scs. soil conservation service (1972), national engineering handbook, section 4, hydrology. us government printing office, washington dc, 544. viessman w, knapp j, lewis g (1989), introduction to hydrology. harper and row publishers. new york 149-355. the journal of engineering research (tjer), vol. 15, no. 1 (2018) 88-101 doi: 10.24200/tjer.vol15iss1pp88-101 performance optimization of centrifugal pump for crude oil delivery s.a.i. bellarya, a. husain*, b, a. samadc and r.a. kanaia a annasaheb dange college of engineering and technology, ashta, sangli-416301, india. b department of mechanical and industrial engineering, sultan qaboos university, po box 33, al-khoudh, pc 123, muscat, sultanate of oman. c department of ocean engineering, indian institute of technology madras, chennai-600036, india. abstract: crude oil transport is an essential task in oil and gas industries, where centrifugal pumps are extensively used. the design of a centrifugal pump involves a number of independent parameters which affect the pump performance. altering some of the parameters within a realistic range improves pump performance and saves a significant amount of energy. the present research investigated the pump characteristics by modifying the number of blades and the exit blade-angles. reynolds-averaged navier-stokes equations with standard k-ε two-equation turbulence closure were used for steady and incompressible flow of crude oil through the pump. the experimental set-up was installed and the pump performance calculated numerically was compared with the experiments. the investigations showed that the number of blades and the exit blade-angles have a significant influence on the head, shaft power, and efficiency. the vortical flow structures, recirculation and reverse flow characteristics around the impeller were investigated to explain the flow dynamics of impeller and casing. a larger number of blades on the rotor showed dominant streamlined flow without any wake phenomena. the combined effect of the number of blades and exit blade angle has led to an increase in head and efficiency through the parametric optimization. keywords: centrifugal pump; exit blade-angle; number of blades; recirculation; reverse flow; vortical structures; pump efficiency. اخلام النفط توزيع يف املركزي الطرد ملضخة األمثل األداء أ ، عبد الصمد ، رفيق أبو كاناي *ب، أفظل حسني أ سيد امحد عمران بالري : يعترب نقل النفط اخلام من املهام األساسية يف صناعات النفط والغاز ، حيث تستخدم مضخات الطرد املركزي خلخ امل عخلى نطاق واسع. وينطوي تصميم مضخة الطرد املركزي عخلى عدد من املعامالت املستقخلة اليت تؤثر عخلى أداء املضخة. ويعمل أداء املضخة وتوفري كمية كبري من الطاقة. ويقوم البثث ااحالي تغيري بعض هذه املعامالت ضمن نطاق واقعي عخلى حتسني ستوكس مع -وزوايا خمارجها. و قد مت استخدام معادالت نافري الريشبدراسة خصائ املضخة عن طريق تعديل عدد ل مروره باملضخة. ومت املعيارية ذات املعادلتني لتدفق النفط اخلام بشكل ثابت وغري قابل لالنضغاط من خال 3املعادلة كيه الدراسات أن عدد الريش أظهرت. والنتائج العمخلية تركيب النظام التجرييب و مقارنة أداء املضخة بالطريقة العددية مع هلا تأثري كبري عخلى الرأس وقو الدوران والكفاء . ومت فث هياكل التدفق الدوراني وإعاد التدوير ها وخمارج زوايا اخلارجي. وأظهر وجود عدد كبري من طارلعكسي حول الدافع ملعرفة ديناميكية تدفق الدافع واالوخصائ التدفق ا . وقد أدى التأثري املشرتك لعدد حدوث وقائع غريبة يف اعقاب ذلكيعمل عخلى تدفق كبري منتظم دون الريش عخلى احملرك .األمثل لخلمعامل إىل زياد يف الرأس والكفاء من خالل االداء زواياهاالريش وخمارج اهلياكل ،التدفق العكسي ،تدويرالإعاد ،عدد الريش ،خمرج زاوية الريش ،مضخة الطرد املركزي: ملفتاحيةالكخلمات ا كفاء املضخة ،الدورانية * corresponding author’s e-mail: afzal19@squ.edu.om mailto:afzal19@squ.edu.om s.a.i. bellary, a. husain, a. samad and r.a. kanai 89 list of symbols h : head generated, m ∆h : hydraulic losses, m m : power consumed by pump, w n : impeller speed, rpm q : volume flow rate, m3/s ui : velocity of fluid, m/s z : number of blades greek symbol β : exit blade angle η : efficiency, % µ : dynamic viscosity, n.s/m2 µt : turbulent viscosity, n.s/m2 ρ : density of fluid, kg/m3 σ : slip factor subscripts act : actual h : hydraulic in : input man : manometric mech : mechanical motor : motor out : output theo : theoretical ov : overall 2 : outlet acronyms cfd : computational fluid dynamics rans : reynolds-averaged navier-stokes performance optimization of centrifugal pump for crude oil delivery 90 1. introduction transportation of crude oil is one of the major tasks of the e & p (exploration and production) companies and centrifugal pumps are extensively used for this application. these pumps transport liquids or mixtures of liquids having a wide range of viscosity and density (bellary and samad 2014; srinivasan 2008). an impeller, a major part of the pump, transfers energy to the fluid and the fluid flow is characterized by diffusion, vortices, strong swirl, separation, reverse flow and recirculation (tuzson 2000). the centrifugal pump performance is dependent upon the impeller diameter, the width of the impeller, the number of blades (z) mounted on the impeller, blade profile and inlet and the exit blade-angles (β) etc. (srinivasan 2008; bellary and samad 2016). several researchers investigated the effects of various geometric parameters on the performance of centrifugal pumps. geometric parameters should be altered carefully so that the basic design of the pump cannot be disturbed (srinivasan 2008; gulich 2010). to do so, one has to change the shape of the impeller rather than changing the key dimensions. proper design parameters and operating range selections are important to enhance the pump performance. during the last few decades, the effects of geometric parameters on the pump performance have been extensively investigated. kamimoto and matsuoka (1956) investigated the internal flow mechanism experimentally and have shown 40% improvement in hydraulic efficiency of an impeller with 30º eba. varley (1961) reported that a higher β gives a higher head for a double suction centrifugal pump, but they skipped reporting the flow mechanism in detail. aoki et al. (1987) and ohta and aoki (1996) have reported that higher β improves the performance of the pump with water as working fluid (varley 1961). murakami et al. (1980) measured flow patterns in centrifugal pumps at both design and off-design conditions and concluded that the velocity and pressure distributions of seven-blade impeller coincide, whereas for three-blade impeller the deviation was large. fard and boyaghchi (2007) and li (2011) reported similar works results on the centrifugal pump. fard et al. (2012) investigated the effect of β for oil as working fluid, both experimentally and numerically, and showed that the increase of β improves the performance. shigemitsu et al. (2011) investigated flow through a mini turbopump and the effects of β, and reported that the best efficiency point shifts towards higher values for higher β. experiments on water and viscous oils show that the z influenced the performance of oil pumps and their effects decreased as fluid viscosity increased (li 2011). houlin et al. (2010) numerically investigated the flow characteristics of a centrifugal pump with different z and compared their results with those of the experiments. an increase in z leads to an increase of pressure fluctuations on the pressure side and on the whole diffuser vane surface (feng et al. 2007). the experimental and numerical analysis show that z affects the head generation (rababa 2011; bellary et al. 2016). kim et al. (2014) performed design optimization of a centrifugal pump using cfd-based analysis and response surface method. the optimized design showed approximately 0.6% higher efficiency and 1.0 m higher head. yan et al. (2017) redesigned a low efficiency and high vibration pump using cfd analysis. the redesigned pump showed about 3% higher efficiency than the baseline pump. alemi et al. (2015) investigated volute geometry effects on the pump head, efficiency and radial force of a low-specific-speed pump. the authors concluded that the circular cross section volute casing with radial diffuser provides higher head and efficiency. more often, industries transport fluids having high viscosities. several researchers have reported the effect of viscosity on pump performance. siddique et al. (2017) reported that the high viscosity skin-friction losses decreased with the increase of surface roughness. bai et al. (2017) analyzed and compared the submersible pump head and efficiency to a different combination of number-of-blades of the impeller and diffuser vanes. korkmaz et al. (2017) performed experiments on different blade angles and concluded that the impeller with less blade angle has superior performance as compared to that of large exit blade angle. the results revealed that the less number of impeller and diffuser vanes favored the rise in head and efficiency at the design point. although many researchers reported experimental and numerical investigations of the centrifugal pump with water as working fluid, there is a scarcity of deterministic literature on centrifugal pumps for crude oil transportation. in this study, the performance enhancement of a centrifugal pump and the effects of several design parameters have been s.a.i. bellary, a. husain, a. samad and r.a. kanai 91 investigated with detailed description for complex flow behavior, velocity, and pressure distributions. an attempt has been made to understand the combined effect of the blade numbers (z) and the exit blade angle (β) for crude oil transportations. 2. experimental and numerical model 2.1. pump facility, experimental setup and test conditions the centrifugal pump used in this study to optimize the performance is shown in fig. 1. the pump has un-vanned single volute casing with circular cross section. the pump has a rotational speed of 1340 rpm (n) and a specific speed of 70 (ns). the total head is 5.05 m (h) at the design point which is having a mass flow rate of ṁdes = 2.171 kg/s. the design parameters and test conditions are presented in table 1. the experimental set-up was installed in the wave energy and fluids engineering laboratory of the department of ocean engineeringiit madras (fig. 2). two differential pressure gauges measured the pump suction and delivery pressure. a needle valve located between the pump outlet and flow meter regulated the flow and a flow meter measured the flow rate. a 3-phase dc-induction motor operated by a variable frequency drive, drove the pump at a constant speed (n=1340 rpm). the change in pressure at the suction and the delivery sides by adjusting flow control valves were recorded. a voltmeter and an ammeter gave the input power. the losses in the motor were considered to calculate the pump input power. the pump head was calculated by measuring pressures at the inlet (p1) and the outlet (p2), i.e.         g pp h  12 (1) including slip factor, the pump head is estimated by the following relation (japikse et al. 1997); 2 1 cotm man cu h g u          (2) where u is the peripheral velocity and cm is the meridional component of velocity at the exit due to slip. wiesner (1967) related z with slip factor as; 0.7 sin 1 z     (3) hydraulic efficiency is given by; theo theo theo act h h hh h h   (4) relationship between overall efficiency and hydraulic efficiency is given by; hvolmechov   (5) following relationship was used to calculate the input power to the pump by measuring input voltage and current including motor losses; motorin eim   cos3 (6) where, e and i are the applied voltage and current to the motor. the output power was determined by the formula; qhgm manout   (7) then, the pump efficiency was determined by the relationship; in out pump m m  (8) figure 1. schematic of centrifugal pump impeller and casing. performance optimization of centrifugal pump for crude oil delivery 92 (a) (b) figure 2. (a) schematic of the experimental facility, and (b) experimental setup of the pump system. table 1. specifications of the pump and, details of the experimental setup. parameter dimension shaft diameter, ds 24 mm inlet diameter, d1 63 mm inlet blade width, b1 23 mm outlet blade width, b2 10 mm inlet blade angle at shroud, βshr 11o inlet blade angle at hub, βhub 31° outlet blade angle, β 36o blade number, z 5 blade thickness, t 3 mm outlet diameter, d2 142 mm base circle diameter of volute casing, d3 147 mm volute width at base circle, b3 15 mm suction pipe diameter, dsuc 50 mm delivery pipe diameter, ddel 32 mm s.a.i. bellary, a. husain, a. samad and r.a. kanai 93 the crude oil of the experiment was obtained from oil and natural gas corporation (ongc), karaikal, india. its properties were extracted from the gas hydrate and flow assurance laboratory of the department of ocean engineeringiit madras. 2.2. numerical model in the present study, reynolds-averaged navier-stokes (rans) equations were solved for the incompressible and steady-state flow using a computational fluid dynamics (cfd) solver and k-ε turbulence closure model was used (ansys solver theory 2010). the impeller and casing geometries, having dimensions given in table 1, were constructed in bladegenv13 and catia-v5, respectively. these geometries were assembled in a design modeler. it was meshed with unstructured grids in ansys turbogrid as shown in fig. 3. table 2 shows the number of nodes, elements, boundary conditions, and the convergence criteria. figure 3. computational domain with unstructured mesh. table 2. details of flow modeling and boundary conditions. parameter dimension flow domain impeller and volute casing interface frozen rotor mesh/nature unstructured total nodes 150000 total elements 500000 fluid nature water and crude oil turbulence model k-ε inlet pressure outlet mass flow rate residual convergence value 1×10-5 time taken for single simulation 6 h mass imbalance 1×10-4 kg/s performance optimization of centrifugal pump for crude oil delivery 94 the flow was simulated at various design and off-design points. although the mesh was not resolved finely for boundary layer capturing, it provided reasonably accurate results for the overall pump performance characteristics. at the impeller leading and trailing edges, near the blade surface and in the volute casing, fine meshes were applied. a fine mesh on the blade surface captures flow behavior more accurately and helps achieving linear velocity distribution in the viscous sublayer. the rans equations can be written in tensor form as; 0   i i x u (9) 2 ( )i i j t j i j j j u up u x x x x x                        (10) here, turbulent viscosity is modeled by the kε model (turbulence model) as; 2 t k c      (11) the equations for turbulent kinetic energy and dissipation rate can be written as; t j k j j k j k k u p x x x                       (12) 1 2 ( )t j k j j j u c p c x x x k                            (13) here, j i i j j i tk x u x u x u p                  (14) the values of the closing constants are cμ=0.09, cε1=1.44, cε2=1.92, σk=1.0 and σε=1.3. the k-ε turbulence closure model uses the scalable wall-function to improve the robustness and accuracy. the scalable wall function allows solution on arbitrarily fine near wall grids, which is a significant improvement over standard wall functions (ansys solver theory 2010; husain et al. 2013). it allows consistent mesh refinement and remains accurate for a fairly wide range of y+ values. the solver employs a multigrid approach which solves the hydrodynamic equations for velocity and pressure (u, v, w, p) as a single system and uses a fully implicit discretization of the equations for a given time step. also, it uses a multigrid accelerated incomplete lower upper factorization technique for solving the discrete system of linearized equations. 3. results and discussion the grid dependency of the numerical scheme was checked before carrying out cfd analysis. the reference parameter for the grid analysis was the pump head at the design conditions. the pump head showed asymptotic behavior with an increase of node numbers as shown in fig. 4. once the number of nodes reached a value of ~150,000, it affected the head generation insignificantly. the convergence criterion for the iterative solutions was a maximum residual of 10-5. several turbulence closure models are available and their suitability depends upon cfd applications. bradshaw (1966) reported that the turbulence selection model depends upon: a) the physical nature of the problem, b) the quality of attained results and c) the computing power. in spite of this, the traditional rans analysis with k-ε or k-ω is widely used and satisfactory results are obtained. two turbulence closure models, namely k-ε and sst, were compared for their performances at design point under similar conditions in this problem. experimental results produced 5.05 m head at the best efficiency point. the head produced by using k-ε and sst turbulence models are 5.42 and 5.95m, respectively (fig. 5). the k-ε with scalable wall function produced the least error (6.8%) and was chosen for further analysis. the computations were carried out at several flow rates ie. at design and off-design conditions of water and crude oil, gave reasonably accurate results as compared to the experimental ones. at the design and off-design points, the calculated heads were overestimated by 6.8% and 4.8% as compared to those of the experimental results, respectively. qualitatively, the present results satisfy the accuracy level as reported in the literature (gupta et al. 2011; patel and ramakrishnan 2006), and the discrepancies reported by (gulich 2010). both experiments and simulations were performed with several flow rates at 1340 rpm. oil industries handle different fluids with varying viscosities. in the present investigation, two fluids, namely, water and crude oil, which have s.a.i. bellary, a. husain, a. samad and r.a. kanai 95 figure 4. grid test typically for rans analysis using k- turbulence model. figure 5. comparison of results using turbulence models with experimental data. (a) (b) figure 6. comparison of (a) pump head and (b) efficiency, against flow rate obtained from cfd and experiments for water (wa) and crude oil (co) as working fluids. performance optimization of centrifugal pump for crude oil delivery 96 the crude oil then the water. the decrease in the efficiency and head was caused by the high viscosity, which gives high disc friction losses over the impeller shroud and hub as obvious from eq. (4). these results in fig. 6 are in good agreement with the available literature of gulich (2010) and li (2000) which represents variation in efficiency (η) and head with flow rate. the pump shows the maximum efficiency (η) at the design-point, at which hydraulic losses, profile losses, and the secondary losses are minimum. on the contrary, at all other points (off-design points) shock losses and secondary losses increase, resulting into higher hydraulic losses and decrease in hydraulic efficiency, which consequently results in a decrease in overall efficiency. at the low flow rates, reverse flow and inlet recirculation are prominent and they increase with a further reduction in flow rate. the results for water as working fluid showed higher values of head and efficiency than that of crude oil (fig. 6). the flow characteristics of crude oil as a fluid medium were investigated in detail through velocity, vorticity and pressure distribution on several x-y planes. fig. 7 shows vorticity distribution on x-y planes in the axial direction (z-direction) in the middle of the impeller at z = 30mm. in general, higher vorticity distributions are present near the inner and outer rings of the impeller, though smaller z exhibited higher vorticity towards outer rings. with the increase of z higher vorticity regions were developed near the inner ring due to the increasing boundary layer effects. an increase in β reduced the vorticity distribution near the inner ring as shown in fig. 8. higher fluid velocity regions appeared towards the impeller outer ring because of the higher radial distance as shown in fig. 9. the velocity increases suddenly as the fluid moves into the volute casing. while moving into the volute casing, the kinetic energy changed into pressure energy, which consequently increased static pressure of the fluid as shown in fig. 10. figure 11 represents the cfd characteristics of the pump head and efficiency with the change of z and β at the designed flow rate. it is clearly reflected that the pump head decreased with increasing volume flow rate due to a decrease in liquid pressure drop. in addition, the pump head increased with increasing z. this can be explained by the fact that when the volume flow rate was kept constant, an increase in z leads to decrease in liquid pressure drop, which consequently increases the pump head. also, impeller outlet velocity was increased and the velocity distribution became smoother with an increase in z and converted into highpressure energy through impeller-volute. the effective area of interest at impeller-volute intersection became more obvious which improved the discharge capability of the pump and consequently higher head was generated. with the increase in z, the flow within the impeller becomes more uniform and reduces the hydraulic losses due to mixing fluid particles. the uniform flow suppresses jet growth and wake phenomena. in addition, smooth pressure and velocity distribution were observed near the impeller exit and volute inlet as shown in fig. 9 and 10. hence, the pump efficiency was increased. the impeller with large β resulted in an increase of head generation. it is due to increase in absolute flow velocity at exit (c2) and corresponding peripheral velocity (cu2) ie. dynamic part of the head. eq. (2) clearly, indicates the impact of the two design parameters cm/u2 and β2 on the head. an increase in β2 reduced the cm/u2 and as a result, the manometric head increased. additionally, the influence of slip factor has to be considered. slip factor value decreases with an increase in exit blade angle and partly curtails the effect of β (bacharoudis et al. 2008). as a total effect, any increase in β results in an increase of head generation. the change in slip factor value affects the shape of the h-q curve and more significantly for β greater than 25o (bacharoudis et al. 2008). it is evident from eq. (3) that the higher z facilitates an increase in slip factor and consequently head augmentation. the hydraulic losses at design point differ by a small amount from off-design conditions; a measure of hydraulic efficiency is directly linked with hydraulic losses. the overall efficiency decreases due to a decrease of hydraulic efficiency. moreover, the pump brake horsepower increases due to the augmentation in β and z. this is due to the increase in the requested shaft torque associated with the increased β (djerroud et al. 2011). table 3. properties of water and crude oil. fluid viscosity [n.s/m2] density [kg/m3] water 0.001 997 crude-oil 0.0025 823 s.a.i. bellary, a. husain, a. samad and r.a. kanai 97 (a) (b) (c) (d) figure 7. vorticity distribution in z-direction on x-y plane at z = 30 mm from inlet for exit blade angle β = 50o and number of blades (a) z = 4, (b) z = 6, (c) z = 8 and (d) z = 10 for pumping crude oil. (a) (b) (c) (d) fig.ure 8. vorticity distribution in z-direction on x-y plane at z = 30 mm from inlet for number of blades z= 8 and exit blade angle (a) β = 20o, (b) β = 30o, (c) β = 40o and (d) β = 50o for pumping crude oil. performance optimization of centrifugal pump for crude oil delivery 98 (a) (b) (c) (d) figure 9. velocity distribution in z-direction on x-y plane at z = 50 mm from inlet for exit blade angle β = 50o and number of blades (a) z = 4, (b) z = 6, (c) z = 8 and (d) z = 10 for pumping crude oil. (a) (b) (c) (d) figure 10. pressure distribution on x-y plane at z = 50 mm from inlet for exit blade angle β = 50o and number of blades (a) z = 4, (b) z = 6, (c) z = 8 and (d) z = 10 for pumping crude oil. s.a.i. bellary, a. husain, a. samad and r.a. kanai 99 the effect of β on the pump performance used for crude oil was explored through cfd for the change of flow rate as shown in fig. 12. once the z (which is 8) was frozen, the β was altered from 30°-50°. similar trends were observed in which an increase in β has augmented the head generation significantly. any increase in β always leads to broader outlet cross sections (li 2011; djerroud et al. 2011) which results in a fall of fluid pressure loss and consequently the head is augmented. it is observed that any increase in β has a substantial influence on the hydraulic efficiency, which increased continuously until the design point is reached. at the off-design point, an increase in β does not increase much of the hydraulic efficiency, instead, decreases due to several losses (srinivasan 2008). (a) (b) (c) figure 11. cfd results on the effects of number of blades and exit blade angle on the pump performance for crude oil at design point flow rate: (a) head vs z, (b) efficiency vs z and (c) efficiency vs head. figure 12. cfd characteristic of head and efficiency with a change of flow rate for crude oil. performance optimization of centrifugal pump for crude oil delivery 100 4. conclusion the centrifugal pump performance investigation and optimization has been carried out for water and crude oil as working fluids at the design and off-design flow rates by experimental and numerical analysis techniques. several combinations of exit bladeangles (β) and number of blades (z) were investigated for the parametric optimization. on the basis of the present study, the following conclusions can be derived; • the increase in z has a significant influence on the head and efficiency of the pump for both water and crude oil. the distribution of vortices was more uniform with less magnitude near the impeller inlet and its distribution increased randomly with high magnitude near impeller exit for high z. with less z the vortices became prominent and more obvious near the impeller outlet. • due to non-mixing of fluid particles at higher z, jet wake phenomenon is absent and more streamlined flow was achieved which has contributed in head augmentation. • the β has a considerable influence on the head and pump efficiency for different viscosity liquids. an increase in β increases the pump efficiency until it reached a maximum rate at the design point. at off-design points, an increase in β decreases its efficiency because of increased hydraulic losses. higher viscosity liquid or crude oil has a lower head generation at the same β because of disc friction, shock and secondary losses. • combined effect of increase in β and z increases head generation with negligible increase in hydraulic efficiency. conflict of interest the authors declare no conflicts of interest. funding the authors would like to acknowledge indian institute of technology madras for the nfsc grant (grant code: oec/1011/529/nfsc/abdu). acknowledgment the authors acknowledge the support of manager of oil and natural gas corporation (ongc)-karaikal for conducting this research the facilities received from gas hydrate and flow assurance lab in the department of ocean egineeringiit madras to determine the properties of crude oil. also, the authors acknowledge the support of sultan qaboos university for conducting this research. references alemi h, nourbakhsh sa, raisee m, najafi af (2015), effects of volute curvature on performance of a low specific-speed centrifugal pump at design and off-design conditions. journal of turbomachinery 137:041009-1-10. ansys13.0. (2010) ansys inc. aoki k, ohta h, and nakayama y (1987), study on centrifugal pump for high viscosity liquids (the 1st report, effect of impeller output angle and number of blades on the pump performance of closed type impellers). proceedings of the school of engineering, tokai university 27(2):151-158. bacharoudis ec, filios ae, mentzos md, margaris dp (2008), parametric study of a centrifugal pump impeller by varying the outlet blade angle. the open mechanical engineering journal 2:75-83. bai y, kong f, xia b, liu y (2017), effect of blade number matching of impeller and diffuser in high-speed rescue pump. advances in mechanical engineering 9(5):1-12. bellary sai, samad a (2014), numerical analysis of centrifugal impeller for different viscous liquids. international journal of fluid machinery and system 8(1):36-45. bellary sai, samad a (2016), pumping crude oil by centrifugal impeller having different blade angles and surface roughness. journal of petroleum and exploration and production technology 6(1):117-127. bellary sai, adhav r, samad a, chon bh, kenyery f (2016), applications of computational fluid dynamics and surrogate-coupled evolutionary computing to enhance centrifugal pump performance. engineering applications of computational fluid mechanics 10(1):172-182. bradshaw p (1996), turbulence modeling with application to turbomachinery. progress in aerospace sciences 32(6): 575-624. djerroud m, ngoma gd, ghie w (2011), numerical identification of key design parameters enhancing the centrifugal pump s.a.i. bellary, a. husain, a. samad and r.a. kanai 101 performance: impeller, impeller-volute, and impeller-diffuser. isrn mechanical engineering, article id 794341, 16 pages, doi:10.5402/2011/794341. fard mhs, boyaghchi fa (2007), studies on the influence of various blade outlet angles in a centrifugal pump when handling viscous fluids, american journal of applied sciences 4(9):718-724. fard mhs, tahani m, ehghaghi mb, fallahian ma, beglari m (2012), numerical study of the effects of some geometric characteristics of a centrifugal pump impeller that pumps a viscous fluid. computers and fluids 60:61-70. feng j, benra fk, dohmen hj (2007), numerical investigation on pressure fluctuations for different configurations of vaned diffuser pumps. international journal of rotating machinery, article id 34752, 10pages. doi:10.1155/2007/34752. gulich jf (2010), centrifugal pumps, 2nd edition, springer publications, berlin. gupta m, kumar s, kumar a (2011), numerical study of pressure and velocity distribution analysis of centrifugal pump. international journal of thermal technologies 1(1):117-121. houlin l, yong w, shouqi y, minggao t, kai w (2010), effects of blade number on characteristics of centrifugal pumps. chinese journal of mechanical engineering 23: 1-6. husain a, kim sm, kim jh, kim ky (2013), thermal performance analysis and optimization of microjet cooling of highpower light-emitting diodes. journal of thermophysics and heat transfer 27(2): 235245. japikse d, marscher wd, furst rb (1997), centrifugal pump design and performance concepts eti, inc., vermont, usa. kamimoto g, matsuoka y (1956), on the flow in the impeller of centrifugal type hydraulic machinery (the 2nd report). transactions of the jsme, series 3, 22(113): 5. kim j-h, lee h-c, kim j-h, kim s, yoon j-y, choi y-s (2015), design techniques to improve the performance of a centrifugal pump using cfd. journal of mechanical science and technology 29(1): 215-225. korkmaz, e, golcu, m, and kurbanoglu, c (2017), effects of blade discharge angle, blade number and splitter blade length on deep well pump performance. journal of applied fluid mechanics, 10(2): 529-540. li wg (2000), effects of viscosity on the performance of a centrifugal oil pump and the flow pattern in the impeller. international journal of heat and fluid flow 21(2): 271-275. li wg (2002), the influence of number of blades on the performance of centrifugal oil pumps. world pumps 427: 32-35. li wg (2011), blade exit angle effects on performance of a standard industrial centrifugal oil pump. journal of applied fluid mechanics 4(2): 105-119. murakami m, kikuyama k, asakura e (1980), velocity and pressure distributions in the impeller passages of centrifugal pump. asme j. of fluids eng. 102: 420-426. ohta h, aoki k (1996), effect of impeller angle on performance and internal flow of centrifugal pump for high viscosity liquids. proceedings of the school of engineering, tokai university 36(1): 159-168. patel k, and ramakrishnan n (2006), cfd analysis of mixed flow pump. international ansys conference proceedings. rababa ks (2011), the effect of blades number and shape on the operating characteristics of groundwater centrifugal pumps. european journal of scientific research 52(2): 243-251. siddique mh, samad a, husain a (2017), combined effects of viscosity and surface roughness on electric submersible pump performance. journal of power and energy 0(0): 1-14. srinivasan km (2008), rotodynamic pumps. new age international (p) ltd., new delhi. shigemitsu t, fukutomi j, nasada r, kaji k (2011), the effect of blade outlet angle on performance and internal flow condition of mini turbo-pump. journal of thermal science 20(1): 32-38. tuzson (2000), centrifugal pump design, john wiley and sons, inc., 1st edition, new york. 5-59. varley fa (1961), effects of impeller design and surface roughness on the performance of centrifugal pumps. proc. instn. mech. engrs, 175(21): 955-969. wiesner fj (1967), a review of slip factors for centrifugal impellers. asme j. eng. power 89: 558–566. yan p, chu n, wu d, cao l, yang s, wu p (2017), computational fluid dynamicsbased pump redesign to improve efficiency and decrease unsteady radial forces. journal of fluids engineering 139: 011101-111. the journal of engineering research (tjer), vol. 15, no. 1 (2018) 01-13 doi: 10.24200/tjer.vol15iss1pp1-13 effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system m.h. albadi*, f.n. al farsi, n. hosseinzadeh, and a. h. al badi department of electrical and computer engineering, sultan qaboos university, p.o. box 33, muscat 132, sultanate of oman. received 18 september 2016; accepted 26 january 2017 abstract: economic dispatch is an important optimization problem in power system planning. this article presents an overview of the economic dispatch formulation, its objective, loss coefficient determination, and a case study. in the case study, different scenarios of the economic dispatch in the main interconnected system (mis) of oman were considered to highlight the importance of considering losses in the optimization process. keywords: economic dispatch; loss coefficients; power systems; main interconnected system. دراسة حالة الشبكة الرئيسة -لألمحال االقتصادي التوزيع الفاقد على احتساب تأثري املرتبطة يف سلطنة عمان ،ف. ن. الفارسي، ن. حسني زاده، أ. ح. البادي*لبادي ا .م.ح املقالة على تقديم حملة يعد التوزيع االقتصادي لألمحال هواملشكلة االهم يف ختطيط نظام الكهرباء. تعمل هذه :امللخص عامة عن صيغة االتوزيع االقتصادي لألمحال وحتديد هدفها و معامل الفقدان و دراسة حالة. و يتم اخذ االعتبار يف دراسة احلالة هذه، سيناريوهات خمتلفة للتوزيع االقتصادي لالمحال يف النظام الرئيسي املرتابط يف عمان و ذلك لتسليط الضوء احتساب الفواقد املرتتبة على عملية التحسني.على أهمية .الشبكة الرئيسة ،أنظمة الطاقة ،معامالت الفقدان،: التوزيع االقتصادي املفتاحية الكلمات * corresponding author’s e-mail: mbadi@squ.edu.om m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 2 symbols and abbrevations ipp : independent power producer lfc : load frequency control mis : main interconnected system of oman oetc : oman electricity transmission company opwp : oman power and water procurement company saoc : closed joint stock company 𝑎𝑗 , 𝑏𝑗 , & 𝑐𝑗 : coefficients of the quadratic cost function 𝐵, 𝐵0, & 𝐵00 : coefficients of the b-loss formula 𝐹𝑙 𝑡 : apparent power flowing through transmission line l during the interval t 𝐹𝑙 𝑚𝑎𝑥 : upper limit of the power flow along line l 𝐹𝑐𝑜𝑠𝑡 : total operating cost 𝐹𝑗 : operating cost of generation unit j 𝑁𝑔 : total number of generation units in the power system 𝑃𝐷 : total power demand 𝑃𝑗 : output power of generation unit j 𝑃𝑗 𝑚𝑖𝑛 : minimum output limit of unit j 𝑃𝑗 𝑚𝑎𝑥 : maximum output limit of unit j 𝑃𝐿 : total power losses 𝑆𝑗 𝑡 : spinning reserve contribution of unit j during time interval t 𝑆𝑅𝑡 : system spinning reserve requirement for interval t effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 5 1. introduction power systems have different types of generation facilities and load profiles. to maintain the stable operation of power systems, the aggregated output of online generation units and the total demand must be balanced in real time. when the net generation is higher than the net demand, the system frequency increases and vice versa (zhu 2015). the total loads in power systems experience different chronological variations: seasonally, on weekends and weekdays, and within each day (during peak and off-peak hours) (albadi and el-saadany 2011). variations in demand occur even on a second-by-second basis. to cope with these chronological variations, the outputs of the different online generation units experience cycles of operation according to the forecasted net load. power systems have different operational planning time frames: unit commitment (weeks and days) (padhy 2004), economic dispatch (hours) (gaing 2003), and frequency regulation (minutes and seconds) (kumar and kothari 2005). unit commitment is the optimum selection of online dispatchable generation units to satisfy different operating constraints. once online generation units are selected, economic dispatch is used to schedule the output of each unit to follow hourly load variations. it is worth mentioning that different utilities conduct economic dispatch at different time intervals. some utilities have an hourly dispatch while others consider shorter dispatch intervals. during real-time operation, the mismatch between demand and scheduled generation is covered using regulation services. these include turbine-governor control and load frequency control (lfc). figure 1 shows a system that consists of ng number of generating units serving an aggregate electrical load. since the combined capacity of these units is normally higher than that of the load, there exists an infinite number of output combinations to satisfy the load. the objective of the economic dispatch is to determine the output of each generation unit so that the total operating cost is minimized while relevant constraints are satisfied (zhu 2015). other objectives of the economic dispatch include the minimization of emissions, the maximization of profit by reducing the total cost, and the maintenance of system stability and security. if the transmission network of the power system under study is considered, the total demand that the generation units must satisfy includes transmission losses (george 1943). the value of transmission losses is a function of the economic dispatch decision variables, which can be determined using power-flow techniques (saadat 1999). however, transmission losses can be included in the optimization procedure using different methods. one of these methods is the loss coefficient technique. the main contributions of this article includes modeling and simulating the main interconnected system (mis) of oman for economic dispatch purposes, calculating the loss coefficients of the system, and including losses in the economic dispatch of the mis. following this introduction, the article presents a review of the economic dispatch formulation in section ii. system data are presented in section iii, and simulation results are discussed in section iv. finally, the conclusions are presented in section v. 2. economic dispatch formulation 2.1 mathematical formulation of economic dispatch the objective function of an economic dispatch problem is to minimize the total operating cost (𝐹𝑐𝑜𝑠𝑡 ), which is the summation of individual generation units’ operating/fuel costs (𝐹𝑗 (𝑃𝑗 )) (zhu 2015). 𝐹𝑐𝑜𝑠𝑡 = ∑ 𝐹𝑗 (𝑃𝑗 ) 𝑁𝑔 𝑗=1 (1) the operating cost of each generation unit is given as a function of its output power (𝑃𝑗 ) and is modeled as a quadratic function. 𝐹𝑗 (𝑃𝑗 ) = 𝑎𝑗 + 𝑏𝑗 𝑃𝑗 + 𝑐𝑗 𝑃𝑗 2 (2) where 𝑎𝑗 , 𝑏𝑗 , and 𝑐𝑗 represent the cost coefficients of the jth generating unit, 𝑃𝑗 represents the real output of the jth generating g boiler/burner turbine generator fuel g g 1 2 p1 p2 pn pload . . . . ng figure 1. generating units serving an electrical load. 3 m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 4 unit (in mw), and 𝑁𝑔 is the total number of generation units in the power system under study. these cost coefficients can be obtained from the heat rate data of generation units at different operating points and the price of fuel. the optimization problem has the following constraints:  power balance constraints ∑ 𝑃𝑗 𝑁𝑔 𝑗=1 = 𝑃𝐷 + 𝑃𝐿 (3) where 𝑃𝐷 is the power demand and 𝑃𝐿 represents the power losses.  generation unit limit 𝑃𝑗 𝑚𝑖𝑛 < 𝑃𝑗 < 𝑃𝑗 𝑚𝑎𝑥 𝑗 = 1 … 𝑁𝑔 (4) where 𝑃𝑗 𝑚𝑖𝑛 and 𝑃𝑗 𝑚𝑎𝑥 are unit j minimum and maximum limits, respectively.  spinning reserve requirement ∑ 𝑆𝑗 𝑡 ≥𝑛𝑗=1 𝑆𝑅 𝑡 𝑡 = 1 … 𝑁 (5) where 𝑆𝑗 𝑡is the spinning reserve contribution of unit j during the time interval t, and 𝑆𝑅𝑡 is the system spinning reserve requirement for interval t. 𝑆𝑗 𝑡 can be expressed as 𝑃𝑗 𝑚𝑎𝑥 − 𝑃𝑗 𝑡 where 𝑃𝑗 𝑡 is unit j output during time t.  network security constraints (voltage limit constraints, e.g. +/10% in the mis (oetc 2010).  line capacities 𝐹𝑙 𝑡 ≤ 𝐹𝑙 𝑚𝑎𝑥 ; 𝑡 = 1 … 𝑁𝑇 (6) where 𝐹𝑙 𝑡 is the apparent power flowing through transmission line l during the interval t, and 𝐹𝑙 𝑚𝑎𝑥 is the upper limit of the power flow along line l. it is worth mentioning that these limits are thermal limits of the transmission lines and can be found in oetc capability statement (oetc 2014. 2.2 loss formula when transmission line distances are small, transmission losses can be ignored and the optimal dispatch of generation units can be achieved without considering the transmission system losses. however, in large interconnected systems, transmission losses play a major role in the optimal dispatch of generation units. hence, the transmission line losses of large networks need to be considered in the optimal dispatch of generation units. there are several methods by which losses become part of the dispatch decision. george was the first to develop the simplest form of the loss equation (george 1943): 𝑃𝐿 = ∑ ∑ 𝑃𝑚 𝐵𝑚𝑛 𝑃𝑛 𝐾 𝑚=1 𝐾 𝑛=1 (7) alternatively, pl = ∑ ∑ b𝑖𝑗 p𝐺𝑖 p𝐺𝑗 ng j=1 ng i=1 (8) where b𝑖𝑗 and 𝐵mn are called loss coefficients. attempts to obtain a more accurate expression of power system losses were made by adding linear terms and a constant to the original expression. these resulted in the b matrix loss formula, which was introduced in the beginning of 1950s as a practical method for the computation of losses and incremental losses (wood and wollenberg 2012). after the addition of linear terms and a constant to the original expression, the following loss formula was obtained: 𝑃𝐿 = 𝑃 𝑇 [𝐵]𝑃 + 𝐵0 𝑇 + 𝐵00 (9) where 𝑃 is the vector of all generator bus net mw, [𝐵] is the square matrix of the same dimension as p,𝐵0 is the vector of the same length as p, 𝐵00 is the constant. this equation can be rewritten as follows: 𝑃𝐿 = ∑ ∑ 𝑃𝑖 𝐵𝑖𝑗 𝑃𝑗 + ∑ 𝐵𝑖0𝑃𝑖 + 𝐵00𝑖𝑗𝑖 (10) alternatively, it can be rewritten as follows: 𝑃𝐿 = ∑ ∑ 𝑃𝑚 𝐵𝑚𝑛𝑃𝑛 𝐾 𝑚=1 𝐾 𝑛=1 + ∑ 𝑃𝑛𝐵𝑛0 𝐾 𝑛=1 + 𝐵00 (11) there are three main methods for obtaining the loss formula coefficients (george 1943; hill and stevenson 1968; yang, hosseini, and gandomi 2012). the first method is the tensor analysis method (yang, hosseini, and gandomi 2012), the second method is the 𝐴𝑗𝑛 method (hill and stevenson 1968), and the third method is the kron-kirchmayer method, which kron developed and kirchmayer adapted (abdelaziz et al. 2008; saadat 1999; wood and wollenberg 2012). kron 1951) presented the mathematical formulation of the tensor analysis method, and hill and stevenson (1968) explained the mathematical formulation of the 𝐴𝑗𝑛 method. furthermore, george (george 1943) explained kron-kirchmayer method. many researchers have used the kron-kirchmayer method to effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 5 determine the loss coefficients (dike, adinfono and ogu 2013; su and lin 2000; yang, hosseini and gandomi 2012). the same method was used to find the loss coefficients for the system under consideration, that is, the main interconnected system of oman. once the coefficients were obtained, it was possible to determine the economic dispatch. it is worth mentioning that loss coefficients are load-specific; therefore, they should be updated for different loading conditions. 3. systems data the test system (the main interconnected system of oman) data were obtained from two sources: the oman power and water procurement company saoc (opwp) (opwp 2016a) and the oman electricity transmission company saoc (oetc) (oetc 2016). 3.1 generation units the transmission system is supplied with electricity generated at eleven gas-based power stations (open cycle and closed cycle) at ghubra, rusail, sur, wadi al jizzi, manah, al kamil, barka i, barka ii, barka iii, sohar i, and sohar ii. in addition, the transmission system is connected directly to large customers with generation facilities, for instance, sohar aluminium, pdo, sohar refinery, omco, and omifco (oetc 2014). some details of the existing generation units are available in the recent issue of “opwp 7year statement 2016-2022” (opwp 2016b) and in previous issues. other details are available in oetc’s “five-year annual transmission capability statement (2014-2018)” regarding the mis system (oetc 2014). table 1 shows a summary of the eleven power stations connected to the mis of oman and their net generation capacity in 2014. the fuel cost coefficients of all the generation units were based on estimated values from the heat rate curves received from opwp. it is worth mentioning that, to preserve confidentiality, the heat rate values were indicative rather than actual. a price of $1.5/mmbtu for the fuel (natural gas) was used to obtain the fuel cost data. then, a curve-fitting technique was used to obtain the fuel cost parameters for each generation unit. an example of a single-cycle generation unit is shown in table 2 and figure 2. the same procedure was conducted for all generation units in the system. 3.2 transmission system the oetc owns and operates the mis’s transmission network. moreover, the oetc is responsible for balancing generation and demand at all times of the day for the economic dispatch of power. it accomplishes this role through the oetc load dispatch centre (ldc), which is located in al mawaleh, muscat. currently, the system has three voltage levels: 132 kv, 200 kv, and 400 kv. the transmission system data are available in the oetc’s 5-year annual capability statements, for instance, the 2014-2018 statement (oetc 2014). a summary of the transmission system in 2014 is shown in table 3. the oetc transmission system is also interconnected with the sohar aluminium system at 220 kv and the pdo transmission network at 132 kv via a single-circuit overhead line. the available data include the length of every line (km) between the substations, the resistance (ω/km), the inductive reactance (ω/km), and the capacitive susceptance (µs/km). the power-flow model of the transmission system was built in matlab. details of the mis transmission system data are given in the appendix (tables a1 and a2). 4. simulation results a case study of the main interconnected system of oman was conducted, and it included losses within the dispatch decision. it is worth mentioning that this simulation was conducted to determine the optimal dispatch for the peakhour load of 2014. the lambda (λ) iteration method, based on lagrange relaxation, is used to solve this optimization problem. this method starts by assuming an initial value for the system incremental fuel cost (λ), and calculating output power of all generation units (pj) as a function of λ. then, the value of λ is updated based on newton-raphson method considering the power balance mismatch. the process continues until the prescribed tolerance is reached (saadat 1999). the load data for all buses were obtained from the oetc annual capability statement (oetc 2014). the details of the load are given in the appendix (table a3). the matlab software package was used in this study. three scenarios were considered in the economic dispatch. m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 6 table 1. power stations connected to the mis of oman. power plant net power generatio n (mw) remarks al kamil 282 al kamil is an ipp and was commissioned in 2002. the station comprises three ge pg9171e gas turbines (site rating: 94.1 mw) operating in open cycle manah 273 the united power company owns manah. when commissioned in 1996, it was the first ipp to be built in oman. the station comprises three ge pg6541b gas turbines (with each site rated at 28.8 mw) and two ge pg9171e gas turbines (ratings of 93.5 mw) wadi al jizzi 324 wadi al jizzi power station comprises eleven gas turbines (with gross site ratings in the range of 19.2-32.5 mw) and all operate in open cycle. the units were installed progressively from 1982. sohar i 605 sohar i power plant is an ipp and was commissioned in 2007. there are three gas turbines, each with a 138.7-mw capacity, and a steam turbine with a 220-mw capacity. all units have been operational since 2007. sohar ii 745 sohar ii power plant is an ipp and was commissioned in 2012. there are two gas turbines, each with a 247.5-mw capacity, and a steam turbine with a 249.9-mw capacity. they have all been operational since 2013. barka i 434 aes (acwa now) commissioned barka 1 in 2003. it was developed as an independent water & power plant (iwpp). the power plant comprises two ansaldo v94.2 gas turbines (manufactured under licence from siemens) and a steam turbine operating in combined cycle. barka ii 681 barka ii is an independent water & power producer (iwpp) and was commissioned in 2009. there are three gas turbines, each with a 130-mw capacity, and two steam turbines, each with a 161-mw capacity. barka iii 745 barka iii is an ipp and was commissioned in 2012. there are two gas turbines, each with a 247.5-mw capacity and a steam turbine with a 249.9mw capacity. they have all been operational since 2013 sur 2000 sur is an ipp and was commissioned in 2014. there are five gas turbines, each with a 244.4-mw capacity, and three steam turbines with capacities ranging between 156.8 and 310.7 mw. rusail 687 the power station has eight frame 9e gas turbines installed and operating in open cycle. the gross site rating of the units varies from 81.5 mw to 95.9 mw. the units were installed progressively between 1984 and 2000. ghubra 406 ghubra power plant has 2 steam turbines (each with a 29.2-mw rating) and 12 gas turbines (with ratings ranging from 16-88.6 mw). table 2. example of estimated cost curve data for a 90-mw single-cycle generation unit. operating point from rating heat rate (mbtu/kwh) fuel consumption (mmbtu/hr) fuel cost ($/hr) 30 % 16,467 395.5 593.20 65 % 10,784 630.8 946.26 100 % 10,431 938.8 1408.2 figure 2. estimated fuel cost parameters of a single-cycle generation unit. effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 7 in scenario 1, losses were considered and calculated using b-coefficients in the kronkirchmayer method.  scenario 2 involved simulating the current practice in the mis, where losses are considered as a fixed load and are not included in the optimization procedure.  scenario 3 involved simulating the system performance when network losses were not considered at all in the dispatch procedure. figures 4 and 5 show the output of power generation for different units in the rusail, ghubra, al kamil, manah, and barka i power plants. figure 4 shows the optimal output of power generation units for the sohar i, sohar ii, barka ii, barka iii, and sur power plants in the three scenarios. there are some differences between the outputs of the generation units in the three scenarios. it is clear that the total dispatched power when the network losses are considered as a fixed load (scenario 2) is higher than the total dispatched power associated with the two other scenarios in most of the power plants. it is worth mentioning that considering losses in the optimization process (scenario 1) requires slightly more execution time than other scenarios. a comparison of the number of iterations and execution time for the three scenarios are presented in table 4. figure 5 shows the total dispatched power in mw for the three scenarios. it is clear that the overall power generation output when one considers the losses as a fixed load (scenario 2) is higher than the optimal dispatch power generation when one considers the network loss using b coefficients (scenario 1) by 48 mw. this result demonstrates that transmission losses should be considered in the economic dispatch process in the mis. figure 6 shows the overall fuel cost comparison of scenarios 1 and 2. the overall cost of power generation when one considers the losses using b coefficients (scenario 1) is lower than the overall cost of the optimal dispatch of power generation when one considers the losses as a fixed load (scenario 2) by $645.62/hr. however, as was mentioned, the generation units’ cost data were merely indicative values, not their actual values. figure 7 presents a comparison of the results of the current dispatch practice and those of the proposed dispatch practice that considers the network losses for each power station in the mis. table 3. oetc transmission system assets. asset size quantity/units overhead transmission lines 220 kv 1,454.4 circuit-km 132 kv 3,051.06 circuit-km underground cables 220 kv 61.6 circuit-km 132 kv 98.681 circuit-km transformer capacity 220/132 8,630 mva 220/33 320 mva 132/33 10,541 mva 132/11 150 mva interconnection grid stations 220 five grid stations 220/132 two 220/33 one 220/132/33 seven 132/33 forty 132/11 one m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 8 figure 3. dispatched power for the three scenarios (rusail, ghubra, al kamil, manah, and barka i power plants). figure 4. dispatched power for the three scenarios (sohar i, sohar ii, barka ii, barka iii, and sur power plants). figure 5. comparison between the three scenarios’ total dispatched power values. 0 20 40 60 80 r u sa il g t 1 r u sa il g t 2 r u sa il g t 3 r u sa il g t 4 r u sa il g t 5 r u sa il g t 6 r u sa il g t 7 r u sa il g t 8 b a rk a i g t 1 b a rk a i g t 2 m a n a h g t 1 m a n a h g t 2 m a n a h g t 3 m a n a h g t 4 m a n a h g t 5 a l k a m il g t 1 a l k a m il g t 2 a l k a m il g t 3 g h u b ra s t 5 g h u b ra s t 6 g h u b ra g t 1 2 g h u b ra g t 1 3 p o w e r o u tp u t (m w ) scenario 1 scenario 3 scenario 2 50 100 150 200 250 p o w e r o u tp u t (m w ) scenario 1 scenario 3 scenario 2 4 4.2 4.4 4.6 t o ta l d is p a tc h e d p o w e r (m w ) t h o u sa n d s scenario 3 scenario 1 scenario 2 effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 9 figure 6. comparison between the overall costs associated with scenario 1 and scenario 2. figure 7. comparison of the current dispatch practice and the proposed dispatch practice. table 4. computational requirements of different scenarios. scenarios scenario 1 scenario 2 scenario 3 number of iteration 9 6 5 execution time 6.03 s 5.74 s 4.97 s simulation results show that the outputs of some power plants, such as sohar ii, barka iii, barka ii, and manah, are more in line with scenario 1 (the proposed practice) than the current practice of oetc ldc. on the other hand, the current practice results in the production of more power in other power plants, for instance, the rusail, ghubra, wadi al jizzi, al kamil, barka i, and sohar 1 power plants. these differences are attributable to (1) the inclusion of losses in the optimization procedure of scenario 1 and (2) failure to consi der water production and voltage support requirements in the simulation. 5. conclusion in this paper, the economic dispatch problem, its formulation, and its objectives and constraints were reviewed. the main interconnected system (mis) of oman was modelled to find the optimal dispatch of power generation at a selected hour on a selected day using the matlab software package. furthermore, different methods for obtaining network loss coefficients were discussed. the loss coefficients of the mis were determined and the optimal dispatch of the power generation units was found, taking the losses into consideration. this was compared with the current practice of dispatching generation units. it was demonstrated that considering the losses in the optimization procedure resulted in the 50 51 52 53 54 55 56 57 58 59 60 f u e l c o st ( t h o u sa n d s $ /h r) scenario 1 scenario 2 0 200 400 600 800 p o w e r p la n t o u tp u t (m w ) oetc ldc dispatch simulation results of scenario 1 m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 10 reduction of the total dispatched power, reducing the cost by $645/hr. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. acknowledgment the authors would like to acknowledge the oman electricity transmission company (oetc) and the oman power and water procurement company (opwp) for supporting this work and for providing the data used in the simulation. references abdelaziz, ay, sf mekhamer, mal badr, and mz kamh. 2008. 'economic dispatch using an enhanced hopfield neural network', electric power components and systems, 36: 719-32. albadi, mh, and ef el-saadany. 2011. 'comparative study on impacts of wind profiles on thermal units scheduling costs', iet renewable power generation, 5: 26-35. dike, damian obioma, moses izuchukwu adinfono, and george ogu. 2013. 'economic dispatch of generated power using modified lambda-iteration method', iosr journal of electrical and electronics engineering, 7: 49-54. gaing, zwe-lee. 2003. 'particle swarm optimization to solving the economic dispatch considering the generator constraints', ieee transactions on power systems, 18: 1187-95. george, ee. 1943. 'intrasystem transmission losses', electrical engineering, 62: 153-58. hill, eugene f, and william d stevenson. 1968. 'a new method of determining loss coefficients', ieee transactions on power apparatus and systems: 1548-53. kron, gabriel. 1951. 'tensorial analysis of integrated transmission systems part i. the six basic reference frames', transactions of the american institute of electrical engineers, 70: 1239-48. kumar, prabhat, and dwarka p kothari. 2005. recent philosophies of automatic generation control strategies in power systems, ieee transactions on power systems, 20: 346-57. oetc (2010), the grid code, version 2. oetc (2014), five-year annual transmission capability statement (2014 2018). muscat. oetc (2016), oman electricity transmission company saoc website. https://www. omangrid.com/sitepages/home.aspx. opwp (2016a), oman power and water procurement company saoc website. http://www.omanpwp.com/new/default. aspx. opwp (2016b), opwp 7-year statement (20162022). muscat. padhy, narayana prasad. 2004. 'unit commitment-a bibliographical survey', ieee transactions on power systems, 19: 1196-205. saadat, hadi. 1999. power system analysis (wcb/mcgraw-hill). su, ching-tzong, and chien-tung lin. 2000. 'new approach with a hopfield modeling framework to economic dispatch', ieee transactions on power systems, 15: 541-45. wood, allen j, and bruce f wollenberg. 2012. power generation, operation, and control (john wiley & sons). yang, xin-she, seyyed soheil sadat hosseini, and amir hossein gandomi. 2012. 'firefly algorithm for solving non-convex economic dispatch problems with valve loading effect', applied soft computing, 12: 1180-86. zhu, jizhong. 2015. optimization of power system operation (john wiley & sons). http://www.omanpwp.com/new/default.aspx http://www.omanpwp.com/new/default.aspx effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 11 appendix table a-1. mis 2014 overhead line data. from substation to substation voltage (kv) length (km) r [ω/km] x [ω/km] b [µs/km] barka power station filaj 220 10 0.0258 0.321 3.82 barka power station filaj 220 10 0.0258 0.321 3.82 filaj airport heights 220 34 0.0258 0.321 3.82 airport heights misfah 220 16 0.0258 0.321 3.82 misfah msq 220 26 0.0258 0.321 3.82 misfah jahloot 220 54.9 0.0258 0.321 3.82 mis sis 220 107 0.0258 0.321 3.82 sps sis 220 38 0.0258 0.321 3.82 al wasit sis 220 66 0.0258 0.321 3.82 blue city mis 220 45.3 0.0258 0.321 3.82 filaj blue city 220 33.2 0.0258 0.321 3.82 al wasit ibri 220 145 0.0258 0.321 3.82 sohar ipp-2 sis 220 36.9 0.0258 0.321 3.82 barka ipp-3 misfah 220 58.2 0.0258 0.321 3.82 al kamil jbb ali 132 55 0.04283 0.2821 3.98 al kamil mudharib 132 51.2 0.04283 0.2821 3.98 al kamil sur 132 73.1 0.04283 0.2821 3.98 barka main filaj 132 6.3 0.04283 0.2821 3.98 dank al hail 132 52.6 0.04283 0.2821 3.98 filaj muladha 132 46.5 0.04283 0.2821 3.98 izki mudaybi 132 62.1 0.04283 0.2821 3.98 izki nizwa 132 31.1 0.04283 0.2821 3.98 manah nizwa 132 19.7 0.04283 0.2821 3.98 airport heights seeb main 132 7.8 0.1472 0.4111 2.6 mis khabourah 132 53.4 0.04283 0.2821 3.98 mabalah barka main 132 11.6 0.04283 0.2821 3.98 ms qaboos jahloot 132 44 0.04283 0.2821 3.98 ms qaboos wadi adai 132 8.2 0.04283 0.2821 3.98 mudhabi mudharib 132 60.1 0.04283 0.2821 3.98 muladha mis 132 11.4 0.04283 0.2821 3.98 nizwa bahla 132 32 0.0857 0.3948 2.85 nizwa ibri 132 123.5 0.04283 0.2821 3.98 rusail mabalah 132 13.1 0.04283 0.2821 3.98 rusail sumail 132 31.2 0.04283 0.2821 3.98 rustaq muladha 132 29.5 0.04283 0.2821 3.98 sis sohar grid 132 27.5 0.04283 0.2821 3.98 sohar grid wadi jizzi 132 24.7 0.04283 0.2821 3.98 sumail izki 132 61 0.04283 0.2821 3.98 wadi adai al falaj 132 3 0.04283 0.2821 3.98 wadi jizzi al wasit 132 36.7 0.0972 0.3168 3.8 wadi kabir wadi adai 132 6 0.0857 0.3948 2.85 manah adam 132 47 0.04283 0.2821 3.98 boushar ms qaboos 132 1.7 0.04283 0.2821 3.98 ghubrah ms qaboos 132 1.7 0.04283 0.2821 3.98 wadi jizzi liwa 132 28 0.04283 0.2821 3.98 liwa shinas 132 20 0.04283 0.2821 3.98 khabourah saham 132 40 0.04283 0.2821 3.98 saham sis 132 30 0.04283 0.2821 3.98 m.h. albadi, f.n. al farsi, n. hosseinzadeh and a. h. al badi 12 al wasit wadi sa'a 132 32.0 0.1503 0.4101 2.83 wadi sa'a dank 132 56.5 0.1503 0.4101 2.83 jahloot yitti 132 25.1 0.0258 0.321 3.82 rusail misfah 132 10.0 0.04283 0.2821 3.98 misfah wadi adai 132 36.0 0.04283 0.2821 3.98 filaj nakhal 132 26.0 0.04283 0.2821 3.98 al wasit buraimi 132 30.73 0.04283 0.2821 3.98 ms qaboos qurum 132 10 0.04283 0.2821 3.98 qurum muttrah 132 10 0.04283 0.2821 3.98 ibri dank 132 55 0.04283 0.2821 3.98 jahloot quriyat 132 50 0.04283 0.2821 3.98 table a-2. 2014 mis cable data. from substation to substation voltage (kv) length (km) r [ω/km] x [ω/km] b [µs/km] sps sis 220 3 0.01074 0.312 91.11 sps sohar industrial area 'a' 220 3 0.01074 0.321 91.11 blue city mis 220 6 0.01074 0.321 91.11 filaj blue city 220 6 0.01074 0.312 91.11 sohar ipp-2 sis 220 2.6 0.01074 0.321 91.11 barka ipp-3 misfah 220 10.2 0.01074 0.321 91.11 airport heights seeb main 132 9 0.04282 0.08361 122.522 mawellah rusail 132 7.3 0.04282 0.08361 122.522 sohar industrial area 'a' sohar refinery co. 132 2.2 0.04 0.08 122.52 wadi kabir wadi adai 132 1.63 0.04 0.143 64.7 ghubrah boushar 132 2.323 0.04 0.08 122.52 boushar ms qaboos 132 1.5 0.04 0.08 122.52 ghubrah ms qaboos 132 4.8 0.04 0.08 122.52 airport heights wave 132 8.806 0.04 0.08 122.52 al wasit buraimi 132 5.47 0.04 0.08 122.52 qurum muttrah 132 3.5 0.04 0.08 122.52 table a-3. peak load and capacitor data in 2014. grid stations p load (mw) q load (mvar) cap (mvar) barka-132 kv 165 16 25 nakhal-132 kv 62 6 25 blue city – 220 kv 53 16 0 mis – 132 kv 141 23 0 muladah – 132 kv 134 5 30 rustaq – 132 kv 117 7 0 khabourah-132 kv 153 23 10 saham 132 kv 112 15 0 sia 220 kv -40 -4.9 0 sia 132 kv 47 8 0 sohar main 132 kv 164 20 20 liwa 132 kv 138 24 0 shinas 85 15 25 mahdah – 132 kv 12 7 0 burimi 1 132 kv 50 8 0 burimi 2 – 132 kv 50 8 35 wadi saa – 132 kv 11 6 0 effect of considering transmission losses in economic dispatch – a case study of oman’s main interconnected system 13 dank -132 kv 18 2 10 al hail – 132 kv 52 9 10 ibri – 132 kv 104 9 10 nizwa and pdo load connected to nizwa 139 18 0 bahla 104 9 0 izki 61 14 0 adam 39 2 0 samail 78 17 0 mudabi 110 18 0 mudayrib 141 4 0 jbb 143 1 0 sur and omifco load 117 4 0 msq – 132 kv 135 16 35 old gubrah 20 11 0 new ghubrah 74 33 0 bosher – 132 kv 226 36 40 airport hight – 132 kv 87 9 30 mwalleh north – authibah 14 1 0 rusail ps – 132 kv 121 34.2 0 misfah 5 13 0 wadi adai 121 19 35 al falaj 68 10 35 wadi kabeer 110 18 35 matrah 47 25 0 qurum 18 9 20 mwalleh south-1 210 16 20 seeb 132 kv 133 42 0 mobillah 104 12 30 jahloot 76 7 25 yitti 15 3 0 quryat 45 -11 0 figure a-1. mis transmission system in 2014. the journal of engineering research (tjer), vol. 15, no. 2 (2018) 142-154 doi: 10.24200/tjer.vol15iss2p142-154 virtual casting: state of the art in metal casting simulation tools m.a.a. khan* and a.k. sheikh mechanical engineering department, king fahd university of petroleum and minerals, dhahran, saudi arabia. received 20 september 2017; accepted 29 january 2018 abstract: the demands on the productivity and robustness of metal casting processes for high quality components are continuously increasing. moreover, the financial considerations necessitate meticulous and reliable planning of the entire casting process before it is actually put into practice. a holistic approach to perform cradle to grave analysis of cast products is simulation-based metal casting. this method allows engineers to model, verify, and validate the process followed by its optimization and performance prediction in virtual reality. this paper provides insights on state of the art in simulation-based metal casting with reference to some case studies. casting simulations software, mathematical models and solution methods, and casting process simulation together with the results obtained are clearly explained. the current practices revealed extensive utilization of simulation packages for defect minimization, yield maximization, and improved quality. the ongoing research on integration of casting simulations with mechanical performance simulations makes it possible to analyze the serviceability of cast parts. the reliability of cast part in service with dynamic loading of varying thermal and mechanical load cycles can be predicted through this integration. however, more rigorous work is needed in this area, particularly by developing the reliability prediction modules embedded in advanced simulation tools. keywords: metal casting; simulation; software; design; defects. الطرق يف سبك املعادنالقولبة أالفرتاضية : أحدث عمل القوالب اخلاصة باألدوات و أ.ك. شيخ *م.أ.أ. خان إن الطلب على إنتاجية ومتانة السبائك املعدنية إلنتاج مكونات عالية اجلودة يتزايد على حنو. كما أن االعتبارات :امللخص ولذا فان النهج الشامل .موضع التطبيق الفعلي املالية تستلزم ختطيطا دقيقا وموثوقا لعملية السبك بأكملها قبل وضعها يسمح للمهندسني هذا إلرساء أساسا لتحليل متزن للمنتجات املسبوكة يكون مبنيا على عملية القولبة االفرتاضية. وتقدم .عي األمثل واألداء الواقباألداء القائمني عليها بوضع منوذجها والتحقق من صحة انتاجها و اجازتها وبعد ذلك التنبؤ هذه الورقة نظرة ثاقبة على أحدث ما توصلت إليه جتارب السبك املعدني القائم على القولبة مع اإلشارة إىل بعض دراسات سبك القوالب و النماذج احلسابية وطرق احلل وعملية سبك القوالب جنبا إىل جنب مع احلالة. و مت القيام بشرح واف لربامج صول عليها. وأظهرت املمارسات احلالية وجود استخدام مكثف حلزم القولبة لتقليل العيوب وزيادة النتائج اليت مت احل و تعمل البحوث اجلارية القائمة على دمج سبك القوالب مع األداء امليكانيكي على التمكن من . االنتاجية وحتسني النوعية التنبؤ مبدى صالحية اجلزء املسبوك للعمل من خالل دمج ميكن للعمل , كما حتليل مدى صالحية أألجزاء املسبوكة التحميل الديناميكي لدورات احلمل احلرارية وامليكانيكية املتغرية ، غري أنه يلزم القيام مبزيد من العمل اجلاد يف هذا .اجملال ، وال سيما من خالل تطوير وحدات التنبؤ مبدى موثوقية أدوات القولبة املتقدمة .العيوب،التصميم ،القرص املرن،: سبك املعادن ات املفتاحيةالكلم * corresponding author’s e-mail: azharali@kfupm.edu.sa m.a.a. khan and a.k. sheikh 143 1. introduction metal casting is one of the simplest and direct methods of producing near net shape products. the process essentially requires a mold cavity made up of sand, ceramic, or steel, where molten metal is poured and solidified to obtain cast products. a number of casting processes have been developed and modified over time to meet the challenges associated with the new materials to be cast, operating temperatures, complexity of shapes, and high quality etc. the advancements in casting processes are reflected by an upward trend in world cast production during the last decade as shown in fig. 1(a). in 2014, the estimated amount of metal processed by metal casting is reported to be 105 million metric tons, which is a 2.3 % increase compared to the previous year (afs 2015; afs 2016). moreover, the widespread use of metal casting around the globe is presented in fig. 1(b), where top ten casting countries are shown with their average production per plant in thousands of metric tons. it is important that such large production volumes should produce, with higher casting yield, minimum defects, and require mechanical properties. for this reason, metal casting industries have adapted the simulation-based casting methodology, which analyzes the entire casting process and its optimization in a virtual domain before it is actually put into practice. the developments in advanced casting simulation tools have supplanted the conventional trial-and-error approach by a more scientific proof-of-concept approach through computer-aided modeling of the part to be cast, pattern design, methoding, process simulation and virtual optimization (behera et al. 2010). moreover, defects prediction and minimization via simulation ensure integrity and repeatability in castings together with faster production rate and minimum production cost. this paper first explains how a casting simulation software works, associated mathematical models and solution methods, and the results obtained by simulation. next, the current practices in metal casting to minimize defect and improve casting quality are presented. case studies on gating, runner, and riser design, systematic and autonomous optimization, stress and strain simulations, and integration of casting simulations with mechanical performance simulations are included, discussed and conclusions drawn. 2. casting simulations casting simulations are extensively utilized in foundries for all cast materials, processes and complex casting shapes. with continuous developments, these simulations are currently used in different stages of product development such as casting design, process determination, flow pattern, design of tooling, quality control and product stress analysis to name a few. casting design is important as it influences all subsequent stages of product development. computer-aided casting designs allow for figure 1. (a) total casting tonnage by alloy (b) average production per plant (thousands of metric tons)(afs 2015; afs 2016). virtual casting: state of the art in metal casting simulation tools 144 optimum casting geometries and features, which can be confirmed by filling simulation, solidification analysis, and stress distribution within the cast product. from process determination perspective, the flow characteristic in a mold is revealed and subsequent solidification behavior can be analyzed, which confirms the optimum mold design, appropriate process routing and process parameters. defect minimization and quality improvements are also possible by simulating the filling and solidification behavior and recognizing the underlying factors affecting product quality and mechanisms of defect formations. 2.1 casting simulation softwares modelling and simulation of casting processes is complicated due to a host of parameters involved; such as fluid velocity, pressure, geometry of the mold, gating and runner system etc. a range of softwares that have emerged over time are a result of understanding the physical phenomena occuring during a casting process. the relevant mathematical models are either developed or modified and then implemented into computer programs to develop a software (ravi 2008). some of the most commonly used casting simulation softwares, which are currently available to researchers and foundrymen, are procast, flow-3d cast, magmasoft, novasolid/flow, autocast, solidcast, castcae, and capcast etc. a casting simulation project using commercial softwares generally comprise five stages as shown in fig. 1 (ravi 2010). data gathering refers to all information needed related to cad model of casting, cast metal properties, mold properties and process parameters etc. methods design and modeling primarily focuses on how to convert the as cast part model into a three dimensional mold which contains cavities, gating system, runners, risers, cores and feed-aids. next, the numerical simulation is done after generating the optimum mesh and defining boundary conditions. visualization of results is done by a post processing module in each simulation software. with the simulation results, it is often possible to identify defects such as hotspots, microporosity, shrinkage, cold shuts and others. therefore, a step forward in simulation is methods optimization which includes modifications in gating and riser designs, process parameters and material properties, and even in part model to minimize the defects. the final stage, which is termed project closure, includes complete documentation of results, generating methods and analysis reports, capturing images and animations for demonstration at the later stage. 2.2 mathematical modelling and solution methods casting simulations begin with modeling the physical phenomena through mathematical equations. in a mathematical perspective, models are expressed as governing equations and boundary conditions. owing to the nonlinearity of models in terms of both geometry and material properties, numerical methods have to be used to avoid non-linearity and thus formulate simultaneous and algebraic equations. the set of developed equations is then used to explain casting process in the form of action-behavior-property relationship (fu et al. 2009). for a metal casting process, the action is supplying molten material to the mold, the behavior is the flow of molten metal within the mold, and the behavior is further decided by properties of the molten metal. from modeling point of view, three important phenomena in any casting process simulation are mold filling, solidification and cooling, and stress and strain profile of the cast parts. in each of these phenomena, a certain set of governing equations are employed. for example, mold filling is modeled by continuity equation, momentum equation (navier-stokes equation) and energy equation (jolly 2003). figure 2. casting simulation and optimization protocol (ravi 2010). m.a.a. khan and a.k. sheikh 145 equations (1) to (3) represent the complete forms of continuity, momentum and energy equations respectively and thus modified accordingly if the density is assumed to be constant and/or the molten metal is a newtonian fluid.   0u t j xj         (1)     i j i xjxi ij xj i t g x u uuu                          (2)     q x t x tuc x tc t jj jp j p                    (3) a variety of methods for solving mathematical equations are available, the most common of which are finite difference method (fdm), finite volume method (fvm), finite element method (fem), vector element method (vem), cellular automation method (ca) etc. (jolly 2002). finite difference method: it is a numerical method where a complex problem is solved by discretizing the complete region of problem (also known as domain) into a finite number of small portions (also known as control volumes). material properties are assumed to be constant throughout the volume. therefore, for high accuracy of the results, the domain should be divided into a maximum number of control volumes possible, taking into account the computational time. fdm is a differential scheme, which is approximation of taylor series expansion. calculations are iterative and done at a predetermined time-step. the results can be stored at the end of each time-step or after a predetermined numbers of steps (jolly 2003). finite volume method: unlike fdm, fvm is an integral scheme. although the idea of discretizing the domain into small control volumes remains the same, the use of integral formulations is advantageous in treating the neumann boundary conditions as well as that of discontinuous source terms due to their reduced requirements on the regularity or smoothness of solution. finite element method: fem discretizes the complete domain of the problem into small pieces; however, they are now termed as elements. each element is made up of nodes (corner points) and edges, which store material properties to be used in computation. the solution is done by using these values to determine a quantity for these specific points (also known as gauss points) within the elements. the position of these points in elements is a function of the integration applied, initial coordinates of the nodes, and the element shape (jolly 2003). values of variables, which are considered to be constant in fdm/fvm across the elements, are calculated using some interpolation function. however, the treatment of time in an iterative and step-wise manner is similar to fdm/fvm. vector element method: this approach to casting simulations is based on determining the largest thermal gradient at any point inside the casting, which is given by the vector sum of flux vectors in all directions from that point (amin et al. 2014). vector element method is relatively simple when compared to other numerical techniques but provides reliable and robust results (ravi 2005; sutaria et al. 2014). in some instances a combination of two or more techniques may also be employed, the examples of which are cellular automation finite element (cafe) method proposed by (rappaz et al. 1998), and a hybrid method for casting process simulation by combining fdm and fem (si et al. 2003). the final simulation results, however, are representative of casting process and properties, qualities and defects of cast products irrespective of the type of solutions discussed above. 2.3 simulation results casting simulations result in three major aspects of a casting process: filling, solidification and stress analysis. mathematical modeling and simulation of conservation of mass, momentum and energy in a filling process provides information about velocity of molten material within the mold cavity, the direction of flow, temperature and pressure at various instances within the mold. physical and thermal characteristics of the filling process are derived from these results, some of which are flow front progression, turbulence in flow, filling evenness, air and gas entrapment, temperature profile, filling sequence, velocity profile of molten metal, odd behaviors such as splashing, misruns, cold shuts etc. during filling. solidification in the casting process is generally complex, where physical, thermal and metallurgical phenomena take place simultaneously. solidification simulation provides information about how these virtual casting: state of the art in metal casting simulation tools 146 phenomena are occurring in process conditions together with the defects that might arise during solidification phase. some key findings of a solidification simulation include cast area that solidifies last, solidification sequence, validation of cooling design, validation of runner design, defects due to shrinkage and microporosity, appropriate riser geometry, size and location within the mold. the stress and strain simulation demonstrates the state of cast parts after ejecting from the mold. the results of these simulations may include identification of dimensional inaccuracies, residual stress generation and distribution in cast part; defects arise due to stress and strain, temperature profile in ejected cast part, design improvements in casting design such as modifications in riser design to reduce stresses etc. 3. current practices in simulationbased metal casting the current practices in simulation-based metal casting brought solutions to many problems in foundries. high quality defect free castings are now produced by developing an in-depth understanding of the areas such as: (a) filling and solidification during several casting processes, (b) design of runner and gating systems, (c) design of feeding systems (locations and number of risers), (d) casting process parameters (thermo-physical data, injection parameters etc.), (e) stress distribution in cast products, and (f) quality control and assurance of the cast products. some of the efforts in these areas are as follows. 3.1 gating, runner, and riser design optimization mi et al. (2009) used casting simulations to modify gating and riser system of a valve block. the original design with two in gates and rounded-rectangular riser geometry as shown in fig. 3(a) resulted in defects i.e. shrinkage and cold shut. the new casting design, as shown in fig. 3(b), comprised a single in gate together with cylindrical risers as they allow more liquid metal during solidification. chills were added to help solidification in the bottom region of casting and shrinkage was overcome via modified riser design. the final casting was reported to be free from surface cracks on the valve body. sun et al. (2008) presented multiple objective optimization for gating system of a cylindrical magnesium alloy casting. the optimum gating system was selected by changing four parameters: ingate height, ingate width, runner height, and runner width as shown in fig. 4(a). the main criteria for casting quality were filling velocity, shrinkage porosity and yield. bottom filling approach was employed as shown in fig. 4(b). it was concluded that runner with small height and large width is effective in reducing metal velocity at the ingate which is consistent as reported by mi et al. (2009); and hu et al. (2000). figure 3. casting design for a valve body (a) original and (b) modified. (mi et al. 2009). m.a.a. khan and a.k. sheikh 147 kermanpur et al. (2008) simulated the filling and solidification sequence of brake disc and flywheel in multi-cavity molds. it was reported that a symmetrical mold configuration for flywheel, as shown in fig. 5(a), provides uniform filling contrary to the one for brake disc presented in fig. 5(b). simulated hotspot reflected as micro-shrinkage in brake disc for which lower superheat temperature was recommended. in another work by choudhari et al. (2014), the hotspot defect was eliminated by simulating an exothermic sleeve around the riser which, delayed solidification time as shown in fig. 6. riser system design was simulated based on quality, feeding yield and feeding efficiency followed by experimental validation which also resulted in cast products free from shrinkage defect. liu et al. (2007) simulated the decomposition of expanded polystyrene (eps) pattern in lost foam casting process using flow 3d. the conventional constant heat transfer coefficient (chtc) approach is modified by incorporating the variability in the heat transfer coefficient (vhtc), which resulted in a better agreement figure 4. (a) gating parameters (b) gating system design and (c) shrinkage porosity prediction in cylindrical magnesium casting (sun et al. 2008). figure 5. final results of filling (left) and solidification (right) for (a) flywheel and (b) brake disc. virtual casting: state of the art in metal casting simulation tools 148 of simulation and experimental results. the effect of temperature on mold filling time was also studied and it was reported that mold filling time predicted by vhtc model were found close to the experimental results. defects in castings are predicted via both chtc and vhtc models where the main mechanism of defect formation was the meeting of metal fronts entering from different in gates. the results of defect predictions with three in gates as well as a modified two in gates system are presented in fig. 7, which confirms reductions in casting defects with fewer in gates. nimbulkar et al. (2016) optimized the gating and feeding system design for casting a wear plate using autocast. the original design for casting utilized a vertical gating system, which resulted in casting defects. in this study, the vertical gating system was replaced by a horizontal gating system design. the existing and new casting designs were simulated by autocast followed by their experimental validation. the final cast products were examined through ultrasonic inspection. it was found that vertical gating system was not suitable for thick casting components due to multiple casting defects observed during solidification. however, the horizontal gating system with symmetrical gates and risers enabled uniform flow of the melt together with minimized porosity in the final casting. in total, the gating and feeding system related defects are found to be reduced by 30%. 3.2 systematic autonomous optimization a recent approach to improve simulationbased casting is systematic and autonomous optimization, where softwares are used as test field or virtual experimentation. sikorski et al. (2012) optimized the gating system of a simple casting layout using magma soft, which follows the sequence as shown in fig. 8(a). owing to a multi-cavity mold, it was aimed to reduce the filling time between the cavities to minimize defects due to air entrapment. the results of initial situation and final optimized solution using the filling time difference criterion are presented in figs. 8(b), and 8(c) respectively. the final casting after autonomous optimization is found to be free from casting defects especially due to air entrapment and oxides. figure 6. effect of exothermic sleeve on the position of hotspot in stepped plate casting (choudhari et al. 2014). figure 7. formation of casting defects by (a) chtc model; (b) vhtc model; and (c) modified model with two in gates (blue and red represents lowest and highest probability for defects respectively) (liu et al. 2007). m.a.a. khan and a.k. sheikh 149 hahn et al. (2010) compared the design of experiments (doe) with virtual autonomous optimization for steel casting. the number, location and dimension of the chills and feeders were investigated as shown in fig. 9(a) to check their effect on the shrinkage distribution. the results, as shown in fig. 9(b), demonstrate that autonomous optimization led to further reduction in shrinkage by fine tuning the chill size. the autonomous optimization was also done for a ductile iron wind turbine hub to control shrinkage porosity. moreover, the configuration of runners in a multi-cavity mold for casting six parts was also optimized in the same study (hahn et al. 2010). sturm et al. (2007) performed the autonomous optimization for casting a head cap in a multi-cavity mold. the actual mold design contained a single cavity where successful filling was achieved through specifically designed gate and a slightly slanted runner as shown in fig. 10(a). a multi-cavity mold was designed and optimized to achieve similar melt velocities and directions as it was already obtained in single cavity mold. moreover, minimization of air entrapment was also set as another criterion for optimization. in total, only 1 out of 106 designs for right-hand runner and 2 out of 97 designs for left-hand runner were qualified meeting all objectives of optimization. the flow patterns for single and multi-cavity molds are presented in fig. 10(a) along with the actual head cap castings in fig. 10(b). figure 8. (a) schematic sequence of optimization in magmasoft (b) initial situation and (c) final optimized solution (sikorski et al. 2012). figure 9. (a) feeder and chills in cope (left) and drag (right) (b) shrinkage distribution by autonomous optimization (left) and doe (right) (hahn et al. 2010). virtual casting: state of the art in metal casting simulation tools 150 3.3 stress and strain simulations the application of numerical simulation to distortion and stress-related problem in casting was presented by egner-walter et al. (2013). the study first explained the formation of residual stresses and distortion in casting followed by simulation of die cast rear door lock panel for a passenger car. figure 11(a) represents the temperature field at ejection and it forms the basis for the formation of stresses and deformation during solidification and cooling to room temperature. the runner being hotter contracted more in the subsequent cooling phase and pulled the casting inwards toward it as shown in fig. 11(b). the swan neck moved in an upward direction, as the upper side of the casting, which faced the shot slug, had a higher temperature than the lower side after ejection. another problem of hot tears, especially important for high pressure die castings, was also investigated by egner-walter et al. (2013) by simulating the hot tear criteria for a flywheel as shown in fig. 12. it was suggested that hot tears can be avoided by modifying process parameters and cooling conditions, or by minor changes in the geometry of casting. figure 10. (a) identical flow patterns in single and multi-cavity molds and (b) actual castings after autonomous optimization (sturm et al. 2007). figure 11 (a) temperature distribution in casting and runner at ejection and (b) deformation of the casting after cooling to room temperature: original geometry (semitransparent grey) and deformed geometry (egner-walter et al. 2013). m.a.a. khan and a.k. sheikh 151 hartmann (2013) studied the stress generation and distribution due to material combinations in light-weight cast components. engine blocks are one such application where the grey iron liners are set into the molds and preheated before pouring molten aluminum. the molten metal gets in contact with liner cools rapidly and solidifies quickly which sometimes results in incomplete filling of the mold. certainly, residual stresses are generated due to a non-uniform temperature distribution in the cast part. since, aluminum shrinks onto the iron liners, tensile stress is generated in aluminum whereas the liner experiences compressive stress, together with other internal stresses as shown in fig. 13. it is reported that the tensile residual stress observed in aluminum between the liners is high, however, any crack in this area does not affect the rigidity of cylinder block at operating temperature. for stress minimization, it was suggested to use the high pre-heat temperatures of the liners to decrease the tensile residual stress in the seam between liners. 3.4 casting simulations integrated with mechanical performance simulations recently, casting simulations are combined with mechanical performance simulations to investigate the life of cast products in service. the effect of casting quality on service life is analyzed by taking the predictions of defects such as shrinkage and porosity from casting simulation results and maps them with stress and fatigue life simulations. figure 12. prediction of hot tears in a flywheel casting using hot tear criterion (egner-walter et al. 2013). figure 13. internal stress in gap between two liners and uneven stress around the liner. high stress (left) in closer distances and low stress (right) in larger distances of the liners. (hartmann 2013). virtual casting: state of the art in metal casting simulation tools 152 bordas et al. (2006) studied the integration of foundry process simulations, non-destructive evaluation, stress analysis, and damage tolerance simulations to casting design as shown in fig. 14. casting simulations are done to determine any porosity-related defects by using a radiographic inspection simulation tool (xrsim). failure caused by the predicted defects is determined by a fatigue crack growth simulation based on extended finite element method. this integrated approach enabled for an “accept or reject” criteria to be set at an early design stage and allows for damage tolerant design strategies. dørum et al. (2007) studied the effects of porosity and surface quality on structural behavior by mapping results of casting simulation by magmasoft with a two-dimensional (shell element) finite element model developed in ls-dyna. the study was primarily focused on mechanical performance of thin walled cast magnesium components subjected to quasi-static loading conditions. olofsson et al. (2012) presented software that incorporates the prediction of casting simulation software into a finite element method. the software is validated through a test case of ductile iron component. in another study by hardin et al. (2012), the fatigue life of a cast steel component in service is predicted by utilizing the porosity predictions of magmasoft in fe simulations. a blockdiagram of the methodology used in the analysis is presented in fig. 15. this approach was first validated for a simple geometry of tensile test specimen by comparing experimentally obtained and the predicted crack initiation fatigue lives. the difference was reported to be within one decade that is thought to be very good for fatigue life prediction. figure 14. integration of nde simulation, casting modeling, and damage tolerance simulation (bordas et al. 2006). figure 15. integration of magmasoft results in abaqus for fatigue life prediction. m.a.a. khan and a.k. sheikh 153 4. conclusion the key conclusions and recommendations from this study are summarized as follows: (i) casting simulations are capable of examining the effect of several process variables in producing sound castings. (ii) modeling boundary conditions precisely in a casting simulation software followed by its experimental validation is challenging. (iii) cast products are often produced with residual stresses the accurate prediction of which can help in suggesting appropriate heat treatments. (iv) autonomous optimization must be utilized at its best to improve quality. (v) reliability of cast part in service with dynamic loading of varying thermal and mechanical load cycles can be predicted through integration of casting simulations with mechanical performance simulations. however, more rigorous work is needed in this area particularly by developing the reliability prediction modules embedded in advanced simulation tools. (vi) all casting simulation softwares are not created equal, based upon the in-house capabilities of foundry skills, cad expertise available, and level of accuracy and integrity needed in cast products, it is imperative to very carefully see and compare all features available in each software and buy the software which meet your goals and provide best benefit over cost ratio. conflict of interest the authors declare no conflicts of interest. funding this work is funded by the national science, technology and innovation plan (nstip) grant number 14-adv890-04-r. acknowledgment the authors greatly acknowledge the support provided by king fahd university of petroleum and minerals in this research. references afs (2015), 49th census of world casting production modest growth in worldwide casting market. afs (2016), modern casting census of world casting production – steady growth in global output. amin ld, patel s, mishra p, joshi d (2014), rapid development of industrial castings using computer simulation. indian foundry journal 60(8). behera r, das s, dutta a, chatterjee d, sutradhar g (2010), comparative evaluation of usability of femand vembased casting simulation software. in 58th indian foundry congress, ahmedabad, india bordas spa, conley jg, moran b, gray j, nichols e (2006), a simulation-based design paradigm for complex cast components. engineering with computers 23(1): 25–37. choudhari cm, narkhede be, mahajan sk (2014), methoding and simulation of lm 6 sand casting for defect minimization with its experimental validation. procedia engineering 97: 1145–1154. dørum c, hopperstad os, lademo og, langseth m (2007), numerical modelling of the structural behaviour of thin-walled cast magnesium components using a throughprocess approach. materials & design 28(10): 2619–2631. egner-walter a, kothen m (2013), using stress simulation to tackle distortion and cracking in castings. metallurgical science and tecnology 24(2). fu mw, yong ms (2009), simulation-enabled casting product defect prediction in die casting process. international journal of production research 47(18): 5203–5216. hahn i, sturm jc (2010), autonomous optimization of casting processes and designs. in world foundry congress, hangzhou, china october, 16–20. hardin ra, beckermann c (2012), integrated design of castings: effect of porosity on mechanical performance. iop conference series: materials science and engineering 33(1): 012069. hartmann g (2013), material combinations in light weight casting components. casting plant and technology 4: 32–37. hu bh, tong kk, niu xp, pinwill i (2000), design and optimization of runner and gating systems for the die casting of thinwalled magnesium telecommunication parts through numerical simulation. journal of materials processing technology 105(1–2): 128–133. jolly m (2002), casting simulation: how well do reality and virtual casting match? state of the art review. international journal of cast metals research 14(5): 303–314. virtual casting: state of the art in metal casting simulation tools 154 jolly m (2003), castings. in comprehensive structural integrity 377–466. elsevier kermanpur a, mahmoudi s, hajipour a (2008), numerical simulation of metal flow and solidification in the multi-cavity casting moulds of automotive components. journal of materials processing technology 206(1–3): 62–68. liu xj, bhavnani sh, overfelt ra (2007), simulation of eps foam decomposition in the lost foam casting process. journal of materials processing technology 182(1–3): 333–342. mi g, liu x, wang k, fu h (2009), application of numerical simulation technique to casting process of valve block. journal of iron and steel research, international 16(4): 12–17. nimbulkar sl, dalu rs (2016), design optimization of gating and feeding system through simulation technique for sand casting of wear plate. perspectives in science 8: 39–42. olofsson j, svensson il, (2012), incorporating predicted local mechanical behavior of cast components into finite element simulations. materials and design 34: 494–500. rappaz m, rettenmayr m (1998), simulation of solidification. current opinion in solid state and materials science 3(3): 275–282. ravi b (2005), metal casting: computer-aided design and analysis. new delhi: prenticehall of india. ravi b (2008), casting simulation and optimization: benefits, bottlenecks, and best practices. indian foundry journal january (special issue). ravi b (2010), casting simulation – best practices. in transactions of 58th ifc, ahmedabad, india sih m, cho c, kwahk sy (2003), a hybrid method for casting process simulation by combining fdm and fem with an efficient data conversion algorithm. journal of materials processing technology 133(3): 311– 321. sikorski s, dieckhues gw, sturm jc (2012), systematic optimization of aluminum sand casting gating systems. american foundry society. sturm c, seefeldt r (2007), two for one– transferring proven filling characteristics from a single to a two cavity die casting die. casting plant & technology 3: 44–52. sun z, hu h, chen x, wang q, yang w (2008), gating system design for a magnesium alloy casting. journal of materials science and technology 24(1): 93–95. sutaria m, ravi b (2014), computation of casting solidification feed-paths using gradient vector method with various boundary conditions. the international journal of advanced manufacturing technology 75(1–4): 209–223. microsoft word paper 2.docx the journal of engineering research (tjer), vol. 14, no. 1, (2017) 10-22 out of axis movement of an auv inside a water pipeline m. moonesun*, a, a. mahdianb, y.m. korolc and a. brazhkod *a national university of shipbuilding admiral makarov (nuos), faculty of ship design, ukraine. b malek university of technology (mut), mechanical engineering department, iran. c national university of shipbuilding admiral makarov (nuos), faculty of ship design, ukraine. d national university of shipbuilding admiral makarov (nuos), faculty of ship design, ukraine. received 23 december 2014; accepted 26 june 2016 abstract: the research intends to evaluate the variation in the resistance and the lift of a torpedo shaped auv brought about by the wall effect inside the pipe as it moves out of the axis inside a water pipeline. movement of an auv at the axis of a pipe causes minimum resistance and lift forces, but when the auv moves at a position parallel with the axis of the pipe (out of axis of the pipe), the hydrodynamic forces especially the lift force changes. the auv must be able to move a float inside the pipe and perform non-contact inspection. in water pipes having limited diameters, there is the wall effect. the added resistance and the lift have to be calculated accurately, which is a necessary requirement for the determination of the vehicle speed, power demand, control, range and duration of the operation. according to the findings of this paper, when moving at the center of pipe the ratio of auv diameter to pipe diameter is equal to 12. this value can be considered for the determination of "the critical pipe diameter" which gives zero resistance. the results of this study can be applied for torpedo movement inside the torpedo tube. the analysis is performed by the flow vision (v.2.3) software based on the cfd method and solving the rans equations. key world: auv, pipeline, hydrodynamic, resistance, lift, wall effect. אאאא  k،*k،kk،אk، אwאאאאאאauvא אאאאאאאkאא אאא  א؛ אא kא אאauvאאאאאkא אאא אאאkא،אא،א،אא ؛ אא אא   k؛אא אאאא١٢،אאא אאאkאאאאאk אאא(v.2.3)אcfdransk    אאw،א،אא،،א،אkאא  * corresponding author’s email: m.moonesun@gmail.com m. moonesun, a. mahdian, y.m. korol and a. brazhko   11 nomenclature a distance from pipe axis to auv axis (m) a1 section area of the model a2 section area of water between model and pipe wall auv autonomous underwater vehicles d maximum diameter of auv (m) d diameter of pipe (m) e ratio of out of axis to the pipe radius (e=a/r) ihss iranian hydrodynamic series of submarines l overall length of hull (m) l overall length of pipe (m) p average pressure of fluid between pipe and model (pa) re reynolds number rf frictional resistance (n) rp pressure resistance (n) rt total resistance (n) r0 total resistance in free stream (without pipe wall stream) (n) * other notations are referred. out of axis movement of an auv inside a water pipeline   12 1. introduction water or petroleum pipelines are extended millions of kilometers all over the world. a severe breakdown of a pipeline may bring about large cost which might lead to pollution and accidents as well. hence routine inspection and maintenance of the pipelines are necessary so that they can function trouble-free. water supply is an essential public service and therefore inspection tasks cannot be compromised on continuity or quality. an inspection system must have the capacity to operate while the pipeline is in-service (painumgal et al. 2013). today several different pipeline inspection robots exist (roman and pellegrino 1993; suzumori et al. 1998; hirose et al. 1999; muramatsu et al. 2000; roh et al. 2001, roh and choi 2005; moghaddam and hadi 2005). the pipeline inspection robots presently are used by the contact with the walls for motion and positioning in the centre of the pipe (okamoto et al. 1999). figure 1 (painumgal et al. 2013), shows the autonomous underwater vehicle (auv) inside a pipeline. this is a simple auv with conical ends, but today modern auvs are usually torpedo shaped. discussion about submarine shape design is presented by joubert (20040 and (moonesun 2014). the aim of developing this inspection system is to enable non-destructive and non-disruptive inspection of water pipelines. recently torpedo shaped ocean going auvs (najjaran 2005) has been modified for inspecting large pipelines. the vehicles have been used to inspect very huge pipelines having a diameter of more than 2 meters. it is important for the pipeline internal inspection system to enter and leave the pipeline with the least disruption or damage to the pipe system. a great pipeline network means a great amount of pipeline network inspection. hence, autonomous and free operation will be ideal for long distance inspection so as to avoid cable entanglement. therefore, autonomous underwater vehicles (auv) are ideal tools for pipeline internal inspection (painumgal et al. 2013). 2. description of the problem and necessity the flow field around a submarine inside a pipe or ducted space is different from that of a free stream. the proximity to the interior boundary of the hull induces wall effect on the fluid flow. in such cases, the boundary layer develops all over the circumference. the initial development of the boundary layer is similar to that over the flat plate. at some distance from the entrance, the boundary layers merge and further changes in velocity distribution become impossible. the velocity profile beyond this point remains unchanged. the distance up to this point is known as the entry length, which is about 0.04re×d. the flow beyond this amount is said to be fully developed. the velocity profiles in the entry region and fully developed region are shown in fig. 2. the flow is observed to be laminar until a reynolds number value of about 2300. the reynolds number is calculated based upon the basis of the diameter (ud/ν). in pipe flow, it is not a function of length. as long as the diameter is constant, the reynolds number depends upon the velocity for a given flow. figure 1. schematic of pipeline inspection system and sample auv (pictan) (painumgal et al.2013). m. moonesun, a. mahdian, y.m. korol and a. brazhko   13 figure 2. boundary layer and velocity distribution inside a pipe in laminar flow. hence, the value of velocity determines the nature of flow in pipes for a given fluid. the value of the flow reynolds number is decided by the diameter and the velocity. as shown in fig. 2, region (a) is the non viscous flow not affected by the boundary layer, but region (b) is the boundary layer region. after some length, the boundary layers merge and the flow becomes fully developed. the entry length in turbulent flow is about 10 to 60 times the diameter. the velocity profile in the fully developed flow remains constant and is generally flatter compared to laminar flow for parabolic case. now it seems clear that the flow through the pipe is different from that of a free stream. should there be an initial flow stream inside the pipe, the entrance length must be regarded but if there is no flow speed and only the auv moves, entrance length could be ignored. in the present research, only the auv moves, and there is no initial fluid speed. when an auv moves inside the pipe, it will experience the wall effects, especially in low diameter pipes. this causes an increase in resistance. the lower the pipe diameter, the higher the resistance. this added resistance should be calculated accurately, since it is a necessary requirement for the determination of vehicle speed, power demand, range and duration of operation. another important parameter is out of axis movement of the auv. movement of an auv parallel to the axis (at the axis) of a pipe, causes minimum resistance and minimum lift forces, but when auv moves out of axis of the pipe, the hydrodynamic forces, especially the lift force changes. 3. specifications of the model and pipe for this research only the bare hull is to be studied. the base model considered here, is an axis-symmetric auv similar to a torpedo, without any appendages. it helps to halve the cfd modeling of the body, which is a time saver. the total length is 2 meters, the diameter is 0.25 m, the wetted surface area is 1.35 m2, the volume is 0.08 m3 and the fineness ratio (l/d) is 8. the specifications of the model are presented in fig. 3. dimensions of the model are constant, but the tube diameter (d) is different. for the present study, 10 pipe diameters are considered as follows; d/d=1.22, 1.41, 1.58, 1.73, 2, 3, 4, 10, 12 and 13 are shown in fig. 4. the ratio of the water section area (a2) to the section area of the model (a1) is important. some of the values of d/d are considered according to a2/a1. the values of d/d= 1.22, 1.41, 1.58 and 1.73 are respectively equivalent to a2/a1=0.5, 1, 1.5 and figure 3. general configuration of the models. irrotational velocity developing fully developed flow region boundary layer velocity profile velocity profile hydrodynamic entrance region fully developed region out of axis movement of an auv inside a water pipeline   14   2. three different speeds are considered; 1, 3 and 10m/s. these speeds are considered so that all reynolds number (ud/ν) values are more than 2300. this provides fully turbulent flow or transition zone inside the pipe. the usual speed of auvs inside the tube is in the range of 1~3m/s but the speed of 10m/s is considered for such high speed vehicles as; the ejection of torpedo from the torpedo tube. the out of axis movement of auv, is evaluated for two values: d/d=4 and 12. the parameter "e", shows the ratio of out of axis to the radius of pipe (e=a/r). the parameter "a" is calculated from the axis of the pipe (fig. 5). in the present study, four values of "e" are studied: e=0, 0.25, 0.5, 0.75. 4. cfd method of study governing equations  viscosity or friction between fluid layers result in transfer of momentum from one fluid layer to another; this is a molecular level effect (due to rubbing of adjacent d (m) d/d a2/a1 0.304 1.22 0.5 0.36 1.41 1 0.394 1.58 1.5 0.432 1.73 2 0.5 2 2.9 0.75 3 7.8 1 4 14.7 2.5 10 97.1 3 12 140.3 3.25 13 164.8 a1= /4 , a2= /4 figure 4. cross section area of auv and pipe (hatched area: water between auv and pipe). e = a/r 0 0.25 0.5 0.75 figure 5. out of axis movement of inside the pipe. auv molecules).  turbulent mixing resulting in additional apparent stress or reynolds stresses (after osburn reynolds in 1880); this is a macroscopic effect due to bulk motion of fluid elements. momentum (navier-stokes) equation (3 components in cartesian coordinates) in conservative form are as eqn.1: ρ ∂u ∂t ∂ u ∂x ∂ uv ∂y ∂ uw ∂z ∂p ∂x ∂ ∂x μ ∂u ∂x ∂ ∂y μ ∂u ∂y ∂ ∂z μ ∂u ∂z ρ ∂v ∂t ∂ v u ∂x ∂ v ∂y ∂ v w ∂z ∂p ∂y ∂ ∂x μ ∂v ∂x ∂ ∂y μ ∂v ∂y ∂ ∂z μ ∂v ∂z ρ ∂w ∂t ∂ w u ∂x ∂ w v ∂y ∂ w ∂z ∂p ∂z ∂ ∂x μ ∂w ∂x ∂ ∂y μ ∂w ∂y ∂ ∂z μ ∂w ∂z (1) another form by using einstein notation (sum each repeated index over i, j, and k) for cartesian coordinates and the x-component where (xi , xj , xk) = (x , y , z) and (ui , uj , uk)= (u , v , w) are as eqn. 2: ρ μ (2) in terms of shear stress by using: τ μ ∂u ∂x ∂u ∂x ,or mean strain rate, m. moonesun, a. mahdian, y.m. korol and a. brazhko   15 s τ 2μ , the equation is reformed to eqn. 3: ρ ∂ u ∂t ∂ u u ∂x ∂p ∂x ∂τ ∂x or ρ ∂ u ∂t ∂ u u ∂x ∂p ∂x ∂ ∂x 2μs (3) mean and fluctuating velocities and pressure can be represented as: u x,y,z, t u x,y,z u x,y,z, t and p x,y,z, t p x,y,z p x,y,z, t . mean velocity is defined as: u u dt substitute in mean and fluctuating variables and expand to get the form of eqn. 4: ρ ∂ u u′ ∂t ∂ u u′ . u u′ ∂x ∂ p p′ ∂x ∂ ∂x μ ∂ u u′ ∂x ρ ∂u ∂t ∂ u′ ∂t ∂ u u ∂x ∂ u u′ ∂x ∂ u u′ ∂x ∂ u′ u′ ∂x ∂p ∂x ∂p′ ∂x ∂ ∂x μ ∂u ∂x ∂u′ ∂x (4) u u u u   u u u u .u u .u 0 ∂u ∂x ∂u ∂x u u u u 0 ∂u ∂t ∂ ∂x u ∂ ∂x u u′ ∂ ∂x u 2uu u′ many terms cancel to give reynolds averaged navier stokes (rans) equations as eqn. 5: ρ ∂u ∂t ∂ u u ∂x ∂p ∂x ∂ ∂x μ ∂u ∂x ρu u ρ ∂u ∂t ∂ u u ∂x ∂p ∂x ∂ ∂x 2 μ s ρu u (5) the term ( ρu u ) is named reynolds stresses. reynolds stresses are positive because the velocity fluctuations are correlated through conservation of mass such that u u 0. reynolds stresses include 9 elements (only 6 independent) due to τ τ ρu u (where i = j for normal stress) as eqn. 6: ′ ′ = ′ ′ ′ ′ ′ ′ ′ ̅ ′ ′ ′ ′ ′ ′ ′ ′ (6) there are 7 unknowns (6 unknowns for velocities and one for pressure) but 4 equations (3 navier stokes and 1 continuity). therefore there 3 transfer equations such as turbulent models such as k-ε. the eddy viscosity is also commonly called the turbulent viscosity and it is normally written as μ . turbulent viscosity is defined based on reynolds stress as: τ ρu′v′ μ . therefore τ τ τ μ μ and μ=μ μ , where μ is viscosity of laminar flow and μ is turbulent viscosity. for calculation of μ two kinds of turbulent models are used: 1) eddy viscosity essential equations such as: boussinesq, speziale, launder and 2) eddy viscosity models included 6 models: standards k-ε, extended k-ε, rng k-ε, anisotropic k-ε, wilcox k-ω and sst k-ω. in this study, the equations employed are: boussinesq and standards k-ε model. boussinesq equation is stated as: 2 (7) finally, for calculation of μ the model standards k-ε uses two parameters of k and ε: , , , , (8) 14 out of axis movement of an auv inside a water pipeline   14   all notations of these equations are represented by sanieenezhad m (2003). 4.2 condition “a” first, the resistance of the auv at the center of the pipe is analyzed. under this condition the lift force is zero, due to symmetry. the analysis is performed by flow vision (v.2.3) software based on cfd method and solving the rans equations. generally, the validity of the results of this software have been verified using several experimental test cases. nowadays the software is accepted as a practical and reliable tool in cfd activities. the present research utilizes finite volume method (fvm) in order to model the cases. a structured mesh with cubic (hexahedral) cell has been used to map the space around the auv. transition of laminar layer to the turbulent layer in the boundary layer, and the flow separation are very determining factors in calculating the resistance. the two significant parameters in cfd for modeling the boundary layerare y+ and mesh numbers, which to be selected with great care. for modeling the boundary layer near solid surfaces, the first cell height was placed such that the y+ is kept in the range of 30~100. it must be clarified that to reach a fine cells near the wall, each mesh near the solid surface is divided into eight smaller meshes. it means that each dimension of hexahedral mesh is divided into two from the middle, and therefore, each mesh converts to eight smaller fine meshes. in order to select the proper quantity and dimensions of the cells, for one specific condition (d/d=1.41, a2/a1=1) and v=3m/s, seven different type of meshes were selected and the results of the resistance forces were compared. the results remained almost constant after 1.9 millions meshes, showing that the results are independent of meshing (fig. 6). finally all models the mesh numbers considered are more than 2.1 million.     figure 6. mesh independency evaluations. due to the axis-symmetric shape and the axis-symmetry of the flow current, a quarter of the body and the domain are modeled. in the domain, there is an inlet (with uniform flow), a free outlet, a symmetry (in the two faces of the symmetric plane) and a wall (for the body of auv and for the pipe interior boundary) (fig. 7). in the present study it is supposed that the water inside the pipe remains calm having no speed, and only auv moves. the turbulence model is k are turbulent scale is considered 0.1m and y+ 30~100. the considered flow is incompressible fluid (fresh water) at 20 degrees centigrade and velocities of 1, 3, 10m/s. settings of the simulation are collated in table 1. (a)   figure 7. (a) domain and general dimensions (b) structured grid and very tiny cells near the wall for boundary layer modeling. 16 m. moonesun, a. mahdian, y.m. korol and a. brazhko   17   table 1. settings of the simulation (inside the pipe). elements boundary conditions descriptions domain cylinder (quarterly) conditions fully submerged modeling (without free surface) quarter and half modelingdomain with inlet, outlet, symmetry and wallwithout heat transfer. dimensions 7l (distance before: 2ldistance after: 4l)*16r*16r grid structured gridhexahedral cellsfine cell near wall meshes more than 2.1 millions. settings iterations more that 700time step=0.01 s. fluid incompressible fluid (water)reynolds number, is different in each pipeturbulent modeling: standard k-εfresh watertempreture:20 degρ=999.841 kg/m3. object wall bare hull of auvvalue 30<y+<100 roughness=0no slip input inlet velocity=1,3,10m/snormal (along x)in 1 face output free outlet zero pressurein 1 face symmetry symmetry in 2 faces boundaries wall for modeling the pipe wallno slip condition the selection of the proper "time step" for each iteration, depends upon three parameters: 1) speed, 2) model length, and 3) mesh numbers along the main direction of the movement, as the transfer of network should be stopped on each section. for example, if v=1m/s, then the boundary layer will pass 2m the length of the body in 2 seconds. the direction of velocity is along the axis, and we considered every 1 cm, one station of mesh, which means 200 longitudinal station along the body (not the whole domain). in order to stop the steam in each station, the time step is 2/200=0.01 seconds. on the other hand, the minimum time step is governed by: body length/speed longitudinal station numbers on the body 9 to select the suitable iteration number, it should be considered that boundary layer could travel the domain, from the beginning to the end. for example, if the full length of domain is taken to be 21m, and v=3m/s, 7 seconds is needed to traverse the total length, and if "time step is taken 0.01 s", a minimum number of 7 0.01=700 iterations are needed. these conditions are collated in table 1. to analyze the pipe wall effect upon resistance, simulating the free stream, without the pipe and the wall effect is needed. the domain and the simulation are shown in fig. 8. in the domain, there is an inlet (having uniform flow), free outlet, asymmetry (in the four faces of the box) and a wall (for the submarine body). the domain is a box with the dimensions of 12*2*2 m (or 6l*16r*16r). 4.3 condition "b" in this part, the out of axis position of auv in the pipe is analyzed. here, the quarter-model cannot be used, therefore, the half domain model is utilized (fig. 9). other conditions of the model remains same as that of "condition a". figure 8. free stream modeling (without pipe wall effects). out of axis movement of an auv inside a water pipeline   18   5. cfd results and analysis 5.1 pressure variation in this case, there are two main factors: 1) pipe wall effect of the pipe and 2) change in flow speed. the pipe wall effect gives rise to boundary layer effects on the fluid, and a zero speed just on the wall. fluid speed change appears due to the limited cross section area between the body and the pipe. the minor distance between the body and the pipe means minor cross section area (a2). the flux of fluid is kept constant, therefore the fluid speed should be increased. an increase in the speed, according to bernoulli's law, is equivalent to a decrease in pressure. for smaller values of a2, the change in the fluid speed arising from the flux is more than the effect of boundary layer at the pipe wall. in next step, the variation in resistance and lift force of the model (auv) is be discussed. the total resistance (rt) is the summation of pressure resistance (rp) and viscous resistance (rf). pressure resistance presents the effect of the shape of the body on the total resistance. on the other hand, this type of resistance is not dependent on the viscosity and boundary layer concepts. in our problem, the governing factor is pressure resistance, figure 9. out of axis model in flow vision. because it wholly depends upon the pressure distribution over the body. in each stage, the total and the pressure resistance are presented. the pressure variation under the condition of out of axis as per d/d=4 and speed of 3m/s for several values of "e" is shown in fig. 10. as the figure shows, at e=0, the pressure distribution is uniform and axis symmetric. hence, the lift force is expected to be zero. by getting out of axis, the fluid velocity at the top of the body decreases and then, the pressure increases. at the low end of the body, these variations are inverse: the velocity increases and the pressure decreases. therefore, there are visible changes in the lift force. 5.2 results of analysis for condition "a" the total and pressure resistance of auv for several different diameters of tubes are presented in table 2. the results are presented for speeds of 1, 3, and 10 m/s. as is evident, at all speeds and for d/d <1.41 (or a2/a1<1), there is a jump in the resistance diagram. therefore, it can be suggested that in all the pipes or torpedo tubes, d/d >1.41 (or a2/a1>1) should be considered. in the swim out of the self-prope figure 10. pressure variation for several ratios of out of axis "e" for d/d=4 and v=3m/s. m. moonesun, a. mahdian, y.m. korol and a. brazhko   19 lled system of the torpedoes in submarines, the torpedo diameter is 533 mm and the torpedo tube (launcher) is 640 mm, which eventuates measurements of a1 = 0.223 m2 and a2 = 0.0985 m2. for easier ejection of a torpedo from a tube, a minimum tube diameter of 750 mm is recommended, which results in a1 = a2. the main limitation for increasing the tube diameter is the architectural arrangement inside a submarine or a naval ship, and the volume of water required for filling the volume between the torpedo and the tube. this volume of water must be maintained at a high level inside the submarine tanks, but in pipelines, there is no such restriction. the diameter of the auv can be designed according to the diameter of the pipe. the variation shows that, after d/d = 1.41, there is slight variation. logically, by increasing the pipe diameter, there should be a decrease in resistance values. at the speeds of 1m/s and 3m/s, the results after d/d=12 remain constant and is equal to results of large diameters. the values of "d/d=∞" is related to condition "b", which models the free stream condition. this means that after this limit, the added resistance and wall effect of pipe is negligible. this diameter is "the critical diameter". at the speed of 10m/s, the critical diameter occurs at d/d=13. since there exists no high speeds in pipes, it can be concluded that d/d=12 is related to the critical diameter. the ratio of the pressure resistance to total resistance is shown in fig. 11. as mentioned above, the pressure resistance has a unique role in determining total resistance. this diagram shows that in value of d/d<1.41, the pressure resistance is about 90 percent of total resistance. in large diameters, this value is about 45 percent. this means that the wall effect of the pipe, has brought about twice as much pressure resistance. by increasing the diameter, the portion of the pressure resistance will be decreased gradually. to interpolate the other values of d/d, the best curve fitted to these points is provided by curve expert software and least mean square (lms) method. the extracted formula is: figure 11. ratio of pressure resistance to total resistance: (a) at different velocities and (b) average value. table 2. values of total resistance and pressure resistance in several d/d at e=0.      d/d v=1 m/s v=3 m/s v=10 m/s rt (n) rp (n) rt (n) rp (n) rt (n) rp (n) 1.22 145.2 125.2 1110.8 963.6 12820 11624 1.41 34.8 26.4 258 198 2392 1838 1.58 22 15.6 160.8 114 1099.2 820 1.73 15.2 10 115.2 76 1075.2 715.6 2 10.4 6 78.4 46.4 741.2 445.6 3 7.2 4 50.4 26 490 266 4 6 3 46 23.6 444 238 10 6 3.36 37.6 16.8 358 172 12 4.6 2 36.8 17.2 352.4 166 13 4.6 2 36.6 16.8 348 158 infinity 4.6 2 36.6 16.8 348 158 rp/rt average rp/rt v=1 m/sec v=3 m/sec v=10 m/sec r p /r t [] a v er ag e r p /r t [] d/d [-] (a) (b) out of axis movement of an auv inside a water pipeline   22   figure 12: ratio of total resistance to free stream resistance. 43.5 / 0.61 / 10 for a better understanding of pipe wall effect on the resistance, the ratio of total resistance to free stream resistance is presented in fig. 12. the amount of rt/r0=1 shows that the wall effect is negligible, and the free stream condition is established. to interpolate the other values of d/d, according to the best fitted curve by least mean square (lms) method, the related formula is as follows: 1.01 1 2.06 . 11   as mentioned above, the ratio of d/d=12 can be regarded as critical diameter, neglecting the added resistance, and the pipe wall effect. 5.3 results of analysis for condition "b" the off-axis movement effect is studied for two values of d/d=4 and 12, at speeds of 1,3,10 m/s for values of e=0, 0.25 and 0.75. the compared results are shown in figs. 13 and 14. for d/d=4, the variation of resistance at low and medium speeds are very little, but for e=0.75, it is considerable, as the body gets very figure 13. diagram of hydrodynamic forces for d/d=4 at different velocities. close to the pipe wall. the changes in the lift force is very stiff owing to the asymmetric distribution of pressure, above and beneath the body. for every step of "e", approximately 300 percent of increase is found, except at "e=0.75", which is different, because at this position, because the body becomes tangent to the pipe wall. this results in changes in the pressure domain. as was shown earlier, at d/d=12, a free stream at the center of the axis can be represented, but by getting out of the center of the pipe, the free stream becomes in valid. as before, the variation in resistance all speeds and all values of "e" is negligible. the changes in the lift force is very stiff too. for every step of "e", found approximately 3-10 times increase in the lift force is observed. rt/ro line of free streamer at rt/ro v=1 m/sec v=3 m/sec v=10 m/sec average of rt/r0 (vs) d/d 20 d/d [-] v [m/s]   v [m/s] v=1 m/s v=3 m/s v=10 m/s v [m/s] m. moonesun, a. mahdian, y.m. korol and a. brazhko   21 it can be concluded that the out-of-axis movement has little effect on the resistance, because the principal factor is the pressure figure 14. diagram of hydrodynamic forces for d/d=12. resistance. the pressure resistance is a function of fluid velocity around the body, which is the function of the flux. by changing the location of auv inside the pipe, the fluid flux remains constant. therefore, the changes in the pressure resistance and subsequently the total resistance is insignificant. the out-of-axis movement imposes a stiff variation in the lift force of the body. it is imperative neutralize this negative lift force in order to prevent hitting to the wall and the consequent damage. for this purpose, the hydroplanes of the auv should be designed in such a way to produce inverse lift to neutralize the body lift force. that is, the hydroplanes should be designed to be able to produce such a big lift force. it is vital to keep the equilibrium and stability of the auv while operating inside a pipe. 6. conclusion in conclusion, as regards to the added resistance of an auv moving inside a pipe, arising from the wall effect, the followings can be suggested: 1the ratio of d/d=12 can be regarded for the determination of "the critical pipe diameter" with added resistance zero. 2the ratio of d/d<1.41 (or a2/a1<1) causes a stiff increase along with the resistance. therefore, the values lower than this ratio are not recommended. 3in the swim out launching system, for 533mm torpedoes, the minimum tube diameter of 750mm is recommended. 4pressure resistance has a crucial role in the total resistance of about 45~90 percent. 5if there is an initial water speed inside the tube, the entrance length and the developed boundary layer has to be taken into account. 6the fully developed condition of the stream, after the entrance length, provides about 5 percent decrease in the total resistance and a 10~20 percent decrease in the pressure resistance. 7out-of-axis movement of the auv inside a pipe causes some change in resistance force. 8out-of-axis movement of the auv inside a pipe imposes a stiff increase in lift force. hydroplanes of auv should be able to neutralize this body lift force. references hirose s, ohno h, mitsui t, suyama k (1999), a direct pneumatic stepping motor for robots: design of inpipe inspection vehicles for ø25,ø50,ø150 pipes. ieee international conference on robotics and automation. 23092314. joubert pn (2004), some aspects of submarine design: part 1: hydrodynamics, australian department of defense. joubert pn (2004), some aspects of submarine design: part 2: shape of a submarine 2026, australian department of defense. kormilitsin yn, khalizev oa (2001), theory of submarine design. saint petersburg state maritime technical university, russia 185221. v m/s t o ta l r es is ta n ce ( n ) v m/s v m/s p re ss u re r es is ta n ce ( n ) l if t f o rc e (n ) out of axis movement of an auv inside a water pipeline   22   moghaddam m, hadi a (2005), control and guidance of a pipe inspection crawler (pic), 22nd international sysnposium on automation and robotics in construction 1-5. moonesun m, javadi m, charmdooz p, korol um (2013), evaluation of submarine model test in towing tank and comparison with cfd and experimental formulas for fully submerged resistance. indian journal of geomarine science 42(8): 1049-1056. moonesun m (2014), introduction of iranian hydrodynamic series of submarines (ihss). j. of taiwan society of naval architecture and marine engineering 33(3): 155-162. moonesun m, mikhailovich ky, tahvildarzade d, javadi m (2014), practical scaling method for underwater hydrodynamic model test of submarine. j. of the korean society of marine engineering vol. 38(10): 850~857. muramatsu m, namiki n, koyama r, suga y (2000), autonomous mobile robot in pipe for piping operations. ieee international conference on intelligent robots and systems 2166-2171. najjaran h (2005), irc creating an underwater robotic vehicle to inspect in service water mains. construction innovation 10(2): 11. okamoto j, adamowskia jc, tsuzukia m, buiochia f, camerinib cs (1999), autonomous system for oil pipelines inspection. mechatronics. elsevier 731743(13). painumgal u, thornton b, tamaki u, nose y (2013), positioning and control of an auv inside a water pipeline for non-contact in service inspection. www.mtsjournal.org / papers / pdfs/130503-241.pdf, roh sg, ryew sm, yang jh, choi hr (2001), actively steerable in-pipe inspection robots for underground urban gas pipelines. ieee international conference on robotics and automation 1: 761-766. roh sg, choi hr (2005), differential-drive inpipe robot for moving inside urban gas pipelines. ieee transactions on robotics and automation. 21(1): 1-17. roman h, pellegrino ba (1993), pipe line crawling inspections: an overview. ieee transactions on energy conversion 8(3): sanieenezhad m (2003), an introduction to turbulent flow and turbulence modeling. cfd group publication iran. seif ms (2012), fluid mechanics. fadak publication, iran. suzumori k, hori k, miyagawa t (1998), a direct pneumatic stepping motor for robots: designs for pipe inspection microrobots and for human care robots. ieee international conference on robotics and automation 30473052. the journal of engineering research (tjer), vol. 15, no. 2 (2018) 102-113 doi: 10.24200/tjer.vol15iss2p102-113 technical loss reduction in rural areas the case of saih al khairat m.h. albadi*,a, a.h. al maghdria, a.s. al hinaia, e.f. el-saadanyb, m.s. awladthanic, s. al hinaia, a. al jabria, k. al azania, and a. al maqrashia a department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, muscat-123, sultanate of oman. b department of electrical and computer engineering, university of waterloo, 200 university ave w, waterloo, on n2l3g1, canada. c real areas electricity company, p.o. box 1166, a’seeb 133, sultanate of oman. received 31 october 2016; accepted 6 march 2017 abstract: this article investigates the potential for reducing technical losses in the rural area network of saih al khairat in thumrait, oman. based on the available network data, a power flow model of the system under study is built and the system performance is studied. to reduce losses and improve the voltage profile, different candidate solutions are investigated in coordination with the distribution system operator. an economic evaluation of the different options is conducted. keywords: distribution systems; rural areas; loss minimization; power flow; reactive compensation. حقلة سدح اخلريات-التقندة يف املنقطق الريافدة افقيدداحلد من امل ،*م. ح. البقدي أ ، أ. ، أ اهلنقئي س.، جثقني أوالد ص.م.، بنيعدلسا ف..إ، أ يئهلنقا ع. س، أ ملغدري. ا حأ. أ قرشيملا .وأ أ زانيالع خ.، أ جلقبريا التقندة يف شبكة املنقطق الريافدة يف سدح اخلريات يف منطقة افقيددتبحث هذه املققلة يف إمكقندة احلد من امل :امللخص واستنقدا إىل بدقنقت الشبكة املتقحة، مت بنقء منوذج لتدفق الطقية للنظقم يدد البحث وتقددم أداء .مثريت بسلطنة عمقن اجلهد، يتم البحث يف احللول املقرتحة املختلافة بقلتنسدق مع مشغل شكلوحتسني افقيددأجل التقلدل من املومن .النظقم .كمق جيري تقددم ايتصقدي ملختلف اخلدقرات .نظقم التوزيع .التافقعلي التعويض الطقية، تدفق املافقيدد، تقلدل الريافدة، املنقطق ،عالتوزي أنظمة: املافتقحدة الكلمقت * corresponding author’s email: mbadi@squ.edu.om mailto:mbadi@squ.edu.om m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 103 nomenclature aer authority of electricity regulation ami advanced metering infrastructure dvr dynamic voltage restorer dps dhofar power system 𝐼 current in amperes mis main interconnected system raec rural areas electricity company npv net present value omr omani riyals p real power in watt pf power factor pp payback period in years q reactive power in var s apparent power in va upfc unified power-flow controller 𝐸𝑠𝑜𝑢𝑟𝑐𝑒 measured supplied energy in kwh 𝐸𝑙𝑜𝑎𝑑 measured consumed energy in kwh |𝑉𝑖 | voltage magnitudes at bus 𝑖 in per unit 𝛿𝑖 voltage angle at bus 𝑖 in radian |𝑌𝑖𝑗 | magnitude of y-bus element in per unit 𝜃𝑖𝑗 angle of y-bus element in radian 𝑆𝐿 𝑖𝑗 losses in the branch i-j in per unit 𝐹𝐿𝑠 loss factor 𝐹𝐿𝐷 load factor 𝑃𝑉𝐴 present value of the recurring annuity a. 1. introduction losses in power systems are classified into two categories: technical losses and non-technical losses al-mahroqi, metwally et al. (2012). nontechnical losses result from actions external to power systems, for instance, human manipulation or mistakes in meter reading. electricity theft is one of the main causes of nontechnical losses in some systems antmann (2009). to reduce losses of this type, some utilities use automated meter reading and advanced metering infrastructure (ami). in addition, it is possible to use data mining and intelligence-based techniques to detect and reduce non-technical losses nagi, mohammad et al. (2008). even without ami, non-technical losses can be estimated by comparing the measured energy consumption in a feeder with the energy that the utility bills plus technical losses neto and coelho (2013). in comparison, technical losses are related to the physical characteristics and functions of the electrical network that result in the dissipation of electrical energy as heat. these types of losses occur mainly in low-efficiency equipment and in transmission and distribution lines. technical losses can be classified as generation losses, transmission losses, and distribution losses. the causes of the technical losses include a low power factor, long lines, unbalanced loading, and overloading. an assessment of technical losses can be made with engineering calculations based on the design of the system. the majority of avoidable technical losses occur where the current is high. technical losses represent economic loss for utilities as generating more energy results in higher costs of generation. in addition, as losses result in the generation of more electrical energy to satisfy the generation-load balance requirement, high technical losses contribute to global warming. according to the authority of electricity regulation (aer), the total losses (technical and non-technical) in oman’s electricity sector were estimated to be 10.2% in 2015, a decrease from 11.6% in 2014. moreover, losses in the main interconnected system (mis) decreased from 11.6% in 2014 to 10% in 2015, whereas the rural areas electricity company’s (raec’s) losses increased from 9.2% in 2014 to 10.7% in 2015 (aer 2016). the main contributions of this article are listed below: 1. modelling of a practical system in oman during peak load condition. the system is known to have an under-voltage problem. 2. simulating the system performance in terms of voltage profile and technical losses using etap® software package. 3. simulating the system performance considering five solutions that aim for reducing system losses and improving voltage profile. these options were coordinated with the network operator. 4. conducting a cost-benefit study of potential savings due to loss reduction for the considered options. these options involve adding reactive compensation elements at selected buses as well as network reconfiguration by adding a new 11 technical loss reduction in rural areas the case of saih al khairat kv feeder. more details of these options are presented in section 5. following this introduction, the paper presents a survey of sources of and mitigation techniques for technical losses in distribution systems. section 3 presents the data of the system under study. as for section 4, it presents and discusses the results of the simulation of the network performance with different options. finally, section 5 presents a summary of the main conclusions. 2. technical loss sources and mitigation techniques technical losses in a power system result naturally from current flow through resistive materials as well as the nonlinear characteristics of some equipment in the grid. the most common example of technical loss is the power dissipated in transformers and transmission lines due to their internal resistance. for example, the losses in a transmission line can be calculated by determining the difference between the measured energy injected from the source into the transmission line (𝐸𝑠𝑜𝑢𝑟𝑐𝑒 ) and the measured energy consumed (𝐸𝑙𝑜𝑎𝑑 ) by loads located at the end of that transmission line. 𝐿𝑜𝑠𝑠𝑒𝑠 = 𝐸𝑠𝑜𝑢𝑟𝑐𝑒 − 𝐸𝑙𝑜𝑎𝑑 (1) 2.1 poor power factor in general, losses that occur in conductive materials can be decreased by reducing the current or by reducing the resistance. however, reducing the current is more effective as the loss formula (i2r) shows. the magnitude of the current is a function of the apparent power (s), which, in turn, has two components: real power (p) and reactive power (q). |𝐼| =|s|/√3|v| (2) 𝑆 = 𝑃 + 𝑗𝑄 (3) the power factor (pf) is related to the cosine of the angle between the voltage and the current or the ratio of the real power to the apparent power. 𝑃𝐹 = 𝑃 𝑆 (4) 𝑃𝐹 = cos (tan−1 𝑄 𝑃 ) (5) the pf decreases as the ratio of reactive power to the real power increases. it is possible to achieve an improvement in pf by using devices, such as capacitors (switched/fixed) that deliver reactive power. a case study involving the reduction of losses using power factor correction was presented in phetlamphanh, premrudeepreechacharn et al. (2012). figure 1 is a simple illustration of the reactive compensation concept. distribution utilities require customers to maintain a good (high) power factor to reduce losses. for example, industrial customers in oman are obliged to maintain a power factor of at least 0.9 (aer 2016). 2.2 unbalanced loads distribution network losses can vary depending on the level of the balancing of the load. in an unbalanced operation mode, voltages and currents are not equally distributed between phases. different factors can result in unbalanced operation modes. they include unequal phase loading and different line parameters in different phases. unbalance commonly occurs in three-phase distribution systems. however, it can be harmful to the operation of power systems. on the generation side, current asymmetry negatively affects efficiency. unbalance reduces transmission capacity and efficiency albadi, hinai et al. (2015). in addition, it reduces the capacity and efficiency of the transformers. zero sequence current is converted into a circulating current in a delta/wye-grounded transformer, resulting in increased losses. there are several approaches to reduce the effects of unbalance. it is essential to adopt regulations and standards to ensure that all system components are designed and manufactured to be symmetrical. these components include generators, transformers, sbefore (va) saft er (va) qaft er (var) qbefore (var) p (w) figure 1. illustration of the reactive compen sation concept. 104 m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 107 transmission lines, three-phase motors, and switching equipment. in addition, imposing standards related to voltage and current unbalance levels is essential. these unbalance levels should be defined for both utilities and customers albadi, hinai et al. (2015). another approach involves revising the connection of single-phase loads on the utility and customer sides. in addition, unbalance can be reduced by using balancing equipment such as single-phase voltage regulators, a dynamic voltage restorer (dvr), surge-protection devices, unified powerflow controllers (upfc), and energy storage devices kazibwe, ringlee et al. (1990), jouanne and banerjee (2001). 2.3 transformers losses in power distribution networks, the losses in transformers can reach 3% of the total electrical power generated (ltd 2006). the transformer efficiency can be increased by reducing these losses. the losses in transformers can be classified in two different categories: the core loss or no-load loss category and the load or copper (winding) loss category al-badi, elmoudi et al. (2011). load losses are not highly sensitive to grid voltage changes, but they are highly sensitive to temperature variations. in the new distribution transformers, the secondary winding takes the form of a cylindrical sheet of aluminium. this is an important consideration in the adjustment of losses for temperature variation. 2.4 network / feeder reconfiguration in some cases, distribution network restructuring to minimize losses is highly costefficient. this option is of interest to efficiencyconscious electric utilities. distribution feeders can be reconfigured by opening and closing switches while maintaining all load requirements (ramesh, chowdhury et al. 2009). to reduce power losses and improve voltage profile, both optimal feeder reconfiguration and optimal capacitor placement has been studied extensively using different optimization techniques chang (2008). for example, genetic algorithms are used in swarnkar, gupta et al. (2010), farahani, vahidi et al. (2012). to overcome genetic algorithms drawbacks encountered in radial feeders’ capacitor placement problem, authors in fu-yuan and men-shen (2005) used evolutionary programming. in farahani, vahidi et al. (2012), simulated annealing is used for joint reconfiguration and capacitor placement optimization. an ant colony algorithm is used in kasaei and gandomkar (2010) to solve feeder reconfiguration and capacitor placement optimization problem. authors in khalil, gorpinich et al. (2013), and sedighizadeh, dakhem et al. (2014) used a selective particle swarm optimization to solve the optimization problem. 3 system data and modelling the rural areas electricity company (raec) is an omani closed joint stock company registered under the commercial companies law of the sultanate of oman. the company commenced its operations on the 1st of may, 2005, following the implementation of a decision that the ministry of national economy issued pursuant to the regulation and privatization of the electricity and related water sector law, which was promulgated by royal decree 78/2004 (raec 2016) (raec 2016). raec serves customers who are not connected to the main interconnected system (mis) and dhofar power system (dps). its license and business activities are associated with generation, transmission, and distribution (raec 2016). 3.3 system data the system under study consists of 12 diesel generators (43 mw and 81 mw) and 103 transformers, as table 1 indicates. the network has four feeders. one is a 33 kv feeder and three are 11kv feeders. the 33 kv feeder is operated and owned and operated by a large customer and was not modelled due to missing data. three types of overhead transmission lines are used in the network. they are panther, wolf, and dog acsr conductors. moreover, the underground cables in the 11kv feeders come in different sizes (50, 185, 240, and 500 mm2). in the 33 kv feeder, the underground cables have a size of 300 mm2. table 1. network data summary. component numbers diesel generators 12 transformers 95 busses 262 overhead lines (km) 175.750 cables (km) 15.790 measured peak kw 4 975 measured peak kvar 3,784 105 technical loss reduction in rural areas the case of saih al khairat there are six 11 kv/415 v transformers, as table 2 indicates. appendices a and b present the details of the loading on these transformers. the single-line diagram is presented in the appendix c, as is the per unit line data. 3.4 power flow model power flow analysis is widely used in power system operation and planning. the power flow model of the system can be built using relevant network, load, and generation data. the outputs of the power flow model include voltages at different busses and line flows in the network. these outputs are obtained by solving power balance equations: 𝑃𝑖 = ∑ |𝑉𝑖 ||𝑉𝑗 ||𝑌𝑖𝑗 | cos(𝛿𝑖 − 𝛿𝑗 − 𝜃𝑖𝑗 ) 𝑛 𝑗=1 (6) 𝑄𝑖 = ∑ |𝑉𝑖 ||𝑉𝑗 ||𝑌𝑖𝑗 | sin(𝛿𝑖 − 𝛿𝑗 − 𝜃𝑖𝑗 ) 𝑛 𝑗=1 (7) where |𝑉𝑖 | and |𝑉𝑗 | are the magnitudes of the voltage at bus 𝑖 and 𝑗, respectively; 𝛿𝑖 and 𝛿𝑗 are the associated angles; |𝑌𝑖𝑗 | is the magnitude of the y-bus element between the two busses; and 𝜃𝑖𝑗 is the corresponding angle. these power balance equations are nonlinear; therefore, iterative techniques such as the newton-raphson, the gauss-seidel, and the fast-decoupled methods are commonly used saadat (1999). in this case, etap® software package was used to model and evaluate the case study. the system losses can be calculated once the power flow problem is iteratively solved. for example, the losses in the branch i-j are the algebraic sum of the power flows. 𝑆𝐿 𝑖𝑗 = 𝑆𝑖𝑗 + 𝑆𝑗𝑖 (8) where 𝑆𝑖𝑗 = 𝑉𝑖 𝐼𝑖𝑗 ∗ and 𝑆𝑗𝑖 = 𝑉𝑗 𝐼𝑗𝑖 ∗ 3.5 load data the load data of the power flow model were validated with the recorded data as shown in table 2. transformer data summary. kva numbers x/r % z% 1,000 6 5.10 4.75 500 5 5.1 4.75 315 69 3.97 4.75 200 9 3.37 4.75 100 5 2.32 4.75 50 1 3.97 4.75 fig. 2. the difference between the measured data and that obtained from the model was about 8.7%, 2.35%, and 0.2% for feeders 1, 2, and 3, respectively. these differences were due to records missing for some of the transformers. in addition, the recorded peak loads did not occur at the same time. individual service transformer (11 kv/415 v) load data were included in the model to obtain the feeder load data. 4 results and discussion 4.1 system technical losses simulation results presented in table 3 show that technical real power losses represented about 5.18% of the generated power, whereas the reactive power losses were 11.28% of the generated power. the distributions of these losses are given in figs. 3 and 4. figure 2. recorded versus modelled data. figure 3. distribution of power losses in the system. figure 4. distribution of losses in the feeders. 0 0.5 1 1.5 2 feeder 1 feeder 2 feeder 3 p o w e r (m w ) actual 1% 76% 23% cables lines transformer 16.43 % 75.24 % 5.48 % feeder 1 feeder 2 feeder 3 106 m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 107 it was clear that most losses occurred in the overhead line. this result was attributable to the fact that overhead lines dominated the rural area system. another observation was that three quarters of the losses occurred in feeder 2. this result was attributed to the long distances connecting the scattered loads in this feeder. 4.2 voltage profile the omani distribution code mandates that the voltage profile be within 6% of the nominal value in distribution networks (33 kv, 11 kv, and 415 v) mjec, mzec et al. (2005). the omani grid code allows for variation from the nominal value of up to 10% in transmission networks (132 kv, 220 kv, and 400 kv) (oetc 2010). the voltage profile at all busses is given in fig. 5. the nominal voltage in this system was 11 kv. the voltage drop varied from one bus to another, depending on the loading of each bus and the distance from the power house. in general, the voltage drop was mostly due to the long distances between the service transformers in the network. the 6% limit that the distribution code specified was violated at many busses, especially in feeder 2. the main reasons of having this low voltage problem are the growing demand and extension of feeders to connect new customers in this small isolated system. these substantial changes in this area are due to transferring some agricultural activities from albatinah governorate to saih al khairat in al najd area, where an underground water reservoir was discovered. this transfer is aiming to conserve underground water in albatinah governorate and reduce air pollution caused by some agricultural activities. therefore, to connect new customers, feeders were extended. 4.3 candidate solutions different options for reducing losses and improving the voltage profile were studied in coordination with raec. these options are listed below: table 3. total supply, demand, and losses. supply demand losses p (mw) 4.975 4.553 0.422 q (mvar) 3.7844 3.017 0.641 s (mva) 6.092 5.325 0.7674 option 1: adding 500 kvar to the existing 500 kvar pole-mounted capacitor bank at bus 66. option 2: installing a 1 mvar pole-mounted capacitor bank in feeder 2 at bus 46 in addition to the existing 500 kvar capacitor bank at bus 66. option 3: installing two pole-mounted capacitor banks (1 mvar each), one in feeder 2 at bus 46 and the other in feeder 1 at bus 40, in addition to the existing 500 kvar capacitor bank at bus 66. option 4: installing 3 pole-mounted capacitor banks, two 1mvar capacitor banks in feeder 2 at bus 46 and one 500 kvar capacitor bank in feeder 1 at bus 40, in addition to the existing 500 kvar capacitor bank. option 5: network reconfiguration by disconnecting feeder 2 at bus 34 and connecting a new 11 kv feeder (length: 30.97 km) from the power station to the disconnected point. the candidate buses for reactive compensation options are selected based on the network voltage profile and the network topology. the sizes of switched capacitors are based on standard sizes used by raec. the suggested new feeder connection point (bus 34) is based on distribution of loads and actual network topology. the voltage profiles of these options in the peak loading condition are shown in fig. 6. 4.4 system performance with different options figure 6 presents the improvements in voltage at different locations of the network using the different proposed options. it is clear that option 2, option 3, and option 4 led to significant improvements in voltage. hence, they will result in better compliance with ±6% voltage limits, which, in turn, will result in increases in equipment lifetime and customer satisfaction. 4.5 economic evaluation of different options tables 4 and 5 show the losses and associated energy costs for the aforementioned options. the energy calculation can be performed by determining the loss factor (𝐹𝐿𝑠) using the following equation (gonen 2008). 𝐹𝐿𝑆 = 0.84 𝐹𝐿𝐷 2 + 0.16 𝐹𝐿𝐷 (9) where 𝐹𝐿𝐷 is the load factor. technical loss reduction in rural areas the case of saih al khairat 108 figure 5. voltage profile at all busses. figure 6. voltage profile for different scenarios. once 𝐹𝐿𝑠 is found, the average power loss can be calculated using the peak losses obtained from the power flow simulation. the load factor used in this study is 0.82. based on this value, the loss factor is 0.7. average loss = fls ∗ peak loss (10) the annual energy loss can be calculated using the following formula: annual energy loss = average loss * 8760 (11) raec estimated the capital cost of the different options. it was estimated that the costs of the switched capacitor banks and the 11 kv feeder were 14,000 omr/mvar and 13,000 omr/km, respectively. using these values and the results from the previous table, the annual savings, payback period, and net present value were calculated based on the 10.4% discount rate and 25-year lifetime that raec employed. the net present value (npv) was calculated for each option. the npv of a project is the difference between revenues and costs in the current monetary value. in any comparison of investment options, the project with the maximum npv is the winning one. for a recurring constant annual income, the present value can be found using the following formula: table 4. reduction in losses. scenario loss (mw) loss reduction (%) existing network 0.400 ------ option 1 0.377 5.75 option 2 0.351 12.25 option 3 0.336 16.0 option 4 0.337 15.75 option 5 0.233 41.75 60 70 80 90 100 110 b u s 1 b u s 6 b u s 1 1 b u s 1 6 b u s 2 1 b u s 2 6 b u s 3 1 b u s 3 6 b u s 4 1 b u s 4 6 b u s 5 1 b u s 5 7 b u s 6 3 b u s 6 9 b u s 7 4 b u s 8 1 b u s 8 6 b u s 9 6 b u s 1 0 1 b u s 1 0 6 b u s 1 1 1 b u s 1 1 6 b u s 1 2 1 b u s 1 6 3 b u s2 1 5 b u s2 2 0 b u s2 2 5 b u s2 3 3 b u s2 3 9 v o lt a g e ( % ) existing lower limt (%v) upper limt (%v) 60 70 80 90 100 110 b u s 1 b u s 6 b u s 1 1 b u s 1 6 b u s 2 1 b u s 2 6 b u s 3 1 b u s 3 6 b u s 4 1 b u s 4 6 b u s 5 1 b u s 5 7 b u s 6 3 b u s 6 9 b u s 7 4 b u s 8 1 b u s 8 6 b u s 9 6 b u s 1 0 1 b u s 1 0 6 b u s 1 1 1 b u s 1 1 6 b u s 1 2 1 b u s 1 6 3 b u s2 1 5 b u s2 2 0 b u s2 2 5 b u s2 3 3 b u s2 3 9 v o lt a g e ( % ) existing option 1 option 2 option 3 option 4 option 5 m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 109 𝑃𝑉𝐴 = 𝐴 (1+𝑑𝑟)𝑁−1 𝑑𝑟(1+𝑑𝑟)𝑁 (12) where 𝑃𝑉𝐴 is the present value of the recurring annuity, a. in this context, the annuity refers to the annual savings with the implementation of different options. the payback period (pp) is defined as the length of time required to recover the initial investment in a project. the shorter the length, the more economically attractive to investors the project is. another benefit of reducing losses is that it brings down fuel costs and subsidies. additionally, the reduction of losses results in improved conditions for the immediate environment due to the reduction in power generation and co2 emission. the results presented in table 6 show that any of the first four options would recover the costs within 2.5 years. the best option, which had the highest npv, was option 4. although option 5 had the largest annual savings, it was not economically attractive. this was because it was associated with a high capital cost. 5. conclusion the objective of the article was to investigate different candidate solutions for reducing technical losses and improving the voltage profile of a rural area distribution system. a model of the saih al khairat network, which the rural areas electricity company owns, was developed, and the load flow solution was obtained using etap® software package. the network data was collected from the rural areas electricity company and the equipment manufacturers. to improve the voltage profile and reduce losses, five options were studied and ranked according to their economic feasibility. the best option among them was the installation of three pole-mounted capacitor banks: two in feeder 2 and one in feeder 1. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. acknowledgment the authors would like to thank the rural area electricity company for providing the system data and for their support in achieving this work. references aer (2016), authority of electricity regulation annual report 2015. table 5. cost of losses for different scenarios. scenario average loss (mw) energy loss (kwh/year) cost (omr/year) existing network 0.28 2452800 110,376 option 1 0.2639 2311764 104,029 option 2 0.2457 2152332 96,855 option 3 0.2352 2060352 92,716 option 4 0.2359 2066484 92,992 option 5 0.1631 1428756 64,294 table 6. economic evaluation for different scenarios. scenario installation cost (omr) savings (omr/year) payback period (years) net present value (omr) option 1 7,000 6,347 1.5 48,881 option 2 14,000 13,521 1.5 105,052 option 3 28,000 17,660 2.5 127,496 option 4 21,000 17,384 1.5 132,066 option 5 402,610 46,082 24.1 3,138 technical loss reduction in rural areas the case of saih al khairat 110 al-badi a, elmoudi a, metwally i, al-wahaibi a, al-ajmi h, al bulushi m (2011), losses reduction in distribution transformers. proceedings of the international multiconference of engineers and computer scientists, citeseer. al-mahroqi y, metwally i, al-hinai a, albadi a (2012), reduction of power losses in distribution systems. world academy of science, engineering and technology 6: 498505. albadi m.h, hinai a.s.a, al-badi a.h, riyami m.s.a, hinai s.m.a, abri r.s.a (2015), unbalance in power systems: case study. ieee international conference on industrial technology (icit). antmann p (2009), reducing technical and non technical losses in the power sector. background paper for the world bank group energy sector strategy, the world bank. chang c.f (2008), reconfiguration and capacitor placement for loss reduction of distribution systems by ant colony search algorithm. ieee transactions on power systems 23(4): 1747-1755. farahani v.bm vahidi, abyaneh h.a (2012), reconfiguration and capacitor placement simultaneously for energy loss reduction based on an improved reconfiguration method. ieee transactions on power systems 27(2): 587-595. fu-yuan, h. and t. men-shen (2005), a multiobjective evolution programming method for feeder reconfiguration of power distribution system. proceedings of the 13th international conference on, intelligent systems application to power systems. gonen t (2008), electric power distribution system engineering. california, taylor and francis group. jouanne a.v, banerjee b (2001), assessment of voltage unbalance. ieee transactions on power delivery 16(4): 782-790. kasaei m.j, gandomkar m (2010), loss reduction in distribution network using simultaneous capacitor placement and reconfiguration with aant colony algorithm. 2010 asia-pacific power and energy engineering conference. kazibwe w.e, ringlee r.j, woodzell g.w, sendaula h.m (1990), power quality: a review. ieee computer applications in power 3(1): 39-42. khalil, t.m, gorpinich a.v, elbanna g.m (2013), combination of capacitor placement and reconfiguration for loss reduction in distribution systems using selective pso. 22nd international conference and exhibition on electricity distribution (cired 2013). ltd d.e.c.p (2006), best practical manualtransformers. vadodara 390007, india. mjec, mzec, medc (2005), the distribution code, version 1.000. nagi j, mohammad a.m, yap k.s, tiong s.k, ahmed s.k (2008), non-technical loss analysis for detection of electricity theft using support vector machines. 2008 ieee 2nd international power and energy conference: pecon, 1(3): 907-912. neto e.a.c.a, coelho j (2013), probabilistic methodology for technical and nontechnical losses estimation in distribution system. electric power systems research 97: 93-99. oetc (2010). the grid code, version 2. phetlamphanh v.s, premrudeepreechacharn, ngamsanroaj k (2012), technical losses reduction of electrical distribution system in vientiane capital. 2012 international conference on renewable energy research and applications (icrera). raec (2016), rural area electricity company. available: http://www.reefiah.com/." ramesh l, chowdhury s.p, chowdhury s, natarajan a.a, gaunt c.t (2009), minimization of power loss in distribution networks by different techniques. international journal of electrical, computer, energetic, electronic and communication engineering 3(4): 7. saadat h (1999), power system analysis, mcgrawhill. sedighizadeh m, dakhem m, sarvi m, kordkheili h.h (2014), optimal reconfiguration and capacitor placement for power loss reduction of distribution system using improved binary particle swarm optimization. international journal of energy and environmental engineering 5(1): 1-11. swarnkar a, gupta n, niazi k.r (2010), optimal placement of fixed and switched shunt capacitors for large-scale distribution systems using genetic algorithms. 2010 ieee pes innovative smart grid technologies conference europe (isgt europe). http://www.reefiah.com/ m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 109 appendix a: branch data table a1: the per unit system branch data from to r (pu) x (pu) b (pu) m1 457 0.326492562 0.521906224 1.36722e-05 457 213 0.056652893 0.091773945 5.74309e-07 457 33 0.673791736 1.091498115 6.83045e-06 33 215 0.064961983 0.105234123 6.58541e-07 33 35 0.574082645 0.929975972 5.81966e-06 35 217 0.389771901 0.631404739 3.95124e-06 35 38 0.143520661 0.232493993 1.45492e-06 38 39 0.034324793 0.057795373 5.76642e-06 39 40 0.270347107 0.568832638 3.79526e-06 40 44 0.123909091 0.260714959 1.73949e-06 44 45 0.054208264 0.061759603 3.4956e-07 45 47 0.255230579 0.290784798 1.64585e-06 47 49 0.347836364 0.396290786 2.24301e-06 44 50 0.074536364 0.084919454 4.80645e-07 50 233 0.422372727 0.48121024 2.72366e-06 44 51 0.101380165 0.213312239 1.42322e-06 51 52 0.006758678 0.014220816 9.48815e-08 52 54 0.074049587 0.064793388 3.90785e-05 54 236 0.016198347 0.018016529 1.23554e-05 54 235 0.026727273 0.029727273 2.03863e-05 52 55 0.234094215 0.21715301 0.000103544 55 56 0.195702479 0.171239669 0.000103279 56 59 0.032793388 0.028694215 1.73062e-05 55 57 0.221350413 0.252185046 1.42737e-06 57 239 0.170082645 0.102396694 4.97665e-05 57 58 0.584997521 0.66648905 3.77234e-06 58 242 0.149072727 0.169838908 9.6129e-07 40 41 0.328826446 0.365735535 0.000250814 41 42 0.151272727 0.132363636 7.98318e-05 42 221 0.128 0.112 6.75499e-05 42 224 0.058181818 0.050909091 3.07045e-05 m1 71 0.498016529 0.799352537 1.60206e-05 71 72 0.883785124 1.431673536 8.95922e-06 72 73 0.146542149 0.237388603 1.48555e-06 72 74 0.166181818 0.269203571 1.68464e-06 71 70 0.072515702 0.117470649 7.35115e-07 70 62 0.045322314 0.073419156 4.59447e-07 62 63 0.15107438 0.244730519 1.53149e-06 63 64 0.083090909 0.134601785 8.4232e-07 64 67 0.203950413 0.330386201 2.06751e-06 63 61 0.083090909 0.134601785 8.4232e-07 61 58 0.092155372 0.149285617 9.34209e-07 58 59 0.048343802 0.078313766 4.90077e-07 58 57 0.099709091 0.161522143 1.01078e-06 77 78 0.355024793 0.57511672 3.599e-06 57 51 0.696452893 1.128207693 7.06017e-06 51 52 0.101219835 0.163969448 1.0261e-06 51 46 1.661818182 2.692035709 1.68464e-05 46 47 0.400347108 0.648535875 4.05845e-06 47 48 0.203950413 0.330386201 2.06751e-06 48 50 0.15107438 0.244730519 1.53149e-06 48 49 0.007553719 0.012236526 7.65745e-08 46 45 0.377685951 0.611826298 3.82873e-06 45 44 0.190353719 0.374437694 2.34318e-06 44 44g 0.028704132 0.046498799 2.90983e-07 44g 44gh 0.012247934 0.003421488 1.01281e-06 44g 43 0.090644628 0.146838311 9.18894e-07 43 34 0.490991736 0.795374187 4.97734e-06 34 35 0.045322314 0.073419156 4.59447e-07 35 36 0.21452562 0.347517337 2.17472e-06 36 37 0.021150413 0.034262273 2.14409e-07 37 38l 0.045322314 0.073419156 4.59447e-07 38l 39 0.039279339 0.063629935 3.98187e-07 38l 40 0.024171901 0.039156883 2.45038e-07 40 42 0.045322314 0.073419156 4.59447e-07 40 41 0.199418182 0.323044285 2.02157e-06 34 32 0.019639669 0.031814967 1.99094e-07 32 40 0.249272727 0.403805356 2.52696e-06 32 31 0.241719008 0.39156883 2.45038e-06 31 30 0.143520661 0.232493993 1.45492e-06 30 29 0.203950413 0.330386201 2.06751e-06 29 26 0.211504132 0.342622727 2.14409e-06 26 27 0.188842975 0.305913149 0 27 28 0.001510744 0.002447305 1.53149e-08 26 25 0.037768595 0.06118263 3.82873e-07 25 17 0.007553719 0.012236526 7.65745e-08 17 18 0.042300826 0.068524545 4.28817e-07 18 19 0.009064463 0.014683831 9.18894e-08 19 20 0.051365289 0.083208376 5.20707e-07 20 21 0.13898843 0.225152077 1.40897e-06 20 22 0.099709091 0.161522143 1.01078e-06 22 23 0.086112397 0.139496396 8.72949e-07 23 24 0.143520661 0.232493993 1.45492e-06 17 16 0.045322314 0.073419156 4.59447e-07 16 15 0.06345124 0.102786818 6.43226e-07 15 14 0.045322314 0.073419156 4.59447e-07 14 14h 0.001510744 0.002447305 1.53149e-08 14 12 0.061940496 0.100339513 6.27911e-07 12 13 0.339917355 0.550643668 3.44585e-06 12 9 0.078558678 0.12725987 7.96375e-07 9 10 0.046833058 0.075866461 4.74762e-07 10 11 0.256826446 0.416041882 2.60353e-06 9 7 0.030214876 0.048946104 3.06298e-07 7 8 0.069494215 0.112576039 7.04485e-07 7 6 0.052876033 0.085655682 5.36022e-07 6 5 1.6210281 2.625958469 1.64329e-05 5 4l 1.577216529 2.554986618 1.59888e-05 4l 3 0.175246281 0.283887402 1.77653e-06 3 2 0.080824793 0.130930828 8.19347e-07 2 1 0.114723141 0.18891979 3.0825e-06 m1 80 0.33938843 0.542385492 1.44125e-05 80 81 0.148052893 0.239835909 1.50086e-06 81 82 0.128413223 0.208020941 1.30177e-06 81 83 0.389771901 0.631404739 3.95124e-06 83 163 0.166181818 0.269203571 1.68464e-06 84 85 0.206971901 0.335280811 2.09814e-06 85 87 0.067983471 0.110128734 6.89171e-07 85 86 0.086112397 0.139496396 8.72949e-07 84 88 0.317256198 0.51393409 3.21613e-06 88 95 0.261358678 0.423383798 2.64948e-06 88 84 0.093666116 0.151732922 9.49524e-07 80 96 0.61185124 0.991158602 6.20253e-06 96 97 0.250783471 0.406252662 2.54227e-06 96 98 0.077047934 0.124812565 7.8106e-07 98 162 0.054386777 0.088102987 5.51336e-07 98 99 0.211504132 0.342622727 2.14409e-06 99 100 0.018884298 0.030591315 1.91436e-07 99 101 0.098198347 0.159074837 9.95469e-07 101 102 0.024171901 0.039156883 2.45038e-07 101 103 0.419986777 0.680350843 4.25754e-06 103 104 0.21452562 0.347517337 2.17472e-06 104 105 0.392793389 0.636299349 3.98187e-06 104 106 0.096687603 0.156627532 9.80154e-07 106 107 0.022661157 0.036709578 2.29724e-07 106 108 0.543867769 0.881029868 5.51336e-06 108 109 0.312723967 0.506592174 0 108 110 0.045322314 0.073419156 4.59447e-07 110 111 0.249272727 0.403805356 2.52696e-06 110 110h 0.03172562 0.051393409 3.21613e-07 103 112 0.403368595 0.653430486 4.08908e-06 112 113 0.160138843 0.25941435 1.62338e-06 113 114 0.007553719 0.012236526 7.65745e-08 114 115 0.126902479 0.205573636 1.28645e-06 112 116 0.090644628 0.146838311 9.18894e-07 116 117 0.225100826 0.364648473 2.28192e-06 117 118 0.01208595 0.019578442 1.22519e-07 118 119 0.013596694 0.022025747 1.37834e-07 118 120 0.191864463 0.310807759 1.94499e-06 116 121 0.143520661 0.232493993 1.45492e-06 121 122 0.280998347 0.455198765 2.84857e-06 121 125 0.746307438 1.208968764 7.56556e-06 111 technical loss reduction in rural areas the case of saih al khairat 108 appendix b: transformer loading data table b1: feeder 1 data sl. no tx. no. tx. kva load (kw) load (kvar) 1 58 315 38 29 2 44 315 36 29 3 34 315 34 26 4 21 200 34 25 5 7 315 34 25 6 92 315 67 51 7 93 315 67 50 8 98 315 8 6 9 64 315 67 50 10 66 315 67 50 11 97 315 67 50 12 3g 1000 159 121 13 1g 1000 159 119 14 1g 1000 159 119 15 88g 500 40 30 16 72 315 3 2 17 5g 1000 158 121 18 65 315 35 26 19 67 315 102 77 20 1g 1000 81 61 21 4g 500 81 60 22 3g 1000 33 25 23 2g 500 81 60 table b2: feeder 3 data sl. no tx. no. tx. kva load (kw) load (kvar) 1 83 315 4 3 2 82 315 3 2 3 28 315 4 3 4 18 315 54 40 5 29 315 14 11 6 71 315 2 1 7 59 315 15 11 8 35 68 40 30 9 17 315 40 30 10 16 315 71 54 11 27 315 27 20 12 79 315 33 25 13 81 315 1 1 14 9 315 38 28 15 80 315 16 12 16 104 315 68 51 17 23 315 44 33 18 69 315 127 95 19 26 315 30 22 20 60 315 145 109 21 61 315 88 66 22 53 315 53 40 23 10 315 67 51 table b3: feeder 2 data sl. no tx. no. tx. kva load (kw) load (kvar) 1 14 500 183 137 2 13 100 14 11 3 12 100 0 0 4 50 315 12 9 5 38 200 19 15 6 45 200 40 30 7 2 200 30 22 8 24 200 75 56 9 3 50 7 5 10 19 315 0 0 11 54 315 76 8 12 95 100 0 0 13 39 200 11 34 14 55 315 22 20 15 4a 315 45 16 16 4 315 27 20 17 78 315 21 16 18 1 315 42 31 19 101 200 7 5 20 48 315 55 42 21 11 315 27 20 22 25 315 21 16 23 6 315 99 74 24 49 315 23 17 25 70 315 0 26 22 200 91 68 27 37 315 36 27 28 5 500 94 71 29 5a 200 89 67 30 107 315 0 31 32 315 70 53 32 36 315 11 8 33 57 315 0 0 34 33 315 64 48 35 73 100 39 30 36 41 315 41 31 37 40 315 67 50 38 46 315 51 38 39 15 315 43 32 40 42 315 41 31 41 68 315 18 13 42 62 315 31 24 43 51 315 10 7 44 56 315 0 0 45 43 315 8 6 46 63 315 36 27 47 20 315 113 85 48 30 315 67 50 49 31 315 24 18 112 m.h. albadi, a.h. al maghdri, a.s. al hinai, e.f. el-saadany, m.s. awladthani, s. al hinai, a. al jabri, k. al azani, and a. al maqrashi 109 appendix c: single-line diagram figure c1: feeder 1 sld. figure c2: feeder 3 sld. figure c3: feeder 2 sld. 113 civl050525.qxd the journal of engineering research vol. 4, no.1 (2007) 37-47 1. background it is a well known belief that the traffic system depends on three elements: engineering, enforcement, and education. these three elements have been developed very much during the few last decades. this paper discusses the elements of traffic enforcement, in particular, the development of a relatively new technique, in this part of the world, using cameras at signalized intersections to snap red light phase violators (it will be referred to as "red light camera" in this paper). ________________________________________ * corresponding author's e-mail: aaljanahi@eng.uob.bh as the red light camera is a relatively new technique in the region, there have been lots of discussions and debates about its efficiency in reducing the number of the red light violators. looking into the experiences of other countries might give a more comprehensive view. the red light camera was found to be very effective in virginia, usa (ruby and hobeika, 2003). the violation rates were reduced by 36% during the first three months of installation, and even, improved to 69% after six months of camera operation. another encouraging result was reported from california, usa (retting et al. 1999). a). the red light violation rates were reduced by approximately 42%. the other good revelation was that even at the nonevaluation of red light camera enforcement at signalized intersections abdulrahman aljanahi department of civil and architectural engineering, director, centre for transport and road studies (ctrs), college of engineering, university of bahrain, p.o. box. 32038 isa town, kingdom of bahrain received 25 may 2005; accepted 9 may 2006 abstract: the study attempts to find the effectiveness of adopting red light cameras in reducing red light violators. an experimental approach was adopted to investigate the use of red light cameras at signalized intersections in the kingdom of bahrain. the study locations were divided into three groups. the first group was related to the approaches monitored with red light cameras. the second group was related to approaches without red light cameras, but located within an intersection that had one of its approaches monitored with red light cameras. the third group was related to intersection approaches located at intersection without red light cameras (controlled sites). a methodology was developed for data collection. the data were then tested statistically by z-test using proportion methods to compare the proportion of red light violations occurring at different sites. the study found that the proportion of red light violators at approaches monitored with red light cameras was significantly less than those at the controlled sites for most of the time. approaches without red light cameras located within intersections having red light cameras showed, in general, fewer violations than controlled sites, but the results were not significant for all times of the day. the study reveals that red light cameras have a positive effect on reducing red light violations. however, these conclusions need further evaluations to justify their safe and economic use. keywords: red light cameras, signalized, intersection, transportation, violation, accident ájqhôe ägqé°téh iohõÿg äé©wé≤àdg ≈∏y áñbgôÿg ägòeéc ≥«ñ£j º««≤j »ëæ÷g øªmôdg~ñy áá°°uuóóÿÿggüƒ∏°sg ´éñjg ” .úødéıg ú≤fé°ùdg πñb øe agôª◊g ägqé°t’g rhéœ iôgéx øe ~ë∏d áñbgôe ägòeéc ωg~îà°sg á«dé©a i~e áaô©ÿ ádhéfi øy iqéñy á°sgq~dg √òg : äé©wé≤j â∏ª°t ¤h’g áyƒªûg :äéyƒª› çók ¤g á°sgq~dg ™bgƒe ⪰ùb .øjôëñdg áµ∏‡ ‘ ájqhôe ägqé°téh √ohõÿg äé©wé≤àdg ‘ áñbgôÿg ägòeéc ωg~îà°sg á°sgq~d »ñjôœ …g ¿h~h é©wé≤j â∏ª°ûa áãdéãdg áyƒªûg éeg ,§≤a ájgqé°ùe øe ~mgƒd áñbgôe ägòeéc ≈∏y …ƒàëj é©wé≤j â∏ª°t á«féãdg áyƒªûg ,é¡jgqé°ùe ™«ªl ‘ áñbgôe ägòeéµh iohõe â∏é°s ägòeéµdéh é¡àñbgôe â“ »àdg ägqé°ùÿg ¿g âë°vhg á°sgq~dg √òg èféàf ré«fé°ümg äéfé«ñdg √òg π«∏– ” ∂dp ~©h ºk äéfé«ñdg ™ª÷ á≤jôw çg~ëà°sg ” .áñbgôe ägòeéc .agôª◊g iqé°s’g rhéœ øy áœéædg ájqhôÿg äéødéıg ‘ ≈∏y’g ∫~©ÿg πé°s áñbgôe ägòeéµh ohõe ò¨dg ™wé≤àdg .âbƒdg º¶©e ‘ agôª◊g iqé°t’g ägrhééàd é° øîæe ’~©e i~fé°ùe çéëhg ¤g áléëh á°sgq~dg √òg ¿g ’g agôª◊g iqé°t’g rhéœ iôgéx øe π«∏≤à∏d ∫é©a »hééjg ôkg äé©wé≤àdg ≈∏y áñbgôÿg ägòeéµd ¿g ™wéb πµ°ûhh äô¡xg á°sgq~dg √òg .éjqhôeh éjoé°üàbg üƒ∏°s’g gòg á«dé©a âñãj ≈àm áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.çoém ,áødéfl ,äó°ugƒe ,™wé≤j ,qhôe ägqé°téh ohõe,agôª◊g iqé°t’g ägòeéc : 38 the journal of engineering research vol. 4, no.1 (2007) 37-47 equipped intersections with red light cameras, the violations reduced. another study (retting, et al. 1999 b) supported those findings, which reported that after installing red light cameras in fairfax, virginia, u.s.a, the situation showed a significant reduction in red light violations after one year of installation. again in san francisco, california it was reported that the red light violators dropped by 40% after 6 months of installing the red light cameras (fleck and smith, 1999). similar results were found in the united kingdom (harbord and jones, 1996). the automatic detection by red light running cameras was found to be useful on motorway m 25. on the other hand kea, et al. (2002) were more conservative in advocating the use of new technologies as the experiment was still new and needed more time to be evaluated. some other reports even went further to present conflicting results for red light violators after installing red light cameras. the before-and-after study by schattler, et al. (2003) showed mixed results, as no significant differences were found in the violation rates. miller (2003) found that the yellow change interval might have an effect on red light violations. other researchers indicated that to gain better results in cutting down the number of red light violators, education should be combined with the use of automated detectors. for example, tanawneh, et al. (1999), found that combining enforcement with education did have an impact in reducing the number of red light violations. others (blakey, 2003) determined that traffic engineers should play a vital role in selecting red light camera sites to ensure best results. this experience was supported by the case of san francisco (fleck and smith, 1999) where installing red light cameras included the interaction of engineering, enforcement, educational and legislation affairs in producing an integrated system. as was observed in some studies, the legal side of using red light cameras should be thoroughly considered. for example a red light violation camera system is still not legalized in many parts of usa (datta, et al. 2000). this opinion is echoed by lawson, et al. (1992), who advised that the operators should be aware in developing policies so as to go hand-in-hand with the legal aspects and recommended further areas of research. some researchers (wissinger, et al. 2000) in this field have been trying to explore the public acceptance of this technique. they found that there were divisions in the public opinion at taking extremes. one group of public was asking for a grace period for violators while the other extreme group was asking for a zero-tolerance for the violators. almost all the opinions were supporting an automated enforcement. according to retting, et al. (1999), the public support reached 80% for this automated enforcement. another study which was conducted in u.k showed that public support was weak for a similar device, namely a speed radar, as it was perceived as a tool to collect revenues for the local agency (keenan, 2001). the same study emphasized the proper site selection for the device as an important condition for increasing its effectiveness. regardless of public opinion, it is important to use the new technologies in traffic enforcement (magee, 1999; chen, et al. 2000). other benefits of red light cameras were experimented. red light cameras have been successfully used in data collection to assist the feasibility of installing such systems (banghman, 2002). it seems that red light violators can be grouped according to certain characteristics. it was found that drivers who were young, less likely to wear seat belts, have poorer driving records, and drove smaller and older vehicles were more likely to violate red light than other drivers (retting and williams, 1996). also, it seems that drivers tend to violate red light if they were alone compared to a situation where they were accompanied by other passengers (ponter and derry, 2001). the other results were that one of each five drivers violated red light; younger drivers tend to violate more than older drivers; and being in hurry not in stress, and were reason number one for violating. among the population surveyed, only 6% were caught and punished. only one out of ten respondents was exposed to a crash due to running red lights. 2. introduction traffic signals were first introduced in 1868 in the uk, in a primitive way and it was re-introduced in new york, usa in 1918 to regulate traffic flow (salter, 1988). they proved to enhance the flow of traffic at intersections. new generations of traffic signals were introduced with a computer processing power to give better performance and more functions. traffic signal effectiveness depends mainly on the road users, especially the drivers observing these signals. the majority of the drivers do observe, but unfortunately, a minority violates these signals. the accident severity of such violations is high (fleck and smith, 1999). the enforcement camera at signalized intersections works on the following basic principle: when the sensors (ie. loops under the surface of the road) detect the movement of a vehicle during the red-phase of that approach, the enforcement camera is activated and photographic /digital shots are taken. the focus of the red light will be on the plate number of the vehicles (ruby and hobeika, 2003). 3. objective the objective of this paper is to measure the effect of red light camera enforcement on the number of red light violators at signalized intersections. the effect was measured for different intersection approaches monitored with and without red light cameras. 4. methodology to fulfill the objective of the study, the following methodology was applied: 39 the journal of engineering research vol. 4, no.1 (2007) 37-47 4.1 the hypothesis the first hypothesis was that the presence of red light cameras at an approach located at signalized intersection would lower the number of red light violators compared to approaches monitored without red light cameras, depending on the direct and continuous enforcement provided by red light cameras. the second hypothesis was that an approach without a red light camera located within a signalized intersection with a red light camera at other approaches would have less red light violators than an approach located at a signalized intersection without using a red light camera at all, depending on the indirect enforcement effect. 4.2 selection of the sites a number of signalized intersections were chosen. the selected sites have similar geometric, traffic, and environmental conditions. figs. 1, 2, 3, and 4 illustrate plans of the sites. the sites were divided into three categories according to the hypothesis: type i (r/d and g/d: approaches located at signalized intersections with red light cameras installed. they are called "direct-effect" sites. type ii (r/id and g/id: approaches with no red light cameras, however; but located at signalized intersections that have red light cameras at another approach. they are called "indirect-effect" sites. type iii (s/n and j/n): approaches with no red light camera located at signalized intersections. they are called "no-effect" (controlled) sites. number of violators at direct-effect sites (type i) was compared to no-effect sites (type iii). the results would test the first hypothesis. another comparison to the second hypothesis would be conducted to compare the number of violators at indirect-effect sites (type ii) with no effect sites (type iii). 4.3 observations 4.3.1 observation times the observations were carried out for half an hour during the morning peak (ie. 7:00 8:00), outside morning peak, during evening peak (14:00 15:00), and during night time (ie. 21:00 22:00). each observation was repeated three times on three different working days. the observation outcomes were combined to represent each interval. at g/d and intersection no observations were taken during night time due to external factors beyond the investigator's control (no observation was allowed for direct effect approach at that site during night due to security reasons). 4.3.2 data collected total number of vehicles that passed through each phase during each time interval were recorded, tables 16. all lanes, straight and left turning, were considered in the analysis as the objective of this study concentrated on red light violations. further studies may test the effect of different lane orientation on red light violation. the proportion of red light violators as a percent of the total traffic entering the intersection at the studied approach was calculated as the number of red light violators was divided over the total number of vehicles passing through the intersection expressed as a percentage. the results are presented in tables 1 to 6. however, fig. 5 summarizes the percent of violators for all the evaluated sites. violating vehicles were defined as vehicles crossing the stopping white line at the start of the red light. all vehicles crossing the stopping white line during yellow-light were recorded even though they were not considered as violators according to the local regulations. one member of the team was responsible for observing the change of lights during the phase. 4.3.3 conditions of observations the observers were instructed to take positions far enough from the traffic so that the drivers would not be distracted by them. the observers were supplied by ready-made log sheets for data entry at each site. the tasks were distributed randomly among the observers to ensure smooth operation. the log sheet was checked in the office to ensure its accuracy. finally, the data were entered to the computer using pre-designed excel sheets. 5. analysis 5.1. variable determination for each site, number of red light violators in each interval was converted to proportions (section 4.3.2). 5.2. statistical test 1. the samples must be independent of each other which was the case here; and 2. np and nq must be 5 or more which was the case as well (the sample size n is more than 30). the formula used was: ) n 1 n 1 (pq )pp()pp( z 21 21 ^ 2 ^ 1 + −−− = z-test was used (bluman, 1992) to compare the differences statistically, if any, between the studied sites. in this case the comparison was not between absolute numbers but between percentages (i.e. percentage of drivers violating the red light), so z-test for comparing two proportions was used. the normal approximation to the binomial is applied to each population, so that the estimators of the population p1 ^ and p2 ^ had approximately normal distribution with two conditions to be fulfilled: 40 the journal of engineering research vol. 4, no.1 (2007) 37-47 figure 1. a plan for r/d and r/id site figure 2. a plan for g/d and g/id site figure 3. a plan for s/n site figure 4. a plan for j/n site 0 1 2 3 4 5 6 % o f vi ol at or s s/n type j/n type g/id type r/id type r/d type r/d type iii iii ii ii i i observbation sites figure 5. proportion of red light violators to total traffic passing through the evaluated approachs for all sites during yellow phase during red phase during yellow and red phases 41 the journal of engineering research vol. 4, no.1 (2007) 37-47 signal time morning peak morning offpeak afternoon peak night time total vehicles green time 1536 558 1068 698 3860 yellow time 13 10 7 17 47 red time 5 3 1 4 13 total veh. 1554 571 1076 719 3920 table 1. total number of vehicles passing during different traffic signal times at r/d signlized intersection (type i) signal time morning peak morning o ffpeak afternoon peak night time total vehicles green time 1309 509 555 516 2889 yellow time 24 25 15 15 79 red time 9 11 5 7 32 total veh. 1342 545 575 538 3000 table 2. total number of vehicles passing during different traffic signal times at r/id intersection (type ii) phase morning peak morning o ffpeak afternoon peak night time total vehicles green time 1066 506 517 n/a 2089 yellow time 7 21 19 n/a 47 red time 1 8 4 n/a 13 total veh. 1074 535 540 n/a 2149 table 3. total number of vehicles passing during different traffic signal times at g/d signalized intersection (type i) signal time morning peak morning o ffpeak afternoon peak night time total vehicles green time 949 588 688 293 2518 yellow time 31 26 18 3 78 red time 8 5 13 11 37 total veh. 988 619 719 307 2633 * n/a: not applicable table 4. total number of vehicles passing during different traffic signal times at g/id signalized intersection (type ii) signal time morning peak morning o ffpeak afternoon peak night time total vehicles green time 358 334 333 455 1480 yellow time 9 8 11 8 36 red time 13 6 5 5 29 total veh. 380 348 349 468 1545 table 5. total number of vehicles passing during different traffic signal times at s/n signalized intersection (type iii) phase morning peak morning off peak afternoon peak night time total green 849 583 548 282 2262 yellow 40 25 16 11 92 red 21 12 3 6 42 total 910 620 567 299 2396 table 6. total number of vehicles passing during different traffic signal times at j/n signalized intersection (type iii) 42 the journal of engineering research vol. 4, no.1 (2007) 37-47 where: n1 : sample size of the first group n2 : sample size of the second group this gives: test value = where: observed value = p1^ p2^ , expected value = p1 p2 , and the test was carried out at confidence level (x) of .05 (as usually accepted for such analysis) and the critical values were + 1.96 and 1.96. 5.3 analysis methodology 5.3.1 direct-effect sites (r/d and g/d intersections) these sites were compared to no-effect sites (s/n and j/n intersection). similar observation periods were compared as well as the average proportion of red light violators for all periods at each site. the results of the test of hypothesis on both proportions are presented in tables 710. 5.3.2 indirect-effect sites (r/id and g/id intersections) these sites were compared with no-effect sites (s/n and j/n intersections) to detect the second hypothesis of psychological effect of having red light cameras at the intersection even though there was no red light camera at the approach itself. the results of the test of hypothesis on both proportions are presented in tables 11-14. 5.3.3 direct-effect sites (r/d and g/d intersections) these sites were compared with the indirect-effect sites (same) to explore the difference between approaches of the same intersection where one had a red light camera and the other one did not have one. the results of the test of hypothesis on both proportions are presented in tables 15 and 16. 5.3.4 same type sites they were compared to see if there is a significant difference concerning red light violators regardless of the presence of red light camera. the results of the test of hypothesis on both proportions are presented in tables 17 and 18. 6. results & discussion 6.1 selected same type sites they were compared to detect if there was any inherent significant difference between the proportions of red light violators which might have been the result of unforeseen factors. the results obtained were: 1. no significant difference was obtained between proportions of red light violators at the direct effect sites at both r/d and g/d intersections as presented in table 17. 2. no significant difference was found between both controlled sites (s/n and j/n intersection) as shown in table 18. 6.2 comparison between direct-effect sites (type i) and no-effect sites (type iii) the results of this comparison were: 1. proportion of red light violators at r/d site (type i) was significantly less than the proportion of red light violators at the s/n-controlled site (type iii). the reduction was more than eighty percent. the red light cameras had a significant effect on reducing red light violations. going through the periods of the day, the significant impact was noticed during morning and afternoon peak hours. it seems that driver behavior during peak hours tends to be more irritated as they rush to reach their work or drop their children at schools or when coming back from work so they take chances in violating red light. in this case, it seems where: z = test value p1 ^ = sample proportion used to estimate the first population proportion p2 ^ = sample proportion used to estimate the second population proportion p1 = probability of the first population proportion p2 = probability of the second proportion p = weighted estimate of p q = weighted estimate of q also, 21 21 nn xx p + + = 1 1 1 n x p = pq −=1 2 2^ 2 n x p = errordardnsta valuectedexpevalueobserved − standard error= ) n 1 n 1 (pq 21 + 43 the journal of engineering research vol. 4, no.1 (2007) 37-47 site morning peak morning off-peak afternoon peak night time all day regency type i 0.32175 0.525394 0.092937 0.556328 0.331633 sheraton type iii 3.421053 1.724138 1.432665 1.068376 1.877023 z-value -5.63995 -1.78999 -3.35874 -0.99387 -5.89109 significance yes no yes no yes table 7. a summary of the z-test of proportions of red light violators at r/d (type i) and s/n/ (type iii) intersections site morning peak morning off peak afternoon peak night time all day regency type i 0.32175 0.525394 0.092937 0.556328 0.331633 jufair type iii 2.307692 1.935484 0.529101 2.006689 1.752922 z-value -4.65618 -2.18005 -1.70548 -2.13707 -5.89911 significance yes yes no yes yes table 8. a summary of the z-test of proportions of red light violators at r/d (type i) and j/n (type iii) intersections site morning peak morning off peak afternoon peak night time all day gulf type i 0.09311 1.495327 0.740741 n/a 0.604933 sheraton type iii 3.421053 1.724138 1.432665 n/a 1.877023 z-value -5.70961 -0.26598 -1.00637 n/a -3.59715 significance yes yes no n/a yes table 9. a summary of the z-test of proportions of red light violators at g/d (type i) and s/n (type iii) intersections n/a: not available site morning peak morning off peak afternoon peak night time all day gulf type i 0.09311 1.495327 0.740741 n/a 0.604933 jufair type iii 2.307692 1.935484 0.529101 n/a 1.752922 z-value -4.6938 -0.57182 0.444033 n/a -3.53396 significance yes no no n/a yes table 10. a summary of the z-test of proportions of red light violators at g/d (type i) and j/n (type iii) intersections n/a: not available site morning peak morning off peak afternoon peak night time all day regency type ii 0.670641 2.018349 0.869565 1.301115 1.066667 sheraton type iii 3.421053 1.724138 1.432665 1.068376 1.877023 z-value -4.2145 0.313759 -0.80204 0.339155 -2.2489 significance yes no no no yes table 11. a summary of the z-test of proportions of red light violators at r/id (type ii) and s/n (type iii) intersections site morning peak morning off-peak afternoon peak night time all day regency type ii 0.670641 2.018349 0.869565 1.301115 1.066667 jufair type iii 2.307692 1.935484 0.529101 2.006689 1.752922 z-value -3.32515 0.101445 0.68973 -0.79104 -2.15364 significance yes no no no yes table 12. a summary of the z-test of proportions of red light violators at r/id (type ii) and j/n (type iii) intersections 44 the journal of engineering research vol. 4, no.1 (2007) 37-47 site morning peak morning off-peak afternoon peak night time all day gulf type ii 0.809717 0.807754 1.808067 3.583062 1.405241 sheraton type iii 3.421053 1.724138 1.432665 1.068376 1.877023 z-value -3.5187 -1.28973 0.447022 2.407941 -1.18064 significance yes no no yes no table 13. a summary of the z-test of proportions of red light violators at g/id (type ii) and s/n (type iii) intersections site morning peak morning off-peak afternoon peak night time all day gulf type ii 0.809717 0.807754 1.808067 3.583062 1.405241 jufair type iii 2.307692 1.935484 0.529101 2.006689 1.752922 z-value -2.65796 -1.70617 2.054346 1.174946 -0.99032 significance yes no yes no no table 14. a summary of the z-test of proportions of red light violators at g/id (type ii) and j/n (type iii) intersections site morning peak morning off peak afternoon peak night time all day regency type i 0.32175 0.525394 0.092937 0.556328 0.331633 regency type ii 0.670641 2.018349 0.869565 1.301115 1.066667 z-value -1.34983 -2.23996 -2.49844 -1.40281 -3.76983 significance no yes yes no yes table 15. a summary of the z-test of proportions of red light violators at r/d (type i) and r/id (type ii) intersections site morning peak morning of f-peak afternoon peak night time all day gulf type i 0.09311 1.495327 0.740741 n/a 0.604933 gulf type ii 0.809717 0.807754 1.808067 n/a 1.405241 z-value -2.46598 1.103645 -1.624 n/a -2.70644 significance yes no no n/a yes table 16. a summary of the z-test of proportions of red light violators at g/d (type i) and g/id (type ii) intersections n/a: not available site morning peak morning off peak afternoon peak night time all day regency type i 0.32175 0.525394 0.092937 n/a 0.331633 gulf type i 0.09311 1.495327 0.740741 n/a 0.604933 z-value 1.207262 -1.62447 -2.21174 n/a -1.559 significance no no no n/a no table 17. a summary of the z-test of proportions of red light violators at r/d (type i) and g/d (type i) intersections n/a: not available site morning peak morning off -peak afternoon peak night time all day sheraton type iii 3.421053 1.724138 1.432665 1.068376 1.877023 jufair type iii 2.307692 1.935484 0.529101 2.006689 1.752922 z-value 1.137912 -0.23357 1.427326 -1.06597 0.285958 significance no no no no no table 18. a summary of the z-test of proportions of red light violators at s/n (type iii) and j/n (type iii) intersections 45 the journal of engineering research vol. 4, no.1 (2007) 37-47 that the red light camera affected their behavior and resulted in less red light violators at the r/d intersection compared to the s/n intersection. during other periods of the day, there were no significant differences even though the proportion of red light violators at the r/d site was always less than the s/n site throughout the day. so the effect of red light camera still prevailed but not significantly. generally, the effect of red light cameras was significant through out the whole day. 2. a significant reduction in red light violators was observed at the r/d site (type i) compared to the j/n site (one without redlight cameras). the reduction was around eighty percent (table 8). the significant reduction was observed during all times of the day except in the afternoon peak. the proportion of red light violators was less at the r/d site compared to the j/n site in the afternoon peak but not significantly. 3. proportions of red light violators during all times of the day were statistically more at the s/n site (without red light cameras) compared to the g/d site (type i), except for the afternoon peak where the difference was not significant. the reduction was about two-thirds. 4. throughout the times of the day the difference between the proportions of red light violators between the g/d site and the j/n site was not significant, except for the morning peak hour. the sum of the proportion of red light violators at the g/d site was two thirds less than the juffair site (table 10). during the time of the day, the red light violators at the g/d site were less than the j/n site, except during the afternoon peak when the opposite occurred but again it was not significant. 6.3 comparison between direct-effect sites (type i) and indirect-effect sites (type ii) 1. at the r/id intersection, the approach with red light cameras (type i) showed two-thirds less red light violators than the other approach at the same intersection without red light cameras (type ii). all results were significant. through times of the day morning off-peak and afternoon peak showed significant results (table 15). 2. generally, at the g/d and g/id intersection the proportion of red light violators at the approach with red light cameras had less red light violators (reduced to half) than the other approach within the same intersection without red light cameras (type ii). the difference was statistically significant as shown in table 16. during morning off-peak period, the red light violators were less in type ii approach compared to type i approach. except for the morning peak, the other periods showed a similar pattern but all of them were insignificant. it seems that the direct presence of red light cameras has more effect than the indirect effect of red light cameras. 6.4 comparison between direct-effect sites (type ii) and no-effect sites (type iii). the comparison was carried out to test the extended effect of the presence of red light cameras at one approach on the observed proportion of red light violators at other approaches within the same intersection. 1. throughout the day the proportion of red light violators at the controlled site (s/n intersection) was higher by about 25% than the proportion at type ii (r/id). the results were significant. during the times of the day, the pattern was inconsistent and showed almost no significance as presented in table 11. 2. analysis results had the same pattern similar to that of the previous case at both r/id and j/n intersections with a reduction of about 40%. however, there was no significant impact of the times of the day on the results (table 12). 3. the total proportion of the observed red light violators was less at the g/id site compared to the s/n site, but not significantly as presented in table 13. during different times of the day, the proportion of red light violators fluctuated, with no consistent behavior as it was in the case of r/id site (type ii). 4. over all, the proportion of red light violators at the j/n site was 20% higher than the g/id site but the differences were not significant as presented in table 14. during different times of the day, no consistent pattern was observed regarding the red light violators and most of the results were not significant. generally, the presence of the redlight cameras in the intersection would have an extended effect to approaches other than the particular approach that has red light cameras, but with less significant influence. 7. conclusions conclusions derived from this study include: 1. in general the study shows that red light cameras (direct-effect) installed at signalized intersections have an effect on reducing the proportion of red light violators compared to approaches in signalized intersections without a red light camera at all (no-effect). the reduction in the proportion of red light violators averaged between 66% and 84% compared to red light violators at signalized intersections without red light cameras. these results confirm earlier findings (ruby and hobeika, 2003; retting et al, 1999 a; fleck and smith, 1999). the significant reductions in the proportion of red light violators occurred during most of the day, and especially during peak hours. 2. it seems that the effect of the presence of red light cameras at an approach within a signalized intersection influenced other approaches in the same intersection (indirect-effect) in terms of reducing the proportion of red light violators, which agreed with other studies (retting et al. 1999 b). the amount of reduction was less compared to the direct-effect approaches. the results showed less significant reductions in red light violations compared both to direct-effect 46 the journal of engineering research vol. 4, no.1 (2007) 37-47 approaches, as well as to no-effect approaches. 3. red light cameras have an impact on reducing the proportion of red light violators. the size of this impact varies during the times of the day. most of the effect could be observed during the peak hours. the effect of the presence of red light cameras might extend to approaches within the same intersection. the results were not firmly conclusive, but the indicators were observed and recorded. 8. recommendations the following recommendations are warranted of this: 1. it is strongly recommended to proceed with the installation of red light camera devices and to upgrade them with the latest technologies as to ensure better performance and reliability of the system. 2. these results might be reflecting the ambiguity of the local situation. red light cameras are a relatively new technology. drivers are not properly educated about this new system. moreover, it is still not included in the training program of new drivers. no significant awareness campaigns were launched either before or after installing the new system. 3. it is believed that more reductions in the number of red light violators would be achieved if more attention is given to education and awareness campaigns. as a result, red light cameras will be more effective in reducing red light violators. 9. scope of future work additional future work items include: 1. more sites should be evaluated once new red light cameras are installed to extend our data base. this might lead to more conclusive results. 2. the results of this study should be correlated with traffic safety records at the same intersections. 3. an economic analysis should be made to justify the installation of red light cameras. references baughman, c., 2002, "a data collection method for red light running," proceeding of the international conference on applications of advanced technologies in transportation engineering, pp. 207214, cambridge, ma, usa. blakey, l. t., 2003, "red-light cameras: effective enforcement measures for intersection safety," ite journal (institute of transport engineer), vol. 73(3), pp. 34-36, 43, usa. bluman, a. g., 1992, "elementary statistics," wm. c. brown publisher (wcb). chen, g., wilson, j., meeckle, w. and cooper, p., 2002, "evaluation of photo radar program in british columbia," accident analysis & prevention, vol. 32(4). datta, t. k., schattler k. and datta, s., 2000, "red light violations and crashes at urban intersections," transportation research record, no. 1734, pp. 52-58, washington, d.c., usa. fleck, j. l. and smith, b. b., 1999, "can we make redlight runners stop? red light photo enforcement in san francisco, california," transportation research record, no. 1693, pp. 46-49, washington, d.c., usa. harbord, b. and jones, j., 1996, "variable speed limit enforcement the m25 controlled motorway pilot scheme," proceedings of the 1997 iee colloquium on camera enforcement of traffic regulations, london, uk. kea, m. d., povey, l. j. and firth, w. j., 2002, "further results from a trial comparing a hidden speed camera programme with visible camera operation," accident analysis & prevention periodical, vol. 34(6). keenan, d. j., 2001, "measuring the effect of speed enforcement cameras on traffic speed violations," institute of transport studies (its) university of leeds, uk lawson, s. d., morris, h. t., handy, r. w. and howard, a. c., 1992, "red light and surveillance cameras policy issues related to accident reduction and enforcement," iee conference publication, no. 355, pp. 38-42, london, uk. magee, a. r., 1000, "a review of the potential benefits of speed control using att applications," institute of transport studies (its)-university of leeds, uk. miller, j. s., 2003, "two methodological considerations for evaluating safety impacts of photo-red enforcement," ite journal (institute of transportation engineers), vol. 73(1), pp. 20-24, usa. ponter, b. e. and derry, t., 1996, "a nation wide survey of self-reported red light running: measuring prevalence, predictors & perceived consequences," accident analysis & prevention: vol. 33(6). retting, r. a. and williams, a. f., 1996, "characteristics of red light violators: results of a field investigation," j. of safety research, vol. 27(1), pp. 9-15, usa. retting, r. a., williams, a. f., farmer, c. m. and feldman, a. f., 1999, "evaluation of red light camera enforcement in fairfax, virginia, usa," ite journal (institute of transportation engineers), vol. 69(8), pp. 30-34, usa. retting, r.a., williams, a. f., farmer, c. m. and feldman, a. f., 1999, "evaluation of red light camera enforcement in oxnard, california," accident analysis and prevention, vol. 31(3), pp. 169-174, usa. ruby, d.e. and hobeika, a.g., 2003, "assessment of red 47 the journal of engineering research vol. 4, no.1 (2007) 37-47 light running cameras in fairfax country, virginia," transportation quarterly, vol. 57(3), pp. 33-48, usa. salter, r. j., 1988, "highway traffic analysis and design," macmillan education, uk. schattler, k. l., datta, t. k. and hill, c. l., 2003, "change and clearance interval design on red light running and late exits," transportation research record, no. 1856, 03-2943, pp. 193-201, washington, d.c., usa. tarawneh, t. m., singh, v. a. and mccoy, p. t., 1999, "investigation of effectiveness of media advertising and police enforcement in reducing red light violations," transportation research record, no. 1693, pp. 37-45, washington, d.c., usa. wissinger, l. m., hummer, j. e. and milazzo, j. s., 2000, "using focus groups to investigate issues of red light running," transportation research record, no. 1734, pp. 38-45, washington, d.c., usa. the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 cost estimation practices prevalent in construction projects in oman m. al-siyabia, a. alnuaimib, m. khademc, and k. alzebdeh*c a technical affairs department, sultan qaboos university, p.o. box 33, al-khod, 123, muscat, oman. b department of civil and architecture engineering, sultan qaboos university, p.o. box 33, al-khod, 123, muscat, oman. c department of mechanical and industrial engineering, sultan qaboos university, p.o. box 33, al-khod, 123, muscat, oman. abstract: this paper reports findings of a study that captures cost estimation practices and techniques prevalent in the construction industry in oman. it also identifies the success and risk factors that are encountered in the cost estimation process. eighty structured questionnaires were distributed to some professionals working in construction projects, selected randomly from large and medium size enterprises across the sultanate. a response rate of 56% was achieved. the results showed that the unit rate method, the use of historical data and benchmarking are the most common methods used for cost estimation. success and risk factors as reported by respondents were analyzed and ranked accordingly. the main sources of risk were reported by respondents as “unclear reading and understanding of specifications”, “lack of clear documents of the project scope of work”, and “unavailability of resources”. this study provides an insight understanding of revealed practices and techniques and globally share the findings with researchers and professionals. keywords: construction management; cost estimation; techniques; practices; risk. مشاريع البناء في ُعمانممارسات في الكلفة تتقدير م. السيابي أ يميع. النأ , ب , م. خادم ت ة, ك. الزبد ت تقدم هذه الورقة البحثية تقارير عن نتائج دراسة للوقوف على تقدير تكلفة الممارسات والتقنيات السائدة في : الملخص صناعة البناء في سلطنة عمان. كما انها تحدد أيًضا عوامل النجاح والمخاطر التي تتم مواجهتها في عملية تقدير التكلفة. مركبا على المهنيين العاملين في مشاريع البناء، حيث تم اختيارهم بشكل ا بتوزيع ثمانين استبيانً الدراسةقمنا في هذه ٪. وأظهرت النتائج 65وكان معدل االستجابة . عشوائي من الشركات الكبيرة والمتوسطة الحجم في جميع أنحاء السلطنة لتقدير التكلفة. و حللنا عوامل أن طريقة معدل الوحدة واستخدام البيانات التاريخية والقياس هي أكثر الطرق المستخدمة النجاح والمخاطر حسب إجابات المشاركين وصنفنا النتائج وفقًا لذلك. وأفاد المجيبون على االستبيان أن مصادر الخطر ح للمواصفات" و "عدم وجود مستندات واضحة لنطاق عمل المشروع" و الرئيسية هي نتيجة "قراءة وفهم غير واض رد" كما توفر هذه الدراسة فهًما عميقًا للممارسات والتقنيات التي تم الكشف عنها ومشاركة النتائج عالميًا "عدم توفر الموا مع الباحثين والمهنيين. .طراخالم؛ ؛ الممارسات تقنياتال؛ لفةتقدير التك ؛ إدارة اإلنشاءات: المفتاحية الكلمات * corresponding author’s e-mail: alzebdeh@squ.edu.om doi: 10.24200/tjer.vol16iss2pp115-129 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 1. introduction construction project outcome is characterized in terms of time, cost and quality achieved. project clients demand their projects be completed on time, within budgeted cost and at appropriate quality or specification. project cost estimation is primarily concerned with the cost of material and human resources needed to design, construct and complete scheduled activities. in addition, the cost estimate should consider the running cost of the constructed facilities including use and maintenance. the accuracy of any construction cost estimate is measured by how well the estimated cost compares to the actual cost (oberlender and trost 2001; toutounchian et al. 2018). elfaki et al. (2014) analysed some proposals that used intelligent techniques for construction cost estimation. they collected and analysed data over a period of 10 years and suggested a methodology that can be used as a standard benchmark for construction cost estimation proposals. lim et al. (2016) studied the factors influencing cost estimation in queensland, australia and its relevance to estimation accuracy. they found that many factors of cost uncertainty for large-scale projects are not relevant for the contractors of smallscale residential construction projects. akintoye and fitzgerald (2000) studied cost estimating practices in the uk and found that the contractors use the labour and material rates to calculate prices for individual items in the bills of quantities. they also found that the major causes of inaccuracy in cost estimating are: (1) lack of practical knowledge of the construction process by those responsible for the estimating function, (2) insufficient time to prepare cost estimates, (3) poor tender documentation, and (4) wide variability of subcontractors' prices. cost estimation techniques can be categorized into qualitative and quantitative techniques (niazi et al. 2006). qualitative techniques use past historical data and judgment to subjectively estimate the cost of a project (project management institute 2008). in contrast, quantitative techniques not only rely on prior data and skilled knowledge, but can also analyze project designs, processes, and individual attributes to explore cost functions. therefore, analytical models offer a reliable way to determine project costs with a reasonable accuracy (artto et al. 2001; chou et al. 2006; touran and asce 2003; wang 2004). saha (2010) proposed a full quantitative technique to predict the potential benefits of simulation-based alternatives over conventional methods. a case study by marco et al. (2005), described the practicability and predictability of traditional estimates at completion based on early progress measurement by presenting empirical results from previous construction projects. in reference to project life cycle cost estimation, nassar (2006) recommended a three-step approach which includes: preparing background information like a system / service describing assumptions and constraint; communication and coordination considering cost and requirements analysis precedence to cost estimate preparation. the third step includes selection methods/models to collect, validate and adjust data estimate costs. kim and kim (2010) proposed a preliminary cost estimation model using case-based reasoning and genetic algorithm. they concluded that their model is more reliable than the conventional cost estimation methods. kujala et al. (2014) found that the cost management functions are impacted by the size, complexity, uncertainty and uniqueness of projects. sequeira and lopes (2015) proposed a spreadsheet method for project cost development for budget allocation. they stated that their method provides a simplified decision tool for assessing the construction cost with high accuracy. olatunji and sher (2014) compared the building information modelling (bim) techniques with the conventional cost estimation methods and concluded that estimators need to respond to the challenges that bim poses to traditional estimating practices. to enhance performance of bim, a recent study by heigermoser et al. (2019) proposed linking bim to lean construction as a construction management tool in attempt to increase productivity and reduce construction waste. wu et al. (2014) examined the cost estimating practices and procedures in the uk and analyzed the impact of using bim. they identified a number of challenges in terms of information exchange, model quality and uk standards. in addition, they developed a methodology to evaluate the ability of the existing bim technology to support the uk quantity surveying practices. several case studies in relation to the accuracy of project budgeting are reported in literature. oyedele (2015) evaluated the construction cost estimation methods in nigeria and recommended an improved procedure to achieve a quality estimate. aibinu and pasco (2008) examined the accuracy of pre-tender building cost estimates in australia, by investigating 56 projects and surveying 102 firms. their work showed that estimation in the construction industry in australia is significantly affected by project size. in small projects, cost is usually over-estimated by a large amount rather than underestimated. moreover, accuracy of estimation has not improved over time, which implies that lack of experience plays a small role in biased cost estimation. al-mohsin and al nuaimi (2013) developed a model to estimate the construction cost of villas in oman at the conceptual design stage and after the preliminary design stage. they found that their model gives good correlation between the selected variables and the actual cost with r 2 = 0.79 in the case of conceptual estimate and r 2 116 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh 0.601 for preliminary estimate. kiziltas et al. (2009) emphasized the importance of obtaining and referring to historical databases in order to reduce estimator’s errors. popovic (2004) stated that implementing data mining techniques provides a number of solutions towards improving cost estimation process in construction projects. they found that the best practice is to arrange the data in a manner per type of work and incorporate it into various contract documents. on the other hand, hwang (2009) proposed two dynamic regression models for the prediction of the construction cost index. yu et al. (2006) used web based intelligent cost estimator (wice) and achieved a higher accuracy estimate compared to existing conceptual estimation systems. jonny and lichtenberg (2016) developed a new method utilizing a successive principle, to achieve a reliable and unbiased cost in infrastructure projects in norway and denmark. recently, juszczyk (2017) proposed an approach based on the concept of nonparametric cost estimation with the use of artificial neural networks. a number of studies have utilized probabilistic estimation method (pem) to establish possible ranges of project costs, helping relevant personnel to apply pem to assess project risks (chou et al. 2009; yang 2008). the applications include a statistical framework for cost estimation in construction projects in 5 states in the usa swei et al. (2017). also, monte carlo simulation (mcs) has been applied to examine numerous project risk analyses regarding cost estimates (chou 2011). rather than estimating project costs, several researchers used an economical approach to determine life-cycle costs of infrastructure construction projects (herbold 2000; salem et al. 2003; tighe 2001; heralova 2017). the results included a decision-making process regarding pavement and rehabilitation scenarios selection. a deterministic life-cycle cost model was developed in which end-user specify the values of input parameters (salem et al. 2003). albalushi et al. (2013) blamed the absence of value engineering for cost and time overruns in projects in oman. very little information was found in the literature about actual practices or procedures used for estimating cost of construction projects in oman. the aim of this study was two folds; exploring practices and techniques prevalent in construction projects in oman, and comparing them with worldwide practices and linking the results of analysis to actual cost and time spent on the project. 2. research methodology in order to achieve the specified objectives, a framework was required to conduct the study. the overall framework that this research followed included the following elements:  identifying cost estimate practices, methods, techniques, tools and factors of project development phases worldwide.  identifying cost estimate practices, methods, techniques, tools and factors of project development phases in the sultanate of oman.  discussing and comparing the practices, methods, techniques, tools and factors in use in the sultanate of oman with those around the world.  identifying and comparing findings, and analysis of results, linking them to final cost estimate and the ending of the project.  linking the best cost estimate practices and techniques and providing recommendations for the same to be implemented in the sultanate of oman for future projects. generally, the nature of each project is an important element as it may determine when to use what practice and method or tools and to those particular practices and methods should be implemented. upon establishment of the research framework, data needed to address the research target had to be collected. initially key points were generated and checked by the research team and then a series of interviews were scheduled and conducted with various personnel in both public and private sector organizations. during the interview, the subject responses were expanded upon and further suggestions and comments were noted. the information collected from these small organizations proved to be insightful. after interviewing the smaller organizations, other larger organizations were interviewed to confirm the information previously collected. the use of a survey assisted in collecting the data needed to address the research target. surveying is a powerful technique in establishing real-world practices. once appropriate contacts were identified, they were electronically forwarded a copy of the survey. this work was followed by a pilot survey via a questionnaire conducted on 10 professionals working directly in the field of estimating construction cost and responsible for project biddings. as a result, some modifications were carried out on the questionnaire to eliminate ambiguities, potential misunderstanding or linguistic issues. the revised questionnaire was then distributed by hand to 80 professionals selected from different organizations across the whole industry ensuring that no more than one person per organization is allowed to participate. a total of 45 completed responses were received, giving a response rate of 56%. the use of a structured-questionnaire survey served as the main tool in collecting the data needed to address the research target. the responses were stored and statistically analyzed using statistical package for social sciences (spss) software. the respondents included clients, consultants and contractors. 117 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 3. analysis of data and discussion of results the data collected from interviews, recent literature review and the detailed survey responses, assisted in identifying the practices, methods and techniques used in project cost estimation in oman. based on this data, a comprehensive listing of both the effective and ineffective practices used by organizations was generated. this listing subsequently led to development of risk and critical factors influencing cost estimation. the practices, methods and factors were matched to the similarities found in the literature. finally, a comparison of current cost estimate practices prevalent in oman could be made with those used around the world. the first part of the questionnaire identified general information about the organizations that participated in filling out the survey. the information was categorized, analyzed and the result are shown in table 1. three distinct categories were used to differentiate between the respondent sectors, namely: public sector (26.67%) private sector (62.22%) and mixed sector (11.11%). the respondents were mixed sector (11.11%). the respondents were categorized as client, consultant or contractor. it was found that the majority of the organizations surveyed (mainly consultants and contractors) belong to the private sector. the respondents operating in the public sector mainly comprise the clients. the organizations were further identified by their primary field of operation as shown in fig. 1. it can be deduced that the majority of the organizations interviewed have operated in the fields of construction, and oil and gas. coincidently, both fields happen to comprise the largest business sectors in oman’s economy according to the ministry of economy, sultanate of oman (statistical year book 2011). in addition, the respondents from manufacturing and engineering fields were also identified. table 2 shows the average value of projects which have been completed in the period 2009 – 2014. it was found that most of the completed projects were valued over 10 million omani riyals (one omani riyal = 2.59 usd) (the majority of these projects were issued by clients and involved contractor participation). the successful completion of high value projects suggests that these organizations have an efficient cost estimating process and construction performance. table 3 presents the range of variations between the estimated and the actual costs of projects in the case of cost overruns. the cost overruns were grouped into five categories based on the value percentage. the data obtained showed that 51.1% of respondents claimed that the variations were minor, table 1. organizations by sector. figure 1. organization field of business by category. no response 7% oil and gas 40% manufacturing 4% construction 42% engineering 7% categories client consultant contractor total % no response 0 0 0 0 0 public sector 12 0 0 12 26.7 private sector 0 14 14 28 62.2 mixed sector 3 1 1 5 11.1 total 15 15 15 45 100 118 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh ranging between 0% 10%. this is considered to be an optimistic outcome as it suggests that the projects were estimated accurately. about 35.5% of the respondents claimed presence of variations between 10% 30%, indicating cost overrun and inefficient project cost estimation. 3.1 practice and methodology for project cost estimation the second part of the survey collected information on various practices of cost estimation. when asked if the organizations have a dedicated cost estimation department, 78% of the respondents stated that they do (fig. 2). they further elaborated on this, describing the structure of the stated department. there exists a department head under whom a number of experienced personnel who carry out relevant data analysis and cost estimation. pricing information is regularly reviewed, updated and an open environment of communication exists, thereby allowing estimators to enhance their knowledge and gain experience. about 18% of the respondents reported that they did not have a dedicated cost estimation department. it is clear that clients give less importance to a dedicated department, which may lead to a poor project cost estimation process. when asked about the existence of standard operating procedures, 80% of the respondents replied in the affirmative (fig. 3). not surprisingly, 66% of the remaining 20% that responded in the negative were clients. it should be noted that there would be an added benefit of producing a manual of standard estimating practices to be used by all clients, consultants and contractors. with such a guide, the estimators would be able to perform estimates in less time since many of their questions could be addressed simply by referring to the manual. the benefits would substantially exceed the cost of capital investment of regular updates of the manual. approximately 53.3% of the respondents claimed that top management has direct involvement in the preparation of the cost estimation process (table 4). about 20% are indirectly involved like advising to those personnel preparing the costing and about 22% of top management are involved at the last stage of preparation of cost estimation process. top management involvement is important since the decisions made at all stages of the cost estimation process will often determine the way in which the project is carried out. it is also important that while estimating and preparing the documents of the cost proposal, the requirements should be understood and the information be robust and clear. top management involvement also provides confidence on the project cost estimate accuracy. figure 2. organizations that have a dedicated cost estimation department. figure 3. status of organizations that have standard operating procedures for the cost estimation department. 4% 16% 13% 0% 0% 29% 4% 0% 0% 33% 0% 0% 4% 78% 18% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% no response yes no others client consultant contractor total 0% 20% 13% 0% 0% 29% 4% 0% 0% 31% 2% 0% 0% 80% 20% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% no response yes no others client consultant contractor total 119 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 approximately 53.3% of the respondents claimed that top management has direct involvement in the preparation of cost estimation process (table 4). about 20% are indirectly involved like advising to those personnel preparing the costing and about 22% of top management are involved at the last stage of preparation of cost estimation process. top management involvement is important since the decisions made at all stages of the cost estimation process will often determine the way in which the project is carried out. it is also important that while estimating and preparing the documents of the cost proposal, the requirements should be understood and the information be robust and clear. top management involvement also provides confidence on the project cost estimate accuracy. 3.2 techniques for project cost estimation the third part of the survey collected information on the techniques of project cost estimation. first, the factors that determine the proper estimation technique to be used were sorted out (table 5). about 62.2% of the respondents considered the following factors while preparing an effective estimate: (1) level of scope of work available, (2) size of the project, (3) time constraints, and (4) availability of software. it was noticed, however, that contractors (70%) favored the level of scope of work available as a criterion while cost estimating. among all surveyed organizations, 77% reported that they use the unit rate method for estimating construction costs (fig. 4). the use of top-down and historical data based techniques is of a second choice. bench marking, bottom-up and expert-opinion were found to be the least used. 55% of the respondents justified the common use of the unit rate method by claiming that it was the most accurate. this opinion was further justified by 31.1% of the respondents who claimed that it was the safest method to employ when the scope of work was unclear (table 6). 55.5% of the respondents were in agreement that the historical bid based estimate was of second preference in oman (table 7). 40% of the respondents cited the following revision methodologies used in validating historical data: (1) revision according to the nature of the project, (2) revision as a market survey, and (3) revisions as a new process of the project (table 8). 33.3% of the respondents favored the use of revisions according to the nature of the project. 53.3% of the respondents disagreed with the table 2. average values of projects completed during the years (2009-2014) by category. categories total no response 4.44% < r.o. 1 million 11.11% r.o. 1 million r.o. 5 million 35.56% r.o. 6 million r.o. 10 million 6.67% > r.o. 10 million 42.22% table 3. ranges of variations between the actual and estimated cost. table 4. top management involvement in the estimation process. categories client consultant contractor total % no response 1 0 0 1 2.22% direct involvement 4 12 8 24 53.33% indirect involvement 3 1 5 9 20% last stage involvement 6 2 2 10 22.2% others 1 0 0 1 2.22% total 15 15 15 45 100% categories total no response 2.22% no project overrun 11.11% 0% 10 % 51.1% 10%20% 20% 20% 30% 15.56% more than 30% 0 120 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh figure 4. commonly used cost estimation techniques. table 5. criteria to be considered while cost estimating. categories client consultant contractor total % no response 2 0 0 2 4.44% level of scope of work available 2 2 9 13 28.9% size of project 2 0 0 2 4.44% time constraint 0 0 0 0 0 availability of software 0 0 0 0 0 all of the above 9 13 6 28 62.22% total 15 15 15 45 100% table 6. respondents reasoning for using rate method. categories client consultant contractor total % no response 2 0 0 2 4.44% very accurate method 10 6 9 25 55.56% safest method when scope of work is unclear 3 5 6 14 31.11% most recommended by management 0 1 0 1 2.22% other methods are risky 0 3 0 3 6.67% total 15 15 15 45 100% table 7. is the historical bidbased estimate the most commonly used in oman? categories client consultant contractor total % no response 2 0 0 2 % agree 8 9 8 25 55.56% not agree 5 6 6 17 37.78 others 0 0 1 1 2.22% total 15 15 15 45 100% 0% 24% 2% 7% 0% 0% 0% 0% 0% 0% 29% 2% 0% 0% 0% 0% 2% 0% 4% 78% 7% 7% 0% 0% 0% 2% 2% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% n o r e s p o n s e u n it r a te h is to ri c a l d a ta b a s e d b e n c h m a rk in g b o tt o m u p t o p d o w n e x p e rt o p in io n p a ra m e tr ic t re n d a n a ly s is client consultant contrator total 121 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 opinion that benchmarking was a commonly used methodology in oman (table 9). both clients (53%) and consultants (73%) strongly disagreed with the use of benchmarking, as they believe that the method is highly inaccurate. they did however point out that it is the fastest method used in arriving at an estimate. 3.3 the use of software for project cost estimation the use of software in the estimating process is beneficial in many ways. software provides estimators with the tools that enable them to save time. software assists estimators in calculating the total net cost of the project. in addition, it helps those who have to make decisions based on the estimator's report. it also assists in implementing company procedures through standardization (brook 2010). 48.8% of the respondents stated that they do not use any computer software for cost estimating, as opposed to 46.6% who said that they do (table 10). 37.7% of the respondents stated that historical average prices are developed by the input of rates based on bid data (table 11). 31% disagreed with the statement above. around 49% of the respondents reported that software was not based on historical low bid cost data (table 12). historical cost databases are always considered when there is a requirement for readily available construction information. historical cost databases serve as a guide during the early stages of project development. table 8. revision methodology used in validating historical data. categories client consultant contractor total % no response 2 0 0 2 4.44% revision according to nature of project 5 5 5 15 33.33% revision as market survey 6 0 1 7 15.56% revision as new process of project 0 1 2 3 6.67% all of the above 2 9 7 18 40% total 15 15 15 45 100% table 9. is benchmarking the most commonly used method in oman? categories client consultant contractor total % no response 0 0 3 3 6.67% agree 6 4 5 15 33.33% not agree 8 11 5 24 53.33% others 1 0 2 3 6.67% total 15 15 15 45 100% table 10. organizations using software for cost estimation. categories contractor consultant client total % no response 0 0 2 2 yes 7 8 6 21 46.6% no 8 7 7 22 48.8% total 15 15 15 45 table 11. rates updated in software based on bid data to develop average prices. categories client consultant contractor total % no response 3 0 0 3 6.66% yes 4 8 5 17 37.7% no 3 6 5 14 31% others 5 1 5 11 24.44% total 15 15 15 45 100% table 12. software based on historical low bid cost data. categories client consultant contractor total no response 3 0 0 6.66% yes 2 6 0 17.77% no 5 8 9 48.88% others 5 1 6 26.66% total 15 15 15 100% 122 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh there are various factors that play a role in interpreting cost databases as discussed earlier in the interpretation of the results. to name a few: (1) project size, (2) year being built, (3) inflation rate, (4) local market conditions, (5) complexity of the project, (7) procurement method, (8) site conditions. modern historical database mining is all about better utilization and re-use of existing data. 48.9% of the respondents claimed that unit rate analysis was the basis for pricing rates (table 13). this finding indicates that the rates used in the software are revised and analyzed according to the latest project scope of work available. 22.2% of the respondents claimed that pricing rates are based on the project process. in either case, pricing rates are not totally dependent on historical databases. 48.9% of the respondents claimed that it was the responsibility of project estimator to define the parameters and variables which serve as inputs for cost estimation software (table 14). 17.8% of the respondents claimed that this responsibility falls on the project manager. the reason could attributed to the fact that such organizations do not have a dedicated estimation department and no particular estimators are assigned specifically for creating project cost estimate. most project managers are appointed to a project as a whole. historical budgeting data offers valuable information on the success of past construction project, while current computer technology makes it easier to develop database oriented programs (brook 2010). although 53.3% of the organizations reported that they do not use olap (on-line analytical processing) facility in their project cost estimate practice, studies have proven that using the olap environment can help in understanding the risk in the estimated cost and provide a means for forecasting more reliable construction cost estimates (moon and kown 2007). olap provides a secure, web-based workplace for people spread around the world, thereby allowing them to share information and make analysis. 35% of the respondents stated that they deal with uncertainty in the models and software very carefully as they realized that the uncertainties can cause the cost to fluctuate erroneously (table 15). table 13. the basis for pricing rates. categories client consultant contractor total no response 4 0 2 6 unit rate analysis 6 6 10 22 fixed standard notes 1 5 1 7 updating according to project process 4 4 2 10 total 15 15 15 45 table 14. fixed parameters and variables which serve as input for cost estimation software. categories client consultant contractor total no response 4 0 0 4 project manager 2 4 2 8 project estimator 8 6 8 22 software in charge 0 0 0 0 others 1 5 5 11 total 15 15 15 45 table 15. dealing with uncertainties in models and software. categories client consultant contractor total no response 4 0 1 5 ignoring them as they have minimal effect 3 3 1 7 dealing with them very carefully as they might change the cost 5 4 7 16 considering them as unclear output and including them in the final cost 2 2 1 5 others 1 6 5 12 total 15 15 15 45 123 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 3.4 risk factors in project estimation process this section of the survey explores the various risk factors encountered in project cost estimation. after having identified the factors influencing cost estimation during all project stages, they are categorized in order of risk. then these risk factors were designated likert scale values ranging from 4, reflecting the highest degree of risk, to 1 reflecting the lowest degree of risk. the responses for each factor were then added and divided by the total number of responses. in doing so, the mean values for each factor were obtained. the factors were then arranged in order of priority as per client, consultant and contractor (tables 16-18). this is done to exercise a degree of control, thereby ensuring effective cost estimation for conventional construction projects. according to the clients, an unclear reading and understanding of specifications scores the highest critical risk factor (3.667). unavailability of resources comes as the second highest factor (3.200), followed by market fluctuation (3.066) (table 16). lack of experienced human resources contributes to committing mistakes during estimating process while changing in prices of raw materials, like cement and steel, results in an unreliable estimation. consultants reported that lack of clear documents of the project scope of work (3.667), followed closely by unavailability of resources (3.600) and an unclear reading and understanding of specifications (3.400), were the most critical risk factors in developing of an estimate (table 17). as can be noticed, there is an agreement between clients and consultants on the main causes of risks while making estimates. remedies to such problems include more time and revisions of scope work with consultation of local and international experts. the contractors reported that the most critical factors as being an unclear reading and understanding of the specification (3.466), lack of clear documents of the project scope of work (3.4667) and lack of experienced project cost estimators (3.400) (table 18). they also reported that the vague specifications and scope of work is the main cause of risk that leads to crude estimates. a study by el-sayegh (2014), identified forty-two risk factors which cause time and cost overrun in uae construction industry. most of these factors are applicable to oman construction industry due to similarity of the gulf countries construction environment. table 16. order of mean values of risk factors as per clients. risk factors in the project estimation process mean values of factors per client priority unclear reading and understanding of specifications 3.6667 1 unavailability of resources 3.2000 2 market fluctuation 3.0667 3 lack of experienced project cost estimator's 3.0000 4 lack of clear documents of the project scope of work 2.8000 5 lack of project planning and monitoring 2.5333 6 political influence 2.5333 7 aggressive competition at tendering stage 2.4667 8 urgency emphasized by owner while issuing tender 2.3333 9 uniqueness project activities 2.2667 10 table 17. order of mean values of risk factors as per consultants. risk factors in the project estimation process mean values of factors per consultant priority lack of clear documents of the project scope of work 3.6667 1 unavailability of resources 3.6000 2 unclear reading and understanding of specifications 3.4000 3 lack of project planning and monitoring 3.1333 4 market fluctuation 2.9333 5 lack of experienced project cost estimator's 2.8667 6 aggressive competition at tendering stage 2.7333 7 urgency emphasized by owner while issuing tender 2.6000 8 uniqueness project activities 1.8667 9 political influence 1.6667 10 124 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh 3.5 factors critical to a successful project cost estimation twenty factors have been identified as being critical to achieve a successful cost estimate. using the factors identified as driving variables, structured questions were made to explore their actual impact. the respondents were then asked to rate each of the factors in order of importance based on their cost estimation experience, by considering all types of projects they have been involved in. likert scale with values ranging from 4, reflecting most critical, to 1, reflecting least critical, was used for evaluating the success of the project. the responses for each of the factors were then added and divided by the total number of responses to result the mean values of each factor. tables 19-21 show the factors’ priorities as per the responses from the clients, consultants, and contractors respectively. according to the clients, accuracy in taking quantities (4.066), availability of materials, skilled labour, equipment and time (4.066) and accuracy of data (3.933), were the most important factors to obtaining a successful cost estimate (table 19). again, accuracy in estimating the quantities requires reasonably high experience and involvement by estimator. failure to control the cost of both human and materials leads to bad consequences on the overall estimated budget. table 18. order of mean values of risk factors as per contractors. risk factors in the project estimation process mean values of factors per contractor priority unclear reading and understanding of specifications 3.4667 1 lack of clear documents of the project scope of work 3.4667 2 lack of experienced project cost estimator's 3.4000 3 market fluctuation 3.2000 4 unavailability of resources 3.1333 5 aggressive competition at tendering stage 3.0000 6 lack of project planning and monitoring 2.9333 7 urgency emphasized by owner while issuing tender 2.8000 8 uniqueness project activities 2.4000 9 political influence 2.0000 10 table 19. mean values of critical factors to successful project cost estimation as per the clients. factors critical to successful project cost estimation mean /client priority accuracy in taking quantity 4.0667 1 availability of materials, skilled labor, equipment and time 4.0667 2 availability and accuracy of data 3.9333 3 top management involvement 3.8000 4 the contractor / consultant work load 3.8000 5 collection of quotations during the estimation process 3.7333 6 wide range of options and choices of different material 3.6667 7 pre-bid vendor negotiation 3.6000 8 involvement of project oriented personnel in the estimation process 3.6000 9 deviating from the original scope of work 3.5333 10 reliability of the materials quotations provided by suppliers 3.5333 11 contractors keen on the job 3.4000 12 clear documentation of projects 3.4000 13 size and value of the project 3.3333 14 level of awareness of competition in the omani market 3.3333 15 timely decision by the owner/ his staff 3.2667 16 chances of improper understanding of client requirement 3.2000 17 in-house logistics 3.2000 18 existence of unit rate analysis 3.1429 19 outsourcing certain services all though available in-house 2.9333 20 125 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh the consultants reported that accuracy in taking quantities (4.200), availability and accuracy of data (4.066) and deviation from the original scope of work (4.000) were the most important factors to reaching a successful cost estimate (table 20). contractors reported that their interest in the job (4.400), the involvement of project-oriented personnel of the estimation process (4.400) and clear documentation of projects (4.333) were the most important factors to obtaining a successful cost estimate (table 21). the factors cited above fall inline with the results of a study liu and zhu (2007) table 20. mean values of critical factors to successful project cost estimation as per the consultants. factors critical to successful project cost estimation mean values of factors per consultant priority accuracy in taking quantity 4.2000 1 availability and accuracy of data 4.0667 2 deviating from the original scope of work 4.0000 3 collection of quotations during the estimation process 4.0000 4 in-house logistics 3.9333 5 availability of materials, skilled labor, equipment and time 3.8000 6 top management involvement 3.7333 7 clear documentation of projects 3.6667 8 pre-bid vendor negotiation 3.3333 9 contractors keen on the job 3.2667 10 existence of unit rate analysis 3.2000 11 outsourcing certain services all though available in-house 3.2000 12 reliability of the materials quotations provided by suppliers 3.1333 13 timely decision by the owner/ his staff 3.0667 14 the contractor / consultant work load 3.0000 15 level of awareness of competition in the omani market 2.8667 16 wide range of options and choices of different material 2.7692 17 involvement of project oriented personnel in the estimation process 2.7333 18 chances of improper understanding of client requirement 2.6667 19 size and value of the project 2.2667 20 table 21. mean values of critical factors to successful project cost estimation per the contractors. factors critical to successful project cost estimation mean values of factors per contractor priority contractors keen on the job 4.4000 1 involvement of project oriented personnel in the estimation process 4.4000 2 clear documentation of projects 4.3333 3 availability of materials, skilled labor, equipment and time 4.3333 4 accuracy taking quantity 4.2667 5 collection of quotations during the estimation process 4.2667 6 in-house logistics 4.1333 7 deviating from the original scope of work 4.0000 8 existence of unit rate analysis 3.9333 9 outsourcing certain services all though available in-house 3.8667 10 pre-bid vendor negotiation 3.8667 11 size and value of the project 3.8000 12 level of awareness of competition in the omani market 3.8000 13 wide range of options and choices of different material 3.7333 14 reliability of the materials quotations provided by suppliers 3.7333 15 the contractor / consultant work load 3.6000 16 availability and accuracy of data 3.5333 17 top management involvement 3.4000 18 timely decision by the owner/ his staff 3.4000 19 chances of improper understanding of client requirement 3.2667 20 126 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh mentioned in the literature review. the authors have shown that cost estimates may be improved by using phased cost factors. this was achieved by grouping influential factors into three: (1) input control factors, (2) behavioural control factors, and (3) output control factors. the responses obtained show that almost all the factors stated above have the same influence in the construction industry in oman. 4. conclusion this paper surveyed and assessed project management practices and techniques used in construction cost estimation across oman. it also explored risk factors that contributed to unsuccessful cost estimation process in these projects. inadequate or poor cost estimation has been identified as one of the major causes for project cost overruns in oman. although a number of high value projects have been successfully completed in oman during the period 2009-2014, opinion amongst clients, consultants and contractors suggests that clients pay the least amount of attention to the cost estimate. feedback from face-to-face interviews and the literature review have helped to identify current best practices and techniques that can aid in achieving a successful project cost estimate. also, identified are factors that can lead to risk and subsequent success or failure in the cost estimation process. as shown in table 3, cost overrun in projects especially in the 20% – 30% range is prevalent amongst clients and consultants. contractors, on the other hand, claim to have a maximum of only 10% cost overrun. the research results showed that:  around 35% of projects were completed with variation within the estimated budget in the range of 10%-30% (table 2 and 3).  the most commonly used cost estimation technique was the unit rate method (fig. 4).  using a web-based software which updates unit rate price proved to be effective in reaching adequate cost estimate.  top management involvement provides confidence on the project cost estimate accuracy.  both clients and contractors believe that unclear understanding of specifications is the highest risk factor toward a cost overrun.  both clients and consultants consider accuracy in taking quantities is the most important factor that contributes to a successful estimation process. it is recommended that clients need to invest further into the methods and procedures of cost estimation in order to have a better visualization of their projects. it is also recommended that future studies in cost estimation practices be carried out more precisely, as per cost estimate techniques matched to a sector. research on each cost estimating method that examines the pros and cons of the method would be more effective. a distinction needs to be made whereby the techniques used in specific industry are not necessarily ported over to another industry and vice versa. methodologies should be specific to the industry the company operates in. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references aibinu a.a. pasco t (2008), the accuracy of pre‐ tender building cost estimates in australia. construction management and economics, 26(12): 1257–1269. akintoye a. fitzgerald e. (2000). product cost estimation: technique classification and methodology review. construction management and economics. construction management and economics, 18(2): 161–172. al-mohsin m. al-nuaimi a.s. (2013), modeling of construction cost of villas in oman. the journal of engineering research, tjer, 11(1): 34-43. albalushi i. usman f. alnuaimi a.s. (2013), construction cost overrun and variations: investigation on its causes and consequences. australian journal of basic and applied sciences, 7(14): 311-323. artto k.a lehtonen j.m. saranen j. (2001), managing projects front-end: incorporating a strategic early view to project management with simulation. international journal of project management, 19(5): 255-264. brook m. (2010), estimating and tendering for construction work, 4 th ed. butterworth-heineman. chou j.s. peng m. persad k. o’connor j. (2006), quantity-based approach to preliminary cost estimates for highway projects. transportation research record: journal of the transportation research board, 1946(january 2006), 22–30. chou j. (2011), cost simulation in an item-based project involving construction engineering and management. international journal of project management, 29(6): 706-717. chou j. yang i. kiong w. (2009), probabilistic simulation for developing likelihood distribution of engineering project cost. automation in 127 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 115-129 construction, 18(5): 570-577. el-sayegh, s. m. (2014). project risk management practices in the uae construction industry. international journal of project organisation and management, 6(1-2), 121-137. elfaki a.o. alatawi s. abushandi e. (2014), using intelligent techniques in construction project cost estimation: 10-year survey. advances in civil engineering. heigermoser d. soto b.j. abbott e.l.s. chua d.k.h. (2019), bim-based last planner system tool for improving construction project management. automation in construction, (104): 246-254. heralova r.s. (2017), life cycle costing as an important contribution to feasibility study in construction projects. procedia engineering, (196): 565-570. herbold k.d. (2000), using monte carlo simulation for pavement cost analysis. public roads, 64(3): 2-3. hwang s. (2009), dynamic regression models for prediction. journal of construction engineering and management, 135(5): 360-367. institute p.m. (2008), a guide to the project management body of knowledge. jonny o. lichtenberg s. (2016), successive cost estimation – successful budgeting of major projects. in procedia social and behavioral sciences 226: 176-183. juszczyk m. (2017), the challenges of nonparametric cost estimation of construction works with the use of artificial intelligence tools. procedia engineering, 196: 415-422. kim k.j. kim k. (2010), preliminary cost estimation model using case-based reasoning and genetic algorithms. journal of computing in civil engineering, 24(6): 499-505. kiziltas s. ph.d. akinci b. (2009), contextual information requirements of cost estimators from past construction projects. journal of construction engineering and management, 135(9): 841-852. kujala j. brady t. putila j. (2014), challenges of cost management in complex projects. international journal of business and management, 9(11): 4858. lim b.m.p.n. skitmore m. xiong b. (2016), drivers of the accuracy of developers’ early stage cost estimates in residential construction. journal of financial management of property and construction, 21(1): 1366-4387. liu l. zhu k. (2007), improving cost estimates of construction projects using phased cost factors. journal of construction engineering and management, 133(1): 91-95. marco a. de. briccarello d. rafele c. (2009), cost and schedule monitoring of industrial building projects: case study. journal of construction engineering and management 135(9): 853-862. moon s. kwon k. (2007), effectiveness of olap based cost data management in construction cost estimate. automation in construction, (16): 33644. nassar a (2006), project life cycle cost estimate. intercontinental networks white paper, january 2006. niazi a. dai j.s. balabani s. seneviratne l. (2006), product cost estimation: technique classification and methodology review. journal of manufacturing science and engineering, 128(2): 563-575. oberlender g.d. trost s.m. (2001), predicting accuracy of early cost estimates based on estimate quality. journal of construction engineering and management, 127(3): 173-182. olatunji o.a. sher w. (2014), perspectives on estimating practices modelling. australasian journal of construction economics and building, 14(4): 32-53. oyedele o.a. (2015), evaluation of factors affecting construction cost estimation methods in nigeria. the wisdom of the ages to the challenges of the modern world, 1-30. popovic z. (2004), implementation of data mining techniques in construction estimating. neural, parallel and scientific computations, 12(1): 3752. saha s.k. (2010), modelling the effectiveness of computer-aided development projects in the semiconductor industry. international journal of engineering management and economics, 1(2/3): 162-178. salem o. abourizk s. ariaratnam s. (2003), riskbased life-cycle costing of infrastructure rehabilitation and construction alternatives. journal of infrastructure systems, 1(march): 6–15. sequeira s. lopes e. (2015), simple method proposal for cost estimation from work breakdown structure. procedia computer science, 64: 537544. statistical year book. (2011). ministry of economy, sultanate of oman. swei o. gregory j. dana t. kirchain r. (2017), construction cost estimation: a parametric approach for better estimates of expected cost and variation. transportation research part b: methodological, 101: 295-305. tighe s. (2001), guidelines for probabilistic pavement life cycle cost analysis. transportation research record: journal of the transportation research board, 1769: 28-38. touran a. asce m. (2003), probabilistic model for cost contingency. journal of construction engineering and management, may/june(june): 280-285. toutounchian s. abbaspour m. dana t. abedi z. (2018), design of a safety cost estimation parametric model in oil and gas engineering, procurement and construction contracts. safety science, 106: 35-46. wang w.c. (2004). supporting project cost threshold 128 m. al-siyabi, a. alnuaimi, m. khadem, and k. alzebdeh decisions via a mathematical cost model. international journal of project management, 22(2): 99-108. wu s. wood g. ginige k. jong s.w. (2014), technical review of bim based cost estimating in uk quantity surveying practice, standards and tools. journal of information technology in construction itcon, 19: 534-562. yang i. (2008), distribution-free monte carlo simulation : journal of construction engineering and management, 134(5): 352-360. yu w. lai c. lee w (2006), a wice approach to real-time construction cost estimation. automation in construction, 15(1): 12-19. 129 the journal of engineering research (tjer), vol. 19, no. 1, (2022) 22-32 *corresponding author’s e-mail: aissa@squ.edu.om doi:10.53540/tjer.vol19iss1pp22-32 investigation of noise exposures, perception, and health effects in different microenvironments in a university community patrick amoatey1*, issa al-harthy 1, khalifa al-jabri 1, abdullah al-mamun1, and mahad said baawain2 1department of civil and architectural engineering, college of engineering, sultan qaboos university, muscat, oman. 2ministry of labour, muscat, oman abstract: this study aims to assess noise levels in selected outdoor and indoor microenvironments in a university community in oman. the perception of noise levels within the sultan qaboos university campus was investigated through a survey study. also, the effect of exposed noise levels on annoyance and sleep disturbance were predicted including their potential risk on cardiovascular health. among all the measured parameters, it was found that outdoor (41.6%-50%) and indoor (38.5%-46.2%) microenvironments have exceeded the critical levels of 55 db during morning and afternoon periods. the respondents (698 people) identified traffic and indoor building-related activities as the main sources of noise levels but the majority (44%) of them rated their impact as low. however, more than 30% of the respondents considered traffic as the main contributor to university noise levels. the percentage of highly annoyed persons was predicted to be high in outdoor areas especially in the residential (25%) and near the hospital (13%) areas. however, indoor environments including construction materials and structures labs (14%) showed similar annoyance rates. also, the percentage of high sleep disturbed persons was found higher in residential areas (7.4%) areas compared to hospital areas (5.3%) locations. the study concluded that there might be an association between the exposed noise levels and the risk of developing cardiovascular diseases. this is the first study that has provided a high spatial variability noise exposure levels across a university environment in oman, this will contribute to designing future sustainable mitigation strategies to improve the health and well-being of the exposed population. the study has provided a baseline knowledge needed for future epidemiological studies. keywords: noise levels; annoyance; sleep disturbance; health effects; university campus; oman. حرمفي ببیئات داخلیة وخارجیة الناتجة عنھ ، واآلثار الصحیةدراسة تأثیر الضوضاء على االدراك جامعي الحارثي*، خلیفة الجابري، عبد� المأمون، محاد باعوین عیسىباتریك أمواتي، ولتحقیق. قابوس السلطان جامعة حرم في وخارجیة داخلیة بیئات في الضوضاء مستویات تأثیر دراسة إلى الدراسة ھذه تھدف: الملخص واضطراب االنزعاج على الضوضاء مستویات بتأثیر التنبؤ تم كما الجامعي، المجتمع رأي الستقصاء االستبیان نظام استخدام تم أھدافھا، حوالي قیاسھا، تم التي العوامل جمیع بین من أن النتائج أظھرتو . الدمویة واألوعیة القلب صحة على المحتملة ومخاطرھما النوم المسموحة الحدود تجاوزت قد الداخلیة البیئات في٪) 46.2و٪ 38.5(و الخارجیة البیئات في الضوضاء مستویات من٪) 50و٪ 41.6( والمباني المرور بحركة المتعلقة األنشطة) شخًصا 698( المشاركون حدد. الظھر وبعد الصباح أوقات في) دیسیبل 55( عالمیا للضوضاء ٪) 30( من أكثر فإن ذلك، ومع٪). 44( الغالبیة قبل من منخفض تأثیرھا تصنیف تم ولكن وضاءالض لمستویات رئیسیة كمصادر الداخلیة نسبة أن كذلك الدراسة وبینت. الجامعة في الضوضاء مستویات لرفع الرئیسي السبب ھي المرور حركة أن یرون المشاركین من االنزعاج نسبة( السكنیة المناطق في وخاصة الخارجیة، البیئات في عام بشكل مرتفعة كانت شدید انزعاج من یعانون الذین األشخاص معامل مواد أظھرت فقد الداخلیة، بالبیئات یتعلق فیما اما٪). 13= الشدید االنزعاج نسبة( الجامعة مستشفى وبالقرب٪) 25= الشدید في المشاركین بعض لدى النوم اضطراب أن دوج كما٪). 14= الشدید االنزعاج نسبة( النتائج من مماثلة نسب البناء واالنشاءات ھناك یكون قد أنھ الى الدراسة خلصت وقد٪). 5.3( الجامعي المستشفى مع بالمقارنة٪) 7.4( السكنیة المناطق في مرتفعة الدراسة والمستشفى) 1.11= الفردیة النسب( السكنیة المناطق في الدمویة واألوعیة القلب أمراض وحدوث الضوضاء مستویات بین ارتباطا ). 1.02= الفردیة النسب( الجامعي .مستویات الضجیج؛ إزعاج؛ اضطراب النوم؛ آثار صحیة؛ حرم الجامعھ؛ سلطنة عمان الكلمات المفتاحیة: investigation of noise exposures, perception, and health effects in different microenvironments in a university community 23 nomenclature cvd cardiovascular disease db decibel dba decibel average imes indoor microenvironments ipd in-patients department omes outdoor microenvironments or odds ratios slm sound level meter squ sultan qaboos university squh sultan qaboos university hospital who world health organization %ha percentage of highly annoyed %hsd percentage of high sleep disturbed 1. introduction according to the world health organization (who), community noise exposure is one of the serious environmental health problems facing the general population especially among the vulnerable group (who 1999). this is attributed to the diverse sources of these noise levels including traffic, railways, industries, and the nearby neighbourhoods, thereby making it difficult to control (al-mansour 2006; liu et al. 2017; sieber et al. 2018). also, indoor activities in construction, religious and recreational areas were recognized as another source of noise pollution (ali 2013; pierrette et al. 2012; ramazani et al. 2018). the health impacts of noise may be exacerbated especially in microenvironments such as schools and hospitals where the population (i.e. students, staff, patients) spend a substantial amount of time (i.e. average of 7 12 hours) (rogers et al. 2004). in the case of health facilities, in-patients could spend > 90% of their time (> 100 hours) in indoor environments especially in hospital wards (amoatey et al. 2018; henshall et al. 2018; upso 2017). thus, considering the longer durations and high frequency of noise exposures from multiple sources, it is therefore expected that its adverse health impact and annoyance levels may worsen (al harthy 2006; kamal et al. 2020). in a systematic review study involving more than 5 million subjects, an association was observed between residential noise exposures (median noise levels of 56.7 dba) and an increase in the risk of hypertension among the adult population (dzhambov and dimitrova 2018). similar studies have also reported an increase in diabetes mellitus (zare sakhvidi et al. 2018), blood pressure in children (dzhambov and dimitrova 2017), and arterial stiffness (foraster et al. 2017) due to community noise exposures. the health effect of noise levels from multiple sources including traffic, aircraft, railways, neighbourhood, sports facilities, construction activities, and hospitality industries was investigated (dreger et al. 2015) among a total of 1,185 school children in germany by estimating their relative risks (rr). the study found a risk of incidence of mental health problems (e.g. inability to conduct themselves (rr=1.62), hyperactivity (rr=1.69), and emotional symptoms (rr=1.69) among the students. in spain, díaz et al. (2020) observed that an increase in 1 dba of traffic noise with median leqday levels of 61.9 dba was associated with emergency hospital admissions from anxiety and depression. a cohort study involving a total of 52,758 residents was conducted across two major cities (aarhus and copenhagen) in denmark to determine how road traffic noise exposures affect cardiovascular health among the population (thacher et al. 2020). it was estimated that, per increase in noise levels at the interquartile range of 10 dba, there was an increase in the risk of incidence of stroke with a hazard ratio of 1.11, and 1.06 for ischemic heart diseases, and 1.13 for cardiovascular diseases among the exposed population (thacher et al. 2020). there is also evidence from recent studies that multi-pollutants exposures including noise and air pollutants could result in several health impacts including the effect on mental health performance levels (klompmaker et al. 2019), childhood obesity (bloemsma et al. 2019), and stillbirth (smith et al. 2020). over the recent years, noise pollution was recognized as one of the serious public health concerns in schools including health care environments as they could have profound effects on the quality of learning and exacerbate the health conditions of patients, especially those with underlying chronic health conditions (al-dorzi et al. 2020; alkhanjari et al. 2014). a noise exposure study involving monitoring of more than 34 different locations across a university campus in nigeria showed daily noise levels ranging from 42 to 97 db. it was concluded that these high noise levels were attributed mostly to high road traffic, the existence of several car parks, and commercial activities occurring in closer proximity to the university campus (okolie et al. 2020). this is deemed as a major occupational risk factor to cause noise to induce hearing loss problems to university staff, and students who spend most of their time in such environments. it was found in kuwait that a maximum 8-hour noise exposure level of more than 100 dba in a school environment was greater than the recommended national institute of occupational safety and health standard of 90 dba (yassin et al. 2016). the study found that 38%, 63%, and 30% of the teachers have complained about poor sleep quality, vocal problems due to excessive shouting during teaching as a result of high background outdoor noise levels, and headache problems, respectively. in qatar, shaaban and abouzaid (2021) reported a positive association between traffic volume and noise levels of 61-72 dba near school environments during the daytime which was higher than qatar and who (1999) permissible limit of 55 dba. according to caviola et al. (2021), noise can have a long-term effect on mathematics performance levels of school-age children as result of listening difficulties caused by frequent exposure to environmental noise. however, as the child grows, these cognitive effects can reduce significantly. patrick amoatey, issa al-harthy , khalifa al-jabri , abdullah al-mamun, and mahad said baawain 24 therefore, it has been recommended that application of effective mitigation measures including the use of acoustic barriers, and growing of vegetation near buildings can improve the learning performances of school children (margaritis et al. 2018; ow and ghosh 2017; umbas et al. 2021). most arid middle east countries including, oman have a high outdoor temperature (> 30ºc) and humidity (> 90%) levels especially during the summer seasons. therefore, to reduce exposure to such harsh environmental conditions, many road networks have been constructed to ease the transportation of people to their residences, schools, hospitals, and shopping malls. however, the presence of these road networks could also serve as potential sources of community noise levels. also, due to the limited public transportation systems in these countries, there has been an increase in the number of private vehicle users, and this could potentially increase community noise levels (abdul-wahab and fadlallah 2014). to date, there are limited studies assessing noise levels and the potential adverse health effects in most schools, academic institutions, and hospital facilities in oman. therefore, the objective of this study is to conduct a comprehensive assessment of indoor and out exposure levels noise levels in different microenvironments of a university community in oman. the annoyance and health impact assessment of the noise levels was evaluated using both social survey and dose-response model. it is expected that this study will provide valuable datasets for future community noise mitigation policies and also as a model for similar studies across the globe. 2. research methodology study area the study was carried out within sultan qaboos university (squ) and squ hospital (squh) communities in muscat (23° 36' 51.5808 ''n, 58° 32' 43.0224'' e), the capital city of oman (amoatey et al. 2020).squ is the oldest and the largest university in the country. the rapid growth of the university community was partly due to an increase in the number of residential buildings as most female students are offered on-campus accommodation throughout their entire study duration. also, since squ does not currently have any other campuses, there has been an increase in the density of students, staff population, and residential buildings over the years. currently, squ has a total of nine (9) colleges, 1 teaching hospital (squh) and thirteen (13) research centres, which are involved in providing teaching, research, and consultancy services at both local and international levels. figure 1 illustrates the map of muscat and an aerial view of squ. squ community has one of the effective on-campus transportation systems with several buses transporting students within and outside the university campus. in addition, due to limited public transportation systems in oman, most of the staff and students use private vehicles thereby increasing road traffic volumes in the university and consequently causing environmental problems (noise and air pollution). it is estimated that the average number of vehicles entering the squ community is about ˃16,000 per day (abdul-wahab and fadlallah 2014). this high influx of vehicles during the day is deemed as a major source of environmental noise exposure. study population the overall estimated number of exposed population in the squ community is 58,997 people consisting of students, administrative staff, academic staff, hospital staff. out of this, the majority are within squh (31,987 people) with hospital in-patient population contributing to the highest number of people within the squ community (upso 2017). the large population size of these vulnerable groups shows that potential noise levels near and within the squh area are an important public health concern. the high number of undergraduates (66.3%) in the squ campus also indicates how excessive noise levels may cause difficulties in learning and cognitive performance levels among them (woolner and hall 2010). table 1 shows a detailed breakdown of the proportion of different population categories found within the squ community. noise assessment areas to accurately assess noise levels in areas where the majority of people spend most of their time and also take into consideration the closer proximity to roads/streets to these areas, noise levels were measured across residential facilities, student apartments, lecture theatres, libraries, and in-patients department (ipd) within the squ community. in addition, to better understand the noise exposure levels, the 25 measurements were taken from both indoor and outdoor locations across twenty-five (25) microenvironments within the squ community within two months (febuary and april 2018) duration. the detailed description of all the measured noise points and their corresponding locations are shown in table 2 and figure 1, respectively. the noise level in each of the locations was measured three times daily: morning (7:30-8:00 am), noon(1:00-2:00 pm), and evening (6:00-8:00 pm) time. noise measurements at each of the 25 microenvironments (figure 1), noise levels were measured with a sound level meter (slm) (extech®, model hd 600) installed on a triploid stand at a height of 1.6 meters (extech 2015). slm was equipped with a high precision microphone (electret condenser) with a precision of ± 1.4 db and could measure sound pressure levels within a specified minimum (30 db) and maximum (130 db) range at the highest frequency of 8 khz (chew and wu 2016). investigation of noise exposures, perception, and health effects in different microenvironments in a university community 25 figure 1. location of muscat and the measured microenvironments in the squ community. table 1. a detailed description of sub-population data of squ community, source: upso (2017). (a) squ (n) (%) undergraduates 15,878 66.3 postgraduates 1,652 6.9 squ teaching staff 3,364 14.0 squh staff 3,058 12.8 total 23952 100.0 (b) squh (n) (%) medical doctors 454 14.8 nursing staff 1,371 44.8 technical staff 579 18.9 administrative staff 654 21.4 total 3058 100.00 (c) number of in-patients 31,987 table 2. detailed all measured noise locations within squ community. indoor microenvironments l1 college of science: corridors l2 college of science: computer labs l3 college of science: chemical labs l4 classrooms corridors block (a) l5 classrooms at block (a) l6 conference hall l7 exhibition hall l8 classrooms at block ( e ) l9 block (e) corridors l10 inside the main library l11 civil engineering construction lab l12 civil engineering computer lab l13 student service area outdoor microenvironments l14 road opposite to the college of science l15 road opposite to the block (a) l16 road opposite to the building (c) l17 out of conference hall l18 botanical garden l19 the road between the mosque and the college of economy and political science l20 around block (e) l21 around the cultural center l22 al andalus garden l23 around the squ hospital l24 squ residential area l25 main entrance the slm used in this study was certified by international electrotechnical commission (iec) as class ii and american national standards institute (ani) as a type ii noise measurement instrument (zaw et al. 2020). in-house calibration was done on slm before and after monitoring each selected location and the instrument was operated continuously for 1 min as start-up time. the device was controlled to measure a resolution of 10 seconds continuously for 3 minutes where the readings of the noise were automatically recorded and saved in the device memory. in each location (l1-l25), measurements were taken 2 times. at the end of each measurement, the average weighted noise value (laeq) was considered as a representative noise level. the social survey squ community noise perception levels were assessed across all the university populations including students, academic staff, and administrative personnel. based on the total number of the squ population n= 58,997 and considering a sample error of 5% at a 95% confidence interval, a sample size of 852 people was estimated (paiva et al. 2019). the questionnaires were designed using google forms in both arabic and english. the questionnaires were distributed to the respondents through the squ email system to the target respondents. a 4-point likert scale was used to assess the respondent's perception about the degree of sources of noise levels for both the selected indoor and outdoor microenvironments (table 2). when the question of "how would you rate each of the following; trucks, buses, traffic horns, speed vehicles, aircraft as sources of outdoor noise in squ, responses were evaluated by using simple metrics such as "very high", "high", "low " and " very low”. a similar question was also asked about the perception about the degree of potential sources of indoor noise levels in the university community. respondents were asked “ what is the degree of noise levels from air conditioning, laboratory instruments, vacuum cleaners, and maintenance activities in indoor environments in squ”. these questions were also evaluated by using the same 4-scale points responses such as "very high", "high", "low " and " very low”. health impact assessment in this study, the percentage of highly annoyed (%ha) and high sleep disturbed (%hsd) persons were quantified using exposure-response models recommended by who (2011). these models were developed based on the results of a meta-analysis from several epidemiological studies conducted across the globe. the %ha model eqn. (1) was based on lden and noise exposure levels ≥ 45 db and ≤ 75 db, noise levels that do not fall within these ranges were excluded due to the risk associated with limited very high, and low noise data. also, %hsd dose-response model was developed as a function of lnight noise patrick amoatey, issa al-harthy , khalifa al-jabri , abdullah al-mamun, and mahad said baawain 26 exposure levels eqn. (2) and was based on noise data within the range of 45-65 dba only due to high levels uncertainties associated with noise levels that do not fall within these defined ranges (who 2011). in addition, the association between exposure to noise levels and incidence of cardiovascular disease (cvd) was determined using the odds ratio (or) estimates as indicated by eqn.3. this model was developed from studies where day (lday) noise levels were limited to 55-80 dba (who 2011). the above health impact assessments were focused on the population living near squ residential and squh areas due to the longer durations people spend in these areas at night times (table 2). %ha = 9.868*10-4 (lden-42)3 – 1.436*10-2 (lden-42)2 + 0.5118 (lden-42) (1) %hsd = 20.8 -1.05 (lnight) + 0.01486 (lnight)2 (2) cvdor = 1.63-0.000613* (lday)2 + 0.00000736*(lday)3 (3) where, lden = lday +2.3 db, and lnight = levening 4.4 db according to conversion formulae developed by brink et al. (2017). 3. results measured noise levels in indoor microenvironments the study assessed environmental noise exposure levels within the squ community by focusing on indoor microenvironments (imes) considering the longer durations spent by the students and staff. as detailed in table 2, imes (l1-l13) include laboratories, classrooms, libraries, and enclosed corridors. table 3 indicates all the measured noise levels observed across the selected imes in squ for the morning, day, and evening times. out of the thirteen (13) selected imes (l1-l13), both morning and daytime noise levels in most laboratories (l3 = 57 and 56.7db, l11= 62.9 and 61.2 db) exceeded the who critical limits of 55 dba (table 3). similarly, noise produced in building corridors (l4 = 58.8 and 57.1db) and student services buildings (l13 = 57.3 and 58.1db) were also observed to be high. overall, all the measured ime noise levels at evening times satisfied the who's threshold of 55 db compared to morning/day times (table 3). the lower noise levels at the latter are due to the closure of university activities such as lectures, laboratories, and student services. overall, the average noise levels of all imes for the morning (53.8 db), noon (54.9 db), and evening (42.6 db) durations were very low compared to who's acceptable threshold limits (55 db). measured noise levels in outdoor microenvironments the study also estimated the outdoor noise exposure levels with the university community. twelve selected (l14-l25) outdoor microenvironments (omes) consisting of near roads, around staff residential apartments, squ hospital areas, gardens, and forecourt of buildings were considered. this is due to the presence of several road networks around these locations which are deemed as one of the main sources of community noise levels. during the morning and daytime, the noise levels near the majority of the roads (l14= 59.6 and 61.2 db, l16= 61.8 and 56 db) within the university were high, this was similar to the evening noise levels observed near the various roads (i.e l14 and l16) as indicated in table 4. however, near the hospital (l23= 67.9 and 60.4 db), residential apartments (l24= 62.3 and 67.8 db) and the university entrance (l25 = 64.5 and 62.8 db) areas showed the highest noise levels during the morning and day times (table 4). all these locations have exceeded the who’s limits of 55 db indicating that noise levels from road sources and around most residential locations including hospital areas pose serious environmental health concerns to the university community. the study also found that the majority of the noise levels during the evening times were very low, and were found below the thresholds limits (table 4). while noise level was high in the aforementioned areas, most of the remaining omes including l17, l18, l19, l21, and l22 had very low noise levels and were found acceptable per who limits with average noise pressure levels across all the omses slightly beyond the who limits of 55 db. perception of noise exposures a total of 698 individuals participated in the online noise survey, representing a response rate of 82% when compared to the estimated sample size of 852 people. table 5 indicates a brief socio-demographic profile and respondent perception about sources of noise within squ. in this study, the age range of the respondents who took part in the survey was within 18-65 years. the study found that majority of respondents were males (58.4%), students (73.5%) with about 25.9% representing employees/staff (table 5). regarding perception about outdoor noise levels within the squ community, an average of 29.8% are of the view that noise levels from traffic-related activities are high while 35.8% considered it to be low. out of these, the sources of noise levels due to both car speeding (33.8%) and intercampus bus activities (30.6%) including traffic horning (26.8%) were reported to be high and very high, respectively (table 5). contrastingly, the majority of the interviewees agreed that the contribution of noise from aircraft (35.3%) and buses (43.1%) within the squ community was low as provided in table 5. regarding noise levels from indoor environments, 27% and 23.7% of the respondents reported that indoor noise levels from maintenance and vacuum cleaners’ activities, respectively are high. however, noise generated as a result of the operation of laboratory equipment (49.9%) was found to be low and that of air conditions (30.6%) was very low. investigation of noise exposures, perception, and health effects in different microenvironments in a university community 27 high annoyed persons results of %ha persons due to noise exposures within the two main microenvironments (imes and omes) using who’s exposure-response relationships are shown in figures 2 and 3. overall, the highest %ha persons were found in omes compared to imes which may be attributed to high noise levels produced by road traffic. in imes, the highest %ha persons were mostly in exhibition areas (l9 =15%), construction materials and structures laboratories (l11=14%), and student service centres (l13 =11%). interestingly, in microenvironments that require sedentary activities and which subsequently tend to produce low noise levels, especially in computer laboratory (l12), and library (l10), the lowest %ha persons were 4% and 5%, respectively. for omes, residential areas (l24 =25%) and near the university entrance (l25=16%) which serve as main hotspots of vehicular noise revealed the highest %ha persons. this was followed by the hospital (l23 =13%) and near other individual road networks (l14 = 14%, l15=12%) linking to the various squ facilities. the study found that most omes especially l21 (3.3%), l18 (3.8%), l19 (3.8%) showed the lowest %ha persons compared to several imes as illustrated in figures 2 and 3. sleep disturbance and cardiovascular disease as indicated in table 6, the %hds persons were estimated using who’s exposure-response model in some specific omes such as near the hospital (l23) and residential areas (l24) of the university. the assessment of %hds persons was limited to only l23 and l24 since they serve as homes for the university staff/students and hospitalization of patients (i.e ipd capacity is about 31,436 patients) (table 1). the developed model revealed that %hds persons were 7.4% and 5.32%, for l24 and l23, respectively. the %hds person estimates were consistent with the measured noise levels observed in l24 (58.4 db) and l23 (54.1) during the evening times. the association between noise exposures and the incidence of cardiovascular diseases was also determined based on who's exposure-response model (table 6). the odds ratio (or) estimates from the model showed that the current noise levels at l24 may be associated with the incidence of cardiovascular diseases with an or of 1.11. however, in the case of l23 (or = 1.02), the exposure may not show an association with the incidence of cardiovascular diseases (table 6). figure 2. percentage of highly annoyed (%ha) persons in the indoor microenvironments. figure 3. percentage of highly annoyed (%ha) persons in the outdoor microenvironments. table 3. measured noise levels at different indoor microenvironments (imes) code a laeq (db) morning noon evening l1 50.3 51.6 42.5 l2 51.2 52.1 39.2 l3 57.0 56.7 46.8 l4 58.8 57.1 38.8 l5 54.2 52.3 43.3 l6 48.1 55.6 41.9 l7 56.0 55.0 38.6 l8 52.7 54.1 46.9 l9 52.9 61.1 46.8 l10 48.3 49.8 37.8 l11 62.9 61.2 38.6 l12 49.8 48.6 46 l13 57.3 58.1 46.9 average 53.8 54.9 42.6 a the italicized noise levels exceeded the who limits of 55 db table 4. measured noise levels at different outdoor microenvironments (omes). code a laeq (db) morning noon evening l14 59.6 61.2 56.3 l15 52.6 59.3 56.3 l16 61.8 56.0 49.0 l17 47.0 53.2 42.1 l18 48.3 48.0 44.0 l19 50.0 48.0 45.0 l20 51.6 53.4 46.2 l21 48.5 46.9 47.6 l22 55.0 49.2 43.9 l23 67.9 60.4 54.1 l24 62.3 67.8 58.4 l25 64.5 62.8 50.9 average 56.0 56.0 49.5 a the italicized noise levels exceeded the who limits of 55 db bitalics indicates values above the who limits 5.7 6.0 9.4 9.8 6.1 8.5 8.2 7.3 14.1 4.7 14.2 4.1 10.7 0.0 5.0 10.0 15.0 l1 l2 l3 l4 l5 l6 l7 l8 l9 l1 0 l1 1 l1 2 l1 3 % h a locations of indoor microenvironments 14.2 11.9 8.5 6.7 3.8 3.8 6.9 3.3 4.4 13.2 24.9 16.3 0.0 5.0 10.0 15.0 20.0 25.0 30.0 l14l15l16l17l18l19l20l21l22l23l24l25 % h a locations of outdoor microenvironments patrick amoatey, issa al-harthy , khalifa al-jabri , abdullah al-mamun, and mahad said baawain 28 table 5. respondents profiles, outdoor and indoor noise perceptions. indicators number (n) percentage (%) gender male 408 58.4 female 290 41.6 total 698 100 exposed groups students 513 73.5 employees 181 25.9 other residents 4 0.6 total 698 100 perception of noise a. outdoor responses traffic very high (%) high (%) low (%) very low (%) trucks 19.3 31.9 36 12.8 buses 15.1 30.6 43.1 11.2 traffic horns 26.8 29.4 30.5 13.4 speed of vehicles 21 33.8 34.1 11.2 aircraft 21 23.3 35.3 20.3 average 20.64 29.8 35.8 13.78 b. indoor responses none-traffic very high (%) high (%) low (%) very low (%) air conditioning 10.8 21 37.6 30.6 lab instruments 6.4 19.5 49.9 24.3 vacuum cleaners 11.4 23.7 39.5 25.4 maintenance activities 22.6 27 36.8 13.6 average 12.22 22.3 41.46 24.06 table 6. estimates of the percentage of highly disturbed sleep and the incidence of cardiovascular diseases from noise exposures. microenvironment %hsd odds ratio (or) cvd around the squ hospital (l23) 5.32 1.02 squ residential area (l24) 7.43 1.11 4. discussion of results the measured noise levels across the different microenvironments in the squ community showed relatively high levels ranging from 56 68 db for both morning and noon times. thus on average, closer to 50% of the individual university environments exceeded the critical levels of 55 db. the exceedance of these current noise levels has the potential of creating acoustic discomfort to the university community. a similar study carried out on a university campus in brazil also reported that more than 80% of the measured noise points exceeding the recommended limits (zannin et al. 2013). another noise exposure level exceeding 76 db was observed following oneweek continuous monitoring of noise in the university of jos, nigeria. the exceedance of about excees 21db of noise level compared to who’s limit of 55 db shows that noise is an important public health problem that requires urgent mitigation measures (akintunde et al. 2020). similar results were also found in a university campus in turkey where the average levels ( 62.7 dba) slightly exceeded the critical limits (ozer et al. 2014). the evidence from these studies calls for urgent needs for effective urban planning and sustainable land use management policies in schools, hospitals, and residential areas to manage the burden of community noise exposures. to reduce the impact of environmental noise levels on the general public, it is imperative to employ an acoustic mapping tool (zytoon 2016). this approach could show the dispersion of noise levels across several locations including sensitive areas such as schools and hospitals. thus, a noise map is deemed as an effective tool for noise pollution management as it could forecast and give alerts about current and future acoustic conditions of a particular location thereby safeguarding the health of the exposed population. although community noise is difficult to control due to its diverse sources (including road traffic, railway, aircraft, industries, etc), it is, therefore, important to adhere to the specific noise control guidelines that are established by who to help reduce health impacts relating to speech interference, sleeping difficulties, and disruption of tranquil environments (who 1999). in the case of this study, among the evaluated points, about 38.5% and 46.2% of indoor locations and 41.6% and 50% for outdoor areas for morning and afternoon times, respectively exceeded the critical levels of 55 db. it is therefore important for continuous evaluation of environment-specific (e.g. laboratories, corridors, residential apartments) noise exposure levels in university communities to aid in the reduction of source-specific noise pollution levels. the survey study among the squ community revealed that noise pollution from traffic (30%) and indoor activities (22%) was viewed as high, while about 44% of the respondents considered them as low. the 80% response rate mostly among students is an indication that community noise issues were well received. this is an important indicator for the successful implementation of future noise mitigation programs. thus, environmental regulations that could reduce vehicle population levels within the squ community may substantially reduce noise levels. the application of indoor building noise abatement technologies could also reduce noise levels in most buildings (garg et al. 2013). annoyance and sleeping difficulties due to community noise exposures have been recognized as investigation of noise exposures, perception, and health effects in different microenvironments in a university community 29 serious environmental health concerns. this is because, in an academic environment, excessive annoyance from noise exposures could interfere with comprehension in learning and cognitive performance levels of students (klatte et al. 2015). in university communities where there are kindergartens, despite children having well-developed psychophysiological reaction levels compared to adults, annoyance could disrupt their communication, reading, singing and can also affect their sleep (tesoriere et al. 2018). a study evaluated students' activity levels and academic achievements among 336 residence and 450 nonresidence students who have been exposed to > 60.3 dba (leq24h) of noise (onchang and hawker 2018). the study concluded that students’ overall grade point average scores were adversely affected by noise exposures due to frequent sleeping disturbances, reading difficulties, and poor cognitive performance levels compared to off-campus students. this study found that the %ha persons in the squ community were generally high in omes especially at the university residence (%ha = 25%) and hospital areas (%ha = 13%). in the same locations, sleep disturbances were highly affected by the former compared to the latter. however, imes in construction materials and structures laboratories (%ha = 14%) showed a similar % of high annoyance persons. community noise annoyance is attributed to discomfort, irritation, anxiety, and stress (di and xu 2017; khaiwal et al. 2016; pohl et al. 2018; who 1999), and these effects are also influenced by the economic and sociodemographic factors of the exposed population. a study conducted by (ahmed and ali 2017) assessed annoyance levels and hearing impairment problems in college campuses where maximum noise levels were in the range of 67-79 dba. the findings showed that the majority (80%) of students have faced some levels of annoyance problems. out of this, more than half reported some degree of hearing difficulties. based on these aforementioned health impacts, several studies have recommended acoustic control techniques and sustainable measures to control community noise exposure levels (ahmadi and dianat 2019; mukate 2013; tristán-hernández et al. 2016). furthermore, there is a growing body of evidence revealing the incidence of cardiovascular diseases (khosravipour and khanlari 2020), respiratory diseases (recio et al. 2016), and an increase in adiposity (an et al. 2018) from long-term exposure to the noise level. the health risk estimates from this study show that the current noise exposures levels are likely to cause the incidence of cardiovascular diseases in squ residential areas, however, no association or health risk was established within the hospital areas, despite being recognized as the most sensitive areas in the university community as far as noise pollution is a concern. these findings imply that most staff who have lived in squ for several years may be vulnerable to noise pollution-related health problems. while this study has assessed the noise exposure levels in both indoor and outdoor environments including, annoyance levels, sleep disturbances, and incidence of cardiovascular diseases, there are several limitations associated with this study. thus, the noise exposure assessment was short-term, and could not account for noise levels during vacations and school sessions to understand the temporal noise levels within the university. in addition, the dose-response model used in this study was developed from meta-analysis, the population of those studies may have different socioeconomic and demographic profiles compared to the exposed population under this study. these factors are therefore considered as important indicators for a particular noise exposure health outcome, sensitivity, irritation, annoyance, and sleep disturbance levels. it is expected that future studies would utilize clinical/health data including potential cofounders (e.g. meteorological factors, green space, economic, and socio-demographic data) to better understand the burden of environmental noise exposures among populations within university campuses and other communities where noise levels are considered as an important environmental health issue. 5. conclusion community noise exposures are recognized as a serious public health concern considering the annoyance, irritation, sleeping difficulties including cardiovascular and cardiometabolic diseases it poses to the general population. in this study, noise exposure assessment via field measurement was conducted across various indoor and outdoor environments in the squ community. perceptions of sources of noise levels, annoyance, sleep difficulties, and their effect on cardiovascular health were also assessed using both field survey and dose-response models. on average, about half of the measured points in both indoor and outdoor environments have exceeded the critical threshold levels. a significant number of the respondents identified traffic and indoor buildingrelated activities as main sources of noise levels but their impact was rated as low by the majority. however, more than a quarter of the respondents view traffic as the main contributor to university noise levels. the study found that highly annoyed persons were high in most indoor and outdoor microenvironments especially in residential areas and in construction materials and structures laboratories of the university. it was also revealed that persons who were affected by sleep disturbances were higher in residential areas compared to that of the hospital vicinity. this study also found an association between noise levels and incidence of cardiovascular diseases in residential areas of the university, however, no association was established within the hospital areas. this is the first study that has provided a high spatial variability noise exposure levels across a university environment in oman, this will contribute to designing patrick amoatey, issa al-harthy , khalifa al-jabri , abdullah al-mamun, and mahad said baawain 30 future mitigation strategies to improve the health and well-being of the exposed population. it is expected that future studies will employ a comprehensive epidemiological approach including the use of clinical/health data and their potential cofounders to better understand the health effects due to noise exposures. this will provide more reliable datasets to design mitigation techniques and intervention policies for reducing the burden of health impacts from community noise exposures. conflict of interest the authors wish to declare no competing conflict of interest. funding no funding was received for this research. references abdul-wahab sa, fadlallah so (2014) a study of the effects of vehicle emissions on the atmosphere of sultan qaboos university in oman. atmospheric environment 98:158-167 doi:10.1016/j.atmosenv.2014.08.049 ahmadi o, dianat i (2019) qualitative aspects of traffic noise in tabriz city, iran: effects, habituation, and possible improvements. international journal of environmental science and technology 16:2009-2016 doi:10.1007/s13762-017-1623-5 ahmed ho, ali wj (2017) noise levels, noise annoyance, and hearing-related problems in a dental college. archives of environmental & occupational health 72:159-165 doi:10.1080 /19338244.2016.1179169 akintunde ea, bayei jy, akintunde ja (2020) noise level mapping in university of jos, nigeria. geojournal doi:10.1007/s10708-019-10135-w al-dorzi h, al-matroud a, al-moamar t, itani a, alchin j, arabi y (2020) 1372: noise level in an adult intensive care unit. critical care medicine 48 al-khanjari z, al-kindi k, al-zidi a, baghdadi y (2014) m-learning: the new horizon of learning at squ. the journal of engineering research [tjer] 11:15-26 doi:10.24200/tjer.vol11iss2pp1526 al-mansour ai (2006) effects of pavement skid resistance on traffic accidents. the journal of engineering research 3:75-78 doi:10.24200/tjer. vol3iss1pp75-78 al harthy ih noise annoyance responses of students and teachers in different schools in muscat city, oman. in: institute of noise control engineering of the usa 35th international congress and exposition on noise control engineering, internoise 2006, 2006. pp 4541-4550 ali saa (2013) study effects of school noise on learning achievement and annoyance in assiut city, egypt. applied acoustics 74:602-606 doi: https://doi.org/10.1016/j.apacoust.2012.10.011 amoatey p, omidvarborna h, baawain ms, almamun a (2018) indoor air pollution and exposure assessment of the gulf cooperation council countries: a critical review. environment international 121:491-506 doi:https://doi.org /10.1016/j.envint.2018.09.043 amoatey p, omidvarborna h, baawain ms, almamun a (2020) evaluation of vehicular pollution levels using line source model for hot spots in muscat, oman. environmental science and pollution research 27:31184–31201 doi:10.1007/s11356-020-09215-z an r, wang j, ashrafi sa, yang y, guan c (2018) chronic noise exposure and adiposity: a systematic review and meta-analysis. american journal of preventive medicine 55:403-411 doi:https://doi.org/10.1016/j.amepre.2018.04.040 bloemsma ld et al. (2019) the associations of air pollution, traffic noise and green space with overweight throughout childhood: the piama birth cohort study. environmental research 169:348-356 doi:https://doi.org/10.1016/j.envres.2018.11.026 brink m, schäffer b, pieren r, wunderli j (2017) conversion between noise exposure indicators leq24h, lday, levening, lnight, ldn and lden: principles and practical guidance. international journal of hygiene and environmental health 221 doi:10.1016/j.ijheh.2017.10.003 caviola s, visentin c, borella e, mammarella i, prodi n (2021) out of the noise: effects of sound environment on maths performance in middleschool students. journal of environmental psychology 73:101552 doi:https://doi.org/10.1016/j.jenvp.2021.101552 chew yr, wu bs (2016) a soundscape approach to analyze traffic noise in the city of taipei, taiwan. computers, environment and urban systems 59:78-85 doi:https://doi.org/10.1016/j.compenvurbsys.2016. 05.002 di g, xu y (2017) influences of combined traffic noise on anxiety in mice. science of the total environment 579:1439-1445 doi:https://doi.org /10.1016/j.scitotenv.2016.11.144 díaz j, lópez-bueno ja, lópez-ossorio jj, gónzález jl, sánchez f, linares c (2020) short-term effects of traffic noise on suicides and emergency hospital admissions due to anxiety and depression in madrid (spain). science of the total environment 710:136315 doi:https://doi.org/10.1016/j.scitotenv. 2019.136315 dreger s, meyer n, fromme h, bolte g (2015) environmental noise and incident mental health investigation of noise exposures, perception, and health effects in different microenvironments in a university community 31 problems: a prospective cohort study among school children in germany. environmental research 143:49-54 doi:https://doi.org/10.1016 /j.envres.2015.08.003 dzhambov am, dimitrova dd (2017) children's blood pressure and its association with road traffic noise exposure – a systematic review with metaanalysis. environmental research 152:244-255 doi:https://doi.org/10.1016/j.envres.2016.10.024 dzhambov am, dimitrova dd (2018) residential road traffic noise as a risk factor for hypertension in adults: systematic review and meta-analysis of analytic studies published in the period 2011–2017. environmental pollution 240:306-318 doi:https://doi.org/10.1016/j.envpol.2018.04.122 extech (2015) digital datalogging sound level meter model hd600. extech. http://www.extech.com/products/resources/hd600 _um-en.pdf. accessed 29/07/2020 foraster m et al. (2017) exposure to road, railway, and aircraft noise and arterial stiffness in the sapaldia study: annual average noise levels and temporal noise characteristics. environmental health perspectives 125:097004 doi:10.1289/ehp1136 garg n, kumar a, maji s (2013) significance and implications of airborne sound insulation criteria in building elements for traffic noise abatement .applied acoustics 74:1429-1435 doi:https:// doi.org/10.1016/j.apacoust.2013.05.012 henshall d et al. (2018) a prospective observational study of emergency department presentations following novel psychoactive substance use. scottish medical journal 63:003693301876076 doi:10.1177/0036933018760761 kamal nm, jinca y, wunas s (2020) the effect of traffic volume on noise of housing and shopping area of manggala district. iop conference series: earth and environmental science 473:012127 doi:10.1088/1755-1315/473/1/012127 khaiwal r, singh t, tripathy jp, mor s, munjal s, patro b, panda n (2016) assessment of noise pollution in and around a sensitive zone in north india and its non-auditory impacts. science of the total environment 566-567:981-987 doi:https:// doi.org/10.1016/j.scitotenv.2016.05.070 khosravipour m, khanlari p (2020) the association between road traffic noise and myocardial infarction: a systematic review and meta-analysis. science of the total environment 731:139226 doi:https://doi.org/10.1016/j.scitotenv.2020.13922 6 klatte m, spilski j, mayerl j, moehler u, lachmann t, bergström k (2015) effects of aircraft noise on reading and oral language abilities in german children near frankfurt/main airport: results of the norah (noise-related annoyance, cognition, and health)-study. conference euronoise, maastricht klompmaker jo et al. (2019) associations of combined exposures to surrounding green, air pollution and traffic noise on mental health. environment international 129:525-537 doi:https://doi.org/10.1016/j.envint.2019.05.040 liu y, xia b, cui c, skitmore m (2017) community response to construction noise in three central cities of zhejiang province, china. environmental pollution 230:1009-1017 doi:https://doi.org/10. 1016/j.envpol.2017.07.058 margaritis e, kang j, filipan k, botteldooren d (2018) the influence of vegetation and surrounding traffic noise parameters on the sound environment of urban parks. applied geography 94:199-212 doi:https://doi.org/10.1016/j.apgeog.2018.02.017 mukate s (2013) street traffic noise in university of pune indian. journal of environmental protection 33:100-108 okolie c, alademomi a, ojegbile b, onyegbula j, adepo r, ademeno w (2020) spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos .the journal of engineering research [tjer] 17:75-88 onchang r, hawker dw (2018) community noise exposure and annoyance, activity interference, and academic achievement among university students. noise and health 20:69-76 doi:10.4103/ nah.nah_54_17 ow lf, ghosh s (2017) urban cities and road traffic noise: reduction through vegetation. applied acoustics 120:15-20 doi:https://doi.org /10.1016/j.apacoust.2017.01.007 ozer s, zengin m, yilmaz h (2014) determination of the noise pollution on university (education) campuses: a case study of ataturk university. ekoloji 23:49-54 doi:10.5053/ekoloji.2014.906 paiva km, cardoso mra, zannin pht (2019) exposure to road traffic noise: annoyance, perception and associated factors among brazil's adult population. science of the total environment 650:978-986 doi:https://doi.org/10.1016/j.scitotenv.2018.09.041 pierrette m, marquis-favre c, morel j, rioux l, vallet m, viollon s, moch a (2012) noise annoyance from industrial and road traffic combined noises: a survey and a total annoyance model comparison. journal of environmental psychology 32:178-186 doi:https://doi.org/10.1016/j.jenvp.2012.01.006 pohl j, gabriel j, hübner g (2018) understanding stress effects of wind turbine noise – the integrated approach. energy policy 112:119-128 doi:https://doi.org/10.1016/j.enpol.2017.10.007 ramazani me, mosaferi m, rasoulzadeh y, pourakbar m, jafarabadi ma, amini h (2018) temporal and spatial evaluation of environmental noise in urban area: a case study in iran. international journal of environmental science and technology 15:11791192 doi:10.1007/s13762-017-1462-4 recio a, linares c, banegas jr, díaz j (2016) road traffic noise effects on cardiovascular, respiratory, and metabolic health: an integrative model of biological mechanisms. environmental research patrick amoatey, issa al-harthy , khalifa al-jabri , abdullah al-mamun, and mahad said baawain 32 146:359-370 doi:https://doi.org/10.1016/j.envres.2015.12.036 rogers ae, hwang w-t, scott ld, aiken lh, dinges df (2004) the working hours of hospital staff nurses and patient safety. health affairs 23:202212 doi:10.1377/hlthaff.23.4.202 chloé sieber, martina s.ragettli, mark brink, toyib olaniyan, roslyn baatjies, apolline saucy, danielle vienneau, nicole probst-hensch, mohamed aqiel dalvie, martinröösli (2018) comparison of sensitivity and annoyance to road traffic and community noise between a south african and a swiss population sample. environmental pollution 241:1056-1062 doi:https://doi.org/10.1016/j.envpol.2018.06.007 rachel b. smith, sean d. beevers, johngullivera, daviddajnak, danielafecht, martablangiardo, margaret douglass, anna l.hansell, h. ross anderson, frank j.kelly, mireille b.toledano (2020) impacts of air pollution and noise on risk of preterm birth and stillbirth in london. environment international 134:105290 doi:https://doi.org /10.1016/j.envint.2019.105290 tesoriere g, campisi t, canale a, severino a (2018) the effects of urban traffic noise on children at kindergarten and primary school: a case study in enna. aip conference proceedings 2040:140005 doi:10.1063/1.5079194 jesse d. thacher, ulla a. hvidtfeldt, aslak h. poulsen, ole raaschou-nielsen, matthias ketzel, jørgen brandt, steen s. jensen, kim overvad, anne tjønneland, thomas münzel, mette sørensen (2020) long-term residential road traffic noise and mortality in a danish cohort. environmental research 187:109633 doi:https://doi.org/10.1016/ j.envres.2020.109633 tristán-hernández e, pavón i, cantón i, kolosovas e, lopez navarro jm (2016) evaluation of psychoacoustic annoyance and perception of noise annoyance inside university facilities. international journal of acoustics and vibration (ijav) 23 doi:10.20855/ijav.2018.23.11059 umbas jcg, bintang ak, aulina s, bahar a, akbar m (2021) the effect of white noise on high school students’ sleep quality at unit b of rajawali girls dormitory makassar. medicina clínica práctica 4:100209 doi:https://doi.org/10.1016/j.mcpsp.2021.100209 upso (2017) university profile statistical overview (upso). sultan qaboos university https://www.squ.edu.om/portals/0/pdf/facts%20a nd%20figures.pdf?ver=2019-04-09-133458-170. accessed 17/09/2021 who (1999) guidelines for community noise. world health organisation. accessed 23/07/2020 who (2011) burden of disease from environmental noise : quantification of healthy life years lost in europe. world health orgnisation regional office for europe. https://www.who.int/ quantifying_ehimpacts/publications/e94888/en/. 16/05/2021 woolner p, hall e (2010) noise in schools: a holistic approach to the issue. international journal of environmental research and public health 7:3255-3269 doi:10.3390/ijerph7083255 yassin mf, almutairi h, alhajeri n, al-fadhli f, al rashidi ms, shatti t (2016) assessment of noise exposure and associated health risk in school environment. international journal of environmental science and technology 13:20112024 doi:10.1007/s13762-016-1035-y zannin pht, engel ms, fiedler pek, bunn f (2013) characterization of environmental noise based on noise measurements, noise mapping and interviews: a case study at a university campus in brazil. cities 31:317-327 doi:https://doi.org/10.1016/j.cities.2012.09.008 zare sakhvidi mj, zare sakhvidi f, mehrparvar ah, foraster m, dadvand p (2018) association between noise exposure and diabetes: a systematic review and meta-analysis. environmental research 166:647-657 doi:https://doi.org/10.1016/j.envres.2018.05.011 zaw ak, myat am, thandar m, htun ym, aung th, tun km, han zm (2020) assessment of noise exposure and hearing loss among workers in textile mill (thamine), myanmar: a crosssectional study. safety and health at work 11:199206 doi:https://doi.org/10.1016/j.shaw.2020.04.002 zytoon ma (2016) opportunities for environmental noise mapping in saudi arabia: a case of traffic noise annoyance in an urban area in jeddah city. international journal of environmental research and public health, 13(5), 496 doi:10.3390/ijerph13050496 1. introduction 2. research methodology study area study population noise assessment areas noise measurements the slm used in this study was certified by international electrotechnical commission (iec) as class ii and american national standards institute (ani) as a type ii noise measurement instrument (zaw et al. 2020). in-house calibration was done on slm... the social survey health impact assessment 3. results measured noise levels in indoor microenvironments measured noise levels in outdoor micro-environments perception of noise exposures high annoyed persons 4. discussion of results 5. conclusion conflict of interest funding references civl50327.qxd the journal of engineering research vol. 4, no.1 (2007) 11-16 1. introduction foundation design should satisfy two criteria; one deals with ultimate bearing capacity of the soil under the foundation and the other is concern with the limit of soil deformation. an excessive settlement, may induce a differential settlement, which causes structural distresses of the superstructure. thus, it is required in many conditions to reduce the settlement to acceptable limits. the improvement of the bearing capacity of subgrades and, hence, modifying the stress-strain relationship by utilization of horizontal reinforcement has been reported by binquet and lee (1975). they studied the effect of horizontal reinforcement of soils on the bearing capacity of ________________________________________ *corresponding author’s e-mail: aghbari1@squ.edu.om foundations by conducting a series of experiments. akinmusuru and akinbolade (1981), fragaszy and lawton (1984), and verma and char (1986) have also contributed to this investigation with their experimental studies. samtani and sonpal (1989), and mahmoud and abdrabbo (1989) have studied the effect of vertical reinforcements of the soil on the bearing capacity of foundations. they conducted a series of tests, and concluded that the bearing capacity of a foundation can be increased to about one-and-a-half to two times that of the unreinforced soil. they also suggested this method for improving the bearing capacity of existing foundations. various methods of soil stabilization are known. however, they are sometimes prohibitively expensive and restricted by site specific conditions. in some cases they are difficult to apply to existing foundations. it is, theresettlement of shallow circular foundations with structural skirts resting on sand m.y. al-aghbari dept. of civil and architectural engineering, sultan qaboos university, p.o. box 33, al-khoud 123, muscat, oman received 27 march 2005; accepted 10 april, 2006 abstract: this paper presents the findings of an experimental study concerning a method of reducing the settlement of shallow circular foundations on sand. it involves the use of structural skirts fixed to the edges of foundations. the experiments were performed in a large tank setting and the footing was instrumented in order to measure normal stresses and settlement. a series of tests were conducted to study the settlements of a circular footing with and without structural skirts. test results indicate that this type of reinforcement reduces the settlement of subgrade and modifies the stress-displacement behaviour of the footing. a settlement reduction factor (srf) was proposed, which takes into account the influence of various parameters that affect settlements. results show that the use of structural skirts can produce enhanced settlement reduction in the range of 0.1 to 1.0 depending on stress applied and skirt depth. given these levels of settlements reduction, it is concluded that the use of structural skirts to reduce the settlement of shallow foundations on dense sand is of practical significance. further testing is recommended for different foundation shapes with structural skirts resting on different soil types. keywords: structural skirts, foundations, settlement, shallow, circular, sand ∫éeôdg ≈∏y i~æà°ùÿÿg á«fé°ûff’g ±gƒ◊g ™e á«ë£°ùddg ájôfg~dg ~ygƒ≤dg ∫hõf …èz’g .… .ω ::áá°°uuóóÿÿgg.i~yé≤dg áaé◊ á≤°sóe á«fé°ûfg ±gƒm ωg~îà°séh »æ©j á≤jô£dg .∫éeôdg ¥ƒa ájôfg~dg ~ygƒ≤dg ∫hõf π«∏≤j á≤jô£h ≥∏©àj ájèfl á°sgqo èféàf ¢vô©j ábqƒdg √òg èféàædg .á«fé°ûf’g ±gƒ◊g òz øeh ™e ájôfg~dg ~ygƒ≤dg ∫hõf á°sgq~d âjôlcg ~b üqééàdg øe áyƒª› .∫hõædgh §¨° dg ¢sé«≤d âàñk i~yé≤dgh òñc ¿gõn ‘ äòøf üqééàdg ‘ ôkƒj »àdg πegƒ©dg ¿éñ°ù◊g ‘ ònéj …òdgh ∫hõædg π«∏≤j πeéy ìîbcg .i~yé≤∏d ∫hõædgh §¨° dg úh ábó©dg ø°ùëjh áhîdg qhõf π∏≤j ájƒ≤àdg øe ´ƒædg gòg ¿g âæ«h òn’g ™e .á«fé°ûf’g ±gƒ◊g ∫ƒwh §¨° dg iƒb ≈∏y ~ªà©jh 1^0 ‹g 0^1 úh ée ∫hõædg ‘ òñc π«∏≤j èàfƒj ¿g øµá á«fé°ûf’g ±gƒ◊g ωg~îà°sg ¿g âàñkg èféàædg .∫hõædg ìî≤j .á«≤«ñ£j öfgƒl é¡d áø«ãµdg ∫éeôdg ‘ á«ë£°ùdg ~ygƒ≤dg ∫hõf øe π«∏≤à∏d á«fé°ûf’g ±gƒ◊g ωg~î°sg ¿g êéàæà°s’g éææµá ∫hõædg π«∏≤j øe áñ°ùædg iòg qéñày’éh .áø∏àfl áhôj ¥ƒa ~æà°ùjh á«fé°ûf’g ±gƒ◊g ™e áø∏àfl ~ygƒ≤d á«aé°veg üqéœ agôlg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .πeq ,…ôfgo ,»ë£°s ,∫hõf ,~ygƒb ,á«fé°ûfg ±gƒm : 12 the journal of engineering research vol. 4, no.1 (2007) 11-16 fore, suggested that an alternative approach is required for improving the bearing capacity of soils and reducing the settlements. this approach is based on the use of "structural skirts" fixed to the edges of the foundation. this method of improvement does not require an excavation of the soil and, thus, it is not restricted by the presence of a high water table. structural skirts fixed to the edges of shallow foundations have been used for a considerable time, principally to increase the "effective depth" of the foundations in marine and other situations where water scours may be a problem. bransby and randolph (1998) have provided a valuable detailed consideration of their use in such applications. however, the use of such structural skirts in conjunction with conventional shallow structural foundations has not been widely employed, nor have the improvements in settlement resulting from their use been investigated in detail. on this basis, an experimental study was undertaken to establish the degree of improvement in settlements to be gained from increasing the effective depth of shallow foundations by fixing structural skirts to their edges. in this paper, the background, methodology and findings of an experimental study of the settlement of shallow circular foundations with structural skirts on sand subjected to central vertical loads are presented. the various factors influencing the settlement improvements to be derived from the use of these skirts are identified. the benefits identified are combined into a parameter called settlement reduction factor (srf), and it is suggested that this parameter can be introduced into the settlement prediction equation for circular foundations resting on sand. 2. test apparatus and procedure an experimental investigation was carried out using a laboratory model. the experimental set-up consists of a rigid test tank, sand raining apparatus, a loading mechanism, an instrumented footing and a data acquisition system. the principal dimensions and layout of the apparatus are shown in fig. 1. the tank was 1000 x 1000 mm in cross section and 800 mm in height. it was made of wood and was stiffened with steel frames. the circular footing was a metal plate that had a diameter of 120 mm and a thickness of 30 mm. the tank was about 8 times the diameter of the footing and it was envisaged that there was no edge effects or that they would be minimal, if there any. the depth of the foundation was adjustable according to the requirement of the surcharge depth (df). at the middle top of the foundation, a small recess was made for locating a metal ball for the application of the load. the ball allows the foundation to tilt. the sand was placed in the tank by a sand raining technique. a special sand raining apparatus was designed to suit the tank size. this sand raining technique causes the sand particles to fall on the whole bed area from a receptacle placed above the receiver tank (butterfield and andrawes, 1970). a range of densities could be obtained by altering the intensity and the height of fall of the sand. in order to eliminate the effect of the later, this distance was kept constant at 800 mm in order to achieve a constant density. furthermore, a series of tests were carried out to check the uniformity of the sand bed using density pots, which were placed in various parts of the tank. for the foundations with skirts, the skirts were placed first in a vertical position on the top of the pre-placed sand. after the skirts were placed in the tank, sand was deposited using the same procedure as outlined above. then the sand raining was stopped, when the required sand level reached the level of foundation base. finally, the surface of sand was levelled and the foundation was placed on top of the skirt. the skirts comprised stanley steel plates with a modulus of elasticity (e) of 210 kn/mm2. the thickness of the skirt was 6 mm and it was welded to form an open cylinder shape. the footing was placed on the top of the skirt. the foundation was loaded by means of a motorized 10 ton capacity hydraulic jack at a constant rate of displace(b) figure 1. (a) dimensions of testing tank and (b) set-up of the loading celll and lvdts on the footing 120 mm 420 mm footing skirt 800 mm sand 1000 mm (a) 13 the journal of engineering research vol. 4, no.1 (2007) 11-16 ment of 12 mm per hour. the applied load was measured using a load cell of 10 kn capacity placed on the top of the footing. two lvdt transducers having a minimum resolution of 0.04 mm were placed at two different locations on top of the footing for measuring the settlement. the output voltage of each individual electrical measuring circuit was recorded automatically in one minute intervals using a data logging system. 3. material a coarse river sand was used throughout this experimental investigation. the particle size distribution of the sand is shown in fig. 2. the sand has d10, d30, d60, and d50 vales of 0.45, 0.65, 0.85 and 0.80 mm, respectively. these yeild a coefficient of uniformity (cu) of 1.89 and a coefficient of curvature (cc) of 1.11. thus the sand was uniform or poorly-graded. the unit weight of the sand in the tank was 16.5 kn/m3 and specific gravity was 2.65. in order to investigate the strength properties of the sand, triaxial tests were carried out to determine the angle of internal friction. the sand specimens were 38 mm in diameter and 76 mm high. different values of confining pressures were used in the range 40 to 200 kn/m2. the samples preparation and test procedure were according to bs.1377:1990. the results showed that the average peak friction angle was 42o for all ranges of confining pressures. 4. results of tests without structural skirts in order to establish that the data from the tank test facilities were reproducible and in conformity with the findings of previous investigations, two tests were conducted on surface foundations (df/b = 0) without structural skirts and another two tests for shallow foundations (df/b = 0.5). in all cases, footings without structural skirts were tested. the stress-settlement ratio relationships for the tests are shown in fig. 3. the settlement ratio is defined as the ratio between the settlement of the footing (s) to the footing width (b).the overall behaviour observed shows that the two tests for both footing depths gave similar results. for the same depth, the difference between the two tests was within 2%. therefore, the reproducibility of the test results was considered to be satisfactory. thus, they may be used as the basis for determining the improvements to be derived from the use of structural skirts. the results from the two surface footings tests were further analyzed and compared with the predicted values using some of the existing methods by terzaghi and peck (1967), schmertmann (1970), bazaraa (1967), and meyerhof (1965). to predict the settlement of footing by these methods, the standard penetration test (n) value of the sand was required. the n value was determined using the relationship between angle of friction and n-value proposed in the literature (das, 1999). a comparison between the experimental settlement at values 0.5 peak load and theoretical predictions of a surface footing (df/b = 0 ) are presented in table 1. shallow foundations are generally designed with safety factors of 2-3 with respect to bearing capacity. a factor of safety of 2 implies 50% of the peak load and, hence, the 0.00 20.00 40.00 60.00 80.00 100.00 0.01 0.1 1 10 grain size (mm) p er ce nt p as si ng figure 2. grain size distribution of the sand 0 50 100 150 200 250 300 0 5 10 15 20 s/b % a p p li e d s tr e s s ( k n /m2 ) tes t 1a, df /b= 0.0 tes t 1b, df /b= 0.0 tes t 2a, df /b= 0.5 tes t 2b, df /b= 0.5 s/b % 0 5 10 15 20 300 250 200 150 100 50 0 a pp lie d st re ss ( kn /m 2 ) figure 3. stress-settlement relationship of tests without skirts method settlement calculated, scal (mm) scal / smeasured terzaghi and peck (1967) 0.16 0.71 bazaraa (1967) 0.22 0.99 schmertmann (1970) 0.25 1.13 meyerhof (1965) 0.84 3.7 table 1. comparison between calculated and measured values of settlement of 0.5 peak load of a surface footing grain size (mm) pe rc en t p as si ng 14 the journal of engineering research vol. 4, no.1 (2007) 11-16 0.5 peak load is used for comparing actual and calculated settlements. it can be concluded that the terzaghi and peck method (1967) underestimates the settlement of the model footings. similar results were also reported by hanna and abdel-rahman (1987), and ameen (1991). the methods proposed by schmertmann's (1970) and meyerhof (1965) overestimate the settlement, while the method proposed by bazaraa (1967) gave the closest values to those obtained in this study. based on an extensive analysis of settlement results, sivakugan et al. (1998) reported that the methods proposed by terzaghi and peck (1967) and schmertmann (1970) overestimate the settlements by 218% and 339%, respectively. as reported by sivakugan and johnson (2002, 2004), the scatter associated with the settlement prediction methods is quite high. they proposed a probabilistic approach to take this scatter into consideration. 5. results of tests with structural skirts the second part of the test program involved the testing of the footing with structural skirts. several experimental tests were carried out for different skirt depth ratios (ds/b) of 0.05, 0.25, 0.5, 0.75, 1.0, 1.25 and 1.5. the stress-settlement relationships data obtained from these tests are shown in fig. 4. data were further analyzed and compared with the results obtained from the tests for a surface footing (df/b=0) and a shallow footing (df/b=0.5) without the skirt. the settlement of the footing at a given stress with and without skirt, ss and sf respectively, was obtained from the stress-settlement relationship. the settlement reduction factor (srf), which is defined as srf = ss/sf , was used throughout this paper to assist in comparing the results. the tests results for a surface footing (df/b =0) and a shallow footing (df/b=0.5) at different skirt depths are summarized in tables 2 and 3, respectively. the comparison of settlement for a footing with and without skirt is made for different stress levels at the peak load. from the comparisons given in the tables 2 and 3, it is clear that structural skirts have a significant influence on the reduction of footing settlement at surface and shallow depths. the results show that for a given depth of a skirt, the reduction factor generally decreases as the stress increases. this is because of non-linearity of the stressstrain relationship of the soil (see figs. 3 and 4). in fig. 3, the footing without skirts curved more to the right for higher stresses while the curves for footing with skirts (fig. 4) were less curved and they were more in the elastic range. using a regression analysis with a coefficient of correlation r2 = 0.95, an equation was proposed to calculate the settlement reduction factor (srf) as: where ds/b : skirt depth ratio σ : stress applied on the footing ss/b % a pp lie d re ss ur e (k n /m 2 ) figure 4. stress-settlement relationship of tests with skirts table 2. srf for surface footing at different stress levels ) b sd18.0( esrf σ− = test no. ds/b stress (kn/m2) sf ss srf 1a 0 25 50 75 100 0.7 1.9 3.0 5.0 1.00 1.00 1.00 1.00 3 0.05 25 50 75 100 0.7 1.9 3.0 5.0 0.65 1.70 2.80 4.20 0.92 0.89 0.93 0.84 4 0.25 25 50 75 100 0.7 1.9 3.0 5.0 0.55 1.40 2.00 3.00 0.78 0.74 0.67 0.60 5 0.5 25 50 75 100 0.7 1.9 3.0 5.0 0.45 0.90 1.10 1.80 0.64 0.47 0.37 0.36 6 0.75 25 50 75 100 0.7 1.9 3.0 5.0 0.40 0.60 0.70 0.90 0.57 0.31 0.23 0.18 7 1.0 25 50 75 100 0.7 1.9 3.0 5.0 0.30 0.40 0.43 0.60 0.43 0.21 0.14 0.12 8 1.25 25 50 75 100 0.7 1.9 3.0 5.0 0.20 0.25 0.30 0.50 0.28 0.13 0.10 0.10 9 1.5 25 50 75 100 0.7 1.9 3.0 5.0 0.13 0.20 0.22 0.32 0.18 0.11 0.07 0.06 15 the journal of engineering research vol. 4, no.1 (2007) 11-16 a comparison between the proposed equation and the experimental results was made and the results are shown in fig. 5. it is clear that the proposed equation is consistent with the experimental results and thus, it can be used to estimate the settlement reduction factor (srf) of footings with skirts. 6. conclusions and recommendations references akinmusure, j.o. and akinbolade, j.a., 1981, "stability of loaded footings on reinforced soil," j. of geotechnical engineering division, american society of civil engineering, vol. 107(gt6), pp. 819827. ameen, s. f., 1991, "settlement of inclined loaded footings on sand layers," proceedings of the 9th asian regional conference on soil, pp. 197-200. bazaraa, a.r., 1967, "use of the standard penetration test for estimating settlements of shallow foundations on sand," ph.d. thesis, university of illinois, usa. binquet, j. and lee, k. l., 1975, "bearing capacity analysis of reinforced earth slabs," j. of geotechnical engineering division, american society of civil engineering, vol. 101(gt12), pp. 1257-1275. bransby, m.f. and randolph, m.f., 1998, combined loading of skirted foundation," geotechnique, vol. 48(5), pp. 637-655. british standard 1377, 1990, method of test for soils for test no. ds/b stress (kn/m2) sf ss srf 2 0 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 1 1 1 1 1 1 1 1 5 0.5 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 0.45 0.70 1.10 1.50 2.00 2.20 2.40 2.60 0.56 0.47 0.47 0.42 0.36 0.36 0.37 0.37 6 0.75 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 0.40 0.50 0.60 0.60 1.20 1.80 1.80 2.00 0.50 0.33 0.26 0.17 0.22 0.30 0.27 0.28 7 1.0 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 0.30 0.40 0.43 0.45 0.70 1.00 1.20 1.50 0.38 0.27 0.19 0.13 0.13 0.17 0.18 0.21 8 1.25 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 0.20 0.25 0.28 0.30 0.50 0.70 0.90 1.00 0.25 0.17 0.12 0.08 0.09 0.11 0.14 0.14 9 1.50 025 050 075 100 150 175 200 230 0.80 1.50 2.30 3.50 5.50 6.00 6.50 7.00 0.13 0.20 0.22 0.32 0.40 0.60 0.70 0.80 0.16 0.13 0.10 0.09 0.07 0.10 0.11 0.08 table 3. srf for shallow footing (df/b=0.5) at different stress levels 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 empirical settlement reduction factor ( srf) e xp er im en ta l s et tle m en t r ed uc tio n f ac to r ( s r f ) df/b=0 df/b =0.5 empirical settlement reduction factor (srf) e xp er im en ta l s et tle m en t r ed uc tio n fa ct or ( sr f) figure 5. relationship between the proposed empirical srf and the experimental srf • test results indicate that the use of structural skirts reduces the settlement of footings and modifies the load–displacement behaviour. • the terzaghi and peck method underestimates the settlement significantly by about 30%. • a simple linear equation is proposed to estimate the settlement reduction factor. the experimental results have a good correlation with the proposed equation. • further investigations are recommended for different foundation shapes with structural skirts resting on different soil types. 16 the journal of engineering research vol. 4, no.1 (2007) 11-16 civil engineering purposes, british standard institution, london. butterfield, r. and andrawes, k.z., 1970, “an air activated sand spreader for forming uniform sand beds," geotechnique, vol. 20(1), pp. 97 100. das, p. m., 1999, "principles of foundation engineering. fourth edition," pws publishing. fragaszy, r.j. and lawton, e., 1984, "bearing capacity of reinforced sand subgrades," j. of geotechnical engineering division, american society of civil engineering, vol. 110(10), pp. 1500 1507. hanna, a..m. and abdel-rahman, m., 1987, "experimental evaluation of foundation settlement in sand," proc. int. conf. foundations and tunnels, university of london, vol. 1, pp. 34-37. mahmoud, m.a and abdrabbo, f.m., 1989, "bearing capacity on strip footing resting on reinforced sand subgrades," canadian geotechical journal, vol. 26, pp. 154 159. meyerhof, g.g., 1965, "shallow foundations," j. of soil mechanics and foundation engineering, asce, vol. 91( sm2), pp. 21-31. samtani, n.c. and sonpal, r.c., 1989, "laboratory tests of strip footing on reinforced cohesive soil," j. of geotechnical engineering division, asce, vol. 115(9), pp. 1326-1330. schmertmann, j.h., 1970, "static cone to compute static settlement over sand," j. of soil mechanics and foundation engineering, asce, vol. 96(se3), pp. 1011-1043. sivakugan, n., eckersley, j.d. and li, h., 1998, "settlement predictions using neural networks," australian civil engineering transactions, ce 40, pp. 49-52. sivakugan, n. and johnson, k., 2002, "probabilistic design chart for settlements of shallow foundations in granular soils," australian civil engineering transactions, ce 43, pp. 19-24. sivakugan, n. and johnson, k., 2004, "settlement predictions in granular soils: a probabilistic approach," geotechnique, vol. 54(7), pp. 499-502. terzaghi, k. and peck, r.b., 1967, "soil mechanics in engineering practice," john wiley and son inc. verma, b.p. and char, a.n.r., 1986, "bearing capacity test on reinforced sand subgrades," j. geotechnical engineering division, asce, vol. 112(7), pp. 701706. microsoft word paper 2.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 115-123 experimental investigation of biogas production from kitchen waste mixed with chicken manure h. mousaa,b,* a. obaidatb, h.b. khaledb, a. alawanehb and a. tarawnehb a department of petroleum and chemical engineering, college of engineering, sultan qaboos university, oman. b department of chemical engineering, jordan university of science and technology, jordon. received 23 october 2015; accepted 8 december 2015 abstract: biogas produced from solid kitchen waste (kw) mixed with chicken manure (m) at different mass ratios was investigated. the effect of the ratio of the amount of water to the mixed solid waste on the amount of biogas produced was studied. the results showed that at a fixed ratio of water-to-solid waste, the amount of biogas increased as the amount of chicken m increased. at a fixed m-to-kw ratio, the amount of biogas produced increased as the solid content increased and then decreased, reaching its maximum value at a solid waste-to-water ratio of 1:1. the ph of the bioreactor containing the kw-m mixture dropped with time, resulting in a decrease in the amount of biogas produced. controlling the ph value by titrating with naoh solution improved the production of biogas. investigating biogas produced from sludge showed that the ph of the reactor slightly decreased and then increased slightly. the results also showed that the amount of biogas produced from sludge containing 3% solid waste was larger than the amount produced from sludge containing 6% solid waste. keywords: biogas, waste to energy, methane, renewable energy, bioreactor, fermentation, biodegradation. אאאאא  ،،،א،،אאא،      אkאאאאאאאאw אkאאאאאאאאא      kא     א א א א   א א א א ،אאאאאאאאאא  א   א א א א      ١w١  א    k אאא،אאאאאאא א א  א    kא א  א    אאאאkאאא אkאאאאאאאא       א א    ٣א   א     ٪ ٦k٪    אא،،،،،א،w.    * corresponding author’s email: hasana@squ.edu.om. h. mousa, a. obaidat, h.b. khaled, a. alawaneh and a. tarawneh  116 1. introduction one of the most important factors in human wellbeing is energy. providing sufficient amounts of reasonably priced energy is necessary to reduce poverty, improve life conditions, and promote living standards. currently, the main source of energy for the world’s industries and individuals is fossil fuel; however, as it is a nonrenewable resource, it is becoming more and more precious. another concern related to the use of fossil fuels is the environmental pollution it causes, resulting in global warming, and water, air, and soil contamination. as such, enormous efforts are being directed towards using renewable energy sources, mainly solar, wind, and bioenergy, in place of fossil fuels. biogas production from solid waste mainly in the form of kitchen waste (kw), has received special attention by governments, societies, and researchers. the literature contains vast amounts of research focusing on biogas produced from a combination of kw, sludge, and agricultural waste and determining the optimum conditions for biogas production. liu et al. (2008) investigated methane production from a mixture of stover and food waste mixed with activated sludge (inoculums) in a batch reactor under mesophilic (35°c) and thermophilic (55°c) conditions. the effect of the ph and solid content on the biogas productivity was investigated. the results showed that biogas productivity under mesophilic conditions yielded a ph of 7 with a higher solid content. similarly, biogas produced from grass (cocksfoot, tall fescue, reed canary, and timothy) was studied by seppl et al. (2009). they found that the methane yield per hectare of the first harvest was always higher than that of the second harvest; however, all the grass types were suitable producers of biogas. bouallagui et al. (2003) studied the production of biogas from fruit and vegetable waste under mesophilic conditions in a semicontinuous tubular digester. the effect of the hydraulic retention time (hrt) and feed concentration on the extent of the degradation of the waste was investigated. the results showed that the biogas produced contained 64% methane. similar results were found by lianhua et al. (2010) and sánchez et al. (2000). production of biogas from wastewater sludge was studied by ferrera et al. (2008) under thermophilic conditions in a semi-continuous reactor. their results showed that biogas production increased by 30% upon pretreatment of the sludge under thermophilic conditions. noutsopoulos et al. (2012) investigated the increase in biogas production by co-digesting lipids in sewage sludge. they found that biogas production increased when the lipid was mixed with the sludge, and the production was proportional to the lipid concentration. the effect of adding fatty acids to sludge on biogas production under thermophilic conditions was studied by lins and illmer (2012). it was found that after certain concentrations, fatty acids initially prevented biogas production; however, production then continued. sludge pretreatment at 70˚c in a microwave showed that biogas production could be improved up to 35% (kuglarz et al. 2013). castrillón et al. (2013) studied methane production from cattle manure (m) supplemented with crude glycerin from the biodiesel industry in a continuously stirred tank reactor (cstr) and induced bed reactor. the results showed that biogas production greatly depends on the organic loading rate (olr). biogas production from whey mixed with poultry m was investigated by gelegenisa et al. (2007a) in a cstr with an olr of 4.9g cod/ld. it was found that biogas production increased by 40%. similar results were found by castrillón et al. (2011) when glycerin was mixed with cattle m. biogas production from cow and llama m was studied by alvareza et al. (2006), showing that the llama m gave better results than cow m since it contains a higher volatile solid (vs) content. biogas production from food waste mixed with animal m was also investigated by marañón et al. (2012); budiyono et al. (2014); and azadeh and jalal (2011). the objective of the current research was to investigate the production of biogas from solid kw mixed with chicken m at different kw-m mass ratios. the effect of the solid content and ph on the amount of biogas produced was also studied. biogas production from sludge and sludge mixed with kw was also investigated. 2. methodology 2.1 setup one-liter squeeze bottles were used as bioreactors. the required amount of water, solid waste prepared a priori, and m were well mixed and placed in the bottles. the bottles’ experimental investigation of biogas production from kitchen waste mixed with chicken manure   117 temperatures were held constant at 35 ± 1˚c by immersing the bottles in a water bath where the temperature was held constant by an external heater and controller. the bottle mouth was connected to a syringe to measure the volume (v) of the biogas produced over time. under the assumption that the gas was ideal, the moles of the biogas produced (n) were calculated using the ideal gas law (1) where t is the temperature, p is the atmospheric pressure, and r is the universal gas constant. 2.2 materials a mixture of vegetables and fruits (200 g of cucumber, tomatoes, orange peel, rice, cabbage, apple, and carrot in equal amounts) represented the kw and was used in this study. the vegetables and fruits were cut into small species (0.5 cm × 0.5 cm x 0.5 cm) and mixed with the required amount of chicken was then added as required, and the mixture was well agitated and introduced to the bottles, which were submersed in the water bath. 2.3 procedure the effect of the solid content was tested by adding the required amount of water to the kw-m mixture, which was termed as solid (s). the following s-water ratios were tested: 1:0 (100% s; no water added), 3:1, 1:1, and 1:3, hence obtaining a s content of 100%, 75%, 50%, and 25%, respectively. the effect of the amount of m to the kw was tested by adding certain amounts of m to the kw so that the ratio of kw:m was 1:0, 3:1, 1:1, and 1:3, hence obtaining percentages of kw in the kw-m mixture of 100% (pure kw), 75%, 50% and 25%, respectively. the produced gas samples were analyzed for methane content using an eagle gas analyzer (rki instruments, union city, california, usa). the ph of the mixture was measured via a ph meter wtw wissenschaftlich-technische werkstätten gmbh, weilheim in oberbayern, germany). the vs contents were measured according to the standard procedures. 3. results and discussion 3.1 effect of solid: water ratio the effect of the s: water ratio on the amount of biogas produced is depicted in fig. 1. when the ratio of kw: m in the solid mixture was 1:1, the vs content in the solid mixture was 78%. figure 1 shows that the production rate was fast on the first day and then slowed down, possibly because the acidogensis bacteria degraded the manure faster compared to the vegetables. the fatty acid concentrations increased in the mixture as a result of this step. accordingly, the ph dropped, reducing the activity of the methanogenic bacteria and, consequently the production of biogas (beam 2011; dearman et al. 2006; raposo et al. 2006; xu et al. 2002; yadvika et al. 2004). figure 1 also shows that an optimum s:water ratio of 1:1 existed at the point where the highest production of biogas was obtained, which is a result similar to that obtained by gelegenis et al. (2007b). 3.2 effect of kw: m ratio the effect of the ratio of kw: m on biogas production when the s: water ratio equaled 1:1 is depicted in fig. 2. as expected, the amount of biogas produced increased as the amount of m increased, or as the ratio of kw: m decreased, since m degrades faster than kw. it should be noted that the vs content of the solid mixture (kw + m) was 78%. 3.3 biogas production from sludge biogas produced from sludge containing 94% water and 97% water is portrayed in fig. 3. the sludge was obtained from a nearby waste water treatment plant and contained 94% water. the sludge was then diluted further with water to obtain a saturation level of 97%. the rate of biogas production was faster for the diluted sludge, and a larger amount of biogas was obtained. this result is in accordance with the results obtained by abdalaal (2012). 3.4 effect of ph on the biogas production as the fermentation process proceeds, fatty acids are produced from organic waste by the acidogenesis of bacteria; accordingly, the ph value of the mixture drops. this causes a drop in the activity of the bacteria, especially the ethanogenic bacteria (dearman et al. 2006; raposao et al. 2006; ryan 2011; yadvika et al. 2004). a measurement of the ph of the mixture as a function of time during the digestion of the kwm concoction is shown in fig. 4. the rate of biogas production was high at the beginning of the process, but after day four it stopped. the ph value, however, dropped from 6.5 to 3.5 and h. mousa, a. obaidat, h.b. khaled, a. alawaneh and a. tarawneh  118 figure 1. amount of biogas produced versus time at 35˚c for various solids, water ratios with kw: m ratio of 1:1. figure 2. effect of kw : m ratio on the amount of biogas produced with solid : water ratio of 1:1. 0 1 2 3 4 5 6 7 0 2 4 6 8 (n /g v s )x 1 07 time (day) solid:water = 1:1 solid:water = 1:3 solid:water = 3:1 0 5 10 15 20 25 30 35 0 2 4 6 8 10 12 14 16 18 20 (n /g v s )x 1 08 time (day) kw:m = 1:3 kw:m = 1:1 kw:m = 3:1 experimental investigation of biogas production from kitchen waste mixed with chicken manure   119 figure 3. the amount of biogas produced versus time from waste water sludge at 35˚c. then stopped, indicating that there was no bacterial activity after four days, which also reflected on the biogas production as no biogas was produced after four days. to study this point further, the same experiment was repeated under the same conditions except that the ph was kept constant at 6 by adding a naoh solution. the results are illustrated in fig. 5, where it can be seen that the process of biogas production continued for 12 days and then stopped. a similar experiment on sludge with a water content of 94% was performed. no effort was made to control the ph. the results are shown in fig. 6, where it can be seen that the rate of biogas production continued until day 12 after which it slowed down. it can be also seen that the ph dropped slightly after the first day and then slowly increased. this result was different from that obtained for kw-m (fig. 4) where the ph continuously decreased. this might be attributed to the high amount of water in the sludge sample. this result was in accordance with the results of abdalaal (2012) and kalloum et al. (2011). however, the current researchers believe that this point needs further investigation. 3.5 effect of temperature the effect of temperature is shown in fig. 7 where, as expected, biogas gas production rates increased as the temperature increased. this result is in accordance with other researches’ results (beam 2011; chua et al. 2013). 4. conclusion biogas produced from the anaerobic digestion of kw mixed with chicken m was investigated. the effect of the ratio of kw mass to that of m in a mixture containing a constant amount of water on the production rate of biogas was studied. the results showed that the lower the ratio of kw:m the higher the quantity of biogas that was produced. investigating the effect of water on biogas production showed that the amount of biogas produced versus the s:w ratio went through a maxima which existed at a s:w ratio of 1:1. controlling the ph at ~6 resulted in the highest yield of biogas. acidic or basic conditions negatively affected the biogas production. sludge with a lower solid content yielded a better biogas production rate. biogas production at 35˚c was greater than at 25˚c. 0 0.2 0.4 0.6 0.8 1 1.2 0 5 10 15 20 25 30 (n /g v s )x 10 7 time (day) water content=97% water content=94% h. mousa, a. obaidat, h.b. khaled, a. alawaneh and a. tarawneh  120 figure 4. measured the ph and the amount of biogas produced for kw : m ratio of 1:1 and solid : water ratio of 1:1 at 35˚c. figure 5. measured amount of biogas produced at constant ph value for a kw : m ratio of 1:1 at 35˚c. 0 1 2 3 4 5 6 7 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 1 2 3 4 5 6 7 8 9 10 11 12 p h (n /g v s )x 1 05 time (day) 0 2 4 6 8 10 12 0 0.5 1 1.5 2 2.5 3 3.5 0 2 4 6 8 10 12 14 16 18 20 p h (n /g v s )x 1 07 time (day) experimental investigation of biogas production from kitchen waste mixed with chicken manure   121 figure 6. amount of biogas produce from sludge sample of water content of 94% at t=35˚c. figure 7. effect of temperature on the amount of gas produced solid : water = 1:1 and kw : m=1:1. 6 6.1 6.2 6.3 6.4 6.5 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 2 4 6 8 10 12 14 16 18 20 22 p h (n /g v s )x 1 07 time (day) 0 5 10 15 20 25 0 5 10 15 20 (m o l/ g v s )x 1 0 8 time (day) t=35oc t=45oc t=35˚c t=45˚c h. mousa, a. obaidat, h.b. khaled, a. alawaneh and a. tarawneh    122 references abdalaal r (2012), potential production of biogas from the sludge of the wastewater treatment plant at jordan university of science and technology, m.sc. thesis. alvareza r, villcaa s, lidénb g (2006), biogas production from llama and cow manure at high altitude. biomass and bioenergy 30: 66–75. azadeh b, jalal s (2011), effect of organic loading rates (olr) on production of methane from anaerobic digestion of vegetables waste. world renewable energy congress, linkoping, sweden, may 8-13. beam r (2011), enhanced biogas production through the optimization of the anaerobic digestion of sewage sludge, m.sc. thesis. the university of alabama. bouallagui h, ben r, cheikh b, marouani l, hamdi m (2003), mesophilic biogas production from fruit and vegetable waste in a tubular digester. bioresource technology 86: 85–89. budiyono n, widiasa n, seno j, sunarso l (2014), increasing biogas production rate from cattle manure using rumen fluid as inoculums. international journal of science and engineering 6: 31-38. castrillón l, fernández-nava y, ormaechea p, marañón e (2013), methane production from cattle manure supplemented with crude glycerin from the biodiesel industry in cstr and ibr. bioresource technology 127: 312–317. castrillón l, fernández-nava y, ormaechea, p, marañón e (2011), optimization of biogas production from cattle manure by pretreatment with ultrasound and co-digestion with crude glycerin. bioresource technology 102: 7845–7849. chua k, cheah w, tan c, leong y (2013), harvesting biogas from wastewater sludge and food waste. 4th international conference on energy and environment (icee 2013). earth and environmental science 16: 012118. dearman p, marschner r, bentham h (2006), methane production and microbial community structure in single-stage batch and sequential batch systems anaerobically codigesting food waste and biosolids. applied microbiol biotechnology 69: 589–596. ferrera i, ponsáb s, vázquezc f, xavier f (2008), increasing biogas production by thermal (70°c) sludge pre-treatment prior to thermophilic anaerobic digestion. biochemical engineering journal 42: 186–192. gelegenis j, georgakakis d, angelidaki i, christopoulou n, goumenaki m (2007a), optimization of biogas production from oliveoil mill wastewater, by codigesting with diluted poultry-manure. applied energy 84: 646–663. gelegenisa j, georgakakisb d, angelidakic i, mavrisa v (2007b), optimization of biogas production by co-digesting whey with diluted poultry manure. renewable energy 32: 2147– 2160. kalloum s, bouabdessalem h, touzi a, iddou a, ouali m (2011), biogas production from the sludge of the municipal wastewater treatment plant of adrar city (southwest of algeria). biomass and bioenergy 35: 2554-25560. kuglarz m, karakashev d, angelidaki i (2013), microwave and thermal pretreatment as methods for increasing the biogas potential of secondary sludge from municipal wastewater treatment plants. bioresource technology 134: 290-297. lianhua l, dong l, yongming s, longlong m, zhenhong y, xiaoying k (2010), effect of temperature and solid concentration on anaerobic digestion of rice straw in south china. international journal of hydrogen energy 35: 7261-7266. lins p, illmer p (2012). effects of volatile fatty acids, ammonium and agitation on thermophilic methane production from biogas plant sludge in lab-scale experiments. folia microbiologica, 57: 313-316. liu c, yuan x, zeng g, li w, li j (2008) prediction of methane yield at optimum ph for anaerobic digestion of organic fraction of municipal solid waste. bioresource technology 99: 882–888. marañón e, castrillón l, quiroga g, fernándeznava y, gómez l, garcía m (2012), codigestion of cattle manure with food waste and sludge to increase biogas management. waste management 32: 1821–1825. noutsopoulos c, mamais d, antoniou k, avramides c (2012), increase of biogas production through co-digestion of lipids and sewage sludge. global nest journal 14: 133-140. raposo f, banks c, siegert i, heaven s, borja r (2006), influence of inoculum to substrate ratio on the biochemical methane potential of maize in batch tests. process biochemistry 41: 1444–1450. ryan g (2011), enhanced biogas production through the optimization of the anaerobic digestion of sewage sludge. phd thesis, the experimental investigation of biogas production from kitchen waste mixed with chicken manure   123 university of alabama, department of chemical and biological engineering. sánchez e, borja r, weiland p, travieso l, martõán a (2000), effect of temperature and ph on the kinetics of methane production, organic nitrogen and phosphorus removal in the batch anaerobic digestion process of cattle manure. bioprocess and biosystems engineering 22: 247-252. seppl m, paavola t, lehtomki a, rintala j (2009), biogas production from boreal herbaceous grasses — specific methane yield and methane yield per hectare. bioresource technology 100(12): 2952-2958. xu h, wang j, tay j (2002), a hybrid anaerobic solid-liquid bioreactor for food waste digestion. biotechnology letters, 24: 757–761. yadvika s, sreekrishnan t sangeeta kvr (2004), enhancement of biogas production from solid substrates using different techniques – a review. bioresource technology 95: 1–10.   the journal of engineering research (tjer), vol. 18, no. 1, (2021) 20-25 corresponding author’s e-mail: 40071@uotechnology.edu.iq doi: 10.53540/tjer.vol18iss1pp20-25 predicting the flow velocity at the toe of a labyrinth stepped spillway jaafar s. maatooq civil engineering department, university of technology, baghdad, iraq abstract: the velocity at the toe of a spillway is a major variable when designing a stilling basin. reducing this velocity leads to reduce the size of the basin as well as the required appurtenances needed for dissipating the surplus kinetic energy of the flow. if the spillway chute is able to dissipate more kinetic energy, then the resulting flow velocity at the toe of the spillway will be reduced. typically, a stepped spillway is able to dissipate more kinetic energy than that of a smooth surface. in the present study, the typical uniform shape of the steps has been modified to a labyrinth shape. it is logical to expect that the labyrinth shape will lead to dissipate more kinetic energy. this impression comes through creating more regions of circulation and turbulence along the lateral sides of each step in addition to those that occur towards the streamwise. this action can also reduce the jet velocities near the surfaces, thus minimizing cavitation. at the same time, the increase in circulation regions will maximize the opportunity for air entrainment, which also helps to dissipate more kinetic energy. the undertaken physical models consisted of three labyrinths of stepped spillways with magnification ratios (width of the labyrinth to width of conventional step) wl/w are 1.1, 1.2, and 1.3 as well as testing a conventional stepped spillway (wl/w=1). two empirical forms of the coefficient are proposed, one for labyrinth shape stepped spillway denoted kl and another for conventional stepped spillway denoted ks. once the value of the coefficient is known, the actual flow velocity at the toe of a stepped spillway can easily be computed without having to resort to measurements on-site. it is concluded that the spillway chute coefficient is directly proportional to the labyrinth ratio and its value decreases as this ratio increases. keywords: stepped spillway; labyrinth spillway; energy dissipation; labyrinth shape; magnification ratio; chute coefficient. حساب سرعة الجریان عند قدمة المطفح المدرج على شكل متاھة جعفر صادق معتوق تعتبر السرعة عند قدمة المطفح من المتغیرات الرئیسیة عند تصمیم أحواض التسكین. وبالتالي التقلیل من تلك :الملخص السرعة یعني التقلیل من الحجم المطلوب لحوض التسكین وكذلك من الملحقات الخاصة بتبدید طاقة الجریان الحركیة. أكبر من طاقة الجریان الحركیة حینھا ستكون سرعة الجریان الواصلة فعندما یكون سطح المطفح قادر على تبدید مقدار عموما المطافح المدرجة لھا القابلیة على تبدید للطاقة الحركیة أكثر من تلك ذات األسطح الملساء. في عند القدمة واطئة. أن یكون شكل المتاھة لھ لنموذجي الى مدرج بشكل متاھة. من المنطقي التوقعھذه الدراسة تم تطویر السطح المدرج ا القابلیة على تبدید المزید من الطاقة الحركیة للجریان من خالل زیادة التداخل بین تفرعات الجریان التي سوف تؤدي الى ھذا الوقت المزید من الدوامات. ھذا التأثیر سوف یقلل من بثق الجریان بالقرب من السطح وبالتالي التقلیل من النخر. في زید من فرصة تداخل الھواء مع الجریان الذي بدوره یساعد على تبدید المزید من الطاقة تطق الدوامات سوف زیادة منا الحركیة. النموذج الفیزیائي المعتمد في ھذذه الدراسة قد تضمن ثالثة نماذج لشكل المتاھة مع نسب تكبیر لعرض المطفح )w/lw( 1.1 لمدرج التقلیديفح اطباألضافة الى نموذج الم 1.3و 1.2و ) ذو النسبةw/lw(. تم أقتراح معاملین التقلیدي. في حال معرفة قیمة المعامل بالنموذج أحدھما خاص بنموذج شكل المتاھة واآلخر للمطفح المدرج تجریبیین من خالل التجریبي یمكن عندھا حساب سرعة الجریان عند قدمة المطفح بسھولة دون الحاجة ألجراء قیاسات موقعیة النتائج یمكن القول بأن معامل المطفح قد تأثر بشكل مباشر مع نسبة التكبیر وبالتالي قیمتھ تقل مع الزیادة في تلك النسبة. نسبة التكبیر. ؛معامل المنزلق ؛شكل المتاھة ؛تبدید الطاقة ؛المطفح المدرج :الكلمات المفتاحیة predicting the flow velocity at the toe of a labyrinth stepped spillway 21 1. introduction the behaviour of flow down conventional uniform stepped spillways has been investigated both experimentally and numerically by numerous researchers. some studies have been conducted into modifications of the conventional configuration of a stepped spillway to improve its hydraulic performance by increasing the kinetic energy dissipation and/ or enhancing the resistance to cavitation. such modifications included; inclined upward steps, installing end sills, adding baffle blocks at the downstream end of each step, turbulence manipulators and macro-roughness systems consisting of concrete blocks (manso and schleiss, 2002; chanson and gonzalez, 2004; el-jumaily and al-lami, 2009). with respect to the values of the flow velocity at the toe of the spillway, bradley, and paterka (1957) and more recently zahra et al. (2015) have applied a correction coefficient, cf to the theoretical velocity to give the actual velocity. in these studies, the hydraulic effects and the roughness along the chute of the spillway are implicitly reflected in this coefficient. abdul-mehdi et al. (2016) from the laboratory study which has been conducted on steps roughed by three different sizes of gravel. the authors concluded that the energy dissipation increases with the increasing the size of the gravel and the maximum per cent of this increase was 16.73% as compared with the traditional steps without roughness by gravel for the same geometric and hydraulic conditions. moghaddam et al. (2017) adopted the reverse gradient with sill installed at the edge of each step of the stepped spillway as an attempt for development. the experimental program has been conducted to test the advantage of this development. the change of; reverse gradient, sill height and sill thickness are the aims to increase the ability of the stepped surface of the spillway to dissipate more kinetic energy. the results show appreciable advantages of such attempts especially with higher reverse gradient and less thickness of sill. from all the aforementioned researches, the key advantage of stepped spillways is increasing the performance in energy dissipation. this action will definitely reduce the flow velocity at the spillway toe, which is positively reflected in the choice of the size and type of the stilling basin. this has led to an investigation of the best configurations and features to provide standardized hydraulic design criteria. recently the features of labyrinth weir have been adopted by maatooq and ojaimi (2014) and maatooq (2016) to modify the step shape in a stepped spillway. the first aim of this modification was to utilize the labyrinth shape to increase the effective width of each step (wl) in comparison with the standard step width (w), where wl= 3ω (4a+2lc), as evident in fig. 1(a). the labyrinth shape has the potential to reduce the jet velocities near the flow surface and to reduce the potential of cavitation. at the same time, increasing circulation regions will indeed increase the opportunity of air entrainment, which has the incentive to dissipate more kinetic energy. the second aim was to reduce the volume of construction materials need for each step compared to the traditional step configuration thereby reducing construction cost and time frame. accordingly, the aim of this research was to investigate a stepped spillway configuration that can reduce the incoming froude number by reducing the flow velocity at the spillway toe. the incoming froude number is an indicator used to classify the type of hydraulic jump often created at the toe of a spillway, and the type of the stilling basin was identified accordingly. physical model testing was undertaken of three labyrinths stepped spillways with the labyrinth magnification ratios of wl/w= 1.1, 1.2, and 1.3 as well as testing the conventional stepped spillway (wl/w = 1). this study has focused on investigating the impact of the wl/w ratio on the actual flow velocity at the toe of the spillway. 2. models and procedure physical modelling was undertaken in a flume located at the hydraulic laboratory of the university of technology-baghdad. this flume has a working section 0.3m wide, 0.45m in depth and its length is 12.5m. three physical models of the stepped spillway at chute angles, θ=35o, 45o, and 55o were adopted with a crest fixed at height 0.32m (i.e., the height of dam), these inclination angles gave ratios of step height (h) to step tread length (s), h/s= 0.7, 1, and 1.428, respectively. with each inclination, four-step configurations were tested. the first was a traditional stepped spillway configuration. the other three models were labyrinth step shapes at n= three cycles each of width equal, ω to give wl/w=1.1, 1.2, and 1.3, where the lateral length of labyrinth step wl= n (4a+2lc), and w is representing a lateral length of traditional step, see fig. 1. the twelve models were divided into three groups based on the spillway angle. in all models the step rise, h was kept at 0.04m, while the tread lengths were, s= 0.057m, 0.04m, and 0.028m. the dimensions as illustrated in fig.1 were selected according to the available hydraulic conditions and flume dimensions. the flow rate was measured using an electromagnetic flowmeter, which can record flows between 0 to 30 l/s for accuracy within ±0.01 l/s. based on the limitation ten different discharges ranged between 2 l/s and 16 l/s were selected to achieve flow conditions between nappe and skimming flow regimes. for each run after the stabilizing of the flow, the depths of flow over the crest and at the toe of the spillway were recorded using a digital point gauge with accuracy ±0.1mm. the pitot tube mounted with a water manometer installed on the centre line of the flume was used to measure the velocity at the toe of the spillway for each predicting the flow velocity at the toe of a labyrinth stepped spillway 22 run. this reading is the actual velocity denoted as va. after calibrating the pitot tube and applying the calibration coefficient, velocity was calculated by; va = c�2gδh (1) where δh, is the water level difference of the manometer and c=0.9 in this study as determined from calibration. this actual velocity is compared with the theoretical velocity which calculated from the relationship has been published by bradley and paterka (1957), in the form of a chart, its ordinate based on the fall from the reservoir level to the stilling basin floor and its abscissa is the ratio of the actual velocity to the theoretical velocity at the entrance of the stilling basin (toe of spillway). accordingly, bradley and paterka (1957) proposed the following formula for calculating the theoretical velocity; vt = �2g �z − h 2 � (2) where, z is the vertical distance between the elevation of the head, h, over the crest of the spillway and the floor level at the toe of the spillway (see fig. 2). the aim of the present study is to assess the functional relationship for the labyrinth coefficient (kl) which represents the ratio of the actual velocity, va to the theoretical velocity, vt; kl= va/vt (3) this coefficient depends on the geometrical properties of the spillway surface and the working head over the crest. the lower the value of this coefficient the greater the dissipation of kinetic energy along the surface of the spillway. 3. results and discussion under conditions of skimming flows down a stepped spillway the flow rotation in the triangular area at a front of the step and on the back of the tread can have a significant impact on energy dissipation. in a conventional stepped spillway, this zone of flow rotation has the same influencing action transversely (i.e., across the width of the spillway). a labyrinth step configuration disrupts this transverse uniformity of flow. the labyrinth step shape can increase the interlocking between a streamwise recirculation as well as a secondary circulation created locally in the labyrinth cycle which definitely increases the ability to dissipate more kinetic energy. the flow conditions and the hydraulic performance of a stepped spillway are governed by the step height, step length, and discharge (chanson, 2002; bose and hager, 2003; ohtsu et al., 2004; khatsuria, 2005). the new concept introduced here is the labyrinth ratio, wl/w. in fig. 3 the variation of kl as the dependent variable under the influence of, h/l is illustrated, where l is the length of the chute (measured along the chute surface) as illustrated in fig. 1(b). as expected kl increases i.e., the velocity at the toe of the spillway becomes closer to the theoretical velocity, as the h/l ratio increases, wherewith increasing the head (h) over the crest at the same length of the chute (l) the actual velocity (va) at the toe definitely increased. (a) figure 1. physical model of labyrinth stepped spillway; (a) plane view and (b) 3-d view (maatooq and ojaimi 2014). figure 2. parameters provided for eqn. (1) (bradley and paterka, 1957). jaafar s. maatooq 23 this physical point is hydraulically known for such structure through increasing the value of incoming froude number for the hydraulic jump when it occurs just at the toe of a spillway as evident through a sketch in fig. 1. figure 3 illustrates the hydraulic impact of the labyrinth steps where the value of kl decreased with increasing the labyrinth ratio wl/w for any given h/l. this effect seems greater at higher h/l. this finding is also demonstrated in fig. 4, where the labyrinth coefficient dramatically decreases as the labyrinth ratio increase for all spillway slope angles. in terms of the effect of inclination or slope angle of the spillway on the labyrinth factor, the h/s=1.428 (θ=55o) has a greater influence. the values of kl however, nearly coincide for labyrinth ratio wl/w=1.3 when the spillway inclination is between θ=35o and 45o as shown in fig. 5. in general, fig. 5 indicates that for any given value of the labyrinth ratio, wl/w the h/s does not have an effective influence on the kl, but at the same time the noticeable influence was inversely proportional instead of the case with the traditional steps. the statistical software (statistica-10) was used to derive the following functional relationship based on experimental data by multiple nonlinear regression analysis with r2=0.921. kl = 1.143 � h l � 0.329 �wl w � −0.875 �h s � −0.228 (4) equation 4 can be applied with high reliability when the rise of step to the tread (h/s) is ranged between 0.7 and 1.4, the labyrinth does not exceed 3-cycle, the labyrinth magnification ratio (wl/w) up to 1.3, and the lower head over the crest (h) and its maximum value are 5% to 25% of the chute length (l) respectively. figure (6) demonstrates the relationship between the predicted values of the labyrinth coefficient based on eqn. (4) and the observed values of this coefficient that resulted through the application of the experimental data at eqns.1 and 2. it should note here that eqn. (4) is solely applicable for the labyrinth shape (i.e., wl/w> 1). for conventional stepped spillway, wl is equal to w. thus the ratio wl/w does not have an effect on kl. since the spillway chute here is consisting of traditional steps, the coefficient should take another characterization and denote as ks. in this situation the effect of two parameters h/l and h/s has been used in regression analysis to get the following relationship with r2=0.938; ks = 1.491 � h l � 0.449 �h s � −0.174 (5) figure 3. effect of head over crest related to chute length on labyrinth coefficient (kl). figure 4. effect of labyrinth ratio on kl. figure 5. effect of inclination of the chute on kl. figure 6. the predicted versus observed value of labyrinth coefficient. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.05 0.1 0.15 0.2 0.25 0.3 k l h/l wl/w=1 wl/w=1.1 wl/w=1.2 wl/w=1.3 0.4 0.5 0.6 0.7 0.9 1 1.1 1.2 1.3 1.4 k l wl/w for h/s=0.7 for h/s=1 for h/s= 1.428 0.4 0.5 0.6 0.7 0.6 0.8 1 1.2 1.4 1.6 1.8 2 k l h/s for wl/w=1 for wl/w=1.1 for wl/w=1.2 for wl/w=1.3 0.3 0.4 0.5 0.6 0.7 0.3 0.4 0.5 0.6 0.7 pr ed ic te d k l observed kl predicting the flow velocity at the toe of a labyrinth stepped spillway 24 statistical indices including the root mean square error, rmse, and the mean bias error, mbe, were also calculated to test the reliability of eqns.4 and 5. these indicators present error in the units (or square units) of the constituent of interest which aid in the analysis of results. it was first presented by moriasi, et al. (2007), then cited and adopted by tariq and latif (2011), and maatooq (2016). zero values of these indicators refer to a perfect fit. another statistical indicator that determines the relative magnitude of the residual variance compared to the variance of the measured data is the nash-sutcliffe efficiency coefficient (nsec) which recommended by the asce, 1993. when nsec has a value between 0 and 1 is viewed as an acceptable performance level, whereas when it being less than zero is an indication of unacceptable performance (less reliability). the statistical measures values of indices for eqns. 4 and 5 respectively are; rmse=0.0196 and 0.0252, mbe=0.000074 and 0.000031. the small values of these indices refer to acceptable correlations. whereas the nsec=0.921 and 0.938 respectively refer to acceptable reliance. based on these values of indices it is concluded that eqns.4 and 5 could be used as predictive equations to calculate the labyrinth coefficient and stepped spillway chute coefficient (for conventional stepped spillways). once the value of this coefficient is known the actual flow velocity at the toe of a stepped spillway can easily computing by using eqn. (2). it should worth noting that the difference between eqn. (4) and eqn. (5) is the former used just when wl/w> 1, i.e. for labyrinth stepped spillway whereas the latter has been derived based on the traditional stepped spillway data. 4. conclusion the velocity at the toe of a spillway is a major variable when designing a stilling basin. reducing this velocity reduces the required size of a stilling basin with dissipater appurtenances which usually need to dissipate the surplus energy of the flow. if the spillway chute is able to dissipate more kinetic energy then the resulting flow velocity at the toe of the spillway will be reduced. typically stepped spillway is able to dissipate more kinetic energy than a smooth chute. in the present study, the typical uniform shape of the steps in a spillway has been modified and a labyrinth shape has been adopted. physical models of three labyrinth stepped spillways with magnification ratios of wl/w= 1.1, 1.2, and 1.3 has been used for experimental work as well as a conventional stepped spillway at which wl/w= 1. the experimental results show the significant and direct effect of the labyrinth configuration on decreasing the values of flow velocity at the toe of the spillway. the aim of the experimental work is to establish a simple empirical formula to facilitate the calculation of the flow velocity at the toe of the spillway. consequently, two empirical forms of the coefficient are proposed, one for labyrinth shape stepped spillway denoted kl and another for conventional stepped spillway denoted ks. once the value of the coefficient is known the actual flow velocity at the toe of a stepped spillway can easily computing by using eqn. (2). it should be noted that the labyrinth coefficient calculated by eqn. (4) has practical limitations such as the labyrinth cycle and the magnification ratio do not exceed 3 and 1.3 respectively along the head over the crest do not exceed the 0.25l. these boundary conditions are established to further researches to expand the scope of applicability. conflict of interest the authors declare no conflicts of interest. acknowledgment the author would like to thank the engineering and technician staff of the hydraulic laboratory in the civil engineering dept. at the university of technology for the help in preparing and installation of physical models. the author acknowledged that some of the readings and recording of the experimental data were conducted by the msc. student mr ojaimi. references asce (1993), criteria for evaluation of watershed models. journal of irrigation and drainage engineering, asce 119(3): 429–442. abdul-mehdi tr, al-mussawy ha, al-madhhachi ast (2016), a laboratory study attempt of flow and energy dissipation in stepped spillways. journal of engineering 22(12):48-64. boes rm, hager wh (2003b), two-phase flow characteristics of stepped spillways. journal of hydraulic engineering, asce, 129(9):661-670. bradley jn, paterka aj (1957), the hydraulic design of stilling basins. proceeding, asce, 83: 29-31. chanson h (2002), the hydraulics of stepped chutes and spillways. balkema publishers, steenwijk, the netherlands. el-jumaily kk, al-lami mk (2009), study of conveniency of using stepped spillway in roller compacted concrete dams (rccd). engineering and technology journal 27(16), http://www.uotechnology.edu.iq/tec_magaz/issues 2009.htm. khatsuria rm (2005), hydraulic of spillway and energy dissipators. marcel dekker, new york, usa, isbn: 0-8247-5789-0. maatooq js, ojaimi tj (2014), evaluation the hydraulic aspects of stepped labyrinth spillway. engineering and technology journal 32(9) (ppart-a): 2174-2185. maatooq js (2016), kinetic energy dissipation on labyrinth configuration stepped spillway. tikrit http://www.uotechnology.edu.iq/tec_magaz/issues2009.htm http://www.uotechnology.edu.iq/tec_magaz/issues2009.htm jaafar s. maatooq 25 journal of engineering sciences 23(3): 12-24. manso pa, schleiss aj (2002), stability of concrete macro-roughness linings for overflow protection of earth embankment dams. canadian journal of civil enggineers 29(5): 762-776. moghddam ak, ketabdar m, ahmadian sa, hoseini p, pishdadakhgari m (2017), experimental survey of energy dissipation in nappe flow regime in stepped spillway equipped with inclined steps and sill. international journal of research and engineering 4(5): 161-165. moriasi dn, arnold jg, van liew mw, bingner rl, harmel rd, veith tl (2007), model evaluation guidelines for systematic quantification of accuracy in watershed simulations. transaction of asabe 50(3): 885–900. ohtsu i, yasuda y, takahashi m (2004), flow characteristics of skimming flows in stepped channels. journal of hydraulic engineering, asce, 130(9): 860-869. tariq ja, latif m (2011), flexibility analysis of irrigation outlet structures using simulation of irrigation canal hydrodynamic model. irrigation science 29: 127-134, doi 10.1007/s00271-010-0222-8. zahra a, afzalimehr h, singh vp, fattahi r (2015), prediction of flow velocity near inclined surfaces with varying roughness. international journal of hydraulic engineering 4(1): 1-9. 1. introduction 2. models and procedure 3. results and discussion 4. conclusion conflict of interest acknowledgment references microsoft word paper 6.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 160-171 performance of cement-based patch repair materials in plain and reinforced concrete members a.h. al-saidy* department of civil & architectural engineering, college of engineering, sultan qaboos university, oman received 2 march 2015; accepted 12 april 2016 abstract: structural elements such as beams, slabs, and columns may require strengthening or repair during their service life. different repair materials (rms) are available and it is usually difficult to choose the best ones, especially when considering the cost of such materials. this paper presents the results of an experimental investigation of patch rms on plain concrete prisms as well as on reinforced concrete beams. three cement-based rms available in the market with different mechanical properties and an ordinary portland cement (opc) mix produced in the lab were used in the study. damage was induced in prisms/beams and then repaired using different materials. the experimental work included assessment of the flexural strength of damaged/repaired plain concrete prisms; slant shear (bond) strength between the concrete and the rm; axial strength of damaged/repaired plain concrete prisms and bond of the repair materials in damaged/repaired reinforced concrete beams loaded to failure. the test results showed that all rms performed well in restoring the strength of damaged plain concrete. compatibility of the rms with substrate concrete was found to be more important in the behavior than superior mechanical properties of the rms. no difference was noted in the behavior between the rms in repairing reinforced concrete beams at the tension side. keywords: patch repair, repair materials, rehabilitation, retrofitting. אאאאאאאאאא אא* א wאאאאא  k  אאא אא  אא ،א  א  אאkאא אאאאא אאאאאאkאאאאא אאאאאאאkא אאאאkאאא  א،אאאא،אאא  kא אk    אאאאאאאא אא kא אאאאאאאאאאא אאאאאאאאאk אאאאאאאאא אאk אאwא،אאא،א،אאא * corresponding author’s email: alsaidy@squ.edu.om a.h. al-saidy 161 1. introduction repair and rehabilitation of reinforced concrete structures are common requirements in modern construction. structural elements such as beams, slabs, and columns may require strengthening or repair during their service life due to aging, increased loads, and deterioration caused by the surrounding harsh environment. steel corrosion is a major cause of deterioration which disrupts the cover zone of reinforced concrete. the conventional approach to repair a corroded reinforced concrete member is the use of patches to repair in the zones where cracking and spalling take place. repairs are generally undertaken by removing the damaged concrete beyond the steel reinforcement, and reinforcements are then cleaned in cases of minor corrosion or replaced if severely damaged. finally, a prepackaged cementitious repair material (rm) is applied. many types of rms are available commercially, and the selection for reinstating a deteriorated concrete structure is usually done on a relatively ad hoc basis. the repair and rehabilitation of reinforced concrete structures can only be successful if the new materials interact effectively with the parent concrete and form a durable barrier against the ingress of carbon dioxide and chlorides. problems may arise whenever a dimensionally unstable rm is placed against an aged (and, therefore, dimensionally stable) substrate concrete. several materials are available in the market for the repair of deteriorated concrete structures. such rms are classified, for example, as cement, epoxy resins, polymer latex, polyester resins, and polyvinyl acetate, with cement-based and epoxy-based rms being the two most widely used. however, manufacturers are reluctant to provide complete details of their materials, which makes the selection of appropriate rms difficult. potential users are faced with a wide choice of materials and little guidance on their properties and long-term performance (cabrera and al-hasan 1997). the design of patch repair is usually based on the experience of specialist contractors, and selection of rms is made based on their relative short-term properties such as strength, bond, and early age plastic shrinkage. although these properties indicate the immediate performance of the repair, they give little information on its long-term performance with respect to cracking and composite action with the substrate concrete to carry loads and prevent deformations. numerous studies in the literature deal with the evaluation of different rms. a great deal of the research has focused on durability issues associated with rms such as chloride permeability, electrical resistivity, and carbonation (al-zahrani et al. 2003; mangat and limbachiya 1995; ueda et al. 2011). such studies have concluded that there is no clear difference between the cementand polymerbased repair mortars with regard to the chloride permeability. however, carbonation has been identified as an issue in cementand polymerbased repair mortars except in dense mortars which has been attributed to the addition of silica fume, fibers, and/or other additives. other studies have focused on the properties that affect the stability of rms such as shrinkage, thermal properties, and modulus of elasticity (mangat and limbachiya 1997; margan 1996). polymer-based repair mortars have been found to have lower shrinkage strains compared to cement-based repair mortar. compatibility of the modulus of the elasticity of rm with that of the substrate concrete were found to have a great influence on load sharing or the contribution of rms in resisting the applied load of the repaired member. rms with a modulus of elasticity close to that of the substrate concrete tends to provide a stable repair under applied loads. the efficiency and durability of a repaired system depends on the bond between the concrete substrate and rm. by increasing the surface roughness, the surface treatment of concrete substrate can promote mechanical interlocking, which is one of the basic mechanisms of adhesion. the durability of a bond depends on several factors, each acting with different degrees of influence. the five major factors as reported by researchers in the literature are micro cracking, the absence of a laitance layer, cleanliness before an overlay placement, and compaction and curing procedures. the first three parameters are directly related to substrate characteristics, which can be modified by surface treatment. the treatment of a concrete substrate is commonly used for cleaning, removing the laitance layer, and roughening the surface. however, it can induce micro cracking if it is not well operated with regard to the quality and strength of concrete (courard et al. 2006; courard et al. 2014; garbacz et al. 2005; liu et al. 2014; qian et al. 2014). the performance of the rms in service under structural loading has received relatively little performance of cement-based patch repair materials in plain and reinforced concrete members 162 attention from the research community, especially in the case of patch repairs. some studies have considered the behavior of repaired reinforced concrete members under service loads (abu-tair 1997; alfarabi et al. 2006; ghassan and zai-ul-hasan 1999; rio et al. 2005). however, there is no agreement between researches regarding the properties, which should be considered in the evaluation. this lack of agreement can have serious effects where differences in mechanical properties can lead to shortor long-term performance or durability problems. this paper presents the results of three commercially available rms on the omani market and one laboratory-prepared material. the experimental work presented here includes bond strength; flexural and axial strength of damaged/repaired prisms; bond and strength behavior of damaged/repaired reinforced concrete beams. 2. repair materials four rms were used in this study. a brief description of each of these materials follows. 2.1 repair materials (rm1, rm2, and rm3) the three rms were pre-packed commercial cement-based materials supplied as ready-touse dry powders which require only the addition of clean water to produce a highly consistent, lightweight repair mortar suitable for general purpose concrete and masonry repairs. each material is packaged in 20 kg bags. to get the required mix, water at a measurement of 15–18% of the powder weight was added according to the manufacturer's instructions. however, the difference between the three materials is mainly in the mechanical properties (table 1). table 1. properties of the repair materials after 28 days. repair material fc (mpa) ft (mpa) e (gpa) fr (mpa) rm1 54.3 4.4 31.7 7.8 rm2 52.5 4.3 31.1 8.1 rm3 63.7 4.2 34.4 7.8 opc 51.2 4.2 30.0 7.7 fc = cube compressive strength ft = cylinder tensile strength e = modulus of elasticity fr = prism flexural strength 2.2 ordinary portland cement (opc) this rm was obtained by mixing opc with water at a ratio of w/c of 0.25. opc was chosen in order to have a consistency similar to that of the other rms. 2.3 mechanical properties a) compressive and tensile strength according to american society for testing and materials (astm) standard c109. b) flexural strength according to astm standard c78. c) elastic modulus according to astm standard c469 . 3. specimen details to evaluate the performance of each rm under service load, four different tests were carried out. the bond of the rm to the substrate concrete was evaluated through the slant shear test. behavior of the rm applied to plain concrete was evaluated using damaged/repaired prisms tested in flexure and axial compression. to simulate the behavior of a real field situation, reinforced concrete beams were damaged and then patch-repaired using each of the rms. the reinforced concrete beams were tested to failure to observe the bond of the rm through all loading stages. 3.1 slant shear specimens the bond strength test (slant shear test) was done according to astm standard c882 using 75 x 150 mm concrete cylinders. each cylinder was cut as specified in the standard at 30° by a concrete cutting machine after curing for 28 days (fig. 1a). after cutting these cylinders, each piece was replaced in a cylindrical mold to cast the rm above it. three specimens were used for each rm. figure 1b shows the combined specimen after casting the rm. 3.2 damaged/repaired prisms to test the flexural strength, six 100 x 100 x 500 mm prisms were made from each mix or rm. two prisms were undamaged, two were damaged, and the other two were damaged but then repaired with the rms (damaged /repaired). a similar number of prisms were also made to test the axial strength: two undamaged, two damaged, and two damaged/repaired. in order to produce the same damage in all specimens, a preformed mold was used while casting the prisms to a.h. al-saidy 163 (a) (b) figure 1. slant shear test: (a) concrete cylinder after cutting, and (b) combined cylinder of repair material and concrete substrate. a) casting of white cement in damaged area. b) removing the white cement patch. c) placing the white cement patch in the prism molds before casting. figure 2. preparation of damage mold to induce uniform damage in prisms. performance of cement-based patch repair materials in plain and reinforced concrete members 164 create a damaged section measuring 200 mm in length with a maximum depth of 25 mm within the middle third of the prism (fig. 2). to accomplish this specific type of damage, a concrete prism was damaged by chipping the concrete with 200 mm width and a maximum depth of 25 mm. to simulate the actual damage in concrete, the damage made by chipping the concrete was not uniform. after chipping, the damaged area was patched with white cement mortar. once the white cement patch had set, it was removed from the concrete with all of the rough surfaces imprinted on the surface of the patch. the patch was then placed in the prism mold before casting (fig. 2(c)). after casting, this patch was removed from the prism, creating the required damage. the same procedure was repeated with the other specimens. in this way, the damage in all specimens was uniform in order to make a fair comparison between all rms. 3.3 damaged/repaired reinforced concrete beams in order to study the behavior of reinforced concrete elements, six reinforced concrete beams were cast in the laboratory (table 2). the beams consisted of one undamaged beam, one damaged beam, and four damaged/repaired beams. the beams were reinforced for tension at their bottom sides using two bars, each 10 mm in diameter. each beam was 2,700 mm long with a cross-section of 100 x 150 mm and a cover of 25 mm. for the damaged and repaired beams, the damage was induced during casting with approximately the same shape and size by the same procedure described in section 3.2 for prism preparation. the damage zone was 400 mm long and 40 mm deep (25 mm cover + 10 mm bar diameter + 5 mm above the table 2. description of beams. beam description c control beam without damage d damaged beam drm1 damaged and batch repaired using rm1 drm2 damaged and batch repaired using rm2 drm3 damaged and batch repaired using rm3 drm4 damaged and batch repaired using opc reinforcement) at the middle of the beam. after casting, each beam was cured for 28 days using polythene sheets. four of the damaged beams were repaired using the selected rms and were cured for an additional seven days prior to testing. 4. test setup all specimens were tested after curing using a wet hessian cloth for 35 days (28 days for concrete substrate + seven days for rm). a bond strength test (slant shear test) was done according to the astm standard c882. each cylinder of the combined material (concrete + rm) was placed under the compression machine, and the failure load was recorded. the flexural strength of the undamaged, damaged, and damaged/repaired plain concrete prisms was measured according to astm standard c78. each prism was loaded under four-point bending; the failure load was recorded and the location of the failure was observed. in addition to flexural strength, a compressive strength test was done on the plain concrete prisms. six prisms were tested for each rm: two undamaged, two damaged, and two damaged/repaired prisms. the failure load was recorded for each prism, and the compressive strength was calculated. figure 3 shows the test setup of plain concrete prisms under axial compression. the reinforced concrete beams were simply supported and subjected to four-point bending (fig. 4). the beams were supported over 2,400 mm. linear variable differential transducers (lvdts) were used to measure deflection at the mid-span of the beam. all beams were loaded at a rate of 0.06 kn/s until the failure load was reached. figure 3. test setup of plain concrete prisms under axial compression. a.h. al-saidy 165 figure 4. test setup of reinforced concrete beams. 5. test results and discussion 5.1 bond test the specimens were loaded axially in compression, which caused failure along the line where the two layers are jointed (fig. 1). bond strength was calculated by dividing the failure load by the interface area (ellipse area). the average bond strength (using three specimens) in n/mm2 were 3.81, 4.97, 5.78, and 4.05 for rm1, rm2, rm3, and opc, respectively. as can be observed, the maximum bond strength was 5.78 n/mm2 for rm3, whereas the minimum bond strength value was 3.81 n/mm2 for rm1. there was not much difference between rm1 (3.81 n/mm2) and the control rm opc (4.05 n/mm2). the mode of failure was almost the same for rm1, rm2, and rm3, and failure occurred at the interface surface between the two materials (fig. 5(a)). failure of the opc specimen was due to crushing failure (fig. 5(b)). 5.2 flexural strength of damaged/repaired plain concrete prisms the mechanical properties of the concrete substrate which received the rms are shown in table 3. these are the properties of the concrete used in the prisms that later were repaired using each of the four rms. it should be noted that the flexural strength was assessed for an undamaged concrete prism. each result was an average of three specimens. a summary of the results of the flexural strength of the undamaged, damaged, damaged/repaired prisms is shown in fig. 6. here the failure load which caused the prism to table 3. properties of concrete substrate of prisms after 28 days. prism fc (mpa) ft (mpa) e (gpa) fr (mpa) used with rm1 36.2 2.6 25.3 6.9 used with rm2 35.4 3.1 24.5 7.0 used with rm3 34.1 2.5 24.1 6.2 used with opc 37.4 2.6 26.0 6.8 fc, ft, e, fr : as defined in table 1. p 150 950 500 950 150 spreader beam 100 150 25 2-ø 10 mm 400 performance of cement-based patch repair materials in plain and reinforced concrete members 166 a) bond failure of rm1, rm2 and rm3 b) crushing failure of opc figure 5. modes of failure of the specimens observed during slant shear bond test. figure 6. flexural strength of prisms. break was reported instead of the flexural stress or modulus of rupture, since the latter might be misleading in the case of damaged prisms where the cross-sectional area was reduced as a result of the damage. it can be seen that the damage caused large reduction in the flexural strength by 45% of the undamaged prism. however, all repaired prisms failed at a higher load than that of the undamaged prisms. an increase in the failure load of 30%, 14%, 36%, and 5% were observed in prism repaired using rm1, rm2, rm3, and opc, respectively. this clearly shows that the rms are very effective when loaded under tension. in all prisms, no debonding failure was observed. the location of the cracks and the failure lines of the prisms are shown in fig. 7. rm1 and rm3 were so strong that the concrete cracked just at the end of the patch where the substrate concrete was weaker in tension than the rm (fig. 8). 5.3 compressive strength of damaged/ repaired plain concrete prism this group of prisms had the same damage as the ones that were tested under flexure and 13.91 14.05 12.43 13.31 8.2 7.42 7.32 8.02 18.18 16.12 16.86 13.85 0 2 4 6 8 10 12 14 16 18 20 rm1 rm2 rm3 opc f le x u ra l f ai lu re l o ad ( k n ) repair material undamaged damaged repaired repaired material a.h. al-saidy 167 a) failure mode of rm2, and opc b) failure mode of rm1 and rm3 figure 7. schematic of failure modes in damaged/repaired prisms. (a) rm1 (b) rm2 (c) rm3 (d) opc figure 8. modes of failure observed during flexural test of prisms for all repaired materials. repaired similarly; however, they were tested under axial compression. a summary of the results of the compressive strength of the undamaged, damaged, damaged/repaired prisms is shown in fig. 9. as expected, all damaged prisms failed at a lower load than the undamaged and repaired prisms due to the reduced cross-sectional area caused by the damage. however, all damaged/repaired prisms showed equal or higher failure loads compared to the undamaged prisms, indicating that the rm contributed to the axial load carrying capacity of the damaged/repaired prisms. performance of cement-based patch repair materials in plain and reinforced concrete members 168 figure 9. compressive strength of prisms. (a) rm1 (b) rm2 (c) rm3 (d) opc figure 10. modes of failure observed during compressive test for all repaired materials. in the compressive strength test, the specimens repaired with opc failed due to compression without detachment of the repair material from their locations, while, the other specimens repaired with rm1, rm2 and rm3 failed due to shear because the rms detached from their locations as indicated by the inclined failure line in fig. 10. the opc rm showed better bond behavior under compression which was also observed in the slant bond test discussed in section 5.1. 5.4 repaired reinforced concrete beams unlike plain concrete, which is very brittle, reinforced concrete behaves in a ductile manner. therefore, it was possible to observe the behavior of the combination of repair + concrete in all loading stages as encountered in the field. load versus mid-span deflections are compared in fig. 11, showing the undamaged, damaged, and damaged/repaired beams. from fig. 11, it can be observed that all of the repaired beams resisted an ultimate load slightly higher than the undamaged beam. all beams failed in a 199.5 297.5 191 276.6 123.5 196.8 140.1 189.5 203.2 285.7 201.2 266.6 0 50 100 150 200 250 300 350 rm1 rm2 rm3 opc c o m p re ss iv e f a il u re l o a d ( k n ) repair material undamaged damaged repaired repaired material c o m p re ss iv e f a il u re l o a d ( k n ) a.h. al-saidy 169 figure 11. load vs. mid span deflection of undamaged, damaged, and damaged/repaired beams. figure 12. repaired beam after failure. ductile manner as in the case of under reinforced concrete beams. failure occurred as the steel yielded, and ultimately the concrete was crushed in the compression zone. all beams showed similar behaviors until failure as expected. however, the objective was to observe the behavior of the patch rm under all loading stages, including in the inelastic (postyielding) stage. a summary of the results is shown in table 4. in all beams, the first cracks appeared in the concrete substrate and then in the rm. in beam drm1, the cracks first appeared in concrete at 3.3 kn whereas in the rm the cracks appeared at 5.6 kn; the other repaired beams also followed the same pattern (table 4). the variations in crack appearance can be attributed to the difference in the tensile strength as the rms have higher tensile strength than the concrete substrate. in all beams, the rms remained attached to the concrete until the beam failed (figs. 12 and 13), and no major difference was observed among the repairs made with different materials. mid span deflections (mm) l o a d ( k n ) performance of cement-based patch repair materials in plain and reinforced concrete members 170 table 4. summary of test results of repaired reinforced beams. beam load for first crack in concrete (kn) load for first crack in repair material (kn) maximum load (kn) maximum deflection (mm) c 19.2 47.0 d 18.5 45.0 drm1 3.3 5.6 19.4 41.0 drm2 4.0 6.2 19.3 44.0 drm3 4.4 6.7 19.3 35.5 drm4 5.0 6.9 19.7 43.0 figure 13. close up view of the repair material (rm2) at failure. 6. conclusion this study presented test results on the behavior of patch rms in plain and reinforced concrete. four rms were used and each differed in the mechanical properties of compressive strength, tensile strength and modulus of elasticity. based on the test results, it can be concluded that:  damage in plain concrete reduced the flexural and axial capacity considerably.  all rms performed very well in restoring the strength lost due to damage and shared the load with the substrate concrete in flexure and axial compression of plain concrete.  compatibility of the mechanical properties between the rm and the substrate concrete is very important. rm4 made with opc was the nearest in mechanical properties to that of the substrate concrete and hence more compatible. the failure mode of rm4 in slant shear and axial compression was the best among all rms in terms of adhesion to the substrate concrete at failure, indicating the importance of compatibility between the substrate concrete and the rms in enhancing the performance of the composite material.  in reinforced concrete beams, no difference was observed in the behavior of the rms and all rms showed good bond till failure. repair material concrete a.h. al-saidy 171 references abu-tair ai (1997), heat cycling of epoxy mortar repaired beams. aci materials journal 94(2): 129-133. al-farabi s, kalimur rm, al-gahtani a, hameeduddin m (2006), behaviour of patch repair of axially loaded reinforced concrete beams. cement and concrete composites 28: 734–741. american society for testing and materials, astm c109. standard test method for compressive strength of hydraulic cement mortars. american society for testing and materials, astm c469. standard test method for static modulus of elasticity and poisson's ratio of concrete in compression. american society for testing and materials, astm c78. standard test method for flexural strength of concrete (using simple beam with third-point loading). american society for testing and materials, astm c882. standard test method for bond strength of epoxy-resin systems used with concrete by slant shear. al-zahrani mm, maslehhuddin m, ibrahim m (2003) mechanical properties and durability characteristics of polymerand cement-based repair materials. cement and concrete research 25: 527-537. cabrera jg, al-hasan as (1997), performance properties of the concrete repair materials. construction and building materials 11(5-6): 283-290. courard l, schwall d, garbacz a, piotrowski t (2006), effect of concrete substrate texture on the adhesion properties of pcc repair mortar. in: aguiar jb, jalali s, camões a, ferreira r.m. (eds). proceedings ispic 2006 international symposium polymers on concrete guimarães, oficinas gráficas de barbosa and x a v i e r , l d a , b r a g a , p o r tu g a l 9 9–1 1 0 . courard l, piotrowski t, garbacz a (2014), near-to-surface properties affecting bond strength in concrete repair. cement and concrete composites 46: 73–80. garbacz, a., górka, m., courard, l., (2005), on the effect of concrete surface treatment on adhesion in repair systems. magazine of concrete research 57(1): 49–60. ghassan n, zai-ul-hasan c (1999), reinforced concrete repairs in beams. construction and building materials 1: 195-212. liu yp, wang fz, liu my, hu sg (2014), a microstructural approach to adherence mechanism of cement and asphalt mortar (ca mortar) to repair materials. construction and building materials 66: 125–131. mangat ps, limbachiya mk (1995), repair material properties which influence longterm performance of the concrete structure. construction and building materials 11(2): 8190. mangat ps, limbachiya mk (1997), repair materials properties for effective structural application. cement and concrete research 27(4): 601-617. margan dr (1996), compatibility of concrete repair materials and systems. construction and building materials 10(1): 57-67. qian j, you c, wang q, wang h, jia x, (2014), a method for assessing bond performance of cement-based repair materials. construction and building materials 68: 307–313. rio o, andrade c, izquierdo d, alonso c (2005), behavior of patch-repaired concrete structural elements under increasing static loads to flexural failure. journal of materials in civil engineering 17: 168-177. ueda h, tamai y, kudu t (2011), evaluation of the durability of cement-based repair materials. quarterly report of railway technical research institute, japan 52(2): 92-96. chen18904.qxd the journal of engineering research vol. 3, no. 1 (2006) 19-30 __________________________________________ *corresponding author’s e-mail: anishf@ksu.edu.sa sensitivity studies for propane oxidative dehydrogenation to propylene in circulating fluidized bed a.h. fakeeha*, m.a. soliman and a.a. ibrahim chemical engineering department, college of engineering, king saud university, p.o. box 800, riyadh 11421, saudi arabia received 18 september 2004; accepted 25 april 2005 abstract: in this paper, simulation and sensitivity studies for propane oxidative dehydrogenation to which were propylene in a circulating fluidized bed is undertaken using a previously developed model. various experimental kinetics, obtained by several investigators for the reaction using different catalysts, are employed in this study. a comparison is made for the performance of certain catalysts when used in a circulating fluidized bed reactor. the effects of changing reaction temperature, solid circulation rate, feed composition, pressure, and gas superficial velocity on reactant conversion and product selectivity and yields are studied. it was found that the catalyst prepared by ramos et al. is superior to the others with respect to yield. keywords: circulating fluidized bed, modeling, oxidation, propane and propylene notation ci,c = concentration of the ith species in the core, mole/l ci,a = concentration of the ith species in the annulus, mole/l dp = mean particle diameter, m dt = riser diameter, m g = gravitational acceleration constant, m/s2 gs = overall solids circulation rate, kg/(m2 s) k = proportionality constant in the acceleration zone, m-1 kg = gas mass-transfer coefficient from core to annulus, m/s kc = reaction rate constant k = equilibrium constant, mol/l lacc = length of the acceleration zone, m p = total pressure, pa ri = rate of the ith reaction, mol/(g.s) rc = core radius, m qgho ™«‡ ~¡e ‘ ú∏hhèdg ¤g ¿éhhèdg πjƒëàd úlhq~«¡∏d áyréædg √~°ùcód á«°sé°ùm äé°sgqo ¬¡«µa .ì.cg@º«ggôhcg .cg.cg h ¿éª«∏°s .cg.ω , ::áá°°uuóóÿÿggâe~îà°sg ~≤d .é≤hé°s √ôjƒ£j ” »°véjq êpƒ‰ ωg~îà°séh qgho ™«‡ ~¡e ‘ ú∏hhèdg ¤g ¿éhhèdg πjƒëàd úlhq~«¡∏d áyréædg i~°ùc’g á°sgqo åëñdg gòg øª° àj ‘ é¡eg~îà°sg ~æy ägõøùg √òg ago’ áfqé≤ã ωé«≤dg ”h .ägõøùg øe o~©d úãméh i~y ¥ôw øe égôjƒ£j ” »àdgh á«f髪«µdg á«côë∏d á«ñjôœ êpé‰ i~y á°sgq~dg √òg ‘ .á«léàf’g h ájqé«ÿgh á∏jƒëàdg ≈∏y ré¨dg áyô°sh §¨° dgh º«≤∏dg õ«côjh áñ∏°üdg ioéÿg ôjh~j ∫~©eh πyéøàdg iqgôm álqo øe πc òkéj á°sgqo â“ ~≤dh .qgh~dg ™«ªÿg ~¡ÿg .á«léàf’g ¢üîj 骫a èféàædg π° ag »£©j ¬béaqh ¢ùegq ᣰsgƒh ô° ùg õøùg ¿g ~lhh ::áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.ú∏hhôh ,¿éhhôh ,i~°ùcg ,álò‰ ,iqgh~dg á©«ªÿg ~¡ÿg 20 the journal of engineering research vol. 3, no. 1 (2006) 19-30 1. introduction alkanes such as propane are available in larger quantities than propylene since it is one of the components of natural or associated gas. the lower cost of propane, encourages further research to develop new catalysts and processes for propane utilization to produce many products such as propylene. the reaction can be carried out either in a fixed bed or circulating fluidized bed reactor (cfb), which is shown schematically in fig. 1. several investigators studied oxidation reaction in cfb both theoretically and experimentally. the theoretical simulation studies include the work of fakeeha et al. (2001) for the ammoxidation of propane to acronitrile, pugsley et al. (1992) and fakeeha et al. (2000) for nbutane oxidation to maleic anhydride, pugsley and berruti (1996) and pannek and mleczko (1998) for oxidative coupling of methane, pugsley and malcus (1997) for the partial oxidation of methane to synthesis gas, patience and mills (1994) for partial oxidation of propylene to acrolein and gainetto et al. (1990) for the ammoxidation of propylene into acrylonitrile and toluene to benzonitrile. on the contrary, golbig and werther (1997) carried out experimentally the partial oxidation of n-butane to maleic anhydride in cfb. in this paper simulation and sensitivity studies for propane oxidative dehydrogenation to propylene in cfb reactor will be carried out using three different kinetics based on mg-v-sb oxide catalysts (michaels et al. 1996; creaser and anderson, 1996; ramos et al. 2001) to determine the most suitable catalyst to carry out the reaction in cfb. sensitivity analysis of the operating conditions such as temperature, superficial velocity, pressure, solid circulation rate as well as propane and oxygen feed compositions will be performed. 2. model development the model consists of a couple of differential mass balance equations for each component, one for the mass balance in the core and the other for the mass balance in the r = riser radius, m re = reynolds number = dpu0ρg/µg rep = particle reynolds number = dputρg/µg t = temperature, k u0 = riser superficial gas velocity, m/s us = average solids velocity, m/s usc = core solids velocity, m/s ut = terminal settling velocity of a single solids particle, m/s x = axial location in the riser, m greek letters figure 1. schematic diagram of a circulating fluidized bed as catalytic reactor 0.3m riser stripper regenerator off gas regenerator air propane + air 20 m γ = constant in acceleration zone εann = annular voidage εavg. = average axial voidage εb = apparent voidage at riser bottom εc = core voidage εmf = voidage at mi nimum fluidization conditions ε∞ = voidage at t he end of the acceleration zone ρg , ρs = gas and solid particle densities, kg/m 3 21 the journal of engineering research vol. 3, no. 1 (2006) 19-30 annulus. this model requires knowledge of the reaction kinetics and the hydrodynamics, therefore the combined models of berruti and kalogerakis (1989) and wong et al. (1992) is used for the description of the hydrodynamics. accordingly, the hydrodynamic model equations are: us = gs / [ρs (1εavg)] (1) εavg = (u0 ρs) / (2 gs+u0ρ) (2) (3) (4) (5) where the apparent voidage at the riser bottom (εb) is calculated from a constant γ by solving the quadratic equation (6) with: εb = 0.714 (gs/ρs u0)-0.02528 dt-0.0794 rep-0.12016 (7) equation (6), being quadratic, has two roots. we choose the root which is less than one. theoretical development leading to eq. (9) requires γ to be a fraction. in eq. (6), k is given by: k = -ln (0.01)/lacc (8) finally we obtain the average voidage in the acceleration zone as (9) the cfb is assumed isothermal and reactions occur in core and annular regions. the gas input to the annulus is due solely to cross flow from the core. the following parameters and operating conditions are considered in the study which are catalyst properties: dp = 75 µm ; ρs = 1500 kg/m3 ; εmf = 0.5 ; ut = 2 m/s the simulation reactions require solution of two mass balance equations for each species: a core region mass balance (10) b. annular region mass balance (11) here the annular voidage, εann is assumed to be equal to the voidage at minimum fluidization conditions, εmf. the kg value is in the range of 0.015 0.094 m/s as used by pugsley et al. (1992). 3. numerical algorithm the height of the cfb riser (20 m) is divided into elements of a length of 0.2 m. testing the base-case with smaller elements length does not improve the accuracy of the simulation results. at each step, the hydrodynamic model is first solved to determine the length of acceleration zone, axial voidage profile, core porosity and core radius of elements. the concentration gradient δc/δx is discretized in the spatial direction and a forward implicit finite difference is used to solve the mass balance eqs. (10) and (11). since there are variations in gas volume due to the reaction at each element, the increase of gas velocity is determined and the hydrodynamic model is solved at the end of each element. the system of non-linear algebraic equations resulting from discretization is difficult to solve. to provide a good initial guess for the solution of the non-linear system of equations the nonlinear kinetics is linearized around the inlet conditions and the system of approximate linear equations is solved for every step in the spatial direction. then the system of non-linear algebraic equations are solved for the concentrations of components in the annulus and core regions at each step using the solution obtained from the linearized model as initial guess. having obtained the concentration of different components of reactants and axial conversion, selectivity and yield of products are calculated. 4. kinetics of oxidative dehydrogenation of propane to propylene many investigators studied the kinetics and mechanism of oxidative dehydrogenation of propane to propylene on different catalysts (ramos, et al. 2001; grabowski et al. 2003; creaser et al. 2000; chen et al. 2000; barsan and thyrion, 2003). in this paper the comparison of the per( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ε⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ρ ε−ρ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎣ ⎡ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ −+ −=ε t 2 cc o 2r s tmfs 2r s c u r ru r c u1 r c 1g 1 mfc mfav 2r r c ε−ε ε−ε =⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ 1 1 1 1 2− − = + −∞ ε ε b su k g γ γ( ) )kx(exp1 1( 1 ) avg −γ− − −= ∞ ε ε 0)cc( r k2 r)1()cu( x a,ic,ic g icsc,io =−−−+− ερδ δ 0)cc( rr rk2 r)1( a,ic,i2 c 2 cg ianns =− − +− ερ97.028.0 g0s 21.1 gs 76.0 pt 8 t acc re)u/g1()/( )d/d(1092.7 d l −− − + ×=⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ρρρ 22 the journal of engineering research vol. 3, no. 1 (2006) 19-30 formance of three fore mentioned catalysts in oxidation of propane based on their kinetics will be performed. michaels et al. (1996) studied the kinetics and mechanism of propane oxidation on mg2v2sbox. the temperature range used was 723-803 k. the network of reaction is shown in fig. 2. this catalyst will be called m catalyst in the subsequent work. it is noticed that the oxidative dehydrogenation and deep oxidation of the hydrocarbon occur by two distinctly different mechanisms. (1) mars-van krevelen mechanism fast, reversible (12) rate-determining (13) fast (14) where: ads denotes an adsorbed species, σhc is hydrocarbon adsorption site, oo and vo = are lattice oxygen and lattice, oxygen vacancy, respectively. (2) langmuir-hinshelwood mechanism fast,reversible (15) fast, reversible (16) rate-determining (17) where σo is oxygen adsorption site. rate expressions for oxidative dehydrogenation and deep oxidation of propane are given next. (18) (19) (20) where: r = reaction rate, k = rate constant, x = mole fraction, k = adsorption equilibrium constant and (21) where pro stands for propane and the entropy of adsorption value are 0.1 e.u. and -6.3 kj/mol respectively as shown in table 1. creaser and anderson (1996) studied the kinetics of propane oxidation on v-mg-o catalyst with 60 wt % mgo and 40 wt % v2o5. the temperature range used was 783823k. this catalyst will be called c catalyst in the subsequent work. the mechanism of the reaction is as follows. 1) propane reacts directly with surface oxygen according to (22) 2) the absorbed propylene reacts further with surface oxygen to give carbon oxides. (23) the rate equations derived take the form (24) (25) (26) )ads(hchc 63hc83 ⇔σ+ vohco)ads(hc 63o83 +→+ oo2vo2o 2 →+ → is irreversible, ⇔ is reversible and equilibrium, co x is carbon oxides )ads(hchc 83hc83 ⇔σ+ )ads(o22o o2 ⇔σ+ x83 co)ads(o)ads(hc →+ propanepropane propanepropane propane xk1 xk r + = propro 2/1 2oproco co xk1 xxk r + = propro 2/1 2opro2co 2co xk1 xxk r + = temperature, oc 450 474 500 530 108 )s/mol(k propane× 78 173.3 286.7 438.3 108 )s/mol(k co× 113.3 181.7 245 435 104 )s/mol(k 2co× 250 390 500 796.7 prok 2.58 3.55 3.27 2.37 e (activation energy) for partial oxidative dehydrogenation 103.3 kj /mol e (activation energy) for d eep oxidation co 79.1 kj/mol e (activation energy) for d eep oxidation co 2 65.3 kj/mol enthalpy -6.3 kj/mol entropy 0.1 e.u. table 1. kinetic model parameters (michaels et al. 1996) where n = 1 or 2, o = surface oxygen site, = free surface site. 08h3c11 pkr θ= 06h3c22 kr θθ= where the fractional cove rages oθ , 6h3cθ and the adsorption equilibrium constant for oxygen are given below respectively ; the kinetic parameters are given in table 2. 22 22 oo oo 0 pk1 pk + =θ figure 2. reaction network for propane oxidation (michaels et al. 1996) propro propro propane xk1 xk r + = ( ) ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∆− ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∆ = rt h exp r s xpeadsorptionk adsadspro propane propylene co, co2 23 the journal of engineering research vol. 3, no. 1 (2006) 19-30 (27) (28) where : tm = reference temperature, r = reaction rate, k = rate constant, k = adsorption equilibrium constant at temperature t, k0 = adsorption equilibrium constant at temperature tm. ramos et al. (2001) studied the kinetics of propane oxidation on v-mg-o with 24 wt% v2o5 and 76 wt% mgo. the range of temperature was 450550 k. this catalyst will be called r catalyst in the subsequent work. the reaction scheme suggested is given in fig.3 below: the power law model takes the form: i = 1, 2, 5., (29) (30) where: tm = 773.15 k, r = reaction rate, k0 = rate constant at reference temperature. the kinetic parameters for ramos et al. (2001) are presented in table 3, where: ea = activation energy, rc = regression coefficient. the reactor model and its numerical solutions used in this work were presented by fakeeha et al. (2000, 2001) before. 5. results and discussion the base case parameters used in this study are shown in table 4. the results for the base case for different catalysts at the end of the riser height of 20 m are shown in table 5. this case will be used in the sensitivity analysis as a basis for comparison. changing one parameter of the base case is performed each time while keeping the other parameters constant in order to carry out the sensitivity study. the effects of changing temperature, solid circulation rate, feed composition, pressure and superficial gas velocity are studied. the variation of the mass transfer coefficient kg in the range of 0.015-0.094 ms-1 was found to have a very slight effect on conversion. 5.1 effect of temperature for catalysts c and m, as temperature increases, the conversion increases linearly; while using catalyst r, the conversion along the riser increased initially at a sharp rate, then the rate of conversion increase becomes slow. however r catalyst gives the highest conversion, e.g. at a temperature of 823 k, the conversion with r catalyst could reach about 12%, c catalyst reaches about 5.5% and the m catalyst reaches only 1.0% (fig. 4). on the other hand, the selectivity with m catalyst is the highest and constant along the riser length, attaining values in the range of 97-98% at highest temperature. for catalyst c, the activation energy for the product k2kc3h6 , for co and co2 formation is lower than that for k1 for propylene formation, thus, we obtain higher selectivity for propylene at higher temperature (fig. 5). while for catalyst r, the changes of selectivity in the first 0-1m length of the riser are too small to be significant. moreover, the activation energies for the formation of co and co2 from propylene (r4, r5) are higher than that for propylene formation (r1). thus we obtain lower selectivity for propylene at higher temperature. in general the r catalyst gives the highest yield among the catalysts used due to its higher values of conversion and selectivity. 5.2 effect of solid circulation rate as solid circulation rate increases, conversion increases for all catalysts due to exposure to more catalyst. the increase in conversion has similar trend as that of the temparameter value confidence interval (95%) units 10k 2.50 × 10 -4 ±1.42 × 10-5 mol s-1 (g cat) -1bar-1 20k 2.35 × 10 -5 ±1.37 × 10-5 mol s-1 (g cat)-1 1ae 1.23 × 10 1 ±1.25 × 102 kj mol-1 2ae 2.34 × 10 1 ±1.31 × 102 kj mol-1 63 hc h∆ -2.29 × 10 2 ±1.53 × 102 kj mol-1 2o h∆ 3.56 × 10 2 ±3.59 × 102 kj mol-1 0 2o k 3.58 × 10 0 ±7.18 × 100 table 2. kinetic parameters of creaser and anderson (1996) 6363 6363 63 hchc hchc hc pk1 pk + =θ ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ − ∆− = m o0 oo t 1 t 1 r h expkk 222 figure 3. reaction network for propane oxidation (ramos et al. 2001) b hc 2 i p a opkir = ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ − − = m 0ii t 1 t 1 r e expkk 24 the journal of engineering research vol. 3, no. 1 (2006) 19-30 perature. the results are shown in figs. 6 and 7, for conversion and selectivity respectively. the highest conversion obtained for catalyst r is 5.5% at 783ok. for m and c catalysts selectivity increase is very slight and nearly constant with increase of solid circulation rate. the selectivity of propylene did not change apparently because the contribution of the deep oxidation of propylene is small due to the lower conversion of propane to propylene. however for r it decreases substantially for solid circulation rate. the yield for m and r catalysts increases with increase in circulation rate. for the c catalyst, the yield increases up to 600 kg/m2s but then decreases at 800 kg/m2s. 5.3 effect of feed composition for m and c catalysts, the increase in propane mole fraction does not affect the conversion, because the reaction is first order with respect to propane. for r catalyst, as propane mole fraction increases, conversion increases slightly because of propane partial pressure in the kinetic expression is having an exponent of 1.11. for all catalysts, there is no change in selectivity due to changes in propane mole fraction. the non-variance of selectivity on propane mole fraction indicates the same order of dependence on propane mole fraction for propylene, co and co2 formation. the selectivity for r and c catalysts decreases substantially along the riser length. the yield for m catalyst increases slightly and that for c catalysts is not affected by propane mole fraction as may be inferred from figs. 8 and 9, since there are no appreciable changes in their selectivities and conversions. while for r catalyst the yield also increases slightly with the increase in mole fraction. for c and r catalysts in fig. 10, conversion increases a little with oxygen mole fraction while for m catalyst conversion does not change because the propane rate of reaction to propylene does not depend on the oxygen mole fraction. all catalysts show a decrease in selectivity with the increase in oxygen mole fraction due to formation of carbon oxides as shown in fig. 11. the yield for m catalyst is not affected by the increase in oxygen mole fraction but those for c and r catalysts increase. 5.4 effect of pressure and superficial velocity all catalysts show increase in conversion with increase in pressure as depicted in fig. 12. however, the increase for m catalyst is fairly small. selectivity to propylene decreases with increase in pressure due to the formation of carbon oxides as shown in fig. 13. in all cases the yield increases with the increase in pressure, while m catalyst gives small increase. when the superficial velocity increases, conversion decreases in all cases due to lower residence time. for m catalyst, as velocity increases, selectivity decreases. for c and r catalysts, the selectivity initially decreases with the increase in velocity, but after a short distance along the riser it increases with increase in velocity. in all cases the yield decreases with increase in velocity. reaction rates 6 0 10×ik (mol/g.s.bar a+b) ae (kj/mol) a b rc 1r 2.30±0.11 157±5 0.05±0.04 1.11±0.06 0.987 2r 0.51±0.12 151±23 0.55±0.33 1.13±0.28 0.973 3r 1.27±0.13 97.5±11 0.63±0.19 1.03±0.14 0.962 4r 6.23±0.49 241±13 0.27±0.04 1.27±0.11 0.980 5r 11.8±0.48 202±6.7 0.21±0.02 1.14±0.06 0.985 table 3. kinetic parameters of (ramos et al. 2001) parameters values pressure, atm 1.0 mass transfer coefficient, kg m s -1 0.017 temperature , k 783.0 superficial gas velocity, u 0, m s -1 2.0 riser length , m 20.0 riser diameter, m 0.3 solid circulation rate, g, kg m-2 s-1 400.0 feed composition (%) c3h8 5.0 o2 2.0 table 4. the base case feed condition catalyst type % 63 hc conversion % 63 hc selectivity % 63 hc yield c 1.46 35.35 0.52 m 0.73 98.23 0.72 r 4.57 73.25 3.35 table 5. the results for the simulation of the base case at the exit of the reactor 25 the journal of engineering research vol.3, no. 1 (2006) 19-30 figure 4. effect of temperature on propane conversion figure 5. effect of temperature on propylene selectivity 26 the journal of engineering research vol.3, no. 1 (2006) 19-30 figure 6. effect of solid circulation rate on propane conversion figure 7. effect of solid circulation rate on propylene selectivity 27 the journal of engineering research vol.3, no. 1 (2006) 19-30 figure 8. effect of propane concentration on propane conversion figure 9. effect of propane concentration on propylene selectivity 28 the journal of engineering research vol.3, no. 1 (2006) 19-30 figure 10. effect of oxygen concentration on propane conversion figure 11. effect of oxygen concentration on propylene selectivity 29 the journal of engineering research vol.3, no. 1 (2006) 19-30 figure 12. effect of pressure on propane conversion figure 13. effect of pressure on propylene selectivity 30 the journal of engineering research vol. 3, no. 1 (2006) 19-30 6. conclusions sensitivity studies of three different catalysts for oxidative dehydrogenation of propane to propylene were carried out using a circulating fluidized bed reactor. the following conclusions are made. increasing temperature, solid circulation rate, and pressure, and decreasing the gas velocity increase the conversion. however, the catalysts studied behave differently with respect to selectivity. r catalyst is, in general, superior in terms of yield with respect to the other catalyst although its selectivity towards propylene formation is not the best. references barsan, m.m. and thyrion, f.c., 2003, "kinetic study of oxidative dehydrogenation of propane over ni-co, olybdate catalyst," catal. today., vol. 81, p. 159. berruti, f. and kalogerakis, n., 1989, "modeling the internal flow structure of circulating fluidized beds," can. j. chem. eng., vol. 67, p. 1010. chen, k., bell, a.t. and iglesia, e., 2000, "kinetics and mechanism of oxidative dehydrogenation of propane on vanadium, molybdenum, and tungsten oxides ," j. phys. chem. b., vol. 104, p. 1292. creaser, d. and andersson, b.o., 1996, "oxidative dehydrogenation of propane over v-mg-o: kinetic investigation by nonlinear regression analysis," appl. catal. a: general., vol. 141, p. 131. creaser, d.c., andersson, b., hudgins, r.r. and silveston, p.l., 2000, "kinetic modeling of oxygen dependence in oxidative dehydrogenation of propane," canadian j. of chem. eng., vol. 78, p. 183. fakeeha, a.h., soliman, m.a. and ibrahim, a.a., 2001, "comparison between vpo catalysts for production of maleic anhydride from butane in a circulating fluidized bed," afinidad lviii., vol. 494, p. 261. fakeeha, a.h., soliman, m.a. and ibrahim, a.a., 2000, "modeling of a circulating fluidized bed for ammoxidation of propane to acrylonitrile," chem. eng. and proc., vol. 39, p. 161. gianetto, a., pagliolico, s., rovero, g. and ruggeri, b., 1990, "theoretical and practical aspects of circulating fluidized bed catalytic reactions (cfbrs) for complex chemical systems," chem. eng. sci., vol. 45, p. 2219. goibig, k.g. and werther, j., 1997, "selective synthesis of maleic anhydride by spatial separation of nbutane oxidation and catalyst reoxidation," chem. eng. sci., vol. 52, p. 583. grabowski, r., sloczynnski, j. and grzesik, n.m., 2003, "kinetics of oxidative dehydrogenation of propane over v2o5/tio2 catalyst", appl. catal. a: general., vol. 242, p. 297. michaels, n.j., stern, d.l. and grasselli, r.k., 1996, "oxydehydrogenation of propane over mg-v-sboxide catalyst. ii reaction kinetics and mechanism," catalysis letters, vol. 47, p. 139. pannek, u. and mleczko, l., 1998, "reaction engineering simulation of oxidative coupling of methane in a circulating fluidized bed reactor," chem. eng. technol., vol. 21, p. 10. patience, g.s. and mills, p.l., 1994, "modeling of propylene oxidation in a circulating fluidized bed reactor," in: v.c. corberan, s.v. bellon (eds.), new developments in selective oxidation ii, elsevier, new york, p. 1. pugsley t.s. and berruti, f., 1996, "the circulating fluidized bed catalytic reactor model validation and simulation of the oxidative coupling of methane," chem. eng. sci., vol. 51, p. 2751. pugsley, t.s. and malcus, s., 1997, "partial oxidation of methane in a circulating fluidized bed catalytic reactor," ind. eng. chem. res., vol. 36, p. 4571. pugsley, t.s., patience, s.g., berruti, f. and chaoki, j., 1992, "modeling the catalytic oxidation of n-butane to maleic anhydride in circulating fluidized bed reactor," ind. eng. chem. res., vol. 31, p. 2652. ramos, r., pina, m.p., menendez, m., santamaria, j. and patience, g.s., 2001, "oxidative dehydrogenation of propane to propene, 1: kinetic study on v/mg," canadian j. of chem. eng., vol. 79, p. 891. ramos, r., pina, m.p., menendez, m., santamaria, j. and patience, g.s., 2001, "oxidative dehydrogenation of propane to propene, 2: simulation of a commercial inert membrane reactor immersed in a fluidized bed," canadian j. of chem. eng., vol. 79, p. 902. wong, r., pugsley, t. and berruti, f., 1992 "modeling the axial voidage profile and flow structure in the riser of circulating fluidized beds," chem. eng. sci., vol. 47(9-11), p. 2301. chen041127.qxd the journal of engineering research vol. 3, no.1 (2006) 31-37 1. introduction essential oils can be broadly defined as volatile oils that differ fundamentally from the fixed fatty oils such as linseed, coconut and olive in being more mobile and volatile. many essential oils, apart from being used for first aid or for treatment of common complaints, are also ideal as bath oils, perfumes or room fresheners. even when they are used purely for aesthetic purposes, they still play a positive preventive and therapeutic role (lawless, 1997). conventional extraction of essential oils includes: tumbling and shaking, oil infusion, cold expression, steam distillation and soxhlet extraction can take from hours to days, and requires significant amounts of ________________________________________ *corresponding author’s e-mail: amer@curtin.edu.my solvent. following the results of the various experiments carried out, it becomes obvious that extraction using microwave technology is a good alternative to conventional extraction techniques. the extraction by microwave as a new extraction technique has its own specific parameters that need to be characterized for every plant that contains essential oil. when a material is exposed to microwaves, the polar molecules within the matrix become excited. then they orient themselves in the direction opposite to the field. the frequency of the change of polarity of the electric field determines the frequency at which the molecules will rotate within the field. as a result of this rotation, temperatures in the vicinity of the rotating molecule increase. the applied field prevents the molecules from reaching an equilibrium position and, thus, the polar molecules store potential energy while they microwave-assisted extraction of essential oil from eucalyptus: study of the effects of operating conditions a.a. saoud*1, r.m. yunus2 and r.a. aziz3 1curtin university of technology, school of engineering and science, cdt 250, 98009 miri, sarawak, malaysia 2faculty of chemical & natural resources engineering, university of teknologi, 81310 skudai, johar, malaysia 3chemical engineering pilot plant, faculty of chemical & natural resources engineering, university teknologi, 81310 skudai, johar, malaysia received 27 november 2004; accepted 28 may 2005 abstract: classical extraction of essential oil such as soxhlet and steam distillation is still a formidable and time-solvent consuming. microwave assisted process (map) is used to accelerate the extraction process of target compounds. it can be used for the extraction of compounds from various plants and animal tissues, or the extraction of undesirable components from raw materials. the investigation of microwave extraction of eucalyptus (globules) essential oil using ethanol as solvent was carried out. the influence of material (eucalyptus)/solvent (ethanol) ratio, required doses of microwave, and time of microwave exposure on extraction efficiency, was studied. keywords: essential oil, microwave, ethanol, eucalyptus, extraction 𫨰ûààdg ±hôx ägòkéàd á°sgqo :á≤«b~dg êggƒe’g i~yé°ùãã qƒaéµdg øe qé«£dg âjõdg ¢uóîà°scg oƒ©°s .cg.cg1@¢ùfƒj .ω.q ,2õjõy .cg .q h3 ::áá°°uuóóÿÿggi~yé°ùã ¢uóîà°s’g á≤jôw ω~îà°ùj .áñjòÿg ioéÿgh âbƒ∏d éaõæà°ùeh éñ©°u qéîñdéh ò£≤àdg h â∏¡°ùcƒ°s πãe áj~«∏≤àdg ¥ô£déh qé«£dg âjõdg ¢uóîà°sg ∫gr ée øe é¡«a üƒzôÿg òz äéfƒµÿg ádgr’ hg á«fgƒ«m h á«jéñf áé°ùfg øe äéñcôe ¢uóîà°s’ é¡eg~îà°sg øµáh áa~¡à°ùÿg äéñcôª∏d ¢uóîà°s’g áyô°s ioéjõd á≤«b~dg êgƒe’g iaéøc ≈∏y á«déàdg πegƒ©dg òkéj á°sgqo h öjòªc »∏«ã«ÿg ∫ƒëµdg ωg~îà°séh á≤«b~dg êgƒe’éh qƒaéµdg äéñ«ñm øe qé«£dg âjõ∏d ¢uóîà°sg á«∏ªy agôlg ” ~≤d .ωéÿg ogƒÿg .é¡d ¢vô©àdg øer h á≤«b~dg êgƒe’g øe áhƒ∏£ÿg äéyô÷gh (»∏«ã«ÿg ∫ƒëµdg) öjòÿg ¤g (qƒaéµdg) ioéÿg áñ°ùf :¢uóîà°s’g ::áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.¢uóîà°s’g ,qƒaéµdg ,»∏«ã«ÿg ∫ƒëµdg ,á≤«b~dg êgƒe’g ,qé«£dg âjõdg 32 the journal of engineering research vol. 3, no.1 (2006) 31-37 are within the field. when the field is removed, the molecules relax to an equilibrium position and release excess kinetic energy in the form of heat. microwave applicability for the extraction of various types of compounds from plant material, food and soil was investigated by ganzler et al. (1986). the production of volatile material from plant exposed to microwave energy in an air stream has been discussed by craveiro et al. (1989). chen and spiro (1994) studied the microwave heating characteristics of extraction of rosemary and peppermint leaves suspended in hexane, ethanol and mixture of hexane and ethanol. pare (1995) patented a general extraction method for biological matter using microwave energy and a microwave transparent solvent. weseler et al. (2002) has studied the antibacterial activity of australian tea-tree oil against several strains of malassezia pachydermatis. different solvents such as ethanol, trichloromethane, cyclohexane, nhexane, were used to study the effect of solvency on the extraction of artemisinin from artemisia annua using microwave-assisted process by jin-yu hao et al. (2002). marie et al. (2004) used a new microwave technique -solvent free microwave extraction(sfme) to extract essential oil from basil (ocimum basilicum l.), garden mint (mentha crispa l.), and thyme (thymus vulgaris l.). the findings showed the aromatic profile of microwave exposure are similar to those obtained by hydrodistillation for 4.5 h. this work is concerned with the optimization of microwave parameters: eucalyptus/ethanol ratio, required doses of microwave, heating then cooling that lead to sequence of sudden expansion to help breaking the gland faster and, thus, released more solutes into the extracting solvent, and time of exposure. different ratios of eucalyptus leaves/ethanol have been examined in order to obtain the optimal material/solvent ratio gives highest yield of extracted essential oil. the required number of microwave doses that provides an accomplishing extraction process has been ascertained. the optimal time of microwave was determined. 2. experimental procedure 2.1 materials and chemicals standard samples of the major eucalyptus constituents (cineole, α-pinene, camphene, cymene, limonene and γ terpinene) of high purity (purchased from sigma-aldrich) were used in this study for gc calibration. the calibration was done by preparing known concentrations of each component in ethanol solution and the samples were introduced to the gc for analysis. the peaks obtained provide the retention time and the corresponding area, representing the component concentration. leaves of eucalyptus plant were supplied by nasuha enterprise plantation (malaysia). ethanol of 96 v/v % was supplied by flukaswitzerland. 2.2 equipment in this study, gas chromatography (gc, perkin elmerauto system xl gas chromatograph) was utilized for the analysis of the extracts. computerized microwave solvent extraction system (milestone-ethos sel microwave lab station) was used for microwave extraction. 2.3 procedure in order to measure the optimal material/solvent ratio, required number of microwave doses, and exposure time, the following procedure was pursued: samples of eucalyptus leaves were placed in sealed containers of 270 ml of each. equal volumes (100 ml) of solvent (ethanol) was used throughout the tests. a magnetic stirrer agitated the solution. in between each dose of exposure, the solutions were allowed to cool back to room temperature (28oc). sampling of 2 ml was carried out at the end of each dose of exposure, thus, the solute concentrations were corrected taking into account the change in solution volume. the extraction was done in sealed vessels, where no evaporation was observed. all solute concentrations are reported in mg/l. power of 1000 w was used in the experiments. equation (1) was used to correct the concentration due to sampling: (1) where c = corrected concentration (mg/l) cn = concentration obtained directly from chromatogram (mg/l) n = number of sampling v = initial volume of solvent (100 ml) 3. results and discussions 3.1 influence of eucalyptus/ethanol ratio on microwave extraction tables 1-5 show the constituent concentrations for different ratios of eucalyptus/ethanol at same conditions of time and power. the total amounts of cineole, α-pinene, limonene, cymene, γ-terpinene, and camphene that can be extracted by microwave extraction, showed an increasing trend of extracted constituents with the increase in exposure doses. obviously, it can be deduced that ethanol can extract eucalyptus essential oil for the ratio of 2, 4 and 6 g of eucalyptus/100 ml ethanol, but for the ratios of 8 and 10 g, ethanol could not extract more essential oil due to over saturation. figures 1-6 show the concentrations of each constituent at different ratios and successive ten doses per one g of leaves in order to make comparison among different ratios used. cineole could be detected from first dose at all ratios while α-pinene could not be detected for the ratio of 2 g, but it appeared for the rest of the ratios. for the ratios of 2 and 4 g, limonene was not observed; but at 6 and 8 g it needs minimum of 8 doses to appear; while at 10 g it could be taken out from the second dose. cymene was extracted at all ratios, but it needs 10 doses for the first ratio (2 g); while for the ratio of 4 g, 7 doses were sufficient. for ratios of 6 and 8 g, it needed v/]1n(2v[cc n −−= 33 the journal of engineering research vol. 3, no.1 (2006) 31-37 five doses; whilst at the ratio of 10 g, it was appreciable from the first dose. camphene and γ-terpinene showed similar readiness of extraction for all ratios but at different doses. the results showed that at the ratio of 6 g leaves/100 ml ethanol, maximum extraction could be obtained. at this ratio, the yield of essential oil was 3.576 mg/g eucalyptus of fresh leaves. the percentage of major constituent (cineole) was 68; while for α-pinene and limonene it was 16 % and 8 % respectively, the rest of constituents (cymene, γ-terpinene, and camphene) represent 8%. 3.2 influence of exposure doses on microwave extraction microwave extraction technique works by exposing the material to a sudden heat caused by the microwave field, to integrate this process, a successive exposing and cooling are required. after each microwave exposure, the no. of doses constituents 1 2 3 4 5 6 7 8 9 10 cineole 3.27 3.36 3.52 3.81 3.90 4.17 4.21 4.26 4.87 4.89 α-pinene limonene cymene 0.12 γ-terpinene 0.32 0.33 camphene 0.12 0.13 table 1. constituent concentrations (mg/ml) of the ratio 2 g eucalyptus/100 ml ethanol, 60 seconds exposure and 1000 w no. of doses constituents 1 2 3 4 5 6 7 8 9 10 cineole 5.49 5.66 5.72 6.09 6.13 8.13 8.22 8.56 9.42 9.44 α-pinene 1.81 2.30 2.32 limonene cymene 0.13 0.18 0.25 0.25 γ-terpinene 0.33 0.44 0.53 0.64 0.66 camphene 0.13 0.17 0.23 0.25 0.25 table 2. constituent concentration (mg/100 ml) of the ratio 4 g eucalyptus/100 ml ethanol, 60 seconds exposure and 1000 w no. of doses constituents 1 2 3 4 5 6 7 8 9 10 cineole 8.27 8.96 9.52 10.81 11.90 12.70 13.10 14.25 14.57 14.58 α-pinene 1.83 2.41 2.84 3.10 3.39 3.42 limonene 1.08 1.68 1.70 cymene 0.13 0.20 0.26 0.31 0.37 0.38 γ-terpinene 0.32 0.47 0.59 0.66 0.75 0.88 0.95 0.98 camphene 0.13 0.24 0.30 0.35 0.39 0.39 table 3. constituent concentration (mg/100 ml) of the ratio 6 g eucalyptus/100 ml ethanol, 60 seconds exposure and 1000 w no. of doses constituents 1 2 3 4 5 6 7 8 9 10 cineole 10.31 11.36 11.52 12.81 13.90 14.70 15.10 15.25 16.82 16.89 α-pinene 1.87 2.15 2.63 2.70 2.81 2.96 3.64 3.65 limonene 1.12 1.80 1.83 cymene 0.13 0.19 0.26 0.32 0.41 0.41 γ-terpinene 0.33 0.38 0.43 0.50 0.61 0.78 0.86 0.96 1.10 1.12 camphene 0.14 0.18 0.22 0.28 0.31 0.35 0.38 0.40 table 4. constiturent concentration (mg/100 ml) of the ratio 8 g eucalyptus/100 ml ethanol, 60 seconds exposure and 1000 w 34 the journal of engineering research vol. 3, no.1 (2006) 31-37 no. of doses constituents 1 2 3 4 5 6 7 8 9 10 cineole 13.30 14.32 15.57 15.90 16.58 17.62 18.83 19.60 20.74 20.76 α-pinene 2.56 2.83 3.11 3.32 3.65 3.76 3.91 4.21 4.84 4.85 limonene 1.11 1.30 1.41 1.46 1.49 1.55 1.72 1.86 1.88 cymene 0.13 0.20 0.27 0.30 0.31 0.33 0.36 0.44 0.53 0.55 γ-terpinene 0.38 0.39 0.43 0.46 0.62 0.78 0.85 0.93 1.15 1.16 camphene 0.13 0.14 0.19 0.24 0.28 0.32 0.37 0.41 0.41 table 5. constituent concentration (mg/100 ml) of the ratio 10 g eucalyptus/100 ml ethanol, 60 seconds exposure and 1000 w figure 1. cineole cumulative concentration-microwave dose plot at 1000 w and 60 seconds at different eucalylptus/ethanol ratios figure 2. α pinene cumulative concentration-microwave dose plot at 1000 w and 60 seconds at different eucalyptus/ethanol ratios 35 the journal of engineering research vol. 3, no.1 (2006) 31-37 figure 3. limonene cumulative concentration-microwave dose plot at 1000 w and 60 seconds at different eucalyptus/ethanol ratios figure 4. cymene cumulative concentration-microwave dose plot at 1000 w and 60 seconds at different eucalyptus/ethanol ratios 2 4 6 8 10 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 γ-terpinene concentration mg/100 ml g material/ 100 ml solvent irradiation dose figure 5. γ-terpinene cumulative concentration-microwave dose plkot at 1000 w and 60 seconds at different eucalyptus/ethanol ratios 36 the journal of engineering research vol. 3, no.1 (2006) 31-37 solution was allowed to cool back to room temperature (28 oc). the natural cooling process helped releasing more solutes from the solid matter and allowed the solute contained gland to return to its original size. the number of microwave doses and corresponding constituents' concentrations are listed in tables 1-5. ten doses of exposure are required due to the adjacent values of concentrations obtained at 9 and 10 doses. hence, the required number for successive exposing and cooling that provides an accomplishing extraction process to achieve the equilibrium of extraction is ten. 3.3 influence of microwave exposure time on eucalyptus essential oil extraction microwave exposure was tested for times of 1 to 5 minutes to inspect the influence of exposure time on the essential oil extraction at 1000 w, material /solvent ratio of 6 g/100 ml ethanol and 10 successive doses. the temperature increases (∆t) measured were 25, 40, 56, 68 and 80oc. figure 7 shows the concentration of the constituents measured with an exposure time of 3 minutes, which brings about equilibrium and gives highest yield. cineole concentration increased with time and reached the equilibrium at 3 minutes of exposure; while the other constituent concentrations showed plateau condition, which indicates that equilibrium condition has been reached much earlier (during the first minute of exposure). optimization of time of exposure yielded 4.50 mg/g of eucalyptus from the fresh leaves. the content of cineole, α-pinene and limonene were 72%, 13.6% and 7.1% respectively. 4. conclusions material-solvent ratio has an important influence on microwave extraction of eucalyptus leaves in ethanol. the results showed that the highest concentration was figure 6. camphene cumulative concentration-microwave dose plot at 1000 w and 60 seconds at different eucalyptus/ethanol ratios figure 7. eucalyptus essential oil constituent concentrations per g of eucalyptus leaves at different microwave exposure times 37 the journal of engineering research vol. 3, no.1 (2006) 31-37 obtained at the ratio of 6 g leaves/100 ml ethanol. ten exposure doses were required to achieve the equilibrium extraction. microwave exposure time of 3 minutes was enough for microwave extraction process. the yield of eucalyptus essential oil was 4.5 mg/g fresh leaves under the conditions: 6 g/100 ml ethanol, 10 doses of exposure, microwave power delivered at 1000w, exposure time of 3 minutes. references chen, s. and spiro, m., 1994, "study of microwave extraction of essential oil constituents from plant materials," j. microwave power electromagnetic energy, vol. 29, pp. 231-241. craveiro, a.a., matos, f.j.a. and alencar, j.w., 1989, "microwave extraction of an essential oil," j. flavor and fragrance, vol. 4, pp. 43-44. ganzler, k., salgo, a. and volko, k., 1986, "determination of lead contented on particulate matter filters by microwave extraction and analysis by atomic absorption spectrometry," j. chromatography, vol. 371, pp. 299-306. hao, j., han, w., huang, s., xue, b. and deng, x., 2002, "microwave-assisted extraction of artemisinin from artemisia annua l," j. separation and purification technology, vol. 28, pp. 191-196. lawless, j., 1997, "the complete illustrated guide to aromatherapy," element books limited, pp. 18-25. marie, e. l., farid, c. and jacqueline, s., 2004, "solvent free microwave extraction of essential oil from aromatic herbs: comparison with conventional hydro-distillation," j. of chromatography, vol. 1043, pp. 323-327. pare, j.r.j., 1995, "microwave-assisted extraction from materials containing organic matter," us patent, vol. 5, pp. 458, 897. weseler, a., geiss, h.k., saller, r. and reichling, j., 2002, "antifungal effect of australian tea tree oil on malassezia pachydermatis isolated from canines suffering from cutaneous skin disease", j. schweizer archiv fur tierheilkunde, vol. 144, pp. 215-221. microsoft word paper 7.docx the journal of engineering research (tjer) vol. 14, no. 1 (2017) 74-84 performance comparison of digital circuits using subthreshold leakage power reduction techniques b. kalagadda*, a, n. muthyalab and k.k. korlapatic *a, b department of electronics and communication engineering, kakatiya university college of engineering and technology, kakatiya university, warangal-506001, telangana, india. c department of electronics and communication engineering, national institute of technology, warangal-506004, telangana, india. received 8 june 2015; accepted 2 october 2016 abstract: complementary metal-oxide semiconductors (cmos), stack, sleep and sleepy keeper techniques are used to control sub-threshold leakage. these effective low-power digital circuit design approaches reduce the overall power dissipation. in this paper, the characteristics of inverter, twoinput negative-and (nand) gate, and half adder digital circuits were analyzed and compared in 45nm, 120nm, 180nm technology nodes by applying several leakage power reduction methodologies to conventional cmos designs. the sleepy keeper technique when compared to other techniques dissipates less static power. the advantage of the sleepy keeper technique is mainly its ability to preserve the logic state of a digital circuit while reducing subthreshold leakage power dissipation. keywords: sub-threshold leakage, stack, sleep, sleepy keeper, static power. אאאאאאאא  kא،*k،k، wא א       א אvlsi  א kcmosא ، א ، אאאאאkאאאאא  א  א  א א א א؛   k   א nand אא٤٥،١٢٠،١٨٠   א   cmos  א  א     kא אאאkאאאאאא kא   wאאkאא،א،א،א،אא   * corresponding author’s email: kalagaddaashu@kakatiya.ac.in b. kalagadda, n. muthyala and k.k. korlapati 75 1. introduction integrated circuit technology enables the innovative devices and systems that change the essence of life. the continuous development of high performance integrated circuits has led to an explosion in communication and computation systems. major concerns with very-large-scale integration (vlsi) designs recently have shifted to power considerations along with area and speed. power management for vlsi design has become vital due to a reduction in the geometry of transistors and also exponential growth in the number of transistors on a single chip. this trend developed because of special portable devices and applications. the change in technology has been the result of development in vlsi techniques, which has emerged as a rapidly growing technology influencing the world in encouraging innovation. improvements in manufacturing have led to ever-smaller transistors which can accommodate a greater number of transistors on a chip. the integrated circuits industry has increased the functionality of larger chips by allowing a steady path for constantly shrinking device geometries. each new generation has approximately doubled logic circuit density and increased performance by about 40%. due to advanced ic technology, the minimum feature size of vlsi circuitry continues to decrease. the feature size of transistors has entered the nano-scale region due to its shrinking capacity. the transistors with smaller feature sizes have the capability of increasing the speed of the circuit and yield of manufactured chips. the latter is due to the smaller silicon area of the chip, which lowers the average number of defects per chip. vlsi chip manufacturers effectively utilize the advantages of possible reductions in feature size by scaling (shrinking) the existing designs of integrated circuits (ics) (borkar 1999). the better performance and rich feature sets on a single chip have evolved with an increase in integration level. as the technology scaling has continued, many design techniques mainly from the field of power management must be employed in order to attain a balanced design to meet end user needs while maintaining the performance trend of a vlsi system. the need to dissipate low power for electronic devices in order to conserve battery life and meet packaging reliability constraints has recently arisen. in portable devices, the reduction in power consumption is essential for improving battery life and reliability and reducing heat removal costs mainly in high performance systems. complementary metal oxide semiconductor (cmos) is a technology for constructing integrated circuits. cmos devices possess important characteristics like high noise immunity and low static power consumption. in vlsi circuits, cmos technology has become the most frequently used technology for implementation purposes because cmos allows for a high density of logic functions on a chip. recently much progress has been made in estimating power dissipation; power dissipation has thereby driven the design of new low power systems, and accurate and efficient power estimation techniques have been used (bursky 1995). cmos transistors, as they constantly switch, dissipate power. due to this feature, dynamic power consumption is related to the number and size of transistors and also to the rate at which they switch. when the feature size shrinks below 180nm, the transistors leak a significant amount of current even in the off state. the chips, therefore, draw static power even when they are idle. the main challenge of vlsi design is to make a good tradeoff between performance and power for a particular application. to achieve this, the methods, which reduce power leakage, have to be adopted with conventional cmos circuits. this paper aims to detail a method to reduce sub-threshold leakage power since it is the most dominant among all the leakages in the deep sub-micron technologies (saini and mishra 2012). in this paper, several leakage reduction methodologies like stack, sleep, and sleepy keeper techniques were applied with the conventional digital cmos circuits like an inverter, a two-input nand gate, and a half adder in 180nm, 120nm, and 45nm technologies. the performance of these circuits was evaluated among those leakage reduction techniques. the current paper is organized as follows: section 2 describes the kinds of power dissipation, section 3 discusses several leakage power reduction mechanisms, section 4 presents the results, and section 5 concludes the paper. performance comparison of digital circuits using subthreshold leakage power reduction techniques 76 2. power dissipation power dissipation in cmos circuits is generally due to dynamic dissipation and static dissipation. the total power in a circuit is given by staticdynamictotal ppp  (1) where ptotal is the total power dissipation of the circuit, pdynamic is the dynamic power consumption and pstatic is the static power consumption. 2.1 dynamic power dissipation the dynamic power dissipation mainly occurs due to the switching activity of the transistors and the continuous charging and discharging of the mos capacitors. this also occurs due to short-circuit current when both pmos and nmos stacks are partially on. 2.2 static power dissipation the static power dissipation is mainly due to the sub-threshold leakage current through off transistors, gate leakage through gate dielectric material, junction leakage from source or drain diffusions, and contention current in ratioed logic circuits. the static power exists even when the chip is not in active mode. the power is also described in terms of active, standby, and sleep modes. the power that is utilized when the chip is on is the active power, which is dominated by switching activity. when the chip is in a non-active state, standby power exists. the standby power is set by leakage in case of halted clocks and disabled ratioed logic circuits. the sleep mode intends to avoid leakage by disconnecting the power supply to unwanted circuits on a chip. in nanometer technologies, nearly one-third of the total power is considered leakage power. since sub-threshold leakage is considered the most dominant of all leakages, a detailed explanation of it is given in the next section. 2.3 subthreshold leakage power subthreshold leakage is the dominant source of static power. subthreshold leakage or conduction occurs due to thermal emission of carriers over the potential barrier, which is set by the threshold (boray et al. 2007). subthreshold leakage current flows even when a transistor is nominally off. sub-threshold leakage increases due to the downscaling of threshold voltages. in weak inversion, below threshold voltage the transistors are not completely off and the sub-threshold current relies on threshold voltage. the sub-threshold leakage increases exponentially with a decrease in threshold voltage or rises with temperature, which is a major problem for chips using low supply and threshold voltages and operating at high temperatures. the sub-threshold leakage current is given as:            t ds t sbdstogs v v nv vkvvv 0dsds e1eii (2) where 0dsi is the current at threshold and is dependent on process and device geometry. 8.12 t0ds evi  (3) the process-dependent term n is affected by depletion region characteristics; the term ds v indicates a reduction of threshold voltage by drain-induced barrier lowering; vgs, vds, and vsb indicate gate-to-source, drain-to-source, and source-to-bulk voltages of a metal-oxidesemiconductor (mos) transistor, respectively; tv is the thermal voltage; vto indicates a zerobased threshold; k is the body effect coefficient, and the e1.8 term is found empirically. the sub-threshold current path is shown in fig. 1. leakage generally accounts for a third of the total active power when low figure 1. subthreshold leakage current path. b. kalagadda, n. muthyala and k.k. korlapati 77 figure 2. method to measure subthreshold leakage current through (a) nmos transistor, and (b) pmos transistor. threshold voltages and thin gate oxides are applied in nanometer processes. except in very low power applications, leakage is insignificant at above 180 nm. in 45 nm, high–k gate dielectrics is used so that sub-threshold leakage is not dominated by gate leakage. necessary reduction techniques are then employed to reduce sub-threshold leakage. static power comes from the leakage and circuits with a path from the vdd (voltage drain) to the gnd (ground drain). when operated at high threshold voltages, the complementary cmos gates dissipate almost zero static power, so the circuits consume low power. the leakage increases as feature size decreases making static power consumption dominant along with dynamic power. 2.4 method to measure subthreshold leakage current in a general metal–oxide–semiconductor field-effect transistor (mosfet) sub-threshold current is found by switching off the gate terminal. the current from vdd to gnd in this case is the subthreshold leakage current. the scenario is shown in fig. 2. 3. reduction techniques of subthreshold leakage the low power design is mainly to reduce the predominant factor of power dissipation. so in nanometer processes, overall leakage becomes an important design constraint (roy et al. 2003). the sub-threshold leakage which is dominant in leakage or static power dissipation is to be reduced. the following are techniques to reduce sub-threshold leakage so that power dissipation can be decreased, which is a vlsi design constraint. 3.1 stack technique the leakage trough of two or more transistors is reduced on account of the stack effect. stacks with three or more off transistors have even lower leakage levels. a circuit can be stacked into two half-width size transistors but propagation delay rises with number of transistors (johnson et al. 2002). figure 3 shows the stack technique. in this technique, the pmos transistors’ width in the pull-up network is divided and stacked into two half-width size transistors. similarly, the width of the nmos transistors in the pull-down network is also divided and stacked into two half-width size transistors. the major disadvantage with this technique is that it increases the overall propagation delay. 3.2 sleep technique the sleep technique uses sleep transistors with high threshold voltages. figure 4 depicts the sleep technique. the sleep signal is used to figure 3. stack technique. gnd gnd gnd gnd performance comparison of digital circuits using subthreshold leakage power reduction techniques 78 figure 4. sleep technique. drive the sleep transistors. in this technique, the pull-up and pull-down networks are separated from supply and ground respectively by sleep pmos and nmos transistors (kumar et al., 2012). in sleep state, the transistors are off which reduces the sub-threshold leakage current. this technique comes under the state destructive technique. this implies that the circuit state is lost but the leakage is reduced. the disadvantages of this technique are the destruction of the state of the circuit and the fabrication of high threshold voltage sleep transistors with conventional mos transistors on a single-die required triple well process. 3.3 sleepy keeper technique the sleepy keeper technique is a statepreserving technique. it has advantageous characteristics over the sleep technique. figure 5 describes the sleepy keeper technique. in this technique, high threshold voltage pmos and nmos transistors are additionally used. this is the same as the sleep technique, but the additional high threshold voltage nmos transistor is placed parallel to the sleep transistor. similarly, the additional high threshold voltage pmos transistor is placed parallel to the sleep nmos transistor. when the circuit is in sleep mode, the high threshold voltage nmos transistor is connected to the figure 5. sleep keeper technique. supply and the high threshold voltage pmos transistor is connected to ground. the sleepy keeper technique is advantageous when compared to all other techniques. it reduces the sub-threshold leakage current along with maintaining its state. 4. results and discussion the performance of digital circuits like inverters, two-input nand gates, and a half adder have been studied in the current article. microwind (microwind, toulouse, france) was used for layout and simulation purposes using a bsim4 model in 45nm, 120nm, and 180nm technologies. stack, sleep and sleepy keeper sub-threshold reduction techniques were used to calculate dynamic power dissipation, static power dissipation, and propagation delay at 27oc. the channel widths and lengths were maintained according to the technology node. the results were compared against the parameters obtained with conventional cmos designs. figure 6 shows the layout of a conventional cmos inverter. figures 7–9 correspond to the layouts of the inverter with stack, sleep, and sleepy keeper techniques, respectively, which were simulated in 45nm technology. gnd gnd b. kalagadda, n. muthyala and k.k. korlapati 79 figure 6. layout of conventional cmos inverter 45nm technology. figure 7. inverter layout with stack technique in 45nm technology. figure 8. inverter layout with sleep technique in 45nm technology. performance comparison of digital circuits using subthreshold leakage power reduction techniques 80 figure 9. inverter layout with sleepy keeper technique in 45nm technology. table 1 compares the performance of inverter using 180, 120 and 45 nm technologies with different leakage reduction mechanisms applied on a conventional cmos inverter (fig. 10). table 1. inverter performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. technology 180nm 120nm 45nm technique dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) sub-threshold leakage power (µw) conventional cmos 9.917 2.12 2.473 0.469 0.078 0.021 stack 9.672 1.919 2.425 0.426 0.073 0.018 sleep 7.78 1.723 1.072 0.329 0.086 0.015 sleepy keeper 9.123 1.694 2.364 0.271 0.094 0.013 table 2 compares the performance of a two input nand gate in 180, 120 and 45 nm technologies with different leakage reduction mechanisms applied on a conventional cmos two input nand gate and is depicted in fig. 11. b. kalagadda, n. muthyala and k.k. korlapati 81 figure 10. inverter performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. table 2. two input nand gate performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. technology 180nm 120nm 45nm technique dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) subthreshold leakage power (µw) conventional cmos 12.01 2.921 3.661 0.737 0.121 0.030 stack 11.693 2.465 2.978 0.621 0.119 0.026 sleep 9.875 1.992 1.708 0.507 0.131 0.023 sleepy keeper 11.01 1.802 2.905 0.391 0.134 0.019 µ w performance comparison of digital circuits using subthreshold leakage power reduction techniques 82 figure 11. two input nand gate performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. table 3 compares the performance of a half adder in 180, 120 and 45 nm technologies with different leakage reduction mechanisms applied on a conventional cmos half adder and is depicted in fig. 12. table 3. half adder performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. technology 180nm 120nm 45nm technique dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) subthreshold leakage power (µw) dynamic power dissipation (µw) subthreshold leakage power (µw) conventional cmos 15.644 4.23 6.740 0.912 0.726 0.103 stack 14.232 3.87 5.359 0.851 0.699 0.101 sleep 12.047 2.789 4.820 0.702 0.796 0.092 sleep keeper 13.879 2.3 5.024 0.589 0.912 0.073 sleepy keeper µ w b. kalagadda, n. muthyala and k.k. korlapati 83 figure 12. half adder performance in 180, 120, 45nm technologies for different leakage reduction mechanisms. from tables table 1-3, it can be observed that the sub-threshold leakage power has been reduced with sleepy keeper technique compared with the other leakage reduction mechanisms for the digital cmos circuits like inverter, a two input nand gate and a half adder. but the dynamic power dissipation associated with that technique is increased due to additional high threshold voltage transistors. however, when static dissipation reduction is required according to the application, the sleepy keeper technique is preferred over other techniques. in deep sub-micron and nanoscale technologies, this technique can be utilized for designing digital circuits with improvement in the overall static power dissipation, which is mainly due to subthreshold leakage power dissipation. as the technology node is reduced to 45nm technology, the power dissipation has reduced when compared to other technologies like 120, 180 nm. the sleepy keeper technique is thus advantageous in static power reduction along with preserving the state. 5. conclusion the performance of inverter, two-input nand gate, and half adder digital circuits were studied for sub-threshold power leakage. even though each and every leakage reduction mechanism has its own advantages and disadvantages, choosing a leakage reduction mechanism depends on the type of circuit and end user application. in this paper, the sleepy keeper approach was found to be the best leakage reduction mechanism for several digital circuits compared to stack and sleep approaches. the sleepy keeper approach resulted in a sub-threshold power leakage of 38%, 42.2%, and 61% for 180nm, 120nm, and 45nm technologies, respectively, for inverters. for a two-input nand gate, the reduction in subthreshold leakage current was 38.3%, 46.9%, and 36.6% for 180nm, 120nm, and 45nm technologies, respectively. the sleepy keeper approach resulted in a sub-threshold leakage reduction of 45.6%, 35.4%, and 29.1% for 180 nm, 120 nm, and 45 nm technologies, respectively, for half adders. references boray s, deepak s, nunez a (2007), analysis of subthreshold leakage reduction in cmos digital circuits. 50th midwest symposium on circuits and systems 1400-1404. borkar s (1999), design challenges of technology scaling. ieee micro journal 19(4): 23-29. bursky d (1995), power reduction schemes promise cool digital ics. electronic design journal 43(1): 51. sleepy keeper µ w performance comparison of digital circuits using subthreshold leakage power reduction techniques 82 johnson mc, somasekhar d, roy k (2002), leakage control with efficient use of transistor stacks in single threshold cmos. ieee transactions on very large scale integration (vlsi) systems 10(1). manish k, md anwar hussain, sajal kp (2012), performance of a two input nand gate using subthreshold leakage control techniques. journal of electron devices 14: 1161-1169. roy k, mukhopadhyay s, mahmoodi mh (2003), leakage current mechanisms and leakage reduction techniques in deep sub micrometer cmos circuits. proceedings of the ieee 91(2): 305-327. saini p, mishra r (2012), leakage power reduction in cmos circuits. international journal of computer applications 55(8): 42-48. se hk, mooney vj (2006), sleepy keeper: a new approach to low-leakage power vlsi design. international conference on very large scale integration 367-372. 84 the journal of engineering research (tjer), vol. 19, no. 1, (2022) 54-62 *corresponding author’s e-mail: t.etri1@squ.edu.om doi:10.53540/tjer.vol19iss1pp54-62 a tidal flow model of the western coast of libya talal etri 1, ahmed f. abugdera 2, and abdelrahman elata 3 1 department of civil and architectural engineering, sultan qaboos university, po box 33 al-khoudh, muscat 123, oman, oman, sultanate of oman 2 civil engineering department, school of engineering and technology, aljafara university, tripoli, libya 3 engineering college, libyan academy, tripoli, libya abstract: this paper presents the hydrodynamics on the western coast of libya. the investigated area, which is a part of the mediterranean sea, is one of the most critical and active coastal regions in the country. a 2dh process-based model for flow based on the delft3d modelling system from deltares is constructed for the study area. extensive field data concerning the tidal constituents were used. the flow model that is necessary to understand the hydrodynamics of the area was calibrated and validated using field measurements. in this paper, only the water levels and tidal components for the astronomical tide are presented. calibration and validation of the numerical flow model show that the results of the water level represent the field conditions well. the present study gives insight into the basic hydrodynamic processes of the investigated area. it should help designers and the decision-makers maintain the region for any other economic and social activities. the flow model for the investigated area can be also coupled with any other models like wave, sediment transport, morphodynamic and water quality. keywords: western coast libya, hydrodynamics, astronomical tide, and process-based model. لیبیا -للتیارات المائیة للساحل الغربي نموذج رقمي عبد الرحمن عثمان العطا حمزة و أحمد فرج قدیرة ي والعترطالل عاشور الساحل الغربي للیبیا. حیث تعد منطقة الدراسة جزء من البحر األبیض المتوسط وواحدة ھایدرودینامیكیةتعرض ھذه الورقة :الملخص لمنطقة الدراسة لدراسة التیارات (2dh)من أكثر المناطق الساحلیة نشاًطا في البالد. تم إنشاء نموذج للمحاكاة الرقمیة ثنائي االبعاد واسعة النطاق تتعلق بمكونات المد والجزر للتیارات میدانیة . تم استخدام قیاسات deltaresمن delft3dالمائیة باستخدام نظام النمذجة القیاسات المیدانیة. في ب بمقارنتھاوالتحقق من صحتھ الرقمينموذج اللفھم ھذه التیارات المائیة في منطقة الدراسة تم معایرة والمائیة. نموذج أن نتائج ال. تُظھر معایرة بدون االخذ في االعتبار تأثیر الریاحمستویات المیاه ومكونات المد والجزر فقط دراسةتم الدراسة،ھذه أكثر تعطي نظرة أنھا الدراسة ھذه استخدامات نتائجومن أھم . مع القیاسات المیدانیة بشكل جیدالمتحصل علیھا متطابقة مستوى الماء أن یمكنھذه النتائج . والمتمثلة في التیارات المائیة للمد والجز على العملیات الھیدرودینامیكیة األساسیة للمنطقة التي تم فحصھا دقة یمكن أن یقترن ذلك كألي أنشطة اقتصادیة واجتماعیة أخرى. في حالة التوسع تساعد المصممین وصناع القرار في الحفاظ على المنطقة لدراسة تأثیر التیارات المائیة الناتجة عن األمواج أو تلك التي تسبب في خرى األمحاكاة النماذج من بأي المستخدم التدفق الرقمينموذج .في منطقة الدراسة جودة المیاهالى دراسة مدى باإلضافة وتغییر طبوغرافیة قاع البحرقل الرواسب ن .التیارات المائیة؛ تیارات المد والجز؛ نماذج المحاكاة الرقمیة الحدیثة ھایدرودینامیكیة؛ ؛الساحل الغربي للیبیا المفتاحیة:الكلمات a tidal flow model of the western coast of libya 55 1. introduction the libyan coast is of great importance to the population. approximately 90% of the population lives in an area of only 10% of the total area of the country (world bank, 2012). most of the people are living on the mediterranean coast, providing the use of many commercial, industrial and tourist activities. tripoli, as the capital of libya and is located on the western coast, represents an important research area. this considers the activities built on it from the port of tripoli, sea and many of the investment, which is the destination of the local population. also added facilities such as conference centre, luxury hotels and a cruise station. where the area was reclaimed from the coast of andalusia to establish the marina within the luxury tourist facilities. the small harbour is located near outlet pipes of wastewater and the fall of a valley. moreover, heavy traffic and pollution close to the sea give extra importance to the study area. all these factors make the modelling of the tide as a primary database an essential work. it is also critical to identify some biological and ecological aspects in the region. therefore, the tripoli coast is extremely important to both the economy and the ecology of the region. at the same time, the model, which is based on free source code delft3d from deltares, and freely available data sets, will support researchers, scientists and institutions to understand the coastal processes. this kind of numerical model would give a good understanding not only of the hydrodynamics of the region but also of the environment. this will lead to help in making decisions that will affect the ecology of the study area. to achieve a good understanding of the biological and water quality in the region, a very well-calibrated and validated hydrodynamic model is required. this includes the information on the velocity, water level, and fluxes for each grid cell within the domain under discussion. then the quantitative analysis of the water quality would be easy to perform by the researchers. it also provides an understanding of the behaviour of the environment, including determining the fate of different compounds. in this paper, a hydrodynamic flow model using delft3d software developed by deltares in delft, the netherlands is set up. the modelling used the tidal movement in tripoli station. moreover, the modelling will be evaluated and calibrated for a one-year evaluation. however, the grid was conducted and distributed by 14 monitoring stations. the tripoli tidal station was with high concern in the study area. the data has taken from delft dashboard (ddb), which provided the open-source project for a network data access protocol. also, the topography of the bed and the open boundary is defined by the general bathymetric chart of the ocean from the general bathymetric charts of the ocean (gebco 08 grid). the hydrodynamic flow model was validated, calibrated and evaluated with observations, which is reliable and useful for further model studies. the goal of this paper is to develop a highresolution model that can describe the hydrodynamics of the tripoli marina located in western coastal libya using delft3d. this model should be able to predict the tidal and water level as appropriately as possible. it should also be able to deal with wet and dry computations. in other words, the model can handle the wet/drylands concerning the water depth. 2. area of study the andalusia marina is a luxury hotel and cruise station. it is situated in tripoli on the west coast of libya. the study area is a part and is located at the southern of the mediterranean sea (fig. 1). the mediterranean sea is a semi-enclosed sea that leads to high salinities, temperatures, and densities. the only connection to the atlantic ocean in the west is through about 13km wide gibraltar strait. the other connection in the northeast is through the dardanelles to the sea of marmara and black sea. the rivers from the northern shores are bringing about 92% (15000 m3/s) of the input water (bryden et al., 1994). on the southern shore, the nile river drains water where most of the water evaporates. on the other hand, the drainage basin of the southern-eastern mediterranean, including a big part of the libyan coast, is mostly desert with low water inputs (ludwig et al. 2009). this amount of water is about one-third of the amount of water that it loses by evaporation. the water balance will have approximately 3250 km3/year losses but also the atlantic waters (bryden et al., 1994). this also will not have a significant effect on the hydrology of the study area. close to the study area, the exchange will be limited due to the shallow sicily stairs which has a depth of about 500m (crise et al., 1999). in the study area (arid and semi-arid) is common to have floods due to lasting rainfall during spring and fall, affecting small coastal catchments (unep/map/med pol, 2003). 3. purpose and methodology of the present study studies of hydrodynamics in the western libyan coast so far have been exploratory and detailed study is required for the functioning of the andalusia marina. this paper presents a 2dh model for flow based on delft3d software developed by deltares. figure 1. area of study. talal etri, ahmed f. abugdera, and abdelrahman elata 56 the flow model was used to compute current velocities, water levels at several locations on the western coast of libya. the flow model was set up for astronomical conditions, calibrated and validated using relevant tide gauges’ measurements at different locations in the area. this paper confines itself to the presentation of results relevant to the flow field in the area of investigation. the flow model when it is completed should be useful to set up sediment transport, morphodynamic, and water quality models in arriving at decisions pertaining to the andalusia marina operation. 4. delft3d flow model the use of the flow model is to simulate the multidimensional hydrodynamic flow and transport phenomena, including sediments (deltares, 2018). delft3d flow model solves the two-dimensional (depth-averaged or 2dh) or three-dimensional (3d) unsteady shallow water equations by applying the hydrostatic pressure assumption. transport and deposition of sediments are computed simultaneously with the hydrodynamics, creating direct feedback between hydroand morphodynamics (deltares, 2018). delft3d flow model calculates non-steady flow resulting from tidal and meteorological forcing. the main purpose is the two-dimensional (2dh, depthaveraged) and three-dimensional (3d) simulation of tidal and wind-driven flow by solving the unsteady shallow water equations. usually, the purpose of the model affects the choice of the grid and the bathymetry resolutions, and the period of the simulation. the model designed in this research is to link data and model in real-time. the delft3d modelling system is designed to simulate wind shear, wave forces, tidal forces, density-driven flows and stratification due to salinity and temperature gradients and atmospheric pressure changes in coastal, river and estuarine areas (lesser g.r., et al., 2004). delft3d flow model is suited with a flooding and drying algorithm. grid cells are activated when water levels exceed a flooding threshold, while grid cells are de-activated when local water levels drop below half this threshold (deltares, 2018). the flow model equations could be normally solved on a cartesian or spherical staggered grid. it is using the generalized lagrangian mean (glm) in the same way as the eulerian equations (lesser g.r., et al., 2004). lesser g.r., et al., 2004 described the governing equations of the flow model as the following: 𝑈𝑈 = 𝑢𝑢 + 𝑢𝑢𝑠𝑠 𝑉𝑉 = 𝑣𝑣 + 𝑣𝑣𝑠𝑠 � (1) where u and v are glm velocity components, u and v are eulerian velocity components and we and vs are the stokes’ drift components (lesser g.r., et al., 2004). the hydrostatic pressure is based on the assumption of shallow water equations where the vertical momentum equation is expressed only with the hydrostatic pressure. at the same time, the vertical component for the acceleration will be neglected. therefore, the final description for the hydrostatic pressure will be 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = −𝜌𝜌𝜌𝜌ℎ (2) where σ is the vertical layer thickness as a ratio of total depth, his the total depth, and 𝜌𝜌 is the local fluid density (including salinity, temperature and sediment). the horizontal momentum equations can be described in two equations 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝑈𝑈 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝑣𝑣 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝜔𝜔 ℎ 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 − 𝑓𝑓𝑉𝑉 = − 1 𝜌𝜌𝑜𝑜 𝑃𝑃𝜕𝜕 + 𝐹𝐹𝜕𝜕 + 𝑀𝑀𝜕𝜕 + 1 ℎ2 𝜕𝜕 𝜕𝜕𝜕𝜕 �𝑣𝑣𝑉𝑉 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 � 𝜕𝜕𝑉𝑉 𝜕𝜕𝜕𝜕 + 𝑈𝑈 𝜕𝜕𝑉𝑉 𝜕𝜕𝜕𝜕 + 𝑉𝑉 𝜕𝜕𝑉𝑉 𝜕𝜕𝜕𝜕 + 𝜔𝜔 ℎ 𝜕𝜕𝑉𝑉 𝜕𝜕𝜕𝜕 − 𝑓𝑓𝑈𝑈 = − 1 𝜌𝜌𝑜𝑜 𝑃𝑃𝜕𝜕 + 𝐹𝐹𝜕𝜕 + 𝑀𝑀𝜕𝜕 + 1 ℎ2 𝜕𝜕 𝜕𝜕𝜕𝜕 �𝑣𝑣𝑉𝑉 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 � ⎭ ⎪⎪ ⎬ ⎪⎪ ⎫ (3) where 𝜔𝜔 is the vertical velocity, f is coriolis parameter, 𝑣𝑣𝜕𝜕 and 𝑣𝑣𝑉𝑉 are the eulerian velocity components in cartesian coordinates (x and y), 𝑀𝑀𝜕𝜕 and 𝑀𝑀𝜕𝜕 are the external sources or sinks of momentum, and the horizontal pressure px and py are given by using boussinesq approximations 1 𝜌𝜌𝑜𝑜 𝑃𝑃𝜕𝜕 = 𝜌𝜌 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝜌𝜌 ℎ 𝜕𝜕𝜌𝜌𝑜𝑜 ∫ � 𝜕𝜕𝜌𝜌 𝜕𝜕𝜕𝜕 + 𝜕𝜕𝜕𝜕 ′ 𝜕𝜕𝜕𝜕 + 𝜕𝜕𝜌𝜌 𝜕𝜕𝜕𝜕′ � 𝑑𝑑𝜎𝜎′ 0 𝜕𝜕 1 𝜌𝜌𝑜𝑜 𝑃𝑃𝜕𝜕 = 𝜌𝜌 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝜌𝜌 ℎ 𝜕𝜕𝜌𝜌𝑜𝑜 ∫ � 𝜕𝜕𝜌𝜌 𝜕𝜕𝜕𝜕 + 𝜕𝜕𝜕𝜕 ′ 𝜕𝜕𝜕𝜕 + 𝜕𝜕𝜌𝜌 𝜕𝜕𝜕𝜕′ � 𝑑𝑑𝜎𝜎′ 0 𝜕𝜕 � (4) where 𝜌𝜌𝑜𝑜 is the reference density of water, 𝜎𝜎′ is the scaled vertical coordinate, and the horizontal reynold’s stress fx and fy are determined by the eddy viscosity concept with simplification, as shown in eqn. (5). 𝐹𝐹𝜕𝜕 = 𝑣𝑣𝐻𝐻 � 𝜕𝜕2𝜕𝜕 𝜕𝜕𝜕𝜕2 + 𝜕𝜕 2𝜕𝜕 𝜕𝜕𝜕𝜕2 � 𝐹𝐹𝜕𝜕 = 𝑣𝑣𝐻𝐻 � 𝜕𝜕2𝑉𝑉 𝜕𝜕𝜕𝜕2 + 𝜕𝜕 2𝑉𝑉 𝜕𝜕𝜕𝜕2 � � (5) mx and my are describing any external sources or subside of momentum like external forces due to hydraulic structures, discharge or withdrawal of water waves stress, etc. the continuity equation will be defined as depthaveraged in eqn. (6). 𝑆𝑆 = 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 + 𝜕𝜕 [ℎ𝜕𝜕�] 𝜕𝜕𝜕𝜕 + 𝜕𝜕 [ℎ𝑉𝑉�] 𝜕𝜕𝜕𝜕 (6) where s is the discharge or withdrawal of water, evaporation and precipitation per unit area, 𝜁𝜁 is the water level above some horizontal plane, and t is the a tidal flow model of the western coast of libya 57 time. for details and verification of the described equation, it is better to refer to the paper from walstra et. al., 2000. to solve the continuity and momentum equations for the flow model from delft3d, the alternating direction implicit (adi) is used (leendertse, 1987). the most important advantage form the adi method is that a system of equations with a small bandwidth will be developed. many improvements have been done by stelling and leendertse, 1991 makes the method used to solve the equations computationally efficiently. this stability can be seen at a courant number of up to about ten and most the second-order accurate (lesser g.r., et al., 2004). 5. flow model set up the domain definition is the first and the most important step for the model setup. it should be taken into account the area of interest and the location of the open sea boundaries. on the other hand, the computational requirements in terms of stability should be fulfilled. figure 1 shows the entire western libyan coast and also the area of interest. the open sea boundaries have been selected away from the area of interest to ensure that the hydrodynamics is well captured. moreover, any fluctuations will not affect the area of interest. to set up the flow model, the tool delft dashboard (ddb) from deltares is used. the ddb is a matlab based tool and several open-source datasets are embedded. these datasets include tidal information in many locations around the world (including the area of interest). it is also including the general bathymetric chart of the ocean (gebco) bathymetric dataset. this dataset is controlled by the support of the international hydrographic organization (iho) and the intergovernmental oceanographic commission (ioc) of unesco (van ormondt et al., 2020). the grid for the domain is generated also using the mentioned ddb tool. the numerical model domain, as it is shown in figure 2, is about 472976 km2. it consists of a rectangular grid with about 564 km cross-shore by about 1204 km longshore at the western coast of libya. the upper boundary is located just south of italy and the lower boundary is west of tunisia from the study area in libya. the grid spacing is about 4 km, in both directions, with about 32164 total number of active cells (301×141 cells total cells). the bathymetry, that has been used for the flow model was generated using the ddb tool with gebco bathymetric dataset. the maximum depth is about 4000 m close to the north open boundary (the deepest point in the mediterranean sea is about 5267 m) and about 500 m in the area of interest. two main open boundaries have been defined for the flow model. the west open boundary is located between the southern part of italian and the tunisian coast (west boundary). the northern open boundary is located between southern italy and the eastern libyan coast (northboundary) (fig. 2). thirteen astronomical components, from the ddb tool, have been applied along the open boundaries using the tpxo tide model. the tpxo models include complex amplitudes of msl relative sea surface elevations and transports/currents for eight primaries (m2, s2, n2, k2, k1, o1, p1 and q1), two long periods (mf and mm) and 3 non-linear (m4, ms4 and mn4) harmonic constituents (egbert and erofeeva, 2002). where m2 is the main semi-diurnal component, s2 is the main solar semidiurnal component, n2 is the lunar component due to monthly variation in moon’s distance from the earth, k2 is the soil-lunar constituent due to changes in declination of sun and moon throughout their orbital cycle, k1 is the soillunar component, o1 is the main lunar diurnal component, p1 is the main solar diurnal component, q1 is the larger lunar elliptic component, mf is moon’s biweekly component, mm is lunar monthly component, m4, ms4 and mn4 are the non-linear tidal components described as functions of m2, s2, and n2. all the tidal components along the open boundaries have been modified, the amplitude and the phase, to improve the flow model performance (more details will be described in detail in the calibration section). 6. evaluating the performance of the flow model any numerical model requires several sequential input parameters tuning steps (numerical and physical). these steps should lead to an acceptable and trusted model. to ensure a good model performance, sensitivity, validation and calibration processes should be taken place. the sensitivity studies for the flow model is dealing with input parameters and their degree of effect on the model results. any tuned parameters will not be included in the evaluation concerning the model quality. the importance of this stage is in the definition of the parameters to be calibrated. figure 2. domain, open boundaries, and observation points. talal etri, ahmed f. abugdera, and abdelrahman elata 58 the calibration procedure includes the defined parameters during the sensitivity analysis. these parameters will be tuned again within the physical limits to get a better agreement between the model results and measured data. the last step of the model evaluation is the validation for the model concerning different data sets in terms of time. this includes longer and different simulation periods but without changing any of the calibrated parameters. to evaluate the model performance properly, 14 observation stations will be used during the sensitivity, calibration, and validation processes (fig. 2). for the sensitivity analysis, no model quality will be investigated but only the water levels as time series will be considered. from the above-described evaluation performance steps, the grid spacing, time step, bottom roughness parameters have been investigated for the sensitivity study of the one month (spring and neap tide) during june 2016. this period has been used to avoid the effect of the wave due to wind (calm wind conditions, no storms no high waves and no wave-induced currents). three grid sets, 2×2 km, 4×4 km, and 6×6 km (fine, medium, and coarse) have been investigated. the model results for the three conditions do not show major differences in terms of water level. the only difference that has been noted is the computational time. the finer grid set makes the longer computational time. therefore, the medium grid set will be the optimal choice. it will have reasonable computing time and a good representation of the domain and observation stations. five-time steps 5, 10, 20, 60 and 120 minutes have been tested. all of them showed small differences in the water level results with longer computation time for the 5 minutes time step and less for the coarse one. in this study, a 10 minutes time step will be used to have reasonable computation time with fewer numerical errors. the bottom roughness for the flow model has been taken into account using chezy formula, uniform across the domain. also, three sets of uniform values of the bottom roughness (55, 65, and 70) are investigated. the results showed no effect would be observed with changing the bottom roughness. therefore, the value set 65 (default from deltars flow model) is used. table 1 concludes the final setting for the model sensitivity study, which will be used during the calibration and the validation procedures. the simulation period for the calibration has also been chosen like the sensitivity analysis. the sensitivity analysis showed that the most significant parameter, which should be taken in detail during the calibration study, is the open boundaries. therefore, the calibration procedure will only be focusing on the modification of the tidal components in terms of amplitude and phase. june 2016 will be considered for the calibration study (the comparison between the model results and the measurements). in this study, the quality of the flow model simulations was evaluated by using the mean square error (mse) and mean absolute error (mae), as shown in eqns. (7) and (8). the mae is more suitable and practical to evaluate the hydrodynamic models (sutherland et al., 2004). therefore, the mae will be used mainly in this paper to evaluate the model performance and the mse with the scatter diagrams will be used to assist the evaluations. 𝑀𝑀𝑀𝑀𝑀𝑀 = 〈|𝐻𝐻𝑐𝑐 − 𝐻𝐻𝑚𝑚|〉 (7) 𝑀𝑀𝑆𝑆𝑀𝑀 = 〈(𝐻𝐻𝑐𝑐 − 𝐻𝐻𝑚𝑚)2〉 (8) where, hc is computed (modelled) water level, and hm is measured water level. as a result of the sensitivity analysis, the calibration analysis will focus on the modification of the open boundaries in terms of phase and amplitude. since it is difficult to adjust the two open boundaries at the same time, the modification by several attempts of both boundaries will take place. these modifications will be by using different percentages (increase or decrease) of the amplitude and the phase till the model results, and the measurements in most of the observation points will have a good agreement. to simplify the adjustment of the open boundaries, an increase and decrease in the phase with a different percentage for each tidal component (semi-diurnal and diurnal components only) are applied. table 2 shows a summary of the modified open boundaries. it can be seen that the adjustments for the amplitude (meter) were required only by adding 10 to 35% to the open boundaries. but less modification for the phase (degree) is required (between 6 to 17%). the simulations also showed that the modification of the semi-diurnal and diurnal tidal components have major effects on the simulated water level. this is because the tide in the area of interest is a semi-diurnal tide (m2, s2, n2 and k2 components). on the other hand, less influence from the remaining tidal components has been seen. table 1. flow model settings. parameter value description ∆t 10 computational time step (minute) ρw 1025 water density (kg/m3) ρair 1 air density (kg/m3) g 9.81 gravitational acceleration (m/s2) ν 1 horizontal eddy viscosity (m2/s) c 65 chezy roughness coefficient threshold depth 0.1 threshold depth for exposure and flooding (m) a tidal flow model of the western coast of libya 59 table 2. adjustment of the northern open boundary. component amplitude (%) phase (%) se m idi ur na l m2 +35 +9 s2 +35 +8 n2 +35 +9 k2 +35 +8 d iu rn al k1 +10 +11 o1 +10 +17 p1 +10 +11 q1 +10 +6 in this study, the evaluation for the model performances during the calibration period will be only for the time series, the model qualification using mae and mse (eqns. (7) and (8)), and the scatter diagrams for water levels from the model results and the measurements. the comparison between the modelled and measured tidal components (semidiurnal and diurnal) will be inefficient for one month (too short). the model results for the 14 observation stations (fig. 2) showed a very good agreement with the measurements in most of the stations. table 3 shows the quality of the model results using the mse and mae qualifications described before using eqns. (7) and (8). the results from the modelled and the measured water level in most of the observation stations showed good quality. but in some observation points (adjim, gabs, and humtsuk) some differences in the results between the modelled and the measured water levels can be seen. this could be due to the location of the observation points, where the three stations are located in a protected and relatively shallow area (1.4 to 5.6m water depth). on the other hand, the observation points close to the open boundaries showed very good agreements (like capopassero and mazzaradelvallo). the results showed for these observation points mae is in the order of a few millimetres. moreover, the results in the area of interest (tripoli station) showed clear excellent model results compared to the measurements in terms of water level, where the mae is about 11mm mae and no phase lag. in this paper, only four observation stations will be shown including tripoli station and three close stations (sfax, el abassia and turgoenssmarsa). figure 3 shows the comparison between the modelled and measured water levels in the four stations and fig. 4 shows the scatter diagrams for the same stations. from both of the mentioned figures, it can be seen that tripoli station showed very good agreements without phase lag with mae about 11 mm and mse almost zero (fig. 3 and table 3). at the same time for the same station, the scatter diagram showed a high to the perfect correlation between the modelled and the measured water levels (fig. 4). the same conclusion can be conducted for the observation point turgoenessmarsa (185 km west from tripoli station) where the mae and mse are about 27 mm and 0.0008 m2, respectively. sfax and el abassia stations are located about 250 and 220 km west of tripoli station, respectively. both of the stations showed fewer results in an agreement between the modelled and the measured water levels, but still acceptable for the mae 94 mm for sfax and 58 mm for el abassia. from the scatter diagram, the sfax observation station showed some deviation in the model results for high and low tides. about 38% less deviation at the el abassia observation station have been seen. moreover, the deviation in el abassia station has been observed in the low tide more than in the high tide. the possible reason for these observed results in sfax and el abassia could be due to the locations of the two stations where both of them are located in a protected coastal region. this reveals that the more open sea stations have fewer errors to the more protected stations. the results for the stations close to the open boundaries, as is expected, have very good agreements between the model and the measurements. in some of them, the calculated mae and mse were about 1mm and zero, respectively (table 3 and figure 2). this leads to that the model is well-calibrated and no more improvements are required. table 3. quality of the flow model during the calibration period. station no. station mae (m) mse (m2) 1 tripoli 0.011 0.0001 2 sfax 0.094 0.0089 3 el abassia 0.057 0.0033 4 turgoenessmarsa 0.027 0.0008 5 valetta harbour 0.002 0.0000 6 adjim 0.284 0.0808 7 zarzis 0.030 0.0009 8 gabs 0.127 0.0162 9 humtsuk 0.104 0.0110 10 marsala 0.025 0.0006 11 mersaelbrega 0.013 0.0002 12 porto empedocle 0.014 0.0002 13 capopassero 0.001 0.0000 14 mazzaradelvallo 0.008 0.0001 talal etri, ahmed f. abugdera, and abdelrahman elata 60 figure 3. modelled vs measured water level. figure 4. scatter diagram for modelled and measured water level. to understand the possible errors in the model results due to the effect of the spring-neap tides for different months in one year, simulation for one year (2016) has been taken place. moreover, a longer simulation period will minimize the effect of the wind on the model results. these simulations have been conducted using the well-calibrated model. this process is defined as the validation of the model. since the simulation is for one year, the evaluation of the model will be for measured and the modelled tidal constituents (diurnal and semi-diurnal components) in tripoli station only (area of interest). therefore, the time series evaluation using mae or mse will not be applicable. the tidal components for the model results have been generated using the tool tide from (deltares systems) where this tool is a major part of the delft3d from deltares. figure 5 shows the results for the validated model for 2016 (january till december) in terms of tidal components at the tripoli observation point. the agreement between the modelled and the measured tidal components are very good for both the amplitude and the phase. this good agreement can be seen for the semi-diurnal tidal constituents (m2, s2, n2 and k2) for both tide amplitude and phase. the differences in the amplitude for measured and modelled semi-diurnal components were in the order of 1cm for one-year model results. also, the phase results showed the same conclusion, where the differences were in the order of 2 to 4 degrees except n2, the lunar component due to monthly variation in the moon’s distance from the earth tidal component (20 degrees). but it is still in the acceptable range. the results, for the semidiurnal tidal constituents, clearly show that the considered calibration period is much better than the diurnal tidal constituents. a tidal flow model of the western coast of libya 61 figure 5. modelled and measured tidal components for tripoli station (the year 2016). this can be seen in the same figure for the diurnal tidal constituents (k1, o1, p1 and q1) without omission the fact that the tide in the area of interest is semidiurnal tide. although of these results, the model showed generally good agreement between the measured and the modelled results. but for some tidal components, the model results showed slightly overestimated amplitude and phase compared to the measurements. on the other hand, the overestimated results will not have a significant effect on the semidiurnal tidal components. 7. conclusion the 2dh flow model of delft3d from deltares was set up in conjunction with field measurements for tidal constituents on the western coast of libya. the model simulations account for the tides to compute hydrodynamics in the area of interest. sensitivity and calibration studies show that the effect of wind during the calm periods on the hydrodynamics in the area of interest is not very important. where the results for the modelled water level were very close and identical in some locations for calm periods (june 2016). results of the validation of the model over a longer period (the year 2016) showed that the flow model is capable of predicting the hydrodynamics agreeably with the observations. some adjustments to the derived tidal components were found necessary to account for calm meteorological conditions. the results of the simulations serve to improve our understating of the underlying physical processes in the area of interest due to tides. although a good understanding of hydrodynamics was achieved, the application of the model to the wind from different years and different seasons may help to improve our understanding of storm effects on the hydrodynamic and morphological evolution. although it was found that the 2dh approximation seems to be appropriate for describing the water level, comparisons of the results for the current velocities with those using a 3d approximation and field measurements are recommended. acknowledgment the work presented in this paper is a result of a cooperation between sultan qaboos university, oman, aljafara university and the libyan academy, libya. the authors gratefully acknowledge the continuous support from all contributors. conflict of interest the authors declare that there are no conflicts of interest regarding this publication. funding no funding was received for this study. rferences bryden h. l., candela j., and kinder t. h., (1994): exchange through the strait of gibraltar. progress in oceanography, (33), 201–248. crise a., allen j.i., baretta j., crispi g., mosetti r., and solidoro c., (1999). the mediterranean pelagic ecosystem response to physical forcing. progress in oceanography, (44), 219-243. deltares, delft3d-flow simulation of multidimensional hydrodynamic flows and transport phenomena, including sediments, (2018), user manual. egbert g. d. and erofeeva s. y., (2002), efficient inverse modeling of barotropic ocean tides. journal of atmospheric and oceanic technology, (19), 183– 204. leendertse j.j., (1987). a three-dimensional alternating direction implicit model with iterative fourth-order dissipative nonlinear advection terms. wd-333-neth, the netherlands rijkswaterstaat. lesser g.r., roelvink j.a., van kester j.a.t.m. and stelling, g.s., (2004), development and validation of a three-dimensional morphological model, coastal engineering, (51), 883-915. ludwig w., dumont e., meybeck m., and talal etri, ahmed f. abugdera, and abdelrahman elata 62 heussner s., (2009): river discharges of water and nutrients to the mediterranean and the black sea: major drivers for ecosystem changes during past and future decades. progress in oceanography, (80), 199– 217. stelling g.s., leendertse j.j., (1991). approximation of convective processes by cyclic aoi methods. proceeding of the 2nd asce conference on estuarine and coastal modelling, tampa. asce, new york, 771– 782. sutherland j., peet a. h., and soulsby r.l., (2004), evaluating the performance of morphological models, coastal engineering, (51), 917-939. unep/map/med pol, (2003). riverine transport of water, sediments and pollutants to the mediterranean sea. united nations environment programme, mediterranean action plan, athens. van ormondt m., nederhoff k., and van dongeren a., (2020), delft dashboard: a quick set-up tool for hydrodynamic models. journal of hydroinformatics, (22), 510-527. walstra d.j.r., roelvink j.a., and groeneweg j., (2000). calculation of wave-driven currents in a 3d mean flow model. edge, b. coastal engineering, (2). asce, new york, 1050– 1063. world bank (2012), the world bank annual report 2012, (1), main report, washington, dc, world bank. طلال عاشور العتري و أحمد فرج قديرة و عبد الرحمن عثمان العطا حمزة 1. introduction 2. area of study 3. purpose and methodology of the present study 4. delft3d flow model 5. flow model set up 6. evaluating the performance of the flow model 7. conclusion acknowledgment conflict of interest rferences the journal of engineering research (tjer), vol. 15, no. 2 (2018) 114-123 doi: 10.24200/tjer.vol15iss2p114-123 evaluation of actual creep length under heading up structures aprons m.a. el mollaa, n.y. saadb,*, and g.s. ezizahb a civil engineering department, higher technological institute, cairo, egypt. b irrigation and hydraulics department, faculty of engineering, ain shams university, cairo, egypt. received 13 january 2016; accepted 26 november 2017 abstract : in this paper a 2d finite element model (seep2d) is implemented to study the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure. different scenarios for the thickness of the pervious layer under the apron, the length of the apron and the depths of the upstream and downstream sheet piles are studied. results show that assuming the outer and inner faces of the sheet piles have the same weight for estimating the creep length while designing aprons of hydraulic structures is weak. design equations for the actual head loss along the outer and inner faces for both the upstream and downstream sheet piles are driven. these equations can be used as a tool in the practical design for aprons of heading up structures formed on pervious soil and provided with upstream and downstream sheet piles at its ends. keywords: seep2d; seepage; heading-up structures; sheet piles; creep length. اهلياكل فوق مئزرتقييم طول الزحف الفعلي ب، غادة س عزيزة ب،*يوسف سعد .نيفني ،أحممد علي املوىل لدراسة الفقدان الفعلي للقمة (seep2d) احملدود الطبقة التحتيةيتم تطبيق منوذج لعنصر البحثية،يف هذه الورقة :امللخص سيناريوهات خمتلفة لسمك الطبقة دراسة ويتمالعلوية اهلياكلاملتواجدة على طول أكوام االطبقات املثبتة يف نهايات مئزر النتائج أن االفرتاض بأن روتظه .والسفليةوطول املئزر وكذلك أعماق أكوام الطبقات العلوية املئزر،القابلة لالخرتاق حتت طول الزحف املقدر عندما يكون تصميم مآزر اهلياكل نفس وزناألوجه اخلارجية والداخلية ألكوام الطبقات هلا للقمة على طول األوجه اخلارجية والداخلية للخسارة الفعليةويتم عمل معادالت التصميم .ضعيفهو إفرتاض اهليدروليكية هذه املعادالت كأداة يف التصميم العملي ملآزر ترتكز على ستخداما والسفلية. وميكنلكل من أكوام الطبقات العلوية .نهاياتها العلوية والسفلية يفومزودة بأكوام من الطبقات لالخرتاق،هياكل متكونة على تربة قابلة .طول الزحف ،مرتاكمةطبقات ،هياكل علوية ،تسرب :الكلمات املفتاحية * corresponding author’s email: neveen_yousif@hotmail.com mailto:neveen_yousif@hotmail.com m.a. el molla, n.y. saad, and g.s. ezizah 115 notations the following symbols are used in this paper. d1 depth of upstream sheet pile (l); d2 depth of downstream sheet pile (l); ho1 head lost along the outer face of upstream sheet pile (l); ho2 head lost along the outer face of downstream sheet pile (l); hi1 head lost along the inner face of upstream sheet pile (l); hi2 head lost along the inner face of downstream sheet pile (l); g gravitational acceleration (l/t2); h head difference between upstream and downstream the apron (l); k hydraulic conductivity of the homogeneous pervious stratum (l/t); l horizontal distance between the upstream and downstream sheet piles (l); p1 head at point (1) (l); p2 head at point (2) (l); p3 head at point (3) (l); p4 head at point (4) (l); p5 head at point (5) (l); p6 head at point (6) (l); t thickness of pervious stratum under the apron (l); ρ density of seeping water (m/l3); evaluation of actual creep length under heading up structures aprons 116 1. introduction seepage is one of the most important factors that affects the stability of a heading-up structure. using sheet piles under the aprons of heading-up structures increases the percolation length, decreases the hydraulic gradient, and hence provides more safety against piping and uplift under the apron of a structure. figure (1) represents the concept of creep length and hydraulic gradient diagram. numerical methods (finite element, finite difference, boundary element and total variation diminishing method ) can be used to solve seepage problems with different degrees of complexities and accuracies harr (1962); serge leliavsky (1965); u.s. army corps of engineers (1986). ahmed and ellboudy (2010) carried out a number of numerical analysis to study the influence of different sheet pile configurations on the seepage losses, the uplift force on the apron, and the exit gradient at the end of the apron. they concluded that driving a sheet pile under a hydraulic structure that surrounds the downstream floor apron from all sides, has greatly reduced the exit gradient at the end toe of the floor. however, this was accompanied with some increase in the uplift force. ahmed (2011) studied different sheet piles configurations and flow through canal banks. kamble et al. (2014) described the potential of different approaches such as geological and geotechnical methods, dam instrumentation, geophysical methods, tracer techniques, nuclear logging and mathematical modeling for monitoring, detecting and analyzing seepage in hydraulic structures. three dimensional modeling techniques (3d) have recently been attempted to study seepage beneath and around hydraulic structures for different purposes. eftekhar and barani (2013) used seep3d to pinpoint the best locations of cutoff walls. they revealed that a cutoff wall installed at the lower toe of the structure eliminates piping and the one located at the upper toe reduces the uplift pressure. guerra et al. (2012) suggested that complex junctions in levee systems (i.e. 90 degrees bend in the levee alignment as well as a box culvert through the levee) should be studied using 3d models. koltuk and iyisan (2013) compared between the 2d and 3d modeling techniques in studying a rectangular shaped cofferdam taking into account the vertical anisotropy of the permea bility of soil layers. he concluded that the difference between the values obtained from 2d and 3d analysis decreased with increasing the length to width ratio of the cofferdam. gad et al. (2016) conclude that 3d simulations are essential in studying seepage under large dams especially in complicated 3d configurations and lateral heterogeneity. el tahan and elmolla (2013) used the electric analogue method to investigate models of hydraulic structures' aprons provided with two equal cutoffs. they investigated the efficiency of both faces (outer and inner) of these cutoffs. they concluded that the estimation of the percolation length under the hydraulic structure must consider a weighted length factor for various faces of the cut-offs. seep2d is a finite element program that has been applied in many researches to study seepage and has proved to be an efficient tool for seepage analysis (ozkan 2003; noori and ismaeel 2011; anas 2012; aboulatta et al. 2014). elmolla (2001) used seep2d to investigate seepage under the aprons of heading up structures provided with a single cut-off. in this paper, seep2d is used to investigate the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure formed on a pervious soil. equations of the relation between head loss along both faces (outer and inner) of upstream (u.s.) and downstream (d.s.) sheet piles and the involved variables will be driven to be used in creep length designing purpose which, to the best of the authors’ knowledge, has not been covered in the literature yet. figure 1. line of creep and hydraulic gradient diagram (after serge leliavsky, 1965). m.a. el molla, n.y. saad, and g.s. ezizah 117 in this study sensitivity analysis for the variables involved in the problem as well as different scenarios for the thickness of pervious layer under the apron, the length of the apron, the depths of the upstream and downstream sheet piles, and the hydraulic conductivity are studied. 2. dimensional analysis in the present study, all the variables involved in the problem can be expressed as (refer to fig. (2)): f (h, d1,d2,l,t,k,r,g,p1,p2,p3,p4, p5,p6 = 0 (1) where: h = head difference between upstream and downstream the apron; d1 & d2 = depth of upstream and downstream sheet pile respectively; l = horizontal distance between the upstream and downstream sheet piles; t = thickness of pervious stratum under the apron; k= hydraulic conductivity of the homogeneous pervious stratum under the apron; ρ= density of seeping water; g = gravitational acceleration; p1, p2, p3, p4, p5, p6 = head at points(1,2,3,4,5,6). applying buckingham's π theorem, the relationship between the above different variables can be presented as follow: 0) h h , h h , h*g k , d t , t l , d d ( 2i 2o 1i 1o 22 1  (2) where 32 21 1 1 pp pp h h i o    (3) and 54 65 2 2 pp pp h h i o    (4) ho1/ hi1 = the ratio between the head loss along the outer face and the inner face of the u.s sheet pile, and ho2/hi2 = the ratio between the head loss along the outer face and the inner face of the d.s sheet pile. 3. description of the model the seep2d software was developed by the united states army engineer waterways experiment station to model a variety of problems involving seepage. the governing equation used in this model is the laplace equation. laplace equation is the governing differential equation for two dimensional, steady state incompressible, isotropic flow in the xy plane which is 0 y h x h 2 2 2 2       (5) where h is the head at any point in the flow domain. transient or time varying problems and unconfined plan view (aerial ) models cannot be modelled using seep2d. seep2d allows for different hydraulic conductivities along the major and minor axes (anisoropic conditions) to be defined. heterogeneous models can be created by specifying different values of hydraulic conductivity for the elements representing the different layers or regions. post-processing includes contouring of the total head (equipotential lines), drawing flow vectors, and computing flow potential values at the nodes. these values can be used to plot flow lines together with the equipotential lines (i.e. flow nets). the phreatic surface can also be displayed (seep2d primer 1998). figure 2. the variables involved in the problem. evaluation of actual creep length under heading up structures aprons 118 in a typical modelling problem involving the seep2d software, a series of tasks are performed in a specific sequence as follows: 1. mesh generation 2. setting boundary conditions; 3. seep2d execution; and 4. postprocessing the output. 4. model application the dimensions for different variables as recommended from a previous study ezizah et al. (2000) are as follows: d1 = (0.2 to 0.6)*l, d2/d1 = (0.7 to 0.2) , h/l = (1/2, 1/3, 1/6, 1/8) , d1/t = (0.1 to 0.5) or t= 1.5l. the model consists of a heading up structure formed on a pervious homogeneous isotropic soil layer with thickness t (t=15, 22.5,30, 60 m). two values of head difference (h) between u.s and d.s the structure are considered (h=3, 7 m). the apron of the structure is of length l (l=10, 15, 20 m) and provided with two sheet piles at its ends driven to depths d1 and d2, where d1 is the depth of the u.s sheet pile (d1= 3,4,5,6 m) and d2 is the depth of the d.s sheet pile (d2=1.5, 2.5 m). figure (3) shows a finite element mesh used for one of the simulations. figure (4) shows a sample of flow net obtained from seep2d (seep2d output). 4.1 boundary conditions all external edges of the problem were modelled as impermeable boundaries including the upstream and downstream edges and the bottom of the problem. the apron of the hydraulic structure was also modelled as impermeable boundary. all upstream nodes of the structure in the canal bed were allocated a prescribed head (h) (h=3 or 7 m). nodes in the canal bed downstream of the structure were assigned a prescribed head value of zero. 4.2 calibration and verification of the model experimental readings of a previous study el tahan and elmolla (2013) were used to calibrate the model to choose a reasonable cell size and check that the input data were entering correctly. three different mesh cell sizes were figure 3. sample of seep2d mesh. figure 4. sample of flow net obtained from seep2d (seep2d output). m.a. el molla, n.y. saad, and g.s. ezizah 119 used in order to calibrate the model in an attempt to match the experimental data. then mesh would be generated as smaller mesh cell sizes were chosen around and between the two sheet piles in order to accurately show the changes in head when using sheet piles. the experimental study includes two equal sheet piles under the apron of the heading up structure. two values of h were used h= 3, 6 m, l was taken 15 and 20m while d1 and d2 were taken 5m. figure (5) illustrates a sample of the model calibration using the chosen best cell size, which represents the head readings at points p1 to p6. it is obvious that there is a good matching between the experimental readings and the numerical model. using the calibrated mesh size, the model is verified by another set of electric analogue experimental readings (aboulatta et al. 2010). the length of apron l= 19m, the depth of the u.s and d.s sheet pile d1= 5m, d2= 2.5m. two values for h (3, 7.5m) were used. figure (6) illustrates a sample of the model verification, which represents the head readings at points p1 to p6. it is obvious that there is a good matching between the experimental readings and the numerical model. 4.3 sensitivity analysis sensitivity analysis was carried out in order to find the most effective parameters for different variables. from the sensitivity analysis it is noticed that the variables which have the biggest effect on the head loss ratio along the upstream sheet pile are the thickness of pervious soil layer under the apron (t) and the depth of u.s sheet pile (d1). for the d.s sheet pile, it is noticed that the thickness of pervious soil layer under the apron (t) and the depth of downstream sheet pile (d2) have the biggest effect on the results. moreover, the head difference h and the hydraulic conductivity k have negligible effect on the ratio of the head lost along the faces for the both sheet piles. figure 5. sample of model calibration h= 6m, l= 20m, d1 & d2= 5m. figure 6. sample of model verification h= 7.5 m, l= 19 m, d1 = 5m, d2= 2.5m. evaluation of actual creep length under heading up structures aprons 120 figures (7) and (8) show the effect of each variable on the ratio between the head loss along the outer face and the inner face of u.s and d.s sheet piles respectively (ho/ hi). 5. results analysis the readings obtained from seep 2d are investigated and analyzed in order to determine the effect of changing the length of apron, the thickness of soil layer and the depths of upstream and downstream sheet piles on the ratio between the head loss along the outer and the inner faces for both the upstream and downstream sheet piles. plotting the head loss ratio along the outer face and the inner face for the u.s sheet pile ((ho1/ hi1) against the depths ratio of u.s and the d.s sheet pile (d1/d2) for different values of (t/d2) is shown in fig. (9). it is noticed that (ho1/ hi1) decreases with the increase of the ratio between (d1/d2) for all (t/d2) values. so, the efficiency of the inner face of the u.s sheet pile improves by 8-16% with increasing the ratio d1/d2. also, it is obvious that the ratio of (t/d2) is directly proportional to (ho1/ hi1). that means the efficiency of the inner face of u.s sheet pile improves by decreasing the values of t/d2. moreover, changing t/d2 affects ho1/ hi1 values for bigger values of d1/d2 than smaller ones. from fig. 9, it was found that the best relation between (ho1/hi1) and (d1/d2) is given by linear equation with correlation equal r (r measures the strength of the relationship between variables): figure 7. effect of each variable on the ratio (ho1/hi1). figure 8. effect of each variable on the ratio (ho2/hi2). m.a. el molla, n.y. saad, and g.s. ezizah 121 2 1 1 1 d d ba h h i o  (6) where a, b are coefficients depending on t/d2. the value of a and b are given in the table 1. so, the general empirical head loss equation along the u.s. sheet pile could be presented as follows: 2 1 2 2 1 21i 1o d d 184.0 d t 006.0 d d d t 005.018.2 h h   (7) plotting the ratio of the head loss along the outer and the inner face for the d.s. sheet pile ((ho2/ hi2) against the depths ratio of u.s. and the d.s. sheet pile (d1/d2) for different values of (l/t) is shown in (fig. 10). it is noticed that ((ho2/ hi2) decreases with the decrease of (d1/d2) for all values of (l/t). that means the efficiency of the inner face of the d.s. sheet pile improves by 1-3% with decreasing the ratio (d1/d2). moreover, comparing the ho2/ hi2 values at the same depth ratio values, we can conclude that as (l/t) values increase the ((ho2/ hi2) values decrease by 2-4%. from fig. 10, it was found that the best relation between (ho2/hi2) and (d1/d2) is given by linear equation with correlation equal r. figure 9. design chart for the ratio between the head lost along the outer and the inner faces of u.s sheet pile. figure 10. design chart for the ratio between the head lost along the outer and the inner faces of d.s sheet pile. table 1. the value of a and b constants. t/d2 6 9 10 12 15 20 a -0.16 -0.14 -0.13 -0.1 -0.11 -0.09 b 2.2 2.26 2.24 2.24 2.3 2.28 r 0.92 0.96 0.97 0.99 0.96 0.98 evaluation of actual creep length under heading up structures aprons 122 2 1 2 2 d d ec h h i o  (8) where c, e are coefficients depending on l/t. the value of c and e are given in the table (2). so, the general empirical head loss equation along the d.s sheet pile could be presented as follows: 2 1 2 1 2i 2o d d 016.0 t l 181.0 d d t l 027.0251.2 h h   (9) for the studied variables range, the ratio (ho1/ hi1) for the upstream sheet pile ranges from 1.55 to 2.15, and the ratio (ho2/hi2) for the downstream sheet pile ranges from 2.08 to 2.3. 6. conclusion seep2d is implemented to study the actual head loss along the sheet piles fixed at the ends of an apron of a heading up structure. the following main conclusions may be drawn:  for the u.s. sheet pile, the variables, which have the biggest effect on the head loss ratio along both of its faces, are the thickness of pervious soil layer (t) and the depth of u.s. sheet pile (d1).  for the d.s. sheet pile, the thickness of pervious soil layer (t), the apron length (l) and the depth of d.s. sheet pile (d2) have the biggest effect on the head loss ratio.  the outer and inner faces of the sheet piles do not have the same weight for estimating the creep length. actually, the creep length along the inner face of the u.s. pile is about 55% of its supposed value and for the d.s. pile is about 45% of its supposed value for most of the tests ranges.  increasing the sheet pile depth ratio (d1/d2=4) improves the efficiency of the inner face of the u.s. sheet pile by 816%.  the improvement of the d.s. sheet pile efficiency by decreasing d1/d2 or increasing l/t is small. so, it is better to increase the sheet pile depth ratio to improve the efficiency of the u.s. sheet pile, although it undermines the downstream sheet pile efficiency by small amount.  design equations for the actual head loss along the outer and inner faces for both the upstream and downstream sheet piles are driven. these equations may be used as a tool in practical design for aprons of heading up structures formed on pervious soil and provided with upstream and downstream sheet piles at the ends of the apron. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references aboulatta n, ezizah g, yousif n, anas d (2010), study of the efficiency of downstream blanket in heading-up structures. nile basin water science and engineering 3(1): 79-89. aboulatta n, ezizah g, el molla d (2014), modelling seepage effects in heterogeneous soil under heading-up structures using an experimental and numerical methodology. ph.d. thesis, faculty of engineering, ain shams university, cairo, egypt. ahmed aa, ellboudy am (2010), effect of sheet pile configuration on seepage beneath hydraulic structures. asce, international conference on scour and erosion (icse-5) san francisco, california, united states. ahmed aa (2011), design of hydraulic structures considering different sheet pile configurations and flow through canal banks. computers and geotechnics 38(4): 559-565. anas d, gad ma, samy g, aboulatta n (2012), a comparative analysis between electric and numerical seepage modeling techniques. the seventh conference on environmental hydrology, american society of civil engineers (asce), cairo, egypt. eftekhar, afzali s, barani ga (2013), application of the seep 3d software in m.a. el molla, n.y. saad , and g.s. ezizah 123 modeling the effects of cutoff walls on the reduction of the seepage pressure and exit hydraulic gradient under hydraulic structure. water engineering journal 5(15): 79-90. el molla am (2001), new trend for evaluating the percolation length under aprons of hydraulic structures provided with cut-off and founded on isotropic soil. cerm, faculty of engineering, al -azhar university 23(1). el tahan am, el-molla am (2013), effects of cut-offs depth ratio on uplift and efficiency of front and back faces of upstream cut-off. 21st csce canadian hydrotechnical conference, alberta, canada. ezizah gms, el-molla am, abdellateef m, and el-niazy hammad m (2000), seepage under hydraulic structures founded on complex formations. ph.d. thesis, faculty of engineering, ain shams university, cairo, egypt (this for ratios). gad ma, anas d, atta na, samy g (2016), worth of the 3d simulation of seepage in the vicinity of heading-up structures. ksce journal of civil engineering 20 (2): 679-686. guerra p, priestley d, fredlund m (2012), 3d groundwater seepage analysis of a levee intersection. canadian dam association (cda) annual conference. harr me (1962), groundwater and seepage. mcgrawhill, new york. kamble rk, muralidhar b, hanumanthappa ms, patil av, edlabadkar js (2014), multiple approaches to analyse and control seepage in hydraulic structures", ish journal of hydraulic engineering, 20(1): 7–13. koltuk s, iyisan r (2013), numerical analysis of groundwater flow through a rectangular cofferdam. electronic journal of geotechnical engineering (ejge), 18. noori bma, ismaeel khs (2011), evaluation of seepage and stability of duhok dam. alrafidain engineering journal 19(1). ozkan s (2003), analytical study on flood induced seepage under river levees. ph.d dissertation, louisiana state university, the department of civil and environmental engineering. seep2d primer (1998), brigham young university, engineering computer graphics laboratory. serge leliavsky (1965), design of dams for percolation and erosion. volume iii, oxford and ibh publishing co. u.s. army corps of engineers, washington dc (1986), seepage analysis and control for dams. engineer manual 1110-1901. microsoft word paper 1.docx the journal of engineering research (tjer), vol. 14, no. 1 (2017) 01-09 estimating drilling cost and duration using copulas dependencies models m. al kindi*, m. al-lawati and n. al-azri department of mechanical and industrial engineering, college of engineering, sultan qaboos university, oman. received 1 september 2015; accepted 8 june 2016 abstract: estimation of drilling budget and duration is a high-level challenge for oil and gas industry. this is due to the many uncertain activities in the drilling procedure such as material prices, overhead cost, inflation, oil prices, well type, and depth of drilling. therefore, it is essential to consider all these uncertain variables and the nature of relationships between them. this eventually leads into the minimization of the level of uncertainty and yet makes a "good" estimation points for budget and duration given the well type. in this paper, the copula probability theory is used in order to model the dependencies between cost/duration and mri (mechanical risk index). the mri is a mathematical computation, which relates various drilling factors such as: water depth, measured depth, true vertical depth in addition to mud weight and horizontal displacement. in general, the value of mri is utilized as an input for the drilling cost and duration estimations. therefore, modeling the uncertain dependencies between mri and both cost and duration using copulas is important. the cost and duration estimates for each well were extracted from the copula dependency model where research study simulate over 10,000 scenarios. these new estimates were later compared to the actual data in order to validate the performance of the procedure. most of the wells show moderate weak relationship of mri dependence, which means that the variation in these wells can be related to mri but to the extent that it is not the primary source. keywords: archimedean copula, monte carlo simulation, mechanical risk index (mri). אאאא kא*{אk{אאk wא  א א kא א   א א א  א  א א   א אא   א א،א א،א، ،א א  kא א   k א   א אא   א א    א    א  א    א   k   א  אאאאאאאfmrieאאk א   א א א  א   ، א،  אא א א   א א אא   k    א א   kאאאאאאאא  א  א א אcopulasא   אא    א א  k copula١٠٠٠٠אאאאאk  א א   א  א  kאא א א   א   אאאkאאא  אאw א،،אאfmriek  * corresponding author’s e-mail: kindim@squ.edu.om m. al-kindi, m. al-lawati and n. al-azri 2 1. introduction quantifying dependencies between random variables has been used recently in many finance and insurance risk analysis. copula dependency model is a superior technique to model dependencies and yet it has rarely been used in petroleum applications (al-harthy et. al. 2007). in general, copula is a function that links the marginal distributions of the random variables and generate the joint distribution given their dependence level. a copula can model the dependence structure given any type of marginal distribution, which is not possible with other correlation measures because it is able to separate the marginal distribution from the correlation. the two random variables are the mri, and cost and duration. cost and duration are two economically detrimental factors to be considered in decision making before venturing well drilling projects. developing reliable procedures and methods for estimating cost and time in well drilling has been an active field of research. early attempts started with regression analysis and adaptive methods (noerager et al. 1987; zoller et al. 2003). however, these basic methods overlooked the possibilities of the alternative scenarios that could happen and render the estimates far from reality. another very intuitive flaw with those models is their dependence on linear regression which is only valid for linearly dependent variables (aas 2006). hence, the linear correlation coefficient is a meaningful measure of dependence but may induce misleading conclusions. an attractive alternative method of capturing the dependence can be the use of copula theory which can separate the marginal distribution from the correlation. the use of copula theory in financial applications was introduced by embrechts et al. (2002). they introduced copula theory for two reasons: first, because it is a principal tool in indicating the drawbacks of linear regression; second, it is an approach for understanding the general concept of dependency. copula theory has also been very common in the oil and gas industry. chiyoshi (2004) applied archimedean copulas in order to estimate the time for drilling wells. archimedean copula has been used because it has many families that are capable of representing different types of dependencies. another research conducted by al-harthy et. al. (2007) found that the two common methods used to model the dependency in the petroleum industry i.e. envelop method and the imanconover method, failed to address the dependence structure. from this point of view, they illustrated and discussed the benefits of using archimedean copulas to model dependencies on the reserves problem. therefore, addressing the dependencies between two economical random variables with the mri value demonstrates a new approach to give managerial insights to decision makers. 2. copulas copula is a multivariate distribution whose one-dimensional margins are uniform on the interval [0, 1] (pradier 2011). the definition of a d-dimensional copula is a multivariate distribution, c, with uniformly distributed marginal u (0; 1) on [0, 1]. sklar’s theorem states that every multivariate distribution f with marginal , …… can be expressed in the form: ,…. . , , ,…, (1) the popular families of copulas are called archimedean copulas which is one of the two main families of copula along with the copulas of normal mixture distributions. archimedean copulas the archimedean copulas is defined as: , (2) for , ∈ [0,1], , is the copula function with and represents the uniform distributions, is the generator and is the inverse generator. the main input for the generator for the different families of archimedean copulas function is known as kendall’s tau correlation,  : (3) where is the sample size , is the number of pairs that are moving in the same direction and is the number of pairs that are moving in the opposite direction of each other. the considered archimedean copulas (clayton, gumbel, and frank copulas) fall into the class of the so-called laplace transform archimedean copulas. each family type of these estimating drilling cost and duration using copulas dependencies models 3 copulas has its own parameters. table 1 and table 2 show the values of these parameters. substituting the values of the generators in eqn. (2) will give a different copula equation for each family. clayton copula: the clayton copula is a systematic copula that exhibits greater dependence in the negative tail than it does in the positive one. this copula is mostly used to study correlated risks because of its ability in capturing lower tail dependence (mahfoud 2012). clayton copula is expressed as: ѳ , max ѳ ѳ 1 ѳ,0 (4) where ѳ,belongs to 1,∞ ∖ 0 , is a parameter that controls the dependence where zero is not included. perfect dependence is obtained when  tends to infinity. in contrast, when  approaches zero then independence is implied. the main drawback associated with this type of copulas is the fact that it cannot account for negative dependence. gumbel copula: the gumbel copula is also an asymmetric copula that exhibits greater dependence in the positive tail than it does in the negative one. this copula is expressed as: table 1. archimedean copulas with their generators and ranges. family generators ranges of the tau clayton (1978) ѳ 1) 1,∞ ∖ 0 gubmel (1960) ln ѳ 1,∞ frank (1979) -ln ѳ ѳ ) ∞,∞ ∖ 0 table 2. kendall tau values as a function of theta for each family and their ranges. family ranges tau clayton (1978) 1,∞ ∖ 0 ѳ ѳ 2 gubmel (1960) 1,∞ 1 1 ѳ frank (1979) ∞,∞ ∖ 0 1 4 ѳ 1 ѳ ѳ , exp ln ѳ ln ѳ ѳ (5) where 1≤ѳ <∞ is a parameter that governs the dependence relationship. frank copula: the frank copula is suitable for modeling data characterized by weak tail dependence. this copula is given by: ѳ , ѳ 1 ѳ 1 ѳ 1 ѳ 1 (6) where -∞< ѳ <∞ is a parameter that governs the dependence relationship. 2.1 estimation several methods can be used for selecting the best copula family that fits the dependence relation. these methods are categorized as nonparametric and semi parametric. chiyoshi (2004) presented the procedure in order to implement these methods. the copula selection method in this work will depend on the semi parametric estimate that is known as the likelihood function. this function is extended to include the akaikes information criteria (aic) which was presented by chiyoshi (2004). given a random sample, x ,…,x :k 1,…. ,n from distribution f x ,…. . ,x c f x ,…,f x , the usual procedure is to select the parameter that maximizes the pseudo loglikelihood, ∑ log , , }]. (7) where is the copula density and is the rescaled empirical distribution function for each 1 . and is given by: 1 1 1 (8) after obtaining the likelihood, aic will be utilized such that: aic 2 log likelihood 2 k (9) where k is the number of the parameters which in our case is one parameter. the best model is the one with the lowest aic. m. al-kindi, m. al-lawati and n. al-azri 4 3. case study cost and time estimates in well drilling are very crucial in deciding the feasibility of a drilling project. financial crisis has become more probable than before, which is a fact that rendered organizations more vulnerable to bankruptcy because of faulty decision making. hence, organizations have started to invest more on optimal budgeting and tracking their consumption patterns, which will help in their future decision making. there are different methodologies proposed in the literature for optimal budgeting. in a survey on drilling cost and the complexity of estimation models, kaiser and pulsipher 2007, described the development of cost and complexity metrics in well drilling and studied several methodologies like, joint association survey (jas), energy information administration (eia), mechanical risk index (mri), directional difficulty index (ddi), and difficulty index (di). according to their review, the joint association surveys (jas) and mechanical risk indexes (mri) are the most popular methods used in evaluating the cost and complexity of drilling in the gulf of mexico. mri was developed in the late 1980s when conoco inc. engineers were tasked to compare offset drilling data for a collection of offshore wells in the gulf of mexico (kaiser and pulsipher 2007). mri takes into consideration, water depth (wd), measured depth (md), kick off point (for sidetracks), true vertical depth (tvd) in addition to mud weight (mw) and horizontal displacement (hd) at md (williams et al. 2001). 3.1 drilling data the historical data used in this work represents more than 300 different wells drilled in different parts of oman between 2008 and 2012. the data contained information on well drilling direction, well function, various rigs operating in the well, well actual cost, well estimated cost, time to completion, estimated time to complete, and the mri log for each well. therefore, the first step was to sort the data according to well type names. 3.2 copulas and simulation mechanical risk index (mri) has been used as the key input parameter in order to estimate the cost and time of drilling. in order to understand the relationship between actual cost/duration and mri, coefficient of determination method was utilized. however, it was found that this coefficient cannot explain the variation within the different groups of mri and the actual cost/duration data as shown in table 3 and table 4. therefore, copula was used in order to have better view on the nature of the relationship between actual cost/duration and mri. therefore, it is very essential to input the random mri and the actual cost/duration data in a way that fits copula input variables such that values of and / . these values are the input to the generator equation of the best copula family. the flow chart of the model shows how copula has been implemented in order to reach to new estimates for each well type, which is shown in fig. 1. table 3. coefficient of determination values between actual cost and mri per category. mri range 0-300 11.8% 300-600 28.6% 600-900 1.3% 900-1200 0.0% 1200-1500 0.0% 1500-1800 23.3% 1800-2100 26.0% 2100-2400 8.7% 2400-2700 0.0% 2700-3000 7.5% 3000-3300 0.0% 3300-3600 0.0% table 4. coefficient of determination values between duration and mri per category. mri range 0-300 17.4% 300-600 24.4% 600-900 0.7% 900-1200 0.3% 1200-1500 0.0% 1500-1800 22.6% 1800-2100 13.6% 2100-2400 12.1% 2400-2700 0.0% 2700-3000 0.0% 3000-3300 0.0% 3300-3600 0.0% estimating drilling cost and duration using copulas dependencies models 5 figure 1. copula model flow chart. m. al-kindi, m. al-lawati and n. al-azri 5 the initial step is to know the type of copula family that represents the dependence relationship between mri with actual cost and mri with duration. it was found that clayton copula represents the relation between mri and the actual cost, while gumbel copula represents the relationship between mri and duration as shown in table 5 and table 6. the tables show the best fit for the copulas models using @risk software. different goodness of fit criteria was compared such as: akaike information criterion (aic), hannan–quinn information criterion (hqic), and schwarz information criterion (sic). after learning/understanding the best copula family that fits the nature of the dependence, a 15000scenario simulation was run. each well type data had different value for the kendall’s tau that represents the relationship between mri and actual cost/duration. therefore, each will had its own scenarios based on this dependency structure. figure 2 shows some examples of the simulation outcomes. table 5. copula best fit between mri and actual cost. table 6. copula best fit between mri and duration. figure 2. mri with cost/duration copula simulations examples. bivariate copulas copulabiclayton copulabigumbel copulabifrank copulabit copulabinormal mle fits clayton(3.22,4) gumbel(2.61,1) frank(8.41,1) t(7,0.8) normal(0.88) goodness of fit aic -2498.650074 -2484.865354 -2156.425193 -1938.887777 -1675.124018 aic ranking 1 2 3 4 5 sic -2487.331731 -2473.547012 -2145.10685 -1927.569434 -1669.463904 sic ranking 1 2 3 4 5 hqic -2494.510635 -2480.725916 -2152.285755 -1934.748339 -1673.053356 hqic ranking 1 2 3 4 5 log likelihood 1251.327863 1244.435503 1080.215423 971.4467147 838.5629507 log likelihood ranking 1 2 3 4 5   bivariate copulas copulabigumbel copulabiclayton copulabifrank copulabit copulabinormal mle fits gumbel(2.608,1) clayton(3.21,4) frank(8.39,1) t(7,0.869) normal(0.86) goodness of fit aic -2375.643145 -2365.262746 -2167.439404 1905.495511 -1711.04269 aic ranking 1 2 3 4 5 sic -2364.324803 -2353.944403 -2156.121062 1894.177168 -1705.382576 sic ranking 1 2 3 4 5 hqic -2371.503707 -2361.123307 -2163.299966 1901.356072 -1708.972028 hqic ranking 1 2 3 4 5 log likelihood 1189.824399 1184.634199 1085.722528 954.7505817 856.5222866 log likelihood ranking 1 2 3 4 5   50 100 150 200 250 300 5 10 15 20 25 mri d u ra ti o n 6 estimating drilling cost and duration using copulas dependencies models 7 from this simulation, three values were extracted for each well type. these values represent the low (10th percentile), base (50th percentile) and high values (90th percentile) of the expected estimates of drilling cost and duration. then, sensitivity analysis was conducted using monte carlo simulation in 100 iterations where the probability distribution of the historical data for each well was fitted. the fitting was based on anderson darling test. this test depends on calculating the p-value for the goodness of fit test, which leads towards determining the best distribution that fits the data. the selection of the distribution depended on the one with the lowest anderson darling statistic value. the selected distribution was utilized with the three output values from copula model in order to obtain the optimum value for the estimates. after running the simulation, it was found that the optimal value is the median. this is supported by the output of monte carlo sensitivity analysis for well a as shown in fig. 3 and fig. 4. the next step was to verify the estimates with the actual data. additional data points were figure 3. monte carlo sensitivity analysis for actual cost estimates for a given well. figure 4. monte carlo sensitivity analysis for duration estimates for a given well. provided in order to validate the calculated estimates. four different indicators were used to test the efficiency of the presented model. the indicators are: (1) whether the actual cost/duration is within the range of the three predicted estimates, (2) the number of data points that show a lower deviation from the actual values compared to the company forecasted estimates when the base value obtained from copula model is used to compare with the actual values, (3) the number of data points that has a lower deviation from the actual values when the budgeting team picks the best value from the three estimates assuming that the team would pick the closest value to the actual one from the three values provided, and (4) the number of points that show a lower deviation if the team would choose the closest value to the company forecasted estimates from these three values. 4. simulation results the historical data for the years 2008-2012 were used during the implementation of the model in order to get the new estimates. these new estimates were representing the upcoming actual cost and time required when the same well is drilled. the final step was to compare our model output with the real data in order to validate our model performance. new data was provided by the company for this purpose. however, the new dataset included few wells that are common with those recorded from 2008 to 2012 which limited the validation procedure to only 110 points and almost 20% of the number wells from the previous dataset. table 7 below shows the results of the comparison through utilizing the four different indications table 7. the performance of copula model estimates. indicator copula results (1) (2) (3) (4) cost results 94 out of 110 (85%) 35 out of 110 (31.8%) 52 out of 110 (47.3%) 37 out of 110 (33.6%) durati on results 65 out of 110 (59%) 34 out of 110 (31%) 83 out of 110 (75.5 %) 75 out of 110 (68%) under-estimation over-estimation m. al-kindi, m. al-lawati and n. al-azri 8 in order to validate the accuracy of the model, we choose four performance indicators. indicator (1) is to test whether the actual new data point is within the predicted range for both cost and duration. the second indicator (2) is to show if the predicted value using the model is better than the deterministic model used by the expert opinion in the company. the third indicator (3) is to let the expert choose one point (i.e. either low, base, or high value) and compare with the actual result. the last indicator is to choose the nearest prediction to the new actual data (i.e. if low prediction value is close to the new actual then it will be selected). the model presented in the paper shows a better prediction compared to previous work by valdes et al. (2013) for a different data set. however, no previous literature is available to benchmark our results for the other remaining indictors. 5. conclusion in this paper, copula probability theory was implemented to model the dependency between mri and drilling cost and duration. typically, the archimedean copulas was selected due to its ability to represent the different types of dependencies. the model selection depended on using the akaikes information criteria method. this method was the expansion for the likelihood function. it was found that clayton copula represents the dependence relationship between mri and cost with gumbel copula represents the dependence relationship between mri and duration. the model was extended to provide new estimates for the cost and time required to drill wells. at the end, validation of the estimates was done by utilizing several indicators. the results from the validation showed that the model was able to predict the new values through utilizing the historical data. acknowledgment the authors would like to thank sultan qaboos university and petroleum development oman (pdo) for their support of this research # cr/eng/mied/12/03. references aas k (2006), technical report on “modeling the dependence structure of financial assets: a survey of four copulas”. norwegian computing center, oslo, norway. al-harthy m, begg s, bratvold rb (2007), copulas: a new technique to model dependence in petroleum decision making. journal of petroleum science and engineering 57(1): 195-208. chiyoshi fy (2004), modeling dependence with copulas: a useful tool for field development decision process. journal of petroleum science and engineering 44(1): 83-91. clayton d g (1978), a model for association in bivariate life tables and its application in epidemiological studies of familial tendency in chronic disease incidence. biometrika 65(1): 141-151. embrechts p, mcneil a, straumann d (2002), correlation and dependence in risk management: properties and pitfalls. risk management: value at risk and beyond 176223. frank mj (1979), on the simultaneous associativity of f(x,y) and x+y-f(x,y). aequationes math, 1: 194-226. genest c, ghoudi k, rivest lp (1995), a semiparametric estimation procedure of dependence parameters in multivariate families of distributions. biometrika 82(3): 543-552. gumbel ej (1960), bivariate exponential distributions. journal of the american statistical association 55(292): 698-707. jaworski p, durante f, hardle wk, rychlik t (2010), copula theory and its applications. new york: springer. kaiser mj, pulsipher ag (2005), rigs-to-reef programs in the gulf of mexico. ocean development and international law, 36(2): 119-134. mahfoud m (2012), bivariate archimedean copulas: an application to two stock market indices. vrije universiteit amsterdam, bmi paper, amsterdam. nelsen rb (1998), an introduction to copulas (lecture notes in statistics). noerager ja, norge e, white jp, floetra a, dawson r (1987), drilling time predictions from statistical analysis. society of petroleum engineers doi:10.2118/16164-ms. patton aj (2012), a review of copula models for economic time series. journal of multivariate analysis 110: 4-18. pradier e (2011), copula theory: an application to risk modeling. technical report, grenoble inp-ensimag. valdes a, mcvay da, noynaert sf (2013), uncertainty quantification improves well estimating drilling cost and duration using copulas dependencies models 9 construction cost estimation in unconventional reservoirs. proceedings of spe unconventional resources conference canada. 5-7 november, calgary, canada. williams c, mason js, spaar j (2001), operational efficiency on eight-well sidetrack program saves $7.3 million vs historical offsets in mp 299 / 144 gom. society of petroleum engineers doi: 10.2118/67826-ms. zoller sl, graulier jr, paterson aw (2003), how probabilistic methods were used to generate accurate campaign costs for enterprise's bijupira and salema development. in the proceedings of the spe/iadc drilling conference, amsterdam, the netherlands. spe elibrary 79902. the journal of engineering research (tjer), vol. 19, no. 1, (2022) 13-21 *corresponding author’s e-mail: alawim@squ.edu.om doi:10.53540/tjer.vol19iss1pp13-21 modeling and testing of a maintenance project using simphony: case study mubarak al alawi*, mohammed al shahri, and zainab bacal department of civil and architectural engineering, sultan qaboos university, muscat, oman abstract: unlike building a new project, the maintenance and rehabilitation of an existing project are complex. the complexity is driven by different factors such as limited access to the site, limited working hours, restricted storage area, a limited number of allowed crews to work simultaneously, etc. considering all of these factors is essential for construction process planning. hence, testing the maintenance project under different project scenarios will help to identify the optimum resources configuration. in this paper, we present a case study of a maintenance project located in al khodh, oman. it is required to model and simulate the maintenance operation, given many client-based work constraints. the project was modelled using a simphony simulation environment. the simulation model was tested in different resource configurations, including several floor accesses, other tiles breaking and tiles installation teams, and a different number of waterproofing teams. this study demonstrated that the tiles installation team is the most influencing resource in the project. given one floor access at a time and using two tiles installation produces the most economical resources configuration. three tiles installation teams are the best selection for two-floor accesses, and seven tiles installation teams selection is best for three floors. keywords: construction operation; simulation; simphony; maintenance project; discrete event simulation دراسة حالة: simphony نمذجة واختبار مشروع صیانة باستخدام زینب بقالمبارك العلوي* ومحمد الشحري و . یرجع التعقید إلى عوامل مختلفة مثل معقدةعتبر عملیة ت صیانة وإعادة تأھیل مشروع قائم جدید،على عكس بناء مشروع :الملخص المسموح بھا أطقم العمل وعدد التخزین،ومحدودیة مساحة المحدودة،ساعات العمل إلى و ،صعوبة حركة المعدات الثقیلة في الموقع من وبالتاليلصیانة، امتطلبات مشروع للعمل في وقت واحد وما إلى ذلك. یعتبر النظر في كل ھذه العوامل ضروریًا لتخطیط وتقدیر یقع في سكني لمشروع صیانة منزل تبحث ھذه الدراسة عملیة نمذجةمختلفة. عمل المھم اختبار مشروع الصیانة في ظل سیناریوھات ، محددة من قبل صاحب المشروعضوء العدید من قیود العمل المحاكاة عملیة الصیانة في وتھدف إلى عمان في سلطنةمدینة الخوض . تم تصمیمھذه الدراسة إلى تحدید خیارات الموارد البشریة المثلى إلتمام المشروع خالل فترة زمنیة قصیرة وبتكلفة مالیة مثالیةوتسعى . ركة البناء التي تعمل في المشروعتم توفیر مدخالت النموذج من قبل شو simphonyباستخدام بیئة محاكاة نموذج محاكاة المشروع ، وعدد تركیب البالطو إزالةوعدد فرق كعدد الطوابق المسموح العمل بھا مدخالت مختلفة في ظلج المحاكاة م اختبار نموذأیضا ت في حالة السماح بالعمل في الطوابق بصورة مختلف من فرق العزل المائي. لقد وجد أن استخدام فریقین لتركیب البالط ھو أفضل اختیار استخدام متزامنة ویفضلفي حالة العمل في طابقین بصورة كیب البالط ھي أفضل اختیارثالثة فرق تر متسلسلة ووجد أیضا استخدام .في حالة العمل في ثالث طوابق بصورة متزامنةسبعة فرق تركیب البالط .محاكاة األحداث المنفصلة ؛أعمال صیانة ؛سیمفوني ؛محاكاة ؛أعمال بناء :الكلمات المفتاحیة modeling and testing of a maintenance project using simphony: case study 14 1. introduction construction activities pose complexity driven by the complex nature of the construction resources required. construction productivity may be subjected to different factors that hinder the optimum uses of resources such as work environment, project nature, project resources, and project management (al alawi, 2021). labour, material, equipment, and budget should be utilized efficiently to maintain the project milestones. construction resources must work in harmony to achieve the projected cost, time, and quality defined for the project. therefore, it is vital to understand the work of the construction activities, build a model, and simulate the construction operation to identify the optimum number of resources to utilize. many optimization research studies covering different areas in construction were covered in the literature. for example, osama and alshibani (2009) optimized construction equipment using the genetic algorithm technique. alavi and abourizk (2017) and zhou et al. (2009) integrated genetic algorithms with simulation to optimize construction site layout variables. albataineh et al. (2013) used a scenario-based simulation analysis for project planning and decision support for a tunnelling construction project. lu et al. (2008) optimized the project schedule using a simplified simulation-based scheduling system. the system combined a discrete event simulation approach (sdesa) and particle swarm optimizer (pso). it formulates resources scenarios and finds the shortest project duration. zahraee et al. (2013) used discrete event simulation to model a concrete pouring process and identify the optimum resources. similarly, hassan and gruber (2008) used discrete event simulation to model concrete paving operations. the model was used to study the impacts of the resources on the flow of operation and the construction process's costeffectiveness. al-alawi, bouferguene, and mohamed (2018) applied distributed simulation with hla to integrate weather effects in simulating earthmoving operations. riga, et al. (2020) used mixed integer programming to optimize the tower crane and storage areas on construction sites. yassin et al. (2020) used agent-based modelling to optimize the workflow of robotic steel and concrete 3d printers. the selection of the best simulation method or the optimization algorithm to be used depends on many factors, such as the type of project, data availability, total operation time, and costs. the simulation allows replicating a system's behaviour, and forecasted changes may exert if tested under different testing scenarios. in this study, a maintenance project in al khodh oman is modelled to find the optimum number of resources required to finish. different work plans are to be tested and a different number of resources configurations. the project poses sequential nature, and thus discrete-event simulation has been selected to build the simulation model. a symphony simulation environment was adopted in this study. the study starts by introducing a brief overview of the simulation. the maintenance project overview and its process description were presented. the testing scenarios were explained, and their results were discussed. 2. simulation overview the simulation system can be either monolithic or distributed. the monolithic simulation models contain components that are executed and terminated simultaneously. meanwhile, the simulation component in the distributed simulation model has its start and end simulation time. it interacts with the other simulation components by sharing information to replicate a real system (abourizk, hague, and ekyalimpa, 2016). in addition, depending on its characteristics, a real system can be simulated using different simulation techniques such as continuous simulation, time-based discrete simulation, and discrete event simulation. the difference between the discrete event simulation and the other two techniques is that the model's state is affected/changed by the state's change. meanwhile, the continuous and time-based discrete simulation state of the model depends on the progress of time (brito and botrter, 2011). different simulation environments have been introduced and used in research; some are listed as follow: abc (beaumont et al., 2002) symphony (abourizk and hajjar, 1998) slam (pritsker and o'reilly, 1997) stroboscope (martinez, 1996) micro cyclone (halpin, 1977) symphony is a simulation tool developed by abourizk and hajjar (1998). it has built-in simulation elements, see table 1, used to construct the operation flow of events. furthermore, it allows system development to introduce simulation templates. a simulation model in simphony starts with identifying and abstracting the case study and its overall process. then the sequences diagram is plotted, and the time parameters for each event's duration are collected from the real operation. one of the main challenges in developing a reliable simulation model is collecting real-time data of the modelled process. however, most companies keep records of durations, productivity rate, and other information, which reduces the time required for model building. if no data exists, the studied problem shall be surveyed, and data recording tools such as cameras can collect the necessary data. 3. maintenance project overview the maintenance project scope of work is to change all the tiles, waterproof, and heat insulation for a villa located in a residential compound in al khodh city, sultanate of oman. the selected case study is situated within a residential compound with eight typical occupied villas, as shown in fig. 1. mubarak al alawi, mohammed al shahri, and zainab bacal 15 table 1. symphony simulation elements names and description. element name description “create” a “create” element is responsible for creating entities and introducing them into the model “resource” a resource element is responsible for defining a shared resource. “preempt” the preempt element force capture the servers of a resource to an entity “capture” the capture element is responsible for granting the exclusive use of one or more servers of a resource to an entity “release” the release element allows an entity to return servers it has previously captured to the pool of available servers. “file” a file element is responsible for defining a queue in which entities wait for a shared resource. “task a task element is responsible for modelling an activity. it achieves this by holding the entity for a while, as specified in its duration property” “batch” a batch element aggregate n number of entities and release them as one entity “unlatch” the unbatch element breaks down the entity to its original number of entities. “generate” a generate element creates one or more clones of a passing entity. “consolidate” a consolidate element blocks an entity arriving via the upper branch until one or more entities arrive via the lower branch. “counter” a counter element is used to record important milestones in the lifecycle of an entity” the project maintenance operation is conducted on all three floors of the project (ground floor, first floor, and roof floor). the complexity associated with this project is related to the operational constraints set by the client. the residential complex is fully occupied except for the studied villa. time constraints such as the number of working hours and the start and end time of the workday have been set by the client. due to limitations at the construction site, construction was allowed only during the daytime for eight hours. the project shall also be fully covered with a green mesh to protect the surrounding from dust pollution. the project's working clearance was set to be one meter from all sides except for the backside of the villa, which was set to be 2 meters to allow for construction waste storing before moving it outside of the residential compound, loading it on the hauling truck. the contractor decides to allocate a construction waste collection area on each floor, see fig. 2 and fix a construction waste dumping tube extending from the roof floor to the ground floor. the contractor also decided to use part of the backside of the project to store some of the construction materials. in addition to the constraints mentioned above, no trucks were allowed to enter the residential compound because of the client's request to provide minimum disturbance to the occupants and the limited manoeuvring area available on the project site. as a result, the contractor allocates another construction waste collection point for the trucks outside the project location. the construction waste transportation from the construction site to the landfill is subcontracted to a different contractor by the primary contractor. it is because of the limited accessibility of the truck to the site and to focus the contractor’s resources in the maintenance operation only. figure 1. top view of the residential compound (google earth-2021) modeling and testing of a maintenance project using simphony: case study 16 figure 2. first-floor project layout 4. maintenance project process description the modelled maintenance operation is built in collaboration with the contractor’s construction engineer. the maintenance operation is split into three sub-processes: roof floor process, first-floor process, and the second-floor process. this splitting process is mainly designed to allow the model to test the operation under different work start operations. for example, one floor at a time, two floors at a time, or three floors at a time scenarios. to achieve the flexibility to test the project operation under different work start times scenarios, a resource called “floor access” was created in the model. the roof floor is the first to start the maintenance operation because the preempt element “get roof floor access”, as shown in fig. 3, is a higher priority than the same element in the first and ground floors sub-processes. once the floor access is granted, the generated element “generate 116 m2” will generate the total area of the roof, 116 m2. the 116 m2 will undergo six significant activities, which are: 1. breaking roof tiles 2. transporting and dumping the construction waste through waste dump tube 3. cleaning the roof to prepare for the waterproofing 4. waterproofing 5. waterproof testing 6. heat boards and tiles insulation. three crews were added to the simulation model to perform the maintenance activities: the breaking team, waterproofing team, and tiles installation team. the selection of those three resources is based on the contractor request. it is worth mentioning that the local contractors classify their work according to the local construction trades in the market. the definition of the breaking work role set by the contractor is that the breaking team shall stop breaking when four m2 are completed and shall transport the construction waste to the dump pipe before starting the breaking work again. once all tiles breaking is complete, the breaking team shall prepare the roof floor for waterproofing and clean and put concrete screeds on all edges. the waterproofing activity comes next and will acquire the waterproof crew to do the work. once the waterproofing process is complete, the roof floor will be filled with water for 48 hours for testing purposes. heat boards and tiles installation are combined in one task in the project and will be performed by the tiles installation team. the first-floor maintenance operation is similar to the roof floors; however, the main difference is the resources' priority role. the model will test the operation under different resources and start work scenarios, and higher priorities are given to the upper floors. for example, if two-floor accesses are provided, the roof floor and first floor will start concurrently. if only one breaking team is available, it will be assigned to the roof floor because of its high priority. although the project has rooms and bathrooms, which can be modelled individually, the construction engineer prefers to finish the task because of the limited workspace. the ground floor work is no different from the previous floors. however, the main difference is that no waterproofing is required, but concrete screeds on all ground floor areas will be added before the start of tiles works. table 2 shows all simulation elements’ initial properties, such as individual task's duration and the number of crews in each element. the construction engineers specify the time of each task extracted from the historical database. a uniform distribution function for all activities durations was implemented. it is noteworthy that the contractor measures and records the productivity of the resources in a daily manner. however, this study is converted to an hourly rate by dividing the daily output by eight working hours. in addition, the waterproof installation and testing are considered constant because of the project's small scale, and their values are one day and two days, respectively. 5. discussion of results the maintenance project was tested under different testing scenarios. in the first scenario, the contractor has only one-floor access. the contractor shall perform the maintenance activities sequentially, starting from the roof and ending at the ground. mubarak al alawi, mohammed al shahri, and zainab bacal 17 f ig ur e 3. m ai nt en an ce p ro je ct s im ul at io n m od el u si ng s im ph on y. modeling and testing of a maintenance project using simphony: case study 18 table. 2. simulation processes and their initial parameters floor process time (min) roof floor breaking 1 m2 of roof tiles uniform (19.5, 24) transporting every 4 m2 broken roof tiles to dump pipe 4 cleaning the roof 480 min (one day work) waterproofing the roof 480 min (one day work) testing the waterproof of the roof 2880 min (left for six days) installation of roof tiles uniform (19.5, 24) first floor breaking 1 m2 of roof tiles uniform (12,16) transporting every 4 m2 broken roof tiles to dump pipe 4 clean all rooms 480 min (one day work) waterproofing wet areas 1440 min testing the waterproof of the roof 2880 min (left for six days) installation of first-floor tiles uniform (30,40) ground floor breaking 1 m2 of ground floor tiles uniform (12,16) cleaning every 4 m2 4 clean all floor 480 min (one day work) concrete screed 480 min (one day work) installation of ground floor tiles uniform (30,40) furthermore, a different number of resources were applied in this scenario to examine their sensitivity correlation with the project duration. for example, the tiles breaking team number of resources were increased by one resource in each simulation run, and the total project duration was recorded. similarly, the total project duration was recorded when the same steps were repeated in the waterproofing and tiles installation teams. in the second scenario, two roof accesses were granted to the contractor, which means that the contractor can work on two floors simultaneously conditioned by the availability of resources. similarly, each type's number was increased by one, and the total project duration was recorded. the third testing scenario is similar to the second scenario. the main difference is that the third scenario allows the contractor to work on all floors simultaneously and is conditioned by resources' availability. also, the priority of utilizing the resources is given in the following sequence: the roof floor has the highest priority, the first floor is the second, and the ground is placed last. figure 4 shows the maintenance total project duration based on one-floor access and a different number of resources. the total project duration has demonstrated a sensitivity to an increase in the number of resources except for the waterproofing team. this might be due to the fact that this team is called for a specific job task during the operation, and its work duration is generally constant. the original project duration was 58 days, given only one team is used for each process. however, it is reduced to 46 days if the breaking team number is increased to 6, and no more reduction in the duration can be achieved further. if the tiles installation team is expanded to seven, the project duration can be reduced to 39 days with no extra reduction if more tiles installation teams are employed. figure 5 shows the simulation model results when tested under two-floor accesses and a different number of resources. the original project duration was found in this case to be equal to 39 days. this means that the freedom to move from one floor to another has drastically improved the total project duration. similar to the first testing scenario's findings, the increase in the waterproofing team number does not affect the project duration. the project duration can be reduced to 33 days if six tiles breaking teams are employed and can be reduced to 24 days if eight tiles installation teams are utilized. it is worth mentioning that the model is tested by fixing the number of resources of two teams and changing one team at a time. figure 4. project duration based on one-floor access and under different number of resources scenarios figure 5. project duration based on two-floor accesses and under different number of resources scenarios. 30 35 40 45 50 55 60 1 3 5 7 9 d ur at io n (d ) number of resources project duration (day) under different number of breaking team members project duration (day) under different number of water proofing team members project duration (day) under different number of tiles installation team members 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 d ur at io n (d ) number of resources project duration (day) under different number of breaking team members project duration (day) under different number of water proofing team members project duration (day) under different number of tiles installation team members mubarak al alawi, mohammed al shahri, and zainab bacal 19 figure 6 shows the third testing scenario results, which include three floors accesses. the original project duration in this scenario is 39 days. the project duration can be reduced to 11 days if nine breaking teams are employed and 19 days if seven tiles installation teams are utilized. similar to the first two testing scenarios, no effect on the total project duration is reported despite more waterproofing teams being employed in the project. in addition to analyzing the project duration under different floor accesses and a different number of team numbers, the direct labour cost was also analyzed. the analysis investigated the direct labour cost's sensitivity under different floor accesses and different team numbers. the rate for each team is given by the contractor and listed in table 3. figure 7 shows the total labour direct cost is given one-floor access and a different number of teams. the original total labour cost was omr 1987, with a total project duration equal to 58 days. the minimum labour direct cost under a different number of breaking teams was omr 1986. it was achieved by utilizing six breaking teams and resulted in a total project duration of 46 days. using two tiles installation teams and onefloor access further reduced the direct labour cost to omr 1864 and the project duration to 44 days. based on the above findings, the contractor recommended the following to achieve the lowest total direct cost given the construction works constraints defined by the client: one-floor access optimum: use two tiles, installation teams. two floors access: use three tiles installation accesses. three floors access: use seven tiles, installation teams. figure 8 shows the direct labour cost given twofloor access and under a different number of breaking and tiles installation teams. like the previous testing scenario, employing six breaking teams will provide the minimum cost, omr 1982, and allow the project to finish in 33 days. however, the project duration can be further reduced to 28 days with the exact cost if three tiles installation teams are utilized. figure 9 shows the total labour direct cost given three-floor accesses and under different teams numbers. the minimum labour cost can be reduced to omr 847 from omr 1991 if ten breaking teams are utilized, and it will allow the project to finish in 19 days. on the other hand, assigning seven tiles installation teams will result in a labour direct cost equal to omr 1631 and allow the project to finish in 19 days. figure 6. project duration based on three-floor accesses and under different number of resources scenarios table. 3. maintenance teams rates. maintenance activity price rate (omr/h) mosaic tiles breaking (roof tiles) 5 tiles breaking 5 mosaic tiles installation (roof tiles) 3.375 tiles installation 4.375 waterproof operation 12 figure 7. total labour cost under different number of resources given one-floor access. 10 15 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 d ur at io n (d ) number of resources project duration (day) under different number of breaking team members project duration (day) under different number of water proofing team members project duration (day) under different number of tiles installation team members 1600 1650 1700 1750 1800 1850 1900 1950 2000 2050 1 2 3 4 5 6 7 8 9 10 t ot al la bo ur c os t ( o m r ) number of resources total labour cost under different breaking team number total labour cost under different tiles installation team number modeling and testing of a maintenance project using simphony: case study 20 figure 8. total labour cost under different number of resources given two-floor accesses. figure 9. total labour cost under different number of resources given three-floor accesses. 6. conclusion a maintenance project located in al khod was modelled using discrete event simulation. the model was built using the simphony simulation environment. the simulation model was built using input given by the construction engineer of the maintenance project. the project was tested under different resource configurations, including floor accesses, other tiles breaking and tiles installation teams, and a different number of waterproofing teams. the total project duration was 58 days, given one resource in each resources category. it was found that the project duration was sensitive to the tiles breaking and tiles installation teams. the project duration can be reduced to 24 days in case of eight tiles installation teams, and two floors access are available to the contractor. employing nine tiles breaking teams and allowing the contractor to work on all floors simultaneously can further reduce the total project duration. the total labour direct cost was also analyzed to select the optimum project duration. the analysis that using two tiles installation teams and given one-floor access will produce the optimum cost. three tiles installation teams are the best selection for two-floor accesses, and seven tiles installation teams’ selection is best for three floors. the project considered direct labour cost; however, the indirect cost may also influence the sensitivity analysis. also, the time and cost analyses were performed independently from each other. therefore, the future direction for the project is to perform a time-cost tradeoff analysis. conflict of interest the authors declare no conflict of interest. funding no funding was received for this research. references abourizk, s. m., & hajjar, d. (1998). a framework for applying simulation in construction. canadian journal of civil engineering, 604-617. abourizk, s. m., hague, s. a., & ekyalimpa, r. (2016). construction simulation: an introduction using simphony. edmonton: university of alberta. al alawi, m. (2021). delay in payment efects on productivity of small and medium construction companies in oman: exploration and ranking. asian journal of civil engineering, 1347-1359. al alawi, m., bouferguene, a., & mohamed , y. (2018). integrating weather in modelling earthmoving mining operation using distributed simulation with hla standards. 30th european modeling and simulation symposium (pp. 1-6). budapest : curran associates. al-bataineh, m., abourizk, s., & parkins, h. (2013). using simulation to plan tunnel construction. journal of construction engineering and management, 564-571. beaumont, m. a., zhang, w., & balding, d. j. (2002). approximate bayesian computation in population genetics. genetics, 2025-2035. brito, t. b., & botrter, r. c. (2011). a conceptual comparison between discrete and continuous simulation to motivate the hybrid simulation methodology. proceedings of the 2011 winter simulation conference (pp. 3915-3927). phoenix, arizona : winter simulation conference. halpin, d. w. (1977). cyclone — method for modeling job site processes. journal of the construction division, 489-499. 500 700 900 1100 1300 1500 1700 1900 2100 1 2 3 4 5 6 7 8 9 10 t ot al la bo ur c os t ( o m r ) number of resources total labour cost under different breaking team number total labour cost under different tiles installation team number 1960 1970 1980 1990 2000 2010 2020 1 2 3 4 5 6 7 8 9 10 t ot al la bo ur c os t ( o m r ) number of resources total labour cost under different breaking team number total labour cost under different tiles installation team number mubarak al alawi, mohammed al shahri, and zainab bacal 21 hassan, m. m., & gruber, s. (2008). simulation of concrete paving operations on interstate-74. journal of construction engineering and management, 2-9. lu, m., lam, h. c., & dai, f. (2008). resourceconstrained critical path analysis based on discrete event simulation and particle swarm optimization. automation in construction, 17(6), 670-681. martinez, j. (1996). stroboscope — state and resource based simulation of construction process. ann arbor: university of michigan. moselhi, o., & alshibani, a. (2009). optimizing of earthmoving operations in heavy civil engineering projects. journal of construction engineering and management, 948-954. pritsker, a. b., & o'reilly, j. j. (1997). simulation with visual slam and awesim. new york: john wiley sons. razavialavi, s., & abourizk, s. (2017). site layout and construction plan optimization using an integrated genetic algorithm simulation framework. journal of construction engineering and management, 31(4)04017011. riga, k., jahr, k., thielen, c., & borrmann, a. (2020). mixed integer programming for dynamic tower crane and storage area optimization on construction sites. automation in construction, 103259. zahraee, s. m., hatami, m., yusof, n. m., rohani, j. m., & ziaei, f. (2013). combined use of design of experiment and computer simulation for resources level determination in concrete pouring process. teknologi, 43-49. zhou, f., abourizk, s., & al-battaineh, h. (2009). optimization of site layout using a hybrid simulation-based system. simulation modelling practice and theory, 348-363. مبارك العلوي* ومحمد الشحري وزينب بقال 1. introduction 2. simulation overview 3. maintenance project overview 4. maintenance project process description 5. discussion of results 6. conclusion conflict of interest funding references copy (2) of mech26_05_04.qxd the journal of engineering research vol. 3, no.1 (2006) 10-18 ____________________________________________ *corresponding author’s e-mail: jalaljalil@yahoo.com heat transfer enhancement in a ribbed duct j.m. jalil*, t.k. murtadha, h.m. kadom educational technology department, university of technology, p.o. box 35010, baghdad, iraq received 26 may 2004; accepted 21 march 2005 abstract: the rib enhancement of heat transfer in a duct is studied numerically and experimentally, where hot air passes through a duct (0.04 x 0.16 x 1.15 m3) with different rib arrangement. the arranments are lower 12-rib arrangement; upper 12 rib arrangement and 24 rib staggered arrangement. the staggered arrangement gives better performance than the others. also, the angle of attack was studied for lower arrangement, three different values were tested (45°, 60° and 90°). angle of 60° gives better performance. numerically, the three-dimension continuity, navier-stokes and energy equations are solved by finite volume method of flow of air through (0.04 x 0.16 x 0.6 m3). validation of the code was performed by comparing the numerical result with the results obtained experimentally for staggered arrangement only. the agreement seems acceptable. the numerical studies were extended to study the case of cold air passing through hot ribbed duct. keywords: cfd, heat transfer enhancement, finite volume ≥fgƒy hp iô› ‘ √qgô◊g ∫é≤àfg ú°ù– ::áá°°uuóónn iô› ∫ón øné°s agƒg qgôeg ¢sqo ~≤d ,é«∏ªy .é«∏ªy h éjo~y ¢sqo ~b ≥fgƒ©dg ᣰsgƒh iqgô◊g ∫é≤àfg ú°ù– ¿g3m).∞∏àfl ö«jôj ™e (1^15*0^16*0^04 ~lh ~bh 90 ,60 ,45 éjghõ∏d h »∏ø°ùdg ö«jî∏d ωƒé¡dg ájhgr â°sqo ∂dòc .agog π° ag ≈£yg ∞déîàÿg ö«jîdg .24 ∞déîàeh ,12 …ƒ∏y ,12 »∏ø°s »g ≥fgƒ©dg äéñ«jôj (0^04*0^16*1^15) iô› ∫ón agƒg ¿éjô÷ io~ùg ωƒé◊g á≤jôw ᣰsgƒh ábé£dgh ºnõdgh á∏àµdg ßøm ä’oé©e πm ” éjo~y .60 ∫ ¿éc agog π° ag ¿g3m. .øné°s iô› ∫ón oqéñdg agƒ¡dg ¿éjôl ádém πª°ûàd ájo~©dg èféàædg äo~e ºk .ádƒñ≤e áfqé≤ÿg âféc ~bh »hƒ°sé◊g èeéfèdg ≥«b~àd ájo~©dg èféàædg ™e âfqƒb á«∏ª©dg èféàædg ::áá««mmééààøøÿÿgg ääggooôôøøÿÿggio~ùg ωƒé◊g ,iqgô◊g ∫é≤àfg ú°ù– ,™fgƒÿg 鵫eéæjo ‘ äébé«°ùdg symbol description ab,ae,af,an,as,aw = coefficient s in general finite-volume equation cµ, c1ε, c2ε = constants in turbulence model dh = hydraulic diameter , m g = gravitational acceleration , m/s2 hr,wr = rib height and width, mm h = height of the d uct, m i,j,k = position indices for the x -, and y, and z -directions iu = turbulence intensity notation 11 the journal of engineering research vol. 3, no.1 (2006) 10-18 1. introduction rib turbulator has numerous applications in practical engineering systems where it is used in internal cooling passages of modern gas turbine blades that must be protected from hot gas streams. thus, the accurate prediction of the characteristics of the flow field in these regions is of great importance. ribs on channel walls can increase heat transfer coefficients in the cooling passages. these ribs which are called turbulators, increase the level of mixing of the cooler air with the warmer air close to channel wall, thereby enhancing the cooling capability of the passage through compressing the coolant flow rate. the rib turbulator was studied extensively in recent years experimentally and numerically. experimentally, different techniques were used to evaluate the aerodynamic of fluid flow and heat transfer. han (1988) studied the effect of the channel aspect ratio on the distribution of the local heat transfer coefficient in rectangular channels with two opposite ribbed walls (to simulate turbine airfoil cooling passages). taslim, et al. (1996) used a liquid crystal technique to measure heat trsnsfer coefficients in twelve test sections with square and trapezoidal cross sectional areas representing blade mid-chord cooling cavities in a modern gas turbine. gao and sunden (2001) used liquid crystal thermography in the heat transfer experiment to demonstrate detailed temperature distribution between a pair of ribs on ribbed surfaces. lee and moneim (2001) examined heat transfer and flow behavior on a horizontal surface with a two-dimensional transverse rib using a cfd model. kim and kim (2001) performed numerical optimization coupled with reynolds-averaged navierstokes analysis of flow and heat transfer for the design of rib-roughened surface. raisee and bolhasani (2003) investigated the fluid flow and heat transfer through ducts roughened by arrays of either attached or detached ribs. the numerical approach used in this work is the finite-volume method. hadhrami and han (2003) studied the effect of various 45o angledrib turbulator arrangements. chang, et al. (2004) presented an experimental study of heat transfer in a rectangular channel with two opposite walls roughened by 45o staggered ribs swinging about two orthogonal axes under single and compound modes of pitching and rolling oscillations. tanda (2004), used repeated ribs on heat exchange surfaces to promote turbulence and enhance convective k = turbulent kinetic energy = ( )222 wvu 2 1 ′+′+′ , m 2/s2 l = length of test section , m p = mean statis pressure, p a, n/m 2 pe = peclet number pr = prandtl number re = reynolds number í hav du . = sö = source term t, tb, tin, tw = mean inlet, bulk and wall temperatures , oc u,v,w = velocities in x ,y,z-directions , m/s uav = average velocity , m/s x,y,z = cartesian coordinates, m ε = dissipation rate of turbulent kinetic energy ρ = density of air, kg/m3 σk, σε = turbulent prandtl numbers for k and ε µ = dynamic viscosity , n.s/m2 ν = kinematic viscosity, m2/s νt = eddy or turbulent viscosity , m 2/s νe = effective kinemat ic viscosity , m 2/s γ = diffusion coefficient, ,/ σµ=γ n.s/m2 γe = effective diffusion coefficient, ee / σµ=γ , n.s/m 2 φ = general dependent variable greek symbols 12 the journal of engineering research vol. 3, no.1 (2006) 10-18 heat transfer. onbasioglu and huseyin (2004) used a liquid crystal based experimental investigation of heat transfer enhancement by ribs on a vertical plate. in this work, a duct roughened with ribs was investigated numerically and experimentally. in both cases, the effect of rib turbulator on fluid flow and heat transfer was studied and analyzed. 2. mathematical model the governing equations are: (patanker, 1980) continuity, (1) navier stokes, (2) (3) (4) also the energy equation: (5) the standard k-ε model (launder and spalding) has two equations, one for k and one for ε. it uses the following transport equations for k and ε. (6) (7) (8) where ε, is the dissipation term. the parametric values in this study are given in table 1. the general partial differential equation has the form: (9) where the three terms on the lefthand side are convection terms and the four terms on the righthand side are diffusion and source terms. the source term appearing in the above governing equation is given in table 2. 3. experimental investigation in this study, flow in rectangular channel with rib turbulator was investigated experimentally for different air bulk velocities of 4.2, 5.8, 8.2 and 10 m/s. figure 1 shows the rib locations along the duct for five arrangements. figure 2 shows a schematics of the experimental apparatus. the test duct has a cross section of 160x40 mm2 with a corresponding hydraulic diameter, dh = 64 mm, and a length of 18 dh . the side walls of the entire test duct were made of plexiglas plates to provide optical access for measurements. 0= ∂ ∂ + ∂ ∂ + ∂ ∂ z w y v x u ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ ρ −= ∂ ∂ + ∂ ∂ + ∂ ∂ z u zy u y x u xx p1 z uw y uv x u ee e 2 ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ ρ −= ∂ ∂ + ∂ ∂ + ∂ ∂ z v zy v y x v xy p1 z vw y v x vu ee e 2 ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + y w zy v yy u x eee ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ ρ −= ∂ ∂ + ∂ ∂ + ∂ ∂ z w zy w y x w xz p1 z w y vw x wu ee e 2 ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ + z w zz v yz u x eee ννν ( ) ( ) ( ) ε−+⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ σ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ σ∂ ∂ + ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ σ∂ ∂ = ∂ ∂ + ∂ ∂ + ∂ ∂ g z kv zy kv y x kv x kw z kv y ku x k t k t k t ( ) ( ) ( ) kkz v zy v y x v x w z v y u x 2 c2gc1 tt t ε − ε +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ε∂ σ∂ ∂ +⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ε∂ σ∂ ∂ + ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ε∂ σ∂ ∂ =ε ∂ ∂ +ε ∂ ∂ +ε ∂ ∂ εε εε ε ⎥ ⎥ ⎦ ⎤ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ∂ ∂ + ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ +⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ + ⎢ ⎢ ⎣ ⎡ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ = ν 222 222 t y w z v x w z v x u y v z w 2 y v 2 x u 2g ( ) ( ) ( ) φφφφ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ φ∂ γ ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ φ∂ γ ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ φ∂ γ ∂ ∂ =φρ ∂ ∂ +φρ ∂ ∂ +φρ ∂ ∂ s zzyyx x w z v y u x cµ c1ε c2ε σk σε 0.09 1.44 1.92 1.00 1.30 table 1. empirical constants in the k-ε ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ γ ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ γ ∂ ∂ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ γ ∂ ∂ = ∂ ∂ + ∂ ∂ + ∂ ∂ z t zy t y x t xz wt y vt x ut ee e to solve the governing eqs. (1)-(5), mathematical expression s for effective kinematic viscosity, eν , and effective diffusion coefficient , γe, are required through use of a turbulence model. ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + x w zx v yx u x eee 13 the journal of engineering research vol. 3, no.1 (2006) 10-18 figure 1. rib locations along the duct figure 2. schematic drawing of the overall experimental system 1. blower 2. heater 3. mesh 4. test section 5. pitot tube 6. manometer 7. control system 8. interface 9. computer 10. temp. control (thermostat) 11. thermocouples 12. ribs (a) lower ribs at α = 90o (b) lower ribs at α = 60o (c) lower ribs at α = 45o (d) staggerd ribs at α = 90o (e) upper ribs at α = 90o 25 cm 30 cm 193 cm 14 the journal of engineering research vol. 3, no.1 (2006) 10-18 the top and bottom are made from steel sheets. the plexiglas rectangular ribs of 5.2-mm width, 6.8-mm height were positioned along the duct at twelve stations. five arrangements were tested: 1. lower wall ribs with α = 45°. 2. lower wall ribs with α = 60°. 3. lower wall ribs with α = 90°. 4. upper wall ribs with α = 90°. 5. staggered arrangement where the ribs are positioned in lower and upper walls with α = 90°. temperature distribution at various locations in longitudinal and lateral directions was measured using insulated copper-constantan thermocouple wire type t (0.27 mm). the rib turbulators were placed on opposite walls with angles of 45°, 60°, 90°. twelve rectangular ribs were tested in this investigation where rib height to duct hydraulic diameter ratio (e/h) is 0.106 and rib width to hydraulic diameter is 0.078. the rib turbulators were rectangular cross section rods (hr = 6.8 mm and wr = 5 mm). the inlet flow to the duct was heated to specified temperatures and the local temperatures were measured in lateral and longitudinal directions using fifty thermocouples: twenty-five thermocouples for measuring the temperature of the flow and twenty-five for measuring the surface temperature of the channel. 4. results and discussions 4.1 experimental results forty tests were carried out. four values of inlet velocities are studied (4.2, 5.8, 8.2 and 10 m/s) with two different inlet temperatures of 40 and 50ºc. in the experimental work, the hot air with controlled temperature enters the duct with ribs. the temperature of the air decreases all the way to the end of the duct because of the heat transfer from hot air to the duct wall which will be enhanced due to turbulence effect of ribs. this enhancement in heat transfer, usually, is accompanied with additional pressure drop. to study the effect of different angles of attack, three angles were tested 45o, 60o and 90o. in 90o, higher pressure drops are expected. the differences in pressure drop are explained as follows, in 90o the flow of air will be stopped completely in front of rib while in 60o, the rib allows the air in front of the rib to turn and take side passage, parallel to the rib. in 45o, easier passage for air is allowed parallel to the rib. the high pressure drops at 90o are not necessarily accompanied by higher heat transfer. in 60o and 45o angles, the heat transfer will be higher due to longer ribs. the rib lengths are 16 cm for lower 90o, 18.5 cm for lower 60o and 22.7 cm for lower 45o. the longer ribs with angle of attack will have two advantages over the short 90o rib. first it causes lower pressure drop, which is the main parameter in fluid flow design. the second advantage is that, the larger rib means higher turbulence effect of consequently higher heat transfer enhancement. the only disadvantage is that longer rib means more material and higher cost. the effect of different angles of attack on heat transfer performance was evaluated through the temperature difference. the high temperature difference means high advantage gain from rib arrangement. angle 60o seems to be the best value. han (1988) also mentions the same for optimum value of angle of attack for rib arrangement. four different inlet air velocities were tested: 4.2, 5.8, 8.2 and 10 m/s. this is achieved through changing the power supply to the fan. due to limitation of heating equipment, only two different inlet air temperatures of 40o and 50o were used. this was done by changing the power supply to the heater working elements. clearly, the higher inlet air temperatures lead to higher air and wall temperatures along the duct. equation φ γφ sφ continuity 1 0 0 u-momentum u νe ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ − x w zx v yx u xx p eee v-momentum v νe ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ − y w zy v yy u xy p eee w-momentum w νe ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ +⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ ν ∂ ∂ + ∂ ∂ − z w zz v yz u xz p eee temperature t γe 0 kinetic energy k γk g-ε dissipation rate ε γε k cg k c 2 21 ε − ε εε table 2. source term in the governing (pdes) 15 the journal of engineering research vol. 3, no.1 (2006) 10-18 4.2 validation of the code the first run of the computer program was for the condition similar to the experimental arrangement where hot air going through the duct and the wall temperature was calculated in addition to air temperature by conduction. one case was tested to compare the results with experimental part. the test case is staggered ribs with 90º, inlet air velocity 4.2 m/s, inlet air temperature 40 ºc. only half length of the duct was simulated by computer program, with 12 ribs in staggered arrangement for half duct only (l = 0.6 m) with same cross section area as in experimental part (0.4 m x 0.16 m). figure 3a shows the velocity vector field, figure 3b shows the isotherm contours. for isotherm contours, the high temperature is shown at inlet and it decreases with very low rate toward the walls and rib. for comparison with experimental results, the air and wall temperatures are compared and good agreement was achieved between the two. 4.3 numerical results the numerical part will be devoted to different physical conditions. usually in most papers, cold air passes through isothermal higher temperature duct walls which we chose to follow. in this part, air with different velocities (2.5, 5, 7.5 and 10 m/s) are passed through a duct similar to experimental part where the cross section is 0.16 x 0.04 m2 with a length of 0.6 m (half the experimental duct). the air inlet temperature was fixed at 20ºc and the wall plus rib temperatures were fixed at 40ºc. number of ribs are 6 for lower and upper arrangements, and 12 for staggered arrangement. three arrangements were tested, staggered, lower and upper. five heights of ribs were tested, 3, 4, 5, 6 and 7 mm. figure 4 presents the distribution of three dimensional flow field in staggered case for re = 25000 and hr = 7 mm. it can be seen from the figure that the introduction of ribs locally reduces the cross section resulting in a main flow acceleration around these obstacles. the sudden expansion downstream of the ribs leads to a separation zone behind them, after reattachment the entrained flow builds up a new boundary layer. the latter is accelerated by the mainstream through shear forces and impinges on the next rib. the flow deflection around the ribs is responsible for high velocity vectors, finally indicating the existence of third small vortex on the top of the ribs. in contour figures, the three-dimensional plot of isotherm contours are plotted for only three levels in zdirection, one near the right wall and second at mid of the duct and the third near the left wall. the air enters with temperature of 20oc and the temperatures of the walls of the duct and ribs are isothermal at 40oc. temperature of the air increases through the duct from 20oc at inlet to 40oc at the walls. the increase in air temperature depends on many factors. the main one is the existence of ribs will work in two different ways; first as surface area with temperature of 40oc. this point was eliminated in experimental part where the ribs are made of plexiglass; but in numerical calculation, the ribs will work as surface area. second the turbulence effect of ribs will increase the heat transfer. the other parameter that increases the air temperature was the rib height. as the height increases, increased area and work and more turbulence is expected. the third parameter is the inlet air velocity. as the inlet velocity increases, less time is available for heat transfer. then the decrease in inlet air velocity will increase the temperature of air through the duct. finally, the kind of arrangement of ribs will affect the process of heat transfer. figure 5 shows the contours for lower arrangement at constant velocity 2.5 m/s with different rib heights 3, 4, 6 and 7 mm. the effect of rib height is clear as the lead to higher air temperatures. the air bulk temperatuee increase as ribs, ribs with higher air velocity decreases. figure 6 shows the variation of air bulk temperature at constant air velocity of 5 m/s and different rib heights of 3, 4, 5, 6 and 7 mm: the higher temperatures at higher heights. the average nusselt number for each of four walls at each x position, can be calculated by, right wall, (10) for lower ribs with 90o, fig. 7 shows plots for fourinlet air velocities. each plot shows the variation of nusselt number with duct length. higher nu results with high rib height. the final correlation equation for air (at constant prandtl number) for each arrangement is: nu = a reb xc hrd (11) the values of constants are given in table 3: 5. conclusions and recommendations the following are concluded: 1. increasing rib height enhances the nusselt number. 2. high reynolds number enhances the nusselt number. 3. the optimum angle of attack was 60o for heat transfer enhancement. ( ) dy dx )k,j,i(t1 nu l 0 − ∫= rib arrange ment a b c d lower 1.478088 0.243154 -0.296225 0.004205 upper 1.974745 0.329244 -0.240705 0.033619 stagge red 4.061650 0.272817 -0.203860 0.118233 table 3 the values of constants in eq. (11) 16 the journal of engineering research iii (2005) 10-18 (3d ) ⏐ 22 may 2005 ⏐ zx y (3d ) ⏐ 22 may 2005 ⏐ 20 30 40 30 30 20 20 20 20 30 30 20 40 40 zx y(3d ) ⏐ 22 may 2005 ⏐ figure 3. velocity vector field and isotherm contours along the duct with staggered ribs, v = 7.5 m/s and hr = 3 mm (3d) ⏐ 23 may 2005 ⏐ x y z (3d) ⏐ 23 may 2005 ⏐ figure 4. velocity vector along the duct with staggered ribs for v = 10 m/s and hr = 7 mm 17 the journal of engineering research iii (2005) 10-18 (3d) ⏐ 23 m ay 200 5 ⏐ 20 20 20 30 303030 30 30 40 40 40 40 40 3030 20 20 zx y (3 d) ⏐ 23 may 2005 ⏐ 20 20 203030 3030 30 30 30 40 40 40 4040 40 20 20 zx y (3d) ⏐ 23 may 2005 ⏐ 20 20 20202030 30 30 30 30 30 4040 40 40 zx y(3d) ⏐ 23 may 2005 ⏐ figure 5. isotherm contours along duct with lower ribs at v = 2.5 m/s, a) hr = 3 mm, b) hr = 5 mm, c) hr = 7 mm figure 6. bulk temperature variation with duct length for different rib heights for v = 5 m/s and lower arrangement 3 mm 4 mm 5 mm 6 mm 7 mm 28 26 24 22 20 0.1 0.2 duct length (m) 0.3 0.4 b ul k te m pe ra tu re ( c ) 18 the journal of engineering research vol. 3, no.1 (2006) 10-18 the following future work is recommended: 1. studying further rib geometries like pitch and width. 2. experimental study of cold air through ribbed duct. references chang, s.w., su, l.m. and yang, t.l., 2004, "heat transfer in a swinging rectangular duct with two opposite walls roughened by 45º staggered ribs," int. j. of heat and mass transfer, vol. 44(11), pp. 287-305. gao, x. and sunden, b., 2001, "heat transfer and pressure drop measurements in rib-roughened rectangular ducts," experimental thermal and fluid science, vol. 24(1-2), pp. 25-34. han, j.c., 1988, "heat transfer and friction characteristics in rectangular channels with ribs turbulators," j. heat transfer, vol. 110. pp. 321-328. hadhrami, l. a. and chin han, j.c., 2003, "effect of rotation on heat transfer in two-pass square channels with five different orientations of 45o angle rib turbulator," int. j. of heat and mass transfer, vol. 46(4), pp. 653-669. kim, k. and kim, s., 2001, "numerical optimization of rib shape to enhance turbulent heat transfer," proceedings of the 2nd int. conference on computational heat and mass transfer. lee, c.k. and moneim, s.a., 2001, "computational analysis of heat transfer in turbulent flow past a horizontal surface with two dimensional ribs," int. communications in heat and mass transfer, vol. 28(2), pp. 161-170. onbasioglu, s. u. and huseyin onba, h., 2004, "on enhancement of heat transfer with ribs," applied thermal engineering, vol. 24(1), pp. 43-57. patanker, s. v., 1980, numerical heat transfer and fluid flow, mcgraw-hill, new york. raisee, m. and bolhasani, m.r., 2003, "computation of turbulent flow and heat transfer in passages with attached and detached rib array," dept. of mech. engi., faculty of engineering, university of tehran. taslim, m. e., li, t. and spring, s. d., 1998, "measurements of heat transfer coefficient and friction factors in passages rib-roughened on all walls," transactions of asme, vol.120, pp. 564-570. tanda, g., 2004, "heat transfer in rectangular channels with transverse and v-shaped broken ribs," int. j. of heat and mass transfer, vol. 47(2), pp. 229-243. figure 7. variation of nusselt number with duct length for lower arrangement and v = 10 m/s and different ribs height 36 34 32 30 28 26 24 22 20 18 16 n us se lt n um be r 10 20 30 40 50 60 duct length (m) the journal of engineering research (tjer), vol. 15, no. 2 (2018) 155-162 doi: 10.24200/tjer.vol15iss2p155-162 numerical assessment of electromagnetic radiated emissions from an uwb radiators located inside metallic enclosures with small air-flow openings m.m. bait-suwailm*, a. boudelliou, and n. tarhuni electrical and computer engineering department, college of engineering, sultan qaboos university, muscat sultanate of oman. received 31 may 2016; accepted 1 february 2018 abstract: shielding effectiveness of electronic devices and computing devices is essential at design stages. devices should pass electromagnetic interference/compatibility tests for certification purposes. in this paper, we investigate the electromagnetic radiation leakage from modern personal computers and mobile electronic devices due to an ultra-wide band source of noise. openings and slots are always unavoidable in electronic devices, for instance openings in computers for air-flow and heat dissipation. three different designs of enclosure’s covers with apertures are modeled for this purpose. the three covers cases comprise the use of rectangular arranged circular holes, hexagonal arranged circular holes, and square openings. for comparison purposes, a shielded enclosure with no openings is considered. the shielding effectiveness is numerically assessed here by capturing electric field strength that is modeled using probes located at two different distances, nearand far-away from the metallic enclosure. keywords: electromagnetic interference; m etallic enclosure; radiated emissions; shielding effectiveness; and uwb noise. التقييم العددي لالنبعاثات الكهرومغناطيسية املنبثقة من مشعات النطاق العريض الفائق اليت تقع داخل حاويات معدنية ذات فتحات صغرية لتدفق اهلواء ، ناصر الرتهوني، أمساء بودليوه *حممد بيت سويلم جيب أن جتتاز . حيثاألجهزة اإللكرتونية وأجهزة احلوسبة أمرا ضروريا جدا يف مراحل التصميمعد فعالية تدريع ُت: امللخص هذه األجهزة اختبارات التداخل / التوافق الكهرومغناطيسي من أجل اعتمادها. فنحن نقوم يف هذه الورقة البحثية، بفحص واألجهزة اإللكرتونية املتنقلة الناجتة عن مصدر تسرب اإلشعاع الكهرومغناطيسي من أجهزة احلاسوب الشخصية احلديثة ُيعد وجود فتحات ومداخل يف األجهزة اإللكرتونية أمرا ال ميكن جتنبه دائما، جمموعة كبرية جدا من املوجات املتداخلة ميم ثالثة مناذج ولقد عملنا على تص .فعلى سبيل املثال، يوجد فتحات يف أجهزة احلاسوب من أجل تدفق اهلواء وتبديد احلرارة وتشمل احلاالت الثالث هذه استخدام فتحات دائرية منسقة على شكل .خمتلفة من األغطية الضميمة للفتحات هلذا الغرض وألغراض املقارنة، عملنا على عمل علبة حممية بدون .مستطيل وفتحات دائرية منسقة على شكل سداسي وفتحات مربعة التدريع عدديا هنا من خالل نقاط شدة اجملال الكهربائي اليت يتم منذجتة باستخدام جمسات يتم تقييم فعالية حيث .فتحات تقع على مسافتني خمتلفتني: القريبة والبعيدة عن العلبة املعدنية : التداخل الكهرومغناطيسي، الضميمة املعدنية، االنبعاثات املشعة، فعالية التدريع، جمموعة كبريةالكلمات املفتاحية .جدا من املوجات املتداخلة * corresponding author’s e-mail: msuwailem@squ.edu.om m.m. bait-suwailm, a. boudelliou, and n. tarhuni 156 1. introduction in recent years, there has been a huge demand from society in technological advancement in computers and electronic systems packaging. however, the dramatic rise in operating frequency and increased speed of electronic devices, including personal computers has indeed led to an increase in the amount of electromagnetic radiated emissions and interference (emi) related issues. such encountered unintentional electromagnetic radiation from devices can be attributed to several sources of noise. for instance, the rapid state switching of high-speed output drivers, which result in the simultaneous switching noise (ssn) (ott h 1988; senthinathan r and john l 1994). as a consequence, the disturbance from such noise may result in sever performance and integrity degradation of high-speed printed circuit boards, and even microprocessors of personal computers and mobile laptop units. thus, it is essential to propose low-cost yet effective shielded enclosures. however, the integrity of metallic enclosures (chassis) is compromised by openings, which are needed for input/output transmission cables, air-flow and even heat dissipation. possible radiated emissions from such perforations have to be maintained as low as possible to comply with the electromagnetic interference (emi) radiated emissions regulations (senthinathan r and john l 1994; kubik z and skala j 2013). previous research works were carried out to investigate the effect of electromagnetic radiated emissions from high-speed printed circuit boards and enclosures with openings, for instance emi significance in personal computers due to narrowband sources of noise (baitsuwailam m, ramahi o 2012; bait-suwailam et. al 2014; bait-suwailam et. al 2015; cerri g et. al 1992; hao j et. al 2016; li m et. al 2000; lin l, ramahi o 2002; qin j et. al 2007; shahparnia s and ramahi o 2004; slyke t et. al 2008, among others). evaluation of electromagnetic field leakage from high-speed printed circuit boards was explored in (shahparnia s and ramahi o 2004; qin j et. al 2007; bait-suwailam m, ramahi o 2012), while leakage from enclosures with openings was evaluated and assessed in (li m et. al 2000; lin l, ramahi o 2002; bait-suwailam et. al 2014; bait-suwailam et. al 2015). however, significance of shielding enclosures due to unintentional ultra-wide band (uwb) sources of noise needs a revisit. recently, sudden growth in microprocessor’s clock speed to several gigahertz has been noted. the associated higherorder harmonics of the fundamental clock frequency of the microprocessor will also be a source of disturbance that can easily couple (talk) to openings with same resonance frequency. hence, this will result in significant electromagnetic radiation from the enclosure’s apertures behaving as a secondary source of radiation. this paper presents numerical assessments of the emi radiation leakage from airflow openings in personal computers metallic enclosures, due to uwb sources of noise in order to quantify the shielding effectiveness of such enclosures as well as higher-order harmonics of the uwb sources of noise, that may constructively add up to existing enclosure's resonances and harmonics. hence, the system would become densely complex and can result in severe performance degradation to receivers whenever placed nearby the enclosure when placing uwb source of noise inside the enclosure. three different cases of slots arrangement mimicking airflow openings that are found in many personal computers brands are numerically assessed using the full-wave solver of cst mw studio (computer simulation technology 2014). numerical results are computed and compared against a reference case comprising a sealed shielded enclosure (with no openings) at both nearand far-field regions. an ultra-wideband low-profile antenna is considered here to mimic an uwb source of noise that resides inside the metallic enclosure. 2. modeling and analysis an intentional uwb refers to a radio technology with a bandwidth exceeding 500 mhz or a fractional bandwidth greater than 20% of the center frequency, according to the u.s. federal communications commission (fcc) (feral communications commission report). technically, uwb devices operate by employing short duration pulses that then result in a wide transmission bandwidth. due to the additionally posed emi emissions limits, by the fcc, in some applications within uwb regime, it is then not surprising to expect emi related problems when placing sensitive receivers in close proximity to an uwb source of noise. hence, proper shielding of devices, including enclosures of personal computers and alike, is very essential. as can be seen from fig. 1, uwb emitters must be kept below the allowable emission mask limit of 41.3 dbm/hz. figure 2 shows the proposed model to investigate the emi threats influenced on sensitive devices, that may represent high-speed printed circuit boards or memory card that are numerical assessment of electromagnetic radiated emissions from an uwb radiators located inside metallic enclosures with small air-flow openings 159 placed inside personal computers. for convenience, an uwb source of noise is modeled here as a circular microstrip patch antenna, where the antenna is fed by a 50  microstrip transmission feed line that is fed through a waveguide port as shown in fig. 3 (liang j et. al 2005). the antenna is printed on an fr-4 substrate with relative permittivity of 4.7, loss tangent of 0.02, thickness of 1.5 mm and board size of l = 50 mm, and w = 42 mm. the metallic patch has a circular shape with a radius, r = 10 mm. the ground plane length is 20 mm and the horizontal gap from the edge of the circular radiator to ground plate is 0.3 mm. the antenna resonates quite well below -10 db over the whole uwb frequency spectrum. figure 4 shows the reflection coffeciont of the modeled uwb antenna, that will represent the source of noise in this study. good antenna impedance matching can be seen, where |s11|  10 db for the operating uwb band, covering 3.1 ghz – 10.6 ghz. due to the limited in-house laboratory resources to only narrowband measurement equipment, modeling scenarios are numerically investigated by means of cst microwave studio. furtheremore, measurement validations over narrowband frequency spectrum were reported in (bait-suwailam m et. al 2015). in all studied scenarios, a perfectly matched layer, which is an absorbing boundary, is placed around the numerical computational domain that surrounds the enclosure and probes in order to avoid any reflections from the truncated space of the modeled cases. note that the computational domain is relatively large along the y-axis (see fig. 2), where it extends to 3.5 meters long in order to accommodate the farfield probe, which is necessary to obtain the radiated emissions from the enclosure. an adaptive mesh is considered to provide accurate measures of the electric field strength that is expected to leak out from the openings. the electric field leakage radiated out from the modeled enclosure through the aforementioned openings cases is captured in near and far fields from the enclosure surface. note that several probes were located outside the enclosure: two at near field region (20 mm and 90 mm) and the other two are far-distant from the cover’s openings (1and 3 meters). among others, two of the factors that can lead to significant em leakage in metallic enclosures are highly resonant modes of metallic enclosure (behaves as a cavity), and unavoidable slots within the enclosure's surface. the possible modes (resonant frequencies) inside an unloaded cavity with conducting boundaries (ie. metallic enclosure here) can be analytically calculated using the following well-known formula (pozar d 2011): figure 1. the fcc uwb power spectral density mask. figure 2. a 3d view of the modeled metallic enclosure with the uwb source of noise placed inside the enclosure. a cover case with hexagonal circular slots is also shown. figure 3. an uwb circular microstrip antenna modeled as a source of noise. circular microstrip 157 m.m. bait-suwailm, a. boudelliou, and n. tarhuni 160 fres = 1 2√𝜀𝜇 √( 𝑚 𝑙 )2 + ( 𝑛 𝑙 )2 + ( 𝑝 ℎ )2, (1) where  is the electric permittivity,  is the magnetic permeability, m, n, p are positive integers representing the possible propagating modes inside the metallic enclosure, and w, l and h represent the width, length and height of the metallic enclosure along the y, x, z respectively. the first 12 resonant modes (above 3.0 ghz) inside the metallic enclosure are displayed in table 1. some degenerate resonant modes are expected to exist within the metallic cavity. in order to assess the shielding effectiveness of the metallic enclosure in presence of slots, electric field probes are used in order to monitor and quantify the amount of the electric field leakage through the openings of the enclosure. in this study, we focus on near-field and far-field strengths, specifically; probes are monitored at 20 mm in near-field and far-distant from the enclosure in far-field region. the shielding enclosure considered in this work has been scaled down to an acceptable size that will still maintain sufficient usage of memory through the numerical simulations runs. the enclosure width, w = 20 cm, length, l =10 cm, and height, h =10 cm, and metal thickness of 5 mm, where w, l and h represent the width, length and height of the enclosure along the y, x, z, respectively. the enclosure is made of a perfect electric conducting material to save up memory and the uwb source of noise (antenna) was placed 25 mm away from the metallic ground of the enclosure along the z-direction. the antenna was excited with a discrete lumped port and the simulations were obtained using the transient analysis, time solver of cst mws. a reference figure 4. simulated reflection coefficient, |s11|, of the uwb antenna. table 1. analytical resonant modes inside an unloaded metallic enclosure. note that frequencies have been selected. mode number resonant frequency (ghz) 1 3.09 2 3.35 3 3.44 4 3.44 5 3.67 6 3.67 7 4.31 8 4.50 9 4.80 10 4.97 11 5.46 12 5.61 (a) (b) (c) figure 5. side view representation of the studied enclosure with various airflow openings: (a) rectangular arranged circular holes cover (d1=16 mm, d2=6 mm and d3=4 mm), (b) hexagonal arranged circular holes cover (d4=6 mm, d5=6 mm and d6=4 mm), and (c) square openings cover (d7=16 mm, d8=6 mm and d9=4 mm). 158 numerical assessment of electromagnetic radiated emissions from an uwb radiators located inside metallic enclosures with small air-flow openings 159 case representing a solid enclosure without any openings was considered for comparison with other proposed scenarios. as shown in fig. 5, the suggested enclosure's openings of the four cases representing the enclosure’s airflow openings are: rectangular-arranged circular holes cover (case 1), hexagonal arranged circular holes cover (case 2) and square openings cover (case 3). 3. results and discussion this section presents numerical results from frequency domain analysis. in order to provide a baseline for this comparative study of enclosure’s openings, the electric field leakage from a shielded enclosure with no openings is captured as shown in fig. 6. in both near and far field regime, the magnitude of the electric field was below 180 db, sufficient to quantify that there is no emi leakage from the metallic enclosure. field probes, that are located outside of the metallic enclosure, are used to capture the amount of field leakage through openings from the enclosure over the uwb frequency spectrum. next, the radiated emissions from the suggested openings in this study are numerically assessed as shown in fig. 7. the electric field strength from opening cover of case 3 (square openings cover) provides less shielding effectiveness as compared to the other two cases. in other words, enclosure's resonances with cover of case 3 were seen high enough to reach above -25 db on average as compared to the other two cases. on the other hand, case 1 (rectangular arranged circular holes cover) provides the best shielding effectiveness of 30 db on average. furthermore, as can be seen from table 1, the shaded resonances agree quite well with the resonant peaks of the captured electric field strength as shown in both figs. 7, 8 and 9. figure 6. the electric field captured by the probes for the shielded enclosure with a solid cover (reference case). figure 7. the electric field captured by the probes for the three cases at 20 mm distance away from the enclosure’s opening. figure 8. the electric field captured by the probes for the three cases at 90 mm distance away from the enclosure’s opening. figure 9. the electric field captured by the probes for the three cases at 1 meter distance away from the enclosure’s opening. m.m. bait-suwailm, a. boudelliou, and n. tarhuni 160 when placing the probe far distant away from the enclosure’s openings, the emi radiation leakage from the openings, as expected, starts to decay for all the three cases as shown in fig. 8. the outputs are closer to each other; however, the first case of rectangular arranged circular holes cover provides the best emi shielding effectiveness of less than 50 db on average in comparison to the other two cases at 90 mm as well as 1 meter (see fig. 9) away from the enclosure's opening. next, it is important to study the electromagnetic field radiated emissions from the uwb source of noise that is placed inside the studied metallic enclosure, mimicking a personal computer, with various openings. in fact, radiated emission requirements for uwb systems are already set clearly in code of federal regulations, telecommunications, which helps in setting up guidelines for maximal allowable electromagnetic radiated emissions from uwb source of noises (federal communications commission revision report). the federal communications commission (fcc) regulations for radio frequency equipment defines clearly power and radiation limits in terms of electric field strength measured in volts/meter at a distance of 3 meter from the source of disturbance. in this study, to calculate the em radiated emissions, the electromagnetic field strength at 3 meter away from the enclosure's opening is computed. in other words, broadband far-field monitors with sufficient number of discrete frequencies are evaluated, after which, radiated emissions are numerically computed using (sanders f 2001): |e|(dbv/m) = pobserved(dbm) + 77.2 db + 20*log10(f) – gant. (2) where pobserved is the power recorded at 3 meter away from the source in dbm, f is the operating frequency in mhz, and gant. is the peak gain of the antenna in db, while the constant 77.2 db is a correction factor. figure 10 depicts the electric field strength that is captured in far-field, at 3 meters away from the enclosure. as can be seen from the figure, within the uwb frequency regime, 3.1 ghz – 10.6 ghz, the enclosure's opening with circular holes that are arranged in rectangular fashion, case 1, shows the minimal electromagnetic radiated emission in comparison to the other two cases. it is also observed that hexagonal circular holes show better shielding performance as compared to the case of rectangular holes (case 3, see fig. 3c), except around 8 ghz, where radiated emission was high reaching 53 dbv/m. furthermore, significant radiated electromagnetic emission from case 3 is observed within the upper frequency band of the uwb regime, roughly from 5.5 ghz – 6.5 ghz as well as above 10 ghz. numerical results from figure 10 correlates well with both figs. 7, 8 and 9. table 2 presents quantitative measures of electromagnetic radiated emissions at 3 meters distance away from the perforated enclosure for 3 selected frequencies within the uwb operating regime of the source of noise. it is observed that at 10 ghz, case 2 provides slightly better shielding performance than case 3. overall, enclosure with cover of case 1 performs better than the other two cases. it is clear from figure 9 that the fcc class b emi limit of radiated emissions is around 54 dbv/m, in which all figure 10. the electric field strength captured in far-field at 3 meters. the recommended fcc mask is also shown, where class b refers to the standard applied to non-residential use of personal computers/portable laptops. table 2. comparison of electromagnetic radiated emissions from the perforated metallic enclosure for the studies scenarios at an observation point placed 3 meters away from the enclosure. frequency = 4 ghz case type radiated emission strength (in dbv/m) case 1 35.78 case 2 40.42 case 3 43.54 frequency = 7 ghz case 1 32.20 case 2 32.76 case 3 42.30 frequency = 10 ghz case 1 41.16 case 2 46.60 case 3 50.36 numerical assessment of electromagnetic radiated emissions from an uwb radiators located inside metallic enclosures with small air-flow openings 159 studied perforated enclosures had maintained their shielding strength below the mask limit, except at few discrete frequencies. 4. conclusion in this paper, a comprehensive study of the emi effectiveness from different airflow openings scenarios due to an ultra-wide band source of noise is conducted. the amount of electromagnetic field leakage from the enclosure's opening was numerically assessed using electric field probes that monitor field strength at locations within nearand far-field region from the enclosure's opening. three different scenarios of openings for continuous air-flow are considered. an uwb source of noise is modeled and presented here using a circular microstrip patch antenna that is located inside the enclosure. based on full-wave numerical assessments, results show that the rectangular arranged circular holes cover case provided better shielding effectiveness and a reduction of -5 db in the amount of electric filed strength as compared to other proposed cases. conflict of interest the authors declare no conflicts of interest. funding this work is financial supported by internal research grant (ig/eng/eced/14-01). acknowledgment the authors would like to thank sultan qaboos university, oman for providing the financial resources needed for this work. references bait-suwailam m, ramahi o (2012), ultrawideband mitigation of simultaneous switching noise and emi reduction in highspeed pcbs using complementary split-ring resonators. ieee transactions on electromagnetic compatibility 54(2): 389– 396 bait-suwailam m, alavikia b, ramahi o (2014), reduction of electromagnetic radiation from apertures and enclosures using electromagnetic bandgap strucutres. ieee transactions on components, packaging and manufacturing technology 4(5): 929-937. bait-suwailam m, al-abri d, teirab a, jervase j, nadir z (2015), electromagnatic interference (emi) radiation from airflow openings in personal computers shielding enclosures: an experemental study. in the 8th ieee gcc conference & exhibition 1-4. cerri g, de leo r, primiani v (1992), theoretical and experimental evaluation of the electromagnetic radiation from apertures in shielded enclosure. ieee transactions on electromagnetic compatibility 34(4): 423–432. computer simulation technology (cst), microwave studio (mws), 2014 [online]. available: http://www.cst.com. federal communications commission report, fcc 02-48 [online]. available: https://apps.fcc.gov/edocs_public/attachm atch/fcc-02-48a1.pdf federal communications commission revision report (2002), revision of part 15 of the commission's rules regarding ultrawideband transmission systems. et docket 98-153. hao j, jiang l, gong y, and fan j (2016), study of the shielding effectiveness of double rectangular enclosures with apertures excited by an internal source. progress in electromagnetics research m, (47): 67-76. kubik z, and skala j (2013), influence of the cavity resonance on shielding effectiveness of perforated shielding boxes. in the proceedings of the 8th international conference on compatibility and power electronics 260– 263. li m, nuebel j, drewniak j, dubroff r, hubing t, van doren t (2000), emi from airflow aperture arrays in shielding enclosuresexperiments, fdtd, and mom modeling. ieee transactions on electromagnetic compatibility 42(3): 265–275. liang j, chiau c, chen x, parini c (2005), study of a printed circular disc monopole antenna for uwb systems. ieee trans. on antennas and propagation 53(11): 3500-3504. lin l, ramahi o (2002), analysis and reduction of electromagnetic field leakage through loaded apertures. in the 2002 international ieee antennas and propagation society symposium 102-105. ott h (1988), noise reduction techniques in electronic systems, 2nd edition, john wiley and sons inc. pozar d (2011), microwave engineering., 4th edition, john wiley & sons, inc., 284-287. qin j, ramahi o, granatstein v (2007), novel planar electromagnetic bandgap structures for mitigation of switching noise and emi reduction in high-speed circuits. ieee 161 http://www.cst.com/ https://apps.fcc.gov/edocs_public/attachmatch/fcc-02-48a1.pdf https://apps.fcc.gov/edocs_public/attachmatch/fcc-02-48a1.pdf m.m. bait-suwailm, a. boudelliou, and n. tarhuni 160 transactions on electromagnetic compatibility 49(3): 661– 669. sanders f (2001), conversion of power measured in a circuit to incident field strength and incident power density, and corrections to measured emission spectra for non-constant aperture measurement antennas. appendix c of ntia report 01-383. senthinathan r, john l (1994), simultaneous switching noise of cmos devices and systems, norwell, ma: kluwer. shahparnia s, ramahi o (2004), electromagnetic interference (emi) reductionfrom printed circuit boards (pcb) using electromagnetic bandgapstructures. ieee transactions on electromagnetic compatibility 46(4): 580– 587. slyke t, kuhn w, natarajan b (2008), measuring interference from a uwb transmitter in the gps l1 band. in the ieee radio and wireless symposium 887-890. 162 civl50604.qxd the journal of engineering research vol. 4, no.1 (2007) 1-10 1. introduction a lot of money is spent each year in maintenance and rehabilitation of roads, where most of the road problems are due to fatigue cracking and rutting (muniandy, et al. 2002). however, a lack of specialized human expertise in pavement design (goh, 1993) and as a result of inexperienced road design practitioners who lack pavement knowledge and assessment and exposure to some other pavement materials often result in the inappropriate selection of road materials for construction. moreover, the road design standards of manual of pavement design by public works department of jkr arahan teknik 5/85 is not updated or improved consistently (jkr, 1986). due to the previous distresses experienced, jkr's standard and methodology should be reviewed (emby and mustafa, ____________________________________________ *corresponding author’s e-mail: hsalih@eng.upm.edu.my 1996). in view of these problems, a study was undertaken with the aid of intensified research in priority areas (irpa) funding to develop a prototype expert system for pavement materials selection in order to assist the pavement designer and contractors in selecting pavement materials alternatives. the system incorporates the properties of soil along with the selected road alignments, where pavement materials are severely over stressed. the type and soil conditions, and expected load repetitions are other parameters found to be indispensable in the formulation of a material selection system for road construction. the material performance and the selection of the appropriate construction materials for the various layers under certain traffic loading, and under different types of terrains such as flat, rolling and mountainous were incorporated as important parameters. such conditions govern the duration of axle loading on the pavement. a prototype expert system for the selection of road construction materials s. hassim*, k.t. teh1, r. muniandy1, h. omar1, and a. hassan2 *1civil engineering department, universiti putra malaysia, p.o. box 43400, serdang, selangor d.e., malaysia 2kelantan public work department, p.o. box 15990, kota bharu, kelantan, malaysia received 5 june 2004; accepted 29 january 2006 abstract: a prototype for an expert system in road construction material selection system, which is based on the outcomes of friedman and multiple comparisons statistical methods was developed. the outcomes were acquired through questionnaires from selected pavement experts. the factors affecting pavement materials under each particular site condition were incorporated into the specific rules of the system. the system knowledge-base was extracted from the statistical testing outcomes and then rearranged and compiled prior to the development of the system. visual basic 6.0 was adopted as the programming tool for development of the system, while the knowledge-base of the separate system was kept in microsoft access 2000. the prototype expert system can be used to emulate part of the professional reasoning capabilities based on the knowledge of a pavement expert or a specialist to solve problems on materials selection. the system can help road designers to improve their professional ability to evaluate all available materials even before carrying out any laboratory tests. keywords: material selection, expert system, pavement, road construction ¥ô£dg aé°ûffg ogƒe qé«àn’ òñn ωé¶f êppƒ‰ ø°ùm ,…oé«fƒe ,— ,º«°ùm áá°°uuóóÿÿggèféàædg ™ªl ” ~bh .io~©àÿg á«fé°üm’g äéfqé≤ÿg ¥ôwh ¿ée~jôa èféàf ≈∏y ~ªà©j …òdgh ¥ô£dg aé°ûfg ‘ áe~îà°ùÿg ogƒÿg qé«àn’ òñn ωé¶f êpƒ‰ ôjƒ£j ” : â«æh ~bh .ωé¶æ∏d io~ùg ~ygƒ≤dg ‘ ™bƒe πµh á°uéÿg ∞«°uôdg ogƒe ≈∏y ôkƒj »àdg πegƒ©dg úª° j ”h .¥ô£dg ∫é› ‘ ú°üàıg øe o~y øe ⩪l äéfé«ñà°sg ωg~îà°séh ωé¶ædg ôjƒ£àd6^0 ∂°ù«h ∫gƒé«a ‹’g ö°sé◊g á›ôh á¨d ωg~îà°sg ”h .ωé¶ædg ôjƒ£j πñb ⩪lh á«fék âñjq ∂dp ~©hh á«fé°üm’g qéñàn’g èféàf ≈∏y ωé¶ædg á«aô©ÿg i~yé≤dg »≤£æÿg òµøàdg ≈∏y iq~≤dg øe aõl iécéfi ‘ √ôjƒ£j ” …òdg òñÿg ωé¶ædg êpƒ‰ ωg~îà°sg øµáh .2000 ¢ù°ùcg âaƒ°shôµjée ωé¶f ‘ á«aô©ÿg äéfé«ñdg i~yéb â¶øm éªæ«h ogƒÿg πc º««≤j ‘ á«æ¡ÿg º¡jgq~b ú°ù– ‘ ¥ô£dg »ªª°üe ωé¶ædg ~yé°ùj ¿g øµáh .ogƒÿg qé«àng πcé°ûe π◊ ∞«°uôdg »°ué°üàng hg agèn áaô©e ≈∏y ~æà°ùÿg ±îùg .á«∏ª©ÿg ägqéñàn’g agôlg πñb ≈àm iôaƒàÿg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.¥ô£dg aé°ûfg ,∞«°uôdg ,òñÿg ωé¶ædg ,ogƒÿg qé«àng : 2 the journal of engineering research vol. 4, no.1 (2007) 1-10 the objective of this study is to develop a pavement material selection system and to incorporate it into the development of the road construction material selection system (rc-mss). the materials selected and positions for a particular site condition were obtained via a questionnaire sent to selected pavement experts. the friedman test was adopted to analyze the questionnaire output. rejected null hypothesis tests were then analyzed using multiple comparisons and ranking for each material under each rejected null hypothesis site condition was done. an expert system is a computer system that emulates the decision ability of a human expert, which is capable of acting in all aspects like a specialist to provide specialized knowledge to solve problems at the level of a human expert. figure 1 shows the basic concept of an expert system function. expert systems today are applied to a variety of fields such as medicine, science, engineering, and business (omar, 2001). 2. road construction material selection system (rc-mss) framework the road construction material selection system (rcmss) was developed based on the survey results obtained from road practitioners, and selected pavement experts. in the surveys to the selected pavement experts, factors and ranges of each category suggested, and ratings were identified each suggested material at each particular site conditions. it is a rating based on the durability, performance and ease of construction at each particular site conditions. materials rating was then determined using multiple comparisons of each suggested material to see if it is significant to reject null hypothesis of differential in durability, performance and ease of construction in each site condition at 95% significant level using the friedman statistical test method. same position was assigned to the site conditions that it is not significant to reject null hypothesis at 95% significant level using the friedman statistical test method. it means that all the materials suggested at that particular site condition have the same performance. the rc-mss framework was developed before incorporating it in the development of the expert system. appendix i shows the position of each suggested pavement materials by the experts in site condition 10 of poor soil conditions regardless of traffic loading and the critical area. 3. factors considered in the system this study looked into the anticipated traffic loading, soil conditions, type of terrain and geometric features which were considered in the system as the factors in the development of pavement materials selection. each factor considered in the system was divided into each particular category, which was obtained via documented sources and through interviews with the selected pavement experts. several interviews were conducted with selected pavement experts to assure that the range of each particular category in each factor is appropriate. 3.1 anticipated traffic loading conditions commercial traffic loading was used in the system because pavement starts to deteriorate once it opens to traffic. failure is caused by heavy axle loads on poor pavement materials. the system incorporates the anticipated equivalent single axle load (esal) as an input factor and it is classified into three categories as illustrated in table 1. 3.2 soil conditions soil condition is another factor considered in pavement design and materials selection. the strength of a road subgrade is commonly assessed in terms of the california bearing ratio (cbr). a higher value of cbr for a soil would be suitable for use as a subgrade. a cbr value of less than 5% should be considered a poor subgrade soil. the system will use the cbr value as an input data. the data are then classified into three categories as shown in table 2 (huang, 1993; lay, 1990). 3.3 geometric features climbing lanes and sharp horizontal curves where multiple heavily loaded vehicles firmly brake, travel at slow speeds or remain standing for periods of time, induce the pavement to deteriorate at a faster rate due to high stresses (harun, 1992). however, these factors are only considered in geometric design and are less of concern in structural development design. hence, geometric features knowledge base inference engine fact experts user figure 1. basic concept of an expert system function layout (omar, 2001) traffic loading low moderate high anticipated esal < 300,000 300,000 – 10,000,000 > 10,000,000 denoted symbol e1 e2 e3 table 1. classification of anticipated esal for a 20year design life soil condition good fair poor cbr, % > 10 5 – 10 ≤ 5 denoted symbol s1 s2 s3 table 2. soil cbr value divided into three categories 3 the journal of engineering research vol. 4, no.1 (2007) 1-10 such as vertical gradient and degree of horizontal curvature were considered in the system and classified into two categories of critical and non-critical areas as presented in table 3. geometric features of horizontal curvature group and vertical gradient group were regrouped into non-critical areas (assigned symbol is c1) and critical areas (assigned symbol is c2) as presented in fig. 2. for a non-critical vertical gradient, the degree of vertical gradient measured longitudinally is less than 5%. it is established as a relationship between vertical rise or fall for a unit of horizontal distance. while, a non-critical situation for a horizontal curve is defined as the subtending a 100-foot arc along the curve is less than seven degrees. a critical area means that the area is highly stressed, where multiple heavily loaded vehicles firmly brake, travel at slow speeds or remain standing for periods of time causing the pavement to deteriorate faster. a real topographic profile such as type of terrain is crucial in pavement design and it is generally divided into three groups: flat, rolling and mountainous which is determined by the percentage of longitudinal slope, as shown in table 4 (jkr, 1986). the terrain type is incorporated into the system to determine the type of material suitable for the stretch. 4. system database and denoted symbols from the questionnaire given to the pavement experts, the ranges of each factor in each category were divided into 11 design conditions as shown in table 5. the 11 design conditions were given designated symbols. the database for design condition '0', was obtained via a documented source. design condition '10' is used for poor soil conditions and is considered critical regardless of the anticipated esal and geometric features. each design condition is then separated into three more conditions for flat, rolling and mountainous terrains, resulting in a total of 33 site conditions. the rc-mss framework is shown in fig. 3. design conditions consisted of anticipated esal group (denoted as e), soil cbr group (denoted as s) and critical area group (denoted as c) as shown in fig. 3. these design conditions would then follow the path indicated by the symbol e, s, c to m. design conditions consisting of symbol s3 (poor soil group), will go to the next station of a treated subgrade (tsg block), which is the adopted measure for poor soils. poor soils need to be treated before proceeding to the next step. the modified design conditions after tsg would be considered equivalent to a good soil group as denoted by the symbol s1 before proceeding to the c group to reach materials suggested database denoted as m. if the previous design condition of symbol designated as tsg block, the database to reach is "material suggested database" with asterisk (denoted symbol is 'm*'). block symbol tsg was extracted from design condition '10'. the database of 'm*' has the same suggested pavement materials for subbase, base course and surface course as database of m except for the subgrade where 'm*' adopted the material from block symbol tsg. 5. example problem a simple example was run to illustrate the framework of rc-mss. the site conditions were as follows.: site conditions: geometric feature non-critical critical vertical gradient < 5 % ≥ 5 % denoted symbol (vertical gradient) v1 v2 degree of horizontal curve ≤ 7º > 7º denoted symbol (horizontal curve) h1 h2 table 3. geometric features classification h2h1 v2 v1 c1 c2 c horizontal curvature vertical gradient figure 2. regrouping of horizontal curvature and vertical gradient groups to critical and non-critical groups note: symbol ‘c’ denoted as a critical condition type of terrain flat rolling mountainous cross slope ≤ 3 % 3 – 25 % ≥ 25 % denoted symbol t1 t2 t3 table 4. terrain type • anticipated esal = 12,500,000 of 18 kip axle repetitions • soil cbr = 3% • vertical gradient = 3% • horizontal curvature = 8o • type of terrain = flat 4 the journal of engineering research vol. 4, no.1 (2007) 1-10 soil anticipated esal geometric features database symbol d es ig n c on di tio n c on di tio n sy m bo l c on di tio n sy m bo l c on di tio n sy m bo l good s1 low e1 non-critical area c1 good s1 moderate e2 non-critical c1 m1 0 fair s2 low e1 non-critical c1 poor s3 low e1 non-critical area c1 m1* 0* poor s3 moderate e2 non-critical c1 m2 1 good s1 high e3 non-critical area c1 m2* 1* poor s3 high e3 non-critical area c1 m3 2 fair s2 moderate e2 non-critical area c1 m4 3 fair s2 high e3 non-critical area c1 m5 4 good s1 low e1 critical area c2 m5* 4* poor s3 low e1 critical area c2 m6 5 good s1 moderate e2 critical area c2 m6* 5* poor s3 moderate e2 critical area c2 m7 6 good s1 high e3 critical area c2 m7* 6* poor s3 high e3 critical area c2 m8 7 fair s2 low e1 critical area c2 m9 8 fair s2 moderate e2 critical area c2 m10 9 fair s2 high e3 critical area c2 tsg 10 poor s3 ⎯ ⎯ ⎯ ⎯ note: 1. *poor soil condition 2. dense-graded mix will remain the same for a binder course table 5. designed site conditions and denoted symbols rc-mss framework e1 e2 e3 tsg m 1, m 1* m 6 m 6* m 3 m 9 m 2, m 2* m 7, m 7* m 4 m 10 m 5, m 5* m 1 m 8 m 1, m 1* s1 s2 s1 s3 s2 s3 s1 s2 c1 c1 c1 c1 c1 c1 c2 c2 c2 c2 c2 s3 c2 note: symbol ‘m’ denoted as materials suggested database, and * represented material suggested database directed from the path from denoted symbol ‘tsg’. figure 3. direction path of design conditions to suggested materials database 5 the journal of engineering research vol. 4, no.1 (2007) 1-10 every site condition was classified into each group as shown in table 6. the geometric feature of the vertical gradient was non-critical while the horizontal curvature was critical. as mentioned earlier in fig. 2, once either one of the geometric features of vertical gradient or horizontal curvature is denoted as 'v2' or 'h2', it would be considered as a critical region. therefore, the example has been classified as symbol 'c2' of a critical area due to h2. also the soil condition was classified as poor which was denoted by the symbol 's3'. as discussed in the previous section of rc-mss framework, the designer would enter the station denoted symbol 'tsg' then exit to station s1 before reaching the 7th asterisk "material suggested database" denoted as symbol 'm7*'. the whole flow diagram is illustrated in fig. 3, which is given a thicker plotted line. in the asterisk "materials suggested database", the database of pavement materials suggested for each layer would then refer to the flat terrain section denoted by symbol 't1', except for the subgrade material which would be referred to the station denoted by symbol 'tsg'. the flow path can be illustrated in fig. 4 and the suggested materials are summarized in table 7. no rank is assigned to the database 'm7*' at 't1', where these materials are equal in performance, durability and ease of construction under these site conditions. the results were analyzed using friedman statistical method. table 7present a summary of the data obtained for the example problem. 6. development of a prototype expert system microsoft visual basic 6.0 (vb6) was selected as a tool due to the simplicity of the program and its ability to provide a development of interface (deitel, et al. 1999). the system framework was then translated into a numerical form or specific rule, which was incorporated into the development of an expert system of rc-mss. 6.1 specific rules based on the data collected from the experts and documented information, the results of the studies were translated into seven sets of rules: the system start with the data available in section nos. 1, 2, 3 and 4 and use the inference rules. an inference engine searches the inference rules until it matches the ifclause which is known to be true. it concludes with the then-clause and the material suggested data available will determine which inference rules are used as shown in section no. 5. some data provide more than one material, so the user would select the material, and the selected one will display in section no. 6. the user can also proceed to section no. 7 for materials cost calculation for the selected ones. the inference rules in each category will be shown in each section as discussed below. in section 1, the designed traffic loading volume (esal), was divided into three categories, low, moderate and high, denoted as e1, e2 and e3, respectively. design condition classification denoted symbol anticipated esal high e3 soil condition poor s3 vertical gradient non-critical v1 horizontal curvature critical h2 type of terrain flat t1 table 6. classification of design condition in example problem e3 s3 tsg c2 s1 m7* m7* referred to t1 with subgrade referred to tsg. figure 4. whole flow path diagram in example problem subgrade, sg subbase, sb base, bc binder, sbc surface, sc adopted measured denoted symbol (positions ) type of material denoted symbol type of material denoted symbol type of material denoted symbol type of material denoted symbol sb1 bc1 sbc sc1 sb1 bc1 sbc sc2 sb1 bc1 sbc sc3 sb1 bc2 sbc sc1 sb1 bc2 sbc sc2 sb1 bc2 sbc sc3 tsg3 (1); tsg5 (2); tsg1 (3); tsg2 (3), and tsg4 (3) sb1 bc4 sbc sc3 table 7. example problem summary using database ‘m7*’ and ‘tsg’ at ‘t1’ • rule no.1: designed traffic loading volume (esal) • rule no. 2 : soil condition • rule no. 3 : highly stressed area • rule no. 4: type of terrain • rule no. 5: pavement materials suggested • rule no. 6: pavement material s elected • rule no. 7: material cost calculation 6 the journal of engineering research vol. 4, no.1 (2007) 1-10 the designed esal was converted into expert system rules in the form as stated below: rule 1: if designed esal <300,000, then the traffic volume is low (e1). rule 2: if designed esal is between 300,000 and 10,000,000, then the traffic volume is moderate (e2). rule 3: if designed esal >10,000,000, then the traffic volume is high (e3). the soil conditions in section 2 is based on cbr values. from framework rc-mss, cbr values were divided into three categories: good, fair and poor soils. each of the categories are given symbols s1, s2, and s3, respectively. the cbr value is then converted into three rules. the three rules are: rule 1: if cbr >10%, then the soil is good (s1). rule 2: if cbr value is more than 5 % and less than 10 %, then the soil is fair (s2). rule 3: if cbr <5%, then the soil is poor (s3). highly stressed areas (section 3) are divided into two groups, which are vertical gradient and horizontal curve (refer to table 3). each of these groups is further divided into two conditions: highly stressed and not highly stressed regions. v1 and h1 represent the symbol of non highly-stressed areas and v2 and h2 represent the symbols of highly-stressed areas for vertical gradient and horizontal curve, respectively. two rules were used in defining vertical gradient conditions as shown below: similar to vertical gradient, horizontal curve was also divided into two rules as presented below: the two stressed area conditions are then combined and converted into four combinations of rules. the four rules are: input data section no. 1 design traffic loading (esal) section no. 5 pavement materials suggested section no. 3 highly stressed areas section no. 2 soil conditions section no. 6 pavement materials selected section no. 4 type of terrain section no. 7 materials cost calculation figure 5. rules sets rule 1: if vertical gradient < 5%, then it is a non highly stressed area (v1). rule 2: if vertical gradient ≥ 5%, then it is a highly stressed area (v2). rule 1: if horizontal curvature < 7 º, then it is a nonhighly stressed area (h1). rule 2: if horizontal curvature ≥ 7 º, then it is a highly stressed area (h2). 7 the journal of engineering research vol. 4, no.1 (2007) 1-10 rule 1: if vertical gradient is v1 and horizontal curvature is h1, then it is non-highly stressed area which is c1. rule 2: if vertical gradient is v1 and horizontal curvature is h2, then it is a highly stressed area which is c2. rule 3: if vertical gradient is v2 and horizontal curve is h1, then it is a highly stressed area which is c2. rule 4: if vertical gradient is v2 and horizontal curve is h2, then it is a highly stressed area which is c2. type of terrain in section 4 is divided into three types: flat, rolling and mountainous as follows: rule 1: if a flat terrain is selected, then it is defined as t1. rule 2: if a rolling terrain is selected, then it is defined as t2. rule 3: if a mountainous terrain is selected, then it is defined as t3. once the four rules are selected, matching of each rule under particular design conditions would then be entered to match the database stored in microsoft access 2000 as shown in section 5. the matched design conditions to the particular database of suggested pavement materials would then be extracted as mentioned in section 3, which is suggested materials. when the desired material is selected, material ranking would then be displayed in the form as shown in section 6. the last section of materials cost calculation (section 7) is an optional part, in case the user wants the system to calculate the materials cost for the selected pavement materials in each layer used in a particular project. 6.2 databases the database is the place where all the data are recorded and kept in the expert system. the reasons for setting databases are to store huge data into an inventory and to facilitate its use for future references; checking the findings of the es; editing data; updating; adding; or even deleting the data. 6.3 development of rc-mss rc-mss was developed into an expert system, es, for a pavement materials selection system. it was compiled and executed from an executable file. figure 6 shows the flow chart of rc-mss. it is currently a first version road materials selection system (rc-mss) for a new flexible pavement construction. the major task in building the knowledge-based system is to acquire and encode the expertise and knowledge of experts into the knowledgebase. rc-mss consists of two main components: the rcmss vb6.0 file and rc-mss database file. the first component was the system rules, preset commands and rule execution to reach the database. it is the shell of the system. rc-mss database file is located in microsoft access 2000. this program can associate with vb6.0 where both programs operate under a microsoft operating system. all the materials suggested in addition to the matching group from the input (design data) would be stored in rc-mss database file. output result: suggested pavement materials and selected pavement materials once the input data information are completed, the specific rules in each input data would be sorted by an inference engine to match the knowledge-base. the process is similar to the framework of the rc-mss shown in the fig. 3. some suggested materials were ranked by the system to show the difference in performance in that particular input data condition. figure 6b in appendix ii shows the selected pavement materials by the user. materials costs calculation rc-mss also provides a rough overall materials costs calculation for the selected pavement materials. the accumulated materials costs would be displayed after every input icon is entered as presented in fig. 6d in appendix ii. the description of each interface is presented in table 8. figure no. figure file description 6a rc-mss data input diagram (geometric design and traffic estimation) input data of design conditions. the input data consist of : 1.estimated traffic loading (esal), 2.geometric feature of vertical gradient and horizontal curvature, 3. type of terrain, and 4. soil cbr. 6b rc-mss suggested pavement materials output of the suggested pavement materials extract from microsoft access 2000 for matched rules. figure 6b shows the suggested pavement and the soil cbr is more than 5%. 6c rc-mss selected pavement materials the selected pavement materials would display in this form and material description button would show the selected material description. 6d rc-mss materials cost calculation input icon of layer thickness and cost of materials needed to enter before each layer materials and the total material cost can be calculated. table 8. figures 6a to 6d description in appendix ii 8 the journal of engineering research vol. 4, no.1 (2007) 1-10 suggested materials and rank print selected materials output soil input esal input terrain input horizontal and vertical input input data select pavement materials materials cost calculation ultimate output implement inference engine rule to determine the matching matrix from the knowledge-base knowledge-base for materials database and rule facts section (i+1) project name company name number of section load existing project project input enter rc-mss log in main menu log out end 7. conclusions with this prototype development of a pavement materials selection system, the road designer is more likely to view more proper pavement materials alternatives in the initial stage of materials selection. moreover, the system is easy to use and a brief cost materials calculation could be done via this rc-mss. with this system, the road designer has more materials choices and it is not necessary only to stick to a dense-graded mixture of asphalt concrete. since this is a prototype material selection system, a comprehensive system is needed which comprises the mechanistic-design method. further investigations of weather effect on materials properties would make the database of the system more effective, valid, precise and reliable. acknowledgments we would like to acknowledge the funding provided by the ministry of science, technology and environment, malaysia. we would also like to show appreciation to those road design engineers and pavement experts who helped in data collection and analysis. references deitel, h.m., deitel, p.j. and nieto, t.r., 1999, "visual basic 6: how to program," new jersey: prentice hall. emby, j. and mustafa, m.s., 1996, "the quality of asphaltic concrete mixes in malaysia," proceeding of the second malaysia road conference, 2nd mrc malaysian road conference. goh, a.t.c., 1993, "advisory expert system for flexible pavement design," artificial intelligent in engineering 0954 1810/93/506.00, vol. 8, pp. 47 56. elsevier science publishers ltd. harun, m.h., 1992, "the performance of various bitumen modifiers on climbing lanes," jkr, mesyuarat penolong pengarah jalan 3/92. kuala lumpur: public work department. huang, y.h., 1993, "pavement analysis and design," new jersey: prentice-hall. jkr, 1986, "a guide on geometric design of roads," arahan teknik (jalan) 8/86. kuala lumpur: public work department. lay, m.g., 1990, "transportation studies volume 8," handbook of road technology (2nd ed.) vol. 1, planning and pavements. muniandy, r., vasudevan, j. and omar, h., 2002, "laboratory evaluation of malaysian cellulose oil palm fiber for use in stne mastic asphalt mixes," international journal of pavements, vol. 1(3), pp. 13 21. omar, h., 200, "development of expert systems," qualifying exam, upm, seri kembangan. figure 6. hierarchy of an expert system in road construction materials system (rc-mss) 9 the journal of engineering research vol. 4, no.1 (2007) 1-10 denoted symbol as tsg site condition 10: soil=poor (cbr < 5 %), regardless of traffic loading and critical area. t er ra in c on di tio n no.adopted measure to the existing s p os iti on 1 tsg1 3 2 tsg2 3 3 tsg3 1 4 tsg4 3 t 1 5 tsg5 2 1 tsg1 4 2 tsg2 3 3 tsg3 1 4 tsg4 4 t 2 5 tsg5 2 1 tsg1 4 2 tsg2 3 3 tsg3 1 4 tsg4 3 t e n / t sg t 3 5 tsg5 2 section 1: input data input reference figure 6a. rc-mss data input diagram (geometric design and traffic estimation) appendix i road construction material selection system (rc-mss) database suggested pavement materials select preferable pavement material rank section design condition input data figure 6b. rc-mss suggested pavement materials 10 the journal of engineering research vol. 4, no.1 (2007) 1-10 material description command button pavement layer cross section selected pavement material figure 6c. rc-mss selected pavement material calculated value for each layer material cost enter data calculate button figure 6d. rc-mss materials cost calculation the journal of engineering research (tjer) vol. 14, no. 2 (2017) 145-155 doi: 10.24200/tjer.vol.14iss2pp145-155 assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country m.b. ramzana, s.m. qureshi*b, m. ullahc, m.s. memond and m.a. siddiquib adept. of garment manufacturing, national textile university, faisalabad, pakistan. *bdept. of industrial & manufacturing engineering, ned university of engg. and technology, karachi, pakistan. cdept. of industrial engineering, university of engg. and technology, peshawar, pakistan. ddept. of industrial engineering mehran university of engg. and technology, jamshoro, pakistan. received 19 march 2017; accepted 1 october 2017 abstract: in developed countries, the application of the quality management system (qms) is widely proven. however, in developing countries, like pakistan, industries are not mature enough to understand and implement the system. the aim of this study is to examine the perception of quality, implementation of qms and cost of quality (coq) program in value-added garment sector, which is one of the major subsectors of textile industry in pakistan. the methodology of the study is a questionnaire survey. the findings suggest that the majority of the respondent firms have well implemented quality management (qm) practices. trend of results also depicts that most of the organizations wish to acquire iso certification, which shows that the industrial processes are customer-centered thus striving to fulfil customer requirements. it can also be concluded that most of the organizations lack the coq concept, its understanding and appropriate implementation. keywords: quality management system; cost of quality; garment; textile; pakistan. حالة دراسة –تقييم مدى تطبيق مفاهيم نظام إدارة اجلودة )قمس( وتكلفة اجلودة )كوق( من بلد نام ه، مدحت عليد، حممد سعد ميموندج، مصباح أواله،ب* ، شهرييار حمسن قريشأحممد بابار رمزان يف البلدان املتقدمة على نطاق واسع. اال انه يف البلدان النامية مثل (qms)ثبت تطبيق نظام إدارة اجلودة لقد : امللخص (. 2014باكستان ، فأن الصناعات ليست ناضجة بشكل كاف الستيعاب وتنفيذ نظام إدارة اجلودة )قريشي وآخرون وتكلفة (qms) طبيق برنامج نظامي إدارة اجلودةوتهدف هذه الدراسة اىل الوقوف على مدى إدراك مفهوم اجلودة وت يف قطاع القيمة املضافة )املالبس(، والذي هو واحد من القطاعات الفرعية الرئيسية لصناعة الغزل والنسيج (coq)اجلودة يف باكستان. و تنبين منهجية هذه الدراسة على مسح استبياني. وتشري النتائج إىل أن غالبية الشركات اليت اجابات على كما يظهر التوجه إىل أن معظم الشركات ترغب يف احلصول . (qm)ارسات جيدة يف إدارة اجلودة هذا االستبيان لديها مم على شهادة األيزو، مما يدل على أن العمليات الصناعية تتمحور حول العمالء وبالتالي تسعى إىل تلبية متطلباتهم. كما وفهمها والطرق املناسبة لتطبيقها (qms)ة اجلودة ميكن أيضا استنتاج أن معظم الشركات تفتقر إىل وجود فكرة تكلف : اجلودة؛ نظام ادارة اجلودة؛ تكلفة اجلودة؛ ؛ املالبس , النسيج ؛ باكستان الكلمات املفتاحية * corresponding author’s e-mail: sheheryar@neduet.edu.pk mailto:sheheryar@neduet.edu.pk m.b. ramzan, s.m. qureshi, m. ullah, m.s. memon and m.a. siddiqui 146 1. introduction changes in the dynamic business environment followed by intense competition require companies to thrive for existence and survival. manufacturing companies’ efforts to win any competition is directed towards “quality”. this scenario has also motivated manufacturing and service industries to employ methodologies like statistical quality control (sqc), total quality management (tqm), total productive maintenance (tpm), quality management system (qms), 5s and kaizen, iso 9000 quality standard and six sigma (köksal et al. 2011). qms provides a framework for the application of advanced methods of sqc, reliability engineering and design of experiments (doe) (bisgaard 2008). in order to grip the qms framework, it is important to be familiar with the four pillars of qms, namely quality planning (qp), quality assurance (qa), quality control (qc) and quality improvement activities (purushothama 2011). total quality improvement and its conservation is the key objective of every industry in present dynamic and intense competitive business environment. tqm is a well-known philosophy for improving business and operations activities of any industry through adopting qms procedures and competencies. this process of evolution starts from inspection, moves towards qc and qa and finally approaches the philosophy of tqm. tqm is an approach of planning, organizing and understanding every activity of an organization and removing all the wastes from efforts and resources. as part of tqm, every organization must adopt qms which ensures that two significant requirements – customer expectations and organizations requirements – are fully met. the coq program is one of the effective tools of qm. it measures the cost incurred in achieving a better and improved product or service quality. the implementation of a coq program enhances the competitive advantage of the company and helps in being cost effective, both on long and short term basis. pakistan is the 4th largest cotton producer and 3rd largest cotton consumer in the world. the textile exports contribute 57% of the country’s exports that stands pakistan as the 8th largest textile exporter in asia. pakistan fulfils 9% of the global textile need and the is ranked in the world as the 10th textile producer. the textile industry of pakistan is mainly based on locally grown cotton and has four subsectors namely yarn manufacturing, fabric manufacturing, textile processing and garment manufacturing. the garments sector contributes 48% of pakistan’s total exports, 30% of value-added in large-scale manufacturing and 40% of industrial employment. further highlighting the importance of this sector, it is pertinent to state that in 2013, the european union (eu) granted generalised system of preferences (gsp) plus status to pakistan. the gsp plus status would allow almost 20% of pakistani exports to enter the eu market at zero tariff and 70% at preferential rates. eu trade concessions would benefit textile and garment industry most by enabling its products to compete with regional rivals like bangladesh and sri lanka, which already have duty free access to the bloc’s market. furthermore, due to the rising labor cost in china (lin 2012) and income rise in the large, fast-growing economies of china and india, the textile and garments export potential of the country would receive huge benefits (hamid and zafar 2014). however, in order to realize the export potential, textile and garment sectors need to realign their structure by adopting modern and effective methodologies. it is observed that in developing countries, like pakistan, industries are not mature enough to understand and implement qms (qureshi et al. 2014). thus the aim of this study is to examine the perception of quality, implementation of qms and coq program in value-added garment sector, which is one of the major subsectors of the textile industry in pakistan. for investigation, a questionnaire was designed according to the guidelines by dillman (1991). the questionnaire comprised two sections: the first section gathered respondents’ demographic information whereas the second one contained comprehensive questions to accomplish the objectives of the study described above. three major cities, karachi, lahore and faisalabad, were selected for the survey because the majority of textile industries are located in these cities. 2. literature review although tqm is not a new concept in the textile and clothing industry, it has received increased attention in the last few years (maia et al. 2015). being a part of tqm philosophy, an increasing trend of implementing qms is observed (iso 2013, towers and mcloughlin assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country 147 2005). it is widely accepted that qms can be implemented in any organization regardless of size or type. despite this benefit, organizations, like small and medium-sized enterprises (smes), hesitate to implement qms due to cost and maintenance difficulties. in developed countries, acceptance and implementation of qms is very high. on the contrary, in developing countries, there is a gap between actual and true implementation (scott and collins 1998). towers and mcloughlin (2005) conducted a survey across uk based textile smes to study the effect of qms on cost, quality and productivity. two-thirds of smes considered qms as a good contributor of increased customer satisfaction, awareness of quality and teamwork in organizations. different studies (fatima and ahmed 2005a; fatima and ahmed 2006, fatima and ahmed 2006b; hasan et al. 2013; shafiq 2012) have assessed the implementation of tqm in textile and garment sector of pakistan and found that most companies have acquired iso certification, but actual practices are still not being followed in real production environment. a survey study by fatima and ahmed (2005) on readymade garment industry observed that the segment was well aware of advanced qm concepts and enhanced product quality after adopting tools like kaizen. however, new strategic directions based on modern management concepts should be adopted by organizations to increase the exports of readymade garments. another survey study results revealed positive and significant effects of tqm practices on firm performance in textile sector of pakistan (hassan et al. 2014). organizations seek to implement effective techniques that satisfy customer requirements. one of the most effective techniques that is used to measure the costs incurred in achieving high quality is the coq program (qureshi et al. 2014). coq ensures improvement in quality and reduction in manufacturing cost by decreasing the defect rate (desai 2008). much work has been done on coq to demonstrate its philosophy. qureshi et al. (2014) summarize all definitions of coq that have been referred to by different researchers in the last three decades. correctly measuring and then reporting coq is also a critical issue. there are different models which can be used to categorize, measure and report coq. omar and murgan 2014; schiffauerova and thomson 2006) discuss these models along with cost categories and case studies indicating the implementation of coq in different sectors. regardless of the type of coq model applied, accurate measurement is the key aspect that gives potential benefits to organizations. these benefits include highlighting the weak areas that require improvement, help controlling the overall quality and increasing firm’s competitive advantage by decreasing cost (yang 2008). the philosophy behind coq is now being understood and accepted in different organizations that include smes, different manufacturing and service industries and supply chains. even in textile sector, estimating quality related costs is important as it helps to improve the overall quality and productivity of textile goods. but the aforementioned studies hardly discuss the implementation of the coq program. integration of different tools and techniques enhances the effectiveness of tools further. this integration concept is widely used to enhance the effectiveness of qms by integrating with lean manufacturing, six sigma and lean six sigma in manufacturing industries (chiarini 2011; karthi et al. 2011a; karthi et al. 2011b; karthi et al. 2012; karthi et al. 2013). all these studies develop integrated models to improve the effectiveness of qms with well-known tools like lean and six sigma. it is suggested that textile organizations in pakistan should think beyond tqm and implement other tools and techniques such as lean manufacturing and six sigma (shafiq 2012). hussain et al. (2014) explore the application of six sigma methodology for reducing the defect percentage in fabric manufacturing sector. they found that by the suitable application of six sigma tools, the sigma level was improved from 2.2 to 3. however, the review of literature provides no evidence of any comprehensive study of coq on textile sector of pakistan. the present study aims to assess the implementation of qms and its effect on coq. for this purpose, the garment sector of pakistan has been investigated by the following objectives:  to assess awareness of quality and qms in the pakistani garment industry.  to assess the level of acceptance and implementation of qms.  to assess awareness regarding measuring coq.  to assess the impact of implementing qm practices on coq. m.b. ramzan, s.m. qureshi, m. ullah, m.s. memon and m.a. siddiqui 148 3. research methodology in order to address the aforementioned research objectives, data were collected from garment industry of pakistan by developing a survey questionnaire. the questionnaire comprised two sections: the first segment gathered respondents’ demographic information (table 1), the second section contained comprehensive questions to accomplish the study objectives described above. in order to collect data, 153 garment manufacturing industries were selected from three main industrial cities of pakistan i.e. karachi, lahore, and faisalabad. out of 153 questionnaires, 102 were returned and 72 were finally considered valid for analysis, giving a response rate of 47%. data were analyzed using spss 22. before analysing the survey results, the gof and reliability of dataset were assessed. χ2 test to check the gof and cronbach's α to test the reliability of data were carried out. table 2 shows the χ2 test statistics for variables in discussion with a p-value indicating the significance. a value of 0.700 of cronbach's α indicates that the reliability and internal consistency of the dataset are acceptable (table 3). table 1. respondents’demographics. characteristics category frequency percentage location lahore 27 38% faisalabad 40 56% karachi 05 07% gender male 47 65% female 25 35% designation general manager / senior manager 04 06% manager/deputy manager 13 18% assistant manager 25 35% officer/trainee engineer 30 42% department qc/qa 11 15% production 17 24% marketing/merchandising 23 32% industrial engineering 21 29% industry (type) knitwear 56 78% denim/woven 12 17% others 04 06% total respondents 72 table 2. χ2 test results. test statistics chi-square (χ2) df p-value quality is related to product only 8.562 1 .003 quality is every department's responsibility 65.575 3 .000 quality is everyone's responsibility 38.000 2 .000 have you implemented qms? 8.562 1 .003 did you find qms effective? 27.597 1 .000 did you benchmark for adopting qms? 9.986 1 .002 are improvements measured, recorded and displayed? 40.500 1 .000 are you an iso certified company? 3.959 1 .047 do you measure coq? 40.500 1 .000 do you wish to measure coq? 54.069 1 .000 is your accounting system good enough for coq? 9.058 1 .003 do you think qm practices influence coq? 99.658 3 .000 assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country 149 table 3. cronbach's α result. cronbach's alpha n of items 0.700 14 4. results and discussion according to the study objectives, the survey findings are described in five sections comprising awareness of quality; awareness of qms; iso certifications; awareness of coq; and ideas to implement coq program. a. awareness of quality this section is designed to assess the respondents' awareness of their concept of quality. it is divided into two parts: perception of quality; and understanding of quality. i. perception of quality the respondents were first asked about their perception of quality and their responses are summarized in table 4. the majority of the respondents (61%) perceived quality as “customer satisfaction,” along with the degree of excellence (15%), fitness for use (7%) and conformance to customer's quality standard (6%). this shows that the conventional definition of meeting/exceeding customer satisfaction prevails in the study sample. this is because of the fact that the pakistani garment industry mainly manufactures products for different customers according to their require ments and product specification. while a few industries manufacture products keep in view the end users, the fitness for use or degree of excellence were perceived by fewer respondents. ii. understanding of quality in the second part, the respondents were asked different questions to evaluate their understanding of quality. in response to the question, “is quality related to product quality only?”, the respondents were found to strongly agree (13%) or agree (28%) with this statement (table 5). this, nevertheless, does not support the true concept of quality (qureshi et al. 2014). on the other hand, a significant number of respondents disagreed (42%) or strongly disagreed (11%) with this statement. one reason could be that clothing industry focusses more on process quality to obtain the required/desired product quality. the respondents were then asked about the responsibility for quality. the results summarized in table 5 indicate that the majority is in agreement with the concepts that quality is an organization-wide responsibility. however, it is astonishing to observe that 96% of respondents believe that quality is both individual and departmental responsibility, which is against the soul of tqm. tqm ensures that quality is not a departmental responsibility and true practices of the philosophy encourage everyone to play his positive role (qureshi et al. 2014). b. awareness of qms the objective of this section is to evaluate the acceptance level of qms in the garment industry of pakistan. for this purpose, the section is organized in two parts: implementation of qms; and perception of qms. table 4. respondents’ perception regarding the term “quality”. perception frequency percentage customer satisfaction 44 61% degree of excellence 11 15% fitness for use 5 7% conformance to customer's quality standard 4 6% zero defects 3 4% segregating of good one from bad one 3 4% continuous improvement with no end line 2 3% m.b. ramzan, s.m. qureshi, m. ullah, m.s. memon and m.a. siddiqui 150 table 5. respondent’s understanding of quality. response quality is related to product only quality is everyone's responsibility quality is every department's responsibility frequency % frequency % frequency % strongly agree 9 13 43 60 45 63 agree 20 28 24 33 24 33 disagree 30 42 4 6 3 4 strongly disagree 8 11 1 1 0 0 missing 5 7 0 0 0 0 table 6. implementation of qms. response have you implemented qms? if “no”, do you wish to implement qms? frequency % frequency % yes 38 53% 30 94 no 32 44% missing 2 3% 2 6 table 7. significance of qms. response did you find qms effective? did you benchmark for adopting qms? are improvements measured, recorded and displayed? frequency % frequency % frequency % yes 55 76 35 49 62 86 no 12 17 36 50 9 13 missing 5 7 1 1 1 1 i. implementation status of qms according to the results summarized in table 6, 53% respondents claimed that they had implemented qms in their organizations. while 44% did not implement qms but they showed their willingness to adopt it. the survey (iso 2013) also revealed that pakistanis are far away from their regional as well as international competitors in the true implementation of qms. ii. perception about qms in order to evaluate the respondents' perception of the benefits of qms, the questions were asked about the effectiveness of qms, display of improvement and benchmarking. most of the companies (76%) found qms as an effective tool (table 7) and observed its positive impact on overall performance. this performance includes improvement in quality and process along with the implementation of standard methods and procedures (o’neill et al. 2016). the respondents were further asked about whether or not they had benchmarked any organization for qms implementation t. the results showed that almost 50% of the respondents benchmarked some organizations, while 50% did not. benchmarking is a recommended activity that facilitates organizations to investigate their strengths and weaknesses by comparing them with the best organization. this practice enables organizations to learn useful things from the best, working in the same field, for improving their own performance (barber 2004). as the firms claimed that they had well-implemented qms, then they were recommended to benchmark other successful organizations or let other organizations benchmark them to learn from the best practices. the findings also show that 86% of the respondent garment manufacturing firms displayed their improvements. qms encourages such practices since they allow everyone to evaluate individuals as well as organizational performance. in garment manufacturing, where organizations depend much on human labour, such performance displays would create a competitive environment (qureshi et al. 2014). c. status of iso certification the objective of this section is to obtain the status of iso certification in the garment assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country 151 industry. the results are divided into two parts: iso certified companies; and non-certified companies. i. iso certified companies the results indicate that 63% of the respondents belonged to the iso certified companies (table 8). among iso certified companies, 64% were iso 9001:2008 certified, which is a widely accepted quality program (kim et al. 2011). however, some organizations (33%) were still not certified, which indicated that the significance of these certifications had not been properly understood by these organizations. theoretical review suggests that the trend of acquiring iso certifications in pakistan has been increasing for the last 10 years, but this growth is still behind competitors like china and india (iso 2013). ii. non-certified companies however, among the non-certified companies, 52% showed their desire to acquire iso certification (table 9). 91% of them opined that iso 9001 was the best-suited certification for their organization to meet customer expectations. however, by comparing this situation with the leading textile manufacturing countries, iso 9001 has already been widely accepted (fonseca 2015). thus, there is a huge gap in implementing the iso based qms when compared with the rest of the world (iso 2013). table 8. iso certified companies. response are you an iso certified company? if "yes", which iso certification do you have? frequency % certification frequency % yes 45 63 iso 9001:2008 29 64 no 24 33 iso 14000 10 22 missing 3 4 others 6 13 table 9. non-certified companies. response do you wish to get iso certification? if "yes", which iso certification would you go for? frequency % certification frequency % yes 22 52 iso 9001:2008 20 91% no iso 14000 2 9% missing 2 5 others 0 0% table 10. measurement of coq. response do you measure coq? do you wish to measure coq? is your account -ing system good enough for coq? do you think qm practices influence coq? frequency % frequency % frequency % frequency % yes 9 13 56 90 21 34 54 75 no 62 86 0 0 37 60 4 6 missing 1 1 6 10 4 6 14 19 d. awareness of coq this section was designed to evaluate awareness of coq. it comprises two main parts: measurement, categories and elements of coq; and techniques to identify coq elements. i. coq measurement in order to evaluate whether or not the industry under study had implemented coq program, different questions were asked. the results are summarized in table 10. despite being familiar with the importance of this m.b. ramzan, s.m. qureshi, m. ullah, m.s. memon and m.a. siddiqui 152 program, the majority of the garment industry (86%) does not measure coq. 13% of the respondents agreed that coq was measured in their companies. however, even within this group, there was no evidence of true implementation as they could not describe their adopted coq model. it was also highlighted that most of the companies (60%) did not have a proper accounting system for coq measurement. although 34% of industries mentioned that their accounting system was capable enough, it could only be evaluated when coq program was fully adopted. these findings show that although garment industry of pakistan seems aware of the importance of qms and iso certifications, coq program is not properly/truly implemented in majority organization (qureshi et al. 2014). furthermore, the results also show that 75% of the respondents observed a significant influence on coq due to effective qm practices. therefore, many clothing companies were informed of the strong relationship between qms and coq (bisgaard 2008). besides this analysis, it is important to observe that 90% of those organizations who did not measure coq showed their willingness to implement the program in the future. ii. categories and elements of coq the respondents were also asked about coq categories and elements. although a significant number (29%) did not respond, the rest of the respondents (71%) gave their suggestions that suit their organizational environment (table 11). in the case of categories, the majority of industries would like to adopt paf (prevention, failure and appraisal) coq categories (43%) followed by conformance/non-conformance (21%) and value added/non-value added (4%) categories. although most of the respondents did not depict sufficient knowledge about coq, the possible reason of selecting paf model was due to its worldwide acceptance in most of the manufacturing industries (su et al. 2009). on the other hand, table 12 depicts that garment industry wants to pay more attention to internal failure (21%) and external failure (25%) costs. this shows poor understanding of coq as recent researches encourage manufacturers to focus more on prevention and appraisal elements to reduce failure costs (omar and murgan 2014). table 11. categories of coq. coq category frequency % paf 31 43 conformance and non-conformance 15 21 value added and non-value added 3 4 others 2 3 missing 21 29 table 12. elements of coq. coq elements frequency % prevention 8 11 appraisal 10 14 internal failure 15 21 external failure 18 25 missing 21 29 iii. techniques to identify coq elements in the last part of this sub-section, the respondents were asked about the techniques to identify coq elements as shown in table 13. according to the respondents, brainstorming (75%), input/output analysis (72%), benchmarking (65%), fishbone diagram (56%) and pareto analysis (49%) are the most commonly used techniques to identify coq elements. 4.1 implementation of coq program in the last part of the survey, the respondents were asked to suggest their ideas for propagating awareness of coq program in garment manufacturing industry of pakistan. the suggested recommendations and ideas are summarized in four key statements (table 14). according to the respondents, the reason behind the lack of implementation of coq program is unfamiliarity with the method and cost element behind this concept. thus, training programs must be initiated to familiarize people assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country 153 with its benefits in terms of quality and process improvement. since there are different types of coq models, training will also help to select a suitable model according to the requirements of the industry. respondents also highlighted that the implementation and training costs of such programs were high and hence normally top management avoided them. however, training programs could motivate people to realize and acknowledge the long-term benefits of implementing coq system. 5. conclusion the aim of this study was to assess the extent of implementing qms and coq concepts in the garment manufacturing industry of pakistan. for this purpose, a survey of garment sector was conducted keeping in view four objectives of this study as mentioned earlier. the findings suggest that the garment industry of pakistan is well aware of the principles of qms to develop and practice good quality culture. moreover, the philosophy of tqm is widely accepted to standardize the methods and processes of garment manufacturing. trend also depicts that most of the organizations wish to acquire iso certifications, which shows that the industrial processes are customer-centred and thus they strive to fulfil customer requirements. however, with reference to certain responses, it is also observed that the principles of qms are not being followed in their true sense by some organizations. these companies wish to implement qms for the sake of acquiring certifications only. the reason behind this narrow view is the extensive involvement of manpower in textile manufacturing processes that are not properly trained on qms principles and procedures. this work also reveals that coq program is not truly implemented in the textile and clothing sector of pakistan. most of the industries are aware of the importance of coq and familiar with its models, along with coq categories and elements. from the table 13. techniques to identify coq elements. techniques frequency % brainstorming 54 75 input/output analysis 52 72 benchmarking 47 65 fishbone diagram 40 56 pareto analysis 35 49 nominal group techniques 20 28 others 2 3 table 14. suggestions and recommendations for implementing coq program. respondents’ recommendations frequency % training programs of coq must be conducted to highlight its real benefits 25 35% cost of training is higher and true practice is difficult to implement 18 25% garment industry should pay more attention to the true implementation of programs like coq 11 15% this is a powerful tool that will identify the expenditures that occur on quality at each stage in a production process 10 14% no recommendation or idea suggested 08 11% m.b. ramzan, s.m. qureshi, m. ullah, m.s. memon and m.a. siddiqui 154 findings, it can be concluded that elementary knowledge of coq system does exist along with the will to implement it, yet most of the organizations lack coq concepts including understanding and appropriate implementation strategy. thus, proper training and formal reporting are still to be familiarized for the true implementation of coq program. finally, respondents also recorded their ideas about promotion of coq program. although it is a bit of a costly process due to initial expenditures of training and execution, it will surely bring longterm benefits, not only for textile business, but also for other manufacturing and service industries. furthermore, the barriers and difficulties encountered in this exercise may also be investigated separately in the context of developing countries (rasamanie and kanapathy 2011). the findings of the study also suggest that existing accounting systems also represent one of the barriers in implementing coq system. this study also highlights the gap that exists between textile industry of pakistan and the market leaders of the world textile business in the context of qms and coq. however, a lack of a deeper and more comprehensive statistical analysis may be considered as a potential limitation of this study. therefore, further work can be done with a larger sample size, adding other textile-rich geographical locations as well, packed with more rigorous statistical analysis will help to draw further conclusions on other dimensions of qms and coq program implementation. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references barber e (2004), benchmarking the management of projects: a review of current thinking. international journal of project management 22(4): 301-307. bisgaard s (2008), quality management and juran's legacy. quality engineering 20(4): 390-401. chiarini a (2011), integrating lean thinking into iso 9001: a first guideline. international journal of lean six sigma 2(2): 96-117. desai da (2008), cost of quality in small-and medium-sized enterprises: case of an indian engineering company. production planning and control 19(1): 25-34. dillman da (1991), "the design and administration of mail surveys. annual review of sociology 225-249. fatima m, ahmed e (2005a), quality management in pakistan's readymade garments' industry. quality engineering 17(3): 459-465. fatima m, ahmed e (2006b), quality management in pakistan's knitwear industry. quality engineering 18(1): 15-22. fonseca lm (2015), from quality gurus and tqm to iso 9001: 2015: a review of several quality paths. international journal for quality research 9(1): 167-180. hamid nin, zafar r (2014), the textiles and garments sector: moving up the value chain. the lahore journal of economics 19: 283-306. hasan iu, sohail mm, piracha jl, ahmad k (2013), implementation status of tqm practices in textile and apparel industrial organization: a case study from faisalabad, pakistan. british journal of economics, management and trade 3(3): 201223. hassan mu, nawaz ms, shaukat s, hassan s (2014), an empirical assessment of tqm dimensions and their relationship with firm performance: evidence from the textile sector of pakistan. world applied sciences journal 30(6): 696-705. hussain th. jamshai, sohail a (2014), reducing defects in textile weaving by applying six sigma methodology: a case study. international journal of six sigma and competitive advantage 8(2): 95-104. iso (2013), iso survey. from http://www.iso.org/iso/iso-survey. karthi ss, devadasan, murugesh r (2011a), lean six sigma through iso 9001 standardbased quality management system: an investigation for research. international journal of productivity and quality management 8(2): 180-204. karthi ss, devadasan, murugesh r (2011b), integration of lean six-sigma with iso 9001: 2008 standard. international journal of lean six sigma 2(4): 309-331. http://www.iso.org/iso/iso-survey assessment of the extent of implementation of quality management system (qms) and cost of quality (coq) concepts – a case from a developing country 155 karthi ss, devadasan r, murugesh c, sreenivasa, sivaram n (2012), global views on integrating six sigma and iso 9001 certification. total quality management and business excellence 23(3-4): 237-262. karthi ss, devadasan, k. selvaraju, n. sivaram and c. sreenivasa (2013). "implementation of lean six sigma through iso 9001: 2008 based qms: a case study in a textile mill." journal of the textile institute 104(10): 10891100. kim dy, kumar v, kumar u (2011). "a performance realization framework for implementing iso 9000. international journal of quality and reliability management 28(4): 383-404. köksal g, batmaz i, testik mc (2011), "a review of data mining applications for quality improvement in manufacturing industry. expert systems with applications 38(10): 13448-13467. lin jy (2012), from flying geese to leading dragons: new opportunities and strategies for structural transformation in developing countries. global policy 3(4): 397-409. maia lca, alves c, leão cp (2015), "how could the triz tool help continuous improvement efforts of the companies? procedia engineering 131: 343-351. omar mk, murgan s (2014), an improved model for the cost of quality. international journal of quality and reliability management 31(4): 395-418. o’neill p, sohal a, teng cw (2016), quality management approaches and their impact on firms׳ financial performance–an australian study. international journal of production economics 171: 381-393. purushothama b (2011), a practical guide to quality management in spinning. woodhead publishing india. qureshi sm, majeed s, khalid r (2014), where do we stand in cost of quality awareness?– pakistan’s case. international journal of quality engineering and technology 4(4): 273289. rasamanie m, kanapathy k (2011), the implementation of cost of quality (coq) reporting system in malaysian manufacturing companies. difficulties encountered and benefits acquired. international journal of business and social science 2(6): 243-247. schiffauerova a, thomson v (2006), a review of research on cost of quality models and best practices. international journal of quality and reliability management 23(6): 647-669. scott l, collins p (1998), evaluation of the iso 9000 series: benefits and problems of implementation and maintenance in production companies. journal of the textile institute 89(1): 90-109. shafiq m (2012), implementation of quality management systems and business excellence frameworks in pakistani textile companies. journal of quality and technology management 7(2): 11-23. su q, shi jh, lai sj (2009), research on the trade-off relationship within quality costs: a case study. total quality management 20(12): 1395-1405. towers n, mcloughlin j (2005), effective total quality management in the textile fashion retail supply chain: a pilot survey of the uk textile manufacturers. journal of the textile institute 96(2): 87-92. yang cc, (2008), improving the definition and quantification of quality costs. total quality management 19(3): 175-191. the journal of engineering research (tjer), vol. 15, no. 1 (2018) 14-25 doi: 10.24200/tjer.vol15iss1pp14-25 vibrations of circular plates resting on elastic foundation with elastically restrained edge against translation l.b. raoa,* and c.k. raob a school of mechanical and building sciences, vit university, chennai campus, vandalur-kelambakkam road, chennai600127, tamil nadu, india. b department of mechanical engineering, school of engineering, chaudari guda, korremulla ‘x’ roads, near narapalli, ghatkesar (m), hyderabad-500 088, india. received 6 july 2016; accepted 26 january 2017 abstract: the present paper deals with exact solutions for the free vibration characteristics of thin circular plates resting on winkler-type elastic foundation based on the classical plate theory elastically restrained against translation. parametric investigations are carried out for estimating the influence of edge restraint against translation and stiffness of the elastic foundation on the natural frequencies of circular plates. the elastic edge restraint against translation and the presence of elastic foundation has been found to have a profound influence on vibration characteristics of the circular plate undergoing free transverse vibrations. computations are carried out for natural frequencies of vibrations for varying values of translational stiffness ratio and stiffness parameter of winkler-type foundation. results are presented for twelve modes of vibration both in tabular and graphical form for use in the design. extensive data is tabulated so that pertinent conclusions can be arrived at on the influence of translational edge restraint and the foundation stiffness ratio of the winkler foundation on the natural frequencies of uniform isotropic circular plates. keywords: plate; frequency; elastic edge; translational stiffness; elastic foundation. اثناء احلركةاملرونة اهتزازات االلواح الدائرية املثبتة على قاعدة مرنة مع حافة مقيدة ب، تشيالبيال كامسوارا راو*،أ لوكافارابو بهاسكارا راو املثبتة على قاعدة قيقةلاللواح الدائرية الر ةاحلر اتخلصائ االهنزاز صحيحةالبحثية احللول ال : تتناول الدراسة لص امل معاملية حباث ا. مت إجراء لتقليديةا لواحمبوجب نظرية االاحلركة اثناءرونة امل مقيدة ذات حافة مرنة من نوع وينكلر و الدائرية. لأللواحصالبة القاعدة املرنة على الرتددات الطبيعية و احلركة اثناءيد احلافة يتأثري تق مدىلتقدير (بارامرتية) مع وجود القاعدة املرنة هلا تأثري كبري على خصائ اهتزاز ركةاحلعملية أثناءاىل أن احلافة مقيدة املرونة توصل البحث يعية لالهتزازات لقيم متفاوتة من تعرض الهتزازات عرضية حرة. ومت إجراء احلسابات للرتددات الطبيوح الدائري اليت لال اتاالهتزاز أمناطالنتائج الثين عشر منطا من و قدم البحثمعدل الصالبة املتعدية ومعامل صالبة القاعدة من نوع وينكلر. إىل منها حبيث ميكن التوصل شاملةيف شكل جداول و رسومات بيانية الستصدامها يف التصميم. كما مت جدولة بيانات نوع وينكلر على الرتددات الطبيعية للوحات من استنتاجات ذات صلة بتأثري تقييد احلافة املتنقلة ونسبة صالبة القاعدة الدائرية املتماثلة املوحدة. قاعدة مرنة ،نقلةصالبة مت ،حافة مرنة،الرتدد ،: لوح فتاحيةملا كلماتال * corresponding author’s e-mail: bhaskarbabu_20@yahoo.com mailto:bhaskarbabu_20@yahoo.com l. b. rao and c.k. rao 15 nomenclature h thickness of a plate, mm a radius of a plate, mm  poisson’s ratio e young’s modulus, n/mm2  density of a material, kg/mm3 ),( rw transverse deflection of the plate, mm d flexural rigidity of a plate, n.mm2 t k translational spring stiffness, n/mm w k stiffness of winkler foundation, n/mm2/mm t translational spring stiffness ratio  foundation stiffness ratio mn  natural frequency, rad/sec mn  eigenvalue without foundation, cycles/sec * mn  eigenvalue with winkler foundation, cycles/sec nm, positive integers corresponding to the number of concentric circles and nodal diameters in each flexural mode. vibrations of circular plates resting on elastic foundation with elastically restrained edge against translation 16 1. introduction circular plates resting on elastic foundation have a wide range of applications in the static and dynamic design of linear/nonlinear vibration absorbers, dynamic exciters, telephone receiver diaphragms, computer discs, printed circuit boards etc. (leissa 1969). due to the essential use of vibration data in the computation of stresses in such structures, reliable prediction of vibration data is of great importance. in view of its importance in engineering design, the problem of vibration of circular plates on elastic foundation has attracted the focus and attention of many researchers. some of the recent studies have reinstated the classical approach efficiency in analyzing the vibrations of variety of structures. circular plate problems allow for significant simplification in view of their symmetry, but still many difficulties arise when the plate boundary conditions become complex involving linear and rotational restraints. a recent review of literature shows that very few studies exist on the study of circular plates resting on elastic foundations. wang and wang (2003), who observed the switching between axisymmetric and asymmetric vibration modes, have recently investigated the effect of internal elastic translational supports. the vibration characteristics of plates resting on an elastic medium are different from those of the plates supported only on the boundary. leissa (1993) discussed the vibration of a plate supported laterally by an elastic foundation. leissa deduced that the effect of winkler foundation merely increases the square of the plate natural frequency by a constant. salari et al. (1987) speculated the same conclusion. ascione and grimaldi (1984) studied unilateral frictionless contact between a circular plate and a winkler foundation using a variational formulation. leissa (1969), who tabulated a frequency parameter for four vibration modes of a simply supported circular plate with varying rotational stiffness, presented one of the earliest formulations of this problem. kang and kim (1996) presented an extensive review of the modal properties of the elastically restrained beams and plates. zheng and zhou (1988) studied the large deflection of a circular plate resting on winkler foundation. ghosh (1997) studied the free and forced vibration of circular plates on winkler foundation by an exact analytical method. (chang and wickert (2001); kim et al. (2001) and tseng and wickert (1994) studied the dynamic characteristics of bolted flange connections involving circular plates displaying beating type of repeat frequencies and typical mode shapes of vibration. bolted flange connections are practically the best examples for the elastically restrained boundary conditions of circular plates on partial or continuous winkler type elastic foundation. the most general soil model used in practical applications is the winkler (1867) model in which the elastic medium below a structure is represented by a system of identical but mutually independent elastic linear springs. recent investigations have reiterated the efficiency of the classical approach (soedel 1993) in analyzing the behavior of structures under vibrations. there are other works (weisman 1970; dempsey et al. 1984; celep et al. 1988) dealing with the study of plates on a winkler foundation. in general, those dealing with vibrating plates, shells and beams are concerned with the determination of eigenvalues and mode shapes (leissa 1969). a good number of studies was conduced (wang and lin 1996; kim and kim 2001; yayli et al. 2014) using the method of fourier series for estimating the frequencies of beams with generally restrained end conditions including the effect of elastic soil foundation. the method includes the use of stoke’s transformation in suitably modifying the complex boundary conditions. very much similar to the dynamic stiffness matrix approach, the elements of the matrix involving infinite fourier series are explicitly obtained in these studies. the determinant of this matrix for each case considered leads to the frequency equation and the same can be solved using well known numerical methods. the results obtained for various elastically restrained beam cases in these studies tallied well with those available in the literature establishing the efficiency of this method. in view of the necessity of using complex combinations of rotational and translational springs at the circular plate boundary to suitably simulate the practical non-classical boundary connections being adopted in a wide range of industrial applications (bhaskara rao et al. 2009; bhaskara rao et al. 2010; lokavarapu and chellapilla 2013; bhaskara rao et al. 2015; lokavarapu et al. 2015; rao et al. 2016), the use of exact method of solution becomes imperative l. b. rao and c.k. rao 17 and hence the same is adopted in this paper. even though the method adopted here is classical, the particular case of vibration of elastically restrained circular plate resting on elastic foundation considered here is not addressed within the available literature. utilizing the classical plate theory, this paper deals with an exact method of solution for the analysis of thin circular plate free transverse vibrations that is elastically restrained against translation and resting on winkler-type elastic foundation. for estimating the influence of edge restraint against the elastic foundation translation and stiffness on circular plates natural frequencies, parametric investigations are carried out by varying the values of elastic edge restraint stiffness against the elastic foundation translation and stiffness. the results obtained on natural frequencies of vibration clearly show that the vibration characteristics of the circular plate undergoing free transverse vibrations are found to be profoundly influenced by these variations. computations are carried out for natural frequencies of vibrations for varying values of translational stiffness ratio and stiffness parameter of winkler-type foundation. the results that are presented for twelve modes of vibration both in tabular and graphical forms are believed to be quite useful for designers in this area. 2. mathematical formulation of the system the considered elastic thin circular plate is supported on a winkler foundation as shown in fig. 1. in the classical plate theory (leissa 1969), the following fourth order differential equation describes free flexural vibrations of a thin circular uniform plate: 0 2 t/)t,,r(w 2 h)t,,r(w 4 .d  (1) where ) 2 1(12/ 3 ehd  is the flexural rigidity of a plate and  ,e,,h,a are the plate’s radius, thickness, density, young’s modulus and poisson’s ratio, respectively. the homogeneous equation for kirchhoff’s plate on one parameter elastic foundation is given by the following equation:  )t,,r(wwk)t,,r(w 4 .d (2) 2 t/)t,,r(w 2 h  =0 displacement in (2) can be presented as a combination of spatial and time dependent components as follows: let ti erwtrw   ),(),,(  (3) now substitute (3) in (2) 0),r(w). 2 hwk(),r(w 4 .d   (4) the solution of the equation takes the following form     ncos. a rmn nimnc a rmn nj mna),r(mnw                            , n > 0 (5) where mna and mnc are constants, nj is bessel function of the first kind of first order and ni is modified bessel function of the first kind of first order, indexes m and n are positive integers and correspond to the number of concentric circles and nodal diameters in each flexural mode. considering an elastically supported plate as shown in fig. 1, boundary conditions can be formulated at ar  , in terms of translational stiffness ( tk ) as follows: 0),a(rm  (6)     ,awtk,arv  (7) where the kelvin-kirchhoff and bending moment are defined as follows: figure 1. a thin circular plate with translational elastic edge restraint and supported on elastic foundation. vibrations of circular plates resting on elastic foundation with elastically restrained edge against translation 18                                      2 ),a(w 2 2 r 1 r ),a(w r 1 2 r ),a(w 2 .d),a(rm       (8)                                         ),a(w 2 r 1 r ),a(w 2 r 1 r 1 )1(),a(w 2 r .d),a(rv (9) by applying eqs. (6) and (8), we obtain the following equation 0 2 ),a(w 2 2 r 1 r ),a(w r 1 2 r ),a(w 2                              (10) from eqs. (5) and (10), we derive the following equation )( 4 2 2 )( 4 2 2 2 2 2 2 mnn mn mn mn mn mnn mn mn mn mn mn i n st j n pq c                              (11) where mn p = )mn(1nj)mn(1nj   ; )mn(2nj)mn(2njmnq   ; )mn(1ni)mn(1nimns   ; )mn(2ni)mn(2nimnt   ; from eqs. (7) and (9), we get the following ),a(w.tk ),a(w 2 r 1 r ),a(w 2 r 1 r 1 )1(),a(w 2 r .d                                          (12) from eqs. (5) and (12), we derived the following equation )mn(nit 3 mn 8 mn 4 3 mn 2 n)3(8 mns 2 mn 2 n)2(44 3 mn mnt2 mnu )mn(njt 3 mn 8 mn 4 3 2 n)3(8 mnp 2 mn 2 n)2(44 3 mn mnq2 mnr mnc                                                          (13) where, d ka t t 3  )()( 11 mnnmnnmn jjp    ; )mn(2nj)mn(2njmnq   ; )()( 33 mnnmnnmn jjr    ; )()( 11 mnnmnnmn iis    ; )mn(2ni)mn(2nimnt   ; )()( 33 mnnmnnmn iiu    ; if  t k then this case becomes a simply supported boundary condition as shown in fig. 2. the frequency equation can be calculated from eqs. (11) and (13), which allows determining eigenvalues mn  . the mode shape parameters mnc can be determined corresponding to these eigenvalues by using either eq. (11) or eq. (13). the amplitude of each vibration mode in eq. (5) is set by the normalization constant mna determined from the following condition. nqmpmnpq a mn mrdrdrwrw    ),().,( 2 0 0 (14) where, mn m is a mass of the plate, 1 nqmp  if qnpm  , and nqmp = 0 if m ≠ p or n ≠ .q figure 2. a simply supported thin circular plate resting on elastic foundation. l. b. rao and c.k. rao 19 the normalization constant mn a can be derived using eqs. (5) and (14) as given below: 1 2 0 a 0 rdrd 2 ncos. a rmn ni.mnc a rmn nj . 2 a 1 mna                                                                  (15) in eq (4), the natural frequency is defined as mn                   h d 2 a 2 mn   (16) it is clear from eq. (16) that the natural frequency parameter mn  is dependent on the plate radius ‘a’. from eq. (16) we can express d 2 mn 4 ha4 mn    (17) 244*   mnmn (18) where d ak w 4 2  (19)  4 1 24*   mnmn (20) 3. solution using matlab programming, computer software with symbolic capabilities, solves the above set of equations. the program determines eigenvalues ( *mn ), for a given range of boundary conditions. the boundary’s linear translational non-dimensional restraint parameter can be defined as follows: d ak t t 3  (21) d ak w 4 2  (22) the following represent the input parameters to the program; (i) translational stiffness ratio ( t ); (ii) foundation ratio (  ); (iii) poisson’s ratio ( ); (iv) upper bound for eigenvalues ( n ); (v) suggested accuracy for eigenvalues ( d ); (vi) number of mode shape parameters ( n );. the program finds eigenvalues *mn by using matlab root finding function. 4. results and discussion the code developed is used to determine eigenvalues of any set or range of translational and foundation constraints. this code is also implemented for various plate materials by adjusting poisson’s ratio. such a wide range of results is not available in the literature. the eigenvalues for the plate edge, which is elastically restrained against translation and fully resting on the elastic foundation, at various values of the translational stiffness ratios, are computed and the results are given in table 1. the effects of the translational stiffness ratios are plotted in fig. 3. as seen from fig. 3, eigenvalues increase with an increment in the translational stiffness ratio, and the plates become unstable in the region when the translational stiffness ratio exceeds a certain value. twelve vibration modes are presented in fig. 3. the smoothened stepped variation is observed in fig. 3. the stepped region increases with increase in translational stiffness ratio and vibration modes. the location of the stepped region with respect to t changed gradually from the range of 0.01526 † [9.9997] ‡ – 5587.5316 [10] to 16.62296 [14.6739] – 611824.96917 [16.75055]. here †.represents translational stiffness ratio and ‡ represents eigen values throughout the text. here the value in the bracket represents eigenvalue. the simply supported boundary conditions (fig. 2) could be accounted for by setting (  t k ) shown in fig. 1. the frequency in this case is 2.23175 and this is in good agreement with the results published by wang (2005). another result, considered for comparison, is from rao and rao (2009) on a study of the case of vibrations of elastically restrained circular plates supported on partial winkler foundation. when the table 1. eigenvalues for different translational stiffness ratio for  =100 & =0.33. log10 𝑇 𝜆00 𝜆10 𝜆20 𝜆01 𝜆11 𝜆21 𝜆02 𝜆12 𝜆22 𝜆03 𝜆13 𝜆23 -3 10 10.0205 10.3517 10.10358 10.79568 12.46774 10.29701 11.43857 13.52984 10.63672 12.2505 14.67332 -2 10 10.02051 10.35171 10.10359 10.79568 12.46775 10.29702 11.43857 13.52985 10.63673 12.25051 14.67332 -1 10.00005 10.02061 10.3518 10.10368 10.79576 12.46779 10.29711 11.43864 13.52989 10.63682 12.25056 14.67335 0 10.00047 10.02162 10.35263 10.10462 10.79649 12.46827 10.29803 11.43927 13.53026 10.63769 12.25109 14.67365 1 10.003 10.03287 10.36125 10.11468 10.804 12.47302 10.30745 11.44568 13.53403 10.64649 12.25645 14.67667 2 10.00568 10.1199 10.46887 10.2319 10.89095 12.52351 10.42003 11.5759 13.57333 10.74585 12.31326 14.70772 3 10.00614 10.203 11.00421 10.49407 11.61576 13.1793 10.91134 12.27823 14.08459 11.41839 12.99719 15.0966 4 10.00619 10.21322 11.14468 10.53794 11.94459 14.15323 11.03916 12.88272 15.33948 11.70347 13.90786 16.53545 5 10.00619 10.21423 11.15725 10.54212 11.97273 14.2438 11.05093 12.93527 15.47979 11.72934 13.99516 16.74092 6 10.00619 10.21433 11.15849 10.54254 11.97546 14.25215 11.05209 12.94027 15.49231 11.73186 14.00329 16.75866 12 10.00619 10.21434 11.15862 10.54258 11.97576 14.25307 11.05222 12.94082 15.49369 11.73214 14.00418 16.7606 2 0 v ib ration s of c ircu lar p lates r estin g on e lastic f ou n d ation w ith e lastically r estrain ed e d g e a g ain st t ran slation table 2. eigenvalues for different foundation stiffness ratio for t =100 &  =0.33. table 3. eigenvalues for different translation and foundation stiffness ratios for  =0.33. log10 𝜉 & log10 𝑇 𝜆00 𝜆10 𝜆20 𝜆01 𝜆11 𝜆21 𝜆02 𝜆12 𝜆22 𝜆03 𝜆13 𝜆23 -3 0.2115 3.0115 6.2054 4.52915 7.73687 10.9091 5.93655 9.18564 12.3826 7.27469 10.57846 13.80849 -2 0.37648 3.01187 6.20544 4.52925 7.73689 10.90911 5.93659 9.18566 12.38261 7.27471 10.57846 13.8085 -1 0.67602 3.0156 6.20584 4.53032 7.73709 10.90918 5.93708 9.18578 12.38266 7.27499 10.57855 13.80854 0 1.30374 3.067 6.21074 4.54336 7.73963 10.91008 5.94302 9.18733 12.38328 7.27834 10.57959 13.80899 1 3.25317 3.90325 6.34932 4.87899 7.81237 10.93614 6.10341 9.23131 12.40119 7.36838 10.60869 13.82199 2 10.00568 10.1199 10.46887 10.2319 10.89095 12.52351 10.42003 11.5152 13.57333 10.74585 12.31326 14.70772 3 31.62297 31.62939 31.65958 31.63958 31.68744 31.78104 31.65573 31.72292 31.85232 31.67796 31.76831 31.94926 log10  𝜆00 𝜆10 𝜆20 𝜆01 𝜆11 𝜆21 𝜆02 𝜆12 𝜆22 𝜆03 𝜆13 𝜆23 -3 2.18341 4.70075 6.69703 5.56683 7.98678 10.99197 6.50356 9.33292 12.43922 7.59879 10.67534 13.84974 -2 2.18341 4.70075 6.69703 5.56683 7.98678 10.99197 6.50356 9.33292 12.43922 7.59879 10.67534 13.84973 -1 2.18365 4.70077 6.69704 5.56684 7.98679 10.99197 6.50357 9.33292 12.43922 7.59879 10.67535 13.84974 0 2.20704 4.70315 6.69786 5.56827 7.98727 10.99216 6.50447 9.33323 12.43935 7.59936 10.67555 13.84983 1 3.3284 4.92488 6.77875 5.7064 8.03541 11.01074 6.5926 9.36352 12.45218 7.65514 10.69583 13.84914 2 10.00568 10.1199 10.46887 10.2319 10.89095 12.52351 10.42003 11.5159 13.57333 10.74585 12.31326 14.70772 3 31.62296 31.62664 31.63867 31.63037 31.6549 31.73756 31.63691 31.68259 31.81038 31.6491 31.72496 31.90972 l . b . r ao an d c .k . r ao 2 1 vibrations of circular plates resting on elastic foundation with elastically restrained edge against translation 22 support position is in full span, which means that when 1b , the case becomes a circular plate having full foundation support with elastically restrained edge against translation. for this case, the frequency is 2.1834 and that is in good agreement with the frequency of 2.18341 obtained from the present study. the eigenvalues at various values of the foundation stiffness ratios for 33.0&100  t are computed and the results are given in table 2. the effects of the foundation stiffness ratio on eigenvalues are plotted in fig. 4. as seen from fig. 4, the eigenvalue increases with increase in the foundation stiffness ratio, and the plate becomes stiffer and stronger as the value of foundation stiffness becomes greater than 102. as seen from the tables 1 and 2, the influence of foundation stiffness ratio on eigenvalue is relatively greater than that of the translation stiffness ratio in increasing the overall natural frequencies of the plate support system. as seen from fig. 4, for all the modes considered here, up to a value of 10 the eigenvalues stay constant and beyond this value all the curves tend to converge to a constant eigenvalue as the foundation stiffness ratio increases up to 103. the convergence starts from 1.07897 [2.0325779] and continues up to a constant value of 9.63274 [13.84796]. the eigenvalues at various values of the translational stiffness ratios and foundation stiffness ratios are computed and the results are given in table 3. the effects of the translation and foundation stiffness ratios on eigenvalues are clearly observed in fig. 5, eigenvalues increases with an increment in both the translational and foundation stiffness ratios. as observed from the table 1 and 3, the influence of foundation stiffness ratio on eigenvalue is more predominant than that of translation stiffness ratio alone. from the results presented in tables 1 and 3, we can see that the influence of foundation stiffness ratio on eigenvalues is more predominant than that of translation stiffness ratio. from the results given in tables 1 to 3, one can easily find that the eigenvalues become lower for lower values of foundation and translation stiffness ratios. as seen from fig. 5, all the curves are stable up to a certain region beyond which the curves tend to converge for increasing values of translation and foundation stiffness ratios. figure 3. effect of translational stiffness ratio t on eigenvalues, mn  . l. b. rao and c.k. rao 23 figure 4. effect of foundation stiffness ratio,  on eigenvalues, .mn figure 5. effect of translational, t and foundation,  stiffness ratios on eigenvalues, .mn table 4. eigenvalues for different poisson ratios. ν 1000t 10,100  t 1000,10  t 10,1  t 50,50  t 0 31.62925 4.92456 10.03027 3.62553 7.28924 0.1 31.62929 4.92466 10.03111 3.64904 7.2896 0.2 31.62934 4.92476 10.0319 3.6708 7.28993 0.3 31.62938 4.92485 10.03265 3.69097 7.29025 0.4 31.62942 4.92494 10.03336 3.70974 7.29056 0 5 10 15 20 25 30 35 -4 -3 -2 -1 0 1 2 3 4 e ig e n v a lu e s , λ m n foundation parameter,log10 ξ λ00 λ10 λ20 λ01 λ11 λ21 λ02 λ12 λ22 λ03 λ13 λ23 vibrations of circular plates resting on elastic foundation with elastically restrained edge against translation 24 figure 6. effect of poisson ratio,  on eigenvalues, .mn 5. conclusion this work deals with a method of computation of eigenvalues of flexural vibrations of a circular plate with translational edge supported and resting on winkler foundation using a specifically written matlab code. the computed numerical results are presented in a tabular format to enable estimating the accuracy of approximate methods being used by other researchers for solving such problems. twodimensional plots of eigenvalues are drawn for a wide range of translational and foundation stiffness ratios facilitating their use in design. from the numerical and graphical results presented in this paper, it can be easily observed that the eigenvalues remain constant only for a limited range of constraints specific to each vibration mode and then steeply increase with the increasing values of foundation stiffness ultimately converging towards a constant value. it is also observed that the influence of foundation stiffness ratio on eigenvalues is more predominant than that of translational stiffness ratio. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references ascione l, grimaldi a (1984), unilateral contact between a plate and an elastic foundation. mechanica 19(3): 223-233. bhaskara rl, kameswara rc (2009), buckling of circular plates with a ring support and elastically restrained edge. advances in vibration engineering 8(1): 61-69. bhaskara rl, kameswara rc (2010), buckling analysis of circular plates with elastically restrained edges and resting on internal elastic ring support. mechanics based design of structures and machines 38(4): 440-452. bhaskara rl, kameswara rc (2015), vibrations of a rotationally restrained circular plate resting on a concentric rigid ring support. international journal of acoustics and vibration 20(4): 220-225. celep z (1988), circular plate on tensionless winkler foundation. journal of engineering mechanics 114(10): 1723-1739. chang jy, wickert ja (2001), response of modulated doublet modes to traveling wave excitation. journal of sound and vibration 242(1): 69-83. dempsey jp, keer lm, patel nb, glasser ml (1984), contact between plates and unilateral supports. journal of applied mechanics 51(2): 324-328 ghosh ak. (1997), axisymmetric dynamic response of a circular plate on an elastic foundation. journal of sound and vibration 205(1): 112-120. kang kh, kim kj (1996), modal properties of beams and plates on resilient supports with l. b. rao and c.k. rao 25 rotational and translational complex stiffness. journal of sound and vibration 190(2): 207-220. kim m, moon j, wickert ja (2000), spatial modulation of repeated vibration modes in rotationally periodic structures. journal of vibration and acoustics, transactions of asme 122(1): 62-68. kim km, kim ms (2001), vibration of beams with generally restrained boundary conditions using fourier series. journal of sound and vibration 245(5): 771-784. leissa aw (1969), vibration of plates (nasa sp-160). office of technology utilization, washington, dc. leissa aw (1993), vibration of plates. acoustical society of america, sewickley, pa. lokavarapu brl, chellapilla kr (2013), fundamental buckling of circular plates with elastically restrained edges and resting on concentric rigid ring support. frontiers of mechanical engineering 8(3): 291-297. lokavarapu brl, chellapilla kr (2015), buckling of circular plate with foundation and elastic edge. international journal of mechanics and materials in design 11(2): 149156. rao ck, rao lb (2009), vibrations of elastically restrained circular plates resting on partial winkler foundation. the open acoustics journal 2: 68-74. rao ck, rao lb (2016), vibrations of circular plate supported on a rigid concentric ring with translational restraint boundary. engineering transactions 64(3): 259-269. salari m, bert cm, striz ag (1987), free vibration of a solid circular plate free at its edges and attached to a winbkler foundation. journal off sound and vibration 118(1): 188-191. soedel w (1993), vibrations of shells and plates. marcel dekker, inc., new york. tseng jg, wickert ja (1994), on the vibration of bolted plate and flange assemblies. journal of vibration and acoustics, transactions of asme 116(4): 468-473. wang jts, lin cc (1996), dynamic analysis of generally supported beams using fourier series. journal of sound and vibration 196(3): 285-293. wang cy, wang cm (2003), fundamental frequencies of circular plates with internal elastic ring support. journal of sound and vibration 263: 1071-1078. wang cy (2005), fundamental frequency of a circular plate supported by a partial elastic foundation. journal of sound and vibration 285(4-5): 1203-1209. weisman y (1970), on foundations that react in compression only. journal of applied mechanics 37: 1019-1030. winkler e. (1867), die lehre von der elasticitaet and festigkeit, prag. dominicus. yayli mö, aras m, aksoy s (2014), an efficient analytical method for vibration analysis of a beam on elastic foundation with elastically restrained ends. shock and vibration 2014: 17. zheng xj, zhou yh (1988), exact solution of nonlinear circular plate on elasticfoundation. journal of engineering mechanicsasce 114: 1303-1316. the journal of engineering research (tjer), vol. 15, no. 1 (2018) 61-72 doi: 10.24200/tjer.vol15iss1pp61-72 constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system n. hosseinzadeh*, a. al shereiqi, m. al badi, and a. al hinai department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, pc 123, muscat – al-khoud, oman. received 27 january 2017; accepted 13 april 2017 abstract: this paper investigates the possibility of constructing multi-microgrids by interlinking the rural area systems in the al wusta governorate of the sultanate of oman, which are currently being supplied by diesel generators. it is proposed to enhance the rural system under study by switching off small diesel stations and replacing them with wind turbines. the microgrids formed in this way are then interlinked together to create multi-microgrids. the paper studies the interlinked multimicrogrids under different scenarios; in terms of voltage profiles and power flow using the etap software package. this study contributes to the feasibility study of retiring some diesel power plants and using renewable energy resources in rural oman. keywords: microgrids; distributed generation; renewable energy; distributed energy resources; multi-microgrids; rural electricity; oman. الريفية يف كهرباء املناطقالطاقة املتجددة يف نظام باستخدام صغريةانشاء شبكة متعددة عمان عامر اهلنائي، ، عبد اهلل الشريقي، حممد البادي*ناصر حسني زاده ربط أنظمة املناطق الريفية يف حمافظة شبكة صغرية متعددة و ذلك بانشاء تبحث هذه الورقة يف إمكانية خلخ: امل باستخدام مولدات الديزل. ومن املقرتح تعزيز النظام الريفي قيد الدراسة تعمل حاليا ، واليت الوسطى بسخلطنة عمان الشبكات الصغرية املنشأة و يتم ربط عن طريق إيقاف حمطات الديزل الصغرية واستبداهلا بالتوربينات الرحيية. عخلى دراسة الشبكة الصغرية الورقة كما تعمل .متعددة ببعضها و ذلك لتكوين شبكة صغرية بهذه الطريقة باستخدام حزمة برامج و ذلك كهرباءاجلهد وتدفق ال نوعحتت سيناريوهات خمتخلفة من حيث املتعددة املرتابطة الديزل واستخدام ب اليت تعمل كهرباءبعض حمطات توليد ال ايقافيف دراسة جدوى بحثال اساهم هذيإيتاب. و ق الريفية يف عمان.الطاقة املتجددة يف املناط موارد ، شبكة دقيقة متعددة ،املوزع، الطاقة املتجددة، موارد الطاقة املوزعة توليد، الشبكة دقيقة فتاحيةالكخلمات امل كهرباء الريف، عمان. * corresponding author’s e-mail: hosseinz@squ.edu.om mailto:hosseinz@squ.edu.om n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 62 1. introduction in any society, an adequate supply of energy is required for the process of modernization, increase in productivity and improvement of standards of living. electrical energy is used more than other forms of energy because it is clean and easier to transport. therefore, both the design and construction of electrical system have to be done professionally and according to international standards. supplying electricity to rural areas, particularly in countries where a small population is dispersed within a large area, is considered to be a challenge. it is argued that extension of the main grid to cover all rural areas is usually a very expensive option compared to other solutions (mahapatra and dasappa 2012). in most cases, micro grids ‎(marnay and venkataramanan 2006) in either stand-alone mode (kobayakawa and kandpal 2016) or, if possible, in a gridconnected mode are a more viable option. other forms of cost-effective extension of power grids to rural areas have been proposed in the literature (hosseinzadeh et al. 2011), but with recent advances in electrical generation using renewable energy resources, it appears that developing microgrids in rural areas is a very promising solution for electrification of remote communities from the main power grid. renewable resources especially solar and wind energy are recommended for use in oman according to the previous research reports. it has been reported that many studies have indicated that oman could benefit from its rich solar and wind energy resources (al hatmi and tan 2013). microgrids for different applications with different levels of control were developed around the globe. many types of microgrids with power management capabilities or various control systems have been introduced in the literature, such as (khodayar et al. 2012; sun et al. 2015; chuang et al. 2016; li et al. 2016; meng et al. 2016; mohan et al. 2016). in this paper, a proposal will be presented to modify the electricity supply of three small towns in the rural area of oman’s electricity network, namely, hijj, najdah and khaloof in the al wusta governorate in the sultanate of oman. electric power is currently supplied separately to these three towns using diesel generators. as diesel generators have high operating costs and contribute to environmental pollution, it is aimed to reduce their usage by retiring the small diesel generators, adding renewable energy resources, and interlinking the three electricity networks. 2. microgrid generally, any typical power network consists of power generation, transmission and distribution. naturally, various energy resources, such as diesel, gas, wind and solar, could be used for generating electrical power. investing in these resources could be done by adopting the concept of microgrid to develop the electricity system. a microgrid (hatziargyriou, asano et al. 2007) is a small system which can generate and distribute electrical power to local consumers and can also be connected to a grid. therefore, it can work in grid-connected mode or islanded mode. the latter mode may be used in failure occurrences in the main grid. in the absence of a grid in the area, the microgrid may be designed to operate completely off-grid. microgrids in power systems have been in use for a long time, but recently they are drawing more attention due to expansion of renewable energy resources (hemmatpour, mohammadian et al. 2016; xu, nthontho et al. 2016). the lack of sufficient resources to balance out the rapid increase in demand and the availability of local generators in the form of renewable-energy power plants have made it possible for electricity utilities to use microgrids for supporting the main grid. microgrids may use distributed generators (dg) from different resources. one essential requirement will be controlling and coordinating these dgs by taking into consideration the technical and economic aspects of microgrid operation (esmaeli 2016; khederzadeh et al. 2016; moradi et al. 2016; nikmehr and ravadanegh 2016). a multi-microgrid is formed by connecting some adjacent microgrids (schwaegerl and tao 2014). in this paper, the first phase of forming a multimicrogrid by connecting the microgrids of hijj, najdah, and khaloof is proposed. a microgrid represents energy distribution architecture from distributed generators to the customer and the interconnection between these sources and the main grid while they offer considerable control capabilities over the network operation (costa, matos et al. 2008; jiayi, chuanwen et al. 2008; shahnia 2011). a simple microgrid structure, which is suitable for oman rural areas, is shown in fig. 1. for the constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system 63 case under study, three such microgrids linked together are proposed. however, in this study, storage units are not included and renewable resources are limited to wind-energy systems. furthermore, the system is currently not connected to the utility grid. the dashed rectangle in fig. 1 shows that the microgrid should be a single controllable entity. controllers in a microgrid are required to keep functioning in required operational modes (lasseter 2011). a microgrid, which is connected to the main grid acts like a plug that can easily isolate itself during any abnormality or interruption in the main grid and can continue supplying the load in islanded mode. in gridconnected mode, the microgrid is connected to the main grid via a switch at the point of common coupling (pcc) as shown in fig. 1. rural areas in oman are far away from the main interconnected system (mis). diesel generators are currently used in most of these areas. accordingly, an increase in the demand could lead to a huge cost increase in operation and maintenance of the small scattered power stations. additionally, some of the power stations are old, costly to operate and have poor power efficiency. interlinking the small power stations, leaving bigger ones and retiring small power stations are some of the suggested solutions, which aim to reduce maintenance costs, improve system reliability, reduce operational costs, increase efficiency, and also to reduce reliance on government subsidies. for retiring some diesel generators, renewable resources are considered as auxiliary supporters to the rural power generation. it is well known that most of the rural areas in oman have ample wind and solar resources. in the case of this study, a combination of the existing diesel power stations with wind turbines for constructing a microgrid could be one of the best solutions to reduce the use of diesel generators while keeping proper voltage profiles in the power system under study. it is important to mention that different variable-speed generators may be used in the wind turbines, e.g. squirrel cage induction generators (scig), synchronous generators (sg), permanent magnet synchronous generators (pmsg) and doubly fed induction generators (dfig). the dfig technology is mostly used in modern wind turbines since it allows a variable-speed operation over a large range. it contains a wind rotor connected to the grid through a converter and stator circuit which is directly connected to the grid. another advantage of the dfig technology is the ability of power electronic converters to generate or absorb reactive power, thus controlling the power factor (senthilkumar and narmadhadevi 2013). the expected outcome from this study is that a multi-microgrid will make a major contribution to increase systems efficiency and reliability especially by deploying renewable energy systems. load generator renewables storage m icrogrid utility grid switchable connect ion figure 1. a microgrid structure suitable for rural areas of oman. n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 64 3. case study this case study is based on modelling diesel and wind resources using etap software in order to study the impact of this hybrid system on the power flow in an existing network and to explore the possibility of constructing multimicrogrid. the case study is divided into two parts, the first part is about running the system with diesel generators only and the second one introduces wind turbines diesel generators. 3.1 part i: interlinking systems using diesel generators only possibility of switching off diesel generators in the smaller villages of khaloof and najdah, and interlinking the three systems using bigger diesel generators in hijj station only. the original system consists of three isolated microgrids with diesel generators in hijj, khaloof, and najdah (see fig. 2). the first study suggests interlinking khaloof and najdah to hijj by switching off the smaller diesel generators in khaloof and najdah, and adding the dotted lines f1-k and f1-n with the associate step-up and step-down transformers as in fig. 3. in this part, the focus will be on studying the figure 2. existing systems in khaloof, hijj and najdah. figure 3. full scheme for the proposed microgrid. --- dashed lines are short underground cables. constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system 65 the dashed lines in the figure represent the short cables of about 200m in length used inside the power station from bus-bar to the step-up transformers. the idea behind this interlinking is to minimize operational cost and to reduce the number of low-efficient power plants. hijj plant is the main power station in this proposal, whereas najdah and khaloof power plants are kept on standby to be used as and whenever needed. one important objective that should be considered in constructing power stations is fuel efficiency. in 2013, fuel consumption in liter/kwh in 2013 was 3.54, 3.0 and 3.0 for hijj, khaloof and najdah power stations, respectively. the retired diesel generators can be used as backup power supplies during power interruption and can also be used during maintenance work in the main power station. it is worth mentioning here that the rural area electricity company (raec) system is not currently connected to the main grid. thus, the pcc in fig. 2 is in the open state, i.e. the microgrid is operated in stand-alone (or offgrid) mode. in the future, the raec system may get interconnected with the mis. additionally, the distributed generators inside the microgrid are supposed to control opening and closing of the switches according to the circumstances and needs of the supply. however, this is beyond the scope of this paper. controllers are needed for dispatching and regulating load among the dgs. however, at present, there is no scada system in the existing network. so, for the time being, the supply can be controlled manually by the system operators. 3.2 part ii: interlinking systems using hybrid wind-diesel resources this case study is a proposal for combined use of wind and diesel energy to determine whether wind energy can help reduce usage of diesel generators while providing enough energy for demand at all times. this combination will be helpful to minimize the use of diesel generators in the seasons and times when there is sufficient wind speed, which should improve voltage profile at the same time of reducing the use of diesel generators. in the present scenario, stand-alone diesel generators in hijj power station feed the load. due to the drawbacks of diesel generators and advantage of wind energy, it is proposed that wind turbines be used in the same location of the switched off stations which are najdah and khaloof. this reduces the operational and maintenance cost of the existing diesel generators in hijj power station and enhances the advantages of using renewable resources in rural area systems. although developing a methodology to design and implement hybrid wind-diesel microgrid could be very useful, the main objective of this paper is to do a feasibility study of a multi-microgrid from a system point of view. therefore, the rated power of wind turbines chosen here are simply selected to be close to the rate of the existing diesel generators in khaloof and najdah power stations. the wind turbines are connected to the first bus of khaloof system and the first bus of najdah system with a rated power of 2.5 mw each to replace the switched off diesel station. start is wind speed � cut-in speed ? is available wind power � demand ? dispatch wind power without using diesel generators dispatch wind power and provide extra supply using diesel generators supply the whole demand by diesel generators end yes no yes no figure 4. flow chart of the process of supplying load by the micro-grid. n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 66 the process of supplying the load using both diesel and wind resources is summarized in the diagram of fig. 4. in this flowchart, it is assumed that some storage units are available to store extra energy produced by wind turbines when wind speed is high and demand is low. the study in this paper is mainly carried out using power flow analysis. the parameters of wind energy system affecting the power flow are presented in table 1. identical parameters are used for khaloof and najdah. the factors affecting wind turbine electricity production are: wind speed (v), air density (ρ) and area swept by the rotor blades (a). the power available in the wind could be described by eq. (1) whereas the output power from the turbines is described by eq. (2), where cp is the power coefficient. pwind = 0.5(ρav3) (1) pout = cp(pwind) (2) wind systems were assumed to work at their full capacity, according to the available data for najdah and khaloof stations or nearby sites. in this case, the available data for joba, a nearby village, was used since it is the closest weather station to najdah and khaloof. the air density ρ = 1.175 kg/m3 is used, taken from (senthilkumar and narmadhadevi 2013), and the average wind speed is 5.2 m/s according to the output of the study carried out by oman authority for electricity regulation. the turbine page of the etap software shown in fig. 5 is used to define the characteristics of the turbine by entering the required data manually and representing the power curves automatically. the air density value in the sea level is usually adjusted to 1.225 kg/m3, but it is calculated as a function of air pressure and air temperature to get the value of 1.175 kg/m3. the pitch angle shown in the turbine page does not affect the power flow directly as there is a pitch controller, which keeps the rotor speed within operating limits. the wind characteristics page in fig. 5 is used to enter wind-related factors that etap can use to create a wind profile for a wind farm. as per the annual mean speed records in the weather station, a height of 40 meters above the ground is needed to reach an average speed of 6 m/s. therefore, the tower height should not be less than 40 meters and the rotor diameter should be suitable with that height. the tower height chosen in this study is 65 meters, 60 meters of which is the rotor diameter. in addition, to investigate the possibility of using smaller wind turbines, simulation study is repeated for 39 meters rotor diameter to compare the output power with 60 meters rotor. the output voltage of the wind turbines is chosen as 11 kv to reduce losses and avoid the need for a transformer. the system in fig. 3 was modified to represent the diesel/wind hybrid system shown in fig. 7. the only difference with fig. 3 is the replacement of the retired diesel generators at khaloof and najdah by wind energy systems. table 1. technical specifications of the wind turbine. description selected wind energy system rated power (each turbine) 0.5 mw rated wind speed 15 m/s cut-in wind speed 4 m/s cut-out wind speed 25 m/s rotor diameter 60 m tower height 65 m swept area (π r2) 2828 m2 rated voltage 12 kv no. of turbines 5 no. of blades 3 constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system 67 figure 5. aerodynamics page of wind turbine. figure 6. wind characteristics page of wind turbine. figure 7. full scheme for the proposed microgrid using wind and diesel. n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 68 4. results 4.1 part i: interlinking systems using diesel generators only in this case, the three systems are interlinked for constructing the microgrid, with najdah and khaloof stations switched off and the microgrid fed from the diesel generators in hijj station, as shown in fig. 3. on running power flow using etap, as shown in fig. 8, a warning indicator was noticed for overload in generation by more than 4.2% from the specified maximum output, which is given as 12.766 mw at 0.9 power factor. therefore, the overload issue has to be solved by upgrading the generators inside hijj power station by 6.4% from their existing capacity. this means that the ratings of the power station need to be upgraded from 12.776 mw (14.2 mva) to 13.59 mw (15.1 mva). after upgrading the generators, the installation of step-up and step-down transformers in both systems has been taken into consideration. power flow analysis showed that without transformers, both khaloof system and najdah system suffered from a voltage drop as high as 19%. there are different approaches to improve the voltage profile and reduce the voltage drop; the easiest way is to improve the voltage profile is by the installation of step-up or step-down transformers (vandoorn, de kooning et al. 2013). for interlinking hijj, khaloof and najdah systems, 2 mva step-up and step-down transformers were used. as shown in table 2, no significant voltage drop or overload issue appeared in the system and the apparent power was within the specified limits. permissible voltage drop in 11, 33 and 0.415 kv systems is ±6% of the rated voltage according to oman distribution code. the current rating of 11 kv bus bar at hijj power station is 630 a for each feeder. 4.2 part ii: interlinking systems using hybrid wind-diesel resources after adding wind energy systems to khaloof and najdah, the system performance improved considerably. the power flow results are shown in table 3 and fig. 9. the voltage profile improved in both systems and the overall losses in the system became 0.262 – j0.543 mva, which is obviously better than merely using diesel generators. in this case, najdah and khaloof diesel generator stations are switched off and the microgrid is fed from the diesel generators in hijj station and local wind generators in najdah and khaloof. however, while running the power flow using etap, an alert was noticed for overload in generation by more than 4.2% from the specified maximum output. this is due to the fact that wind energy systems usually produce less power output than their capacity. however, the overload issue can be solved by upgrading the generators inside hijj power station by 6.4% from their existing capacity. this means that power station rate will change from 12.776 mw to 13.59 mw. the contribution of the sources is shown in table 4, which shows that the hijj power station has been upgraded to 13.59 mw, 0.92 power factor, and each wind generator supplies only 4.3% of the total required generation. this, of course, depends on wind speed and some other conditions. figure 8. power flow for the outgoing feeders from hijj station in scenario 1. constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system 69 table 3. power flow using wind turbines in khaloof in najdah stations. f1-n f1-k f3-h f2-h f1-h feeder 0.9+j0.2 0.7+j0.08 2.1+j1 2.5+j1.2 4.5+j2.2 s(kva) at (beginning of feeder) table 4. contributions of wind turbines in khaloof in najdah stations. source type rating mw active mw % local supply contribution % pf % total supply contribution gen1 syn. 13.59 9.926 100 91.97 91.6 wtg1 wind. 2.5 0.456 60.5 90.07 4.2 wtg2 wind. 2.5 0.456 59.6 90.07 4.2 table 5. increasing wind turbines in khaloof in najdah stations. source type rating mw active mw % pf voltage (% mag.) gen1 syn. 13.59 9.114 90.1 100 wtg1 wind. 6.0 1.095 90.4 104.49 wtg2 wind. 4.0 0.730 91.1 101.07 table 6. daily wind turbines contribution in generation. morning evening station gen. power at min. speed (3.2 m/s) gen. power at max. speed (4.5 m/s) gen. power at min. speed (5.8 m/s) gen. power at max. speed (8.1 m/s) s(mva) s (mva) s (mva) s (mva) khaloof j 0.2 0.29+j0.2 0.6+j0.2 1.72+j0.2 najdah j 0.2 0.29+j0.2 0.6+j0.2 1.72+j0.2 a) najdah b) khaloof figure 9. power flow results of wind turbines in khaloof and najdah. n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 70 wind energy systems are only proposed for najdah and khaloof, and not for hijj. this can be considered in the future. the revised table 4 shows that in najdah 506 kw is supplied by wind turbines, i.e. 60.5% of the required power in najdah. similarly, 59.6% of power used in khaloof is supplied by wind turbines. however, only 8.4% of the total required power in hijj, khaloof and najdah systems is supplied by wind energy systems. this is due to the fact that no renewable resource is proposed for hijj power station. in all the above cases, annual average wind speed is used. however, the daily wind speed gives a clearer picture of wind turbines’ contribution of wind turbines in supplying power. it is expected that in some seasons and times, wind energy systems contribute much more to the generation of the electrical power needed to supply the loads. studying the power flow using daily wind speed, as shown in table 6, shows that the maximum output from turbines is during the evening when they comply with maximum power demand. rotor diameter is one of the parameters that affects the output power. there is proportionality between the output power and rotor diameter. table 7 presents the output from the wind turbines using 39 meters rotors. in comparing table 6 with table 7, it is observed that the output power of wind turbines with 39 meters rotors for the same wind speed is reduced to about half of the output of 60 meters rotors. this suggests the selection of 60 meters rotor instead of 39 meters rotor if single wind turbine is used. unquestionably, another suggested option, in this case, is to use more wind turbines. so, in that case, it is expected that even with 39 meters rotors, a good portion of the required power can be supplied by wind energy systems. 5. proposing future work increasing continuously. it is proposed that the rural power system of three small towns can be interlinked to work as a multi-microgrid. this system needs to be made smart by using renewable resources with advanced control systems (hamzeh, emamian et al. 2016). at the same time, switching off non-efficient diesel stations has to be the priority of the electricity company authorized in rural areas of oman. with respect to this, there are some arising issues (see below) which have not been addressed in this paper and will be studied as future research.  controlling the microgrid (sheng, wang et al. 2010; shafiee, guerrero et al. 2014) in a practical way applied to the case study by constructing scada systems and building local and central controllers with advanced technologies for controlling the power (senthilkumar and narmadhadevi 2013) and dispatching it between the sources of a microgrid (rodriguez-amenedo, arnalte et al. 2002; varaiya, wu et al. 2011).  since there is a plan to make the grid available close to the rural network under study, future research may also focus on the microgrid in the grid-connected mode instead of the study carried out for islanded mode. currently, oman’s main grid is around 250 km away from the hijj system. there is a possibility of extending 132 kv up to hijj, which will make the shutting down of the diesel power station possible.  studying the financial and environment aspects of the constructed microgrid.  developing a methodology to design and implement hybrid wind-diesel microgrid.  constructing the proposed wind turbines in hijj and test them in practice.  a regulatory framework is needed for encouraging consumers to participate in power generation using renewable resources. the demand of power in rural areas is table 7. sources output using 39 meter rotor diameter. source rating (mw) active (mw) reactive (mvar) % pf % generation wtg1 najdah 2.5 0.193 0.22 65.9 2.1 wtg2 khaloof 2.5 0.193 0.22 65.9 2.1 constructing a multi-microgrid with the inclusion of renewable energy in oman's rural power system 71 6. conclusion in this paper, it is proposed to use wind energy conversion systems (wecs) as a part of a multimicrogrid for the existing rural networks in hijj, khaloof and najdah in the sultanate of oman. the wecs technology is used to reduce the usage of diesel generators, which have high operating costs and also add to environmental pollution. it is proposed that some small power stations that use diesel generators be retired and instead power flow be managed by interlinking existing isolated power systems plus adding wecs where feasible. the system performance is studied mainly by power flow analysis of the system before and after interlinking between the aforementioned systems. the study is carried out on an actual power system in rural oman with all parameters obtained from a real system in operation. it was shown that microgrid could be enhanced by using wind turbines in place of switched-off power stations. these wind turbines ensure the system reliability while producing less pollution compared to diesel generators. however, more advanced control is needed for converting these intermittent sources of energy to dispatch-able generators. this paper contributes to the technical feasibility study of using renewable energy sources and reducing the usage of diesel generators in rural oman. the financial and environment aspects of the constructed microgrid need to be studied too. these studies are essential before making a final decision on constructing the proposed microgrid. to study the financial aspects pertaining to this microgrid, market price has to be provided in addition with actual bill of quantity needed for construction or upgrade of power station, feeders, substations, etc. this is under study by the authors. conflict of interest the authors declare no conflicts of interest. funding this work is funded by the research council (trc) grant org-squ-e1-11-009 (rc/eng/eced/12/02). acknowledgment the authors would like to thank squ, the research council, and oman rural area electricity company (raec) for their support to this research work. references al hatmi y, tan c (2013), issues and challenges with renewable energy in oman. gas (bcm) 4(9): 212-218. ceh ehc (2013), introduction to microgrids. chuang sj, hong cm, chen ch (2016), design of intelligent control for stabilization of microgrid system. international journal of electrical power & energy systems 82: 569578. costa pm, matos ma, lopes jp (2008), regulation of microgeneration and microgrids. energy policy 36(10): 3893-3904. esmaeli a (2016), stability analysis and control of microgrids by sliding mode control. international journal of electrical power and energy systems 78: 22-28. hamzeh m, emamian s, karimi h, mahseredjian j (2016), robust control of an islanded microgrid under unbalanced and nonlinear load conditions. ieee journal of emerging and selected topic in power electronics 4(2): 512-520. hatziargyriou n, asano h, iravani r, marnay c (2007), microgrids. power and energy magazine ieee 5(4): 78-94. hemmatpour mh, mohammadian m, gharaveisi aa (2016), optimum islanded microgrid reconfiguration based on maximization of system loadability and minimization of power losses. international journal of electrical power and energy systems 78: 343-355. hosseinzadeh n, mayer je, wolfs pj (2011), rural single wire earth return distribution networks–associated problems and costeffective solutions. international journal of electrical power and energy systems 33(2): 159-170. jiayi h, chuanwen j, rong x (2008), a review on distributed energy resources and microgrid. renewable and sustainable energy reviews 12(9): 2472-2483. n. hosseinzadeh, a. al shereiqi, m. al badi and a. al hinai 72 khederzadeh m, maleki h, asgharian v (2016), frequency control improvement of two adjacent microgrids in autonomous mode using back to back voltage-sourced converters. international journal of electrical power and energy systems 74: 126-133. khodayar me, barati m, shahidehpour m (2012), integration of high reliability distribution system in microgrid operation. ieee transactions on smart grid 3(4): 19972006. lasseter rh (2011), smart distribution: coupled microgrids. proceedings of the ieee 99(6): 1074-1082. li c, liu x, cao y, zhang p, shi h, ren l, kuang y (2016), a time-scale adaptive dispatch method for renewable energy power supply systems on islands. ieee transactions on smart grid 7(2): 1069-1078. marnay c, venkataramanan g (2006), microgrids in the evolving electricity generation and delivery infrastructure. paper presented at the 2006 ieee power engineering society general meeting. meng l, luna a, diaz er, sun b, dragicevic t, savaghebi m, andrade f (2016), flexible system integration and advanced hierarchical control architectures in the microgrid research laboratory of aalborg university. ieee transactions on industry applications 52(2): 1736-1749. mohan v, singh jg, ongsakul w, suresh mr (2016), performance enhancement of online energy scheduling in a radial utility distribution microgrid. international journal of electrical power and energy systems 79: 98107. moradi mh, eskandari m, hosseinian sm (2016), cooperative control strategy of energy storage systems and micro sources for stabilizing microgrids in different operation modes. international journal of electrical power and energy systems 78: 390400. nikmehr n, ravadanegh sn (2016), reliability evaluation of multi-microgrids considering optimal operation of small scale energy zones under load-generation uncertainties. international journal of electrical power and energy systems 78: 80-87. oman-aer (2008), study on renewable energy resources. regulation afe, (2008). study on renewable energy resources, from http://www.aer-oman.org rodriguez-amenedo jl, arnalte s, burgos jc (2002), automatic generation control of a wind farm with variable speed wind turbines. ieee transactions on energy conversion 17(2): 279-284. schwaegerl c, tao l (2014), the microgrids concept. microgrids: architectures and control 1-24. senthilkumar a, narmadhadevi v (2013), real and reactive power control of doubly-fed induction generator during voltage sag. international journal of engineering research and technology. shafiee q, guerrero jm, vasquez jc (2014), distributed secondary control for islanded microgrids—a novel approach. ieee transactions on power electronics 29(2): 10181031. shahnia f (2011), analysis and correction of voltage profile in low voltage distribution networks containing photovoltaic cells and electric vehicles. queensland university of technology. sheng sr, wang xy, xu hz, zhu gq, zhou yf (2010), anatomy of large animal spines and its comparison to the human spine: a systematic review. euro spine journal 19(1): 46-56. sun q, zhou j, guerrer jm, zhang h (2015), hybrid three-phase/single-phase microgrid architecture with power management capabilities. ieee transactions on power electronics 30(10): 5964-5977. vandoorn tl, de kooning jd, meersman b, guerrero jm, vandevelde l (2013), voltagebased control of a smart transformer in a microgrid. ieee transactions on industrial electronics 60(4): 1291-1305. varaiya pp, wu ff, bialek jw (2011), smart operation of smart grid: risk-limiting dispatch. proceedings of the ieee 99(1): 40-57. xu z, nthontho m, chowdhury s (2016), rural electrification implementation strategies through microgrid approach in south african context. international journal of electrical power and energy systems 82: 452465. http://www.aer-oman.org/ microsoft word paper 4.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 137-148 a comparison of trends in extreme rainfall using 20-year data in three major cities in oman l.n. gunawardhana* and g.a. al-rawas department of civil and architectural engineering, college of engineering, sultan qaboos university, oman. received 25 november 2015; accepted 9 march 2016 abstract: many regions in the world have recently experienced more frequent and intensive disasters such as flash floods and persistent droughts. the sultanate of oman is no exception to this. we analyzed twodecade long daily precipitation records in three major cities, namely, sohar, muscat and salalah, mainly focusing on extremes. a set of climate indices defined in the rclimdex software package was used. moreover, annual maximum 1-day precipitations in three study areas were analyzed using the generalized extreme value (gev) distribution function. results showed significant changes in the precipitation regime in recent years. the annual total precipitation in sohar and salalah decreased, while that in muscat shows statistically week increasing trend. however, all indices analyzed indicate enhanced extreme precipitation toward 2010 in muscat and salalah. as a result, the contribution from extreme events to the annual total rainfall steadily increases in both study areas. a clear conclusion could not be made based on selected indices for sohar due to consistent drier years occurred from 1999 to 2005. frequency analysis indicates that the annual the maximum 1-day rainfall estimated in sohar and muscat for 5 and 10 year return periods are approximately same (70 mm/day and 108 mm/day, respectively) but about two-fold greater than that in salalah (29 mm/day and 60 mm/day, respectively). keywords: gev distribution, rclimdex, rx-1day, sohar, muscat, salalah. אאאאא٢٠  אאא،אא א:א א  א א אא  א א  א  א א א   wאאאkאkא א،אאאאאאkאא אאאאאאאאkאא kאאאאאאkאאאאאא  א א      ،      א א    אאאאא،kאאא٢٠١٠א kאאאאאאא،k אאאאאאאאאאאא ١٩٩٩٢٠٠٥אאאאאאk٥١٠א f٧٠l١٠٨fא،eאא،l٢٩l٦٠،l keאא  אאwאא،אאאא،א،،،  * corresponding author’s e-mail: luminda@squ.edu.om l.n. gunawardhana and g.a. al-rawas   138 1. introduction in recent decades, weather has gotten more extreme in many parts of the world. these changes have been detected in the arabian peninsula (alsarmi and washigton 2013), asia (liu et al. 2015), america (coopersmith et al. 2014), and europe (haren et al. 2013). in a regional scale analysis, groisman et al. (2005) reported that in the mid-latitudes there has been a notable increase in the frequency of very heavy precipitation (upper 0.3% of daily precipitation) during the past 50–100 years. alexander et al. (2006) reported a significant decrease in the annual occurrence of cold nights and a significant increase in the annual occurrence of warm nights over two-thirds of the global land area. these extreme weather events often cause extensive damage to economies, the environment, and human life (tezuka et al. 2014). for example, exposure to extreme heat can cause heat stroke, which may lead to death or permanent disability. changes in a precipitation regime may result in flash floods and long-term droughts. moreover, some plants and animals are highly dependent on specific climate conditions. any changes in the climate of an area can affect them and the entire ecosystem (tank et al. 2009). consequences of extreme weather events affect regions in different ways, and are projected to worsen conditions in water-scarce countries like oman. gunasekara et al. (2013) analyzed the combined impact of population change and climate change, and identified the middle east as one of four regions that will be exposed to extremely negative impacts that could lower fresh water availability by 50% by 2100 compared to the situation in 2000. in oman, changes in precipitation and temperature have already begun to be detected. for example, the super cyclone gonu in 2007 caused extensive damage in coastal cities, with rainfall recorded at 256 mm/day at seeb airport in muscat governorate. the cyclone caused about usd$4.4 billion in damages at the time and 69 deaths. numerous studies based on global circulation model (gcm) and regional climate model (rcm) projections indicate that anthropogenic climate change could cause an increase in the frequency and intensity of extreme weather events (beniston et al. 2007; fowler et al. 2005). under changing climate conditions (past and future), exclusive use of climatic norms derived from historical data may no longer be appropriate and could render infrastructure vulnerable by leading to design with insufficient capacity or by prompting the adoption of policy measures that become difficult to practice over time. therefore, understanding the occurrence of past extremes is of great importance to avoid or at least reduce the damage due to catastrophic floods and prolonged periods of drought. the objective of this research is therefore to evaluate the trends of extreme precipitation in three of oman’s major cities— muscat, salalah, and sohar—over recent decades. 2. study area oman is located in the southeastern corner of the arabian peninsula. it covers an area of 309,500 km2 and encompasses a diverse range of topography, including mountain ranges, arid deserts, and fertile plains. the climate varies mainly from semi-arid to hyper-arid. on average, annual rainfall is less than 100 mm compared to a global annual mean of 1,123 mm. the city of sohar is located in al batinah north governorate in northern oman (fig. 1). the annual total precipitation average between 1991–2010 was approximately 106 mm (fig. 2), which suggests a strong negative trend in rainfall in the 21st century (67 mm/decade, p = 0.05). a p-value of <0.05 leads a researcher to reject the null hypothesis and conclude that there is a significant linear trend. cold frontal troughs originating in the north atlantic ocean or mediterranean sea bring nearly 75% of figure 1. three study areas. a comparison of trends in extreme rainfall using 20 years data in three major cities in oman  139 sohar’s total rainfall from november to april (fig. 3). the number of wet days, which are defined as days with precipitation >1 mm, occur in sohar about 11.7 days/year, which is higher than the number of wet days that occur in muscat (7 days/year), but is significantly lower than the number of wet days in salalah (25 days/year). sohar’s monthly mean minimum temperature fluctuates between 14–29c, while the monthly mean maximum temperature ranges between 24–36c. muscat, the capital city, accommodated approximately 29.5% of the total population in oman in 2010. the annual total precipitation average from 1991–2010 was approximately 87 mm, showing a statistically weak trend, with precipitation increasing at an average rate of 18 figure 2. variations of annual total precipitation in three study areas. figure 3. monthly precipitation and temperature variations in two study areas (muscat and salalah). 0 50 100 150 200 250 300 350 400 1990 1995 2000 2005 2010 a n n u a l  to ta l  p re ci p it a ti o n  ( m m ) sohar salalah 0 50 100 150 200 250 300 350 1990 1995 2000 2005 2010 a n n u a l  to ta l  p re ci p it a ti o n  ( m m ) muscat 0 5 10 15 20 25 1 2 3 4 5 6 7 8 9 10 11 12 m o n th ly t o ta l p re ci p it at io n ( m m ) 0 5 10 15 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 11 12 m o n th ly m ea n t em p er at u re s (° c ) t-min t-max   0 5 10 15 20 25 1 2 3 4 5 6 7 8 9 10 11 12 m o n th ly t o ta l p re ci p it at io n ( m m )   0 5 10 15 20 25 30 35 40 45 1 2 3 4 5 6 7 8 9 10 11 12 m o n th ly m ea n t em p er at u re s (° c ) t-min t-max a u g ja n f eb m ar a p r m ay ju n ju l s ep o ct n o v d ec a u g ja n f eb m ar a p r m ay ju n ju l s ep o ct n o v d ec muscat salalah a u g ja n f eb m ar a p r m ay ju n ju l s ep o ct n o v d ec a u g ja n f eb m ar a p r m ay ju n ju l s ep o ct n o v d ec l.n. gunawardhana and g.a. al-rawas   140 mm/decade (p = 0.57) (fig. 2). about 73% of the annual total precipitation occurs from november to april when the cold frontal troughs are active (fig. 3). in addition, tropical cyclones originating in the arabian sea bring intense rainfall during the pre-monsoon (mayjune) and post-monsoon (october-november) seasons. even though these cyclones are uncommon and their contribution is not apparent in long-term averages, they occasionally bring heavy rains to the coastal areas, causing substantial damages. the monthly mean maximum temperature in muscat fluctuates between 25–40c, and the monthly mean minimum temperature ranges between 17–29c (fig. 3). the city of salalah is located in the dhofar governorate in southern oman (fig. 1). summer months are cooler than in the northern regions of the country; thus, salalah attracts tourists from the other parts of oman and neighboring countries. from late june to september, the region experiences monsoon precipitation, known locally as the khareef, which brings approximately 61% of the total precipitation for the year (fig. 3). high humidity and drizzle are common through the khareef season. as a result, salalah experiences about 25 wet days per year. the annual total precipitation average from 1991–2010 is 82 mm, which shows a statistically weak negative trend of -18 mm/decade (p = 0.51) (fig. 2). the monthly mean maximum temperature fluctuates between 27–33c, and the monthly mean minimum temperature ranges between 19–28c (fig. 3). 3. methodology oman’s daily precipitation records between 1991–2010 were used. in the current study, rclimdex software developed by the expert team on climate change detection, monitoring and indices (etccdmi) was used to calculate nine indices that analyze different aspects of precipitation under changing climate conditions (alexander et al. 2006). the quality control procedure in rclimdex was applied to identify errors in data processing, and a temporal homogeneity test was conducted using the rhtest software package. climatic time series often exhibit abrupt changes and/or gradual shifts due to station relocation or changes in instruments or observing practices. however, no artificial step changes were detected in the time series of temperature or precipitation. even though all nine indices were calculated, discussion is limited to the indices with hydrologically significant trends only. definitions of the selected indices are presented below. sdii: simple precipitation intensity index, where rrw is the daily precipitation amount on a wet day and w represents the number of wet days (eqn. 1). ∑ (1) r10mm: annual count of days when daily precipitation is ≥10 mm. rx-1day: annual maximum one-day precipitation. r95ptot: annual total precipitation when rrw> 95p. here, 95p is the 95th percentile of precipitation on wet days in the base period (1991–2000). 95 ∑ where 95 (2) prcptot: annual total precipitation on wet days ∑ (3) this analysis was further extended to investigate the spatial variation of extreme rainfall for different return periods. the yearly maxima of the daily precipitation were analysed using the gev distribution function (demaria et al. 2013; stahli et al. 2011). the cumulative probability distribution function for the gev is given as in eqn. 4. ; , , exp 1  / (4) for {1+[(x-)/] 0}. for the location (), scale () and shape (), parameters of this distribution were estimated by the maximum likelihood estimation (mle) method (jenkinson 1955). the goodness-of-fit of the fitted distribution was tested using the kolmogorov-smirnov (k-s) test, which examines the null hypothesis that sample data follow the specified distribution. the characteristics of the precipitation distributions were assessed in terms of the return values and return periods of their annual extremes. given an annual exceedance a comparison of trends in extreme rainfall using 20 years data in three major cities in oman  141 probability of p(xx), the return period is calculated by the reciprocal of p(xx). return period  (5) a t-year return value is defined as the threshold that is exceeded by an annual extreme in any given year with the probability p = 1/t (kharin et al. 2007). 4. results and discussion the trends of selected indices are not consistent in showing whether precipitation will increase or decrease over the three study areas (table 1). of the three precipitation time series examined, muscat and salalah exhibited similar characteristics in trends analysis, while findings for sohar revealed results that are distinct from the other two cities. this can be attributed to the consecutive number of dry years which occurred for seven years, beginning in 1999, in sohar (fig. 2). for example, the total annual precipitation in sohar from 1991–1998 was 180 mm/year but was reduced to 56 mm/year from 1999–2010. therefore, calculations made relative to reasonably higher annual total precipitation rates in the past resulted in notably smaller estimations in recent years, suggesting a negative trend for precipitation indices overall (figs. 4 and 5). even though muscat and salalah depict similar trends for most selected indices, their magnitudes were found to be significantly different. for example, the average annual rx1day rate over 20 years in muscat and sohar was approximately 42 mm/day and 48 mm/day, respectively, while in salalah it was only 22 mm/day (fig. 4). similarly, the average wet day intensity (sdii) estimated in muscat and sohar was approximately 14 mm/day and 12 mm/day, respectively. in salalah, this figure decreased by more than three-fold (3.6 mm/day). this difference has been mainly attributed to a comparatively higher number of rainy days and the drizzly, low-intensity rainfall events common during the khareef season in salalah. the 95th percentile of the precipitation rates on rainy days during the study period were calculated. rainfall events that result in precipitation rates above the 95th percentile are generally accepted as extreme (alexander et al. 2006). this threshold in salalah is approximately 13 mm, while in muscat it is approximately 43 mm. therefore, the 10 mm/day threshold determined when calculating the r10mm index (fig. 4) can be considered a mild event for muscat but an extreme event for salalah. table 1. a summary of calculated indices. index linear regression fitted using the least squares approach sohar muscat salalah rx-1day negative (r2=0.05, p-value = 0.001) positive (r2=0.31, p-value = 0.016) positive (r2=0.39, p-value = 0.090) r10mm negative (r2=0.10, p-value = 0.197) negative (r2=0.14, p-value = 0.123) positive (r2=0.005, p-value = 0.197) sdii negative (r2=0.39, p-value = 0.001) positive (r2=0.39, p-value = 0.090) positive (r2=0.39, p-value = 0.073) negative (r2=0.31, p-value = 0.010) positive (r2=0.005, p-value = 0.197) positive (r2=0.042, p-value = 0.250) l.n. gunawardhana and g.a. al-rawas   142 figure 4. trends in rx-1day and r10mm indices: (a) sohar, (b) muscat, and (c) muscat. all indices examined in muscat indicated increases in precipitation intensity over the past 20 years. the annual rx-1day (fig. 4) and average sdii (fig. 5) suggested positive trends with reasonably high statistical significance (pvalue = 0.016 and 0.005, respectively). moreover, the number of days with mild rainfall intensity as calculated by the r10mm index depicts a statistically weak negative trend. to understand the effect of extreme precipitation on the total annual precipitation, the ratio between r95ptot and the total annual precipitation (prcptot) was calculated. as the r95ptot index gives the total annual precipitation generated by extreme events, the r95ptot/prcptot ratio indicates the contribution from the extreme rainfall to the total annual precipitation. this ratio in muscat (fig. 5) increased between 1991 – 2010, suggesr² = 0.14 p‐value = 0.123  0 2 4 6 8 10 12 1990 1995 2000 2005 2010 r 1 0 m m  ( d a y s/ y e a r) r‐10mm 3‐yrs ma linear (3‐yrs ma) r² = 0.31 p‐value = 0.016 0 40 80 120 160 200 240 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1 day 3‐yrs ma linear (3‐yrs ma) r² = 0.10 p‐value = 0.197 0 2 4 6 8 10 12 1990 1995 2000 2005 2010 r 1 0 m m  ( d a y s/ y e a r) r‐10mm 3‐yrs ma linear (3‐yrs ma) (b) r² = 0.08 p‐value = 0.254 0 50 100 150 200 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1 day 3‐yrs ma linear (3‐yrs ma)   r² = 0.06 p‐value = 0.317 0 2 4 6 8 10 12 1990 1995 2000 2005 2010 r 1 0 m m  ( d a y s/ y e a r) r‐10mm 3‐yrs ma linear (3‐yrs ma) (c)   r² = 0.50 p‐value = 0.001 0 50 100 150 200 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1day 3‐yrs ma linear (3‐yrs ma) (a) a comparison of trends in extreme rainfall using 20 years data in three major cities in oman    143 figure 5. trends in contribution from very wet days and average wet day intensity: (a) sohar, (b) muscat, and (c) muscat. ting enhanced extreme precipitation events over the study period. all indices calculated for salalah indicate evidence for increased precipitation intensity, even though the trend estimations were not found to be statistically significant (figs. 4 and 5). however, despite the decreases in total annual precipitation estimates for salalah (fig. 2), the r95ptot / prcptot ratio suggests that the contribution from extreme rainfall to the total annual precipitation has steadily increased. intensity-duration-frequency (idf) curves are used in designing hydraulic structures such as dams and open channels. the extreme value theory that produces the relationship between the intensity and return periods for selected durations requires the time series to be       (a) (b) contribution from very wet days contribution from very wet days average wet day intensity average wet day intensity   r² = 0.090 p‐value = 0.197 0.0 0.2 0.4 0.6 0.8 1.0 1990 1994 1998 2002 2006 2010 r 95 p t o t / p r c p t o t   r² = 0.39 p‐value = 0.005 0 10 20 30 40 50 60 70 80 90 1990 1994 1998 2002 2006 2010 s d ii  ( m m /d a y ) sdii 3‐yrs ma linear (3‐yrs ma) contribution from very wet days average wet day intensity (c) r 95 p t o t / p r c p t o t r 95 p t o t / p r c p t o t l.n. gunawardhana and g.a. al-rawas   144   stationary. extreme events observed within the study period of the current research (eg. cyclone gonu in 2007), however, lie outside the envelope of constant variability assumed under stationary climate conditions. in such situations, the world meteorological organization (wmo) recommends conducting an analysis, assuming the non-stationarity within the time periods is sufficiently small, or using more advanced methods in which the parameters in the statistical models vary over time to describe the temporal evolution of the extremes (tank et al. 2009). inclusion of a trend in one or more parameters of the gev distribution implies a return level for a particular return period change over time; therefore, no return-level plot can be produced. the researchers in the current study used the second method, introducing a linear trend to one of the parameters in the gev distribution. the most suitable parameter to consider the time variant was decided by the negative log-likelihood value produced by the mle method. table 2 shows the constrained parameters in their study areas. furthermore, fig. 6 shows the estimated rx-1day for the 10year return period with observed rx-1day values plotted over the study period. in all three study areas, no major change in return levels was observed between 1991–2010. inspection of the standard error (table 2) shows that the value for the time variant parameter is much lower than its standard error. it is therefore concluded that the consideration of a time variant trend for more than one parameter could further increase the uncertainty of the results. these results also suggest that the assumption of stationarity is reasonably valid to generate idf curves in the three study areas. in order to understand the differences in frequency and intensity of extreme precipitation in the three study areas, gev parameters were estimated under a stationary assumption (table 3). the standard error in parentheses reveals that the shape parameter is always larger than the standard deviation, which rejects a twoparameter distribution, also called a gumbel distribution, for all the time series. the k-s test results indicated good agreement between observations and modelled probabilities. a reasonably higher shape parameter in salalah may represent an increase in the annual rx1day precipitation compared to the other two study areas. in contrast, the location and smaller scale parameters in salalah compared to those in muscat and sohar can be attributed to the smaller intensity rainfall events. the plot of results from the frequency analysis allows an estimate of the expected return levels (annual rx1-day rainfall) for different return periods (fig. 7). moreover, fig. 8 was produced to compare the annual rx1-day rainfall estimated for 5-, 10-, and 20-year return table 2. estimated parameters of the gev distribution under non-stationary condition.   study area parameters of the gev distribution location () scale () shape () sohar 22.28+[-0.138(year1990)] {13.24, 0.922} 20.48 {5.24} 0.5099 {0.2580} muscat 21.02+[0.124(year1990)] {8.66, 0.774} 16.32 {3.43} 0.3431 {0.1575} salalah 7.44+[-0.066(year-1990)] {1.57, 0.040} 5.16 {2.17} 1.298 {0.4534} standard errors are shown in parentheses {} table 3. estimated parameters of the gev distribution under stationary climate conditions. study area parameters of the gev distribution k-s test significant at 5% level reject gumbel hypothesis location () scale () shape () sohar 20.45 {5.22} 20.31 (5.11) 0.5290 {0.2341} yes yes muscat 22.23 {3.99} 16.30 (3.43) 0.3449 {0.1572} yes yes salalah 7.43 {1.61} 6.02 (2.14) 1.0397 {0.3417} yes yes standard errors are shown in parentheses. a comparison of trends in extreme rainfall using 20 years data in three major cities in oman    145 (a) (b) (c) figure 6. the 10-years return level estimated under non-stationary condition: (a) sohar, (b) muscat, and (c) salalah. 0 20 40 60 80 100 120 140 160 180 200 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1day rx‐1day for 10 yrs rp 0 50 100 150 200 250 300 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1day rx‐1day for 10 yrs rp 0 20 40 60 80 100 120 1990 1995 2000 2005 2010 r x‐ 1 d a y  ( m m /d a y ) rx‐1day rx‐1day for 10 yrs rp l.n. gunawardhana and g.a. al-rawas   146   (a) (b) (c) figure 7. return periods of annual maximum 1-day precipitation estimated by gev distribution: (a) sohar, (b) muscat, and (c) salalah. y = 0.0005x2 + 0.0258x + 0.8682 r² = 0.9999 0 5 10 15 20 25 0 50 100 150 200 r e tu rn  p e ri o d  ( y e a rs ) annual maximum 1‐day rainfall (mm/day) y = 1.1098e0.0212x r² = 0.9543 0 50 100 150 200 250 300 0 50 100 150 200 250 300 r e tu rn  p e ri o d  ( y e a rs ) annual maximum 1‐day rainfall (mm/day) y = ‐4e‐05x2 + 0.1568x + 0.4488 r² = 0.9999 0 2 4 6 8 10 12 14 16 18 20 0 20 40 60 80 100 120 r e tu rn  p e ri o d  ( y e a rs ) annual maximum 1‐day rainfall (mm/day) hurricane gonu  a comparison of trends in extreme rainfall using 20 years data in three major cities in oman  147 figure 8. annual maximum 1-day precipitation estimated for different return periods in three study areas. periods in the three study areas. results show that the annual rx1-day rainfall estimated for the 5and 10-year return periods in sohar and muscat are approximately the same, but are about two-fold greater than in salalah. however, the rainfall for the 20-year return period in all three cities was estimated to be higher and reasonably close. this is because all of the study areas are affected by tropical cyclones, which bring occasional but heavy rainfall. 5. conclusion in this study, precipitation extremes in three major cities in sultanate of oman were examined using a set of climate indices. moreover, a frequency analysis was conducted to understand the differences in annual maximum 1-day precipitation estimated for different return periods in the three study areas. in the end, the following conclusions were obtained. the annual total precipitation in sohar and salalah decreased, especially in recent years, while that in muscat slightly showed an increasing trend. this difference in muscat is mainly because of the heavy precipitation caused by hurricane gonu in 2007. the average wet day precipitation intensity in salalah is approximately 3.6 mm/day, which is significantly smaller than that in sohar and muscat (12 and 14 mm/day, respectively). this is mainly attributed to the frequent but low intensity rainfall events common in the khareef season in salalah. despite the decrease in annual total precipitation estimated in salalah, all indices analysed indicate enhanced extreme events toward 2010. consequently, the contribution from extreme events to the annual total rainfall steadily increases in salalah. the results in muscat also lead to similar conclusion, which indicates that two-decades long wetting signal apparent in total precipitation can be attributed largely to the increases in extreme precipitation in recent decades. however, a comprehensible conclusion could not be made based on the selected indices for sohar due to consistent drier years occurring from 1999 to 2005. an analysis using long-term data with focus on potential drought is therefore recommended for sohar. consideration of time variant parameter in gev distribution produced no significant changes in return levels in all three study areas. frequency analysis conducted under the assumption of stationarity indicates that the annual maximum 1-day rainfall estimated for 5 and 10 year return periods in sohar and muscat are approximately the same but about two-fold greater than that in salalah. however, the rainfalls for 20 year return period in all three cities were estimated to be higher and reasonably close. these results are expected to be useful for planning countermeasures for flash floods and persistent droughts in these study areas. 0 20 40 60 80 100 120 140 160 180 200 0 5 10 15 20 25 a n n u a l m a xi m u m  1 ‐d a y  r a in fa ll  ( m m /d a y ) return period (years) sohar muscat salalah linear (sohar) linear (muscat) l.n. gunawardhana and g.a. al-rawas   148   acknowledgment this research was supported by the “oman national climate change strategy” project funded by the ministry of environment and climatic affairs (meca) in sultanate of oman. references alexander lv, zhang x, peterson tc, caesar j, gleason b, tank amgk, haylock m, collins d, trewin b, rahimzadeh f, tagipour a, kumar kr, ravadekar j, griffiths g, vincent l, stephenson db, burn j, aguilar e, brunet m, taylor m, new m, zhai p, rusticucci m, aguirre jlv (2006), global observed changes in daily climate extremes of temperature and precipitation. journal of geophysical research d05109, doi:10.1029/2005jd006290. alsarmi sh, washington r (2013), changes in climate extremes in the arabian peninsula: analysis of daily data. international journal of climatology 34: 1329-1345. beniston m, stephenson db, christensen ob, ferro cat, frei c, goyette s, halsnaes k, holt t, jylha k, koffi b, palutikof j, scholl r, semmler t, woth k (2007), future extreme events in european climate: an exploration of regional climate model projections. climatic change 81: 71-95. coopersmith ej, minsker bs, sivapalan m (2014), patterns of regional hydroclimatic shifts: an analysis of changing hydrologic regimes. water resources research 50: 19601983. demaria emc, maurer ep, thrasher b, vicuna s, meza fj (2013), climate change impacts on an alpine watershed in chile: do new model projections change the story? journal of hydrology 502: 128-138. fowler hj, ekstrom m, kilsby cg, jones pd (2005), new estimates of future changes in extreme rainfall across the uk using regional climate model integrations, 1. assessment of control climate. journal of hydrology 300: 212233. groisman py, knight rw, easterling dr, karl tr, hegerl tc, razuvaev vn (2005), trends in intense precipitation in the climate record. journal of climate 18: 1326-1350. gunasekara nk, kazama s, yamazaki d, oki t (2013), the effects of country-level population policy for enhancing adaptation to climate change. hydrology and earth system sciences 17: 4429-4440. haren rv, oldenborgh gjv, lenderink g, collins m, hazeleger w (2013), sst and circulation trend biases cause an underestimation of european precipitation trends. climate dynamics 40: 1-20. jenkinson af (1955), the frequency distribution of the annual maximum (or minimum) values of meteorological elements. quarterly journal of the royal meteorological society 81: 158-171. kharin vv, zwiers fw, zhang x, hegerl gc (2007), changes in temperature and precipitation extremes in the ipcc ensemble of global coupled model simulations. journal of climate 20: 1419-1444. liu b, chen c, lian y, chen j, chen x (2015), long-term change of wet and dry climatic conditions in the southwest karst area of china. global and planetary change 127: 1-11. stahli m, badoux a, ludwig a, steiner k, zappa m, hegg c (2011), one century of hydrological monitoring in two small catchments with different forest coverage. environmental monitoring and assessment 174: 91-106. tank amgk, zwiers fw, zhang x (2009), guidelines on analysis of extremes in a changing climate in support of informed decisions for adaptation (wcdmp-72, wmotd/no.1500), 1-55. tezuka s, takiguchi h, kazama s, sato a, kawagoe s, sarukkalige r (2014), estimation of the effects of climate change on floodtriggered economic losses in japan. international journal of disaster risk reduction 9: 58-67. the journal of engineering research (tjer), vol. 15, no. 2 (2018) 129-141 doi: 10.24200/tjer.vol15iss2p129-141 variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings n.a. al-azri*,a, and s. al-saadib a department of mechanical and industrial engineering, sultan qaboos university, po box 33, al-khoud 123, oman. b department of civil and architectural engineering, sultan qaboos university po box 33, al-khoud 123, oman. received 20 june 2016; accepted 28 january 2018 abstract: typical meteorological years (tmys) are widely used for the analysis and simulation of energyintensive systems. the reliability of a developed typical year depends on the accuracy of the historical record of weather data as well as the fitness of the developed approach to the application. in this work, a tmy for seeb area in the muscat governorate, oman was developed using different approaches. the developed tmys are compared to the current commonly used tmy which is based on 1985-2001 records that have many gaps and anomalies and hence have intensive interpolation treatment. the different tmys were compared by simulating energy consumption of a typical residential building and also by studying applicability of passive cooling strategies. the findings showed that the variation in energy consumption is minimal for the different tmy development approaches for the same set of historical records but the difference is very significant when the comparison is based on the two sets from the two periods of records. keywords: typical meteorological year (tmy); weather data; passive cooling. سلطنة عمان وتأثريها على السيب،يف التطورات املتغرية لسنوات األرصاد اجلوية النمطية حماكاة الطاقة للمباني السكنية بي ، صاحل السعد أ *العزري أمحدناصر تعتمد لطاقة الكثيفة.استهالك انظمة أعلى نطاق واسع لتحليل وحماكاة النمطية: تستخدم سنوات األرصاد اجلوية امللخص النموذجية السنة بناءاملتقدمة على دقة السجل التارخيي لبيانات الطقس باإلضافة إىل مدى مالءمة أسلوب طيةموثوقية السنة النم عمان باستخدام مسقط،ملنطقة السيب يف حمافظة النمطية االرصاد سنة مت تطوير العمل،يف هذا .املستهدف لالستخدام املستخدمة بشكل شائع النمطية االرصاد سنة املطورة مع سنوات األرصاد اجلوية النمطية مقارنة تمت حيث .أساليب خمتلفة .مكثفة ملعاجلة حتتاج بالتالي اليتالثغرات وو الشذوذ العديد من بهااليت و م2001إىل 1985من االعوامواليت تستند إىل سجالت بنى سك ي وموذجي وأياا من لال ملختلفة من لال حماكاة استهالك الطاقة امل نمطيةالوية اجلرصاد األسنوات ومتت مقارنة ساليبضئيل بالنسبة ال أظهرت النتائج أن التباين يف استهالك الطاقةو . دراسة إمكانية تطبيق اسرتاتيجيات التربيد السليب عندما تكون املقارنة مبنية كبريالفرق هذاسجالت التارخيية ولكن اللنفس جمموعة التطوير املختلفة لـسنة االرصاد النمطية .على جمموعتني من فرتتي التسجيل .، بيانات الطقس، التربيد السليب طيةاألرصاد اجلوية النم ةسن :الكلمات املفتاحية * corresponding author’s e-mail: nalazri@squ.edu.om mailto:nalazri@squ.edu.om n.a. al-azri and s. al-saadi 130 1. introduction typical meteorological years (tmys) are used by a wide spectrum of professionals in designing and studying climate-dependent cases. a tmy is a set of year-long data, statistically selected from a historical record of metrological data and it is regarded as the best representation of the climatic behavior in a given location throughout the year. tmys were first developed in the united states for 229 locations using solar radiation and weather records from 1948 to 1980. that collection was updated twice using data for the periods 1961-1990 and 1991-2005 for 1020 locations in the second update. typical meteorological years are essential in simulating long-term weather-dependent projects such as estimating cooling and heating loads and building renewable energy harvesting facilities. basing simulation on weather data from a random actual year will make the results susceptible to underestimation and overestimation due to the likely variation in weather behavior. hence, tmys are only good when an average weather behavior is under consideration and they should not be used for extreme weather conditions. the reliability of tmys depends on the accuracy of the historical record and the fitness of the development methodology for the targeted application. the typical meteorological year is composed of the typical months in the historical records eg. typical january is january of one of the years in the full record that is statistically the best representative of the common trend in the weather during the month of january. the typical month is basically the month whose weather parameters are the closest to the mean of all months in the historical record. the more relevant parameters are assigned the higher weight in the statistical methodology and hence have higher impact on whether to select a given month. the tmy currently used for seeb area in muscat governorate, oman is based on a historical record from 1985 to 2001. like most of the records around that period in oman, it suffers from many gaps and anomalies. this work aims at updating the tmy used in seeb, governorate of muscat, oman and assessing the impact of using tmys that are based on different historical records and developed using different approaches. in this work, the developed typical metrological year is based on a more accurate and recent record from 2000 to 2014. the new tmy is compared to the old one based on the energy consumption and capital cost of a typical residential building in order to study the variation when two historical records are used. moreover, different development methods are compared based on the 2000-2014 record. passive cooling strategies are based on bioclimatic charts which are based on dry bulb temperature and dew point, usually taken from a typical metrological year. in order to study the impact of the application on the reliability of the developed tmy, a tmy is developed with all the weight assigned to the two relevant parameters and zero weight assigned to solar radiation. bioclimatic charts and the applicability of passive cooling strategies are then based and compared on the two sets of tmys. this research sheds light on the reliability of the currently used tmys in oman and the need for updating them. it also tackles the impact of using different approaches in developing tmys on the simulation results of energy systems. it also addresses the impact of using tmys based on inaccurate records on bioclimatic charts and it stresses the urgency for examining the quality of the data set when developing tmys. 2. tmy development the development of the tmy is based on statistical methodology in which data from historical archives are selected to represent a typical year. the tmy was first developed by sandia laboratories, usa and it was intended for solar heating and cooling purposes habte et al. (2014). the sandia procedure is thoroughly explained in the work of sawaqed et al. (2005). the method in its original version is based on the daily averages of the nine parameters listed in table 1 except the direct radiation. each of these parameters is statistically analyzed over a recommended period of 12-24 years. then finkelstein-schafer statistic finkelstein and schafer (1971) is used to indicate the deviation of the short term daily mean of the nine parameters (using cumulative distribution function) from their daily averages over the value for all years in the record, which is known as long term cumulative distribution function. then, a weighted sum of the average daily deviations (finkelstein-schafer statistic) is calculated. for each month in the year, five best candidate years are selected according to the variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 131 least weighted deviations. in order to reassure that this selection was not a result of some anomalies, further screening is performed by checking the deviation of the mean and median for each month from the long term means and medians and also checking the run count and length in each month as described by sawaqed et al. (2005). most tmy development methods are based on the sandia method but mainly differ in the weights assigned to each of the contributing parameters in the calculation of finkelsteinschafer statistic as shown in table 1. these weights are usually selected based on the purpose for which tmy is developed. the weights shown in table 1 are intended for general simulation of thermal applications and are not valid for case analysis at extreme weather conditions. the weights suggested by sandia were adopted in the generation of the first version of tmys in the united states and also in many other works, eg. petrakis et al. (1996). other applications of sandia used different weights. for instance, pissimanis et al. (1988) gave no weight to all parameter ranges in addition to the minimum wind speed. sawadeq et al. (2005) adopted the same weights as in sandia but gave no weight to the minimum wind speed. however, sawadeq et al. (2005) replaced the dew point with relative humidity. subsequent generations of tmys in the united states; tmy2 (marion et al. 1995) and tmy3 (wilcox et al. 2008) also excluded the ranges and the minimum wind speed but added direct radiation to the list of parameters. in addition to the purpose of use, the selection of these weights is also intended for increasing the agreement with the historical record. for example, the adoption of direct radiation in tmy3 improved the agreement between annual direct normal radiation for the typical year and the record of the 30-year annual average by a factor of two (wilcox and marion, 2008). moreover, the wind velocity is given less weight while increasing the weights of dry bulb and dew point temperatures, which are more important in solar energy conversion systems and buildings. hence, this modification has rendered tmy3 inappropriate for wind energy applications (wilcox and marion, 2008). in addition to the sandia method, which is the classical development algorithm, heuristic methodologies are also used, in which, machine learning algorithms are used in developing weather files (chakraborty et al. 2016). table 1. the different weights used in different works in the development of the tmy. parameter sandiaa pissimani et al.b tmy2/3c sawaqed et al.d,e [14] max dry bulb temp 1/24 1/24 1/20 1/22 min dry bulb temp 1/24 1/24 1/20 1/22 mean dry bulb temp 1/24 2/24 2/20 1/22 dry bulb temp range 1/24 0 0 1/22 max dew point temp 1/24 1/24 1/20 1/22 min dew point temp 1/24 1/24 1/20 1/22 mean dew point temp 1/24 2/24 2/20 1/22 dew point temp range 1/24 0 0 1/22 max wind velocity 1/24 2/24 1/20 1/22 min wind velocity 1/24 0 0 0 mean wind velocity 1/24 2/24 1/20 1/22 wind velocity range 1/24 0 0 1/22 global radiation 12/24 12/24 5/20 11/22 direct radiation 0 0 5/20 0 ahall et al. (1978) bpissimanis et al. (1988) cmarion and urban (1995) and wilcox and marion (2008) dsawaqed et al. ( 2005) ethe relative humidity is used in place of dew point temperature. n.a. al-azri and s. al-saadi 132 3. building simulation and weather files building simulation software programs are primarily used to perform energy and thermal analyses of buildings. energy is analyzed based on hourly or sub-hourly basis in order to capture the dynamic behavior of building systems. in the energy analysis procedure, a typical weather file should be derived from raw data of many years. many typical weather data sets are internationally recognized for use in detailed energy simulation programs. in north america, well-known weather data sets are used which include: typical reference year (try, the american version) (ncdc, 1976), typical metrological year (tmy, tmy2 and tmy3) wilcox and marion (2008), weather year for energy calculation (wyec), canadian weather for energy calculations (cwec) watsun (1992), and weather year for energy calculation version: 2 (wyec2) (huang j, 1998; stoffel and rymes, 1998). for the u.s.a. tmy files are commonly developed for hundreds of locations and some states like california has its own weather data set (california thermal zone version2: ctz2) which is used for the state code compliance. in canada, cwec weather data set is commonly used but wyec2 format is also used for some locations. in europe, the standard method of generating test reference year (isotry version) is proposed under iso standard pren iso 15927-4 (levermore and doylend, 2002). another form of try is proposed by chartered institution of building services engineers (cibse) in the u.k. (holmes and hitchin, 1978). the iso-try is proposed for near-extreme plant design or near-extreme performance assessment whereas the cibsetry format is for average energy estimation and analysis. therefore, iso-try is more appropriate for hvac design and short-term performance analysis. many endeavors were dedicated to establishing suitable weather data files for energy simulation programs (clarke 2001). the american society of heating, refrigerating, and air-conditioning engineers (ashrae) developed typical weather years, which are known as international weather for energy calculation (iwec), for nearly three thousand sites outside us and canada (ashrae 2015). due to the lack of solar radiation data, solar radiation algorithms were used to generate the iwec. the weather data years are selected based on the standard methodology of tmy2 (marion and urban, 1995), which is also based on sandia method. the types of different weather data sets and the selection methodologies of a typical weather year have brought about many issues related to the reliability of energy simulation. this problem has been addressed by crawley et al. (1999). in their study, a typical office building was simulated in eight u.s. locations using different reference years (try, tmy, tmy2, wyec, wyec2) based on a 30-year-period of actual hourly weather data. crawley studied the impact of using various weather data sets on the annual energy use and costs and the annual peaks of electrical demand, heating load, and cooling load. the study concluded that the american version of try-type weather data should be avoided since no single year can represent the typical long-term weather patterns. tmy2 and wyec2 were found more rigorous in predicting the energy consumption since improved solar models are used in simulating solar radiation. therefore, they are regarded as the best representative of long-term average climatic conditions. huang (1998) conducted another study in which he evaluated the influence of different sets of weather data, also based on a 30-year-period of historical records, on residential buildings. the study concluded that tmy2 and wyec2 weather data gives results within 5%. the study justified the use of the typical weather year to simulate the peak building loads because the selection process doesn’t eliminate the peak design conditions. it was also concluded that the american version of try is less reliable in replicating the average historical conditions. erba et al. (2017) studied the effect of weather datasets on building energy simulation outputs for a public social housing block located in the area of milan, italy. in their study, they found a difference in the output results when they used two commonly used weather datasets. in their conclusion, they asserted the need for updating weather files for more accurate simulation. 4. using tmy for passive cooling passive cooling techniques are used in minimizing the load on mechanically-driven cooling devices. in general, the selection of building passive thermal design strategies is based on dry bulb temperature and humidity. identifying suitable strategies for a given variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 133 location can be made using bioclimatic charts. olgyay’s bioclimatic chart (olgay 1963) developed in the 1950s, was one of the first attempts for identifying the best passive cooling methods. nowadays, there are a number of approaches for developing bioclimatic charts, and these mostly have outlined zones for each passive method overlaid on a psychrometric chart. figure 3 shows giovoni’s bioclimatic chart in which there are five passive cooling zones: comfort zone, natural ventilation, high mass, high mass with night ventilation and evaporative cooling. extensive description on passive cooling techniques and the development of bioclimatic charts is available in the literature (al-azri et al. 2013). another more popular bioclimatic chart is that of givoni (1992; 1994). this chart is based on the linear relationship between the temperature amplitude and vapor pressure of the outdoor air. givoni’s chart identifies the suitable cooling technique based on the outdoor climatic condition. five zones are identified on givoni’s chart: thermal comfort, natural ventilation, high mass, high mass with night ventilation and evaporative cooling. givoni’s bioclimatic charts are utilized by first identifying the average monthly conditions. for each month, the average of the daily maximum temperature is calculated and matched with the average of the minimum daily absolute humidity to form the point (�̅�𝑚𝑎𝑥, �̅�𝑚𝑖𝑛). likewise, the average of the daily minimum temperature is matched with that of the average daily minimum absolute humidity to form (�̅�𝑚𝑖𝑛, �̅�𝑚𝑎𝑥). the placement of the line segment connecting the two points will determine the proper passive cooling strategy for that month. givoni’s chart is mainly applicable to residential and office buildings where heat gain is minimal. modifications to givoni’s chart that suit nondomestic buildings can also be found in the literature (lomas et al. 2004). the natural ventilation zone on givoni’s chart assumes that the indoor mean radiant temperature and the vapor pressure are the same as those at outdoor conditions; an assumption that limits the application to buildings with medium to high thermal structure (watson 1981). a more accurate representation of the typical metrological year for the development of the bioclimatic charts can be developed by eliminating irrelevant parameters, such as wind speed and solar radiation, in sandia procedure and dedicating all the weight to dry bulb temperature and dew point (al-azri 2016). 5. the study weather and solar radiation data for the seeb area were collected for the period 2000-2014 which feature better accuracy and fewer gaps and anomalies compared to the data used in previous studies. the time reference point for the data was taken at the beginning of the hour and hence data taken at fractions of the hour were interpolated with the adjacent readings using linear interpolation provided that the end points of the interpolation period does not exceed 60 minutes. the data was then qualitychecked by investigating missing points and anomalies. unrealistic anomalies were removed and regarded as missing points. missing points were found to be around 2% in the whole set and their consecutive occurrences extend for different periods mostly for a few hours. periods of missing points up to two hours where lineally interpolated with the adjacent readings. consecutive missing points for longer periods where linearly interpolated using readings from the adjacent days at the same period of time in the day. in the development of the tmy, the same procedure described by sawaqed et al. (2005) was followed but using the different weights shown in table 1. in this paper, different approaches for typical metrological years are studied for seeb area in muscat governorate, oman. the weather data were taken for muscat airport station, which is located at 23.583o latitude and 58.283o longitude and it is 8.4 meter above sea level. two sets of records for this station were used; the old 1985-2001 record and the new 2000-2014 record. the two sets have distinct features. the old record, upon which the current tmy was developed, has many anomalies and its solar radiation data were calculated using the daily mean values. 6. simulation of a typical residential unit simulation programs are commonly used for the design of energy-efficient buildings. these tools are primarily used for energy analysis but can also be helpful when sizing air conditioning systems. for the purpose of this study, quick energy simulation tool (equest) is utilized for yearly simulation (lbnl and jjh 1998). a typical residential building shown in fig. 1 is n.a. al-azri and s. al-saadi 134 assumed to be the representative house in oman. the physical, thermal characteristics and construction details are defined in table 2. figure 1: 3d view of the typical housing in equest. table 2. characteristics of building systems for a typical residential housing in oman. characteristics description of the base case location seeb, oman orientation front elevation facing north plan shape rectangular number of floor two thermal zones: living area and sleeping area floor to floor height 3.35 m (11 ft) floor area 240 m2 (2582 ft2) floor dimension 12.5 x 19.2 m (41.5 x 62.3 ft) window area 15% of the gross wall area, uniformly distributed window type of glass: 4 mm single green tinted glazing frame: aluminum with no thermal break solar absorbance 0.50 for external walls (medium color) 0.50 for the roof ( medium color ) exterior walls 15mm cement plaster + 200 mm cmu hollow block + 15mm cement plaster roof tiles + 10 mm mortar + 150 mm reinforced concrete slab + 15 mm cement plaster floors ground floor: 300 mm soil+ 200 mm slab on grade, intermediate floor: 150 mm reinforced concrete slab shading interior shades when zones are occupied occupancy density 6 people lighting power density 5 w/m2 (for ground and 1stfloor) equipment power density 8.5 w/m2(for ground and 1st floor) infiltration 0.75 ach system type split air-conditioning units (constant-volume dx ac) thermostat setting 24ºc for cooling (no heating is provided) cop 2.8 variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 135 figure 2 shows the monthly electrical consumption and energy cost for the house when a weather file from meteonorm software is used (meteonorm 7, 2015). for comparison, a recent survey study for 50 houses in seeb, oman indicated that actual annual energy consumption ranges between 16-82 mwh of electricity in a year, with an average of 44 mwh/year (al-hashim 2013). the equest software predicts an electrical consumption of 48.2 mwh/year for the base case model considered in this study. in addition to the weather file generated by meteonorm software (meteonorm 7, 2015), the building simulation model is used with different weather files developed in this research based on approaches described in the literature (hall et al. 1978; pissimanis et al. 1988; sawaqed et al. 2005; wilcox and marion, 2008). using the tmy approach described by sawaqed et. al. (2005), two weather files using two different raw weather datasets were generated; one tmy file based on weather data from the period 1985–2001, and another tmy file is based on weather data for the period 2000-2014. the other tmy files described by hall et al. (1978), pissimanis et al. (1988), sawaqed et al. (2005) and, wilcox and marion (2008) were all generated using weather dataset of 2000-2014. the development of the weather files are based on the weights given in table 1. table 3 shows the typical years selected by the different approaches. despite the variations in the assigned weights, the different approaches merged to almost an identical typical years except for a few months. in general, wilcox’ approach exhibits the most variation as it differs from the others in five months which can be attributed to the fact that this approach carry an additional parameter, namely, direct radiation, assigned 25% of the weight which has influenced the selection process. this variation, however, might not have any impact on the simulation as the hourly variation in the climatic readings of the different months might behave in a similar manner whose impact on energy analysis is minimal. this fact will be more elaborate in the results of the simulation. figure 2. monthly electrical consumption/cost for the house using weather file from meteonorm software. table 3. the typical years selected for each month using different approaches. tmy approach jan feb mar apr may jun jul aug sep oct nov dec sandia 2009 2012 2007 2014 2013 2012 2010 2009 2005 2002 2012 2007 pissimanis 2009 2002 2007 2014 2013 2012 2010 2009 2005 2002 2012 2007 tmy2/3 2005 2002 2011 2007 2013 2012 2010 2009 2001 2006 2012 2007 sawadeq, 2005a 1995 1994 1993 1988 1994 1987 1996 1989 1994 1989 1986 1987 sawadeq, 2005b 2009 2012 2007 2014 2013 2012 2010 2009 2005 2005 2012 2007 a based on 1985-2001 dataset and b based on 2000-2014 dataset. n.a. al-azri and s. al-saadi 136 meteonorm weather files are commonly utilized in building simulation tools. hence, it is used here as the base case for comparison. two performance indicators are extracted from the equest simulation files: the annual space cooling energy (kwh/year) and the peak space cooling load (tons). the associated costs for these two indicators are provided for comparison using the local electrical cost of 0.026 $/kwh and capital equipment cost of $ 260/ton. table 4 shows the results when using all six tmy weather files. the absolute and relative differences for the associate costs are calculated. the negative sign of the relative difference indicates that the results of the simulated cases are lower than the base case (ie. meteonorm tmy weather file). regardless of the tmy approach used, the results based on the 2000-2014 dataset show low deviation from the reference case (lower by 1.92.4%). for the same dataset and its derived tmy weather files, the result for the peak cooling load is identical and lower by ~16% when compared to the base case. in all tmy files, the peak month, which is used for peak calculation is the same, june 2012, which explains the similarity amongst the results. the results predicted by the tmy file developed from 1985-2001 dataset using the sawaqed’s (2005) approach is 14% and 11% lower than the base case for annual cooling energy and peak cooling load, respectively. this dataset was based on an older record and its solar radiation data was modeled using monthly averages. in terms of accuracy, the 2000-2014 dataset is better and hence the results using tmy based on this set should be more credible. 7. passive cooling in seeb in order to study the impact of variation in typical meteorological years, bioclimatic charts are developed for seeb using the same dataset but different approaches and using the same approach with different datasets. the two results are compared to the base tmy of meteonorm. the first dataset was taken from 1985-2001 in which the approach of sawadeq et al. (2005) was used. the other dataset was taken from 2000-2014 and features minimal missing data and hourly solar radiation measurements, unlike the older set in which radiation was modeled from monthly averages. as shown in table 3, the behavior of the different approaches and weights is almost identical when the same dataset is used in the development of the tmy. hence, sawadeq et al. (2005) approach is used for both sets; 1985-2001 and 2000-2014. figures 3 and 4 shows the bioclimatic charts using both datasets. table 4. impact of weather files on annual cooling energy consumption, peak cooling load and the associated costs. reference dataset annual cooling ( kwh) annual cooling ( $) difference in annual cooling (%$) peak load (tons) ac capital cost, $ difference in capital (%$) (meteonorm 7, 2015)a 2000-2009 31,587 821.3 20.1 5,226 (pissimanis et al., 1988) 2000-2014 30,823 801.4 19.9 (-2.4%) 16.9 4,394 832 (-15.9%) (wilcox & marion, 2008) 2000-2014 31,092 808.4 12.9 (-1.6%) 16.9 4,394 832 (-15.9%) (sawaqed et al., 2005) 1985–2001b 27,032 702.8 118.5 (-14.4%) 17.8 4,628 598 (-11.4%) (sawaqed et al., 2005) 2000-2014 30,833 801.7 19.6 (-2.4%) 16.9 4,394 832 (-15.9%) (petrakis et al., 1996) 2000-2014 30,823 801.4 19.9 (-2.4%) 16.9 4,394 832 (-15.9%) areference case, bsolar radiation is not measured in this station and based on 1985-2001 dataset. variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 137 the solid lines in both graphs indicate the results projected by the tmy based on meteonorm while the dashed lines are those taken from the two records of datasets. when compared with each other, the distribution pattern is the same for the meteonorm tmy and those developed from the two datasets. however, the representation of the older dataset is more similar to that of meteonorm where all counterpart months are close enough and intersect one another. the observation is so different with the newer dataset (2000-2014) that in some cases a technique is suggested by one tmy but not by the other. for instance, the two lines of june lie far apart; while in meteorm set it suggests mechanical air conditioning, the 2000-2014 set suggests that high mass will work at high temperature values. the comparison is quantified in table 5 which shows the percentage of readings falling on each zone using metenorm tmy (set a), tmy developed from 1985-2001 dataset (set b) and tmy developed from 2000-2014 dataset (set c). the three sets, in many cases, exhibit wide differences. however, sets a and b are closer to each other than a and c. figure 3. givoni bioclimatic chart using tmy developed by sawaqed et al. (2005) using 1985-2001 dataset. figure 4. givoni bioclimatic chart using tmy developed by sawaqed et al. (2005) using 2000-2014 dataset. n.a. al-azri and s. al-saadi 138 unlike the simulation of active cooling system, the deviation from the base set (a) here is exhibited more in the 2000-2014, set (c), than it is in the 1985-2001 set (b). this is due to the fact that in the bioclimatic charts only humidity and temperature that are considered. the development of the bioclimatic chart is dependent based on the average minima and maxima of these parameters; while in the active cooling simulation, radiation was also part of the simulation. figure 5 and fig. 6 show the monthly mean dry bulb temperature in seeb for all years in the 2000-2014 record and the year chosen to be the typical meteorological year (bold curve). in figure 6, all parameters are considered in the development of the tmy, with the mean dry bulb temperature carries 10% of the weight based on tmy2/3 approach. figure 7 considers dry bulb temperature and dew point only, which are the relevant ones to the development of the bioclimatic charts with mean dry bulb temperature caries 20% of the weight while the 80% are assigned to mean dew point (20%), minimum, maximum and range of dry bulb and dew point (10% each). table 5. the percentage of hours falling on each zone using meteonorm (a), 1985-2001 (b) and 20002014 (c) tmy. comfort zone natural ventilation high mass high mass /night ventilation evaporative cooling a b c a b c a b c a b c a b c 1 55 64 31 58 66 33 56 65 33 56 65 33 56 65 33 2 55 55 37 69 74 48 66 61 48 66 61 48 65 60 49 3 53 43 48 90 90 80 84 63 80 85 64 82 77 59 82 4 28 7 21 75 63 76 78 61 76 85 73 98 68 51 89 5 2 0 9 34 26 65 43 23 65 81 49 94 29 15 69 6 0 0 8 28 11 53 11 9 53 28 32 93 1 8 55 7 0 0 0 31 6 17 0 4 17 1 27 41 0 4 4 8 0 0 0 48 58 31 1 1 31 1 2 54 0 0 19 9 0 3 11 66 45 63 2 15 63 2 19 98 0 10 53 10 10 9 49 77 71 83 53 35 83 63 44 96 22 28 86 11 41 36 45 95 98 80 74 67 80 74 67 80 59 54 77 12 55 67 39 76 71 48 65 70 48 65 70 48 62 69 48 a: meteonorm b:1985-2001 c:2000-2014 figure 5. the typical year based on all parameters using tmy2/3 approach. variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 139 it is clear that the selection of the typical years for each month has changed. figure 6 shows more conformity between the typical year and the crowd of the means of all months in the record since the mean temperature is assigned a higher weight. however, deviation is still present in both figures because of the existence of other parameters. this comparison shows the importance of only considering the relevant parameters in the intended purpose. 8. conclusion typical meteorological years were developed for seeb area in muscat governorate, oman using different sources of weather data and different development approaches. each developed tmy is used for analyzing energy consumption in a building and the variation in the accuracies of the different sources has shown greater impact than the variation resulting from using different approaches for the same data source. hence, the accuracy of the source data is the main playing factor. moreover, although using different tmy development methods has minimal impact on annual energy consumption, the impact can be significant on the capital cost which is due to the fact that equipment sizing is greatly dependent on the peak points of the external heat gain. the tmy is composed of typical months in the record and the selection of these months are based on calculations of averages where peak points are hardly detectable. until very recently, weather data was very scarce in oman and most works were based on records from 1985 to 2001 which had many gaps and anomalies. when it is based on this record, the analysis of the presented case study has shown significant difference in capital cost and yearly energy consumption compared to the analysis based on the recent record which is more accurate and well-kept. hence, with the development of more accurate measurement and logging technologies, it is recommended that tmys be redeveloped from time to time to reassure their reliability. bioclimatic charts were developed using the different tmys for studying the applicability of the different passive cooling strategies. since bioclimatic charts only involve dry bulb temperature and dew point, involving solar radiation in developing their typical year should involve disturbance to their accuracy. this fact is proven when comparing the applicability of passive cooling strategies based on tmys developed with and without solar radiation. the involvement of solar radiation poses significant variation when the tmys are based on the same set of historical record. it is hence recommended that for bioclimatic charts and passive cooling, tmys should be developed without solar radiation. figure 6. the typical year based on dry bulb and dew point temperatures only. n.a. al-azri and s. al-saadi 140 conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references al-azri n (2016), development of a typical meteorological year based on temperature and humidity for passive cooling applications. energy for sustainable development 33: 61-74. al-azri n, zurigat yh, al-rawahi n (2013), development of bioclimatic chart for passive building design. international journal of sustainable energy 32(6): 713-723. al-hashim a (2013), masters report: guidelines for improving the thermal performance of typical residential houses in oman: the case of integrating shading into building codes in muscat city. (master’s report), the university of arizona. ashrae. (2015). url: https://www.ashrae.org/resources--publications/bookstore/iwec2[accessed: may/ 23/2015]. chakraborty d, elzarka h, bhatnagar r (2016), generation of accurate weather files using a hybrid machine learning methodology for design and analysis of sustainable and resilient buildings. sustainable cities and society 24: 33-41. clarke ja (2001), energy simulation in building design (2nd ed.). london butterworthheinemann. crawley d, winkelmam f, lawrie l, pedersen c (1999). energyplus: a new-generation building energy simulation program. building energy simulation user news 20(1). erba s, causone f, armani r (2017), the effect of weather datasets on building energy simulation output. 9th international conference on sustainability in energy and buildings, seb -17, 5-7 july 2017, chania, crete, greece. givoni b (1992), comfort, climate analysis and building design guidelines. energy and buildings 18(1): 11-23. doi: http://dx.doi.org/10.1016/0378-7788 (92) 90047-k. givoni b (1994). passive low energy cooling of buildings: wiley. habte a, lopez a, sengupta m, wilcox s (2014), temporal and spatial comparison of gridded tmy, tdy, and tgy data sets (report no. nrel/tp-5d00-60886 united states10.2172/ 1126297mon may 12 11:16:12 edt 2014nrelenglish). http://www.osti. gov/ scitech//servlets/purl/1126297/. hall i, prairie r, anderson h, boes e (1978), generation of typical meteorological years for 26 solmet stations. sand78-1601. albuquerque, nm: sandia national laboratories. holmes mj, hitchin er (1978), an example weather year for the calculation of energy demand in buildings. cibse building services eng, 45(2): 186-189. huang j (1998), impact of different weather data on simulated residential heating and cooling loads. ashrae transactions 104(2): 516-527. lawrence berkeley national laboratory, and james j. hirsch & associates (jjh). (1998). overview of doe-2.2, report by simulation research group. levermore gj, doylend no (2002), north american and european hourly based weather data and methods for hvac building energy analyses and design by simulation. ashrae transactions, 108(2): 1053-1062. lomas kj, fiala d, cook mj, cropper p.c (2004), building bioclimatic charts for nondomestic buildings and passive downdraught evaporative cooling. building and environment 39(6): 661-676. doi: http://dx.doi.org/10.1016/j.buildenv.2003.1 2.011 marion w, urban k (1995), user’s manual for tmy2s typical meteorological years (report no. nrel/sp-463-7668). meteonorm 7. (2015). url: http:// meteonorm.com. [accessed: may/23/2015] national climate data center, (1976), test reference year (try), tape reference manual (report no. td-9706 ). asheville, north carolina. olgyay v, olgyay a (1963), design with climate: bioclimatic approach to architectural regionalism: some chapters based on cooperative research with aladar olgyay: princeton university press. petrakis m, lykoudis s, kassomenos p (1996), a software tool for the creation of a typical meteorological year. environmental software 11(4): 221-227. doi: http://dx.doi.org /10.1016/s0266-9838(96)00006-8. http://dx.doi.org/10.1016/j.buildenv.2003.12.011 http://dx.doi.org/10.1016/j.buildenv.2003.12.011 variant developments of typical meteorological years (tmys) for seeb, oman and their impact on energy simulation of residential buildings 141 pissimanis d, karras g, notaridou v, gavra k (1988), the generation of a “typical meteorological year” for the city of athens. solar energy 40(5): 405-411. doi: http://dx.doi.org/10.1016/0038-092x (88)90095-3. sawaqed nm, zurigat yh, al-hinai h (2005), a step-by-step application of sandia method in developing typical meteorological years for different locations in oman. international journal of energy research 29(8): 723-737. doi: 10.1002/er.1078. stoffel tl, rymes md (1998), production of the weather year for energy calculations version 2 (wyec2) data files. ashrae transactions 104(2). watson d (1981), analysis of weather data for determining appropriate climate control strategies in architectural design. in r. haisley (ed.), international passive and hybrid cooling conference. miami beach, florida (u.s.a.): solar energy association. watsun simulation laboratory. (1992). engineering data sets of hourly weather observations in wyec2 format (wyec2 files) and canadian weather for energy calculations (cwec files), user’s manual, prepared for environment canada atmosphere environment service and national research council canada. waterloo, ontario. wilcox s, and marion w (2008). users manual for tmy3 data sets (revised) (report no. nrel/tp-581-43156). microsoft word paper 7.docx the journal of engineering research (tjer), vol. 13, no. 2 (2016) 172-186 single cfta based current-mode universal biquad filter s.v. singhb, r.s. tomar a, * and d.s. chauhanc b department of electronics and communication engineering, jaypee institute of information technology, u.p. india *a department of electronics and communication engineering, anand engineering college, u.p. india c department of electrical engineering, g.l.a. university received 13 january 2016; accepted 15 may 2016 abstract: this paper introduces a new current-mode (cm) universal biquad filter structure with optimum number of active and passive elements. in the design, the proposed circuit uses a single active element namely, current follower trans-conductance amplifier (cfta) and two grounded capacitors as passive elements. the main feature of the proposed circuit is that it can realize all five standard filtering functions such as low pass (lp), band pass (bp), high pass (hp), band stop (bs) and all pass (ap) responses across an explicit high impedance output terminal through the appropriate selection of three inputs. in addition, the same circuit is also capable to simultaneously realize three filtering functions (lp, bp and hp) by the use of single current input signal. moreover, the proposed structure is suited for low voltage, low power operations and offers the feature of electronic tunability of pole-frequency and quality factor. further to extend the utility of the proposed circuit block higher order current-mode filters are also realized through direct cascading. a detailed non-ideal and parasitic study is also included. the performance of the circuits has been examined using standard 0.25 μm cmos parameters from tsmc. keywords: analog signal processing, biquad-filters, current-mode, tunable. אאאאא ، ∗ ،א    אא       א   w א אא  א   eא    f     א אא   ،א  kא k     إن  א ،א א א   א אא  אkאא،،،،אאאא، אאאאאkאא אא،אא،אאאkאא אאאאאאא،אאאאאk אא   אא אא א    kא  א א א ٢٥}٠ k  אאwةالفالتر الھوائي، القياسية معالجة اإلشارات ، ،א * corresponding author’s email: r_tomar15@rediffmail.com s.v. singh, r.s. tomar and d.s. chauhan 173 1. introduction analog circuit designers have paid a significant accent on the study and synthesis of currentmode (cm) continuous-time (ct) filter structures using variants of current conveyors suited for high performance analog signal processing applications (toumazou et al. 1990; kacar and kuntman 2011). in the last few decades, circuit designers primarily focused on designing low cost, low supply voltage operable simple filter structures realizing all the standard filtering responses. these filter structures are termed as universal filters. further, in the perspective of less chip area, cost reduction and lower average power consumption, the concluding remarks state that the circuit topology must be intended for an optimum use of active and passive elements in the biquad design (satansup and tangsrirat 2014). another concern pertinent to the integrated circuit realization point of view is the electronic tuning aspects of filter parameters used for the fine adjustment of filter parameters against large manufacturing tolerances, which can also be used to compensate for deviations in the circuit performance parameters due to process tolerance, port parasitic, temperature drift, imperfections and aging etc. (karybakas and papazoglou 1999; metin 2007). in the process, enormous papers on the synthesis of currentmode biquad filter circuits having one or more active elements are reported in the available literature (keskin et al. 2006; tomar et al. 2014a). the circuits reported in references are capable of realizing all the five standard filtering functions in current-mode either simultaneously (cicekoglu 2001; maheshwari and khan 2004; maheshwari and khan 2005) or one at a time (singh et al. 2014; tomar et al. 2014b; singh et al. 2012) or both way (wang et al. 2001) without changing the circuit structure. beside it, most of the circuit structure except few (cicekoglu 2001; wang et al. 2001) is resister less and canonical. however, all these circuits require more than one active element to realize filtering functions. with a view to being adaptable towards the simplicity, low cost, lower power consumption and space saving of the circuits, it is beneficial to design the biquad filter by the use of single active element and minimum number of passive elements. therefore, we have focused here only on the single active element based current-mode filters from the available literature (arsian et al. 2006; prasad et al. 2009) for the comparative purpose and hence, the detail features of all single active elements based current-mode biquad filters are compared and summarized in table 1. some of the circuits among the currentmode filters based on single active element as discussed in table 1 are configured as single input three output (sito) (mangkalakecree et al. 2009; prasad et al. 2009) and remaining are configured as three input single output (tiso) (arsian et al. 2006; satansup et al. 2014). the sito filters can realize at the most three filtering functions simultaneously whereas tiso filters can realize all five filtering functions but one function explicitly at a time. unfortunately, most of the filter circuits (arsian et al. 2006; chang et al. 2007; prasad et al. 2009; tangsrirat 2010) and except (satansup et al. 2012; satansup et al. 2014) use at least three passive elements in the design. beside this, some of the circuit configurations also employ floating passive elements (arsian et al. 2006; biolek et al. 2009; chang et al. 2007; prasad et al. 2009; sirirat et al. 2010; singh et al. 2013; tangsrirat 2010; yuce et al. 2004) and, hence, are not favorable for ic implementation. moreover, few circuits require inverted current input signals and/or component matching constraints (arsian et al. 2006; chang et al. 2007; chaturvedi and maheswari 2011; prasad et al. 2009; sirirat et al. 2010; satansup et al. 2014) too for the filtering functions realization. furthermore, none of the cm filters endorsed in table 1 can be configured as single input multiple output (simo) as well as multiple input single output (miso) simultaneously from the same structure and are inept to operate at low supply voltages with lower power consumption. the current-mode filter structure proposed in this paper consists of single cfta and two grounded capacitors and can realize lp, bp, hp, bs and ap filter functions in three input single output (tiso) and lp, bp, and hp filter functions in single input three output (sito) configurations. the active element cfta is slightly modified to get dual electronic tunability feature. the proposed structure is a resistor-less cmos topology which neither requires inverted current input nor matching constraints for the realization of any filter function(s). with suitable impedance levels the proposed filter blocks are directly cascaded to obtain higher order filters ( kacar and kuntman 2011; tomar et al. 2014a). the proposed circuits are evaluated through p-spice programs on cadence tools using 0.25μm tsmc cmos parameters (prommee et al. 2009). table 1. comparison among recent single active element based current-mode filters (prasad et al. 2009; arsian et al. 2006). attributes active element config. passive element s functions realized floating passive elements expli cit outp uts electronic tunability of ω and q designed pole frequency matching condition required inverted input required power supply power consumed reference (prasad et al. 2009) docdta simo 2r+2c lp,bp,hp 2r 1 yes na yes no ± 1 v na (chaturvedi et al. 2011) dvcc 2r+2c lp,bp,hp none none no 9.7 mhz yes no ± 2.5 v 4.9 mw (yuce et al. 2004) ccii 2r+2c lp,bp 1r none no na no no na na (herencsar et al. 2009) ccta 2c+2r lp, bp, hp none 1 yes 1 mhz no no ± 2 v na (prasad et al. 2013) vdta 1r+2c lp,bp,hp none 1 yes na no no ± 1 v na (singh et al. 2013) mocfta 1r+2c lp,bp,hp 1r 2 yes 1.35 mhz no no ± 1.5 v 1.18 mw (kacar et al. 2012) fdccii 2r+2c lp,bp or hp,bp none 2 no 10 mhz no no ± 1.3 v na (biolek et al. 2009) cdta 1r+2c lp,bp,hp 1r 1 yes 10 mhz no no na na (tangsrirat 2010) cfta 1r+2c lp,bp,hp 1r 1 no 159 khz no no ± 3 v (satansup et al. 2012) vdta 2c lp,bp,hp none 1 yes 3.03 mhz no no ± 2 v na (mangkalakecree et al. 2009) ccccta 2c lp,bp,hp none 2 yes 572 khz no no ± 1.5 v 1.42 mw (satansup et al. 2014) vdta miso 2c lp, bp, hp, bs, ap none 1 yes 3 mhz no yes ± 2 v 1 mw (chang et al. 2007) cccii 1r+2c lp, bp, hp, bs, ap 1r 1 yes 1 mhz yes yes ± 1.65 v na (kacar et al. 2012) fdccii 2r+2c lp, bp, hp, bs, ap none none no 1.59 mhz no no ± 3 v na (pandey and paul 2011) dvccta 1r+2c lp, bp, hp, bs, ap none none no 1.59 mhz no no ± 1.25 v na (sirirat et al. 2010) cfta 1r+2c lp, bp, hp, bs, ap 1r 1 no 1.12 mhz yes yes ± 3 v na (arsian et al. 2006) ccii 2r+2c lp, bp, hp, bs, ap 1c 1 no na yes no na na proposed cfta both simo/ miso 2c lp, bp, hp, bs, ap none none yes 8 mhz no no ± 0.75 v 0.6 mw 174 single c ft a b ased c urrent-m ode u niversal b iquad filter s.v. singh, r.s. tomar and d.s. chauhan 175 2. basics of cfta since after its inception as current-mode active element and its terminal impedance suitability for current-mode analog signal processing applications, cfta have been received considerable attention from the circuit designers (satansup and tangsrirat 2011; tomar et al. 2014a; singh et al. 2012; singh and maheswari 2013). the conventional cfta can be further modified for additional advantage of dual electronic tunability. the modified cfta with symbolic diagram is shown in fig. 1 and can be described through the following set of equations figure 1. schematic symbol of modified cfta. 22 11321 2 1 ,,,0 zmx zmxfzzzf vgi vgiiiiiv = ===== ± ± (1) realization of the modified cfta using cmos is depicted in fig. 2. the transconductance parameters gm1 and gm2 of the cmos implemented cfta (tomar et al. 2014b; singh et al. 2012) can be related with the biasing currents is1 and is2, respectively as: 11 snm ig β= and 22 snm ig β= (2) where, )l/w(coxnn μ=β is the process parameter of nmos transistors m8-m9, m15-m16 forming differential pairs at the transconductance stages of employed cfta. 3. proposed current-mode filter the proposed current-mode biquad filter circuit as shown in fig. 3 employs single cfta as the active element and two grounded capacitors. the use of grounded capacitors is elegant for integrability (bhusan and newcomb 1967). the proposed circuit in fig. 3 can be configured as either three input single output or single input three output configurations through appropriate selection of input and output current signal(s). the workability of the proposed current-mode filter structure can be described in following two different cases: 3.1 case-i: tiso filter configuration if the proposed circuit in fig. 3 is analyzed with three current input signals (iin1, iin2, iin3), the expression for current output (iout) can be obtained as: )s(d i)ggscgscc(sicgiscc i 3in2m1m12m 2 212in12m1in 2 21 out +++−− = (3) where, 211221 2)( mmm ggscgccssd ++= (4) figure 2. cmos implementation of cfta. single cfta based current-mode universal biquad filter 176 figure 3. proposed current-mode biquad filter. it can be easily retrieved from equation (3) that different cm filtering responses can be obtained at current output iout by selecting current inputs appropriately, as follows: • for iin1 = iin2 = iin3 = iin a non-inverted lp response is obtained. • for iin1 = iin and iin2 = iin3 = 0 an inverted hp response is obtained. • for iin2 = iin and iin1 = iin3 = 0 an inverted bp response is obtained. • for iin2 = iin3 = iin and iin1 = 0 a non-inverted bs response is obtained. • for iin2 = 2iin, iin3 = iin and iin1 = 0 a noninverted ap response is obtained. 3.2 case-ii: sito configuration with single current input (iin2=iin), the proposed circuit in fig. 3 is capable to realize lp, bp and hp responses, simultaneously across three different current outputs ilp, ibp, ihp. the transfer functions for received filtering responses are as follows: )( 12 sd scg i i m in bp −= (5) )( 21 sd gg i i mm in lp = (6) )( 2 21 sd scc i i in hp = (7) from cases i and ii, it can be remarked that proposed circuit in fig. 3 can realize all the standard filter functions (lp, bp, hp, bs, and ap) across an explicit high impedance current output terminal by the appropriate selection of three current input signals. in addition, the same structure realizes lp, bp, hp responses, simultaneously by the use of single current input signal. thus, the proposed circuit can be used as either tiso or sito current-mode biquad filter without changing the structure and neither requires any component matching constraint nor inverted current input signal for any of the response realization. the filter’s characteristic parameters like pole frequency (ω), quality factor (q) and bandwidth (bw) can be derived from the transfer function as: 21 21 cc gg mm=ω , 21 12 m m gc gc q = , 2 2 c g bw m= (8) on substituting trans-conductance parameters gm1 and gm2 from equation (2) we find that equation (8) turns out to be: , )( 21 2/1 21 cc ii ssnβω = 2/1 2 1 1 2       = s s i i c c q (9) from equation (9), it is evident that the pole frequency can be tuned precisely, without distressing quality factor, if the ratio of biasing currents is1 and is2 is kept to be constant. the band width (bw) of the proposed filter circuit can also be expressed as: 2 2/1 2 )( c i bw sn β = (10) equation (10) shows that the bw can solely be controlled by only biasing current is2. from equations (9) and (10), it is clearly seen that filter’s parameters ω and q can be electronically adjusted by regulating the only biasing current is1 without influencing bw. 4. non-ideal and parasitic effects 4.1 non-ideal analysis this section describes the non-ideal aspects of the proposed biquad structure. for this a non-ideal cfta can be modeled through modified set of equations, as follows: s.v. singh, r.s. tomar and d.s. chauhan 177 2222 1111321 ,,,0 zmx zmxfzzzf vgi vgiiiiiv γ γα ±= ±===== ± ± (11) where, εα −= 1 with )1( <<εε represents the current tracking errors between f to zi (i =1, 2, 3) terminals of cfta and depends on the ratio of pmos trans-conductance. with the use of identical pmos transistors at the input stage, α may approach to unity. similarly, γ1 and γ2 are the trans-conductance inaccuracies between z1 to ± x1 and z2 to ± x2 terminals, respectively. on considering these non-ideal factors of the cfta and re-analyzing the proposed filter of fig. 3, we get the similar denominator for each current output expressions as: 212112221 2)(' mmm ggscgccssd γαγαγ ++= (12) with involved non-idealities, the modified ω and q are obtained from equation (12) as: 21 2121 cc gg mmγαγω = , 212 121 m m gc gc q αγ γ = (13) from equation (13), it can be seen that minor deviations may occur in ω and q due to involved non-ideal factors. though these slight deviations can be ignored as non-ideal factors α, γ1 and γ2 are imminent to one at operational frequency. however, to what an extent the filter’s characteristic may actually be deviated due to possible changes in the design can be visualized by sensitivity coefficients which are as follows: , 2 1 2121 ,,,, =ω γγα mm ggs 2 1 s 21 c,c −=ω (l4) 2 1 1,21 , =q gc ms γ , 2 1 sq g,c, 2m12 −=γ (15) hence, we can state that the entire sensitivity coefficients are within ‘half’ in magnitude. 4.2 parasitic effects in this section, the performance of the proposed current-mode filter circuit in the presence of diverse parasitic impedance effects of the cfta are to be considered. a practical cfta, similar to any other active including various ports parasitic is shown in fig. 4. these are port z1, z2 and z3 parasitic in the form of rz1//cz1, rz2//cz2 and rz3//cz3 respectively, and ports x1, x2 parasitic in the form of rx1//cx1, rx2//cx2 and port ‘f’ parasitic in the form of rf. in fig. 4 the x2 port parasitic rx2 and cx2 appear between the high impedance terminal and ground. to abolish their effect, the cfta should be well designed to comprise a very low port parasitic rf. ideally, the value of rf is zero and terminal ‘f’ is virtually grounded. since these parasitic impedances are connected between true ground and virtual ground, these are, therefore, almost ineffective (singh et al. 2012). the external capacitances c1 and c2 can be chosen to be much greater than the parasitic capacitances at the z1, z2, x1 and x2 terminals of cfta i.e. c1, c2>> cz1, cz2, cx1, cx2. figure 4. modified cfta with port parasitic. to visualize the effects of various parasitic impedances on the performance of the proposed current-mode filter circuit, the proposed circuit is reanalyzed in the presence of various ports parasitic of cfta. for this purpose, currentmode bp response is taken in to consideration. by taking the above discussion and parasitic effects into account and on further reanalyzing the proposed current-mode filter, we get following expression for the bp filter at iout. ][][ ) 1 ( 21 21 2 22 112 2 m cc gg l c g ss rc s c g i i mmm z m in bp ++ + = (16) where, 121 2 2 11 1 zmcm rgc c rg l ++= (17) and 12111 11 1 zcmmzm rrggrg m ++= (18) with, rc = rz2 // rx1 (19) and hence, filter parameters in equations (8) to (10) are changed to: single cfta based current-mode universal biquad filter 178 mωω =* (20) ][][ 2 2 * l q l c g q m ωω ==      (21) and         =         = l m q l m cg cg q m m 12 21* (22) in equations (16)-(22), undesirable factors are yielded due to various ports parasitic of cfta. by considering c1=c2, and gm1=gm2 in the design, these undesirable factors can be eliminated or minimized and hence, the proposed filter may approach towards ideal response, if we choose the design criterion given in equation (23). , , 1 1 1 zr c >>ω (23) 5. simulation results in this section, the current-mode universal filter in fig. 3 has been simulated using pspice programs to authenticate the theoretical expectations. the employed cfta as given in fig. 1 was implemented using standard cmos technology (tsmc 0.25µm process parameters) as shown in fig. 2. the dimensions of mos transistors were obtained as summarized in table 2. the circuit was biased with power supply vdd = -vss= 0.75v and vbb = -0.32v and designed to realize biquadratic characteristics with f = 8 mhz at q = 1. in design, the active and passive components value were determined as is1 = is2 = 40 µa and c1 = c2 = 10 pf, which resulted in total average power consumption of about 0.6 mw. fig. 5 shows the simulated and ideal current-gain responses of bp, lp and hp filter functions for the sito configuration of the proposed current-mode filter circuit. similarly, for the tiso configuration the simulated and ideal current-gain and phase responses for lp, hp, bp, bs and ap filtering functions are shown in fig. 6. in fig. 5 and fig. 6 the difference between ideal and simulated results is clearly visible along with distortion (spike) beyond 670 mhz or so. which is obviously due to involved non ideal factors and parasitic effects dominating at higher frequencies. however, at working frequencies these non ideal factors were found to be unified as discussed in section 4. for different filtering responses the pole frequency was measured as 7.94 mhz, which deviated only by 0.75% from the numerical designed value of 8 mhz. to show the electronic tuning aspects of the proposed current-mode filter, the circuit is further simulated to obtain various bp responses at different sets of is1 and is2 in such a way so that is1=is2 = 5µa, 10µa, 30µa, 150µa, which result in to the pole frequency variation as 1.25 mhz, 3.22 mhz, 7.25 mhz and 12.03 mhz, respectively, at constant q = 1. the simulated result showing the pole frequency tuning aspect independent of q is shown in fig. 7. similarly, the simulation results showing the tuning feature of q independent of f0 for the sito configuration is also shown in fig. 8, which were obtained by maintaining the product of is1 and is2 to be a constant as (is1 =10µa, is2 =160 µa), (is1 =16µa, is2 =100 µa), (is1 =32µa, is2 =50 µa), and (is1 =64µa, is2 =25 µa) respectively. in fig. 9, the time-domain analysis of bp filter (sito configuration) is displayed, which was obtained by applying a sinusoidal input of 80µa peak to peak at 7.94 mhz. similarly, the timedomain analysis of lp filtering function (tiso configuration) is also shown in fig. 10, which was obtained by applying a sinusoidal input of 80µa peak to peak at 200 mhz. next, the thds of the lp response of the proposed circuit were also measured by applying a current input sinusoidal signal of peak amplitude of 40 µa and having varying frequency from 50 khz to 900 khz. it can be observed that the thd figures are adequate and not more than 3%, causes the output signal is not distorted significantly for an applied input signal of varying frequency as shown in fig. 11. to observe the effect of passive component mismatching on the filter’s performance, monte-carlo analysis has been performed for 100 samples. for this, the bp filter of sito configuration was simulated by selecting the values of capacitors c1 and c2 with 5% gaussian deviation. the statistical results in the form of plots are depicted in fig. 12, the simulated mean, median and standard deviations were respectively, 8.24 mhz, 8.22 mhz and 164.66 khz, which reveal that with respect to the simulated pole frequency of 7.94, the proposed filter is less sensitive to the change in capacitors value and thus offers reasonable passive sensitivity. 2 1 1 1<< m zg r 2 1 1<< m cg r s.v. singh, r.s. tomar and d.s. chauhan 179 figure 5. current-gain responses of lp, hp and bp filters for sito configuration of the proposed filter in figure 3. (a) (b) (c) single cfta based current-mode universal biquad filter 180 (d) (e) figure 6. simulated and ideal current-gain and phase responses for tiso configuration of the proposed filter in figure 3 (a) lp (b) hp (c) bp (d) bs and (e) ap filter functions. figure 7. simulated bp responses of sito configuration of the proposed filter in figure 3, showing variation in pole frequency, and keeping q = 1. s.v. singh, r.s. tomar and d.s. chauhan 181 figure 8. simulated bp responses of sito configuration showing variation in q, by keeping pole frequency at 7.94 mhz. figure 9.time domain response for bp filter of sito configuration of the proposed filter in figure 3. figure 10. time domain response for lp filter of tiso configuration of the proposed filter in figure 3. single cfta based current-mode universal biquad filter 182 figure 11. thd performance of lp response of the sito configuration for a sinusoidal current input of amplitude 40μa, for a frequency range 50 to 900 khz. figure 12. statistical results of monte-carlo analysis for the bp response of sito configuration of the proposed circuit with 5% deviation in capacitor values. 6. application: higher order filters to further support the proposed cm filter topology in fig. 3, it is used as a general purpose structure for realizing higher order current-mode filter (kacar and kuntman 2011; tomar et al. 2014a). for example, nth order (where n = even) bp filter and hp filter are shown in fig. 13 and fig. 14, which are implemented through direct cascading (maheshwari et al. 2004) of m = n/2, similar proposed blocks without using any impedance matching or coupling. direct cascading facilitates with no lower cut-off frequency limitations and circuit can operate from low to high frequency range. simulations are next performed for the 4th, 6th and 8th order fourth-order filter to illustrate the frequency domain performance of the circuit and corresponding results are shown in fig. 15 and fig. 16. in this design, the component values for each of the filter block (1, 2…..m) were chosen as ism= 40 µa and c1m = c2m = 10 pf. the voltage supply rails were used as vdd = -vss = 0.75 v with a bias vbb = 0.32 v. from different simulation results, it is evident that the higher order circuits realized through direct cascading are attractive choice and operated efficiently at low supply rails with least power consumption. s.v. singh, r.s. tomar and d.s. chauhan 183 figure 13. nth order band pass (bp) filter, derived through direct cascading of the proposed circuit blocks in figure 3. figure 14. nth order high pass (hp) filter, derived through direct cascading of the proposed circuit blocks in figure 3. table 2. dimensions of mos transistors in figure 2. mos transistors dimensions w (μm)/l (μm) m1, m2 1/0.25 m3-m7, m10-m14, m17-m21 3/0.25 m8,m9, m15, m16 15/0.25 m22-m27, m29-m33, m35m41 5/0.25 m24 10/0.25 m28, m34 4.2/0.25 single cfta based current-mode universal biquad filter 184 figure 15. current-gain and corresponding phase response for the nth-order (n = 2, 4, 6, 8) bp response of figure 13. figure 16. current-gain and corresponding phase response for the nth-order (n = 2, 4, 6, 8), hp response of figure 14. s.v. singh, r.s. tomar and d.s. chauhan 185 7. conclusion in this paper, a single cfta and two grounded capacitors based current-mode biquad filter is proposed and verified using simulation results. in addition, non ideal aspects and parasitic impedance effects of the cfta were also considered in details. the simulation results agree quite well with the theoretical expectations. the proposed circuit has a number of noticeable advantages as discussed below. the proposed circuit shows versatility by realizing lp, bp, hp, bs and ap functions in tiso as well as lp, bp, hp responses as the sito configuration. the use of a single active element makes the circuit topology simpler. moreover, the use of two grounded capacitors provides a canonical structure, leads to the option for integrated circuit implementation and for parasitic reduction. moreover, the circuit neither requires inverted current input(s) nor any component matching conditions for response realization, and thus eliminates the need of complex external circuitry. orthogonal control of q and ω via biasing currents suited for practical adjustments in music and speech synthesis. the proposed circuit is a low voltage and low power structure, hence it is suited for battery operated systems. expansion possibility of the proposed circuit for higher order realization through direct cascading validates its usefulness for analog signal processing. the cmos compatibility of the circuit is fully suited for low cost integrated circuit implementation. moreover, the circuit offer low active and passive sensitivity performance. a careful inspection of table 1 and introductory section of this study describes the superiority of the proposed filter block over other single active element based designs and finally concludes that none of the earlier reported current-mode filters offers all the above described features simultaneously. references arsian e, metin b, ciekoglu o (2006), multiinput single-output cascadable current-mode universal filter topology with a single current conveyor. third iasted international conference circuits, signals, and systems 62-66. bhusan m, newcomb rw (1967), grounding of capacitors in integrated circuits. electronics letters 3: 148-149. biolek d, biolkova v, kolka z (2009), currentmode biquad employing cdta. indian journal of pure and applied physics 47: 535-537. chaturvedi b, maheshwari s (2011), currentmode biquad filter with minimum component count. active and passive electronic components: 1-7. chang cm, huang th, tu sh, hou cl, horng jw (2007), universal active current filter using single second-generation current controlled conveyor. international journal of circuits, systems and signal processing 1: 194198. cicekoglu o (2001), current-mode biquad with a minimum number of passive elements. ieee transaction on circuits and systems-ii: analog and digital signal processing 48: 221222. fabre a, alami m (1995), universal current mode biquad implemented from two second generation current conveyors. ieee transaction on circuits and systems-i: fundamental theory and applications 42: 383385. herencsar n, koton j, vrba k (2009), single ccta-based universal biquadratic filters employing minimum components. international journal of computer and electrical engineering 1: 307-310. kacar f, kuntman h (2010), new current-mode filters using single fdccii with grounded resistor and capacitors. the online journal on electronics and electrical engineering (ojeee) 2(4): 341-345: ref. no.-w10-0040. kacar f, kuntman h (2011), a new improved cmos realization of cdta and its filter applications. turkish journal of electrical engineering and computer sciences 19(4): 631642. kacar f, yesil a, kuntman h (2012), currentmode biquad filters employing single fdccii. radioengineering journal 21: 12691278. karybakas ca, papazoglou ca (1999), low sensitive ccii based biquadratic filters offering electronic frequency shifting. ieee transactions on circuits and systems-ii 46: 527539. keskin aü, biolek d, hancioglu e, biolkova v (2006), current-mode khn filter employing current differencing transconductance amplifiers. international journal of electronics and communications 60: 443-446. maheshwari s, khan ia (2004), novel cascadable current-mode translinear-c single cfta based current-mode universal biquad filter 186 universal filter. active passive electronic component 27: 215-218. maheshwari s, khan ia (2005), high performance versatile translinear-c universal filter. active and passive electronic devices 1: 41-51. mangkalakecree s, duangmalai d, siripruchyanun m (2009), current-mode khn filter using single cc-ccta. in the seventh psu engineering conference: 306-309. metin b (2007), electronic tunability in currentmode filters. phd thesis. minaei s, türköz s (2004), current-mode electronically tunable universal filter using only plus-type current controlled conveyors and grounded capacitors. etri journal 26(4): 292-296. pandey n, paul sk (2011), vm and cm universal filters based on single dvccta. active and passive electronic components: 1-7. prasad d, bhaskar dr, singh ak (2009), universal current-mode biquad filter using dual output current differencing transconductance amplifier. international journal of electronics and communication 63: 497-501. prasad d, bhaskar dr, srivastava m (2013), universal current-mode biquad filter using a vdta. journal of circuits and systems 4: 32-36. prommee pp, angkeaw k, somdunyakanok m, dejhan k (2009), cmosbased near zerooffset multiple inputs max-min circuits and its applications. analog integrated circuits and signal processing 61: 93-105. satansup j, tangsrirat w (2011), single input five output electronically tunable currentmode biquad consisting of only zc-cftas and grounded capacitors. radioengineering 20: 273-280. satansup j, tangsrirat w (2012), single vdtabased current-mode electronically tunable multifunction filter. in proceedings of the iseec 2012: 266-272. satansup j, tangsrirat w (2014), compact vdta-based current-mode electronically tunable universal filters using grounded capacitors. microelectronics journal 45: 613618. senani r, singh vk, singh ak, bhaskar dr (2004), novel electronically controllable current mode universal biquad filter. ieice electronics express 1(14): 410-415. singh b, singh ak, senani r (2012), new universal current-mode biquad using only three zc-cftas. radioengineering 21(1): 273280. singh sv, maheshwari s (2012), currentprocessing current controlled universal biquad filter. radioengineering 21 (1): 317-323. singh sv, tomar rs, chauhan ds (2013), single mo-cfta based current-mode sito biquad filter with electronic tuning. international journal of computer science and information security (ijcsis) 11(4): 77-81. singh sv, tomar rs, chauhan ds (2014), current tunable current-mode tiso biquad filter consisting of two mcftas and minimum number of grounded capacitors. in proceedings of international conference on signal processing and integrated networks (spin) 555-560. sirirat j, prasetsom d, tangsrirat w (2010), high output impedance current-mode electronically tunable universal filter using single cfta. in proceedings of the iscit 2010 tokyo 200-203. tangsrirat w (2010), novel current-mode and voltage-mode universal biquad filters using single cfta. indian journal of engineering & material sciences 17: 99-104. tangsrirat w (2011), single input three output electronically tunable universal currentmode filter using current follower transconductance amplifiers. international journal of electronics and communication 65: 783-787. tomar rs, singh sv, chauhan ds (2014a), cascadable low voltage operated currentmode universal biquad filter. wseas transactions on signal processing 10: 345-353. tomar rs, singh sv, chauhan c (2014b), fully integrated electronically tunable universal biquad filter operating in current-mode. in proceedings of international conference on signal processing and integrated networks (spin), 549-554. toumazou c, lidgey fj, haigh dg (1990), analogue ics design: the current-mode approach. peter peregrinus, london, uk: chapter 13. wang hy, lee ct (2001), versatile insensitive current mode universal biquads using current conveyors. ieee transaction on circuits and systems-ii: analog and digital signal processing, 48: 409-413. yuce e, metin b, cicekoglu o (2004), currentmode biquadretic filters using single ccii and minimum number of passive elements. frequenz: journal of rf-engineering and telecommunication 58: 225-227. the journal of engineering (tjer), vol. 14, no. 2 (2017) 124-136 doi: 10.24200/tjer.vol.14iss2pp124-136 influence of exposure to an aggressive environment on cyclic fatigue response and life of an alloy steel s. gowda, k. manigandan, a. patnaika and t.s. srivatsan* department of mechanical engineering, the university of akron , akron, ohio 44325, usa received 20 march 2017; accepted 28 september 2017 abstract: corrosion or environment-induced degradation often occurs in structural steel as an electrochemical process which leads to gradual loss in mass over a period when subjected to prolonged exposure to an aggressive environment. immediate effects of this environment-induced degradation, also referred to in industry circles as corrosion, includes a progressive reduction of the cross section, which in turn has a detrimental influence on stiffness and load carrying capacity of the components in a structure, such as a bridge, a stiffened panel or a building. due to its high strength, low alloy a572 grade 50 steel is a potentially viable candidate for a wide range of applications in the construction industry. however, like in other high strength alloy steels, a572 is vulnerable to the effects of degradation-induced by the environment owing to its chemical composition. this paper discusses the details of tests conducted to determine the fatigue properties of a572 steel after inducing uniform environment-induced degradation or corrosion. flat (rectangular dog-bone shaped) specimens, conforming to specifications detailed in astm e8 standard, were used in this study. a technique that was developed by the astm and general motors corporation (gm) [called gmw14872] for a controlled corrosion process based on use of the spray technique was used to induce accelerated corrosion on selected test specimens in an environment chamber. stress-controlled high cycle fatigue tests were conducted on the corroded test specimens and compared with the asnew, uncorroded counterpart. keywords: a572 steel; corrosion; gmw14872; orientation; microstructure; fatigue life; degradation. على تفاعل االجهاد الدوري و عمرسبائك الصلب وعرةتأثري التعرض لبيئة ، أنيل باتنايك و ت.س. سريفاتسان، ك. مانيغاندان*سونيل غودا كهروكيميائية حيدث التآكل أو التدهور الناجم عن البيئة , يف كثري من األحيان , يف الصلب اهليكلي كعملية :امللخص مما يؤدي إىل نقصان تدرجيي يف الكتلة مبرور الوقت وعندما يتعرض لفرتات طويلة لبيئة قاسية. و يشار أيضا إىل اآلثار املباشرة هلذا التدهور الناجم عن البيئة يف دوائر الصناعة بالتآكل، ويشمل النقصان التدرجيي يف املقطع العرضي، والذي ضارا على صالبة و قدرة االجزاء احلاملة ملكونات اهلياكل مثل اجلسور و اللوحات الصلبة أو املبانى. بدوره يؤثر تاثريا , فهو املرشح املمكن و القابل لالستخدام يف جمموعة 50رقم a572ونظرا للقوة الشديدة لسبائك الصلب املنخفضة نوع وكما هو احلال مع غريه من السبائك شديدة القوة ، فإن نوع كبرية من التطبيقات يف صناعة البناء والتشييد. ومع ذلك، a572 يكون عرضة لتأثريات التاكل الناجم عن البيئة بسبب تركيبته الكيميائية. وتناقش هذه الورقة تفاصيل ئة. ومت بعد إحداث احنالل أو تآكل منتتظم بسبب البي a572االختبارات اليت أجريت لتحديد خصائص إجهاد نوع الصلب ألمريكية اجلمعية اعايري استخدام عينات مسطحة )مستطيلة على شكل عظام الكلب( و مطابقة للمواصفات املفصلة يف م ومت استخدام تقنية مت تطويرها من قبل أستم و شركة جنرال موتورز ( يف هذه الدراسة. astm e8لالختبارات واملواد ) (gm) [gmw14872 تآكل خاضعة للسيطرة ومبنية على أساس استخدام تقنية الرش الحداث التآكل [ لعملية ]تدعى املتسارع على عينات اختبار خمتارة يف غرفة البيئة. وأجريت اختبارات االجهاد ذات الدورة عالية الضغط اخلاضعة للسيطرة على عينات اختبار متآكلة ومت مقارنتها مع نظرياتها اجلديدة غري املتآكلة. .، التآكل، ، والتوجيه، االجزاء االجزاء الظاهرة، ,عمر االجهاد ، التاكل572: الصلب أ كلمات املفتاحيةال * corresponding author’s e-mail: tsrivatsan@uakron.edu mailto:tsrivatsan@uakron.edu s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 125 1. introduction based both on economy and strength, structural steel can be safely categorized to be one of the most economically viable and preferentially attractive choice for selection and use in a spectrum of infrastructure-related applications to include buildings, bridges, transportation ways, and even industrial equipment (banerjee 1965). these steel structures are often built using a combination of various standard shapes, the dimensions of which are both defined and documented by the american institute of steel construction [aisc] (american institute of steel construction 2005) to be wide flanged sections, angled members, channel sections and even hollow structures. since these steel structures are often exposed to environments spanning a range of aggressiveness, a gradual loss of metal due to environment-induced degradation, referred to in industry relevant circles as corrosion, becomes a critical issue during the service life of a structure. the loss of metal, or metal-based material, results in a progressive decrease in strength of the structure, which eventually leads to failure of the structural element, or structure, and concomitant risk to safety and overall well-being (banerjee 1965). common types of corrosion encountered by steel-based structures include the following: (i) uniform corrosion, (ii) galvanic corrosion, (iii) pitting corrosion, (iv) stress corrosion cracking, and (v) crevice corrosion (fontana 1986). upon identification of the type of corrosion it is up to the discretion of the engineer to determine viable ways to either mitigate or totally prevent them from occurring and thereby suppress deterioration and/or degradation. the scope of this research study was limited to evaluating and understanding the uniform corrosion behavior of a high strength alloy steel. the influence of corrosion in a sodium chloride (nacl) aqueous solution and cyclic loading is known to affect the long term mechanical properties and cyclic endurance of metallic alloys belonging to both the ferrousbased and non-ferrous-based families (konda 1979; xe et al. 2008; laird and duquette 1973; pelloux and genkin 2008; plin-luc et al. 2010; ebara et al. 1980; mcadam 1927; devos et al. 2006; metals handbook: properties and selection 1990; may et al. 2012 and liu et al. 2011). the mechanical response of structural steels is governed by the conjoint and mutually interactive influences of chemical composition, history of processing used to engineer the starting material, intrinsic microstructural effects and test temperature. structural astm a572 grade 50 steel was chosen for this research study with the primary purpose of determining the influence of corrosion on cyclic stresscontrolled high cycle fatigue response. the steel plates used in this study were manufactured and provided by alro steel (akron, oh). this alloy steel is a high performance structural steel based on its chemical composition and mechanical properties (metals handbook: properties and selection 1990). specific details pertaining to secondary processing treatment given to the as-provided steel stock were proprietary by the manufacturer. the minimum strength allowed for astm a572 grade 50 steel is yield strength [ σys] of 345mpa and ultimate tensile strength [σuts] of 450mpa. general, or uniform corrosion, is the most prevalent form of corrosion r e s u l t i n g i n overall loss of surface material, which culminates in progressive thinning of the structural member. this type of corrosion often accounts for the largest percentage of corrosioninduced damage experienced by steel structures that are often exposed to aggressive aqueous environment and/or gaseous environment (fontana 1986 and kondo 1979). in fact, general or uniform corrosion has been regarded as one of the most easily identifiable form of corrosion due to the overall nature of the corrosion process. since there does occur a uniform loss of material, the corrosion rate is also considered to be uniform. besides an observable loss in mass there also occurs a degradation in strength and a noticeable drop or decrease in ductility of the chosen material. the magnitude of degradation experienced by the material being proportional to the extent and severity of corrosion resulting because of materialenvironment interaction. uniform corrosion has been regarded by researchers t o b e less dangerous and/or damaging type of corrosion due to its easily identifiable nature (fontana 1986; laird and duquette 1973; pelloux and genkin 2008; plin-luc et al. 2010; ebara et al. 1980 and mcadam 1927). one of the simplest examples of uniform corrosion is the formation of an oxide product that protects the steel from weathering. a newly weathered steel is generally coated with very fine in size mill scale that gradually flakes off because of continued weathering action and corrosion. this in turn tends to expose the base metal to the aggressive environment. a gradual progression of the corrosion process occurs until influence of exposure to an aggressive environment on cyclic fatigue response and life of an alloy steel 126 such time the surface is covered with its own corrosion product. presence of corrosion product on the surface does tend to reduce, or decelerate, the rate of corrosion by forming a protective barrier between the metal and the environment (devos et al. 2006 and metals handbook: properties and selection 1990). uniform corrosion has been observed to occur in bridges made from structural steel where in structural components often having a large surface area are easily prone to either attack or oxidation because of exposure to an aggressive environment (devos et al. 2006 and metals handbook: properties and selection 1990). usually, the small pits on the surface of the steel tend to grow and eventually coalesce. this can safely be categorized as uniform corrosion rather than pitting corrosion. the effect of corrosion on cyclic stresscontrolled high cycle fatigue response of a high strength martensitic stainless steel, which is available for use in aeronautical applications was shown by may and co-workers (may et al. (2012). based on fractographic observations and interpretations coupled with results obtained from mechanical tests, the mechanisms for fatigue crack initiation in the environments of laboratory air and aqueous sodium chloride (nacl) solution were recorded. it was found that repeated rupture of the “local” passive film was largely responsible for favoring the occurrence of crack initiation during cyclic loading. the film experiences repeatedly rupturing by strain in the material that is induced by the applied cyclic stress during cyclic fatigue testing. this then causes the material, i.e. steel, to experience increased environment-induced attack at the ‘local’ level. in another study, the initiation of fine microscopic cracks was found to accelerate by preferential corrosion at locations of high “local” plastic strain of the high strength steel 38crmoal. the effect of using an accelerated pre-corrosion salt spray environment on fatigue behavior was investigated. the fatigue life response characteristics of the pre-corroded test specimens were studied using statistical probability methods, and both mathematical expectations and standard tolerances for fatigue life of the material following different precorrosion times were determined. it was concluded from this study that presence of corrosion-induced pits accelerated the fatigue crack initiation and exacerbated fatigue crack growth while concurrently reducing fatigue life of the material. the objective of this research paper is to present and discuss the results of a recent study aimed at examining the influence of exposure to an aggressive aqueous environment on cyclic fatigue response and life of high strength alloy steel a572 and thereby assess the severity of degradation in fatigue life arising from exposure to an aggressive environment. 2. experimental procedures 2.1 corrosion [procedure: gmw14872] the gmw14872 (gmw14872: cyclic corrosion laboratory test 2006) procedure is an accelerated laboratory corrosion test for evaluating the corrosion properties of metals. concentration of salt solution, temperature, and humidity are important factors that must be considered in studying the accelerated corrosion behavior of metals. the gm14872 was categorized to be an effective method for evaluating different types of corrosion ranging from general (uniform), to galvanic and including crevice corrosion. the individual parameters, such as: (i) concentration of salt solution, (ii) test temperature, and (iii) humidity, can be varied with the primary purpose of achieving a desired level of damage due to corrosion. the test is run in terms of a cycle. each cycle represents 24 hours (1 day) of testing. the required loss in mass experienced by a coupon is obtained for a targeted exposure time. often in this test, 15 cycles of exposure were carried out, which is divided into 3 categories. this cycle test also considers the weekends. 2.1.1 preparation of salt solution the testing solution is a combination of sodium chloride (nacl), calcium chloride (cacl2) and sodium bicarbonate (nahco3). the sodium chloride is of the reagent type or food grade. calcium chloride is also of reagent type and sodium bicarbonate can be either baking soda type or comparable product type. both cacl2 and nahco3 chemicals were dissolved separately in water, and then gradually added to solution of sodium chloride. 2.1.2 initial cleaning of the test coupon 1. the coupons were first washed using a required amount of acetone with the primary purpose of removing any dye or paint on the surface of the metal. this was followed by: (i) rinsing with distilled water, s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 127 (ii) drying the coupon with a lint free towel, and (iii) gentle scrubbing. 2. the samples then had their surfaces touched up using 600-grit emery paper with the purpose of removing any oil, mill scale and other corrosion products adhering to the surface. subsequently, they were labeled, taped, thickness measured, and precision weighed. 3. it is critical that all forming or preservation oils/lubes be removed prior to exposure to facilitate the occurrence of general/uniform corrosion of the coupon. this process was aided by using a commercial grade degreaser prior to cleaning the surface with either methanol or acetone. 2.1.3 post corrosion cleaning according to procedures detailed in astm standard (astm g1-03 2011), standard practice for preparing, cleaning and evaluation corrosion test specimens, this essentially involved cleaning of the corrosion products resulting as a direct consequence of the corrosion process. these can be categorized to be (i) mechanical, (ii) chemical, and (iii) electrolytic. however, the scope of this research is limited to the chemical method used for purpose of removal of corrosion products. an ideal procedure should remove only the corrosion products and should not result in removal of the base metal. to determine the loss in mass experienced by the base metal when removing the corrosion-induced products, duplicate un-corroded control specimens were also cleaned using the same procedure that was used for the test specimens. by weighing the control specimen, both before and after cleaning, the extent of loss in mass, experienced by the metal as a direct consequence of cleaning can be utilized to correct for the mass loss resulting due to environment-induced corrosion. chemical procedures involved immersion of the corrosion test specimen in a specific solution that is designed to remove the corrosion products with minimal dissolution of any base metal. a few procedures are listed in astm g1-03 (2011). the exact choice of the chemical procedure to be used can be categorized to be one of trial and error with the primary objective of establishing the most effective method for a specific metal and type of product scale due to corrosion. in this study, the c 3.2 corrosion procedure was used for purposes of cleaning. a description of this cleaning procedure is summarized in table 1. the effect of different cleaning procedures on tensile strength of cylindrical samples [astm e8] is shown in fig. 1. from this graph, it can be inferred that the astm c 3.2 cleaning procedure for purpose of removing corrosion products from the surface of a steel specimen is not a true representation of the cleaning process as it is not only removes rust on the surface but also a considerable amount of the base metal resulting in a noticeable drop in tensile strength of the test specimen. instead, the astm c 3.1, also called as clarke’s solution was used for removing corrosion products from the surface of the test specimen as it does not adversely influence the base metal. figure 1. effect of astm cleaning procedures on the tensile strength of a572 steel. influence of exposure to an aggressive environment on cyclic fatigue response and life of an alloy steel 128 table 1. chemical cleaning procedures for removal of corrosion products according to astm g1 [0]. c 3.1 is adopted for this study. designation material solution time temperature remarks c 3.2 iron and steel 50 g naoh 200g granulated zinc or zinc chips reagent water to make 1000 ml 30 to 40 mins 80 to 90oc caution should be exercised in the use of any zinc dust since spontaneous ignition upon exposure to air can occur. c 3.1 iron and steel 1000 ml hcl, 20g antimony trioxide, 50g stannous chloride, 50g naoh 1 to 25 mins 20 to 25oc solution should be vigorously stirred or specimen should be brushed. longer times may be required in certain instances. 2.1.4 preparation of coupon rack prior to the initiation of the test; a coupon rack to hold sufficient coupons was prepared. for this test, two types of racks were used. the rack for the rectangular coupons had an angled slot to ensure that the test solution stays on the surface without having to run down, which aids in the occurrence of uniform corrosion of the metal surface. the rack for the cylindrical specimens had grooved slots for facilitating exposure of test specimen to the aggressive aqueous environment. 2.1.5 process of testing the gmw14872 (gmw14872: cyclic corrosion laboratory test 2006) corrosion cycle was repeated every day until requirements for test exposure were met. the testing was continued through the weekends for decreasing the test time provided the number of cycles (n) and mass loss requirements were met. for each salt mist application, a spray apparatus was used to mist the test samples and coupons until all areas were thoroughly wet and/or dripping. the quantity of spray applied was adequate to visibly rinse away any salt accumulation or residue left from the previous spray. at the end of each cycle (24 hours), the coupons were: (i) removed from the test chamber, (ii) washed with de-ionized water (di water), (iii) sprayed with the test solution, and (iv) put back into the chamber. the different stages [ambient stage, humid stage and dry stage) were automatically controlled by the test chamber. 2.2 mechanical testing the tensile and cyclic fatigue tests were performed on a fully automated, servohydraulic instron testing machine equipped with a 100kn load cell. all the tests were performed in the room temperature, laboratory air environment (t = 27oc, relative humidity of 55 pct.). an axial 12.5mm gage length clip-on type extensometer was attached to the test specimen at the gage section using elastic bands. the stress and strain measurements parallel to the load line and resultant mechanical properties were provided as an output by the control unit of the servo hydraulic test machine. the stress amplitude-controlled high cycle fatigue tests were performed using a sinusoidal waveform and in conformance with procedures outlined in astm: e466-07 [17]. cyclic fatigue tests were conducted over a range of stress amplitudes (δσ /2), at a constant frequency of 5 hz and at a load ratio of 0.1, to determine the fatigue life (nf). from the test results, the following relationships were established: a) variation of fatigue life (nf) with stress amplitude (σ/2). b) variation of fatigue life (nf) with maximum elastic strain (σ max /e). s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 129 3. results and discussion 3.1 initial microstructure optical microstructure of the chosen astm a572 sample (chemical composition given in table 2) is as shown in fig. 2 for both the longitudinal (l) and transverse (t) orientations. the observed microstructure is typical of high carbon, low alloy content steel. the primary micro-constituent is the carbon depleted white region that is referred to as ferrite. the grains are distinct with well-defined grain boundaries, small and non-uniform in shape. the secondphase particles primarily the carbides were dispersed randomly through the microstructure. the presence and morphology of the ferrite micro-constituent is governed both by chemical composition and the processing techniques, both primary and secondary, used to engineer this steel. these intrinsic features, namely: the size and morphology of the ferrite micro constituent along with randomly distributed carbide particles, are an important factor that determines the hardness, tensile properties and cyclic fatigue response, quantified in terms of fatigue life. 3.2 tensile properties the tensile properties for both the longitudinal (l) and transverse (t) orientations for the chosen structural steel, for the two different exposure times, i.e. 7 days and 14 days, are summarized in table 3. results reported are the mean values based on duplicate tests. for testing in laboratory air and at the two chosen exposure times; that is, 7-day exposure and 14-day exposure, the yield strength and ultimate tensile strength was consistent, i.e. well within the range of 5-10 percent, in both the longitudinal (l) and transverse (t) orientation. in both orientations, i.e. longitudinal and trans verse, the tensile strength was noticeably higher than the yield strength indicating the occurrence of strain hardening beyond yield. upon exposure for time periods of 7-days and 14-days to an aggressive aqueous environment, the tensile yield strength and ultimate tensile strength of the chosen alloy steel revealed an observable decrease in comparison to values for the test specimen in laboratory air (27oc) and no exposure to an aggressive environment. for a fixed exposure time to the aggressive aqueous environment the elongation was consistent in both the longitudinal (l) and transverse (t) orientations. time of exposure to the aggressive aqueous environment was observed to have minimal influence on ductility, quantified by elongation over a gage length of 12.5 mm, of the chosen alloy steel. reduction in test specimen cross-section area, another measure of ductility, was identical in both the longitudinal (l) and transverse (t) orientations, for a given exposure time to the environment. figure 2. optical micrographs showing microstructure of a572 steel in the two orientations: (a) longitudinal at 200x, and (b) transverse at 200 x. s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 130 further, the reduction in test specimen crosssection area was observed to noticeably decrease for both the longitudinal (l) and transverse (t) orientations with an increase in time of exposure to the aggressive aqueous environment upon comparing with the values obtained for the tests done in room temperature (270c), laboratory air environment. table 2. nominal chemical composition of the chosen alloy steel a572 (in weight percent) c mn p s a572 0.25 1.35 0.04 0.05 table 3. uniaxial tensile properties of a572 alloy steel for the different conditions. exposure condition orientation elastic modulus yield strength uts elongation (%) reduction in area (%) ksi gpa ksi mpa ksi mpa no exposure longitudinal 30457 210 68 471 73 505 27 41 transverse 30312 209 65 447 75 516 28 41 7-day exposure longitudinal 30167 208 64 441 68 473 25 33 transverse 30457 210 58 405 68 474 26 33 14-day exposure longitudinal 30167 208 53 365 58 403 26 23 transverse 30167 208 56 387 66 459 23 23 figure 3. influence of test specimen orientation on variation of maximum stress (σmax) with fatigue life (nf) for a572 steel when cyclically deformed for as-new specimens and at a load ratio of 0.1. fatigue life (n f ) (cycles) 104 105 106 107 m a x im u m s tr e s s ( m p a ) 250 300 350 400 450 500 longitudinal transverse s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 131 3.3 cyclic stress amplitude controlled fatigue 3.3.1 as-received, unexposed test samples the test results of the stress amplitudecontrolled cyclic fatigue tests for both the longitudinal (l) and transverse (t) orientations in the room temperature (27oc) environment are shown in fig. 3 as the variation of cyclic fatigue life (nf) with maximum stress (σmax). the curves in this figure reveal a gradual increase in cyclic fatigue life (nf) with a decrease in maximum stress (σmax). this is very much in conformance with the trend shown by most ferrous alloys and their composite counterparts. to have a better understanding of the variation of maximum stress (σmax) with cyclic fatigue life (nf) in the two chosen orientations [longitudinal (l) and transverse (t)] the variation of maximum stress with fatigue life [nf] is plotted as the variation of maximum elastic strain [σmax /e] with fatigue life [nf]. this is shown in fig. 4 and reveals a linear trend. at equivalent values of maximum elastic strain [σmax /e] the transverse (t) orientation had higher cyclic fatigue life than the longitudinal (l) orientation. this clearly reveals the specific role and contribution of intrinsic microstructural features and resultant intrinsic microstructural effects in governing stress versus fatigue life (nf) response of the un-exposed test samples as a function of test specimen orientation. figure 4. influence of test specimen orientation on variation of maximum elastic strain (σmax/e) with fatigue life (nf) for a572 steel when cyclically deformed for as-new specimens and at a load ratio of 0.1. figure 5. influence of test specimen orientation on variation of max stress (σmax) with fatigue life (nf) for the a572 steel when cyclically deformed at a load ratio of 0.1 following 7-days of exposure to environment. fatigue life (n f ) (cycles) 104 105 106 107 m a x im u m e la s ti c s tr a in [ m a x /e ] 0.0012 0.0014 0.0016 0.0018 0.0020 0.0022 0.0024 0.0026 0.0028 longitudinal transverse fatigue life (n f ) (cycles) 104 105 106 107 m a x im u m s tr e s s ( m p a ) 250 300 350 400 450 500 longitudinal transverse influence of exposure to an aggressive environment on cyclic fatigue response and life of an alloy steel 132 3.3.2 exposure for 7 days to the aqueous environment the variation of maximum stress (σmax) with cyclic fatigue life (nf) is shown in fig. 5 for the test samples that were cyclically deformed following seven days of exposure to the aggressive aqueous environment. this behavior or trend shown is quite like behavior observed for the as-received, unexposed text samples. the cyclic fatigue life (nf) increased with a decrease in maximum stress (σmax). at equivalent values of maximum stress, the fatigue life of the test specimen taken from the longitudinal (l) orientation was less than the transverse (t) counterpart (fig. 5). variation of maximum elastic [σ max/e] strain with cyclic fatigue life (nf), shown in fig. 6, reveals a linear trend, and at equivalent values of maximum elastic strain the longitudinal (l) orientation was inferior to the transverse (t) orientation by more than one-order of magnitude. 3.3.3 exposure for 14 days to aggressive aqueous environment the test results for samples that were exposed for full 14 days to the aggressive aqueous environment are shown in figs. 7 and 8. variation of maximum stress with cyclic fatigue life (nf) is as shown in fig. 7. the chosen steel following prolonged exposed to an aggressive environment reveals a pattern of increasing fatigue life with a decrease in maximum stress, a trend shown by most other ferrous alloys and to even their composite counterparts. at near equivalent values of the maximum stress, the test specimens taken from the longitudinal (l) orientation revealed inferior fatigue life (nf) when compared to the test specimens taken from the transverse orientation (t). from a design engineering perspective, the variation of maximum elastic strain [σmax/e] with fatigue life is linear and at equivalent figure 6. influence of test specimen orientation on variation of maximum elastic strain (σmax/e) with fatigue life (nf) for the a572 steel when cyclically deformed at a load ratio of 0.1 following 7-days of exposure to the environment. figure 7. influence of test specimen orientation on variation of max stress (σmax) with fatigue life (nf) forthe a572 steel when cyclically deformed at a load ratio of 0.1 following 14-days of exposure to the environment. fatigue life (n f ) (cycles) 104 105 106 107 m a x im u m e la s ti c s tr a in [ m a x /e ] 0.0012 0.0014 0.0016 0.0018 0.0020 0.0022 0.0024 0.0026 0.0028 longitudinal transverse s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 133 values of maximum elastic strain the test samples from the longitudinal (l) orientation revealed at least an order of magnitude inferior fatigue life when compared with the samples taken from the transverse (t) orientation (fig. 8). influence of time of exposure to aggressive aqueous environment for test specimens taken from longitudinal (l) orientation of the asprovided plate stock is shown in fig. 9 as the variation of maximum stress (mpa) with fatigue life (nf). minimal to no difference is observed for the trend shown by the unexposed samples and the test samples exposed for 14 full days to the aggressive aqueous environment. however, the test specimens that were exposed for seven days to the aggressive environment revealed an observable decrease in cyclic fatigue life (nf) at equivalent values of maximum stress. this can be ascribed to the formation and presence of a thin layer of corrosion deposits on the surface of the test sample coupled with the occurrence and presence of fine microscopic pits that are detrimental to crack initiation during continued cyclic loading. influence of time of exposure to aggressive aqueous environment on test specimens prepared from the transverse (t) orientation of the as-provided alloy steel (a572) stock is shown in fig. 10. this figure reveals that at equivalent values of maximum stress the test specimens that were exposed for seven days to the environment revealed marginally better cyclic fatigue life (nf) when compared to the unexposed counterpart. this is attributed to the protective nature of the corrosion-induced deposits on the sample surface upon exposure to an aggressive environment. however, an increase in exposure time of the test specimen to the aggressive aqueous environment was observed to be detrimental to cyclic high cycle fatigue resistance by way of decreased fatigue life. at equivalent values of maximum stress the decrease in cyclic fatigue life increased at the lower values of maximum stress and resultant enhanced fatigue life (nf). this can essentially be attributed to rupture of the brittle corrosion deposits that form on the sample surface under the extrinsic influence of sustained cyclic loading. 3.4 viable mechanisms governing failure at the fine microscopic level in this research paper, an attempt has been made to present and briefly discuss the conjoint and mutually interactive influences of time of exposure to an aggressive environment and test specimen orientation on high cycle fatigue response and life of astm a572 high strength alloy steel. the observed fracture behavior is quantified based on both the macroscopic mode and the underlying mechanisms governing deformation and fracture occurring at the fine microscopic level. under conditions of cyclic loading in tensiontension [positive load ratio (r) of 0.1] the introduction and presence of dislocations in the metal matrix coupled with their gradual build up, or accumulation, at locations of: (i) the grain boundaries. (ii) interfaces of the soft and plastically deforming ferrite matrix and the hard and elastically deforming second-phase, which in this case is predominantly pearlite, and (iii) presence of other hard and elastically deforming second-phase particles, namely the carbide particles, that are randomly distributed through the microstructure, are conducive for an early initiation of fine microscopic voids during cyclic deformation. the formation of voids is favored to occur when the local strain caused by the gradual build-up of dislocations at the interfaces of the soft and plastically deforming metal matrix with the hard and elastically deforming secondphase particle reaches a critical value. the conditions prevailing for an early initiation of fine microscopic voids become appropriate upon the occurrence of yielding at the “local” level. with continued cyclic loading, a few of the hard and essentially elastically deforming second-phase particles tend to easily fail by cracking because their intrinsic brittleness when compared one-on-one with the soft and plastically deforming metal matrix, i.e., ferrite micro-constituent. since gradual growth of the fine microscopic voids coupled with extension of the crack is favored to occur at the prevailing high local stress intensities, the presence of several microscopic and macroscopic voids contribute to decrease in fatigue life (nf) of the test specimen [18]. s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 134 figure 8. influence of test specimen orientation on variation of maximum elastic strain (σmax/e) with fatigue life (nf) for the a572 steel when cyclically deformed at a load ratio of 0.1following 14-days exposure to the environment. figure 9. influence of tine of exposure to aggressive aqueous environment on the variation of maximum stress (mpa) with cyclic fatigue life (nf) for test specimens taken from the longitudinal orientation of the chosen a572 steel. fatigue life (n f ) (cycles) 104 105 106 107 m a x im u m e la s ti c s tr a in [ m a x /e ] 0.0012 0.0014 0.0016 0.0018 0.0020 0.0022 0.0024 0.0026 0.0028 longitudinal transverse fatigue life (n f ) (cycles) 105 106 m a x im u m s tr e s s ( m p a ) 280 300 320 340 360 380 400 unexposed exposed for 7 days exposed for 14 days s. gowda, k. manigandan, a. patnaika and t.s. srivatsan 135 figure 10. influence of tine of exposure to aggressive aqueous environment on the variation of maximum stress (mpa) with cyclic fatigue life (nf) for test specimens taken from the transverse (t) orientation of the chosen a572 steel. 3.5 potential future area of research study based entirely on the experiments executed systematically with precision and the results obtained, there now exists a thought by the authors to pursue numerical modeling through finite element analysis the prediction of behavior or response of these structures when exposed to aggressive environments for prolonged periods of times extending far and beyond 14-day exposure. based on this preliminary experimental study we plan on using the numerical approach to provide valuable information on structural response when exposed to aggressive environments for an extended period. 4. conclusion this study was undertaken with the objective of understanding the specific role of time of exposure to an aggressive environment and test specimen orientation on tensile properties, cyclic high cycle fatigue response and final fracture behavior of the high strength low alloy steel a572 following are the key findings. 1. light optical micrographs reveal a microstructure that is rich in the microconstituent ferrite, i.e. the carbon depleted region. the grains are small and of varying shape. the second-phase particles are randomly dispersed through the microstructure. 2. at a given value of maximum stress, test specimens taken from the longitudinal [l] orientation had noticeably lower cyclic fatigue life than the test specimens machined from the transverse [t] orientation. this is attributed to the role and/or contribution from intrinsic microstructural effects during cyclic loading. 3. the noticeable inferior high cycle fatigue resistance for the test specimens of the alloy taken from the longitudinal [l] orientation when compared one-on-one with the transverse [t] counterpart is evident at all values of the ratio of maximum stress and maximum elastic strain for the chosen alloy steel (a572) in the specific orientation. 4. for test specimens taken from both the longitudinal [l] and transverse [t] orientations, that were exposed for 168 hours (7 days) to the environment prior to cyclic fatigue testing, the test specimens taken from the longitudinal [l] orientation of the alloy steel plate had an order of magnitude lower fatigue life than the test specimens prepared from the transverse orientation. 5. for test specimens of alloy steel a572 taken from both the longitudinal [l] and fatigue life (n f ) (cycles) 104 105 106 m a x im u m s tr e s s ( m p a ) 360 380 400 420 unexposed exposed for 7 days exposed for 14 days influence of exposure to an aggressive environment on cyclic fatigue response and life of an alloy steel 136 transverse [t] orientations and exposed for 14 full days to the aggressive environment prior to cyclic fatigue testing, both orientations show a trend of increase in fatigue life with decrease in maximum stress. conflict of interest the authors declares no conflicts of interest. funding the research work presented in this paper was partially funded by the dod technical corrosion collaboration (tcc), u.s. dod office of corrosion policy and oversight. the research was administered by the u.s. construction engineering research laboratory under agreement number fy12 w9132t-11-c-0035. references american institute of steel construction (2005), (aisc): steel construction manual in unified 13th edition, chicago, il, usa. astm e466 (2011), standard test method for conducting constant amplitude axial fatigue tests on metallic materials. astm, american society for testing and materials, philadelphia, pa, usa. astm g1-03 (2011), standard practice for preparing, cleaning and evaluating corrosion test specimens. american society for testing and materials, philadelphia, pa, usa. bajer aj, laura ej, wei rp (1965), structure and properties of ultrahigh strength steels. astm, stp 370, american society for testing and materials 3-14. banerjee br (1965), structure and properties of ultrahigh strength steels, astm, stp 370, american society for testing and materials, philadelphia, pa, usa. 94-109. devos o, gabrielli g, tribollet b (2006), simultaneous eis and in situ microscope observation on partially blocked electrode application to scale electrodeposition. electrochemical acta, 51: 1413-1422. ebara r, yamada y, goto y (1980), corrosion fatigue behavior of 13cr stainless steel and ti-6al-4v ultrasonic frequency. fatigue ultrasonore, the minerals, metals and materials society (tms), warrendale, pa, usa (1980). fatigue strength of martensitic stainless steel x12crnimov12-3. international journal of fatigue (doi: http: // dx.doi.org / 10.1016 / j.ijfatigue.2012.09.018. fontana mg (1986), corrosion engineering, mcgraw-hill publishers. gmw14872 (2006), cyclic corrosion laboratory test. general motors corporation, michigan, usa. kondo y (1979), prediction of fatigue crack initiation life based on pit growth. corrosion 45: 7-11. laird c, duquette dj (1973), mechanisms of fatigue crack nucleation. in corrosion fatigue: chemistry, mechanics and microstructure, [editors: o. devereux, a.j. mcevily and r.w. stachle], national association corrosion engineering (nace) – 2 88. may me, palin-luc t, saintier n, devos o, (2012), effect of corrosion on the high cycle fatigue strength of martensitic stainless steel x12crnimov12-3. international journal of fatigue (2012), doi: http: // dx.doi.org / 10.1016/j.ijfatigue.2012.09.018. mcadam dj (1927), corrosion fatigue of metals. journal transaction american society steel treat, 11: 355-79. metals handbook: properties and selection (1990), classification and designation of carbon and low alloy steels. tenth edition, asm international, materials park, ohio, usa. palin-luc t, pérez-mora r, bathias c, domínguez g, paris pc, luis aran j (2010), ‘‘fatigue crack initiation and growth on a steel in the very high cycle regime with sea water corrosion. engineering fracture mechanics 77: 1953-1962. pelloux r, genkin jm (2008), chapitre 10: fatigue-corrosion’’, in ‘‘fatigue des matériaux et des structures’’ (editors: c. bathias and a. pineau), 2: 141-152. xu s, wu xq, han eh, ke w, katada y (2008), crack initiation mechanisms for lowcycle fatigue of type 316ti stainless steel in high temperature water. materials science and engineering 490: 16-25. the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 an assessment of the effect of construction errors during the implementation of rc t-beams a.g. asran a , m.a. eliwa a , m.a. alkersh a , and e.s.a. bayoumi *b,c a department of civil engineering, faculty of engineering, al-azhar university, egypt. b department of civil engineering, engineering college, qassim university, ksa. c engineering expert at ministry of justice, egypt. abstract: this paper investigates the effects of construction errors during the implementation of reinforced concrete t-beams. these errors are classified into two main sections. the first focuses on the position and ratio of reinforcing bars, while the other is related to the concrete strength. a total of ten specimens of t-beams were tested to assess the effect of the possible defects in the construction sites, viz. impact of misplacement of slab reinforcement, irregular arrangement of slab reinforcement, the change in bar diameter of slab reinforcement and the effect of casting method of concrete on the structural behavior of t-beam sections. the results indicated that the faulty placement of slab reinforcement leads to a lower bending moment capacity of the slab (brittle behavior) and the steel strain of slab decreases as the height of slab reinforcement decreases. the irregularity of the reinforcing bars in concrete slab affects the ultimate load carrying capacity of the slab. also, it was found out that well-arranged distribution of reinforcement improves the ductile behavior of the slab and reduces the corresponding deflections. keywords: construction errors; t-beams; faulty placement; irregular arrangement; bar diameter. t-beamالخرسانيةثناء تنفيذ الكمرات أتقييم تأثير االخطاء االنشائية أحمد جمعة عسران أ ، دمحم عبدالعظيم عليوة أ ، دمحم عبدالصمد القرش أ ، السعيد عبدهللا بيومي *ت، ب االخطاء إلى ه، حيث يتم تصنيف هذt-beamاثار االخطاء االنشائية اثناء تنفيذ الكمرات الخرسانية فحصيهدف هذا البحث إلي :الملخص عشر عينات من تم اختبار فقد التسليح ، بينما يرتبط اآلخر بمقاومة الخرسانة. سياخقسمين رئيسيين. يركز األول منها على موضع ونسبة أ t-beam وح البالطة الدراسة على تتبع تأثير الوضع الخاطئ لحديد تسلي هتقييم تأثير العيوب المحتملة في مواقع االنشاء ، حيث شملت هذل تأثير التغيير وسواء كان الحديد متالصق أو متباعد )سوء توزيع فى رص حديد تسليح البالطة( , منتظم لحديد تسليح البالطة الأثر الترتيب غير لخلل فى أشارت النتائج إلى أن او قد.t-beam sectionأثر طريقة صب خرسانة البالطة على السلوك االنشائي وفى قطر حديد البالطة ، أثر على قيمة حمل الكسر وإنفعال الخضوع لهذة العينات بالمقارنة كما ، ألحمالالبالطة لتحمل كفاءة تقليل وضع حديد تسليح البالطة أدى الى كما وجد ان التوزيع مع العينة المثالية. كذلك كان سوء توزيع اسياخ حديد التسليح فى البالطة مؤثراً على قدرة تحمل البالطة لألحمال المؤثرة ، المنتظم لحديد التسليح أدى الي تحسين تعزيز البالطة فى مقاومة األحمال. .قطر الحديد ؛ االنتظامعدم ؛الوضع الخاطئ ؛ تيكمرات ؛أخطاء األنشاء الكلمات المفتاحية: * corresponding author’s e-mail: saidbay80@hotmail.com doi: 10.24200/tjer.vol16iss2pp103-114 mailto:saidbay80@hotmail.com the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 1. introduction reinforced concrete (rc) has continued to be the most widely used structural material for over a century all over the world. reinforced concrete buildings consist of several structural components. the basic components of a reinforced concrete building are slabs and roof systems, beams, column, walls and foundations. these structural components can be classified into horizontal components (slabs, roofs, and beams) and vertical components (columns and walls). reinforced concrete (rc) slab is a horizontal concrete plate which carries loads perpendicular to its plane. it transfers these loads to beams which further transfers it to columns. this load transfer generates mainly bending moment in the slab. both concrete slabs and beams act together in resisting the applied loads. the top portion of the beam that is merged with slab is called flange and the portion of the beam below the slab is called the web. to consider a slab and a beam as a t-section, it is necessary to ensure interaction between these elements through a solid connection. the connection between the slab and the beam must be capable of ensuring a proper resistance to longitudinal and transverse shear forces. t-section beams have the distinct advantages of easy construction and costs efficiency, hence these have been extensively used in rc structures (katarzyna et al. 2017; al-khaburi and amoudi 2018; zhang et al. 2018). a few researchers studied the defects that occur in the concrete structures and the cracking behavior in rc structures, especially in the concrete slabs. the defects can be classified into two main sections; the first section focuses on the defects that occur in reinforcing steel detailing and cracking in rc elements, while the other focuses on concrete strength. cracks are common in concrete construction, which affect the buildings artistic form and destroy the wall’s integrity. this affects the structural safety and even poses a serious threat to the durability of structure. cracks occur because of the low tensile strength of concrete. these cracks, however, have a significant influence on the structural performance of concrete elements including tensile and bending stiffness, energy absorption capacity, ductility, and corrosion resistance of reinforcement. cracks develop due to the deterioration of concrete or corrosion or reinforcement bars due to poor construction, poor instruction by the supervisor and lack of worker training or inappropriate selection of constituent material and by temperature and shrinkage effects (eccs203, 2007; young-jin kang 1987 and watstein and parson 1943; byung hwan 1987). these cracks can be classified into three groups: structural, nonstructural, and due to fire load. there are many reasons for the development of non-structural cracking in the plastic stage of freshly cast concrete. however, structural cracking influences load carrying capacity of the structure. the cracks significantly decrease the structural stability and safety. these cracks may even lead to possible failure of structures during and after the construction. some possible causes of structural cracking are design failure, change of serviceability, an increase of design load, the poor quality material used, poor construction technology and impact loads. the cracks due to fire load can be considered as both structural and non-structural (nurul and mydin 2014; duinkherjav, bayar 2011, and aci committee report 1993). there are several parameters which also affect cracking in reinforced concrete structures, including the properties of concrete constituents, concrete cover, the diameter of main steel and its ratio, distribution of reinforcement and the characteristics of applied loads. cogurcu (2015) and peansupapa and rothmony (2015) investigated the construction and design defects in the residential buildings. there are three causes for failure of residential buildings that can be identified as: (i) faulty design, (ii) construction stage errors (poor workmanship, low strength of materials, inadequate transverse reinforcement–stirrup usage, defective and inadequate interlocking length, lack of control), (iii) errors in usage. experimental investigations were performed to investigate the behavior of reinforced concrete t-beams with different types of concrete in web and flange and effect of flange geometry by amr h. zaher et al. (2015); r. al-mahaidi et al. 2011 and ghailan 2010. it has been observed that t-beams play an important role in slabs and this technique of construction is widely used and suitable for residential, commercial, prefabrication and industrial buildings especially for large spans. a web reinforced shear critical reinforced concrete t-beam subjected to a concentrated point load will fail by one of two mechanisms. the first is a beam shear mechanism in which a diagonal tension crack continues from the web and penetrates into the flange. while the second is a punching shear mechanism wherein the applied load punches through the flange. an increase in the ratio of flange width to web width is shown to produce an accompanying increase in the ultimate strength of a reinforced concrete t-beam. this increase in shear resistance with an increase in the ratio of flange width to web width continues until the flange is wide enough to allow the formation of a failure mechanism whereby the load point punches through the flange. the slab contributed to increase in the shear resistance of the tbeams, where the shear failure loads increased by 42% for rectangular section in the t-beams without stirrups, by 43% in the t-beams with ordinary web stirrups and by 54% in the t-beams with flange stirrups. the shear resistance increases with the increase in slab thickness. when the ratio of slab thickness to beam thickness increases from 13% to 27% the shear failure loads increase by about 45%. human errors are predominant and are one of the many factors responsible for the failure of structures. human errors arise in the form of lapses and shortcomings that remain unnoticed during the design 104 http://www.engineersdaily.com/2014/04/concrete-floor-finishes.html a.g. asran, m.a. eliwa, m.a. alkersh, and e.s.a. bayoumi and construction of structures. it is obvious that there are very few research studies that investigate the human errors and implementation defects that occur during the execution of concrete structures. these defects are in the displacement of the reinforcing steel in the appropriate position, the irregular distribution of reinforcing bars in the concrete slabs or beams, casting method and increased water/cement ratio in the concrete mix. these defects are quite common, especially in egypt. these defects need to be investigated more deeply to determine precisely the effect of these parameters on the structural behavior of concrete slabs and connected beams (hong and he 2015; stewart 1993 and kaminetzky 1991). al.khuzaie and atea (2018) studied a hollow reinforced concrete t-beam under the effect of pure tension experimentally, which are made of reactive powder concrete (rpc). they concluded that the addition of 2% steel fibers to concrete mix increased the cracking and ultimate torque of the rpc hollow tbeam. an increase of 184% in cracking torque and 66% in ultimate torque for hollow section was achieved while the other properties were kept constant, while alexandra et al. (2018) concluded that the most common error in the mixing process is the balling effect that can occur in the production process of self-compacting concrete. it was found out that the mix design and the production process are more sensitive to the influence of the concrete constituents compared to that for ordinary concrete. the present study investigates the effect of position and ratio of negative reinforcement and change in compressive strength of concrete on the behavior of concrete slabs and connected beams. this research aims at introducing a rational evaluation of the common shortcomings in the implementation of reinforced concrete slabs under flexural stresses. a total of ten specimens were tested. these specimens were divided into four groups. the first group deals with the impact of misplacement of slab reinforcement, the second group investigates the effect of the irregular arrangement of slab reinforcement, the third one examines the effect of a change in bar diameter of slab reinforcement while the last discusses the effect of casting method of concrete on the efficiency of t-beam sections. 2. experimental investigation the experimental program involved testing nine fullscale reinforced concrete t-beam sections fabricated with defects in erecting practices were studied, in addition to the control specimen, which mainly fails under flexural stresses. the shape of the specimen consisted of beam (web) and slab (flange) that was cast monolithically. thus, slab and beam act together in resisting the applied loads of specimen dimensions were flange width is equal to 950 mm, flange thickness is equal to 100 mm, while beam depth is equal to 300mm and web thickness is equal to 150 mm. longitudinal and cross-sectional dimensions of the specimens are shown in the fig. 1. reinforcement in the t-beam comprised of normalgrade bars (mild-steel) 240/370 while high-tensile steel 360/520 was used in main and secondary steel of the projected beam. for all specimens; the amount of main longitudinal bottom and top reinforcement were kept constant. the vertical stirrups were also constant. the main longitudinal bottom reinforcement or the secondary reinforcement involves three bars of 10mm diameter high tensile steel. the vertical stirrups reinforcement was 8mm diameter mild steel bar spaced at 200mm acting as transverse reinforcement. slab reinforcement was changed from specimen to specimen according to the type of group. all specimens are 2 m long and loaded by the uniformly distributed load (line load) at the end of the slab from two sides. figure 2 displays the set-up of the t-beam section in the experimental program. the first group consists of four specimens, gιm-1 (control specimen), gιm-2, gιm-3, and gιm-4. the variable used for all specimens is slab depth reinforcement. the misplacement ratio of slab reinforcement (tmis./ts) is varied as 0.2, 0.4, 0.6 and 0.8, percent respectively where tmis. is slab misplacement reinforcement while ts is slab thickness. control specimen (gιm-1) was made with standard requirements of good compaction using a mechanical vibrator, enough concrete cover, well-arranged reinforcement. no splices in slab beam reinforcement were used in this control specimen. all specimens were constructed in the laboratory at the faculty of engineering, al-azhar university of egypt. figure 1. longitudinal and cross-sectional dimensions the tested specimens (units: mm). 105 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 the second group contains two specimens, in addition to the control specimen. in this group, eccentricity of the main steel in slab was the major parameter. the area of steel for slab was not changed (13ø8 mm on the length of the slab), but the distribution of steel was varied for two specimens (unequal distribution of slab reinforcement). in the first specimen (giia-1), irregular arrangement of slab reinforcement was used with three reinforcement bars at the mid-span of the slab at 50 mm while two bars were placed at a distance of 260 mm from two sides from the previous three bars keeping the distance between these bars as 50 mm. moreover, there are two bars from two sides at distance from the end of the slab. at the end 260 mm of the specimen one bar was erected. figure 3 illustrates the plan of reinforcement distribution of slab specimen (gιιa-1). in another specimen (gιιa-2), the eccentricity of slab reinforcement was in three groups, every group comprises three bars, and there are two distances between the three bars. this distance was 100 mm while the distance between the groups was 290 mm. at the end of the specimen, two bars were erected from two sides, and the distance was 100 mm. plan of slab reinforcement distribution for specimen (gιιa-2) is shown in fig. 4. the third group comprises three specimens (giiid-1), (giiid-2), in addition to the control specimen gim-1, and identifies the change in bar diameter of slab reinforcement. the ratio of steel was not changed, but the diameter only was changed. the first specimen was 6mm diameter whereas the 10mm diameter was used in the second specimen. in the first specimen (giiid-1), the steel used was 11ø6/m while the second specimen (giiid-2); it was 4ø10/m. in last group, the method of casting has been studied to simulate the method of buildings construction in egyptian society. the first specimen slab givc-1 was cast on two layers, the thickness of each layer was equal to 50mm and the time interval between castings of the two layers was equal to 20 minutes. the casting cart was moved during this interval on the concrete of slab. while the second specimen slab givc-2 was cast in three layers, the thickness of the first and second layer was equal to 30mm and the third layer was equal to 40mm. the interval between the first, second and third layers was 20 minutes, respectively. during the interval, the casting cast was moved to each layer. this method happens in some of the execution sites in egypt, especially in some places where there are no ready-mix concrete plants. details of all tested specimens are shown in table 1. 3. materials properties the cement was procured from torah cement factory which complied with the requirements of the egyptian standard specifications (es 4756/1-2007) for portland cement. crushed stone (coarse aggregate) and fine aggregate (sand) are used in the experimental program. water used in all mixes was clean drinking fresh water, free from impurities. the trial of mixes was carried out until the required workability was achieved. a water/cement ratio = 0.50 was finally used. mix proportions: the concrete mix used in all specimens was designed according to the egyptian code of practice. the concrete mix was designed to obtain target strength of 25 n/mm 2 at the age of 28 days for all specimens. the mix proportions by weight (kg/m 3 ) are presented in table 2. figure 2. set-up of the t-beam section in the experimental program. figure 3. plan of slab reinforcement distribution for specimen (gιιa-1). to determine the compressive strength of concrete after 7 and 28 days from casting, eighteen standard cube tests (150×150×150) mm 3 had been made; nine 106 a.g. asran, m.a. eliwa, m.a. alkersh, and e.s.a. bayoumi concrete cubes were tested after 7 days while the remaining concrete cubes were tested after 28 days. the cubes were filled with concrete in three layers while tamping each layer with a steel rod for twentyfive times according to the egyptian code of practice. the average compressive strengths of concrete cubes measured at 7 days was 23.44 mpa while at 28 days, the average compressive strength was 29.76 mpa. 4. instrumentations and loading system all the tested tbeam sections were simply supported with a span of 1800 mm and in a vertical position under concentrated vertical load from hydraulic jacks on two spreader beams which have a uniform load (line load) on two ends of the slab up to failure. the load was applied through spreader beam on the specimen using 100kn capacity hydraulic jacks. two linear variable displacement transducers (lvdt's) were installed vertically at the middle of the beam to measure the vertical deflection at midpoint of the flange at the end of the slab. table 1. summary of the different specimens. table 2. concrete mix design. cracks were detected through visual observation during testing all specimens, as well as marking the propagation of cracks at each load increment. the cracking and ultimate loads were accurately recorded during each test. the test specimens were loaded to failure and load-deflection response, modes of failure, and the strain values in the main lab and beam reinforcement were observed and recorded. 5. experimental results of tbeam section and discussions the measured deflection curves were plotted against the corresponding applied loads, from the start of applying the load, and up to failure. cracking load pcr, ultimate load pu, ultimate deflection at failure ∆f and toughness for all tested specimens are shown in table 3. 5.1 load-deflection analysis at the end of slab figure 5 shows the relationship between load and deflection of the tested specimens of group i. it was observed that the experimental specimens manifested the linear deflection behavior before cracking. upon cracking, specimen stiffness decreases as the load is increased. for specimens gim-2, gim-3, and gim-4, lower values of the ultimate load and deflection were observed compared with the load-deflection curve up to failure is called toughness. toughness is the ability of the material to withstand or absorb mechanical energy as shown in table 3. it can be concluded from comparison of group i that the rate of increase for the ratio of slab reinforcement figure 4. plan of slab reinforcement distribution for specimen (gιιa-2). group specimen notation notes control specimen gim-1 control specimen (tmis./ts) = 20% misplacement of slab reinforcement (gim) gim-2 (tmis./ts) = 40% gim-3 (tmis./ts) = 60% gim-4 (tmis./ts) = 80% effect of irregular arrangement of slab reinforcement (giia) giia-1 ـــــــــــ giia-2 ـــــــــــ effect of change in bar diameter of slab reinforcement (giiid) giiid-1 slab rft = 11ø6/m giiid-2 slab rft = 4ø10/m effect of change in casting method of concrete (givc) givc-1 casting on two layers givc-2 casting on three layers constituents mix proportions by weight for 1 m 3 crushed stone 1256kg graded sand 628kg water 150liters water/cement ratio 0.50 107 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 misplacement was less than the ultimate load value of the tested specimen, but the rate of decrease in the ultimate load was not the same as the rate of increase in the percentage of the wrong placement in slab reinforcement. a lower position of reinforcement leads to a lower bending moment capacity of the slab and can also lead to a brittle behavior in case of collapse. it can be concluded from the results of group ii as shown in fig. 6 that well-arranged distribution of reinforcement improves the ductile behavior of the slab and reduces the corresponding deflections. meanwhile, the eccentricity of main steel creates a sort of non-uniform stress distribution over the section and accelerates the failure. figure 7 shows the results of group iii. it can be seen that the use of 10 mm diameter in the reinforcement of the slab exhibited high resistance to loads while, on the contrary, the 6 mm diameter reinforcement offered a weak resistance to the loads affecting the slab. load-deflection responses for specimens gim-1 and giiid-2 showed approximately the same trend, and no significant difference was observed at low loading level, while the third specimen giiid-1 exhibited a significant difference in deflection values from the beginning of loading. it is concluded from these results that an increase in the diameter of slab reinforcement while keeping reinforcement ratio constant, enhanced the behavior of t-beam to withstand the loads and increased the ductility of the t-beams. it also improves the efficiency of t-beam section under loading effect. the minimum bar diameter for slab reinforcement is 8 mm because the 6 mm diameter reinforcement was found to be weak in resisting the loads. load versus deflection graph of specimens for group iv is shown in fig. 8. the load-deflection relationship was a nearly bilinear response up to failure. the specimens had approximately a similar loading up to the initiation of the first crack, followed by a reduction in the stiffness for all slabs but with different tendencies due to the cracking of concrete. after the cracking stage, tested slabs stiffness was dependent on the main reinforcing bars of slab. control specimen gim-1 exhibited more post-cracking flexure loading and more deflection compared with the specimens givc-1 and givc-2 due to specimen givc-1 and givc-2 being cast on different layers. on the other hand, specimen givc-1 showed a higher value in the deflection than the specimen givc-2 because these specimen were cast on three layers. the toughness of control specimen was higher than that of the other specimens where specimen gim-1 exhibited a value of toughness 22.58% and38.60% higher than specimens givc-1 and givc-2, respectively. for group iv specimens, it was found that the casting of slab on different layers and non-use of vibrators in the compaction of concrete slab led to presence of honey combing in the slab, which affected slab compressive strength against loads. the honeycombed specimens givc-1 and givc-2 had relatively lower stiffness compared with control specimen gim1 and this could be due to the early initiation of cracks in between concrete pores. 5.2 load-deflection analysis at the mid-span of the beam figure 9 displays the applied load against the vertical deflection at mid-span of the reinforced concrete beams for group i. it is apparent that the shape of the load-deflection curves in the elastic region before cracking be the same for the all specimens. however, it appears that after cracking, both specimens gim-1 and gim-2 produced higher values of deflection than specimens gim-3 and gim-4 for the same level of loading. the maximum deflection of specimens gim-1, gim-2, gim-3, and gim-4 was 5.9mm, 6.7mm, 7.55mm, and 6.3mm, respectively at the failure load. it could be claimed that the effects from the wrong position of slab reinforcement are more serious on the behavior of slab and the attached beam than the correct place for reinforcing steel for the slab. table 3. summary of the experimental results for all tested specimens. 108 108 a.g. asran, m.a. eliwa, m.a. alkersh, and e.s.a. bayoumi figure 5. load-deflection relationships of the tested specimens of group i. figure 6. load-deflection relationships of the tested specimens of group ii. figure 7. load-deflection relationships of the tested specimens of group iii. figure 8. load-deflection relationships of the tested specimens of group iv. figure 9. load-deflection relationship at mid-span of the beam for specimens of group i. figure 10. load-deflection relationship at mid-span of the beam for specimens of group ii. 109 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 the load-deflection curves for specimens of group ii are shown in fig. 10. there is no significant difference between three specimens in the values of deflection, especially at the beginning of loading before the initiation of cracks. the maximum deflection for the specimens gim-1, giia-1 and giia2 at the failure load was 5.90 mm, 6.80 mm, and 5.75mm, respectively. it is evident that the irregularity of slab reinforcement does not have any significant effect on the efficiency of a concrete beam connected with the slab. furthermore, the beam attached to slab was not significantly affected by the irregularity of the slab reinforcement shape. figure 11 shows the comparison between specimens of group iii. it was found out that the use of 10mm diameter in reinforcing the slab in t-section significantly improved the flexural behavior of the slab to resist the load. thus, the behavior of the beam connected to the slab was improved to withstand the loads. also, there is no clear difference between the behavior of specimens gim-1 and giiid-2 with slab reinforcement with diameters 8mm and 10 mm in load-deflection values. the maximum deflection value for the beam at the failure was 5.9 mm, 8.75 mm and 10.25 mm for specimens gim-1, giiid-1, and giiid2, respectively. specimen giiid-2 demonstrated higher deflection than specimen gim-1 and giiid-1 where this specimen was reinforced with a diameter of 10 mm, showed the beam connected to the slab to have a high resistance to loads. therefore, it is preferable to use 10 mm diameter in reinforcing the slabs. figure 12 represents the applied load against the vertical deflection at mid-span of the reinforced concrete beams for group iv specimens. the deflection of beam for both specimens givc-1 and givc-2 which casted on layers produced greater values of deflection than specimen gim-1 for the same level of loading. the ultimate beam deflection for specimens gim-1, givc-1 and givc-2 was 5.9 mm, 5.75 mm, and 5.55 mm, respectively at the failure load. finally from these results, it could be claimed that the presence of construction defects in the slab, especially the harmful effect from the honey-combed concrete are more pronounced. the harmful effect of casting of slab on the layers led to the reinforcing bars of slab to resist the loading early. the performance of slab reinforcement was affected on the resistance of both slab and connected beam (t-section) to the loads. 5.3 load versus strain values in main reinforcement of slab the values of strain in group i (higher than the yield strain εy =2000micro-strain), at the maximum load was equal to 32kn (maximum load for specimen gim-1), the specimen gim-1 showed 5.88%, 28.43% and 43.13% higher strain than the gim-3, and gim-4, respectively. meanwhile, the results of group ii displayed that the specimen gim-1 demonstrated 11.17% and 9.80% higher strain than the specimen giia-1 and giia-2, respectively, whereas the test results of group iii indicated that; the specimen gim-1 exhibited 4.37% lower strain than the specimen giiid2 and 18.95% higher strain than the specimen giiid-1. examining the strain results in the last group, specimen gim-1 demonstrated 24.21% and 35.68% higher strain than the specimens givc-1 and givc-2, respectively. 5.4 load versus strain values in main reinforcement of beam for s first specimens group at the failure load of control specimen gim-1, the steel strain increases to 17.78%, 20.44% and 26.67% for the specimens gim2, gim-3, and gim-4, respectively. the maximum recorded strain of the longitudinal reinforcement of specimens gim-2, gim-3, and gim-4 did not reach the yield, whereas the strain values of control specimen gim-1 were higher than that of the yield figure 11. load-deflection relationship at mid-span of the beam for specimens of group iii. figure 12. load-deflection relationship at mid-span of the beam for specimens of group iv. 110 a.g. asran, m.a. eliwa, m.a. alkersh, and e.s.a. bayoumi strain value. furthermore in the second group, the steel strain value in the main steel of the beam increases in specimen giia-1 and giia-2 with 15.33% and 14.17%, respectively. for the third group, the specimen giiid-2 showed 5.11% higher strain than the specimen gim-1, and the specimen giiid-1 displayed 4.44% lower strain than the specimen gim1. moreover, strain value at failure load for specimens of the last group, the specimen gim-1 showed 6.97% and 11.78% higher than the strain values of specimens givc-1 and givc-2. 6. mode of failure first group specimens: for the control specimen gim-1, the first noticeable crack started in the borderline between slab (flange) and beam (web) at a load equal to 10 kn on both sides of the slab and this is classified as flexural failure. these cracks were inclined in the direction of the slab towards the loading effect, i.e. the crack initiated in the region of the maximum tensile stress. these cracks extended on the boundary between the slab and the beam on the overall length of the specimen. then, by increasing the loading value, the existing cracks grow wider and deeper until the failure occurred in the slab and then cracks became inclined towards tensile stresses trajectories. at later stages of loading, cracks concentrated on the section of the maximum bending moment, and the failure occurred in the slab at the load of 32kn in the connected region between the slab and beam as shown in fig. 13. for the remaining specimens of group i, it was observed that the first cracks visible to naked eyes were at one side of the border line between slab (flange) and beam (web) and spread along the length of the specimen. these cracks began at a load of 5kn, 4.5kn and 6kn for specimens gim-2, gim-3, and gim-4, respectively. in gim-4, the initial cracks were slightly delayed where they appeared on the left side due some defects in loading system. at the failure, the cracks width was noticeably wider and propagated at the face of the intersection between slab and beam where the height of reinforcement in slab is reduced. the failure occurred at a load equal to 25kn, 20.5kn and 18.5kn for gim-2, gim-3, and gim-4 specimens, respectively. based on test observations, the failure of gim-2, gim-3 and gim-4 specimens could be classified as pronounced, sudden and flexural mode of failure in the slab where the misplacement of slab reinforcement affected the efficiency of the slab, leading to a faster failure in the slab. figures 14, 15 and 16 illustrate the photo of gim-2, gim-3, and gim-4 specimens, respectively after the failure. second group specimens: the specimens giia-1 and giia-2 exhibited basically the same cracking pattern and final mode of failure in nature of loading. the failure of these specimens was flexural tensile at maximum bending moment region (the vicinity region between slab connection with the beam) where the moment is concentrated in this region. also, the cracks started to appear in the region where there are no reinforcing bars in the slab and propagated towards the loading points. the first crack took place at load 7.5 kn and 12kn for specimens giia-1 and giia-2, respectively. as the load was further increased, the crack became wider and extended at both sides of the beam on the overall length of the specimen. the specimens giia-1 and giia-2 failed primarily at the ultimate conditions in a flexural mode at loads equal to 28 kn and 30.5 kn, respectively. the failure occurred on the entire border line between the beam (web) and slab (flange) interface. the cracking patterns of tested specimens giia-1 and giia-2 at failure are shown in figs. 17 and 18, respectively. third group specimens: the crack patterns for giiid-1 and giiid-2 specimens are shown in figures 19 and 20, respectively. the first cracks were horizontal flexural cracks in the vicinity of the tension zone within and near the constant moment region at figure 13. mode of failure of specimen gim-1. figure 14. mode of failure of specimen gim-2. 111 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 the connection of slab (web) with a beam (flange) at a load of about 4.60 kn and 7 kn for specimens giiid-1 and giiid-2, respectively. these cracks continued on the overall length of the specimen (ie. parallel to the loading effect). at higher loading stages for specimen giiid-1, the rate of formation of new cracks significantly decreased. moreover, the existing cracks grow wider, especially the first formed cracks. the specimen failed at a lower loading value at 19 kn in the region of negative moment concentration affecting the t-beam section where tension reinforcement of the slab yielded, indicating that 6 mm diameter reinforcement was weak in load resistance. while for the second specimen giiid-2, with the increase in load, cracks appeared on the borderline between the beam and the slab. the width of the cracks increased with the increase in the loading effect till the specimen failed at a value of 38 kn. crushing of the concrete in one corner of the specimen occurred when load was increased. this specimen exhibited high resistance to the loads compared with specimens gim-1 and giiid1. figure 15. mode of failure of specimen gim-3. figure 16. mode of failure of specimen gim-4. figure 17. mode of failure of specimen giia-1. figure 18. mode of failure of specimen giia-2. figure 19. mode of failure of specimen giiid-1. 112 a.g. asran, m.a. eliwa, m.a. alkersh, and e.s.a. bayoumi fourth group specimens: the first crack was seen in specimen givc-1 in the slab on both sides of the beam (at the maximum bending moment region from slab) while for the second specimen givc-2, it occurred in the slab on the right side from the connected beam with slab. figure 20. mode of failure of specimen giiid-2. figure 21. mode of failure of specimen givc-1. figure 22. mode of failure of specimen givc-2. the first cracks were flexural crack and occurred at the tension side of the slab, visible to naked eyes at a load of 7 kn and 6 kn for specimens givc-1 and givc-2, respectively. as the load was increased, tensile cracks developed faster in concrete and propagated easily in between concrete pores. the presence of these pores in the honey combed concrete helped the propagation of cracks with more easiness. moreover, the width of existing cracks in the slab became wider and deeper with increase in the load on the specimens and affected the efficiency of the slab and finally caused failure in flexure. figures 21 and 22 illustrate the cracking of the tested specimens givc-1 and givc-2 at failure. for specimen givc-1, the failure occurred in the slab on both sides of the beam while the failure occurred in the slab on the right side from the connected beam with slab in the second specimen givc-2. at ultimate load, specimens were not capable of resisting any further load. on the other hand, a noticeable drop in the ultimate load of the honey combed specimens givc-1 and givc-2 was found compared to the control specimen gim-1. the ultimate load for specimens gim-1, givc-1 and givc-2 was 32kn, 27.6kn and 23.5kn, respectively. 7. conclusion the following conclusions can be drawn from the test results of the experimental investigation: 1. misplacement of slab reinforcement ratio, if increased from 20% to 40%, results in reduction of the ultimate load of 21.87%, while when this percentage reached 60%, the reduction equals 35.93%, and when the percentage increases in the place of misplacement in the reinforcement of slab is 80%, the reduction in the maximum load is 42.18%. 2. the toughness decreases with increase in the ratio of misplacement of reinforcing bars of slab as 71.52%, 60.11%, and 50.57% for misplacement ratio (tmis./ts) for reinforcing bars of slab equal to 0.40, 0.60 and 0.80, respectively. 3. the control specimen gim-1 recorded an increase of 4.68% and 12.5% in the ultimate load over specimens giia-1 and giia-2, respectively. 4. well-arranged distribution of reinforcement improves the ductile behavior of the slab and reduces the corresponding deflections. meanwhile, eccentricity of main steel creates a sort of non-uniform stress distribution over the section and accelerates the failure. 5. to improve the efficiency of t-beam section under the loading effect, the minimum bar diameter for the reinforcement of slab is 8 mm 113 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 103-114 because the 6 mm diameter reinforcement was found to be weak in offering resistance to the loads. 6. non-using of vibrators in the compaction during the casting of concrete slab led to the presence of honey combing in the slab. the load carrying capacity of specimen gim-1 was 32kn, which is 13.80% and 26.60% higher than the load carrying capacity of the specimens givc-1 and givc-2, respectively. 7. honeycombed concrete presented one of the most serious defects on the behavior of reinforced concrete t-sections as it caused a considerable decrease in both the ultimate load and toughness values. the toughness of control specimen was higher than the other specimens where specimen gim-1 exhibited a value of toughness 22.58% and 38.60% higher than specimens givc-1 and givc-2, respectively. 8. the harmful effect of slab casting in the layers led to the reinforcing bars of slab to resist the loading early. the honey-combed specimens givc-1 and givc-2 had relatively lower stiffness compared with control specimen gim-1 and this could be due to the early initiation of cracks in between pores of concrete. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references aci committee report 224.1r-93 (1993), causes, evaluation and repair of cracks in concrete structures (aci 224.1r-93). american concrete institute journal 1-22. al khuzaie h.m. atea r.s. (2018). investigation of torsional behavior and capacity of reactive powder concrete (rpc) of hollow t-beam. j. mater. res. technol. 1:1-9. alexandra c, bogdan h, camelia n and zoltan k (2018), mix design of self-compacting concrete with limestone filler versus fly ash addition. procedia manufacturing. 22: 301-8. al-khaburi sakina and amoudi omar (2018), analysis of accident causes at construction sites in oman. jordan journal of civil engineering. 12(22 : 279-294. amr h. zaher, wael m montaser, and ahmed k elshenawy (2015), shear behavior of light weight concrete t–beams. international journal of emerging technology and advanced engineering. 5(12): 12. byung hwan o.h. and young-jin kang (1987). new formulas for maximum crack width and crack spacing in reinforced concrete flexural members. aci journal 103-112. cogurcu m.t. (2015), construction and design defects in the residential buildings and observed earthquake damage types in turkey. nat. hazards earth syst. sci., 15: 931-945. dhia b. ghailan (2010), t-beam behavior in flexure with different layers of concrete in web and flange. kufa journal of engineering. 2: 53-61. duinkherjav y. javkhlan b (2011), the influence of concrete cover to protect reinforcing bar on load carrying capacity of floor slab. the twelfth east asia-pacific conference on structural engineering and construction, procedia engineering 14: 22542259. egyptian code of practice for design and construction of reinforced concrete structures (eccs203-2007). housing and building research center, giza, egypt. hong h.p. he, w. x. (2015), effect of human error on the reliability of roof panel under uplift wind pressure. structural safety, 52: 54-65. kaminetzky d (1991). design and construction failures: lessons from forensic investigations.: mcgraw-hill inc., new york, usa. katarzyna c, maciej s, and jacek ścigałło (2017), the numerical analysis of the effective flange width in t-section reinforced concrete beams. www.sciencedirect.com. procedia engineering 172: 178-185. nurul n. b. and mydin a.o. (2014), general building defects: causes, symptoms and remedial work. european journal of technology and design, 3: 417. riadh al-mahaidi, geoff taplin, and craig giaccio (2011), experimental study on the effect of flange geometry on the shear strength of reinforced concrete t-beams subjected to concentrated loads. canadian journal of civil engineering 29: 911-918. stewart m.g. (1993), modeling human performance in reinforced concrete beam construction. journal of construction engineering and management. 119: 622. vachara peansupapa and rothmony ly (2015), evaluating the impact level of design errors in structural and other building components in building construction projects in cambodia. procedia engineering 123: 370-378. watstein d, parson d.e. (1943), width and spacing of tensile cracks in axially reinforced concrete cylinders. journal of research national bureau of standards 31: 1-24. zhang yannian, xie jun and wang liu (2018), experimental study on rc t-section beams strengthened with bottom steel plates. jordan journal of civil engineering. 12(3): 502-515. 114 http://www.sciencedirect.com/ the journal of engineering research (tjer), vol. 15, no. 2 (2018) 163-174 doi: 10.24200/tjer.vol15iss2p163-174 electromagnetic heating effects in power distribution cables under different operating conditions m. eladawya,*, and i.a. metwallyb a electrical engineering department, faculty of engineering, mansoura university, mansoura 35516, egypt. b department of electrical and computer engineering, college of eng. sultan qaboos university, muscat 123, oman. received 8 april 2018; accepted 22 april 2018 abstract: this paper presents a finite element simulation by comsol multiphysics package to investigate the temperature distribution inside three-phase, three-core, 33 kv underground power cables (ugc) through a coupled electromagnetic-thermal modelling. the simulations are very controlled and fine realistic details can be added to the model such as the temperature conductivity dependence of any metallic layer and armour permeability. distributions of magnetic field, current density, resistive losses and temperature inside ugc different layers are calculated at different operating conditions. the exponential increase in conductor temperature with increasing the conductor current limits the single-phasing operation of such cables. therefore, they must be derated, otherwise their lifetime will be reduced exponentially. finally, the effect of current harmonics on the temperature distribution inside the insulation material and hence its lifetime is calculated using matlab. it is found that higher steady-state conductor temperatures are expected for cables with larger conductor cross-sectional areas, using aluminium core rather than copper, or using 6-pulse rectifiers rather than a higher pulse types. keywords: power cables; heating; electromagnatic modelling. التأثريات الكهرومغناطيسية احلرارية يف كابالت توزيع القدرة حتت ظروف التشغيل املختلفة ب، إبراهيم متوىلأ*حممد العدوى بشكل كبري حيث ميكن إضافة تفاصيل واقعية ك.ق. 33لكابالت القوى جهد احملاكاة يف هذا البحث متت: امللخص دقيقة إىل النموذج مثل اعتمادية التوصيلية الكهربية ألي جزء معدنى على درجة احلرارة وأيضا النفاذية املغناطيسية للصفائح فاقيد الناجتة املعدنية الواقية للكابل. و يتم احتساب كل من نسب توزيع اجملال املغناطيسي و كثافة التيار الكهربى و امل وقد . املختلفة التشغيل ظروف حتت( ugcعن املقاومة الكهربية ودرجة احلرارة داخل الطبقات املختلفة املكونة للكابل ) الوجة أحادي التشغيل عملية من حتد قد فيه الكهربى التيار لزيادة كنتيجة املوصل حرارة درجة يف األسية الزيادة أن وجد يتضاعف سوف وإال ، الكابل داخل به املسموح الكهربى التيار ختفيض عملية يف الشروع ينبغي لذلك،. الكابالت هلذه توزيع على الكهربى التيار توافقيات تأثري احتساب مت ، وأخرًيا. سريعة بصورة الكابالت لتلك األفرتاضى العمر نقصان ووقد. matlab برنامج باستخدام األفرتاضى العمر حساب وبالتالي للكابل املكونة العازلة املادة داخل احلرارة درجة للموصل العرضي املقطع مناطق ذات للكابالت الثابتة احلالة موصل حرارة درجات يف ارتفاع وجود املتوقع من أنه تبني الذبذبات ذات األنواع من بداًل ذبذبات 6 ذات حموالت باستخدام أو ، النحاس من بداًل األلومنيوم باستخدام وذلك ، األكرب بداًل األلومنيوم تستخدم والتى كبرية مقطع مساحة ذو موصالت على حتتوى التى الكابالت حالة فى أعلى تكون ل.األعلى .النبضات من أكرب عدد على تعتمد التى األنواع من بداًل نبضات 6 على تعتمد التى التيار موحدات حالة فى أو النحاس، من .كهرومغناطيسية منذجة تسخني، القوى، كابالت :املفتاحية الكلمات * corresponding author’s e-mail: madawy@mans.edu.eg https://scholar.google.com/citations?view_op=view_org&hl=en&org=6041225520940426446 m. eladawy and i.a. metwally 164 1. introduction significant growing demand of electrical power has been a great challenge in transmission and distribution systems over the years due to the progressive increase of power consumers. underground cables (ugcs) are widely used by utilities to transmit and distribute electric power at corresponding different voltage levels from generating stations to urban areas in order to provide an environment-friendly and reliable pathway for this power flow. consequently, this leads the utilities to evaluate the performance of ugcs considering their design, reliability, lifetime, manufacturing and installation costs. sufficiently accurate modelling is, in general, more difficult to obtain for ugcs than that for overhead transmission lines due to the complexity of geometrical configuration of ugcs, which leads to high sensitivity of cable parameters to errors. advanced diagnostic and simulation techniques can be used for accurate assessment of ugc insulation condition and consequently monitoring the probable reasons related to failure process of the insulation system (moore 1997). numerical modelling (huang et al. 2015) of ugcs provides, especially when calibrated with empirical results, an attractive tool to study the electromagnetic heating effect inside them, improve design, identify the dielectric response of the insulation system and determine temperature distribution through the different layers of the ugc in order to reduce manufacturing costs and improve design. however, the dielectric material of medium voltage cables is usually polymeric materials. one of the major problems, which shortens its lifetime, is the increase of its operating temperature than the maximum permissible one. however, this increase of temperature may arise from any imposed harmonics on the maximum allowable current carrying capacity (ampacity) or any surrounding heat source such as other neighbouring ugcs (anders et al. 1999). the simulation of the coupled electromagnetic-thermal model of ugcs can provide a simple tool to identify any increase in temperature due to the normal and abnormal operating conditions, which could be harmful and may shorten its lifetime over the long-term operation. accordingly, the purpose of this work is to provide a numerical model of electromagnetic heating effect in a three-phase medium-voltage ugc under different operating conditions through finite element simulation using comsol multiphysics package (2013). this model allows realistic details to be added and to study their effects on the temperature distribution inside different ugc layers moreover, different modes of operations can be studied such as unintended single-phasing operation of a three-phase ugc. firstly, these calculations are considered for pure sinusoidal current waveforms. thereafter, the investigations are followed by considering the effect of harmonic contents on the temperature rise and consequently the lifetime of ugcs. 2. transient electromagnetic heating modelling for a relatively uniform surrounding environment and when the cable length is much larger than its overall diameter, a twodimensional model can be used for this simulation as shown in fig. 1. to study the electromagnetic heating characteristics in ugcs, comsol multiphysics package (2013) is used to build a model for a 33 kv, 240 mm2, threephase, screened three-core belted armoured ugc buried at 0.8 m, which represents the mainstream of installed ugcs (saog). in general, heat transfer mechanisms comprise conduction, convection and radiation. when the cables are directly buried, heat transfer mainly depends on the conduction mechanism due to soil thermal characteristics and burial depth, whereas convection due to air and solar radiation effect can be ignored for instance (moore 1997). figure 1. a typical 2d cross section of threephase screened three-core belted underground cable. electromagnetic heating effects in power distribution cables under different operating conditions 165 consequently, the time-dependent heat transfer equation has the following form:    tk t t ca,tq p      (1) q heat source (w/m) t conductor temperature (k) a magnetic vector potential  mass density (kg/m3) cp specific heat capacity at constant pressure (j/kg.k) k thermal conductivity (w/m.k) the heat source q generally incorporates any resistive heating and magnetic losses through the poynting’s theorem (kovetz 1990). this can be calculated from the definition of magnetic vector potential and the application of maxwellampere’s law taking into account the temperature dependence of electrical conductivity as follows:     100 tt1    (2) e 2 jaajj   (3)    aajj 12e   (4)     ** hbjre 2 1 ajjre 2 1 q   (5) where:  electrical conductivity (s/m) 0  reference resistivity (ω.m) α temperature coefficient of resistivity (k-1) 0 t reference temperature (293 k) j total current density (a/m2) e j current density due to an external source (a/m2)  angular frequency (rad/s)  permittivity (f/m)  permeability (h/m) re real operator of complex quantity moreover, for a fixed frequency and relative phase shift of 120 between phases, the resulting three-phase currents distribution depends on the conductivity and permeability of each material layer of ugc through both skin and proximity effects. boundary conditions are critically important with the equations describing the electromagnetic heating model. since the magnetic field can extend outside the ugc with copper armour tape, a relatively large domain of soil outside the cable has been modelled. however, the size of soil region was chosen by studying progressively larger region in order to reduce the error difference to be smaller than 104 between the two consecutive simulations when increasing the region size. the magnetic potential at the soil boundaries, which sets the tangential components of the magnetic potential is considered to be zero as follows (comsol 2013): 0an  (6) where n is the normal component vector. in this model, the ambient soil temperature at the ground level is set as 20c and the calorific flow continuity has been considered at the separation surface between any two different materials as follows (comsol 2013): n t k n t k 21      (7) where k1 and k2 are the thermal conductivities of the two adjacent media 1 and 2, respectively. in this model, the ambient soil temperature at the ground level is set as 20c and the calorific flow continuity has been considered at the separation surface between any two different materials. in order to increase the computation accuracy, a non-uniform mesh refinement process is a base for further analysis whereas the mesh of minimum size near the boundaries and hence higher gradients significantly influence the calculation accuracy. furthermore, the mesh of minimum size is subjected to a parametric study in order to reduce the error difference between two consecutive runs of simulation. however, this difference was smaller than 10-4% without any considerable impact on the calculated field or thermal distributions in the bulk of materials (huang et al. 2015). 3. modelled three-phase three-core xlpe insulated ugc table 1 summarizes the geometrical characteristics of a 33 kv, three-phase, threecopper conductor ugc shown in fig. 1. however, the insulation systems mainly comprise three parts; conductor insulation, belt m. eladawy and i.a. metwally 166 insulation and the filling material in the interstice between the cores. the conductor insulation is based on cross-linked polyethylene insulation (xlpe) while the filler and belt insulation is based on ethylene propylene rubber (epr). for the purpose of calculation, it is assumed that the conductor surrounding insulation materials are uniform and the conductor has a smoothed and circular cross section. this is a valid postulation since in the case of stranded conductor; a semiconductor shield is applied to regulate any concentration of stress that may result due to stranded wires. in ugc, semiconducting materials are widely used to mitigate the electrical stress at the interfaces between the conductor and insulation and between the insulation and metallic copper screen layer as shown in fig. 1. typically, ethylene vinyl acetate (eva) and ethylene ethyl acrylate copolymers (eea) are widely adopted as semiconducting screens (moore 1997). the metallic sheath in such cable (saog) consists of a double steel tape armour covered by an outer pvc jacket to protect against environmental hazard. moreover, the relative permeability of steel tape armour is of interest due to its effect on the magnetic field distribution and consequently its thermal performance. some of the physical parameters of the materials used for this calculation are summarized in table 2. soil characteristics, such as its type and moisture content, significantly control heat dissipation outside ugc surface. when ugc surface temperature exceeds its critical value, a boundary zone of soil surrounding the cable starts to dry out resulting in higher thermal resistivity than prior to drying out (anders 2005; gouda et al, 2010). this in turn increases the temperature of the cable sheath. however, table 1. geometrical characteristics of a 33kv, 240 mm2, three-phase, three-core copper conductor, xlpe insulated, steel armored cable (saog). quantity value unit conductor diameter 18.4 mm nominal insulation thickness 8.0 mm nominal separation sheath thickness 2.2 mm nominal steel tap thickness 0.8 mm nominal outer pvc sheath thickness 4.0 mm approximate cable overall diameter 86.0 mm dc resistance of conductor at 20c 0.0754 ω/km current rating for 0.8 m laying depth and thermal resistivity of soil= 1.5c m/w. 431 a table 2. physical constants of the materials used in ugc (moore 1997). material density (kg/m3) specific heat (j/(kg.k)) relative permittivity thermal conductivity (w/(m.k)) copper 8700 385 1 400 eva 925 1400 2.25 0.34 xlpe 930 1900 2.3 0.286 epr 860 2000 2.35 0.286 steel 7780 460 1 18 pvc 1760 900 2.9 0.2 soil 770 1500 4 0.667 electromagnetic heating effects in power distribution cables under different operating conditions 167 the rest of soil has uniform thermal resistivity outside the dry zone corresponding to its moisture content. for simplicity, the entire soil surrounding the cable is considered to have uniform thermal characteristics prior to drying out. for more details see (anders 2005; eladawy 2017). 4. results and discussion 4.1 signatures of magnetic field and resistive losses maxwell’s equations in frequency domain govern the electromagnetic field diffusion inside the cable and the surrounding soil because of current application to the conductors. with the specified magnitude of each carrying current and phase angles of 0 for upper, 120 for right and -120 for left conductors in the case of power frequency range (50 hz), the total current density is based on the source current and induced eddy current densities. as a reference case for simulation, a constant relative permeability of the steel tape armour is considered to be 500. figure 2a shows a surface plot of current density inside the metallic parts of the ugc. a non-uniform current density distribution is observed in the adjoining conductors due to both skin and proximity effects of the flowing core currents that induce eddy currents in all the metallic parts of the cable. however, the non-uniform current density inside the copper conductors (cores) and their metallic copper screens can be attributed to the current displacement phenomenon which diverges the currents inside the adjacent conductors. it can be observed that the maximum eddy current density takes place inside the three copper conductors followed by the closest individual copper sheath boundaries to the steel tape armour with 120 phase shift of each (metwally 2010). this constant phase shift has the effect of transferring constant power over each current cycle. figure 2b shows a surface plot of the resistive losses due to these current distributions with proportional arrow representation of the surrounding magnetic field intensity inside the cable and the surrounding soil. it is clear that the resistive losses follow the same trend of current distribution as observed in fig. 2a. however, its value depends on the current distribution and the specific resistance of the metallic parts and its variation with tempara (a) surface plot of current density (a/m2). (b) surface plot of resistive losses (w/m3) with proportional arrow direction of magnetic field intensity (a/m). figure 2. distribution of current density and corresponding resistive losses of the metallic parts in medium voltage ugc after 3000 h of simulation. (a) t = 10 h. (b) t = 3000 h. figure 3. surface plot and level contours of temperature distribution inside ugc at two times of simulation until steady-state thermal equilibrium. m. eladawy and i.a. metwally 174 ture. it is worth mentioning that the losses in the copper screen (sheath) of each core also depend on the bonding arrangement (moore 1997). this issue is not currently studied in this paper. 4.2 transient temperature evolution thermal behaviour of ugc has received a lot of attentions from numerous manufacturers and researchers in order to improve cable design and electric installation guidelines for safe ampacity calculation of conductor and reliable operation of the constructive insulation material. however, this behaviour is a robust function of loading current, burial depth, soil thermal characteristics, skin and proximity effects, sheath and armour losses, internal and external thermal resistances and current distribution inside the cable parts (metwally 2012; anders et al. 1999; iec 60287 1994, anders 2005). according to iec standard method for the thermal rating of ugc (iec 60287 1994), several approximations and simplifications of this function may lead to underestimation of more than 5% in the calculated maximum continuous conductor temperature (sedaghat et al. 2014). coupled finite element electromagneticthermal modelling offers an elegant method for temperature evolution through different parts of the ugc starting from current conservation principle until thermal equilibrium of heat transfer. hence, accurate estimation of thermal distribution at any point through the modelled domain. therefore, the simulation results are subjected to extensive comparison with the iec standard 60287 (1994) in order to ensure the validity of this modelling. figure 3 shows temperature evolution at two times inside the cable and the soil until thermal equilibrium is achieved at 3000 h of simulation for the reference case in this work. however, the approximately constant temperature of all conductors refers to constant transferred power throughout each current cycle. figure 4 illustrates the evolution of temperature profiles until steady-state temperature occurs at different layers in the cable for temperature dependent conductivity ( =  (t)) of the metallic parts compared with the iec standard calculations for dynamic rating of ugc summarized in anders (2005; and eladawy (2017). the imposed rms current value of 431 a generates a steady-state temperature of 84.4 c at the conductor surface at 3000 h from current application for a 100% load factor, which is lower than the maximum continuous temperature, (90c) (moore 1997). there is a decrease in conductor temperature by about 11% when considering constant conductivity ( = constant) of the metallic parts. however, the iec standard 60287 (1994) calculation shows a conductor steady-state temperature of 80.46c which is 4.6% lower than the calculated temperature with the finite element model as shown in fig. 4. therefore, the steady-state temperature severely depends on the applied current, temperature dependent alternating current resistance of the metallic parts, number of phases, thermal resistance of each constructive domain and loss factors of metallic sheaths. the transient part of temperature profile highly depends on the soil thermal diffusivity, burial depth and heat capacitance of constructive metallic parts of the cable (anders 2005). consequently, this complete signature of the coupled electromagnetic-thermal behaviour provides designers and manufactures with the ability to study multiple operating and environmental conditions, and estimate ugc lifetime. 4.3 relative permeability of armour layer magnetic field distribution and losses in armour material have attracted the interest of many researchers, e.g. anders (2005); and lucca (2016) since the relative permeability of armour material may disturb the emanated magnetic field distribution. however, the relative permeability has different values from different cables’ manufacturers depending on the type of the applied armour material (saog). figure 4. temperature evolution profiles for different points in the core and outer points of pvc jacket until steady-state after 3000 h of simulation compared with the iec standard calculation. 168 electromagnetic heating effects in power distribution cables under different operating conditions 169 generally, most cable manufacturers take the relative permeability of 300 (moore 1997; saog). the larger relative permeability of steel armour is, the lower magnetic field outside the cable (lucca 2016). in addition, the larger the relative permeability of steel armour, the larger is the induced current density due to high intensification of magnetic field inside the highly permeable material and consequently reducing it in the neighbouring regions. in reference case (r= 500), a maximum flux density of 0.22 t is observed in the steel armour near the closest individual copper sheaths due to this high-imposed relative permeability. the flux density on the conductor surface reaches a value of 0.02 t where its value has 2.07 t at the ground surface. it is worth mentioning that the latter value reaches up to 13.4 t when r = 1 for the armor layer. inside the touching domain of individual copper sheaths, the current density seems to be constant since the magnetic flux is the resultant of three balanced equidistance-varying fields. calculation of losses in armour is not straightforward which depends on its material, cable type and installation method. the relative permeability (r) of steel tape armour in fig. 1 is changed without making any modifications of geometrical dimensions (table 1) or physical parameters of constructive materials (table 2). figure 5 shows the effect of relative permeability of steel armour on the temperature rise at different points inside the cable at steady state after 3000 h of simulation. for a nonmagnetic steel armour (r= 1), the steady-state temperature at the conductor surface reaches 75.8c and continuously increases with the increase of relative permeability up to 500 by 11.35 %. it is clear that the entrapment of magnetic field inside the cable leads to a significant rise of temperature at different layers of the cable because of increasing current density distribution and consequently power losses inside the metallic parts of the ugc. 4.4 single-phasing operation when the three-phase ugc is connected to a three-phase induction motor, unintended fault removal of one phase leads to a single phase running of the machine with drawing a greater current than the winding rated one. this type of fault may be due to motor's open-winding as result of burned connection, open circuit fault of fused contactors in the transformer-motor connection, fuse blowing, …etc. in oil industry (metwally et al. 2008), it is common to keep running the electric submersible pumps (esp) even if one phase of the motor is damaged or one of the fuses is blown out to keep oil production to the last minute before a complete failure of the motor. these failures have a very high cost because electric submersible pump’s failures result in oil production losses or what is so-called “oil deferment” affecting revenue, additionally pulling out the motor from down-hole requires mobilization of a rig, which has a very high renting cost. however, phase removal represents the worst case of unbalanced current condition, which leads to a current increase by a factor of √3 in the remaining two phases of the ugc (metwally 2010). it is found that current unbalance gives low contribution of power losses, and hence it is ignored in this paper. irrespective of the overcurrent and thermal protection procedures, this single-phasing hazard may not be discovered until the burning of either the ugc or the running esp occurs. the traditional analytical method for ugc thermal rating (anders 2005; sedaghat 2014) may fail to predict the temperature level with the desired accuracy in the remaining phases of single phasing operation mode due to the complexity of thermal insulation resistance and parameter calculations. the coupled electromagnetic-thermal model offers a simple way of predicting this thermal performance due to single phasing of ugc by removing a current from one phase and changing the rated current of the other two phases. figure 5. effect of relative permeability of steel armour on the steady-state temperature rise inside different parts of the ugc. m. eladawy and i.a. metwally 174 however, the simulation results show that when increasing the rated current by a factor of √3 as a result of phase removal, it would result in high conductor temperature values which exceed the maximum permissible value. this is due to the exponential increase in conductor temperature with increasing the conductor current. hereby, the lower left conductor current of fig. 1 is relaxed and the remaining two conductors are loaded by 0.5√3𝐼𝑛 phase currents with 120 of phase shift. figure 6a shows the resistive losses distribution inside the metallic parts of the ugc single-phasing operation. the maximum observed magnetic flux density in the steel armour is 0.26 t and it is located on steel periphery containing the two energized cores. the maximum induced current density and consequently resistive losses in the copper sheaths of energized conductors are increased to 247% and 662% of their values at a half rated current operation, respectively. from fig. 6a, it can be seen that the energized conductors and their copper sheaths have these higher values of current density and resistive losses. fig. 6b shows a comparison of steady-state temperature between normal operations of three-phase ugc at a half rated current and a single-phasing operation. all the three conductors attain a value of 34.5c temperature at a steady state in case of normal operation with a half load current. in case of a single-phasing operation, the temperature of the de-energized conductor is 56c because of the induced current density by the proximity effect, while it reaches up to 60.1c at the surface of the upper core conductor, which is lower than the permissible temperature value for the other energized conductors. any increase of the operating current in a single-phasing case would result in overheating of the insulation material and hence abrupts failure of the ugc (metwally et al. 2008). 4.5 current harmonics current harmonic distortion can be generated by power electronic devices, which have been widely penetrated into industry. the preliminary effect of this disturbance is the introduction of additional power losses in ugc, and hence increases its operating cost and reduces its lifetime due to additional thermal stresses (rasoulpoor et al. 2017). harmonic filter seems to be a simple solution and its higher cost limits its application. therefore, this harmonic current perturbation must be accounted for the proper ampacity derating and accurate calculation of thermal loading of cable constructive insulation material (tofoli et al. 2006). in the above simulation, a pure sinusoidal current waveform is assumed flowing through the conductors. in case of harmonic perturbation current, however, the rms current irms of the nonsinusoidal waveform is calculated as follows (tofoli et al. 2006): 2 i1 1h 2 hrms thd1iii     (8) where i1 is fundamental current component, ih is the current of harmonic order h, and thdi is the current total harmonic distortion. increasing the harmonic order would result in conductor resistance increase, and consequently power loss. the effect of superimposed harmonics on the sinusoidal current can be accounted for (a) surface plot of resistive losses (w/m3) with proportional arrow direction of magnetic field intensity (a/m). (b) steady-state temperature at different layers inside the upper core and at the cable surface at half rated current for normal and single-phasing operations. figure 6. single-phasing operation of ugc at half rated current. 170 electromagnetic heating effects in power distribution cables under different operating conditions 169 resistance calculation as follows (tofoli et al. 2006):  psdcac kk1rr  (9) where, rac and rdc are the ac and dc resistances at the operating temperature in ω/km, and ks and kp are the skin and proximity effect, respectively. the skin effect parameter x can be calculated as the following (iec 60287 1994; tofoli et al. 2006): dcr f 01528.0x   (10) where, f is the frequency in hz,  is the magnetic permeability of the conductor. the skin and proximity effect parameters can be approximated as follows tofoli et al. (2006): 100x2 if, 63.769x9.1061 x500x343.83 x616.6x2.0 10k 2x if, 0166.0x2764.0 x447.1x24.3 x24.8x04.1 10k 23 45 3 se 23 45 3 se                                     (11)                        2 c se 2 c sepe s d 312.0 27.0k 18.1 s d kk (12) where, dc is the conductor diameter and s is the axial spacing between conductors. metallic sheath and armour losses in xlpe ugc are about 5% of the three-conductor losses and can be ignored in per-unit power loss calculations (gandhare et al. 2013). considering only odd harmonics, the power loss p in all conductors can be calculated as follows (gandhare et al. 2013):            3h )h(ac 2 h)1(ac 2 1 rhri3p (13) where, rac(1) and rac(h) are the ac resistances for the fundamental and harmonic frequencies, respectively, and hh is the percentage of harmonic load. different cable sizes and conductor materials, namely copper and aluminium, are studied in order to investigate the effect of current harmonics on the power losses inside the ugc. the rating current for different cable sizes and conductor materials, namely copper and aluminium, are studied in order to investigate the effect of current harmonics on the power losses inside the ugc. the rating current for each copper conductor cable is applied on the same corresponding size of aluminium conductor cable. figure 7 shows the variation of conductor resistance with the harmonic order for two different materials and for different cable cross-sectional areas. increasing the harmonic order would result in increasing the conductor resistance, which depends on the cable dimensions due to both skin and proximity effects as observed by (9). the harmonic spectrum of the three different types of pulse rectifiers, namely 6-, 12and 18pulse rectifiers with current total harmonic distortion thdi of 30.82%, 14.63% and 9.6% (wu et al. 2017), respectively, has been imposed through matlab program on the results of electromagnetic-thermal model to study the effect of current harmonics on the steady-state temperature rise of ugc. figure 8 shows the effect of the studied harmonic distortion on the steady-state temperature rise for the considered different cable configurations. it is worth mentioning that the greater the harmonic distortion, the larger is the number of harmonics. consequently, each current harmonic generates its own individual heating effect. it can be seen that this increase in temperature depends on the harmonics signature of each type of rectifier and the rated current for each cable type. increasing the 6pulse rectifier thdi to 43.6% would result in figure 7. effect of harmonic order on the ac resistance for different sizes and conductor materials of ugc. 171 m. eladawy and i.a. metwally 174 higher temperature as shown in fig. 8. the 6pulse rectifier gives the highest conductor temperature rise for all considered cable configurations. considering the pure sinusoidal waveform (thdi = 0) as the reference case of ugc (240 mm2), there is an increase of temperature by 10% and 19%, 2% and 1% for 6(thdi = 30.82% and 43.6%), 12and 18pulse rectifiers, respectively. in addition, the conductor temperature exceeds the maximum continuous conductor temperature in case of having 6-pulse rectifier. therefore, the operating current must be decreased in order to meet the thermal limit of ugc, which results in cable derating due to this current harmonics distortion. 4.6 cable lifetime ugc lifetime is one of utilities’ major aspects in real operating conditions (zhoun et al. 2017). it is mainly detected by the thermal stresses of the insulation material (zarchi et al. 2016). normally, cable lifetime tlife is not affected when the conductor temperature is lower than the maximum continuous conductor temperature. however, it exponentially decreases when the temperature rises above its maximum limit according to arrhenius model as follows (gandhare et al. 2013):          1 0 1a 0life tt k w exptt (14) where t0 is the reference maximum continuous conductor temperature (363.15 k), t is the accelerated ageing temperature in k, wa is the activation energy for thermal degradation of xlpe (wa=1.1 ev) (zarchi et al. 2016), k is boltzmann’s constant in ev/k and t0 is the cable lifetime at t=t0. figure 9 shows the expected per-unit lifetime of the ugc under different operating conditions when increasing the current rating and current total harmonic distortion. in the three-phase operation of the reference cable (240 mm2) and for pure sinewave, overloading exponentially decreases the lifetime. in the case of the single-phasing operation, the sharp rise of temperature rapidly degrades the cable above half of its current rating. in the case of the rated current for the 6-pulse rectifier (thdi = 30.82%), the cable lifetime is reduced to 43% and 49% of its reference lifetime for the 240 and 50 mm2 copper conductor cables, respectively. therefore, the cable derating factor is necessary depending on the thdi loading of the cable in order to meet the thermal limits of its constructive insulation material and consequently its lifetime. this issue will be discussed in the next section. 4.7 cable derating factor due to current harmonics cable derating factor (kd) can be calculated when the conductor temperature exceeds its maximum continuous value, which may be due to current harmonic distortion. in this study, for a known harmonic signature, kd is defined as the ratio of rms value of the current distorted by harmonic (irms,d) to the rms value of a pure sinusoidal current (irms,p) with fundamental frequency which produces the same losses as follows (tofoli et al. 2006): figure 8. effect of current total harmonic distortion of the steady-state conductor temperature rises for different sizes and ugc materials figure 9. expected per-unit lifetime of the cable under different operating conditions versus current rating and thdi. 172 electromagnetic heating effects in power distribution cables under different operating conditions 169     1h dc )h(ac2 h dc)1(ac p,rms d,rms d r r h r/r i i k (15) as long as the conductor temperature is lower than the maximum continuous temperature, there is no derating of cable’s ampacity (kd =1). for the case of reference cable (240 mm2) and 6-pulse rectifier with thdi = 43.6%, the calculated derating factor (kd = 0.9136) as a result of higher temperature than the maximum continuous one. figure 10 shows the effect of different current harmonic distortion of 6-pulse rectifier on the conductor temperature of 240 mm2 ugc with the corresponding derating factor. it can be noticed that exceeding maximum continuous conductor temperature would result in a higher derating process. 5. conclusion a coupled electromagnetic-thermal method based on finite-element modelling is introduced to investigate the behaviour of ugc under different operating conditions using comsol multiphysics software. distributions of magnetic field, current density, resistive losses and temperature inside different layers of the ugc are calculated. some realistic parameters are studied, such as temperature-varying conductivity and relative permeability of steel armour. a single-phasing operating condition is considered to be one of the severest unbalanced cases. it significantly raises the temperature of the cable insulation material even at half of its rated current. precautions should be considered in the case of this unintended operation mode owing to the rapid increase of magnetic field perturbation and excessive rise to temperature inside the cable, and hence shortening the cable lifetime. current harmonics distortion can also give rise to temperature because of high-frequency contents, which increases the resistance and consequently steady-state temperature. higher steady-state conductor temperatures are expected in cables with larger conductor crosssectional areas, using aluminium core rather than copper, or using 6-pulse rectifiers rather than higher pulse types. expected lifetime exponentially decreases with an increase above maximum continuous conductor temperature due to accelerated thermal aging. therefore, it is tremendously important to limit the order and the percentage contents of the load current harmonics to reduce the harmonic loss and the harmonic temperature rise. finally, cable derating process should be taken into consideration in order to meet the thermal limits of insulation material. figure 10. effect of current total harmonic distortion on the conductor temperature and the corresponding derating factor for the 240 mm2 ugc. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references anders gj, napieralski a, kulesza z (1999), calculation of the internal thermal resistance and ampacity of 3-core screened cables with fillers. ieee trans. power deliv, 14(3): 729734. anders gj (2005), rating of electric power cables in unfavorable thermal environment. (wiley-ieee press). comsol group ltd., stockholm, sweden (20133), comsol multiphysics software package version 4.4. eladawy m (2017), dynamic derating of threephase underground power cable under current harmonic distortion. 19th international middle east power system conference mepcon’19. 173 m. eladawy and i.a. metwally 174 gandhare wz, patil kd (2013), effects of harmonics on power loss in xlpe cables. energy and power engineering 5(4b): 12351239. gouda oe, amer gm, el dein az (2010), effect of the formation of the dry zone around underground power cables on their ratings. ieee trans. power deliv, 26(2): 972-978. huang z, pilgrim ja, lewin p, swingler s, gregory tg (2015), thermal modelling and analysis for offshore submarine high-voltage direct current cable crossings. iet gener. transm. distrib, 9(16): 2717-2723. international standard iec 60287 (19944), electric cables calculation of the current rating – part 1-1: current rating equations (100% load factor) and calculation of lossessection 1: general. 1994-12. kovetz a (1990), the principles of electromagnetic theory. cambridge university press, 1990. lucca g (2016), electromagnetic interference at power frequencies: shielding factor related to an urban environment. iet sci. meas. technol. 10(6): 614–620. metwally ia, gastli a (2008), correlation between eddy currents and corrosion in electric submersible pump systems. int. j. thermal sciences (ijts), 47(6): 800-810. metwally ia (2010) electrostatic and magnetic field analyses of 66-kv cross-linked polyethylene submarine power cable equipped with optical fiber sensors. electric power components and systems 38(4): 465–476. metwally ia (2012), the evolution of medium voltage power cables. ieee potentials, 31(3: 20-25. moore gf (1997), electric cables handbook. (blackwell science, 3rd edn.). oman cables industry (saog): medium voltage cables. oci/pbmvc/rev. 001/010410, available: www. omancables.com. rasoulpoor m, mirzaie m, mirimani sm (2017), effects of non-sinusoidal current on current division, ampacity and magnetic field of parallel power cables. iet sci. meas. technol, 11(5): 553–562. sedaghat a, de león f (2014), thermal analysis of power cables in free air: evaluation and improvement of the iec standard ampacity calculations. ieee trans. power deliv, 29(5): 2306-2314. tofoli fl, sanhueza smr, de oliveira a (2006), on the study of losses in cables and transformers in nonsinusoidal conditions. ieee trans. power deliv, 21(2): 971-978. wu b, narimani m (2017), high-power converters and ac drives. (wiley-ieee press, 2nd edn.). zarchi da, vahidi b (2016), optimal placement of underground cables to maximise total ampacity considering cable lifetime. iet gener. transm. distrib, 10(1): 263-269. zhoun c, yi h, dong x (2017), review of recent research towards power cable life cycle management. iet high voltage 2(3): 179 – 187. the journal of engineering research (tjer), vol. 14, no. 2 (2017) 191-205 doi: 10.24200/tjer.vol.14iss2pp191-205 newtonian and non-newtonian pulsatile flows through an artery with stenosis p. haldera, a. husaina, m. zunaidb and a. samad*,c a department of mechanical and industrial engineering, sultan qaboos university, muscat, pc-123, oman. b department of mechanical engineering, delhi technological university, delhi-110042, india. c department of ocean engineering, indian institute of technology madras, chennai-600036, india. received 14 march 2017; accepted 16 october 2017 abstract: the current study presents three-dimensional modeling and analysis of blood flow through artery stenosis under several variants of pulsatile flow to mimic the atherosclerosis artery disease. the study considered newtonian and non-newtonian blood flow models and solved the threedimensional laminar and unsteady navier-stokes (ns) equations for different inlet velocity profiles. the carreau yasuda model was used for hemodynamics modeling of fluid viscosity. a comparative analysis was carried out for newtonian and non-newtonian fluid models under several simple pulsatile, equivalent pulsatile and physiological velocity profiles. the non-newtonian fluid exhibited a higher centerline velocity. wall shear stresses, shear strain rates, velocity fields, and vortex distribution for newtonian and non-newtonian fluid flows show significant differences. a higher wall shear was noticed near the stenosis. the vortex formed near the stenosis shifts to upstream and downstream of the stenosis with the change of velocity pulse. the velocity profile deflates more for non-newtonian flows than the newtonian flows. keywords: non-newtonian fluid; blood flows; artery stenosis; physiological flows; pulsatile flows. ضيقمصاب بالشريان عرب ةنيوتونيالوغري ةنيوتونيتدفقات النبض ال *ج،، عبد الصمدبزنيد، حممد أحسني ال، أفضأباريش هالدر : قامت الدراسة احلالية بتقديم منذجة ثالثية األبعاد وحتليل لتدفق الدم عربشريان مصاب بضيق امللخص حتت عدة تدفقات نبض متغرية و ذلك حملاكاة مرض تصلب الشرايني. وأخذت الدراسة يف االعتبار وذلك مناذج تدفق الدم النيوتونية وغري النيوتونية وحلت الرقائق ثالثية األبعاد و معادلة نافيري ستوكس )نس( و يو ياسودا للزوجة السوائل من معادالت االشكال املختلفة للسرعة عند املدخل.وقد مت استخدام منوذج كار أجل منذجة ديناميكا الدم .و مت اجراء حتليل مقارن لنماذج السوائل النيوتونية وغري النيوتونية حتت عدة الفسيولوجيي. وقد أظهر السائل غري النيوتوني سرعة أعلى النبض نبضات بسيطة مضاهية الشكال سرعة ة يف إجهاد القص، ومعدالت القص و مسارات السرعة يف خط الوسط. كما تبني وجود اختالفات كبري وتوزيع كتلة الدوامة على السوائل النيوتونية وغري النيوتونية. و قد لوحظ وجود ارتفاع يف جدار القص بالقرب من منطقة الضيق و تنقل كتلة الدوامة املتكونة بالقرب من منطقة الضيق من املدخل اىل ة النبض واخنفاض يف درجة التدفقات غري النيوتونية بسرعة أكثر من التدفقات املخرج مع تغري يف سرع النيوتونية. تدفقات الدم. ضيق الشريان. التدفقات السوائل غري النيوتونية ،: التدفقات النابضة الكلمات املفتاحية الفسيولوجية * corresponding author’s e-mail: samad@iitm.ac.in p. halder, a. husain, m. zunaid and a. samad 192 1. introduction the primary function of the cardiovascular system is to transport nutrients and oxygen to different body parts through blood circulation in arteries. atherosclerosis or artery stenosis is an artery disease that causes artery constriction due to the decomposition of smooth muscle cells and to the presence of lipids in the artery that form plaque in the artery wall. these plaques grow gradually with time and block the flow path. the hemodynamic effects of artery stenosis are far reaching beyond the reduced artery section, and dictates normal and shear stress distribution and low-pressure zones near the constricted region. the atherosclerosis plaque forms in the high shear region; therefore, it necessitates the investigation of the dynamic effect of the constriction on blood flow before and after the formation of atherosclerotic lesions. the cardiovascular system of bifurcated tubes supplies nutrition to different organs of the body with the help of heart pump. by increasing or decreasing their cross-sectional area, the arteries that are far from the inert tubes adapt to changing pressure and blood flow conditions to accommodate the hemodynamic demands. it is important to study the blood flow in arteries with physiological conditions, where the blood flow demand changes under disease conditions. several numerical and experimental studies have been reported to comprehend the behavior of blood flow through arteries. mcdonald (1955) studied pressure oscillations generated flow through the femoral artery of a dog. young and tsai (1973) experimentally investigated the effect of axisymmetric and non-symmetric plastic models with different area ratios and constriction lengths. the test results showed that geometry has a strong influence on the flow characteristics, which includes size and shape of the stenosis. deshpande et al. (1976) carried out computational analysis for steady blood flow through an artery with axisymmetric-contoured constriction and presented comparison with experimental data. the reported results were well-matched with the existing experimental results (young and tsai 1973). the flow field in human vasculature is relatively complex, and the carotid sinus exerts a high impact on the hemodynamics at the bifurcation (bharadvaj et et al. 1982a; bharadvaj et al. 1982b). mazher and ekaterinaris (1988) modled three-dimensional blood flows as newtonian fluid flows incorporating secondary and separated flows under steady conditions. most of the above studies are based on steady flow analysis whereas the heart delivers a pulsating flow, which is unsteady in nature. later, several authors investigated pulsatile newtonian and non-newtonian flows in artery bifurcation. perktold et al. (1991) reported numerical results for newtonian and nonnewtonian fluid flow through a carotid artery bifurcation model. the result showed that the fluid variables in the sinus get affected by the angle between the internal carotid axis and the common carotid axis. perktold and rappitsch (1995) analyzed the effect of the distensible artery wall on the local flow field and determined the mechanical stresses on the artery wall for a non-newtonian inelastic fluid. further, ku et al. (1985) measured pulsatile flows in a scaled model of the human carotid bifurcation by a laser doppler velocimetry. tu et al. (1992) computationally investigated flow through a stenosis in a rigid circular tube based on a two-dimensional galerkin finite element method. the paper reported the efffects of constriction severity, stricture length and reynolds number on flow vortex and pressure distribution. the sharp peaks of wss (wall shear stress) were present near the throat of the stenosis and presusre recovered partly at downstream of the stenosis. deplano and siouffi (1999) experimentally and numerically investigated different types of constriction shapes and the two results were in good agreement. zendehbudi and moayeri (1999) indicated that the behaviour of the pulsatile flow is completely different from physological flow. razavi et al (2011) investigated the flow dynamics related to a common carotid artery with symmetric stenosis varying severity 30– 60% under newtonian and non-newtonian viscosity models. furthermore, chakravarty and mandal (1994) developed a mathematical model for blood flow in a stenosed arterial segment using a pulsatile pressure gradient method, and showed that stenosis affects the wss distribution. tu and deville (1996) used herschel-bulkley, bingham, and power-law models of non-newtonian fluids to analyze the flow through arterial stenosis and discussed pressure distribution, flow pattern, and wss distribution. liao et al. (2004) investigated a bellshaped stenosis for three different pulsatile flows and concluded that the fluctuations of flow parameters affect the flow field for the newtonian and non-newtonian pulsatile flows through an artery with stenosis 193 physiological flow. srivastava and rastogi (2010) investigated an axially nonsymmetrical stenosis blood flow through a narrow catheterized artery. the results showed that the impedance enhances with the catheter, the hematocrit and the stenosis size decreases with the shape of parameter. nadeem et al. (2011) investigated tapered artery with stenosis using power law fluid model and examined velocity, impedance (flow resistance), wss and stress at the stenosis throat. ellahi et al. (2014) studied the mathematical models of jeffrey fluid with nanoparticles in the tapered atherosclerotic arteries. the authors solved mathematical model for mild stenosis. mekheimer and el kot (2015) investigated the axisymmetric flow of blood through curved coaxial tubes. a significant difference was noticed between the flows through curvature and non-curvature annulus flows through catheterized stenosed artery for the axial velocities, the stream function, the impedance, pressure drop and the wss distribution. due to the complex nature of fluid and flow modeleing, most of the above studies presented conservative two-dimesional analysis. further, research was carried out to investigate threedimensional flow dynamics of artery blood flows. ang and mazumdar (1997) developed a mathematical model for blood flow through an asymmetric arterial stenosis. a higher shear stress was noticed on the wall with stenosis as compared to the other side with no protuberance for the flow with reynolds number ranging from 100 to 1000 and for four degrees of stenosis severity. long et al. (2001) investigated pulsatile flow through asymmetrical and axisymmetrical stenosis models with 25%, 50% and 75% constrictions and reported that the stenosis influence length is shorter for asymmetrical models. hye and paul (2015) studied the spiral and non-spiral blood flows with axisymmetric arterial stenosis by using two-equation standard k-ω model and large eddy simulation (les). the results showed that the spiral flow disturbs the turbulence kinetic energy in the post stenosis region. chiastra et al. (2017) studied the coronary bifurcation anatomical features on both near-wall and intravascular flow features for a healthy and diseased artery. the bifurcation angle effects are small on hemodynamics in both cases of artery with or without stenosis. a lot of research has been carried out to conservative modeling of artery bifurcation and stenosis, there is a lack of deterministic modeling of flow through diseased artery with realistic fluid model, flow conditions, physiological pulse and elastic artery wall. the present research is part of comprehensive numerical analysis carried out to investigate flow through stenosed artery with realistic flow modeling and physiological conditions. in this study, hemodynamics of newtonian and nonnewtonian fluids in an artery with a symmetric stenosis for simple and equivalent pulsatile and physiological blood flows are studied through three-dimensional numerical modelling assuming rigid artery walls. the comparative analysis is presented for newtonian and nonnewtonian flows under variants of pulsatile flows. the several flow characteristics have been investigated for velocity, vortex, wall shear stress and shear strain rates to demonstrate blood flow dynamics at different intervals of flow cycle under different flow conditions. 2. mathematical model and numerical method an artery with a symmetrical stenosis was taken for numerical analyses as shown in fig. 1. the cylindrical coordinate system is considered with variable coordinates (r, z). the stenosis profile can be defined mathematically as follows (liao et al. 2004):          0z z cos1 0r2 1 0r r  00 zzz  (1) where, r is the artery radius with the function of z in the stenosis region, l is the length of the artery and r0 is the radius of the artery. in this case, δ = 0.5 and z0 = 1.0 (liao et al. 2004). figure 1 represents the computational domain for numerical analysis. the cad model was setup in ansys design modular. the mesh generation tool ansys mesh was employed to build the structured mesh for the computational domain. the number of mesh elements are increased gradually from 141102 to 840164 to check the grid dependence of the solution. the results showed that there is no a significant difference in velocity profiles for the change of mesh elements from 325017 to 840164. therefore, a grid with about 600000 elements is considered as optimum one (fig. 2). no slip boundary condition was applied for the artery p. halder, a. husain, m. zunaid and a. samad 194 wall whereas time dependent velocity was assigned at the inlet boundary. the outlet was assigned as an opening type boundary condition with pressure outlet. the static pressure at the outlet was kept as zero atm, while the reference pressure was one atm. the governing equations for the model are as follows: the continuity equation yields 0).(    u t   (2) and momentum equations     .. )(    puu t u (3) where the stress tensor, τ is related to the strain rate by 2 ( ) . 3 t u u u              (4) figure 2. grid test. 0 0.2 0.4 0.6 0.8 1 -0.5 0 0.5 1 1.5 2 2.5 3 r (m m ) u (m/s) element: 1,41,102 element: 3,25,017 element: 6,89,768 element: 8,40,164 figure 1. (a) schematic diagram and computational domain, and (b) grid distribution. (b) (a) r o r z l newtonian and non-newtonian pulsatile flows through an artery with stenosis 195 in the above equations, ρ, u, p and t represent density, velocity, pressure and time, respectively. a finite volume based solver ansys-cfx (ansys cfx 2010) was used to solve the unsteady three-dimensional incompressible navier-stokes (ns) equations to model the flow in artery. it is a coupled pressure based solver which solves velocity and pressure equations as a single system (ansys cfx 2010) and uses simplec (semi-implicit method for pressure linked equations-consistent) algorithm for pressure correction (van doormaal and raithby 1984). the equations were discretized by the cell-centered finite volume approach. high resolution and first order discretization schemes were implemented in the computations. the discretized system of equations was solved using a multigrid accelerated ilu (incomplete lower-upper) factorization procedure at any given time step and the solution was marched in time through fully implicit procedure. for each time step, the spatial field was solved through several iterations to reach to the conversion criteria. the required root mean square (rms) convergence criteria for the momentum and continuity equations were 1×10-9. blood density depends on age and sex and it is varied within the range 1030–1070 kg/m3 (jozwik and obidowski 2008). for conservative analysis, the density of the blood was assumed constant as 1058 kg/m3 for both newtonian and non-newtonian fluids. the flow analysis was done for both newtonian and non-newtonian fluids. the dynamic viscosity for the newtonian fluid was kept constant (0.004 pa.s) whereas for the non-newtonian fluid the carreau-yasuda model was implemented which is expressed as:     a n a o       1 )(1     (5) here, low shear viscosity, µo = 0.056 pa.s; high shear viscosity, µ∞ = 0.00345 pa.s; time constant, λ = 3.3313 s; power law index, n = 0.3568; yasuda exponent, a = 2 (ahsaas and tiwari 2016a). the carreau-yasuda model is continuous for a wide range of flow shear rates (�̇� ≥ 0). for high shear thinning the fluid viscosity approaches to 𝜇∞, whereas for low shear thinning it approaches to 𝜇0 which mimics well blood flow modeling (razavi et al. 2011; ahsaas and tiwari 2016; gijsen et al. 1999). 3. results and discussion in this work, three common velocity profiles, i.e. simple pulsatile, equivalent pulsatile and physiological flows, were used to describe the arterial flows (mcdonald 1955). the cyclic variations of the velocities are given in fig. 3. the simple pulsatile and the equivalent pulsatile flows are expressed as: simple pulsatile flow: ( ) 1 sin( *)u t t  (6) equivalent pulsatile flow: 1 (1 cos( * ))u t     (7) where t* is the non-dimensional time (2𝜋𝑡/𝑇), t is the time period, α is the amplitude and ϕ is the phase angle. the variation of velocity for simple pulsatile flow is symmetric about zero velocity over the time, whereas equivalent velocity profile has positive velocity for the larger part of the cycle. the numerical model was validated and the velocity profiles were compared with the ones reported in literature by deshpande et al. (1976) and liao et al. (2004) for fully developed laminar flow through artery stenosis at two locations (z = 0.001 and z = 10.26). the velocity profiles predicted by present model are in good agreement with the reference profiles except the slight deviation near the walls which appears due to the three-dimensional nature of the flow considered in the present study as compared to the two-dimensional behavior reported in the reference studies (deshpande et al. 1976; liao et al. 2004) (fig. 4). the steady state hat fully developed flow field is generated through simulations and the results were compared with newtonian and non-newtonian fluids. fig. 5 illustrates velocity and streamline distribution at the mid plane of the artery. the higher velocity region extends longer downstream to stenosis along with greater vertical regions near the walls for the newtonian fluid than the non-newtonian fluid. the non-newtonian fluid gives a deflated velocity profile and high velocity gradients are concentrated near the walls; whereas, the newtonian fluid exhibits a high velocity core downstream to the stenosis due to shear thinning. the separation regime spreads downstream of the stenosis much longer for the newtonian fluid than the non-newtonian fluid. p. halder, a. husain, m. zunaid and a. samad 196 both fluids produce a higher shear strain rate in the wall near the artery stenosis, but the strain rate is higher at the proximal upstream and lower at the proximal downstream of the stenosis (fig. 6). the non-newtonian fluid shows high wall shear stresses appeared in the larger part of the artery wall because of the higher viscosity, while, the newtonian fluid shows a higher shear strain rate. figure 7 shows velocity distribution for steady state fully developed flow at different streamwise locations, i.e. upstream to stenosis plane at z = -0.001 m, middle of the stenosis, z = 0.0 m, and downstream to stenosis planes, z = 0.001 m and z = 0.01 m. the velocity distribution is almost similar at the upstream of the stenosis for both fluids. a deflated velocity profile for the non-newtonian fluid appears at the proximal and distal downstream of the stenosis due to shear thinning. furthermore, transient flow analysis was carried out for both newtonian and nonnewtonian fluids for simple pulsatile, equivalent pulsatile and physiological flows and -1.5 -1 -0.5 0 0.5 1 1.5 0 0.2 0.4 0.6 0.8 1u /u m a x t/t simple pulsatile flow equivalent pulsatile flow physiological velocity figure 3. velocity profile with time for simple pulsatile, equivalent pulsatile and physiological flows. p q r s (a) z=0.001 (b) z=10.26 figure 4. validation of numerical model with existing computational data at re = 200 (deshpande et al. 1976; liao et al. 2004). 0 0.2 0.4 0.6 0.8 1 -0.5 0 0.5 1 1.5 2 2.5 3 r (m m ) u (m/s) computational data, deshpande et al. computational data,liao et al. present result, newtonian present result, non-newtonian 0 0.2 0.4 0.6 0.8 1 -0.5 0 0.5 1 1.5 2 2.5 r (m m ) u (m/s) computational data, deshpande et al. computational data,liao et al. present result, newtonian present result, non-newtonian figure 5. velocity distribution and streamlines at the mid-section for newtonian and non-newtonian flow. newtonian fluid non-newtonian fluid newtonian fluid non-newtonian fluid newtonian and non-newtonian pulsatile flows through an artery with stenosis 197 flow characteristics at some discrete locations in the artery are compared as shown in fig. 8. a higher velocity appears at the middle of the stenosis (point c) and proximal downstream to stenosis (point d) than the other upstream and downstream locations for the simple pulsatile flow for newtonian and non-newtonian fluid due to the constriction near the stenosis. for equivalent pulsatile and physiological flows, the velocity is maximum at and proximal to stenosis during the forward pulse, which is typically similar to simple pulsatile flow, whereas during the backward pulse the location of the maximum velocity (umax) is shifted towards upstream side. a significant velocity fluctuation appears at point e over the time due to the growth of flow disturbances (fig. 8(b) and (c)). the maximum velocity (umax) is observed at the middle and proximal downstream of stenosis (point c and point d) for all the cases. the umax appears twice during the cycle for simple pulsatile flow, during the forward pulse for equivalent pulsatile and during the first forward pulse for physiological flows. the minimum velocity (umin = 0) is observed at all points at different time intervals, i.e. at 3t/4 for simple pulsatile flows, at 0t, t/2 and t for equivalent pulsatile and at several time instants for physiological flows as shown in fig. 8. the flow field is further analyzed for streamlines, shear strain and shear stress at different time intervals for both newtonian and non-newtonian flows with simple pulsatile, equivalent pulsatile and physiological flows. figure 9 shows streamlines for the simple pulsatile flow at different time intervals for both newtonian and non-newtonian fluids at the middle plane of the artery. the flow field development is almost similar to both the fluids because of the similar velocity waveform in the first and second half-cycle. the mean velocity of the flow is higher in the forward half-cycle than the backward one. the asymmetry of the flow appears during the forward half cycle of newtonian fluid due to the growing recirculation and disturbance in the flow downstream to stenosis, whereas the higher viscosity resulting from the higher shear thinning suppresses the flow disturbances for non-newtonian fluid (fig. 9). though, both fluids produce recirculation downstream of the stenosis, due to lower viscosity, newtonian fluid produces a larger recirculation regime. figure 6. shear strain rate and shear stress distribution at the wall. newtonian fluid non-newtonian fluid newtonian fluid non-newtonian fluid figure 7. velocity distribution at several data reduction planes. newtonian fluid non-newtonian fluid z=-0.001 z=0.0 z=0.001 z=0.01 z=-0.001 z=0.0 z=0.001 z=0.01 plane 3 plane 4 plane 1 plane 2 p. halder, a. husain, m. zunaid and a. samad 198 point a point b point c point d point e (a) newtonian fluid (b) non-newtonian fluid 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e 0 1 2 3 4 5 6 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e 0 2 4 6 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e newtonian fluid (c) non-newtonian fluid 0 1 2 3 4 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e 0 1 2 3 4 0 0.2 0.4 0.6 0.8 1 v e lo c it y ( m /s ) t/t point a point b point c point d point e figure 8. (a) data reduction point probes, and velocity distribution for (b) simple pulsatile, (c) equivalent pulsatile, and (d) physiological flows, at different locations for newtonian and non-newtonian flows. newtonian fluid (d) non-newtonian fluid newtonian and non-newtonian pulsatile flows through an artery with stenosis 199 figure 10 shows the wall stress shear (wss) and shear strain rate distribution at different time intervals for simple pulsatile flows. the induction of wss facilitates atherosclerotic processes that affect the formation and rupture of atheromatous plaque. for the non-newtonian fluid having higher viscosity, a higher wss is produced near the bifurcation region, and the maximum wss lies between z = -0.001m to z = 0.0m (fig. 10). for newtonian fluid, a higher shear strain rate is noticed near the throat during the forward cycle. figure 11 shows the velocity distribution along the streamwise direction at different locations (z=0.001, 0.0, 0.001 and 0.01) and at different time intervals. for the newtonian fluid, the flow parameter changes significantly, and the changes are more dominant at z=0.001 and z=0.01. the symmetry of the flow is broken as the flow past stenosis due to the growing disturbances created by velocity pulse for the newtonian fluid which necessitates threedimensional modeling of the flow. as for nonnewtonian fluid, these disturbances are suppressed and symmetry is maintained. the transient effect of the velocity pulse on the artery flow field is analyzed through velocity and pressure distribution along the centerline in the artery at several time intervals for simple pulsatile flow (fig. 12). the velocity increases gradually along the flow direction up to the middle of the stenosis and sharply near the stenosis due to artery constriction. furthermore, velocity decreases gradually post stenosis until the end of the artery. the maximum and the minimum velocity in the artery appear at t = 0.3 s, and 0.7 s, respectively. t/8 5t/8 3t/8 7t/8 figure 9. streamlines at the mid-plane along the flow direction for simple pulsatile flows. newtonian fluid non-newtonian fluid t/8 5t/8 3t/8 7t/8 newtonian fluid figure 10. wall stress shear and shear strain rate distribution at different time for simple pulsatile flows. non-newtonian fluid newtonian fluid non-newtonian fluid p. halder, a. husain, m. zunaid and a. samad 200 the pressure in the artery first decreases gradually until the stenosis center then decreases sharply in the middle of the stenosis, which is then followed by gradual increase until the end of the artery. the minimum pressure appears at t = 0.3 s at proximal downstream of the stenosis indicating the regime prone to atherosclerotic processes. the streamlines at different time instances for the equivalent pulsatile flow are shown in fig 13. the newtonian fluid shows a uniform velocity profile in the core of the artery, whereas the non-newtonian fluid shows a velocity balloon near the stenosis. for the newtonian fluid, a large vortex covers the central core in both forward and backward pulses. the nonnewtonian fluid shows a small vortex only at the proximal region to the stenosis, i.e., downstream during the forward cycle and upstream during backward cycle. the nonnewtonian fluid flow produces a higher wss at all instances, a higher shear strain rate at all locations other than the downstream of the stenosis (fig. 14). both fluids give higher wss near the stenosis. the pressure and velocity distribution along the central line for equivalent pulsatile flow at different time intervals are shown in fig. 15. at t = 0.3 s, the peak velocity appears at the throat of the stenosis for the forward cycle. figure 16 shows streamlines for physiological flows for both fluids at different t/8 3t/8 newtonian fluid figure 11. velocity distribution at different locations in the flow direction for simple pulsatile flows. z=-0.001 z=0.01 z=0.001 z=0.01 z=-0.001 z=0.01 z=0.001 z=0.01 non-newtonian fluid (a) 0 1 2 3 4 5 6 7 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s -8000 -4000 0 4000 8000 12000 16000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s (b) figure 12. velocity and pressure distribution on the central line of the artery for (a) newtonian and (b) non-newtonian pulsatile flows at different time intervals. 0 1 2 3 4 5 6 7 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s -8000 -4000 0 4000 8000 12000 16000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x t=0.1s t=0.3s t=0.5s t=0.7s (mm) newtonian and non-newtonian pulsatile flows through an artery with stenosis 201 time intervals. the location of the vortex, which is proximal to the stenosis, keeps changing its location as the velocity pulse reverses its direction during the cycle. both fluids give a larger vortex during the first quarter. the temporal characteristics at the central line in the artery are shown in fig. 17 at different time intervals. the maximum velocity which is associated with the minimum pressure appears at t = 0.125 s for physiological flows, while the peak velocity is occurred at t = 0.3s for simple and equivalent pulsatile flows. the present research reports numerical modelling of blood flows through an artery with a stenosis for hemodynamic characteristics of atherosclerosis. the model was set up for newtonian and non-newtonian fluid flows for three commonly used velocity profiles, i.e. simple pulsatile, equivalent pulsatile, and physiological flows. the computational model was validated and results were compared with the results available in the open literature. the following conclusions can be drawn based on the findings of the present study: t/8 5t/8 3t/8 7t/8 newtonian fluid non-newtonian fluid figure 13. streamlines at the mid-plane along the flow direction for equivalent pulsatile flows. t/8 5t/8 3t/8 7t/8 newtonian fluid non-newtonian fluid newtonian fluid non-newtonian fluid figure 14. wall stress shear and shear strain rate distribution at different time for equivalent pulsatile flows. p. halder, a. husain, m. zunaid and a. samad 202  the newtonian fluid flow produces a large separation regime with vortex for uniform inlet velocity.  the newtonian fluid flow gives a higher shear strain rate, whereas the nonnewtonian fluid flow gives a higher wall shear stresses near the bifurcation regime.  in general, the maximum velocity appears at the middle of the artery. the nonnewtonian fluid pulsatile flow gave a deflated velocity profile as the flow past the stenosis. the newtonian fluid produces asymmetric velocity distribution as the flow past the stenosis whereas the nonnewtonian fluid gave symmetric flow field under the investigated flow conditions. (a) 0 0.5 1 1.5 2 2.5 3 3.5 4 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s -4000 -2000 0 2000 4000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s (b) figure 15. velocity and pressure distribution on the central line of the artery for (a) newtonian and (b) non-newtonian fluids equivalent pulsatile flows. 0 0.5 1 1.5 2 2.5 3 3.5 4 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s -4000 -2000 0 2000 4000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x (mm) t=0.1s t=0.3s t=0.5s t=0.7s t/8 5t/8 3t/8 7t/8 newtonian fluid non-newtonian fluid figure 16. streamlines at the mid-plane along the flow direction for physiological flows. newtonian and non-newtonian pulsatile flows through an artery with stenosis 203  the pressure drops sharply near the stenosis and recovers gradually as the flow past the stenosis.  the vortex and low-pressure zone remains proximal to the stenosis for the equivalent pulsatile flow. the vortex location keeps changing with the forward and backward flows during the pulsating cycle. the physiological flow produces the longest vertical regime among the types of flows investigated in the present study. conflict of interest the authors declare no conflicts of interest. funding the authors acknowledge the support of sultan qaboos university through collaborative research program (cl/squ-uaeu/15/02) for conducting this research. references ahsaas s, tiwari s (2016), numerical simulation of blood flow through asymmetric and symmetric occlusion in carotid artery. in: proceedings of the 3rd international conference on fluid flow, heat and mass transfer (ffhmt’16). ottawa, canada 170– 8. ang c, mazumdar jn (1997), mathematical flow through modelling of threedimensional an asymmetric arterial stenosis. mathl. comput. modelling 25: 19-29. ansys cfx (2010), release 11.0: ansys cfxsolver theory guide. ansys. bharadvaj bk, mabon rf, giddens dp (1982a), steady flow in a model of the human carotid bifurcation. part i-flow visualization. journal of biomechanics 15:349– 362. doi: 10.1016/0021-9290(82)90057-4. bharadvaj bk, mabon rf, giddens dp (1982b), steady flow in a model of the human carotid bifurcation. part ii-laser-doppler anemometer measurements. journal of biomechanics 15:363–378. doi: 10.1016/00219290(82)90058-6. (a) 0 0.5 1 1.5 2 2.5 3 3.5 4 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.125s t=0.375s t=0.625s t=0.875s -2000 -1000 0 1000 2000 3000 4000 5000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x (mm) t=0.125s t=0.375s t=0.625s t=0.875s (b) figure 17. velocity and pressure distribution on the mid line of the artery for (a) newtonian and (b) non-newtonian fluids physiological flows. 0 0.5 1 1.5 2 2.5 3 3.5 4 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 v e lo c it y ( m /s ) x (mm) t=0.125s t=0.375s t=0.625s t=0.875s -2000 -1000 0 1000 2000 3000 4000 5000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 p re s s u re ( p a ) x (mm) t=0.125s t=0.375s t=0.625s t=0.875s p. halder, a. husain, m. zunaid and a. samad 204 chakravarty s, mandal pk (1994), mathematical modelling of blood flow through an overlapping arterial stenosis. mathematical and computer modelling 19:59–70. doi: 10.1016/0895-7177(94)90116-3. chiastra c, gallo d, tasso p, iannaccone f, migliavacca f, wentzel jj, morbiducci u (2017), healthy and diseased coronary bifurcation geometries influence near-wall and intravascular flow: a computational exploration of the hemodynamic risk. journal of biomechanics 58:79–88. doi: 10.1016/j.jbiomech.2017.04.016. deplano v, siouffi m (1999), experimental and numerical study of pulsatile flows through stenosis: wall shear stress analysis. journal of biomechanics 32:1081–1090. doi: 10.1016/s0021-9290(99)00098-6. deshpande md, giddens dp, mabon rf (1976), steady laminar flow through modelled vascular stenoses. journal of biomechanics 9:165–174. doi: 10.1016/0021-9290(76)900014. ellahi r, rahman su, nadeem s (2014), blood flow of jeffrey fluid in a catherized tapered artery with the suspension of nanoparticles. physics letters, section a: general, atomic and solid state physics 378:2973–2980. doi: 10.1016/j.physleta.2014.08.002. gijsen fjh, van de vosse fn, janssen jd (1999), the influence of the non-newtonian properties of blood on the flow in large arteries: steady flow in a carotid bifurcation model. j biomech 32: 601–608. hye ma, paul mc (2015), a computational study on spiral blood flow in stenosed arteries with and without an upstream curved section. applied mathematical modelling 39:4746–4766. doi: 10.1016/ j.apm.2015.04.019. jozwik k, obidowski d (2008), geometrical models of vertebral arteries and numerical simulations of the blood flow through them. in: proceedings of the 3rd frontiers in biomedical devices conference and exhibition. june 18-20, irvine, california, usa. 75–79. ku dn, giddens dp, zarins ck, glagov s (1985), pulsatile flow and atherosclerosis in the human carotid bifurcation. arterioscler thromb vasc biol 5:293–302. doi: 10.1161/01.atv.5.3.293. liao w, lee ts, low ht (2004), numerical studies of physiological pulsatile flow through constricted tube. international journal of numerical methods for heat and fluid flow 14:689–713. doi: 10.1108/09615530410539991. long q, xu xy, ramnarine k v., hoskins p (2001), numerical investigation of physiologically realistic pulsatile flow through arterial stenosis. journal of biomechanics 34:1229–1242. doi: 10.1016/s0021-9290(01)00100-2. mazher ahk, ekaterinaris ja (1988), numerical simulation of blood flow through arteries. in: proceedings of the 1988 fourteenth annual northeast bioengineering conference. 10-11 march; durham, nh, usa, usa, 273–276. mcdonald da (1955), the relation of pulsatile pressure to flow in arteries. j physiol i27:533–552. doi: 10.1113 / jphysiol. 1955. sp005275. mekheimer ks, el kot ma (2015), suspension model for blood flow through catheterized curved artery with time-variant overlapping stenosis. engineering science and technology, an international journal 18:452–462. doi: 10.1016/j.jestch.2015.03.010. nadeem s, akbar ns, hendi aa, hayat t (2011), power law fluid model for blood flow through a tapered artery with a stenosis. applied mathematics and computation 217:7108–7116. doi: 10.1016/j.amc.2011.01.026. perktold k, rappitsch g (1995), computersimulation of local blood-flow and vessel mechanics in a compliant carotid-artery bifurcation model. journal of biomechanics 28:845–856. doi: 10.1016/0021-9290(95) 95273-8. perktold k, resch m, peter ro (1991), threedimensional numerical analysis of pulsatile flow and wall shear stress in the carotid artery bifurcation. journal of biomechanics 24:409–420. doi: 10.1016/00219290(91)90029-m. razavi a, shirani e, sadeghi mr (2011), numerical simulation of blood pulsatile flow in a stenosed carotid artery using different rheological models. journal of biomechanics 44:2021–2030. doi: 10.1016/j.jbiomech.2011.04.023. srivastava vp, rastogi r (2010), blood flow through a stenosed catheterized artery: effects of hematocrit and stenosis shape. computers and mathematics with applications 59:1377–1385. doi: 10.1016/j.camwa.2009.12.007. tu c, deville m (1996), pulsatile flow of nonnewtonian fluids through arterial stenoses. newtonian and non-newtonian pulsatile flows through an artery with stenosis 205 journal of biomechanics 29:899–908. doi: 10.1016/0021-9290(95)00151-4. tu c, deville m, dheur l, vanderschuren l (1992), finite element simulation of pulsatile flow through arterial stenosis. journal of biomechanics 25:1141–1152. doi: 10.1016/0021-9290(92)90070-h. van doormaal jp, raithby gd (1984), enhancements of the simple method for predicting incompressible fluid flows. numerical heat transfer 7:147–163. doi: 10.1080/01495728408961817. young df, tsai fy (1973), flow characteristics in models of arterial stenosis i. steady flow. journal of biomechanics 6:395–410. doi: 10.1016/0021-9290(73)90099-7. zendehbudi gr, moayeri ms (1999), comparison of physiological and simple pulsatile flows through stenosed arteries. journal of biomechanics 32:959–965. doi: 10.1016/s0021-9290(99)00053-6. the journal of engineering research (tjer), vol. 17, no. 1, (2020) 41-46 *corresponding author’s e-mail: algarny@iau.edu.sa doi:10.24200/tjer.vol17iss1pp41-46  potential enrichment of physical properties of gypsum adding palm tree ash and saw dust ali mohammad al-garny* and mohammed a. alhefnawi college of architecture and planning imam abdulrahman bin faisal university saudi arabia abstract: natural architecture uses locally produced materials to assure the strength of the construction elements. palm tree is widely planted in the gulf countries and its annual abounded residues constitutes a greater environmental burden of being neglected and rarely used or recycled in efficient ways. the paper discusses the potential changes in the physical properties of pure gypsum cubes when utilizing additives such as the ash and saw dust of palm tree residues. the additives are added in various proportions by weight. four groups of different laboratory tests are conducted on four different groups of cubic samples made of pure gypsum powder, admixture of gypsum and palm tree ash, and admixture of gypsum and palm tree saw dust. results showed that both sawdust and ash of palm tree have significantly enhanced the physical properties of gypsum cubes tested in this paper, weight, porosity, and compressive strength. keywords: ash; building materials; gypsum; palm tree; properties; saw dust. أشجار النخيل نشارة بإضافة رماد و فرص تحسين الخصائص الفيزيائية للجبس علي محمد القرني* و محمد الحفناوي تُزرع شجرة النخيل على والهندسة المعمارية الطبيعية المواد المنتجة محليا لضمان قوة عناصر البناء. تستخدم :الملخص عاد تدويرها بطرق يُ ونادراً ما تُستعمل أو ير، تشكل بقاياها السنوية الكبيرة عبئًا بيئيا كب كمانطاق واسع في دول الخليج التغيرات المحتملة في الخواص الفيزيائية لمكعبات الجبس النقي عند استخدام مواد مضافة البحثية الورقة هذه فعالة. تناقش أجريت أربع مجموعات من كما ،بنسب مختلفة حسب الوزنخلط المواد المضافة بقايا أشجار النخيل. تم ونشارة مثل الرماد ، (مجموعتين) من مسحوق الجبس النقي مصنوعة بةاالختبارات المعملية على أربع مجموعات مختلفة من عينات مكع . أظهرت النتائج (مجموعة) نخيلال ةشجر ةالجبس ونشارومزيج من ،(مجموعة) شجرة النخيل اد ومزيج من الجبس ورم أن كال من نشارة الخشب ورماد شجرة النخيل قد عززتا بشكل كبير الخواص الفيزيائية لمكعبات الجبس المختبرة في هذه الوزن والمسامية وقوة االنضغاط.من حيث الورقة .نشارة الخشب ؛مواد البناء ؛رماد اشجار النجيل؛ خصائص الجبس: لمات المفتاحيةالك potential enrichment of physical properties of gypsum adding palm tree ash and saw dust 42  1. introduction gypsum is an important building material that has many commonly known uses, such as gypsum boards and paints. gypsum is at the forefront of basic materials used in building and construction (eg 2017). it has many important characteristics such as: fire resistance, sound absorption, heat insulation, considerable bending strength, light weight, low cost, long life, fast solidification and ease of composition (kami and kami 1995). current necessities for gypsum compounds for execution or mortaring comprise organized setting time, workability, sag resistance, compressive and flexural strength, ideal bond to bricks and concrete, waterproofing, and better thermal and acoustic. these advances are accomplished by the diligence of some chemical admixtures and mineral additives; among them are superplasticizers, water-soluble polymers and the admixtures responsible for retarding and airentraining (arikan and sobolev, 2002). in the last few years, considerable interest has been demonstrated in the investigation of date palm waste products as reinforcing and insulating materials to raise the mechanical properties of building materials. consequently, much literature has been published on gypsum and the effects of some additives on its associated properties. this paper is conducted to investigate the properties of gypsum cubic blocks by adding the sawdust and ash separately to gypsum powder. 2. related work danso el al. (2015) investigated the characteristics of mud blocks stabilized with leftover agricultural fibers, and performed the tests of density, shrinkage, water absorption, compressive and tensile strength, corrosion and wearing. the results revealed general enhancement in the physical, mechanical and durability properties of the blocks with the 0.5 wt. % fiber content. sharma el al. (2016) presented an effort to develop the low durability of clay blocks by reinforcing it with the rural natural waste materials of grewia optiva and pinus roxburghii. the results showed an increase by 72 % to 68 % in the durability of the blocks reinforced with fibers of grewia optiva and pinus roxburghii. sharma el al. (2015a) investigated improving the compressive strength of adobe using natural fibers of grewia optivia and pinus roxburghii. the reinforcement with these fibers resulted in an increase in the compressive strength by about 94–200 % and 73–137 % for grewia optivia and pinus roxburghii respectively. sharma el al. (2015b) investigated enhancing soil sustainability and stabilizing it with natural fibers of pinus roxburghii and grewia optiva. results revealed an increase in the compressive strength by 131–145 % when adding fiber p. roxburghii and 225–235 % when adding fiber g. optiva for the cubical and cylindrical specimens respectively. zak el al. (2016) presented an experiment to enhance the compressive strength of earth bricks using different admixtures of earth, cement, gypsum, hemp and flax fibers. the results indicated that the impacts of adding fibers hemp and flax is very limited on the compressive strength. cement and gypsum reduced the binding force of the clay and highly decreased strength. villamizar el al. (2012) investigated the effect of adding coal-ash and cassava skins on the properties of mud blocks. results showed that stabilizing earth blocks with coal-ash improved the compressive and bending properties when added as less than or equal to 5 %. ahmed el al. (1988) investigated the properties of date palm frond and concluded that tensile strength of the stalk walls is double that of the cores. the elasticity modulus of stalks is between 10 to 30 kn/mm2. arikan and sobolev (2002) studied the effect of different admixtures on the compressive strength consistency. the additive admixtures are a superplasticizer, a water-soluble polymer, a retarding, and an air-entraining. the results proved that admixtures slightly increased the setting time, whereas, the application of a retarding admixture allowed control of a setting time of 1 to 3 hours. li el al. (2003) stated that the addition of cotton stalk to gypsum had a substantial impact on its mechanical properties. cotton stalk fiber is treated with emulsion to increase its combination with gypsum. the results illustrated that, although the composite of gypsum had lower strength, its insulation and fire resistance properties had improved. on the other hand, wu (2004) studied the shear performance of gypsum panels when being longitudinally reinforced with glass fiber. the attained results showed that persistence of the longitudinal reinforcement had no significant issue on the shear strength, whereas it caused a small impression on their early stiffness, stiffness degradation, ductility ratio, and energy dissipation capacity. abu-sharkh and hamid (2004) studied the properties of a composite material produced by mixing palm leaves, polypropylene, and stabilizers. the results proved that uncompatibilized specimens are more stable than compatibilized ones. they justified their results that maleate polypropylene has low stability. kriker el al. (2005) tested the properties of palm date fiber reinforced concrete. it is found that increasing percentage and length of fiber in curing improved the flexural strength and the toughness coefficient and reduced the compressive strength. turgut and algin (2007) mixed limestone powder and wood sawdust. the properties of the obtained material such as weight, water absorption, compressive strength, flexural strength and ultrasonic velocity satisfied the appropriate international standali mohammad al-garny and mohammed a. alhefnawi 43  ards. therefore, the product had potentials for uses for walls, ceiling panels, sound panels, wooden board substitute, and a cheap alternate to concrete blocks. liangyuan and zongli (2007) debated the features inducing the performance of gypsum-based composite material such as water/cement ratio, and retarder quantities. it is concluded that the strength and coefficient of softness increased compared to gypsum. portland cement quickened the thickening and improved the strength and softness coefficient of composite, but over dosage caused the fall of these coefficients. jia-yan el al. (2008) studied the connection between wood-gypsum ratio and water gypsum ratio and explored the effects of temperature on hydration process and the gypsum morphology. it is concluded that both water-gypsum ratio and wood-gypsum ratio are crucial to improve the properties of the boards in 40℃ and to shorten the hydration time and strengthen the boards. okunade (2008) examined the impacts of adding sawdust and wood ash admixtures to clay mix to produce bricks. the results showed that the key influence of the sawdust admixture is reducing the dry density of the brick. furthermore, the wood ash admixture attained denser products with higher compressive strength. brencis el al. (2001) produced an energy saving material made of foam gypsum and fibrous hemp in which sound absorption coefficient increased. it is also found that short fibers’ reinforcement increased the density of foam gypsum, while long fibers decreased it. hai alami (2013) studied the properties of clay mixture with desert solid wastes such as fronds and pits. it is proved that crushed pits expressively improved the strength of clay bricks adding more 85 % toughness, and 15 % strength. zhou el al. (2013) studied the properties of gypsum-based composites produced by mixing the powder of gypsum with latex and polyvinyl acetate (pva). release time, density and impact strength are investigated. it is proved that pva contributed to delaying of time release and increased impact strength. on the other side, sawdust reduced the density. sari (2014) compared between two mixtures of polyethylene glycol with gypsum and clay in regard to low temperature-thermal energy storage. it is proved that compatibility between components in regard to capillary and surface tension enabled the composite for passive thermal energy storage applications in buildings. dai and fan (2015a) developed a complex from gypsum and sawdust with a water-based epoxy spray. it is proved that sawdust water increased the flexural and compressive strength of gypsum by 10 % and 7 % respectively. furthermore, analysis with optical microscopy confirmed that decrease of water uptake enlarged the gypsum covering ratio from 42 % to 68 %. dai and fan (2015b) studied the mechanical properties of a bio-composite of wood sawdust and gypsum. it is concluded that the saw dust ratio of 20% gained better flexural and compressive strengths compared with the 30% ratio. algarny el al. (2016) surveyed the feasibility of adding palm tree sawdust to gypsum powder to produce blocks for indoor usage. it is resolved that the shrinkage value, weight, and porosity value are significantly reduced by increasing the ratio of sawdust to gypsum powder, whereas compressive strength is increased at the same time. 3. materials and methods the materials used are pure gypsum powder, palm tree saw dust, and the ashes resulting from the burnt saw dust of the local palm tree fronds in the eastern province of the kingdom of saudi arabia. different mixing ratios of ash or saw dust to gypsum are added to produce gypsum cubes of 5×5×5 cm, table 1. three different laboratory tests are conducted at the same time for every two groups of the samples: weight, porosity, and compression strength. humboldt mortar laboratory mixer of 5l, which supports the astm c305 standard, is used. it is a bench top type that provides two mixing speeds of 140±5 rpm and 285±10 rpm. it has the dimensions of 235×396×568 mm. materials described are mixed together in the dry state. then, water is added and mixed until the mixture became homogeneous before conducting the tests. mixing time is not to exceed one minute, so as not to get water evaporates. each test is equipped with a number of 25 specimens, at a rate of 5 specimens for each mixing ratio. the result recorded in the test is the average value of the 5 specimens of each mixing ratio. the total number of the tested specimens are 150 specimens. unlike a number of other specimens that are excluded for defects during the casting or the processing of the test raising, the number of specimens that are actually consumed to more than 200 specimens. in the weight test, the specimens are casted in their molds and left for 24 hours. then, they are taken out and weighed by a digital scale. the arithmetic average is calculated for each mixing ratio using the recorded weights of 5 different specimens. the results are shown in fig. 1. in order to conduct the porosity test, the specimens are undergone the following steps: 1. the dry weights of the specimens are measured by a digital scale. 2. the specimens are immersed in water for 24 hours. 3. the wet weights of the specimens are measured by a digital scale. 4. the following formula is applied to calculate the value of porosity in the specimen (astm 2000): 𝑃, % 𝑊 𝐷 /𝑉 100 p: porosity w: saturated weight d: dry weight v: specimen dry volume (1) potential enrichment of physical properties of gypsum adding palm tree ash and saw dust 44  the results are shown in fig. 3. in the compressive strength test, for each admixture, five cubic specimens from each mixing ratio are prepared in a total of 50 specimens required for the test. the specimens are compressed in the compression strength instrument and the results are shown in fig.4. 4. results and discussion 4.1 weight test as shown in fig. 1, adding ash or sawdust reduced the specimens’ weights. similar results have been reported by the researchers algarny el al. (2016), li el al. (2003), and arikan and sobolev (2002). the addition of saw dust reduced the weight of the specimen more effectively than adding ash, as shown in the values of r2 for the trend lines of both results. it can be justified as the saw dust has less density and less weight for a specific volume compared to the ash. taking into account the constant volume of the cubic specimen tested, adding more saw dust reduces more weight of the specimen than adding more ash. results indicate that the maximum weight reductions are 3.62% and 4.9% of the control specimen for ash and saw dust respectively at 6% mixing ratio. so, saw dust-gypsum admixtures would have more benefits for the construction industry and reduce the internal loads and stresses on the structural and construction elements of the buildings (fig. 2). 4.2 porosity test as shown in fig. 3, the addition of ash and sawdust significantly reduced the porosity value of the specimens. lower porosity is recorded in higher mixing ratios. the result is consistent with the work reported by algarny el al. (2016), and okunade (2008). the porosity values are reduced by 4.80 % and 3.77 % in the 6 % admixtures of gypsum-ash and gypsum-sawdust respectively compared to the control specimens. sawdust has a limited impact on decreasing the porosity of gypsum cubes compared to the ash, which might be because sawdust has a lower density and bigger particle size compared to ash. therefore, for a specific weight of ash or saw dust added to the gypsum powder, sawdust has a relatively higher volume than ash leading to keeping more gaps between the particles of the gypsum powder compared to the case of adding ash which has a compacted higher density compacted form that introduces smaller quantities of gaps among the gypsum particles. so, gypsum-ash cubes are more convenient to applications that need less porosity compared to sawdust-gypsum cubes that suits the need for more porous materials. table 1. mixing ratios. gypsum (gm) ash / sawdust water (gm) 3000 % (gm) 1950 (65 % of gypsum weight) 0 0 1 30 2 60 3 90 6 180 figure 1. relationship between the weights of the specimens and the applied mixing ratios of the additives. figure. 2. relationship between the weight loss of the specimens and the applied mixing ratios of the additives. weight loss is directly proportional to mixing ratios. figure 3.  relationship between the porosity of the specimens and the applied mixing ratios of the additives. ali mohammad al-garny and mohammed a. alhefnawi 45  figure. 4. relationship between the fracture loads of the specimens and the applied mixing ratios of the additives. fracture load is directly proportional to mixing ratios. 4.3 compressive strength test the results in fig. 4 show that both the sawdust and ash have significantly increased the compression strength of the specimens and improved their failure resistance, which is consistent with the work reported by algarny et al. (2016), arikan and sobolev (2002), and okunade (2008). the resistance to fracture has increased by125.83 % and 62.9 % for the specimens of 6 % mixing ratio of sawdust and ash content respectively compared to the control specimen which is of pure gypsum. this result is tremendously important and can divert the use of gypsum in new building and construction applications due to its increased resistance to fracture in compression. adding sawdust to the gypsum cubes will raise its compressive strength to almost double the value it will get if adding ash instead. the fibers content in the sawdust might be responsible for providing stronger bonds among the gypsum particles compared the case of adding ash which has already lost its fiber during the burning process. the justification of the previous results can be explained as palm contains wax, phenol compounds and pectin (al-dosary 2009), which have chemical properties to enhance the physical properties of gypsum powder. wax solidifies at normal temperature and melts in heat (eb 2019a). when adding water, the wax melts from the emitted heat and spreads within the gypsum particles, then the wax hardens increasing the cohesion of the admixture blocking the pores of gypsum, and reducing the porosity value of the cubes, while increasing its strength. on the other hand, pectin material gives the rigidity to the plant (al-dosary 2009), and when added to gypsum it increases its hardness and compression strength. in addition, phenol is soluble in water (eb 2019b), and it interacts with gypsum powder to produce salty water and stimulate the interaction between the granulated particles of gypsum and sawdust and ash of palm granules reducing porosity as well. 5. conclusion both sawdust and ash of palm tree have significantly enhanced the physical properties of gypsum cubes tested in this paper. both induced weight loss but sawdust achieved more reduction. the weight of gypsum specimen is inversely proportional to the amount of sawdust or ash content; as lower weights are recorded with higher content of the additives in the specimen. in line with the previous result, porosity ratio in the specimen of gypsum cubes is inversely proportional to the amount of ash or sawdust. both additives reduced the porosity of gypsum cubes, but ash has a broader reduction. on contrary, compressive strength of the samples of gypsum cubes is directly proportional to the amount of ash or sawdust, where higher values of compressive strength are recorded in higher ratios of any of them in the specimen. the compressive strength the cubes gain when adding sawdust is almost double of that gained when adding the ash. the overall findings are encouraging for promoting the use of leftover palm tree residues to enhance the physical properties of some building and finishing materials which leads to adding an economic value and securing more environmental protection against soil and air pollution. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references abu-sharkh b f, hamid h (2004), degradation study of date palm fibre/polypropylene composites in natural and artificial weathering: mechanical and thermal analysis. journal of polymer degradation and stability 85: 967–973. ahmad f, abdel-rahman h, al-juruf r, alam i (1988), physical, mechanical and durability characteristics of date palm frond stalks as reinforcement in structural concrete. international journal of cement composite and lightweight concrete 10(3): 175-180. al-dosary n h (2009), effect of leaf pinaes chemical composition of different date palm cultivar on infection by whit scale insect parlatoria blanchardii targ. homoptera: coccinea: diaspidae, date palm research centerbasrah university, basrah, iraq. algarny a, al-naimi i, alhefnawi m (2016), enhancing the properties of gypsum as an indoor 0 5 10 15 20 0% 2% 4% 6% 8% f ra ct u re l o ad s (k n ) mixing ratioash saw dust potential enrichment of physical properties of gypsum adding palm tree ash and saw dust 46  finishing material using palm tree residues. kku of basic and applied science 2(2), 111-118. arikan m, sobolev k (2002), the optimization of a gypsum-based composite material. journal of cement concrete research 32 (11): 1725–1728. astm designation: c 20 – 00 (2000), standard test methods for apparent porosity, water absorption, apparent specific gravity, and bulk density of burned refractory brick and shapes by boiling water. astm. pa. united states. brencis r, skujans j, iljins u, ziemelis i, osits o (2011), research on foam gypsum with hemp fibrous reinforcement. journal of chemical engineering transformation 25:159-164. dai d, fan m (2015a), preparation of gypsum/sawdust green composite with spray coating. rsc advances journal. an online article at: http://pubs.rsc.org/en/content/articlelandin /2015/ra/c5ra18707a#!divabstract. last accessed on 22/2/2019. dai d, fan m (2015b), preparation of bio-composite from wood sawdust and gypsum. journal of industrial crops and products 74 (15): 417–424. danso h, martinson d, ali m, williams j (2015), physical, mechanical and durability properties of soil building blocks reinforced with natural fibers. journal of construction and building materials 101 (1): 797-809. eg eurogypsum (2017), living with gypsum: from raw material to finished product. an online article at: www.eurogypsum.org. last accessed on 11/2/2019. encyclopedia britannica eb (2019a), wax. an online article at: http://www.britannica.com/ technology/wax. last accessed 30/2/2019. encyclopedia britannica eb (2019b), phenol chemical compounds. an online article at: http://www.britannica.com/science/phenol. last accessed 30/2/2019. hai alami a (2013), experiments on unfired masonry clay bricks mixed with palm fronds and date pits for thermal insulation applications. journal of renewable and sustainable energy 5 (2): 31-36. jia-yan l, zhen-hua h, yu-he d, de-xin z, wen l, ling x, mei z (2008), the effects of woodgypsum ratio, water-gypsum ratio and temperature on the properties of the gypsum sawdust board. journal of nanjing university (natural science edition). an online article at: en.cnki.com.cn/article_en. last accessed on 20/2/2019. karni j, karni e (1995), gypsum in construction: origin and properties. journal of materials and structures 28 (2): 92-100. kriker a, debicki g, bali a, khenfer m m, chabannet m (2005), mechanical properties of date palm fibres and concrete reinforced with date palm fibres in hot-dry climate. journal of cement and concrete composites 27 (5):554-564. li g, yu y, zhao z, li j, li c (2003), properties study of cotton stalk fiber / gypsum composite. journal of cement and concrete research 33 (1): 43–46. liangyuan l, zongli s (2007), research on contributing factors of gypsum-based composite material performance. new building materials. an online article at: http://en.cnki.com.cn/article_en/cjfdtotalxxjz200705000.htm. last accessed on 06/2/2019. okunade e a (2008), the effect of wood ash and sawdust admixtures on the engineering properties of a burnt laterite clay brick. journal of applied sciences 8(6): 1042–1048. sari a (2014), composites of polyethylene glycol with gypsum and natural clay as new kinds of building pcms for low temperature-thermal energy storage. journal of energy and buildings 69: 184–192. sharma v, marwaha b, vinayak h (2016), enhancing durability of adobe by natural reinforcement for propagating sustainable mud housing. international journal of sustainable built environment 5 (1): 141-155. sharma v, vinayak h, marwaha b (2015a), enhancing compressive strength of soil using natural fibers. journal of construction and building materials 93: 943-949. sharma v, vinayak h, marwaha b (2015b), enhancing sustainability of rural adobe houses of hills by addition of vernacular fiber reinforcement. international journal of sustainable built environment 4 (2): 348-358. turgut p, algin h (2007), limestone dust and wood sawdust as brick material. journal of building and environment 42 (9): 3399–3403. villamizar m, araque v, reyes c, silva r (2012), effect of the addition of coal-ash and cassava peels on the engineering properties of compressed earth blocks. journal of construction and building materials 36: 276-286. wu yf (2004), the effect of longitudinal reinforcement on the cyclic shear behavior of glass fiber reinforced gypsum wall panels. jour-nal of engineering structures 26 (11):1633–1646. zak p, ashour t, korjenic a, korjenic s, wu w (2016), the influence of natural reinforcement fibers, gypsum and cement on compressive strength of earth bricks materials. journal of construction and building materials 106: 179-188. zhou s b, xiao a g, chen y z, chen g, hao a p, chen y d, huang x b (2013), the preparation and performance of gypsum-based composites. journal of applied mechanics and materials 310: 46-50. microsoft word paper 3.docx the journal of engineering research (tjer), vol. 14, no. 1 (2017) 23-38   strengthening of historical stone masonry buildings in oman using textile reinforced mortars a.h. al-saidy*, a.w. hago, s. el-gamal and m. dawood department of civil and architectural engineering, college of engineering, sultan qaboos university, oman. received 27 august 2015; accepted 29 june 2016 abstract: most historical buildings and structures in oman were built using unreinforced stone masonry. such structures are prone to damage from seismic activity as these structures have negligible resistance to tensile stresses. this paper presents an exploration of the physical and chemical properties of local building materials and the results of experiments with a new strengthening technique using textile reinforced mortar. limestone and sarooj were the main building materials used historically in these buildings. sarooj is a local omani term for artificial pozzolana produced by calcining clays. two mortar mixes were developed, one as the jointing mortar and the second as the matrix mortar. the second part of the study consisted of the construction of three masonry walls and two columns of 0.35 x 0.25 x 1.2 m (width x depth x height). the three walls were strengthened using externally bonded textiles using three different combinations of mortar and textile. one column specimen was kept as a control and the other was fully wrapped with carbon textile reinforced mortar. the walls were tested under out-of-plane four-point bending. all the wall specimens showed a significant improvement in strength up to three times its self-weight and reasonable deformation before failure. the column specimens were tested under axial compression. the failure load and displacement of the strengthened columns increased by seven and three times, respectively, compared to the control column. keywords: strengthening, historical, stone masonry wall, textile reinforced mortar, rehabilitation. אאאאאאא אא*،א،אא،א אאאwאאאאא   א א   א א  א   k א אא  א   א  k  ؛  א א  אא א  א   אאאאאאאאא kאאאא אאאא א א א           א א א אאאאאkא א،אאא אא א א א א    kא    אkאאאאאאא אאאאאאאאאאאאא kאאאאאאאאא אאwא،א،،،אאk * corresponding author’s email: alsaidy@squ.edu.om a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   24 1. introduction oman has a long and rich history that has defined every aspect of its culture. in standing witness of this civilization, there exists a great number of historical buildings distributed all over the country. a variety of historical, geographical, social, and economic events has influenced the omani building style, both from structural and architectural perspectives. these historical buildings are of great value to oman as they highlight the development of omani civilization over the years from social, cultural and economic points of view, which when combined contribute to the current omani identity. preserving these buildings is a national duty to keep omanis linked to their history (ministry of heritage, 1994). an inspection of oman’s historical buildings indicates that omanis of the past utilized locally available materials in their buildings due to their low cost and compatibility with the surrounding environment (al-mashani 1997). oman is a relatively large country with significant variations in topography, geography, geology, and climate. these variations influenced the materials used, and unique methods of construction followed. the importance of the buildings and their intended uses were major factors in the selection of the materials and construction methods. unreinforced masonry walls (urm) are the main load-carrying elements in oman’s historical buildings. these walls are limited in their ability to withstand compressive vertical loads (bernat-maso et al. 2014). unreinforced masonry walls are prone to damage from seismic activity, which produces in-plane and out-of-plane loadings. therefore, masonry walls need to be strengthened for both flexure and shear loads. due to the recent threat of seismic activity in the region, oman’s structures may be subjected to tensile forces. therefore, solutions for strengthening these structures are crucial. the walls in historical structures are typically constructed with stones or mud bricks and bonded with locally produced mortar made by mixing sarooj, lime, and water. sarooj is an omani term for artificial pozzolana produced by calcining clays. calcined clay mixed with lime and water has been used as a cementing material for a long time (al-rawas and hago 2006). in addition, unreinforced masonry columns are critical members in omani historical buildings and require upgrading to enhance their axial compressive strength. various conventional methods have been used for strengthening masonry walls such as the construction of single or double jackets made of steel-reinforced shotcrete or cast-insitu concrete and grout injections (ehsani et al. 1997; triantafillou 1998). these methods may be problematic due to the extra weight added to the building that increases the threat of failure under cyclic loading. such reinforcement has drawbacks, including the eventual deterioration of steel reinforcement due to corrosion, the need for large site work and scaffolding, and the effect on the appearance of the building which does not preserve the architectural value of the historical buildings (ehsani et al. 1997; triantafillou 1998). the application of fiber reinforced polymers (frp) in upgrading and reinforcing civil engineering structures has received special attention in recent years due to frps’ favorable properties and their extremely high strengthto-weight ratio, corrosion resistance, ease and speed of application, and minimal change in the geometry. frp composites are made of fibers (carbon, glass, and aramid) held by polymer resin (epoxy, polyester, or vinyl ester). the conventional method of applying frp composites involves bonding them to previously prepared surfaces (concrete or masonry) by means of an epoxy resin. however, this technique has a few drawbacks which are attributed to the resins used to bind or impregnate the fibers. these drawbacks include poor behavior of epoxy resins at high temperatures; the requirement of special and expensive fire protection measures; the relatively high cost of epoxies; a lack of vapor permeability, which may cause damage to the concrete or masonry structure, and the incompatibility of epoxy resins and substrate materials (harajli et al. 2010; kolsch 1998; papanicolaou et al. 2011; pinho et al. 2014). the solution to overcome the drawbacks associated with these physicochemical incompatibilities (water-vapor permeability) is to replace the epoxy resin with an inorganic strengthening of historical stone masonry buildings in oman using textile reinforced mortars   25   binder such as cement-based mortars. furthermore, in order to ensure a good matrixcore adhesion, grid textiles may replace continuous fiber sheets. this composite material based on textiles embedded in a mortar matrix is known as textile reinforced mortar (trm). kolsch (1998) performed a study of strengthening urm walls with fiber reinforced mortar (frm). in his study, full scale masonry walls constructed from sand-lime bricks and strengthened with a carbon fiber cement matrix system (three layers of unidirectional carbon textile in a polymer-modified cement binder) were tested under out-of-plane loading. cyclic horizontal loading and unloading was exerted on the walls using a pressurized air bag. the wall systems exhibited a considerable increase in the out-of-plane bending capacity compared with the unstrengthened walls. harajli et al. (2010) investigated urm walls strengthened with textile reinforced mortar. specimens (n = 25) were constructed of different stones—hollow block, brick, and sandstone—bonded by a mortar. textile meshes made out of either bitumen-coated eglass fibers (coated to provide an alkaline resistance) or coated/uncoated basalt fibers were attached to the tension surface by mortar, which was either lime mortar of normal and high strengths or cement-based mortar. regardless of the causes of failure, trm walls resulted in a considerable increase to their outof-plane flexural and displacement capacities when subjected to static loading. papanicolaou et al. (2011) investigated the effectiveness of trm jacketing for out-of-plane loading of masonry specimens. the main parameter explored in the investigation was the number of grid layers used to form the jackets (one or two layers, applied on both sides for all specimens). the results indicated that even the weakest trm configurations (low-grade textiles combined with low-strength mortars), when adequately anchored, resulted in a more than 400% increase in strength and a 130% increase in deformability. pinho et al. (2014) worked on strengthening rubble stone masonry walls. this strengthening solution used three-dimensional (3d) ties comprising zinc coated steel wires and a textile reinforced mortar rendering (composed of a 30 mm layer of lime, cement, and sand [1:1:6] mortar, reinforced with glass fiber mesh) applied to both sides of the specimens. specimens were subjected to axial compression force. results indicated an up to three times improvement in both compressive strength and stiffness of the strengthened walls relative to the unstrengthened walls. however, the ductility of the walls was not improved. buckling was observed in all strengthened walls. nevertheless, the nature of failure was significantly improved as complete collapse did not occur compared to the reference walls. the strengthening system was also found to be compatible with the masonry walls in terms of porosity, permeability and water capillary. to preserve omani historical buildings using compatible materials, local materials should be used as much as possible. to implement advance strengthening techniques such as trm, mortar made from sarooj was used since this material had been used in the original construction of such historical structures. hago and al-rawas (1997) indicated that sarooj properties are dependent on the mineral composition of the clay used in terms of the reactive silica and alumina composition. the binding nature of sarooj results from its reactive silica and alumina reaction with lime and water to produce calcium silicates and aluminates hydrates similar to those in portland cement and high aluminate cement, respectively. sarooj’s burning temperature and duration was found to influence the pozzolanic-reactivity of the resultant sarooj. hago et al. (2002a) indicated that sarooj’s reactivity is not only a function of its chemical composition and burning conditions but also of its physical properties, including particle size distribution, density, shape, and fineness, which also influence the sarooj properties. it was observed that increasing the sarooj-lime mixture’s fineness increased the compressive strength and decreased workability and setting times. hago et al. (2002b) investigated different curing conditions on the compressive strength of sarooj-cement mixes. it was found that for the prevailing climate conditions in the gulf a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   30   region, sun and air curing are the best to expedite the strength that can be gained from sarooj mortars. in the same study, the optimum sand/sarooj and cement/sarooj ratios were also investigated for a variety of mixes. the study concluded that a sand/sarooj ratio of 0.3 and a cement/sarooj ratio of 0.6 represent the optimum ratios that provide the highest compressive strength for all mixes. the experimental work presented is a pilot study in the strengthening of historical buildings through construction and strengthening of representative elements of walls and columns. to the authors’ knowledge, this is the first study that deals with strengthening local omani historical buildings. the study consisted of a collection of lime stones and sarooj and conducting chemical analysis and physical tests to determine various properties of the materials. scaled masonry walls and columns were then constructed. strengthening schemes were applied to these elements using glass fiber and carbon fiber textiles bonded with compatible mortar made of sarooj. out-of-plane loading was used to test strengthened walls while axial compression was used to test columns. 2. material characterization 2.1 stone limestone is defined according to the american society for testing and materials (astm) standards c51 (2011) as an initially sedimentary rock consisting mainly of calcium carbonate or of the carbonates of calcium and magnesium. limestone may be high in calcium, magnesium, or dolomiticas. due to limestone’s wide availability, its ease of cutting, its good heat and humidity insulation, and its aesthetics, it was widely used historically in oman’s historical buildings. the limestone used in this study was obtained from wadi alkhoud, and the stones were crushed to fragments of random sizes (fig. 1a). the stone was verified to be limestone by applying hydrochloric acid (hcl) and listening for a strong fizzing to occur due to the co2 produced from the reaction of hcl with caco3 (fig. 1b). a sample of the limestone crushed into particles and passed through a sieve sized 0.075 mm was prepared for chemical analysis. astm standard c 25-11 procedures were followed to determine its chemical composition. limestone’s main chemical constituents are summarized in table 1. the stone’s uniaxial compressive strength was measured in accordance with astm standard d7012 (2004). in this test, 10 cores of 37 mm diameter and 80 mm in height were extracted from random limestone samples. compressive load was applied through two circular plates at a rate of 0.5 kn/s until failure occurred. the average compressive strength was found to be 50 mpa for the air-dried limestone cores. water absorption of the stones was determined following astm standard (a) (b) figure 1. a) limestone , b) limestone reaction with hcl. table 1. main chemical compounds of limestone and sarooj. compound limestone (%) sarooj (%) cao 39.99 28.92 sio2 3.18 21.46 al2o3 0.59 6.49 fe2o3 6.61 3.66 mgo 1.5 3.28 loi 41 27.51 25 26 strengthening of historical stone masonry buildings in oman using textile reinforced mortars   25   d6473 (2010). the stones were weighed after they had been soaked in water. the stones were later oven dried and weighed. water absorption was then calculated and found to be 2.1%. 2.2 sarooj sarooj is a traditional omani construction material and a main component of historical buildings, especially those made of stones. sarooj is manufactured by subjecting clay to the calcination process. previous studies have identified different factors that affect the reactivity of sarooj, including physical, mechanical, chemical, and geotechnical properties. the sarooj used in this study was obtained from the ministry of heritage and culture (mhc) where it was produced in the nakhal sarooj factory following traditional production methods. nakhal-produced sarooj is used extensively in the restoration work of historical buildings administered by mhc. a sample of sarooj passing 0.075 mm was prepared for chemical analysis. standard procedures were followed to determine its chemical composition (astm c 114-13). the chemical analysis data is an important indicator of the sarooj pozzolanic activity. sarooj’s main chemical constituents are summarized in table 1. three major components—sio2, al2o3 and fe2o3— constitute 31.61% of the sarooj. astm standard c618 and indian standards (varshney 1982) require that, for a good pozzolana, the total of the three components should be greater than 70 and 50%, respectively. from the chemical composition alone, it seems that the sarooj is not likely to have high pozzolanic activity. the chemical composition depends on the source of the calcined clay and the calcination process (temperature, duration etc.). the latter is not well controlled in the traditional method of producing sarooj (al-rawas and hago 2006). 2.3 jointing mortar historical data indicate that mixtures of coarse sarooj, sand, and water were used to bind stones in the construction of omani masonry structures. workability was the main parameter in designing the mix as no historical records of the exact mix proportions is available. in the current study, sarooj passing through a 4.75-mm sieve was used along with coarse normal sand passing through a five-mm sieve and water with proportions of 2:1:0.67, respectively. the slump for the mix was found to be 85 mm. the mix was prepared and cast in 70.7 x 70.7 x 70.7 mm3. the cubes were removed from the mold after 24 hours and left to cure in the air under lab temperatures of 20– 22°c for 28 days before testing under compression. three cubes were tested under uniaxial compression at a rate of 0.25 mm/min for 28 days for compressive strength. the average 28 days’ compressive strength was found to be 0.76 mpa. the long setting time and low strength is mainly attributed to the low pozzolanic activity of the sarooj produced by traditional methods as indicated by chemical compounds—mainly sio2, al2o3, and fe2o3 as discussed in section 2.2. this also requires a long curing period for the sarooj to gain strength as reported in several studies (alrawas and hago 2006; hago and al-rawas 1997; hago et al. 2002a; hago et al. 2002b). 2.4 matrix mortar in order to bond the textiles to the stone masonry specimens, and develop adequate bonding strength, durable and compatible mixes for plastering were developed. since this was to be applied externally to an existing structure, some modification in the sarooj mix to enhance the bonding strength was determined to be acceptable, but this could not be done in the joining mortar as it was considered to be part of the existing condition of the historical structure. the optimum design mix ratios of sarooj, sand, and cement to produce the maximum strength were adopted from a study by hago et al. (2002b). a sand: sarooj ratio of 0.3 and cement/sarooj ratio of 0.6 were used in the matrix mortar. to enhance the strength and bond properties of the matrix mortar, very fine sarooj and sand were used. accordingly, two mixes of sarooj and cement, and sand and water were prepared using normal and dune sand alternatively. the first mix (m1) was made of sarooj: cement: sand: water with proportions of 1:0.6:0.3:0.52 using dune sand and sarooj, both of which passed through a 0.3-mm sieve. the second mix (m2) was prepared with sarooj: cement: sand: water in proportions of 1:0.6:0.3:0.5 using normal sand passing through a 0.6-mm sieve and sarooj passing through a 0.3-mm sieve. cubes of 70.7 x 70.7 x 70.7 mm were cast, removed after one day from the molds and kept under 27 a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   30   air curing at a laboratory temperature of 20– 23°c for 28 days before testing in compression. prisms of 100 x 100 x 500 mm were cast, removed from the molds after one day and air cured under laboratory temperatures for 28 days in order to be tested for flexural tensile strength. both mixes resulted in an equal slump of 70 mm. mix m1 yielded an average 28 days compressive and flexural tensile strengths of 14.77 mpa and 1 mpa, respectively. mix m2 resulted in average 28 days compressive and flexural tensile strengths of 18.12 mpa and 1.21 mpa, respectively. normal sand matrix mortar (m2) showed slightly higher compressive and flexural tensile strength compared to the dune sand matrix mortar (m1), which may be attributed to the granular surface of the normal sand compared to the smooth surface of the dune sand. 2.5 textile carbon fiber (cf) and glass fiber (gf) bidirectional textiles were used for the strengthening of the specimens. cf textile has a roving dimension of 1.1 mm wide by 0.5 mm thick in the longitudinal direction with 12 mm spacing between the rovings; and a 0.5 mm wide by 0.5 mm thick dimension in the transverse direction with 6 mm spacing between the rovings. gf textile has a roving dimension of 1.3 mm wide by 0.5 mm thick in both directions and spacing between the rovings of an average of 5 mm in both directions (fig. 2). the textile was characterized in a laboratory by means of uniaxial tensile tests according to astm standard d5034–09 (2013). textile fibers specimens were cut into 220 mm lengths and 80 mm widths to fit the testing machine (fig. 3). the test was done on a load control rate of 10 kn/s. table 2 shows the tensile test results of the textile specimens reported as an average value of three test samples. the peak load f is reported in kn/m as a force per unit width based on the 80 mm width and strain ε defined as the elongation of the specimen at peak load relative to the original length of the specimen. 3. stone masonry specimens construction and test setup five specimens were constructed: three walls and two columns. the dimensions of the specimens were 0.35 x 0.25 x 1.2 m (width x (a) (b) figure 2. a) carbon fiber textile, and b) glass fiber textile. figure 3. photograph of pure tensile tests of textile specimens. 28 strengthening of historical stone masonry buildings in oman using textile reinforced mortars   25   depth x height). a professional stone mason was hired to construct the specimens. the walls and columns were constructed using an assembly of limestone bonded by sarooj mortar. the specimens were constructed over a wooden platform covered by a polyethylene sheet for easy mobilization of the specimens during testing. they were constructed vertically in two phases: the layers of stones and mortar constructed to a height of 0.6 m and then left for the mortar to set to avoid squeezing the mortar at the lower layers. stones were cut into appropriate sizes and arranged in layers, and mortar was used to bind them together (fig. 4). the constructed specimens were left to air cure for 60 days. table 2. textile tensile properties. textile f (kn/m) ε (%) carbon 35 4 glass 25 6 (a) (b) figure 4. (a) construction of stone masonry specimens, (b) specimens air curing after construction completion. 3.1 strengthening of wall specimens after the specimens had been cured for almost 30 days, walls were strengthened with a bonding textile on the tension side. carbon fiber and glass fiber textiles were used in combination with two matrix mortars (table 3). the specimens were cleaned and rubbed to even the surface and increase the bond with the matrix mortar or plaster. the surfaces to be plastered were sprayed with water prior to plastering to maintain the water content in the plaster mortar. after the face of interest had been prepared, a layer of plaster mortar of 1–2 cm was applied, and then the textile was bonded on the surface, pushed into the plaster layer, and surfaced with a thin plaster layer to fix it in place. then a second layer of mortar of 1–1.5 cm was applied. the variation in the plaster or mortar thickness was due to the uneven surface of the wall. figure 5 shows a schematic representation of the strengthening procedure. the specimens after application of the textile were kept for 28 days under air curing at laboratory temperatures before testing. 3.2 strengthening of column specimens two columns were constructed: one specimen was used as a reference specimen while the other was strengthened by fully wrapping carbon fiber textile around the perimeter of the whole column (table 3). carbon has a higher stiffness (stress/strain ratio) compared to glass. the confinement effect is more efficient using carbon fiber as the fibers reach higher stresses at relatively small strain compared to glass fiber. the table 3. specimens’ designation and description. designation description wgm1 wall strengthened by glass fiber textile using dune sand mortar (m1) wgm2 wall strengthened by glass fiber textile using normal sand mortar (m2) wcm2 wall strengthened by carbon fiber textile using mortar (m2) cc control column ccm2 fully wrapped column using carbon textile with mortar (m2) 29 a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   30   figure 5. schematic representation of strengthening procedure for walls. strengthening of the column followed a similar procedure as in the wall specimens, except that for the column the process was done on the four sides of the column. a first layer of plaster mortar in the range of 1–2 cm was applied on two adjacent faces, and then the textile was bonded on the plastered area. it was pushed into the plaster layer and surfaced with a thin plaster layer to fix it. the same procedure was repeated for the other two faces. the textile was overlapped on the first face to the center of this face to ensure full confinement around the column. a second layer of plaster of 1–2 cm was then applied around the perimeter of the column. the sequence of bonding the textile is shown in fig. 6. the column was then capped on top with the same mortar to level the surface for later axial load application. the strengthened specimen was kept for another 28 days under air curing at laboratory temperatures before testing. 3.3 test setup the three wall specimens were tilted and flipped such that the strengthened face was at the bottom and transported to the testing area. all wall specimens were tested under out-ofplane four-point bending. the specimens were simply supported with span length in between supports of one meter (fig. 7). at the wall’s mid-span, two displacement transducers were placed to measure deflection at both sides of the wall. loading was then applied using a 250 kn compression actuator at a rate of 1 mm/min until failure. the columns were tested under axial compression. the columns were capped with a rubber sheet followed by a steel plate. a compression load was then applied using a 4000 kn utm at a rate of 2 mm/min until ultimate load was reached. the load and displacement at the top of the column were recorded. the test setup is shown in fig. 8. 4. test results and discussion of strengthened stone masonry elements 4.1 strengthened wall specimens the load versus mid-span deflection for the strengthened walls is shown in fig. 9. as can be noted, there are multiple load peaks in each curve in all strengthened walls’ behaviors. all the wall specimens failed due to diagonal shear adjacent to the supports within the shear span (fig. 10). cracks developed diagonally between the support and the loading point and, were expanded until a separation of the masonry wall occurred. as diagonal cracks developed, a drop in the applied load was observed. then the load picked up again with further deflection with an increase in wall curvature causing cracks and gaps to come closer to enhancing the interlocking action of the limestone and further increase in the applied load. the diagonal shear failure in all walls was due to the weak sarooj jointing mortar, which has very low strength as discussed in section 2.3. no failure in the textile or the matrix mortar was observed. all textile types and matrix mortar performed well in terms of tensile and bond strength, and all failures were caused by the weak jointing mortar. nevertheless, all strengthened walls attained high strength and ductility when compared to plain or unreinforced masonry walls. the bending moment due to the loading scheme used in the test of the walls caused tensile stresses at the bottoms of the walls, which was entirely resisted by the textile layer at the bottom of the strengthened wall. however, unreinforced masonry wall will fail when subjected to tensile stresses as a result of plain masonry layer first plaster layer textile layer second plaster layer strengthening of historical stone masonry buildings in oman using textile reinforced mortars 31 figure 6. stages of column strengthening: (a) wetting of the masonry surface replaced by, (b) applying the first mortar layer, (c) bonding the textile, (d) fixing the textile with thin mortar layer, (d) applying the second mortar layer, and (f) finishing up the surface and capping the column.   figure 7. walls test setup. (a) (b) (c) (d) (e) (f) a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   32 figure 8. columns test setup. figure 9. load vs. mid span deflections for the tested walls. mid span deflection (mm) l o ad ( k n ) strengthening of historical stone masonry buildings in oman using textile reinforced mortars   33   (a) (b) (c) figure 10. shear failure: (a) wall wgm1 replaced by, (b) wall wgm2, and (c) wall wcm2. a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   34 out-of-plane or a bending load since the jointing mortar cannot resist tensile stresses. the mass of each wall was measured to be approximately 250 kg on average, resulting in a self-weight of 2.45 kn. therefore, an unreinforced wall was assumed to fail in a brittle manner even under its own weight when subjected to tensile stresses due to out-of-plane loading. it must be mentioned here that no control specimen in the form of an unreinforced wall was tested as it was already known that it would have crumbled into pieces due to tensile stresses once it was tilted to one side (i.e. laid horizontally) as was the case in testing the strengthened walls. as shown in fig. 9, wgm1 reached an ultimate load of 6.95 kn corresponding to a 25.2-mm mid-span deflection. wgm2 reached an ultimate load of 7.05 kn corresponding to a mid-span deflection of 8.8 mm. wcm2 reached an ultimate load of 8.35 kn at a mid-span deflection of 20.3 mm. the overall deflections at failure in the three walls specimens were 28 mm, 32 mm, and 35 mm for wgm2, wgm1, and wcm2, respectively. each strengthened wall was able to reach an ultimate load in addition to its own weight of 2.45 kn. compared to an unreinforced wall which, it was assumed, would have failed under its own weight of 2.45 kn, the increase in ultimate load capacity with respect to the unreinforced wall were 383%, 387%, and 440% in walls wgm1, wgm2, and wcm2, respectively. the same argument can be made about the deflection of strengthened walls compared to unreinforced walls, which would have failed in a brittle manner without any deformation under just its own weight. it must be mentioned here that given the fact that the sarooj jointing mortar has better strength, the performance of the strengthened walls was expected to be much better. as reported by alal-rawas et al. (1998), the sarooj produced at other locations in oman possessed better properties with much higher strength than the sarooj from nakhal and wadi al mawal. the results obtained in this study clearly show the effectiveness of this strengthening technique even with very low-strength sarooj jointing mortar. it was noted that the cracks in wall wgm1 started at a lower load of 5.3 kn compared to 7 kn when the cracks appeared in wall wgm2 (figure 10 a and b). in addition, wall wcm2 resisted a slightly higher load than wgm2. nonetheless, the main failure was due to shear in the jointing mortar and, therefore, it is very difficult to make a fair comparison on the effect of the matrix mortar m1 and m2 as well as the effect of the glass fiber textile in comparison with the effect of carbon fiber textile. in other words, both strengthening schemes would work since the weakness was in the jointing mortar rather than the trm layer. 4.1 strengthened column specimens column specimens were tested under axial compression. the load versus column top displacement is shown in fig. 11. the control column (cc) reached an ultimate compression load of 37.30 kn at a displacement of 8.8 mm (fig. 11 and table 4). initially the response of the column showed an increase in displacement without a corresponding increase in load. this behavior may be attributed to the low-strength sarooj jointing mortar. the load then picked up once the jointing mortar between the stone layers could not be further compressed and the stone layers became effective in the load carrying mechanism. upon reaching the ultimate load, longitudinal cracks started to develop in the column sides as a result of the lateral tensile stress or poisson’s effect (fig. 12). the cracks then enlarged and expanded until separation in the masonry column occurred and portions of the columns collapsed, causing brittle failure of the column. the strengthened column (ccm2) showed superior performance from the start due to the confining effect of the textile reinforced mortar (trm) jacket around the column (fig. 11). it has been reported by many studies (saadatmanesh et al. 1994; shirmohammadi et al. 2015) that using steel jackets or frp wrapping around reinforced concrete column leads to a multifold increase in the axial load in table 4. summary of test results. designation maximu m load (kn) deflection @max. load (mm) maximum deflection (mm) wgm1 6.95 25.2 28 wgm2 7.05 8.8 32 wcm2 8.35 20.3 35 cc 37.3 8.8 15 ccm2 262.8 26.2 35 strengthening of historical stone masonry buildings in oman using textile reinforced mortars 35 figure 11. load versus top displacement for columns cc and ccm2. (a) (b) figure 12. column cc: (a) longitudinal cracks development, and (b) at failure of column. capacity due to the hoop stresses developed by the jacket around the column which applies a confining stresses responsible for an increase in the axial stress capacity. column ccm2 reached an ultimate compression load of 262.8 kn at a corresponding top displacement of 26.2 mm. the failure was initiated by cracks which developed at the column sides followed by the outer plaster layer peeling off. the inner plaster layer then started to crumble at the top top displacement (mm) a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood 36 (a) (b) figure 13. column ccm2; (a) outer mortar layer peeled off; (b) trm jacket opened up at the overlapping and failure of the column. of the column as the masonry expanded in the transverse direction and due to the high compressive stress at this stage. the textile started to rupture at one of the column edges in a localized manner due to stress concentration at the corners. the trm jacket finally debonded at the overlap, causing the wrapping to open up and leading to the column’s partial collapse (fig. 13). comparing the response of the cc and the ccm2 shows a considerable improvement in both the axial load capacity as the ccm2 was seven times that of the cc and reflected a 600% increase in the axial load capacity. similarly, the displacement at the ultimate load attained in ccm2 was 3.2 times that of the cc, or a 216% increase, indicating an increase in the ductility of the strengthened column. due to the failure of the wrapping at the overlapping part, it is recommended to extend the overlap to cover the entire face of the column in which the wrapping of the textile has started. 5. conclusion the presented experimental work is a pilot study to investigate the efficiency of a new strengthening technique using textilereinforced sarooj mortar for the rehabilitation of plain stone masonry structures. the study included strengthening plain stone masonry walls for out of plane loading and full wrapping of stone masonry columns subjected to axial compression. based on the presented experimental work, the following remarks are drawn:  when subjected to out-of-plane bending, all the wall specimens failed due to diagonal shear failure, which can be attributed to the weak strength of the jointing sarooj mortar.  all strengthened masonry walls were able to resist 3–4 times their own weight or a 300–400% increase in the load-carrying capacity compared to unstrengthened masonry walls.  all strengthened walls showed a ductile behavior before failure compared to the brittle failure of a plain masonry wall.  plain stone masonry columns failed in a brittle manner because of the low strength strengthening of historical stone masonry buildings in oman using textile reinforced mortars   37   jointing mortar and due to lateral tensile stresses caused by poisson’s effect.  strengthening masonry columns by full wrapping with textile reinforced mortar resulted in a 600% and 130% strength and ductility improvement, respectively, compared to the reference column. references al-mashani, a (1997), traditional construction in dhofar. 1st ed, in arabic. al-rawas a, hago a, corcoran t and al ghafri k (1998), properties of omani artificial pozzolana (sarooj). applied clay science 13: 275–292. al-rawas a and hago a (2006), evaluation of field and laboratory produced burnt clay pozzolans. applied clay science 31: 29– 35. american society for testing and materials, astm c 114-13, standard test methods for chemical analysis of hydraulic cement. american society for testing and materials, astm c 25-11, standard test methods for chemical analysis of limestone, quicklime and hydrated lime. american society for testing and materials, astm c51, 2011. standard terminology relating to lime and limestone (as used by the industry). american society for testing and materials, astm c618 (1980), specifications for fly ash and raw or calcined natural pozzolana for use as a mineral admixture in portland cement. 375–378. american society for testing and materials, astm d5034–09, 2013, standard test method for breaking strength and elongation of textile fabrics (grab test). american society for testing and materials, astm d6473, 2010, standard test method for specific gravity and absorption of rock for erosion control. american society for testing and materials, astm d7012, 2004, standard test methods for compressive strength and elastic moduli of intact rock core specimens under varying states of stress and temperatures. bernat-maso e, escrig c, aranha c and gil l (2014), experimental assessment of textile reinforced sprayed mortar. construction and building materials 50: 226–236. ehsani m r, saadatmanesh h, al-saidy a (1997), shear behavior of urm retrofitted with frp overlays. journal of composites. for construction 2(1): 17-25. hago a, al-rawas a (1997), properties of the omani sarooj. engineering journal of university of qatar 10: 81-91. hago a, al-rawas a, al-sidairi a (2002a), effect of the fineness of artificial pozzolana (sarooj) on the properties of limepozzolana mixes. science and technology 7: 251-258. hago a, al-rawas a, al-riyami a (2002b), effect of varying cement content and curing conditions on the properties of sarooj. building and environment 37: 45-53. harajli m, elkhatib h, san-jose j (2010), static and cyclic out-of-plane response of masonry walls. journal of materials in civil engineering 22(11): 1171-1180. kolsch h (1998), carbon fiber cement matrix (cfcm) overlay system for masonry strengthening. journal of composites for construction 2:105-109. ministry of heritage and culture, office of deputy minister for ministry affairs (1994), forts and castles in oman. sultanate of oman: studies department, in arabic. papanicolaou c, triantafillou t, lekka m (2011), externally bonded grids as strengthening and seismic retrofitting materials for masonry panels. construction and building materials 25: 504–514. pinho f, luco v, baiao m (2014), rubble stone masonry walls strengthened by threedimensional steel ties and textile reinforced mortar render under compression. international journal of architectural heritage 8(5): 670–689. saadatmanesh h, ehsani m, li m (1994), strength and ductility of concrete columns externally reinforced with fiber composite straps. aci struct. journal. 91(4): 434–447. shirmohammadi f, esmaeily a, kiaeipour z (2015), stress–strain model for circular concrete columns confined by frp and conventional lateral steel. engineering structures 84: 395–405. triantafillou t (1998), strengthening of masonry structures using epoxy-bonded a.h. al-saidy, a.w. hago, s. el-gamal and m. dawood   38 frp laminates. journal of composites for construction 2: 96-104. varshney rs (1982), concrete technology. oxford: ibh company. civl051122.qxd the journal of engineering research vol. 4, no.1 (2007) 23-36 ____________________________________________ *corresponding author’s e-mail: alnuaimi@squ.edu.om direct design of t-beams for combined load of bending and torsion a. s. alnuaimi*, k.s. al-jabri and a.w. hago department of civil and architectural engineering, sultan qaboos university, p.o. box 33, al khodh 123, muscat, oman received 22 november 2005; accepted 24 april 2006 abstract : tests were conducted on five reinforced concrete t-beams subjected to combined load of bending and torsion. elastic stress field in conjunction with nielsen's 2d yield criterion for reinforced concrete subjected to in-plane forces were used in the direct design method for the design of reinforcement. the beam dimensions were: flange width = 600mm, flange thickness = 150mm, web width = 200mm, total depth = 600mm and beam length = 5.2m. required reinforcement calculated using the direct design method was compared with the aci and bsi codes. it was found that the direct design method requires longitudinal reinforcement similar to the aci code but less than the bsi code. in the transverse direction, the direct design method requires much less reinforcement than both codes. the main variable studied was the ratio of the maximum twisting moment to the bending moment which was varied between 0.6 and 1.5. good agreement was found between the design and experimental failure loads. most of the longitudinal and transverse steel yielded or reached near yield stress when the design load was approached. all beams failed near the design loads and undergone ductile behaviour until failure. the results indicate that the direct design method can be successfully used to design reinforced concrete t-beams for the combined effect of bending and torsion loads. keywords: t-beams, direct design method, bending, torsion, reinforced concrete notations ac gross area of the whole cross-section ao area of the cross-section enclosed by the shear flow centreline ax,ay steel areas per unit width in x and y directions respectively e young's modulus of elasticity fcu concrete cube compressive strength fsx steel stress in the longitudinal direction fsy steel stress in the transverse direction agƒàd’gh aéæëf’g ∫éªm’ á°vô©ÿg áø°ûddg äghp ¿gõ«é∏∏d ô°téñÿg º«ª°üàdg ƒég ~mgƒdg~ñyh …ôhé÷g ¬ø«∏n h »ª«©ædg »∏y áá°°uuóóÿÿggáaé°vg »wé£e oé¡lg π≤m ωg~îà°sg ” ~bh .agƒàdgh aéæëfeg øe áfƒµe ∫éªmg áyƒªû â©° ncg áø°t äghp áë∏°ùÿg áfé°sôÿg øe ¿gõ«l á°ùªn ≈∏y ägqéñàn’g âjôlcg : oé©hcg âféch .í«∏°ùàdg ~j~m ä髪c ≈∏y ∫ƒ°üë∏d ô°téñÿg º«ª°üàdg á≤jôw ‘ ájƒà°ùÿg iƒ≤dg òkéàd á°vô©ÿg áë∏°ùÿg áfé°sôÿg ´ƒ° îh ¢uéÿgh oé©h’g »féæk ú°ù∏«f qé«©ÿ áfqé≤e â“h .«∏e 5200 õfé÷g ∫ƒwh «∏e 600 »∏µdg ∂ª°ùdgh «∏e200 iôjƒdg ¢vôyh «∏e 150 áø°ûdg ∂ª°sh «∏e 600 áø°ûdg ¢vôy :»j’éc ¿gõ«÷g áhƒ°ùùg ‹ƒ£dg √霒g ‘ í«∏°ùàdg ~j~m ä髪c ¿g ~lhh .êé£jèdgh »µjôe’g øjoƒµdg ᣰsgƒh ö°ù– »àdg ∂∏àh ô°téñÿg º«ª°üàdg á≤jô£h áhƒ°ùùg í«∏°ùàdg ~j~m ä髪c ö∏£àj ô°téñÿg º«ª°üàdg á≤jôw ¿eéa á°vô©à°ùÿg á¡÷g ‘h .êé£jèdg oƒµdg é¡ñ∏£àj »àdg ∂∏j øe πbg é¡æµd »µjôe’g oƒµdg é¡ñ∏£àj »àdg ä髪µdg πké“ ô°téñÿg º«ª°üàdg á≤jô£h â檰 jh aéæëf’g ωõy ¤g ≈°üb’g agƒàd’g ωõy áñ°ùf ¿éc »°ù«fôdg ò¨àÿg ¿eéa á«dé◊g á°sgq~dg ‘h .êé£jèdg h »µjôe’g ¿goƒµdg áñ∏£àj é‡ òãµh πbg í«∏°ùàdg ~j~m øe ä髪c éªc é¡à°sgqo â“ »àdg ¿gõ«é∏d qé«¡f’g ∫éªmgh º«ª°üàdg ∫éªmg úh iòñc äéaóàng oƒlh ω~y ¤g á°sgq~dg â∏°uƒj ~bh 1^5 ¤g 0^6úh áñ°ùædg √ò¡d io~©àe 骫b á°sgq~dg ~æy äqé¡fg ~b ¿gõ«÷g πc ¿g éªc ,º«ª°üàdg ∫éªmg øe ∫éªm’g üî≤j ée~æy ´ƒ° ÿg oé¡lg üqéb hg π°uh ~b ¢vô©à°ùÿgh ‹ƒ£dg √霒g ‘ í«∏°ùàdg ~j~m ö∏zg ¿g ~lh á«fé°sôÿg ¿gõ«÷g º«ª°üj ‘ ô°téñÿg º«ª°üàdg á≤jôw ∫骩à°sg øµá áfg ¤g á°sgq~dg â°ü∏n ~bh .égqé«¡fg ~æy ≈àm ¿ôe ∑ƒ∏°s äghp âféch º«ª°üàdg ∫éªmg øe üî≤j ∫éªmg .agƒà∏dgh aéæëf’g ∫éªm’ á°vô©ÿgh áø°ûdg äghp áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.áë∏°ùÿg áfé°sôÿg ,agƒà∏dg ,aéæëfg ,ô°téñÿg º«ª°üàdg á≤jôw ,áø°ûdg äghp ¿gõ«÷g : 24 the journal of engineering research vol. 4, no.1 (2007) 23-36 1. introduction the present code design equations for concrete structures (especially for shear and torsion) are largely based on experimental results which lead to continuous changes in the code. calculation of reinforcement for each load is done separately and then summed algebraically which leads to conservative design. for combined loading, several interaction theories based on ultimate strength criterion have been developed (ie. theories on truss-analogy and skew-bending), but the major weakness of this approach lies in the fact that it does not present a general design procedure capable of being applied to design for a given general stress state. many researchers investigated the behaviour of reinforced and pre-stressed concrete t-beams under different loading conditions (ie. bending, torsion and shear) and the contribution of the sections' elements (ie. flange and web) on resisting applied forces has been also reported. the experimental behaviour of twenty six reinforced concrete t-beams without web reinforcement under combined bending, shear and torsion was reported by farmer and ferguson, (1967). torque-to-shear ratios of 1.5, 3 and 6 were investigated. results indicated that span length and cross-sectional warping was not found to influence the behaviour greatly. it also suggested that torsion lowered the shear strength of the beam much less than ordinary plastic theory would suggest. nawy and potyondy (1971) studied the flexural cracking behaviour of pre-tensioned iand t-beams. they tested twenty two iand t-beams with a span of 9-ft (2.74m) and pre-stressed with 1/4 in (6.35mm) diameter 7-stranded wires. from the results, it was concluded that the confining reinforcement had no influence on the magnitude of load at first crack and tsections cracked at a lower load than i-sections. kirk and loveland (1972) observed, after testing eighteen unsymmetrical reinforced t-beams under bending and torsion, that the location of longitudinal steel at the top or the bottom of the section can have significant influence on the torsional resistance. the effect of reinforced flanges on the torsional capacity of reinforced concrete t-beams was investigated by zararis and penelis, (1986). eighty two tbeams with various flange widths and percentages of reinforcement were tested in combined torsion and bending. results showed that there is significant contribution from the reinforced flanges to the torsional capacity of the tbeams especially when torsion prevailed. experimental studies by kotsovos et al. (1987) also showed that the shear resistance of t-beams with shear-span-to-depth ratios greater than 2.5 is provided by the flange and not as widely accepted by the web. gurfinkel and fonseca (1996) investigated the aci code requirement for additional ductility in the t-beam sections and reported that "there is little justification for code-imposed additional demand on minimum ductility of t-beams sections". recently experimental investigation was carried out by rahal and collins, (2003) in order to evaluate the aci and aashto-lrfd design provisions for combined torsion and shear. it was found that the aci provisions give very conservative results if the recommended value of 45degree is used for the inclination of the compression diagonals. fy yield strength of longitudinal steel fyv yield strength of transverse steel i moment of inertia of the section l.f. load factor = (ti/td + mi/md)/2 at any load increment i le/ld failure load ratio = (te/td + me/md)/2 for the last (failure) load increment n1 concrete resisting force in the principal direction 1, (=σ1 t) n2 concrete resisting force in the principal direction 2, (=σ2 t) nx applied in-plane force per unit length in the x direction on a element with thickness t, (=σxt) nxy applied in-plane shear force per unit length on a element with thickness t, (= τxyt) ny applied in-plane force per unit length in the y direction on a element with thickness t, (=σyt ) nsy resisting forces in x direction steel (= axfsx) nsy resisting forces in y direction steel (= ayfsy) t thickness of the element td, md design torsion and bending moment respectively te, me experimentally measured torsion and bending moment at failure ti, mi experimentally measured torsion and bending moment at load increment i y distance from the neutral axis to the location where stress to be calculated σ1 concrete principal stress in direction 1 σ2 concrete principal stress in direction 2 σx, σy applied normal stresses in x and y directions respectively ε measured steel strain at any load increment i εy longitudinal steel yield strain (= fy/e) εyv transverse steel yield strain (= fyv/e) τxy applied elastic shear stress 25 the journal of engineering research vol. 4, no.1 (2007) 23-36 it is clear from the above literature and others that large area of disagreement exists among the researchers on the design of t-beams when shear and torsion are involved. one reason for this disagreement is the fact that the design equations used in the code are not based on sound theoretical principles like the theories of mechanics (ie. theory of plasticity). the direct design method, which is based on the lower bound theorem of the theory of plasticity, presents a solution to some of the above pitfalls. this procedure satisfies the fundamental requirements of classical theory of plasticity (equilibrium condition, yield criterion and mechanism condition) and it reduces the ductility demand assumed for metals. in contrary to the existing codes of practice, this method precludes the use of empirical equations and therefore, it is less vulnerable to changes and may form a basis for a unified design method. previous work by bhatt and ebireri (1989) had shown that the 'direct design method' could be applied to the design of hollow and solid beams with square cross-section subjected to combined bending and twisting moments. alnuaimi and bhatt (2004a, b) used the direct design method for the design of eight reinforced concrete square hollow beams subjected to combined bending, torsion and shear. the main variables studied were the ratio of twisting moment to the bending moment and the ratio of the shear stress due to torsion to the shear stress due to shear force. they concluded that the direct design approach used, gave satisfactory results with beams failing within the range of the design loads and behaving in a ductile manner until failure. many other researchers have successfully used the direct design method for the design of different types of structures (kemp 1971; nielsen 1974; morley 1977; memon 1984; el-nuonu 1985; hago and bhatt 1986; abdel-hafez 1986; el-hussein 1994; bhatt and bensalem 1996a, b; bhatt and mousa 1996; and elarabi 1999). in this research, the direct design method was used for the design of five reinforced concrete t-beams. the beams were subjected to combined load of bending and torsion. the required reinforcement obtained using the direct design method was compared with that obtained using the aci (2002) and the bsi (1997) codes. 2. research significance a theory of plasticity-based, design procedure for the design of reinforced concrete t-beams subjected to combined loading of bending and torsion is presented as replacement for the empirical-based methods used by the codes of practice. the required reinforcement from this procedure, direct design method, is compared with the required reinforcement using the aci and the bsi codes. test results from five t-beams designed using the direct design method are presented. 3. test beams five reinforced concrete t-beams were designed for combined action of bending and twisting moments using the direct design method. the beams were 5200 mm long with overall depth of 600 mm. the dimensions of the tsection were 600 x 150 mm flange and 450 x 200 mm web. table 1 shows loads that each beam was designed to resist. the main variable in this series was the ratio of the design twisting moment to the bending moment, td /md, which varied between 0.6 and 1.5. figure 1 shows the geometry and dimensions of the tested beam. 4. comparison between the aci and bsi codes and the direct design method table 2 shows the required reinforcement for the tbeams used in this research based on the aci and bsi codes and the direct design method, ddm. it can be seen that the direct design approach leads to longitudinal steel requirement close to the aci code while the bsi code is more conservative. in the transverse reinforcement, the direct design procedure required less reinforcement than both codes in most cases, especially the bsi code. the direct design method was used for the design of this research beams. 5. design of t-beams the following steps were used for the design of tbeams, and a numerical example is given in appendix a. detailed derivation of equations can be found elsewhere alnuami and bhatt, (2004a, b). td m d t md d/ beam no. kn.m kn.m ratio tb1 36 60 0.60 tb2 75 100 0.75 tb3 50 60 0.83 tb4 72 72 1.00 tb5 108 72 1.50 table 1. design load combinations 150 a a b b 600 4000 600 60 0 15 0 600 200 60 0 sec. a -a 600 60 0 sec. b-b all dimensions in mm figure 1. beam geometry and dimensions 26 the journal of engineering research vol. 4, no.1 (2007) 23-36 the required reinforcement at mid-span was provided over a test span length of 1200mm in the middle of beam. figure 3 shows typical reinforcement arrangements in a tested t-beam and fig. 4 shows details of provided reinforcement in the test span of each beam. the solid dots in the figure indicate the locations of bars, which were strain-gauged. to ensure failure at mid-span, additional reinforcement was provided in the regions outside the test span. further, to reduce the stresses due to distortion of the cross section and produce good griping mechanism in the torsional arms, the end section was made 600x600 mm solid square for a length of 400mm from each end. 6. casting of beams after the steel cage was erected including all longitudinal and transverse reinforcement, the beam was cast in a 1. calculate the ultimate (factored) design bending moment m d and torsion t d at the section based on the loading condition, support locations and geometry of the beam. 2. divide the cross -section into different levels (regions) as shown in fig. 2. 3. find the neutral axis , the moment of inertia and the area enclosed by the shear flow centreline . 4. calculate normal stress xσ due to applied load at the centre of each region (level) using elastic stress distribution, i/myx =σ . 5. calculate the shear stresses due to torsion, ota2 t xy =τ . 6. assume 0y =σ (no out of plane bending is considered). 7. calculate the ratios xy x τ σ and xy y τ σ for each region. 8. select design equat ions as explained in the numerical example (appendix a) . 9. calculate the required reinforcement. aci bsi ddm beam no. as (mm2) asv(mm 2/m) as (mm2) asv(mm 2/m) as (mm2) asv(mm 2/m) tb1 1136 914 1325 978 1091 606 tb2 1903 1848 2479 1978 1908 1226 tb3 1484 1269 1746 1358 1516 842 tb4 1841 1741 2457 1956 1932 1155 tb5 2633 2611 3542 2934 2897 1732 table 2. required reinforcement using the aci, bsi and ddm methods a b c d e f g h i j k l figure 2. regions (levels) in the cross-section see x -section y10 @ 100 mm y10 @100 mm 2000 mm 1200 mm 2000 mm 5200 mm test span figure 3. typical arrangement of reinforcement 2y12 2y12 2y12 3y12 2y10 2y12 y 10 @ 26 0 tb-1 y 10 @ 12 8 2y12 2y12 3y12 3y12 4y12 2y12 2y12 tb-2 y 10 @ 18 6 2y12 2y12 2y12 4y12 2y12 2y12 tb-3 y 10 @ 13 6 2y12 2y12 3y12 3y12 2y12 4y12 tb-4 y 10 @ 90 3y12 4y12 4y12 4y12 2y12 4y12 tb-5 figure 4. provided reinforcement at test span 27 the journal of engineering research vol. 4, no.1 (2007) 23-36 specially designed wooden and glass reinforced plastic (grp) formwork. one day after casting, the sides of the mould were disassembled leaving the bottom part bellow the beam. six concrete cubes (100mm) and six cylinders (150 x 300mm) were cast with each beam from the concrete used in the beam. a vibrating table was used for compaction of the cubes and cylinders while the beam was vibrated using an external vibrator attached to the shuttering. the beam and the cubes and cylinders were cured using wet hessian cloth for about one week and then in room temperature for another three weeks. all cubes and cylinders were tested on the day the beam was tested to determine the cube and cylinder compressive strengths and cylinder tensile strength. only 10mm and 12mm high yield steel bars were used as longitudinal reinforcement and 10mm bars were used for stirrups. from the tensile test results, the average yield stress of steel was 545n/mm2 for longitudinal reinforcement and 513n/mm2 for stirrups. ordinary portland cement (opc), fine aggregate, coarse aggregate (10mm) and fresh water were used in the concrete mix. table 3 shows the average of some measured material properties. 7. experimental set-up figure 5 shows the testing rig with a typical beam installed. the installation process required inserting the beam square ends into the torsional arms (fig. 6) which are attached to shafts and bearings resting on pedestals (fig. 7). the torsional arm ends are hooked to the strong floor by inextensible cables with hydraulic actuators and load cells. on the top face of the beam, a 1200mm long spreader was installed. the spreader was centred at the mid-span of the beam with its centreline coinciding with beam centreline. the bending moment in the middle 1200mm test span is acquired by applying a point load at the middle of the spreader. the torsional moment which is constant over the entire length of the beam, was applied through torsional arms clamped at the ends. the load from the spreader was transferred to the beam through two perpendicular rollers in one side while a roller and a pin were installed at the other side. this support and loading arrangements produced a simply supported condition allowing free rotation about the longitudinal axis of the beam and horizontal displacements with a rigid body motion being prevented by the pin on the top face. the test span region is mainly subjected to bending and twisting moments while outside the test span shear is added to the twisting and bending moments. strains in the reinforcement were monitored using electrical resistance strain gauges. figure 8 shows the legend cuf yf yvf beam no. n/mm² n/mm² n/mm² tb1 36 500 500 tb2 30 596 515 tb3 45 500 500 tb4 29 565 525 tb5 30 565 525 table 3. measured average material properties figure 5. test rig with a typical beam installed figure 6. torsional arm while testing figure 7. close lookup of end support and torsional arm with actuator and load cell cl fl11 fl12 fl13 ba c dfs1 fs2 cl rl11 rl12 rl13 f e g h rs1 rs2 (a) front side (b) rear side figure 8. legend for strain gauge locations in longitudinal and transverse steel (see table 4) 28 the journal of engineering research vol. 4, no.1 (2007) 23-36 and arrangement of the strain gauges in the longitudinal and transverse reinforcement and table 4 gives the location of strain gauges for each beam. at each specified location on each bar, two strain gauges were installed opposite to each other. figure 9 shows typical locations of the longitudinal bars and table 5 shows vertical distances between the longitudinal bars. linear variable differential transducers (lvdt) were used to measure the beam deflection (fig. 10). the concrete surface strains were measured by demec (digital) gauges. all strain gauges and lvdts were hooked to a data logger. the crack widths were measured using a crack width measuring microscope. for each experiment, the design load was divided into load increments, 10% of the design load for each of the first three increments and 5% for each of the following increments until failure. the first step was to zero all load cells and record instrument readings with minimum possible loads on the model. flexural loads were normally applied before torsional load at each load level. to allow for stable deformation to take place after each load increment, an interval of about two minutes was used before recording the instrument readings. after the steel strains and displacement readings were recorded, cracks were marked and demec readings were taken before the next load was applied. the beam was considered to have collapsed when it could resist no more loads. this usually happened after a major crack spiralled around the beam near the mid-span dividing the beam into two parts connected by the longitudinal reinforcement. 8. experimental observations this section summaries the observed behaviour in the test span region of the tested t-beams at significant stages in the behaviour. the angle of crack was measured from the horizontal axis. vertical displacement was measured at the mid-span at the bottom of the web. strain ratios in the bottom layer of the longitudinal steel and in the transverse steel of mid-span are presented. the quoted load factor is a percentage of the experimentally measured load at each increment to the design load, l.f. = (mi/md + ti/td)/2 including the self-weight. the failure load is a percentage of the experimentally measured failure load to the design load, le/ld = (me/md + te/td)/2. space a b c d e f g h beam mm mm mm mm mm mm mm mm tb1 260 260 130 130 260 260 130 130 tb2 128 128 64 64 128 128 64 64 tb3 186 186 93 93 186 186 93 93 tb4 136 136 68 68 136 136 68 68 tb5 90 90 45 45 90 90 45 45 vertical dist. for the stirrup’s strain gauges is 300mm from the bottom of the web. table 4. location of strain gauges (see figure 8) 1 2 3 4 5 6 7 figure 9. typical vertical clear spaces between longitudinal bars (see table 5) 2 test 1 3 4 5 6 7 8 9 a b c mid span a: bottom side transducers b: front sdie transducers c: rear side transducers figure 10. locations of lvdt location 1 2 3 4 5 6 7 beam mm mm mm mm mm mm mm tb1 20 100 100 120 20 20 tb2 20 80 80 100 100 20 20 tb3 20 100 100 120 20 20 tb4 20 100 100 120 20 20 tb5 20 100 100 120 20 20 table 5. vertical clear distances between longitudinal bars (see fig. 9) test span 29 the journal of engineering research vol. 4, no.1 (2007) 23-36 the average strain ratio in the longitudinal ε/εy or transverse ε/εyv reinforcements is the average of the measured strain at each load increment to the yield strain. at each strain-gauge location on the bar, the average of the readings from the two opposite strain gauges is considered. tb1: td=36knm, md=60knm, td/md=0.6 (1) the first few cracks were noticed at 45% of load at the bottom of the web near mid-span. the cracks were about 80o inclined. more cracks developed by the increase of the load until about 90% of the design load with angle of inclination being reduced to an average of 50o and average width of crack of about 0.4 mm. after this stage, few major inclined cracks started out side the test span. near failure, prominent cracks spiralled round the beam meeting near the mid-span. the failure happened at 110% of the design load. the average crack width just before failure was about 0.6 mm. figure 11 shows the displacement at mid-span of the beam and figs. 12-15 show strain ratios in the longitudinal and transverse reinforcement. the average strain ratio reached near failure in the longitudinal steel was 0.94 and in the stirrups, the ratio was 0.92. figure 16 shows crack development in the test-span of the beam. tb2: td=75knm, md=100knm, td/md=0.75 (2) at 40% of load, several inclined hair cracks (about 60o) were noticed in the test span region close to the bottom of the web. with the increase of the load, many fine cracks developed until about 90% of load when two major cracks widened up (45o inclination) and spiralled around the beam causing failure at 105% of the design load. the average crack width near failure load was about 0.5 mm. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 5 10 15 20 25 30 deflection(mm) l .f . figure 11. deflection at mid-span (tb1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.4 0.8 1.2 ε/εy l. f. fl11 fl12 fl13 figure 12. front side longitudinal steel strain ratios (tb1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.4 0.8 1.2ε/εy l. f. rl11 rl12 rl13 figure 13. rear side longitudinal steel strain ratios (tb1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 ε/εyv l .f . fs1 fs2 figure 14. front side transverse steel strain ratios (tb1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 ε /ε yv l .f . rs1 rs2 figure 15. rear side transverse steel strain ratios (tb1) figure 16. crack development (tb1) 30 the journal of engineering research vol. 4, no.1 (2007) 23-36 figure 17 shows the displacement at mid-span and figs. 18-19 show the strain ratios in the longitudinal and transverse bars. the average strain ratio in the longitudinal bars 0.9 and in the stirrups was 0.61. figure 20 shows crack development in this beam. tb3: td=50knm, md=60knm, td/md=0.83 (3) the development of cracks in this beam was similar to that of beam tb2 with number of cracks being larger in tb3 than tb2. failure happened at 112% of the design load. figure 21 shows displacement at mid-span and figs. 22-25 show steel strain ratios in the longitudinal and transverse steel. figure 26 shows crack development in this beam. the average width of cracks near failure was 0.45 mm and the average strain ratio near failure was 0.91 in the longitudinal steel and 0.85 in the stirrups. 0 0.2 0.4 0.6 0.8 1 1.2 0 4 8 12 16 20 24 28 deflection(mm) l .f . figure 17. deflection at midspan (tb2) 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2ε/εy l .f . fl1 rl1 figure 18. longitudinal steel strain ratios (tb2) 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1εε//εεyv l .f . fs2 rs2 figure 19. transverse steel strain ratios (tb2) figure 20. crack development (tb-2) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 5 10 15 20 25 30 deflection(mm) l .f . figure 21. deflection at midspan (tb3) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4ε/εy l .f . fl11 fl12 fl13 figure 22. front side longitudinal steel strain ratios (tb3) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.4 0.8 1.2εε /εε y l .f . rl11 rl12 rl13 figure 23. rear side longitudinal steel strain ratios (tb3) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0ε/εy v l .f . fs1 fs2 figure 24. front side transverse steel strain ratios (tb3) 31 the journal of engineering research vol. 4, no.1 (2007) 23-36 tb4: td=72knm, md=72knm, td/md=1.0 (4) at 25% of load, few inclined cracks (about 45o) started in the lower part of the web. the number of cracks increased largely with the increase of the load but the width was smaller than in the previous beams. the beam failed at 100% of the design load with few cracks widened slightly larger than the others did. figure 27 shows displacement at mid-span and figs. 28-31 show steel strain ratios in the longitudinal and transverse steel. figure 32 shows crack development in this beam. the average width of cracks near failure was 0.4 mm and the average strain ratio near failure was 1.05 in the longitudinal steel and 0.94 in the stirrups. tb5: td=108knm, md=72knm, td/md=1.5 (5) this beam behaved in a similar manner like tb4 but 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0ε/εyv l .f . rs1 rs2 figure 25. rear side transverse steel strain ratios (tb3) tb-3 figure 26. crack development (tb3) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 2 4 6 8 10 12 14 deflection(mm) l .f . figure 27. deflection at midspan (tb4) 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 ε/ε y l .f . fl11 fl12 fl13 figure 28. front side longitudinal steel strain ratios (tb4) 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 ε/εy l .f . rl11 rl12 rl13 figure 29. rear side longitudinal steel strain ratios (tb4) 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2ε/εy v l .f . fs1 fs2 figure 30. front side transverse steel strain ratios (tb4) 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 ε/εyv l .f . rs1 rs2 figure 31. rear side transverse steel strain ratios (tb4) tb4 figure 32. crack development (tb4) 32 the journal of engineering research vol. 4, no.1 (2007) 23-36 with larger number and finer cracks. there was no major cracks but the beam failed by losing strength due to the large number of cracks at a load of 98% of the design load. figure 33 shows displacement at mid-span and figs. 34-35 show steel strain ratios in the longitudinal and transverse steel. figure 36 shows crack development in this beam. the average width of cracks near failure was 0.3 mm and the average strain ratio near failure was 1.04 in the longitudinal steel and 0.92 in the stirrups. 9. discussion of results in all beams, the longitudinal and transverse steel either yielded or reached nearly the yield strain. this is to be expected since the beams were designed according to the lower bound theory of plasticity. beams loaded dominantly in bending (ie. low td/md ratio) showed more ductile behaviour than those with large td/md ratios (table 7). in general, all beams showed ductile behaviour as expected for under-reinforced sections. furthermore, there was no crushing of concrete before yielding of steel. all beams failed near their design loads for all variations of td/md ratios (column 3 of table 6). beams loaded dominantly in bending (td/md <1) had vertical cracks started to develop first at the bottom of the web, then progressed as inclined cracks in succeeding load increments until reaching the flange at about 80% of the design load. beams loaded dominantly in torsion (td/md >1) had inclined cracks started at mid-depth in the web before they extend into the bottom of the web. generally, cracks formed earlier in beams loaded dominantly in torsion (td/md >1) than other beams. in these beams, cracking started at a load as early as 25% of design load, while in other beams, they started at about 40% of design. 0 0.2 0.4 0.6 0.8 1 0 4 8 12 16 20 24 28 deflection(mm) l .f . figure 33. deflection of midspan (tb5) 0 0.2 0.4 0.6 0.8 1 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 ε/εy l .f . fl11 fl12 fl13 figure 34. front side longitudinal steel strain ratios (tb5) 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1ε/εyv l .f . fs2 figure 35. front side transverse steel strain ratios (tb5) figure 36. crack development (tb5) load td/md le/ld beam ratio ratio tb1 0.60 1.1 tb2 0.75 1.05 tb3 0.83 1.12 tb4 1.00 1.00 tb5 1.50 0.98 average 1.05 table 6. experimental/design failure load ratios load at first crack average max. crack width near failure max. disp. near failure beam no. ratio mm mm tb1 0.45 0.6 29.6 tb2 0.4 0.5 25.3 tb3 0.4 0.45 27.6 tb4 0.25 0.4 13.3 tb5 0.25 0.3 25.5 average 0.35 0.39 24.26 table 7. load-crack and displacement measurement 33 the journal of engineering research vol. 4, no.1 (2007) 23-36 in beams where bending was dominant, the cracks in the test span were less in number and wider than in the beams where torsion was dominant. this is due to the fact that torsion was constant throughout the entire length while maximum bending moment existed over a small distance at the middle of the beam. 10. conclusions from the results shown in this paper, it can be concluded that the direct design method produces a safe design for reinforced concrete t-beams under combined loading of bending and torsion. most of the beams failed near design loads and undergone ductile behaviour until failure. the reinforcement yielded or reached near yield when the design load was approached. the direct design method required less longitudinal reinforcement than the bsi code and close to the reinforcement required by the aci code. it required less transverse reinforcement than both bsi and aci codes. references abdel-hafez laila, m., 1986, "direct design of reinforced concrete skew slab," phd thesis, university of glasgow, 1986. aci committee 318, "building code requirements for structural concrete (aci318m-02) and commentary (aci318rm-02)," american concrete institute, farmington hills, mich. 48333-9094, usa. ali alnuaimi and bhatt, p., 2004a, "direct design of hollow reinforced concrete beams -part i: design procedure," structural concrete journal, vol. 5(4), pp. 139-146. ali alnuaimi and bhatt. p., 2004b, "direct design of hollow reinforced concrete beams -part ii: experimental investigation," structural concrete journal, vol. 5(4), pp. 147-160. bhatt, p. and bensalem, a., 1996a, "behaviour of reinforced concrete flat slab over column support using nonlinear stress field design," structural engineering review, vol. 8(2-3), pp. 201 212. bhatt, p. and bensalem, a., 1996b, "use of non-elastic stress in the design of reinforced concrete deep beams," structural engineering review, vol.8(2-3), pp. 213 225. bhatt, p. and ebireri, j. o., 1989, “direct design of beams for combined bending and torsion,” stavebnicky casopis, vol. 37(4), pp. 249-263. bhatt, p. and mousa, j., 1996, "tests on concrete box beams under non-monotonic loading," structural engineering review, vol. 8(2-3), pp. 227 235. bs8110, 1997, "structural use of concrete, part1: code of practice for design and construction", british standard institution, 389chiswick high road, london, w4 4al. elarabi, h., 1999, "application of the direct design method on reinforced concrete beams subjected to combined torsion, bending and shear," building and road research journal, university of khartom, vol. 2, pp. 47-56. el-hussein, e. a., 1994, "finite element and direct design method in combined torsion, bending and shear of reinforced concrete," computational structural engineering for practice,” civil-comp ltd, edinburgh, scotland, pp. 165-171. el-nuonu, g. f. r., 1985, "design of shear wall-floor slab connections," phd thesis, university of glasgow, uk. farmer, l. e. and ferguson, p. m., 1967, "shear and torsion", aci journal, pp. 757-65. gurfinkel, g. and fonseca, f., 1996, "designing t-beams for less ductility than required by aci", aci structural journal, vol. 93(s12), pp. 116-127. hago, a. and bhatt p., 1986, "tests on reinforced concrete slabs designed by direct design procedure," aci journal, vol. 83-79(6), no.6, nov.dec., 1986, pp. 916-924. kemp, k. o., 1971, "optimum reinforcement in a concrete slab subjected to multiple loadings," publications of the international association of bridge and structural engineering, pp. 93 105. kirk, d. w. and loveland, n. c., 1972, "unsymmetrically reinforced t-beams subject to combined bending and torsion," paper no. 69-44, aci journal, pp. 49299. kotsovos, m. d., bobrowski, j., and eibl, j., 1987, "behaviour of reinforced concrete t-beams in shear," the structural engineer, vol. 65b(1), pp. 1-10. memon, m., 1984, "strength and stiffness of shear wallfloor slab connections," phd thesis, university of glasgow, uk. morley, c. t. and gulvanessian, h., 1977, "optimum reinforcement of concrete slab elements", proc. institution of civil engineers, vol. 63(2), pp. 441-454. nawy, e. g. and potyondy, j. g., 1971, "flexural cracking behaviour of pretensioned iand t-beams,” aci journal, pp. 355-360 nielsen, m. p., 1974, optimum design of reinforced concrete shells and slabs,” structural research laboratory, technical university of denmark, report nr.r44, pp.190-200. rahal, k. n. and collins, m. p., 2003, "experimental evaluation of aci and aashto-lfrd design provisions for combined shear and torsion," aci structural journal, vol. 100(3), pp. 277 282. zararis, p. d. and penelis, g. gr., 1986, "reinforced concrete t-beams in torsion and bending," aci journal, title no. 83-17, pp. 145-55. 34 the journal of engineering research vol. 4, no.1 (2007) 23-36 appendix example: design of t -beams using direct design method tb3: td=50 kn.m md=60 kn.m td/md =0.83 see figs: 1 and 2. • moment of inertia, i = 3 1 [600x225 2+200(600-225)3-753(600-200)] =5.7375x10 4mm4 • normal stress due to bending at the centroid of each level : i imy i =σ , i = level reference and y=distance between level centroid and neutral axis (column 3 of table a1). • assume óyi = 0, no out of plane bending (column 4 of table a1). fig. a3: levels (regions) in the cross section and normal stress. • gross-sectional area (fig. a1) : a=600x150+200x450=180000mm 2 • depth of neutral axis from top of flange: 600x150x150/2 + 200x450x375= (600x150+200x450) x x = 225mm for torsion assume 50mm effective w all thickness (fig. a2) : • area of section within the shear flow centerline, ao = 100x550 + 150x425 = 11.875x 10 4 mm2 25mm shear flow centerline ao fig. a1: geometry of cross -section fig. a2: shear flow centerline 60 0 45 0 x 600 150 200 x • calculate shear stre ss due to torsion, wall thickness 50mm (column 5 of table a1): 2mm/n21.4 410x875.11x50x2 610x50 ota2 t xy ===τ • calculate ó xi / ôxy and ó yi / ôxy (columns 6 and 7 of table a1). • select the design case from the boundary graph for nielsen’s design equations shown in fig. a4 (column 8 of table a1) . • calculate forces in the longitudinal direction per unit length: xynxnsxn += (column 9 of table a1). • calculate forces in the transverse direction per unit length: xynynsyn += (column 10 of table a1). • calculate steel areas for each region in the x and y di rections, assume f y = fyv = 500n/mm 2 (columns 11 and 12). • select the number of bars and bar size based on the available sizes and required areas. • detail calculation for level a is shown below table a1. 35 the journal of engineering research vol. 4, no.1 (2007) 23-36 table a1: calculation of reinforcement for example a 1 2 3 4 5 6 7 8 9 10 11 12 level width (mm) óxi (n/mm 2 ) óyi (n/mm 2 ) ôxy (n/mm 2 ) óxi / ôxy óyi / ôxy case s xn (n /mm) s yn (n /mm) asx (mm2) asy (mm2 /m) a 600 -2.09 0 4.21 -0.496 0 3 106 210 127.2 420 b 600 -1.57 0 4.21 -0.373 0 3 132 210 158.4 420 c 600 -1.05 0 4.21 -0.249 0 3 158 210 189.6 420 d 200 -0.52 0 4.21 -0.124 0 3 184.5 210 73.8 420 e 200 0 0 4.21 0 0 3 210 210 84 420 f 200 0.52 0 4.21 0.124 0 3 236.5 210 94.6 420 g 200 1.05 0 4.21 0.249 0 3 263 210 105.2 420 h 200 1.57 0 4.21 0.373 0 3 289 210 115.6 420 i 200 2.09 0 4.21 0.496 0 3 315 210 126 420 j 200 2.61 0 4.21 0.62 0 3 341 210 136.4 420 k 200 3.14 0 4.21 0.746 0 3 367.5 210 147 420 l 200 3.66 0 4.21 0.869 0 3 393.5 210 157.4 420 detail calculation for level a: col. 3: 2 9 6 /09.2 107375.5 2001060 mmn x xx i my a a −= − = − =σ (c) col. 8: from columns 6 & 7: we have: óxi / ôxy = -0.496 and ó yi / ôxy = 0 ? from fig a 4 the location of this point lies in the region of case 3. col. 9: the force in x direction cal culated as given in fig. a 4: mm/n106)21.409.2(50xynxnsxn =+−=+= . col. 10: the force in y direction calculated as given in fig. a 4: mmnnnn xyy s y /210)21.40(50 =+=+= . col. 11: the steel area in x direction is given by: 2mm2.127600x212.0mm/2mm212.0 500 106 yf s xn xa =⇒==== . col. 12: the steel area in y direc tion is given by: m/2mm4201000x42.0mm/2mm42.0 500 210 yvf s yn ya =⇒==== 4.21n/mm 2 2.09n/mm2 36 the journal of engineering research vol. 4, no.1 (2007) 23-36 (0,0) (-1,-1) || xy x n n case 3 || xynxn s xn += ||22 xynn −= xand ysteel required || xynyn s yn += y-steel only required case 1 0=sxn x n xy n xnn 2 2 += x n xy n yn s yn 2 −= case 2 x-steel only req uired x n xy n xn s xn 2 −= 0=syn y n xy n ynn 2 2 += case 4 no steel required 0=sxn 0=syn 2 4 2 )( 22 1 xyn ynxn ynxn n n + − ± + = ⎭ ⎬ ⎫ || xyn yn fig. a4: boundary graph for nielson's 2d design equations the journal of engineering research (tjer), vol. 14, no. 2 (2017) 166-181 doi: 10.24200/tjer.vol.14iss2pp166-181 design of biomechanical legs with a passive toe joint for enhanced human-like walking o. eldirdiry, r. zaier* and a. al-yahmedi department of mechanical and industrial engineering, college of engineering, sultan qaboos university, muscat, oman. received 5 march 2017; accepted 1 october 2017 abstract: this paper presents the design procedure of a biomechanical leg, with a passive toe joint, which is capable of mimicking the human walking. this leg has to provide the major features of human gait in the motion trajectories of the hip, knee, ankle, and toe joints. focus was given to the approach of designing the passive toe joint of the biomechanical leg in its role and effectiveness in performing human like motion. this study was inspired by experimental and theoretical studies in the fields of biomechanics and robotics. very light materials were mainly used in the design process. aluminum and carbon fiber parts were selected to design the proposed structure of this biomechanical leg, which is to be manufactured in the mechanical lab of the sultan qaboos university (squ). the capabilities of the designed leg to perform the normal human walking are presented. this study provides a noteworthy and unique design for the passive toe joint, represented by a mass-spring damper system, using torsion springs in the foot segment. the working principle and characteristics of the passive toe joint are discussed. four-designed cases, with different design parameters, for the passives toe joint system are presented to address the significant role that the passive toe joint plays in human-like motion. the dynamic motion that is used to conduct this comparison was the first stage of the stance motion. the advantages of the presence of the passive toe joint in gait, and its effect on reducing the energy consumption by the other actuated joints are presented and a comparison between the four-designed cases is discussed. keywords: design procedure; biomechanical leg; passive joint; human-like dynamic; locomotion; gait; toe joint. اكاة مشي االنسانتصميم سيقان الية حيوية بأصبع قدم ذو مفصل تفاعلي لتحسني حم الزايررياض ،عمر الدرديري * اليحمدي عامرو : تعرض هذه الورقة البحثية عملية تصميم الساق االلية احليوية، مع إصبع قدم ذو مفصل تفاعلي، والقادرة على حماكاة مشي امللخص اإلنسان يف مسارات احلركة من مفاصل الورك والركبة االنسان. وهذه الساق من املتوقع أن تعمل على توفري السمات الرئيسية ملشية والكاحل وأصابع القدم. ومت الرتكيز بعمق على نهج تصميم مفصل إصبع القدم التفاعلي للساق االلية احليوية وعلى دوره وفعاليته يف أداء مليكانيكا احليوية والروبوتات. ومت استخدام حماكاة االنسان. وقد استلهمت هذه الدراسة من الدراسات التجريبية والنظرية يف جماالت ا حيث قمنا باختيار مواد من األملنيوم واأللياف الكربونية لتصميم اهليكل املقرتح هلذه ،ميف عملية التصمي مواد خفيفة جدا بشكل أساسي بعرض قدرات الساق املصممة ألداء حماكاة الساق اااللية احليوية اليت سيتم تصنيعها يف خمترب امليكانيكا جبامعة السلطان قابوس. وسنقوم هذه الدراسة على توفري تصميما مميزا وفريدا ملفصل أصابع القدم التفاعلي متمثلة يف نظام الصمام املنظم تعمل مشي االنسان العادي. مفصل إصبع القدم التفاعلي. الزنربكي الشامل وذلك باستخدام زنربكات ملتوية يف جزء القدم. كما تناقش الدراسة مبدأ العمل وخصائص سنقوم بعرض أربع حاالت مصممة بطرق خمتلفة لنظام مفصل إصبع القدم التفاعلي لتناول الدور اهلام الذي يلعبه مفصل إصبع القدم التفاعلي . وسيتم عرض مزايا وجود يف حماكاة حركة اإلنسان. وتعترب احلركة الديناميكية املستخدمة إلجراء هذه املقارنة هي املرحلة األوىل للعمل رنة مفصل إصبع القدم التفاعلي يف املشية، وتأثريه على احلد من استهالك الطاقة من قبل املفاصل احملركة األخرى. وسيتم أيضا مناقشة املقا بني احلاالت املصممة األربع. اإلنسان، التحرك، املشية، مفصل إصبع القدم، إجراءات: الساق االلية احليوية، املفصل التفاعلي، حماكاة حركة الكلمات املفتاحية .التصميم * corresponding author’s e-mail: zaier@squ.edu.om o. eldirdiry, r. zaier and a. al-yahmedi 167 1. introduction during the last few decades, the fields of biomechanics and robotics have witnessed a significant improvement in the designs of the biomechanical legs and prostheses. moreover, many studies have focused on building humanoid robots in order to mimic human motions (denny et al. 2016). there are two classes of legged robots; powered and passive legs. the robotic leg to be considered in this study is a powered one. the joints of the robotic legs of this type are usually driven with electric actuators, unlike the passive legs that rely on the force of gravity to perform their stable gait on the ground. it is essential for passive legged robots to have a declining slope (collins et al. 2001; mcgeer 1990). however, if these legs are needed to operate on the ground level, an external force, must be applied to these legs to give them the initial movement (collins et al. 2005; wisse and van frankenhuyzen 2006). moreover, in some applications, an actuator is necessary to compensate for the energy loss when the foot touches the ground at the heel. usually, a pitching actuator in the hip joint is used for this purpose (alghooneh et al. 2016). there are, generally, three main purposes for designing biomechanical legs. the first objective is to use the biomechanical leg as an orthosis for a human leg. this type of leg is mainly used in rehabilitation, where the orthosis functions as a support for the paralysed leg of a human (allemand et al. 2009; yabunaka et al. 2013). since these types of biomechanical leg are used as a pillar for the leg, they are not designed to look anything but artificial. the second type of biomechanical leg is called a prosthesis, which is a device used to replace a missing part of the human body; as with amputee. very prominent designs for these types are the design of pantoe 1 in peking university in china (zhu et al. 2010), the design of massachusetts institute of technology (mit) the prosthetic foot (au et al. 2007) and the design in lapre et al. (2016). the third type of biomechanical leg is used to study the motion of human gait. examples of these types are the biped robots studied in (hwang et al. 2016; otani et al. 2016; yi et al. 2016; iida et al. 2009 and iida et al. 2007). in this paper, a biped robot with biomechanical legs capable of mimicking the main features of human gait was considered, but, to start with, any suitable biomechanical design, we must take into consideration the importance of ground reaction forces, especially when building a biomechanical leg. therefore, the structure of the foot, that consists of two segments connected together via a toe joint, needs to be designed so that the leg functions like a human leg. 1.1 motivation for design there were many factors that motivated this study. the main one was to address the importance of the toe in human locomotion and to deal with pathological gaits because of the many problems that humans may suffer from such as foot drop. therefore, the design of a biomechanical leg had to be done in such a way that the mechanical leg possesses the main features of a human leg. it was therefore necessary to investigate the role that the toe joint plays during motion. moreover, this study highlights the impact of this joint in the dynamic behavior of the biomechanical leg in gait. in order to study these features, it was important to produce a design procedure for the biomechanical leg, in general, and for the passive toe joint in particular. designing the parameters of the passive toe joint allowed for different study cases to be conducted and allowed for comparison of their dynamic behavior. compared to the large number of biomechanical legs designs, as reported above, very few investigations have considered this toe joint in their studies (sellauoti et al. 2006; wang et al. 2006; yamamoto et al. 2007) and the remarkable design of the toe joint in (piazza et al. 2016). the toe joint itself has been represented as either a powered joint (ezati et al. 2014; hernández-santos et al. 2012; nishiwaki et al. 2002) or as a passive joint (sellauoti et al. 2006). in this paper, a study case of a biomechanical leg with a passive toe joint represented by torsion springs is considered. moreover, the design of this toe joint and the control mechanism of its stiffness are represented so that the behaviour of the foot is relatively close to the behaviour of a human foot. 1.2 design objectives and constraints the objective of this study is to propose the design of a biomechanical leg with the closest essential features needed to generate a gait similar to the human one. furthermore, this leg will be used to study pathological gaits including, but not limited to, antalgic, ataxic, diplegic, drop foot. this paper, mainly, design of biomechanical legs with a passive toe joint for enhanced human-like walking 168 addresses the role of the toe joint in generating gait, and therefore, the following constraints have being identified for the design of the leg and these will be considered in the detailed design:  relatively similar dimensions to the human leg.  made of very light and stiff materials.  all the joints are capable of performing the same range of motion as in the human joint.  the toe joint needs to be passive, with the ability to control its design parameters. this paper represents the design procedure for the proposed biomechanical leg, and it is organized as follows: in section 2, the proposed mechanical design of the biomechanical leg is discussed, where geometry, materials, and motor selections are presented. section 2 also describes the examined range of motion of the biomechanical joints leg design and compares them with the range of motion of the human joints. in section 3, the capability of the simulated biomechanical leg design to perform normal human walking gait is presented. section 4 focuses on the suggested design of the passive toe joint and its proposed control mechanism. the kinematics and the dynamics of the proposed biomechanical leg are presented in section 5 and section 6, respectively. the results of these analyses are discussed in section 7. lastly, a summary is presented in section 8. 2. mechanical design in this section, the design and the components used to build the biomechanical leg are described (fig. 1(a)). figure 1. (a) an isometric view for the designed biomechanical legs and (b) kinematic parameters of the biomechanical legs. o. eldirdiry, r. zaier and a. al-yahmedi 169 2.1 dimensions, weights, and joints range motions table 1 and table 2 compare the average dimensions and weights of human-legs, taken from 100 volunteers in the study (de leva 1996) with the proposed design of the biomechanical leg. the dimensions were selected, so the design was as close as possible to the human leg, in order to make it easier to mimic the human gait. however, the weight was as light as possible, so that it could be easier to add extra weights if required. for this purpose, aluminum and carbon fiber materials were mainly used to build this design. the dimensions l1, l2, and l3 refer to the lengths of the thigh, the shank, and foot, respectively. to study the foot motion in more detail, the dimensions h1, h2, and h3 (fig. 2) refer to some segments in the foot. moreover, the length of the biomechanical leg had to be similar to the length of a human leg (table 1). however, the weight of the biomechanical leg needs not be a pivotal issue in order to mimic the human motion, as long as the motors are capable of carrying the body segments and performing the gait. table 1. dimensions comparison in (m). symbol human leg (de leva 1996) biomechanical leg l1 0.41 0.41 l2 0.44 0.44 l3 0.26 0.26 h1 0.09 0.09 h2 0.05 0.04 h3 0.07 0.06 l1, l2, l3 : thigh, shank, foot length h1 = toe length h2 = horizontal distance from back to the ankle of foot. h3 = vertical distance from ankle to the bottom of the foot. table 2. weights comparison in (kg). segment human leg biomechanical leg thigh 10 0.98 shank 3.1 0.31 foot 1.0 1.41 toe 0.36 de leva's (1996) study was used to determine the proportionality of heights and the weights for the proposed design segments. the averages for the total body height and weight were selected from de leva (1996), which are 173.1 cm and 73 kg, respectively. hence, the percentage of length and the weight of the leg segments with respect to the total height and weight of the body are found to be as follows:  length (thigh 23.2%, shank 24.7%, and foot 4.25%).  weight (thigh 14.16%, shank 4.33%, and foot 1.37%). based on these percentages, the length and the weight of the human leg segments are calculated (table 1 and table 2). the actual length and weight for the biomechanical leg segments, after building the design, are given in the same tables. the weights of the thigh and shank of the biomechanical leg are very light compared to the human leg segments. this is due to the very light materials used to build the biochemical leg. however, although very light materials were used to design the biomechanical leg, the foot segment is slightly heavier than the foot segment in the human foot (table 2), which is due to the large number of parts used to build the foot in the proposed design. many studies contributed to calculating the average jointsrange motion of human leg e.g. (grasso et al. 2000). the joint ranges of the human right leg, taken from grasso et al. (2000), is compared to the joint ranges of the biomechanical leg (table 3). the joint ranges of the biomechanical leg were calculated based on the motion capability of the joints (not causing any collapses) in the cad design in solid works. notice that the joint ranges of a human leg are within the ranges of motion of the biomechanical leg except in the knee (pitching) and the ankle (pitching). all these recorded joint ranges of the biomechanical leg are enough to perform normal human walking. 2.2 motors selection fig. 1(b) describes the kinematic parameters and the motor locations of the biomechanical leg, where three motors (3 dofs) are used in the hip position, for pitching, rolling and yawing motions. the knee joint is represented by 1 dof pitching motion, whereas 2 dof are used in the ankle joint to perform the rolling and design of biomechanical legs with a passive toe joint for enhanced human-like walking 170 (a) (b) figure 2. side views for (a) the biomechanical foot and (b) a sketch of human foot. table 3. joints ranges for right leg. joint human leg (grasso et al. 2000) biomechanical leg hip: roll pitch yaw -20° to +40° -30° to +110° -25° to +195° -180° to +180° -180° to +180° knee: pitch 0° to +150° -105° to +105° ankle: roll pitch -20° to +30° -50° to +20° -35° to +35° -45° to +25° toe (passive): pitch 0° to +70° table 4. maximum toque for the bio-mechanical leg joints. joint torque value (nm) hip: roll pitch yaw 3.87 5.23 1.50 knee: pitch 6.35 ankle: roll pitch 0.26 4.48 toe (passive): pitch 0.61 pitching motions. the toe joint (the only passive joint in this structure) is represented by 1 dof without any actuator. table 4 shows the maximum joint torque values that are required to accomplish normal human gait. dynamixel pro motors, provided by robotis (2017) are used since they can provide the torque values sufficient to perform the gait. it is clear that the pitching movement required more torque than the rolling and yawing movements (table 4). the motor l42-10-s300-r (dynamixel pro motor) was selected to represent the low torque joints, whereas the motor l54-30-s400-r was selected to fulfill the high toque joint (the pitching hip joint and the pitching knee joint required maximum torque of 5.23 nm and 6.35 nm, respectively). the specifications for these two motors are shown in fig. 3. another major factor considered in the selection of these motors is the angular speed of http://www.robotis.us/dynamixel-pro-l42-10-s300-r/ http://www.robotis.us/dynamixel-pro-l42-10-s300-r/ o. eldirdiry, r. zaier and a. al-yahmedi 171 motors required to produce a gait similar to human gait. since the angular velocity for a human joint is, approximately, 20 rpm (riemer and shapiro 2011), which is equal to 120 degree per second, these two motors were selected so that they could perform well at this speed. 3. biomechanical behavior and human walking gaits in order to mimic human motion using the biomechanical leg, it is important to study the cyclic pattern of human motion. the walking step consists of two phases: the stance phase, which represent about 60% of the duration of the whole step, and the swing phase, which takes approximately the remaining 40% (zhu et al. 2010). figure 4 highlights the phases of one step, presented by the proposed biomechanical leg. as shown in the figure, at the beginning of the step, the heel of the biomechanical foot strikes the ground. the bottom part of the leg was made of a rubber material in order to absorb the shock when the heel hits the ground. then, the pitching motor, at the ankle joint, rotates, causing a flat foot orientation, on the ground level. after that, with the help of the upper structure of the biomechanical leg, the ankle joint pitches in the opposite direction to reach the mid-stance phase of the foot and continues its rotation to reach the heel-off phase. because of this motion, the heel of the foot will start rising from the ground level, while the toe remains level on the ground. as a result, the passive toe joint stores energy, in a spring element, which is released at the end of the stance phase and, therefore, contributes in lifting the leg. during the swing motion, the ankle pitching actuator returns to its original orientation so that the foot becomes ready for a new stance phase. it is worth mentioning that during this cyclic motion, the upper motors of the hip and the knee joints contribute in maintaining the right orientation to perform the stance and the swing motions. 4. the passive toe joint to address the importance of the presence of the passive toe joint in the foot segment in the biomechanical leg and to enhance locomotion stability, the effectiveness of the passive toe joint during the stance phase should be investigated. the proposed model of the passive toe joint consisting of damper-spring-mass will be presented first. then, by investigating four values of the stiffness and damping coefficients, a comparison will be made in terms of the amount of energy consumed by the actuators of the ankle and knee joints. 4.1 mechanical structure of the passive toe joint in order to reach the design objectives and to achieve the controllability of the passive toe joint parameters, the design in fig. 5(a) and (b) was selected. in this design, the passive joint consists of five torsion springs, rotating freely around a carbon fiber tube. there are two bearings that support the carbon fiber tube from both sides and allow the pitching rotation in the toe joint. the torque of each spring can be calculated using the following equation: 𝜏 = −𝑘r𝜃 (1) where τ is the torque applied on the spring in (nm), and θ is the twist angle from its original position in radians. kr is the rotational stiffness constant (which is also known as spring's torsion coefficient or, torsion elastic modulus) with units of newton-meters/radian. in equation (1), the negative sign represents the direction of the torque with respect to the twist direction. when five springs work in parallel, their rotational stiffness will sum up and give the following rotational spring constant: 𝑘𝑟𝑒𝑞 = 𝑘r1 + 𝑘r2 + 𝑘r3 + 𝑘r4 + 𝑘r5 (2) in the structure shown above, there is a carbon fiber sheet with five rooms, one room in front of each spring. this sheet works as a stopper to prevent the upper part (side a) of the spring from rotating. all the lower side (side b) of the torsion springs are fixed in the foot. during the gait, when the heel of the biomechanical leg starts to rise and just before the toe leaves the ground (fig. 4), the lower part of the five springs will start to rotate around the carbon fiber tube. meanwhile, the upper part of the five springs will enter the rooms of the stopper. this stopper was designed so that two options of rotational stiffness could be provided. the first option is to have a constant https://en.wikipedia.org/wiki/radian design of biomechanical legs with a passive toe joint for enhanced human-like walking 172 figure 4. a simulation of one biomechanical leg during the stance and swing phases of a normal gait. data performance graph unit value dimension mm (in) 42 x 42 x 72 (1.65 x 1.65 x 2.83) weight kg(oz) 0.269(9.5) nominal voltage v 24 no load speed rpm 28 no load current a 0.52 continuous operation speed rpm 26.0 torque n.m(oz.fin) 1.7(241) current a 0.6 resolution steps/turn 263168 gear ratio 257:1 backlash arcmin 4.2 network interface rs-485 operating temperature 0 c 5~55 (a) data performance graph unit value dimension mm (in) 54 x 54 x 108 (2.13 x 2.13 x 4.25) weight kg(oz) 0.612(21.6) nominal voltage v 24 no load speed rpm 28.7 no load current a 1.19 continuous operation speed rpm 26.9 torque n.m(oz.fin) 2.5(354) current a 1.6 resolution steps/turn 288360 gear ratio 401:1 backlash arcmin 4.6 network interface rs-485 operating temperature 0 c 5~55 (b) figure 3. specifications for the dynamixel pro motors (a) l42-10-s300-r and (b) l54-30-s400-r (robotis 2017). http://www.robotis.us/dynamixel-pro-l42-10-s300-r/ http://www.robotis.us/dynamixel-pro-l42-10-s300-r/ o. eldirdiry, r. zaier and a. al-yahmedi 173 figure 5. (a) an isometric material and (b) a side views for the structure of the passive toe joint part for the biomechanical leg. (c) a supporter for the isotropic material is attached below the stopper, (d) the stopped is hidden and one of the five springs is shown in front of the isotropic material, and (e) a side view for the structure. stiffness when the five torsion springs engage with the carbon fiber sheet rooms simultaneously. the other option is to have a variable stiffness, when the springs are engaged inside the rooms at different times during the gait. it is worth mentioning that the springs’ housings for the first option are the same lengths while in the second option they have different lengths (fig. 5). also, the carbon fiber sheet (i.e. the stopper) can be easily replaced with another sheet in order to switch between the two rotational stiffness. in this study, a constant value of the rotational spring constant will be considered. the energy u is the stored energy in the spring in joules and it can be calculated using the following equation: 𝑈 = 1 2 𝑘req𝜃 2 (3) in this design (fig. 5(c), (d), and (e)) an elastic isotropic material is placed below the stopper and in front of the springs in order to add a damping effect to the system. the springs will be in contact with the isometric material before they reach the stopper. this material can be selected later based on the rotational damping coefficient cr, which will also, be calculated later in this study. 4.2 passive toe joint model during the last stage of the stance phase (fig. 4) described by the motions between the positions “4” and “5”, the location of the toe joint is fixed. the role of the springs comes during this period. the proposed design model for the toe joint is represented by a springdamper system (fig. 6(a)). the mass (𝑀) represents the mass of the whole leg. the line (𝑟) links the toe joint to the heel point, and 𝜃 depicts the angle between the line (𝑟) and the ground level. the motion of the heel point is represented by the arc length (𝑠). when the biomechanical leg performs the motion, from position “4” to position “5” (fig. 4), the value of θ increases. hence, energy will be stored in the compressed torsion springs in the toe joint. this stored energy will be released allowing the foot to contribute to generating the gait from positions “5” and “6” (fig. 4). in order to manifest the roles of the toe joint, the model was described by the following ordinary differential equation: 𝐽�̈� + 𝑐r�̇� + 𝑘𝑟𝑒𝑞 𝜃 = 0, (4) where 𝐽 is the polar moment of inertia of the toe joint about the y-axis, 𝑐r is the rotational damping coefficient, 𝑘𝑟𝑒𝑞 is the rotational spring constant for the five torsion springs. 𝜃, �̇� and https://en.wikipedia.org/wiki/ordinary_differential_equation https://en.wikipedia.org/wiki/ordinary_differential_equation design of biomechanical legs with a passive toe joint for enhanced human-like walking 174 𝜃 ̈ are the angle, the angular velocity, and the angular acceleration of the toe joint, respectively. by neglecting the mass of the links, the mass that is being conceded in this study is the mass of the torso and the total mass of the other swinging leg, which is 𝑀 = 5 kg. therefore, the curved rotational model (fig. 6(a)) is demonstrated as an equivalence model for the linear model (fig. 6(b)). since the mass m is moving vertically without rotating, the equation of motion will depend only on the displacement y. hence, the rotational spring 𝑘req needs to be converted to linear one 𝑘leq. the following relation can be used: 𝑘leq = 𝑘req ( 1 𝑟 ) 2 (5) similarly, the relation between the linear damping coefficient (𝑐l) and the rotational damping coefficient (𝑐r) can be calculated, as follows: 𝑐l = 𝑐r ( 1 𝑟 ) 2 (6) the equation of motion of the mass m can be illustrated in equation (6) in term of the linear elements: 𝑀�̈� + 𝑐l�̇� + 𝑘𝑙𝑒𝑞 𝑦 = 0, (7) where 𝑀 is the mass of the biomechanical leg above the foot segment, including all the mass shown by the area a. 𝑦, �̇� and �̈� are the position, the velocity, and the acceleration of the heel point along the y-axis with respect to the ground, respectively. an important point to mention is that the spring constant 𝑘l and the damping coefficient 𝑐l are linear in this case and they are measured in n/m and the ns/m, respectively. unlike equation (4), the spring constant and the damping coefficient are measured in nm/rad and mns/rad, respectively. 4.3 system design for the passive toe joint in order to design the toe joint, the desired angle 𝜃 that the toe joint can reach is set. assuming that, during the last stage of the stance motion from position “4” to position “5” (fig. 4) the toe joint rotates in the clockwise direction from 00 to 300. if the value of 𝑟=0.13m, then using the relation 𝑦 = 𝑟. 𝑠𝑖𝑛𝜃, the desired change in 𝑦 along the y-axis will vary from 0m to 0.065m. the solution of equation (7) is given as follows: 𝑦(𝑡) = 𝑒 −𝜉𝜔𝑡 (𝑏1 𝑒 𝑗√1−𝜉2𝜔𝑡 + 𝑏2𝑒 −𝑗√1−𝜉2𝜔𝑡 ), (8) where, 𝑏1 and 𝑏2 are arbitrary complex-valued constants of integration that can be found by the initial conditions, 𝜉 is non-dimensional number called damping ratio and 𝜔 is the undamped natural frequency in rad per second. the damping coefficient 𝑐l and the spring constant 𝑘l can be found from the following relations: 𝑐𝑐𝑟 = 2√𝑘l𝑀, (9) 𝑐𝑐𝑟 = 2𝑀𝜔, (10) 𝜉 = 𝑐l 𝑐𝑐𝑟 , (11) where 𝑐𝑐𝑟 is the critical damping value. therefore, the values of the damping coefficient 𝑐l and the spring constant 𝑘l can be calculated with the following equations, for given values of 𝜔 and 𝜉: 𝑘l = 𝑚𝜔 2 (12) 𝑐l = 2𝜉√𝑘𝑀 (13) the effect of the value of 𝜉 on the response of the second order system is demonstrated in inman (2013), where the system is called under damped system when 0<𝜉<1, critically damped system when 𝜉=1, and overdamped system when 1< 𝜉<2. three values, of 𝜔 and 𝜉, were selected in order to investigate the trajectories of the heel point during 1.5 seconds in the second order system (fig. 7 and table 5). for the selected parameters 𝜔 and 𝜉, the values of 𝑐 and 𝑘 were calculated, using the equations (12) and (13), and recorded (table 6). to study the effect of the passive toe joint in the powered ankle and knee joints, a fourth case for the flat foot without a spring and damper elements (fixed toe joint) was introduced. then, both the kinematics and dynamics analysis of the biomechanical leg must be studied. https://www.abebooks.com/servlet/searchresults?an=daniel+inman&cm_sp=det-_-plp-_-author o. eldirdiry, r. zaier and a. al-yahmedi 175 figure 6. (a) a sketch of a human foot with mass-spring damper system model, for the toe joint, which uses torsion spring and damper. (b) equivalent mass-spring damper system model that uses linear spring and damper. figure 7. the vertical displacements of the heel point, in the biomechanical foot, caused by the rotational movement of the toe joint by 300 for case 1, case 2, case 3, and case 4. 5. kinematics analysis for the biomechanical leg. before studying the dynamics and addressing the effect of the passive toe joint on the ankle and knee joints, it is important to solve the inverse kinematic for the whole leg. since the case when the biomechanical leg is at the end of the stance phase is considered, the location of the toe joint will be taken as the base point (fig. 8). the torso part will represent the end-effector of the mechanism as the point (𝑋𝑇 , 𝑌𝑇 ), and it will move based on a predefined motion profile. since the structure of the proposed biomechanical leg is identical to the structure of the biped robot in wang et al. (2006), similar equations and methodology will be developed table 5. three cases of damper and springs for toe joint. symbol (unit) case 1 case 2 case 3 𝜉 1.0 0.2 0.15 𝜔 (hz) 5.0 1.2 2.1 (a) (b) design of biomechanical legs with a passive toe joint for enhanced human-like walking 180 in deriving the dynamics and the kinematics of the leg. in this research, the following two relations represent the desired planar motion of the torso: 𝑋𝑇 = 𝑎𝑡 − 0.1 (14) 𝑌𝑇 = 𝐴 𝑠𝑖𝑛 ( 𝜋𝑡 2 ) + 𝐿2 + 𝐿3 − 𝐿4, (15) where 𝑎 is the forward distance for the torso and 𝐴 is the swing height of the torso motion. in order to achieve a reachable position by the torso during the stance phase, 𝑎 and 𝐴 are set to be 0.3m and 0.1m, respectively. 𝑡 is the time, 𝐿2,𝐿3, and 𝐿4 are the lengths of the links in the mechanism (fig. 8). the angle 𝛼, which is shown in the same figure, is constant in the design and approximately equal to 𝜋 9 . the toe joint is considered to be fixed in the ground and represented with the angle 𝜃1. since the position of the heel point "𝑦" was determined in the previous section for different toe joint cases, the path trajectory of 𝜃1 can be found using the relation 𝜃 = sin−1 ( 𝑦 𝑟 ). therefore, for a given torso motion, the other pitching angles of ankle, knee and hip joints of the supporting leg can be found as following: 𝜃2 = 𝛾 − 𝛽 − 𝜋 2 + 𝜃1 (16) 𝜃3 = 𝑎𝑟𝑐𝑐𝑜𝑠 ( (𝑋𝑇−𝐿1𝐶1) 2+(𝑌𝑇−𝐿1𝑆1) 2−𝐿2 2 −𝐿3 2 2𝐿2𝐿3 ) (17) 𝜃4 = 𝜋 2 − 𝛾 + 𝛽 − 𝜃1 − 𝜃3 (18) from the geometry of the leg, (fig. 8) 𝛾 and 𝛽 can be represented by the following equations: 𝛾 = arctan (𝑌𝑇 − 𝐿1𝑆1 , 𝑋𝑇 − 𝐿1𝐶1) (19) 𝛽 = 𝑎𝑟𝑐𝑐𝑜𝑠 ( (𝑋𝑇−𝐿1𝐶1) 2+(𝑌𝑇−𝐿1𝑆1) 2+𝐿2 2 −𝐿3 2 2𝐿2√(𝑋𝑇−𝐿1𝐶1) 2+(𝑌𝑇−𝐿1𝑆1) 2 ) (20) where 𝑆𝑛 and 𝐶𝑛 in the equations are short forms of 𝑆𝑖𝑛(𝜃𝑛 ) and 𝑐𝑜𝑠(𝜃𝑛), respectively. 6. kinematics analysis for the biomechanical leg. before calculating the torque and the output energy of the joints, it is important to address the potential energy (𝑈) and the kinetic energy (𝐾) of the system. the following equations were used to describe the planar location of the torso (𝑋𝑇 , 𝑌𝑇 ): 𝑋𝑇 = 𝐿1𝐶1 + 𝐿2𝐶2 + 𝐿3𝐶3 (21) 𝑌𝑇 = 𝐿1𝑆1 + 𝐿2𝑆2 + 𝐿3𝑆3 (22) therefore, the velocity of the torso can be written as: 𝑋�̇� = −𝐿1𝑆1𝜃1̇ − 𝐿2𝑆2𝜃2̇ − 𝐿3𝑆3𝜃3̇ (23) 𝑌�̇� = 𝐿1𝐶1𝜃1̇ + 𝐿2𝐶2𝜃2̇ + 𝐿3𝐶3𝜃3̇ (24) where 𝜃�̇� is the angular velocity of the joints. based on the equations of the position and velocity of the torso and considering the gravity as 𝑔 = −9.8 m/s2, the potential and the kinetic energies of the system can be calculated as follows: 𝑈 = 𝑀𝑔𝑌𝑇 = −𝑀𝑔 ∑ 𝐿𝑛 𝑆𝑛 3 𝑛=1 (25) 𝐾 = 1 2 𝑀 ((𝑋�̇� ) 2 + (𝑌�̇� ) 2 ) (26) then, using lagrange’s dynamical equation: 𝑑 𝑑𝑡 ( 𝜕𝐾 𝜕𝜃�̇� ) − 𝜕𝐾 𝜕𝜃𝑛 + 𝜕𝑈 𝜕𝜃𝑛 = 𝑇𝑛 (27) the torque of the system is represented by the following dynamic equation: 𝜏 = 𝐴�̈� + 𝐵𝜃2̇ + 𝐺 (28) where 𝜏 = [𝜏1 𝜏2 𝜏3 ] represents the torque of the first three joints, as 𝜏4 = 0 in the proposed case. the angular acceleration in the equation is �̈� = [𝜃1̈ 𝜃2̈ 𝜃3̈ ] 𝑇, and the square of the angular velocity is 𝜃2̇ = [𝜃1 2̇ 𝜃2 2̇ 𝜃3 2̇ ] 𝑇 . the matrices 𝐴, 𝐵 and g of the dynamic equation (28) are calculated as: table 6. calculated damper and springs coefficients for toe joint. symbol (unit) case 1 case 2 case 3 𝑐l(n.s/m) 50 2.4 3.15 𝑘l(n/m) 125 7.2 22.05 176 o. eldirdiry, r. zaier and a. al-yahmedi 177 figure 8. biomechanical leg model with a fixed toe joint position to perform the first stage of the stance motion. 𝐴 = [ 𝑀𝐿1 2 𝑀𝐿1𝐿2𝐶2−1 𝑀𝐿1𝐿3𝐶3−1 𝑀𝐿1𝐿2𝐶2−1 𝑀𝐿2 2 𝑀𝐿2𝐿3𝐶3−2 𝑀𝐿1𝐿3𝐶3−1 𝑀𝐿2𝐿3𝐶3−2 𝑀𝐿3 2 ] 𝐵 = [ 0 −𝑀𝐿1𝐿2𝑆2−1 −𝑀𝐿1𝐿3𝑆3−1 𝑀𝐿1𝐿2𝑆2−1 0 −𝑀𝐿2𝐿3𝑆3−2 𝑀𝐿1𝐿3𝑆3−1 𝑀𝐿2𝐿3𝑆3−2 0 ] 𝐺 = [ −𝑀𝑔𝐿1𝐶1 −𝑀𝑔𝐿2𝐶2 −𝑀𝑔𝐿3𝐶3 ] , where 𝑆𝑛−𝑚 and 𝐶𝑛−𝑚 are defined as 𝑆𝑖𝑛(𝜃𝑛 − 𝜃𝑚 ) and 𝑐𝑜𝑠(𝜃𝑛 − 𝜃𝑚 ), respectively. by using the toque equation (28) and the angular velocity of each joint 𝜃�̇� , the output energy of the joints can be calculated from the following equation: 𝑊𝑛 = ∫ |𝜏𝑛. 𝜃�̇�| 𝑡 0 𝑑𝑡 (29) 7. results and discussion the role of the toe joint comes into place in the biomechanical leg, during the normal gait, at the end of the stance phase and in the first stage of the swing phase. these two stages are addressed by the movements between position “4” to position “5” and from position “5” to position “6” (fig. 4). the study by wang et al. (2006) showed the advantages of using the passive toe joint in the biped robot without considering any design factors for the passive toe joint. in yamamoto et al. 2007; and zhu et al. 2010, the design factors and the dynamics (i.e. stiffness and damping coefficients) of the passive toe joint were considered for only one case. however, in this study, three different cases were presented, which show the effect of considering different values for the design factors in the passive toe joint in the energy consumption of the biomechanical leg. figure 9 compares the joints’ angles of the biomechanical leg, during the last stage of the stance motion, for the four toe joint cases. 𝜃1, 𝜃2, 𝜃3, and 𝜃4 represent the toe, ankle, knee, and hip joints, respectively. except in case 4, the case with flat-foot (fixed toe-joint), the start and the end values for all joints are the same. however, the trajectories of these joints values are different in the middle. it is very clear from the figure that the minimum changes in the ankle (𝜃2) and the knee (𝜃3) joints appear in case 1. the effect of the presence of the toe joint is apparent in the toque graphs (fig. 10). 𝜏1, 𝜏2, and 𝜏3 represent the toe, ankle, and the knee torques, respectively. in case 1, although torque of the knee joint (𝜏3) starts with a high value compared to the other cases, the torque of the knee joint in case 1, has a lower value compared to the other cases during the whole motion. another feature of case 1 is that the toque of the ankle joint (𝜏2) starts with a very low value and continues to have a minimal toque value during the motion compared to the other three cases. figure 11 shows the output energy curves for the joints during the desired motion for the four cases. 𝐸1, 𝐸2, and 𝐸3 represent the toe, ankle, and the knee energies, respectively. in both fig. 11((a), and (b)) the total energy consumption (𝐸1 and 𝐸2) by the toe and the ankle joints, respectively, during the 1.5 seconds gait for the three designed cases (case 1, case 2, and case 3) are approximately the same. since there is a proportional relation between the energy and the torque, the results for the energy (fig. 11(c)) were expected where the energy of the knee joint (𝐸3), in case 1, started from a higher value compared to the other three cases. however, during the rest of the motion, it maintains a minimal value. case 2 was competing with case 3 in having minimal energy output for knee joint (𝐸3) in the early stage of the motion and the final energy outcome of both cases where it was design of biomechanical legs with a passive toe joint for enhanced human-like walking 178 less than the outcome of case 4 (without the toe joint). in the desired motion, the greatest reduction in energy consumption due to the presence of the toe joint was found in the knee (𝜃3) joint. the greatest reduction was found to be in case 1 with 𝑘l=125 (n/m), and 𝑐l= 50 (n.s/m). since these calculations were based on the linear spring and damper coefficients, then, the rotational spring and damper elements of the system can be calculated, using equations (5) and (6). the calculations of case 1 give kr = 2.1125 the rotational stiffness and 𝑐r = 0.845 (m.n.s/rad) as the damping coefficient for the rotational model. moreover, the overall output energy of the knee joint for the designed case (with the toe-joint) is less than the output energy of the knee joint in the case without the toe joint. 8. conclusion this study presented the design procedure for biomechanical legs with a passive toe joint. the design was made based on the motion capability of normal human gait where the joints’ ranges of human joints are considered. the design of these biomechanical legs was made from very light materials. the dimensions of the designed leg and the motors of the joints were selected so that the biomechanical leg could mimic the gait of normal human. a unique design for a toe joint was presented, in detail, to address the concept of the passive joint and to present its role during the human-like motion. the suggested design of the toe joint allowed for different design cases with different design parameters to be examined and tested. the proposed model for the passive toe joint was presented as a massspring damper system and three designed cases were extracted from this model. another case was introduced for the foot as a flat-foot (fixed toe joint). the results, from the dynamic analysis that compares these four cases, gave a noticeable finding on how the presence of the (a) case 1 (b) case 2 (c) case 3 (d) case 4 figure 9. the calculated changes in the joints angles (𝜃2, 𝜃3, and 𝜃4) for a given torso movement, with different trajectories (different cases) of toe joint (𝜃1). o. eldirdiry, r. zaier and a. al-yahmedi 179 passive top joint can affect the performance of proposed design, the reduction in energy consumption by the actuated joints, due to the presence of the passive toe joint, was proved. the case with the least energy consumption was identified. the particular case will be used when designing the final design for the manufacturing phase. moreover, the results from this paper gave an idea about how the driven trajectory of the passive toe joint, by the powered joints, can affect the dynamic motion of a biomechanical leg. conflict of interest the authors declare no conflicts of interest. (a) (b) (c) figure 10. the calculated torques in the joints: (a) toe, (b) ankle, and (c) knee for a given torso movement, with different trajectories (different cases) of toe joint (𝜃1). (a) (b) (c) figure 11. the calculated energy in the joints (a) toe, (b) ankle, and (c) knee for a given torso movement, with different trajectories (different cases) of toe joint (𝜃1). design of biomechanical legs with a passive toe joint for enhanced human-like walking 180 funding no funding was received for this research. references alghooneh m, wu cq, esfandiari m (2016), a passive-based physical bipedal robot with a dynamic and energy efficient gait on the flat ground. ieee international conference on transactions on mechatronics 21(4): 1977–84. allemand y, stauffer y, clavel r, brodard r (2009), design of a new lower extremity orthosis for overground gait training with the walk trainer. ieee international conference on rehabilitation robotics 550–55. au sk, weber j, herr h (2007), biomechanical design of a powered ankle-foot prosthesis. ieee 10th international conference on rehabilitation robotics 298–303. collins sh, wisse m, ruina a (2001), a threedimensional passive-dynamic walking robot with two legs and knees. the international journal of robotics research 20(7): 607–15. collins s, ruina a, tedrake r, wisse m (2005), efficient bipedal robots based on passivedynamic walkers. science 307(5712): 1082–85. de leva p (1996), adjustments to zatsiorskyseluyanov’s segment inertia parameters. journal of biomechanics 29(9): 1223–30. denny j, elyas m, d’costa sa, d’souza rd (2016), humanoid robots–past, present and the future. european journal of advances in engineering and technology 3(5): 8-15. ezati m, khadiv m, moosavian saa (2014), dynamics modeling of a biped robot with active toe joints. in second rsi/ism international conference on robotics and mechatronics (icrom) 107–12. grasso r, zago m, lacquaniti f (2000), interactions between posture and locomotion: motor patterns in humans walking with bent posture versus erect posture. journal of neurophysiology 83(1): 288– 300. hernández-santos c, rodriguez-leal e, soto r, gordillo jl (2012), kinematics and dynamics of a new 16 dof humanoid biped robot with active toe joint. international journal of advanced robotic systems 9(5): 190. hwang ks, lin jl, li js (2016), biped balance control by reinforcement learning. journal information science and engineering 32(4): 1041-1060. iida f, rummel j, seyfarth a (2007), bipedal walking and running with compliant legs. ieee international conference on robotics and automation 3970–75. iida f, minekawa y, rummel j, seyfarth a (2009), toward a human-like biped robot with compliant legs. robotics and autonomous systems 57(2): 139–44. inman dj (2013), engineering vibration. pearson 4th edition. lapre ak, umberger br, sup fc (2016), a robotic ankle–foot prosthesis with active alignment. journal of medical devices 10(2): 025001. mcgeer t (1990), passive dynamic walking. the international journal of robotics research 9(2): 62–82. nishiwaki k, kagami s, kuniyoshi y, inaba m, inoue h (2002), toe joints that enhance bipedal and fullbody motion of humanoid robots. in ieee/ icra’02 international conference on robotics and automation (3): 3105–10. otani t, hashimoto k, isomichi t, sakaguchi m, kawakami y, lim ho, takanishi a (2016), joint mechanism that mimics elastic characteristics in human running. machines 4(1): 5. piazza c, della santina c, gasparri gm, catalano mg, grioli g, garabini m, bicchi a (2016), toward an adaptive foot for natural walking. in ieee-ras 16th international conference on humanoid robots (humanoids) 1204-1210. riemer r, shapiro a (2011), biomechanical energy harvesting from human motion: theory, state of the art, design guidelines, and future directions. journal of neuroengineering and rehabilitation 8(1): 22. robotis (2017), dynamixel pro [online]. available: http:// en.robotis.com / index / index.php. [accessed: 10oct2017]. sellauoti r, stasse o, kajita s, yokoi k, kheddar a (2006), faster and smoother walking of humanoid hrp-2 with passive toe-joints. in ieee/rsj international conference on intelligent robots and systems 4909–14. wang l, yu z, meng q, zhang z (2006), influence analysis of toe-joint on biped gaits. o. eldirdiry, r. zaier and a. al-yahmedi 181 in 2006 international conference on mechatronics and automation 1631–35. wisse m, van frankenhuyzen j (2006), design and construction of mike; a 2-d autonomous biped based on passive dynamic walking. in adaptive motion of animals and machines, springer tokyo, 143– 54. yabunaka t, kawaguchi n, shinoda n, nakanishi a, kobayashi n, yano ki (2013), development of an improved lower limb orthosis for a motion-assist robot for the lower limb. in 16th international conference on advanced robotics (icar) 1–6. yamamoto k, sugihara t, nakamura y (2007), toe joint mechanism ssing parallel four-bar linkage enabling humanlike multiple support at toe pad and toe tip. in 2007 7th ieee-ras international conference on humanoid robots 410–15. yi sj, zhang bt, hong d, lee dd (2016), whole-body balancing walk controller for position controlled humanoid robots. international journal of humanoid robotics 13(01): 1650011. zhu j, wang q, wang l (2010), pantoe 1: biomechanical design of powered ankle-foot prosthesis with compliant joints and segmented foot. in 2010 ieee/asme international conference on advanced intelligent mechatronics 31–36. microsoft word paper 5.docx the journal of engineering research (tjer) (2017) 49-63   development of a system model for determining optimal personnel interaction strategy in a production industry b. kareem*, a and a. adekiigbeb a b. kareem, department of mechanical engineering, federal university of technology, akure, nigeria. b a. adekiigbe, department of mechanical engineering technology, osun state polytechnic, iree, nigeria. received 29 july 2013; accepted 9 june 2016 abstract: manufacturing organizations have become more complex in recent time as a result of technological advances. communication among production workers operating in an environment marked by increased organizational complexity may require planning for the economically appropriate selection of network channels/media with enhanced productivity. this paper examines traditional and modern communication channels (media) and their comparative advantages over one another in their adoption in manufacturing organizations. in this framework, six media (human messengers, mobile-phones, intranet, fixed-internet, mobile-internet, and private branch exchange [pbx] phone systems) were subjected to analyses using five identified network patterns (all-channel, chain, y, wheel, and circle) of interactions in manufacturing organizations. costs, benefits, and the utility of the channels were integrated into the model and utilized to determine the most sustainable media that could enhance productivity in industry. the developed model was implemented using expert data/information collected from the plastic production industry. the results of an availability assessment showed that the enhancement of productivity could be fully achieved by utilizing mobile phones and internet networks, but when considering overall utility, only mobile phones could bring about the desired productivity with 0.59 probability. the findings suggest that the system developed is robust in revealing how productivity might be affected by means of communication among industrial workers. keywords: communication, channels, personnel, productivity, utility. אאאאאאא  k،*k،   wא א אא א  kא    א   א א  kא  א  א  אא א   אא  א א   אאאא  אאאeאfאא f   א         א א  kא א א א א ،אא א א ،א א ،אאא  א א ، אאאpbx،א،אאfאeyא، א א   אא   אא  א אא א   keאא אא   א אא  א   א א kא  א   א א א א     א  א  kא   אאאאאאאאא ٥٩}٠א  א    א    א  א א  א  k kא  wאאkא،א،א،אא،א * corresponding author’s e-mail: bkareem@futa.edu.ng b. kareem and a. adekiigbe   50 1. introduction the transformation of production resources to useful goods and services requires effective interaction among the principal resources, namely manpower, material, and machinery (drucker 2002). people in organizations need to communicate within and outside of those settings for effective coordination of production processes and service delivery (kleinbaum et al. 2008). interaction (communication, contact, interface, dealings, relations, collaboration) is the process of sharing ideas, information, and messages with others in a particular time and place (kleinbaum et al. 2008). communication may be written or spoken, or expressed in some other way. in another categorization, communication may be nonverbal or verbal. nonverbal communication includes facial expressions, body language, or gestures. verbal and visual communication may be in the form of images or pictures (painting, photography, video, or film) while telephone calls, electronic mail, cable television, satellite broadcasts, direct conversation, etc. are examples of verbal or electronic communication (balschbach et al. 1998; payne et al. 1996). contributions of nonverbal communication system in terms of cost and effectiveness to the production process are difficult to measure because of the seeming intangibility of the system (nucci et al. 2013). however, interaction by means of verbal and electronic communication systems is quantifiable, and its effectiveness to the production process (productivity) is measurable (balschbach et al. 1998; nucci et al. 2013; payne et al. 1996) using appropriate decision tools (saaty 1990). in the current study, a system model is developed for sustainable personnel interaction in a production industry by considering the relative utility of communication channels with the objective of achieving effective service delivery in the organization. interaction in the production industry should be based on a pre-determined design with the target of improving output capacity or enhancing goodwill (gurbaxani and plice 2004). advanced technology has created many communication tools that could be necessary and useful in the improvement process (nagy 2013). companies have a choice to make among various communication channels available in order to sustain their competitiveness (gurbaxani and plice 2004). it is possible that making wise choices in communication channels might improve the agility and efficiency of an industrial system (biggiero 1999; stohl 1995; wada et al. 2013; yates et al. 1999). communication/interaction models have been used throughout history as a means of analyzing the components and structures of effective communication in the industry (opperman 2013; shannon 1948). models that are well known and widespread in the literature (feital et al. 2013; nagy 2013; payne et al. 1996; shannon 1948) for information/ communication technology (ict)-based cost estimation and benefit in organization are the total cost of ownership model (tco) which deals with all costs associated in the lifecycle of any capital investment in hardware or software, from acquisition to disposal (lievrouw 2008; nagy 2013). the jikes rvm adaptive optimization system (shannon 1948) attempted to evaluate the break-even point for each action using an online competitive algorithm (zhao et al. 2013). it relies on an analytic model to estimate the costs and benefits of each selective recompilation action, and evaluates the best actions according to the model predictions online. the enterprise resource planning (erp) system established and implemented administrative software systems that are organizational based. the erp software systems provided are expensive and have long-term financial and human resources consequences (opperman 2013). the business impact analysis (bia) (feital et al. 2013) determined how far to go in protecting the functions of the people, information, and equipment in an organization in order to sustain their reliability (kareem 2015). the quantitative risk model (qrm) (nohria and eccles 1992) was used to analyze expected interaction losses over a pre-determined period (sooksumrarn and krairiksh 2013). the generalized cost consequence (gcc) (nohria and ghoshal 1997; payne et al. 1996) model estimates the total cost of interaction outages as a function of time after an event. the major diagnostic category model (mdc) (kleinbaum et al. 2008) helps to relay information that enable the determination of resources that experience highly severe or very frequent losses in production organizations. in the stated models system, interaction was determined using separate criteria based on lines of authority (line, functional, modified, hierarchical, matrix, flat, and combination) development of a system model for determining optimal personnel interaction strategy in a production industry   51   (central intelligence agency [cia] 2011; jaques 1998; urwick 1943; urwick 1974), patterns of interaction (circle, y, chain, wheel and allchannel) (payne 1996), or the channels of transmission (human messengers, telephones, face-to-face, meetings, memorandum, formal report, teleconferencing, intranet, fixedinternet, mobile internet, pbx phone system (froehle 2006; kuboye et al. 2009; larkin and larkin 2005; payne 1996; stohl 1995). many of these studies were carried out theoretically (bhattacharjee 2007; biggiero 1999; mackenzie 1978; urwick 1974). there is a need for integrating interaction channels and lines of authority into a production system in order to arrive at a realizable and practical system model useful for the sustenance of productivity and effective service delivery. the identified channels/media of telecommunications utilized in nigerian production industries—mtn, zain, glo mobile, visafone, multilinks, starcomms, zoom, and etisalat—control the gsm market (kareem 2014; ndb 2007). past studies (kuboye et al. 2009; olokede 2009) have found some of these channels to be unreliable and exhibit random failures similar to other mobile systems (kareem 2015). the effects of utilizing any of these communication channels /networks on productivity in the industry have not been quantitatively addressed (mtega and msungu 2013; nyambo and ligate 2013). previous studies failed to consider the cost benefit and utility of different ict channels in industry (cevasco and marmolejo 2013; hurley 2008; jackson et al. 2012; larkin and larkin 2005; li et al. 2013; lu and yu 2012). the contribution of this study to the literature is to propose the integration of communication channels and personnel interactions into lines of authority. additionally, the study aims to highlight the responsibilities of a production organisation in minimizing lifecycle costs and maximize the benefits and utility of the selected channels with a target of improving productivity and service delivery. the practical implementation of an integrated system for determining optimal interaction channel(s) in a production industry will be carried out. the study will be extended to cost benefit and utility analyses of the media. a cost-benefit analysis involves weighing the total expected costs against the total expected benefits of one communication channel /medium over the others (bhattacharjee 2007) in order to select the most profitable option. a discount rate will be used to compute all relevant future costs and benefits in present value terms. utility is an alternative measure based on the non-monetary benefit used to express the relative satisfaction of personnel (values) with the communication/interaction media/channels under consideration. utilities of affordability, dependability, maintainability, portability, flexibility, and usability of service equipment have been explained in the literature (adam et al. 1995; bhattacharjee 2007). for this study, only attributes based on affordability (acquisition, development, and operations) and usability (performance, availability, and quality) will be considered because of the low capital outlay and technicalknow-how of the production industries in a developing economy. a technique based on the analytical hierarchy process (ahp) was found to be promising (finnie et al. 1993; saaty 2008) for analyzing the stated attributes because it helped set priorities and the likelihood of arriving at the most optimal decision under complex qualitative and quantitative characteristics (saaty 1990). 2. model formulation this section provides modelling parameters, relationships among the parameters, a modelling strategy/framework, and the implementation of the model with a case study. 2.1 modelling parameters and objectives in developing the model, the parameters considered in regards to the production industry include communication channels/media in use, the number of personnel involved, pattern(s) of interaction, and the cost, benefit and utility of the channels/media utilized. communication channels considered in the model were the use of traditional human messengers, mobile phones, the intranet, fixed internet, mobile internet, and a pbx phone system. the patterns of communication/interaction utilized were chain, y, wheel, circle, and all-channels (kareem 2014). the objective was to obtain (choose) communication channel(s) for the industry with minimum operation costs and maximum benefits and utility. the number of personnel involved was the major constraint besides the number of communication patterns and channels utilised. b. kareem and a. adekiigbe   52 figure 1. interaction pattern, authority and responsibility relationships. 2.2 modelling parameter relationships relationships that exist among the model parameters appear in fig. 1. it consists of channels j j( = 1, 2...6). each channel j option could be utilized by personnel numbering  using interaction / communication network pattern(s) i i( =1, 2, ..., 5) of class g g( = 1, 2, 3). the total number of interactions at class g when using channel j is jgin while the total number of interactions at all classes considered is tn . this is integrated into the lines of authority and responsibility of the production organization in order to minimize lifecycle costs and maximize the benefits and utility of the selected channels j in the industry. 2.3 modelling strategy/framework the generalized interaction relationship based on the authority of personnel and their responsibility (kareem 2014; payne et al. 1996) in the manufacturing organizations is expressed as eqn. (1).  1n1  (1) where, ,n l is the total number of interactions and  is the number of personnel. eqn. (1) shows no restriction in interaction. practically speaking, this is not always possible as effective operation may require confidential information among the top management levels such as directors (fig. 1). the effectiveness of information is weaker down the ladder. ordinarily, there must be a higher level of confidentiality in the information received by managers from managing directors (first class, g = 1) than in the information received by the supervisors from a manager (second class, g = 2). information becomes weakest at the production floor (third class, g = 3). based on the stated characteristics, eq. (1) was modified to take care of three identified categories of interactions, namely direct single (information flows from the managing director to the manager), cross (from manager to supervisors), and direct groups (among supervisors, foremen, or operators) (fig. 1). the cumulative sum of interactions at various levels can be expressed as eqn. (2): j g 5 it i nn  (2) where, nt the total number of interactions in the workplace, j g i n is the total number of interactions in a class or category, g (g = 1, 2, 3) is the collection of network pattern,  5...,,2,1, ii , using a communication channel, ,j  6...,,2,1j . m.d. d d d d s f f o o o o o d, director m.d, managing director s, supervisor f, foreman. o, operator g=1, class 1-wheel, chain (i=1,2) g=2, class 2-y, circle (i=3,4) g=3, class 3-all-channel (i=5) j=1, messenger; j=2, mobile phone; j=3, intranet; j=4, fixed internet; j=5, mobile internet; j=6, pbx g=2 g=3 g=1 j=1 j=6  j=2  development of a system model for determining optimal personnel interaction strategy in a production industry   53   the first class category ( g = 1) was made up of the network pattern wheel and chain,  2,1, ii . the second class category g( = 2) consisted of network pattern y and circle  4,3, ii while the third category g( = 3) comprised an all-channel network pattern  5, ii . the communication channels, j , considered were messengers j( = 1), mobile phone j( = 2), intranet j( = 3), fixed internet j( = 4), mobile internet j( = 5) and pbx j( = 6). 2.4 cost and benefit model the lifecycle cost was calculated for the interaction channels j with reference to data from information technology (it) vendors (kareem 2014, ndb 2007) as: j c c    k 1c j cc (3) where j c c , is the total life cycle costs of using communication channel, j per unit of time. ,c is the counter for elemental costs of the life cycle per unit of time (c=1,2,…,k) for a particular communication channel, j. the life cycle cost elements comprised initial setup cost, acquisition cost, development cost, operations cost, and maintenance cost. the cost of interaction, ,c jt was estimated as the product of lifecycle cost, ,c jc the total number of interactions, nt, and the time taken, .t the optimal channel, j , corresponds to the minimum (min.) of the function in eq. (4). . 4 the benefit function for the interaction channels was estimated as 5 jbb , is the total benefit cost of using communication channel, j, per unit of time, b, is the counter for elemental benefits/savings per unit of time (b = 1, 2…q) for a particular communication channel, j, as utilized by the industry. the benefits/savings comprised overhead, productivity increase, maintenance cost, and operations cost savings. the total benefit of each interaction channel in term of cost savings is: 6 the optimal channel, j, was obtained corresponding to the maximum (max.) benefit and expressed as: . 7 where, b jt  is the total benefit of interaction. the costs and benefits were spread over a period of three years, the time period that was considered one lifecycle, for the interaction channels with a discount rate of 8.75% (kareem, 2014). the net present value (npv) of benefits-costs (bc) was calculated as: pvcpvbnpv  (8) the comparison of the overall costs and benefits and the benefit-to-cost ratio (bcr) was carried out using eq. (9): bcr = dc db (9) represents the present value of cost, which is ∗ , represents the present value of benefit, which is ∗ , represents the cumulative discounted benefits, represents the cumulative discounted costs, represents discount rate, which is a measure of 1 1 (kareem 2014), is the discounted factor, and is the channel life cycle. the channel(s) with the highest bcr was recommended for improved productivity in the industry. any channel having a bcr of less than one could lead to shortages in the organization. post optimality based on utility analysis would assist in determining whether he communication channel(s) could withstand the test of time under usage. b. kareem and a. adekiigbe   5   2.5 network utility performance analysis in order to further ascertain the optimality of the network channel(s) from the benefit-cost ratio (bcr) outcomes, utility performance analysis was carried out. in this sense, the utility of affordability and a usability analysis were carried out to take care of the low-capital nature of the industries in developing economies. in this process, a bcr comparison of the channels was converted to their respective utility values. the objective was to select channel(s) with maximum utility value. the stated objective was achieved by using an ahp (saaty 1990) with an additive utility function (vargas 1990). 2.5.1 hierarchy process model application analytical process the relative specific weights for the communication channels were calculated by the ahp (saaty 1990; vargas 1990). the computations employed by ahp can be decomposed into the following six steps:  a pair-wise comparison matrix j a was formed with the communication channels considered. a                      n nnnn n n w w w w w w w w w w w w w w w w w w w w w w w w ... ......... ... ...... ... 121 2 3 2 2 2 1 2 1 3 1 2 1 1 1 (10)  where each w is the corresponding weight (or importance level) of each individual channel, and each entry within the matrix represents the relative weight ratio between two channels.  the gathered data on level of utility/importance of the channels from industrial experts were converted to the fractional ratio from which the average weight of each channel was calculated in a matrix format.  the matrix product was formed by multiplying, element by element, each row of the first vector, a1,j=1,2,…, (w1/ wn=1,2,…,n) by corresponding elements of the second vector, wj=1,2,…n, and adding. thus, the first element of the product would be: (w1/ w1)* w1+ (w1/ w2)* w2+...+ (11) (w1/ wn)* wn = n w1. similarly, for the second element, the product would be (w2/ w1)* w2+ (w2/ w2)* w2 +.....+ (12) (w2/ wn)* wn = n wn thus, the resulting vector for nth element is nwn.  the eigenvector of matrix a was then computed.  the eigenvector result was standardized by summing up the rows and row totals (this is the first approximation of the eigenvector). 54 development of a system model for determining optimal personnel interaction strategy in a production industry   55                     n nnnn n n w w w w w w w w w w w w w w w w w w w w w w w w ... ......... ... ...... ... 121 2 3 2 2 2 1 2 1 3 1 2 1 1 1 *             n 2 1 w ... w w =             n 2 1 nw ... nw nw   a * w = wn (13)  the system was finally normalized by reiterating steps iii and iv in order to arrive at a consistent matrix with the following features. aik akj = aij for all i, j. (14) in forming the matrix, a comparison of the two attributes was made to take any of the following values: equally important-1; moderately important-2; strongly important-3; very strongly important-4; and extremely important-5. based on the stated model, a comparison of the channels (j) was carried out on the basis of the level of utility attributes (affordability and usability), with additive utility function assigned the value of 0 and 1 to the worst and best conceivable outcomes, respectively. affordability was measured in monetary terms while usability was measured non-monetarily (subjectively) (saaty 2008; vargas 1990). 2.5.2 computation of weighted utility the weighted utility function (saaty 1990) j nu (affordability, a) for the three sub-attributes (n) of acquisition (n = 1), development (n = 2) and operations (n = 3) was represented by the equation 1,2 ,…,6 15 where, j represents the available interaction channel options. similarly, usability in which the three attributes (n) were performance (n = 1), availability (n = 2) and quality (n = 3) was also obtained using the weighted utility function (for usability, s) s j nu and was represented as ∑ 1,2 ,…,6 (16) the cost-benefit ratios across different attributes of affordability and then usability were converted into individual utilities as follows: utility j (bcr) = (bcr – worst value) / (17) (best value – worst value) here, subscript j is the channels for which utility is to be calculated and bcr is the value for that attribute or attribute of usability. the worst value is the lowest bcr value, and the best value is the highest bcr value. a combination of utility by affordability (a) and usability (s) ojnu was used to estimate the overall utility value as: 2⁄ (18) based on the outcome of eq. (18) (utility of affordability and usability), the interaction channel(s) with the highest overall utility value o j nu * will be recommended as the communication network. that is, ∗ . 2⁄ (19) 3. system model implementation an application of the system was carried out with a plastic production industry in osogbo, nigeria, running within a capital base of millions of nigerian naira (n is the symbol for the nigerian currency, and one us dollar is equivalent to approximately n170). the company manufactures polyvinal chloride (pvc) pipes, polyethylene (pe) pipes, fittings, b. kareem and a. adekiigbe   57 table 1. interaction data from the plastic industry. s/n interaction information description 1 organizational structure hierarchical authority 2 interaction patterns circle, chain, y, wheel and all-channel 3 interaction flow downward, upward, & lateral 4 interaction channel memoranda, meetings & face-to-face 5 average production rate (units/year) 1,152,000 6 average production income rate (n/year) 921,600,000 7 staff strength 75 8 average salary and wages (n/year) 40,088,000 9 messengers (no.) 16 10 average salary and wages (16 messengers) including food contributions, transportation costs, social costs, health cost (n/year) 3,063,300 11 average production input (n/year) 841,088,000 12 average fixed capital (n/year) 400,000,000 13 average raw materials costs (n/year) 400,000,000 14 power consumption cost (n/year) 1,000,000 and other plastic products. data were obtained on the organizational structures, including the nature of the division of labor, the type of hierarchical authority, and job descriptions. job descriptions included information about whether jobs were centralized or decentralized. data were also collected on whether the organization was divisional, functional, matrixed, or networked, and on the type of reporting relationships or reward structure that existed, channels and media of interaction, and workforce size. it was noted whether industries utilized messengers to transfer messages from one department/unit to another, and whether the company would like to increase its productivity by finding alternative ways of message transfer by using other means besides messengers. the data collected were tabulated and entered into table 1. the organization chart, including personnel interactions and lines of authority and responsibility to the company, was similar to one presented in fig. 1. the implementation process was carried out in steps. first, interaction levels of the three classes (i.e. first, second, and third) of personnel in the production industry based on its hierarchical structure (organisational charts) were determined using eq. (2). second, with a channels lifespan of three years with discount rates of 8.75%, the cost and benefit of adopting modern communication channels/media in the industry compared to the traditional use of messengers was evaluated using eqns. (3–9). third, the ahp technique was applied to find the utility values of affordability and usability of the channels using the cost and benefit outcomes as presented in eqns. (10–19). last, combinations of communication channels’ effectiveness as related to productivity were assessed and compared monetarily and subjectively using a cost-benefit analysis and ahp techniques, respectively. appendix a (i– iii) shows a detailed analysis of the data using ahp technique. the information used in the ahp was obtained from industrial experts using interviews and questionnaires on their views about the various factors (attributes) under consideration. table 2a shows the communication channels’ final output weighting of the utility values for affordability (acquisition cost/overhead saving, development cost/productivity cost, and operation cost/maintenance savings), and usability (performance, availability, and quality) using ahp during which productivity and availability criteria were given the highest weighting. 4. results and discussion the comparison of the overall cost, benefits, and the benefit-cost ratio is shown in table 2b as compared to messengers as a traditional 56 development of a system model for determining optimal personnel interaction strategy in a production industry   57 table 2a. output weights of the channels’ utility attributes using ahp. affordability utility attributes weight (wn) values acquisition costs/overhead savings w1 0.23 development costs/productivity costs w2 0.58 operations cost/maintenance savings w3 0.19 total 1.00 usable utility attributes performance w1 0.23 availability/reliability w2 0.58 quality w3 0.19 total 1.00 table 2b. comparison of discounted benefit-cost ratio (bcr). relative value comparison channel type discounted cost (n) discounted benefit (n) discounted net (n) bcr mobile phone 3053881.89 13430040.63 10376158.74 4.39 intranet 2056085.64 15615168.51 13559082.87 7.59 fixed internet 3684105 14429658.34 10745553.34 3.91 mobile internet 3931563.14 16837339.42 12905776.28 4.28 pbx system 2023185.74 12248106.76 10224921.02 6.05 bcr = benefit-cost ratio; pbx = private branch exchange table 3. corresponding utility values for various combinations of channels. channel type performance availability quality mobile phone 0.06367 1.00000 0.00000 intranet 0.25093 0.25093 0.25093 fixed internet 0.06367 0.25093 0.25093 mobile internet 0.06367 1.00000 0.25093 pbx 0.25093 0.25093 0.00000 pbx = private branch exchange communication channel. from the table, mobile internet contributed highest to the cost of production (n3931563.14) while the pbx phone system contribute the lowest to costs (n2023185.74). management decisions based on minimum production cost contributions will most likely identify a pbx system as an optimal choice. similarly, mobile internet contributed the highest benefits to production in terms of revenue generation (n16837339.42), while a pbx system contributed the lowest benefits (n12248106.76). based on maximum benefits, the use of mobile internet would be the superior choice. however, the highest discounted net benefit (n13559082.87) was obtained through an intranet channel, and a pbx system was the lowest (n10224921.02). therefore, the choice of an intranet channel was supportive of productivity of the industry at this level. similarly, an evaluation based on a bcr indicated that intranet, having the highest bcr (7.59), would be the best channel for improved productivity. the corresponding utility values for the combinations of the channels are presented in table 3. performance-based utility results showed that intranet and pbx were better for productivity than a combination of mobile phones, fixed internet, and mobile internet with a higher probability of 0.25093. diverse results were obtained from utility of availability. in this case, the mobile phone and mobile internet were found to be better with a probability close to unity because of the usual network availability, than the combination of an intranet, fixed internet, and pbx system. when considering quality, only the intranet, fixed internet, and mobile internet were found to be b. kareem and a. adekiigbe   58   promising for production sustainability in a situation dominated by moderate network quality. the utility values of affordability and usability, and their overall utility are shown in table 4. the overall utility of interaction channels showed that mobile phone networks have the best overall utility value of 0.59. the high utility of mobile phones has resulted in its categorization of it as the best network among a combination of channels usable to enhance productivity. based on the subjective results obtained from the developed system, availability was the leading factor which made the mobile phone and mobile internet networks the best candidates for productivity enhancement in industry (table 3). however, mobile phones were the best channel to adopt (with 0.59 probability) according to an overall utility assessment (table 4). figure 2 indicated that mobile internet possessed the highest usability, followed by mobile phones, the intranet, fixed internet, and a pbx system. 5. conclusion information management has a prominent role to play in effective operation in the manufacturing industry. understanding the organizational structure of a company in terms of active interaction among personnel is paramount for effective evaluation of communication channels that, apart from cost savings, will enhance productivity. in this study, an efficient system model was developed to determine optimal personnel interaction strategies in the plastic production industry. in developing the model, interaction levels of the categories of the personnel in the production industry were obtained. a study of the cost and benefit of adopting modern communication channels/media in industry aside from traditional human messengers was carried out. ahp was applied to find the utility values of affordability and usability of the network channels. the outcomes from the application of the model to the plastic industry indicated that the mobile internet channel, in terms of a bcr, had the highest benefits while the pbx had the lowest. the intranet channel had the highest overall bcr while the mobile phone and mobile internet had the highest utility in terms of affordability and usability, respectively. the findings generally showed that intranet, mobile phones, and mobile internet were the best three productivityenhancing networks, and they could adequately replace traditional communication methods in industry. further studies to investigate system robustness in response to problems associated with staff of different sizes and complexities are encouraged in the future. this is necessary to develop a generalized system, which can accommodate varying organizational complexities and attitudes associated with personnel. table 4. utility values of communication channels. channel type affordability usability overall utility values mobile phone 0.58 0.59 0.59 intranet 0.27 0.25 0.26 fixed internet 0.28 0.21 0.25 mobile internet 0.35 0.64 0.50 pbx 0.45 0.07 0.26 pbx = private branch exchange figure 2. comparison of utility of usability. 0 0.5 1 interaction channelso rd e r o f u sa b il it y comparison of utility of usability mobile phone intranet fixed internet development of a system model for determining optimal personnel interaction strategy in a production industry   57 acknowledgment the author acknowledges the federal university of technology, akure, nigeria, for providing the necessary facilities for the successful completion of this study. the management of the plastics production industry in osogbo, nigeria, is also acknowledged for releasing its data for this study. references adams sm., sarkis j, liles dh (1995), a tool for the development of strategic performance metrics for enterprise activities. proceedings of the fourth annual industrial engineering research conference, institute of industrial engineers. nashville, tn. balschbach g, klinke j, jaehne b (1998), multichannel shape from shading techniques for moving specular surfaces. lecture notes in computer science 1407: 170-184. bhattacharjee d (2007), a proposed cost-benefit analysis model for physical form analysis for a futuristic submarine decision support system. unpublished msc. thesis, system design and management program, massachusetts institute of technology. biggiero l (1999), complexity and organization in modern day. international handbook of it/s, italian conference on systemics. apogeo. milan. cevasco j, marmolejo rf (2013), the importance of studying prosody in the comprehension of spontaneous spoken discourse. revista latinoamericana de psicologia 45(1): 21-33. cia world factbook (2011), https://www.cia.gov / library / publications / the-world-factbook / geos.html access on 27th august, 2011. drucker pf (2002), the discipline of innovation. harvard business review 95-98. feital t, kruger u, dutra j, pinto jc, lima el (2013), modelling and performance monitoring of multivariate multimodal processes, aiche journal 59(5):1557-1569. finnie gr, wittig ge, petkov di (1993), prioritizing software development productivity factors using the analytic hierarchy process. journal of systems and software 22(2): 129-139. froehle cm (2006), service personnel, technology, and their interaction in influencing customer satisfaction. decision sciences journal 37(1): 5–38. gurbaxani v, plice r (2004), a model of network-centric organizations. working paper uci crito consortium, http:// www.crito.uci.edu/consortium accessed 30th november, 2009. hurley s (2008), the shared circuits model (scm): how control, mirroring, and simulation can enable imitation, deliberation, and mindreading. behavioural and brain sciences 31(1): 1-22. jackson p, skirrow p, hare dj (2012), asperger through the looking glass: an exploratory study of self-understanding in people with asperger's syndrome. journal of autism and developmental disorders 42(5): 697-706. jaques e (1998), requisite organization, a tool system for effective managerial organization and managerial leadership for 21st century. cason hall and co pub. north america, usa. kareem b (2014), cost –benefit model for optimal selection of communication channels in manufacturing organisation. ned university journal of research-applied sciences xi(2): 1-13. kareem b (2015), evaluation of failures in mechanical crankshafts of automobile based on expert opinion. case studies in engineering failure analysis 3(1): 25-33. kleinbaum am, stuar te, and tushman ml (2008), communication and coordination in a modern, complex organization. http://www.hbswk.hbs.edu/item/5991.ht ml accessed 30th november, 2009. kuboye bm., alese bk fajuyigbe o (2009), congestion analysis on the nigerian global system for mobile communications (gsm) network. the pacific journal of science and technology 10(1): 262-271. larkin tj, larkin s (2005), communicating big change: using small communication (2nd edition). larkin communication consulting. li c, pickup d, saunders t, cosker d, marshall d, hall p, willis p (2013), water surface modelling from a single viewpoint video. ieee transactions on visualization and computer graphics 19 (7): 1242-1251. lievrouw la (2008), communication. microsoft encarta 2009 [dvd]. redmond, wa: microsoft corporation, 2008. 59 b. kareem and a. adekiigbe   60   lu l, yu x (2012), a multimedia teaching method research of animation production in china, iccse 2012 proceedings of 2012 7th international conference on computer science and education. 1495-1497. mackenzie kd (1978), organizational structure. ahm publishing corporation. ahmedabad, india. ndb-nigeria data bank (2007), nigeria gsm statistic. december 2007. nohria n, eccles r (1992), face-to-face: making network organisations work. in: nohria n. and r. eccles (eds.) networks and organisations: structure, form and action. boston. nohria n, ghoshal s (1997), the differentiated network. organizing multinational corporations for value creation. josseybass, san francisco. mtega wp, msungu ac (2013), using information and communication technologies for enhancing the accessibility of agricultural information for improved agricultural production in tanzania. electronic journal of information systems in developing countries 56(1): 1-14. nagy m (2013), a multi-channel communication framework. communications in computer and information science 347: 72-88. nucci m, grassi m, piazza f (2013), ontologybased device configuration and management for smart homes. smart innovation, systems and technologies 19: 301310 nyambo b, ligate e (2013), smallholder information sources and communication pathways for cashew production and marketing in tanzania: an ex-post study in tandahimba and lindi rural districts, southern tanzania. journal of agricultural education and extension 19 (1): 73-92. olokede ss (2009), reliability analysis of fixed wireless networks in nigeria. the pacific journal of science and technology 10(1): 251255. opperman sb (2013), modernization and rural health in mexico: the case of the tepalcatepec commission. endeavour 37 (1): 47-55. payne ac, chelsom jv, reavill lrp (1996), management for engineers. john wiley and sons, england. saaty tl (1990), how to make a decision: the analytic hierarchy process. european journal of operational research 48(1): 9-26. saaty tl (2008), decision making with the analytic hierarchy process. international journal of services and sciences 1(1): 83-98. shannon ce (1948), a mathematical theory of communication. the bell system technical journal 27(6): 379-423. sooksumrarn p, krairiksh m (2013), uhf wireless communication channel in a tree canopy. ieee antennas and propagation society, ap-s international symposium (digest) article no. 6393914: 311-314 stohl c (1995), organizational communicationconnectedness in action. sage publications london, new delhi. urwick lf (1943), personnel management in relation to factory organization. l. urwick institute of labour management, london. urwick lf (1974), the manager’s span of control. the academy of management journal 10(1): 349-354. vargas lg (1990), an overview of the analytic hierarchy process and its applications. european journal of operational research 48(1): 2-8. wada s, sasaoka n, itoh y, okello j, kobayashi m (2013), speech enhancement based on bias free noise reconstruction method. ispacs 2012 ieee international symposium on intelligent signal processing and communications systems, article no. 6473537: 479-484. yates ja, orlikowski wj, okamura k (1999), explicit and implicit structuring of genres in electronic communication: reinforcement and change of social interaction. organization science 10(4): 83-103. zhao y, xue x, zhang t (2013), receiverchannel based adaptive blind equalization approach for gps dynamic multipath mitigation 2013. chinese journal of aeronautics 26(2): 378-384. development of a system model for determining optimal personnel interaction strategy in a production industry   61 appendix a (i, ii & iii): utility and performance analysis i. determining the utility of affordability: the utility of affordability–individual weights the analytical hierarchy process (ahp) was used to calculate the individual weights among the various attributes. for comparing two attributes, a scale of 1–5 was used to make comparisons proposed in the ahp. among the three attributes of affordability (acquisition, development, and operations), the ranking of the various attributes are assumed as follows: development is as important as operations; acquisition is much more important than development, and acquisition is moderately more important than operations. the order of the three attributes is acquisition followed by development and operations. a 3 x 3 matrix is required for comparing three attributes. the 3 x 3 eigenvalue matrix for the preferences stated above takes the following form:                  0.13231 23 0.121 13120.1 3x3 aa aa aa a (a1) here, a12 refers to comparing acquisition over development. similarly, a32 refers to comparing operations over development. based on the preferences of the attributes, the pair-wise comparison of the attributes would be as follows:            0.125.05.1 0.30.10.2 5.15.00.1 (a2) the normalized matrix is determined by dividing the values in each column by the sum of the column as follows:            15.014.027.0 62.057.054.0 23.029.015.0 (a3) now, the eigenvector is formed as the average of each normalized row:                         19.0 58.0 23.0 3 2 1 1x3 w w w w (a4) finally, the eigenvector is the weights of the three attributes using the ahp model where the weights of all the attributes add up to 1. the different weights are as follows: weight(acquisition costs/overhead savings) = 0.23 weight(development cost/productivity increase) = 0.58 weight(operations costs/maintenance savings) = 0.19 these weights, along with the individual utilities calculated, are taken together for calculating the final utility using the additive utility function. utility of affordability – individual utility: b. kareem and a. adekiigbe   62   utility j (x) = (x – worst value) / (best value – worst value) (a5) utility mobile phone (acquisition) = (2.62 – 1.08)/(331.23 – 1.08) = 0.00467 utility mobile phone (development) = (331.23 – 1.08))/ (331.23 – 1.08) = 1.00000 utility mobile phone (operations) = (1.15 – 1.08)/ (331.23 – 1.08) = 0.00021 utility intranet (acquisition) = (2.33 -1.08)/ (331.23 – 1.08) = 0.00378 utility intranet (development) = (152.37 -1.08)/ (331.23 – 1.08) = 0.45553 utility intranet (operations) = (10.83 – 1.08)/ (331.23 – 1.08) = 0.02955 utility fixed internet (acquisition) = (1.15 – 1.08)/ (331.23 – 1.08) = 0.00021 utility fixed internet (development) = (158.99 – 1.08)/ (331.23 – 1.08) = 0.047869 utility fixed internet (operations) = (1.26 – 1.08)/ (331.23 – 1.08) = 0.00055 utility mobile internet (acquisition) = (1.08 – 1.08)/ (331.23 – 1.08) = 0.00000 utility mobile internet (development) = (198.73 -1.08)/ (331.23 – 1.08) = 0.59909 utility mobile internet (operations) = (3.05 -1.08)/ (331.23 – 1.08) = 0.00597 utility pbx (acquisition) = (2.03 – 1.08)/ (331.23 – 1.08) = 0.00289 utility pbx (development) = (248.42 – 1.08)/ (331.23 – 1.08) = 0.74969 utility pbx (operations) = (24.24 – 1.08)/ (331.23 – 1.08) = 0.07019 these are the individual utility values for the various attributes categorized under various interaction channels. the next step is to calculate the individual weights. utility of affordability – additive utility: the weighted utility function for the three attributes of acquisition, development, and operations is represented by the equation: utility j (acquisition, development, operations) = weight (acquisitions) * utility i (acquisitions) + weight (development) * utility j (development) + weight (operations) * utility j (operations) (a6) here subscript j represents the various available options: mobile phone, intranet, fixed internet, mobile internet, and pbx systems. based on this additive utility function, the utility of the various options is as follows: utility mobile phone (acquisition, development, operations) = 0.23 * 0.00467 + 0.58 * 1 + 0.19 * 0.00021 = 0.5811141 utility intranet (acquisition, development, operations) = 0.23 * 0.00378 + 0.58 * 0.4553 + 0.19 * 0.02955 = 0.2705579 utility fixed internet (acquisition, development, operations) = 0.23 * 0.00021 + 0.58 * 0.47869 + 0.19 * 0.00055 = 0.277793 utility mobile internet (acquisition, development, operations) = 0.23 * 0.0000 + 0.58 * 0.59909 + 0.19* 0.00597 = 0.3485563 utility pbx (acquisition, development, operations) = 0.23 * 0.00289 + 0.58 * 0.74969 + 0.19 * 0.07019 = 0.4488211 therefore, the cbrs obtained earlier were converted to equivalent utility values. the utility values of affordability for the various display options are as follows: utility mobile phone (affordability) = 0.58 utility intranet (affordability) = 0.27 utility fixed internet (affordability) = 0.28 development of a system model for determining optimal personnel interaction strategy in a production industry   63   utility mobile internet (affordability) = 0.35 utility pbx (affordability) = 0.45 ii. utility of usability utility of usability for the channels of interaction were obtained in similar way as that of affordability. the outcomes are as follows: utility mobile phone (usability) = 0.59 utility intranet (usability) = 0.25 utility fixed internet (usability) = 0.21 utility mobile internet (usability) = 0.64 utility pbx (usability) = 0.07 iii. overall utility value in this paper, it was assumed that affordability is equally important to usability. since both the attributes are equally important, the ahp model will generate the same weight for both attributes. therefore, weight (affordability) = 0.5, and weight (usability) = 0.5. using the additive utility function, the overall utility values are as follows: overall utility mobile phone (affordability, usability) = 0.5 * 0.58 + 0.5 * 0 .59 = 0.585 overall utility intranet (affordability, usability) = 0.5 * 0.27 + 0.5 * 0.25 = 0.26 overall utility fixed internet (affordability, usability) = 0.5 * 0.28 + 0.5 * 0.21 = 0.245 overall utility mobile internet (affordability, usability) = 0.5 * 0.35 + 0.5 * 0.64 = 0.50 overall utility pbx (affordability, usability) = 0.5 * 0.45 + 0.5 * 0.07 = 0.26 therefore, the order of utility is as follows: mobile phone, mobile internet, intranet, pbx, and fixed internet. c:\users\raoh\desktop\paper 4.xps the journal of engineering research vol. 12, no. 1 (2015) 44-52 microwave-assisted adsorptive desulfurization of model diesel fuel using synthesized microporous rare earth metal-doped zeolite y n. salahudeena, a.s. ahmeda, a.h. al-muhtaseb*b, b.y. jibrila, r. al-hajria, s. m. waziria, m. daudac, j. al-sabahid a department of chemical engineering, ahmadu bello university, zaria, nigeria. b department of petroleum and chemical engineering, college of engineering, sultan qaboos university, muscat, oman. c department of mechanical engineering, ahmadu bello university, zaria, nigeria. d central instrument laboratory, college of agricultural and marine sciences, sultan qaboos university, muscat, oman. received 21 november 2014; accepted 30 december 2014 abstract: the microwave-assisted adsorptive desulfurization of model fuel (thiophene in n-heptane) was investigated using a synthesized rare earth metal-doped zeolite y (re y). crystallinity of the synthesized zeolite was 89.5%, the silicon/aluminium (si/al) molar ratio was 5.2, the brunauer–emmett–teller (bet) surface area was 980.9 m2/g, and the pore volume and diameter was 0.3494 cm3/g and 1.425 nm, respectively. the results showed that the microwave reactor could be used to enhance the adsorptive desulfurization process with best efficiency of 75% at reaction conditions of 100 °c and 15 minutes. the high desulfurization effect was likely due to the higher efficiency impact of microwave energy in the interaction between sulfur in thiophene and ho-la(osial). keywords: desulfurization; thiophene; microwave; diesel; zeolite re y. re y re y y89.5si/al5.2 bet980.90.3494 1.425 7510015 ho-la(osial) re y corresponding author’s email: muhtaseb@squ.edu.om nurudeen salahudeen, abdulkarim s. ahmed, ala’a h. al-muhtaseb, baba y. jibril, rashid al-hajri, saidu m. waziri, mohammed dauda, jamal al-sabahi 1. introduction diesel fuel has become one of the most popular energy sources for driving industrial, domestic, and transportation engines due to its high efficiency (dasgupta et al. 2013). the exponentially growing demand for diesel has increasingly heightened the environmental concerns arising from environmental pollution caused by diesel engine exhaust (hans et al. 1989; lin et al. 2009; zhang et al. 2001). diesel engine exhaust contains high levels of sulfur-containing gasses in the form of sulphur oxides (sox), which have been identified as detrimental to human health and the environment. these detrimental effects include adverse reactions in the human respiratory system, contributions to the greenhouse effect, and acid rain. most developed countries in the world such as the usa, canada, and japan, and various countries in europe have enacted environmental regulatory policies to control sulphur levels in the diesel used by highway vehicles and restrict emissions to a bare minimum. in the usa, for instance, the environmental protection agency (usepa) has mandated a maximum sulpur content of 15 ppm in highway diesel since 2006 and, in europe, a maximum sulphur content of 10 ppm was instituted as the eu-wide limit for diesel used in road transportation (dasgupta et al. 2013). due to these increasingly stringent regulations, the desulfurization of diesel fuel has become an important research subject. an earlier industrial technique for the desulfurization of diesel was the hydro-desulfurization (hds) process (babich and moulijin 2003; duarte et al. 2011; zannikos et al. 1995). in hds, sulfur-containing compounds were converted to hydrogen sulfide (h2s) and other hydrocarbons via extreme temperature and pressure conditions using hydrogen as the converter. therefore, the hds process was not only very expensive but was also environmentally dangerous due to the highly inflammable nature of hydrogen. recently, several alternative, greener desulfurization processes have been proposed (ahmad et al. 2009; marin-rosas et al. 2010; salem and hamid 1997; seredych and bandosz 2010; shakirullah et al. 2009; velu et al. 2003; weitkamp et al. 1991). some of the techniques reported include extractive desulfurization (eds), selective adsorptive desulfurization (sad) (subhan et al. 2012), and oxidative desulfurization (ods) (dharaskar et al. 2014; lü et al. 2014; wei et al. 2009). some of these techniques have limitations. for instance, the extractant used in the ods and eds processes are usually flammable and contain volatile organic compounds (vocs) which are also known for their environmental and safety impacts (gao et al. 2010; gui et al. 2010). the adsorption-based approaches appear particularly attractive due to their simpler flow sheets and lower hydrogen requirements (ahmad et al. 2014; dasgupta et al. 2013; hussain et al. 2012; sisani et al. 2014). due to the possibility of selective heating, microwave energy has attracted considerable attention in the field of desulphurization, where it has been found that microwave heating makes the process more effective. with microwaves, energy can be delivered to reacting molecules at a much greater rate compared to conventional heating, which is a more efficient way to heat since microwaves allow faster and more uniform heating. ibe et al. (2007) applied microwave heating to facilitate the desulphurization of poisoned mono-nitrogen oxide (nox) storagereduction (nsr) catalysts. they concluded that high catalyst temperatures did not need to be reached in order to achieve the desorption of sulfur species. therefore, the use of microwave heating provided an efficient way to promote the regeneration of the nsr catalysts. wei et al. (2009) investigated the simultaneous desulfurization and denitrification from stimulated flue gas by a microwave reactor with potassium permanganate and zeolite. they found that the microwave reactor could be used in the oxidation of sulphur dioxide (so2) to sulfate with a desulphurization efficiency of 96.8%. the addition of zeolite to microwave potassium permanganate increased the microwave removal efficiency of so2 from 16.5% to 43.5%. microwave-assisted oxidative desulfurization of sour natural gas condensate via a combination of sulfuric and nitric acids was investigated by moaseri et al. (2014). they found that that the ods efficiency of sore condensate 45 microwave-assisted adsorptive desulfurization of model diesel fuel using synthesized microporous rare …. 46 figure 1. flow diagram for the synthesis of zeolite y. by the combination of sulfuric acid (h2so4) and nitric acid (hno3) was improved by 30% in the presence of microwave radiation. this work was aimed at the desulfurization of model fuel using a cost effective and environmentally friendly technique. the suitability of a synthesized zeolite y impregnated with lanthanum metal for microwave-assisted selective adsorptive desulfurization of a model fuel was investigated. 2. materials and methods 2.1 synthesis of zeolite nay zeolite y in the nay form was synthesized by adapting the verified method of the synthesis commission of the international zeolite association (harry, 2001; ginter et al. 1992). this process involved preparation of seed and feedstock gels, and a general gel using sodium hydroxide, sodium aluminate, aluminium hydroxide, and distilled water. this general gel was crystallized through a template-free, low temperature crystallization method inside a polypropylene bottle (fig. 1). 2.2 transformation to zeolite re y the synthesized zeolite nay was protonated by an ion exchange reaction to make zeolite hy. a 1.0 m solution of nh4cl was prepared and exchanged with zeolite nay at 80 °c using a ratio of 10 ml of solution to 1 g of solid and stirred for 20 minutes (du et al. 2013; gao et al. 2012). after the exchange reaction, the slurry was suction filtered using a vacuum pump and washed thoroughly with distilled water, then dried at 110 °c for six hours. the zeolite hy formed was impregnated with rare earth metal (la). the reaction was carried out using an ion exchange wet-impregnation method: a 0.5 m solution of lanthanum chloride (lacl3) was prepared and mixed with the protonated zeolite (hy) at 80 °c using a ratio of 12.5 ml of solution to 1 g of solid and stirring the solution for one hour (du et al. 2013; gao et al. 2012). the product was then filtered using a vacuum pump and washed thoroughly with distilled water. the wet cake of zeolite was hydrothermally calcined at n. salahudeen, a.s. ahmed, a.h. al-muhtaseb, b.y. jibril, r. al-hajri, s. m. waziri, m. dauda, j. al-sabahi 47 650 °c in air for two hours, resulting in zeolite re y. 2.3 characterization x-ray diffraction (xrd) patterns were measured using a miniflex 600 x-ray diffractometer at 40 kv and 10 ma with cu kα radiation (λ = 0.15418 nm,) (rigaku, tokyo, japan). the samples were scanned from 20–80 °c (2θ) at a rate of 0.02 °c/min. x-ray fluorescence (xrf) analysis was carried out using a minipal 4 x-ray fluorescence analyser (panalytical b.v., almelo, the netherlands). fourier transform infrared (ftir) spectroscopy was applied mainly to study the chemical structure of the catalysts. the infrared spectra of the catalyst samples were measured at wavelengths ranging from 400–4000 cm−1 using a frontier ft-ir spectrometer (perkinelmer, waltham, massachusetts, usa). the morphology of the samples was analyzed with field emission scanning electron microscope jsm7600f (jeol, inc., tokyo, japan) and by a tem1230 high contrast transmission electron microscope (tem) (jeol). the bet surface area and total pore volume were measured using autosorb-1 nitrogen adsorption apparatus (quantachrome instruments, boynton beach, florida, usa). all samples were degassed at 300 °c for four hours prior to the measurement. 2.4 desulfurization of model diesel the re y form of the synthesized zeolite y was used as the catalyst for the desulfurization of the model diesel using the microwave-assisted technique. the model oil used was prepared by dissolving 200 ppm of thiophene (sigma-aldrich, co., st. louis, missouri, usa) in n-heptane (sigma-aldrich) and pouring 10 g of model oil into the pressure vessel of a microwave reactor (milestone, inc., sorisole, italy) and 1 wt% of the re y zeolite was also added. the pressure vessel was torqued and inserted in the workstation of the reactor. the reactor’s thermocouple was then inserted in the thermowell and the reactor was programmed for variable time and temperature runs. after each run, the reactor content was allowed to cool for 30 minutes. at the end of this process, the torque was loosened, and the vessel content was removed. the resulting mixture was centrifuged at 10,000 rpm for 10 minutes in an avantitm j-251 high performance centrifuge (beckman coulter, brea, california, usa). the resulting raw liquid product was analysed using a clarus 600c gc/ms system (perkinelmer). 3. results and discussion 3.1 catalyst characteristics figure 2 shows the xrd patterns of the commercial and synthesized zeolite y. the zeolite y peaks could be observed in both xrd patterns at bragg’s angles of 6, 10, 12, 16, 19, 20, 24, 27, 31, and 32° (harry 2001; ginter et al. 1992; treacy and higgins, 2001). the intensities of the characteristic peaks for zeolite y at bragg’s angles of 6.13, 10.0, 11.8, and 15.5˚ were 1795, 718, 527 and 693 counts, respectively, for the synthesized zeolite nay, whereas those of the commercial zeolite y were 2176, 847, 416, and 698 counts, respectively. therefore, using the commercial zeolite y as a reference, and using equation (1) (pal et al. 2013), the crystallinity of the synthesized zeolite y was evaluated as 89.5%. figure 2. xrd patterns of the synthesized and commercial zeolite y. table 1 presents the xrf result of the synthesized zeolite y, reflecting the fact that the zeolite contained low concentrations of impurities, with a measurement generally below 1 wt%. the si/al molar ratio was 5.2 and the loss on ignition (loi) was 26 wt%. figures 3a–d show the structural mophology of the synthesized zeolite nay at magnifications of 8,000, 15,000, 30,000, and 45,000, respectively. the macrographs show that the zeolite y at magnifications of 100,000, 200,000, 300,000 and 400,000 times. the clear boundary shapes of the crystals are very obvious. although the crystal microwave-assisted adsorptive desulfurization of model diesel fuel using synthesized microporous rare …. 46 crystallinity intensity of the characteristic xrd peak of the productintensity of the characteristic peak of the reference xrd pattern×100 (1) table 1. chemical composition of the synthesized zeolite y. metal oxides wt % al2o3 17.98 sio2 54.75 k2o 0.04 cao 0.27 tio2 0.09 cr2o3 0.04 mno 0.03 fe2o3 0.3 cuo nd loi 26 nd: not determined figure 3. sem images for the synthesized zeolite nay at various magnifications: (a) 8000; (b) 15000; (c) 30000 and (d) 45000. 48 n. salahudeen, a.s. ahmed, a.h. al-muhtaseb, b.y. jibril, r. al-hajri, s. m. waziri, m. dauda, j. al-sabahi 50 table 2. bet results for the synthesized zeolite y. sample surface area (m2/g) pore volume (cm3/g) pore diameter (å) -synthesized 980.9 0.3494 14.25 commercial 601.8 0.2500 16.40 figure 4. tem images for the synthesized zeolite nay at various magnifications: (a) 100000; (b) 200000; (c) 300000 and (d) 400000. average crystal size was estimated as 200 nm. figure 4 also shows the tem images of the synthesized zeolite y at magnifications of 100,000, 200,000, 300,000 and 400,000 times. the clear boundary shapes of the crystals are very obvious. although the crystal images overlap extensively, the boundary edges of the unit tetrahedral crystal shape is obvious. the average crystal size was estimated as 200 nm. bet analysis for the synthesized and commercial zeolite y shows that the specific surface area of the synthesized zeolite y was 63% higher and the pore volume was 40% higher than that of the commercial zeolite y (table 2). however, the pore diameter of the commercial zeolite y was 15% higher than that of the synthesized zeolite y. 3.2 desulfurization of model diesel the sulfur reduction at variable reaction times and temperatures for the microwave-assisted desulfurization of model petroleum oil using the prepared zeolite re y catalyst is shown in figs. 5 and 6, respectively. figure 5 shows the sulfur 49 microwave-assisted adsorptive desulfurization of model diesel fuel using synthesized microporous rare …. reduction for reactions carried out at 100 °c at variable times. the maximum sulfur reduction was observed at 15 minutes when the sulphur reduction was 74%. figure 6 shows the sulfur reduction for the reactions carried out at 10 minutes for variable temperatures. there was a continuous progressive increase in sulfur reduction throughout the runs, with the maximum sulphur reduction (74%) occurring at 150 °c. figure 5. trend of microwave desulfurization of model diesel at variable time. figure 6. trend of microwave desulfurization of model diesel at variable temperature. figures 7 and 8 show the ftir spectra of the synthesized catalyst before and after the reaction. the spectra were similar except that the spent zeolite showed the presence of strange bands at wavelengths 1450 and 2930 cm-1 for both the variable temperatures and time curves. these bands were due to the adsorption of sulfur (song et al. 2013) and were completely absent in the fresh catalyst spectrum. it could also be observed that for the variable time runs, the sample at 15 minutes showed a larger band area while for the variable temperature runs, the sample at 150 °c showed the largest band area. hence, the synthesized catalyst showed a potent effect on the desulfurization of the model diesel fuel, while the ftir analysis of the catalyst before and after the reactions showed that the mode of desulphurization employed by the catalyst was adsorptive desulfurization. the extent of desulfurization recorded by this work compares closely with the work of fallah et al. (2015), who reported using a different approach. figure 7. ftir spectra of the fresh and used catalysts for the desulfurization at variable time. figure 8. ftir spectra of the fresh and used catalyst for the desulfurization at variable temperatures. 4. conclusions a microporous zeolite y having 89.5% crystallinity, a 980.9 m2/g surface area, a 0.3494 cm3/g pore volume, and a 200 nm crystal size was synthesized. the si/al molar ratio of the synthesized zeolite was 5.2 and the loi was 26%. the synthesized zeolite was effective for adsorptive desulfurization of a model fuel, having the best efficiency of 75% in a microwave-assisted desulfurization carried out at 100 °c for 15 minutes. the high desulfurization effect was likely 5150 n. salahudeen, a.s. ahmed, a.h. al-muhtaseb, b.y. jibril, r. al-hajri, s. m. waziri, m. dauda, j. al-sabahi 53 due to the higher efficiency impact of microwave energy in the interaction between sulfur in thiophene and ho-la(osial). acknowledgments the authors gratefully acknowledge the financial and technical support of the petroleum development trust fund (pdtf) of abuja, ahmadu bello university in zaria and the college of engineering at sultan qaboos university in oman. references ahmad w, ahmad i, ishaq m, ihsan k (2014), adsorptive desulfurization of kerosene and diesel oil by zn impregnated montmorollonite clay. arabian journal of chemistry, in press. babich iv, moulijn ja (2003), science and technology of novel processes for deep desulfurization of oil refinery streams: a review. fuel 82:607–631. dasgupta s, gupta, pa, nanoti a, goswami an, garg mo, tangstad e, vistad ob, karlsson a, stöcker m (2013), adsorptive desulfurization of diesel by regenerable nickel based dsorbents. fuel, 108:184–189. dharaskar sa, wasewar kl, varma mn, shende dz, tadi kk, yoo ck (2014), synthesis, characterization, and application of novel trihexyl tetradecyl phosphonium bis (2,4,4trimethylpentyl) phosphinate for extractive desulfurization of liquid fuel. fuel processing technolog 123:1–10. du x, zhang , li x, tan z, liu h, gao x (2013), cation location and migration in lanthanumexchanged nay zeolite. chinese journal of catalysis 34:1599–1607. duarte fa, mello pa, bizzi ca, nunes mag, moreira em, alencar ms, motta hn, dressler v, flores em (2011), sulfur removal from hydrotreated petroleum fractions using ultrasound-assisted oxidative desulfurization process. fuel 90,:2158–2164. fallah rn, azizian s, dwivedi ad, sillanpää m (2015), adsorptive desulfurization using different passivated carbon nanoparticles by peg-200. fuel processing technology 130:214– 223. gao hs, guo c, xing jm, zhao jm, liu hz (2010), extraction and oxidative desulfurization of diesel fuel catalyzed by a brønsted acidic ionic liquid at room temperature. green chemistry 12:1220–1224. gao x, qin z, wang b, zhao x, li l, zhao h, liu h, shen b (2012), high silica rehy zeolite with low rare earth loading as high-performance catalyst for oil conversion. applied catalysis a: genetics 413:254–260. ginter dm, bell a, radke cj (1992), synthesis of microporous materials, molecular sieves, vol. 1. new york: van nostrand reinhold 6. gui jz, liu d, sun zl, liu ds, min d, song b, peng xl (2010), deep oxidative desulfurization with task-specific ionic liquids: an experimental and computational study. journal of molecular catalysis a: chemical 331:64–70. hans b, annemie b, gerd e, ruud s, julian rhr (1989), lithium-vanadium bronzes as model catalysts for the selective reduction of nitric oxide. catalysis today 4:139–154. harry r (2001), verified synthesis of zeolitic materials; linde type y, synthesis commission of the international zeolite association. amsterdam: elsevier. hussain m, abbas n, fino d, russo n (2012), novel mesoporous silica supported zno adsorbents for the desulfurization of biogas at low temperatures. chemical engineering journal 188:222–232. ibe m, gomez s, malinger ka, fanson p, suib sl (2007), microwave-assisted desulfurization of nox storage-reduction catalyst. applied catalysis b: environmental 69:235–239. lin zz, niu h, he h, ji y (2009), simultaneous desulfurization and denitrification by microwave reactor with ammonium bicarbonate and zeolite. journal of hazardous materials 162:837–841. lü h, deng c, renb w, yang x (2014), oxidative desulfurization of model diesel using [(c4h9)4n]6mo7o24 as a catalyst in ionic liquids. fuel processing technology 119:87–91. marin-rosas c, ramirez-verduzco lf murrietaguevara fr, hernandez-tapia g, rodriguezotal lm (2010), desulfurization of low sulfur diesel by adsorption using activated carbon: adsorption isotherms. industrial & engineering chemistry research 49: 4372–4376. moaseri e, shahsavand s, bazubandi b (2014), microwave-assisted oxidative desulfurization of 51 microwave-assisted adsorptive desulfurization of model diesel fuel using synthesized microporous rare …. sour natural gas condensate via combination of sulfuric and nitric acids. energy & fuels 28:825−831. pal p, das jk, das n, bandyopadhyay s (2013), synthesis of nap zeolite at room temperature and short crystallization time by sonochemical method. ultrasonics sonochemistry 20:314–321. salem ab, hamid hs (1997), removal of sulfur compounds from naphtha solutions using solid adsorbents. chemical engineering & technology 20:342–347. seredych m, bandosz tj (2010), adsorption of dibenzothiophenes on nanoporous carbons: identification of specific adsorption sites governing capacity and selectivity. energy & fuels 24:3352–3360. shakirullah s, ahmad i, ishaq m, ahmad w (2009), study on the role of metal oxides in desulfurization of some petroleum fractions. journal of the chinese chemical society 56:107– 114. sisani e, cinti g, discepoli g, penchini d, desideri u, marmottini f (2014), adsorptive removal of h2s in biogas conditions for high temperature fuel cell systems. international journal of hydrogen energy, in press. song h, wan x, dai m, zhang j, li f, song h (2013), deep desulfurization of model gasoline by selective adsorption over cu–ce bimetal ionexchanged y zeolite. fuel processing technology 116:52–62. subhan f, liu bs, zhang y, li xg (2012), high desulfurization characteristic of lanthanum loaded mesoporous mcm-41 sorbents for diesel fuel. fuel processing technology 97:71–78. treacy mmj, and higgins jb (2001), collection of simulated xrd powders for zeolites, published on behalf of the synthesis commission of the international zeolite association. amsterdam: elsevier. velu s, ma x, song c (2003), selective adsorption for removing sulfur from jet fuel over eolitebased adsorbents. industrial and engineering chemical research 42:5293–5304. wei z, niu h, ji y (2009), simultaneous removal of so2 and nox by microwave with potassium permanganate over zeolite. fuel processing technology 90:324–329. weitkamp j, schwark m, ernst s (1991), removal of thiophene impurities from benzene by selective adsorption in zeolite zsm-5. journal of chemical society, chemical communications, 1133–1134. zannikos f, lois e, stournas s (1995), desulfurization of petroleum fractions by oxidation and solvent extraction. fuel processing technology 42:35–45. zhang xl, hayward do, lee cd, mingos dmp (2001), microwave-assisted catalytic reduction of sulfur dioxide with methane over mos2 catalysts. applied catalysis b: environmental 33:137–148. 52 c:\users\raoh\desktop\paper 6.xps the journal of engineering research (tjer) vol. 12,no. 1, (2015) 69-80 electromagnetic field detector circuit for lowfrequency energy applications m.m. bait-suwailam department of electrical and computer engineering, sultan qaboos university, p.o. box 33, pc 123, muscat, oman received 7 september 2014; accepted 11 january 2015 abstract: this study details an electromagnetic (em) field radiation detector system that was developed for near-field low-frequency energy applications. the prototype constitutes the use of a dual-band monopole antenna system as a probe along with a detecting circuit. furthermore, the prototype was equipped with a qualitative em radiation strength display unit at its output stage. for proof of concept, the detecting probe was implemented on a printed-circuit board. both numerical simulations were based on pspice software (cadence design systems, inc., san jose, california, usa) and measurements are presented and discussed. the em field detector aimed to sense any potential sources of em radiation from mobile phone units as well as wifi access points, simultaneously, which is accomplished with the use of the dual-band antenna system. such a sensitive detector has useful application as a stand-alone monitoring probe for troubleshooting as well as to identify sources of em radiation interference threats for industrial high-speed electronic devices. additionally, such a sensor is a potentially useful tool for site testing and scanning for optimal locations of base station masks for telecommunication service providers. other prototypes are also presented to illustrate the usefulness of such detectors in some of the aforementioned applications. keywords: antenna, electromagnetic radiation, detector circuit, probe, sensor. pspice . corresponding author’s e-mail: msuwailem@squ.edu.om m.m. bait-suwailam 1. introduction in the atmospheric space around us, there are various types of electromagnetic radiation which are present due to the operation of high-speed electronic devices, microwave systems / subsystems, and telecommunication devices, as well as the use of radio frequency (rf). with the rapid growth in the wireless, industrial, and medical sectors nowadays, there has been an increase in the presence and use of mobile communication devices. additionally, the excessive installation of high-powered base station towers, ubiquitous electronic devices and sensors, and the constant use of broadcast radio and tv transmitters, and satellite systems have led to public concern. this concern stems from the possibility of hazardous and adverse health effects associated with exposure to rf radiation originating from such systems and high-speed devices (advisory group on non-ionizing radiation, 1992; advisory group on non-ionizing radiation, 1994; hanna et al. 2009; international commission on non-ionizing radiation protection, 1998; ieee standards association, 1994; mann et al. 2001; world health organization, 2011). hence, there is a need for monitoring systems that can detect and predict sources of em radiation along with their energy strength. it is well known from classical physics theory that electromagnetic (em) waves can propagate and travel at the speed of light in free space or slower than the speed of light in any other respective media. such waves behave as plane waves when far away from the source of radiation, in which the electric field, magnetic field, and direction of propagation form a mutual orthogonal triplet. in fact, as frequency of the propagating wave is increased, em waves have strong energy levels which may be hazardous. in general, the human eye can detect light in free space at wavelengths between 400–700 nanometers (ulaby et al. 2014). hence, this spectrum range is known as visible light. however, the objects around us can also emit other types of em radiation that cannot be seen by the human eye, such as rf waves. such em radiation exists in the universe surrounding us. furthermore, the energy carried by em waves emanating from mobile devices can easily interfere with other devices that operate within close proximity to the operating frequency of such mobile devices, for instance, around the gsm-900 megahertz (mhz) frequency band. another example of an originator of emradiated emissions is conductive coupling paths. for example, connectors, like wires and cables, radiate at a high frequency; thus, their em wavelengths cannot be neglected for their contribution in increasing frequency so that it is close to the ghz range. such is the case of today’s connectors and accessories for high-speed electronic devices, such as power cables (archambeault et al. 2001; bait-suwailam et al. 2014; brench and brench 1989; buchanan et al. 1994; park et al. 2013). in sum, em waves are everywhere and cannot be ignored. their strength and existence will dramatically increase in the coming years with the rapid growth of technology and employed services. an example of this is, 4g service, which is available at present, and the upcoming 5g service, which will no doubt result in an increase in the number of base station towers and an additional increase in their power strength in order to provide broadband mobile services to customers (datang telecom technology & industry group, 2013). in order to provide satisfactory safety margins for the public, several international organizations worldwide have set the health and safety guidelines which are currently in place. these guidelines’ ability to limit public exposure to radio waves from base stations and mobile phones are well documented, having been set by the international commission on non-ionizing radiation protection (icnirp) (1998). the amount of em radio wave energy absorbed by the human body is quantified as specific absorption rate (sar). in practice, mobile devices and high-speed electronic devices, for instance touch pads, are assessed for potentially excessive of em radiation by a series of sar measurements prior to certification in order to meet the federal communications commission (fcc) regulations (icnirp, 1998). the spatial peak sar limit set by the fcc in the us is 1.6 w/kg over 1 gram of actual tissue, while it is 2 w/kg in europe (bernardi et al. 1998; rowley and waterhouse 1999; seabury 2005). although knowledge of sar levels and appliances’ em output is important, the public generally lacks knowledge of these limits, and compatibility testing equipment is not easily found. therefore, it is believed that alternative, easy-to-find-and-use techniques would attract 70 electromagnetic field detector circuit for low-frequency energy applications 71 public attention. such alternatives rely on developing highly sensitive em-detecting units that could be of interest in identifying potential sources of electromagnetic radiation interference emi). hence, an electrician, or even a homeowner, might employ this device to locate buried electrical lines or even defects within high-speed electronic devices. in (lin, james 2009), em radiation detectors using electronic circuitry were briefly developed. herein, further results covering applications to detect malfunctions in power meters as well as high-speed electronic circuits are presented. based on qualitative results, the probes are found to be highly sensitive to any em radiation within lowfrequency range. it is believed that the stand-alone detector circuit will also be useful as a measurement tool to predict em radiated emissions from high-speed chips/circuits, especially for certification of electronic devices in industries as well as in medical sector. in addition to the aforementioned applications, such detector is of no doubt beneficial to predict any unauthorized use of mobile phones in examination halls in academia as well as in security rooms. this paper is organized as follows. section 2 covers details of the developed detector circuit that targets three applications, namely: power line meters, extremely low-frequency energy targeting printed circuit boards, and lowfrequency regime covering mobile phone devices and wifi access points. both simulation and experimental results are then presented and discussed in section 3. section 4 concludes with a brief summary of the findings. 2. the developed detector the detecting prototype that is developed here constitutes the use of an antenna system, followed with a detecting circuitry and ends up with an indicator (led) as well as a qualitative field strength display unit. figure 1 shows block diagram of the aforementioned stages of the developed detector. at an early stage through the design and numerical simulations, a pre-defined alternating current (ac) signal was used instead of the antenna system to mimic an incoming electromagnetic wave. note that in all the simulated detectors, an ac voltage source used as an input to mimic the detected electromagnetic wave captured by the antenna. in this research work, a quarter-wavelength ( /4) thin metallic monopole antenna is considered in all prototypes. a view of such antenna is shown in fig. 2, where the chassis is used as a ground plane for the antenna. the resonance frequency of the monopole antenna is determined by its length d, which is given by: /4, (1) where 0 is the free-space wavelength of the electromagnetic wave at the operating frequency. note that in practice, the resonance frequency will in fact need to be slightly less than a quarterwavelength. an added advantage of the use of monopole antennas is their omnidirectional radiation pattern in the horizontal plane (h-plane), in which the antenna will be able to detect incoming electromagnetic waves arriving from any direction (balanis, c. 2005). the radiation pattern of monopole antennas has a typical donut shape in the vertical plane (e-plane). in order to identify the resonance of the monopole antenna, the reflection coefficient (or standing wave ratio swr) is studied. figure 3 shows the voltage swr for one of the vertical monopole antenna that resonates within the 900 mhz band, where the electrical length of the antenna is nearly 8.3 cm. it is essential to highlight here that several design constraints and challenges were encountered, including mostly the employed active devices, for instance operational amplifier and transistors. such components were found to be band limited and sensitive to certain operating frequency range. as such, three prototypes were developed targeting several applications, with emphasis on the radiation out from mobile devices and wifi access points. the detecting prototypes are discussed and presented next. 3. numerical results and experimental validation 3.1 detectors for power transmission line meters figure 4 depicts the developed detector circuit. for simulation purposes, a pre-defined ac signal was assumed in order to mimic an incoming radiated em wave in space. m.m. bait-suwailam figure 1. block diagram showing the stages of the em field detector circuit. figure 2. a quarter-wavelength monopole antenna mounted on top of a metallic chassis. figure 3. voltage standing wave ratio for the monopole antenna resonating at the 900 mhz band. 72 electromagnetic field detector circuit for low-frequency energy applications figure 4. the detector circuit as modeled in pspice to sense any possible em field radiated emissions from 60-hz power meters. the circuit was modeled and simulated using pspice circuit simulator. an operational amplifier (ua741) was used and found suitable for such extremely low-frequency applications. the detecting circuit needs only a dc supply of 12 volts. the output is then captured and recorded. both the input and output signals out from the simulation model are shown in fig. 5, while experimental results are displayed in fig. 6. the output signal captured by the detector circuit has a frequency of 60 hz, which was the original input signal to the circuit. note that the output waveform is captured at the output waveform is captured at the output terminals of the general-purpose operational amplifier. in fact, some ripples were noticed and appeared at the output terminal of the detector circuit as shown in fig. 6. such ripples are attributed to the power line signal that was picked up by the detector circuit. nonetheless, the 60-hz input signal was detected at the output of the detector circuit, despite the small coupling that was observed due to the em wave radiated emissions from the power line cords in laboratory. this observation of electromagnetic interference issues has been clarified in section 1. such coupling effect is attributed mainly to the magnetic field of radiated emissions from the power line cords which was inherently induced on the detector circuit. 3.2. detectors for extremely low-frequency electronic circuits in this part, we examine the ability of the detector circuit to capture electromagnetic waves at the extremely low-frequency range which is of interest in electronic circuits troubleshooting. figure 7 depicts another detecting circuit for low frequency applications, targeting high electromagnetic exposure from extremely lowfrequency electronic devices and systems. in this circuit, a 9-v dc supply was found sufficient. the experimental setup along is illustrated in fig. 8. for demonstration purposes, the circuit was implemented in general-purpose breadboard. the captured output signal is shown in fig. 9, when an input waveform at a frequency of 1.0 mhz was considered. it is worth noting here that other signals of higher frequency up to 3.5 mhz were easily captured by the detecting circuit, beyond which noise started to appear at output waveform. this is attributed to the limited capabilities of the general-purpose transistor and operational amplifier. 3.3 detectors for mobile phones and wifi access points in this part, the detector circuitry, as shown in fig. 10, is developed to predict potential electromagnetic radiation from mobile devices and wifi access points simultaneously. in order to accomplish this, a dual-band monopole antenna system is designed to capture the energy of such incoming waves of the dual-resonance system. figure 11 shows the simulation results when considering an ac signal with a frequency of 900 mhz. both input and output signals are shown. it can be seen that the circuit was able to identify the 900 mhz signal, as evident from the period of the output waveform. figure 12 shows the implemented circuit (see fig. 10) in breadboard. in the implemented detector circuit, a single high-profile monopole antenna was used, where it was designed to resonate and capture the energy of 900 mhz em 73 m.m. bait-suwailam (a) (b) figure 5. the detector circuit as modeled in spice to sense any possible em field radiated emissions from 60-hz power meters. figure 6. experimental setup for the 60-hz detector circuit (a) input waveform; and (b) captured signal at output terminals of the detector circuit. wave radiated out of mobile phones. both led as well as a qualitative em field strength display unit will turn on once a call/text message is either transmitted or received. interestingly, the em field strength display unit was highly sensitive to incoming signals from either mobile phone or wifi router, which is indicated by the levels in the display unit (see fig. 12). in order to avoid any sources of external electromagnetic interference, the same circuitry has been built in the lab using printed circuit board (pcb) technology. 74 electromagnetic field detector circuit for low-frequency energy applications figure 7. detecting circuit that detects electromagnetic exposure from low-frequency electronic devices. figure 8. experimental setup of detecting circuit for low-frequency applications. figure 9. measured output waveform of the low-frequency detecting circuit at 1.0 mhz, visualized by an oscilloscope. 75 m.m. bait-suwailam figure 10. detecting circuit that detects em radiation from mobile phones and wifi access points, simultaneously through the use of a dual-band monopole antenna. figure 11. the simulation results for the detecting circuit modeled in pspice. note that input waveform is shown on top, whereas the output is on bottom. figure 12. the built-in detecting circuit in breadboard with a single monopole antenna. notice the blinking of the led as well as strength display unit being on when a call was either placed or received. 76 electromagnetic field detector circuit for low-frequency energy applications figure 13 shows details of the detector circuit implemented on pcb, where two dual-antenna elements were mounted at front end of the detecting circuit in order to simultaneously detect em radiation coming from either the 900 mhz gsm band or wireless devices at 2.45 ghz. it is instructive to note here that the selection of the lumped elements for the detector circuit in part (c), for instance, was not straight forward. in fact, the selection was done carefully through series of a trained “trial-and-error” procedure, as some of the components were found to be quite sensitive to the operating frequency of the incoming em wave, including the use of specialized transistors and operational amplifiers to fit the operating frequency regime. for instance, the general purpose operational amplifier was found not suitable for this kind of applications as frequency is relatively high. special amplifiers, lm358 and ad-845, were employed in the detectors circuits of parts (b) and (c) (lm358 datasheet), although other alternative wideband amplifiers could be used. furthermore, a generic 2n2222 npn transistor was used for the circuits in parts (a) and (b), while a special transistor, bc548 (bc548 datasheet), that fits highfrequency applications was adopted in the detector circuit of part (c). several sensitivity analyses were also conducted in laboratory in order to quantify the sensitivity of the antenna elements to em radiation. it was found out that the detector targeting mobile phone devices can detect em radiation up to almost 1.0 meters, which is quite useful as it reasonably detects within both near and far-field regime from the em source of radiation. however, it was observed that the detector system is highly sensitive to em radiation within near-field proximity to the source of radiation as compared to the far-field region from the same source of radiation. the results of this study are summarized and tabulated as presented in table. 1. for this study, the energy of em waves radiated out from mobile devices were captured and detected when either transmitting or receiving calls. as expected, em field strength was decreased as we get away from the source of radiation. figure 14 shows the detected mobile phone signal by the developed dual-band probe. the detected signal strength was around 915 mhz. in this particular case, only mobile phone was in vicinity of the detecting probe where no wifi signals were in close proximity. hence, a peak that falls around the 900 mhz band is visible in the spectrum analyzer. based on the accomplished simulated and measured results, there are issues that need to be addressed here. due to the sensitivity of the detector circuits to operating frequency of the incoming electromagnetic waves, several prototypes were developed to fit three frequency energy, covering real-world applications, ranging from as low as power line meters, to mobile phone devices and several bands of telecommunication services. in future, it will be essential to come up with a wideband detecting circuitry that can fit a wide spectrum of energy applications. in addition to that, incorporating an em field strength meter is essential in order to provide a quantitative measure of em radiated emissions strength from potential sources of interference. furthermore, the use of high-profile antennas may not fit other applications, especially for biomedical applications, eg. implanted devices. therefore, the use of low-profile miniaturized antennas yet being wideband are highly desirable. the aforementioned issues will be the subject of future work. table 1. measured alternating current voltage and current at output stage of the detecting system at various distances from mobile phone unit. distance (cm) voltage (v) current (ma) 0 2.3 5 14 1.5 5 30 1.2 2 40 1.1 2 50 0.8 2 60 0.6 1 80 0.5 1 100 0.3 1 >100 0 0 77 m.m. bait-suwailam figure 13. the detector system with an integrated dual-band monopole antenna, as shown in both images. details of two applications of detecting em waves generated from both mobile phone (top image) and wifi access points (bottom snapshot) are shown. figure 14. the detected em radiation out from a mobile phone using the spectrum analyzer. 78 electromagnetic field detector circuit for low-frequency energy applications 4. conclusions in this research work, low-frequency electromagnetic field detector system was developed, numerically simulated and experimentally validated. the proposed detector was integrated with a dual-band monopole antenna system that targets detection of electromagnetic radiated emissions from both mobile phone units (gsm-900 mhz band) and wireless devices (wifi access points) operating at the 2.45 ghz band. the prototype was simulated, and validated against experimental data for the aforementioned applications. the measured results agree with expected numerical results. other prototypes were presented that can easily predict and sense possible malfunctions in power line meters as well as em interference threats in highspeed printed circuit boards. although the incorporated antenna system is narrowband, the dual-band antenna elements were quite sensitive to capture the electromagnetic radiated emissions from low-energy devices as evident from the experimental prototype (see fig. 12). the use of wideband antenna system is useful for a standalone monitoring probe system that will find useful applications, for instance for mobile communication service provides as a test drive scanning tool, or as an alarming system for diagnosis and industrial troubleshooting for potential emi threats. acknowledgments the author would like to thank sultan qaboos university for providing the financial resources needed for the project. the author would also like to thank issa al-balushi, mohammed al-alawi, mohammed al sadairi, and salim al-risi for carrying out couple of measurements for this work. references advisory group on non-ionizing radiation (1992), electromagnetic fields and the risk of cancer report. national radiological protection board 3(1):1–138. advisory group on non-ionizing radiation (1994), health effects related to the use of visual display units report. national radiological protection board 5(2):1–75. archambeault b, ramahi o, brench c (2001), emi/emc computational modeling handbook, 2nd ed, boston: kluwer. bait-suwailam m, alavikia b, ramahi o (2014), reduction of electromagnetic radiation from apertures and enclosures using electromagnetic bandgap structures. ieee transactions on components, packaging, and manufacturing technology 4(5):929–937. bait-suwailam m, al-balushi i, al-sadairi m, alalawi m, al-risi s (2014), electromagnetic radiation probe using a detector circuit with a dual-band antenna system. proceedings of the 2nd international conference on electronics design (iced), penang, malaysia. balanis c (2005), antenna theory: analysis and design 3rd ed. hoboken, nj: john wiley, inc. bernardi p, cavagnaro m, pisa s, piuzzi e (1998), sar distribution and temperature increase in an anatomical model of the human eye exposed to the field radiated by the user antenna in a wireless lan. ieee transmission on microwave theory and techniques 46(12):2074–2082. brench c, brench b (1989), effects of cable and peripheral placement on radiated emissions. proceedings of the ieee national symposium on electromagnetic compatibility may 23-25, 351–356. buchanan wj, gupta nk, arnold jm (1994), application of 3d finite-difference time-domain (fdtd) method to predict radiation from a pcb with high speed pulse propagation. proceedings of the ninth international conference on electromagnetic compatibility, university of manchester 287-291. datang telecom technology & industry group (2013), evolution, convergence, and innovation—white paper on 5g technology. retrieved from: www.datanggroup.cn/ upload/accessory/201312/201312919445526537 2.pdf. fairchild (2014), bc546/bc547/bc548/ bc549/ bc550 npn epitaxial silicon transistor. retrieved from: hwww.fairchildsemi.com/ ds/ bc/bc547.pdf. hanna s, motai y, varhue w, titcomb s (2009), very low-frequency electromagnetic field detector with data acquisition. ieee transmission on instrumentation and measurement 58(1):129–140. international commission on non-ionizing radiation protection (1998), guidelines for 79 m.m. bait-suwailam limiting exposure to time-varying electric, magnetic and electromagnetic fields (up to 300 ghz). health physics journal 74(4):494–522. ieee standard std. (1994), procedures for measurement of power frequency electric and magnetic fields from ac power lines. ieee std 644. doi: 10.1109/ieeestd. 1995. 122621. lin j (2009), specific absorption rates (sar) induced in head tissues by microwave radiation from cell phones. ieee antennas and propagation magazine 42(5):138–140. mann j, lee r, aragon t, bhatia r (2001), radio frequency radiation from broadcast transmission towers and cancer: a review of epidemiology studies. retrieved from: www.sfdph.org/dph/ files/reports/ studies data/ radiofreqradrpt032001.pdf. park h, park h, lee h (2013), a simple method of estimating the radiated emission from a cable attached to a mobile device. ieee transmission on electromagnetic compatibility 55(2):257–264. 80 rowley j, waterhouse r (1999), performance of shorted microstrip patch antennas for mobile communications handsets at 1800 mhz. ieee transmission on antennas and propagation 47(5):815–822. seabury d (2005), an update on sar standards and the basic requirements for sar assessment. retrieved from: www.ets-lindgren.com/pdf/ sar_lo.pdf. texas instruments (2014), lmx58-n low-power, dual-operational amplifiers. retrieved from: http://www.ti.com/lit/ds/symlink/lm158n.pdf. ulaby ft, michielssen e, ravaioli u (2014), fundamentals of applied electromagnetics 6th ed. london: pearson education limited 30–31. world health organization (2011), electromagnetic fields and public health: mobile phones. fact sheet no. 193. retrieved from: http://www.who.int/mediacentre/factsheets/f s193/en/. the journal of engineering research (tjer) vol. 15, no. 1 (2018) 42-60 doi: 10.24200/tjer.vol15iss1pp42-60 simulation-optimization model for design of water distribution system using ant based algorithms r. moeini*, a, and s.a.m. moulaeib adepartment of civil engineering, faculty of civil engineering and transportation, university of isfahan, 81746-73441, isfahan, iran. bbaft imam hussein faculty of engineering, professional technical university, 7851915848, azadegan street, baft, kerman, iran. received 28 january 2015; accepted 7 february 2017 abstract: in this paper, the ant colony optimization algorithm (acoa) is applied to solve water distribution system design optimization problem proposing two different methods. considering pipe diameters as decision variables of the problem, ant system and max-min ant system, referred to acoa1 and acoa2 respectively, are applied to determine pipe diameters. in proposed methods, the ant-based models are interfaced with epanet as simulator for the hydraulic analysis. three benchmark test examples are solved with proposed methods and the results are presented and compared with those obtained with other existing methods. the results show the capability of the proposed methods to optimally solve the design optimization problem in which best results are obtained with acoa2 in comparison with other available results. furthermore, the results show the superiority of the proposed acoa2 over than the acoa1 in which the trade-off between the two contradictory search characteristic of exploration and exploitation is managed better by using maxmin ant system. keywords: water distribution system; ant colony optimization algorithm; epanet; optimal design; pipe diameter. تقنية االحتماالت من خالل منوذج احملاكاة األمثل لتصميم نظام توزيع املياه باستخدام حساب خوارزمية حتسني مستعمرة النمل(البيانية )الرسومات ب و سيد أمني موسوي مواليي *أ رامتني مويين : قام هذا البحث بتطبيق حساب تقنية االحتماالت من خالل الرسومات البيانية )أكوا( حلل مشكلة حتسني تصميم امللخص االعتبار أقطار األنابيب كمعامل يف قرار حل املشكلة. قام نظم توزيع املياه واقرتاح طريقتني خمتلفتني لذلك . و مت أخذ يف على التوالي. و يف هذه 2و اكوا 1صطلحني : اكوا ملالبحث بتطبيق نظام النمل و ماكس مني النمل. ، املشار إلىه با ج حملاكاة التحليل الطرق أملقرتحة مت توصيل النماذج املستندة إىل نظام النمل باستخدام نظام برجميات إيبانيت كنموذ اهليدروليكي. وهنا يعمل البحث على حل ثالث مناذج اختبار معيارية بالطرق املقرتحة ويقوم بعرض النتائج ومقارنتها مع تلك اليت مت احلصول عليها باستخدام األساليب القائمة األخرى. وقد أظهرت النتائج قدرة األساليب املقرتحة على حل مشكلة باملقارنة مع غريها من النتائج املتاحة. كما اظهرت 2ل. حيث مت احلصول على أفضل النتائج من نظام اكواالتصميم األمث الذي مت فيه املفاضلة بني طريقتني متناقضتني من حيث طرق 1النتائج تفوق نظام اكوا أملقرتح على نظام اكوا م ماكس مني النمل.االستكشاف واالستغالل ملعرفة ايهما افضل باستخدام نظا قطر ،التصميم األمثل،نظام برجميات إيبانيت ،خوارزمية حتسني مستعمرة النمل،: نظام توزيع املياه الكلمات املفتاحية األنابيب. * corresponding author’s e-mail: r.moeini@eng.ui.ac.ir r. moeini and s.a.m. moulaei 43 1. introduction a water distribution system (wds) is a manmade system that is used to convey water from the source to the user. the main aim of every wds is to deliver water to the user when it is necessary, in the correct quantity and in accordance with relevant quality standards. wdss are therefore considered to be essential infrastructure in every modern community. as wdss continue to get aged and cities continue to grow, the design of new wdss and the rehabilitation or upgrading of existing wdss will continue to be a central problem. construction of wds is very expensive due to cost of wds components. a quite small change in the wds components, therefore, leads to an enormous saving. the general wds design optimization problem, therefore, involves minimizing the system cost subject to hydraulic constraints such as meeting minimum allowable pressure and/or maximum allowable velocity constraints under design demand levels. a significant amount of research has focused on the optimal design, upgrade or rehabilitation of wdss. in general, the available methods can be classified as 1) linear programming (lp), 2) nonlinear programming (nlp), 3) dynamic programming (dp) and 4) evolutionary algorithm (ea) such as genetic algorithm (ga), simulated annealing (sa), shuffled frog leaping algorithm (sfla), tabu search (ts), ant colony optimization algorithm (acoa), harmony search (hs), particle swarm optimization (pso), cross entropy (ce), honey-bee mating optimization (hbmo) and differential evolution (de) in which they were usually used for optimal design of wds. haghighi et al. (2011) listed different methods applied to design of wds such as enumeration (gessler 1985; lp (bai et al. 2007; nlp xu and goulter 1999; dp schaake and lai 1969; ga jian and yanbing 2010; bi et al. 2015; sa costa et al. 2000; sfla baoyu et al. 2011; ts lippai et al. 1999; acoa afshar 2007; hs yang et al. 2012); pso babu and vijayalakshmi 2013; hbmo mohan and babu 2010; ce shibua and janga reddya 2012; de dong et al. 2012 and other ea and hybrid methods zhou et al. 2016; sheikholeslami et al. 2016) in which some of them are presented in this paper. each of these optimization methods has its own limitations. for example, calculating derivatives or requirement of an initial policy to start off the solution process are some important limitations of the conventional mathematical optimization methods (mays and tung 1992). in addition, dp is known to suffer from the “curse of dimensionally” when large scale problems are attempted to solve. nowadays, eas, therefore, have originally been proposed to overcome the limitations of the conventional optimization methods. however, some limitations such as the element of randomness used in generation of initial solutions and generating alternative solutions inherent in most of eas may reduce their efficiency. in many large scale cases, generally, eas require additional amount of computer memory which is based on the fact that they need more objective function evaluations than conventional optimization methods. in addition, some eas such as acoa have discrete nature and some others have continuous nature. therefore, acoa is suitable algorithm for solving discrete problem such as wds design optimization problem considered here. acoa is one of the eas that has specific characteristic for solving np-hard combinatorial discrete optimization problems with reasonable computational time cost. for wds design optimization problem, the pipe diameters are discrete variables in design process and, therefore continuing optimization methods cannot determine them properly. the specific characteristic of acoa is that in acoa each artificial ant incrementally builds a solution by adding opportunely defined solution components to a partial solution under construction. this specific unique feature, namely incremental solution building capability, is very useful for solving optimization problems of sequential nature. this algorithm has been inspired in the real ant colony behaviour for finding the shortest path from food source to the nest. starting with ant system (as) (colorni et al. 1991), a number of algorithmic approaches based on this characteristic of real ant was developed and applied with considerable success to a variety of combinatorial optimization problems from academic as well as from real world applications. many other algorithms have been proposed to improve the performance of as, such as ant colony system (acs), elitist ant system (aselite), elitist-rank ant system (asrank) and max-min ant system (mmas). advancements have been made on the as to improve the operation of the decision policy and the manner in which the policy incorporates new information to help explore simulation-optimization model for design of water distribution system using ant based algorithms 44 the search space of the problem. these developments have primarily been aimed at managing the trade-off between the two contradictory search characteristic of exploration and exploitation. ali et al. 2009; ostfeld 2011; stutzle et al. 2011; and chandar mohan and baskaran 2012) reviewed an amount of research work in the field of using acoa to solve different optimization problems in the last 20 years. a review of the literature reveals that studies on the application of acoa to different optimization problems has not been satisfactorily fruitful and is worth further investigation. numerous papers on the acoa application for optimization of water engineering problems such as sewer network design (moeini and afshar 2013a), wds design (zeccchin et al. 2007) and reservoir operation (moeini and afshar, 2013b) have been published in recent years. in this paper, one of the eas with discrete nature means acoa is used to effectively solve one of the most important discrete problem means optimal design of wds by proposing two different methods. in the proposed methods, the acoa is used to determine pipe diameters as the discrete decision variables of the problem. as and mmas referred to acoa1 and acoa2 respectively, are applied to determine pipe diameters. mmas is applied here to overcome the premature convergence problem to suboptimal solution which is well trained in as by managing the trade-off between the two inconsistent search characteristic of exploration and exploitation. by determining the pipe diameters of wds, a hydraulic analysis is required. for the hydraulic analysis, in this paper, the acoa-based models are interfaced with epanet as simulator. three benchmark test examples are solved here with proposed methods and the results are presented and compared with those obtained with other existing methods. it is worth noting that many different eas have been used to solve this problem. however, based on the discrete nature of design of wds, here, acoa is used to solve this problem in order to overcome the limitations of pervious works. in some researches, the hydraulic analysis is based on the simplified assumptions such as considering inaccurate value for hydraulic value coefficients equations. however, epanet simulator is used here for the hydraulic analysis. in addition, in some researches the continues nature of eas are used to solve discrete nature problem of wds design and in some other researches the equations of simulation model are solved using iterative numerical or traditional methods and therefore some of the wds components such as pump cannot be easily modelled. here, this fact is regarded considering third test example in which the wds is fed by pump. the efficiency of these innovations is fully presented and highlighted in section “model application and results” by comparison of the results obtained with proposed methods with other available results. 2. ant colony optimization algorithm (acoa). the acoa is based on the natural foraging behaviour of a colony of real ants to find the shortest path between food source and their nest. when ants are travelling, they deposit a substance called pheromone which is more attractive for other ants to follow them. ants can smell pheromone and they incline to choose probability paths marked by strong pheromone concentrations when choosing their path. it has been shown experimentally that this pheromone trail following behaviour can give rise to the emergence of the shortest paths once employed by a colony of ants. based on the real ant behaviour, the acoa was developed (colorni et al., 1991). the original and simplest version of acoa is as. later, many other algorithms such as mmas have been developed to overcome the limitations and improve the performance of the as. here, as and mmas are used to solve wds design optimization problem proposing two different methods. definition of proper graph is the first step of process for solving an arbitrary optimization problem using acoa. a graph g=(dp,op,co), therefore, can be defined, where dp=  idp,....,2dp,1dp is the set of decision points, op=  ijop is the set of options j ( j=1, 2,…,j) at each of the decision point i (i=1,2,…,i) and finally co=  ijco is the set of costs associated with options op=  ij op . (moeini and afshar, 2009). 2.1. ant system (as) the original and simplest version of acoa is as. the basic steps of solving optimization problem using as can be defined as follows (colorni et al. 1991): r. moeini and s.a.m. moulaei 45 1. at the start of the process, some proper values are initialized to the amount of pheromone trail on all options ijop . 2. a colony with the size of m is selected and at one time, each of them is placed randomly on the decision points of the problem. 3. by using a transition rule for selection proper option at each decision point i, (eq. 1), each arbitrary ant, m, starts its movement from one decision point to the next and the solutions are incrementally constructed. this procedure is repeated until all decision points of the problem are covered.            j 1j ]ij[)]t(ij[ ]ij[)]t(ij[ )t,m(ijp (1) in eqn. 1, )t,m(ijp is the probability that the ant m selects option j of the ith decision point, ijop , at iteration t; )t(ij is the concentration of pheromone on option ij op at iteration t; ij is the heuristic value of option ij op , and  and  are pheromone and heuristic sensitivity parameters, respectively. 4. the cost of the solution, m)(f  , is calculated when ant m creates a complete solution, m )( . 5. when steps 3 and 4 are repeated for all ants, m, at each iteration t, the pheromone is updated using the general form of eq. 2 for the pheromone updating rule. ij)t(ij)1t(ij  (2) in eqn. 2, )1t(ij  is the amount of pheromone trail on option ijop at iteration t+1; )t(ij is the concentration of pheromone on option ijop at iteration t; )10(  is the pheromone evaporation coefficient and ij  represents the change in pheromone concentration associated with option ijop which is defined as eqn. 3:                otherwise 0 ant best by the chosen is joption if m )(f rm 1m m 1m m ijij (3) in eqn. 3, m )(f  is the cost of the solution produced by the ant m and r is pheromone reward factor. 6. steps 2 to 5 are continued until convergence criterion is met. 2.2. max-min ant system (mmas) the most important issue that can be experienced by all acoas is premature convergence to suboptimal solutions. to overcome this problem, the mmas was developed by stutzle and hoos (2000). the basis of mmas is the provision of dynamically evolving bounds on the pheromone trail intensities ( )t(ij ) which are defined as follows (eqns. 4 and 5): best )(f r 1 1 )t(max   (4) i/1 )bestp.(avgj ) i/1 )bestp(1( )t(max)t(min   (5) in eqs. 4 and 5, )t(min and )t(max represent the lower and upper limit on the pheromone trail strength at iteration t, respectively; best )(f  is the cost of the solution produced by the best ant at each iteration, bestp is the probability that the best solution is constructed again; javg represents the average number of options available at decision points of the problem; and i is the number of decision points. in mmas, only the cost of iteration-best ant is used for calculating the change of pheromone simulation-optimization model for design of water distribution system using ant based algorithms 48            otherwise 0 ant best by the chosen is joption if best )(f r best ijij (6) ij which is defined as eqn. 6. in eqn. 6, best)(f  is the cost of the solution produced by the best ant. it is worth noting that application of mmas to some benchmark optimization problems has shown that mmas overcomes the stagnation problem, hence improves the performance of the acoas (afshar and moeini 2008). 3. water distribution system design optimization problem wdss consist of interconnected elements such as pipes, tanks, pumps and other components. such systems are used to transfer treated water from one or more sources to users spread over a wide area. designing an effective wds with minimum cost is a quite complex task which can be achieved by formulating and solving it as an optimization problem. optimal wds design requires optimal determination of wds components such as pipe diameters, tank and pumping station positions and heights leading to a complex optimization problem with a large number of hydraulic constraints. having a suitable and cost-effective wds is normally interpreted as finding the solution for this problem that minimizes total cost without violating the constraints. the standard description form of a wds design optimization problem with a prespecified layout can be presented as follows. the objective function of this problem can be formulated as eqn. 7 in which it is minimizing the total construction cost of wds.    n 1l lllctcminimize (7) in eqn. 7, tc is the total cost of the pipes in the wds; l l is length of the lth pipe; l c is per unit cost of the lth pipe and n is the total number of existing pipes. the cost function is to be minimized under the following constraints. continuity equation: a continuity constraint should be satisfied for each node as follows:      )k(inl )k(outl k,.......,1kkqlqlq (8) in eqn. 8, qk is the required demand at consumption node k; l q is the flow rate in pipe l; )k(in is the list of the pipes incoming to node k; )(kout is the list of the pipes out coming from node k; and k is the total number of exiting nodes in the wds. energy equation: the energy constraint for each loop in the network of wds can be written as follows: p,......,1p0 pl lj   (9) in eqn. 9, lj is the head loss in the lth pipe and p is the total number of loops in the wds. several equations can be used to evaluate the hydraulic parameters of wds such as head losses in the pipes. here, hazen-williams equation is used to find hydraulic parameters. therefore the head losses can be calculated as follows: n,......,1l l d lch lq lllj             (10) in eqn. 10, lj is the head loss in the lth pipe; ld is the diameter of pipe l; lch is the hazenwilliams coefficient of pipe l; l q is the flow rate in pipe l and  ,, are constant coefficients of hazen-williams equation. nodal pressure head: the maximum and minimum values should be defined for the nodal pressure head of wds as eq. 11 in which the nodal pressure head of wds should be within this range. k,......,1kmaxhkhminh  (11) in eqn. 11, kh is the pressure head at node k; maxh is the maximum required pressure head; and min h is the minimum required pressure head. flow velocity: the maximum and minimum values should be defined for flow velocity of wds pipes as eq. 12 in which the flow velocity of wds pipes should be within this range. 46 r. moeini and s.a.m. moulaei 47 n,......,1lmaxvlvminv  (12) in eqn. 12, lv is the flow velocity in the lth pipe; maxv is the maximum flow velocity; and minv is the minimum flow velocity. commercial pipe diameters: a set of commercially available wds pipe diameters should be defined for wds pipe diameters as eqn. 13 in which the wds pipe diameters should be selected from it. n,.....,1lld  d (13) in eqn. 13, ld is the diameter of pipe l and d denotes the set of commercially available pipe diameters. it should be noted that the penalty method is often used to effectively guide solutions of optimization problems from one infeasible solution area to a feasible solution one. here, the penalty method is applied for formulation of the wds as an unconstrained optimization problem. in the used method, therefore, nodal pressure head and flow velocity constraints are added in the original objective function. therefore, a new problem can be defined by the minimization of the penalized objective function of eq. 14 subject to the constraints defined in eqns. (8) to (13): )hcsvvcsv(p n 1i lllcpc    2 n 1l 1 maxv lv 2 n 1l minv lv1vcsv                      2 k 1k 1 maxh kh 2 k 1k minh kh1hcsv                      (14) in eqn. 14, pc is the penalized total cost function of wds; p is the penalty parameter with a large enough value to ensure that any infeasible solution will have a higher total cost than any feasible solution; vcsv is a measure of the flow velocity constraint violation of the trial solution and hcsv is a measure of the nodal pressure head constraint violation of the trial solution. it should be noted that in calculating the csv, the summation ranges over those nodes and pipes at which a violation of constraints (11) and (12) occurs, that is the terms in parentheses are positive. other hydraulic constraints are satisfied among using simulation program (epanet). in other words, epanet satisfies the continuity and energy conservation equations while calculating the flow rate in each pipe and the pressure head at each node. furthermore, here, the flow velocity and the nodal pressure head constraint violations have the same penalty parameter value due to the fact that these constrains are normalized. 4. methodology the wds design optimization problem is solved here by proposing the simulationoptimization approach. in the proposed approach, the epanet software is used as a simulator and acoa is used as an optimization method with the goal of finding pipe diameters. here, the optimization model is coded in matlab and linked with epanet. in the proposed approach, at first, a proper graph should be defined to formulate the wds design as an optimization problem using acoa. this graph consists of a set of nodes referred to as decision points and edges referred to as options available at each decision points. here the decision variables of the problem are the pipe diameters of the wds which are determined in acoa1 and acoa2 methods using as and mmas, respectively. in the proposed methods, the pipes of wds are considered as decision points. the options available at each decision point are, therefore, defined by a finite number of commercially available pipe diameters. figure 1 represents the problem graph defined for acoa1 or acoa2, where vertical lines show the decision points, dashed small lines show the components of commercially available pipe diameters (j=1,…..j) at each decision point i (i=1,…..i), and finally the bold small lines show a trial solution on the graph constructed by an arbitrary ant. by determining the pipe diameters of wds, the acoa-based models are interfaced with epanet for the hydraulic analysis. for both methods, eqn. 14 is used to calculate the total cost of the trial solutions generated. finally, optimal pipe diameters are determined using proposed methods. figures 2 and 3 illustrate the process of solution constructions in acoa1 and acoa2 formulations, respectively. it is worth noting that the different hydraulic softwares have been developed for pipe network analysis such as epanet, water cad, water gems, mikeurban and so on. each of these software’s has its advantages and simulation-optimization model for design of water distribution system using ant based algorithms 48 limitations. due to more advantages of epanet software, this software is used in this paper. in other words, epanet is freeware, user friendly and free source software that everybody can simply use it. due to these noted facts and its accuracy and availability, epanet is so popular in the field of water distribution analysis and therefore it is used here. epanet water distribution system simulator is the most commercial hydraulic model which has been developed by the us environment protection agency. the epanet performs extendedperiod simulation of hydraulic and water quality behaviour within pressurized pipe networks. this well-documented and tested public domain simulator provides a convenient platform for implementing the simulationoptimization approach (rossman 2000). here, the epanet software is linked to matlab with the help of epanet toolkit. epanet toolkit is a dynamic link library (dll) of functions and an extension of epanet simulation package that allows developers to customize epanet software accordingly (rossman 1999). if other pipe network analysis software can be linked to matlab, these software can be easily considered as simulators. 5. model applications and results in order to show the ability and performance of the proposed methods for the optimal design of wds, three benchmark examples are applied here with different scales. therefore, small (test example i), medium (test example ii) and large scale (text example iii) test examples are presen ted and solved here using proposed methods. furthermore, in the third test example, the wds is fed by a pump; however, in the first and second test examples the wds is fed by gravity. the first test example (i) as shown in fig. 4 is the two-loop wds (alperovits and shamir 1977). this wds network has 7 nodes, 8 pipes and two loops in which it is fed by gravity from a reservoir with a 689 (ft) fixed head. all wds network pipes lengths are 3281 (ft) and their hazen-williams coefficient are 130. the minimum nodal pressure head is 98.4 (ft). fourteen commercial candidate pipe diameters for each pipe with corresponding cost values are listed in table 1. details of the network including nodal ground elevations and water demand are given in table 2. the second test example (ii) is the new york city water supply tunnels (nycwst) problem. figure 5 shows the network of this problem in which it consists of 20 nodes connected via 21 pipes (schaake and lai 1969). table 3 represents the pipe lengths, existing diameters, minimum nodal required pressure head and water demands of this network. the original form of this wds has pressure constraint violations at nodes 16, 17, 18, 19, and 20. therefore, this wds is to be modified by adding duplicate pipes in parallel with the existing pipes to meet the minimum nodal pressure head requirements. for each duplicate pipe, there are 15 different diameter sizes and the option of no pipe as shown in table 1 with corresponding cost values. this wds is fed by gravity from a reservoir with a 300 (ft) fixed head. all pipes hazen-williams constant are equal to 100. figure 1. problem graph defined for acoa1 or acoa2. dp1 dp… j=1 2 j-1 dpi dpi j dp… r. moeini and s.a.m. moulaei 49 figure 2. process of solution constructions in acoa1. define the problem graph start from arbitrary node choose an option (pipe diameter) using eq. (1) and move to the next option (pipe) stop start no yes simulation of wds (epanet toolkit) define input data and model parameters (ant number (m), iteration number (t), ) set pheromone=1 for all set set all decision points are covered? compute objective function and constraint violations (eq. 14) no update pheromone of (eqs. 2 & 3) yes yes no t = t + 1 m = m + 1 simulation-optimization model for design of water distribution system using ant based algorithms 54 figure 3. process of solution constructions in acoa2. define the problem graph start from arbitrary node choose an option (pipe diameter) using eq. (1) and move to the next option (pipe) stop start no yes simulation of wds (epanet toolkit) define input data and model parameters (ant number (m), iteration number (t), ) set pheromone=1 for all set set all decision points are covered? compute objective function and constraint violations (eq. 14) no update pheromone of (eqs. 2 & 5) t = t yes yes no t = t + 1 m = m + 1 compute (eqs. 4 & 5) 50 r. moeini and s.a.m. moulaei 53 table 1. candidate pipe diameters and corresponding cost of test examples. test example candidate pipe diameters cost of candidate pipe diameters i {1,2,3,4,6,8,10,12,14,16,18,20,22,2 4} (in) {2,5,8,11,16,23,32,50,60,90,130, 170, 300, 550} (units/meter) ii {36,48,60,72,84,96,108,120,132,14 4,156, 168, 180, 192, 204} (in) {93.5,134.0,176.0,221.0,267.0,316.0,365.0,417.0, 469.0,522.0,577.0,632.0,689.0,746.0,804.0}(dollar/ foot) iii {80,100,125,150,200,250,300,350} (mm) {37890,38933,40563,42554,47624,54125,62109,715 24} (won/meter) table 2. nodal data for test example i. nodal data ground elevation (m) demand (m3/h) node 210 reservoir 1 150 100 2 160 100 3 155 120 4 150 270 5 165 330 6 160 200 7 figure 4. network of two-loop wds (test example i). 1 2 3 4 5 6 7 [1] [2] [3] [4] [5] [6] [7] [8] 51 simulation-optimization model for design of water distribution system using ant based algorithms 54 table 3. pipe and nodal data for test example ii. nodal data pipe data min head (ft) demand (cft/s) node existing diameter (in) length (ft) pipe 300 reservoir 1 180 11600 1 255 92.4 2 180 19800 2 255 92.4 3 180 7300 3 255 88.2 4 180 8300 4 255 88.2 5 180 8600 5 255 88.2 6 180 19100 6 255 88.2 7 132 9600 7 255 88.2 8 132 12500 8 255 170 9 180 9600 9 255 1 10 204 11200 10 255 170 11 204 14500 11 255 117.1 12 204 12200 12 255 117.1 13 204 24100 13 255 92.4 14 204 21100 14 255 92.4 15 204 15500 15 260 170 16 72 26400 16 272.8 57.5 17 72 31200 17 255 117.1 18 60 24000 18 255 117.1 19 60 14400 19 255 170 20 60 38400 20 72 26400 21 51 52 r. moeini and s.a.m. moulaei 53 table 4. pipe and nodal data for test example iii. nodal data pipe data ground elevation (m) demand (cmd) node length (m) pipe 71 reservoir 1 165 1 56.4 153 2 124 2 53.8 70.5 3 118 3 54.9 58.5 4 81 4 56 75 5 134 5 57 67.5 6 135 6 53.9 63 7 202 7 54.5 48 8 135 8 57.9 42 9 170 9 62.1 30 10 113 10 62.8 42 11 335 11 58.6 37.5 12 115 12 59.3 37.5 13 345 13 59.8 63 14 114 14 59.2 445.5 15 103 15 53.6 108 16 261 16 54.8 79.5 17 72 17 55.1 55.5 18 373 18 54.2 118.5 19 98 19 54.5 124.5 20 110 20 62.9 31.5 21 98 21 61.8 799.5 22 246 22 174 23 102 24 92 25 100 26 130 27 90 28 185 29 90 30 simulation-optimization model for design of water distribution system using ant based algorithms 54 table 5. values of acoa1 and acoa2 parameters for all test examples. method iteration ant    bestp acoa1 1000 100 2 0.2 0.95 acoa2 1000 100 2 0.2 0.95 0.2 table 6. maximum, minimum and average solution cost values over 10 runs. test example method cost value scaled standard deviation no. of runs with final feasible solution cpu time for each run (s) minimum maximum average i acoa1 419000 (units) 487000 (units) 433200 (units) 0.0480 10 128 acoa2 419000 (units) 441000 (units) 421900 (units) 0.0163 10 137 ii acoa1 71.27 (m$) 105.98 (m$) 83.99 (m$) 0.1188 10 307 acoa2 38.64 (m$) 53.63 (m$) 45.87 (m$) 0.1180 10 327 iii acoa1 175783163 (won) 176018101 (won) 175856569 (won) 0.0006 10 367 acoa2 175783163 (won) 175783163 (won) 175783163 (won) 0 10 392 figure 5. network of new york wds (test example ii). 1 9 2 3 1 1 5 [1] [4] [5] [2] [3] 4 6 7 8 10 17 [6] [7] [8] [9] [16] 1 2 1 3 1 4 1 5 1 6 2 0 1 8 1 9 [15] [14] [13] [12] [11] [10] [19] [20] [21] [17] [18] r. moeini and s.a.m. moulaei 55 figure 6 shows the third test example (iii) of goyang network which is first constructed in south korea (kim et al. 1994). this wds network consists of 22 nodes, 30 pipes, 9 loops and is fed by a pump (4.52 kw) from a reservoir with a 71-m fixed head. details of the wds including pipe lengths, and nodal ground elevations and water demands are given in table 4. the hazen-williams constant is equal to 100 for all pipes. the minimum nodal pressure head is 15 meter above ground level. eight commercial candidate diameters for each pipe with corresponding cost values are listed in table 1. a set of preliminary runs is first conducted to find the proper values of as and mmas parameters referred to acoa1 and acoa2 respectively, as shown in table 5 for all three test examples. for both methods of all test examples, a colony size of 100 with maximum number of 1000 iterations amounting to maximum number of 100,000 function evaluations is used. table 6 shows the results of 10 runs carried out using different randomly generated initial guess for all test examples along with the scaled standard deviation, the number of final feasible solutions and cpu time for each run. it is clearly seen from table 6 that all measures of the quality of the final solutions such as the minimum, maximum and average cost and the scaled standard deviation are improved when using acoa2 compared to acoa1. it should be noted that, due to the fact that premature convergence does not occur in acoa2 by providing dynamically evolving bounds on the pheromone trail intensities, the acoa2 has been able to outperform the acoa1. in other words, the results show the superiority of the mmas, in conjunction with the acoa2, over than as, in conjunction with the acoa1, in which the trade-off between the two contradictory search characteristic of exploration and exploitation is managed better using mmas. in order to show the efficiency of proposed methods, these results are compared with those obtained with other existing methods. comparison of the results shows that the best results are obtained using acoa2 with no extra computational time cost due to the unique feature of acoa. therefore, the proposed method is useful to solve np-hard combinatorial discrete optimization problems such as optimal design of wds. figure 6. network of goyang wds (test example iii). 1 2 4 11 19 [1] [14] [21] [2] [16] 3 5 16 8 14 10 [18] [4] [7] [5] [23] 6 15 21 13 9 12 [10] [9] [22] [26] [20] [27] [6] [28] [24] [25] p 22 20 [8] [12] [13] 18 [15] [17] [3] [19] 7 17 [29] [30] [11] simulation-optimization model for design of water distribution system using ant based algorithms 56 nodal pressure head is one of the most important constraints of wds design optimization problem. here, the maximum and minimum pressure head of all test examples obtained with acoa2 are presented to show the efficiency of proposed methods. for text example i, the maximum pressure head occurred at node 2 is 174.72 ft and the minimum pressure head occurred at node 6 is 99.87 ft. for text example ii, the maximum pressure head occurred at node 2 is 294.207 ft and the minimum pressure head occurred at node 19 is 255.054 ft. in addition, in case study ii, the pressure head at node 16 and 17 are 260.077 and 272.868, respectively. finally, for text example iii, the maximum pressure head occurred at node 3 is 27.94 meter and the minimum pressure head occurred at node 15 is 15.1 meter. it is worth noting that the obtained pressure heads are satisfied nodal pressure head for all test examples. different methods have been proposed and used to solve text examples i,ii and iii. at first, the best results obtained with the proposed acoa2 for test example i are compared with some other available results. the comparison of the results shows that the optimal solution of 419000 units is obtained at the expense of 4700 function evaluations using acoa2. it is worth noting that the acoa2 has been able to outperform the methods of abebe and solomatine (1999) and savic and walters (1997). furthermore, this compares favourably with about 250000 function evaluations required by the savic and walters (1997) using ga1, about 53000 function evaluations required by cuncha and sousa (1999) using sa, about 100000 function evaluations required by of prasad and park (2004) using ga, about 11155 function evaluations required by of eusuff and lansey (2003) using sfla, about 5100 function evaluations required by of afshar (2007) using aco and about 7467 function evaluations required by wu et al. (2001) using fast messy genetic algorithm (fmga) to get the least cost solution of 419000 units. the comparison of these results indicates that the acoa2 has been able to outperform other methods with no extra computational time cost due to the unique presented feature of this proposed method. in addition, the results of test example ii obtained with the proposed method are compared here with some other available results using different value for constant coefficient of hazen-williams equation )( . the acoa2 is able to get the optimal solution of 38.64 (m$) in 9900 function evaluations )669.10(  . this can be compared approvingly with about 200000 function evaluations required by murphy et al. (1993) to find the solution of 38.80 (m$) using ga )669.10(  , about 1000000 function evaluations required by savic and walters (1997) to find the solutions of 40.42 (m$) and 37.13 (m$) using ga2 and ga1, respectively, )5088.10(  , about 46016 function evaluations required by lippai et al. (1999) to find the solution of 38.13 (m$) using ts )669.10(  , about 1000000 function evaluations required by cunha and sousa (1999) to find the solutions of 37.13 (m$) using sa )5088.10(  , about 37186 function evaluations required by wu et al. (2001) to find the solutions of 37.83 (m$) and 37.13 (m$) using two formulations of fast messy genetic algorithm named fmga1 and fmga2, respectively, )5088.10(  , about 31267 function evaluations required by eusuff and lansey (2003) to find the solutions of 38.13 (m$) using sfla )669.10(  , about 6000 function evaluations required by geem (2006) to find the solutions of 36.66 (m$) using hs )5088.10(  , about 7760 function evaluations required by afshar (2009) to find the solutions of 38.64 (m$) using compact ga (cga) )669.10(  , about 18200 function evaluations required by afshar (2007) to find the solutions of 38.64 (m$) using aco )669.10(  , about 13928 function evaluations required by maier et al. (2003) to find the least-cost solution of 38.64 (m$) using acoa )669.10(  and finally about 22635 function evaluations required by zecchin et al. (2006) to find the least-cost solution of 38.64 (m$) using mmas )67.10(  . it is worth noting that the solutions of lippai et al. (1999) and eusuff and lansey (2003) are infeasible. in other words, the minimum required pressure head is not satisfied at nodes 17 and 19 when obtained diameters are simulated in epanet software. in addition, a smaller value for constant coefficient of hazen-williams equation )( leads to smaller head losses and therefore smaller pipe diameters in which these solutions are infeasible when they are simulated in epanet software. comparison of the results shows that acoa2 has been able to outperform the other used methods with no extra r. moeini and s.a.m. moulaei 57 computational time cost due to the unique feature of this method which is presented before. in addition, the near optimal solution of maier et al. (2003), zecchin et al. (2006) and afshar (2007), 38.64 (m$), is obtained here using proposed acoa2 with smaller computational time. finally, the results of test example iii obtained with the proposed method are compared with some other available results. the acoa2 is able to find the optimal solution of 175783163 (won) in just 8600 function evaluations. this compares favourably with those obtained with other existing methods. kim et al. (1994) solved the problem using a projected lagrangian algorithm supported by gams/minos and then converted the continuous diameters to discrete commercial diameters. they obtained 179142700 (won) while the original cost was 179428600 (won). furthermore, the present result can be compared with about 10000 function evaluations required by geem (2006) to get the solutions of 177135800 (won) using hs. it is worth noting that the acoa2 has been able to outperform all existing methods with no extra computational time cost due to the unique presented feature of this proposed method. superior performance of the acoa2 compared to acoa1 is already attributed to the provision of dynamically evolving bounds on the pheromone trail intensities and methodology of calculating the change of pheromone ij . this fact reflected in the typical convergence curves shown in figs. 7, 8 and 9 for test example i, ii and iii, respectively. it is seen that the population of acoa2 has solution costs way below that of acoa1 due to the fact discussed before. 6. conclusion in this paper, the ant colony optimization algorithm (acoa) was applied to solve water distribution system (wds) design optimization problem using two different methods named acoa1 and acoa2. pipe diameters of the wds network were determined in both acoa1 and acoa2 methods using ant system (as) and max-min ant system (mmas), respectively. for the hydraulic analysis, here, the acoa-based models were interfaced with epanet as simulator. three benchmark test examples were solved using proposed methods and the results were presented and compared. while two proposed methods showed good performance in solving the problems under consideration, the acoa2 was observed to produce better results for wds design optimization problem and to be less sensitive to the randomly generated initial guess required to start the solution process represented by the scaled standard deviation of the solutions produced in ten different runs. in other words, the results showed the superiority of the acoa2 over other the acoa1 in which the trade-off between the two contradictory search characteristic of exploration and exploitation was managed better using mmas which is in conjunction with the acoa2. it was also shown that the acoa2 consistently gave better quality solutions than existing traditional and heuristic search methods with less computational effort due to the unique presented feature of this proposed formulation. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. references abebe aj, solomatine dp (1999), application of global optimization to the design of pipe networks. 3rd international conference on hydroinformatics, copenhagen, balkema, netherlands, 989-996. afshar mh (2007), application of ant algorithm to pipe network optimization. iranian journal of science and technology, transactions b: engineering 31(b5): 487-500. afshar mh (2009), application of a compact genetic algorithm to pipe network optimization problems. scientia iranica. transactions a: civil engineering 16(3): 264271. afshar mh, moeini r (2008), partially and fully constrained ant algorithms for the optimal solution of large scale reservoir operation problems. journal water resource management 22(1): 1835-1857. ali m, pant m, abraham a (2009), a hybrid ant colony differential evolution and its application to water resources problems. in proceedings of the 2009 nature and simulation-optimization model for design of water distribution system using ant based algorithms 58 figure 7. variation of average solution cost values of test example i using acoa1 and acoa2. figure 8. variation of average solution cost values of test example ii using acoa1 and acoa2. figure 9. variation of average solution cost values of test example iii using acoa1 and acoa2. biologically inspired computing world congress 1133 – 1138. alperovits e, shamir u (1977), design of optimal water distribution systems. water resources research 13(6): 885-900. babu k, vijayalakshmi d (2013), self-adaptive pso-ga hybrid model for combinatorial water distribution network design. journal pipeline system engineering practice 4(1): 57– 67. bai d, yang pj, song lx (2007), optimal design method of looped water distribution network. systems engineering-theory and practice 27(7): 137-143. baoyu z, yufei y, xin-hua z (2011), optimization of water distribution system based on improved shuffled frog leaping algorithm. china water and wastewater 27 (9): 45-49. 0.4 0.42 0.44 0.46 0.48 0.5 0 100 200 300 400 500 600 700 800 900 1000 iteration c o s t (m u n it s ) acoa1 acoa2 40 50 60 70 80 90 100 0 100 200 300 400 500 600 700 800 900 1000 iteration c o s t (m $ ) acoa1 acoa2 175.5 176.5 177.5 0 100 200 300 400 500 600 700 800 900 1000 iteration c o s t (m w o n ) acoa1 acoa2 r. moeini and s.a.m. moulaei 58 bi w, dandy gc, maier hr (2015), improved genetic algorithm optimization of water distribution system design by incorporating domain knowledge. environmental modelling and software 69: 370-381. chandra mohan b, baskaran r (2012), a survey: ant colony optimization based recent research and implementation on several engineering domain. expert systems with applications 39: 4618–4627. colorni a, dorigo m, maniezzo v (1991), distributed optimization by ant colonies. proceeding 1st european conference on artificial life (ecal'91), france, paris 134142. costa alh, medeiros jl, pessoa flp (2000), optimization of pipe networks including pumps by simulated annealing. brazilian journal of chemical engineering, 17 (4-7): 887896. cunha mc, sousa j (1999), water distribution network design optimization: simulated annealing approach. journal of water resources planning and management 125(4): 215-221. dong x, liu s, tao t, li sp, xin kl (2012), a comparative study of differential evolution and genetic algorithms for optimizing the design of water distribution systems. journal of zhejiang university science a 13(9): 674686. eusuff mm, lansey ke (2003), optimization of water distribution network design using the shuffled frog leaping algorithm. journal water resource planning management 129(3): 210–225. geem zw (2006), optimal cost design of water distribution networks using harmony search. engineering optimization 38(3): 259– 280. gessler j (1985), pipe network optimization by enumeration. water resource research 23(7): 977-982. haghighi a, samani hmv, samani zmv (2011), ga-ilp method for optimization of water distribution networks. water resource management 25(7): 1791-1808. jian l, yanbing j (2010), optimal design of water distribution system by genetic algorithm. water and wastewater engineering 27(5): 36-39. kim jh, kim tg, kim jh, yoon yn (1994), a study on the pipe network system design using non-linear programming. journal of korean water resources association 27(4): 5967. lippai i, heaney jp, laguna m (1999), robust water system design with commercial intelligent search optimizers. journal of computing in civil engineering 13(3): 135143. maier hr, simpson ar, zecchin ac, foong wk, phang ky, seah hy, tan cl (2003), ant colony optimization for design of water distribution systems. water resources planning and management 129(3): 200–209. mays lw, tung yk (1992), hydrosystems engineering and management. new york, usa: mcgraw-hill. moeini r, afshar mh (2009), application of an ant colony optimization algorithm for optimal operation of reservoirs: a comparative study of three proposed formulations. scientia iranica. transactions a: civil engineering 16(4): 273–285. moeini r, afshar mh (2013a), constrained ant colony optimization algorithm for the layout and size optimization of sanitary sewer networks. urban water journal 10(3): 154-173. moeini r, afshar mh (2013b), extension of the constrained ant colony optimization algorithms for the optimal operation of multi-reservoir systems. journal of hydroinformatics 15(1): 155-173. mohan s, babu k, (2010), optimal water distribution network design with honey-bee mating optimization. journal of computing in civil engineering 24(1): 117–126. murphy lj, simpson ar, dandy gc (1993), design of a network using genetic algorithms. water 20: 40-42. ostfeld a (2011), ant colony optimization for water resources systems analysis – review and challenges. ant colony optimization methods and applications. chapter 17, published by intech. prasad td, park ns (2004), multiobjective genetic algorithms for design of water distribution networks. journal of water resource planning and management 130(1): 73–82. rossman la (1999), the epanet programmer’s toolkit. in proceedings of water resources planning and management division annual specialty conference, asce, tempe, az, 1-74. rossman la (2000), epanet 2 users manual. national risk management research laboratory, office of research and 59 http://link.springer.com/search?facet-author=%22xiao-lei+dong%22 http://link.springer.com/search?facet-author=%22sui-qing+liu%22 http://link.springer.com/search?facet-author=%22tao+tao%22 http://link.springer.com/search?facet-author=%22shu-ping+li%22 http://link.springer.com/search?facet-author=%22kun-lun+xin%22 http://link.springer.com/journal/11582 http://link.springer.com/journal/11582 http://link.springer.com/journal/11582/13/9/page/1 simulation-optimization model for design of water distribution system using ant based algorithms 60 development, u.s. epa, cincinnati, oh 45268. savic da, walters ga (1997), genetic algorithms for least-cost design of water distribution networks, journal of water resources planning and management 123(2): 67-77. schaake j, lai d (1969), linear programming and dynamic programming – application of water distribution network design. report 116, mit press, cambridge, ma. sheikholeslami r, zecchin ac, zheng f,talatahari s (2016), a hybrid cuckoo– harmony search algorithm for optimal design of water distribution systems. journal of hydroinformatics 18(3): 544-563 shibua a, janga reddya m (2012), reliabilitybased optimal design of water distribution networks under uncertain demands using cross-entropy method. ish journal of hydraulic engineering 18(3): 258-268. stutzle t, hoos h (2000), max-min ant system. future generation computer systems 16(8): 889-914. stutzle t, lópez-ibáñez m, dorigo m (2011), concise overview of applications of ant colony optimization. wiley encyclopedia of operations research and management science, john wiley and sons inc, 896-911. wu zy, boulos pf, orr ch, ro jj (2001), using genetic algorithms to rehabilitate distribution systems. journal of american water works association 93(11): 74-85. xu c, goulter i (1999), reliability-based optimal design of water distribution networks. journal of water resource planning and management 125(6): 352–362. yang l, sui j, hua z (2012), harmony search algorithm for optimal design of water supply networks. journal of theoretical and applied information technology 46(2): 526-536. zecchin ac, simpson ar, maier hr, leonard m, roberts aj, berrisford mj (2006), application of two ant colony optimization algorithms to water distribution system optimization. mathematical and computer modelling 44: 451–468. zecchin ac, maier hr, simpson ar, leonard m, nixon jb (2007), ant colony optimization applied to water distribution system design: comparative study of five algorithms. journal of water resources planning and management 133(1): 87-92. zhou x, gao dy, simpson ar (2016), optimal design of water distribution networks by a discrete state transition algorithm. engineering optimization 48(4): 603-628. http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&searchtype=journal&result=true&prevsearch=%2bauthorsfield%3a(shibu%2c+a.) http://www.tandfonline.com/action/dosearch?action=runsearch&type=advanced&searchtype=journal&result=true&prevsearch=%2bauthorsfield%3a(janga+reddy%2c+m.) http://www.tandfonline.com/loi/tish20?open=18#vol_18 http://www.tandfonline.com/toc/tish20/18/3 http://www.tandfonline.com/author/zhou%2c+xiaojun http://www.tandfonline.com/author/gao%2c+david+y http://www.tandfonline.com/author/simpson%2c+angus+r d:\desktop\paper_4.xps tjer 2013, vol. 10, no. 1, 41-49 _______________________________ *corresponding author’s e-mail: farooq@mi.utoronto.ca expansion of tubular with elastomers in multilateral wells md veldena and fks al-jahwari*b asenior well engineer, concept & design, shell oil, seria, brunei bmechanical engineering department, university of toronto, canada received 29 january 2012; accepted 30 september 2012 abstract: the use of solid expandable tubular technology during the last decade has focused on solving many challenges in well drilling and delivery including zonal isolation, deep drilling, conservation of hole sizes, etc. not only as pioneered solution but also providing cost effective and long lasting solutions. concurrently, the technology was extended for construction of multilateral in typical wells. the process of horizontal tubular expansion is similar to the vertical expansion of expandable tubular in down-hole environment with the addition of uniformly distributed force due to its weight. the expansion is targeted to increase its diameter such that post expansion characteristics remain within allowable limits. in this study a typical expandable tubular of 57.15 mm outer diameter and 6.35 mm wall thickness was used with two different elastomer seals of 5 and 7 mm thickness placed at equal spacing of 200 mm. the developed stress contours during expansion process clearly showed the high stress areas in the vicinity of expansion region which lies around the mandrel. these high stresses may result in excessive wear of the mandrel. it was also found out that the drawing force increases as the mandrel angle, expansion ratio, and friction coefficient increases. a mandrel angle of 20o requires minimum expansion force and can be considered as an optimum geometrical parameter to lower the power required for expansion. keywords: multilateral well, horizontal expansion, finite element analysis, oil well, tubular 42 md velden and fks al jahwari 1. introduction the core issue in developing the solid expandable systems is to reduce installation risks, while driving down construction and well repair costs and minimizing drilling hazards. solid tubular expansion technology is essentially a down-hole cold working process of a solid steel tubular aiming to increase its diameter with the mean of a hardened mandrel through the use of pressurized fluid or mandrel on stick approach. in both cases the effective internal forces must not exceed the allowable tubular strength with the provision of reasonable safety limit. the expansion process subjects the solid tubular to a large plastic deformation leading to variations in tubular thickness and length, which may result in premature and unexpected failures. research work in expandable tubular technology started to provide solutions for many problems such as zonal isolation, deep drilling, conservation of hole sizes, etc. (stewart et al. 1999; daigle et al. 2000; owoeye et al. 2000). there is no doubt that the most significant benefit of expandable tubular technology is its enabling capacity to drill deeper, and reach hard to access and small scattered pockets of reservoirs. it had been proven that the technology can be extended to accomplish multilateral wells using special arrangement at the branching point from the main borehole (zhang et al. 2009). for a number of years, the industry has sought to use single diameter expandable liner extensions as an alternative to conventional casing designs. a single diameter expandable liner extension enables the drilling of deeper exploration and production wells with larger hole sizes. a solid expandable system developed in collaboration between baker oil tools and british petroleum allows for the isolation of reactive shale zones, subsalt environments, low-fracturegradient formations or other drilling situations without reducing the casing and subsequent drilled-hole size into the reservoir (coolkidge et al. 2007). weatherford (2009) has successfully designed and developed another variant of tubular technology called metal skin system. the system was developed to tackle four main functions. these include: 1) open-hole liner; 2) mono-bore open-hole liner; 3) cased-hole liner; and 4) mono-bore open-hole clad. most of these systems can be used for multiple applications, to address both pre-planned and remedial well construction challenges. however, there are slight differences in the problems targeted by the individual system versions. another operator has developed one of its fields using open-hole, multi-lateral completions in longreach horizontal wells. water production problems and the inability to re-enter open-hole sidetracks led to the use of solid expandable tubular and intelligent well technology to control the wells. the horizontal well was transformed into a tri-lateral forked well. binggui et al. (2009) developed a finite element model sing a commercial finite element analysis software to analyze the stress, strain, and residual stress in an expandable tubular without elastomers and applied it in casing repair of damaged wells, as well as to monitor the changes on internal surfaces of the tubular. however, their model simulates expansion of vertical tubular without the inclusion of weight effect. akhtar et al. (2012) compared different elastomer material model when used with expandable tubular for expansion. they developed a suitable hyper-elastic material model for elastomer by best fitting the tensile behavior of the elastomers commonly used with expandable tubular. the experiments were conducted on exposed elastomers and the experimental data was then fitted to neo-hookean model to be used for finite element simulations. the numerical simulation of expandable tubular in horizontal wells for a combined tubular-elastomer system was not investigated by many researchers due to the complexity at tubular/mandrel/formation contact surfaces. pervez et al. (2007; 2005) simulated the expansion of tubular with elastomers for both horizontal and vertical cases. the results were summarized with findings of significance and relevance to field engineers. it was found that the drawing force increases with an increase in mandrel angle, expansion ratio, and friction coefficients at interfaces. tubular wall thickness decreases as the mandrel angle, expansion ratio, and friction coefficient increase. tubular length often shortens with a reduction in thickness to compensate the increase in its diameter and maintain the conservation of mass. this paper presents an extension to this study with more insight to the control parameters of the expansion process, and visualization and better understanding of stress variations at tubular/mandrel interface. 2. finite element modeling the horizontal tubular expansion with elastomer seals had been modeled in abaqus, commercial finite element analysis software, using explicit solver that adequately accounts for the nonlinearity in the system. figure 1 shows a cross-sectional view of the tubular-elastomer geometric model. figure 2 shows 3d representation as was developed in finite element model. it consists of the tubular surrounded by 200 mm long elastomer seal of either 5 or 7 mm thickness. the elastomer seals are placed in cascades with a gap of 200 mm. the tubular is 1500-mm long and has an 43 expansion of tubular with elastomers in multilateral wells inner diameter and thickness of 57.15 mm and 6.35 mm, respectively. tubular expansion was achieved by pushing an analytical rigid mandrel through the tubular by prescribed mandrel displacement in the software. formation was modeled as analytical rigid body which implies that the surrounding medium is rocklike material. seals were assumed to be perfectly bonded to the tubular ie. no relative motion between seal and tubular is allowed while the interaction takes place between mandrel, tubular and formation. the contact conditions at these interfaces were modeled using coulomb friction law. linear brick finite element having 8-nodes per element with reduced integration was used for the whole system with optimized mesh size and reasonably enough through-thickness elements to account for the tubular-seal elongation at the expense of reduction in cross-sectional area of each. due to the symmetry only one half of the tubular was considered while the axi-symmetric approximation was excluded because of gravitational effects. three different mandrel angles of 10, 20 and 45 degrees were used to study the effect on expansion behavior. for each of these mandrel angles, the parameters of interest such as expansion ratio, seal thickness and friction coefficients were varied to see their effect on expansion process. different expansion ratios were governed by the tubular diameter after expansion. table 1 shows a summary of the parameters used in this study and their corresponding values. for field needs, effects on the drawing force, length and thickness of the tubular were investigated by varying the friction coefficients at the tubular-mandrel and sealformation interfaces, mandrel angle (geometrical factor that significantly affect the expansion process), seal thickness, and expansion ratio. 3. material model tubular was modeled as an elastic-plastic material obeying the power hardening law due to the large plastic deformation resulting from the expansion process. typical steel data was used with modulus of elasticity e = 200 gpa, poisson's ratio = 0.3, material yield strength sy = 500 mpa, and hardening parameter n = 0.25. subsequently, actual stress-strain data obtained through experiment was used to investigate similar problems under realistic conditions. the elastomer seal was considered as a non-linear elastic material due to high compression ratios. elastomeric material of the seal was modeled using arruda-boyce model in abaqus. in general, an elastomer is very flexible for tensile and shear loading, but extremely stiff for hydrostatic deformations. arruda-boyce model parameters were obtained by the abaqus built-in curve fitting procedure using experimental nominal stress vs. nominal strain data. the stability of the solution is achieved by carefully selecting the step sizes in abaqus. however, neo-hookean model would be more accurate as suggested by akhtar et al. (2012) but may not necessarily yield a numerically stable model. 200 mm 200 mm elastomer seal seal thickness = 5, 7 mm tabular formation figure 1. cross-sectional view of horizontal tabular with seals and formation table 1. values of parameters used in simulation of horizontal solid tabular expansion 44 md velden and fks al jahwari 4. boundary and loading conditions due to symmetry all nodes lying on the plane of symmetry were constrained in the direction perpendicular to the plane. in addition, the left end of the tubular was held fixed hence the tubular expands under tension. the right end of the tubular was free to move. the mandrel moved from left to right at a constant low speed of 1.5 m/min. this was achieved by defining a reference node at the center of the mandrel which followed a pre-defined path during expansion process. due to low mandrel speed, the strain rate effect was assumed to be negligible. all six degrees of freedom for the reference node of the formation were fixed. the mandrel was also constrained from rotation about its own axis. the gravity force was applied on the tubular and mandrel, and assumed negligible for seals. this was due to much smaller weight of seals compared to the tubular. 5. results and discussions results type and representation were carefully selected to support the requirements of field engineers. these results can be of importance in decision making by well engineers at the start and during expansion process. the important expansion parameter studied in this paper is the expansion or drawing force. geometrical and contact factors affecting it are also investigated. figure 3 shows the stress contours of effective stresses for four different expansion ratios, for coefficient of friction = 0.1 and mandrel angle = 10o. although the stress contours look similar indicating that the expanded zone reaches yielding in all cases, the tubular thickness variation changes with the expansion ratio and becomes thinner as the expansion ratio increases to 22% as was also concluded by pervez et al. (2007). the stress contour patterns for 5 or 7 mm thick seals look similar but differ in values. the drawing force required for 7 mm thick seal is higher than 5 mm thick seal. figures 4 and 5 show the axial, radial, shear, and effective stress contours for various tubular expansion ratios and indicate that the stresses vary along the tubular thickness and length. stresses concentrate in the mandrel middle section forming a ring. this indicates the possibility of mandrel/tubular wear due to high friction at interfaces. three important operational parameters that affect the force required for expansion are investigated. those parameters are expansion ratio, friction coefficient, and the mandrel angle. the expansion ratio is selected as per the production well requirements and conditions. however, friction coefficient is always aimed to be minimized within a reasonable cost limit but the mandrel angle may be optimized without significant economical affect. the drawing force increases with the friction coefficient and expansion ratio irrespective of mandrel angle as shown in fig. 6. it is interesting to note that the variation in expansion force elastomer seal tabular/formation interface tabular mandrel mandrel angle figure 2. 3-d representation of tabular and mandrel model 45 expansion of tubular with elastomers in multilateral wells figure 4. plot of stress contours for = 0.2, = 10 for 15% expansion ratio figure 5. plot of stress contours for = 0.2, = 10 and 22% expansion ratio figure 3. effective stress contours for = 0.1 and = 10o 46 md velden and fks al jahwari becomes very small for 15% expansion ratio when the value of coefficient of friction increases to 0.3 or more. this suggests that low cost coating has to be used on tubular internal surfaces when the mandrel is 45 degree and desired maximum expansion ratio is 15% or less. figure 7 show the variation in expansion forces with respect to friction coefficient for different mandrel angle. the variation pattern is similar to that of fig. 6 with one exception. all four figures, representing expansion ratios from 15% to 22% indicate that the best mandrel angle for lower expansion force is 20o. figure 8 shows the variation in expansion forces with respect to the mandrel angle at four different values of coefficient of friction. it is clear from the figure that the expansion force decreases as the mandrel angle increases. the lowest expansion force is attained in the vicinity of 20 degree mandrel values. the mandrel angle is not the same for all coefficients of friction and varies between 18 and 20 degrees. however, as a general guideline, the mandrel angle of 20 degree will requires less power or expansion. it reinforces the conclusion from fig. 7. again, at higher value of friction coefficient (µ = 0.4), increase in mandrel angle beyond figure 6. variation in drawing forces w.r.f friction coefficient for different expansion ratios 47 expansion of tubular with elastomers in multilateral wells 20 degree does not alter the required expansion forces. hence, a mandrel with 20 degree angle is desirable due to reduced manufacturing cost. 6. conclusions the simulation of solid expandable tubular, using finite element analysis, in horizontal wells for a combined tubular-elastomer system was performed for different contact conditions representing various field scenarios. the seals of 200 mm length and 5 or 7 mm thick were placed at equal intervals of 200 mm. stress contours showed that the expanded zone reaches yielding in all cases and form a high stress ring around the effectively expanding edge of the mandrel strengthening the view of excessive wear of mandrel. expansion forces were found to increase as the mandrel angle, expansion ratio, and friction coefficient increase. the effect of one parameter over the other is negligible. figure 7. variation in drawing w.r.f friction coefficient for different mandrel angles 48 md velden and fks al jahwari references akhtar m, qamar sz, pervez t, khan r, al-kharusi msm (2012), elastomer seals in cold expansion of petroleum tubulars: comparison of material models. journal of materials and manufacturing processes 27:715-720. binggui x, zhang y, wang h, hongwei y, jia t (2009), application of numerical simulation in the solid expandable tubular repair for casing damaged wells. journal of petroleum exploration and development 36(5):651-657. coolidge r, baker rv, stockmeyer ce, adam m, emerson b (2007), bp, baker run first expandable monobore liner extension system. practical drilling technology special report, oil and gas journal 12:1-5. daigle cl, campo db, naquin cj, cardenas r, ring lm, yourk p (2000), expandable tubulars: field examples of application in well construction and figure 8. variation in drawing force w.r.t the mandrel angle for various friction coefficients 49 expansion of tubular with elastomers in multilateral wells remediation. spe paper # 62958, spe annual technical conference and exhibition, dallas, texas. owoeye oo, aihevba lo, hartmann ra, ogoke vc (2000), optimization of well economics by application of expandable tubular technology. iadc/spe paper # 59142, iadc/spe drilling conference, new orleans, louisiana. pervez t, seibi ac, al-hiddabi sa, al-jahwari fk, qamar sz, marketz f (2007), solid tubular expansion in horizontal wells. spe paper # 105704, 15th middle east oil and gas show and conference, meos, bahrain. pervez t, seibi ac, karrech a (2005), simulation of solid tubular expansion in well drilling using finite element method. journal of petroleum science and technology 23:775-794. stewart rb, marketz f, lohbeck wcm (1999), expandable wellbore tubulars. spe paper # 60799, spe technical symposium, dhahran, saudi arabia. weatherford meta skin system. 2009. http://www.weatherford.com/ecmweb/groups/ web/documents/weatherfordcorp/wft098992.pdf. zhang y, rongquan r, wang h, wang j (2009), multilateral drilling and completion technology based on solid expandable tubular fixing system. journal of petroleum exploration and development 36(6):768-775. mech050808_l.qxd the journal of engineering research vol. 4, no.1 (2007) 48-63 1. introduction in recent years, the increased usage of air conditioning systems in hot/humid climates has resulted in a significant increase in demand for primary energy during summer months. in some countries, blackout situations have occurred during periods of peak load due to overburdened distribution networks. most air-conditioning systems operate using a vapour compression cycle and are electrically driven by means of fossil fuel. this results in a large amount of co2 being released to the atmosphere, which contributes strongly to global warming. the steady rise in the temperature of the earth's atmosphere will of course exacerbate the demand for air-conditioning, thus creating a vicious cycle in terms of cooling and electricity demand. use of solar energy to power air-conditioning systems would allow this cycle to be broken as peak cooling loads coincide with maximum available solar power. various cycles can be used to provide cooling. each has some advantages, but also some limitations in comparison with other cycles. the following section presents a brief _____________________________________________ *corresponding author’s e-mail: saffa.riffat@nottingham.ac.uk description of each cycle followed by a comparison with other methods. the technical, environmental and economic issues associated with various solar-driven cooling/refrigeration systems are also discussed. 2. types of cooling cycle 2.1 conventional refrigeration (vapour compression cycle) the most widely used refrigerators are those, which use a liquefiable vapour as the refrigerant. the evaporation and condensation processes take place when the fluid is receiving and rejecting latent heat at constant temperature and pressure. fig. 1-b shows the schematic diagram of a vapour-compression cycle, while fig. 1-a illustrates the cycle on a p-h diagram. the ideal cycle can be described as follows: at point 1, the refrigerant is a low-pressure vapour. the vapour pressure as well as the temperature increases to point 2 by the effect of the compressor through path 1-2 (isentropic compression). from point 2 to point 3, the vapour passes through the condenser where heat is removed, condensing the vapour to a sub-cooled liquid (or saturated liquid) at constant pressure. from point 3 to point 4 the high-pressure liquid refrigerant is throttled through an expansion device (valve), where solar-driven air-conditioning cycles: a review a. m. abu-zour, s. b. riffat* school of the built environment, university of nottingham, nottingham, ng7 2rd, uk. received 8 august 2005; accepted 9 may 2006 abstract: most conventional cooling/refrigeration systems are driven by fossil fuel combustion, and therefore give rise to emission of environmentally damaging pollutants. in addition, many cooling systems employ refrigerants, which are also harmful to the environment in terms of their global warming potential (gwp) and ozone depletion potential (odp). development of a passive or hybrid solar-driven air-conditioning system is therefore of interest as exploitation of such systems would reduce the demand for grid electricity particularly at times of peak load. this paper presents a review of various cooling cycles and summarises work carried out on solar-driven air-conditioning systems. keywords: solar energy, cooling cycles, air-conditioning system, and coefficient of performance (cop) ¢vgô©à°sg / ¢ùª°ûdg ≈∏y á«æñÿg agƒ¡dg ∞««µj πªy ägqho äé©aq õh .¢s ,qhr ƒhcg .ω .cg áá°°uuóóÿÿgg᪶fcg øe ~j~©dg ¿g ,∂dp ¤g ±é° j .é«ä«h iqé° dg äékƒ∏ÿg çé©ñfg ¤g …odƒj ‹éàdéhh iqƒ° m’g oƒbƒdg ¥gîmg ᣰsgƒh πª©j áj~«∏≤àdg ~jèàdg ᪶fg º¶©e ¿g : ‘ qƒ£àdg ¿g .¿hrhc’g á≤ñw ±gõæà°sgh iqgô◊g álqo ´éøjqéc á«ÿéy á«ä«h πcé°ûã é¡ññ°ùj á«méf øe áä«ñdg ≈∏y gô£n πµ°ûj é° jg »gh ,ioèÿg πfgƒ°ùdg ω~îà°ùj ~jèàdg ábqƒdg √òg .ihqòdg äébhg ‘ é°uƒ°ünh aéhô¡µdg ≈∏y ö∏£dg π∏b ~b ᪶f’g √òg πãe ∫ó¨à°sg ¿g ¢sé°sg ≈∏y ᫪g’g øe íñ°ug ¢ùª°ûdg ≈∏y ≈æñÿg ∞««µàdg ωé¶f úé¡j .¢ùª°ûdéh á∏eé©dg agƒ¡dg ∞««µj ᪶fcg ≈∏y áh ωé«≤dg ” »àdg πª©dg ¢üî∏jh ~jèàdg ägqho ∞∏àfl ¢vô©à°ùj áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.ago’g πeé©e ,agƒ¡dg ∞««µj ωé¶f ,~jèàdg äé°sgqo ,á«°ùª°ûdg ábé£dg : 49 the journal of engineering research vol. 4, no.1 (2007) 48-63 some liquid flashes to vapour, forming a two-phase mixture at low pressure and temperature. the two-phase (liquid-vapour) mixture flashes to vapour on entering the evaporator, where the heat necessary for evaporation is taken from the space being conditioned. for the cycle to be termed ideal, all the individual processes would need to behave exactly as specified above. in the actual cycle however, pressure loss occurs in the evaporator and the condenser. in addition, the compressor does not operate isentropically (adiabatic and reversible), ie. the compression process is an irreversible process (fig.2). the refrigerant does not enter the compressor in a dry saturated condition. generally, it has a small degree of superheating to avoid droplet formation on entering the compressor that could cause impact damage. the actual process can be represented in figure 2-a by shifting point 1-a little to the right. the condensed vapour can be cooled at constant pressure (condenser pressure) to a temperature below the saturation temperature, ie. line 3-4, which represents the throttling process and can be moved to the left. this results in an increase in the refrigeration effect. the effect of under-cooling/sub-cooling of the refrigerant after the condenser would offset the performance reduction slightly. most existing cooling and refrigeration systems employ a vapour compression cycle due to its high cop. however, these systems are now less attractive as the chlorofluorocarbons (cfcs) refrigerants have a high ozone depletion potential (odp) and a high global warming potential (gwp). other problems associated with vapour compression cycles include: adaptation of the vapour compression cycle to allow use of environmentally non-damaging refrigerants has been encouraged to protect the ozone layer and reduce global warming. the use of thermal energy, such as industrial waste heat, geothermal and solar energy has also been investigated as a mean of alternative source of energy for the refrigeration cycle in order to reduce the demand of (b) actual (a) ideal cycle 1 2 3 4 enthalpy pressure isentropic lines fig. 1 -a p-h diagram for vapour compression cycle fig. 1 -b schematic diagram for vapour compression cycle condenser evaporator expansion c 1 2 3 4 input power input power figure 1-a p-h diagram for vapour compress cycle figure 1-b schematic diagram for vapour compression cycle fig. 2 classification of solar refrigeration cycle s. solar cooling / refrigeration system solar -photo -voltaic system solar thermal system bryton cycles stirling cycles rankine cycles vapour compression cycles vapour absorption cycles vapour adsorption cycles vapour ejector cycles thermo -electric cycles gas refrigeration cycles figure 2. classification of solar refrigeration cycles • hazardous working fluids. • high consumption of energy, since the compressor is powered by electrical energy. • lifetime of the system is relatively short, due to wear of t he moving parts. • electrical energy is usually supplied by fossil fuel combustion, producing greenhouse gases and other pollutants. • noise and vibration due to moving parts. 50 the journal of engineering research vol. 4, no.1 (2007) 48-63 primary energy from fossil fuels. a number of refrigeration cycles have been devised to utilize solar energy as a clean and renewable energy source. the following section provides an overview of refrigeration cycles that could be integrated with solar energy and utilized in place of a vapour compression refrigeration cycle (see fig. 2). the primary benefit of solar powered systems is reduced demand for grid electricity in the afternoon when load is at its peak. a solar-driven vapour compression cycle is one solution to some of the problems, mentioned previously, ie; high consumption of primary energy and negative environmental impact of some emissions. the rankine cycle and photovoltaic can be regarded as the two main techniques for driving a vapour compression cycle using solar energy. many refrigerants traditionally used in vapour-compression systems exhibit excellent thermo-physical properties, but have been banned because of their adverse environmental impact, ie. high odp and gwp. some refrigerants have negligible odp but still have a considerable gwp as illustrated in table 1 ( us epa http://www.epa.gov/ ). this has created the need to identify new, environmentally friendly refrigerants for vapour compression systems. the most important feature of these refrigerants include: 1. non-toxic, 2. non-flammable, 3. non-corrosive 4. very low freezing temperature (well below evaporator temperature) 5. very high boiling temperature (well above condenser temperature) for example, the first two criteria are not satisfied by ammonia, which also corrodes copper pipes, and co2 does not fulfill the fifth condition. 2.1.1 rankine vapour compression cooling cycle a rankine vapour compression cooling cycle consists of two component cycles, ie. the rankine power cycle drives the vapour compression cooling cycle. the basic principle of this combined cycle is illustrated in fig. 3. the working fluid (liquid state) is heated under high pressure in the boiler of the rankine cycle by the collected solar radiation. the high-pressure vapour is then expanded through a turbine to a lower pressure and temperature, which results in mechanical work at the turbine shaft. the vapour is then condensed and returned to a liquid state under the same pressure. pumping the liquid back to the boiler allows the cycle to be repeated. the mechanical work generated at the turbine shaft is used to drive the compressor of the vapour compression cycle. coupling these two cycles could be carried out mechanically through a common shaft, or electrically using electric generator, which would supply electricity to the compressor and other electrical loads. flat plate and evacuated solar collectors would be suitable for use with the rankine power cycle (curran, 1992). 2.1.2 photovoltaic-driven vapour compression cooling cycle this cycle has two main configurations. the first uses direct current, dc, supplied directly from the photovoltaic, pv, and the second uses alternating current, ac, supplied by the use of an inverter to convert the dc generated by the pv. the second configuration is preferable, as it is possible to connect it to the national grid (with higher efficiency) without the need for expensive battery storage. dcdriven vapour compression systems are suitable for rural areas, where buildings may be remote from the electricity grid. the rankine cycle is the most efficient thermodynamic cycle, compared with stirling and bryton cycles, for driving a vapour compression cycle. on the other hand, it is significantly more complex than a photovoltaic driven vapour compression cycle from an energy conversion refrigerant type refrigerant odp* gwp** co2 r744 0 1 r11 1 4000 cfc r12 1 8500 r113 0.8 5000 hcfc r22 0.05 1700 r401b 0.035 1190 r507 0 3800 r123 0.02 93 hfc r134a ~0 3200 r23 0 12100 *odp quoted relative to r11 and r12, the odp of which has been designated as1. ** gwp quoted relative to co2, the gwp of which has been designated as1. table 1. odp and gwp of cfc, hcfs and hfc refrigerants condens evaporato c boiler condenser t heat supplied from solar collector expansion valve rankine cycle vapour compression cycle figure 3. schematic diagram of rankine-vapour compression cooling cycle boiler condenser evaporator condenser t c 51 the journal of engineering research vol. 4, no.1 (2007) 48-63 aspect. in other words, a rankine cycle involves two more conversion processes, compared with a photovoltaic cycle, to operate the cooling cycle as illustrated in the fig. 4. in addition to providing direct electrical energy, a pv system has several other advantages: ept can be defined as the number of years that are required for the cell modules to generate the same amount of electric power as was consumed in their fabrication. this depends on conversion efficiency and production volume. at a production volume of 10 mw/year in the case of a 8% conversion efficiency solar cell, the ept is estimated to be 1.2 years for a-si solar cell and 4 years for poly-si solar cell (ohnishi et al., 1995). however pv systems also have some disadvantages: it should be noted that the cost of pv technology is likely to fall significantly as production volumes increase and improvements in efficiency are also probable as technology develops. as an example, a protection against over-current for a pv driven air-conditioning system has been developed (wagdy et al., 1995). pv-cells can also be integrated in buildings as architectural element in an aesthetically pleasing manner, so land displacement and visual intrusion can be alleviated. cooling of pv cells in order to maintain their efficiency can be accomplished by attaching a pcm (phase change material) to the back of the modules, or by passing a stream of air through a ducting from the back side so that the hot air generated may be used in other utilities (morgan and prasad, 2001). jubran et al. (2003) has carried out a feasibility study on various pv systems namely; stand-alone pv (sapv), a hybrid grid pv (gpv), grid pv wind turbine (gpvw) and pv wind turbine (pvw) for supplying the power requirements of a window-type air conditioning system under hot-arid climat (oman climate condition on 2003). it was found that the hybrid system (gpvw) was the preferred system with minimum net present cost but higher capital cost depending on the wind speed. in other words, the gpvw system has the least total net present cost among all systems at around $33,000 with a capital cost of $39,500 for 5 m/s wind speed. increasing the wind speed to 8 m/s resulted in a further decrease in the net present cost to less than $10,000, while the capital cost increased from $39,500 to around $56,500. 2.2 absorption cooling cycle the absorption refrigeration cycle is well known. ferdinand carre (chad et al., 1995) developed the first absorption machine in the mid-nineteenth century. the key difference between this cycle and a vapour-compression cycle is the process by which the low-pressure refrigerant (vapour) is transformed to a high-pressure vapour, path 1-2 (see fig.5-a or 5-b). in an absorption cycle, the low-pressure vapour is absorbed into an absorber (either solid or liquid) at low pressure, pumped to a high pressure, and then heated to produce a high-pressure vapour. the compressor function in an absorption cycle is commonly referred to as a thermal compressor, chemical compressor, or thermo-chemical compressor (see fig.5-b), whereas in a vapour-compression cycle, any type of mechanical compressor (reciprocating, rotary, screw, …etc.) can be used. in order that the sequence of events should be continuous, it is necessary for the refrigerant to be separated from the absorbent and subsequently condensed before being returned to the evaporator. path 1-2 in the absorption cycle can be described as follows: the low-pressure, low-temperature refrigerant vapour enters the thermal compressor at point 1, and is absorbed into a solution (for example) at (2-0). the absorbent-refrigerant mixture is then pumped, via the heat exchanger, to the generator at point (2-3), where the refrigerant vapour is separated from the absorbent solution by heating, this process generates high-temperature high-pressure refrigerant vapour which is then passed to the condenser and the weakened cooling cycle heat energy (in the boiler) mechanical energy solar energy (solar collector) electrical energy solar energy (solar cell) electrical energy rankine route photovoltaic route fig. 4 schematic diagram of rankine cycle and photovoltaic cycle routes. figure 4. schemtic diagram of rankine cycle and photovoltaic cycle routes no co2 emission very long service life due to the absence of moving parts (estimated as longer than 20 years) reliable; since it operates at very low irradiance (even on cloudy days) energy payback time (ept) is very much shorter than its lifetime (ohnishi et al., 1995) higher initial cost land displacement visual intrusion. reduction of performance at higher temperatures. 52 the journal of engineering research vol. 4, no.1 (2007) 48-63 absorbent solution is returned to the absorber. a (p-h) diagram cannot be used to graphically show the absorption cycle. instead, a chart that includes the solution concentration -a ptx equilibrium charthas been used. fig. 3-a shows the ptx equilibrium chart, which shows the relationship between the refrigerant vapour pressure and its corresponding temperature, the solution temperature and the concentration of the solution. currently two absorbent-refrigerant systems are widely employed, namely, lithium bromide (libr)-water (aqueous) system and water (aqueous)-ammonia system (aar). in the first system, the libr, a salt, is the absorbent and the water is the refrigerant, while in the second system, the water is the absorbent and the ammonia is the refrigerant. absorption machines are categorized either by the number of effects, as in libr-water refrigeration system, or by the number of stages, as in an aqueous-ammonia refrigeration (aar) system, to increase their cop. in this guide, an effect refers to the number of times the input heat is used by the absorption machine, either directly or indirectly (chad, et al. 1995). in single-effect system, the input heat is used once. in double-effect system the input heat is used twice and so on. stage refers to the number of evaporator/absorber pairs at different temperatures in an absorber machine. the cooling outputs of these two systems are different. the cooling output for a libr system is (4)-(38) oc for a single-effect and (4)-(27) oc for a double-effect, and this is suitable for air-conditioning applications. for an aar system the cooling effect is (-51) (4) oc and is suitable for refrigeration applications. the popularity of the absorption system has been influenced by economic conditions and technology breakthroughs from competing technology. the benefits of absorption systems over vapour-compression systems can be summarized as follows: existing absorption systems however, do have some limitations in that they can be used only with a few refrigerant/absorbent combinations, principally water/libr. concentrated libr solutions have a relatively high crystallization temperature and can freeze inside the chiller unit. they are corrosive to metals and are very expensive. many refrigerant/absorbent combinations have been mentioned in the literature, but these have similar disadvantages to libr solution. however, many limitations can be overcome by the use of a new working fluid namely, water/potassium formate (h2o/hcook) in absorption systems (riffat et al., 1999). the physical properties of potassium formate present many benefits, ie. it is less corrosive, less expansive to manufacture and has a lower density and viscosity than absorbents such as libr. other challenges facing solar-driven absorption systems are: over the last three decades various attempts have been solar heat in thermal compressor fig. 5-b schematic diagram absorption refrigeration cycle (single effect). metering & expansion device condenser evaporator 1 2 generator [2-3] heat exchanger absorber [2-0] 3 4 2-1 2-2 2-4 2-5 fig. 5-a ptx chart for absorption refrigeration cycle. (for single effect) generator solution temperature crystallization r ef ri ge ra nt v ap ou r p re ss ur e re fr ig er an t v ap ou r t em pe ra tu re 1&4 2-1 2-2 2-3 2-4 2-5 2-0 2&3 0% libr (100%water) evaporator absorber diluted concentrated high libr concentration condenser figure 5-a ptx chart for absorption refrigeration cycle (for single effect) figure 5-bschematic diagram absorption refrigeration cycle (single effect) • absorption systems do not contribute to ozone depletion and may have less impact on global warming. • they are quiet and vibration free and hence have a longer operational lifetime. • they have a lower electrical need compared to vapour-compression systems (ie. much less work is required to compress liquid rather than compress vapour). • diversity of energy source (ie. absorption systems can be driven by waste heat, solar energy or hybrid sources (e.g. solar/gas). • high initial cost, • low cop • bulky and heavy • difficulty in storing high temperature heat (i.e. short period of solar irradiance per day, hence not yet competitive with gas-fired or electricity driven a/c systems. 53 the journal of engineering research vol. 4, no.1 (2007) 48-63 made to improve the performance of solar-driven absorption systems through improvement of the major components of the system, eg. absorbent material, chiller, solar collector and storage tank. in order to achieve continuous operation and improve system performance, a hot water storage tank is essential. it has been reported by (kreider and kreith 1981) that using two hot storage tanks maintained at different temperatures would allow the system to operate more efficiently. partitioning a single hot water storage tank into an upper part and a lower part has also been investigated (li and sumathy, 2001). in the morning, the upper part would operate the system, while the whole tank would serve as a heat reservoir in the afternoon. the efficiency of the absorption cycle can be increased by raising the operating temperature through multieffects. the typical operating temperature for a single effect is in the range of 60 to 80 oc, while for a double effect it is 100 to 160oc. these ranges of operating temperature are highly suitable for solar energy applications using different solar collector techniques. double or triple effect systems however, are extremely expensive, due to the cost of high temperature collectors. but the question arises here: are multi-effect systems of higher cop, more suitable for solar powered cooling than their single-effect counterpart, with the availability of high temperature gasfired systems. this question has been studied (grossman, 2002) by conducting a comparison between single-, double-, and triple-effect systems, using libr-water chillers producing chilled water at 7 oc with cooling water supplied at 30 oc (refer to table 2). it is indicated that the total system cost is mainly dependent on the solar part of the system. it is also noticed that a double-effect system is cost and performance competitive with the single -effect system. 2.3 desiccant cooling and open cycle system a desiccant cooling system is basically an open absorption cycle, utilizing water as the refrigerant in direct contact with air. the term 'open' indicates that the refrigerant is disposed from the system after supplying the cooling effect (ie. without air circulation), while the same amount of refrigerant is replaced in an open-ended loop as illustrated in fig. 6. add in another configuration, an open solid desiccant cycle could be operated in a circulation mode, where the room air is circulated and the ambient air is used only for regeneration (ie. does not enter the room). the main advantages of these systems over closed absorption systems are as follows (grossman, 2001): it is worth mentioning that the major merits of soliddesiccant cooling systems include the improved indoor air quality and simplicity of construction and maintenance. on the other hand, the major limitation of desiccant systems is the low cop (0.5-1.0) relative to the closed system, due to inefficient regeneration. also their performance may be influenced by variation in weather conditions. periodic replacement of the desiccant is also necessary owing to contamination by air pollutants. a desiccant wheel (honeycomb rotary bed impregnated with sorbent material such as silica gel, zeolites or carbon) may be used in the solid desiccant system to ensure the continuity of the dehumidification process (ie. adsorb the water vapour from the moist air passing through it). both solid and liquid desiccants can be used to remove latent heat (ie. dehumidification process), while the evaporative cooler removes the sensible heat. this is preferable to maintain comfortable humidity levels in the conditioned space. the disadvantage of the liquid desiccant (sorbent) may be the carryover potential of sorbent in the air, which may pollute the air and reduce the amount of available desiccant. liquid desiccant systems may employ packed towers, packed beds or a spray chamber (mongar, 1997). this type of system has a number of advantages over solid desiccants, including: a number of solar-driven air-conditioning systems (1) typical cop (2) heat source temp. [oc] (3) solar heat required [kw]* (4) type of collectors (5) typical collector cost [s/m2] (6) heat collected per day [kwh/m2] (7) collector area required [m2] (8) solar system cost [s]* (9) absorpti on chiller cost [s]* (10) total system cost [s]* singleeffect 0.7 85 1.43 flat-plate 110 1.53 7.48 1234 200 1434 double effect 1.2 130 0.83 flatplate/cpc 160 1.31 5.07 1216 175 1391 triple effect 1.7 220 0.95 evac-tube/ concentrating 400 1.05 4.49 2694 165 2859 * indicates figures per kw of cooling capacity table 2. comparison of single-, doubleand triple-effect absorption chillers for solar-powered air conditioning systems (grossman, 2002) • they operate at ambient temperature, and not at vacuum or an elevated pressure • heat and mass transfer between the air and desiccant takes place in direct contact • cooling and dehumidification of the conditioned air may be provided in variable quantities. 54 the journal of engineering research vol. 4, no.1 (2007) 48-63 using liquid desiccant have been constructed (wood, 1986; lenz et al., 1985). a novel-solar powered open cycle dehumidifier-evaporatorregenerator (der) system has been developed (gershon grossman, 2002) and in contrast with a closed-cycle absorption system of double or triple effect more this fase found to be suitable for use with relatively low temperature heat sources such as a flat plate solar collector. it produced both chilled water, and cold, dehumidified air in variable quantities. a novel airconditioning system using a liquid desiccant, in an open cycle, has also developed (armando et al., 1999). the system employed needle-impeller rotors to improve heat and mass transfer in the absorber and evaporator units (riffat, patent) as shown in fig. 7. the system has a substantially lower initial cost than a solid desiccant system (using desiccant wheel), due to the reduction in system volume. it also has the potential to compete with a conventional vapour compression system owing to its energy efficiency, low capital cost, and its compactness. 2.4 adsorption cooling systems the basic principle of operation of adsorption and absorption cycles is similar. they do however, differ in two respects (anyanwu, 2004). the first is in the nature of the sorbent. an adsorption cycle uses only solid adsorbent, and as a result is characteristically intermittent in operation. an absorption cycle on the other hand, may use either solid or liquid absorbent and can therefore operate intermittently or continuously. the second difference is the process of refrigerant accumulation on the adsorbent, since the refrigerant is trapped in the micro-pores of the solid adsorbent. the duration of the cooling cycle is also significantly longer for single bed adsorption systems. for the multi-bed system, the cycle time could be short. the main advantages of an adsorption system over existing absorption systems are as follows: on the other hand, the cop of the adsorption chillers is lower than that of absorption systems, and as they are more expensive, their commercial availability is still limited. this is in addition to less efficient heat transfer in solid adsorbent beds. generally, adsorption is a process in which fluid molecules are trapped on the surface and within the pores of a solid. the pore structure of interest for refrigeration applications is micro-pore (radii <15 å). there are two main types of solid adsorption process: physical and chemical adsorption (physisorption and chemisorption, respectively). physisorption occurs as a result of relatively weak van der waals' attraction forces binding the adsorbing molecules to the solid phase, and predominants at low temperatures (anyanwu, 2004). chemisorption is accompanied by a chemical reaction, which occurs when covalent or ionic bonds are formed between the adsorbing molecules and the solid substances. this is therefore characterized by comparatively high energies of adsorption. there are many adsorbent-adsorbate pairs used for airconditioning/refrigeration purposes, as illustrated in table 3 (sumathy et al., 2003) the commonly used adsorbents in air-conditioning are silica gel and zeolite, with water as an adsorbate (refrigerant). silica gel/water is the most common pairing particularly in solar driven system, due to its lower regeneration temperature, which is below 100 oc,. for a zeolite/water system this temperature is 170 oc (papadopoulos et al., 2003). fig. 6 desiccant cooling system heat supply from solar or waste heat evaporative cooler desiccant wheel exhaust air supply air 1 2 3 4 5 6 7 8 1 6 2 7 4 5 3 8 room h ea t e xc ha ng er figure 6. desiccant cooling system lower regeneration temperatures (60 – 80 oc for liquid desiccant compared to100 –150 oc for solid desiccant) (ashrae, 1997). suitability for dust removal by filtration greater mobility by pumping the sorbents lower pressure drop of air across the desiccant. dehumidification of greater quantities of air per unit mass than solid desiccant (eg. lithium chloride can absorb up to 1000% its dry mass weight of water (ashrae, 1997). • simple construction and can be built on a small scale • can utilize heat at lower temperatures • no corrosion problem • no need for solution pump • the system units are totally autonomous. in some cases for the absorption systems, precise temperature regulation is required, which may be difficult to achieve with solar heating (meunier, 1994). 55 the journal of engineering research vol. 4, no.1 (2007) 48-63 adsorption systems may be intermittent or continuous. the use of a single bed results in intermittent cycle operation due to changing between adsorption and desorption stages. a description of one-bed cycle operation is as follows: the refrigerant vapour, which has been evaporated in the evaporator, is adsorbed into a solid adsorbent bed, which acts as a sponge. when all the refrigerant vapour has left the evaporator, it can be recovered from the adsorbent by heating it in the adsorber/generator chamber. the refrigerant vapour is extracted and then cooled and condensed in the evaporator/condenser chamber. intermittent adsorption can be overcome by using multiple adsorbent beds with one in operation while the other is regenerated. critoph and tamainottelto (1997) have studied the performance of a solar adsorption refrigerator experimentally for three different configurations of glass cover of a flat plate solar collector. similarly, the performance of a solarpowered continuous (two bed) adsorption air-conditioning figure 7. desiccant air-conditioning open cycle and air evolution in temperature (t)-water content (x) graph (armando et al., 1999) x t 51 3 1 indoor space evaporator water evaporator water heat pipes 2 absorber a vapour generator heaterdesiccant solution heat exchanger adsorbent adsorbate heat of adsorption (kj kg -1) remarks activated alumina water 3000 water is applicable except for very low operating pressures zeolite (various grades) water 3300-4200 natural zeolites have lower values than synthetic zeolites ammonia 400-600 carbon dioxide 800-1000 methanol 2300-2600 silica gel methyl alcohol 1000-1500 not suitable above 200oc water 2800 used mostly for desiccant cooling charcoal c2h4 1000-1200 reacts at approximate 100 oc. ammonia and methanol are not compatible with copper at a high temperature ammonia 2000-2700 water 2300-2600 methanol 1800-2000 c2h5oh 1200-1400 table 3. absorbent/adsorbate pairs and heat of adsorption (sumathy et al. 2003) 56 the journal of engineering research vol. 4, no.1 (2007) 48-63 system driven by a flat plate solar collector has been investigated (yong and sumathy, 2004). collectors used in the water heating system, during winter, were utilized as the heat source to energize the air-conditioning system during summer. the working principle of the system is shown in fig. 8. in the desorption phase, adsorber a is connected to tank 1 to heat and desorb the refrigerant from the adsorbent. the adsorption phase is conducted in adsorber b, which is connected to tank 2, to adsorb the refrigerant vapour leaving the evaporator at low pressure. the condensed refrigerant leaves the condenser and flows to the evaporator. in order to increase the efficiency of the system and to attain continuous cooling, the two adsorbers are connected to each other after each desorption cycle, to recover heat. considerable effort has been made to improve the heat regeneration efficiency of the system. typical cycles are listed as follows (hai-ming lai, 1999): 1. heat recovery cycle (meunier, 1985; van, 1995): heat is recovered by an additional regenerative period: circulating a heat transfer fluid (htf) between the two adsorbers until they reach a common temperature. 2. thermal wave cycle (shelton, et al. 1989): heat is stored and regenerated using sharp temperature fronts (thermal wave) in the adsorbent beds. theoretically this would obtain more than 80% of heat recovery from the bed being cooled. 3. flow reversal cycle (dong, 1995; lai and li, 1996): in terms of the concept of thermal wave, internal heat regeneration is produced within one bed by forced unsteady-state operation achieved by periodically switching the flow direction of the htf. 4. convective thermal wave cycle (critoph,1994): this is different to the traditional means of heat transfer carried out by heat conduction described as above three cycles (1-3).the heat and mass transfer of this cycle are intensified by using forced convection between the refrigerant and adsorbent. the heat regeneration is performed via a gas-gas heat exchanger or inert packed beds. a detailed review of heat regeneration cycles of adsorption systems has been presented by (sumathy et al., 2003). a periodic reversal forced convective cycle has been suggested (lai and li., 1999) based on the combination of cycles 3 and 4 listed previously, and a mass recovery cycle has been found to be effective for low regenerating temperatures (pons et al., 1999; wang, 2001). a novel strategy in mass recovery has been proposed (akira et al., 2004) to improve cooling capacity. this is completely different from the conventional mass recovery cycle, in that the heating and cooling processes during the mass recovery process, do not exist. in this cycle, the adsorber kept in cooling process, while the desorber is continuously provided with hot water. this cycle is very applicable for solar heat source/low temperature waste. a brief description of the working principle of the two-bed system is shown by the diagrams in fig. 9. the cycle has six modes, which can be divided into two similar parts (ie. the first three modes are analogous to the next three modes), where the adsorption and desorption processes are switched. in the first mode the evaporator and heat exchanger (hex1) are connected via valve 2 (v2) in the adsorption-evaporation process, while the condenser and (hex2) are connected via v3 in desorption-condensation process. all other valves are closed. in the second mode, the desorbed refrigerant in hex2 moves to the adsorber (hex1) through v5 by the effect of differential operating pressure on the two heat exchangers, while all other valves are closed. the third mode is a relax mode, where all the valves are closed and the heat exchangers are switched (ie. hex1 is warmed up and hex2 is cooled down until they become almost equal pressures to the con and eva, respectively). the fourth mode is then started, where v1 and v4 are opened. this mode is similar to mode 1. actually, the only difference between this cycle and the conventional single stage cycle is v5 and the pipe connecting the heat exchangers (ie. the 2nd and 5th modes do not exist). the same investigators (alam et al., 2004) have also studied a four-bed mass recovery system, which provided improved performance. the cascading feature of the heating and cooling source developed different pressure levels, which in turn enhanced the refrigerant mass circulation. combination of commonly used solar collectors (flat plate, evacuated tube and solar air collectors) with sorption refrigeration/cooling systems can be evaluated by a collector efficiency-driving heat temperature diagram as shown in fig. 10 (henning, 2000). it can be seen that only vacuum tube collectors are applicable for a double effect absorption system, while flat plate or evacuated tube collectors can be used for adsorption and single effect absorption systems. it is clear from the diagram that the generation temperature for desiccant cooling is relatively low, and so a solar air-collector would be suitable for this system. adsorber b condenser evaporator hot water tank (1) cold water tank (2) adsorber a solar collector hot water cooling water refrigerant heat recovery water fig. 8 solar powered continuous (two beds) adsorption air -conditioning system (li yong and k. sumathy, 2003) figure 8. solar powered continuous (two beds) adsorption air-conditioning system (yong and sumathy, 2004) 57 the journal of engineering research vol. 4, no.1 (2007) 48-63 2.5. ejector cooling cycle an ejector-cooling cycle is one of the alternatives to a vapour-compression cycle, which retains the condenser and evaporator but achieves refrigerant compression by other means. the ejector compressor is considered as the heart of the ejector cooling cycle and determines the overall performance and cooling capacity. an ejector can be described as a device which uses the high pressure stream, which can be called 'primary fluid' or 'motive fluid', to drive a second fluid stream, which can be called 'secondary fluid' or 'suction fluid', by direct mixing. the resulting mixture can be discharged at a pressure that lies between the driving pressure and the suction pressure. a steam ejector refrigerator was first developed by blance and parsons around 1901 (gosney, 1982). to better understand how a typical ejector functions, a description of its operation is given in fig. 11 (chunnanod and aphornratana, 2003). normally, a steam ejector consists of four principle parts, which are the primary nozzle, mixing chamber, the ejector throat and the subsonic diffuser. the high-pressure primary stream (point p) starts to accelerate as it enters a convergent section of the nozzle and reaches the sonic level at the nozzle throat (i). as the primary stream expands through a divergent section of the nozzle, its speed is further increased and reaches supersonic speed at the exit plane, creating a low-pressure region (ii). this causes the primary stream (the expanded wave) to entrain the secondary fluid in the mixing chamber (s), where the secondary stream is accelerated and completely mixed with the primary stream at point (iii). if the mixed stream is still supersonic at (iii), a shock will occur in the throat (iv), resulting in a subsonic stream at a point determined by the intersection point of fanno and rayleigh lines (alexis and rogdakis, 2002). the mixed stream is then brought to near zero velocity (ie. stagnation condition) as it passes the subsonic diffuser (b). the primary refrigerant of high pressure and temperature (point 2, see fig. 12-a & 12-b), evolved from the generator, expands through the primary nozzle of the ejector. this expansion creates a low-pressure region, as explained above, at the primary nozzle exit plane (3) that results in an entrainment of the secondary fluid from the evaporator. the mixed stream is discharged via the diffuser to the condenser (4). after condensation, the refrigerant is pumped back to the boiler, whilst the remainder is expanded through the throttling valve to the evaporator (6). the ejector assists vaporisation of the evaporator refrigerant at low temperature by creating the low-pressure area at the primary nozzle exit. the main advantages of the ejector refrigeration system are its few moving parts (ie. no associated noise or vibration compared to a mechanical compressor), long service life, simple construction, low initial cost and no chemical corrosion. in addition, water can be used as a working fluid. the only limitations of this system are its low cop and cooling capacity. the first solar-driven ejector air-conditioning system was demonstrated by sokolov and hershgal (1993). refrigerant r-114, which is now banned an environmental grounds, was used in this system. it was shown that for moderate climates, this system could produce hot water, air-conditioning and space heating. the later could be accomplished by passing the solar heat from the collector to the evaporator. more recent work has been carried out using other refrigerants. better performance was obtained using r-141b (huang et.al., 1998), however the overall cop did not exceed 0.25, since the ejector cop= 0.5 and the solar collector efficiency = 0.5. arbel and m. sokolov (2004) replaced r-114 with r-142b as a working fluid. the results indicated that better performance was achieved with r-142b. particular care must be given when selecting the solar collector and type of refrigerant. using a higher performance solar collector would increase the generator temperature and consequently the system's efficiency, however this might also adversely affect the properties and chemical stability of the refrigerant. fig. 9-a ptx diagram for mass recovery cycle tads-end tads-beginning tdes-end tdes-beginning qmax qmin fig. 9-b schematic of mass recovery adsorption refrigeration cycle, second mode (k. alam, 2004) chilled water in/out temperature chilled water in/out concentration w at er v ap or p re ss ur e figure 9-a. ptx diagram for mass recovery cycle figure 9-b. schematic of mass recovery absorption refrig4eration cycle, second made (alam, 2004) 58 the journal of engineering research vol. 4, no.1 (2007) 48-63 10 0.8 0.6 0.4 0.2 25 50 75 100 125 150 175 double effect absorption single effect absorption disiccant cooling absorption generator temperature (oc) figure 10. collector efficiency versus driving heat temperature for various sorption refrigeration systems (henning, 2000) mixing chamber throat subsonic diffuser primary nozzle primary flow secondary flow expansion wave v el oc ity p re ss ur e p i s ii iii iv b sonic velocity distance along axis figure 11. typical steam ejector and flow characteristic in the ejector (chunnanod and aphornratana, 2003) pressure isentropic lines 3 4 2 1 5 6 3’ enthalpy saturation line fig. 12 -a p-h diagram of ejector cycle 3’ fig. 12 -b schematic diagram of ejector cycle added head boiler 2 ejector 3’ 3 4 condensor rejected heat refrigeration evaporator 6 5 figure 12-a. p-h diagram of ejector cycle figure 12-b. schematic diagram of ejector cycle 1 59 the journal of engineering research vol. 4, no.1 (2007) 48-63 a solar powered passive ejector cooling system, which contains no moving parts, has been successfully tested in a uk climate (nguyen et al., 2001). some active components were used in the cooling and hot water circuits. a gravity pump was used for refrigerant (water) transfer from the condenser to the boiler and the cop of this design reached 0.32. research on the solar powered ejector cooling system is still on going. the cop of the system is expected exceed 0.6 when water is used as a refrigerant and a concentrating parabolic trough solar collector is utilized. generally speaking, the performance of an ejector cooling system is influenced by the working fluid selected, the thermodynamic cycle, and the design of the ejector, especially the compression ratio pdischarge/pentrainment) (korres et al., 2002). 2.6 thermoelectric cooling cycle refrigeration can also be achieved using the thermoelectric effect, which converts electrical energy into a temperature gradient. this phenomenon was discovered by peltier (1834). peltier refrigeration occurs when a direct current (dc) is passed from a negative to positive type semiconductor material. the temperature tc of the interconnecting conductor decreases and heat is absorbed from the environment (cooling effect). this cooling effect occurs, as shown in fig. 13-a, when electrons pass from a low energy level in the positive type material through the interconnecting conductor to a higher energy level in the n-type material (riffat, 2003). the potential uses of this system range from cooling of the electronic components, to domestic refrigerators and air conditioners for cooling/heating a room space. two types of multi-couple thermoelectric devices that are commercially available are shown in fig. 13-b. these usually consist of a series of pand n-type bulk semiconductor thermo-elements connected electrically by copper strips and sandwiched between two electrically insulating, and thermally conducting ceramics plates. thermoelectric devices offer several distinct advantages over other technologies (riffat and ma, 2003): a thermoelectric solar-driven air conditioner would however be expensive in terms of its capital and operating cost) as it uses a large amount of electricity. its low cop (0.38-0.45) (riffat and qiu, 2004) is an additional factor limiting its application for domestic cooling. thermoelectric air-conditioners are however very suitable for cooling of small enclosures, such as cars where the power consumption is low. mei et al. (1993) studied a solar-assisted automobile thermoelectric air-conditioner and other studies related to solar thermoelectric refrigerators have been reported. (dai et al., 2003). 3. coefficient of performance normally, refrigerator/air-conditioner and heat-pump performance is defined by means of coefficient of performance. the coefficient of performance for cooling applications is defined as the amount of cooling delivered by a machine divided by the amount of energy input required to produce the cooling (ashrae, 1997). the best cop is given by a carnot cycle, which is operating between given temperature conditions. such a theoretical reversible cycle uses a wet vapour as a working substance, since a superheated vapour cannot fulfill the necessary requirement of the carnot cycle (ie. constant pressure heat supply and heat rejection are made at constant temperature) figs. 14 -a and 14-b). the vapour compression cooling cycle is the real (irreversible) cycle based on the carnot theoretical cycle (refer to fig. 2). it is worth mentioning that the low cop of the solar-powered cooling cycles compared with the vapour compression cycle does not preclude these cycles, as different power inputs and consequently different definitions of cop are appropriate. for example, for a vapour-compression cycle, the input power is determined by the amount of mechanical work required to compress the vapour: in a heat-driven cooling cycle, the cop is determined by the amount of heat required (in the generator) and the pump work: to make a fair comparison of the cop of two different systems with the same cooling capacity, the energy input needs to be estimated on the same basis. electricity is the main power driving a conventional air-conditioning system, whereas thermal power and parasitic electric power are used to drive solar-powered systems. therefore, to compare different systems in terms of thermal and electrical energy inputs, it is reasonable to convert the electric • almost maintenance–free due to the absence of the moving parts • environmentally friendly, ie. contain no cfcs. • long lifetime; life testing has shown their capability exceeds 100’000 h of steady state operation. • they are not position dependent (portable). • precise temperature control to within ± 0.1 oc can be maintained. • changing the polarity of the dc power supply causes heat to be pumped in the opposite direction ie. a cooler can become a heater. • they have a wide range of applications, e.g., they could be used as cooler/heaters, power generators or thermal energy sensors. • a thermoelectric solar-driven air-conditioner can be powered directly by a dc electric source, such as a pv cell. compressor evaportor q q cop.ie = compressor evaportor w q cop.ie = pumpgenerator evaporator wq q cop.ie + = 60 the journal of engineering research vol. 4, no.1 (2007) 48-63 fig. 13-b schematic diagram of multicouple thermoelectric m (a) with large inter-thermo-element separation (b) with small inter-thermo-element separation fig. 13-a schematic of thermo-electric module operation (cooling mode), (s.b. riffat, xiaoli ma, 2002). v heat dissipated to heat sink qd dc source p type n type ½-j heat ½-j heat conducted heat conductor tc (cold end) heat absorbed from environment (cooling) qin peltier cooling peltier heating + th (hot end) figure 13-a. schematic of thermo-electricmodule operation figure 13-b. schematic diagram of multicouple (cooling mode), riffat and ma, 2003) thermoelectric modulus (a) with large inter-thermo-element separation (b) with small inter-thermo-element separation (a) (b) electrical insulating layer aluminium conducting strip ceramic plate thermoelements copper conducting strip compressor condenser evaporator 1 2 3 engine 4 3 2 1 4 s2=s1 s4=s3 t4=t3 t2=t3 t s w q1 figure 14-a. carnot cycle, t-s diagram figure 14-b. carnot cycle schematic diagram figure 15. schematic diagram of a solar-powered combined ejector-vapour compression ccle (sun, 1997) 61 the journal of engineering research vol. 4, no.1 (2007) 48-63 power back into heat to provide an estimation of the actual cop, losses in electric power consumption and calculate the energy saving potential of solar-driven systems on the basis of heat input. some research effort has been directed to combine two or more refrigeration cycles to improve the coefficient of performance (cop) of the system. sozen, et al. (2003) investigated performance prediction of a solar driven ejector-absorption refrigeration system (ears) using fuzzy logic. the maximum calculated cop of the system was 0.62, 0.69, and 0.79 at evaporator temperatures of -10, 0, and +10 c, respectively with a deviation of 1.2% or less. adopted direction of a combination of two cooling cycles might be an effective solution to overcome the limitations of many cooling cycles. this combination phenomenon has been applied to a mechanical vapour compression cycle combined with a solar powered ejector cycle (sun, 1997). the combined cycle, as shown in fig.15, brings together the advantages and eliminates the disadvantages of the two conventional cycles. the thermodynamic analysis shows that this combined cycle offers the following advantages: elsafty and al-daini (2001) has presented an economic comparison between a solar-powered vapour absorption system (for both a single and double effect system) and a conventional vapour compression system, (elsafty and daini, 2001). the cost analysis covered the initial and operating costs of each system and considered the interrelationship between economic and thermodynamic aspects, such as the dependence of operating cost on the surrounding climate conditions. the results showed that the double-effect vapour absorption system was the preferred option for its minimum present worth value, as well as the equivalent annual cost. mistuhiro et al., (2001) discussed the feasibility of a vapour compression/absorption hybrid refrigeration cycle for energy saving and utilization of waste heat. the cycle employed propane as a natural refrigerant and refrigeration oil as an absorbent. the experimental results were in close agreement with the mathematical calculation, which showed that the hybrid cycle using waste heat (fig. 16) had the potential to achieve a higher performance (cop) than the vapour compression cycle. a novel solar-powered adsorption-ejection refrigeration system has been studied (li, et al. 2002) in order to overcome the intermittence of adsorption refrigeration. the ejection cycle operates during the day, while the adsorption cycle operates at night time. 4. conclusions the following conclusions can be drawn from this study: • improvement of system efficiency by more than 50% over the vapour compression cycle. • environment-friendly system, as water is used in the ejector cycle and hfc-134a in the vapour compression cycle, ie. odp = 0. also the energy saving of more than 50% means that less electricity is used in the system, and consequently less co2 is released to the atmosphere by the combustion of fossil fuels. • the main additional cost for the combined cycle is in the solar panels. this would be offset by the saving in the electrical energy. condenser generator refrigerent heater rich solution weak solution expansion valve evaporator compressor absorber solution pump solution heat exchanger figure 16. schematic diagram of a vapour compression/absorption hybrid cycle. the absortion unit is connected to the compressor exit in series (mistuhiro et al., 2001) • several technologies for solar powered airconditioning have been described. in comparison with conventional vapour-compression systems, these systems usually have a higher initial cost, and it is this economic barrier, rather than technical difficulties, which has limited the emergence of these technologies. as the collectors are normally the most expensive component of a solar-operated system, it would be cost effective to use them for more than one purpose, eg. collectors used in a water-heating system during the winter, could be utilized to power an air-conditioning system during the summer. • it is more realistic to base a comparison of the cop of the vapour compression system with a solar-driven systems on primary energy input rather than shaft work input. the reason for the high values of cop quoted for vapour compression cycles is then clear, as no account for the heat input in power generation has been made for typical analyses of vapour compression cycles. • the use of combined cooling cycles could be an effective solution to overcome the limitations of many cooling cycles, e.g., the low cop. the high capital cost of solar systems would fall as mass production techniques evolve in line with market penetration. 62 the journal of engineering research vol. 4, no.1 (2007) 48-63 references akira, akahira, alam k.c.a., yoshinori hamamoto, atsushi akisawa, takao kashiwagi., 2004, "mass recovery adsorption refrigeration cycle-improving cooling capacity," int. j. of refrigeration, vol. 27(3), pp. 225-234. alam, k.c.a., akahira, a., hamamoto, y., akisawa, a. and kashiwagi, t., 2004, "a four-bed mass recovery adsorption refrigeration cycle driven by low temperature waste/renewable heat source," renewable energy, vol. 29(9), pp. 1461-1475. anis, w. r., metwally, h. h. b. and el-samahy, a. m., 1997, "microprocessor-based overcurrent protection for pv powered air-conditioning system," energy conversion and management, vol. 38(1), pp. 29-37. anyanwu, e. e., 2004, "review of solid adsorption solar refrigeration ii: an overview of the principle and theory," energy conversion and management, vol. 45(7-8), pp. 1279-1295. arbel, a. m., sokolov, 2004, "revisiting solar-powered ejector air conditioner-the greener the better," solar energy, vol. 77(1), pp. 57-66 ashrae handbook fundamentals, 1997. curran, h., 1992, “mechanical systems and componentns for active solar systems,” mit press, g. lof. critoph, r. e. and tamainot-telto, z., 1997, "solar sorption refrigerator," renewable energy, vol. 12(4), pp. 409-417. critoph, r. e., 1994, "forced convection enhancement of adsorption cycles heat recovery system and chp," vol. 14(4) pp. 343-350. dai, y. j., wang, r. z. and ni, l., 2003, "experimental investigation and analysis on a thermoelectric refrigerator driven by solar cells," solar energy materials and solar cells, vol. 77(4), pp. 377-391. chad, b., dargon, steven, p., leight and charles, e. dargon., 1995, "application guide for absorption cooling/refrigeration using recovered heat," ashrae. dong, y. m., lai, h. m., 1995, "theoretical studies on flow reversal cycle process of ca cl2 /ch3oh chemical heat pump," chemical engineering (chinese) vol. 23(1), pp. 57-61. elsafty, a. and al-daini, a. j., 2002, "economical comparison study between a solar-powered vapour absorption air-conditioning system and a vapour compression system in the middle east," renewable energy, vol. 25(4), pp. 569-583. fukuta m., yanagisawa t., iwata h. and tada, k., 2002, "performance of compression/absorption hybrid refrigeration cycle with propane/mineral oil combination," int. j. of refrigeration, vol. 25(7), pp. 907-915. gosney, w. b., 1982, "principle of refrigeration," cambridge university press. gershon grossman., 2002, "solar powered system for cooling, dehumidification and air-conditioning," solar energy, vol. 72(1), pp. 53-62. hai-ming lai., 2000, “an anhanced absorption cycle operated by periodic reversal forced convection,” applied thermal engineering, vol.20(7), pp. 595617. henning m. h., 2000, "air-conditioning with solar energy," servitec meeting, barcelona. henry, c., 1992, "mechanical systems and components from active solar systems," mit press,g.lof. jubran, b. a., al-hinai, h. a., zurigat, y. h., and alsalti, s., 2003, "feasibility of using various photovoltaic systems for window-type airconditioning units under hot-arid climates," renewable energy, vol. 28(10), pp. 1545-1553. korres, c. j., papaioannou, a. t., lygerou, v. and koumoutsos, n.g., 2002, "solar cooling by thermal compression: the dependence of the jet thermal compressor efficiency on the compression ratio," energy, vol. 27(8), pp. 795-805. kreider, j. f. and kreith, f., 1981, "solar system for space cooling," in solar energy handbook, mcgraw-hill, new york, chapter 15. lai, h. m. and li, c. y., 1996, "application of periodic reversal flow reactors to chemical heat pump systemsbased on solid/vapor non-catalytic reaction," chem. eng. sci vol. 51(11), 2971-2976. lenz, t. g., löf, g.o.g., lyer, r. and wenger, j., 1985, "open cycle lithium chloride cooling system," final reports on doe contracts for 1982/83, 1983/84, solar energy application laboratory, colorado state university. li, c.h., wang, r.z. and lu, y.z., 2002, "investigation of a novel combined cycle of solar powered adsorption-ejection refrigeration system," renewable energy, pp. 26:611-22. li, z. f. and sumathy, k., 2001, "experimental studies on a solar powered air conditioning system with partitioned hot water storage tank," solar energy, vol. 71(5), pp. 285-297. little, a. d., 1993, "the market for solar photovoltaic (pv) technology," technical report tr-102290, inc,epri, palo alto, ca. mei, v.c., chen, f.c, mathiprakasam, b. and heenan, p., 1993, "study of solar-assisted thermoelectric technology for automobile air conditioning," j. of solar energy engineering, transactions of the asme, vol. 115(4), pp. 200-205. meunier, f., 1985, "second law analysis of a solid adsorption heat pumpon a reversible cascade cycles: application to the zeolite water pair," heat recovery systems and chp 5, pp. 131 -141. • combined cooling cycles could also be used for energy saving and reduced co2 emission. • problems of the gwp and odp of many refrigerants used in vapour-compression systems could be eliminated/reduced by the use of solarpowered or combined cooling cycles. 63 the journal of engineering research vol. 4, no.1 (2007) 48-63 meunier, f., 1994, " sorption solar cooling," renewable energy, vol. 5(1-4), pp. 422-429. mongar, a. l., 1997, "liquid desiccant technology state of the art," in: proc. desiccant cooling seminar, cibse. morgan, d. b. and deo prasad., 2001, "modelling of phtovoltaic heat recovery system and its role in a design decision support tool for building professionals,". nguyen, v. m., riffat, s. b. and doherty, p. s., 2001, "development of a solar-powered passive ejector cooling system," applied thermal engineering, vol. 21(2), pp. 157-168. ohnishi, m., takeoka, a., nakano, s. and kuwano, y., 1995, "advanced photovoltaic technologies and residential applications," renewable energy, vol. 6(3), pp. 275-282. oliveira, a. c., afonso, c. f., riffat, s. b. and doherty, p. s., 2000, "thermal performance of a novel airconditioning system using a liquid desiccant," applied thermal engineering, vol. 20(13), pp. 12131223. papadopoulos, a.m., oxizidis, s. and kyriakis, n., 2003, "perspectives of solar cooling in view of the developments in the air-conditioning sector," renewable and sustainable energy reviews, vol. 7(5), pp. 419-438 pons, m, poyelle f., 1999, "adsorption mechanics with advantaged cycles for heat pumping or cooling applications.," int j. refrigeration, vol. 22, pp. 2737. riffat, s. b. and qiu, g., 2004, "comparative investigation of thermoelectric air-conditioners versus vapour compression and absorption airconditioners,". riffat, s. b., international patent pct/gb96/02704, university of nottingham, uk. riffat, s. b., ma, x., 2003, "thermoelectrics: a review of present and potential applications," applied thermal engineering, vol. 23(8), pp. 913-935. riffat, s. b., wong, c. w. and james, s., 1999, "absorption recompressing chiller using a new working fluid,". shelton, s. v., wepfer, w. j. and miles, d. j., 1989, "squar wave analysis of the solid-vapor adsorption heat pump," heat recovery systes and chp 9(3), pp. 233 247. sokolov, m. and hershgal, d., 1993, solar-powered compression enhanced ejector air conditioner," solar energy, vol. 51, pp. 183-194. sozen, a., mustafa, k., akcayol, a. and ozalp m., 2004, "performance prediction of a solar driven ejectorabsorption cycle using fuzzy logic," renewable energy, vol. 29(1), pp. 53-71. sumathy, k., yeung, k.h. and yong, li., 2003, "technology development in the solar adsorption refrigeration systems," progress in energy and combustion science, vol. 29(4), pp. 301-327. sun, d.w., 1997, "solar powered combined ejectorvapour compression cycle for air conditioning and refrigeration," energy conversion and management, vol. 38(5), pp. 479-491. us epa http://www.epa.gov/ van benthem, g.h.w., cacciola, g. and restuccia, g., 1995, "regenerative adsorption heat pumps: optimization of the design," heat recovery system and chp 15(6), pp. 531 -544. wagdy, r., anis., hamid, metwally and abd elshkour m. el-samamy., 1997, “microprocessor-based overcurrent protection for pv powered airconditioning systems,” energy conversion and management, vol.(38(1), pp. 29-37. wang, r. z., 2001, "performance improvement of adsorption cooling by heat and mass recovery operation," int j. refrigeration, vol. 14, pp. 602-11. wood, b. d., 1986, "open cycle absorption solar cooling," final report no. erc-r86020d, college of engineering and applied sciences, arizona state university. yong, l. and sumathy, k., 2004, "modeling and simulation of a solar powered two bed adsorption air-conditioning system," energy conversion and management, vol. 45(17), pp. 2761-2775. mech81003.qxd the journal of engineering research vol. 2, no. 1 (2005) 69-76 1. introduction worker productivity and satisfaction improvements are a major concern in industry, especially for repetitive tasks such as short-cycled assembly. these tasks are considered boring, monotonous, fatiguing and de-motivating. this, in turn, results in reduced worker productivity, poor work quality, higher absenteeism and causes detrimental effects on worker physical and mental well being shikdar and das, (1995). improving worker productivity in such tasks is, therefore, a major challenge, especially for the management. ergonomics is concerned with making the workplace as efficient, safe and comfortable as possible. effective _______________________________________ *corresponding authors e-mail: ashraf1@squ.edu.om application of ergonomics in work system design must achieve a balance between worker characteristics and task demands. this can enhance worker productivity, provide worker safety as well as physical and mental well-being and job satisfaction. many research studies have shown the positive effects of applying ergonomic principles in workplace design, machine and tool design, and environment and facilities design (hasselquist, 1981; schnauber, 1986; ryan, 1989; das, 1997; resnik and zanotti, 1997; burri and helander, 1991; shikdar and das, 1995; das and sengupta, 1996; das and shikdar, 1999). research studies in ergonomics have produced data and guidelines for industrial applications. the features of ergonomic design of machines, workstations, and facilities are well known (grandjean, 1988; konz, 1995; das and grady, 1983; salvendy, 1987; melamed et al. 1989; sanders and mccormick, 1992; wilson and corlett, 1992; operator performance and satisfaction in an ergonomically designed assembly workstation a.a. shikdar* and m.a. al-hadhrami department of mechanical and industrial engineering, sultan qaboos university, p.o. box 33, al-khod 123, muscat, sultanate of oman received 8 august 2003; accepted 2 october 2004 abstract: the objective of this study was to investigate the effects on operator performance and satisfaction of an ergonomically designed workstation for performing a repetitive industrial assembly task. experiments were conducted in a company with industrial workers using existing and newly developed workstations. operator performance on the ergonomically designed workstation was 27% higher compared to the existing non-ergonomically designed workstation. worker satisfaction score was also improved by 41% in the ergonomically designed workstation condition. the new workstation for a repetitive assembly task had highly significant positive effect on worker performance and satisfaction. special features of the ergonomically designed assembly workstation were an adjustable and adequate worktable, an adjustable and ergonomically designed chair, ergonomically designed hand tools and a systematic layout of the workstation components. keywords: cycle time, operator performance and satisfaction, ergonomically designed workstation, assembly task ¢¢üü∏∏îîàà°°ùÿÿgggg” å«m .™fé°üÿg ‘ iqôµàÿg ™«ªéàdg ∫éªyc’ »eƒfƒzqeg º«ª°üj hp πªy á£fi ωg~îà°séh ∫骩dg é°vqh agog ‘ ägôkdƒÿg øy åëñdg ‘ á°sgq~dg iòg ±~¡j : .πª©dg ¢ùøæh ¿ƒeƒ≤j ø‡ ∫éªy ωg~îà°séh á°sgq~dg √ò¡d √qé«àng ” …òdg ™æ°üÿg ‘ »eƒfƒzqh’g º«ª°üàdg hp i~j~÷g h á«dé◊g πª©dg á£fi ωg~îà°séh üqééàdg πªy ¿é«ñà°s’g èféàf äô¡xcg éªc .á«dé◊g πª©dg á£ëã áfqé≤e i~j~÷g πª©dg á£fi ωg~îà°séh %27 áñ°ùæh ∫骩dg i~d á«léàf’g ‘ kgqƒ£j èféàædg äô¡xg üqééàdg ∫ón øe ‘ »hééj’g ôkc’g é¡d ¿éc qôµàÿg ™«ªéà∏d áe~îà°ùÿg »eƒfƒzqh’g º«ª°üàdg hp i~j~÷g πª©dg á£fi ¿g .i~j~÷g πª©dg á£fi ωg~îà°séh %41 áñ°ùæh äogr ∫骩dg é°vq ¿éh º«ª°üàh ájh~j äghocg ,á«eƒfƒzqhg ägõ«ã »°sôc .πª©∏d á«aéc ámé°ùe oƒlhh ádhé£dg ´éøjqg §ñ°v ƒg i~j~÷g πª©dg á£fi õ«á ée ºgcg .∫骩dg i~d é°vôdgh á«léàf’g ioéjr .πª©dg á£fi agõlc’ º¶æe ™jrƒjh »eƒfƒzqhcg gáá««mmééààøøÿÿgg ääggooôôøøÿÿ.™«ªéàdg πªy ,»eƒfƒzqhg º«ª°üàh πªy á£fi ,…oéy º«ª°üàh πªy á£fi ,πeé©dg é°vqh agog ,iqh~dg øer : »eƒfƒzqhc’g º«ª°üàdg hp ™«ªéàdg á£fi ωg~îà°sg ‘ πeé©dg aé°vqh agocg 70 the journal of engineering research vol. 2, no. 1 (2005) 69-76 mcleod, 1995). however, there is still a low level of acceptance and limited application of ergonomics in industry, especially in developing countries. the main concern of work system design is the improvement of machines and tools. inadequate or no consideration is given to the design of the work system as a whole. therefore, poorly designed systems are a common place in industry (konz, 1995; das, 1987). neglect of ergonomic principles brings inefficiency and pain to the workforce. an ergonomically deficient workplace can cause physical and emotional stress, low productivity and poor quality of work (ayoub, 1990a, 1990b). a workstation should be laid out such that it minimizes the working area so that while carrying out the operations the worker uses shorter motions and expends less energy. this will reduce fatigue. the concept of workspace design and the application of anthropometric data had been reviewed by das and grady (1983). it was established that an adjustable chair was highly desirable for the workplace as well as a standard size workbench. however, the standard height of the workbench cannot be defined without the anthropometric data of the user population. it was believed that application of ergonomics in the design of repetitive assembly tasks, including redesigning workstations and tasks, would not only improve worker safety and work quality, but it would also reduce cycle time and thus improve worker productivity significantly. an earlier study conducted in the laboratory with a simulated repetitive manufacturing task showed significant improvement in operator performance shikdar et al. (1997). however, this study did not consider a fully adjustable workstation and a real life task situation. the objective of this research was to study the effects on operator performance and satisfaction of an ergonomically designed workstation for a repetitive industrial assembly task. 2. methodology the methodology adopted in the conduct of this study was to select a representative real life industrial assembly task, study the workstation thoroughly and task performance on the workstation, redesign the workstation ergonomically and conduct experiments to assess the effects on operator performance (measured in terms of work cycle time) and satisfaction. 2.1. assembly task the selected task was an assembly of an electrical switch (molded case circuit breaker, mccb) that consisted of 14 parts. usually, simulated tasks chosen for research purpose does not represent real life industrial tasks. manual assembly of switches is a common task in the electrical industry. the selected assembly task was a highly repetitive task. it was performed on workstations that were not designed ergonomically. also, the task method was not designed following ergonomic principles. the assembly task involved picking up the base and cover from the incoming bins, assembling all the inside parts in the base, putting the cover on, tightening the assembly using a power screwdriver and placing it in the outgoing bin. considerable maneuvering and correcting motions were required for the task performance. the assembly steps were modified in the new design by incorporating motion study and ergonomic principles. 2.2. participants the participants were ten assembly operators from a local electrical manufacturing company. they were selected based on their experience in the selected and similar tasks, adequate training in the selected task, education and commitment to complete the study. all the participants were omani and indian male and their average age was 25.2 yrs, with a standard deviation of 6.18 yrs. the mean stature of the sample was 1679 mm with a standard deviation of 55.47 mm. education of the participants ranged from primary to trade diplomas. the average experience of the participants in this and similar tasks was 2.5 yrs. this was considered adequate for this study. the participants had full support from the management. 2.3. existing workstation an existing workstation was provided by the company and brought to our ergonomics laboratory. the workstation was thoroughly analyzed to identify ergonomic deficiencies with respect to work height, work areas, seating, posture, clearances, hand tools and layout. in general, it was found that no ergonomic consideration was given to the existing workstation. the task was performed in a sitting posture using a chair which did not have an adjustable back support, tilt mechanism and appropriate cushioning. the table height was too low and fixed (h=740 mm) with inadequate work area for laying out components. the operators were unable to maintain a natural work posture. working under such conditions for prolonged period could lead to shoulder stress and back pain (grandjean, 1988). the hand tools were poorly designed and not suitable for operator's ease of use and comfort. these were not ergonomically designed, including no handle for the file and an inadequate handle for the screwdriver. environment was not given adequate consideration as noise level was high and lighting was poor. the layout of the workplace did not consider any systematic guidelines. the components were placed on the worktable without considering functional use and normal and maximum work areas. the incoming bin was placed on a side table and outgoing bin was placed on the floor on the same side. figure 1 shows the layout of the existing workstation with the normal and maximum work areas superimposed. it can be seen that the bins were laid outside the maximum reach area. 2.4. ergonomically designed workstation the workstation was then redesigned using ergonomic 71 the journal of engineering research vol. 2, no. 1 (2005) 69-76 12 8 16 19 6 15 3 5 18 21 17 4 2 9 7 10 11 1 20 legend: 1. incoming bin (600x400); 2. terminal lever bin (320x205); 3. terminal clamp bin (140x110); 4. 3-pole body bin; 5. handle base bin; 6. handle springs, long screws, cover parts bin; 7. handles bin; 8. tripping lever bin; 9. tripping lever springs bin; 10. chutes; 11. covers bin; 12. outgoing bin; 13. maximum reach; 14. normal reach; 15. grease container; 16. assembly area; 17. fixture; 18. tool area; 19. assembly table; 20. side table; 21. power screwdriver. figure 1. plan view of the existing workstation with normal and maximum reach superimposed. all dimensions are in mm. figure 2. a conceptual ergomas model of the workstation 72 the journal of engineering research vol. 2, no. 1 (2005) 69-76 principles and data. a conceptual design was developed using ergonomic layout and optimization of manufacturing systems (ergomas) software package. the software is capable of developing workstation designs using a population anthropometry. workstation components could be selected from a wide range of products in the database. the conceptual model is shown in fig. 2. the model used 90 percentile anthropometric dimensions of the us military male population as local anthropometric data is not available and it closely matches the dimensions of the user population sample in the company (ergomas, 2001). considering the nature of the complexity of the assembly task it was decided to provide the worker with a fully adjustable ergonomic chair and an adjustable table so that the work could be performed in a posture that relieved the operator from unnecessary motions and fatigue. a footrest was provided for the worker. the existing hand tools were either modified or replaced with ergonomically designed ones. the workplace layout was made according to recommended dimensions for table height, seat height, thigh clearance, and reach envelopes of normal and maximum work areas for a male population al-haboubi (1992). however, the worker had the flexibility to adjust the table height or the chair to his comfort. figure 3 shows some of the specific features of the ergonomically designed workstation. an improved work method was developed for the task performance. 2.5. the experiments a methods-time measurement (mtm) analysis was conducted to establish the cycle time for the performance of the task on both workstations. they varied considerably. experiments were conducted in the company using existing non-ergonomically designed workstation and new ergonomically designed workstation. data on ten cycle times per participant (100 cycle times) were collected at random times using the existing non-ergonomic workstation condition before the intervention. the ergonomically designed workstation was taken to the company and installed on the shopfloor. participants were trained on the new workstation for two weeks and then allowed them to familiarize with the new method. the second set of experiments were scheduled three months later. ten cycle times (in minutes) of each participant to assemble the product were collected under the new condition. a total of 200 cycle times were collected in two experimental conditions (100 cycle times for each condition). 2.6. satisfaction questionnaire satisfaction questionnaire was conducted at the end of both experimental conditions in order to measure worker satisfaction in using the existing and the ergonomically designed workstations. it consisted of 23 questions in three areas: satisfaction with regard to workstation, satisfaction with regard to task method, and satisfaction with regard to health attributes. the human factor satisfaction questionnaire was modified to suit this particular study carlopio, (1991). the specific questions measured actual worker perception in the attributes using a likert-type scale of 1 to 5, with 1 being very dissatisfied to 5 very satisfied. an example of satisfaction question with regard to workstation was 'how satisfied are you with the work space on the table?'. the data were summarized and analyzed using statistical package for social sciences (spss) software. the analyses were conducted with the cycle times and worker satisfaction scores. 3. analysis of results 3.1. operator performance operator performance was measured in terms of work cycle time, converted to operator performance. a work cycle is defined as the time taken by an operator to complete one unit of the product. the mean cycle times of task performance in the two experimental conditions are presented in fig. 4. the mean cycle time under the existing workstation design was 13.38 min. with a standard deviation of 5.23 min. in contrast, the mean cycle time under the ergonomically designed workstation condition was 9.73 min with a standard deviation of 3.79 min. an analysis of variance (anova) on the cycle time data showed that the difference between the two conditions was highly significant (f=334.72, p<0.01) in terms of cycle time (table 1). a comparison between condition 2 (ergonomically designed workstation) and condition 1 (existing non-ergonomically designed workstation) showed that the mean cycle time of the participants working on the ergonomically designed workstation condition was significantly less compared to the mean cycle time of the participants working on the non-ergonomic workstation condition. the operator performance of the ergonomically design workstation condition (condition 2) was 27% higher compared to worker performance in the nonergonomically designed system. the higher worker performance was due to incorporating ergonomic principles and data in the design of the workstation and the task. participants consistently performed better in the ergonomically designed workstation. figure 5 shows the performance of participants on the two workstations. major changes that were incorporated in the redesigned workstation and the task were: adequate table surface area with height adjustment mechanism, fully ergonomically designed chair with a comfortable seat pan, height adjustment, adjustable armrests, and an adjustable back rest with tilt mechanism, workspace layout in the normal and maximum work areas, ergonomically designed hand tools, and an improved work method. since the worktable was adjustable, a flexible posture in the task performance was possible by the operators. the ergonomic changes incorporated to the design of the repetitive assembly task and the workplace made the work more comfortable, less fatiguing and more efficient. this is evident from the results of the satisfaction questions in section 3.2. an example of these questions was ' how satisfied are you 73 the journal of engineering research vol. 2, no. 1 (2005) 69-76 15 10 18 9 8 7 14 19 16 6 5 4 3 17 1 2 11 legend: 1. incoming bin (600x400); 2. arc chutes bin (140x110); 3. terminal clamp bin; 4. tripping lever bin; 5. cover parts bin; 6. handle springs and long screws bin; 7. handle and handle base bin; 8. line terminal bin; 9. 3-pole body bin; 10. outgoing bin (600x400); 11. assembly table; 12. maximum reach; 13. normal reach; 14. fixture; 15. assembly area; 16. grease container; 17. tool area 1; 18. tool area 2; 19. power screwdriver. figure 3. plan view of the ergonomically designed workstation. all dimensions are in mm. 0 4 8 12 16 existing condition ergonom ic condition m ea n c yc le t im e (m in u te s figure 4. mean cycle time for two conditions (minutes) source df ms f pr>f* r2** model 20 1559.33 783.08 0.01 0.99 conditions 1 666.53 334.72 subjects 9 397.71 199.73 conditions x subjects 9 21.65 10.87 errors 180 1.99 total 200 df – degrees of freedom; ms – mean square; f – f value; *pr>f = probability that a random f value would be greater than or equal to the observed value; **r 2 = coefficient of determination. table 1. analysis of variance (anova) of cycle time data m ea n cy cl e ti m e (m in ut es ) 74 the journal of engineering research vol. 2, no. 1 (2005) 69-76 0 5 10 15 20 25 1 2 3 4 5 6 7 8 9 10 observations c yc le -t im e (m in ut es ) existing workstation ergonomic workstation figure 5. observed cycle times of typical participant on workstations 0 1 2 3 4 5 ta ble h eig ht w ork sp ac e a rea th igh -ta ble c lea ran ce w ork tab le de sig n ch air h eig ht ad jus tab ilit y ch air c om for t fo otr est bi ns l ay ou t o n t he ta ble po sit ion & h eig ht of inc om ing b in po sit ion & h eig ht of ou tgo ing b in ha nd le of the to ols ea se of us e ov era ll w ork sta ion d esi gn w or ke rs s co re existing workstation ergonomically designed workstation figure 6. operator satisfaction with design attributes on two workstations satisfaction attribute areas mean satisfaction scores* non-ergonomically designed workstation ergonomically designed workstation workstation design 3.14 4.52 task method 3.03 4.35 health related issues 3.03 3.95 * max possible score = 5.0 table 2. mean operator satisfaction scores 75 the journal of engineering research vol. 2, no. 1 (2005) 69-76 with the posture you have to adopt in the task performance?'. 3.2. operator satisfaction the mean operator satisfaction scores for the workstation, task method and health attributes are presented in table 2. it can be seen that the mean scores for the ergonomically designed workstation were higher compared to the mean scores for the non-ergonomically designed workstation condition. results of anova in table 3, showed that the difference between the conditions was highly significant with regard to worker satisfaction (f=305.15, p<0.01). the improvement in worker satisfaction with regard to workstation design attributes is 45% (fig. 6). a similar analysis of the satisfaction scores with regards to the task method and health attributes showed significant improvements in the ergonomically designed workstation. worker satisfaction scores with regard to task attributes improved by 46%, while health attributes improved by 33%. overall worker satisfaction improved by 41%. the improvement was due to incorporation of ergonomics in the workstation design and task method that was well perceived by the workers. 4. conclusions the ergonomically designed workstation for performing an assembly task had a highly significant positive effect on reducing the cycle time. the reduction in the mean cycle time was 27% in using ergonomically designed workstation. worker satisfaction scores also improved significantly (i.e. 41%) on the newly designed workstation. a fully adjustable ergonomically designed system for assembly task was possible by using an adjustable worktable, an adjustable chair and a systematic layout. the new workstation could be used as a sit-stand workstation. the major changes made in the original workstation were in providing an ergonomic seat (i.e. an adjustable chair) an adjustable and adequate worktable, proper layout of bins and tools within the normal working area, ergonomically designed hand tools and an appropriate work method. ergonomic principles and data should be applied systematically in the design of repetitive assembly tasks to improve worker performance and satisfaction by making the tasks more comfortable, less fatiguing and more interesting. in closing, the current research was conducted for a short period, taking random cycles of assembly time for each operator. future research should focus on validation using a longer time period. to maximize operator performance, the optimum number of units that could be handled by the operator in a cycle should be investigated. worker satisfaction with regards to safety, health and satisfaction with the ergonomically designed workstation condition should also be investigated. acknowledgments the authors wish to acknowledge the financial support provided by squ through research grant eng/00/02. references al-haboubi, m.h., 1992, "anthropometry for a mix of different populations," applied ergonomics, vol. 23(3), pp. 191-196. ayoub, m.a., 1990a, "ergonomic deficiencies: i. pain at work," j. of occupational medicine, vol. 32(1), 52-57. ayoub, m.a., 1990b, "ergonomic deficiencies: ii. probable causes," j. of occupational medicine, vol. 32(2), pp. 131-136. burri, g.j. and helander, m.g., 1991, "a field study of productivity improvements in the manufacturing of circuit boards," int. j. of industrial ergonomics, vol. 7, pp. 207-215. carlopio, j.r., 1991, "construct human factors satisfaction questionnaire," working paper 91-001, australian graduate school of management, the university of new south wales, australia. das., b, 1987, "an ergonomic approach to designing a manufacturing work system," int. j. of industrial ergonomics, vol. 1(3), pp. 231-240. das, b. and grady, r.m. 1983, “industrial workplace layout design: an application of engineering anthropometry,” ergonomics, vol. 26(5), pp. 433-443. das, b. and sengupta, a., 1996, " industrial workstation design: a systematic ergonomic approach," applied ergonomics, vol. 27(3), pp. 157-163. source df ms f pr>f* r² ** model 20 200.70 496.98 0.01 0.98 conditions 1 123.24 305.15 0.01 subjects 9 4.15 10.27 0.01 conditions x subjects 9 5.08 12.58 0.01 error 240 0.40 total 260 variables are same as in table 1. table 3. anova of operator satisfaction scores with workstation design attributes 76 the journal of engineering research vol. 2, no. 1 (2005) 69-76 das, b. and shikdar, a., 1999, "participative versus assigned production standard setting in a repetitive industrial task: a strategy for improving worker productivity," int. j. of occupational safety and ergonomics, vol. 5(3), pp. 417-430. ergomas., 2001, "ergonomic design and optimization of manual work system," training manual, delmia, germany. grandjean, e., 1988, "fitting the task to the man: an ergonomic approach," taylor and francis, london, uk. hasselquist, r.j., 1981, “increasing manufacturing productivity using human factors principles,” proceedings of the human factors society 25th annual meeting, pp. 204-206. konz, s., 1995, "work design: industrial ergonomics,” (2nd edition), grid columbus, ohio. mcleod, d., 1995, "the ergonomics edge: improving safety, quality and productivity,” john wiley, new york. melamed, s., luz, j., najenson, t., jucha, e. and green, m., 1989, "ergonomic stress levels, personal characteristics, accident occurrence and sickness absence among factory workers," ergonomics, vol. 32(9), pp. 1101-1110. resnik, m.l. and zanotti, a., 1997, "using ergonomics to target productivity improvements,” computers and industrial engineering, vol. 33(1-2), pp. 185-188. ryan, j.p., 1989, "a study of selected ergonomic factors in occupational safety," advances in industrial ergonomics and safety i, ed. anil mital, taylor and francis, pp. 359-364. salvendy, g., 1987, "handbook of human factors," john wiley, new york. sanders, m.s. and mccormick, e.j., 1992, "human factors in engineering and design," 6th edition, mcgraw hill, new york. schnauber, h., 1986, "ergonomics and productivity as reflected in a new factory. trends in ergonomics/human factors iii," ed., karwowski, w., elsevier science publishers, pp. 459-465. shikdar, a., das, b. and hall, r., 1997, "ergonomics and worker productivity: a study with a repetitive manufacturing task," proceedings of the international workplace health and safety congress and ergonomics society of australia conference, cd rom 1-5. shikdar, a.a. and das, b., 1995, "a field study of worker productivity improvements,” applied ergonomics, vol. 26(1), pp. 21-27. wilson, j.r. and corlett, e.n., 1992, "evaluation of human work: a practical ergonomics methodology," taylor and francis, philadelphia. d:\desktop\paper_4.xps tjer 2013, vol. 10, no. 2, 46-51 _______________________________ *corresponding author’s e-mail: abdelaziz970@yahoo.fr a new scheme for crack growth modeling by coupling modified quarter point crack-tip element and the level set method y abdelaziz lms laboratory, department of mechanical engineering, university of bechar, algeria received 31 may 2011; accepted 8 december 2012 abstract: in this paper, an efficient, numerical procedure is presented to track crack growth modeling without remeshing. the method's key feature is the coupling of a modified quarter-point crack tip element (mqpe) with the level set method (lsm) for crack growth problems. the lsm was used to represent the crack location, including the location of crack tips. the mqpe was used to compute the stress and displacement fields necessary for determining the rate of crack growth. numerical test cases including various geometrical exceptions (the center-crack plate specimen, the single edge-crack plate specimen, and the double-edge crack plate) demonstrate the accuracy, robustness, and efficiency of the mqpe/lsm coupling. the extrapolation technique was used to estimate numerically the calibration factor for various specimens. this work confirms the feasibility of the mqpe/lsm to model accurately the singularity existing in the vicinity of the cracks. it allows an economic and adequate calculation of the stress intensity factors, which can be introduced into the various criteria of fracture or laws of propagation of the crack. the new method reduces the need for remeshing, and results agree well with reference data. keywords: fracture, crack, singularity, sifs, mqpe, level set method, calibration factors mqpelsm ) mqpe mqpe/lsm 47 y abdelaziz 1. introduction computational mechanics is a flourishing subject in science and engineering. it is a field in which physical mechanics problems are solved through mechanics, computers, and various numerical methods. the finite element method (long et al. 2009; oñate 2009) has been recognized widely as a well-established numerical method as it has applications in virtually are areas of science and engineering. however, for fracture mechanics (nestor 2004; anderson 2005), one of the major difficulties in simulating crack problems numerically resides in the singular nature of the asymptotic crack tip fields. in order to capture this singularity using conventional finite element method, high mesh density in the crack front region is required. also, there is a need for remeshing during crack growth simulations. however, this task is not always straightforward, especially for complex geometries. over last few decades, many approaches have been proposed to solve modeling of crack problems. for example, there have been methods based on the quarter-point finite element (barsoum 1974), the enriched finite element method (gifford and hilton 1978), the boundary collocation method (newman 1971), the integral equation method (sneddon 1973), the body force method (nisitani 1985), the boundary elements method (cruse 1988), and the dislocation method (chen and hasebe 1995), plus mesh-free methods such as the element-free galerkin method (fleming et al. 1997). to avoid remeshing in modeling crack problems, diverse techniques were proposed, including the incorporation of a discontinuous mode on an element level (oliver 1995), a moving mesh technique (rashid 1998), and an enrichment technique based on a partition-of-unity x-fem (belytschko and black1999). barsoum (1974) discovered that by taking the midside nodes of an element that are adjacent to a crack tip and moving them to the quarter point of the element side, the singular stress field which occurs at the crack tip could be produced (fig. 1). recently, various studies (gray et al. 2003; alashoaibi and arifin 2006; souiyah et.al. 2009; abdelaziz 2010) were carried out to develop this type of element. these efforts are justified by the effectiveness of the quarter-point finite element to reproduce a singularity and by its simplicity of integration in major calculation programs without modifications. specifically, gray et al. (2003) proposed a modification of the finite element with a quarter of a point in the vicinity of the cracks and used this new type of element in the two-dimensional analysis of the fracture. the standard singular element was adjusted so that the near-tip crack opening displacement satisfied a known constraint: the coefficient of the term which is linear in the distance to the tip must vanish. the two-dimensional standard quarter point finite element (qpe) is based upon the three-noded quadratic element. fort [0,1], the shape functions for this element are given by: (1) (2) (3) gray et al. (2003) defined new shape functions (fig. 2) by adding a cubic term: (4) (5) (6) 2. coupling the level set method with modified quarter-point crack tip element (mqpe) the level set method (lsm) was developed by the american mathematicians (osher and sethian 1988). it has become popular in many disciplines, including image processing, computer graphics, computational geometry, optimization, and computational fluid dynamics. the aim of this technique is to track the motion of an interface by embedding the interface as the zero level set of the signed distance function. the motion of the interface is matched with the zero level set of the level set function, and the resulting initial value partial differential equation for the evolution of the level set function. in this work, we propose a numerical scheme coupling the lsm with the modified quarter-point crack tip element. the lsm was used to represent the crack location including the location of crack tips. the mqpe was used to compute the stress and displacement fields necessary for determining the rate of crack growth. a crack is described by two level sets (fig. 3): 1 2 3 t = 0 t = 1 l l/4 crack tip figure 1. standard quarter point finite element (qpe) 48 a new scheme for crack growth modeling by coupling modified quarter point cracktip elemenet ... 1. a normal level set, (x), which the signed distance to the crack surface, and 2. a tangent level set (x), which is the signed distance to the plane including the crack front and perpendicular to the crack surface. in a given element, min and max are, respectively, the minimum and maximum nodal values of on the nodes of that element. similarly, let min and max be the minimum and maximum nodal values of on the nodes of an element, if min max= 0 and min max = 0, the elements around the crack tip become degenerated finite elements. 3. numerical examples in order to be able to compare our results with those in the literature, we determined the calibration factors of some samples of an isotropic material. three physical problems were modeled using the mqpe/lsm approach. they are a center-crack plate (ccp) specimen, a single edge-crack plate (secp) specimen, and a double-edge crack plate (decp) specimen, as shown in fig. 5. as these specimens had clean symmetry, we discounted their studies as a whole. figure 6 describes the mesh and boundary conditions applied to the specimens. the analytical calibration is proposed: (7) where ki = stress intensity factor; f = charge applied; w = width of the specimen; b = thickness of the specimen; a = length of the crack. the general expression of stress intensity factors (sifs) by means of the extrapolation technique is given by: (8) where r = outdistance face of the crack v = transverse displacement e = young's modulus for plane strain, k = (3-4 ) for plane stress, k = (3)/ (1+ ) = poisson's ratio the analytical calibration factor y(a/w) can be computed based on the equation in tada's handbook (2000) as follows: (a) the ccp specimen (b) the secp specimen (c) the decp specimen the results of this study are shown in figs. 6-8 and are compared with those of the analytical solution (tada and paris 2000) (table 1). the caliberation fact figure 3. construction of initial level set functions (9) (10) (11) 49 y abdelaziz tor increases with an an increase in crack width. the results are in good agreement with analytical solution. as can be seen from figures, the numerical study highlights the consistent agreement between current results obtained using mqpe/lsm approach, and the available analytical results. besides the consistent agreement of results, it is necessary to emphasize two very important points: * the mqpe was automatically introduced. * the analysis was carried out without remeshing, and the initial mesh was maintained despite the change in the ratio (a/w). 4. conclusions solving crack problems in fracture mechanics is imperative when quantifying and predicting the behavior of cracked structures under service conditions. to this end, the accurate evaluation of fracture parameters such as the stress intensity factors (sif) is required. the finite element method has proved to be very wellsuited for the study of fracture mechanics. however, one of its main drawbacks is that remeshing techniques are required to investigate and follow crack propagation patterns. moreover, the crack tip singularity needs to be accurately represented using approximation methods. due to the fact that the standard finite element methods are based on piecewise differentiable polynomial approximations, they are not well suited to problems with discontinuous and/or singular solutions. typically, finite element methods require significant mesh refinement to get accurate results. in response to this deficiency, modified finite elements have been developed. in this work, an initial attempt was made in developing a crack growth prediction tool under the assumption of linear elastic fracture mechanics. the tool was based upon the level set method, a mqpe crack-tip element, and the extrapolation technique for evaluating a calibration factor. the lsm was used to represent the crack location, including the location of crack tips. the mqpe was used to compute the stress and displacement fields necessary for determining the figure 4. various crack positions figure 5. mesh and boundary conditions ccp secp decp ccp secp decp 50 a new scheme for crack growth modeling by coupling modified quarter point cracktip elemenet ... rate of crack growth. in this paper, the performance of this technique is illustrated and compared with existing solutions for three examples. the results obtained are compared to reported solutions and showed an excellent agreement. this study has demonstrated the utility and potential of the mqpe elements as a computational fracture tool to study the singularity near the crack-tip. moreover, mqpe coupled with lsm makes possible the accurate solution of engineering problems in complex domains, which may be difficult to solve using the standard finite element method. references abdelaziz y, benkeira s, rikioui t, mekkaoui a (2010), a double degenerated finite element for modeling the crack tip singularity. applied mathematical modeling 34:4031-39. alashoaibi a, arifin k (2006), finite element simulation of stress intensity factors in elastic-plastic figure 6. calibration factor y(a/w) for ccp specimen figure 7. calibration factor y(a/w) for secp specimen crack specimen length/width (a/w) crack specimen length/width (a/w) table 1. normalized factors crack specimen length/width (a/w) figure 8. calibration factor y(a/w) for decp specimen 51 y abdelaziz crack growth. j zhejiang univ sci. a:1336-42. anderson tl (2005), fracture mechanics: fundamentals and applications. boca raton, florida, usa, crc press. barsoum rs (1974), application of quadratic isoparametric finite elements in linear fracture mechanics. int j fract. 10:603-5. belytschko t, black t (1999), elastic crack growth in finite elements with minimal remeshing. int j numer meth eng. 45:601-20. chen y, hasebe n (1995), new integration scheme for the branch crack problem. eng fract mech. 52:791-801. cruse ta (1988), boundary element analysis in computational fracture mechanics. norwell, massachusetts, kluwer academic publishing. fleming m, chu y, moran b, belytschko t (1997), enriched element-free galerkin methods for crack tip fields. int j numer methods eng. 40:14831504. gifford j, hilton p (1978), stress intensity factors by enriched finite elements. eng fract mech. 10:48596. gray l, phan a, paulino h, kaplana t (2003), improved quarter point crack tip element. eng fract mech. 70:269-83. long y, cen s, long z (2009), advanced finite element method in structural engineering. wiesbaden, germany, springer science + business media. nestor p (2004), fracture mechanics. dordrecht, the netherlands, kluwer academic publishers. newman j (1971), an improved method of collocation for the stress analysis of cracked plates with various shaped boundaries. technical report tn d-6376 nasa. nisitani h (1985), body force method for determination of the stress intensity factors. j aeronautical soc india (special issue on fracture mechanics) 37:21-41. osher s, sethian ja (1988), fronts propagating with curvature-dependent speed : algorithms based on hamilton-jacobi formulations. j comput. phys 79:12-49. oliver j (1995), continuum modeling of strong discontinuities in solid mechanics using damage models. comput mech. 17:49-61. oñate e (2009), structural analysis with the finite element method. linear statics: vol. 1, basis and solids. wiesbaden, germany, springer science + business media. rashid m (1998), the arbitrary local mesh refinement method: an alternative to remeshing for crack propagation analysis. comput meth appl mech eng. 154:133-50. sneddon i (1973), integral transform methods, in: methods of analysis and solutions of crack problems. leyden, the netherlands, nordhoff international. souiyah m, muchtar a, alshoaibi a, ariffin ak (2009), finite element analysis of the crack propagation for solid materials. j app sci 6:13961402. tada h, paris p (2000), the stress analysis of cracks handbook. new york, usa, asme press. d:\desktop\paper_7.xps tjer 2013, vol. 10, no. 1, 80-87 ___________________________________ *corresponding author’s e-mail: rashidkh@neduet.edu.pk anisotropic deformation behavior of al2024t351 aluminum alloy r khan mechanical engineering department, ned university of engineering and technology, karachi, pakistan received 15 october 2011 ; accepted 28 july 2012 abstract: the objective of this work was to investigate the effects of material anisotropy on the yielding and hardening behavior of 2024t351 aluminum alloy using isotropic and anisotropic yield criteria. anisotropy may be induced in a material during the manufacturing through processes like rolling or forging. this induced anisotropy gives rise to the concept of orientation-dependent material properties such as yield strength, ductility, strain hardening, fracture strength, or fatigue resistance. inclusion of the effects of anisotropy is essential in correctly predicting the deformation behavior of a material. in this study, uniaxial tensile tests were first performed in all three rolling directions, l, t and s, for smooth bar specimens made from hot rolled plate of al2024 alloy. the experimental results showed that the land t-directions yielded higher yield strengths and a greater percentage of elongation before fracture than the s-direction. subsequently, finite element analysis of tensile specimens was performed using isotropic (von mises) and anisotropic (hill) yield criteria to predict the onset of yielding and hardening behaviors during the course of deformation. hill's criterion perfectly fitted with the test data in the s-direction, but slightly underestimated the yield strength in l-direction. the results indicated that the hill yield criterion is the most suitable one to predict the onset of yielding and hardening behaviors for 2024t351 aluminum alloy in all directions. keywords: deformation behaviour, anisotropy, downhole expandable tubular, finite element analysis al2024t351 2024t351 'l''t''s' ai2024'l' 't' 's' 's''l' 2024t351 81 r khan 1. introduction aluminum alloys are well-known for their excellent mechanical properties such as a high strength-toweight ratio, resistance to corrosion, and strength integrity at low temperatures. these properties have led to its extensive use in aerospace and cryogenic applications. for decades, thin rolled sheets of aluminum alloy 2024t351 have been used for aircraft applications (starke 1996; nakai 2000). like other aluminum alloys, the 2024 alloy exhibits anisotropy in its deformation and fracture behaviors. although extensive experimental work, along with analytical and numerical analyses, has been performed (lam 2010; lee 1999; mimaroglu 2003; schmitt 1998) for predicting the isotropic and anisotropic deformation behaviors of aluminum 2024 alloys, so far its deformation-induced anisotropy has not been completely understood. anisotropy is induced in a material during manufacturing through processes like rolling or forging. the induced anisotropy results in directionaldependent material properties such as yield strength, ductility, strain hardening, fracture strength, and fatigue resistance. numerous studies have been done to evaluate the material properties of 2024 aluminum alloys (bron 2006, 2004b; srivatsan 2007; shanmuga 2010). a thorough understanding of the effects of induced anisotropy is essential to correctly predicting the deformation behavior of a material. in thin plates made of 2024 alloy, the properties are strongly dependent on the rolling direction (steglich 2008), while the other two orientations (ie. thickness and width) are not critical in predicting the deformation behavior. however, for thick plates the anisotropy along thickness and width directions cannot be neglected in estimating the deformation behavior under different loading conditions. the onset of yielding, the progression of plastic deformation, and work hardening are not the same in all directions. researchers have carried out numerous studies and developed various yield criteria, plastic flow rules, and work/strain hardening hypotheses to predict deformation-induced anisotropy for ferrous alloys. these same criteria, rules, and hypotheses are assumed to be equally applicable for aluminum alloys with some modifications. however, the yield criterion which is most applicable for 2024 alloy for all directions, and which agrees well with experimental results, remains a topic of research. there is no reported literature on this subject and filling this void is the main objective of the current study. downhole tubular expansion technology in the oil and gas industry has demonstrated its enormous influence on the overall drilling cost of wells (campo 2003). many issues remain unresolved in spite of its successful implementation. one of these issues is related to induced anisotropy in downhole tubular material during the cold expansion process. in the expansion process, the tube might not become deformed uniformly due to the presence of non-homogeneities as a result of the manufacturing process. the non-uniform expansion process results in the anisotropic behavior of downhole tubes and ultimately affects their structural integrity. therefore, a careful analysis of induced anisotropy is necessary to estimate its effects on the structural integrity of downhole tubes. this work focused on analyzing an induced anisotropy in a material during deformation, while the relation between an induced anisotropy in a downhole tube with its structural integrity will be included in future work. also, this research is an addition to the work reported by pervez (2008), where the postexpansion properties of downhole tubular materials like expansion force, thickness reduction, length shortening, and contact pressure were compared in two different types of materials steel and aluminum. the authors concluded that aluminum is a better option for downhole tubular construction due to its lower energy requirement for expansion, high formability, and good corrosion resistance. this work contains the effects of anisotropy on the deformation behavior of downhole tubular materials, while assuming that the downhole tube is made up of al2024 aluminum alloy. a law defining the onset of yield under a combined state of stress is known as yield criterion. the generalized von mises yield criterion for isotropic materials (1913) accurately predicts the onset of yielding for most polycrystalline metals as compared to the criterion of (tresca 1864). an extension of von mises's yield function was first proposed by hershey (1954), and then was generalized by hosford (1972). the loci of yield surface of isotropic yield function by hosford lies between that proposed by von mises and tresca. the equivalent stress by hosford is given as: 1 2 3 3 1 1 2 (1) 82 anisotropic deformation behavior of al2024t351 aluminum alloy (2a) where, rij is defined as (2b) (3) (4) (4a) (4b) (4c) (4d) (2) bk k 83 r khan the current work is aimed at the prediction of yielding and hardening behaviors, resulting from different loading conditions and specimen geometries for the al2024t351 alloy using different yield criteria. uniaxial tensile tests were performed on smooth bar specimens. fe simulations were done in fe code abaqus for smooth bar specimen using different yield criteria to find out which single yield criterion could be used to fit the hardening curve in all three directions. the numerical results were validated using experimental data. 2. experimental study uniaxial tensile tests were performed on smooth bar specimens made from a hot rolled plate of 2024 alloy in l-, tand s-directions. the material orientations with respect to rolling and loading directions are shown in figures 1a and b. before cutting the sample, the plate was solution heat treated, air quenched, and stress relieved by cold stretching. all samples had a uniform thickness of 100 mm. the gauge length of the smooth bar specimen was 50 mm with a diameter of 10 mm. the main alloying element in 2xxx series of aluminum alloy was copper, while magnesium and manganese were added to improve the quenching properties. the composition of the al2024 alloy is given in table 1. 3. finite element analysis 4. results and discussions the experimental results for the uniaxial tensile tests of smooth bar are shown in fig. 2. for smooth bar, the yield point remained the same in land tdirections, but was of a lower value in the s-direction. for smooth bar, the percentage elongations in l, t and s-directions were 22, 19.6, and 9, respectively. this (4e) 84 anisotropic deformation behavior of al2024t351 aluminum alloy means that the material exhibited less ductility in the thickness direction as compared to the rolling and width directions. 4.1 smooth bar specimen table 1. composition of 2024 aluminium alloy by weight percent (%) (a) (b) figure 1. (a) thin rolled plate of aluminium 2024t351 alloy; (b) material orientation with respect to loading directions (1,2,3) and (x,y,z) are global and local coordinate systems, respectively figure 2. experimental results of tensile tests for smooth bar specimens, ss01 has an s-direction smooth bar in experiment 1 and ss03 has an s-direction smooth bar in experiment 3. the findings are similar in the land t-directions) (a) (b) figure 3. force versus elongation in the s-direction using von mises yield criterion; (a) 2-d finite element analyses (fea); (b) 2-d/3-d fea (a) (b) figure 4. force versus elongation in the s-directions using hill’s yield criterion; (a) fea based on yield strengths and (b) ultimate strengths 85 r khan 86 anisotropic deformation behavior of al2024t351 aluminum alloy (5) the given coefficients in eqn. (5) are a clear indication of the presence of anisotropy along the x and z axes only, but not along the y axis, which is the loading direction in s-material orientation. it indicates that only planar anisotropy is present. also, no anisotropic effect on shearing was considered (r12=1, r13=1, r23=1). the comparison between simulation and experimental results for this case is shown in fig. 4(a). it is evident from fig. 4(a) that the hill's criterion resulted in a close match with experimental results for the s-direction but overestimated the force at the onset of yield by approximately 7% for the l-direction. based on ultimate tensile strengths in sand l-directions, the anisotropic coefficients for hill's yield criterion are given in eqn. (6): (6) it is obvious from fig. 4(b) that the simulation results perfectly fit the experimental hardening behavior in the l-direction while underestimating the onset of yielding. it can be concluded that using hill's yield criterion, both the onset of yielding and hardening behaviors in the l-direction cannot be accurately predicted. therefore, the choice of anisotropic coefficients in hill's yield criterion is mainly dependent on the application's need. for designing products involving plastic deformation, hardening behavior is more important as compared to the onset of yielding, while the reverse is true in the case of designing within the elastic limit. the results prove that there is no isotropic effect in shear for smooth bar. 5. conclusions isotropic and anisotropic yield criteria were used to predict the yielding and hardening behaviors for sand l-directions in smooth bar using uniaxial tensile test specimens. hardening behavior was determined only for s-direction. three dimention finite element analyses for smooth bar were performed using isotropic and anisotropic yield criteria. anisotropic parameters for hill's criterion were determined using mechanical test data of smooth bar in the l-direction. hill yield criterion slightly overestimated the hardening curve for the s-direction, while the effect was more pronounced in the l-direction. it can be concluded that by using hill's yield criterion, yielding and hardening behaviors in the l-direction cannot be predicted precisely. therefore, the selection of hill's anisotropic coefficients is mainly dependent on application. the present work could be extended by including other anisotropic yield criteria to predict deformation behavior more accurately. furthermore, the analysis of notched bar specimens could also be included to investigate the effects of notch radius on the anisotropy of materials. references barlat f, becker rc, hayashida y, maeda y, yanagawa m, chung k, brem jc, lege dj, matsui k, murtha sj, hattori s (1997), yielding description for solution strengthened aluminum alloys. international journal of plasticity 13:385401. barlat f, lege dj, brem jc (1991), a six-component yield function for anisotropic materials. international journal of plasticity 7:693-712. bron f, besson j ( 2004), a yield function for anisotropic materials: application to aluminum alloys. international journal of plasticity 20(45):937-963. bron f, besson j (2006), simulation of the ductile tearing for two grades of 2024 aluminum alloy thin sheets. engineering fracture mechanics 73(11):1531-1552. bron f, besson j, pineau a (2004), ductile rupture in thin sheets of two grades of 2024 aluminum alloy. materials science and engineering a 380(12):356-364. campo d, williams c, fillippov a, cook l, brisco d, dean b, ring l (2003), monodiameter drilling liner from concept to reality paper spe 79790, presented at the drilling conference held in amsterdam, the netherlands. hershey av (1954), the plasticity of an isotropic aggregate of anisotropic face-centered cubic crystals. journal of applied mechanics 21:241-249. 87 r khan hill r (1954), the mathematical theory of plasticity. clarendon press, oxford, uk. hosford wf (1972), a generalize isotropic yield criterion. journal of applied mechanics 39:607-609. hosford wf (1996), on the crystallographic basis of yield criteria. textures microstructure 26-27:479493. karafillis ap, boyce mc (1993), a general anisotropic yield criterion using bounds and a transformation weighting tensor. journal of mechanics of physics and solids 41:1859-1886. lam df, menzemer cc, srivatsan ts (2010), a study to evaluate and understand the response of aluminum alloy 2026 subjected to tensile deformation. materials and design 31(1):166-175. lee ht, shaue gh (1999), the thermomechanical behavior for aluminum alloy under uniaxial tensile loading. materials science and engineering a 268(1-2):154-164. logan rw, hosford wf (1980), upper-bound anisotropic yield locus calculations assuming <111> pencil glide. international journal of mechanical sciences 22:419430. mimaroglu a, yenihayat of (2003), modelling the superplastic deformation process of 2024 aluminium alloys under constant strain rate: use of finite element technique. materials and design 24(3):189-195. nakai m, eto t (2000), new aspect of development of high strength aluminum alloys for aerospace applications. material science and engineering a 285(1-2):62-68. pervez t, qamar sz, seibi ac, al-jahwari fk (2008), use of set in cased and open holes: comparison between aluminum and steel. materials and design 29:811-817. schmitt c, lipinski p (1998), modelling of the elasto-plastic behavior of rolled aluminium alloy for forming applications. journal of materials processing technology 74(1-3):14-22. shanmuga sn, murugan n (2010), tensile behavior of dissimilar friction stir welded joints of aluminium alloys. materials and design 31(9):41844193. srivatsan ts, vasudevan s, park l (2007), the tensile deformation and fracture behavior of friction stir welded aluminum alloy 2024. materials science and engineering a 466(1-2):235-245. starke ea, staley jt (1996), application of modern aluminum alloys to aircraft. progress in aerospace science 32(2-3):131-172. steglich d, brocks w, heerens j, pardoen t (2008), anisotropic ductile fracture of al 2024 alloys. engineering fracture mechanics 75:3692-3706. tresca h (1864), on the flow of solid bodies subjected to high pressures. in: comptes-rendus de l'académie des sciences paris 59:754-756. von mises r (1913), mechanik der festen körper im plastischdeformablen zustand. nachrichten von der gesellschaft der wissenschaften zu göttingen, mathematisch-physikalische klasse 582592:doi:citeulike-article-id:6897288. the journal of engineering research (tjer), vol. 15, no. 1 (2018) 73-87 doi: 10.24200/tjer.vol15iss1pp73-87 capture aware channel access protocol in wireless network s. mustafa* college of engineering, salahaddin university, iraq 21 january 2015; accepted 1 may 2017 abstract: spatial reuse in wireless networks is limited by the sinr threshold and it might be feasible to capture a packet in the presence of ongoing foreign transmission. this work considers a new capture aware channel access protocol by incorporating global channel state information in the decision making process for the channel access jointly with adaptive power framework. the protocol employs power heterogeneous ad-hoc networks; it assigns different transmission power level to individual nodes based on dynamic observation of the network traffic. it exploits spatial heterogeneity of flows at a given channel allocation and sets up either symmetric or asymmetric carrier sensing which in turn schedules the data packets transmission to maintain adequate service quality and fairness enjoyed by a user. it stands atop capture capable phy to leverage the channel reuse which is of paramount importance in the design of high capacity ad-hoc networks. through extensive simulations, the paper demonstrates the efficacy of the new protocol. it delivers high network utilization and also provides fair access to the media. keywords: dcf; carrier sensing; mac mechanism; power control; capture effect; wireless network. بروتوكول الدخول اىل القنوات و الدراية حبدوث التقاط االرسال يف الشبكة الالسلكية *مساح مصطفى )سينر(. إشارة نسبة التدخل: ان إعادة االستخدام املكاني يف الشبكات الالسلكية يكون حمدودا بسبب حاجز امللخص وقد يكون من املمكن التقاط حزمة ما يف وجود ارسال أجنيب مستمر. ويأخذ هذا العمل يف االعتبار بروتوكوال جديدا للوصول إىل القنوات من خالل حالة دمج املعلومات بالقناة العامة يف عملية صنع القرار من أجل الوصول إىل القناة وباالستفادة ويستخدم الربوتوكول قدرة غري متجانسة لشبكات خمصصة , حيث أنه حيدد مستوى قدرة .فمن إطار القدرة على التكي اإلرسال املختلفة بالنسبة إىل الُعقد الفردية على أساس مراقبة ديناميكية حلركة مرور الشبكة. فهو يستغل عدم التجانس ماثل أو غري متماثل للموجة احلاملة وحيدد املكاني للتدفقات عند ختصيص قناة معينة ، وحيدد سواء كان االستشعار مت بدوره إرسال رزم البيانات للحفاظ على وجود جودة خدمة كافية وعادلة يتمتع بها املستعمل. وهو يقف فوق عملية االلتقاط . من القادرة على االستفادة من إعادة استخدام القناة اليت تعترب ذات أهمية قصوى يف تصميم شبكات خمصصة عالية السعة خالل احملاكاة واسعة النطاق. تربهن هذه الورقة على فاعلية الربوتوكول اجلديد. وهي توفر استخدام عال للشبكة وتوفر أيضا إمكانية الوصول العادل إىل وسائط اإلعالم. وسائط اإلعالمالتحكم بالوصول إىل آلية ،االستشعار الناقل،) دي سي اف( وظيفة التنسيق املوزع :ة كلمات املفتاحيال تأثري االلتقاط. شبكة السلكية.،التحكم يف الطاقة ،) ماك( * corresponding author’s e-mail: samah.eecs@gmail.com mailto:samah.eecs@gmail.com s. mustafa 74 1. introduction a key problem in ad-hoc wireless networking is the design of distributed media access control (mac) mechanism to discover and negotiate the access to the reusable physical channels such that the transmission can occur without collision. a representative mechanism of mac is the carrier sense multiple access/collision avoidance (csma/ca) algorithm that forms the basis of the dominant wireless multiple access protocol distributed coordination function (dcf) in 802.11 (kurth and redlich 2009). physically a channel is sensed busy either when the detected energy level is above a certain threshold ζed, when a signal with the same phy characteristics is detected, or a combination of both (larroca and rodríguez 2014). carrier sensing is the main interference mitigation mechanisms used in the phy/mac layers of 802.11 wlans, and its efficacy becomes a key factor in determining network capacity. a station defers its transmission for a random period upon detecting a carrier on the channel. however, the absence of a carrier does not mean a transmission will succeed, nor does the presence of a carrier mean that a transmission will interfere. hereby carrier sense approach can unnecessarily reduce network throughput and strongly affect the delay characteristics at a station (jamieson et al. 2005; zhu et al. 2004; ma et al. 2009). many studies have been done to illustrate that physical carrier sense adversely limits the effective bandwidth utilization in the network because of fixed carrier sense threshold and inadequately chosen carrier sense range. extensive research efforts advocate the tuning of carrier sense threshold for utilization improvement (ma et al. 2009; zhu et al. 2004; zhu e et al. 2006; haghani et al. 2010); however, many of today’s 802.11 mac implementations do not allow the threshold to be independently tunable or even accessible (zhu et al. 2004). moreover, (pelechrinis 2009) claims that tuning carrier sense threshold opens the door for selfish users with lower back-off times to ignore other transmission and therefore obtain a higher unfair share of the spectrum. there is a clear trade-off between the probability of interference and the utilization in setting carrier sense range. larger sense range decreases both the probability of interference and the bandwidth utilization, whereas small carrier sense range will be more subject to interference. exposed node prevents a successful transmission while hidden node wastes a transmission opportunity due to a collision. as has been pointed out in (ma et al. 2009), the optimum value of carrier sense threshold is that at which the carrier sense range of the transmitter just covers the interference range of the receiver. the setting of interference range is rather heuristic and remains an open problem. many studies on wireless networks have largely considered an interference range as twice the transmission range, and others rely on equal interference and transmission ranges. physically the interference range is defined to denote the range within which an interfering node will lower signal-to-interference ratio (sir) below the threshold for successful decoding of a transmission (maheshwari et al. 2008). it has been depicted in (chen et al. 2007 and lee et al. 2010), an interferer’s impact becomes serious when its preamble is detectable such that the intended end-user device could engage with, even if it is not able to decode the frame correctly. hereby the author recommends characterizing the interference range within which the preamble of a frame is detectable and the carrier sense range should be equal to this range. recall that spatial channel reuse relies on the carrier sense threshold and the transmit power each station uses. mac like 802.11 dcf uses the same transmission power at all nodes regardless the link distance between transmitter-receiver pair as well as interference level at the receiver (li et al. 2009). moreover, it always transmits at full power which in turn potentially maximize the link utility from the medium. however, this selfish behavior leads to excessive mutual interference and degrades the network performance. in contrast, some proposed mac mechanisms (gomez et al. 2001; lim and yoshida 2005) transmit at minimum possible power to leverage the spatial reuse hence maximizing the interference probability. the sinr threshold for successful reception depends on the timing and the relative order of signal and the interference at the receiver (chen et al. 2007; lee et al. 2010; and mustafa 2015). the radio captures a new stronger frame above a predefined preamble capture threshold γp, if it arrives during the first frame’s preamble reception. 802.11's phy layer continuously monitors the received signal strength even the layer in data reception state. it enables a radio to correctly receive a second frame, even after it capture aware channel access protocol in wireless network 79 has already synchronized with the first one, if the ratio of powers is sufficiently high. this is called message in message mim mode which is not exploited before in capture aware mechanisms. a typical ratio of powers required for body capture is γd=10db (chen et al. 2007 and lee et al. 2010). thus, the sinr at the intended receiver and transmission order are crucial factors for successful delivery. this work presents an approach to leverage ad-hoc network's capacity relying on a distributed capture aware framework to coordinate the transmit power and time schedule the traffics to exploit the phy capture effect with ultimate goals of improving utility and spectrum fairness. time schedule the traffic is done by assigning transmit power level to set up either symmetric or asymmetric carrier sense scenario that can potentially schedule the data packets transmission. asymmetric carrier sense situation is neither actually introduced nor applied in previous works. it is known that a hidden link would starve in presence of sensed link with high load traffic. the simulation study offers a new prospective regarding the asymmetric carrier sense scenario and demonstrates its benefit to time schedule the hidden traffic without an extra timing signal. in this work, capture aware multiple access protocol (cama) is proposed. it requires disseminating the locally measured link gain at individual stations throughout the network neighborhood for identifying links that can be concurrent and allowing them to coexist, thereby enhanceing the effectiveness of channel reservation. it doesn’t modify the carrier sense threshold. cama exploits spatial heterogeneity of flows at a given channel allocation by means of dynamic power control to manage the cochannel interference and allow for higher and fair utilization of resources. the rest of the paper is organized as follows: section 2 goes through works related to this paper. section 3 describes cama mechanism that adjusts the transmission power of the nodes to maximize the benefits of the capture effect. section 4 presents the simulation results that describe the behavior of the default and cama mechanisms, and demonstrate the benefits of the proposed protocol. finally, section 5 concludes the paper with general observations and suggestions for future research. 2. related work extensive works have subjected different aspects of 802.11mac protocol to enhance perflow and aggregate throughput. this section presents an overview of some existing work that aims to improve network capacity by spatial diversity through tuning transmit power. ding et al. 2005 proposes transmit power control protocol to adjust the transmit power for each frame under the same decision rule as ieee 802.11 which in turn achieves limited improvement. mhatre et al. 2007 claim that different power levels at transmitters use the same carrier sense threshold in 802.11 networks introduces asymmetric links which in turn lead to throughput starvation. they demonstrate that cross layer approach by joint tuning of the clear channel assessment cca threshold and transmit power ensures starvation free control system. furthermore, the clients use the same transmit power and cca threshold as their associated access point regardless of different channel state each sees. the current work demonstrates that asymmetric link do not necessarily introduce starving nodes. the traditional access control cannot handle multiple packets without declaring a collision. a capture aware mac mechanism was proposed in (santhapuri et al. 2007), where primary node transmits request-to-send (rts)/clear-to-send (cts) and waits for one preamble time interval before transmitting data while secondary node(s) that overhears rts will initiate its own transmission either during wait time for the primary user or during primary's data transmission depending on the estimated sinr at its receiver and at the primary user. however, this scheme doesn't address the synchronization issue due to the existence of propagation delay, and it incorporates that receiving the preamble of the interference frame helps to detect the frame of interest that arrives later while in contrast has been proved in (lee et al. 2010). in (li et al. 2009), nodes transmit rts/cts at the same maximum power, and transmit data/ack at the power based on the link distance, as well as the interference level at the receiver. based on the overhead rts or cts, the protocol enables a concurrent transmission at a possible highest power level as long as it does not interfere with the ongoing ones. although, 75 s. mustafa 78 the protocol considers the aggregate interference level, but it assumes equal contribution from the neighboring nodes, no matter how close they are as well as the transmit power they use. there are some limitations on these mechanisms: hidden terminal problem still exists with rts/cts, and possible loss in ack packets is ignored to benefit from fewer collisions and build inaccurate results. various phy layer capture aware mac mechanism are proposed for wireless area local area networks wlans. to meet fair access and high overall throughputs, (jeong et al. 2013) changes either the contention window or arbitration inter-frame space or transmission opportunity. the technique in (patras et al. 2014) uses different power levels to result in packet capture in wlans and improve the performance. however, an ideal channel condition is assumed where losses are only caused by collisions. the work in (takahashi et al. 2015) configures the minimum contention window of the stations to reduce the collision in each network. these studies considered infrastructure mode where nodes are within carrier sense range of one another. further, these mechanisms don’t exploit mim mode and the assignment of concurrent links set satisfies a single sinr constraint. the study in (mustafa 2015) demonstrates the inaccuracy of the well-known interference physical model. 3. capture aware multiple access cama framework the 802.11 dcf (li et al. 2009) is strong transmit power mac design wherein all nodes transmit with the same power no matter how close the sender and receiver are. the current work here claims that by moving beyond the maximum fixed transmit power, the fairness and channel utilization can be improved. further, successful unicast transmission depends on channel condition at both the receiver and sender due to downstream data traffic and upstream mac layer acknowledgment ack traffic. however, suggestion of simply extending the carrier sensing mechanism to the receiver does not address the mac limitations of inefficient and unfair spectrum utilization. a capture aware multiple access mechanism is proposed based on broad channel observation to tune power level of each node to successfully retrieve the transmitted signal and make a decision to access the channel, while carrier sense range is fixed and set equal to interference range. in environment with users of homogenous technology competing for resources, the interference range in this work is characterized within which the preamble of a frame is detectable. 3.1 two sender-receiver pairs consider two sender-receiver pairs: (s1;r1) and (s2;r2) transmitting over communication links l1 and l2 respectively, and sharing a single wireless channel as shown in fig. 1. 𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅𝑖 experienced by ri due to concurrent data data-data transmissions from si and sj is given by 𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 𝐺𝑗,𝑖 𝑆𝑅 𝑃𝑆𝑗 + 𝜂 (1) figure 1. two flows sharing a wireless channel. depending on propagation paths between the four stations, link(s) may or may not be established between the nodes. psi denotes transmit power level at si and η is the noise floor. 𝐺𝑖,𝑗 𝑆𝑅 denotes si-rj link gain where a propagation model suited for the surrounding environment is used to describe the fluctuation in the received signal strength such that the mean received power falls off as (1/d)α. α is a path loss exponent. a receiver can engage with a delivered packet if the received signal power is greater than the receiver sensitivity threshold ζs, i.e. 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 ≥ ζ𝑆 (2) whether parallel data transmissions of data packets are successful or failure depends on the s1 r1 s2 r2 𝐺1,1 𝑆𝑅 𝐺2,2 𝑆𝑅 𝐺1,2 𝑅𝑅 𝐺2,1 𝑆𝑅 𝐺1,2 𝑆𝑅 𝐺1,2 𝑆𝑆 76 capture aware channel access protocol in wireless network 79 observed sinr basically at receivers that should be higher or equal to γd. it enables the radio to correctly receive an intended frame irrespective of the timing relation between overlapped transmissions. γd is usually set to 10 that map to ten times the signal power over interference power, as stated earlier in first section. this condition holds if ri is closer to si than to both sj and rj by a sufficient factor, and vice versa. it may happen that this condition holds even though csma/ca would prevent the second pair from transmitting. to illustrate this possibility, consider the following example. the four nodes are on a line in the order, from left-to-right r1; s1; s2; r2. their locations on the x-axis are 0; 1; d+1; d+2, respectively. assume all the nodes transmit with a unit power and (𝐺𝑖,𝑖 𝑆𝑅 /𝐺𝑗,𝑖 𝑆𝑅 ) ≥ 10db. a capture-aware protocol that would let the sender transmit even if it sense the other transmission could potentially have a throughput almost twice as large as the standard dcf protocol. cama looks for a minimum possible power level that guarantees correct packet delivery and hide the sender from other. further, it is able to throwing out carrier sense line by disabling randomized backoff and allows two senders to transmit packets continuously and simultaneously. now, consider the four nodes on the x-axis in the order from left to right s1; r1; r2; s2, at locations 0; 1; d+1; d+2, respectively. depending on the channel state the nodes see, data-ack and data-data packets collision are possible. for example, 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅𝑖 due to simultaneous transmission of data and ack packets from si and rj, respectively is given by 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 𝐺𝑗,𝑖 𝑅𝑅 𝑃𝑅𝑗 + 𝜂 (3) pri denotes the transmit power of ri. assume the nodes transmit with a unit power over a single channel of state such that 𝐺𝑖,𝑖 𝑆𝑅 /𝐺𝑗,𝑖 𝑅𝑅 is less than a sufficient factor ζd. ζd=5db is the sinr threshold to guarantee a successful packet delivery if the receiver is not engaged with another packet (ieee std. 2007, chen et al. 2007, lee et al. 2010). it may happen that the senders are hidden and don't preclude data transmission if any. failed transmission not only wastes energy but also has the potential to corrupt other transmissions. the proposed approach evaluates the state and makes an efficient decision to access the channel. it assigns a transmit power to set up mutual carrier sensing and dependently schedule the transmission opportunity. homogenous links in terms of channel gain are assumed in previous examples, nevertheless they are not usually. cama exploits spatial heterogeneity of flows and assigns different power level at each station. recall that link s2r2 has higher quality than link s1-r1. while the default solution for sender(s) out of the capture range of the intended receipt is to preclude the concurrent transmissions, a more efficient solution is addressed by achieving asymmetric carrier sensing through heterogeneous power to schedule the transmission time. high power node s2 does not sense the transmissions of low power node s1. cama looks for possible minimum and different power levels at each sender to set [𝐏𝐒𝟐𝐆𝟐,𝟐 𝐒𝐑 /(𝐏𝐒𝟏 𝐆𝟏,𝟐 𝐒𝐑 + 𝜼)] ≥ 𝟏𝟎𝐝𝐁 and [𝐏𝐒𝟏𝐆𝟏,𝟏 𝐒𝐑 /(𝐏𝐒𝟐𝐆𝟐,𝟏 𝐒𝐑 + 𝜼)] ≥ 𝟓𝐝𝐁, and lets the senders achieve high throughput. transmit power plays a vital role in our mechanism to enable either the mim or capture effect at the nodes and set possible concurrent transmissions on a single wireless channel. s1 (being hidden from s2) transmits only if the channel is idle and commences a data transmission prior to s2 which doesn’t block any transmission request from an upper layer on sensing a clear channel. it can potentially initiate transmission that overlap with lower power transmission. thereby, r1 captures the first coming packet if there is an overlapping transmission on the channel and r2 will begin receiving s1’s packets first, and later re-lock onto s2’s new packet which is much stronger than s1’s. the senders can successfully and fairly utilize the channel if the following conditions hold:  𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅1 and 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅1 exceed a threshold guarantees a successful packet detection threshold ζd.  𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅2 and 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅2 exceed γd and ζd, respectively. further constraint should be followed since it is also possible that a data packet is received correctly, but the ack packet is lost 𝑆𝐼𝑁𝑅|𝐴−𝐷 𝑆𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑅𝑖 𝐺𝑗,𝑖 𝑆𝑆 𝑃𝑆𝑗 + 𝜂 ≥ ζ𝐷 (4) briefly, the channel is utilized efficiently only if s1 commence transmission before s2. 77 s. mustafa 78 flow of higher transmit power can afford to start later. it is important to highlight that we are not using a special timing signal to schedule the data transmission; we only depend on carrier sense scenario to time order the traffic. it is well understood that asymmetric links in power heterogeneous network unfairly degrades the utility experienced by low power links (mhatre et al. 2007, shah et al. 2007). however, our simulation study confirms that diversity in transmit power does not necessarily starve low power users if some constraints are followed. 3.2 n sender-receiver pairs the intended stations are responsible for estimating and disseminating the channel quality around to make a joint prediction about whether more than a flow can transmit simultaneously through minimum possible transmit power at each station to mitigate the interference and optimize the per-flow and overall network throughput on a collision free channel. if the prediction result is positive, the stations should not block their own transmissions, if any. otherwise, an approach that considers tuning of transmit power as a phy parameter, should be applied to set a suitable sense interaction and subsequently schedule the access to a shared medium. simulation results in section 4 illustrate that such an access scheme performs better than the csma/ca even at high traffic loads, and applying it has the benefit of improving perflow and aggregate utility in wireless networks. to estimate the complexity of such a protocol, assume there are n sender/receiver pairs willing to access the media. the nodes in pair i should estimate the channel gain over link(s) that may establish; 𝐺𝑖,𝑗 𝑆𝑅 , 𝐺𝑖,𝑗 𝑆𝑆 and 𝐺𝑖,𝑗 𝑅𝑅 ∀ neighbor pair j, either periodically or whenever it has data to transmit and/or node(s) observes degradation in the performance. they then post these values to a distributed control system to coordinate the channel access and transmit power of the nodes according to the specified policy. the stations may exchange simple control signaling to initiate the process and run the optimization algorithm. despite that the optimal algorithm provides optimal utility of the channel, it is rather computationally expensive for a large network. to guarantee successful concurrent transmissions that would not change the result of the capture effect, sinr constraints should consider joint and maximum interference strength at a station. 𝑆𝐼𝑁𝑅1|𝐷−𝐷 𝑅𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 ∑ 𝐺𝑗,𝑖 𝑆𝑅 𝑃𝑆𝑗 𝑁 𝑗=1,𝑗≠𝑖 + 𝜂 ≥ ζd (5) 𝑆𝐼𝑁𝑅2|𝐷−𝐷 𝑅𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 𝑚𝑎𝑥𝑗≠𝑖 (𝐺𝑗,𝑖 𝑆𝑅 𝑃𝑆𝑗) + 𝜂 ≥ γd (6) if the simultaneous transmissions could ban the capture on a subset of m links, higher transmit power should be assigned at the senders of these links that being oblivious of lower power flows within their sensing range. sinr constraints should satisfy the earlier inequalities at the receivers of the m links. sinr experienced at the receivers of the remaining nm links should exceed ζd as given by 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑆𝑖 ∑ 𝐺𝑗,𝑖 𝑅𝑅 𝑃𝑅𝑗 𝑁 𝑗=1,𝑗≠𝑖 + 𝜂 ≥ ζd (7) further, sinr experienced at the senders should exceed ζd for successful delivery of ack packets, as given by 𝑆𝐼𝑁𝑅|𝐴−𝐷 𝑆𝑖 = 𝐺𝑖,𝑖 𝑆𝑅 𝑃𝑅𝑖 ∑ 𝐺𝑗,𝑖 𝑆𝑆 𝑃𝑆𝑗 𝑁 𝑗=1,𝑗≠𝑖 + 𝜂 ≥ ζ𝐷 (8) cama determines the appropriate minimum transmit power that ensures that the sender can sustain a data rate to reach the intended receipt and the interference level perceived at other nodes can be mitigated thereby higher number of concurrent transmissions can be achieved. 4. simulation results in this section, we present simulation-based studies using ns-2.34 which incorporates the modeling details of the ieee 802.11 mac and phy modules (chen et al. 2007). the phy module includes cumulative sinr computation, preamble and plcp header processing and capture, and frame body capture. the mac accurately models transmission and reception coordination, back off management and channel state monitoring in a structured and modular manner to presents the csma/ca mechanism. all nodes implement the 802.11g technology. the senders and receivers are placed in an indoor environment, and the radio propagation reflects shadowing model with path loss exponent of 4 over the distance (srinivasan and capture aware channel access protocol in wireless network 79 (a) s2 s1 r2 r1 s2 s1 r2 r1 (b) (c) s2 s1 r2 r1 haenggi 2009). parameters as antenna gain and system loss are assumed to be fixed. the default values of configuration parameters are used. compared to preamble detection threshold ζp where preamble detection starts to work, energy detection threshold ζed in terms of snr is higher (lee et al. 2010). table i lists the main parameters in our simulation study. the minimum phy bit rate δ=6mbps is considered in the simulation. hence the minimum sinr which guarantees a reliable packet communication ζd=ζp=5db, as a hardware defined threshold (ieee std. 2007; chen et al. 2007; and lee et al. 2010). the simulation is restricted to one-hop unicast udp traffic. each link carries constant bit rate (cbr) traffic. each simulation runs for 60 seconds with disabled rts/cts virtual carrier sensing. the nodes are configured to be always backlogged with packets to send and each mac data frame to be 1028 bytes long. during the simulation we consider utility in terms of mac layer goodput at the receiver and channel usage efficiency cue in terms of the ratio of mac layer goodput at the receiver to phy layer throughput at the transmitter over a given channel as two distinct metrics. we first have activated one link to find the maximum achievable throughput, goodput and cue of unicast traffic that are approximately 5.19 mbps, 5.05 mbps and 0.973 respectively due to mac/phy layers overhead. low cue indicates that a node has gained access with failed transmission which consumes energy and may cause interference to neighbors over the same channel. fairly shared spectrum efficiency (fsse) index is considered to measure the portion of system spectral efficiency sse that is shared equally among all active users. in case of scheduling starvation, fsse would be zero during certain time intervals. in case of equally shared resources, fs se would be equal to sse. if fsse is maximized, the max-min fairness can be achieved (eriksson 2001). 4.1 performance evaluation of default dcf determining how well carrier sense works in wireless networks is the focus of the simulation studies in this section. we show few examples where the default bandwidth utilization by the 802.11mac is far from optimal setting. carrier sense may sometimes make incorrect access decisions thereby leading to inefficient channel usage, as explained before when exposed and/or hidden terminals are present. utility in terms of goodput and cue are our basic metrics for comparison of link performance across various carrier sensing settings and for measuring the efficiency of default channel access scheme. consider a scenario of a simple four nodes topology in a 100x100m2 indoor environment to compose two wireless links, with symmetric incomplete view of channel state in fig. 2a. the senders are within sense range of each other and within the capture range of their intended receiver. hence, if s1 commences a data transmission, s2 should not block its own transmissions, if any, and vice-versa despite the presence of the ongoing data delivery. they could simultaneously transmit without causing excessive interference at the receivers to disrupt table 1. configuration parameters in the simulation study. parameters configuration value γp 4db γd 10db ζed 10db ζp= ζd 5db δ 6mbps η -96dbm figure 2. three scenarios of two flows sharing a wireless channel. in scenario (a) the intended links could always present, given that senders are in capture range of their respective receiver. scenario (b) portrays hidden terminal case where the intended links could not present simultaneously. scenario (c) shows a topology of two heterogeneous wireless links in terms of channel gain, given that the senders are hidden terminals and s1-r1 link only is always present. http://en.wikipedia.org/wiki/scheduling_starvation s. mustafa 80 successful communication. however, the default mac mechanism based on local channel assessment may mispredict the channel state and consequently waste the transmission opportunity. figure 3 illustrates the normalized utility on two communication links l1 and l2 when the senders depend on local channel observation only or extend the observation to the intended receiver. having a single exposed node potentially drops the overall throughput to 50% of the optimal setting. further complications arise in cases where hidden nodes are likely, scenario (b). suppose s1 is oblivious of ongoing flow from s2 and viceversa. the receivers are in close proximity such that s1 (s2) data transmission and/or r1 (r2) ack transmission could corrupt r2 (r1) reception. the incomplete view of channel state possessed by the sender causes hidden terminal where the absence of a carrier at the sender does not mean a transmission will not collide. both links experience packet loss due to the concurrent transmissions from the hidden terminals. the senders double their contention window size for each failure of data transmission; hereby an obvious drop in the channel utilization is illustrated in fig. 3. misprediction of channel state can drop the utility significantly than if an access decision is extended to the intended receiver. it is obvious that exposed and hidden terminals in the previous scenarios affect the two flows equally. the third scenario in fig. 2 considers a topology of hidden terminals with heterogeneous link quality. s1 and s2 are out of each other’s sensing range, and s2 cannot corrupt r1 reception. however, s1 may initiate a transmission while s2 is transmitting, and disrupt the reception at r2. if s1 occupies the channel for a long time, s2 will starve. the simulation in fig. 3 shows extreme unfairness due to asymmetric interference and diverse transmission opportunity the senders have, which cannot be addressed in this scenario with extending the channel observation to the intended receiver. the unfairness in access opportunity which is undesirable as it can adversely affect delay sensitive applications cannot be addressed without coordination with the surrounding nodes. next, we have evaluated the performance of hidden terminals in the topology of fig. 2c with different traffic load on the wireless links and default mac mechanism. s1 is located within the capture range of its intended receiver while s2 is not always depending on the traffic rate on the interfering link l1. regarding the simulation results in fig. 4 in terms of normalized utility on l2, r2's default mac layer has recorded nothing at traffic load of 3mbps. it could work better at low traffic load wherein the target utility of 1mbps is attained. however, cue metric considered here is not satisfying even with low traffic load. failed transmissions not only alleviate cue and waste energy but also have the potential to corrupt other transmissions. generally, the overall performance claim that the default mechanism doesn't work efficiently even with low traffic load, and simply extending the observation to the intended receiver don't address the limitation. an intuition to maximize the utility and channel usage efficiency is to set up capture aware mac scheme jointly with adaptive transmit power as has been described in section 3. figure 3. normalized utility for the scenarios in fig. 2, in terms of mac's layer good put with default channel assessment at the senders and extended assessment to the intended receiver. 0 0.2 0.4 0.6 0.8 1 1.2 l1 l2 l1 l2 l1 l2 n o rm a liz e d u ti lit y default local observation extended observation capture aware channel access protocol in wireless network 81 4.2 performance evaluation of cama in this section we assess the performance of our proposed framework to mitigate the interference and leverage the spatial reuse via a broad picture of the channel state. in this paper, we assume that information exchange works perfectly as well as the coordination. we conducted three different sets of simulations using different number of sender-receiver pairs use cbr applications on the same channel. 4.3.1 first scenario first, we have conducted a number of simulations using two links for illustrative purpose. the position of nodes is sampled from a two-dimensional distribution in 100x100 m2 indoor environment, as shown in fig. 5. the receivers are deployed in overlapping area of the senders and each tolerates a certain level of interference depending on channel gain and the transmit power. the unicast traffic considered is generated and flowed at a rate sufficient to saturate the medium. the link gain is assumed as being fixed for the duration at which power updates are performed in cama framework. fig. 6a shows the normalized utility experienced by users to illustrate the performance improvement compared to the default dcf at different power levels. the default mac mechanism depends on local channel assessment to decide whether to commence a transmission or to defer it. with default dcf and transmit power level of 19dbm (case1), the senders are hidden and transmit independently. l2 hits the maximum utility with sender in capture range of the intended receiver, whereas l1 is seriously affected by co-channel interference. fig. 6b shows inefficient channel usage by s1, in contrast to s2 that being oblivious of the collisions experienced by r1. figure 4. performance evaluation in terms of normalized utility and channel usage efficiency on l2 for the scenario (c) in fig. 2 at different traffic loads. in case1, traffic loads ʈ1= ʈ2 = 3mbps. case2 considers ʈ1=2mbps and ʈ2=1mbps, while in case3, ʈ1 = ʈ2 = 1mbps. figure 5. network model and topology as used in the simulation. the nodes disseminate the estimated link gain to get broader view and perform joint optimization. 0 0.2 0.4 0.6 0.8 1 1.2 case1 case2 case3 normalized utility s. mustafa 82 higher transmit power can be used to set symmetric carrier sensing between the senders, for example 27dbm (case2). the goal is to silence a sender when the other initiates a transmission. slight different performance: 0.48 and 0.59 of the maximum utility is hit, respectively over the transmission links. collisions occur at r1 when two senders choose the same slot in the contention window to transmit their frames since l1 observes less degree of channel quality; however, this case doesn’t happen as often compared to the case when the senders are hidden from each other. while the default solution for sender out of the capture range of the intended receiver, is to preclude the concurrent transmissions, cama framework shows the existence of optimal transmit power for each node in this scenario to set up asymmetric carrier sense and satisfy sinr constraints, where an obvious improvement in per-flow and aggregate utility is resulted (case 3). by switching for example to transmit power 27dbm at s1 and 19dbm at s2, r1 and r2; s2 is able to carrier sense data transmissions from s1 but not vice-versa. the sender s2 can access the channel during time interval between successive transmitted frames of high power sender within its sensing range. s1 doesn’t block any transmission request from an upper layer that could overlap with existing flow. moreover, if s1’s frame arrives when s2 is in transmission state, s2 will not be able to hear this frame even if its transmission would end very soon after, since the received energy from s1 is below energy detection threshold and s2 would have missed the preamble and plcp header. we confirmed the above argument by tracing the transmission order and inter-transmission time ns-2.34 trace file. albeit the transmission time of the terminals could overlap and eventually the frames based on the traffic size, the concurrent transmissions with the new setting of transmit power don't introduce enough interference to deprive the reception. low power data flow survives the collision with the early interference frame from s2 with 𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅1 ≥10db and hits the maximum utility. higher power flow attains that utility with 𝑆𝐼𝑁𝑅|𝐷−𝐷 𝑅2 ≥5db that guarantees data packet delivery: the first captured frame from s2 tends to survive the collision with a later interference packet from s1. other sinr constraints: 𝑆𝐼𝑁𝑅|𝐷−𝐴 𝑅𝑖 ≥5db and 𝑆𝐼𝑁𝑅|𝐴−𝐷 𝑆𝑖 ≥5db are also met to preclude data and ack packets collision. the senders successfully and fairly utilize the channel wherein normalized utility and cue results in fig. 6 are almost one. the results demonstrate no starving with asymmetric links in power heterogeneous network, and confirm the efficacy of cama framework to provide collision free channel. our simulation results with unicast traffic in contrast to the results have been observed in (rao and stoica 2005) with broadcast traffic over asymmetric sensing links. extended inter frame space (eifs) is supposed to be the main reason for the unfairness in the transmission opportunity. figure 7 shows fsse and sse results wherein cama mechanism obviously has much better spectral efficiency and fairness since it gets off hidden and exposed node scenarios, which are the well known sources for the limitation of the default mechanism. the results shows that coordinated and collaborative framework enables efficient use of the wireless resource and achieves potential improvement over dcf. figure 6. performance evaluation in terms of a) normalized utility, and b) cue of possible scenarios with default dcf and cama. 0 0.2 0.4 0.6 0.8 1 1.2 case1 case2 case3 n o rm a liz e d u ti lit y l1 l2 0 0.2 0.4 0.6 0.8 1 1.2 case1 case2 case3 c h a n n e l u sa g e e ff ic ie n cy c u e l1 l2 (a) (b) s. mustafa 83 4.3.2 second scenario we substantiated the proposed cama mechanism on larger networks of senderreceiver pairs distributed in three-dimensional distribution in 100x100x100m3 indoor environment with disparity in link quality. the probability with which a given frame could be captured by intended receipt is a function of the number of ongoing transmission. the energy received from the intended packet should be higher than total accumulated energy received from the remainder parallel transmission over the capture interval, as defined in the equations (5)-(8). the optimal solution consists of defining a set of willing end-users to transmit, and apply cama mechanism to enable possible concurrent transmissions over the channel. to portray the impact of the aggregate interference, consider a set of four senderreceiver pairs (fig. 8). the distance between the senders s1, s2, s3 and s4 is roughly 40m except the distance between s3 and s4 is 70m. the position of the receivers is sampled such that r3 is within the interference range of s1 and s2, i.e. either r3 is able to decode the foreign frame correctly or not, it could engage with overheard s1 or s2's preamble. r4 is also within the interference range of s1. each flow is sending separate cbr flows to the corresponding receiver on the same channel. in a scenario where the nodes have no knowledge of ongoing hidden transmission (case1), s1 and s2 monopolize the channel and achieve the maximum utility at the expense of the starving nodes s3 and s4, as simulation results in terms of utility and fsse demonstrate in figs. 9 and 10, respectively. average cue in terms of the ratio of the aggregate mac layer good put at the receivers to the aggregate phy layer throughput at the corresponding transmitters, is also presented in fig. 10. we have recorded 10.4mbps an aggregate utility. in the next case (case2), mutual sensing leverages fsse and drops sse with an aggregate utility reaches roughly 7.2mbps. figure 7. performance evaluation in terms of fairness and overall spectral efficiency of the scenarios in fig. 6. figure 8. four sender-receiver pairs distributed in three-dimensional distribution in 100x100x100m3 indoor environments with disparity in link quality. 0 0.1 0.2 0.3 0.4 0.5 case1 case2 case3 fsse (bps/hz) sse (bps/hz) capture aware channel access protocol in wireless network s. mustafa 84 later, we have applied cama mechanism in case 3 to find the appropriate transmit power for each node depending on the interfering environment it sees, such that parallel uplink and downlink transmissions cannot introduce enough interference to corrupt the reception at neighbor send-receive pairs. higher transmit power of 23dbm and 21dbm are set at s3 and s4, respectively due to the link quality each sees. lower power level 13dbm is set at the remaining nodes such that:  s1 and s2 are sensing s3 and s4's traffic but not vice versa.  s1 and s2 transmit independently as well as s3 and s4 since they don't experience peer interference. power heterogeneous network is required here to schedule the possible largest number of successful concurrent transmissions in a single time slot on a single wireless channel. various power levels are set at the nodes to drive a sender in or out of sensing range of others herby schedule the time to transmit, and to satisfy the sinr constraints in equation (5-8). an obvious improvement is portrayed in the simulation based results. energy detection threshold here is set at the default value. individual transmission of s3 and s4 results in snr higher than ζed and ζp, respectively at s1 and s2 that enables either s3's preamble or energy detection, and s4's energy detection if s1 or s2 are listening. it is obvious that their concurrent transmission is also detectable. as for individual links, we observe that power adaptation according to cama maximizes s3 and s4’s utility only, and their saturated traffic especially of s3 reduces obviously other's opportunity to access the channel. nevertheless, obvious improvement in fsse and sse is attained. later in case4, higher ζed is configured such that s1 and s2 do not sense a delivered energy of ongoing transmission when the corresponding preamble is missed. preamble detection only is required to schedule the instantaneous transmissions of senders with homogenous technology. s1 and s2 knock the channel frequently to achieve the maximum utility and subsequently lead to the maximum fairness, though the power levels are not changed. fairly and efficiently channel utilization demonstrates the efficacy of cama framework. figure 9. performance evaluation of different scenarios of four heterogeneous wireless links in terms of channel gain, and distributed in 3d indoor environment. figure 10. performance evaluation in terms of fairness, overall spectral efficiency and average channel usage efficiency. 0 0.2 0.4 0.6 0.8 1 1.2 case1 case2 case3 case4 n o rm a liz e d u ti lit y l1 l2 l3 l4 0 0.2 0.4 0.6 0.8 1 1.2 fsse (bps/hz) sse (bps/hz) capture aware channel access protocol in wireless network 85 4.3.3 third scenario we next randomly distribute 20 and 50 nodes in a 100x100 m2 indoor environment, respectively. each sender has its own receiver chosen randomly from the set of immediate neighbors. a random traffic model at each sender is used with a sufficiently high load such that a sender to be always backlogged with packets to send. a high ζed is configured at each node such that they do not sense a delivered energy of ongoing transmission when the corresponding preamble is missed. in this case study, cama uses a genetic algorithm ga approach for optimal power distribution in a network. the ga is applied to find the optimal power level for each user under different operating conditions based on the constraints defined earlier. the problem is formulated as a multi-objective optimization problem which aims at maximizing the utilization and fairness. ga can work even when the objective function is not exactly known since it relies only on an objective function’s evaluation. it is well-known for their remarkable generality and versatility, and has been applied in a wide variety of settings in wireless networks. the author skipped the details of ga since it is not in the scope of the work rather than for applying an optimization algorithm in cama framework. for details on ga, the author recommends (sivanandam and deepa 2007). figure 11 illustrates the aggregate one-hope utility achieved using 20 and 50 nodes in a random topology using the default dcf and cama as a channel access mechanism. the aggregate utility is found as the aggregate throughput of concurrent one hop communication in the network. the results are averaged over three independent simulation runs. our approach offers clear improvement against dcf. figure 12 illustrates the average value of the fsse and cue using the default dcf and cama in random topology of 50 nodes. performance evaluation in terms of fairness and the channel usage demonstrates the significant improvement using cama based on a broad observation of the channel state. figure 11. the average of the aggregate one-hope utility in random topology of 20 and 50 nodes. the results are achieved using default dcf and cama mechanism. figure 12. performance evaluation of 50 nodes in a random topology in terms of fairness fsse and average channel usage efficiency cue, using the default dcf and cama mechanism. 10 40 70 100 130 160 20 nodes 50 nodes a g g rg a te o n e -h o p e u tl it y m b p s default dcf 0 0.2 0.4 0.6 0.8 1 default dcf cama fsse avg. cue s. mustafa 86 through the investigated use cases, asymmetric carrier sense is introduced and applied in cama. the simulation results portray the benefit of asymmetric sense scenario to time schedule the overlapped traffic without an extra timing signal. cama employs power heterogeneous adhoc network to optimize criteria involving throughput and fairness. this requires obtaining effective spatial reuse while satisfying the interference constraints. 5. conclusion and future work based on local channel assessment, the default dcf mispredicts transmission failures and wastes potential transmission opportunities when they exist. it ignores the possibility of parallel transmissions that make use of the capture effect. it can be ineffective even at low traffic load. this paper proposes a new captureaware mechanism for ad-hoc networks while their benefit is amplified by properly adjusting the transmission power of the nodes based on broad observation of the channel. through simulations, such a mechanism is shown to deliver high network utilization and provide fair access to the media than the standard dcf that these networks use. using the simulator trace files, we have found that tuning transmit power is helpful in setting specific sense scenario and subsequently schedule the access to a shared medium without modifying carrier sense threshold, as has been exploited in cama framework in an interference-free fashion. the nodes cooperate in power updating process via exchanging simple control signaling in power heterogeneous network. further, based on the simulation results, we have found that throwing energy detection carrier sense line offers higher network utility. in environment with heterogeneous technologies, foreign preambles are undetectable and energy detection may results in lower per-flow utility. however, the target sinr by cama at the corresponding heterogeneous nodes is expected to be lower since a receiver would not engage to a first coming foreign preamble, which in turn enables more concurrent transmissions and records higher aggregate utility and overall spectral efficiency. our future work will demonstrate it. we also propose to explore how a similar mechanism can be used to adjust more phy parameters like the phy rate or/and contention window size as mac parameter. conflict of interest the author declares no conflicts of interest. funding no funding was received for this research. references chen q, eisenlohr fs, jiang d, moreno mt, delgrossi l, hartenstein h (2007), overhaul of ieee 802.11 modeling and simulation in ns-2. proc. acm mswim'07: 159-168. ding p, holliday j, celik a (2005), demac: an adaptive power control mac protocol for ad-hoc networks. proceeding of ieee 16th international symposium on personal, indoor and mobile radio communications pimrc 2: 1389-1395. eriksson m (2001), dynamic single frequency networks. ieee journal on selected areas in communications 19(10): 1905-1914. gomez j, campbell at, naghshineh m, bisdikian c (2001), conserving transmission power in wireless ad hoc networks. proceeding of the ninth international conference on network protocols icnp: 11-14. haghani e, krishnan mn, zakhor a (2010), adaptive carrier-sensing for throughput improvement in ieee 802.11 networks. proceeding of ieee global telecommunications conference globecom: 1-6. ieee std. 802.11 (2007), part 11: wireless lan medium access control (mac) and physical layer (phy) specifications, ieee std. 802.11. jamieson k, hull b, miu a, balakrishnan h (2005), understanding the real-world http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=10989 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=10989 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=10989 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=10989 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=5682081 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=5682081 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=5682081 capture aware channel access protocol in wireless network 87 performance of carrier sense. proceedings of 2005 acm sigcomm workshop on experimental approaches to wireless network design and analysis, e-wind’05: 52–57. jeong j, choi s, yoo j, lee s, kim ck (2013), physical layer capture aware mac for wlans. wireless networks 19(4): 533546. kurth m, redlich jp (2009), carrier sensing and receiver performance in indoor ieee 802.11b mesh networks. proceeding of international conference on wireless communications and mobile computing: connecting the world wirelessly iwcmc’09: 726-732. larroca f, rodríguez f (2014), an overview of wlan performance, some important case-scenarios and their associated models. wireless personal communications 79(1): 131-184. lee j, ryu j, lee sj, kwon tt (2010), improved modeling of ieee 802.11a phy through fine-grained measurements. computer networks 54(4): 641-657. li p, geng x, fang y (2009), an adaptive power controlled mac protocol for wireless ad hoc networks. ieee transaction on wireless communications 8(1): 226-233. lim a, yoshida s (2005), a power adapted mac (pmac) scheme for energy saving in wireless ad hoc networks. ieice transaction on fundamentals of electronics, communication and computer sciences, e88-a (7):1836– 1844. ma h, vijayakumar r, roy s, zhu j (2009), optimizing 802.11 wireless mesh networks based on physical carrier sensing. ieee/acm transactions on networking 17(5): 1550-1563. maheshwari r, jain s, das sr (2008), a measurement study of interference modeling and scheduling in low-power wireless networks. proceedings of the 6th acm conference on embedded network sensor systems sensys’08: 141-154. mhatre vp, papagiannaki k, baccelli f (2007), interference mitigation through power control in high density 802.11 wlans. proceeding of ieee infocom: 535-543. mustafa s (2015), interference estimation and mitigation in wireless networks. international journal on computing and digital system 4(4): 233-243. pelechrinis k, guanhua y, eidenbenz s, krishnamurthy sv (2009), detecting selfish exploitation of carrier sensing in 802.11 networks. proc. ieee infocom: 657-665. patras p, qi h, malone d (2014), mitigating collisions through power-hopping to improve 802.11 performance. pervasive and mobile computing 11: 41-55. rao a, stoica i (2005), an overlay mac layer for 802.11 networks. proceedings of the 3rd international conference on mobile systems, applications, and services acm mobisys: 135-148. santhapuri n, nelakuditi s, choudhury rr (2007), capture-aware staggering for concurrent transmissions. proceedings of the 2007 acm conext conference. shah v, gelal e, krishnamurthy sv (2007), handling asymmetry in power heterogeneous ad hoc networks. computer networks 51: 2594-2615. sivanandam sn, deepa sn (2007), introduction to genetic algorithms. springer. srinivasan s, haenggi m (2009), path loss exponent estimation in large wireless networks. information theory and applications workshop: 124-129. takahashi k, obata h, murase t, ishida k (2015), throughput improvement method exploiting capture effect in densely placed wlans. proceeding of the 9th international management and communication imcom'15. http://link.springer.com/journal/11276 http://ieeexplore.ieee.org/xpl/mostrecentissue.jsp?punumber=5061887 http://www.sciencedirect.com/science/journal/15741192 http://www.sciencedirect.com/science/journal/15741192 http://www.sciencedirect.com/science/journal/15741192/11/supp/c http://ita.ucsd.edu/workshop/09/files/paper/paper_628.pdf http://ita.ucsd.edu/workshop/09/files/paper/paper_628.pdf http://ita.ucsd.edu/workshop/09/files/paper/paper_628.pdf http://ita.ucsd.edu/workshop/09/files/paper/paper_628.pdf microsoft word paper 7.docx the journal of engineering research vol. 11, no. 2 (2014) 79-88 0 on directed edge-disjoint spanning trees in product networks, an algorithmic approach a.r. touzene* and k. day department of computer science, college of science, sultan qaboos university, p.o. box 36, postal code 123, al-khodh, muscat, sultanate of oman. received 11 december 2013; accepted 15 september 2014 abstract: in (ku et al. 2003), the authors have proposed a construction of edge-disjoint spanning trees edsts in undirected product networks. their construction method focuses more on showing the existence of a maximum number (n1+n2-1) of edsts in product network of two graphs, where factor graphs have respectively n1 and n2 edsts. in this paper, we propose a new systematic and algorithmic approach to construct (n1+n2) directed routed edst in the product networks. the direction of an edge is added to support bidirectional links in interconnection networks. our edsts can be used straightforward to develop efficient collective communication algorithms for both models store-and-forward and wormhole. keywords: product networks, directed edge-disjoint spanning trees, interconnection networks. אאאאאwא  אאאgא    א wא  א   א א edsts  k   א  אא(n1+n2-1)edst א אאn1n2edstsאk   (n1+n2)   edsts  א אא  א א   kא     kא edstsא א א א   אא  א  אkאא אאw،אאא،אאא *corresponding author’s e-mail: touzene@squ.edu.om a.r. touzene and k. day 80 1. introduction there has been increasing interest over the last two decades in product networks (day, and al-ayyoub 1997; ku et al. 2003; x and yang 2007; imrich et al. 2008; klavar and špacapan 2008; jänicke et al. 2010; hammack et al. 2011; chen et al. 2011; ma et al. 2011; cheng et al. 2013; erveš and žerovnik 2013; govorčin and škrekovski 2014). the cartesian product is a well-known graph operation. when applied to interconnection networks, the cartesian product operation combines factor networks into a product network. graph product is an important method to construct bigger graphs, and plays a key role in the design and analysis of networks. a number of spanning trees of a graph are edge-disjoint if no two trees contain the same edge. edge-disjoint spanning trees (edsts) have many practical applications including enhancing interconnection network fault-tolerance and developing efficient collective communication algorithms in distributed memory parallel computers (fragopoulo and akl 1996; johnsson and ho 1989; touzene 2003). in (ku et al. 2003), the authorshave studied construction of maximum edge-disjoint spanning trees(n1+n2-1) edsts in undirected product network of two graphs, where factor graphs have respectively n1 and n2 edsts. the presented construction is more about showing the existence of a maximum number of spanning trees. they did not provide a straight-forward algorithmic way for their construction. in this paper, we propose a new systematic and algorithmic approach to construct (n1+n2) directed rooted edge-disjoint spanning tree in product networks. we assume that the factor graphs are connected graphs and have respectively n1 and n2 edsts. directed rooted edge-disjoint spanning trees have been discussed for different graphs such as the ndimensional hypercube (johnsson and ho 1989), k-ary-n-cube (touzene 2003), star graphs (fragopoulo and akl 1996), etc. we assume directed edges: if a and b are two nodes in the graph, the edge (a, b) is different from the edge (b, a). directed edges support bidirectional links in interconnection networks. the advantage of our method is the direct use of our trees to develop collective communication procedures in product interconnection networks. the remainder of this paper is organized as follows: in section 2, notations and preliminaries are presented. in section 3, the construction of edge-disjoint spanning trees in product networks is proposed. in section 4, we conclude this paper. 2. notations and preliminaries the cartesian product g =g1×g2 of two undirected graphs g1 = (v1, e1) and g2 = (v2, e2) is the undirected graph g = (v, e), where v and e are given by: v= { <x1, x2> | x1∈v1 and x2∈v2}, and for any u =<x1, x2> and v = <y1, y2> in v, (u, v) is an edge in e if, and only if, either (x1, y1) is an edge in e1 and x2 = y2, or (x2, y2) is an edge in e2 and x1 = y1. the edge (u, v) is called a g1-edge if (x1, y1) is an edge in g1, and it is called a g2edge if (x2, y2) is an edge in e2. x1 is called the g1component of u and x2 is called the g2component. in all what follows we consider directed edges in the sense that the edge (u, v) is different from the edge (v, u). 3. construction of edsts in a product network consider two graphs g1= (v1, e1) and g2 = (v1, e1) having the following properties: the graph g1 contains n1 edst all rooted at x denoted: x1(x), x2(x), … , xn1(x). each xi(x) tree is assumed to be formed of an edge (x, xi), where xi is the ith neighbor of x, and a sub-tree denoted xi(x)/x rooted at xi that spans all the g1 nodes other than x (fig. 1.a). the graph g2 contains n2 edst all rooted at y denoted: y1(y), y2(y), … , yn2(y), each yj(y) tree is assumed to be formed of an edge (y, yj), where yj is the jth neighbor of y, and a sub-tree denoted yj(y)/y rooted at yj that spans all the g2 nodes other than y (figure 1.b). in fig. 1 (a, b) straight lines correspond to g1edges and dashed lines correspond to g2-edges. on directed edge-disjoint spanning trees in product networks an algorithmic approach 81 xi (x/x)yi(y/y) figure 1.a. ith edstxi (x) rooted at x in g1 figure 1.b. ith edst yi (y) at y in g2 and its xi (x) sub-tree. and its yj (y)/y sub-tree. in what follows, we fix a specific node <x0, y0> in g as a desired root for the edst to be constructed. we denote by <xi, y0>, i = 1,…, n1, the n1 neighbors of <x0, y0>in g reached from <x0, y0> via g1-edges, and by <x0,yj>, j = 1, …, n2, the n2 neighbors of<x0, y0> reached from <x0, y0>via g2-edges. for a given node x in g1 and a given tree y in g2, we denote by <x, y> the tree in g1×g2 obtained by fixing the g1-component to x and following the edges of tree y in g2. similarly, <x, y> denotes the tree in g1×g2 obtained by following the edges of a tree x in g1 while the g2-component is fixed to node y. 3.1 the st1 and st2 edst for g we present a construction algorithm of n1+n2-2 edst (without using non-tree edges (ku et al. 2003) for the product graph g: n1-1 edst for g denoted st1(i), i = 2.. n1 and n2-1edst for g denoted st2(j), j = 2.. n2. 3.2 construction of st1(i), for any i 2 ≤ i ≤ n1 1. connect <x0, y0>to its neighbor <xi, y0> (see edge labeled 1 in fig. 2(a)). 2. attach to <xi, y0> the sub-tree <xi(x0)/x0, y0>(see sub-tree labeled 2 in fig. 2(a)). 3. connect <xi, y0> to its neighbor <xi, y1> (see edge labeled 3 in fig. 2(a)). 4. to <xi, y1> attach the sub-tree <xi, y1(y0)/y0>(see sub-tree labeled 4 in fig. 2(a)). 5. to each node <xi, y> in the sub-tree <xi, y1(y0)/y0> (including its root <xi, y1>) attach the tree <xi(x0)/x0, y>(see sub-tree labeled 5 in fig. 2(a)). 6. connect each node <xi, y> in the subtree<xi, y1(y0)/y0> (including its root <xi, y1>) to its neighbor <x0, y1>(see edge labeled 6 in fig. 2(a)). 3.3 construction of the tree st2(j), j=2, .. n2 1. connect <x0, y0>to its neighbor <x0, yj> (see edge labeled 1 in fig. 2(b)). 2. attach to <x0, yj> the sub-tree <x0,yj(y0)/y0>(see sub-tree labeled 2 in fig. 2(b)). 3. connect <x0, yj> to its neighbor <x1, yj> (see edge labeled 3 in fig. 2(b)). 4. to <x1, yj> attach the sub-tree <x1(x0)/x0, yj> (see labeled 4 in fig. 2(b)). 5. to each node <x, yj >in the sub-tree <x1(x0)/x0, yj > (including its root <x1, yj>) attach the tree <x, yj(y0)/y0> (see sub-tree labeled 5 in fig. 2(b)). 6. connect each node <x, yj >in the subtree<x1(x0)/x0, yj > (including its root <x1, yj>) to its neighbor <x1, y0>(see edge labeled 6 in figure 2(b)).in figure 2(a, b), straight lines are g1-edges and dashed lines are to g2-edges. theorem 1: the set {st1(i), 2 ≤ i ≤ n2}∪{st2(j), 2 ≤ j ≤ n2} is a family of (n1+n2-2) edge-disjoint panning trees in g = g1×g2. proof: we show that all the nodes <x, y> of the product graph are reached in the (n1+n2-2) edgedisjoint spanning tree using different edges. x xi y y yj a.r. touzene and k. day 82 figure 2. construction of spanning trees st1 (i) and st2 (j). • case 1: nodes <x0, y> are reached by different g1-edges (<xi, y>, <x0, y>), i = 2, ..., n1, in the different trees st1(i) (edges labeled 6 in figure 2(a)). in trees st2(j), j = 2, ..., n2, these nodes are covered by g2-edges of the sub-trees <x0, yj(y0)/y0> (edges labeled 2 in fig. 2(b)). • case 2: nodes <x, y0>, similar proof as in case 1 (symmetrical). • case 3: nodes <xi, y>, i = 2, ..., n1 are covered in four different ways: 1. in sub-trees <xi, y1(y0)/y0> , i = 2, ..., n1 of trees st1(i) using y1 tree edges ( labeled 4 fig. 2(a)). 2. in sub-tree <x, yj(y0)/y0> , j = 2, ..., n2 of the trees st2(j). these nodes are covered using yj trees edges (j>1), (labeled 5 in fig. 2(b)). 3. in sub-trees <xi(x0)/x0,y>, i = 2, ..., n1of trees st1(j) using xi tree edges (labeled 5 in fig. 2(a)) . 4. in sub-tree <x1(x0)/x0,yj>j = 2, ..., n2 of the trees st2(j)using x1treeedges(labeled 4 in fig. 2(b)). • case 4: nodes <x, yj>, similar proof as in case 3 (symmetrical). • case 5: nodes <x, y>, x ≠ xi , y ≠yjare covered using different g1-edges in sub-trees <xi(x0)/x0,y>, i = 2, ..., n1 of trees st1(i) (sub tree labeled 5 in fig. 2(a)). these nodes are covered using g2-edges in the sub-trees <x, yj(y0)/y0> , j = 2, ..., n2 in the trees st2(j) (labeled 5 in fig. 2(b)). 3.4 the special t1 and t2 edsts for g we present a construction algorithm for the directed edsts in the product graph g denoted t1 and t2. 3.5 construction of t1 1. connect <x0, y0>to its neighbor <x1, y0> (see edge labeled 1 in fig. 3(a)). 2. attach to <x1, y0> the sub-tree <x1(x0)/x0, y0>(see sub-tree labeled 2 in fig. 3(a)). 3. connect <x1, y0> to its neighbor <x1, y1> (see edge labeled 3 in fig. 3(a)). 4. to <x1, y1> attach the sub-tree <x1, y1(y0)/y0>(see sub-tree labeled 4 in fig. 3(a)). 5. to each node <x1, y/yj>, j=1,…,n2in the sub-tree <x1, y1(y0)/y0> (including its root <x1, y1>) attach the tree <x1(x0)/x0, y> (see sub-tree labeled 5 in fig. 3(a)). 6. connect each node <x1, y> in the subtree<x1, y1(y0)/y0> (including its root <x1, y1>) to its neighbor <x0, y1>(see edge labeled 6 in fig. 3(a)). <x0, y0> <x0, y0> <x0, y> < xi,y1(y0)/y0> 1 < x0, y> <x1, yj> <x0, yj(y0)/y0> <x, yj> 2 6 <x, y0> <x, yj(y0)/y0> (b) st2(j), j=2.. n2 <x0, yj> 5 3 4 (a) st1(i), i=2.. n1 1 <xi, y0> <x, y0> <xi(x0)/x0, y0> <xi, y1> <xi(x0)/x0, y> <xi, y> 2 6 5 3 4 on directed edge-disjoint spanning trees in product networks an algorithmic approach 83 <x0, y0> <x0, y0> <x1(x0)/x0 , y1 > <x0, y> < x1,y1(y0)/y0> 1 < x0, y/y1> <x1, y1> < x0, y1(y0)/y0> < x, y1> 2 6 <x, y0> <x/xi, y1(y0)/y0> (b) t2 <x0, y1> 5 3 4 (a) t1 1 <x1, y0> <x/x1, y0 <x1(x0)/x0, y0> <x1, y1> < x1(x0)/x0, y/yj > <x1, y> 2 6 5 3 4 <x1, y/yj> < x/x1, yj > < x/xi, y1> <xi, y/y1 > 7 7 figure 3. construction of spanning trees t1 and t2. 7. connect each node <x1, y> in the subtree<x1, y1(y0)/y0> (including its root <x1, y1>) to its neighbor <x0, y1>(see edge labeled 6 in fig. 3(a)). 8. connect each node <x/x1, y0> in the subtree <x1(x0)/x0, y0> to the node <x/x1, yj>(see label 7 in fig. 3(a)). 3.6 construction of the tree t2 1. connect <x0, y0>to its neighbor <x0, y1> (see edge labeled 1 in fig. 3(b)). 2. attach to <x0, y1> the sub-tree <x0,y1(y0)/y0>(see sub-tree labeled 2 in fig. 3(b)). 3. connect <x0, y1> to its neighbor <x1, y1> (see edge labeled 3 in fig. 3(b)). 4. to <x1, y1> attach the sub-tree <x1(x0)/x0, y1>(see labeled 4 in fig. 3(b)). 5. to each node <x/xi, y1 >, i=1,…,n1in the sub-tree <x1(x0)/x0, y1 > (including its root <x1, y1>) attach the tree <x, y1(y0)/y0> (see sub-tree labeled 5 in fig. 3(b)). 6. connect each node <x, y1 >in the subtree<x1(x0)/x0, y1 > (including its root <x1, y1>) to its neighbor <x1, y0>. 7. connect each node <x0, y/y1> in the subtree <x0,y1(y0)/y0>to the node <xi, y/y1>(see label 7 in fig. 3(b)). note that in t1 the edges (<x, y0>, <x, yj>) are used but in t2(j), 2 ≤ j ≤ n2 , the opposite direction edges (<x, yj>) , <x, y0> ) are use. similarly, in t2 the edges (<x0, y>, <xi, y>) are used but in t1(i), 2 ≤ i ≤ n1 , the opposite direction edges (<xi, y>, <x0, y>) are used. it is easy to see that using a similar proof as in theorem 1, the trees t1 , t2 , st1(i), 2 ≤ i ≤ n2 and t2(j), 2 ≤ j ≤ n2 is a family of (n1+n2) directed rooted edge-disjoint spanning trees in g = g1×g2. to illustrate our construction algorithm, we give a complete example of product of two interconnection networks the 3-cube (3 directed rooted edts’s (johnsson and ho 1989)) and a ring with three nodes (a, b and c) (2 directed rooted edst’s). dark circles represents the root node of the trees and the numbers on the edges a.r. touzene and k. day 84 figure 4. three edsts of the 3-cube and two edsts of the ring (3 nodes). y2 y1 a b c a c b on directed edge-disjoint spanning trees in product networks an algorithmic approach 85 figure 5 (a).spanning tree st1(1). figure 5 (b). spanning tree st1(2). a.r. touzene and k. day 86 figure 5 (c). spanning tree st1(3). figure 5 (d). spanning tree st2(1). on directed edge-disjoint spanning trees in product networks an algorithmic approach 85 figure 5 (e). spanning tree st2(2). represent the dimension number relative to the 3-cube, see figs. 4 and 5. the trees are directed from the root nodes to leave nodes. 4. conclusions in this paper, we presented a new systematic and algorithmic approach to construct n1+n2 (without using non-tree edges) directed rooted edges-disjoint spanning trees for product networks. the previous work on undirected edsts of the product networks (ku et al. 2003) focuses more on the existence of n1+n2-1 but did not provide an explicit algorithmic way for their construction. our n1+n2 edsts can be used straight-forward to develop efficient collective communication algorithms for both models store-and-forward and wormhole using bidirectional links. references chen m, guox, zhai s (2011), the optimal strong radius and optimal strong diameter of the cartesian product graphs. applied mathematics letters 24(5):657-660. cheng cw, lee cw, hsieh sy (2013), conditional edge-fault hamiltonicity of cartesian product graphs. ieee transactions on parallel and distributed systems 24 (10):1951-1960. day k, al-ayyoub ae (1997) , the cross product of interconnection networks. ieee trans. parallel and distributed systems 8(2):109-118. erveš r, žerovnik j (2013), mixed fault diameter of cartesian graph bundles. discrete applied mathematics 161 (12):1726-1733. fragopoulo p, akl sg (1996), edge-disjoint spanning trees on the star network with application to fauttolerance. ieee trans. computers 45(2):174-185. govorčin j, škrekovski r (2014) , on the connectivity of cartesian product of graphs. arsmathematicacontemporanea 7(2):293297. hammack r, imrich w, klavžar s (2011), handbook of product graphs,discrete mathematics and its applications, second ed. crc press boca raton. imrich w, klavžar s, rall df (2008) , topics in graph theory: graphs and their cartesian product, ak peters, ltd., wellesley, ma. a.r. touzene and k. day 88 jänicke s, heine c, hellmuth m, stadler pf, scheuermann g (2010), visualization of graph products. ieee transactions on visualization and computer graphics 16(6):1082-1089. johnsson sl, ho ct (1989) , optimal broadcasting and personalized communication in hypercubes. ieee trans. computers 38(9):1249-1268. klavar s, špacapan s (2008), on the edgeconnectivity of cartesian product graphs. asian-european journal of mathematics 1(1):93-98. ku s, wang b , hung t (2003), constructing edge-disjoint spanning trees in product networks. ieee trans. parallel and distributed systems 14(3):213-221. ma m,xu jm, zhu q (2011),the menger number of the cartesian product of graphs. applied mathematics letters 24(5):627-629. touzene a (2004), optimal all-port collective communication algorithms for the k-ary ncube interconnection networks. journal of systems architecture 50:221-231. xu jm, yang c(2007), fault diameter of product graphs. information processing letters 102(6):226-228. d:\desktop\comp101031(4).xps tjer 2012, vol. 9, no. 2, 36-45 1. introduction despite wind power's intermittent nature, strong growth has occurred in its frequency of install action because of environmental concerns about other sources of electricity, and because of developments in wind turbine technologies. due to wind speed variability, a wind turbine rarely operates at its rated output. therefore, the capacity factor (cf) of a turbine is com___________________________________________ *corresponding author’s e-mail: mbadi@squ.edu.om monly used to estimate its average energy production, which in turn can be used for the economic appraisal of wind power projects at potential sites. moreover, cf models can be used by manufacturers and wind power project developers for optimum turbine-site matching, and for the ranking of potential sites (taiher and li 2008; rau and jangamshetti 2001; jangamshetti and rau 2001; salameh and safari 1992; jangamshetti and rau 1999). comparative study on impacts of power curve model on capacity factor estimation of pitch-regulated turbines mh albadi*a and ef el-saadanyb adepartment of electrical and computer engineering,, college of engineering, sultan qaboos university, p.o box 33, postal code 123, al-khoud, muscat, sultanate of oman bdepartment of electrical and computer engineering, university of waterloo, 200 university ave. w, waterloo, on, n2l3g1,canada received 31 october 2010; accepted 11 january 2012 abstract: the amount of energy produced by a turbine depends on the characteristics of both wind speed at the site under investigation and the turbine's power performance curve. the capacity factor (cf) of a wind turbine is commonly used to estimate the turbine's average energy production. this paper investigates the effect of the accuracy of the power curve model on cf estimation. the study considers three cf models. the first cf model is based on a power curve model that underestimates the turbine output throughout the ascending segment of the power curve. to compensate for the aforementioned discrepancy, the weibull parameters, c and k, which are used to describe wind profile, are calculated based on cubic mean wind speed (cmws). the second cf model is based on the most accurate generic power curve model available in open literature. the third cf model is based on a new model of power performance curve which mimics the behavior of a typical pitch-regulated turbine curve. as the coefficients of this power curve model are based on a general estimation of the turbine output at different wind speeds, they can be further tuned to provide a more accurate fit with turbine data from a certain manufacturer. keywords: wind power, turbine curve modelling, capacity factor estimation cf (c and k) 37 mh albadi and ef el-saadany the amount of energy produced by a turbine depends on the characteristics of both wind speed at the site under investigation and the turbine's power performance curve. wind speed at any site is commonly modeled by the weibull probability density function (pdf), which is characterized by two parameters: the scale factor, c, and the shape factor, k. the turbine's power performance curve can be described by three parameters: cut-in, nominal, and cut-out speeds. this paper presents a novel model to estimate the cf of modern pitch-regulated wind turbines, based on a new turbine's power performance curve and the weibull parameters of wind speed. an early version of this paper was published in albadi and el-saadany (2009). after this introduction, the paper proceeds with a literature survey devoted to wind power output modeling, which includes wind speed modeling, turbine output modeling, and the existing cf model. a new power curve model is then proposed in section iii. in section iv, the two new cf models are derived, and a case study is presented in section v. finally, conclusions are presented. 2. wind power modeling 2.1 wind speed modeling wind speed is commonly modeled by the 2-parameter weibull pdf shown in the following formula: (1) where v is the wind speed in m/s; k is the shape factor, and c is the scale factor. the shape factor, k, is related to the variance of the wind speed; therefore, it is location specific. the weibull parameters can be obtained using the mean and the standard deviation of wind speed at the chosen site. the mean wind speed (mws) can be calculated using the following formula: (2) the above equation can be written as follows. (3) where is the complete gamma function given by (4) the standard deviation of wind speed measurements is calculated using the following equation. (5) the above formula can be written as follows: (6) knowing the mean and standard deviation of wind speed data at the potential site, one can estimate the two parameters of the weibull function by solving (3) and (6) iteratively. 2.2 turbine output modeling the power output of a wind turbine is given as follows: (7) where p(v) is wind turbine output power; is the air density in kilograms per cubic meter (kg/m3); a is the swept rotor area in square meters (m2), and cp is the turbine coefficient of performance. the power curve of a pitch-regulated wind turbine is characterized by three speeds: cut-in, nominal, and cut-out speeds. when the wind speed is below the cut-in speed (vc), the output power is zero, and the rotor cannot be loaded. at its nominal speed, (vr), the power output is at the rated value (pr). in response to the power control mechanisms, the power output remains constant as wind speed increases until the cut-out speed (vf), at which point the turbine will be turned off to prevent mechanical damage. therefore, eqn. 7 can be written as follows: (8) where pasc(v) is wind turbine output power throughout the ascending segment of the power curve. an example of a turbine output is presented in figure 1. for this specific turbine vestas turbine vc, vr, and vf take figure 1. a graphical comparison of the quadratic and cubic models vestas v90 1.8 mw cubic model 1 manufacturer data cubic model 2 quadratic model 2 0 5 10 15 20 25 wind speed (m/s) 1800 1200 600 0 38 comparative study on impacts of power curve model on cf estimation of pitch-regulated turbines values of 3, 12, and 25 m/s, respectively. between vc and vr, the turbine output increases as wind speed increases. manufacturer data show a point of inflection in the ascending segment of the power curve. this point indicates that the turbine efficiency experiences a change at this point. despite the single point of inflection in the ascending power curve segment, cp is not constant for most of the speed range. this phenomenon is demonstrated in figure 2, in which the cp and power output of nordex n90-2300 although the cp is unique for each turbine and, as a result, is difficult to generalize, there have been some attempts to represent the ascending segment of the power curve with a generic model. generic here means that a turbine’s output, as a percentage of rated power, is described using the cut-in and nominal speeds only, without requiring the specific turbine output throughout the ascending segment. generic models available in open literature include linear, quadratic, and cubic ones. below is a brief description of each model: 1) the linear model assumes a linear increase in the turbine output between the cut-in and the nominal speeds. this model, generally, overestimates wind potential. the linear model is given by the following equation: (9) 2) cubic model 1 implicitly assumes a constant overall efficiency of the turbine throughout the ascending segment of the power curve. this model is given by the following formula (dialynas and machias 1989): (10) 3) cubic model 2, considered by jangamshetti and rau (1999) and salameh and safari (1992) is very similar to cubic model 1. the only difference is the absence of vc from this model. cubic model 2 is given by the following equation: (11) 4) quadratic model 1 was originally proposed by justus et al. (1976), and its coefficients were calculated by giorsetto and utsurogi (1983). these coefficients were determined based on the assumption that the output of the turbine follows the cubic model, eq. 10, between (vc+vr)/2 and vr (justus et al. 1976). (12) 5) quadratic model 2, presented by pallabazzer (1995), does not have the (a1v) term of the previous model. (13) as demonstrated in figure 1 and table i, the quadratic model presented in pallabazzer (1995) is the most accurate generic model to represent manufacturer data throughout the ascending segment of the power curve; therefore, it is used in the derivation of one of the two cf models proposed in this paper. a better representation of manufacturer data can be achieved by using a higher order polynomial function as described by the following equation, where n is the order of the polynomial function: (14) however, due to the unique and nonlinear behavior of cp, the coefficients, an, are turbine specific and difficult to generalize. the authors in chang and tu (2007) and celik (2003) use a third order polynomial function to represent the turbine output in the ascending power curve segment, and regression is used to find the coefficients (an). for the manufacturer data presented in figure 1 and table i, the polynomial coefficients take values of 0.5734, -0.3537, 0.0639, and 0.0026 for a0, a1, a2, and a3, respectively. n90-2300 power output cp 0.5 0.4 0.3 0.2 0.1 0 2500 2000 1500 1000 500 0 0 5 10 15 wind speed (m/s) figure 2. coefficient of performance (cp) of n90-2300 39 mh albadi and ef el-saadany 2.3 capacity factor modeling the average power produced by a wind turbine can be calculated by integrating the power curve model multiplied by the weibull function represented by eqn. (1). (15) the capacity factor is the ratio between the average and the rated power of the turbine. the authors in jangamshetti and rau (1999), salameh and safari (1992) used cubic model 2, represented by eqn. (11), the cf model can be written as follows: (16) jangamshetti and rau (2001) compared the values of the cf obtained from (16) to the measured ones, and found that the model significantly underestimated wind potential at the site under study. to compensate for the mismatch between the modeled and the measured values, jangamshetti and rau (2001) investigated the effect of using the root mean square wind speed (rmsws) and the cmws to estimate the weibull function parameters of the wind profile. jangamshetti and rau (2001) found that using the cmws resulted in a better estimation of the cf, at the site under study than when using the original (arithmetic) mws and the rmsws. the mrsws and cmws are defined by the following formulas: (17) (18) however, when one compares the original wind profile, obtained using the arithmetic mws, with that obtained using the rmsws or the cmws, a significant difference in the profile is observed (albadi and el-saadany 2010). actually, using the rmsws or the cmws shifts the original wind speed data towards higher values, as illustrated in figure 3. tai-her and li (2008) solved the integral presented in (16) and devised a cf model as a function of the main turbine curve parameters, vc, vr, and vf, and the two parameters of the weibull function, c and k, that are obtained based on the cmws. (19) table 1. comparison of different models for v90-3mw * based on mws=6 m/s and k=2 n i i n i ii rmsws 40 comparative study on impacts of power curve model on cf estimation of pitch-regulated turbines where is the lower incomplete gamma function given by (20) 3. proposed power curve model figures 4 to 6 present examples of the power curve of pitch-regulated turbines from three manufacturers, vestas, fuhrlander, and the national renewable energy laboratory. similar to figure 2, all turbine curves show their point of inflection in the second half of the ascending segment. this point is attributable to the decrease in turbine efficiency after it has reached its maximum value. because none of existing power curve models consider this property, all of them underestimate the power curve in the second half of the for a more accurate representation of the ascending segment of the power curve, a four order polynomial a0 to 4, of the polynomial function are obtained by solving (21) where original wind data (c=8.0076 & k=2.0765) reproduced using cmws (c=9.8054 & k=2.3503 0 5 10 15 20 25 wind speed (m/s) 0.12 0.08 0.04 0 figure 3. the effect of using the cmws to obtain c and k (data are from) 0 5 10 15 20 25 wind speed (m/s) vestus v90-1.8 m v90-3 m v90-2 m v100-1.8 m 3000 2500 2000 1500 1000 500 0 figure 4. vestus pitch-regulated turbines n80-2500 n90-2300 a70-1500 n100-2500 nordex 0 5 10 15 20 25 wind speed (m/s) 2500 2000 1500 1000 500 0 figure 5. nordex pitch-regulated turbines fl2500-100 fl2500-90 fl2500-80 fuhrlaender 0 5 10 15 20 25 wind speed (m/s) 2500 2000 1500 1000 500 0 figure 6. fuhlaender pitch-regulated turbines 41 mh albadi and ef el-saadany normalized turbine curves for different cut-in and nominal speeds are presented in table 2. an example of the proposed power curve model is presented in figure 7, and a comparison of the proposed model and quadratic model 2 for turbines with vc = 3 m/s and vr = 13 m/s is presented in figure 8. as the coefficients of this model are based on a general estimation of turbine output at different wind speeds, they can be further tuned to obtain a better fit of turbine data from a certain manufacturer. from figure 8, one can conclude that although the power curve in the ascending segment is turbine specific, the proposed model represents manufacturer data better than quadratic model 2, which is the best generic model available in open literature. for v90-1.8 turbine data, presented in table i, the error in annual energy estimation of the proposed model is about 0.3%. 4. new capacity factor model based on the new power curve model proposed in the previous section, the new cf model for pitch-regulated wind turbines is given by the following equation. (22) where f(v) is the weibull pdf, given in (1), and its parameters are based on the mws rather than the cmws. using integration by substitution and by parts (bird 2003), the new cf model is derived as described by albadi and el-saadany (2010). equation (23) can be simplified as follows: (24) vr=15 m/s vr=14 m/s vr=13 m/s vr=12 m/s vr=11 m/s vc = 3 m/s 0 2 4 6 8 10 12 14 16 wind speed (m/s) 1 0.8 0.6 0.4 0.2 0 figure 7. proposed model when the cut-in speed is 3 m/s and different nominal speeds v90-2 m fl2500-90 n90-2300 s70-1500 s77-1500 n100-2500 proposed model quadratic model 2 0 5 10 15 20 25 wind speed (m/s) 1.2 1 0.8 0.6 0.4 0.2 0 figure 8. proposed model versus quadratic model turbine with vc = 3 m/s and vr = 13 m/s (23) 42 comparative study on impacts of power curve model on cf estimation of pitch-regulated turbines table 2. normalized turbine curves for different turbines characgterized by cut-in (vc) and nominal speed (vr) based on the proposed power curve model 43 mh albadi and ef el-saadany because p(vc) = 0 and p(vr) = 1 pu, the above equation could be further simplified as follows: (25) the above equation is independent of a0 and can be used for any higher order polynomial representation of the power curve. for sites at which wind profile is represented by a raleigh distribution function (k = 2), the scale factor, c, can be approximated by the following equation: therefore, eq. (25) can be written as follows: (26) similarly, the cf model based on quadratic model 2 can be calculated by substituting a1 = a3 = a4 = 0, and a2 = 1/(vr2 vc2) in eqn. 26 (27) 5. case study the power curve parameters, represented by vc, vr, and vf, of 12 different turbines are presented in table 3. using both the existing and the proposed models, seven wind speed scenarios, presented in table 4, are used to calculate the cf for each turbine. in table 5, the cf of each turbine is calculated using the three models. from table 4, one can make the following observations and draw certain tentative conclusions: 1. due to the accuracy of the proposed power curve model over that of quadratic model 2, the cf values calculated using the cf model that is based on quadratic model 2 of the turbine curve eqn. 27 is always lower than those obtained using the cf model that is based on the proposed power curve model (eqn. 25) for all turbines. 2. although the existing model eqn. 19 is based on cubic model 2, which underestimates turbine output as demonstrated in table i, cf values calculated using eqn. 27, are always lower than those obtained using the existing cf model, for mws scenarios of 6 and 9 m/s. this result is attributable to the fact that the existing model is based on cmws. using cmws to estimate the weibull function parameters shifts the original wind speed data towards higher values, as illustrated in figure 3. additionally, cf values obtained using eqn. 25 are often lower than those calculated using the existing cf model eqn. 19 due to the same reason. 3. for an mws scenario of 12 m/s, using the cmws in eqn. 19 for cf calculation resulted in lower cf values for most turbines than the values calculated using eqns. 25 and 27. this phenomenon is attributable to the fact that using the cmws in the existing model results in estimating more non-captured wind energy than actually happens. this is due to wind speeds exceeding vf. thus, the value of the cf estimated using the existing model peaks at 10 m/s, figure 9. 6. conclusions the existing cf model is based on a cubic model for turbine output, which underestimates the turbine output throughout the ascending segment of the power curve. to compensate for this mismatch, cmws is commonly used to estimate the weibull function parameters. this paper demonstrates that the use of cmws shifts the original wind speed data towards higher values. as a result, the existing cf model tends to overestimate the cf values for sites with low wind speeds and underestimate the values for other sites with high wind speeds. this paper presents two new cf models. the first one is based on an existing power curve model that yields a better estimation of the turbine output in the ascending segment of the power curve. the other proposed cf model is based on a new power curve model. the coefficients of this power model are calculated based on a general estimation of the turbine output at different wind speeds; therefore, they can be further tuned to obtain a better fit with turbine data from specific manufacturers. applications of the proposed models include wind power potential and turbine-site matching studies. 44 comparative study on impacts of power curve model on cf estimation of pitch-regulated turbines table 3. turbines used in the case study table 4. wind speed characteristics existing cf (eq. 19) cf1 (eq. 25) cf2 (eq. 27) t1, k = 1.50.6 0.5 0.4 0.3 0.2 5 6 7 8 9 10 11 12 mws (m/s) figure 9. cf of t1 for different mws scenarios table 5. cf calculation using different models for different wind speed scenarios 45 mh albadi and ef el-saadany references albadi mh, el-saadany ef (2010), new method for estimating cf of pitch-regulated wind turbines. electric power systems research 80:1182-1188. albadi mh, el-saadany ef (2009), effect of power curve model accuracy on cf estimation of pitchregulated turbines. cigré canada conference on power systems, toronto, ontario, canada. bird j (2003), engineering mathematics. elsevier newnes. celik an (2003), energy output estimation for smallscale wind power generators using weibull-representative wind data. j. of wind engineering and industrial aerodynamics 91:693-707. chang tj, tu yl (2007), evaluation of monthly capacity factor of wecs using chronological and probabilistic wind speed data: a case study of taiwan. renewable energy 32:1999-2010. dialynas en, machias av (1989), reliability modeling interactive techniques of power systems including wind generating units. archiv fuer elektrotechnik 72:33-41. giorsetto p, utsurogi kf (1983), development of a new procedure for reliability modeling of wind turbine generators. power apparatus and systems. ieee transactions on pas-102:134-143. jangamshetti sh, rau vg (2001), normalized power curves as a tool for identification of optimum wind turbine generator parameters. ieee transaction on energy conversion 16:283. jangamshetti sh, rau vg (1999), site matching of wind turbine generators: a case study. ieee transaction on energy conversion 14:1537-1543. jangamshetti sh, rau vg (2001), optimum sitting of wind turbine generators. ieee transaction on energy conversion 16:8-13. justus cg, hargraves wr, yalcin a (1976), nationwide assessment of potential output from wind-powered generators. j. of applied meteorology 15: pallabazzer r (1995), evaluation of wind-generator potentiality. solar energy 55:49-59. salameh zm, safari i (1992), optimum windmill-site matching. ieee transaction on energy conversion 7:669-676. tai-her y, li w (2008), a study on generator capacity for wind turbines under various tower heights and rated wind speeds using weibull distribution. ieee transaction on energy conversion 23: 592602. microsoft word paper 8.docx the journal of engineering research, vol. 11, no. 2 (2014) 89-97 analysis and design of a new dual band microstrip patch antenna based on slot matching y-shaped r. walia, s. ghnimi*a, a.g. handb and t. razbanb a high frequency electronics circuits and systems, science faculty of tunis, university tunis el manar, tunis, tunisia b ietr polytech nantes, universite de nantes, site de la chantrerie, rue christian-pauc, nantes, france received 4 june 2013; accepted 30 december 2014 abstract: a new compact microstrip slot antenna with y-shaped coupling aperture is proposed. this antenna is based on a rectangular form with a microstrip fed line for excitation, and consists of slots on the edge of the radiation patch to provide dual-band operation. the design and simulation of the antenna were performed using cst microwave studio simulator. a prototype of the proposed antenna has been constructed and tested. in order to validate the return loss of the prototype antenna, the experimental results are presented. the results show two available bands with –10 db band s at 2.7 ghz and band c at 5 ghz. also, good radiation performance and antenna gain over the two frequency ranges have been obtained. computation results confirm the experimental findings. keywords: microstrip antenna, dual band antenna, slot antenna, cst microwave studio simulator. אאאאאא y  אאאא   אw אאאאאאyאאאk אאאאאkא אאאאאcstkא kאאkאאאאא  j10s2.7c5kא kאאאאאkאא א w א א א،א  א ،אא  ،א   אcst *corresponding author’s e-mail: said.ghnimi@fst.rnu.tn r. wali, s. ghnimi, a.g. hand and t. razban 90 1. introduction the demands on radar application performance have increased rapidly over the last few years. future requirements for advanced radar antennas include the need for dual-band operation (allen et al. 2005; christina et al. 2009; zhilong et al. 2012; qu et al. 2006; andan et al. 2009; khunead et al. 2007; khaled and saad 2008; abdelaziz 2006; gai et al. 2010; saidatul et al. 2009; anden et al. 2010; balanis 1997 and ghnimi et al. 2013). it is essential to integrate multiple bands for use in one device. in the literature, numerous antennas for dual band radar applications have been used. in (allen et al 2005), a dual tapered meander slot antenna is designed for use as an element in a phased array antenna system. in (christina et al. 2009), a novel broadband monopole antenna designs with dual-band circular polarization are presented. in (zhilong 2012), the author presents a method for reducing radar cross section of a patch antenna in two frequency bands. this method is based on the metamaterial composed of double-square-loop structure, which is loaded around the exciting source. in (qu et al. 2006), a novel dual-band dual-polarized micro-strip antenna array is presented. in adnan et al. 2009), a loop-loaded printed dipole antenna array is introduced for a dual-band radar application operating at 3.0/5.5 ghz bands. in (khunead et al. 2007), a rectangular slot antenna composed of two conductor strips located in a slot cut in the ground plane was investigated for dual frequency. several approaches are developed in order to design multi-band micro-strip patch antennas such as slot matching concept (khaled and saad 2008), as two different radiating elements connected together through a matched section in abdelaziz 2006), as a pair of u-shaped strips in (gai et al. 2010), as a fractal planar inverted f antenna in (saidatul et al. 2009). in this paper, the proposed antenna presents a simple technique; indeed the use of a slot on the patch with a simple substrate of low cost and larger height allows a good performance. the main purpose of the antenna designing is to obtain a directional radiation and a high gain with widest bandwidth as much as possible. the organization of this paper is as follows: in section 2; the design of the rectangular microstrip patch antenna is given. the parametric study for this proposed antenna is presented in section 3; the fabrication and the measurement of s-parameter, radiation pattern and gain for validation are in section 4. section 5, contains conclusions and recommendations for further studies. 2. antenna design designing a simple patch antenna involves the choice of materials before calculating the dimensions of the patch and the width of the feed line (which depends on the input impedance). all are simulated using cst microwave studio software. the model is then adjusted and optimized for best results. antennas for radar applications must have a large bandwidth and to ensure this condition, the substrate should be thick and having low dielectric constant. this ensures the desirable qualities; better efficiency, larger bandwidth and better radiation (balanis 1997). this process provides a reasonably accurate starting point but it does not provide the final dimensions of the patch (balanis 1997). the dimensions of the antenna for a chosen resonant frequency can be calculated using cavity model (saidatul et al. 2009). a prototype of the antennas was designed (fig. 1) and its characteristics is verified with theory and simulation results. the structure of this antenna is configured to cover the band s for radar application (2 ghz 4 ghz). the studied antenna is printed on the fr4_epoxy which is a very commonly used substrate, which allows a slight compromise between size and losses of the antenna. the metallization level is t = 0.035 mm with a dielectric loss tan δ = 0.002, a relative permittivity εr = 4.6 and a thickness h = 1.58 mm. analysis and design of a new dual band microstrip patch antenna based on slot matching y-shaped 91 figure 1. geometry of the proposed antenna: (a) top view, (b) side view. figure 2. simulated reflection coefficient as a function of frequency. the transverse dimension of this antenna is 50 × 45 mm2. the patch size is 21 × 19.5 mm², and it is excited by a feed line with dimension 15.5 × 3 mm². the calculations are done for a patch antenna with a ground plane. by removing a part of the ground plane and reducing it to a dimension w1=14.5 mm printed on a portion of the underside of the substrate provide an increase of the bandwidth. after the simulation of this antenna with dimensions given above, we found the reflection coefficient, as a function of the frequency (fig. 2). it can be seen from fig. 2 that the antenna resonates at 2.89 ghz with a minimum reflection coefficient of -24.73 db. the reflection coefficient is less than -10 db between 2.4 and 4.26 ghz. this bandwidth obtained for the optimized structure covers the s-band radar applications. 3. parametric study of dual band antenna our objective is essentially the design of a dual-band antenna to cover two radar bands. indeed, several techniques are used including the disturbance slot geometry in the radiating element, which allows the creation of new resonances (adnen et al. 2010 and ghnimi et al. 2013). based on this idea, we created slots in the simulated antenna geometries given in the previous section. the band that we want to add with the insertion of a slot in the radiating element is the c band radar applications which uses frequencies ranging from 4 to 8 ghz, while maintaining the first band around 3 ghz (fig. 3). r. wali, s. ghnimi, a.g. hand and t. razban 92 figure 3. geometry of the dual band antenna and slot parameters: (a) top view, (b) side view. figure 4. simulated reflection coefficient of the antenna with slot for different values of ‘k’. figure 5. simulated reflection coefficient of the antenna with slot for different values of ‘e’. analysis and design of a new dual band microstrip patch antenna based on slot matching y-shaped 93 frequency (ghz) figure 6. simulated reflection coefficient of the antenna with slot for different values of ‘g’. frequency (ghz) figure 7. simulated reflection coefficient of the antenna with slot for different values of ‘m’. frequency (ghz) figure 8. simulated reflection coefficient of the antenna with slot for different values of ‘i’. after insertion of the slot, we are interested in optimizing the printed antenna for a radar antenna in both bands. to do this, we used a parametric study of slots dimensions to see r. wali, s. ghnimi, a.g. hand and t. razban 94 which one affects the most the performances of the proposed antenna. the dimensions of the proposed antenna are chosen to obtain the lowest return loss and the widest bandwidth and one parameter is changed at a time while the others are kept constant. figures 4 to 8 show the parametric effect of the slot on electrical characteristics of the proposed antenna. all the results in these figures show that this antenna has two resonant frequencies (c and s) for radar application. it can be seen from fig. 4 that with an increase in the parameter ‘k’, the bandwidth of the lower band decreases, while the upper resonant frequency shifts down. on the other side the parameter ‘e’ does not affect the impedance characteristics significantly, as shown in fig. 5. however, a slight degradation in performance at the upper frequencies can be observed. however return loss decreases by 4db when e increases from 6 mm to 10 mm at f=4.8 ghz. that is why, the choice of k = 14 mm and e = 10 mm are the most appropriate for the best performance. figures 6 and 7 show that the variation of ‘g’ and ‘m’, allows us to have a very significant effect on the coverage of the second bandwidth of the antenna and its adaptation. indeed, when the parameters 'm' and 'g' increase, the reflection coefficient in the second band increases and the resonant frequency decreases. thus, for a better performance, it is advisable to choose m = 5.5 mm and g = 10 mm. the parameter ‘i’ as shown in fig. 8, is opposite to the aforementioned effect of the other parameters. indeed, we note that the bandwidth is not influenced by the different values of 'i'. however, the reflection coefficient is slightly affected in the second band. from the point of view of the resonant frequencies, the microstrip patch antenna is considered a single resonant cavity functioning mainly at the band s. but, with a good choice of the y shaped slot dimension; we can give other possibility to cover other c radar bands. following the simulation of the optimized antenna with the final dimensions of the slot, we were able to get a dual-band aspect for better performance of the antenna (fig. 9). it covers both the first band from 2.27 ghz to 3.13 ghz lying in the s band radar and the second band between 4.4 ghz and 5.8 ghz which is in the band c for radar application. moreover, the antenna admits respectively a reflection coefficient of -21.7 db at 2.63 ghz and 4.65 ghz -33.22 db at the first and second operating bands. 4. simulated and measured results to verify the validity of the analysis presented in the previous section, measurements were performed to study the electric and radiation characteristics for the proposed antenna. figure 10 shows the prototype of the performed monopole dual band antenna. it is printed on a fr4_epoxy substrate with dimensions of 50 × 45 mm2, the relative permittivity εr = 4.6 and a thickness h = 1.58 mm. 4.1 measured reflection coefficient figure 11 illustrates the evolution of the reflection coefficient simulated and measured within the frequency band of 1-12 ghz. according to fig. 11, we observe a good agreement between the simulated and measured reflection coefficients around the s band. however there is a slight difference in c band. this confirms that the variation of the second band is affected by the insertion of the slot. indeed, the variation of the measured reflection coefficient shows two frequency bands: the low-band resonant frequency is located at about 2.7 ghz with -10 db impedance bandwidth from about 2.3 to 2.5 ghz. however, the high-band resonant frequency is located at about 5 ghz with -10 db impedance bandwidth from about 4.9 to 6.0 ghz. the resonant frequencies are adapted for -25 db at 2.7 ghz and -30 db at 5 ghz respectively. furthermore, we see the emergence of two frequency bands at 8 and 11 ghz with very low adaptation. however, we believe that the appearance of these additional bands is attributed to the effect of the bnc connector. analysis and design of a new dual band microstrip patch antenna based on slot matching y-shaped 95 frequency (ghz) figure 9. evolution of the simulated reflection coefficient as a function of frequency. figure 10. photograph of the proposed dual band antenna: (a) top view, (b) side view. s parameter magnitude (db) frequency (ghz) figure 11. simulated and measured reflection coefficient of the proposed antenna. r. wali, s. ghnimi, a.g. hand and t. razban 96 (a) (b) figure 12. simulated and measured radiation patterns of antenna in e-plane at (a) 2.63 ghz and (b) 4.65 ghz. 4.2 radiation patterns the simulated and measured e-plane (copolarization) radiation patterns at resonant frequencies are shown in figs. 12(a) and 12 (b), respectively. the measurement results show an excellent agreement with the calculated data at the frequency 4.65 ghz. at 2.63 ghz, the measurement results are also correlated with calculations. the radiation patterns in e-plane are nearly bidirectional at 4.65 ghz, but slightly distorted at 2.63 ghz. however, the observed differences in the form of radiation patterns are substantially due to the differences in the amplitude scales in the plots. for the simulated radiation pattern at 2.63 ghz, the main lobe gain is 2.4 dbi in perpendicular direction to the antenna surface (position 180° in the plot) and the angular width is 84.4 degrees. but for the measured radiation pattern, the main lobe has an offset of 20° from perpendicular direction (position 160° in the plot) and the angular aperture is 70 degrees. the radiation of the opposite side is also shifted by about 20 degrees. from fig. 12(b) we can see that the measured maximum side lobe level is -1.2 dbi; the simulated maximum side lobe level is -1.7 dbi. note also that the main lobe magnitude for simulated radiation pattern is -2.2 db with an angular width of 78.4 degree in 169 degree of direction, which covered 90 degree in the same direction for the measured radiation patterns. 5. conclusions in this work, the new design of dual-band microstrip antenna has been successfully implemented and studied. this antenna is achieved by etching a y-shaped slot matching concept on the radiating element and have been functioning in two bands s and c . the electrical and radiation characteristics are given by the simulated and measured s-parameters and the radiation patterns of both bands. good agreement between simulated and measured 5 10 15 20 30 210 60 240 90 270 120 300 150 330 180 0radiation pattern (φ = 90°) θ (° ) e measured 5 10 15 20 30 210 60 240 90 270 120 300 150 330 180 0radiation pattern (φ = 90°) θ (° ) e measured analysis and design of a new dual band microstrip patch antenna based on slot matching y-shaped 97 results has been achieved in the first operation band s. but, in the second operation band c, there is a slight difference explained by the insertion of the slot. this insertion creates other multi-resonant cavity, and each cavity has a new frequency band. the first band is appropriate for the original antenna and the other band is appropriate for the proposed antenna. references abdelaziz aa (2006), bandwidth enhancement of microstrip antenna. progress in electromagnetics research, pier 63:311–317. adnan s, mustafa hb, ucar, yunus ee (2009), loop-loaded printed dipole array design for a dual-band radar application. antennas & propagation conference, lapc loughborough. 529–532. adnen r, ghnimi s, ali g (2010, electrical characteristics of a dual-band microstrip patch antenna for gsm/ umts / wlan operations. international journal of communication networks and information security (ijcnis) 2(1):54-59. allen cm, eldek aa, elsherbeni az, smith ce (2005), dual tapered meander slot antenna for radar applications. ieee transactions on antennas and propagation 53(7):23242328. balanis ca (1997), antenna theory: analysis and design. 2nd edition, john wiley & sons, new york, usa. christina f, jou jww, chien-jen w (2009), novel broadband monopole antennas with dual-band circular polarization. ieee transactions on antennas and propagation 57(4):1027-1034. gai s, jiao yc, yang yb, li cy, gong jg (2010), design of a novel microstrip-fed dual-band slot antenna for wlan applications. progress in electromagnetics research letters 13:7581. ghnimi s, rawia w, ali g, razban t (2013), a new design of an s/x dual band circular slot antenna for radar applications. journal of microwave power and electromagnetic energy 47(2):138-146. khaled eem, saad aar (2008), multiwide band compact microstrip patch antenna based on slot matching. progress in electromagnetics research c 4:169–177. khunead g, nakasuwan j, songthanapitak n, anantrasirichai n (2007), investigate rectangular slot antenna with l-shaped strip. piers online 3(7):1076-1079. qu x, zhong s, zhang ym (2006), dualband dual-polarised microstrip antenna array for sar applications. electronic letters 42(24):1376-1377. saidatul na, azremi aah, ahmad rb, soh pj, malek f (2009), multiband fractal planar inverted f antenna (f-pifa) for mobile phone application. progress in electromagnetics research b 14:127–148. zhilong m, cheng h, qin z, mingbo p, xiaoliang m, xiangang l (2012), design of a patch antenna with dual-band radar cross section reduction. international conference on microwave and millimeter wave technology (icmmt) 5:1–3. c:\users\raoh\desktop\paper 3.xps the journal of engineering research vol. 12, no. 1 (2015) 32-43 investigations on hysteresis-based current control techniques for grid connected photovoltaic systems v. rajini department of electrical and electronics engineering, sri sivasubramaniya nadar (ssn) college of engineering, kalavakkam, india received 21 july 2014; accepted 16 december 2014 abstract: this paper focuses on delivering an effective performance qualification (pq) control strategy for a grid connected photovoltaic (pv) system. a pv-based distributed generation system with a capacity of 10 kilo volt amps (kva) was connected to the utility side of a power grid operating at 415 v. the power grid was modelled with a capacity of 100 kva, 11 kv, and 100 km transmission line length. the pq control strategy was implemented with three hysteresis-based current control techniques. to overcome the drawbacks of conventional hysteresis-control techniques, adaptive tuning of the hysteresis band was carried out, and a vector-based hysteresis current control is proposed to improve the results. keywords: pq control, pv-based distribution grid system, vector-based hysteresis current control, adaptive hysteresis current control, conventional hysteresis current control. pq 10 41510011 100pq pq corresponding author’s e-mail: rajiniv@ssn.edu.in v. rajini 33 1. introduction the growing demand for power and an impending energy crisis have revealed the acute need for the expansion of clean energy to better meet the world’s heightened power needs. integration of solar power into the distribution grid (dg) is one of the greatest concerns in engineering. efficient harnessing of this energy by a photovoltaic (pv) system and directly feeding it into the grid has always been the key solution to this concern. but such a system’s low efficiency and poor controllability have remained the major drawbacks of a pv-based dg system (blaabjerg et al. 2004). for a grid-connected pv system, both a voltage source pulse-width modulation (pwm) inverter and a system of closed loop power control are required. three-phase pwm voltage source inverters have been used in high performance ac motor drives, high power factor converters, active filters, etc. (rahim and mekhilef 2002). in most of these applications, the magnitude of alternating current (ac) voltage is a function of the voltage across the direct current (dc) link capacitor engaged in the input side of the inverter (kwon et al. 1998; mohseni and islam 2010; vahedi et al. 2011). regulated pwm voltage source inverters where the inner current loop determines the control over the injected power are widely employed in grid-connected dg systems in order to achieve independent control over the active and reactive power, and improve load dynamics and power quality (chitti babu et al. 2008; vahedi et al. 2011). the application of dynamic control for power injection is examined in the current study because of the varying demand for power in the utility grid due to the erratic variation in load requirements (vahedi et al. 2011; babu et al. 2012). employing a performance qualification (pq) control strategy can aid with the varying power requirements of the utility grid, and hence generate reference vectors for the inner current control loop of the voltage source pwm inverter. effective control of the grid parameters to determine the real and reactive power references is presented and the simulation results analysed. a vector-based hysteresis control with adaptive tuning is proposed in this paper. the basic model of a grid-connected pv system employs three different stages, including the utility grid, the dg source side control, and the three-phase pulse-width modulation voltagesource invertor (pwm-vsi). the utility grid includes a three-phase power source, the distribution line, and a three-phase resistive inductive (rl) load which varies periodically. figure 1 is a block diagram of a grid-connected pv system. the pv source—the solar array—is modelled mathematically through the basic parameters of a pv cell (rahim and mekhilef 2002). an incremental conductance-based maximum power point tracking (mppt) algorithm was employed to obtain maximum power from the source at all stages. the output of the solar array was interfaced through a dc-dc boost converter and connected through a dc-link to the inverter. the three-phase pwm-vsi was connected to the utility grid in order to inject a three-phase current which was controlled by a pq controller. figure 2 shows the inverter topology connected to the grid with a filter. the solar photovoltaic (spv) system is modelled using the five-parameter model with equations 1–4. a detailed modelling of the spv system can be found in references 10 and 11. i exp ∗ 1 1 i i exp (2) 0 3 100 4 2. performance qualification (pq) strategy the paper aims to deliver a generic pq figure 1. block diagram of a grid-connected photovoltaic (pv) system. figure 2. inverter topology. control technique for any dg source. the time variables of threephase current and voltage vectors are transformed into space vectors in a rotating reference frame (d-q) (fig. 3). hence, the real and reactive powers (p and q) are controlled in the synchronous reference frame (d-q). in order to determine the phase angle (ɵ), clark’s transformation was performed to obtain the voltage and current vectors in the stationary frame (α-β). this can be given by the clark’s transformation matrix for a three-phase system. where ct is the clark’s transformation matrix. the current and voltage vectors are hence given by eqns. (7) and (8). vdc r+jxl r+jxl r+jxl s1 s3 s5 s6s4s2 va vb vc ia ib ic 34 v. rajini 35 figure 3. stationary and rotating frames. iii 23 1 12 120 √32 √321√2 1√2 1√2 5 iii 6 park’s transformation was employed to convert the stationary two-phase (iα, iβ) system into a rotating (id, iq) system using the unit vector angle (ɵ) (brod and novotny 1985) as follows: cosθ sinθcosθ sinθ 7 using the same transformation, the voltage vectors (vd) and vq were obtained using eqn. (8). cosθ sinθcosθ sinθ (8) the real and reactive powers are given by eqns. (9), (10) and (11). 9 (10) (11) the d-axis of the synchronous frame is aligned to rotate with the voltage vector (vd) and, correspondingly, the value of vq becomes zero. hence, the power eqns. (9) and (10) can be rewritten as eqns. (12) and (13). (12) 13 the value of vd corresponds to the voltage magnitude, which is a constant. hence, precise control of the real and reactive powers to be injected can be achieved by controlling the magnitudes of id and iq. two different constants, kp and kq, are considered such that the reference current vectors along the direct and quadrature axes are given by eqns. (14) and (15), ∗ 14 ∗ 15 where kp and kq are the real and reactive power coefficients of the pv source, to be tuned in accordance with the real and reactive power demands, respectively. by performing suitable inverse transformations, the above current reference variables are transformed to the a, b, c reference frame (iaref, ibref, icref), and are used in the inner current control loop in order to generate the pwm pulses for the vsi. the pq control system can be so defined in order to deliver preferred ratios or desired values of real power (pi) and reactive power (qi). in the case of a pv-based dg source of maximum capacity (ppv), the values of kp and kq are selected so that the total power conservation stays inviolate. this can be ensured by following eqn. (16). 16 3. current control the control of the inverter current by using the generated three-phase reference wave is employed for the calculated injection of power into the utility grid. i d βq d θ i β i q α θ i s i α investigations on hysteresis-based current control techniques for grid connected photovoltaic systems 36 3.1 conventional hysteresis current contro ller a conventional hcc technique is preferred for its dynamic performance and ease in implementation. in this technique, an error signal is generated by direct comparison of the inverter current with the reference current value (fig. 4). the switching pulses are so generated that the current error is kept within a predefined hysteresis band. this method aids the control of switches in an inverter by asynchronously varying the current through the load so that it tracks a reference current signal. figure 4. gate pulse generation but one major drawback of this technique is the occurrence of a high switching frequency at a lower modulation index due to a lack of coordination between three-phase hysteresis comparators (esram and chapman 2007). hysteresis band violations of up to twice the bandwidth are also prone to occur. one of the much squared solutions for this problem is the use of a dual hysteresis band wherein there are two bandwidths: δ and , such that < δ. this method provides a modest constructive effect in case of transients and peak overshoots. figure 5 shows the error band region. 3.2 adaptive bandwidth hysteresis current control (hcc) since the switching frequency in the hcc technique depends on the rate of change of the line current, it varies with the inverter current waveform (ic). this current also varies with the line inductance (l) and the dc link voltage (vdc). the bandwidth of the hysteresis band determines the tolerance of current error. thus, figure 5. the error band region for conven tional hysteresis current control (hcc). the bandwidth determines the rate at which the inverter current gets switched within the hysteresis band and, hence, the switching frequency (fs). it can be concluded that the relation between the bandwidth and switching frequency is: δα 1 f 17 even though the increase in switching frequency (fs) can improve the inverter current (ic), it augments the switching losses in the system, which is undesirable. therefore, the bandwidth (δ) is selected to balance the switching losses, switching frequency (fs), and inverter current (ic). in order to adaptively tune the bandwidth (δ), the fact that the switching frequency (fs) is a function of the rate of change of the line current ( ) must be considered. thus, in a switching condition where the switch on time is ton and the switch off time is toff, the time period (t) and switching frequency (fs) are given by: 1 18 the switching of voltage and the line current (il) for one of the phases at the switching frequency mentioned in eqn. (18) is illustrated below: iref iact ie gate pulse iref iβ iα ie δ δ δ v. rajini 37 figure 6. switching of inverter line current within the hysteresis band. figure 6 is analysed in order to determine the adaptive tuning of the hysteresis bandwidth (vahedi et al. 2011). considering the differential rise and fall in line currents, the line inductance (l), phase voltage (vph) and dc input voltage (vdc) are as follows: 2 (19) 2 (20) by approximating values in fig. 7, the equations of the bandwidth are as follows: 2 21 2 22 thus, by subtracting eqn. (21) from (22), eqn. (23) can be derived: 1 23 by adding eqns. (21) and (22), the following is derived: 4 24 by analysing eqns. (19), (20), (23), and (24), bandwidth (δ) can be represented as 1 12 4 251 2 3.3 vector-based hysteresis current control (hcc) the selection of optimum bandwidth (δ) was carried out in the case of the adaptive hcc technique. but this technique carries a prime disadvantage of exhibiting high switching frequencies at the zero crossing points of the reference current. even though this approach yields an almost fixed switching frequency and minimizes inverter current oscillations, this improved performance is usually obtained at the expense of extra signal processing and control complexity requirements, which compromises the simplicity of conventional hccs (esram and chapman 2007). moreover, we may observe stability problems and limited transient performance in these adaptive tuning techniques. as an alternate to achieving reduced switching frequencies, vector-based hcc techniques can be used. optimal vector-based methods systematically select zero and non-zero voltage vectors to follow the reference current vector, thereby significantly reducing the inverter switching frequency. a vector-based hcc that employs two levels of hysteresis comparators of bandwidths δ and δ+ δ, to generate an optimal switching pattern that restricts the current error vector within a hexagonal tolerance region, is discussed. vdc/ 2 -vdc/2 2 investigations on hysteresis-based current control techniques for grid connected photovoltaic systems 38 table 1: switching table for vector hcc. sa1 sb1 sc1 sa2 sb2 sc2 000 001 010 100 011 101 110 111 100 v01 v01 v01 v1 v01 v6 v2 v01 110 v00 v00 v3 v1 v00 v00 v2 v00 010 v01 v01 v3 v01 v4 v01 v2 v01 011 v00 v5 v3 v00 v4 v00 v00 v00 001 v01 v5 v01 v01 v4 v6 v01 v01 101 v00 v5 v00 v1 v00 v6 v00 v00 this technique retains the simplicity and ease of implementation showcased by the conventional hcc. furthermore, it offers a considerably reduced switching frequency and inverter current oscillations, thereby exhibiting improved transient and steady state performance. figure 7. error band region for vector hysteresis current control (hcc). in order to understand the theory behind vector-based hcc schemes, an understanding of the basic voltage equations of the three phases in case of a standard inverter topology is necessary (fig. 2). 13 2 13 2 13 2 (26) figure 8. tracking current error. iref iact iβ iα δ δ +∆δ ie v2 v3 v4 v5 v6 vopt v0,v7 , v1 v. rajini 39 the vector form of the above equations based on clark’s transformation can be given as 1 27 where vn is the output voltage vector and eo is the vector notation of back-electromagnetic field (emf) (kwon et al. 1998; mohseni and islam, 2010). the current error can be noted as the difference between the reference and the actual values of current. (28) using eqns. (27) and (28), we can denote the derivative of error current as follows: 29 by eliminating the r term in equation (29), the equation for the voltage vector corresponding to the optimal switching state, such that the error current vector ie is zero, results. 30 equations (29) and (30) denote the derivative of the current error vector (fig. 9). this derivative can be written as (31) in the case of steady state conditions, vectorbased hcc techniques usually follow control strategies to apply voltage vectors such that the tracking error is kept minimal. in figure 8, where the optimal voltage vector lies in sector 1, of the six non-zero and two zero vectors (vectors v0–v7) (fig. 9). the switching states corresponding to the two zero vectors and the adjacent non zero vectors (v1 and v2) are the best to be applied. in contrast, during transient conditions it may be necessary to switch back the actual current iact within the hysteresis band as quickly as possible irrespective of the tracking error value. the switching table for this technique is given in table 1. 4. simulation results the pq control strategy is employed using the three current control techniques, in a three-phase 415 v system, rated 11 kv at the generation end and transmission lines of length 100 km. a resistor–capacitor circuit (rc) filter is designed to eliminate a band of harmonics at the inverter end to improve power quality. figure 10 represents the phase voltage at the point of common coupling (230 vrms). to illustrate the effect of the proposed controller, the load power, power supplied by the grid, and that generated by spov are varied and the details of variations are given in table 2. when the real power demand is 8000 watts (w), a peak current of 29.55 amps (a) is supplied to the load out of which 14.49 a peak comes from the grid. when the real power demand changes to 16,000w as indicated in table 2, a peak current of 49 a is supplied to the load out of which 43 a comes from the grid. both these current wave forms are given in figs. 11(a) and (b). similarly, the power demand and the power delivered by the main grid are shown in the figs. 12(a) and (b). the sum of the power delivered by the main grid and dg source equals the power demand. figure 9. space vector representation. the load is rated to demand 8000 w of real power and 12,000 volts-amperes reactive (var) of reactive power for the time period of 0–0.1 s. an additional load is connected to the system at 0.1 s, so that the demand rises to 16,000 w and 24,000 var for the time period 0.1–0.2 s. hence the pq controller (section 2) allows the dg system to inject 8,000 w of real power and 6,000 var of investigations on hysteresis-based current control techniques for grid connected photovoltaic systems 40 reactive power for 0–0.1 s (kp = 1; kq = 0.5). once the demand rises, the controller allows the system to deliver its total capacity (10,000 w) as real power (kp = 0.625; kq = 0). table 2. power allocation. periods/parameters 0 – 0.1 0.10.2 pl(w) 8000 16000 ql(var) 12000 24000 pi(w) 6000 10000 qi(var) 6000 0 pg(w) 2000 6000 qg(var) 6000 24000 kp 1 0.625 kq 0.5 0 figure 10. voltage at the point of common coup ling. figure 11(a). load current required. figure 11(b). current drawn from the main grid. figure 12(a). real and reactive power demand of the load figure 12(b). real and reactive power delivered the main grid. the inverter current waveform and the real and reactive power delivered by the dg source for the three current control techniques are shown in the figs. 13 and 14, respectively. it can be inferred from the figures that the inverter current and power waveforms of the vector hcc technique are smoother compared to the other two techniques. v. rajini 41 figure 13(a). current delivered by the distributed generation (dg) source—conventional hysteresis current control (hcc). figure 13(b). current delivered by the distributed generation (dg) source— adaptive hysteresis current control (hcc). figure 13(c). current delivered by the distributed generation (dg) source—vector hysteresis current control (hcc). figure 15(a) seems to resemble fig. 15(b). figure 16 shows the error bandwidth in the stationary frame. figure 15 shows that the error band region is more confined and the magnitude of the error vector is restricted in the case of a vector-based technique. figure 15(b) shows a tuned bandwidth resembling fig. 7. figure 14(a). real and reactive power delivered by the distributed generation (dg) source— conventional hysteresis current control (hcc). figure 14(b). real and reactive power delivered by the distributed generation (dg) source—adaptive hysteresis current control (hcc). figure 14(c). real and reactive power delivered by the distributed generation (dg) source—vector hysteresis current control (hcc). investigations on hysteresis-based current control techniques for grid connected photovoltaic systems 40 5. conclusions an effective real and reactive power control strategy for the distribution side power injection of a pv dg source has been discussed. this pq control strategy (a) (b) (c) figure 15(a). current error in dq plane— conventional hysteresis current control (hcc), (b) adaptive hcc and (c) vector hcc. figure 16. adaptive band tuning. table 3: comparison of simulation results parameters switching loss (watts) thd of inverter current (%) conventional 170 5.4 adaptive hcc 130 3.2 vector hcc 60 2.1 was executed through a conventional hcc technique. having observed the disadvantages in implementing a conventional hcc technique, an adaptively tuned hcc and a vector-based hcc were employed in the system, their simulation results were then compared and the system was analysed for the three current control techniques. based on the observations carried out on parameters such as total harmonic distortion and the switching loss of the system, the vector-based control technique was found to be more appropriate and effective. references babu d, kn, ramaprabha r, rajini v (2012), mathematical modeling and simulation of grid connected solar photovoltaic system. international journal of electrical and electronics engineering 20(5):73–77. blaabjerg f, chen z, kjaer s (2004), power electronics as efficient interface in dispersed power generation systems. ieee transactions on power electronics 19(5):1184–1194. brod dm, novotny dw (1985), current control techniques of vsi-pwm inverter. ieee transactions on industrial applications 562– 570. chitti babu b, mohapatra m, jena m, naik a (2008), dynamic performance of adaptive hysteresis current controller for mainsconnected inverter system. ieee transactions on power electronics 22(3):197– 306. esram t, chapman pl (2007), comparison of photovoltaic array maximum power point tracking techniques. ieee transactions on energy conversion 22(2):439–449. kwon bh, kim tw, youn jh (1998), a novel svm-based hysteresis current controller. ieee transactions on power electronics 13(2):297– 307. mohseni m, islam sm (2010), a new vector-based hysteresis current control scheme for threephase pwm voltage-source inverters. ieee transactions on power electronics 25(9). rahim na, mekhilef s (2002), implementation of three-phase grid connected inverter for photovoltaic solar power generation system. proceedings ieee powercon, 1:570–573. 42 v. rajini 43 salmi t, bouzguenda m, gastli a, masmoudi a (2012), matlab/simulink-based modelling of solar photovoltaic cell. international journal of renewable energy research 2(2). vahedi h, sheikholeslami a, bina mt (2011), a novel hysteresis bandwidth (nhb) calculation to fix the switching frequency employed in active power filter. ieee applied power electronics colloquium (iapec), 156. appendix transmission line parameters: resistance: 0.2 ohms/km inductance: 0.5mh/km transformer rating:11kv/415v load parameters: from 0-0.1 s: real power demand: 8000 w reactive power demand: 12000 var from 0.1-0.2 s: real power demand: 16000 w reactive power demand: 24000 var dc link voltage (vdc):700 v adaptive tuning of hysteresis band: line inductance (l): 5 mh switching frequency (fs): 12 khz c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 7.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 60-67 drag reduction properties of nanofluids in microchannels h.a. abdulbari* and f.l.w. ming centre of excellence for advanced research in fluid flow; faculty of chemical engineering and natural resources, universiti malaysia pahang, gambang 26300, kuantan, pahang, malaysia. received 21 november 2014; accepted 1 july 2015 abstract: an experimental investigation of the drag reduction (dr) individualities in different sized micro channels was carried out with nanopowder additives (nas) (bismuth(iii) oxide, iron(ii/iii) oxide, silica, and titanium(iv) oxide) water suspensions/fluids. the primary objective was to evaluate the effects of various concentrations of nas with different microchannel sizes (50, 100, and 200 µm) on the pressure drop of a system in a single phase. a critical concentration was observed with all the nas, above which increasing the concentration was not effective. based on the experimental results, the optimum dr percentages were calculated. the optimum percentages were found to be as follows: bismuth iii oxides: ~65% dr, 200 ppm and a microchannel size of 100 µm; iron ii/iii oxides: ~57% dr, 300 ppm, and a microchannel size of 50 µm; titanium iv oxides: ~57% dr, 200 ppm, and a microchannel size of 50 µm, and silica: 55% dr, 200 ppm, and a microchannel size of 50 µm. keywords: microchannels, pressure drop, drag reduction, nanopowder additives. خواص // 50100 200 65200100 /5730050 572005055 20050 *corresponding authors’ e-mail: hayder.bari@gmail.com h.a. abdulbari and f.l.w. ming 61 1. introduction the study of drag reduction (dr) in microchannels has raised considerable interest in researchers all over the world. dr additives (dras), such as polymers, surfactants, and microbubbles, have proven their effectiveness and are appreciated by many industries for their economic value (nghe et al. 2010). the use of diverse polymers such as dras has been reported previously (abubakar et al. 2014; alsarkhi 2012; matras and kopiczak 2015; edomwonyi-outu, chinaud and angeli 2015; hong et al. 2015; iaccarino et al. 2010; resende et al. 2011) and surfactants (drzazga et al. 2013; li et al. 2008; różański 2011; yu and kawaguchi 2006; tuan mizunuma 2013; qi et al. 2011) depending upon the polarities and different behaviors in a turbulent flow (tarn and pamme 2014), while the impact of other colloidal suspensions, such as nanofluids, in reducing the pressure drop has not been widely studied. the dispersion quality of the nanoparticles in the base fluid and the stability of the suspension play a crucial role in most applications of practical interest (rivet et al. 2011; xie et al. 2003; choi et al. 2007; ganguly et al. 2009). zhao et al. (2009) studied the viscosity of silicon dioxide nanofluids with different particle sizes and ph values. the results revealed that the nanoparticle diameter is of crucial importance to the viscosity of a nanofluid. the smaller the nanoparticles, the larger the viscosity and the greater the dependence on the nanofluid volume fraction. zhao et al. (2009) studied the effect of silica-based nanofluids as dras. the results suggested that the dr enhanced increasing nanofluid concentrations up to a critical concentration, above which no significant dr was observed. pouranfard. (2014) observed an increase in the viscosity of a fluid with increasing concentrations of nanofluid dras and decreases in temperature. kostic (2013) reported a comparative study of silica and carbon nanotubes (cnts), where the cnts were found to be better dras (75%) than silica with a maximum reduction of 60% (yang et al. 2005). very few studies have been done on dr using nanoparticles compared to polymers and surfactants. therefore, in this study, nanofluids at different operating conditions were injected through microchannels, and the effects of different parameters on dr were investigated. pressure drop was calculated as a function of flow rate (rpm) in different sized microchannels (50, 100, and 200 µm) by varying the concentration of different types of oxide nanoparticles, including bismuth, iron, silica, and titanium. the pressure drop was later applied in calculating the dr efficacy (%dr) of a particular nanofluid. in addition, the effect of flow rate and channel size on dr was also investigated. 2. experimental procedure 2.1 materials bismuth(iii) oxide, iron(ii, iii) oxide, titanium (iv) oxide, and silicon(ii) oxide were purchased from sigma-aldrich (st. louis. missouri, usa) and used without any alterations [table 1]. deionized water was used to prepare samples and make dilutions. table 1. physical properties of nanoparticles used. working fluids m. weight (g/mol) size density (g/ml) bismuth(iii) oxide (bi2o3) 465.96 90– 210 0.5–1.1 iron(ii,iii) oxide (fe3o4) 251.53 50– 100 4.8–5.1 silica (sio2) 60.08 200– 300 0.037 titanium(iv) oxide (tio2) 79.87 <100 - 2.2 nanofluid solution preparation nanoparticle solutions were prepared in deionized water for each oxide at five different concentrations, ranging from 100–500 ppm. the fluids were homogenized at high speed for 10 minutes with a stirrer and then left overnight on low speed to achieve maximum dispersion. before running each experiment, the solution was stirred for two hours at 100 rpm to avoid any agglomeration. 2.3 experimental setup straight microchannels of 50, 100, and 200 µm with a length of 58.5 mm (topas advanced polymers, frankfurt-höchst, germany) were employed in this study. the solution was transferred through two syringes via a syringe pump (model sn-50f6) connected to a t-junction connection whose outlet was connected to another t-junction connection. a pressure transmitter (model: stk336) and the inlet of the microchannel were connected to the outlet of the second tdrag reduction properties of nanofluids in microchannels 62 junction connection. the solutions were pumped into the microchannel via connecting tubes. the flow rate of the solutions was controlled using the syringe pump. the flow of the liquid in the microchannel was observed using an hbo 50 microscope (zeiss international, oberkochen, germany) connected to a camera. the pressure drop across the microchannel was observed using the pressure transmitter and the results were recorded. the %dr was calculated using eqn. 1. %dr ∆p ∆p∆p 100 1 where, ∆p = pressure drop before adding dra ∆p = pressure drop after adding dra 3. results and discussion a series of experiments were performed to study the effects of nanofluid concentration, microchannel size, and flow rate on pressure drop and %dr. it was found that the %dr rises with increasing nanofluid concentration. however, there is a critical concentration above which no more reduction can be attained. in the present study, the critical concentration obtained was almost the same for all oxide nanoparticles (200 ppm) with the exception of fe3o4 (300 ppm) [table 2]. figure 1 presents a comparison of the pressure drop at the critical concentrations of nanofluids and deionized water versus the flow rate in the microchannels. a regular pressure drop trend was observed with an increasing flow rate in all cases. increasing the flow rate increases the degree of turbulence which provides a better platform for dras to perform (stone and ajdari 2014). sio2 showed a maximum pressure drop at its critical concentration, which may be due to the fact that the mechanism of dr by sio2 nanoparticles is mainly a surface modification. the sio2 nanoparticles have good rigidity and reduce the friction factor through dispersion into the liquid phase and contact with the pipe surface, so a higher pressure drop was observed (zhao et al. 2009). figures 2–5 show the %dr versus the rate of flow in the different sized microchannels and at different oxide nanoparticle concentrations. the dr is greater with smaller particles as the smaller sized particles with larger surface areas result in the migration effect, where a bond is formed between the transporting fluid molecules and the dr agents (khadom and abdul-hadi 2014). it was observed that the use of nanofluids at low concentrations did not cause a serious reduction in the pressure drop which is due to the small size of the nanoparticles (abdulbari, shabirin and abdurrahman 2013). the diameter of the microchannels does not seem to affect the %dr greatly in contrast to the behavior usually observed with a reduction in diameter (lee and mudawar 2007). the effect of solution velocity (m) on the %dr was studied in terms of the volumetric flow rate. the results show that the %dr increases with increasing fluid velocity. increasing the fluid velocity means increasing the degree of turbulence inside the pipe, which will provide a better medium for the drag reducer to be more effective (byrne, hart and da silva 2012). figure 2 presents the %dr while keeping the concentration constant for all the nanofluids at 100 ppm with varying microchannel sizes (50, 100, and 200 µm). an increase in %dr was observed when increasing the microchannels’ sizes from 50 to 200 µm. this can be explained by the presence of large figure 1. a comparative study of pressure drop between deionized water and oxides used in the study. h.a. abdulbari and f.l.w. ming 63 table 2. parameters affecting the drag reduction in different nanofluids. oxide channel size critical con. flow rate pressure drop drag reduction µm ppm ml/hr kpa % 50 200 1350 9.93 56.9 titanium 100 200 1250 10.74 53.3 200 200 1350 9.34 51.3 50 200 1400 8.28 63.0 bismuth 100 200 1400 7.35 64.8 200 200 1200 8.67 57.5 50 200 1350 10.35 55.1 silicon 100 200 1400 9.99 52.1 200 200 1400 9.68 57.6 50 300 1450 9.59 56.8 iron 100 300 1350 9.66 56.5 200 300 1350 9.94 48.2 µm = micrometer; con. = concentration; ppm = parts per million; kpa = kilopascals figure 2. variations of %dr with a nanofluid concentration of 100 ppm. 42 44 46 48 50 52 54 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-100ppm-50µm bismuth (iii) oxide-100ppm-50µm silicon dioxide-100ppm-50µm iron (ii,iii) oxide-100ppm-50µm 39 41 43 45 47 49 51 53 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-100ppm-100µm bismuth (iii) oxide-100ppm-100µm silicon dioxide-100ppm-100µm iron (ii,iii) oxide-100ppm-100µm 34 36 38 40 42 44 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-100ppm-200µm bismuth (iii) oxide-100ppm-200µm silicon dioxide-100ppm-200µm iron (ii,iii) oxide-100ppm-200µm (a) (b) (c) drag reduction properties of nanofluids in microchannels 64 figure 3. variations of %dr with nanofluid concentration of 200 ppm. figure 4. variation in %dr at a nanofluid concentration of 300 ppm. 45 50 55 60 65 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-200ppm-50µm bismuth (iii) oxide-200ppm-50µm silicon dioxide-200ppm-50µm iron (ii,iii) oxide-200ppm-50µm 38 43 48 53 58 63 68 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-200ppm-100µm bismuth (iii) oxide-200ppm-100µm silicon dioxide-200ppm-100µm iron (ii,iii) oxide-200ppm-100µm 34 39 44 49 54 59 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-200ppm-200µm bismuth (iii) oxide-200ppm-200µm silicon dioxide-200ppm-200µm iron (ii,iii) oxide-200ppm-200µm 32 37 42 47 52 57 62 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-300ppm-50µm bismuth (iii) oxide-300ppm-50µm silicon dioxide-300ppm-50µm iron (ii,iii) oxide-300ppm-50µm 32 37 42 47 52 57 62 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-300ppm100µm bismuth (iii) oxide-300ppm100µm 25 30 35 40 45 50 55 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-300ppm-200µm bismuth (iii) oxide-300ppm-200µm silicon dioxide-300ppm-200µm iron (ii,iii) oxide-300ppm-200µm (a) (b) (c) (c) (a) (b) h.a. abdulbari and f.l.w. ming 65 figure 5. variations in %dr at nanofluid concentrations of 400 ppm. eddies in the larger microchannels which absorb a large amount of energy from the flow. in smaller microchannels, more smaller eddies are formed, but some small eddies are unable to absorb enough energy and are incapable of overcoming the viscosity resistance (kim et al. 2000; khadom and abdul-hadi 2014; al-sarkhi and hanratty 2001). sio2 exhibited the highest %dr of all the microchannels, which may be due to the larger sized silica particles settling down quickly and further decreasing the diameter of the microchannels. figure 3 depicts the maximum dr achieved in all the oxides except for fe3o4 nanofluids at a concentration of 200 ppm, which is a clear indication of reaching a critical concentration. bi2o3 exhibited the highest %dr (~65%) at the critical concentration with a microchannel size of 100 µm at 1400 rpm, which reduced to 200 µm-sized microchannels (~58%) with a flow rate of 1200 rpm. in the case of sio2 and tio2, the maximum drag achieved at the critical concentration was ~57.5 and ~57% with 200 and 50 µm-sized microchannels, respectively. the iron oxide nanofluid exhibited maximum dr at a concentration of 300 ppm in all the microchannels [fig. 4]. maximum dr (~ 56.8 %) was achieved with 50 µm-sized microchannels and a flow rate of 1400 ml/hr. in cases of 100 µm-sized microchannels, a nearly identical increasing trend was observed and maximum drag reduction (~56.5 %) was observed at a slightly lower flow rate (1350 ml/hr). a decreasing trend in drag was observed for all the other oxides, which confirmed that all the oxides had reached their critical concentration limit; increasing the concentration further had no more effect on the pressure drop in a flow. a gradual decrease in the dr properties of the iron oxide nanofluid was observed at 400 ppm [fig. 5], emphasizing that iron oxide nanofluids are more effective in the concentration range of 300–400 ppm. all the other nanofluids exhibited a very low dr; thus, it is obvious that, after a critical limit has been reached, increasing the concentration has no positive effect on dr efficiency and a negative impact on the flow of a system. 4. conclusions to conclude, the dr performances of different nanofluids were studied at diverse concentrations in 22 27 32 37 42 47 52 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-400ppm-50µm bismuth (iii) oxide-400ppm-50µm silicon dioxide-400ppm-50µm iron (ii,iii) oxide-400ppm-50µm 32 37 42 47 52 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-400ppm-100µm bismuth (iii) oxide-400ppm-100µm silicon dioxide-400ppm-100µm iron (ii,iii) oxide-400ppm-100µm 22 27 32 37 42 1100 1200 1300 1400 1500 flowrate (ml/hr) titanium (iv) oxide-400ppm-200µm bismuth (iii) oxide-400ppm-200µm silicon dioxide-400ppm-200µm iron (ii,iii) oxide-400ppm-200µm drag reduction properties of nanofluids in microchannels 66 different sized microchannels. it was observed that all the nanofluids exhibited a certain critical concentration above which increasing the concentration had no greater impact on the pressure drop or the dr of the system. increasing the flow rate also had an impact on dr efficiency, but it was noticed that, above the critical concentration limit, the flow rate did not have much effect on the %dr. the channel size was found to have no effect on the dr efficiency of the nanofluids. acknowledgments authors are grateful to universiti malaysia pahang, for financial assistance. references abubakar a, al-wahaibi t, al-wahaibi y, alhashmi ar, al-ajmi a (2014), roles of drag reducing polymers in singleand multi-phase flows. chemical engineering research and design 92(11): 2153–2181. al-sarkhi a (2012), effect of mixing on frictional loss reduction by drag reducing polymer in annular horizontal two-phase flows. international journal of multiphase flow 39: 186–192. al-sarkhi a, hanratty tj (2001), effect of pipe diameter on the performance of drag-reducing polymers in annular gas-liquid flows. chemical engineering research and design 79(4): 402–408. byrne m, hart r, da silva a (2012), experimental thermal hydraulic evaluation of cuo nanofluids in microchannels at various concentrations with and without suspension enhancers. international journal heat mass transfer 55: 2684–2691. choi cs, park sj, choi hj (2007), carbon nanotube/polyaniline nanocomposites and their electrorhelogical characteristics under an applied electric field. current applications in physics 7: 352– 355. drzazga m, gierczycki a, dzido g, lemanowicz m (2013), influence of nonionic surfactant addition on drag reduction of water based nanofluid in a small diameter pipe. chinese journal of chemical engineering 21(1): 104–108. edomwonyi-otu lc, chinaud m, angeli p (2015), effect of drag reducing polymer on horizontal liquid–liquid flows. experimental thermal and fluid science 64: 164–174. ganguly s, sikdar s, basu s (2009), experimental investigation of the effective electrical conductivity of aluminum oxide nanofluids. powder technology 196: 326–330. hong c, choi hj, zhang k, renou f, grisel m (2015), effect of salt on turbulent drag reduction of xanthan gum. carbohydrate polymers 121: 342–347. iaccarino g, shaqfeh esg, dubief y (2010), reynolds-averaged modeling of polymer drag reduction in turbulent flows. journal of nonnewtonian fluid mechanics 165(7–8): 376–384. khadom aa, abdul-hadi aa (2014), performance of polyacrylamide as drag reduction polymer of crude petroleum flow. ain shams engineering journal 5(3): 861–865. kim nj, lee jy, yoon sm, kim cb, hur bk (2000), drag reduction rates and degradation effects on synthetic polymer solution with surfactant additives. journal of industrial and engineering chemistry 6: 412–418. kostic mm (2013), friction and heat transfer characteristics of silica and cnt nanofluids in a tube flow. proceedings of the 8th annual international conference on energy and environment, rhodes island, greece. lee j, mudawar i (2007), assessment of the effectiveness of nanofluids for single phase and two-phase heat transfer in micro-channels. international journal of heat mass transfer 50(3-4): 452–463. li fc, kawaguchi y, yu b, wei jj, hishida k (2008), experimental study of drag-reduction mechanism for a dilute surfactant solution flow. international journal of heat and mass transfer 51(3–4): 835–843. matras z, kopiczak b (2015), intensification of drag reduction effect by simultaneous addition of surfactant and high molecular polymer into the solvent. chemical engineering research and design 96: 35–42. nghe p, tabeling p, ajdari a (2010), flow-induced polymer degradation probed by a high throughput microfluidic set-up. journal of nonnewtonian fluid mechanics 165(7–8): 313–322. pouranfard ar, mowla d, esmaeilzadeh f (2014), an experimental study of drag reduction by nanofluids through horizontal pipe turbulent flow of a newtonian liquid. journal of industrial and engineering chemistry 20(2): 633–637. qi y, kesselman e, hart dj, talmon y, mateo a, zakina jl (2011), comparison of oleyl and elaidyl isomer surfactant–counterion systems in drag reduction, rheological properties and nanostructure. journal of colloid and interface science 354(2): 691–699. resende pr, kim k, younis ba, sureshkumar r, pinhoa rt (2011), a fene-p k–ε turbulence model for low and intermediate regimes of polymerinduced drag reduction. journal of non-newtonian fluid mechanics 166(12–13): 639–660. h.a. abdulbari and f.l.w. ming 67 rivet c, hirsch a, hamilton s, lu h (2011), microfluidics for medical diagnostics and biosensors. chemical engineering science 66: 1490– 1507. różański j (2011), flow of drag-reducing surfactant solutions in rough pipes. journal of non-newtonian fluid mechanics 166(5–6): 279–288. stone ha, ajdari a (2004), engineering flows in small devices: microfluidics toward a lab-on-achip. annual review of fluid mechanics 36: 381–411. tarn, md, pamme n (2014), microfluidics, in reference module in chemistry, molecular sciences and chemical engineering elsevier. tuan na, mizunuma h (2013), high-shear drag reduction of surfactant solutions. journal of nonnewtonian fluid mechanics 198: 71–77. xie h, lee h, youn w, choi m (2003), nanofluids containing multi walled carbon nanotubes and their enhanced thermal conductivities. journal of applied physics 94: 4967–4972. yang y, zhang z, grulke ea, anderson wb, wu g (2005), heat transfer properties of nanoparticle-in-fluid dispersions (nanofluids) in laminar flow. international journal of heat mass transfer 48(6): 1107–1116. yu b, kawaguchi y (2006), parametric study of surfactant-induced drag-reduction by dns. international journal of heat and fluid flow 27(5): 887–894. zhao jf, luo zy, ni mj, cen kf (2009), dependence of nanofluid viscosity on particle size and ph value. chinese physics letters 26: 066202. the journal of engineering research (tjer), vol. 19, no. 1, (2022) 41-53 *corresponding author’s e-mail: s121293@student.squ.edu.om doi:10.53540/tjer.vol19iss1pp41-53 recent cnn-based techniques for breast cancer histology image classification arunadevi karuppasamy1,*, abdelhamid abdesselam1, rachid hedjam1, hamza zidoum 1, and maiya al-bahri2 1 department of computer science, sultan qaboos university, 2 department of pathology, sultan qaboos university hospital, muscat, sultanate of oman abstract: histology images are extensively used by pathologists to assess abnormalities and detect malignancy in breast tissues. on the other hand, convolutional neural networks (cnn) are by far, the privileged models for image classification and interpretation. based on these two facts, we surveyed the recent cnn-based methods for breast cancer histology image analysis. the survey focuses on two major issues usually faced by cnn-based methods namely the design of an appropriate cnn architecture and the lack of a sufficient labelled dataset for training the model. regarding the design of the cnn architecture, methods examining breast histology images adopt three main approaches: designing manually from scratch the cnn architecture, using pre-trained models and adopting an automatic architecture design. methods addressing the lack of labelled datasets are grouped into four categories: methods using pre-trained models, methods using data augmentation, methods adopting weakly supervised learning and those adopting feedforward filter learning. research works from each category and reported performance are presented in this paper. we conclude the paper by indicating some future research directions related to the analysis of histology images. keywords: breast cancer; cnn; deep learning; histology image classification; machine learning. لتصنیف صور سرطان الثدي cnn المعتمدة على شبكةالتقنیات الحدیثة و عبدالحمید عبدالسالم و راشد حجام و حمزة زیدوم و میا البحري *آرون دیفي كربوسامي ة في أنسجة الثدي. من یستخدم علماء األمراض صور األنسجة على نطاق واسع لتقییم التشوھات واكتشاف األورام الخبیث :الملخص قمنا ج الممیزة لتصنیف الصور وتفسیرھا؛ بناًء على ھاتین الحقیقتینالنماذ تعد الشبكات العصبیة التالفیفیة إلى حد بعیدناحیة أخرى لتحلیل صورة أنسجة سرطان الثدي. یركز المسح على مشكلتین رئیسیتین عادة ما تواجھھما cnnبمسح الطرق الحدیثة المعتمدة على المناسبة وعدم وجود مجموعة بیانات مصنفة كافیة لتدریب النموذج. فیما یتعلق cnnوھما تصمیم بنیة cnnاألسالیب المعتمدة على ، استخدام نماذج مدربة cnnلھندسة ثة مناھج رئیسیة: التصمیم الیدوي، تتبنى طرق فحص صور أنسجة الثدي ثالcnnبتصمیم بنیة ق التي تعالج نقص مجموعات البیانات المصنفة في أربع فئات: األسالیب التي مسبقًا واعتماد تصمیم ھندسي تلقائي. یتم تجمیع الطر تستخدم النماذج المدربة مسبقًا، والطرق التي تستخدم زیادة البیانات، والطرق التي تعتمد التعلم الخاضع لإلشراف الضعیف، وتلك التي ة من كل فئة واألداء المبلغ عنھ. نختم الورقة باإلشارة إلى بعض تتبنى التعلم التوضیحي. في ھذه الورقة، نقوم بعرض األعمال البحثی اتجاھات البحث المستقبلیة المتعلقة بتحلیل صور األنسجة. .األليالتعلم ؛تعلم اآللة ؛تصنیف صور األنسجة ؛الشبكات العصبیة التالفیفیة ؛سرطان الثدي الكلمات المفتاحیة: recent cnn-based techniques for breast cancer histology image classification 42 1. introduction breast cancer is the most common cancer among women in developed countries (iarc, 2020) (intercollegiate and network, 2005) (sickles, 1997). currently, pathologists examine histology images to interpret the tissue appearance and detect abnormal conditions. these images usually have a large resolution, different magnification scales, various acquisition processes, and different staining methods, which set them apart from other medical images (see figure 1). besides, the result of a histology image analysis may vary from different pathologists, and sometimes from different analyses done by the same pathologist (gurcan et al., 2009, germain, 2017), which leads to inter-observer and intra-observer variability and uncertainties in the decision-making process. hence, there is a need for developing effective and efficient computer-aided techniques to analyze and interpret histology images. machine learning (ml) is a discipline of artificial intelligence that has proven to be very efficient in solving classification problems. it allows computers to learn from data without being explicitly programmed. supervised learning is an ml category in which a model is trained on a set of inputs (features) with known outcomes (labels). once the training is completed, the defined model will be capable of making predictions when fed with new unseen data. traditionally, features are explicitly selected and extracted by the user (see figure 2). extracted features are mostly related to intensity, morphology, and texture (boucheron, 2008) (dundar et al., 2011) (liu et al., 2011). examples of these features include: (i) gabor-wavelet filters, and density measures features (marugame et al., 2009), (ii) gaussian markov random field and fractal dimension features (al-kadi, 2010), (iii) intensity, morphological, co-occurrence, and run-length features (irshad, 2013), (iv) haar-like features and gaussian filters features (vink et al., 2013), and (v) local binary patterns, morphometric features, entropy features and graylevel co-occurrence matrix (tashk et al., 2015) (bruno et al., 2016), (peikari et al., 2017). with the advent of deep learning (dl), image classification and object recognition achieved unprecedented levels of accuracy (lecun et al., 2015). similarly, its application to healthcare especially for analyzing medical images showed excellent performance in solving various problems such as segmentation, interpretation and registration (kim m et al. 2019). the success of dl is mainly due to the ability of modern computers to process huge amounts of data, and extract features automatically at different abstraction levels. convolutional neural network (cnn), is a dl technique specially adapted to process images and videos. the first cnn, lenet (y. lecun et al., 1998) was proposed in 1998 to classify handwritten digits, but the real surge in cnn popularity started with alexnet (krizhevsky et al., 2012) when it won the imagenet challenge 2012 and since then several deeper and more performant cnns were developed (vgg16 (k. simonyan and a. zisserman, 2014), googlenet (c. szegedy et al., 2015 and 2016), and resnet(he et al., 2016). figure 1. benign and malignant samples from breakhis dataset of various magnification (f. spanhol, 2016). figure 2. the traditional machine learning process for breast cancer histology image classification. arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 43 figure 3. the number of papers addressing the problem of breast cancer histology image classification published between 2013 and 2021. cnn-based methods, including those addressing histopathology image classification, usually face two major challenges: 1)there is no systematic approach for designing an appropriate cnn architecture and 2) there is a need for large-scale training dataset that is, unfortunately not available for many real-world problems, including those related to the medical field. this review aims at exploring how these two major challenges have been addressed in the literature, especially by those research works using breast cancer histopathological images. to the best of our knowledge, this is the first article conducting such a study. the selection of the reviewed papers is based on the search keywords “deep learning, histopathology and breast cancer” in springer, sciencedirect, and ieee explore during the period 2013-2021. the total number of publications in the breast cancer histopathology classification gradually increased during that period (see figure3). the search returned with 987 research papers, then we removed 136 duplications, and reported only the papers related to the classification of breast cancer using cnn. the remainder of the paper is organized as follows: section 2 describes the main components of a cnn model and introduces the datasets used by the methods included in this survey. section 3 describes the proposed models for breast cancer histology image classification and categorizes them based on the way their architecture is designed. section 4 describes methods proposed to address the problem of the lack of labelled datasets. the performance, in terms of accuracy and training time, of all surveyed works, were reported. the paper concludes by summarizing the main characteristics of the proposed approaches and presenting some future research directions related to the analysis of histology images. 2. convolutional neural network for histology image classification cnns are capable of learning automatically features from raw input data. they consist of several feature extraction layers where the first layers extract basic features such as edges and blobs and deeper layers extract more complex and abstract features. a cnn includes three main building blocks: (i) convolutional and pooling layers, (ii) fully-connected layers and (iii) a classification layer (see figure 4). the convolution operation is the most expensive computation of a cnn, it extracts information by convolving the input data with a set of filters. a feature map is obtained by applying an activation function to the output of the convolution operation. this ensures non-linearity and enforces a sparse representation of the feature map. early models used sigmoid and hyperbolic tangent functions as activation functions (lecun et al., 2012). recent works adopted other activation functions such as relu (krizhevsky et al., 2012), leaky relu(lrelu) (maas et al., 2013), parametricrelu(prelu) (he et al., 2015), randomised-relu (rrelu) (xu et al., 2015), s-shaped relu(s-relu) (jin et al., 2016), maxout (goodfellow et al., 2013), and exponential linear unit (elu) (trottier et al., 2017). a pooling operation usually follows convolutional and activation operations to reduce the size of the feature maps. there are different pooling types in cnn (maas et al., 2013) such as max-pooling, mean-pooling, stochastic pooling, spatial pyramid pooling, and deformation pooling. max pooling (boureau, bach, et al., 2010)(boureau, ponce, et al., 2010) is the most used method; it consists of replacing a pooling window by the maximum value in the window. then come a set of fully-connected layers, where each neuron in a layer is connected to all neurons of the following layer. the last feature map is passed to the first fully connected layer in a form of a 1-d vector (kröse et al., 1993) to produce an output more appropriate to the classification task that is performed either by applying a function like a softmax for multi-label classification or a sigmoid for binary classification. one of the traditional classification models such as logistic regression (lr), support vector machines (svm) or random forests (rf) can also perform the classification. figure 4. convolutional neural network for breast cancer histology image classification. pu bl ic at io n year recent cnn-based techniques for breast cancer histology image classification 44 table 1. dataset and links for breast cancer histology classification. dataset links icpr-2012 http://ludo17.free.fr/mitos_2012/ tupac2016 http://tupac.tue-image.nl/ breakhis 2016 https://web.inf.ufpr.br/vri/databases/ breastcancer-histopathological camelyon16 https://camelyon16.grandchallenge.org/ patchcamelyon https://patchcamelyon.grandchallenge.org/download/ bach-iciar18 https://iciar2018-challenge.grandchallenge kaggle data repository https://www.kaggle.com/c/histopatho logic-cancer-detection most of the models discussed in this paper are trained and tested on breast cancer datasets publicly available on the internet. their links are shown in table 1. three main performance metrics are used by the methods surveyed in this paper. classification accuracy, f-score metric and area under curve (auc), where 1. accuracy= # 𝑜𝑜𝑜𝑜 𝑐𝑐𝑜𝑜𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑜𝑜𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑜𝑜𝑐𝑐𝑐𝑐 𝑇𝑇𝑜𝑜𝑐𝑐𝑐𝑐𝑐𝑐 𝑐𝑐𝑛𝑛𝑛𝑛𝑛𝑛𝑐𝑐𝑐𝑐 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑡𝑡 𝑐𝑐𝑛𝑛𝑐𝑐𝑖𝑖𝑐𝑐𝑐𝑐 (1) 2. f-score= = 2 x precision x recall precision+recall , (2) where precision is the fraction of correctly classified images among those labelled as positive, and recall is the fraction correctly classified images out of all positive images in the test dataset. 3. area under the curve (auc) indicates how well the positive classes are classified compared to negative classes. (3) f-score and auc are more appropriate for evaluating the classification performance on datasets with imbalanced classes. 3. cnn architecture design based on the way architectures are designed, we can categorize cnn-based methods for classifying cancer histology images into three categories: (i) manuallydesigned cnn architectures, (ii) pre-trained cnn architectures, and (iii) automatically-designed cnn architectures. • manually-designed cnn architectures most cnn-based histology image classification methods are of this category as shown in table 2. they use relatively shallow architectures (small number of layers) and backpropagation algorithms for filter learning. they are trained mostly on datasets of limited size using a limited number of epochs and achieve moderate to good levels of performance. (cirecsan et al., 2013) proposed a 6-layer cnn (four convolutional and pooling layers and two fullyconnected layers) to detect mitosis in a dataset of 50 images. it was trained for one day with less than 30 epochs in an optimized gpu implementation. this method won the icpr2012 competition by achieving an f1-score of 0.78. (cruz-roa et al., 2014) designed a 3-layer cnn to automatically detect invasive ductal carcinoma from whole-slide images (wsi) of breast cancer. the dataset consists of 162 wsis from which about 165000 patches of size 100x100 were extracted and used for model training and testing. the proposed cnn was trained with 25 epochs and achieved an f1score of 0.71. (wang et al., 2014) proposed a cascade ensemble that combines hand-crafted features (morphology, colour, and texture features) with features extracted from a 3-layer cnn. these two feature sets were passed to a random forest classifier to detect mitotic nuclei on the icpr dataset consisting of 50 high power field images (scanned into 2084x2084 pixels rgb images). the model was trained for 9 epochs and it required nearly 18 hours without gpu implementation. they reported an f1-score of 0.73. (litjens et al., 2016) proposed a patch-based cnn model to classify prostate cancer and detect micrometastases and macro-metastases breast cancer in sentinel lymph nodes. their dataset consists of 271 wsis from which 2.6 million patches of size 128x128 were obtained through data augmentation. for metastases detection, the cnn was trained with 12 epochs, each epoch requiring 200 minutes on geforce gtx970, and obtained an auc of 0.88. (araújo et al., 2017) proposed an 8-layer cnn model that classifies histology images into four categories: i) normal tissue, ii) benign lesion, iii) in situ carcinoma, and iv) invasive carcinoma. their dataset consists of 269 high resolution images (2049x1536 pixels) obtained from the bio-imaging 2015 breast histology classification challenge dataset from which 70000 (512x512) patches were extracted. the cnn model was trained with 50 epochs and achieved a validation accuracy of 80.6% for binary classification (non-carcinoma and carcinoma). the accuracy improved to 83.3% when using the cnn as a feature selection tool and an svm as a classifier. (bayramoglu et al., 2017), proposed two 6-layer cnn architectures for breast cancer histology image classification characterized by their independence of the magnification level. a single task cnn (3 convolutional layers followed by 3 fully connected layers) was used to predict malignancy and a multi-task cnn (3 convolutional layers followed by two sequential fully connected layers) is used to predict both malignancy and image magnification level simultaneously. the cnns were trained with images from different magnification levels (40x, 100x, 200x, and 400x) of breakhis dataset. http://ludo17.free.fr/mitos_2012/ http://tupac.tue-image.nl/ https://patchcamelyon.grand-challenge.org/download/ https://patchcamelyon.grand-challenge.org/download/ https://iciar2018-challenge.grand-challenge/ https://iciar2018-challenge.grand-challenge/ arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 45 table 2. overview of manually-designed cnn architectures used in breast cancer histology image classification. the authors reported an average recognition accuracy of 83.25% for single cnn and averages of 82.13% and 80.10% for benign or malignant classification and magnification detection respectively. (wahab et al., 2017) proposed a two-phase cnn to handle imbalanced histology data. the first phase of the cnn was trained with 80x80 pixels non-mitosis patches to classify them into three imbalanced classes: easy, normal and hard using 5 epochs. the resulting classes are then balanced and combined with augmented mitosis patches to be used for retraining the cnn with 25 additional epochs during 15 hours on i7-3770 cpu of 3.4ghz to classify the patches into mitotic and nonmitotic nuclei. the patches were extracted from two datasets (50 hpfs(high-power fields) from the icpr2012 context dataset and 73 hpfs from the tupac16 dataset). the authors reported an f1-score of 0.79. (cruz-roa et al., 2018) proposed a framework called hashi for invasive breast cancer detection in wsi. the training data is obtained by applying a regular sampling on 600 labelled wsis. the patches resulting from a pseudorandom sampling of the new unseen wsis are passed to a 2-layer cnn classifier; resulting predictions are used to build an interpolated probability map. dense sampling is further applied to regions with high uncertainty, which produces an improved probability map estimation. the authors reported an auc of 0.90. (roy et al., 2019) proposed an 8-layer cnn model for binary (non-malignant and malignant) and multi-class (normal, benign, in-situ and invasive) classification on the iciar-2018 dataset, which consists of 500 images (400 for training and 100 for testing) of size 2048x1536. the model was trained with about 70 epochs on 128gb ram of cpu based system with xenon processor to yield an accuracy of 92.5% and 90% for binary classification and multiclassification, respectively. • pre-trained cnn architectures this category of methods implements transfer learning, where features learned to solve a problem from one domain constitute the baseline for the features used to solve a problem from a different but related domain (mehra, 2018). these techniques take an existing cnn architecture usually trained on a large dataset, and retrain it on a smaller dataset related to the problem under investigation. during the training, only the weights of the last few layers are updated using the backpropagation approach for filter learning. these cnns are usually deep (more than seven layers), initially trained for a long time and with a large number of epochs which make them very performant in solving the original problem. retraining takes usually, much less time (less epochs) on much smaller datasets and resulting models reference cnn architecture dataset performance (cirecsan et al., 2013) 6-layer cnn(4-convolutional layers and 2 fully-connected layers) icpr2012(50 images) f1-score :0.78 (cruz-roa et al., 2014) 3-layer cnn (2-convolutional layers and 1-fully-connected layer) hospital of university of pennsylvania and the cancer institute of new jersey dataset (162 images ) f-score : 0.71 acc: 84.23% (wang et al., 2014) 3-layer cnn (2-convolutional layers and a fully-connected layer) icpr dataset (50 images) f-score: 0.73 (litjens et al., 2016) 6-layer cnn(4-convolutional layers and 2 fully-connected layers ) radboud university medical center dataset(173 images) auc: 0.88 (araújo et al., 2017) 8-layer cnn(5 convolutional layers and 3 fully-connected layers) bioimaging 2015 breast histology classification challenge dataset (269 images) acc: multi-class:77.8% and binary-class: 83.3% (bayramoglu et al., 2017) two 6-layer cnns breakhis dataset (7909 images) average recognition rate:83.25% (wahab et al., 2017) two-phase cnn model icpr2012 dataset(50 images) and tupac16 dataset(73 images) f-score : 0.79 (cruz-roa et al., 2018) 2-layer cnn hospital of the univ. of pennsylvania(239 images), case western reserve univ(110 images), cancer institute of new jersey(52 images), and the cancer genome atlas(195 images) acc: 0.90 dice coefficient: 76% (roy et al., 2019) patch-based classifier with cnn (6-convolutional layers and 2-fully connected layers) iciar-2018 dataset (500 images) acc: binay-class: 92.5% multi-class:90% recent cnn-based techniques for breast cancer histology image classification 46 table 3. overview of pre-trained cnn architectures used in breast cancer histology image classification. reference cnn architecture dataset performance (fabio alexandre spanhol et al., 2016) alexnet breakhis dataset (7909 images) acc:85.6 (40x), 83.0(100x), 83.1(200x), 80.8 (400x) (fabio a spanhol et al., 2017) bvlc caffenet model (pre-trained alexnet as a cnn) along with decaf features breakhis dataset (7909 images) acc: 84.3 (40x), 84.7 (100x), 84.1(200x), 81.6 (400x) (song et al., 2017) component selective encoding cnn (combination of fisher vector and vggnet19) breakhis dataset (7909 images) acc: 87.5 (40x), 88.6(100x), 85.5(200x), 85.0(400x) (bejnordi et al., 2017) cas-cnn( a combination of wideresnet and vggnet) radboud university medical centre dataset(224 images) auc: 0.96(binary) acc: 89%(binary) acc: 81.3%(multiclass) (han et al., 2017) class structure-based deep convolutional neural network (googlenet) breakhis dataset (7909 images) acc: 95.8 (40x), 96.9 (100x), 96.7 (200x), 94.9 (400x) (vang et al., 2018) inceptionv3 and dual-path network iciar-2018 dataset(400 images) acc: 87.5% (motlagh et al., 2018) resnet152 tma database(6,402) and breakhis dataset(7909 images) tma acc:99.8% breakhis acc:98.7%(binary) acc:94.6%(multiclass) (vesal et al., 2018) resnet50 bach-2018 dataset (320 images) acc: 97.50% (kausar et al., 2019) vggnet16 iciar 2018 (400 images) and breakhis dataset(7909 images) acc: 98.2%(iciar2018) 96.85%(40xbreakhis) budak et al. (2019) fcn based on alexnet and bi-lstm (bidirectional long shortterm memory) breakhis dataset(7909 images) acc: 95.69(40x), 93.61(100x), 96.32(200x), 94.29(400x). alom et al. (2019) inception recurrent residual cnn (irrcnn) bioimaging challenge 2015 acc: 99.05% (for binary) and 98.59%(for multi) vo et al. (2019) inception and gradient boosting trees breakhis dataset(7909 images) acc: 93.5(40x), 95.3(100x), 96.1(200x), 91.1(400x) (anwar et al., 2020) resnet50,hog,wpd breakhis dataset (7909 images) acc: 97.10(40x), 97.56(100x), 96.41(200x), 94.32(400x) (chakraborty et al., 2020) resnet50 kaggle data repository (220,025 patches of images) acc:96.48% f1-score:96.32% laxmisagar and hanumantharaju (2021) mobilenet2.10ex bioimaging challenge 2015 acc: 88.92% munien and viriri (2021) efficientnet iciar2018 dataset acc: 98.33% arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 47 are usually of high performance. the most re-used cnn architectures are alexnet (krizhevsky et al., 2012), vggnet (simonyan and zisserman, 2014), googlenet (szegedy et al., 2015), and resnet (he et al., 2016) (see table 3). alexnet is an 8-layer cnn trained on an ilsvrc dataset (deng et al., 2009). the dataset consists of images of 1000 different classes. 1.2 million images are used for training, 50,000 images for validation and 150,000 images for testing. it took 6 days for a 90 epoch training and achieved a winning error rate of 15.3%. the second version of alexnet trained on cifar-10 dataset that contains 60000 images (50000 for training and 10000 for testing) from 10-classes has also been used as a pre-trained model (krizhevsky and hinton, 2010) (krizhevsky, 2009). vggnet is available in two architectures, 16 layers and 19 layers; they were trained on the ilsvrc14 challenge dataset for 74 epochs and won second place by obtaining an error rate of 14.7% and 7.3% respectively. googlenet is a 22layer architecture also trained on the ilsvrc14 challenge dataset and won the competition by obtaining an error rate of 6.7%. resnet is a deep network that is available in different versions, the most common ones have 50, 101, and 152 layers respectively. the resnet50 was trained on the imagenet dataset and achieved an error rate of 5.25%. the resnet101 model was trained using 80k iterations for detection and segmentation of the coco 2015 challenge dataset and won the coco 2015 competitions by achieving a 28% relative improvement on object detection. the resnet152 model was trained on the ilsvrc15 challenge dataset for 60 × 104 iterations and won the ilsvrc15 competitions by achieving an error rate of 3.57%. classification of histology images with pre-trained networks yields better performance compared with manually designed cnns. fabio alexandre spanhol et al. (2016) applied the alexnet pre-trained on the cifar-10 dataset on a multimagnification dataset called breakhis that was introduced by (fabio a spanhol et al., 2015). a subset of 1000 patches of size 64x64 was used to train the model for about 80,000 iterations to obtain image-level accuracy of 85.6 ± 4.8% (40x), 83.0 ± 3.9% (100x), 83.1 ± 1.9% (200x), and 80.8 ± 3 % (400x). (fabio a spanhol et al., 2017) applied a modified version of alexnet (the order of pooling and normalization layer was exchanged) pre-trained on the imagenet dataset. on the breaks dataset their model achieved an imagelevel accuracy of 84.3 ± 2.9% (40x), 84.7 ± 4.4% (100x), 84.1 ± 1.5% (200x), and 81.6 ± 3.7% (400x). (song et al., 2017) proposed a model, which extracts features from the final convolutional layer of vggnet19, encoded by a fisher-vector to feed an svm classifier. the proposed model was evaluated on the breakhis dataset and obtained a binary classification accuracy of 87.5 ± 1.6 (40x), 88.6 ± 3.6 (100x), 85.5 ± 2.0 (200x), and 85.0 ± 4.6 (400x). (bejnordi et al., 2017) proposed a cascaded cnn (cas-cnn) model, which is a combination of a modified version of resnet called wide-resnet (zagoruyko and komodakis, 2016) and vggnet. the proposed cas-cnn model was trained and tested on 224 wsi (100 normal/benign, 69 dcis, and 55 idc wsis) from radboud university medical center, netherland. an accuracy of 89% was obtained for binary classification (benign and cancer) and 81% for multi-classification (benign, ductal carcinoma, and invasive ductal carcinoma). (han et al., 2017) proposed a class structure-based deep cnn (csdcnn), which is based on a pre-trained googlenet for binary and multi-classification. the data augmentation method was applied on breakhis dataset to increase the training images. moreover, an over-sampling based on gaussian distribution was applied to deal with the imbalance of the classes. the training took 10 hours on the augmented dataset on intel i7nividia quadro k22200 gpu and achieved image-level accuracies of (95.8 ± 3.1 (40x), 96.9 ± 1.9 (100x), 96.7 ± 2.0 (200x), and 94.9 ± 2.8 (400x)). (vang et al., 2018) used the pre-trained googlenet model on the iciar-2018 challenge dataset, which contains 400 microscopy images of size 2040x1536 pixels. the training was conducted on 4 gpus (2 nividia titanx gpus and 2 nividia gtx 1080ti) for 30 epochs to achieve an accuracy of 87.5% for a multi-classification (normal, benign, in-situ carcinoma and invasive carcinoma). (motlagh et al., 2018) used the pre-trained resnet152 to extract hierarchical features on tma (tissue micro array) and breakhis datasets. the model was trained on asus geforce gtx1080 for 3000 epochs. it achieved an accuracy of 98.7% and 96.4%, for binary and multi-classification, respectively. (vesal et al., 2018) applied resnet50 on the bach-2018 challenge dataset, which consists of 400 wsi of size 2040x1536 pixels. from the 320 training images, 67,200 patches of size 512x512 were extracted for training. the model was trained for about 100 epochs and achieved an accuracy of 97.5%. (kausar et al., 2019) proposed a modified version of vggnet16 by replacing the last three fully connected layers with a global pooling layer. the input of the model consists of haar wavelet-decomposed images. iciar2018 and breakhis datasets have been used to train the model for about 150 epochs on nvidia tesla m40 gpu to achieve an accuracy of 98.2% and 96.85% on the two datasets, respectively. (budak et al., 2019) used an integrated model combining alexnet, and bilstm (bidirectional long short-term memory) to classify breakhis dataset. the authors reported that pre-trained alexnet was re-trained for 20 epochs in the nvidia quadro p6000 gpu, and achieved the accuracy of 95.69% (40x), 93.61% (100x), 96.31%(200x), and 94.29%(400x). (alom et al., 2019) proposed a model that combines inception-v4, resnet and recurrent cnn. the model was re-trained for 150 epochs on the breakhis dataset using a 56g of ram gpu with nvidia geforce gtx-980 ti. it achieved an accuracy of 97.95% (40x), 97.57% (100x), 97.32 (200x), and 97.36 % (400x). (vo et al., 2019) proposed recent cnn-based techniques for breast cancer histology image classification 48 a model based on inception-v3 and resnet-152. it was trained on breakhis dataset for 50 epochs on geforce gtx 1080 ti that yields an accuracy of 93.5% (40x), 95.3(100x), 96.1%(200x), and 91.1%(400x). (anwar et al., 2020) proposed a model that combines features produced by resnet50, wavelets packet decomposition, and histograms of the gradient. pca is then applied to reduce the feature dimensionality. a data augmentation was applied to the breakhis dataset to obtain 237270 training patches. the authors reported accuracies of 97.10 (40x), 97.56 (100x), 96.41 (200x), and 94.32 (400x). (chakraborty et al., 2020) proposed a dual channel residual cnn (dcrcnn) which decomposes input images into two separate stain channels (eosin and hemotoxylin) (ruifrok, 2019). stain channels are fed into two separate resnet50; resulting features are then fused for classification. the proposed model was applied on kaggle data repository (kaggle dataset, 2018), which contains nearly 220,025 images of size 96x96. the authors trained the model for 50 epochs and achieved an accuracy, and f1-score of 96.48% and 96.32%, respectively. laxmisagar and hanumantharaju (2021) proposed a model that incorporated mobilenet2.10e with fully convolutional deep neural network on bio-imaging challenge 2015 dataset. this model was trained for 280 epochs in google colab with gpu 2496 cuda, and achieved an accuracy of 88.92%. (munien and viriri, 2021) proposed a model based on efficientnet-b2 trained for 50 epochs in google colab with nvidia tesla k80 gpu, and achieved an accuracy of 98.3% in iciar2018 dataset. • automatically-designed cnn architecture network architecture search (nas) is a state-of-theart approach for designing efficient cnn-based solutions; it was recently introduced to automate and optimize the hyper-parameters and the architecture design of neural networks (zoph and le, 2017)(wistuba, 2019). nas methods can be categorized according to three dimensions (elsken et al. 2019): search space, search strategies and performance evaluation strategies. the search space defines which architectures a nas method might discover, the search strategy defines the way the search space is explored to discover an optimized architecture in terms of some parameters such as number of layers, number of filters, size of the filters and pooling window, and strides while a performance evaluation strategy aims at estimating the performance of a neural network architecture in terms of predefined criteria such as the accuracy. the reader is referred to (elsken et al. 2019) for more details on these strategies. (koohbanani et al., 2018) proposed a model that applies the nas approach to design a neural network for breast cancer histology image classification. in this work, a bayesian optimization approach (zoph and le, 2017) was applied to estimate the optimal values of the main hyper-parameters of the model (filter size, dropout layer, learning rate, and momentum). the optimized model was used to detect metastasis on camelyon16 dataset that consists of 400 wsi (270 wsi for training and 130 wsi for testing). the authors reported an auc of 0.99. (oyelade and ezugwu, 2021) proposed a nas based model using ebola optimization search algorithm for search space to optimize the hyper-parameters of deep neural networks that applied on the breast cancer histopathology images of breakhis and bach dataset. the optimized model trained for 500 epochs on both datasets and attained a 100% accuracy 4. lack of labeled histology datasets to achieve high prediction performance, a cnn model requires large labelled training datasets which is, unfortunately, not available for many problem domains including the medical domain. there are two main reasons behind the lack of labelled histology image datasets: 1) labeling histology images is timeconsuming and requires medical experts, which make the task very expensive. 2) medical data, in general, are usually confidential and therefore not easily available to the research community. several approaches have been adopted to address this problem; the most common methods used for histology image classification are patch generation, data augmentation and pre-trained cnns. the two other methods addressing this problem are weakly supervised learning and feed-forward filter learning. • commonly used approaches dealing with small histology datasets histology images are in general of large size (wide slide images 2040 x 1536 pixels) which allows increasing the number of the dataset samples by generating thousands of smaller labelled images (patches) from the original wsi dataset (litjens et al., 2016) (araújo et al., 2017) (wahab et al., 2017) (vesal et al., 2018). data augmentation is another commonly adopted method for increasing the size of image datasets. it consists of generating additional labelled images by applying various image transformations, such as rotation, translation, zooming, and cropping, to the original dataset. (fabio alexandre spanhol et al., 2016) (vang et al., 2018) (anwar et al., 2020). finally, pre-trained models are often adopted because they can lead to high accuracy even when re-trained on relatively small datasets (song et al., 2017) (bejnordi et al., 2017). • cnn using weakly-supervised learning weakly supervised learning includes three categories (zhou, 2018): 1) incomplete supervision, in which only a subset of training data is given with labels; 2)inexact supervision, in which the training data are given with only coarse-grained labels, and 3) inaccurate supervision, in which the given labels are not always ground truth (albarqouni et al., 2016). arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 49 table 3. overview of methods dealing with the lack of large labelled datasets in breast cancer histology classification. reference supervision/ filter learning dataset performance (albarqouni et al., 2016) inaccurate supervision miccai-amida13 challenge dataset (311 images) f1-score: 0.74 auc: 0.869 (das et al., 2018) inexact supervision (mil) breakhis dataset (7909 images) acc:89.52(40x), 89.06 (100x), 88.84(200x), 87.67(400x) (sudharshan et al., 2019) inexact supervision (mil) breakhis dataset (7909 images) acc:86.9(40x), 85.7 (100x), 85.7 (200x), 83.4 (400x) (jaiswal et al., 2019) incomplete supervision patchcamelyon (327680 patches of images) auc: 0.98 (shi et al., 2016) pca mgh dataset (66 images) auc: 0.88 acc: 81.51±2.91 (huang et al., 2017) pca netherland cancer institute (778 images), vancouver general hospital (664 images) and public dataset. [stanford tissue microarray consortium web portal and the ucsb bio-segmentation benchmark dataset] (45 images) acc: 85% incomplete supervision is implemented via two major techniques, namely active learning which involves a human expert to label selected unlabeled instances considered as the most valuable, and semisupervised learning that makes use of labelled data and some basic assumptions about data distribution to predict labels of unlabeled data. for more details, the reader is referred to (zhou, 2018). inexact supervision uses the multiple-instance learning (mil) technique introduced by (dietterich et al., 1997). in mil, cnns inputs consist of multiple instances of images (called a bag of images), and the bag labels are obtained using two methods that are instance-based and bag-based methods. for more details, the reader is referred to (herrera et al., 2016). in inaccurate supervision, training data may contain erroneous labelling. a typical scenario is when some of the labels are obtained through the crowdsourcing technique introduced by (brabham, 2008). in this technique, non-expert users registered to a common platform provide labels for the unlabeled data. all published works applying methods grouped under this category are using the backpropagation approach for filter learning. albarqouni et al. (2016) proposed a 5-layer cnn model (aggnet) for mitotic detection. a crowdsourcing platform, called crowdflower was developed to annotate the unlabeled images. users registered on this platform are able to access the samples of labelled images and participate in the annotation process. this model was trained for 100 epochs in geforce gt 750m on the miccaiamida13 challenge dataset that consists of 318 hpf images for training and 22 hpf images for testing, along with 5500 labelled patches from the crowdflower platform (550 patches annotated by 10 participants). the authors reported an auc of 0.87. (das et al., 2018) proposed a mil-cnn with new multiple instance-pooling layers that extract features from the final fully connected layer for classification. a bag with multiple instances is obtained by extracting regions of interest (roi) from wsi. each patient is assigned a bag containing 25 to 60 randomly chosen patches of size 224x224 extracted from the breakhis dataset. their model achieved patient-level accuracies of 89.52% (40x), 89.6% (100x), 88.4% (200x) and 87.67 (400x). (sudharshan et al., 2019) applied on breakhis dataset milcnn model initially proposed by (sun et al. 2016). in their work, they considered two different settings; in the first set, each patient is considered as a bag which is labelled by its diagnosis while in the second set 1000 patches are randomly extracted from each image to constitute bags assigned the image label. the model was trained for 80 to120 epochs and achieved patient-level accuracies of 86.9 ± 5.4% (40x), 85.7 ± 4.8% (100x), 85.9 ± 3.9% (200x), and 83.4 ± 5.3% (400x) and image-level accuracies of 86.1 ± 4.2% (40x), 83.8 ± 3.1% (100x), 80.2 ± 2.6% (200x), and 80.6 ± 4.6% (400x). (jaiswal et al., 2019) applied a modified densenet-201 proposed by (g. huang and weinberger, 2018) where they substituted the last two fully connected layers with a global-max and global-average pooling layers. the model was trained with a small labelled dataset then the trained model is applied on a batch of the unlabeled dataset to obtain a pseudo-labelled dataset that was used to explore discriminating features of the unlabeled dataset. during the training, the backpropagation algorithm optimizes the objective function that combines the loss from the predictions of the pseudolabelled dataset with that from the labelled dataset. recent cnn-based techniques for breast cancer histology image classification 50 their model was applied to pcam (patchcamelyon) dataset that includes labelled and unlabeled images consisting of 327,680 patches of size 96x96 pixels. they trained the model for 5 to 7 epochs on nvidia tesla p100 and reported an auc of 0.98. • methods using forward filter learning in this approach, predefined filters are used in the convolutional layers instead of learning them through the traditional backpropagation algorithm. these methods are trained for a short period at a low computational cost and they can perform well with small-sized datasets. two pioneer research works have adopted this approach for learning filters, namely scatnet (bruna, j. and mallat, s. 2013) and pcanet (chan et al., 2015). the formerly used wavelet operators to define the filters while the latter used pca method to define them. shi et al. (2016) proposed a method inspired by pcanet called color pattern random binary hashingbased pcanet (c-rbh-pcanet). in this method, filters learned using a pca-based technique were applied to a 2-block cnn (each including a convolutional layer, non-linearity layer and a pooling layer). they reported an auc and accuracy of 0.88 and 81.51%, respectively on a dataset that consists of 1000 patches from 66 wsi of size 800x1800 from the mgh (massachusetts general hospital) dataset. (huang et al., 2017) proposed a 5-layer cnn model. the filters were learned using principal component analysis on the public datasets (university of pittsburgh school of medicine,upsm and the ucsb bio-segmentation, ucsb-bs benchmark dataset), and a sample of different types of histology images obtained through the google search engine. they achieved an accuracy of 85% on randomly selected 12,000 patches of size 50x50 pixels from three datasets (netherland cancer institute dataset, vancouver general hospital dataset, and the upsm and ucsb-bs public datasets). 5. discussion in this paper, the focus was on how cnn-based methods addressed the two major challenges faced by any cnn-based model namely the design of an appropriate cnn architecture and the acquisition of an adequate labelled dataset. regarding the cnn architecture design, the survey showed that the early methods adopted a handcrafted approach that builds from scratch the adequate cnn architecture based on trial and error. these methods are characterized by their relatively small depths (mostly less than 7 layers), reasonable training time and epoch number and moderate to good accuracy (mostly from 70% to 90%). tuning the values of the cnn hyperparameters such as the number of layers, number of filters and their sizes, etc. is tedious and timeconsuming. besides, the identified architectures are difficult to adapt to new datasets. with the event of performant deep cnns such as alexnet (8 layers), vggnet (16 and 19 layers), googlenet (22 layers) and resnet (152 layers), transfer learning became the preferred approach for designing image classifiers. the proposed methods attained better accuracies on breast cancer histology images (from mid-80% to mid-90%) with much less design effort, as most of the time the design consists of retraining and /or modifying the few last layers of a predesigned cnn architecture. the main drawback of these methods is the large number of learned weights that have to be stored. very recently network architecture search (nas) techniques have been proposed. they aim to automatically design optimized cnn architectures with efficient hyperparameter setting that suits the dataset under consideration. designed cnn can achieve very high classification accuracies, the only work we found addressing breast cancer histology image classification reported an accuracy of 99%. the main drawback of these methods resides in the significant search time required to identify the appropriate architecture. regarding the acquisition of an adequate labelled dataset, the survey showed that proposed breast cancer histology image classification methods addressed the problem in different ways. some methods make use of the large size of the whole-slide images or wsi (about 2040 x 1536) to extract a significant number of small labelled patches of a few hundred pixels in each dimension. the resulting images are then used to train and test the designed cnn models which lead to achieving accuracies in the range of mid-80% to mid90%. other methods use pre-trained models that usually do not need a large number of images for training because of the huge size of the dataset used to design the pre-trained cnn. when there is a need to further increase the size of the dataset, two approaches are applied: data augmentation and weakly supervised learning. in data augmentation, new images/patches are generated from collected labelled data by applying image transformations such as rotations, zooming, illumination variations, cropping etc. obtained classification accuracies are usually very high (from 85% to 99%). weakly supervised learning acquires labelled images through three techniques crowdsourcing, bagging or the semi-supervised technique proposed by chapelle et al. methods adopting these techniques achieved good classification accuracies (from 83% to 89%). recently, feedforward approaches for learning filters were proposed to overcome the drawbacks of the traditional backpropagation method including the need for large labelled datasets. proposed methods achieve good accuracy levels (from 81% to 85%) with a few layers and reasonable size of labelled datasets. 6. conclusion in this paper, we surveyed recent cnn-based methods for breast cancer histology image classification. the surge in high-performance gpu computing coupled with the availability of a large amount of training data has made cnn achieve state-of-the-art performance in arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 51 the classification tasks. although the feature representation from pre-trained models has been learned from natural images, these models achieve high classification accuracy on histological images and therefore are widely applied. recently, training with few samples is getting more attention from the research community using histology images because of the limited size of this type of dataset. this survey showed that the use of automatic architecture design and feedforward filter learning approaches have not been fully explored for the analysis of breast histology images despite the impressive results obtained by the former in classifying images and the promising perspectives offered by the latter to deal with the lack of labelled histology image datasets. therefore, we believe these two directions need to be further explored and could be of great potential and value in the future. conflict of interest the authors declare that there are no conflicts of interest regarding this publication. funding no funding was received for this study. reference albarqouni, s., baur, c., achilles, f., belagiannis, v., demirci, s., and navab, n. (2016). aggnet: deep learning from crowds for mitosis detection in breast cancer histology images. ieee transactions on medical imaging, 35(5), 1313–1321. https://doi .org/10.1109/tmi.2016.2528120 alom, m. z., yakopcic, c., nasrin, m. s., taha, t. m., and asari, v. k. (2019). breast cancer classification from histopathological images with inception recurrent residual convolutional neural network. journal of digital imaging, 32(4), 605–617. anwar, f., attallah, o., ghanem, n., and ismail, m. a. (2020). automatic breast cancer classification from histopathological images. 2019 international conference on advances in the emerging computing technologies, aect 2019, 16–21. https://doi.org/ 10.1109/aect47998.2020.9194194 araújo, t., aresta, g., castro, e., rouco, j., aguiar, p., eloy, c., polónia, a., and campilho, a. (2017). classification of breast cancer histology images using convolutional neural networks. plos one, 12(6). bayramoglu, n., kannala, j., and heikkila, j. (2017). deep learning for magnification independent breast cancer histopathology image classification. proceedings international conference on pattern recognition, 2440–2445. https://doi.org/10.1109 /icpr.2016.7900002 bejnordi, b. e., zuidhof, g., balkenhol, m., hermsen, m., bult, p., van ginneken, b., karssemeijer, n., litjens, g., and van der laak, j. (2017). context-aware stacked convolutional neural networks for classification of breast carcinomas in whole-slide histopathology images. 1–13. https://doi.org/10.1117/1.jmi.4.4.044504 boureau, y.-l., bach, f., lecun, y., and ponce, j. (2010). learning mid-level features for recognition. 2010 ieee computer society conference on computer vision and pattern recognition, 2559–2566. boureau, y.-l., ponce, j., and lecun, y. (2010). a theoretical analysis of feature pooling in visual recognition. proceedings of the 27th international conference on machine learning (icml-10), 111– 118. brabham, d. c. (2008). crowdsourcing as a model for problem-solving: an introduction and cases. convergence, 14(1), 75–90. https://doi.org/10.1177 /1354856507084420 budak, ü., cömert, z., rashid, z. n., \csengür, a., and ç\ibuk, m. (2019). computer-aided diagnosis system combining fcn and bi-lstm model for efficient breast cancer detection from histopathological images. applied soft computing, 85, 105765. chakraborty, s., aich, s., kumar, a., sarkar, s., sim, j. s., and kim, h. c. (2020). detection of cancerous tissue in histopathological images using dual-channel residual convolutional neural networks (dcrcnn). international conference on advanced communication technology, icact, 2020, 197–202. https://doi.org/10.23919/icact48636.2020.9061289 cire\csan, d. c., giusti, a., gambardella, l. m., and schmidhuber, j. (2013). mitosis detection in breast cancer histology images with deep neural networks. international conference on medical image computing and computer-assisted intervention, 411– 418. cruz-roa, a., basavanhally, a., gonzález, f., gilmore, h., feldman, m., ganesan, s., shih, n., tomaszewski, j., and madabhushi, a. (2014). automatic detection of invasive ductal carcinoma in whole slide images with convolutional neural networks. 9041(216), 904103. https://doi.org/10.1117 /12.2043872 cruz-roa, a., gilmore, h., basavanhally, a., feldman, m., ganesan, s., shih, n., tomaszewski, j., madabhushi, a., and gonzález, f. (2018). highthroughput adaptive sampling for whole-slide histopathology image analysis (hashi) via convolutional neural networks: application to invasive breast cancer detection. plos one, 13(5). das, k., conjeti, s., roy, a. g., chatterjee, j., and sheet, d. (2018). multiple instance learning of deep convolutional neural networks for breast histopathology whole slide classification. proceedings international symposium on biomedical imaging, 2018-april(isbi), 578–581. https://doi.org/10.1109/ isbi.2018.8363642 deng, j., dong, w., socher, r., li, l.-j., li, k., and fei-fei, l. (2009). imagenet: a large-scale hierarchical image database. 2009 ieee conference on computer recent cnn-based techniques for breast cancer histology image classification 52 vision and pattern recognition, 248–255. dietterich, t. g., lathrop, r. h., and lozanopérez, t. (1997). solving the multiple instance problem with axis-parallel rectangles. artificial intelligence, 89(1–2), 31–71. https://doi.org/10.1016/s0004-3702 (96)00034-3 goodfellow, i. j., warde-farley, d., mirza, m., courville, a., and bengio, y. (2013). maxout networks. arxiv preprint arxiv:1302.4389. han, z., wei, b., zheng, y., yin, y., li, k., and li, s. (2017). breast cancer multi-classification from histopathological images with structured deep learning model. scientific reports, 7(1), 1–10. https://doi.org/10.1038/s41598-017-04075-z he, k., zhang, x., ren, s., and sun, j. (2016). deep residual learning for image recognition. proceedings of the ieee conference on computer vision and pattern recognition, 770–778. he, k., zhang, x., ren, s., and sun, j. (2015). delving deep into rectifiers: surpassing human-level performance on imagenet classification. proceedings of the ieee international conference on computer vision, 1026–1034. herrera, f., ventura, s., bello, r., cornelis, c., zafra, a., sánchez-tarragó, d., vluymans, s., herrera, f., ventura, s., bello, r., cornelis, c., zafra, a., sánchez-tarragó, d., and vluymans, s. (2016). multiple instance learning. in multiple instance learning. https://doi.org/10.1007/978-3-319-477596_2 huang, y., zheng, h., liu, c., ding, x., and rohde, g. k. (2017). epithelium-stroma classification via convolutional neural networks and unsupervised domain adaptation in histopathological images. ieee journal of biomedical and health informatics, 21(6), 1625–1632. https://doi.org/10.1109/jbhi.2017.2691738 iarc. (2020). globocan 2020: new global cancer data. https://www.uicc.org/news/globocan2020-new-global-cancer-data intercollegiate, s., and network, g. (2005). sign guideline no. 84 management of breast cancer in women. december. jaiswal, a. k., panshin, i., shulkin, d., aneja, n., and abramov, s. (2019). semi-supervised learning for cancer detection of lymph node metastases. http://arxiv.org/abs/1906.09587 jin, x., xu, c., feng, j., wei, y., xiong, j., and yan, s. (2016). deep learning with s-shaped rectified linear activation units. thirtieth aaai conference on artificial intelligence. kaggle dataset. (2018). https://www.kaggle.com /c/histopathologic-cancer-detection kausar, t., wang, m. j., idrees, m., and lu, y. (2019). hwdcnn: multi-class recognition in breast histopathology with haar wavelet decomposed image based convolution neural network. biocybernetics and biomedical engineering, 39(4), 967–982. https://doi.org/10.1016/j.bbe.2019.09.003 koohbanani, n. a., qaisar, t., shaban, m., gamper, j., and rajpoot, n. (2018). significance of hyperparameter optimization for metastasis detection in breast histology images. in computational pathology and ophthalmic medical image analysis (pp. 139–147). springer. krizhevsky, a. (2009). learning multiple layers of features from tiny images. … science department, university of toronto, tech. …, 1–60. https:// doi.org/10.1.1.222.9220 krizhevsky, a., and hinton, g. (2010). convolutional deep belief networks on cifar-10. unpublished manuscript, 1–9. http://scholar. google.com/scholar?hl=enandbtng=searchandq=intitl e:convolutional+deep+belief+networks+on+cifar -10#0 krizhevsky, a., sutskever, i., and hinton, g. e. (2012). imagenet classification with deep convolutional neural networks. advances in neural information processing systems, 1097–1105. kröse, b., krose, b., van der smagt, p., and smagt, p. (1993). an introduction to neural networks. laxmisagar, h. s., and hanumantharaju, m. c. (2021). design of an efficient deep neural network for multi-level classification of breast cancer histology images. in intelligent computing and applications (pp. 447–459). springer. lecun, y. a., bottou, l., orr, g. b., and müller, k.-r. (2012). efficient backprop. in neural networks: tricks of the trade (pp. 9–48). springer. litjens, g., sánchez, c. i., timofeeva, n., hermsen, m., nagtegaal, i., kovacs, i., kaa, c. h. van de, bult, p., and ginneken, b. van. (2016). deep learning as a tool for increased accuracy and efficiency of histopathological diagnosis. nature publishing group, january, 1–11. https://doi.org/10.1016/ j.media.2020.101813 maas, a. l., hannun, a. y., and ng, a. y. (2013). rectifier nonlinearities improve neural network acoustic models. proc. icml, 30(1), 3. mehra, r. (2018). breast cancer histology images classification : training from scratch or transfer learning ? ict express, xxxx. https://doi.org/10.1016/j.icte.2018.10.007 motlagh, n. h., jannesary, m., aboulkheyr, h., khosravi, p., elemento, o., totonchi, m., and hajirasouliha, i. (2018). breast cancer histopathological image classification: a deep learning approach. biorxiv, 242818. munien, c., and viriri, s. (2021). classification of hematoxylin and eosin-stained breast cancer histology microscopy images using transfer learning with efficientnets. computational intelligence and neuroscience, 2021. oyelade, o. n., and ezugwu, a. e. (2021). a bioinspired neural architecture search based convolutional neural network for breast cancer detection using histopathology images. scientific reports, 11(1), 1–28. https://doi.org/10.1038/s41598021-98978-7 roy, k., banik, d., bhattacharjee, d., and arunadevi karuppasamy, a. abdesselam, r. hedjam, h. zidoum , and maiya al-bahri 53 nasipuri, m. (2019). patch-based system for classification of breast histology images using deep learning. computerized medical imaging and graphics, 71, 90–103. https://doi.org/10.1016/j.compmedimag.2018.11.003 ruifrok, a. c. (2019). quantification of histochemical staining by color deconvolution. journal of chemical information and modeling, 53(9), 1689– 1699. shi, j., wu, j., li, y., zhang, q., and ying, s. (2016). histopathological image classification with color pattern random binary hashing-based pcanet and matrix-form classifier. ieee journal of biomedical and health informatics, 21(5), 1327–1337. sickles, e. a. (1997). breast cancer screening outcomes in women ages 40 – 49 : clinical experience with service screening using modern mammography. 99–104. simonyan, k., and zisserman, a. (2014). very deep convolutional networks for large-scale image recognition. 1–14. https://doi.org/10.1016/j.infsof .2008.09.005 song, y., zou, j. j., chang, h., and cai, w. (2017). adapting fisher vectors for histopathology image classification. 2017 ieee 14th international symposium on biomedical imaging (isbi 2017), 600– 603. spanhol, f. (2016). a dataset for breast cancer histopathological a dataset for breast cancer histopathological image classification. 63(november 2015), 1455–1462. https://doi.org/10.1109/ tbme.2015.2496264 spanhol, fabio a, cavalin, p. r., oliveira, l. s., petitjean, c., and heutte, l. (2017). deep features for breast cancer histopathological image classification. 1868–1873. spanhol, fabio a, oliveira, l. s., petitjean, c., and heutte, l. (2015). a dataset for breast cancer histopathological image classification. ieee transactions on biomedical engineering, 63(7), 1455– 1462. spanhol, fabio alexandre, oliveira, l. s., petitjean, c., heutte, l., cavalin, p. r., oliveira, l. s., petitjean, c., heutte, l., sudharshan, p. j., petitjean, c., spanhol, f. a., eduardo, l., heutte, l., and honeine, p. (2016). breast cancer histopathological image classification using convolutional neural networks. proceedings of the international joint conference on neural networks, 2016-octob, 2560– 2567. https://doi.org/10.1109/ijcnn.2016.7727519 sudharshan, p. j., petitjean, c., spanhol, f., eduardo, l., heutte, l., and honeine, p. (2019). multiple instance learning for histopathological breast cancer image classification. expert systems with applications, 117, 103–111. https://doi.org/10.1016 /j.eswa.2018.09.049 szegedy, c., liu, w., jia, y., sermanet, p., reed, s., anguelov, d., erhan, d., vanhoucke, v., and rabinovich, a. (2015). going deeper with convolutions. proceedings of the ieee conference on computer vision and pattern recognition, 1–9. trottier, l., gigu, p., chaib-draa, b., and others. (2017). parametric exponential linear unit for deep convolutional neural networks. 2017 16th ieee international conference on machine learning and applications (icmla), 207–214. vang, y. s., chen, z., and xie, x. (2018). deep learning framework for multi-class breast cancer histology image classification. lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), 10882 lncs, 914–922. https://doi .org/10.1007/978-3-319-93000-8_104 vesal, s., ravikumar, n., davari, a., … s. e.-… c. i., and 2018, undefined. (n.d.). classification of breast cancer histology images using transfer learning. springer,1,https://link.springer.com/chapter/10.1007/ 978-3-319-93000-8_92 vo, d. m., nguyen, n. q., and lee, s. w. (2019). classification of breast cancer histology images using incremental boosting convolution networks. information sciences, 482, 123–138. https://doi.org /10.1016/j.ins.2018.12.089 wahab, n., khan, a., and lee, y. s. (2017). twophase deep convolutional neural network for reducing class skewness in histopathological images based breast cancer detection. computers in biology and medicine, 85(november 2016), 86–97. https://doi.org /10.1016/j.compbiomed.2017.04.012 wang, h., cruz-roa, a., basavanhally, a., gilmore, h., shih, n., feldman, m., tomaszewski, j., gonzalez, f., and madabhushi, a. (2014). mitosis detection in breast cancer pathology images by combining handcrafted and convolutional neural network features. journal of medical imaging, 1(3), 034003. https://doi.org/10.1117/1.jmi.1.3.034003 wistuba, m. (n.d.). a survey on neural architecture search. xu, b., wang, n., chen, t., and li, m. (2015). empirical evaluation of rectified activations in convolutional network. arxiv preprint arxiv:1505.00853. zagoruyko, s., and komodakis, n. (2016). wide residual networks. british machine vision conference 2016, bmvc 2016, 2016-septe, 87.187.12. https://doi.org/10.5244/c.30.87 zhou, z.-h. (2018). a brief introduction to weakly supervised learning. national science review, 5(1), 44–53. zoph, b., and le, q. v. (2017). neural architecture search with reinforcement learning. 5th international conference on learning representations, iclr 2017 conference track proceedings, 1–16. 1. introduction 2. convolutional neural network for histology image classification 3. cnn architecture design • manually-designed cnn architectures • pre-trained cnn architectures • automatically-designed cnn architecture 4. lack of labeled histology datasets • commonly used approaches dealing with small histology datasets • cnn using weakly-supervised learning • methods using forward filter learning 5. discussion 6. conclusion conflict of interest funding d:\desktop\paper_7.xps tjer 2013, vol. 10, no. 2, 78-86 ____________________________________ *corresponding author’s e-mail: amtex.uitm@gmail.com analysis on development of transverse sectioned weld zone using fem verified with multipulsed resistance seam welding n muhammad and yhp manurang* faculty of mechanical engineering, universiti teknologi mara, 40450 shah alam, malaysia received 1 june 2011; accepted 2 march 2013 abstract: abstract: this paper details an investigation, through an experimental study, of the development of weld nuggets and a heat-affected zone (haz) in resistance seam welding (rsew) using a numerical simulation approach. sysweld software for the simulation of heat treatment, welding, and welding assembly was utilized for the simulation process. the integrated spot weld advisor (swa) in sysweld was applied to simulate the rsew model using a two-dimensional axis-symmetric fe model with customized electrode meshing. the thermal-mechanical-electrical characteristic and contact condition were taken into account throughout this study. the developed model comprised a transverse cross section for welding two layers of low carbon steel with a thickness of 1 mm. for the experimental verification, three-pulsed rsew with two different current stages was carried out. it was discovered that this program code, spotweld advisor, when used with the meshing method, was capable of offering results that were in agreement with physical experiments. keywords: customized electrode meshing, finite element method (fem), resistance seam welding, swa sysweld, visual-mesh hazrsew sysweld (swa sysweld(rsew rsewadvisor spotweld 79 n muhammad and yhp manurang 1. introduction resistance seam welding (rsew) is widely employed for the fabrication of automotive gas tanks which require water or air tightness. highly reliable welding procedures which produce a high quality result are required, as leakage within the system renders the tank unsafe. in general, the quality of the weld is strongly dependent on the maintenance of the rsew machine and the selection of optimum welding conditions (manurung et al. 2010; murakawa and minami 2002 and murakawa 2003). rsew was originally a modification of resistance spot welding (rsw), which produces a continuity of overlapping spot welds by means of rotating wheel electrodes. this method forms a continuous, leak-tight joint. in order to obtain a good quality weld nugget, many experiments have to be conducted which can require significant cost and time. one effective solution to this problem is to use a computer simulation (zhang 2003; eisazadeh et al. 2010; srikunwong et al. 2002). in our study, a finite element method (fem) for seam welding was developed, and appropriate conditions were studied throughout the simulation process. there are two possible two-dimensional (2-d) models of electrode geometries for seam weldinglongitudinal cross section and transverse cross section (murakawa and minami 2002). in this research, only the transverse section of electrode geometries and sheet metal were developed using customized electrode meshing with mesh generator software visualmesh 5.2 (esi group, paris, france) followed by exporting the generated mesh into sysweld (esi group, paris, france) for the numerical simulation of seam welding. since there is no specific software to replicate the seam welding process, the integrated spot welding advisor (swa) in sysweld 2009 was utilized. within this tool, all common parameters are readily available for spot welding and can be manipulated using customized meshing methods to simulate seam welding. 2. basic principle of rsew 2.1 governing equation and boundary conditions the general principle of rsew is the joule heating law (song et al. 2005) where the heat, q, is generated depending on three basic factors as expressed in the following formula: (1) where i is the current passing through the metal combination, r is the resistance of the base metals and the contact interface, and t is the time of the current flow. force is applied before, during, and after the application of electric current to maintain the electrical connectivity and to provide the pressure necessary to produce defect-free seam welds. this process involves strong interactions between electrical, thermal, metallurgical, and mechanical fields (feng et al. 1998; feulvarch et al. 2004; hou et al. 2007). rsew shares a similar principle regarding spot welding with different electrode geometry. two rotating disc electrode wheels are used to apply current and force, with the discs rotating either continuously or intermittently. the seam welding electrode is expected to be similar to spot welding as can be seen from the side view (fig. 1). in this investigation, a flat electrode wheel tip was used in which the contact radii were considered constant. according to the axis-symmetry and 2-d problems with full coupling between the electrical and thermal phenomena, the coupling system governing equations are as follows (srikunwung et al. 2002): (2) (3) where t and v are the temperature and the scalar electrical potential respectively. , and represent the density, the thermal conductivity and the electrical conductivity of the media with respect to temperature dependency characteristic. h is the enthalpy with wheel electrode material wheel electrode figure 1. transverse sectioned scam electrode and weld 80 analysis on development of transverse sectioned weld zone using fem verified with multipulsed ... temperature dependency and q is internal heat generation. the current flow in an electrical conductance generates joule heating effect. the internal heat dissipation is represented in term of gradv. .gradv. since a 50 hz sine wave ac current was used, the welding current can be described with standard sinusoidal signal. therefore, the root-mean square value of the welding current can be given by (li et al. 2007). (4) (5) in this simulation, a three-pulsed current was applied consisting of 1 cycle for the first, 3 cycles for the second, and 1 cycle for the third current stages (fig. 2). in general, multi-pulsed current is applied when either too thick or too thin a base material is to be welded. using multi-pulsed technology, the current needs a certain amount of cooling time in order to avoid the occurrence of an oxide layer (cuff 1998). the materials used in this research were low carbon steel with a thickness of 1.0 mm and electrodes made of copper alloy. table 1 shows the material composition obtained from the data base of the fem software, international standards, and an investigation using a spark emission spectrometer. 2.2 electrothermal contact conditions one of the most important considerations in modeling rsew is the contact condition at the electrodesheet and sheet-sheet interfaces (murukawa and minami 2002). under the application of the electrode force, the contact element becomes in contact. when it is in contact, enough stiffness is given to the contact element and the same values of the electric and thermal conductivity of the bulk materials are given to the contact element. the contact resistance is a function of contact pressure (feng et al. 1998). as pressure increases, the resistance reduces and the heat generation across the object boundary is also reduced (chigurupati et al. 2010). the electrical and thermal contact resistance of the electrodes and the sheets are temperature dependent. sysweld 2009 defines the function of contact resistances with respect to temperature, using fortran. in this simulation study, the thermal and electrical contact resistance between electrode-sheet and sheet-sheet was considered a weak coupling. in the weak coupling option, the thermal analysis was performed during the current application, followed by mechanical analysis at the end of the current application. 3. simulation procedure and experimental set-up 3.1 sysweld's customized meshing strategy sysweld in swa offers two different methods for solving the rsw problem: a predefined method and a customized meshing method. in the predefined method, a standard electrode form is selected and sysweld generates the meshing automatically after parameters such as curvature, radius of the electrode, internal radius, external radius, and height of the electrode are defined. while using the customized electrode meshing, electrode skins and geometry must be defined first and subsequently meshed according to the actual dimensions. in this research, the later meshing method was used to simulate seam welding created using visual-mesh, version 5.2 as shown (esi group, paris, france) (fig. 3). the characteristics and functions of the curves are summarized as follows (manurung et al. 2010). 10ka 9ka 10ka 11ka 9ka 11ka weld time 1 cycle 3 cycles 1 cycle figure 2. schematic illustration of multipulsed weld current table 1. material composition of low carbon steel 81 n muhammad and yhp manurang i. wat_el_tra_1 contains elements where thermal transfer between electrodes and water is modeled. ii. cure_ appl_1 contains elements where current density and pressure are applied. iii. el_ri_side_1 contains elements where thermal transfer between electrodes and air is modeled. iv. r_ctc_inf_1 contains elements where thermal transfer between electrodes and air is modeled, with additional sliding contact elements between electrodes and sheets. v. l_ctc_inf_1 contains elements where thermal transfer between electrodes and sheets is modeled, with additional sliding or sticking (depending on the pressure application) contact elements between electrodes and sheets as well. vi. elect_1 contains 2-d elements where electrothermo-metallurgical and mechanical properties are affected. 3.2 experimental setup in this study, electrode size, material thickness, weld force, squeezing cycle, and weld cycle of the simulation and the experimental study were set to be constant throughout the investigation. the variable parameter was the applied three-pulse weld current, which was 10-9-10 ka and 11-9-11 ka. the details of the welding parameters are shown in table 2. the experimental investigation on rsew was conducted using nash 10-100 (fig. 4). to measure the width and height of the weld nugget, and the heat-affected zone (haz), the welded metal was cut transversely from the middle position using a common cutting machine. in order to assure the precision of the specimen dimension, it was etched using a 2% nital solution. the macrograph of the weld zone was captured using a metallurgical microscope interfaced with an image analysis system. the development of the weld nugget and haz using the fem simulation approach is represented in figs. 5 and 6, respectively. in the simulation for 5 weld cycles, a three-pulsed weld current of 10-9-10 ka and 11-9-11 ka were applied in which the first, second, and third pulses were 1 cycle, 3 cycles and 1 cycle for respective current stages. figure 3. meshing strategy of electrode and group definition of skin elements table 2. simulation and experimental process parameters 82 analysis on development of transverse sectioned weld zone using fem verified with multipulsed ... 4. results and discussions figure 5 shows that, for the multi-pulsed weld current of 10-9-10 ka, the nugget width development started at 0.13 seconds, reaching the maximum value of 4.40 mm at 0.19 seconds, while for 11-9-11 ka, it started earlier (0.12 seconds) and reached the maximum value of 4.60 mm at 0.19 seconds. the final nugget was formed at the end of the hold time because there was some deformation in the welding zone when the well-meant cooled due to the electrode pressure and material shrinkage. figure 4. rsew machine (brand: nash, type: 10-100) figure 5. graph for development of weld nugget time (s) width of weld nugget 10.9.10ka multipulsed weld current 11.9.11ka multipulsed weld current 0.00 0.10 0.20 0.30 0.40 0.50 5.00 4.00 3.00 2.00 1.00 0.00 max: 4.60mm, 0.19s max: 4.39mm, 0.19s 83 n muhammad and yhp manurang figure 5 also shows that the nugget size increased with the weld current. after applying the weld current, the temperature of the work pieces rose by the generation of joule heat. in the joule heating formula, heat generation due to electrical current is more considerable compared with heat due to time or force parameters. therefore, a higher weld current generated more heat and formed a bigger weld nugget. figure 6 shows the development of haz. as the width of the weld nugget increased, the width of the haz also increased in proportion to the weld current. figure 7 shows the simulated temperature profile and nugget growth at the time when the nugget started to form. at the start of the welding process, the temperature at the center of faying surface increased very quickly. the highest temperature was always at the time (s) figure 6. graph for development of haz width of haz figure 7. temperature distribution at the time nugget started to form: (a) 10-9-10 ka weld current and (b) 11-9-11 ka weld current (a) (b) 10.9.10ka multipulsed weld current 11.9.11ka multipulsed weld current 0.00 0.10 0.20 0.30 0.40 0.50 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 max: 5.67mm, 0.18s max: 5.98mm, 0.39s 84 analysis on development of transverse sectioned weld zone using fem verified with multipulsed ... center of faying surface during the whole welding process. the nugget was shaped as a very flat ellipse. melting occurred first at the faying surface and then expanded to the surrounded material. as the temperature rose, the nugget kept growing. when the electrical current was switched off, the well-meant started to cool down so the temperatures of the nugget decreased quickly due to the loss of heat by convection of the air and cooling water. figure 8 shows the simulation result of the nugget height development, which reached 1.60 mm for 10-910 ka and increased to 1.70 mm for 11-9-11 ka. results from the simulations showed good comparison with the experiments (fig. 9). the experimental investigation results are presented in the form of macrographs as shown in fig. 9. from the macrographs, the width and height of the weld nugget and the haz were measured. the results obtained by simulation and experimentally are summarized in table 3. the results indicate that the width of the weld nugget, and the haz itself is proportional to the increasing welding current, in both simulations and the experimental process with the rsew. according to hamedi and pashazadeh, at a constant welding time, an increase of welding current affects the rise in nugget temperature (hamedi and pashazadeh 2008). welding current is the most dominant factor that affects the nugget area, such that increasing the welding current simultaneously increases the nugget area (darwish and al-dekhial 1999). 5. conclusions a simulation using customized electrode meshing, sysweld's swa, and experimental process of rsew on two sheet layers of low carbon steel with similar thicknesses was studied. it was discovered that a 2-d axis-symmetrical fem could be developed using this meshing method to model the thermomechanical-electrical coupling of the rsew for a transverse cross section of electrode geometry and material. it shows that the seam welding process can be simulated using swa in sysweld. this tool was used throughout this research in order to investigate figure 8. simulation result of nugget development: (a) 10-9-10 ka weld current (b) 11-0-11 ka weld current (a) (b) table 3. simulation and experimental results 85 n muhammad and yhp manurang the effect of various parameters on the formation of transverse-sectioned weld nugget. from this study, it can be concluded that fem simulation using sysweld's swa presents good agreement with experimental result in terms of the width and height of the weld nugget and haz. furthermore, the information on the temperature and time profile was also available. the fem simulation approach of the rsew provides a better understanding of the effects of welding parameters on the nugget formation. it provides a precise analysis, and saves both cost and time over the welding process. it is well known that the rate of weld nugget growth is mostly influenced by welding current. it is recommended that this study be continued further with other parameters such as materials, number of layers, sheet thicknesses, weld time, squeezing time, and electrode force. it is also recommended that various types of electrode meshing be investigated to determine their effect on nugget development and formation. furthermore, the experimental investigation should be enhanced by using optical technology to measure the real-time development of the weld zone. acknowledgments the authors would like to express their gratitude to the faculty of mechanical engineering and the advance manufacturing technology excellence centre (amtex) at the universiti technologi mara (uitm) in shah alam, malaysia, for the facilities and technical support. a special thank you is addressed to the malaysian ministry of higher education (mohe) for their financial support. references chigurupati p, chun bk, bandar a, wu wt (2010, finite element modeling of resistance spot welding process. international j. mater form 3:991994. cuff r (1998), using multiple-impulse (pulsation) resistance welding for coated material, the fabricator 36-39. darwish sm, al-dekhial sd (1999), statistical models for spot welding of commercial aluminum sheets. international j. machine tools manufacturing 39:1589-1610. eisazadeh h, hamedi m, halvaee a. 2010. new parametric study of nugget size in resistance spot welding using finite element method. j. of material and design 31:149-157. feng z, babu ss, santella ml, riemer bw, gould je (1998), an incrementally coupled electrical-ther figure 9. the etched macrographs with 2% nital for multipulsed weld current: (a) 10-9-10 ka and (b) 11-9-11 ka (a) (b) 86 analysis on development of transverse sectioned weld zone using fem verified with multipulsed ... mal-mechanical model for resistance spot welding. 5th international conference on trends in welding research, pine mountain, ga. feulvarch e, robin v, bergheau jm (2004), resistance spot welding simulation: a general finite element formulation of electrothermal contacts conditions, j. of materials processing technology 153-154:436-441. hamedi m, pashazadeh h (2008), numerical study of nugget formation in resistance spot welding. international j. mechanics 2:11-15. hou z, kim is, wang y, li c, chen c (2007), finite element analysis for the mechanical features of resistance spot welding process. j. of material processing technology 185:160-165. li yb, lin zq, hu sj, chen gl (2007), numerical analysis of magnetic fluid dynamics behaviors during resistance spot welding. j. of applied physics 101. manurung yhp, muhammad n, haruman e, abas sk, tham g, salleh km, chau cy (2010), investigation on weld nugget and haz development of resistance spot welding using sysweld's customized electrode meshing and experimental verification. asian j. of industrial engineering 2:63-71. murakawa h, minami h (2002), development of finite element method for seam welding and its application to optimization of welding condition. proceedings of the twelfth international offshore and polar engineering conference, kitakyushu, japan 344-347. murakawa h, minami h, kato t (2003), finite element simulation of seam welding process, trans. jwri 30:111-117. murakawa h (2003), simulation of resistance welding for selection of optimum welding conditions and process control. trans. jwri 32:215-218. song q, zhang w, bay n (2005), an experimental study determines the electrical contact resistance in resistance welding. welding journal 73-76. srikunwong c, dupuy t, bienvenu y (2002), a numerical and experimental study in resistance spot welding process, ecole des mines de paris, france. zhang w (2003), design and implementation of software for resistance welding process simulations. society of automotive engineers, inc. 105-113. microsoft word paper 2.docx the journal of engineering research, vol. 11, no. 2 (2014) 15-26 m-learning: the new horizon of learning at squ z. al-khanjari*, k. al-kindi, a. al-zidi and y. baghdadi department of computer science, college of science, sultan qaboos university, p.o. box 36, postal code 123, sultanate of oman received 11 september 2013; accepted 30 april 2014 abstract: m-learning extends the theory and practice of learning and mobility in converging technological environments. developing a smart course in order to improve the standard of education at sultan qaboos university (squ) is one of the main aims of the current authors. this proposal requires developing innovative applications for ubiquitous, mobile technologies for learning. this kind of development covers new educational and technological methods and concepts for supporting formal and informal learning. squ is currently using moodle as the open source e-learning management system to support and enhance traditional learning. although elearning in squ has proven its importance in enhancing traditional learning, it is limited to areas and locations where a personal computer exists. this constraint is a burden to many e-learning users who are staff or students, especially if they live in rural areas of oman. to overcome this drawback, an investigative survey of the importance of m-learning was designed and distributed to squ students. the results showed that almost all students favored and supported the idea and requested the implementation of a m-learning application. as a step forward, this paper proposes an extension for e-learning—a new m-learning tool to support learners who use mobile device technologies. our goal in introducing m-learning at squ is not to replace e-learning but to complement and improve it so both modalities are available, since each grants certain advantages to users. keywords: traditional learning, e-learning, m-learning, mobile application, j2me. אwאא  אאאאא   אאאאאאwאאאאk אאאאאא   א א    k  א    אא    אא  א א א   א  א אאאאkא kא  א،אאwאא،א،אא،.j2m *corresponding author’s e-mail: zuhoor@squ.edu.om 16 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi 1. introduction the importance of m-learning emerged because many learning outlets can be provided via mobile devices, which are a part of the daily culture of almost every student and teacher. m-learning, as a new learning technology, is defined as any sort of learning that happens when the learner is not at a fixed, predetermined location, or learning that happens when the learner takes advantage of the learning opportunities offered by mobile technologies (o’malley et al. 2003). m-learning uses new technologies to develop a simulated student culture and enhance the learning process. it also solves the formality of traditional learning and enhances the selfconfidence and self-esteem of students (khalaf et al. 2009). interactive m-learning could be supported by various mobile communication mechanisms such as voice communication access to a learning portal on the internet or learning through short message service (sms) (ting 2005). enabled by improved software, improved hardware, and evolving habits of mobile device users, the opportunities to develop smart courses have increased even more significantly in past few years. the continuous proliferation of personal smart phones with advanced web browsers has created an incredible opportunity that cannot be ignored. it is now possible to deliver course content across many platforms using mobile browsers. this alternative method has the potential to be more effective and gain wider acceptance by users over computer-based e-learning (haag 2011). with the smart course, learners would develop critical thinking skills, learning activities and creativity. the aim of the smart course is to improve the quality of learning, training, course organization and student presentation. this paper presents an ongoing research project which proposes a new m-learning tool for students to resolve e-learning limitations. the study focuses on the student-content learning relationship experimented at squ with m-learning. it aims to provide anywhere and anytime learning tools for squ students such that they may reach course content at any time and from any place. almost all squ students possess mobile devices. indeed, those devices enable students to view teacher announcements or other course material even while on the move. m-learning would also provide a creative environment for students as they enjoy technology’s advantages and create a social communication platform between students and their course content. this research is based on two questions: to what extent does m-learning at squ add value to the interactions of students in relation to content and teachers? and how can mlearning complement e-learning at squ? to answer these questions, a survey was conducted among squ students. the survey consisted of five sections: general information, the purpose of using mobile phones, pedagogical aspects, distance learning difficulties and improving learning difficulties. 2. related work m-learning started around 2000 when (sharples 2000) discussed the potential for new designs in personal mobile technologies which could enhance lifelong learning programs. since then, many educators have indicated positive perceptions about the potential of mobile learning. articles evaluating mlearning technologies are increasing. there are many examples of projects which have encouraged the use of m-learning technology and shown its advantages. in 2007, the international association for development of the information society (iadis) held the international conference on cognition and exploratory learning in the digital age (celda 2007). one paper presented a pilot case study involving learning aspect processes within a m-learning environment: math4mobile. math4mobile is a cellular application for mathematics learning. the paper discussed math4mobile, which could be installed on most cellular phones on the market, as an application designed to support mathematics learning. the researchers watched two pairs of students who studied a mathematics methods course using mobile applications. the case study found that the contribution of the mobile environment not only helped in making dynamic mathematical applications more available, but it also helped 17 m-learning: the new horizon of learning in squ to support the execution of mathematical tasks that were closer to the students’ experiences, which has the potential to enhance experiential learning (lai et al. 2007; galit and michal 2007; botzer and yerushalmy 2007). in another 2007 project, the higher education admission center in the sultanate of oman provided grade 12 students in omani schools with sms service to access the center (khalaf et al. 2009; al-mamari 2009). students sent an sms from their mobile phones to be able to access and modify their personal information, ask about the offered programs in universities and colleges, register with programs and get final grade results. in 2008, a new project on mobile math (mmath) commenced. it was a web application designed to run on both mobile devices and personal computers in malaysian primary schools by using open source technology. mmath focused on students aged 11 and 12 and facilitated learning mathematics using handheld devices. this m-learning application used a wireless access protocol (wap) gateway to enable communication between three groups of users: students, teachers and administrators and was designed for two learning environments: web browsers on personal computers and micro browsers on mobile devices. the computer application was different from the mobile application in that the users of the computer application had to login into the main page. teachers were able to set questions for quizzes and tests, analyze the student performance in student list and graph view, and post an announcement. the administrator had full privilege to control the system and users of the system. the server side of the system was developed using extensible hypertext markup language (xhtml), wireless markup language (wml), a hypertext preprocessor (php) and mysql, the world's second most widely used open-source relational database management system. by including the uniform resource locator (url) dialing function in the implementation, the student was able to call his teacher. a jpgraph module (asial corp, tokyo, japan) was used to implement the graph generating functionality. the application provided a progress tracking function that allows users to view their progress and student performance from time to time. also, there was an announcement function, which allowed teachers to post any news or update about the course. the user tests showed a significant improvement in terms of student satisfaction with and performance for the m-learning environment compared to a more traditional learning approach (saipunidzam et al. 2008; saipunidzam et al. 2010). myartspace, another m-learning project, engaged students through personal mobile technology to make connections between learning within the classroom and the outside environment (vavoula et al. 2009). in 2010, the department of computer science at the university of alcala in spain introduced a new mobile application designed for auto-assessment that allowed students to test their knowledge and expertise on a specific topic using questionnaires designed by their teachers. the system was developed using java technology and java micro edition (jme) for the mobile application (oracle corp., redwood city, california, usa) and extensible style sheet language transformation (xslt) sheets. the system consisted of three subsystems: a web server to store, deliver and evaluate online tests; a mobile application that students employ to connect to the server, download questionnaires and complete them; and a webbased from-end that is used to offer different functionalities to each kind of user (students, teachers and administrators) (luis et. al. 2010; de-marcos et al. 2010). existing programs that leverage mobile course delivery as part of their overall training strategy have been in existence in the corporate sector for several years. few studies have been conducted on the general effectiveness of this specific form of delivery in a military setting (haag 2011). in another study, wu and colleagues mentioned that mobile phones and personal digital assistants (pdas) are the most commonly used devices for m-learning but they may be replaced in the future by new emerging technologies (wu et al. 2012). also, trucano, who is the world bank's senior information and communication technology (ict) and education policy specialist, serves as the world bank's focal point within the education sector on issues at the intersection of technology use and education. he talked about the world bank’s infodev project. the world bank secured funding to do a global survey of 18 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi the use of mobile phones in education in developing countries based on the belief that the increasing availability of the small, connected computing devices more commonly known as “mobile phones” was going to have year 8.7 23.9 65.2 1st year 2nd / 3rd year 4th year and above discipline 59.2 40.8 science-based non-science-based figure 1. e-learning usage in squ. figure 1 shows that the final years and science-based students are more adopted to computers. therefore, it became obvious that they use the e-learning facilities more than others. (a) entertainment (games 22.4 30.6 34.7 12.3 monthly weekly daily not at all (b) communica tion (sending s mss to friends and rela tives) 2 8.2 89.8 0 monthly weekly da ily not at all (c) business 18.4 26.534.7 20.4 monthly weekly daily not at all figure 2. mobile phone usage. figure 2 displays that most students use their mobile phones daily in their entertainment, communication and business. increasing relevance to school systems around the world (trucano 2012). recently, gluga and colleagues contributed to finding foundations for mobile university curricula in terms of multiple learning goal sets (gluga et al. 2013). 3. development of squ m-learning application tools students at squ use and support e-learning in their studies. although this is the case, they do not complain about the lack of available elearning resources when they are away from campus. this lack of e-learning opportunities cannot help them to achieve their educational objectives. this section first checks the importance of introducing a m-learning application tool. then it provides the design of such tool. finally, it shows the implementation of the m-learning tool. 3.1 the importance of introducing a mlearning application tool due to the lack of e-learning facilities that can reach all users in all areas of the sultanate of oman, a questionnaire was distributed to squ students from all colleges. the aim of this survey was to show the potential significance of introducing new m-learning services in squ in order to improve the existing e-learning facilities. results of the survey are demonstrated in figs. 1 and 2. 19 m-learning: the new horizon of learning in squ (b) search for learning (search for information) 22.4 20.418.4 38.8 monthly weekly daily not at all figure 3. frequent of using mobile phones for study purposes. figure 3 shows that most of the students would like to use their smart phones for education purposes. but unfortunately, they aren’t able to do so because of the lack of m-learning applications in squ. the study showed that establishment of such application is very important for the students to overcome the limitations of e-learning. 3.2 design of the m-learning application tool the design of the m-learning application tool is described with the use case diagram illustrated in fig. 4. the application is divided into two separate subsystems: a) interface program for teacher, and b) interface application for student. teacher/instructor is able to login, add study material, upload assignment, post grades and post announcement. on the other hand, the student figure 4. use case diagram of the m-learning application tool. is able to login, display/download study material, download/submit assignments, view announcement and view grades. 3.3 implementation of the m-learning application tool implementation is the part of the process where software engineers actually program the code for the project. also, the implementation phase provides the conceptual foundation required for the design and development of complex distributed and real-time applications. the tool is implemented using visual basic language in microsoft visual studio 2010 environment, java2 micro edition in netbeans ide 9.6.1 environment and sql server 2008 r2. these facilities enable us to design the interface and implement the code easily. figure 5 provides the architectural design of m-learning application tool with the two 20 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi implemented interfaces for both subsystems mentioned in the use case diagrams. first, the student has to login via the main page mobile application interface by entering his/her username and password. as a result, valid login case menu screen will be displayed. the student will then be able to select any of the offered choices. mobile application interface for students is designed using j2me however, the teacher’s interface is developed using visual basic programming language. figure 5. the architectural view of the squ m-learning application. figure 6. an application on student’s mobile phones as client side. mobile application interface for students is designed using j2me (java for mobile application) (enrique ortiz and giguere, 2001). however, the teacher’s interface is developed using visual basic programming language. asp.net is used to create the web service, which is connected to oracle sql database. this connection is important to store announcements, study materials, quizzes, assignments and grades added by the teacher in the database. the web service then gets these required data by connecting to the same database to transform them to the application in students smart phones. sql web services courses and users student’s pageteacher’s page 21 m-learning: the new horizon of learning in squ the proposed tool comprises the following four different parts to form the complete system. they are explained further in the following sections. • an application on mobile phone as client (eg. student) side. • an application on teacher personal computer. • a web service connected to both applications. • sql database for storing and retrieving information about students and courses. 3.2.1 student interface students can download and save available study materials, announcements, assignments and grades while connecting to figure 7. the login screen of the teacher. figure 8. menu screen for the teacher. figure 9. open file dialog when teacher clicks on “browse” button. the internet. they can disconnect from internet after completing the download to view announcements and grades or print course materials and assignments. figure 6 shows some of the application screens in the students mobile phones (al-zeidi et al. 2011). 3.2.2 teacher interface teachers can prepare/upload course materials, store announcement and prepare/manage quizzes, assignments and grades. figure 7, shows the first form (eg. login form), which appears once the application program runs. this form includes two textboxes for entry of teacher username and 22 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi password. to login to the tool and establish connection to the database, the teacher should provide his/her username and password, which should be followed by clicking on the “login” button. the tool then checks the validation of the teacher entries. if the program failed to connect with database then an error message should appear. otherwise, if the program didn’t find a match of the username, password or both in the teachers’ data table, a message should appear requesting the teacher to reenter valid entries to be able to go to the next window. if the teacher entries are valid, a menu window appears and the teacher can select the wanted choice as shown in fig. 8. for each choice, a related window is displayed. for example, if the teacher clicks on “add course material” then the add course material window is displayed as shown in fig. 9. in order to upload the course material, the teacher should click on the “browse” button shown in fig. 9. as a result, open window is displayed through which the required files could be uploaded. this window is also displayed in fig. 9. teacher then selects the file and clicks on “upload” button to store the file in the database. 3.3.3. web service in visual basic environment we implemented a web service using asp.net. the web service written in vb language and has four functions. the function in asp.net is declared as web method() so that before any function declaration the statement <webmethod()> _ is included. to connect to the required database in the sql server we imported in the following libraries as indicated in the fig. 10 below. figure 10. code of connecting asp.net to sql server the code of connecting asp.net to sql server is similar to that used in connecting vb to sql server. in the following code, we connected to a local database. the following figure shows an example the data from table “tablename”. 3.3.4. j2me implementation in the environment of j2me we dragged the item we wanted to appear in the phone screen from the list of items. this is when we choose the tap screen. there is another tap which is “flow” which enabled us to make the relationship and connect channels between different screens. the code is written by clicking on “source” tap. the following figure is of our flow. to connect j2me to the web service we can use one of the following two ways: • implement connection using ksoap. • the second way is to use get http as shown in the following figure. 3.3.5. sql server implementation creation of the database is very direct in sql server. also, the table and columns creation were done quickly without needing to write any code. we created quires by clicking on “new query” tab and used them for altering tables and adding primary and foreign keys. after creating the database and tables we inserted values in id and password columns in both student and teacher tables. we also inserted values for cid (course id) and cdesc (course description). figures 16, 17, 18 and 19 provide the details about tables and columns and examples of used queries. 4. conclusions mobile technologies can play an important role in squ students' academic lives. mobile devices connect users to the world instantly, heightening access to information and enabling interactivity with others. applications such as m-courses that run on these devices let users not only consume but also discover and produce content. this can enhance the students learning and influence their learning preferences, both within and outside the classroom. to successfully adopt mobile technologies across the university, this paper presented the imports system.configuration imports system.data imports system.linq imports system.data.sqlclient 23 m-learning: the new horizon of learning in squ ongoing research work in developing mobile learning services for squ students. it discussed the needs for m-learning for achieving several goals, which could lead to a great success in learning process in our country. students in squ have heard about mobile learning technology and they are waiting for the time to be able to enjoy its benefits and facilities. for the future work, teachers’ interface could be enhanced by preparing it as a web page using vb.net. also, the tool could be improved by using sms facility to send messages to students whenever a new material is posted on the site of the smart course. on the other hand, instructors could be informed via sms when any assignment or project is submitted by the student. figure 11. web method in asp.net. figure 12. flow of j2me application. figure 13. code of import ksoap. <webmethod()> public function connect ()as dataset dim conn as sqlconnection dim dadauthors as sqldataadapter dim s as dataset = new dataset dim ds as dataset try conn = new sqlconnection ("data source=fyp06-lab5\mlearning;initial catalog=squmlearning;integrated security=sspi") dadauthors = new sqldataadapter("select filedirectory from tablename ",conn) conn.open() ds = new dataset dadauthors.fill(ds,"tablename") conn.close() return ds catch ex as exception msgbox(ex.message) end try return s end function import org.ksoap.*; import org.ksoap.transport.*; import org.ksoap.soapobject; 24 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi figure 14 shows the entire code of adding the url to connect with and store data in string buffer. figure 14. code of connect to url. figure 15. code of connect to url using get http. figure 16. databases list in sql server. figure 17. course table and its column specifications. figure 18. student table and its column specifications. public void dowebservice(string url) throws exception { stringbuffer sb = new stringbuffer(); string name = "phobos"; textbox textbox = null; soapobject client = new soapobject(url, "sayhello"); client.addproperty("name", name); httptransport ht = new httptransport(url, public void run() { httpconnection hc = null; datainputstream in = null; try { string url = getappproperty("imageloader-url"); hc = (httpconnection)connector.open(url); int length = (int)hc.getlength(); byte[] data = null; if (length != -1) { 25 m-learning: the new horizon of learning in squ figure 19. teacher table and its column specifications. references al-mamari ks (2009), the effect of mobile learning technologies on students’ listening skills in higher education in the sultanate of oman. thesis for the degree of master in education. department of curriculum and instruction. college of education. sultan qaboos university, oman. al-zeidi a, al-kindi k, al-khanjari z (2011), squ future: from e-learning to m-learning application development. the proceedings of the international information systems conference (iisc). college of commerce, sultan qaboos university, oman. botzer g, yerushalmy m (2007), mobile application for mobile learning. iadis international conference on cognition and exploratory learning in digital age (celda 2007):313-316. de-marcos l, josé-ramón h, eva g, antonio g, josé-javier m, josé-maría g, roberto b, salvador o, josé-antonio g, elena v, miriam m, salvador e (2010), a mobile learning tool to deliver online questionnaires. computer science department, university of alcalá, spain: 319-320. enrique oc, giguere e (2001), j2me: the complete reference. john wiley & sons, usa. galit b, michal y (2007), “mobile application for mobile learning. the international conference on cognition and exploratory learning in digital age (celda 2007): 313-316. gluga r, kay j, lever t (2013), foundations for mobile university curricula in terms of multiple learning goal sets. ieee transactions on learning technologies 6(1). jason h (2011), from elearning to mlearning: the effectiveness of mobile course delivery. interservice/industry training, simulation, and education conference (i/itsec). khalaf r, mukhi n, weerawarana s (2009), a pervasive distributed architecture for mlearning is presented based. journal of mobile learning and organization 3(2):128-147. lai ch, yang jc, chen fc, ho cw, chan tw (2007), affordances of mobile technologies for experiential learning: the interplay of technology and pedagogical practices. journal of computer assisted learning, 23(4):326-337. luis jrh, eva g, antonio g, josé-javier m, josé-maría g, roberto b, salvador o, joséantonio g, elena v, miriam m, salvador e (2010), a mobile learning tool to deliver online questionnaires. department of computer science, university of alcalá, spain 319–320. o’malley c, vavoula g, glew j, taylor j, sharples m, lefrere p (2003), mobilearn wp 4 – guidelines for learning /teaching /tutoring in a mobile environment. mobilearn / uon,uob,ou / d4.1/1.0. parsons d, ryu h, cranshaw m (2007), a design requirements framework for mobile learning environments. the journal of computers. 2(4):. saipunidzam m, mohammad n, mohamad i, shakirah m (2008), open source implementation of m-learning for primary school in malaysia. international journal of human and social sciences 3(4):309–313. saipunidzam m, mohammad n, mohamad i, shakirah m (2010), m-learning: a new paradigm of learning mathematics in malaysia. international journal of computer science & information technology (ijcsit), 2(4):76-86. 26 z. al-khanjari, k. al-kindi, a. al-zidi and y. baghdadi sharples m (2000), the design of personal mobile technologies for lifelong learning. computers and education 34:177-193. ting yl (2005), mobile learning-current trend and future challenges. proceedings of the fifth ieee icalt’05. trucano m (2012), mobile learning and textbooks of the future, e-reading and edtech policies: trends in technology use in education in developing countries. world bank’s edutech blog, 4. vavoula g, sharples m, rudman p, meek j, lonsdale p (2009), myartspace: design and evaluation of support for learning with multimedia phones between classrooms and museums. computers & education, doi: 0.1016/j.compedu.2009.02.007. wu w, wu y, chen c, kao h, lin c, huang s (2012), review of trends from mobile learning studies: a meta-analysis. computers & education 59(2):817-827. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 1.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 01-14 arginine-amino acid interactions and implications to protein solubility and aggregation a. r. shaikh*a and d. shahb *a department of chemical sciences and engineering, kobe university, japan b petroleum and chemical engineering department, sultan qaboos university, oman received 23 february 2014; accepted 23 february 2015 abstract: arginine, useful in protein refolding, solubilization of proteins, and suppression of protein aggregation and non-specific adsorption during formulation and purification, is a ubiquitous additive in the biotechnology and pharmaceutical industries. in order to provide a framework for analyzing the molecular level mechanisms behind arginine/protein interactions in the above context, density functional theory was used to systematically examine how arginine interacts with naturally occurring amino acids. the results show that the most favorable interaction of arginine is with acidic amino acids and arises from charge interactions and hydrogen-bond interactions. arginine is also shown to form stacking and t-shaped structures with aromatic amino acids, the types of cation– and n–h… interactions, respectively, known to be important contributors to protein stability. the analysis also shows that arginine-arginine interactions lead to stable clusters, with the stability of the clusters arising from the stacking of the guanidinium part of arginine. the results show that the unique ability of arginine to form clusters with itself makes it an effective aggregation suppressant and support the interpretations of the current study using experimental and molecular dynamics results available in the literature. the results also contribute to understanding the role of arginine in increasing protein solubility, imparting thermal stability of important enzymes, and designing better additives. keywords: protein aggregation, protein solubility, arginine, quantum chemical calculation, amino acid. . . corresponding author’s e-mail: abdulrajjak.shaikh@people.kobe-u.ac.jp a. r. shaikh and d. shah 2 1. introduction amino acids are critical to life as they play a central role in biology and physiology, both as building blocks of proteins and as intermediaries or vehicles in various metabolisms. proteins catalyze almost all intracellular reactions and also control most of the cellular processes. they are active only in their respective, unique three-dimensional (3d) structures. interactions between amino acids, contained within the protein sequence, carry the necessary information to fold the proteins correctly and stabilize the active structure. for example, a point mutation in the amino acid sequence within a protein can substantially change the interactions and thus the structure and functionality of the protein molecules (leippert et al. 1997). therefore, there is considerable interest in a systematic examination of amino-acid/amino-acid interactions in order to understand the structural stability of proteins. the problem of protein aggregation is intimately tied to protein folding and the stability of the folded shape and is important not only in cell biology but also in pharmaceutical, recombinant proteins, and the food industry. during the past two decades, the advent of recombinantly expressed therapeutic proteins in the pharmaceutical industry has highlighted the importance of issues such as the chemical and physical stability of the resulting proteins and their stability during long-term storage (frokjaer and otzen 2005). proteins tend to degrade through aggregation with even a slight change in environmental conditions, including temperature, ph, buffers, and ion concentration. an increasing number of disorders, including alzheimer’s and parkinson’s disease, the spongiform encephalopathies, and type two diabetes, also are directly associated with the deposition of protein aggregates in tissues, including the brain, heart and spleen (cleland 1991; cellamare et al. 2008). a commonly used technique to prevent protein misfolding and aggregation is the use of lowmolecular-weight additives, termed “artificial molecular-chaperones”. artificial molecular chaperones have been developed to prevent aggregation in vitro and promote efficient refolding of denatured proteins via the addition of chemical compounds, including amino acids (arginine, proline, alanine, etc.), denaturants (typically guanidine and urea), poly(ethylene glycol), surfactants, and so on (chen et al. 2009; lee et al. 2006; ghosh et al. 2009; matsuoka et al. 2007; shiraki et al. 2002; takano et al. 2009). clearly, the intermolecular interactions between these stabilizers and the amino acids on the protein surface play a dominant role in stabilizing the proteins against aggregation and, therefore, deserve careful attention so that educated choices for stabilizers could be made for a given protein. one of the most commonly used additives for protein stabilization is arginine, which is effective in suppressing protein aggregation, increasing the refolding yield of proteins, and enhancing the solubility of aggregation-prone proteins as required for pharmaceutical applications (shiraki et al. 2002). apart from this, arginine also finds a number of applications in improving the yield of ion and affinity chromatography (ejima et al. 2005). arginine is also used to protect enzymes from thermal denaturation. an example is phytase, an enzyme which is added to animal feed and then subjected to thermal processing in which the feed is heated to 60–90 c. however, as phytase is a heatunstable enzyme which loses its activity during processing, arginine is added to the pre-feed formulation in order to prevent thermal denaturation. the addition of even as little as 1% arginine to the feed formulation has been shown to remarkably enhance the stability of enzyme (ryu and park 1998). although arginine is a frequent choice as an aggregation suppressor, refolding enhancer, and protein solubilizer, the exact stabilization mechanism of how arginine preserves enzyme stability is still unclear, although it is suspected that arginine binds to some regions of the protein via ionic interactions. a number of different hypotheses to explain the mechanism have been reported, including preferential interactions with the protein surface, increase of surface tension, and protein hydration (arakawa et al. 2007). different propositions recently have been advanced to explain the stabilizing effect of arginine. for example, a recent experimental study by das et al. (2007) suggests that arginine stabilizes proteins against aggregation by reducing the hydrophobic interactions between the proteins through hydrophobic association between the aliphatic chain of arginine and hydrophobic patches on the protein, whereas ghosh et al. (2009) suggest that strong interactions of arginine with tryptophan residues are the ones responsible for the stabilization. in general, the presently available studies do not explain or illustrate the mechanistic basis of arginine as a protein stabilizer adequately. it is well known that the amino acid sequence of a particular protein contains all the information needed to fold protein in a specific conformation, arginine-amino acid interactions and implications to protein solubility and aggregation 3 which ensures effective interaction within itself as well as with specific receptor sites required to trigger the biological response or functioning of the protein. interactions with the aqueous solvent, known as the hydrophobic effect, result in residues with non-polar side chains typically being buried in the interior of a protein. conversely, polar amino acid side chains tend to collect on the surface of a protein, where they are exposed to the aqueous medium. hence, it was assumed that when therapeutic protein is in its native conformation, interaction between the side chain of exposed amino acids and the arginine will play an important role in stabilizing protein from aggregation. considering this fact, only a truncated model of a side chain was used in the present calculations. such a truncated model has been used successfully to study amino acid side chain interactions. herewith, in order to develop a framework for subsequent analyses of the role of arginine in the above contexts (suppressing protein aggregation, preventing thermal denaturation, and enhancing protein solubility), the atomic level interactions between arginine and each amino acid are studied using density functional theory and a comparative picture of the interaction of arginine with each amino acid is presented. the present analysis sheds light on the different kinds of hydrogen bonds and the dominant interactions that exist between arginine and amino acids in a protein, some of which have been neglected in the past (for example, n–h… interactions) so that one can identify which of the many possible interactions are either favorable or unfavorable for a particular application. since, in general, favorable interactions lead to an increase in solubility of the protein and suppression of protein aggregation (shiraki et al. 2002), and unfavorable interactions have the reverse effect (timasheff 2002), the present study and similar studies for other additives can assist in designing protein-specific strategies to prevent the aggregation and misfolding of proteins, enhance protein solubility, and develop generally accepted guidelines and tests so that unforeseen adverse effects can be avoided by choosing additives most appropriate for a given protein. 2. computational methods all the calculations presented here were carried out using the dmol3 numerical density functional theory program (delley 1990; delley 2000) computed in materials studio 4.3 (accelrys inc., san diego, california, usa). the double numerical polarized (dnp) basis set, which includes all occupied atomic orbitals plus a second set of valence orbitals and polarized d-valence orbitals, was employed. geometrical optimizations were carried out using the local density approximation (lda) and the vosko-wilk-nusair (vwn) exchange correlation functional. optimized geometries were refined further by calculating single-point energy using the general gradient potential approximation (gga) and the becke-leeyang-parr (blyp) exchange correlation functional. charges and spin populations were calculated by the hirshfeld and mulliken population analysis. the effect of the aqueous environment was accounted approximately by computing singlepoint energies using the conductor-like screening model (cosmo) (klamt and schuurmann 1993; klamt 1995) at the gga-becke-lee-yang-parr (blyp) level. frequency calculations were performed to ensure that structures obtained were indeed minimum energy structures with zero imaginary frequencies. charge transfer and chemical bonds are, in most cases, well described within design for testing (dft) with either the local density approximation (lda) or the gga to account for the exchange correlation energy of the electrons. however, dft does not include london-dispersion forces (ortmann et al. 2005; chakarova-kack et al. 2006), which represent the dominant stabilization energy contribution required for stacking-type interactions. while it has recently become possible to include such effects in a non-empirical way in dft calculations for systems of practical interest (chakarova-kack et al. 2006), such calculations are still computationally very expensive. however, dispersion interactions are compensated in dmol3 by other terms in the interaction energy (andzelm et al. 2003; natsume et al. 2006). in ab initio calculations of interaction energies, the basis set superposition error (bsse) approach is used to account for incomplete atomic basis sets. the bsse contribution is expected to be small, as the dmol3 program uses numerical functions that are far more complete than the traditional gaussian functions (andzelm et al. 2003). the configurational binding energy eb between arginine and an amino acid was calculated by the following equation: a. r. shaikh and d. shah 4 [ :: ] [ ] [ ]be arg aa arg aa (1) where [arg] stands for the energy of the arginine molecule, [aa] is the energy of the amino acid, and [arg::aa] represents the energy of the arginineamino acid complex. the interaction energy includes the effect of zero-point vibrational energy (zpve), which was obtained from normal mode analysis. gibbs free energy at 298.15 k (room temperature) was calculated using the vibrational analysis from dmol3. the focus of the current study is on the interactions of the guanidinium group of arginine with the side-chains of amino acids with the arginine molecule truncated at the methyl guanidine part in order to focus particularly on the interactions of the guanidinium group of arginine with various protein residues and reduce the computation time and complexity arising from the flexibility of the side-chain of arginine. further, focus has been restricted to the interactions of the methyl guanidine part of arginine with the sidechains of different residues, as guanidine is known to not interact with the protein backbone (lim et al. 2009). similar truncated systems have been successfully used previously to demonstrate some of the interactions between amino acids (melo et al. 1999). different possible orientations of each amino acid relative to arginine were studied in order to identify the maximum binding energy, both at 0°k and at room temperature. because of the positive charge on arginine at neutral ph, the protonated form of arginine was used in the calculations; however, the neutral form was also considered wherever appropriate. 3. results and discussion section 3.1 begins with the results of the dft calculations for the binding energies for arginine’s interactions with itself and with the various amino acids. the strengths of binding and the most favorable orientations and structures are discussed. this is followed by brief discussions of other interactions (ie. beyond interaction of arginine with a single amino acid) such as interactions with more than one amino acid simultaneously in section 3.2. section 4 then focuses on a particular application, namely, implications of some observations of the role and mechanism of arginine as a protein stabilizer against aggregation, in protein solubility, and the thermal stability of enzymes. table 1. classification of amino acids based on their side chains. group classification amino acids 1 acidic and their amides asp, glu, asn, gln 2 ring containing trp, phe, tyr, pro 3 non-aromatic hydroxyl r-groups ser, thr 4 basic groups arg, lys, his 5 sulfur groups cys, met 6 aliphatic groups ala, val, leu, ile it is convenient to classify the amino acids into six different groups and calculate the binding energy of each amino acid with arginine. group 1 consists of amino acids with acidic and aminecontaining side chains (asp, glu, asn, and gln); group 2 contains amino acids with rings (phe, tyr, trp, and pro); group 3 is non-aromatic amino acids with hydroxyl r-groups (ser and thr); group 4 is basic amino acids (lys, arg, and his); group 5 is amino acids with sulfur-containing side chains (cys and met); and group 6 contains amino acids with aliphatic r-group side chains (gly, ala, val, leu, and ile) [table 1]. 3.1 interaction of arginine with single amino acids group 1: acidic amino acids and their amides the interaction of the side-chains of amino acids in proteins is a determining factor of the mechanisms behind a wide variety of biological phenomena such as antigen/antibody recognition and enzyme-substrate interactions. in particular, the interactions involving ionic groups of opposite charges are expected to be more dominant because of the electrostatic contributions. one special case of these interactions involves the guanidinium group of the arginine and carboxylate group of acidic amino acids—forming a guanidiniumarginine-amino acid interactions and implications to protein solubility and aggregation 5 carboxylate salt bridge. a detailed theoretical study of the interaction of arginine with acidic amino acids has been presented previously by melo et al. (1999), who showed that the zwitterionic form of amino acids is more stable in an aqueous environment while the neutral form is more stable in vacuo. a similar situation was observed in the present calculations as well as strong interactions of arginine with acidic amino acids in the aqueous environment. the calculations of the current study indicate that preferred conformations of the arginine/amino acid pair are those where it is possible to establish two hydrogen bonds of the type n–h…o c, where the two oxygen atoms of the carboxylate group of acidic amino acids share a partial negative charge and the two n–h groups of arginine carry a partial positive charge. there are two such interactions possible involving different nitrogen atoms of arginine. the results show a pronounced difference between the two conformations, with the hydrogen bonds of carboxylate with 1n and 2n being preferred over the hydrogen bonds with n and 2n of the arginine. the difference between the binding energies of the two conformations is 26.21 kcal/mol. interactions between arginine and amino acids with amide side chains (asn and gln) are weaker than those with acidic amino acids (figures 1a and b). the difference in binding energies between acidic and amino-side-chain-containing amino acids was found to be ~10 kcal/mol [table 2], since only one hydrogen bond is possible in the latter case. in summary, arginine has a stronger interaction with acidic amino acids than with amido-side-chain-containing amino acids. the results shown above are based on calculations done at ground state (0 °k). additional calculations were performed to include the effects of entropy and temperature (298.15°k). the gibbs free energies of binding (gb298.15k) of arginine with the various amino acids are also presented in table 2, which shows that even at 298.15°k, the interactions between arginine and acidic amino acids dominate those with other amino acids. group 2: amino acid-containing rings arginine forms parallel stacking interactions with aromatic amino acids (flocco and mowbray 1994). these types of aromatic-arginine interactions, which are often found in locations critical to the activities of proteins, apparently serve to orient the arginine side chain without interfering with its ability to form hydrogen bond elsewhere. longrange interactions between arginine and aromatic residues have been reported by martis et al. (2008); here, the cation from arginine interacts with the cloud of the aromatic rings, forming a stacking cation interaction. in this type of interaction, the side chains of aromatic amino acids provide a surface of negative electrostatic potential that can bind to a wide range of cations through predominantly electrostatic interactions. martis et al. (2008) performed detailed analyses on a number of different proteins to show that 31% of the total trp residues within the protein structure participate in such interactions with arginine. these types of cation– interactions are more hydrophobic in nature than electrostatic. martis et al. (2008) also observe that, amongst all the aromatic residues, the average energy of arg-trp interactions is the highest; however, phe is shown to have the maximum number of such interactions. recently, ottiger et al. (2009) also studied hydrogen bonding interactions between amino and aromatic moieties. they suggested that the formation of such a t-shape interaction is due to the hydrogen bonding of a n–h… nature. these interactions are amongst the weak intermolecular interactions found in proteins and play an important role in biological systems. although flocco and mowbray (1994) suggest that parallel stacking interactions between arginine and aromatic residues are commonly found in protein structures and have high binding energies, the dft calculations of the current study showed that t-shape interactions are more favored than the parallel stacking interactions by approximately 1.0 kcal/mol. this indicates that there is a competition between the parallel stacking and n–h… hydrogen bond (t-shape) interactions (figure 1e and f). the highest eb of arginine with trp was observed in group 2 amino acids [table 2] because the indole group of trp provides a larger negative electrostatic potential than benzene or phenol, thus making trp constitute a more attractive cation-binding site. proline is a small molecule with a non-aromatic ring, and has therefore been included in the group 2 classification. the binding energy of arginine with proline is –19.67 kcal/mol, higher than the one corresponding to tyrosine (–18.75 kcal/mol). the binding energy between arg and pro falls in between those of parallel stacking and t-shape interactions due to the flexible nature of the ring [fig. 1h]. a. r. shaikh and d. shah 6 in summary, the binding energy of arginine with respect to amino acids with rings is in the following order [table 2]: trp > pro > phe > tyr ito et al. (2011), based on x-ray studies, reported that lysozyme mostly interacts with arginine. shah et al. (2011; 2012) reported similar results using experimental and molecular dynamics studies. group 3: amino acids with non-aromatic hydroxyl rgroups the hydroxyl side chain of serine/threonine in a protein can establish additional intramolecular hydrogen bonds; in particular, it can act as a proton donor as well as a proton acceptor. calculations from the current study show that the binding energy for arginine with serine is –22.70 kcal/mol and for threonine it is –21.11 kcal/mol. strong binding energies were observed between arginine table 2. binding energy (eb at 0 k) and gibbs free energy of binding (gb at 298.15 k) of each amino acid with arginine calculated at gga-pw91 level. all the energies are in kcal/mol.a group amino acid eb 0k gb298.15k 1 asp –35.52 –20.84 glu –34.72 –21.03 asn –23.81 –12.51 gln –24.39 –13.69 2 trp –17.82 –4.17 tyr –17.50 –4.18 phe –17.34 –4.50 pro –18.01 –4.95 3 ser –19.83 –10.03 thr –18.61 –5.28 4 arg –30.57 –20.84 lys –12.54 –1.23 his –23.11 –10.64 5 cys –23.42 –10.10 met –18.67 –5.14 6 ala –15.05 –6.13 val –13.49 –0.69 leu –14.31 –1.36 ile –12.92 –0.04 note: a both the binding energy and the gibbs free energy reported here are inclusive of zpve correction. arginine-amino acid interactions and implications to protein solubility and aggregation 7 figure 1. optimized geometries of arginine-amino acid complexes. only representative complexes are shown. and ser/thr in a conformation where the hydrogen in the hydroxyl group of ser/thr points away from the arginine guanidinium group, hence making an n–h…o type of hydrogen bond [fig. 1d]. thus, electrostatic interactions and hydrogen bonding play a dominant role in the stabilization of such complexes. also were calculated interactions of these amino acids with the neutral form of arginine; however, the binding energies of the resulting complexes are smaller (–9.20 kcal/mol for serine, – 4.90 kcal/mol for threonine) than the ones corresponding to the cationic form of arginine. hence, ser/thr also demonstrates a favorable interaction with arginine through n-h…o electrostatic interactions. group 4: amino acids with basic groups although it is intuitive that arginine would have repulsive interactions with basic amino acids because both arginine and the basic amino acids have similar charges, a strong binding was observed in the form of parallel stacking alignment due to van der waals interactions. we begin with the interaction of arginine with itself, which has a very high energy of –35.15 kcal/mol [table 2, fig. 1i]. such stacking interactions are often found in protein structures. also evaluated was the binding energy for the complex with the two arginine molecules placed face-to-face. such selforganization is, however, not favored as it leads to charge repulsions (+3.90 kcal/mol). the formation of arginine clusters was studied using quantum calculations and experiments in the gas phase (julian et al. 2001), where it was suggested that cyclic arginine trimers possess exceptional stability. however, in the solvent environment, it was observed that stack interactions also possess good stability. as shown in table 2, arginine-arginine interactions remain prominent even at room temperature. similarly, the parallel stacking interactions for lysine with arginine are also very stable. the binding energy of neutral histidine with arginine is high (–25.89 kcal/mol) due to the presence of a n– h(arg)…n(his) interaction in the complex (fig. 1g). the results presented in this section highlight the unique tendency of arginine to form stable clusters in the form of stacks with other arginine molecules. in addition, the results show that arginine also interacts with lysine and histidine through such stacking interactions. group 5: amino acids with sulfur-containing groups cysteine plays a crucial role in determining the structure as well as functions of many proteins and has a high reactivity due to the presence of the thiol group. the disulfide bond formed between two cysteines has an important role in protein folding. sulfur-containing functional groups of cysteine and methionine are normally considered hydrophobic moieties or weak hydrogen bond acceptors in a folded protein structure. gregoret et al. (1991) studied the hydrogen bonds involving sulfur atoms in proteins and found the interaction between arginine and sulfur-containing amino acids to be the n–h…s type. detailed analyses of these types of s…x interactions (where x = n or o) are a. r. shaikh and d. shah 8 presented by iwaoka et al. (2002). these interactions contribute greatly to the stability of proteins. the binding energy for cysteine with arginine is –24.88 kcal/mol, while for methionine it is –20.43 kcal/mol. these interactions are similar in nature to those of serine and threonine. here the hydrogen (– sh) of cys and the methyl group (–sch3) of met point away from the arginine amino group, while the n–h of arginine forms a hydrogen bond with the sulfur [fig. 1c]. this hydrogen bond could be due to the n(s) *(n–h) orbital interaction in which the n atom tends to approach the lone pair of the s atom. such a hydrogen bond is usually weaker than the n–h…o type hydrogen bond. the lone n–h pair type interaction could also be hydrophobic in nature since, although sulfur is quite polarizable, it is less electronegative. group 6: amino acids with aliphatic groups aliphatic interactions, also termed hydrophobic interactions, play a major role in protein folding and protein aggregation processes. most of the hydrophobic groups are believed to be exposed to the solvent environment once a protein denatures. therefore, as arginine is used as a protein stabilizer, the interactions between arginine and the aliphatic amino acids become very important. although the aliphatic groups do not contain any functional group and thus lack any specific interactions with arginine, favorable interactions of these amino acids with arginine because of induced-dipole interactions were observed [fig. 2]. the binding energy varies from -15.62 kcal/mol to -18.19 kcal/mol, with the energy of interaction for alanine being the highest and the one for isoleucine being the lowest in group 6. it is believed that the reason for the observed trend is the larger size of isoleucine, in which the induced dipole gets distributed throughout the molecule. it is notable that the energy for glycine was not computed as it does not have any side chain. figure 2 shows the induced dipole for the case of alanine. although the binding energy of the guanidinium group of arginine and aliphatic amino acids is low, it should be noted that the hydrophobic side chain of arginine interacts favorably with aliphatic amino acids through the van der waals interaction. figure 2. optimized geometries for a) arg-ala complex, b) arg, and c) ala. the hirshfeld charges are also shown. the presence of arginine induces a dipole in alanine. 3.2 additional special interactions of different amino acids with arginine in addition to the interactions of arginine with single amino acids, it was observed that arginine can act as a strong link between two amino acids since the side chain of arginine (ie. the guanidinium part) has two n atoms that can take part in hydrogen bonding. selective combinations of arginine with glu, asp, trp, phe, and tyr were examined and it was found that arginine can form two strong salt bridges with two acidic amino acids simultaneously. binding energies for these interactions are -46.33 kcal/mol [fig. 3a]. in addition, the binding energies for trp-arg-trp, tyr-arg-tyr, and phe-arg-phe complexes were arginine-amino acid interactions and implications to protein solubility and aggregation 9 computed in parallel stacking arrangements [fig. 3b]. these energies are approximately -10.0 kcal/mol. the t-shape geometry (-30.30 kcal/mol) is more favored than the stacking interactions in the type of bridging interactions above. hence, the stacking interactions of arg with aromatic residues are weaker than the salt bridges with the acidic amino acids. these types of interactions can play a key role in protein aggregation. next the formation of arginine clusters was probed. as reported in the literature, arginine displays a strong stacking interaction with itself. for example, the dft studies of julian et al. (2001) suggest that arginine forms cyclic trimer clusters in the gas phase. in these clusters, protonated guanidinium of an arginine interacts with the carboxylate group of another arginine to form stable non covalent complexes in the form of cyclic trimer coordinated to either a cation or an anion. further, the experimental studies by das et al. (2007) confirmed the formation of arginine clusters and reported that, in a solution environment, hydrophobic interactions (from the aliphatic chain of arginine) play a dominant role in the formation of such clusters. although hydrophobic interactions arising from the aliphatic side chain of arginine may contribute to the formation of clusters, the current study focuses on the interaction involving the guanidinium group, since, as discussed further below, it is the guanidinium group that is critical to the formation of clusters. arginine can also form different types of clusters (eg. julian et al. 2001), but the focus of the current study is only on the clusters in the form of stacks, as these clusters would be more hydrophobic in nature and might be useful to mask exposed hydrophobic residues on the protein surface. figure 3. binding energy (eb) and relative (to a stack of 2 arginine molecule) lumo–homo (l–h) gap in the arginine stack. all the energies are in kcal/mol. a. r. shaikh and d. shah 10 table 3. binding energy (eb) and relative (to a stack of 2 arginine molecule) lumo-hom) (l-h) gap in the arginine stack. all the energies and in kcal/mol. no. of arg in stack eb relative l–h gap 2 –35.15 0.00 3 –49.11 –0.88 4 –63.99 –1.32 5 –75.82 –1.25 the stacking interactions between the guanidinium groups of arginine help in stabilizing the arginine-arginine complex. in order to quantify the optimum arginine pairing in the stacking interactions, dft calculations for up to five arginine molecules in a stack were carried out; the binding energy increased linearly with the number of arginine molecules and the homo-lumo gap of the complex attains a minimum for a stack of five arg [table 3; fig. 3c]. das et al.14 suggested that typically 8–10 molecules are present in an arginine cluster. these clusters can be formed by stacking or parallel or both. calculations in the current study show that clusters containing up to seven arginine molecules are stable based on the binding energy even at 298.15°k [table 4]. table 4. comparison of binding energies between various complexes at 0 k and room temperature (298.15 k). complex eb (0 k) gb (298.15 k) asp-arg-asp trp-arg-trp (stack) trp-arg-trp (t-shape) his-arg-his arg molecules in stack 2 3 4 5 6 7 4. implications to arginine’s role as a stabilizer or a solubilizer here the implications and use of the above results in the context of a few typical uses of arginine are discussed. 4.1 arginine as an aggregation suppressant the results reported above shed light on the mechanism behind arginine’s role as an effective suppressant of protein aggregation as calculations clearly suggest that the guanidinium group of arginine interacts with a variety of amino acids. in particular, it forms strong salt bridges with the acidic amino acids and strong stable clusters with itself and can also interact favorably with the aromatic amino acids. protein structures contain mostly hydrophilic amino acids on the surface, which include group 1 and group 4 amino acids, according to the classifications of the current study. the current calculations show that arginine binds strongly with these residues. the aliphatic, methylene groups in the side chains of arginine molecules bound to group 1 and group 4 amino acids can easily mask any exposed hydrophobic patches on the proteins, with the polar end-caps on the side chains providing a solvent-friendly shield. in addition, molecular clusters of arginine can also shield the hydrophobic patches on a protein, thereby further reducing the tendency of the proteins to aggregate. the importance of the arginine clusters for stability was also highlighted by das et al. (2007) based on their light scattering and chromatographic analyses of clustering in arginine-amino acid interactions and implications to protein solubility and aggregation 11 arginine solutions and the effect of arginine on heat-induced aggregation of alzheimer’s beta amyloid a 1-42 protein. das et al. (2007) suggested that the arginine clusters are stabilized through hydrophobic interactions among the methylene groups. however, while the current researchers’ dft studies also predict clustering of arginine and, in fact, the guanidinium group of arginine contributes to the stability of the resulting clusters. the role of the guanidinium group in the formation of such clusters becomes more evident when one compares the behavior of arginine against that of lysine in solution. lysine, although quite similar to arginine in structure, lacks the guanidinium group (lysine has an amino group at the end of its side chain), does not form stable clusters, indicating that a similar side chain alone is insufficient to promote the formation of clusters. it is worth noting that the above conclusion concerning the role of the guanidinium group is also consistent with the recent molecular dynamics simulations of vondrasek et al. (2009) on short dior decaarginine peptides. in particular, vondrasek et al.’s (2009) results show significant pairing of the guanidinium groups, which is further confirmed by independent ab initio calculations. the results of the current study indicate that the planar geometry of guanidinium promotes its alignment with the guanidinium group of another arginine molecule and causes the formation of arginine clusters in a solution. one might expect, however, that the hydrophobic interactions among the methylene groups might further stabilize the stacks as they are formed, but such interactions alone are insufficient by themselves to sustain clustering. apart from stacking clusters, as noted in the above calculations, other possibilities of arginine clustering do exist. for example, carboxylate part of an arginine molecule can interact with the guanidinium group of another forming a cluster, but the current researchers believe that only clusters that could create enhanced hydrophobicity (arising from the methylene group of arginine) contribute to protein stability. in summary, it is proposed that the interactions discussed here contribute to the stability of the proteins in a number of ways. first, the aliphatic side-chains of arginine clusters can mask the exposed hydrophobic patches on the proteins. in addition, the hydrogen bonding interactions between arginine and the hydrophilic residues of a protein, especially the interactions with acidic amino acids, promote the binding of arginine to the proteins (through the guanidinium group), and the aliphatic chains of such bound arginine molecules can also shield any neighboring hydrophobic residues on the protein. 4.2 thermal stabilization of proteins as noted in section 1, arginine is also used for the thermal stabilization of proteins. the binding energies for arginine-amino acid interactions reported above in section 3 shed some light on the potential interactions at play in the thermal stabilization of phytase (ryu and park 1998). for instance, the gibbs free energy for interaction between arg-asp is about 17.6 kcal/mol even at 90 c when the dft results are adjusted for the effect of temperature (a reduction of about 20% from the magnitude at room temperature), whereas the corresponding magnitudes of gibbs energies for the interactions of arginine with the other amino acids diminish even more, with some becoming almost zero or even positive. hence, the ionic interactions between arginine and acidic amino acids might be responsible for inducing stability to these enzymes at high temperatures. the interaction of arginine with various amino acids [table 2] also identifies the factors that can enhance the stability of proteins in extreme conditions. this is of particular interest because it raises the possibility of engineering enzymes with enhanced high-temperature stability and catalytic efficiency for industrial applications. one of the approaches to achieve this goal would be to use site-directed mutagenesis to replace surface hydrophobic residues with acidic amino residues. 4.3 arginine as a solubilizer arginine is also widely used to increase the solubility of various proteins, although how arginine enhances the solubility is not known currently. favorable interactions of amino acid residues with an additive will results in enhancement of the solubility and vice-versa. the current study’s dft calculations adjusted to room temperature suggest that arginine has fairly strong interactions with all amino acids except valine and isoleucine [table 2]. this is in agreement with the experiments carried out by arakawa et al. (2007) on the solubility of various amino acids in the presence of arginine. arakawa et al. observed that ile and val, are indeed less soluble in the presence of arginine, while other amino acids, like tyr and trp, a. r. shaikh and d. shah 12 are more soluble. although an absolute quantitative correlation between the binding energies was not observed in the present study, the solubility enhancement of amino acids as reported by arakawa et al. was calculated; although similar interaction energies of arginine with try, trp, and phe were observed, the solubility of trp in an arginine solution is higher than that of phe. the theoretical framework used here can provide an estimate of the solubilizing efficiency of an additive. the above discussions illustrate that the dft study presented here not only provides information about important interactions which occur within a protein molecules (ie. between the different amino acid side chains and arginine) but also sheds light on the crucial role played by arginine in the suppression of protein aggregation, in increasing the solubility of proteins and thermal stability of important enzymes, and in separation techniques. 5. conclusions in summary, the researchers carried out detailed first-principles density functional theory calculations of the structures and the binding energies of the 20 different amino acids with arginine. the theoretical calculations undertaken show that arginine has the strongest interactions with acidic amino acids, followed by interactions with itself. detailed explanations of the types of interactions and conformation have been discussed above. the binding energies of arginine with all the other amino acids fall in a close range, from -15.62–22.49 kcal/mol (at a ground state) below the binding energies with acidic amino acids. various interactions such as n–h… , a hydrogen bond, van der waals, induced-dipole, etc. have also been identified as the cause of the stability of the interactions of arginine with the other amino acids. most of these interactions occur within the protein structure, and the comparative study presented provides a framework for identifying the dominant interactions that help in the stabilization of the protein structure. as noted in the introductory section of the paper, arginine is a popular choice as an additive for effective suppression of protein aggregation and protein solubilization; however, the mechanistic basis behind the stabilization action of arginine is still not fully understood. the calculations in the current study provide a basis for understanding the role of arginine in this respect. arginine favorably interacts with the side chains of many amino acids. the results also reveal the strong interaction of arginine with the acidic amino acids through the formation of salt bridges. acidic amino acids are generally present on a protein surface because of their hydrophilic nature; thus, they are easily accessible to the arginine molecules to interact with. these interactions are believed to be major contributors to increase the solubility of proteins. however, as discussed earlier, which of the numerous other interactions detailed in the paper are also likely to influence stabilization/solubilization depends on the distribution of surface residues and the relative magnitudes of the binding energies in their interactions with arginine. although a variety of factors influence protein stability and solubility, charge interactions between arginine and acidic amino acids appear to play a major role. to the best of the authors’ knowledge, the present study is the first to examine the interaction between arginine with all other amino acids. in addition, the parallel stacking of arginine with itself as clusters in a solution environment was also analyzed, which would further help in elucidating the functioning mechanism of this unique additive. finally, it is noteworthy that the approach outlined here can be used for additives other than arginine to analyze systematically how those additives interact with proteins and to design better additives and mixtures of additives. references andzelm j, govind n, fitzgerald g, maiti a (2003), dft study of methanol conversion to hydrocarbons in a zeolite catalyst. international journal of quantum chemistry 91(3): 467-473. arakawa t, ejima d, tsumoto k, obeyama n, tanaka y, kita y, timasheff sn (2007), suppression of protein interactions by arginine: a proposed mechanism of the arginine effects. biophysical chemistry 127(1-2): 1-8. cellamare s, stefanachi a, stolfa da, basile t, catto m, campagna f, sotelo e, acquafredda p, arginine-amino acid interactions and implications to protein solubility and aggregation 11 carotti a (2008), design, synthesis, and biological evaluation of glycine-based molecular tongs as inhibitors of aβ1–40 aggregation in vitro. bioorganic and medicinal chemistry 16: 48104822. chakarova-kack sd, schroder e, lundqvist bi, langreth dc (2006), application of van der waals density functional to an extended system: adsorption of benzene and naphthalene on graphite. physical review letter 96(14): 46107-(14). chen d, martin zs, soto c, schein ch (2009), computational selection of inhibitors of abeta aggregation and neuronal toxicity. bioorganic and medicinal chemistry 17: 5189-5197. cleland jl (1991), mechanisms protein aggregation and refolding. phd thesis, massachusetts institute of technology, cambridge. das u, hariprasad g, ethayathulla as, manral p, das tk, pasha s, mann a, ganguli m, verma ak, bhat r, chandrayan sk, ahmed s, sharma s, kaur p, singh tp, srinivasan a (2007), inhibition of protein aggregation: supramolecular assemblies of arginine hold the key. plos one 2(11): e1176. delley b (1990), an all-electron numerical method for solving the local density functional for polyatomic molecules. journal of chemical physics 92(1): 508-517. delley b (2000), from molecules to solids with the dmol3 approach. journal of chemical physics 113(18): 7756-7764. ejima d, yumioka r, arakawa t, tsumoto k (2005), arginine as an effective additive in gel permeation chromatography. journal of chromatography a 1094(1-2): 49-55. flocco mm, mowbray sl (1994), planar stacking interactions of arginine and aromatic side chains in proteins. journal of molecular biology 235: 709717. frokjaer s, otzen de (2005), protein drug stability: a formulation challenge. nature review drug discovery 4: 298-306. ghosh r, sharma s, chattopadhyay, k (2009), effect of arginine on protein aggregation studied by fluorescence correlation spectroscopy and other biophysical methods. biochemistry 48(5): 1135-1143. gregoret lm, rader sd, fletterick rj, cohen fe (1991), hydrogen bonds involving sulfur atoms in proteins. proteins structure function genetics 9: 99-107. iwaoka m, takemoto s, okada m, tomoda s (2002), weak nonbonded s•••x (x = o, n, and s) interactions in proteins. statistical and theoretical studies. bulletin of chemical society of japan 75: 1611-1625. ito l, shiraki k, matsuura t, okumura m, hasegawa k, baba s, yamaguchi h, kumasaka t (2011), high-resolution x-ray analysis reveals binding of arginine to aromatic residues of lysozyme surface: implication of suppression of protein aggregation by arginine. protein engineering design and selection 24(3): 269-74. julian rr, hodyss r, beauchamp jl (2001), salt bridge stabilization of charged zwitterionic arginine aggregates in the gas phase. journal of american chemical society 123: 3577-3583. klamt a (1995), conductor-like screening model for real solvents: a new approach to the quantitative calculation of solvation phenomena. journal of physical chemistry 99(7): 2224-2235. klamt a, schuurmann g (1993), cosmo: a new approach to dielectric screening in solvents with explicit expressions for the screening energy and its gradient. journal of chemical society perkin transaction 2(5): 799-805. lee sh, carpenter jf, chang bs, randolph tw, kim ys (2006), effects of solutes on solubilization and refolding of proteins from inclusion bodies with high hydrostatic pressure. protein science 15: 304-313. leippert m, beck e, weiland f, pfaff e (1997), point mutations within the betag-betah loop of footand-mouth disease virus o1k affect virus attachment to target cells. journal of virology 71: 1046-1051. lim wk, rosgen j, englander sw (2009), urea, but not guanidinium, destabilizes proteins by forming hydrogen bonds to the peptide group. proceedings of natural academy of sciences usa 106(8): 2595-2600. martis rl, singh sk, gromiha mm, santhosh c (2008), role of cation-pi interactions in single chain 'all-alpha' proteins. journal of theoretical biology 250: 655-662. matsuoka t, tomita s, hamada h, shiraki k (2007), amidated amino acids are prominent additives for preventing heat-induced aggregation of lysozyme. journal of bioscience and bioengineering 103(5): 440-443. melo a, ramos mj, floriano wb, gomes janf, leão jfr, magalhães al, maigret b, nascimento 13 a. r. shaikh and d. shah 12 mc, reuter n (1999), theoretical study of arginine-carboxylate interactions. journal of molecular structure theochem 463(1-2): 81-90. natsume t, ishikawa y, dedachi k, tsukamoto t, kurita n (2006), dft study of the electronic properties of dna–dna and pna–dna double strands. international journal of quantum chemistry 106(15): 3278-3287. ortmann f, schmidt wg, bechstedt f (2005), attracted by long-range electron correlation: adenine on graphite. physical review letter 95(18): 186101-(1-4). ottiger p, pfaffen c, leist r, leutwyler s, bachorz ra, klopper w (2009), strong n-h...pi hydrogen bonding in amide-benzene interactions. journal of physical chemistry b 113: 2937-2943. ryu s, park tg (1998), thermal stabilization of aspergillus phytase by l-arginine. biotechnology bioprocess engineering 3(1): 32-34. shah d, shaikh ar, peng x, rajagopalan r (2011), effects of arginine of heat-induced aggregation of concentrated protein solutions. biotechnology progress 27(2): 513-520. shah d, li j, shaikh ar, rajagopalan r (2012), arginine–aromatic interactions and their effects on arginine-induced solubilization of aromatic solutes and suppression of protein aggregation. biotechnology progress 28(1): 223-231. shiraki k, kudou m, fujiwara s, imanaka t, takagi m (2002), biophysical effect of amino acids on the prevention of protein aggregation. the journal of biochemistry 132: 591-595. takano k, higashi r, okada j, mukaiyama a, tadokoro t, koga y, kanaya s (2009), proline effect on the thermostability and slow unfolding of a hyperthermophilic protein. the journal of biochemistry 145(1): 79-85. timasheff sn (2002), protein hydration, thermodynamic binding, and preferential hydration. biochemistry 41(46): 13473-13482. vondrasek j, mason pe, heyda j, collins kd, jungwirth p (2009), the molecular origin of likecharge arginine-arginine pairing in water. journal of physical chemistry b 113: 9041-9045. 14 d:\desktop\paper 8.xps tjer vol. 11, no. 1, 69-76 ________________________________________ *corresponding author’s e-mail: albusaid@squ.edu.om complementing digital logic design with logisim sa al-busaidi department of electrical & computer engineering, college of engineering, sultan qaboos university, po box 33, pc 123, al-khoud, muscat, sultanate of oman received 30 april 2013; accepted 23 march 2014 abstract: classified as free and open source software (foss), logisim is a delightful tool that can easily be used to reinforce a solid understanding of the theoretical concepts related to a digital logic design course. unlike logicworks, one of the most attractive features of logisim is its ability to include user built libraries. this can result in the development of a library that models the complete set of integrated circuits (ics) required for a digital logic design course. as a consequence, numerous merits can be observed regarding a student's learning level within such a course. keywords: data sheet, digital logic design, free and open source software (foss), integrated circuit (ic), logicworks, logisim. (foss), (ics) (foss)circuits .(ic) integrated 70 sa al-busaidi 1. introduction a digital logic design course was developed at sultan qaboos university (squ) to introduce students to the fundamental theory and basic design blocks of digital circuits. students of theoretical fundamentals delve into boolean algebra, a subarea of algebra associated with digital circuits. the fundamental gates, namely the and, or, not, exclusive-or gates (xor), nor and a combination of and and not gates (or nand gates), are a direct product of the use of boolean algebra. these gates constitute the basic units for all digital logic design. to complement both the theoretical background and the design of digital circuits, the course includes experiments that progress at an equivalent pace with the course lectures. each experiment can be divided into two components: a pre-lab that is associated with a software simulation tool and the actual physical implementation of the experiment. the significance of the pre-lab is that it obligates students to practice the theory through simulation prior to conducting the experiment. on this basis, students not only conduct their experiments in a shorter time frame but also save resources through minimizing incorrect connections that could lead to the destruction of an integrated circuit (ic). accordingly, software packages that can be utilized for simulation purposes are a core requirement for the successful delivery of digital logic design courses (ahmad et al. 2013(a); ahmad et al. 2013(b); ahmed and ruelens 2013; ahmed et al. 2012;ahmed and al-abri 2012; ahmad and al-abri 2010; allawati and ahmad 2003 and al-busaidy 2013). this paper aims to share an experience using the free and open source software (foss) tool logisim for a digital logic design course. in section 2, a detailed account of the topics covered in the course is given. section 3 outlines the lab experiments that are assigned to complement the course's theoretical content. this is then followed in section 4 by a brief history and the current deployment of logisim. section 5 outlines the method by which logisim was adopted within this course. in sub-section 5(a), an account of the basic capabilities of logisim is given while in subsection 5(b) the method of extending logisim's library is outlined. the results of adopting this tool within the course are then given in section 6, followed by a conclusion in section 7. 2. digital logic design course this section recounts in detail the topics that are covered within the digital logic design course at squ. the initial part of the course touches upon the concepts of signed and unsigned decimal number representation in binary, binary ones complement and binary twos complement. the conversion process of numbers is then expanded to include the conversion to and from the octal and hexadecimal bases. on the other hand, the representation of the alphanumeric keys associated with the keyboard is demonstrated through the american standard code of information interchange (ascii) codes. the concept of including the parity bit for the purpose of detecting errors linked to ascii code transmissions is also introduced. the course also includes arithmetic operations on pairs of numbers that are represented in binary. of particular interest is the key concept of performing the binary subtraction operation using the binary twos complement. the outcome of this method demonstrates the underlying ingenuity of utilizing a single arithmetic operation, addition, to conduct two different operations, addition and subtraction. by representing numbers in the binary twos complement, the subtraction operation can be converted into the addition operation. justifiably, in this case only an adder is required to manipulate the pair of binary numbers. the next portion of the course introduces the concept of boolean algebra along with its associated theories and propositions. through boolean algebra, a complete set of basic functions for two variables is derived by means of a truth table. among the functions are the well known and, or, not, exclusively-or (xor), nand, nor and exclusively-nor (xnor) functions. these functions, along with their schematic gate representation, are clearly shown at this stage. this is followed by introducing the concept of transforming and, or and not gates into either a nand or nor-only equivalent. as a consequence of this equivalent representation, both gates are coined as universal gates. the combination of different sets of the basic functions permits for the emergence of new and more complex functions. these new functions can be made to reflect upon the ability of solving specific real life problems in light of digital design. in deriving these functions, the truth table that relates both input variables to the output function is essential. the function can be read directly from the truth table in either the sum of standard product form or the product of standard sum form (tocci et al. 2010). a function in this form requires only two gate levels, which is essential if a gate propagation delay is critical to the design. on the other hand, by mapping every possible output of the functions from the truth table, the concept of the karnaugh (k)-map is introduced as a methodology to 71 complementing digital logic design with logisim derive the minimized number of literal equivalent functions. although this could be alternatively achieved through boolean algebra manipulations, that may be a long and exhaustive procedure; however, the k-map is a graphics tool that is both intuitive and simple. the outcomeminimized number of literal function that is derived can be represented in either the sum-of-product form or the product-of-sum form (tocci et al. 2010). a function derived using this method not only requires only two gate levels but also requires a minimum number of gates. this translates into the reduction of both propagation delay and the required power consumption of the design. furthermore, at this stage of the course, functions derived from the combination of and, nor and not gates are converted into nor and nand equivalents. two methods can be used for the conversions: boolean algebra or a graphics method. after establishing these basics, the course then focuses on numerous designs of digital circuits that perform useful functions. in effect, at this stage the focus is on application of digital circuits. for such reasons, it is essential to differentiate between two types of digital circuits-namely combinational circuits and sequential circuits. whereby the input signals in combinational circuits flow in only one direction and lack memory devices, the defining attributes of sequential circuits are the signal feedback and memory devices. from a different perspective, sequential circuits can be viewed as a combinational circuit with the two aforementioned attributes. as a consequence, the course delves primarily into combinational circuits before introducing the concepts of memory and sequential circuits. a number of functional circuits are introduced in the combinational design portion of the course. this includes the half and full adder and subtractors, two bit multiplier, decoder, encoder, multiplexer (mux) and comparator circuits. it is worthwhile noting at this stage that the students experience a transition from theory to design. as it had been earlier stated, a subtractor can be replaced by an adder through converting the binary number into its twos complement form. henceforth, at this stage, it is clearly shown how this can be achieved using the full adder circuit with an external xor gate to perform both addition and subtraction operations. on the other hand, the concept of the latch and flipflop (ff) along with its different types are introduced in the sequential circuit portion. this includes three ff types-namely the data-flip-flop (d-ff), jack kilby flip-flop (jk-ff) and toggle flip-flop (t-ff). unlike its combinational circuit counterpart, the analysis of sequential circuits requires state tables instead of truth tables. state tables comprise the current state of the memory devices along with the input variable from one end, and the next state along with the output variable from the other end. in relating the next state to the current state it becomes necessary to determine the input function of each ff used within the design. interestingly, the analysis of sequential circuits can, therefore, be conducted in two stages. first, a combination stage, in which all memory elements are virtually removed from the design and the concern, is in obtaining the input ff equations for each memory element. second, in the next state stage, analysis is conducted only on every individual memory element; hence the complete circuit is virtually removed with only the memory devices remaining. the analysis of sequential circuits finally leads to a graphic representation of the state table in the form of a state diagram. the course then focuses on the reverse process, or the design process. initially, a state table is derived from a specific state diagram. the state table is derived taking into consideration a specific configuration of memory elements. the choice of memory elements will dictate the design of each ff input equation according to the ff characteristic table. from the state table, the design can subsequently be obtained. after this theoretical background, the design of sequential circuits in the form of registers and counters are demonstrated. two designs of registers in particular are studied-namely the parallel load parallel shift register and the serial load serial shift register. for counters, two designs of counters are studied: the asynchronous ripple counter and synchronous up/down counter. finally, the course terminates by briefly introducing different programmable devices such as random access memory (ram), programmable logic device (pld), programmable logic array (pla) and programmable array logic (pal). at this stage, the design methodology of deriving a read only memory (rom) device using a decoder and implementing a function using a pla are both demonstrated. 3. lab experiments this section highlights the five lab experiments conducted at the same pace as the theoretical aspect of the course. in the first experiment, the newcomer is introduced to the experimentation work area, the breadboard, and the various types of ics within the lab environment. this can be viewed as no more than a mere familiarization exercise. experiment 2 includes two parts. the aim of the first part is to convert a logical expression into a working digital circuit. as for the second part, the aim is to demonstrate how boolean algebra can be utilized to express functions in equivalent forms. the focus in experiments 3 and 4 is on designing combinational circuits. experiment 3 is used to design 72 sa al-busaidi half and full adders, while experiment 4 demonstrates how to successfully include decoders and multiplexers in a design. at this stage, the k-map is used to simplify the function to be implemented in the experiment. finally, the focus in experiment 5 is on the design of sequential circuits. in previous years, the experiment involved designing and implementing a 2-bit up/down counter using the d-ff. this year, as a simple upgrade, the requirement was to design two sequential circuits that could count the number of ones in an 8-bit sequence. for each design, the main constraint factor was the types of ics that had to be included within each design. the first design required the use of a single parallel load serial shift register, a full adder and four d-ffs, while the second design required a single parallel load serial shift register and a 4-bit synchronous up/down counter. from the list of experiments, it is clear that students are exposed to a plethora of ics through this course. table 1 gives an extensive list of the ics related to both the course content and experiments. 4. a brief history and current deployment of logisim logisim was introduced in 2000 by carl burch who was, at the time, based in the department of computer science at the college of st. benedict and at st. john's university. burch realized that students require a simulator to build logic circuits without the superfluous knowledge linked to its physical building. the outcome of his realization was a key motivating factor to his development of logisim. another motivating factor was the unjustifiable cost that is usually tagged onto commercial products. this is especially true in circumstances in which institutions only benefit from an extremely small portion of the complete capabilities of an extensive software product. such a case can be directly related to the software simulation product requirement for a basic course such as digital logic design. the next logical step for the author, therefore, was to build the digital logic software logisim. burch´s concern for adopting appropriate software for digital logic design was shared by other institutes. this shared concern could be regarded as one of the main drivers behind spreading the logisim project beyond its point of origin, and the use of logisim as a teaching aid has experienced phenomenal growth during the twelve years since its creation. not only has it become popular in the united states, but it also has crossed international borders. in the us, it has been adopted by 75 different institutions, including brown university, the georgia institute of technology, and princeton university. outside of the us, it has been adopted by more than 30 institutes. it has spread across 15 european countries, including germany, france and the united kingdom. on the other side of the pacific, it is used in seven institutions in both australia and new zealand. in latin america, 11 institutions use logisim in countries including brazil and argentina. however its popularity in asia is still lagging, with only 8 asian institutes having adopted logisim, including those in saudi arabia, singapore and, more recently, oman. as of this publication, no institute in africa has accepted logisim. these findings have been compiled in table 2, which summarizes the global spread of logisim. 5. integrating logisim into the course in previous years, the software package logicworks from capilano computing systems (richmond, table 1. a list of required ic used in the digital logic design course (datasheet, 2000). 73 complementing digital logic design with logisim british columbia, canada) was the main simulation tool used in labs. on the positive side, it has a free version targeting academic institutions. however, logicworks also has limitations within the university context. although logicworks is a versatile piece of software, it includes an extensive component library which is unnecessary for the course. furthermore, as logicworks is a commercial product, graduating students either would have to purchase a copy of the software to pursue any digital design circuit simulation based on logicworks, or would be forced to work at a company or institute with access to the software in order to continue using it after graduation. both scenarios are highly improbable given the relatively small electronics industry in the sultanate of oman. furthermore, while building experiments for the digital logic design course around logicworks, two important deficiencies were observed. first, within the digital logic design course, students design and analyze circuits with direct connections to single gates. unfortunately, this would rarely be the situation in the real world. circuits are naturally built around ics which include, in most cases, multiple gates within a single chip. ideally, the simulations of all experiments must be conducted using the ics found within the lab. this, however, could not be attained using logicworks as it did not include all of the desired ics within the lab environment. second, if a particular ic did exist in the logicworks library, its implementation was a source of confusion to the newcomer. this can be explained by observing the equivalent ic pin assignments of both the actual and simulation components. unfortunately, the software company did not attempt to match the pin layout of the actual component when including the ic in the library. figure 1 clarifies this problem using as an example the 3-to-8 decoder, ic 74138 (data sheet 2000). figure 1a is the actual component layout while figure 1b shows the logicworks model layout. as a result of this ic layout mismatch, the task of experimenting with and the simulation of digital circuits becomes all the more daunting and confusing to new students. to contain the three constraining factors found in logicworks, logisim was recently adopted at squ to fit the digital logic design course simulation niche. the best part is that logisim is foss which complies with the general public license (gpl). under this license, students not only experience using the software as a computer aided design (cad) tool for the course as undergraduates, but are further encouraged to take it along with them after graduating. accordingly, this software can be considered a personal cad tool that can assist graduates in the pursuit of building digital logic circuits at both professional and personal levels. moreover, logisim is light weight software in that it requires only a small amount of computing estate and is deployable in windows, linux and mac operating systems. this is attributed to the fact that the software was created using java. furthermore, the built in component libraries comprises most of the basic units required for designing digital logic circuits, which is more than required to successfully deliver the course as shall be outlined in the next section. finally, logisim has been built with the capability of expanding its built-in library with user-defined libraries. this final feature can be considered one of the most attractive of this software. consequently, any instructor of digital logic design can readily tailor the software package according to both the course requirement and lab experiments. 5.1 (a) logisim's basic capabilities this section is devoted to showing the various builttable 2. the global spread of logisim. 74 sa al-busaidi in libraries of logisim from a primer installation perspective. logisim includes seven built-in libraries, namely a wiring, gates, plexers, arithmetic, memory, base and input/output library. in the wiring library, the most important components are the input and output probes, clock, power and ground. as for the gates library, it includes all of the basic gates. the multiplexer (mux), demultiplexer (demux) and decoder all form part of the plexer library. included within the arithmetic library are the four basic arithmetic operators along with the comparator component. all three types of ffs, including the counter, register, shift register, random access memory (ram) and read-only memory (rom), form part of the memory library. while in the base library, the tools poke, edit, select, wire, text and label are readily available to the user. finally, the input/output library contains a button, joystick, keyboard, light emitting diode (led), and seven-segment and hexadecimal displays. 5.2 (b) extending logisim's library although the built-in library included within logisim can suffice to conduct most of the experiments, the built-in library components and the actual ics have different layouts. this layout difference can be a major cause of confusion for the newcomer to the field. furthermore, certain ics require a particular sequence of pin configurations to operate correctly. as an example, consider the parallel load serial out shift register, ic 74165. its package layout and pin assignment taken from an in-house built logisim library is depicted in fig. 2. to operate this ic correctly, its corresponding function table is shown in table 3 (datasheet 2002). to successfully shift 8-bits from the input pins 3 to 6 and 11 to 14, to output pin 9, it is first necessary to load the data correctly. the parallel load operation of bits a to h into the ic requires that pin 1, indicated by sh/ld', be initially low. this is followed by assigning pins 2 and 15 to low while subsequently changing pin 1 to high. the loaded bits can now be shifted out to pin 9. for every clock cycle, a transition from low to high will shift out one bit at a time, commencing from loaded bit h up to loaded bit a. finally, as a bit is transferred to an adjacent pin, a new value from pin 10 is loaded at every clock cycle. from this example, it can be stated that it is imperative that a component within logisim operates in line with its function table from the ic's data sheet. this correspondence between the actual and simulation components has clearly not been included with the built-in library components of logisim. however, included within logisim is a particularly delightful feature that allows users to build their own libraries. to achieve this, a user simply designs a circuit in the usual manner and saves it. upon opening a new project, the saved file can be included as a logisim library. it was therefore, a logical step to build up a library inclusive of all the components found within the digital lab. as a result, the ics listed in table 1 were all modeled in accordance to each individual ic function table and pin layout. 6. gained experiences of using logisim in contrast to the previous years where the pre-lab were designed at the gate level, the fall semester of 2012 has seen a shift from this routine. given that all the necessary ics had their equivalent model in our inhouse built library ic0v2.cir, every pre-lab was simulated at the ic level. by enforcing this shift, a number of visible merits were clearly observed. first, students realized that if their pre-lab operated as specified, then the physical work was no more than a mere copy and paste operation from the logisim work pane onto the breadboard. from this perspective, the meaning of simulation gained a completely new dimension and students started to appreciate the concept of simulating prior to building. moreover, the time students spent pondering figure 1. the difference between the actual chip layout (a) and logicworks layout (b). (a) (b) figure 2. a logisim representation of the in-house built parallel load serial out shift register ic 74165. 75 complementing digital logic design with logisim over their connections on the breadboard was also shortened. second, it was observed that the number of damaged components could be significantly reduced. third, by exposing students to the function tables within data sheets, it was evident that students managed successfully to operate ics with which they had had no prior experience. this had been achieved through the mindset that an unfamiliar ic should be treated as a black box which operates according to its function table only. this was especially true under the revised experiment 5 that required students primarily to understand the operation of each individual ic from its data sheet prior to connecting the ics together. in this final experiment, students had to think in terms of designing a complete circuit composed of multiple ics and viewing each ic as a simple block achieving one specific task. fourth, given that the components used in the simulator were identical to the real world components, groups of students could easily work together at their own leisure. this permitted the labs to be less frequented by digital logic design students, especially at the end of semester when students flock to conduct experiments in preparation for the lab exam. finally, as logisim follows in the footsteps of foss, students were encouraged to take their personal copies of the software and the in-house built ic library. with this final merit, it is hoped that all digital logic design students will continue using the software to solve real life problems through simulation prior to building a functional product. 7. conclusions in the foss world, there are applications that suit almost every field. this includes with no exception the field of cad tools required for digital logic design courses. one particularly nice piece of software is logisim. on the one hand it is light weight and can run on practically every operating system. on the other, it can be tailored and expanded in accordance with digital logic design course requirements. in our course, it was possible to change students' perspectives towards the importance of simulation by having them build an extensive ic library which includes ics that are linked to both the course and its experiments; this was only achievable given logisim's capability to include new libraries. this permitted modeling of ics in their actual pin configuration while realizing the ic's function table. through this, students gained greater insight into how individual ic components should be treated and connected to build a functional system. finally, as an added benefit of foss, students were encouraged to use this cad tool to tackle new problems of the type that may be encountered in their future engineering careers. references ahmad a, ruelens d, ahmad s (2013a), development of verification tool for minimal boolean equation. ieee technology and engineering education, (itee) 8(4):29-34. ahmad a, al-busaidi ss, al-mushrafi mj (2013b), on properties of pn sequences generated by lfsr a generalized study and simulation modeling. indian journal of science and technology 10(10):5351-5358. ahmad a, ruelens d (2013), development of digital logic design teaching tool using matlab & simulink. ieee technology and engineering education 8(1):7-11. ahmad a, al-abri d, al-busaidi ss (2012), adding pseudo-random test sequence generator in the test simulator for dft approach. computer technology and applications 3(7):463-470. ahmad a, al-abri d (2012), design of a pseudo-random binary code generator via a developed simulation model. aceee international journal on information technology 2(1):33-36. ahmad a, al-abri d (2010), design of an optimal test simulator for built-in self test environment. journal of engineering research 7(2):69-79. al-busaidi s (2013), complementing digital logic design with logisim. proceedings of free and open source software conference (fossc-13), held at sultan qaboos university, muscat, oman 18-19 february, 4-9. table 3. function table of ic 74165. 76 sa al-busaidi al-lawati a, ahmad a (2003), realization of a simplified controllability computation procedure a matlab-simulink based tool. journal for scientific research science and technology 8:131-143. burch c (2000), logisim: a graphical system for logic circuit design and simulation. journal of educational and resources in computing 2(1):516. data sheet of "ic dm74ls138 decoder / demultiplexer", fairchild semiconductor, revised march 2000. data sheet of "ic 74165 parallel-load 8-bit shift register", ti, sdls0620, revised february 2002. http://maps.google.com/maps/ms?ie=utf8&oe=utf 8&msa=0&msid=209845857912254026337.0004 9c67b154d0e5433a0. tocci rj, widmer ns, moss gl (2010), digital systems: principles and applications. eleventh edition, prentice hall. ece060214.qxd the journal of engineering research vol. 4, no.1 (2007) 69-74 1. introduction amplifiers for fiber-optic receivers require very high performance features in order to achieve the high gain, wideband, low noise characteristics required for optical communication systems. some very elegant called "current-mode" optical transimpedance amplifiers (tias) using bicmos process have been reported (halkias et al., 2000; vansiri and toumazou, 1995; toumazou and park, 1996). these tias are based on a common-base (cb) architecture, employing current shunt feedback. two main advantages of the current-mode approach are the very wide bandwidth and the low-input resistance of the circuit. this renders the amplifier bandwidth insensitive to the capacitance of the pin photodiode at the input. in pin photodiode receivers, the tia has to achieve wide band _______________________________________ *corresponding author’s e-mail: touatif@squ.edu.om width, low noise, high and accurate gain, and low power consumption. however, these requirements conflict with each other, requiring tradeoffs to be made to suit a particular application. a current mode cb configuration has the potential to achieve simultaneously high bandwidth and high feedback resistors, potentially improving the amplifier gain and noise performance (vansiri and toumazou, 1995; toumazou and park, 1996; ikeda et al. 2001). in this paper, new design approaches of bicmos-cb tias in a current-feedback configuration for high-speed optical communication applications are presented. the tias are design optimized for implementation with a standard, commercially available, 0.8-µm bicmos technology. a detailed analysis on the effect of moving toward a more-fet based design is presented. the results are compared to recent similar designs. high-performance bicmos transimpedance amplifiers for fiber-optic receivers f. touati*1 and m. loulou2 *1 department of electrical and computer engineering, sultan qaboos university, p. o. box 33, al-khoud 123, muscat, sultanate of oman 2 ecole nationale d’ingenieurs de sfaz, tunisie received 14 february 2006; accepted 10 june 2006 abstract: high gain, wide bandwidth, low noise, and low-power transimpedance amplifiers based on new bicmos common-base topologies have been designed for fiber-optic receivers. in particular a design approach, hereafter called "a morefet approach", added a new dimension to effectively optimize performance tradeoffs inherent in such circuits. using conventional silicon 0.8 µm process parameters, simulated performance features of a total-fet transimpedance amplifier operating at 7.2 ghz, which is close to the technology ft of 12 ghz, are presented. the results are superior to those of similar recent designs and comparable to ic designs using gaas technology. a detailed analysis of the design architecture, including a discussion on the effects of moving toward more fet-based designs is presented. keywords: optical receiver, transimpedance bicmos amplifier, 0.8µm silicon technology, low power amplifier äéªî° etransimpadencebicomsá«fƒ° dg ∑ó°s’g èy ∫éñ≤à°sg iõ¡lc’ á«déy äéø°ugƒe ägp »jgƒàdg .±1ƒdƒd .ω h2 áá°°uuóóÿÿgggg äéî° e º«ª°üj ” :transimpadence é«lƒ∏æµj ≈∏y ~ªà©j i~j~lbicoms.áñfƒ° dg ∑ó°s’g èy ∫éñ≤à°sg iõ¡l’ ∂dph ácî°ûÿg i~yé≤dg hp »ª°ùj »àdgh º«ª°üà∏d i~j~÷g á≤jô£dg âféch''''fet-approachre mo"a∂∏j ‘ á«©«ñ£dg ägoé° àÿg ™e í‚cg áø°üh πeé©àdg á«ø«µd g~j~l g~©h âaé°vcg ~b ¿ƒµ«∏°ùdg é«lƒ∏æµj ωg~îà°séhh .äéªî° ÿgm0.8µ äéªî° e agocg ¢üfé°üÿ iécéùg èféàf ¢vôy ”total-fet ooôj ≈∏y π¨à°ûj 7.2ghzøe áñjôb »g »àdgh á∏ª©à°ùÿg é«lƒ∏æµà∏d iƒ°ü≤dg ᪫≤dg.12ghzé«lƒ∏æµj ω~îà°ùj iôncg ™e áhqé≤àeh é«lƒ∏æµàdg ¢ùøæd áãj~m iông çéëhg èféàf ¥ƒøj é¡«∏y π°üëàÿg èféàædg âféc gaas ≈∏y é«∏c ~ªà©j º«ª°üj øe ügîb’g òkéj á«ø«c á°ûbéæe ™e i~j~÷g ä骫ª°üàdg ∂∏àd ™°sƒe π«∏– ¢vôy ” áfg éªc .ìcg áø∏µe øµdh ´ô°sg »g »àdgfet≈∏y .äéªî° ª∏d ago’g ¢üfé°ün áá««mmééààøøÿÿgg ääggooôôøøÿÿgg ºî° e ,»fƒ°v ∫éñ≤à°sg ré¡l :transimpadencebicoms ¿ƒµ«∏°ùdg é«lƒ∏æµj ,m0.8µ .áø«©°v iq~b ∂∏¡à°ùj ºî° e , 70 the journal of engineering research vol. 4, no.1 (2007) 69-74 2. quantitative description of architecture the new cb topologies are shown in figs. 1 and 2 (design 1 and 2), respectively. the basic architecture consists of a cb input stage q1 followed by a common-source amplifier stage m2, and q2 and q3 are buffers. another cb design reported in (toumazou and park, 1996) is also shown in fig. 3 for comparison. what is special about the new designs compared to previous cb schemes is that it is a total or quasi-fet based approach. that is, solely fets are used to bias the amplification stages (q1 and m2). in the present designs (1 and 2), fet current mirrors are used to bias stages, where the initial current source of 5 ma must be very stable against external factors (temperature, biasing). therefore, bandgap reference-based techniques must be used to implement that 5-ma current source. also, the feedback resistor is connected from the output of q2 via rf to the input of m2 and not to the emitter of q1 (ie. the amplifier input), as would be the conventional case. as a result, two straightforward advantages emerge: 1) the total power consumption is lower for the same gain and 2) these fets should be sized in such a way to increase the gain while optimizing noise performance of the amplifier. this constitutes a new dimension added to optimize performance tradeoffs inherent in such designs while maintaining device processing yield high. the input transistor is a bipolar transistor. therefore, the collector bias current required is less than the drain bias current required in a common-gate configuration for the same transconductance gm, hence leading to a lowerpower consumption. alternatively, for the same supply voltage the transconductance gm is higher in the bipolar transistor case and hence the noise contribution is lower. the amplifier stage is selected to be an fet (m2) because the input current noise is less than that of a bipolar transistor and because the optimum size can be selected. going from design 3 via 2 to 1, the circuit becomes more and more fet-based. we will see what impact this would have on performance features. 3. dc analysis the accurate prediction of the dc operating point of such circuits is of critical importance because the various stages are bias interdependent. extensive dc analysis and optimization of the new cb design circuits was performed. the resistive shunt-feedback reduces the circuit sensitivity to external factors such as biasing, process tolerances, and temperature. also, the bjt at the input stage gives a higher transconductance for the same supply voltage as explained above, resulting in two main advantages: 1) a higher transimpedance gain (hence low noise) and 2) a low input impedance (almost 1/gm1). the main objective to accommodate a low dc power consumption, a high transimpedance gain, low-noise, and to insure bias insensitivity within a relatively large range of supply voltages has been successively met. the positive and negative supply voltages vcc and vee (in figs. 1 and 2) can be varied within the range from +3.0 to 6.0 v and from -5.8 to -4.6 v, respectively, without degrading performance. the total dc power consumption (no load) for design 1, 2, and 3 is 56.6, 51.8, and 72.3 mw, respectively. this indicates that the new designs (1 and 2) are better optimized for low power consumption. figure 1. bicmos cb tia (design 1) figure 2. bicmos cb tia (design 2) figure 3. bicmos common-base tia (design 3) (toumazou and park, 1996) 71 the journal of engineering research vol. 4, no.1 (2007) 69-74 4. ac analysis the common-base topology adopted makes the -3 db bandwidth of the amplifier totally independent of the photodiode input capacitance, which determines only a nondominant pole since the input resistance of m1 is small (toumazou and park, 1996). figure 4 shows the small signal hybrid π model of the common-base transimpedance amplifier of fig. 2 at high frequency. here the base spreading resistances of bipolar junction transistors were neglected. the transfer function of this amplifier is governed by five major time constants: at the input (tin), associated with the miller feedback resistance at the input of m2 (trfmin), associated with the miller feedback resistance at the output of m2 (trfmout), associated with the base-emitter capacitance of q2 (tcbe2), and at the output (tout). let us assume that trfmin is the dominant time constant (which often has a much lower value) when compared to the other time constants. therefore, the -3 db bandwidth of the amplifier is approximately determined by the cut-off frequency as: (1) therefore, by combining all the above equations eq. (1) becomes: (2) the dominant pole of the amplifier, which determines the -3 db bandwidth, depends mainly on the input capacitance of m2 with miller effect, the input capacitance of q1, the drain capacitance of m4, and the feedback resistor rf. since m2 is a much smaller device, its net input capacitance is small and hence rf can be made larger for the same equivalent bandwidth, which would result in a lower overall noise. however, smaller device means higher resistance which would reduce the bandwidth. this would impose a careful sizing of m2 when dealing with the tradeoff between low noise and wide bandwidth. on the other hand, m4 should be made narrow in order to reduce its noise contribution while preserving adequate bias collector current for q1. in fig. 1, the size of m5, load of m2, can be selected to optimize av2 for wide bandwidth. a critical aspect of the design phase was the optimization of the feedback resistor rf as well as the size of the fet transistors in a tradeoff between gain, bandwidth, noise, and dc power consumption. it was shown (abidi, 1988) that optimum noise performance may be obtained in transimpedance amplifiers employing submicron fet input stages by choosing the width of the input device such that its gate capacitance is one-fifth of the sum of the photodiode and stray capacitance. based on this criterion, it was concluded that for design 1 (design 2), the optimum value of rf = 2.3 kω (rf = 2.1 kω) and w2 = 30 µm (w2 = 30 µm) for l = 0.8 ?m allow the desired transimpedance gain, minimum input noise current density, and low power consumption over the bandwidth of dc-7.2 ghz (4.4 ghz) to be achieved. this was obtained through successive cycles of dc and rf simulations. simulated results of transimpedance gain of the new designs (1 and 2) are shown in fig. 5. also shown for comparison, the curve reported in (toumazou and park, 1996) (design 3). state-of-the-art 12 ghz ft silicon bicmos parts of a conventional 0.8-µm bicmos technology were used. it can be seen that going toward a more fet-based configuration (ie. from design 3 via 2 to 1), the transimpedance gain as well as the -3 db bandwidth improves. this trend is more prominent when one compares design 1 (thick solid line in fig. 5) with design 3 (fine thin solid line in fig. 5). this can be accounted for by the increasingly higher feedback resistor rf = 1.4kω, 2.1kω, and 2.3kω which was achieved for design 3, 2, and 1, respectively. this would lead to a lower thermal noise contribution from the feedback resistor. this high transimpedance gain is also due to using more and more active loads from design 3 to design 1 instead of resistive loads, since the closed-loop gain depends on rf and the openloop gain. fminr db3 t2 1 f π ≅− where: tcminrfminrt = with 2va1 fr minr + = . av2 is the voltage gain of m2 and 2minc2gsc4gdc1ctc +++= µ . the capacitance cì 1 is the base-to-collector capacitance of q1, cgs2 and cgd2 are the gate-to-source and gate-todrain capacitances of m2, cgd4 is the gate-to-drain capacitance of m4, and cmin2 is the miller capacitance of the gate-to-drain capacitance cgd2 of m2 at its input and given by ( ) 2gd2v2min ca1c += . ft 2v db3 rc2 )a1( f π + ≅− figure 4. small signal hybrid π model of the common-base amplifier of fig. 2 72 the journal of engineering research vol. 4, no.1 (2007) 69-74 5. noise analysis the noise model, used to calculate the noise components, combines the conventional fet rms noise theory with the optical preamplifier (oeic) noise theory (minasian, 1987; smith and personick, 1980). the total input-referred equivalent noise current spectral density of designs 1 and 2 is approximately given by: (3) in eq. (3), k is the boltzmann constant, t is the absolute temperature, rbb1 is the base spreading resistance of q1 (0.3 ω), cd is the capacitance of the pin photodiode (0.32 pf which is typical for a pin photodiode), rbe1 = kt/qib1 with ib1 (27 ua) is the base current of q1, cbe1 (0.8 pf) is the base-emitter capacitance of q1 and cµ1 (defined above) = 0.2 pf, β1 (100) is the ac current gain of q1, ig2 is the gate leakage current of m2, cgs2 = 1.0 pf, cgd2 = cgd4 = 0.3 pf, gm2 = 7.0 ms, gm4 = 15.0 ms, q is the electron charge, γ is the excess noise factor equal to 1.7 (abidi, 1988) to account for the short-channel effects. the first three terms in eq. (3) represent thermal noise contributions by rf, re1, and rbb1, respectively. the next two terms are base and collector shot noise current of q1, the sixth term is noise contribution by gate leakage current of m2, which can be neglected (halkias et al. 2000). the last two terms represent the channel thermal noises of m2 and m4, respectively. figure 6 shows simulated results of the total rms inputreferred noise current of the new designs, as calculated using eq. (3) over a bandwidth of 10 ghz. also shown for comparison, the results reported in (toumazou and park, 1996) (design 3). state-of-the-art 12 ghz ft silicon bicmos parts of a conventional 0.8 µm bicmos technology have been used. two important features are clear. first, above around 5.0 ghz the new designs offer a lower total noise performance compared to that of (toumazou and park, 1996), whereas the opposite occurs otherwise. second, moving toward higher frequencies, the noise current becomes increasingly lower when we move toward a more fet-based design (solid line). this is important since the lower noise at higher frequencies offered by the total-fet based design will result in a lower ber at high data rates. preliminary results showed that the addition of the inductor l = 2-15 nh (realizable monolithically as transmission lines) as shown in fig. 1 results in a significant noise reduction. this can be explained by the reduction of noise contribution of m4 (last term in eq. (3)) by adding the inductor. m4 noise spectral density with the inductor becomes: (4) from eq. (4), with gm4 = 30 ms, and l = 15 nh, the inductor will reduce the noise spectral density of m4 by a factor of 1/3 at 1 ghz. table 1 summarizes performance features of the new cb designs (1 and 2) and those of (toumazou and park, 1996). it is clear that the results obtained with the new cb designs are superior to those reported in (toumazou and park, 1996) and in (park and yoo, 2003). also, these results are better than those reported in (haralabidis et al., 2000) and fairly approach those obtained with gaas mesfet and inp-hbt (huber et al. 2000; yoneyama et al. 2000; minasian, 1987). figure 5. transimpedance gain versus frequency ( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ +++ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ ++++ + ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ ++++ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ +++= 2 1bed 2 2 1e1 1be 1 1be4m 2 4gd12gs 2 2 f2m 2g 2 1bed 2 2 1e1 1be 1be 2 d 2 2 1e 1bb 1ef eq )cc()f2( r 1ktr2rg )ccc()f2( r 1 g kt4 qi2cc)f2( r 1ktr2 r kt2 )c()f2( r 1 ktr4 r kt4 r kt4 )f(s π ββ π γ π β π µ figure 6. total mean input-referred noise current versus frequency ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ ++ + 2 1bed 2 2 1e1 1be 1 1be 2 4m 4m )cc()f2( r 1ktr2r )fl2g(1 g π ββπ 73 the journal of engineering research vol. 4, no.1 (2007) 69-74 figure 7 shows the transient response of the previous cb tia designs when driven by a 2.5 gbps input current pulse train. it can be clearly seen that moving toward a more fet-based design, the output swing widens and the response becomes faster. although the settling time in this case is bigger, still the response is faster and settles within the time frame of the driving pulses period. for the total-fet based design (thick solid line), the response clearly reaches 200 mv output swing for a 100 µa input current. according to that, one can expect a sufficiently open eye diagram at 2.5 gbps for nrz synchronous links. noise simulations of the new designs have also been conducted using 0.6 µm bicmos process parameters. results showed even lower total rms noise current; more than 1 db lower in the frequency range above 4.0 ghz. this may indicate the validity of the new design approach for shorter channels also. further investigations are in order in this regard. 6. conclusions high-performance bicmos common-base transimpedance amplifiers for fiber-optic receivers have been improved. in particular, a more fet-based approach was found effective in potentially optimizing performance features. simulated results showed improved transimpedance gain, noise characteristics, bandwidth, and power consumption when compared to recent similar designs. a transimpedance gain as high as 65.2 db over a bandwidth of 7.2 ghz, which is close to the technology ft of 12 ghz, was achieved. the 3.5 ghz and 7.0 ghz total equivalent input-referred noise current have been minimized to achieve 17.7 and 28.7 pa/hz0.5, respectively. also the total power consumption of the circuit was optimized to 56.6 mw. these performance features are fairly comparable to those reported using gaas mesfet. while optimizing the design for gain, noise and power, little attention has been given to dynamic range and output swing. however, now that the new approach is demonstrated, much study is aimed at optimizing other performance features. acknowledgments the author wish to thank sultan qaboos university (squ) for providing the facilities and resources. also, the authors wish to thank the tunisian ministry of scientific research technology and development of competencies for its support and grant in the frame of the program of cooperation with the tunisian researchers abroad. references abidi, a., 1988, "on the choice of optimum fet size in wide-band transimpedance amplifiers," j. of lightwave technology, vol. 6(1), pp. 64-66. halkias, g., haralabidis, n., kyriakis-bitzaros, e. d. and katsafouros, s., 2000, "1.7 ghz bipolar optoelectronic receiver using conventional 0.8?m bicmos process," ieee int. symp. on circuits & systems, pp. v417-v420. haralabidis, n., katsafouros, s. and halkias, g., 2000, "a 1 ghz cmos transimpedance amplifier for chipto-chip optical interconnects," ieee int. symp. on circuits & systems, pp. v421-v424. huber, d., bauknecht, r., bergamaschi, c., bitter, m., huber, a., morf, t., neiger, a., rohner, m., schnyder, i., schwarz, v. and jackel, h., 2000, "inpingaas single hbt technology for photoreceiver oeic's at 40 gb/s and beyond," j. of lightwave technology, vol. 18, pp. 992-999. ikeda, h., ohshima, t., tsunotani, m., ichioka, t. and kimura, t., 2001, "an auto-gain control transimpedance amplifier with low noise and wide input dynamic range 10-gb/s optical communication systems," ieee j. of solid-state circuits, vol. 36(9), pp. 1303-1308. minasian, r., 1987, "optimum design of a 4gbit/s gaas mesfet optical preamplifier," j. of lightwave technology, vol. lt-5(3), pp. 373-379. park, s. and yoo, h., 2003, "2.5 gbit/s cmos transimpedance amplifier for optical communication applications," electronics letters, vol. 39(2), pp. 211-212. gai n (db) bw (ghz) noise (pa/hz0.5) @ 3.5 7 ghz total power (mw) new design 1 65.2 7.2 17.7 28.7 56.6 new design 2 62.1 4.4 18.3 29.4 51.8 design 3 [3] 59.8 4.2 15.2 35.5 72.3 table 1. bicmos cb tia design parameters figure 7. transient response of the cb bicmos tias at 2.5 gbps for a 100µa pulse input current signal 74 the journal of engineering research vol. 4, no.1 (2007) 69-74 smith, r. and personick, s., 1980, "receiver design for optical fiber communication systems", chapter 4. publisher: springer-verlag. toumazou, c. and park, s., 1996, "wideband low noise cmos transimpedance amplifier for gigahertz operation," electronics letters, vol. 32(13), pp. 1194-1996. vansiri, t. and toumazou, c., 1995, "integrated high frequency low-noise curret-mode optical transimpedance preamplifiers: theory and practice," ieee j. of solid-state circuits, vol. 30(6), pp. 677685. yoneyama, m., miyamoto, y., otsuji, t., toba, h., yamane, y., ishibashi, t. and miyazawa, h., 2000, "fully electrical 40gb/s tdm system prototype based inp hemt digital ic technologies," j. of lightwave technology, vol. 18, pp. 34-43. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 5.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 41-50 kinetics and mechanism of bulk polymerization of vinyl chloride in a polymerization reactor a. s. ibrahim*, y. a. ali, h. m. saad and i. h. amur chemical engineering department, dhofar university, sultanate of oman received 1 march 2015; accepted 4 may 2015 abstract: polyvinyl chloride (pvc) is the third most commonly produced polymer and is important because of its mechanical characteristics. the most common method of pvc manufacturing is the process of suspension. although, there are several benefits associated with suspension, this study will focus on the bulk polymerization of vinyl chloride; highlight the physical and chemical properties of pvc, which can be changed through an estimation of the optimum ratio that exists between the hydrophilic and hydrophobic parts of the polymer’s surface, and propose a new mathematical model which will be helpful for the conversion of pvc into a useful form. the result will be the proposal of a new dynamic mathematical model for the three-phase structure model. all particles have been taken into account in the proposed model, which helped contribute to the reaction in gel, solid, and liquid phases, emphasizing the use of mercury (hg) as a catalyst. the proposed mathematical model considers the heat and mass transfer between the liquid, gel, and solid phases with chemical reactions that occur between the liquid and solid phases, and between the gel and solid phases. the effect of the catalyst and volumetric flow rates of vinyl chloride monomer (vcm) on the system have been evaluated through the proposed mathematical model. furthermore, the study’s experimental data have been compared with the findings of the suggested model in the context of concentration and temperature reaction. obtained results show good agreement between the proposed mathematical model and the actual plant data. keywords: mathematical model, catalyst, polymerization reactor, chemical reaction, heat and mass transfer. corresponding author’s e-mail: ahmadsaadi@gmail.com a. s. ibrahim, y. a. ali, h. m. saad and i. h. amur 42 nomenclature symbol definition, unit a cross-sectional area of the reactor, (m2)c concentration of vinyl chloride monomer, mol/m3c concentration of poly vinyl chloride, mol/m3c concentration of gel phase, mol/m3c heat capacity, kj/mol.km mass rate of vinyl chloride monomer, mol/secmw catalyst molecular weight, kg/mol mw molecular weight of vcm, kg/molk mass transfer coefficient from liquid to gel phases, 1/seck mass transfer coefficient from gel to solid phases, 1/sec ∆h heat enthalpy of reaction, kj/mol.kf volumetric flow rate, m3/s u bubble gas velocity, m/s t feed temperature, kt temperature of gel phase, k r overall rate of reaction, mol/ kg cat.m2 ρ catalyst density, kg/ 3m to temperature of reaction, kv volume of reactor, m3 t time, sec cat catalyst ki rate of activation kin rate of initiation kp rate of propagation kliv rate of live polymer kdead rate of dead polymer kinetics and mechanism of bulk polymerization of vinyl chloride in a polymerization reactor 43 1. introduction polyvinyl chloride (pvc) is considered one of the most versatile polymers due to its mechanical properties and is the third most frequently produced polymer. in 2014, the quantity of pvc produced reached more than 50,000,000 tons because of its wide demand in different fields. furthermore, the production of pvc is expected to increase in upcoming years due to its significant and effective performance in different fields. according to ambrozek et al. (2012), appropriate and suitable techniques must be adopted for the conversion of pvc into a usable, practical form. a study by lăzăroaie et al. (2010) established that pvc can be generated commercially through different approaches and techniques, including suspension emulsion, solution polymerization, and bulk processing. according to ibrahem (2012), the process of suspension is one of the most widelyused techniques for pvc manufacturing because an emission technique is used for the production of almost 12% of pvc. additionally, 80% of pvc is manufactured with the help of polymerization, 7.5% is manufactured through the emulsion process, and 0.5% is manufactured by solution. among all of these processes and approaches, the process of suspension has been found to be the most common and effective approach due to the fact that the suspension process provides the advantage of inexpensive dying and enables inexpressive separation of resin. additionally, the suspension process also helps in providing strong control over the entire manufacturing process. besides focusing on the advantages of the suspension process, this paper will emphasize the bulk process of the polymerization of vinyl chloride. kamo et al. (2012) stated that polymerization of monomers is carried out without the need of diluents in the bulk process. the research further asserted that one of the major benefits associated with the bulk process is polymer fabrication. it has been ascertained that fabricated polymers do not contain any impurities. on the other hand, lathia et al. (2004) identified a number of disadvantages associated with the bulk process, including limited and/or inadequate control of excessive removal of heat and rate of reaction. in this regard, (blazevska and spasesk 2010) proposed a mathematical model which can be used for predicting the best possible range of temperature— between 40–70°c for a pvc polymerization system. the proposed mathematical model was dependent on the ratio of the mass transfer coefficient between liquid and gel phases. an evaluation of the research by balakrishnam et al. (2005) revealed that the analysis and study of reaction conditions is highly significant due to the fact that it assists in directing the pressure, catalyst flow rate, and temperature effects. all of these actions play a major and effective role in ensuring the reduction of the effects of pollution within the environment. a number of studies have sought to analyze and examine thermal properties of pvc in normal environmental conditions. in order to carry out such analyses, the research mathematical model systems were employed. chavadej et al. (2008) developed an integrated and comprehensive mathematical model so as to examine the properties and behaviors of industrial pvc. chattopadhyay and madras (2002) documented a number of studies to assess different characteristics of pvc, including its stability, degradation rates, and modification surfaces of pvc for several applications in medical research (blazevska-gilev and spasesk 2010; chattopadhyay and madras 2002; cohan 1975; ibrahem 2012; kamo et al. 2012; maezawa et al. 2004; scott et al. 1990). zuga and cincu (2006) stated that the physical and chemical properties of pvc can be altered by carrying out calculations on the optimum ratio between the hydrophilic and hydrophobic parts of a polymer’s surface. according to (blazevska and spasesk 2010), this calculation can assist in converting pvc into a valuable and utilizable form and can be deployed in several different applications. it seems necessary to discern whether the scaled-up pilot plan has been deployed so as to carry out the comparison of simulation results obtained through experimental study. it also appears essential that the modeling of pvc’s adsorption systems must be taken into account. such modeling and comparison will help in assessing the effects of the catalyst flow rate on the reaction’s temperature; specifically, the impact of potassium/titanium oxide (pt/tio2) must be analyzed. in addition to this, it would also help in determining the variation in the concentration of emulsion reactants with the passage of time (blazevska and spasesk 2010; field 1973; kiparissides et al. 1994; zuga and cincu 2006). the current research will elaborate a newly proposed mathematical model which will be useful for the transformation of pvc into a utilizable form. the study further seeks to determine whether the established mathematical model can be applied to gel, liquid, and solid phases so as to carry out comparisons between the experimental results. despite the mathematical model, this study intends to carry out a simulation so as to evaluate as well as a. s. ibrahim, y. a. ali, h. m. saad and i. h. amur 44 examine the flow rate of a vinylchloride monomer (vcm). additionally, a consideration of simulation results will also help in identifying the impact of a catalyst’s flow rate on reaction temperature and concentration (chattopadhyay and madras 2002; cohan 1975; field 1973; ibrahem 2012; kamo et al. 2012). in addition to this, the descriptive behavior of the newly proposed mathematical model for pvc in a continuous stirred tank reactor (cstr) will also be highlighted in this study. 2. proposed descriptive behavior of the new mathematical model for the pvc cstr process in this account, several assumptions will be taken into account in order to assure the integrity and reliability of the process [table 1]. vcm will be mixed with water in order to run the experiment, which will be carried out by allowing the entry of a mixture reactant from the top of the cstr. the reaction will be accomplished in the liquid phase by allowing the catalyst, hg, to flow down within the reactor. this process will be carried out with the help of five major mechanisms: initiation, rates of activation, propagation, a living polymer transfer. and the death of a polymer chain. a chemical reaction will be activated between the reactants and the hg catalyst in the liquid phase. however, almost 70–90% of the conversion of vcm is generally in the solid phase, while ~10–30% is transferred in the gel phase. afterwards, the vcm will be converted into the product of the pvc solid phase (ambrozek et al. 2012; balakrishnan et al. 2005; chattopadhyay and madras 2002; field 1973). this proposed model accounts for the effects of all of the most important variables in the system [fig. 1]. subsequently, the product will flow down into the vcm which is not converted. catalyst liquid and pvc solid phases then will he separated from the bottom of the cstr. 3. assumptions of the modified mathematical model it is essential that several different assumptions are taken into account when modifying the mathematical model as these assumptions will help in assessing the credibility and reliability of the final findings. in light of this, it has been presumed that cstr includes the gel, solid, and liquid phases. also assumed is that reactions occur with catalysts during the liquid phrase (kiparissides et al. 1994). in addition to this, it is also observed and identified in the new mathematical model that all three phases take place in controlled conditions. there are several assumptions which need to be taken into account and include minor and insignificant temperature and concentration gradients within the cstr. the altered mathematical model also takes into consideration the possibility that the liquid phase is not comprised of two main elements—vm or hg. the study further assumes that the dynamics of different reactions are explained by the reaction rates of different processes (zuga and cincu 2006). reaction rates of activation, initiation, propagation, living polymer transfer, and the death of polymer chains are some of the most prominent processes parameter which must be defined. table 1. modified mathematical model assumptions. 1. the cstr comprises the liquid, gel, and solid phases. 2. reactions occur within the catalyst liquid phase. 3. all phases are at controllable conditions. 4. there are negligible radial temperatures and concentration gradients in the cstr. 5. the liquid phase is composed of vcm and hg. 6. the dynamics of reactions are represented by the rate of reaction rates of activation, initiation, propagation, the transfer of living polymers and the death of polymer chains. cstr = continuous stirred tank reactor; vcm = vinyl chloride monomer; hg = mercury 4. reaction kinetics when the reaction kinetics were analyzed, the rate of reaction while using the catalyst, involved five main mechanisms. these five mechanisms include rates of activation, initiation, and/or propagation; the transfer of living polymers; and the death of polymer chains. these steps, which are included in the reaction, involve mass and heat transfer, the diffusion of reactants from the bulk of vcm, and activation using a hg catalyst flow rate inside process (maezawa et al. 2004). all of these steps can significantly influence the overall reaction and mathematical model design for the cstr process [fig. 1]. kinetics and mechanism of bulk polymerization of vinyl chloride in a polymerization reactor 45 figure 1. the mechanism of the chemical reaction. 5. kinetic reaction the kinetic reaction includes five mechanisms of the free radical method: 1. rate of activation to activate two free radical alkyl groups [eqn. 1] ………… ……….. (1) 2. rate of initiation of vcm r + . …………… …………. (2) 3. rate of propagation to produce a radical-free chain of the units of the monomer . . ……… … (3) 4. rate of the living polymer to transfer to the monomer ..…… … (4) 5. rate of death in the live polymer to stop the propagation and life of polymer chains . . (5) 6. from equations 1–5, the overall rate of reaction (r ) can be determined by the following eqn. 6: i + ……… ..... (6) a. s. ibrahim, y. a. ali, h. m. saad and i. h. amur 46 where, exp eg = activation energy aer = arrhenius number (7) r = gas constant, ie. 8.314 kj/kmol.k t = temperature, ie. °k 6. cstr reactor modeling according to balakrishnam et al. (2005), the main reaction process within the cstr contains several steps, including the three major steps which are incorporated in this process: 1. heat and mass conversion of vcm from a liquid phase directly to the solid phase of pvc through a chemical reaction. 2. mass and heat transformation of vcm from a liquid phase directly to the gel phase of pvc without a chemical reaction. 3. mass and heat transfer of vcm from a gel phase to a solid phase of pvc through a chemical reaction. 7. mass and heat transfer of vcm from a liquid phase directly to a solid phase of pvc with a chemical reaction in this state, 70–90% of vcm transfers directly to the pvc. mass and heat transfers with a chemical reaction represent five rates of reactions (eqns. 1–7) and are represented in eqns. 8 and 9, respectively. c c r aρ (8) m c t t r ∆h amw (9) ρ vcmw dtdt 8. mass and heat transfer of vcm from a liquid phase directly to a gel phase of pvc without a chemical reaction mass and heat transfer without chemical reactions are represented in eqns. 10 and 11, respectively. k c c (10) m c t t (11) 9. mass and heat transfer of the gel phase of vcm to the solid phase of pvc through a chemical reaction in this state, 75–90% of the vcm from a gel phase transfers directly to the pvc. mass and heat transfers via a chemical reaction representing the five rates of reactions [eqns. 1–7] are represented in eqns. 12 and 13, respectively. k c c r aρ (12) m c t t r ∆h amw (13) ρ fmw ρ vcmw dtdt 10. model solution and analysis equations 1–13 represent a previously revealed process model which encapsulates the parameter values included in table 2. the differential algebraic equation (dae) solver of matlab software (the mathworks, inc., natick, massachusetts, usa) was used to determine values. in addition to the dae, the modified fourth order of the runge-kutta method’s variable step size was also incorporated. the operating parameters as well as physical constants were taken into account [table 2] and deployed in a mathematical model system. simulation of the process was carried out to determine the impacts of the velocity of vcm flow rate on temperature and concentration reactions in the model. the simulation makes use of variable time and evaluates the concentration and temperature of the intermediate gel state. for different phases of the system, the simulation results will be described in the following section of the paper. the effect of the vcm flow rate on the concentration in the process with respect to time about 10–30% of vcm transfers to a gel phase. kinetics and mechanism of bulk polymerization of vinyl chloride in a polymerization reactor 47 table 2. physical constants and operating parameters for the mathematical model system. 6.15 m2 3.12×1012 sec-1 ρ 13534 kg/m3v 75 m3c 1050 j/kg.k t 35oc tg 55oc t 63oc ∆h 91.6 kj/mol ea 107033 kj /mol mw 200.6 kg/kmol ρ 860 kg/m3mw 62.5 kg/molkk 0.57 kk 0.69100% ki 2.889 × 105 sec-1 kin 5.779× 105 sec-1 kp 5.779× 105 sec-1 klive 1.506 × 102 sec-1 kdead 3.125× 102 sec-1 and the conversion of the reaction is proportional to the flow rate of vcm [fig. 2]. the conversion is equal to 7.2% when the vcm flow rate is 0.004m3/second at 2.16 hours. the conversion is equal to 12% when the vcm flow rate is 0.02 m3/second at 1.634 hours, and the conversion is equal to 21.3% when the vcm flow rate is 0.05 m3/second at 0.56 hours. pvc production is proportional to the vcm flow rate [fig. 2]. gel concentration is inversley proportional to the vcm flow rate [fig. 3], especially for the first three hours of the reaction due to about 90% of the vcm being transferred to pvc. figure 3 shows two different vcm flow rates (0.004 m3/second and 0.02 m3/second) and their effects on the concentration of the gel phase. the concentration increases more for vcm = 0.004 m3/second as compared to the vcm = 0.02 m3/sec. vcm consumption and the production of pvc and of gel phases increased with time [fig. 4]. concentrations of pvc and the gel phase depend on the vcm flow rate and catalyst flow rate [figs. 2–4]. the temperature of the gel phase increased with time until it reaches the reaction temperature of the system [fig. 5]. it was found that the temperature of the reaction decreased for the first 10 minutes of the reaction because a parallel reaction forms the gel phase and absorbs the energy at the beginning of reaction. at that point, the overall reaction temperature of the system increased until it reached one temperature which represents the entire system [fig. 6]. figure 7 shows a comparison between the gel phase and reaction temperatures at the first ten minutes. figure 2. effects of vcm flow rate in the liquid and solid phases with variable time for reaction concentration. 0 1 2 3 4 5 6 7 8 9 10 7.5 8 8.5 9 9.5 10 time (hr) vcm=0.004 m3/sec vcm=0.02 m3/sec vcm=0.05 m3/sec a. s. ibrahim, y. a. ali, h. m. saad and i. h. amur 50 figure 3. effect of vcm flow rate on the gel concentration changes with the time. figure 4. variation in concentrations of vcm and gel with the time. figure 5. variation in gel temperature with the time. figure 6. variation in reaction temperature with the time. figure 7. variation in reaction and gel temperatures with the time. 11. model validation study the accuracy and reliability of the steady state behavior of this proposed mathematical model can be seen by making a comparison with the published experimental data presented by chattopadhyay and madras (2002). figures 8 and 9 illustrate the concentration and reaction temperature of pvc, respectively. the condition related to ten different samples presented by chattopadhyay and madras (2002) are compared in the two figures. in figs. 8 and 9, for instance, the first sample differs by only about 0.2 kmol/m3 of concentration and 3.2°k of reaction temperature with respect to the actual plant results, respectively. in fig. 10, the comparison between experimental data and model predictions is given for a typical polymerization. the comparison was performed in the presence of all rates of reaction. data illustrated by both figures show good agreement between the results obtained from the proposed model and the actual experimental results. 48 kinetics and mechanism of bulk polymerization of vinyl chloride in a polymerization reactor 49 figure 8. actual plant versus new model predicted concentration values. figure 9. actual versus model predicted reaction temperature. figure 10. comparison between experimental measurements and theoretical mathematical model predictions on reaction temperature. 12. conclusions a new dynamic mathematical model for a threephase structure model has been proposed in this research. the proposed model takes into account the presence of particles contributing to the reaction in the liquid, gel, and solid phases. these phases depend on catalyst feed and volumetric flow rates of vcm. in addition to this, mass and heat transfers between the two phases were found, with the solid and gel phases considered in calculations of both heat and mass transfer. according to observations, this model can be conveniently used as a predictive tool in order to determine and examine the impacts of hydrodynamics, and operating and kinetic parameters on the performance of reactors as well as on the properties of pvc. references ambrozek b, nastaj j, gabrus e (2012), modeling of adsorptive drying of n-pvc. drying technology: an international journal 30: 10. balakrishnan b, kumar ds, yoshida y, jayakrishnan a (2005), chemical modification of pvc using poly ethylene glycol to improve blood compatibility. biomaterials and artificial organs, 18(32): 6335–6342. blazevska-gilev j, spasesk d (2010), formal kinetic analysis of pvc thermal degradation. journal of the university of chemical technology and metallurgy 45(3): 251-254. chattopadhyay s, madras g (2002), degradation of pvc properties. polymer degradation and stability 71: 273–278. chavadej s, phuaphromyod p, gulari e, rangsunvigit p, sreethawong tm (2008), photocatalytic degradation of 2-pvc by using pt/tio2 prepared by microemulsion technique, chemical engineering journal 137(3): 489-495. cohan gf (1975), industrial preparation of polyvinyl chloride. environmental health perspectives 11: 53–57. field ma (1973), predication of optimum temperature profiles for vinyl chloride polymerization using mathematical model. ici internal report. ibrahem sa (2012), chemical reaction engineering and design by using matlab software. saarbrücken, lap lambert academic publishing ag and co. kamo t, yamamoto y, miki k, sato y (2012), conversion of waste pvc to useful chemicals, samplenumber 0 2 4 6 8 10 -15 -10 -5 0 5 10 proposed model zero residual a. s. ibrahim, y. a. ali, h. m. saad and i. h. amur 50 report of national institute for resources and environment. kiparissides c, achilas ds, chatzi e (1994), dynamic simulation of primary particle size distribution in vinyl chloride polymerization. journal of applied polymer science 54: 1423–1438 . lathia jd, el-sherif d, wheatley ma (2004), surface modification of polymeric contrast agents for cancer targeting. pharmaceutical engineering 23(6): 142-143. lăzăroaie c, rusen e, mărculescu b, zecheru t, hubcă g (2010), chemical modification of pvc polymer matrices with special properties. upb science bulletin 72(2): 127–140. maezawa y, tezuka f, inoue y (2004), study and evaluation of kinetic analysis of pvc thermal degradation. polymer degradation and stability 81(2): 187–196. scott ds, czernik sr, radlein ds (1990), thermal analysis of pvc in the environment. energy and fuels 4(4): 407–411. zuga d, cincu c (2006), polymer composites by reclaiming rubber wastes resulting from finishing the rubberized rolls used in printing industry. upb science bulletin, series b, 68(1): 25-30. c:\users\raoh\desktop\paper 7.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 80-88 production of paperboard briquette using waste paper and sawdust j.k. odusote*, a, s.a. onowumab and e.a. fodekeb a* department of materials and metallurgical engineering, university of ilorin, ilorin, nigeria. bdepartment of mechanical engineering, university of ilorin, ilorin, nigeria. received 30 december 2014; accepted 11 november 2015 abstract: nigeria has abundant supplies of biomass resources and agro-forestry residues (waste), including sawdust, whose potentials are yet to be fully utilized for economic advantage. this study was undertaken to investigate the properties of paperboard briquettes produced from a mixture of shredded waste paper (pulp) and sawdust using starch as binder. the paperboard briquettes were produced manually by compressing soaked mixtures of pulp and sawdust in a wooden mould (28 cm x 28 cm x 1 cm) with a compressive load of 26 kg in the mixing ratios (by weight) 100:0, 70:30, 50:50, 70:30, and 0:100 of pulp to sawdust, respectively. the compressed mixtures were sun dried for three days to ensure proper drying and free from moisture. density, compressive and tensile tests were carried out on the test samples. the results showed that mass and density of the paperboard increased with increased amount of the sawdust in the mixture. test results showed that compressive strength decreases as the sawdust content increases. sample a, which is 100% paper, had a compressive strength of 5215 kpa, while sample e, which is 100% sawdust, had a compressive strength of 22.02 kpa. it was also observed that sample b, which is 70% pulp and 30% sawdust, had the highest tensile strength of 0.629 kpa. however, the tensile strength reduces as the sawdust ratio in the mixture increases. the tensile strength of sample e (100% sawdust) could not be determined because the sample failed to bind properly. it was concluded that paperboard briquettes can be produced using paper and sawdust, which are generally considered wastes. keywords: pulp, sawdust, compressive strength, tensile strength, paperboard. 100:0 ٪ ٪ ٪٪ 0.629 * corresponding author’s e-mail: jaminkolawole@gmail.com j.k. odusote, s.a. onowuma and e.a. fodeke 81 1. introduction paper is a thin material produced by pressing together moist fibres, typically cellulose pulp derived from wood, fibrous materials and nonwood fibrous materials, and drying them into flexible sheets (cocca et al. 2011). paper is one of the necessities of civilization and it is almost impossible to imagine the continuance of a world without the printed books and newspapers. though paper, is most commonly used for writing and printing upon, but also found applications include usage as a packaging material, industrial and construction purposes among others (brozek and novakova 2013). these and many more of its uses make it almost impossible to have a halt in its production. in addition, wastes generated through the paper usage are also transformed into other value-added products such as fiber-cement composites (thaemngoen et al. 2003), granular and sheet activated carbon (littrell et al. 2000; uradei et al. 2000; khezami et al. 2007; masahiro et al. 2004; malikov et al. 2007; yorgun et al. 2009), and activated carbon monoliths (kercher and nagle 2003; nakagawa et al. 2007). sawdust or wood dust, which is a by-product of cutting, drilling, sanding, or otherwise pulverizing wood with a saw or other tools, is constantly in production in every workshop where wood work is carried out (rizki et al. 2010). wood dust can also be by-product of certain animals, birds and insects which live in wood, such as the woodpecker and carpenter ant (kuti 2009). sawdust has quite a number of practical uses, including production of particle board, absorbent materials used in laboratories and poultry farms, fuel sources, among others. over the years, several attempts have been made towards reducing, recycling and reusing industrial, commercial, as well as domestic wastes and this has been a subject of research hitherto. the prevalent habit of getting rid of the waste in nigeria is by burning as depicted in fig. 1(a & b), which is hazardous to the environment and human health. sawdust is indiscriminately disposed into the environment in and around the wood processing industries (fig. 1), and is often hazardous to health and safety. sawdust becomes a potential health problem when, for example, the wood particle from processes such as sanding become airborne and inhaled. sawdust is a known human carcinogen (aina et al. 2009). having seen the rate of production and use of paper and sawdust, it has become imperative to develop ways of processing their waste through recycling, reusing and transforming them into value-added products. many products have been developed from waste paper and sawdust over the years (cocca et al. 2013), and these include products like fibre-cement composites, granular and sheet activated carbon, activated carbon monoliths, particle board and paper briquette (tangsathit and sanongrai 2012), sawdust briquette (aina et al. 2009; emerhi 2011). this project work will be considering the production and analyses of paper boards from waste paper and sawdust with varying composition of paper and sawdust. this project work will be of great importance because it will serve as a way of conversion of solid wastes to useful product. the present work is significant because it will solve the problems associated with waste paper disposal in academic environment and convert them into other useful products in the same environment. these products will find application as notice board in offices and around the university community. it can also serve as particle board for demarcation of class rooms and laboratories. the waste paper will be mixed with saw dust which is also readily available from local sawmill and furniture workshops. this will help in reducing atmospheric pollution resulting from burning, which is the common way of getting rid of waste paper and sawdust. another important aspect of this project is that since materials needed are readily available, it will be economical and may serve as a means of livelihood for the unemployed. 2. mater ials and method the materials used for the production of the paper board include waste papers (a4), sawdust (mahogany wood) and starch. the waste papers are shredded manually using razor blade and scissors prior to soaking in water as shown in figs. 2a and 2b. the sawdust was also soaked in water to allow for softening before further processing (fig. 2c). the shredded paper was soaked for 3 hours which is the most appropriate soaking time for paper (tangsathit and sanongraj 2012) prior to pounding to obtain fine pulp. the sawdust was also soaked for about 4 hours in a separate container to ensure an eventual proper and uniform mixing with the pulp. after soaking and pounding, the pulp was collected and drained to remove the water content. the same was also done for the soaked sawdust. production of paperboard briquette using waste paper and sawdust 82 the drained water can be reclaimed for re-usage or used for other purposes such as block making. this was followed by thorough mixing of the pulp and soaked sawdust, in varying proportion by mass as shown in table 1 and 2 with 200 g of starch as the figure 1. (a) burning of wood waste, and (b) disposal of wood waste along drainage. figure 2. (a) shredded paper, (b) soaked shredded paper, and (c) soaked sawdust. binder. the briquettes were prepared by pouring the mixture into moulds of equal size (28 cm x 28 cm x 1 cm), and they were compressed by placing a mass of 26 kg on each of them. the mass placed on the five samples are left for two days after which the boards with the moulds were sun dried for 3 days for complete drying (fig. 3). 3. sample testing the five briquette samples were tested to determine the following parameters: mass, density, percentage moisture loss, tensile stress and rupture force at point of failure. 3.1 mass the mass of each of the briquettes before and after sun drying was determined using a weighing balance. each of the samples was weighed three times for accuracy and the average values were obtained from the readings (fig. 4). 3.2 density the volume of the boards is uniform and was calculated to be 756.25 cm3. hence, the density of each of the briquette was determined as the ratio of mass against volume. j.k. odusote, s.a. onowuma and e.a. fodeke 83 table 1. composition of pulp and sawdust in the paperboard briquette. sample label pulp (wt. %) sawdust (wt. %) a 100 b 70 30 c 50 50 d 30 70 e 100 table 2. mean mass of the wet and dry briquette and the mean moisture loss of different samples after sun drying. sample label mean mass of the briquettes (g) wet (mi) dry (mo) difference (mi mo) moisture loss (%) a (100% pulp) 936.0 ± 2 180.0 ± 1 756.0 ± 1 80.77 ± 0.07 b (70% pulp, 30% sawdust) 1038.0 ± 2 274.0 ± 1 764.0 ± 1 73.60 ± 0.05 c (50% pulp, 50% sawdust) 1122.0 ± 1 275.0 ± 1 847.0 ± 0 75.49 ± 0.07 d (30% pulp, 70% sawdust) 1072.0 ± 2 335.0 ± 1 737.0 ± 1,, 68.75 ± 0.04 e (100% sawdust) 956.0 ± 1 376.3 ± 1.53 579.7 ± 0.58 60.63 ± 0.12 3.3 moisture loss the masses of the briquettes are determined before and after sun drying. moisture content is calculated using eqn. 1. 100 % (1) where is the mass of the paperboard briquette before sun drying, and is the mass after sun drying. 3.4 tensile strength each briquette samples was cut into 3 sections. the length and width of each section is measured. the maximum tensile load of each sample was determined using the centre loading method according to the british standard, bs 4415-1:1992 (british 1992). the samples were incrementally loaded until failure occurred. when the sample test broke, the maximum load used was recorded. the tensile stress of each section is calculated according to eqn. 2 (tangsathit and sanongraj 2012). the average values of the tensile stresses of the samples are determined. tensile stress (pa) = (2) 3.5 compressive test the method used for the present work has been previously reported for testing briquettes made from different non-metallic (novakova and brozek 2011) and metallic materials (plistil et al. 2004). the briquettes were weighed and measured according to the british standard, bs 1so 12192:2002 (british 2002), before placing each of them between the plates of the universal tensile testing machine and continuously loaded until the briquette ruptures as shown in fig. 5. the test is completed when the briquette ruptures which is accompanied with the rapid load decrease. the maximum load prior to rupture is noted from the load indicator (brozek and novakova 2013). 4. results and discussion the photographs of the final dried briquette samples are presented in figs. 6a to 6e, which revealed the color, shape and size of the finished briquette samples. sample a contains 100% pulp, sample b contains 70% pulp and 30% sawdust, while sample c has equal amount of pulp and sawdust. sample d contains 30% pulp with 70% sawdust and sample e is 100% sawdust. 4.1 mass of the briquette samples mass of briquette before sun drying was determined and the values recorded are given in production of paperboard briquette using waste paper and sawdust 84 figure 3. isometric view of the wooden mold. figure 4. measurement of the briquette mass on a weighing balance. table 2. the mass of wet briquette samples were obtained from the difference between the mass of mould + wet briquette and the mass of empty mould. from the table, it was observed that wet sample c (50% pulp, 50% sawdust) has the highest mass followed by the wet sample b (70% pulp, 30% sawdust) and sample d (30% pulp, 70% sawdust). wet sample a (100% pulp) has the least mass slightly below the values of sample e (100% sawdust). the results indicated that the water absorption tendency of both pulp (soaked waste paper) and sawdust are very similar with the sawdust having relatively higher absorption property than the wet waste paper. 4.2 moisture loss after sun drying, the masses of the briquette samples were determined by direct weighing and the results were compared to the mass of the samples before sun drying to determine the moisture loss using eqn. 1. the results are presented in table 2, with the deviation from the mean values. from the table, it is observed that there was reduction in the masses of the samples after sun drying. this is due to the loss in the moisture content after the samples were dried. sample a (100% pulp) has the highest percentage of moisture loss which is 80.77% (fig. 5), indicating poor retention of water as a result of faster rate of drying. this observation may also be due to the low density of the sample a (100% pulp) compared to other briquette samples as revealed in fig. 6. the lower the density of any material, the higher the presence of pores and the more the ability to loss moisture, that is, the less the water retention ability since there will be more spaces for the water/moisture to escape. generally, as the j.k. odusote, s.a. onowuma and e.a. fodeke 85 figure 5. compression test set up. figure 6. pulp and sawdust briquette samples: (a) 100% pulp, (b) 70% pulp and 30% sawdust, (c) 50% pulp and 50% sawdust, (d) 30% pulp and 70% sawdust, and (e) 100% sawdust. composition of pulp in the briquette samples increases, the percentage of moisture loss also increases as shown in table 2. however, sample with equal amount of pulp and sawdust showed slightly higher moisture loss compared with sample with 70% pulp possibly due to homogeneity and better compaction. sample e (100% sawdust) retained more water and thus have lower percentage of moisture loss possibly due to slow rate of drying as shown in fig. 7. this observation agrees with fig. 8, which shows that sample e has the highest density among all the briquette samples. it is evident from the figure that sawdust retains more water than pulp due to better compaction resulting in improved density and lesser pores as compared to pulp. 4.3 density the density of each of the product was determined by dividing the mass of the dry briquette by its volume, which was obtained from the volume of the mould. the densities of each of the samples are presented in fig. 8. it is observed that the density of the briquette samples increases as the composition of sawdust increases. the density of sample a, which is purely made of waste paper (100% pulp), is 238.0 kg/m3, while that of sample e, which is 100% sawdust, is 497.6 kg/m-3. production of paperboard briquette using waste paper and sawdust 86 figure 7. percentage moisture of loss of paperboard samples. figure 8. density of various samples of paperboard. this indicates that wood waste, that is, sawdust is denser than paper. mahogany wood has been reported to have a density of between 495 and 850 kg/m3, agreeing with the result of the present study. from the results, it can be concluded that sawdust imparts higher density to the paperboard compared to waste paper (pulp) due to better resulting from their relatively uniform grain sizes. 4.4 compressive strength the compressive load of each of the samples was determined. samples were made in cube with all the sides being 8cm as shown in fig. 3 prior to testing. the ratio of the load to the surface area of each cube was obtained to determine the compressive strength. the results are given in table 3. it is clear from the table that sample a has the highest compressive strength. the compressive strength of the samples decreases as the composition of sawdust increases, indicating poor compressive strength of sawdust. sample e, which is 100 % sawdust, has the least compressive strength of 22.0 ± 2.0 kpa. this may be due to its higher water retention capability resulting from poor compaction and presence of pores, and thus making the sample more brittle and easily shattered. compressive strength has not been j.k. odusote, s.a. onowuma and e.a. fodeke 87 reported in most of the previous work found in the literature and thus making the comparison of the result in the current study to those in literature not feasible. 4.5 tensile strength the tensile strength of each of the samples was determined using the centre loading system, where the load is uniformly distributed across the entire sample. the load at the point of failure is recorded. the mean tensile stress is then calculated and presented in table 4. the table revealed that sample b, which is 70% pulp and 30% sawdust, has the highest tensile strength of 628.9 ± 0.8 pa. however, the tensile stress reduces as sawdust content in the paper board increases probably due to the poor tensile strength of sawdust resulting from its brittleness. similar observation has not been reported in the literature on waste paper and sawdust briquettes. the tensile stress of sample e (100% sawdust) could not be determined because the board did not bind properly. this shows that the sample is very brittle hence, it could not be loaded. table 3. mean compressive strength of the paperboard briquette samples. sample label mean compressive force (n) mean compressive strength (kpa) a (100% pulp) 33375.0 ± 0 5215.0 ± 0 b (70% pulp, 30% sawdust) 14462.7 ± 222.5 2260.0 ± 35.0 c (50% pulp, 50% sawdust) 11347.7 ± 222.5 1773.0 ± 35.0 d (30% pulp, 70% sawdust) 3782.7 ± 222.5 591.0 ± 35.0 e (100% sawdust) 141.3 ± 12.7 22.0 ± 2.0 table 4. mean tensile strength of the paperboard briquette samples. sample label mean tensile force (n) mean tensile strength (pa) a (100% pulp) 15.3± 0.04 426.7 ± 1.4 b (70% pulp, 30% sawdust) 22.5 ± 0.03 628.9 ± 0.8 c (50% pulp, 50% sawdust) 9.2 ± 0.03 257.0 ± 0.8 d (30% pulp, 70% sawdust) 5.2 ± 0.03 146.3 ± 0.8 e (100% sawdust) 5. conclusion boards were produced from waste paper and sawdust using different mixing ratios and starch as binder. after production, tests were carried out on the boards to determine their properties. from the tests carried out and analyses done, it was observed that sawdust did not retain water as much as paper did. this explains why boards with higher paper content lost more mass than those with higher sawdust content. also, it is observed that sample b which is made of 70% paper and 30% sawdust has the highest tensile strength. this means it will be able to withstand bending to a reasonable extent without failing. it also has a relatively low density which makes it light in weight compared to other samples. thus, 70% paper and 30% sawdust is recommended as the best mix ratio for production of quality and durable paper board. references aina om, adetogun ac, iyiola ka (2009), heat energy from value-added sawdust briquettes of albizia zygia. ethiopian journal of environmental studies and management 2(1): 2009. british standard institution (1992), determination of the tensile properties of paper and board. constant rate of loading method. bs 4415-1:1992, iso 1924-1: 1992. british standard institution (2002), paper and board. compressive strength. ring crush method. bs 1so 12192:2002. brozek m, novakova a (2013), briquettes from recovered paper and board. jelgava 23-24.05.2013. cocca m, d’arienzo l, d’orazio l (2011), effects of different artificial agings on structure and properties of what man paper samples. international scholarly research network isrn materials science 2011: id 863083, 7 pages doi:10.5402/2011/863083. production of paperboard briquette using waste paper and sawdust 88 emerhi ea (2011), physical and combustion properties of briquettes produced from sawdust of three hardwood species and different organic binders. advances in applied science research 2(6): 236-246. kercher aa, nagle dc (2003), monolithic activated carbon sheets from carbonized medium-density fiberboard. carbon 41: 3-13. khezami l, ould-dris a, capart r (2007), activated carbon from thermo-compressed wood and other lignocellosic precursors. bioresources 2(2): 193-209. kuti oa (2009), performance of composite sawdust briquette fuel in a biomass stove under simulated condition. a.u. j.t. 12(4): 284-288. littrell kc, khalili nr, campbell m, sandi g, thiyagarajan p (2000), structure characterization of activated carbon adsorbents prepared from paper mill sluge. chemical material 14(1): 327-333. malikov in, noskova ya, karaseva ms, perederii ma (2007), granulated sorbents from wood waste. chemical solid fuel 41(2): 100-106. masahiro s, iada t, kawarada k, chiba y, mamoto t, okayama t (2004), pore structure and adsorption properties of activated carbon prepared from granular molded waste paper. journal of mater cycles waste 6: 111-118. nakagawa y, molina-sabio m, rodriguez – reinoso f (2007), modification of the porous structure along the preparation of activated carbon monoliths with h3po4 and zncl2. microprocessor mesopor. material 103: 29-34. nováková a, brožek m (2011), briquettes from paper waste. in.: ecology and farming technologies: agro-engineering approaches. saint-petersburg-pavlovsk: russian academy of agricultural sciences et al. 219-225. plíštil d, brožek m, malaták j, heneman p (2004), heating briquettes from energy crops. research in agricultural engineering 50: 136-139. rizki m, tamai y, takashi y, terazawa m (2010), scrutiny on physical properties of sawdust from tropical commercial wood species: effects of different mills and sawdust’s particle size. journal of forestry research 7(1): 20-23. tangsathit s, sanongraj s (2012), technique for production of paperboard briquette from wastepaper. suranaree journal of science and technology 19(2): 79-92. thaemngoen c, sungwornpatansakul w, asavapisit s (2003), the potential for utilization of bagasse and eucalyptus pulp slugs from wastewater treatment plants of the pulp industries to produce fiber-cement composites. suranaree journal of science and technology 11: 125-131. uradei y, technol, taniwatashi r, kubo s (2000), activated carbon sheet prepared from softwood acetic acid lignin. journal of wood science 46: 5258. yorgun s, vural n, demiral h (2009), preparation of high-surface-area activated carbons from paulownia wood by zncl2 activation. microporous and mesoporous materials 122: 189-194. d:\desktop\paper_3.xps tjer 2013, vol. 10, no. 2, 33-45 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks using a genetic algorithm vv juli*a and j rajab *adepartment of electronics and communication engineering, anna university of technology, tamilnadu, india bdepartment of electronics and communication engineering, anna university of technology, tiruchirapalli, tamilnadu, india received 14 february 2012; accepted 4 december 2012 abstract: abstract : wireless sensor networks extend the capability to monitor and control far-flung environments. however, sensor nodes must be deployed appropriately to reach an adequate coverage level for the successful acquisition of data. modern sensing devices are able to move from one place to another for different purposes and constitute the mobile sensor network. this mobile sensor capability could be used to enhance the coverage of the sensor network. since mobile sensor nodes have limited capabilities and power constraints, the algorithms which drive the sensors to optimal locations should extend the coverage. it should also reduce the power needed to move the sensors efficiently. in this paper, a genetic algorithm(ga) based sensor deployment scheme is proposed to maximize network coverage, and the performance was studied with the random deployment using a matlab simulation. keywords: coverage, deployment, network, population size and sensor _____________________________________________ *corresponding author’s e-mail: violetjuli.prince@gmail.com 34 vv juli and j raja 1. introduction a wireless sensor network (wsn) is a type of network in which tiny sensor nodes are spread out into smaller groups around premeditated positions and connected by wireless media. the sensor nodes have the ability to sense the phenomenon, process the data, and communicate with other sensor nodes and are placed one by one, either manually or by robot. the deployment of a large number of sensor nodes needs very attentive handling of topology maintenance. in military applications or when the environment was unknown the sensors are scattered randomly throughout the field, whether on land or water. the sensor nodes could be deployed statically or in a highly mobile environment, but generally the nodes are thrown out en masse from aircraft or ballistic propulsion systems to acquire satisfactory coverage. as a result, the density of distribution is not uniform using this method (chen and sun 2009). of course, the method for initial deployment should reduce the installation cost, eliminate the need for pre-organization and pre-planning, enhance the flexibility of arrangement, and promote self-organization and fault tolerance (akyildiz et al. 2002). failure of a sensing device is a general and standard occurrence due to energy depletion or destruction. invariably, there are changes in topology as well as changes in the nodes' positions, reachability, malfunctioning, and task details that transpire after deployment. those malfunctioning nodes could be replaced by redeploying additional nodes at any course of time, but the addition of new nodes requires a need to reorganize the network. as a result of this, the sensor network topologies undergo frequent and continuous fluctuations after deployment as the deployment of nodes influences the efficiency of the network. a good deployment will improve resource allocation, communication, and information gathering. however, because of inevitable device failure, topology maintenance has turned into a challenging task (chen and sun 2009). coverage is considered one of the basic and fundamental difficulties in sensor networks. getting and providing complete network coverage has been a challenging problem. in order to achieve complete coverage, it has been required to deploy a massive number of sensors in the field. besides the execution of a wide set up, coverage formulations should endeavor to detect weak areas in the monitoring domain and suggest schemes for the future utilization and reconfiguration of sensors to improve the overall coverage of the network. today, sensing devices are freely and easily transported from one place to another, possibly by being attached to robots, vehicles, tanks, animals, and even human beings. despite the fact that mobile devices are of high operational value, they also are associated with high power consumption and increased collisions in wireless transmissions. however, it is possible that mobile sensors can be used as energy sources to facilitate in energy transfers to stationery devices. in a mobile network, sensor nodes or the sink can be moved; in some networks, both can be kept mobile. in this paper, we address the problem of moving the randomly dropped sensors from their initial locations to optimal locations using a genetic algorithm (ga). the main goal of the paper is to better the distribution pattern of the nodes by redeployment of the same from crowded areas to nearby uncovered areas, thereby maximizing coverage with minimal displacement of already active nodes. this paper is organized as follows: the deployment objectives and existing deployment algorithms are presented in section 2. the problems associated with sensor node deployment and desired properties of algorithms is presented in section 3. the proposed algorithms and solutions are described in section 4. the performance evaluation of deployment strategies is narrated in section 5. finally, the conclusion and future works are highlighted in section 6. 2. deployment algorithms 2.1 deployment intentions despite several deployment objectives that can exist when establishing sensor networks (eg. coverage, lifetime, routing, etc), maximizing coverage is an area of primary concern. cardei and wu (2004) identified three types of sensor coverage: area, point, and barrier. gage (1992) sorted the same into barrier, blanket, and sweep coverage. moreover, (huang and tseng 2003) mentioned that coverage was considered a decision problem which identified whether every point in the region was covered by at least some minimum number of sensors. in their study, an algorithm determined the level of coverage in each location, and application areas of the algorithm were also discussed. the same concept was examined by (zhou and gupta 2004). in their study, a minimum size connected kcoverage problem was addressed. the study found that the problem resulted in fault tolerance and energy conservation techniques because only a few selected sensors were made active at a particular time. rabie et al. (2006) addressed the coverage question of how to monitor the chiefly imperative areas in the meadows to which the sensors had been deployed by deploying extremely reliable sensors. maximizing the 35 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... sensor network life was the second most important objective in that study of sensor deployment. zhao et al. (2007) proposed an algorithm in which integer planning was adopted to deploy the relay sensors in static heterogeneous networks. greater energy efficiency was achieved by reducing the path length. the sensors' batteries as well as the quantity of data sent per sensor played a major role in increasing the life of the sensor network. reducing the power needed for routing and communication was another key goal of deployment. as a result, sensor communication range played the most important role in maintaining quality network connectivity. in our study, deployment algorithms were categorized as random, incremental, or movement assisted according to the method of sensor deployment. 2.2 random deployment random deployment was the method used to place the sensor nodes when the environment was unknown or subject to a severe change in conditions. sensors could be "scattered" randomly from a helicopter or thrown using flying robots. clouqueur et al. (2002) used a random deployment method for target detection where the sensors were deployed sequentially. a restricted number of sensors were deployed in each step until the required revealing probability was achieved. a mechanism for sensor collaboration to perform target detection was proposed, with the goal of maximizing exposure of the least exposed path in the region. the cost function used in that study depended on the number of sensors deployed in each step and the cost of each deployment. the cost function was minimized by appropriately choosing the sensors. they found that the optimal number of sensors deployed in each step varied with the relative cost assigned to deployment and sensors. morteza and massoud (2005) considered the problem of energy efficient random deployment of sensor nodes. the objective of the deployment was to find the node density at every point inside the specified deployment region, subject to restriction of the quality of monitoring (qom) and network lifetime. using random methods, the cost of and time needed for deployment were reduced. however, the actual landing positions of the sensors were different from the intented ones due to the presence of wind, moving objects, and other obstacles present in the environment. their finding was that deploying the sensor nodes randomly might not produce a uniform distribution of nodes and could possibly result in an inefficient wsn. as a result, thorough coverage was not guaranteed and possibly would not have been able to meet the application requirements. 2.3 incremental or centralized deployment algorithms in the incremental deployment, the nodes were launched one by one into the target field, with a powerful centralized sink node responsible for node deployment. each node maintained bidirectional communication with the sink node. information about the previously deployed nodes was communicated by the sink node to the sensor nodes and that information was used by the nodes to determine their ideal locations. howard et al. (2002) proposed an incremental deployment algorithm in which the sensor nodes were added one at a time. they discussed four phases of the algorithm; namely, initialization, selection, assignment, and distribution. they also considered the obstacles present in the target area while calculating the coverage probability. the network coverage was maximized with the constraint that nodes should maintain line of sight with their neighboring nodes during the deployment process. wu et al. (2006) proposed delaunay triangulation based on an algorithm called the delaunay triangulation (dt-) score. the goal was to maximize the coverage area of an obstacle-ridden sensor field. in that method, the sensors were added incrementally. the coverage increased when the number of deployed sensor nodes increased. for all random scenarios, the deployment method increased coverage. in the incremental deployments, the nodes were positioned in optimal locations during each step of the deployment process. hence, when the nodes were moved to improve coverage, only a minimum amount of energy was consumed for sensor mobility. however, the deployment time increased. scalability was one of the main problems and, since a large number of sensors was used in the networks, numerous messages were simultaneously reported to the centralized node. generally, in highly dynamic networks, the cost and power consumption of the network increases due to the large number of messages being communicated. hence, such algorithms were prone to single point failure. in such situations, the mobility assisted distributed deployment algorithms would provide better solutions for solving the scalability problem. 2.4 mobility assisted deployment algorithms in previous studies, random and incremental deployment strategies resulted in poor network performance and reduced network lifespan. hence, in this study, the mobile sensor nodes were deployed in the field to improve coverage and other network performances such as energy consumption and network lifetime. the sensors were capable of self-deploying after the initial random deployment and arranging themselves in the monitored field based upon a local 36 vv juli and j raja decision. here the powerful sink node was not involved and the nodes made their own decisions in determining their optimal locations. there were mainly two approaches: a computational geometry-based approach and a virtual force-based approach. 2.4.1 computational geometry-based approach in the geometrical approach, the sensor regions were represented by a set grid or i-polygon. the grids and polygons changed when the sensor nodes were moved. luo et al. (2005) proposed a novel method called the grid method. the deployment area was divided into individual grids. the locations of the predeployed nodes and obstacles were already known and marked by the base station. the total weighted value of each grid was calculated through the consideration of factors such as boundary, pre-deployed nodes, hot zones, and obstacles. the section of the grid with the lowest weighting score was deployed with a new mobile node. coverage and uniformity improved rapidly when mobile nodes were placed in the target field without altering the distribution of previously deployed nodes. the calculations involved were simple. wang et al. (2006) proposed two sets of distributed protocols for controlling the movement of sensors to achieve target coverage: basic and virtual movement protocols. in the basic movement protocols, sensors detected coverage holes using a voronoi diagram. to heal the holes, the nodes were moved. in the virtual movement protocols, sensors did not perform iterative physical movements. instead, after calculating the target locations, sensors moved virtually and exchanged those new virtual locations with the sensors which would have been their neighbors if they had actually moved. the actual movement occurred only when the communication cost to reach their logical neighbors was too high, or when they determine their final destinations. those protocols were effective in terms of coverage, deployment time, and movement. in computational geometry methods, the nodes were uniformly distributed in the sensor field, but the presence of obstacles was not taken into consideration and their geometrical areas were not modeled. in each iteration, since each node was moved after calculating the optimal location, the energy consumption for driving the sensor was high. 2.4.2 virtual force based approach virtual force-based techniques have been used to solve navigation problems in mobile robotics. in these methods, the nodes and obstacles were considered points with attractive and repulsive forces. the resultant force depended on the distance between the neighboring nodes and obstacles. the nodes move from denser areas to the uncovered area similar to the way in which a charged particle would move in an electrostatic field. howard et al. (2002) considered the problem of deploying a mobile sensor network in an environment that may be both hostile and dynamic. they presented a potential field-based approach to deployment in which nodes would be treated as virtual particles subject to virtual forces. these forces repelled the nodes from each other and from the obstacles. hence, the nodes were forced to spread throughout the target field, and the approach was both distributed and scalable. the nodes were quickly spread out to maximize the coverage area of the network. in addition to these repulsive forces, the nodes were subjected to a viscous friction force. this force was used to ensure that the network would eventually reach a state of static equilibrium where all nodes would ultimately come to a complete stop. if something was moved, the network would automatically reconfigure itself for the modified environment before returning once again to a static equilibrium. thus, nodes moved only when it was necessary to do something and saved considerable energy. the algorithm was both robust and highly scalable. sameera and gaurav (2004) studied a sensor deployment strategy that maximized the area coverage of the network with the constraint that each of the nodes had at least k neighbors, where k was a userspecified parameter. they proposed an algorithm based on artificial potential fields which were distributed and scalable. the algorithm did not require a prior map of the environment. it included two different kinds of forces. one was a repulsive force that tried to maximize coverage while the other was an attractive force that imposed a constraint of k degrees. as a result of these forces, a group of nodes placed close together spread out into a network to maximize the coverage while satisfying the constraint of k degrees. the repulsive forces from the obstacles were considered to avoid obstacles during the movement of the nodes. the algorithm was effective in terms of energy and coverage in the presence of obstacles. in virtual force-based methods, the presence of obstacles were also identified and considered during node movement. the degree of coverage could be controlled by the setting threshold on the distance between sensor nodes. but in those methods, a powerful cluster head node was responsible for collecting information and organizing node movement. in many environments, such as disaster areas, the cluster head node might not be present. in such situations, organizing nodes into clusters became difficult and resulted in network partitions. 2.4.3 other approaches for mobility wu et al. (2007a) suggested three sensor relocation 37 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... algorithms to match the mobility degree of sensor nodes. the particle swarm optimization algorithm (psoa), relay shift based algorithm (rsba), and energyefficient fuzzy optimization algorithm (efoa) were proposed to relocate the sensors. in psoa, the mobility of the sensors was not limited. but in rsba and efoa, the sensors' mobility was limited by setting a threshold in mobility so that the energy consumption was reduced. the algorithms improved the coverage. wu et al. (2007b) presented a method for the redeployment of sensor nodes in a hybrid network which consisted of both static and mobile sensor nodes. the sensors were moved without altering the position of static sensors. the direction of node movement was based on the analytical hierarchy process (ahp). four factors, namely coverage hole, hot spot, obstacle, and the node boundary were considered for optimal deployment. when the number of mobile nodes was increased, the coverage also increased since the field became more flexible due to the movement of the sensors. the coverage improved with limited node movement without compromising connectivity. li and lei (2009) presented the improved particle swarm optimization (pso) strategy for optimizing the position of mobile nodes and increasing the whole coverage area. after initial random deployment, the nodes were able to communicate with the cluster head node. the cluster head executed the pso algorithm and managed a one-time movement of sensor nodes to the desired location. the coverage of the sensor was represented by a number of intersection points on the established grid. the coverage ratio was used as the fitness function. the connectivity between nodes was ensured. the algorithm provided better coverage performance than the vfa, and a reduced network. in our work, we aspired to better the distribution pattern of the sensor nodes in the target field. the nodes were moved from the crowded areas to the uncovered areas using a ga so that the coverage was improved. 3. problem specification the deployment of the sensors could not be carried out effectively by humans in various unfavorable working environments, such as remote harsh fields, disaster areas, and toxic urban regions. to accomplish that task, it was suggested to scatter the nodes by aircraft. but while implementing that technique, many obstacles and hindrances such as the presence of wind, trees, and buildings were encountered, and the actual landing positions of the sensors were altered. that improper landing of devices led to poor network coverage. the density of the nodes was not uniform throughout the sensor field. to overcome those difficulties, it was imperative to consider the mobile sensors for the usage because of their capability of moving to the right place for providing the required coverage. while congregating the algorithms, care must be taken not to jar the nodes while moving them. such jarring creates a dead zone. an ultimate solution could be obtained by putting a maximum limit on the algorithm iterations and runtimes. gathering information about the neighboring nodes was the biggest hurdle to overcome. in such work, a sophisticated global positioning system(gps) based network scenario was assumed. the following were some of the desired properties of the proposed solution: * convergence and connectivity: the algorithm should converge and provide an ultimate solution for the specified objective, and the whole network should be connected. * coverage: the final layout of the network should endow the area with the best coverage of the agreed target area with the fewest "dead zones". * minimalism: the algorithm should not pave the way for the nodes to spend excessive time finding their environment and calculating a solution since this causes enormous drain on node energy. * scalability: the algorithm protocol should operate with an arbitrary number of nodes and should properly adjust to work when the nodes are added or removed from the network. * competence: the modus operandi ought to evade loops where nodes incessantly move in a cycle and constantly move between the same positions. in this study, the problem of placing mobile sensors in the premeditated positions to maximize the coverage of the network was explored. the ga-based protocol was proposed for controlling the movement of sensors to improve network coverage. in that protocol, sensors moved iteratively and eventually reached the final destination. in the ga-based method, in each iteration (generation), the fitness of each set of locations of the population was calculated to select the two best solutions (genes) based on the coverage. those two selected genes would undergo crossover and mutation operations to create a new population with a new set of potential solutions. the procedure was repeated for a number (n) of generations or until acquiring an optimum solution (ie. coverage). in this study, a ga-based sensor deployment scheme was implemented and the results were compared with the random deployment scheme using a matlab (mathworks, inc. nattick, massachusetts) simulation. 38 vv juli and j raja 4. ga-based sensor node deployment in this study, an autonomous deployment scheme using gas was described detailing a method to move the sensors from the regions of dense deployment to the areas of sparse deployment requiring enhanced coverage. 4.1 genetic algorithm (ga) ga is used in computing near perfect solutions for optimization. techniques stimulated by evolutionary biology such as inheritance, mutation, selection, and crossover (also called recombination) are used by gas. these types of algorithms are applied as a computer simulation in which a population of conceptual representations (ie. chromosomes, the genotype, or the genome) of candidate solutions (ie. individuals, creatures, or phenotypes) are compared to an optimization problem in a movement toward developing better solutions. binary values as strings of 0s and 1s are usually adopted to represent the solutions but there are also possibilities for other encodings. a population of randomly generated individuals is the initial point for the progression to start and progress in generations. the robustness of every individual in the population is assessed. manifold individuals are stochastically selected from the current population purely based on their fitness, and then are customized to form a new population in each generation, and they are used in the subsequent iteration of the algorithm. at that point, when a suitable fitness level is obtained for the population or when a maximum number of generations have been produced, the algorithm is terminated. 4.2 the proposed ga-based sensor node deployment figure 1 shows the design of the ga-based node deployment algorithm. the two-dimensional sensor field is assumed and hence the z-direction is assumed as zero. in the sensor field, 100 nodes are randomly deployed. the step-by-step procedure is discussed below: step 1: initial population to change the location of a node, one has to specify the displacement as well as the directions of the nodes. so, for coding the displacement and direction, four numbers will be needed-two for the x-direction and two for the y-direction. the four attributes (tx, ty, xd, yd) are coded with binary notation of bit string, where tx and ty are the small displacements in the xand y-directions, xd and yd represent the xand ydirections. since there are n nodes and n sets of displacements, directions are used to handle the movement of all the nodes in the network. if the population size, or future locations for each node, is n, then there should be n x n x 4 numerical values to represent the n number of n sets of node movements. if we use 4 bits for representing the maximum possible movements (15), then two bits will be used to represent the xand y-directions. later, during calculating the fitness of a solution, if the direction bit is 0 then the small distance will be added to the corresponding coordinate of the location of the node. if the direction bit is 1, then the small distance will be subtracted from the corresponding coordinate of the location of the node. as far as the problem is concerned, if we need to optimize the location of only one node then there will not be a change in direction of the other nodes. but in this case, we will be changing the location of 100 nodes. hence, there are many chances for a change in direction of the sensor nodes during each generation because of the change in location of all the neighbors. further, a node cannot go and fill a distant coverage hole. it can be allowed to move only a short distance and fill the coverage hole to minimize the power needed for moving the sensors. that is why the node is limited to 15 movements. step 2: evaluation after creating random locations, the first step is to calculate the fitness value of each location in the population based on the similarity of the previous optimum locations. the process of evaluating the fitness of a location consists of following three steps: * conversion of the chromosome's genotype to its phenotype. that means converting the binary string into corresponding real values. * evaluation of the objective function. the objective function is the sensor network coverage and one has to maximize the coverage in each generation. * conversion of the value of objective function into fitness value. step 3: objective function values and fitness the set of the nodes' new locations are calculated using the current location of the nodes, displacements, and directions given by the ga. the objective function (coverage area) values of f (fitness values) are calculated using the new location of the nodes as follows: 39 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... evaluation is the process which compares the total coverage achieved by n number of nodes when each node is located in any one of the assumed future locations n. if there is n number of nodes, there will be n corresponding locations. for each node, we consider n possible future locations. out of n, we consider the first best two locations, which contribute to maximizing the coverage achieved by n number of nodes in the target field. the best set of new locations as well as their previous displacements and directions would be preserved for further processing. step 4: creation of a new population after evaluation, a new population from the current generation is created using the previous set of displacements and directions of the best two solutions to generate the new population. the three operations, reproduction, crossover, and mutation, are used. the crossover and mutation rates should be selected in accordance with the convergence factor. otherwise, the algorithm results in a primitive random search. the figure 1. ga model 40 vv juli and j raja population size is fixed with respect to the convergence factors. for that, the previously selected best displacements and directions are also included in the new population. reproduction: for each node, from the newly found locations in the population of a generation, the locations with the best fitness and the second best fitness are allowed to exist and are selected as new parents to produce the next generation. the best set of new locations as well as their previous displacements and directions would be preserved for further processing. crossover: the crossover used here is the one-cutpoint method, which randomly selects one cut-point and exchanges the right parts of two parents to generate offspring. the crossover point would be chosen selectively with respect to the convergence factors. mutation: after crossover, the mutation is carried out. in accordance with the convergence factors, the mutation level is selected. one or more genes (direction bits) are modified by mutation with a probability equivalent to the mutation rate, and a sequence of random numbers rk is generated. (the number of bits in the whole population) if ri is 1, change the ith bit in the whole population from 1 to 0 or from 0 to 1. the chromosomes (directions) reproduced are not subject to mutation, so after mutation, they should be restored. after the finishing point of a single iteration of the ga, a new population is created. then the entire process is repeated for as many times (eg. ten) as preferred until the best value of the objective function in each generation is evaluated. ultimately, the two best sets of optimum sensor locations based on the best fitness are reached. 5. results and discussions the random and proposed ga-based sensor network deployment scenarios were simulated using matlab and the performance of deployment algorithms were analyzed in terms of displacement of nodes, cumulative displacement and coverage of network. the following network is considered: sensor network size : 600 m x 600 m total number of nodes : 100 sensing range of the sensors : 50 m initially the sensors were randomly dropped in a 600 meter x 600 meter field. figure 2 shows the randomly dropped sensor in a field. the blue color points represent the randomly dropped sensors in the field. the green circles represent the sensing range of the sensors. one can see that the nodes are densely deployed in some areas and sparsely deployed in some other areas. hence the nodes are not uniformly distributed. in figure 3, the rectangles represent the areas where nodes are densely deployed and the ellipses represent the areas which are uncovered by the sensor nodes. in figure 4, the green area represents the covered area by the randomly dropped sensors and the length of the field in meters figure 2. randomly dropped sensors length of the field in meters figure 3. field with densely and sparsely deployed sorsors length of the field in meters figure 4. covered and uncovered regions of sensor field 41 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... white area represents the uncovered areas. during each iteration of ga, the system was programed to discover better locations. to achieve the results in table 1, the following ga parameters were used: total population size : 200 total number of generations : 10 mutation level : 0.20 crossover rate : 0.20 figure 5 shows the sensor field after the tenth iteration of the ga-based sensor node deployment method. one can clearly see that the nodes had moved from the denser areas to the uncovered areas. the covered and uncovered regions after optimization were shown in fig. 6. figure 7 shows the variations in node displacement during each iteration of the algorithm. the displacement of nodes decreased when the iteration was increased. in the initial little iteration, the displacement was increased. figure 8 represents the cumulatable 1. results of ga based sensor deployment length of the field in meters figure 5. sensor field after optimization length of the field in meters figure 6. covered and uncovered regions after optimization 42 vv juli and j raja tive total displacement of sensors in meters. after each iteration, we calculated the area covered by the 100 sensor nodes. the coverage was described by the ratio of the union of covered area of each node in the area of interest. if the node was located well inside the target field, the complete area of that particular node was included in computation of coverage. if the node was located near the boundary regions of the field, then only the partial area covered by the node was included in computation. figure 9 shows the variation of the percentage of area covered by the sensors with respect to the iteration. until the seventh iteration there was significant improvement in coverage. we analyzed the improvement in coverage of sensor network achieved iteration by iteration. the basis for the first iteration was the coverage achieved in random deployment. the basis for the second iteration was the results of first iteration ie. coverage achieved in first iteration. similarly the basis for the third iteration was the results of second iteration. thus all the iterations carried out progressively with the results of previous iteration for the subsequent iteration. for each iteration, we calculated the coverage improvement index using in = an + 1 / an where in is the coverage improvement index for the nth iteration and an is the coverage area achieved in the nth iteration. (n = 1, 2, 3 …10 as the iteration number). the coverage performance improvement index is tabulated in table 1 and the variation in improvement index is graphically represented in fig.10. in iteration 1, the index was initially maximized and slowly reduced when the iteration was increased. it means that the algorithm gives optimal results in the initial few iterations itself. the distribution of nodes was better; thus, the coverage improved. figure 11 represents the comparison of maximum coverage achieved by the random and gabased sensor node deployments, respectively. the coverage 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 1 2 3 4 5 6 7 8 9 10 iteration no. figure 7. iteration vs. displacement of nodes 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 x 104 1 2 3 4 5 6 7 8 9 10 iteration no. figure 8. iteration vs. cumulative displacement of nodes 98 96 94 92 90 88 86 1 2 3 4 5 6 7 8 9 10 iteration no. figure 9. iteration vs. percentage of covered area 1.035 1.03 1.025 1.02 1.015 1.01 1.005 1 0.995 1 2 3 4 5 6 7 8 9 10 iteration in no. figure 10. iteration vs. coverage improvement index 43 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... improved from 84.55% to 96.57% through ga-based node deployment. we obtained the results from a single run but confirmed the results by running the simulation many times. we also used the algorithm to various random topologies. in all topologies, we deployed 100 nodes randomly, applied the ga, and analyzed the coverage. the results are tabulated in table 3. the comparison of coverage achieved by the random and ga-based deployments after 10 iterations for 10 random topologies is shown in fig. 12. the ga-based algorithm gave better results for all topologies studied. thus, the topology could be selected as per the coverage requirement. 6. conclusions the study of ga-based sensor node self-deployment by using the matlab simulation technique improved the effectiveness of the coverage of the sensor network in comparison with that of random deployment. the conclusions of the research study were summed up as follows: 1. the coverage improved through ga-based deployment. 2. the redeployment of the sensors was accomplished using nodes from the nearby crowded areas only; hence, the power consumption for sensor movement was less than what it would be if moving them from far off places. 3. this energy saving would, in turn, prolong the life of the battery and consequently enhance the life of the network. 4. the combination of points 1 through 3 would result in greater cost effectiveness for the entire operation of sensor deployment and data collection. as an extension of this study, hybrid sensor deployment algorithms could be incorporated for furthering the already achieved benefits of the application of ga for the deployment of sensors. figure 11. coverage area table 3. results for various random topologies 98 96 94 92 90 88 86 84 82 80 78 84.55 96.57 random ga random ga deployment method 44 vv juli and j raja references akyildiz if, su wl, sankarasubramaniam y, cayirci (2002), a survey on wireless sensor networks. ieee communication magazine 40(8):102-114. cardei m, wu j (2004), coverage in wireless sensor networks. in handbook of sensor networks, crc press, 0-8493:1968-4. chen jm, sun yx (2009), the deployment algorithm in wireless sensor network: a survey. information technology journal 8:293-301. clouqueur t, phipatanasuphorn v, ramanathan p, saluja k (2002), sensor deployment strategy for target detection. proceedings of the 1st acm international workshop on wireless sensor networks and applications 42-48. gage dw (1992), command control for many robot systems. proceedings of the 19th annual technical symposium auvs. pages 22-24. reprinted in unmanned systems magazine 10:2834. howard a, mataric mj, sukhatme gv (2002), an incremental self-deployment algorithm for mobile sensor networks. autonomous robots special issue on intelligent embedded systems 13:113126. huang cf, tseng yc (2003), the coverage problem in a wireless sensor network. mobile networks and applications 10:519-528. li zm, lei l (2009), sensor node deployment in wireless sensor networks based on improved particle swarm optimization. proceedings of the international conference on applied superconductivity and electromagnetic devices 215-217. luo rc, tu lc, chen o (2005), auto-deployment of mobile nodes in wireless sensor networks using grid method. ieee 359-364. mortesa m, massoud p (2005), qom and lifetimeconstrained random deployment of sensor networks for minimum energy consumption. ieee 293-300. rabie r, al-nawaiseh a, el-rewini h, khaled a (2006), parallel meta-heuristic approaches for deployment of heterogonous sensing devices. pdcs, accepted at the isca 19th international conference on parallel and distributed computing systems. sameera p, gaurav ss (2004), constrained coverage for mobile sensor networks. proceedings of the ieee international conference on robotics and automation 165-172. wang gl, cao gh, la porta tf (2006), movementassisted sensor deployment. ieee transactions on mobile computing 5:640-652. random deployment ga based development 1 2 3 4 5 6 7 8 9 10 topology number 100 95 90 85 80 75 70 figure 12. comparison of coverage for different topologies 45 mobility assisted sensor node self-deployment for maximizing the coverage of wireless sensor networks ... zhao c, yu z, chen p (2007), optimal deployment of nodes based on genetic algorithm in heteroge neous sensor networks. ieee 2745:4. zhou zh, das gupta h (2004), connected k-coverage problem in sensor networks. proceedings of the 13th international conference on computer communication and networks (ic3n). wu ch, chung l, chung yc (2006), a delaunay triangulation based for wireless sensor networks. wu xl, cho js, d'auriol bj, lee sy, hee yy (2007a), self-deployment of mobile nodes in hybrid sensor networks by ahp. lecture notes in computer science 4611:663-673. wu xl, cho js, d'auriol bj, lee sy (2007b), mobility assisted relocation for self-deployment in wireless sensor networks. the institute of electronics information and communication engineers (ieice), transaction on communication e90-b(8):2056-2069. microsoft word paper 3.docx the journal of engineering research, vol. 11, no. 2 (2014) 27-38 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing h.i. al-abdul wahhab* and m.a. dalhat civil & environmental engineering department, king fahd university of petroleum and minerals, dhahran 31261, saudi arabia. received 4 july 2012; accepted 5 may 2014 abstract: recycling as an economic disposal process for many hazardous waste materials has become a popular means of conserving our planet’s scarce and diminishing natural resources. this paper is a study of the influence of oil sludge (os) on the physical behavior and performance of asphalt filled with heavy oil fly ash (hofa), cement kiln dust (ckd) and limestone dust (lmd). conventional asphalt consistency tests in addition to a new bond strength (bs) test were conducted on the modified asphalt mastics. the results were statistically analyzed and assessed in accordance with american society for testing and materials (astm) d 332 and astm d 449 specifications. too much os resulted in strength deterioration of the asphalt mastic, which can be compensated for by filling the mastic with hofa. os interacts constructively with the fillers to improve their effectiveness in raising the softening point (sp) and viscosity of the asphalt, and also in reducing its penetration and ductility. even though sludge mastics hold promise as suitable composites for damp proofing and waterproofing, the resulting low flash point (fp) and sp of some of these mastics make their suitability for roofing applications questionable. keywords: oil sludge, oil fly ash, asphalt mastic, bond strength, waterproofing, roofing. א אאאאאא א  אאא  אwאאאאאאאא א אkאאאאאאאא אאא،אאאאאאאא אאkאאאastmd 332  astm d 449kאאאאא hofaאאאאאk  kא  אא    א  א     א א  אאאאאאאא. אw אא،אא،אאא،א،א. *corresponding author’s e-mail: hawahab@kfupm.edu.sa 28 h.i. al-abdul wahhab and m.a. dalhat 1. introduction the rapid growth in the asphalt waterproofing material market is creating more demand for asphalt. a 4.1% continuous annual increase in asphalt consumption is forecasted for the coming years (world asphalt 2012). nearly 15% of the global asphalt demand is generated from the roofing sector, and this demand rises annually (world asphalt 2010). high demand, coupled with a limited supply of asphalt, has resulted in a continuous rise in cost. this fact, together with the need to conserve our scarce natural resources, calls for the exploration of a cheap local material supplement. in kingdom of saudi arabia (ksa), more than 35,000 m3 of tank bottom oil sludge (os) is produced from crude oil storage tanks on a yearly basis. additionally, about 340,000 m3 of heavy oil fly ash (hofa) waste is generated annually, along with more than 12 megatons of combined cement kiln dust (ckd) and limestone dust (lmd). air-blown asphalt is the main conventional bituminous component of most asphalt-based roofing and waterproofing products. but as roofing chemistry has become more sophisticated, various formulations with different viscosity ranges and physical and mechanical properties (for horizontal and vertical applications) have been developed for specific uses from regular asphalt/bitumen through chemical and mineralogical content modifications such as pourable sealers (pitch pocket mastics) and elastomeric sealants (mastics for high movement joints and terminations). this is due to the fact that the improvements achieved on the bituminous materials’ durability and extensibility (especially at lower temperatures) by adding polymers and other additives will result in a modified material with superior physical properties, which surpasses alternative materials but costs the same amount (paul 2002). many published works have explored roofing asphalt polymer modifications (lucke 1989; terry et al. 1999; singh et al. 2002; martinalfonso et al. 2008 a and b; fang et al. 2009), and sufficient insight into the subject has been established. but besides polymers, mineral fillers are another major additive that can also dictate the behavior of waterproofing asphalt mastic. studies have been carried out to investigate the effect of various mineral fillers on certain asphalt binders’ properties (johansson and isacsson 1998; recasens et al. 2005; chen et al. 2008), but most of these works were directed towards the behavior and performance of asphalt concrete (ac) pavement and not roofing or waterproofing applications. os is waste that accumulates at the bottom of crude oil storage tanks or separation vessels. it was previously used as an asphalt extender by combining comminuted refinery os with liquid asphalt that is either hot asphalt or an asphalt emulsion (eric 2006). the resulting composition is similar in performance to trinidad lake asphalt and can be used as a binder for asphalt-aggregate compositions. some of the mastic formulation was found to comply with astm d 1227 type iii roof coatings’ specifications. the tank bottom sludge was also utilized as a sole binder for a moderate traffic paved and unpaved road mix design (taha et al. 2007). hot mixing yields a concrete with about 12 kn stability, satisfying the requirements for low (3.3 kn) or medium (5.3 kn) trafficked surfaces or base layers according to asphalt institute specifications. taha et al. (2001) examined the environmental and structural impact of os as a substitute in asphalt paving mixtures. experimental results indicated that mixtures containing up to 22% os could meet the necessary criteria for a specific ac-wearing course or bituminous base course. results from the chemical analysis of leachates from a 50% os-containing mixture indicated an insignificant concentration level of toxic metals. in another environmentally-friendly disposal technique study of tank bottom os, al-futaisi et al. found that the toxicity characteristics leaching procedure (tclp) results show that no extracts exceeded the established tclp maximum limits set by the us environmental protection agency (epa). the naturally occurring radioactive minerals (norm) activity values obtained from sludge road mix and solidified sludge mixtures were found to be very low when compared to norm standard values of 100 bq/g; thus, it was concluded that the os for this application 29 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing was non-hazardous. in previous similar research by veenstra et al. (2000), the findings were similar: none of the road base samples was determined to be hazardous according to tclp analyses, but leachates from all the samples were toxic according to the 48-hour acute toxicity test. however, none of the samples had unacceptable levels of norm. there was no disagreement whatsoever among these studies, but the characteristics of the sludge from different oil sources differed in their constituent proportions. heavy oil fly ash (hofa) is a byproduct of the heavy fuel (diesel, cracked fuel, etc.) combustion process. hofa is recycled for a wide variety of purposes, and its potential as a filler reinforcement in low density polyethylene (ldpe) polymer composites has been reported by khan et al. (2011). their results show enhancement in rheological properties of the modified ldpe. the use of hofa to improve the performance of an asphalt binder and concrete was patented by al-methel et al. (2011). hofa-asphalt mixes satisfy the asphalt performance grade limits with potential improvement in durability and strength. traditional asphalt physical tests such as the ductility test (astm d 113), penetration test (astm d 5), sp test (astm d 36), and flash and fire point test (astm d 92) still form the basis for general specifications in, among others, roofing asphalt cements, asphalt for waterproofing and damp proofing applications (astm d 312 2006; astm d 449 2008). there is a lack of adequate standard testing to assess and characterize appropriately the asphalt mastic on its own at the semi-finished product level for this application (goikoetxeaundia et al. 2007). in this study, a new mastic bond strength (bs) test for assessing the performance of asphalt mastic was devised. the objective of this research was to establish the effect of os on the physical behavior and tensile bond strength (bs) of roofing asphalt filled with mineral fillers. both the influence of os on the sp, ductility, penetration, flash point and viscosity of asphalt filled with hofa, ckd and lmd and the tensile bs of the resulting asphalt mastic were examined. 2. materials and methods 2.1 materials the asphalt used in this study, obtained from riyadh’s refinery, is the only locally available grade. it is a 60/70 penetration grade asphalt, with a 64-10 performance grade. the physical properties of the asphalt are shown in table 1. table 1. asphalt physical properties. property magnitude ductility (cm) 150 + penetration (dmm) 67.2 softening point (°c) 52 flash point (°c) 342 viscosity (cp) 575 the os was obtained from rastanura crude storage tanks of the saudi arabian oil company (aramco). the metal and sediment composition of the sludge is given in table 2. the os contained about 70% water, 22% sediments and 8% hydrocarbon oils. water was separated by the centrifuge method (astm d 4007). the dehydrated sludge was subjected to analyses for sulfur (astm d 4294), metals (sulfur, nickel and vanadium; astm d 6481) and sediments (calcium oxide [cao], iron rust [fe2o3] and copper oxide [cuo]; astm d 6481). the dehydrated os was also extracted with carbon disulfide (cs2) for oil fraction and subjected to high temperature distillation (sim; astm d 6352) to obtain the fuel oil fractions. table 2. metal and sediment composition of oil-sludge. contents method unit amount sulfur astm d4294 wt. % 1.4 nickel astm d6481 ppm 5694 vanadium astm d6481 ppm 38 cao astm d6481 wt. % 11 fe2o3 astm d6481 wt. % 4.9 cuo astm d6481 wt. % 5.73 30 h.i. al-abdul wahhab and m.a. dalhat cement kiln dust (ckd) and limestone dust (lmd), which are by-products of limestone quarrying and cement manufacturing, were obtained from a local road construction company. hofa was collected from rabiq thermal power station located on the western cost of the ksa, along the red sea. table 3. oil-sludge-filler mastic experimental design. generalized mastic formulation a binder type % filler (hofa/ckd/lmd) 0 % 10% 15% 20% 25% 0% sludge plain asphalt 200g _ asphalt + 20g _ filler 200g _ asphalt + 30g _ filler 200g _ asphalt + 40g _ filler 200g _ asphalt + 50g _ filler 20% sludge 160g _ asphalt + 40g _ sludge + 0g _ filler 160g _ asphalt + 40g _ sludge + 20g _ filler 160g _ asphalt + 40g _ sludge + 30g _ filler 160g _ asphalt + 40g _ sludge + 40g _ filler 160g _ asphalt + 40g _ sludge + 50g _ filler 40% sludge 120g _ asphalt + 80g _ sludge + 0g _ filler 120g _ asphalt + 80g _ sludge + 20g _ filler 120g _ asphalt + 80g _ sludge + 30g _ filler 120g _ asphalt + 80g _ sludge + 40g _ filler 120g _ asphalt + 80g _ sludge + 50g _ filler a note that, the sludge composition is by asphalt weight, and the filler composition is by asphalt weight in the case of the plain asphalt or by sludge modified asphalt weight in the case of sludge-filler-asphalt blends. 2.2 sample preparation the raw os was first dried in an oven at 70 °c for 48 hours. this temperature was selected to prevent the volatilization of the sludge’s oily constituents; hence, the required slow water evaporation. the filler materials were also placed in an oven for 24 hours at 100–105 °c prior to mixing in order to eliminate the absorbed moisture. neat asphalt (800 g) was poured into a 1000 ml mixing can, which was then placed in an oil bath at 140 °c. the asphalt was continuously stirred with the aid of a high speed shear mixer (1500 rpm) until thermal equilibrium between the bath and the container had been established. an appropriate amount of os (20% and 40%), determined by the asphalt’s weight, was introduced into the asphalt, and the stirring continued for 10 minutes. afterwards, 200 g of the final mix was poured into four new mixing cans and stored temporarily inside the oven at 140 °c. the four blends were then mixed in the same manner as previously mentioned with appropriate filler content (10%, 15%, 20% and 25%) for five minutes. test samples were cast immediately for each blend to avoid a prolonged storage additive settlement or separation. the generalized mastic composition matrix is shown in table 3. the styrene butadiene styrene (sbs) polymer was mixed for 20 minutes at 200 °c, subsequent to the warm soaking of the sbs powder in the asphalt at 160 °c for two hours. 2.3 properties measurement conventional asphalt physical tests such as the ductility test (astm d 113), penetration test (astm d 5), sp test (astm d 36), flash and fire point test (astm d 92) and an additional viscosity test (astm d 4402) were first conducted on the samples. then, selective blends were evaluated using the bs test, which was developed in this study to assess the tensile bs of asphalt mastics to smooth surfaces. 2.4 bond strength test the bs test was devised to measure the asphalt mastic ability to resist tensile force and to transfer a sizable amount of force between two bonded surfaces. it is also an indicator of the maximum bs the mastic can surmount when subjected to tension. the test involves loading a 30 mm x 20 mm x 6 mm prepared sample of the asphalt mastic in tension at a rate of 1.3 mm/minute at 25 °c (astm d 545-08, 2008). the load magnitude and its corresponding deformation are measured in the process. the tensile strength is reported as the maximum stress recorded, which is 31 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing obtained by dividing the highest load carried by the sample before it fails by the plate area. the current active specification test method for asphalt base expansion joint filler (astm d 545-08: standard test methods for preformed expansion joint filler for concrete construction) prescribed the means of assessing the suitabili ty of joint sealant performance and durability through various tests, including the failure due to compression test. however, it failed to include the tensile failure test. the fact is that joint openings are bound to widen during winter just as they are likely to narrow in the summer season, since concrete does not only expand but also contracts; therefore; an additional tensile test is relevant. another major common defect exhibited by an asphalt waterproofing membrane used in a flat roofing system is the formation of blisters (finch et al. 2010). blisters propagate more easily after the initial bond loss between the membrane and the roofing deck if the membrane is lacking in tensile strength. a membrane with low tensile strength that unsticks will certainly swell and bulge at the slightest opportunity and, as a result, facilitates failure of the whole process. 2.4.1 apparatus the apparatus consisted of two 30 mm x 20 mm x 6 mm plates. there was a mechanism that held and stretched the sample while the load was applied [figure 1] and a hydraulic, or screwed-up, device. a sample grip having a wedge-like edge slot matching the size and shape of the plate was fixed to the load mechanism’s main frame upper block with the aid of a short steel rod that was supplemented with a spring bearing to help eliminate any unnecessary compressive force while the sample was being inserted. the upper part of the mechanism rested on a bearing or spring suspension system which eliminated any additional load on the tested sample due to the self-weight of the upper frame. 2.4.2 sample casting the two plates were preheated, spaced 6 mm apart and fixed in position with the aid of figure 1. bond strength test setup. a holder; three sides of the arrangement were wrapped with non-sticking paper. the mastic was heated to a workable state capable of filling the 6 mm × 20 mm × 30 mm space without voids and also to promote sticking to the plate wall with full strength. the material was allowed to cool sufficiently before unwrapping it (≥15 minutes). when necessary, the sample was put in a freezer after 30 minutes for 5 minutes to enable the smooth removal of the non-stick paper without disturbing the sample. then the sample was put in a 25 °c water bathe for at least 90 minutes before testing. 3. results and discussion 3.1 preliminaries the os increased the penetration value of the asphalt at doses above 10%. an average uniform increase of about 20 dmm for every 20% increase in os could be observed (fig. 2), while when the content was below 10%, the penetration tended to decline. this can be attributed to the critical filler/additive content that will reverse the original asphalt matrix, when the filler particles interlock within the asphalt network (asphalt continuous phase), to asphalt patches dispersed within the mesh of the os additive. asphalt’s sp drops with an increase in the amount of os for quantities above 15% by weight of the asphalt, and rises below 10% content (fig. 2). the os acts as a filler when the content is below 10% and as a flux when it is 32 h.i. al-abdul wahhab and m.a. dalhat oil sludge 0% 10% 20% 30% 40% p en et ra tio n, d m m @ 2 5o c 50 60 70 80 90 100 110 120 penetration s of te ni ng p oi nt , o c 46 48 50 52 54 56 softening point figure 2. penetration and softening point vs. oil-sludge. above 10%. this can also be confirmed from the viscosity plot of the asphalt (fig. 3). initially, the solid constituents of the os, which function as fillers, dictate the os’s role at lower quantities, but as the percentage of the os increases, its waxy component reaches significant levels and takes charge of the os in a dominant role. at sufficient levels, the dissolved wax from os lubricates the composite interlayer surfaces, causing a drop in viscosity. thus, 20% and 40% os contents were selected for its role as an asphaltextending material. oil sludge 0% 10% 20% 30% 40% v is co si ty , c p @ 1 35 o c & 2 0r pm 520 540 560 580 600 620 640 660 680 700 sludge vs viscosity figure 3. viscosity vs. oil-sludge. 3.2 softening point (sp) a measure of the temperature at which the material will begin to flow (or soften), the sp gives an upper temperature range within which the asphalt mastic can function stably. hofa raised the sp value appreciably but became less effective when combined with sludge when the hofa content was 15% or below (fig. 4). however, there was a constructive interaction between the two substances when the hofa content was 20% or above, which yielded higher sp values than those of the hofa-only blends. each 20% increase in os caused a decrease of about 3 °c in the sp value for sludge-only blends, but combining 25% hofa with 40% os yielded a material composite having about a 60% higher sp than the neat asphalt. the ckd and os demonstrated a similar constructive interaction when combined within the asphalt medium, with a positive effect on the sp-value, but it was not as significant as in the case of hofa-os blends. figure 4 shows that there was just a 7 °c increase in sp for 25% ckd and 40% os blend compared to pure asphalt. figure 4. softening point of the oil-sludge asphalt mastics. there seems to be little or no interaction between the lmd and os for the given content ranges. lower sp values were observed for higher os-content blends containing the same amount of lmd than blends having no os (fig. 4). the 25% lmd and 40% os blend had a lower sp relative to the neat asphalt. 3.3 viscosity hofa has an exponential positive effect on the asphalt viscosity, which tended to increase with more os for doses below 10%, after which it started declining until it fell a bit fillers (hofa, ckd and lmd) 0% 5% 10% 15% 20% 25% s of te ni ng p oi nt , o c 40 50 60 70 80 90 ckd 0% sludge ckd 20% sludge ckd 40% sludge lmd 0% sludge lmd 20% sulfur lmd 40% sludge hofa 0% sludge hofa 20% sludge hofa 40% sludge 33 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing below the original asphalt viscosity (by 37.5 centi-poise) at a 40% os content (fig. 3). however, the resulting viscosities of the hofa-os blends were always higher than that of hofa-only mastics (fig. 5). this signifies a constructive interaction between the two additives. fillers (hofa, ckd and lmd) 0% 5% 10% 15% 20% 25% v is co si ty , c p @ 1 35 o c , 2 0r pm 0 2000 4000 6000 8000 10000 12000 14000 ckd 0% sludge ckd 20% sludge ckd 40% sludge lmd 0% sludge lmd 20% sludge lmd 40% sludge hofa 0% sludge hofa 20% sludge hofa 40% sludge figure 5. softening point of the oil-sludge asphalt mastics. 3.4 ductility ductility is a measure of the ease with which the material can be deformed plastically at room temperature. os had a negative effect on the ductility of the asphalt material. the initial level of os (20%) caused a tremendous ductility loss (~75%) when compared to the fresh asphalt, after which the rate of decrease slowed considerably (fig. 6). blends containing both os and hofa exhibited a normal interaction between the two additives. the ductility decreased uniformly with the addition of os and hofa. fillers (hofa, ckd and lmd) 0% 5% 10% 15% 20% 25% d uc til ity , c m @ 2 5o c 0 20 40 60 80 100 120 140 160 ckd 0% sludge ckd 20% sludge ckd 40% sludge lmd 0% sludge lmd 20% sulfur lmd 40% sludge hofa 0% sludge hofa 20% sludge hofa 40% sludge figure 6. ductility of the oil-sludge asphalt. normal interaction could also be observed for blends containing both ckd and os additives (fig. 6). the ductility decreased with an increase in both ckd and os when used individually and in combination. this trend was maintained for the hybrid blends, with 20% and 40% os curves showing a uniform transition rate in ductility. the ductility also decreased with more lmd for lmd-only blends, with an almost linear uniform declining trend. but when combined with os, the lmd had little effect on ductility in the sense that the ductility value of lmd-os blends was more or less equal to that of the os-only composite. this can be seen from the almost horizontal 40% and 20% os curves of fig. 6. 3.5 penetration penetration is a measure of the material hardness at room temperature. the individual trend of both os and hofa was maintained even in composites containing both additives, but the decrease in penetration with increasing hofa was more significant in blends containing high os (40%) than those containing 20% os (fig. 7). in (fig. 7), an abnormal rise in penetration value with increasing ckd content can be observed for a 20% sludge curve. instead of declining with more filler content as usual, it turns out that the penetration keeps rising with the addition of up to 20% ckd and then drops slightly at 25%. one possible explanation for this different trait could be the relatively higher particle size of ckd compared to the other fillers. this would result in uneven and sparsely distributed filler-grains which produce a weaker ckd-asphalt-sludge monolith with weaker asphalt-sludge threedimensional (3d) spots at a lower ckd content. when the penetration needle was released, it passed through these weak spots and easily pushed down any ckd particle blocking its path. so, even when the ckd quantity increased, the result was a weaker adhesion of the asphalt-os fluid to the more numerous ckd grains. however, as these fines increased further, their downward displacement by the needle tends to slow down, thus resulting in a relatively lesser penetration value (25% ckd). but in cases of higher sludge content (40% sludge), the 3d matrix is more stably compact, with large amounts of sludge solid fines. this resulted in a continuous decrease in downward and 34 h.i. al-abdul wahhab and m.a. dalhat lateral displacement of the ckd grains as they became situated in a highly filled asphalt matrix. similar but not obvious behavior can be observed in the lmd-os blend (fig. 7), but due to the finer lmd particles, the trend reversed immediately after the lmd content reached 10%. lmd tended to decline in the penetration reading when added to the sludgeasphalt composite, even though the effect was very slight. figure 7. penetration of the oil-sludge aspha lt mastics. 3.6 flash point (fp) flash point (fp) is a measure of the minimum temperature at which material will ignite. it serves as an indicator for fire risk assessment in terms of safety when handling and in service. the os had a decreasing effect on the fp of the asphalt which is due to the presence of some volatile flammable hydrocarbons in the waxy constituent of the os. the 20% content resulted in a decline of about 80 °c, but this effect became less pronounced with the addition of os (fig. 8). as for the os-hofa composites, the fp remained almost unaffected with the addition of hofa compared to the os-only mastic. figure 8. flash point of the oil-sludge asphalt mastics. the ckd tended to affect the fp more significantly in the os-asphalt medium as compared to hofa. an additional decrease of about 30 °c was observed for 20% and 40% os curves at 10% and 25% ckd contents, respectively (fig. 8). all in all, there seemed to be little interaction between the substances. adding lmd to an os-asphalt blend did not seem to change its original fp value by more than 15 °c in a mixture with 25% content (fig. 8). the 20% and 40% os curves are almost straight lines. 3.7 tensile bond strength (bs) the tensile bond strength (bs) increased with os initially, and then started to decline at higher content levels (fig. 9). when the composition contained only a small amount, os particles served as a filler reinforcement engulfed within the sufficiently thick asphalt network surrounding them. but as the sludge’s solid particles increased in quantity along with its wax, the asphalt network responsible for binding them thinned out. this was also accompanied by reduced cohesion, which gradually led to phase inversion and a continuous decline of the asphalt’s strength. the neat asphalt’s bs was slightly above 25 kn/m2, and adding filler to the asphalt fillers (hofa, ckd and lmd) 0% 5% 10% 15% 20% 25% p en et ra tio n, d m m @ 2 5o c 0 20 40 60 80 100 120 140 ckd 0% sludge ckd 20% sludge ckd 40% sludge lmd 0% sludge lmd 20% sludge lmd 40% sludge hofa 0% sludge hofa 20% sludge hofa 40% sludge fillers (hofa, ckd and lmd) 0% 5% 10% 15% 20% 25% f la sh p oi nt , o c 180 200 220 240 260 280 300 320 340 360 ckd 0% sludge ckd 20% sludge ckd 40% sludge lmd 0% sludge lmd 20% sludge lmd 40% sludge hofa 0% sludge hofa 20% sludge hofa 40% sludge 35 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing oil sludge 0% 10% 20% 30% 40% b on d s tr en gt h kn /m 2 12 14 16 18 20 22 24 26 28 30 32 34 oil-sludge vs bs figure 9. bond strength vs. oil-sludge. generally resulted in an increase in bs (fig. 10). both ckd and lmd produced composites with at least a 100% increase in bs compared to the original asphalt at 25% content. mixtures with 25% hofa yielded a material with a bs 12 times that of pure asphalt. neat asphalt 40% oil-sludge 5% sbs b on d s tr en gt h, k n /m 2 0 100 200 300 400 500 600 plain asphalt 25% lmd 25% ckd 25% hofa figure 10. bond strength of fillers, oil-sludge and sbs asphalt mastics. in higher quantities, the os yielded a composite with very little bs, and reviving its strength by adding 25% fillers seemed not to help improve the bs much (fig. 10). but the hofa-sludge mastic showed a considerable increase in bs when compared to neat asphalt, which had a bs above 100 kn/m2. for reference purposes, conventional sbs polymer mastics for asphalt roofing were also examined. the 5% sbs-modified asphalt exhibited almost twice the bs possessed by the 25% ckd and lmd containing asphalt. adding ckd or lmd to the 5% sbs blend nearly tripled its bs, but the sbs-hofa mastic resulted in a much higher bs (fig. 10). the 40% os mastic-containing hofa possessed a bs higher than the 5% sbs of modified asphalt. 3.8 results of statistical analysis of variance a two-way analysis of variance (anova) was conducted on the physical test results obtained with the additives used as effects/factors, using minitab statistical software (minitab ltd., coventry, uk) version16. this was to ascertain the relative effectiveness of the fillers within the sludgemodified asphalt and the os’s level of influence on the asphalt properties. however, before anova was selected for the analysis, a model adequacy check was performed on the data. tests such as the equality of variance test and normality check test were carried out to ensure the relevance and appropriateness of the selected test method. table 4 shows the results of the statistical analysis of the test results. os did not have a significant influence on the sp when used along with hofa and ckd fillers, unlike when it was used with lmd. but the os filler’s interaction was highly significant, and caused a more significant rise in the sp value when paired together than when the filler was utilized alone. in other words, the small loss in sp due to the use of sludge could be more than compensated for by the addition of ckd or hofa. os significantly affected the penetration when used along with the fillers, but ckd in particular showed an insignificant influence on the penetration compared to the rest of the fillers. viscosity was slightly affected by os, but all fillers significantly affected viscosity when paired with os. ductility was considerably influenced by os and the fillers. 3.9 astm specifications according to astm d 449, which is the standard specification for asphalt used in damp proofing and waterproofing, all the mastics containing only hofa are type ii and those containing ckd/lmd are type i material. most of the sludge-hofa mixes are considered type ii class, but some containing higher sludge/hofa levels failed to meet the minimum ductility limit. the majority of ckdsludge blends that should have been classified as type i material exhibited a slightly deficient 36 h.i. al-abdul wahhab and m.a. dalhat table 4. summary of statistical analysis result. analysis of variance result obtained at 5% significance level factors/additives tabular fvalue calculated fvalue p-value inference s o f t e n in g p o in t (o c) oil sludge hofa 4.4590 0.32 0.736 insignificant 3.8379 6.96 0.010 significant oil sludge ckd 4.4590 0.24 0.795 insignificant 3.8379 15.76 0.001 significant oil sludge lmd 4.4590 61.99 0.000 significant 3.8379 10.07 0.003 significant d u c t il it y ( cm ) oil sludge hofa 4.4590 1.72 0.239 insignificant 3.8379 2.82 0.099 insignificant oil sludge ckd 4.4590 5.84 0.027 significant 3.8379 2.17 0.163 insignificant oil sludge lmd 4.4590 7.96 0.013 significant 3.8379 1.30 0.349 insignificant p e n e t r a t io n (d m m ) oil sludge hofa 4.4590 0.24 0.795 insignificant 3.8379 15.76 0.001 significant oil sludge ckd 4.4590 16.99 0.001 significant 3.8379 1.40 0.316 insignificant oil sludge lmd 4.4590 98.57 0.000 significant 3.8379 4.44 0.035 significant v is c o s it y ( cp ) oil sludge hofa 4.4590 3.34 0.088 insignificant 3.8379 11.54 0.002 significant oil sludge ckd 4.4590 6.56 0.021 significant 3.8379 22.92 0.000 significant oil sludge lmd 4.4590 3.66 0.074 insignificant 3.8379 37.03 0.000 significant table 5. recommended mastic additive content. roofing application sludge hofa lmd/ckd 0–40% 10%−2 5% 15% −35% waterproofing and damp proofing application 0−40% 0%−15% 0%−25% fp value, which relegates them to a position outside the class range type. all the 20% sludge-lmd blends were type i, but the 40% sludge-lmd materials which should also have been classified as type i could not pass the ductility tests and some did not even pass the penetration standards. the sludge-hofa blends fall under various material classes according to astm d 312, which is the standard specification for asphalt used in roofing, but some have a fp slightly 37 oil-sludge extended asphalt mastic filled with heavy oil fly ash and cement waste for waterproofing below the minimum set point. most of the double-additive blends containing lmd or ckd in addition to sludge did not pass the sp criteria and all failed to satisfy the fp minimum set value. 4. conclusions all the fillers (hofa, ckd and lmd) produced composites with at least a 100% increase in bs as compared to the original asphalt with 25% content. a large quantity of os in a mixture resulted in deterioration of the asphalt mastic strength, but this loss in strength could be more than compensated for by filling the mastic with hofa. unlike the lmd and ckd sludge blends, the hofasludge mastic showed a considerable increase in bs when compared to neat asphalt. os interacted positively with the fillers (especially hofa and ckd) to improve their effectiveness in raising the sp and viscosity of the asphalt, and also in reducing its penetration and ductility. the resulting low fp of some of the sludge mastics as per roofing asphalt specification astm d 312 put into question their suitability for this application, even though these mastics hold the promise of becoming suitable composites for damp proofing and waterproofing. table 5 shows the recommended ranges of additive content for the two categories of waterproofing applications. finally, the bs test could encourage the bonding ability of the modified asphalt mastic which the other conventional asphalt tests such as the viscosity test could not do. the viscosity results show the ability of the sludge to thicken the asphalt mastic, a trend which is consistent with other fillers like hofa, ckd and lmd. but while the viscosity increased as a result of the os, yielding a lower bs, the increase in viscosity due to the fillers had a positive effect on the bs of the mastic. acknowledgments the authors acknowledge the support provided by king fahd university of petroleum and minerals and aramco, dhahran, ksa, in carrying out this research. references al-futaisi a, ahmad j, basma y, ramzi t (2007), assessment of alternative management techniques of tank bottom petroleum sludge in oman. journal of hazardous materials 141:557-564. al-methel m, al-abdul wahhab hi, hussein ia (2011), utilization of heavy oil fly ash to improve asphalt binder and asphalt concrete performance. patent no. 8:062-413 b1. astm d5 / d5m (2013), standard test method for penetration of bituminous materials. astm international, west conshohocken, pa, doi: 10.1520/d0005_d0005m-13. astm d36 / d36m (2012), standard test method for softening point of bitumen (ringand-ball apparatus). astm international, west conshohocken, pa, doi:10.1520/d0036_d0036m-12. astm d92 (2012b), standard test method for flash and fire points by cleveland open cup tester. astm international, west conshohocken, pa, doi: 10.1520/d009212b. astm d113 (2007), standard test method for ductility of bituminous materials. astm international, west conshohocken, pa, doi: 10.1520/d0113-07. astm d 312 (2000), standard specification for asphalt used in roofing. astm international, west conshohocken, pa, doi: 10.1520/d0312-00r06. astm d 449 (2003), standard specification for asphalt used in damp proofing and waterproofing. astm international, west conshohocken, pa, doi: 10.1520/d0449_d0449m. astm d 545 (2008), standard test methods for preformed expansion joint filler for concrete construction. astm international, west conshohocken, pa, doi: 10.1520/d0545-08. astm d 4007 (2011), standard test method for water and sediment in crude oil by the centrifuge method. astm international, west conshohocken, pa, doi 10.1520/d4007. astm d 4294 (2010), standard test method for sulfur in petroleum and petroleum products by energy dispersive x-ray fluorescence spectrometry. astm international, west conshohocken, pa, doi: 10.1520/d4294-10. 38 h.i. al-abdul wahhab and m.a. dalhat astm d 6352 (2012), standard test method for boiling range distribution of petroleum distillates in boiling range from 174 to 700°c by gas chromatography. astm international, west conshohocken, pa, doi: 10.1520/d6352-12. astm d 6481, (2010), standard test method for determination of phosphorus, sulfur, calcium, and zinc in lubrication oils by energy dispersive x-ray fluorescence spectroscopy. astm international, west conshohocken, pa, doi: 10.1520/d648199r10. chen js, kuo ph, lin ps, huang cc, lin ky (2008), experimental and theoretical characterization of the engineering behavior of bitumen mixed with mineral filler. materials and structures 41:1015-1024. eric pj (2006), asphalt mastic utilizing petroleum refinery waste solids. patent no. us 7,025,822 b2. fang c, zhou s, zhang m, zhao s (2009), modification of waterproofing asphalt by pvc packaging waste. journal of vinyl and additive technology 15(4):229-233. finch g, hubbs b, bombino r (2010), moisture transport by osmotic flow through waterproofing membranes – towards the development of osmosis resistant membranes. proceedings of buildings xi conference, clear water, florida, usa. goikoetxeaundia g, gonzalez o, munoz me, pena jj, santamaria a (2007), dynamic viscoelastic characterization of bitumen/polymer roofing membranes. macromolecular materials and engineering 292: 715-722. johansson ls, isacsson u (1998), effect of filler on low temperature physical hardening of bitumen. construction and building materials. 12:463-470. khan mj, al-juhani aa, ul-hamid a, shawabkeh r, hussein ia (2011), effect of chemical modification of oil fly ash and compatibilization on the rheological and morphological properties of low-density polyethylene composites. journal of applied polymer science 122:2486-2496. lucke h (1989), single polyurethane-modified bitumen composition. us pat. no. 4795760. martin-alfonso mj, partal p, navarro fj, garcia-morales m, gallegos c (2008a), use of mdi-functionalized reactive polymer for the manufacture of modified bitumen with enhanced properties for roofing application. european polymer journal 44:1451-1461. martin-alfonso mj, partal p, navarro fj, garcia-morales m, gallegos c (2008b), the role of water in the development of new isocyanate-based bituminous products. industrial & engineering chemistry research 47(18):6933-6940. paul mm (2002), history of polymer modified and cold process roofing. (http://www.rcionline.org/interface/2002-11-mitchell.pdf). recasens rm, martinez a, jimenez fp, bianchetto h (2005), effect of filler on the aging resistance of asphalt mixture. journal of transport research board 1901:10-17. singh b, gupta m, tarannum h (2002), mastic of polymer-modified bitumen and poly(vinylchloride) wastes. journal of applied polymer science 90:1347-1356. taha r, ba-omar m, pillay ae, roos g, alhamdi a (2001), recycling of petroleumcontaminated sand. journal of environmental monitoring 3:417-420. taha r, hassan h, al-rawas a, yaghi b, alfutaisi a, jamrah a, al-suleimani y (2007), use of tank bottom sludge to construct and upgrade unpaved roads. transportation research record: journal of the transportation research board 1989 208-214. terry ce, berard ra, pinholster jr, df (1999), polyurethane-modified bitumen coating composition. us patent 5, 981,010. veenstra jn, sanders da, snethen dr (2000), evaluation of road base material derived from tank bottom sludge. integrated petroleum environmental consortium (ipec). world asphalt (2010), industry study with forecasts for 2013 & 2018. freedonia study, #2602. world asphalt (2012), industry market research for business leaders, strategists decision makers. freedonia, sku: fg6830349. mech10_11_03_tb.qxd the journal of engineering research vol. 2, no. 1 (2005) 43-52 1. introduction many engineering problems lead to the consideration of contact between surfaces. these include studies related to friction-induced vibration and noise, thermal and electrical contact resistance, and mechanical seals, bushing, fasteners, etc. examination of contact characteristics encompasses equivalent contact stiffness and damping for vibration and noise, and true contact area for mechanical seals, and electrical and thermal resistance. accounting for contact characteristics inherently necessitates characterization of surface topography and development of probabilistic models, which relate contact area, contact stiffness, contact load and separation of surfaces. these models are based on the presumption that a surface can, in effect, be represented by a distribution of asperities. as two surfaces are brought into contact, the macroscopic contact characteristic in question is a cumulative effect of localized interactions of the asperities. this approach has required the statistical formulation of a surface and statistical summation of microscopic contact effects to obtain probabilistic macroscopic expectation of ______________________________________ *corresponding author e-mail: jdabdo@squ.edu.om the contact characteristic (contact area, load, and stiffness). greenwood and williamson (1966) pioneered "asperity-based model." the existing literature shows extensions of the greenwood and williamson (gw) model over the last three decades. the present probabilistic models of contact may be viewed with respect to the premise of elastic and plastic contact. the elastic models primarily rely on the hertz theory of contact between two elastic bodies (greenwood and williamson, 1966; greenwood and tripp, 1967, 1970; hisakado, 1974; bush et al. 1975 and mccool, 1966). these models differ in their assumptions related to surface and asperity geometry and material properties. these extensions have included, for instance, the inclusion of the surface curvature effects greenwood and tripp, (1967), allowance for non-uniform curvature of asperity summits hisakado, (1974) and presumption of average elliptic paraboloidal representation of asperity bush et al. (1975). other works presented advanced models for anisotropic materials mccool, (1986). of particular interest to the future goal of the present work is the surface model in which asperities are allowed to form contact on shoulders greenwood and tripp, (1967). the plastic models are based on the presumption that a modified approach in modeling and calculation of contact characteristics of rough surfaces j.a. abdo* mechanical and industrial engineering department, college of engineering, sultan qaboos university, p.o. box 33, al-khod 123, muscat, sultanate of oman received 10 november 2003; accepted 30 july 2004 abstract: a mathematical formulation for the contact of rough surfaces is presented. the derivation of the contact model is facilitated through the definition of plastic asperities that are assumed to be embedded at a critical depth within the actual surface asperities. the surface asperities are assumed to deform elastically whereas the plastic asperities experience only plastic deformation. the deformation of plastic asperities is made to obey the law of conservation of volume. it is believed that the proposed model is advantageous since (a) it provides a more accurate account of elasticplastic behavior of surfaces in contact and (b) it is applicable to model formulations that involve asperity shoulder-toshoulder contact. comparison of numerical results for estimating true contact area and contact force using the proposed model and the earlier methods suggest that the proposed approach provides a more realistic prediction of elastic-plastic contact behavior. keywords: elastic-plastic, contact, friction, rough surfaces, contact model ¢¢üü∏∏îîàà°°ùÿÿgg~m ¤g πng~àj ¿g ¢vîøj »àdgh áfôÿg òz äcgƒàædg ∞jô©j ∫ón øe ” á«°véjôdg ä’oé©ÿg ¥é≤à°tg .áæ°ûÿg í£°s’g ¢sé«≤d »°véjq êpƒ‰ ω~≤j ábqƒdg iòg : åëñdg gò¡d .ºé◊g ßøm ¿ƒféb ™ñàj å«ëh ¿ôe òz πµ°ûh »æëæj áfôÿg òz äcgƒàædg éªæ«h ¿ôe πµ°ûh »æëæj ¿cg ¢vîøj äcgƒàædg í£°s .»≤«≤◊g äcgƒàædg í£°s ‘ ú©e é° jg åëñdg ω~≤j .»ñfé÷g ¢séªà∏d ä’oé©e oééjg ‘ ∫骩à°s’gh ≥«ñ£à∏d πhéb (ü) .¿ôÿg òzh ¿ôÿg í£°ùdg ¢sé“ ∑ƒ∏°s üé°ù◊ ècg ábo »£©j (cg) :úññ°ùd á«∏° ag .é≤hé°s â“iông äé°sgqo èféàfh »°véjôdg êpƒªædg gòg ωg~îà°sg øe œéædg úh ¢séªà∏d á«≤«≤◊g iƒ≤dgh ámé°ùÿg ôj~≤àd ájo~y áfqé≤e ««mmééààøøÿÿgg ääggooôôøøÿÿgg .¢séªà∏d »°véjq êpƒ‰ ,áæ°ûÿg í£°s’g ,∑éµàm’g ,¿ôÿg òz¿ôÿg ¢séªàdg : á áæ°ûÿg í£°sód ¢séªàdg ¢üfé°ün üé°ù◊ ∫~©e »°véjq êpƒ‰ 44 the journal of engineering research vol. 2, no. 1 (2005) 43-52 contact is dominated by plastic flow. such models may be best suited for load ranges that warrant large degree of plastic flow. the earliest work on plastic contact model is attributed to abbott and firestone (1933). this model utilizes the geometrical intersection of a plane with a rough surface and presumes that contact flow pressure exists over the area of contact, obtained from geometrical intersection. later works include two dimensional random process model nayak, (1971), and investigations leading to the postulation that volume conservation dictates plastic flow (pullen and williamson, 1972; nayak, 1973). the research work of francis (1976, 1977), dealing with plastic deformation of contact, include empirical characterization using the results for spherical indentation. liu et al. (2000) derived exponential functions from a fitting procedure applied to numerical results of the gaussian height distribution offering analytical expression for the greenwood and williamson (1966); elastic model, change, et al. (1986), elastic-plastic model and horng, (1998) elliptic elastic-plastic model. whereas the elastic and plastic models are seen to be advantageous for extreme cases of loading, in a large number of engineering applications, contact loads may fall within ranges that do not warrant adequate representation by either elastic or plastic model. this fact has led researchers to consider what is referred to as elastic-plastic models (ishigaki et al. 1979, chang et al. 1997). in the latter work, elastic-plastic model of contact is proposed based on the conservation of volume during plastic deformation. the work proposed by chang et al. (1997) treat contact as elastic-plastic at macroscopic scale while at microscopic level it views an asperity to experience initially elastic deformation followed by purely plastic deformation when a critical interference is reached. more recent works (kucharski et al. 1994; chang, 1986) have considered and modified contact models to treat elasticplastic behavior of metallic coating. finite element method solutions to elastic-plastic contact problems were also utilized. liu et al. (2000) developed a finite element method solution for an elastic-plastic contact problem. the finite element method presented by kogu and etsion (2002) suggested that the evolution of the elastic-plastic contact could be divided into three distinct stages. kogu and etsion (2003) presented an improved finite element-based model for the contact of rough surfaces. the contact parameters for a single asperity contact were predicted. the present work proposes a modified elastic-plastic model for surfaces in contact. the critical interference proposed by chang et al. (1986) is utilized to define fictitious asperities. in this manner two sets of asperities are defined: (1) those that are the actual surface asperities and (2) the fictitious asperities which can only deform plastically. the concept of elastic and plastic asperities allows formulation of asperity deformation model which is also elastic-plastic at microscopic scale; in contrast to the model by chang et al. (1986). volume conservation is applied only to the plastic (fictitious) asperities. comparisons of the contact load/separation values predicted by the proposed model with that obtained through experiments as reported by kucharski et al. (1994) demonstrate the effectiveness of the method in providing accurate prediction of contact characteristics. 2. the mathematical model consider first the contact between a rough surface covered with a number of asperities, that have spherical shape at their summits with average radius r and a nominally flat surface shown in fig. 1. the summit heights, z, are z ys w d statistical representation (b) rough surface plane surface mean of asperity heights mean of surface heights (a) wc average asperity fictitious asperity (c) figure 1. contact between a flat and a rough surface 45 the journal of engineering research vol. 2, no. 1 (2005) 43-52 assumed to have probability density function φ(z). the probability that an asperity develops contact with the flat plane is (greenwood and williamson, 1966): (1) where d is the separation based on asperity heights. for n asperities, the expected number of contacts will be: (2) where the total number of asperities, n, the density of asperities, η, and the nominal area, an, are related according to (3) for this type of contact, the interference w may be described as (greenwood and williamson, 1966): (4) in analyzing the contact, the laws describing the dependence of the local contact area, ao, and the local contact load, po, on w are employed. hence, (5) the expected total contact area, a1, and the expected total contact load, p, are the statistical sums of the local contributions of each asperity. therefore, (6) (7) the hertzian contact area, ao, contact load, po, and the maximum contact pressure between one asperity, having interference w with a plane, pm, are given as follows: (8) (9) (10) where, (11) in the equations above r is the average equivalent radius of curvature. in general, the maximum contact pressure is related to the hardness of the softer material, h, through (12) where k is the maximum contact pressure factor. tabor (1977) showed that the plastic flow is reached when pm= 0.6h. the critical interference at the inception of plastic deformation wc is defined by substituting eq. (12) in eqs. (10) and (14). (13) for w > wc, the contact is plastic. using the critical interference and imposing the conservation of volume, chang et al. (1997) derived the modified equations describing the contact area and load on an asperity: (14) and (15) it is appropriate here to emphasize the difference between the present work and that proposed by chang et al. (1997). in the derivation of the equations, the authors assume that an asperity behavior is initially elastic. as the load is increased the elastic behavior continues to describe the deformation until a critical interference is reached. at this critical load and beyond, the asperity deforms as a purely plastic body. in the work of chang et al. (1997) the elastic and plastic behavior do not occur simultaneously for an asperity. hence, their formulation can only be characterized as an elastic-then-plastic model at microscopic scale. the method proposed in this paper is shown to represent more accurately an elastic-plastic model of the contact, through the introduction of a fictitious surface that can only deform plastically. as shown in fig. 1, the critical interference, wc, is used to define a second surface. this second surface is obtained by displacement of every point on the surface by wc along the direction normal to the surface (see fig. 2). as illustrated in fig. 2, to obtain the mathematical description of the plastic asperity, the mapping of a point, a, on the surface to a point, b, on the plastic asperity must be considered. it is also noted that an asperity is described fig. 2 in terms of a frame of reference whose origin is at the asperity peak and ordinate points towards the mean plane greenwood and tripp, (1967). therefore, ρy-frame is used to describe the original asperity, whereas, xy-frame is employed for the plastic asperity. ∫=> ∞ d dzzdz )()(prob φ ∫= ∞ dc dzznn )(φ nan η= dzw −= ,)(00 waa = )(00 wpp = dzzdzaada d nt )()()( 0 φη −∫= ∞ dzzdzpadp d n )()()( 0 φη −∫= ∞ rwa π=0 2321 0 3 4 werp = 21 0 0 2 2 3 ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ == r we a p pm π 2 2 2 1 2 1 111 eee νν − + − = khpm = ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −== w w rw a a c2 4 2 0 π π ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −== w w rw a a c2 4 2 0 π π kh w w rwp c ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −= 20 π the respective positions of points a and b are denoted by ar and br , as depicted in fig. 2. the position of point b on the fictitious asperity is : 46 the journal of engineering research vol. 2, no. 1 (2005) 43-52 (16) (17) equations (16) and (17) may be used to obtain the equation describing the plastic asperity (see appendix a for detail). to the first approximation, the equation of the plastic asperity is given as (appendix a): (18) since the plastic asperities only deform plastically, their introduction allows the reduction of eqs. (14) and (15) to the well-known forms: (19) (20) where, rp represents the summit radius of curvature of the plastic asperity. based on eq. (18), this radius of curvature is (21) clearly the formulation of the area of contact and contact force for two surfaces involves interactions of two sets of asperities: the original surface asperities which are assumed to deform elastically and plastic asperities that deform plastically. the next section presents this new mathematical model of contact. 3. elastic-plastic contact between a rough surface and a plane the contact between one asperity on a rough surface and a plane is considered. figure 3 illustrates two types of interactions. the first is the elastic contact between the plane and the surface asperity. if the interference, w, exceeds the critical interference wc, then the interaction also includes plastic contact. it is noted that the shaded volume representing the interference of the plastic asperities and the plane contribute to the plastic portion of contact whereas the remaining volume of interference contributes to the elastic contact. therefore if we let q be the characteristic of contact (area or load) then it may be obtained by appropriately accounting for the aforementioned interactions, that is: (22) where qe1 corresponds to the elastic contact between the plane and surface asperity. however this includes, as shown in fig. 3, a portion of the interference which is plastic. therefore, the contribution due to elastic interferfigure 2. plastic (fictitious) asperity shape ncacb uwrjwr ++−= where, nu is the unit normal vector to the original asperity at point a. as usual , the unit vectors i and j are defined along ρ (or x) and y axes, respectively. employing the notation of greenwood and tripp (1967), j r ira 2 2 1 ρρ += )(2 2 cwr x y − = wra pπ20 = wkhrp pπ20 = )( cp wrr −= 221 pee qqqq +−= figure 3. elastic-plastic interaction of an asperity with a plane 47 the journal of engineering research vol. 2, no. 1 (2005) 43-52 ence between the plane and the plastic asperity qe2 must be subtracted to obtain the net elastic contribution. the contribution from plastic interaction due to the plastic interference of the plane and plastic asperity qp2 is, then, added to the result. using this approach the area of contact may be described as: (23) where, (24) (25) (26) (27) (28) (29) it is noteworthy to mention here that the plastic asperity peaks can be viewed as being farther away from the plane by wc. we have made the approximation as to the mean plane of plastic asperity being wc below the mean plane of the surface asperities. therefore the limits of integration are shifted by wc as presented in eqs. (25) and (26). furthermore, in these equations the summit curvature corresponding to plastic asperities are used. proceeding in a similar manner, the contact load may be written as: (30) where, (31) (32) (33) 4. results: evaluation of the proposed model the effectiveness of the proposed model is evaluated using the data and results given by chang et al. (1997) and kucharski et al. (1994). for convenience we shall refer to the former as ceb and the latter as kkpk models. the proposed model in this paper will be referred to as afm model. in the ensuing discussion, results are presented in two subsections. in the first subsection the afm model is evaluated and compared with the ceb model. in the second subsection afm, ceb and kkpk models are tested against experimental results given by kucharski et al. (1994). 4.1 comparison of afm and ceb models in the ceb model, as it is the case in our model, a dimensionless form of the probability density function is introduced: (34) to combine the material and surface topographic properties in contact, the plasticity index, ψ, is introduced according to greenwood and williamson (1996): (35) the relation between separation h and σ of the surface microgeometry model and d and σs of the asperity-based model is given as: (36) a form for plasticity index is obtained by substituting eqs. (13) and (36) in (35), to give: (37) ])(5.0exp[)()2()( 2221* ss ss σ σ σ σ πφ −= − * 2 * 2 * 1 * )( epet aaaha −+= ∫= ∞ − ** *** 1 )()( syh e dsswha φπβ ∫= ∞ +− *** *** 2 )()( cwsyh pe dsswha φπβ ∫= ∞ +− *** *** 2 )(2)( cwsyh pp dsswha φπβ where φ(s) is the dimensionless form of the probability density function for summit height distribution . β, βp, w* and the dimensionless height of asperity , s, are defined as follows : σηβσηβ )(, cp wrr −== *** syhsw +−= σ z s = * 2 * 2 * 1 ** )( epe ppphp −+= ∫⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = ∞ − ** 23* 21 ** 1 )(3 4 )( syh e dsswr hp φβ σ ∫⎟ ⎠ ⎞ ⎜ ⎝ ⎛ = ∞ +− *** 23* 21 ** 2 )(3 4 )( cwsyh pe dsswr hp φβ σ ∫= ∞ +− *** *** 2 )(2)( cwsyh pp dsswe h khp φβπ 21)( −= s cw σ ψ 22 4 22 10717.3 r s η σσ −× += 41 2 4 21 ) 10717.3 1()( 2 β σ π ψ −× −= rkh e the data employed by chang et al. (1997) pertaining to steel on steel material is employed with the following parameters: mpaee 521 1007.2 ×== , brinell hardness mpah 1960= , 29.021 == νν . the maximum contact pressure is taken to be based on 6.0=k , and β and σ/r are selected for different values of plasticity index from the experimental work of nuri and halling (1975). these values and the values of *cw and σ/σs, calculated from eqs. (35) and (36), respectively, are shown in table 1. figure 4 illustrates the contact area versus separation, both given in dimensionless form, for the afm and ceb models. while both models predict similar contact area for low plasticity index, the afm model pr edicts higher values for materials of higher plasticity index. figure 5 depicts the dimensionless 48 the journal of engineering research vol. 2, no. 1 (2005) 43-52 figure 4. contact area ration versus separation: comparison between aft and ceb models figure 5. dimensionless separation versus contact load: comparison between aft and ceb models ψ ηη r σ σ/r σ/σs 0.5 0.0302 8.75x10-5 1.299192 0.8 0.374 2.00x10-4 1.001331 2.5 0.601 1.77x10-3 1.000515 table 1. plasticity index and surface topography c on ta ct a re a r at io , a t/a n dimensionless mean separation d im en si on le ss m ea n se pa ra tio n dimensionless load, p/ane 49 the journal of engineering research vol. 2, no. 1 (2005) 43-52 (38) therefore material with low plasticity index may be approximated as a purely elastic body. on the other hand, for high plasticity index (softer material), the contact is approximately totally plastic and the total expected dimensionless contact load for elastic-plastic contact of eq. (30) is approximately: (39) figure 6 depicts the contact area ratio versus dimensionless contact load for different values of, φ, as predicted by the afm and ceb models. 4.2 comparison of afm, ceb and kkpk models it is the intent of this subsection to evaluate the proposed model (aft) and present a comparative study of the model with those proposed by ceb and kkpk. in doing so we present the experimental results by kucharski et al. (1994) as the basis of these comparisons. they presented a finite element model of elastic-plastic contact (referred to as kkpk). they also performed measurement of contact load, area and approach. in their work, steel specimens are employed with young's modulus e1 = 200,000 mpa, poisson's ratio v1 = 0.3 and tensile yield strength y = 400 mpa. the three dimensional profilometry of sand-blasted surface (e60), average over three samples, are given as follows kucharski et al. (1994): as in kucharski et al. (1994), 3y/2 is employed instead of kh value in calculating the plastic contact contribution. moreover, dimensionless approach is used instead of separation in presenting the results. in this case, approach is obtained by kucharski et al. (1994): (40) where α is the approach, and zmax and h are the maximum peak height and separation. figure 7 illustrates the results. in this case dimensionless approach is with respect to the maximum summit height, msh. clearly, the proposed (afm) model presents the most favorable agreement between the predicted contact load/approach values to that obtained by measurements. the accuracy of the afm model is also attested to by the results shown in fig. 8. the results show that the afm model provides significantly more accurate prediction of contact load and contact area than ceb and kkpk models. 4.3 elastic-plastic contact for rough surfaces in this section direct formulation of contact of roughon-rough surfaces based on the work of greenwood and tripp (1967) is presented. the purpose here is to demonstrate the adaptability of the proposed contact model using plastic asperity concept and also a formal mathematical treatment of contact between two rough surfaces. greenwood and tripp (1967) introduced elastic as well as plastic models for two rough surfaces covered with paraboloidal asperities. consideration of contact between two asperities results in the following description of interference (41) where z1 and z2 are the heights of asperities on surfaces one and two, respectively. r is the radial distance denoting the offset between the central lines of the two asperities. for a summit radius of curvature r, the presumption of hertzian contact leads to: (42) (43) the expected dimensionless contact load for elastic contact is (44) and that for plastic contact is (45) where, )( 15 216 )( **25 2** syhfr hp −⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = σ πβ separation h* versus dimensionless load ea p n for different values of plasticity index , ϕ , as predicted by the afm and ceb models. it is clear from the figure that the separation increases with the plasticity index for a given load. for low plasticity index (harder material), the contact is approximately totally elastic. as summarized in table 1, lower plasticity index corresponds to larger critical interference, cw . this, in turn, increases e ffective separation for plast ic asperity as seen in eqs. (32) and (33) . the result is diminished contributions of plastic asperities. hence , in eq. (30) 2ep and 2pp attain smaller value s, resulting in the following approximation : * 1 ** )( ephp = * 2 ** )( pphp = • areal summit density 4.655=η mm-2 • mean summit radius r = 30.0µm • 3d standard deviation of summit height distribution σ = 2.3µm • 3d maximum peak height zmax = 13.5 µm • 3d standard deviation of the surface height distribution σs = 2.95µm • 3d maximum summit height msh = 9.9µm hza −= max )2/(221 rfzzw −+= w r a 20 π= 23 21 0 23 4 w r ep ⎟ ⎠ ⎞ ⎜ ⎝ ⎛= )(2)( **2 22** syhfe h hp −= βπ dssusuf u n n )()()( φ∫ −= ∞ 50 the journal of engineering research vol. 2, no. 1 (2005) 43-52 figure 6. contact area ratio versus contact load: comparison between aft amd ceb models 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.00e+00 5.00e-04 1.00e-03 1.50e-03 dimensionless load, p/ane d im en si on le ss a pp ro ac h aft ceb experiment fem figure 7. contact load versus approach: comparison between aft, ceb, experimental results of kkpk, and fem model of kkpk figure 8. contact area versus dimensionless load: comparison between aft, ceb, experimental results of kkpk, and fem model of kkpk c on ta ct a re a r at io , a t/a n c on ta ct a re a r at io , a t/a n d im en si on le ss a pp ro ac h dimensionless load, p/ane dimensionless load, p/ane dimensionless load, p/ane 51 the journal of engineering research vol. 2, no. 1 (2005) 43-52 hence, using the proposed method, the total dimensionless expected contact load for elastic-plastic behavior is: (46) where, (47) (48) (49) and σ denotes the standard deviation of sum of asperity heights on the two surfaces. 5. conclusions a mathematical formulation of elastic-plastic contact has been presented in this paper. using the definition of critical interference, the concept of elastic and plastic asperities have been developed in which a rough surface is represented by two surfaces. the first surface is the actual physical surface and is assumed to deform as a purely elastic body. the second surface is a fictitious one and it is derived from the physical surface and the critical interference. this surface is assumed to deform as a purely plastic body. the development of the elastic and plastic surfaces has facilitated the mathematical formulation of elastic-plastic contact of rough surfaces. comparison of the proposed model with the existing models for elastic-plastic contact have been performed. the measurement results of previously performed experiments on sand-blasted steel surface (kucharski et al. 1994) have shown that the proposed model provides significant improvement over previous models in the prediction of contact load and area of contact. the proposed model (aft) gives better prediction than ceb since in ceb for every asperity the elastic quantity due to plastic pressure qe2 is not considered. hence, the elastic and plastic behaviors do not occur simultaneously for an asperity. therefore, their formulation can only be characterized as an elastic-then-plastic model at microscopic scale and elastic-plastic model at macroscopic scale. appendix a consider an asperity and assume that its shape is quadratic as proposed by greenwood and tripp (1971). the equation of the surface asperity is given by (a1) (a2) the unit tangential vector is obtained as (a3) hence the unit normal vector is (a4) then the description of the plastic asperity is obtained by (a5) or (a6) (a7) let then (a8) therefore, the shape of the plastic (fictitious) asperity is given by: (a9) * 2 * 2 * 1 ** )( epet ppphp −+= )( 15 216 )( **25 2** 1 se yhfr hp −⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = σ πβ )( 15 216 *** 25 2* 2 cspe wyhfr p +−⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = σ πβ )(2 ***2 22* 2 cspp wyhfe h p +−= βπ 2 2 1 ρ r y = j r ijyira 2 2ρ ρρ +=+= )( 1 1 2 2 j r i r rd rd u a a t ρ ρ + + == )( 1 1 2 2 ji r r un +− + = ρ ρ jwuwrr cncab −+= jw r w r i r r w r c c c b ) 1 2 () 1 1( 2 2 2 2 2 − + ++ + −= ρ ρ ρ ρ j r w r i r w r ccb ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −+⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −= 1 2 1 2ρ ρ )(2 2 cwr x y − = as shown in fig. 2. the figure also illustrates fictitious plastic asperity whose shape is obtained by a displacement of cw along the normal to the quadratic curve. let tu and nu represent t he tangential and normal unit vector to the quadratic at point a. the position of point a is given by vector ar as: ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ −= r w x c1ρ for small r wc and r ρ , br may be approximated by : j wr x ixr c b )(2 2 − += 52 the journal of engineering research vol. 2, no. 1 (2005) 43-52 references abbott, e. j. and firestone, f.a., 1933, "specifying surface quantity-method based on accurate measurement and comparison," mech. engr. vol. 55, pp. 569-577. bush, a. w., gibson, r. d. and thomas, t. r., 1975, "the elastic contact of a rough surface," wear, vol. 35, pp. 87-111. chang, w.r., 1997, "an elastic-plastic contact model for a rough surface with an ion-plated soft metallic coating," wear, vol. 212, pp. 229-237. chang, w.r., etsion, i. and bogy, d.b., 1986, "an elasticplastic model for the contact of rough surfaces," j. of tribology, vol. 109, pp. 257-263. francis, h. a., 1977, "application of spherical indentation mechanics to reversible and irreversible contact between rough surfaces," wear, vol. 45, pp. 221-269. francis, h. a., 1976, "phenomenological analysis of plastic spherical indentation," asme j. of engineering materials and technology, vol. 98, pp. 272-281. greenwood, j. a. and williamson, j. b. p., 1966, "contact of nominally flat surfaces," proc. roy. soc. (london) vol. a295, pp. 300-319. greenwood, j. a. and tripp, j. h., 1967, "the elastic contact of rough sphere," asme j. of appl, vol. 34, pp. 153-159. greenwood, j. a. and tripp, j. h., 1970, "the contact of two rough nominally flat rough surfaces," proc. instn. mech. engrs. vol. 185, pp. 625-633. hisakado, t., 1974, "effect of surface roughness on contact between solid surfaces," wear, vol. 28, pp. 217-234. horng, j. h., 1998, "an elliptic elastic-plastic aspirity microcontact model for rough surfaces," asme j. of tribology, vol. 120, pp. 82-88. ishigaki, h., kawaguchi, i. and mizuta, s., 1979, "a simple estimation of the elastic-plastic deformation of contacting asperities," wear, vol. 54, pp. 157-164. kogu, l. and etsion, i., 2002, "elastic-plastic contact analysis of a sphere and a rigid flat," asme j. of appl. mech, vol. 69, pp. 657-662. kogu, l. and etsion, i., 2003, "a finite element based elastic plastic model for the contact of rough surfaces," trib. trans, pp. 383-390. kucharski, s., klimczak, t., polijaniuk, a. and kaczmarek, j., 1994, "finite-element model for the contact of rough surfaces," wear, vol. 177, pp. 113. liu, z., neville, a. and reuben, r. l., 2000, "an analytical solution for elastic and elastic-plastic contact models," asme j. of tribology, vol. 43, pp. 627-633. liu, g., zhu, j.l., yu, l. and wang, q.j., 2001, "elasticplastic contact of rough surfaces,” asme j. of tribology, vol. 44, pp. 437-443. mccool, j. i., 1986, "predicting microfracture in ceramics via a microcontact model," asme j. of tribology, vol. 108, pp. 380-386. nayak, p. r., 1973, "random process model of rough surfaces in plastic contact," wear, vol. 26, 305-333. nayak, p. r., 1971, "random process model of rough surfaces," asme j. of lubrication technology, vol. 93, pp. 398-407. nuri, k. a. and halling, j., 1975, "the normal approach between rough flat surfaces in contact," wear, vol. 32, pp. 81-93 pullen j. and williamson, j.b.p., 1972, "on the plastic contact of rough surfaces," proc. roy. soc (london), vol. a327, pp. 159-173. tabor, d.j., 1977, “collide interface science,” proc. roy. soc (london), vol. 58(2), pp. 2-10. c:\users\raoh\desktop\paper 1.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 01-21 prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes h.i. al-abdul wahhab*, a and g.m.s. abdullahb a department of civil engineering, king fahd university of petroleum and minerals, dhahran 31261, saudi arabia b department of civil engineering, najran university, najran, saudi arabia received 24 november 2014; accepted 12 august 2015 abstract: this paper presents a study carried out to evaluate and compare the permanent deformation of marginal soils. under consideration were marl, sabkha and dune sand stabilized with foamed and emulsified sulfur asphalt (fsa, esa) with mixes of the same soils stabilized with conventional foamed and emulsified asphalt (fa, ea) for road base construction. designed mixes at their optimum asphalt content were evaluated for the dynamic resilient modulus (mr) at 22°c and 40°c and dynamic triaxial at three levels of deviatoric stress and at 22°c and 40°c. the wheel tracking (wt) test was carried out at 22°c. permanent deformation of stabilized mixes with optimum binder contents was modeled and simulated using dynamic triaxial and wt tests. the developed models were calibrated to predict rutting using vesys 5w software. results indicated that the fsa increased rutting resistance as compared to conventional fa mixes. on the other hand, esa increased the rutting susceptibility of marl soil as compared to ea. the calibrated models of rutting prediction were found to predict the rut depth with 90% accuracy. keywords: rutting, dynamic triaxial, wheel tracking, calibrated models, vesys. oo ooo (vesys 5w) * corresponding author’s e-mail: hawahab@kfupm.edu.sa h.i. al-abdul wahhab and g.m.s. abdullah 2 1. introduction a major stressor in flexible pavement is rutting, as indicated by the permanent deformation, or rut depth, sometimes found along wheel paths. rutting in flexible pavement usually consists of longitudinal depressions in the wheel path due to the accumulation of small amounts of unrecoverable deformation caused by each load application (fig. 1) (hafeez and imran h 2009; asphalt institute 1996). many factors affect the width and depth of the rut, such as structural characteristics of pavement layers (thickness and material quality), traffic loads and environmental conditions. the accumulation of load-induced permanent deformation developed from all individual pavement layers, including the subgrade forms of surface rutting. a large percentage of the total rutting will mostly come from the under layers of pavement such as the base, subbase, and subgrade when the surface layer is thin. several models have been developed to predict rutting in asphalt concrete (ac) pavement layers and were evaluated in national cooperative highway research program (nchrp) publication 1-26 (barenberg et al. 1990). the study reported that those models which are related to the log of permanent strain to the log of load repetition appear to be the most suitable and versatile for practical use. this power model is often fitted to the accumulated permanent deformation curve and is definitely the most commonly used permanent deformation formula. the power model usually plots a straight line on a log-log scale. the slope and intercept of the power model, when plotted on a log-log scale, are used as indicators of permanent deformation resistance (garba 2002). figure 1. accumulated plastic strains in pavements (asphalt institute 1996). the basic permanent strain to load repetition model is expressed as: ɛp = anb (1) this model was initially suggested for subgrade and unbound materials by monismith, carl lm (1976) and was first used for ac mixes by khedr, safwan ak (1986). various researchers consequently used the same model for ac (diyaljee and raymond 1982; vuong and peter 1991; gholam and william 1996), where a and b are the intercept and slope coefficients and n is the load repetition. figure 2 presents graphically the curve of a power model on a log-log scale between load repetition and permanent strain. the progress of rutting with load repetition can be measured using layer elastic and viscoelastic theory. in order to get the cumulated permanent strain in each layer, the following equation is used (hafeez, imran h 2009): (2) where r ab α = 1b (3) p = permanent strain (rut value) n = number of load application a = intercept coefficient of accumulated permanent strain versus number of load repetitions curve on log-log scale; b = slope coefficient of accumulated permanent strain versus number of load repetitions curve on log-log scale; and εr = resilient strain εr = resilient strain µ = ratio of plastic to elastic response α = rate of change of the plastic response. thus, the total rut depth can now be calculated as: ipi n id hr 1 1 (4) figure 2. log-log form of power model. prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 3 where, hi is thickness of the layer. according to bevan (2002), α and µ are the stressand temperature-dependent nonlinear parameters and can be used for modeling permanent deformation of mixes. thus, in order to predict the total permanent deformation of a pavement structure, one has to determine the properties of material in each layer and predict the layer rut. the vesys model was adopted in this study since it is considered suitable to predict the layer and total rut depths of the pavement structure. the model parameters depend on the results of the experimental work on the material for each layer. therefore, the dynamic triaxial repeated load test, which is a suitable lab test to obtain the vesys model parameters, was conducted on the investigated materials. 2. experimental work marginal soils cover most of the eastern provinces of saudi arabia. in order to investigate the possible treatment of the marginal soils to be used in road construction, dune sand, marl, and sabkha soils were collected, subjected to basic characterization, and then stored for use in the experiments. the basic engineering properties of the soils were assessed by conducting preliminary characterization tests, including specific gravity, plasticity tests, and grain size distribution. standard foamed asphalt (fa) and 30/70 foamed sulfur asphalt (fsa) were produced using a calibrated laboratory scale fa plant wlb 10. the laboratory mix design procedure was carried out as outlined in the wirtgen cold recycling manual to determine the optimum asphalt content for the marl, sabkha, and dune sand soil mixes. after adding fa, 2% cement was added to all mixes to prevent effecting the optimum moisture content (wahhab et al. 2012). a set of test specimens 63.5 ± 6 mm high and 101.6 mm in diameter were prepared over a range of residual asphalt content. mixes were designed using soaked indirect tensile strength (its) test. a relationship between soaked its and residual asphalt content was plotted and the optimum residual asphalt content, which provides the maximum soaked its, was determined. for emulsified soils mixes, a set of test specimens (63.5 ± 6 mm high and 101.6 mm in diameter) were prepared over a range of residual asphalt contents. test mixtures were prepared in various increments of residual asphalt contents using previously determined amounts of premix water and the optimum water content required for mixing and compaction. prepared mixtures were compacted in a marshall mold with 75 blows for each side. after that, the molds containing the compacted specimens were placed on a perforated shelf in a 60°c (140°f) forced draft oven for 48 hours (asphalt emulsion manufacturers association, 2004). specimens were soaked under 50 mm hg of vacuum pressure in water for one hour and without vacuum pressure for another hour and then tested for soaked stability. based on the results, a relationship between soaked stability and residual asphalt content was plotted, and the optimum residual asphalt content providing the maximum soaked stability was determined. designed fsa and esa at their optimum binder contents including marl, sabkha, and dune sand in addition to the mixes of these soils with conventional fa and ea were subjected to mr (aashto t-307), dynamic triaxial and wheel tracking (wt) tests to evaluate their permanent deformation behavior. 3. results the basic properties of the marginal soils (marl, sabkha and dune sand) used in this study were determined and the soils were classified. stabilized soils treated with fsa and esa, as well as the conventional fa and ea were designed utilizing indirect tensile strength (its) and marshall stability. optimum asphalt contents for all mixes are summarized in table 1. the designed mixes at their optimum asphalt content were subjected to mr and dynamic triaxial (creep) at 22°c and 40°c, representing relevant temperatures that base layers are subjected to in the summer and winter and wheel racking tests at 22°c. 3.1 basic soils properties the dry and wet grain size distribution curves for marl, sabkha, and dune sand soils are shown in fig. 3. wet sieving of the sabkha was done with sabkha brine rather than distilled water in order to simulate field conditions. the investigated soils were non-plastic, based on atterberg limit tests. the specific gravity values of the marl, sabkha, and dune sand soils were 2.69, 2.71 and 2.63 kg/m3, respectively. h. i. al-abdul wahhb and g.m.s. abdullah 4 table 1. optimum asphalt contents of soils mixes. soil type additive type optimum asphalt (%) marl fa 9 fsa 9 ea 8.1 esa 7.2 sabkha fa 8 fsa 7 ea 4 esa 3.6 sand fa 7 fsa 8 ea 5.4 esa 5.4 figure 3. grain size distribution of soils. marl and sabkha soils were classified as sm and a3 according to the unified soil classification system (uscs) and american association of state highway and transportation officials (aashto) soil systems, respectively, based on both dry and wet sieving methods. however, dune sand was classified as sp and a-3 based on both dry and wet sieving methods. it is clear from dry and wet sieve analysis curves for marl and sabkha soils that the wet sieving curve is always above that of the dry curve. this is ascribed to the fact that water tends to dissolve the salts between particles of the soil; thus, the proportion of wet materials passing through a particular sieve is consistently more than that for dry sieving. this difference would be higher if sabkha was sieved with distilled water instead of sabkha brine. however, in sand soil, it can be seen that there is almost no variation between grain size distributions calculated by both the dry and wet sieving methods. this is ascribed to the fact that sand is made up of quartz which is not very affected by washing. prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 5 3.2 resilient modulus (mr) a dynamic triaxial test was conducted at 22°c and 40oc to measure the dynamic mr at these two temperatures for marl, sabkha, and sand mixed with fa, fsa, ea and esa. the dynamic mr was measured according to the aashto t-307 procedure. the specimens were tested under different combinations of confined pressure measuring 21–138 kpa (3–20 psi) and deviatoric stress measuring 34–276 kpa (5-40 psi) to simulate the traffic loading that the granular base and subbase materials are subjected to in the road structures. the variations of the mr with the deviator stress for all fa/fsa mixes at 22°c are presented in fig. 4. the results indicate that there is a significant effect of the deviator stress variation on mr for all mixes. the mr increased greatly with the increase in the deviator stress. these findings are consistent with the results reported by li and liu (2010). furthermore, the marl soil mixed with fa or fsa has the highest mr values, followed by sabkha and sand mixes. however, fsa mixes have lower mr values than the fa mixes. similarly, fa and fsa mixes were tested for the mr at 40oc. the relationships between the mr and the deviator stress are presented in fig. 5, which shows that, upon testing at 40 c, the mr values dropped by a magnitude of about 30% for fa mixes and 5–10% for fsa mixes, which is insignificant. the slight reduction in the mr of fsa mixes under an increased temperature as compared to fa may be attributed to the stiff bond between soil particles of the mix, which is improved due to the small-dispersed stiff droplets of fsa as a result of the foaming and mixing processes. at low ambient temperature, stiffness increases at the points where mineral particles are bounded by sulfur asphalt droplets while, where there were no sulfur asphalt droplets, the connection between particles did not change (li and liu 2010). thus, when temperatures increase, only the sulfur asphalt-bounded particles are affected, which are fewer in comparison to the mineral particles that are not bounded by sulfur asphalt droplets. furthermore, sulfur asphalt has a higher stiffness and less susceptibility to temperature differences as compared to plain asphalt, resulting in a stiffer mixes, which might justify its lower sensitivity to temperature variation. the variations of mr with the deviator stress at 22°c for emulsified soils mixes are presented in fig. 6, which clearly shows that the mr increased greatly with an increase in deviator stress. additionally, the marl soil treated with ea or esa has the highest mr, followed by the sabkha and lastly by the sand soils treated with the same stabilizers. the difference, however, is insignificant. figure 7 presents the variation of the mr, measured at 40 c, for the same soils mixed with ea and esa. again, there is a significant effect of deviator stress on the mr results. the mr increased significantly with the increase in the applied deviator stress; this is the same trend noticed for the mr measured at 22°c. the results also show that there is a drop in the mr values measured at 40°c compared to those measured at 22 c. this reduction in the mr values appeared to be small in the esa mixes compared to the ea mixes. thus, the sulfur modified ea mixes are less sensitive to the temperature effect which is a great advantage. in comparison, between the mr for the ea and esa mixes, the moduli for the marl-esa and sabkha-esa are slightly higher than those for the marl-ea and sabkha-ea mixes, whereas the inverse is true for the sand soil. 3.3 dynamic triaxial (creep) a dynamic triaxial repeated load test was conducted on specimens 100 mm in diameter by 200 mm in height prepared at the optimum water and fa, fsa, ea and esa contents and compacted to the maximum dry density. the specimens were tested under 10 psi (68.95 kpa) of confining pressure and a 40–80 psi range of deviator stress (276–552 kpa) to simulate the traffic loading that the granular base and subbase materials are subjected to in the roads. the deviatoric stress was applied in the form of a sinusoidal (haversine) wave pulse with a loading time of 0.1 second followed by a reset period of 0.9 seconds. repeated load tests were applied for treated mixes samples at lab temperatures of 22°c and 40°c and with deviator stress levels of 60, 70, and 80 psi (414, 483, and 552 kpa) for stiff materials and 40, 50, and 60 psi (276, 345, and 414 kpa) for less stiff materials. the results of the dynamic triaxial tests for the marl, sabkha and dune sand soils treated at the optimum percentages with fa and fsa and tested at 22°c are presented below (figs. 8–10). it is clear that the marl soil has a high resistance to rutting, followed by sabkha and sand soils. furthermore, the fsa mixes, except sand soil, reflect high rutting resistance in comparison to the fa mixes, which is considered a great advantage for newly investigated fsa. however, for the sand soil, the difference is not as high, and the permanent deformation is high either for the standard fa mixes or the modified fsa mixes. h.i. al-abdul wahhab and g.m.s. abdullah 6 figure 4. variation of mr with deviator stress for the foamed soils at 22°c. figure 5. variation of mr with deviator stress for foamed soils at 40°c. deviator stress, kpa deviator stress, kpa prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 7 figure 6. variation of mr with deviator stress for emulsified soils at 22°c. figure 7. variation of mr with deviator stress for emulsified soils at 40°c. deviator stress, kpa deviator stress, kpa h.i. al-abdul wahhab and g.m.s. abdullah 8 figure 8. dynamic triaxial test results for marl-fa/fsa at 22°c. figure 9. dynamic triaxial test results for sabkha-fa/fsa at 22°c. figure 10. dynamic triaxial test results for sand-fa/fsa at 22°c. load repetitions load repetitions load repetitionsload repetitionsload repetitions prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 9 the increase in the resistivity of the sulfur modified mixes to the permanent deformation may be ascribed to the increased asphalt stiffness. similarly, figs. 11–14 show the results of the dynamic triaxial test for the same soils tested at 40 c. the same trend as was found at 22°c is noticeable for the sabkha soil and marl, in which fsa mixes have higher resistance to permanent deformation. in addition to that and based on the results of the dune sand mixes shown in figs. 13 and 14, in spite of the weakness of the sand mixes, the fsa mix is more resistant to permanent deformation than the fa mix when tested at 40°c, particularly under low to medium loading stresses. figure 11. dynamic triaxial test results for marl-fa/fsa at 40°c. figure 12. dynamic triaxial test results for sabkha-fa/fsa at 40°c. load repetitions load repetitions h. i. al-abdul wahhb and g.m.s. abdullah 10 figure 13. dynamic triaxial test results for sand-fa at 40 c. figure 14. dynamic triaxial test results for sand-fsa at 40 c. figures 15–17 show the results of the dynamic triaxial tests of ea and esa mixes conducted at 22oc. contrary to the foamed mixes, the addition of the modified esa to the soils increased the possibility of rutting, especially for marl and sand soils. sabkha soil esa mixes showed less permanent deformation than the ea mixes, but they reached to the failure stage earlier than ea mixes, particularly at high stress levels. such behavior may be attributed to the reduction in the shear strength parameters (c and ф) of the soils when treated with esa rather than conventional ea. it should be noted that the internal friction angle of traded soils did not change significantly due to the treatment type while the cohesion has dropped considerably when esa was used instead of ea. figures 18–20 show the dynamic triaxial test results for the same three soils tested at 40°c. based on the results for the marl soil, the same trend found when testing at 22°c is clearly present [fig. 18]. here, the mixes remained in the second stage of the creep curve and did not reach to the failure stage as happened when testing at 22°c and under 80 psi (552 kpa) of loading, which means that sulfur mixes perform well at higher temperatures. on the other hand, sabkha and sand soils show the inverse trend. esa mixes are less sensitive to rutting in comparison to conventional mixes which were found to fail rapidly (figs. 19–20). load repetitions load repetitions prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 11 furthermore, the mixes did not reach the failure stage, particularly at low to medium loading for sand and for all levels of loading in the case of the sabkha soil. in addition, the slope (b) and intercept (a) of the permanent deformation curves are low which leads to a high and low µ for rutting model parameters, resulting in low rutting prediction. hence, the sulfur modified emulsion mixes have lower temperature susceptibility than conventional mixes. figure 15. dynamic triaxial test results for marl-ea/esa at 22°c. figure 16. dynamic triaxial test results for sabkha-ea/esa at 22°c. load repetitions load repetitions h. i. al-abdul wahhb and g.m.s. abdullah 12 figure 17. dynamic triaxial test results for sand-ea/esa at 22°c. figure 18. dynamic triaxial test results for marl-ea/esa at 40°c. figure 19. dynamic triaxial test results for sabkha-ea/esa at 40°c. load repetitions load repetitions load repetitions prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 13 figure 20. dynamic triaxial test results for sand-ea/esa at 40°c. 3.4 wheel tracking (wt) a wt test is considered one of the best tests to simulate field conditions. thus, rutting behavior of compacted base mixes was simulated using a wheel track tester. two slabs for each mix were compacted to their maximum dry density, which is the same density as in the dynamic triaxial test specimens, using dynamic compaction. from the marl, sabkha or sand mixed with fa, fsa, ea, and esa, 24 slabs measuring 45 cm x 22 cm x 10 cm were prepared. the slabs were then cured at 40°c for 72 hours for foamed mixes and at 60°c for 48 hours for emulsified mixes and tested dry under a wheel load of 80 psi (552 kpa) at 22°c. figure 21 shows samples of some tested slabs. the results shown in figs. 22–24 are the measured rut depths for the faand fsa-treated slabs of marl, sabkha, and sand soils, respectively. results clearly show the same rankings as were obtained from the dynamic triaxial test; marl soil had the highest rutting resistance followed by sabkha and sand soil. furthermore, the modified fsa mixes had less rutting susceptibility than the conventional foam mixes. results also showed that the sand treated with fa or fsa is very sensitive to rutting and exhibited higher permanent deformation (>15 mm within the first thousand load repetitions) compared to the other marl and sabkha mixes. the test section was considered failed when the vertical deformation was ≥25mm (1 inch). the results shown in figs. 25–27 are the measured rut depths for ea and esa mixes for marl, sabkha, and sand, respectively. again, the same trend and ranking as in the dynamic triaxial was observed in which marl soil has the highest rutting resistance, followed by sabkha and sand soils. this means that rutting can be predicted from any of the test methods and compared reasonably well with other. results also indicate that the esa mixes showed lower rutting resistance than the emulsion mixes. in comparison to either normal or sulfur modified foam mixes, emulsified mixes are less resistant to permanent deformation than foamed mixes. this can be attributed to the fact that the fa treatment mechanism is different from that of the ea treatment. while ea tends to coat soil particles with a very fine asphalt layer, fa works more like a soil stabilized with lime whereas fa mostly binds the finer particles of the soil, improving its structure. as such, fsa has a stiff bond between fine soil particles of the mix (droplets constituted of fsa). results also showed that sand treated with ea or esa is very sensitive to rutting and exhibited higher permanent deformation (>15 mm within the first thousand load repetitions) as compared to marl and sabkha mixes. generally, marl asphalt mixes are more resistant to permanent deformation than sand or sabkha mixes due to the higher internal friction angle of marl asphalt mixes, resulting in a longer secondary stage of permanent deformation. the low friction of dune and sabkha soils is attributed to the spherical nature of the particles. in general, the relative performance ranking of accelerated pavement tests (apt) represented here by a loaded wt test and a dynamic triaxial repeated load test was found to be the same for all investigated mixes in this study. load repetitions h. i. al-abdul wahhb and g.m.s. abdullah 14 figure 21. permanent deformation in wt samples. figure 22. results of the permanent deformation for the marl-fa/fsa mixes using a wheel track machine, dry at 22°c. figure 23. results of the permanent deformation for the sabkha-fa/fsa mixes using wheel track machine, dry at 22°c. load repetitions load repetitions prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 15 figure 24. results of the permanent deformation for the sand-fa/fsa mixes using a wheel track machine, dry at 22°c. figure 25. results of the permanent deformation for the marl-ea/esa mixes using a wheel track machine, dry at 22 c. figure 26. results of the permanent deformation for the sabkha-ea/esa mixes using wheel track machine, dry at 22 c. load repetitions load repetitions load repetitions h. i. al-abdul wahhb and g.m.s. abdullah 16 figure 27. results of the permanent deformation for the sand-ea/esa mixes using wheel track machine, dry at 22 c. 4. prediction of permanent deformation using the vesys model the vesys model includes two different flexible pavement rutting models: system and layer rutting. one advantage of the layer rutting model is its capability to predict both surface rutting and the permanent deformation in each layer of the pavement structure. the permanent deformation in each finite layer can be estimated as the product of the layer material permanent deformation law associated with that layer and the elastic compression at that layer which, in layer theory, is given by the difference in deflections of the top and bottom of the layer. thus, the rut depth in any finite layer can be calculated from the following equation (zhou and scullion 2002): rd (n) = (w+-w-)* n (1αi) (5) where, rd = the permanent deformation (rutting) level after n load repetitions; w+, w= the elastic deflection amplitudes of the top and bottom surfaces of the layer, respectively; µi, αi = the laboratory permanent deformation parameters for the each layer (i) material; μ , α = 1b (6) εr = (7) in this study, vesys 5w software was used to analyze the pavement structure and conduct performance analysis. the two main input parameters in the vesys 5w rutting model are the permanent deformation parameters, α and µ, for each layer. to calculate the mr of the stabilized mixes, regression models with high coefficients of correlation in terms of deviator stress, confining pressure, and temperature were developed using minitab software, version 16 (minitab, inc., state college, pennsylvania, usa) based on the results of the mr tests. tables 2 and 3 present the mr models for fa and ea mixes, respectively. the permanent deformation parameters ( and for each mix were obtained by representing the accumulated permanent strain versus the number of load repetitions, represented as a curve resulting from the dynamic triaxial test on a log-log scale. the intercept and slope coefficients were determined and then and µ were calculated according to eqn. 6. regression models were created using minitab software, version 16 to express these two parameters as functions of stress levels and temperature for foamed and emulsified mixes as shown in tables 4 and table 5, respectively. load repetitions prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 17 table 2. regression models of mr for fa mixes. material type type of additive mr r 2 marl fa mr = 260 6.47 t + 0.004 c + 4.92 d 0.989 fsa mr = 147 1.77 t 0.692 c + 4.00 d 0.982 sabkha fa mr = 256 6.16 t + 0.079 c + 4.51 d 0.986 fsa mr = 115 1.53 t 0.393 c + 3.82 d 0.988 sand fa mr = 135 3.57 t + 0.152 c + 4.36 d 0.977 fsa mr = 112 2.20 t 0.371 c + 3.88 d 0.991 mr = resilient modulus; fa = foamed asphalt; fsa = foamed sulfur asphalt. table 3. regression models of mr for ea mixes. material type type of additive mr r 2 marl ea mr = 128 2.26 t 0.513 c + 4.52 d 0.98 esa mr = 119 2.49 t 0.257 c + 4.61 d 0.993 sabkha ea mr = 158 3.07 t 0.346 c + 4.05 d 0.988 esa mr = 139 2.39 t 0.417 c + 4.25 d 0.99 sand ea mr = 151 3.21 t 0.093 c + 3.79 d 0.99 esa mr = 86.1 1.05 t 0.329 c + 4.01 d 0.992 mr = resilient modulus; ea = emulsified asphalt; esa = emulsified sulfur asphalt. table 4. regression models of μ and α for fa mixes. soil type stabilizer μ r 2 α r2 marl fa μ = 5.81 + 0.0161 t + 0.0133 d 0.97 α = 1.15 + 0.000037 t 0.000598 d 0.95 fsa μ = 1.53 + 0.0189 t + 0.00325 d 0.934 α = 1.04 0.00115 t 0.000334 d 0.974 sabkha fa μ = 3.80 + 0.0502 t + 0.0125 d 0.95 α = 0.805 + 0.000926 t 0.000181 d 0.962 fsa μ = 1.88 + 0.0204 t + 0.00776 d 0.882 α = 0.865 0.000741 t 0.000181 d 0.892 sand fa μ = 14.7 + 0.345 t + 0.0261 d 0.85 α = 1.93 0.0247 t 0.00154 d 0.902 fsa μ = 15.1 + 0.232 t + 0.0346 d 0.678 α = 1.37 0.00148 t 0.00149 d 0.505 fa = foamed asphalt; fsa = foamed sulfur asphalt. h. i. al-abdul wahhb and g.m.s. abdullah 18 table 5. regression models of μ and α for ea mixes. soil type stabilizer μ r2 α r2 marl ea μ = 0.113 -0.00005 t + 0.00113 d 0.914 α = 0.812 + 0.00326 t + 0.00004 d 0.949 esa μ = 10.9 + 0.037 t + 0.0278 d 0.826 α = 1.14 0.000296 t 0.000518 d 0.916 sabkha ea μ = 3.31 + 0.0247 t + 0.0137 d 0.867 α = 1.17 0.00444 t 0.000979 d 0.715 esa μ = 0.91 0.0719 t + 0.0157 d 0.811 α = 0.341 + 0.0146 t + 0.000036 d 0.915 sand ea μ = 5.63 + 0.107 t + 0.00897 d 0.714 α = 2.17 0.0280 t 0.00178 d 0.918 esa μ = -2.51 0.0454 t + 0.0124 d 0.706 α = 0.735 + 0.0181t 0.00123 d 0.895 ea = emulsified asphalt; esa = emulsified sulfur asphalt. table 6. calibration factors for α and µ. soil type treatment type αcf µcf marl fa 0.8904 1.058 fsa 0.9638 2.50 sabkha fa 1.0959 1.429 fsa 0.8267 0.351 sand fa 0.6604 1.16 fsa 0.7843 1.098 marl ea 0.7912 2.941 esa 0.9412 1.331 sabkha ea 1.321 1.77 esa 0.60 0.903 sand ea 0.93 1.191 a = foamed asphalt; fsa = foamed sulfur asphalt; ea = emulsified asphalt; esa = emulsified sulfur asphalt. where, and permanent deformation properties of the materials; t = temperature in degree d = deviator stress in kpa 4.1 validation and calibration of rut depth prediction models simulation in accelerated pavement tests (apt), such as through a wt test, is the most effective manner to simulate the impact of traffic loading and environmental conditions on pavement configurations. these tests are also known as the hamburg wt and used for the superpave evaluation process. the rutting prediction models derived in this study are based on the dynamic triaxial test, which is different from wheel tracking (wt) in many respects such as boundary conditions and traffic loading. however, both methods remain only a simulation of actual behavior of the material (jawad et al. 2013). it is necessary to verify and calibrate these models with the results of the wt test, which presents a better simulation of actual field conditions. figure 28 shows the procedure of the calibration process. first, the permanent deformation properties of the materials ( and µ) were calculated from the developed triaxial models at a temperature of 22 c and under a deviator stress of 552 kpa (similar to a wt test) for each mix. data were entered into vesys 5w software and rut depths were predicted. the predicted rut depths were compared with the rut depths measured using the wt tests. if the predicted rut depth was similar to the measured rut depth (within 90% of the calculated value), then it was determined that prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 19 there was no need for calibration and the model could be used for rutting prediction. otherwise, and µ were multiplied by the ratio of the recent predicted rut depth (rd+1) to the rut depth predicted in the previous step (rd) and called i+1 and μi+1 as shown in the flowchart. i+1 and μi+1 were re-entered into the vesys 5w software and the process was repeated until the predicted rut depth was within 90% of the wt-measured rut depth. finally, calibration factors for the permanent deformation properties ( and µ) for each mix were determined (table 6). it is worth mentioning that these factors were used only for the marginal soils (marl, sabkha, and dune sand) whose basic properties and classifications are listed in this paper. figures 29 and 30 show the wt-measured and vesys 5w-predicted rut depth curves for the marl treated with fsa and esa, respectively, as an example of the rutting prediction using calibrated models. it is clear from the figures that the calibrated models have predicted the rutting values close to that measured using a wt test. figure 28. model calibration flowchart. h. i. al-abdul wahhb and g.m.s. abdullah 20 figure 29. measured and predicted for marl-fsa mixes. figure 30. measured and predicted for marl-esa mixes. 5. conclusion this paper summarizes a laboratory investigation to evaluate the permanent deformation of marl, sabkha and dune sand soils treated with fsa and esa as well as conventional fa and ea by conducting dynamic triaxial and wt tests and developed and calibrated models for rutting prediction. based on the results of this study, the following conclusions are drawn: 1. the mr of fsa mixes are slightly less than those of fa mixes. however, the mr of esa mixes are slightly higher than ea mixes. 2. fsa mixes showed superior rutting resistance compared with fa mixes and less temperature susceptibility. 3. esa was found to increase the permanent deformation susceptibility of soil mixes. 4. permanent deformation prediction models for the investigated mixes were developed and calibrated with the wt test results, and the calibrated models were found to predict the rut depth with 90% accuracy. 5. the vesys 5w program is suitable to predict rut depth since it can predict the layer rut depth and the total rut depth of the pavement structure with a 90% degree of accuracy. prediction of permanent deformation of foamed and emulsified sulfur asphalt soils mixes 21 acknowledgment the authors would like to acknowledge the support provided by king fahd university of petroleum and minerals (kfupm) for the execution of this research. references abdullah and wahhab, gamil msa, hamad i. alabdul w (2015), evaluation of foamed sulfur asphalt stabilized soils for road applications. construction and building materials 88: 149–158. abdullah, gamil msa (2014), modeling the behavior of sulfur modified foamed and emulsified asphalt soils mixes for local road applications. ph.d dissertation, king fahd university of petroleum and minerals, dhahran, saudi arabia. asphalt emulsion manufacturers association (aema) (2004), a basic asphalt emulsion manual. manual series (ms) no. 19. asphalt institute (1996), superpave tm mix design. superpave series sp-2, lexington, kentucky, usa. barenberg, ernest jb, marshall rt (1990), calibrated mechanistic structural analysis procedures for pavements phase 1 & ii of nchrp project 1-26. national cooperative highway research program”. transportation research board. washington, dc. bevan ws (2002), development of flow number and flow time candidate simple performance test for asphalt mixtures. ms thesis. department of civil and environmental engineering, arizona state university. diyaljee, vishnu ad, gerald pr (1982), repetitive load deformation of cohesionless soil. journal of the geotechnical engineering division. proc. asce 108(10): 1215-1229. garba, rabbira g (2002), permanent deformation properties of asphalt concrete mixtures. phd thesis, norwegian university of science and technology. gholam ab, william oy (1996), determination of elastic and plastic subgrade soil parameters for asphalt cracking and rutting prediction. transp. res. rec. 1540, transportation research board, washington dc. 97-104. hafeez, imran h (2009), impact of hot mix asphalt properties on its permanent deformation behavior. 05uet/ phd-ce-22, university of engineering and technology, taxila. jawad h, douglas jw, theunis fph, david a (2013), comparing results between the repeated load triaxial test rand accelerated pavement test on unbound aggregate. journal of materials in civil engineering 26(3): 476–483. khedr, safwan ak (1986), deformation mechanism in asphaltic concrete. journal of transportation engineering asce112(1): 29-45. li, liu, peng l, juanyu l (2010), characterization of asphalt treated base course material. (no. fhwa-ak-rd-10-07). alaska university transport centre. alaska, usa. monismith, carl lm (1976), rutting prediction in asphalt concrete pavements. transp. res. rec. 616, transportation research board, washington, d.c. 2-8. vuong b, peter a (1991), repeated load triaxial testing on the subgrade from mulgrace alf site. australian road research board, wd ri91/023. wahhab, hamad ia, mirza mb, isam ma, hisham mk (2012), study of road bases construction in saudi arabia using foam asphalt. construction and building materials 26: 113–121. zhu and scullion, fujie z, thomas s (2002), vesys5 rutting model calibrations with local accelerated pavement test data and associated implementation. report no. fhwa/tx-03/91502-01-2, texas transportation institute, college station, tx. microsoft word paper 5.docx the journal of engineering research, vol. 11, no. 2 (2014) 50-63 finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars s.e. el-gamal department of civil & architectural engineering, sultan qaboos university, p.o. box 33, postal code 123, al-khodh, muscat, sultanate of oman received 10 june 2014; accepted 11 september 2014 abstract: due to their non-corrosive nature, high strength and light weight, fiber reinforced polymers (frp) are being widely used as reinforcement in concrete bridges, especially those in harsh environments. the current design methods of concrete deck slabs in most bridge design codes assume a flexural behavior under traffic wheel loads. the load carrying capacities of concrete bridge deck slabs, however, are greatly enhanced due to the arching action effect developed by lateral restraints. this study presents the results of a non-linear finite element (fe) investigation that predicts the performance of frp reinforced concrete (rc) deck slabs. the fe investigation is divided into two main parts: a calibration study and a parametric study. in the calibration study, the validity and accuracy of the fe model were verified against experimental test results of concrete slabs reinforced with glass and carbon frp bars. in the parametric study, the effect of some key parameters influencing the performance of frp-rc deck slabs bars was investigated. these parameters include the frp reinforcement ratio, concrete compressive strength, slab thickness and span-to-depth ratio. keywords: finite element analysis, concrete, bridge deck slabs, fiber reinforced polymers, reinforcement ratio, compressive strength, span to depth ratio. אאאאאאאא אא k א  אwאאא،אא א אאאאkאאאאאא א     א א     kאא        kא א  א א אאא א  אאא  א א אאאאאאאkאwאא אאאאkא،אאאאאא  א   אא  א א  א א kא אאא،אאאאאא אאאא אkאא א،א،א،אkא    אאwאא ،،א،אא،א{א، א،אkא corresponding author’s e-mail: sherif@squ.edu.om s.e. el-gamal 51 1. introduction one of the most common bridge designs used for highway reinforced concrete (rc) bridges is the slab-on-girder type. this type consists of a rc slab supported on a number of parallel longitudinal girders. different design codes (can/csa 2006; aashto 1996) specify the flexural design method for designing rc deck slabs. different experimental investigations and theoretical studies, however, have concluded that, when subjected to wheel loads, punching shear is the failure mode of restrained bridge deck slabs having a certain range of span/depth ratio. many researchers have reported that the punching strength of a concrete slab could be greatly enhanced by the arching action that results when its edges are restrained against movement. the magnitude of strength enhancement depends on the degree of lateral restraint; the stiffer the lateral restraint the greater the enhancement (fang et al. 1990; graddy et al. 2002; hewitt and batchelor 1975). consequently, the canadian highway bridge design code (chbdc) (can/csa-s6, 2006) allows under some conditions the use of an empirical method to design concrete deck slabs supported on girders. this method requires four layers of steel reinforcement with two at the top and two at the bottom. each layer has a steel reinforcement ratio of only 0.3%. the corrosion of steel reinforcement has typically led to significant deterioration and rehabilitation needs for rc structures in harsh environmental conditions. concrete bridge deck slabs are one of the main bridge components that rapidly deteriorate due to direct exposure to de-icing salts and the traffic environment. fiber reinforced polymer (frp) bars, with their corrosion resistance, light weight and high strength are a promising alternative to traditional steel reinforcement in bridge deck slabs (el-gamal et al. 2004; bouguerra et al. 2011; alsayed et al. 2012). due to the increased use of frp bars, especially in bridges, the chbdc allows their use in barriers, beams and slabs. however, the behavior of frp-rc deck slabs has not been fully explored and the current design of such slabs is still based on tests of deck slabs reinforced with steel. therefore, research studies and investigations on the behavior of such deck slabs are needed. an extensive experimental study has been carried out by the author to investigate the performance of deck slabs reinforced with several types and ratios of frp composite bars. although accurate and reliable results can be obtained from experimental tests, they are limited, expensive and time consuming. therefore, the finite element (fe) method is an attractive alternative method. this paper presents the results of a fe study that investigates the performance of frp-rc bridge deck slabs. the fe investigation is divided into a calibration study and a parametric study. the validity and accuracy of the fe model were verified against the experimental test results of two full scale concrete deck slabs reinforced with glass and carbon frp bars. in the parametric study, the effect of some key parameters that influence the behavior of bridge deck slabs reinforced with glass fiber reinforced polymer (gfrp) bars was investigated. these parameters include frp reinforcement ratio, concrete compressive strength, slab thickness and span-to-depth ratio. 2. finite element (fe) modeling 2.1 program capabilities fe modeling was conducted using the anatech concrete analysis program, version 3.0 (anatech corp., san diego, california, usa). the concrete material model was based on a smeared cracking methodology which states that when a crack occurs, the stress normal to the crack direction decreases to zero, resulting in redistribution of stresses around the crack. cracks are allowed to form in three directions and, once a crack forms, it may close and re-open but it cannot heal. the model includes residual tension stiffening algorithms for the gradual transfer of load to reinforcement during crack formation and for shear retention to simulate the effect of crack roughness through aggregate interlocking (mohamed and rizkalla 1999). the program accounts for the shear stiffness reduction due to cracking and also considers further decay as the crack opens. the program models the reinforcement as individual sub-elements within the concrete elements. the anchorage loss is modeled as an effective stiffness degradation of the rebar as a function finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 52 of the concrete strain normal to the rebar. more details about the program can be found in the anacap concrete analysis program theory manual (james 2004). 2.2 validation of the model 2.2.1 description of test specimens the finite element model (fem) was calibrated using experimental tests conducted by the author (el-gamal 2005). through this program, several full scale bridge deck slabs (3,000 mm long × 2,500 mm wide × 200 mm deep) were constructed and tested. different types and ratios of frp bars were used in the slabs. the test results of two slabs were used for validation of the fem. the two slabs contained identical gfrp reinforcement in all directions (no.16 @ 200 mm) except in the bottom short direction where one slab (gfrp) was reinforced with no.19 (db=19 mm) gfrp bars and the other slab (cfrp) was reinforced with no.10 (db=9.5 mm) cfrp bars. table 1 shows the reinforcement details and concrete compressive strength of the two slabs. the slabs were constructed using a ready-mixed concrete with an average 28-day concrete compressive strength of 45 mpa. sand-coated glass and carbon frp bars were used. the mechanical properties of the bars were determined by performing tensile tests on frp specimens according to aci440 (2006). the properties of the bars used are listed in table 2. the slabs were tested under a concentrated load with an area of 600 mm × 250 mm to represent the foot print of a truck wheel (cl-625 truck, as specified by clause 3.8.3.1 in the chbdc) acting on the center of the slab. the slabs were supported on steel girders using steel bolts. steel angles were used to prevent the lateral movement of the girders (fig. 1). more details about test specimens, instrumentation and test set-up can be found in (el-gamal 2005; el-gamal et al. 2005a, 2007). 2.2.2 description of the finite element model (fem) due to symmetry, one-quarter of the slab and supporting elements was modeled. the slab thickness was divided into three unequal thickness layers. the thicknesses of the layers were selected to produce a finer mesh in the compression zone near the top surface of the slab. a finer mesh with smaller element sizes was also used within the loading area (fig. 2). twenty-node iso-parametric brick elements were used to model the supporting steel girders, concrete slab and top steel channel. the cross frames were modeled using spring element with stiffness equal to the axial stiffness of the steel angles. reinforcing bars and steel bolts were modeled using embedded bar elements. the load was applied as a distributed load over the loaded area incrementally. figure 3 shows the finite element mesh and reinforcement of one quarter of the deck slab. at the expected cracking and failure load levels, small load increments were used to increase the accuracy of the analysis and to closely capture the behavior. in addition to the material properties of concrete and reinforcing bars given before, the modulus of elasticity and cracking strength of concrete was calculated using eq. 1 and 2, respectively. 5.1' )2300/)(69003000( ccc fe γ+= (can/csa-s6 2006) (eq.1) '4.0 cr ff = (eq.2)  poission’s ratio of concrete, νc, was assumed to be 0.2.  poission’s ratio of steel, νs, was assumed to be 0.3.  poission’s ratio of frp bars, νfrp, was assumed to be 0.25. 2.3 comparison between the fem model and experimental results table 3 summarizes the predicted cracking and failure loads, deflections and reinforcement strains using the fem compared to the experimental results of the two deck slabs (cfrp and gfrp). it can be noticed that the fem predicted well the cracking and punching failure loads, load-deflection behavior and loadstrain behavior. the two frp-reinforced deck slabs failed in punching shear. at failure, the concrete compressive strain at a distance of d/2+y from the center of the loaded area reached a value of 1,900 micro-strains, where d is the diameter of the idealized load area which is defined as a s.e. el-gamal 53 circle that has the same perimeter as the rectangular loaded area, y is defined as the location of the center of rotation of concrete wedges, causing punching which is approximately 0.1 of the thickness of the slab measured from the center of the load (hassan et al. 2004). this distance and strain level (1,900 micro-strains) have been considered by several researchers (kinnunen and nylander 1960; mufti and newhook 1998; mohamed and rizkalla 1999; hassan et al. 2004) as an indication of failure due to concrete crushing, which results in punching shear failure. table 1. details of slab reinforcement and concrete compressive strength. slab f′c (mpa) ρact (%) of bottom main rein. axial stiffness (ρact × e) (n/mm2) reinforcement configuration bottom short (main) all other directions gfrp 49.1 1.2 528 no.19 @ 150 mm no. 16 @ 200 mm cfrp 49.6 0.34 415 no. 10 @ 125 mm no. 16 @ 200 mm table 2. the mechanical properties of the fiber reinforced polymer (frp) bars. bar type diameter (mm) area (mm2) modulus of elasticity (gpa) tensile strength (mpa) ultimate strain (%) gfrp no. 16 15.9 198.6 44.6 ± 0.8 727 ± 9 [700]+ 1.65 ± 0.03 [1.56]! no. 19 19.0 283 44.5 ± 1.3 637 ± 15 [592]+ 1.37 ± 0.03 [1.28]! cfrp no. 10 9.5 71.0 122 ± 2.4 1444 ± 18.0 [1390]+ 1.23 ± 0.07 [1.02]! +guaranteed tensile strength = average tensile strength minus three times the standard deviation !guaranteed tensile strain = average tensile strain minus three times the standard deviation figure 1. test set-up (el-gamal 2005). bolts 25 mm @ 250 mm 630 3 cross frames @ 1500 mm 2000 neoprane pads 250 250 410 (2 l 55 55 6) 326 5003000500 support support i beam slab 27 section a-a section b-b steel channel 200 bolts 25 mm @ 250 mmhydraulic jack (capacity 1500 kn) 10-mm thick neoprane bad 3-mm thick neoprane pad plate 600 x 250 x 60 mm support b b a a support finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 54 2.3.1 cracking and failure loads the cracking and failure load predictions from the fem compared to the experimental measures of the two deck slabs are shown in table 3. it shows that the fea predicted with high accuracy (>96%) the cracking load of the two deck slabs. the predicted failure loads were 675 and 695 kn for cfrp and gfrp slabs, respectively. this indicates that the fem predicted with good accuracy (95 and 100.7%) the punching capacity of the two deck slabs. figure 2. mesh dimensions using a threedimensional 20-node brick element model. (a) finite element mesh (b) model showing the reinforcement figure 3. the finite element model (fem) of one quarter of the slab and the set-up. 2.3.2 load-deflection and strains relationships figures 4 and 5 show the plots of load versus deflection, reinforcement strains and concrete strains from the fem compared with the experimental results. good agreement can be noted between the obtained fe predictions and experimental results for deflection and strains. before cracking, the fea predictions were slightly lower than those from the experimental tests. table 3. summary of experimental and fea results. 125 287.5 287.5 287.5 287.5 37.2 61.4 101.4 125 100 250 250 250 250 250 125 125 87 82 73 16 73 82 21 8. 5 21 8. 5 10 0 10 0 10 0 10 0 27 576 27 applied load27 slab cracking load (kn) failure load(kn) max. deflection (mm) max. strains in reinforcement (micro-strain) ultimate design load (208.25 kn) failure ultimate design load (208.25 kn) failure exp fea exp fea exp fea exp fea exp fea exp fea cfrp 115 115 674 679 3.67 3.62 22.60 22.87 2367 2651 8500 8210 gfrp 115 115 732 695 3.47 3.69 21.20 21.41 2270 2168 8080 7458 s.e. el-gamal 55 this may be attributed to the existence of some micro cracks in the slabs due to concrete shrinkage, which can reduce the stiffness of the slabs. however, after cracking until failure, the fem predictions matched the test results very well. for the cfrp-rc deck slab, the maximum predicted deflection at factored design load (206.25 kn according to the chbdc) and at failure were 3.67 and 22.87 mm, respectively, which were about 99% and 101% of the measured values (3.62 and 22.6 mm), respectively. the maximum predicted reinforcement strains were 2,651 and 8,210 micro-strains at factored design load and at failure, respectively, which were about 111% and 97% of the measured values (2,367 and 8,500 micro-strains). for the gfrp-rc deck slab, the maximum predicted deflection at the factored design load and at failure were 3.69 and 21.41 mm, respectively, which were about 106% and 101% of the measured values (3.47 and 21.2 mm), respectively. the maximum predicted reinforcement strains were 2,161 and 7,458 micro-strains at the factored design load and at failure, respectively, which were about 96% and 93% of the measured strain values (2,270 and 8,080 micro-strains). (a) load-maximum deflection. (b) load-max. reinforcement strains. (c) load-maximum concrete strains. figure 4. comparison of fem and experimental results of the cfrp-reinforced slab. (a) load-maximum deflection. (b) load-max. reinforcement strains. (c) load-maximum concrete strains. figure 5. comparison of fem and experimental results of the gfrp-reinforced slab. 0 100 200 300 400 500 600 700 800 0 5 10 15 20 25 30 l o a d (k n ) deflection (mm) exper. (cfrp) fem (cfrp) 0 100 200 300 400 500 600 700 800 0 1500 3000 4500 6000 7500 9000 l o a d (k n ) strains (micro-strain) exper. (cfrp) fem (cfrp) 0 100 200 300 400 500 600 700 800 -2500 -2000 -1500 -1000 -500 0 l o a d (k n ) strains (micro-strain) exper. (cfrp) fem (cfrp) 0 100 200 300 400 500 600 700 800 0 5 10 15 20 25 30 l o a d (k n ) deflection (mm) exper. (gfrp) fea (gfrp) 0 100 200 300 400 500 600 700 800 0 1500 3000 4500 6000 7500 9000 l o a d (k n ) strains (micro-strain) exper. (gfrp) fem (gfrp) 0 100 200 300 400 500 600 700 800 -2500 -2000 -1500 -1000 -500 0 l o a d ( k n ) strains (micro-strain) exper. (gfrp) fem (gfrp) finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 56 2.3.3 cracks propagation the crack propagations at different load levels are shown in fig. 6. the figure shows that the predictions from the fem match the behavior from the experimental tests. the predictions show that the first cracks occurred under the load and propagated longitudinally (fig. 6a). by increasing the load, other cracks propagated in the radial direction (fig. 6b). on the top face of the slabs, the first crack occurred above the supporting beams at high load levels (about 480 kn) (fig.6c). at failure, all the bottom surface of the deck slabs cracked, especially under the loaded area, which is evidence of punching failure [fig. 6d]. these cracking propagation stages were in close agreement with that obtained from the experimental test. 3. parametric study following the verification of the fem, the developed model was used to study the effect of key parameters that influence the behavior of frp-rc bridge deck slabs. these parameters included the frp reinforcement ratio, concrete compressive strength, slab thickness and spanto-depth ratio. 3.1 frp reinforcement ratio among the different types of frp bars, glass frp composite bars have recently drawn much attention as reinforcement for concrete infrastructures due to their low cost compared to the other available types (carbon and aramid) allowing their wide use in bridge deck slabs. the experimentally tested slab gfrp (second slab in table 3), was considered a reference specimen for fems. the concrete dimensions, concrete covers, boundary and loading conditions were kept unchanged. the only parameter was the gfrp reinforcement ratio in different directions. five different reinforcement configurations were investigated in this study (table 4). 3.1.1 effect of bottom transverse reinforcement three different reinforcement ratios for the bottom transverse reinforcement were investigated as shown in table 4 (cases 1, 2 and 3). the three cases have the same reinforcement in all directions (0.6%) except in the main bottom direction, which has three different reinforcement ratios (1.2%, 0.9% and 0.6%). the first case represents almost the same bottom transverse reinforcement specified by the chbdc (can/csa 2006). cases 2 and 3 represent lower bottom transverse reinforcement ratios. the predicted load-deflection curves of the three slabs are shown in fig. 7a. the predicted deflection at service load level ranged between 0.4 and 0.44 mm, which is well below the allowable code limits (span/800). the maximum predicted deflections at failure were 21.3, 25.5 and 27.45 mm for cases 1, 2 and 3, respectively. this means that decreasing the reinforcement ratio of the main bottom reinforcement to half increased the maximum deflection by about 9% and 22% at service and ultimate load, respectively. the predicted load-maximum reinforcement strains in all three cases are shown in fig. 7b. the predicted reinforcement strains at service load were 215, 329 and 398 micro-strains for cases 1, 2 and 3, respectively. at failure, these strains were 7,538, 8,444 and 10,930 microstrains, respectively. this means that the maximum strains increased by about 85% and 45% at service and ultimate load, respectively, when the bottom transverse reinforcement was decreased to half. this indicates that the bottom transverse reinforcement ratio had a significant effect on the frp reinforcement strains. thus, decreasing the bottom main reinforcement ratio by 1.2 %, which is allowed by the chbdc, is not recommended until approved under different spans and loading conditions. all the slabs failed by concrete crushing resulting in punching shear. the ultimate load values were 692, 687 and 674 kn for cases 1, 2 and 3 , respectively, which indicates that the punching capacity decreased by about 2.6% as a result of decreasing the main bottom reinforcement by 50%. this indicates that the bottom main reinforcement had an insignificant effect on the punching capacity of frp-rc bridge deck slabs. 3.1.2 effect of top reinforcement three different reinforcement ratios for the top reinforcement in both directions were investigated as listed in table 4 (cases 1, 4 and 5). s.e. el-gamal 57 table 4. slab configurations (effect of reinforcement ratio). case no. bottom reinforcement top reinforcement transverse longitudinal (both directions) 1 no. 19@150 mm (1.2%) no. 16@200 mm (0.6%) no. 16@200 mm (0.6%) 2 no. 19@200 mm (0.9%) 3 no. 16@200 mm (0.6%) 4 no. 19@150 mm (1.2%) no. 16@200 mm (0.6%) no. 12@200 mm (0.36%) 5 --------- (0.00%) (a) just after cracking (120 kn). (b) at 220 kn. (c) at 480 kn. (d) at failure load (695 kn). figure 6. typical crack propagation for frp-rc deck slabs (gfrp slab). (a) load-maximum deflection. (b) load-maximum reinforcement strains. figure 7. the effect of bottom transverse reinforcement. 0 100 200 300 400 500 600 700 800 0 5 10 15 20 25 30 35 l o a d (k n ) deflection (mm) case 1 (1.2 %) case 2 (0.9 %) case 3 (0.6 %) 0 100 200 300 400 500 600 700 800 0 2000 4000 6000 8000 10000 12000 l o a d (k n ) strain (micro-strain) case 1 (1.2 %) case 2 (0.9 %) case 3 (0.6 %) finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 58 (a) load-maximum deflection. (b) load-maximum concrete compressive strains. figure 8. effect of top reinforcement the top reinforcement in both directions had three different reinforcement ratios of 0.6%, 0.34% and 0.0% for cases 1, 4 and 5, respectively. the predicted load-deflection curves and loadmaximum compressive concrete strains under the loaded area of the three slabs are shown in fig. 8. the three slabs gave almost the same behavior. all the slabs failed in punching shear failure. the concrete compressive strains under the loaded area reached 4,700 micro-strains which were 55% higher than the maximum compressive strain of concrete. this can be related to concrete confinement under the loaded area. the ultimate load values were 693, 684.5 and 679.5 kn for cases 1, 4 and 5, respectively, which indicates that without top reinforcement, the punching capacity decreased by only 0.6%. this means that the top reinforcement had almost no effect on the overall behavior of restrained frp-rc bridge deck slabs. 3.2 concrete compressive strength concrete compressive strength is a major factor, if not the most important, in all punching shear code equations and design provisions. most of the punching shear code equations (aci 318-08 2008; csa-a23.3 1994; jsce 1997; elgamal et al. 2005b) are based on the assumption that the punching shear is proportional to the square root of concrete compressive strength. other codes and design provisions (bs8110 1997) assume that the punching shear is proportional to the cubic root of concrete compressive strength rather than the square root. in addition, these code equations are based on the test results of slabs made with relatively low compressive strength, varying mostly from 14 to 40 mpa (marzouk and hussien 1991). to investigate the effect of concrete compressive strength on the behavior of gfrp-rc deck slabs, different concrete compressive strengths were investigated. the dimensions, concrete covers, reinforcement and boundary and loading conditions were kept unchanged as in the gfrp slab. only the concrete strength was changed from 20 to 80 mpa (in increments of 10 mpa), consequently, the modulus of elasticity and cracking strength of concrete were changed and calculated using eqs. 1 and 2, respectively. 3.2.1 deflection and cracking behavior the load-deflection behavior using different concrete compressive strengths is given in fig. 9a. increasing the concrete compressive strength increased the cracking load, which was 85 kn for the slab with a concrete compressive strength of 20 mpa. it increased by 30%, 52% and 88% when the concrete strengths increased to 40, 60 and 80 mpa, respectively. after cracking, linear deflection behavior continued until failure. all the load-deflection curves had the same slope; however, increasing the concrete strength decreased the deflections as a result of different cracking loads. at the factored design load, the predicted deflections were 6.49, 4.73, 3.5 and 2.37 mm for slabs with concrete compressive strengths of 20, 40, 60 and 80 mpa. 3.2.2 reinforcement strains the predicted curves of load-maximum bottom transverse reinforcement strains for different concrete compressive strengths are 0 100 200 300 400 500 600 700 800 0 5 10 15 20 25 30 35 l o a d (k n ) deflection (mm) case 1 (0.6 %) case 4 (0.36 %) case 5 (0.0 %) 0 100 200 300 400 500 600 700 800 -5000 -4000 -3000 -2000 -1000 0 l o a d (k n ) concrete strain (micro-strain) case 1 (0.6 %) case 4 (0.36 %) case 5 (0.0 %) s.e. el-gamal 59 shown in fig. 9b. it can be noticed that increasing the concrete compressive strength decreased the predicted reinforcement strains before cracking. a sudden increase in the reinforcement strains was noted just after cracking and then the slabs behaved linearly up to failure. all the slabs had almost similar reinforcement strains after cracking. at the factored design load of 208.25 kn, the maximum predicted reinforcement strains were 2,515, 2,485, 2,977 and 2,289 micro-strains for slabs with concrete strengths of 20, 40, 60 and 80 mpa, respectively. this indicates that the maximum reinforcement strains at the factored design load decreased by about 5.3%, 10% and 16.4% when the concrete compressive strength increased to 40, 60 and 80 mpa. at a load level of 445 kn (the failure load of the slab with the minimum concrete compressive strength of 20 mpa), the maximum predicted reinforcement strains were 5,530, 5,365, 5,484 and 5,671 micro-strains (almost equal values) for slabs with concrete strengths of 20, 40, 60 and 80 mpa, respectively. therefore, it can be concluded that the concrete compressive strength has some effect on the bottom transverse reinforcement strains before cracking; however, after cracking it has an insignificant effect. 3.2.3 failure load the effect of concrete compressive strength on the punching shear capacity is illustrated in fig. 9c. the punching capacity increased by increasing the concrete strength. the failure load was found to be 445 kn for the slab with a concrete strength of 20 mpa. it can be noted that the values of the failure loads increased by about 39%, 75% and 95% when the concrete strength was increased to 40, 60 and 80 mpa, respectively. the failure loads did not increase linearly with respect to the concrete strength. the trend line of the results in fig. 9c shows that the punching capacity is proportional to the square root of the concrete compressive strength as concluded by the north american codes. 3.3 deck slab thickness the effective depth of the concrete slab and, consequently, the slab thickness is a major factor that affects the punching capacity. all current design codes (aci 318-08 2008; csa-a23.3 1994; jsce 1997; bs8110 1997) take into account the effect of slab thickness on the punching shear capacity. in order to investigate the effect of deck thickness on the behavior of frp-rc bridge deck slabs, five thicknesses were used (150, 175, 200, 225 and 250 mm). all the slabs had 2000 mm span. the thicknesses studied provided a span/depth ratio ranging between 8 and 13.3. (a) load-maximum deflection. (a) load-maximum reinforcement strains. (b) punching capacity. figure 9. effect of concrete compressive strength. 0 100 200 300 400 500 600 700 800 900 1000 0 5 10 15 20 25 30 35 40 l o a d (k n ) deflection (mm) fc=20 mpa fc=40 mpa fc=60 mpa fc=80 mpa 0 100 200 300 400 500 600 700 800 900 1000 0 2000 4000 6000 8000 10000 l o a d (k n ) strains (micro-strain) fc=20 mpa fc=40 mpa fc=60 mpa fc=80 mpa 445 535 620 697 780 840 870 y = 98.077x0.5021 300 400 500 600 700 800 900 1000 10 20 30 40 50 60 70 80 90 100 l o a d (k n ) concrete compressive strength (mpa) fea power (fea ) finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 60 the model shown in figs. 2 and 3 was used for the five slabs. for all cases, the slab thickness was divided into three layers. the thickness of the layers was selected to produce a finer mesh in the compression zone near the top surface of the slab to increase the accuracy of the analysis. a concrete compressive strength of 40 mpa was used for all cases. all other parameters (reinforcement, concrete covers, loading and boundary conditions) were kept the same as that of the gfrp slab. 3.3.1 deflection and cracking behavior the predicted load-deflection curves of the five cases are given in fig. 10a. it can be noticed that the deflections decreased with increasing deck thickness as a result of the significant increase in the stiffness of the slabs. at the service load level, the maximum predicted deflections were 3.8, 2.1, 1.48, 0.29 and 0.21 mm for slabs with thicknesses of 150, 175, 200, 225 and 250 mm, respectively. this indicates that only the slab with the 150 mm thickness (span/depth = 13.33) has higher deflections than those allowed by the aashto code (span/800 = 2.5 mm) (aashto 1996). the predicted cracking load was 75 kn for the 150 mm thick deck slab. the cracking load increased by about 13%, 47%, 80% and 127% when the slab thickness increased to 175, 200, 225 and 250 mm, respectively. 3.3.2 reinforcement strains the predicted load-maximum bottom transverse reinforcement strains behavior using different slab thicknesses is shown in fig. 10b. it can be observed that the reinforcement strains decreased by increasing the slab thickness. at the factored design load, the maximum predicted reinforcement strains were 3,145, 2,765, 2,485, 2,050 and 1,259 micro-strains (about 24 to 10% of the ultimate strain of the gfrp bars) for slabs of 150, 175, 200, 225 and 250 mm thickness, respectively. this indicated that, at the factored load, the maximum bottom transverse reinforcement strains of the slab with the 150 mm thickness decreased by about 12, 21, 34 and 60% by increasing the slab thickness to 175, 200, 225 and 250 mm, respectively. the maximum predicted reinforcement strains at the failure load of each slab were very close and ranged between 6,890 and 7,410 micro-strains (about 55% of the ultimate strain of the gfrp bars). this indicates that all the slabs reached the same level of strains at failure regardless of the slab thicknesses. 3.3.3 failure load the effect of slab thickness on the punching shear capacity is illustrated in fig. 10c. as the slab thickness increased, both the effective depth and the critical perimeter increased, resulting in higher failure loads. the failure load was 476 kn in the 150 mm thick. it can be observed that the values of the failure load increased by about 13%, 30%, 53% and 71% when the slab thickness increased to 175, 200, 225 and 250 mm, respectively. the failure loads increased almost linearly with respect to the slab thickness as shown by the trend line in fig. 10c. 3.4 span-to depth ratio the punching shear behavior of restrained concrete deck slabs reinforced with steel has been studied by many researchers (graddy et al. 2002; fang et al. 1990; hewitt and batchelor 1975). they concluded that punching shear failure is the mode of failure of restrained bridge deck slabs with a span/depth ratio of 9 to 15. in order to investigate the effect of span/depth ratio on the overall behavior and failure mode of frp-rc bridge deck slabs, six span/depth ratios were investigated (8, 10, 12, 14, 16 and 18). all the slabs had the same depth of 200 mm. the investigated span/depth ratios provided spans ranging between 1,600 and 3,800 mm. a concrete compressive strength of 40 mpa was used for all cases. all other parameters (dimensions, reinforcement, concrete covers, loading and boundary conditions) were kept the same as that of slab gfrp. 3.4.1 deflection and cracking behavior the predicted load-deflection curves of the six cases are shown in fig. 11a. increasing the span-to-depth ratio decreased the cracking loads of the deck slabs. the predicted cracking load was 125 kn for the deck slab with a span-todepth ratio of eight. the cracking load decreased by about 12, 20, 32, 40 and 46% when the spanto-depth ratio increased to 10, 12, 14, 16 and 18, respectively. s.e. el-gamal 61 (a) load-maximum deflection. (b) load-maximum reinforcement strains. (c) punching capacity. figure 10. effect of slab thickness. it can be also noticed that the deflections increased by increasing the span/depth ratio as a result of the increase in the bending moments applied on the deck slabs. 3.4.2 reinforcement strains the predicted load-maximum bottom transverse reinforcement strains behavior using different span-to-depth ratios is shown in fig. 11b. the reinforcement strains increased as the span/depth ratio increased. at the factored design load, the maximum predicted reinforcement strains were 1,795, 2,485, 2,665, 3,745, 4,025 and 4,395 micro-strains (about 13% to 32% of the ultimate strain of gfrp bars) for slabs with span/depth ratios of 8, 10, 12, 14, 16 and 18, respectively, which corresponded to spans of 1.6, 2, 2.4, 2.8, 3.2 and 3.6 m, respectively. this indicates that, at the factored design load, increasing the span-to-depth ratio by 25, 50, 75, 100 and 125% (from 10 to 18) increased the bottom transverse reinforcement strains by about 38, 48, 108, 125 and 144%, respectively. the maximum predicted reinforcement strains at failure were 5,220, 6,925, 7,783, 8,504, 9,499 and 9,741 micro-strains, respectively. these values are about 40 to 75% of the ultimate strain of the gfrp bars. 3.4.3 failure load the effect of span/depth ratio on the punching shear capacity is illustrated in fig. 11c. as the span/depth ratio increased, the bending moments increase resulting in higher concrete compressive strains and lower failure loads. all the slabs failed due to concrete crushing, leading to punching shear failure. the failure load was found to be 700 kn of the slab with a span/depth ratio of eight. it was observed that the values of the failure load decreased by about 11.5%, 20%, 25%, 30% and 36% when the spanto-depth ratio increased to 10, 12, 14, 16 and 18, respectively. figure 11c shows that the failure loads did not decrease linearly with respect to the span/depth ratio. this may be attributed to the arching action in bridge deck slabs. for lower span-to depth ratios, the induced in-plane compressive membrane forces are higher, leading to concrete confinement. consequently, the failure loads increased. (a) load-maximum deflection. 0 100 200 300 400 500 600 700 800 900 1000 0 5 10 15 20 25 30 35 l o a d (k n ) deflection (mm) 150 mm 175 mm 200 mm 225 mm 250 mm allowable deflection service load level 0 100 200 300 400 500 600 700 800 900 1000 0 2000 4000 6000 8000 10000 12000 l o a d (k n ) strains (micro-strain) 150 mm 175 mm 200 mm 225 mm 250 mm 476 541 620 730 816 y = 2.1233x1.0761 300 400 500 600 700 800 900 1000 125 150 175 200 225 250 275 l o a d (k n ) slab thickness (mm) fea power (fea ) 0 100 200 300 400 500 600 700 800 0 10 20 30 40 50 60 70 l o a d (k n ) deflection (mm) s/d=8 s/d=10 s/d=12 s/d=14 s/d=16 s/d=18 finite element analysis of concrete bridge slabs reinforced with fiber reinforced polymer bars 62 (b) load-maximum reinforcement strains. (c) punching capacity. figure 11. effect of span to depth ratio. 4. conclusions this paper presents a non-linear fea investigation of frp-rc bridge deck slabs to study their behavior at different load levels. the validity and accuracy of the fea were verified against experimental test results conducted by the author. using the fea, the effect of some key parameters that influence the behavior of frp-rc bridge deck slabs was investigated. these parameters included the frp reinforcement ratio, concrete compressive strength, slab thickness and span-to-depth ratio. based on the results of this study, the following points can be concluded: 1. the fe analysis conducted using the anacap program was capable of predicting the behavior, punching capacity and mode of failure of frp-rc bridge deck slabs. 2. the bottom main reinforcement had an insignificant effect on the punching strength of restrained bridge deck slabs; however, it had a small effect on the deflection behavior. 3. the top reinforcement had a negligible effect on the deflection behavior, strains and punching capacity of deck slabs. 4. the concrete compressive strength, slab thickness and span-to-depth ratio had a significant effect on the overall behaviour of frp-reinforced bridge deck slabs. 5. increasing the concrete strength decreased deflections and increased the punching capacity. the fem proved that the punching capacity is proportional to the square root of the concrete compressive strength, which is in agreement with several design codes. 6. increasing the slab thickness significantly increased the punching capacity and decreased the deflections and reinforcement strains. at the factored load, the maximum bottom transverse reinforcement strains of slab with a 150 mm thickness decreased by about 12, 21, 34 and 60% by increasing the slab thickness to 175, 200, 225 and 250 mm, respectively. the failure loads increased almost linearly with respect to the deck slab thicknesses. 7. increasing the span-to-depth ratio from 8 to 18 significantly decreased the punching capacity and increased the deflections and reinforcement strains. at the factored design load, increasing the span-to-depth ratio by 25, 50, 75, 100 and 125% (from 8 to 10, 12, 14, 16 and 18) increased the bottom transverse reinforcement strains to by about 38, 48, 108, 125 and 144%, respectively. acknowledgment thanks are due to the anonymous referees for their valuable comments, which improved the quality of the paper. references aashto (1996), standard specifications for highway bridges. american association of state highway and transportation officials, washington d.c. aci 440.1r-06 (2006), guide for the design and construction of concrete reinforced with frp 0 100 200 300 400 500 600 700 800 0 2000 4000 6000 8000 10000 l o a d (k n ) strain (micro strain) s/d=8 s/d=10 s/d=12 s/d=14 s/d=16 s/d=18 700 620 565 522 488 448 y = 2143.3x-0.537 300 400 500 600 700 800 6 8 10 12 14 16 18 20 l o a d (k n ) span to depth ratio fea power (fea) s.e. el-gamal 63 bars. american concrete institute, farmington hills, michigan, usa. aci 318-08 (2008), building code requirements for reinforced concrete. american concrete institute, farmington hills, michigan 427. alsayed s, al-salloum y, almusallam t, elgamal s, aqel m (2012), performance of glass fiber reinforced polymer bars under elevated temperatures. composites part b 43:2265– 2271. bouguerra k, ahmed e, el-gamal s, benmokrane b (2011), testing of full-scale concrete bridge deck slabs reinforced with fiber reinforced polymer (frp) bars. construction and building materials journal 25:3956–3965. british standards institution (1997), structural use of concrete, bs 8110: part 1code of practice for design and construction, london. csa-a23.3 m-94 (1994), design of concrete structures for buildings. canadian standards association, rexdale, ontario. can/csa-s6 (2006), canadian highway bridge design code. chbdc, canadian standard association, rexdale, ontario, canada. el-gamal s, el-salakawy e, benmokrane b (2004), behaviour of frp reinforced concrete bridge decks under concentrated loads. proceedings of the 4th international conference on advanced composite materials in bridges and structures, calgary, alberta, canada. el-gamal s (2005a), behavior of restrained concrete bridge deck slabs reinforced with reinforcing bars under concentrated load. ph.d. thesis, sherbrooke (quebec): department of civil engineering, université de sherbrooke, canada. el-gamal s, el-salakawy e, benmokrane b (2005a), behavior of concrete bridge deck slabs reinforced with frp bars under concentrated loads. aci structural journal 102(5):727-735. el-gamal s, el-salakawy e, benmokrane b (2005b), a new punching shear equation for two-way concrete slabs reinforced with frp bars. aci special publication sp-230(50):877894. el-gamal s, el-salakawy e, benmokrane b (2007), influence of reinforcement on the behavior of concrete bridge deck slabs reinforced with frp bars. asce, journal of composites for constructions 11(5):449-458. fang i, worley j, burns n, klingner r (1990), behavior of isotropic reinforced concrete bridge decks on steel girders. journal of composites of construction, asce 116(3): 659-678. graddy j, kim j, whitt j, burns n, klingner r (2002), punching shear behavior of bridge decks under fatigue loading. aci structural journal 99(3):257-266. hassan t, rizkalla s, abdelrahman a, rochelle r (2004), frp bars for concrete bridge decks. proceedings of the 4th international conference on advanced composite materials in bridges and structures, calgary, alberta, canada. hewitt b, batchelor b (1975), punching shear strength of restrained slabs. journal of structural division, asce 101(st9):1837 1853. james, r.g.(2004), anacap concrete analysis program theory manual, version 3.0, anatech corporation, san diego, ca, 2004. japan society of civil engineers, jsce (1997), recommendation for design and construction of concrete structures using continuous fibre reinforcing materials. concrete engineering series 23, edited by a machida. kinnunen s, nylander h (1960), +. transactions of the royal institute of technology, stockholm, sweden. norway 112. marzouk h, hussien a (1991), punching shear analysis of restrained high strength concrete slabs. canadian journal of civil engineering 18:954-963. mohamed k, rizkalla s (1999), behavior of concrete bridge decks reinforced with frp. technical report, isis canada, university of manitoba, manitoba, canada. mufti a, newhook j (1998), punching shear strength of restrained concrete bridge deck slabs. aci structural journal 95:(4):375-381. the journal of engineering research (tjer), vol. 16, no. 1 (2019) 11-17 assessing software reliability using genetic algorithms r. jain *,a , and a. sharma amity university uttar pradesh, noida, department of it, indira gandhi delhi technical university for women (igdtuw), india. abstract: the role of software reliability and quality improvement is becoming more important than any other issues related to software development. to date, we have various techniques that give a prediction of software reliability like neural networks, fuzzy logic, and other evolutionary algorithms. a genetic algorithm has been explored for predicting software reliability. one of the important aspects of software quality is called software reliability, thus, software engineering is of a great place in the software industry. to increase the software reliability, it is mandatory that we must design a model that predicts the fault and error in the software program at early stages, rectify them and then increase the functionality of the program within a minimum time and in a low cost. there exist numerous algorithms that predict software errors such as the genetic algorithm, which has a very high ability to predict software bugs, failure and errors rather than any other algorithm. the main purpose of this paper is to predict software errors with so precise, less time-consuming and cost-effective methodology. the outcome of this research paper is showing that the rates of applied methods and strategies are more than 96 percent in ideal conditions. keywords: evolutionary algorithms; genetic algorithm; faulty /non-faulty data analysis; software reliability engineering. الجينية اتالخوارزمي باستخدام البرمجيات موثوقية تقييم شارما أ. و أ،* جاين . ر ب بتطوير متعلقة أخرى مسألة أي من أهمية أكثر جودتها وتحسين البرمجيات موثوقية دور لفد أصبح :الملخص: الملخص الضبابي والمنطق العصبية الشبكات مثل البرمجيات بموثوقية تنبًؤا تعطي التي التقنيات من العديد لدينا فاليوم. البرمجيات أحد أن ,نعلم وكما. بوثوقية البرمجيات للتنبؤ جينية خوارزمية استكشاف تمقد التطورية. و الخوارزميات من وغيره في مهما مكانا تحتل البرمجيات هندسة فإن ، وبالتالي ، البرمجيات موثوقية يُدعى البرمجيات لجودة الهامة الجوانب في والخطأ بالخلل يتنبأ نموذجا بتصميم نقوم أن يلزم ، البرمجيات موثوقية زيادة أجل فمن البرمجيات . صناعة البرمجيات في أقل أداء بؤدي الى التحسين من بالعمل على تصحيحها وبالتالي و يقوم المبكرة المراحل في البرمجيات والتي ، الجينية الخوارزمية مثل البرمجيات بأخطاء تتنبأ التي الخوارزميات من العديد منخفضة. هذا ويوجد وبتكلفة وقت أخرى. و عليه فإن خوارزمية أي من أكثر واألخطاء والقصور البرمجيات أخطاء توقع على جدًا فائقة بقدرة تتمتع للوقت , استهالكا أقل الدقة عالية منهجية باستخدام البرمجيات بأخطاء التنبؤ هو الورقة البحثية هذه من الرئيسي الغرض لهذا المطبقة واالستراتيجيات الوسائل معدالت أن البحثية الورقة هذه نتائج هذا وتظهر. التكلفة حيث كما انها مناسبة من المثالية الظروف في المائة من مثيالتها في 69 الغرض هي أعلى من موثوقية هندسة الخاطئة و غير الخاطئة؛ البيانات الجينية؛ تحليل تطورية؛ الخوارزمية خوارزميات : المفتاحية الكلمات .البرمجيات corresponding author’s e-mail: jain.rac16@gmail.com doi: 10.24200/tjer.vol16iss1pp11-17 mailto:rjain1@amity.edu the journal of engineering research (tjer), vol. 16, no. 1 (2019) 11-17 12 1. introduction software quality prediction is a very vast and curtail issue in software improvement because it helps to find out various solutions for minimizing the cost of product development even if the product is time effective and needs fewer efforts to implement. there is a very much hot discussion on the notion that software quality characteristics (briand. l. et.al 2000) cannot be directly measured. however, it can be measured with the help of other software attributes like coupling, size of the code, and of course, the complexity of the software product. there is a relationship between the attributes that are measurable or those which may be unmeasurable, though these software models are very inflexible to simplify and recycle on very innovative and undetected software as their accuracy failure has been found to be significant. to overcome this process, this paper proposed a new method that, together with a genetic algorithm which acclimates such models, get to a new type of tests data. it gives an experimental evidence that clarifies this approach, which upbeats the decision trees machine learning algorithms (quinlan, j.r 1993) like c4.5 and random presumption. to predict software reliability, we have to use software development models with respect to a software testing process. the parametric models are currently being used in software fault prediction, in which the most commonly used model is known as the markov model. this model is sometimes also known as non homogeneous poisson process (wang j. et.al 2014). basically, software quality assessment includes evaluating architecture results attributes and combination of these small attributes in order to achieve the best quality control and features. for this reason, the best way to go with software testing while using various test designs or models with genetic algorithm includes the white box testing methods and some other methods that are also available like black box testing method, robust testing, alpha and beta testing etc. in these proposed methods, a control structure of the software programme (that is actually a combination of the graph, use case diagram, and some other control flow diagram) has been prepared. basically, a control flow diagram or control flow graph (yamada s and ohba 1983) is graphical representation where each node represents a program instruction and every edge represents a transfer of control between the instructions used for the same nodes using that particular edge. many software reliability prediction and growth models have been designed and developed. we can use spc (statistical process control) for monitoring the process of software reliability and track all the actions which had been taken during the software failure process (satya prasad r. et.al 2012). now the role of genetic algorithm basically becomes more important because software testing is a key for software quality assessment. in this work, genetic algorithm and genetic programming are used to optimize the test data and generate reports on complex data analysis. for this reason, we use a fitness function that will be considered for following three criteria of population production as hereunder: 1. all path and traversal, that behaves as an independent path. 2. what percentage of an independent path is traversing? 3. the production time of the test data is not more than the controlled time. 2. background choi sang-hyeon and lee ikjin (2018), have given a detailed introduction about sequential optimization and reliability assessment. this is based on rbdo terminology, which has a higher success rate than conventional double loop rbdo methods although it is not more effective than the single loop approaches. mishra, dubey s.k. (2016) applied fuzzy approach for error detection and prediction. the authors used ck metrics for fault prediction and continuously made efforts to meet the challenge of maintaining software reliability in object-oriented languages with the help the fuzzy logic approach and ck metrics. according to amin a, grunske l, colman a. (2013) reliability is the main constraint software quality assessment. there were various models that have already been designed and used to estimate and predict the reliability based on the result of software testing methods. in this contrast the software reliability growth models are the most commonly used model to achieve this goal. sharma, dubey s.k. (2015), conducted a survey on various techniques that extract from various journals and conferences and analyze the widely used methodologies and techniques used to predict software reliability. kumar r and gupta n (2015); in order to establish a relationship between reliability and complexity, as used the respective influence of the relationship between complexity and reliability. bishnu p.s and bhattacherjee v (2011) proposed an idea about kd-tree methodology for predicting the fault at the early stage of software programming development with the help of the application of kmedoids with kd-tree. they applied a learning approach for the prediction of faults in the software module. it is also called supervised technique that is also used for prediction effort. rauf et. al. (2010) have presented a gui testing and analysis methodology on genetic algorithm. in order to expose the event-driven characteristics of gui, they used event-flow graph technique together with automated gui testing. the genetic algorithm uses various parameter combustions like coverage function, fitness function, crossover and mutation that ensure the correctness of test data. 12 r. jain and a. sharma for improving software testing efficiency, researchers srivastava and tai-hoon (2009) fond out the most critical paths after applying the genetic algorithm technique. to find the clusters in a program a method for optimizing software testing efficiency was identified. to achieve this, the researchers used wcf (weighted control flow graph) to test data generation by using the machine learning algorithm called genetic algorithm. there was another approach that found out the entire path in the underlying test called path testing that covers every possible path in the software program. the researchers concluded that, for improving software testing efficiency and performing exhaustive search, we can apply genetic algorithm easily and get meaningful results. yong c (2009) compared a test data for automatic path coverage by using genetic algorithm for generating new test data for path testing. he found that while using genetic algorithm less time is required for generating the meaningful test cased for path testing. ghiduk g (2008) introduced another method for reducing the software testing cost. it is a concept of relationships between the nodes of control flow graph that overcome the cost of testing the data. this is done with the help of a new fitness function that is defined by using dominance relationship for evaluation of the test data. the cost of testing is reduced by evaluating the effective fitness function and simultaneously comparing it to the random testing techniques for evaluating the effectiveness of the new fitness function as well as the methods that are used to overcome the cost of software testing. the proposed genetic algorithm and random technique were used with various object-oriented languages and programs, like structure oriented programming languages. as proposed by rajappa et. al. (2008), the use of graph theory with genetic algorithm approach will provide effective test cases for software testing. this is done by using the directed graph of all states of the program for the expected behavior, and the population was generated with the help of all nodes of the graphs. a combination of nodes was also selected from the population to perform crossover function and mutation for obtaining child nodes. this will lead to generating test cares in the real-time system. gupta, rohil (2008) used the genetic algorithm methodology to generate test cases for an object oriented software program and other structured languages with the help of tree data structure. in an experimental step, this approach was used to generate test cases for java classes. in the field of evolutionary algorithm set, the genetic algorithm is one of the most powerful and multi-purpose optimization tools which gives the right direction and recognition to the principles of the evolution. the genetic algorithm is very much capable of offering ideal solutions even in the most complex research environment de jong et al. (1989). at the beginning of 1960’s, the genetic algorithm was first proposed by john holland and at that time it was basically used for the issues that dealt with the complex search space programs and structural behavior catal et al. (2009). the genetic algorithm was used to simulate the progression of living things through finding out a suitable answer related to all the problems related to it. these issues necessitated the requirement of such type of algorithm that generates the populations of test, goldberg d.e. et. al. (1992) analyzed a test or a series of tests, developing test data generators that can be structuralized as functional testing (goldberg 1989; dua et. al. 2002). basically, a genetic algorithm typically has five main parts as hereunder:  chromosome  pooling of chromosomes  fitness function  selection function  mutation and crossover operator we can understand chromosome as a string of binary data, or as a data structure. initial pool of chromosome can be manually created or may be randomly produced by some logic tricks and evaluation process. the selection function is responsible for deciding the initial progress stage of the genetic algorithm provided by the mutation and crossover operators in which chromosomes will participate. however, the crossover and mutation operators used to exchange the genes from two chromosomes and create two new chromosomes for generating population. a pseudo-code for genetic algorithm can be defined as: ga : genetic algorithm starts { initialize : initialize the population (children) evaluate : calculate the fittest function applied on the population if (condition! = accepted criteria): implemen-tation of loop until termination criteria reached { selection function: selection of fitness function operators : select operators like lro (linear range operators) evaluation process : applied fitness function on operators and evaluate results } } 13 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 11-17 basically the genetic algorithm is a used for bounded and unbounded optimization. 3. general fault prediction approach the defect test cases and test data regarding measurement that have been collected from software development efforts are used to construct a prediction model. in comparing the actual defectiveness and predicted defectiveness of the software modules in the test data, one can get the model performance. figure 1 shows the common defect prediction procedure explained by sunghun kim et.al. (2011). 3.1 tagging the data defect can be collected for preparation of a prediction model. this is a process that used to extract the instances of data items from different s/w documentations. 3.2 mining features and creating training sets in this phase, there is one important thing that is taking place which is prediction of extraction of features and prediction of labels of illustrations both are coming from the prediction of fault test data. defect prediction sheta a (2006) has some complex components like metrics, keywords, dependencies and change that may be structural and nonstructural. the training set can be produced by combining the labels and features of instances that are used by the machine learning algorithms. 3.3 constructing prediction models we can use general machine learning algorithms goldberg, d.e. (1987) such as support vector machines or bayesian network for building a training set and prediction model. this model can use two levels as “true” and “false” and obtain a new instance. when the prediction of the model is calculated pham et al. (1999) it requires testing data sets not training sets. this can be obtained by a prediction model that evaluates and compares the real labels and prediction. we can also separate the training sets from the testing sets by using the 10-fold crossvalidation methods. 4. proposed model: a genetic algorothm approach for fault prediction while using a genetic algorithm, we must use three components to create a prediction model that can predict the fault. these are as follows: 1. creation of a model of the software 2. generating effective test cases 3. calculation of software test adequacy a. generation of test data for test case generation in this paper genetic algorithm is used to generate test cases that predict the software fault data set for prediction of software reliability oliveira e et. al. (2005). for this reason, we assume that the chromosome is one that contains the input values. the proposed algorithm can evaluate the test data berndt d et. al (2003) by implementing the program including the test data as inputs. figure 1. general fault prediction approach using machine language sunghun kim et.al. (2011). figure 2. a genetic algorithm approch for software prediction.variables used: 14 r. jain and a. sharma algorithm used: generatetdata input: part of a software program cdgcontrol dependencies graph for test code ipol: initial population used for genetic algorithm test_r list of test requirements output: final: set of test cases, used for fault prediction. variables used: cdg_pth: a set of all paths in cdg t_update: a record of satisfied test data cdg_pth: a set of all paths in cdg t_update: a record of satisfied test data old_pop & n_pop: set of test cases motive: which test cases are to be generated? max_a(): function that returns maximum attempts for single target oft (): function that returns true when time limit reached and false otherwise. step: 1 initialization and setup create software model create control dependencies graph initialize initial population step: 2 generate test cases while ((some, (r, unmarked) ԑ test_r and not oft ()) do select unmarked target from test_r while target not marked and not max_attempt () do compute fitness function sort the initial population according to the fitness select parents of new population generate new populations execute program for each population end while end while step: 3 generate the final effective test cases satisfy (test_r) return (final, test_r) end the typical use of the genetic algorithm is crucial due to this procedure. this algorithm works according to the assigned goal and each test case has its own goal rather than having the fitness function unchanged during the test case generation houck et. al. (1995). the generation of test data has an input named as “program” where a prototype can be used for generating the test data. let us see the following code as an example code of a program to generate test cases using a genetic algorithm for software predictability and fault prediction. let us consider a programming code written in clanguage that has three integers i.e. x1, x2 and x3. now generate test cases randomly and also create a cfg for this program code. the test cases are represented by t1, t2, t3, t4, t5, t6, t7 and t8. 1. intiger ( x1, x2, x3) 2. read x1,x2,x3; 3. if( x1<x2) 4. if( x2<x3) 5. x1=x3; 6. else x3=x1; 7. end if 8. end if 9. print x1, x2, x3; 10. end suppose that the algorithm we are using selects the following test cases in order { (0, 6, 9), (0, 1, 4), (0, 1, 4), (0, 1, 4) }, then by using one point crossover method we can get t5( 1, 6, 4 ), t6 (0, 1, 9), t7 (0, 6, 4) & t8 (5, 1, 4), after applying genetic algorithm, the two t5 and t7 satisfy the target and it is found that t6 and t8 are a faulty data set. figure 3. a control flow diagram for programming code written in c-language. 15 15 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 11-17 after implementation of genetic algorithm we find following test data: table 1. data set. test cases input (x, y, z ) statement executed t1 0, 1, 4 1, 2, 3, 4, 6 t2 1, 6, 9 1, 2, 3, 4, 6 t3 5, 0, 1 1, 2, 3 t4 2, 2, 3 1, 2, 6 t5 1, 6, 4 1, 2, 3, 4, 6 t6 0, 6, 9 1, 2, 6 t7 0, 6, 4 1, 2, 3, 4, 6 t8 5, 1, 4 1, 2, 3 5. proposed algorithm the proposed algorithm has been applied to the test data as per the follwing stepts written hereunder. 1. take a program pseudo code written in any programming language (in this paper c language) 2. generate the test data of the program 3. find all the possible conditions in the programing code. 4. generate the random set of test cases and apply it, 5. define the fitness function to calculate the path coverage of the programing code using cfg. 6. check for the path coverage in cfg. 7. if path coverage is satisfied, then stop. otherwise, go to next step. 8. select the initial population with the rank based selection process. 9. apply the ga’s operations (crossover and mutation) to generate the new population. 10. goto step 6 6. measurement of software reliability set of test-cases the set of test cares are derived from data selection, which is a collection, of path traversal from one node to another. mutants test cases to generate the mutants data, the volume of code along with mutant’s operator are used. the test adequacy can be calculated by the formula given below: adequacy = coverage of test × the mutation score the range of calculated adequacy is in the sort of [0, 1]; and it is useful in completing a accurate approximation of software reliability fault prediction. if the test model is an exact representation of the system under test, then the reliability of the software can be estimated by the following ratio: casestestof.totalno verifiedisthatcasestestof.no adequacy  (1) the program that is under the test can be assumed as a diagram shown (fig. 2) containing the verified (specified by the symbol ⇒) and non-verified (specified by the symbol →) and arbitrarily persuaded fault is the dart thrown at it. figure 4. example of path in a given program, where the path t2 is verified and reliable. 7. conclusion to achieve the software reliability, a code verification tool should comprehensively investigate every piece of code and verify its reliability against all possible data values. the study on programs has revealed the following results: software reliability fault prediction is actually based on the number of defects that had already been present in the software before delivery or dispatch. while dealing with safetycritical software there was no situation found in which one can find the warnings or even errors during dynamic analysis. but in the present study, it was found that the average warnings or errors observed during the dynamic analysis decrease exponentially as the reliability increases. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. t1 t2 t3 t2 16 r. jain and a. sharma references berndt d, fisher j, johnson l, pinglikar j, watkins a (2003), breeding software test cases with genetic algorithms. in system sciences, proceedings of the 36th annual hawaii international conference on. ieee 338–347. bishnu p.s, bhattacherjee v (2011), application of k-medoids with kd-tree for software fault prediction. ieee transactions software engineering 321-452. briand l, daly w, wüst j, porter d (2000), exploring the relationships between design measures and software quality. journal of systems and software 245-273. caoy, hu c, li l (2009), an approach to generate software test data for a specific path automatically with genetic algorithm. international conference on reliability, maintainability and safety 888-892. catal c, sevim u, diri b (2009), clustering and metrics thresholds based software fault prediction of unlabelled program modules. sixth international conference on information technology: new generations 199-204. de j, spears w.m (1989). using genetic algorithms to solve np-complete problems. in icga 124-132. dua a, mishra g (2002), stochastic search technique for solving constrained optimization problems with multiple objectives. proc of national conference on emerging convergent technologies and systems 399-404. goldberg d.e (1987), genetic algorithms in search, optimisation, and machine learning. proc. of the 2nd int. conf. on genetic algorithms, addison wesley 41 –49. goldberg d (1989), genetic algorithms in search, optimization and machine learning. addisonwesley, reading, massachusetts 60-309. goldberg d.e, deb k, clark j.h (1992), genetic algorithms, noise, and the sizing of populations. complex systems 333–362. ghiduk, ahmed s, girgis m.r (2010), using genetic algorithms and dominance concepts for generating reduced test data, informatica slovenia 377-385. gupta n.k, rohil m.k (2008), using genetic algorithm for unit testing of object oriented software. international conference on emerging trends in engineering and technology 308-313. houck christopher r, joines a, kay j.g.m (1995), a genetic algorithm for function optimization. a matlab implementation 1-10. kim s, zhang h, wu r, gong l (2011), dealing with noise in defect prediction. cse‟11, waikiki, honolulu, hi, usa 481-490. kumar r, gupta n (2015), reliability measurement of object oriented design: complexity perspective. international advanced research journal in science, engineering and technology 38-44. mishra, dubey s.k (2016), reliability of object oriented software using fuzzy approach. iso/iec 9126 model and ck metrics 398-401. nirpal, premal b, kale k.v (2010), comparison of software test data for automatic path coverage using genetic algorithm. internal journal of computer science and engineering technology 42-48. oliveira e, pozo a, vergilio s.r (2006), using boosting techniques to improve software reliability models based on genetic programming. 18th ieee international conference on tools with artificial intelligence 643-650. oliveira e, silia c (2005), modelling software reliability growth with genetic programming. proceedings of the 16th iee international symposium on software reliability engineering 39-121. pham h, nordmann (1999), a general imperfect software-debugging model with s-shaped faultdetection rate. ieee trans. reliability 169–175. quinlan j.r (1993), programs for machine learning. morgan kaufmann publishers 235-240. rajappa v, biradar a, panda s (2008), effective software test case generation using genetic algorithm based graph theory. first international conference on emerging trends in engineering and technology 298-303. rauf a, anwar s, jaffer m.a, shahid a.a (2010), automated gui test coverage analysis using ga. 7 th international conference on information technology new generations 1057-1062. sharma c, dubey s.k. (2015), a perspective approach of software reliability models and techniques. arpn journal of engineering and applied sciences 7300-7308 srivastava p.r. kim t.h (2009), application of genetic algorithm in software testing. international journal of software engineering and its applications 87-96. sheta a (2006), reliability growth modelling for software fault detection using particle swarm optimization. ieee congress on evolutionary computation 10428–10435. yamada s, ohba (1983), s-shaped reliability growth modelling for software error detection. ieee trans. reliability 475–484. 17 comp080726.qxd the journal of engineering research vol. 6, no. 2 (2009) 47-51 1. introduction the multi-carrier mc technology has received much attention in modern communication systems. in mc systems a number of data symbols are transmitted at separated sub-carriers in parallel thus increasing the symbol length, which reduces the sensitivity to interference. multi-carrier modulation mcm divides the available frequency band into a large number of orthogonal tones which can be implemented in all digital realization by exploiting discrete fourier transform dft methods as in dmt and ofdm multi-carrier modulation hnzo et al. 2004; reimers 2001 and al-dhahir and minn, 2006). discrete multi-tone dmt a currently standard for asynchronous digital subscriber line adsl units manages bandwidth for data transmission. it is one of the technologies that provides high speed internet access in residence and offices economically via wire technology (gagnaire, 2003 and daly 2003). orthogonal frequency division multiplexing ofdm gives high resistance to fre_____________________________________________ corresponding author’s e-mail: sameh-eng@yahoo.com quency selective fading. it is often used in mobile radio systems such as digital video broadcasting dvb/ digital audio broadcasting dab (reimers, 2001 and al-dhahir and minn, 2006). it is proposed for 4th generation wireless communication systems (bria, 2001). however, due to the large number of sub carriers, the mc signal owns a large peak to average power ratio papr. thus, the transmitted signal is sensitive to non-linear distortions that will degrade the error performance and introduce high adjacent channel interference with spectrum regrowth. several techniques for reducing papr of multi-carriers signal have appeared in the literature (mustafa, 2007 and friese, 1996). however, most of these techniques introduce additional complexity. moreover, there have been intense research efforts aimed at designing different multi-carrier transceiver scheme with different spectral containment. among those designs are the discrete cosine transform dct-mcm (al-dhahir and minn, 2006 and ak-dhahir and minn, 2005) and discrete wavelet multi-tone dwmt (daly, 2003 and farhang-boroujeny and chin, 2000). in this paper dct-based multi-carrier system was consimulation of a multi-carrier system in a non-linear flat fading channel samah a.b. mustafa department of electrical engineering, university of salahaddin, erbil, iraq received 26 july 2008; accepted 15 february 2009 abstract: this paper is on multi-carrier wireless communication system, focusing on the physical layer in such modulation system. multi-carrier mc system has received much attention in modern communication technology and is finding attractive applications. the paper introduces and discusses the discrete cosine transform dct-based mc. similarities and differences with respect to the fourier transform-based system are pointed out. it was found that a reduction up to 5db in peak power can be gained over the conventional system. also, the simulation over flat fading channel depicted that qam signaling in mc scheme based on dct performed better than that in fft-based scheme, although it depends strongly on the channel parameters. keywords: multi-carrier transmission, mcm, discrete multi-tone dmt, orthogonal frequency division multiplexing ofdm, peak-to average power ratio papr in mc system, flat fading channel flat fading channel ƒà°ùe ƒñn ägp á«£n’ iéæb ‘ πegƒ◊g o~©àe ωé¶æd iécéfi ≈ø£°üe ì骰s áá°°uuóóÿÿggüòl πegƒ◊g o~©àÿg ωé¶ædg ¿g .»æ©ÿg úª° àdg ωé¶f ‘ á«féjõ«ødg á≤ñ£dg ≈∏y õcîj h ¢s.ω πegƒ◊g o~©àe »µ∏°sódg ∫é°üj’g ωé¶f øy á°sgqo åëñdg øª° àj : äé¡hé°ûàdg ¤g é° jg iqé°t’g ” .ä.¢s.o ™£≤àÿg ωéªàdg ö«l πjƒ– ≈∏y ~ªà©ÿg πegƒ◊g o~©àe ωé¶f ¢ûbéæjh åëñdg ¢vô©j .áàø∏e äé≤«ñ£j ~lhgh ä’é°üjg á«æ≤àc √éñàf’g ¥ƒa ago’g π«∏– ¿gh ,oéà©ÿg ωé¶ædg ™e áfqé≤e ¬ñ°ùc øµá ≈ª¶©dg iq~≤dg ‘ 5o.ü ¤g π°üj é°véøîfg ¿g ~lh ~≤dh .ôjqƒødg πjƒ– ≈∏y ~ªà©ÿg ωé¶ædg ™e äéaóàn’g h .ä.±.± ≈∏y ~ªà©ÿg §£ıg ‘ ƒg é‡ π° ag .ä.¢s.o ≈∏y ~ªà©ÿg §£ıg ‘ .ω.cg.∑ úª° àdg hp πegƒ◊g o~©àe iqé°tg agog ¿g ¤g ò°ûj í£°ùe ƒñn hp iéæb .iéæ≤dg ägîeg qéh ≈∏y òñµdg iqé°t’g agog oéªàyg øe ºzôdéh áá««mmééààøøÿÿgg ääggooôôøøÿÿggiq~≤dg §°sƒàe ¤g ihqòdg ^ωg …o ±g hg ~eé©àÿg ooîdg º«°ù≤àh ∫é°sq’g o~©àe ^ä ω o π°üøæÿg ᪨ædg o~©àe ^ ω ¢s ω ^πegƒ◊g o~©àe ∫é°sqeg : .ƒà°ùe ƒñn hp iéæb ^ ¢s ω ωé¶f ‘ qg »h …g »h 48 the journal of engineering research vol. 6, no. 2 (2009) 47-51 sidered. many simulations have been carried to outline the differences over linear awgn, non-linear awgn and a flat faded channel with respect to the dft-based system. section ii reviews in brief the properties and representations of the dct, and section iii depicts dct-based scheme. simulation results and some further discussions of the error performance over different channel parameters are presented in section iv. finally, section v concludes the results. 2. discrete cosine transform the dct has been considered as one of the best tools in digital signal processing and therefore, it has many applications, e.g., in the area of multimedia and telecommunications. in this section, an introduction to the dct is given in order to provide background and motivation for our work. the orthogonal dct is classified into many different types with slightly different even/odd boundary conditions at the two ends of the matrix. the first definitions of the forward and inverse dct were given by: (1a) (1b) the dct operators map an n-size real sequence into another n-size real sequence. it is a linear fourier-related transformation similar to the dft using only real numbers of the dft. by repeating the samples in a time reversed order and performing a dft on the length sample set a dct is obtained. also, if we represent the n-size sequence as x=[x1,…., xn] and y=[y1,…., yn] and denote the nxn dct matrix by: (2) thus eq. 1 can be rewritten as y=xc and x=yc-1. dct is computationally simpler and the dct matrices are orthogonal, ie. cct=i, the inverse transform matrix are obtained with a matrix transpose (where t denotes transposition and i is the identity matrix) (strang, 1999; oppenheim, et al. 1999 and rao and yip, 2003). 3. mc system model figure 1 illustrates the mcm system used for multicarrier modulation. in dct-based scheme, the ifft & fft blocks are simply replaced by an idct & dct or vice versa, respectively, since idct matrix is the transpose of dct matrix. in mc-based transceiver, the binary data is encoded into a set of m-psk or m-qam symbols, called sub-symbols, and converted into lower rate sequences via serial to parallel conversion, which are then multiplexed by an idct. the outputs are serialized and transmitted. at the receiver, inverse operations take place in reverse order. in fourier-based system, the hermition symmetry is enforced to ensure a real waveform at the output of the ifft. this can be achieved by conjugate mirroring the complex symbols and transmitting zero signals on the dc and nyquist tones of ifft block sequence. however, as complex signals are mapped onto orthogonal sub-channels without imposing the conjugate symmetry condition, the multi-carrier system transmits twice the data in twice the bandwidth required when the condition is imposed. also, it is not necessary in our proposed system, as even a real signal can be mapped onto orthogonal sub-channels. dct converts real signals to real signals (as depicted earlier), and hence binary signaling or parsed in-phase and quadrature components of complex data word must be used in each sub-channel individually to keep the same data rate in dct systems. 4. simulation results much of this paper compares the error performance of dct-based mc system, which transmits a real waveform with that of the fourier-based; therefore the conjugate symmetry condition is imposed on fft-based system under the same transformation size n=256 to provide fairest comparison possible. in the simulation we also considered a system operating without guard space in the resulted time domain symbols. the scheme was simulated in matlab ver. 7 with a constant bit allocation of log2(m) bits per sub-channel is assumed. the number of simulated bits was about 106 bits and the results was computed and averaged over 3 iterations for the demodulator. 4.1 linear awgn channel the error performance curves presented in fig. 2 are based on the signal-to-noise ratio snr of the channel, assuming the channel is a pure awgn channel. as seen in the figure, a differential phase shift keying (dpsk) in both mc system will limit the number of bits per symbol and results in an 11db loss in snr. by comparison the error curves, a gain in snr can be observed at the low 1 2 1 1 2 nk k n n w ( n ) ( k ) y x c o s nn π ∑ = − − = ⋅ 1k , . . . . . , n= 1 1 2 1 1 2 n n k k k ( n )( k ) x w y c o s nn π ∑ = − − = ⋅ ⋅ 1n , . . . . , n= 1 1 2 2k kwhere w k n =⎧= ⎨ ≤ ≤⎩ 2 1 1 2 kw ( n )( k )c cos nn π − − = ⋅ mapper mapper idct p/s i q i q figure 1. block diagram of dct-based mc modulator 49 the journal of engineering research vol. 6, no. 2 (2009) 47-51 error rates using dct-based scheme with qam modulated tones. the resulted real and imaginary parts from the constellation are used separately to modulate different tones to guarantee real time domain samples without transmitting zero signals on the dc and nyquist tones as have been applied in ifft block sequence. thus the transmission data rate is improved only by 2*log2m than that of fft-based system. however, a dct scheme can be used with all the subcarriers allocated to transmit only the estimated phase of the different constellation points (qpsk, dpsk), where the amplitude is constant, thereby improving the data rate to the double. 4.2 non-linear awgn channel we also analyzed the distribution of time domain transformed amplitude as depicted in fig. 3 & fig. 4. the probability density function (pdf) depends on the probability of occurrence of each discrete sample level. the distribution of mc signal with 256 sub-carriers and 16-qam is founded where the amplitude sn has rayleigh distribution with pdf given by: (3) where sn the transformed (using ifft or idct) time domain samples have real parts only as stated before. for rayleigh distribution, as seen the signal levels around the mean value have higher probability than other levels, while the occurrence of the large signals has the smallest probability and moreover different level in dct and fft-based schemes, as shown in fig. 3 & fig. 4. it is reasonable cast a way to change the statistic of the amplitude for the benefit of papr reduction. notice that papr is a random variable for each transmitted block. it has been found from a large number of independent runs for the dct and fft-based mc scheme that within each block, the peak power is reduced using dct-based system of about 5db and a reduction of papr of about 1.3db are achieved. for the simulated error in fig. 5 a non-linear power amplifier is applied with am/am response of clipping scheme at the saturation point at three different output back off obo; 4, 5 & 6db with no compensation of the non-linearity in the conventional mc system. obo is defined as the ratio of the maximum possible amplifier output power to the average output power. this indicates the power efficiency of the amplifier (al-dhahir and minn, 2006). obo=6db guarantees linear transmission while obo=4 & 5db in fft-mcm scheme generates an almost flattened error curves indicating saturation of the power amplifier. these results can be directly compared with the results for mc scheme using dct. error curves with obo=4, 5 & 6db are also measured. it is found that when obo=5db the simulated approaches the linear figure 2. pe in dct-based and fft-based mc scheme versus snr of linear awgn channel figure 3. distribution of dct-based scheme signal amplitude figure 4. distribution of fft-based scheme signal amplitude 2 22 2 0 0 s /( )n n s n sn s e s p( s ) otherwise σ σ −⎧ ≥⎪⎪= ⎨ ⎪ ⎪⎩ 50 the journal of engineering research vol. 6, no. 2 (2009) 47-51 transmission due to the overall net improvement in the papr using dct, while when obo decreases to 4db the performance diverges from the linear case also, indicating non-linear distortion occurs. 4.3 flat fading channel flat fading has an impulse response given by: (4) where g (t, τ) is the impulse response at observation time t to an impulse applied at time t-τ. this channel is considered to be slowly time-varying such that the amplitude β(t) and the phase shift θ(t) can be considered constant during one symbol interval. to compare and contrast the performance of the schemes over fading channel, the error curves are examined at similar parameters. fig. 6 shows that, for qam multi-carrier signal the performance are incompatible exactly for 16-qam signal where the dct-based system approaches the linear transmission. the distortion arose from fading channel can be compensated by increasing snr, while multi-amplitude signaling based on fft without channel estimation induces irreducible error curves. as depicted the dpsk signals in both mc systems have the same performance depending on the phase tracking of the decoder. the simulated error rates as shown in fig. 7 are depicted for various flat fading parameters, which affects on the overall dct-based mc system performance and the error performance limit. it is clearly observed that the amplitude and phase shift of the reflected signals have a much stronger influence on the systems using qam than on dpsk modulation systems. the results could be greatly improved to meet the specification requirement that could be implemented in the future to further enhance the simulator, where the results can be extended to any system which uses the mcm technique. 5. conclusions extensive computer simulations have been carried out to demonstrate and compare the performance of dct multi-carrier scheme with that of fft scheme. the results indicate quite similar performance over linear awgn channel. concerning the overall system performance, qam is often more robust and more spectrally efficient than dpsk modulation. performance enhancement of the proposed system has been recorded over non-linear awgn channel due to the reduction in papr power of the dct multi-carrier signal. it is more flexible to present linear transmission with no compensation of the non-linearity in the system. moreover, the simulated results concludes that the new scheme over flat faded channel without channel estimation performs better corresponding to figure 5. pe in fft-based mc scheme versus snr of non-linear awgn channel j ( t )g( t , ) ( t ).e . ( t )θτ β δ= figure 6. pe in dct-based and fft-based mc scheme versus snr over flat fading channel figure 7. pe in dct-based mc scheme versus snr with 16-qam (discrete line) and 64-qam (solid line) modulated sub-carriers over three different flat fading channel characters (ch.1, ch.2, and ch.3) 51 the journal of engineering research vol. 6, no. 2 (2009) 47-51 that of the traditional scheme, especially with qam modulated tones depending too much on the reflected path amplitude and phase. such modulated tones in fft-based scheme results in an irreducible error even at high snr. however, the results could be greatly improved to meet the specification requirement that could be implemented in the future to further enhance the simulator, where the results can be extended to any system which uses the mcm technique. references al-dhahir, n. and minn, h., 2006, "optimum cdtbased multicarrier transceivers for frequencyselective channels," ieee transaction on communications, vol. 54(5), pp. 911-921. al-dhahir, n. and minn, h., 2005, "a new multicarrier transceiver based on the discrete cosine transform," wireless communications and networking conference, vol. 1, issue, pp. 45-50. bria, a., 2001, "4th-generation wireless infrastructures: scenarios and research challenges," ieee personal communications magazine, vol. 8, pp. 25-31. daly, d., 2003, "efficient multi-carrier communication on the digital subscriber loop," ph. d thesis, national university of ireland. farhang-boroujeny, b. and chin, w.h., 2000, "time domain equalizer design for dwmt multicarrier transceivers," electronics letters, vol. 36(18), pp.1590 1592. friese, m., 1996, "multicarrier modulation with peak-toaverage power ratio reduction and error correction," electronic letters, vol. 32(8), pp. 713714. gagnaire, m., 2003, "broadband local loops for high speed internet access," artech house. hnzo, l., munster, m., choi, b.j. and keller, t., 2004, "ofdm and mc-cdma," ieee communication socity. mustafa, s., 2007, "performance analysis of coded orthogonal frequency division multiplexed signal," information and communication technologies international symposium ictis'07, pp. 566 -569, morocco. oppenheim, a.v., schafer, r.w. and buck, j.r., 1999, " discrete-time signal processing, 2nd ed., prentice hall, new jersey. rao. k.r. and yip, p., 2003, "discrete cosine transform: algorithms," advantages, applications. boston: academic press. reimers, u., 2001, "digital video broadcasting (dvb)," iinternational standard for digital television, new york: springer-verlag. strang, g., 1999, "the discrete cosine transform," siam rev., vol. 41(1), pp. 135-147, 1999. d:\desktop\paper_8.xps tjer 2013, vol. 10, no. 1, 88-96 this technology subjects solid tubes and sand control screens to large radial plastic deformations through expansion. the expansion process is usually performed either mechanically or hydraulically using conical mandrels. tubular (casing) expansion has been performed down hole for well rehabilitation to repair damaged zones, corroded casing sections, and shut-off perforations. in addition, expansion of sand control 1. introduction solid expandable tubular technology was introduced to the oil industry in the early nineties with the aim of achieving mono-diameter wells as alternatives to existing telescopic well configurations. this led to tremendous savings of as much as 50% of the original cost of the telescopic well configurations. _______________________________________ *corresponding author’s e-mail: aseibi@pi.ac.ae stress concentration factor of expanded aluminum tubes using finite element modeling l mhamdia, ac seibi*b, a karrechc, s el-borgia and i barsoumb amechanical engineering department, tunisia polytechnic school, tunis, tunisia bmechanical engineering program, petroleum institute, abu dhabi, united arab emirates cschool of civil and resources engineering, the university of wastern australia, australia received 23 may 2012 ; accepted 31 october 2012 abstract: this paper discusses the development of semi-empirical relations for the maximum stress concentration factor (scf) around circular holes embedded in aluminum tubes under various expansion ratios and mandrel angles. finite element models were developed to study the expansion of a typical aluminum tube with embedded holes of various sizes. an elastic perfectly-plastic material behaviour was used to describe the structural response of the tubes under expansion. various hole-diameter-to-tubewall-thickness ratios, tube expansion ratios, and mandrel angles were considered to determine the stress state around the hole at zero and 90 degree locations from which the maximum scf was determined. semi-empirical relations for the maximum scf using the lagrange interpolation formulation were developed. the developed relations were found to predict the scfs accurately. keywords: stress concentration factor, expansion process, semi-empirical formulations, tube expansion 89 l mhamdia, ac seibi, a karrech, s el-borgi and i barsoum screens has become a normal practice used by field engineers to increase productivity by enlarging the cross-sectional area. sand control screens, which are generally used in horizontal wells, consist of tubes possessing slots ranging from fine at the outer surface to considerably larger-sized holes at the base plate. the presence of holes can be problematic and presents a source of stress concentration. this may reveal some weaknesses in the structure, especially when expanded to higher expansion ratios. chitwood et al. (2005) showed that a perforated 13cr base tubular fractured after 25% expansion and cracks propagated between the perforations, causing a loss of structural integrity. pilkey 1997 and toubal et al. 2004 studied the problem of stress concentration around holes in flat plates under axial loading. similarly, (van dyke 1965; tan 1994) studied pressure vessels and their related components under internal pressure and/or axial loading. steele et al. 1986 and xue et al. 1991 developed an efficient numerical algorithm to determine stress variations in cylindrical vessels with large openings. subsequently, dennis and palazotto 1990 extended the approach to composite structures with cut-outs. however, according to the knowledge of the authors, stress concentration around holes in expanded tubes has never been studied before. therefore, the present work, which deals with the estimation of stress concentration around holes in expanded tubes, is the first of its kind. this work is complex in nature and represents a moving boundary problem in terms of mandrel motion within the tube passing by the embedded hole while expanding the tube to a prescribed expansion ratio. the stress concentration may induce crack initiation at the notch root which can result in loss of integrity. for this reason, the tangential stress at zero and 90 degree locations around the hole received great attention through the conduct of a finite-element analysis from which scfs based on the maximum stress were determined. 2. finite element modeling finite element analysis was performed on solid tubes subjected to various expansion ratios, mandrel angles, and hole-diameter-to-tube-wall thickness ratios. tubular expansion occurs when the mandrel moves downward and expands the solid tube to a defined expansion ratio (do di)/di*100 (fig. 1). figure 1a shows the geometrical model of a tube with an inner diameter of di = 54.6 mm, a wall thickness of t = 6 mm, and an embedded circular hole of variable diameter, d. the figure also shows a conical mandrel with mandrel angle and an outer diameter of do, which can be varied according to expansion requirements. points 1 and 2 are located at zero and 90 degree positions around the hole and are the focus of this study, where the variations of the tangential stress at both locations were taken as the mandrel moves downward to expand the tube. this expansion process subjects the tube to a compression state ahead of the mandrel, due to the end conditions. a three-dimensional brick element and an analytic rigid body element were used to model the tube and mandrel, respectively, as shown in fig. 1b. the mandrel/tube interaction was modeled using a coulomb friction model where a friction coefficient of 0.1 was used. proper boundary conditions representative of lab and field conditions are essential to obtaining accurate results. special attention must be focused on the region in the vicinity of the hole where stress concentration occurs. the zoomed window section in fig. 1b shows the finite element mesh of the tube where fine meshes were used in the vicinity of the hole due to the high stress concentration experienced during expansion in this region. an elastic perfectly-plastic material model was used to describe the tube's behaviour. the material properties used for the tube were elastic modulus e = 67 gpa, poisson's ratio = 0.3, and tensile yield stress y = 150 mpa. tube expansion in this study was performed by constraining all nodes lying at the bottom side of the tube from moving in any direction, whereas all other nodes were free to move. since the mandrel was modeled as a rigid body, a vertical downward displacement was given to its reference point, which was located along its centre line at the top face to move the mandrel down. 3. case study in this study, the friction coefficient was taken as 0.1 for the entire finite element analysis. three mandrel angles of = 10, 22.5, and 30o, as well as five expansion ratios of e = 5, 10, 15, 20, and 25% were used. in addition, eight hole-diameter-to-tube-wallthickness ratios were used: d/t = 0.1, 0.2, 0.6, 0.8, 1.0, 1.2, and 1.5. the tangential stress of zero and 90 degrees was taken from the finite element results from which the nominal and peak stress values were computed. the generated data were used to determine the maximum scf at both locations. peak values of the tangential stress were used to estimate the scf defined by chitwood et al. (2005) as: (1) 90 stress concentration factor of expanded aluminum tubes using finite element modeling where max (x)|(0.90) is the maximum stress at zero and 90 degrees, and x denotes the mandrel position along the tube during the expansion process. the stress ave is defined as the average stress of the stress variation before the mandrel reaches the hole. it is worth noting that eqn. (1) differs from the traditional definition of the stress concentration where the nominal stress is taken into account instead of the average stress used in this study. 4. results and discussions the finite element results focus on the stress state at zero and 90 degrees at locations around the hole from which the maximum scf can be determined. figure 2 shows the variation of the drawing force as a function of the mandrel position for various expansion ratios and mandrel angles. in all cases, the figure shows that the force increases with an increase in expansion ratio and mandrel angle. figure 2d shows a summary of the average drawing force in terms of expansion ratio for three mandrel angles, indicating that the force increases with an increase in both expansion ratio and mandrel angle. typical stress variations of 22 and 11 in terms of mandrel position are shown in figs. 3 and 4. it is worth noting that the tangential stress at points 1 and 2 are respectively compressive and tensile when the mandrel is far away from the hole, since tube expansion was performed under compression. the two stresses become more compressive as the mandrel approaches the hole and changes sign to become tensile, reaching peak values while expanding the hole. as the mandrel passes the expanded hole, residual compressive stresses were induced at both locations. this behavior is shown in fig. 5 where the shape variation of the hole during the three stages of expansion is clearly observed. it can be seen that the hole is under compressive and tensile stress state at points 1 and 2, respectively, when the mandrel is far away from the hole. as the mandrel approaches the hole and starts expanding the region in the vicinity of the hole, the hole becomes oval (fig. 5b) and the stress state at both locations becomes tensile due to the hole's enlargement. as the mandrel passes the expanded zone of the hole, the hole tends to recover; therefore, the residual stress becomes compressive at both locations. peak values of 11 and 22 were used to compute the maximum kmax(0.90) at points 1 and 2 in terms of a hole-diameter-to-tube-wall-thickness ratio as well as mandrel angle and expansion ratio. (a) geometric model (b) finite element model figure 1. geometric and finite element models of the tubular expansion process 91 l mhamdia, ac seibi, a karrech, s el-borgi and i barsoum 92 stress concentration factor of expanded aluminum tubes using finite element modeling to the increase of the nominal stress nom and the decrease of the maximum stress max as of the ratio of d/t increases. it is worth mentioning that the scf based on max decreases as the expansion ratio increases from 5 to 25% which is also attributed to the decrease of nominal stress as the expansion ratio increases from 5 to 25%. 4.1 semi-analytical formulations of stress concentration this section aims to develop expressions for the scfs as functions of the parameters d/t, , and e used in this study. table 1 shows expressions for the stress concentration factor kmax in terms of hole-diameterto-tube-wall thickness for one set of data. figure 4. tangential stresses versus mandrel position for an expansion ratio of 25%, mandrel angle of 30o, and when d/t = 0.6 mandrel displacement (mm) mandrel dislacement (mm) (a) stress at point 1 (b) stress at point 2 table 1. scf for a mandrel angle = 10o % expansion kmax 93 l mhamdia, ac seibi, a karrech, s el-borgi and i barsoum (c) after hole expansion figure 5. stress contours and shapes of the hole during the complete expansion process (a) before reaching the hole (b) while expanding the hole 94 stress concentration factor of expanded aluminum tubes using finite element modeling these seventh-order polynomials are valid within the range used in this study (0.1 d/t 0.15). to develop such relations, the lagrange interpolation method, which is an efficient numerical technique, was used. the advantage of this technique is that the data set points need not be arranged in any particular order as long as they are mutually distinct. the analysis started with a data set consisting of one set of either parameter (d/t, , or e) and their corresponding values of (ktmax) for a well-determined case. several cases figure 6. scf as a function of the hole-diameter-to-tube-thickness ratio for 5% expansion ratio and mandrel angles of 0 and 90 degrees (a) point 1 (b) point 2 figure 7. scf as a function of hole-diameter-to-tube-thickness ratio, expansion ratios and mandrel angles 95 l mhamdia, ac seibi, a karrech, s el-borgi and i barsoum were studied to determine the effect of each parameter on the scf. similar expressions can be developed for other parameters. figures 8 and 9 show a comparison between the finite element results and the developed semi-analytical model for kmax for different expansion ratios at points 1 and 2, respectively. both figures show good agreement between the finite element results and the developed model. figure 8 shows that the stress magnification factor is plotted at point 2 due to higher values of stress. the stress at point 2 experienced a significant change in value because the stress state changed from compression (pre-mandrel position) to tension (post-mandrel position) as the mandrel traveled along the tube during expansion. hole diameter to tube thickness ratio hole diameter to tube thickness ratio (a) scf at point 1 (b) scf at point 2 figure 8. scf kmax versus hole-diameter to tube-thickness ratio for an expansion ratio of 5% and mandrel angle of 22.5o hole diameter to tube thickness ratio hole diameter to tube thickness ratio (a) scf at point 1 (b) scf at point 2 figure 9. scf kmax versus hole-diameter to tube-thickness ratio for an expansion ratio of 25% and mandrel angle of 22.5o 96 stress concentration factor of expanded aluminum tubes using finite element modeling 5. conclusions the results from three-dimensional elastic perfectly-plastic finite-element analyses were presented to compute the scfs around circular holes in perforated tubes subjected to a compressive expansion process. over 120 finite element models were run in order to study the effect of the parameters d/t, , and e on the scfs. it was found that both scfs were higher at point 2 (90o location) indicating that failure might originate at this location. in addition, the results showed that the scf decreased as the hole-diameterto-tube-thickness ratio and mandrel angle increased. moreover, generalised semi-empirical relations of both scfs in terms of d/t, , and e were developed using the lagrange interpolation method. these expressions can reproduce any of the finite element results with the range of the parametric values undertaken in this study. references chitwood gb, eisinger nc, puckett bc (2005), corrosion resistant and long-term reliability of super-austenitic alloy 27-7mo for use as expandable tube. spe95094. spe international symposium on oilfield corrosion, aberdeen, united kingdom. dennis st, palazotto an (1990), static response of a cylindrical composite panel with cutouts using a geometrically nonlinear theory. journal of american institute of aeronautics and astronautics 28(6):1082-1088. pilkey wd (1997), peterson's stress concentration factors. 2nd edition, wiley interscience, john wiley and sons, isbn: 978-0-471-53849-3. steele cr, steele ml, khathlan a (1986), an efficient computational approach for a large opening in a cylindrical vessel. asme journal of pressure vessels technology 108(4):436-443. tan sc (1994), stress concentrations in laminated composites. crc press, taylor and francis group, usa. toubal l, karama m, lorrain b (2004), stress concentrations in a circular hole in composite plate. composite structure 68:31-36. van dp (1965), stress about a circular hole in a cylindrical shell. journal of american institute of aeronautics and astronautics 3:1733-1742. xue md, deng y, hwang kc (1991), some results on the analytical solution of cylindrical shells with large openings. asme journal of pressure vessels technology 113(2):291-308. (2) (3) c:\users\raoh\desktop\paper 7.xps the journal of engineering research (tjer) vol. 12, no. 1 (2015) 81-91 experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the genetic algorithm method r.s. al-maamaria*, g. vakili-nezhaada* and m. vatanib a* department of petroleum and chemical engineering, college of engineering, sultan qaboos university, p.o. box 33, pc 123, muscat, oman. b department of chemical engineering, amirkabir university of technology (tehran polytechnic), tehran, iran. received 24 february 2014; accepted 22 january 2015 abstract: achieving an estimation of viscosity in crude oil binary mixtures is often difficult because the relationship of viscosity, to the fraction of each crude oil, and many other parameters and constants is comply. this relationship can be expressed by mathematical models with different variables. besides the known models for predicting the viscosity of crude oil mixtures, the petroleum industry demands other models which give accurate predictions. in this work, two new empirical models have been developed for the calculation of the viscosity of binary crude oil blends. two techniques—least square (ls) and genetic algorithm (ga)—were used to determine the parameters of the proposed models. dynamic viscosity of 12 sets of crude oil blends at 298.15 k and 25 different shear rates were measured, resulting in 300 sets of binary data. moreover, 700 sets of kinematic viscosity binary data were collected from literature sources and used along with 200 of the 300 sets of experimental binary data with a wide range of american petroleum institute (api) gravity (9.89–41.2) and viscosity (1.054–165,860 cst) to examine existing available models as well as the newly developed models in this study. the remaining 100 experimental data points which were not used in the regression process were used for validating the models. the results in terms of the absolute average relative deviation (aard%) were 33.546 and 14.195 for the ls method and 13.113 and 13.672 for the ga method for proposed models one and two, respectively. the results of statistical parameters based on the ga and ls methods showed that the ga is a superior method for new model parameter estimation as compared with the traditional ls technique. the ga-based models developed in this study provided the highest accuracy for viscosity calculation of the crude oil blends over all existing models in the literature. keywords: viscosity, crude oil, binary mixture, genetic algorithm, least square, optimization. *corresponding author’s email: rsh@squ.edu.om; vakili@squ.edu.om r.s. al-maamari, g. vakili-nezhaad and m. vatani 82 12298.1525 300700 200300 api9.8941.21.054165.860cst 100 aard% 33.54614.19513.11313.67212 1. introduction heavy crude oil is highly viscous under normal reservoir conditions, but its viscosity is reduced to avoid problems in transportation and reduce pumping costs. heavy crude oil or bitumen is mixed with a solvent or diluent oil to achieve a crude oil blend with a certain viscosity. thus, the viscosity of a crude oil blend depends on the viscosity and fraction of each crude oil. predicting the viscosity of a crude oil blend is one of the greatest challenges in the petroleum industry, although the prediction can be expressed by mathematical models. several correlations have been proposed for calculating the viscosity of mixtures. centeno et al. (2011) used 17 mixing rules reported in the literature to predict the viscosity of sample oil. they classified mixing rules as pure, with a viscosity blending index and with additional parameters. pure mixing rules are easy to apply as they require experimental viscosity of the components and composition of mixtures in terms of volume, mole, or weight fractions. models include those of arrhenius (1887), bingham (1914), and koval (1963). the models based on mixing rules with a viscosity blending index are (prakash 2003; refutas baird 1989; maxwell 1950; wallace and henry 1987; chevron riazi 2005 and cragoe 1933). mixing rules with additional parameters include the calculation of extra parameters that are experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the …… usually obtained by mathematical methods. miadonye et al. (2000) and han et al. (2007) developed models for predicting the viscosity of a crude oil mixture based on additional parameters. al-besharah (1989) correlated a model based on viscosity, mass fraction, and the api gravity of each component. the calculation of model parameters is a critical task in developing mathematical models of such applications. concerning parameter calculations, many different methods have been suggested in the literature. most of the existing methods for solving problems are local or traditional search methods such as the least square (ls) objective function or maximization likelihood criterion (stragevitch and davila 1997). in such problems, the objective function is in the form of a nonlinear expression containing several extremum points which are both local and global minima and within the specified bounds of variables (sahoo et el. 2006). traditional local search methods often give local optima, and a global optimization routine often works for such problems. some popular stochastic optimization techniques based on global optimization methods include the genetic algorithm (ga) (holland 1992); simulated annealing (sa) (kirkpatrick et al. 1983), and differential evolution (de) (price et al. 2005). these methods belong to the evolutionary algorithms category, which mimics the principle of survival of the fittest. in recent years, stochastic global optimization methods have been used extensively in fluid phase equilibrium problems and parameter estimation of models. among these methods, ga is the most widely used (alvarez et al. 2008; bonilla-petriciolet et al. 2013; chamkalani et al. 2013; maddinelli and pavoni 2013; rangaiah 2001; singh et al. 2005; sahoo et al. 2006; sahoo et al. 2007; tan et al. 2014; vakili-nezhaad et al. 2013; vakili-nezhaad et al. 2014; vatani et al. 2012a; vatani et al. 2012b; xue et al. 2014). in this work, two new empirical models have been developed for calculation of the viscosity of binary crude oil blends. ls and ga techniques were used to determine the parameters of the proposed models. an accurate approach to determine the adjustable parameters of the models was found, and the results of the proposed models are compared with the results obtained using models available in the literature. 2. genetic algorithm (ga) ga is considered a reliable algorithm for complex engineering calculations. ga is based on natural selection—the process that drives biological evolution—and is recommended for use with optimization problems that are not well suited to traditional optimization algorithms, especially problems in which the objective function is discontinuous, non-differentiable, or highly nonlinear. ga differs from traditional optimization algorithms as follows: traditional algorithms use only a single point in each iteration, while ga explores the search space using randomly generated multiple points. derivative-based algorithms generate a new point by a deterministic computation while ga creates a new population by a probabilistic computation. the outcome of traditional algorithms depends both on initial guess work and on the step size used in the algorithm. for the optimization of problems, the fitness function must be defined; for a standard optimization algorithm, this is the objective function. the algorithm begins by creating random initial populations and modifies a population of individual solutions between the values of the lower and upper bounds of the successive variables. the current population size determines the subsequent size of the population at each resultant generation. increasing the population size enables the ga to search more points, thereby obtaining better results. as the population size increases, the computation time also increases. at each step, the ga selects individuals from the current population stochastically to be the parents of the children for the next generation based on genetic operators including selection, crossover, and mutation by exploring all regions of the search space. in the first step, an initial population is generated. each individual is evaluated for its fitness in order to proceed, and the best individuals are chosen from the selection step. individuals with a high probability have a greater likelihood of producing offspring, which are then generated by a combination of selected individuals using a crossover step. in the mutation step, random changes are applied to some individuals. the purpose of the mutation operator is to prevent the ga from converging to a local 84 83 r.s. al-maamari, g. vakili-nezhaad and m. vatani 84 minimum and to introduce new possible solutions into the population (chamkalani et al. 2013; preechakul and kheawhom, 2009). the algorithm continues to find the minimum of the fitness function. a more complete discussion of the ga, including extensions and related topics, can be found in matlab’s “genetic algorithm and direct search toolbox” (the mathworks, inc., 2004). 3. experimental and literature database three oil samples (light, medium, and heavy crude oils) were obtained from different omani oil fields. twelve binary mixture samples were prepared using different weight fractions (20, 40, 60, and 80%). the viscosity of each binary mixture and three pure crude oils were measured at 25 different shear rates (10–100 s-1) with a rotational rheometer: rheolabqc (anton paar gmbh, graz, germany), while the density for each sample was measured by a dma-5400m density meter (anton paar, gmbh). all of the measurements were carried out at 298.15 k. the viscosity and density of all samples are listed in tables 1 and 2, respectively. the crude oil mixtures used in this work cover a wide range of api gravities (9.89–41.2) and viscosities (1.054–165,860 cst) at various temperatures. as noted, the viscosity of 12 types of crude oil binary mixtures at 25 different shear rates were measured, and 300 binary data points were obtained. prepared samples weighed 5–7 grams. in addition to the experimental data, 700 pieces of data were obtained from the literature (al-besharah, 1989; centeno et al. 2011; diaz et al. 1996; rahmes and nelson 1948; yuan et al. 2005). from these 1,000 data points, 900 binary data were used in regression of the proposed models, and 100 experimental data were used to validate them. 4. viscosity models among all existing models for predicting the viscosity of mixtures, the most well-known, including 13 models, are presented here and are a function of api gravity, mass, or volume fraction and the viscosity of each component (table 3). in these models, v is the kinematic viscosity, and v and w are the volume and mass fraction, respectively. subscript i denotes the binary mixture components, so subscripts a, b, and mix are the more viscous component, the less viscous component, and the binary mixture viscosity, respectively. 5. proposed models after trying multiple models and regression analysis, two empirical models were developed (eqns. (1)–(8)). these models are able to predict the viscosity of crude oil binary mixtures with high accuracy. the regression analysis of these models was performed based on the ls and ga methods. the model coefficients of the two methods are presented in table 4. in the first model, the viscosity of the binary mixture is a function of mass fraction, api gravity, and the viscosity of each component and presented as: )ln(ln )ln(ln )ln(ln )ln(ln )ln(ln lnln 53 53 53 53 43 43 ccc ccw ccc ccw ccc cc bb bb aa aa mixmix mix (1) 1 2 ( )a a b b ac c api w c api api (2) 1 2 ( )b b a a bc c api w c api api (3) mix a a b bc w c w c (4) in the second model, the viscosity of the binary mixture is a function of the volume fraction and 3 1 2exp exp ( ) /mix vi c c c (5) where iv can be obtained from eqns. (6) to (8) and c1 to c3 are model constant parameters. a bv a v b v i v i v i (6) 1 2 3ln lnav ai c c c (7) 1 2 3ln ln( )bv bi c c c (8) in these models, w, v, v and api are mass fraction, volume fraction, kinematic viscosity and api degree gravity, and subscripts a or b and mix denote to the each crude oil and binary mixture respectively. experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the …… 85 table 1. experimental viscosity of different crude oils and their mixtures. note: a:80, b denotes a blend of 80% a and 20% b, etc. table 2. experimental density of different crude oils and their mixtures. components density (g/cm3) a:100 0.940 b:100 0.830 c:100 0.854 a:80, b 0.930 a:60, b 0.901 a:40, b 0.882 a:20, b 0.854 a:80, c 0.938 a:60, c 0.918 a:40, c 0.910 a:20, c 0.881 b:80, c 0.836 b:60, c 0.841 b:40, c 0.846 b:20, c 0.852 note: a:80, b denotes a blend of 80% a and 20% b, etc. r.s. al-maamari, g. vakili-nezhaad and m. vatani 86 table 3. models for the viscosity prediction of binary crude oil mixture. no. model name model ref. 1 arrenius log log logmix a a b bv v arrhenius (1887) 2 bingham 1 1 1 mix a a b bv v bingham (1914) 3 koval 0.25 0.25 0.25 mix a a b bv v koval (1963) 4 parkash 157.43 exp exp 0.93425 376.38 : 157.43 376.38 ln ln( 0.93425) a b i p mix p a p b p p i i where i v i v i i parkash ( 2003) 5 refutas 10.975 exp exp 0.8 14.534 : 10.975 14.534 ln ln( 0.8) a b i r mix r a r b r r i i where i w i w i i baird (1989) 6 maxwell 59.58959 exp exp 0.8 21.8373 : 59.58959 21.8373 ln ln( 0.8) a b i m mix m a m b m m i i where i v i v i i maxwell (1950) 7 wallace and henry 1 0.01exp : 1 ln 0.01 a b i mix wh wh a wh b wh wh i i where i w i w i i wallace and henry (1987) 8 chevron 3 1 10 : log 3 log ic ic a b i mix c a c b c i c i where i v i v i i riazi (2005) 9 cragoe 1000 ln(20) 0.0005 exp : 1000 ln(20) ln 0.0005 a b i mix cr cr a cr b cr cr i i where i w i w i i cragoe (1933) experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the …… 85 10 latour exp exp (1 ) ln 1 : ln ln ln 1 0.09029 0.1351 n mix b b a b b b a w where a n miadonye et al. (2000) 11 shan-peng #1 log log log : 0.0613 log log 0.134 mix a a b b ab a b ab a b v v c v v where c shan-peng et al. (2007) 12 shan-peng #2 log log log log log log : 0.0644 log log log log 0.1706 mix a a b b ab a b ab a b v v c v v where c shan-peng et al. (2007) 13 al-besharah 3 2ln ln ln 4.976 10 ( )mix a a b b a b a bv v v v api api al-besharah (1989) table 4. tuned coefficients of new proposed models based on least square (ls) and genetic algorithm (ga) methods. coef. model # 1 model # 2 ls ga ls ga c1 0.0892 0.0776 0.6727 1.7637 c2 0.1092 0.1000 2.0105 1.2641 c3 -0.3441 11.7985 -0.4117 -0.8944 c4 3.1558 4.5531 c5 -0.0436 4.5531 6. results and discussion the accuracy of all models, including the new proposed models, has been examined using available experimental data. for this task, 900 kinematic viscosity binary data were taken from the database and the calculated values were compared with experimental data points. these comparisons were in terms of average absolute relative deviation percent (aard %) which is expressed as follows: exp exp 1 1 % 100 n cal i i ii aard n (9) the published models given in table 3, were used for calculating the viscosity of crude oil mixture. the statistical parameters of these models are presented in table 5 which show that koval and parkash models with aard<14% are the best models for the blend viscosity calculation. the new proposed models in the present work were developed based on the ls and ga methods using the same set of data. the objective function as the fitness function was defined. parameter estimation is considered as a minimization of an objective function which minimizes the deviation between the experimental and predicted binary viscosities with the best model parameters. the objective function used in this work read as: exp exp 1 . n cal i i ii v v o f v (10) where v is the kinematic viscosity of crude oil blends and superscripts exp. and cal. refer to the experimental and calculated viscosity of blends, respectively. a flowchart for the calculation procedure for one run based on ga method is shown in fig. 1. 87 r.s. al-maamari, g. vakili-nezhaad and m. vatani 86 to find the minimum objective function value, more than 10 runs were applied with the random initial values of population at different values of the lower and upper bounds and various input parameters (which include crossover constant, mutation probability, and population size) in ga toolbox. each new generation in ga decreases the objective function values. due to different performance of each operator, the obtained results by each operator showed different values of the objective function. the algorithm continues to find the minimum of the evaluated fitness function until termination criteria were reached. the most frequently used termination criterion is the specified maximum number of generations. a more complete discussion and flowchart of ga toolbox can be found in a previous work (vakilinezhaad et al. 2014). finally after more than 10 runs, the lowest values of the objective function along with the corresponding parameters were selected as the final results. the final results in terms of equations (9) are presented in table 5. these findings showed that ga could produce more accurate and global results compared to the ls method, so parameters obtained using ga fitted the experimental data with higher accuracy. to validate the new models, the remaining 100 experimental data for the binary mixtures of b and c which were not used in the regression analysis were tested. table 5 represents the aard% values of the new models as well as available models for this data set. as shown in table 5, the ga based models proposed in this work provided the minimum aard% compared to all available models in the literature. a comparison between aard% for each model is shown in fig. 2. 7. conclusions based on the existing literature and our experimental data, two new models were developed for predicting the viscosity of crude oil binary mixtures. in the first model, the table 5. aard% of this study compared with other models. model aard% 900 data 100 data arrenius 41.275 7.362 bingham 33.976 7.342 koval 13.807 7.204 parkash 13.868 7.208 refutas 15.187 7.385 maxwell 14.289 7.242 wallace 14.899 7.398 chevron 16.209 7.256 cragoe 19.595 7.432 latour 14.094 7.671 shan-peng #1 36.721 7.637 shan-peng #2 29.423 23.408 al-besharah 18.949 6.111 model #1 (ls) 33.546 17.123 model #1 (ga) 13.113 7.125 model #2 (ls) 14.195 7.146 model #2 (ga) 13.672 7.080 viscosity of mixture is a function of mass function, api gravity and viscosity of each component, while the second model is a function of volume fraction and viscosity of each component. adjustable parameters of the proposed models were calculated based on two approaches; genetic algorithm and least square methods. according to the calculated aard%, the genetic algorithm based new proposed models have higher accuracy compared to the available published models, which shows the superiority of ga method over the traditional ls method for parameter estimation of the empirical models. the present study while ls method uses only the single joint in each iteration, generates new point for next step by a de3terministic computation and outcome of it depends on initial guess, ga explores the search space using 88 experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the …… 86 no yes figure 1. flow chart for the calculation procedure based on ga method. figure 2. average absolute relative deviation percent (aard %) for each model. print adjustable parameters input fraction and viscosity of each component estimate adjustable parameters by ga calculate viscosity of mixture by model ( ) compare and by objective functions are termination criteria reached? 9089 r.s. al-maamari, g. vakili-nezhaad and m. vatani 85 randomly generated multiple points, creates new population by a probabilistic computation and the outcome of it, is not related to initial guess, so the accuracy of ga is higher than ls in optimization problems and the results of the present work approved it. the proposed models can be used for better prediction of crude oil blends physical properties that are needed in the calculations of power required for crude oil transportation in pipelines. references al-besharah jm (1989), the effect of blending on selected physical properties of crude oils and their products. phd dissertation: university of aston, birmingham, alabama. alvarez vh, larico r, ianos y, aznar m (2008), parameter estimation for vle calculation by global minimization: the genetic algorithm. brazilian journal of chemical engineering 25:409–418. arrhenius sa (1887), uber die dissociation der in wasser gelosten stoffe. zeitschrift fur physikalische chemie-leipzig 1:631–648. baird ct (1989), iv guide to petroleum product blending. austin (tx): hpi consultants, inc. bingham ec (1914), the viscosity of binary mixtures. the journal of physical chemistry 18:157–165. bonilla-petriciolet a, fateen sek, rangaiah gp (2013), assessment of capabilities and limitations of stochastic global optimization methods for modeling mean activity coefficients of ionic liquids. fluid phase equilibria 340:15-26. centeno g, sánchez-reyna g, ancheyta j, muñoz jad, cardona n (2011), testing various mixing rules for calculation of viscosity of petroleum blends. fuel 90:3561-3570. chamkalani a, maesoumi a, sameni a (2013), an intelligent approach for optimal prediction of gas deviation factor using particle swarm optimization and genetic algorithm. journal of natural gas science and engineering 14:132-143. cragoe cs (1933), changes in the viscosity of liquids with temperature, pressure and composition. process world petroleum congress london 2:529–541. diaz rm, bernardo mi, fernbndez am, folgueras mb (1996), prediction of the viscosity of lubricating oil blends at any temperature. fuel 75:574-578. holland jh (1992), algoritmos geneticos. investigacion y ciencia 192:38–45. kirkpatrick s, gelatt cd, vecchi mp (1983), optimization by simulated annealing. science 220:671–680. koval ej (1963), a method for predicting the performance of unstable miscible displacement in heterogeneous media. spe journal 3:145-154. maddinelli g, pavoni s (2013), method for predicting the properties of crude oils by the application of neural networks. us 20130103627 a1. mathworks (2004–2006), genetic algorithm and direct search toolbox: users guide. copyright by the math works, natick, mass. maxwell jb (1950), data book on hydrocarbons: application to process engineering. princeton, new jersey: d. van nostrand company, inc. miadonye a, latour n, puttagunta vr (2000), a correlation for viscosity and solvent mass fraction of bitumen-diluent mixtures. petroleum science and technology 18:1–14. parkash s (2003), refining processes handbook. oxford, uk: elsevier (gulf professional publishing). preechakul c, kheawhom s (2009), modified genetic algorithm with sampling techniques for chemical engineering optimization. journal of industrial and engineering chemistry 15:110118. price kv, storn rm, lampinen ja (2005), differential evolution: a practical approach to global optimization. springer, berlin heidelberg. rahmes vh, nelson wl (1948), viscosity blending relationships of heavy petroleum oils. analytical chemistry 20:912-915. rangaiah gp (2001), evaluation of genetic algorithms and simulated annealing for phase equilibrium and stability problems. fluid phase equilibria 187–188:83–109. riazi mr (2005), characterization and properties of petroleum fractions, first ed. american society and testing materials, astm manual series: mnl50, usa. sahoo rk, banerjee t, ahmad sa, khanna a (2006), improved binary parameters using ga for multi-component aromatic extraction: nrtl model without and with closure equations. fluid phase equilibria 239:107–119. 90 experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the …… 91 sahoo rk, banerjee t, khanna a (2007), uniquac interaction parameters with closure for imidazolium based ionic liquid systems using genetic algorithm. the canadian journal of chemical engineering 85:833–853. shan-peng h, wen-xue j, jin-jun z (2007), approaches to predict viscosities of crude oil blends. journal of central south university of technology 14:466-470. singh mk, banerjee t, khanna a (2005), genetic algorithm to estimate interaction parameters of multicomponent systems for liquid–liquid equilibria. computers and chemical engineering 29:1712–1719. stragevitch l, davila sg (1997), application of a generalized maximum likelihood method in the reduction of multicomponent liquid-liquid equilibrium data. brazilian journal of chemical engineering 14:41–52. tan m, zou y, wang x, guo y (2014), a novel wireline logs based approach for assessment of mineral concentrations in organic shale. oil shale 31(2):132–146. vakili-nezhaad g, vatani m, asghari m (2013), calculation of the binary interaction and nonrandomness parameters of nrtl, nrtl1, and nrtl2 models using genetic algorithm for ternary ionic liquid systems. chemical engineering communications 200:1102–1120. vakili-nezhaad g, vatani m, gujarathi am (2014), application of genetic algorithm to calculation of three-suffix margules parameters in ternary extraction ionic liquid systems. international journal of thermodynamics 17:1-6. vatani m, asghari m, vakili-nezhaad g (2012a), application of genetic algorithm to the calculation of parameters for nrtl and twosuffix margules models in ternary extraction ionic liquid systems. journal of industrial and engineering chemistry 18:1715–1720. vatani m, asghari m, vakili-nezhaad g (2012b), application of genetic algorithm to parameter estimation in liquid-liquid phase equilibrium modeling. the journal of mathematics and computer science 5(1):60-66. wallace d, henry d (1987), a correlation for correcting the viscosity of solvent-extracted bitumen to a solvent-free basis. aostra journal of research 3:239-249. xue y, cheng l, mou j, zhao w (2014), a new fracture prediction method by combining genetic algorithm with neural network in lowpermeability reservoirs. journal of petroleum science and engineering 121:159–166. yuan w, hansen ac, zhang q, tan z (2005), temperature-dependent kinematic viscosity of selected biodiesel fuels and blends with diesel fuel. journal of the american oil chemists society 82:195-199. c:\users\raoh\desktop\paper 5.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 58-71 performance analysis of a multiple micro-jet impingements cooling model a. husain*, a, n.a. al-azria, a. samadb, k.y. kimc *a department of mechanical and industrial engineering, sultan qaboos university, muscat, oman. b department of ocean engineering, indian institute of technology, madras, chennai, india. c department of mechanical engineering, inha university, incheon, republic of korea. received 14 april 2014; accepted 21 october 2015 abstract: the present study investigates the thermal performance of a multiple micro-jet impingements model for electronics cooling. the fluid flow and heat transport characteristics were investigated for steady incompressible laminar flow by solving three-dimensional (3d) navier-stokes equations. several parallel and staggered micro-jet configurations (ie. inline 2 × 2, 3 × 3 and 4 × 4 jets, and staggered five-jet and 13-jet arrays with the jet diameter to the channel height ratios from 0.25–0.5) were analyzed at various flow rates for the maximum temperature rise, pressure drop, heat-transfer coefficient, thermal resistance, and pumping power characteristics. the parametric investigation was carried out based on the number of jets and the jet diameters at various mass flow rates and jet reynolds numbers. temperature uniformity and coefficient of performance were evaluated to find out the trade-off among the various designs investigated in the present study. the maximum temperature rise and the pressure drop decreased with an increase in the number of jets except in the case of staggered five-jet array. a higher temperature uniformity was observed at higher flow rates with a decrease in the coefficient of performance. the performance parameters, such as thermal resistance and pumping power, showed a conflicting nature with respect to design variables (viz. jet diameter to stand-off ratio and interjet spacing or number of jets) at various reynolds numbers within the laminar regime. keywords: jet impingement, thermal resistance, pumping power, temperature uniformity, heat transfer coefficient, coefficient of performance. رینولد * corresponding author’s e-mail: afzal19@squ.edu.om a. husain, n.a. al-azri, a. samad and k.y. kim 59 nomenclature as surface area of the substrate base, m2 cp specific heat at constant pressure, j kg-1k-1 cop coefficient of performance d diameter of the nozzle, m h height of the channel, m h heat transfer coefficient, w m-2 k-1 k thermal conductivity, w m-1 k-1 l length of the nozzle, m lx, ly, lz length, width and height of the silicon substrate, respectively, m n number of observations for calculating the mean value n number of jets p, δp pressure and pressure-drop, respectively, pa p pumping power, w pumping power flux, w m-2 q heat flux, w m-2 rth thermal resistance, k w-1 s interjet spacing, m t, δt temperature and temperature-rise, respectively, k ts thickness of the substrate base. u, v, w orthogonal components of fluid velocity, m s-1 v velocity vector volume flow rate, m3 s-1 x, y, z orthogonal coordinate system m mass flow rate, kg s-1 greek symbols µ dynamic viscosity, kg s-1 m-1 density, kg m-3 σ standard deviation, k subscripts 4,5,9,13,16 number of jets avg area-average value f fluid i inlet j jet max maximum value mean mean value n vector in wall/plane normal direction o outlet s substrate "p v performance analysis of multiple micro-jet impingements cooling model 60 1. introduction the growing demand for higher heat flux removal and the occurrence of hot spots from ultra-large scale integrated (ulsi) circuits of power electronics, concentrated photovoltaic, and highpower light emitting diodes (leds) have posed a major challenge to heat transfer researchers. the sensitivity, reliability, performance and efficiency of electronic devices are highly temperature dependent and a high temperature may lead to irreparable loss of electronics (hamann et al. 2007). the increase in junction temperature decreases the life span of leds exponentially (narendran and gu 2005). the heat generated at the p-n junction is transferred to the substrate material which must be dissipated to allow proper functioning of electronics. these challenges have encouraged researchers to develop new, alternative cooling techniques at the miniand micro-levels. as the air cooling techniques reach their limits, the liquid flow through parallel micro-channels provides a potential solution for the high heat-flux management of electronics. tuckerman and pease (1981) were the first to examine liquid-flow microchannel heat sinks. the liquid flows from one end to the other through narrow parallel channels, leaving higher temperatures at the inlet side and lower temperatures at the outlet side (husain and kim 2008a; kawano et al. 1998; knight et al. 1992; liu and garimella 2005; toh et al. 2002; qu and mudawar 2002). the hot-spot occurrence and high temperaturegradients lead to temperature non-uniformity on the heated surface and, consequently on the chip surface, inducing thermal stresses which eventually reduce the life span of the electronics. jet impingement cooling offers an efficient solution for high heat-flux management of an entire surface as well as hot spots in power electronics with relatively more uniform temperature distribution over the chip surface. the recent developments in electronics cooling have directed researchers’ endeavors towards achieving higher temperature uniformities as well as hot-spot management by the application of micro-jet impingements on the heated substrate (lee and vafai 1999; wu et al. 1999). wu et al. (1999) examined experimentally the heat transfer characteristics of 500 μm and 550 μm single jets as well as a jet-array with h ranging from 200–3000 μm. they obtained an area-averaged heat transfer coefficient of 320 w/m2 k for a 500 μm single jet with h = 750 μm at a pressure-drop of about 34.5 kpa. they also reported a higher cooling efficiency at a lower driving pressure, which decreased with further increase in the driving pressure. a comparative analysis of the microchannels and jet impingement cooling shows that the micro-channel cooling is preferable for a heated surface size smaller than 70 mm × 70 mm; and the jet impingement is comparable to or better than the micro-channel cooling for a large target plate with a proper arrangement of spent flow after impingements (lee and vafai 1999). the potential of micro-jet impingement cooling in electronics has motivated researchers to examine micro-jet impingements cooling models both numerically and experimentally. moreover, with the advent of advanced fabrication processes, the production of micro-scale jets has become more convenient than before (jackson 2006). wang et al. (2004) investigated experimentally singleand twophase flow micro-jet impingement cooling systems. they fabricated a single and multi-jet array with circular orifice diameters ranging from 40–76 µm, which removes 90 w using a four-jet array at a flow rate of eight ml/min with a temperature rise of 100 oc for a 1 cm × 1 cm chip. fabbri and dhir (2005) examined impinging jet arrays in a circular pattern with three different radial and circumferential pitches of 1, 2 and 3 mm. they investigated free jets of both water and fc-40 at reynolds numbers ranging from 73–3813, and reported heat transfer coefficients ranging from 6–60 kw/m2 k. sung and mudawar (2006) proposed a hybrid cooling scheme for electronics cooling, which combines the cooling benefits of both microchannel and jet impingement. they carried out experimental and numerical analyses of slot-jet impingements in the microchannels which showed low rise and small temperature gradients across the surface. michna et al. (2009) experimentally investigated an array of micro-jets for stagnation heat transfer and obtained a maximum 11 × 106 w/m2 heat flux for the rise of a surface temperature less than 30oc. they observed that the reynolds number, the prandtl number, and the area ratio (total area of jets divided by the surface area) can have significant effects on heat transfer. browne et al. (2010) reported micro-jet arrays of jet diameters 52 µm and 112 µm and a jet spacing of 250 µm for heat transfer characteristics. they observed similar trends for micro-jet arrays for conventionaland macro-sized jet arrays. husain et al. (2013a) investigated multiple micro-jet impingement heat sinks and carried out parametric optimization to find out optimal designs. in other studies, husain et al. (2013b; 2013c) proposed turbulent and laminar flow numerical optimization models of jet impingement system for high power leds and a. husain, n.a. al-azri, a. samad and k.y. kim 61 ulsi circuits cooling, respectively. wang and peles (2014) investigated heat transfer enhancement through passive and active flow control by a jet issued from a micro pillar in a channel. the combined effect of micro-jet and pillar-enhanced spatially-averaged the nusselt number by 80%. macro-jet impingement cooling has been the subject of intense research with its applications in macro-heat sinks and turbomachines before the micro-jet impingement was applied in electronics cooling and micro-systems. the macro-jet impingement has a rich history, with most of the investigations based on air jet impingements (garimella and rice, 1995; webb and ma 1995; womac et al. 1993; womac et al. 1994;). womac et al. (1993) proposed correlations for heat transfer and concluded that the h/d ratio has a negligible effect on heat transfer, but an increase in velocity improves heat transfer. garimella and rice (1995) characterized a single submerged and confined jet for local heat transfer. they examined the effects of d, h/d, l/d, and the flow rate on heat transfer and proposed correlations based on the jet reynolds number. heo et al. (2011) investigated an inclined elliptic jet for cross-flow analysis through numerical methods, observing that the size of the recirculation region just upstream of the jet and the length of the potential core of the jet significantly affect the heat transfer process. in view of the limited literature on the liquid flow multiple micro-jet impingement cooling and optimal jet-array configuration,, the present study is devoted to investigating thermal performance and carries out parametric analyses of a liquid-flow multiple micro-jet impingements cooling model built on the backside of a silicon chip substrate. the effectiveness of several multi-jet configurations (ie. inline 2 × 2, 3 × 3, and 4 × 4 jets and staggered fivejet and 13-jet arrays with nozzle diameters of, 50 µm, 76 µm, and 100 µm) were analyzed at different flow rates to determine the thermal and pressuredrop characteristics. 2. mathematical model and simulation strategy 2.1 geometrical model a multiple micro-jet impingements cooling model consisting of a fluid channel can be fabricated on a silicon substrate and micro-nozzles on a cover plate (fig. 1). the design can be implemented at the backside of a silicon chip that emits heat flux from the integrated circuitry. the nozzles were designed on the substrate, and a water channel was created to allow liquid to pass through the nozzles as discussed by wang et al. (2004). the liquid jets take the heat while striking at the heated surface. the flow then turns to the radial direction and reaches the outlet ports (fig. 1). for a conservative analysis, the substrate on which these nozzles were designed was not taken for analysis, whereas the nozzles with a fluid channel and substrate base were included in the computational domain (fig. 1(b)). the nozzles and outlet channel of the substrate can be fabricated separately and bonded together with epoxy to form the designed structure as discussed by wang et al. (2004). the overall dimensions of the micro-jet impingements domain (fig. 1) are 10 mm × 10 mm × 0.6 mm. the thickness of the substrate base is 100 µm and the height of the liquid channel is 200 µm. the thickness of the nozzle plate in which multiple nozzles are designed is 300 µm. one half of the entire domain was taken as the computational domain due to the symmetry about the central plane (fig. 1). the various parameters that affect the performance of the model are the thickness of the substrate base (ts), nozzle diameter (d), length of the nozzle (l), interjet spacing (s), number of nozzles (n), channel height (h), and the distribution of the jets (ie. the inline and staggered arrays) (fig. 2). the geometric parameters, l and ts, were kept constant due to their unidirectional effects on the model performance (liu and garimella 2005; michna et al. 2009), whereas the channel height (h) was kept constant to keep concise the parametric investigations. 2.2 mathematical model and boundary conditions three-dimensional (3d) numerical analyses were performed for steady state incompressible laminar flow and conjugate heat transfer. the coolant, deionized ultra-filtered (diuf) water, was allowed to flow through the micro-nozzles and channel. the governing equations of the model (viz. conservation of mass, momentum and energy) for fluid flow and conjugate heat transfer can be written in vector form as follows: ( ) 0fv (1) ( ) ( )f fpv v v (2) for the fluid ,( ) ( )f p f f f fc t k tv (3) for the substrate ( ) 0s sk t (4) performance analysis of multiple micro-jet impingements cooling model 62 figure 1. schematic of multiple micro-jet-impingements model. (a) cover plate with jet holes, and (b) computational domain with boundary conditions. the numerical simulations were carried out using a commercial computational fluid dynamics (cfd) code that employs coupled algebraic multigrid method (cfx theory 2006; raw 1996). the thermophysical properties (incropera and dewitt 2002) of the coolant diuf water were allowed to vary with temperature and updated after every iteration to take micro-scale effects into account a. husain, n.a. al-azri, a. samad and k.y. kim 63 figure 2. various micro-jet configurations on the nozzle plate. (a) inline 2 × 2 jet array, s4 = 5 mm, (b) staggered five-jet array, s5 = 7.0 mm, (c) inline 3 × 3 jet array, s9 = 3.5 mm, (d) staggered 13-jet array, s13 = 4 mm, and (e) inline 4 × 4 jet array, s16 = 2.5 mm (herwig an mahilikar 2006; husain and kim 2008b; toh et al. 2002) during the iterative solution. a nonuniform hexahedral mesh was used in a specified computational domain to implement a numerical scheme (fig. 3(a)). a high resolution scheme was applied to discretize advection terms of the governing equations. the top and bottom faces of the fluid domain interfaced with the substrate base and cover plate, respectively. the interior walls of the nozzles were assigned with adiabatic boundary conditions along with a no slip ( 0v ) boundary condition. for conservative analysis, a constant heat flux was assigned as a heat source at the base of the substrate. a no-slip condition ( 0v ) was applied at the interior walls of the nozzles and channel. the thermal boundary conditions are: s t k q z at 0z , 2 2 x xl lx , and 0 2 yly 0s t k z at 2 2 x xl lx , and 0 2 yly 0s t k n at the nozzle walls. the boundary conditions on the symmetric plane are as follows: 0s t k n , 0v , 0 u n , 0 p n , and 0 w n for the solid-liquid interface, the conditions are s ft t and fs s f tt k k n n the overall thermal resistance is defined as max th s t r qa (5) the maximum temperature-rise in the substrate is defined as max ,max ,s f it t t (6) s16s13 s9s4 d s5 performance analysis of multiple micro-jet impingements cooling model 64 (a) (b) figure 3. (a) unstructured hexahedral grid, and (b) comparison of numerical results with the results of wang et al. (2004) for an inline 2 × 2 array and staggered 13-jet array. the pumping power required to drive the fluid through the model can be evaluated as follows: p v p (7) the pressure-drop is defined as: o ip p p (8) the reynolds number at the jet inlet was defined as: ,re j i n v d (9) where vj,i is the velocity at the inlet. the heat transfer coefficient was evaluated as ,( )j i s q h t t (10) the coefficient of performance (cop) can be represented as the ratio of the heat flux to the pumping-power flux: q cop p (11) where p is defined as s p p a . power (w) 5 10 15 20 20 40 60 80 d/h = 0.38 wang et al. (2004) d/h = 0.38 present scheme d/h = 0.2 wang et al. (2004) d/h = 0.2 present scheme flow rate 8 ml/min inline 2x2 array staggered 13-jet array a. husain, n.a. al-azri, a. samad and k.y. kim 65 the degree of temperature uniformity at the impingement surface can be assessed in terms of the standard deviation (σ) as follows: temperature uniformity, 12 where t1, t2, t3 ……….….. tn are the temperatures at n sample points on the surface, and tmean is the mean surface temperature. the temperature uniformity (-σ) at the surface increased with a decreasing σ value. for a conservative analysis and simplification of the numerical model, radiation effects were ignored. the convergence was checked by the residual target of 10-6. 3. results and discussion the grid dependency tests were carried out for both solid and liquid domains, and a grid of approximately 1,000,000 elements was used for a typical design of d/h = 0.33 in an inline 2 × 2 jetarray. for a grid of 500,000 elements, the changes in pressure-drop (δp) and maximum temperature-rise (δtmax) were more than 4%, whereas the changes in δp and δtmax were less than 2% for grid systems of 1,500,000 and 2,000,000 elements. the validation of the present numerical scheme for a laminar flow in a channel and jet impingement, respectively, was reported in the authors’ previous works (husain and kim 2008a; husain et al. 2013a). further validation of the micro-jet impingement system is presented in fig. 3(b) for an inline 2 × 2 jet-array and staggered 13-jet array. the numerical reproduction of the data showed reasonable agreement with the experimental results reported by wang et al. (2004). the temperature distribution and uniformity on the heated surface were presented by local temperature plots and the overall standard deviation of the temperature at the solid-fluid interface (fig 3). figure 4 shows the temperature distribution at the impingement surface at a constant mass flow rate. a substantially small temperature range and higher temperature uniformity were observed for the design with the higher number of jets. the high temperature zones were present in the interjet spacings away from the stagnation region. with the increased number of jets, the interjet spacing decreased, resulting in high fluid mixing and the occurrence of secondary flow. the strength of the secondary flow increased with the further decrease in the interjet spacing. the increase in number of impingements resulted in increased temperature uniformity and decreased temperature-gradient monotonously. the maximum-temperature-rise increases with the increase in jet diameter to the channel height ratio (fig. 5) shows the temperature distribution and uniformity at the impingement surface for d/h = 0.2–0.5 and 45.32 10m kg.s-1. at a constant mass flow rate, the jet velocity and jet reynolds number decreased with the increase in d/h ratio. the reduced jet reynolds number led to a decrease in the heat transfer coefficient and an increase in the δtmax (fig. 5). the performance of the various jet configurations (ie. inline 2 × 2, 3 × 3, and 4 × 4 jets and staggered five-jet and 13-jet arrays) were analyzed for the area-averaged heat transfer coefficient (havg), δtmax, and pressure drop at a constant mass flow rate of 1.33×10-4 kg s-1 for d/h = 0.25, d/h = 0.33 and d/h = 0.5 (fig. 6). at a constant mass flow rate, increasing the number of jets led to decreasing the mass flow rate per jet and the jet reynolds number. in general, increasing the number of jets decreased the δtmax except for the increase in number of jets from 4 to 5 (fig. 6(a)) where interjet spacing increased significantly even with the increase in the number of jets as shown in figure 2’s configuration. the increase in interjet spacing developed high temperature zones between the adjacent jets. the heat transfer coefficient and the pressure-drop decreased consistently with an increase in the number of jets, showing the weakened effects of stagnation heat transfer due to the decrease in jet velocity for all parameters explored in the present study (figs. 6(b) and (c)). the increase in the number of jets significantly decreased the δtmax and the pressure drop. the inline 4 × 4 jet array design was selected for further analysis of performance parameters. the designs with lower δtmax were achieved with higher pressure-drop penalties (fig. 7). in view of the limited pumping powers available at the microlevel and to maintain flow within the laminar regime, the range of the mass flow rates was set to vary from 1.33 × 10-4 to 7.98 × 10-4 kg s-1. the δtmax and pressure-drop change nonlinearly with the d/h ratio and mass flow rate (fig. 7(c)). the higher jet velocities, which stem from the increased mass flow rate, increased the local as well as the area-averaged heat transfer coefficients, which finally decreased the δtmax. with increasing performance analysis of multiple micro-jet impingements cooling model 66 (a) (b) (c) (d) (e) figure 4. temperature distribution and uniformity (-σ) on the jet impingement surface for various microjet configurations at d/h = 0.5 and a mass flow rate of 1.33×10-4: (a) inline 2 × 2 jet array, -σ = -9.81, (b) staggered five-jet array, -σ = -13.37, (c) inline 3 × 3 jet array, -σ = -6.6, (d) staggered 13-jet array, -σ = -5.24, and (e) inline 4 × 4 jet array, -σ = -3.4. coolant flow rates, the δtmax decreased and pressure drop increased consistently, but higher coefficient of performance (cop) values were associated with the low flow rates (table 1). although an increase of mass flow rate decreased the cop of the model, it improved the temperature uniformity (-σ) at the heated surface. table 1 highlights the conflicting nature of the cop and the temperature uniformity with the increasing mass flow rate for an inline 4 × 4 jets array. the functional characteristics between the δtmax and pressure drop showed that a higher d/h ratio gives a lower maximum-temperature rise, lower temperature uniformity and pressure drop but a higher cop, making it advantageous within the range investigated in this study. y x a. husain, n.a. al-azri, a. samad and k.y. kim 67 (a) (b) (c) figure 5. temperature distribution and uniformity (-σ) on the jet impingement surface for inline 4 × 4 jet array at a mass flow rate of 5.32 × 10-4 kg/s. (a) d/h = 0.25, -σ = -2.57, (b) d/h = 0.33, -σ = -2.71, and (c) d/h 0.5, -σ = -2.66. (a) (b) (c) figure 6. effect of the micro-jet configuration and number of jets on thermal performance at a mass flow rate of 1.33 × 10-4 kg/s and d/h = 0.25–0.5. (a) area-averaged heat transfer coefficient (havg), (b) maximum temperature-rise (δtmax), and (c) pressure drop (δp). n 5 10 15 5000 5500 6000 6500 d/h = 0.25 d/h = 0.33 d/h = 0.5 flow rate 1.33e-4 kg/s n 5 10 15 60 70 80 d/h = 0.25 d/h = 0.33 d/h = 0.5 flow rate 1.33e-4 kg/s n 5 10 15 0 40 80 120 160 d/h = 0.25 d/h = 0.33 d/h = 0.5 flow rate 1.33e-4 kg/s y x performance analysis of multiple micro-jet impingements cooling model 68 (a) (b) (c) figure 7. effect of the mass flow rate on the maximum temperature rise (δtmax) and pressure drop (δp) characteristics for an inline 4 × 4 jet array configuration. (a) maximum temperature rise (δtmax), (b) pressure drop (δp), and (c) maximum temperature rise (δtmax) versus pressure drop (δp). the jet impingement flow field can be assumed to be laminar for a jet reynolds number <1000 (zuckerman and lior 2006). the increase in the jet reynolds number caused an increase in the stagnation heat transfer coefficient, which consequently increased the area-averaged heat coefficient of the jet impingements. the areaaveraged heat transfer coefficient increased monotonously with the increasing jet reynolds number (fig, 8(a)) for an inline 4 × 4 jet array with d/h = 0.33–0.5, resulting in a consistent decrease in the thermal resistance (rth). the pressure drop flow rate (kg s-1) 0.0002 0.0004 0.0006 0.000 20 30 40 50 60 d/h = 0.25 d/h = 0.33 d/h = 0.5 4x4array, 16-jet flow rate (kg s-1) 0.0002 0.0003 0.0004 0.0005 0 40 80 120 160 d/h = 0.25 d/h = 0.33 d/h = 0.5 4x4 array, 16-jet p (kpa) 101 102 20 40 60 d/h = 0.25 d/h = 0.33 d/h = 0.5 4x4 array, 16-jet 5.32e-4 kg s-1 3.99e-4 kg s-1 2.66e-4 kg s-1 1.33e-4 kg s-1 6.65e-4 kg s-1 7.98e-4 kg s-1 a. husain, n.a. al-azri, a. samad and k.y. kim 69 (a) (b) (c) (d) figure 8. characteristic variation of area-averaged heat transfer coefficient (havg), thermal resistance (rth), and pumping power (p) with reynolds number (re). (a) area-averaged heat transfer coefficient (havg) versus reynolds number (re), (b) thermal resistance (rth) versus reynolds number (re), (c) pumping power (p) versus reynolds number (re), and (d) thermal resistance (rth) versus pumping power (p). table 1. maximum temperature rise (δtmax), coefficient of performance (cop), and temperature uniformity (-σ) of a multiple micro-jet-impingements cooling model with an inline 4 × 4 jet array at various mass flow rates. mass flow rate (kg s-1) δtmax (k) cop -σ (k) d/h = 0.25 d/h = 0.33 d/h = 0.5 d/h = 0.25 d/h = 0.33 d/h = 0.5 d/h = 0.25 d/h = 0.33 d/h = 0.5 1.33e-4 59.92 92.79 65.26 6603 33672 92439 -3.76 -3.58 -3.40 2.66e-4 36.21 39.68 42.76 1355 6777 18607 -3.15 -3.09 -2.96 3.99e-4 27.38 30.34 33.70 535 2636 7251 -2.85 -2.87 -2.79 5.32e-4 21.95 24.92 28.07 278 1349 3725 -2.57 -2.71 -2.66 6.65 e-4 21.32 23.95 803 2229 -2.57 -2.54 7.98 e-4 18.84 20.96 525 1468 -2.45 -2.44 δtmax = maximum temperature rise; cop = coefficient of performance; -σ = temperature uniformity. re 200 400 600 800 1000 6000 10000 14000 18000 22000 26000 d/h = 0.25 d/h = 0.33 d/h = 0.5 inline 4x4 jet array re 200 400 600 800 1000 1 2 d/h = 0.25 d/h = 0.33 d/h = 0.5 inline 4x4 jet array re 200 400 600 800 1000 0 0.02 0.04 0.06 0.08 0.1 d/h = 0.25 d/h = 0.33 d/h = 0.5 inline 4x4 jet array p (w) 0 0.02 0.04 0.06 0.080.6 1 1.4 1.8 2.2 2.6 d/h = 0.25 d/h = 0.33 d/h = 0.5 inline 4x4 jet array performance analysis of multiple micro-jet impingements cooling model 70 increased consistently with the increase in the reynolds number due to the higher friction losses in the impingement system, consequently resulting in a higher pumping power (p) to drive the fluid through the domain (fig. 8(c)). the velocity of the fluid passing through the nozzles increased with the decrease in d/h ratio at the constant jet reynolds number, which resulted in an increase in the pressure drop and, consequently, an increase in pumping power (fig. 8(c)). on the other hand, higher jet velocities increased the stagnation as well as area-averaged heat transfer coefficient (martin 1997), which consequently reduced the convective thermal resistance of the system. the two performance parameters of thermal resistance and pumping power showed a conflicting nature with the change of the design variable or the reynolds numbers (fig. 8(d)). the information in fig. 8(d) suggests that lower thermal resistance can be obtained at the expense of higher pumping power and vice versa. the optimum designs can be obtained from the trade-off curves based on the available pumping power and required thermal resistance. 4. summary and conclusion this study examined the thermal and pressuredrop characteristics of a silicon-based micro-jet impingement cooling model for electronic cooling through 3d numerical analyses. the different micro-jet configurations (ie. inline 2 × 2, 3 × 3, and 4 × 4 jets and staggered five-jet and 13-jet arrays with various jet diameter-to-channel height ratios) were examined to determine the temperature and pressure drop, heat transfer coefficient, thermal resistance and pumping power characteristics for a range of flow rates and reynolds numbers. the convection and radiation at the outer walls were neglected to maintain a conservative thermal analysis. the multiple impingements produced higher temperature uniformity on the heated surface for the jets of all diameters. for the constant flow rates across the domain, the increase in jet diameter increased the maximum-temperature rise in the substrate. in general, the temperature rise and pressure drop decreased with an increase in number of jets. the increase in flow rates, though, increased the temperature uniformity but decreased the coefficient of performance of the model. the thermal resistance and pumping power showed a conflicting nature with the design parameters and jet reynolds numbers. acknowledgment the authors would like to acknowledge the support of sultan qaboos university through an internal research grant (ig/eng/meid/14/03) for conducting this research. references browne ea, michna gj, jensen mk, peles y (2010), experimental investigation of single-phase microjet array heat transfer. journal of heat transfer 132: 041013–19. cfx-11.0 (2006), solver theory, ansys europe ltd. fabbri m, dhir vk (2005), optimized heat transfer for high power electronic cooling using array of microjets. journal of heat transfer 127: 760–769. garimella sv, rice ra (1995), confined and submerged liquid jet impingement heat transfer. asme journal of heat transfer 117(4): 871–877. hamann hf, weger a, lacey ja, hu z, bose p, cohen e, wakil j (2007), hotspot-limited microprocessors: direct temperature and power distribution measurements. ieee journal of solidstate circuits 42(1): 56–65. heo mw, lee kd, kim kw (2011), optimization of an inclined elliptic impinging jet with cross flow for enhancing heat transfer. heat and mass transfer 47: 731–742. herwig h, mahulikar sp (2006), variable property effects in single-phase incompressible flows through microchannels. international journal of thermal science 45: 977–981. husain a, kim ky (2008a), optimization of a microchannel heat sink with temperature dependent fluid properties. applied thermal engineering 28: 1101–1107. husain a, kim ky (2008b), shape optimization of micro-channel heat sink for micro-electronic cooling. components and packaging technologies, ieee transactions 31(2): 322–330. husain a, kim jh, kim ky (2013a), thermal performance of a silicon-based multiple micro-jet impingement heat sink, proceedings of asme 2013 international technical conference and exhibition on packaging and integration of electronic and photonic microsystems, interpack2013, july 16-18, burlingame, ca, usa, ipack2013-73063. husain a, kim sm, kim jh, kim ky (2013b), thermal performance analysis and optimization of multiple micro-jet impingements cooling of high power leds. journal of thermophysics and heat transfer 27(2): 235–245. a. husain, n.a. al-azri, a. samad and k.y. kim 71 husain a, kim sm, kim ky (2013c), performance analysis and design optimization of micro-jet impingement heat sink. heat and mass transfer 49(11): 1613–1624. incropera fp, dewitt dp (2002), fundamentals of heat and mass transfer. new york, ny: john wiley and sons, inc. jackson mj (2006), microfabrication and nanofabrication. boca raton, fl: crc press. kawano k, minakami k, iwasaki h, ishizuka m (1998), development of micro channels heat exchanging. in: nelson ra, jr., swanson lw, bianchi mva, and camci c (eds), application of heat transfer in equipment systems and education. asme, new york, htd-361-3/pid-3: 173–180. knight rw, hall dj, goodling js, jaeger rc (1992), heat sink optimization with application to microchannels. ieee transactions in components, hybrids, and manufacturing technologies 15(5): 832–842. lee dy, and vafai k (1999), comparative analysis of jet impingement and microchannel cooling for high heat flux applications. international journal of heat and mass transfer 42: 1555–1568. liu d, garimella sv (2005), analysis and optimization of the thermal performance of microchannel heat sinks. international journal of numerical methods for heat and fluid flow 15(1): 7– 26. martin h (1997), heat and mass transfer between impinging jets and solid surfaces. advances in heat transfer 13: 1–60 michna gj, browne ea, peles y, jensen mk (2009), the effect of area ratio on microjet array heat transfer. international journal of heat and mass transfer 54: 1782–1790. narendra n, gu y (2005), life of led-based white light sources. ieee/osa (optical society of america) journal of display technology 1(1): 167– 171. qu w, mudawar i (2002), experimental and numerical study of pressure drop and heat transfer in a single-phase micro-channel heat sink. international journal of heat and mass transfer 45: 2549–2565. raw mj (1996), robustness of coupled algebraic multigrid for the navier-stokes equations. in: 34th aerospace and sciences meeting and exhibit, january 15-18, aiaa 96-0297, reno, nv. sung mk, mudawar i (2006), experimental and numerical investigation of single-phase heat transfer using hybrid jet impingement/microchannel cooling scheme. international journal of heat and mass transfer 49: 682–694. toh kc, chen xy, chai jc (2002), numerical computation of fluid flow and heat transfer in microchannels. international journal of heat and mass transfer 45: 5133–5141. tuckerman db, pease rfw (1981), highperformance heat sinking for vlsi. ieee electronic device letters 2(5): 126–129. wang y, peles, y (2014), an experimental study of passive and active heat transfer enhancement in microchannels. journal of heat transfer 136: 031901–111. wang en, zhang l, jiang l, koo jm, maveety jg, sanchez ea, goodson ke, kenny tw (2004), micromachined jets for liquid impingement cooling of vlsi chips. journal of microelectromechanical systems 13(5): 833–842. webb bw, ma cf (1995), single-phase liquid jet impingement heat transfer. advances in heat transfer 26: 105–217. womac dj, ramadhdhyani s, incropera fp (1993), correlating equations for impingement cooling of small heat sources with single circular liquid jets asme. journal of heat transfer 115(1): 106–116. womac dj, incropera fp, ramadhdhyani s (1994), correlating equations for impingement cooling of small heat sources with multiple circular liquid jets. asme journal of heat transfer 116(2): 482– 486. wu s, mai j, tai yc, ho cm (1999), micro heat exchanger by using mems impinging jets. in: proc. of the 12th ieee international conference on micro electro mechanical systems (mems), orlando, fl, 171–176. zuckerman n, lior n (2006), jet impingement heat transfer: physics, correlations, and numerical modeling. advances in heat transfer 39: 565–631. the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 performance enhancement of solar still by using reflectors-jute cloth-improved glass angle b.b. sahoo a,* , and c. subudhi b a school of mechanical engineering, vellore institute of technology vellore, vellore-632 014, india. b department of mechanical engineering, synergy institute of engineering and technology, dhenkanal-759 001, india. abstract: this paper is aimed to improve the productivity of an existing conventional single slope solar still with modifications. in this study, the conventional solar still is modified by using reflectors, jute cloth and improved glass angle. the modified still productivity and performance are investigated experimentally and compared to that of conventional solar still. the results demonstrated that solar still performance is improved with the 10 o glass angle due to better water flow properties and higher solar radiation fall on the still. the results reveal that, with all combined modifications, the water productivity and efficiency of modified solar still increase by 72.18% and 41.51% respectively more than the conventional one. this implies that modified solar still could be adopted for higher water productivity at a lower additional cost. keywords: solar still; efficiency; jute cloth; glass angle; reflector; productivity. التقطير الشمسي باستخدام العواكس و نسيج الجوت و تعديل زاوية الزجاججهاز تحسين أداء ب.ب. ساهو وسي. سوبودي *أ، ب الحالي باجراء بعض ير الشمسي التقليدي أحادي الميل التقطجهاز : يهدف هذا البحث إلى تحسين إنتاجية الملخص التقطير الشمسي التقليدي باستخدام العواكس ، و نسيج الجوت و جهاز التعديالت. ففي هذه الدراسة ،نعمل على تعديل جهاز ا في التقطير الشمسي مخبريا ومقارنتها بمثيالتهجهاز حيث أننا نقوم بدراسة إنتاجية وأداء تعديل زاوية الزجاج. و ذلك يرجع 01التقطير الشمسي يتحسن مع زاوية الزجاج جهاز التقطير الشمسي التقليدي. وقد أظهرت النتائج أن أداء و أشارت النتائج إلى أنه مع كل التقطير.جهاز إلى الخاصية األفضل لتدفق الماء واإلشعاع الشمسي العالي الساقط على على ٪10.10و ٪81.07التقطير الشمسي المعدل تزداد بنسبة جهاز إنتاجية المياه وكفاءة تلك التعديالت مجتمعة إلى أن التقطير الشمسي المعدل لزيادة إنتاجية جهاز التقطير التقليدي. وهذا يعني أنه يمكن تبني جهاز التوالي أكثر من مثيالتها في المياه بتكلفة إضافية أقل. .اإلنتاجية ؛ العاكس ؛ زجاجية زاوية ؛ الجوت نسيج ؛ الكفاءة ؛ الشمسي التقطير جهاز: المفتاحية الكلمات corresponding author’s e-mail: bbs.iitg@gmail.com doi: 10.24200/tjer.vol16iss1pp1-10 mailto:bbs.iitg@gmail.com the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 2 notations ag area of the glass plate (m 2 ) cm centimeter e east lts litere ml millilitre n north qs solar radiation (w/m 2 ) twi water inlet temperature ( o c) o degree λ transmissivity cfd computational fluid dynamics css conventional solar still hfg latent heat of water vapour (j/kg) mcss modified conventional solar still mv mass of distillate output (kg/s) qin solar input (w) t temperature ( o c) w watt o c degree centigrade ηst solar still efficiency (%) 1. introduction the shortage of clean, drinking water because of global warming, climate change and population growth is a concern for the world today. therefore, the challenge for both developing and developed countries is to provide good quality drinking water. of the total water available on the earth, about 97.5% contains salt and harmful bacteria that cannot be directly used for healthy drinking purpose kumar et al. (2015). hence, we need a technology which can produce purified drinking water by distilling brackish and saline water. an efficient, cheap and environment friendly technique to supply clean and safe drinking water is the ‘solar still’. it is a solar energy system technology that is capable of removing a very wide variety of contaminants from water in just one step distillation process bhattacharyya (2013). the solar still is a simple, inexpensive and efficient system. the equipment primarily consists of a shallow basin with a transparent glass or plastic cover. when the radiation heat from the sun falls inside the glass covered chamber, the basin water is heated. then the higher sun temperature evaporates the heated water in the basin. moisture rises and the water vapour condenses when it reaches the cool glass or plastic cover. the water droplets run down into a collection trough to give salts, minerals, germs and impurities free pure water sahoo et al. (2008). in the past, several attempts were made by many researchers to increase water productivity and utilize the maximum solar radiation on the still to increase the system efficiency. iranean scientists rahbar and esfahani have shown splendid interest in the solar still research in all respects; theoretical, numerical and experimental. the researchers developed a new correlation for estimating the convective heat transfer coefficient which showed good agreement with the published well-known models rahbar and esfahani (2012-2). the same researchers improved the solar still productivity at different operating conditions by using cfd simulation work (rahbar and esfahani 2013; rahbar et al. 2015). the authors and their colleagues carried out extensive experimental work to improve the performance of various types of solar still. utilizing a heat pipe and thermoelectric module to a solar still, the authors improved the productivity (rahbar and esfahani 2012-1; rahbar et al. 2016). it was due to the reduction in the start of distillation point with a thermoelectric cooler. the productivity of the solar still can be enhanced by adopting different techniques like partitioning of solar still rashidi et al. (2016) and the use of porous media rashidi et al. (2018-1) and nano fluid rashidi et al. (2018-2) in a solar still. the same researchers also performed energy and cost analysis of the modified solar stills (rashidi et al. 2017; rashidi et al. 2018-a). the cost of water from the modified stills were found to be much less than the conventional solar still. the glass inclination angle is a design parameter of the solar still that influences the productivity from a solar still kaviti et al. (2016). the increase in the still inclination produces a higher surface area and a thin water surface. hence, the still productivity increased for a higher glass inclination angle murugavel et al. (2013). in an experimental work, five different glass cover inclination solar stills tilted at different angles from 15 o to 45 o were studied. the results showed that 15 o glass cover inclination system was the best among the five models hashim et al. (2010). in another work on a single basin solar still, two glass inclination combinations, 30 o and 23 o , were tested. the 30 o inclination of the solar still was found to be more efficient dinesh et al. (2013). the effect of glass inclination angle was investigated in two previous investigations with covering glass angles of 4 o , 7 o , 10 o , 13 o , 15 o , 30 o and 45 o respectively (edioja et al. 2013; singh et al. 2013). ten solar stills with different glass inclination angles, 10–55 in steps of 5 o were experimented for still performance. the results demonstrated that the angles 20 o and 25 o provide the optimum still performance azooza and younis (2016). the inclusion of reflectors, either externally or internally, to a solar still increases the solar irradiation to basin water and it raises the still efficiency omara et al. (2017). to enhance the performance of solar stills, the use of internal reflectors were tried by the scientists (abdallah et al. 2008; tankaka, 2009; omara et al. 2016). the water temperature of solar still was increased due to addition of solar reflection into basin water. in 2016, omara et al. (2014) presented two solar still modifications using a double layer wick material and internal reflectors. they found 145.5% higher productivity for the modified stills as compared to the conventional solar still omara et al. (2014). 2 b.b. sahoo and c. subudhi increasing the basin water surface area, water temperature increases and hence, increases in the evaporation rate from the top layer of the water. this caused an increase in water productivity from solar still (manokar et al. 2014; samuel et al. 2016). this concept was tried by placing sponge cubes over the water surface of the solar still and it caused an increase of yield by 18% bassam et al. (2003). few researchers have conducted experiments on this concept on solar still by using jute cloths or wick at different positions, horizontal, vertical and on the water surface (shukla and sorayan 2005; bouchekima et al. 1998; sakthivel et al. 2010; mona et al. 2002). with all positions of jute cloth, the modified solar still yield and efficiency improved due to an increase in surface area and rate of evaporation. in some other experimental works, the absorption materials like black rubber mat, charcoals, black dye, black gravel and black ink were tried to enhance the still efficiency (mona et al. 2002; nafey et al. 2001; bilal et al. 1998). the still yield increased from 35% to 60% with charcoal and wick combined modifications. from the above literature survey, it is found that “solar still” is one of the very important equipment for producing drinking water. also, several researchers demonstrated the enhancement of simple solar still productivity and efficiency with appropriate modifications in the design of glass cover and basin. with this in mind, in the present work, various combinations of solar still performance enhancement techniques are presented. then, these combinations are investigated experimentally to find out water productivity rate and efficiency from the modified solar still. 2. work objective previous studies have attempted various modifications to improve the productivity of solar stills. the objective of the present work is to conduct experimental investigations on three combined modifications on a conventional single slope solar still. the modifications done here to the original solar still are: (i) changes in glass inclination angle; (ii) fixing reflectors to both sides of still; and (iii) placing jute cloth horizontally on the basin water. in this communication, three modified still combinations were examined experimentally and discussed with respect to the system performance. 3. theoretical aspect 3.1 solar input measurement in this work, energy input to the solar still was from sun only. transmissivity (λ) of the glass was determined by measuring the ratio of solar intensity over and under the glass. the average value of λ was calculated as 0.847 and it was approximated to 0.85 for calculation purpose. the area of glass plate was assumed equal to the area of the solar still in the calculation of solar input to the still. solar input (qin) to the still was evaluated by multiplying the solar radiation by the transmissivity of the glass cover is presented in equation (1) tiwari et al. (2009). qin = τ ag qs, w (1) where λ = transmissivity of the glass (= 0.85), ag = area of the glass plate (1.10 m× 1.10 m = 1.21 m 2 ), and qs = solar radiation [w/m 2 ]. solar radiation (qs) was measured hourly by a ‘lux meter’ during the solar still testing period of the day. its average value of the total testing hours was considered for the solar input calculation. the lux unit reading of the instrument was converted into the unit of w/m 2 by multiplying by a conversion factor of 0.0161028. the specifications of testo make lux meter are presented in table 1. table 1. specifications of lux meter. sl. no. name of the instrument specification of the instrument 1 lux meter make: testo – 540 measuring range: 0 to 99999 lux accuracy: ±3 lux or ±3 % resolution: 1 lux (0 to 19999 lux) 10 lux (remaining range) measuring rate: 0.5 second 3.2 still efficiency calculation it is the ratio of the heat energy utilized for water evaporation to the total solar radiation energy received by the still tiwari et al. (2009). it was assumed that a small amount of sensible energy consumed by distillate water is compared to its latent heat and hence, it was neglected in the still efficiency calculation. due to the glass small angle, it was assumed that the amount of solar radiation on the solar still to slope of glass is unaffected.    sqgafghvmst   (2) where mv = mass of distillate output (kg/s), hfg = latent heat of water vapor (=2257x10 3 j/kg. 3.3 measurement of distillate output the hourly collected still distillate output was collected and measured in a scaled measuring glass jar on volume basis. this volumetric distillate output was then converted into mass basis and is taken as mv in kg/s for calculation purpose. 3 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 4 3.4 measurement of temperature the temperature was recorded during the tests with the help of alcohol in-glass thermometer of 0 – 100 o c range at three different places, water, outside glass and ambient temperature. 4. experimental investigation the tested experimental setup is installed at the mechanical department, synergy institute of engineering and technology, dhenkanal, india. the schematic diagram of the single slope solar still setup is shown in fig. 1. the black painted still basin is made from 3 mm thick fibre-reinforced plastic with a maximum of 25 litres water capacity. the top of the still is covered by 6 mm thick and 1.10 m× 1.10 m area glass and placed at 8 o inclination to the horizontal. the leakage of heat and water vapour from the solar still are prevented by filling putty between the glass cover and side edges of still space. 4.1 test matrix one set of baseline tests was carried out on the existing conventional solar still. the still was modified with the following alteration by: (i) increasing glass inclination angle to 10o for better water flow conditions; (ii) fixing 20 cm height reflectors at 60o angle to both sides of still top layer to increase solar radiation on water maambo and simate (2016); (iii) placing jute cloth horizontally on the basin water to increase evaporation surface area. these three sets of experimentations were conducted on the modified conventional still. the complete particulars on the test matrix adopted for the experiments on solar still are presented in table 2. the schematic diagrams of the modified solar still operations setup of mcss-2 and mcss-3 the modes are presented in fig. 2 and fig. 3 respectively. figure 1. schematic diagram of conventional solar still (css). table 2. test matrix. designation of operation features of solar still quantity of water (lts) css conventional solar still (8 o inclination) 15 mcss-1 modified conventional solar still – 10 o inclination 15 mcss-2 modified conventional solar still – 10 o inclination, reflector fixed at 60 o angle to both sides of the still 15 mcss-3 modified conventional solar still – 10 o inclination, reflector fixed at 60 o angle to both sides of the still, placing jute cloth on the basin water 15 4 b.b. sahoo and c. subudhi 4.2 experimental procedure to the test matrix, all tests were carried out in the dhenkanal district of odisha state with geographical coordinates of 20.8424° n, 85.4376° e, india. the experiments were carried out on the south faced solar still from 10 am to 03 pm during february – march, 2017. the amount of water was kept as 15 liters in the still basin during the experimentation. the amount of water was fed into the still by the fabricated hose pipe arrangement. the initial temperature of water and atmosphere were recorded. the basin water received the solar radiation, got heated and evaporated. the water vaporized and condensed against the glass cover. the glass temperature and solar radiation were recorded at hourly interval basis. the water molecules slide down through the glass inclination to the collecting channel and fell on collected measuring jar. the hourly distillate from the solar still was collected and measured. three sets of experiments were conducted for each test matrix and the average observed parameters were used for performance analysis of the solar still. the same experimental procedures were followed to evaluate the modified sill performances as per the test matrix and the average recorded parameters were considered for analysis purpose. 4.3 measurement errors nst = f ( mv, ag , qs , twi ) (3) the uncertainties associated with the calculation of solar still efficiency were estimated by using sequential perturbation techniques kline and mcclintock (1953). it included contributions from uncertainties in measurement volume of collected distillate (1%), initial water temperature (0.5%), area of the glass plate (0.1%) and solar radiation (0.1%). based on the above values, the still efficiency was believed to be accurate within  2%. 5. results and discussion the various temperatures, solar radiation and distillate output were measured on hour intervals on the day of experimentation on the solar still with 15 liters water capacity. the results were analyzed and compared for both the modified and the conventional solar stills as per defined test matrix. figure 2. schematic diagram of modified solar still (mcss-2). figure 3. schematic diagram of modified solar still (mcss-3). 5 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 6 5.1 solar radiation the solar radiation is increased gradually from morning hours as the day progresses, as shown in fig. 4. for all modes of still operation, its highest value is reached at 12 noon and then decreases until 3 pm. the solar intensity values are varied from 583 x 10 3 to 1000 x 10 3 lux units during the experimentation. 5.2 ambient temperature figure 5 shows that the ambient temperatures are varied from 27 o c to 41 o c in the days of still operation. for all the modes, the highest value of ambient temperature is obtained at 1 pm of the day. 5.3 glass temperature in general, the solar radiation on the earth begins to increase from morning hours, reaches to maximum values around mid-day and then decreases in the afternoon. due to this solar intensity variation, it is observed that the glass temperatures for all modes go up as the day progresses and reaches a maximum value at 2 pm for all the operation days (fig. 6). the mcss-3 mode of operation produced highest glass temperature of 50 o c followed by mcss-2 (49 o c), mcss-1 (44 o c) and css (44 o c). the increase in solar energy absorption due to reflectors with mcss2 and mccs-3 modes increase the glass temperatures. the increase in solar radiation increases the evaporation and condensation of the basin water. so, the glass temperatures in both mcss-2 and mccs-3 modes are higher compared to css and mccs-1 modes. in mcss-3 mode, the increase in surface area due to jute cloth absorbs more heat energy and hence, it showed the highest glass temperature. the maximum glass temperatures are recorded equal with css and mcsss-1 modes. however, the difference in maximum and minimum glass temperatures of mcss1 mode is found higher than that of css mode. to achieve maximum glass temperature and best performance from the still, the glass angle should be increased further up to the place latitude angle of 20.8424°. figure 4. variation of solar radiation on the days of operation. figure 5. variation of ambient temperature on the days of operation. 6 b.b. sahoo and c. subudhi 7 5.4 distillate output the hourly distillate output variation with the day progress is shown in fig. 7. it is observed that the distillate output began at nil in the morning at 10 am, increased and till 2:00 pm of the day. after 2 pm, the water productivity of css and mcss-1 modes are reduced. but, mcss-2 and mcss-3 modes showed enough water production potential even after 2:00 pm. in general, the distilled water productivity of the still varies proportionally to solar radiation and ambient temperature. due to more solar absorption capability, more water yield is found with mcss-2 and mcss-3 modes modified solar stills. the maximum distillate output is recorded 189 ml and 181 ml per hour from mcss-2 and mcss-3 modes respectively around 3:00 pm. the ability of higher evaporation and condensation rates of the modified stills (mcss-2 and mcss-3) produced more water yield than css and mcss-1 modes of operation. this is the same reason that generates more glass temperatures. so, the daily water production from the mcss-2 (707 ml) and mcss-3 (718 ml) are found more than that of the other modes css (417 ml) and mcss-1 (664 ml) as shown in fig. 8. the water yield output of mcss-1 and mcss-2 are increased by 59.23% and 69.54% respectively compared with the css mode. the distillate water yield in the mcss-3 mode is increased by 72.18 % as compared to the css mode. this is due to the combined modifications effect of 10 o glass inclination, reflectors and jute cloth operation of mcss-3 mode which produced increased solar radiation absorption, higher evaporation and condensation water rate. 5.5 still efficiency figure 9 shows the comparison of efficiency for all four modes of operations solar still. the still efficiencies are found to be 10.6%, 12.31%, 13.6% and 15% for the modes css, mcss-1, mcss-2 and mcss-3 respectively. the mcss-3 mode operation of solar still showed 42.31% higher efficiency than that of the original css mode. the increase in the glass inclination to 10 o provided better water flow slope and also improved solar radiation fall on basin water as compared to the original 8 o inclination. the amount of solar radiation on water is increased due to reflectors. jute cloth inside the basin increased the water evaporation surface area. combining the benefits all together, the water productivity and, hence, the still efficiency is increased by a huge margin of 42.31% for mcss-3 mode as compared to the original css setup. figure 6. variation of glass temperature on the days of operation. figure 7. variation of hourly distillate water output on the days of operation. 7 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 8 figure 8. average distillate water output for all modes operation. figure 9. variation of solar still efficiency for all modes of operation. table 3. comparison of water productivity of different single slope solar stills. still specifications increase in productivity still with asphalt and sprinkler [badran, 2007] 51 % still with internal mirrors and step-wise basin [abdallah et al. 2008] 180 % stepped solar still with internal reflectors [omara et al. 2013] 75 % still with fins at bottom [velmurugan et al. 2008] 45.5 % still with water coral fleece with stepped wire mesh absorber [hansen et al. 2015] 71.2 % present work (still with increased glass angle, jute cloth and reflectors) 72.18 5.6 comparison of productivity of different single slope solar stills the comparison of water productivity of previous researchers’ works and present work is represented in table 3. 6. conclusion this study deals with the performance enhancement of a conventional solar still by using low cost modifications such as reflectors, jute cloth and improved glass angle. the collected data during the experimentation days were analyzed and the findings were summarized below.  the modified solar still with 10 o inclination angle in mcss-1 mode produces better water angular flow water properties with 16.31% more still efficiency than that of the original css mode. also, the water productivity of mcss-1 mode moves up by 59.23% as compared to css mode.  by fixing reflectors on the sides of the still (mcss-2 mode), the water productivity and still efficiency increases by 69.54% and 28.3% respectively as compared with the css 8 b.b. sahoo and c. subudhi 9 operation. water production increases by 290 ml with mcss-2 mode for the solar still hours of operation as compared to that of css mode.  the mcss-3 mode produces a maximum glass temperature of 50 o c when solar radiation absorbing capacity of the still is increased by using jute cloth to the basin water area along with reflectors and improved glass angle.  water productivity of the mcss-3 mode is 718 ml on the day of operation which is an enhancement amount of 72.18% over css mode.  the mcss-3 mode produces 41.5% higher still efficiency with all combined modifications as compared to that of 10% efficient css mode. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. acknowledgment the authors wish to thank and acknowledge their b.tech and m.tech students, vellore institute of technology, for their assistance towards experimenttation and data collection for the entire experiment days. references abdallah s, badran o, abu-khader mm (2008), performance evaluation of a modified design of a single slope solar still. desalination 219: 222–230. abu–hijlehbassam ak, rababa'h hm (2003), experimental study of solar still with sponge cubes in basin. energy conversion management 44: 1411–1418. akash a, mohsen ms, omar o, yaserelayan (1998), experimental evaluation of a single-basin solar still using different absorbing materials. renewable energy 14: 307–310. amitava b (2013), solar stills for desalination of water in rural households. international journal of environment and sustainability 2(1): 21–30. azooza a, younis gg (2016), effect of glass inclination angle on solar still performance. journal of renewable and sustainable energy 8(3): 033702-1-033702-12. badran o.o. (2007), experimental study of the enhancement parameters on a single slope solar still productivity. desalination 209: 136–143. bouchekima b, gros b, ouahes r, diboun m (1998), performance study of the capillary film solar distiller. desalination 116: 185–192. dinesh k, himanshu p, zameer a (2013), performance analysis of single slope solar still. international journal of emerging technology and advanced engineering 3(3): 66–72. edeoja o, unom f, edeoja ja (2013), investigation of the effect of angle of cover inclination on the yield of a single basin solar still under makurdi climate. international journal of engineering science 2(7): 131–138. hansen rs, narayanan cs, murugavel kk (2015), performance analysis on inclined solar still with different new wick materials and wire mesh. desalination 358: 1–8. hashim ay, al asadi mj, tahaalramdhan wa (2010), an attempt to solar still productivity optimization; solar still shape, glass cover inclination and inner surface area of a single basin solar still, optimization. basrah journal of science 28(1): 39–48. kaviti ak, yadav a, shukla a (2016), inclined solar still designs: a review. renewable and sustainable energy reviews 54: 429–451. kline sj, mcclintock f.a. (1953), describing uncertainties in single-sample experiments. mechanical engineering 75: 3–8. kumar pv, kumar a, prakash o, kaviti ak (2015), solar stills system design: a review. renewable and sustainable energy review 51: 153–181. maambo h, simate i (2016), performance improvement of solar water stills by using reflectors. journal of humanitarian engineering 4(2): 8-15. manokar a.m, murugavel k.k, esakkimuthu g(2014), different parameters affecting the rate of evaporation and condensation on passive solar still– a review. renewable and sustainable energy reviews 38: 309–322. murugavel k.k, anburaj p, samuel r.h, elango t (2013), progresses in inclined type solar stills. renewable and sustainable energy reviews 20: 364–377. naim m.m, abdei kawi m.a (2002), nonconventional solar stills part-1 nonconventional solar stills with charcoal particles as absorber medium. desalination 153: 55–64. nafey a.s, abdelkader m, abdelmotalip a, mabrouk a.a (2001), solar still productivity enhancement. energy conversion management 42: 1401–1408. omara z.m, kabeel a.e, younes m.m (2013), enhancing the stepped solar still performance using internal reflectors. desalination 314: 67–72. omara z.m, kabeel a.e, younes m.m (2014), enhancing the stepped solar still performance using internal and external reflectors. energy conversion management 78: 876–881. 9 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 1-10 10 omara z.m, kabeel a.e, abdullah a.s, essa f.a (2016), experimental investigation of corrugated absorber solar still with wick and reflectors. desalination 381: 111–116. omara z.m, kabeel a.e, abdullah a.s (2017), a review of solar still performance with reflectors. renewable and sustainable energy reviews 68: 638–649. rahbar n, esfahani j.a (2012-1), experimental study of a novel portable solar still by utilizing the heatpipe and thermoelectric module. desalination 284: 55–61. rahbar n, esfahani j.a (2012-2), estimation of convective heat transfer coefficient in a single-slope solar still: a numerical study. desalination and water treatment 50: 387–396. rahbar n, esfahani j.a (2013), productivity estimation of a single-slope solar still: theoretical and numerical analysis. energy 49: 289-297. rahbar n, esfahani j.a, fotouhi-bafghi e (2015), estimation of convective heat transfer coefficient and water-productivity in a tubular solar still – cfd simulation and theoretical analysis. solar energy 113: 313-323. rahbar n, esfahani j.a, asadi a (2017), an experimental investigation on productivity and performance of a new improved design portable asymmetrical solar still utilizing thermoelectric modules. energy conversion and management 118: 55-62. rashidi s, bovand m, esfahani j.a (2016), optimization of partitioning inside a single slope solar still for performance improvement. desalination 395: 79-91. rashidi s, esfahani j.a, rahbar n (2017), partitioning of solar still for performance recovery: experimental and numerical investigations with cost analysis. solar energy 153: 41-50. rashidi s, rahbar n, valipour m.s, esfahani j.a (2018-1), enhancement of solar still by reticular porous media: experimental investigation with exergy and economic analysis. applied thermal engineering 130: 1341-1348. rashidi s, bovand m, rahbar n, javadabolfazliesfahani j.a (2018-2), steps optimization and productivity enhancement in a nanofluid cascade solar still. renewable energy 118: 536-545. sakthivel m, shanmugasundaram s, alwarsamy t (2010), an experimental study on a regenerative solar still with energy storage medium —jute cloth. desalination 264: 24–31. sahoo b.b, sahoo n, mahanta p, borboral, kalita p, saha u.k (2008), performance assessment of a solar still using blackened surface and thermocol insulation. renewable energy 33: 1703–1708. samuel d.g.h, nagarajan p.k, arunkumar t, kannan e, sathyamurthy r (2016), enhancing the solar still yield by increasing the surface area of water—a review. environmental progress and sustainable energy 35(3): 815–822. shukla s.k, sorayan v.p.s (2005), thermal modeling of solar stills: an experimental validation. renewable energy 30: 683–699. singh a, gaikwadrai a.k (2013), effect of inclination angle and water temperature on internal heat transfer coefficients and productivity of single slope solar still: an indoor simulation. international journal of scientific research 2(7): 142–146. tanaka h (2009), experimental study of a basin type solar still with internal and external reflectors in winter. desalination 249: 130–134. tiwari g.n, dimri v, chel a (2009), parametric study of an active and passive solar distillation system: energy and exergy analysis. desalination 242: 1–18. velmurugan v, gopalakrishnan m, raghu r, srithar k (2008), single basin solar still with fin for enhancing productivity. energy conversion and management 49: 2602–2608. 10 template for electronic submission to acs journals the journal of engineering research (tjer), vol. 16, no. 1 (2019) 35-43 doi: 10.24200/tjer.vol16iss1pp35-43 flexural behavior of rc beams strengthned with cfrp sheets using different strengthening techniques s.e. el gamal a,* , a.al nuaimi b , a. al-saidy b , and k. al-shanfari b a civil and architectural engineering department, sultan qaboos university, al-khoudh, muscat, oman, on leave from menoufia university, menoufia, egypt. b civil and architectural engineering department, sultan qaboos university, al-khoudh, muscat, oman. abstract: due to the advantages of fiber reinforced polymer (frp) materials, they have been utilized to strengthen several reinforced concrete (rc) elements such as slabs, beams and columns. in this paper, five rc beams (200 mm width, 300 mm height, and 2750 mm length) were constructed. four of these beams were strengthened with cfrp sheets whereas the last beam was used as a reference. test parameters include the amount of frp and the strengthening technique. three strengthening techniques were used including the externally bonded technique (eb), the near surface mounted (nsm) technique using folded cfrp sheets inserted in near surface grooves, and a hybrid technique. all beams were tested under four point bending setup until failure. the control beam failed by the yielding of the tension steel followed by concrete crushing. the strengthened beams failed by steel yielding followed by either rupture or debonding of cfrp sheets at higher loads compared to the reference one. the stiffness after steel yielding and the ultimate capacity increased as the amount of cfrp increased. the strengthening technique affected the ultimate capacity of the strengthened beams. the nsm beam showed the lowest increase in the ultimate capacity (25.2%) whereas the hybrid beam showed the best performance with the highest increase in the ultimate capacity (58%) compared to the reference beam. keywords: carbon fiber reinforced polymers; experimental work; flexural behavior; reinforced concrete beams; strengthening techniques. باستخدام تقنيات تقوية البوليمرية السلوك املرن للعوارض اخلرسانية املسلحة املقواة بألياف الكربون خمتلفة أ، اجلمل .ش * ب الشنفري .، خب السعيدي. ، ع ب النعيمي .، ع ة املسلحة مثل البالطات يعناصر اخلرسانالفقد مت استخدامها لتقوية العديد من : نظرا ملزايا مواد البوليمر املقوى باألليافامللخص من تلك مت تقوية أربع ومم. 0572× 022× 022 والعوارض واألعمدة. يف هذا البحث، مت بناء مخسة عوارض خرسانية مسلحة بأبعاد سةعوامل الدر تتضمنواخلامسة بدون تقوية كمرجع. يف حني مت استخدام العارضة يةالبوليمر ألياف الكربونعوارض باستخدام ال مت استخدام ثالث تقنيات تقوية مبا يف ذلك تقنية وقد وكذلك التقنية املستخدمة يف التقوية. البوليمرية كل من كمية ألياف الكربون باستخدام نسيج ألياف (nsm) تقنية الرتكيب القريب من السطح و بني األلياف و السطح اخلارجي للعوارض (eb) الرتابط اخلارجي . مت اختبار مجيع العوارض من التقنيتني السابقتني املطوية اليت مت إدخاهلا يف أخاديد على سطح العوارض ، وتقنية هجينة الكربون خضوع حديد التسليح متبوًعا بتهشم اخلرسانة. قواة عن طري املانهارت العارضة املرجعية غري ف باستخدام محلني مركزين حتى اإلنهيار. بينما انهارت العوارض املقواة على أمحال أعلي من العارضة املرجعية عن طري خضوع حديد التسليح متبوعا إما بتمزق ألياف الكربون القصوى للعوارض املقواة زيادة كل من جساءة و قوة التحملكذلك لوحظ وقد العوارض اخلرسانية. بأو تفكك التصاقها البوليمرية أثرت على قوة التحمل القصوى تت الدراسة أيضا أن تقنية التقوية املستخدمة بأث . كماالبوليمرية ألياف الكربون عدد طبقاتبزيادة أعطت أقل زيادة يف قوة التحمل القصوى للعوارض املقواة (nsm)تقنية الرتكيب القريب من السطح هذا وقد أعطت للعوارض املقواة. أداء من حيث قوة التحمل القصوى للعوارض أفضل % مقارنة بالعارضة املرجعية. بينما التقنية اهلجينة أعطت0750بنسبة زيادة بلغت .مقارنة بالعارضة املرجعية %75املقواة بنسبة زيادة بلغت تقنيات التقوية5 ؛العوارض الخرسانية المسلحة ؛السلوك المرن ؛دراسة معملية ؛ية البوليمر الكربون ليافأ مفتاحية:الكلمات ال *corresponding author’s e-mail: sherif@squ.edu.om the journal of engineering research (tjer), vol. 16, no. 1 (2019) 36-44 36 1. introduction due to harsh environmental conditions such as high humidity and temperatures, several reinforced concrete (rc) structures in the arabian gulf region, especially those close to shorelines, suffer from steel corrosion problems. this requires strengthening and/or rehabilitation of these structures almusallam et al. (2013). in other cases, strengthening of rc structures is required to overcome design or construction mistakes, to upgrade the capacity of an existing structure, or to fulfill revisions in codes and standards al-salloum et al. (2013). several strengthening/rehabilitation techniques of rc structures have been used in the last decades. due to their high strength, less weight, corrosion resistance, and easy in application, carbon fiber reinforced polymer (cfrp) sheets have been widely used for strengthening of several rc elements including rc columns, beams, and slabs el-gamal et al. (2012). several studies have investigated the behavior of rc beams strengthened in flexure with frp composites. many research studies have used the near surface mounted (nsm) technique where frp bars are inserted and bonded to near surface longitudinal grooves in the tension surface of the beam using bonding materials (al mahmoud et al. 2009; soliman et al. 2011; sharaky et al. 2014; el-gamal et al. 2014). other studies used the externally bonded (eb) technique where frp sheets are bonded to the tension surface of the beams (al-tamimi 2011; mostofinejad and shameli 2011, 2013; jiangfeng et al. 2012; attari et al. 2012; dong et al. 2013; hawileh et al. 2014; al-saidy et al. 2015; el-gamal et al. 2017). furthermore, other studies used both techniques (choi et al. 2007; jung et al. 2007; seracino et al. 2007; rasheed et al. 2010; ceroni 2010; el-gamal et al. 2016). among those studies was the investigation conducted by choi et al. (2011) who investigated the behavior of seven t-shaped rc beams strengthened with both eb-cfrp plates and nsm-cfrp bars. their test parameters included bonded length (fully and partially bonded length) and strengthening technique (eb and nsm). in the eb specimens with partially bonded length, they used transverse anchorage frp sheets to increase the bond strength. test results of the eb beams showed that the fully bonded specimens failed by frp debonding. whereas the partially bonded beams with transverse anchorage failed at higher load by frp rupture. the nsm specimens, however, utilized the full strength of the cfrp bars and the failure modes varied between frp bar slipping after sustaining a very large deflection in fully bonded beam and concrete crushing at mid-span for the partially bonded beams. they concluded that the effectiveness of the nsm system was better than the eb system. the eb and nsm strengthening techniques increased the ultimate loading capacity by 6% to 23% and 36% to 53%, respectively, compared to the reference beam. jung et al. (2007) investigated the flexural behavior of eight rc beams strengthened with cfrp strips and rods using eb, nsm and nsm with mechanical interlocking (mi) grooves, and nsm with prestressing strengthening techniques under four-point bending setup. test results showed that the maximum measured strains in the frp reinforcement for the eb specimens was about 30% of the ultimate strain, 82-87% for the nsm specimens, and approximately 100% for the nsm specimens with the mi grooves and the pre-stressed specimens, which proved that nsm system had utilized frp reinforcement sufficiently compared to eb specimens. the pre-stressed nsm specimens showed a behavior similar to the nsm specimens with the mi grooves. mostofinejad and shameli (2011, 2013) conducted two experimental studies to investigate the performance of new flexural strengthening techniques using cfrp sheets. they called the first technique “externally bonded reinforcement on grooves” (ebrog) where the frp sheet is laid over longitudinal grooves filled with epoxy resin with the same concept of eb technique. the second technique is called externally bonded reinforcement in groove (ebrig). this technique is similar to the nsm technique; however, the frp sheets are bonded to the internal surfaces of the grooves as well as to the tension face of the beams. in both studies, they constructed small-scale concrete beam specimens with no internal flexural reinforcement. they used only internal steel stirrups to prevent shear failure. cfrp sheets with different number of layers (one to three layers) were used for strengthening the beams using four strengthening techniques i.e. eb, nsm, ebrog and ebrig. they concluded that the beams strengthened with ebrog and ebrig techniques had considerable higher capacities compared to the beams strengthened with eb technique. in addition, the ebrig technique permitted for higher failure loads and displacements compared to the ebrog technique. dong et al. (2013) conducted an experimental study to investigate the flexural behavior of seven rc beams strengthened with one or two layers of cfrp sheets. the aim of the study was to investigate the effectiveness of strengthening on cracking load, ultimate load, strains and deflections. all the beams were simply supported over a clear span of 1500 mm and were tested under four-point bending. the test result showed that the strengthened beams failed due to either rupture or debonding of frp sheets. the overall flexural capacity of all strengthened beams varied between 41% and 125% over the control beams. they also concluded that frp sheets controlled cracks and increased the ductility of the beams. rasheed et al. (2010) constructed six beams and strengthened them using four strengthening techniques and materials: eb cfrp sheets, nsm cfrp strips, eb steel reinforced polymer (srp) sheets, and nsm stainless steel bars. they used 36 s.e. el gamal, k. al-shanfari, a. al nuaimi and a. al-saidy different configurations of external transverse reinforcement to improve ductility and control debonding of the strengthening systems. test results showed that all strengthened beams showed much higher capacities compared to the reference beam. they concluded that the use of external transverse anchoring reinforcement further increased the flexural capacity of the strengthened members and allowed for better utilization of the high strength properties of the strengthening materials. ceroni (2010) constructed 21 rc beams to investigate the behavior of rc beams strengthened with eb cfrp sheets or nsm cfrp bars. all beams were tested under four point bending setup. test results indicated that most of the eb beams without anchoring devices failed by debonding. an adequate ductility was attained by introducing anchorage devices that eliminated or delayed the debonding. they also concluded that the best result was reached by applying distributed u-shaped frp strips. the increased strength in the eb beams varied between 18% and 51%. the nsm beams achieved higher load capacity (46–72%) greater than the eb beams having similar equivalent reinforcement percentage. hawileh et al. (2014) conducted an experimental study on five rc beams strengthened with eb frp sheets to investigate their flexural performance. the beams were strengthened with combinations of eb glass and carbon frp sheets and were tested under four point bending setup. in addition, they developed an analytical model to predict the load–deflection response and the carrying capacities of the beams and they compared the model with their experimental results and with the aci 440.2r-08 predictions. they recorded an increase in the ultimate capacity of the strengthened beams ranging between 30.7 and 98% compared to the control beam. they observed that the beams strengthened with hybrid glass and carbon frp sheets developed higher ductility than those strengthened with a single carbon sheet, however, the beams strengthened with a single glass frp sheet developed the highest ductility among all strengthened beams. they concluded that using a hybrid system of glass and carbon frp sheets was the best as it combined both the high strength of the carbon sheets that improved strength and the low stiffness of glass sheets that improved ductility. they also concluded that the aci provisions were accurate for the beams with one layer of strengthening sheet; however, they were less accurate for hybrid specimens as the number of strengthening layers increased. el-gamal et al. (2014, 2016) conducted experimental studies to investigate the flexural behavior of rc beams strengthened with glass and carbon frp composites using nsm and eb techniques. the experimental investigation included four parameters; strengthening technique, type of frp, amount of frp, and steel reinforcement ratio. test results indicated that all strengthened beams showed an increase in the ultimate load capacity ranging between 55 and 133% compared with the reference beam. the nsm-cfrp strengthened beams showed higher capacities than the nsm-gfrp strengthened ones; however, they showed much more brittle behavior. they recommended conducting further experimental studies to investigate other parameters. it can be noticed that most of the above mentioned research studies used the nsm technique where frp bars are inserted into longitudinal grooves on the tension surface of the beams or the externally bonded (eb) technique where the cfrp sheets are attached to the tension surface of the beams. however, to the best knowledge of the authors, very limited research studies investigated the use of cfrp sheets inserted in nsm grooves mostofinejad and shameli (2013) or used a hybrid technique (nsm and eb). this research study aims to fill this gap by investigating the behavior of rc beams strengthened with cfrp sheets using different techniques. this includes the regular eb technique, a modified nsm technique using cfrp sheets that were folded and embedded into near surface grooves, and a hybrid technique (eb and nsm). the experimental work includes the construction and testing of five rc beams. one beam was a reference beam without strengthening. another beam was strengthened with two cfrp sheets folded and inserted into two grooves, which is similar to the nsm technique. two beams were strengthened with one/two layers of eb cfrp sheets. the last beam was strengthened with a hybrid technique where one cfrp sheet was inserted in a near surface groove while a second cfrp sheet was bonded to the surface of the beams using the eb technique. the measurements included mid-span deflection, cracking, ultimate capacity, and mode of failure. 2. test specimens and set-up 2.1 materials the five beams were cast using a ready mix concrete. standard concrete cylinders of 150 mm diameter and 300 mm height were taken from the concrete mix during casting and were cured with the beams. standard compressive and splitting tests were conducted on the cylinders after curing (fig. 1). the measured concrete compressive strengths was 56.3 mpa, whereas the tensile strength was 3.2 mpa. deformed steel bars of 12 mm diameter were used for the bottom longitudinal reinforcement of the (a) (b) figure 1. (a) brazilian test; (b) compressive strength test. 37 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 35-43 beams, whereas 8 mm diameter bars were used for the top reinforcement and stirrups. the steel bars were tested in the laboratory to determine their tensile properties. test results showed that the average yield strength of the steel bars was 480 mpa for both diameters. unidirectional cfrp sheets were used to strengthen the beams. the ultimate tensile capacity of the cfrp sheets was 350 kn/m-width as given by the manufacturer. fig. 2 shows a photo of the cfrp sheets used in this study. epoxy resin was used to bond the cfrp sheets with concrete. the technical information of both cfrp sheets and epoxy resin can be found in (sika group, 2018; basf, 2018), respectively. 2.2 description of test specimens five rc beams with typical reinforcement and dimensions were constructed. all the beams had dimensions of 2760×300×200 mm (length × depth × width). all beams were reinforced with two steel bars of 8 and 12 mm diameters in the top and bottom longitudinal direction, respectively. steel bars of eight mm diameter spaced at 100 mm center-to-center were used in all beams as shear reinforcement to prevent any shear failure. fig. 3 shows the reinforcement details and dimensions of the beams. figure 3. elevation and cross section of the beams (all dimensions are in mm). out of the five constructed beams, the first beam was used as a control specimen without strengthening (fig. 3). all the other four beams were strengthened (in the bottom face) with cfrp sheets (200 mm width) using different procedures. two beams (eb1 and eb2) were strengthened using the well-known eb technique. beam eb1 was strengthened with one cfrp sheet, whereas beam eb2 was strengthened with two cfrp sheets as shown in fig. 4a,b. the third beam (nsm) was strengthened with two cfrp sheets folded and inserted into two near surface grooves (10 mm width and 20 mm depth) in the tension side of the beam. in this technique, which is similar to the nsm technique, the cfrp sheets were unconventionally embedded into the near surface grooves, which were filled with epoxy as shown in fig. 4c. the fourth beam (hyb) was strengthened by a hybrid technique using one cfrp sheet inserted in a groove and one cfrp eb sheet. table 1 lists the matrix of beams, whereas fig. 4 shows the cross sections of all strengthened beams. table 1. schedule of test specimens. (a) (b) (c) (d) figure 4. cross sections of strengthened beams: (a) eb1; (b) eb2; (c) hyb; (d) nsm (all dimensions are in mm). beam strengthening reinforcement strengthening technique control eb1 one eb cfrp sheet externally bonded eb2 two eb cfrp sheet externally bonded nsm two cfrp sheets imbedded into two near surface grooves nsm hyb 1 cfrp sheet imbedded in near surface groove and one eb cfrp sheet hybrid (nsm and eb) figure 2. unidirectional cfrp sheets used in this study. load t8@100 mm 2t8 mm 2t12 mm 2760 mm 2360 mm 930 mm 500 930 mm 212 28 one eb cfrp sheet 212 28 two eb cfrp sheets 212 28 one nsm cfrp sheet one eb cfrp sheet 212 28 85 two folded nsm cfrp sheets 15 42. 5 20 260 20 25 300 200 2t12 2t8 38 s.e. el gamal, k. al-shanfari, a. al nuaimi and a. al-saidy 2.3 construction of test specimens after preparing the steel cages of the five beams, the cages were placed into wooden molds before concrete casting using a ready mix concrete. fifteen standard concrete cylinders were cast with the beams for measuring compressive and tensile strength of concrete. all beams were cured for 28 days before testing. fig. 5 shows the specimens before and after casting. 2.4 strengthening of test specimens for the beams with external layers of cfrp sheets (eb beams), the concrete surface was prepared before applying the fibers. the bottom surfaces of the beams were roughened and cleaned well before adding the epoxy on the surface and installing the cfrp sheets. for the eb2 beam, a second layer of epoxy was added before installing the second cfrp sheet. fig. 6a shows one of the eb specimens after installing the sheets. for the nsm beam, each groove was cleaned and partially filled with the epoxy resin. afterwards, the dry cfrp sheet was folded and inserted into the groove then covered with more epoxy resin. the same procedure was repeated for the second groove. fig. 6b shows the nsm beam after installing the cfrp sheets. for the hyb beam, the first cfrp sheet was inserted in the groove as described in the nsm beam above. after that, a second cfrp sheet was installed on the surface as described in the eb beams. 2.5 test set-up all beams were tested under four-point bending setup with a clear simply supported span of 2360 mm as shown in fig. 3. three linear variable differential transformer (lvdts) were used at mid and quarter spans of the beams for deflection measurements. two concrete strain gauges were installed at the top midspan of the beams for concrete compressive strains measurements. load was applied gradually at a rate of 1 mm/min until failure. all data were automatically recorded using a data acquisition system connected to a computer. fig. 7 shows one beam during testing. 3. test results and discussions 3.1 description of test results table 2 shows the main results of test specimens. the ultimate capacity of the control beam was 76.4 kn and the maximum mid-span deflection was 53.4 mm. the first crack was recorded at a load of about 24 kn. figure 8 shows the load versus mid-span deflection curve of the control beam. it can be noticed that the curve shows three different stages. the first part represents the behavior of the beam with its gross inertia before cracking. the second part shows the behavior after cracking until steel yielding. the third part presents the behavior of the beam after steel yielding until failure. as expected, the control beam failed by yielding of tension steel followed by crushing of concrete after large deflections. all strengthened beams had almost the same cracking load as the control beam (24 kn to 25 kn), which depends mainly on the gross inertia of the cross section of the beam. however, table 2 shows a big difference between the control beam and the strengthened beams in terms of yield load, maximum load, mid-span deflection, number of cracks at failure, and failure mode. the strengthened beams revealed higher capacities (95.7 to 120.9 kn) compared to the control beam (76.4 kn) as shown in (b) eb beams (a) nsm beams figure 5. test specimens before and after casting. figure 6. installation of cfrp sheets. figure 7. photo of one beam during testing. 39 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 35-43 fig. 9. the increase in the ultimate capacity of the strengthened beams ranged between 25.3% in the nsm beam and 58.3% in the hyb beam. the control beam, however, showed higher deflections at maximum load. three different failure modes were recorded. the control beam failed in a ductile manner by tension steel yielding followed by concrete crushing after large deflections. beam eb1 failed by yielding of tension steel followed by rupture of cfrp sheet, whereas the other three strengthened beams (eb2, nsm, hyb) failed by yielding of steel followed by cfrp debonding. figure 10 shows the beams after failure. it can be seen that the strengthened beams had more cracks at table 2. main test results. beam pcr kn py kn pmax kn max mm capacity increase (%) failure mode control 24 64 76.4 53.4 sycc eb1 24 72 101.2 22.8 32.6 syrup eb2 25 100 119.2 17.2 56.1 sydeb hyb 24 79 120.9 26.0 58.3 nsm 24 90 95.7 21.1 25.3 where pcr is the cracking load, py is the yield load, pmax is the maximum load, max = mid-span deflection at maximum load, sy = steel yielding; cc = concrete crushing; rup= rupture of cfrp; deb = debonding of cfrp. failure compared to the control beam. the number of recorded cracks ranged between 18 and 24 cracks in the strengthened beams compared to only 14 cracks in the control beam. 3.2 effect of the amount of cfrp reinforcement figure 11 shows mid-span deflection versus load in beams eb1, eb2, and the control beam. the figure shows that the capacity of the strengthened beams increased as the amount of cfrp increased. using one cfrp sheet (in eb1) increased the capacity by about 32.6%, whereas using two cfrp sheets (in eb2) increased the capacity by about 56.1% compared to the control beam. both beams failed in a brittle manner; however, the mode of failure was not similar. beam eb1 failed by cfrp rupture, which means that the strains in the cfrp sheet reached the ultimate strain of the cfrp fibers. beam eb2, however, failed by cfrp debonding before reaching the ultimate strength of the sheets, which was in agreement with several research studies. this means that the capacity of this beam could be further increased if the debonding problem was delayed or eliminated using transverse anchorage as described by different researchers (choi et al. 2007; rasheed et al. 2010; ceroni, 2010) or by using ebrog system as given by mostofinejad and shameli (2011, 2013). therefore, it is recommended to conduct further investigations on the best anchorage technique in a future study. the strengthened beams showed a slightly higher stiffness after cracking compared to the control beam, which is expected as the stiffness at this stage depends on both steel and frp material. after steel yielding, fig. 11 shows that the stiffness significantly increased as the amount of cfrp increased as the stiffness at this stage is mainly depending on the frp material. it can be concluded from this parameter that the amount of frp reinforcement did not show a significant effect on the behavior of the beam at the elastic stage. however, it affected the mode of failure, the ultimate capacity and the stiffness after steel yielding. both ultimate capacity and stiffness of the beams after steel yielding increased as the amount of the cfrp increased. 76.6 101.23 119.23 120.87 95.5 0 25 50 75 100 125 150 control beam eb1 eb2 hybrid nsm u lt im a te l o a d ( k n ) figure 8. load–deflection curve of the control beam.. figure 9. comparison between the ultimate loads of tested beams. control figure 10. tested beams after failure. 40 s.e. el gamal, k. al-shanfari, a. al nuaimi and a. al-saidy figure 11. effect of the amount of cfrp reinforcement. figure 12. effect of strengthening technique. 3.3 effect of the strengthening technique load versus mid-span deflection curves of beams eb2, nsm and hyp are given fig. 12. in the three strengthened beams, the same amount of cfrp (two sheets) were used. however, the strengthening technique was different. the three strengthened beams had a similar failure mode (yielding of tension steel followed by cfrp debonding). debonding started at the end of the cfrp sheets in beams eb2 and hyb, however, it started from the middle part of the sheet in the nsm beam. the ultimate capacities of the three strengthened beams were higher than that of the control beam, which demonstrates the advantage of using frp as an effective strengthening material. although the same amount of cfrp was used in the three beams, the capacities were not similar. compared to the control beam, the nsm beam gave the lowest increase in the ultimate capacity (25.1%) among the three beams. in the eb2 and hyb beams, the increase in the capacity was about 56.1 and 58.3% compared to the control beam, respectively. this demonstrated that the nsm technique using cfep sheets imbedded in near surface grooves was not effective and had a limited effect on the ultimate load enhancement. this could be related to the strengthening procedure used in this technique, which might be prevented the epoxy resin from saturating all the cfrp fibers and resulted in an early debonding of the cfrp in one groove at low load level followed by a debonding in the second groove. the nsm beam, however, showed the best ductile behavior among all strengthened beams, which was the only advantage of this technique in this research study. it is recommended, when using this technique in the future, to fully saturate the cfrp sheets in resin before inserting them into grooves. this will increase the bond and consequently result in higher capacities. the ultimate capacity of the hyb beam was almost similar to that recorded in the eb2 beam. however, it can be seen from fig. 12 that the hyb beam showed more ductile behavior compared to the eb2 beam. the recorded mid-span deflection at maximum load was about 26 mm in the hyb beam compared with an only 17.2 mm in the eb2 beam. in general, it can be concluded that strengthening technique had a significant effect on the behavior of the strengthened beams and that additional studies are required to find the best way to delay the debonding of the frp sheets and consequently increase the capacity and the ductility of the strengthened beams. 4. conclusion based on this research study, the following concluding remarks can be drawn: 1all strengthened beams showed higher ultimate capacities compared to the control beam. this increase ranged between 25.3 and 58.3%. 2the amount of cfrp reinforcement did not show a significant effect on the behavior of the beams at the elastic stage. however, it affected the mode of failure, the ultimate capacity and the stiffness after steel yielding. the stiffness after steel yielding and the ultimate capacity increased as the amount of cfrp increased. 3the strengthening technique affected the ultimate capacity of the strengthened beams. the hybrid beam showed the highest improvement in the ultimate capacity (58.3%). it also showed better ductile behavior than eb2 beam. this indicates that the hybrid beam gave the best performance in all strengthened beams tested in this study. 4the nsm beam strengthened with cfrp sheets inserted in near surface grooves gave the lowest ultimate capacity among all strengthened beams. this could be related to the procedure used in this study where the dry fibers were folded and inserted into the grooves. it is recommended, when using this technique in the future, to fully saturate the cfrp sheets in resin before inserting them into the grooves. this will increase the bond and the ultimate capacities of the strengthened beams. 5 the debonding of the cfrp sheets was the main reason of failure in the strengthened beams with two cfrp sheets. more studies are still needed to find the best way to eliminate the debonding problem and increase the capacity and the ductility of the beams. 41 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 35-43 confilct of interest the author declares no conflicts of interest. funding this research study was funded by sultan qaboos university, sultanate of oman, through the internal research grant no. ig-eng-caed-14-01. acknowledgment the authors would like to acknowledge the civil and architectural engineering department and sultan qaboos university for their support. the authors would also like to thank the engineers and technicians at the concrete and structural laboratory for their help. references al mahmoud f, castel a, françois r, tourneur c (2009), strengthening of rc members with nearsurface mounted cfrp rods. compos. struct. 91(2): 138-147. almusallam t.h, al-salloum y.a, alsayed sh, elgamal se, aqel m (2013), tensile properties of glass fiber-reinforced polymer bars embedded in concrete under severe laboratory and field environmental conditions. journal of composite materials, 47(4): 393-407. al-saidy a.h, saadatmanesh h, el-gamal s.e, aljabri k.s, waris b.m (2015), structural behavior of corroded rc beams with/without stirrups repaired with cfrp sheets. materials and structures, 49(9): 3733-3747. al-salloum y, el-gamal se, almusallam t, alsayed s, aqel m (2013), effect of harsh environmental conditions on the tensile properties of gfrp bars. composites: part b, 45(1): 835-844. al-tamimi1 a, hawileh r, abdalla j, rasheed h (2011), effects of ratio of cfrp plate length to shear span and end anchorage on flexural behavior of scc rc beams. j. compos. constr, 15(6): 908-919. attari n.a, amziane b.m, chemrouk c (2012), an experimental study on the flexural response of partially bonded frp strengthened concrete beams. construction and building materials, 37: 746-757. basf mbrace saturant (2018) available at: http://www.basf-cc.com.sg/en/products/ concrete repairand protection systems/ mbrace/ mbrace saturant/pages/default.aspx, (accessed: 18 feburary 2018). ceroni f (2010), experimental performances of rc beams strengthened with frp materials. construction and building materials, 24(9): 1547-1155. choi h.t, west j.s, soudki k.a (2011), partially bonded near-surface mounted cfrp bars for strengthened concrete t-beams. construction and building materials. 25(5): 2441-2449. dong j, wang q, guan z (2013), structural behavior of rc beams with external flexural and flexural– shear strengthening by frp sheets. composites part b: engineering, 44(1): 604-612. el-gamal s.e, al-salloum y, alsayed s, aqel m (2012), performance of near surface mounted gfrp bars in concrete. journal of reinforced plastics and composites. 31(22): 1501-1515. el-gamal s.e, al-nuaimi a, al-saidy a, al-lawati a (2014), flexural strengthening of rc beams using near surface mounted fiber reinforced polymers. in proc. of brunei international conference on engineering and technology. institut teknologi brunei, brunei darussalam, 13 november 2014. doi:10.1049/cp.2014.1082. el-gamal se, al-nuaimi a, al-saidy a, al-lawati a (2016), efficiency of near surface mounted technique using fiber reinforced polymers for the flexural strengthening of rc beams. construction and building materials. 118: 5262. el-gamal s.e, al-nuaimi a, al-saidy a, al-shanfari a (2017), flexural behavior of reinforced concrete beams strengthened with carbon fiber reinforced polymer sheets using different bonding techniques. in proc. of the 8th international conference on structural engineering and construction management (icsecm), kandi, srilanka. hawileh r, rasheed h, abdalla j, al-tamimi a (2014), behavior of reinforced concrete beams strengthened with externally bonded hybrid fiber reinforced polymer systems. materials and design. 53: 972-982. jiangfeng d, qingyuan w, zhongwei g (2012), structural behavior of rc beams with external flexural and flexural–shear strengthening by frp sheets. composites part b. 44: 604-612. jung w, park j, park y (2007), a study on the flexural behavior of reinforced concrete beams strengthened with nsm prestressed cfrp reinforcements, in proc. of the 8th international symposium on fibre-reinforced polymer reinforcement for concrete structures (frprcs8); patras, greece. mostofinejad d and shameli s (2011), performance of ebrog method under multilayer frp sheets for flexural strengthening of concrete beams. procedia engineering. 14: 3176-3182. mostofinejad d and shameli s.m (2013), externally bonded reinforcement in grooves (ebrig) technique to postpone debonding of frp sheets in strengthened concrete beams. construction and building materials. 38: 751-758. rasheed h, harrison r, peterman r, alkhrdaji t (2010), ductile strengthening using externally bonded and near surface mounted, composite structure. 92: 2379-2393. seracino r, saifulnaz m.r, oehlers d.j (2007), generic debonding resistance of eb and nsm plate-to-concrete joints. j compos constr, 11(1): 62-70. sharaky i.a, torres l, comas j, barris c (2014), flexural response of reinforced concrete (rc) 42 http://www.sciencedirect.com/science/journal/13598368/45/1 http://www.sciencedirect.com/science/journal/13598368/45/1 http://www.sciencedirect.com/science/journal/13598368/45/1 http://www.sciencedirect.com/science/journal/13598368/45/1 file:///c:/users/raoh/desktop/tjer_submitted%20paper%20(r)/civl180220/basf%20-%20mbrace%20saturant%20(2018)%20available%20at:%20http:/www.basf-cc.com.sg/en/products/%20concrete%20repairandprotectionsystems/mbrace/mbracesaturant/pages/default.aspx,%20(accessed:%2018%20feburary%202018) file:///c:/users/raoh/desktop/tjer_submitted%20paper%20(r)/civl180220/basf%20-%20mbrace%20saturant%20(2018)%20available%20at:%20http:/www.basf-cc.com.sg/en/products/%20concrete%20repairandprotectionsystems/mbrace/mbracesaturant/pages/default.aspx,%20(accessed:%2018%20feburary%202018) file:///c:/users/raoh/desktop/tjer_submitted%20paper%20(r)/civl180220/basf%20-%20mbrace%20saturant%20(2018)%20available%20at:%20http:/www.basf-cc.com.sg/en/products/%20concrete%20repairandprotectionsystems/mbrace/mbracesaturant/pages/default.aspx,%20(accessed:%2018%20feburary%202018) file:///c:/users/raoh/desktop/tjer_submitted%20paper%20(r)/civl180220/basf%20-%20mbrace%20saturant%20(2018)%20available%20at:%20http:/www.basf-cc.com.sg/en/products/%20concrete%20repairandprotectionsystems/mbrace/mbracesaturant/pages/default.aspx,%20(accessed:%2018%20feburary%202018) file:///c:/users/raoh/desktop/tjer_submitted%20paper%20(r)/civl180220/basf%20-%20mbrace%20saturant%20(2018)%20available%20at:%20http:/www.basf-cc.com.sg/en/products/%20concrete%20repairandprotectionsystems/mbrace/mbracesaturant/pages/default.aspx,%20(accessed:%2018%20feburary%202018) https://www.sciencedirect.com/science/journal/13598368 https://www.sciencedirect.com/science/journal/13598368 https://doi.org/10.1049/cp.2014.1082 https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic https://scholar.google.com/citations?view_op=view_citation&hl=en&user=w1blmocaaaaj&citation_for_view=w1blmocaaaaj:m3nemzrmikic http://www.sciencedirect.com/science/journal/13598368 http://www.sciencedirect.com/science/journal/13598368 http://www.sciencedirect.com/science/journal/13598368 http://www.sciencedirect.com/science/journal/13598368 s.e. el gamal, k. al-shanfari, a. al nuaimi and a. al-saidy beams strengthened with near surface mounted reinforced polymer (frp) bars. computational structures 109: 8-22. sika group, sikawrap-230 (2018) available at: http://gcc.sika.com/en/solutions_products/constr uction/02a013/02a013sa06/02a013sa06100/02a0 13sa0613.html, (accessed: 18 february 2018). soliman s, el-salakawy e, benmokrane b (2011), bond performance of near-surface-mounted frp bars. journal composite construcions 15(1): 103-111. 43 http://gcc.sika.com/en/solutions_products/constructi http://gcc.sika.com/en/solutions_products/constructi civl28103_latest.qxd ______________________________________ *corresponding author e-mail:makis@mix.meng.auth.gr the journal of engineering research 1 (2004) 07-18 facility location for solid waste management through compilation and multicriterial ranking of optimal decentralised scenarios: a case study for the region of peloponesse in southern greece a. karagiannidis*, g. perkoulidis, n. moussiopoulos and m. chrysochoou laboratory of heat transfer and environmental engineering, department of mechanical engineering, box 483, aristotle university, gr 54124 thessaloniki abstract: the present paper addresses the problem of locating solid waste management facilities. specifically, it studies and proposes optimal alternative solutions for the greek region of peloponnese, by examining facilities for transferring, sorting, treating and landfilling of wastes. quantitative and qualitative databases concerning the current solid waste management at the region have been created and used by the model. a customized mixed-integer linear network model has been developed and solved for various evaluation criteria on a single-criterion basis by the use of a location-allocation modeling framework. the solutions resulting from the parametrical application of the multicriterial method electre iii are then ranked for the entire criteria-spectrum. the best alternative scenario is presented for the region in accordance with current legislation on waste management, which maximizes environmental benefits and promotes recycling, in the frame of sustainable waste management. keywords: solid waste management, location, facilities, mixed integer linear model, multicriterial method, electre iii √ò¡d égqéñànéh πãe’g ägqé«ÿg âmîbgh ábqƒdg √òg â°ûbéf ~j~ëàdéhh áñ∏°üdg äéjéøædg iqgog ≥agôe ™bgƒe ~j~– á∏µ°ûe ¢ûbéæj ábqƒdg √òg :¢ü∏îà°ùÿg á°sgq~dg â– áñ∏°üdg äéjéøædg iqgog ≥aôÿ á«ø«ch ᫪c äéfé«h i~yéb aé°ûfg ” .äéø∏îª∏d »°vq’g ωoôdgh ,á÷é©ÿgh ,∞«æ°üàdg h ,π≤ædg å«m øe ≥agôÿg êpƒªædg aé°ûfg ‘ á£∏àıg og~y’g ägp á«£ÿg äéµñ°ùdg êpƒ‰ ∞jô©j ioéyg ” .êpƒªædg ‘ √òg äéfé«ñdg i~yéb âe~îà°sgh ¿éfƒ«dg üƒæéh ioƒlƒe »gh áœéædg êpƒªædg ∫ƒ∏m ö«jôj ” ºk .™bgƒÿg ~j~– álò‰ πµ«g ωg~îà°séh √~m ≈∏y πc áø∏àıg òjé©ÿg ~j~ëàd ¬dó¨à°sg ”h .á°sgq~dg ‘ ω~îà°ùÿg io~©àÿg òjé©ÿg á≤jôw ) qé«©e πc øy ¢ jƒ©àdg øere iiielect¬°shq~ÿg ™bgƒÿg ‘ ≥«ñ£à∏d ¬°vôyh ƒjqéæ«°s π° ag qé«àng ” ájé¡ædg ‘h .( .äéjéøædg iqgog ∫é› ‘ ôjh~àdg ioéy’ ºyoh á浇 á«æ«h ~fgƒa ≈°übg âféc áé«àædgh ,äéjéøædg iqgo’ á«dé◊g ú«fgƒ≤∏d é≤agƒe ¿ƒµj å«ëh io~©àÿg òjé©ÿg á≤jôw ,á£∏àıg og~yc’g hp »£ÿg êpƒªædg .≥agôe ,™bgƒe ,áñ∏°üdg äéjéøædg iqgoeg :á«méàøÿg ägoôøÿg am coefficient for calculating the material recovery from waste management facilities t·day /t waste·year ãe energy recovery coefficients concerning corresponding waste management facilities (rotary kiln and rdf facilities) mwh·day/t waste·year ååî energy recovery coefficients concerning corresponding waste management facilities (landfills) mwh·day/t waste·year ãî ghe emission coefficients for greenhouse effect concerning rotary kiln and rdf facilities eq. tco 2·day/twaste·year åï ghe emission coefficients fo r greenhouse effect concerning landfills eq. tco 2·day/twaste·year ä landfill capacity twaste/day notation 1. introduction environmental protection gained importance in legal, economical and technical terms in greece (european parliament and council, 1994, joined ministerial decision 114218, 1997; joined ministerial decision 69728/824, 1996; joined ministerial decision 113944, 1997). the location of facilities for integrated solid waste treatment and disposal has evolved into a complex issue during the recent years. furthermore, public acceptance and social opposition are exercising pressure that brings such issues at the top of the political agenda. this trend (in greece) gradually leads into interesting developments in the field of municipal waste management. in the late nineties, research focused on implementing the integrated management of municipal waste in various areas in greece (karagiannidis and moussiopoulos, 1997; karagiannidis, 1998; perkoulidis et al., 1998). furthermore, according to the new european legislation (european parliament and council, 1994a), several studies were conducted with regard to integrated solid waste management at prefectural level (karagiannidis et al., 1998; moussiopoulos et al., 1998; anatoliki, 1999, association of thessaloniki municipalities, 1999, perkoulidis et al., 1999; kouras et al., 2001, moussiopoulos, 2000; moussiopoulos et al., 2002). in order to reduce costs resulting from waste collection and material recovery, fixed and mobile transfer stations are proposed at the region of peloponnese, thus discouraging illegal dumping to the over 1,000 existing uncontrolled landfills at this region (drossos and terzopoulos, 1999). it must be pointed out that some material recovery facilities and sanitary landfills are currently in operation at this region. these facilities should be taken into account during the regional planning for solid waste management, as this is outlined by greek legislation (official journal of the greek government, 1999). 2. background of modelling techniques various deterministic mathematical programming models have been used for planning solid waste management systems. linear programming (hsieh and ho, 1993; lund and tchobanoglous, 1994), mixed integer programming (anderson, 1968; fuertes at al., 1974; gottinger, 1986; zhu and revelle, 1990), dynamic programming (huang et al., 1992) and multiobjective programming (perlack and willis, 1985) are included in these deterministic modelling techniques. a review of the various methods regarding sitting of facilities related with municipal solid waste management, was presented by karagiannidis et al., 2002. anderson and nigam (1967) examined closely the cost minimisation of waste flows from transfer stations to landfills through the implementation of a branch and bound system. furthermore, they developed an in-kilter algorithm, which disregarded, however, the existence of mass holes (gottinger, 1988). a mathematical model for selection 8 karagianidis et al. / the journal of engineering research 1 (2004) 07-18 ì local transfer stations typologies í material recovery facilities typologies î incineration facilities typologies o typologies of landfills ð local transfer station system ñ material recovery facility ó waste-to-energy facility ò uncompressed waste flow from local transfer station to the material recovery facility twaste/day ô landfill ö ì ð binary variable for locating local transfer stations 0 or 1 ÷í ñ binary variable for locating material recovery facilities 0 or 1 ø î ó binary variable for locating waste -to-energy facilities 0 or 1 ù ï ô binary variable for locating landfills 0 or 1 bð flow of waste from producer to transfer station ð twaste/day c flow of waste to material recovery facility twaste/day d flow of waste to waste -to-energy facility twaste/day e flow of waste to landfill twaste/day fv efficiency degree of material recovery facilities % fî efficiency degree of energy recovery facilities % g upper limit to the capacity of material recovery facilities t/day h upper limit to the capacity of waste -to-energy facilities t/day k upper limit to the capacity of local transfer station t/day mrj objective function concerning the material recovery from wastes t/year qî residues from waste -to-energy facility t/day sí wastes and residues from material recovery facility t/day u upper limit to the landfill capacity t/day between alternative solid waste management systems with linear constraints and a non-linear objective function was presented by helms and clark (1971). waste-to-energy facilities and sanitary landfills were considered as candidate facilities by the system. fixed costs were related to binary variables (0-1), while linear transportation and treatment costs to continuous variables. helms and clark considered that residues from waste-to-energy facilities were driven to existing landfill sites. marks and liebman (1970) dealt with the problem of transfer stations location by minimization of the total cost. a brunch and bound system was adopted using the fulkerson algorithm (1961). rossman (1971) extended the aforementioned study by adding waste-to-energy facilities in the set of candidate facilities. esmaili (1972) proposed an optimisation model for minimizing total waste management costs. gottinger (1998) and kirca and erkip (1988) consider the objective function of transportation costs as linear. erkut and neuman (1992) created a multi-objective model for locating undesirable facilities through minimization of the total cost. finally, caruso et al. (1993) studied the entire range of components that make up the integrated solid waste management system. the selection of the multicriterial method for the evaluation of solid waste management systems could be characterized as a post-multicriterial problem. saaty and alexander (1981) compared candidate areas for locating an undesirable facility through implementation of the analytic hierarchy process (ahp). vuk and koželj (1991) presented promethee and gaia for the selection of a landfill in slovenia, while hokkanen and salminen (1994) implemented electre iii for the selection of a solid waste management system in finland. karagiannidis and moussiopoulos (1997) compared integrated solid waste management systems for the greater athens area, greece, taking into account 25 criteria with electre iii, which was also used for the same purpose in other regions in greece (perkoulidis, 2001). 3. objectives the main objective of the work presented in this paper is the determination of the best location for solid waste management facilities (transfer stations, material recovery facilities, waste-to-energy facilities, landfills), as well as the allocation of wastes (from municipal waste producers and transfer stations) and residues (from treatment facilities). the following steps were followed for the selection of the best alternative solid waste management scenario in the region of peloponnese: . creation of quantitative and qualitative databases on produced amount of wastes as well as on residues and quantitative analysis respectively per waste producer (municipality). this data was obtained by recording the current status of solid waste management in the region of peloponnese. . elaboration of the databases with the main characteristics of the existing and candidate solid waste management facilities. these concern mainly the capacity, the investment and operation cost, the amount of recovered energy and the emissions of each facility. . definition of linear objective functions through regression analysis. for the determination of these functions, the aforementioned databases (main characteristics of facilities) were used. each objective function was sequentially adapted to each from a set of criteria; its value was determined as performance. . calculation of performances of each criterion. these resulted from the optimisation of the aforementioned objective functions. . application of a multicriterial method, in order to evaluate the aforementioned performances. . sensitivity analysis in order to derive the frequency distribution of each scenario's ranking from best to worst and lead to the proposal of the best alternative solid waste management scenario. 4. methodology the hierarchical facility location system comprises the following four primary levels (european parliament and council, 1994; fishbein and gelb, 1992 figure 1): (level-a) transfer stations, (level-b) material recovery facilities, (level-c) waste-to-energy facilities, and (leveld) sanitary landfills. a customized mixed-integer linear model is derived in a spreadsheet environment for the studied area by the use of a recent location-allocation, modelling framework (karagiannidis and moussiopoulos, 1998; perkoulidis et al., 1998). binary variables set equal to 1 (treated as constants) are the two existing facilities (one landfill in patra and one material recovery facility in kalamata), as well as the thirty-six considered transfer stations. the latter is assumed due to the relatively bad rural road network, which makes the use of transfer stations necessary. as waste producers the municipalities at the region of peloponnese were considered. an objective function is sequentially adapted to each criterion from a set of five evaluation criteria used in the present study: criterion-1 greenhouse effect (ghre): co2 equivalent of co2 and ch4 emitted from facilities operation and residue transportation (kt/year to be minimized equation 1, c.f, figure 1). the values of γξghe and eo ghe are given in table 1. (1) 9 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 ∑∑ ∑∑ ⋅ε+ ⋅γ= ο τ ξ σ ξ σξ )( )(min e dghe ghe o ghe (criterion-2) final disposal, i.e. the amount of solid wastes from waste producers and transfer stations, as well as the residues from material recovery facilities and wasteto-energy facilities that are finally disposed at sanitary landfills (kt/year to be minimized equation 2, cf. figure 1). (2) (criterion-3) energy recovery i.e., the recovered energy amount from waste-to-energy facilities and sanitary landfills (mwh/year, to be maximized equation 3, cf. figure 1, table 2). (3) (criterion-4) material recovery i.e., the amount of recovered materials from material recovery facilities (kt/year to be maximized equation 4, cf. figure 1, table 3) (4) criterion-5) total specific financial cost, i.e. facility investment-/operating cost, as well as transportation cost (€/t, to be minimized equation 5, figure 2, table 4). 10 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 figure 1. four primary levels of a hierarchical facility allocation system ο τ ξ ο σ τ ξο στ ο τ µ ο π τ µο πτ ο τ ο τ ο τ ωδ ωι ωη ⋅+ ⋅+ ⋅= ∑∑∑∑ ∑∑∑∑ ∑∑∑ i ifidimin ∑∑ ∑∑ ∑∑ ⋅ε+ ⋅γ+ ⋅α= ο τ ο τ ξ σ ξ σ ν ρ ν ρ )( )( )(max e d cer e j e j e jj ∑∑ ∑∑ ⋅γ+ ⋅α= ξ σ ξ σ ν ρ ν ρ )( )(max d cmr m j m jj emission coefficient facility ãî ghe åï ghe rotary kiln 53,89 rdf 53,83 landfill 12,28 table 1. emission coefficients for greenhouse effect (equivalent ktco2/year kolar, 1990; baldasano and cremades 1995; ipts, 1999). facility (j) coefficient rdf rotary kiln landfill ãej 96,7 93,6 ååj 42,7 table 2: energy recovery coefficients (mwh · day/twaste · year) concerning corresponding waste management facilities (karagiannidis, 1996). facility (j) coefficient am j ã m j material recovery facility 120,0 rdf 75,0 table 3: coefficients for calculating the material recovery from waste management facilities (karagiannidis, 1996) [5] where το cffi j represents the fixed cost per each one solid waste management facilities (€ /t) and cfvi j represents the variable cost for transport, treatment and final disposal of the wastes (€ / day). each solution resulting from the 5 aforementioned linear models are hence called 'scenario'. the constraints of each derived model refer to: (a) service demand, i.e. the produced amount of wastes is equal to the sum of capacities from the facilities of the system (equation 6, cf. figure 1). [6] (b) facility capacity. this constraint concerns the relationship between the planning capacity of the facility (upper capacity limit in table 4) and the arriving flows (wastes or residues): b1) bµπ < κµπ. ϕµπ (local transfer stations), b2) cνρ < gνρ . χνρ (material recovery facilities), b3) dξα < hξα . ψξα (energy recovery facilities) and b4) eoτ < uoτ . ωoτ (landfills). (c) mass input-output relation at facilities. in case this constraint refers to local transfer stations, then the equation bµπ = tµπ (mass conservation) is in effect. in case of a treatment facility, then the equations fv.cνρ = sνρ and f ξ . dξσ = q ξ σ are in effect for material recovery facilities and wte facilities, respectively. the f ν and f ξ coefficients are chosen from databases, which resulted from bibliographic data elaboration of material recovery facilities and incinerators (michos and pazvanti, 1999; kampataidis, 1998; and katsameni and korakis, 1996). (d) compactorand truck-capacity, maximum allowed gross truck weight and speed limits. solid wastes were allowed to either leave transfer stations uncompressed (in the case that they are sent to material recovery facilities, by mostly opencontainers), or compressed (in the cases they are sent to waste-to energy facilities or landfills). the location of waste management facilities becomes more complicated due to the resulting necessity for multicriteria evaluation and classification of alternative solutions. in this work, a knowledge base, figure 3, in combination with a database of solid waste management facilities in greece (moussiopoulos et al., 2000) and a relative model generator (karagiannidis and moussiopoulos, 1998), was used for the formulation of the location system and the multicriterial analysis of waste flow allocation. 11 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 figure 2. total cost for the transportation, treatment and final disposal of solid wastes in accordance with facility capacity, (kamgiannidis, 1996, perkovlidis, 2001) [ ] [ ] [ ] [ ]∑∑ ∑∑ ∑∑ ∑∑ ⋅+⋅+ ⋅+⋅+ +⋅+⋅+ ⋅+⋅= ο τ ο τ ο ω ο τ ο ω ξ σ ξ σ ξ σ ξ σ ξ ψ ν ρ ν ρ ν χ ν ρ ν χ µ π µ π µ ϕ µ π µ ϕ ω ψ χ ϕ ecfvcff dcfvcff ccfvcff bcfvcffftmax ∑∑∑∑ ∑∑∑∑ ++ += ο ω ο ω ξ σ ξ σ ν ρ ν ρ µ π µ π ηζ εα ii iiia 5. the case study area the region of peloponnese has a population of 1,171,.000 and lies on the southern edge of mainland greece, figure 4. from an administrative point of view, the region is composed of 7 prefectures: ilia, korinthia, argolis, achaia, laconia, arkadhia and messinia. the region is mostly highland, while there are some fertile plains, which mainly stretch along coastal zones. there are intense inter-district inequalities at the region, due to its separation to industrial and highland zones. three main road networks lead to places with abundant physical beau12 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 figure 3. proposed generic structure of the model-generating framework for integrated waste management models (karagiannidis, 1996) facility capacity range (t/d) investment cost range (€/t) operational cost range (€/t) total cost range (€/t) cfvi j cffi j landfill 50-3,000 57.00-18.00 -0.01 57.66 local transfer station 5-25 2.31-0.71 4.11 (const) 6.42-4.82 -0.08 6.82 material r ecovery facility 80-880 11.42-7.78 48.06-7.33 59.48-15.11 -0.06 63.92 rdf-s 530-3,000 21.08 (const) 61.10-41.26 82.18-62.34 -0.01 86.44 rotary kiln 100-2,650 26.66 (const) 71.52-10.99 98.18-37.65 -0.02 100.55 table 4. cffij and cfvij coefficients for calculating the total cost for the transportation, treatment and final disposal of solid wastes (karagiannidis, 1996, karagiannidis at al., 1996, perkovlidis, 2001) 13 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 figure 4. case-study area (stippled) figure 5. existing and candidate facilities in case study area ty and important archaeological sites. the solid waste collection system in the region of peloponnese covers 90% of its population. here, it should be mentioned that the number of existing uncontrolled landfills is estimated at about 1,000 at the time of the study. the produced amount of wastes is 400,000 t/year. concerning solid waste management facilities, the municipality of patra has proposed the creation of a multicollecting recycling program, where the construction of one material recovery facility is included. there is also already one sanitary landfill in operation in patra. its new cell, with a budget of 2,932,600 €, operates since 1995 and it is estimated that it will cover the needs of municipality for 15 years. furthermore, there is one material recovery facility in kalamata with a total capacity of 400 t/week, which serves the homonymous municipality since 1997. finally, reported plans concerning the construction of one sanitary landfill in korinthos, were also taken into consideration, figure 5. in this work, the 37 municipalities of the region of peloponesse were considered as main waste producers. related quantitative data is given at table 5. due to the dispersion of producers, the transport of produced wastes to distant facilities was proposed to take place by local transfer systems. finally, it was assumed that energy recovery is performed at all proposed landfills through collection and treatment of the produced biogas, thus also reducing the emitted methane, which contributes significantly to the greenhouse effect (energy information administration, 1999). the region of peloponesse is scheduled to be connected in the near future to the natural gas national network, but at the time the study was conducted it depended solely on the use of lignite for its energy demands. for this reason, the creation of wte facilities in industrial areas with increased energy demand was proposed (cf. figure 5). from the application of the model, five decentralized scenarios resulted as alternative solutions concerning the management of solid wastes for each of the five evaluation criteria: 1) minimal impact to the greenhouse effect (scenario 1), 2) minimal amount of landfilled wastes and residues (from the treatment of wastes) to landfills (scenario 2), 3) maximal energy recovery (scenario 3), 4) maximal material recovery (scenario 4) and 5) minimal total cost for the management of produced wastes (scenario 5). here it should be mentioned that due to the increased dispersion of solid waste management facilities and also due to the emphasis on local management at the period of the study, all five scenarios have a strongly local character (i.e. one landfill per prefecture figure 6). the capacities of existing facilities (landfill in patra and material recovery facility in kalamata) were set as constant. the results are presented at table 6. 6. multicriterial evaluation of results in the following multicriterial approach, all five aforementioned criteria were used. the performances of the five scenarios in all criteria were calculated by means of a pre-compiled knowledge base (karagiannidis et al., 1996). it must be emphasized that these were quite close to each other in all scenarios (cf. table 6) due to their decentralized character and the assumption that a great number of facilities existed variable capacity, which was in accordance with the expressed preferences of local decision makers (drossos and terzopoulos, 1999). therefore, apart from the already existing facilities, local authorities of the region have already expressed their interest for specific candidate facilities (cf. figure 5). for global evaluation and ranking purposes, the multi14 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 prefecture number of municipalities produced wastes (t/day) prefecture number of municipalities produced wastes (t/day) ilia 22 210 laconia 22 108 korinthia 15 166 arkadhia 23 122 argolis 16 111 messinia 31 196 achaia 23 436 total 152 1.349 table 5. produced wastes in the region of peloponnese greenhouse effect (equivalent ktco2/year) final disposal in landfill ( kt/year) energy recovery (gwh/year) material recovery (kt/year) total financial cost (€/t) 1 311 388 69 17 87 2 376 356 63 17 86 3 339 389 67 17 87 4 316 387 67 18 87 5 376 358 63 17 86 table 6. performance of the five scenarios criterial approach was chosen. multi-attribute utility theory, compromising programming, goal programming and discrete methods have been developed as specific multicriterial techniques. in this context, the well-established electre iii multicriterial method (rogers and bruen, 1998a-b; roy, 1986; 1990) was applied in the present study. a total of 243 independent weight combinations were derived from the sensitivity analysis, which resulted from 50%-step on the 5 criterion-weights (i.e. 0%, 50% and 100%). the frequency distribution of the five positions of each scenario is given in figure 7. 7. conclusions and discussion scenario 1 (min of greenhouse effect) was ranked best at 26% (63 cases) in the sensitivity analysis. it was followed by scenario 2 (min final disposal), which appeared at the top 17% of all rankings (cf. figure 7). scenario 1 combines recycling and thermal treatment of wastes. here, it should be mentioned that the same candidate facilities were proposed as existing in all scenarios. thus, the best alternative scenario also gives the optimal facility capacities (cf. figure 6). furthermore, scenario 1 is second-best 15 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 figure 6. best alternative solution for location of facilities and allocation of wastes flows figure 7. frequency distribution of each scenario’s ranking (position 1: best; position 5: worst) for the 243 considered weight combinations used in sensitivity analysis in terms of cost, following scenarios 2 and 5 (table 6). as far as the second-best position is concerned, scenario 1 (min greenhouse effect) is placed there more frequently than all the others (i.e. 35% of all cases). the proposed construction of a material recovery facility near the landfill of patra is in agreement with recent european legislation concerning the diversion of biodegradable wastes from landfills. furthermore, its existence will facilitate the continuous and steady flow of residue-derived-fuel to the waste-to-energy facility that is proposed to be located close to the city of patra, an area with high energy demand. three waste-to-energy facilities are proposed to be located in industrial areas with high energy demand. finally, seven landfills are proposed in the entire region, one for each prefecture. the proposed thirty-six local transfer stations can also be used as drop-off centers for recycling and sending compressed and/or uncompressed wastes to facilities at relatively large distances. they can also contribute to the decommissioning of 'wild' landfills (uncontrolled tipping sites), where the open incineration of wastes causes various forms of pollution. obviously, their existence reduces the total cost of solid waste management. furthermore, it should be mentioned that the best alternative solution concerning the proposed construction of a waste-to-energy facility and a material recovery facility in patra will combine incineration and recycling programs and therefore greatly reduce the volume of landfilled material. by operating effective recycling and collection programs, the prime concerns associated with msw incineration, i.e. the emissions of heavy metals, acid gases and toxic substances will be minimized (morrison et al., 1997). in conclusion, an optimal alternative solid waste management scenario was presented in this paper for a region in southern greece in accordance with current legislation (both hellenic and european) concerning waste management (common ministerial decision 114218, 1997; common ministerial decision 69728/824, 1996; common ministerial decision 113944, 1997; directive 369, 1989; directive 105, 1997). the final placement of facilities and the allocation of wastes and/or residues to them were influenced by: a) the morphological characteristics of the region, b) the existing road network, c) the physiognomy of grounds (i.e. industrial, agricultural, urban areas) and d) the preferences of local authorities involved in the waste management activities. finally, future research should consider a more centralized location of facilities (i.e. one landfill or treatment facility that will be used by at least two prefectures). thus, inter-prefectural collaborations could also be enhanced in the frame of regional integration of solid waste management. references anatoliki, s. a., 1999, “study of thessaloniki’s prefectural framework for integrated solid waste management,” anatoliki s.a. development agency of eastern thessaloniki, funded by the prefecture of thessaloniki, according to the new european legislation 91/156, 94/62 (in greek). anderson, l.e., 1968, “a mathematical model for the optimization of a waste management system,” university of california at berkeley, sanitary engineering research laboratory, serl report, usa, no. 68-1. anderson, l.e., and nigam, a.k. 1967, “a mathematical model for the optimization of a waste management system,” orc 67-25, operations research centre, university of california, berkeley. association of thessaloniki municipalities., 1999, “planning of solid waste management in the prefecture of thessaloniki,” (in greek). baldasano, j., and cremades, l., 1995, “atmospheric impact of a municipal wastes incinerator,” international journal of environment and pollution, vol. 5, no. 4-6, pp. 538-547. bontoux, laurent (1999), “the incineration of waste in europe: isasues and perspectives,” a report for the committee for environment, public health and consumer protection of the european parliament, european commission joint research centre, eur 18717 en, ipts, march. caruso, c., colorni, a., and paruccini, m., 1993, “the regional urban solid waste management system, a modelling approach,” european journal of operational research, vol. 12, pp. 19-28. common ministerial decision 69728/824, 1996, “ measurements and terms for solid waste management,” (in greek), government paper 358. common ministerial decision 113944, 1997, “national planning for solid waste management,” (in greek), government paper 1016. common ministerial decision 114218, 1997, “formation of norm standards and general programs for solid waste management, national planning for solid waste management, general directions of solid waste management,” (in greek), government paper 1016. common ministerial decision 14312/1302, 2000, “completion and specialization of common ministerial decision 113944/1944/1997 on national planning for solid waste management (general directions of solid waste management),” government paper 723, 1016/1997 (in greek). council directive 91/156/eec, 1991, amending directive 75/442/eec on waste, 18 march. directive 369 (1989), on the prevention of air pollution from new municipal waste incineration plants. directive 105, 1997, “proposal direction of the council concerning the leadfilling of wastes”. drossos, e., and terzopoulos, p., 1999, “multicriterial evaluation of best alternative solid waste management activities: implementation of electre iii at the region of peloponnese,” diploma thesis, laboratory of heat transfer and environmental engineering, department of mechanical engineering, aristotle university of thessaloniki (in greek). energy information administration, 1999, “emissions of 16 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 greenhouse gases in the united states 1998,” office of integrated analysis and forecasting, doe/eia0573, u.s. department of energy, washington, october. erkut, e., and neuman s., 1992, “a multiobjective model for locating undesirable facilities,” annals of operations research, vol. 40, pp. 209-227. esmaili h., 1972, “facility selection and haul optimization model,” journal of sanitary engineering, asce, 98/sa6. european parliament and council, 1994, directive 62, official journal of the european communities, no l 365/10, 31/12/1994. european parliament and council, 1994a, directive on packaging and packaging waste (94/62/ec). fishbein, b., and gelb c., 1992, “making less garbage,” a planning guide for communities, inform, new york. fuertes, l.a., hudson, j.f., and mark, d.h., 1974, “solid waste management: equity trade-off models,” journal of urban planning development, asce 100(2) (1974) 155-171. fulkerson, d.r., 1961, “an out-of-kilter method for minimal cost-flow problems,” j. soc. industrial and applied math., vol. 9, pp. 18-27. gottinger, h.w,. 1986, “a computational model for solid waste management with applications,” application of mathematical modeling vol. 10, pp. 330-338. gottinger, h.w., 1988, “a computational model for solid waste management with application,” european journal of operational research, vol. 35, pp. 350364. helms, b.p,. and clark, r.m., 1971, “locational models for solid waste management,” journal of urban planning and development division, 97/sa2. hokkanen, j., and salminen, p., 1994, “the choice of a solid waste management system by using electre iii mcda method, in: applying mcda to environmental management,” (m.paruccini ed.), luwer academic publishers, ecsc, eec, eaec, brussels and luxembourg, pp. 111-153 (isbn 07923-2922-8). hsieh, h.n., and ho, k.h., 1993, ”optimization of solid waste disposal system by linear programming technique,” journal of resource management and technology vol. 21 (4), pp. 194 201. huang, g.h., baetz, b.w., and patry, g.g,. 1992, “a grey linear programming approach for municipal solid waste management planning under uncertainty,” civil engineering systems, vol. 9, pp. 319-335. hellenic official gazette, joint ministerial decision (jmd) 69728/824, 1996, “measures and terms for swm,” hog 358b/17-5-1996. hellenic official gazette, joint ministerial decision (jmd) 113944/1302, 1997, “national planning of solid wm-general guidelines for solid wm, hog 1016b/17-11-1997. hellenic official gazette, joint ministerial decision (jmd) 114218/1302, 1997, titled “tutoring of the specification framework and general programs for solid wm,” hog 1016b/17-11-1997. hellenic official gazette, joint ministerial decision (jmd) 14312/1302,2000, titled “national planning of solid wm general guidelines for solid wm,” hog 723b/9-6-2000. kampataidis, chr., 1998, “evaluation of best alternative activities concerning the solid waste management activities at central macedonia region, greece,” diploma thesis, laboratory of heat transfer and environmental engineering, department of mechanical engineering, aristotle university of thessaloniki (in greek). karagiannidis, a., 1996, “modelling integrated waste management,” dissertation thesis submitted to the department of mechanical engineering of aristotle university thessaloniki (in greek). karagiannidis, a., 1998, “waste management and life cycle analysis: two against interactive thematic unities. life cycle analysis,” laboratory of heat transfer and environmental engineering, thessaloniki, july, pp. 98-95 (in greek). karagiannidis, a., p. korakis, l., katsameni, and moussiopoulos, n., 1996, “a data– and criteriabase for evaluating regional management scenarios for solid waste,” proc. 5th national congress on renewable energy sources, athens, vol. 2, pp. 341349. karagiannidis, a., and moussiopoulos, n., 1997, “application of electre iii for the integrated management of municipal solid wastes in the greater athens area,” european journal of operational research, pp. 439-449. karagiannidis, a., and moussiopoulos, n., 1998, “modelling framework for solid waste regional management with explicit consideration of local features,” location science, vol. 6, pp. 281-305 (in greek). karagiannidis, a., perkoulidis, g., and moussiopoulos, n., 1999a, “applying ahp for the optimal locational-policy selection of municipal solid waste management facilities,” proc. 5th int. conf. ”decision sciences institute,” (eds. d. despotis and c. zopounidis), athens, greece, 4-7 july 1999, vol. 2, pp. 1320-1322. karagiannidis, a., mousiopoulos, n., samaras, z., koufodimos, g., verganelakis k., kouimtzis, t., voutsa, d., and kouras, a., 1999b, “optimisation of waste management in thessaloniki,” final report-, research project funded by: general secretariat of research and technology, thessaloniki (in greek). karagiannidis, a., moussiopoulos, n., panagiotakopoulos, c., and perkoulidis, g,. 2002, “locating facilities for treatment and disposal of solid wastes: a review of current literature,” proceedings of the 1st international conference on solid waste management, athens, 28 february-01 march, cd (in greek). katsameni, m. l., and korakis, p., 1996,” formation of a criteria data bank for evaluating prefectural solid waste management scenarios,” diploma thesis, 17 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 laboratory of heat transfer and environmental engineering, department of mechanical engineering, aristotle university of thessaloniki (in greek). kolar j., 1990, stickstoffoxide und luftreinhaltung, berlin. kouras, a., voutsa, d., and kouimtzis, th., 2001, “characterization of bottom ash produced from combustion of fermentable matter of municipal solid wastes from northern greece,” fresenious environmental bulletin, vol. 10 (8), pp. 669-673. lund, j.r., and tchobanoglous, g., 1994, “linear programming for analysis of material recovery facilities,” journal of environmental engineering, asce 120 (5). marks, d.h., and liebman, j.c., 1970, “mathematical analysis of solid waste management,” usphs, bureau of solid waste management. michos, s., and pazvanti, e., 1999, “multicriterial analysis of best alternative waste management scenarios. implementation of the analytical hierarchy process (ahp) at western macedonia region,” greece, diploma thesis, laboratory of heat transfer and environmental engineering,” department of mechanical engineering, aristotle university of thessaloniki (in greek). morrison, k., robert, h., and cone, m., 1997, “the role of a municipal solid waste incinerator in a small community, municipal solid wastes, problems and solutions,” ed. r. e. landreth, p. a. rebers, crc press, inc. moussiopoulos, n., karagiannidis, a., salonikidiou, a., and perkoulidis, g., 1998, “waste management in thessaloniki – existing situation and future perspectives considering recycling, working paper in recycling,” diavlos, network of waste derived material recovery focused on plastics and packaging, thessaloniki (in greek). moussiopoulos, n., 2000, “creation and development of a bank with information and data concerning waste management in greece,” final report, hellenic general secretariat of research and technology (1998-2000), aristotle university thessaloniki – department of mechanical engineering, national technical university of athens – department of chemical engineering, university of aegean – department of environment, thessaloniki (in greek). moussiopoulos, n., 2000, “integrated information system for solid waste management,” funded by hellenic general secretariat of research and technology, gnomon s.a. (1999-2000) (in greek). moussiopoulos, n., samaras, z., kouimtzis, th., karagiannidis, a., voutsa, d., perkoulidis, g., koufodimos, g., and kouras, th., 2002, “sampling, field work and laboratory analyses in the frame of integrated investigation of municipal solid waste management in thessaloniki,” greece, proceedings of the 1st environmental conference of macedonia, 1 – 4 march, thessaloniki, pp. 206-212. official gazette of the greek government, 1999, national planning for the solid waste management, 6/1999. perkoulidis, g., karagiannidis, a. and moussiopoulos , n., 1998, “spreadsheet modeling of solid waste management,” proc. int. conf. protection and restoration of the environment iv, (eds. k.l. katsifarakis, g. p. korfiatis, y. a. mylopoulos and a. c. demetrakopoulos), sani, july, vol.2, pp. 628-634. perkoulidis, g., karagiannidis, a., and moussiopoulos, n., 1999, “applying mixed-integer linear programming for the evaluation of alternative integrated municipal solid waste management schemes in the region of western macedonia greece,” studies in regional and urban planning, vol. 7, pp. 75-90. perkoulidis, g., 2001, “elaboration and implementation of a multicriterial analysis and location system for investigating the perspectives of thermal treatment of solid wastes in greece,” dissertation thesis submitted to the department of mechanical engineering of aristotle university thessaloniki (in greek). perlack, r.d., and willis, c.e., 1985, “multiobjective decision-making in waste disposal planning,” journal of environmental engineering, asce 111(3), pp. 373-385. rogers, m., and bruen, m., 1998a, “choosing realistic values of indifference, preference and veto thresholds for use within environmental criteria with electre,” european journal of operational research, vol. 107, pp. 542-541. rogers, m., and bruen, m., 1998b, “a new system for weighing environmental criteria for use within electre iii,” european journal of operational research, vol. 107, pp. 552-226. rossman, l.a., 1971, “a general model for solid waste management facility location,” department of civil engineering, university of illinois, urbana. roy, b., (1986), “classement et choix en presence de criteres multiples (la methode electre),” riro, vol. 8, pp. 57-75. roy, b., 1990, “the outranking approach and the foundations of electre methods,” in c.a. bana e costa (ed.), readings in multiple criteria decision aid, springler verlag, pp. 155-183. saaty, t.l., and alexander, j.m., 1981, “thinking with models,” pergamon press, oxford. vuk, d., and koželj, b., 1991, “application of multicriterial analysis on the selection of the location for disposal of communal waste,” european journal of operational research, vol. 55, pp.211-217. zhu, z., and revelle, c., 1990, “a cost allocation method for facilities sitting with fixed –charge cost functions,” civil eng. systems vol. 7, pp. 29-35. 18 karagiannidis et al. / the journal of engineering research 1 (2004) 07-18 c:\users\raoh\desktop\paper 4.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 42-57 sustainability in small and medium-sized manufacturing enterprises: an empirical study i. garbie* department of mechanical and industrial engineering, sultan qaboos university, muscat, sultanate of oman department of mechanical engineering, helwan university, helwan, cairo, egypt received 21 january 2015; accepted 10 april 2015 abstract: the main goal of this paper is to analyze and investigate sustainable practices in small and medium-sized manufacturing enterprises (smes). a comprehensive analysis and a mathematical framework are used to assess the sustainability indexes (sdis) of each aspect/issue and pillar/dimension, and of the whole manufacturing enterprise. sustainability in the whole manufacturing enterprise is represented by the sustainable development index (sdi). the results show that there is a significant difference in sdis between aspects and dimensions, with economic sustainability representing the highest percentage in the sdi. also, the results show that industrial companies are adopting sustainable practices and applying them to most of the issues/aspects of the dimensions, and they can direct manufacturing companies in the industrial sector to develop strategies for sustainability. this paper introduces a new understanding of the practices and implementation of sustainability/sustainable development by smes through assessing the sdis. keywords: sustainability, sustainable development, manufacturing enterprises. / / / * corresponding author’s e-mail: garbie@squ.edu.om i. garbie 43 1. introduction the world co mmission on environment and development (wced) defined sustainability as development that meets the needs of the present without affecting the ability of future generations to meet their own needs (1987). sustainability/sustainable development (sd) is based on and is used to balance between three dimensions: economy, society, and environment. to achieve sustainability for the survival of productivity and a community’s well-being, some manufacturing enterprises must change their products or manufacturing processes which are related to events and processes outside their direct control, including their supply chains and the standard measure of sustainability performance. the environmental impact can be reduced through end of life product management, and a green design should be introduced with a manufacturing process that requires little energy and emits zero waste. to extend product life, preventive maintenance should be employed. small and medium-sized manufacturing enterprises (smes) are considered pivotal to growth and development, especially in developing countries (urban and naidoo 2012). smes are poised to solve socio-economic problems in developing countries as the manufacturing sector is one of the most significant contributors to gross domestic product (gdp), in spite of the fact that smes are characterized and challenged by uncertain markets and high failure rates. on the other side of the economic spectrum, developed countries such as the usa have markets that are dominated by smes with fewer than 500 employees (theyel and hofmann 2012). therefore, analysis and investigation of sd in smes is very important and should be appreciated. sd always emphasizes the satisfaction of people’s basic need so that they may enjoy a better quality of life (qol) without compromising t h e qol o f future generations (wced, 1987). sd means making decisions that integrate and incorporate the main dimensions of sustainability and is important and urgent as the recognition of its value is fundamental to implementing sustainable practices, especially in smes. sd has a lot of drivers and barriers (kulatunga et al. 2013). the drivers, or motivational factors, are numerous and represented by risk reduction, pressure from the market (customer demand), its potential to be used as a marketing tool (iso certification), government regulation (eg. legislation), limitations in the existing process improvement techniques, success stories of sustainable manufacturing in other organizations, economic benefits, and the availability of funds for green projects. implementing these drivers will reduce the possible risks of the smes and increase their positive reputations, resulting in a positive relationship with stakeholders and an increase in customer loyalty and, consequently, economic growth. the smes should be developed through product and process innovations, and increased market share and motivation of employees, with a simultaneous reduction in material and water consumption, and scrape rates. smes also face barriers such as the lack of awareness of sustainability concepts (garbie 2015a) within society at large, and of the potential rewards/benefits at the governmental level; a lack of consensus at the top level, a lack of awareness amongst members of localities of sustainability programs; negative attitudes towards the concept of sustainability; lack of support from employees; a lack of funds for green projects to offset the costs of unsustainable economic conditions, and difficulties in operation and maintenance (garbie 2015b). although the concept and practices of sd were developed over the last 30 years, assessment of its index at different levels (starting from aspect/issue and dimension/pillar) still needs more attention from academics and manufacturers. assessing sustainability practices is the main challenge faced by smes. this assessment also can be used to test manufacturing enterprises for sustainability and their potential to compete economically in the future. this paper will help smes’ stakeholders to measure each aspect and dimension of sd, and identify the weakness of each one. stakeholders in the smes will suggest solutions for the existing weakness. the main goals of this paper are: understanding sustainability’s concepts in smes, understanding the most significant aspects and dimensions of sustainability, considering the need for s m e s to become more sustainable and globalized, and new evaluations of smes in terms of sd. to achieve this, a case study will be analyzed and evaluated for sustainability through the three dimensions/pillars of sd based on a combined approach published by garbie (2013, 2014) and winroth et al. (2012). this paper is organized into several sections. section 2 discusses the literature sustainability in small and medium-sized manufacturing enterprises: an empirical study regarding sustainability concepts, practices, and assessment. the proposed sustainability analysis will be presented in section 3. sustainability assessments will be introduced in section 4. section 5 explains the implementation of the proposed analysis and measurements of sustainability in a real-life case study. conclusions and recommendations for further work will be discussed in section 6. 2. literature review many examples exist in the literature of sd as it applies to smes based on the number of employees in the company (lower than 500) according to theyel and hofmann (2012). sustainability definitions and better communication in the process toward sd were presented by glavic and likman (2007). sustainable engineering was recommended as a new educational course in engineering schools by davidson et al. (2010). hasna (2010) analyzed sustainability information in print press journals, periodicals and textbooks in order to provide a guideline for understanding sustainability science. garetti and taisch (2012) developed a fully detailed discussion about sustainable manufacturing which showed it to be one of the most pressing issues regarding sd. liu et al. (2011) reviewed recent advances in the decision making process for supporting sustainable manufacturing operations through design, manufacturing, and supply chain management. bi (2011) broadly discussed manufacturing system requirements, both clarifying their limitations and discussing their potential bottlenecks. rosen and kishawy (2012) investigated the integration of sustainability with the lines of different objectives, including function, competitiveness, profitability and productivity, while rosen (2012) identified key requirements for engineering sustainability, including resources, processes, increased efficiency, and reduced environmental impact. lozano (2012) provided an analysis of the most widely known initiatives which aimed to embed sustainability into companies’ systems. theyel and hofmann (2012) studied stakeholders’ effects on companies’ adoption of sustainability for social and environmental purposes by interacting with product and process innovation. the importance of operational skills regarding business sustainability was conducted empirically in the smes of south africa through five operations skills: inventory and operations management, quality control, productivity techniques, and support systems (urban and naidoo 2012). they indicated that entrepreneurs lack managerial skills. lee (2009) explored and investigated the process of green management in smes for both the short-term (internal) and long-term views (external). he found that smes can make themselves greener by making strategic and organizational changes. martins et al. (2011) used group reporting initiative (gri) to evaluate the efficiency of sustainability performance in a group of brazilian manufacturing companies. raising companies' awareness was suggested by using sustainability indicators as a new methodology by veleva et al. (2001) and veleva and ellenbecker (2001). ron (1998) suggested a five-step assessment procedure for sustainability. a framework for sustainable assessment was presented through the three dimensions of sustainability (garbie 2013, 2014). the performance of sustainable production indicators by using fuzzy measure and the analytical network process (anp) was evaluated (tseng et al. 2009). a methodology for designing manufacturing metrics was developed, taking into consideration two specific concerns: the availability factor and time remaining (weiser et al. 2008). a framework to assess the sustainability of operations in the manufacturing sector (labuschagne et al. 2005) and for social assessment (labuschagne and brent 2008) was proposed. leszczynska (2012) adopted the gri standard and methodology as a point of reference to analyze the reports of sustainability published by multinational organizations. performance assessment was developed by using a tool which considered a framework for sustainability indicators (azapagic 2004). integrating between sustainability pillars was investigated and studied to show which is more significant than the other (garbie 2015c). 3. sustainability analysis analysis of sustainability in manufacturing enterprises is based on the dimensions of the triple bottom line (tbl): economy, society, and environment. each dimension/pillar consists of several aspects/issues and each aspect/issue is based on several indicators. each indicator is assessed by using performance metrics. 3.1 economic analysis the essentials of economic sustainability considered in this section are business operations, employees, customer, research and development, operations management, and suppliers. 3.1.1 business operations (a1) it is important to improve the business 44 i. garbie 45 operations and financial returns of smes. as appropriate and timely business operations are made, sales and profit are considered major drivers/motivations to run a business, and the smes must focus on developing this issue. 3.1.2 employees (a2) employees must be motivated and their productivity must be monitored. employees’ feelings concerning job security, motivation and workplace comfort, and the accessibility of employees’ skills are considered indicators of employee satisfaction. 3.1.3 customers (a3) customers are the ultimate target for smes and determine company success as they provide sales and profit. ignoring customer satisfaction will lead to lost customer loyalty and a subsequent reduction in company revenue. therefore, the rates of customer complaints and the addition of new customers per year are considered indicators for a smes’s sustainability index (sdi). 3.1.4 research and development (a4) research and development (r&d) and innovation can increase creativity and services offered by a company. r&d also can help smes compete and remain sustainable at the market level. there are two indicators for measuring si regarding r&d: the number of new products related to the total number of existing products and the percentage of a company’s annual budget that is devoted to r&d. 3.1.5 operations management (a5) operation and production management includes activities related to manufacturing management. these activities are overall equipment efficiency, productivity, the performance rate of manual labor, the utilization of manual labor, delivery precision, manufacturing lead time, operation and production flexibility (eg. manufacturing flexibility), and a maintenance plan. these activities are also considered and used as indicators for the operation and production management si. 3.1.6 suppliers (a6) suppliers play an important role in the process of any smes as they offer raw materials, semifinished products, and equipment. thus, the delivery precision of suppliers is very highly appreciated. pauses in production due to suppliers’ failings are used as indicators for measuring si as it relates to suppliers. 3.2. social analysis the second pillar of the tbl is the social dimension. there are five aspects represented by this pillar: employees’ health, employees’ knowledge, labor management relationships, diversity, and human capital. each aspect has its own indicators. 3.2.1 employees’ health (b1) health care services for employees within a workplace are extremely important. the investment in health benefits for employees must cover all smes because it is considered the major component of a sustainable industry. the number of accidents, employee absences due to on-the-job injury, and the rate of elimination of work place hazards are considered indicators for assessing sustainability regarding health and safety aspects. 3.2.2 employees’ knowledge (b2) employees’ knowledge depends on employees’ levels of education, number of training hours received annually, and participation ratio in improvement activities. these are considered indicators of employees’ knowledge. 3.2.3 labor management relationship (b3) the analysis of the labor-management relationship in the smes involves the investigation of rules and policies which govern and organize employment. these rules and policies are measured through two main indicators: the rate of temporary workers and number of employees who are shareholders. 3.2.4 diversity (b4) understanding equity/equality means understanding the diversity between the social and cultural contributions of individuals. therefore, providing different and/or flexible services for all employees is important: equal access and opportunity for individuals and groups who have specific needs should be met. the presence of diversity and equal opportunity, maleto-female ratio, cross-functional teams for improvements, and the prevalence of discrimination are used as indicators to assess the si for diversity and equal opportunity. 3.2.5 human capital (b5) the human capital aspect represents the value of smes in a community. smes are found in specific regions and lead to job generation. they focus on the capacities of groups of people and their interactions. in this paper, some indicators are modified to assess the sustainability of human sustainability in small and medium-sized manufacturing enterprises: an empirical study capital. these indicators are company wages as they relate to the local minimum wage, the number of new employees hired per year, employee satisfaction rate, support for employees’ physical activities, health care and medical benefits offered, turnover rate, responsibility and competence, the clarity of employees’ job descriptions and promotion opportunities for all employees. 3.3. environmental analysis the environment can be affected and damaged by manufacturing-related causes. these causes include instability/inefficiency of equipment and a high rate of natural resource consumption. in order to attain environmental sustainability, smes may require a set of changes in these issues which are represented by manufacturing/industrial processes, type of equipment, amount of natural resources consumed, and waste management. therefore, smes may face many challenges to adapting and implementing changes. the five aspects/issues for environmental sustainability analysis will be discussed below. 3.3.1 natural resources (c1) natural resources are defined as land use, water consumption, the use of recycled water, purification of waste water, and rate of resource reuse. the availability of these resources plays an important role for smes. there are many smes which are considered high water consumers including refineries, petrochemical plants, special chemical producers, pulp and paper industries, electric equipment, and the food industry. some smes have reduced water consumption by reusing wastewater, regenerating water supplies, and implementing recycling or process changes. 3.3.2 energy (c2) the manufacturing sector in smes is a major consumer of energy which results in greenhouse gas emissions. in order to achieve energy efficiency and protect the environment, emissions of greenhouse gases and their resultant air pollution as well as fossil fuel consumption can be reduced, and the percentage of renewable energy use can be increased. in addition, it also measures renewable energy used by manufacturers. sometimes it is difficult to use renewable energy for the main area of production due to heavy power consumption, but it easily can be applied to lighting and/or air conditioning use. three indicators are used to determine the si of energy: the use of renewable energy sources, net energy use, and idle energy losses. 3.3.3 materials (c3) raw materials mainly are extracted from the earth, meaning supplies will be limited. therefore, the used material can be reduced by adopting new technologies such as recycling and remanufacturing. the recycling/reassembly of materials can extend their useful life beyond discarding them. there are four indicators used to assess si in relation to raw materials: net material usage, scrape rate, the rate of packing material, and the use of process additives. 3.3.4 waste and emissions (c4) waste and emissions are identified inefficiencies in the production process. environmental protection incorporates new techniques and processes for the treatment and disposal of waste. smes can achieve environmental sustainability by waste and disposal elimination, thereby improving efficiency. in this paper, six indicators are suggested to assess si regarding waste and emissions: total solid waste, weight of hazardous waste, emission of ozone-depleting substances, emissions causing acid rain, and the emission of particles of carbon dioxide. 3.3.5 environmental management (c5) the si of environmental legality and standard compliance is assessed by using four indicators. these indicators are represented as environmental accidents, cost for environmental health and safety, iso 14001 compliance, and environmental management systems. 4. sustainability assessments sustainability assessments are measured through the definition of proposed aspects and indicators. sustainable indicators have specific characteristics which come into consideration when designing a framework for assessing sustainability: measurable (quantitatively and/or qualitatively), relevant, understandable, and reliable (garbie 2013, 2014; rosen and k ishawy 2012). a procedure to assess the dimensions of sustainability is introduced and used in assessing the strengths and weaknesses of the smes. the triple bottom line (tbl) is used to identify the intersection between pillars of sustainability (fig. 1), suggesting that the intersection between sustainability pillars is not strong enough and needs more explanation to ensure the meaning of this value. figure 2 illustrates the strong relationship between sustainability pillars not only in interactions but also in incorporation. this means that economic sustainability is the core 46 i. garbie 47 figure 1. interacting between the triple bottom line (tbl) of sustainability pillars. economic sustainability social sustainability environmental sustainability a1: business operations b1: employees health c1: natural resources a2: employees b2: employees knowledge c2: energy a3: customers b3: labor management relationship c3: material a4: development b4: diversity c4: waste and emission a5: operations management b5: human capital c5: environmental management a6: suppliers figure 2. incorporating between sustainability pillars. 1 1 1 1 1 1 6 5 4 3 2 1 654321 5 6 4 6 3 6 2 6 1 6 6 5 4 5 3 5 2 5 1 5 6 4 5 4 3 4 2 4 1 4 6 3 5 3 4 3 2 3 1 3 6 2 5 2 4 2 3 2 1 2 6 1 5 1 4 1 3 1 2 1 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w a a a a a a a aaaaaa (1) sustainability/sustainabl e development index (sdi) sustainability in small and medium-sized manufacturing enterprises: an empirical study 48 aspect of sustainability, followed by society and environment. 4.1. framework design the main objective of designing a framework is to present a procedure for smes to assess their sustainability levels. the issues/aspects of each pillar were chosen and filtered after looking thro ug h different references throughout this framework based on the proposed mathematical framework published by garbie (2013, 2014) and an analysis proposed by winroth et al. (2012). indicators were assigned based on an appropriate unit of measurement. the unit for each indicator was selected to be either quantitative or qualitative. the proposed framework is based on determining the existing performance and the target value for each indictor. this procedure assists smes in understanding the difficulties within an existing status. this procedure can be also used to increase companies’ awareness of sustainability. the framework suggests a matrix that reflects the relative weights representing the relationship between all aspects in each pillar. these relationships also indicate the importance of each aspect relative to other aspects as shown in the following matrices. the relative weight of each aspect is measured using an analytical hierarchy process (ahp) (garbie et al. 2008a and b; garbie 2012; garbie and shikdar 2011a and b). the relative weights using ahp are estimated and changed frequently according to the new circumstance determined by a decision maker or a group of decision makers (garbie 2013) as shown in eqns. (1-3). these groups are represented by senior management, manufacturing and/or production management, and plant managers. the proposed framework of each dimension is shown in tables 1– 3 individually. table 1. economic sustainability framework. dimension aspect/issue indicator unit performance measures title code title code existin g value target value economy (a) business operations a1 profit, profitability a11 financial measures employees a2 value added/employee (productivity) a21 monetary units/# employment cost in relation to income sales a22 yes/no/degree access to skilled personnel a23 yes/no employee cost per hour a24 monetary units/# customers a3 rate of customers complaints a31 # of customers /unit # of new customers per year a32 #/year developme nt a4 # of new products related to total # of products a41 # of new products/total # of products % of annual budget to r&d a42 % operations manageme nt a5 overall equipment efficiency a51 % productivity a52 #/hours performance rate of manual labor a53 actual time /ideal time utilization of manual labor a54 % utilization delivery precision a55 % on time delivery manufacturing lead time a56 days, hours flexibility a57 range metric, time maintenance plan a58 # of days, hours suppliers a6 stops caused by suppliers a61 # of stops i. garbie 49 table 2. social sustainability framework. dime nsion aspect/issue indicator unit performance measures title code title code existing value target value social (b) employees health b1 # of accidents b11 # absence due to injuries or relate illness b12 # of days elimination of hazard work places b13 yes/no/de gree employees knowledge b2 # of training hours/employee b21 # of hours participation ratio in improvement group b22 # of group/# of employees level of education b23 average level labormanagement relationship b3 rate of temporary workers b31 # of temporary/ # of permanent rate of employees that are share holders b32 % diversity b4 equal opportunity b41 yes/no/de gree male to female ratios b42 # of female/# of employees cross functional teams b43 # of teams discrimination b44 yes/no/de gree human capital b5 company wage relate to local minimum wage b51 % # of new employees per year b52 # employee satisfaction rate b53 % support for employee physical activity, health care and medicine b54 yes/no employees turnover b55 % annual turnover rate responsibility and competence b56 yes/no/de gree clear job descriptions b57 yes/no/de gree promotion opportunities b58 yes/no/de gree sustainability in small and medium-sized manufacturing enterprises: an empirical study 50 table 3. environment sustainability framework. dimension aspect/issue indicator unit performance measures title code title code existin g value targ et value environment (c) natural resources c1 land consumption c11 square meter water consumption c12 cubic meter recycled water c13 % of total consumption purification of waste water c14 % of total consumption share of reuse or recycled c15 % of total consumption energy c2 use of renewable energy c21 % of total energy energy use c22 kwh energy use per unit idle of energy losses c23 kwh idle time energy use material c3 material usage c31 kg or cubic meter per unit scrap rate c32 % of material usage rate of packing material c33 % of material usage use of process additive c34 % of material usage waste and emission c4 total solid waste c41 kg or cubic solid waste weight of hazardous waste c42 kg or cubic meter of hazardous waste emission of ozonedepleting substances c43 kg or cubic meter ozone-depleting substances emission causing acid rain c44 kg or cubic meter of emission of nox emission of particles c45 kg emission of particles emission of carbon dioxide c46 kg emission of carbon dioxide environ mental manage ment c5 environmental accidents c51 # of accidents cost of ehs compliance c52 monetary units compliance with iso 14001 c53 yes/no environmental impact assessment is used c54 yes/no i. garbie 49 1 1 1 1 1 5 4 3 2 1 54321 4 5 3 5 2 5 1 5 5 4 3 4 2 4 1 4 5 3 4 3 2 3 1 3 5 2 4 2 3 2 1 2 5 1 4 1 3 1 2 1 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w b b b b b b bbbbb (2) 1 1 1 1 1 5 4 3 2 1 54321 4 5 3 5 2 5 1 5 5 4 3 4 2 4 1 4 5 3 4 3 2 3 1 3 5 2 4 2 3 2 1 2 5 1 4 1 3 1 2 1 c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w w c c c c c c ccccc (3) 4.2 mathematical approach for assessing sustainability the mathematical model for measuring the sdi was developed by garbie (2013, 2014) and is used to assess the proposed framework, especially in the eqns. 4 and 6. it contains a set of equations that can be used to calculate the sdi of each dimension/pillar through the calculation of each aspect/issue as follows: jm j j j y m m m m j e t sdi 1 (4) where: jsdi = sustainable development index for aspect/issue j, mjy = exponent of the change towards sustainability for indicator m in each aspect j, mjt value of indicator m towards sustainability, mje value of indicator m regarding the existing status, and jm number of indicators of aspect/issue j. jjj mmm ety log (5) the sdi for each dimension/pillar is estimated by using eqn. 6. in j jji sdiwsdi 1 (6) where: isdi = sustainable development index for dimension/pillar i (economic sustainability, social sustainability, environmental sustainability), relative weight for each aspect j, jn = number of aspects in the dimension i. jw = relative weights between aspects/issues for each pillar and they are determined as references (garbie et al. 2008a and b; garbie 2012 and 2015d; garbie and shikdar 2011a and b). the general sdi is estimated as eqn. 7. ccbbaa i ii sdiwsdiwsdiwsdiwsdi 3 1 (7) where, sdi is based on three pillars of sustainability (economic, social, and environmental) and the iw = the relative weight associated with sustainable development for each dimension/pillar i. 5. case study and implementation business operations (a1) table 4 s h o w s that the abc company’s output is a b o u t 96.6% through the use of eqn. 4. a s a result, the abc company needs 96.6% more effort (time and cost) than the existing value regarding business operations aspects. employees (a2) the sdi for employees is 129.18%. this value means that, in this respect, the abc company needs 129.18% more effort than the existing value. with respect to upgrading the sdi in this aspect, the abc company would have to increase the rate of employees’ pay. the abc company had to search extensively for highly skilled workers, showing that their pay rates were too low. it would be important to increase pay to gain more loyalty from employees. customers (a3) in this respect, the abc company has to increase the effort by 1.54 times (sdi = 154.80%) over the existing value. it was mentioned that the number of customers per year was two while the willing number was forecast to reach up to five. 51 sustainability in small and medium-sized manufacturing enterprises: an empirical study 52 table 4. sustainability indexes for aspects and economic dimensions. indicators code performance measures aspect/ issue index economic sustainability index existing target a11 3.37 4 0.966 1.470 a21 70 90 1.291 a22 yes yes a23 yes yes a23 2 2.25 a31 0.1 0 1.548 a32 2 5 a41 10 20 2.213 a42 5 7 a51 80 87 2.792 a52 10 15 a53 86 93 a54 75 81 a55 90 100 a56 2 1.6 a57 60 80 a58 1 0.8 a61 no no 0.0 10.10.10.133.10.1 0.110.167.00.20.1 0.10.110.133.183.0 50.150.10.110.20.1 75.050.075.05.015.0 0.10.120.10.10.21 6 5 4 3 2 1 654321 a a a a a a a aaaaaa (8) therefore, the company would have to expend much more effort in this respect based on marketing and advertising. development (a4) in regard to money spent to produce a new product or improving the existing one, it can be noted that the results is 213.30%. this means more effort towards sustainability is required than the existing one (table 4). operations management (a5) operations management includes the activities in the plant and was identified by sev eral indicators. the sdi of production manag ement was 279.20%, and the abc company has to work hard in this respect. the abc company requires almost three-and-a-half times more effort beyond what they make currently. in this respect, the abc company struggles because of late delivery and customer complaints. it can be noted that productivity and efficiency is an area of high consideration for employees and machines, respectively. suppliers (a6) the sdi of the suppliers is almost zero, and the abc company does not need any more effort for this aspect as it has several suppliers of aluminum. the relative economic weights of the aspects appear in the matrix represented by eqn. (8) and were estimated by using ahp with computing consistency and random indexes as 0.1884, 0.1056, 0.1967, 0.166, 0.1720 and 0.1713 for business operations, including employees, customers, development, operation management, and suppliers, respectively (table 4). the results are acceptable based on the inconsistency ratio which is less than 0.10. finally, after using eqn. 6, the economic sustainability index will be calculated as 1.470 (0.966 * 0.1884 + 1.2918 * 0.1056 + 1.548 * 0.1967 + 2.213 * 0.166 + 2.792 * 0.172 + 0 = 1.470). this value means that the abc company needs to exert 1.47 times (147%) more effort towards economic sustainability. sustainability in small and medium-sized manufacturing enterprises: an empirical study 53 table 5. sustainability indexes for aspects and social dimension. indicators code performance measures aspect/ issue index social sustainability index existing target b11 7 1~0 0.048 1.018 b12 7 1~0 b13 yes yes b21 7 9 1.1936 b22 5 7 b23 2 3 b31 n/a n/a 2.635 b32 0-1 5 b41 yes yes 1.468 b42 10/350 15/350 b43 5 7 b44 no no b51 80 90 0.73 b52 10 12 b53 80 95 b54 yes yes b55 20 10 b56 yes yes b57 yes yes b58 yes yes 133.10.250.00.1 75.010.567.05.0 5.02.015.05.0 0.25.10.210.1 0.10.20.20.11 5 4 3 2 1 54321 b b b b b b bbbbb (9) 5.2. the dimension of social sustainability in this analysis of the abc company, the collected data from five aspects were analyzed in areas of employee health, employee knowledge, labor-management relationships, diversity, and human capital (table 5). employees health (b1) the output results show that the abc company puts about 4.8% of its effort toward sustainability as compared with the existing one in regards to safety and health. the abc company gives more attention to the dangerous material and hazard, spending approximately ro 20,000 us$ 50,000 per year in safety and health. the abc company has an appropriate level of safety as well as instructors who do a good job helping employees avoid hazards. the company provides adequate safety equipment’s. employees’ knowledge (b2) the resulting value of social sustainability regarding employees’ knowledge is about 119%. the data show that the numbers of training hours, levels of education and participation ratios in the improvement group are high. it can be seen that the average level of education in the abc company is low with others having achieved only a high school degree. labor management relationship (b3) the data show that the value of labor management relationship is about 2.635 times. the reasons behind this low value is that the abc company has almost no shareholders among its employees and also because it has a small number of temporary workers. labor management relationship aspect is a very important in achieving social sustainability; the abc company should make a greater effort in this area. diversity (b4) the estimated value of diversity and equal opportunity needed to reach social sustainability is 146.80%. in this aspect, more effort is needed in the number of cross functional teams in order to sustainability in small and medium-sized manufacturing enterprises: an empirical study 54 achieve improvements in the abc company. also, the company needs to increase the number of females it employs. however, the abc company provides equal opportunities for all employees and does not discriminate based on gender, disability, or race. human capital (b5) the value of social sustainability regarding human capital in the abc company is 73%. this value reflects the percentage of improvement that the company should achieve in order to reach social sustainability in human capital. the abc company has good human capital but it needs to exert more effort to accomplish higher employee satisfaction. it needs also to increase the number of new employees each year. the relative weight of social sustainability based on the matrix represented by eqn. 9 were estimated as 0.2423, 0.2638, 0.094, 0.2064, and 0.1930 for employee’s health, employee’s knowledge, labor management relationship, diversity, and human capital, respectively while the results is accepted based on the inconsistency ratio which is less than 0.10. finally, and after using eqn. (6), the social sustainability index will be calculated as 1.018. the abc company has to achieve almost 1.018 times (101.80%). this value means that the social aspects need double more effort especially in labor management relationship, employees’ knowledge, and human capital. 5.3. the dimension of environmental sustainability according to the survey of the abc company (table 6), the data were analyzed as follows: natural resources (c1) the results show that 148.66% of the effort is needed regarding natural resources towards sustainability as compared with the existing value. also, since the abc company uses large amounts of water, recycling this resource will reduce consumption. energy (c2) the estimated value of the energy sustainability aspect index is 17.50%. the company uses regular, fossil fuel-driven energy sources due to the manufacturing processes and does not have plans to change to renewable energy. thus, because renewable sources cannot provide enough energy at a suitable cost, this concept was misunderstood by the managers of the company. material (c3) the value of the environmental sdi regarding material is about 77.70%. this means that the abc company has to increase its effort towards sustainability by 77.70% over the existing value. the main raw material used is aluminum, which is imported from dubai city. in addition, they import different kinds of paint based on customer orders. the abc company has a good recycling plan which involves sending the materials back to dubai for recycling and reuse. waste and emission (c4) the resulting value of the environmental sdi regarding waste and emissions was about 131.20%. the collected data show that the total emissions from solid waste and ozone depleting co2 and no3 are not high. environmental legal and standard compliance (c5) it can be observed that the value of the environmental sdi regarding environmental legal and standard compliance was 197.90%. this area does not need more effort. the relative weights of environmental sustainability were estimated based on the matrix represented by eqn. 10 and using ahp: 0.2494, 0.1168, 0.2494, 0.20, and 0.1834 for natural resources, energy, materials, waste and emissions, and environmental system management, respectively (see table 6) while the results is accepted based on the inconsistency ratio which is less than 0.10. finally, and after using eqn. (6), the environmental sustainability index will be calculated as 1.21. this means that the abc company needs to exert 121% more than their existing effort towards environmental sustainability when compared with the existing value. 5.4. overall sustainability index (sdi) the overall sustainability index for the abc company equals 123% [ (1.470 + 1.018 + 1.210)/ 3=1.23)] based on eqn. 7, with equal relative weights between the pillars/dimensions of sustainability. table 7 is used to illustrate the percent of each pillar/dimension of sustainability index with respect to the whole value. i. garbie 55 table 6. sustainability indexes for aspects and environmental dimension. indicators code performance measures aspect/ issue index environmental sustainability index existing target c11 49000 55000 1.4866 1.21 c12 3650 3500 c13 65 70 c14 0 0 c15 1 2 c21 n/a n/a 0.175 c22 1260950 1386825 c23 600 250 c31 1500 1700 0.777 c32 35 20 c33 80 90 c34 2 3 c41 1000 850 1.312 c42 50 20 c43 0.05 0.01 c44 n/a n/a c45 3 1 c46 5 4 c51 7 0 1.979 c52 30000 50000 c53 no no c54 no no table 7. percentage of pillar/dimension sustainability. pillar/dimension value % economy (a) 1.470 40.00 society (b) 1.018 27.50 environment (c) 1.210 32.50 15.00.150.00.1 0.2150.050.067.0 0.10.210.10.1 75.050.050.015.0 0.10.20.10.21 5 4 3 2 1 54321 c c c c c c ccccc (10) determination the importance of flexible weights between dimensions of sustainability pillars in general/overall sustainability index is shown in table 8 based on the different perspectives of the stakeholders. this flexibility is representing the sensitivity analysis of the relative weights between sustainability pillars/dimensions. table 8 is used to illustrate the differences between the relative weights between sustainability pillars and the overall sustainability index (sdi). it can be noticed from table 8 that when the importance of economic dimension increase (25-40%), the overall sustainability index is also increased (from 1.198 to 1.267). this means that economic value is important and significant in sustainable development and it can be achieved within its requirements. the associated changing of social pillar was decreased from 40.0% to 25.0%. this is no significant changes and significant in sustainable development and it can be achieved within its requirements. the associated changing of social pillar was decreased from 40.0% to 25.0%. this is no significant changes in social pillar comparing with economic and/or environmental dimension which keeps constant values. sustainability in small and medium-sized manufacturing enterprises: an empirical study 56 table 8. comparing between different perspectives. perspective # aw bw cw ai bi ci si 1 0.25 0.40 0.35 1.470 1.018 1.210 1.198 2 0.30 0.35 0.35 1.470 1.018 1.210 1.220 3 0.33 0.33 0.33 1.470 1.018 1.210 1.230 4 0.35 0.30 0.35 1.470 1.018 1.210 1.243 5 0.40 0.25 0.35 1.470 1.018 1.210 1.267 6. conclusion, contribution and recommendations for future work in this paper, an approach for assessing sustainability in smes was analyzed and investigated through a case study applied to different dimensions of sustainability (economy, society and environment). the case study shows that economic sustainability represents the highest value of sustainable development (40%) followed by environmental sustainability (32.50%) and social sustainability (27.50%) based on equal relative weights between sustainability pillars. this means more effort is needed to achieve economic sustainability than social and environmental sustainability. for the economic dimension, more work is still needed, especially in operations and production management; customers’ issues and development expenditures. with respect to environmental sustainability, some efforts are required in the areas of environmental, legal, and standard compliance. these results might change when applying the proposed approach in different studies. this means that sdi in any manufacturing enterprise, and especially smes, requires both human and institutional capacities in innovation, development, assessment, adaptation and diffusion. these capacities are considered requirements for supporting an industrial organization’s mission. the main contribution of this paper is to analyze and investigate the sustainability/sustainable development index in manufacturing enterprises through the real case study. this paper also illustrates how it is possible to model and assess sustainability indicators, and the proposed approach has become very useful assessment tool. the author is looking to apply the proposed approach with its associated questionnaires among numerous industrial companies to identify the index of sustainability/sustainable development index and to show the most common and the feasibility of sustainability indicators and performance measures. references azapagic a (2004), developing a framework for sustainable development indictors for the mining and minerals industry. journal of cleaner production 12: 639–662. bi z (2011), revisiting system paradigms from the viewpoint of manufacturing sustainability. sustainability 3: 1323–1340. davidson ci, hendrickson ct, matthews hs, bridges mw, allen dt, murphy cf, allenby br, crittenden jc, austin s (2010), preparing future engineers for challenges of the 21st century: sustainable engineering journal of cleaner production 18(7): 698–701. garbie ih (2012), design for complexity: a global perspective through industrial enterprises analyst and designer. international journal of industrial and systems engineering 11(3): 279–307. garbie ih (2013), dfsme: design for sustainable manufacturing enterprises (an economic viewpoint). international journal of production research 51(2): 479–503. garbie ih (2014), an analytical technique to model and assess sustainable development index in manufacturing enterprises. international journal of production research 52(16): 4876–4915. garbie ih (2015a), sustainability awareness in industrial organizations. procedia cirp, 26: 64–69. garbie ih (2015b), sustainability optimization in manufacturing enterprises. procedia cirp, 26: 504–509. garbie ih (2015c), integrating sustainability assessments in the manufacturing enterprises. international journal of industrial and systems engineering 20(3): 343–368. garbie ih (2015d), fundamental requirements for sustainability practices and implementation: an analytical modeling and empirical investigation. international journal of sustainable manufacturing 3(4): 333-362. garbie ih, parsaei hr, leep hr (2008a), a novel approach for measuring agility in manufacturing firms. international journal of computer applications in technology 32(2): 95–103. i. garbie 57 garbie ih, parsaei hr, leep hr (2008b), measurement of needed reconfiguration level for manufacturing firms. international journal of agile systems and management 3(1-2): 78–92. garbie ih, shikdar aa (2011a), analysis and estimation of complexity level in industrial firms. international journal of industrial and systems engineering 8(2): 175–197. garbie ih, shikdar aa (2011b), complexity analysis of industrial organizations based on a perspective of systems engineering analysts. the journal of engineering research (tjer) squ 8(2): 1– 9. glavic p, lukman r (2007), review of sustainability terms and their definitions. journal of cleaner production 15: 1875–1885. garetti m, taisch m (2012), sustainable manufacturing: trends and research challenges. production planning and control 23(2-3): 83–104. hasna am (2010), sustainability classifications in engineering: discipline and approach. international journal of sustainable engineering 3(4): 258–276. kulatunga ak, jayatilaka pr, jayawickrama m (2013), drivers and barriers to implement sustainable concepts in sri lankan manufacturing sector. the 11th global conference on sustainable manufacturing berlin, germany 177–182. labuschagne c, brent ac, erck rp (2005), assessing the sustainability performances of industries. journal of cleaner production 13: 373– 385. labuschagne c, brent ac (2008), an industry perspective of the completeness and relevance of a social assessment framework for project and technology management in the manufacturing sector. journal of cleaner production 19: 253–262. lee kh (2009), why and how to adopt green management into business organization—the case study of koren smes in manufacturing industry. management decisions 47(7): 1101–1121. leszczynska a (2012), towards shareholders’ value: an analysis of sustainability reports. industrial management and data systems 112(6): 911–928. liu s, leat m, smith mh (2011), state-of-the-art sustainability analysis methodologies for efficient decision support in green production operations. international journal of sustainable engineering 4(3): 236–250. lozano r (2012), towards better embedding sustainability into companies’ systems: an analysis of voluntary corporate initiatives. journal of cleaner production 25: 14–26. martins ra, araujo jb, ometto ar (2011), investigation on use sustainability performance measures by brazilian manufacturing companies. proc. of the 2011 industrial engineering research conference, institute of industrial engineers, may 21–25, 2011, reno, nv, usa, 1121–1129. ron ajd (1998), sustainable production: the ultimate result of a continuous improvement. international journal of production economic 56-57: 99–110. rossen ma (2012), engineering sustainability: a technical approach to sustainability. sustainability 4: 2270–2292. rosen ma, kishawy ha (2012), sustainable manufacturing and design: concepts, practices and needs. sustainability 4: 154–174. theyel g, hofmann k (2012), stakeholder relations and sustainability practices of us small and medium-sized manufacturers. management research review 35(12): 1110–1133. tseng ml, divinagracia l, divinagracia r (2009), evaluating firm’s sustainable production indicators in uncertainty. computers and industrial engineering 57: 1393–1403. urban b, naidoo r (2012), business sustainability: empirical evidence on operational skills in smes in south africa. journal of small business enterprises development 19(1): 146–163. veleva v, hart m, greiner t, grumbley c (2001), indictors of sustainable production. journal of cleaner production 9(5): 447–452. veleva v, ellenbecker m (2001), indicators of sustainable production: framework and methodology. journal of cleaner production 9(6): 519–549. weiser cr, vijayaraghavan a, dornfeld d (2008), metrics for sustainable manufacturing. proceedings of the 2008 international manufacturing science and engineering conference (msec2008), evanston, illinois, usa, 1–9. winroth m, almstrom p, andersson c (2012), sustainable indicators at factory level—a framework for practical assessment. proceedings of the 2012 industrial and systems engineering research conference new orlando, florida, usa. world commission on environment and development (wced), (1987), our common future. oxford. uk: oxford university press. civl11203.qxd 1. introduction generally, the impact of most research results in the field of construction shows that it has not had a significant impact on the construction industry as noted by some authors amor et al. (2002). most construction it solutions such as integrated building models, including complex modular and process models, require highly organized and standardized project environments, which are not found in real-life construction projects. partners in a construction project consortium are typically at very dif_______________________________________ *corresponding authors e-mail: mnjadid@yahoo.com ferent levels of organization and it use. therefore, they are forced to use mutual digital communication at quite a low level. typically, isolated it solutions which did have a tremendous impact on the construction industry such as e-mail and telefax machines connected at different sites, made it possible to quickly share drawings, specifications and other documents needed in a technical environment during the cycle of a construction project. the widespread use of mobile phones was another it solution which made it possible to reach nearly anywhere at any time, thus, providing significant support for an undisturbed flow of information, but once again with an automation level of nearly zero, unfortunately. during the past few years, another device has evolved to a mature level the handheld perusing mobile computing and information technology in civil engineering construction projects m.n. jadid*1 and m.m. idress2 1department of building science & technology university campus projects, general supervisor for the university campus projects, p.o. box 30973, alkhobar 31952 saudi arabia 2 university campus projects it consultant, p.o. box 1982, dammam, saudi arabia received 1 december 2003; accepted 14 april 2004 abstract: mobile computing technology has an enormous potential in civil engineering, mainly in the field of construction management. this paper presents an approach by using mobile computing in a civil engineering construction project, exploring how information can be exchanged at a construction site, and how billing of quantities (boq), drawings and specifications are shared for the benefit of the project. one of the major obstacles in using information technology (it) when applied to civil engineering is due to the dynamic changes in the construction activities and the diversity of construction methods as compared to other industries. this has proved to be a great disadvantage when trying to give construction sites adequate it support. computers require infrastructure, which can be transferred easily to the head office, but rarely to the construction site itself in a systematic way. therefore, the impact of it on the way the construction industry actually works has not been significantly utilized. this paper consists of two main parts. the first part defines the concept of mobile computing, describes some recent research, and stresses the potential for the construction industry. the second part is focuses on an approach, implementing mobile computing in a real construction site. keywords: mobile computing, networking, personal digital assistant (pda), site, soil. vol. 2, no. 1 (2005) 25-31ring researchrnal of engineethe jou ::¢¢üü∏∏îîàà°°ùÿÿggá∏≤æàÿg á«æ≤àdg øe ioéøà°s’g á≤jôw ábqƒdg √òg ω~≤jh .™jqé°ûÿg iqgog ∫é› ‘ á°uénh á«f~ÿg á°s~æ¡dg ∫é› ‘ á∏fég äé«féµeg á∏≤æàÿg á浫ÿg á«æ≤j ôaƒj .´hô°ûÿg ídé°üd äéeƒ∏©ÿg ácqé°ûe ±~¡h ∂dph äéø°ugƒÿgh ä飣ıgh ä髪µdg ∫hg~l ¢üîj éã áhƒ∏£ÿg äéeƒ∏©ÿg ∫oéñj é¡æe øµá »àdg á«f~ÿg á°s~æ¡dg ™jqé°ûe ‘ iôn’g äéyéæ°üdg ™e áfqé≤ÿéh áyƒæàÿg aéæñdg ᣰûfgh qgôªà°séh ᫵«eéæj~dg ägò¨àdg ‘ á∏ãªàe á«f~ÿg á°s~æ¡dg ∫ééã á∏≤æàÿg á«æ≤àdg ωg~îà°sg ‘ á«°ù«fôdg ≥fgƒ©dg øeh ô≤e ¤g é¡dé°üjéh äéhƒ©°u ~lƒj éªæ«h »°ù«fôdg ô≤ÿg ¤g é¡dé°üjg π¡°ùj á«à– á«æh aéæh ¤g êéà– äéñ°sé◊g iõ¡lg ¿g .äéeƒ∏©ÿg á«æ≤j ∫é°üjcéh äéhƒ©°üdg ≈∏y ∫~j π‡ á浫ÿg ωƒ¡øã ±ô©j ∫h’g aõ÷g :ú«°ù«fq úfõl ábqƒdg √òg øª° àjh .™jqé°ûÿg ‘ ~«l πµ°ûh é¡æe ioéøà°s’g ºàj ⁄ äéeƒ∏©ÿg á«æ≤j ¿g ∫ƒ≤dg øµá ∂dòd ,´hô°ûÿg ωg~îà°sg á≤jôw ≈∏y êéãdg aõ÷g õcôj éªæ«h ,iôn’g äéyéæ°üdg ‘ ¬jéeg~îà°sg á«féµeg ‘ õ«cîdg ™e ∫éûg gòg ‘ áãj~◊g çƒëñdg ¢ ©h ¤g é° jg ¥ô£àjh á∏≤æàÿg .ò«øæàdg â–h áªféb ™jqé°ûe ‘ á∏≤æàÿg á浫ÿg ≥«ñ£jh ::áá««mmééààøøÿÿgg ääggooôôøøÿÿgg »°üî°ûdg »ªbôdg ~yé°ùÿg ,áµñ°t ,¬›èÿg äéjƒ∏ÿg, (pda) .áhôj ,™bƒe á«f~ÿg á°s~æ¡dg ägaé°ûfg ™jqé°ûe ‘ äéeƒ∏©ÿg á«æ≤jh á∏≤æàÿg á浫ÿg á«æ≤j ∫骩à°scg sonal computer (hpc), or personal digital assistant (pda) which is often resorted to. its speed, memory capacity, communication possibilities, reliability, small size and long power independence, as well as its level of hardware and software standardization gives the pda a powerful potential in the information chain of a construction project at the site. however, it does not operate on an isolated individual level, but rather in an integrated and systematic way, which means an organized combination of pdas, mobile phones providing wireless data transmission, and network-based document management. 2. an overview of the wireless lan and mobile computing wireless lan based mobile computing enables users to establish and maintain a wireless network connection throughout or between buildings, without the limitations of wires or cables gast, (2003). with a wlan, the covered area is analogous to a standard local area network. this might be a building or a campus environment, or even a small office or home environment. in large applications, by deploying a series of access points throughout a building or campus, you can achieve full coverage, providing the benefits of high-speed data rates, the freedom of mobility, and the ability to access broadband data anywhere within the local area network. wireless lansolutions are becoming quite prevalent in many horizontal applications in the construction, general office and vertical industries, such as health care, manufacturing, universities and retail. here are important but introductory facts about wireless lans: * most wireless lans operates in multiple frequency ranges unlicensed 902-928 ism (industrial, scientific and medical) band, unlicensed 2.4 ghz band and licensed 18-23 ghz band. * ieee 802.1x delivers a centrally managed, standardsbased, and open wireless network security scheme. in addition, the eap framework is extensible to wired networks, enabling an enterprise to use a single security architecture for every access method * there are two technologies used in wireless lans: direct sequence (ds) coding or frequency hopping (fh). * key hardware components of a wireless lan are a wireless lan adapter along with an antenna that resides in the mobile computer device (notebook or pda) and an "access point" which acts as the hub and connects multiple mobile devices to a server. you can realistically support only 10-30 mobile devices per access point, depending on the network traffic. * when the 14 db antennas are added, the outdoor wireless configuration is extended to 3.5 miles at an impressive 11 mbps throughput. when combined with any third party 24 db antenna, distances of 25 miles at 11 mbps can be achieved. 2.1 network configuration the access point's default configuration is as a root unit on a wired lan. the other two possible roles, repeater unit and central unit in an all-wireless network, require specific changes to the default configuration. 2.1.1 central unit in an all-wireless network in an all-wireless network, an access point acts as a stand-alone root unit. the access point is not attached to a wired lan; it functions as a hub linking all stations 26 the journal of engineering research vol. 2, no. 1 (2005) 25-31 figure 1. base station point to point configuration (c-spec’s web site) together as shown in fig. 1. the access point serves as the focal point for communications. 2.1.2 repeater unit that extends wireless range an access point can be configured as a stand alone repeater to extend the range of your infrastructure or to overcome an obstacle that blocks radio communication (fig. 2). the repeater forwards traffic between wireless users and the wired lan by sending packets either to another repeater or to an access point connected to the wired lan (fig. 3). 2.2 mobile computing architecture wireless lan consists of a hardware access point that generates wireless signals in a specific region (fig. 3). pda's wireless notebook or cellelular devices are used to access the features. a user can access text/email, text/email, web browsing and all internet features. the ieee 802.1x protocol is used in this technology. 2.3 availability of software and hardware mobile computing fundamentally differs from desktop computing. mobile devices including pdas, mobile phones and digital cameras, when compared to desktop computers have low computational power, small memory and often no mass storage. communication links to other mobile devices or to a stationary network are usually wireless. 27 the journal of engineering research vol. 2, no. 1 (2005) 25-31 figure 2. use of repeater (cisco-aironet) hand held computer laptop computer access point www cluster cellular device in te ro pe ra tib ili ty mobile computing technilogies graphics/ pictures text /email web browsing instant messaging word processor spread sheet application server database sql server applications exchange server firewall figure 3. mobile computing architecture four groups responsible for the implementation of activities in parallel are: * existing hardware, software and organization on site and in related offices, * mobile computers and suitable application software, * wireless communication technology (local networks, mobile phones and services), * project management, implementation of document management system. 2.4 data formats supported through wireless communication * text file, * pictures, * drawing, * video /multimedia. figure 4 shows an emulation of foreman's equipment demonstrating mobile cad running on pda at king faisal university construction site. 2.5 general software (cad, pms, gis) general software includes * island applications, * wickes lumber construction-management software (ordering of material), * sherpa (pipling operations), * bidcom (recording project data), * airwavz, * job cost. 3. mobile computing in construction industry there are a number of project tasks that can be automated in the construction industry a few common examples are provided above. some more applications are discussed in the following section. 3.1 site inspection in construction industry site inspection and project status reporting is considered as a high value mobile application by construction engineers because it reduces the number of site visits by supervising engineers by half or greater. while the construction industry has been slow in adopting mobile computing and automation in the field, things are changing quite fast now, since low-cost handheld devices are available and several construction automation applications have been developed by vendors. 3.2 impact of changes of site locations it is a fact that the locations of construction sites frequently change which can cause a delay in collecting data from the field, and therefore produces a great need and importance for mobile computing. the field client not only collects data, but it also provides information to the progress engineer. the annual report by daito (2000) mentioned four main benefits of the system: * it eliminates redundancy in project task operations, * reduces the response waiting time, * greatly limits revision of job tasks, * enables access to new construction standards. but, how many companies working together on a construction project would be able to follow this approach? it becomes evident that standard devices and services have to be available on the market at affordable prices in order to be attractive. special devices that could be highly applicable to construction activities include all the necessary input and out28 the journal of engineering research vol. 2, no. 1 (2005) 25-31 figure 4. king faisal university construction site (the foreman’s equipment) put devices where the user's hands are left free xybernaut, (2000) (fig. 5). the difficulty of controlling mobile computers by voice, which is also a requirement for wearable computers, has excited the interest of some researchers. in the field of civil engineering there are two interesting reports, one dealing mostly with inspection-oriented applications garrett, (2000), and other focusing on navigation through drawings reinhardt and et al. (2000). 3.3 construction industry applications * dispatch of construction crew to a job site, * project status monitoring and reporting, * on-site estimation of small jobs, * scheduling of construction equipment like graders, trucks, etc., * ordering of rental equipment to meet sudden demands, * transfer of equipment from one site to another, * tracking of labour time spent and the material used for cost plus projects, * crew card automation and transmittal through wireless networks. 4. mobile use at a construction site the new site for the king faisal university (kfu) campus in dammam, saudi arabia is located on the coast between cities of dammam and alkhobar. these two developing cities along with dhahran city are expected to evolve into a single integrated urban complex that will be the administrative and industrial center of the eastern region of saudi arabia. the total area of the campus is about 3,000,000 m2 and will be constructed in several phases. the total building area at the completion of the contract is about 570,000 m2. the new campus will accommodate male and female students in two distinct areas. each area will carry its own educational facilities, public services and housing. the main campus is divided into five main areas as shown in fig. 6 while table 1 shows the subdivision for each area. this division was made in order to acquire an efficient outcome in mobile computing at this construction site. with mobile computing, engineers, architectures, technicians and workmen can take advantage by means of timely and targeted information. the jobs can transmit through a fast medium like wireless devices to a particular engineer or technician, as well as collect data or daily 29 the journal of engineering research vol. 2, no. 1 (2005) 25-31 figure 5. daily activity plan (bill of quantities-boq), excel & pocket excel sheet figure 6. master plan for king faisal university (divided into five main areas) 30 the journal of engineering research vol. 2, no. 1 (2005) 25-31 1. main hospital area the psychiatric center main hospital the dental clinic the rehabilitation center the mosque the nurses’ resident 2. female medical center and administration college of dentistry college of medicine college of nursing college of applied science administration building basic science college female housing and activity center food services, telephone and telecommunication building 3. male medical center and administration college of dentistry college of medicine college of nursing college of applied science administration building basic science college male housing and activity center food services telephone and telecommunication building 4. sports area and the utility building open sports facilities student housing open sports fields closed sports facilities stadium housing utility building 5. housing unmarried faculty housing married faculty housing married student housing married staff housing table 1. colleges and facilities within the areas servers client site office consultant site office contractor site office hub et he rn et access point access point repeater supervisor / foreman hand held computer infrastructure construction in progress infrastructure construction in progress building construction in progress building construction in progress building construction in progress drawings documents drawings documents drawings documentsaccess point repeater flow of information through mobile computing in a construction site. figure 7. logical flow of information suggested at the kfu construction site reports from the field. the mobile computing technique can be used effectively in civil engineering in general. 4.1 infrastructure and buildings site conditions the preliminary geotechnical consultant's report (saudi services consultant, 2001) has revealed that the soil is very loose to loose, silty sand up to minimum of 4.0 m depth, followed by dense to very dense, silty sand up to 12 m. the resulting sub-grade shall be compacted by a heavy vibratory roller to a minimum of 70% of the relative density for cohesionless soil or to a minimum of 95% of modified proctor density for cohesive and well-graded soil. the recommended fill by the designer should be placed in layers not exceeding 300 mm each and compacted according to the type of material used. the above information revealed that the soil condition for the site is poor therefore soil treatment is recommended by the consultants. the scope of the infrastructure and buildings for the site includes but not limited to: * excavation and backfilling including slope protection, * roads and parking that requires base course, asphalting marking and signage, * walkways, curbs and gutters, * storm water drainage, * water, sewer, irrigation and chilled water networks, * electrical network and lighting systems, * sewer and storm pump stations, * reverse osmosis (ro) plant, * sewage treatment plant, * hospital, academic buildings, administration, housings, sport areas and the utility building. to overcome these shortcomings of the site conditions, an approach was suggested to provide mobile communication in order to utilize the flow of information between the client, consultant and the contractor effectively. the main approach proposed consists of the following tasks: * exploring the characteristics and potential of commercially available hand held devices for the construction industry, * evaluating the use of wifi or blue tooth technology with palms for data transmissions at the construction project site, * ensuring flow of information among king faisal university supervisors, consultants, contractors and the engineers at the construction site, * providing a document management system for mobile computing, transfer of text, drawings and spreadsheets through the software of latest hand held devices. the logical flow of the suggested information is shown in fig. 7, which represents the mobile computing works in construction process at the kfu site. 5. conclusions construction management can be achieved by using current mobile computing and web services. their full potential has not been utilized and there are rooms for improvement expected in the future. the commercially available pdas already offer abundant functionality and processing speed, however, suitability under stress conditions like dust, strong light, rain, which is normal at a construction site, is another problem which needs to be addressed seriously in the future. there is also the ever-present problem of screen size: either the screen is too small for a good information overview, or too big to fit in a pocket. it is clear that devices not fitting into a foreman's pocket will not be carried around, but rather left back at the site office. utilizing the latest in mobile computing at the construction site will provide innovation and customized solutions which would allow individuals in the construction site to maximize the efficiency when collecting data and accessing important information. acknowledgments the authors express their sincere thanks to king faisal university for providing the necessary data and information. references amor, r., betts, m., and coetzee, g., 2002, “information technology for construction: recent work and future directions,” itcon 17: pp. 245-258. daito trust construction co. (2000), annual report, http://www.kentaku.co.jp/e/tcs/. garrett, j.h., buergy, c., reinhardt, and sunkpho, j., 2002, "an overview of the research in mobile/wearable computer-aided engineering systems in the advanced infrastructure systems lab at carnegie mellon university," in the cd proceedings of the bauen mit computern (computer-aided construction) conference, bonn, germany, april 11-12, 2002. matthew, s. and gast, 2003, "802.11 wireless networks: the definitive guide,” (o'reilly networking). reinhardt, j., garrett, j.h. and scherer, j., 2000, "the preliminary design of a wearable computer for the supporting construction progress monitoring," internationales, kolloquium uber die anwedung der informatik under mathematik in architecture und weimar, germany, june 2000. saudi consulting services, 2001, king faisal university permanent campus, soil report, saudi arabia. xybernaut wearable computers. www.xybernaut.com/ 31 the journal of engineering research vol. 2, no. 1 (2005) 25-31 template for electronic submission to acs journals the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 doi: 10.24200/tjer.vol16iss1pp53-62 development and evaluation of ceramic tiles using wastes and solid minerals z.u. elakhame *,a , y.l. shuaib-babata b , and i.o. ambali c a department of prototype design and development, federal institute of industrial research, lagos, nigeria. b department of materials and metallurgical engineering, university of ilorin, ilorin, nigeria. c department of materials and metallurgical engineering, university of ilorin, ilorin, nigeria. abstract: this paper focuses on the need for domesticating the production of building materials like tiles to address the problem of nigeria over dependence on imported goods, despite adequate availability of mineral resources like clay, quartz and feldspar in the country. clay, quartz and feldspar, and milled glass were respectively obtained from ogijo in ogun state, okpila in edo state and oshodi, lagos state of nigeria were characterized using astm c71 as a guide. samples of the ceramic tile were produced from varying mixtures of clay, quartz, feldspar and milled glass. the chemical and physio-mechanical properties of the samples were also determined. the results showed that the clay sample belong to alumino-silicate group with 59.20% silica and 21.25% alumina. the samples’ properties met the required standards; hence, the materials were found suitable for production of ceramic tiles of acceptable standards. though, the porosity of the sample tiles increased as the percentage weight of the clay material in the mixture decreased the proportion of quartz, feldspar and milled glass increased. the sample containing 70% clay and 30% feldspar had better formulation properties than others. keywords: apparent porosity; bulk density; ceramic; tile; milled glass; and solid minerals. باستخدام النفايات واملعادن الصلبة الفخاري بالط السرياميكالوتقييم انتاج ، .يو. االخاميد ز *أ اتاباب بشعي .ال .واي ب اليبام .ه وأى. او ج على السلع ريبالكنيجرييا اعتماد مشكلة لحل حمليا تركز الدراسة على احلاجة إىل إنتاج مواد البناء مثل البالط: امللخص الطني وقد حصلنا على. يف البالد سيليكات االلومنيومتوافر املوارد املعدنية الكافية مثل الطني والكوارتز ورغم املستوردة والية اوكيال يف منطقة و اوغن والية يفأوجيو منطقةوالزجاج املطحون على التوالي من سيليكات االلومنيوم والكوارتز و بإنتاج عينات من بالط و قمنا كدليل. سي 7 استيم يف نيجرييا باستخدام الغوس اوشودي يف والية منطقةايدو و بتحديد اخلصائص قمناوالزجاج املطحون. كما االلومنيومسيليكات والكوارتز والسرياميك من خليط من الطني أظهرت النتائج أن عينة الطني تنتمي إىل جمموعة ألومينو سيليكات قد و .الكيميائية والفيزيائية امليكانيكية للعينات املواد دنا ان وجاستوفت خصائص العينات املعايري املطلوبة ؛ وبالتالي ، وقد . ألومينا ٪92.90سيليكا و 02.95بنسبة البالطات قد ازدادت مع اخنفاض عينة على الرغم من ذلك ، فإن مسامية. و مناسبة إلنتاج بالط السرياميك باملعايري باملقبولة وكان والزجاج املطحون. سيليكات االلومنيوم النسبة املئوية لوزن املادة الطينية يف اخللطة املختلطة وزادت نسبة الكوارتز و .أفضل من غريها تكوينخصائص سيليكات االلومنيوم من ٪ 05الطني و من ٪ 05حتتوي على اليت عينة لل املعادن الصلبة.؛ الزجاج املطحون ؛ البالط ؛ سرياميكال ؛الرئيسة الكثافة ؛الظاهرة املسامية مفتاحية:الكلمات ال *corresponding author’s e-mail: ezeberu@yahoo.com the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 1. introduction in nigeria, natural mineral resources, such as clay, feldspar, quartz are available in large quantities shuaib-babata and mudiare (2017), but the country is still highly being challenged with economic problems. some studies have attributed the present nigeria economic problem to over-dependence on foreign goods, oil and gas and to negligence of other sectors (chete et al. 2017; projectsxtra 2017; shuaibbabata et al. 2017a). it is claimed that an identified veritable tool towards any nation development is industrialization (shuaib-babata et al. 2017b; shuaib-babata et al. 2017c). availability of mineral resources in nigeria has been the backbone of the nation’s economic and industrial development aspirations foraminifera (2017). presently, there are several calls for diversification of nigeria’s economy and significant reduction of its overdependence on oil and gas to come out permanently of the present economic recession foraminifera (2017). production capacity from locally available raw materials used in development of technology in the transformation of the raw materials to finished products is a significant factor towards a successful industrialization (shuaib-babata et al. 2017a, shuaibbabata et al. 2017b). the identified bases for the development of any industrial sector include access to raw materials, labor force, funds and technology hughes (2008). effective industrialization could be significant to reduce overdependence of a nation on foreign goods and enhance export base shuaibbabata et al. (2017a). worldwide, production and consumption of ceramic tiles has recently increased all over the world as result of rapid development of real estate industry wifi ceramics (2017). the tiles used in the estate development are of different quality; it ranges from high to middle and low quality, which is imported to nigeria, especially from china. with the present call for economic diversification in nigeria, the need for local production of ceramic tiles with good quality is highly essential. this will be of help to proffer a solution to the problem and sustain the country economy through industrialization. there have been numerous attempts to utilize either industrial or agricultural wastes in the manufacture of wall and floor ceramic tiles in the past last two decades (elakhame et. al. 2016; cesar et al. 2011; juan, et al. 2010). industrial wastes are considered to be an important part of the materials used. cement kiln dust (ckd) as a great trouble maker waste, was utilized in ceramic tile bodies by youssef (2002), to replace half the feldspar required for ceramic wall tile body with a double target of cost reduction and environment protection. in that study, verification parameters of the fired samples was measured and the water absorption was within both iso and egyptian standards limits of wall tiles. ckd in the mixture makes it less dense, more porous and lowers the compressive strength (safiuddin et al. 2010; yoshizawa et al. 2004; central pollution control board (cpcb), 2000; torgal and jalali 2010). for better quality tiles, the basic raw materials employed for the production of ceramic tiles include quartz and feldspar (which serve as a source of silica and alumina to give the tiles the desired strength) and clay from a good source to serve as binder. in this study, milled glass and aluminum were considered as a primary source of alumina and silica. this study aims at the development of standard composite ceramic tiles-based (cct) using available mineral resources and wastes to aid local production of building materials in nigeria. it also focuses on the use of cheaper and durable materials to fabricate the cct to enhance its properties by variation of fillers to identify the best mechanical and physical properties of the produced tiles. in the end, this should assist in solving the problem of unemployment among youths. it should also help in diversification of economy and solve the problem of over dependence on the oil and gas sectors, which had led to economic recession in the country. 2. materials and methods 2.1 materials in this study, the major materials used include clay (cy), quartz (qz), feldspar (fr) and milled glass (mg). the clay material and the milled glass bottle was sourced from ogijo-shagamu area, ogun state and oshodi, lagos nigeria respectively. the quartz and feldspar were obtained at different deposits in okpila, edo state, nigeria. samples of the raw materials for the production of ceramic tiles in solid forms are shown in fig 1. 2.2 methods 2.2.1 chemical analysis the chemical compositions of the representative samples of the clay, quartz, feldspar and milled glass were determined at engineering materials development institute (emdi), akure, nigeria using atomic absorption spectrometer (aas) unicam 929. this is in line with the data reported by (ryan 1978; yoshizawa et al. 2004; and safiuddin, et al. 2010). clay feldspar quartz glass bottles figure 1. samples of the raw materials used for the production of ceramic tiles. 54 z.u. elakhame, y.l. shuaib-babata, and i.o. ambali 55 2.2.2 preparation of materials 2.2.2.1 clay materials to remove debris and other unwanted materials in the clay sample, the clay was soaked in water for five (5) days, sieved with a mesh of 350µm and dried at room temperature for one week (7 days). subsequently, the dried sample was milled and sieved into 100 µm sizes in line with the practice of elakhame et al. (2016b). 2.2.2.2 quartz, feldspar and milled bottle. the quartz and feldspar and the milled glass bottle (additive materials) were crushed using a hammer milling machine (model; 000t, puissane; 1.5kv, s/n; 13634) and milled into fine size particles using ball-milling machine (model; 87002…. limoges-france, type; a50---43), and then sieved with a vibro-sieve (model; fritsch gmbh, d55743.1 dar-oberstein germany) into 100 µm particles size. 2.3.2 raw materials mixtures in forming the raw materials mixtures for the production of ceramic tiles, milled bottle, quartz and felspar were added to the clay materials at different proportion as presented in table 1 (a & b). table 1a. proportion of clay and other materials in forming the raw mixes for the production of ceramic tiles (formulation a – c). samples sample formulation a sample formation b sample formation c cy (%) qz (%) cy (%) fr (%) cy (%) mg (%) 1 100 0 100 0 100 0 2 80 20 80 20 80 20 3 70 30 70 30 70 30 4 60 40 60 40 60 40 5 50 50 50 50 50 50 6 40 60 40 60 40 60 7 30 70 30 70 30 70 8 20 80 20 80 20 80 table 1b. proportion of clay and other materials in forming the raw mixes for the production of ceramic tiles (formulation d. formulation sample d cy (%) qz (%) fr (%) mg (%) 50 0 40 10 50 10 30 10 50 40 0 10 50 5 5 40 0 35 35 30 note: clay = cy, quartz = qz, feldspar = fr, and milled glass = mg 2.3.3 tiles production in development of ceramic tile samples, various raw materials mixes were processed sequentially using the process technology flow chart presented in fig. 2. figure 2. process technology flow chart. the milled powder clay, quartz, feldspar and glass bottles were individually measured appropriately as indicated in table 1 using a 0.001mg sensitivity weighing scale. the samples were sieved with 100 µm mesh and mixed to homogeneity. subsequently, the mixed material (granulated powder mixture) was poured into a prepared mould, 5cm × 5cm size (fig. 3) to form shaped test specimens and then uniaxially compacted using hydraulic press (weberhydraulic, ac-8800 vibrorg, denmark. type; p-16h, capacity; 16t, s/n; 29580) under a pressure of 50 mpa, in accordance with guidelines in mould pressed ceramic tiles specification (first revision of is 15622) ics91.100.23. the pressed specimens were held overnight and then dried at 90 100°c for 48 hours in an oven. dried specimens were fired in a laboratory type electric furnace (nabertherm, more than heat 303000°c) at the rate of 5°c/ min. the technological parameters values were measured after firing steps. the firing procedure used involves heating the sample at a temperature of 1170°c, then soaked for 3 hours and naturally cooled correlia et al. (2004). the procedures were repeated for all the formulations (a 55 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 56 d). the produced samples of the ceramic tile specimens produced are shown in fig. 4. 2.4 determination of the specimens’ properties some properties of the material mixes (moulded specimens) were tested for proper evaluation of their suitability. the properties include fired shrinkage, apparent porosity, bulk density, porosity and compressive strength, chemical composition and microstructural test. for the products to be in compliance with national and international standards, these properties were evaluated using various national and international test procedures and standards as presented in table 2. 2.4.1 porosity of the brick sample porosity of the brick sample was determined using astm jisd4418-1996 guidelines by soaking the bricks in a bath of water at 1000c for 8 hours. the mass of each of the samples before and after figure 3a. the mould. figure 3b. samples of ceramic tile specimens. table 2. various national and international test procedures and standards used for evaluation of the samples. s/n test procedures standards 1 specific gravity ms 474: part 1: 2003 2 compressive strain test method ms iso 6310:2003, part 4 3 determination of modulus of rupture and breaking strength ms iso (part 6): 2006 (first revision) 4 shear test procedure ms iso 6312:2003, part 6 5 determination of thermal conductivity iso 7882: 1986 6 porosity & apparent porosity measurement jis d 4418 – 1996 7 determination of linear thermal expansion (part 4): 2006 (first revision) 9 determination of water absorption and bulk density. (part 2): 2006 (first revision) doc no. ced 5(7888), title: pressed ceramic tiles – specification (first revision of is 15622) ics91.100.23. soaking was determined (in mg) and recorded. the porosity of the samples were then determined using equation 1. porosity (p) 100 v 2m1m            (1) where m1 is the mass (g) of brick before soaking, m2 is the mass (g) of brick after soaking water, ρ is the density of the liquid and v is the volume of sample (cm 3 ). 2.4.2 apparent porosity and bulk density of the brick samples the apparent porosity of each of the brick samples was also determined in accordance with jis d 4418 – 1996. brick samples (5 × 5 cm 2 ) dimension were used. the samples’ dried weight in air (wsw) and the saturated (wsa) of each sample were determined. the apparent porosity (pa) and bulk density (ρb) of each brick was determined using the relationships in equation 2 and 3 respectively. apparent porosity (pa) 100 swwsaw dawsaw     (2) bulk density (ρb) swwsaw daw   (3) 2.4.3 the compressive property the compressive property test was conducted on the brick samples using testometeric universal testing machine [tuf – c1000 kn (si)] as specified by the ms iso (part 6):2006 standard. brick samples with diameter of 5 × 5 cm 2 were loaded gradually in compression until the brick failed to offer further resistance to deformation which was indicated by the bricks fracture. various properties (stress, strain, extension and energy at different stages) of the samples were recorded from the display unit (monitor) attached to the machine. 2.4.4 measurement of fired shrinkage the shrinkage properties of the bricks were determined by measuring both the green and fired dimensions of the 5 × 5 cm 2 brick samples, using length comparator with digital indicator in accordance with the guidelines in astm c227 / and as specified by the iso 10545-2 standard. both sides of the specimens were measured and the average linear shrinkage was calculated. the linear shrinkage of each side was calculated as a percentage of the original green dimension as expressed in equation 4. linear shrinkage (ls) 100 gl flgl    (4) where lg and lf are green and fired length of bricks respectively. 56 z.u. elakhame, y.l. shuaib-babata, and i.o. ambali 57 3. results and discussion 3.1 materials chemical composition the results of the chemical analysis in table 3 show the major constituents of the clays, quartz and feldspar. 3.2 porosity and bulk density the porosity and bulk density values of the various produced moulded ceramic tiles samples are as shown in fig. 4 and 5 respectively. the specimens’ average porosity measurements ranged between 0.1 and 2.7%. it is shown in the results (table 2) that 80% and the above constituents of the clay and feldspar were silica (sio2) and alumina (al2o3). clay possessed 59.2% silica and 21.25% alumina, while fledspar had 49.16% and 34.02% of silica and alumina respectively. therefore, the clay and feldspar belong to alumino-silicate group in line with the view of (shuaib-babata 2016; abolarin et al. 2006; hassan 1993; encyclopedia britannica (2018a; and bastin 2018). quartz also had silica as its major constituents with 99.96% value. hobalt (2018) and encyclopedia britannica (2018b) affirmed that quartz primarily consists of silica. it is observed from fig. 4 that the fraction of the volume of the voids over the surface of the specimens produced were very closed-cell foam. however, the results revealed that specimens 6 and 8 in formulation c had the lowest porosity values of 0.1%. this might be the result of reduction in the clay percentage in the composition of the specimens. materials with low table 3. the chemical compositions of all the raw solid minerals. figure 4. average porosity values of different moulded ceramic tiles samples with varying compositions. figure 5. average bulk density values of different moulded ceramic tiles samples with varying composition. 0 0.5 1 1.5 2 2.5 3 1 2 3 4 5 6 7 8 p o ro si ty ( % ) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 1 2 3 4 5 6 7 8 b u lk d e n si ty ( g /c m 3 ) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) s/n parameters level of detection (wt%) clay quartz feldspar 1 sio2 59.200 99.96 49.16 2 tio2 0.005 0.047 0.01 3 al2o3 21.250 0.041 34.02 4 fe2o3 15.700 0.006 0.74 5 cao 1.920 0.100 2.87 6 mgo 0.880 0.007 0.08 7 na2o 0.050 0.000 2.63 8 k2o 0.040 0.000 8.40 9 mno 0.010 0.008 0.002 10 moisture 0.002 0.001 0.005 57 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 figure 6. average linear shrinkage values of different moulded ceramic tiles samples with varying compositions. figure 7. apparent porosity values of different moulded ceramic tiles samples with varying compositions. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 1 2 3 4 5 6 7 8 li n e a r s h ri in k a g e ( m m ) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) 0 20 40 60 80 100 120 1 2 3 4 5 6 7 8 a p p a re n t p o ro si ty ( % ) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) figure 8. stress – strain relationship in formulation a. 58 z.u. elakhame, y.l. shuaib-babata, and i.o. ambali porosity value are more preferred for production of tiles elakheme et al. (2016). generally, the average porosity values were within the recommended value of 0.13.5% for tiles (national research council of italy, 2017). the specimens’ average bulk densities were within 1.033 and 1.875 g/cm 3 (fig. 5), which fall within the recommended standard bulk density range of 1.0 to 2.5 g/cm 3 for tiles (hassan, 2011; national research council of italy, 2017) and 1.7 –2.1g/cm3 for dense ceramic materials hassan (2001). sample 4 in formulation d had the lowest bulk density value of 1.033 g/cm 3 , while samples 5 & 6 in body formulation c recorded the highest value of 1.87 g/cm 3 . materials with low density would be better for production of tile. 3.3 fired shrinkage and apparent porosity the samples’ linear shrinkage ranged between 0.08 and 0.40 mm (fig. 6), which were within the accepted range of 0.02 and 0.80 mm (national research council of italy, 2017). the samples’ linear shrinkage values at 1100 o c were significantly low. this is very advantageous for wall and floor tiles composition, which may be due to the presence of appreciable sio2 in the composition. however, the magnitude of shrinkage increases with clay contents but the limit is reached for higher specific water contents boivin et al. (2004). the samples’ calculated percentage apparent porosity ranged between 45.0 to 98.1%. these values were within the specified 30.0 to 60.0% values (national research council of italy, 2017). sample 8 in formulation a (80% and 20% quartz) had the lowest apparent porosity value of 45%, which is considered to be the best. this might be attributed to high proportion of sio2 in the composition. high shrinkage resulted to reduced apparent porosity in the closure of internal pores. 3.4 compressive strength test figs. 8 to 11 show the stress-strain relation of the samples, which are better illustrated in figs. 12 and 13 as an indicator for durability of the samples in service. the variations of calculated apparent porosity values in various samples of the composite are as presented in fig. 7. figure 9. stress–strain relationship in formulation b. figure 10. stress–strain relationship in formulation c. 59 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 figure 11. stress–strain relationship in formulation d. figure 12. average compressive strength values of different moulded ceramic tiles samples with varying composition. figure 13. average energy at maximum stress by different moulded ceramic tiles samples with varying composition. table 4. summary of the sample’s properties compared with the available information on the properties of tiles. s/n properties tiles {cy + qz & dolomite} [national research of italy, 2017] new formulation lab. tiles {cy + qz + mg & alumina} 1 compressive strength (mpa) 5.000 20.800 1.0117 14.331 2 bulk density (g/cm3) 1.0-2.5 1.2 1.8 3 porosity measurement (%) 0.1-3.5 0.1 2.7 4 apparent porosity % 30 – 60 45 – 98 5 shrinkage measurement 0.02-0.80 0.08 0.4 0 2 4 6 8 10 12 14 16 1 2 3 4 5 6 7 m a x im u m c o m p re ss iv e s tr e ss ( m p a ) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) 0 2 4 6 8 10 12 14 1 2 3 4 5 6 7 e n e rg y a t m a x im u m s tr e ss ( j) various samples of the composites clay+quartz (a) clay+feldspar (b) clay+milled glass ( c ) clay+m.g+q+f (d) 60 z.u. elakhame, y.l. shuaib-babata, and i.o. ambali within the formulation a, b, c and d, sample 4, 2, 5 and 4 exhibited the highest compressive strength values of 6.98156 mpa, 14,33139 mpa, 11.45160 mpa and 8.83201 mpa respectively in each formulation group. these strength values were within the recommended range of 5.00 to 20.8 mpa for tiles krivandin and markiov (1980). however, other samples that had strength values within the recommended values include samples 1, 2, 4, 5 and 6 in formulation b, samples 1, 2, 3 and 4 in formulation b, and samples 2, 4 and 5 in formulation c (fig. 12). from the obtained results, sample 2 in formulation b consisting of 80% clay and 20% feldspar exhibited the highest average compressive strength values (14,33139 mpa). this might be as a result of combination of high sio2 with al2o3 from both clay and feldspar. samples 4, 2, 5 and 4 respectively in formulation a, b, c and d also had energy at maximum stress of 4.74391 j, 11.77584 j, 9.53256 j and 7.5682 j respectively, in each formulation group. generally, the energy values exhibited by the samples ranged between 0.02818 and 11.77584 j. the majority of these energy values were considered to be adequate for sustainability of the tiles in service. to determine the stability of the samples, it is essential to compare the properties exhibited by the samples with that of available information on the properties of ceramics tiles as presented in table 4. thorough critical study of the data in table 3 shows that various properties exhibited by sample 4 (formulation a), sample 2 (formulation b), sample 5 (formulation c) and sample 4 (formulation d) met the required standards for tiles. therefore, the new formulation for production of ceramics tiles using locally available raw materials (clay, quartz, feldspar and milled bottles) proved suitable and adequate for local production of composite ceramic tiles in terms of strength and other properties. this will assist to address the nation’s economic problems and help towards job creation. 4. conclusion from this study, the following conclusions are drawn: 1. the clay and feldspar used in this study were found to belong to alumino-silicate group. 2. suitable combinations of nigerian natural resources and wastes such clay, feldspar, quartz and milled glasses were found suitable for the production of standard ceramic tiles. the suitable proportions of the materials in production of the tiles: 60: 40%, 80:20%, 50:50% and 60:5:5:30% of clay and quartz, clay and feldspar, clay, quartz and feldspar, quartz, feldspar and milled glasses respectively. 3. clay, quartz, feldspar and milled glasses were characterized to ascertain for their potentials for ceramic tiles, refractory and others ceramics applications. the results obtained showed that they meet the criteria for use as refractory/ceramic raw materials in the entire characteristic investigated via chemical composition, fired shrinkage, porosity, apparent porosity, bulk density and compressive strength. the clay can be used as a fired refractory, while quartz, feldspar and milled bottles could be used as a source of silica and alumina for ceramic products. 4. adequate utilization of available natural resources and wastes in nigeria such as clay, quartz, feldspar and milled glasses will help in domesticating production of building materials like tiles. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references abolarin m.s. olugboji o.a. ugwuoke i.c. (2006), determination of moulding properties of locally available clays for casting operations. assumption university journal of technolog, 9(4): 238-242 bastin e.s. quartz and feldspar: quartz and feldspar of maine. contributions to economic geology, part i. available online: https:// pubs.usgs.gov/ bul/ 0315l/report.pdf, accessed on december 20, 2018. central pollution control board (cpcb) (2000), report on management of municipal solid wastes, india, delhi: central pollution control board. cesar m, sánchez de rojas m. i. moisés f, andrés j (2011), using ceramic materials in ecoefficient concrete and precast concrete products. advances in ceramics – electric and magnetic ceramics, bioceramics, ceramics and environment (24): 533550, doi:10.5772/21302 chete l.n. adeoti j.o. adeyinka f.m. ogundele o (2017), industrial development and growth in nigeria: lesson and challenges, leaning to complete working paper no. 8, african growth initiative at brookings institute/african development bank group / united nations, university unu-wider world institute for development. available online: https:// www.brookings.edu/wp-content/uploads/ 2016/07/ l2c_wp8_chete-et-al-1.pdf, accessed on august 22, 2017. elakhame z. u. ifebhor f, asotah w. a. (2016), development and production of ceramic tiles from waste bottle powder (milled glass), kathmandu university journal of science, engineering and technology (kuset), 12(ii): 50-59. elakhame z, bello s. a. agunsoye j.o. patric i. otitoju o (2016b), characterization of ujogba clay deposits in edo state, nigeria for refractory applications. kathmadu university journal of 61 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 53-62 science, engineering and technology (kuset), 12(ii): 71-82. encyclopedia britannica (2018a), feldspar mineral – chemical composition available online: https://www.britannica.com/ science/feldspar/ chemical-composition, accessed on december 10, 2018. encyclopedia britannica (2018b), quartz mineral. available online: https://www.britannica.com/ science/quartz, accessed on december 22, 2018. foraminifera (2017), kaolin mining in nigeria, the pre-feasibility report. available online: https:// foramfera.com/ marketresearchreports/ miningand-quarrying/kaolin-mining-in-nigeria-the-prefeasibility -report/, accessed on december 24, 2017). finelib (2017), feldspar mineral deposits in nigeria and the enriched states. available online: https://www.finelib.com/about/nigeria-naturalresources/feldspar-mineral-deposits-in-nigeria-andthe-enriched-states/150, accessed on december 24, 2017. hassan s.b. (1993), refractory properties of some nigerian clays. nigerian society of engineers transactions 28(3): 21-25. hughes, j.r.t. (2008). industrialization. international encyclopedia of the social sciences, thompson gale, canada. available online: http://www. encyclopedia.com/history/united-states-andcanada / us-history/industrialization, accessed on june 12, 2017. hobalt m.k. (2018), what is quartz? available online: https://geology.com/minerals/quartz.shtml, accessed on may 22, 2018. juan a. medina c. morán j.m, guerra m.i. aguado p.j. isabel m. sánchez de rojas m.i. frías m. rodriguez o (2010), reuse of ceramic wastes in construction. ceramic materials, 1 st edition, in technology 197–214. intechopen, london, se19sg, u.k. krivandin v and markov b (1980), metallurgical furnaces. 1st edition, moscow: mir publishers. national research council of italy (2017). ceramic raw materials, institute of science and technology for ceramics. national research council of italy. available online: http://www.istec.cnr.it/ index.php/eng/our-research/high-tech-and-nano technology/389-ceramic-row-material, accessed on december 26, 2017. projectsxtra (2017), effect of imported goods on nigeria economy (comparative study). available online: http://www.projectsxtra.com/resources/ 302.html, accessed on november 3, 2017. ryan w. c. (1978), properties of ceramic raw materials, u.s.a.: oxford paragon press, 6. safiuddin m. d. jumaat mz. salam m.a. islam m.s. hashim r (2010), utilization of solid wastes in construction materials. international journal of the physical sciences 5(13): 1952–1963. shuaib-babata y.l. mudiare e. egwim c.e. (2016). suitability of using ado ekiti, akerebiata (ilorin) and birni gwari (kaduna) clays for production of household ceramic water filter. journal of engineering researc, 21(1): 77-79. shuaib-babata y.l. mudiare e (2017), characterization and evaluation of the refractory properties of local clay deposits in kwara state, nigeria. nigerian journal of mechanical engineering 7(1): 1-21. shuaib-babata y. l., abegunde a. j. and abdul j. m. (2017a), suitability of ado-ekiti (nigeria) natural moulding sands for use as foundry sands in production of aluminum alloy cast. journal of production engineering, 20(2): 91-100. shuaib-babata y.l. yaru s.s. abdulkareem s. ajayi s. busari y.o. ajao k.s. ibrahim h.k. ambali i.o. mohammed g.a. (2017b), suitability of some selected ado-ekiti (nigeria) natural moulding sands’ properties for sand casting. covenant journal of engineering technology (cjet), 1(2): 53-64. shuaib-babata y. l. abegunde e. ambali i.o. (2017c), evaluation of chemical and physicomechanical properties of ado-ekiti (south west, nigeria) natural moulding sands for foundry applications. arid zone journal of engineering, technology and environment (azojete), 13(5): 718-733. torgal f. p. jalali s (2010), reusing ceramic wastes in concrete. construction and building materials 24: 832-838. wifi ceramics (2017). the development of ceramic tile industry is inseparable from innovation. available online: http://www.wificeramics.com /en/ show news.hb?id=17&dir=1&nid=4&pnid=4, accessed on december 24, 2017. yoshizawa s. tanaka m. shekdar a.v. (2004), global trends in waste generation. in: recycling, waste treatment and clean technology. tms mineral, spain: metals and materials publishers 1541-1552. youssef n.f. (2002), utilization of cement kiln dust in the manufacture of wall tiles. industrial ceramics 21(1): 1-8. 62 https://www.britannica.com/ https://www.britannica.com/ https://geology.com/minerals/quartz.shtml http://www.istec.cnr.it/ http://www.projectsxtra.com/resources/ http://www.wificeramics.com/ c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 3.xps the journal of engineering research (tjer), vol. 12, no. 2 (2015) 22-31 efficient desalination of brackish ground water via a novel capacitive deionization cell using nanoporous activated carbon cloth electrodes k. laxmana,b, m.t.z myinta, m. al abric, l. al-gharibia, b. al namania, h. bourdoucenb,d and j. dutta*a *a chair in nanotechnology, water research center, sultan qaboos university, pob. 17, muscat, pc 123, oman b department of electrical and computer engineering, sultan qaboos university, pob. 33, muscat, pc 123, oman c department of petroleum and chemical engineering, sultan qaboos university, pob. 33, muscat, pc 123, oman d communication and information research center (circ), sultan qaboos university, pob. 17, muscat, oman received 24 september 2014; accepted 18 march 2015 abstract: sea water intrusion in ground water sources has made brackish water desalination a necessity in oman. the application of capacitive deionization (cdi) for the deionization of ground water samples from wells in al-musanaah wilayat is proposed and demonstrated. a cdi cell is fabricated using nanoporous activated carbon cloth (acc) as the electrodes and is shown to be power efficient for desalting ground water samples with total dissolved solids (tds) of up to 4,000 mg/l. the cdi cell was able to remove up to 73% of the ionic scaling and fouling contaminants from brackish water samples. the power consumption for deionization of brackish water was estimated to be 1 kwh/m3 of desalinated water, which is much lower than the power required to process water with equivalent tds by the reverse osmosis processes. the cdi process is elaborated, and observations and analysis of the ion adsorption characteristics and electrical properties of the capacitive cell are elucidated. keywords: capacitive deionization, activated carbon cloth, brackish water desalination. 4000 73٪ 1 . *correspondence author’s e-mail: dutta@squ.edu.om k. laxman, m.t.z myint, m. al abri, l. al-gharibi, b. al namani, h. bourdoucen and j. dutta 23 1. introduction seawater intrusion into ground water sources is a major problem in coastal regions worldwide. groundwater is the primary source of fresh water for more than two-thirds of the world’s population living in these regions. particularly in the sultanate of oman, which has more than 2,000 kilometers of coastline, rapid population growth and increased agricultural activities have severely stressed groundwater sources, which constitutes about 65% of available fresh water in the sultanate. increased demand for groundwater (1,650 mm3/annum) compounded by falling water tables has led to rapid sea water intrusion, costing the sultanate an estimated us$19–36 million annually (2012; 2006). to supplement the groundwater supply, the government has prioritized water desalination, enabling it to grow from 3% of the total water supply in 1999 to 13% in 2008 (gastli et al. 2010). although reverse osmosis (ro), with its modular technology, is at the forefront of this industry, the membranes used in ro are generally prone to fouling by dissolved macromolecular organic substances, soluble inorganic compounds, and suspended particles and microorganisms, leading to reduction in water recovery and membrane life (van der bruggen et al. 2002; vrouwenvelder et al. 2003). generally, extensive pretreatment like coagulation, flocculation, ph adjustment, and the addition of antiscalants, disinfectants, and media filtration are required to increase water recovery, which further increases operational costs of ro plants (ghabayen et al. 2004; hassan et al. 1998). typically, ro plants desalinating brackish water can remove 98% of the salt at total dissolved solid (tds) levels of 2,500–3000 mg/l, using pressures of 13.6– 17 atmospheric pressure (atm) and a flux of 24 l/m2/day. recent technological advancements like high efficiency pumps and energy recovery systems have reduced the expected energy requirements of brackish water ro to ~2–3 kwh/m3 (zhao et al. 2013). however, the ro process is restricted to certain operating conditions, including maximum temperatures of 45oc, ph values of 2–10, a silt density index (sdi)<3, and mandatory dechlorination, amongst others. this makes changes due to variations in water quality uneconomical in operational ro plants (nicolaisen, 2002). additionally, water recovery in ro is governed by the scale formation potential of brackish water, which is rich in scaling elements including calcium carbonate, calcium sulfate, magnesium sulfate, strontium sulfate, silicates, and aluminosilicates (amiri and samiei 2007; antony et al. 2011; farhat et al. 2012; ghafour 2003; tzotzi et al. 2007). scaling reduces productivity and the water recovery ratio, which are necessary to maximize water production due to restrictions on brine reinjection in oman (ahmed et al. 2001; mohamed et al. 2005). an alternative and emerging desalting technique which can circumvent these issues is called capacitive deionization (cdi), wherein emphasis is laid on removing salt from the water rather than separating the water from the salt as is done in ro (porada et al. 2013; zhao et al. 2010). since the volume fraction of salt is considerably lower than the volume of water in ground or sea water, the process has shown promise as a good alternative to ro, especially for brackish water deionization (lee et al. 2006; mossad et al. 2013; mossad and zou 2013; zhang et al. 2013). salt removal by cdi involves the electrosorption (potential induced ion adsorption) of ions on electrodes and can be reasonably fitted with the langmuir adsorption isotherm (mossad et al. 2013; ryoo et al. 2003; welgemoed and schutte, 2005) and modified donnan model, where equilibrium achieved between the adsorption and desorption of ions defines the salt adsorption capacity of the electrodes. since ion adsorption takes place on the surface of the electrodes, desalting efficiency is directly proportional to the surface area of the electrodes (zhao et al. 2010). the magnitude of the applied potential for cdi is generally kept below 2.0 v dc to avoid hydrolysis of the aqueous media (biesheuvel et al. 2009; oren 2008). in a cdi cell, one electrode is connected to the positive potential terminal and the other to the negative terminal of the dc power supply. when salt water is passed through the cell, anions (negative ions) will be electrosorbed to the positive terminal while cations (positive ions) will be electrosorbed to the negative terminal, thus depleting the water of ions and leading to desalted water at the output. since the adsorption is a surface related phenomenon, continuous deionization leads to saturation of the electrodes and subsequent inability to adsorb any further. to circumvent this problem, every deionization cycle is followed by a regeneration cycle, during which the applied potential is removed, leading to a release of the electrosorbed ions from the activated carbon cloth (acc) surface into the flowing solution, leading to a highly concentrated solution, known as brine, at the output. a key advantage of electrosorption is that the ions are efficient desalination of brackish ground water via a novel capacitive deionization cell using nanoporous activated carbon cloth electrodes 24 figure 1. (a) location of the al batinah region, which accounts for more than 50% of oman’s agriculture production and (b) salinity increase due to sea water intrusion in the al batinah region over the years (ministry of agriculture and fisheries, 2012). figure 2. representation of total land area affected by sea water intrusion showing that al-musanaah was heavily intruded (ministry of agriculture and fisheries, 2012). k. laxman, m.t.z myint, m. al abri, l. al-gharibi, b. al namani, h. bourdoucen and j. dutta 25 physisorbed onto the electrode surface, minimizing electrode fouling and leading to longer operational life of the electrodes. further, the absence of ion exchange membranes and the chemically inert nature of acc enables the facile removal of any organic or microbial contamination without adversely affecting the electrode integrity by washing in an acid bath (laxman et al. 2015; mossad et al. 2013). this manuscript demonstrates the use of cdi incorporated with acc electrodes for efficient removal of salt from brackish water collected from wells in al-musanaah wilayat in oman. emphasis is laid on the effect of an ionic charge on desalting capacity along with power consumed for the deionization of a well water sample to drinking water standards. 2. methods analytical grade hydrochloric acid 34% (hcl) (merck kgag, darmstadt, germany) was used without any further purification. fm-100 acc zorflex (calgon carbon corporation, pittsburgh, pennsylvania, usa) with a mean pore diameter of 0.7 nm and a total pore volume of 1.417 cm3/g was used. micropores accounted for 0.56 cm3/g and macropores for 0.85 cm3/g, while mesopores were very few (<0.001 cm3/g), making the material predominantly microporous. the acc also possesses a ribbed surface with a hierarchical pore structure which reduced the pressure drop across the electrode and is preferable for flow systems like cdi. nuclear magnetic resonance measurements estimated the surface area of pristine acc to be ~1000 m2/g. in addition to the high surface area, previous reports have suggested that the sheet resistance of symmetrically cut acc varies between 4–5 ω/sq cm, all of which make it suitable for application in cdi (subrenat et al. 2001). the electrode substrates (brasquet and le cloirec 1999; shim et al. 2001) were cleaned with highly concentrated 2 m hcl and heated to 115 ˚c, which is slightly above the boiling point of hcl solution, for 12 hours to remove mineral contaminants. subsequently, samples were thoroughly rinsed with deionized water, dried in a vacuum oven at 150°c for 12 hours, and stored in a desiccator until needed for further use. the cdi device (cell) used for the deionization consisted of two 10 x 10 cm acc electrodes separated by a spacer having a thickness of 650 μm. it was comprised of a reservoir made up of polymethyl methacrylate (pmma) from the sylgard® 184 silicone elastomer kit (dow corning, midland, michigan, usa), a current collector (graphite electrode), and an acrylic plate to support the cdi cell [figs. 3a–b]. for brackish ground water deionization, well water from two different wells located in al musanaah of the al batinah region in oman’s northwest was chosen. the al batinah region as a whole accounts for 50% of the country's agricultural production, and 95% of the abstracted water in the governorate is used for agriculture and livestock. however, this area has experienced a steep increase in groundwater salinity due to a high abstraction rate and seawater intrusion [figs. 1a and b; fig. 2]. of al musanaah’s farms, 40% are irrigated with 2–5 ms/cm brackish water, while only 35% are irrigated with <2 ms/cm of ground water (2012). the surface morphology of acc electrodes was studied with a jsm-7200 field emission scanning electron microscope (fesem) working at 20 kv (jeol usa, inc., peabody, massachusetts, usa). water conductivity was determined using eutech con 700 conductivity probe with a cell volume of 5 ml (thermo fisher scientific, inc., waltham, massachusetts, usa) which was modified for online real time recording. active surface area measurements were conducted using a nuclear magnetic resonance (nmr) relaxation technique in xigonanotools equipment (bethlehem, pennsylvania, usa) with water as the solvent. the changes in the t2 relaxation time after perturbation was measured and analyzed by acorn area software (xigonanotools) to interpret the surface area of the electrode using the specific surface relaxivity of activated carbon in contact with the solvent. cations were determined by inductively coupled plasma optical emission spectrometer (icpoes) 710 (varian, medical systems, palo alto, california, usa). power consumption was calculated by integrating the charging current decay curve in matlab (mathworks, natick, massachusetts, usa) to obtain the total charge deposited in one cycle. the charge deposited was multiplied by the applied voltage to obtain the work done in joules, which was subsequently converted to watts by dividing it by the duration of the deionization cycle. rates of ion adsorption were obtained by fitting the ion adsorption rates (with respect to time and number of deionization cycles) as obtained by icp with a cubic polynomial function with an r2 value greater than 99%, followed by taking the first derivative of the cubic fitting function. efficient desalination of brackish ground water via a novel capacitive deionization cell using nanoporous activated carbon cloth electrodes 26 figure 3(a). prototype of the cdi cell comprising two 100 cm2 acc electrodes separated by a spacer and (b) building blocks of the cdi cell showing the placement of the electrodes, spacer and electrical contacts. 3. capacitive deionization experiments the deionization/regeneration experiments were conducted using flow through capacitive deionization cell. capacitive deionization experiments were performed in a batch mode, where 50 ml of feed solution was passed through the cell at a flow rate of 5 ml/minute in a continuous loop for 15 minutes. a peristaltic pump drive 5201 (heidolph instruments, schwabach, germany) was used to maintain a constant flow rate of the feed water into the cdi cell. during this period, a dc potential of 1.6 v dc was applied across the cdi electrodes, leading to electrosorption of the ions on the acc electrodes [fig. 4a]. water samples were collected every five minutes during the deionization period to measure the ion content using icp-oes. the change in solution conductivity, current and voltage across the cell was monitored using gw instek gdm-396 online multimeters (good will instrument co., taipei, taiwan) for power consumption analysis. after 15 minutes of deionization, regeneration was carried out by removing the applied potential across the electrodes and letting the previously desalinated water flow through the system at five ml/minute. during this period, ions which were previously electrosorbed onto the electrode surface were released and removed from the cell [fig 4b]. the regeneration cycle was continued until original feed water feed water conductivity was recovered, confirming complete regeneration of the electrodes. in a second experiment, the above deionization period of 15 minutes was followed by a regeneration period of 20 minutes to complete one deionization-regeneration cycle. a total of seven such cycles were conducted on water from well two to attain water fit for drinking. water samples were collected after every cycle to measure the ion content and adsorption characteristics at different cycles. 4. results and discussion a fesem image shows the interwoven fiber structure of the conductivity of the acc electrodes [fig. 5] used for the deionization process. each fiber is composed of macro, meso, and micropores, which lead to a highly active surface area available for ion adsorption (garcía-quismondo et al. 2013; huang et al. 2013). the deionization performance of the cdi cell was studied with ground water samples from two different wells in al-musanaah. the analyzed composition and properties of the as received well water samples are given in table 1. k. laxman, m.t.z myint, m. al abri, l. al-gharibi, b. al namani, h. bourdoucen and j. dutta 27 figure 4. schematic showing the functioning of a capacitive desalination cell where (a) the desalination step involves the application of a 1.6 v dc potential across the electrodes and the subsequent electro-sorption of ions from the inlet feed water on to the acc electrode surface and (b) is followed by the regeneration step wherein the surface adsorbed ions are released to the inlet water feed by removing the applied potential to regenerate the electrode surface prior to the next desalination step. figure 5. field emission scanning electron microscopy image of activated carbon cloth (acc) showing the inter-woven fibers. table 1. measurements of properties of ground water samples from two wells in al-musanaah. sample conductivity (ms/cm) ph sodium (na) (ppm) calcium (ca) (ppm) potassium (k) (ppm) chloride (cl) (ppm) well 1 5.5 8.2 293 330 10 784 well 2 3.1 8.0 252 244 8 507 table 2. cation removal efficiency in the cdi cell for water from well one after a 15-minute deionization cycle as determined from inductively coupled plasma measurements. the conductivity of the water reduced from 5.5 ms/cm to 3.6 ms/cm. ion type initial concentration (ppm) final concentration (ppm) % removal sodium (na+) 293 190 35 calcium (ca2+) 330 184 44 potassium (k+) 10 7.7 23 efficient desalination of brackish ground water via a novel capacitive deionization cell using nanoporous activated carbon cloth electrodes 28 table 3. cation removal efficiency in the cdi cell for water from well two after 7 deionization cycles as determined from inductively coupled plasma measurements. the conductivity of the water was reduced from 3.2 ms/cm to 0.84 ms/cm. ion type initial concentration (ppm) final concentration (ppm) % removal sodium (na+) 252 39 85 calcium (ca2+) 242 35 86 potassium (k+) 8 3 63 figure 6. (a) desalination and regeneration performance of a batch mode process for water from well 1, showing 35% overall salt removal efficiency. the desalination was carried at a flow rate of 5 ml/min for 15 minutes at an applied potential of 1.6 v dc; (b) normalized ion adsorption curve for sodium (na+), potassium (k+) and calcium (ca2+) with time, suggesting that there was no surface saturation of the electrode during the period of desalination and (c) ion adsorption process was relatively constant for high concentration ions like sodium and calcium. water from well one has a tds of ~4 g/l with an initial conductivity of 5.5 ms/cm. from icp measurements, the calcium content in well one (330 ppm) was found to be more than that of sodium (293 ppm), while the potassium content was low (10 ppm). on being passed through the cdi cell continuously for 15 minutes (without any regeneration), the solution conductivity was reduced to ~3.6 ms/cm, effectively a 35% reduction in tds of the water with an electrode-specific salt adsorption capacity of ~8 mg/g [fig. 6a]. icp analysis of the desalinated water showed that na+ was reduced by 35%, ca2+ by 44%, and k+ by 23% following the wellestablished adsorption trend of higher selectivity for divalent and higher concentration ions over monovalent and low concentration ions (hou and huang 2013) [fig. 6b and table 2]. na+ ions had preferential selectivity for adsorption over k+ ions although they had the same ionic charge. in cases where the magnitude of charge is the same, ions with a higher concentration, like na + in this case, will be selectively adsorbed due to a stronger driving force mediated by the concentration gradients, resulting in easier ion transport through the acc pores (hou and huang 2013). figure 6b shows the comparison of normalized ion adsorption curves for the three ions, which confirms that ca2+ has a higher adsorption capacity as compared to the other two ions, and that the electrode surface is not saturated after the 15-minute deionization period. this was further confirmed as the first order differential plots of the ion adsorption curves showed that the slope of the tangent lines for na+ and ca2+ were either constant (na+) or still increasing (ca2+), which would not be the case if the electrode was saturated [fig. 6c]. the absence of electrode saturation is an important aspect of the cdi operation to maximize the charge efficiency, which is expressed as a ratio of electrical charge input to ionic charge adsorbed. the charge efficiency was found to be ~68%, expressing the k. laxman, m.t.z myint, m. al abri, l. al-gharibi, b. al namani, h. bourdoucen and j. dutta 31 power efficiency of the deionization process (huang et al. 2012; zhao et al. 2010). based on these results, the deionization period of further experiments was fixed at 15 minutes. to estimate the total power and time required to bring a saline ground water sample to drinking water standards, water from well two was passed through the cdi cell in multiple deionizationregeneration cycles. the deionization was carried out for seven cycles in a repeated batch mode for 200 ml of water from well two, where each cycle comprises a deionization period of 15 minutes followed by a regeneration period of 20 minutes. the solution conductivity was reduced from 3.2 ms/cm to 0.84 ms/cm, effectively removing 73% of the ion content from the brackish water from well two, with an average electrode salt adsorption capacity of 6.2 mg/g [fig. 7a]. the rate of ion adsorption was determined by taking the first order differential of the normalized ion adsorption curve, plotted with respect to deionization cycles. as the adsorption function continuously reduces with time, its derivative will have a perpetual negative slope, where a high negative value indicates a high adsorption rate. similarly, a positive slope (rising) indicates a reducing adsorption rate and a negative slope (falling) an increasing rate. as observed, the adsorption rate curves suggest that for both singly and doubly charged ions, the adsorption rate initially decreases with time to reach a steady state where the rate shows a value close to zero and then increases again when the solution ion concentration reduces further [fig. 7b]. this trend suggests that adsorption rates and solution ion concentrations are not constant and change with each cycle. in essence, the water was successively depleted of the salt ions in each cycle, with a total na+ removal of 85%, ca2+ removal of 86%, and k+ removal of 63% [table 3]. as the conductivity of the water sample decreased with an increasing number of deionization cycles, the power consumption also decreased [fig. 8], showing that the power was directly related to the ionic strength of the solution being desalinated. thus, water with a high ion concentration will require higher power, limiting the application of cdi for low to moderate saline ground water deionization (cohen et al. 2013; zhao et al. 2013). extrapolating the power consumption data from the desalinated one m3 of well two water to drinking water quality (from 3.2 ms/cm to 0.84 ms/cm), 1 kwh of power would be required, excluding power consumed for pumping. this in effect would be lower than the power required for ro of a similar water sample. the absence of a membrane in the design, small footprint with low manufacturing costs, and chemically robust nature of the acc electrode will further reduce the recurring costs of the system. additionally, the low voltage and current requirements of the cdi cell as compared to ro make it more effective as a portable deionization unit for use in remote locations (zhang et al. 2013). figure 7. (a) desalination performance of a multiple batch mode process for water from well 2, leading to 73% overall salt removal efficiency. the desalination was carried at a flow rate of 5 ml/min for 7 desalination (followed by regeneration) cycles of 15 minutes each at an applied potential of 1.6 v dc; (b) the ion adsorption rates for sodium, potassium and calcium increases and subsequently reduces with time, suggesting that ion removal rate varies linearly with the ion concentration in solution, with lower ion concentration leading to lower removal rates. 29 efficient desalination of brackish ground water via a novel capacitive deionization cell using nanoporous activated carbon cloth electrodes 30 figure 8. (a) integration of the area under the desalination current curve for 1 cycle to calculate the total charge delivered to the cdi cell during the desalination process. this charge multiplied by the applied potential (1.6 v dc) and divided by the time duration of the desalination cycle (15 minutes) to give the power (watt) consumption which was converted to watt-hour format; (b) graph showing the change in conductivity per desalination cycle and its associated power consumption for the desalination cycle in watt-hours, which suggests that power consumption reduces with conductivity (ion concentration). 5. conclusions the viability of the cdi device in desalinating ground water samples from the al-musanaah region of oman has been shown. the ion content and conductivity of 200 ml of brackish water (well two) was reduced from its initial value of 3.2 ms/cm to 0.84 ms/cm, making it drinking water quality, with just 0.2 x 10-3 kwh of power consumption. it is estimated that one m3 of water from well two could be desalinated to drinking water standards using one kwh of power, which is lower than that required by ro. additionally, scaling and fouling ions were shown to be efficiently removed during the process, making the cdi a viable option for brackish water deionization. acknowledgments the authors would like to thank the chair in nanotechnology, the research council of oman (trc), and sultan qaboos university for financial support. references ahmed m, shayya w, hoey d, al-handaly j (2001), brine disposal from reverse osmosis desalination plants in oman and the united arab emirates. desalination 133: 135–147. amiri m, samiei m (2007), enhancing permeate flux in a ro plant by controlling membrane fouling. desalination 207: 361–369. antony a, low j, gray s, childress a, le-clech p, leslie g (2011), scale formation and control in high pressure membrane water treatment systems: a review. journal of membrane science 383: 1–16. biesheuvel pm, van limpt b, van der wal a (2009), dynamic adsorption/desorption process model for capacitive deionization. journal of physical chemistry c 113: 5636–5640. brasquet c, le cloirec p (1999), effects of activated carbon cloth surface on organic adsorption in aqueous solutions. use of statistical methods to describe mechanisms. langmuir 15: 5906–5912. cohen i, avraham e, soffer a, aurbach d (2013), water desalination by capacitive deionization advantages limitations and modification. ecs transactions 45: 43–59. farhat s, kamel f, jedoui y, kallel m (2012), the relation between the ro fouling membrane and the feed water quality and the pretreatment in djerba island plant. desalination 286: 412–416. garcía-quismondo e, santos c, lado j, palma j, anderson ma (2013), optimizing the energy efficiency of capacitive deionization reactors working under real-world conditions. environmental science technology 47: 11866– 11872. gastli a, charabi y, zekri s (2010), gis-based assessment of combined csp electric power and seawater desalination plant for duqum oman. renewable and sustainable energy reviews 14: 821–827. ghabayen s, mckee m, kemblowski m (2004), characterization of uncertainties in the operation and economics of the proposed seawater desalination plant in the gaza strip. desalination 161: 191–201. k. laxman, m.t.z myint, m. al abri, l. al-gharibi, b. al namani, h. bourdoucen and j. dutta 31 ghafour e (2003), enhancing ro system performance utilizing antiscalants. desalination 153: 149–153. hassan a, al-sofi m, al-amoudi a, jamaluddin a, farooque a, rowaili a, dalvi a, kither n, mustafa g, al-tisan i (1998), a new approach to membrane and thermal seawater desalination processes using nanofiltration membranes (part 1). desalination 118: 35–51. hou ch, huang cy (2013), a comparative study of electrosorption selectivity of ions by activated carbon electrodes in capacitive deionization. desalination 314: 124–129. huang w, zhang y, bao s, song s (2013), desalination by capacitive deionization with carbon-based materials as electrode: a review. surface review and letters 20: 1330003-1330023. huang zh, wang m, wang l, kang f (2012), relation between the charge efficiency of activated carbon fiber and its desalination performance. langmuir 28: 5079–5084. laxman k, myint mtz, al abri m, sathe p, dobretsov s, dutta j (2015), desalination and disinfection of inland brackish ground water in a capacitive deionization cell using nanoporous activated carbon cloth electrodes. desalination 362: 126–132. lee jb, park kk, eum hm, lee cw (2006), desalination of a thermal power plant wastewater by membrane capacitive deionization. desalination 196: 125–134. mohamed a, maraqa m, handaly j, (2005), impact of land disposal of reject brine from desalination plants on soil and groundwater. desalination 182: 411–433. mossad m, zhang w, zou l (2013), using capacitive deionisation for inland brackish groundwater desalination in a remote location. desalination 308: 154–160. mossad m, zou l (2013), study of fouling and scaling in capacitive deionisation by using dissolved organic and inorganic salts. journal of hazardous materials 244–245: 387–393. ministry of agriculture and fisheries (2012), oman salinity strategy, muscat, oman. ministry of regional municipalities and water resources (2006). nicolaisen b (2002), developments in membrane technology for water treatment. desalination 153: 355–360. oren y (2008), capacitive deionization (cdi) for desalination and water treatment — past, present and future (a review). desalination 228: 10–29. porada s, zhao r, van der wal a, presser v, biesheuvel pm (2013), review on the science and technology of water desalination by capacitive deionization. progress in materials science 58: 1388–1442. ryoo mw, kim jh, seo g (2003), role of titania incorporated on activated carbon cloth for capacitive deionization of nacl solution. journal of colloid interface science 264: 414–419. shim jw, park sj, ryu sk (2001), effect of modification with hno3 and naoh on metal adsorption by pitch-based activated carbon fibers. carbon 39: 1635–1642. subrenat a, baléo jn, le cloirec p, blanc pe (2001), electrical behaviour of activated carbon cloth heated by the joule effect: desorption application. carbon 39: 707–716. tzotzi c, pahiadaki t, yiantsios s, karabelas a, andritsos n (2007), a study of caco3 scale formation and inhibition in ro and nf membrane processes. membrane science 296: 171–184. van der bruggen b, braeken l, vandecasteele c (2002), evaluation of parameters describing flux decline in nanofiltration of aqueous solutions containing organic compounds. desalination 147: 281–288. vrouwenvelder j, kappelhof j, heijman s, schippers j, kooij d (2003), tools for fouling diagnosis of nf and ro membranes and assessment of the fouling potential of feed water. desalination 157: 361–365. welgemoed tj, schutte cf (2005), capacitive deionization technology™: an alternative desalination solution. desalination 183: 327–340. zhang w, mossad m, zou l (2013), a study of the long-term operation of capacitive deionisation in inland brackish water desalination. desalination 320: 80–85. zhao r, biesheuvel pm, miedema h, bruning h, van der wal a (2010), charge efficiency: a functional tool to probe the double-layer structure inside of porous electrodes and application in the modeling of capacitive deionization. journal of physical chemistry letters 1: 205–210. zhao r, porada s, biesheuvel pm, van der wal a (2013), energy consumption in membrane capacitive deionization for different water recoveries and flow rates, and comparison with reverse osmosis. desalination 330: 35–41. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 4.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 32-40 variation in methane concentration produced from anaerobically digested vegetables h. mousa*a,b and m. silwadic a sultan qaboos university, department of petroleum and chemical engineering, p.o. box 33, al khoud p.c. 123, muscat, oman. b jordan university of science and technology, department of chemical engineering, p. o. box 3030, irbid 22110, jordan. c higher college of technology, applied science department, p. o. box 74, al-khuwair, p.c. 133, oman. received 17 february 2015; accepted 16 april 2015 abstract: anaerobic fermentation is a highly promising technology for converting biomass waste into methane, which then may directly be used as an energy source. attempts have been made to optimize various parameters in order to determine the most favorable recipe for maximum biogas production from fermented vegetable waste. the biogas production from many types of vegetable waste such as zucchini, orange peel, tomato, potato, and rice was studied in batch digesters. the effect of adding chicken dung and sludge to vegetable waste on the concentration of methane in the produced biogas was investigated. the experiments were conducted at room temperature (20°c) and at 35°c. the results revealed that methane concentration goes through maximum value with time. this maximum value is obtained faster when the rate of digestion is faster. the concentrations of methane in the biogas produced are ranked as follows: potato>rice>tomato>zucchini>orange peels. the concentrations of methane gas increased as chicken dung and sludge were mixed with the vegetables. the maximum value of methane concentration is reached faster in a mixture of chicken dung and sludge. for both chicken dung and sludge, the maximum value is reached at the same time. keywords: biogas, methane, carbon dioxide, renewable energy, fermentation, chicken dung, sludge, anaerobic digestion. 2035 . hasana@squ.edu.omcorresponding author’s email: * h. mousa and m. silwadi 33 1. introduction due to its high biodegradability, the anaerobic digestion (ad) of organic waste could play a role as an unconventional nonrenewable source of energy. a great deal of research has been conducted to study biogas production from different mixtures of organic waste (bouallangui 2005; berlian et al. 2013; demirbas 2006; singh et al. 2012; velmurugan and ramanujam 2011; yogita et al. 2012). various studies have been performed investigating biogas production from fruits and vegetable such as rice (okeh et al. 2014; ofoefule et al. 2011), potatoes (cheng et al. 2011; bernd 2006; kryvoruchko et al. 2009; mayer 1998; parawira et al. 2004; parawira et al. 2005), orange peels (nguyen 2013; wikandari et al. 2014; periyasamy and nagarajan 2012), zucchini (salter 2007), tomatoes (saev et al. 2009; sarada and joseph 1994), olive pomace (tekin and dalgic 2000), banana peels (nirmala et al. 1996) as well as many other types of organic waste (velmurugan and ramanujam 2011). past research has shown that on average 200-400 ml of biogas can be produced per gram of volatile organic solids (velmurugan et al. 2011). the productivity of the biogas is affected by the ph of the reactor, the organic loading rate, the temperature of the waste, the type of organic waste used, and the mixture of the reactor and water contents (babaee and shayegan 2010; bouallagui et al. 2004; bouallagui et al. 2005; weiland 2011). the biogas produced by the ad process contains 60% methane and 40% co2 as well as traces of other gases such as hydrogen sulfide, ammonia, nitrogen, and hydrogen. the anaerobic conversion of organic waste to biogas goes through four main steps: hydrolysis, acidogenesis, acetogenesis, and methanogenisis, with the latter responsible for methane gas production. accordingly, the digestion time, called hydraulic retention time (hrt), is a crucial parameter in maximizing the quantity of the biogas and its methane content. accordingly, the methane composition of the biogas is expected to increase with time, reaching a maximum value at which it is recommended to withdraw the biogas for energy generation purposes. in this paper, the maximum concentration of methane gas in the biogas produced from potatoes, zucchini, orange peels, rice, and tomatoes will be determined. the effect of mixing such vegetables with sludge and chicken manure as well as the digestion temperature and the time at which the maxima is reached will also be investigated. 2. experimental setup the experiments were carried out in 2.25 l plastic bottles, which were used as bioreactors. to ease gas sampling, the caps of the bottles were pierced, allowing hollow screws to go through. the screws were tightened well to the bottles’ cap to prevent any gas leak. plastic pipes were connected tightly to the outer part of the screws and sealed by clamps. when gas analysis was needed, the clamps were slightly loosened to allow the gas to be released [fig. 1]. the study was conducted using potatoes, zucchini, tomatoes, and orange peels which were bought from the local market, cut into pieces ~2 cm in dimension. then, 500 grams of a given vegetable were placed in a bottle and 500 grams of tap water were added so that the ratio of the solid waste to water was 1:1. three sets of these bottles were prepared. to each bottle in the second set, 25 ml of chicken dung was added and mixed with the water and vegetable waste. likewise, 25 ml of sludge was added to each bottle in set three and mixed with its contents. this allowed a study of the effect of chicken manure and sludge on the concentration of methane in the produced biogas. the bottles were prepared in triplicate in order to assure reproducibility of the results. the chicken dung was obtained from a nearby agricultural school and the sludge was obtained from the waste water treatment plant at jordan university of science and technology (just). the experiments were performed at room temperature (20°c) by placing the bottles on the shelves in the lab and at 35°c by placing the third set of bottles in an incubator. the produced biogas was analyzed by withdrawing the gas via a 20 ml syringe and then introducing the withdrawn sample to the gas analyzer (eagle type gas analyzer, union city, california, usa). the volatile solid content (vs) of each batch of vegetable waste, chicken dung, and sludge was determined a priori according to standard methods (apha, 1995). 3. results and discussion 3.1 water and volatile solids content the percentage of water of the selected organics is shown in table 1. the dried samples were analyzed for their volatile solid (vs) contents and the results are presented as kilograms of vs/kg of dry solids [table 1]. the results indicate that rice contains the lowest water content and the maximum vs whereas wastewater sludge has the highest amount of water and the lowest amount of vs. variation in methane concentration produced from anaerobically digested vegetables 34 figure 1. a photograph of the bottle containing solid waste. table 1. water content of tested solid wastes. waste type % water content vs (vs/kg dry solid) x 100% rice 65.82 ± 3.3 99.14 ± 0.25 potato 81.46 ± 2.44 95.15 ± 0.42 tomato 95.19 ± 0.36 82.44 ± 1.59 zucchini 94.55 ± 0.53 83.62 ± 0.82 orange peel 77.17 ± 0.02 96.84 ± 0.8 chicken dung 20.96 ± 0.43 86.28 ± 0.49 sludge 97.05 ± 0.15 71.28 ± 0.8 3.2 ch4 concentration in the produced biogas a comparison between the concentrations of methane gas in the biogas produced by the solid waste tested at 35oc is shown in fig. 2. potato was found to give biogas with a higher ch4 content compared to rice. table 2 shows that potato has higher potassium content compared to rice, explaining the higher content of methane in the biogas produced (machnica et al. 2008). orange peels, however, have a higher vs content but contain d-limonene oil that upsets the function of the anaerobic bacteria which explains the low methane content of the biogas (martin et al. 2010). the acidity of tomatoes negatively affects the productivity of the bacteria causing lower methane production. figure 2 shows that methane concentration is achieved through a maximum time value which is reached at different times according to the solid used. it can be noted that the faster the biodegradation process, the higher the maximum. the time necessary follows an order from high to low: potatoes > rice > tomatoes > zucchini > orange peels. this point needs further investigation and must be related to the amount of biogas produced to draw a solid conclusion. the amount of biogas produced and its methane concentration are important factors to determine the optimum hrt. the effect of temperature on the amount of bigas produced is presented in figs. 3a–d for zucchini, potatoes, tomatoes, and orange peels, respectively. the results are in accordance with other researchers’ findings where biogas of higher methane concentrations is produced at higher temperatures (beam 2011; nallathambi 2004; chua h. mousa and m. silwadi 35 figure 2. methane concentration in the biogas produced from the various solid wastes at 35°c. figure 3a. methane concentration in biogas produced from zucchini at 20°c and 35°c. figure 3b. methane concentration in the biogas produced from potato at 20°c and 35°c. figure 3c. methane concentrations in the biogas produced from tomatoes at 20°c and 35°c. variation in methane concentration produced from anaerobically digested vegetables 36 figure 3d. methane concentration in the biogas produced from orange peels at 20°c and 35°c. et al. 2013). except for orange peels, biogas of a higher methane concentration is produced at 35°c compared to 20°c. the effect is strong for potato and tomato and weak for zucchini. moreover the maxima are achieved at shorter times when the temperature is higher and the rate of degredation is faster. for orange peels, biogas of a lower methane concentration is produced at 35°c as compared to 20°c because d-limonene becomes unstable at low temperatures. since d-limonene negatively affects the bacterial activity, oxidation causes its concentration to drop, enhancing the activity of the bacteria and leading to higher biogas production at 20°c compared to 35°c (martin et al. 2010; paramita et al. 2010). the effect of adding sludge to the vegetable waste on the concentration of methane in the biogas is depicted in fig. 4a–d. the presence of sludge increases the concentration of methane gas produced and its maximum is achieved faster. this is obvious since sludge enhances the rate of methane production a result similar to that obtained by (liu et al. 2009; satayanarayana et al. 2008; gelegenis et al. 2007). a comparison between the effect of adding sludge and chicken dung to zucchini is depicted in fig. 4a. biogas with a higher methane concentration is produced when chicken dung is added; however, chicken dung and sludge achieve maximum methane levels at the same time. a careful inspection of the figures shows that after the methane gas concentration drops from its maximum value, it reaches a minimum value and again increases. a possible explanation is that methanogenic bacteria transforms the volatile fatty acids (vfa) formed by the acidogenesis bacteria into methane gas. accordingly, the vfa concentration drops and so does the methane gas level. this drives the activity of acidogenesis bacteria to produce vfa, producing an acidic environment (low ph). as a consequence of this drop, the activity of acidogenesis bacteria drops and metthanogenic bacteria utilize vfa to produce methane, causing an increase in its concentration and the cycle repeats. this point is not proven in this work but worth investigation in future research. 4. conclusions in the current study, anaerobic digestion of vegetable waste was carried out in batch digesters. the concentration of methane in the biogas produced from zucchini, potatoes, tomatoes, rice, and orange peels was measured. the effect of adding chicken dung and sludge to the above named vegetables on the concentration of methane was investigated. the work was carried out at 20°c and 35°c. the concentration of methane excreted from the vegetables studied is as follows: potato>rice>tomato>zucchini>orange peel. the concentration of methane achieves a maximum, and that maximum is achieved faster in tandem with the rate of digestion. sludge, chicken dung, and temperature improved the concentration of methane in the biogas, and the maximum value is reached earlier. the produced biogas has a higher methane concentration when chicken dung is used; however, a maximum methane concentration is reached at nearly the same time whether chicken dung or sludge is used. h. mousa and m. silwadi 37 table 2. chemical composition per 100 grams of vegetables studieda. component rice potatoes zucchini orange peels tomatoes protein (g) 7.1 2.0 1.2 1.5 0.9 fat (g) 0.66 0.09 0.2 0.2 0.2 carbohydrates (g) 80 17 2.9 25 4.7 fiber (g) 1.3 2.2 1.1 10.6 1.2 sugar (g) 0.12 0.78 2.6 calcium (mg) 28 12 21.4 161 13 iron (mg) 0.8 0.78 0.35 0.8 0.5 magnesium (mg) 25 23 17 22 11 phosphorus (mg) 115 57 51.4 21 24 potassium (mg) 115 421 262 212 237 sodium (mg) 5 6 10 3 3 zinc (mg) 1.09 0.29 0.29 0.25 copper (mg) 0.22 0.11 .051 manganese (mg) 1.09 0.15 0.175 0.114 selenium (μg) 15.1 0.3 0.58 vitamin c (mg) 0 19.7 17 136 14 thiamin (mg) 0.07 0.08 0.12 0.037 riboflavin (mg) 0.05 0.03 0.09 niacin (mg) 1.6 1.05 0.487 0.9 0.594 pantothenic acid (mg) 1.01 0.30 0.155 vitamin b6 (mg) 0.16 0.30 0.218 0.176 0.08 folate total (μg) 8 16 48 vitamin a (iu) 0 2 420 42 vitamin e, alpha-tocopherol (mg) 0.11 0.01 0.25 0.54 vitamin k1 (μg) 0.1 1.9 7.9 beta-carotene (μg) 0 1 449 lutein + zeaxanthin (μg) 0 8 123 saturated fatty acids (g) 0.18 0.03 0 0.024 monounsaturated fatty acids (g) 0.21 0.00 0 0.036 polyunsaturated fatty acids (g) 0.18 0.04 0.1 0.04 a united states department of agriculture (usda), national nutrient data base for standards reference release 27, http://ndb.nal.usda.gov/ndb/search. variation in methane concentration produced from anaerobically digested vegetables 40 figure 4-a. methane gas concentration in the biogas produced from zucchini in the presence of sludge, and chicken dung as an inoculum 35°c. figure 4-b. methane gas concentration in the biogas produced from potatoes with and without sludge (t=35°c). figure 4-c. methane gas concentration in the biogas produced from orange peels with and without sludge (t = 35°c). figure 4d. methane gas concentration in the biogas produced from rice with and without sludge (t = 35°c). 38 h. mousa and m. silwadi 39 references american public health association (1995), standard methods for the examination of water and waste water. 19th ed. edited by american public health association, washington, dc. babaee a, shayegan j (2010), biogas production: current state and perspectives. appl. microbiol biotechnology 85: 849–860. beam r (2011), enhanced biogas production through optimization of anaerobic digestion of sewage sludge. phd thesis, the university of alabama, department of chemical and biological engineering. berlian s, sukandar a, seno d (2013), biogas recovery from anaerobic digestion process of mixed fruit -vegetable wastes. energy procedia, 32: 176-182. bernd l (2006), kinetic study of thermophilic anaerobic digestion of solid wastes from potato processing. biomass and bioenergy 30: 892–896. bouallagui h, haouari o, touhami y, cheikh r, marouani l, hamdi m (2004), effect of temperature on the performance of an anaerobic tubular reactor treating fruit and vegetable waste. process biochemistry 39: 2143–2148. bouallagui h, touhami y, cheikh r, hamdi m (2005), bioreactor performance in anaerobic digestion of fruit and vegetable wastes. process biochemistry 40: 989–995. cheng f, kanokwan b, irini a (2011), biogas production from potato-juice, a by-product from potato-starch processing, in up flow anaerobic sludge blanket (uasb) and expanded granular sludge bed (egsb) reactors. bioresource technology 102: 5734–5741. chua k, cheah w, tan c, leong y (2013), harvesting biogas from wastewater sludge and food waste. 4th international conference on energy and environment, earth and environmental science 16: 012118. demirbas a (2006), biogas production from the organic fraction of municipal solid waste. energy sources, part a, 28: 1127–1134. gelegenis j, georgakakis d, angelidaki i, christopoulou n, goumenaki m (2007), optimization of biogas production from olive-oil mill wastewater by codigesting with diluted poultry-manure. applied energy 84: 646–663. kryvoruchko v, machmuller a, bodiroza v, amon b, amon t (2009), anaerobic digestion of by-products of sugar beet and starch potato processing. biomass and bio energy 33: 620– 627. liu y, miller s, safferman s (2009), screening codigestion of food waste water with manure for biogas production. biofuels, bioproducts and biorefining 3: 11–19. ma j, van wambeke m, carballa m, verstraete w (2008), improvement of the anaerobic treatment of potato processing wastewater in a uasb reactor by co-digestion with glycerol. biotechnology letters 30: 861–867. machnicka a, suschka j, grübel k (2008), the importance of potassium and magnesium ions in biological phosphorus removal from wastewater. water environmental research 80: 617–623. martín a, siles a, el bari h, chica f, martín a (2010), orange peel: organic waste or energetic resource.ramiran.net/ramiran2010/docs/ramir an 2010_ 0083_ final.pdf (accessed 15.02.2015). mayer f (1998), potato pulp: properties, physical modification and applications. polymer degradationa and stability 59: 231–235. nallathambi v (2004), biochemical methane potential of fruits and vegetable solid waste feedstocks. biomass and bioenergy 26: 389–399. nasir im, ghazi t, omar r (2012), production of biogas from solid organic wastes through anaerobic digestion: a review. applied microbiology and biotechnology 95: 321–329. nguyen h (2013), biogas production from solvent pretreated orange peel. m.sc. thesis, department of chemical and biological engineering, chalmers university of technology, goteborg, sweden. nirmala b, deepak s, sunil k (1996), biomethanation of banana peel and pineapple waste. bioresource technology 58: 73–76. ofoefule a, onyeoziri m, uzodinma o (2011), comparative study of biogas production from chemically-treated powdered and un-powdered rice husks. journal of environmental chemistry and ecotoxicology 3(4): 75–79. okeha o, onwosib c, odiboa f (2014), biogas production from rice husks generated from various rice mills in ebonyi state. nigeria renewable energy 62: 204–208. paramita v, iida k, yoshii h, furuta t (2010), effect of feed liquid temperature on the structural morphologies of d-limonene microencapsulated powder and its preservation. journal of food science 75(1): 39–45. parawira w, murto m, read j, mattiasson b (2005), profile of hydrolases and biogas production during two-stage mesophilic anaerobic digestion variation in methane concentration produced from anaerobically digested vegetables 40 of solid potato waste. process biochemistry 40: 2945–2952. parawira w, murto m, zvauya r, mattiasson b (2004), anaerobic batch digestion of solid potato waste alone and in combination with sugar beet leaves. renewable energy 29: 1811–1823. periyasamy e, nagarajan p (2012), biogas production from co-digestion of orange peel waste and jatropha de-oiled cake in an anaerobic batch reactor. african journal of biotechnology 11(14): 3339–3345. saev m, koumanova b, simeonov i (2009), anaerobic co-digestion of wasted tomatoes and cattle dung for biogas production. journal of the university of chemical technology and metallurgy 44: 55–60. salter a (2007), crops for biogas production; yields, suitability and energy balances iea bioenergy, 15th european biomass conference, berlin. sarada r, joseph r (1994), studies on factors influencing methane production from tomatoprocessing waste. bioresource technology 47: 55–57. satyanarayana s, murkutea p, ramakant a (2008), biogas production enhancement by brassica compestries amendment in cattle dung digesters. biomass and bioenergy 32: 210–215. singh a, kuila a, sunita a, moumita b, rintu b (2012), utilization of vegetable wastes for bioenergy generation. agricultural research 1(3): 213–222. song z, yang g, guo y, zhang t (2012), comparison of two chemical pretreatments of rice straw for biogas production by anaerobic digestion bioresources 7: 3223–3236. tekin a, dalgic a (2000), biogas production from olive pomace, resources. conservation and recycling 30: 301–313. velmurugan b, ramanujam r (2011), anaerobic digestion of vegetable wastes for biogas production in a fed-batch reactor. international journal of emerging sciences 1(3): 478–486. weiland p (2011), effect of organic loading rates (olr) on production of methane from anaerobic digestion of vegetables waste. world renewable energy congress, sweden. wikandari r, nguyen h, millati r, niklasson c, taherzadeh m (2014), improvement of biogas production from orange peel waste by leaching of limonene. biomed research international, article id 494182, 1–6. ye j, li d, sun y, wang g, yuan z, zhen f, wang y (2013), improved biogas production from rice straw by co-digestion with kitchen waste and pig manure. waste management 33: 2653–2658. yogita s, sumit b, radhika s (2012), evaluation of biogas production from solid waste using pretreatment method in anaerobic condition. international journal of emerging sciences 2(3): 405–414. zhua b, gikasb p, zhang r, lord j, jenkins b, li x (2009), characteristics and biogas production potential of municipal solid wastes pretreated with a rotary drum reactor. bioresource technology 100: 1122–1129. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 2.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 15-21 dynamics of a carriage system of comparator for calibrating the line standards of length a. kilikevičius, v. vekteris and v. mokšin* department of mechanical engineering, vilnius gediminas technical university, j. basanavičiaus str. 28, 03224 vilnius, lithuania received 14 april 2014; accepted 10 march 2015 abstract: this paper presents research dynamic properties of a calibration comparator which is used to calibrate high precision line standards of length. for this purpose, multi-body dynamic and mathematical models of a carriage system of the comparator were presented. calculated amplitude-frequency responses and modes of oscillations allowed a determination of resonant frequencies of the system. keywords: vibration amplitude, resonant frequency, carriage, line standard of length, calibration comparator. .. . . 1. introduction length calibration is one of the most complex, important, and fundamental issues facing science and high-tech industries. the meter is the most commonly used length measurement unit and the need for reliable, precise measurements to ensure unity of measurements worldwide constantly grows. recently, linear displacement transducers increasingly have been used to precisely measure displacements, particularly in automated industrial equipment (sawabe et al. 2004; thalmann 1997). at present, line standards technology is well developed, and precise scales and gratings have replaced interferometers in displacement measuring systems (demarest 1998; jakstas et al. 2014; kasparaitis et al. 2012; kasparaitis et al. 2015). *corresponding author’s email: vadim.moksin@vgtu.it the relevance of the accurate and fast detection of divisions of line standards of length is primarily determined by the constantly increasing requirements of precision scale calibration productivity (bipm 2003; raeng, 2004; swyt 2001). the new generation of precision scale production and calibration systems must meet significantly higher accuracy and efficiency requirements (swyt 2001; webb and jones 2004; bosse and flugge 2002; kasparaitis et al. 2012). to fulfill these requirements, accuracy parameters of precision scales should be metrologically linked to those of higher etalons already in place during the manufacturing process. these needs can be satisfied by creating qualitatively new systems which reflect modern scientific and technological achievements and optimally meet the specific calibration requirements, as well as by improving existing calibration systems in accordance with the fundamental principles of precision engineering (nakazava 1994). a. kilikevičius, v. vekteris and v. mokšin 16 in designing precision length calibration systems which fulfill specific requirements, it is necessary to solve a number of new scientific and technical problems, in particular with regard to the calibration uncertainty. modern scientific and technical achievements allow length measurements at a nano-level; therefore, a qualitative and quantitative description of the determinants of the precision calibration process requires complex mathematical models and the use of numerical methods. this is especially true for an evaluation of the influence of thermal and contact deformations on calibration accuracy, the compensation of geometrical and optical dynamic errors, the measurement signal processing, and so on. it is known that two scale calibration techniques can be used in calibration comparators: static and dynamic. if the former method is used, the calibrating element usually is scanned many times; therefore, it is easy to accumulate measurement data and reduce the influence of random noise on calibration uncertainty. however, during the line scanning process, the scale position in respect to the microscope must be fixed. therefore, the static method is potentially more accurate, but slower than the dynamic method. it is used in high precision calibration systems such as the nist (usa), metas (switzerland), ptb (germany) comparators, as well as in precise chip patterns comparators (beers and penzes 1999; metas 2000). this method provides an edge detection accuracy of less than 1 nm. recent progress in measurement technology allows precise calibration under dynamic conditions. in such a case, calibration is performed during the motion of a microscope or scale. the dynamic calibration method benefits are especially noticeable if long standards are calibrated to greater than 1 meter. the advantages of the latter method can be presented as follows: the calibration process is much faster (productivity increases); the process lends both a simple comparator’s design and control, because it is not necessary to accurately capture the position of scale with respect to the microscope; the calibration accuracy is less influenced by a change of environmental conditions due to the reduced duration of the calibration procedure. despite the mentioned advantages, the dynamic regime of calibration can lead to dynamic calibration errors due to vibrations originating in the calibration system. dynamic effects become extremely important because vibrations directly influence the reading accuracy. new demanding and contradictory requirements of high-speed and accurate precision line scale calibrations calls for an in-depth analysis of the noise and vibrations associated with drivers, dampers, and the moving parts of the calibration system (jakstas et al. 2011). this paper analyzes the dynamics of line standards in a length calibration comparator’s carriage system in which a line detection system and interferometer mirrors were installed. carriage vibration analysis allowed the establishment of harmful resonant frequency levels and an evaluation of the system's response to external forces. 2. research objectives the length calibration comparator (kasparaitis et al. 2008; kaušinis et al. 2009; kilikevičius et al. 2009; kilikevičius and sabaitis 2008) consists of nine main parts: a laser interferometer, gauges for measuring environmental parameters, a microscope with a ccd camera, a driver system, a controller, a data accumulating and processing computer, a compensation system, granite guides, and a carriage system consisting of the force and precision carriages [fig. 2]. the base part of the comparator [fig. 1] is massive—four meters in length with a granite frame with guides. the frame sits on four pneumatic supports, which dampen high frequency vibrations. carriages are placed on rigidly mounted aerostatic supports and preloaded by means of springy supports mounted from the opposite sides. it enables an adjustment of required clearances in aerostatic supports and an increase in the rigidity of the system. the carriage system moves along the frame guides. the carriage system is pulled along the guides by means of the program-controlled friction driver [fig. 1]. the construction of the carriage and the driver system enables the elimination of the influence of driver errors on the accuracy of the longitudinal motion. to this aim, the carriage system is composed of the force and precision carriages [fig. 2]. dynamics of a carriage system of comparator for calibrating the line standards of length 17 figure 1. scheme of the comparator. figure 2. the force carriage of the comparator. the driver is connected to the force carriage; however, all measuring systems used to measure carriage motion and detect the lines on the scale are connected to the precision carriage. the force and precision carriages are connected by a special force transmitting element. this element is designed in such a way that the direction of the force transmitted by it coincides with the direction of the motion of the carriages along guides, where forces in the transverse direction are not applied. 3. dynamic modeling of the carriage system equations of motion can be derived by dividing the comparator model into two subsystems: the force carriage subsystem with links to the granite frame (o1) and the precision carriage subsystem with links to the granite frame and the force carriage (o2) [fig. 3]. the integration of subsystems into the whole carriage system is done by means of auxiliary links expressed using the auxiliary (surplus) coordinates. a. kilikevičius, v. vekteris and v. mokšin 18 a) b) c) figure 3. a dynamic model of the carriage system of the comparator. the views in fig. 3 are: a) xz plane, b) yx plane, and c) yz plane. o1 and o2 are the origins of the coordinate systems of separate subsystems of the model which coincide with mass centers of corresponding parts m1 and m2. i1j, i2j, are moments of inertia, where j corresponds to angles α, β, and γ. kn are coefficients of rigidity in the direction of linear axes (n = 1, 2, ..., 6), while cn are coefficients of rigidity in the direction of rotational axes (n = 1, 2, ..., 6), hn are damping coefficients in the direction of linear axes (n =1, 2, ..., 6), and rn are damping coefficients in the direction of rotational axes (n = 1, 2, ..., 6). η expresses the kinematical excitation function, or oscillations of the foundation. the comparator is complex dynamic system with many degrees of freedom, is elastic and dissipative [fig. 3]. in order to derive differential equations describing the behavior of such objects, it is expedient to use the second order lagrange equation (augustaitis 2000) as appears in eqn. 1, )(tq qd d qd d qd dt qd dt dt d (1) where t, π and φ are the kinetic and potential energies; the dissipative function of the system, qq , , are the vectors of displacements and velocities, and tq is the vector of external excitation forces. the kinetic and potential energy of individual subsystems—force carriage o1 and precision carriage o2—can be expressed as follows: 2222222 iiiiiiiiiii iiizyxmt , i = 1, 2 (2) 2 013 2 011 2 012 2 013 2 012 2 01112 cccyyk zzkxxk (3) dynamics of a carriage system of comparator for calibrating the line standards of length 19 2 026 2 024 2 025 2 0326 2 025 2 02422 ccc yyykzzkxxk (4) where, 000000 ,,,,, zyx are the amplitudes of kinematical excitations acting from the foundation in the direction of the corresponding coordinate axis. dissipative functions can be expressed as follows: 2 013 2 011 2 012 2 013 2 012 2 01112 rrr yyhzzhxxh (5) 2 026 2 024 2 025 2 0326 2 025 2 02422 rrr yyyhzzhxxh (6) the mathematical model of the whole carriage system is composed of the system of second order differential equations (augustaitis 2000), (7) where, [a], [b], [c] are the matrices of inertia, damping and rigidity, respectively; qqq ,, are the vectors of displacements, velocities and accelerations, and tq is the vector of external forces, including forces due to kinematical excitation tqtqtq qextern 0. . matrices [a], [b] and [c], vectors qqq ,, and can be written as follows: 1313213113 1322221 311211 aaa aaa aaa a 1313213113 1322221 1311211 bbb bbb bbb b 1313213113 1322221 1311211 ddd ddd ddd c (8) 13 2 1 q q q q 13 2 1 q q q q 13 2 1 q q q q (9) )10(0121213 llyy the data required for modeling the rigidity of the aerostatic bearings of the carriages, the driver system connecting the force carriage with the frame, and the elements connecting the force carriage with the precision carriage, as well as amplitudes of excitation of the frame rising due to vibrations of the foundation was determined experimentally. for this purpose, vibrations excited by the impact hammer at the characteristic points of both carriages were analyzed, and vibrations of the frame supports were measured and analyzed. the damping coefficients were evaluated as a result of the viscous friction for the entire model. additionally, for the connection of the force and precision carriages, it was assumed that the friction force is proportional to the displacement. simulation and analysis of the model were carried out using simulink and matlab software (the mathworks, inc., natick, massachusetts, usa). the y axis is extremely important for the accuracy of the comparator because this axis coincides with the measuring direction during the scale calibration process. figure 4 illustrates a system response to the most important excitation—external force applied to the force carriage. two resonance zones can be distinguished in fig. 4: a 20–25 and 280–310 hz frequency interval. modes of vibrations at resonant frequencies are presented in fig. 5. it can be seen that oscillations along coordinates y1 and y2 are counteracted at these frequencies. 4. conclusions 1. a created dynamic model of the carriage system of the comparator allowed an analysis of its dynamic characteristics as well as their influence on calibration accuracy. 2. two harmful resonant frequencies of the system—24 hz and 296 hz—were determined. 3. vibrations of the precision and force carriages at resonant frequencies act in opposite directions. a. kilikevičius, v. vekteris and v. mokšin 20 figure 4. amplitude-frequency response to the excitation applied to the force carriage in the y2 direction. line 1 indicates a precision carriage along the coordinate y1, while line 2 indicates a force carriage along the coordinate y2, and line 3 shows the granite frame along the coordinate y0. figure 5. modes of vibrations of the carriage system at resonant frequencies of (a) 24 hz and (b) 296 hz. references augustaitis vk (2000), fundamentals of mechanical vibrations. vilnius, žiburio leidykla. beers js, penzes wb (1999), the nist length scale interferometer. journal of research of the national institute of standards and technology 104(3): 225– 252. bosse h, flugge j (2002), requirements and recent developments in high precision length metrology. nw-verlag. demarest f (1998), high-resolution, high-speed, low data age uncertainty, heterodyne displacement measuring interferometer electronics. measurement science and technology 9(7): 1024–1030. evolving needs for metrology in trade, industry and society and the role of the bipm (2003), bureau international des poids et measures (bipm). 20 dynamics of a carriage system of comparator for calibrating the line standards of length 21 jakstas a, kausinis s, barauskas r, kasparaitis a, barakauskas a (2011), investigation of dynamicsinduced errors of long line scale calibration systems. measurement 44(5): 976–987. jakstas a, kausinis s, barauskas r, kasparaitis a, barakauskas a (2014), thermal error analysis in precision length measurements. measurement 51: 133–146. kasparaitis a, kilikevičius a, barakauskas a, mokšin v (2012), experimental research of vibrations of angle measurement comparator. journal of vibro engineering 14(4): 1760–65. kasparaitis a, kilikevičius a, vėžys j, prokopovič v, makarskas v (2015), dynamic research on precision angle measurement comparator. solid state phenomena 220–221: 554–557. kasparaitis a, vekteris v, kilikevičius a (2008), line scale comparator carriage vibrations during dynamic calibration. journal of vibro engineering 10(3): 347–354. kaušinis s, kasparaitis a, barakauskas a, barauskas r, jakštas a, kilikevičius a (2009), line scale calibration in non-ideal measurement situation. solid state phenomena 147–149: 682– 685. kilikevičius a, sabaitis d (2009), stability of length measurement comparator’s carriage under artificial excitation. ultragarsas 63(1): 33–37. kilikevičius a, vekteris v, kasparaitis a (2009), precision measurement of the carriage vibrations at the different speeds. solid state phenomena 147–149: 690–695. nakazava h (1994), principles of precision engineering. oxford, oxford university press. nano4: 1d gratings. (2000), final report, swiss federal office of metrology (metas). nanoscience and nanotechnologies: opportunities and uncertainties (2004), the royal society and the royal academy of engineering (raeng). sawabe m, maeda f, yamaryo y, simomura t, saruki y, kubo t, sakai h, aoyagi s (2004), a new vacuum interferometric comparator for calibrating the fine linear encoders and scales. precision engineering 28(3): 320–328. swyt da (2001), length and dimensional measurements at nist. journal of research of the national institute of standards and technology 106(1): 1–23. thalmann r (1997), a new high precision length measuring machine, progress in precision engineering and nanotechnology. proceedings of the 9-ipes/ume4, braunschweig, germany 1: 112–115. webb ce, jones, jdc (2004), handbook of laser technology and applications. new york, taylor and francis. d:\desktop\paper_6.xps tjer 2013, vol. 10, no. 2, 69-77 solution methods for the periodic petrol station replenishment problem c triki mechanical and industrial engineering department, college of engineering, sultan qaboos university, po box 33, pc 123, al-khoud, muscat sultanate of oman received 19 june 2012; accepted 20 february 2013 abstract: abstract: in this paper we introduce the periodic petrol station replenishment problem (ppsrp) over a t-day planning horizon and describe four heuristic methods for its solution. even though all the proposed heuristics belong to the common partitioning-then-routing paradigm, they differ in assigning the stations to each day of the horizon. the resulting daily routing problems are then solved exactly until achieving optimalization. moreover, an improvement procedure is also developed with the aim of ensuring a better quality solution. our heuristics are tested and compared in two real-life cases, and our computational results show encouraging improvements with respect to a human planning solution keywords: petrol delivery, periodic constraints, vehicle routing problem __________________________________________ *corresponding author’s e-mail: chefi@squ.edu.om 1. introduction the problem of planning petrol delivery to distribution stations is well-recognized in the operations research literature under the subject name petrol station replenishment problem (psrp). it consists of simultaneously making several decisions, including determining the minimum number of trucks required, assigning the stations to the available trucks, defining a feasible route for each truck, settling whether drivers' work overtime is required, etc. the objective to be achieved is usually defined as the minimization of the traveled distance by the trucks to serve all of the distribution stations. most of the studies presented in the literature solve this problem over a time period of one single day, ignoring the possibility that solving the psrp over a tday planning horizon may yield delivery savings for the company (cornillier et al. 2012). however, most of the psrps are characterized by the fact that the stations should not be served at each day of the t-day planning horizon but rather through a specified number of times. this would introduce a further characterization of the psrp; namely, the periodic nature of the application. specifically, each station (i) must be 70 c triki served (ri) on certain days within the time horizon, and these service days are assigned to i by selecting one of the feasible combinations of ri service days with the objective of minimizing the total distance traveled by the delivery trucks. in this paper, we define a new variant of the above described problem that we will denote as the periodic petrol station replenishment problem (ppsrp). we propose a novel three-phase heuristic method for its solution. our approaches consist of first assigning the service combinations by solving different integer programming models, then defining the routing for each day and, finally, using a local improvement technique to further reduce the delivery distance. a real-life application was chosen for testing and assessing the performance of our solution methods. for some of the proposed heuristics, satisfactory results were obtained with respect to the solution currently adopted by the company. the paper is organized as follows: the next section, section 2, will be devoted to surveying the related works proposed in the literature and highlighting the original contribution of this paper. section 3 will formally define the ppsrp and describe the periodicity constraints to be implemented. section 4 will be dedicated to the development of our heuristic approaches to solve the ppsrp. the computational performance of the heuristics will be discussed in section 5 and, finally, some concluding remarks and future developments will be drawn in section 6. 2. literature review the problem introduced in this paper finds its root in two different classes of distribution problems that are broadly studied in the literature; namely, the psrp and the periodic vehicle routing problem (vrp). from one perspective, the ppsrp is a natural extension of the psrp as it prolongs the planning horizon to cover a t-day period. the 1-day variant of the problem has been subjected to extensive research activity, and several mathematical models and numerical methods have been proposed for its solution. a limited list of these works includes a collection of papers by (cornillier et al. (2008a, 2008b, 2009, 2012) and the contributions of (brown and graves 1981; brown et al. 1987; ben abdelaziz et al. 2002; rizzoli et al. 2003; ng et al. 2008; surjandari et al. 2011 and boctor et al. 2012). only a few papers have considered the multi-period aspect in the context of psrp, but none of them included explicitly the periodic nature of the problem. taqa allah et al. (2000) specifically have extended the planning horizon of the psrp to the multi-period case and have proposed several heuristic methods for the solution of the so-called mpsrp. in their paper, they examined a single depot and an unlimited homogeneous truck fleet. malepart et al. (2003) also tackled the multi-visit petrol station replenishment problem (mpsrp) in a real-life context and included a case in which the distribution company would decide for some stations the quantity to be delivered at each visit. cornillier et al. (2008b) penned the most recent work, which considered a multi-period horizon and proposed a heuristic that maximizes the profit, which is calculated as the difference between the delivery revenue and the sum of regular and overtime drivers' costs. thus, to the best of our knowledge this paper represents a first tentative step in tackling a psrp which is characterized by the periodic service requirements of the petrol stations. on the other hand, the ppsrp can find its roots in the literature of the periodic vrp that attracted the interest of many researchers over the previous two decades (gaudioso and paletta 1992; chao et al. 1995; cordeau et al. 1997). the two problems have many similarities, but the main difference between the two consists of the fact that the ppsrp should include additional constraints (for example, those related to the drivers' shifts' lengths). 3. problem description the objective of the ppsrp consists of finding a minimum distance routing to serve all the stations when considering an extended planning horizon of tdays. 71 solution methods for the periodic petrol station replenishment problem the resulting network model can be expressed in terms of combinatorial optimization having two sets of decision variables. the first set defines the routing solution and the corresponding truck assignment, (ie. it will give clear insight on which truck will serve which stations in which day in order to satisfy all the demand). the second set of variables will choose for each station the best periodicity combination that will minimize the overall delivery distance. the integer model with all its constraints is cumbersome and represents only slight differences with respect to the mathematical model reported (cornillier et al. 2008b). for this reason and for brevity we avoid reporting the whole model in this paper and we focus in the sequel on the innovative heuristic methods developed for its solution. 4. heuristic algorithms since ppsrp instances are generally characterized by a high number of variables and constraints due mainly to the extended time horizon and also to the large number of stations to be served, we have chosen to solve the problem heuristically. we propose here four different heuristics all based on the idea of splitting the task of assigning the periodicity combinations to the stations from that of performing the routing rather than dealing with both the tasks simultaneously (fig. 2). the routing task then is carried out by solving the resulting vrps-one for each day of the planning horizon. finally, all of the heuristics are concluded by running an improvement procedure that is common to all four approaches. 4.1 combination assignment procedure the procedure of assigning feasible combinations to the stations is carried out by solving exactly, using any general purpose software, one of the integer programming models proposed below. the notation to be used through this section is the following: figure 1. an example of a region to be served (left) and its corresponding connected network (right). the stations correspond to the nodes of the graph and the roads are the undirected edges table 1. periodic sequence combinations for t = 6 72 c triki 4.1.1. heuristic 1: minimum daily demand the idea here is to minimize the total demand to be served on each single day of the planning horizon. in this way, the workload will be as balanced as possible and, consequently, the number of vehicles will be minimized on a daily basis. while this procedure has the advantage of being easy to understand and implement, it has the disadvantage of not taking into account the geographical position of the stations. the proposed integer program has yir as the decision variables that define the periodicity combination to be assigned to each station, as follows: yir is 1 if combination r is assigned to station i, 0 otherwise. moreover, by denoting by dt the total quantity of petrol to be served on day t, the mathematical model can be expressed as: (1) (2) (3) (4) the objective function (1) minimizes the highest quantity of petrol to be served over all the days of the planning horizon. it is expressed here as a min-max objective function but it can be easily transformed to a standard minimization problem by using conventional operations research techniques. constraints (2) force the model to choose only one periodic combination among all the feasible combinations. constraints (3) restrict the daily quantity (dt) to be delivered to match at least the total demand to be served on day t. finally, constraints (4) impose the domain of the decision variables. 4.1.2. heuristic 2: minimum daily stations the idea of this heuristic is to minimize the per day total number of stations to be served. even in this case, the attempt aimed at balancing the workload was expressed in terms of the number of stations and, consequently, at minimizing the number of vehicles to be used on a daily basis. this heuristics shares the same advantage as its predecessor of being easy to understand and implement and, moreover, ensures a balanced workload for the trucks over the t days of the horizon. both the heuristics also share the same disadvantage of ignoring the geographical position of the stations. by using the same binary variables yir as defined previously and by introducing a new variable wnt defining the number of stations to be visited during day t, the mathematical model is expressed as: (5) (6) (7) (8) (9) the objective function (5) makes use of the new variable wnt in order to minimize the highest number of stations to be served over all the days of the planning horizon. constraints (6) assigns as before only one periodic combination to each station. constraint (7) imposes logic relations between the new variable wnt and the assignment variables yir on a daily basis, and the domain of the decision variables is defined through constraints (8) and (9). 4.1.3. heuristic 3: depot minimum distance the aim of this heuristic is to solve a relaxation of a pvrp-like process in order to minimize the total distance between the depot (node 0) and all the stations to be served. even though the resulting model is a simplified version of the pvrp, it will maintain some of the complexity but on the other hand it has the advantage of making use of more information since it takes into consideration the geographical position of the stations with respect to the depot. the integer model has two sets of variables. besides, the binary variables yir as previously used we need to introduce also the following binary variables: x0it is 1 if station i is assigned to the depot (node 0) on day t, and 0 otherwise. n1,...,i1y 73 solution methods for the periodic petrol station replenishment problem the optimization model can be summarized as follows: (10) (11) (12) (13) (14) (15) the objective function (10) minimizes the total distance of all the stations with respect to the depot over the days of the planning horizon. constraints (11) has the same role as in the previous models, whereas constraint (12) imposes logic relations between the two sets of variables x0it and yir. the inequalities of (13) are capacity constraints that ensure, for each day t, the respect of the trucks' capacity and, finally, the integrity of the decision variables that are defined by constraints (14) and (15). it should be noted that this model is similar to that in (1) through (4). the main difference between the two problems lies in the objective function (1), which becomes (10). 4.1.4. heuristic 4: minimum distance clusters this heuristic is an extension of the previous one, allowing the minimization of the total distance of the stations with respect to virtual centers, with one for each day of the horizon, rather than considering only the depot. this clustering strategy allows better use of the information about the distances between stations and, hopefully, generates more compact sets of stations from a geographical viewpoint. the obvious disadvantage is the complexity of the resulting model involving a large number of binary variables. for simplicity, the virtual centers are chosen here among a set of specific positions that coincide with some of the stations, so that the topology of the network remains unaffected. the integer model makes use of the variables yir and, moreover, introduces the following two sets of binary variables: xijt is 1 if station i is assigned to the virtual centre j in day t, and 0 otherwise zjt is 1 if virtual centre j is assigned to day t, and 0 otherwise the integer model corresponding to heuristic 4 is as follows: (16) (17) (18) (19) figure 2. the four heuristic solution approaches heuristic 1 heuristic 2 heuristic 3 heuristic 4 minimum daily demand model minimum daily stations model minimum daily distance model minimum daily clusters model 74 c triki (20) (21) (22) (23) similar to the previous heuristic, the objective function (16) minimizes the total distance of all the stations with respect to the t virtual centres (one for each day of the planning horizon). all the other constraints have the same meaning and role as in the previous heuristic except for some slight adjustments due to presence of t virtual centres and the introduction of the new set of variables zjt. we should mention that the notation may be further simplified in the above model by merging the indexes j and t since they represent the same set of horizon days and have a one-to-one relationship but we preferred, for the sake of clarity, to distinguish between the temporal index t and the spatial index j. finally, we should note that the last two heuristics seem to be more challenging computationally since they include the geographical aspect of the problem and, thus, involve additional decision variables with respect to the former two methods. possible higher quality solutions to be yielded by the latter heuristics may be paid by longer computational time when solving large-scale problems. this time-quality trade off may be very useful when solving several instances with different sizes. 4.2. solving the vrps: routing procedure once the hard task of assigning the stations to the days of the planning horizon while keeping in mind the periodicity constraints on each station has been solved, we need to define the routes for each day. the routing procedure takes as input the sets of stations corresponding to each day (ie. the output of any of the models of subsection 4.1) and generates a set of routes for the available trucks. this requires nothing more than solving independently one t standard vrp for each day of the horizon. since each resulting vrp is relatively small with respect to the original ppsrp, we decided to solve these t problems exactly by using a standard vrp model and a general purpose optimization software package. 4.3. getting a better solution: improvement procedure the final step of our approach consisted of trying to improve the quality of the solution by further minimizing the overall distance through the application of a local search technique. theoretically, room for its improvement should be available since the assignment of the stations to the t days has been done heuristically without any guarantee to get the best feasible solution. starting from a set of routes assigned to the available trucks for every day of the planning horizon, the technique employed here is based on switching any two stations assigned to two different service days and checking if any improvement is taking place. while checking the improvement in the total distance at each switching station, we should check also the feasibility of the solution expressed in terms of service periodicity of each station and the capacity of the corresponding trucks. this procedure is easy to understand but requires considerable effort for its implementation, since no standard software is available to achieve this task. 5. computational results this section is dedicated to the application of our heuristic methods to solving a real-life test problem, and to analyzing and comparing their performances. all the models introduced above were implemented using the lingo commercial package, version 8.0 (lindo systems, inc., chicago, ill., usa) as modeling language that has been interfaced with the state-ofthe-art solver cplex, version 12.1 (ibm ilog, gentilly, france) for the solution of the resulting integer programs. on the other hand, the improvement procedure required specific programming skills and, thus, the language java has been chosen for its implementation. a computer with an intel core 2 duo processor 2.53 ghz and 4 gb of ram was used to run all the resulting codes. 5.1. real-life test problem to our knowledge, there are no standard test problems available as a benchmark for the ppsrp, since we believe it is the first time this problem is being tackled. for this reason, we will compare the quality of our heuristics' solutions from one side with respect to the solution currently being adopted by the delivery company, and from the other with respect to the best feasible solution that can be reached by the exact solver. this case study consisted of a regional company located in the south of italy having a central petrol depot from which a heterogeneous fleet of trucks starts daily trips to serve 38 petrol stations located in different places throughout the regional territory. the planning horizon for this problem is a week consisting of 6 work days (t = 6). the weekly demand for petrol varies from one station to another requiring in some cases only one delivery visit; however, most depots require a multi-visit service fractioned into 2 or 3 deliveries per week. the available trucks have differ75 solution methods for the periodic petrol station replenishment problem ent capacities, allowing the model to choose for each route the truck that best fits the model's constraints. concerning the drivers, generally speaking the company has enough drivers available during each weekday except for the sixth working day (ie. the first weekend day) in which the number of drivers is limited because of the unavailability of most of them to work. this limitation has been considered in our models by dropping, whenever is possible, the sequence combinations that involve day 6. for example, when ri = 2, we have considered only the set ci = {(1,4. .2,5)} as feasible combinations and we disregarded the combination (3, 6) in order to take into account the drivers' unavailability. we kept, however, both the sequences ci = {(1,3,5. .2,4,6)} for the stations requiring 3 visits per week. the integrated ppsrp model corresponding to our application resulted in a large-scale problem having 84,740 constraints and 84,701 variables. the state-ofthe-art cplex software package was not able to yield an exact solution for the problem but generated at least a feasible upper bound solution of 13,975 km/week. indeed, after more than 30 cpu hours and a huge number of iterations, the software presented a faulty memory error due to the large number of sub problems generated during the branch-and-bound method. we also collected data for a different aspect of the test problem in order to ascertain information about the performance of our heuristic procedures. the different aspect refers to another working week with different demands for petrol and also different frequencies of periodicity. 5.2. results and discussions we started by assessing the solution quality of each approach (including the exact method) before and after applying the improvement procedure and the corresponding cpu time. the results collected in table 2, show, from one side, the superiority of heuristic 4, both in its solution quality and cpu time, with respect to the other heuristics and also with respect to the feasible upper bound obtained by cplex. from the figure 3. comparison between our solutions and the company’s solution figure 4. savings obtained with respect to company’s solution other side, the positive effect of the improvement procedure on all the methods, except heuristic 4, is clear. this means that the solution generated by heuristic 4 is probably close enough to the optimal solution so that no margin was left for improvement. the second set of experiments deals with the assessment of our approaches with respect to the solution currently adopted by the company. such a solution has not been obtained through the use of any decision support tool but just as the result of a long experience of the company's human operators in managing not only the loading/unloading operations but also the transportation phase. for the data corresponding to the week we took into consideration, the actual distance run by the trucks for serving all 38 stations was 14,820 kms. it is evident from table 3 that all our approaches improve the human operator solution and, particularly, heuristic 4 will reach, after improvement, a relative distance saving of 17.7% with respect to the company's decisions. these same results are also depicted in figs. 3 and 4. finally, for the sake of completeness, we tested our heuristic procedures on a different set of data for the same distribution network. this kind of sensitivity analysis experiment has the aim of checking the robustness of the obtained solutions, and specially verifying the conclusions related to heuristic 4. this new set of data has been purposely collected for another week belonging to a different season and with a different demand pattern with respect to the previous data. the two sets differ only by the amount of petrol to be served to each station and, consequently, by their frequency of service. the results reported in table 4 show that these new results substantially confirm the trend observed in the previous test (except for heuristic 1). more specifically, heuristic 4 once again outperforms all the other approaches. moreover, even in this case the improve76 c triki table 2. solution costs (in terms of distance) and cpu times table 3. comparison between our solutions and the company’s solution table 4. comparison of the solutions for different test data 77 solution methods for the periodic petrol station replenishment problem ment procedure has no effect on the initial solution obtained, confirming our intuitive conclusion that heuristic 4 generates a nearly optimal solution. from another point of view, the results reported in the final column of table 4 confirm the improvement that three of our heuristics can achieve with respect to the solution adopted by the company for that week. 6. conclusions in this paper, we presented and discussed the ppsrp. we proposed four heuristic approaches based on three phases each: the assignment of periodicity combinations to each station, the construction of trucks' delivery routes, and, finally, the routes' improvement phase. the performance of the proposed heuristics was tested on two different instances of a real-life periodic problem. both instances showed the superiority of heuristic 4 in relation to the others, reaching an average improvement of more than 15% in terms of traveled distance, with respect to the solutions actually adopted by the company while keeping the execution time within reasonable limits. future developments in this direction should consist of solving another more challenging real-life problem on a national scale for a petrol distribution company operating in oman. besides the difficulties that will be introduced by the problem's size, the new application also might be characterized by further complexities such as uncertainty related to the drivers' availability and regulatory restrictions on the citizenship of the drivers to accomplish some routes. acknowledgment thanks are due to the anonymous referees for their valuable comments, which improved the quality of the paper. references ben abdelaziz f, roucairol c, bacha c (2002), deliveries of liquid fuels to sndp gas stations using vehicles with multiple compartments. in: systems managment and cybernetics ieee international conference. hammamet, tunisia. boctor f, renaud j, cornillier f (2011), trip packing in petrol stations replenishment. omega 39:86-98. brown gg, ellis cj, graves gw, ronen d (1987), real-time wide area dispatch of mobil tank trucks. interfaces 17:107-120. brown gg, graves gw (1981), real-time dispatch of petroleum tank trucks. management science 27:19-32. chao im, golden bl, wasil e (1995), an improved heuristic for the period vehicle routing problem. networks 26:25-44. christofides n, beasley je (1984), the period routing problem. networks 14:237-256. cordeau jf, gendreau m, laporte g (1997), a tabu search heuristic for periodic and multi-depot vehicle routing problems. networks 30:105-119. cornillier f, boctor f, laporte g, renaud j (2008a), an exact algorithm for the petrol station replenishment problem. journal of the operational research society 59:607-615. cornillier f, boctor f, laporte g, renaud j (2008b), a heuristic for the multi-period petrol station replenishment problem. european journal of operational research 191:295-305. cornillier f, laporte g, boctor f, renaud j (2009), the petrol station replenishment problem with time windows. computers and operations research 36:919-935. cornillier f, boctor f, renaud j (2012), heuristics for the multi-depot petrol station replenishment problem with time windows. european journal of operational research 220:361-369. gaudioso m, paletta g (1992), a heuristic for the periodic vehicle routing problem. transportation science 26:86-92. ibm ilog cplex optimizer. (2011), available at: 01.ibm.com/software/integration/optimization/cpl ex-optimizer/. accessed may 2012. lindo systems inc. (2006), optimization modeling with lingo. available at: www.lindo.com. accessed may 2012. malépart v, boctor f, renaud j, labilois s (2003), nouvelles approches pour l'approvisionnement des stations d'essence. revue française de gestion industrielle 22:15-31. ng wl, leung sh, lam jp, pan sw (2008), petrol delivery tanker assignment and routing: a case study in hong kong. journal of the operational research society 59:1191-1200. paletta g, triki c (2004), solving the asymmetric traveling salesman problem with periodic constraints. networks 44:31-37. rizzoli a, casagrande n, donati a, gambardella l, lepori d, montemanni r, pina p, zaffalon m (2003), planning and optimisation of vehicle routes for fuel oil distribution. in: modsim international congress on modelling and simulation. townsville, australia. surjandari i, rachman a, dianawati f, wibowo rp (2011), petrol delivery assignment with multiproduct, multi-depot, split deliveries and time windows. international journal of modeling and optimisation 1:375-379. taqa allah d, renaud j, boctor ff (2000), le problème d'approvisionnement des stations d'essence. apii-jesa journal européen des systèmes automatisés 34:11-33. c:\users\raoh\desktop\paper 5.xps the journal of engineering research vol. 12, no. 1 (2015) 53-68 applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman k. alzebdeh*a, h.a. bashirb and s.k. al siyabic a department of mechanical and industrial engineering, college of engineering, sultan qaboos university, sultanate of oman. b department of industrial engineering and management, university of sharjah, sharjah, united arab emirates. c oman tank terminal company, muscat, sultanate of oman. received 5 may 2014; accepted 6 january 2015 abstract: cost overruns in construction projects are a problem faced by project managers, engineers, and clients throughout the middle east. globally, several studies in the literature have focused on identifying the causes of these overruns and used statistical methods to rank them according to their impacts. none of these studies have considered the interactions among these factors. this paper examines interpretive structural modelling (ism) as a viable technique for modelling complex interactions among factors responsible for cost overruns in construction projects in the sultanate of oman. in particular, thirteen interrelated factors associated with cost overruns were identified, along with their contextual interrelationships. application of ism leads to organizing these factors in a hierarchical structure which effectively demonstrates their interactions in a simple way. four factors were found to be at the root of cost overruns: instability of the us dollar, changes in governmental regulations, faulty cost estimation, and poor coordination among projects’ parties. taking appropriate actions to minimize the influence of these factors can ultimately lead to better control of future project costs. this study is of value to managers and decision makers because it provides a powerful yet very easy to apply approach for investigating the problem of cost overruns and other similar issues. keywords: project, cost overrun, interpretive structural modelling (ism). ي اإلحصائیةاألسالیبواستخدامالتجاوزاتھذهأسبابالدراساتمنالعدید، ركزت لتأثیراتهاوفقالتصنیفها اتخاذالمشروعأطرافبینالتنسیقوضعف تؤديأنیمكنالعواملھذهتأثیرمنللحدالمناسبةاإلجراءات المماثلةالقضایامنوغیرھا *corresponding authors email: alzebdeh@squ.edu.om k. alzebdeh, h.a. bashir and s.k. al siyabi 54 1. introduction cost overrun is one of the most critical risks in projects. it is defined as an excess of actual cost over budget and is common to all types of projects. for instance, flyvbjerg et al. (2003) sampled 258 transport infrastructure projects in 20 countries and reported that 9 out of 10 projects had cost overruns, with overruns of 50–100% being common. azhar et al. (2008) ascertained that cost overrun is a frequent phenomenon associated with almost all projects in the construction industry. this trend is more severe in developing countries where overruns sometimes exceed the anticipated cost of the project by 100%. several studies have been conducted worldwide to investigate and attempt to minimize the severity of the project cost overrun problem. these studies analysed cost overruns in projects in many different industries including civil and structural engineering construction (mansfield et al. 1994; kaming et al. 1997; frimpong et al. 2003; flyvbjerg et al. 2004; koushki et al. 2005; creedy 2006; azhar et al. 2008; enshassi 2009; kaliba et al. 2009; le-hoai et al. 2008; ali and kamaruzzaman 2010; wang and yuan 2011). one of the common objectives of these studies was to identify the major factors causing project cost overruns in a particular country through opinion surveys of contactors, clients, and consultants. researchers used statistical methods to rank the contributing factors according to their impacts without considering the interactions among these factors. moreover, the studies suggested solutions and remedies without identifying the root causes of the overruns. any attempt to address the cost overruns would be futile unless the root causes are identified because cost overruns are a net result of complex relationships between one or more factors. given the limitations of previous works as highlighted above, this study proposes that interpretive structural modelling (ism) be used to identify and analyse the interaction among the various factors causing project cost overrun, thus developing a hierarchy of factors that will help to identify the root causes of the problem. ism was first developed to identify the interrelationship among elements making up the structure of a system (warfield, 1974; 1979). the ism approach is multifaceted: it is interpretive, as the judgment of the group establishes the relationships among causal elements. it is structurally-oriented because the overall structure is extracted from a complex set of elements. as a modelling technique, it highlights specific relationships and portrays the overall structure in a digraph model. it serves as a communication tool for analysing complex systems and providing an understanding of their interacting elements. ism is a powerful technique that depicts the structure of a complex system through graphics as well as words. it helps one understand the complex relationships among elements within a system by documenting their order and the influence of one element on another. in a broad range of applications, ism has been used extensively since its inception. examples of its use in policy analysis appear in the early literature (hawthorne & sage, 1975; brand et al., 1976; kawamura & christakis, 1976), and recent improvements in computer-based programs have expanded its applications to various fields. for example, using ism, sharma et al. (1995) developed a hierarchy of actions required to meet the objectives of a waste management project in india. mandal and deshmukh (1994) developed an interpretive structural model to analyse vendor selection criteria and demonstrate interrelationships between these criteria and their levels of importance. ism was used by singh et al. (2003) in knowledge management to study the complexity of relationships among elements of a manufacturing system. bolanos et al. (2005) applied ism in order to improve the decision making process among executives working in different functional areas. the ism methodology was implemented by ravi and shankar (2005) to model and analyse the interactions among barrier elements in reverse logistics in automobile supply chains. in a study by farooquie and khan (2007), ism was used to develop the structural relationship among planning factors that contribute to the success of heating, ventilation, and air-conditioning (hvac) projects in india. hasan et al. (2007) applied ism to explore and establish a relationship among various barriers to adopting agile manufacturing. wang et al. (2008) applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 55 applied ism to investigate the obstacles to energy savings in china. bashir (2010) adapted the methodology to remove redundant relationships in an activity on network (aon) project network. ism was used by iyer and sagheer (2010) to identify the nature and degree of the interrelationship between risks during the development phase of public-private partnership projects in the indian road sector. by applying ism, (sahney et al. 2010) proposed a quality framework for administrative staff in india's higher education system. shankar et al. (2011) introduced ism as a tool that can be used to help understand and manage urban planning issues. talib et al. (2011) applied ism to analyze the interaction among the barriers to total quality management implementation. govindan et al. (2012) adopted ism for selecting the best reverse logistics provider. most recently, (mathiyazhagan et al. 2013) developed a structural model of the barriers to implementing green supply chain management in the automobile industry in india. also, (ansari et al. 2013) applied ism to analyze thirteen competing barriers to implement solar power installation in india. the objective of this paper was to identify and rank factors responsible for cost overruns in construction projects. applying ism to identified factors involves modelling interactions among them and determining the root factors that act as the driving cause for initiating the cost overrun problem. before proceeding, it is important to note that in ism terminology, factors are called elements. therefore, these two terms are used interchangeably in the remainder of the paper. 2. methodology of ism several professionals and project managers belonging to substantially large construction projects were contacted. responses on key factors that they encountered during the course of planning and construction phases of their projects were collected. furthermore, their opinions on the interrelations among the causing factors were recorded. to further analyze the interactions among these factors along with the goal of identifying the root causes of cost overruns, ism was selected as a tool for the study; it has been proven effective and simple in other fields, including manufacturing, logistics, and socioeconomics. the technique also acts as a educating tool at the managerial level, facilitating better insight into the system under study. the logical process flow of the ism technique as applied to project cost overruns is presented in fig. 1, which shows that the ism methodology involves five major steps: 1. identifying the factors causing cost overruns 2. developing a structural self-interaction matrix 3. constructing a reachability matrix 4. partitioning the reachability matrix into different levels 5. forming the ism digraph these steps were applied to factors that caused cost overruns in large construction projects carried out in oman. a typical construction project’s performance was deemed unsatisfactory if the project’s final cost of completion exceeded the original budgeted estimate by more than 30%. statistics on average cost escalation in construction projects worldwide are available in the literature (flyvbjerg et al. 2003; abdul azis et al. 2013). 2.1 identifying the factors causing cost overruns the first step in applying ism to the problem of project cost overrun is for experts to identify the causal factors and their interrelationships using methods such as brainstorming. according to (rawlinson 1986), brainstorming is probably the best known of all the techniques available for creative problem solving. it is worthwhile to emphasize that studying variability in interrelationships among factors is beyond the scope of the current study. based on a brainstorming session conducted with professionals involved in projects under study, thirteen critical factors were identified as major causes of cost overruns. these factors are the instability of the us dollar, poor cost estimation, changes in governmental regulations, increased material costs, delays in material delivery, increased wages and equipment costs, poor communication within the project team and among project stakeholders, design changes during the construction phase, high staff turnover, schedule delays, and procurement changes. these factors will be detailed individually below. k. alzebdeh, h.a. bashir and s.k. al siyabi 56 figure 1. flow diagram for developing ism. 2.1.1 instability of the us dollar because oman’s currency is tied to the us dollar at a fixed exchange rate, it is common practice to use the us dollar as the primary currency in construction projects in oman. unfortunately, during the construction phase of the projects under consideration in the current study, the us dollar declined in value against other world currencies causing inflation; this resulted in a high turnover of staff and higher costs for staff wages, materials, and equipment. this can be attributed to the fact that materials and equipment were purchased from europe and other foreign countries. similarly, inflation forced experienced labourers to seek more attractive opportunities in construction and other fields. identifying the factors causing cost overrun establishing contextual relationship (xi,j) between problem variables (i, j) developing a structural selfinteraction matrix (ssim) partitioning the reachability matrix into different levels developing the reachability matrix in its conical form removing transitivity from the digraph replacing variables nodes with relationship statements is there any conceptual inconsistency? representing relationship statement into model to establish the hierarchy of variables constructing a reachability obtaining expert opinion no yes forming ism digraph applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 57 2.1.2 poor cost estimation many studies have identified poor cost estimation as one of the major causes of project cost overruns (mansfield et al. 1994; kaming et al. 1997; le-hoai et al. 2008). the initial cost of some projects under study was estimated using an analogy approach, one of the traditional estimation techniques. however, using only one estimation technique is limited in producing realistic estimates that become the basis of a project plan (bashir 2008). 2.1.3 changes in governmental regulations changes in governmental regulations are common external factors that affect the final cost of projects (frimpong et al. 2003). during the execution phase of the projects considered in the current study, the omani government introduced new customs regulations and procedures that caused changes in procurement and negatively impacted both costs and material delivery. 2.1.4 increased material costs during the current study, the prices of materials increased significantly during the construction phase for several reasons, including declines in the value of the us dollar, the imposition of new customs regulations and procedures, and procurement changes due to significant design changes. 2.1.5 delays in material delivery since most of the construction materials and equipment required for the analyzed project were imported from overseas, their delivery times were negatively impacted by the new customs regulations and procedures introduced by the local government. the materials’ delivery times were also negatively impacted by procurement changes which resulted from significant design changes during the construction phase. 2.1.6 increased wages all the previously reviewed published research (ogunlana and promkuntong 1996; kartam and kartam 2001; mansfield et al. 1994; assaf et al. 1995; ogunlana et al. 1996; kaming et al. 1997; odeh and battaineh 2002; frimpong, oluwoye and crawford 2003; assaf and al-hejji 2006; sambasivan and soon 2007; moura teixeria and pires 2007; le-hoai, lee and lee 2008; azhar, farooqui and ahmed 2008; el-sayegh 2008) have identified the increased wages as one of the risk factors in construction projects. during the execution phase of the projects under consideration in the current study, the industry in oman experienced high staff turnover. therefore, to attract new employees and retain existing ones, wages were substantially increased. in addition, the flourishing construction business contributed to a high demand for workers, which further pushed up their wages. 2.1.7 increased equipment costs controlling costs has become a difficult task for construction equipment users in recent years. along with the general upward trend in construction costs, including both materials and labor, construction equipment prices increased significantly during the current study’s construction phase. costs of mechanical equipment installed as part of the facility also increased by about 10%. again, market volatility driven by the weakness of the us dollar has been a key driving force behind such cost increases. in addition, the equipment costs were affected by procurement changes. 2.1.8 poor communication within the project team proper communication among team members is crucial for the success of any project. however, projects examined in the current study were plagued by miscommunication among team members. this was mainly due to the fact that the project team members had different cultural backgrounds and work experience. the communication problem was compounded by the high staff turnover. rework during the construction stage was one of the consequences of poor communication. any problem with communication can lead to severe misunderstandings and, therefore, delays in the execution of a project (sambasivan and soon, 2007). azhar et al. (2008) ranked the severity of the coordination within the project team as having a medium effect on the total cost of the project in pakistani construction projects. miscommunications between a project team’s members were identified as one of the factors which led the project to time and cost overruns (ogunlana and promkuntong 1996; kartam and kartam 2001; sambasivan soon 2007). k. alzebdeh, h.a. bashir and s.k. al siyabi 58 2.1.9 design changes during the construction phase excessive design changes in the construction phase can cause significant disruption to a project and require changes in scheduling and procurement. in several projects, excessive change orders were issued as a result of poor communication among project stakeholders. change orders were also identified by creedy (2006) as one of the main causes of cost overrun in construction projects. 2.1.10 high staff turnover turnover, the rate at which a company gains and loses employees, can have a negative impact on project scheduling and costs. when a company loses employees, one or both of the following measures are usually taken: the working hours of current staff are extended or new employees are hired. the first measure may increase the stress on the remaining team members and lead to an increase in mistakes. as a result, the amount of rework may increase and the completion deadline may need to be extended. the second measure also causes schedule delays because it takes significant time to make new hires productive on a project as they are required to go through training programs (bashir and thomson 1999). according to (schlesinger and heskett 1991), the estimated cost of employee turnover for a company can reach up to 150% of employees' total remuneration. at a typical project examined during the current study, cost overrun was compounded by the high turnover of project staff. this was mainly due to employee dissatisfaction with their compensation packages in light of the high inflation rates that had affected those regions and countries, including oman, that tie their currencies to the us dollar. 2.1.11 schedule delays schedule delays are one of the most common problems in the construction industry worldwide. for instance, (al-khalil and al-ghafly 1999) reported that a high proportion (about 70%) of projects in saudi arabia experienced schedule delays. another study in saudi arabia carried out by assaf and al-hejji (2006) reported that 70% of experienced time overrun; their study also found that 45 out of 76 projects were delayed. according to (faridi and al-sayegh 2006), 50% of construction projects in the united arab emirates (uae) experience delays. in malaysia, (sambasivan and soon 2007) reported that 17.3% of construction projects experience delays of more than three months. in addition to its negative impact on timely project delivery, project delays can cause cost overruns and disputes between project stakeholders. if not resolved, disputes can lead to arbitration and litigation. the construction projects under study experienced a significant delay due to late material delivery, design changes during the construction phase, high staff turnover, procurement changes and poor communication among project stakeholders. 2.1.12 procurement changes procurement can be defined as the process used to acquire goods or services. it involves identifying the items to be procured, selecting suppliers, and contracting. according to several studies (arditi et al. 1985; mansfield et al. 1994; ogunlana et al. 1996; sweis 2008), procurement is critical to the successful completion of any project. deficient procurement management and/or an excessive number of procurement changes can cause project delays and cost overruns. during the construction phase of the projects in the current study, significant procurement changes had to be made due to new customs regulations and design changes; these, in turn, caused delays in material delivery and scheduling, and increases in material and equipment costs. 2.1.13 poor communication among project stakeholders there is no doubt that proper communication among project stakeholders is crucial for a project’s success. consistent with the findings of (sambasivan and soon 2007), poor communication between the contractors, clients, and consultants for a project causes design changes during the construction phase and, subsequently, schedule delays. applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 59 figure 2. digraph depicting the interrelationships among project cost overruns and causal factors. 2.2 developing the structural self-interaction matrix (ssim) the interrelationships among the fourteen elements (project cost overrun and the above mentioned thirteen causal factors) are represented in the diagram shown in fig. 2, which represents the problem and its causal system. this could be further simplified and rearranged through a systematic methodology calling for converting the digraph into the structural self-interaction matrix (ssim). the utility of the ssim is that it can be used to develop a hierarchical restructuring of the diagraph. to develop this matrix, three symbols are used to denote the direction or influence of the relationship between factors i and j: f means that factor i leads to factor j (a forward relationship) b means that factor j leads to factor i (a backward relationship) n means that factor i and factor j are unrelated (no relationship). using the above-mentioned symbols, the ssim for the project cost overrun is presented in table 1. table 1. structural self-interaction matrix (ssim). factors 14 13 12 11 10 9 8 7 6 5 4 3 2 1 1. instability of the us dollar n n n n f n n f f n f n n 2. poor cost estimation f n n n n n n n n n n n 3. changes in governmental regulations n n f n n n n n n f f 4. increased material costs f n b n n n n n n n 11 1 2 3 4 5 6 7 8 9 10 12 13 14 k. alzebdeh, h.a. bashir and s.k. al siyabi 60 2.3 constructing the reachability matrix after development, the ssim is transformed into a binary matrix—the initial reachability matrix—by substituting f, b, and n with 1 and 0. the rules for the substitution of ones and zeros are as follows: if the (i, j) entry in the ssim is f, then the (i, j) entry in the reachability matrix becomes 1 and the (j, i) entry becomes 0. if the (i, j) entry in the ssim is b, then the (i, j) entry in the reachability matrix becomes 0 and the (j, i) entry becomes 1. if the (i, j) entry in the ssim is n, then the (i, j) entry in the reachability matrix becomes 0 and the (j, i) entry becomes 0. based on these rules, the initial reachability matrix is shown in table 2. the final reachability matrix is then obtained by incorporating the transitivity links. that is, if element x is related to y and y is related to z, then x is necessarily related to z. the final reachability matrix, as well as the driving power and dependence level of each factor, are shown in table 3. the driving power of a particular factor is the total number of factors which it may cause. the dependence level is the total number of factors which may contribute to causing it. for instance, element 1 (instability of the us dollar) has the highest driving power, whereas element 14 (project cost overrun) has the highest dependence level. table 2. initial reachability matrix. 5. delays in material delivery n n b f n n n n n 6. increased wages f n n n b n n n 7. increased equipment costs f n b n n n n 8. poor communication within the project team n n n n b f 9. design changes during the construction phase n b f f n 10. high staff turnover n n n f 11. schedule delays f b b 12. procurement changes n n 13. poor communication among project stakeholders n 14. project cost overruns factors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. instability of the us dollar 1 0 0 1 0 1 1 0 0 1 0 0 0 0 2. poor cost estimation 0 1 0 0 0 0 0 0 0 0 0 0 0 1 3. changes in governmental regulations 0 0 1 1 1 0 0 0 0 0 0 1 0 0 4. increased material costs 0 0 0 1 0 0 0 0 0 0 0 0 0 1 applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 61 table 3. final reachability matrix. factors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 driving power 1. instability of the us dollar 1 0 0 1 0 1 1 1 0 1 1 0 0 1 8 2. poor cost estimation 0 1 0 0 0 0 0 0 0 0 0 0 0 1 2 3. changes in governmental regulations 0 0 1 1 1 0 1 0 0 0 1 1 0 0 6 4. increased material costs 0 0 0 1 0 0 0 0 0 0 0 0 0 1 2 5. delays in material delivery 0 0 0 0 1 0 0 0 0 0 1 0 0 1 3 6. increased wages 0 0 0 0 0 1 0 1 0 0 1 0 0 1 4 7. increased equipment costs 0 0 0 0 0 0 1 0 0 0 0 0 0 1 2 8. poor communication within the project team 0 0 0 0 0 0 0 1 1 0 1 1 0 0 4 9. design changes during the construction phase 0 0 0 1 0 0 1 0 1 0 1 1 0 1 6 10. high staff turnover 0 0 0 0 0 1 0 1 1 1 1 1 0 1 7 11. schedule delays 0 0 0 0 0 0 0 0 0 0 1 0 0 1 2 12. procurement changes 0 0 0 1 1 0 1 0 0 0 1 1 0 1 5 5. delays in material delivery 0 0 0 0 1 0 0 0 0 0 1 0 0 0 6. increased wages 0 0 0 0 0 1 0 0 0 0 0 0 0 1 7. increased equipment costs 0 0 0 0 0 0 1 0 0 0 0 0 0 1 8. poor communication within the project team 0 0 0 0 0 0 0 1 1 0 0 0 0 0 9. design changes during the construction phase 0 0 0 0 0 0 0 0 1 0 1 1 0 0 10. high staff turnover 0 0 0 0 0 1 0 1 0 1 1 0 0 0 11. schedule delays 0 0 0 0 0 0 0 0 0 0 1 0 0 1 12. procurement changes 0 0 0 1 1 0 1 0 0 0 1 1 0 0 13. poor communication among project stakeholders 0 0 0 0 0 0 0 0 1 0 1 0 1 0 14. project cost overruns 0 0 0 0 0 0 0 0 0 0 0 0 0 1 k. alzebdeh, h.a. bashir and s.k. al siyabi 62 13. poor communication among project stakeholders 0 0 0 1 0 0 1 0 1 0 1 1 1 1 7 14. project cost overruns 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 dependence level 1 1 1 6 3 3 6 4 4 2 10 6 1 12 2.4 partitioning the reachability matrix in this step, the final reachability matrix is partitioned by finding the reachability and antecedent sets for each element. the reachability set contains the element itself and other elements which it may cause, whereas the antecedent set contains the element and other elements which may cause it. then, the intersection of the two sets is derived for all elements. the element(s) for which the antecedent set and intersection set are equal are placed at the bottom level in the ism hierarchy. once the bottom level elements are identified, they are separated from the rest of the elements. this process is repeated to find the next higher level of elements. the partitioning step is completed when the highest level of elements is reached. these identified levels are used to help build the digraph and final model. in the present case study, this process was accomplished in eight iterations. tables 4 and 5 show the elements, along with their reachability sets, antecedent sets, and intersection sets and levels for the first two iterations. in table 4, elements 1 (instability of the us dollar), 2 (poor cost estimation), 3 (changes in governmental regulations) and 13 (poor communication among project stakeholders) are found at level i. therefore, they are placed at the bottom level of the hierarchy in the ism model. after discarding elements 1, 2, 3, and 13 from table 4 of the first iteration, element 5 (delays in material delivery) and element 10 (high staff turnover) are found at level ii in the next iteration as shown in table 5. thus, they will be positioned at the second to bottom level of the hierarchy of the ism model. the iteration process is repeated until the last (top) level is reached (level vii), which contains element 14 (project cost overruns) as shown in table 6. 2.5 formation of the ism diagram once the hierarchical levels of all factors and both direct and indirect interrelationships as determined from the reachability matrix and after including transitivity links are known, a preliminary model called a diagram is developed. the indirect links between the factors are removed to obtain the interpretive structural model. 3. discussion the aforementioned steps of ism were applied to the cost overruns problem. a portrayal of the final ism model is shown in fig. 3, where the 13 factors are segregated in a hierarchy structure showing the sequence of cause and effect relationships among them. clearly, the ism model appears to be cohesive and easy to interpret, and proves that the project cost overrun is not a result of individual factors; rather, it is the net result of many interacting factors occurring in sequence or parallel. the bottom level factors of the hierarchy (instability of the us dollar, poor cost estimation, changes in governmental regulations, and poor communication among project stakeholders) are the immediate causes of the problems of high staff turnover and delays in material delivery in the second highest level of the hierarchy structure. subsequent factors in level ii are causing the factors in level ii and so on. as such, the factors at the bottom (level i) are the root cause of the cost overruns. by directing corrective measures at the root causes, the risk of problem recurrence is more likely to be minimized. the result of the current study was compared with the findings of research in other developing countries as reported in the literature. the k. alzebdeh, h.a. bashir and s.k. al siyabi 63 table 4. iteration 1. element reachability set antecedent set interaction set level 1 1,4,6,7,8, 10, 11,14 1 1 i 2 2,14 2 2 i 3 3,4,5,7,11,12 3 3 i 4 4,14 1,3,4,9,12,13 4 5 5,11,14 3,5 5 6 6,8, 11 1,6, 10 6 7 7,14 1,3,7,9,12,13 7 8 8,9,11,12 1,6,8,10 8 9 4,7,9,11,12,14 8,9,10,13 9 10 6, 8,9,10,11,12,14 1, 10 10 11 11,14 1,3,5,6,8,9,10,11,12,13 11 12 4,5, 7,11,12,14 3,8,9,10,12,13 12 13 4,7,9,11,12,13,14 13 13 i 14 14 1,2,4,5,7,9,10,11,12,13,14 14 table 5. iteration 2. element reachability set antecedent set interaction set level 4 4,14 4,9,12 4 5 5,11,14 5 5 ii 6 6,8, 11 6, 10 6 7 7,14 7,9,12 7 8 8,9,11,12 6,8,10 8 9 4, 5, 7,9,11,12,14 8,9,10 9 10 6, 8,9,10,11,12,14 10 10 ii 11 11,14 5,6,8,9,10,11,12 11 12 4,7,11,12,14 8,9,10,12 12 14 14 4,5,7,9,10,11,12,14 14 table 6. iteration 8. element reachability set antecedent set interaction set level 14 14 14 14 vii thirteen identified factors that cause cost escalations in construction projects in oman are typical of similar projects in the uae, kuwait, malaysia, vietnam, ghana, and others. of note is that “inflation and sudden change in prices” has been ranked the number one risk factor in emirati construction projects (el-sayegh 2008). in contrast, “coordination with subcontractors” and “quality of work” were among the highest risk factors in the kuwaiti construction industry (kartam and kartam 2001). the findings of a major cost overrun risk study in china (wang and yuan 2011) ranked the “crude cost estimation” to be number six among a group of 13 factors leading to cost overruns. similarly, “inaccurate estimation” was the most important risk factor in construction projects in vietnam (long et al. 2004). furthermore, “poor contract management” and “inflation” were ranked number 2 and 4, respectively ,in studying cost overruns in construction of groundwater projects in developing countries (frimpong et al 2003). “information and communication” was ranked the second most severe contributor of cost overrun in construction in malaysia (abdul azis 2013). in this study, the four factors found as the root causes of the cost overruns were in good agreement with previous studies which were conducted in similar developing countries. on the other hand, factors were analysed according to their driving power and dependence level as shown in table 3, where any entry of 1 applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 64 along the rows indicates that that factor has a driving power in causing other factors while any entry of 1 along the columns indicates a dependence on other factors. subsequently, a diagram for the driving power-dependence level was constructed (fig. 4). based on the values in table 3, one can observe that factor 1 has a driving power of 8 and a dependence level of 1. therefore, in fig. 4, the value is positioned in a location corresponding to a driving power of 8 and dependence level of 1. as such, these elements are classified into four clusters according to the levels of driving power and dependence (fig. 4). the first cluster (i) consists of autonomous factors, which have both a weak driving power and dependence level. these factors are relatively disconnected from the system, and have only a few links that may be considered strong. the second cluster (ii) consists of the dependent factors that have weak driving power but a strong dependence level. the third cluster (iii) has linkage factors with both a strong driving power and dependence level. these factors are unstable because any actions taken on them will affect others. the fourth cluster (iv) includes independent factors that have strong driving power (an ability to influence other elements), but a weak dependence level (slightly influenced by other elements). a factor with a very strong driving power is called the key element, and falls into the category of independent or linkage factors. factor 1, instability of the us dollar; factor 3, changes in governmental regulations; factor 9, design changes during the construction phase; factor 10, high staff turnover, and factor 13, poor coordination among the project parties, fall into the independent category, requiring greater intervention to minimize their effect. 4. conclusions in assessment of effective solutions for the cost overrun problem, it is necessary to identify the root causes. this can be achieved by identifying and modeling the complex relations among the causal factors. for this purpose, the current study proposes to use ism, which is characterized by its ability to simplify a complex set of interrelated variables into a simple structural model. the ism was applied to develop a model for factors that caused costs to escalate for construction projects carried out in oman. thirteen factors causing project cost overruns and their interrelationships were determined using an expert opinion approach. then, interactions among these factors were organized by the ism procedure into a simple yet very easy to interpret hierarchy structure consisting of seven levels. the bottom level factors—two external and two internal factors—were the root causes of the cost overrun problem in the current study. the two external factors are the instability of the us dollar and changes in governmental regulations. the two internal factors are faulty cost estimation and poor coordination among the project parties. to better control future projects’ costs, remedies for these causes should be the first focus of attention. the conclusion drawn from the ism model provided several important implications for project managers and can be used in case studies to deal with similar problems in future projects. specifically, managers can better deal with the instability of the us dollar—an external factor —by adopting a hedging policy that mitigates risks in foreign currency exchanges when multiple currencies are used in procurement and budgeting. the effects of changes in governmental regulations can be minimized by setting aside contingency funds during the planning stage. these funds should be retained to pay for mandatory and optional modifications initiated by the owner or the contractor after the contract has been awarded. a more accurate cost estimate process could be achieved by using an analogy approach in combination with cost estimate techniques such as parametric methods (bashir 2008). coordination between parties could be enhanced by adopting quality function deployment (qfd) procedures during the planning phase, a common approach in the manufacturing industry that has also found its way into the construction industry (dissanayaka et al. 2007). careful attention should be given to developing a transparent and mutuallybeneficial relationship between the owner and the contractor that enables the scope of k. alzebdeh, h.a. bashir and s.k. al siyabi 65 figure 3. ism-based model for the cost overrun problem. instability of the us dollar (1) poor communication among project stakeholders (13) poor cost estimation (2) changes in governmental regulations (3) high staff turnover (10) increased wages (6) poor communication within the project team (8) design changes during the construction phase (9) procurement changes (12) schedule delays (11) increased material costs (4) increased equipment costs (7) delays in material delivery (5) project cost overruns (14) level ii level iii level iv level v level vi level vii level i k. alzebdeh, h.a. bashir and s.k. al siyabi 66 11 10 9 8 1 iv iii 7 13 6 3 10 9 5 12 4 6 8 3 5 2 2 i 4,7 ii 11 1 1 2 3 4 5 6 7 8 9 10 11 dependence level figure 4. driving power and dependence level diagram. the work and plan execution to be successful. this would enable a high degree of cooperation and a team spirit to develop between project participants. as current research has found ism to be effective in identifying and modeling the complex interrelationships among factors causing project cost overruns, the methodology should be extended to other industries. however, the ism methodology may have some limitations, even though it has demonstrated itself to be highly reliable and practical. one drawback is that it relies on experts’ opinions, limiting the use of sensitivity analysis. moreover, implementing the ism manually tends to be rather time consuming. this could be addressed by putting the entire system into a computer application. references abdul azis a, memon a, abdul rahman i, karim a (2013), controlling cost overrun factor in construction projects in malaysia. research journal of applied sciences and technology 5(8), 2621–2629. ali as, kamaruzzaman sn (2010), cost performance for building construction projects in klang valley. journal of building performance 1(1), 110–118. al-khalil mi, al-ghafly ma (1999), delay in public utility projects in saudi arabia. international journal of project management. 17(2), 101–106. ansari m, kharb rk, luthra s, shimmi sl, chatterji s (2013), analysis of barriers to implement solar power installations using interpretive structural modelling technique. renewable and sustainable energy reviews 27:163–174. arditi d, akan gt, gurdamar s (1985), reasons for delays in public projects in turkey. construction management and economics 3(2):171–181. assaf sa, al-hejji s (2006), causes of delay in large construction projects. international journal of project management 24:349–357. azhar n, farooqui ru, ahmed sm (2008), cost overrun factors in construction industry of pakistan, in proceedings of the first international conference on construction in developing countries. karachi, pakistan, august 4–5. bashir ha (2008), modelling of development time for hydroelectric generators using factor and multiple regression analyses. international journal of project management 26(4):457–464. driving power applying interpretive structural modeling to cost overruns in construction projects in the sultanate of oman 67 bashir ha (2010), removal of redundant relationships in an aon project network for evaluating schedule complexity. journal of construction engineering and management 136(7):787–793. bashir ha, thomson v (1999), metrics for design projects: a review. design studies 20(3):263–277. bolanos r, fontela e, nenclares a, paster p (2005), using interpretive structural modeling in strategic decision making groups. management decision 43:877–895. brand jr. dh, irwin dm, kawamura k (1976), implementation of interpretive structural modelling in a state-level planning context, in proceedings of the seventh annual conference on modelling and simulation, pittsburgh, pennsylvania, usa, april 26–27. creedy gd. (2006), risk factors leading to cost overrun in highway construction projects, in sidwell, a.c. (ed) proceedings of queensland university of technology research week international conference, brisbane, australia. unpublished doctoral dissertation, queensland university of technology, australia, july 4–8. dissanayaka sm, delgado-hernandez dj, bampton ke, anspinwall e (2007), quality function deployment in construction. construction management and economics 25(6):597–609. el-sayegh s (2008), risk assessment and allocation in the uae construction industry. international journal of project management 26(4):431–438. enshassi a, al-najjar j, kumaraswamy m (2009), delays and cost overruns in the construction projects in the gaza strip. journal of financial management of property and construction 14(2):126–151. fairdi a, el-sayegh sm (2006), significant factors causing delay in the uae construction industry. contraction management and economics 24(11):1167–1176. farooquie p, khan j (2007), an interpretive structural model for project planning and success, in ieee international conference on industrial engineering and engineering management, singapore, december 2–5, 1022–1026. flyvbjerg b, skamris holm mk, buhl sl (2004), what causes cost overrun in transport infrastructure projects? transport reviews 24(1):3–18. flyvbjerg b, skamris holm, mk, buhl sl (2003), how common and how large are cost overruns in transport infrastructure projects? transport reviews 23(1):71–88. frimpong y, oluwoye j, crawford l (2003), causes of delay and cost overruns in construction of groundwater projects in a developing countries; ghana as a case study. international journal of project management 21(5):321–326. govindan k, palaniappan m, zhu q, kannan d (2012), analysis of third party reverse logistics provider using interpretive structural modelling. international journal of production economics 140:204–211. hasan ma, shankar r, sarkis j (2007), a study of barriers to agile manufacturing international. journal of agile systems and management 2:1–22. hawthorne rw, sage ap (1975), on applications of interpretive structural modelling to higher education program planning. socio-economic planning sciences 9(1)31–43. iyer k, sagheer m (2010), hierarchical structuring of ppp risks using interpretative structural modeling. journal of construction engineering and management 136(2):151– 159. kaliba c, muya m, mumba k (2009), cost escalation and schedule delays in road construction projects in zambia. international journal of project management 27(5):522–531. kaming p, olomolaiye p, holt g, harris f (1997), factors influencing construction time and cost overruns on high-rise projects in indonesia. construction management and economics 15(1):83–94. kartam n, kartam s (2001), risk and its managements in the kuwaiti construction industry: a contractors' perspective. international journal of project management 19(6):325–335. kawamura k, christakis an (1976), the role of structural modelling in technology assessment, in proceedings of the second international congress on technology assessment, ann arbour, usa, october. koushki pa, al-rashid, k, kartam n (2005), delays and cost increases in the construction of private residential projects in kuwait. construction management and economics 23(3):285–294. le-hoai ly, lee, d, lee jy (2008), delay and cost overruns in vietnam large construction projects: a comparison with other selected k. alzebdeh, h.a. bashir and s.k. al siyabi 68 countries. ksce journal of civil engineering 12(6):367–377. long nd, ogunlana so, quang t, lam kc (2004), large construction projects in developing countries: a case study from vietnam. international journal of project management 22(7):553–561. mandal a, deshmukh sg (1994), vender selection using interpretive structural modelling. international journal of operations & production management 14(6): 52–60. mansfield nr, ugwa oo, doran t (1994), causes of delay and cost overrun in nigerian construction projects international journal of project management 12(4):254–260. mathiyazhagan k, govindan k, noorul haq a, geng y (2013), an ism approach for the barrier analysis in implementing green supply chain management. journal of cleaner production 47(2):283–297. ogunlana s, promkuntong k, jearkjirm v (1996), construction delays in a fast growing economy: comparing thailand with other economies. international journal of project management 14(1):37–45. ravi v, shankar r (2005), analysis of interactions among the barriers of reverse logistics. technological forecasting and social change 72(9):1011–1029. rawlinson jg (1986), creative thinking and brainstorming. surrey, united kingdom: ashgate. sage ap (1979), methodology for large-scale systems. new york: mcgraw-hill. sahney s, banwet dk, karunes s (2010), quality framework in education through application of interpretive structural modeling: an administrative staff perspective in the indian context. the tqm journal 22: 56–71. sambasivan m, soon yw (2007), causes and effects of delays in malaysian construction industry. international journal of project management 25(5):517–526. schlesinger la, heskett j (1991), breaking the cycle of failure in services. mit sloan management review 33(3):17–28. shankar c, transue m, felder fa (2011), analysis of barriers to development in landfill communities using interpretive structural modelling. habitat international 35:246–253. sharma hd, gupta ad, sushil pv (1995), the objectives of waste management in india: a future inquiry. technological forecasting and social change 48(3):285–309. singh md, shankar r, narain r, agarwal a (2003), an interpretive structural modelling of knowledge management in engineering industries. journal of advanced research in management 1(1):28–40. sweis g, sweis r, abu hammad a, shboul a (2008), delays in construction projects: the case of jordan. international journal of project management 26(6):665–674. talib f, rahman z, qureshi mn (2011), analysis of interaction among the barriers to total quality management implementation using interpretive structural modeling approach. benchmarking: an international journal 18:563–587. wang g, wang y, zhao t (2008), analysis of interactions among the barriers to energy saving in china. energy policy 36(6):1879– 1889. wang jy, yuan hp (2011), major cost-overrun risks in construction projects in china. international journal of project organization and management 3(4):226–242. warfield jw (1974), developing interconnected matrices in structural modelling. ieee transactions on systems, man, and cybernetics 4(1):51–81. civl010604.qxd the journal of engineering research vol. 2, no. 1 (2005) 53-68 ______________________________ *corresponding authors e-mail: alnuaimi@squ.edu.om 2d idealisation of hollow reinforced concrete beams subjected to combined torsion, bending and shear a.s. al-nuaimi*1 and p. bhatt2 1department of civil and architectural engineering, college of engineering, sultan qaboos university, p.o. box 33, al-khodh 123, muscat, sultanate of oman 2honorary senior research fellow, department of civil engineering, glasgow university, glasgow, g12 8qq, uk received 1 june 2004; accepted 4 september 2004 abstract : this paper presents a finite element model for idealisation of reinforced concrete hollow beams using 2d plane elements. the method of ensuring compatibility between the plates using two-dimensional model to analyze this type of structures is discussed. cross-sectional distortion was minimised by incorporating end diaphragms in the fe model. experimental results from eight reinforced concrete hollow beams are compared with the non-linear predictions produced by a 2d in-house fe program. the beam dimensions were 300x300 mm cross section with 200x200 mm hollow core and 3800 mm long. the beam ends were filled with concrete to form solid end diaphragms to prevent local distortion. the beams were subjected to combined bending, torsion and shear. it was found that the two-dimensional idealisation of hollow beams is adequate provided that compatibility of displacements between adjoining plates along the line of intersection is maintained and the cross-sectional distortion is reduced to minimum. the results from the 2d in-house finite element program showed a good agreement with experimental results. keywords: 2d analysis, finite element method, hollow beams, bending, torsion, shear, combined load, numerical model notation b = the strain matrix composed of derivatives of shape functions d = the elasticity matrix {f} = the equivalent nodal forces for the continuum ec = the initial modulus of elasticity of concrete for uniaxial loading. es = the secant modulus of elasticity at peak stress es = σp / εp fcu = concrete cube compressive strength f'c = concrete cylinder compressive strength ft’ = tensile strength of concrete {fi} = the total applied load vector ¢¢üü∏∏îîàà°°ùÿÿggáfé°sôÿg øe áyƒæ°üÿg áaƒûg qƒ°ù÷g πªëàd iƒ°ü≤dg iq~≤dgh ∑ƒ∏°ùdg π«ãªàd oé©h’g »féæk oh~ùg ô°üæ©dg ωg~îà°séh »hé°ùm êpƒ‰ ábqƒdg iòg ¢vô©à°ùj : íféø°u úh •é≤æ∏d á«≤a’g ámgr’g ¿g .äé°ûf’g øe ´ƒædg gòg π«∏ëàd oé©h’g »féæk êpƒªædg ωg~îà°séh íféø°üdg úh ≥agƒàdg ¿éª° d âe~îà°sg á≤jôw ábqƒdg ¢ûbéæjh .áë∏°ùÿg èféàædg ™e oé©h’g »féæk oh~ùg ô°üæ©dg èeéfôh ωg~îà°sg øy áœéædg á«£nódg ägƒäñæàdg èféàf âfqƒb .√霒g ¢ùøf ‘ äémgr’g §hgƒ°v ™°vh ≥jôw øy é¡«∏y ô£«°s qƒ°ù÷g á≤jô£h qƒ°ù÷g √òg ™«æ°üj ”h .º∏e 200* 200 ™£≤ã æqéa ∞jƒœ oƒlh ™e º∏e 3800 ¬dƒwh º∏e 300* 300 ô°ù÷g ™£≤e ¿éc .áaƒ› qƒ°ùl á«féªk o~©d á«∏ª©ÿg »féæk ô°üæ©∏d »hé°ù◊g èeéfèdg èféàf äô¡xg ~≤d .»hé°ù◊g π«∏ëàdg ‘ aõ÷g gòg π«ã“ ” ~bh »∏ùg agƒàd’g ™æÿ ∂dph áfé°sôÿéh âä∏e å«m áરüe ô°ùl πc »àjé¡f π©œ ‘ é¡«∏y oéªày’g øµá oé©h’g á«féæk ioh~fi ô°uéæy ωg~îà°séh π«ãªàdg á≤jôw ¿g ≈∏y ∫~j é‡ …èıg πª©dg ‘ â°ù«b »àdg ∂∏j øe 顪«b ‘ áñjôb á«©bƒj º«b ,oé©h’g .qƒ°ù÷g øe ´ƒædg gò¡d iƒ°ü≤dg iq~≤dgh ∑ƒ∏°ùdg äé©bƒj üé°ùm ««mmééààøøÿÿgg ääggooôôøøÿÿgg.»hé°ùm êpƒ‰ ,áë∏°ùe áfé°sôn ,oh~fi ô°üæy ,oé©h’g »féæk π«ã“ ,qƒ°ùl :á ¢ü≤dgh »æãdgh ≈∏dg iƒ≤d á°vô©ÿgh áë∏°ùÿg ¬fé°sôÿg øe áyƒæ°üÿg ázôøÿg qƒ°ùé∏d oé©hc’g »féæk π«ãªàdg 54 the journal of engineering research vol. 2, no. 1 (2005) 53-68 1. introduction considering the complex behaviour of hollow beams, a detailed analysis would normally require a full threedimensional finite element model especially with box-sections for large girder bridges. however, a study of the structural behaviour of typical thin-walled concrete beams indicates that the main stress conditions are those of direct stresses in the plates of the box beam. the forces involved in out-of-plane bending are very small and can be ignored. the distortion of cross-section is prevented by the use of reasonably thick plates and end diaphragms. this suggests that the main stresses are in-plane ones and, therefore, plane stress elements can be used to account for the major stresses. zero stiffness is assumed for out-of-plane bending action of the component plates. figure 1 shows the state of stress in a typical box beam subjected to bending, torsion and shear. the wall thickness used for torsional resistance was 50 mm which is 1/6 of the beam depth (thurliman, 1979; macgregor and ghoneim, 1995). bhatt and beshara (1980) studied the behaviour of bridge box girders using similar method with plane elements. in their work, out of plane bending was considered due to the large size of the box girder. the advantages of using 2d approach over a full 3d finite elements solution as studied by abdel-kader (1993) is that it is easier and leads to cheaper computations while at the same time the main stresses are obtained with reasonable accuracy. in this study, a 2d in-house finite element program was used to analyse eight reinforced concrete box beams. the actual beam ends were filled with concrete to form end diaphragms. the diaphragms were included in the numerical analysis. the predicted behaviour was compared with the experimental results. lc = computational failure load = (mc+tc)/2 ld = design ultimate load = (md+td)/2 le = experimental failure load = (me+te)/2 l.f. = load factor l.f. = (me/md + te/td)/2 md,td,vd = design ultimate bending moment, twisting moment and shear force, respectively me,te,ve = experimental failure bending, torsion and shear force, respectively mc,tc,vc = computed failure bending, torsion and shear force, respectively {ri} = the residual force vector at ith iteration α = the ratio of the principal stresses σ2 / σ1 {δ } = the nodal displacements of the continuum εcc = concrete strain at peak stress εcr = the tensile crack strain (corresponds to the peak tensile stress) εmax = maximum compressive strain εn = a fictitious strain normal to the crack plane εp = the strain at the peak (maximum) compressive stress of the concrete σp ε/εy = steel strain ratio = measured strain at l.f. load / measured yield strain v = the poisson's ratio σ and ε = the current stress and strain σp = the ultimate strength of concrete in compression, equal fc’ τtor = shear stress due torsion τshr = shear stress due shear force. torsionbending shear figure 1. state of stress in a typical beam 55 the journal of engineering research vol. 2, no. 1 (2005) 53-68 2. research significance a method using 2d plane elements to model 3d hollow beams is introduced. the compatibility of displacement between adjoining plates was maintained by introducing constraints on the displacement. the cross-sectional distortion was minimized by incorporating the end diaphragms into the in-house fe program. this work is restricted to the analysis of hollow cross-sections with inplane stresses. 3. geometrical relationship between displacements the two-dimensional idealisation of box girders is adequate provided compatibility of displacement between adjoining plates along the line of intersection is maintained and cross-sectional distortion is reduced to a minimum. to achieve these objectives, the following steps were adopted as shown in fig. 2: 1. to ensure shear transfer between adjoining plates of the beam, compatibility of displacement along the line of intersection at the common edge of adjoining plates is maintained by introducing geometrical constraints. 2. to reduce cross-sectional distortion, end diaphragms are introduced into the analysis. to illustrate this technique, consider top flange, front web and left diaphragm of a typical beam as shown in fig. 3. for ease, only corner nodes of some elements are shown in this illustration. if the out-of-plane bending is ignored, then the web and the flange can be considered as thin plates in a state of plane stress. however, the displacements of both plates along the joining line are equal to each other. the displacements perpendicular to the joining line are independent of each other. the same applies to the lines joining the plates with the diaphragm. it is therefore, necessary in this analysis to enforce geometrical constraints to ensure compatibility along the lines of intersections. this is done by giving the same freedom number for those equal displacements. other freedoms which are independent are given different numbers. in other words, every pair of displacements in the x-direction (direction of the beam axis) of the joining line between the flange and the web will be having the same freedom number and, therefore, the same displacement value for that pair. the displacements perpendicular to this line will have different numbers. in addition to the freedom numbering, attention should be given to plate orientation when the whole structure is assembled, to prevent contradicting directions of displacements. the rigid body movement is prevented by proper restraints, which are dependent on the load conditions and support locations. 4. a 2-d finite element program in modelling the linear and non-linear responses of concrete as a continuum, the non-linear elasticity approach was adopted. in this approach, the bulk modulus, shear modulus, poisson's ratio and young's modulus of concrete are expressed in terms of stress/strain variables, such as deviatoric stresses or strains, stress or strain invariants, normal and octahedral strain ... etc. the moduli were used to formulate an isotropic matrix to represent the behaviour of concrete at a certain load level (alnuaimi, 1999; kotsovos and pavlovic, 1995; chen and saleeb, 1994). no bar dual action or bar kinking were included in the analysis. the above idealisation was implemented in a 2d finite element program originally developed for carrying out non-linear analysis of solid rectangular beams abdelkader, (1993). in this program, eq. 1 was used to represent the ascending portion of the uniaxial compressive stress-strain curve. this equation was tested by darwin and pecknold, (1977) on many experimental results and used by bhatt and abdel-kader, (1995) to numerically analyse many reinforced and pre-stressed concrete beams from different experimental investigations. (1) in modelling the non-linear stress-strain relationship of the concrete in the principle stress direction, eq. (2) was adopted as proposed by liu et al. (1972) and used by bhatt and abdel-kader, (1995). this equation accounts for non-linear behaviour of concrete in biaxial compression and takes the form: (2) if α = 0, i.e. for uniaxial state of stress, eqs. (1) and (2) become identical. equation (2) was used to generate the stress-strain behaviour of concrete in biaxial compression up to peak strain εp, after which this equation ceases to be valid due to softening deformation. a linear descending curve was adopted by bhatt and abdel kader (1995) as given by eq. (3) after the maximum stress was reached until a maximum strain of 0.0035 after which the stress drops to zero. (3) figure 4 shows typical stress-strain curve for combined ascending and descending parts plotted using eqs. (1) and (3). smeared crack model with orthogonal cracks was adopted. in this model, it is assumed that before cracking the concrete is a homogeneous, isotropic and linear elastic 2 21 . ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ +⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ −+ = pps c c e e e ε ε ε ε ε σ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ +− − +− = 2 )2 1 1 (1)1( . pps c c e e v v e ε ε ε ε α α ε σ 0.0035< )1.0( )1.0( ' ε ε ε σ c cc f − − = 56 the journal of engineering research vol. 2, no. 1 (2005) 53-68 web compatibility of displacement top flange diaphragm figure 2. imposed displacement constraint a b c 9 3 6 15 18 1 2 4 5 7 8 10 11 14 13 16 17 12 a b c a 5 7 9 23 11 13 15 17 2119 20 22 24 25 2627 3940 2 4 656 53 54 192740 55 top flange front web left diaphragm figure 3. freedom constraints 57 the journal of engineering research vol. 2, no. 1 (2005) 53-68 material and after cracking the concrete remains a continuum but becomes orthotropic with one of the material axis orientated along the direction of the crack. the idea of orthogonal cracks was adopted, where cracks are allowed to open only in directions orthogonal to the existing cracks. these fixed orthogonal cracks are governed by the direction of the first principal (tensile) stress that exceeds the cracking strength. when a first crack occurs, it is assumed that direct tensile stresses cannot be supported in the direction normal to the crack and therefore, the modulus of elasticity corresponding to this direction (normal to crack) is reduced to zero unless tension stiffening is considered. the reduced shear stiffness due to aggregate interlock is accounted for by a reduced shear modulus of elasticity. on further loading, orthogonal cracks occur when the tensile stress parallel to the first cracks becomes greater than the concrete tensile strength f’t (alnuaimi, 1999; kotsovos and pavlovic, 1995; chen and saleeb, 1994; abdel-kader, 1993). the effect of the shear retention factor was considered as given by eq. (4) while eq. (5) was used for modelling the tension stiffening: (4) (5) figure 5 shows typical shear retention curve and fig. 6 shows tension stiffening curve used in this model. the program was based on the finite element displacement method, load control, where the displacements are the prime unknowns, with stresses being determined from the calculated displacement field. the non-linear problems were solved satisfying the basic laws of continuum mechanics: equilibrium, compatibility and constitutive relationship of materials. in this program, the continuum (structure) is divided into a series of distinct non-overlapping eight nodded 2d isoparametric elements with nine gauss points for integration. element stiffness matrix is given by eq. (6). (6) the summation of terms in eq. (6) over all elements leads to the continuum stiffness matrix [k]. this is used in a system of equilibrium equations for complete continuum given by (7) this system of equations is solved using direct gaussian elimination algorithms in conjunction with frontal method of equation assembly and reduction to yield the nodal displacements. the main feature of this method is that it assembles the equations and eliminates the variable at the same time. hence, the complete structural stiffness is never formed. this reduces computer storage significantly. once this is done, the strains {ε} and thereafter the figure 4. typical combined stress-strain curve 05.0 /4.0 ≥= ncr εεβ ∫= ev te dvbdbk ]][[][][ }]{[}{ δkf = 2/')4( mmntfn cr t ε εσ = eq. (1) eq. (3) 58 the journal of engineering research vol. 2, no. 1 (2005) 53-68 εn åcr 05.04.0 ≥⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ = n cr ε ε β figure 5. shear retention curve εεn ◊◊ t f′′cr εcr ' 4 t n cr t f⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = ε ε σ figure 6. tension stiffening curve σt εn εn β εcr βmin (11) and the norm of the applied loads fi* is given by eq. (12) (12) when a reinforced concrete member is subjected to a tensile stress exceeding its tensile strength, the concrete cracks at discrete locations. the total force is then transferred across the crack by the tensile steel. the reinforcing bars were modelled as embedded in orthogonal directions. perfect bond (full compatibility) between steel and concrete was assumed. the displacement of any point on the bar was obtained from the displacement field of the isoparametric element in the bar direction. the stiffness matrix of the bar ks was calculated separately and then added to the stiffness matrix of the concrete kc to form the element stiffness matrix eq. (13) (13) in addition, the following parameters were used as constants: 5. validation of the proposed model the program was used in a parametric study for the analysis of eight hollow reinforced concrete beams tested at the university of glasgow, uk (alnuaimi, 1999; alnuaimi, 2002; alnuaimi, 2003) under combined bend59 the journal of engineering research vol. 2, no. 1 (2005) 53-68 stresses {σ} in each element are evaluated using eqs. (8) and (9) respectively (8) (9) incremental iterative procedure with modified newtonraphson method was used for the numerical solution of the non-linear problems. in this approach, the total load is divided into a number of load increments and the solutions are obtained iteratively until equilibrium is achieved to an acceptable level of accuracy. for every increment the stiffness matrix was updated at the first iteration. to ensure adequate convergence to the exact solution, continuity of displacement between adjoining elements, material constitutive relation as well as equilibrium are maintained by constraints on nodal points and ensuring that at any load level, stresses are consistent with displacement field and material constitutive relationship. this is obtained by successive linear solutions until specified accuracy is reached. if the material is within the elastic region, the relationship between the nodal forces and the displacement is linear and the stiffness matrix is unchanged. however, if the material yields then the relationship is non-linear and the stiffness matrix has to be updated. this is done by a succession of linear approximations considering the new material law. this piece-wise linearization is used to form a global non-linear solution for the problem. a yield criterion for biaxial stress conditions based on non-dimensionalised values of the principal stresses σ1 and σ2 with respect to the uniaxial cylinder strength of concrete f’c’ was adopted. this criterion was originally proposed by kupfer et al. (1969) and tested and used by many researchers. detailed derivations of compressioncompression, compression-tension and tension-tension criteria can be found in many references (i.e. alnuaimi, 1999; kotsovos and pavlovic, 1995; chen and saleeb, 1994; kupfer et al. 1969). to ensure gradual elimination of out-of-balance (residual) forces and to terminate the iterative process when the desired accuracy has been achieved, a convergence criterion based on the out-of-balance load norms was used. since it is difficult and expensive to check the decay of residual forces for every degree of freedom, an overall evaluation of convergence was used. this was done by using the residual force norm in each iteration i. this criterion assumes that convergence is achieved if: (10) { } { }eb δε ][= { } { }εσ ][d= 04.0* * ≤ ∆ i i f r where the norm of residuals *ir∆ at each iteration i, is given by e q. (11) }{}{* i t ii rrr =∆ }{}{* i t ii fff = sce kkk += • concrete cube compressive strength: cuf as measured for each beam. • concrete cylinder compressive strength: cuc ff 8.0 ' = n/mm2. • young’s modulus: 2' /5000 mmnfe cc = . • compressive strain at peak stress: 2500 ' c cc f =ε . • maximum compressive strain: 0035.0max =ε . • tensile strength of concrete: '' 54.0 ct ff = n/mm2. • mesh size: 84x84mm. • maximum number of load increments: 30. • maximum number of iteration s per increment: 100. • load increment: 10% for the first five increments and 5% thereafter. • steel strain points for comparison with experimental results were located as in the actual beam. 60 the journal of engineering research vol. 2, no. 1 (2005) 53-68 ing, shear and torsion. the main variables studied were the ratios of twisting moment to bending moment which varied between 0.19 to 2.62 and the ratio of shear stress due torsion to shear stress due shear which varied from 0.59 to 6.84 as shown in table 2. 5.1 test beams the tested beams were 300x300 mm cross section with 200x200 mm hollow core and 3800 mm length. the outer 400 mm of each end was filled with concrete to make it solid to form a diaphragm to prevent distortion. the beams were subjected to bending moment, twisting moment and shear force. the middle 1200 mm of the beam was considered as test span. this is the region where both maximum moment and shear occur and with least effect of concentrated stresses near the ends. figure 7 shows loading and support arrangement and fig. 8 shows close look of the torsional arm and concrete filled end of beam diaphragm. the beam was simply supported by a set of two perpendicular rollers at each support to allow axial displacement and rotation about a horizontal axis at the soffit of the beam fig. 9. bending and shear were direct result of mid-span point down-ward load while constant torsion was applied by means of the torsion arms. during testing the load was applied in increments as a percentage of the design load, ld, 10% for the first three increments, in anticipation of crack initiation, and then 5% until failure. the beam was considered to have collapsed when it could resist no more loads. table 1 shows average measured material properties. figure 10 shows the reinforcement provided along the test span of each beam. the solid circles refer to bars on which strain gauges were stuck while the hollow ones refer to bars without strain gauges installed. 5.2. comparison between experimental and the 2d computational results here some experimental and computational results are compared. the comparison was carried out using the following criteria: * load displacement relationship, * longitudinal steel strain, * transverse steel strain, * failure load, * crack pattern and mode of failure. 5.2.1 load-displacement relationship figure 11 shows displacements at mid-span of the beam. it is clear from this figure that, in general, an acceptable agreement between experimental and computational results was achieved in most cases. it should be noted that in few cases, due to technical problems, it was not possible to record displacement readings just before failure (i.e. btv6). 5.2.2 longitudinal steel strain figure 12 shows that a very good agreement between experimental and computational results was obtained for longitudinal steel strains. the reported strain ratios were closest to the mid-span of the beam. 5.2.3 transverse steel strain figure 13 shows that a very good agreement between experimental and computational results was obtained for transverse steel strains. the reported strain ratios were at the mid-depth of the beam section, closest to mid-span. 5.2.4 failure load the ratios of experimental to computational failure fcu f'c fy fyv age beam no unit n/mm2 n/mm2 n/mm2 n/mm2 days btv1 39 33 495 516 10 btv2 37 24 490 472 7 btv3 38 27 490 472 7 btv4 42 33 480 472 7 btv5 35 27 490 472 8 btv6 35 28 490 472 7 btv7 54 34 500 472 7 btv8 53 36 500 472 8 average 41.6 30.2 492 477.5 7.6 table 1. average measured material properties 61 the journal of engineering research vol. 2, no. 1 (2005) 53-68 test span 1000mm 900mm 900mm 1000mm 1200mm figure 7. loading and support arrangement figure 8. torsion arm with a jack and load cell installed (see the solid end of the beam) 1200 mm 1000 mm 900 mm 900 mm 1000 mm test span 62 the journal of engineering research vol. 2, no. 1 (2005) 53-68 figure 9. set of perpendicular rollers btv1 btv2 btv3 btv4 btv5 btv6 btv7 btv8 y8 @66 mm c /c 4y12 2y8 2y10 2y8 y8 @60 mm c /c 4y12 2y8 2y10 2y10 y8 @93 mm c /c 5y10 2y8 2y10 2y8 y8 @84 mm c /c 5y12+1y8 2y8 2y10 2y8 3y10 2y8 2y8 2y8 y8 @173 mm c /c y8 @120 mm c /c 4y10 2y8 2y10 3y10 2y8 2y8 2y10 y8 @92 mm c /c y8 @80 mm c /c 4y10+1y8 2y8 2y10 2y8 figure 10. reinforcement provided in the test span 63 the journal of engineering research vol. 2, no. 1 (2005) 53-68 0 0.2 0.4 0.6 0.8 1 1.2 1.4 disp.(m m ) l. f. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 disp(mm) l .f . -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 disp.(m m ) l. f. 0 0.2 0.4 0.6 0.8 1 1.2 disp.(m m ) l. f. comput. exp 10mm 10mm btv 3 btv 4 10mm 10mm btv 1 btv 2 20mm 20mm btv 5 btv 6 20mm 20mm btv 7 btv 8 figure 11. vertical displacement at mid-span 64 the journal of engineering research vol. 2, no. 1 (2005) 53-68 0 0.2 0.4 0.6 0.8 1 1.2 1.4 ε/εy l. f. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 ε/εy l. f. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 ε/εy l. f. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 ε/εy l. f. comput exp 1 1 btv1 btv2 1 1 btv3 btv4 1 1 btv5 btv6 1 1 btv7 btv8 figure 12. longitudinal steel strain ratios 65 the journal of engineering research vol. 2, no. 1 (2005) 53-68 0 0.2 0.4 0.6 0.8 1 1.2 1.4 ε/εy l. f. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 ε/εy l. f. 0 0.2 0.4 0.6 0.8 1 1.2 ε/εy l. f. -0.2 0 0.2 0.4 0.6 0.8 1 1.2 ε/εy l. f. comput exp 1 1 btv 1 btv 2 1 1 btv 3 btv 4 1 1 btv 5 btv 6 1 1 btv 7 btv 8 figure 13. transverse steel strain ratios 66 the journal of engineering research vol. 2, no. 1 (2005) 53-68 loads le/lc are given in table 2. a very good agreement was attained in failure loads. in most cases, the experimental results were slightly larger than the computed ones. wide range of t/m and τtor/τshr ratios did not result in large differences between experimental and computational results. 5.2.5 crack pattern and mode of failure when the principle tensile stress at any point in the structure exceeds the maximum tensile strength of the concrete a crack is initiated perpendicular to the direction of the principle tensile stress. figure 14 shows typical computed and observed crack patterns of the tested beams. the first graph in each pair represents the predicted location of cracks and their orientations in the test span just before failure load. it can be seen that a good agreement was achieved between the experimental and computational results on the crack concentration and orientation. all beams failed in a ductile manner as can be seen from figures on displacements and steel ratios (figs. 11 13). steel yielded in most cases or reached near yield before the concrete crushed. the load was transferred from the concrete to steel at about 35% of failure load. enough fine cracks have developed in each case long before major cracks development near the failure load. beam td md vd td/md τtor/τshr le/lc no. knm knm kn ratio ratio ratio btv1 13 14.89 21.08 0.87 2.28 0.99 btv2 13 32.89 41.08 0.4 1.17 1.1 btv3 13 50.89 61.08 0.26 0.79 1.14 btv4 13 68.89 81.08 0.19 0.59 1.12 btv5 26 14.89 21.08 1.75 4.56 0.97 btv6 26 32.89 41.08 0.79 2.34 0.88 btv7 39 14.89 21.08 2.62 6.84 1.06 btv8 39 32.89 41.08 1.19 3.51 1.01 table 2. design, experimental and computed failure loads figure 14a. computed and observed crack development in the back web (btv1) 67 the journal of engineering research vol. 2, no. 1 (2005) 53-68 figure 14b. computed and observed crack development in the front web (btv3) figure 14c. computed and observed crack development in the front web (btv6) figure 14d. computed and observed crack development in the front web (btv8) 68 the journal of engineering research vol. 2, no. 1 (2005) 53-68 6. conclusions from the results presented in this paper, it can be concluded that the 2-d idealisation of hollow beams using plane elements is adequate for cross-sections with inplane stresses. the 2-d in-house finite element program used for the non-linear analysis gave good results when compared with experimental ones. wide range of ratios of torsion to bending, t/m, and shear stress due to torsion to shear stress due to shear force, τtor/ τshr, did not result in large differences between experimental and computational failure loads. references abdel-kader, m.m.a., 1993, “prediction of shear strength of reinforced and prestressed concrete beams by finite element method,” ph.d. thesis, university of glasgow. alnuaimi, a.s., 1999, "direct design of reinforced and partially prestressed concrete beams for combined torsion, bending and shear," ph.d. thesis, university of glasgow, uk. alnuaimi, a.s., 2002, "parametric study on the computational behaviour of hollow beams designed using the direct design method material factors," proceedings of high performance structures and composites, seville, spain, wit press publishers, march, pp. 605-614. alnuaimi, a.s., 2003, "parametric study on the computational behaviour of hollow beams designed using the direct design method numerical factors," proceedings of international conference on advances in structures, asscca'03, sydney, australia, a.a., 22-25 june, balkema publishers, vol. 2, pp. 1017 1021. bhatt p. and abdelkader, m., 1995, "prediction of shear strength of reinforced concrete beams by nonlinear finite element analysis," developments in computational techniques for civil engineering. civil-comp press, edinburgh, scotland, uk. pp. 191-204. bhatt, p. and beshara, a.w., 1980, "in-plane stresses in end diaphragms of box girder bridges," journal of the institution of engineers (india), part cv; civil engineering division; sep. vol. 61, pp. 103-109. chen, w.f. and saleeb, a.f., 1994, “constitutive equations for engineering materials,” elsevier science b.v., sara burgerhartstraat 25, p.o. box 211, 1000 ae amesterdam, the netherlands, p. 1129, isbn: 0444884084. darwin, d. and pecknold, d. a., 1977, "nonlinear biaxial stress-strain law for concrete," j. of the engineering mechanics division, asce, vol. 103 (em2), april, pp. 229-241. kotsovos, m.d. and pavlovic, m.n., 1995, “structural concrete,” thomas telford publications. 1heron quay, london e14 4jd, uk, p. 550, isbn: 0727720279. kupfer, h. hilsdorf, h.k. and rusch, h., 1969, "behaviour of concrete under biaxial stresses," aci journal, august, pp. 656-666. liu t.c.y., nilson, a.h., and slate, f.o., 1972, "stressstrain response and fracture of concrete in uniaxial and biaxial compression," aci journal, may, pp. 291295. macgregor, j.g. and ghoneim, m.g., 1995, "design for torsion," aci structural j, no. 92-s20, mar.apr., pp. 211-218. thurlimann, b., 1979, "torsional strength of reinforced and prestressed concrete beams-ceb approach," institut fur baustatik und konstruktion, eth. zurich., no. 92. june, pp. 117-143. mech24503_current.qxd 1. introduction requirements for comfortable living and working conditions have resulted in an increased demand for air conditioning. the latter is normally carried out using a vapour compression system powered by electricity and incorporating a chlorofluorocarbon (cfc) refrigerant. generation of electricity by conventional means results in vast quantities of co2 being released to the atmosphere. concerns over global warming and the environmental damage caused by cfc refrigerants have stimulated research to develop alternate systems for air conditioning. one such system is based on evaporative cooling. this technique eliminates environmental damage and offers energy savings in running costs. among the various methods of evaporative cooling, indirect evaporative cooling has been used widely in industrial and commercial air conditioning (joudi and mehdi, 2000; suri, 2001), not only because it operates at very low energy cost, but also because it cools the air without adding moisture to it. indirect evaporative cooling may be applied to buildings in the form of a chilled ceiling. this technique is becoming increasingly popular in many european countries as an alternative means of air conditioning (costelloe and finn, 2003; gasparella and longo, 2003). ___________________________________________ *corresponding author email:saffa.riffat@nottingham.ac.uk indirect evaporative coolers use evaporation of moisture in the outside air to produce a cooling effect. this is transferred to the supply air via a heat exchaner or thermal conductor without the addition of moisture to the supply stream. 2. methodology 2.1 cooler description the cooler consists of a porous ceramic and a heat pipe (riffat, 2001), figure 1. the cool source is a porous ceramic filled with water. when outside dry air passes across the surface of the ceramic, it causes water to siphon to its surface through the multitude of pores. the water then evaporates and creates a cooling effect. the heat pipe used in this system acts as the heat exchanger, and transfers the coolness to the supply air. the condenser section of the heat pipe is embedded in the porous ceramic, figure 2. the air supply stream is cooled when it passes across the evaporator section of the heat pipe. its heat is absorbed by the heat pipe’s working fluid, causing the latter to evaporate. the vapour then condenses at the condenser section of the heat pipe and surrenders its heat to the water in the ceramic. 2.2 porous ceramic this experimental investigation studied the effects of the journal of engineering research 1 (2004) 46-52 experimental investigation of an indirect evaporative cooler consisting of a heat pipe embedded in porous ceramic s. riffat* and j. zhu school of the built environment, university of nottingham, ng7 2rd, uk received 24 may 2003; accepted 5 november, 2003 abstract: an indirect evaporative cooler incorporating a porous ceramic and a heat pipe was described. experiments were carried out to examine the effects of various ceramic properties, such as porosity, wall thickness, and outside diameter. use of the cooler for a chilled ceiling in an environmental chamber was also investigated. a temperature drop of 3.8oc per square metre of ceramic surface area was achieved for an 18 m chamber. keywords: heat pipe, porous ceramic, chilled ceiling, indirect evaporative cooling :¢ü∏îà°ùÿÿgqéîødg ¢üfé°ün òkéj qéñàn’ üqéœ âjôlg .péøf qéîa ‘ øª° e …qgôm üƒñfg øe ¿ƒµªÿ ô°téñeódg ôîñàdéh oèÿ éø°uh ábqƒdg ω~≤j 3ω 18 á©°s ¢sé«b áaôz ~jèj ‘ oèÿg á«∏yéa á°sgqo â“ éªc .oèÿg á«∏yéa ≈∏y …qéîødg üƒñfód »lqéÿg ô£≤dgh ¿gq~÷g ∂ª°sh ájòaéædéc .qéîødg øe ™hôe îe πµd ¢ù«°ù∏°s ¢sé«≤e ≈∏y álqo 3h8 qg~≤ã iqgô◊g álqo ‘ ¢véøîfg ßmƒd ~bh :á«méàøÿg ägoôøÿg.ô°téñeódg ôîñàdéh ~jèàdg ,~jèj ∞≤°s ,péøf qéîa ,…qgôm üƒñfg several factors on the performance of the ceramic cooler. 2.2.1 ceramic porosity the ceramics were made from clay material with the chemical composition shown in table 1. however, the processes of manufacture were different and ceramics with different apparent porosities were obtained from batches fired at different temperatures, table 2. the expressed porosity is a percentage by volume, and it was determined from the amount of water absorbed by the ceramic. 2.2.2 ceramic wall thickness ceramics with different wall thickness, i.e., 10mm, 15mm and 20mm, were investigated. the outside dimensions of the ceramics were kept constant and so consequently the inside volumes were slightly different in each case. 2.2.3 ceramic outside dimension cylindrical ceramics with diameters of 60mm, 120mm and 240mm were investigated with other parameters being kept the same. the cylindrical surface area for water evaporation therefore increased proportionately. 2.2.4 thermal conductivity copper wool was packed between the ceramic and the heat pipe when the investigation of thermal conductivity was carried out. heat was therefore transferred from the heat pipe to the ceramic by both water and copper wool under this condition. 47 riffat and zhu / the journal of engineering research 1 (2004) 46-52 figure 1. cooler with heat pipe embedded in porous ceramic water vapour cold water warm air cold air porous ceramic condenser section of heat pipe evaporator section of heat pipe fins supply water wick figure 2. indirect evaporative component composition (%) sio2 69.39 al2o2 19.23 fe2o3 0.47 tio2 0.26 mgo 0.18 cao 0.84 k2o 1.62 na2o 0.36 loi 6.91 table 1. ceramic composition porosity % firing temperature (oc) firing cycle (hr) 29.7 1020 6.5 32.8 950 5.5 34.6 810 4.5 table 2. ceramic fired 2.3. cooler construction 2.3.1 cooler unit the sealed tube of the heat pipe was constructed of phosphorus deoxidized non-arsenical copper. the tube had an outside diameter 20mm and length 950mm; the fin height was 15mm and the length of the evaporative section was 350mm. the length of the heat pipe condenser section (embedded in ceramic) was 400mm. a screen mesh was used as the wick of the heat pipe, and the working fluid was distilled water. the outside length of the ceramic was 420mm. 2.3.2 chilled ceiling a group of indirect evaporative coolers was assembled to form the chilled ceiling, figure 3. ceramics with a porosity of 32.8%, an outside diameter of 60mm and a wall thickness of 10mm, without copper wool, were used. only the evaporative section of the heat pipe was fixed in the frame. the chilled ceiling dimension was 450mm (l) × 420mm (w) × 340mm (h). the evaporator section of the heat pipe was in the room and a fan was used to blow the room hot air through the heat pipe evaporative section from the bottom of the chilled ceiling. the cooled air left from both sides of the ceiling and returned to the room. after becoming hot, the air was recycled. the adiabatic heat pipe section was in the wall of room. the porous ceramics were on the outside of the room. the total surface area for water evaporation was 0.712 m2. 2.4 experiments 2.4.1 cooler test the cooler apparatus largely consisted of two wind tunnels that were thermally isolated from their surroundings, figure 4. the ceramics were installed in the outside air stream tunnel. the evaporator section of the heat pipes was installed in the supply air stream tunnel. the adiabatic section of the heat pipe was between the two tunnels. each tunnel had an air velocity sensor, and two temperature sensors for inlet and outlet air streams. in the outside air stream tunnel, two humidity sensors were installed for measuring the inlet and outlet air humidities. all sensors were connected to a datataker that sent the data to a computer. the air stream velocities were adjusted via two fans. a humidifier or dehumidifier controlled the inlet humidity of the outside air stream. heaters regulated the inlet air stream temperatures of the two tunnels. a total of nine types of ceramic were tested in this study. they were divided into three groups; one group with different porosities, a second group with different wall thicknesses, and the third group with different outside diameters. each group of ceramics was tested under the same inlet air stream parameters, e.g., temperature, velocity and humidity. the thermal conductivity test was carried out using ceramics with copper wool. ceramics with, and without, copper wool were tested under the same conditions. the cooling capacity q is defined as: 48 riffat and zhu / the journal of engineering research 1 (2004) 46-52 cool air hot air cool air evaporator section of heat pipe fan figure 3. chilled ceiling [1] herein, v3 is the supply air stream velocity, s3 is the crosssectional area of the supply air stream tunnel, ρ is the air density of the supply air stream, cp is the specific heat capacity of the supply air stream, t3 and t4 are the inlet and outlet air temperatures of supply air stream, respectively, and sc is the total surface area of ceramic for water evaporation in outside air stream tunnel. 2.4.2 chilled ceiling test the chilled ceiling test was carried out using an environmental chamber, figure. 5. two identical rooms in the chamber were selected based on the cooler performance. the walls of the rooms were made of a sandwich of 3mm wood board, 50mm polystyrene board and 3mm wood board. the roofs were 3mm wood board, 50mm polystyrene board and 20mm wood board. the inside dimensions of each room were 3m (l) × 2.4m (w) × 2.5m (h). the chilled ceiling was installed in room 1. seven temperature sensors and two humidity sensors were installed in the environmental chamber in order to measure the surrounding temperature and humidity. there were eight temperature sensors in each room to allow an average room temperature to be obtained. all sensors were connected to a datataker attached to a computer. the two rooms had the same heat loads throughout the experiments. prior to the tests, they had the same temperature. the doors and windows of the two rooms were closed during the tests. 49 riffat and zhu / the journal of engineering research 1 (2004) 46-52 porous ceramic outside air stream tunnel supply air stream tunnel evaporator section of heat pipe adiabatic section of heat pipe figure 4. experimental set-up for cooler environmental chamber room 2 room 1 porous ceramics figure 5. environmental chamber 50 riffat and zhu / the journal of engineering research 1 (2004) 46-52 outside air tunnel supply air tunnel inlet outlet velocity inlet outlet velocity q porosity (%) t1(oc) rh1(%) t2 ( oc) rh2 (%) v1 (m/s) t3( oc) t4( oc) v3 (m/s) (w/m 2) 35.4 22.4 31.8 33.6 1.08 35.0 32.2 0.36 29.70 35.2 22.9 31.9 32.2 1.56 34.9 31.8 0.37 33.58 29.7 35.2 23.1 32.2 31.6 2.03 35.2 31.9 0.39 36.05 34.9 40.3 31.4 57.1 1.07 35.2 32.0 0.44 41.03 35.6 36.6 32.0 51.5 1.52 35.3 31.9 0.46 44.68 32.8 35.3 37.6 31.9 51.2 1.97 35.4 31.9 0.46 46.49 35.1 35.0 30.9 52.3 1.14 35.1 31.1 0.46 52.52 35.3 32.8 31.4 47.6 1.58 35.1 31.0 0.46 53.35 34.6 35.6 32.9 31.6 46.7 2.01 35.2 30.5 0.43 59.21 table 3. effect of ceramic porosity outside air tunnel supply air tunnel inlet outlet velocity inlet outlet velocity q thickness (mm) t1(oc) rh1 (%) t2 ( oc) rh2 (%) v1 (m/s) t3( oc) t4( oc) v3 (m/s) (w/m 2) 35.3 32.8 31.4 47.6 1.58 35.1 31.0 0.46 53.35 35.6 32.9 31.6 46.7 2.01 35.2 30.5 0.43 59.21 10 35.2 33.2 31.4 46.3 2.31 34.9 29.8 0.42 62.40 35.4 27.0 30.8 44.6 1.53 35.5 31.8 0.46 48.94 35.8 31.6 31.8 46.2 2.01 35.8 31.7 0.46 53.72 15 35.6 31.4 31.7 45.2 2.31 35.7 31.4 0.46 56.71 35.3 29.5 31.7 42.8 1.57 35.0 31.5 0.43 43.63 35.3 30.5 31.7 43.2 2.02 35.1 31.2 0.42 47.83 20 35.6 32.5 32.4 43.3 2.35 35.5 31.5 0.46 52.67 table 4. effect of ceramic wall thickness outside air tunnel supply air tunnel inlet outlet velocity inlet outlet velocity q dia-meter (mm) t1( oc) rh1(%) t2( oc) rh2 (%) v1 (m/s) t3 ( oc) t4 ( oc) v3 (m/s) (w/m 2) 60 35.7 33.6 33.8 42.7 0.59 35.2 32.6 0.42 65.65 120 35.4 37.0 30.8 61.6 0.43 35.4 32.3 0.44 40.24 240 35.2 37.7 27.2 72.7 0.54 35.4 31.1 0.45 28.17 table 5. effect of ceramic outside diameter outside air tunnel supply air tunnel inlet outlet velocity inlet outlet velocity with copper wool t1( oc) rh1(%) t2 ( oc) rh2 (%) v1 (m/s) t3( oc) t4( oc) v3 (m/s q (w/m2) 35.2 33.5 30.1 59.4 0.43 35.2 32.0 0.46 42.26 35.0 33.3 30.7 51.9 1.07 34.9 31.4 0.45 45.80 35.4 27.0 30.8 44.6 1.53 35.5 31.8 0.46 48.94 35.8 31.6 31.8 46.2 2.01 35.8 31.7 0.46 53.72 no 35.6 31.4 31.7 45.2 2.31 35.7 31.4 0.46 56.71 35.6 27.1 29.7 55.5 0.49 35.7 32.4 0.48 45.15 35.4 27.2 30.2 47.9 1.04 35.5 31.9 0.47 48.45 35.8 33.9 31.1 51.0 1.52 35.3 31.4 0.46 52.36 34.8 33.9 30.5 49.5 2.02 35.2 30.7 0.45 59.09 yes 36.0 30.0 31.2 45.1 2.30 35.6 31.0 0.48 63.27 table 6. effect of thermal conductivity 3. results and discussion 3.1 cooler 3.1.1 effect of ceramic porosity it was found that cooling capacity increased with ceramic porosity due to increased evaporation for ceramics with high permeability, table 3. for example, cooling capacity increased from 44.68 w/m2 to 53.35 w/m2 as the porosity increased from 32.8% to 34.6% under the same inlet air conditions (including outside air tunnel temperature t1, relative humidity rh1 and velocity v1, and supply air tunnel t3 and v3). the average rate of increase of cooling capacity was 4.89 (w/m2)/% within this air velocity range. 3.1.2 effect of ceramic wall thickness cooling capacity decreased with increasing ceramic wall thickness, as shown in table 4. for example, cooling capacity decreased from 59.21 w/m2 to 53.72 w/m2 as ceramic wall thickness increased from 10 mm to 15 mm under the same inlet air conditions. this was due to the increased resistance to water permeance at greater wall thicknesses. the average decreasing rate of cooling capacity was 1.03 (w/m2)/mm. the ceramic wall should therefore be as thin as possible taking into account design strength limitations. 3.1.3 effect of ceramic outside diameter cooling capacity decreased with increasing outside diameter as shown in table 5, even though the total cooling ability (q multiply surface area) increased. for example, cooling capacity decreased from 65.65 w/m2 to 28.17 w/m2 as ceramic outside diameter increased from 60 mm to 240 mm under the same inlet air conditions. this was due to increasing heat transfer resistance between the heat pipe condenser and the ceramic as the ceramic diameter increased. this demonstrated that it is not possible to enhance the cooler performance continuously by solely increasing the surface area of ceramic. the heat transfer resistance needs to be considered simultaneously. in these tests, the ceramic outside dimensions had an influence on air flow modes in the outside tunnel. this uncertainty effect will be considered in future investigations. 3.1.4 effect of thermal conductivity cooling capacities were enhanced when copper wool was inserted into the ceramic, table 6. for example, the cooling capacity increased from 48.94 w/m2 to 52.36 w/m2 under the same inlet air conditions. this was because the thermal conductivity of water is very low compared with the heat pipe. the main heat transfer resistance was due to the water in the ceramic. copper has a high thermal conductivity, and so the heat transfer ability was improved by packing copper wool into the ceramic. the average increase in the rate of cooling capacity was 8.2% in these tests. 3.2 chilled ceiling the variations in the temperatures of the two rooms were the same, figure 6. the heat transfer and internal heat load were the same for the two rooms. the temperature difference of the two rooms was therefore achieved by 51 riffat and zhu / the journal of engineering research 1 (2004) 46-52 15 20 25 30 35 40 45 0 30 60 90 120 150 180 time [min] t [d eg ] o r r h [% ] environment-t room2-t room1-t environment-rh figure 6. room temperature in environmental chamber the chilled ceiling. initially, the temperatures of the rooms were the same. the temperature of room 1 fell quickly at the beginning but after 3 minutes the temperature difference became almost constant. a temperature drop of 3.8 oc per square metre of ceramic surface area was achieved under the average temperature 39.4 oc and the average relative humidity 20.3% of environmental chamber. 4. conclusions the performance of an indirect evaporative cooler with a porous ceramic and a heat pipe was investigated. the application of the cooler to a chilled ceiling was also studied. the investigation revealed that cooling capacity increased with ceramic porosity, the average rate of increase was 4.89 (w/m2)/%. cooling capacity decreased with ceramic wall thickness, the average rate of decrease was 1.03 (w/m2)/mm. the ceramic wall thickness should therefore be as small as possible within the constraints of structural strength requirements. cooling capacities can be enhanced by inserting copper wool into the ceramic; the average rate of increase was 8.2%. a temperature drop of 3.8 oc per square metre of ceramic surface area was achieved for a room with an internal volume of 18 m3 fitted with a chilled ceiling using the indirect evaporative cooler. references costelloe, b., and finn, d., 2003, “indirect evaporative cooling potential in air-water system in temperature climates,” energy and buildings, vol. 35, pp. 573-591. gasparella, a., and longo, g. a, 2003, “indirect evaporative cooling and economy cycle in summer air conditioning,” international journal of energy research, vol. 27, no. 7, pp. 625-637. joudi, a. k., and mehdi, m. s., 2000, “application of indirect evaporative cooling to variable domestic cooling load,” energy conversion & management, vol. 41, pp. 1931-1951. riffat, b. s., 2001, “cooling apparatus (porous pot),“ patent no. pct/gb00/04025. suri, r. k., 2001, “energy-saving potential of an indirect evaporative cooler,” applied energy, vol. 69, no. 1, pp. 69-76. acknowledgements this project was funded by the engineering and physical sciences research council, uk. 52 riffat and zhu / the journal of engineering research 1 (2004) 46-52 d:\desktop\paper 7.xps tjer vol. 11, no. 1, 64-68 __________________________________________ *corresponding author’s e-mail: gelvan_17@yahoo.co.in effect of cryogenic treatment on microstructure and micro hardness of aluminium (lm25) sic metal matrix composite g elango*a, bk raghunathb and k thamizhmarana *a department of mechanical engineering, vrs college of engineering and technology, arasur, tamilnadu, india b department of manufacturing engineering, anamalai university chidambaram, tamilnadu, india received 4 june 2013 ; accepted 16 march 2014 abstract: the basic aim of this paper is to increase awareness amongst the researchers and to draw their attention towards the present approach to deal with the cryogenic treatment for the nonferrous metals. cryogenic treated nonferrous metals will exhibit longer wear and more durability. during metal making process, when solidification takes place, some molecules get caught in a random pattern. the molecules do move about at subzero and deep cryogenic treatment slowly. in this experimental study, the effect of cryogenic treatment on microstructure changes and the hardness properties varies for lm25 alloy and lm25-sic metal matrix composite at -196°c. it is analyzed for different durations. the execution of cryogenic treatment on both alloy and mmcs changed the distribution of precipitates. the xrd crystallogram reveals that the cryogenic treatment can change the diffraction peak intensity of some crystal planes in mmcs. the influences of different volume fraction of reinforcement and cryogenic process parameters on microhardness of lm25 alloy and composite were compared with alloy and composite without cryogenic treatment and the results showed that the cryogenic treatment improves the hardness of lm25/sic composites. keywords: mmcs, cryogenic treatment, microstructure, micro hardness. m25l sic m25 lm2-sicl mmcsxrd mmcsm25 l m2/sicl mmcs 65 g elango, bk raghunath and k thamizhmaran 1. introduction the attractive performance of metal matrix composites (mmcs) has been noted in a wide range of applications, from toys to high performance-requiring areas like the automobile and aerospace industries. this is due to their weight-saving characteristics and the fact that they can provide greater benefits than existing materials (mahadevan et al. 2008). to increase the properties of these composites, several processes, like heat treatment for whole materials and surface processing are also being carried out. on the other hand, cryogenic treatment, also known as subzero treatment, is a very old process that has been used widely for high precision parts and objects and especially for the ferrous materials mentioned earlier (sendooran and raja 2011). subjecting materials to extreme cold hardens and strengthens; this method has been used for centuries (bensely et al. 2007). now cryogenic treatment is widely used in the automotive, aerospace, electronic and mechanical engineering industries to improve mechanical strength and the dimensional stability of various components (zhirafar et al. 2007). for the past few years, in order to improve properties, a cryogenic treatment for nonferrous metals such as aluminium and magnesium alloys has been used (kaveh et al. 2009). the mechanical properties and microstructure of metals and alloys in cryogenic treatment have drawn the attention of researchers. lulay et al. (2002) and jiang et al. (2009) showed the beneficial effects of cryogenic treatment on nonferrous metal aluminium. when considering the wear performance of copper alloy, cryogenic treatment yields the least significant changes (guozhi et al. 2010). however, woodcraft and adam (2005) showed a significant improvement in the mechanical properties of the strength, hardness, and toughness of aluminium alloy when subjected to cryogenic treatment. this has led to the idea of analyzing individual alloys' properties when mmcs undergo cryogenic treatment. this field is rapidly growing and is being used by many manufacturers. the present work intends to construct a facility to research the process and results of cryogenic treatment in order to create standards for both processing and testing, which are currently unavailable; hence it is important that mechanical properties of mmcs being developed are evaluated at cryogenic temperatures. to satisfy this requirement, in this experimental work cryogenic treatment was applied to strontium modified al-7%si/silicon carbide (lm25/sic) mmcs to study its effect on its microstructure and hardness. 2. experimentation the metal matrix composite was prepared by the stir casting method by taking matrix material as lm25 alloy and the particulate reinforcement as sic up to a volume fraction of 20%. the chemical composition of the lm25 alloy is given in table 1. the reinforcement particles were commercial sic with 99.5% purity. sic sized 30-50 µm were used as particle reinforcement in the composite material. the reinforcement particle varied in volume fraction by 5-20% in order to manufacture different composites. the reinforcement particle was preheated to a temperature of 550°c to remove moisture before adding it into the molten aluminium. to fabricate the composite, the various volume fraction of reinforced sic composite material was created through the liquid metallurgy technique. in this technique, the preheated sic particle was introduced into the molten pool in the vortex created in the melt by the use of a power operated stirrer, which had been coated with alumina to prevent the migration of ferrous ions from the stirrer material to the molten metal. the stirrer's speed was maintained at 550 rotations per minute (rpm) at two-thirds the depth of the molten metal. the resulting mixture of lm25 alloy and sic was tilt poured into the preheated permanent mould (elango et al. 2013). the cryogenic treatment of samples was performed by placing a lm25 alloy and lm25/sic specimen in a cryogenic chamber. this chamber was progressively immersed in a liquid nitrogen reservoir. the sample temperature was monitored by a k-type thermocouple which was used to operate a stepper motor which lowered the sample and maintained a temperature decline at the rate of 1°c/min. the temperature was lowered to -196°c over nearly 4 hours. the cryogenic processing method we followed, as outlined by (kaveh et al. 2009), is a painstakingly slow, microprocessor-controlled process (fig. 1) table 1. chemical composition (%) of lm25 aluminium alloy. 66 effect of cryogenic treatment on microstructure and micro hardness of aluminium (lm25) sic metal matrix composite which eliminates the probability of thermal shock and micro-cracking. specimens were held at -196°c for various durations (eg. 1, 5, 10, 20, 30, 40, and 50 hours) and then slowly brought up to approximately +25°c. after the completion of the cryogenic processing, the specimen was prepared for microstructure analysis according to the american society for testing and materials (astm) e3 standards. the samples were subjected to grinding and polishing followed by etching by nital. optical microscopy was taken using a metallurgical microscope and then the specimen was washed with acetone and dried thoroughly for the hardness test. the micro hardness test was conducted using a leitz micro hardness tester (leica microsystems, wetzler, germany) equipped with a vickers diamond pyramid indenter. the load applied was 1 newton. 3. results and discussion figures 2 and 3 show the microstructure images of the lm25 alloy and its composite before and after cryogenic processing with a 10% volume fraction of sic. the cryogenic processing resulted in significant changes in the microstructure of mmcs and led to the transformation of -al to the (mg17al12) phase. in the lm25 alloy, the phase, whose main strengthening effect on a-mg-based alloys at room temperature was proved by mehta et al. (2004), exhibited irregular morphologies (eutectic phase) and tiny laminar shaped morphologies. the lower mechanical properroom temperature descend ascend soak figure 1. cryogenic processing. (a) (b) figure 2. microstructure of (a) lm25 alloy and (b) lm25+sic mmcs before cryogenic treatment. (a) (b) figure 3. microstructure of (a) lm25 alloy and (b) lm25+sic mmcs after cryogenic treatment. 67 g elango, bk raghunath and k thamizhmaran ties at elevated temperatures is due to the low melting point of these alloys (kaveh et al. 2009). the cryogenic treatment of the mmcs of lm25+sic led to the changes in microstructure as shown in fig. 3(b) whereby it can be seen that the coarse divorced phase penetrated the matrix alloy. this improvement in hardness was the strengthening of the matrix against propagation of the existing defect, which is due to the important role of precipitates in the microstructure. these provide the main strengthening effect at room temperature. the xrd pattern before and after cryogenic treatment was studied. it reveals that cryogenic treatment can change the diffraction peak intensity of the crystal planes in these alloys. figure 4(a) shows the xrd pattern of lm25+sic mmcs before cryogenic treatment has a range of incident angle between 30-100°. the target is cu ka, the tube voltage is 40kv and the electric current is 60 ma. the properties of mmcs are related to its xrd patterns gained from the surface of mmcs. figure 4(b) shows that virgin surfaces of al mmcs are mainly consistent with the standard pattern for fcc al. in the xrd pattern after cryogenic treatment, all the peaks are consistent except for the half width of the (111) peak decrease. this indicates that the grains in mmcs become larger after cryogenic treatment. govindan et al. (2000) showed that the crystallization strengthens after the specimens have been cryogenic treated. due to this change in microstructure, the hardness of the cryogenic-treated samples was increased as compared with the cast specimen with no treatment as shown in figs. 5 and 6. figure 5 shows the percentage of reinforcement and cryogenic duration on the hardness properties of mmcs. it clearly depicts the increase in hardness for the increase in reinforcement percentage and the cryogenic duration in hours. the results show that hardness decreases in the lm25 alloy with increasing cryogenic treatment. this improvement was attributed to the strengthening of the matrix against the propagation of the existing defects, which is due to the important role of precipitates in the microstructure. the precipitates contribute the main strengthening effect at room temperature. since the precipitates are mainly distributed at the grain boundaries, the morphology of the particles after cryogenic treatment helps in stabilization of internal microstructure. in lm25, alloys discontinue precipifigure 5. effect of reinforcement on hardness of the lm25+sic mmcs. figure 6. effect of cryogenic treatment duration on microhardness of lm25 alloy and lm25+ sic mmcs. percentage of a1203 cyrogenic treatment duration in hours 30 40 50 60 70 80 90 100 100 500 0 (a) 2000 3000 0 30 40 50 60 70 80 90 100 2 [degree] (b) figure 4(a) & (b). xrd patterns of lm25 + sic mmcs before and after cryogenic treatment. 68 effect of cryogenic treatment on microstructure and micro hardness of aluminium (lm25) sic metal matrix composite tation of -dominance at grain boundaries, weaken the grain boundaries and reduce the hardness with increasing cryogenic duration. the xrd observation showed the cryogenic treatment caused the grain boundaries to break, resulting in many grains sized 1-3µm. the broken eq-uiaxial grains mmcs improve the hardness of the matrix alloy. 4. conclusions in conclusion, cryogenic treatment changes the morphology of precipitates in both the lm25 alloy and lm25+sic composite. the hardness of the cryogenic-treated mmcs samples can improve due to the coarse eutectic phase present in the matrix as compared with cast samples. the hardness increases with the increasing volume percentage of reinforcement for the mmcs for the same cryogenic condition. the application of cryogenic treatment has increased the effects of particulates with an increase in particulate percentage but pure al decreased the effect of cryogenic treatment. micro structural changes occurred during initial cryogenic treatment which is felt in the changes in the diffraction pattern in xrd. references bensely a, senthilkumar d, mohan ld, nagarajan g, rajadurai a (2007), effect of cryogenic treatment on tensile behavior of case carburized steel815m1. materials characterization 58(5):485491. elango g, raghunath bk, palanikumar k (2013), sliding wear of lm25 aluminium alloy with 7.5% sic + 2.5% tio2 and 2.5% sic + 7.5% tio2 and 2.5% sic + 7.5% tio2 hybrid composites. doi: 10.1177/0021998313496592 1-10. govindan ppk, nageswara rb, srivastava vk (2000), residual strength of aluminum-lithium alloy center surface crack tension specimens at cryogenic temperatures. cryogenics 40:789-795. guozhi ma, ding c, zhenhua c, wei li (2010), effect of cryogenic treatment on microstructure and mechanical behaviors of the cu-based bulk metallic glass matrix composite. journal of alloys and compounds 505(1):319-323. jiang xq, ning li, hong he, zhang xj, hao y (2009), effect of cryogenic treatment on mechanical properties and microstructure of 3102 al alloy. materials science forum 546-49(2):845848. kaveh ma, alireza t, farzad k (2009), effect of deep cryogenic treatment on microstructure, creep and wear behaviors of az91 magnesium alloy. materials science and engineering a 523:27-31. khomamizadeh f (2009), effect of deep cryogenic treatment on microstructure, creep and wear behaviors of az91 magnesium alloy. materials science and engineering a 523:27-31. lulay ke, khan k, chaaya d (2002), the effect of cryogenic treatments on 7075 aluminum alloy. journal of materials engineering and performance 11(5):479-480. mahadevana k, raghukandanb k, pai uts, pillai bc (2008), influence of precipitation hardening parameters on the fatigue strength of aa 6061sicp composite. journal of materials processing technology 198:241-247. mehta ds, masood sh, song wq (2004), investigation of wear properties of magnesium and aluminum alloys for automotive applications. journal of materials processing technology 155156:1526-1531. sendooran s, raja p (2011), metallurgical investigation on cryogenic treated hss tool. international journal of engineering science and technology 3(5):3992-3996. woodcraft, adam l (2005), recommended values for the thermal conductivity of aluminium of different purities in the cryogenic to room temperature range and a comparison with copper. cryogenics 45(9):626-636. zhirafar s, rezaeian a, pugh m (2007), effect of cryogenic treatment on the mechanical properties of 4340 steel. journal of materials processing technology 186(1-3):298-303. pdf compressor d:\desktop\paper_5.xps tjer 2013, vol. 10, no. 1, 50-64 tomer physically swelling. upon contact with fluids such as oil and water and, under the correct conditions, elastomers will almost always swell. however, depending upon the base elastomer, this can vary from an almost negligible amount to a quite large volume 1. introduction all elastomers can be and are affected by the different chemical substances with which they come into contact. the effects of this contact can vary; however, one of the most common effects is that of the elas________________________________________ *corresponding author’s e-mail: r.hibberd@oblique.eu the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends r seyger, s resink, h harms and r hibberd* ruma products bv, zeppelinstraat 5, 7903 br hoogeveen, the netherlands received 24 august 2012 ; accepted 28 november 2012 abstract: swelling elastomers have gained acceptance as very effective products for creating sealing in various industries, including those creating energy from fossil fuels and geothermal resources. this paper outlines the research and development work being conducted not only in the application of these elastomers but also in the development work required to create new generations of elastomers. it touches on fundamental research into the mechanics of swelling with the intent to create a better and more predictable sealing as well as more advanced elastomers. it lifts the veil on the direction of work being done on new elastomers being developed in order to enable a better control of swelling. by doing so, the research is opening up field of applications for new equipment designs and mechanical possibilities in the future. additionally, it addresses the need for a better and more in-depth dialogue between both chemical and mechanical engineers, and the elastomer companies and their customers on the potential that both swelling and non-swelling elastomers can offer to the industry as a whole. keywords: swelling elastomers, sealing, water-swell, oil-swell 51 r seyger, s resink, h harms and r hibberd swell of 100-150%. the degree of swell is dependent upon the fluid conditions and the elastomer type and design. the swell is effectively a result of oil or water interacting with one or more of the chemical ingredients used in the production of the elastomer. these ingredients were originally intended to enhance or influence one or another of the elastomer's mechanical properties used during the manufacturing process. over the years, a huge amount of work and research has been conducted to lower or eliminate the degree of swell while maintaining the desired mechanical properties. this swell is regarded and experienced as an undesirable effect. however, because elastomers can swell in very selective fluids and this swell can be to different degrees, it sometimes has been enhanced so as in some industries or applications to be considered a potentially useable characteristic. for years, this swell has been utilized in mining and civil engineering to create seals, and for sealing against fluid ingress (pal et al. 2010). whilst these so-called swelling elastomers have been around so long and are used in various forms by some major industries, their general acceptance and use in the oil and gas fields has been relatively recent. after initially proving their worth, they have eventually gained acceptance in the oil patch, and a whole new series of applications have been created for them which center around their unique nature. this has opened up new economic production opportunities and created new use areas to operators. probably the most widely known and talked about of these areas has been the step change they have brought as "fracturing" seals to the shale oil and gas industries (anjum 2010). here the available conventional mechanical techniques to open up the tight gas formations had technical limitations and in turn were economically restricting the production of vast reserves of so called "tight gas". the availability of swelling seals, where these could just be run and left to swell, rather than the conventional mechanical packers that needed to be operated via intervention from the surface, enabled simple quick and easy multilateral design and installation for reservoir fracturing to occur. this resulted in the freeing up of relatively cheap and substantial reserves of hydrocarbon gas and, in some newer shale oil fields, cheap and effective production. a new worldwide resurgence in interest in the long known of but economically unviable reserves has been created. in the key use area of the usa, this has led to a glut in gas and with the resulting low price has come a radical change in other industries and their use of gas in general. in the usa, gas is now replacing coal in power generation and is being used as feed stock for many existing products (young 2011). however, in spite of the acceptance by the oil and gas industry as a whole, little real new development has occurred to swelling elastomers themselves. rather, the work has centered only on the technical application of the elastomers. most of the elastomer changes, where they have happened, are just small refinements to the base elastomers' performance. basically, the elastomers have not substantially changed in the last 5 to 10 years (wack 2011). as designers and manufacturers of elastomers for use in products, we are involved not just with the present generations of swelling compounds, but we are also doing both fundamental and applied research, and development into the next generation and beyond. 2. temperature ranges, low and high while there have been a multitude of different swelling seals that will swell in various different fluids, and some very specific ones that swell only in acids or alkaline fluids, this study will look at the work done in swelling in oil and/or water as this is the major market that is of interest to most people at the moment. so where are we as a company going with the development of the next generations of elastomers? the market for swelling elastomers has moved from simple mining and civil engineering applications through to the first few oil, water, and hybrid swelling elastomer seals trialed and introduced into operations by national and international oil companies. this introduction has been into a market that has been predicted to approach or, in some studies, exceed over a third of a million seal items per year by 2014-15. this reflects quite a huge use increase, occurring in a little less than a 10-year time frame. it is difficult, however, to think that this is just a drop in the bucket when considering what is actually coming. the demand is out there and is growing for a new generation of elastomers designed to deal with more extreme environments and conditions, with demand for both higher and lower temperature ranges than have been the application norm to date. the first of these demands is for a much lower temperature application. most existing swelling elastomers perform reasonably well in temperatures down to about 40-50° celsius but, at this low temperature, the rate of swelling slows substantially. at any temperature much lower than this, standard seals and especially oil swell elastomers start to become a problem as most oils at low temperatures are quiet viscous and the diffusion gradient causing swell is quite low. when coupled with the hardness of the elastomer at lower temperatures, the result is a slow swell coupled with a low level of absolute swelling. while oil swell is proving more of a problem at low temperatures, water swell fares slightly better with reasonable swelling down to 20-25°c as long as the fluid salinity is not too high. however, to be able to improve performance, a deep 52 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends understanding of elastomer swelling is required, especially as all of these improvements are taking place in a continually price-orientated market. elastomer swelling and mechanical performance on the one side has to be coupled with improved manufacturing performance of the elastomer on the other. this enables elastomers to have not only improved technical performance but also to be commercially attractive to users. as the operating envelopes move both up and down, the quality of the ingredients as well as the quality of the manufacturing of the elastomers demand improved ingredient specifications as well as a far tighter quality control in manufacturing. if these standards are not met, the performance envelope for the elastomer becomes too variable. additionally, chemical ingredients are being rationed by manufacturers and suppliers, resulting in a frequent need to find and test comparable products from new or different sources to ensure quality is maintained or improved upon. 3. how do they work? 3.1 basic principles we are aware that many engineers are not comfortable in their understanding of the mechanics of how swelling elastomers work. they know that they do swell, but the explanation as to why is often one that a chemical engineer would understand but is not easily grasped by a mechanical engineer. conversely, the swelling phenomenon is often presented by supplier sales staff as a form of magic that is performed. this notion can be advanced simply because many do not understand the process themselves. this lack of understanding is partially due to scientists' different disciplines and backgrounds, but can also be explained by the fact that some terms in chemistry differ from those commonly used in mechanical engineering. whether it is an oil or water swell or, for that matter, any other type of fluid-motivated swelling, it is the result of an entropy difference between the elastomer and the environment in which it is being used. this energy disparity manifests itself in an effort to achieve an energy balance by creating a diffusion gradient between the elastomer and the fluid around it. for details of fick's first and second laws of diffusion, please see the appendix (wikipedia 2012). figure 1 shows how by modeling the various physical mechanical parameters of the elastomer and chemical parameters of the elastomer in the motive fluid, it is possible to predict the anticipated performance of swelling elastomers prior to actually mixing the ingredients. this then enables rapid prototype modeling on paper to precede the compound design prior to making and testing in the lab, thus speeding up and optimizing the compounding process while reducing development costs and unnecessary research time. in this way, the work centers around confirming the desired characteristics and those results are being achieved. 3.2 oil swell elastomers in oil swelling, one or more of the distillate fractions in the oil has an affinity for the base elastomer and starts via the diffusion gradient to dissolve the hydrocarbon chain that forms a part of the elastomer. this results in an oil fraction being pulled into the elastomer and a dissolving and breaking down of the long chains that form the elastomer. because the motive fluid (oil) is incompressible, as is the elastomer, this pulling in of oil fractions results in a substantial force within the elastomer. this elastomer mix is then stretched much like blowing up a balloon where the air in a balloon is being replaced by a soup of hydrocarbons made up of the oil fraction and dissolving elastomer. the resulting effect is a physical swelling of the elastomer. (note that in water swell the elastomer is not directly taking part in the process so is not itself changing in volume unless the base elastomer type is an oil swelling one as used in multifunctional hybrid elastomers). the term "base elastomer" refers to the elastomer matrix that is used and with it the different elastomers and elastomer mixtures that can be used such as acrylonitrile-butadiene rubber (nbr), hydrogenated nitrile butadiene rubber (hnbr), ethylene propylene diene monomer rubber (epdm), silicones, etc. figure 2 shows how a change in elastomer characteristics (different elastomer types 1 to 5), together with an alternative base test oil such as an intermediate restorative material (irm) or diesel, or the same elastomer and fluids but with a different temperature alters the swelling characteristics achieved in oil swelling elastomers. this makes accurate swell performance predictions for oil swell elastomers very difficult without actually lab testing the specific motive fluids. 3.3 water swell elastomers water swell works in the same way as oil swell, (ie. a diffusion gradient being created, except that the water now has an affinity via a diffusion gradient for one or more of the ingredients in the elastomer). however, it does differ in that the base elastomer is not affected by the motive water as is the case with oil swelling. there is a widely held belief that the swelling is due to specific ingredients being added to create the swell; however, this is not necessarily the case. the water's affinity can be for the plasticizers used, the vulcanization enhancers or even one of the many ingredients used to enable the elastomer to flow easily during the molding or production processes. in 53 r seyger, s resink, h harms and r hibberd 54 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends modates the fluid that is actually creating the swell. please see the appendix for a simple explanation about the working of saps (m2 polymer technologies 2012). because the swell is a physical interaction between the motive fluid and the base elastomer, the design of the swelling elastomer base is critical. if ingredients are being added in water swelling applications specifically to create the swell, a number of things become very important. the more ingredients that are added, the less elastomer per unit volume remains. it goes without saying that the elastomer imparts strength to the product. this means that to maintain the strength of the whole, the elastomer remaining has to have increased strength (more cross linking). the higher the crosslink density the more rigid the elastomer; the more it resists the load created by the fluid, the less the swell becomes. this is a very complex inter-relationship with a fine balance between swell performance and mechanical integrity for both sealing and pressure containment. the internal pressure being generated by reducing the diffusion gradient through swelling is what is being utilized to create the basis for the sealing force. ingredient quality, mixing order, and vulcanization quality and control are vital to enable a swelling product to be created with a well-designed and consistent sealing performance envelope. the effects low and high temperatures have on diffusion gradients as well as elastomer mechanical performance are crucial for the creation of commercially technically viable products with a working life that is comparable to so called non-swelling elastomers in mechanical packers. poorly designed or tested seals are easy to achieve but can be a very disappointing product in terms of longterm mechanical integrity and performance. while conventional mechanical packers are created by mechanical engineers and utilize one of the worst mechanical characteristics of elastomers, swelling elastomer seals are radically different in that they are created by chemical engineers using chemical engineering and utilize its best mechanical characteristics. this engineering change is something very new to the oil field. figure 3 illustrates the different absolute swell that can occur in an elastomer in a fixed salinity with a change in fluid temperatures versus time. what is apparent is the very steep rapid swell that occurs at the start of the swell cycle and how this eventually flattens off into a more gradual swell. this also highlights the dangers that a one-size-fits-all approach can create, and different circumstances require a different range of elastomers. 3.4 low temperature swelling as explained above, diffusion gradients and mechanical elastomer characteristics interact to create the actual physical elastomer swelling. with low temperatures, diffusion is very low and slow and the elastomer base is less flexible, which limits performance dramatically. absolute swell is also very low (this is the absolute amount an elastomer will swell at those figure 3. amount of swelling for one elastomer against swelling time under the same medium at different temperatures 55 r seyger, s resink, h harms and r hibberd conditions irrespective of how thick the swelling elastomer actually is. it will never swell more than the absolute swell because of the effects swell has on lowering the already low diffusion gradient). to create swell at lower temperatures, it is important that the diffusion gradient be improved and that the base elastomer mechanical characteristic be altered to utilize the available diffusion gradient better along with a balancing of the internal swelling pressure. a new generation of oil swell elastomer is in the pipeline, which relies on the use of the new oilswelling sap technology which is similar to those used in some of the low-end water swelling elastomers. these saps were initially developed to help oil spill clean-up but are now being applied to help to open up this lower temperature range oil swell requirement. they are proving to be an important technology in helping to improve the swell speed performance and when coupled to newly designed base elastomers are enabling swell speeds and swell levels to approach those experienced in oil swelling temperature ranges of more than 40°c. these new commercial elastomer products are appearing and being used in low temperature fields where their effective working in real conditions has been demonstrated. additionally, it means elastomer bases can be used that are not affected by the oil, as the bases are resistant to oil degradation. this has always been a limitation with oil swelling where only certain base elastomer families could be used as other base families would not react with the oil to actually swell. now it is possible to use non-oil based swelling elastomer bases that are the same as those found in water swelling and are also not affected by their reaction with the motive fluids. this will dramatically improve their ultimate potential performance. 3.5 high temperature elastomers the latest major developments have occurred at the higher end of the temperature range where demand is growing for elastomers well above 180°c. what we see is that there is now a very distinct splitting of requirements for high temperature applications. for the likes of geothermal applications, swelling at this high temperature is more or less the requirement. however, where the requirement is for the use of steam, in for example tertiary recovery in oil wells, a lower temperature swell is required, but as the steam injection occurs, a resistance to these now much higher temperatures by the already swollen elastomer is vital. these are effectively two very different characteristics made all the more difficult by the probability of the steam well probably being cycled (stopping and starting) during steam injection. as we move up to these and even higher temperatures levels, we hit two substantial walls. the base elastomer development takes considerably more time not just to develop, but also more importantly to test adequately. probably even more important though is the development of bonding agents and systems suitable for use at these high temperatures. it is all well and good to have an elastomer that will work at these conditions but it has to stay stuck to the pipe and continue to hold a seal. however, the chemical bonding along with mechanical bonding is now proving to be as hard, if not harder to develop than the elastomer itself. this is especially so when one bears in mind that chemical bonding is very sensitive to base-pipe metallurgy and base elastomers. this applies to all elastomers whether used in high temperatures or not. long-term bonding requires considerable knowledge and testing to achieve bonding on different substrates commensurate with the expected life expectancy of the seal. it is frequently overlooked, but it is the bonding that holds the swelling forces in the axial direction; the load that is applied to the bonding is much higher in a 12-foot seal than in a 1-foot seal, even though the bonding strength remains the same. efforts to restrict swelling at the end of a seal do this at the cost of the sealing capacity and often have little to do with any seal pressure extrusion. this then, coupled with the type of system needed to both measure and test the elastomers, means a considerable step into a new world. while base elastomers are constantly being developed, the higher the temperature of the swelling elastomer, the deeper we enter into a totally new field of science with chemical reactions taking place between elastomer ingredients. at the same time, one can witness reactions between these ingredients and the swelling fluids (at temperatures and pressures), all of which influence the mechanical properties of the elastomer immediately and over an extended period of time. this means that not only do we need to test the elastomer as such for mechanical properties at elevated temperatures as well as later seal design versus pressures, but we also must create new tests and testing parameters to ensure that we understand what is being seen and measured. we must ensure that we are actually creating a test that gives meaningful and understandable results. the design of the test for both accuracy and repeatability is crucial to the performance of the elastomer. swell curves have to be structurally created which enable the performance envelope boundaries to be determined, as these boundaries change according to differences in salinity and temperature. understanding the conditions and compensation factors required, ensures that the results are both meaningful and properly understood. it is not just enough to plunk a piece of rubber in a pot and then measure the swell. things happen during swelling that require compensation in the measurements being made. some of these are quite simple, such as the change in salini56 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends ty of the sample water during swelling which is caused by the sweet water being extracted from the brine and entering the elastomer. this does, however, result in a salinity increase in the motive brine and, depending on the volume of the initial motive fluids, can affect the results dramatically. likewise, the results can be affected by compensating for salinity changes between brines created at room temperature that are then heated and used at elevated temperatures. these and a few others are the factors that have to be understood and examined in detail when creating testing regimens. it goes without saying that while the above is important for determining swell rates (swell versus time) and ultimate swell against time, this must also be borne in mind when we look to the design of full scale testing for pressure holding capacity. swell is a one-toone factor; that is, the elastomer swells a unit volume by removing that equivalent unit volume from the motive fluid in which it exists. when designing the pressure testing facilities, especially for large elastomer seal volumes, that removal of large volumes of fluid from the system to achieve swell also happens. this means that the test facilities have to have this capacity in the fluid circuits and the pressure vessel to ensure that the elastomer swell is not being influenced by a change in motive fluid composition caused by the swell that is actually occurring. failure to do this results in great inaccuracy in the amount of swell present and consequently in the sealing pressure tests results actually achieved. for optimum effect we internally use 3-5 samples of standardized format bonded to steel plate samplesthe precise number depends heavily on the elastomer and application. individually, in separate measurement pots, the samples are swollen over time in order to assess swelling. we facilitate this swelling for motive fluids from distilled water by making small salinity changes in steps, beginning with sweet water and adding brine until the solution achieves a saturated state. this is then repeated with new samples, increasing the sample temperature by 20°c from under the lower end of the swell range for that elastomer, until elastomer failure is achieved at a high end temperature. thereby we determine the full operating envelope of that elastomer for the full range of salinity and temperatures. the results are then compared for repeatability and only when found to be repeatable within predetermined tolerance is the formulation under test then frozen and the results averaged to create the envelope used in our swell calculator. bonding results before and after swelling are compared to ensure that the swelling forces during swelling do not affect the bonding, as well as to ensure that the temperature and fluid have not affected the level of bonding achieved or caused any significant bonding aging. standardized international standards organizations (iso) tests are used for the bonding test measurements for each of the different substrates required. 4. remote automated tester as the operating temperatures increase to above 100° celsius for water swelling elastomers or above the oil flashpoint for oil swelling elastomers, the difficulties and dangers in measuring performance accurately increase immensely. even things that are technically simple at lower temperatures can prove difficult when testing real performance at elevated temperatures and high pressure conditions. we have done considerable work to design and build an automated unit that can be used to measure swell and sealing in high pressure situations and temperatures up to 300°c. the system is designed to be a hands-off system that allows for little operator intervention and measures swelling in real temperature and pressure situations. for the first time, it is allowing researchers to gain insight into how swelling at elevated conditions takes place. at the same time, sample dimensions have to be sufficiently large to be representative of working elastomers. if the sample dimensions are too small, the results cannot be up scaled and are not representative of larger seal size performances. the ability to achieve and then repeat swell results, as well as to confirm accurately the amount of swell is vital. in the past, it was clearly evident that when lab samples were cooled and measured for swell, results, especially those ascertained at higher temperatures yielded higher error margins when compared to those collected in live measurement situations. additionally, it was and is vital to eliminate operator variations. these variations may creep into results when using other methods both for time issues and for pressurerelated items and disallow accurate and repeatable swell rate curves to be generated. the measurement system that has been created is quite a simple method for measuring swell against time. the system works with a set rubber thickness on the outside of a shaft in a swell chamber. the thickness is kept constant but the shaft diameter is varied, bringing the sample closer or further away from the cell wall. as the rubber swells, it will reach a point where it will contact the measuring cell wall. by causing the pressure under the seal to pulsate, and measuring the pressure above and below the test piece, it is possible to compare the pressure pulse time below the sample with the pressure pulse time above the sample. as the elastomer swells, a time delay occurs between pulses both in the lower and upper sections of the shaft. this delay is caused by fluid resistance between the upper and lower parts and is caused by the elastomer's contact with the test cell wall. through automation of the testing system, the height of the pressure pulse changes as the time delay is met. in this way, first contact can be determined and, with a variable time element, the point of sealing can also be reliably determined. by varying the shaft diameter and keeping the rubber thickness constant, it is possible to set up and measure a percentage of swell diameter versus time in live conditions. by repeating this for different thicknesses, salinities and temperatures, it is possible to create swell curves for various conditions without the need for operator intervention as would have been necessary in the old way of doing these tests. freeing up the compound development engineer enables a reduction in the cost of development. it also allows the engineers to develop and fine tune the compound. with the different known swell curves, it is now possible to replace the swell curve samples with test samples and, using the time obtained by the swell curves, allow the samples to swell. additionally, an automatic test cycle allows a pressure test to be conducted for two limits: 1) pressure to failure, where the pressure is increased in steps until the sample fails, and 2) extrusion degree, where an operating pressure is achieved and the sample maintained or cycled under pressure over an extended period of time, until either failure occurs or the test is stopped. comprehension of both limits gives a much deeper understanding of the elastomer sealing performance over load change, time, and temperature. this is because swelling is not sealing, sealing has to be created by understanding how swelling occurs. the ability to test automatically and to format data directly from the data logger is vital for the development of second generation seals where the shape can contribute more to the sealing than the swelling. this is doubly so when low and high temperature sealing is required and the seal mechanics and physical characteristics heavily influence potential sealing performance. it is often forgotten that swelling is 3-dimensional, which means the seal also gets longer as it swells. this length change puts huge demands upon the bonding, especially when very long seals are created. this, together with some other proprietary measurement systems, have enabled us to develop a new generation of patent pending seals where seal-shape geometry has contributed to not only improved sealing performance but also reduced running hang up and debris interference in horizontal well open-hole liner systems. the use of this system also enables engineers to test the effects of surface treatment on finished samples' swelling rates and performance. data show a huge variation in initial swell-rate, which has been observed in samples deliberately made differently in order to ascertain the most ideal manufacturing process and the necessary quality requirements for production. simple factors, such as surface grinding, can have a huge impact on swelling rate if they are done incorrectly or poorly. what is clear is that this basic but necessary fundamental research proves that a soundly engineered compound, together with a rigorous and defined testing program, can ensure customer satisfaction. this is especially necessary in an industry where there are no testing standards or compound qualifications that would otherwise ensure a product based on sound testing and technology. 5. the classic swelling elastomer work continues on measuring and developing classic swelling elastomers. these studies are aimed primarily at extending and improving their envelopes. however, major trends in the industry will soon start to move rapidly away from these rather basic elastomers. extensive efforts are put to look into and research some of the more useful chemical tricks that are possible with elastomers and has patents pending on a number of its research project studies which have looked more deeply into the chemistry of elastomers and their swelling performance. 5.1 triggering and/or controlling swelling classic swelling elastomers swell when they come into contact with their designated motive fluids. it is a given fact, and unless some error has been made in the design of the application, swelling is guaranteed to occur. this can, however, be a potential problem on complex liners and completion systems where swelling seals are included, and extra time is required for these completions to be run with their liners into the well. whilst the elastomer types and dimensions are engineered for this, should anything untoward happen and cause a delay in achieving depth, the swelling seal can cause the liner to get stuck during the running in process. in the worst case, this potentially can lead to an inability to remove the liner from the well. 5.2 controlled swell depending on operational constraints, in most wells water-swelling elastomers can be slowed down if required by decreasing the water temperature and/or increasing the salinity by timely fluid circulation. however, the swelling can only be slowed and this approach must only be seen as a very last resort solution. with this concern in mind, we have looked into this as an integral part of our on-going retrievable swelling packer development program, which is especially timely considering the increased use of swelling 57 r seyger, s resink, h harms and r hibberd 58 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market seals in open-hole situations. we have studied in great depth different ways of stopping, starting, and even, where possible, reversing swell, and now have a number of patents pending based upon our work here. we have researched and developed new, relatively simple, controllable techniques to stop and start water swelling on demand. this now gives us the ability to actually delay the onset of swelling where desired until it is required. this in turn means that we are able to increase elastomer dimensions, ensuring a commensurate reduction in swell time to first contact and, ultimately, to full seal. whilst the development of a retrievable swelling packer is still some way off, this spin-off from the study is a valuable tool for the future. figure 4 shows the results of a simple test where four identical samples have been placed in fluid. the yellow and blue samples were placed first in distilled water for two days and then transferred to water containing a swelldelaying fluid. the green and red samples were placed in the water with the swell-delaying fluid and then transferred to distilled water. clearly the swell-delaying fluid had an effect on both sets of samples. 5.3 triggered swell we have developed a second and very exciting new technique that almost indefinitely delays swell. this technique is able to delay the occurrence of swell and then trigger it on demand to start whenever it may be desired or required do so. although this technology is still in the experimental, laboratory stages, it will ultimately lead to a third generation of new elastomer product groups with what should be vastly enhanced abilities. figures 5 and 6 illustrate the degrees of volume swell being achieved using newly developed triggered swell elastomer/fluid combinations (note this is not dimensional swell, which is roughly 1/3 of volume swell). the first shows a slight swell without triggering fluid addition, but at a low volume. these are the inherent swelling characteristics caused by the ingredients in the base elastomer. the effects of the triggering fluids can be clearly seen. additionally, the swell control that the percentage of trigger fluid can be engineered to create can be seen. in the above graph, the effect that temperature has on a fixed swell trigger fluid percentage can also be seen. by managing the trigger fluid, it is possible to totally alter the swell curves to create a system-managed swell to suit the applications required. these new offerings, together with probable developments by others in the future, will amount to a potential second-step change in the way wells are built. this step is as large as if not larger than the one that which was created by the pioneering, introduction, and application of water swelling elastomers almost 10 years ago. 6. the way forward elastomers, and elastomer and chemical engineering are not really areas that have received the focus figure 4. elastomer swelling in water and in swell-delaying fluid 59 r seyger, s resink, h harms and r hibberd and attention in well design that they deserve. whilst most well engineers have a sound and deep understanding of metallurgy as well as well mechanics, they do tend to have a very poor to almost non-existent understanding of the potential that elastomers and elastomer chemistry can bring to well engineering. most of them only have had contact with elastomers via packer elements that utilize what is possibly the weakest of the characteristics of elastomers. what is clear is the potential that chemical engineering offers to mechanical engineers in their quest for ever simpler and more cost effective sealing in figure 5. effect of triggering amount on volume swelling figure 6. effect of temperature on volume swelling for 20% trigger 60 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends wells. this technology is only beginning to be understood and used. elastomer engineers working with mechanical equipment designers and well engineers can alter the face of well design and building very substantially, as can be seen by the impact seen of swelling seals on shale gas and oil recovery. hard wearing elastomers can potentially replace hard facing on drilling equipment. they can create hard wearing and cost effective corrosion layers on equipment prone to corrosion attack, at a cost substantially lower than that associated with exotic metallurgy. what is clear is that well engineers do not know enough about elastomers, and elastomer engineers do not know enough about well design to automatically create solutions to problems. there is clearly a large deal of education that needs to be given by elastomer engineering companies on both swelling and nonswelling elastomers and their potential uses to completion equipment designers and equipment users. simultaneously, the completion equipment designers and users clearly need to educate the elastomer companies on their potential requirements, wishes, and dreams. there is still a real disconnect between both parties as to what they can mean to each other and the future market and trends that can offer joint options and commercial as well as technical possibilities. this disconnectedness needs to be addressed by all parties to help the future of cost effective hydrocarbon production operations make a move into the future. it was, after all, elastomer engineers who created the solution that has freed up shale gas and oil, while it was the well engineers who saw the potential and grasped it. references anjum mg, shafeeq u, younas ma (2010), solid expandable tubular combined with swellable elastomers facilitate muli-zonal isolation and fracturing. spe paper # 131489-ms, international oil and gas conference and exhibition, beijing, china. m2 polymer technologies incorporated. (2012), how sap works. from: www.m2polymer.com/ html/ chemistry_sap.html. pal k, rajasekar r, kang dj, zhang zx, pal sk, das ck, kim jk (2010), effect of fillers on natural rubber/high styrene rubber blends with nano silica: morphology and wear. materials and design 31(2):677-686. wack h, hintemann d, michael h, buschner n (2011), preparation and properties of swellable thermoplastic elastomer alloys based on elastomeric powder, polypropylene, and superabsorbent polymer. journal of applied polymer science 120(3):1290-1296. wikipedia (2012), fick's first and second laws. from http://en.wikipedia.org/wiki/fick%27s_law. young d, sevre ak, smith pe, al-madlouh m, awang mz (2011), effects of rapid gas decompression on swellable rubber and common oilfield rubber compounds. spe paper # 149076-ms, spe/dgs saudi arabia section technical symposium and exhibition, al-khobar, saudi arabia. 61 r seyger, s resink, h harms and r hibberd (a1-1) (a1-2) (a1-3) (a1-4) (a2-1) 62 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends (a2-2) (a2-3) (a2-4) (a2-5) (a2-6) 63 r seyger, s resink, h harms and r hibberd 64 the future of swelling elastomers: an elastomer manufacturer's view of swelling elastomer developments and market trends and here’s what a super absorbent network looks like on a larger scale. c:\users\raoh\desktop\paper 2.xps the journal of engineering research vol. 12, no. 1 (2015) 15-31 comparative analysis of maximum power point tracking controllers under partial shaded conditions in a photovoltaic system r. ramaprabhaa and s.p. chitra*b department of electrical and electronics engineering, ssn college of engineering, rajiv gandhi salai, kalavakkam-603110, chennai, tamilnadu, india. received 18 march 2014; accepted december 2014 abstract: mismatching effects due to partial shaded conditions are the major drawbacks existing in today’s photovoltaic (pv) systems. these mismatch effects are greatly reduced in distributed pv system architecture where each panel is effectively decoupled from its neighboring panel. to obtain the optimal operation of the pv panels, maximum power point tracking (mppt) techniques are used. in partial shaded conditions, detecting the maximum operating point is difficult as the characteristic curves are complex with multiple peaks. in this paper, a neural network control technique is employed for mppt. detailed analyses were carried out on mppt controllers in centralized and distributed architecture under partial shaded environments. the efficiency of the mppt controllers and the effectiveness of the proposed control technique under partial shaded environments was examined using matlab software. the results were validated through experimentation. keywords: partial shaded pv system, efficiency, artificial neural network, centralized controller, distributed controller. : corresponding author’s email: chitraadk@gmail.com r ramaprabha and s.p. chitra 16 1. introduction photovoltaic (pv) energy generation technology has been investigated intensively in the past decades due to continuously growing demands for renewable, environmentally friendly, and costefficient energy resources. at any given time and environmental condition, there is a single maximum power point (mpp) in pv panel characteristics. maximum power point tracking (mppt) is an electronic form of tracking that utilizes algorithms and control circuits to search for a maximum power point in a pv panel, thus allowing a converter circuit to harvest the maximum power available from a pv panel. generally, the maximum power extracted from a pv panel depends on insolation conditions, load profile, and temperature. from the output and the v-i characteristics of commercial pv panels, it can be observed that the temperature changes mainly affect the pv output voltage while the irradiation/insolation levels affect the pv output current. mpp changes due to both sources of the variations (masoum et al. 2002). basically, the mppt algorithm can be classified as direct or indirect control. the indirect methods are based on the use of a database, where the parameters are calculated from the pv curves or by using mathematical functions. even though they are simple and inexpensive, they are not versatile with respect to load profile. in contrast, the direct control measures either pv voltage or current. the optimum operating point is calculated by taking into account the variations of the operating point. in the current study, the operating point was independent of temperature, insolation, and degradation levels but the problems were the possibility of undesirable errors and difficulties with tracking accuracy. according to this basic principle, different methodologies, like differentiation, feedback voltage, perturb-and-observe, and incremental conductance, should be implemented. the efficiency and accuracy can still be improved by employing modern controllers using fuzzy logic or neural network (walker 2001). mppt in a uniformly illuminated condition is simple and easy to implement as there is a single optimum peak. nowadays there is an increasing trend toward using pv panels in urban rooftops or installing building integrated pv (bipv) systems. in such applications, pv panels are mounted to outer-fit the roof/wall of the buildings. in these cases, different panels may receive different levels of insolation, which resembles conditions of partial shading. in partial shaded environments, to prevent panels being damaged by a hotspot (koutroulis et al. 2001; femia et al. 2005), antiparallel bypass diodes are used. this leads to the occurrence of multi-peaks in characteristics (jain & agarwal, 2004; patel & agarwal, 2008). to track the global peak, a complex mppt is required; many researchers have been working in this area and have developed algorithms (esram & chapman 2007; silvestre et al. 2009; herrmann et al. 1997). some researchers have addressed this problem by using distributed architecture where each panel has a separate converter (walker & semia, 2004; walker & pierce, 2006; kasper et al. 2014). this increases the modularity, maximum power extracted, flexibility, and stability. in spite of the above advantages, the switching losses and converter costs are quite high when compared with centralized architecture where a single mppt controller is employed. this paper investigates the effectiveness and efficiency of mppt controllers in distributed architecture and compares it with mppt controllers in centralized architecture. the paper is organized as follows: section ii deals with the effects of partial shading. section iii discusses the artificial neural network (ann) control algorithm and its training methodology. sections iv and v describe the mppt controller in centralized and distributed architecture. section vi presents the simulation results and their experimental verification. section vii concludes with the comparative analysis of mppt controllers and their suitability in specific applications. 2. effects of partial shading to understand the behavior of a solar cell, the equivalent circuit of the pv panel is derived based on discrete electrical components. a single diode model with five parameters is considered in this work (ramaprabha & mathur, 2008; karatepe et al. 2007; villalva et al. 2009). this model offers a good compromise between simplicity and accuracy. the modeling equations are: comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 17 shidiphipvi (1) shr serpvipvv 1 tv serpvipvv expriphipvi (2) ref ttα1refph,i ref g g ph i (3) 1 tnv refttβrefoc,v exp refttαrefsc,i r i (4) generally, pv panels are connected in a series in order to achieve usable voltage. partial shaded conditions is a critical issue in series-connected pv panels, because the shaded cells may get reverse biased and start consuming power instead of generating power. the resulting power loss in the shaded cells leads to local heating and creates thermal stress on neighboring panels. in a worst case scenario, the reverse bias across the shaded panels will exceed the breakdown voltage and get open circuited, and the entire array may be affected. in order to protect the panels, a bypass diode is connected across each panel, but these bypass diodes introduce multiple peaks in the characteristic curves, and it becomes essential to identify the global peak. in this paper, the effect of partial shaded conditions is analyzed and characteristic curves are obtained by considering a series of three connected panels along with a bypass diode (fig. 1). the solar panel was modeled using 1–4 in a matlab m-file (mathworks, mattick, massachusetts, usa). this can be extended to a number of solar panels (m) in a series and a number of solar panels (n) running in parallel (array size of m × n) with a different shading pattern (ramaprabha & mathur 2012). in this paper, a small pv array (size 3 × 1) was considered (fig. 1). the simulated characteristic curves were obtained for various insolation levels. the characteristic curve for ten different random shading profiles is shown in fig. 2. figure 1. electrical representation of three panels connected in a series with partial shading. 3. artificial neural network controller there are a number of mppt algorithms present in the literature, but in this paper, a neural network-based mppt algorithm was used. artificial neural networks (anns) are best suited to the approximation of non-linear systems. nonlinear systems can be exactly emulated by multilevel neural networks, which yield better results than other algorithms (torres 1998; alamoudi & zhang 2000; hiyama & kitabayashi 1997; aït cheikh et al. 2007). because of parallel computation of anns, the computation is fault tolerant (ie. deterioration of a few weights and/or missing links will not significantly deteriorate the output signal’s quality). in addition, anns have noise or harmonic filtering properties. the most commonly used feed forward multilayer back propagation network was chosen for this work. the architecture of the anns for this work is shown in fig. 3. the network is normally called a multilayer neural network and contains three layers: the input layer, the hidden layer, and the output layer. a neural network has to be configured such that the application of a set of inputs produces the desired set of outputs. various methods to set the strengths of the connections exist. one way is to set the weights explicitly, using a priori knowledge. another way is to train the neural network by feeding it r ramaprabha and s.p. chitra 18 figure 2. characteristic curves with partial shaded conditions. figure 3. a multilayer feed forward neural network. teaching patterns and letting it change its weights according to some learned rule. in the present study, the neural network was first trained in the offline mode which was provided with two inputs: a temperature and an insolation pattern. the desired output, or mpp, was first calculated practically for a particular insolation and temperature. temperature and insolation levels were measured with a thermocouple and a pyrometer. the data were stored in a memory module via a data logger (model dwr 8102, stack, ltd., bicester, uk). the experimental data for various insolation levels were computed by repeating the experiments in different weather conditions. the neural network was presented with different patterns of insolation and desired outputs for the corresponding pattern. its weights were adjusted so as to provide the desired outputs. in this paper, anns were trained with 250 data points. the training results obtained for the ann block using matlab are shown in fig. 4. at the end of the training, a feed forward multilayer back propagation ann network incorporated in this work was trained to detect the mpp under various insolation’s and temperatures. the maximum power extracted from the panel varied according to pv panels’ ages and the accumulated thickness of dirt and dust. the incorporated ann network can be trained to track performance degradation due to panels’ aging and the amount of dirt and dust that is accumulated on the panels by feeding the network with a degradation factor. the degradation factor can be determined with the help of the 2012 national comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 19 figure 4. training results of the artificial neural network (ann). figure 5. the artificial neural network (ann) with maximum power point tracking (mppt). renewable energy laboratory (nrel) report provided by jordan and kurtz. however a network might not be able to track performance variation due to panel damage as such a phenomenon is unpredictable. the mpp of the pv panel was tracked using a simple boost converter as a power conditioning unit. the response was improved by incorporating a proper pi controller. the initial values of the integral and proportional constants (kp and ki) were determined using the ziegler-nichols method. then the values were tuned to get the required response with the values of kp = 0.0002 and ki = 125, respectively. the controlled error was compared with a high frequency bipolar triangular carrier to produce pulse-width modulation (pwm) signals for the boost converter. the block diagram of ann-based mppt control used in this work is shown in fig. 5. 4. maximum power point tracking (mppt) in a centralized configuretion in a centralized configuration, the pv panels are connected in a series to a single power conditioning unit as in fig. 6. in this type of configuration, multiple peaks are observed in the characteristics due to the presence of bypass diodes. hence, it is required to identify the global peak. this can be achieved by the continuous scanning of characteristics. this approach to calculating the mpp is referred to as a scanned maximum power point tracking (smppt) system (itako 2012). in a smppt, a single offline trained ann controller is employed. the ann controller is fed with (2 × n) inputs, where n is the number of panels. the controller is trained to identify the global peak for any given set of temperatures and insolation values from the three solar panels. the ann controller, along r ramaprabha and s.p. chitra 20 figure 6. maximum power point tracking (mppt) in centralized architecture. figure 7. maximum power point tracking (mppt) in distributed architecture. with the pi controller, provides pulses to the power conditioning unit such that the maximum possible power is extracted. the training algorithm is quite complex since it has to settle down at an optimum point. a centralized configuration results in disproportionate losses. this is due to the fact that mppt in a centralized configuration becomes confused, stopping on a local maximum point and settling in a sub-optimal point of the voltage-topower configuration; second, the voltage point of comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 21 the mpp can be very diverse due to irregular conditions, going beyond the scope and voltage range of the centralized mppt. 5. maximum power point tracking (mppt) in a distributed configuration in the case of distributed architecture, each pv panel has separate power conditioning units and controllers, and the entire setup is connected in a series as in fig. 7. in such a case, ann controllers are employed to extract the maximum available power from a particular panel. each controller has only two inputs and calculates the mpp corresponding to each panel. as the number of inputs lessens, training of an ann controller becomes quite simple and fast. this makes the system simpler, modular, and stable (jiang et al. 2012) since the performance and control of one panel is effectively decoupled from the other. at the same time, there is no need to detect the global peak as in the case of centralized architecture. the control algorithm is also quite simple. distributed mppt can enhance and boost the performance of panels independently. one of the key advantages of having distributed plant architecture is the increased availability of a power plant, since a failure in a single direct current (dc)-dc converter or the disconnection or malfunction of a group of strings does not lead to a complete plant shutdown. another important factor to be considered is the number of converters employed in distributed architecture and their cost factor. moreover, the higher efficiency introduced in distributed architecture is greatly reduced by the static and dynamic losses of the power conditioning units as they are operated at a high switching frequency. 6. simulation results in order to verify the performance of the annbased mppt controller, a simulation was carried out on a three-series connected pv panel in the matlab environment. in each case, irradiance given to each panel was different from the neighboring panel. the simulink model (mathworks, mattick, massachusetts, usa) used for the smppt is given in fig. 8. the temperature and insolation levels for the three panels were continuously sensed by the single-trained mppt controller and provided the mpp for each pattern. the pwm technique was employed to vary the width of the pulse given to the power conditioning unit, such that it moved the operating point towards the mpp from each panel at any instant. the insolation patterns for study are shown in table 1. in pattern i, all of the panels were subjected to standard insolation and temperature to provide a base for comparison. apart from that, random patterns were chosen such that panel i was always subjected to standard insolation and temperature. panel ii was subjected to standard insolation with different temperature inputs to study the effect of temperature variation. panel iii was subjected to varied insolation and temperature patterns in order to study their combined effect on performance variation. the input power extracted from the pv panels and the power delivered to the load is presented in table 2. the simulated waveforms of the centralized architecture (fig. 8) with a scanned mppt controller for pattern v are shown in fig. 9. the matlab-simulink model of an mppt controller in a distributed environment is shown in fig. 10. in order to facilitate comparison the same patterns of insolation levels are presented to pv panels. in this case, three separate offline trained mppt controllers were employed to provide mpp for each panel. the maximum power from each panel was derived by separate power conditioning units using the pwm technique. the input power extracted and output power delivered by individual panels in a distributed environment is given in table 3. the simulated waveforms of a pv panel with a dmppt controller for pattern v are shown in fig. 11. from the simulated waveforms, it can be inferred that a panel with an irradiance of 600 watts/m2 is being by passed. r ramaprabha and s.p. chitra 22 figure 8. a simulink model of centralized architecture. table 1. patterns of insolation and temperature. pattern panel 1 panel 2 panel 3 g1(w/m2) t1oc g2(w/m2) t2oc g3(w/m2) t3oc i 1000 25 1000 25 1000 25 ii 1000 25 1000 38 1000 25 iii 1000 25 1000 25 800 38 iv 1000 25 1000 31.9 800 38 v 1000 25 1000 35.9 1000 25 table 2. input and output parameters in a scanned maximum power point tracking (smppt) system. pattern input parameters output parameters vpv (v) ipv (a) ppv (w) vo (v) io (a) po (w) i 48.77 2.287 111.53 63.40 1.732 109.80 ii 43.24 2.019 87.30 56.21 1.519 85.38 iii 42.06 1.965 82.65 54.67 1.481 80.97 iv 29.95 1.39 41.63 35.10 1.155 40.54 v 34.21 1.593 54.49 35.43 1.501 53.18 comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 25 figure 9. input and output waveforms of a scanned maximum power point training (smppt) system for pattern v. table 3. input and output parameters in distributed maximum power point tracking (dmppt) systems. pattern input side output side input power 1 ppv1 (w) input power 2 ppv2 (w) input power 3 ppv3 (w) total input power ppv (w) output power 1 po1 (w) output power 2 po2 (w) output power 3 po3 (w) total output power po (w) i 37.3 37.2 37.3 111.8 35.41 35.70 35.44 106.55 ii 35.8 18.1 35.8 89.70 34.12 16.34 34.12 84.58 iii 24.6 35.0 24.6 84.20 23.14 33.52 23.14 79.80 iv 19.6 1.1 23.1 43.80 18.51 0.04 21.98 40.53 v 25.3 1.26 29.7 56.26 24.06 0.06 28.42 52.54 23 r ramaprabha and s.p. chitra 24 figure 10. simulink model of distributed architecture. table 4. efficiency comparison in a simulated environment. pattern ƞ in centralized architecture (%) ƞ in distributed architecture (%) i 98.45 95.30 ii 97.80 94.29 iii 97.97 94.77 iv 97.38 92.53 v 97.59 93.39 table 5. devices used for hardware implementation. devices/components specification/values switches irfp450 diode in5408 inductor(e-core) 160 µh,200 µh filter capacitors 330 µf,220 µf frequency 20 khz comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 25 figure 11. input and output waveforms of distributed maximum power point tracking (dmppt) for pattern v. the efficiency of the mppt controller in both centralized and distributed configurations was calculated with data obtained from simulations and was then compared for different patterns. the results were tabulated and are presented in table 4. it can be observed that the power conversion efficiency of a smppt was around 4% greater than that of the dmppt. 7. experimental verification to verify the simulation results and to observe the characteristics of the mppt controller practically, a series string system was built. the string system was comprised of three 37.08 wp pv panels (solkar solar industry limited, chennai, tamil nadu, india). the control algorithm was implemented with a digital signal processor (dsp) controller tms320lf2407 (texas instruments, inc., houston, texas, usa) and an opto-coupler transistor circuit for isolation. a step down transformer and an integrated circuit (ic) regulator 7812 was used to supply power to the controller circuit. the pulses from the controller were given to the power conditioning unit. temperature and insolation levels were measured with a thermocouple and pyrometer via a stack data logger (model dwr 8102). the key system components used for hardware implementation are listed in table 5. the control scheme used in the hardware is shown in fig. 12. figure 13 shows the hardware set up of a smppt system, and fig. 14 displays the control circuit implemented for a smppt system. figure 15 shows the i-v curve for a 3 × 1 pv array with and without shading. the multiple steps in v-i characteristics lead to multiple peaks in v-p characteristics which are clearly identified in a partial shaded system. the shading is introduced artificially by covering part of the pv panels. r ramaprabha and s.p. chitra 26 figure 12. the control scheme of the hardware implemented. figure 13. hardware setup of a scanned maximum power point training (smppt) system. figure 14. control circuit of a scanned maximum power point training (smppt) system. comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 25 figure 15. i-v curve with uniform illumination and non-uniform illumination. figure 16. voltage and current waveforms in a scanned maximum power point training (smppt) system without shading. figure 17. voltage and current waveforms in a scanned maximum power point training (smppt) system with partial shading. figure 16 shows the waveforms for input pv voltage, pv current, output voltage, and output current with uniform insolation of g = 978 w/m2 and t = 31o c. the parameters observed were as follows: vpv = 48.8 v; ipv = 2.2 a; ppv = 107.36 w; vo = 68.01 v; io = 1.54 a; po = 104.73 w. figure 17 shows the waveforms of input pv voltage, pv current, output voltage, and output current with 27 r ramaprabha and s.p. chitra 28 partial shaded conditions in a smppt system. artificial shading was made by covering part of the two panels while one panel was allowed to receive the full insolation level. upon coverage, the conditions were, g1 = 978 w/m2; g2 = 682 w/m2; g3 = 682 w/m2 and t = 31o c. the parameters observed were as follows: vpv = 49.29 v; ipv = 1.57 a; ppv = 77.38 w; vo = 68.69 v; io = 1.07 a; po = 73.49 w, and the power conversion efficiency was η = 94.97%. to compare the performance of centralized and distributed structures, the same shading pattern was retained and the parameters or variables were observed using the hardware set up of the dmppt system, with three separate power conditioning units (fig. 18). figure 19 shows the waveforms of input pv voltage, pv current, output voltage and output current with partial shaded conditions in the dmppt system. the values observed were vpv = 49.01 v; ipv = 1.55 a; ppv = 75.97 w; vo = 68.53 v; io = 1.01 a; po = 69.20 w, and η = 91.08%. figure 20 exhibits the variation of the width of the control pulses produced by the dsp processor for different insolation levels and temperatures. figure 18. hardware setup of the distributed maximum power point tracking (dmppt) system. figure 19. input and output waveforms in the distributed maximum power point tracking (dmppt) system. comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 29 figure 20. pulse width variation for different insolations and temperatures. table 6. comparison of maximum power point tracking (mppt) controllers. insolation levels ( w/ m2) smppt dmppt g1 g2 g3 ppv (w) po (w) ƞ (%) ppv (w) po (w) ƞ (%) 845 736 721 68.66 67.17 97.83 70.59 66.35 93.99 625 736 571 46.02 44.74 97.22 48.03 44.56 92.78 735 640 345 43.33 42.17 97.32 44.51 40.52 91.04 645 420 305 23.59 22.67 96.10 24.95 22.47 90.06 724 720 680 61.9 60.46 97.67 63.85 59.82 93.69 the input and output powers of both smppt and dmppt systems were measured directly using a single phase clamp on a power quality fluke 345 clamp meter (fluke corporation, everett, washington, usa) for different shading patterns practically, and readings were tabulated with power conversion efficiency (table 6). table 6 demonstrates that the efficiency of the smppt system was higher (around 4–5%) than that of the dmppt system. the losses in the dmppt system are due to an increase in converter circuits which cause more switching and power conversion losses. 8. conclusions this work analyses the performance of anntrained mppt controllers in a centralized and distributed environment. in the past, it has been inferred that a smppt controller was more efficient than a dmppt controller. even though a dmppt harvests more power from the pv panel, the efficiency is reduced due to static and dynamic losses introduced by the converters. it must be kept in mind that a dmppt controller increases the modularity, stability, and reliability. hence, an mppt controller for specific applications has to be preferred. this work comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 30 suggests that a smppt controller is more appropriate for low power and non-critical loads like residential loads due to its low cost and high efficiency. in contrast, dmppt controllers are preferred for critical loads like hospitals, where stability and reliability are of concern. this work can be extended to improve the performance of a dmppt with interleaved soft switching converters. acknowledgment the authors wish to thank the management of ssn college of engineering, chennai, india, for providing all the experimental and computational facilities to carry out this work through internal funding. references aït cheikh ms, haddadi m, zerguerras a (2007), design of a neural network control scheme for the maximum power point tracking (mppt). revue des energies renouvelables 10(1):109– 118. al-amoudi a, zhang l (2000), application of radial basis function networks for solar-array modeling and maximum power-point prediction. ieee proceeding generation, transmission and distribution 147(5):310–316. de medeiros torres a, antunes flm, dos reis fs (1998), an artificial neural network-based real time maximum power tracking controller for connecting a pv system to the grid. proceeding of ieee the 24th annual conference on industrial electronics society 1:554–558. esram t, chapman pl (2007), comparison of photovoltaic array maximum power point tracking techniques. ieee transaction on energy conversion 22(2):439–449. femia n, petrone g, spagnuolo g, vitelli m (2005), optimization of perturb and observe maximum power point tracking method. ieee trans.power electron 20(4):963–973. herrmann w, wiesner w, waassen w (1997), hot spots investigations on pv modules—new concepts for a test standard and consequences for module design with respect to by-pass diodes. proceedings of the 26th ieee photovoltaic specialists conference 1129–1132. hiyama t, kitabayashi k (1997), neural network based estimation of maximum power generation from pv module using environment information. ieee transaction on energy conversion 12(3):241–247. jain s, agarwal v (2004), a new algorithm for rapid tracking of approximate maximum power point in photovoltaic systems. ieee power electron. lett. 2(1):16–19. jiang yc, jaber a, qahouq a, orabi m (2012), ac pv solar system distributed architecture with maximum power point tracking. ieee telecommunication energy conference 1–5. jordan d, kurtz sr (2012), photovoltaic degradation rates—an analytical review. nrel report number ja-5200-51664. karatepe e, boztepe m, colak m (2007), development of suitable model for characterizing photovoltaic arrays with shaded solar cells. solar energy 81(8):977–992. kasper m, bortis d, kolar jw (2014), classification and comparative evaluation of pv panel-integrated dc–dc converter concepts. ieee transactions on power electronics 29(5):2511–2526. kazutaka i (2012), new i-v characteristics scantype mppt control method for pv generation system. journal of technology innovations in renewable energy 1:87–91. koutroulis e, kalaitzakis k, voulgaris nc (2001), development of a microcontroller-based, photovoltaic maximum power point tracking control system. ieee trans power electron 16(1):46–54. masoum mas, dehbonei h, fuchs ef (2002), theoretical and experimental analyses of photovoltaic systems with voltageand current-based maximum power-point tracking. ieee transactions on energy conversion 17(4):514–522. patel h, agarwal v (2008), matlab-based modeling to study the effects of partial shading on pv array characteristics. ieee trans energy conversion 55(4):1689–1698. ramaprabha r, mathur bl (2008), modeling and simulation of solar pv array under partial shaded conditions. proc. of ieee int. conf. on sustainable energy technologies 12–16. ramaprabha r, mathur bl (2012), a comprehensive review and analysis of solar photovoltaic array configurations under partial shaded conditions. international journal of comparative analysis of maximum power point tracking controllers under partial shaded conditions in a ….. 31 photo energy, special issue on recent developments in solar energy harvesting and photo catalysis 1–16. silvestre s, boronat a, chouder a (2009), study of bypass diodes configuration on pv modules, applied energy 86(9):1632–1640. villalva mg, gazoli jr, filho er (2009), comprehensive approach to modeling and simulation of photovoltaic arrays. ieee transactions on power electronics 24(5):1198– 1208. walker gr, pierce jc (2006), photovoltaic dcdc module integrated converter for novel cascaded and bypass grid connection topologies design and optimization, ieee power electronics specialists conference 1–7. walker gr, semia pc (2004), cascaded dc-dc converter connection of photovoltaic modules. ieee transactions on power electronics 19(4):1130–1139. walker gs (2001), evaluating mppt converter topologies using a matlab pv model. journal of electrical & electronics engineering, australia 21(1):49–56. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 8.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 68-78 overdesign of villa structures in oman a.s. alnuaimi*a, m. al mohsinb, a. hagoc and s. el gamald a,c,d department of civil and architectural engineering, college of engineering, sultan qaboos university, oman. b department of civil and architectural engineering, college of engineering, university of buraimi, oman. received 24 march 2015; accepted 15 july 2015 abstract: residential projects represent the largest section of the construction industry in oman. it is widely claimed that designers in oman tend to unnecessarily overdesign structures, thereby aggravating construction costs. the aim of this research was to investigate the existence and extent of structural overdesign through the investigation of 31 villas which were approved by muscat municipality between 2000–2010. the specified villas were structurally redesigned based on the british code of practice using the same geometry and material strengths recorded in the drawings. the required reinforcement resulting from the redesign work was compared with the provided reinforcement as per the drawings. the results showed clear cases of extra reinforcement in almost all villas. the difference between required and provided reinforcements in the villas ranged from 2.3– 104.8%, with an average of 48.5% and a standard deviation of 24.0. the required and provided weights of reinforcement per square meter of the built-up area ranged from 25.7–71 and 40.9–87.9 kg/m2, respectively. the largest differences between the required and provided reinforcement were in the slabs, followed by the footings. a field survey of consultants and municipality engineers indicated that the most notable reasons for overdesign are the absence of authorized agents specialized in reviewing structural design, a lack of an engineering licensing system in oman, and the phenomenon of experienced engineers leaving design offices for better paying jobs elsewhere. finally, the cost implication of overdesign is minor compared to the overall cost of building. keywords: villa design, structural design, oman, municipality design approval. 3120002010 2.3104.848.5240 25.771040.987.9 1 23 4 . *corresponding author’s e-mail: alnuaimi@squ.edu.om a.s. alnuaimi, m. al mohsin, a. hago and s. el gamal 69 1. introduction with cost escalation in building construction, owners seek ways of reducing both the cost and time spent on providing decent residential buildings. designers should be responsible for providing a safe and cost-effective building to suit their clients’ needs. however, the objectives of maximizing the degree of technical performance and safety while minimizing cost are usually in conflict. every design code explicitly stipulates certain safety factors that have been determined to be adequate for loads and material strengths. overdesigning, or creating a building that goes beyond these limits, is a waste of resources as it provides unjustified levels of safety. a small number of researchers have examined overdesign practices occurring in the gulf cooperative council (gcc) countries. sadek et al. (2006) used american concrete institute (aci) provisions for the structural redesign of six typical villas in kuwait and stated that due to the absence of a unified national building code, larger variation in the quality of design practices exercised by private consultants confirms the nonuniformity of designs and structural overdesign of residential units when compared to aci codes. they found the average percentage of extra reinforcement to be 30% for slabs and 60% for beams and columns. columns were substantially oversized in terms of concrete dimensions as the carrying capacity of these elements as per the existing design was found to be higher than the actual applied loads by as much as 240% on average. the concrete material, steel strength, soil bearing capacity, and structure dimensions were maintained as per the existing design, keeping the reinforcement ratio as the only variable. arafah et al. (1999) studied problems related to reinforced concrete buildings due to the absence of a unified national code in saudi arabia, starting from the high cost of maintenance to complete structural collapse. they emphasized the need for a unified structural design code in saudi arabia to avoid discrepancies in design, maintain quality, and assure safety and serviceability. alnegheimish et al. (2002) studied the design of 41 residential buildings in riyadh, saudi arabia, to evaluate the design practices in the district’s small design offices. they concluded that gross and wasteful simplifications were common practice and emphasized the urgent need for a national code of practice for design and quality control. no information was found in the literature about the design practice for villas in oman; sadek et al. (2006) seem to be the only researchers investigating the presence and extent of overdesign in the gcc. residential projects represent the largest section of the construction industry in oman; ministry of economy (2011). table 1 shows the annual number of building permits issued during the period 20002010 in muscat; (muscat municipality (2011)), which is home to 40% of oman’s population of approximately 3.6 million. with this statistic in mind, the objectives of this study are to investigate the existence and extent of overdesigned villa structures in muscat, study the effect of such design on the total weight of reinforcement, determine the causes of overdesign in muscat’s villas, and propose solutions to the issue of overdesign. the significance of this study originates from the fact that it investigates the appropriateness of the design of villa-type housing, which is the most popular kind of accommodation in oman and the gcc. table 1. number of building permits issued during the period from 2000 to 2010 in muscat. location year mutrah al-amerat al-seeb baushar qurayat total 2000 118 96 715 456 90 1475 2001 110 136 835 436 72 1589 2002 131 197 1060 574 88 2050 2003 147 247 1169 565 110 2238 2004 118 247 894 545 115 1899 2005 124 333 924 553 127 2061 2006 142 430 1081 558 129 2340 2007 146 469 1234 574 142 2565 2008 196 812 2251 809 226 4294 2009 215 807 2228 767 198 4215 2010 224 869 2015 663 200 3971 total (2000-2010) 28697 overdesign of villa structures in oman 70 2. scope of work and methodology the study data gathered from muscat municipality was comprised of drawings of 31 villas approved for construction in various locations around muscat from 2000–2010. the drawings included architectural, structural, plumbing, electrical, and material information for construction purposes. the villas had been designed by different small to medium-size consultancy offices and lacked signs of the firms have used design software or a named code of practice. because the british code is widely used for structural design in oman, the villas were redesigned as part of this research using british code british standard (bs) 8110:97 (1997). a detailed structural analysis of all the villas was performed to find the bending moments, shear forces, and reactions of the supports using staad.pro structural analysis and design engineering software (2007) (bentley systems, inc., exton, pennsylvania, usa). the redesign followed bs8110 using reinforced concrete council (rcc) spread sheets (2012) for the design of slabs and footings and staad.pro for the design of beams and columns. the materials’ characteristic strengths and soil bearing capacity were kept as specified in the approved drawings, and the dimensions of the structural elements, geometry of the building and locations of footings were also maintained according to specifications. the weight of the materials and the dead load were calculated based on the type and thickness of finishes specified in the drawings. live loads were selected based on the bs6399-part 1 (1996) code requirements for different uses. the concrete density was 24 kn/m3 and the characteristic yield strength of the longitudinal and transverse reinforcement was 460 mpa. the concrete cover for reinforcement was also kept as directed in the approved drawings. the factors of safety for dead and live loads were 1.4 and 1.6, respectively. the only variable was the amount of reinforcement. random design checks were carried out for some elements using both manual calculations and based on the bs8110:97 code to ensure correctness of results. our use of the term “provided” refers to the amount of reinforcement provided in the drawings approved by the muscat municipality, while the term “required” represents the amount of reinforcement required, based on the redesign carried out as part of this research. the provided and required amounts of reinforcement for all structural elements forming the skeleton were compared. all the reinforced concrete slabs were either one-way or two-way with thicknesses ranging from 120–200 mm. all the beams were reinforced concrete plinth, downstand, and hidden beams. the width of the plinth beams was 200 mm and their depth ranged from 300–700 mm. the width of the downstand beams ranged from 200– 400 mm with a depth ranging from 320–900 mm. the width of the hidden beams ranged from 300– 1400 mm and their depth was from 240–420 mm. most of the columns used in the selected villas were rectangular. there were also some circular columns, mainly in the open halls. the rectangular dimensions varied from 200–1200 mm and the circular columns’ diameters ranged from 200–300 mm. all footings were pad or combined with lengths ranging from 0.9–3.4 m, widths ranging from 0.9–3 m, and thicknesses ranging from 0.3– 0.75 m. the length of the combined footings ranged from 1.6–6.4 m, the width ranged from 1.2– 3.8 m, and the thickness ranged from 0.3–0.75 m. the rcc spreadsheets were used to calculate the required reinforcement for the footings and slabs, while the staad.pro program was used for the design of the columns and beams. 3. results the required reinforcement that resulted from the redesign of the 31 villas was compared with the provided reinforcement given in the approved drawings. the percentage of difference between the provided and required reinforcement was calculated as (provided – required)/required) x 100) the characteristic concrete strength used ranged between 25–35 n/mm2, and the soil bearing capacity ranged from 150–250 kn/m2. the concrete cover to reinforcement of slabs and beams was between 25–30 mm, while for columns it was between 30–40 mm, and 50 mm for the footings. the comparison relates to individual elements and the total required and provided reinforcements for the whole villa. 3.1 slabs figure 1 shows the total required and provided reinforcements for each villa and makes clear that almost all the slabs were provided with more reinforcement than required. the total required and provided reinforcement for the slabs of the 31 villas were 159.9 tons and 273.1 tons, respectively. the percentage of difference ranged from -2.8– a.s. alnuaimi, m. al mohsin, a. hago and s. el gamal 71 162.0%, with an average of 75.6% and a standard deviation (sd) of 42.8. 3.2 beams figure 2 shows the required and provided beam reinforcements. it can easily be seen that all beams have more reinforcement than required. the total required and provided reinforcement for the beams of the 31 villas were 269.8 tons and 387.3 tons, respectively. the percentage of difference ranged from 0–102.91% with an average of 44.2% and a sd of 32.0. 3.3 columns figure 3 shows the required and provided reinforcements of columns. with the exception of one villa (v26), all the columns were provided with more reinforcement than required. the total required and provided reinforcement for the columns of the 31 villas were 180.5 tons and 246.63 tons, respectively. the percentage of difference ranged from -13.3–78.3%, with an average of 37.4% and a sd of 24.0. there were some cases of underdesign. 3.4 footings figure 4 shows the required and provided reinforcements of footings. apart from one villa (v30), all the footings were provided with more reinforcement than required. the total required and provided reinforcement for the footings of the 31 villas were 114.6 tons and 169.0 tons, respectively. the percentage of difference ranged from 0–182.4%, with an average of 57.6% and a sd of 45.1. 3.5 total reinforcement figure 5 shows the total required and provided reinforcements for all the villas; all were provided figure 1. required and provided slabs’ reinforcements for all villas. figure 2. required and provided beams’ reinforcements for all villas. overdesign of villa structures in oman 72 figure 3. required and provided columns’ reinforcements for all villas. figure 4. required and provided footings’ reinforcements for all villas. figure 5. comparison between total required and total provided reinforcement in all villas. a.s. alnuaimi, m. al mohsin, a. hago and s. el gamal 73 table 2. ratios of total amount of reinforcement to built-up area. villa no. number of floors total built up area (m2) ratio of required reinforcement (kg/m2) ratio of provided reinforcement (kg/m2) % dif. v 1 2 631.86 41.46 58.84 41.91 v 2 3 911.40 34.08 48.56 42.50 v 3 2 334.56 44.42 64.38 44.95 v 4 2 367.96 44.24 69.36 56.76 v 5 3 623.10 40.70 64.58 58.68 v 6 2 392 41.38 62.60 51.29 v 7 3 1036.8 32.89 45.66 38.83 v 8 2 626.24 39.86 71.12 78.45 v 9 2 350.68 42.03 69.41 65.13 v 10 3 707.40 34.80 62.77 80.34 v 11 3 1122.00 25.69 40.86 59.06 v 12 2 353.72 37.26 65.59 76.02 v 13 3 1078.35 34.74 71.15 104.81 v 14 3 630.00 34.76 52.29 50.41 v 15 3 697.89 33.56 62.39 85.91 v 16 3 810 40.00 67.16 67.90 v 17 3 725.16 35.99 54.77 52.18 v 18 3 597.6 34.77 52.51 51.01 v 19 2 719.94 53.06 75.56 42.41 v 20 2 225 62.49 82.13 31.44 v 21 3 698.72 70.96 87.59 23.44 v 22 2 430.00 44.84 57.72 28.73 v 23 2 447.40 44.48 62.00 39.40 v 24 2 285.84 42.12 74.94 77.91 v 25 2 305.38 43.34 69.37 60.07 v 26 3 846.45 48.18 55.49 15.16 v 27 2 298.33 44.95 58.74 30.69 v 28 2 476.65 47.17 58.75 24.54 v 29 2 300.60 44.61 51.20 14.77 v 30 2 290.17 42.05 44.76 6.46 v 31 3 525.89 45.09 46.12 2.30 average (kg/m2) 42.13 61.56 48.50 min (kg/m2) 25.69 40.86 2.30 max. (kg/m2) 70.96 87.59 104.81 stdv 8.54 10.89 24.00 with total reinforcement greater than required. the total required and provided reinforcement for the 31 villas were 724.8 tons and 1076.0 tons, respectively. the percentage of difference ranged from 2.3–104.8%, with an average of 48.5% and a sd of 24.0. 4. discussion of the results table 2 shows the ratio of the reinforcement-tounit in the built-up floor area of each villa. only two villas received less than 10% extra steel. this indicates that the reinforcement was underutilized but the cost and dead load were increased. at the same time, the bond between concrete and reinforcement may have been reduced due to insufficient space between bars. table 3 shows that the largest percentage of extra reinforcement was used in the slabs followed by the footings, and the lowest percentage was found in the columns. adding 75.6% extra reinforcement in the slabs aggravated the problem, as slabs naturally receive the largest amount of reinforcement. on the other hand, some columns were under-designed, which should be a concern because columns are critical for structural integrity and the possibility of failure should be remote. in fact, overdesign in columns should be considered acceptable for overdesign of villa structures in oman 74 structural reasons (zhao et al. 2004). additionally, columns receive a relatively low amount of reinforcement and concrete, resulting in only a negligible cost increase. it is clear that, in most cases, municipality approval almost guarantees the safety of structures under working load conditions, but the economy of the construction receives insufficient attention. the high sd for the extra reinforcement provided for individual elements indicates obvious variability in the design methods and codes used [table 3]. table 3. percentages of extra reinforcement in different members. slabs beams columns footings average of % dif. 75.63 44.17 37.42 57.61 standard deviation % dif. 42.84 31.96 24.03 45.06 max. value of % dif. 162.04 102.91 78.28 182.35 min. value of % dif. -2.83 0 -13.26 0 5. field survey results the results from the above study led to a field survey using a questionnaire prepared by the authors to explore causes of overdesigning and suggest solutions. the first part of the questionnaire covered general information about the respondents’ educational backgrounds, experience with design code of practice, and the software used in design. the second and third parts explored the possible causes of overdesign and suggested solutions, respectively. the causes and solutions associated with the issue of overdesign were suggested by the authors based on face-to-face discussions with 15 consulting and governmental engineers in the field of design and construction of villas in oman. the respondents were asked to rank the priority of each reason or solution using a five-point likert scale ranging from strongly disagree to strongly agree. first, a pilot study was conducted with five elements of each group. accordingly, the questions were modified to eliminate possibly confusing wording, dual definitions, or ambiguities prior to the actual study presented here. responses from 56 engineers in consulting offices and 24 municipality engineers were analyzed. the selected participants from the consultants and municipalities had a minimum of a bsc degree in civil engineering with at least three years of field experience. to maximize the spread of the distribution, the selection criteria allowed the participation of only one member from each consulting office and at most two engineers from each municipality office. the respondents from the consulting offices had a wide range of experience (≥5 years) in the design of villas and multi-story buildings, and educational degrees from different countries. the majority of them adopted the british code bs8110 (:1997) and/or the american code aci318:08 (2008) in their designs and used structural analysis software, especially staad.pro and/or sap. respondents from the municipalities had varying experiences (≥3 years) in studying design drawings of villas and multi-story buildings. the majority of them did not use any structural analysis program and relied on knowledge gained from experience in the municipality offices and training courses offered by their employers. the questionnaire was issued to and collected from the respondents by hand, with a face-to-face meeting to explain the questions, if required. microsoft excel software (microsoft, inc., redmond. washington, usa) was used to analyze the resulting data and produce descriptive statistics. tables 4 and 5 show the relative importance of causes of overdesign based on consulting offices’ and municipalities’ respondents’ opinions, respectively. the consultants emphasized the “need for a design review agent” as a top priority to ensure adequacy of design submitted by the consultants for approval. the other three most important reasons (reasons 2–4 in table 4) are related to professional licensing and the qualifications of designers who are allowed to practice structural design. it is interesting to note that these reasons were sequentially followed by “fear of structural failure” (reason 5 of table 4). this is possibly due to mistrust of workmanship at construction sites. the shortage of engineers in the municipality was a relatively important reason (reason 6 of table 4) for overdesign in omani villas. the majority of respondents from the municipalities were aware of possible structural overdesign in villas which the respondents justified with the statement “minor cost implication” (reason 1 in table 5). the third most important reason indicated for villa overdesign was weak site supervision. a.s. alnuaimi, m. al mohsin, a. hago and s. el gamal 75 table 4. relative importance (ri) of reasons of overdesign based on respondents from consulting engineers. no. priority of causes by consultants ri 1. there is no authorized agent that is specialized in reviewing the consultant structural design 0.704 2. non-availability of engineering licensing for engineers in oman to maintain the quality of consultancy services. 0.669 3. engineers who gain good experience leave design offices for better paying jobs 0.665 4. the local universities do not offer enough structural design courses 0.658 5. fear of structural failure 0.654 6. the engineers in municipality are few and can’t review each step of the design 0.638 7. the owners believe that more steel lead to safer building which influences the designer to increase the amount of steel 0.635 8. poor communications between consultant and municipality 0.631 9. municipality’s engineers ask for more steel than the codes require. 0.627 10 the engineers in municipality are not willing to take risk when it comes to reduce the amount of steel suggested by the designer. even if they know that the mount of steel is much more than what actually required. 0.627 11. municipality’s engineers do not have enough experience in visualizing the cost implication of overdesigned structures 0.619 12. the duration that it takes to get approval from the municipality is one of the common reasons behind the practice of possible over design as the consultants tend to increase the amount of steel to easily gain approval of the municipality 0.588 13. design error 0.588 14. the owner’s scope requirements during the design stage are unclear or not well defined 0.585 15. most of the consultants do not have collective experience about constructions in oman 0.581 16. some civil engineers are not specialized in structural design 0.565 17. diversity of codes’ requirements for load and material factors of safety 0.519 table 5. relative importance (ri) of reasons of overdesign based on respondents from municipalities’ engineers. no. priority of causes municipality ri 1. allow overdesign because the cost impact of overdesign is very minor compared to the overall cost of the building 0.8 2. engineers at the municipality are comparing the amount of steel provided in the drawings with the recommendation of the maximum and the minimum steel from the bs or aci codes without carrying out any calculation 0.765 3. minor over designing has its advantage to overcome problems of insufficient supervision by consultants 0.739 4. the overall contractor’s poor skills in the country have influence on the overdesign phenomena. as the designer tend to increase the amount of steel to offset any shortcoming of contractor 0.652 5. the grouping process of the structural elements is not accurate as the elements with less required steel are grouped with elements with highly required steel and that is to save some time in designing. 0.643 6. engineers at the municipality are checking the structural design based on their experience with similar structures and not based on formal code procedures 0.6 as far as the fourth ranked reason is concerned, the municipality engineers justified overdesign by blaming contractors’ poor workmanship. for the last item, the engineers admit that they have no formal code procedure for checking designs. tables 6 and 7 show the relative importance of solutions suggested by the authors as ranked by the consultants and the municipality respondents, respectively. the consultants assumed that the overdesign of villa structures in oman 76 table 6. relative importance (ri) of suggested remedies based on respondents from consulting offices. no. priority of remedies by consultants ri 1. a special agent shall be formed and authorized to approve structural designs consisting of highly experienced, qualified engineers (bsc, msc, phd) other than municipality. 0.781 2. the structural design of any structure should be checked and approved by a specialized engineer authorised by a special agent 0.777 3. no civil engineer shall be allowed to practice structural design without a professional license 0.777 4. a specialized structural engineer should be assigned to each large construction project 0.750 5. a common learning database system should be shared among all consultancy offices and municipality 0.746 6. a standard procedure should be established for all stages of project starting from design to completion and handover. 0.727 7. municipality should force all consultancies to use unified code of practice for structural design. 0.608 table 7. relative importance (ri) of suggested remedies based on respondents from muscat municipality. no. priority of remedies by municipality ri 1. no civil engineer shall be allowed to practice structural design without having a professional license 0.896 2. a specialized structural engineer should be assigned to design large projects 0.896 3. a special agent shall be formed to approve structural designs consisting of highly experienced, qualified engineers (bsc, msc, phd) other than municipality. 0.852 4. a standard procedure should be established for all stages of project starting from design to completion and handover. 0.85 5. a common learning database system should be shared among all consultancy offices and municipality 0.843 6. the structural design of any structure should be checked and approved by a specialized engineer 0.835 7. a national standard should be established to include all construction regulations and permits required for construction projects in oman. 0.817 8. a technical committee should be established to regularly review design procedures and manuals, and follow their implementation 0.774 9. municipality should force all consultancies to use unified code of practice for structural design. 0.765 most important solution for villa overdesign would be to “hire or recruit the services of an agent in-charge for reviewing structural design”. unlike the consultants, the municipality engineers gave equal priority to the importance of “attaining a professional license to perform structural design” and the “dedication of a specialized structural engineer to design large projects”. the third most important solution in the consultants’ ranking was in line with the first: a villa’s design should be reviewed by a specialized engineer authorized by the agent. the municipality engineers ranked as second the solution that a “specialized structural engineer should be assigned to design large projects”. the solution ranked fourth by the consultants was a legal and procedural issue to mandate that no design could be submitted to the municipality unless the designer is licensed by a supreme authority in oman. the municipality engineers indicated their belief that “a special agency should be formed to approve structural designs consisting a.s. alnuaimi, m. al mohsin, a. hago and s. el gamal 77 of highly experienced and qualified engineers other than municipality” would be the third most important solution. it is clear from the discussion of causes and solutions that the consultants and the municipality engineers had different opinions on the causes of overdesign; however, they shared almost identical opinions as to the solutions, albeit with different weights for each proposed solution. 6. concluding remarks this study focused on the structural redesign of 31 villas in the muscat area using the british code. the aim was to find the required reinforcement based on redesign work and to compare the results with the provided reinforcement in the approved drawings. based on an analysis of the results, the following concluding remarks can be made: most structural elements of almost all studied villas were structurally overdesigned, which was represented by the provided reinforcement being more than required by building codes. the extra reinforcement totaled 162% in the slabs, 103% in the beams, 78.3% in the columns, and 182.4% in the footings. the percentage of total extra reinforcement in the 31 villas ranged from 2.30–104.81%, with an average of 48.50% and a standard deviation of 24.00. a field survey of consultants and municipality engineers indicated that the most important reasons for overdesign in oman are an absence of authorized agents who specialize in reviewing structural design and the lack of an engineering licensing system. additionally, respondents indicated that engineers who gain good experience leave design offices for better paying jobs. in the same survey, respondents indicated the three most important solutions to overdesign in omani villas. first, the respondents expressed their belief that a special agency of highly experienced and qualified, non-municipality affiliated engineers should be authorized to approve structural designs. second, civil engineers should be banned from practicing structural design without a professional license. third, a specialized structural engineer should be assigned to design large projects. acknowledgment the authors are indebted to the building permit department at muscat municipality for providing the drawings and statistical data used in this research. the cooperation of the staff concerned is gratefully acknowledged and appreciated. assistance in data collection and structural design by al-mazrui k. m. and al kasbi f. a, airoon s. m, al kharousi h. h. al mujaini a. a. and aljabri w. z., is highly appreciated. references al-negheimish ai, shuraim ab and al-tayyar as (2002), structural design practice for residential buildings in riyadh: an overview. the 6th saudi engineering conference, king fahad university of petroleum and minerals, dhahran 3: 47-62. american concrete institute aci318 (2008), building code requirements for structural concrete and commentary. 318-08, 604 farmington hills, mi, usa. arafah a, al-zaid r, al-haddad m, ahmed ae, al sulaimani g and wafa f (1999), basic studies towards the development of a rc building code in saudi arabia. the arabian science and engineering 24(1): 3-18. loading for buildings, part 1 code of practice for dead and imposed loads bs6399-1, (1996), british standard institution, london. ministry of economy, sultanate of oman (2011), statistical year book no. 39. muscat municipality, directorate general of technical affairs (2011), report on building permits. sadek aw, al-fadala s and al-mutairi n (2006), improving the design of residential buildings in kuwait. emirates journal for engineering research 11(2): 59-65. spread sheets for structural design to bs8110, (2012), reinforced concrete council (rcc) british cement association. uk. retrieved from http://www.structural-engineering.fsnet. co.uk/rcc.htm. structural analysis and design software, staad.pro (2007), research engineering international, bentley systems, inc., exton, pennsylvania, usa. structural use of concrete—part 1 (1997), code of practice for design and construction. bs8110:97, british standard institution, london. zhao yg, ono t and yoshhara k (2004), an investigation on column overdesign factors avoiding weak story mechanisms of steel overdesign of villa structures in oman 78 structures. the 13th world conference on earthquake engineering. vancouver, bc, canada, august 1-6, paper no. 3448. c:\users\raoh\desktop\paper 3.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 33-41 semi interlocking masonry as infill wall system for earthquake resistant buildings: a review y. totoev*, a and a. al harthyb *a centre for infrastructure performance and reliability, university of newcastle, new south wales, australia. b civil and architectural engineering department, college of engineering, sultan qaboos university, oman. received 20 january 2015; accepted 30 august 2015 abstract: masonry is one of the most popular building materials. it has many excellent material properties and proven durability. over time, masonry structures have evolved from massive walls, which work mainly through compression, to more slender walls, which could also experience tension and shear. earthquake-induced tensile and shear stresses often exceed the capacity of traditional unreinforced masonry resulting in substantial damage and failure. a new mortar-less masonry system called semi-interlocking masonry (sim) is presented in this paper. it has reduced stiffness and susceptibility to damage and utilized a special method of interlocking bricks that allows relative sliding of brick courses in-plane of a wall and prevents out-of-plane relative movement of bricks. the paper reviews the research on sim to date and identifies new research opportunities in this area. keywords: masonry, wall, interlocking masonry, sim, infill. * corresponding author’s e-mail: yuri.totoev@newcastle.edu.au y. totoev and a. al harthy 34 1. introduction masonry is one of the most popular building materials. it has many excellent material properties and proven durability. over time, masonry structures have evolved from massive walls, which work mainly through compression, to more slender walls, which could also experience tension and shear. earthquake-induced tensile and shear stresses often exceed the capacity of traditional unreinforced masonry resulting in substantial damage and failure. reinforced masonry has better earthquake resistance; however, it is more expensive and requires expertise not always available in developing countries. the design of practical masonry with improved earthquake resistance still presents a challenge for structural engineers. oman and other countries in the arabian peninsula and arabian gulf are part of the arabian plate that comes into collision at its northeastern margin with the zagros mountains of iran and at the indian ocean in the east with the makran trench. these margins are marked by intense earthquake activities. for example, the 2008 bandar abbas earthquake measuring 6.3 on the richter scale causing seven deaths and 30 injuries occurred only 100 km from khasab, a town in the north of oman. the destructive earthquakes of sistan and baluchistan in 2013, bushehr in 2013 and bam in 2003 occurred, respectively, at 425 km, 550 km and 1200 km from khasab. the very destructive 2001 gujarat earthquake on the western indian coast occurred at about 1000 km from the coast of oman. earthquake monitoring in oman and countries in the gulf cooperative council (gcc) has started recently and earthquakes of even minor intensity have been felt and recorded (qamaruddin and al harthy 2000). as the occurrence of an earthquake in oman is probable, earthquake-resistant design provisions in building codes need to be prepared and used for buildings to resist earthquake loads, especially for important buildings such as hospitals, schools and telecommunication towers. buildings in oman and many parts of the middle east are made of reinforced concrete frame with infill masonry walls made of concrete blocks which, if properly designed and detailed, can resist moderate earthquake loads. the first author of this paper has developed a new masonry system called semi-interlocking masonry (sim). it has reduced stiffness and susceptibility to damage and an increased capacity to dissipate earthquake energy as compared with traditional masonry. the main objective of this paper is to introduce sim panels as an option for an earthquake-resistant framed masonry building. the paper will also summarize the research on sim conducted to date, describe benefits of and problems with sim and identify new research opportunities. 2. semi-interlocking masonry (sim) sim is an innovative building system for mortarless walls which utilizes a special method of interlocking sim bricks that allows relative sliding of brick courses in-plane of a wall and prevents out-of-plane relative movement of bricks (australian patent application no. 2010905681, 2010). two different methods of semi interlocking have been developed: using specially shaped bricks, or topological sim and using conventionally shaped bricks with special perforations and dowels, or mechanical sim (fig. 1). traditional brick molding technology can be easily adopted for making topological sim units. mechanical sim units are designed to utilize existing brick extrusion technology. the structural performance of these two sim types is essentially identical (wang et al. 2014). topological sim, however, appears to have better resistance to water penetration (forghani et al. 2014). several possible structural and non-structural applications of sim include infill panels in multistorey frame structures, walls in confined masonry structures, masonry skins of a reverse brick veneer system, robotically prefabricated masonry walls, and do-it-yourself (diy) masonry. there are many different interlocking brick/block masonry systems on the market. they have all been developed to build structural or nonstructural walls without mortar. some of them are dry set like sim; others use various adhesives to bond units into a monolithic wall. the main semi interlocking masonry as infill wall system for earthquake resistant buildings: a review 35 figure 1. different methods of semi interlocking: a. topological and b. mechanical (totoev et al. 2014). difference of sim is that, unlike all of these systems, it avoids connecting units into a monolith. in fact, its purpose is quite the opposite: it makes walls pliable and deformable. to better understand the novelty of sim, recall the definitions of a structure and a mechanism. a structure is a body or an assembly of bodies that form a system capable of supporting loads. a mechanism is an assembly of moving parts capable of performing a complete functional motion. sim is designed for the relative motion of bricks without necessarily supporting loads. therefore, some sim walls, including infill panels, are not structures but energy dissipating mechanisms. this paper discusses the use of sim as infill panels in multistory frame structures. 3. structural application of sim as infill panels semi interlocking masonry (sim) is an innovative building system that uses engineered mortar-less masonry panels to improve energy dissipation of frame structures during earthquakes. the energy dissipation occurs through friction between bricks as they engage in relative sliding by the frame vibrating during earthquake. sim panels are designed for inclusion in new earthquake-resistant structures as well as seismic rehabilitation or retrofitting of existing structures. traditional masonry infills are either architectural walls or structural panels designed to brace frame structures. they are not intended for energy dissipation. energy dissipation in these infills during earthquakes mostly relates to micro and macro-structural cracking and the plastic behavior of materials. the capacity of traditional infills to dissipate energy in this way before failing is quite limited. the novel purpose of sim infill panels is to provide frame structures with artificially added damping. in sim panels, energy dissipation occurs mostly through friction between bricks of the panel. sim is a unique system, which utilizes masonry infills as effective energy-dissipation devices (edd) to improve the earthquake resistance of frame structures. superficially, a sim infill looks like any other masonry infill panel. however, it is conceptually different from all other masonry infill types. consider the classical equation of motion for a structure under earthquake load to demonstrate this difference: (1) where is the vector of dynamic displacements (vibrations); is the vector of velocities, is the vector of accelerations, is the acceleration of the ground, is the stiffness matrix, is the damping matrix, and is the mass matrix. all common types of masonry infills structurally are various forms of frame bracing. they minimize vibrations mainly by increasing the stiffness of the structure represented in eq. 1 above by the stiffness matrix. often this is achieved at the expense of lowering the yield displacement and displacement ductility of the structure. sim infills also aim to minimise vibrations but in a different way. being energy dissipation devices, they achieve this objective by changing the damping matrix without detrimental effect on the yield displacement and displacement ductility of the structure. the invented system in 2010 (australian patent application no. 2010905681, 2010) was first introduced in print in 2011 (lin et al. 2011). various elements of it are not new. in fact, one could trace its heritage to the dry set stone masonry of the mesolithic era with elements of interlocking such as the mortise-and-tenon joints of stonehenge. y. totoev and a. al harthy 36 another ancient example of topologically interlocking masonry is the multifaceted stones of machu picchu. ancient egyptians, romans, incas, and khmers used metal masonry block connectors. slotted holes are very common in steel construction for relative sliding of connected parts. the concept of a masonry wall designed not as a monolith structure but as a mechanism where bricks slide against each other is entirely new, however. sim does not possess self-recentering capability. therefore, it could exacerbate the residual distortion of the frame after earthquakes. it would be advisable to combine sim panels with other structural systems to reduce possible residual distortion. 4. previous research on dry stack masonry some research has been done previously on dry stack masonry. lourenco with his colleagues performed a series of tests (lourenco et al. 2004; lourenco and ramos 2004) and concluded that the failure criteria of dry stack stone is a type of mohrcoulomb failure. a number of cyclic tests and shaking table tests both on dry stack stone and mortar stone wall were also carried out (lourenco et al. 2005; vasconcelos et al. 2006; vasconcelos and lourenco 2009). from those tests, the type of wall boundary conditions and the vertical compression level were confirmed as two important factors for the failure mode. considerable nonlinear deformations were attained (storey drift of 2.5%). however, because of the rocking failure mechanism, it was concluded that unframed dry stack walls were unable to dissipate energy. uzoegbo (2003) have researched both the inplane and out-of-plane seismic behavior of dry stack masonry wall (uzoegbo et al. 2003; 2004). according to this research, the strength of dry stack units does not make a significant difference in the resistance to lateral loads; the interlocking and friction between units govern the lateral load bearing capacity. the compressive strength of the panel is directly proportional to the strength of masonry units. they also observed rocking of the dry wall before failure. a shake table test was conducted on the dry stack system, which demonstrated that the dry stack masonry structure could resist ground acceleration of up to 0.3g (uzoegbo and senthivel 2009). all previous research on framed dry stack masonry infills was conducted at the university of newcastle in australia and harbin institute of technology (shenzhen graduate school). 5. research results on sim 5.1 types of sim infill panels a narrow gap between the top of a sim panel and the frame girder is difficult to avoid during construction of panels within the frame. special packing should be used when this gap is undesirable. the presence of this gap and its width play a key role in the structural response of sim panels to earthquake-induced vibrations. there are three main types of sim panels: sim with open gap. this type of sim panel is built hard against the columns but has a gap between the top of the panel and the girder (fig. 3a). the frame interacts with the sim panel only in trough columns. the gap does not close during earthquake-induced vibrations. assuming a sin shape for deforming columns, the critical gap width dgap can be calculated as illustrated in fig. 2. 4796 2sin 29 3 37 45 )(sin0 x xxl x (2) 82.3 4796 0 37 45 )(sinl (3) ult ultult ababgap hld 34.022 82.3 (4) where in eq. 4, δult is the ultimate storey drift, lab is the length of distorted column ab, and hab is its height. for a sim infill panel with the gap always open, its width must conform to the following conditions: (5) frame girders never clamp a sim panel of this type in a vertical direction. it provides mainly energy dissipation to the structure. its strengthening effect is limited to the maximum friction force developed on the bead joints of the panel due to self-weight. sim without a gap. there are no gaps between this type of sim panel and the frame. it is in contact with the girder as well as columns (fig. 3b) (6) therefore, panels are clamped between girders at all amplitudes of vibrations. this has the dual effect of providing some bracing to the frame semi interlocking masonry as infill wall system for earthquake resistant buildings: a review 37 figure 2. estimation of the critical gap width above a sim infill panel. a b figure 3. different types of sim panels displaying a) with an open gap and b) without a gap. through the diagonal clamping zone and providing a higher level of energy dissipation compared to the previous type of sim panel due to higher compression/friction on the bead joints. sim with a closing gap. this is a combination of the first two types. it has a very narrow gap between the top of the panel and the girder (7) this type of sim panel provides mainly energy dissipation to the structure during small amplitude vibrations when the gap remains open. however, as the amplitude increases, the gap closes, the clamping is activated, and the panel begins to provide additional bracing to the frame as well as higher energy dissipation. vertical post-tensioning of sim panels through regular perforations in bricks could have several benefits to structural performance. to illustrate these let us consider the sim with open gap type panels. post-tensioning would increase compression on the bead joints and, therefore, friction between bricks; increase the bracing capacity of the panel; allow the design of a sim panel for a certain slip load, and improve out-ofplane stiffness and the strength of the panel. 5.2 experimental work initial tests on sim have included compressive tests on sim units and sim prisms. cyclic friction tests on sim triplets (lin et al. 2012) [fig. 4a] were performed on prototype mechanical sim units using a modified triplet shear test (i.o.f. pren1052-3, 1993). an average friction coefficient of 0.66 was determined for concrete units at a clamping stress of 0.1–0.5 mpa. however, at higher levels of clamping stress, the friction coefficient reduced to 0.55. this value is recommended for analysis as it is more conservative. in-plane cyclic displacement tests were performed on the full-scale reduced size rc frame infilled with the prototype mechanical concrete sim with a closing gap panel (2 x 2 m; type 3; 227 x 113 x 80 mm concrete units) [fig. 4b]. detailed results are reported in kun et al. (2011) and lin et al. (2011a). these tests identified three main response mechanisms for a frame with a sim infill panel: constant friction response, the mohr-coulomb response, and a plastic response. in-plane cyclic displacement tests were performed on the full-scale reduced size steel frame infilled with two different sim panels: that with an open gap (type 1) and that without a gap (type 2). both measure 2.4 x 2.4 m and are made of 230 x 110 x 76 mm topological concrete units [fig. 4c]. these tests confirmed in-plane response mechanisms. an outof-plane monotonic airbag test was performed after the last cyclic test on the same panel [fig. 4d]. the out-of-plane displacement capacity of the square type 2 sim panel was more than 1.5 times the thickness of the panel. detailed results have been reported by totoev and wang (2013). all in-plane y. totoev and a. al harthy 38 a b c d figure 4. experimental tests on sim, with a) displaying a friction test, b) an in-plane cyclic test on a type 3 panel, c) an in-plane cyclic test on type 1 and 2 panels, and d) an out-of-plane test. cyclic tests also have provided experimental data for calculation of the frictional energy dissipation and estimation of damping. 5.3 numerical modelling numerical modelling of sim panels was done using the microstructural approach with displacement analyzer (diana) finite element (fe) software (informer technologies, inc., walnut, california, usa) (witte and kikstra, 2002) and the super-element approach through the seismostruct fe program, version 6.5 (seismosoft, ltd., pavia, italy). both models were verified using experimental results described above. seismostruct was selected for numerical simulations for multi storey frames as it is a more practical program. four fe models were created for the tree bay four story rc frame, including: 1. rc frame without infill panels, 2. rc frame with type 1 sim infill panels. 3. rc frame with type 2 sim infill panels, and an 4. rc frame with traditional urm infill panels. the first numerical simulation was the nonlinear response due to monotonic load (pushover analysis). this simulation determined the yield, ultimate, and structural displacement ductility for all models. the second numerical simulation was the response history analysis under synthetic earthquake ground motion (totoev et al. 2014). figure 5 shows selected results of these numerical simulations. semi interlocking masonry as infill wall system for earthquake resistant buildings: a review 39 a b c d figure 5. numerical simulations of a) a fe model, b) pushover results, c) synthetic ground acceleration, and d) response history analysis results (totoev et al. 2014). 6. design implication of sim infill panel. a sim infill panel is a kind of a wall and as such must be designed to perform typical architectural functions, providing shelter and security, and dividing internal space. properties such as sound insulation, fire resistance, thermal conductivity, and water penetration, however, are beyond the scope of the following discussion, which will be limited to the engineering design of the sim infill panel. it is important to stress that a sim infill panel is not a frame bracing structure but a mechanical energy dissipation device and must be designed accordingly. a good design procedure for passive energy dissipation devices like a sim panel is outlined in asce fema 356 (2000). this document also provides recommendations for testing such devices. the damping effect afforded by a sim panel could be calculated as follows: (8) where ζef is the effective damping in the structure, ζframe is the hysteretic damping in the frame only (typically taken as 5%), ζ*sim is the equivalent viscoelastic damping for sim panels (about 8.5%) (totoev and lin 2012), ufsim is the frictional energy dissipation in all sim panels during one cycle of vibrations at designed target displacement, and usframe is the maximum strain energy in the frame. in the standard procedure of direct displacement-based design, increasing the effective damping of the structure would increase its effective stiffness for the same target displacement. this would allow savings to be made by reducing cross sections and reinforcement of the frame. 7. summary of the structural benefits of and problems with sim infill panels sim is a new system which has not yet been used in practice. the majority of the structural benefits of y. totoev and a. al harthy 40 sim listed below were observed during laboratory experiments. high frictional energy dissipation, high yield displacement and displacement ductility, improved out-of-plane structural integrity, discontinued stress field and, therefore, low tensile stress, no shear failure mode—sim is designed to slide, reduced danger of compressive failure—sim (type 1) is not load bearing, no crack propagation from damaged units (therefore, localized damage), not dependant on the compressive strut action; therefore openings are not critical. sim also has properties, which are of concern. these properties are mostly related to the mortar-less nature of sim and are perceived rather than observed: consistent dimensions of sim units were difficult to achieve during prototype unit making in the laboratory conditions. it is expected to be a problem during commercial manufacturing too. as a result, a sim panel may have gaps up to about 2 mm wide at the joints; the water and air tightness of sim’s joints is expected to be higher than in traditional masonry and in some regions could be of concern; sim does not possess re-centering capabilities. while the slip friction on the joints contributes to the energy dissipation, the static friction could be a cause of small residual structural distortion after an earthquake. 8. further research directions new opportunities in this research area naturally relate to the sim’s problems listed above: it would be useful to find/develop a gap filler for sim joints which could improve the buildability of sim and its water and air tightness without bonding units into a rigid composite. following the previous point, the friction capacity of joints with gap fillers will need to be reassessed. the confining frame could assist with the recentering of sim panels after an earthquake. a special study would be required to evaluate and compare the strain energy in the frame to the static friction in the sim. 9. conclusion in this paper, a review on experimental and numerical studies carried out on semi interlocking masonry (sim) used as infills wall system for earthquake resistant buildings has been reported. the role of the gap between the sim wall and the frame girder on the bracing of the building is explained. the paper summarizes the structural benefits and problems of sim infill walls. references american society of civil engineers (2000), fema 356 prestandard and commentary for the seismic rehabilitation of buildings. asce for the federal emergency management agency, washington, d.c. australian patent application no. 2010905681, (filing date dec. 24, 2010), mortarless masonry (newcastle innovation limited, applicant: totoev, yuri z., inventor). forghani r, totoev yz, kanjanabootra s (2014), experimental investigation of the water penetration through semi interlocking masonry (sim) walls. proc. of the annual meeting of architectural institute of japan, kobe, japan 889– 890. international organization for standardization: standardization, i.o.f., pren 1052-3 methods of test for masonry: determination of initial shear strength. 1993. lin k, liu h, zarevich ty (2012), behavior of mortar-less masonry joint under cyclic shearcompression loading. research report 44(008), harbin university 6–10. lin k, totoev yz, hong jl (2011), in-plane cyclic test on framed dry-stack masonry panel. advanced material research journal, 163(167): 3899–3903. lin k, totoev yz, liu h (2011a), energy dissipation during cyclic tests in framed dry stack unreinforced masonry panels. proceedings of the 9th australasian masonry conference, queenstown, new zealand 205–214. lourenço pb, oliveira dv, roca p, orduna a (2005), dry joint stone masonry walls subjected to in-plane combined loading. journal of structural engineering 131(11): 1665–1673. semi interlocking masonry as infill wall system for earthquake resistant buildings: a review 41 lourenço pb, ramos lf (2004), characterization of cyclic behavior of dry masonry joints. journal of structural engineering 130(5): 779–786. lourenço pb, ramos lf, vasconcelos g (2004), on the cyclic behaviour of stone dry masonry joints. proceedings of the 13th international brick/block masonry conference, madrid, spain 1049–1058. qamaruddin m, al-harthy a (2000), earthquake hazard potential in oman. proc. 12th world conference on earthquake engineering 12wcee, paper 1943, auckland, new zealand. seisomosoft. seismostruct, version 6.5 (2013), a computer program for static and dynamic nonlinear analysis of framed structures.” pavia, italy. totoev yz, lin k (2012), frictional energy dissipation and damping capacity of framed semi-interlocking masonry infill panel. proc. 15th international brick/block masonry conference, florianapolis, brazil. totoev yz, wang z (2013), in-plane and out-ofplane tests on steel frame with sim infill, proc. 12th canadian masonry symposium, vancouver, british columbia, canada. totoev yz, williamson d, wang z (2014), vibrations of multi-storey rc frame with sim panels; numerical simulation. proc. 9th international masonry conference, guimaraes, portugal. uzoegbo hc (2003), in-plane and out-plane behaviour of dry-stack block walling systems. proc., the ninth north american masonry conference, clemson, south carolina. uzoegbo hc, senthivel r (2009), an effective retrofitting system for dry-stack masonry subject to seismic loading. the masonry society journal 22(3): 71–84. uzoegbo hc, senthivel r, ngowi, jv, (2004), loading capacity of dry-stack masonry walls. the masonry society journal 25(1): 41–52. vasconcelos g, lourenco pb (2009), in-plane experimental behavior of stone masonry walls under cyclic loading. journal of structural engineering 135(10): 1269–1277. vasconcelos g, lourenço pb, mouzakis h, karapitta l (2006), experimental investigations on dry stone masonry walls. proceedings of the 1st international conference on restoration of heritage masonry structures, cairo, egypt. wang z, totoev yz, lin k (2014), experimental study on rc and steel frames with sim infill. proceedings of the 9th international masonry conference, guimaraes, portugal. witte fcd, kikstra wp (2002), diana user's manual. california. p. 38. c:\users\raoh\desktop\tjer_vol.12(2)_2015_pdfs_press\paper 6.xps the journal of engineering research (tjer) vol. 12, no. 2 (2015) 51-59 loss reduction in isolated rural area distribution network using photovoltaic system m. h. albadi*, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, muscat-123, sultanate of oman received 21 july 2014; accepted 30 june 2015 abstract: this paper investigates the utilization of photovoltaic (pv) systems in a rural area electricity network. the rural area electricity company of oman is planning to install a 304 kw pv system at the powerhouse of the almazyonah network. based on the available network data, a power flow model of the system under study was built and the system’s performance was studied. the location and installed capacity of the planned pv system were optimized based on loss reduction. keywords: loss reduction, rural area networks, power flow, photovoltaic systems. 304 *corresponding author’s e-mail: mbadi@squ.edu.om m. h. albadi, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri 52 1. introduction natural gas-based generation facilities are used to generate electricity in the main interconnected system (mis) of oman. however, as oman is a thinly populated country, connecting all remote rural areas to the mis is not economical. therefore, diesel generators are used to provide electricity for remote rural areas. this service is provided by the rural area electricity company saoc (raeco) under a license issued by the authority for electricity regulation (aer) to generate, transmit, and distribute electricity in these remote rural areas. as of 2014, more than 45 such diesel-based isolated rural networks are in operation. a renewable energy study published by the aer in june 2008 recommended the immediate implementation of renewable energy-based generation facilities in rural areas. two years later, the aer announced six renewable energy pilot projects to be implemented within raeco networks. potential benefits from these projects include economic benefits in the form of avoided costs of diesel generation, the transfer of knowledge and experience to oman, and capacity building within raeco and other omani companies (authority for electricity regulation, 2015). by the end of 2012, the aer had issued new regulations to encourage the implementation of renewable energy projects in oman’s rural areas. however, according to the aer, the implementation of renewable energy pilot projects has been delayed because of two significant barriers. the first barrier is the absence of a policy framework and policy instruments to encourage and support the economic deployment of renewable energy projects. the second barrier is the fossil fuel subsidies, which make renewable energy-based generation appear more expensive in comparison with fossil fuel-based electricity (authority for electricity regulation, 2015). one of the renewable energy pilot projects approved by the aer was a photovoltaic (pv) system in almazyonah in the dhofar governorate. according to raeco, the power purchase agreement of the 304-kw pv project was finalized in 2013 (rural areas electric company, 2015), and the first rural area pv system came online in may 2015. solar pv systems are classified as renewable energy-based distributed generation resources. despite the benefits of pv systems and other renewable energy-based distributed generation (albadi, bourdoucen 2011; elhadidy and shaahid 2009), their presence in distribution systems alters the power flow. consequently, the operation of pv systems may have technical and economic impacts on distribution systems. therefore, it is important to optimize the location and size of any added pv system (atwa and el-saadany 2010; abu-mouti and el-hawary 2011). power losses are proportional to the resistance of the power path and the loading of the line. unlike central plants, distributed generation (dg) is located near consumers; therefore, the resistance in the current path is much lower than that between central plants and electricity consumers. moreover, on-site dg can reduce the loading of transmission lines and relieve heavily loaded lines, leading to more reduction in power loss. in kai et al. (2012), a distribution system expansion planning strategy encompassing renewable dg systems with schedulable and intermittent power generation patterns is presented. in chen and duan (2014), the proposed cost function of the optimization problem included two cost components: operation and investment. a strategy for dg allocation in radial distribution networks under uncertainties of load and generation using an adaptive genetic algorithm (ga) was presented by ganguly and samajpati (2015). it has been demonstrated that loss reduction depends on dg penetration level, location, and reactive power control capability (albadi and el-saadany 2008; attwa et al. 2009). in this work, the location and size of the pv system was optimized to minimize losses using ga for the first pv pilot project installed in raeco networks. after this introduction, the paper proceeds with presenting network data from the almazyonah site. the model of that network is presented in section 3. the simulation results of the network performance with and without the pv system are presented and discussed in section 4. finally, a summary of the main conclusions is presented in section 5. 2. system data the almazyonah network consists of nine diesel generators (3 2-mw, 3 1-mw, and 3 1-mw mobile engines) and 51 transformers [table 1]. there are five sizes of 11 kv/415 v transformers [table 2]. detailed loading on these transformers and per unit line data are presented in the appendices a and b, and a single-line diagram is presented in fig. 1. loss reduction in isolated rural area distribution network using photovoltaic system 53 2 1 tx 57 11 13 14 tx 19 65 99 tx 11 100 47 48 tx 58 101 tx 29 tx 14 103 49 50 tx 30 104 51 tx 26 105 tx 59 107 52 tx 2 106 tx 40 95 33 34 tx 55 96 35 tx 20 97 36 tx 41 98 tx 19 58 4 tx 21 57 5 tx 51 60 tx 3 61 6 tx 22 62 tx 4 63 8 3 tx 10 59 10 7 9 tx 15 64 tx 12 66 15 tx 18 68 16 tx 25 67 tx 53 69 tx 52 70 tx 36 71 tx 56 72 20 22 17 18 19 21 tx 46 73 12 75 74 24 23 tx 44 tx 47 tx 9 25 26 76 27 tx 23 77 28 tx 16 78 tx 24 79 29 32 tx 54 30 81 tx 42 31 80 tx 39 82 37 38 tx 45 83 42 43 tx 27 84 tx 28 85 39 40 tx 17 86 87 tx 38 41 89 45 90 46 44 tx 35 tx 34 tx 33 tx 47 91 53 54 tx 48 92 93 55 tx 49 56 94 tx 50 102 figure 1. almazyonah single-line diagram. table 1. network data. component quantity diesel generators 9 transformers 51 busses 107 oh lines 43 cables 14 measured peak (kw) 1,823 measured peak (kvar) 1,136 table 2. transformers’ data. kva quantity x/r % z% 1,000 9 5.57 4.75 500 5 5.1 4.75 315 9 3.97 4.75 200 13 3.37 4.75 100 15 2.32 4.75 in this study, the proposed pv plant is modelled as load bus with a negative output. the output of the pv system coincides with the system peak load at which the power flow analysis was conducted. in this study, a 3% derating of the pv system output is assumed to consider the inverter efficiency and dust effect, and the output at peak load is assumed to be 295 kw. different derating factors for a 5 mw plant are discussed in al-badi et al. (2011). in the current study, for the purpose of testing the actual pv system performance in local weather conditions, two types of pv technologies were used: thin film and crystalline technologies. m. h. albadi, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri 54 3. system modeling 3.1 power flow model power flow analysis is widely used in power system operation and planning. the power flow model of the system can be built using relevant networks, loads, and generation data. outputs of the power flow model include voltages at different busses and line flows in the network. these outputs are obtained by solving power balance equations as follows: ∑ | | cos , (1) ∑ | | sin , (2) where, | | and are the magnitudes of the voltage at bus and , respectively; and are the associated angles; is the magnitude of the y-bus element between the two busses; and is the corresponding angle. these power balance equations are nonlinear; therefore, iterative techniques such as the newtonraphson, gauss-seidel, and fast-decoupled methods are commonly used (al-badi et al. 2011). in this work, the network model was built using the matlab load flow toolbox (mathworks, natick, massachusetts, usa). input data consisted of two input matrices: a bus data matrix that shows each type of bus, voltage magnitude of voltage controlled busses, and generation and load data; and a branch data matrix that gives information on all connections between busses, including branch impedances. the system losses can be calculated once the power flow problem is iteratively solved. for example, the losses in line i-j are the algebraic sum of the power flow as below: (3) where, ∗ and ∗ the genetic algorithm (ga) toolbox in matlab was used to study the effect of installing the pv system on the total system losses (al-badi et al. 2011). 3.2 genetic algorithms (gas) gas are global search heuristics that are based on the mechanics of natural genetics (al-badi et al. 2011; albadi and el-saadany 2008) and use techniques inspired by evolutionary biology. examples of these techniques are inheritance, mutation, selection, and crossover. ga maintains a population of individuals that represent the candidate solutions to the given problem and has been used to obtain high-quality solutions for many optimization problems in science and technology. the ga process starts with an initial solution from a population of randomly generated individuals [fig. 2]. each individual represents a candidate solution to the optimization problem. the fitness of every individual in the population is evaluated in each generation. the reproduction operator is the first to be applied, which follows darwinian theory. selecting an individual in the new population depends on its fitness value, which in turn depends on the fitness function, which is problem specific. in this study, the total power loss in the system is defined as the fitness function. the survival of the fittest process encourages the propagation of strong individuals. however, this does not necessarily produce better individuals. to find stronger ones, the crossover procedure is used. this is achieved by mating reproduced individuals, therefore combining their features. mutation is used to improve the population diversity and can help in cases where an individual representing a suboptimal solution dominates the population. in such cases, mutation is used to converge to a suboptimal solution by changing the values of the binary digits at random. it is worth mentioning that mutation rates should be kept low to prevent damaging good solutions. ag can be terminated when the fitness level is satisfactory or the limit of the maximum number of generations is reached. 3.3 identification of candidate connection busses to find out the optimum location or the point of connection of the pv system in the almazyonah network, a total of 107 network buses are available. these include both the medium (11 kv) and low voltage (415 v) nodes. because of the size of the pv system, it is connected to the 11 kv network. as a result, 56 medium-voltage nodes remain as candidate connection busses after excluding the low voltage busses. looking at the voltage profile of the system before connecting a new dg, the pv system, in this case, busses that are far away from the source (the power house), has a low-voltage profile compared with the upstream ones. therefore, considering the voltage profile resulted in reducing the number of candidate locations to 10. loss reduction in isolated rural area distribution network using photovoltaic system 56 figure 2. genetic algorithm (ga) flow chart. figure 3. distribution of power losses in the system. 55 m. h. albadi, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri 56 4. results and discussion 4.1. system performance without pv system the calculated real power losses in all transformers’ and lines’ resistive elements represent 2% of the total power generated by the diesel generators (23 kw). the contribution of different network components in the losses is shown in fig. 3. it is worth noting that the pilot pv project (304 kw) was connected to the distribution system at the powerhouse according to the current plan. installing the pv system at the generation station would not alter the power flow; therefore, the value of losses will not be altered either. in this case, the benefits of the pv system are limited to emission and fuel consumption reduction. 4.2. optimal location of current proposed size using ga, the optimum location of the proposed-sized pv system is bus 54 (transformer tx 50). the total losses were reduced to 17.31 kw, yielding a ~25% reduction. in addition, installing the pv system at this location improves the voltage profile at different busses [fig. 4]. 4.3. optimum location and size of pv system the impact of increasing the size of the pv system at different locations was investigated [fig. 5]. on the basis of the results, the optimum location of the pv system was determined to be at bus 39 (transformer tx 28), with a size of 1,200 kw. this connection would reduce the real power losses from 23 to 12 kw yielding a ~50% reduction. losses could be reduced further by considering more than one location. however, the objective of the study was to determine one location for the pv project. the voltage profile of the system was improved [fig. 4]. figure 4. voltage profile at different scenarios. 5. conclusions this study considered the impact of the first pv pilot project in oman on system losses via a system model, which was built according to measured data provided by the rural area electricity company, and system losses were obtained. it was found that optimal sizing and siting of pv systems in distribution networks helps in maximizing their benefits. this is because the presence of pv systems in distribution networks alters the classical downstream power flow; therefore, the presence of pv systems may have technical and economic impacts on the system. installing the pv system at the powerhouse will not change the power flow; therefore, it will not m. h. albadi, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri 57 figure 5. impact of increasing the pv system on losses. affect the value of losses. on the basis of loss minimization, the optimal location of the proposed pv system was determined using ga. installing the pv system at the optimal location reduced system losses by approximately 25% and voltage profile was improved. acknowledgment the authors thank the rural area electricity company for providing the system data and support in achieving this work. references abu-mouti fs, el-hawary me (2011), optimal distributed generation allocation and sizing in distribution systems via artificial bee colony algorithm. ieee transactions on power delivery 26: 2090–2101. al-badi a, bourdoucen h (2011), feasibility analysis of renewable hybrid energy supply options for masirah island. international journal of sustainable engineering vol.5(5): 244-251. al-badi ah, albadi mh, al-lawati am, malik as (2011), economic perspective of pv electricity in oman. energy 36: 226–232. albadi mh, el-saadany ef (2008), the role of distributed generation in restructured power systems. presented at the ieee pes north american power symposium (naps08), calgary, alberta, canada. al-hinai a, (2010), dynamic stability enhancement using genetic algorithm power system stabilizer. presented at the international conference on power system technology (powercon2010), hangzhou, china. attwa ym, el-saadany ef, albadi mh (2009), optimum allocation of wind based dg in unbalanced rural network. presented at the international conference on computer communication and power (icccp09), muscat, oman. atwa ym, el-saadany ef (2010), optimal allocation of ess in distribution systems with a high penetration of wind energy. ieee transactions on power systems 25: 1815–1822. authority for electricity regulation, oman (2015), welcome to the authority for electricity regulation. retrieved from http://www.aeroman.org/. chen cs, duan sx (2014), optimal allocation of distributed generation and energy storage system in microgrids. renewable power generation, iet 8: 581–589. elhadidy ma, shaahid sm (2009), exploitation of renewable energy resources for environment-friendly sustainable development in saudi arabia. international journal of sustainable engineering 2: 56–66. ganguly s, samajpati d (2015), distributed generation allocation on radial distribution networks under uncertainties of load and generation using genetic algorithm. sustainable energy, ieee transactions 6: 688–697. kai z, agalgaonkar ap, muttaqi km, perera s (2012), distribution system planning with incorporating dg reactive capability and system uncertainties. sustainable energy, ieee transactions 3: 112–123. rural areas electricity company (2015), about us. retrieved from http://www.reefiah.com. m. h. albadi, n.a. al-mashaikhi, s. al-hinai, r. s. al-abri, a.s. al-hinai, q.k. al-aamri, a.a. al-mazidi and m.s. al-gafri 58 appendix a. the per unit line data. from to r x b/2 1 2 0.0026 0.0029 3.87e-06 1 3 0.2613 0.2286 0.000202 1 11 0.2824 0.2471 0.000218 2 99 6.7564 22.7700 0 3 10 0.0106 0.0171 1.90e-07 3 4 0.1481 0.2397 2.66e-06 3 58 18.8000 43.6200 0 4 5 0.0597 0.0966 1.07e-06 4 57 18.8000 43.6200 0 5 6 0.0597 0.0966 1.07e-06 5 60 18.8000 43.6200 0 6 7 0.1511 0.2446 2.71e-06 6 61 0.8394 4.6753 0 7 8 0.0529 0.0856 9.50e-07 7 62 18.8000 43.6200 0 8 9 0.9306 1.5069 1.67e-05 8 63 1.8280 9.3220 0 9 64 3.6825 14.6222 0 10 59 18.8000 43.6200 0 11 13 0.0045 0.0073 8.14e-08 12 71 18.8000 43.6200 0 13 14 0.1163 0.1884 2.09e-06 13 47 0.0053 0.0086 9.50e-08 14 15 0.0181 0.0294 3.26e-07 14 65 1.8280 9.3220 0 15 16 0.0295 0.0477 5.29e-07 15 66 3.6825 14.6222 0 16 17 0.0302 0.0489 5.43e-07 16 25 0.0295 0.0477 5.29e-07 16 68 18.8000 43.6200 0 17 18 0.0287 0.0465 5.16e-07 17 67 3.6825 14.6222 0 18 19 0.0242 0.0391 4.34e-07 18 69 0.8394 4.6753 0 19 20 0.0091 0.0147 1.63e-07 19 70 0.8394 4.6753 0 20 21 0.1051 0.1703 1.89e-06 20 12 0.2842 0.0794 3.44e-05 21 22 0.0232 0.0088 4.64e-06 21 23 0.1058 0.1712 1.90e-06 21 72 6.7564 22.7700 0 22 73 0.8394 4.6753 0 23 24 0.1461 0.2366 2.62e-06 23 74 18.8000 43.6200 0 24 75 18.8000 43.6200 0 25 26 0.0653 0.0182 7.91e-06 25 27 0.0166 0.0269 2.98e-07 26 76 1.8280 9.3220 0 27 28 0.0121 0.0196 2.17e-07 27 77 6.7564 22.7700 0 28 29 0.1511 0.2446 2.71e-06 28 78 3.6825 14.6222 0 29 30 0.0085 0.0074 6.54e-06 29 31 0.0373 0.0414 4.16e-05 29 32 0.0032 0.0051 5.70e-08 29 79 3.6825 14.6222 0 30 81 0.8394 4.6753 0 31 80 0.8394 4.6753 0 32 33 0.0106 0.0171 1.90e-07 32 37 0.0158 0.0138 1.22e-05 33 34 0.0136 0.0220 2.44e-07 33 95 3.6825 14.6222 0 34 35 0.0136 0.0220 2.44e-07 34 96 6.7564 22.7700 0 35 36 0.1934 0.3131 3.47e-06 35 97 6.7564 22.7700 0 36 98 3.6825 14.6222 0 37 38 0.0166 0.0269 2.98e-07 37 82 3.6825 14.6222 0 38 42 0.0181 0.0294 3.26e-07 38 83 3.6825 14.6222 0 42 39 0.0121 0.0196 2.17e-07 42 43 0.0166 0.0269 2.98e-07 39 40 0.0091 0.0147 1.63e-07 39 85 18.8000 43.6200 0 40 41 0.0486 0.0293 2.08e-05 40 86 3.6825 14.6222 0 41 87 1.8280 9.3220 0 43 44 0.1284 0.2079 2.31e-06 43 84 3.6825 14.6222 0 44 45 0.2160 0.3498 3.88e-06 44 53 0.0609 0.0533 4.71e-05 44 88 18.8000 43.6200 0 45 46 0.0816 0.1321 1.47e-06 45 89 18.8000 43.6200 0 46 90 18.8000 43.6200 0 47 48 0.0015 0.0024 2.71e-08 47 100 6.7564 22.7700 0 48 49 0.3067 0.4966 5.51e-06 48 101 3.6825 14.6222 0 48 102 18.8000 43.6200 0 49 50 0.0831 0.1345 1.49e-06 49 103 6.7564 22.7700 0 50 51 0.2795 0.4526 5.02e-06 50 104 3.6825 14.6222 0 51 52 0.0967 0.1566 1.74e-06 51 105 18.8000 43.6200 0 52 106 6.7564 22.7700 0 52 107 6.7564 22.7700 0 53 54 0.0224 0.0196 1.73e-05 53 91 0.8394 4.6753 0 54 55 0.0504 0.0441 3.89e-05 54 92 1.8280 9.3220 0 55 56 0.0757 0.0663 5.85e-05 55 93 0.8394 4.6753 0 56 94 0.8394 4.6753 0 loss reduction in isolated rural area distribution network using photovoltaic system 59 appendix b. transformer loading data. name rating (kva) p (kw) q (kvar) tx 2 200 0 0 tx 3 1000 74.96 46.74 tx 4 500 53.35 33.27 tx 5 500 30.71 19.15 tx 9 500 77.25 48.17 tx 10 100 0.00 0.00 tx 11 200 14.30 8.92 tx 12 315 12.16 7.58 tx 14 200 28.97 18.06 tx 15 315 10.10 6.30 tx 16 315 119.99 74.82 tx 17 315 69.06 43.06 tx 18 100 0.00 0.00 tx 19 100 46.88 29.23 tx 20 200 44.31 27.63 tx 21 100 20.27 12.64 tx 22 100 0.00 0.00 tx 23 200 35.15 21.92 tx 24 315 73.74 45.98 tx 25 315 123.99 77.31 tx 26 100 0.00 0.00 tx 27 315 14.16 8.83 tx 28 100 18.79 11.71 tx 29 100 0.00 0.00 tx 30 315 31.89 19.89 tx 33 100 25.22 15.73 tx 34 100 29.75 18.55 tx 35 100 5.24 3.27 tx 36 100 3.85 2.40 tx 38 500 28.00 17.46 tx 39 315 26.69 16.64 tx 40 315 8.64 5.38 tx 41 315 0.00 0.00 tx 42 1000 103.75 64.69 tx 43 100 21.44 13.37 tx 44 100 17.03 10.62 tx 45 315 12.73 7.94 tx 46 1000 85.52 53.32 tx 47 1000 79.25 49.42 tx 48 500 86.57 53.98 tx 49 1000 33.31 20.77 tx 50 1000 57.34 35.75 tx 51 100 0.91 0.57 tx 52 1000 85.54 53.34 tx 53 1000 154.38 96.26 tx 54 1000 12.45 7.76 tx 55 200 2.90 1.81 tx 56 200 14.46 9.02 tx 57 200 28.36 17.68 tx 58 315 0.00 0.00 tx 59 200 0.00 0.00 total 1823.36 1136.92 microsoft word paper 4.docx the journal of engineering research, vol. 11, no. 2 (2014) 39-49 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of asphalt mixes a.m.a. abdo department of civil and environmental engineering, dhofar university, po box 2509, salalah – 211, sultanate of oman received 4 july 2013; accepted 22 may 2014 abstract: during the past decades, many numerical models have been used to predict responses of asphalt mixes under different types of loading. some of these models were simple due to practicality but overestimated the response of asphalt mixes. on the other hand, sophisticated but effective numerical models have been developed to address the shortcomings of the simpler models, and were used mostly in finite element analysis (fea). however, these models were complicated and not user friendly. recently, the approach of the discrete element method (dem) was adopted. unlike traditional fea, dem can simulate crack propagation by allowing the separation of elements in the simulated models. understanding these challenges, this study was initiated to investigate the utilization of a simple visco-elasto-plastic model that had been used successfully in predicting deformation in asphalt mixes using the dem embedded in particle flow code in two dimensions (pfc2d) software simulations. simulation results, when compared to flow time (ft) and number (fn) test results, showed that this model could simulate actual tests, thus predicting deformation of asphalt mixes using the dem on a larger scale. keywords: asphalt mixes, numerical modeling, discrete element method, flow time test, flow number test, deformation. אאאא אאא  k א :אאאאאkא אאkאאא א  א א  א  א      ،  אא  אאאkאkkאא אאאאאאאא אkא،אאאאאאא   א א א  א א   א pfc2dא  k אאאאאאאא אאאאkא אאw،אא،אא،אא،אא א *corresponding author’s e-mail: aabuabdo@du.edu.om 40 a.m.a. abdo 1. introduction flexible asphalt pavements are being subjected to more diverse and harsher loading conditions than ever before. however, even after adjusting design factors to account for these conditions, the adopted design techniques, which were developed using simple models, fall short of predicting failures as efficiently as they were able to do in the past. additionally, it is difficult to take all properties into consideration while designing flexible pavement. asphalt mix is a viscoplastic and anisotropic material. it varies in terms of its temperature, loading time, and aggregate orientation dependency. thus, designers have chosen to overlook those properties in favor of simplifying the design process. plastic deformation, known as rutting, is the major distress pavement designers encounter. in order to prevent this type of premature failure and any undesired effects due to permanent deformations, as well as to account for the complex material properties of asphalt mixes, a more realistic model is essential. incorporating the development and adoption of such material models into design methodologies will result in an improvement of performance and a reduction in maintenance. recent studies simulated asphalt mixes’ responses to different types of loading conditions. by utilizing numerical models to describe a mesh of elements that simulated test setups, numerical solutions based on finite element analysis (fea) were found. a study by buttlar et al. (2003) suggested that most existing analysis models did not directly account for the continuous grading of properties in flexible pavements. they presented the application of a numerical model embedded in fea software for asphalt pavement analysis. masad et al. (2005) used an anisotropic non-associated flow rule based on the drucker–prager yield criterion. the model parameters were related to the experimental measurements of aggregates’ characteristics and microstructure damage, which was measured using x-ray tomography and image analysis techniques. tashman et al. (2005) introduced a microstructure-based viscoplastic continuum model for predicting permanent deformation of asphalt mixes. this model took into consideration strain rate and confining pressure dependency, dilation, aggregate friction, anisotropy, and damage, all of which have a huge impact on the permanent deformation of asphalt mixes at high temperatures. the developed model predictions were in good agreement with the experimental measurements. a disadvantage of fea is that it will not allow the separation of elements during analysis. thus, no formation of micro or macro cracks in the modeling of asphalt mixes under loading will occur. 2. the discrete element method (dem) recently, more attention has been directed to the dem, which was introduced in 1971 to analyze problems with rock mechanics, ice formation and flow in streams, as well as earthquakes, impacts, and explosion damage to structures. the discrete element algorithm models a continuum as a system of distinct, interacting, and general-shaped particles subjected to laws of motion and deformation, which are bonded together. when the maximum stress in a contact bond is reached, the bond is broken and a separation of these particles, evidenced as cracks, can be observed (cundall and strack 1979). in the dem, the complex constitutive behavior of a material is simulated by associating simple constitutive models with each particle contact. shear and normal stiffness, static and sliding friction, and inter particle cohesion are three of the simpler contact models that can be employed (you 2003). you and buttlar (2006) argued that the dem is a fundamental way of looking at the complex behavior and heterogeneity of asphalt mixes. they suggested that the dem could be used to simulate asphalt mixes responses under different loading and temperature conditions. they represented the dem approach as a research tool for modeling asphalt mix microstructure. abbas et al. (2007) utilized the dem to develop a micromechanical model that accounted for viscoelastic behavior of asphalt mixes. asphalt mix microstructure was captured using gray scale images of vertically 41 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of ….. cut sections of the compacted samples. collop et al. (2006) investigated the use of the dem by utilizing the burger model to describe the contact bonds in their analysis under compression. they argued that the behavior of the mixture would be dominated by asphalt binder and complex aggregates interlock effects could be minimized. it was found that tested mixes dilated when the ratio of compressive to tensile contact stiffness increases as a function of loading time. recent studies (yang et al. 2012; liu et al. 2012; cai et al. 2013) suggested that the dem, via particle flow code in two dimensions (pfc2d) and particle flow code in three dimensions (pfc3d) software (itasca international inc., minneapolis, minnesota, usa), could be used to predict the rutting resistance of asphalt mixtures more conveniently, by simulating the different test setups of asphalt mixes. the simulation results showed that the discrete element simulation and laboratory test had good correlation, which verified the applicability of these adopted models. when it was developed, the main function of the dem was to act as a tool to perform research to help with understanding the behavior of granular materials. however, it has been used recently in biomechanics, petrochemical engineering, fluid mechanics, and different structural engineering applications in which particle models were used to simulate material behavior in real engineering problems that involve complicated deformation patterns (shibata et al. 2003; goda and ebert 2005; mas ivars 2006; mahmoud et al. 2010; jinag et al. 2010; lau et al. 2011; krabbenhoft et al. 2012; abu abdo et al. 2012; abraham et al. 2013; nakamura et al. 2013). 3. scope the main purpose of this study was to supplement previous research conducted in the area of hot mixture asphalt (hma) evaluation. this study investigated the utilization of a simple visco-elasto-plastic model that had been used successfully in predicting deformation in asphalt mixes (abu abdo 2012). the goal of simulating laboratory tests in dem was to verify the ability of this simplified predictive model to determine the simplified predictive model and the deformation of asphalt mixes on a larger scale under different loading conditions. 4. particle flow code in two dimensions (pfc2d) software pfc2d is a platform for conducting the complicated calculations of the dem, where representative elements containing several hundred particles are tested numerically. continuum methods are used to solve real problems that involve complicated deformation patterns (itasca 2006). in pfc2d, particles’ interactions are treated as a dynamic process with states of equilibrium developing whenever the internal forces balance. each particle is assigned contact bonds to all adjacent particles and will break when ultimate strength is reached. contact forces and displacements of a loaded body of particles are found by mapping the movements of the individual particles, which is dependable on the physical properties of the discrete system. the calculations performed in pfc2d alternate between the application of newton’s second law of particles and a forcedisplacement law of contacts. newton’s second law is used to determine the motion of each particle arising from the contact and body forces acting on it, while the forcedisplacement law is used to update the contact forces arising from the relative motion at each contact (itasca 2006). most numerical modeling of asphalt mixes under axial compressive loading uses the burger model to describe contact bonds between particles in pfc2d simulations (schwarz and weeks 1977; ye et al. 2009; zelelew and papagiannakis 2009). in this study, the sea ice model (schwarz and weeks 1977; mellor 1981 and abu abdo 2012) was used for its better representation of the contact bonds between particles. commonly, sea ice, which is similar to asphalt mixes, exhibits time-dependency, mainly delayed elastic recovery and creeping. many experiments have shown that ice strength has a dynamic value depending non-linearly upon the strain rate. the strain-stress relation that describes the sea ice behavior under loading is described in eqn. (1). 42 a.m.a. abdo = + . 1 − . +. . (1) where, σ is the applied stress (kpa), e is the mix stiffness (kpa), c is a material constant, b is a time exponent for delayed elastic strain, at is a material constant (s-1), d is an average grain diameter (mm), do is a grain diameter (mm), s is the stress exponent for grain-bound sliding, n is the degree of viscosity power law, and evo is the viscous strain rate (s-1). a study by abu abdo (2012) suggested that the sea ice model could be used effectively in predicting deformation in asphalt mixes under static and cyclic loading conditions. the model parameters were easily determined, and could be adapted to many numerical methods and used effectively in predicting the deformation of asphalt mixes. 5. experimental verification 5.1 test setups 5.1.1 static creep/flow time (ft) test a flow time (ft) test is a triaxial static compressive creep test at which a total straintime relationship for a specimen is measured in the laboratory under unconfined or confined conditions. a ft test is used to determine the instantaneous elastic and plastic components, as well as the viscoelastic and viscoplastic components of the material’s response as shown in fig. 1. the relationship could be divided into three major zones: the primary zone (initial stage); the secondary zone (linear portion); and the tertiary flow zone (the portion in which the strain increases significantly) at which failure occurs. a study by witczak et al. (2002) suggested that the higher the slope of the linear portion of the graph (secondary zone) the higher the permanent deformation under loading would be. furthermore, the ft is defined as the postulated time when shear deformation, under constant volume, starts. in this study, the specimens were subjected to a static axial load of 207 kpa and were conducted without confinement at a testing temperature of 54.4 °c. 5.1.2 flow number (fn) test another approach of measuring permanent deformation of an asphalt mix is the flow number (fn) test. it is a triaxial repeated load test conducted with several thousand repetitions of a cyclic load, while recording the cumulative permanent deformation as a function of the number of load cycles, as shown in fig. 2, where the relation could be divided into three major zones similar to the ft test. the fn is defined as the number of load repetitions at which shear deformation, under constant volume, starts (witczak et al. 2002). specimens were tested using a loading cycle of 1.0 second. the loading cycle consisted of a 0.1 second haversine load of 207 kpa followed by a 0.9 second rest at a testing temperature of 54.4 °c. 5.2 tested mixes properties and samples preparation two superpave mixes with different properties were used and are listed in table 1. samples were compacted using a superpave gyratory compactor (sgc) to a height of 175mm with 9% air voids. these samples were then cored and sawed to obtain a sample measuring 100 mm in diameter and 150 mm in height. these samples had a 7% air void, as per the sample preparation procedure for the ft and number tests (witczak et al. 2002). 5.3 pfc2d simulations and results to conduct simulations using pfc2d software, tested specimens were modeled. the model developed initially for this study was constructed of a wide range of particle (sphere) sizes, thus approximating the irregular matrix of the cross-section of an asphalt mix sample (fig. 3). during this initial phase, the goal was to create a model that simulated actual physical samples. unfortunately, the analysis of such a model did not converge and no solution was obtained. then it was decided to 43 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of ….. model asphalt mix samples as a homogeneous set of 5,000 particles arranged in rows where a solution was obtained (fig. 4). one of the major benefits of using a homogeneous particle configuration is the significant decrease of time needed to analyze the model by excluding packing procedure. then contact bond properties were identified and modeled using the sea ice model eqn. (1). to utilize the sea ice model, the model parameters should be figure 1. typical flow time test results for asphalt mixes (after witczak et al. 2002). figure 2. typical flow number test results for asphalt mixes (after witczak et al. 2002). determined. these parameters were obtained from a previous study (abu abdo 2012) (table 2). actual results of the ft and fn tests were used to determine the model parameters. mixes 1 and 2 had the same model parameters except for the evo. it was suggested that the evo 44 a.m.a. abdo parameter was related to the asphalt binder grade (viscous part of the mix) and since the tested mixes had different binders (table 1) it was expected that the evo would be different for both mixes. figure 3. original element assembly (flow time test model). figure 4. final element assembly (flow time and number test models). similar to ft and fn test setups, a simulated specimen using pfc2d software was loaded by assigning forces (constant for the ft test and cyclic for the fn test) to the bottom particles to 45 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of ….. simulate the lower loading plate in both tests setups (fig. 5). the displacements of top particles were prohibited to simulate fixed top plate that acts as a support. laboratory test results for the ft test (fig. 6) showed that mix 1 failed (tertiary flow zone) at a lower number of cycles (tf = 1025 sec), while mix 2 did not fail. thus, it could be concluded that mix 2 would perform better than mix 1 under the same loading conditions. this could be explained by the better mix properties of mix 2 (ie. larger aggregates sizes and higher binder grade). table 1. mix properties as detailed by abu abdo (2012). parameter mix 1 mix 2 job mix design class 19.0mm 25.4mm binder grade pg 58-34 pg 70-28 asphalt content 4.6% 5.0% gmm 2.801 2.458 %passing sieve size 25mm 100 % 98 % 19mm 96 % 96 % 12.5mm 79 % 82 % 9.5mm 70 % 73 % 4.75mm 46 % 55 % 2.36mm 28 % 36 % 1.18mm 18 % 24 % 0.6mm 12 % 13 % 0.3mm 9 % 9 % 0.15mm 5 % 6 % 0.07mm 3.5 % 4.4 % table 2. sea ice model parameters for asphalt mixes as detailed by abu abdo (2012). parameter mix 1 mix 2 s 1.229378 1.229378 at 3.14x10-04 s-1 3.14x10-04 s-1 c (static loading) 7.5 7.5 c (dynamic loading) 2.0 2.0 b 0.34 0.34 n 3 3 evo 6.59x10-06 s-1 1.78x10-07 s-1 then laboratory test results were compared to the simulated tests. as shown in fig. 6, pfc2d simulation slightly over-predicted the total strain of mix 1 until the ft was reached, where the tested samples started to flow and shear deformation began (tertiary flow zone). on the other hand, pfc2d simulation approximately matched the measured values of mix 2. 46 a.m.a. abdo figure 5. flow time (ft) and number (fn) test setups. figure 6-a. measured vs. fitted model results of total strain determined in flow time test (mix 1). 47 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of ….. figure 6-b. measured vs. predicted results of total strain determined in flow number test (mix 2). figure 7. measured vs . predicted results of total strain determined in flow number test (mix 2). 48 a.m.a. abdo to confirm the effectiveness of the sea ice model under different loading conditions, mix 2 was tested using the fn test (cyclic loading) setup. as shown in fig. 7, pfc2d simulation results over-predicted total strain at an earlier stage (primary zone) and then under-predicted the total strain at the secondary and tertiary flow zones. further, it was observed that the total strain predicted by pfc2d reached an upper limit. the model parameters might need fine tuning to achieve better results. in addition, it was speculated that if pfc2d model was constructed of a wide range sizes of randomly distributed particles (fig. 3), simulation results would be more accurate since it represents an actual cross-section of a tested asphalt mix sample. 6. conclusions nowadays different and harsher loading conditions exist for asphalt pavements. therefore, the current design methods, which are based on simple models and assume that asphalt mixes have an elastic and isotropic nature, are not efficient. these methods cannot accurately predict failure as before. a more realistic and applicable approach is needed to account for the visco-plastic behavior and anisotropic nature of asphalt mixes. to prevent failures and undesired effects due to permanent deformations, many numerical models were developed to predict permanent deformation in asphalt mixes. unfortunately, these models include many specific parameters, which makes their use impractical. this study described the use of a simple viscoelasto-plastic model that has been used successfully in determining sea ice distresses under different wind loads. the model was used in the dem, embedded in the pfc2d software. pfc2d was used to simulate the ft and fn tests. two super pave mixes were used; results of the ft test showed that mix 2 would perform better than mix 1 under the same loading conditions. furthermore, a pfc2d simulation of the ft test results slightly overpredicted the total strain of mix 1 and approximately matched the measured values of mix 2. to validate the utilization of the adopted model under different loading conditions, a simulation of flow number test was conducted using mix 2. results by pfc2d simulations over-predicted the total strain at the early stage then under-predicted the total strain at the final stages. in addition, it was observed that total strain reached an upper limit, it is suggested that the model parameters might need fine adjusting to achieve better results. overall, results from this study indicated that the sea ice model could be utilized effectively in the dem via pfc2d software. additionally, flexible pavement simulation in pfc2d could be adapted to predict deformation on a larger scale and under different loading conditions. references abbas a, masad e, papagiannakis t, harman t (2007), micromechanical modeling of the viscoelastic behavior of asphalt mixtures using the discrete-element method. international journal of geomechanics 7:131–139. abu abdo a, eckwright f, jung sj, bayomy f, nielsen r (2012), evaluation of scnb testing procedure for hot mixture asphalt. proceedings of the institution of civil engineers: transport 167:48–58. abu abdo a (2012), simplified numerical modeling for asphalt mixes. international journal of pavement research and technology (ijprt). 5:40–45. abraham cl, maas sa, weiss ja, ellis bj, peters cl, anderson ae (2013), a new discrete element analysis method for predicting hip joint contact stresses. journal of biomechanics 46:1121–1127. buttlar wg, paulino gh, song sh (2003), application of graded finite elements for asphalt pavement analysis. computational fluid and solid mechanics, proceedings of second mit conference on computational fluid and solid mechanics 157–161. cai w, mcdowell gr, airey gd (2013), discrete element modelling of uniaxial constant strain rate tests on asphalt mixtures. granular matter 15:163–174. collop ac, mcdowell gr, lee yw (2006), modelling dilation in an idealized asphalt mixture using discrete element modelling. granular matter 8:175–184. 49 utilizing a simple numerical model in discrete element analysis to simulate flow time and number tests of ….. cundall pa, strack odl (1979), discrete numerical model for granular assemblies. géotechnique 29(1):47-65. goda tj, ebert f (2005), three-dimensional discrete element simulations in hoppers and silos. powder technology 158:58–68. itasca consulting group inc. (2006), particle flow code in 2 dimensions user's guide. minneapolis, minnesota: itasca consulting group inc. jiang mj, xiao y, chen sl, hu hj, wu xf (2010), discrete element analysis of bearing mechanism of single pile in sand under vertical load. yantu lixue/rock and soil mechanics 31:366–372. krabbenhoft k, lyamin av, huang j, vicente da silva m (2012), granular contact dynamics using mathematical programming methods. computers and geotechnics 43:165–176. lau m, lawrence kp, rothenburg l (2011), discrete element analysis of ice loads on ships and structures. ships and offshore structures 6:211–221. liu y, you z, zhao y (2012), threedimensional discrete element modeling of asphalt concrete: size effects of elements. construction and building materials 37:775– 782. mahmoud e, masad e, nazarian s (2010), discrete element analysis of the influences of aggregate properties and internal structure on fracture in asphalt mixtures. journal of materials in civil engineering 22:10–20. masad e, tashman l, little d, zbib h (2005), viscoplastic modeling of asphalt mixes with the effects of anisotropy, damage and aggregate characteristics. mechanics of materials 37:1242–1256. mas id (2006), water inflow into excavations in fractured rock—a three-dimensional hydro-mechanical numerical study. international journal of rock mechanics and mining sciences 43:705–725. mellor m (1981), fracture toughness measurement for ice. crrel technical memorandum, hanover, new hampshire, usa, u.s. army corps of engineers, cold regions research and engineering laboratory. nakamura h, fujii h, watano s (2013), scaleup of high shear mixer-granulator based on discrete element analysis. powder technology 236:149-156. schwarz j, weeks wf (1977), engineering properties of sea ice. journal of glaciology 19: 499–531. shibata n, tomita n, ikeuchi k (2003), numerical simulations on fatigue destruction of ultra-high molecular weight polyethylene using discrete element analyses. journal of biomedical materials research part a, 64(3):570–582. tashman l, masad e, little d, zbib h (2005), a microstructure-based viscoplastic model for asphalt concrete. international journal of plasticity 21:1659–1685. yang j, zhang x, zhu hr (2012), discrete element simulation on tri-axial shear test of asphalt mixtures. journal of building materials 15:64–68. ye y, yang x, chen c (2009), experimental researches on visco-elastoplastic constitutive model of asphalt mastic. construction and building materials 23:3161–3165. witczak mw, kaloush k, pellinen t, albasyouny m, von quintus h (2002), simple performance test for superpave mix design. nchrp report 465, trb, washington dc, usa. you z (2003), development of a micromechanical modeling approach to predict asphalt mixture stiffness using the discrete element method. phd dissertation, university of illinois at urbana-champaign. you z, buttlar wg (2006), micromechanical modeling approach to predict compressive dynamic moduli of asphalt mixtures using the distinct element method. transportation research record 1970:73–83. zelelew hm, papagiannakis at (2009), dem simulation of asphalt concrete uniaxial creep. sixth international conference on maintenance and rehabilitation of pavements and technological control (mairepav6), turin, italy. the journal of engineering research (tjer), vol. 16, no. 1 (2019) 28-34 evolution and characterizations of aurivillius bismuth titanate by modified solid state processing m. shahnawaz a , s. khan b , and s. mukherjee *,c a department of electronics and communication engineering, heritage institute of technology, kolkata700107, india. b department of metallurgical and materials engineering, jadavpur university, kolkata-700032, india. c department of mechanical and automation engineering, amity university kolkata, kolkata-700135, india. abstract: mild agate mortar activation followed by sintering of precursors, in resemblance to modified solid state process, was studied for synthesis of the aurivillius bismuth titanate. hand on mill activation was carried out for eight hours followed by annealing treatment at 1000°c for 12 hours in presence of air atmosphere to obtain the proper phase. scherrer’s formula was utilized to estimate crystallite size along with planes of orientation. crystallite size was about 65nm while prominent peaks of orientations were (-117), (006), (111), (200) and others. fesem and tem studies were carried to obtain the morphology and estimated grain size of the synthesized aurivillius compound. morphological features execute the material system to be nanocrystalline in nature as estimated from grain size measurement in correspondence with xrd crystallite size estimation. ftir analysis confirms m-o coordinations of synthesized modified perovskite (aurivillius) sample. optical property was evaluated from uv-vis analysis with prominent absorption spectra in the visible region. tauc plot was used to estimate the band gap to be about 2.86ev & 2ev for both direct and indirect transitions. keywords: bismuth titanate; aurivillius; morphology; band gap. لحالة الصلبةل معدلة بمعالجة ثالبزمولتيتنات عنصر اورفوليت صيغة تطور وتوصيفات شاهناواز .م أ خان .، س ب *موخرجي .س و ،ج دراسة عملية تنشيط مالط العقيق الخفيف متبوًعا بتلبيدة معدة سالفا لتكون متشابهة مع عملية الحالة ب لقد قمنا: الملخص بعملية التنشيط هذه بالمصنع لمدة ثماني ساعات تلتها قمنا .ثالصلبة المعدلة لتوليفة اورفوليت لتيتنات عنصرالبزمو ساعة في وجود الهواء الجوي و ذلك للحصول على الحالة 01مئوية لمدة 0111المعالجة الصلبة عند درجة حرارة حوالي المناسبة. و قمنا باستخدام صيغة سكيرر من أجل تقدير حجم البلورات المتجهة نحو السطح. و كان حجم البلورة ( وغيرها.وتم اجراء قياسين بمجهر 111( و )000( و )115( و )001-نانومتر بينما كانت الزوايا البارزة تتجه نحو ) 56 تقدير حجم حبيبات مركب لللحصول على الشكل المورفولوجي وو ذلك إلكتروني متحرك ومجهر إلكتروني ماسح ى جعل نظام المواد ذات طبيعة نانوية كما هو مقدر من قياس حجم هذا و تعمل الخصائص المورفولوجية عل اورفوليت. الحبيبات بالتوافق مع تقدير جهاز قياس نسبة األشعة السينية لحجم البلورات. وقد أكد تحليل جهاز "فورييه" لتحويل طيف الخاصية البصرية من التحليل ( لتوليفة عينة اورفوليت المعدلة . وتم تقييم m-oاألشعة تحت الحمراء "فتير" تنسيقات ) "توك" و قد قمنا باستخداِم مخطط .الطيفي فوق البنفسجي المرئي مع مراحل االمتصاص البارزة في المنطقة المرئية لكل من التحوالت المباشرة وغير المباشرة. 2.86ev & 2ev لتقدير الفجوة بين .فجوة المراحل؛ المورفولوجيا ؛تيتان ث: اورفوليت البزمو المفتاحية الكلمات corresponding author’s e-mail: smmukherjee3@gmail.com doi: 10.24200/tjer.vol16iss1pp28-34 mailto:smmukherjee3@gmail.com the journal of engineering research (tjer), vol. 16, no. 1 (2019) 28-34 29 1. introduction among the bi-ti-o system, bismuth titanate is one of the most important having versatile applications. memory storage, effluent treatment, optical displays, piezoelectric transducers and photocatalysis are the domain where bi4ti3o12 is noted to be one of the promising candidates for execution nogueira e. andré et. al. (2014). intergrowth of fluorite [bi2o2] 2+ and perovskite [bi2ti3o10] 2 in which bi ions occupy twelve-coordinated sites, which were found to be the basis of the structural aspects of bismuth titanate compound chen zhiwu et. al. (2009). generally, fluorite and perovskite can be represented as [m2o2] 2+ and perovskite [an-1bno3n+1] 2, where n is the number of bo6 octahedra. the material has ferroelectric nature with intermediate curie temperature of about 675°c. such material of bi-tio system also has high fatigue endurance, and considerable electro-optic switching behavior. bismuth titanate has the potential for application in non-volatile random access memory, high temperature piezoelectric and electro-optic devices. for fluorite [m2o2] 2+ (m is 8-coordinated) while the three layer member of aurivillius phase (n=3) actually represents bismuth titanate. pavlović nikolina et. al. (2009). it is observed that bismuth titanate has monoclinic structure at room temperature and converts to a tetragonal phase at above curie temperature and possesses low dielectric constant, low coercive field, and a high breakdown voltage yang qunbao et. al. (2003) quite a few process routes have been developed for the synthesis of bismuth titanate. most of the common process routes for synthesis are hydrolysis of metal organic salts osamu y et. al. (1991), co-precipitation horn a. jeffrey et. al. (1999), sol-gel (haoshuang gu et. al. 1998; sedlar m et. al. 1996) hydrothermal (yang qunbao et. al. 2003; shi yanhui et. al. 2000), mechanochemical ng hwee szu et. al. (2002), molten salts method gopalan srikanth et. al (1996) and others. in spite of all these processes a high temperature sintering is required for the crystallization of bi3ti4o12. yang qunbao et. al. (2003). due to the presence of ferroelectric, piezoelectric properties, bismuth titanate is noted as a promising candidate in comparison to the lead zirconium titanate which is used primarily for nonvolatile memory, piezoelectric applications. absence of lead content in the composition of aurivillius structure makes bismuth titante an ecofriendly material system for versatile applications. ng hwee szu et. al. (2002). high density, and proper microstructure are very essential to obtain such electrical properties guha j.p. (1999). solid state mix assisted mechano-chemical activation followed by annealings is used to synthesize aurivillius bismuth titanate as shown in this article. there is a limited research, so far, focusing on synthesis of the compound by such route which is a low cost easy method of synthesis without involvement of complex steps and precursors. the compound is analyzed for phase determination by xrd, morphology from fesem, hrtem bond formation from ftir spectra along with band gap evaluation from uv-vis spectra analysis. 2. experimental stoichiometric amounts of bismuth oxide (bi2o3) and titanium oxide (tio2) ar grade in 2:3 molar ratios were taken in an agate mortar for milling activation. precursor powders were then mixed in the agate mortar-pestle. solid state mixing was carried for eight hours followed by sintering at 1000°c for a 12 hour. soaking period in air atmosphere within a pid controlled furnace to synthesis bismuth titanate. the heating rate was fixed at 6°c per min, 5°c per min, 4°c per min for the range of 0°c to 600°c, 600°c to 800°c and lastly from 800°c to 1000°c for proper synthesis while using a normal cooling rate. the stoichometric equation to synthesize the required compound is given as 2bi2o3 + 3tio2 = bi4ti3o12. the characterization techniques were analysed to determine the phases developed, morphological features, bonding and optical properties of the material. using a wavelength of cu kα 1.54å, 40ma, 50kv as current and voltage ratings within the scan range 2θ 10-80°, normal scan rate of 5°/min, crystallographic analysis of the sample was carried by xrd (rigaku ultima iii) at ambient condition. m-o co-ordinations of the synthesized samples were identified by ftir analysis (ir prestige-21, shimadzu) to obtain information for bonding in the sample. kbr powder was mixed with the sample to form a pellet under pressure of 6t/cm 2 for spectral analysis. to undergo morphological analyses of the samples using fesem (hitachi, s-4800) and hrtem (jeol, jem 2100) operated at 200kv, conducting coating of gold was deposited (by sputtering) to avoid static charge accumulation in order to obtain a better resolution. optical properties were determined from uv-vis spectral analysis followed by band gap evaluation. the samples were prepared by sonication using high pure di water as medium and put into a cuvette for analysis. band gap calculation was carried using tauc plot from uv-vis spectra (perkin elmer, lambda 35). 3. results and discussion figure 1 shows the xrd patterns of the prepared nano-crystallites by modified solid state process. an xrd pattern indicates aurivillus bismuth titanate formation as per jcpds pdf # 802143. thorough xrd pattern analysis indicates nil trace amount of bismuth oxide, titanium dioxide within the synthesized sample as residue or unreacted one. major peaks for growth are observed along (-117) plane while other important peaks are observed at 29 m. shahnawaz, s. khan and s. mukherjee (200), (111), (006), (0014). growth of crystallites is possible when there is a thermodynamic feasibility of phases after proper nucleation along with development of such phases in preferred oriented planes. thus, these observed planes or facets of planes indicate feasibility of the crystallite growth along these mentioned planes as observed from the aforementioned figures. none of the peaks, even in a minor form indicates phase decomposition or transition. absence of precursor peaks or intermediate oxides in xrd spectra confirms purity of the aurivillius phase. the crystallite sizes of the prepared sample is calculated using scherrer’s formula are expressed as t = 0.9 λ / (β cos θ), where t is the particle size, λ is the x-ray wavelength, β is the line broadening at half of the maximum intensity in radians and θ is the bragg angle in degree. major peak (-117) is indexed for the crystallite size estimation. the indices observed are along either all odd or even (h, k, l) indices and such observations get matched with xrd analysis of wang-ming chun et. al. (2008). the average particle size of bit calculated from scherrer’s equation is around 65 nm. as observed, the peaks have are sharp narrow width, indicating high crystalline nature without any humps and deconvolution. spectral patterns of ftir for the synthesized bit are shown in fig. 2 and they vary within the scan range from 400-4000 cm -1 . molecular vibration bond for stretching vibrations, symmetric and asymmetric stretching condition is observed from ftir spectra within the prescribed scan range. major relevant bond formed for the respective aurivillius phase related to m-o coordinations is observed to be within the range 400-700cm -1 . for synthesized material bi-o bond vibration is observed to be about 500cm -1 while for ti-o bond linkage is indicated by 1000 cm -1 . it is evident from fig. 2 that for bit formation distinct absorption intensity is around 500 cm −1 . stretching vibrations of bi–o bonds due to presence of absorption peak is strongly attributed to distorted bio6 octahedral units at around 490 cm −1 in the ftir absorption spectra as reported by rahaman atiar et. al. (2016). the stretched band at around 632700 cm −1 is due to ti–o bond or tio4 tetrahedral molla rahaman atiar et. al. (2016). the isosbestic point is observed at about 700 cm −1 which indicates that equilibrium exists between the oxide and bismuth containing species. small absorbance spectra at around 470cm -1 is possibly due to o-ti-o stretching vibration of tio2 which is used as precursor liu weiliang et. al. (2010). figure 3 indicates that the morphological features obtained from fesem of synthesized aurivillius type oxide after annealing at 1000ºc with a two hour soaking period. all morphological features observed for synthesized samples at different magnification and resolution exhibit agglomeration tendency while the growth pattern is more towards homogenous and uniform in nature. sharp defined edges are noted for the plate type morphology of the particulates in all of the above. multilayer, agglomerated, elongated structure is formed as a whole agglomerate of aurivillius bit. plates with elongated structured morphology are observed to be stacked together in some portions (fig. 3-b) and top view of such stacked platelets (mulilayer) give rise to agglomerated form like irregular shape polygonal or regular polygonal (fig. 3-a,d). layer thickness of individual platelets is noted to be about 0.15µ. the experimental findings have some resemblance with morphological findings by lin xue et. al. (2012). they have worked on photocatalytic activity of bismuth titanate in presence of visible light after synthesis in the same way as by hydrothermal method. in those cases, microsphere with platelets are noted as morphology depending upon the concentration of ti(oc4h9)4 lin xue et. al. (2012). similar results were also noted during hydrothermal synthesis of bi4ti3o12. with increase in temperature regular plate shape also increases in number and the size also increases significantly chen zhiwu et. al. (2009). negligible interconnected pores are noted amongst the agglomerated particulates. agglomerate chunks are around 2.5µ while individual particulates are found to be around 0.02µ-0.075µ. the average size of a single particle, as determined from the fesem micrographs, are in close correspondence to the particle sizes determined using the scherrer’s formula. high resolution images of solid state synthesized aurivillius barium bismuth titanate are noted from hrtem as shown in fig. 4. bit particles have strong agglomeration with elongated plate as shown throughout for hrtem analysis. both figs. 4a) and fig 4c) depict the elongated nature of platelets while fig 4b) and fig 4d) represent the sharp edges of an individual particulate. the elongated particulate has a width around 170nm and comprises of various small particulates. both figs. 4e) and 4f) represents agglomeration tendency as depicted from morphological analysis of fesem studies. the agglomerated forms that have regular or irregular polygonal shapes are observed. the growth pattern of agglomerated form is homogenous in nature. thus both fesem and hrtem analyses are found to be in correspondence with each other. moreover, hrtem studies confirm that particulates are in nano domain as observed from xrd crystallite size estimation by scherrers formula. the agglomerated form varies from spherical to polygonal type morphology. figure 5 depicts the band gap of aurivillus bismuth titanate at 1000°c for 12 hours using tauc plot. uvvis spectra are carried in the scan range of 200-850nm range in absorption mode. powder samples are dispersed in alcohol after ultrasonication for one hour to prepare the samples for optical analysis. tauc relation (αhν) n = c(hν –eg) is used where n =2, for direct allowed transition, n=1/2 for indirect allowed transition with the band structure. 30 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 28-34 figure 1. xrd of sintered bismuth titanate. figure 2. ftir spectra of aurivillius bismuth titanate sintered at 1200°c for 12 hours after 8 hours of milling activation. 31 m. shahnawaz, s. khan and s. mukherjee figure 3. fesem micrographs of bismuth titanate (bit) sintered at 1000 °c for 12 hours. figure 4. typical hrtem images of bismuth titanate (bit) sintered at 1000°c for 12 hours. 32 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 28-34 1 2 3 4 5 6 7 0.00e+000 2.00e+014 4.00e+014 6.00e+014 8.00e+014 1.00e+015 1.20e+015  h  )2 h (ev) 2.86ev direct bit figure 5. direct and indirect band gap of bismuth titanate by tauc plot at 1000oc for 12 hours. extrapolation of the tangent of the curve to the x axis, gives the value of the sample’s band gap. direct and indirect band gap values of the synthesized sample are around 2.86ev and 2ev respectively. hence, the aurivillus sample is observed to be active for energy transition (electron transfer within the band structure) within the visible region. thus, the synthesized material may have the possibility for photocatalytic applications, photoinduced devices in visible spectrum. moreover, the band gap values are quite comparable with oxide semiconductors like zinc oxide, barium titanate, bismuth ferrite and others. 4. conclusion bismuth titanate synthesis was carried after mechanochemical activation for eight hours along with annealing at 1000°c for 12 hrs. aurivillius type structure with crystallite size around 65nm was confirmed from phase analysis by xrd. m-o coordination and the bonding confirm formation of bismuth titanate by ftir analysis. morphological studies from fesem, hrtem confirm agglomeration, elongated structure, polygonal regular or irregular type agglomerates with homogenous distribution. agglomerates were around 2.5µ while individual particulates were around 0.02µ-0.075µ. band gap values estimated that direct and indirect transitions were 2.86ev and 2ev respectively using tauc relation. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references chen zhiwu, yu ying, hu jianqiang, shui anze, he xinhua (2009), hydrothermal synthesis and characterization of bi4ti3o12 powders. journal of the ceramic society of japan 117(3): 264-267. gopalan srikanth, mehta karun, virkar v. anil (1996), synthesis of oxide perovskite solid solutions using the molten salt method. journal of materials research 11(8): 1863–65. guha j.p. (1999), reaction chemistry and sub solidus phase equilibria in lead-based relaxor systems: part 1, formation and stability of the perovskite and pyrochlore compounds in the system pbo–mgo– nb2o5. journal of materials science 34(20): 4985–94. haoshuang gu, chen peizhi, zhou youhua, zhao min, kuang anxiang, li xingjiao (1998), reactions in preparing bi4ti3o12 ultrafine powders by sol-gel process. ferroelectrics 211: 271–280. horn a. jeffrey, zhang s c, selvaraj u, messing l. gary, mckinstry-trolier susan (1999), templated grain growth of textured bismuth titanate. journal of the american ceramic society 82(4): 921–926. lin xue, lv peng, guan qingfeng, li haibo, zhai hongjv, liu chunbo (2012), bismuth titanate microspheres: directed synthesis and their visible light photocatalytic activity. applied surface science 258: 7146-7153. liu weiliang, wang xinqiang, tian dong, xiao chenglong, wei zengjiang, chen shouhua (2010), chemical reaction and crystalline procedure of bismuth titanate nanoparticles derived by metalorganic decomposition technique. materials sciences and applications 1: 91-96. molla rahaman atiar, tarafder anal, karmakar basudeb (2016), fabrication and properties of nd 3+ doped ferroelectric barium bismuth titanate glass-ceramic nanocomposites. journal of alloys and compounds 680: 237-246. ng hwee szu, xue junmin, wang john (2002), bismuth titanate from mechanical activation of a chemically coprecipitated precursor. journal of the american ceramic society 85(11): 2660-2665. nogueira e. andré, longo elson, leite r. edson, camargo r. emerson (2014), synthesis and photo catalytic properties of bismuth titanate with different structures via oxidant peroxo method (opm). journal of colloid and interface science 415: 89–94. 33 m. shahnawaz, s. khan and s. mukherjee osamu y, noboru m, ken h (1991), formation and characterization of alkoxy-derived bi4ti3o12. british ceramic transactions 90: 111–113. pavlović nikolina, kancko dejan, szécsényi mészáros katalin, srdić v. vladimir (2009), synthesis and characterization of ce and la modified bismuth titanate. processing and application of ceramics 3(1-2): 88-95. sedlar m, sayer m (1996), structural and electrical properties of ferroelectric bismuth titanate thin films prepared by the sol–gel method. ceramic international 22(3): 241–247. shi yanhui, cao changsheng, feng shouhua (2000), hydrothermal synthesis and characterization of bi4ti3o12. materials letter 46: 270–273. wang-ming chun, wang feng jin (2008), aurivillius phase potassium bismuth titanate: k0.5bi4.5ti4o15. journal of the american ceramic society 91(3): 918-923. yang qunbao, li yongxiang, yin qingrui, wang peiling, cheng bing-yi (2003), bi4ti3o12 nanoparticles prepared by hydrothermal synthesis. journal of the european ceramic society 23: 161166. 34 microsoft word paper 1.docx the journal of engineering research, vol. 11, no. 2 (2014) 1-14 on sequence lengths of some special external exclusive or type lfsr structures – study and analysis a ahmad* and a al maashri department of electrical & computer engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, sultanate of oman received 14 february 2014; accepted 20 april 2014 abstract: the study of the length of pseudo-random binary sequences generated by linearfeedback shift registers (lfsrs) plays an important role in the design approaches of built-in selftest, cryptosystems, and other applications. however, certain lfsr structures might not be appropriate in some situations. given that determining the length of generated pseudo-random binary sequence is a complex task, therefore, before using an lfsr structure, it is essential to investigate the length and the properties of the sequence. this paper investigates some conditions and lfsr’s structures, which restrict the pseudo-random binary sequences’ generation to a certain fixed length. the outcomes of this paper are presented in the form of theorems, simulations, and analyses. we believe that these outcomes are of great importance to the designers of built-in self-test equipment, cryptosystems, and other applications such as radar, cdma, error correction, and monte carlo simulation. keywords: lfsr, pseudo-random binary sequence, seed, feedback connection, periodicity, exclusive or. אאאlfsr jא  א אwאאאאאאאאlfsr אאאאאאkאlfsrא אאkאאאאאא אlfsr،אאאאאlfsrkא אאlfsrאאאk        א    k א   א      א אאא א  א אאא   א    cmdakא אwlfsr،אאאא،אא،אא،א. *corresponding author’s e-mail: afaq@squ.edu.om a ahmad and a al maashri 2 1. introduction pseudo-random binary sequences (prbss) have been used for various applications. some of the application areas are built-in self-test (bist) for very large scale integration (vlsi) circuits’ design, cryptography applications like stream ciphers, and error correction and detection codes. in addition, prbs have been commonly used in the fields of digital signal processing, wireless communications, direct sequence spread spectrum, scrambling &descrambling, encryption & decryption, steganography, and many more (williams 1984; mccluskey 1985, bardell et al. 1987; nanda et al. 1989, ahmad 1997; jamil and ahmad 2002; ahmad 2005a, 2012, 2013a; hell and johansson 2008, mukherjee et al. 2011 and ayinala and parhi2011). linear feedback shift registers (lfsrs) are usually used for generating prbss (peterson and weldon 1984 and golomb 1981). in fact, lfsrs have been employed in a wide range of applications. this is due to several reasons: 1) lfsrs are well-suited to hardware implementation, 2) lfsrs can produce prbs with good statistical properties, 3) lfsrs can produce sequences of large periods with different frequencies, and 4) because of their structures; lfsrs can be readily analyzed using algebraic techniques. however, there are a number of design issues that need to be considered prior to integrating lfsr to a real application. some of these issues include the size of lfsr ‘n’, the seed ‘s’ (ie. initial state of the lfsr), feedback connection (fb) in the lfsr, and the type of the lfsr (ie. internal or external) using exclusive-or ‘xor’ or exclusive-nor ‘xnor’. some structures of lfsrs are constructed using internal xnor model with respect to their periodicity, which have been analyzed in (ahmad and al-maashri 2008) exploiting the state space model of xnor structures of lfsr (ahmad 2005b). in this paper, we consider internal xor model of lfsr structures to study some of the conditions, which restrict the lfsr to a particular periodicity. our study is based on a derived algebraic modeling of generated prbss by the lfsr. we further validate our results through simulation process. we also present a study on randomness criterion of those lfsr structures. the rest of this paper is organized as appears in sections 2 7. section 2 introduces lfsr model. in section 3 we present the derived algebraic model of an lfsr, whereas section 4 presents the analytical study. simulation model and runs are embodied in section 5, while section 6 presents a study on randomness criterion of prbss. finally, section 7 concludes the paper and discusses future work. also, an appendix (appendix a) is provided for the abbreviations and terminologies used in this paper. 2. lfsr – an introduction an lfsr is a special type of serial-in serial-out (siso) shift register that, when clocked, advances the signal through the register from one bit to the next mostsignificant bit. figure 1 shows an n-bit siso shift register. the key element of siso shift register is d-type flip-flops (ffs). the {q1, q2, qi, …qn-1, qn} are the states of the flipflops {d1, d2, …, di, … dn-1,dn}, respectively. siso shift register has two special features: 1) some of the outputs are combined internally or externally in exclusive-nor or exclusive-or configuration to form a feedback mechanism and 2) it retains the autonomous nature of lfsr that is the last output should be part of feedback mechanism. figure 2 shows an external type xor structure of an n-bit lfsr. the {c0, c1, c2, ….., ci,, …., cn-1, cn}, are the possible feedback connections. on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 3 figure 1. an n-bit siso shift register. figure 2. an n-bit lfsr (external xor type). therefore, an lfsr can be formed by either performing exclusive-or or exclusive-nor operations on the combined outputs of two or more of the flip-flops. the model for exclusive-nor has been presented in (ahmad and al-maashri 2008; ahmad 2005b). in this paper, however, our focus will be towards presenting an algebraic model for an n-bit external exclusive-or type lfsr. in an external exclusive-or type lfsr, the output of the aforementioned operations on the combined outputs of two or more of the ffs and the result is fed to the least significant ff (ie. q1) as shown in fig. 2. figure 3 shows an example of a 3-bit lfsr, which is constructed using external xor functional block. note how the feedback– which is fed as an input to the first ff – is the result of exclusive-or operation of the outputs of the second and third ffs. table 1 visualizes the operations of the lfsr depicted in fig. 3. the table elaborates the next states (ff1_out, ff2_out, and ff3_out) and the output sequence si. the used seed to start the operation is considered as q1 (0) = 1, q2 (0) = 0, and q3 (0) = 1. it is this feedback function that causes the register to loop through repetitive sequences of prbs value. the choice of feedback connections, the seed, and the value of ‘n’ determine the number of prbs values in a given sequence before the sequence repeats this length is known as periodicity ‘p’ of the lfsr (williams 1984;mccluskey 1985; nanda et al. 1989; ahmad 1997; hell and johansson 2008; mukherjee et al. 2011; ayinala and parhi2011; peterson and data in clock d1 q1 d2 q2 di qi dn-1 qn-1 dn qn d ata o ut ... ... q1 q2 qn q3 … qn-1 generated prbs eor function block (external) c0 c1 c2 ... cn-1 cn ff1_out d1 q1 ff1 d2 q2 ff2 d2 q2 ff3 ff2_out ff3_out c0 c2 c3 si figure 3. a 3-bit lfsr (external xor type). a ahmad and a al maashri 4 weldon, jr. 1984; golomb 1981; ahmad and al-maashri 2008; ahmad 2005b; williams et al. 1988; ahmad 1990; ahmad and elabdalla 1997; ahmad 1994; knuth 1997; ahmad et al. 2002; ahmad 2002; krishnaswamy and pillai 2012, peinado and fuster-sabater, 2013; chunqiang et al. 2012; ming-hung 2013). 3. lfsr as prbs generator – an algebraic modeling in this section, we present a generalized algebraic model exclusively for an n-bit external exclusive-or type lfsr based prbs generator. any binary data sequence can be represented in form of polynomial in gf(2). therefore, the feedback connection vector for an lfsr can be represented in the form of a polynomial and is technically known as a characteristic polynomial. eqn. (1) define a general form of a characteristic polynomial and let us call it ( ). ( ) = × (1) let {am}= [a0 , a1 , . . , ai, . .], represent the output sequence generated by the lfsr used as prbs, where ai= 0 or 1. then this sequence can be represented as given in eqn. (2). ( ) = × (2) from the structure of the type of the lfsr shown in fig. 2, it can be seen that if the current state of the ith flip-flop is am-i , for i = 1, 2 , ... , n , then by the recurrence relation an equation can be given as depicted in eqn. (3). = × (3) the generating function ( ) associated with the prbs can be mathematically defined as in eqn. (4). i ii xaxg ×= ∞ =0 )( (4) or ( ) = ∑ ∑ × ( )∑ × (5) or, eqn. (4) can be rewritten as: ( ) = ( ) = ( )( ) (6) the ( ) and ( ) can be written in an expanded form as described by eqns. (8) and (10), respectively. ( ) = ∑ × (7) ( ) = ( × ) + ( × ) + ⋯+ ( × ) (8) table 1.next state sequences (prbs) for the structure of lfsr of figure 3. clock q1 (ff1_out) q2 (ff2_out) q3 (ff3_out) si 0 1 0 1 …11101001110 1 1 1 0 2 1 1 1 3 0 1 1 4 0 0 1 5 1 0 0 6 0 1 0 7 1 0 1 repeats on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 5 or, ( ) = 1 + ( × ) + ⋯+ ( × )(9) ( ) = ( × × ) + ( × × ) +( × × ) + ( × × ) +( × × ) + ( × × ) + ⋯+ ( × × ) + ⋯+ ( × × ) + ( × ) (10) as an example, let us consider the lfsr structure shown in fig. 3. in this structure, n = 1, q1 (0) = 1, q2 (0) = 0, q3 (0) = 1, c0 = 1, c1 = 0, c2 = 1and c3 = 1. hence, ( ) and ( )can be derived in the forms of polynomials as given in eqns. (11) and (12) respectively. ( ) = 1 + ( ) + ( ) (11) ( ) = 1 + ( ) + ( ) = 1 (12) computing ( ) = ( )/ ( ), we get the result as shown in fig. 4. the quotient of this long division process is prbs in a polynomial form. this result is validated by crosschecking with those presented in table 1. figure 4. long-division computation of ( ) = ( )/ ( ). . . . . . . . a ahmad and a al maashri 6 4. analytical study this section presents a study on some special cases where the lfsrs are restricted to generate prbss of limited periodicity. the study covers the roles of all parameters related to the lfsr generating the prbss. these parameters are ‘n’, ‘seed’ and the feedback connection function (‘fb’). the value of n may be either even or odd, seed may vary from (0)10to (2n-1)10.the fb function depends on input connections coming from c1, c2, ….., ci,, …., cn-1, cn links to the xor function block. we present the results of our study in the forms of theorems supported with proofs using algebraic model of lfsr presented in section 2.throughout the study, we consider an nbit xor structure of lfsr. theorem 1: “if the seed value in the lfsr is 0 ( = (0)10), then for any value of n and for any fb function the period ‘p’ of generated prbs by the lfsr will be 1 (p = 1).” proof: since by substituting ( = (0)10), in eqn. (5), the equation reduces to a reduction modular (mod) equation as given below. 0 × = 0 (13) hence the generated prbs is , = 0 which is the seed value. hence this proves that the period ‘p’ of generated prbs by the lfsr is 1. theorem 2: “if the seed value in the lfsr is all ones, = (2n-1)10, where n is odd and fb function is considered from all the links [c1, c2, ….., ci,, …., cn-1, cn], then the period ‘p’ of generated prbs by the lfsr will be 1 (p = 1).” proof: using eqns. (8) and (9) we can write ( ) = 1 + ( × ) + ⋯+ ( × ), ( ) = ( ) + ( ) + ( ) + ⋯+ ( ). the result of the long division process of eqn. (14) or eqn. 15 produces ( ) = (2n-1)10, which proves that the ‘p’ of ( ) can be given as: ( ) = 1 + + ⋯+1 + + ⋯+ (14) or, × × = (15) the value ∑ implicates that the generated prbs by the lfsr structure set in theorem 2 is 1. theorem 3: “if the seed value in the lfsr is all ones, = (2n-1)10, and in the total number of considered links in the fb function from the [c1, c2, ….., ci, …., cn-1, cn] is odd, then the period ‘p’ of generated prbs by the lfsr will be 1 (p = 1).” proof: by substituting the seed value = (2n-1)10, and inserting fb function for the said structure in theorem 3 in eqns. (8) and (9) we get: (1 + ) = (16) therefore, the next state of the lfsr will be the same as the seed and hence it proves that the period of the generated prbs by the lfsr structure of theorem 3 will be 1, ie. (p = 1). on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 7 theorem 4: “if the fb function is considered only from the link [cn] and the seed value in the lfsr is any value except all zeros, = (0)10 or all ones, = (2n-1)10, then the period ‘p’ of generated prbs by the lfsr will be n (p = n).” proof: let n = 3 and q1 (0) = 1, q2 (0) = 1, and q3 (0) = 0. using equation (9), the numerator ( ) can be computed as ( ) = 1 + . therefore, dividing ( ) by ( ), [ ( ) =1 + ] we get: ( ) = 1 + + + + + + ⋯ =(((1 + ) × ) × ). this demonstrate that the period ‘p’ of generated prbs by the lfsr is 3 (i.e. p = n). we can consider any value of n, resulting ( ) = (((1 + ) × ) × )). theorem 5: “if fb function is considered from all the links [c1, c2, ….., ci,, …., cn-1, cn], and the seed value in the lfsr is any value except all zeros, = (0)10 or all ones, = (2n-1)10, then the period ‘p’ of generated prbs by the lfsr will be n +1 (p = n +1).” proof: let n = 3 and q1 (0) = 1, q2 (0) = 1, and q3 (0) = 0. using eqn. (9) the numerator ( ) can be computed as ( ) = 1. therefore, dividing ( ) by ( ), [ ( ) = 1 + + + ] we get ( ) = 1 + + + + ++ ⋯ = (((1 + ) × ) × )). this demonstrates that the period ‘p’ of generated prbs by the lfsr is 3 (ie. p = n). we can consider any value of n, resulting ( ) = (((1 + ) × ) × . theorems 1-5 presented in section 4 have great values. firstly, from point of view of applications of lfsrs and secondly, the described theorems shall help in deducing the lfsr structures of maximal length sequences. considering the interests of practicing engineers, we present table 2 to demonstrate how guidelines and restrictions can be ascertained while using lfsrs for its practical usage. for demonstrating the applicability of theorems 1 and 3-5, all possible lfsr’s structures of order n = 4 are considered. also, the applicability of the study helps in searching the generator for maximal length sequences. as can be visualized from table 2, for n = 4, there exist 8 possible fb functions, out of those 8, 5 of them have restrictions. hence, the search set is reduced to 3 as (1 + + ), (1 + +), and (1 + + ). 5. simulation model a simulation model was developed to validate the analytical study presented in section 3. the model was developed in matlab to simulate the behavior of lfsr structures. figure 5 illustrates the two simulation models that were developed to validate the theorems. figure 5a depicts the model “prb_single_seed”, which is capable of generating prbs ‘prbs’ as function of feedback connection ‘fb’ and seed‘s’. in simulation, the length of the generated sequence is controlled by a set length ‘l’, computed as follows: l = 2*(2n-1). a ahmad and a al maashri 8 (a) (b) figure 5. the two simulation models that are used to validate the theorems presented in analytical study. (a) used to validate the theorems 1, 2, and 3 (b) used to validate the theorems 4 & 5.the model “prb_single_seed” is used to validate the theorems 1, 2, and 3. prb_single_seed fb s l prbs prb_all_seed fb p table 2. demonstrating guideline and restrictions based on theorem 1–5. study possible feedback connection functions of order 4 (n = 4) comment 1+ 1++ 1+ + 1++ + 1+ + 1++ + 1+ ++ 1++++ theorem 1 *g1 *g1 *g1 *g1 *g1 *g1 *g1 *g1 *g1: guideline do not use = (0)10 theorem 3 *r3 *r3 *r3 *r3 *r3: restriction do not use = (2n-1)10 theorem 4 *r4 *r4: restriction restricts p to p = n for any value of except as restricted by theorems 1 and 3 theorem 5 *r5 *r5: restriction restricts p as: p = n + 1 for any value of except as restricted by theorems 1 and 3 theorem 2 n = 5; 1+ + + + + *r2: restriction restricts p as p = 1 for a value of , as = (2n-1)10 n = 7; 1 + + + + + + + n = 9; 1 + + + + + + + + + on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 9 table 3 shows a representative result set out of the simulation results when running the model to validate theorem 1. the table demonstrates that for any possible ‘fb’ the period of ‘prbs’ is 1. similarly, table 4 shows the simulation results for validating theorems 2 and 3. to validate theorems 4 and 5, we use another simulation model called “prb_all_seed” (see figure 5b). unlike the model above, “prb_all_seed” examines the lfsr structures by generating all possible seeds, while requiring only ‘fb’ to compute the output p of the function. the sample results for the runs are given in tables 5 and 6 for validations of theorems 4 and 5, respectively. 6. justification of the study due to their good statistical properties, the lfsrs generating maximal length prbss are widely used in stream ciphers (knuth 1997; ahmad et al. 2001). the maximal length prbss are popularly known as msequence or pseudo noise (pn) sequence. the maximal length prbss have period length of 2n-1 (where n is the length of the lfsr). such lfsrs, which generates msequence, are realized when the corresponding fb to the lfsr is primitive (peterson and weldon, jr. 1984; golomb 1981; ahmad et al. 1990; ahmad and elabdalla1997; knuth 1997; chunqiang et al. 2012; ahmad et. al 2013b; ming-hung 2013). it is imperative for the designers of crypto systems to consider suitable criteria table 3. results of simulation runs for theorem 1. theorem n p theorem 1 2 case 1: fb=[1,1,1]; s=[0,0]; prbs = 0 0 0 0 0 0 case 2: fb=[1,0,1]; s=[0,0]; prbs= 0 0 0 0 0 0 p = 1 for all the cases 3 case 1: fb=[1,1,1,1]; s=[0,0,0]; prbs = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 case 2: fb=[1,0,0,1]; s=[0,0,0]; prbs = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 case 3: fb=[1,0,1,1]; s=[0,0,0]; prbs= 0 0 0 0 0 0 0 0 0 0 0 0 0 0 case 4: fb=[1,1,0,1]; s=[0,0,0]; prbs = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a ahmad and a al maashri 10 fb = feedback connection; s = seed; prbs = pseudorandom binary sequence table 5. results of simulation runs for theorem 4. theorem n p theorem 4 2 fb = [1,0,1]; p = 2 p = n for all the cases 3 fb =[1,0,0,1]; p = 3 5 fb = [1,0,0,0,0,1]; p = 5 21 fb = [1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1]; p = 21 41 fb = [1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,1]; p = 41 fb = feedback connection; p = periodicity fb = feedback connection; p = periodicity table 4. results of simulation runs for theorems 2 and 3. theorem n p theorem 2 3 fb = [1,1,1,1]; s = [1,1,1]; prbs = 1 1 1 1 1 1 1 1 1 1 1 1 1 1 p = 1 for all the cases 5 fb = [1,1,1,1,1,1]; s = [1,1,1,1,1]; prbs = 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 theorem 3 4 fb = [1,0,1,1,1]; s = [1,1,1,1]; prbs = 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 table 6. results of simulation runs for theorem 5. theorem n p theorem 5 16 fb = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]; p = 17 p = n + 1 for all the cases 32 fb = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]; p = 33 64 fb = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]; p = 65 on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 11 for the selection of a key stream generator in the design. some of these design criteria are statistical measures, period and linear complexity. acceptable prbss should exhibit no statistical bias in occurrence of individual symbols or small block of symbols. in this regard, golomb’s postulates defined in (ahmad et al. 2002; ahmad et al. 2013c; golomb 1981)suggest that a prbs that passes the tests for randomness will be acceptable for the use of cryptosystems and other applications like radar, code division multiple access (cdma), error correction and monte carlo simulation. for the sake of completeness, we list the postulates outlined in (ahmad et al. 2002; ahmad et al, 2013c;golomb 1981); namely, 1) prbs length, 2) balance of 1’s and 0’s, 3) run property, and 4) ideal autocorrelation. in this paper, the study of the theorems 1, 2 and 3 provide the boundary situations where lfsrs lock and fail to generate prbss of sufficient length. also lfsr structures and seed combinations are to be avoided in length. because of this, lfsr structures and applications of bist as test pattern generators, and in cryptography as key generators. to demonstrate the level of prohibitions and utilizations of the lfsr structures defined in theorems 4 and 5, we considered the criterions of golomb’s postulates. the ratios of the lengths of generated prbs using the generators defined in theorems 4 and 5 with respect to the maximal length prbs of 2n-1 (where n is the length of the lfsr) are shown in table 7 for n = {2, 3, 4, 5, 6, 7, 8, 16, 32, 64, 128}. in the table, r4ml and r5ml represent the ratios due to the described prbs generators of theorems 4 and 5, respectively. in addition, table 7 demonstrates the balance properties of the lfsr structures defined via theorems 4 and 5. moreover, r4mo and r5mo represent the ratios due to the described prbs generators of theorems 4 and 5, respectively, of maximum possible number of 1’s. whereas, r4mz and r5mz represent the ratios due to the described prbs generators of theorems 4 and 5, respectively, of maximum possible number of zeroes. 7. conclusion and future work prbs serves an important role in a diversified collection of application domains; including cryptography and fault tolerance. this work has investigated the properties of lfsr circuits used for generating prbs periods and highlighted the behavior of some of the lfsr structures. the study has employed the recurrence relations to describe the prbs periods generated by the lfsr structures. a number of theorems have been presented in this paper. these theorems summarize the observations that were outlined throughout the discussion of the analytical model. these observations add some knowledge towards the generation of maximal length prbs. the theorems and their subsequent outcomes were validated using a simulation model. the focus of this study was on prbs generation; however, lfsrs could also be the building block of other correlation functions. as future work, it would be interesting to investigate further properties and observations on the lfsr circuits that could be of use in other application domains and functions. acknowledgments the authors would like to express their great appreciations and gratitude to sultan qaboos university, sultanate of oman for providing research facilities, technical supports and research environment. references ahmad a (1994), critical role of polynomial seeds on the effectiveness of an lfsrbased testing technique. international journal of electronics 77:127–137. ahmad a (1997), achievement of higher testability goals through the modification of shift register in lfsr based testing. international journal of electronics 82:249–260. a ahmad and a al maashri 12 ahmad a (2002), constant error masking behavior of an internal xor type signature analyzer due to the changed polynomial seeds. journal of computers & electrical engineering 28:577–589. ahmad a (2005a), testing of complex integrated circuits (ics)–the bottlenecks and solutions. asian journal of information technology4:816–822. ahmad a (2005b), development of state model theory for external exclusive-nor type lfsr structures. world enformatika society-transactions on engineering, computing and technology 10:12–19. ahmad a (2012), better pn generators for cdma application–a verilog-hdl implementation approach. international journal of information engineering 2:6– 11. ahmad a (2013a), development realization of a better signature analysis scheme by adding a bit to the size of 8k. international journal of information engineering 3:122–128. ahmad a, al-busaidi ss, al-maashri a, awadalla m, rizvi mak, mohanan n (2013b), computing and listing of possible number of m-sequence generators of order n. indian journal of science and technology10:5359–5369. ahmad a, al-busaidi ss, al-mushrafi mj (2013c), on properties of pn sequences generated by lfsr—a generalized study and simulation modeling. indian journal of science and technology10:5351–5358. ahmad a, al-maashri a (2008), investigating some special sequence lengths generated in an external exclusive-nor type lfsr. journal of computers and electrical engineering 34:270–280. ahmad a, al-musharafi mj, al-busaidi s (2002), study and implementation of properties of m-sequences in matlabsimulink—a pass/fail test tool for designs of random generators. squ journal of scientific research–science and technology 7:147–156. ahmad a, al-musharafi mj, al-busaidi s, al-naamany a, jervase ja (2001), an nlfsr based sequence generation for stream ciphers. proceeding international conference on sequences and their applications, bergen, norway, may 1317, 11–12. ahmad a, elabdalla am (1997), an efficient method to determine linear feedback connections in shift registers that generate maximal length pseudo-random up and down binary sequences. computer & electrical engineering—an international journal 23:33–39. ahmad a, nanda nk, garg k (1997), are primitive polynomials always best in signature analysis? ieee design & test of computers 7:36–38. ayinala m, parhi k (2011), high-speed parallel architectures for linear feedback shift registers. ieee transactions on signal processing 59(9):4459–4469. bardell ph, mcanney wh, savir j (1987), built-in-test for vlsi. new york: john wiley. chunqiang h, xiaofeng l, xiuzhen (2012), verifiable multi-secret sharing based on lrsr sequences. journal of theoretical computer science 445:52-62. golomb sw (1981), shift register sequence. walnut creek, ca: aegean park press. hell m, johansson t, maximov a, meier w (2008), new stream cipher designs—the estream finalists, springer 4986:179– 190. hu cq, liao xf, cheng xh (2012), verifiable multi-secret sharing based on lfsr sequences. journal of theoretical computer science (elsevier) 45:52–62. jamil t, ahmad a (2002), an investigation into the application of linear feedback shift registers for steganography. proceeding ieee southeast con, sc, usa. april 2002. kao mh (2013), on the optimality of extended maximal length linear feedback shift register sequences. journal statistics & probability letters 83:1479–1483. on sequence lengths of some special external exclusive or type lfsr structures – study and analysis 13 knuth d (1997), the art of computer programming, semi numerical algorithms. reading, ma: addisonwesley 2. krishnaswamy s, pillai hk (2012), on the number of linear feedback shift registers with a special structure. ieee transactions on information theory 58(3):1783–1790. mccluskey ej (1985), built-in self-test techniques. ieee design & test of computers 2(2):21–28. ming-hung k (2013), on the optimality of extended maximal length linear feedback shift register sequences. journal of statistics and probability letters 83(6):1479-1483. mukherjee n, rajski j, mrugalski g, pogiel a (2011), ring generator: an ultimate linear feedback shift register. ieee computer 44(6):64–71. nanda nk, ahmad a, gaindhar vc (1989), shift register modification for multipurpose use in combinational circuit testing. international journal of electronics 66(6):875–878. peinado a, fuster-sabater a (2013), generation of pseudorandom binary sequences by means of linear feedback shift registers (lfsrs) with dynamic feedback. mathematical and computer modeling 57(6): 2596–2604. peterson ww, weldon ej jr (1984), errorcorrecting codes. 2nd ed., cambridge, ma: massachusetts institute of technology press. williams tw (1984), vlsi testing. ieee computer c-17(10):126–136. williams tw, daehn w, gruetzner m, starke cw (1988), bounds and analysis of aliasing errors in linear feedback shift registers. ieee trans. computer aideddesign cad-7(1):75–83. a ahmad and a al maashri 14 appendix a list of abbreviations and terminology used in the paper prbs pseudorandom binary sequence bist built-in self-test vlsi very-large-scale integration lfsr linear feedback shift register xor exclusive-or xnor exclusive-nor siso serial-in serial-out n lfsr size ci the ith feedback connection ff flip-flop qi the state of the ith ff di the input to the ith ff gf(2) galois field mod 2 d(x) characteristic polynomial of lfsr s generated prbs seed initial state of lfsr d:\desktop\paper 5.xps tjer vol. 11, no. 1, 44-54 __________________________________________ *corresponding author’s e-mail: mmubaraki@jazanu.edu.sa identification of pavement distress types and pavement condition evaluation based on network level inspection for jazan city road network m mubaraki department of civil engineering, college of engineering, the university of jazan, po box 706, saudi arabia received 30 december 2012; accepted 26 september 2013 abstract: the first step in establishing a pavement management system (pms) is road network identification. an important feature of a pms is the ability to determine the current condition of a road network and predict its future condition. pavement condition evaluation may involve structure, roughness, surface distress, and safety evaluation. in this study, a pavement distress condition rating procedure was used to achieve the objectives of this study. the main objectives of this study were to identify the common types of distress that exist on the jazan road network (jrn), either on main roads or secondary roads, and to evaluate the pavement condition based on network level inspection. the study was conducted by collecting pavement distress types from 227 sample units on main roads and 500 sample units from secondary roads. data were examined through analysis of common types of distress identified in both main and secondary roads. through these data, pavement condition index (pci) for each sample unit was then calculated. through these calculations, average pcis for the main and secondary roads were determined. results indicated that the most common pavement distress types on main roads were patching and utility cut patching, longitudinal and transverse cracking, polished aggregate, weathering and raveling, and alligator cracking. the most common pavement distress types on secondary roads were weathering and raveling, patching and utility cut patching, longitudinal and transverse cracking, potholes, and alligator cracking. the results also indicated that 65% of jazan's main road network has an average pavement condition rating of very good while only 30% of jazan's secondary roads network has an average pavement condition. keywords: pavements, maintenance, rehabilitation, distress types, pci. 45 m mubaraki 1. introduction the city of jazan, in saudi arabia, has rapidly developed during the last 10 years. city pavement requires maintenance because of the impact from traffic, cuts in the roadway by utilities, and climatic conditions (al-swailmi 1994). therefore, evaluation of pavement condition is a vital process for municipalities to preserve capital infrastructure and maximize its benefits. generally, pavement condition evaluation may involve the evaluation of structural pavement, roughness, surface distress, and safety or skid resistance. pavement engineers have long recognized the importance of distress information in quantifying pavement quality. this information has been used to document present pavement conditions, chart past performance history, and predict future pavement condition (shahin 2005). pavement distress information is also broadly used as the only quality measure in many pavement management systems (pms). this is particularly true for systems used by municipalities in urban areas where measurements of roughness and skid resistance are not performed because of lack of relative applicability (mubaraki 2004). on the other hand, the interrelationship between structural and functional evaluation for the pavement can be considered because there is no direct relationship. therefore, both functional and structural evaluations of pavements are important and are meant to supplement rather than replace one other. an observation of functionality below the acceptable level is one way to trigger a structural evaluation. the rough pavement may be structurally sound and require only a level up surface layer, or it may be structurally weak and require replacement or a thick overlay (haas et al. 1994). pavement distresses are visible imperfections on the surface of pavements. there are symptoms of the deterioration of pavement structures. most if not all agencies have implemented pms and collect periodic surface distress information on their pavement through distress surveys (haas et al. 1994). surface distress evaluation is the most prevalent way to evaluate pavement condition in saudi arabia major cities of riyadh, dammam, madinah almunawarah, and the holy city of makkah (mubaraki 2010). the officials of jazan municipality are very anxious to upgrade the existing pavement practices to rational, systematic, and practical ones. research has been started by collecting information on the network in order to build a pms. this paper concerns identifying pavement distress types for the entire network. each distress, in almost all distress evaluation methodologies, is specified by severity level (low, medium, high) and an extent level described in measurable units (linear or area). a distress type, and levels of severitty and extent are assigned a deducted value, which is an indication of how this combination, when present, affects the perfect pavement. however, road agencies or municipalities do differ in handling distress data, and it is very difficult to conclude that one system is better than the others. therefore, each agency should select or develop a procedure to deal with distress data so that it suits their municipalities' needs and capabilities. once the distress data for a network are available, maintenance needs and budget can be determined. each type of distress is the result of one or more variables, which provides great insight into the causes of pavement deterioration. thus, for each pavement type, the proper pavement evaluation program and procedure should include the identification of pavement distress types, severity, and the number of distresses (shahin 2005). pavement distress information is usually converted into a condition index. the condition index combines information from all of the distress types, severities, and quantities into a single number. this number can be used at the network level to define the condition state, identify when treatments are needed, rank or prioritize, and to derive the number used to forecast pavement condition. the condition index may represent a single distress such as fatigue cracking or a combination of many pavement distresses which is then usually referred to as a composite index. additional information has also been included in some indices such as traffic levels and highway class to produce priority ranking indices. the u.s. army corps of engineers developed a very complete condition index for a pavement management system in 1976 (baladi and snyuder 1992). this includes pavement condition survey procedures and a detailed method for calculating a pavement condition index (pci), which is still used today by many agencies. the computational procedures for this index will be shown later in this module. condition indices are used in most pavement management systems for the following four basic reasons (deighton and sztraka 1995), in order to trigger treatments, calculate life-cycle costs, evaluate the network condition, and make use of the same relative scale between systems. the main objective of this paper is to identify pavement distress types for the entire jazan network using the paver procedure and, as a result of that, investigating the network pavement condition. to achieve these objectives, the study involves sampling and data collection, identifying and selecting common types of distress, and evaluating pavement conditions. 46 identification of pavement distress types and pavement condition evaluation based on network level ..... 2. pavement condition 2.1 introduction it is common for agencies to describe pavement distress in terms of severity and extent. severity indicates how bad the distress is while extent indicates the quantity of distress. extent can be estimated for an entire section length or over a representative area (such as 100 meters per km), or measured. together, these two parameters can describe a great deal about a particular distress. to transform this data into a meaningful condition index, deduct values are needed. deduct values are points which are used to compute the index based on the severity and extent of the distress represented. the development and calibration of the proper deduct value is the most complicated and critical part in the development of a pci. the index base is calculated by deducting a number of points from the index value of a section of pavement in perfect condition depending on the severity and extent of the deficiency. the value deducted also depends on the condition of the pavement. the number of points deducted is called the deduct value. the deduct values should be scaled such that the resulting condition index threshold value (or action point) occurs at about the middle of the scale (deighton and sztraka 1995). in the past, some have established the "should consider action" level at about 60 % of the scale and the "must consider action" level at 40% of the scale. moreover, the transition of the deduct values through the various levels of the distress matrix should produce a condition index that transitions as smoothly as possible with time. pavement distress is usually observed in the field as a continuous process with time, as the distress progresses through the full range of severity and extent. the trends of the pci in the pms should correspond to the trends observed in the specific pavement section represented by that index. most deficiencies, once they become apparent, tend to increase in both severity and extent at an increasing rate with time. thus, the pci which is the numeric representation of the pavement condition in the field, should have the same trends with time as the deficiency appears to have in the field. in general, those pavements that deteriorate rapidly after the last treatment tend to have a fairly linear form (ie. the rate of change in pavement condition is about the same from one year to the next). the pavement that lasts longer before some distress is observed tends to be more exponential in form. pavements that have lasted an unusually long time before distress occurs tend to deteriorate quite rapidly in the end; thus, they appear to have a very sharp exponential trend. the very simple formula which uses a deduct value to compute a distress index is shown in the following equation: pcii = pcimax-deduct where pcii = individual condition index based on measured condition 1 pcimax= value for perfect condition with no measured defects deduct = deduct value assigned to distress type, severity and extent obviously, using this equation implies that the condition index gets worse as the deduct value increases. assume an agency uses an index with a scale from 0 (bad) to 100 ( perfect = dmax,i ). if the pavement was in perfect condition, the deduct value would be 0, resulting in an index value of 100. if the pavement was in terrible condition, the deduct value would be 100, resulting in an index value of 0. the relative value of the pavement distress index which represents the condition of the pavement and the shape of the resulting pavement deterioration depends entirely upon the development of the deduct values. two basic approaches are often used to develop deduct value: expert opinion, and engineering or mathematical input. astm standard d5340 provides a set of deduct value versus extent curves based on work done by baladi and snyder (1992). the x-axis represents the density or extent, while the y-axis represents the deduct values for each distress individually. whatever method is chosen to assign deduct values for the individual condition indices is up to each agency. each method has its strengths and weaknesses, and, each method can produce a different shape in the performance curve. 2.2 background of paver development the development of paver was begun in 1974 by the army corps of engineers through funding from the military services to manage the vast inventory of pavement in the u.s. department of defense (dod). the intent was to build a pavement database. in 1976, the engineers developed a very important application with the help of technology-a pci for airfield pavement. the pci for roads and parking lots was developed in 1979. the pci, inventory database, and basic reporting capabilities constituted the first version of paver, and an official prototype evaluation test was conducted at fort eustis, virginia, usa. in the early 1980s, the american public works association (apwa) adopted the system for use by cities and counties (shahin and walteher 1990). after development, users in the field began to ask for the ability to analyze the consequence of different budgets on future pavement condition and the backlog 47 m mubaraki of deferred major maintenance and rehabilitation (m&r) projects. as result of this effort, pci prediction models for different pavement types, namely asphalt concrete (ac) and portland cement concrete (pcc), were developed. after over 3 years of intensive data collection and research efforts, it was concluded that there is no such thing as an accurate model that fits all pavements, no matter how many independent variables are included in the prediction models. variables considered were load (frequency and intensity), climate (precipitation, temperature, freezing index, and solar radiation), and calculated stresses and strains in the pavement. as a result of developing a pci, the engineers started to build a prediction engine called paver. the paver prediction engine enables users to develop condition deterioration trends for each family of uniform pavements in their agency. these trends are then used to predict pavement condition and perform the m&r budget consequence analysis. in the mid-1980s, the usa's federal aviation agency (faa) funded the production of desktop paver, and it became widely used by airports. in 1993, the american society for testing and materials (astm) adopted the airfield pavement pci as a standard. this pci was followed a few years later by the adoption of the pci for roads and parking lots. now the paver system is widely used by municipalities and road agencies all over the world. the paver building components included inventory, condition assessment, condition prediction, performance analysis, budget consequence, optimization (work planning), and project formulation. paver inventory management is based on a hierarchical structure composed of networks, branches, and sections, with the section being the smallest managed unit. this structure allows users to easily organize their inventory while providing numerous fields and levels for storing pavement data. to assess pavement condition, paver uses the pci as its primary condition index. pci measures the pavement structural integrity and surface operational condition on a scale from 0 to 100. it is calculated from measured pavement distress types, severities, and quantities and was developed to agree with collective judgment of experienced pavement engineers. in addition to the pci, paver allows users to create other condition indices based on the pci distresses already stored in the system. paver uses the pci, with pci based on subjective observations. the index itself must be both objective and systematic to be of value. the pci is calculated based on the results of a visual condition survey in which distress type, severity, and quantity are identified. field verification of the pci inspection method has shown that the index gives a good indication of a pavement's structural integrity and operational condition. it provides a useful index of both the current condition and an indication of future performance. the degree of pavement deterioration is a function of distress type and severity, and the density of distress. for this reason, deduct values were introduced as a type of weighing factor to indicate the size of the effect that each particular distress type, severity level, and distress density combination has on pavement condition. a pci needs to be based on manageable road sections, a roads inventory, and a classification and rating system for road defects. in order to develop a pci, the road network needs to be divided into manageable segments. sections with a relatively uniform pavement structure, design and traffic volume will have similar performance characteristics. in urban settings, sections should be kept to a manageable length-typically one block long. some road authorities limit the length to 100-150 meters for problematic sections. other authorities will use longer segments for roads that are consistent throughout their length. road sections in rural settings can be considerably longer-in some cases as much as 10 kilometres. each road section needs a unique identification so that the pci observations can be maintained in a database. each road section should have a basic history attached to it including class, length, width, geometry, type and volume of traffic, pavement type (whether it is flexible, rigid, or composite), original construction date, maintenance and rehabilitation history, and current condition based on the last pci. 2.3 pci calculation procedure these steps and the examples have been summarized from the annual book of astm standards (2011). step1: determine deduct values * calculate the density percentage [density = (quantity /sample unit area) * 100] * determine the deduct value from the deduct value curves (given an example for alligator cracking as shown in fig. 1) step 2: determine the maximum allowable number of deducts (m) * determine the maximum allowable number of deducts [m= 1 + (9/98)(100-hdv)] where hdv is the highest deduct value. * reduce the number of deduct values to m. if the deduct values are fewer than m, then all of the deduct values should be taken. * list the deduct values in a descending order, given an example as in table 1. step 3: determine the maximum number of corrected deduct value (cdv) 48 identification of pavement distress types and pavement condition evaluation based on network level ..... * determine the number of deducts >2, q * determine the total deduct values by adding all individual deduct values * determine the cdv form q and total deduct value for the first iteration by looking for the appropriate correction * set the minimum deduct value >2 to 2 in the next iteration and repeat the steps * select the maximum cdv form the calculated cdvs as the largest value step 4: calculate pci by subtracting the maximum cdv from 100 3. sampling and data collection jazan municipality divides jazan city into 25 districts. a district is an area surrounded by four main roads, and each district covers a number of regions. a region is an area that contains a number of secondary roads surrounded by four main roads within a district. there are a total number of 110 regions in the city. jazan's road network is divided into two categories: main and secondary roads. main roads represent 22% of the total network area and are defined as roads with a middle island, or with a total width of more than 30 figure 1. deduct value for alligator cracking. table 1. example of deduct values for typical road degradation. 49 m mubaraki meters without a middle island. secondary roads account for approximately 78% of the total network area and represent the roads inside a defined region surrounded by four main roads. every category from the two roads is divided into branches, and the branches are divided into sections. the section for a main road is a defined distance between two intersections in the main roads. the section for a secondary road is defined as the area surrounded by four main roads. the sections for both main and secondary roads are divided into a number of sample units. a sample unit for a main road is 100 meters per lane from the section. a sample unit of the secondary main road is the distance between two intersections. jazan municipality was started as a project to develop a systematic way for managing the road network. gathering information about the city's road network and pavement condition survey were started to build a good database about the city roads in order to reach to a pms. the survey was based on the paver condition survey procedure. the procedure involves determining distress types and quantities, and severity levels for each sample unit. it is recommended that the sample units to be inspected be spaced equally throughout the section, and the first one be chosen at random. therefore, a systematic random sampling technique was used to select sample units. this technique can be achieved by finding (n) and (n) where n equals the total number of available sample units and n equals the minimum of sample units to be surveyed, and then dividing n by n to get the sampling interval (i). random start (s) is selected at random between sample unit 1 and i. the sample units to be surveyed are identified as s, s + 1, s + 2i, etc. in this paper, the network level inspection was used to conduct the pavement condition evaluation. a network level inspection can be conducted by surveying only a few sample units per section compared to project level inspection. the paper used the following table 1, which is based on eqn.1 (shahin 2005). (eq. 1) figure 2. an example corrected deduct value. table 2. an example for pci calculation. pci = 100 68 = 32 50 identification of pavement distress types and pavement condition evaluation based on network level ..... the degree of sampling presented in table 1 is sufficient for assessing the condition of the pavement (shahin 2005) as well as others, such as maintenance plans. table 1 shows that the recommended values for network level inspection was based on a 95% confidence level, which means the degree of sampling taking in account in this study is reliable with an error of only 5%. where n = total number of sample units in the pavement section e = allowable error in the estimate of the section pci (e = 5) s = standard deviation of pci between sample units in the section (s = 5) this equation provides a consistent method for selecting the number of units to inspect for different size sections and this is the case for this paper's data. the rationale behind this is that management at the project level requires accurate data for the preparation of work plans and contracts. therefore, more sample units were inspected than usually sampled for network level management. using this number, a reasonable estimate of the true mean pci of the section will be obtained. there is 95% confidence that the estimate is within ± 5 points of the true mean pci (shahin 2005). the condition survey was conducted by inspectors. each inspector was sent to a particular section of the network to record the existing distress types, quantities and severity level in each selected sample unit within the section. the inspector's work supervised by an engineer to ensure work quality. out of the 25 districts in jazan city only 5 have pavement evaluation records so far. out of the total 50 main roads, only 11 have pavement evaluation records so far. eleven main roads selected from different locations within the city were included in the analysis. within the selected main roads, 27 sections were considered. the 27 sections were subdivided into 227 sample units with a total length of 22,270 meters. from the five districts that have data, a total of 10 regions were selected. the total numbers of sample units within these regions were 500, with a total area of 769,000 square meters. 4. identification and selection of common types of distress 4.1 common distress on main roads figure 3 shows the percentage of distribution of distress types found on main road networks at three severity levels-low, medium, and high. table 3. network level sampling based on equation 1. figure 3. percentage distribution of distress density on main road networks. distress type percentage distribution of distress density on main roads 51 m mubaraki from this figure, it is clear that 66% of the distresses found on the main roads network at a low severity level is patching and utility cut patching (d11). types of longitudinal and transverse cracking (d10) represent about 55% of the distress types found at the lowseverity level. other distress types found at low severity level include polished aggregate (d12), weathering and raveling (d19), rutting (d15), alligator cracking (d1), depression (d6), and bleeding (d2). about 42% of the distress types at the medium severity level are longitudinal and transverse cracking (d10). another 22% of the distress types at the medium severity level are the patching and utility cut patching (d11). other distress types found at the medium severity level were polished aggregate (d12), weathering and raveling (d19), rutting (d15), block cracking (d3), alligator cracking (d1), bleeding (d2), and depression (d6). about 21% of the distress types at high severity level are the alligator cracking (d1) type. longitudinal and transverse cracking (d10) type represents (18%) and polished aggregate (d12) account for about (14%) of the total high severity distress types. about (11%) of total distress are the patching and utility cut patching (d11) types. rutting (d15) type represents only (7%) of the high severity distress types. other distress types showed at high severity level were weathering and raveling (d19), bleeding (d2), and block cracking (d3). 4.2 common distress on secondary roads the percentage of distribution of distress types found on secondary road networks at the three levels of severity are presented in fig. 4. the analysis indicated that 88% of low severity distress types found on secondary roads is the d19 type. among the distress types with medium severity, d19 represents about 38% of the total distress types found on secondary roads. d11 type represents about 22% of high severity level types of distress. 5. pavement condition evaluation 5.1 evaluation of main roads condition condition evaluation for individual main roads the pci for each main road out of the 11 selected main roads is shown in fig. 5 bar graph. this figure shows that the average pci for the selected main roads ranged from 60 to 91, which reflects a rating of good to excellent. the minimum pci value for all the sample units included in the analysis was found to be 19, with a condition rating of very poor. condition evaluation for all main roads the pci for the main road network were computed by averaging the pcis for the selected roads because all the sample units were of the same size. a total of 22,700 meters of main roads representing 227 sample units was included. the average pci for the main roads was found to be 76, which indicates a general condition rating of very good. the analysis also indicated that 15% of jazan's main roads have an average condition rating of excellent. the finding was that 65%, 15%, and 5% of the networks were very good, good, and fair, respectively as indicated in fig. 6. this good rating reflects the city's newly developed network. it also shows the municipality's great challenge of preserving the network and controlling its rate of deterioration. figure 4. percentage distribution of distress density on secondary roads. 5.2 evaluation of secondary roads condition condition evaluation within each district the results of pavement condition for the secondary roads within each region are presented in fig. 7. the selected regions have been chosen from the north, south, east, west, and centre of the city. the figure shows that the average pci for the selected regions range from 25-85, with a general condition rating of poor to very good. condition evaluation for all regions the pavement condition rating of all secondary roads was computed by averaging the pcis of secondary roads within each region. a total area of 769,000 square meters of secondary roads representing 500 sample units within the selected 10 regions was considered. the average pci for the secondary road network of jazan city was found to be 66, which indicates a general rating condition of good. the percentage distribution of the condition rating of the secondary road network is presented in fig. 8, which shows that about 65% of the secondary road network is in a critical stage and, if these roads are not properly maintained, they will deteriorate with time, and significant cost will have to be incurred to make repairs. a welldesigned pavement is likely to stay in good condition for a long time before it reaches what paver refers to as the "critical condition". the length of time it takes for the pavement to reach critical condition is the pavement life. pavement life can be significantly increased by performing preventative maintenance. on the other hand, once the pavement reaches critical 52 identification of pavement distress types and pavement condition evaluation based on network level ..... figure 6. average conditions rating for jazan main road network. 53 m mubaraki condition, its rate of deterioration significantly increases and the return on investment of preventive maintenance decreases. for example, consider the question of the effectiveness of crack sealing if the pavement already has potholes. therefore, critical pci is the pci value at which the rate of pci loss increases with time. usually this value ranges from 70 to 55. additionally, fig. 8 shows that 10% of the secondary network needs to receive a thick overlay. one of the foremost reasons for using a condition indices pavement management system is to trigger treatments (deighton and sztraka 1995). during a pms analysis, a list of maintenance and rehabilitation strategies is generated. in generating this list, it is important that only feasible treatments are considered; otherwise the list will be impossibly long. to include only feasible treatments on the list, the pms needs to know when a treatment is feasible and when it is not. the process used by most pms can be described as a simple decision process, where decision trees or trig gers are used. the major inputs to this process are the condition indices. for example, with a treatment such as a thick overlay, a condition index is needed to indicate when a road is in a condition that makes applying a thick overlay feasible. feasibility can be examined from an operational and economical perspective, and it is important not to confuse them. from an operational point of view, a thin (25 mm) overlay is sometimes impossible to actually place on the pavement. consider a road with ruts, distortions, and severe roughness. if a pms included a thin overlay treatment on the list of strategies for that pavement section, it would lose credibility. therefore, from an operational perspective, the pms needs condition indices to indicate when a road is outside the operationally feasible zone of receiving a particular treatment. usually, agencies begin their pms by mimicking current practices; thus, this usually defines what is feasible from an operational point of view. in a study by mubaraki (2012), the thick overlay (50 mm) which is the structural overlay is highly recommended when the pci is less than 40. the feasibility of its application should be based on developing pavement conditions, a model, and an application of that model in setting maintenance strategies at the project level for low volume urban roads (secondary roads). in this paper, it was found that 10% of the secondary roads in the jazan city network had an as pci value below 35. therefore, the view of feasibility treatment is important here because of two perspective criteria operations and economics (limited resources). figure 8. average rating for jazan secondary roads network. 54 identification of pavement distress types and pavement condition evaluation based on network level ..... 6. conclusions based on the results obtained from this study, the following may be concluded: the types of distress that commonly exist in jazan city's main road network are patching and utility cut patching, longitudinal and transverse cracking, polished aggregate, weathering and raveling, and alligator cracking. the most common types of distress that exist on jazan city's secondary road network are weathering and raveling, patching and utility cut patching, longitudinal and transverse cracking, potholes, and alligator cracking. the average pavement condition rating of jazan's main road network is very good. the percentage of distribution of the pavement condition rating in jazan city's main road network is 5% (fair), 15% (good), 65% (very good), and 15% (excellent). about 20% of jazan's secondary road network needs immediate major maintenance such as thin and thick overlay. jazan city's main road network is newly developed and, relatively speaking, in very good condition. extensive effort should be made in order to preserve and improve the network. however, on the other hand, jazan's secondary road network is in bad condition and needs immediate action taken to improve its condition. acknowledgment the author would like to acknowledge the assistance he received from the staff of jazan municipality. their assistance and support made the production of this paper possible. references al-swailmi s (1994), framework for municipal maintenance management systems. transportation research board, transportation research board, trr 1442, 3-10, washington dc, usa. annual book of astm standards (2011), astm d5340-11, standard practice for roads and parking lots pavement condition index surveys 04:(03). baladi gy, snyder mb (1992), highway pavements, student workbook publication no. fhwa hi-90026 federal highway administration. national highway institute course no.13114. deighton r, sztraka j (1995), pavement condition. dtv technical guide (vol. 3), published by deighton associates limited. haas r, hudson wr, zaniewski j (1994), modern pavement management. krieger publishing company, malabor, florida. mubaraki m (2004), modeling the common pavement distress types on riyadh streets network. master's thesis, king saud university, saudi arabia. mubaraki m (2010), predicting deterioration for the saudi arabian urban road network. phd thesis, university of nottingham, uk. mubaraki m (2012), maintenance strategies at project level for low volume urban roads. international journal of pavement research and technology 5(4):225-233. shahin my (2005), pavement for airports, roads, and parking lots. springer, new york, usa. shahin my, walteher j (1990), pavement maintenance management for roads streets using the paver system. technical report m-90, u.s. army construction engineering lab. microsoft word paper 2.docx the journal of engineering research (tjer) vol. 13, no. 1 (2016) 22-32 effect of sediment density on the bed topography in a channel bend using numerical modeling m. vaghefi*, a, y. safarpoor and s.s. hashemi civil engineering department, persian gulf university, bushehr, iran received 25 february 2015; accepted 13 august 2015 abstract: in this paper, the bed topography of a channel at a 90-degree bend was studied at the t-shaped spur dike installed at the middle of the outer bank. numerical analyses were performed with the sediment simulation in intakes with multiblock option (ssiim) model. three relative curvatures (r/b = 2, 3, and 4) related to the three sediment densities (p = 2.35, 2.5, and 2.65) were modeled in nine cases and the effects of sediment density on the bed scour patterns were studied. it was observed that the maximum amount of scour occurred near the head of the dike wing at the dike’s upstream, and the maximum amount of sedimentation occurred at the inner bank of the bend exit. by increasing the sediment density when r/b = 4, the maximum scour and sedimentation decreased to 27.41% and 46.15%, respectively. when r/b = 3, the maximum scour and sedimentation decreased to 25.93% and 23.91%, respectively. when r/b = 2, the maximum scour and sedimentation decreased to 25.98% and 10%, respectively. keywords: ssiim model, t-shaped spur dike channel bend, sediment density, relative curvature. אאאאאא {، wאאאאאא٩٠kא   אא kאאssiim k  אא f/z٢،٣،٤ef2.35, 2.5, 2.65e٩،אאא kאאא،א אאאאאkא/z٤،א  אא٤١}٢٧٪١٥}٤٦٪אא ؛  /z٣ ، א א א٩٣}٢٥٪٩١}٢٣٪אא ؛  /z٢ ،  א א א٩٨}٢٥٪١٠٪אאk אא :ssiim،א،א،אkאאk * corresponding author’s e-mail: vaghefi@pgu.ac.ir m. vaghefi, y. safarpoor and s.s. hashemi 23 1. introduction one of the methods employed for river training is the use of spur dike structures. spur dikes divert the mainstream and protect river banks from erosion and have been studied for many years by researchers. numerical model simulations are usually more cost-effective and faster than physical model studies and have no inherent limitations on spatial extent (hua et al. 2006). to study the curved channels, it is necessary to first check the flow pattern affected by cross-flow and secondary flow. shukry (1949) presented an equation for calculating the power of secondary flow in channel bends. nouh and townsend (1979) determined that the secondary flow after leaving the bend not only is lost, but continues a straight distance downstream. gill (1972), by changing the radius of curve, depth of flow and diameter of particles in the direct and bend channels showed that the distance between spur dikes is dependent on the radius of the curve. da silva and yalin (1997) discussed the locations of the sediments and erosion holes in two mild and sharp meander bends. lian et al. (1999) found that the secondary flow in a 180degree bend is stronger than that in a 90-degree one. the curvature of the flow in river bends leads to secondary flow normal to the main flow’s direction. in a river bend, a large secondary flow cell develops, covering most of the cross-section. moreover, in river bends with steep outer banks, a much weaker second, counter rotating, secondary flow cell appears along the upper part of the outside river bank (booij 2002). one of the consequences arising from the installation of a spur dike is the creation of scour around the spur dike bed which is caused by a change in the flow pattern of the channel. elawady et al. (2001) concluded that the scouring around a submerged spur dike is affected by the submergence and opening ratios of a channel. a greater opening ratio leads to a reduction in scour and the erosion of downstream banks rarely occurs. one type of spur dike is a t-shaped spur dike. because of the cross section at the tip of the spur dike (the dike wing), it has a different impact relative to direct and normal spur dikes, and it has a unique effect on the pattern of flow and the scour. this effect occurs because the longer longitudinal section of the channel is blocked. in a t-shaped spur dike, the dike wing reduces the scouring in the bed of a dike web. the effects of the dike and froude number on the scour around a t-shaped spur dike in a 90-degree bend was examined by ghodsian and vaghefi (2009). it was found that the dimensions of the scour hole increased as a result of an increase in the length of the spur dike. additionally, it was determined that by increasing the froude number and length of the spur dike, the amount of scour increases. for bank protection purposes, the spacing-to-length ratio of the spur dikes is limited to a maximum value of six due to the characteristics of the vortex system (azinfar 2010). several numerical models have been used to calculate flow patterns and sediment but in this study, the ssiim numerical model has been employed because of its ability to analyze the flow and scour patterns in the channel with complex geometry. an experimental and numerical simulation of flow using the sediment simulation in intakes with multiblock option (ssiim) model showed that the ssiim model could accurately simulate the flow pattern in 90 degree bends (abhari et al. 2010). in meander bends, the velocity distribution and angle of attack are indicative of secondary currents associated with channel curvature (sclafani et al. 2012). vaghefi et al. (2014) studied the effect of the t-shaped spur dike submergence ratio on the water surface profile using the ssiim model. they showed that the flow velocity over the spur dike crest level will be greater with the increase of t-shaped spur dike submergence. vaghefi et al. (2015) calibrated the ssiim numerical model based on the experimental model in a channel bend; therefore, in this study the ssiim model was used. there has been no thorough investigation of the effect of changes in sediment density on a river bed’s curved channels or a comparison of them in different relative curvatures. what makes this study important is the numerical study of scour pattern in channel bed under the influence of a t-shaped dike. the types of sediments in a river bed in order of density, diameter, and fall velocity affect the calculations of sediment transport; therefore, in this study, the variations of the bed topography were evaluated at the channel bend with three different relative curvatures due to changes in sediment density. 2. materials and methods 2.1 numerical model the ssiim numerical model addresses the navier-stokes equations with the k-ɛ model on a three-dimensional (3d) general non-orthogonal grid. in ssiim, the navier-stokes equations for turbulent flow in a general 3d geometry are solved to obtain water velocity. the navier-stokes equations for noncompressible and flow with constant density can be expressed as follows: effect of sediment density on the bed topography in a channel bend using numerical modeling 24 ( )jiij ii i j i uup xx u u t u ρδ ρ −− ∂ ∂ = ∂ ∂ + ∂ ∂ 1 (1) where x1, x2, and x3 are distances and u1, u2, and u3 are the velocities in three directions. p is pressure and δij is the kronecker delta that is equal to unity for i = j, and zero otherwise. the first term on the left side of the equation is the transient term while the next term is the convective term. the first term on the right-hand side is the pressure term while the second term on the right side of equation is the reynolds stress term and, to evaluate this term, k-ɛ turbulence model is used. the semi-implicit method for pressure-linked equations (simple) was used to compute the pressure term. the effect of the density variations on the water flow field is taken into account by introducing a modified eddy viscosity. the eddy viscosity from the k-ɛ model is multiplied with a factor, taking into account the velocity and concentration gradients. sediment transport is traditionally divided into bed and suspended loads. for suspended load, van rijn (1987) developed a formula for the equilibrium sediment concentration, cbed, close to the bed (eqn. 2). in addition to the suspended load, the bed load, qb, can be calculated. in (eqn. 3), van rijn’s formula for bed load is used (2007). ( ) 1.0 5.1 3.0 2 015.0         −       − = υρ ρρ τ ττ w gwsa d c c c bed (2) ( ) 1.0 5.1 5.1 50 2 3.0 50 053.0         −     − =       − υρ ρρ τ ττ ρ ρρ w gws c c g w ws d q d b (3) the sediment particle diameter is denoted by d, a is a reference level set equal to the roughness height. τ is the bed shear stress, and τc is the critical bed shear stress for movement of sediment particles according to shield’s curve. ρw and ρs are the density of water and sediment, υ is the viscosity of the water, and g is the acceleration of gravity (olsen 1999, 2000, 2009). 2.2 initial conditions analysis was performed in clear water conditions. the diameter of the bed particles (ds50) was 1.28 mm. the standard deviation (ϭ) of the bed particles was 1.3 mm. the discharge of flow was only from upstream (q = 25 liter/second). the water depth in the channel upstream and the start of the straight run (y initial) was equal to 11.6 cm. the froude (fr) number of the flow in the straight direction of the upstream of channel was 0.34. channel walls were rigid and erosion took place only through the channel bed. the spur dike was tshaped and 9 cm long, with an equal ratio of wing to web length. the dike was positioned at a 45degree angle to the channel bend. three relative curvatures related to the three sediment densities (ρ = 2.35, 2.5, and 2.65 g/cm3) are analyzed in this study in nine cases. it took about eight hours to analyze each case for flow pattern and scour. the channel bend is shown as fig. 1, in which ɵ is the angle of each cross-section from the bend upstream. 2.3 boundary conditions for numerical model ssiim-1, version 1 uses a structured grid. in a structured grid, there will always be one more grid line than grid cells in any given direction. because boundary conditions are also needed, there will be a grid cell with zero thickness at the walls. boundary conditions are required on all boundaries for the sediment flow calculation. the two most used types of boundary conditions are zero gradient and dirichlet. zero gradient boundary conditions means the derivative of the variable at the boundary is zero. in other words, the value at the boundary is the same as the value in the cell closest to the boundary. this boundary condition is often used as the outflow boundary for the sediment concentration calculation and can also be used at walls. a zero gradient boundary condition is a type of neumann boundary condition, while a dirichlet boundary conditions means the values of a variable are given at the boundary. the gradient of all parameters in the outputs boundary are zero. in addition, the output rates of discharge must be introduced in the outgoing boundary conditions. the gradient of loss of the kinetic energy as well as the value of the kinetic energy at the water’s surface is zero. the flux passing the bed and walls is zero. the discharge should be introduced in the bend entrance. the velocity gradient towards the wall is often very steep. if it is to be resolved in the grid, this will require too many grid cells. instead, the wall laws area should be used. this means that it is assumed that the velocity profile follows a m. vaghefi, y. safarpoor and s.s. hashemi 25 (a) (b) figure 1. (a) channel bed plan view and (b) magnified spur dike details. figure 2. the boundary conditions at cross section. certain empirical function,called a wall law. as eqn. 4 shows, the wall law of schlichting (1979) is used, where u is the velocity, u* is the shear velocity, k is a constant coefficient equal to 0.4, y is the distance from the wall to the center of the cell, ks is the roughness which is equal to 90% of the particles’ diameter in the bed grading curve (olsen 1999, 2000, 2001). ( )3/30 1 kyln ku u = (4) 2.4 verification for verification of the model, ssiim parameters were calibrated using laboratory model results. the intended experiments have been carried out in tarbiat modares university’s hydraulics laboratory in iran on a laboratory flume with a width of 60 cm and height of 70 cm in a compound straight and bent route (fig. 3). the straight upstream route is 710 cm and is connected to a straight downstream route of 520 cm via a 90-degree bend, with an external radius of 270 cm and an internal radius of 210 cm. the spur dike used in this experiment is tshaped. the length of the wing and web are nine cm with a height of 25 cm. this spur dike is vertical and non-submerged in a 45-degree position. each experiment took the equivalent of 88% of the scour depth on experiment with an equilibrium time of 120 hours. this amount was 24 hours for each experiment (vaghefi et al. 2012). comparisons of the bed profiles in the cross sections at downstream and upstream of the spur dike are shown in fig. 4. the model is capable of providing the bed topography. in other words, the ssiim model could simulate the scour pattern in a 90-degree bend and is applicable in this study. at a range of 0–10 cm from the outer bank, the bed profiles do not match with each other perfectly due to the presence of the spur dike in this area and the presence of the turbulent flow (fig. 4). in the project’s execution, the spur dike bed is protected from erosion by a collar layer; thus, the difference in the position of maximum scour is not significant. the difference in scour amounts in all sections appears in table 1, where de/ds is the dimensionless effect of sediment density on the bed topography in a channel bend using numerical modeling 26 (a) (b) figure 3. the laboratory schematic model (a) before the test and (b) after the test. (e) (f) figure 4. the bed profile in the cross sections of (a) 37.750 (b) 400 (c) 41.250 (d) 420 (e) 43.750 (f) 46.250. v. mohammad, s. yaser and h.s. shaker 27 table 1. difference in scour amounts related to the numerical and laboratory models. section 1 de/ds section de/ds figure 4(a) 1.5% figure 4(d) 0.85% figure 4(b) 0.9% figure 4(e) 0.02% figure 4(c) 0.95% figure 4(f) 0.6% parameter, and de is the difference in maximum scour amounts for the two models and ds is the maximum scour depth. figure 5 illustrates that the two models are a good match in terms of the quantity and range of maximum scour so that the maximum scour occurs in the dike upstream near the tip of its wing. in comparison, results are in good agreement. figure 5 illustrates that rc is the curvature radius of the bend. x and y are the distance of the coordinates from the center of curve. to show the amount of bed changes, the dimensionless quantity (ds / y) is used which represents the ratio of the change in the bed depth to the initial water depth (yinitial = 11.6 cm). (b) (a) figure 5. bed changes around in the spur dike (a) numerical data and (b) laboratory findings. 3. results and discussion to analyze the effect of the bed sediment density on bed changes, the current researchers modeled a 90-degree bend channel with three relative curvatures, with r/b = 2, 3, and 4, where r is the radius of curvature and b is the width of channel. sediment density (ρ) with respect to the empirical values and based on the gender of the river sediments were set at 2.35, 2.5, and 2.65 g/cm3. according to figs. 6–8, in all relative curvatures (r/b), the bed topography varies by increasing the sediment density, but the maximum amount of scour always occurs near the head of dike wing at the dike upstream, and the maximum amount of sedimentation occurs at the inner bank of bend exit. the laboratory study of vaghefi et al. (2012) showed in this case that r/b = 4. as the figs. 6 and 7 illustrate, the bed sediments were moved to the channel downstream from around the spur dike. by increasing the sediment densities from 2.35 to 2.65, the maximum amounts of scour and sedimentation were reduced (table 2). in a channel bend with a relative curvature equal to 3 and 4, the second scour hole is formed at the inner bank of the bend exit; however, when r/b = 2, this hole is not formed because the bend length is less than when r/b = 3 and 4. when r/b = 2, the bend downstream is affected by the main scour hole which is formed at the upstream of the dike wing). this influence continues until the bend exit. effect of sediment density on the bed topography in a channel bend using numerical modeling 28 (b) (a) (c) figure 6. bed topography when r/b = 4 (a) ρ = 2.35 (b) ρ = 2.5 (c) ρ = 2.65. table 2. redacting the maximum scour and sedimentation due to an increase in density. relative curvature reading the scour redacting the sedimentation r/b = 2 25.98 10% r/b = 3 25.93% 23.91% r/b = 4 27.41% 46.15 x /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.49 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.14 x/rc y /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.52 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.35 x/rc y /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.28 0.2 0 -0.2 -0.4 -0.6 -0.8 -0.9 -0.98 v. mohammad, s. yaser and h.s. shaker 29 (b) (a) (c) figure 7. bed topography when r/b = 3 (a) ρ = 2.35 (b) ρ = 2.5 (c) ρ = 2.65. as figs. 6–8 illustrate, for r/b =4 with an increases in sediment density, at the dike upstream the range of sedimentation decreases. this sedimentation was created by transferring a portion of the sediment from the spur dike bed due to the longitudinal vortices. also, downstream of the spur dike and at an angle of about ɵ = 600, a sedimentary bar expands in the middle of the channel width, such as when the density equals 2.65. it has the highest quantity. vaghefi et al. (2015), by investigating the changes in bed topography in different relative curvatures concluded that when ρ = 2.35, by increasing the relative curvature, the maximum amount of scour increases in the dike upstream. in this paper, it can be seen that when ρ = 2.5 and 2.65 g/cm3, by increasing the relative curvature according to table 3, the maximum amount of scour increased as expected. as figs. 9 and 10 show, the maximum amounts of scour and sedimentation are reduced because, by increasing the sediment density, the mass of the sediment increases; therefore, more force is needed to pick up and move the sediment. since the conditions are the same for all analyses, the x / rc y /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.38 0.3 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.13 x / rc y /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.4 0.3 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.31 x / rc y /r c 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ds/y 0.3 0.2 0 -0.2 -0.4 -0.6 -0.8 -0.97 effect of sediment density on the bed topography in a channel bend using numerical modeling 30 (b) (a) (c) figure 8. the bed topography when r/b = 2 (a) ρ = 2.35 (b) ρ = 2.5 (c) ρ = 2.65. table 3. the dimensionless of maximum scour due to increase in density and relative curvature. r/b = 4 r/b = 3 r/b = 2 sediment density g/cm3 ds/y = -1.35 ds/y = -1.31 ds/y = -1.27 ρ = 2.35 (results of vaghefi et al. (2015)) ds/y = -1.14 ds/y = -1.13 ds/y = -1.10 ρ = 2.5 (results of this study) ds/y = -0.98 ds/y = -0.97 ds/y = -0.94 ρ = 2.65 (results of this study) maximum scour decreases. by reducing the amount of scour, less sediment is transported at the bend downstream and the amount of maximum sedimentation is reduced. by increasing the relative curvature of the bend, based on change in sediment density, the trend of change in the scour amount increases (fig. 10) and the greater change of the diagram slope is evident when r/b = 4. 4. conclusion the effect of sediment density (ρ = 2.35, 2.5, and 2.65 g/cm3) on the bed topography of the channel with three relative curvatures (r/b = 2, 3, and 4) illustrates nine cases of installing the t-shaped spur dike in the middle of the outer bank (three relative curvatures related to the three sediment densities). the following results were obtained: x / rc y /r c 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 ds/y 0.29 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.1 x / rc y /r c 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 ds/y 0.3 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.27 x / rc y /r c 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 ds/y 0.27 0.2 0 -0.2 -0.4 -0.6 -0.8 -0.94 v. mohammad, s. yaser and h.s. shaker 31 figure 9. diagram of the maximum scour based on relative curvature and density. figure 10. diagram of the maximum sedimentat ion based on relative curvature and density. 1. in all cases, the maximum amount of scour occurred near the head of the dike wing at the dike upstream, and the maximum amount of sedimentation occurred at inner bank of the bend exit. 2. in a channel bend with a relative curvature equal to 3 and 4, the second scour hole was formed at the inner bank of the bend exit. 3. when r/b = 4 by increasing the sediment density, at the dike upstream, the range of the amount of the sedimentation decreased. 4. by increasing the sediment density from 2.35 to 2.65 (gram/cm3) when r/b = 4, the maximum amounts of scour and sedimentation decreased to 27.41% and 46.15%, respectively. when r/b = 3, the maximum amounts of scour and sedimentation decreased to 25.93% and 23.91%, respectively. when r/b = 2, the maximum amounts of scour and sedimentation decreased to 25.98% and 10%, respectively. references abhari m, ghodsian m, vaghefi m, panahpur n (2010), experimental and numerical simulation of flow in a 90 degree of bend. flow measurement and instrumentation 21: 292–298 azinfar h (2010), flow resistance and associated backwater effect due to spur dikes in open channels. a thesis submitted to for the degree of doctor of philosophy, department of civil and geological engineering, university of saskatchewan saskatoon, canada. booij r (2002) modeling of the secondary flow structure in river bends. in, river flow, bousmar and zech (eds), isbn 90 5809 509 6: 127:133. da silva amf, yalin ms (1997), laboratory measurements in sine-generated meandering channels. international journal of sediment research 12(1): 16–28. elawady e, michiue m, hinokidani o (2001), movable bed scour around submerged spurdikes. annual journal of hydraulic engineering 45: 373–378. ghodsian m, vaghefi m (2009), experimental study on scour and flow field in a scour hole around a t-shape spur dike in a 90° bend. international journal of sediment research 24: 145– 158. gill ma (1972) erosion of sand beds around spur dikes. journal of the hydraulics division 98(9): 91– 98. hua l, roger a, kuhnle b, barkdoll d (2006), countermeasures against scour at abutments. research report no.49, channel and watershed processes research unit, national sedimentation laboratory, usda agricultural research service oxford. lian hc, hsied ty, yang jc (1999), bend-flow simulation using 2d depth–averaged mode. journal of hydraulic engineering 125(10): 1097–1108. nouh ma, townsend rd (1979), shear-stress distribution in stable channel bends. journal of the hydraulics division, american society of civil engineers 105 (hy10) proceedings paper 14898, 1233–1245. olsen nrb (1999), computational fluid dynamics in hydraulic and sedimentation engineering. department of hydraulic and environmental engineering, norwegian university of science and technology, class notes, revision 2nd. olsen nrb (2000), cfd algorithms for hydraulic engineering. class notes, department of hydraulic and environmental engineering, norwegian university of science and technology, norway. -1.4 -1.3 -1.2 -1.1 -1 -0.9 2.35 2.45 2.55 2.65 m ax im um s co ur /i ni ti al d ep th ρ (g/cm3) r/b=2 r/b=3 0.25 0.3 0.35 0.4 0.45 0.5 0.55 2.35 2.45 2.55 2.65 m ax im um s ed im en t/ in it ia l de pt h ρ (g/cm3) r/b=2 r/b=3 r/b=4 effect of sediment density on the bed topography in a channel bend using numerical modeling 32 olsen nrb (2001) cfd modeling for hydraulic structures. norwegian university of science and technology, preliminary 1st edition, 8 may. isbn 82-7598-048-8. olsen nrb (2009), a three-dimensional numerical model for simulation of sediment movement in water intakes with multi-block option. department of hydraulic and environmental engineering, norwegian university of science and technology, ssiim user's manual. sclafani p, thornton ci, cox al, abt sr (2012), methodology for predicting maximum velocity and shear stress in a sinuous channel with bend way weirs using 1-d hec-ras modeling results. colorado state university, engineering research center shukry a (1949), flow around bends in an open flume. american society of civil engineers, no. 2411, 115(1): 785-789. schlichting h (1979), boundary-layer theory. new york: mcgraw-hill. vaghefi m, ghodsian m, salehi neyshabouri saa (2012), experimental study on scour around a tshaped spur dike in a channel bend. journal of hydraulic engineering 138(5): 471–474. vaghefi m, safarpoor y, hashemi ss (2014), effect of t-shape spur dike submergence ratio on the water surface profile in 90 degree channel bends with ssiim numerical model. international journal of advanced engineering applications 7(4): 1–6. vaghefi m, safarpoor y, hashemi ss (2015), effects of relative curvature on the scour pattern in a 90° bend with a t-shaped spur dike using a numerical method. international journal of river basin management 13(4): 501-514. van rijn l (2007), unified view of sediment transport by currents and waves. ii: suspended transport. journal of hydraulic engineering 133(6): 668–689. c:\users\raoh\desktop\paper 6.xps the journal of engineering research (tjer) vol. 13, no. 1 (2016) 72-79 stochastic allocation of transmit power for realistic wireless channel models n. g. tarhuni*, a, m. elmusratib and m. al-nadabia *a electrical and computer engineering department, college of engineering, sultan qaboos university, sultanate of oman. b electrical engineering department, university of benghazi, libya. received 8 march 2015; accepted 29 october 2015 abstract: control of transmitted power is crucial for the successful operation of multi-user wireless channels communications. there are practical situations in which the transmitted power cannot be adjusted by feedback information; hence, only forward transmit power allocation can be applied, especially in situations where a feedback channel is not available in a wireless network or when wireless nodes are only transmit types. conventionally, transmitted power can be fixed. higher gain may be observed if the sensors’ transmitted power is randomized. in this work, random power allocation for a nakagami-m distributed wireless channel model was investigated, and a number of random distributions were evaluated theoretically and tested by simulations. the outage probability was evaluated theoretically and validated by monte carlo simulations. keywords: random power control, nakagami channel, weibull distribution. * corresponding author’s e-mail: tarhuni@squ.edu.om n. g. tarhuni, m. elmusrati and m. al-nadabi 73 1. introduction wireless sensor networks are ideally suited for applications ranging from scientific data gathering, and environmental hazard and battlefield monitoring because of their rapid and inexpensive deployment. many sensor network applications have been investigated in the literature. for example, one can find applications to target tracking, environmental monitoring, and the detection and monitoring of car theft (brooks, ramanathan, sayeed 2003; bonnet, gehrke, seshadri 2000; pottie and kaiser 2000). however, sensor networks have varying characteristics and limitations, and sensors are usually deployed in the region of interest without intensive preplanning. one positive characteristic is that the required quality of service (qos) of sensors is usually much more flexible than mobile networks. the sensors send their data to access points in single or multihop wireless channels. the required data rate depends on the application, but generally it is less than the cellular mobile requirements. further, for intensive sensor deployment, sensed data can be highly redundant which can be exploited in post processing. the maximum allowable delay depends also on the application. for example, if the sensors are part of a closed loop control system, then the maximum delay is critical for the system stability and robustness. sensors are energy-constrained devices as it is not practical to periodically recharge or replace the sensors’ batteries, especially for large sensor networks where hundreds or even thousands of sensors may be deployed in an area of interest. when in receive mode, practical sensors can consume power levels comparable to the consumption rate experienced during transmit mode. therefore, deaf sensors can prolong network life significantly. moreover, the sensor nodes should be efficient in terms of transmitted power and data coding in order to reduce any unnecessary extra power transmission and improve the overall system performance. because of the characteristics of sensor networks, sensors should be designed to be cheap, simple, and highly energy efficient. to keep the total cost and complexity of sensor networks at reasonable levels, it is preferable to design the sensors as broadcasters. in any multiuser wireless network with non-perfectly orthogonal multiple access (such as code division multiple access [cdma], time division multiple access [tdma], or orthogonal frequency-division multiple access [ofdma] with frequency reuse, etc.), power control is essential to limit interference, improve network capacity, enhance performance, improve energy efficiency, and prolong the battery life of devices. prolonging the lifetime of batteries is a critical objective in wireless sensors that have batteries that cannot be recharged or replaced, which is the case of many applications with outdoor sensor network deployment. the optimum transmit power value should be the minimum power needed to achieve a transmit-receive link objective (such as signal-to-noise ratio [snr], throughput, bit error rate [ber], etc.). hence, the transmitter should have enough information about the channel. this information about the channel is called channel state information (csi), which is measured, or estimated, at the receiver side with the assistance, for example, of a pilot signal. nevertheless, the receiver should send full or partial information to the transmitter using a feedback channel. if the coherence time of the channel is longer than the feedback delay, then this csi could be used to optimize the transmit power or other resources such as the modulation level. this procedure is known as closed-loop transmitter power control. these topics have been widely covered in research on wireless networks for more than a decade (elmusrati et al. 2007; fan et al. 2015; zander 1992). however, there are some extreme cases where optimizing the transmit power with closed loop power control is not feasible. two examples follow: first, when the feedback delay is greater than the coherence time of the channel, the csi becomes outdated and even uncorrelated with the channel at the time of transmission. therefore, using csi is useless in this case as the feedback delay can be larger than the coherence time of the channel. when the channel is changing with high dynamics such as in very fast mobiles or when the duty cycle of the feedback channel is very large, such as in some energy-limited sensor devices, using csi is not tenable. the second case is observed when there is a practical obstacle for the feedback channel such as when receivers do not have transmission capabilities, the transmitters do not have reception capabilities, or there is no bandwidth available for the feedback channel. in such situations, algorithms for optimizing the transmit power according to the channel variations cannot be used. this extreme situation of lack of reception capability can be found, for example, in some primitive sensor terminals where the sensors sense some physical quantity and transmit its value stochastic allocation of transmit power for realistic wireless channel models 74 without any knowledge about the csi at the receiver side. in such extreme situations, the conventional and practical power allocation is to use fixed (can be different) power transmission for all sensors. however, there are several problems associated with this fixed power allocation such as high unnecessarily energy consumption and high interference profile. to handle such problems, a randomized power allocation strategy was suggested by kim and kim (2005) whereas uniformly randomly distributed transmitted power levels were used to mitigate the near-far effect in congested systems without any channel feedback. it was based on the so-called “snapshot analysis approach”, neglecting the effects of channel fading. an evaluation of the performance of random power for rayleigh fading channel was discussed by elmusrati et al. (2009, 2012). recently, the randomization of power and modulation resources for device-to-device (d2d) scenarios were investigated by zhang and haenggi (2012); lee, lin et al. (2015); and erturk et al. (2013). in this paper, the performance of random power allocation strategies for wireless sensor networks were investigated taking into account realistic channel fading. the assumed channel model is the nakagami-m model, which includes rayleigh and rician fading channels, and the unfaded channels as special cases. in section ii, the system model is described. in section iii, the analysis of the fixed transmit power allocation is given. in section iv, random transmit power allocation is considered. finally, conclusions are drawn in section v. 2. system model in the considered scenario, the sensors are scattered randomly and uniformly over the area of interest and send their packets over a wireless link to an access point. as shown in fig. 1, the measured quantities (eg. temperature, pressure, speed, etc.) are detected by sensors and then multiplexed. error correction coding is added to improve performance, and then the signal is power amplified to be injected to the antenna for transmission. the amplification factor of the power amplifier stage is controlled by a power allocation scheme which can choose to fix or randomize the power according to a selected probability distribution. every packet contains coded transmitted information, pilot bits, and time and sequence stamps. the transmitted packets arrive from the n sensors to the access point in a multipath manner with or without a dominant path. in this work, a nakagami-m fading channel with mean channel gain is assumed and is presumed to be inversely proportional to the sensor distance as 1/ . the lost packets cannot be retransmitted because there is no feedback channel. however, the likelihood of packet loss can be compensated by increasing the level of redundancy. assuming the envelope of the received signal follows the nakagami distribution, then the link gain connecting transmitter j and receiver i will be gamma distributed as, , 0 (1) where denotes the mean channel gain. if =1, the distribution reduces to exponential distribution corresponding to rayleigh fading. as increases, the channel behaves like rician fading. as →∞ the channel becomes additive white gaussian noise (awgn). the characteristic function of the gamma-distributed link gain is well known and given by papoulis (2002) as ψ 1 (2) for 1/2. if is a positive integer, then the cumulative distribution function for the link gain is (papoulis 2002). 1 ∑ (3) for simplicity, consider a case in which all links have the same channel statistics of for all (i, j). however, it is straightforward to generalize the analysis to the case where also depends on the link (i, j). therefore, a normalized probability density function (pdf) , cumulative density function (cdf) and characteristic function ψ that corresponds to a channel with a unit mean that can be defined. the sinr γ for each sensor node can be written as γ ∑ (4) where is the i-th sensor transmit power and denotes the noise power. 3. fixed transmit power allocation when there is no feedback channel available and no channel information, the conventional way is to n. g. tarhuni, m. elmusrati and m. al-nadabi 75 figure 1. sensor network system layout. set the transmit power of all sensors to a fixed value. in case of single transmitter and fixed transmission power , the outage probability pout can be written as, pr (5) where is the minimum detectable signal-tointerference-plus-noise ratio (sinr). if m is a positive integer, the outage in case of multiuser interference can be derived as where ∑ denotes the interference power at receiver i. let a random variable be defined as . the characteristic function of for fixed transmission power can be directly derived from the characteristic equation of the channel; hence, ψ ∏ ψ (6) the outage probability pr can be evaluated as, f βm df β1 ∑ (7) where denotes the cdf of the random variable . note that by definition, ψ (8) and ψ ψ (9) thus, the outage can be written in the form 1 ∑ ψ 1 (10) stochastic allocation of transmit power for realistic wireless channel models 76 figure 2. outage versus network size. figure 3. outage versus sensor location. the result in eqn. 10 is validated numerically as follows. assume 10 sensors are distributed randomly such that 10 50, where di is the distance of sensor i to the access point. let the sensors be arranged in increasing distance from the access node with d1 spaced at 10 meters and dn at 50 meters. the first simulation is carried out by assuming that the sensors at d1 and d2 are active and the others are set to off. then the outage of the sensor at d2 is evaluated theoretically and numerically. next the sensor at d3 is switched on while keeping the first ones set to on, and its outage is calculated, and so on. this means that the outage of the farthest sensor is evaluated for an increased network size. the transmitted power of all sensors is fixed at 0 dbm, the sinr at 5 db and the noise at -70 dbm. figure 2 shows that the theoretically evaluated outage is consistent with the numerically evaluated outage for m = 1 and 4. figure 3 shows the outage of each sensor of a network with 10 sensors for different values of the nakagami-m parameter m. it can be clearly seen that closer sensors will benefit from stronger paths. on the other hand, the far sensors outage will increase due to the increased interference from closer sensors. n. g. tarhuni, m. elmusrati and m. al-nadabi 77 4. random transit power allocation let denote the power received at receiver i from transmitter j. in the current study, it was assumed that the power values were drawn from a distribution having finite moments. therefore, the characteristic function of the received power can be expressed as: ψ 1 (11) let us define the normalized characteristic function ψ corresponding to the case 1. it follows that the characteristic function for becomes ψ ∏ ψ (12) using the same argument as was used in the last section, the outage probability can be written as for a general random power distribution as 1 ∑ ψ (13) evaluating the outage for a specific tx power distribution is tedious. it involves the evaluation of the integral m 1 derivatives of the characteristic function of which is composed of the product of n 1 characteristic functions of the received signal power. therefore, a numerical evaluation of the outage for some selected tx power distributions was employed and the average outage for the fixed, uniform and weibull distributed transmit power allocations were compared numerically. the weibull pdf was selected for the following reasons. first, it has large shaping flexibility as shown in the next section, meaning that wide statistical behavior of the transmit power can be obtained. second, it has a tractable mathematical form. it should be stressed at this point that optimizing the power allocation distribution is beyond the scope of this work, and other distributions may produce similar or better behavior compared to the weibull allocations. 5. weibull distributed transmit power the pdf of the weibull distribution of the random variable x is given by , 00, (14) where, a > 0 is the scale parameter, and b > 0 is the shape parameter. when applied as the pdf of the transmit power, a truncated weibull distribution is used. also for the sake of easier control of the transmit power distribution, the scale parameter was fixed to a = 1 and let the shape parameter b be a free control variable. then the truncated weibull pdf is given by , 00, (15) figure 4 shows this pdf for several values of the shape parameter with set to 1 dbm. as can be seen by changing the shape parameter, different shapes of the truncated weibull power distribution can be obtained. for b <1 the transmitted power will be more biased toward zero. on the other hand, for b >1, the transmitted power will be closer to . because <1, b = 1 results in an approximately uniformly distributed power. therefore, if b is tuned separately for the different sensors according to their location, it is expected that the outage of the sensors will be equalized such that the far and close sensors to the access node will perform comparably in terms of outage probability. the average power is given by means of a distribution that can be evaluated by 1 , (16) where , exp is the incomplete gamma function. 6. simulation results in this section, the investigated technique of random power allocation was tested via several simulation runs. in the simulation setup, the assumption was a network with 10 sensor nodes distributed randomly but uniformly over an area such that the closest sensor was at a distance of 10 meters and the farthest sensor was at 50 meters. the target sinr of the nodes was set to 5 db, the noise power was set to -70 dbm, and the maximum transmitted power was 0 dbm. the channel was assumed to be a nakagami-m with m = 1 (rayleigh fading channel). the channel realizations were generated by first producing average channel gains following the fourth order path loss factor, such that 1/ in eqn. 1, where dij is the distance to the i-th sensor, and gij is generated using the nakagami-m pdf. stochastic allocation of transmit power for realistic wireless channel models 78 figure 4. truncated weibull pdf for several values of the shape parameter b. next, the transmitted power by each sensor was generated according to the selected distribution and the received signal power was evaluated. hence, the probability of outage for each sensor was evaluated by checking if γ in eqn. 4 was less than the target γ . the simulation was conducted such that 20 network deployments were performed, and for each deployment 105 channels and power realizations were generated. the final outputs were found by averaging the results of the 20 network deployments. five scenarios were simulated to test the following transmit power allocations: scenario i: all sensors transmit at their maximum fixed power of pmax scenario ii: all sensors use a randomly and uniformly distributed power (~0 to pmax) scenario iii: all sensors transmit a truncated weibull random power with fixed b = 0.2 scenario iv: all sensors transmit a truncated weibull random power with fixed b = 4 scenario v: sensors transmit following a truncated weibull random power with b assigned linearly to sensors such that the closest sensor has b = 0.2 and the farthest sensor has b = 4 in all weibull random power scenarios, the scale parameter was fixed at a to one. as depicted in fig. 5, scenario iii showed the worst performance due to low b which tends to produce a low transmit power that is insufficient to combat noise. for most of the nodes, scenarios i, ii and iv resulted in comparable performance with scenario i, showing slightly better results. however, scenario i consumed the highest average power of pmax per node. the total average power of scenarios i–iv was 10 dbm, 7 dbm, 1.8 dbm, and 9 dbm, respectively. in scenario v, it was assumed that the shape parameter b can be controlled before the sensor is deployed to its location such that its value is tuned to lower values for close sensors and higher values for distant sensors. for simplicity and without any effort to optimize its value, b was set to depend linearly on the distance by setting the close sensor to b = 0:2 and the farthest sensor to b = 4. the b for sensors located in between were interpolated linearly from these end points. in this case, the power of close sensor would be biased towards zero and therefore have higher outage compared to fixed and uniform power allocations. on the other hand, far sensors would benefit from regulated power allocation and enhanced performance can be clearly observed in fig. 5. the network size of 10 sensors was used for the sake of demonstration. it is expected that as the network size increases, the performance will deteriorate. the factors affecting the performance are the target sinr, the mean channel gain, the transmit power, and the noise. n. g. tarhuni, m. elmusrati and m. al-nadabi 79 figure 5. outage versus transmit power distribution. 7. conclusion in this paper, the performance of a random power allocation scheme in a nakagami-m wireless channel was analyzed. the nakagami-m model can emulate several scenarios, including the line of sight, the non-line of sight, and the awgn channel as special cases. fixed power allocation was compared to uniform random power and truncated weibull random power allocation schemes. it was found that the fixed power allocation outperforms the other methods with the disadvantage of increased average transmitted power. on the other hand, the weibull distribution outperformed the other methods if the control parameter of the distribution was tuned such that far sensors had higher values and closer sensors had lower values. references bonnet p, gehrke j, seshadri p (2000), querying the physical world. personal communications ieee 7(5): 10-15. brooks rr, ramanathan p, sayeed am (2003), distributed target classification and tracking in sensor networks. proceedings of the ieee 91(8): 1163-1171. elmusrati m, jantti r, koivo hn (2007), multiobjective distributed power control algorithm for cdma wireless communication systems. ieee transactions on vehicular technology 56(2): 779– 788. elmusrati m, tarhuni n, jantti r (2009), performance analysis of random uniform power allocation for wireless networks in rayleigh fading channels. european transactions on telecommunications 20(4): 457–462. elmusrati m, tarhuni n, jantti r (2012), framework of random power allocation for wireless sensor networks. journal of wireless sensor networks 4(3): 76–83. erturk mc, mukherjee s, ishii h, arslan h (2013), distributions of transmit power and sinr in device-to-device networks. ieee communications letters 27(2): 273–276. fan z, bai s, wang s, he t (2015), delay-bounded transmission power control for low-duty-cycle sensor networks. ieee transaction on communications 14: 3157–3170. kim ts, kim sl (2005), random power control in wireless ad hoc networks. ieee communications letters 9(12): 1046–1048. lee ny, lin xq, andrews jg, heath rw (2015), power control for d2d underlaid cellular networks: modeling, algorithms, and analysis. ieee journal on selected areas in communications 33(1): 1–13. papoulis p (2002), probability, random variables, and stochastic processes. 4th ed.” boston: mcgraw-hill higher education. pottie gj, kaiser wj (2000), wireless integrated network sensors. communications of the acm 43(5): 51–58. zander j (1992), performance of optimum transmitter power control in cellular radio systems. ieee transactions on vehicular technology 41: 57–62. zhang x, haenggi m (2012), random power control in poisson networks. ieee transactions on communication 60(9): 2602–2611. microsoft word paper 8.docx the journal of engineering research (tjer) vol. 13, no. 1 (2016) 88-101   optimization of process design problems using differential evolution algorithm a.m. gujarathia*, g. vakili-nezhaada and m. vatanib a* department of petroleum and chemical engineering, college of engineering, sultan qaboos university, p.o. box. 33, muscat 123, sultanate of oman. b department of chemical engineering, amirkabir university of technology (tehran polytechnic), hafez ave., p.o. box 158754413,tehran, iran. received 15 september 2015; accepted 29 november 2015 abstract: a modified differential evolution algorithm (mde) has been used for solving different process related design problems (namely calculation of the nrtl and two-suffix margules activity coefficient models parameters in 20 ternary extraction systems including different ionic liquids and reactor network design problem). the obtained results, in terms of root mean square deviations (rmsd) for these models are satisfactory, with the overall values of 0.0023 and 0.0170 for 169 tie-lines for nrtl and two-suffix margules models, respectively. the results showed that the mde algorithm results in better solutions compared to the previous work based on genetic algorithm (ga) for correlating liquid-liquid equilibrium (lle) data in these systems. mde also outperformed de algorithm when tested on reactor network design problem with respect to convergence and speed. keywords: nrtl, two-suffix margules, parameter estimation, ionic liquid, optimization. אאאאאאא אkkk، j،k   wא אאאאfאmdeeאאאאf nrtlאאtwo-suffix margules٢٠א אאאאkeאאא،אא אאfrmsdeא،٠٠٢٣}٠٠١٧٠}٠١٦٩א nrtlאmarguleskאא،אאאא אאאאאfgaeאאא א jאflleeאkאאא(mde)אא (de)אkאא אאw אאאfmde ،enrtl،margules،אא،א jא، אאא،،אאk * corresponding author’s e-mail: ashishg@squ.edu.om a.m. gujarathi, g. vakili-nezhaad and m. vatani  89 nomenclature list of symbols a interaction parameter cr crossover constant f scaling factor fobj objective function g energy parameter gmax maximum number of generations k distribution ratio l molar split ratio m number of tie lines np population size r universal gas constant t absolute temperature x liquid-phase mole fraction z overall, or make-up molar fraction greek letters  activity coefficient  energy parameter α non-randomness parameter superscript and subscripts cal calculated value exp experimental value i, ii two liquid phases at equilibrium i, j, k component, phase and tie line respectively 1, 2, 3 aromatic, aliphatic and ionic liquid respectively optimization of process design problems using differential evolution algorithm  90 1. introduction phase equilibria play an important role in the synthesis, development, design, and control of chemical and petroleum engineering processes. phase behavior is often analyzed using equation of state and activity coefficient models. though lots of data is available in literature which correlate both the model and experimental data, estimation of accurate model parameters still remain a challenge. local optimization methods (such as newtons method or least square techniques, etc.) are often used to solve phase equilibrium problems. but because the search space is highly nonlinear consisting of local and global minima’s, the local methods may converge to a local optimum point (stragevitch and davila 1997; sahoo et al. 2006). in recent past the stochastic optimization techniques (such as differential evolution (de), genetic algorithms (ga), simulated annealing (sa), particle swarm algorithm, etc.) (holland 1992; kirkpatrick et al. 1983; price et al. 2005) have shown considerable potential in solving complex engineering problems (such as phase equilibrium problems, reaction engineering problems, process calculations related optimization problems etc.). liquid-liquid-equilibrium (lle) has remained a very challenging problem for non-ideal liquid solutions such as ionic liquids (ils). ils are green solvent and they have shown potential applications in many separation processes (lei et al. 2009). for designing any separation processes using ionic liquids, the accurate estimation of activity coefficient parameters is of vital importance. as a quick review on these problems we may refer to the following works: ga has been utilized for estimation of the parameters for two well-known activity coefficient models (ie. nrtl and uniquac) (singh et al. 2005). they proved that their results were better than local search methods. in another study (sahoo et al. 2006; sahoo et al. 2007), ga was used to estimate ternary, quaternary and quinary lle interaction parameters for nrtl and uniquac models. it was reported that ga results were better than other techniques. another group of authors (rashtchian et al. 2007) also used a stochastic method (genetic algorithm) for phase behavior of multiphase and multicomponent systems. they calculated the binary interaction parameters of uniquac, nrtl and wilson models for a number of systems. they also reported their data and compared their data for various systems (including vapor liquid equilibrium, vapor-liquidliquid equilibrium, and liquid-liquid equilibrium) form the literature. in some other studies, ga was successfully applied for estimation of interaction parameters of other complex models for systems of lle containing ionic liquids and also for modeling of viscosity of crude oil binary blends (vatani et al. 2012; vakili-nezhaad et al. 2013; vakili-nezhaad et al. 2014; al-maamari et al. 2015). modified de algorithm is used for calculation of two complex problems. the performance of mde algorithm is also explored by comparing its output with other algorithms. parameters of two activation coefficient models (which include nrtl and twosuffix-margules model) are calculated using experimental data having 20 sets of ternary liquid systems having 20 different ils. the reactor network problem is also solved and both the speed and accuracy aspects of mde algorithm are reported. the statistical parameters of the models including root mean square deviation for the above mentioned set of ternary systems is reported. comparison of the results obtained using mde with our previous work based on ga (vatani et al. 2012) shows that mde is a very reliable algorithm in parameter estimation problems which can give more accurate results compared to ga. abbreviation, full names and list of ternary ionic liquid systems used in this study is given in table 1. 2. problem formulation 2.1. the lle modeling the lle modeling is based on thermodynamic equilibrium condition for mole fractions and activity coefficient between the aliphatic rich phase (i) and the il rich phase (ii). ( ) ( )i i i ix x    (1) the molar component balances lead to the following equation: (1 )i i iz x l x l     (2) where x,  , z and l are mole fraction, activity coefficient, the overall mole fraction and the molar split ratio respectively, and the subscript i denotes the components. considering the distribution ratio of component as below: i i i x k x    (3) and solving eq. (3) with eq. (2), the rachford–rice equation (seader and henley 2006) is concluded as:     table 1. abbreviation, full name and list of ternary ionic liquid systems used in this work. abbreviation full name of ionic liquid sys. no. name of ternary systems t (k) tieline ref. [mebupy][bf4] 4-methyl-n-butylpyridinium tetrafluoroborate 1 propylbenzene(1)+hexadecane(2)+[mebupy][bf4](3) 313.00 10 alkhaldi et al. 2011 [3-mebupy][dca] 3-methyl-n-butyl pyridiniumdicyanamide 2 benzene(1)+hexane(2)+[3-mebupy][dca](3) 303.15 11 hansmeier et al. 2010a [emim][eso4] 1-ethyl-3-methylimidazolium ethylsulfate 3 benzene(1)+hexane(2)+[emim][eso4](3) 313.20 08 garcía et al. 2009 [bmim][mso4] 1-butyl-3-methylimidazolium methylsulfate 4 benzene(1)+hexane(2)+[bmim][mso4](3) 328.20 08 garcía 2010a [bmim][bf4] 1-butyl-3-methylimidazolium tetrafluoroborate 5 benzene(1)+heptane(2)+[bmim][bf4](3) 298.15 08 revelli et al. 2010 [bmim][ntf2] 1-butyl-3-methylimidazolium bis{trifluoromethylsulfonyl}imide 6 benzene(1)+octane(2)+[bmim][ntf2](3) 298.15 13 domínguez et al. 2011 [bmim][pf6] 1-butyl-3-methylimidazolium hexafluorophosphate 7 benzene(1)+undecane(2)+[bmim][pf6](3) 298.15 05 maduro and aznar 2008 [omim][cl] 1-octyl-3-methylimidazolium chloride 8 benzene(1)+dodecane(2)+ [omim][cl](3) 298.20 04 letcher and deenadayalu 2003 [hmim][bf4] 1-hexyl-3-methylimidazolium tetrafluoroborate 9 benzene(1)+hexadecane(2)+[hmim][bf4](3) 298.20 09 letcher and reddy 2005 [pmim][pf6] 1-propyl-3-methylimidazolium h exafluorophosphate 10 benzene(1)+cyclohexane(2)+[pmim][pf6](3) 298.15 08 zhou et al. 2012 ammoeng 102 tetraalkyl ammonium sulfate 11 toluene(1)+heptane(2)+ammoeng102(3) 298.15 08 pereiro and rodriguez 2009 [bmim][scn] 1-butyl-3-methylimidazolium thiocyanate 12 toluene(1)+heptane(2)+[bmim][scn](3) 303.15 07 hansmeier et al. 2010b [bpy][bf4] n-butylpyridiniumtetrafluoroborate 13 toluene(1)+heptane(2)+[bpy][bf4](3) 313.20 08 garcía et al. 2010b [etme][imi3] 1-ethyl-3-methylimidazolium triiodide 14 toluene(1)+heptane(2)+[etme][imi3](3) 318.15 14 selvan et al. 2000 [bmim][dca] 1-butyl-3-methylimidazolium dicyanamide 15 toluene(1)+heptane(2)+[bmim][dca](3) 328.15 07 hansmeier et al. 2010b [hmim][tcb] 1-hexyl-3-methylimidazolium tetracyanoborate 16 toluene(1)+methylcyclohexane(2)+[hmim][tcb](3) 293.15 09 gutierrez et al. 2011 [bmim][tcb] 1-butyl-3-methylimidazolium tetracyanoborate 17 toluene(1)+methylcyclohexane(2)+[bmim][tcb](3) 313.15 09 gutierrez et al. 2011 [omim][pf6] 1-octyl-3-methylimidazolium hexafluorophosphate 18 m-xylene(1)+nonane(2)+[omim][pf6](3) 298.15 04 maduro and aznar 2010 c2ntf2 ethyl(2hydroxyethyl)dimethylammoniumbis{(trifluomethyl)sulfonyl}imide 19 m-xylene(1)+hexane(2)+[c2ntf2](3) 298.15 10 dománska et al. 2007 [dmim][mp] 1,3-dimethylimidazolium methylphosphonate 20 thiophene(1)+heptane(2)+[dmim][mp](3) 298.15 09 revelli et al. 2010 overall 169 a .m . g u jarathi, g . v akili-n ezhaad an d m . v atan i 91 optimization of process design problems using differential evolution algorithm  92   0 )1(1 1     i ii i kl kz (4) considering eq. (1), eq. (3) can be rewritten as /i i ik     , where i is proper activity coefficient models with an appropriate adjustable parameters. 2.2. the nrtl and two-suffix margules models in the present work, the experimental lle data were correlated by nrtl and two-suffix margules models (renon and prausnitz 1968; prausnitz et al. 1999). the activity coefficient i of nrtl model is expressed as: 1 1 1 1 1 1 ln m m ji ji j m k kj kj j j ij k i ijm m m j ki k kj k kj k k k k g x x gx g g x g x g x                         (5) where the binary interaction aij, and nonrandomness parameters αij, have been defined as: ij ij ij g a rt t     , exp( )ij ij ijg    (6) so in general aii= ajj=0, aij ≠ aji and αij = αji. beside nrtl model, the activity coefficients of two-suffix margules model are as below: 2 2 1 12 2 13 3 12 13 23 2 3( )ln a x a x a a a x x      (7) 2 2 2 12 1 23 3 12 23 13 1 3( )ln a x a x a a a x x      (8) 2 2 3 13 1 23 2 13 23 12 1 2( )ln a x a x a a a x x      (9) where, a12, a13 and a23 are the interaction parameters, which were regressed by ternary systems data used in this work. 2.3. objective function and decision variables the single objective optimization study is considered. the experimental data for 20 ternary extraction systems containing different ionic liquids are used. nrtl and two-suffix margules activity coefficient models parameters are estimated using the modified differential evolution algorithm. in general, parameter estimation is considered as a minimization of an objective function value (fobj) (revelli et al. 2010), which minimizes the deviation between the experimental and calculated mole fractions of the components. this part of study deals with the estimation of thermodynamic activity coefficient models parameters for a ternary set of systems, accordingly the problem is formulated as:          i j k cal kjikjiobj xxfminimize ,, exp ,, (10) lower and uppers bounds of parameters used in this work are given in table 2. the criterion for comparison between the experimental and calculated data is given by the following equation. 1 2 exp 2( ) / 6calijk ijk i j k rmsd x x m         (11) where, x and m are mole fraction and the number of tie lines and subscripts i, j and k are component, phase, and tie lines respectively. in this study reactor network design (rnd) problem is considered for optimization study. rnd problem is optimized using variants of differential evolution algorithms namely, de and mde. 2.4. reactor network design problem: this rnd design problem is obtained from ryoo and sahinidis (ryoo and sahinidis 1995) (see fig. 1). two cstr reactors in sequence having consecutive reactions (p→q→r) are considered. the intermediate product concentration leaving the second reactor is maximized in this problem. both the global and local solutions for this problem are reported in literature (angira 2006; babu and angira 2006). the problem is reformulated by removing the equality constraint (eqs. 12-15): maximize. )1(*)1( *)1(*)1( )1(* )1(* 624513 622511 6251 5362 xkxk xkxk xkx kxkxk f      (12) sub. to: 45.06 5.0 5  xx (13) a.m. gujarathi, g. vakili-nezhaad and m. vatani    93 table 2. interaction parameters values used in present study for the systems. component i-j nrtl margules aij aji αij aij 1–2 [1, 2500] [-500, 2500] [0.15, 0.5] [-10, 50] 1–3 [1, 2500] [-500, 2500] [0.15, 0.5] [-10, 50] 2–3 [1, 2500] [1, 2500] [0.15, 0.5] [-10, 50]   figure 1. reactor network design problem. 10-5 ≤ x5 ≤ 16 (14) 10-5 ≤ x6 ≤ 16 (15) where, k11 =0.09755988, k22 =0.99*k1, k13 =0.0391908, k24 =0.9*k3 in this study, de and mde algorithms are used to solve the reformulated problem. 3. modified differential evolution (mde) the survival of the fittest principle is used in both de and mde. against de, mde maintains only one array of variables (bounded by lower and upper bounds). this population gets improved if a better candidate solution is obtained by allowing new solution to take part in the cross over and mutations operations in the same generation. thus it is observed that number of function evaluations are lowered thus improving the convergence of algorithm. it has been found in the previous studies that an updating the single array continuously enhances the convergence speed leading to less function evaluations as compared to de (price et al. 2005). mde also offers advantages of one array consuming less memory and an improvement on cpu time. premature convergence can be avoided by wisely/appropriately choosing the key parameters of mde (such as np, cr and f) (fan and lampinen 2003; angira and babu 2006). both de and mde algorithms are widely used in the field of non-linear chemical processes, computational magnetics, process synthesis and design problems, and computational fluid dynamics, etc. (angira and babu 2006; angira 2006; stumberger et al. 2000; colaco et al. 2004). similarly multi-objective optimization strategies of differential algorithms are successfully applied on selected real world optimization problems (such as moo of ldpe tubular reactor, styrene reactor (both adiabatic and pseudo-isothermal reactor), pta oxidation process, etc.) (gujarathi and babu 2009a; gujarathi and babu 2009b; gujarathi and babu 2010a; gujarathi and babu 2010b; gujarathi and babu 2011). these moo strategies of de algorithms are in general found to outperform some other existing evolutionary strategies of moo. 4. results and discussion 4.1. effect of key parameters (cr, f, and np) modified differential evolution is a population based stochastic type of algorithm. it is observed that the outcome of stochastic algorithms, in general, depends on its control parameters. the performance of de and mde algorithms depends on key parameters, namely, np, cr, and f. three systems (system 1, 6 and k11 k13 k22 k24 p q r p q r p p, q, r q p, q, r cp0 = 1 x1 = cp1 x2 = ca2 x3 = cq1 x4 = cb2 optimization of process design problems using differential evolution algorithm    94 10) are selected randomly. table 3 shows the parameter values used in this study. all the table 3. parameter values used in present study. parameter value population size (np) 120 maximum number of generations (gmax) 150 crossover constant (cr) 0.8 scaling factor (f) random (0,1) optimization runs were run for 150 numbers of generations. various optimization runs are carried out by judiciously varying these control parameters for each of the selected systems. population size is plotted versus objective function and is shown through fig. 2a. there is no fixed guideline about the optimum size of population. if the population size is too low, it becomes difficult for an algorithm to converge   figure 2. effects of control parameters a) population size; b) crossover constant; and c) scaling factor. 0 50 100 150 200 0.000 0.002 0.004 0.006 0.008 0.010 system 1 system 6 system 10 o b je c tiv e f u n c tio n population size a) 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 system 1 system 6 system 10 o b je c tiv e f u n c tio n crossover constant (cr) b)   0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 -0.002 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 0.020 0.022 0.024 0.026 0.028 system 1 system 6 system 10 o b je c tiv e f u n c tio n scaling factor (f) random value of f c) a b c a.m. gujarathi, g. vakili-nezhaad and m. vatani    95 optimally. the computation cost increases by increasing the population size. in this optimization study, population specific runs are carried out by varying the population size for each system. results for three randomly selected systems are plotted and shown in fig. 2a. accordingly population size of 120 is selected and is used during the optimization study in this work. similarly the crossover constant and the scaling factors are also varied and results for three randomly selected systems are shown in fig. 2b, and fig. 2c. similar results could be obtained for all systems in this work. 4.2. comparison of mde, ga and other methods from literature the mde algorithm is used to calculate the interaction parameters of nrtl and two-suffix margules activity coefficient models in lle systems shown in table 1. both the optimization and model simulation codes are written and simulated using matlab (2009) software. during optimization study, three systems are randomly selected to study the effects of parameters on objective function. the objective function value with respect to generation number is plotted for 3 randomly selected systems. fig. 3 shows that irrespective of selected system (in this study); the mde algorithm converges to the minimum function value in less than 90 generations. however, the algorithm is allowed to run for 150 numbers of generations to ensure that the optimum is reached.   figure 3. convergence profile of selected systems using mde. population based stochastic algorithms have a typical feature that they start with population of points and all the points are expected to converge to the optimum points. the results are shown through table 4 and table 5. table 4 shows the calculated parameters of the individual models for each individual system. table 5 shows the results of rmsd obtained using mde and ga and other methods form literature. mde is able to converge to the global optimum having overall rmsd values of 0.0023 and 0.0170 for nrtl and two-suffix margules models. as an example, the experimental and calculated tie-lines were compared in figs. 4 and 5 for the system no.1. 4.3. optimization of cstr reactor design network table 6 shows the results obtained using de and mde for the optimization of reactor network design problem. results are shown considering average best, worst, mean, median, and standard deviation of the optimum values of 25 runs. though the experimental runs are taken for 500 numbers of generations, the comparison of results is carried out at intermediate numbers of functions evaluations (fes) ie. fes value of 10000, 15000, 20000, 50000 and 100000 (corresponding to 50, 75, 100, 250 and 500 generations respectively). taking standard deviation to be one of the measures of the effectiveness of any algorithm, it is observed that mde produces better results than those obtained using de. the consistency with which de and mde has performed is evident from the fact that standard deviation achieved a value of 0 (zero) or close to zero. figure 6 shows that error reduces faster in mde compared to de. table 7 shows fes, cpu time and function value for various problems using strategies of de. mde algorithm finds the global optimum in lesser number of fes than de. the cpu time taken by mde is much lesser compared to de. 5. conclusion binary interaction parameters of two wellknown activity coefficient models are calculated for 20 different extraction systems having ils based on flash systems using modified de and genetic algorithms. the activity coefficient models parameters as well as the rmsd have been obtained and are reported. the mde algorithm successfully converged to the global optimum having rmsd value of 0.0023 and 0.0170 for nrtl and two-suffix margules models respectively for 169 tie-lines. the nrtl   0 30 60 90 120 150 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 system 2 system 3 system 15 o b je ct iv e f u n c tio n generation number a.m. gujarathi, g. vakili-nezhaad and m. vatani    95 table 4. the calculated parameters of nrtl and two-suffix margules models for ternary systems listed in table 1. sys. no com. i-j genetic algorithm differential evolution (present study) nrtl margules nrtl margules aij aji αij aij aij aji αij aij 1 1–2 332.461 -373.942 0.348 0.598 1808.666 245.915 0.318 0.550 1–3 1229.120 -0.160 0.317 1.970 2248.897 723.989 0.304 1.912 2–3 2446.890 1302.380 0.167 31.003 1631.324 1402.217 0.258 17.380 2 1–2 391.263 -303.572 0.265 1.260 585.162 1184.298 0.500 3.055 1–3 1892.510 -460.185 0.282 2.543 1115.590 1147.705 0.485 4.006 2–3 2067.940 494.310 0.160 30.618 1793.492 468.268 0.155 40.318 3 1–2 1166.300 -291.783 0.202 1.481 1253.989 616.033 0.344 1.481 1–3 2287.100 -183.950 0.181 1.954 1770.190 846.111 0.339 1.955 2–3 2349.960 1661.440 0.399 7.870 2212.606 1168.640 0.249 7.874 4 1–2 546.571 -150.197 0.168 2.051 2.051 2.051 2.051 1.481 1–3 2299.100 -72.187 0.269 2.227 2.227 2.227 2.227 1.955 2–3 2349.940 611.617 0.158 6.263 6.263 6.263 6.263 7.874 5 1–2 2203.280 152.150 0.319 2.848 1450.995 228.000 0.468 2.848 1–3 1414.250 -19.610 0.262 2.880 1199.688 171.480 0.410 2.880 2–3 1156.610 902.890 0.199 26.417 1627.191 1495.390 0.476 26.413 6 1–2 718.490 -350.020 0.156 3.474 822.354 373.761 0.486 3.474 1–3 1416.000 -253.554 0.448 3.994 1007.923 237.460 0.490 3.994 2–3 1630.430 570.194 0.274 6.982 2182.877 1171.958 0.392 6.982 7 1–2 1614.230 -317.152 0.202 -3.423 1317.259 348.606 0.393 -3.421 1–3 913.786 1647.690 0.340 2.320 886.443 822.118 0.393 2.321 2–3 1999.940 1982.020 0.331 0.707 1996.124 780.712 0.393 0.708 8 1–2 841.486 1596.257 0.353 2.012 1256.108 2131.636 0.290 2.012 1–3 1452.005 998.742 0.388 2.173 1431.010 1021.001 0.353 2.173 2–3 1316.861 1018.664 0.317 8.219 2079.796 1749.430 0.452 8.219 9 1–2 741.522 -211.792 0.275 2.111 477.960 2350.571 0.459 2.005 1–3 1222.370 549.816 0.480 2.851 1398.317 602.968 0.47 3.208 2–3 1672.420 1349.360 0.297 27.852 2126.787 537.834 0.153 50.000 10 1–2 694.667 -388.136 0.172 3.326 337.318 2191.849 0.478 3.325 1–3 1105.550 -134.000 0.430 3.469 987.739 2022.673 0.490 3.468 2–3 2200.000 1145.020 0.359 5.085 1920.268 952.934 0.325 5.084 11 1–2 -204.989 704.990 0.170 2.285 350.795 503.925 0.404 2.285 1–3 936.160 635.653 0.479 2.839 879.405 853.769 0.436 2.839 2–3 1748.260 791.847 0.329 3.733 2375.034 978.002 0.335 3.733 12 1–2 -200.378 438.288 0.206 1.854 857.393 623.613 0.392 1.457 1–3 1741.060 108.973 0.305 2.443 2084.367 916.730 0.348 2.449 2–3 1737.620 956.303 0.163 23.571 2205.090 917.315 0.176 50 13 1–2 905.115 -272.115 0.171 1.256 1200.128 487.116 0.490 1.226 1–3 2499.900 533.974 0.328 2.242 1579.257 635.045 0.376 2.300 2–3 1839.160 1450.190 0.238 36.998 2124.391 1261.642 0.246 50.000 14 1–2 848.574 1072.305 0.455 2.833 665.977 814.838 0.498 2.832 1–3 1109.720 1479.689 0.429 3.234 2389.324 944.450 0.393 3.234 2–3 2499.746 1320.741 0.360 8.947 2137.053 1261.558 0.340 8.949 15 1–2 340.847 79.208 0.257 1.580 779.009 610.614 0.468 1.381 1–3 1493.610 310.828 0.396 2.485 2036.159 805.412 0.372 2.523 2–3 1989.530 720.570 0.161 37.999 1964.976 663.817 0.152 50.000 16 1–2 674.414 -237.622 0.237 3.003 403.228 1668.606 0.496 3.003 1–3 1257.920 -235.265 0.393 3.540 1767.738 1507.400 0.476 3.540 2–3 1632.750 644.722 0.347 5.327 2128.587 655.524 0.293 5.326 17 1–2 736.220 -378.023 0.293 2.718 294.293 1909.581 0.403 2.718 1–3 1669.349 -361.050 0.306 2.982 2169.359 1729.861 0.413 2.982 2–3 2143.899 436.847 0.240 5.921 1527.369 175.643 0.166 5.921 18 1–2 2045.840 -190.682 0.174 2.383 1800.158 954.327 0.327 2.383 1–3 1122.930 873.039 0.370 2.667 1301.322 1178.783 0.359 2.667 2–3 2101.480 1110.080 0.440 4.883 796.140 522.945 0.480 4.882 19 1–2 719.628 -251.316 0.174 1.428 705.099 325.113 0.418 1.428 1–3 1422.290 475.905 0.389 2.730 2364.234 801.241 0.336 2.730 2–3 2234.870 829.965 0.221 24.076 2078.687 572.206 0.172 24.083 20 1–2 -250.213 815.750 0.170 1.880 804.410 744.982 0.425 1.929 1–3 2308.960 -96.792 0.187 2.295 1976.340 856.057 0.360 2.325 2–3 1943.820 998.124 0.161 35.104 1931.866 1924.642 0.170 42.644 96 a.m. gujarathi, g. vakili-nezhaad and m. vatani  97 table 5. the results of rmsd obtained in this work using mde, ga and other methods in literature. sys. no. nrtl margules literature ga mde this work ga mde this work 1 0.0021 0.0013 0.0006 0.0019 0.0018 2 0.0155 0.0038 0.0028 0.0179 0.0139 3 0.0029 0.0008 0.0006 0.0040 0.0040 4 0.0050 0.0008 0.0011 0.0078 0.0078 5 0.0236 0.0039 0.0035 0.0073 0.0073 6 0.0041 0.0024 0.0036 0.0212 0.0212 7 0.0443 0.0031 0.0013 0.0339 0.0339 8 0.0020 0.0004 0.0016 0.0088 0.0087 9 0.0160 0.0059 0.0027 0.0230 0.0227 10 0.0065 0.0030 0.0029 0.0220 0.0219 11 0.0280 0.0062 0.0046 0.0339 0.0338 12 0.0050 0.0018 0.0010 0.0193 0.0184 13 0.0096 0.0038 0.0010 0.0038 0.0032 14 0.0145 0.0048 0.0024 0.0184 0.0184 15 0.0051 0.0027 0.0019 0.0192 0.0187 16 0.0157 0.0036 0.0037 0.0292 0.0291 17 0.0188 0.0060 0.0046 0.0273 0.0272 18 0.0359 0.0051 0.0041 0.0318 0.0317 19 0.0090 0.0028 0.0011 0.0130 0.0130 20 0.0108 0.0056 0.0009 0.0035 0.0033 overall 0.0159 0.0039 0.0023 0.0195 0.0170 table 6. results obtained for rnd problem using de and mde. algorithm fes best worst mean median std-dev de 10000 0.388811 0.388453 0.388789 0.388793 1.81e-05 15000 0.388811 0.388804 0.388811 0.388811 2.46e-07 20000 0.388811 0.388811 0.388811 0.388811 2.88e-09 50000 0.388811 0.388811 0.388811 0.388811 0 100000 0.388811 0.388811 0.388811 0.388811 0 mde 10000 0.388811 0.388723 0.388807 0.388806 4.62e-06 15000 0.388811 0.3811 0.388811 0.388811 2.92e-08 20000 0.388811 0.388811 0.388811 0.388811 2e-10 50000 0.388811 0.388811 0.388811 0.388811 0 100000 0.388811 0.388811 0.388811 0.388811 0   table 7. number of function evaluations (fes), cpu time and function value problem de mde rnd no. of function evaluations 14283 12699 cpu time (s) 3.634823 2.979619 function value (cost) 0.388811 0.388811 optimization of process design problems using differential evolution algorithm        figure 4. tie-lines for system no. 1, nrtl model, solid lines and full points: predicted tie-lines with ga; dashed lines and empty points: predicted tie-lines with mde.   figure 5. tie-lines for system no. 1, two-suffix margules model, solid lines and full points: predicted tie-lines with ga; dashed lines and empty points: predicted tie-lines with mde. figure 6. error against number of generations for reactor network design problem.   [mebupy][bf4]0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 propylbenzene 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 hexadecane 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0   [mebupy][bf4]0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 propylbenzene 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 hexadecane 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0   15 20 25 30 35 40 45 50 55 60 0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 e rr o r no. of generations de mde 98 a.m. gujarathi, g. vakili-nezhaad and m. vatani    99 model outperformed the two-suffix margules model by predicting more accurate (with one order of magnitude) rmsd values. the results obtained in this work using mde are better compared to the results obtained using ga and other traditional algorithms. this study can be extended for the calculation of some other activity coefficient models using similar approach. for rnd problem mde algorithm outperforms de in terms of number of function evaluations. acknowledgment the authors are grateful to petroleum and chemical engineering department of sultan qaboos university for supporting this study. references alkhaldi khae, al-tuwaim ms, fandary ms, al-jimaz as (2011), separation of propylbenzene and n-alkanes from their mixtures using 4-methyl-n-butylpyridinium tetrafluoroborate as an ionic solvent at several temperatures. fluid phase equilibria 309: 102– 107. al-maamari rs, vakili-nezhaad g, vatani m (2015), experimental and modeling investigations of the viscosity of crude oil binary blends: new models based on the genetic algorithm method. the journal of engineering research 11: 81-91. angira r (2006), evolutionary computation for optimization of selected non-linear chemical processes. ph.d. thesis (bits-pilani, india). angira r, babu bv (2006), optimization of process synthesis and design problems: a modified differential evolution approach. chemical engineering science 61: 4707-4721. babu bv, angira r (2006), modified differential evolution (mde) for optimization of nonlinear chemical processes. computers and chemical engineering 30: 989-1002. colaco mj, dulikravich gs, martin tj (2004), optimization of wall electrodes for electrohydrodynamic control of natural convection during solidification. materials and manufacturing processes 19: 719–736. dománska u, pobudkowska a, krolikowski m (2007), separation of aromatic hydrocarbons from alkanes using ammonium ionic liquid c2ntf2 at t = 298.15 k. fluid phase equilibria 259: 173–179. domínguez i, gonzález ej, gonzález r, domínguez a (2011), extraction of benzene from aliphatic compounds using commercial ionic liquids as solvents: study of the liquid– liquid equilibrium at t = 298.15 k. journal of chemical and engineering data 56: 3376–3383. fan hy, lampinen j (2003), trigonometric mutation operation to differential evolution. journal of global optimization 27: 105–129. garcía j, fernández a, torrecilla j, oliet m, rodríguez f (2009), liquid–liquid equilibria for {hexane + benzene + 1-ethyl-3methylimidazolium ethylsulfate} at (298.2, 313.2 and 328.2) k. fluid phase equilibria 282: 117–120. garcía j, fernández a, torrecilla j, oliet m, rodríguez f (2010a), ternary liquid−liquid equilibria measurement for hexane and benzene with the ionic liquid 1-butyl-3methylimidazolium methylsulfate at t = (298.2, 313.2, and 328.2) k. journal of chemical and engineering data 55: 258–261. garcía j, garcía s, torrecilla j, oliet m, rodríguez f (2010b), liquid−liquid equilibria for the ternary systems {heptane + toluene + n-butylpyridinium tetrafluoroborate or nhexylpyridinium tetrafluoroborate} at t = 313.2 k. journal of chemical and engineering data 55: 2862–2865. gujarathi am, babu bv (2009a), improved multi objective differential evolution (mode) approach for purified terephthalic acid (pta) oxidation process. materials and manufacturing processes 24: 303-319. gujarathi am, babu bv (2009b), optimization of adiabatic styrene reactor: a hybrid multi objective differential evolution (h-mode) approach. industrial and engineering chemistry research 48: 11115–11132. gujarathi am, babu bv (2010a), advances in optimization and simulation of low density polyethylene (ldpe) tubular reactor. in handbook of research on chemo informatics and chemical engineering, edited by a k haghi, nova science publishers, usa 109-138. gujarathi am, babu bv (2010b), multi-objective optimization of industrial styrene reactor: adiabatic and pseudo-isothermal operation. chemical engineering science 65: 2009-2026. gujarathi am, babu bv (2011), multiobjective optimization of industrial processes using elitist multi objective differential evolution (elitist-mode). materials and manufacturing processes 26: 455-463. gutierrez jp, meindersma w, de haan ab (2011), binary and ternary (liquid + liquid) equilibrium for {methylcyclohexane (1) + toluene (2) + 1-hexyl-3-methylimidazolium tetracyanoborate (3)/1-butyl-3methylimidazolium tetracyanoborate (3). the optimization of process design problems using differential evolution algorithm    100 journal of chemical thermodynamics 43: 1672– 1677. hansmeier ar, jongmans m, meindersma gw, de haan ab (2010a), lle data for the ionic liquid 3-methyl-n-butyl pyridinium dicyanamide with several aromatic and aliphatic hydrocarbons. the journal of chemical thermodynamics 42: 484–490. hansmeier ar, minoves ruiz m, meindersma gw, de haan ab (2010b), liquid−liquid equilibria for the three ternary systems (3methyl-n-butylpyridinium dicyanamide + toluene + heptane), (1-butyl-3methylimidazolium dicyanamide + toluene + heptane) and (1-butyl-3-methylimidazolium thiocyanate + toluene + heptane) at t = (313.15 and 348.15) k and p = 0.1 mpa. journal of chemical and engineering data 55: 708–713. holland jh (1992), algoritmos geneticos. investigacion y ciencia 192: 38–45. kirkpatrick s, gelatt cd, vecchi mp (1983), optimization by simulated annealing. science 220: 671–680. lei z, zhang j, li q, chen b (2009), unifac model for ionic liquids. industrial and engineering chemistry research 48: 2697-2704. letcher tm, deenadayalu n (2003), ternary liquid–liquid equilibria for mixtures of 1methyl3 – octyl imidazolium chlorideþbenzeneþan alkane at t=298.2k and 1 atm. the journal of chemical thermodynamics 35: 67–76. letcher tm, reddy p (2005), ternary (liquid+liquid) equilibria for mixtures of 1hexyl-3-methylimidazolium (tetrafluoroborate or hexafluorophosphate)+benzene+alkane at t=298.2 k and p=0.1 mpa. the journal of chemical thermodynamics 37: 415–421. maduro rm, aznar m (2008), liquid–liquid equilibrium of ternary systems 1-butyl-3methylimidazolium hexafluorophosphate + aromatic + aliphatic. fluid phase equilibria 265: 129–138. maduro rm, aznar m (2010), liquid–liquid equilibrium of ternary systems 1-octyl-3methylimidazolium hexafluorophosphate + aromatic + aliphatic hydrocarbons. fluid phase equilibria 296: 88–94. pereiro ab, rodriguez a (2009), application of the ionic liquid ammoeng 102 for aromatic/aliphatic hydrocarbon separation. the journal of chemical thermodynamics 41: 951–956. prausnitz jm, lichtenthaler rn, azevedo eg (1999), molecular thermodynamics of fluid phase equilibria, third ed., prentice-hall, new york. price kv, storn rm, lampinen ja (2005), differential evolution: a practical approach to global optimization. springer, berlin heidelberg. rashtchian d, ovaysi s, taghikhani v, ghotbi c (2007), application of the genetic algorithm to calculate the interaction parameters for multiphase and multicomponent systems. iranian journal of chemistry and chemical engineering 26: 89-102. renon h, prausnitz j m (1968), local compositions in thermodynamic excess functions for liquid mixtures. aiche journal 14: 135–144. revelli al, mutelet f, jaubert jn (2010), extraction of benzene or thiophene from nheptane using ionic liquids. nmr and thermodynamic study. the journal of physical chemistry b 114: 4600-4608. ryoo hs, sahinidis nv (1995), global optimization of nonconvex nlps and minlps with applications in process design. computers and chemical engineering 19: 551–566. sahoo rk, banerjee t, ahmad sa, khanna a (2006), improved binary parameters using ga for multi-component aromatic extraction: nrtl model without and with closure equations. fluid phase equilibria 239: 107–119. sahoo rk, banerjee t, khanna a (2007), uniquac interaction parameters with closure for imidazolium based ionic liquid systems using genetic algorithm. the canadian journal of chemical engineering 85: 833–853. seader jd, henley e j (2006), separation process principles, second ed., john wiley and sons, inc., new york. selvan ms, mckinley md, dubois rh, atwood jl (2000), liquid−liquid equilibria for toluene + heptane + 1-ethyl-3-methylimidazolium triiodide and toluene + heptane + 1-butyl-3methylimidazolium triiodide. journal of chemical and engineering data 45: 841-845. singh mk, banerjee t, khanna a (2005), genetic algorithm to estimate interaction parameters of multicomponent systems for liquid–liquid equilibria. computers and chemical engineering 29: 1712–1719. a.m. gujarathi, g. vakili-nezhaad and m. vatani  101 stragevitch l, davila sg (1997), application of a generalized maximum likelihood method in the reduction of multicomponent liquid-liquid equilibrium data. brazilian journal of chemical engineering 14: 41–52. stumberger g, dolinar d, pahner u, hameyer k (2000), optimization of radial active magnetic bearings using the finite element technique and the differential evolution algorithm. ieee transactions on magnetics 36: 1009–1013. vakili-nezhaad g, vatani m, asghari m (2013), calculation of the binary interaction and nonrandomness parameters of nrtl, nrtl1, and nrtl2 models using genetic algorithm for ternary ionic liquid systems. chemical engineering communications 200: 1102–1120. vakili-nezhaad g, vatani m, gujarathi am (2014), application of genetic algorithm to calculation of three-suffix margules parameters in ternary extraction ionic liquid systems. international journal of thermodynamics 17: 1-6. vatani m, asghari m, vakili-nezhaad g (2012), application of genetic algorithm to the calculation of parameters for nrtl and twosuffix margules models in ternary extraction ionic liquid systems. journal of industrial and engineering chemistry 18: 1715– 1720. zhou t, wang z, chen l, ye y, qi z, freund h, sundmacher k (2012), evaluation of the ionic liquids 1-alkyl-3-methylimidazolium hexafluorophosphate as a solvent for the extraction of benzene from cyclohexane: (liquid + liquid) equilibria. the journal of chemical thermodynamics 48: 145–149. civl061203.qxd the journal of engineering research vol.2, no. 1 (2005) 32-42 ___________________________________________ *corresponding authors e-mail: hossam@squ.edu.om development of asphalt pavement temperature models for oman h.f. hassan*1, a.s. al-nuaimi1, r. taha1, and t.m.a. jafar2 1department of civil and architectural engineering, sultan qaboos university, p.o. box 33, al khod 123, muscat sultanate of oman 2 research and technical studies department, muscat municipality, muscat, sultanate of oman received 6 december 2003; accepted 6 june 2004 abstract : asphalt pavements form an integral part of any transportation system. the structural capacity of the hot mix asphalt concrete layers depends on many factors including its temperature. moreover, temperature can be a major contributor to several types of distresses. therefore, temperature is a significant factor that affects the performance and life span of a pavement. the sultanate of oman's road network expanded at a phenomenal pace from approximately 10 km of paved roads in 1970 to 9,673 km in 2001. with the recent shrp and lttp research findings, it was necessary to investigate the applicability of the models developed from these research studies to oman's environmental conditions and more generally to the arabian gulf climate. this paper presents the research undertaken to develop models to predict high and low asphalt pavement temperatures in oman. a pavement monitoring station was set-up at the sultan qaboos university (squ) campus to monitor air, pavement temperatures and solar radiation. data were collected for 445 days. daily minimum and maximum temperatures were recorded. a regression analysis was used to develop the low pavement temperature model. a stepwise regression was used to develop high temperature models using air temperature, solar radiation, and duration of solar radiation as independent variables. the developed models were compared with the shrp and ltpp models. the shrp and ltpp models were found to be more conservative than the developed models, which are more suitable for predicting pavement temperatures in oman, and more generally in the gulf region. keywords: pavement temperature, models, shrp, ltpp, asphalt concrete, oman ¢¢üü∏∏îîàà°°ùÿÿgggòg .iqgô◊g álqo é¡æª°v øe πegƒy i~y ≈∏y á«à∏ø°sc’g áfé°sôÿg ∞°uq á≤ñw πª– iƒb ~ªà©jh ,π≤f áµñ°t icg ‘ »°ù«fq aõl á«à∏ø°sc’g ¥ô£dg èà©j : ¥ô£dg áµñ°t äqƒ£j ~bh gòg . ∞°uôdg äé≤ñw ôªyh agocg ≈∏y kgôkdƒe kóeéy iqgô◊g èà©j ∂dòdh .∞°uôdg äé≤ñw üƒ«y øe áø∏àfl ´gƒfcg qƒ¡x ≈∏y iqgô◊g òkéàd áaé°ve’éh ωé¶f qƒ£jh qƒ¡x ~©hh .2001 ωéy ‘ ºc 9700 ‹gƒm ¤eg 1970 ωéy ‘ áaƒ°uôÿg ¥ô£dg øe ºc 10 ‹gƒm øe iòñc ä’~©ã ¿éªy ‘ shrpèeéfôh h ltpp¿éªy ±hô¶d êpéªædg √òg áñ°séæe …~e ~j~–h ωé¶ædg gòg ∫ón øe äqƒw »àdg êpéªædg ≥«ñ£j øe ≥≤ëàdg øe ~h’ ¿éc âjôlg »àdg çéëhc’g èféàf qƒ¡x h ∞°uôdg äé≤ñ£d …ô¨°üdg h iƒ°ü≤dg √qgô◊g äélqo ™bƒàd êpé‰ oééje’ √dhgôleg ” …òdg åëñdg èféàf ábqƒdg √òg ¢vô©à°ùjh .»hô©dg è«∏ÿg á≤£æe ñéæÿ ºycg áø°ühh á«néæÿg ~j~–h ωƒj 445 i~ÿ äéfé«ñdg ~°uq ” ~bh .»°ùª°ûdg ´é©°te’g ∂dòch ∞°uôdg äé≤ñwh ƒ÷g iqgôm äélqo ¢sé«≤d á£fi ö«côjh ™bƒe õ«¡œ ” ~bh .¿éªy ‘ »à∏ø°sc’g êpƒ‰ oééje’ ägƒ£n hp …qg~ëfe’g π«∏ëàdg á≤jôwh ,iô¨°üdg √qgô◊g äélq~d êpƒ‰ oééje’ …qg~ëfe’g π«∏ëàdg á≤jôw ωg~îà°seg ”h ,á«eƒ«dg …ô¨°üdgh …ƒ°ü≤dg iqgô◊g äélqo á°sgq~dg √òg ‘ iqƒ£ÿg êpéªædg áfqé≤e ” ~bh .á∏≤à°ùe ägò¨àªc »°ùª°ûdg ´é©°te’g √îa ∂dòch »°ùª°ûdg ´é©°te’gh ƒ÷g iqgôm álqo ωg~îà°seéh iƒ°ü≤dg iqgô◊g äélq~d êpé‰ ™e shrphltppiqgôm äélqo ™bƒàd áªfóe ìccg èà©j »àdgh á°sgq~dg √òg ‘ iqƒ£ÿg êpéªædg øe ké¶ø– ìccg iònc’g êpéªædg ¿g áfqé≤ÿg øe úñj h .è«∏ÿg á≤£æe ‘ áeéy áø°ühh ¿éªy ‘ »à∏ø°sc’g ∞°uôdg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.¿éªy ,á«à∏ø°s’g áfé°sôÿg ,üqé°t ,êpé‰ ,∞°uôdg iqgôm álqo : ¿éªy ‘ »à∏ø°s’g ∞°uôdg iqgôm äélq~d êpé‰ ôjƒ£j 33 the journal of engineering research vol. 2, no. 1 (2005) 32-42 1. introduction asphalt pavements form an integral part of any transportation system. a significant factor that affects the performance and life span of a pavement is the influence of temperature. temperature can contribute to certain common types of asphalt pavement distresses such as permanent deformation or rutting (typically associated with high temperature environments), bleeding, and thermal cracking (associated with low temperature environments). the strategic highway research program (shrp) established the long term pavement monitoring program (ltpp) program in 1987 to support a broad range of pavement performance analyses leading to improved engineering tools to design, construct, and manage pavements diefenderfer, et al. (2002). the seasonal monitoring program (smp) was established as an element of ltpp in 1991 to measure and evaluate the effects of temperature and moisture variations on pavement performance and validate the available models (mohesni, 1998; diefenderfer, et al. 2002). from the initial shrp testing and smp data, several pavement temperature models were developed to assist in the proper selection of the asphalt binder performance grade (mohesni, 1998; mohesni and symons, 1998a; mohesni and symons, 1998b; lukanen, 1998; diefenderfer, et al. 2002). solaimanian and kennedy (1993) proposed an analytical model based on the theory of heat and energy transfer. shao et al. (1997) also developed a procedure based on heat transfer theory to estimate pavement temperatures. regression models based on other sets of data were developed (bosscher et al. 1998; marshall et al. 2001). a simulation model was developed to calculate temperatures during summer conditions based on the heat transfer models developed by solaimanian and kennedy (hermansson, 2000 and 2001; diefenderfer, et al. 2002). al-abdul wahhab et al. (1994) conducted a study in two regions in saudi arabia to manually measure pavement temperatures in different pavement sections. the study concluded that the extreme pavement temperatures in arid environment ranged between 3 and 72°c, while in coastal areas, the temperature ranged between 4 and 65°c. in another study, al-abdul wahhab et al. (1997) recommended five performance graded binder zones for the whole gulf area. the study also proposed modification of the currently used binders to suite the proposed grades. the sultanate of oman as well as the gulf area, in general, possesses a different environment from that of the north america. the applicability of the recent shrp and ltpp developments in the us to oman's or the arabian gulf's environmental conditions needed to be evaluated. this paper presents the research undertaken to develop models to predict high and low asphalt pavement temperatures in oman. a pavement monitoring station was established at the sultan qaboos university (squ) campus to monitor air, pavement temperatures, and solar radiation. data was collected for 445 days. daily minimum and maximum temperatures were recorded. a regression analysis was used to develop the low pavement temperature model. a stepwise regression was used to develop high temperature models using air temperature, solar radiation, and duration of solar radiation as independent variables. the instrumentation used is described and collected data are presented. the developed models were compared with the shrp and ltpp models. notation df = degree of freedom, f = f statistic, lat. = latitude, ms = mean sum of squares, p-value = power of the test, r2adj. = adjusted coefficient of determination r2 = coefficient of determination, see = standard error of estimate, solar = cumulative solar radiation from dawn to the time of occurrence of maximum air temperature , w.hr/m2, ss = sum of squares, std. error = standard error or standard deviation of the coefficient's distribution, tair = air temperature, °c, trad = duration of sunshine from dawn to the time of occurrence of maximum air temperature, hours, t-stat = t statistic, tsurf = pavement surface temperature, °c, α = level of significance. 2. pavement temperature monitoring station a monitoring station was set-up to collect data on air temperature, solar radiation and pavement temperatures at various depths. sensors were connected to a cr10x campbell scientific datalogger housed in a protection enclosure. the enclosure was mounted on an instrumentation tower. the datalogger was operated by a battery. the following section describes the installation and sensor locations. 2.1. installation of station a test area 7x8 m was selected for the station installation inside a newly built parking lot at the sultan qaboos university campus. the pavement cross section was modified from the original design in the area of the test section to be composed of 135 mm of asphalt concrete (ac) wearing course on top of 165 mm of unbound base course on top of compacted subgrade soil. after the construction of the pavement was completed, the installation of the sensors was initiated. the procedure started with the excavation of a 500 mm by 500 mm by 480 mm deep pit. a trench 200 mm wide by 480 mm deep was also excavated from the pit to where the tower was installed. a fence was erected around the tower area to protect the equipment from damage. a 50 mm pvc pipe was laid in the trench to protect the sensor cables extending from the pit to the tower location. all sensors were inserted into the pvc pipe and led to the excavation pit providing adequate slack so as to allow for comfortable installation of the probes at the desired locations. two sets of thermistors referred to as center and right, were installed in the center and right sides of the pit. the right set consisted of eight sensors covering the full depth of the excavated pit. the center set had four sensors that covered only the depth of the ac layer to serve as a duplicate for the first set. at the targeted sensor locations, a mark was made at the appropriate side of the pit. an electrically powered horizontal and 90° angle hammer drill was used to drill holes approximately 280 mm deep into the side of the excavated pit. the 90° angle was used to accommodate the hammer drill and drilling bit inside the limited pit space. the process was repeated for each sensor with the exception of those located at the surface (at 0 mm depth). holes were cleared of loose debris before sensors were inserted. sensors were then inserted all the way into the drilled holes and sealed in place using a silicon sealant to prevent concrete or moisture infiltration when backfilling the pit. figure 1 shows the sensors inserted in the drilled holes at the desired depths. for surface temperature sensors, a grinder was used to cut through the pavement surface such that the probe could be flush with the pavement surface. the probe was positioned in place and a cold sand asphalt mix was used to seal around it. all actual depths of pavement sensors were measured and recorded. figure 2 shows a schematic representation of the installation pit showing the layer thicknesses and the actual depth of the sensors. after adjusting the extra cable lengths, the pit and trench were filled with the subgrade and base material. compaction of the material was performed manually. portland cement concrete was used to backfill the pit up to the surface level instead of asphalt concrete due to the ease of backfilling with concrete in addition to the fear of damaging the sensor cables from the hot mix. it was felt that using the concrete instead of asphalt concrete will not affect the readings from the probes, since the sensing tip of the probe (60 mm long) was buried inside the asphalt concrete side of the pit at least 220 mm away from the interface between the existing pavement and poured concrete. finally, the tower was erected and the datalogger solar radiation sensor and air temperature thermistor were mounted on the tower. the sensors were connected to the datalogger. figure 3 shows an overview of the test area 34 the journal of engineering research vol. 2, no. 1 (2005) 32-42 figure 1. installation of thermistors 35 the journal of engineering research vol. 2, no. 1 (2005) 32-42 section of instrumentation hole (right) (center) section location of sensors section location of sensors plan of instrumentation hole (dimensions are in mm) figure 2. pavement layers and thermistor locations figure 3. general view of experimental test area pit and trench tower and data logger cabinet after the set-up of the station. a program was written to operate the cr10x datalogger based on the available sensors configuration. the reading from the solar radiation sensor is taken every minute, and the average value is stored every 15 minutes. all other sensors are read every 15 minutes. data are stored on the datalogger and downloaded to a computer via a storage module. 3. data analysis the temperature data collected were presented to show the different variability patterns between air and pavement temperatures. pavement temperature models were established using regression analysis on the collected daily maximum and minimum temperatures. the data used to 36 the journal of engineering research vol. 2, no. 1 (2005) 32-42 15 20 25 30 35 40 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 time of day t em pe ra tu re , o c air temperature surface 20 mm 65 mm 125 mm 210 mm 290 mm 375 mm figure 4. air and pavement temperature variation on the 24th and 25th december, 2002 20 25 30 35 40 45 50 55 60 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 time of day t em pe ra tu re , o c air temperature surface 20 mm 65 mm 125 mm 210 mm 290 mm 375 mm figure 5. air and pavement temperature variation on the 2nd and 3rd of may 2003 develop the models were collected in the period from 23rd of december 2002 to 6th of april 2004, with a total of 445 days. 3.1. temperature data figures 4 and 5 show the variation of air and pavement temperatures at different depths with the time of day. fig. 4 shows data obtained for two consecutive days during the cold season; namely the 24th and 25th of december, 2002, while fig. 5 shows this variation for two consecutive days during the hot season; namely the 2nd and 3rd of may, 2003. it can be seen that thermistors placed at 290 mm depth or more (located at the boundary of the subbase and subgrade layers) are less sensitive to changes in the air temperature than sensors located from the surface down to 210 mm deep. the surface layer is the layer most sensitive to temperature fluctuations. there is a noticeable time lag in the heating cycle between the maximum air temperature and pavement temperature. in other words, it takes more time to heat the pavement than the air. the time lag increases as pavement depth increases. time lag is also noticed between the minimum air temperature and pavement temperatures in the cooling cycle. the pavement surface cools quicker than other deeper layers. figures 4 and 5 also indicate that the maximum or minimum pavement temperatures occur some time after the occurrence of maximum or minimum air temperature, respectively. the minimum pavement temperature is recorded at the surface but is generally warmer than minimum air temperature. this agrees with the findings reported in the literature. figures 6 and 7 show the same data only for the air, 37 the journal of engineering research vol. 2, no. 1 (2005) 32-42 10 15 20 25 30 35 40 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 time of day te m pe ra tu re , o c 0 50 100 150 200 250 so la r r ad ia tio n, w .h r/m 2 air temperature surface 20mm depth solar radiation 20 25 30 35 40 45 50 55 60 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 20 0 40 0 60 0 80 0 10 00 12 00 14 00 16 00 18 00 20 00 22 00 0 time of day t em pe ra tu re , o c 0 50 100 150 200 250 300 so la r r ad ia tio n, w .h r/m 2 air temperature surface 20mm depth solar radiation figure 6. temperature and solar radiation for 24th and 25th of december, 2002 figure 7. temperature and solar radiation for 2nd and 3rd of may, 2003 surface and 20 mm sensors, in addition to the solar radiation. fig. 4 illustrates some fluctuation in temperatures detected by the air temperature sensor and the sensors placed at the surface and 20 mm depth. this is more clearly illustrated in fig. 6 where the readings obtained for solar radiation as well as the readings from the surface and 20 mm sensors are plotted. this effect is not seen in hotter conditions (fig. 7). this can be explained by the cloudy weather conditions encountered in winter that shade the sun's rays, as shown by the fluctuation in solar radiation readings presented in fig. 6. it can also be seen that the temperature of the pavement at greater depths is warmer than those placed nearer to the surface at night time. this pattern is maintained until the minimum air temperature is reached, and the pattern is reversed with the surface layers becoming warmer and gain energy from the sun rays during the day while the deeper layers become cooler. 3.2. development of low temperature model the low pavement temperature model was established using air temperature as an independent variable and pavement surface temperature as the dependant variable. the lowest air temperature in any single day was identified for the entire period during which data were collected and the corresponding daily minimum pavement temperature at the surface was obtained. a regression analysis was performed using minitab statistical software to develop the model (neter et al., 1990; minitab, 2000). the model was developed using 445 data points. tsurf = -1.437 + 1.121 tair (1) where, tsurf = minimum pavement temperature, ºc and tair = minimum air temperature, °c. this relationship with the collected data is illustrated in figure 8. the model achieved a goodness-of-fit (r2) of 0.968 and a standard error of estimate (see) of 1.11 °c. table 1 shows the anova results for the model. it is to be noted that the lowest pavement temperatures recorded in oman are well above the low pavement temperatures typically encountered in the us. this is due to the gulf area climate which does not have subfreezing temperatures. 3.3. development of high pavement temperature models the highest air temperaure and the highest pavement temperature at a depth of 20 mm below the surface for each day were identified. a linear regression was performed using the highest air temperature as the independent variable and the highest 20 mm pavement temperature as the dependent variable. using 445 days of collected data, the following equation was obtained: 38 the journal of engineering research vol. 2, no. 1 (2005) 32-42 tsurf = -1.437 + 1.121 t air r2 = 0.968 10.0 15.0 20.0 25.0 30.0 35.0 40.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 air temperature, oc pa ve m en t s ur fa ce te m p. , o c figure 8. developed low pavement temperature prediction model df ss ms f regression 1 16155 16155 13150.93 residual 443 544 1 total 444 16699 coefficients std. error t-stat p-value intercept -1.4373 0.239 -6.02 0.000 variable 1.1208 0.010 114.68 0.00 0 table 1. anova table for low temperature model* *df = degree of freedom; ss = sum of squares; ms = mean sum of squares, f = f statistic, std.error = standard error or standard deviation of the coefficient’s distribution t20mm = 3.160 + 1.319 tair (2) where, t20mm= pavement temperature at 20 mm depth, ºc and tair = maximum air temperature, ºc. figure 9 shows the data collected as well as the regression equation obtained. the equation achieved an r2 of 0.847 and an see of 3.18°c. table 2 shows the anova results for the model. since high pavement temperature is affected by factors other than air temperature, other parameters were considered with the aim of improving r2 of the model. three variables were considered as independent variables: air temperature (tair), cumulative solar radiation from dawn to the time at which the maximum air temperature occurred (solar), and the duration over which solar radiation is accumulated (trad). the pavement temperature at 20 mm depth was taken as the dependant variable. in order to obtain the best model, a stepwise regression was performed. the accumulated solar radiation was considered as this is the energy that the pavement absorbs and contributes to the highest pavement temperature. the duration to the maximum pavement temperature would have been more appropriate to consider, however, the identification of the maximum pavement temperature would not be possible at other weather stations where no pavement sensors would be available. the addition and removal of variables were allowed in the stepwise regression procedure. the partial f-test was used to verify for the entry and removal of variables from the model. the stepwise regression resulted in one additional model. thus, a second model was developed using pavement temperature as the dependent variable, and air temperature and cumulative solar radiation as the independent variables. the developed equation was as follows: t20mm = 2.713 + 1.281 tair + 0.00053 solar (3) where, solar = cumulative solar radiation from dawn to the time of occurrence of maximum air temperature in w.h/m2. r2 increased and see decreased slightly compared to the first model (eq. 2) to 0.854 and 3.12°c, respectively. 39 the journal of engineering research vol. 2, no. 1 (2005) 32-42 t20mm = 3.160 + 1.319 t air r2 = 0.847 20.0 30.0 40.0 50.0 60.0 70.0 80.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 air temperature, oc pa ve m en t t em p. , o c figure 9. developed high pavement temperature prediction model df ss ms f regression 1 24512 24512 2416.44 residual 435 4413 10 total 436 28925 coefficients std. error t-stat p-value intercept 3.160 0.907 3.48 0.001 variable 1.319 0.0268 49.16 0.000 table 2. anova table for high temperature model with air temperature variable* * same notations as in table 1 table 3 shows the results of the stepwise regression. the analysis used the same level of significance (0.05) to add or remove variables. using the partial f-test, if the p value is less than 0.05, then there is enough evidence that the variable is not zero at 0.05 level of significance and the variable should be kept. otherwise, the variable should be removed. the table shows that the p value is less than 0.05 in both models. the third parameter (trad) was removed from the model as the p value was more than 0.05. a correlation between the cumulative solar radiation and the duration (trad) variables was found to be 0.839 as expressed by the pearson product moment correlation coefficient. this shows a strong correlation between the two variables, and supports the removal of the (trad) variable. the use of the developed relationships depends on the availability of data. equation 2 can be used if only air temperature measurements are available. the more accurate model eq. 3 can be used if solar radiation measurements are available. it should be noted that solar radiation may be estimated using the relationship described in details by diefenderfer et al. (2002) which estimates solar radiation using the geographical latitude of any given location and day of the year. 4. comparison of developed models with shrp and ltpp models 4.1. low temperature model figure 10 shows a comparison between the developed low temperature model and shrp and ltpp low temperature prediction models for pavement surface temperature. the collected field data are also shown in fig. 10. the shrp and lttp low temperature models are given by the following relationships, respectively mohseni, (1998): tsurf = 0.859tair + 1.7 (4) tsurf = 1.56 + 0.72tair 0.004lat2 + 6.26 log10 (25) (5) 40 the journal of engineering research vol. 2, no. 1 (2005) 32-42 step 1 2 constant 3.160 2.713 tair 1.319 1.281 t-stat. 49.16 46.14 p-value 0.000 0.000 solar 0.00053 t-stat. 4.32 p-value 0.000 see 3.18 3.12 r2 84.74 85.37 r2 adj. 84.71 85.31 table 3. stepwise regression analysis for high pavement temperature model σ for entry = 0.05, α to remove = 0.05 10 15 20 25 30 35 40 10 15 20 25 30 35 40 air temperature, oc pa ve m en t s ur fa ce te m p. , o c ltpp shrp low temp. model figure 10. comparison between low pavement temperature prediction model and shrp and lttp models where, tsurf = pavement surface temperature, tair = air temperature, and lat=latitude. the latitude was substituted with the seeb international airport (muscat) latitude, which is 23° 35' 00". figure 10 clearly shows that the predicted shrp and ltpp low pavement temperature is lower than both the measured values and the low pavement temperature predicted by the developed model. both models tend to be closer to both the collected data and the developed model at temperatures less than 15°c. this indicates that the developed model is more suitable for oman's climate and that shrp and ltpp models would be more conservative in the performance grade (pg) binder selection. 4.2. high pavement temperature models figure 11 shows a graphical comparison between the developed high pavement temperature prediction model, shrp and ltpp models. the collected data are also presented. shrp and ltpp models are given by the following relationships, respectively mohesni, (1998): t20mm = (tair 0.00618lat2 + 0.2289lat + 42.4) 0.9545 17.78 (6) t20mm = 54.32 + 0.78tair 0.0025lat2 15.14log10(45) (7) where, t20mm = pavement temperature at 20 mm depth, tair = air temperature, and lat=latitude. the latitude was replaced with muscat latitude. figure 11 illustrates that the predicted shrp and ltpp high pavement temperature at 20 mm are higher than both the measured values and the high pavement temperature predicted by the developed model. therefore, the developed model is more representative of oman's climatic conditions. shrp and ltpp models would be expected to result in a more conservative selection of the pg binder. 5. conclusions the following conclusions can be drawn from the analysis of the data collected in this research: 1. a pavement temperature monitoring station was set-up on sultan qaboos university campus. the collected data indicated the expected trends for cycles of heating and cooling of the pavement layers. 2. a low pavement temperature model was developed for oman based on air temperature as an independent variable. the model achieved an r2 of 0.968. 3. the low temperature model, when compared with shrp and ltpp, showed that the two later models provide lower estimate of pavement temperature compared with the developed model. shrp and ltpp models are more conservative in the selection of performance grade (pg) binders, and the developed model would be more suitable for oman's climate. 4. two high pavement temperature models were developed, taking into account air temperature and cumulative solar radiation from dawn to the time of maximum air temperature as independent variables to predict the high pavement temperature at 20 mm depth. a stepwise regression analysis was used to develop the models. the models achieved r2 values of 0.847 and 0.854 with the addition of the air and solar radiation variables, consecutively. the partial f-test indicated that the duration of solar radiation up to the maximum air temperature variable for the three variables model should be removed at 0.05 level of significance. 5. the comparison of high pavement temperature model using only air temperature as an independent variable with shrp and ltpp models indicated that the latter models would estimate higher pavement temperatures 41 the journal of engineering research vol. 2, no. 1 (2005) 32-42 20 30 40 50 60 70 80 15 20 25 30 35 40 45 50 air temperature, oc pa ve m en t t em p. , o c ltpp shrp high temp. model figure 11. comparison between high pavement temperature prediction model and shrp and ltpp models for oman. therefore, these models would be conservative in the selection of performance grade (pg) binders for oman compared with the developed model. 6. the use of the high pavement temperature model would depend on the availability of data. if solar radiation measurements are available or can be estimated, then the two variable model can be used. acknowledgments the authors gratefully acknowledge the financial support by sultan qaboos university through an internal research grant. the authors acknowledge the help provided by the project and maintenance department in the university for the station installation. references al-abdul wahhab, h.i. and balghunaim, f.a., 1994, "asphalt pavement temperature related to arid saudi environment," j. of materials in civil engineering, american society of civil engineers, 6 (1), pp. 1-14. al-abdul wahhab, h.i., asi, i.m., al-dubabe, a. and ali, m.f., 1997, "development of performance-based bitumen specifications for the gulf countries," construction and building materials, 11(1), pp. 15-22. bosscher, p. j., bahia, h. u., thomas, s. and russell, j. s., 1998, "the relation between pavement temperature and weather data: a wisconsin field study to verify the superpave algorithm," transportation research record 1609, trb, national research council, washington, dc, pp. 1-11. diefenderfer, b.k., al-qadi, i.l., reubush, s.d. and freeman, t.e., 2002, "development and validation of a model to predict pavement temperature profile," presented at transportation research board 82nd annual meeting, washington dc. hermansson, a., 2000, "mathematical model for calculation of pavement temperatures: comparisons of calculated and measured temperatures," transportation research record 1699, trb, national research council, washington, dc, pp. 134-141. hermansson, a., 2001, "mathematical model for calculation of pavement temperatures: comparisons of calculated and measured temperatures," transportation research record 1764, trb, national research council, washington, dc, pp. 180-188. lukanen, e. o., han, c. and skok, e. l. jr., 1998, "probabilistic method of asphalt binder selection based on pavement temperature," transportation research record 1609, trb, national research council, washington, dc, pp. 12-20. marshall, c., meier, r.w. and welsh, m., 2001, "seasonal temperature effects on flexible pavements in tennessee," presented at transportation research board 80th annual meeting, washington, dc. minitab statistical software release 13, 2000, user's guide, minitab inc. mohseni, a., 1998, "ltpp seasonal asphalt concrete (ac) pavement temperature models," report fhwa-rd-97-103, federal highway administration, u.s. department of transportation, p. 71. mohseni, a. and symons, m., 1998a, "effect of improved ltpp ac pavement temperature models on superpave performance grades," presented at transportation research board 77th annual meeting, washington, d.c. mohseni, a. and symons, m., 1998b, "improved ac pavement temperature models from ltpp seasonal data," presented at transportation research board 77th annual meeting, washington, d.c. neter, j., wasserman, w. and kunter, m.h., 1990, "applied linear statistical models," third edition, richard d. irwin, inc., boston, p. 1181. shao, l., park, s.w. and kim, y.r., 1997, "a simplified procedure for prediction of asphalt pavement subsurface temperatures based on heat transfer theories," transportation research record 1568, trb, national research council, washington, dc, pp. 114123. solaimanian, m. and kennedy, t.w., 1993, "predicting maximum pavement surface temperature using maximum air temperature and hourly solar radiation", transportation research record 1417, trb, national research council, washington, dc, pp. 1-11. 42 the journal of engineering research vol. 2, no. 1 (2005) 32-42 microsoft word paper 6.docx the journal of engineering research, vol. 11, no. 2 (2014) 64-78 a basic wind speed map for oman a.s. alnuaimia, m.a. mohsinb and k.h. al-riyamic a department of civil & architectural engineering, sultan qaboos university, pob 33, pc 123, muscat, oman b department of civil and architectural engineering, college of engineering, university of buraimi, po box 890, buraimi 512, oman. c senior structural engineer, central design office, royal court affairs, po box 949, muscat 100, oman. received 17 may 2014; accepted 14 september 2014 abstract: the aim of this research was to develop the first basic wind speed map for oman. hourly mean wind speed records from 40 metrological stations were used in the calculation. the period of continuous records ranged from 4–37 years. the maximum monthly hourly mean and the maxima annual hourly mean wind speed data were analysed using the gumbel and gringorten methods. both methods gave close results in determining basic wind speeds, with the gumbel method giving slightly higher values. due to a lack of long-term records in some regions of oman, basic wind speeds were extrapolated for some stations with only short-term records, which were defined as those with only 4– 8 years of continuous records; in these cases, monthly maxima were used to predict the long-term basic wind speeds. accordingly, a basic wind speed map was developed for a 50-year return period. this map was based on basic wind speeds calculated from actual annual maxima records of 29 stations with at least 9 continuous years of records as well as predicted annual maxima wind speeds for 11 short-term record stations. the basic wind speed values ranged from 16 meters/second (m/s) to 31 m/s. the basic wind speed map developed in this research is recommended for use as a guide for structural design in oman. keywords: oman, basic wind speed, basic wind speed map, weather stations. אא  אkאkאkk  אw אאאאאkא   40א א א kא א       437k אאאאאאאאאא gumbel  gringorten אk  אא  gumbel   k      א     א  אא  א א אא  א f א א  א   א48 eא א  א  א א א   kא  א   50   א א29f9אאאאאאאeאא אא11אאאאk16l31kl kאאאאאאא אw،אא،אא،א.  *corresponding author’s e-mail: alnuaimi@squ.edu.om a.s. alnuaimi, m.a. mohsin and k.h. al-riyami 1. introduction 1.1 literature review as oman is not considered a potential earthquake zone, the effects of wind loads on structures are considered the dominant factor when determining lateral loads on edifices such as buildings, chimneys, power transmission towers, and satellite communication towers. wind speed measurement in oman is carried out by the directorate general of civil aviation and meteorology under the ministry of transport and communications. as of the end of 2013, there were 55 meteorological stations throughout the country as shown in fig. 1. the oldest station, at muscat international airport, was constructed in 1977. the stations measure the hourly mean wind speed, an hourly threesecond gust, temperature, precipitation and humidity. wind speed is measured at a 10-meter height in open and level ground by cup anemometers whereas wind vanes are used to determine wind direction. basic wind speed (vb) is the wind speed estimated by different methods to be exceeded on average only once in 50 years (gibbs et al. 1981). guide to the use of the wind load provisions of the american society of civil engineers (asce) 7-02 (kishor and delahay 2004) requires a 2% exceedance probability which equals to a 50-year return period. for crucial facilities such as hospitals, only a 1% exceedance probability on a 100-year return period is allowed. for structures with a low risk to human life if compromised, an exceedance limit of up to 4% is allowed. the minimum design loads for buildings and other structures (asce/sei-7-10) requires different exceedance probabilities in a 50-year return period depending on the location and occupancy category of the structure. it uses the three-second gust in the calculation of basic wind speed. the basic wind speed for the design of buildings in the british standard, annex b of the bs6399-2 (1997), depends on the hourly mean wind speed records. the monthly maximum of the hourly mean is used to obtain the annual maxima, which are used to develop basic wind speed maps. the basic wind speed has an annual risk of exceedance of 2% (a 50-year return period). simiu (2009) stated that movement toward international standards appears to justify the use of 10-minute mean speeds, thus conforming to the practice sanctioned by the world meteorological organization or of other internationally acceptable measures of sustained wind speeds. a procedure for estimating extreme wind speeds at locations where longterm data are not available was reported by (simiu et al. 1982; grigoriu 1984). the method was tested on 67 three-year records in the usa and found to infer the approximate probabilistic behaviour of extreme winds from data consisting of the largest monthly wind speeds recorded over a period of three years or longer. similarly, (kramer and gerhardt 1988) carried out a simulation process to predict extreme wind speed from actual 22 year records. they stated that it is not necessary to have 22 years of continuous records and that even a five-year record is adequate for showing good comparisons between the simulated results and historical data. among many statistical methods, the (gumbel 1958; gringorten 1963) methods are the most popular statistical methods used in meteorology for calculating basic wind speeds. these methods have been used extensively for analyses of wind speed records in many parts of the world (dyrbye and hansen 1999; holmes 2001). harris (1996) re-examined gumbel’s extreme value distribution for analysing annual maxima wind speeds or similar data and suggested an interchange of the axes of this method. he proposed an automated procedure that gives exactly the same results. an and pandy (2005) made a comparison of methods of extreme wind speed estimation. they studied four different methods: gumbel’s method, the modified gumbel distribution, the peak-overthreshold method and cook’s method of independent storms. these four methods were applied to a common data set consisting of six stations in the usa. they concluded that the standard gumbel method tends to provide an upper bound estimate of 50/500 year design wind speed, and cook’s method of independent storms estimates of design wind speed exhibits a 65 a basic wind speed for oman figure 1. locations of meteorological stations in oman (names of stations are as provided by directorate general of civil aviation and meteorology). 66 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami more stable trend with limited threshold sensitivity, which is in contrast with the rapidly fluctuating estimates obtained from the peaks over threshold methods. abohemda and alshebani (2010) used gumble’s method to predict basic wind speed values for libya for a 50-year return period from the data of 22 stations. lakshmana et al. (2009) studied the wind speed data from 70 metrological centres in india and calculated basic wind speeds using gumbel’s approach. al maawali et al. (2008) studied wind data from 10 weather stations in oman. they calculated the basic wind speed using the three-second gust at the locations of 10 stations using the gumbel and gringorten methods. they found that both methods yielded similar results. dorvlo (2002) used the weibull distribution to model wind speeds at four locations in oman. the scale and shape parameters were estimated using pearson’s chi-squared test, the method of moments, and the regression method. it was observed that the estimates using the chisquared test gave the best overall fit to the distribution of the wind data. choi and tanurdjaja (2002) studied extreme wind estimates in singapore using gumbel’s method and cook’s method of independent storms. they concluded that the independent storm method and gumbel’s method both gave quite similar results. kasperski (2002) developed a wind speed map for germany based on a refined extreme value analysis using high gust wind speeds. he divided the country into four basic wind speed zones ranging from 22.5 meters/second (m/s) to 32.5 m/s. akosy et al. (2004) used the wavelet approach to generate hourly mean wind speed data using normal and weibull probability distribution functions. they concluded that the wavelet approach can be used as an alternate to existing generation methods. sahin (2003) developed an hourly mean wind speed exceedance map for turkey using the gumbel-lieblein blue method. in this research, the gumbel and gringorten methods were used for the analysis of hourly mean wind speed records. 1.2 problem statement and motivation due to the recent extensive development of major cities in oman, the trend toward building higher and lighter structures has emerged only during the last two decades. wind load must be taken into consideration during the structural design of high-rise buildings, chimneys, transmission towers, and so on, requiring basic wind speed values. the basic wind speed is determined according to the climatic condition of each region. before this research, the hourly mean wind speed data collected from 55 metrological stations had not been utilized in the development of a basic wind speed map for oman and, as of this article’s publication, no local unified code or written regulations are available for the design of high-rise buildings. the need for a basic wind speed map that covers different regions of the country has become evident. the aim of this research was to develop a basic wind speed map which can be used as a code guideline or regulation for structural design in oman. 1.3 gumbel verus gringorten method gumbel (1958) developed an easily usable methodology for fitting recorded monthly or annual wind speeds. the asce (1991) stated that the gumbel extreme-value procedure is well accepted by other investigators, including (simiu et al. 1998). the gringorten (1963) method is considered a simple modification of gumbel’s extreme-value procedure. the only difference between gumbel and gringorten’s methods is in the calculation of the probability of non-exceedance (p). the procedure of calculating the basic wind speed is as follows: step 1: data are ranked in ascending order (1, 2, 3,.., m,) where m is the largest value. the total number of readings is called n (ie. number of years). step 2: the probability of non-exceedance (p) is determined according to gumbel: = (1) gringorten: = .. (2) step 3: reduced variant, y is calculated as = −ln (− ) (3) i = record identification number. 67 a basic wind speed for oman step 4: the data are plotted against the reduced variant, y, and a line of best fit is drawn, usually by means of linear regression. therefore, if the best fitting line y = ax + u, then u = the mode of distribution = the intercept of the line and a = the slope of the line (scale factor for x). step 5: for different values of the return period, eqn. 4 can be used to calculate the basic wind speed, = + ln (4) where r = return periods in years (eg. 50, 100, etc.) and vb = basic wind speed in m/s. 2. methodology the basic information about oman’s 55 weather stations is given in table 1. more than 8,200 records of monthly maximum hourly mean wind speed were obtained from the directorate general of civil aviation and meteorology in oman. descriptive statistics along with box plot graphs were developed for each station to ensure the homogeneity and independence of raw data, and to exclude extreme outliers. the records were divided into three types: long term, short term and not considered. long term means stations that have continuous records for at least nine years (29 stations); short term means stations that have continuous records of at least four years and less than nine years (11 stations), and not considered (15 stations) means records are not included in the analysis due to discontinuity of the records (less than four continuous years of records are required to be classified as not considered). the monthly maximum and the annual maxima of the hourly mean wind speed records of the 29 long term stations as well as the monthly maximum hourly mean wind speed records of the 11 short term stations were used in the analysis using gumbel’s and gringorten’s methods. the wind speed for each station was plotted against a reduced variate by the method of order statistics (ascending order), and a straight line of best fit was drawn. the intercept and slope of these lines give the mode, u, and slope, a, of the fitted curve to the fisher-tippet type i (ft1) extreme value distribution, respectively. a comparison between the results given by both methods was made. a relationship between the monthly maximum and the annual maxima records was developed and used for predictions of the annual maxima basic wind speeds from the monthly maximum hourly mean wind speed of the 11 short-term stations. finally, a basic wind speed map for oman was developed based on actual records of annual maxima and predicted annual maxima stations records. 3. data analysis and results table 2 shows that the data used from the shortand long-term stations (40 stations with 6,736 records) are consistent, with very few records considered as outliers or extreme outliers based on box plot analysis. the extreme outliers have been excluded from the analysis of basic wind speed calculations. the two worst cases were 1.8% and 1.3% at muscat international airport and the town of saiq, respectively. figure 2 shows typical monthly maximum hourly mean wind speed data for saiq from 1988–2013 based on a box plot analysis. the monthly records ranged from 5.564–21.896 m/s. this range is distributed into three areas to the left and right of the median, with each area representing 1.5 of interquartile ranges (iqr). four records were excluded because they were considered extreme outliers. the equations for line of best fit and basic wind speed for the 40 stations are given in table 3 and were calculated using the gumbel and gringorten methods for a 50-year return period based on the monthly maximum hourly mean wind speed. it is clear that in all cases, gumbel’s method is conservative in predicting slightly larger values of basic wind speeds. the maximum and minimum differences between the two method values were 2.97% and 0.31%, respectively. table 4 shows the equations for line of best fit and basic wind speed values for the 29 longterm stations using the gumbel and gringorten methods for a 50-year return period based on the annual maxima of the monthly maximum 68 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami table 1. stations’ years of records and elevations (names of stations are as provided by directorate general of civil aviation and meteorology). station years of records record type elevation above mean see level (m) adam 16 long term 286 adam airport 4 short term 328 adam-research 4 short term 250 al amrat 4 short term 105 al khaboura 1 not considered 35 al mudhaibi 4 short term 378 bahla 16 long term 592 bidiyah 3 4 short term 316 buraimi 28 long term 299 buraimi new 3 not considered 372 diba 14 long term 10 duqum 11 long term 28 duqum new 3 not considered 102 fahud 22 long term 170 haima 3 not considered 146 ibra 16 long term 469.2 ibra new 1 not considered 528 ibri 14 long term 323 jabal nus 6 short term 706 jabalshamas 11 long term 2764 joba 11 long term 34 khasab airport 10 long term 29 khasab port 28 long term 4 manah 4 short term 345 mahdha 1 not considered 15 majis 31 long term 2 majis new 3 not considered 0 marmul 29 long term 269 masirah 28 long term 19 mina salalah 17 long term 25 mina sabouy 1 13 long term 3 muscat airport 37 long term 8.4 muscat new 1 not considered 12 new bahala 1 not considered 12 new samail 1 not considered 417 nizwa 15 long term 462 qairoon 15 long term 881 69 a basic wind speed for oman qalhat 14 long term 11 qarnalam 11 not considered 139 qumaira 4 short term 633 rasalhad 9 long term 43 rustaq 16 long term 322 saham 1 not considered 24 saiq 26 long term 1986 saiq new 11 not considered 1995 salalah 28 long term 23 samail 16 long term 417 sunaynah 5 short term 257 sur 29 long term 13 suwaiq 4 short term 38 thumrait 28 long term 448 um zamaim 3 not considered 126 wave rider 0 not considered 0 yalooni 15 long term 156 yalooni airport "ajais" 2 not considered 172.3 table 2. outliers and extreme outliers of wind speed data for longand short-term stations. station no. of records outliers extreme outliers 1 adam 120 5 1 2 adam airport 89 none none 3 al amrat 45 1 none 4 al mudhaibi 39 none none 5 bahla 185 none none 6 bidiyah 44 none 1 7 buraimi 288 2 1 8 diba 153 1 none 9 duqum 115 3 none 10 duqum new 58 none none 11 fahud 178 4 none 12 ibra 182 1 none 13 ibri 139 none none 14 jabal nus 59 1 none 15 jabal shams 118 5 none 16 joba 123 none none 17 khasab airport 105 none none 18 khasab port 246 2 none 19 majis 365 5 1 20 marmul 261 4 none 70 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami 21 masirah 335 5 3 22 mina salalah 182 6 1 23 mina sa bouy 1 124 1 none 24 muscat airport 442 15 8 25 nizwa 174 2 none 26 qairoon 168 4 none 27 qalhat 166 1 none 28 qaranalam 93 1 none 29 qumaira 37 1 none 30 ras al haad 93 none 2 31 rustaq 184 7 2 32 saiq 311 2 4 33 saiq new 111 3 2 34 salalah 335 13 1 35 samail 187 5 1 36 sunaynah 50 1 none 37 sur 331 2 none 38 suwaiq 45 1 none 39 thumrait 336 none none 40 yalooni 120 3 none figure 2. monthly maximum wind speed for saiq (1988 – 2013). 71 a basic wind speed for oman hourly mean wind speed. again, gumbel’s method predicted slightly larger values for basic wind speed. the maximum and minimum differences between the two method values were 7.58% and 1.85%, respectively. gumbel’s method was considered for further assessment and a comparison between the annual maxima and the monthly maximum basic wind speeds of the 29 long-term stations. due to a lack of long-term records in some regions of oman, the monthly maximum basic wind speed from the short-term stations was used to predict the long-term basic wind speed in those areas using eqn. 5. vb(annual) = 0.76 + 1.21vb(monthly) (5) r2 = 0.665 where, vb(monthly) = basic wind speed based on monthly maximum hourly-mean (m/s). vb (annual) = predicted annual basic wind speed (m/s). in the coefficient of correlation for eqn. 5, r2 (0.665) shows an acceptable degree of correlation. furthermore, analysis of variance (anova) for eqn. 5 resulted in a satisfactory hypothesis with p value = 0.000 for the given fvalue (53.59) as shown in table 5. therefore, equation 5 can be used as a conversion factor in the 11 short-term stations. table 6 shows the 50-year return period for predicted basic wind speed of the 29 long-term stations based on actual annual maxima of the monthly maximum hourly-mean wind speed as well as for the 11 short-term stations using eqn. 5.return periods for 100 years or more (implying a low-risk level) may have to be extrapolated for exceptionally important structures, such as nuclear power reactors and satellite communication towers. extrapolations of basic wind speed to higher return periods of more than 50 years for ultimate limit state design have not been carried out in this research due to the limitation of historical records, which are generally less than 37 years old. 4. basic wind speed map the basic wind speed values contained in table 6 were used to develop a basic wind speed map of oman using the largest reading in each nearby groups of values as shown in fig. 3. as expected, the coastal regions of oman recorded basic wind speed values higher than the interior regions due to the presence of the al hajer mountains and the reduction of wind speed over the land. this is more pronounced in the musandam region in oman’s far north where the basic wind speed dropped from 21 meters/second (m/s) in diba to 16 m/s in khasab. the basic wind speed values ranged from 16 m/s in khasab, which is surrounded by tall mountains, to 31 m/s in masirah, which borders the indian ocean. it is assumed that the basic wind speed at any location in the country will fall within the range of 16–31 m/s. this implies that for areas in the interior,, basic wind speeds that do not exactly fit in any given value(fig. 3) should be assigned to the nearest higher level value due to similarity in topography and the absence of wind barriers such as mountains. it is worth mentioning that this research has developed basic wind speeds based on hourlymean raw data which is in line with the bs63992 procedure, superseding the previous method of cp3: chapter 5-part 2, which used the threesecond gust which as is (al maawali et al. 2008). 5. conclusion and recommendations 5.1 conclusion hourly mean wind speed records were obtained from 55 metrological stations in oman, and 40 stations with at least four continuous years of records were used in the calculation of basic wind speeds. the monthly maximum and the annual maxima of the hourly mean were used in the analysis using gumbel and gringorten’s methods. it was found that gumbel’s method predicted slightly larger values of basic wind speed than gringorten’s. the maximum and minimum differences between the two method values were 2.97 and 0.31%, respectively, in the case of monthly maximum hourly mean and 7.58 and 1.85%, 72 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami table 3. linear best-fit equations for the 40 stations (monthly maximum hourly-mean). station gumbel method gringorten method gumbel method gringorten method % difference adam 1.3746x+8.7355 1.3396x+8.7451 14.11 13.98 0.91 adam airport 2.0007x + 10.832 1.8699x + 10.864 18.65 18.17 2.66 al amrat 0.8782x + 6.5392 0.8295x + 6.5485 9.97 9.79 1.87 al mudhaibi 0.7689x + 6.5824 0.7467x + 6.5899 9.59 9.51 0.83 bahla 1.2814x + 6.4704 1.244x + 6.4818 11.48 11.35 1.18 bidiyah 1.8098x + 8.8109 1.7025x + 8.8335 15.89 15.49 2.55 buraimi 1.4081x + 8.3545 1.3784x + 8.3637 13.86 13.75 0.79 diba 1.8632x + 8.1027 1.8033x + 8.1203 15.39 15.17 1.43 duqum 2.6691x + 10.336 2.567x + 10.365 20.77 20.40 1.81 duqum new 3.0118x + 11.723 2.8262x + 11.775 23.50 22.82 2.97 fahud 2.062x + 11.224 2.002x + 11.242 19.29 19.07 1.13 ibra 1.4357x + 7.3221 1.3959x + 7.3334 12.94 12.79 1.14 ibri 1.4215x + 10.662 1.3705x + 10.677 16.22 16.03 1.18 jabal nus 2.4533x + 11.995 2.2907x + 12.044 21.59 21.00 2.80 jabal shams 2.6081x + 12.833 2.5159x + 12.857 23.03 22.69 1.52 joba 1.5454x + 11.19 1.485x + 11.208 17.23 17.01 1.32 khasab airport 0.8294x + 8.1281 0.7926x + 8.1394 11.37 11.24 1.19 khasab port 1.3584x + 9.1348 1.3261x + 9.1447 14.45 14.33 0.80 majis 1.7366x + 7.9534 1.7104x + 7.9604 14.74 14.65 0.65 marmul 2.2194x + 11.472 2.1729x + 11.486 20.15 19.98 0.87 masirah 2.0096x + 10.386 1.9926x + 10.391 18.24 18.18 0.31 mina salalah 2.4491x + 6.5257 2.3894x + 6.5406 16.11 15.89 1.38 minasabouy 1 1.3665x + 7.4038 1.3165x + 7.4181 12.75 12.57 1.44 muscat airport 2.743x + 8.7434 2.7187x + 8.7464 19.47 19.38 0.49 nizwa 1.6017x + 8.4896 1.5528x + 8.5046 14.75 14.58 1.21 qairoon 2.6519x + 11.215 2.579x + 11.235 21.59 21.32 1.25 qalhat 1.9233x + 10.277 1.8677x + 10.293 17.80 17.60 1.14 qumaira 1.7374x + 11.369 1.6592x + 11.392 17.80 17.60 1.14 qaranaalam 1.7374x + 11.369 1.7374x + 11.369 18.17 17.88 1.58 ras al haad 1.5294x + 11.476 1.4585x + 11.497 17.46 17.20 1.49 rustaq 1.5368x + 5.5871 1.501x + 5.5956 11.60 11.47 1.15 saiq 1.5533x + 9.1184 1.5212x + 9.1286 15.19 15.08 0.77 saiq new 1.7909x + 8.8229 1.7205x + 8.8426 15.83 15.57 1.64 salalah 1.8437x + 8.224 1.8128x + 8.2326 15.44 15.32 0.73 samail 1.1133x + 6.9427 1.0842x + 6.9508 11.30 11.20 0.87 sunaynah 1.5861x + 11.034 1.4863x + 11.06 17.24 16.87 2.16 sur 2.4025x + 12.569 2.3549x + 12.584 21.97 21.80 0.79 suwaiq 0.9255x + 7.9023 0.8663x + 7.9165 11.52 11.31 1.92 thumrait 4.4675x + 22.839 4.374x + 22.87 20.72 20.50 1.08 yalooni 1x + 12.694 0.8671x + 12.728 15.68 15.47 1.34 73 a basic wind speed for oman table 4. linear best-fit equations for the 29 long-term stations (annual maxima of the monthly maximum hourly-mean). station gumbel method gringorten method gumbel method gringorten method % difference adam y = 1.715x + 11.59 y = 1.595x + 11.63 18.30 17.87 2.40 bahla y = 0.848x + 9.460 y = 0.748x + 9.483 12.78 12.41 2.97 buraimi y = 1.201x + 11.38 y = 1.092x + 11.41 16.08 15.68 2.53 diba y = 1.624x + 12.19 y = 1.421x + 12.24 18.54 17.80 4.18 duqum y = 3.512x + 13.74 y = 2.981x + 13.88 27.48 25.54 7.58 fahud y = 1.413x + 15.79 y = 1.289x + 15.82 21.32 20.86 2.18 ibra y = 1.325x + 10.46 y = 1.180x + 10.49 15.64 15.11 3.56 ibri y = 1.542x + 12.89 y = 1.331x + 12.94 18.92 18.15 4.27 jabal shams y = 2.967x + 18.12 y = 2.513x + 18.24 29.73 28.07 5.90 joba y = 0.918x + 14.63 y = 0.780x + 14.66 18.22 17.71 2.88 khasab airport y = 0.644x + 9.708 y = 0.541x + 9.735 12.23 11.85 3.17 khasab port y = 0.904x + 12.29 y = 0.825x + 12.30 15.83 15.53 1.93 majis y = 2.378x + 11.39 y = 2.190x + 11.43 20.69 20.00 3.48 marmul y = 2.311x + 15.28 y = 2.124x + 15.32 24.32 23.63 2.93 masirah y = 4.100x + 15.05 y = 3.789x + 15.11 31.09 29.93 3.86 mina salalah y = 2.567x + 12. y = 2.310x + 12.07 22.06 21.11 4.53 mina sabouy 1 y = 1.345x + 10.16 y = 1.160x + 10 15.42 14.74 4.64 muscat airport y = 4.555x + 12.08 y = 4.268x + 12.13 29.90 28.83 3.72 nizwa y = 1.091x + 12.17 y = 0.959x + 12.20 16.44 15.95 3.05 qairoon y = 2.209x + 13.98 y = 1.925x + 14.05 22.62 21.58 4.82 qalhat y = 1.614x + 15.22 y = 1.420x + 15.27 21.53 20.83 3.40 ras al haad y = 2.898x + 14.62 y = 2.486x + 14.70 25.96 24.43 6.27 rustaq y = 1.672x + 9.103 y = 1.503x + 9.134 15.64 15.01 4.20 saiq y = 2.509x + 13.54 y = 2.286x + 13.59 23.36 22.53 3.65 salalah y = 1.909x + 13.12 y = 1.743x + 13.16 20.59 19.98 3.05 samail y = 2.083x + 9.659 y = 1.866x + 9.701 17.81 17.00 4.75 sur y = 1.848x + 17.59 y = 1.685x + 17.63 24.82 24.22 2.47 thumrait y = 1.741x + 17.17 y = 1.627x + 17.18 23.98 23.54 1.85 yalooni y = 1x + 12.69 y = 0.867x + 12.72 16.60 16.11 3.04 table 5. analysis of variance for equation 5. source df ss ms f p-value regression 1 481.30 481.30 53.59 0.000 regression error 27 242.47 8.98 total 28 723.77 df = the degrees of freedom in the source ss = the sum of squares due to the source ms = the mean sum of squares due to the source f = calculated f value (ms regression/ ms errors) p-value = the probability of obtaining a test statistic at least as extreme as the one that was actually observed, assuming that the null hypothesis is true. 74 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami table 6. basic wind speed for 50-year return period based on annual maxima of monthly maximum hourly-mean. station basic wind speed 50-year return period adam 18.30 adam airport 23.33* al amrat 12.82* al mudhaibi 12.36* bahla 12.78 bidiyah 19.99* buraimi 16.08 diba 18.54 duqum 27.48 duqum new 29.20* fahud 21.32 ibra 15.64 ibri 18.92 jabal nus 26.88* jabal shams 29.73 joba 18.22 khasab airport 12.23 khasab port 15.83 majis 20.69 marmul 24.32 masirah 31.09 mina salalah 22.06 minasabouy 1 15.42 muscat airport 29.90 nizwa 16.44 qairoon 22.62 qalhat 21.53 qumaira 22.30* qaranaalam 22.75* ras al haad 25.96 rustaq 15.64 saiq 23.36 saiq new 19.91* salalah 20.59 samail 17.81 sunaynah 21.62* sur 24.82 suwaiq 14.70* thumrait 23.98 yalooni 16.60 *predicted using equation 5 75 a basic wind speed for oman figure 3. basic wind speed map. 76 a.s. alnuaimi, m.a. mohsin and k.h. al-riyami respectively, in the case of annual maxima. due to the lack of long-term records in some regions of oman, the monthly maximum basic wind speed from the short-term stations was used to predict long-term basic wind speed. accordingly, a basic wind speed map was developed for the 50-year return period using the basic wind speed from 29 stations with at least nine continuous years of records based on actual annual maxima as well as predicted annual maxima basic wind speed for 11 stations with 4–8 continuous years of records. the basic wind speed values ranged from 16 m/s to 31m/s. 5.2 recommendations • the basic wind speed map developed in this research is recommended for use as a guide for structural design in oman. • for future research, results from different gulf cooperative council (gcc) countries and yemen are compiled and a regional basic wind speed map is developed. • further analysis of future wind speed data need to be carried out every five years. • special analysis of recent odd readings, such as those that resulted from cyclones gonu and phet, should be made if such events recur frequently enough to establish records. acknowledgment the authors are indebted to the directorate general of civil aviation and meteorology under the ministry of transportation and telecommunication for the provision of the data used in this research. the data included station names, locations and hourly mean wind speeds. the concerned staff’s cooperation is gratefully acknowledged and appreciated. references abohemda mb, al shebani mm (2010), wind load characteristics in libya, world academy of science, engineering and technology 63:240-243. akosy h, toprak zf, aytek a, unal ne (2004), stochastic generation of hourly mean wind speed data. renewable energy, elsevier 29:2111-2131. al maawli ss, majid ta, yahya as (2008), determination of basic wind speed for building structures in oman. proceedings of international conference on construction and building technology, iccbt’08, f(22):235-244. american society of civil engineers (asce) (1991), guidelines for electrical transmission line structural loadings, 1801 alexander bell drive, reston, virginia 20191, usa. an y, pandy m (2005), a comparison of methods of extreme wind speed estimation. journal wind engineering and industrial aerodynamics, elsevier 93:535–545. bs 6399-2: (1997), loading for buildings part 2. code of practice for wind loads. british standard institution, london. choi e, tanurdjaja a (2002), extreme wind studies in singapore, an area with mixed weather system. journal wind engineering and industrial aerodynamics, elsevier 90:1611–1630. dorvlo ass (2002), estimating wind speed distribution. energy conversion and management. pergamon 43:311–2318. dyrbye, c, hansen so (1999), wind loads on structures. john wiley & sons. gibbs t, browne h, rocheford b (1981), code of practice for wind loads for structural design. bns-cp28, sponsored by the organization of american states, the national council for science and technology and the barbados association of professional engineers. grigoriu m (1984), estimation of extreme winds from short records. journal of structural engineering asce 110(7):1467-1484. gringorten i (1963), a plotting rule for extreme probability paper. journal of geophysical research 68(3):813-814. gumbel ej (1958), statistics of extreme, colombia university press. new york, nyc, usa. harris ri (1996), gumbel re-visited – a new look at extreme value statistics applied to wind speeds. journal wind engineering and industrial aerodynamics 59: 1-22. holmes j (2001), wind loading of structures, spon press. kasperski m (2002), a new wind zone map of germany.journal wind engineering and industrial aerodynamics 90:1271–1287. kishor cm, delahay j (2004), guide to the use of the wind load provisions of asce 7-02, 140, 77 a basic wind speed for oman asce press. kramer c, gerhardt hj (1988), advances in civil engineering, part 1:8, elsevier science publishing company, 665 avenue of the americas, new york, usa. lakshmanan n, gomathinayagam s, harikrishna p, abraham a, ganapathi sc (2009), basic wind speed map of india with long-term hourly wind data. currant science 96(7):911-922. minimum design loads for buildings and other structures asce/sei-7-10 (2010), american society of civil engineers and the structural engineering institute. asce standard, 1801 alexander bell drive, reston, virginia 20191, usa. ministry of transportation and tele communication (2014), directorate general of civil aviation and meteorology – oman, annual report. sahin a (2003), hourly wind velocity exceedance maps of turkey. energy conversion and management 44:549-557. simiu e (2009), wind loading codification in the americas. fundamentals for a renewal, 11th americas conference on wind engineering, san juan, pr, usa. simiu e, filliben jj, shaver jr (1982), short-term records and extreme wind speeds. journal of structural division asce 108(11):2571-2577. simiu e, heckert n, filliben j, johnson s (2001), extreme wind load estimates based on the gumbel distribution of dynamic pressures: an assessment. structural safety 23:221-229. 78 civl080903.qxd 1. introduction inverse problems often occur in many branches of engineering and mathematics where the values of some physical model parameter(s) must be deduced from observed data. system identification comes under the category of inverse problems. it is the process of determining the parameters of a system based on the observed input and _______________________________________ *corresponding author’s email: skris@iitm.ac.in output (i/o) of the system. in structural engineering, sys tem identification is used to determine unknown parameters such as mass, stiffness and damping properties of a structure. structural identification methods can be classified under various categories, eg. frequency domain and time domain, parametric and nonparametric methods. inverse problems can be solved using classical and nonclassical methods. the journal of engineering research vol. 6, no. 2 (2009) 64-77 structural system identification in the time domain using evolutionary and behaviorally inspired algorithms and their hybrids s. sandesha , abhishek kumar sahua and k. shankar*b a research scholars, department of mechanical engineering, indian institute of technology, madras, chennai 600 036, india *b department of mechanical engineering, indian institute of technology, madras, chennai 60 036, india received 3 september 2008; accepted 1 march 2009 abstract : in this study, parametric identification of structural properties such as stiffness and damping is carried out using acceleration responses in the time domain. the process consists of minimizing the difference between the experimentally measured and theoretically predicted acceleration responses. the unknown parameters of certain numerical models, viz., a ten degree of freedom lumped mass system, a nine member truss and a non-uniform simply supported beam are thus identified. evolutionary and behaviorally inspired optimization algorithms are used for minimization operations. the performance of their hybrid combinations is also investigated. genetic algorithm (ga) is a well known evolutionary algorithm used in system identification. recently particle swarm optimization (pso), a behaviorally inspired algorithm, has emerged as a strong contender to ga in speed and accuracy. the discrete ant colony optimization (aco) method is yet another behaviorally inspired method studied here. the performance (speed and accuracy) of each algorithm alone and in their hybrid combinations such as ga with pso, aco with pso and aco with ga are extensively investigated using the numerical examples with effects of noise added for realism. the ga+pso hybrid algorithm was found to give the best performance in speed and accuracy compared to all others. the next best in performance was pure pso followed by pure ga. aco performed poorly in all the cases. keywords: inverse problem, system identification, genetic algorithm, ant colony optimization, particle swarm optimization é¡jéæ«ég ™e é«cƒ∏°s ¬ª¡∏e h ájqƒ£j äé«erqgƒn ∫骩à°séh øeõdg π≤m ‘ »fé°ûf’g ωé¶ædg ájƒg ~j~– qéµfé°t .∑ h ƒgé°s qéeƒc ∂°ûæ¡hg h ¢t~fé°s .¢s áá°°uuóóÿÿgg¢ü«∏≤j øe á«∏ª©dg ¿ƒµàj .øeõdg π≤m ‘ π«é©àdg äéhééà°sg ∫骩à°séh √ò«øæj ” ~b úgƒàdg h ihé°ù≤dg πãe á«fé°ûf’g ¢ugƒî∏d ájƒ¡dg ¢ü«î°ûj ” á°sgq~dg √òg ‘ : ¿ƒ∏ªl ‘ ájô◊g øe äélqo ô°ûy øe ¬äjéædg ¬∏àµdg agreg ᫪bôdg êpéªædg ¢ ©h ‘ áeƒ∏©ÿg òz ±gôw’g ¿g .é«ñjôœ á°sé≤ÿg ∂∏j h éjô¶f á°sé≤ÿg ¬∏é©ÿg äéhééà°s’g úh ¥ôødg áñ«cîdg agog á°sgqo ∂dòc ” .äé«∏ª©dg ¢ü«∏≤àd é«cƒ∏°s ¬ª¡∏eh ájqƒ£j äé«erqgƒn ∫骩à°sg ” .é¡°ü«î°ûj ” ~b ᣰùñe ᨫ°üh ∫ƒªfih º¶àæe òz πµ°ûh πªfih aé° yg ¬©°ùj øe ) á«kgqƒdg ä骫jqézƒ∏dg .¬æé¡ÿgga) äé«fõ÷g ~«°û– ¿eg .ωé¶ædg ¢ü«î°ûj ‘ é¡d骩à°sg ” ~bh g~«l áahô©e ájqƒ£j ä髪kqéz ƒd »g (pso¬ª¡∏e äé«erqgƒn »gh ( ) ¤g …ƒb …~ëàªc éãj~m ärôh ~b ±ô©àdgga) πªædg iôª©à°ùe á«∏yéa ¬≤jôw .áb~dgh áyô°ùdg ‘ (acoago’g .éæg é¡à°sgqo â“ ~bh º¡∏e ±ô°üj ägp á≤jôw ∂dòc »g ( ) πãe áñcôÿg é¡àæ«ég h √oôøæe ᫪kqézƒd πµd áb~dgh áyô°ùdg hacoâ“ ~b »àdg aéxƒ¶dg ägòkéj ™e ᫪bq êpé‰ ∫骩à°séh áøãµe √qƒ°üh é¡ãëh ” ~b áæé¡e ᫪kqéz ƒd ( ) øé¡ÿg ºkqézƒdg ¿g .á«©bgƒdg πlg øe é¡àaé°vgpso +ga) ƒg ago’g ‘ á«∏j …òdg äéjôn’éh áfqé≤e áb~dg h áyô°ùdg ‘ agog π° ag »£©j ¬féh ~lh ~b (pso¢üdéÿg ( ) ¬©ñàjhga) agog ¿éc ¢üdéÿg (aco .ä’é◊g ™«ªl ‘ ∞«©°v ( áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .äéäjõ÷g ~«°û– á«∏yéa,πªædg iôª©à°ùe á«∏yéa ,»kgqh ”qézƒd ,ωé¶ædg ¢ü«î°ûj ,áhƒ∏≤ÿg ádé°ùÿg : 65 the journal of engineering research vol. 6, no. 2 (2009) 64-77 1.1 classical methods the least square method is perhaps the simplest classical methods for structural identification. it estimates the unknown parameters of structural system by minimizing the sum of squared errors between the predicted outputs and the measured outputs. many other classical methods such as maximum likelihood method, instrument variable method and extended kalman filter method have been developed. koh et al. (1991) used the kalman filter method for identification of stiffness and damping coefficients from measurement of dynamic responses in the time domain. detailed information about classical methods and non-linear identification are presented in the review paper by (gaetan et al. 2006). some classical optimization methods such as gradient methods (typically variations of cauchy, newton and levenburg-marquardt methods) need a good initial guess of the parameters to converge fast and accurately. some studies have been conducted to supply good initial values using non-traditional methods such as ga which would sample the search domain efficiently. thus certain classical methods such as levensburg-marquardt were hybridized with ga to identify system parameters with less computational time and improved accuracy (kishore et al. 2007 and friswell et al. 1998). 1.2 evolutionary algorithms and behaviourally inspired algorithms evolutionary algorithms are robust optimization algorithms based on the heuristic concept of natural selection and genetic operations. optimization algorithms find the minima or maxima of functions in a given function domain. over the past few decades the area of genetic algorithm (ga) has been widely developed and applied for structural identification. unlike classical methods, there is no requirement for calculation of gradients and second derivatives which frequently lead to numerical instability. ga applications in system identification such the identification of elastic properties of composite plates from dynamic test data is presented in (jesiel et al. 1999; chakraborthy and mukhopadhyaya 2002). recently efforts have been made to alter the architecture of ga and to incorporate local search algorithms to further improve its performance, see for example see (koh et al. 2003 and perry et al. 2004). behaviorally inspired optimization algorithms have been developed out of attempts to model the natural behavior of a flock of birds or a colony of ants. social insects have diverse foraging systems that reflect the enormous range of ecological conditions in which they operate. ant colony optimization (aco) and particle swarm optimization (pso) techniques are inspired by real ant colonies and flock of birds respectively. 2. ant colony optimization (aco) ant colonies continually adjust foraging effort to changing conditions. individuals use local information in foraging decisions and colonies can tune foraging effort to the location, quality and abundance of food. aco is a general purpose combinatorial optimization (co) technique. co optimization is one of the youngest and most active areas of discrete mathematics. as the name suggest co deals with finding optimal combination of available problem components. hence, it is required that the problem is portioned into a finite set of components and co algorithm attempts to find the optimal combination. aco, a meta-heuristic which is based on the ant system introduced in the early nineties by (dorigo 1992 and dorigo et al. 1996). it has been used to solve a variety of combinatorial optimization problems, eg. the vehicle routing problem by (bullnheimer et al. 1999) the traveling salesman problem by (dorigo et al. 1997) and the industrial layout problems studied by (hami et al. 2007; corry and kozan 2004 and rajamani and adil 1996). abbospour et al. (2001) proposed the aco-im (inverse modeling) method for estimating soil hydraulic parameters. they compared results of aco-im with levenbergmarquardt (lm) method and finally complimented aco and lm to obtain final solution which was better than pure aco itself. some extensions and variants of aco are presented in the review paper by (blum 2005). recently aco has been applied to structural optimization and topology optimization. the goal of truss optimization is to optimally utilize the geometry and material of the proposed design elements to result in the lightest structure satisfying all the design, manufacturing constraints and other physical constraints. since aco is a discrete co problem, several studies using it to optimize steel trusses for minimum weight using discrete cross sectional areas and other parameters were conducted, see (camp et al. 2004; camp et al. 2005 and serra et al. 2006). these studies mapped the length of the tour of the ant to the weight of the truss and represented the volume of the element as virtual paths in aco model for truss design. a biomechanical application of pso is presented by (jaco et al. 2005) where ankle joint model parameters were identified and compared with gradient methods such as sequential quadratic programming, quasi-newton algorithm and ga. the pso algorithm was found insensitive to design variable scaling and gradient methods are highly sensitive. the parameters of lorenz chaotic system are estimated using pso in reference (qie et al. 2007). it was found that pso converges to exact value with high population size and more effective than ga. a 10-dof structural dynamic model was identified using frequency response function by ga and pso in (mouser and dunn 2005), without using hybrids. pso was found more likely than ga to produce better parametric models. a pso-ga hybrid was used to successfully identify faults in the water supply system in japanese cities by (furuta and yasui 2005). however only a single case is studied, and comparisons with pure ga and pso are not shown for that case study, but only for benchmark equations. recently pso has been successfully applied in many research and application areas such as pattern recognition (peng-yeng et al. 2006), scheduling (binghui et al. (2007)), layout 66 the journal of engineering research vol. 6, no. 2 (2009) 64-77 optimization (zhang et al. (2007), and the design optimization of unmanned aerial vehicle (uav) with respect to flight loads (hu et al. 2004). 2.1 the aco algorithm the inspiration for aco is the foraging behavior of ants, particularly how they find the shortest paths between food sources and their nest. while searching for food, ants initially explore the area surrounding their nest randomly. when ants move they leave a chemical pheromone trail on the ground. when choosing their way, ants tend to favor paths marked by strong pheromone concentrations. as soon as an ant finds a food source, it evaluates the quantity and the quality of the food and carries some of it back to the nest. during the return trip, the quantity of pheromone that an ant leaves on the ground may depend on the quantity and quality of the food. the pheromone trails will guide other ants to the food source. aco model consists of graph g = (v, e) (blum et al. (2005)) where v consists of two nodes, namely vs (nest) and vd (food source). e consists of two links e1 and e2 between vs and vd and corresponding length are l1 and l2 as shown in fig. 1. path e1 represents the short path between vs and vd , and e2 represents the long path. real ants deposit pheromone on the paths on which they move. thus, the chemical pheromone trails are modeled as follows. (1) where pi is the probability of an ant choosing the ith path, τi is artificial pheromone value for each of the two links ei , i = 1, 2. obviously, if τ1 > τ2, the probability of choosing e1 is higher, and vice versa. for returning from vd to vs , an ant uses the same path as it chose to reach vd and it changes the artificial pheromone value associated to the used edge. after all the ants have returned to the nest, the pheromone information is updated using eq. (2) (2) wher p ε (0,1] is evaporation rate, q is a constant, lk is length of the path traversed kth ant and na is number ants in the colony. the aim of pheromone update is to increase the pheromone value associated with good or promising paths. pheromone evaporation is needed to avoid too rapid convergence of the algorithm. implementing the discrete aco algorithm for a continuous problem requires substantial modifications. the physical problems need to be represented in a graphical from as shown in fig. 2 and fig 3. the elemental stiffness value is divided in to 'np' different values randomly with in the specified range. each of these values can be represented as local path between the two levels and each of which leads to global path as shown in fig. 2b and 3b. as mentioned earlier, aco algorithm requires finite set of components, i.e. resolution of the local paths, so each stiffness value is divided in 'np' different paths (stiffness) in the given range. aco algorithm has to find shortest path out of npnv (nvnumber of parameters to be identified) available combinations. in real ant colony initially ants take random paths and deposit pheromone. as the iteration proceeds, all the ants in colony take the shortest path and there by increasing the intensity of pheromone trail on the shortest path. the main goal in structural identification is minimizing the objective function, which in turn decides the amount of pheromone deposition. generally aco minimizes the length of the lk in eq. 2, traversed by kth ant. in structural identification, aco minimizes the fitness value, ε eq. 8, which is analogous to length of the path lk in eq. 2 and the amount of pheromone deposited depends on the value of ε. 3. particle swarm optimization (pso) particle swarm optimization (pso) is a population based continuous optimization technique developed by (eberhart and kennedy 1995), inspired by social behavior of bird flocking or fish schooling. pso shares some similarities with evolutionary computation techniques such as genetic algorithms (ga). the system is initialized with a population of random solutions and searches for optima by updating generations. however, unlike ga, pso has no evolution operators such as crossover and mutation. in pso, the potential solutions, called particles, move through the problem space by following the current optimum particles. the basic pso algorithm consists of the velocity and position equation of the kth generation: (3) (4) i particle number k iteration number v velocity of ith particle x position of the ith particle/ present solution pi historically best position/solution found by ith particle g historically best position/solution found by all particles γ1,2 random number in the interval (0,1) applied to ith particle vs e1, l1 vd e2, l2 figure 1. the aco model 1 2 1, 2.i i p i τ τ τ = = + 1 where(1 ) k knak i i i i k q l τ ρ τ τ τ∑== − + ∆ ∆ = 1 2 ( 1) ( ) ( ( )) ( ( )) i i i i i i i v k v k p x k g x kγ γ+ = + − + − ( 1) ( ) ( 1) i i i x k x k v k+ = + + 67 the journal of engineering research vol. 6, no. 2 (2009) 64-77 the frequently used pso form includes an inertia term and acceleration constants, hence the velocity equation of pso algorithm is modified as: (5) ϕ inertia function α1,2 acceleration constants the inertia function is commonly either taken as constant or as a linearly decreasing function from 0.9 to 0.4. the acceleration constants are most commonly set to both equal to 2 as in (shi and eberhart 1998). the algorithm starts with initializing the i particles randomly with initial zero velocities. pso then searches for optima by updating the positions of particle through several generations. at every generation, the location of each particle is updated based on two "best" values defined as follows. the first is the historically best solution (fitness) achieved so far by all particles and stored as gbest. another "best" value is the historically best value obtained so far by the ith particle in the population and is called pbest. the position of each particle is updated by a quantity which reflects the difference between gbest and pbest (equation (4) and (5)) and eventually all the particles tend to converge to the global best (gbest) position. the superiority of pso over other comparable heuristic algorithms such as ga could be attributed to the explicit tracking and updating of gbest and pbest over the generations. 4. genetic algorithm (ga) genetic algorithms are optimization algorithms based on the mechanism of natural selection and survival of the fittest. over the past few decades ga has been widely developed and applied for structural identification. ga combines the explorative ability of large search domains as well as reasonable guided search to the global optima (michaelwicz 1994). ga creates an initial random sample within the specified domain of variables, called 'population'. it then ranks them in the order of fitness and conducts crossover operations from among a pool of 'parents' through the roulette wheel selection. parents having higher fitness have a greater probability of being selected for crossovers and their offsprings contribute to the next generation. these offspring have marginally better fitness than the parents and over many generations they attain the global maxima. ga can be programmed in the binary or continuous versions. here the continuous (decimal number) version is used to avoid the computationally intensive conversion from binary to decimal and vice-versa. it been indicated in a few studies that continuous ga is superior to binary ga in computational performance (see (michaelwicz 1994 and haupt 1994) ). it may be noted that unlike pso, ga does not explicitly keep track of the globally best solution (gbest) and particle best (pbest) solutions. 5. numerical examples and discussion for structural identification problems it is usually assumed that the mass of the structure is known and the identification is limited to structural stiffness and damping k1 knv m1 mnv knv-1 mnv-1 ants taking different paths local paths global paths figure 2 (a) multi-dof physical mode (b) aco-graphysical model (a) (b) (a) (b) figure 3 (a) physical model of truss (b) aco-model of the truss (for brevity only a few paths are shown 1 1 2 2 ( 1) ( ) ( ) [ ( ( ))] [ ( ( ))] i i i i i i i v k k v k p x k g x k ϕ α γ α γ + = + − + − 68 the journal of engineering research vol. 6, no. 2 (2009) 64-77 properties. it is assumed that the structure is excited by known forces and that the responses of the structure in terms of accelerations, are recorded at some given points. the dynamic equation of a structural system can be written as (6) where m, c and k are the mass matrix, damping matrix and stiffness matrix, respectively, u is the displacement vector and f is the force vector. proportional damping is adopted as follows (7) where α and β are two damping constants, which can be related to the modal damping ratio. in the examples which follow, experiments are simulated numerically using the known parameters of the structure ie. the forward analysis is carried out by numerically solving the dynamic equation using the parameter values (such as in tables 1 and 4). the acceleration responses in the time domain are obtained at desired places and they are treated as experimentally obtained values. noise may be added to them to be more realistic. the difference between the estimated acceleration response and measured acceleration response is used to compute the fitness value (objective function) as given by the eq. (8) given below, which has to be minimized (see reference koh et al. 1991). (8) the parameters used in ga, pso and aco are given in the table 2. they were applied to all the numerical examples. to compare ga and pso they were given the same population size (200) and number of generations (50). these values are taken from koh and shankar (2003) which dealt with identification of structural systems similar to this paper. the crossover rate (40%) and mutation rate (5%) for ga, and pso inertia function value (0.3) and acceleration constant (2) were selected from the best parameters recommended from standard literature on these algorithms (michaelwicz 1994; haupt 1994; shi and eberhart 1998). comparable aco parameters were however more difficult to establish since it is a discrete approach whereas pso and ga are continuous with their resolution decided by the precision of the smallest decimal number which is by default set to double precision in matlab. it is impractical to obtain a comparable precision in aco by discrete division of the search interval (ie. number of paths). thus the crucial aco parameters such as number of ants (400), number of paths (100) and number of iterations (100) were chosen after several trials as a compromise, which would converge in a reasonably comparable time as ga, with mean errors which are acceptable for system identification purposes. each algorithm has been run ten times separately to minimize the objective function, ε, for both impulse and random excitation with noise-free and noisy data and the final results shown is the average. 5.1 example 1: 10-dof lumped mass system in the numerical study, a 10-dof lumped mass system studied in (koh et al. 2003) and shown in fig. 2(a) is considered. the structural parameters are tabulated in the table 1. impulse and random excitation is applied at the 5th level. the impulse was given as an initial velocity of 10 m/s to the first mass, the displacements and velocities of all other dof set to zero. the impulse excitation was simulated by imposing equivalent initial velocity conditions obtained from impulse momentum relation using the method followed in (hanagud 1985). the gaussian random force was applied as with a maximum amplitude of 10 n. accelerations at alternate levels, ie. levels 2, 4, 6, 8, 10 are assumed to be available for the purpose of structural identification. referring to eq. 8, here m=5, and n the time steps are 500 in the range from 0 to 2 seconds. here the objective is to find out the unknown spring stiffnesses k1 to k10. in all the problems hereafter the search limits for unknown parameters are taken as -50% (lower limit) to +100% (upper limit) of the exact value. in ga and pso initial population/particles are generated within this specified range. in aco a matrix of size nv × np is generated within the specified parameter range. aco has to choose best combination of paths which minimize the objective function. the results for this system are shown in the fig. 4. ga and pso have identified the parameters more accurately as shown in fig. 4 (a) and (b). the accuracy of identified values by pso is more than ga which can be noticed in the almost constant plateau in fig. 4 (b) as compared to ( ) ( ) ( ) ( )mu t cu t ku t f t+ + = c m kα β= + 2 1 1 ( , ) ( , ) minimize: * m n mea est i j u i j u i j m n ε = = − = ∑ ∑ where ( , )meau i j and ( , )estu i j are, respectively the measured and estimated responses of ith measurement location at jth time step, m is the number of measurement location and n is the number of time steps. the “measured” responses are obtained from numerical simulation. the “estimated” responses are obtained from the mathematical model (eq. 6 and 7) where the optimization variables are t he now unknown stiffness properties and damping ratio. the values of the variables which minimize eq. 8 give the required structural properties. to excite higher modes for better identification the input forces are impulse and random excitations (the latte r by means of gaussian white noise). the dynamic equations are solved using newmark’s constant acceleration method. two seconds of acceleration response is computed with constant time step of 0.004s in 500 steps. to investigate the effect of noisy data on identification, the i/o time signals are artificially contaminated by zero -mean gaussian white noise. the noise level is defined as the ratio of standard deviation of the noise to the root -meansquare value of the uncontaminated time history; here it is taken as 10%. 69 the journal of engineering research vol. 6, no. 2 (2009) 64-77 ga (fig 4(a)) for stiffness levels 6-10. from fig. 5 and 6 it is clear that ga and pso performs better with presence of noise for both types of excitations (impulse and random). but aco fails to identify the parameters with noisy data. table 3 gives maximum and average errors (averaged over 10 runs) in identification of the stiffness parameters using impulse and random excitation. the maximum errors observed for the noisy case are 15.9% for aco, 9.12% for ga and 4.17% for pso (using random excitation). the mean errors are 12.34%, 3.47% and 2.88% respectively (using random excitation). the table shows that using impulse excitation gives lesser errors than the random type. in such as case pso produces a maximum error of only 3.02% with a mean error of 1.13%. thus it is clear from the fig. 5, fig. 6 and table 3, that the pso gives superior results as compared to the other two algorithms for both types of excitations and with noisy data from the point of view of maximum and mean errors as well as computational time which is similar for both load cases. 5.2 example 2: 11member planar truss in this example an 11-member planar truss is considered for identification of the axial rigidity (ea) of all the 11 (ie. number of variables, nv = 11) members. configuration and structural properties of the truss are given in fig. 3(a) and table 4, respectively. to implement aco, the physical truss has to be represented as a graphical model with axial rigidity (stiffness) of each member discretized into 100 units ie. this would correspond to 100 paths per member. fig. 3(b) shows schematically this division of each member into several paths. for clarity only 3 paths are shown per member. stiffness (kn/m) level 1-5 350 level 6-10 600 mass (kg) level 1-5 500 level 6-10 300 damping critical damping ratio for first 2 modes 5% table 1. structural properties of lumped mass system ga pso aco population size 200 no. of particles 200 no. of ants 400 no. of generations 50 no. of generations 50 no. of iterations 100 cross over rate 40% inertia function 0.3 evaporation rate 30% mutation rate 5% acceleration constants 2 no. of paths 100 constant q 1 table 2. parameters used in the algorithms impulse excitation random excitation algorithms with out noi se with 10% noise with out noise with 10% noise average computational time (sec) max mean max mean max mean max mean ga 3 1.77 4.7 1.90 5.4 2.26 9.12 3.47 95 pso 0.23 0.06 3.02 1.13 2.51 0.58 4.17 2.88 40 aco 4.55 1.45 15.15 7.68 4.39 2.09 15.91 12.34 140 table 3. absolute percentage error in identification of stiffness lumped mass system axial rigidity (mn) 300 cross section area (m 2) 0.0015 density (kg/m 3) 7800 critical damping for first 2 modes 5% table 4. structural properties of truss 70 the journal of engineering research vol. 6, no. 2 (2009) 64-77 (a) (b) (c) figure 4. identified values of the level stiffness v/s no. of runs with noise-free data for impulse excitation (a) ga (b) pso (c) aco (a) (b) (c) figure 5. identified values of the level stiffness v/s no. of runs with noise-free data for impulse excitation (a) ga (b) pso (c) aco 71 the journal of engineering research vol. 6, no. 2 (2009) 64-77 the truss is excited with impulse and random input at node 2 in y-direction and acceleration responses (using the same excitation and times as the previous case) are assumed to be available at nodes 1, 2, 3, 4 and 5 in ydirection. acceleration responses are simulated using properties given in table 4, and 10% noise added to the simulated responses. the truss is modeled using finite element method. the same set of algorithm parameters (table 2) are used in this example also for minimizing the objective function ε (eq. 8). in all the algorithms finite element model is updated till the error between simulated acceleration responses and estimated by the algorithms is minimized. figure 7 shows the typical percentage error in identification of axial rigidity of the truss members for impulse excitation with and without 10% noise. from fig. 7 it can be noticed that the error in identification of all the members is small in the case of pso and ga. but aco appears to oscillate about zero with significant variation giving maximum errors of about 20% 30%. the maximum and mean errors in identification by all algorithms for both excitations are tabulated in table 5. it is seen that when noise is added to the responses the impulse excitation gives better identification than random excitation. the maximum and mean errors and computational time of pso is much better than the others in the presence of noise. the computational time is roughly same for both load cases hence only a typical average value of time is shown to compare between the various algorithms. 5.3 example 3: simply supported beam in this example a simply supported beam, fig. 8, with linearly varying stiffness from support to the middle is considered for element wise stiffness identification (modulus of elasticity) in the time domain. cross section of the beam is square with an area of 9x10-4 m2 and density of 7800 kg/m3. the beam is modeled with 2 noded, 11 euler beam elements with two degrees of freedom per node, translation and rotation. the stiffness variation is 200 gpa at the support to 50 gpa at the center of the beam. impulse excitation is applied at node 3 in transverse direction and acceleration responses are assumed to be available at nodes 1, 4, 7 and 10 for identification. in identification, modulus of elasticity of all eleven elements are considered unknown and damping ratio of 5% is assumed to be known. here also the finite element model is updated as mentioned in the previous example to minimize the objective function given in eq. (8). results are represented in fig. 9 for impulse excitation only. again it is noted that the performance of pso is better than other algorithms. this is seen in the results where the mean error in identification is 5.8%, 3.1% and 15% for ga, pso and aco with noisy data, respectively. in all the examples thus presented it is found that aco cannot identify the parameters with the same degree of accuracy as pso and ga. so in order to combine the advantages of each of these algorithms, a hybrid approach is next investigated. (a) (b) (c) figure 6. identified values of the level stiffness v/s no. of runs with 10% noise data for random excitation (a) ga (b) pso (c) aco 72 the journal of engineering research vol. 6, no. 2 (2009) 64-77 5.4 the hybrid approach the results of hybrid approach are tabulated in the tables 7 to 10. table 7a shows the identification results with known damping for three cases a, b and c which represent the lumped mass, truss and beam models. the extensive data in table 7a is summarized in table 7b by averaging over the three cases. table 7b data can be studied from the point of view of a) average time of solution : from this point of view it is seen that ga+pso hybrid is clearly superior with only 39.7s, followed by pure pso with 66.7s, then aco+pso hybrid with 110s and pure ga at 143s. the worst time performer is pure aco which takes 219s. b) accuracy of solution: we examine the values for mean and maximum error in identification . here also ga+pso hybrid gives the smallest mean and maximum errors (0.72% and 2.47%) followed by pure pso (a) (b) figure 7. percentage error in identification of axial rigidities of the members with impulse excivation (a) with out noise (b) with 10% noise figure 8. simply supported beam table 5. absolute percentage error in identification of axial rigidity truss ga+pso aco+pso aco+ga algorithm parameters ga pso aco pso aco ga population size 100 ----50 no. of generations 10 ----50 no. of particles -50 -50 -- iterations -50 -50 -- no. of ants --200 -200 - iterations --50 -50 - table 6. algorithm parameters for hybrids 73 the journal of engineering research vol. 6, no. 2 (2009) 64-77 1 2 3 4 5 6 7 8 9 10 11 0 50 100 150 200 250 300 element number m od ul u s o f e la s tic ity actual ga pso aco (a) 1 2 3 4 5 6 7 8 9 10 11 0 50 100 150 200 250 300 element number m od ul u s o f e la s tic ity actual ga pso aco (b) figure 9. identified stiffness along the length of the beam for impulse excitation (a) without noise (b) with 10% noise case descripti on pso ga aco ga+pso aco+pso aco+ga case-a, ma x error 3.02 4.7 15.15 0.63 3.0 11.86 case-a, mean error 1.13 1.71 7.68 0.126 0.75 4.47 case-a, time in sec 36 90 137 26 77 89 case-b, ma x e rror .67 1.52 30.3 0.46 7.25 19.71 case-b, mean e rror .39 .7 12.26 0.25 4.65 7.0 case-b, time in sec 88 164 230 34 91 104 case-c, ma x e rror 7.45 12.66 46.06 6.31 10.09 40.13 case-c, mean e rror 2.7 5.8 12.5 1.78 3.45 11.35 case-c, time in sec 76 186 291 59 163 192 table 7a. percentage error of identified values of various cases with known damping and noise added. (case-a is the lumped mass; case-b is the truss and case-c is the beam) 74 the journal of engineering research vol. 6, no. 2 (2009) 64-77 (1.4% and 3.71%) and then pure ga (2.73% and 6.29%). the aco+pso hybrid has marginally less accuracy than pure ga although speed wise it is better. the worst accuracy is shown by aco. the discrete nature of aco and its lack of resolution compared to ga and pso are obvious handicaps. next, the same structural cases are studied, but damping is considered as an additional unknown. tables 8, 9 and 10 deal with these cases. in each table the values of all unknown parameters are presented in detail, in addition to mean error, maximum error and time taken. the accuracy of identification of the unknown damping ratio by the algorithms is also an important parameter considered here. in table. 8 it is seen that the ga+pso hybrid is the fastest (ie. with least time of 23s) and next is pso (34s) followed closely by aco+pso (78s) and ga (80s) and the worst is pure aco(145s). the time trend is the same as in table. 7. damping parameter has been identified most accurately by ga+pso (-1.21% error) and pure pso (-1.05%) followed by aco+pso and ga. pure aco gives the worst value here(16%). regarding the overall accuracy (including damping as well as stiffness) it is seen that ga+pso has identified fairly accurately with a small mean and maximum error of 1.47% and 4.67% followed by aco+pso (1.7%, 4.89%) , pso (1.85%, 7.2%) and ga (2.01% and 5.41%). thus there are some small deviations in the trend of accuracy when compared to table 7. in table. 9 it is also seen that ga+pso is the fastest (60s) followed by pso (82s) then aco+pso (105s) and ga (107s). accuracy-wise the best performer is ga+pso hybrid (mean 0.44% and maximum error 1.05%) followed by pso(2.9%, 5.68%), then ga(3.9%, 6.62%). the accuracy of aco+pso in this case is quite insufficient compared to ga(11.6%, 23.77%). as expected pure aco performed worst in time and accuracy factors. in table.10 the same general trends are observed with respect to time .i.e., ga+pso is the fastest (104s) followed by pso(138s) then aco+pso (175s) . the accuracy of ga+pso is the best (mean error 2.3%, maximum error 6.39%), then pso(4.8%, 11.07%), followed by aco+pso (7.6%,21.04%) and ga(11.6%, 28.35%). pure aco performed worst in both time and accuracy factors. summary of observations: the general trend shown in tables 7 to 10 is as follows. from the point of view of speed (minimum computation time) ga+pso is the best, followed by pure pso, then aco+pso and pure ga. in some cases ga is nearly comparable to aco+pso in speed. pure aco is the worst performer. from the point of aver age over case a, b and c pso ga aco ga+pso aco+pso aco+ga max error 3.71 6.29 30.5 2.47 6.78 23.9 mean error 1.4 2.73 10.8 0.72 2.95 7.6 time in sec 66.7 143 219.3 39.7 110.3 128.3 table 7b. average vbalues obtained from case a, b and c exac t (kn/ m) pso ga aco ga+pso aco+pso aco+ga 350 2.81 2.26 -5.65 0.20 0.87 -0.21 350 1.54 1.10 -1.63 1.52 1.2 1.91 350 0.32 0.49 4.70 0.23 0.59 5.18 350 -2.99 -2.28 10.1 -2.81 -2.72 -3.71 350 0.86 -0.41 -3.90 1.23 -0.55 -0.98 600 -0.21 0.86 7.58 0.25 -1.62 3.27 600 -1.46 -5.41 6.42 -0.53 -2.89 -0.86 600 -0.45 1.48 -9.20 -0.81 -0.5 19.22 600 7.21 -0.25 15.90 4.69 -0.57 14.71 600 -1.41 3.47 19.7 -2.78 4.87 -10.11 da mping 0.05 -1.05 4.05 16 -1.21 -2.46 5.4 abs. ma x. error 7.21 5.41 19.7 4.69 4.89 19.22 abs. mean error 1.85 2.01 9.14 1.47 1.7 5.96 co mputational time in seconds 34 80 145 23 78 90 table 8. percentage error in identified value of 10-dof lumped mass system with unknown damping and noise 75 the journal of engineering research vol. 6, no. 2 (2009) 64-77 view of accuracy, again ga+pso is the best performer followed by pure pso. then follows ga and aco+pso their relative superiority in accuracy varies from case to case. even in these circumstances pure ga would still be preferred as it is simpler to program than the hybrid. the aco+ga hybrid and pure aco was not found promising in speed or accuracy. the observations from this comprehensive numerical study would be useful in future work, such as damage detection of structural members using these algorithms. 6. conclusions the paper studies the application of behaviorally inspired algorithms and their hybrids for structural parametric identification in the time domain. the stiffness and exac t (mn) pso ga aco ga+pso aco+pso aco+ga 300 -1.61 0.15 27.27 -0.4 17.27 18.67 300 5.68 -5.59 45.11 -1.05 -15.28 -24.25 300 3.43 4.92 -13.03 0.64 -17.75 -9.73 300 0.09 3.65 -15.15 -0.44 -10.30 20.25 300 2.73 -4.64 -1.52 0.4 -8.73 -16.38 300 -2.7 6.62 -31.21 0.37 3.12 1.14 300 5.37 -5.31 -11.52 -0.5 -2.27 5.03 300 -5.07 -6.2 3.7 -0.32 20.35 -17.72 300 -5.40 2.11 39.0 0.62 23.77 -6.41 300 1.56 -1.99 -24.1 0.42 -16.9 17.35 da mping 0.05 0.87 5.61 25.67 -0.12 3.54 7.35 abs. ma x. error 5.68 6.62 45.11 1.05 23.77 24.25 abs. mean error 2.9 3.9 19.7 0.44 11.6 12.0 co mputational time in seconds 82 107 166 60 105 137 table 9. percentage error in identified value of 11-member planar truss with unknown damping and noise exac t (gp a) pso ga aco ga+pso aco+pso aco+ga 200.00 -3.33 11.14 -20.15 0.54 0.32 10.97 170.00 -3.58 -2.22 -3.74 -2.96 -3.45 -26.92 140.00 11.07 -5.21 -28.44 -0.43 0.32 -19.7 110.00 4.67 -3.05 -26.45 -1.87 -1.43 -33.61 80.00 6.30 -11.29 25.0 2.52 4.89 1.14 50.00 6.77 7.94 14.55 1.36 9.24 23.72 80.00 -6.25 6.71 25.76 -6.39 -9.98 -13.64 110.00 -5.73 18.04 24.44 -0.81 -13.99 -30.03 140.00 6.39 -17.03 -32.21 5.09 -5.84 -5.63 170.00 -1.18 -28.35 -38.56 -1.13 -21.04 -17.65 200.00 -2.23 27.03 17.28 4.46 19.66 35.70 da mping 0.05 0.84 0.91 3.30 0.47 1.45 20.00 abs. ma x. error 11.07 28.35 38.56 6.39 21.04 35.70 abs. mean error 4.80 11.60 21.67 2.30 7.60 20.0 co mputational time in seconds 138 279 343 104 175 225 table 10. prcentage error in identified value of simply supported beam with unknown damping and noise 76 the journal of engineering research vol. 6, no. 2 (2009) 64-77 damping parameters are predicted from the time domain acceleration responses. impulse and random excitations cases are studied. particle swarm optimization (continuous), genetic algorithm (continuous) and classical ant colony optimization (discrete) are used here. three different hybridizations of ga, pso and aco have been investigated using three numerical models i.e., a lumped mass model, a truss and a non-uniform beam. comparing the performance between pure ga and pso it is found that the latter consistently outperforms the former in computational time and mean error, especially in the presence of noise. the computational implementation of pso is also simpler than ga. unlike ga, pso keeps track of the global best and particle best solution from iteration to iteration which explains its superiority over ga and fast convergence. regarding the performance of hybrid algorithms, the ga+pso has shown its superiority over all the other algorithms in both speed and accuracy. the next best performer is pure pso followed by pure ga. the performance of pure ga is however comparable to aco+pso hybrid in a few cases. however pure ga would be preferred as it is simpler to program than the hybrid. pure aco performed worst in all cases as expected because of its discrete nature. references abbospur, k.c., schulin, r. and van ganuethen, m., 2004, "estimating unsaturated hydraulic parameters using ant colony optimization," advances in water resources, vol. 24, pp. 827-841. binghui, yu., xiaohui yuan and jinwen wang, 2007, "short-term hydro-thermal scheduling using particle swarm optimization method," energy conversion and management, vol. 48, pp. 19021908. bao zhang, hong-fei teng and yan-jun shi, 2008, "layout optimization of satellite module using soft computing techniques," applied soft computing, vol. 8(1), pp. 507-521. bullnheimer, b., hartl, r.f. and ch. strauss, 1999, "an improved ant system algorithm for the vehicle routing problem," annals of operations research, vol. 89, pp. 319-328. christian blum, 2005, "ant colony optimization: introduction and recent trends," physics of life reviews, vol. 2, pp. 353-373. charles v camp and barron bichon, j., 2004, "design of space trusses using ant colony optimization," journal of structural engineering, asce, vol. 130(5), pp. 741-751. charles camp, v., barron bichon, j. and scott stovall, p., "design of steel frame using ant colony optimization," journal of structural engineering, asce, vol.131(3), pp. 369-379. chakraborthy. s. and mukhopadhyay, m., 2002, "determination of physical parameters of stiffened plates using genetic algorithm," journal of computing in civil engineering, vol. 16(3), pp. 206221. dorigo m., 1992, "optimization, learning and natural algorithms," doctoral dissertation, politecnico di milano, italy. dorigo, m.; l. m. gambardella "ant colony system: a cooperative learning approach to the travelling salesman problem", ieee transactions on evolutionary computation, vol. 1(1), 1997, pp. 5366. dorigo, m., maniezzo, v. and colorni, a., 1994, "ant system: optimization by a colony of cooperating agents," ieee transactions on systems, man and cybernetics, vol. 26(1), pp. 29-41. eberhant, r.c. and kennedy, j., 1995, "a new optimizer using particle swarm theory," proceedings of the 6th international symposium on micro machine and human science, nagaya, japan, pp. 39-43. friswell, m.i., penny, j.e.t. and garvey, s.d., 1998, "a combined genetic and eigensensitivity algorithm for the location of damage in structures," computers and structures, vol. 69(5), pp. 547-556. haupt, r. l., 1999, "practical genetic algorithms," john wiley and sons, new york. hitoshi furuta., masahiro yasui. and saeri fukuhara., 2005, "structural health monitoring using evolutionary computing," 6th world congress of structural and multidisciplinary optimisation, brazil. hami, y., amodeo, l., yalowi, f. and chen, h., 2007, "ant colony optimization for solving industrial layout problem," european journal of operation research (article in press). hanagud, v., meyyappa, m. and craig, j.i., 1985, " method of multiple scales and identification of nonlinear structural dynamic systems," aiaa journal, vol. 23(5), pp. 802-807. qie, he., ling wang, and bo liu, 2007, "parameter estimation for chaotic systems by particle swarm optimization," chaos, solitons and fractals, vol. 34, pp. 654-661. hu, y. and zhou z.y., 2004, "the particle swarm optimization algorithm and its application in design optimisation of a mini-uav," journal of flight dynamics, vol. 22 (2), pp. 61-64. jaco, f., schutte byung-il koh., jeffrey, a., reinbolt., raphael, t., haftka, alan, d., george, and benjamin fregly, f., 2005, "evaluation of a particle swarm algorithm for biomechanical optimization," journal of biomedical engineering, vol. 127, pp. 465-474. jesiel, c,. scott, c. and christophe, b., 1999, "application of genetic algorithms for the identification of elastic constants of composite materials from dynamic tests," international journal for numerical methods in engineering, vol. 45, pp. 891-900. gaetan kerschena, keith worden, alexander vakakis, f. and jean-claude golinval., 2006, "past, present and future of nonlinear system identification in structural dynamics," mechanical systems and signal processing, vol. 20, pp. 505-592. 77 the journal of engineering research vol. 6, no. 2 (2009) 64-77 kishore kumar, r., sandesh, s. and shankar, k., 2007, "parametric identification of non-linear dynamic systems using combined levenberg-marquardt and genetic algorithm," international journal of structural stability and dynamics, vol.7(4). koh, c.g., chen, y.f. and liaw, c.y., 2003, "a hybrid computational strategy for identification of structural parameters," computers and structures, vol. 81, pp. 107-117. koh, c.g., see, l.m. and balendra, t., 1991, "estimation of structural parameters in time domain: a substructural approach," earthquake engineering and structural dynamics, vol. 20, pp. 787-801. koh, c,g. and shankar, k., 2003, "substructural identification method without interface measurement," journal of engineering mechanics," vol. 129(7), pp. 769-776. michalwicz, z., 1994, "genetic algorithms + data structures = evolutionary programs" ai series," springer verlag, new york, 1994. mouser, c.r. and dunn, s.a., 2005, "comparing genetic algorithm and particle swarm optimization for inverse problems," anziam journal, vol. 46(e), pp. c89-c101. paul corry and erhan kozan., 2004, "ant colony optimization for machine layout problem," computational optimization and application, vol. 28, pp. 287-301. peng-yeng yin., 2006, "particle swarm optimization for point pattern matching," journal of visual communication image representation, vol. 17, pp. 143-162. perry, m.j., koh, c.g. and choo, y.s., 2004, "modified genetic algorithm for structural identification," computers and structures, vol. 84, pp. 529-540. rajamani, d. and adil, g.k., 1996, "machine loading in flexible manufacturing systems considering routing flexibility," intl j adv manuf technology, vol. 11(5), pp. 372-380. serra, m. and venini, p., "on some applications of ant colony optimization metaheuristic to plane truss optimization," struct multidisc optim , vol. 32, pp. 499-506. shi, y. and eberhart, r., 1998, "parameter selection in particle swarm optimization," proceeding of the 7th annual conference on evolutionary programming, pp. 591-601. the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 new asymmetric 21-level inverter with reduced number of switches v. thiyagarajan a, *, and p. somasundaram b a department of eee, ssn college of engineering, chennai, tamilnadu, india. b department of eee, ceg, anna university, chennai, tamilnadu, india. abstract: multilevel inverter plays an important role in the field of modern power electronics and is widely being used for many high voltage and high power industrial and commercial applications. the objective of this paper is to design and simulate the modified asymmetric multilevel inverter topology with reduced number of switches. the proposed inverter topology synthesizes 21-level output voltage during symmetric operation using three dc voltage sources and twelve switches 8 main switches and 4 auxiliary switches. the different methods of calculating the switching angles are presented in this paper. the matlab/simulink software is used to simulate the proposed inverter. the performance of the proposed inverter is analyzed and the corresponding simulation results are presented in this paper. keywords: mulilevel inverter, symmetric, asymmetric, thd, switching angle, pwm. أقل مفاتيحعدد مع مستوى 12ذو جديد غير متماثل محول ف. ثياغاراجان *أ، . سوماسندارامبو ب متعدد المستويات دوًرا مهًما في مجال اإللكترونيات الحديثة للطاقة ويستخدم على نطاق واسع في محوليلعب ال :الملخص الهدف من هذه الورقة إن . فائقةطاقة كذلك تلك المتطلبة العديد من التطبيقات الصناعية والتجارية ذات الجهد العالي و وتعمل متعدد المستويات مع تقليل عدد المفاتيح.و الغير المتماثل والمعدل محولطوبولوجيا ال محاكاةهو تصميم و ةالبحثي باستخدام ثالثة ةالتشغيل المتماثلعملية أثناء 12 ات الـمستوىال ذو رج اجهد الخالتجميع على المقترحة محولطوبولوجيا ال الطرق المختلفة ويقدم الباحثان . مفاتيح مساعدة( 4مفاتيح رئيسية و 8) مفتاحا و اثني عشرمباشر تيار لمصادر جهد المقترح. محولمحاكاة الل برمجيات ماتالب/سيمالينكباستخدام كما يقوما . حثيةبال لحساب زوايا التبديل في هذه الورقة هة.بمتشاالمقترح وعرض نتائج المحاكاة ال محولتحليل أداء البفي هذه الورقة ويقومان تعديل ؛زاوية التبديل ؛ تشويه التوافقي الكليال ؛غير المتماثل ؛متماثل ؛ المستوياتتعدد ممحول : المفتاحية الكلمات ذبات.الذب نطاق corresponding author’s e-mail: thiyagarajanv@ssn.edu.in doi: 10.24200/tjer.vol16iss1pp18-27 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 19 1. introduction to date multilevel inverters have greatly attracted many researchers due to their promising advantages for many industrial applications meynard et al. (2002). the different advantages of multilevel inverters include lower switching frequency, lower switching losses, lower peak inverse voltage, smaller common mode voltage, lower harmonic distortion, less electromagnetic interference and high voltage capability farakhor et al. (2015). the three main conventional multilevel inverter topologies are diode clamped or neutral point clamped (busquets-monge et al. 2008; nabae et al. 1981), flying capacitor (sadigh et al. 2010; dargahi et al. 2014) and cascaded h-bridge with independent dc sources (babaei et al. 2012; babaei et al. 2015; ramani et al. 2015). a high power rating can be achieved by the multilevel inverter. various resources such as photovoltaic, battery and fuel cells can be used as a dc source, which can be easily interfaced for a high power and medium voltage application. in diode clamped inverter, large numbers of clamping diodes are required and regulating the capacitor voltages makes the control method more complex (busquetsmonge et al. 2008; nabae et al. 1981). in flying capacitor inverter, the size of the inverter is increased because of the requirement of more capacitors to achieve higher levels (sadigh et al. 2010; dargahi et al. 2014). cascaded h-bridge inverter is more modular compared with other topologies. it can be easily expandable for a greater number of output voltage levels. however, to achieve higher output voltage levels, it requires a greater number of individual dc voltage sources for each module and semiconductor devices which increase the number gate drive circuit (banaei et al. 2013; toupchi khosroshahi, 2014; babaei et al. 2007; prabaharan et al. 2015). to overcome these drawbacks, different topologies of multilevel inverters and modified pulse width modulation (pwm) methods have been developed in recent years. the pwm methods for multilevel inverters help to achieve the following objectives: easy implementation, reduced switching loss and minimum total harmonic distortion grahame et al. (2003). the most widely used pwm techniques for multilevel inverters are the carrier based pwm techniques (rahim et al. 2010; jayabalan et al. 2017), selective harmonic elimination pwm (she-pwm) techniques zhang et al. (2009) and the space vector based pwm techniques vafakhah et al. (2010). different topologies of transformer with less inverter are proposed to improve the efficiency and reduce current leakage (thiyagarajan et al. 2016; li et al. 2015; islam et al. 2015). lately, different topologies of multilevel inverters with reduced components have been developed (karasani et al. 2016; alishah et al. 2014; samadaei et al. 2016; babaei et al. 2014). these inverter topologies may be symmetric or asymmetric. however, these topologies do not considerably reduce the power switches. hence, a new topology with reduced number of switching devices and voltage sources is proposed in this paper. unlike the previously mentioned topologies, the proposed inverter topology reduces the total number of switches in the conduction path, thereby minimizes the switching losses. in this paper, a modified multilevel asymmetric inverter topology with a reduced number of power switches is proposed. the proposed inverter uses three dc voltage sources and twelve switches (8 main switches and 4 auxiliary switches) to generate 21-level output voltage during asymmetric mode of operation. the different modes of operation of the proposed 21-level inverter are explained in section-2. section-3 presents the comparison of the proposed topology with the other existing topologies. section-4 discusses the different methods of calculating the switching angles. the simulation results obtained using matlab/simulink are explained in section-5. the detailed conclusion is presented in section-6. 2. proposed inverter topology figure 1 shows the circuit diagram of the proposed 21-level asymmetric inverter. the proposed inverter consists of 3 dc voltage sources, 8 main switches and 4 auxiliary. here, the switches s4 and s5 are bidirectional and other switches are unidirectional. the term "asymmetric" is used as magnitude since the magnitude of each dc voltage sources is different. the proposed asymmetric multilevel inverter consists of two units, namely, a level creator unit and a polarity changing unit. the level creator unit creates ten levels of positive voltage. the polarity changing unit helps to change the polarity of the voltage generated by the level creator unit. the positive levels of voltage are obtained when the switches s9 and s12 are turned on and switches s10 and s11 are turned off. similarly, the negative levels of voltage are obtained with the switches s9 and s12 are turned off and switches s10 and s11 are turned on. the proposed inverter can generate 21level output voltage (i.e. 10 positive levels, 10 negative levels and 1 zero). in mode-1, s1 and s8 are figure 1. proposed 21-level inverter. 19 v. thiyagarajan and p. somasundaram turned on to get level-1 voltage. to get level-2 voltage during mode-2, the switches s4 and s7 are tuned on. the level-3 voltage is obtained during mode-3 where the switches s3 and s7 are turned on. during mode-4 operation, the switches s3 and s8 are tuned on and the level-4 voltage is obtained. the level-5 voltage is obtained during mode-5 operation. during mode-5, the switches s2 and s5 are on. during mode-6, the level-6 voltage is obtained with the switches s1 and s5 are on. (a) v2 (b) v1+v2 (c) v3-v2 (d) v1+v2+v3 (e) (v1-v2) (f) v3 (g) (v2+v3) (h) (v1-v2+v3) figure 2. different output levels of the proposed inverter topology. 20 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 21 the level-7 voltage is obtained when the switches s1 and s6 are on during mode-7. the level-8 voltage is obtained when the switches s4 and s5 are on and the other switches are off during mode-8. the voltage level-9 is obtained during mode-9 operation. in this mode, the switches s4 and s6 are on. during mode-10, the level-10 voltage is obtained when the switches s3 and s6 are on and the other switches are off. the different output levels of the proposed inverter topology with its current conduction path is shown in fig. 2. the switching states for different output voltage levels during positive cycle of the proposed multilevel inverter is given in table 1. 3. comparison the comparison of the output voltage levels with the number of dc voltage sources, the number of switches and the number of on-state switches for different topologies of multilevel inverter are given in table 2. for the proposed inverter topology, the ratio of the number of dc sources (ndc) to the number of levels (n) is 0.14 and the ratio of the number of switches (n s) to the number of levels is 0.57. therefore, it is clear that the proposed multilevel inverter uses minimum number of switches to achieve fifteen level output voltage. the proposed topology used minimum switching devices and dc voltage sources to generate 21-level output voltage waveform. in addition, the number of on-state switches for the proposed inverter topology is very minimal when compared with other topologies. for the proposed topology, the number of switches in the current conduction path is only 4. this will considerably reduce the switching losses and thereby will increase the efficiency of the proposed inverter topology. table 1. switching states for different output levels. mode s1 s2 s3 s4 s5 s6 s7 s8 output voltage mode 0 0 1 0 0 0 0 0 1 0 mode 1 1 0 0 0 0 0 0 1 v2 mode 2 0 0 0 1 0 0 1 0 v1 v2 mode 3 0 0 1 0 0 0 1 0 v1 mode 4 0 0 1 0 0 0 0 1 v1 + v2 mode 5 0 1 0 0 1 0 0 0 v3 v2 mode 6 1 0 0 0 1 0 0 0 v3 mode 7 1 0 0 0 0 1 0 0 v2 + v3 mode 8 0 0 0 1 1 0 0 0 v1 v2+ v3 mode 9 0 0 0 1 0 1 0 0 v1+ v3 mode 10 0 0 1 0 0 1 0 0 v1+ v2+ v3 table 2. comparison of different multilevel inverters. inverter number of dc sources (ndc) number of switches (ns) number of on-state switches number of level (n) ratio (ndc / n) ratio (ns / n) cascaded h-bridge inverter 3 12 6 13 0.23 0.92 alishah et al. 2014 3 7 5 13 0.23 0.54 babaei et al. 2007 3 16 4 13 0.23 1.23 babaei et al. 2012 3 8 5 15 0.20 0.53 babaei et al. 2014 4 12 6 13 0.31 0.92 babaei et al. 2015 4 11 5 9 0.44 1.22 banaei et al. 2013 4 10 6 13 0.31 0.77 jayabalan et al. 2017 3 9 5 13 0.23 0.69 karasani et al. 2016 5 9 4 11 0.45 0.82 thiyagarajan et al. 2017 4 10 4 17 0.23 0.59 proposed inverter 3 12 4 21 0.14 0.57 21 v. thiyagarajan and p. somasundaram 22 4. calculation of switching angles the switching angles play a major role to reduce the of total harmonic distortion (thd) of the output voltage waveform. the different methods of calculating the switching angles for the proposed asymmetrical 21-level inverter are given below luo et al. (2013). method 1 in method-1, the switching angles are distributed averagely over the range 0–90 o and are determined by,        2 1-n 3..., 2, 1, where, i n 180 jj  (1) method 2 in method-2, the switching angles are determined by using the following equation,                 2 1n ,...,2,1j where, 1n 1j21 sin 2 1 j (2) method 3 the switching angles for this method are determined by using the following equation,            2 1n ,...,2,1j where, 1n 180 jj (3) method – 4 the above three methods are capable of arranging the main switching angles where the output waveform is not a sine wave. according to the sine function, a new method to determine the main switching angles were established. the main idea of this method is that when the function value increases to half to the half-height of the output level, the switch angle is set and thus a better output waveform is obtained. in this method, the main switching angles are determined by the formula given below.                2 1-n .... 3 2, 1,j where, 1n 1j21 sinj (4) where, n = number of output levels. this method gives a better output voltage waveform compared with the other methods as the output waveform is more similar to sine wave. the switching angles corresponding to the period 0 to 90 o are called as main switching angles. the other switching angles are obtained using the following relations: 1. for period 0 to π/2 : θ1, θ2, . . . , θ(n-1)/2. 2. for period π/2 to π : θ (n+1)/2. . ., θ (n-1). = (πθ(n1)/2),. . .,(πθ1). 3. for period π to 3π/2 : θn, . . ., θ3(n-1)/2 = (π+ θ1),. . .,(π+θ(n-1)/2). 4. for period 3π/2 to 2π : θ(3n-1)/2, . . ., θ2(n-1) = (2π θ(n-1)/2),. . .,(2πθ1). for the proposed 21-level inverter, there are ten main switching angles and these are given in table 3. the switching pulses obtained using the different methods are shown in fig. 3. it is seen that only two main switches were turned on to synthesize the required output voltage at any time. 5. simulation results the matlab/simulink simulation results of the proposed asymmetrical 21-level inverter are presented in this section. the proposed inverter consists of three dc voltage sources and eight main switches. the values of the different dc voltage sources are v1 = 60v, v2 = 20v and v3= 120v. the magnitude of the voltage sources is selected in the ratio 2:1:3. the maximum voltage obtained as 200 v (i.e. v1 + v2+ v3). the 21-level output voltage obtained for different switching methods is shown in fig. 4. it is seen that the output voltage waveform is triangular in shape for method-1 and method-3 as the switching angles are distributed averagely over the range 0–90 o . for method-2, the output voltage waveform is trapezoidal in shape. it is also observed that the output voltage waveform is sinusoidal for method-4. the fft analysis of the 21 level output voltage waveform for different switching methods are shown in fig. 5. the simulation results show that the harmonic content of the 21-level output voltage waveform for the switching method-4 is less as 3.90% when compared with the other methods. however, it is observed that the fundamental output voltage for the switching method-2 is as high as 240.3 v when compared with the other switching methods. as mentioned in section 3, method-4 gives a better output voltage waveform compared with the other methods. the comparison of the fundamental output voltage and %thd are given in table 4. the load current waveform for the different loading conditions is shown in fig. 6. the comparison of the thd of the load current for different loading conditions is given in table 5. it is observed that an increase in the inductive load decreases the thd of the current waveform, however, the thd of the voltage waveform remains the same for the proposed inverter topology. 22 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 23 (a) (b) (c) (d) figure 3. switching pulses (a) method 1 (b) method 2 (c) method 3 and (d) method 4. 23 v. thiyagarajan and p. somasundaram 6. conclusion a new asymmetric type 21-level inverter with a minimum number of switches has been proposed in this paper. the main advantage of this inverter topology is using only three dc voltage sources and eight main switches to achieve 21-level output voltage during the asymmetric mode of operation as compared with other conventional topologies. the three different methods used to calculate the switching angles are presented and the simulation results are also compared in this paper. the maximum obtained voltage is equal to the sum of the magnitude of the individual voltage. the simulation result shows that the switching angles obtained by method 4 achieves less thd compared with the other methods. however, method 2 achieves high fundamental output table 3. switching angles. angle main switching angles (in degree) method1 method2 method3 method4 θ1 8.5714 1.433 8.1818 2.866 θ2 17.1429 4.3135 16.3636 8.6269 θ3 25.7143 7.2388 24.5455 14.4775 θ4 34.2857 10.2437 32.7273 20.4873 θ5 42.8571 13.3718 40.9091 26.7437 θ6 51.4286 16.6835 49.0909 33.367 θ7 60.0000 20.2708 57.2727 40.5416 θ8 68.5714 24.2952 65.4545 48.5904 θ9 77.1429 29.1058 73.6364 58.2117 θ10 85.7143 35.9026 81.8182 71.8051 table 4. comparison of fundamental output voltage and thd. method fundamental output voltage (v) thd (%) method 1 157.6 16.43 method 2 240.3 20.48 method 3 165.3 15.74 method 4 200.7 3.90 table 5. comparison of thd of the current waveform. load thd (%) method-1 method-2 method-3 method-4 r=100ω 16.43 20.48 15.74 3.90 r=100ω l=5mh 16.10 20.37 15.45 3.05 r=75ω l=10mh 15.72 20.26 15.12 2.01 r=50ω l=5mh 15.79 20.37 15.17 2.20 24 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 (a) (b) (c) (d) figure 4. output voltage (a) method 1 (b) method 2 (c) method – 3 and (d) method 4. (a) (b) (c) (d) figure 5. fft analysis (a) method 1 (b) method 2 (c) method 3 and (c) method 4. 25 25 v. thiyagarajan and p. somasundaram figure 6. output voltage (a) r=100ω (b) r=100ω and l=5mh (c) r=75ω and l=10mh and (d) r=50ω and l=5mh. voltage as compared with the other methods. the major advantage of the proposed 21-level inverter is using a minimum number of switches and hence both the cost and the size of the inverter are reduced. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references alishah rs, nazarpour d, hossein sh, sabahi m (2014), design of new single-phase multilevel voltage source inverter. international journal of power electronics and drive system 5(1): 45-55. babaei e, hosseini sh, gharehpetian gb, tarafdar haque m, sabahi m (2007), reduction of dc voltage sources and switches in asymmetrical multilevel converters using a novel topology. electric power systems research 77: 1073–1085. babaei e, kangarlu mf, mazgar f (2012), symmetric and asymmetric multilevel inverter topologies with reduced switching devices. electric power systems research 86:122–130. babaei e, laali s, alilu s (2014), cascaded multilevel inverter with series connection of novel h-bridge basic units. ieee transactions on industrial electronics 61(12): 6664-6671. babaei e, laali s, bayat z (2015), a single-phase cascaded multilevel inverter based on a new basic unit with reduced number of power switches. ieee transactions on industrial electronics 62(2): 922929. banaei mr, salary e (2013), asymmetric cascaded multi-level inverter: a solution to obtain high number of voltage levels. journal of electrical engineering and technology 8(2): 316-325. busquets-monge s, rocabert j, rodriguez p, alepuz s, bordonau j (2008), multilevel diode-clamped converter for photovoltaic generators with independent voltage control of each solar array. ieee transactions on industrial electronics 55(7): 2713-2723. dargahi s, babaei e, eskandari s, dargahi v, sabahi m (2014), flying-capacitor stacked multicell multilevel voltage source inverters: analysis and modelling. iet power electronics 7(12): 29692987. farakhor a, reza ahrabi r, ardi h, najafi rs (2015), symmetric and asymmetric transformer based cascaded multilevel inverter with minimum number of components. iet power electronics 8(6): 1052-1060. (a) (b) (c) (d) 26 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 18-27 27 grahame dh, thomas al (2003), pulse width modulation for power converters. wiley interscienceinc. islam m, mekhilef s (2015), h6-type transformerless single-phase inverter for grid-tied photovoltaic system. iet power electronics, 8(4): 636-644. jayabalan m, jeevarathinam b, sandirasegarane t (2017), reduced switch count pulse width modulated multilevel inverter. iet power electronics 10(1):10-17. karasani rr, borghate vb, meshram pm, suryawanshi hm (2016), a modified switcheddiode topology for cascaded multilevel inverters. journal of power electronics 16(5): 1706-1715. li w, gu y, luo h, cui w, he x, xia c (2015), topology review and derivation methodology of single-phase transformerless photovoltaic inverters for leakage current suppression. ieee transactions on industrial electronics 62(7): 4537–4551. luo fl,ye h (2013), advanced dc/ac inverters. crc press llc. meynard ta (2002), multicell converters: derived topologies. ieee transactions on industrial electronics 49(5): 978-987. nabae a, takahashi i, akagi h (1981), a new neutral-point-clamped pwm inverter. ieee transactions on industry applications 17(5): 518523. prabaharan n, palanisamy k (2015), investigation of single phase reduced switch count asymmetric multilevel inverter using advanced pulse width modulation technique. international journal of renewable energy research 5(3): 879-890. rahim na, selvaraj j (2010), multistring five-level inverter with novel pwm control scheme for pv application. ieee transactions on industrial electronics, 57(6): 2111–2123. ramani k, sathik ma, sivakumar s (2015) a new symmetric multilevel inverter topology using single and double source sub-multilevel inverters. journal of power electrronics 15(1): 96-105. sadigh ak, hosseini sh, sabahi m, gharehpetian g.b. (2010), double flying capacitor multicell converter based on modified phase-shifted pulsewidth modulation. ieee transactions on power electronics 25(6): 1517-1526. samadaei e, gholamian sa, sheikholeslami a, adabi j (2016), an envelope type (e-type) module: asymmetric multilevel inverters with reduced components. ieee transactions on industrial electronics 63(11): 7148-7156. thiyagarajan v, somasundaram p (2016), performance analysis of photovoltaic array with h5 inverter under partial shading conditions. international journal of innovation and scientific research 22(1): 164-177. thiyagarajan v, somasundaram p (2017), modeling and analysis of novel multilevel inverter topology with minimum number of switching components. cmes: computer modeling in engineering and sciences 113(4): 461-473. toupchi khosroshahi m (2014), crisscross cascade multilevel inverter with reduction in number of components. iet power electronics 7(12): 29142924 vafakhah b, salmon j, knight am (2010), a new space-vector pwm with optimal switching selection for multilevel coupled inductor inverters. ieee transactions on industrial electronics 57(7): 2354-2364. zhang f, yan y (2009), selective harmonic elimination pwm control scheme on a threephase four-leg voltage source inverter. ieee transactions on power electronics 24(7): 16821689. 27 mech080705.qxd the journal of engineering research vol. 7, no. 1 (2010) 22-30 designed. these data are essential when the dynamic behavior of the aircraft is analyzed, indicating the great importance attached to the stability and maneuverability of modern aircraft. the dynamic aero characteristics of 1. introduction the data on aerodynamic characteristics are when the guidance system and the control system of the aircraft are _________________________________________ *corresponding author’s email: qilin@xmu.edu.cn preliminary step towards wire-driven parallel suspension systems for static and dynamic derivatives of the aircraft in low-speed wind tunnels yaqing zhenga,b, qi lina*, xiongwei liuc and peter mitrouchevd a* department of aeronautics, xiamen university, xiamen,361005,china b college of mechanical engineering and automation, huaqiao university, quanzhou, 362021, china c school of computing, engineering and physical sciences, university of central lancashire, preston, uk d g-scop laboratory, grenobleinp-ujf, fre 3028 cnrs, 46 avenue félix viallet, 38031 grenoble cedex1, france received 5 july 2008; accepted 7 june 2009 abstract: the wind tunnel test is one important way to obtain the aerodynamic derivatives of aircrafts. these derivatives are necessary when the guidance and control systems of the aircraft are designed and when the dynamic quality of the aircraft is analyzed as well. the results of experiments of the static derivatives and dynamic derivatives of the aircraft in low-speed wind tunnels have revealed that there are some unavoidable drawbacks such as the interference of the streamline flow brought about by the strut in the traditional strut suspension system. a cable-mounted system is very suitable for experiments of the static derivatives of an aircraft, but it cannot be used in the experiments of dynamic derivatives. in order to use the same wire-driven parallel suspension system to realize the static and dynamic derivates experiments in low-speed wind tunnels, a survey of the research work addressed within the wire-driven parallel suspension systems (wdpss-8) project is presented in this paper. the results show that wdpss-8 can be successfully used in experiments of static derivatives, and that it has potentiality to be used in experiments of dynamic derivatives. in the issues in the theoretical aspect the issues have been handled. however, much work should be done in the experimental aspects. the research outcomes of wdpss-8 will help the chinese set up chinese brands in the field of wind tunnel tests of aircrafts. keywords: wind tunnel, suspension system, wire-driven, static and dynamic derivatives áä«£ñdg áyô°ùdg ägp á«fgƒ¡dg ¥éøf’g ‘ iôfé£∏d á«μ«eéæj~dg h á«μ«jéà°ùdg äé≤à°ûª∏d ájrgƒàÿg ≥«∏©àdg º¶f ‘ »μ∏°ùdg üô° dg ƒëf á«dhc’g ägƒ£æÿg ød »c , ≠æjr èæ«céj,∞«lhôe î«h , ‹ „ƒc áá°°uuóóÿÿggée~æy h iôfé£∏d iô£«°ùdg h á«lƒàdg ωé¶f ºª°üj ée~æy ájqhô°v äé≤à°ûÿg √òëg .ägôfé£∏d á«μ«ªæjg~dg á«fgƒ¡dg äé≤à°ûÿg oééj’ ᪡e á≤jôw ƒg »fgƒ¡dg ≥øædg qéñàng : âø°ûc ~b áä«£ñdg áyô°ùdg agƒ¡dg ¥éøfg ‘ iôfé£∏d á«μ«ág~dg äé≤à°ûÿg h á«μ«jéà°ùdg äé≤à°ûª∏d üqééà∏d ≥«∏©àdg ωé¶f á°uón èféàf .iôfé£∏d á«μ«ªæjg~dg á«yƒædg π«∏– ºàj üqééà∏d g~l ö°séæe é«μ∏°s ∫ƒªùg ωé¶ædg .áj~«∏≤àdg áeéy~dg ‘ ≥«∏©àdg ωé¶f ‘ áeéy~dg πñb øe öñ°ùàÿg º¶àæÿg ¿éjôé∏d πng~àdg πãe é¡ñæœ øμá’ äébƒ©ÿg ¢ ©h ¿g øy äé≤à°ûe üqéœ ré‚’ é«μ∏°s ºë≤ÿg …rgƒàÿg ≥«∏©àdg ωé¶f ¢ùøf ∫骩à°sg πlg øe .á«μ«eéæj~dg äé≤à°ûª∏d üqééàdg ‘ é¡d骩à°sg øμá’ ∂dp ™eh ,iôfé£∏d á«μ«jéà°ùdg äé≤à°ûª∏d ) ´hô°ûÿ èféàædg √òg .åëñdg gòg ‘ é¡á~≤j ”h …rgƒàÿg ºë≤ÿg ∂∏°ùdg º¶f ¢uƒ°üîh åëñdg ∫éªyg í°ùe ” .áyô°ùdg áä«£ñdg ìéjôdg ¥éøfg ‘ á«μ«eéæj~dg h á«μ«jéà°ùdg (wdpss-8) ¿éh âæ«h èféàædg .åëñdg gòg ‘ é¡á~≤j ” ~b(wdpss-8)¬∏ªàfi äé«féμeg é¡d ∂dòch ,»μ«jéà°ùdg ¥é≤à°t’g üqéœ ‘ ìééæh ád骩à°sg øμ‡ ) ∫ƒm åëñdg èféàf .á«ñjôéàdg »mgƒædg ‘ òãμdg πª©dg öéj øμdh é¡©e πeé©àdg ” ájô¶ædg »mgƒædg ,»μ«eéæj~dg¥é≤à°t’g üqéœ ‘(wdpss-8ú«æ«°üdg ~yé°ùà°s .ägôfé£dg ¢üëød á«fgƒ¡dg ¥éøf’g π≤m ‘ á«æ«°üdg áeó©dg ™°vƒd áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .á«μªæjgo h á«μ«jéà°s äé≤à°ûe ,»μ∏°ùdg ™a~dg ,≥«∏©j ωé¶f ,»fgƒg ≥øf : 23 the journal of engineering research vol. 7, no. 1 (2010) 22-30 deep and systematic research work on a 6-dof wiredriven parallel suspension system with 8 wires (wdpss8) for low-speed wind tunnels deep and systematic research (liu et al. 2004; zheng, 2004; zheng et al. 2005, zheng et al. 2007, lin et al. 2006; zheng, 2006; liu et al. 2006; zheng, 2006 and liang et al. 2007). the attitude control and aerodynamic coefficients (static derivatives) of the scale model have been investigated in theory and in experiment. a prototype (wdpss-8) has been realized (built) and tested in an open return circuit low speed wind tunnel with a wind speed of 30 m/s. the experimental section of the wind tunnel is rectangular, with a length of 1 meter, a width of 0.52 meters and a height of 0.42 meters (liu et al. 2006). the authors also found the possibility of using the prototype (wdpss-8) for the experiment of dynamic derivatives by successfully implementing the single-dof oscillation control of the scale model (liang et al. 2007; hu, 2008 and lin et al. 2008). concerning the possibility of using the same wire-driven parallel suspension system to realize the static and dynamic derivates experiments in low-speed wind tunnels, a survey of the research work finished on some key issues of wdpss-8 for the experiments of the static and dynamic derivatives will be addressed. the research results of the experiments, especially those of the experiments of the dynamic derivatives, will provide some criterion of experimental data for free flight and some effective experimental methods about the controllability capability of post-stall maneuvers in the design of large aircraft and new generation of military aircraft. this will help the chinese set up the chinese brands in the field of wind tunnel tests of aircraft. the rest of the paper is organized as follows: the key issues of wdpss-8 for the experiments of static derivatives of aircrafts in low-speed wind tunnels (lswt) are given in the next section 2. the key issues of wdpss8 for the experiments of dynamic derivatives of aircraft in lswt are presented in section 3 by showing the possibility to use the same wdpss-8 for static and dynamic derivates experiments in lswt. finally, discussions and future work are suggested in section 4. 2. wdpss-8 for experiments of static derivatives of aircraft in low-speed wind tunnels 2.1 two wdpss-8 prototypes 2.1.1 a manually-operated wdpss-8 prototype validated in a closed-circuit wind tunnel a wire-driven parallel manipulator is a closed-loop mechanism where the moving platform is connected to the base via wires by a multitude of independent kinematic chains. the number of a moving platform's degree of freedoms (dofs) is defined as the dimension of linearity space which is positively spanned by all the screws of the structure matrix of the manipulator. so the moving platform of a 6-dof’s completely or redundantly restrained the aircraft are closely related with the aerodynamic forces and torques exerted on the aircraft. in analyzing the aerodynamic characteristics, these aerodynamic forces and torques can be expressed by some aerodynamic coefficients and derivatives. the latter are the derivatives of the aerodynamic coefficients to some parameters such as angle of attack, angle of side-slip and mach number which are called static derivatives. the derivatives of the aerodynamic coefficients to some parameters such as pitching angular velocity, rotating angular velocity and yawing angular velocity are called dynamic derivatives (sun, 1999). the wind tunnel test is one important way to obtain the aerodynamic derivatives of the aircraft. during a windtunnel tests it is necessary to place the scale model of the aircraft in the streamline flow of the experimental section of the wind tunnel using some kind of suspension system. the suspension system will have a lot of influences on the reliability of the results of the wind tunnel tests. the traditional rigid suspension systems have some unavoidable drawbacks for the experiments of static and dynamic derivatives such as the serious interference of the strut on the streamline flow (zhang et al. 2006; roas, 2001; bernhardt and williams, 2000; cao and deng, 1998 and du and wang, 2003). the cable-mounted systems for wind tunnel tests, developed in the past several decades, deal with the contradiction between the supporting stiffness and the interference on the streamline flow (bian et al. 1999; shen and huang, 1999; bennett et al. 1978; liu et al. 2001, wang et al. 2004 and griffin 1991). however their mechanism is not robotic and consequently quite different from wire-driven parallel suspension systems in attitude-control schemes and force-measuring principle. in addition they can not be used in dynamic derivatives experiments (graffin, 1992). instead, the freeflight simulation concept in wind tunnels through an active suspension such as six-dof (degrees-of-freedom) wire-driven parallel suspension systems (wdpss), coming from the research improvement in wire-driven parallel manipulator and force control, is suitable for calculating the aerodynamic derivatives of the aircrafts lafourcade et al. 2002; onera.fr/desd/sacso/index.php; liu, 2004 and zheng 2004). the static derivatives can be obtained by researching the static models without winds and with winds (zheng, 2006). some successful achievements in this field have been made in the suspension active pour soufflerie (sacso) project for nearly 8 years and it has been applied in vertical wind tunnel tests with a wind speed of 35 m/s for fighters at the first stage of their conceptual design. however the system can not be used in experiments of dynamic derivatives in low-speed wind tunnels (lafourcade et al. ). in this context our goal is to introduce some contributions in the field of wire-driven parallel suspension systems for static and dynamic derivatives of the aircraft in low-speed wind tunnels. under the sponsorship of nsfc (national natural science foundation of china), the authors and the co-authors have done over 6-years of 24 the journal of engineering research vol. 7, no. 1 (2010) 22-30 found in the experiments that there is little vibration occurring at the end of the scale model which is less than that in the corresponding traditional strut suspension system shown in fig. 1 (f), from which it has been found that the fundamental frequency of wdpps-8 is smaller than that of the corresponding traditional strut suspension. wire-driven parallel manipulator is driven by at least 7 or more wires. meanwhile, a 6-dof wire-driven parallel suspension system (wdpss) is essential for the free fight of the aircraft in a 3-dimensional space wind tunnel. fig. 1(a) shows the concept of a 6-dof wdpss driven by 8 wires (wdpss-8), the geometric definition of which is shown in fig. 1(b) and the geometric parameters of which are listed in table 1. a manually-operated prototype of such a design shown in fig. 1 (c) is built and tested in a closedcircuit lswt. to implement the scenario for the attitude adjustment of the aircraft, a driving mechanism adjusted manually has been developed which allows the aircraft model to maneuver, ie. to permit roll, pitch and yaw motion. for the wdpss-8, each cable will be attached to a driving unit, which consists of a screw bar and a driving nut, as shown in fig.1(d). a commercial load cell interfaced to the cable shown in fig. 1(e) is used to measure the tension of a cable. to avoid extra interference, the strain gage balance and driving unit are attached to the wind tunnel frame on the outside of the tunnel, as shown in fig. 1(c) and (d).however, fig. 1(f) shows the aircraft model mounted on a conventional strut suspension system in the same lswt. the wdpss-8 prototype has been validated by windtunnel tests in a wind speed of 28.8 m/s and it has been (a) the concept of wire-driven parallel suspension system with 8 wires (wdpss-8). (b) geometric definition of the wdpss-8 prototype. (c) wdpss-8 tested in lswt in a wind speed of 28.8 m/s. (d) driving unit of wdpss-8. (e) load cell interface. (f) strut suspension system. figure 1. comparison of 2 different suspension systems. 25 the journal of engineering research vol. 7, no. 1 (2010) 22-30 a test for this wdpss-8 for low-speed wind tunnels is also realized, shown in fig. 2(b), in which the 3 rotational attitude control of the scale model (yaw, roll and pitch) has been accomplished. during the wind tunnel testing for the static derivatives, it is necessary to place the scale model using the suspension system in the streamline flow of the experimental section in wind tunnels, using the suspension system, and the attitude of the scale model should be adjustable. the inverse kinematics problem deals with the calculation of the length of each cable corresponding to the given different attitude of the scale model which will provide the data for the position control. the modeling of inverse pose kinematics of wdpss-8 can be found in references (zheng et al. 2005 and zheng et al. 2006). 2.2 calculation of the static derivatives because the scale model moves in a quasi-static way during the lswt experiment for the static derivatives, it is reasonable to calculate the aerodynamic force and torque exerted on it using the difference of the force and torque exerted on the scale model between without wind and with wind. as a preliminary research, the assumption that all constraints are perfectly applied with no resistance in pulleys or other mechanisms such as point-shaped joints which are required to maintain the geometry of the wires at the base and the scale model is given for convenience. may be this is not really the case, but it is reasonable because the attitude of the scale model is controlled and adjusted in a quasi-static way so that the errors in the mechanism configuration between without wind and with wind could be easily limited to a range which can be neglected. the static model of wdpss-8 without wind can be expressed by: (1) here, t is a vector (t1…t8) with 8 components, 0 is a vector with 6 components, jt is the structural matrix of the manipulator, fg is the gravity vector with 6 components. the static model of wdpss-8 with wind can be expressed by: (2) here, fa is the vector of aerodynamic force and torque with 6 components. 2.1.2 another wdpss-8 prototype tested in an open return-circuit wind tunnel figure 2(a) shows another geometric definition of wdpss-8 and its the structural parameters are listed in table 2. pitch () c1 (x,y,z) c2 (x,y,z) c3 (x,y,z) c4 (x,y,z) -6 -300, 0, -32 300, 0, 32 0, 605, -30 0, -605, -30 0 -302, 0, 0 302, 0, 0 0, 605, -30 0, -605, -30 6 -300, 0, 32 300, 0, -32 0, 605, -30 0, -605, -30 12 -295, 0, 63 295, 0, -63 0, 605, -30 0, -605, -30 p1 (x,y,z) p2 (x,y,z) p3 (x,y,z) p4 (x,y,z) 0, 0, 420 0, 0, -420 0, 605, 0 0, -605, 0 table 1. geometric parameters of the wdpss-8 prototype (unit in mm) (a) another geometric definition of wepss-8 prototype. (b) prototype of wdpss-8 figure 2. another wdpss-8 prototype tested in an open return-circuit wind tunnel. p1 (xp,,yp,,zp) p2 (xp,,yp,,zp) p3 (xp,,yp,,zp) p4 (xp,,yp,,zp) -120, 0, 0 80, 0, 0 0, 50,0 0, -50, 0 b1 (x,y,z) b3 (x,y,z) b5 (x,y,z) b7 (x,y,z) 0, 0, 0 0, 0, 1040 0, -390, 520 0, 390, 520 table 2. structural parameters of the wdpss-8 prototype (unit in mm) jtt+fg =0 26 the journal of engineering research vol. 7, no. 1 (2010) 22-30 2.2.2 experiment of static derivatives using wdpss8 in open return circuit low speed wind tunnel with the experimental set-up in place, the wind tunnel testing by wdpss-8 is run by inducing airspeed of 30m/s, as shown in fig. 7. for calculation of wire tension, the sensor ns-wl1, with a range of 10 kg, was chosen. a group of wind tunnel tests are done for the scale model of wdpps-8 in an open return circuit low-speed wind tunnel and the experimental curves of lift coefficient versus angle of pitch, drag coefficient versus angle of pitch and lift/drag ratio versus angle of pitch are shown. in order to calculate the static derivatives (related to fa), we should measure the tension of all wires and the posture of the scale model when the scale model is position-controlled without wind and with wind. 2 different groups of experiment of static derivatives using wdpss8 with 2 different mechanism configuration parameters have been finished in a closed circuit low-speed wind tunnel and an open return circuit low-speed wind tunnel respectively, which will be stated in the following in detail. 2.2.1 experiment of static derivatives using wdpss-8 finished in a closed low-speed wind tunnel as shown in fig. 1 and fig. 4, the experiments of static derivatives using wdpss-8 and using the conventional strut suspension system have been finished in a closedcircuit low-speed wind tunnel with a wind speed of 28.8 m/s. the attack angle (pitch) can be adjusted from 00 to 180. fig. 4 and fig. 5 show the measured lift and drag forces versus the pitch angles and the measured momentversus the pitch angles using the conventional wind tunnel testing and using wdpss-8 wind tunnel testing respectively. figure 4. wdpss-8 in a closed low-speed tunnel. 0 2 4 6 8 10 12 14 16 -10 -5 0 5 10 15 pitch (degree) f o rc e ( n ) lift drag figure 5(a) measured lift and drag vs, pitch using the conventional wind tunnel testing. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 -10 -5 0 5 10 15 pitch (degree) m om en t ( n m ) figure 5(b) measured moment vs. pitch using the conventional wind tunnel testing. figure 5. conventional wind tunnel testing figure 6(a) measured lift and drag vs. pitch using wdpss-8 wind tunnel testing. figure 6(b) measured moment vs. pitch using wdpss-8 wind tunnel testing. figure 6. wdpss-8 wind tunnel testing. figure 7. wire-driven parallel suspension systems with 8 wires in the experiment of static derivatives. 27 the journal of engineering research vol. 7, no. 1 (2010) 22-30 (4) here, jt v = 0 is satisfied. moreover kd and kv are the different values of the control feedback gain without wind. it can be proven that the control system is stable and robust with the control law mentioned above. it is noted that if another control law is used which can ensure the stability of the control system, the value s of dynamic derivatives calculated from the control system which will be formulated in detail in the next section will be different. a question may arise about the correctness of the method for the dynamic derivatives' calculation, but it is regarded as reasonable after a balance of the analysis of the differences of control schemes and of their robustness is given. much more work in this aspect will be discussed in the future work. moreover the required repeatability of the control system will be indicated and investigated using some kind of tools like robustness. 3.2 dynamic derivatives' calculation taking the pitch oscillation for example (ie. x =p = [0 0 0 0 p]t = [0 0 0 0 p0 sin 0]t), the total dynamic model without wind and with wind can be obtained. in fact, the single-dof oscillation control of the scale model is successfully implemented using the wdpss-8 prototype, the frequency of which is from 0~2 hz and the amplitude of which is 9~10 degree (liang et al. 2007). in addition, concerning the issue that the motion versus time tracks that differs by small fractions of a degree should be provided by a motion control system in order to obtain accurate measurements of the dynamic derivatives. may be a time-varying discrete control system should be built and investigated, but as the preliminary research, it is only regarded as a time-constant continuous control system. moreover, the specifications on the path of the single-dof oscillation of the scale model matching with wind on and wind off should be considered to an extent that target precisions are related to expected accuracies of the measured derivatives, but this issue will not be discussed here. much more work in the aspect will be discussed in future work with the tools like the robustness of the control system. under the condition when the scale model has a pure pitch rotation without wind, eq. (3) can be written as (mo + jt mj) p + (mo+jt m j+jt bj) p = jt f8 (5) from eq. (5), we can get p + b p = c, where: a is the result of adding all the elements of the 5th row of matrix (mo + jt mj ), b is the result of adding all the elements of the 5th row of matrix (mo + jt m j + jt bj), c = kd (pd p ) + kv (pd p ), p pd and pd are the desired pitch angle and angular velocity of the scale model. as shown in fig. 8, there are 3 experimental curves for wind tunnel testing of wdpss-8 with a wind speed of 17.27m/s. though there are no data about the standard model as criteria, the curves are suitable for expressing the aerodynamic characteristics of aircraft. 3. wdpss-8 for the experiments of dynamic derivatives of the aircraft in low-speed wind tunnels in the test platform of wdpss-8, the oscillation control of the scale model is controlled without wind and with wind respectively. the vibration displacement of the scale model is measured by the sensors inside the model such as micro gyro and micro accelerators. the aerodynamic force and torque are calculated by the dynamic equations of the systemm. also the dynamic derivatives can be calculated by the real torques of motors without wind and with wind which can be measured by force sensors mounted on the axis of the motors. the following parts will state the issue in detail. 3.1 dynamic modeling of the system and the scheme of oscillation control the suspension system's dynamic modeling is necessary for the control system's design of the oscillation of the scale model. the following assumptions when carrying out the dynamic modeling are given as follows: (1) the deformation of wires is so small that it can be neglected. the mass of the wires can be neglected as well. (2) the dynamics of the actuators are neglected to simplify the dynamics model of the manipulator. in references zheng et al. 2007 and zheng 2006, the total dynamic modeling of wdpss-8 is written as: (mo + jt mj) x + (mo + jt m j + jt bj) x = jt f8 (3) the total dynamic modeling is a highly coupled and redundantly restrained nonlinear system which should be decoupled and linearized. a control law of actuator vector of motor is designed as follows: here ] 0 0)( [ )33(g33 3333p 0    a i m m is the inertia matrix of the scale model including any attached payload, pm is the mass and ga is the inertia tensor about the gravity center. 8881 r),,(diag  mmm is the inertia matrix of the actuators x is vector of the posture of the scale model, 8881 r),,(diag  bbb is the matrix of viscous friction of the actuators 8t81 r),,(  ô is vector of the torque of the actuators t pg )0,0,0,.,0,0( gmf is the gravity vector of the scale model, and g=9.8 (m/s2). ô (jt)+ (kd ( xx d ) + kv ( .. xx d ) + fg) + v 28 the journal of engineering research vol. 7, no. 1 (2010) 22-30 (7) my (t)on, defined as the oscillation torque vector of the system with wind when the scale model has a pure pitch rotation , satisfies (8) from eqs. (5) and (7), the following equation can be obtained: (9) from eqs. (6), (8) and (9), the following equation can be obtained: (10) (11) hence, eq. (9) can be expressed by: here js is the result of adding all the elements of the 5th row of matrix jt. my (t)off , defined as the oscillation torque vector of the system without wind when the scale model has a pure pitch rotation, satisfies: (6) under this condition when the scale model has a purepitch rotation with wind, eq. (3) can be written as (a) the lift coefficient vs angle of attach (b) the drag coefficient vs angle of attach (c) lift-to-drag ratio vs angle of attach figure 8. experimental curves for wind tunnel testing of wepss-8 fitting curveexperimental value angle of attack /o th e lif t c oe ff ic ie nt c y experimental value fitting curve angle of attack /o th e dr ag c oe ff ic ie nt c d experimental value fitting curve angle of attack /o lif tto -d ra g ra tio y off ( m ) sin ( t )    .. p ym  pè + . p ym  pè + p  ym pè +  p .. t 0 )( èmjjm g 't . t . t . 0 )( fôjèbjjjmjm  p ony tm )( = y off m )sin(  t = 'dk ( ppd   ) + 'vk (   ppd  ) .. p .. p ym + . p . p ym + pp  ym = ( onym )( )sin())(   tm offy ( onym )( )sin())(   tm offy =( ' dk dk ) ( pd p ) + ( ' vk vk ) (   ppd  ) in fact the 5 th element of vector torque ô and 'ô of the motors, y and ' y , can be measured by tension sensors, which can be respectively expressed by: tbta yyy   cossin 11  ; ' y tbta yy   cossin '' ' 1 ' 1  )sin())()((    tmm offony y = tbbjtaaj yyyy   cos)(sin)( 1 ' 151 ' 15 ''  owing to the three equations: t sinp0p  , t cosp0 . p  , t sinp0 2 .. p  , eq. (9) can be rewritten as: .. pp 2   yy mm  = 0 )()( p offyony mm   cos = )( 1 ' 15 ' yy aaj   ; . p ym = 0 )()( p offyony mm   sin = )( 1 ' 15 ' yy bbj   owing to the 3 equations: pym =  ym + y ym  ; .. p ym = . y ym  ; pym =  ym ), eqs. (9) can be rewritten as: . 2 y yy mm   = 0 )()( p offyony mm   cos = )( 1 ' 15 ' yy aaj   (12) 29 the journal of engineering research vol. 7, no. 1 (2010) 22-30 (3) the wire-driven parallel suspension system has opened a new horizon for measuring the static and dynamic derivatives of aircraft in lswt. using the same system based on positon control and force control in robotics allows free flight of the aircraft and obtains the static and dynamic derivatives. however, the results given in this paper can only be considered as a preliminary step in establishing feasibility though the wire-driven parallel suspension system is a very interesting design, and if it is with sufficient may be developed into a rountinely practical system. acknowledgments this research is sponsored by national natural science foundation of china (grant no. 50475099),the youth talents creation foundation of fujian province (grant no. 2006f3083) and the scientific launch foundation for the talents of huaqiao university (grant no. 06bs218). references bennett, r.m., farmer, m.g., mohr, r.l., et al, 1978, "wind-tunnel technique for determining stability derivatives from cable-mounted models," journal of aircraft, vol. 15(5), pp. 304-310. bernhardt, j.e., williams, d.r., 2000, "close-loop control of forebody flow asymmetry," journal of aircraft, vol. 37(3), pp. 491-498. bian, y.z., li, x.r., li, l., et al, 1999, "development of model cable mount and strain gauge balance measuring system in low speed closed wind tunnel," experiments and measurements in fluid mechanics, vol.13(3), pp. 85-90, [in chinese]. cao, y.h. and deng, x.y., 1998, "experimental study of aircraft fuselage aerodynamic forces at high attack angle," acta aerodynamica sinica, vol. 16(4), pp. 394-399, [in chinese]. du, x.q., wang, x.j. and zhou, y., 2003, "investigation of support interference in low speed wind tunnel oscillation tests," experiments and measurements in fluid mechanics, vol.17,no.2,pp.37-40, [in chinese]. griffin, s.a., 1991, "vane support system (vss), a new generation wind tunnel model support system," aiaa91-0398. http://www.onera.fr/dcsd/sacso/index.php. hu, l., 2008, "research on wire-driven parallel suspension systems for low-speed wind tunnels", master's thesis. quanzhou: huaqiao university, october 2008. [in chinese]. lafourcade, p., llibre, m., reboulet, c., 2002, "design of a parallel wire-driven manipulator for wind tunnels," proceedings of the workshop on fundamental issues and future directions for parallel mechanisms and manipulators,quebec city, quebec, pp. 187-194. liang, b., zheng, y.q. and lin, q., 2007, "attitude (13) eqs. (11) and (12) can also be rewritten as eq. (14) and eq. (15) respectively, as follows: (14) (15) in the same way, the dynamic derivatives corresponding to the other 2 single-dof oscillation (roll oscillation and yaw oscillation) are obtained. however, the analysis is just based on the theoretical aspects. the test platform of wdpss-8 for the experiment of dynamic derivatives is still to be built and the precise measuring systems of the vibration angular displacement and the real torque of the motors should be designed and implemented. 4. conclusions and future work the following conclusions can be drawn into after the analysis of the comparision of the 3 suspension systems including the strut suspension system, cable-mounted system and wire-driven parallel suspension system. (1) though strut or sting suspension systems are used in most of low-speed wind tunnels, cable mounted system is also one suitable method for measuring the static derivatives of the aircraft in lswt which allows a large supporting stiffness, small interference of the streamline flow and a high measuring precision for large attack angle, however it can not be used in measuring the dynamic derivatives. (2) for now, the strut suspension systems and rotary balances can be used in measuring the dynamic derivatives of the aircraft in both vertical and horizontal lowspeed wind tunnels successfully, though wire-driven parallel suspension system has a great potentiality, but it is still under investigation. .  ym + y ym  = 0 )()( p offyony mm   sin = )( 1 ' 15 ' yy bbj   . 2 y yy mkm   = ap offyony qsb mm 0 )()(   cos = ap qsb aaj yy 0 1 ' 15 )( '    ( . 2 y yy mkm   ) is called in-phase pitch oscillatory derivatives. v b k a   is called reduced frequency, ab is called mean aerodynamic chord length. v is called free-stream airflow velocity. .  ym + y ym  = kqsb mm ap offyony 0 )()(   sin = kqsb bbj ap yy 0 1 ' 15 )( '    ( .  ym + y ym  ) is called out -of-phase pitch o scillatory derivatives. 30 the journal of engineering research vol. 7, no. 1 (2010) 22-30 sun, h.s., 1999, "the development of 96-test system for measuring dynamic derivatives at high angle of attack," journal of experiments in fluid mechanics, vol. 13(1), pp. 31-37, [in chinese]. wang, y.k., huang, t. and deng, x.y., 2004, "the study of aerodynamic characteristics of after body and effect of cable-mounting," chinese journal of theoretical applied mechanics, vol. 36(3), pp. 257264, [in chinese]. zhang, r.p., wang, x.n. and li, z.x., et al. 2006, "investigation of sting support interference in low speed wind tunnel," journal of experiments in fluid mechanics, vol.20(3), pp. 33-38, [in chinese]. zheng, y.q., lin, q. and liu, x.w., 2007, "initial test of a wire-driven parallel suspension system for low speed wind tunnels," in: proceedings of the 12th world congress in mechanism and machine science, besancon, france, june 17-21, 2007, vol. 5, pp. 8893. zheng, y.q., lin, q. and liu, x.w., 2005, "design methodology of wire-driven parallel support systems in the low speed wind tunnels and attitude control scheme of the scale model," acta aeronautica et astronautica sinica, vol. 26(6), pp. 774-778, [in chinese]. zheng, y.q., 2004, "research on key theoretical issues of wire-driven parallel kinematic manipulators and the application to wind tunnel support systems," phd dissertation, quanzhou:huaqiao university, pp. 4-104, [in chinese]. zheng, y.q., 2006, "feedback linearization control of a wire-driven parallel support system in wind tunnels," proceedings of sixth international conference on intelligent system design and applications, jinan, shandong, china. zheng, y.q., lin, q. and liu, x.w., 2006, "kinematic calibration of a wire-driven parallel support system in wind tunnels," china mechanical engineering, vol.17(6), pp. 551-554. control of the scale model of wire-driven parallel suspension. systems for low-speed wind tunnels," forum on key technique of large aircraft and academic annual meeting of chinese society of aeronautics and astronautics in 2007, shenzhen, china. lin, q., liang, b. and zheng, y.q., 2008, "control study on mmodel aattitude and oscillation by wire-driven parallel manipulator support system for low-speed wind tunnel," journal of experiments in fluid mechanics, vol.22(3), pp. 75-79, [in chinese]. lin, q., zheng, y.q. and liu, x.w., 2006, "modeling and control of a wire-driven parallel support system with large attack angles in low speed wind tunnels," cd proceedings of 25th congress of the international council of the aeronautical sciences, hamburg, germany. liu, t.f., gao, f.x. and lv, w., 2001, "mechanical property calculation of wire-support structure in wind tunnel experiment," mechanics in engineering, vol. 23(1), pp. 46-48, [in chinese]. liu, x.w., q.y., agyemang, b.b., zheng, y.q. and lin, q.,2006, "design of a wire-driven parallel suspension system for wind tunnel based virtual flight testing," proceedings of the 7th international conference on frontiers of design and manufacturing, guangzhou, china. liu, x.w., zheng, y.q. and lin q., 2004, "overview of eire-driven parallel manipulators for aircraft wind tunnels," acta aeronautica et astronautica sinica, vol. 25(4), pp. 393-400, [in chinese]. roos, f.w., 2001, "micro blowing for high-angle-ofattack vortex flow control on fighter aircraft," journal of aircraft, vol. 38(3), pp. 454-457. shen, l.m., shen, z.h. and huang, y., 1988, "a wire type-support system for high angle of attack test in lw speed wind tunnel," experiments and measurements in fluid mechanics, vol. 12(4), pp. 1522, [in chinese]. comp080504.qxd the journal of engineering research vol. 6, no. 2 (2009) 40-46 1. introduction heart failure is a common condition that usually develops slowly as the heart muscle weakens and needs to work harder to keep blood flowing through the body. heart failure develops following injury to the heart such as the damage caused by heart attack, long-term high blood pressure, or an abnormality of one of the heart valves. heart failure is often not recognized until a more advanced stage of heart failure, commonly referred to as congestive heart failure (chf), in which fluid may leak into the lungs, feet, and in some cases the liver or abdominal cavity. physicians often assess the stage of heart failure according to the new york heart association (nyha) functional classification system. this system relates symptoms to everyday activities and the patient's quality of life. according to this system, heart failure is classified into 4 classes (heart failure society of america, http://www.abouthf.org/questions_what_is_hf.htm): _____________________________________________ *corresponding author’s e-mail: abhossen@squ.edu.om 1. class i (mild), symptoms with more than ordinary activity. 2. class ii (mild), symptoms with ordinary activity. 3. class iii (moderate), symptoms with minimal activity. 4. class iv (severe), symptoms at rest. physicians often order a number of tests when exploring a possible diagnosis of heart failure. the most important of these tests is an echocardiogram. echo cardiography is a noninvasive, entirely safe test that uses ultrasound to image the heart as it is beating (congestive heart failure, university of maryland medicine, http:/www.umm.edu/patiented/doc13diagnos.htm). cardiac ultrasounds provide information about: * accurate indications of valve function. * the amount of blood flow through the heart's chambers. * the location of the failure, whether has occurred on the left side, the right side, or both. physicians use information from the echocardiography a pattern recognition technique based on wavelet decomposition for identification of patients with congestive heart failure abdulnasir hossen*a and bader al-ghunaimia department of electrical and computer engineering, college of engineering, sultan qaboos university, p.o. box 33, p.c 123, al-khoud, muscat, sultanate of oman. received 4 may 2008; accepted 15 february 2009 abstract: a pattern recognition technique based on approximate estimation of power spectral densities (psd) of sub-bands resulted from wavelet decomposition of r-r interval (rri) data for identification of patients with congestive heart failure (chf) is investigated. both trial and test data used in this work are drawn from mit databases. two standard patterns of the base-2 logarithmic values of the reciprocal of the probability measure of the approximated psd of chf patients and normal subjects are derived by averaging all corresponding values of all sub-bands of 12 chf data and 12 normal subjects in the trial set. the computed pattern of each data under test is then compared band-by-band with both standard patterns of chf and normal subjects to find the closest pattern. the new technique resulted in an identification accuracy of about 90% by applying it on the test data. keywords: congestive heart failure, pattern recognition, wavelet decomposition, soft-decision, power spectral density êé≤àmc’g ö∏≤dg õéy ≈°vôe øy ∞°ûµ∏d áfõûg äééjƒª∏d ájooîdg ωõ◊g ≈∏y áyrƒÿg iq~≤∏d •é‰cg õ««“ á«æ≤j »ª«æ¨dg q~h h ú°ùm ô°uéædg~ñy áá°°uuóóÿÿggö∏≤dg äé° ñf ∫~©e ò¨àe iqé°t’ äééjƒÿg áfõœ øe áœéædg …ooîdg ∞«£dg ωõm ≈∏y áyrƒÿg á«ñjô≤àdg iq~≤dg áaéãµd •é‰c’g õ««ªàd á«æ≤j oééjg åëñdg gòg øª° àj : é°sée ~¡©e äéfé«h i~yéb øe ¢üëø∏d áe~îà°ùÿg ∂∏jh áhôéà∏d áe~îà°ùÿg äéfé«ñdg äòncg .úª«∏°ùdg ¢uéî°t’g øy ºgõ««“ h êé≤àmc’g ö∏≤dg õéy ≈°vôe øy ∞°ûµdg πlcg øe ≈°vôÿ éªg~mcg ,á«ñjô≤àdg á«ø«£dg iq~≤dg áaéãµd áhƒ°ùùg á«déªàmc’g üƒ∏≤ÿ 2-¢sé°sg »“qézƒd »æëæÿ ú°sé«b ú£‰ ¥é≤à°tg ” .i~ëàÿg äéj’ƒdg ‘ »lƒ∏æµàdg õà«°s ƒ°t 12 o~©d h êé≤àmc’g ö∏≤dg õé©h ¢ jôe 12 o~©d i~fé©dg á«ñjôéàdg äéfé«ñ∏d iôxéæàÿg á«yôødg ωõë∏d iq~≤dg ∫~©e oééjéh ∂dph úª«∏°s ¢uéî°tc’ ônb’g êé≤àmc’g ö∏≤dg õéy êé≤àmc’g ö∏≤dg õéy »àdé◊ ú≤à°ûÿg ú«°sé«≤dg ú£ªædg øe πc ™e áeõëh áeõm é¡àfqé≤e ” ~≤a ¢üëødg â– ádé◊g äéfé«h §‰ üé°ùm ~©h ¬fcéa ¢üëødg aéækcg éecg .º«∏°s ¢üî°t .¢üëødg äéfé«h •é‰cg õ««“ ‘ %90 ¤g π°üj ábo áñ°ùf i~j~÷g á«æ≤àdg √òg â£ycg .é¡©e ∞æ°üàd üôbcg »g áäa ájcg ¤g áaô©ÿ º«∏°ùdg ö∏≤dg h gáá««mmééààøøÿÿgg ääggooôôøøÿÿ .…ooîdg ∞«£∏d iq~≤dg áaéãc ,¿ôÿg qé«ànc’g ,áfõûg äééjƒÿg ,•é‰c’g õ««“ ,êé≤àmc’g ö∏≤dg õéy : 41 the journal of engineering research vol. 6, no. 2 (2009) 40-46 for calculating the ejection fraction (ef), which is the percent of the blood pumped out during each heartbeat. ef is important for determining the severity of heart failure. people with a healthy heart usually have an ef of 50 percent or greater. most people with heart failure, but not all, have an ef of 40 percent or less. the echocardiogram is the most accurate diagnostic test but an expensive one. so there is a need to a noninvasive simple test that helps in identification of patients who need an echocardiogram test. 2. heart rate variability heart rate variability (hrv) is referred to as the beat-to-beat variation in heart rate. instantaneous heart rate is measured as the time in seconds between peaks of two consecutive r waves of the ecg signal. this time is referred to as the rri. the beat-to-beat fluctuations in the rhythm of the heart provide us with an indirect measure of the heart health, as defined by the degree of balance in sympathetic and vagus nerve activity (robert nolan and heart rate variability http://www.biomedical.com/news_display.cfmnewsid=27). hrv analysis serves as a marker for cardiovascular disease because cardiac dysfunction is often manifested by systematic changes in the variability of the r-r interval sequence relative to that of normal controls. the variation of heart rate accompanies the variation of several physiological activities such as breathing, thermoregulation and blood pressure changes (task force of the european society of cardiology and the north american society of pacing and electrophysiology). several hrv abnormalities have been described in patients with chf. it has been shown that patients with heart failure have decreased hrv, particularly those who are at risk of cardiac death (ponikowski et al. 1997). analysis of the hrv using many signal processing techniques can help to detect abnormalities, by manipulating rri data either in the time-domain or in the frequency domain (teich et al. 2001). the frequency spectrum of the rri data is divided into three main bands: 3. rri data groups two different groups of data are used as described below: 3.1 trial data the chf records and normal sinus rhythm (nsr) records are drawn from massachusetts institute of technology (mit) database, which includes 12 records from normal subjects (age 29-64 years) (physionet, the mit-bih normal sinus rhythm database, http://www.physionet.org/physiobank/database/nsrdb) and 12 records from patients (age 22-71 years) with chf (nyha class 3 and 4) (physionet, the bidmc congestive heart failure database, http://www.physionet.org/physiobank/database/chfdb). 3.2 test data this group contains 53 nsr (physionet, normal sinus rhythm rr interval database, http://www.physionet.org/physiobank/database/nsr2db) and 17 chf (physionet, congestive hert filyre rr interval database, heep://www.physionet.org/physiobank/database/ch2db) recordings that are used to test the performance of the classification algorithm. the chf subjects are selected from a larger set that contains 29 long-term recordings. the selected 17 records have chf with nyha class 3 and 4, and the remaining 12 records have nyha class 1 and 2, and therefore have been excluded from the study. the subjects for the selected records are 8 men, aged 39 to 68, and 2 women aged 38 and 59; gender is unknown for the 7 remaining records, but aged between 35 and 64 years. the nsr data of this group contains 53 long-term (about 24 hours) rri records. the subjects are 30 men; aged 28.5 to 76, and 24 women aged 58 to 73. the rri data, for both trial and test groups, are generated from the annotation file, which is associated with each record, using wfdb software (physionet, an nihncrr research resource, http://www.physionet.org/physiotools/wfdb.shtml). in the beginning, all records are truncated to a length of 75821 rr intervals, since the shortest record has such a length. such truncation was also performed by other researchers (teich et al. 2001). it was found in (teich et al. 2001) that with the help of the wavelet transform standard deviation the normal subjects exhibited greater fluctuations at scale window between 16 and 32 samples than those afflicted with heart failure. it was possible to completely discriminate between 12 chf and 12 nsr subjects that represent the trial data in this study. 4. soft-decision technique a method was implemented for screening obstructive sleep apnea and normal controls using the estimated psd by a soft-decision algorithm on decomposed sub-bands (hossen, et al. 2003 and hossen et al. 2005). the same method was also used to screen patients with chf using power spectral ratios lf/hf and vlf/hf as screening parameters on mit trial data only (hossen and alghunaimi, 2004). wavelet decompositions were used instead of subband decomposition to estimate the psd of a signal (hossen 2004). in this work a pattern recognition technique based on psd estimation of wavelet decom• the very low-frequency band (vlf): f ∈ (0.0033 – 0.04) hz. • the low-frequency band (lf): f ∈ (0.04 – 0.15) hz. • the high-frequency band (hf): f ∈ (0.15 – 0.4) hz. 42 the journal of engineering research vol. 6, no. 2 (2009) 40-46 posed sub-bands of rri data is implemented for the purpose of identification of patients having chf from normal controls. a complete analysis is included on both mit trial and test data sets. the work is examined by measuring the classification performances and including a consistency test. 4.1 wavelet-decomposition and band-selection the block-diagram of the wavelet decomposition is shown in fig. 1. the input signal x(n) of length-n is first filtered by lowpass (lpf) and high-pass (hpf) filters and then downsampled by a factor of 2 to produce both the "approximation" a(n) and the "details" d(n). assuming haar-filters are used, a(n) and d(n) can be obtained by: (1) if there is no information about the energy distribution of the input sequence, a band-selection algorithm (hossen and heute, 1993) can be used to decide (as a hard decision) which band is to be computed or kept for more processing. this method depends on the energy comparison between the lowand high-frequency subsequences after the down sampling in fig. 1. (2) according to the sign of b, the decision is taken: if b is positive, the low-frequency band is considered, and if b is negative, the high-frequency band is considered. since we are not interested in the value of b, but only in its sign, a more-simpler equation than eq. 2 can be obtained approximately as (hossen and heute, 1993): (3) 4.2 estimation of power spectral density the decomposition stages in fig. 1 are computed with all branches up to a certain stage m to obtain 2m subbands. all estimator results up to stage m are stored, and a probability measure is assigned to each path (frequency band) to bear the primary information. if j(l) is the assigned probability of the input signal being primarily low-pass, the number j(h) = 1j(l) is the probability that the signal is primarily high-pass. one simple way to make the probability assignments is to use the ratio of the number of positive comparisons between | a(n) | and | d(n) | in eq. (3) to the total number of comparisons for a given stage. at the following stage, the resulting estimate can be interpreted as the conditional probability of the new input sequence containing primarily low (high) frequency components, given that the previous branch was predominantly low (high)-pass. using this reasoning and laws of probability, the assignments for the probability measure of the resulting sub-bands should be made equal to the product of the previous branch probability and the conditional probability estimated at a given stage. fig. 2 shows this step of probability assignment for 8 sub-bands. the probabilities p(bi) derived from the estimator outputs, where i is the index of the band, may be interpreted themselves as a coarse measurement of the psd: the higher the probability value of any band, the higher is its power-spectral content. for m decomposition stages, 2m bands are resulted. each band covers (0.5/2m) hz of the rri spectrum (0-0.5) hz. 5. pattern recognition technique the proposed pattern recognition technique is implemented using the following steps: x(n) lpf hpf v(n) u(n) 2 2 a(n) d(n) figure 1. a single stage wevelet-decomposition 1 2 2 1 2 a( n ) [ x( n ) x( n )]= + + 1 2 2 1 2 d( n ) [ x( n ) x( n )]= − + 1 2 2 2 0 n b ( a( n )) ( d( n )) n − = −∑ = 1 2 0 n sgn( b ) sgn | a( n ) | | d( n ) | n − = −∑ = figure 2. probability assignments of 8 sub-bands 43 the journal of engineering research vol. 6, no. 2 (2009) 40-46 1. the probability measures are computed for each data of the trial set up to stage m to obtain 2m probabilities p(bi), where i is the index of the band. the wavelet filters used are of the type daubechies (db4). the selection of the wavelet filter is a matter of compromise between the complexity and good filter performances. 2. the logarithmic based-2 values of the reciprocal of the probabilities can be obtained as: (4) a staircase approximation of the values of i(bi) can be plotted for all bands. 3. an average plot is found by averaging all i(bi) values of corresponding bands for all 12 chf data and normal data in the trial set to obtain two standard plots for chf and normal as in fig. 3 for 32 sub-bands. a clear increase in the values of i(bi) (clear reduction in the probabilities p(bi) ) in the lf region can be noticed in chf plot compared to that of normal plot. 4. the i(bi) values for each data under test are found for 32 sub-bands. a classification factor cf is then determined as: (5) where ih (bi) and in (bi) are the standard patterns for chf and normal respectively. the rri spectrum in the frequency region (0.0625 hz 0.4375 hz) corresponding to bands 5 to 28 is considered for determining the patterns. the lf and hf regions are used in the recognition, while the vlf and the upper high frequency regions are excluded. depending on the sign of cf, the algorithm can decide whether the data is chf or normal. the data with a negative cf is considered as normal and the data with a positive cf is a chf data. 6. performance evaluation the performance of a classifier is evaluated by three main metrics: sensitivity, specificity, and accuracy as defined below (rangayyan, 2001): sensitivity (%) = tp*100/(tp+fn) (6) specificity (%) = tn *100/(tn +fp) (7) accuracy (%) = (tn +tp)*100/t (8) where, tp, tn, fn, fp are defined as in table 1, and t is the total number of data under test. sensitivity represents the ability of a classifier to detect the positive cases, eg. chf. specificity indicates the ability of a classifier to detect negative cases, eg. normal subjects. accuracy represents the overall performance of a classifier. it indicates the percentage of correctly classified positives and negative cases from the total cases (rangayyan, 2001). index of band (i) 0 a 10 15 20 25 30 35 0 1 2 3 4 5 6 7 8 9 i(b i) normal chf figure 3. standard patterns of chf and normal with 32 sub-bands 1 2i ( b ) log ( )i p( b )i = ( ) ( )( ) ( ) ( )( ) 228 282 5 5 cf i b i b i b i bi h i i n i i i = − − −∑ ∑ = = table 1. the confusion matrix 44 the journal of engineering research vol. 6, no. 2 (2009) 40-46 7. results and discussions figures 4 and 5 show the values of cf for all chf and normal data using 32 sub-bands for trial and test data respectively. the algorithm classifies chf and normal trial data with an accuracy of 100% (fig. 4) and classifies the test data with sensitivity of (13/17) and with a specificity of (50/53), with an overall identification accuracy of (63/70) 90% (fig. 5). 8. consistency of results the leave-one-out method (marques 2002) is used to test the consistency of our results. in this method, which is also called partition method, the available set of data (94 records, 24 trial records (12 chf, 12 normal) and 70 test records (17 chf, 53 normal)) is divided into k=n subsets which rotate in their use of design (trial) and test. the classifier is designed with n-1 subsets and tested on the remaining subset as follows: 1. divide the available data into k=n=5 subsets (groups: g1, g2, g3, g4, g5) of randomly patterns. each subset contains 19 records (6 chf and 13 normal). g3 contains one chf data less than other groups. 2. design the classifier using the first n-1=4 subsets and 0 5 10 15 -15 -10 -5 0 5 10 15 index of data cf chf normal figure 4. classification factor results of trial data with 32 sub-bands 0 10 20 30 40 50 60 -25 -20 -15 -10 -5 0 5 10 15 cf chf normal index of data figure 5. classification factor results of test data with 32 sub-bands 45 the journal of engineering research vol. 6, no. 2 (2009) 40-46 test it on the remaining subset. estimate the performance of the classifier. 3. repeat the previous step rotating the position of the test subset, obtaining there by k=5 estimates. tables 2 shows the results of applying this method on 32 sub-bands. this proves that the results obtained on test data are very consistent. this method was also tested with 16-subband wavelet decomposition. the accuracy result was almost the same with the 32 subband decomposition. for further investigation of the screening capability, three different distance metrics were used: * nmsd: negative mean square distance, which measures the mean square distance of the negative outputs (normal) from the threshold (zero) and is defined as: (9) * pmsd: positive mean square distance, which measures the mean square distance of the positive outputs (chf) from the threshold (zero) and is defined as: (10) * tmsd: total mean square distance, which measures the mean square distance of all outputs (normal and chf) from the threshold (zero) and is defined as: (11) where, cf is the classification factor for 53 (normal) and 17 (chf) test data. those distance metrics are listed in table 3. from this table, it can be concluded that increasing the sub-bands from 16 to 32, improves the results (the distances of classified points from threshold). the pattern recognition method results in identical accuracy values when it is used with 16 or 32 sub-bands, while the consistency of the method is improved as the number of sub-bands increases from 16 to 32. on the other hand the omplexity of the algorithm is increased as the number of sub-bands increases from 16 to 32 subbands. it has been found that the execution time of the complete program of obtaining the pattern of the signal under test and compare it with both standard patterns and finding the result of identification of the subject, is almost doubled when the number of sub-bands increases from 16 to 32, while the time needed only for comparison with both standard patterns and finding the results of identification is almost the same as the number of sub-bands increases from 16 to 32. 9. comparison with other techniques an accuracy of 100% is achieved in (teich et al. 2001) on the same trial data at window scales of 16-samples and 32-samples using some individual wavelets parameters. the test data in our study was also investigated in a timedomain technique in (asyali, 2003), which resulted in an accuracy of 93.2%. both techniques in (teich et al. 2001) and (asyali, 2003) can not be considered as screening techniques because neither of them was implemented on both trial and test data. so their results can be considered as data-dependent. the same drawback is applied on our work in (hossen and al-ghunaimi 2004), since the 100% efficiency was obtained only on trial data. on the opposite, our proposed technique in this paper is implemented on both trial data and test data with an accuracy of 100% and 90% respectively and with a total accuracy of 92.55%. two average standard plots are obtained from trial data and then have been used to identify the test data. 10. conclusions a soft decision algorithm of psd estimation of decomposed wavelet sub-bands is implemented to obtain two standard plots of chf and normal subjects for the purpose of classification between them in a new pattern recognition method. the base-2 logarithmic values of the reciprocal of the estimated psd (with db4 wavelets) have been used to compute the classification factor. the power spectral density of the hrv of chf patients is noticed to be reduced compared to that of normal subject especially at lf region. the frequency range (0.0625 to 0.4375) hz of the psd spectrum of the rri data is used in the recognition. an accuracy of 100% is obtained on 24 mit trail data and 90% on 70 mit test data. references asyali., m.h., 2003, " discrimination power of longterm heart rate variability measures," proceedings of the 25th annual international conference of the ieee, vol. 1, pp. 200-203. congestive heart failure, university of maryland medicine, available at: http://www.umm.edu/patiented/doc13diagnos.html. test group sensitivity specificity accuracy g1 5/6 10/13 15/19 g2 4/6 11/13 15/19 g3 5/5 11/13 16/18 g4 5/6 13/13 18/19 g5 4/6 11/13 15/19 table 2. results of leave-one-out-method using 32 sub-bands 53 2 53 1 i nmsd cf / i = = ∑ = 17 2 17 1 i pmsd cf / i = = ∑ = 70 2 70 1 i tmsd cf / i = = ∑ = sub-bands nmsd pmsd tmsd 16 14.49 15.17 12.38 32 108.85 121.91 58.65 table 3. distance metrics 46 the journal of engineering research vol. 6, no. 2 (2009) 40-46 heart failure society of america, available at: (http://www.abouthf.org/questions_what_is_hf.htm). hossen a., al-ghunaimi b. and hassan, m.o., 2005, "subband decomposition soft decision algorithm for heart rate variability analysis in patients with osa and normal controls," signal processing, vol. 85, pp. 95-106. hossen, a. and al-ghunaimi, b., 2004, " a new method for ccreening of patients with congestive heart failure," international journal of computational intelligence, vol. 1(3), pp. 266-270. hossen, a., 2004, " power spectral density estimation via wavelet decomposition," electronic letters, vol. 40(17), pp. 1055-1056. hossen, a., al-ghunaimi, b. and hassan, m. o., 2003. "a new simple algorithm for heart rate variability analysis in patients with obstructive sleep apnea and normal controls," international journal of bioelectromagnetism, vol. 5(1), pp. 238-239. hossen, a., heute, u., 1993, "fully adaptive evaluation of sb-dft," proceedings of ieee int. symp. on circuits and systems, chicago, illinois, usa. marques, de sa, j.p., 2002, "pattern recognition, concepts, methods and applications," springer verlag. physionet, an nih-ncrr research resource, wfdb software package, available at: http://www.physionet.org/physiotools/wfdb.shtml. physionet, congestive heart failure rr interval database physionet.org/physiobank/database/chf2db/. physionet, normal sinus rhythm rr interval database. physionet.org/physiobank/database/nsr2db/. physionet, the bidmc congestive heart failure database, physionet.org/physiobank/database/chfdb/. physionet, the mit-bih normal sinus rhythm database, physionet.org/physiobank/database/nsrdb/. ponikowski, p., et al. 1997, “depressed heart rate variability as an independent predictor of death in chronic congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy,” am j cardiol, vol. 79, pp. 1645-50. rangayyan, r. m., 2001, "biomedical signal analysis: a case-study approach," ieee press, pp. 466-472. robert nolan and heart rate variability, available at biomedical.com/news_display.cfm?newsid=27. task force of the european society of cardiology and the north american society of pacing and electrophysiology, 1996 heart rate variability, standards of measurements, physiological interpretation, and clinical use, circulation 93, pp. 1043-1065. teich m.c., lowen s.b., jost b.m., vibe-rheymer k., and heneghan c., 2001. heart rate variability: measures and models, nonlinear biomedical signal processing,” dynamic analysis and modeling, ieee press, new york, chapter 6, vo. ii, pp. 159-213. c:\users\raoh\desktop\hadj_paper4.xps the journal of engineering research vol. 7 no. 2 (2010) 59-68 1. introduction a widely used high profile antenna type is the helical antenna. axial-mode helical antenna has become popular due to its low-cost, light weight, wide band, and is used in many applications requiring circularly polarized (cp) radiation patterns and high gain. axialmode helices typically have high gain depending on ___________________________________________ * corresponding author’s e-mail:mbaitsuw@engmail.uwaterloo.ca the number of turns and other factors (kruas 1988). axial-mode helical antennas have their maximum radiation along the axial direction, unlike the normalmode helices where their maximum radiation goes broadside to the helix axis. for high gain applications, multiple antenna elements are employed in an array. enhancement in the performance and capacity of the antenna system can mutual coupling reduction between axial-mode helical antennas using single-negative (sng) magnetic metamaterials m. m. bait-suwailam*a and o.m. ramahi department of electrical and computer engineering, university of waterloo, 200 university ave. w., waterloo, ontario n21, 3g1, canada received 12 may 2009; accepted 19 april 2010 abstract:: this paper investigates the mutual coupling effects between axial-mode helical antenna arrays. single-negative (sng) magnetic metamaterials are developed and used in this work in order to reduce mutual coupling between axial-mode helical antennas used in multiple-input multiple-output systems. the inclusions considered here are composed of broadside coupled split-ring resonators. the magnetic permeability of those sng inclusions have complex effective magnetic permeability response over a frequency band with high negative real part above the resonance and positive real part below the resonance. the sng resonators are inserted between closely-spaced axial-mode helical antenna elements. effective responses of the constitutive parameters of the developed magnetic inclusions are incorporated within the numerical models. it is shown that mutual coupling between the antenna elements can be reduced significantly by incorporating such magnetic inclusions. the sng magnetic resonators work as antenna decoupler, and thus can be applied in a variety of antenna applications. keywords: artificial magnetic materials, axial-mode, helical antennas, metamaterial, and mutual coupling * 60 the journal of engineering research vol. 7 no. 2 (2010) 59-68 be achieved in comparison to a single-input single-output system (siso) (foschini and gans, 1998). such mechanism, or enhancement, is possible if the antenna elements are well-decorrelated since information can be sent through multiple elements at the transmitter and/or the receiver side more efficiently. therefore the performance of antenna arrays or multiple-input multiple-output (mimo) systems can be degraded when the antenna elements are in close proximity. the degradation is due to near-field effects, diffraction from finite-ground planes and strong inductive and capacitive coupling between the elements (balanis, 2005 and vaughan and andersen, 1987). mutual coupling in antenna arrays becomes significant when placing the radiators close to each other. nakano et al. 1984 has investigated the possibility of reducing the mutual coupling between two short (1.5 turns) identical helical antennas by rotating one helix with respect to another. mutual impedance between bifilars of a quadrifilar helix antenna has been previously studied in (amin and chahill, 2005). signal correlation between two normal-mode helical antennas has been studied in (hui et al. 2004). the correlation coefficient was expressed based on the mutual impedances of the antenna system. the performance of mimo antenna elements can be evaluated using the envelope correlation parameter which can be derived from scattering parameters. in many outdoor and indoor wireless communication environments, there are many obstacles and scatterers, like buildings, mobile terminals, and offices...etc. a problem that is encountered in such environments is multipath fading. this fading phenomenon limits the performance and capacity of the antenna system. one of the ways to combat that multipath fading is to use multiple antennas at either transmit, receive or both ends. in mimo systems, more than two antennas are often employed. thus, the correlation between any two antennas within the array is often required and need to be kept as low as possible. in (blanch et al. 2003), the envelope correlation is derived from the scattering parameters of an antenna system. the correlation between different planar inverted f-antenna (pifa) elements has been studied in (thaysen and jakobsen, 2006). it is observed that by using the scattering parameters to compute the envelope correlations in the antenna array is more cost effective in comparison to using the far-field radiation pattern method. to the best of our knowledge, no analysis has been conducted for axial-mode helical antenna arrays incorporating metamaterials as antenna decouplers. in this paper, electromagnetic coupling between closelyspaced axial-mode helical antennas with single-negative (sng) magnetic metamaterials is numerically analyzed and discussed. mutual coupling and envelope correlation between two helical antenna elements are considered. other important antenna parameters such as return loss, axial ratio, and field patterns are also presented. sng metamaterials can have negative permittivity or negative permeability values over a certain frequency band. one way of realizing sng metamaterials is to use split-ring resonators (srr) (pendry et al. 1999 and marques et al. 1996). once these inclusions are excited using proper polarization, an induced electric current develops on the inclusions. if a large number of inclusions are arranged either periodically or aperiodically, the inclusions behave as magnetic dipole arrays. those dipole elements result in having negative effective permeability µeff, over a certain frequency band. this in turn prevents the existence of real propagating modes within the sng metamaterials. thus, it is plausible that an sng material can block electromagnetic energy radiated by one antenna element from being transmitted to a nearby antenna element within an array system. this implies that the ensemble of resonators, comprising the metamaterial, act as a magnetic shielding layer. this paper is organized as follows. section ii gives a conspectus of the theory and motivation behind using single-negative (sng) magnetic metamaterials as magnetic shielding layer. the design methodology for the magnetic materials considered in this work is also discussed, with emphasis on the numerical characterization of the magnetic metamaterials. numerical results are presented and discussed in section iii. brief summary and concluded remarks are given in section iv. 2. single-negative (sng) magnetic metamaterials as magnetic decoupling layer figure 1 shows a classification of materials based on their constitutive parameters, the electrical permittivity , and magnetic permeability µ. in microwave frequencies, and µ are both positive for natural materials, which can be termed double-positive (dps) materials. material with negative permittivity and negative permeability (dng), however, cannot exist naturally over microwave frequencies and engineered artificial materials are used to realize such materials (pendry et al. 1999; marques et al. 1996 and pendry et al. 1996). although there are fundamental differences in the propagation and refraction properties of dps and dng (veselago, 1968), both types of materials can transmit electromagnetic waves. from an electromagnetic interference (emi) point of view, inserting a slab of dps or dng material between two high profile antennas does not have a significant effect. an inter61 the journal of engineering research vol. 7 no. 2 (2010) 59-68 esting phenomenon of strong relevance to the wide range of problems where electromagnetic interference is of concern is the elimination of propagation when one of the constitutive parameters, either eff or µeff, is negative. this can be seen from the second and fourth quadrants of the classification of materials, the -µ domain shown in fig. 1. in such a scenario, transmission of electromagnetic energy would cease since purely evanescent waves exist within the sng medium. therefore, a single-negative (sng) metamaterials would be expected to reduce the mutual coupling and decouple the closely-spaced antenna elements. 2.1 single-negative (sng) magnetic metamaterials design the single-negative (sng) metamaterial considered in this work has a real positive effective electrical permittivity, eff, and a complex effective magnetic permeability, µeff with a negative real part. the magnetic permeability is negative over the frequency band, where we desire decreased coupling between the antennas. in (smith et al. 2000; maslo0vski et al. 2005 and erentok et al. 2005), several types of sng metamaterial designs were discussed. in this work, we consider one of the popular types, namely the rectangular split-ring resonators (srrs). figure 2a shows a unit cell of the single-negative magnetic inclusion. it is figure 1. materials classification based on their constitutive paramaters -µ figure 2. the developed single-negative magnetic metamaterials based on the (bc-srrs): (a) splitring resonator (srr) unit cell with its dimensions, (b) waveguide structure used for the characterization of mng metamaterials. note the e-field points in y-direction, h-field is in z-direction, and the propagation is x-direction 62 the journal of engineering research vol. 7 no. 2 (2010) 59-68 made of two concentric copper rings, each interrupted by a small gap. this capacitive gap in addition to the capacitance between the concentric rings strongly influences the resonance of the srr inclusion. as such, it is possible to realize a resonant magnetic inclusions with dimensions of only /10 of the operating wavelength (ie. in this work, dimensions of roughly /15 are considered as compared to a closed ring which requires a dimension of /4.) if an alternating magnetic field is incident upon the srr plane, the srr behaves as a parallel inductorcapacitor (lc) circuit. this lc circuit exhibits a resonant response at a frequency m=1/ lc, associated with a resonant current circulating around the rings. these resonant circular currents give rise to a resonant magnetic dipole moments, thus srrs inclusions can be characterized by resonant effective permeability, which is usually followed by a negative permeability regime (ie. on the high-frequency side of srrs resonance). this lumped lc behavior is valid since the dimensions of the srr inclusions are much smaller than the operating wavelength. one way to trace the negative permeability regime in srrs or even in metamaterials in general is through transmission measurements or simulations, where this regime appears as a transmission dip. the characterization and retrieval of sng magnetic metamaterials are considered next. the model adopted in our work is based on the retrieval extraction method reported in (chen 2004). basically, one have to compute the scattering parameters of a unit cell srr inclusion, from which the effective refractive index n and impedance z are first computed. the effective permittivity and permeability µ are then calculated from µ = nz and = n/z. in order to numerically characterize the srr inclusions, an air-filled waveguide with a unit cell positioned at the center of the waveguide is used as shown in fig. 2b. notice that the cell is positioned such that the incident magnetic field is perpendicular to the inclusion's trace/surface. the model mimics a transverse electromagnetic mode in both ports of the waveguide (see fig. 2b), with top and bottom sides of the air-filled waveguide assigned as perfect electric conducting (pec) walls (ie. walls orthogonal to the yaxis), while its sides as perfect magnetic (pmc) walls (ie. walls orthogonal to the z-axis). initial design dimensions of the srr inclusion were estimated numerically using the model discussed above. next, the optimized dimensions were obtained such that the inclusion resonance takes place at the middle of the frequency band of interest. the srr rings considered here have equal sides of lengths s = 12 mm. the srr rings with opposite cut openings are etched on the sides of a dielectric substrate (rogers ro4350, r = 3.48, tan = 0.004) having a thickness of 0.762 mm. the cut (gaps) within the metallic rings is 1 mm. the size of the dielectric substrate, l, (see fig. 2b) is 16 mm, which is much less than the operating wavelength at a frequency of 1.24 ghz. the srr rings are made of copper with a thickness of 20 µm. in this work, 8 srr inclusion pairs are etched on both sides of the dielectric substrate. the periodicity of the magnetic inclusions is an important parameter in the design of such inclusions, and it needs to be much smaller than the operating wavelength in order to realize a homogenous medium for the srr resonators. the periodicity, or separation distance between each strip of the developed magnetic inclusions (see fig. 2), corresponds to an electrical thickness of /15 at srr's resonance. figure 3 shows the scattering parameters, s11 and s21, computed using two commercially available fullwave electromagnetic solvers (ansoft.com hfss (ansoft.com), and cst microwave studio (mws) (cst microwave studio, www.cst.com). good agreement between the two solvers is achieved. the real and imaginary parts of the effective electric permittivity and magnetic permeability of the srr resonators are then extracted using the retrieval method discussed above. the extracted parameters are shown in fig. 4. 2.2 effective parameters approach due to computational memory constraints, the effective response of the magnetic inclusions (see fig. 4) is used in order to mimic the response of the µ-negative magnetic inclusions. the artificial magnetic structures are in general inherently dispersive and anisotropic. the mng materials considered in this work are based on the broadside-coupled bc-srrs (marques et al. 2002). the srrs inclusions provide enhanced permeability only in the direction normal to the inclusions plane (z-direction in fig. 2b), and enhanced permittivity in the directions tangent to the plane (xand y-directions). in this work, the orientation of the magnetic inclusions is assumed in xy-plane, as shown in fig. 2b. in such a case, the effective permeability and permittivity of the magnetic structures can take the following tensor forms, respectively. (1) (2) where r represents the permittivity of the host medium ( r =3.48, tan = 0.004 ). it is instructive to note that within the numerical models, the effective permeability in the (x, y, z) direczz o 63 the journal of engineering research vol. 7 no. 2 (2010) 59-68 tions correspond to (1, 1, µzz( )), while the effective permittivity response corresponds to ( xx( ), yy( ), r) in (x, y, z) respectively, as can be seen from (1) and (2). the magnetic sng inclusions are oriented in such a way that the external magnetic field would couple into them, (the external magnetic field is perpendicular to the magnetic inclusions). 3. numerical results figure 5 shows a simple sketch for a typical helical antenna. basically, the antenna is in the form of a helix wire of a certain diameter; however, a strip wire can also be used for ease of integration and miniaturization. the antenna is excited from one end and radiate from the other open end. in this analysis, a strip of width, w = 3 mm is used as the radiating element. the antenna considered here is left-hand circularly polarized (lhcp). a helical antenna is usually specified by its diameter (d), number of turns (n), spacing between each two consecutive turns (s), a pitch angle ( ), and a wire turn length (l). in this work, a helix with 4 turns is considered. the diameter of the helical antenna d = 0.32 0, spacing, s = 0/5, and a circumference of 1 0 is used. figure 6 shows the hfss model for a single axialfrequency (ghz) figure 3. scattering parameters of an sng metamaterials’ unit-cell extracted from the waveguide frequency (ghz) figure 4. effective response for the constitutive parameters of the sng unit cell extracted using the retrieval method 64 the journal of engineering research vol. 7 no. 2 (2010) 59-68 mode helical antenna. a quarter-wavelength transformer was used to match a 50 sma connector to the helix antenna. figure 7 depicts the frequency response of the axial-ratio (ar) for the single helix antenna, where the axial-ratio remains within 3 db over the antenna operating band. thus, it quantifies the circular polarized waves generated out of the antenna. figure 8 shows the model setup used for the coupling analysis between two helical antenna elements separated by a distance of 0.5 0, where single-negative metamaterial strips (sng inclusions) are inserted between the two antennas. a reference geometry comprising two helical antennas spaced by the same distance but without the sng magnetic inclusions is used for comparison. a perfectly conducting (pec) screen is investigated as well in order to quantify the performance of the proposed magnetic shielding screen. note that the width of the pec screen corresponds to the electrical thickness of the developed magnetic strips, o/16. figure 9 shows the computed reflection coefficient, s11, for the two helical antennas in free-space and when using single-negative magnetic inclusions. good impedance matching is observed over a wide frequenfigure 5. design parameters of the helical figure 6. hfss model for a single axial-mode helix antenna frequency (ghz) figure 7. frequency characteristics of axial-ratio in db for the single helix antenna figure 8. hfss model for the two helical antennas with strips of single-negative metamaterials. note that feed center-to-centre distance is o/2 65 cy band of 54% for both cases with and without the magnetic inclusions. the impedance matching deteriorates for the pec screen case over the antennas operating frequency, where total reflection from the pec screen is expected. the mutual coupling, s21, between the two helical antennas with and without the sng magnetic inclusions (see fig. 8) is computed and shown in fig. 10. a reduction of more than 20 db is achieved while creating a slight shift in the resonance frequency (approximately 70 mhz). thus, the coupling between the helical antenna elements has been reduced over a frequency band from 1.26 ghz up to 1.4 ghz. a noticeable dip of s21 is observed at a frequency of 1.31 ghz. moreover, helical antennas show less mutual coupling effects over the frequency band used in this analysis as compared to the monopole antennas studied in (baitsuwailam et al. 2009). nevertheless, the effect of the developed magnetic inclusions in enhancing the gain of the helical antennas is of concern in this work. more isolation between the helical antenna elements is attainable when using the pec screen, as expected, figure 10. magnitude of s21 for the helical antenna system with and without sng metamaterials frequency (ghz) frequency (ghz) 66 the journal of engineering research vol. 7 no. 2 (2010) 59-68 although the use of the sng magnetic inclusions shows higher reduction of mutual coupling around the resonance frequency of 1.31 ghz. next, the performance of antenna arrays applicable to mimo systems is studied. in a rich scattering multipath environment, the correlation between the antenna elements can be evaluated using antennas far-field components (vaughan and andersen, 1987), mutual impedances between the antenna elements derneryd and kristensson, 2004 and hui and lui, 2008), or directly from the scattering parameters measured at the ports (blanch et al. 2003 and thaysen and jakobsen, 2006). under the assumptions of uniform incident waves along the azimuth plane, uncorrelated incident waves impinging from the elevation plane and good impedance matching, the envelope correlation between antenna elements can be calculated from the scattering parameters much more readily than from the far-field patterns. in this work, the use of scattering parameters is incorporated in the envelope correlation computation. the antenna system performance can be determined by the correlation matrix. for better perwell as high isolation between antenna elements. ing parameters as c = i sh and can be represented in a matrix form as where the superscript h denotes the hermitian of the scattering matrix s, and i is the identity matrix. the sented as cii and cij, respectively, where i and correlation cii from antenna i (similarly for cjj cij antenna elements i and j. the envelope correlation e et al. 2003). in the case of a 2 x 2 mimo antenna system (n = 2), the envelope correlation between antenna element i = 1 and j = 2 is given as (6) the envelope correlation between the helical antenna elements has been reduced (see fig. 11) by almost 10 db when using the magnetic inclusions. figure 12 shows the autocorrelation, c11, computed using the antennas scattering parameters. the antenna elements with sng magnetic inclusions gave almost similar correlation to that of the air case. this is attributed to that the antenna elements with and without the sng magnetic resonators have in average low mutual coupling. more reduction in the correlation is observed around 1.31 ghz, where reduction in mutual coupling is observed when using the sng magnetic layer. the cross-correlation, c21, results are shown in fig. 13. the helical antennas with the sng magnetic inclusions are well-decorrelated over a frequency band covfrequency (ghz) figure 11. envelope correlation, e, for the helical antenna system with and without sng air case sng case air case sng case frequency (ghz) figure 12. computed auto-correlation for the helical antenna system with and without sng magnetic metamaterials 67 figure 14 shows the gain pattern in axial direction for the single helix antenna, the two helical antennas with and without the sng magnetic inclusions. a gain of 9.8 db is obtained for the single axial-mode helix antenna in the axial-direction (ie. z-direction in this analysis), while a gain of 12.8 db is obtained for the two helices in free-space. the computed gain for the srr case is 13.0 db. thus, by placing the magnetic inclusions between the helical antenna elements, the gain has been increased, and the mutual coupling between the antenna elements has been reduced by almost 20 db (see fig. 10). conclusions in this paper, mutual coupling between closelyspaced axial-mode helical antennas is investigated with particular focus on multiple-input multiple-outeff, with high negathe developed magnetic inclusions show their two axial-mode helical antenna system has increased as a result of inserting the magnetic resonators between the two radiators. such a decoupling layer is an efficient magnetic shielding wall that can be advantageous in antenna applications and emi problems. acknowledgment this work was supported by sultan qaboos university of oman, research in motion inc., and the national science and engineering research council of canada under the nserc/rim industrial research chair and discovery programs. references amin, m. and cahill, r., 2005. "mutual impedance between bifilars of a quadrifilar helix antenna”, in high frequency postgraduate student colloquium, pp. 66-69. ansoft hfss, http://www.ansoft.com. bait-suwailam, m.m., boybay, m.s. and ramahi, o.m., 2009, "electromagnetic coupling reduction in high-profile antennas using singlenegative metamaterials for mimo applications”, in international conference on communication, computer and power (icccp), pp. 115-120. balanis, c.a., 2005, " antenna theory analysis and design”, john wiley and sons inc. blanch, s., romeu, j. and corbella, i., 2003, "exact representation of antenna system diversity performance from input parameter description”, electron. lett., vol. 39(9), pp. 705-707. chen, x., grzegorczyk, t.m., wu, b.i., pacheco, j. and kong, j.a., 2004, "robust method to retrieve the constitutive effective parameters of metamaterials”, phys. rev. e, vol. 70(1), pp. 016 608.1-016 608.7. cst microwave studio, http://www.cst.com. derneryd, a. and kristensson, g., 2004, "antenna signal correlation and its relation to the air case sng case frequency (ghz) figure 13. computed cross-correlation for the helical antenna system with and without sng magnetic metamaterials figure 14. far-field gain pattern in axial-direction for the single-helix, the two helices with and without the magnetic inclusions, computed at a frequency of 13.1 ghz 68 the journal of engineering research vol. 7 no. 2 (2010) 59-68 impedance matrix”, electron. lett., vol. 40(7), pp. 401-402. erentok, a., luljak, p. and ziolkowski, r., 2005, "characterization of a volumetric metamaterial realization of an artificial magnetic conductor for antenna applications”, ieee trans. antennas propag., vol. 53(1), pp. 160-172. foschini, g.j. and gans, mj., 1998, "on limits of wireless communications in a fading environment when using multiple antennas”, wireless personal communications, vol. 6(3), pp. 311-335. hui, h. and lui, h.s., 2008, "expression of correlation coefficient for two omindirectional antennas using conventional mutual impedances”, electron. lett., vol. 44(2), pp. 1177-1178. hui, h.t., yong, w. and toh, k., 2004, "signal correlation between two normal-mode helical antennas for diversity reception in a multipath environment”, ieee trans. antennas propag., vol. 52(2), pp. 572-577. kraus, j.d., 1988, " antennas”, 2nd ed. mcgraw-hill. marques, r., medina, f. and rafii-el-idrissi, r., 2002, "role of bianisotropy in negative permeability and left-handed metamaterials”, phys. rev. b, vol. 65, pp. 144 440-6. maslovski, s., ikonen, p., kolmakov, i., tretyakov, s. and kaunisto, m., 2005, "artificial magnetic materials based on the new magnetic particle: metasolenoid”, progress in electromagnetic research, vol. 54, pp. 61-81. nakano, h., asaka, n. and yamauchi, j., 1984, "radiation characteristics of short helical antenna and its mutual coupling”, electron. lett., vol. 20(5), pp. 202-204. pendry, j., holden, a., robbins, d. and stewart, w., 1999, "magnetism from conductors and enhanced nonlinear phenomena”, ieee trans. microw. theory tech., vol. 47(11). pp. 20752084. pendry, j.b., holden, a.j., stewart, w.j. and youngs, i., 1996. "extremely low frequency plasmons in metallic mesostructures”, phys. rev. lett., vol. 76(25), pp. 4773-4776. smith, d.r., padilla, w.j., vier, d.c., nematnasser, s.c. and schultz, s., 2000, "a composite medium with simultaneously negative permeability and permittivity”, phys. rev. lett., vol. 84, pp. 4184-4187. thaysen, j. and jakobsen, k., 2006, "envelope correlation in (n, n) mimo antenna array from scattering parameters," microwave and opt. tech. letts., vol. 48(5), pp. 832-834. vaughan, r. and andersen, j., 1987, "antenna diversity in mobile communications," ieee transactions on vehicular technology, vol. 36(4), pp. 149-172, nov. 1987. veselago, v.g., 1968, "the electrodynamics of substances with simulataneously negative values of and µ,", soviet physics uspekhi, vol. 10(4), pp. 509-514. c:\users\raoh\desktop\paper_2.xps the journal of engineering research vol. 8 no 2 (2011) 10-18 1. introduction the study of solitary-wave dates from 1845 when john scott russell (svendssen 2006) reported his experiments. the discovery of the solitary wave have lead to many mathematical investigations that try to understand its properties. in particular, korteweg and ______________________________________ *corresponding author’s e-mail: ksmida@ksu.edu.sa de vries derived their well-known kdv equation for water waves propagating in one direction on shallow water (dean et al. 1991). they also found an exact solution of the kdv equation for a single wave propagating with a uniform velocity without changing formthe solitary wave. cfd analysis of water solitary wave reflection k. smida*a, h. lamloumib, z. hafsiab and k. maalelb adepartment of medical equipments technology, applied medical sciences college, majmaah university, p.o. box 66, al majmaah 11952, kingdom of saudi arabia blaboratoire de modélisation en hydraulique et environnement (lmhe). enit, b.p 37 le belvédère, 1002 tunis, tunisia received 4 april 2011; accepted 19 september 2011 abstract: a new numerical wave generation method is used to investigate the head-on collision of two solitary waves. the reflection at vertical wall of a solitary wave is also presented. the originality of this model, based on the navier-stokes equations, is the specification of an internal inlet velocity, defined as a source line within the computational domain for the generation of these non linear waves. this model was successfully implemented in the phoenics (parabolic hyperbolic or elliptic numerical integration code series) code. the collision of two counter-propagating solitary waves is similar to the interaction of a soliton with a vertical wall. this wave generation method allows the saving of considerable time for this collision process since the counter-propagating wave is generated directly without reflection at vertical wall. for the collision of two solitary waves, numerical results show that the run-up phenomenon can be well explained, the solution of the maximum wave run-up is almost equal to experimental measurement. the simulated wave profiles during the collision are in good agreement with experimental results. for the reflection at vertical wall, the spatial profiles of the wave at fixed instants show that this problem is equivalent to the collision process. keywords: wave generation, head-on collision, solitary waves, source line, wave run-up 11 k. smida, h. lamloumi, z. hafsia and k. maalel the problem of head-on collision of two solitary waves takes its origin from the study of the run-up of these waves on a vertical wall (tsunami case). the tsunamis are the most known for the several types of extreme waves (neelamani et al. 2009) and to simulate a tsunami wave, the best and simplest model is the solitary wave. the study of the dynamics of these waves helps greatly with the protection of coastal upland from erosion which is one of the challenging problems. different types of shore protection structures are in use around the world. indeed, in the absence of viscosity and in the symmetric case, both problems become equivalent. many studies have been conducted on this problem, analytically, numerically and experimentally, including the calculation of the maximum run-up (or superlinear amplitude increase) and the phase shifts. maxworthy (1976) conducted experiments on the collision of two solitary waves; where he showed that the maximum run-up reached a value higher than the sum of waves amplitudes before collision. su et al. (1980) and craig et al. (2006) presented analytical and numerical studies on the head-on collision of two solitary waves and they found that the collision was not elastic, and a small amount of energy was lost by the waves to form secondary waves. wu (1998) presented an analytical study for head-on and following collisions of solitary waves of unequal amplitudes. he distinguished three regimes: the single peak, the double peak and the critical regime. according to lubin et al. (2005), there is a lack of numerical simulation models of the head-on collision waves based on navier-stokes equations. the most known method for linear and non linear wave generation is the internal source region proposed by lin et al. (1999). however, the existing cfd code does not integrate any numerical wave method. power et al. (1984) studied the reflection of a solitary wave on a vertical wall by solving the boussinesq equations analytically as well as numerically. the analytical solution is obtained through an asymptotic matching technique, while the numerical solution is based on a finite difference scheme. they calculated the amplitude of the maximum run-up and the time at which it reached. they found that the incident wave is not reflected immediately as predicted by linear theory. in addition, they showed that the wave undergoes a phase delay during the collision. this phase was found to be inversely proportional to the square root of the initial wave amplitude. cooker et al. (1997) also studied the reflection of a solitary wave on a vertical wall. their numerical approach based on a boundary integral method is used to calculate the fluid potential flow described by the euler equations. they calculated the wall residence time, the time the wave crest remains attached to the wall, a concept introduced for the first time by (temperville 1979). they showed that the wall residence time provides an unambiguous characterization of the wave. they compared their results on the time of attachment and detachment of the wave crest with the asymptotic formulas of (su et al. 1980). further results on the flow were obtained, including the maximum run-up and the instantaneous forces exerted on the wall. their numerical results on the residence time are in accordance with measurements taken from a film on the reflection of a solitary wave on a vertical wall from the experiences of (maxworthy 1976). the proposed numerical wave method is used in the simulation of two-dimensional head-on collision of two solitary waves and the reflection of a solitary wave at vertical wall. it is based on the navier-stokes equations with additional transport scalar equation to describe the free surface evolution (the volume of fluid (vof) equation). the non linear solitary waves are generated by two source lines in internal flow region. at each source line, an internal time dependent inlet velocity is imposed according to the wave characteristics (hafsia et al. 2009). the numerical results are compared to experimental results reported by ming et al. (2003) and non linear wave analytical solution. numerical results given by power et al. (1984) are used to validate the reflection of solitary wave at vertical wall. 2. governing equations and boundary conditions 2.1 transport equations for unsteady flow and incompressible fluid, the mass conservation equation is written in cartesian coordinates: (1) with u and w are the velocity components respectively in the x and z directions. in order to avoid velocity damping in uniform horizontal flow, the damping force in horizontal direction is not considered. the momentum transport equation describing this velocity component is: (2) a dissipation zone is introduced in order to damp the wave at the open boundaries. within this region, a friction source term is added to the momentum trans12 cfd analysis of water solitary wave reflection port equation in the vertical direction: (3) with, p, the pressure; v, kinematic viscosity; , fluid density; g, gravitational acceleration and (x) is a damping function equals to zero except for the added dissipation zone. a linear damping law is adopted: (4) 2.2 free surface treatment the free surface displacements may be treated with a convective transport equation describing the fraction of flow (vof) in each cell of the computation domain: (5) the free surface profile is considered as a twophase flow involving water phase and air phase. we assume that the sliding between the two phases is negligible and that there is no mass exchange across the interface. hence, the velocity field at the free surface is continued. 2.3 initial and boundary conditions the initial condition considered is still water with no wave or current motion. the following boundary conditions are considered when solving the above mentioned transport equations: * for the free surface boundary condition, the normal stress is imposed by setting the pressure p equal to the atmospheric pressure patm (p = patm). * for open boundary condition, a dissipation zone is added in order to avoid wave reflection at each end. within such zone it is advantageous to consider, in addition to the damping friction force, a numerical dissipation by applying coarse grids in the dissipation zones. the neumann boundary condition is specified at the end of each dissipation zone. to generate numerically a given wave, based on the 2-d navier-stokes equations, the source region proposed by (lin et al. 1999) is transformed to source line method (fig. 1). since source region is reduced to a source line, the mass flux per unit time is imposed in terms of vertical velocity wi (t) at the bottom of the internal inlet. the mass source term in the transport equation is then considered as a time dependent inlet boundary condition. 2.4 the internal inlet velocity following to (hafsia et al. 2009), the wave is generated on horizontal source line with vertical pulsating velocity wi (t). the expression of wi (t) at the source line is given by the following time dependent boundary condition: (6) where ls is the length of the internal source line. using this transformation, the mass source region is modeled as an internal inlet imposed at the bottom of the control cell. the velocity wi (t) is upward or downward following that the mass is added wi (t) > 0) or subtracted (wi (t) < 0). the specified internal inlet velocity depends on the desired free surface wave profiles such as linear monochromatic wave or nonlinear solitary wave. assuming that the source line is located at xs = 0, the free surface elevation is function only of the time t. for solitary wave, we have: (7) (a) source region (b) source line wi ls figure 1. internal source for wave generation; (a) the source region method proposed by (lin et al. 1999); (b) the horizontal source line method proposed by (hafsia et al. 2009) 13 k. smida, h. lamloumi, z. hafsia and k. maalel h: wave height; and c the wave celerity given by: (8) the parameter k is given by: (9) the wavelength of a solitary wave is theoretically infinitely long. however, for practical purposes we can define an arbitrarily wavelength as: (10) the apparent wave period is defined by the following ratio: (11) the distance xn is introduced to make the free surface elevation negligible at the initial time and is given by: (12) 3. results and discussions 3.1 introduction two source lines were introduced within the computational domain to generate solitary waves propagating in constant depth (fig. 2). several analytical solutions are suggested in the literature for the head-on according to ming et al. (2003) experiments, two solitary waves were considered with different amplitudes of h1 = 5.70 10-3 m (left wave) and h2 = 5.43 10-3 m (right wave). the still water depth is d = 5.00 10-2 m. the horizontal distance between the two source lines for generating these waves is s12 = 2.00 m. in horizontal direction, the left source line is located at s1 = 6.03 m and the right source at s2 = 8.03 m. these two sources were located at the same elevation from the bottom ds = 0.6 d. the computational domain is discretized into uniform mesh in horizontal direction with x = 2.50 10-2 m. the minimum grid size in vertical direction are chosen near the free surface z = 4.03 10-4 m. the time step is t = 1.33 10-2 s. 3.2 waves generation in order to explain the solitary waves generation, fig. 3 illustrate the free surface profiles for different times. at t = 0 the free surface is horizontal due to that the internal inlet velocity is equal to zero. at t = 0.4 s (fig. 3a), a small free surface elevation is noted. indeed, the internal source perturbation became more influential (figs. 3b and 3c). at t = 0.71 s (fig. 3d), the two waves have remarkable shapes. after t = 0.52 s, the free surface reaches the maximum of amplitude for the two source lines, the potential energy is maximum (fig. 3e). after this instant, the solitary waves separation can be observed (figs. 3f to 3h). at t = 1.400 s, each source line will generate two solitary waves propagating in opposite directions denoted: w1l and w1r for the source s1 and w2l and w2r for the source s2. the generated four solitary waves (fig. 4) are in accordance with the analytical solutions (svendssen 2006). 3.3 head-on collision a sequence of spatial profiles during collision of the two counter-propagating solitary waves is shown in fig. 5. the free surface profiles between the times of t = 1.505 s and t = 1.806 s (figs 5a to 5d) represent the waves propagation before collision. the numerical and experimental results are in a good agreement except some observed discrepancies. for the same head-on collision configuration, based on euler equations, the differences between the predicted results given by (hammack et al. 2004) and experimental data are attributed to the viscous effects and the precision of experimental measurements. at t = 2.307 s, the two solitary waves propagating in opposite directions (w1r and w2l) are merged into a unique wave (fig. 5e). at this instant, the potential energy is maximum (the vertical velocity is negligible), the merging wave seems to be at rest and has a maximum amplitude higher than the sum of both incoming waves. numerical simulation show that the figure 2. computational domain and positions of the two source lines introduced to study the collision of two solitary waves at constant water depth 14 cfd analysis of water solitary wave reflection figure 3. free surface profiles of generated waves at different times: a) t = 0.40 s; b) t = 0.47 s; c) t = 0.52 s; d) t = 0.71 s; e) t = 0.92 s; f) t = 1.00 s; g) t = 1.08 s; h) t = 1.28 s 15 figure 4. comparison between analytical (ooo) and numerical free surface profiles (--) for the two solitary waves w1r and w2l before collision (t = 1.40 s) figure 5. comparison between experimental (ooo) during the head-on collision at different times: a) t = 1.505 s; b) t = 1.603 s; c) t = 1.705 s; d) t = 1.806 s; e) t = 2.307 s (maximum of run-up); f) t = 2.605 s; g) t = 3.004 s; h) t = 3.505 s 16 tary wave at vertical wall is investigated by the source line method. the wall is located at x = 0, the incident solitary wave located far away the wall, is assumed to propathe solitary wave considered has an amplitude of h = 2 10-2 m. the still water depth is h = 2 10-1 m. in horizontal direction, the source line is located at la = 5 m from the vertical wall. the source was located at an elevation from the bottom hs = 0.5h. the computational domain is discretized into uniform mesh in horizontal direction with x = 3.125 10-2 m. the minimum grid size in vertical direction are chosen near the free surface z = 1.1 10-3 m. the time step is t = 1.19 10-2 s. figure 8 represents the reflected wave profile at t = 6.88 s, the reflected wave retrieve its initial shape with small loss of amplitude. in fact, after reflection the solitary wave loses energy to a dispersive wave train and loses height, so ultimately the speed of the reflected wave is smaller than before collision. spatial profiles during reflection of the solitary wave at a vertical wall are shown in fig. 9. from t = 4.74 s to t = 5.17 s (figs. 9a to 9d), the wave runup due to the kinetic energy reduction is observed. once the wave crest reaches the wall, the reflection does not occur immediately as predicted by the classical linear wave theory. the wave crest remains at the wall for a certain time, called the phase delay, to complete the reflection process. during this time delay, the wave amplitude continues to increase. the maximum run-up which is twice greater than the initial wave amplitude is reached at t = 5.31 s (fig. 9e). from this instant to t = 6.02 s (figs. 9f to 9j), the wave amplitude starts to decrease and the wave begins to move away from the wall and propagates in the positive x direction. the numerical profiles agree fairly well with numerical free surfaces profiles given by (power et al. 1984). conclusions we have studied the head collision of two solitary waves with almost equal amplitudes by two internal mass source lines. each source line will generate two solitary waves. the comparison with analytical non linear wave solution shows that the solitary waves are accurately generated. between the two source lines, the two counter-propagating solitary waves are merged into a unique wave with an amplitude higher than the sum of both incoming waves. the navierstokes solution of the maximum wave run-up is almost equal to experimental measurements. after merging, the two separated solitary waves show small differences between their initial amplitude and the simulated one. these differences are attributed to secondary waves. since the system of two head-on colliding waves is equivalent to a single solitary wave hiths figure 6. computational domain and position of the source line introduced to study the figure 7. comparison between numerical results given by (power et al. 1984) (ooo) and free surface profiles ( ) before the reflection of a solitary wave at vertical wall; t = 4.02 s figure 8. comparison between numerical results given by (power et al. 1984) (ooo) and numerical free s urface profiles ( ) fter the reflection of a solitary wave at vertical wall; t = 6.88 s 17 k. smida, h. lamloumi, z. hafsia and k. maalel figure 9. comparison between numerical results given by (power et al. 1984) (ooo) and numerical free surface profiles ( during the reflection of a solitary wave at vertical wall at different times; a) t = 4.74 s; b) t = 4.88 s; c) t = 5.02 s; d) t = 5.17 s; e) t = 5.31 s; f) t = 5.45 s; g) t = 5.60 s; h) t = 5.74 s; i) t = 5.88 s; j) t = 6.02 s 18 cfd analysis of water solitary wave reflection ting a vertical wall, this problem is also treated in this study; using one source line for wave generation. compared to numerical results given by (power et al. 1984), the simulated results are considered satisfactory. as a perspective of this study, we can investigate the head-on and co-propagating collision of solitary waves with unequal amplitudes. also the superposition of other wave types can be treated by this source line generation method. references cooker mj, weidman pd, bale ds (1997), reflection of a high amplitude solitary wave at a vertical wall. j. fluid mech. 342:141-158. craig w, guyenne p, hammack j, hendersen d, sulem c (2006), solitary wave interactions. phys. fluids 18:1-25. dean rg, dalrymple ra (1991), water wave for engineers and scientists. world scientific. hafsia z, ben haj m, lamloumi h, maalel k (2009), internal inlet for wave generation and absorption treatment. j. of coastal engineering 56:951-959. hammack j, henderson d, guyenne p, yi m (2004), solitary-wave collisions. proceedings of the 23rd asme offshore mechanics and arctic engineering (a symposium to honor theodore yao-tsu wu), vancouver, canada. lin p, liu plf (1999), internal wave-maker for navier-stokes equations models. j. of waterway, port, coastal, and ocean engineering 125(4): 207-215. lo dc, young dl (2004), arbitrary lagrangian-eulerian finite element analysis of free surface flow using a velocity-vorticity formulation. j. of computational physics 195:175-201. lubin p, chanson h, glockner s (2010), large eddy simulation of turbulence generated by a weak breaking tidal bore. environ fluid mech. 10:587602. lubin p, vincent s, caltagirone jp (2005), on the navier-stokes equations simulation of the head-on collision between two surface solitary waves. c. r. mecanique 333:351-357. maxworthy t (1976), experiments on the collisions between solitary waves. j. fluid mech. 76:177185. ming yi, haink tu (2003), head-on collision of kdv solitons. http:math.psu.edu/dmh/solitons/. neelamani s, al-salem k, rakha k (2009), extreme gravity waves in the arabian gulf. j. of engineering research 6(1):21-36. power h, chwang at (1984), on reflection of a planar solitary wave at a vertical wall. wave motion 6:183-195 north holland. su cs, mirie rm (1980), on head-on collisions between two solitary waves. j. fluid mech. 98: 509-525. svendsen ia (2006), introduction to nearshore hydrodynamics. world scientific. temperville a (1979), interactions of solitary waves in shallow water theory. arch. mech. 31:177-184. wu ty (1998), nonlinear waves and solitons in water. physica d. 123:48-63. mech080927.qxd the journal of engineering research vol. 6, no. 2 (2009) 52-63 1. introduction since the autonomy of kuwait environment public authority (kuepa) from the jurisdiction of the ministry of health (moh) and the recognition of its legal holding power, pollution prevention in the state of kuwait has _____________________________________________ *corresponding author’s e-mail: ssalem@ns1.prsc.kisr.edu.kw been of a growing concern over urban and suburban areas. decision makers in the state express widespread interest in controlling greenhouse gases (ghgs), volatile organic compounds (vocs) and other airborne pollutants. increased levels of these pollutants, besides eroding materials, are known to have deleterious effects on human health (brunekreef and holgate, 2002, al-salem and bouhamrah, 2006), to cause injury to plants (saitanis et ambient air quality assessment of al-mansoriah residential area in the state of kuwait s.m. al-salem*a, a.a. al-fadhleeb and a.r. khanc *a petrochemical processes program element, petroleum research and studies center, kuwait institute for scientific research, p.o. box. 24885, safat 13109, state of kuwait b planning department, international marketing sector, kuwait petroleum corporation, p.o. box. 26565, 13126, safat, state of kuwait c department of environmental technology management, college for women, kuwait university, p.o. box. 5969, safat 13060, state of kuwait received 27 september 2008; accepted 15 february 2009 abstract : air quality data (ground level ozone (o3), non-methane hydrocarbons (nmhc), methane (ch4), nitrogen oxide (no), nitrogen dioxide (no2), carbon monoxide (co), hydrogen sulfide (h2s) and sulfur dioxide (so2) of a kuwaiti residential area, al-mansoriah, were analyzed to evaluate and determine: (a) the exceedances of air quality from permissible limits set by the kuwait environment public authority (kuepa), (b) the diurnal patterns of air pollutants, (c) the predominant sources of airborne pollutants in the surrounding area, and (d) the ''weekend effect'' on ozone levels. the dataset covered the period of five consecutive years, from january 2000 to december 2004. high levels of ozone were witnessed; recording a number of exceedances. inhabitants of al-mansoriah were exposed, during the period of study, to acute and chronic levels of so2. concerning nmhc, the permissible limit (0.24 ppmrolling average between 6-9 am), was violated in each of the monitoring years. no diurnal pattern showed two distinct strong peaks during the months from october to march, in which the low solar radiation does not cause intense photochemical reactions, which lead to no destruction. a clear h2s source (considered as a primary one) was witnessed from kuwait city (sharq district). the backed up sewage lines and maintenance work coinciding with study period clearly affected almansoriah. a clear "weekend effect", in terms of difference in levels occurring during midweek and weekends, was ascertained. the analysis suggests that al-mansoriah residential area should be considered a nox sensitive region of kuwait. keywords: weekend effect, nox, so2, co, o3, monthly variation, seasonal variation âjƒµdg ádh~h á«æµ°ùddg ájqƒ°üæÿg á≤£æe ‘ »lqéÿg agƒ¡dg ioƒl á°sgqo ¿én øªmôdg~ñy h »∏° ødg ~jéy ,⁄é°ùdg ¿é£∏°s áá°°uuóóÿÿgg~«°ùcg ∫hcg ,¿éã«ÿg aéæãà°séh äéfƒhôchq~«¡dg ´ƒª› ,¿éã«ÿg ,¿hrhc’g réz ä’~©ÿ á檰 àÿg) á«æµ°ùdg ájqƒ°üæÿg á≤£æÿ »lqéÿg agƒ¡dg ioƒl äéfé«h á°sgqo â“ : áeé©dg áä«¡dg oh~m …~©j øe áªléædg äéødéıg o~y (cg:øe πc ~j~ëàd ,(âjèµdg ~«°ùcg êék ,úlhq~«¡dg ~«àjèc ,¿ƒhôµdg ~«°ùcg ∫hg ,úlhî«ædg ~«°ùcg êék ,úlhî«ædg á©héààe ägƒæ°s á°ùªn iîa á«£¨j ” ~bh ,¿hrhc’g réz ≈∏y ´ƒñ°s’g ájé¡f ωéjcg òkéj (o .á≤£æÿg ≈∏y á檫¡ÿg çƒ∏àdg qoé°üe (ê.ákƒ∏ÿg ägré¨∏d ᫪°sƒÿg •é‰c’g (ü ,áä«ñ∏d ä’~©ÿ á≤£æÿg ¿éµ°s ¢vô©j á¶móe ∂dòc ” ~bh .¿hrhc’g réz ¢üîj éã ,äéødéıg øe o~y π«é°ùj ” ~bh .2004 誰ùjo ¤g 2000 ôjéæj úh ée ,á°sgq~dg √òg ‘ á¶móe â“ ~≤a(ájƒ÷g äóyéøàdg ∫~©e ¢véøîfg º°sƒe) ôhƒàcg h ¢sqée ô¡°t úh ée iîødg ‘ ,âjèµdg ~«°ùcc’ »ª°sƒÿg §ªædg ¢üîj éãh .âjèµdg ~«°ùcg êék réz øe á«déy á≤£æe ájqƒ°üæÿg ¿g ¢uóîà°sg øµá á°sgq~dg √òg øeh .úlhq~«¡dg ~«àjèc ré¨d »°ù«fq q~°üªc ¥ô°ûdg á≤£æe ~j~– é° jg ” ~bh .á«fé«ñdg äéeƒ°sôdg ≈∏y ƒ÷g ‘ ré¨dg ∂µøj .úlhî«ædg ~«°séccg ´ƒªû á°sé°ùm áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .¿éã«ÿg aéæãà°séh äéfƒhôchq~«g ´ƒªfi ,úlhî«ædg ~«°ùccg êék ,úlhî«ædg ~«°ùcg ∫hcg ,´ƒñ°s’g ájé¡f ωéjg : 53 the journal of engineering research vol. 6, no. 2 (2009) 52-63 al. 200; saitanis and karandinos, 2002), to reduce crops yield (nali et al. 2002) and to negatively affect ecosystems (bytnerowicz et al. 2002). natural ecosystems adjacent to urban areas are exposed increasingly to air pollutants of urban origin. air pollutants, particularly so2, no2 and the major secondary photochemical oxidant o3 are important threats to plants. their effects comprise many physiological and biochemical changes in plants, which may result in growth reduction and yield loss, even at low levels chronic exposure. concerning human health, tropospheric o3 measured in urban regions and across regional airsheds, are known to affect human health (riga-karandinos and saitanis, 2005). the objective of this work was to investigate the air pollution of a typical residential area in kuwait, namely al-mansoriah. to this purpose we explored the air pollution levels in this area, with emphasis to ground level ozone (o3), non-methane hydrocarbons (nmhc), methane (ch4), nitrogen oxide (no), nitrogen dioxide (no2), carbon monoxide (co), hydrogen sulfide (h2s) and sulfur dioxide (so2). the exceedances of the air quality threshold values were calculated. the diurnal profiles and the seasonal variations of these gases are studied and the relationships between each other were explored. in addition, pollution sources in the surroundings of the area under investigation were identified using a series of concentration and wind roses analyses. the so called ''weekend effect'', concerning the variations of ozone levels between weekdays and weekends, was also investigated. 2. investigated area description al-mansoriah residential area is a luxurious suburb of the city which hosts mainly kuwaiti residents of mid/high class (39,444 residents). figure 1 shows the area from satellite imagery, relative to kuwait city. private and public transport vehicles reach a maximum of 45,000, mostly around the areas main streets. the receptor (monitoring) point chosen for the area was above (at) the polyclinic of al-mansoriah situated near the area's co-op, which is associated with restaurants and other air polluting anthropogenic activities. these pollutants include n-ch4 (resulting from restaurants), nox and co (emissions of automobile and other burning sources) and vocs from the gas and gasoline dispensing station. on the other hand, al-arabi sports club is considered one of the largest sports and family functions facility in kuwait, in terms of children entertainment lounges and other function spaces. cairo st. is situated to the north side of the receptor point. it is one of the busiest streets of kuwait especially during rush hour, linking three main residential areas together in the governorate, ie. al-mansoriah, al-qadisiah and al-daeiah. automobile vehicles operating with gasoline are the main source of pollution associated with traffic jam in the street. rush hours are usually between 7:30 am to 9:00 am (weekdays); and 8:00 pm to 9:30 pm (weekends). 3. data acquirement and methods the dataset of al-mansoriah monitoring station, used in this study, was provided form the kuwait environment public authority (kuepa). the data cover the period from jan 1st 2000 to dec 31st 2004. the station is equipped with a number of air sampling devices and analyzers (thermotecg 08, satcal, al-issa co. ltd.) with accuracy of order 1%, with a measuring frequency of 5 minutes (original data). air probe was approximately 15 m above sea level. all data were stored and manipulated with envidas data acquisition software, which did store up to three months worth of data values. data collected by the station concerned the following pollutants: ch4 (ppm), nmhc (ppm), co (ppm), no (ppb), no2 (ppb), h2s (ppb), so2 (ppb), and o3 (ppb). metrological conditions data were collected through a weather station recording the following parameters: wind speed (ms-1) and direction (o), relative humidity (%) and ambient air temperature (oc). data points were treated and filtered before performing any analysis. filtration procedure was performed as indicated by (al-salem and khan, 2006). no2 and no values exceeding 200 ppb were excluded, in order to eliminate any automobile point source effect on data collection. values of ch4 levels below 1.3 ppm were also removed to avoid insertion of instrumentation signal noise. calibration and span check points were also deleted. many researchers have been occupied with the great number of reactions and types of urban air pollutants (wen and lian, 2009; fenger, 2009; ramanathan and feng, 2009; gildemeister et al. 2007). one of the most important characteristics of urban air ph otochemistry is the oxidation of so 2 and no 2 and their conversion to particulate sulfate 24( so ) − and (gaseous and particulate) nitrates (no 3). moreover, the rate of the conversion of no x to no3 affects ozone formation and the fate of t he nox in the atmosphere. no x is oxidized to nitric acid (hno 3) in the atmosphere, which in turn forms no 3 particles (matsumoto and tanaka, 1996). acidification of precipitation, visibility reduction and deleterious effects on human health and plants, ar e all effects associated with these secondary pollutants. such assessment of urban air quality will surely aid in the planning step of agencies concerned with green industries and keeping up with threshold and permissible limits in populated areas. althou gh air quality was investigated in a number of urban areas in the state of kuwait (al -salem and khan, 2006; al salem et al. 2008a; al -salem et al. 2008b; al -salem et al. 2008c), there is insufficient or lack of information about the air pollution levels in many areas, which host a high percentage of the population of the country. 54 the journal of engineering research vol. 6, no. 2 (2009) 52-63 4. results and discussions 4.1. exceedances of air quality permissible limits/threshold values in order to assess the air quality in al-mansoriah area, measured levels of air pollutants were compared with threshold limits specified in the rules and regulations of kuwait environment public authority (kuepa, 2001) and the european directive standards (eea, 2003a; eea, 2003b). concerning o3, for the protection of both, human health and ecosystem, kuepa rules and regulations set a permissible limit of 80 ppb as an hourly rolling (continuous) average for o3, not to be exceeded twice in a 30 days period. inhabitants of al-mansoriah were exposed to o3 levels above the limit, but only a few days (table 1). in 2000, six exceednaces were recorded based on 8-hrs rolling averages, and one in the year 2004 (table 1). judging from the current dataset of al-mansoriah, o3 levels should be considered potentially phytotoxic in this residential area. investigations in other countries have shown that the nowadays ambient ozone levels can cause injury to plants. for example, (saitanis et al. 2004) used plants of the ozone bioindicator bel-w3 tobacco variety, at 28 sites around a town in greece. in all those sites the used figure 1. al-mansoriah residential area relative to kuwait city. image courtesy of satellite archieves, kgic, kuwait institute for scientific research (kisr) 55 the journal of engineering research vol. 6, no. 2 (2009) 52-63 bioindicator plants were highly injured even after one week of exposure. similar studies, incorporating instrumental monitoring and phytodetection with ozonebioindicator plants, have demonstrated the occurrence of ozone at potentially phytotoxic levels in other rural and forest areas in greece (saitanis and karandinos, 2001; saitanis, 2003; saitanis et al. 2003). kuepa has not yet established any long-term objectives for ozone levels in the near future (5-15 years). long-term objectives have been established for europe after considering the results of the scientific experiments conducted in the framework of the un-ece project over several years in many european countries (rigakarandinos and saitanis, 2005). current permissible limit in europe, has been set to an aot40 value of 3000 ppb*h above which direct adverse effects on trees, crops and natural vegetation may occur (benton et al. 2000). aot40 is the sum of the exceedances of the hourly ozone concentrations above the threshold of 40 ppb, during the daylight hours over a period of three months. the level of 40 ppb (80 µg/m3) of o3 should not be considered as a lower concentration limit in europe for biological effects, since some biological responses may occur below that value (eea, 2003a; eea, 2003b). by comparison with other major european cities, al-manosriah exceeded the number of permissible limit exceedances (8 hr rolling average) in the same duration of monitoring in paris (vardoulakis et al. 2002) and aberdeen (marr et al. 2007). so2 levels in al-mansoriah were generally high, which on the other hand, risks inhabitants of acute and chronic exposure to the pollutant. the kuepa permissible limit (30 ppb) was violated in 2003 (annual threshold) and 5 exceedances were recorded in 2003 based on daily rolling averages. this also was typical of a greek metropolis (volos), were levels were monitored on the basis of a rolling average as well (riga-karandinos and saitanis, 2005). it was also noticed that so2 levels exceeded the european threshold limit for human protection reported previously (eaa, 2003a). concerning no2, 279 average exceedances per year were observed during the five years of monitoring for the short term (1 hr) limit. the corresponding annual permissible limit (30 ppb) was exceeded during the monitoring period (table 1). for carbon monoxide (co) the limit value (8000 ppb) of a daily mean was exceeded 6 times in 2004. pollutant averaging period permissible limit #exceedances /year ozone 8 hrs 60 (ppb) 6/2000 1/2004 ozone hourly 80 (ppb) 6/2000 5/2001 1/2002 1/2003 3/2004 no2 daily 50 (ppb) 95/2000 48/2001 41/2002 93/2003 73/2004 no2 hourly 100 (ppb) 234/2000 237/2001 342/2002 259/2003 323/2004 no2 annual 30 (ppb) 5/(2000-2004) so2 hourly 170 (ppb) 5/2000 8/2002 12/2003 17/2004 so2 annual 30 (ppb) 1/2003 so2 daily 60 (ppb) 5/2003 co daily 8000 (ppb) 6/2004 nmhc 6-9 am 0.24 (ppm) 120/2000 85/2001 178/2002 172/2003 207/2004 table 1. air quality data of al-mansoriah in comparison with the air quality rules and regulations determined by kuepa (kuepsa, 2001) 56 the journal of engineering research vol. 6, no. 2 (2009) 52-63 methane (ch4) and non-methane hydrocarbons (nmhc) levels were monitored as well. kuepa has not yet established any regulation for ch4 in kuwait. it has, however, set a daily permissible limit (0.24 ppm) for the period 6-9 am for nmhc. the analysis of data showed that nmhc recorded 153 exceedances (average) during the five years of monitoring. the number of exceedances increased gradually on an annual basis (table 1). 4.2. diurnal patterns of the pollutants a good way of unraveling the dynamics and patterns of air pollutants is to examine their diurnal pattern. the average diurnal patterns of o3, nmhc, ch4, no, no2, co, h2s and so2 levels for each month of the year were examined (fig. 2). significant variations were observed between the pollutants studied in both, the recorded levels and diurnal patterns. it was observed that, in almansoriah residential area, co, no and no2, exhibit the typical urban daily pattern of primary air pollutants characterized by two peaks, one in the morning and one late in the evening (riga-karandinos and saitanis, 2005). this pattern coincides with the human activities (traffic, open market hours and central heating). so2 exhibited an unstable diurnal fluctuation. so2 levels have been commented to be driven not only for the emissions but also from the levels of o3 and oh (harrison and perry, 1986) and the solar radiation (riga-karandinos and saitanis, 2005). khoder's (2002) results show that the oxidation processes and conversion of so2 to sulfate and of no2 to nitrate (no3) depend on the photochemical oxidation and that the sulfur conversion ratio increases with increasing ozone concentration. the no levels were higher during the months from october to march. during these months, the lower solar radiation intensity does not allow intense photochemical reactions (riga-karandinos and saitanis, 2005), which would lead to no destruction. the concentrations of no, from april to september, were lower than during october to march months. however, during all, high and low no months, the two typical peaks one early in the morning and another in the evening were present the patterns of o3 show a clear seasonal and diurnal variation (increased ozone levels during months from april to september and a strong daytime photochemical ozone buildup due to photooxidation of precursor gasses) showing more intense photochemical processes during the luminous hours of the days and during the high sunlight season (fig. 2). according to (simpson et al. 1995), annual cycle of ozone over mid-latitudes shows mainly two kinds of seasonal behavior, either with a broad summer maximum, generally attributed to local photochemical productions and typical of populated and industrialized areas, or with a spring maximum, of more controversial origin, typical of remote regions and associated to background conditions. the high rush-hour traffic during early morning between 7:30 am to 9:00 am (weekdays); and 8:00 pm to 9:30 pm (weekends), well before sun radiation is strong enough to trigger photochemical reactions of ozone formation leads to ozone destruction shown in fig. 2 as ozone decrease. the witnessed patterns of co were, somewhat, inverse to those of ozone. it was clear that seasonal variations of ozone affected indirectly the seasonal pattern of co. ozone is the photochemical precursor of hydroxyl radicals (oh) in the lower troposphere. reaction with oh removes most gases of natural and anthropogenic origin from the atmosphere (jockel et al. 2003). it is also a fact that the main mechanism for the removal of co from the atmosphere is the reaction of oh. on the other hand, tropospheric oxidation reactions of co in the presence of sufficient no also lead to the production of substantial amounts of ozone. the comparison of the diurnal patterns of o3 and co, suggests that the depletion of o3 and the enhancement of co, during the non-photochemistry hours (early morning and late in the evening), come from the o3 titration when mixed with urban emissions and simultaneously the anthropogenic release of co. this pattern of co may be the typical signature of urban influence (fig. 2). regarding ch4 patterns, a clear increase was witnessed in the early morning hours (5-7 am) associated with restaurants and other gas operated businesses opening hours. the same is clear for non-methane hydrocarbons (nmhc). the levels increase for both pollutants and reach their peak in the three summer break months, ie. july, august and september. this could be attributed to the increasing activities in luxurious areas like almansoriah in the summer time. 4.3. wind and pollutants' concentrations roses in order to investigate the predominant sources affecting al-mansoriah residential area, five years mean wind rose and concentration roses for the air pollutants (o3, nmhc, ch4, no, no2, co, h2s and so2), around the receptor point, (al-mansoriah polyclinic) were analyzed. kuwait is characterized by north/northwestern winds (primary direction) contributing to over 70% of the winds around the year (khan and al-salem, 2007). figure 3 shows the 150 wind speed rose executed for the period of study around al-mansoriah. west/west-northern and north/northwestern winds covering the corridor of wind direction between 285-3300, establish the primary wind direction with average winds exceeding 6 ms-1. all wind and pollutants' concentrations' roses in this study are in a blowing form manner. southeastern winds could be considered as secondary winds ( 6 ms-1 ) in the case of al-mansoriah residential area, as well as the state of kuwait over the past 30 years (al-hajraf et al. 2005). in the case of h2s, the concentration rose reveals a spike value (3450exceeding 5 ppb) in the direction of sharq district in kuwait city (fig. 4). backed up sewage and municipal work has been commenced in the period of study in which pedestrians in the district use to complain 57 the journal of engineering research vol. 6, no. 2 (2009) 52-63 for suffering rotten eggs odor (typical of h2s vented or other type of emissions from municipal manholes) in sharq. other directions on the concentration rose show similar emission strength from wind corridors, which are typical for urban and suburban areas (khan and al-salem, 2007). methane gas followed typical behavior in almost every case (excluding industrial sites), being uniform from all wind directions (fig. 5). ch4 averaged at 2 ppm, which is also typical of kuwaiti urban areas as previously reported in fahaheel (al-salem and khan, 2006; al-salem et al. 2008a) and al-riqa (al-salem et al. 2008b; al-salem et al. 2008c; al-salem and al-fadhlee, 2008). co also showed typical urban behavior and concentration rose trend; being uniform around the receptor point and averaging about 1 ppm (fig. 5). daily activities in almansoriah contribute greatly to both ch4 and co concentration roses. automobile emissions, dry cleaning services, gas dispensing networks and restaurant emissions; all contribute to both pollutants. sources of these emissions 1.9 2.0 2.0 2.1 2.1 2.2 2.2 2.3 2.3 2.4 2.4 1 6 11 16 21 hour c h 4 (p pm ) jan feb mar apr may jun july aug sep oct nov dec 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1 6 11 16 21 hour co (p pm ) 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1 6 11 16 21 hour nm hc (p pm ) 15 25 35 45 55 65 75 85 1 3 5 7 9 11 13 15 17 19 21 23 hour n o 2 (p pb ) 3 5 6 8 1 5 9 13 17 21hour h2 s (p pb ) 0 10 20 30 40 50 60 70 80 1 6 11 16 21hour no (p pb ) 0 5 10 15 20 25 30 35 40 45 1 6 11 16 21 hour o3 (p pb ) 14 19 24 29 34 39 1 6 11 16 21 hour so 2 (p pb ) figure 2. plots of monthly diurnal patterns of air pollutants in al-mansoriah 58 the journal of engineering research vol. 6, no. 2 (2009) 52-63 figure 3. plot (rose) of the average wind speed (ms-1) monitored in al-mansoriah during the period (january 2000 to december 2004) figure 4. plot (rose) of the average h2s (ppb) monitored in al-mansoriah during the period (from january 2000 to december 2004) figure 5. plot (rose) of average ch4, nmhc and co concentration rose (ppm) monitored in al-mansoriah during the period (from january 2000 to december 2004) 59 the journal of engineering research vol. 6, no. 2 (2009) 52-63 are distributed around the receptor point, which explains the trends left by both ch4 and co on the roses. regarding nmhc, the concentration rose shows a peak in the emission strength in the northwestern direction (fig. 5). figure 6 shows the concentration rose for no and no2 gases around the receptor point. northern winds (covering 345-150) were clearly the richest in both pollutants (exceeding 60 ppb for no2 and 70 ppb for no). this direction corresponds to kuwait city, which hosts a number of combustion sources that contribute to no and no2 levels in the ambient. opposite to nox, winds blowing from northwestern and eastern directions showed higher ozone levels (fig. 7). nox titration and other sets of reactions fade ozone from the lower atmosphere, which also explains the no primary direction of north. so2 was present around the receptor point in an almost equal level of 25 ppb (fig. 8). 4.4. weekend/ weekday variations during the past decade, weekend/weekday differences in levels of air pollutants (especially of o3) have been a subject of research interest (marr and harley, 2002; vandaele et al. 2002; fujita et al. 2003, riga-karandinos and saitanis, 2005). although, weekly changes in the emissions caused by human activities are known to affect the weekly cycle of ambient pollutant concentrations, this emission-concentration relationship, in urban, suburban and rural sites, is not well elucidated as it is reflected in controversial reports (blier and winer, 1999; pun et al. 2003). in the state of kuwait, during the period of the study, the weekend was on thursdays and fridays (nowadays the weekend has shift to fridays and saturdays). in weekends (thursdays and fridays), the emissions of anthropogenic pollutants and especially of o3 precursors, are believed to be lower compared to those occurring during weekdays, because car traffic is lower and several polluting plants and factories may be less active or inactive. many approaches have been used for the study of the socalled weekend effect, meaning the occurrence of increased o3 levels and decreased nox during weekends in comparison to weekdays. near constant o3 levels, during weekdays, in the face of strong weekday/weekend variations in nox, should also be viewed as a weekend effect (fujita et al. 2003). after taking into account the shift in the relaxing days and the days of high anthropogenic activities in kuwait, in comparison to the west countries, in this study, we analyzed the monthly averages of mean hourly o3 and no concentrations during sunday-tuesday vs. the corresponding concentrations on wednesday, thursday, friday and saturday, recorded in al-mansoriah, (fig. 9). in table 2, the relevant linear regression equations and the corresponding correlation coefficients are given. concerning o3, the slops of linear regressions revealed that wednesday concentrations were very close to those occurring in during sun-tue (fig. 9; table 2). unexpectedly, thursdays-saturdays o3 concentrations were much lower than sun-tue concentrations. regarding no, during wednesdays and thursday the slops were clearly close to one indicating similar levels occurring during these two days to those occurring during sun-tue. on the other hand, no levels were clearly much lower on fridays than ones occurring during sun-tue (fig. 9; table 2). elevated nox levels can always inhibit o3 production either by no titration of o3 or by reaction of no2 and oh to form hno3 -nitric acid(riga-karandinos and saitanis, 2005). synoptically, o3 decreases more rapidly with reductions in nox emissions, than with reductions in vocs (volatile organic compounds) in nox-sensitive regions (ie. abundant vocs relative to nox). while in vocs-sensitive regions (ie. abundant nox relative to vocs) a decrease in nox emissions may cause an increase in o3 because of reduced titration or reduced formation of hno3. locations where ozone increases during weekends, while nox emissions are significantly lower should be considered vocs-limited (blanchard and tanenbaum, 2003). thus, in the case of al-mansoriah residential area, signs of being a nox sensitive region are very clear although nox levels are not always higher on weekends, suggesting a strongly influential background effect. the reduction of anthropogenic emissions during weekends is expected to reduce proportionally equally anthropogenic vocs (avocs) and nox. in addition to the above-discussed role of vocs in the ''weekend effect'', it is the high early morning rush-hour traffic on weekends that allow ozone concentrations to rise from a significantly lower concentration and therefore reaching lower overall weekend concentrations. a number of studies demonstrate a wide range in the weekday/weekend o 3 behavior. marr and harley (2002) suggest that no x concentrations are consistently lower on weekends relative to weekdays by approximately 30 –40%, while voc changes are less pronounced ( approximately 10%). the explanation of weekday/weekend variation, due to no x reduction, is supported by a wide range of analyses of ambient data and several photochemical studies, while changes in the timing and location of emissions and meteorological factors play sma ller roles in weekend o 3 behavior (heuss et al. 2003; riga-karandinos and saitanis, 2005). analyses of ambient air quality data and emissions forecasts for weekdays and weekends may improve considerably our understanding of the effects of control strategi es and of future changes in emissions on future ambient o 3 concentrations. 60 the journal of engineering research vol. 6, no. 2 (2009) 52-63 figure 6. plot (rose) of average no and no2 concentration rose (ppb) monitored in al-mansoriah during the the period (from january 2000 to december 2004) figure 7. plot (rose) of average o3 concentration rose (ppb) monitored in al-mansoriah during the period (from january 2000 to december 2004) figure 8. plot (rose) of average o3 concentration rose (ppb) monitored in al-mansoriah during the period (from january 2000 to december 2004) 61 the journal of engineering research vol. 6, no. 2 (2009) 52-63 5. conclusions after analyzing the dataset obtained for al-mansoriah residential area, a number of exceedances were recorded, violating kuepa permissible limits of air quality. concerning the exceedances of o3, there were no exceedances violating annual or daily limits. however, inhabitants of al-mansoriah were exposed to acute so2 levels. recorded no2 annual, daily and hourly exceedances in the five years of monitoring. the diurnal patterns of the studied pollutants were investigated and revealed a seasonal variation in ozone levels. an inverse pattern was witnessed for co. on the other hand, methane patterns showed an increase in the early morning hours associated with the anthropogenic activates in the area. series of concentration roses were executed around the receptor point (ie. polyclinic) to determine the influential pollution sources around al-mansoriah. methane and co showed typical urban area patterns, while no and no2 concentration roses revealed a primary source situated in northern wind corridor (covering 345-150). weekend effect analysis revealed that the area under investigation was a nox sensitive one. the causes of the weekend o3 effect are probably the weekend/weekday differences in nox emissions and the complex non-linear photochemistry of ozone. the data of this study would be useful for future comparisons, after the operation of the new infrastructures at kuwait city (sharq district), which certainly will cause changes in pollutants levels. it is obvious that, for future air quality management strategies, the development of accurate, temporally and spatially resolved dayof-week emission inventories, including separate inventories for weekends can promote a better understanding. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 midweek w ed o3 no 0 10 20 30 40 50 60 0 10 20 30 40 50 60 midweek th ur 0 10 20 30 40 50 60 0 10 20 30 40 50 60 midweek fr i 0 10 20 30 40 50 60 0 10 20 30 40 50 60 midweek sa t figure 9. regression plots of the monthly averages of the mean hourly ozone (dashed line) and no (solid line) concentrations during midweek (sun-tues) vs. the corresponding concentrations of the rest of the week day/ (pollutant) wednesday/ o 3 wednesday/ no y = 0.91x y = 1.02x r = 0.28 r = 0.24 thursday/ o 3 thursday/ no y = 0.78x y = 0.97x r = 0.40 r = 0.87 friday/ o 3 friday/ no y = 0.54x y = 0.77x r = 0.64 r = 0.67 saturday/ o 3 saturday/ no y = 0.48x y = 1.12x r = 0.46 r = 0.54 table 2. regression equations of mean hourly ozone concentrations during (sun-tues) vs. the corresponding mean hourly ozone concentratopns on wednesday, thursday, friday, and saturday (x) 62 the journal of engineering research vol. 6, no. 2 (2009) 52-63 references al-hajraf, s., al-ajmi, d., khan, a., tang, h., wahner, a. and klemp, d., 2005, "air quality assessment of ali sabah al-salem urban community," kisr final report #ec031c. al-salem, s.m. and al-fadhlee, a.a., 2008, "source identification of selected airborne pollutants using concentration and wind roses simulated around alahmadi governorate in the state of kuwait," in: proc 3rd international conference on scientific computing to computational engineering (3rd ic-scce), athens, greece. al-salem, s.m. and bouhamrah, w.s., 2006, "ambient concentrations of benzene and other voc's at typical industrial sites in kuwait and their cancer risk assessment," research journal of chemistry and environment, vol. 10(3), pp. 42-46. al-salem, s.m. and khan, a.r., 2006, "methane dispersion modeling and source determination around urban areas in kuwait," in: proc 1st international conference & exhibition on green industry, manama, bahrain. al-salem, s.m., al-fadhlee, a.a. and al-haddad, a.a., 2008b, "monitoring hydrocarbons levels in al-riqa urban area (kuwait) using modern environmental techniques: investigating the effect of petroleum downstream industries on the outdoor air quality," in: proc (recshow '08) middle east recycling, waste and environmental management exhibition & congress, dead sea, jordan, 17th-20th february, pp. 48. al-salem, s.m., al-fadhlee, a.a., al-haddad, a.a. and khan, a.r., 2008c, "identifying the contribution of air pollution sources in al-riqa urban area (kuwait) using chemical mass balance (cmb) modeling," in: proc climate change, energy and environ management conference, johannesburg, south africa, 23rd-25th march. al-salem, s.m., al-haddad, a.a. and khan, a.r., 2008a, "primary pollutants monitoring and modeling using chemical mass balance (cmb) around fahaheel residential area," american journal of environmental sciences, vol. 4(1), pp. 13-21. benton, j., fuhrer, j., gimeno, b.s., skarby, l., palmerbrown, d., ball, g., roadknight, c. and mills, g., 2000, "an international cooperative programme indicates the widespread occurrence of ozone injury on crops," agriculture, ecosystems and environment, vol. 78(1), pp. 19-30. blanchard, c.l. and tanenbaum, s.j., 2003, "differences between weekday and weekend air pollutant levels in southern california," journal of air & waste manage association, vol. 53(3), pp. 816-828. blier, w. and winer, a.m., 1999, "analysis of weekday/weekend differences in ambient air quality and meteorology in the south coast air basin," final report; contract no. 95-334. california air resources board, sacramento, ca. brunekreef, b. and holgate, s.t., 2002, "air pollution and health", lancet, vol. 360(9341), pp. 1233-1242. bytnerowicz, a., godzik, b., fraczek, w., grodzinska, k., krywult, m., badea, o., barancok, p., blum, o., cerny, godzik, m., mankovska, s., manning, b., moravcik, w., musselman, p., oszlanyi, r., postelnicu, j., szdzuj, d., varsavova, j. and zota, m., 2002, "distribution of ozone and other air pollutants in forests of the carpathian mountains in central europe," environment pollution, vol. 116(1), pp. 3-25. eea, 2003a, "exposure of population to exceedances of eu air quality standards (2002)," available online (june 2004). eea, 2003b, "air pollution by ozone in europe in summer 2003-overview exceedances of ec ozone threshold values during the summer april-eu 1999," council directive 1999/30/ec of 22 april 1999 relating to limit values for sulphur dioxide, nitrogen dioxide and oxides of nitrogen, particulate matter and lead in ambient air, official journal of the european communities 29.6.1999, l 163/41. fenger, j., 2009, "air pollution in the last 50 years from local to global," atmospheric environment, vol. 43(1), pp. 13-22. fujita, e.m., stockwell, w.r., campbell, d.e., keislar, r.e. and lawson, d.r., 2003, "evolution of the magnitude and spatial extend of the weekend ozone effect in california south coast air basin, 19812000," journal of air & waste manage association, vol. 53, pp. 802-815. gildemeister, a.e., hopke, p.k. and kim, e., 2007, "sources of fine urban particulate matter in detroit, mi," chemosphere, vol. 69(7): 1064-1074. harrison, r.m. and perry, r., 1986, "handbook of air pollution analysis," second ed. chapman and hall, london, new york. heuss, j.m., kahlbaum, d.f. and wolff, g.t., 2003, "weekday/weekend ozone differences: what can we learn from them," journal of air & waste manage association, vol. 53(4), pp. 772-788. jockel, p., brenninkmeijer, c.a.m. and crutzen, p.j., 2003, "a discussion on the determination of atmospheric oh and its trends," atmospheric chemistry & physics, vol. 3(2), pp. 107-118. khan, a. and al-salem, s., 2007, "seasonal variation effect on airborne pollutants in an urban area of the state of kuwait," journal of environmental research & development, vol. 1(3), pp. 215-218. khoder, m.i., 2002, "atmospheric conversion of sulfur dioxide to particulate sulfate and nitrogen dioxide to particulate nitrate and gaseous nitric acid in an urban area," chemosphere, vol. 49(6), pp. 675-684. kuepa rules & regulations, 2001, "kuwait al-youm appendix 533," kwt gov. press, law 210/2001. 63 the journal of engineering research vol. 6, no. 2 (2009) 52-63 marr, i.l., rosser, d.p. and meneses, c.a., 2007, "an air quality survey and emissions inventory at aberdeen harbour," atmospheric environment, vol. 41 (1), pp. 6379-6395. marr, l. and harley, r.a., 2002, "spectral analysis of weekdayweekend differences in ambient ozone, nitrogen oxide and non-methane hydrocarbons time series in california," atmospheric environment, vol. 36(14), pp. 2327-2385. matsumoto, k. and tanaka, h., 1996, "formation and dissociation of atmospheric particulate nitrate and chloride: an approach based on phase equilibrium," atmospheric environment, vol. 30(4), pp. 639-648. nali, c., pucciariello, c. and lorenzini, g., 2002, "ozone distribution in central italy and its effects on crops productivity," agriculture, ecosystems and environment, vol. 90(3), pp. 277-289. pun, b.k., seigneur, ch. and white, w., 2003, "day-ofweek behavior of atmospheric ozone in three us cities," journal of air & waste manage association, vol. 53(3), pp. 789-801. ramanathan, v. and feng, y., 2009, "air pollution, greenhouse gases and climate change: global and regional perspectives," atmospheric environment, vol. 43(1): 37-50. riga-karandinos, a.n. and saitanis, c., 2005, "comparative assessment of ambient air quality in two typical mediterranean coastal cities in greece," chemosphere, vol. 59(8), pp. 1125-1136. saitanis, c.j. and karandinos, m.g., 2001, "instrumental recording and biomonitoring of ambient ozone in greek countryside", chemosphere, vol. 44(4), pp. 813-821. saitanis, c.j. and karandinos, m.g., 2002, "effects of ozone on tobacco (nicotiana tabacum l.) varieties," journal of agronomy and crop science, vol. 188(1), pp. 51-58. saitanis, c.j., 2003, "background ozone levels monitoring and phytodetection in the greater rural area of corinth-greece," chemosphere, vol. 51(5), pp. 913-923. saitanis, c.j., karandinos, m.g., riga-karandinos, a.n., lorenzini, g. and vlassi, a., 2003, "photochemical air pollutant levels and ozone phytotoxicity in the region of mesogia-attica," greece, international journal of environmental pollution, vol. 19(2), pp. 197-208. saitanis, c.j., katsaras, d.h., riga-karandinos, a.n., lekkas, d.b. and arapis, g., 2004, "evaluation of ozone phytotoxicity in the greater area of a typical mediterranean small city (volos) and in the nearby forest (pelion mt.)-central greece," bulletin of environmental contamination and toxicology, vol. 72(6), pp. 1268-1277. saitanis, c.j., riga-karandinos, a.n. and karandinos, m.g., 2001, "effects of ozone on chlorophyll and quantum yield of tobacco (nicotiana tabacum l.) varieties," chemosphere, vol. 42(17), pp. 945-953. simpson, o., guenter, a., hewitt, c.n. and steinbrecher, r., 1995, "biogenic emissions in europe i: estimates and uncertainties," journal of geophysical research, vol. 100(d11), pp. 22875-22890. vandaele, a.c., tsouli, a., carleer, m. and colin, r., 2002, "uv fourier transform measurements of tropospheric o3, no2, so2, benzene and toluene," environmental pollution, vol. 116(12), pp. 193-201. vardoulakisa, s., gonzalez-flescaa, n. and fisher, b.e.a., 2002, "assessment of traffic-related air pollution in two street canyons in paris: implications for exposure studies," atmospheric environment, vol. 36(1), pp. 1025-1039. wen, y. and lian, z., 2009, "influence of air conditioners utilization on urban thermal environment," applied thermal engineering, vol. 29(4), pp. 670-675. c:\users\raoh\desktop\paper_8.xps the journal of engineering research vol 8 no. 1 (2011) 74-82 1. introduction: the expensive development of wireless communication systems during the last decade has particularly put the design of radio frequency power amplifiers in focus. in general, a power amplifier is designated as the last amplifier in a transmission chain as shown in fig. 1. it is the amplifier stage that typically requires _______________________________________ *corredsponding author’s e-mail: nadir@squ.edu.om the most attention to power efficiency. handsets are battery-operated devices and their talk-time will directly depend on the efficiency of the power amplifier. efficiency and linearity are opposing requirements in the power amplifiers design and much research is focused on how to improve the efficiency class-e amplifier design improvements for gsm frequencies z. nadir* and f. touati department of electrical and computer engineering, college of engineering, sultan qaboos university, p.o. box 33, p.c. 123, al-khodh, muscat, sultanate of oman received 8 september 2009; accepted 8 september 2010 abstract: efficient power amplifiers are essential in portable battery-operated systems such as mobile phones. also, the power amplifier (pa) is the most power-consuming building block in the transmitter of a portable system. this paper investigates how the efficiency of the power amplifier (which is beneficial for multiple applications in communcation sector) can be improved by increasing the efficiency of switching mode class e power amplifiers for frequencies of 900 mhz and 1800 mhz. the paper tackles modeling, design improvements and verification through simulation for higher efficiencies. this is the continuation of previous work by the authors. these nonlinear power amplifiers can only amplify constant-envelope rf signals without introducing significant distortion. mobile systems such as advanced mobile phone system (amps) and global system for mobile communications (gsm) use modulation schemes which generate constant amplitude rf outputs in order to use efficient but nonlinear power amplifiers. improvements in designs are suggested and higher efficiencies are achieved, to the tune of 67.1% (for 900 mhz) and 67.0% (1800 mhz). keywords: class-e power amplifiers, switching mode amplifier, high efficiency, matching network, switching losses 75 the journal of engineering research vol 8 no. 1 (2011) 74-82 of power amplifiers circuits while still satisfying the linearity requirements of a given system (su. et al. 1998 and lee et al. 2004). the need for linear power amplifiers arises in many radio frequency applications. sdr, cognitive radio etc. at present, most linear power amplifiers designed for portable devices, employ a class a output stage and exhibit efficiencies around 30% to 40% (b. razafi et al. 1998). to achieve a higher efficiency, it is possible to begin with a nonlinear power amplifier and apply linearization techniques to the circuit. this paper focuses on improving the efficiency of class-e power amplifiers by reducing the switching losses. the approach proposes a better biasing section of the transistor for a faster switching and optimizes the power transferred to the load through a well-optimized matching network. 2. theory and system design sokal was the first to introduce the highly efficient class-e amplifier (sokal et al. 1975) which achieve significantly higher efficiency than conventional class b or c amplifiers which represents an attractive solution for portable radio devices enabling longer operation times. the power consumed by the transistor is low during the rf period because of switching transitions which are an effective part in meeting high-efficiency requirements. designers make on/off switching transitions as fast as practical with high efficiency techniques which must be accommodated. the high voltage -current product during the switching transitions can also be avoided by using the following strategies: 1. the rise of transistor voltage is delayed until the current is reduced to zero. 2. the transistor voltage returns to zero before the current begins to rise. as a result, waveforms of the high voltage and high current never exist simultaneously (berglund et al. 2006). 2.1 design analysis at 900 mhz figure 2 shows the basic circuit of a class e power amplifier where the ams cmos 0.35 µm process transistor parameters are used (nadir et al. 2009). the circuit is the basic topology of a class e power amplifier circuit which comprises a switch (ie. transistor m1), a grounded capacitor c2, and a series network c1 and l2. l1, which is the radio-frequency choke (rfc). it has high impedance at the frequency of operation and c2 includes the drain junction capacitance of m1. the values of c1, c2, l2, and rl are selected such that the power switching losses of m1 are reduced, which would increase the power efficiency. we assume the following: a center frequency, ƒo = 900 mhz, pout = 20 dbm (100 mw) which is available across the load resistor, vcc = 2.5 v (the power supply) and rl = 50 , which is considered as antennas input impedance. for optimum power of a class e power amplifier, the optimum resistance, rlopt comes out as (lee et al. 2004): figure 1. classic architecture of a transmission channel source transmitter receiver destination noise source figure 2. basic circuit of class e power amplifier 76 the journal of engineering research vol 8 no. 1 (2011) 74-82 (1) rl should be down-converted from 50 to 36.06 using a matching network as shown in fig. 3 for optimum power delivery. where, to convert 50 to 36.06 , let us use rs=36.06 and rp= 50 so: o = 2 ƒo = 5.65 x 109 rad / s (2) (3) (4) (5) the total input impedance looking from the left of the circuit shown in fig. 3 is: (6) as can be seen, for a frequency of 900 mhz, the real part of the impedance is 36.05 whereas the imaginary part is 0.02489 which justifies our previous calculations. theoretically the higher the value of q, the better in terms of selectivity, but since a higher value of q requires a larger inductor this is not preferable for portable radio devices because size is a major issue. the quality factor (or q) of an inductor is the ratio of its inductive reactance to its resistance at a figure 3. down convert matching network matchl matchmatchmatchll total (a) (b) figure 4. (a) real and (b) imaginary parts of t he impedance against ffrequency at 900 mhz 77 the journal of engineering research vol 8 no. 1 (2011) 74-82 given frequency, and is a measure of its efficiency. the q factor is a widespread measure used to characterize resonators. it is defined as the peak energy stored in the circuit divided by the average energy dissipated in it per cycle at resonance. low q circuits are therefore damped and lossy and high q circuits are under-damped. as q is related to bandwidth; low q circuits are wide band and high q circuits are narrow band. in our case, quality factor for the series resonance branch (c1, l2), as shown in fig. 2, is assumed as 5 for a reasonable-sized inductor. it follows that: (7) (8) (9) l1 is chosen to be 30 uh since it must have a sufficiently large impedance at the center frequency. table-1 summarizes the designed reactive elements of the matching network for 900 mhz and 1800 mhz networks (similar design steps are carried out for 850/1900 mhz circuits but is not shown in the current paper). 3. design procedure and results discussion simulation is done for a class e power amplifier for both frequencies 900 mhz and 1800 mhz (dhawyani et al. 2008). 3.1 design results for 900 mhz and 1800 mhz figure. 5 shows the circuit of a class e power amplifier (only for 900 mhz is shown). a sinusoidal input with an amplitude of 1 v, at 900 mhz, is considered. the results show that the output power which is causing low efficiency needs to be improved. these are clearly shown through the graphs. the time delay as visible in graphs suggests faster switching. we focused on improving the efficiency of class-e power amplifiers by reducing the switching losses. the approach proposes a better biasing section of the transistor for a faster switching and optimizes the power transferred to the load through a well-optimized matching network. 3.2 re-designed circuits for 900 and 1800 mhz improved topologies of the circuit in fig. 5 are shown in figs. 12 and 13. however fig. 14 mentions the process of biasing for fast transition. the resistors 4 k and 2 k , as shown in figs. 12 and 13, bias the transistor in the sub-threshold region. this is simply a voltage division procedure adopted to bias the transistor at sub-threshold region. this makes the transistor switch even for a small increment of the input ac signal and hence fosters faster switching. better transitions were obtained as shown in following figs. 15 to 20. these results show the response of pas in order to observe the switching losses. these faster transitions reduced the switching losses and hence improved the power efficiencies. table. 2 summarizes the results of the basic and improved circuits for the two frequencies of 900 mhz and 1800 mhz of interest. the efficiency of a power amplifier is calculated below: the supply power figure: (pdc) = idc vdc pd (drain power) = 34.9ma 2.5v = 87.25 mw average output power obtained from fig. 12 is: poutavg = 56.3 mw the average input power pin is in uw range which is very small compared to pdc and is therefore ignored. similar steps are used for calculating efficiencies for other circuits at different frequencies. for maximum power transfer, matching networks were also designed at the input of the amplifier. without a proper matching network, a significant amount of power will be lost. the addition of a proper value of inductor to the source of the switching trantable 1. summary of the designed passive components for the class e power amplifier and 1800 mhz 78 the journal of engineering research vol 8 no. 1 (2011) 74-82 figure 5. class e power amplifier at 900 mhz time, nsec figure 6. input voltage (vin) and amplified output figure 7. drain current (id) and drain voltage (vd) across the transistor at 900 mhz time, nsec figure 9. input voltage ( vin ) and amplified output voltage ( vout ) at 1800 mhz time, nsec time, nsec 79 the journal of engineering research vol 8 no. 1 (2011) 74-82 sistor makes the input resonant at 900 mhz or 1800 mhz. the value of 30 uh is a reasonable value considering also the current technology limitations. however, that tuning inductor makes the current switching of the transistor slower and hence decreases the efficiency. techniques on how to keep the input of the amplifier matched to the source during fast switching are under investigation. initial results have been encouraging. the matching is an issue to be addressed for maximum power transfer thus increasing efficiency. a pae of around 41% was achieved at 900 mhz ( yoo et al. 2001). also, the drain efficiency was observed around 40% and 45% for 900 mhz and 1800 mhz bands respectively (larcher et al. 2009). the objectives of design improvements for higher efficiency of pas at 900 mhz and 1800 mhz were achieved. these results will be used for further verifications and also for other bands of gsm frequencies which are out of the scope of the current paper. conclusions the class e power amplifier were designed and simulated for gsm frequencies of 900 mhz and 1800 mhz. the redesigned circuits show good improvement in efficiency compared to the efficiencies given by the basic circuits. the dissipated power in the transistors was high due to the transitions between the drain voltage and drain current which led to lower efficiencies. although the figures for the drain voltages and currents for basic and redesigned circuits are less elaborate from visual point of view, the calculation showed significant improvement as tabulated above. also, the use of a proper matching network at the output and better switching techniques achieved by redesigning the bias circuits for proper threshold voltages improved the amplifier efficiency at both frequencies. simulated results show power efficiencies of 67.1% and 67% at frequencies of 900 mhz and 1800 mhz respectively. acknowledgment the authors would like to thank a. kalbani, m. dhawyani, q. mahrooqi, and f. rahbi for their valuable contribution in carrying out the simulation tests and presenting some discussion of the results. the authors would also like to thank sultan qaboos university for the internal grant which helped in achieving the desired results. references berglund, b., johansson, j. and lejon, t., 2006, "high efficiency power amplifiers," ericsson review no. 3, pp. 92-96. dhawyani, m., mahrooqi, q., rahbi, f., kalbani, a., touati, f. and nadir, z., 2008, "improvement in the efficiency of class-e power amplifier for rf," ieee 5th int. multi conference on systems, signals and devices, (ssd'08), pp. 491-495, jordan. figure 10. drain current ( id ) and drain voltage ( vd ) across the transistor at 1800 mhz figure 11. output power (w) for the basic circuit at 1800 mhz time, nsec time, nsec table 2. summary of the efficiency 80 the journal of engineering research vol 8 no. 1 (2011) 74-82 figure 12. improved circuit of class e power amplifier at 900 mhz figure 13. improved circuit of class e power amplifier at 1800 mhz basic circuit redesigned circuit (a) (b) figure 14. (a) the gate of source voltage ( vgs ) for basic circuit and (b) the gate to source voltage ( vgs ) for redesigned circuit vs time (sec) 81 the journal of engineering research vol 8 no. 1 (2011) 74-82 time, nsec figure 15. input vltage ( vin ) and amplified output voltage ( vout ) of the redesigned time, nsec figure 16. drain current (ld) and drain voltage (vd) across the transistor (mni) for the redesigned circuit at 900 mhz figure 17. output power (w) for the redesigned circuit at 900 mhz time. nsec time, nsec figure 18. input voltage ( vin ) and amplified output voltage ( vout ) of the redesigntime, nsec figure 19. drain current ( id ) and drain voltage ( vd ) for the modified circuit at 1800 mhz figure 20. output power (w) for the redesigned circuit at 1800 mhz 82 the journal of engineering research vol 8 no. 1 (2011) 74-82 larcher, l., brama, r., ganzerli, m., iannacci, j., margesin, b., bedani, m. and gnudi, a., 2009, "a mems reconfigurable quad-band class e power amplifier for gsm standard," ieee 22nd int. conference on micro electro mechanical systems, (mems'09), pp. 864-867, australia. lee, t., 2004, "the design of cmos radiofrequency integrated circuits," isbn 139780521835398, cambridge university press, uk. nadir, z. and touati, f., 2009, "design of high efficiency switching mode power amplifier for rf and microwave," ieee 16th int. conference on telecommunications (ict'09), pp. 329-333, morocco. razafi, b., 1998, "rf microelectronics," isbn:0-13887571-5, prentice hall, usa. sokal, n. and sokal, a., 1975, "class e-a new class of high efficiency tuned single-ended switching power amplifiers," ieee j. soild-state circuits, vol. 10(3), pp. 168-176. su, d. and mcfarland, w., 1998, "an ic for linearizing rf power amplifiers using envelope elimination and restoration," ieee j. of solid state circuits, vol 33(12), pp. 2252-2258. yoo, c. and huang, q., 2001, "a common gate switched 0.9w class e power amplifier with 41% pae in 0.25m cmos," ieee j. of solid state circuits, vol. 36(5), pp. 823-830. c:\users\raoh\desktop\paper_5.xps the journal of engineering research vol. 8 no. 1 (2011) 44-52 ________________________________ *corresponding author’s e-mail: lakhwinder_ymca@hot mail.com influence of residual stress on fatigue design of aisi 304 stainless steel l. singh*a, r.a. khanb and m.l. aggarwala *amechanical engineering department, ymca institute of engineering, faridabad -121006, haryana , india. bmechanical engineering department, jamia millia islamia, new delhi-110025. received 5 august 2009; accepted 14 february 2010 abstract: austenitic stainless steel cannot be hardened by any form of heat treatment, in fact, quenching from 10000c merely softens them. they are usually cold worked to increase the hardness. shot peening is a cold working process that changes micro-structure as well as residual stress in the surface layer. in the present work, the compressive residual stress and fatigue strength of aisi 304 austenitic stainless steel have been evaluated at various shot peening conditions. the improvement in various mechanical properties such as hardness, damping factors and fatigue strength was noticed. compressive residual stress induced by shot peening varies with cyclic loading due to relaxation of compressive residual stress field. the consideration of relaxed compressive residual stress field instead of original compressive residual stress field provides reliable fatigue design of components. in this paper, the exact reductions in weight and control of mechanical properties due to shot peening process are discussed. keywords: hardness, fatigue, shot peening, weight reduction nomenclature crsf compressive residual stress field rcrsf relaxed compressive residual stress field a almen 'a' scale sy yield point stress ssp permissible stress ssp/ permissible stress after shot peening fos factor of safety p applied load at the spring end aisi 304 45 the journal of engineering research vol. 8 no. 1 (2011) 46-52 1. introduction shot peening is a method of cold working in which compressive stresses are induced in the exposed surface layers of metallic parts by the impingement of a stream of shots directed at the metal surface at high velocity under controlled conditions. the major applications are related to improvement and restoration of fatigue life and reliability of machine elements by increasing their fatigue strength, straightening and forming of machine elements (metal parts), pretreatment prior to plating, pretreatment for components to be metalized or coated with plastics, enhancement to resistance to stress corrosion cracking and corrosion fatigue, etc. shot peening improves the fatigue and abrasion resistance of metal parts. very fine cracks are generated in plastically deformed layer by shot peening. existing dislocations are interacting with these dislocations in the deformed zone and results in no sharp slip steps. this effect leads to homogeneous deform surface zone and therefore leads to retardation of crack initiation (pandey and deshmukh, 2001). it is applicable to ferrous and non-ferrous parts but is mostly used on steel surfaces. it consists of throwing hardened steel balls at the surface to be peened. the steel balls, or shots, are thrown against the surface either by compressed air or by centrifugal force as rotating wheel fires the shot. the intensity of the process can be varied by regulating the size of shot, the hardness of shot, speed at which it is fired, the length of time and the work is exposed to the shot. the result of the interaction of material parameters with the shot peening parameters are generation of residual stresses, strain hardening of the surface and sub surface layers, changes in microstructure and substructure of material, change in surface conditions and hardening characteristics of the material. mitsubayashi, et al. 1995, suggested a method of improvement in fatigue strength by shot peening and selection of most effective peening condition. shaw, et al. 2003, illustrated the role of residual stress on the fatigue strength of high performance gearing. aggarwal, et al. 2005 and aggarwal, 2006 found that the controlled shot peening improves hardness, causes relaxation of residual stress and increases fatigue strength of en45a spring steel. austenitic stainless steel cannot be hardened by any form of heat treatment, in fact, quenching from 1000oc merely softens them (devdutt and aggarwal, 2006). in the present work, the exact reductions in weight due to relaxation of residual stress and control of mechanical properties by shot peening, are discussed for aisi 304 stainless steel. 2. experimental procedure the chemical composition of austenitic stainless steel aisi 304 used was shown in table 1. the mechanical properties of parent material are: yield strength 210mpa, ultimate tensile strength 560mpa, fatigue limit 176mpa and elongation 40%. the specimens were made according to astm standards. the specimen used for determining the fatigue strength is shown in fig 1. some specimens were shot peened for the purpose of comparing the peening effect. a-type almen strip, 76x19x1.3mm thick was used to measure the intensity of shot peening. w/t = 2 to 6 l w r 8 w to minimize kt of specimens taking, t = 5mm (thickness of the specimens w = 13mm (width of specimen) l = 42mm r = 117mm almen intensity was expressed as arc height of almen strip. 0.15mm camber value was designated as 6a. shot peening was doing with centrifugal wheel system: diameter of wheel 495 mm, operating speed 2250 rpm. the cast steel spherical shots of 0.6mm diameter were used. the hardness of the shots was 56hrc to 60hrc. shot flow rate was varied to obtain various shot peening intensities. the material was treated with almen intensities; 3a, 6a, 8a and 15a. the coverage of shot peening is 90%. in fatigue test a stress ratio (r) equal to 0.1 was used. the specimens are testing in axial fatigue testing machine mts model 810, at a frequency 30hz, at room temperature. l half the length of span n total number of leaves b width of each leaf t the thickness of the each leaf lt total length of the leaf spring density of leaf spring material. 46 the journal of engineering research vol. 8 no. 1 (2011) 46-52 the residual stress induced by shot peening was measured by x-ray diffraction method. a cr tube operating at 30kv with current 8ma was used for projecting k 1 x-rays. the diffraction angle 2 without strain was taken as 135.4o. 3. fatigue design information flow diagram for improvement in fatigue design of a component is shown in fig. 2. the essential characteristics to be known for the fatigue design are: (i) s/n curve (ii) crsf (iii) rcrsf. the data obtained from these characteristics and the mechanical properties of the material helps in fatigue design of the component. basic material data obtained on simple specimen is used for prediction model suggested by (schijve, 2003). the various design factors should be judiciously chosen and the decision is based on data calculated, governing conditions, statistical variations and application. in laboratory testing maximum and minimum load is applied at operating frequency to ascertain whether the component will sustain the definite number of cycles or not. 4. results and discussion fatigue design of shot peened aisi 304 stainless steel leaf spring has been illustrated with an example. the design shows the improvement in fatigue life, fatigue strength, reduction in weight and reliability. also the hardness of the component improves with shot peening. 4.1 mechanical properties three mechanical properties are discussed for reliable fatigue design of shot peened leaf springs: 4.1.1 hardness due to shot peening treatment, the effect of change in hardness was observed in the surface layer. the average value of hardness for base material ie. un-shot peened sample (usp) was 198 hv for aisi 304 stainless steel. depending on the shot peening intensity, the hardness varied from 242 hv to 328 hv for shot peened samples (table 2). the increase in hardness was seen because of the plastic deformation of metals on the metal surface. also after shot peening the austenite phase is partially converted into martensite phase. the microstructure of unpeened austenitic stainless steel (austenite phase) and microstructure of shot peened austenitic stainless steel (martensite + austenite phase) shown in fig. 3 and fig. 4. the phase transformation also improves the hardness (goa yukui, 2004). the effect of shot peening was limited to a very small depth, 0.25 to 0.60 mm from the surface. it may be noticed that the depth of deformed layer is found to vary with shot peening intensity. 4.1.2. damping factor the damping capacity in a system is characterized by a ratio called damping ratio or damping factor. it is the ratio of damping to critical damping. damping factor is an important surface characteristic, which varies with shot peening conditions and depth from the surface of materials (aggarwal, 2008). damping ratio or damping factor of shot peened sample is measured by table 1. chemical composition of aisi 304 (wt %) figure 1. astm flat specimen for fatigue strength 47 the journal of engineering research vol. 8 no. 1 (2011) 46-52 using an experimental set-up consisting of a cro, accelerometer and amplifier. the damping factor is calculated from the logarithmic decrement ( ) (meriam and kraige 2002). damping factor is critical property of the material to control the vibrations. table 3 shows the damping factor of un-treated sample at different depth from the surface. it was also found that with increase in depth of deform layer, damping factor increases (table 4). this was due to the induction of compressive stress on the thin layer of the surface by changing shot peening intensity. 4.2 s/n curves shot peeing improves the fatigue strength of the components. it repairs micro cracks and hence, the initiation of fatigue crack over the surface can be delayed by cold working of the surface with the shots (goa et al 2003 and sharma 2001). s-n curves for the base material and the shot peened specimens were determined using an axial fatigue-testing machine. the four stress levels 210mpa, 225mpa, 240mpa and 265mpa were tested and more tests were conducted near the endurance limit. fifteen specimens (sae annual book, 1977) were tested in order to plot an s/n curve. only the average points were presented for each level (fig. 5). at different stress levels shot peening influence the fatigue strength. the effect of shot peening was negligible if the magnitude of the applied stress is above the endurance limit. it is important to note that there is reduction in fatigue strength at higher almen intensities. at almen intensity 15a fatigue strength of the specimen is less figure 2. flow diagram for improving design table 2. variation of hardness and depth of deformed layer with shot penning 48 the journal of engineering research vol. 8 no. 1 (2011) 46-52 as compare the almen intensity 8a (fig. 5). this reduction in fatigue life at higher almen intensities is due to damage of surface material, which leads to crack initiation on the surface. hence, it is not necessary that increase in shot peening intensity will always increase fatigue strength (goa et al. 2003). 4.3 reduction in weight 4.3.1. compressive residual stress field (crsf) measurement of crsf was carried out by x-ray diffraction method. the measurement of residual stress with x-rays utilizes the interatomic spacing of figure 3. microstructure of austenitic stainless steel without short peening figure 4. microstructure of austenitic stainless steel with shot penning 49 the journal of engineering research vol. 8 no. 1 (2011) 46-52 certain lattice planes as the gage length for measuring strain, fig. 6. in essence, the interatomic spacing (d) for a given lattice plane is determined for the stressfree condition and for the same material containing residual stress. the change in lattice spacing is related to residual stress. bragg's law expresses the relationship between the distance between a give set of lattice planes (d), and the wavelength of x-ray radiation ( ), the order of diffraction (n), and the measured diffraction angle ( ): 2d sin = n (1) the strain normal to the surface, which is measured by x-rays: (2) e = normal strain d1 = lattice spacing with shot peening, a0 d0 = lattice spacing without shot peening, a0 (3) table 3. damping factor for un-treated specimen table 4. effect of shot penning on damping factor figure 5. s/n curve of leaft spring at various shot peened intensities residual compressive stress = e 50 the journal of engineering research vol. 8 no. 1 (2011) 46-52 1a0 = 10-8 cm, e = young's modulus of the material the relaxation behaviour of compressive residual stresses is an essential problem for performance and life prediction of shot peened material and structure. the variation of crsf at different depths for all the peening conditions is shown in fig. 7. it is common to all of the materials that the shot peening induced figure 6. reflextion of x-rays by equidistant parallel atomic planes figure 7. crsf for various shot peening intensities figure 8. relaxation of crsf at different shot peening conditions 51 the journal of engineering research vol. 8 no. 1 (2011) 46-52 compressive residual stress must be released during cyclic loading when the superimposed stress of residual and applied ones is beyond the static and cyclic yield strength. it is usually admitted that the benefit of shot peening is more pronounced for components under a high cyclic region. figure 8 shows the relaxation of stress at different shot peenig intensities. relaxed compressive residual stress field is also measured by x-ray diffraction. 4.3.2. exact reduction in weight the compressive residual stress generated on upper surface due to shot peening protects the part against tensile service load. note that the sign of residual stress is important because if the part were reverse loaded in service to the point of yielding, it would relieve beneficial compressive stress and compromise the part life. the ssp in the component is given as: (4) fos has been applied on crsf because of measurement error, wear of surface and variation in shot size. compressive residual stress induced by shot peening varies with cyclic loading due to relaxation of crsf. relaxation of crsf is due to micro yielding and depends upon ductility of the material, number of cycles and stress applied. therefore, the ssp/ taking into account the relaxation of crsf is stated as: (5) the crsf should be added to yield stress for static loading and rcrsf should be added to yield stress for fatigue loading. aggarwal et al. 2006 illustrated that the results of en45a spring steel specimens and heavy leaf springs used in commercial vehicles, differ from each other. during the present analysis, it is assumed that while working with simple aisi 304 stainless steel leaf springs, results are slightly lower than specimen results and it can be accommodated in factor of safety during fatigue design. a factor of safety 1.6 was chosen from the eq. (6). the variations of the following eight, measure the factor of safety: a) contribution due to material properties (fsm). b) contributions due to incomplete knowledge of load / stress (fss). c) contribution due to geometry resulting from tolerance (fsg). d) contribution due to accuracy of failure analysis and the confidence one has put on them (fsfa). e) contribution because of reliability desired (fsr). f) contribution because of environment (fse). g) factor to consider danger to personnel (fsd). h) factor to consider economic impact (fsel). fos = fsm x fss x fsg x fsfa x fsr x fse x fsd x fse1 (6) the effect of shot peening intensity on fatigue design of semi-elliptical leaf spring with dimensions: p = 300n, l = 425mm, t = 40mm, b = 60mm, n = 6, lt = 3.635 m and = 7900kg/m3, is shown in table 5. bending stress for leaf springs ( b) was calculated using (aaron 1985), b = 6pl / nbt2 (7) weight of leaf spring was calculated as, w = ltbt x (8) 5. conclusions the following conclusions were drawn from the above discussion: 1) the crsf relaxes due to fatigue loading. relaxed compressive residual stress field (rcrsf) due to fatigue loading must be considered for reliable fatigue design of stainless steel aisi 304 leaf springs. 2) the weight reduction is higher in case of static loading but it is lesser in case of dynamic loading. it may be observed that there is a 10-20% weight reduction at optimum shot peening intensity. the designer must include the effect of shot peening, in calculation if he wants to remain competitive. 3) shot peening is also found to be effective process for controlling mechanical properties such as damping factor, hardness and fatigue life in metallic surfaces. 52 the journal of engineering research vol. 8 no. 1 (2011) 46-52 references aaron, d., 1985, "machine design," 3rd ed, macmillan publishing co. usa, pp 750-757. aggarwal, m.l, khan, r.a. and agrawal, v.p, 2005, "influence of shot peening intensity of fatigue design reliability of 65si7 spring steel," indian journal of engg. & materials science, vol. 12(6), pp 515-520. aggarwal, m.l, agrawal, v.p. and khan, r.a., 2006, "a stress approach model for predicting fatigue life of en45a spring steel," int. j. fatigue, elsevier publication, vol. 28(12), pp. 1845-53. aggarwal, m.l., 2008, "shot peened high nitrogen austenitic stainless steel: an alternate to ni rich austenitic stainless steel for medical applications," int. j. of materials and product technology, (accepted). devdutt and aggarwal m.l., 2006, "trends and advances in austenitic stainless steel," national conf. on trends and advances in mech. engg., ymcaie faridabad, india, p. 470. gao, y.k., yin, y.f. and yao, m., 2003, "effects of shot peening on fatigue properties of 0cr13ni8mo2al steel," materials science and technology. vol.19(3), pp. 372-374. gao yukui., 2004, "microstructure changes and residual stresses of each phase caused by shot peening in 304 austenitic stainless steel," journal of aeronautical materials. vol. 24(4), pp. 18-21. meriam, j.l. and kraige, l.g., 2002, "engineering mechanics," vol. 2, 5th ed. (john wiley & sons, inc., singapore), pp. 593-600. mitsubayashi, m., miyata t. and aihara, h., 1995, "prediction of improvement in fatigue strength by shot peening and selection of most effective peening condition," trans. jpn soc. mech. eng. ser a, vol. 61, pp. 1172-1178. pandey, r.k. and deshmukh, m.n., 2001, "shot peening and its effect on fatigue life of engineering components," int. conf. on shot peening and shot blasting, pp. 1-20. sae annual book of astm standards part 10, 1977, pp. 536-554. schijve, j., 2003, "fatigue of structures and materials in the 20th century and the state of the art," int. j. of fatigue, vol. 25(8), pp. 679-700. shaw, b.a., aylott, c., hara, p.o. and brimble, k., 2003, "the role of residual stress on the fatigue strength of high performance gearing," int. j. of fatigue, vol. 25, pp. 1279-1283. sharma, m.c., 2001, "shot peening and blasting," int. conf. on shot peening and sand blasting, pp. 190-195. table 5. effect of almen intensities of fatigue design template for electronic submission to acs journals the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 carrier based pwm technique and adaptive neural network based rotor resistance estimator for the performance enhancement of vector controlled induction motor drives a. venkadesan department of electrical and electronics engineering, national institute of technology puducherry, thiruvettakudy, karaikal-609609, india. abstract: in this paper, the carrier based pulse width modulation (pwm) technique and neural network based rotor resistance estimator are proposed for vector controller induction motor (im) drives. the popular sine pwm is used for induction motor drive. the popular sine pwm has poor harmonic profile and (dc) utilization. the space vector modulation (svm) technique overcomes the disadvantages of sine pwm. but svm is computationally complex. hence a simple pwm technique namely “carrier based pwm technique” similar to svm is identified and proposed for vector controlled im drive. the experimental set up is built up and the performance of carrier based pwm is validated using fpga processor. the adaptive neural network based rotor resistance estimator in predictive mode is proposed for the vector controlled induction motor drive. the performance enhancement of the drive with carrier based pwm and rotor resistance estimator is comprehensively presented. keywords: carrier based pwm technique; rotor resistance estimator; mras; adaptive neural network; vector control; induction motor drive. تقنية تعديل عرض النبضة القائمة على الناقل ومقدار مقاومة الدوران القائم على الشبكة العصبية التكيفية من الحثية موجهة التحكم للمحركاتأجل تحسين األداء فينكاديسانأ. العصبية التكيفية باقتراح تعديل عرض النبضة القائمة على الناقل مقاوم الدوران القائم على الشبكة : قمنا الملخص ويتسم جيب زاوية جهاز تعديل عرض النبضة بسمة توافقية ضعيفة عند استخدام التيار الحثية موجهة التحكم. للمحركات المباشر. ويعمل استخدام تقنية تعديل اتجاه النطاق على التغلب على عيوب جيب زاوية تعديل عرض النبضة. لكن يُتعد بتحديد تقنية بسيطة لتعديل عرض النبضة وهي ق أمرا معقدا من الناحية الحسابية. ومن ثم ، قمنا تقنية تعديل اتجاه النطا "تقنية تعديل عرض النبضة القائمة على الناقل المشابهة لتقنية تعديل اتجاه النطاق، واقترحنا محرك حثي قابل للتحكم فيه. أداء ناقل الحركة القائم على تقنية تعديل اتجاه نطاق و قمنا بإعداد التركيب التجريبي والتحقق من مدى صحة باستخدام معالج تنسيق قابل للبرمجة لنطاق المدخل. ونقترح استخدام تقنية مقاومة الدوران القائم على الشبكة الحركة ر تحسين األداء الحثية موجهة التحكم. و عملنا على عرض شامل لمقدا العصبية التكيفية في الوضع التنبئي للمحركات الحثية موجهة التحكم مع تعديل عرض النبضة القائم على الناقل ومقدّر مقاومة الدوران. للمحركات التكيفية العصبية الشبكة التكيف لنظام مرجعي نموذج الدوران مقاومة مقدار : تعديل عرض النبضةالكلمات المفتاحية . التحكم موجهة الحثية المحركات الناقل متحكم corresponding author’s e-mail: venkadesan@nitpy.ac.in doi: 10.24200/tjer.vol.16iss1pp63-76 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 nomenclature s qsv, s ds v d-axis stator voltage, q-axis stator voltage s qsi, s ds i d-axis stator current, q-axis stator current s qr, s dr  d-axis rotor flux, q-axis rotor flux rr,sr stator resistance, rotor resistance rl,sl stator inductance, rotor inductance ml magnetization inductance st -sampling time r -rotor speed (rad/sec) superscript s -stationary reference frame superscript e -synchronous reference frame 1. introduction three phase induction motor drives are popularly employed for variable speed control applications in many industries. this is due to their ease in maintenance (abdelkarim ammar et al. 2017; ahmed a.z. diab 2014; venkadesan a et al. 2016; venkadesan a et al. 2017; witold pawlus et al. 2017). now-a-days, for high performance applications, vector control provides good alternative as compared to scalar control methods. various pulse width modulation techniques are used to generate pulse for three phase inverter in vector controlled im drives. out of which, space vector modulation technique (benchabane et al. 2012; bimal k. bose 2005; durgasukumar g et al. 2012; govindasamy et al. 2014; hannan m. a et al. 2017; mahmoud gaballah et al. 2013; sabah v.s et al. 2015) is attaining popularity because it has the advantages of having better dc utilization and better harmonic profile as compared to sine-pwm technique. the space vector modulation technique is computationally rigorous and complex as it involves many mathematical calculations (govindasamy et al. 2014; hannan m.a et al. 2017; sabah v.s et al. 2015). the carrier based pwm technique with a common mode voltage injection method is proposed for a three phase inverter keliang zhou et al. 2002). in the paper by keliang zhou et al. (2002), it is mathematically proved that the carrier based pwm technique with the common mode voltage injection method performs similar to svm technique. but the carrier based pwm technique is simple for on-line digital implementation. in the paper keliang zhou et al. (2002), the application of the technique is not demonstrated in vector controlled im drives. the same technique is proposed for open loop speed control of multilevel inverter fed permanent magnet motor shriwastava r.g et al. (2016) and not demonstrated in vector controlled im drives. in this paper, the same technique is proposed for vector controlled im drive. the proposed carrier based pwm technique with the common mode voltage injection method technique (cb-pwm-cmv) is also implemented in fpga (spartan 6 board). the performance of cb-pwm-cmv is also validated experimentally. the vector control especially indirect type, is sensitive to rotor resistance variation (chitra a et al. 2015; karanayil b et al. 2007). hence, rotor resistance estimation using a neural network approach is proposed to track variation in the rotor resistance. the neural network based rotor resistance estimator is proposed for sine pwm based vector controlled im drive chitra a et al. (2015). in the proposed work, neural network based rotor resistance estimator is proposed for cb-pwm-cmv based vector controlled im drive. also in the paper (chitra a et al. 2015; karanayil b et al. 2007), the adaptive nn is used in simulation mode. but the proposed adaptive nn based rotor resistance estimator is used in predictive mode. the idea of using adaptive nn in predictive mode is inspired from the speed estimator proposed by maurizio cirrincione et al. (2005). in the paper by maurizio cirrincione et al. (2005), it is applied for speed estimation. in this proposed work, it is applied for rotor resistance estimation which is novel in this paper. 2. vector controlled im drive the vector control is preferred for high performance applications. the advantages of vector control are well presented in krishnan r (2007). the indirect vector controlled im drive with proposed pwm technique (cb-pwm-cmv) and rotor resistance estimator is shown in fig. 1. for closed operation, the actual speed is compared with the reference speed. the speed error is processed through the pi controller. the torque figure 1. indirect vector controlled im drive showing the rotor resistance estimator. 64 a.venkadesan command ( *t e ) is generated. the reference flux producing component of current ( e*id ) is generated from the reference flux ( φ ref ) using equation (1). the reference torque producing component of current ( e*iq ) is generated using the equation (2). the slip frequency ( sl ω ) is obtained using equation (3). the transformation angle ( e θ ) is obtained using the equation (4). the actual three phase stator current is transformed to actual flux producing component ( eid ) and actual torque producing component ( eiq ). the ( e id ) is compared with ( e*id )and ( e iq ) is compared with ( e*iq ) to produce the corresponding command voltage (vd*, vq*). the reference three phase voltage (va*, vb*, vc*) is generated using the command voltage and transformation angle through d-q transformation theory. the three phase duty cycle profiles are generated using the cb-pwm-cmv. the duty cycle profile/control signal is compared with the triangular signal to generate the switching pulses to the inverter. φ refe*=id l m (1) *4e*=iq 3p φ ref l t r e l m         (2) r sl r e*riq ω = e*lid (3)  e sl rθ = ω +ω dt (4) 2.1 carrier based pwm with common mode voltage injection technique in the carrier based pwm with the common mode voltage injection technique, the common mode voltage is injected into the fundamental sinusoidal reference signal. the resultant signal is used as the modulating signal and it is compared with the carrier signal for the generation of switching pulses to the inverter. the addition of the common mode voltage provides lower harmonic distortion, higher fundamental output voltage srirattanawichaikul w et al. (2011). the common mode voltage can be generated using three phase sinusoidal reference voltages as (5). the duty cycle for each phase can be generated using (6).    max *, *, * min *, *, *a b c a b c * 2 cmv v v v v v v v    (5)  * * , , ,p p cmvd v v p a b c   (6)  * sina av m t (7) 2 * sin 3 b a v m t          (8) 2 * sin 3 c a v m t          (9) where ma is the modulation index. the range of ma is 0<ma ≤1 for under modulation ranges. the fig. 2 shows the block diagram for the extraction of the duty cycle profile. the equations involve only simple arithmetic operations. the equations are simple, less complex and give ease in digital implementation. the fig. 3 presents the duty cycle profile of cb-pwmcmv. figure 2. duty cycle profile generation using cb-pwm cmv technique. the performance of the proposed cb-pwm-cmv technique is compared with the svm and sine pwm technique in terms of dc bus utilization, harmonics profile, efficiency and complexity. the motor is operated with m=0.9, frequency=50hz under a rated load condition. the performance comparison is shown in the table 1. table 1. comparison of cb-pwm-cmv with sine pwm and svm. v d c (v o lt s) m e th o d o u tp u t v o lt a g e v fu n d a m e n ta l (v o lt s) o u tp u t v o lt a g e v t h d (% ) % e ff ic ie n c y c o m p u ta ti o n a l c o m p le x it y 582.5 sine pw m 455.6 79.10% 78.20% simple sv m 528 63.68% 83.43% comple x cbpw mcm v 528 63.68% 83.43% simpler than svm 65 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 (a) (b) (c) (d) figure 3. (a) duty cycle profile of cb-pwm-cmv (b) max(va, vb, vc) (c) min(va, vb, vc) (d) common mode voltage. 66 a.venkadesan the sine pwm is simple, but has poor dc bus utilization and harmonic profile. the svm has better dc bus utilization of 15.89% as compared to sine pwm. it has an output voltage thd of 63.69% which is lesser as compared to sine pwm. the efficiency of the drive with svm and sine pwm is also computed in table 1. for efficiency calculation, the electrical input power of the drive is computed using discrete 3phase positive-sequence active and reactive power block in matlab simulink. it is seen that the efficiency of the drive is improved with svm as compared to the drive with sine pwm. but conventional svm is computationally complex which involves many sector calculations. the proposed cbpwm-cmv shows a similar performance as that of svm but it is simple and less complex. hence, the proposed pwm technique has better dc bus utilization, a better harmonic profile and a better efficiency. besides, it is computationally simple and offers easy on-line digital implementation. the experimental set up is built up to validate the performance of cb-pwm-cmv. the prototype laboratory setup is shown in fig. 4. the cb-pwmcmv is implemented on spartan-6 fpga board. the clock frequency for fpga implementation is chosen as 20mhz. the reference sinusoidal signals are stored in a look up table with 8-bits precision as an integer. the duty cycle profiles are generated with 8-bits precision from the reference sine signals using the proposed cb-pwm-cmv technique. the vhdl coding is developed to realize the cb-pwm-cmv. the gate pulses are given to the three phase inverter through the pulse driver circuit and opto-isolator to drive the induction motor. the switching is chosen as 10khz. the single phase bridge rectifier [db107], buffer ic [cd4081], opto-isolator [6n137] are used for hardware implementation. two coupling capacitors of 330 μf connected in parallel (equivalent capacitor value is 660 μf) are used. the dead time for the inverter is chosen as 6μs. the performance of cb-pwm-cmv fed induction drive is tested for various operating conditions. the sample results obtained are presented. the modulating waveform and switching pulses obtained from the fpga processor are shown in fig. 5(a) and fig. 5(b) respectively. the inverter output voltage obtained for modulation index=0.96 and frequency=50hz is shown in fig. 6(a). the inverter output voltage obtained for modulation index=0.48 and frequency=25hz is shown in fig. 6(b). the two-level inverter is used in this paper. hence, two levels are obtained in the output line-line voltage. it is observed that as soon as the ratio of m/f is reduced to 50%, the fundamental output voltage magnitude and frequency of the line-line voltage is also reduced to around 50%. the effect of variation in the fundamental output voltage magnitude and frequency for various m/f ratios is reflected in the rotor speed. since the rotor speed can be easily measured using tachometer, the rotor speed for various m/f ratios is shown in table 2. the measured speed increases proportionally with increase in m/f ratio. from this table, it is shown that the drive performs very well under a practical condition with cb-pwm-cmv for various m/f ratios. figure 4. hardware setup of svm based inverter fed im drive with fpga controller. 67 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 (a) (b) figure 5. (a) modulating waveform (b) pulses to igbt switches. (a) (b) figure 6. (a) inverter output voltage (l-l) for m=0.96,f =50 hz (b) inverter output (l-l) for m=0.48, f=25 hz. 68 a.venkadesan 2.2 adaptive neural network based rotor resistance estimator the design of an indirect vector controller, to a large extent, depends on the accuracy of the rotor resistance. the slip frequency depends on the knowledge of rotor resistance which is evident from (3). the transformation angle or flux angle depends on the slip frequency which is evident from (4). in real time, the rotor resistance may vary up to 100 due to the rotor heating karanayil b et al. (2007). any mismatch between the rotor resistance of motor and the rotor resistance value in the vector controller leads to the loss of the decoupled control (chitra a et al. 2015; karanayil b et al. 2007). hence to track variation in the rotor resistance, the rotor resistance estimator using model reference adaptive system (mras) with adaptive neural network scheme is proposed. for simulation study, the motor is modeled in d-q frame to introduce variation in the rotor resistance. the proposed mras scheme for rotor resistance estimation is shown in fig. 7. the mras uses a voltage and a current model for rotor resistance estimation. the output from both models is d and qaxis rotor fluxes in stationary reference frame. it is well known that the voltage model is independent of rotor resistance and the flux estimated will be independent of the rotor resistance. hence, the voltage model is used as a reference model. accordingly, the equations (10) and (11) do not contain rotor resistance. the three phase voltages and currents are measured and transformed into d and q axis stationary reference frame and given as inputs to the voltage model. the current model is dependent on the rotor resistance and the flux estimated will be dependent on the rotor resistance. therefore, the current model equation can be used as the adaptive model. consequently, the equations (12) and (13) contain rotor resistance. the three currents and rotor speed are measured. the three phase currents are transformed into d and q axis stationary reference frame. as a result, the inputs to the current model are d and q axis current and rotor speed. the current model equations (12), (13) are represented as a two-layer neural model and used as the adaptive model because it is dependent on the rotor resistance. the weight of the adaptive nn model table 2. measured rotor speed with cb-pwm-cmv technique for various m/f ratios. modulation index frequency (hz) measured rotor speed (rpm) 0.96 50 1478 0.768 40 1176 0.48 25 743 0.325 16.93 498 is the function of the rotor resistance (rr). the error between the reference model flux and the adaptive model flux is back propagated to adjust the weights of the adaptive model to estimate the rotor resistance. the error between the reference and adaptive model is minimized using the bpm (back propagation with momentum algorithm). 2 2 r r s r s m s m m r s l l l l -l ls s s s= ( -r )dtφ v i idr ds ds dsl l l l                (10) 2 2 r r s r s m s m m r s l l l l -l ls s s s= ( -r )dtφ v i iqs qsqsqr l l l l                (11) r m r r r r s φ r l r s sdr s= -ωφ φi qrds drdt l l (12) r m r r r r s φ r l rqr s ss= +ωφ φiqs qr drdt l l (13) the discrete form of current model equations can be obtained using backward euler’s rule. using this rule, the rate of change of rotor flux can be expressed as in (14)    s 's 'st t 1φ φ φ dr dr dr= dt t s   (14) the discrete form of current model equations can be obtained using backward euler’s rule and are given in (14) and (15) s 1 2 3 ds 's 's 's(t) = w (t-1)-w (t-1)+w i (t-1)φ φ φqrdr dr (15) s 1 2 3 qs 's 's 's(t) = w (t-1)+w (t-1)+w i (t-1)φ φ φqr qr dr (16) equations (14) and (15) can be represented as a two-layer neural network with a linear activation function (fig. 8). this is used as the adaptive model. it contains four inputs { 's (t-1)φ dr , 's (t-1)φqr , s (t-1)ids , s (t-1)iqs } and two outputs { 's (t)φ dr , 's (t)φqr }. the adaptive model is used in a predictive mode instead in simulation mode. in predictive mode. in the predictive mode, the past values of rotor flux required as the input for the adaptive model is obtained from the reference voltage model itself rather than from the adaptive model as in simulation mode. this gives ease in digital implementation of the adaptive model and overcomes the problem of instability maurizio cirrincione et al. (2005). the inputs and outputs are connected through the weights w1, w2, w3. the weight w2 is kept constant. the other two weights w1 and w3 are proportional to the rr they are updated using bpm 69 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 to minimize the error e (17). the bpm is coded in matlab-m-file. the learning rate () and momentum factor () are chosen as 5×10 -5 and 2×10 -5 respectively. the e is chosen as 1×10 -6 . the weights of nn are updated using (20) and (21). 2 2s 's φ φ(t) (t)r r 1 e= [ε(t)] =0.5 [ ] 2 (17) 1 d q s sδw (t) = α [ε (t) (t-1) + ε (t) (t-1)] φ φqrdr (18) 3 d q s sδw (t) = [ε (t) (t-1) + ε (t) (t-1)]i iqsds (19) 1 1 1 1 w (t)=w (t-1)+α δw (t)+η δw (t-1) (20) 3 3 3 3 w (t)=w (t-1)+α δw (t)+η δw (t-1) (21) where, s φ (t)r s (t)φ dr = s (t)φqr          ; 'sφ (t)r 's (t)φ dr = 's (t)φqr          ; d s 'sε (t)= (t)(t)φ φ dr dr ; q s 'sε (t)= (t)(t)φ φqr qr the rotor resistance can be obtained either from w1 or w3 (chitra a et al. 2015; karanayil b et al. 2007). r 3 r m s l w r = l t       (22)  rr 1 s l r = 1-w t (23) the effect of sampling time on the performance of the proposed rotor resistance estimator assumes importance for on-line digital implementation. hence, the same is tested for various sampling time. the drive is operated with the speed of 148 rad/s under 100% load torque. the rotor resistance is increased to 50%. the switching frequency of the inverter is chosen as 10khz. hence the switching time is 100µs. figure 7. nn-mras scheme for rotor resistance estimation. figure 8. two-layer neural network. 70 a. venkadesan the steady state error is computed between the actual and estimated rotor resistance for various sampling time and presented in table 3. lesser sampling time leads to more accuracy but increases computation time as the number of samples per cycle is more. on the other hand, larger sampling time decreases computation time but it decreases accuracy as the number of samples per cycle is less. in this paper, the sampling time is chosen same as the inverter switching time for study. the estimator is tested for a step change in rotor resistance. the actual and estimated rotor resistance for step variation (24) is shown in fig. 9(a). the sample weight evolution plot for step change is also shown in fig. 9(b). it is understood that the network learns to estimate the rotor resistance. the weight and the rotor resistance are linearly related. the step change is the extreme case condition. it is used to show the capability of the proposed estimator to estimate rr even in the extreme case condition. but in the practical real time, the rotor resistance varies slowly. accordingly, the proposed estimator is tested for trapezoidal variation (25) in the rr and also with exponential variation beguenane r et al. (1999) as in (26). the result obtained for trapezoidal variation and exponential variation is shown in the fig. 9(c) and fig. 9(d) respectively. the nn-mras tracks the variation in rr for the various rotor resistance variation profile very well with good accuracy.       sec5.11mno,trmno,tr sec5.11mno,tr )t(rr (24)                   sec4tsec3 )mno,rrmno,rr(mno,rr t)mno,rr( sec3tsec2mno,rrmno,rr sec2tsec1 mno,rr)mno,rrmno,rr( )mno,rr(t sec)4t(orsec)1t(mno,rr )t(rr (25) ) t2.1 1(mno,rr)t(tr   e (26) the estimated rr value for various %changes in rr is consolidated and presented in the table 4 along with the steady state error for 100% and 50% load torque. it is found that the proposed estimator tracks, the change in rr with good accuracy. 2.3 adaptable neural network based rotor resistance estimator the performance of the drive with and without a rotor resistance estimator is studied. the motor is operated with the speed of 148 rad/sec, torque of 7.5 nm and rotor flux of 0.9 wb. the step change in rotor resistance is effected at 1.5 sec. without the rr estimator, the rotor flux, torque and three phase stator current are shown in the fig. 10 (a), fig. (b), fig. (c) respectively. it is observed that without the rotor resistance estimator, the error in the rotor flux is found to be 16.66% and the actual flux deviates from the reference flux. the ripple in the machine torque is found to be 30.46% and oscillates about the load torque which is evident from fig. 10(b). the sinusoidal shape of the three phase stator current gets distorted. this type of response is not desirable in high performance drives. with the rotor resistance estimator, the actual rotor flux tracks the reference flux with the error of 0.133% after about 0.4sec as soon as the step change in rotor resistance is applied at 1.5sec which is evident in fig. 11(a). with the rotor resistance estimator, the ripple in the machine torque is reduced significantly to 3.382% which can be observed in fig. 11 (b). the sinusoidal shape of the current is maintained which can be observed from in fig. 11(c). hence, the performance of the cb-pwm-cmv based vector controlled im drive is significantly enhanced with rotor resistance estimator. table 3. effect of sampling time on the rotor resistance estimator for switching frequency 10khz. sampling time actual rr (ohms) estimated rr using proposed nnmras scheme (ohms) %steady state error 50µs 9.127 9.123 0.043826 100µs 9.114 0.142435 200µs 9.052 0.821738 300µs 8.958 1.851649 400µs 8.828 3.275994 500µs 8.657 5.149556 3. conclusion a carrier based pwm technique and neural network based rotor resistance estimator are proposed for the performance enhancement of induction motor drives. the cb-pwm-cmv technique is computationally less complex and gives ease in digital implementation as compared to space vector modulation technique. the cb-pwm-cmv technique is shown to improve the efficiency of the drive as compared to sine-pwm. the cb-pwm-cmv technique is implemented on spartan-6 fpga processor and experimentally validated. a neural network based model reference adaptive system in prediction mode is proposed for the rotor resistance estimation. the proposed nnmras is shown to estimate rr with good accuracy which is evident from table 4. with rotor resistance estimator, the performance of the drive is significantly improved. hence, it is concluded that the performance of the drive is enhanced significantly with cb-pwmcmv and rotor resistance estimator. 71 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 (a) (b) (c) (d) figure 9. (a) rotor resistance for step variation (b) weight evolution plot for step variation (c) rotor resistance for trapezoidal variation (d) rotor resistance for exponential variation. 72 a.venkadesan (a) (b) (c) figure 10. without rr estimator: (a) rotor flux (b) torque (c) current. 73 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 (a) (b) (c) figure 11. with rr estimator: (a) rotor flux (b) torque (c) current. 74 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 table 4. rotor resistance estimation using nn mras for various % changes in rotor resistance. % v a r ia ti o n i n r r a c tu a l r r (o h m s) 100% load torque 50% load torque e st im a te d r r u si n g p r o p o se d n n m r a s s c h e m e (o h m s) % s te a d y s ta te e r r o r e st im a te d r r u si n g p r o p o se d n n m r a s s c h e m e (o h m s) % s te a d y s ta te e r r o r 10 6.69 6.672 0.328653 6.663 0.463101 20 7.30 7.284 0.246508 7.264 0.520405 30 7.91 7.983 -0.92288 7.889 0.265487 40 8.51 8.500 0.223031 8.487 0.375631 50 9.12 9.114 0.142435 9.098 0.317739 60 9.73 9.713 0.236237 9.711 0.256779 70 10.34 10.33 0.096712 10.32 0.193424 80 10.95 10.94 0.091324 10.92 0.273973 90 11.56 11.54 0.17301 11.53 0.259516 100 12.17 12.15 0.164339 12.14 0.246508 confilct of interest author declares no conflicts of interest. funding no funding was received for this research. acknowledgment i thank dr. smrutisikta mishra, assistant professor, department of humanities and social sciences, nit puducherry, karaikal for her valuable suggestions in english language. references abdelkarim ammar, amor bourek, abdelhamid benakcha (2017), sensorless svm-direct torque control for induction motor drive using sliding mode observers. journal of control, automation mand electrical systems 28(2): 189-201. ahmed a.z.d (2014), real-time implementation of full-order observer for speed sensorless vector control of induction motor drive. journal of control. automation and electrical systems 25 (6): 639-648. beguenane r, benbouzid m. e. h (1999), induction motors thermal monitoring by means of rotor resistance identification. ieee transactions on energy conversion 14(3): 566-570. benchabane f, titaouine a, bennis o, yahia k, taibi d (2012), direct field oriented control scheme for space vector modulated ac/dc/ac converter fed induction motor. frontiers in energy 6(2): 129 137. bimal k. bose (2005), modern power electronics and ac drives. prentice hall of india. chitra a, himavathi s (2015), a modified neural learning algorithm for online rotor resistance estimation in vector controlled induction motor drives. frontiers in energy 9(1): 22-30. durgasukumar g, pathak m.k (2012), comparison of adaptive neuro-fuzzy-based space-vector modulation for two-level inverter. electrical power and energy systems 38: 9-19. govindasamy renukadevi, kalyanasundaram rajambal (2014), field programmable gate array implementation of space-vector pulse-width modulation technique for five-phase voltage source inverter. iet power electronics 7(2): 376389. hannan m. a, jamal abd ali, azah mohamed, mohammad nasir uddin (2017), a random forest regression based space vector pwm inverter controller for the induction motor drive. ieee transactions on industrial electronics 64(4): 2689-2699. karanayil b, rahman m.f and grantham c (2007), online stator and rotor resistance estimation scheme using artificial neural networks for vector controlled speed sensorless induction motor drive. ieee transactions on industrial electronics 54 (1): 167-176. keliang zhou and danwei wang (2002), relationship between space-vector modulation and three-phase carrier-based pwm: a comprehensive analysis. ieee transactions on industrial electronics 49 75 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 63-76 (1): 186-196. krishnan r (2007), electric motor drives modeling, analysis and control. prentice hall of india. mahmoud gaballah, mohammed el-bardini (2013), low cost digital signal generation for driving space vector pwm inverter. ain shams engineering journal 4: 763-774. maurizio cirrincione, and marcello pucci (2005), an mras-based sensorless high-performance induction motor drive with a predictive adaptive model. ieee transactions on industrial electronics 52(2): 532-551. peter vas (1998), sensorless vector and direct torque control. oxford university press. sabah v.s, charles baby t, krishna prabhakar lall, balamurugan m,sarat kumar sahoo, chitra a,prabhakar, karthikeyan, i. jacob raglend (2015), a simplified space vector pulse width modulation technique for multilevel inverter fed induction motor drive. ieee international conference on circuit, power and computing technologies (iccpct). shriwastava r.g, daigavane m. b, daigavane p. m (2016), simulation analysis of three level diode clamped multilevel inverter fed pmsm drive using carrier based space vector pulse width modulation. (cb-svpwm), procedia computer science (elsevier). 79: 616-623. srirattanawichaikul w, kumsuwan y, premrudeepreechacharn s, wu b (2011), carrierbased pwm strategy for three-level neutral-pointclamped voltage source inverters. ieee peds 2011, singapore, 948-951. venkadesan a, himavathi s and muthuramalingam a (2016), a novel nn based rotor flux mras to overcome low speed problems for rotor resistance estimation in vector controlled im drives. frontiers in energy 10(4): 382-392. venkadesan a, himavathi s, sedhuraman k, muthuramalingam a (2017), design and field programmable gate array implementation of cascade neural network based flux estimator for speed estimation in induction motor drives. iet electric power applications 11(1): 121-131. witold p, huynh v.k, michael r.h (2017), temperature rise estimation of induction motor drives based on loadability curves to facilitate design of electric powertrains. ieee transactions on industrial informatics 13(3): 985-994. appendix 1 induction motor parameters: 3 phase, 4 poles, 1100w, 415v, 50hz, 7.5nm. rs rr lm lsl lrl j b 6.03ω 6.085ω 0.4893h 0.0138h 0.0138h 0.011787kg.m 2 0.0027kg.m 2 /s 76 c:\users\raoh\desktop\paper_4.xps the journal of engineering research vol. 8 no. 2 (2011) 32-39 1. introduction resistance spot welding (rsw) has been used for decades as a joining method for sheet metal. weldbonded is a combination of resistance spot welding and adhesive bonding, which has gathered wide acceptance as an effective joining method for significant enhancement of static, dynamic and impact resistance of the joint. it also improves the corrosion and noise resistance as well as stiffness of the joint, com __________________________________________ *corresponding author e-mail: ebahkali@ksu.edu.sa pared to those observed in case of conventional resistance spot welding. industrial applications such as automobile and aerospace are good examples of using weld-bonded process. in order to reach an optimum welding quality of a spot welded or a weld-bonded joints, different calibration trials must be conducted to setup the optimum welding parameters, i.e. welding current, electrode force, and welding time (bouyousfi et al. 2007, farukawa et al. 2006). load-displacement curves of spot welded, bonded, and weld-bonded joints for dissimilar materials and thickness e.a. al-bahkali department of mechanical engineering, college of engineering, king saud university, p.o. box 800, riyadh 11421, saudi arabia received 15 april 2011; accepted 21 september 2011 abstract: three-dimensional finite element models of spot welded, bonded and weld-bonded joints are developed using abaqus software. each model consists of two strips with dissimilar materials and thickness and is subjected to an axial loading. the bonded and weld-bonded joints have specific adhesive thickness. a detailed experimental plan to define many properties and quantities such as, the elastic plastic properties, modulus of elasticity, fracture limit, and properties of the nugget and heat affected zones are carried out. experiments include standard testing of the base metal, the adhesive, the nugget and heat affected zone. they also include employing the indentation techniques, and ductile fracture limits criteria, using the special notch tests. complete load-displacement curves are obtained for all joining models and a comparison is made to determine the best combination. keywords: spot welding, adhesive, weld-bonded, finite element modelling, dissimilar material 33 e.a. al-bahkali various techniques have been used to analyze the spot welded and weld-bonded joints. finite element method (fem) is one of these techniques, which has received a wide acceptance among researchers as a tool for modelling and analysis of welding joints. it has the capability to predict stress distribution, stress concentration, and failure modes for both weld-bonded and spot welded nuggets. to model real joint connection using fe, the elastic and plastic mechanical properties of the parts must be taken into consideration. therefore, many researchers worked on weldbonded technology experimentally and numerically in an attempt to represent real joint connection. kang et al. (2006) used tabor-equation to transform the indentation data to true stress-strain values. more sophisticated methods to determine the elastic-plastic properties from indentation are used and reported by (venkatesh et al. 2000 and dao et al. 2001). they used instrumented sharp indentation machine, where the load and displacement during indentation were recorded continuously. li et al. (2005) studied numerically the deformation and fracture initiation of mode i (normal strength and toughness) in adhesive bonding. later, yang et al. (2001) investigated the fracture toughness data for adhesive. in addition, (li et al. 2006) continued the study of the mixed mode cohesive-zone models for fracture of an adhesively bonded polymer-matrix composite. in their work, the load displacement curve of single lap shear joint obtained by fe analysis showed an acceptable agreement with the experimental data. moreover, (cavalli et al. 2004) investigated the finite element model based on cohesive element model for weld-bonded joint of aa-5754 aluminium alloy. it is also important to study the microstructure and mechanical properties at the nugget and the heat affected zone (haz) caused by welding process because the plastic properties of the nugget and haz are different from the base metal. kong et al. (2008) reported that, the elastic-plastic properties have to be defined for each region of spot welded joint, by the combination of indention and extensive fe simulation of the indentation. al-bahkali et al. (2010) have developed a 3d fe model of spot welded, adhesive bonded, and weld bonded joints of austenitic stainless steel sheets of 1.0 mm thickness. their models are based on elastic-plastic properties, and ductile fracture limit criteria for steel, whereas the adhesive bonding is modelled based on traction separation. they have also studied the elastic plastic properties, modulus of elasticity, fracture limit, nugget and haz properties. the load-displacement curves obtained from the fe models are in agreement with the experimental data. spot weld and weld-bonding of dissimilar materials are of interest to engineers and scientists for efficiently joining two different materials. hasanbasoglu and kacar (2007) investigated the influence of the primary welding parameters on the morphology, microhardness, and shear tensile load bearing capacity of two dissimilar steel welds. darwish (2004) analyzed spot-welded and weld-bonded of dissimilar material joints using fem. he showed that the stresses were more concentrated towards the member that had the lowest melting point of the joint. he pointed that, adding an adhesive layer in combination with the spot weld resulted in eliminating the stress concentration and strength ending dissimilar material joints. the aim of present work is to study the load-displacement curves toward the best combination for spot weld, bonded, and weld-bonded for dissimilar materials and thicknesses. 3d fe models are developed for dissimilar spot welded, bonded, and bond-welded joints, based on elastic-plastic properties and ductile fracture limit criteria for each region (base metal, nugget and haz). these developed fe models can predict the deformation and fracture initiation, including the maximum load of the weld-bonded joint as well as spot welded and adhesive bonded for any combination of sheet thickness, overlap area, and nugget diameter. in the fe models, the plastic properties of each welding region are defined by spherical indentation and transformed to true stress-true strain using ahn-equation 2006. furthermore, the ductile fracture limit criteria are employed to determine the fracture initiation point in the fe model (bao 2005; mackenzie et al. 1977; hancock, mackenzie 1976). this method is based on evaluating the stress triaxiality versus equivalent plastic strain conditions. the fracture limits of adhesive bonding are introduced in terms of maximum normal stress and shear stress of traction separation, which is defined experimentally (sun et al. 2009, diehl 2005). 2. finite element model 2.1 geometry in this section, three different finite elements models are considered. these models are a single lap bonded model, a single lap spot weld model, and a single lap weld-bonded model. figure 1 shows the front view (f.v) and top view (t.v) configurations, dimensions, constraints and loading conditions for bonded and weld-bonded models. spot weld model is similar to weld-bonded except that it does not have an adhesive layer around the nugget. the total length of each model is 175 mm. the strips thicknesses are 1mm and 1.5 mm, respectively. the following assumptions are considered throughout the idealization process. the models are threedimensional and because of symmetry half of the model is considered to save computation time. the 34 load-displacement curves of spot welded, bonded, and weld-bonded joints for dissimilar materials and thickness adhesive layer is isotropic and has a thickness of 0.12 mm. there is no adhesive layer in a zone 1 mm around the circumference of the weld nugget and the depth of the indentation is assumed to be 0.1 mm for both strips (baohua et al. 2001). this indentation caused by the electrode of the spot weld machine. all the regions in the fe modeling were connected by sharing nodes. the overlap joint is subjected to a constant velocity of 1 mm/min. the line of action is not initially parallel to the adhesive layer as shown in fig. 2. as the load increases the overlap area bends. consequently, the ends of the adhesive layer peel and shear stresses appear. these stresses often induce joint failure. 2.2 boundary conditions the boundary conditions associated with each finite element model can be summarized as follows. on the edges x=0, clamped boundary conditions are imposed, ux |x=o = uy |x=o = uz |x=o = 0 (1) whereas both strips are subjected to a fixed y-direction boundary condition for 30 mm segment of strip a (x=0 to 30 mm) and at the end 30 mm segment of strip b (x=145 to 175 mm). uy |x=0-30 = uy |x=145-175 = 0 (2) in the overlap area, tie constraints are imposed between components of welded joints; i.e. strip a, strip b, adhesive layer, and weld nugget. by doing so, the translational and rotational boundary conditions of tied surfaces are made identical, regardless of the way these parts are meshed. the model is subjected to a constant velocity (v =1 mm / min.) at the right edges of strip b. vx |x=175 =1 mm / min (3) 2.3 finite element mesh the finite element computation is carried out using abaqus software. the finite-element meshes of these models are generated using eight-node-linear brick reduced integration elements. figure 3 shows the fe meshes for adhesive model (bonded) and weldbonded models, respectively. the mesh of bonded model is straight forward and simpler because of the absence of spot welding, which leads to only the overlap area that needs to be divided into fine mesh. however, the spot and weld-bonded models need further fine mesh on the edges of spot weld and adhesive layer to reduce modeling errors. the numbers of elements for the different models, after several trials of refined meshes to ensure the conversion of fe results, are given in table 1. 3. results 3.1 material properties the material properties used throughout the present figure 1. bonded and weld-bonded models (a) (b) figure 2. lap-joint have (a) a line of action is not initially parallel to the adhesive layer, (b) the overlap area bends as load increses (a) adhesive model (b) weld-bonded model figure 3. finite element mesh for both bonded weld-bonded models 35 e.a. al-bahkali work are given in table 2. these properties were obtained in the laboratory. the elastic-plastic properties of the joint are needed for the fe modeling. this requires identifying and determining accurately each region of the spot welded joint, which is achieved by conducting microhardness measurements across the joint specimens. therefore, vickers micro-hardness measurements of the weld nugget, haz, and base metal of spot welded joints are carried out at a load of 100 gram. the measurements started from the center of the nugget moving outwards step by step to the haz then to the base metal, with an incremental distance of 0.25mm. spherical indentation (2mm diameter) is used to define the plastic properties of the nugget and heat affected zone for only the aisi 304 steel. this is because steel's properties are affected by the heat. however, brass is not affected by the heat during spot welding process. consequently, the true stress-true strain curves for aisi 304 steel in these regions are derived using ahn-equation: (4) (5) where is the true stress, is an empirical factor which is equal to 3.6. p is the load, pm is the mean pressure, ac is the contact radius between the indenter and the material, is the true strain, is the adjustment constant taken as 0.14 (jeon et al. 2006), and r is the indenter radius. the true stress-true strain curves of the base metal, haz, and nugget of spot welded joint for aisi 304 steel are shown in fig. 4. for the brass, being not affected by heat, the true stress-true strain curves is the same for base metal, haz, and nugget of spot welded joint as shown in fig. 5. in order to ensure the micro-hardness results, a micro structure of welding cross section was considtable 1. elements number used in different models tab le 2. material properties of strip and adhesive figure 4. true stress-true strain curves of the base metal, haz, and nugget of spot welded 0 0.2 0.4 0.6 true strain 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0.6 0.4 0.2 0.0 0 0.1 0.2 0.3 0.4 true strain figure 5. true stress-true strain curves for brass 36 load-displacement curves of spot welded, bonded, and weld-bonded joints for dissimilar materials and thickness ered. after grinding and polishing, the specimen was etched by dipping into a solution of 10 ml nitric acid, 10ml acetic acid, 5ml glycerin and 15 ml hydrochloric acid, followed by washing in stream of water. finally, the micro structure was examined for aisi 304 steel using optical microscope as shown in fig. 6. the micro structure of brass is not changed during the process of spot welding which leads to constant hardness for base metal, haz, and nugget areas. the results of micro-hardness testing for aisi 304 steel are plotted in fig. 7. it is clear from the figure that the nugget has the highest hardness, followed by the haz, and finally the base metal. in addition, the dimension of each region is described clearly in the figure. to obtain the fracture initiation point (maximum load) as one of the results of finite element modeling, the fracture initiation limits have to be introduced in the model. these properties can be defined and evaluated from the notch tensile test. in abaqus software, the ductile fracture limits are defined in terms of stress triaxiality and corresponding equivalent fracture strain. the basic theories of the stress triaxiality are described in references (bao 2005; mackenzie et al. 1977; hancock, mackenzie 1976). mackenzie et al. (1977) defined the stress triaxiality for cylindrical specimen as: triaxiality = m / ' (5) where m is the mean principal stress, and it is defined as: m = ( 1+ 2+ 3) / 3 (6) and ' is the effective stress which is defined as: ' = [1/2{( 12)2+( 23)2+( 31)2}]1/2 (7) where 1, 2, and 3 are the principal stresses. -8 -6 -4 -2 0 2 4 6 8 position from center of nugget (mm) figure 7. micro-hardness vickers test results for aisi 304 steel figure 8. notched brass specimens prepared for tinsile test 0 0.2 0.4 0.6 stress triaxiality 0.6 0.5 0.4 0.3 0.2 0.1 0 stainless steel brass figure 9. the ductile fracture limit properties 37 e.a. al-bahkali in order to obtain the stress triaxiality and its corresponding equivalent fracture strain values for sheet specimen, notch tensile specimens were prepared and tested according to the standard tensile test. five notch specimens were prepared with notch radii of 2, 2.5, 3, 4, and 5.1mm, respectively, for each material. figure 8 shows a picture of notched brass specimens before subjected to tensile test. standard tensile tests were conducted on each samples as stated above, a complete load displacement curve is obtained from the tests. the equivalent true strains at the fracture initiation were calculated automatically using the bluehill-software. then, the stress triaxiality at the determined equivalent strain was calculated numerically using abaqus software. the stress triaxiality values and the corresponding equivalent fracture strains for aisi 304 steel and brass are plotted in the fig. 9. the data were extrapolated as an exponential curve, based on a recommendation of previous researchers (bao 2005; mackenzie et al. 1977; hancock, mackenzie 1976). the adhesive layer model is developed based on traction separation mode. after defining all materials properties, the fe models are run to obtain the loaddisplacement curves. 3.2 finite element results in this section, the results of the finite element simulations for bonded, spot weld, and weld-bonded dissimilar materials and thickness joints are examined. the combination dissimilar materials used in the analysis are steel-brass. in the first combination, the material for strip a is steel with 1mm thickness while the material for strip b is brass with 1.5 mm thickness. in the second combination, strip a represents steel with 1.5 mm thickness while strip b represents brass with 1mm thickness. figure 10 shows three different load-displacement curves for joining steel (1 mm thickness) with brass (1.5 mm thickness). when both materials are joined by spot weld, the joint can withstand a maximum load of 4 kn at a displacement of 0.39 mm. when they are joined by adhesive, the load reaches 9.9 kn and the displacement reaches 1.08 mm. adding an adhesive layer to the spot weld, the maximum load is slightly increased to 10 kn and the displacement is increased to 1.28 mm. figure 11 shows three different load-displacement curves for joining steel (1.5 mm thickness) with brass (1mm thickness). when the materials are joined by spot weld, the joint can hold up to a maximum load of 3.8 kn at a displacement of 0.29 mm. when they are joined by adhesive, the load reaches 7.79 kn and the displacement reaches 0.94 mm. adding an adhesive layer to the spot weld, the maximum load is almost the same at 7.78 kn. however, the displacement is reduced to 0.78 mm. figure 11. load displacement curves for joining steel (1.5 mm thickness) with brass (1 mm thickness) the maximum load (p kn) and corresponding displacement ( mm) obtained from both figs. 10 and 11 can be summarized in table 3. the results show that when the thinner strip is steel, the weld-bonded model can withstand larger load and longer extension. in case the thinner strip is brass, both bonded and weld-bonded models have almost the same maximum load, however, bonded model has longer extension. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 displacement (mm) 12 10 8 6 4 2 0 figure 10. load displacement curves for joining steel (1 mm thickness) with brass (1.5 mm thickness) adhesive bonding spot welding weld b4onding 0 0.2 0.4 0.6 0.8 1 1.2 1.4 displacement (mm) adhesive bonding spot welding weld bonding 10 8 6 4 2 0 figure 11. load displacement curves for joining steel (1.5 mm thickness) with brass (1 mm thickness) 38 load-displacement curves of spot welded, bonded, and weld-bonded joints for dissimilar materials and thickness conclusions a three-dimensional finite element model for dissimilar materials and thickness are developed. this includes the 3-d finite element modeling of the spot welded, bonded and weld-bonded joints under axial loading conditions. the combinations of dissimilar materials that are used in the analysis are steel-brass. in the first combination, steel is thinner than brass while in the other combination, the brass is thinner than the steel. for each combination, the load-displacement curves for all three joining types are successfully obtained. when the thinner strip is steel, the bonded model is 2.47 times better than spot model in terms of load carrying capacity. adding the adhesive layer to the spot model will result in a very small improvement in maximum load bearing capacity. however, the displacement is improved by 15.63%. when the thinner strip is brass, both bonded and weld-bonded models have almost the same maximum load, however, it is better to use the bonded model because the displacement in bonded is more than the weld-bonded by 20.5%. in general, for a model with dissimilar materials and thickness, it is better to use the weld-bonded model when the strip of the softer material has the large thickness and to use the bonded model when the strip of the harder material is thick. acknowledgments the author is grateful for the college of engineering research center at king saud university, for the support of this work (project no. 19/428). references al-bahkali e, es-saheb m, herwan j (2010), finite element modeling of weld-bonded joint. the 4th int. conf. on advanced computational engineering and experimenting, paris, france. bao y (2005), dependence of ductile crack formation in tensile tests on stress triaxiality stress and strain ratios. j. of engineering fracture mechanics 72: 502-522. baohua c, yaowu s, liangqing l (2001), studies on the stress distribution and fatigue behavior of weld-bonded lap shear joints. j. of materials processing technology 108:307-313. bouyousfi b, sahraoui t, guessasma s, chaoch k (2007), effect of process parameter on the physical characteristic of spot weld joints. j. of materials and design 28:414-419. cavalli m, thouless m, yang q (2004), cohesivezone modeling of the deformation and fracture of weld-bonded joints. welding journal 133-139. dao m, chollacoop n, van vliet kj, venkatesh ta, suresh s (2001), computational modeling of forward and reverse problems in instrumented sharp indentation. acta materialia 49:3899-3918. darwish s (2004), analysis of weld-bonded dissimilar materials. int. j. of adhesion and adhesives 24: 347-354. diehl t (2005), modeling surface-bonded structures with abaqus cohesive elements: beam-type solution. abaqus user's conference. furukawa k, katoh m, nishio k, yamaguchi t (2006), influence of electrode pressure and welding conditions on the maximum tensile shear load. q, j. of the japan welding society 10-16. hancock jw, mackenzie ac (1976), on the mechanism of ductile failure in high-strength steels subjected to multi-axial stress-states. j. of the mechanics and physics of solids 24:147-169. hasanbasoglu a, kacar r (2007), resistance spot welding of dissimilar materials (aisi 316l-din en 10130-99). j. of materials and design 28:1794-1800. jeon e, kim jy, baik mk, kim sh, park js, kwon d (2006), optimum definition of true strain beneath a spherical indenter for deriving indentation flow curves. j. of materials science and engineering a419:196-201. kang bsj, yao z, barbero ej (2006), post-yielding stress-strain determination using spherical indentation. mechanics of advanced materials and structures 13(2):129-138. kong x, yang q, li b, rothwell g, english r, ren h (2008), numerical study of spot-welded joints of steel. j. of materials and design 29:1554-1561. li s, thouless md, waas am, schroeder ja, zavattieri pd (2005), use of mode i cohesive zone models to describe the fracture of an adhesively-bonded polymer-matrix composite. j. of composite science and technology 65:281-293. li s, thouless md, waas a.m, schroeder ja, zavattieri pd (2006), mixed-mode cohesive-zone table 3. maximum load-displacement for different models combinations 39 e.a. al-bahkali models for fracture of an adhesively bonded polymer-matrix composite. j. of engineering fracture mechanics 73:64-78. mackenzie ac, hancock jw, brown dk (1977), on the influence of state of stress on ductile failure initiation in high strength steels. j. of engineering fracture mechanics 9:167-188. sun c, thouless md, waas am, schroeder j, zavattieri pd (2009), rate effects for mixed-mode fracture of plastically-deforming adhesivelybonded structures. int. j. of adhesion and adhesives 29:434-443. venkatesh ta, van vliet kj, giannakopoulos ae, suresh s (2000), determination of elasto-plastic properties by instrumented sharp indentation: guidelines for property extraction. scripta materialia 42:833-839. yang qd, thouless md, ward sm (2001), elasticplastic mode ii fracture of adhesive koints. international journal of solid structure 38:32513262. the journal of engineering research (tjer), vol. 18, no. 1, (2021) 1-11 *corresponding author’s e-mail: amalhamza31@yahoo.com doi: 10.53540/tjer.vol18iss1pp1-11 improving biogas production using different pretreatment of rice husk inoculated with ostrich dung huda jassim *, and amal h. khalil department of environmental engineering, college of engineering, university of babylon, iraq abstract: the increased demand for fossil fuel due to its huge importance in global energy has caused many problems to the environment both locally and globally. this study investigated the feasibility of the anaerobic co-digestion using rise husks as a biomass waste source and ostrich dung as a source of inoculum for biogas production as renewable energy. samples of grounded rh were pretreated with ultrasound, hydrothermal, combined hydrothermal-ultrasonic, naoh (3% w/v) and combined 3% naoh-ultrasound pretreatment. the cumulative production of biogas for the pretreated samples was 44.19, 39.5, 46.3, 47.16 and 52.38 ml/g vs, respectively. the results stated the increase in biogas generation by 50.72, 34.72, 57.91, 60.85 and 78.65%, respectively, as compared to untreated sample that produced 29.32 ml/g vs of biogas. the results of methane productivity, which were 30.73, 26.78, 32.24, 32.91 and 37.3 ml/g vs, respectively; for the pretreated samples with the same arrangement mentioned above, caused an increment by 66.11, 44.76, 74.27, 77.89, 101.62%, respectively, as compared to methane yield of the untreated sample that resulted in 18.5 ml/g vs. it was observed that there was well compatibility between estimated and predicted values for methane (> 0.95) using modified gompertz model. the improvements in biogas and methane yield pointed out the anaerobic codigestion of rice husks after pretreatment inoculated with ostrich dung is a promising technology to produce renewable energy. the use of ostrich dung encouraged biogas productivity. in addition, biogas can reduce environmental pollution by managing wastes. keywords: biogas; co-digestion; hydrothermal; ultrasound, rice husks; ostrich dung. النعام بفضالت األرز الملقح لقشور مختلفة مسبقةمعالجات باستخدام الحیوي الغاز إنتاج تحسین امال حمزة خلیل ، *ھدى جاسم ان الطلب المتزاید للوقود االحفوري نسبة الى اھمیتھ الكبرى في الطاقة العالمیة ادت الى عدة مشاكل للبیئة محلیا :الملخص الدراسة الحالیة بحثت امكانیة تطبیق الھضم المشترك الالھوائي باستعمال قشور الرز كمصدر للمخلفات الحیویة .وعالمیا وفضالت النعام كمصدر للبكتریا النتاج الغاز الحیوي كنوع من الطاقة المتجددة. تمت معالجة مسحوق قشور الرز بعدة الجة بالبخار،المعالجة المركبة بالموجات فوق الصوتیة معالجات وھي المعالجة بالموجات فوق الصوتیة ،المع 3) ، واخیرا المعالجة المركبة بھیدروكسید الصودیوم w/v) %3%والبخار،المعالجة بھیدروكسید الصودیوم بتركیز ، 47.16، 46.3، 44.19،39.5والموجات فوق الصوتیة. االنتاج المتراكم للغاز الحیوي لھذه النماذج المعالجة ھو مل/غم من المواد العضویة المتطایرة على التوالي مما یعني ان الزیادة الحاصلة في انتاج الغاز الحیوي لھذه النماذج 52.38 % على التوالي مقارنة بالنموذج غیر المعالج. ان انتاجیة المیثان 78.65، 60.85، 57.91، 34.72، 50.72المعالجة ھي % على التوالي للنماذج المعالجة بنفس الترتیب اعاله 101.62، 77.89، 74.27 ،44.76، 66.11تحسنت ایضا بحوالي مقارنة بالنموذج غیر المعالج. لقد اظھرت النتائج بان ھناك تطابق بین النتائج المختبریة والنتائج النظریة عند تطبیق في نتائج الغاز الحیوي والمیثان . ان الزیادة 0.95>وبمعامل ارتباط modified gompertz modelالنموذج الریاضي ت النعام ھي تكنولوجیا واعدة في لمطحون قشور الرز المعالج وفضال تشیر الى ان ان عملیة الھضم المشترك الالھوائي مجال الطاقة المتجددة. ان استعمال فضالت النعام قد ادى الى تعزیز انتاجیة الغاز الحیوي التي لھا دور في تقلیل التلوث البیئي من خالل ادارة المخلفات. فضالت النعام ؛قشور الرز ؛تیةلصوا الموجات فوق ؛راریة المائیةالح ؛المشتركالھضم ؛الغاز الحیوي لكلمات المفتاحیة:ا improving biogas production using different pretreatment of rice husk inoculated with ostrich dung 2 1. introduction the global energy markets depend majorly on fossil fuel like coal, petroleum, diesel, and natural gas. since fossil fuel formation requires trillions of years, its stock is finite and subjects to diminution (saboor et al., 2017). however, ease of access to fossil fuels during almost two centuries has decreased the available fossil fuel reservoirs, causing rising prices. (sari & budiyono, 2014). on the other hand, fossil fuel is considered as one of the main sources of global warming which emits carbon dioxide (co2), methane (ch4), and some other trace gases, which result in many environmental problems and bad effects on human health and ecosystem. therefore, the issue that has become a worldwide concern is the need for renewable, cheap, and clean energy sources, which has grown up with industrialization and global development of technology, as described by noori (2017). anaerobic digestion (ad) has been defined as the biochemical process in which microorganisms breakdown the bulky organic molecule into smaller molecules under anaerobic conditions (srivastava, 2020). lignocellulosic biomass is the earth’s most abundant and renewable resource, and, lignin is its strongest component such as hardwood, softwood, grasses, as well as household, industrial, and agriculture residues that can be used for sustainable methane production (otieno and ogutu, 2020; zieliński et al., 2017). rice husks are agricultural wastes that remain after the processing of the rice crop, like rice straw and rice husk. therefore, rice husks can be improved methane yield, as reported by vivekanandan & kamaraj (2011). lignocellulosic biomass material majorly includes cellulose, hemicellulose, and lignin, the arrangements of the contents inside the biomass makes it an extremely complex structure, as reported by rehman et al. (2013). in order to improve biogas production from lignocellulosic biomass, a pretreatment process is necessary to disrupt naturally recalcitrant carbohydrateelignin shields that impair the accessibility of enzymes and microbes to cellulose and hemicellulose (zheng et al., 2014). besides, this led to the enhancement of lignin and hemicellulose removal, which allows hydrolytic reagents, improved access to the cellulose molecules in the subsequent hydrolysis steps (rehman et al., 2013). from the pretreatments, ultrasound, hydrothermal, combined hydrothermal-ultrasonic, alkaline, and combined alkaline-ultrasound pretreatment. a high alternating voltage is produced by ultrasound energy during the pretreatment of ultrasound. ultrasound waves produce internal negative pressure, which in turn causes forming of small bubbles during the process of cell disruption. ultrasound energy causes a stunning temperature and pressure, which lead to the destruction of the cell membrane (liqian, 2011). hydrothermal pretreatment of lignocellulosic biomass for enhanced bio-ethanol and biogas production is gaining the 21st century. water under high pressure and temperature can penetrate into biomass, hydrate cellulose, and removes hemicellulose and part of lignin, the major advantages are no addition of chemicals and no requirements of corrosion-resistant materials for hydrolysis reactors in the hydrothermal pretreatment process (chandra et al., 2012). alkaline pretreatment, usually employing naoh, is known to break bond linkages between lignin in the lignocelluloses material, causing cellulose swelling which results in greater accessibility of the cellulose fraction (talha et al., 2018). combined pretreatments including hydrothermal-ultrasonic and 3% naohultrasound pretreatment are applied, some studies have found that the synergies of matched and mixed pretreatments might optimize the overall outcome, as reported by liqian (2011). this study was aimed to investigate the influence of single pretreatment) ultrasonic, hydrothermal, naoh (3% w/v)) and combined pretreatment (hydrothermalultrasonic, 3% naoh, -ultrasonic) for rice husks on biogas production and on the characteristics of rice husks. 2. materials and methods 2.1 preparing of substrate and inoculum 2.1.1 preparing of substrate rh was collected from a rice field in al-musayib city (hilla, iraq). after collection, the sample was manually cleaned to remove sand and undesirable particulates. the rice husk was air-dried for three days at 3740 ºc. then, they were milled by using an electrical household grinder and screened by using a mechanical screener to get a size of (0.3-0.6 mm). the size of the sample was selected depending on the previous work (ismail and noori, 2018). the chemical characteristics of rh are shown in table 1. table 1. chemical characterizes of rice husks (rh) and ostrich dung (od). variables rh od ts (%) 94.2 94.4 vs (%) 82.08 32.52 vs/ts 0.88 0.34 cellulose content (%) 33 hemicellulose content (%) 25 lignin content (%) 20 carbon content (%) 48.58 18.86 nitrogen content (%) 0.52 2.33 c/n 93.42 8.09 huda jassim and amal h. khalil 3 2.1.2 preparing of inoculum ostrich dung (od) was used as bacterium sources in this study. six fresh samples (one and two days old) were collected in the year 2019 from the ostrich farm in al-mahawil city, babylon, iraq. then, it was airdried, crushed, and stored in a plastic container. the characteristic of od is clarified in table 1. the inoculum was incubated at 35-37ºc for 1 week to degrade easily degradable materials presented in them and to reduce non-specific biogas generation (saboor et al., 2017; bundó et al., 2017). the slurry of inoculum was provided by blending 300 ml of distilled water with 35 g of crushed od. the mixed slurry was manually homogenized by glass rod, according to the procedure adopted by ismail and noori (2018). 2.2 pretreatment conditions to improve the capability of lignocellulosic rh to digestate through the anaerobic co-digestion, six pretreatments were applied which were ultrasound, hydrothermal, combined hydrothermal and ultrasound, alkaline with naoh (98%, provided by ca, usa), and combined alkaline with ultrasound pretreatment. the details of the pretreatments are explained in table 2. each pretreatment was performed using 30 g from ground rh. ultrasound pretreatment was performed by ultrasound cell crusher (model: sjia-1200w mti, germany). the temperature (25 °c) was selected depending on study kavitha et al. (2016) who reported that the sonication was carried out at room temperature without any temperature regulation. while the time (15 min) was selected depending on study sulʾman et al. (2011) who reported that the optimum conditions for pretreatment of the substrate at which the most complete destruction of the plant material is achieved are an intensity of 368 w/ cm2 and a duration of no more than 15 min., the frequency (20 khz, was selected depending on kavitha et al., 2016), whereas the input power (480) was modified. in electrical autoclave (model: daihan labtech, korea), the hydrothermal pretreatment was done. pretreatment time (15 min) was selected according to (alzate et al., 2012). the temperature (121°c) and pressure (15 psi) were modified. the heating temperature (80 °c) used in alkaline pretreatment was carried out according to the previous research work (hassan et al., 2016a) while the concentration of naoh (3%) and pretreatment time (3hr.) were modified. 2.3 biochemical methane potential (bmp) set up bmp tests were used to compare the anaerobic biodegradability of pretreated and non-pretreated microalgal biomass (fabiana and ivet, 2015), six digesters were performed in a batch system. the digesters included rice husk in case of untreated (rh1) and in case of pretreatments: rh-2, rh-3, rh-4, rh-5, and rh-6. figure (1) states, the arrangement of the lab-scale system used in this study for the anaerobic co-digestion process, which was depending on the previous study (ismail & noori., 2018) with some modifications to dimensions. each digester is represented by 500-ml pyrex borosilicate. inside each digester, the ratio of 1:10 was used according to previous research (ismail & noori., 2018). this ratio indicated 30 g of solid waste: 300 ml of inoculum slurry. a rubber stopper was used to plug each digester. two ports in each one with 5 mm diameter. glass tubes were submerged inside the digester through the two ports. a rubber tube was connected to each glass tube from the other end, which in turn connected to valves to curb the release of generated gas, according to previous research (ismail & noori., 2018). before the start of each batch, the ph of all the reactors was determined after mixing the inoculum and substrate and was adjusted to ph =7 if necessary by adding concentrated hcl solution after adding the inoculums and substrates, according to the procedure adopted by previous studies (sambusiti et al., 2013; saboor et al., 2017). in order to avoid aerobic respiration, oxygen in the headspace was replaced by the purging of nitrogen gas for 3-5mins. the introduced nitrogen will maintain an anaerobic environment and it will be helpful for a survey of methanogenic bacteria for effective biogas production. then the digesters were closed tightly, as explained by kannah et al. (2017). table 2. pretreatment conditions in this study. pretreatment type symbol % naoh temperature (°c) time frequency khz input power w pressure psi untreated rh-1 ultrasound rh-2 25* 15** min 20 hydrothermal rh-3 121 15*** min 15 hydrotherm al-ultrasonic rh-4 the same conditions of hydrothermal and ultrasonic were applied in this pretreatment alkaline rh-5 3 80**** 3 h 480 alkalineultrasound rh-6 all above conditions were applied in this combined pretreatment * depending on study kavitha et al. (2016), **depending on study sulʾman et al. (2011), *** according to alzate et al. (2012), **** according to the previous research work (hassan et al., 2016a). improving biogas production using different pretreatment of rice husk inoculated with ostrich dung 4 figure 1. experimental setup for a biogas production system that composed of the main parts: (1): water bath, (2) activated digester, (3) thermostat, (4) digital thermometer, (5) rubber tube for flushing nitrogen, and (6) manometer. digesters were put in a thermostatic water bath (this arrangement was depending on the previous study (noori, 2017) with some modifications) so that they remain in the required range of temperature which was 50 ºc which was selected according to previous research, as described by deepanraj et al. (2014). 2.4 analytical methods parameters like volatile solids (vs), total solids (ts), total nitrogen (tn), total carbon (tc) and lignocellulosic characteristics were analyzed for untreated and pretreated samples of rh. also, ts, vs, tc and tn (tn was carried out using a kjeldahl digestion system (england)) were estimated depending on the methods described by ismail & noori (2018). the characterizes of lignocellulosic rh which include lignin, hemicellulose, and cellulose were determined depending on van soest̕s method prescribed by van soest et al. (1991) using raw fibres extractor (england). finally, the produced biogas was measured daily by a manometer method, as described by ismail & noori (2018). 2.5 kinetic study for an anaerobic digester operating in a batch mode, the rate of biogas generation corresponds to a specific growth rate of methanogenic bacteria in this digester. accordingly, the predicted rate of biogas production can be calculated using the modified gompertz model (ismail and noori, 2018), the modified gompertz model was carried out using (ibm spss 24.0.1fp2) analytical software. this non-linear regression model has been identified as a good empirical regression model commonly applied in the simulation of methane accumulation, because of its good enough precise prediction for different substrates (talha et al., 2018): g(t) = g0.exp{exp [ rmax.e g0 (λ-t) +1 ]} (1) where, rmax is a maximum production rate of methane, ml/ gvs; g0 is the potential methane production, ml/ gvs; g(t) is the accumulated biomethane at the time t, ml/g. vs; t is measured time, day; λ is the lag-phase, day; and e is exp (1) = 2.7183 (talha et al., 2018; ismail and noori, 2018). 3. results and discussion 3.1 the impact of types pretreatments on the properties of rice husk the goal of the pretreatment process is to break down the lignin structure and disrupt the crystalline structure of cellulose so that acids or enzymes can easily access and hydrolyze cellulose, as reported by kumar et al. (2009). in order to determine the influence of pretreatment with ultrasound (rh-2), hydrothermal (rh-3), combined hydrothermalultrasonic (rh-4), naoh (3% w/v) (rh-5), and combined 3% naoh-ultrasound (rh-6) on the amount of cellulose, hemicellulose and lignin content for rh were explained in table 3. table 3 provides that all pretreatments the chemical composition of rh. the amount of cellulose content was increased from 33% for rh-1 to become 35% for rh-6. the highest cellulose content is again by rh-6 pretreatment. in rh-2 pretreatment, 15 and 20% hemicellulose and lignin reduction were noticed, respectively, compared to an untreated sample (rh1), our results were in good acceptance with sulʾman et al. (2011) who suggested that ultrasound pretreatment caused lignin removal by 11.4% when it's used to treat lignocellulosic biomass. c/n (rh-2) was increased by 3.26% and reduction in ts and vs was obtained by 1.3 and 4%, respectively, compared to an untreated sample (rh-1). for rh-3 pretreatment, all parameters including hemicellulose, lignin, ts, vs, and c/n were decreased by 13.04, 17, 1.55, 3.29 and 3.2%, respectively, as compared with the untreated rh-1. the amount of ts, vs, c/n, hemicellulose and lignin was 90%, 79.67%, 63.19, 20.5% and 15.2% in the rh-4 pretreatment, corresponding to a reduction of 4.46, 2.94, 32.36, 18, and 24%, respectively, as compared to rh-1. in rh-5 pretreatment, the hemicellulose removal was 24% compared to rh-1, while 31% was the removal of lignin, the c/n ratio, ts was decreased by 25.13 and 2.23%, respectively, while vs was increased by 3.34%, as compared with rh-1. our results agree with the study monlau et al., (2015) in which hemicellulose (26%) with subsequent solubilization and removal of lignin (22%), compared to untreated samples because of pretreatment of sunflower stalks with 4 g naoh/ 100 g total solids (ts) at 55 °c for 24 h. huda jassim and amal h. khalil 5 table 3. influence of the pretreatments on the characteristics of rh. variable cellulose content (%) hemicellulose content (%) lignin content (%) ts (%) vs (%) c/n rh-1 33 25 20 94.2 82.08 93.42 rh-2 33.6 21.25 16 92.9 78.82 96.47 rh-3 32.8 21.74 16.6 92.7 79.38 90.43 rh-4 32.5 20.5 15.2 90 79.67 63.19 rh-5 34.5 19 13.8 92.1 84.82 69.94 rh-6 35 17.5 12.6 90.83 83.76 38.85 the highest decrease was obtained during rh-6 pretreatment at 30, 37, 58.41and 3.58% for hemicellulose, lignin, c/n ratio and ts, respectively, in comparison with rh-1. on the other hand, vs was slightly increased by 2.005%. rh-6 pretreatment caused the difference in the chemical compositions and this difference earns it the preference compared to rh-2 and rh-5. our results were agreed with previous research (soontornchaiboon et al., 2016) which showed that the pretreatment by the combination of alkaline and ultrasound changed the chemical composition of all samples (corn cob, pineapple waste, bagasse, rice straw and water hyacinth), which was better than sole alkaline pretreatment. 3.2 the effect of pretreatments for rice husk on the generation of biogas and methane in this study, five pretreatments (rh-2, rh-3, rh-4, rh-5 and rh-6) were used to improve the production of biogas as compared to the control rh-1. table 4 stated the cumulative productions of biogas and methane for all pretreated and untreated samples of rh. the cumulative productions of biogas for rh-2, rh-3, rh-4, rh-5 and rh-6 were 44.19, 39.5, 46.3, 47.16 and 52.38 ml/g vs, respectively; corresponding to biogas increment of 50.72, 34.72, 57.91, 60.85 and 78.65%, respectively, as compared to its value of the control rh-1 which was equal to 29.32 ml/g vs for biogas production, as clarified in figure (2). the cumulative productions of methane for rh-2, rh-3, rh-4, rh-5, and rh-6 were 30.73, 26.78, 32.24, 32.91 and 37.3 ml/g vs, respectively; corresponding to methane increment of 66.11, 44.76, 74.27, 77.89, and 101.62%, respectively, as compared to its value the control rh-1 which was equal to 18.5 ml/g vs for methane yield, as clarified in figure (3). for rh-2 pretreatment, increasing in biogas and methane yield was due to the disruption of the structures of lignocellulosic biomass (hemicellulose and lignin), where lignin transforms most actively during the biodegradation of the pretreated substrate, due to the disruption of its bonds with cellulose upon ultrasonic treatment as reported by sulʾman et al. table 4. effect of pretreatments on cumulative productions of biogas and methane. pretreatment max. biogas production (ml/g.vs) max. methane yield (ml/g.vs) rh-1 29.32 18.5 rh-2 44.19 30.73 rh-3 39.5 26.78 rh-4 46.3 32.24 rh-5 47.16 32.91 rh-6 52.38 37. 3 (2011), and that was confirmed by the hemicellulose and lignin reduction in this study which was achieved in rh-2 pretreatment by 15 and 20%, respectively, compared to the control rh-1, as explained in table 3. rehman et al. (2013) stated that ultrasound generates pressure with rapid cycling. in the rarefaction phase of this pretreatment, the pressure differentials within a solution cause growth and forming of cavitation microbubbles in the sonicated liquid. thus, ultra-sonication produces both physical and chemical effects in the slurry being sonicated. all the above circumstances contribute to the pretreatment of lignocellulosic biomass. our results agree with the previous study deepanraj et al. (2014) reported that the ultrasonic pretreatment at 20 khz in the batch reactor on a palm oil mill and stated that the methane yield enhanced by 16%. and, that increase in biogas production from 3657 l/m3 of was (waste activated sludge) to 4413 l/m3 of was an increase in methane production from 2507 l/m3 of was to 3007 l/m3 of was under optimal ultrasonic pretreatment conditions in a batch reactor. for rh-3 pretreatment, the enhancements, which were stated above for biogas and methane yields, were because of the fact s it is efficient in penetration of the biomass, cellulose hydration, and removal of hemicellulose and part of lignin (hesami et al., 2015). improving biogas production using different pretreatment of rice husk inoculated with ostrich dung 6 the major advantages are that there is no requirement for chemicals and corrosion-resistant material for the reactor. typically, it can remove most of the hemicellulose and part of lignin in biomass by degrading them into soluble fractions and loosen the recalcitrant structure as well (beltrán et al., 2019). the results obtained by rh-4 pretreatment, which showed enhancement in both biogas and methane yields, were because the combined pretreatment was better than individual pretreatment. as rh-4 was combined of two pretreatments which were rh-2 and rh-3, that enhanced its ability to increase biogas and methane production yields. zheng et al. (2014) indicated that combined pretreatment could be beneficial due to higher methane yield and more biomass utilization compared to single pretreatment; however, it may also increase pretreatment costs. the results were in good agreement with the previous study reported by trzcinski et al. (2015) who stated that combination ultrasonication and thermal pretreatment of sewage sludge resulted in a 20% increase in biogas production during the anaerobic digestion of the pre-treated sludge. for rh-5 pretreatment, the results could be due to the fact that reported by zheng et al. (2014) who stated that alkaline pretreatment is thought to be the cleavage and the saponification of carbohydrate-lignin linkages. figure 2. the cumulative production of biogas yield through the digestion process for all untreated and pretreated samples of rh. figure 3. the cumulative production of methane yield through the digestion process for all untreated and pretreated samples of rh. 0 12 24 36 48 60 0 5 10 15 20 25 30 c om ul at iv e bi og as , m l/g v s retention time, days rh-1 rh-2 rh-3 rh-5 rh-4 rh-6 0 8 16 24 32 40 0 5 10 15 20 25 30 c um ul at iv e m et ha ne , m l/g v s solid retention time, days rh-1 rh-2 rh-3 rh-5 rh-4 rh-6 huda jassim and amal h. khalil 7 by the removal of crosslinks, alkaline pretreatment leads to an increase of porosity, an internal surface area, structural swelling, a decrease in the degree of polymerization and crystallinity, disruption of lignin structure, and a breakdown of links between lignin and other polymers. the effective parameters should have increased the porosity of the substrate, which makes the carbohydrates more accessible for enzymes, as stated by liqian (2011). sambusiti et al. (2013) suggested that a 43% increase in methane yield was noticed using pretreatment conditions 10% gnaoh/g ts at 40 ºc for wheat straw in batch mode in comparison with the control sample of wheat straw. the results associated with rh-6 pretreatment stated that it was better than rh-2 and slightly higher than rh-5 in the production of biogas and methane yield. in addition, rh-6 was better in lignin reduction than rh-2 and rh-5, as stated in table (3). this performance of rh-6 pretreatment was due to that the combined pretreatment was better than single pretreatment. as rh-6 was combined with two pretreatments, which were rh-2 and rh-5, made the effect of rh-6 was stronger than the effect of rh-2 or rh-5 in the enhancement of biogas and methane yields. our results had provided an obvious consistency with talha et al. (2018) reported that the combined alkaline and ultrasonic treatment, with different naoh loading rates (0.25%-6%) and pre-treatment times, were applied to enhance methane yield (up to 39.49%) compared to the untreated filter mud. 3.3 kinetic study using modified gompertz model figure (4a-f) and table 5 present the results of the application modified gompertz model, which explained that the estimated values and predicted values were incompatibilities with ρ> 0.95. a: predicted and measure results for rh-1 pretreatment. b: predicted and measure results for rh-2 pretreatment. 0 4 8 12 16 20 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted r2= 0.959 0 7 14 21 28 35 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted rrr2=0.968 improving biogas production using different pretreatment of rice husk inoculated with ostrich dung 8 c: predicted and measure results for rh-3 pretreatment. d: predicted and measure results for rh-4 pretreatment. e: predicted and measure results for rh-5 pretreatment. 0 7 14 21 28 35 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted r r2=0.986 0 7 14 21 28 35 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted r2=0.970 0 7 14 21 28 35 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted r2=0.961 huda jassim and amal h. khalil 9 f: predicted and measure results for rh-6 pretreatment. figure 4. the predicted and measured results for methane yield to all the untreated and pretreated samples of rh for all rectors (a, rh-1); (b, rh-2); (c, rh-3); (d, rh-4); (e, rh-5); and (f, rh-6). table 5. results from performing modified gompertz model. digester no. g(t) exp. (ml methane/g vs) gompertz model parameters r2 λ (day) rmax. (ml methane/g vs) g0 (ml methane/g vs) g(t) model (ml methane/g vs) 1 18.5 1.25 0.93 18.5 16.63 0.96 2 30.73 3.04 1.47 30.73 26.27 0.97 3 26.78 0.53 1.42 26.78 24.69 0.99 4 32.24 1.36 1.59 32.24 28.76 0.97 5 32.91 2.03 1.44 32.91 27.57 0.96 6 37.3 0.17 1.8 37.3 33.88 0.97 4. conclusion the results obtained from anaerobic co-digestion of pretreated rice husks inoculated with ostrich dung were most effective than those of untreated rice husks. in addition, the combined pretreatments for rice husks showed better results than solo pretreatment in biogas and methane production. the combined hydrothermal and ultrasonic pretreatment showed better productivity of biogas and methane than ultrasonic and hydrothermal pretreatments separately. also, the combined alkaline and ultrasonic pretreatment showed better productivity of biogas and methane than alkaline and ultrasonic pretreatments separately. best performance in biogas and methane increase for solo pretreatments was for alkaline pretreatment with 3% naoh by 60.85 and 77.89%, respectively as compared with untreated rice husks. the combined alkaline and ultrasonic pretreatment were preferably recommended for biogas and methane enhancement, it caused an increase by 78.65 and 101.62%, respectively, as compared to untreated rice husks. conflict of interest the authors declare no conflict of interest. funding no funding was received for this study. acknowledgement the authors would like to thank the environmental engineering department and chemistry& microbiology laboratory in the biology department/college of science at the university of 0 8 16 24 32 40 0 5 10 15 20 25 30 m l m et ha ne /g v s solid retention time, days measured predicted r2=0.969 improving biogas production using different pretreatment of rice husk inoculated with ostrich dung 10 babylon. agriculture department/ college of agriculture at the university of baghdad. references alzate m e, muñoz r, rogalla f, fdz-polanco f and pérez-elvira s i (2012), biochemical methane potential of microalgae: influence of substrate to inoculum ratio, biomass concentration and pretreatment. bioresource technology 123: 488494. beltrán j u h, de lira i o h, santos m m c, luevanos a s, terán f h and balagurusamy n (2019), insight into pretreatment methods of lignocellulosic biomass to increase biogas yield: current state, challenges, and opportunities. applied science 9. bundó m s, eskicioglu m g, carrère h, ferrer i (2017), anaerobic co-digestion of microalgal biomass and wheat straw with and without thermo alkaline pretreatment. chandra r, takeuchi h, hasegawa t, and kumar r, (2012), hydrothermal pretreatment of rice straw biomass: a potential and promising method for enhanced biogas production. applied energy 94: 129–140. deepanraj b, velmurugan s and jayaraj, s (2014), "enhancement of biogas production by pretreatmen," ivth international conference on advance in energy research, indian institute of technology bombay, mumbai, pp. 309-319. fabiana p, and ivet f (2015), influence of hydrothermal pretreatment on microalgal biomass anaerobic digestion and bioenergy production. water research 6 8: 3 6 4 -3 7 3. hassan m, ding w, bi j, mehryar e, talha z a a and huang h (2016a), methane enhancement through oxidative cleavage and alkali solubilization pre-treatments for corn stover with anaerobic activated sludge. bioresource technology 200: 405-412. ismail z z, and noori n a (2018), anaerobic codigestion of giant reed for biogas recovery. journal of engineering 24(3): 68-83. kannah r u, kavitha s and banu j r (2017), bmp a tool for assessing energy efficient biogas production. biotechnology and microbiology 5(2): 1-5. kavitha s, brindha g m j, gloriana a s, rajashankar k, yeom i t, and banu j r (2016), enhancement of anaerobic biodegradability potential municipal waste activated sludge by ultrasonic aided bacterial disintegration. bioresource technology 200:161-169. kumar p, barrett d m, delwiche m j, strove p (2009), methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. industrial engineering and chemistry research 48: 3713-3729. liqian w (2011), different pretreatments to enhance biogas production: a comparison of thermal, chemical and ultrasonic methods, m. s. thesis, university of halmstad, sweden. monlau f, kaparaju p, trably e, steyer j p and carrere h (2015), alkaline pretreatment to enhance one-stage ch4 and two-stage h2/ch4production from sunflower stalks: mass, energy and economical balances. chemical engineering journal 260: 377-385. noori n a (2017), sustainable approach biogas recovery from co-digestion of giant reed for clean energy production, m. s. thesis, university of baghdad, iraq. otieno j o and oguto f o (2020), a review of potential of lignocellulosic biomass for bioethanol production in kenya. asian journal of chemical sciences 8(2):34-54. rehman m s u, kim i, chisti y and han j i (2013), used of ultrasound in the production of bioethanol from lignocellulose biomass. energy education science and technology part a: energy science and research 30(2):1931-1410. saboor a, khan s, shah a a, hasan f, khan haji and badshah m (2017), enhancement of biomethane production from cattle manure with codigestion of dilute acid pretreated lignocellulose. international journal of green energy 14: 632-637 . sambusiti c, monlau f, ficara e, carrère h and malpei f (2013), a comparison of different pretreatments to increase methane production from two agricultural substrates. applied energy 104: 62-70. sari f p and budiyono (2014), enhanced biogas production from rice straw with various pretreatment: a review. waste technology 2(1):17-25. soontornchaiboon w, kim s m, pawongrat r (2016), effects of alkaline combined with ultrasonic pretreatment and enzymatic hydrolysis of agricultural wastes for high reducing sugar production. sains malaysiana 45(6): 955–962. srivastava s k (2020), advancement in biogas production from the solid waste by optimizing the anaerobic digestion. waste disposal sustainability energy 2:85-103. sul’man e m, m g sul’man, and prutenskaya e a (2011), effect of ultrasonic pretreatment on the composition of lignocellulosic material in biotechnological processes. biocatalysis 3(1): 2833. talha z, hamid a, guo d, hassan m, mehryar e, okinda1 c and ding w (2018), ultrasound assisted alkaline pre-treatment of sugarcane filter mud for performance enhancement in biogas production. international journal of agricultural and biological engineering 11(1): 226-231. trzcinski a p, tian x, wang c, lin l l and ng w j (2015), combined ultrasonication and thermal pre-treatment of sewage sludge for increasing huda jassim and amal h. khalil 11 methane production. journal of environmental science and health, part a 50: 213-223. van soest p j, robertson j b and lewis b a (1991), symposium: carbohydrate methodology, metabolism, and nutritional implications in dairy cattle, journal of dairy science, 74 (10). vivekanandan s and kamaraj g (2011), the study of biogas production from rice chaff (karukka) as co-substrate with cow dung. indian journal of science and technology 4(6): 657659. zheng y, zhao j, xu f and li y (2014), pretreatment of lignocellulosic biomass for enhanced biogas production. progress in energy and combustion science 42: 35–53. zieliński m, dęowski m, kisielewaska m, noweska a, rokicka m, szware k (2017), comparision of ultrasonic and hydrothermal cavitaions pretreatments of cattle manure mixed with wheat straw on fermentative biogas production. waste biomass ware. تحسين إنتاج الغاز الحيوي باستخدام معالجات مسبقة مختلفة لقشور الأرز الملقح بفضلات النعام 1. introduction conflict of interest funding srivastava s k (2020), advancement in biogas production from the solid waste by optimizing the anaerobic digestion. waste disposal sustainability energy 2:85-103. the journal of engineering research (tjer), vol. 16, no. 1 (2019) 44-52 doi: 10.24200/tjer.vol16iss1pp44-52 development of chicken feather reinforced insulation paperboard from waste carton and portland cement j.k. odusote a,* and k.s. dosunmu b a department of materials and metallurgical university, university of ilorin, ilorin, nigeria. b department of mechanical engineering, landmark university, omu aran, kwara state, nigeria. abstract: cartons and chicken feathers are common wastes which we need to dispose in one way or another. disposal problems associated with these wastes can be solved by processing them into useful products such as insulation and ceiling boards. in this study, chicken feather reinforced ceiling board was developed from waste carton and portland cement. the quantity of the chicken feather was kept constant at 10% based on previous findings, while the cement and waste carton contents were varied to produce 5 samples of different compositions. the density of the board was found to range between 337.8 and 700.7 kg/m 2 , while the thickness swelling ranges between 0.81 and 9.02%. water absorption values of the samples varied between 7.16 and 24.41%, while the compressive strength and modulus of elasticity values varied from 4.8 10.3 n/mm 2 and 1.03 1.60 gpa, respectively. the values of modulus of rupture ranges between 1.34 and 2.2 mpa while the thermal conductivity of the samples ranges from 0.951 to 1.077 w/m.k. density, compressive strength, modulus of elasticity, modulus of rupture and thermal conductivity of the samples increased as the cement content increased, while the thickness swelling and water absorption decreased with increase in cement content. the results revealed that the properties of ceiling boards developed from 80% cement, 10% carton and 10% chicken feather can compete favorably with most ceiling boards available in the market. keywords: carton; chicken feather; cement; physical properties; thermal conductivity; waste management. تطوير لوح ورقي عازل مقوى من ريش الدجاج ونفايات الكرتون وأسمنت بورتالند *اودوستو . ك .ج أ دوسونمو س. .ك ، ب يعد الكرتون وريش الدجاج من نفايات البيئة والتي يجب التخلص منها بطريقة ما أو بأخرى. : الملخص بتحويلها إلى منتجات مفيدة مثل ألواح عزل واسقف. ففي هذه فيمكن حل مشاكل التخلص من هذه النفايات الدراسة ، تم تطوير لوح سقف مدعم بالريش من نفايات الكرتون وأسمنت بورتالند. بقيت كمية ريش ٪ بناًء على النتائج السابقة ، بينما تفاوتت محتويات نفايات الكرتون والخرسانة 01الدجاج ثابتة عند نسة كجم/م 711.7و 7.3..ت من التراكيب المختلفة. وتبين أن كثافة اللوح تتراوح بين عينا 5إلنتاج . ، في و 7.07٪. وتراوحت قيم امتصاص الماء للعينات بين 2.19و 1.30حين يتراوح معدل االنتفاخ بين نيوتن / مم ..01-4.3٪ ، بينما تفاوتت قوة الضغط ومعامل قيم المرونة من 94.40 9 جيجا، 0.71-.0.1و ميجا باسكال ، في حين تتراوح نسة التوصيل 9.9و 4..0على التوالي. وتتراوح قيم معامل التمزق بين واط/م ك. كما زادت كثافة وقوة االنضغاط ومعامل المرونة 1.0177إلى 1.1250الحراري للعينات من حين تناقص سمك االنتفاخ ومعامل التمزق والتوصيل الحراري للعينات مع زيادة محتوى األسمنت ، في وامتصاص الماء مع زيادة محتوى األسمنت. وبينت النتائج أن خصائص ألواح األسقف التي تم تطويرها ٪ ريش الدجاج يمكن أن تتنافس بشكل إيجابي مع معظم ألواح 01٪ كرتون و 01٪ أسمنت و 31من االسقف المتوفرة في السوق. دجاج؛ االسمنت؛ الخصائص الفيزيائية؛ التوصيل الحراري؛ إدارة كرتون؛ ريش ال: المفتاحية الكلمات المخلفات. *corresponding author’s e-mail: odusote.jk@unilorin.edu.ng mailto:odusote.jk@unilorin.edu.ng the journal of engineering research (tjer), vol. 16, no. 1 (2019) 44-52 1. introduction wastes are generated from virtually all human activities in solid, liquid or gaseous forms. carton and chicken feather are solid wastes that can be found across nigeria, causing a serious environmental problem. carton is a form of paper, which has been utilized mainly for packaging in everyday activities for more than 20 decades. after usage, cartons are usually discarded, thereby forming major components of municipal solid wastes. thus, there is a need to develop alternative methods of disposing or re-processing of this waste for re-using. chicken feathers, on the other hand, are waste products from poultry processing industries. it was estimated in 2015 that nigeria has over 180 million chickens (birds) (national veterinary research institute 2015). most of these chickens are supplied to supermarkets and fast food outlets to meet the demand for white meat. however, since about 6% of the total weight of mature chicken is feather, large amount of feathers are generated as waste through the processing of chickens (parkinson 1998; mcgovern 2000). traditional disposal strategies of chicken feathers and cartons/papers, as shown in fig. 1, are unhealthy and uneconomical. they are often burned in incineration plants or buried in landfills. burning of these wastes generates greenhouse gases and poses a great danger to the wellbeing of people living around the dumpsites. in addition, burning can also damage the environment. disposal problems associated with these wastes can be solved by processing them into useful products such as ceiling boards, notice boards, paperboards, etc. odusote et al. (2016). thus, the aim of this work is to utilize corrugated paper (carton) to develop chicken feather reinforced insulation ceiling boards using a varied amount of ordinary portland cement as binder. these boards will be eco-friendly and they are more likely to be cheaper. the boards can also find application in large quantities in construction industry. the use of this eco-friendly ceiling boards in construction industry will reduce the reliance on imported materials for roofing and assist in solving the environmental challenges associated with unhealthy disposal of cartons and chicken feathers. 2. materials and methods 2.1 materials the materials used for this study are corrugated paper (cartons), chicken feathers, portland cement (commercially available dangote 3x portland cement of grade 42.5n with composition as presented in table 1), water, 600 mm x 600 mm fiber glass moulds, trowel, straight edge, a labtech bl20001 electronic compact scale, ball mill (model 48-d0500/d and serial no. 14002201), electric oven model (sdo/225 and serial no. y9c227), and a vibratory sieving machine. according to fraser and parry, (1996), chicken feathers contain 91% protein (keratin), 1% lipids, and 8% water. keratin fibers are strictly nonabrasive, having low density, biodegradable, renewable, eco-friendly, insoluble in organic solvents, hydrophobic, warmth retention and cost effective meyers et al. ( 2008). the density of chicken feathers is usually about 0.89 g/cm 3 , while the specific gravity ranges between 0.7 and 1.2, and the strength varied between 41 – 130 mpa hong and wool (2005). figure 1. traditional disposal method of wastes: (a) chicken feather (b) carton and household wastes. table 1. composition of ordinary dangote portland cement (yahaya 2009). component cao sio2 al2o3 fe2o3 *loi *ir *free cao so3 mgo quantity (wt %) 63.14 16.56 4.78 10.5 5.22 2.03 2.15 1.4 1.45 *loi = loss on ignition; ir = insoluble residue and free cao = free lime. 45 j.k. odusote and k.s. dosunmu 2.2 samples preparation staple pins, cenotaph and other unwanted materials were manually removed from the collected cartons, which were then soaked in about 200 liters of water for a period of 5 days to ensure adequate softening. after soaking, the mixture was first pulverized using mortar and pestle and then followed by ball mill (model 48d0500/d) to obtain the finest texture. the pulverized cartons were then sieved using a 1 mm sieve size and drained to reduce the amount of water present in the formed slurry. the dewatered carton was then dried in the sun for 3 days, and then further pulverized in the ball mill for about 1 h and sieved in the vibratory sieving machine using 500 µm sized sieve. the feathers were washed several times with warm water mixed with detergent to remove blood, manure and extraneous materials. the clean feathers were then spread on aluminum sheet to dry for 3 days. it was further dried in an oven at 150℃ for 1 h before grinding in the ball mill into powder of about 150 µm. the compositions of carton, cement and chicken feather in the 5 different samples of the insulation ceiling boards are presented in table 2. chicken feather was kept constant at 10% for all the formulations based on the result of acda (2010). it was reported that increasing the proportion of chicken feather above 10% resulted in the significant reduction of modulus of elasticity and modulus of rupture. the carton/cement/chicken feather mixture was poured into an already made mould which was lubricated with engine oil to enhance easy removal and prevent the sticking of the board to the plate surface. after pouring, a stainless pipe roller was used to spread the mixture on the mould to allow for uniform mat formation. a pressure of 80 kg/m 3 was applied on the mixture in the mould after uniform spreading for 8 hours for proper setting. the mould was then dismantled to remove the cast ceiling board, which was then dried in the sun for 3 days to allow for proper drying. after drying, the cast was allowed to cure for 28 days. 2.3 testing procedure the board edges were trimmed with a circular saw to avoid edge effect and cut into the required sizes for each test experiment. the following tests were carried out on all the amples: physical appearance (presence of crack); density; thickness swelling; water absorption; compressive strength; table 2. mixing ratios for the insulation boards formulation. formulation (wt %) a b 80 70 10 20 10 10 c d e 60 50 45 30 40 45 10 10 10 and thermal conductivity. 2.4 density the sample for board density evaluation of size 50 mm × 50 mm × 10 mm was cut from the produced board. the samples were weighed and the density was determined using: volume mass density  (1) where mass is in kg and volume is in m 3 . 2.5 thickness swelling this was carried out by measuring the samples initial thickness (t1) with the aid of a mitutoyo digital caliper. the sample was then soaked in water for 24 hours and the final thickness (t2) was measured with the same caliper. the thickness swelling was estimated as a percentage of the change in thickness. 2.6 water absorption the initial weight of the sample of dimensions 50 mm 50 mm 10 mm that was cut out of the produced board was measured using an electronic compact scale, a labtech bl20001 of sensitivity 0.1 g. the samples were soaked in water for 24 hours before they were then surface dried and re weighed for their final weight. the initial weight w1and the final weight w2were noted and the water absorption rate was calculated following astm c 642 -06 standard using equation 2. water absorption = 100 w ww 1 12          (2) 2.7 compressive strength this test was conducted following astm c 10995 standard using a 50 kn testometric computer controlled universal material testing machine, model number m500-50at. each ϕ38 mm 76 mm test specimen, was mounted on the jaws of the machine one after the other. the loading was increased gradually from 0 n at a test speed of 100 mm/min until the sample was fractured. the measurements were displayed on the monitor attached to the testometric machine as shown in fig. 2. 2.8 flexural test flexural test is the measure of the force required to bend a beam under three-point loading. test samples of 600 mm x 25.4 mm x 5 mm were mounted on an improvised test fixture on testrometric universal testing machine one after the other. the samples were placed on two supporting pins as shown in fig. 3, which shows 46 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 44-52 a sample under a load in a three-point bending setup. the load applied at fracture point, the length of the support span, the width, thickness and deflection values were used to calculate the modulus of elasticity (moe) and modulus of rupture (mor) using equations 3 and 4, respectively bansal (2010). moe= 3 3 hbd4 pl (3) and mor= 2 bd2 pl3 (4) where p = ultimate failure load (n) l = the span of board sample between the machine supports (mm); b = width of the board sample (mm); d = thickness of the board sample (mm) and; h = maximum deflection at the center of the beam. 2.9 thermal conductivity the following parameters were obtained from arm field computer compatible linear heat conduction apparatus: heater voltage v (volts); heater current i (amps); heated section at high temperature (℃); heated section at mid temperature (℃); heated section at low temperature (℃); cooled section high temperature (℃); cooled section at mid temperature (℃); cooled section at low temperature (℃); and thickness of the specimen x(m). thermal conductivity of the specimen, ksp= )m( 4 d 2 sp (5) where: heat flow (power of heater), q = vi (watts) (6) cross sectional area of specimen, asp= )m( 4 d 2 sp (7) temperature difference across specimen, )c(ttt coldfacehotfacesp  and temperature of hot face of specimen =   )c( 2 tt tt 32 1hotface    (8) temperature of cold face of specimen =   )c( 2 tt tt 87 6coldface    (9) figure 2. set up for compressive test. figure 3. flexural test showing a sample under a load in a three-point bending setup. figure 4. computer compatible linear heat conduction apparatus setup for determining thermal conductivity. 47 j.k. odusote and k.s. dosunmu where ∆xsp is the thickness of the specimen board. the following constants are applicable: the diameter of the specimen in contact with the bar = 0.025 (m). the distance between each thermocouple = 0.015 (m). the distance between thermocouple t3 or t6 and the end face i = 0.0075 (m). the conductivity of the brass sections is approximately = 121 (w/m °c). 3 results and discussion 3.1 physical appearance the first formulation that was produced is sample e and it was observed that the sample edges cracked, as shown in fig. 5, while the cast was being removed from the mould. this was however taken care of in the subsequent cast by proper lubrication of the mould prior to pouring of the mixed formulation and adequate force was applied to the board before removal from the mould. in addition, the chamfering at the edges of the mould was increased in order to allow for easy removal. samples a, b, c, d and e of the produced ceiling board are shown in fig. 6. it can be observed that there are variation in the samples’ colors which is attributed to the composition process, that is, chicken feather and paper contents in figure 5. first cast (sample e: 45 – 45 -10 weight %) of the ceiling board. each of the samples. sample a is found to be bluish while sample e is brownish, taking the color of carton, since it has the highest carton content of 45% by weight. 3.2 density the densities of the samples are presented in fig. 7. the densities of the samples were found to range between 337.8 and 700.7 kg/m3. the densities increased with the increase in cement content. this indicates that the more the cement content the lesser the voids in the particle boards and thus the higher the density. the densities of the ceiling boards in the present study were lower than those of the other types of ceiling boards such as asbestos ceiling board (1500-1950 kg/ m3), fiber cement flat sheets (1250-1350 kg/ m3) and trilite (1000-1050 kg/ m3) as reported by ataguba (2016). lower density implies that the load that the ceiling board will exert on the walls and the reinforcement of buildings will be lower. however, the density should not be too low so that it can also support the structure. figure 6. produced ceiling board samples showing color variations. 48 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 44-52 3.3 swelling thickness and water absorption thickness swelling and water absorption of the ceiling board decreased with increased quantity of the cement as shown in fig. 8 (a & b). thickness swelling (fig. 8a) has values ranging from 0.81% to 9.02% while the water absorption (fig. 8b) ranges between 7.16 and 21.41%. boards produced from sample e exhibited the highest thickness swelling and water absorption. this is due to the fact that, low density boards have more permeable voids than dense ones so that more water can be absorbed. in addition, large numbers of micro voids present in the sample can accelerate the penetration of water by the capillary action (salem and al-salami 2016). after saturating the cell wall, water then occupies the micro voids. therefore, increasing board density decreases the voids and reduces swelling rate and water absorption. 3.4 modulus of elasticity and modulus of rupture. modulus of elasticity (m.o.e.) and modulus of rupture (m.o.r.) of the ceiling board increased with the increased quantity of the cement as shown in fig. 9 (a & b). modulus of elasticity (fig. 8a) has values ranging from 1.03 gpa to 1.6 gpa while the modulus of rupture (fig. 9b) ranges from 1.34 mpa and 2.2 mpa. boards produced from sample e exhibited the lowest m.o.e. and m.o.r. this is due to the fact that porosities and voids in the composite ceiling board increased as the cement content decreased thereby reducing flexural and load bearing capacity of the board. figure 7. the influence of sample composition on density of the board produced. figure 8. the influence of sample composition on (a) swelling thickness (b) water absorption of the ceiling board. 49 j.k. odusote and k.s. dosunmu . 3.5 compressive strength the compressive strength of the ceiling board is presented in fig. 10 (a). it is observed that the compressive strength decreased with the increase in the carton content of the mix. sample a has the highest compressive strength of 10.3 n/mm 2 . this is probably due to the fact that higher amount of portland cement is present in the sample, since cement has high compressive strength. sample e, which is 45% carton, has the least compressive strength of 4.8 n/mm 2 this is probably due to the fact that the carton content yields a lot of voids and porosity in the samples, leading to a lower compressive strength paramisivam et al. (1984). 3.6 thermal conductivity the thermal conductivity of the samples is relatively low ranging from 0.1077 to 0.0951 (w/m.k), which is decreased with the increase in the carton content of the mix as shown in fig. 10 (b). this is due to the inclusion of supplementary cementitious materials has composites which; according to kim et al. (2003), will greatly reduce the thermal conductivity, but increase porosity. porosity in composite materials has dramatic influence on thermal conductivity; increasing the pore volume will, under most circumstances, result in reduction of thermal conductivity. in fact, many materials that are used for thermal insulation are porous. heat transfer across pores is ordinarily slow and inefficient. internal pores normally contain still air, which has an extremely low thermal conductivity. figure 9. the influence of sample composition on: (a) modulus of elasticity; and (b) modulus of rupture of the ceiling board. figure 10. the influence of sample composition on: (a) compressive strength; and (b) thermal conductivity. 50 the journal of engineering research (tjer), vol. 16, no. 1 (2019) 44-52 as the carton content in the samples increased, the thermal conductivity and the density decreased, because the extent of voids in the samples increases considerably. the thermal conductivity of a porous medium is inversely proportional to the level of voids in the specimen, so it can be concluded that the thermal insulation efficiency of a material is inversely proportional to its density (salas alvarez 1986). in asbestos cement sheets, asbestos cement board, asbestos loosely packed and asbestos mill board, the values for thermal conductivity are 0.166, 0.744, 0.15 and 0.14 (w/m.k), respectively (www.engineering toolbox.com). comparing these values with the samples in the present study revealed that they have better thermal conductivity properties than all the asbestos boards. table 3 compares the density and thermal conductivity of insulating ceiling board produced in this study with asbestos and nigerite board. 4. conclusion waste carton and chicken feather in cement bonded composites can be used to produce ceiling boards and also for any other non-structural applications in low cost housing projects in nigeria. sample a (80%-cement, 10% chicken feather, 10% carton) gave the lowest water absorption and swelling thickness, and thus recommended for ceiling board applications. sample e (45% cement, 10% chicken feather, 45% carton) has weight advantage over other insulation board samples, and can hereby be used as notice board in offices. the produced board samples produced were better in terms of density and thermal conductivity compared with other available ceiling boards including nigerite boards. cartons and chicken feathers are readily available in our environment and they are often discarded as wastes. thus, insulation ceiling boards produced from these wastes are expected to be relatively cheaper than most commercially available boards and should be more environmentally friendly. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. acknowledgment the authors wish to thank the technical staff of the department of mechanical engineering as well as the management of landmark university omu aran, kwara state, nigeria, for their technical support and provision of facilities for the laboratory tests, respectively. references acda m (2010), sustainable use of waste chicken feather for durable and low cost building materials for tropical climates. sustainable agriculture: technology, planning and management 353-366. ataguba c.o. (2016), properties of ceiling boards produced from a composite of waste paper and rice husk. international journal of advances in science, engineering and technology (2): 117121. astm c642-97 (1997), standard test method for density, absorption, and voids in hardened concrete. astm international, west conshohocken, pa, available online at www.astm.org. astm c 109-95 (1995), standard test method for compressive strength of hydraulic cement mortars. astm international, west conshohocken, pa, available online at www.astm.org. table 3. properties of insulation boards produced compared with asbestos and nigerite board. material density (kg/m3) thermal conductivity (w/mk) reference(s) asbestos board 1500 – 1950 0.15 -2.07 ataguba, (2016), nigerite board 1350 0.2 http://www.nigeritelimited.com a 700.7 0.1077 this study b 547.0 0.0973 this study c 443.5 0.0962 this study d 373.9 0.0958 this study e 337.8 0.0951 this study 51 http://www.engineering/ http://www.astm.org/ http://www.astm.org/ http://www.nigeritelimited.com/ j.k. odusote and k.s. dosunmu bansal r.k. (2010), engineering mechanics and strength of materials, laxmi publications 678. bs en iso 8990 (1996), thermal insulations determination of steady-state thermal transmission properties – calibrated and guarded hot box. british standards institute. engineering tool box (2003), thermal conductivity of common materials and gases. available at: https://www.engineeringtoolbox.com/thermalconductivity-d_429.html accessed: 20th june 2016. fraser r.d.b, and parry d.a.d. (1996), the molecular structure of reptilian keratin. international journal of biological macromolecules 19(3): 207-211 hong c.k. wool r.p. (2005), development of a bio-based composite material from soybean oil and keratin fibers. journal of applied polymer science 95: 1524-1538. khedari j, nankongnab n, hirunlabh j. teekasap s (2004), new low cost insulation particleboard from mix of durian peel and coconut coir. build environ, 39: 59-65. kim d, naaman a, el-tawil e (2008), comparative flexural behavior of four fibers reinforced cementitious composites. cem concer compos, 917-928. mcgovern v (2000), recycling poultry feather. environmental health perspective, 108: 336339. meyers m.a, chen p.y, lin a.y, seki y (2008), biological materials structure and mechanical properties. progress in materials science 53(1): 200-206. nigerite (2000), http://www.nigeritelimited.com, accessed: 18th december 2016). odusote j.k, onowuma s.a. fodeke e.a. (2016), production of paper board briquette using waste paper and sawdust. the journal of engineering research 13(1): 80-88. pakravan h.r, jamshidi m, latifi m. pachecotorgal f (2013), cementitious composites reinforced with polypropylene, nylon and polyacrylonitile fibres. materials science forum 271-276. rajput r (2010), a text book of engineering thermodynamics. new delhi: laxmi publications. salem r.m. al-salami a.e. (2016), preparation of waste paper fibrous cement and studying of some physical properties. civil and environmental research 8(3): 42-54. salas j, alvarez m, vessa j (1986), lightweight insulation concrete with rice husk. international journal of cement composite lightweight concrete 8(3): 171-180. usda foreign agricultural service l. (2008), global agriculture information. usda foreign a gricultural s ervices, livestock and poultry. washington, dc: world markets and trade. vachira s, panya s, nuchthana p (2012), repulped waste paper-based composite building materials with low thermal conductivity. vachira sanggrutsamee, panya srichandr and nuchthana poolt jourmal of asian architecture and building engineering 151(05): 56-61. yahaya m.d. (2009), physico-chemical classification of nigerian cement. australian journal of technology 12(3): 164-17. 52 http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html http://www.nigeritelimited.com/ comp060419.qxd the journal of engineering research vol. 4, no.1 (2007) 95-102 1. introduction during the last decade, several standards for mobile telecommunications have been developed. the most common is the global system for mobile communications (gsm), standardized by the european telecommunications standards institute (etsi), as a cellular communications system operational in many countries around the globe. as gsm depends mainly on integrated digital services network (isdn) standard (rahnema, m., 1993), it is a successful platform for voice calls and data messages. despite the exceptional features of gsm, it could not be considered as a typical stand for public safety and military applications requirements (tetra touch, nokia tetra customer newsletter, 2002; ___________________________________________ *corresponding author’s e-mail: m.dababneh@surrey.ac.uk heikonen, k. et al. 2002). thus, a new brother for gsm is the trunked mobile radio system that is being implemented in europe and is known as, terrestrial trunked radio (tetra), it is also developed by (etsi) providing pure digital information technology for the transmission of speech and data. tetra has been chosen as the platform for the operation of nationwide trunked radio networks in europe; thus offering efficient and flexible multimedia communication services from both private mobile radio (pmr) and public-access mobile radio (pamr) users (etsi, terrestrial trunked radio, 2005). as the cellular mobile communications market experiences hot competition, it is essential to compare the performance of the competing system types designed for data transmission such as the general packet radio system (gprs) stemming from gsm and others with tetra to pin point exactly the real usage and applications of the performance of tetra and gsm\gprs as a platform for broadband internet and multimedia evolution m. i. dababneh*1 and e. s. al-khawaldeh 1* department of electrical engineering, alisra university, jordan received 19 apr 2006; accepted 30 september 2006 abstract: in this paper, a novel approach for the comparison between tetra system and other contenders is presented showing the relative potentials, thence applications and possible new services. tetra voice plus data (v+d) and gprsbased-gsm performance are also compared for defined multimedia transmission to revise the capability of both systems to afford multimedia applications to large number of subscribers with the shortest time possible while maintaining the optimal channel utilization. performance measurements for data transmission illustrate that tetra exhibits faster time response for connection setup time than the gprs\gsm applications can provide. simulation results also confirm that tetra can support larger number of mobile stations (mss) than the gsm\gprs for the same time delay period, yet tetra requires minimal network resources in comparison with gsm\gprs for the same supported service. indeed this is important in the present time where the demand on wireless communication is increasing dramatically and this what stimulates the concept of this research. also, the hypothesis that certain communications systems utilizes spectrum efficiency more effectively than others and the fact that radio spectrum is a scarce recourse have further encourages the start of this research. keywords: tetra, gsm, gprs, call setup-time, throughput ωé¶f agocggsm/gprs andratet.ä’éûg o~©àÿg h á° jô©dg ájooîdg á©°ùddg ägp âfîf’g qƒ¡¶d √~dgƒÿg .¢s .´ h1@á«æhého .´ .ω áá°°uuóóÿÿgg ωé¶f áfqé≤d √ôµàñeh i~j~l á≤jôw ¢vôy ” ,ábqƒdg √òg ‘ :ratet,áj~÷g äée~ÿg á«féµegh äé≤«ñ£àdg ¢vôy ”h éªc á«ñ°ùædg ägaéøµdg úæ«ñe ¬«°ùaéæeh ωé¶f agog áfqé≤e ”hratet ωé¶f ™e 䃰üdgh äéeƒ∏©ÿg qƒ¶æe øegprsωé¶f ≈∏y g~ªà©egsm~jgõàÿg øe~îà°ùÿg o~y á«£¨àd ᪶f’g á©lgôe ”h gòg , ωé¶f ¿g í°vƒj äéeƒ∏©ÿg åh ag~d ¢sé«b .á浇 ájooôj iaéøc ≈∏yg ≈∏y á¶aéùéhh øµ‡ âbh ô°übéhratetωé¶æd ∂dp ™e áfqé≤ÿéh π«°uƒà∏d ´ô°sg ühéœ âbh ∂∏á gprs ωé¶f ¿céh éª∏y ,úeé¶ædg úh …hé°ùàe ònéàdg âbh ¿g úñj iécéùg èféàf ∂dòchratetéæàbh ‘ g~l º¡e ∂°t óh gòg πbg ᫵ñ°t äé«féµeg ¤g êéàëj øe øµªàæd á浇 á≤jôw π° acéh ooîdg ∫ó¨à°sg öéj á«∏yh oh~fi q~°üe ägp ägooîdg ¿g ¤g áaé°v’éh ,᫵∏°sódgh ájƒ∏ÿg ä’é°üj’g ≈∏y òñc ~jgõj ∑éæg ¿g á°uénh ô°vé◊g .ájƒ∏ÿg ä’é°üj’g ᪶fg ≈∏y i~jgõàÿg ö∏£dg áà«£¨j áá««mmééààøøÿÿgg ääggooôôøøÿÿgg ,a飩dg ,áÿéµÿg ¢ù«°séj âbh :ra, gsm, gprstet 96 the journal of engineering research vol. 4, no.1 (2007) 95-102 various systems depending on the associated characteristics, and to explore further services that were not originally planned for these systems. therefore, the aim of this paper is to study tetra performance under the operating mode of v+d. thus, comparing it with the gsm\gprs network regarding the time delay and spectrum channel utilization needed to access the data by the mss. it is essential however; that the network operators supply the subscribers with the required data under demand while maintaining the optimal channel utilization and maximum possible data rate (throughput). this paper starts in section 2 with an outline of tetra together with an overview of the gsm\gprs systems. tetra network simulation analysis is then presented in section 3 followed in section 4 by performance evaluation of the gsm\gprs and complete description of the system. section 5 describes the relative performance between tetra and gprs networks. finally, the conclusion section together with a set of recommendations concludes the paper. 2. overview of tetra (v+d) and gsm/ gprs systems this section introduces the main features for tetra (v+d) and gsm\gprs systems. more information about both systems' protocols, signaling and architecture are beyond the subject of this paper and can be obtained from (rahnema, m., 1993; tetra touch, nokia tetra customer newsletter, 2002; etsi, terrestrial trunked radio, 2005), and (bettstetter, c, et al. 1999). 2.1 overview of tetra (v+d) system tetra is a fully digital cellular private mobile radio (pmr) architecture designed for reliable multimedia applications. the main reason to offer the new standard is that the need to provide a standardized network base for public safety and security (pss) that requires a network capable of support the following five main features detailed in (tetra touch, nokia tetra customer newsletter, 2002; heikonen, k. et al. 2002): these above requirements are entirely covered by tetra. however, two modes of operations are defined for tetra, one of which is the (v+d) mode of operation that allows a unique combination of point-to-point communication, point-to-multipoint communication, mobile telephony, and mobile data services accessible from one radio service (etsi, terrestrial trunked radio, 2005). tetra assigns uplink and downlink traffic separately. it assigns a user up to four time slots per time division multiple access (tdma) frame. when each slot transmits at a rate of 7.2 kbps, the four time slots, when combined, increases the data rate to 28.8 kbps in an operation known as slot stealing, bandwidth on demand, or multislot operation. this feature makes tetra an efficient standard for mobile communications having a frequency spectrum efficiency equivalent to 6.25 khz per channel per user. thus, tetra is considered as a typical platform for data and multimedia applications such as short data messages, status messages and internet protocol for packet data services. tetra also supports generic and tailored data applications such as command and control systems, automatic vehicle locations, database queries, reporting, wireless access protocol (wap) solutions, image transfer, and video transmission (lammerts, e, et al. 1999 and shearer, e.h.s., 1995). 2.2 overview of gprs based-gsm system a practical extension of gsm is satisfied since 2001 when the service of general packet radio system (gprs) is implemented in europe and soon around the globe. gprs is a new service bearer for gsm that greatly improves and simplifies wireless access to packet data networks, e.g., to the internet. it transfers user data packets in an efficient way between gsm mobile stations and external packet data networks. packets can be directly routed from the gprs mss to packet switched networks. networks based on the internet protocol (ip) and x.25 networks are supported in gprs (bettstetter, c, et al. 1999 ). as an improvement over gsm, gprs supports multislot operation, in a similar way used in tetra, such that each time slot allows data to be transmitted at a rate of 9.05 to 21.8 kbps and up to 171.2 kbps for the total 8 tdma slots. as in tetra, gprs uplink and downlink traffics are allocated separately to provide higher efficiency. under mss request for transmission, gprs cell allocates physical channels to support gprs users called packet data channels (pdch)s where each pdch contains multi logical channels for traffic and signaling (bettstetter, c, et al. 1999). 3. performance evaluation of tetra 3.1 performance criteria and simulation model 3.1.1 call setup time for pmr subscribers, one of the most important performance criteria is the call setup time. it is essential that a subscriber can request calls and send\download data as fast as possible. this depends critically on the call setup procedure used to access a channel. the frame timing for the call setup procedure considered in tetra is shown in fig. 1, which shows the call procedure for two mobile stations; ms a calling ms b via the base station. the call setup time is defined as the time between the initial access • group communications, • average response time in fractions of a second for both individual and group calls, • seamless radio coverage throughout the whole served area, • uncomposed and well-recognized high voice quality, • secure and authenticated connections. 97 the journal of engineering research vol. 4, no.1 (2007) 95-102 request (u-setup message) sent by ms a and the allocation of traffic channel resources (d-connect and dconnect acknowledge messages). however, this call procedure is used for individual (point-to-point) calls. the procedure is different in group (multipoint) call communications where no acknowledgement message is required from ms b. thus, the setup this time can be defined as the time elapsed between the sending of the (u-setup message) and the transmission of the (d-setup message). in both procedures, if the system is loaded, mss may try several attempts to gain access to the network. this will increase the delay such that the actual delay will equal the sum of the call setup time defined above and the waiting time for the ms to gain access to the traffic channel. 3.1.2 traffic throughput and traffic load performance criterias another important parameter to be considered in the analysis is the traffic throughput. it is defined for control channel during random access as the average number of successful call setups per frame; it is fully simulated in this analysis. the throughput is affected by the traffic load in the network. the traffic load is defined as the number of rejected calls per tdma frame until 95% of the total call requests are succeeded. in other words, the offered traffic is an indication of the number of mss requesting access to the network resources. as the number of mss access the network increases, the traffic load is increased. thus, it is expected that call setup time to be enlarged by the increased traffic load. this will become more apparent in section 3.2. the aim of the simulation study carried out in this section is to analyze the performance of tetra network for call setup time delay and the throughput for both individual and group communications. the simulation of tetra system protocols however, is performed with the following constraints in place: 3.2 analysis and results of tetra performance a computer simulation is performed based on the simulation parameters above for both individual (point to point) and group (multipoint) calls. figure 2 shows the complementary distribution function (cdf) of percent call success against the setup time for point-to-point calls under different traffic load conditions. the statistical nature of fig. 2 shows that to reach 95% of call success, an average of 35 tdma frames (2 seconds) are required for the 0.1 traffic condition (13 success calls). this value increases to reach 140 tdma frames (8 seconds) for 0.5 offered traffic (126 success calls). this yields that faster success call rate builds-up when the network exhibits less traffic load. figure 1. frame timing of call setup time for tetra system for point-point calls air interface air interface ms a base station ms b u-setup d-setup d-cellprocessing u-connect d-connect d-connect acknowledge • tetra operates under (v+d) mode of operation • 2500 mss are uniformly distributed throughout the cell • control channel activity is captured over 10,000 frames • one tdma frame duration = 56.67 ms • call generation is modeled as poison process • given the distance between a mobile and the base station, the path loss is calculated as the sum of the average path loss (hata model) and a 1-path rayleigh fading channel • the transmission of each message is affected by noise and interference from other mobiles in contention. the statistical behaviour of this interference is assumed similar to additive white gaussian noise • power control has not been applied • control channel capacity is reserved to support call set-up signaling • it is assumed that sufficient traffic channels are available to support the offered traffic • aloha access protocol is applied for uplink and downlink access with a back off time of 3 frames, an access window size of 6, and the maximum number of transmission attempts (if failed) is 5 both uplink and downlink access. setup time (no. of frame) c m m ul at iv e p er ce nt c al l s uc ce ss figure 2. set-up time for point-to-point calls 98 the journal of engineering research vol. 4, no.1 (2007) 95-102 it is clear that, in the case of group (multipoint) calls, however, the situation is better. the setup call procedure, discussed in section 3.1, for multipoint communications drops the number of messages required to establish a call and this in turn cuts down on loading the control channels hence leaving more capacity for initial random access. therefore, it is interesting for this research to compare the performance between the point-to-point and the multipoint call procedures by considering fig. 3. the results presented in fig. 3 shows that, for the multipoint call procedure setup time measurements, to reach only 95% cumulative, only less than 2 frames are required for the 0.1 traffic (9 success calls) and around 15 frames for 0.5 traffic load (250 success calls). that is there is a drop in setup time for the multipoint call procedure by a large percentage reaching 96% for the same traffic load. accordingly, much faster call setup times and enhanced tdma frames utilization are achieved leaving the door open for more mss to be served more willingly than the situation of point-point calls. for this purpose, the multipoint call procedure is more suitable for group (multipoint) communications applications supported by tetra. the effect of the offered traffic load on the throughput is graphically depicted in fig. 4 for point-to-point and multipoint calls. in the case of point-to-point calls, it is evident that the throughput increases from 1 to 15 for increasing traffic condition from 0.1 to 0.5. for the same traffic load, then the throughput increases from 10 to 47 in the case of multipoint calls. as a result, traffic throughput is clearly enhanced in the situation of multipoint calls. nevertheless, the throughput performance is degraded for an increasing traffic load for point-to-point and multipoint call procedures. a summary of the above is shown in table 1. it is obvious from table 1 that the multipoint call procedure decreases the setup time required to reach 95% success call rate. taking the 0.5 traffic load condition as a case of study, for the point-to-point call procedure, tetra network rejects 126 mss from 2506 mss within 140 tdma frames giving a throughput of 17 success call per frame. on the other hand, the multipoint call procedure rejects only 14 mss from 264 mss requesting access within the setup time of 5.3 tdma frames, giving a throughput of 47 success calls per tdma frame. thus, for such condition of heavy traffic load, the throughput is enhanced by 176%. 4. performance evaluation of gprs-basedgsm 4.1 performance criteria and simulation parameters gprs is a service applied over gsm cellular mobile communications which has been established to transmit data in a rate of up to 171.2 kbit/s. the simulation provides analysis for the gsm network based on gprs service having two types of data applications; the e-mail and the www. the traffic in this case deals with two cases: one with pure www and e-mail, and other with mixed www and e-mail applications. www is based on hypertext transfer protocol (http) layer which manages the transfer of the hyper text markup language (html) documents such as web pages (shearer, e.h.s., 1997). each webpage contains a number of objects with certain object size. e-mail depends mainly on the simple mail transfer protocol (smtp) and the post office protocol version 3 (pop3). http, smtp and pop3 protocols are located in the application layer but the www and email traffic are carried through tcp\ip protocols (bruce, a. mah, 1997 and loshin, p., 2000). gprs network performance can be measured using the following three main performance criteria's: setup time (no. of frame) c um m ul at iv e pe rc en t c al l s uc ce ss figure 3. set-up time for multipoint calls • mean application response time : defined as the difference between the time when a user is requesting a web page, or e -mail and the time when it is completely received . • downlink ip throughput per user : is the number of ip bits transmitted in each tdma frame. • packet data channel (pdch) utilization : is the number of mac blocks utilized for mac data and control data and control blocks normalized to the sum of data, control and idle blocks. offered traffic t hr ou gh pu t figure 4. throughput for point-to-point and multipoint calls 99 the journal of engineering research vol. 4, no.1 (2007) 95-102 4.1.1 gprs simulation parameters a computer based simulation is implemented with parameters stated as below: the traffic load parameters of www and email are described in table 2. 4.1.2 gprs simulation environment the core of the simulator used in this paper is described in fig. 5. the environment of the simulator is designed to comprise a ms model, a signal generator and a base station (bs) receives ms requests through the air interface (um), a serving gprs support node (sgsn). all the components are integrated together to feed the statistical evaluation module. 4.2 analysis and results of gprs-based gsm performance the performance of the gsm network under gprs service is analyzed using computer simulation for both mixed and pure www and e-mail traffic. figure 6 shows the results for the mean application response time for the case of pure www\e-mail traffic. the response time increases from approximately two seconds at low traffic load to 20 seconds for the situation of heavy load. it is clear that www exhibits larger time period in compare with that of the e-mail time; as it contains more data. traffic load throughput setup time to reach 95% call success [no. of frames] no. of succeeded mss = no. of success calls no. of rejected mss = no. of rejected calls total no. of mss requesting for access = total number of calls 0.1 1 35 35 2 37 0.2 4 60 240 13 253 0.3 6 88 528 28 556 0.4 13 118.3 1538 81 1619 point-topoint 0.5 17 140 2380 126 2506 0.1 10 0.88 9 1 10 0.2 20 1.77 36 2 38 0.3 30 2.65 80 5 85 0.4 40 3.53 142 8 150 multipoint 0.5 47 5.3 250 14 264 table 1. relationship between the traffic load, number of mss and setup time for point-to-point and multipoint call procedures www parameter distribution mean variance pages per session geometric 5.0 20.0 intervals between pages [s] negative exponential 12.0 144.0 objects per page geometric 2.5 3.75 object size [byte] log2-erlang-k (k = 17) 3700 1.36 × 106 e-mail parameter distribution mean variance e-mail size (lower 80 %) [byte] log2-normal 1700 5.5 × 106 e-mail size (upper 20 %) [byte] log2-normal 15700 62.9 × 109 base quota [byte] constant 300 0 table 2. www and e-mail traffic parameters • cluster size (7). • cell radius = 3000 m. • velocity of mss = 6 km/h. • acknowledged e-mail and www applications • multi-slot capability with one time slot for uplink and four time slots for downlink. • media access control (mac) protocol is applied. • the e-mail size is (1700 – 15700) byte. • the webpage size is (9250) byte. • 4 packet data channels (pdch) are considered. • 30% of the traffic is assumed to be for e-mails and the rest is assumed to be for the www in the case of mixed traffic analysis. 100 the journal of engineering research vol. 4, no.1 (2007) 95-102 figure 7 shows the application response time for the mixed www\e-mail traffic. as in the case of pure traffic, the response time does increase as the number of active mss increases, and in this case the time increases from 2 to 7 seconds compared with the case of pure traffic. as a result, the mixed traffic condition greatly speeds up the response time by 80% for heavy load conditions. on the other hand, for the case of pure traffic, the reason for the large increase in the response time is apparent from the pdch utilization measurements shown in fig. 8 where in, approximately 100% utilization is reached when 15 mss are competing to use the available pdch channels. in fig. 9 however, the pdch utilization seems to have linear relationship with the number of mss. therefore, instead of 100% utilization attained for pure traffic to serve 15 mss, the mixed traffic decreases the pdch utilization down to 70% for 20 mss. thus, the mixed traffic condition saves 30% of the available pdchs and enables more mss to access channel resources, thus resulting in better spectrum utilization. figure 10 shows graphically the measurements of the mean downlink ip throughput per user during transmission periods. it would be appropriate to discuss fig. 10 and fig. 8 of pdch utilization together. the situation of 2 mss illustrates that the network can offer high data rate (around 22kbit\s) while utilizing 20% of the available pdchs. as expected, the situation becomes poorer once figure 5. grps simulation environment www e-mail number of ms 0 5 10 15 20 30 25 20 15 10 5 0 m ea n a pp lic at io n r es po ns e ti m e (s ) figure 6. mean application response time for pure www\e-mail traffic 0 5 10 15 20 number of ms 30 25 20 15 10 5 0 m ea n a pp lic at io n r es po ns e ti m e (s ) figure 7. mean application response time for mixed www\e-mail traffic www e-mail 101 the journal of engineering research vol. 4, no.1 (2007) 95-102 the network is loaded by 15 mss, in which the data rate is truncated down to be less than 5kbit/s and the pdchs utilization is in excess of 80%. the throughput measurements seen in fig. 11, indicates that its rate of decrease is slower than the case of pure traffic analysis shown in fig. 10. therefore, it is evident that the throughput decreases as the number of mss increases, but this time from approximately 20kbit/s down to 10kbit/s rather than 5kbit/s for pure traffic. though, at heavy traffic load, the mixed traffic condition increases the downlink throughput data rate by approximately 100%. furthermore, and for the case mixed traffic condition, the enhanced performance in the response time and throughput are due to the fact that better pdch channel utilization is applicable for mixed traffic situation. this is obvious from fig. 8 in comparison with fig. 10. finally, comparing e-mail application with www application yields that the e-mails exhibits shorter application response time than the www. this comes from the fact that www web pages contain more data than the emails. this shows that www application has poorer pdch utilization and throughput performance. 5. relative performance between tetra and gsm\gprs an assessment for the study of tetra and gsm/gprs networks is provided in this section. also, simulation models for tetra and gprs networks under defined conditions have been implemented. the relative performance criteria measurements for both network types are compared together. the setup time delay for the tetra (v+d) network is compared to the www\e-mail response time for the gprs network. the relationship between the time delay measurements of both the throughput and the number of mss is also discussed and fully analyzed. 5.1 setup time in tetra versus application response time in gprs a summary of the setup time delay in tetra and the 0 5 10 15 100 80 60 40 20 0 number of ms d ow nl in k pd c h % u til iz at io n figure 8. mean downlink pdch utilization for pure www\e-mail traffic www e-mail www e-mail 0 5 10 15 20 number of ms 100 80 60 40 20 0 figure 9. mean downlink pdch utilization for mixed www\e-mail traffic 0 5 10 15 20 number of ms 25 20 15 10 5 0 m ea n do w nl in k ip th ro ug hp ut p er u se r [k b it/ s] www e-mail m ea n do w nl in k ip th ro ug hp ut p er u se r [k b it/ s] figure 10. mean downlink ip throughput for user per pure www\e-mail traffic 0 5 10 15 20 www e-mail number of ms 25 20 15 10 5 0m ea n do w nl in k ip th ro ug hp ut p er u se r [k bi t/s ] figure 11. mean downlink ip throughput for mixed www\e-mail traffic 102 the journal of engineering research vol. 4, no.1 (2007) 95-102 application response time in gprs is provided in table 3. time delay measurements shown are chosen for different traffic load conditions. the e-mail traffic load condition is chosen as it consumes less response time than the www application. as a sample of measurements, to serve around 250 mss, tetra needs around (3.4 s) and (300 ms) for point-to-point and multipoint call procedures respectively. on the other hand, the gprs network serves 18 mss during (30 s) and (5 s) for pure and mixed traffic respectively. it is apparent therefore that tetra, in comparison to gprs, serves much larger number of mss for the same time duration and this results in shorter delay time. the assessment of the two networks can not be completed until the throughput evaluation is revised with its relationship to the traffic load, time delay measurements as well as channel utilization. 5.1.1 throughput versus the number of mss in tetra and gprs the definition of throughput performance criteria differs between tetra and gprs networks. in tetra it is defined for control channel during random access as the average number of successful call setups per tdma frame. in gprs however, it is defined for the downlink ip packets; as the number of bits transmitted per tdma frame per user. in general, the throughput is an indication of the data rate of transmission or the speed of access and it is related to the number of mss (traffic load) in the network. to cope with the definition of the traffic throughput, defined in section 3.1 for tetra network; as the average number of mss per total application response time is calculated in the gprs network. procedures in tetra support higher capacity for the offered traffic load. this is a manifest that tetra strongly supports high capacity traffic and group calls suitable for pss requirements. the results of this comparison are shown in figure 12 for tetra network and in fig. 13 for gprs network. in fig. 12, note that the throughput definition for tetra network is measured per one second rather than per tdma frame. as a summary, fig. 11 shows the measurements of the number of success calls per second, whereas fig. 13 measures the number of completed applications per one second. therefore, it is clear that multipoint call procedure enhances the capacity for the random offered traffic compared to the point-to-point call procedure. in comparison to gprs, it is evident that both call procedures in tetra support higher capacity for the offered traffic load. this is a manifest that tetra strongly supports high capacity traffic and group calls suitable for pss requirements. 6. conclusions it is obvious from the simulation analysis that the gsm\gprs network performance degrades as the number of mss is increased with limited number of pdch channels. the mixed traffic condition shows superior performance relative to pure www and e-mail applications in the response time, ip throughput, and pdch utilization performance criterias. thence, under heavy loaded network situations and mixed traffic condition the gsm\gprs network saves approximately 30% of the available pdchs, increases the data rate by around 100% and deceases the response time by approximately 80%. the comparison between the results obtained from the gsm\gprs network and the tetra network states that tetra performance is better in both: time delay and the number of mobile stations in the network. this leads to assert that tetra strongly supports multimedia applications compared to gsm\gprs network for the same time delay conditions. in accordance with the lower number of control messages needed in the call procedure, it is found that tetra call setup time for voice and data traffic in the case of group communications procedure is faster than in the case of individual calls. furthermore, the throughput performance criterion for group call procedure enhances the network performance by 176% for the case of heavy traffic load of 0.5, thence leave more traffic channels idle for random access. comparing the throughput measurements for both network types shows that, network performance degrades as traffic load in the network increases. references bettstetter, c. vögel, h. and eberspacher, j., 1999, "gsm phase 2+ general packet radio service gprs: architecture, protocols, and air interface," ieee communications surveys, http://www.comsoc.org/pubs/surveys, vol. 2(3). bruce, a. mah., 1997, "an empirical model of http network traffic," infocom, pp. 592, infocom. etsi, 2005, "terrestrial trunked radio (tetra) voice plus data (v+d). part1, general network design; final draft," en 300.3921 vol..3.0. heikonen, k. and saaristo, t., 2002, "you and your tetra radio, tetra as a tool for public safety," it press edita publishing ltd. lammerts, e. slump, c. h. et a., 1999, "realization of a mobile data application in tetra," stw, 10 1901:038. loshin, p., 2000, "essential email standards; rfcs and protocols made practical," john wiley & sons. rahnema, m., 1993, "overview of the gsm system and protocol architecture," communication magazine, ieee vol. 31(4), pp. 92-100. issn 0163-6804. shearer, e.h.s., 1995, "tetra a platform of multimedia," philips telecom pmr cambridge. "tetra touch, 2002, nokia tetra customer newsletter" www.nokia.com/tetra, vol. 2. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 17, no. 2, (2020) 126-134 doi:10.24200/tjer.vol17iss2pp126-134 pedestrian midblock crossing safety development modeling in dohuk city road network ayman a. abdulmawjoud *, and abdulkhalik a. al-taei civil engineering department, college of engineering, university of mosul, mosul-iraq abstract: pedestrians are vulnerable users walking on foot, having no or little protection-usually provided by police officers-when trying to cross from a midblock section in urban areas. a lot of conflicts and sometimes real pedestrian crashes probably took place in some locations in dohuk city midblock crosswalks. for this purpose, and to have a comprehensive idea about the safety situation in this city, and what are the main safety countermeasures needed at some hazardous pedestrian crossing locations, this study was proposed. data were collected from twenty crosswalk locations in central business district (cbd), and suburban areas related to pedestrian movement using two video camera photography to collect data. data were compiled and presented to measure pedestrian speeds, space, unit flow, crossroad width, etc., meanwhile, vehicular traffic data related to crossing volume, speed…etc. were presented too. safety and security data related to conflicts of pedestrians with cross vehicles and some actual crash data were obtained from the qualified offices controlling traffic in the city. the exponential model was found as best-fitting conflict rates and pedestrian spaces with high r2 in both urban and suburban areas. meanwhile, the quadratic model was found to best fitting conflict rates and pedestrian speeds with r2 values of 0.764, and 0.818 in both urban and suburban areas respectively. conflict rates are found as best correlating the real crash frequency with a high r2 value of 0.924 in suburban areas. conflict rates in both cbd and suburban areas are going to be decreased as the pedestrian provided space, speed, crossed road width, and 85th percentile vehicular speed, but increased with the amount of waiting time experienced by pedestrians at crosswalk edge with r2 values of 0.84, and 0.90 for cbd and suburban crosswalks respectively. keywords: conflicts; crash; midblock crossing; pedestrian; sidewalks. في مدینة دھوكنمذجة تطویر السالمة لعبور السابلة على شبكة الطرق عبدالخالق مال هللا الطائيو ،ایمن عبدالھادي عبدالموجود في الطریق على األقدام ، وال یتمتعون بأي سیرونمحصنین والذین یالغیر ي الطریق : السابلة او المشاة ھم مستخدمالملخص التي یتم توفیرھا من الى الحمایةاغلبھا حمایة عند محاولة العبور على ضفتي الطریق في المناطق الحضریة للمدینة حیث تفتقر ابلة في مدینة قبل مسؤلي المرور في بعض االحیان. یحدث الكثیر من التعارض وأحیاناً حوادث حقیقیة في بعض مواقع عبور الس دھوك. وألخذ مفھوم كامل حول وضع السالمة في ھذه المدینة وماھي التدابیر الالزمة في بعض مواقع عبور المشاة الخطرة, تم صویر بكامیرتي تالباستخدام ھاضواحیوموقع لعبور السابلة من مناطق في مركز المدینة 20جمع أنواع مختلفة من البیانات من ، وفي الوقت نفسھ تم جمع بیانات تخص المركبات كالسرعة، والفضاء، وتدفق الجریان، ... إلخبیانات السابلة ع فیدیو. تم تجمی ... إلخ. تم الحصول على بیانات السالمة المروریة المتعلقة بالحوادث الفعلیة من المكاتب المؤھلة تھاوسرع ھامنھا حركة مرور تم تثبیت بعض البیانات األخرى المتعلقة بعرض الشوارع والالزمة لفھم أسباب التي تتحكم في حركة المرور في المدینة. جمیع ھذه البیانات من خالل إجراء تحلیل احصائي لربط المتغیرات المختلفة حللت في سالمة السابلة في مدینة دھوك ثم القصور م الحصول علیھا تعتبر التعارض فعال في وصف للحصول على نماذج تجریبیة تصف الموقف في المواقع المحددة. النتائج التي ت الحوادث وكذلك معدالت التعارض تعتمد على سرعة السابلة وعرض موقع العبور. .المماشي ؛سابلة ؛معبر الوسط ؛اصطدام ؛تعارضال الكلمات المفتاحیة: *corresponding author’s e-mail: aymanmawjoud@uomosul.edu.iq mailto:aymanmawjoud@uomosul.edu.iq mailto:aymanmawjoud@uomosul.edu.iq pedestrian midblock crossing safety development modeling in dohuk city road network 127 1. introduction traffic safety is known as a critical social and a huge economic matter in kurdistan region-iraq. the rapid economic growth has produced increased motorization, particularly in urban areas. in parallel, the improvements in the infrastructure have enhanced high traffic mobility and speed. as a result of a lack of comprehensive pedestrian traffic safety countermeasures by the government and people’s low sensibility of traffic safety, the proportion of traffic crashes and fatalities had constantly increased over the last five years. these impacts involve tragic loss in effort, time, money, and human lives. moreover, they create growing feelings of a great lack of safety in all facilities of human lives. nchrp, 2008, defined traffic conflicts as an event involving two or more road users, in which one user performs some typical or unusual activity such as a direction or speed that places other users in jeopardy or crash unless an evasive manoeuvre is undertaken. hemasiri, and edirisinghe, 2010 conducted a study in sri-lanka at pedestrian crossings as most of the drivers in the country don't follow the rule of ''give way to pedestrians''. three main types of field data were collected: video survey, observer measurements, and drivers' interview. six different zebra crossings located within the limits of the kandy-peradeniya-gelioya road segment were selected for field surveys. the behaviour of drivers with pedestrians was summarized in this study to use as a guideline to improve pedestrian safety. hussein and neham, 2008, investigated three "t", and, four "cross" legs signalized intersections in iraqi capital, baghdad, to study the traffic conflicts as a surrogate to traffic crashes. they considered shapes of conflicts mostly took place at signalized intersections. conflicts were correlated to approach spot speed, stopped vehicles near signals control, delay time, and some other traffic geometric parameters of the intersection approach such as lane width, median width, and length of the auxiliary lanes constructed near left-turn movement. the study concluded that conflicts are mostly correlated linearly to stopped delay time with r2=0.978, meanwhile, an exponential model described the relationship between conflict rate and traffic volume with r2 from 0.77 to 0.94 acceptable on 90% significance level. multiple regression analysis showed that conflict rates increased with increasing approach spot speed, but reduced with increasing lane width, median width, and availability of auxiliary lanes. diogenes and lindau, 2010, evaluated the pedestrian safety and analogy of midblock crossing with the highest number of pedestrian crashes in porto alegre city, brazil. twenty-one crosswalks were chosen for evaluation. results show that pedestrian crash hazard was affected by a combination of interacting hazard factors, such as the presence of bus stops, number of lanes, road width, and the volume of vehicles and pedestrians. zhang et al., 2019, investigated the behaviour and safety of pedestrian crossing at midblock crosswalks. for this goal, twelve midblock crosswalks located in wuhan, china were chosen to gather data through field investigation. descriptive statistics were applied to resolve pedestrian crossing attitudes and the allocation of pedestrian-vehicle conflicts. three models for pedestrian-vehicle conflicts analysis were found to measure the influence of various factors on pedestrian safety. also, it was pointed out that pedestrian behavioural properties such as rolling gap mode, crossing with others would obviously increase the potential of pedestrian-vehicle conflicts. 1.1 pedestrian safety problem pedestrian safety problem in dohuk city is merely related to the sprawl in both population, and vehicles, especially in the last five years. terrorist attacks to wide lands around the kurdistan region in mosul, anbar, and kirkuk cities caused the migration of a lot of people to the region congesting both roads, and crosswalks due to the unusual movement of both people and vehicles in dohuk city road network. figure 1 is showing the trend of car ownership in dohuk city between the period 2002 up to 2014. the average number of growths in car ownership was increased by 12 per cent between 2010 and 2014 periods (accident data files, 2015). vehiclepedestrian conflicts increased as it seems clear from the crash statistics compiled by the directorate of traffic police in dohuk city crash files. 1.2 study purpose to reveal some of the tragedy related to the safety of the pedestrians, or the unprotected users walking on foot, the following goals are planned: • to prepare a database related to pedestrian safety in dohuk city for future research or development. • to collect crash data related to pedestrians and where mostly the problem is probably going to accumulate. • nature of traffic conflicts produced between vehicles and crossing people, especially at peak service hours; • to correlate pedestrian crashes with conflict rates on midblock pedestrian crossing zones; • to correlate conflict rates with other pedestrian crossing parameters, and some road geometric parameters. 2. research methodology to execute the main goals of this study, a certain area was selected including all the urban, and suburban areas ayman a. abdulmawjoud and abdulkhalik a. al-taei 128 figure 1. number of vehicles owned by the people in duhok city during the period (2002-2014), (accident data files, 2015). of dohuk city. figure 2 shows the selected areas, which is a type of gis map taken from the city (golden guide, 2010). figure 3 shows the flow chart of the proposed methodology, which contains the main types of data collected. traffic conflicts is an event that can be observed showing the interaction of vehicle against pedestrians while crossing the road. conflicts used by glauz and migletz, 1980, as surrogates for the prediction of crashes at some hazardous locations are selected to study pedestrian crossing behaviour. using the conflict data collection form used by glauz and migletz, 1980, thousands of conflict data were observed between vehicles and pedestrians in both cbd and suburban crossing locations selected for analysis in this study. in this form, conflicts are classified into major conflicts (m.c.), and secondary conflicts (s.c.), major conflicts are those related to turning or through movement contacts with pedestrians represented by some dangerous action or longhorn. meanwhile, minor conflicts are represented by brake light during pedestrian crossing as a parameter to draw out some conclusions. conflict rates per 1000 vehicles were computed for this correlation using the following equation given by glauz and migletz, 1980: conflict rate/1000 vehicle = (nco/tvo) × 10 (1) where: nco: number of conflicts observed during three hours of study; and tvo: total number of vehicles counted during the same period of time in thousands. conflict rates during the period of observing pedestrians were recorded in order to take an opinion about the critical situations. pedestrians were usually subjected to driver frustrations when passing the road in the existence of mechanized traffic. crash data was compiled from crash data files from the directorate of traffic control in dohuk city, 2015, through the period (2009-2014) which is the up-to-date provided crash data to represent the safety behaviour of pedestrians crossing from different midblock locations. crash data were provided only for suburban area pedestrian crossing locations meanwhile; cbd area crashes were not registered. location selected on both areas within dohuk city in this study is midblock segments not near to any congested signalized or unsignalised intersection, and those candidates as dangerous by police officers. other data related to pedestrian flow parameters such as unit flows, speeds, spaces, etc., were collected from video filming and manually transferred into computer files. the huge volume of the compiled data was analyzed using spss program version 24, 2017, and microsoft excel package. figure 4 shows two locations one of them in cbd and the other in suburban dohuk city area where data cameras had been fitted on illumination poles or at some vantage points for photography. 3. analysis of results and discussions in order to start the stage of results analysis and discussions, the twenty locations were investigated, ten of them are in the cbd area and the other ten is in suburban areas. to start the data analysis in this study, peak hour demands between (8:30-10:30 a.m.), and (5:30-7:30 p.m.) were selected as peak morning and evening periods for both vehicles and pedestrians and limited exactly in order to concentrate data collection and analysis. pedestrian midblock crossing safety development modeling in dohuk city road network 129 figure 3. research methodology workflow diagram. (a) (b) figure 4. two photos showing two of the pedestrian crosswalks in both cbd and suburban areas: (a) panorama centre commercial. cbd pedestrian crosswalk. (b) shelan hospital (qazi mohammad st.). commercial suburban pedestrian crosswalk. data collection from 20 selected midblock pedestrian crosswalk selected crossing locations in suburban area pedestrian flow variables, volumes, speeds, delay time, space, crossing time, geometrics vehicle-pedestrian conflict rates from observations using special data form statistical analysis and model crash frequency from police data files conclusions and recommendations selected crossing locations in cbd area location of pedestrian observations (e/37.3°n/43.9°, e/37.5°n/43.15°) figure 2. dohuk city study area with 20 data collection locations, golden guide, 2010. ayman a. abdulmawjoud and abdulkhalik a. al-taei 130 figure 5. effect of pedestrian space on the conflict rate in midblock crosswalks in dohuk city cbd and suburban areas. 3.1 pedestrian conflict and pedestrian space at cbd and suburban crossing locations pedestrian conflicts (i.e., both major and secondary conflicts) were computed in rates per 1000 vehicles from eqn. (1) and correlated with pedestrian provided space in m2/pedestrian (m2/ped.) which is the number of pedestrians occupying an area of one square meter at the crosswalk during the crossing process stated by the hcm, 2010. the space variable was counted automatically from an area of one square meter using photoshop and excel sheet limited with photo zoom picture by picture. figure (5) shows the variation of conflict rates with space. a relationship is described by exponential models with a good correlation of determination coefficients r2 of 0.906 and 0.866 for both cbd and suburban locations, respectively. equations (2) and (3) reflect the fact that as pedestrian move closer, conflicts with the adjacent vehicles will increase. the form of the models derived are as follows: y1 = 683.35e−0.07x (2) y2 = 745.55e−0.042x (3) where: y1: conflict rate/1000 vehicles in dohuk city cbd crosswalks; y2: conflict rate/1000 vehicles in dohuk city suburban crosswalks; and x: pedestrian space in m2/ped. correlations between conflict rate and pedestrian space were implemented using the above statistical analysis by both simple and multiple regression to show the effect of the independent on the dependent variable and what shapes of empirical models are correlating them. 3.2 pedestrian conflict and pedestrian speed at cbd and suburban crossing locations conflict rates and pedestrian speed variations on the crosswalks located in both cbd and suburban areas are shown in fig. 6. coefficients of correlation r2 are 0.764 and 0.818 for cbd and suburban crosswalks respectively. the best models are 2nd degree polynomial models with the following mathematical forms: y1 = 770.9x2 + 2332x − 1259 (4) y2 = −4044.7x2 + 12492x − 9136.2 (5) where: y1: conflict rate/1000 vehicles for cbd crosswalks; y2: conflict rate/1000 vehicles for suburban crosswalks; and x: pedestrian speed, m/s. figure 6 shows that at an optimum conflict rate of about 500 conflicts/1000 vehicles, the pedestrian speed is around 1.55 m/sec. both trends are going to increase with the increase of pedestrian speed up to the abovementioned number of conflict rate and then goes down to reduce. the conflict rate decreases for suburban after a certain speed, since the number of pedestrians in the suburban is low with higher walking speed. figure 6. effect of conflict rate on pedestrian speed on midblock crossing sections in dohuk city cbd and suburban areas. figure 7. pedestrian crash frequency and conflict rates correlation on midblock crossings located in dohuk city suburban area. pedestrian midblock crossing safety development modeling in dohuk city road network 131 3.3 pedestrian traffic crash frequency and conflict rate correlations in order to detect how much traffic conflicts, represent the chance of traffic crashes to occur on pedestrian crosswalks, crash data and conflict rates are correlated in this study to know the empirical relationship combining them. the analysis was implemented for suburban locations only using a step-wise form of the spss program; as the cbd investigated locations did not have enough crash data to perform the analysis required. crash frequency and conflict rates data in a suburban area was plotted to find if there is any correlation with the conflict rates at suburban crosswalks. figure 7 shows the predicted model obtained, which indicates that crashes increase as conflict rate increases on suburban crosswalks, up to a peak value of 43 crash at 365 conflicts rate/1000 vehicles, then decrease as conflict rates increase as drivers are going to be more sensitive and awake to drive obeying traffic rules and ordinances. this correlation is describing an empirical crash field data registered by police officers and conflicts from photography filming with manual data counting. in this crosswalk a special zebra type is needed, in addition to an automatic push-button pedestrian crossing signal installation will be very useful to survive people's life, especially school students and elderly people. some of the cross midblock locations investigated in this study are more than 10 meters wide, and some pedestrians are not so fast to cross, in this case, an overpass is required as a safe structure to provide more safety and security to some weak type pedestrians. the mathematical model is a 3rd degree polynomial with r2 of 0.924 as follows: 𝑌𝑌 = −1𝐸𝐸 − 06𝑋𝑋3 − 0.0003𝑋𝑋2 + 0.7346𝑋𝑋 − 117.6 (6) where: y: crash frequency of pedestrians on suburban crosswalks, and x: conflict rate of pedestrians on suburban crosswalks in conflicts/1000 vehicles. this result is supporting many other results found elsewhere, that crashes could be represented or replaced by conflict rates in other safety research or safety supporting projects such as those implemented by khalik, 1997. 3.4 analysis of conflict rates with pedestrians and other traffic and geometric characteristics of the crossing locations pedestrians are usually coming to crosswalks to pass safely as soon as possible with minimum delay time and without interaction with the passing vehicles. this condition is highly affecting the behaviour of the pedestrian, which may lose his/her patience to cross without care. in this study, conflict rates are correlated with five main factors to know how they affect pedestrian safety in both cbd and suburban areas. the main factors selected for step-wise correlation analysis with pedestrian conflict rates (as a dependent variable) are provided pedestrian spaces, the pedestrian speed at a crossing, road width which is the total approach width from curb to curb distance without considering median width as a geometric variable, 85th percentile vehicular traffic speed, and pedestrian waiting time. the analysis is performed for both cbd and suburban crosswalk locations included in this study. the analysis was performed using the spss program. the model variable introduced are described as below: y1 conflict rate in conflicts/1000 vehicles on cbd crosswalks; y2 conflict rate in conflicts/1000 vehicles on suburban crosswalks; x1 pedestrian space measured in m2/ped; x2 pedestrian speed at crosswalks in m/sec; x3 total width of a crossed road in (m); x4 85th percentile vehicular traffic speed on main crossed road in km/h; and x5 pedestrian waiting time on road curbstone in sec/ped. 3.4.1 pedestrian conflict rate modelling in dohuk city cbd area crosswalk step-wise regression analysis correlating conflict rates on cbd crosswalk, and other variables listed above is shown below: y 1 = 813.958 − 4.871x1 − 22.251x2 − 17.344x3 − 12.318x4 + 45.766x5 (7) from the above model, it is shown that some of the variables are going to decrease conflict rates, meanwhile, only one is going to increase it. pedestrian space, and speed, is going to decrease it, and the same effect seems for road width, and vehicle speed too, but pedestrian waiting time on curbstone on cbd is going to increase them. collinearity table which shows the interaction among the five independent variables, x1, x2, x3, x4, and x5 is shown in table 1. analysis of variance anova is shown in table 2 for the predicted model. the derived model produced multiple correlation coefficient for the total model of r2= 0.84. vif is best showing that collinearity is very weak as its value is less than 5.0, and the p-value of 0.838% is less than 5% required insignificance level (anderson, 2008). in table 2, the f-calculated value is more than the standardized value of f at 5% insignificance level of 3.84, this means that regression coefficients are the best-fitted values as the null hypothesis will be rejected. the partial correlation matrix shows no collinearity among the five independent variables, and proof model validity as it is shown in table 3. figure 8 shows the uniform ayman a. abdulmawjoud and abdulkhalik a. al-taei 132 table 1: collinearity matrix of independent variables x1, x2, x3, x4, and x5 for step-wise regression analysis of conflict rates on dohuk city cbd crosswalks. table 2: analysis of variance of variables included in the regression analysis of y1, x1, x2, x3, x4, and x5 of conflict rates on dohuk city cbd crosswalks. table 3: partial correlation matrix of the step-wise regression of y1, on the other variables x1, x2, x3, x4, and x5 of conflict rates on dohuk city cbd crosswalks. figure 8. expected and observed regression validation of conflict rates around normal probability diagonal line on dohuk city cbd midblock crosswalks. figure 9. expected and observed regression validation of conflict rates around normal probability diagonal line on dohuk city suburban midblock crosswalks. predictor coefficients standard deviation student t-value probability of insignificant variable inflation factor (vif) constant 813.958 976.666 0.833 0.451 x1 -4.871 7.386 -0.659 0.546 2.667 x2 -22.251 405.281 -0.055 0.959 2.153 x3 -17.344 24.569 -0.706 0.519 1.193 x4 -12.318 10.757 -1.145 0.316 2.558 x5 45.766 41.532 1.102 0.332 3.349 source of variation d.f. ss ms f p regression 5 141363.85 28272.77 4.12 0.00838 error 4 27449.28 6862.32 total 9 168813.13 variables y1 x1 x2 x3 x4 x5 y1 1.00 x1 -0.20 1.00 x2 -0.117 -0.416 1.00 x3 -0.110 -0.017 -0.007 1.00 x4 -0.233 -0.307 -0.395 -0.022 1.00 x5 0.241 0.474 -0.685 0.221 0.439 1.00 pedestrian midblock crossing safety development modeling in dohuk city road network 133 table 5: analysis of variance of variables included in the regression analysis of y1, x1, x2, x3, x4, and x5 for conflict rates on dohuk city suburban crosswalks. table 6: partial correlation matrix among y2, and other variables x1, x2, x3, x4, and x5 for conflict rates on dohuk city suburban crosswalks. distribution of sse values around the diagonal of the normal probability plot around the zero value of the sse, which represents the model validation as plotted using the spss programme. 3.4.2 pedestrian conflict rate modelling in dohuk city suburban crosswalk step-wise regression analysis correlating conflict rates on dohuk suburban crosswalk and other same variables listed above is shown in the model represented in eqn. 8 as below: 𝑌𝑌2 = 698.479 − 3.815𝑋𝑋1 − 301.984𝑋𝑋2 − 12.829𝑋𝑋3 − 0.966𝑋𝑋4 + 1.016𝑋𝑋5 (8) from the model, it is shown that some of the variables are going to increase conflict rates, meanwhile, others are going to decrease it. pedestrian space, pedestrian speed, total road width, and 85th percentile vehicle speed is decreasing conflict rates but, pedestrian waiting time on the curbstone of dohuk city suburban area is increasing it. the collinearity table representing the interaction among the five independent variables x1, x2, x3, x4, and x5 is shown in table 4. figure 9 shows the uniform distribution of sse values around the diagonal of the normal probability plot around the zero value of the sse, which represents the model validation criteria as plotted using the spss program. 4. conclusion according to the size of data collected and the techniques used to analyze it and decide what is necessary to solve the pedestrian safety problem in dohuk city in both cbd and suburban crosswalks. the following conclusions could be drawn out: • vehicle conflicts and pedestrian space are related in exponential models with r2=0.906 and r2=0.866 in cbd and suburban crosswalks respectively. • vehicle conflicts and pedestrian speed are related in quadratic polynomial models with r2=0.764 and r2=0.818 in cbd and suburban crosswalks respectively. the optimum value was 500 conflicts /1000 vehicles at pedestrian speeds of 1.55 m/s. • crash and conflict rates correlations in the suburban area show a cubic polynomial model with a high coefficient of multiple determination r2 of 0.924. the function is best fitted with a peak crash frequency at 365 conflict rates per 1000 vehicles at an optimum crash frequency of 43. the model concludes that in suburban area crosswalks conflict rates could be considered as a surrogate to crash frequency in describing the safety condition in dohuk city suburban area. • conflict rate per 1000 vehicles happened in dohuk city cbd and suburban crosswalks decrease with pedestrian provided space, speed, crossed road total approach width, and 85th percentile vehicular speed, but increases with the amount of waiting time experienced by a pedestrian at crosswalk edge with r2 values of 0.84 and 0.90 for cbd and suburban crosswalks respectively. a lot of changes are found necessary to be recommended to develop pedestrian movement safety in dohuk city in both cbd and suburban crosswalks, such as pedestrian crossing automatic signs, zebra, and pelican crossing markings where they are justified in order to reduce or eliminate conflicts, and/or crash occurrence near or at crosswalks and make crossing easier for pedestrians. formal analysis obtained a lot of results that recommend some locations need overpass bridge structures to isolate pedestrians from the rapid hazardous traffic vehicles travelling with high speeds and, this selection is needing more study by decisionmakers to put down the type of each solution needed. source of variation d.f. ss ms f p regression 5 150203.28 30040.656 7.3 0.0048 error 4 16460.632 4115.158 total 9 166663.912 y2 x1 x2 x3 x4 x5 y2 1.00 x1 -0.550 1.00 x2 -0.450 0.295 1.00 x3 -0.182 -0.151 0.589 1.00 x4 -0.233 -0.159 0.282 0.065 1.00 x5 0.013 0.252 0.262 0.522 0.217 1.00 ayman a. abdulmawjoud and abdulkhalik a. al-taei 134 conflict of interest the authors declare that there are no conflicts of interest as regards this article. funding this study did not receive any funding from the public, private or not-for-profit sectors. acknowledgment the authors wish to acknowledge the help provided by the staff of the dohuk governorate traffic police department and, the staff of the traffic engineering unit of dohuk municipality directorate for their assistance in the collection of traffic data. references anderson, dr., sweeney, dj., and williams, ta. 2010. statistics for business and economics, 11th edition, south-western college publication. diogenes, mc., and lindau, la. 2010. “evaluation of pedestrian safety at midblock crossings, porto alegre, brazil”. transportation research record: journal of the transportation research board, no. 2193, transportation research board of the national academies, washington, d.c., 37–43. dohuk city traffic police directorate, 2015. accident data files (1995-2014); dohuk governorate, 2010. guide and information center gic, golden guide. glass, wd., and migletz, dj. 1980. “application of traffic conflict analysis at intersections”. nchrp report 219, transportation research board. hemasiri, ag., and edirisinghe, j. 2010. “safety on pedestrian crossing”. university of peradeniya, peradeniya 20400, sri lanka. highway capacity manual, 2010. transportation research board (trb), national research council, washington d.c. hussein, ae., and neham, ss. 2008. “traffic conflict technique: a tool for traffic safety study at threeleg signalized intersections”. journal of karbala university, vol. 6, no.1 scientific. khalik, am. 1997. “conflict analysis at unsignalized unchannelized t-intersections”. scientific journal of tikrit university, engineering science vol.4 national center highway research program, (nchrp), 1980. “application traffic conflict technique at urban intersection”. transportation research record no.776. spss version 24, 2017. “statistical package for social sciences”. zhang, c., chen, f., and wei, y. 2019. “evaluation of pedestrian crossing behavior and safety at uncontrolled mid-block crosswalks with different numbers of lanes in china”. accident analysis prevention, vol.123, pp. 263-273. 1. introduction conflict of interest references civl080303.qxd the journal of engineering research vol 7, no. 1 (2010) 11-21 1. introduction hot mix asphalt (hma) is the most common material used for paving applications. it consists primarily of an asphalt binder and mineral aggregates. the binder acts as a gluing agent that binds aggregate particles into a cohesive mass. when bound by an asphalt binder, a mineral aggregate acts as a stone framework that provides strength and toughness to the system. several mixture design methods have been developed over time, in an effort to create a mixture that is capable of providing acceptable performance based on certain predefined set of criteria. the most recently developed mixture design method ____________________________________________ * corresponding author e-mail: alshamsi@squ.edu.om is the superpave method. superpave stands for superior performing asphalt pavements and represents a basis for specifying component materials, asphalt mixture design and analysis, and pavement performance prediction. it includes several processes and decision points. a key feature in the superpave system is laboratory compaction. in the mix design procedure, a superpave gyratory compactor (sgc) is used to carry out the compaction of the asphalt mixture specimens in the laboratory. sgc was found to be effective in simulating field compaction and ensuring that the properties of the samples compacted in estimating optimum compaction level for dense-graded hot-mix asphalt mixtures khalid al shamsi*a and louay n. mohammadb *a department of civil and architectural engineering, college of engineering, sultan qaboos university, p.o box 33, postal code 123, al-khoud, muscat, sultanate of oman b department of civil and environmental engineering and louisiana transportation research center, louisiana state university, 4101 gourrier ave, baton rouge, la 70808, usa received 3 march 2008; accepted 18 may 2009 abstract: a critical step in the design of asphalt mixtures is laboratory compaction. laboratory compaction should reflect field compaction and should produce mixtures that are economical and possess high structural stability. during the compaction process, asphalt mixtures are subjected to certain amount of compaction energy in order to achieve the required density. the superpave volumetric mix design is based on compacting hma mixtures to a specified compaction level described by the number of gyrations from the superpave gyratory compactor (sgc). this level is termed ndes and represents the required energy (based on the traffic level expected) to densify the mixture to a 4% air voids level. this paper re-examines the superpave compaction requirements through extensive laboratory investigation of the response of a number of asphalt mixtures to the applied compaction energy. it also presents an alternative method to estimate the number of gyrations at which a mixture first reaches an optimum aggregate interlock and hence prevents overcompaction problems that might result in unstable aggregate structures or dry asphalt mixtures. a total of 12 hma mixtures were studied. during compaction, force measurement was made using the pressure distribution analyzer (pda). the compaction characteristics of the mixtures were analyzed using data from the pda and the traditional superpave gyratory compactor (sgc) results. keywords: locking point, mix design, asphalt mixtures, pavement materials, laboratory compaction á«dé©dg áaéãμdg ägp á«à∏ø°sc’g äé£∏î∏d …èıg ∂e~∏d πãe’g iƒà°ùÿg ~j~– ~ªfi …dƒd , ¢ùeé°ûdg ~dén ::áá°°uuóóÿÿggá«°s~æg ¢üfé°ün ägp á«à∏ø°scg äéæ«y êéàfg ¤g …odƒj ¿cg »¨ñæj …èıg ∂e~dg .á«à∏ø°s’g äé£∏ÿg º«ª°üj ‘ á«°sé°s’g ägƒ£ÿg øe …èıg (§¨° dg) ∂e~dg ~©j …èıg ∂e~∏d πãe’g iƒà°ùÿg ~j~ëàd iôμàñe á≤jôw ω~≤j á°sgq~dg √òg .≥jô£dg ~««°ûj ‘ áe~îà°ùÿg §¨° dg äg~©e ¬k~– …òdg ∂e~dg øe áœéædg á«à∏ø°sc’g ioéª∏d ¬¡hé°ûe qƒ£àe ré¡éh áfé©à°s’g é° jcg h áø∏àfl á«à∏ø°scg á£∏n 12 ¢üfé°ün π«∏– h º«ª°üj â∏ª°t á°sgq~dg .iqƒ£àe ájèfl äé°uƒëa ωg~îà°séh ∂dp h á«dé©dg áaéãμdg ägp á«à∏ø°sc’g äé£∏î∏d .∂e~dg aéækcg áæ«©dg πngo §¨° dg ™jrƒj ¢sé«≤d ::áá««mmééààøøÿÿgg ääggooôôøøÿÿgg á«à∏ø°scg á£∏n ,∂e~dg ,â∏ø°scg ,∞°uôdg 12 the journal of engineering research vol 7, no. 1 (2010) 11-21 the laboratory are to some extent, similar to the mix placed in the field (cominsky, 1994). figure 1 is a schematic illustration of the sgc and fig. 2 presents a typical output from it. a hydraulic or mechanical system applies a load to the loading ram, which applies 600 kpa compaction pressure to the specimen. the loading ram diameter nominally matches the inside diameter of the compaction mold, which is normally 150 mm for design purposes. the sgc base rotates at a constant rate of 30 revolutions (gyrations) per minute during compaction, with the compaction mold positioned at a compaction angle of 1.25º. the ram pressure is monitored by a pressure gauge during compaction. as the specimen densifies, the pressure gauge and loading ram maintain compaction pressure. the specimen height during compaction is monitored and recorded. in the superpave system, asphalt mixtures are designed at a specific level of compactive effort. this is a function of the design number of gyrations (ndesign). the design number of gyrations depends upon the traffic level for which the mix is designed. traffic is represented by the design-equivalent single axle loads (easls). four traffic categories ranging from 0.3 to greater than 30 million easls are used. higher compaction energy is applied to mixtures in the heavy traffic category. the analysis of the compacted samples is done in terms of percentage of the theoretical maximum specific gravity of the mixtures. the data from the sgc are generally used in computing volumetric properties such as density or air void content as a function of compaction gyrations. however, several attempts have been recently made to analyze the densification curve obtained from the sgc in order to evaluate the asphalt mixtures’ workability and resistance to permanent deformation. the initial number of gyrations (ninitial) and the slope of the initial portion of the sgc compaction curve have been hypothesized to reveal certain mixture properties such as tenderness of the mixtures and the strength of aggregate structure (mcgennis, 1995). vavrik (vavrik, et al. 2000) suggested the evaluation of mixture compaction characteristics based upon the locking point or the point during compaction at which the mixture exhibits a marked increase in resistance to further densification. the alabama department of transportation (alabama dot, 2002) have adopted the locking point mix design concept. they define the locking point as the point where the sample being gyrated loses less than 0.1 mm in height between successive gyrations. similarly, the georgia department of transportation (georgia dot, 2003) use the concept of locking point in designing hma mixtures. they define the locking point as the number of gyrations at which, in the first occurrence, the same height has been recorded for the third time. for georgia, typical locking points are reported to be in the range between low 60's to high 80's measured with the superpave gyratory compactor. recently, the oregon department of transportation changed the superpave design gyration levels for mixes used in the state highway system (asphalt, 2007). oregon highway mixes will be designed at 4.0 percent air voids at 65, 80, and 100 gyrations. for highways with low truck volumes (odot level 2), gyrations have been rolled back from 75 to 65. mixes to support moderate truck traffic (odot level 3) will be designed at 80 gyrations instead of 100. interstate and other highways with heavy truck traffic (odot level 4) will be built with mixes designed at 100 gyrations rather than 125 gyrations. guler et al. (guler, et al. 2000) developed a gyratory load cell and plate assembly (glpa) for measuring hma shear resistance during compaction with any sgc. it is a thin cylindrical device that is inserted on top of the mixture in the compaction mold that gives a continuous measure of shear resistance under gyratory loading during compaction. they hypothesized that bulk shear resistance from the glpa is a good indicator of the compactability of hma mixtures and their potential resistance to rutting under traffic. it was concluded that the device offers potential as a low-cost tool to complement volumetric properties from the sgc by evaluating the compactability of asphalt mixtures as well. control and data acquisition panel loading ram mold reaction frame rotating base figure 1. superpave gyratory compactor (asphalt institute, 2001) figure 2. typical data output from the sgc umber of gyrationns h ei gh t of s pe ci m en , m m 13 the journal of engineering research vol 7, no. 1 (2010) 11-21 2. objective and scope the primary objective of this study was to develop a method for estimating the number of gyrations at which an asphalt mixture reaches its optimum aggregate interlock. to achieve this objective, an extensive laboratory investigation was carried out to study the response of a number of asphalt mixtures to the applied compaction energy when compacted in the superpave gyratory compactor (sgc). a total of 12 hma mixtures were evaluated. during compaction, force measurement was made using a small device inserted in the compaction mold called the pressure distribution analyzer (pda). the compaction characteristics of the mixtures were analyzed using data from the pda and the traditional superpave gyratory compactor (sgc) results. the following sections describe in detail the materials used and the methodology applied to achieve the objective of the study. 3. materials an asphalt mixture is a composite material that is largely made of two main components, aggregate and asphalt cement. sources of aggregate were selected to encompass a wide range of aggregates commonly used in asphalt mixtures. three aggregate types were used. these are: different stockpiles from each type of aggregates were acquired. natural coarse sand was used whenever necessary in the final design blends. aggregates were acquired in 50-gallons barrels and kept properly sealed from any moisture intrusion. detailed laboratory evaluation procedures of individual stockpiles were conducted to determine the basic aggregate properties such as specific gravity, gradation, and other superpave consensus properties. larger-sized stockpiles were sieved into individual size fractions. materials retained on 1", 3/4", 1/2", 3/8", no. 4, and passing no. 4 sieves were stored in separate containers so that the required gradations could be batched directly from the individual size fractions. this method of aggregate separation, while somewhat time and laborintensive, allows for strict control and exact replication of mixture's aggregate gradation. three aggregate structures were designed for each aggregate type: coarse (c), medium (m), and fine (f). figure 3 shows an example of the aggregate structures used in the study. sb polymer-modified asphalt binders pg76-22m were used in all the mixtures evaluated in this study. 4. mixture design mixture design was performed on all the aggregate structures that were formulated. the superpave mixture design method was followed except for the vma requirement. all the mixtures were designed for high-volume traffic (ndes = 125 gyrations at 1.25° angle of gyration). the optimum asphalt content was determined as the asphalt content required to achieve 4.0 percent air voids at ndes. optimum asphalt contents ranged from 3.0% to half inch granite 0 10 20 30 40 50 60 70 80 90 100 sieve size, mm pe rc en t p as sin g coarse medium fine 19.0 12.5 9.5 4.75 2.36 1.18 0.60 0.30 0.150 0.075 figure 3. example of aggregate structure used in the study  hard agg regates; crushed granite (gr)  water-absorptive, high -friction aggregate; sandstone (sst) and  low friction low-water-absorption aggregate; limestone aggregate (ls). 14 the journal of engineering research vol 7, no. 1 (2010) 11-21 5.1% . the coarse mixtures had higher optimum asphalt contents for all the aggregate types considered. tables 1 and 2 provide the job mix formula for each of the designed mixtures. 5. mixtures compactability the densification curve obtained from the sgc was used to evaluate mixture resistance to the compaction energy applied by the sgc. the behavior of the mixtures during compaction was also captured using the pressure distribution analyzer (pda). this is a simple accessory that measures the force applied to the mixtures using three load-cells equally spaced at an angle of 120°. the loadcells allow measuring of the variation of forces during gyration such that the position or eccentricity of the resultant force from the gyratory compactor can be determined in real time. the two-dimensional distributions of the eccentricity of the resultant force can be used to calculate the effective moment required to overcome the internal shear frictional resistance of mixtures when tilting the mold to conform to the 1.25º angle. based on the data from the load-cells, the two components of the eccentricity of the total load relative to the center of the plate (ex and ey) can be calculated. the calculations are simply done with general moment equilibrium equations along two perpendicular axes passing through the center of one of the load-cells as shown in fig. 4 using the following equations: (1) mixture name1 ls coarse ls medium ls fine sst coarse sst medium sst fine gr coarse gr medium gr fine mix type 12.5 mm 12.5 mm 12.5 mm 12.5 mm 12.5 mm 12.5 mm 12.5 mm 12.5 mm 12.5 mm binder type pg 76-22m pg 76-22m pg 7622m pg 76-22m pg 76-22m pg 76-22m pg 76-22m pg 76-22m pg 76-22m design ac content, volumetric properties, and densification % gmm at ni 85.1 86.2 88.0 86.0 86.4 88.0 87.3 87.3 87.1 % gmm at nm 97.2 97.4 97.3 97.0 97.1 97.4 97.5 97.2 97.0 design binder content, % 5.1 4.0 3.5 5.1 3.6 3.9 4.8 4.5 4.3 design air void, % 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 vma, % 12.8 14.5 16.6 12.3 11.8 13.8 12.7 11.3 10.9 vfa, % 71.0 62.7 58.5 69.0 50.0 54.0 66.2 62.4 60.6 metric (u.s.) sieve gradation, (% passing) 19 mm (¾ in) 100 100 100 100 100 100 100 100.0 100.0 12.5 mm (½ in) 97.1 97.0 97.2 96.0 96.6 97.2 97.3 97.7 98.3 9.5 mm (3/8 in) 80.3 80.2 81.7 80.7 83.8 86.5 79.3 82.5 86.8 4.75 mm (no.4) 46.9 55.2 59.8 48.6 57.6 64.7 46.7 54.4 65.0 2.36 mm (no.8) 31.5 39.6 46.1 32.8 41.6 48.4 33.2 39.5 49.0 1.18 mm (no.16) 21.8 27.9 34.7 22.2 31.5 36.9 23.3 27.8 35.4 0.6 mm (no.30) 15.3 19.7 25.6 16.2 23.7 27.8 16.6 19.7 25.5 0.3 mm (no.50) 9.3 11.1 14.4 12.1 15.9 17.7 10.1 11.7 14.6 0.15 mm (no.100) 6.6 7.4 9.3 6.7 11.2 12.1 6.6 7.4 9.0 0.075 mm (no.200) 5.5 6.0 7.2 4.2 8.4 9.1 4.8 5.4 6.5 table 1. job mix formula 12.5 mm mixes ndes = 125 gyrations 1 ls: limestone, sst: sandstone, gr: granite 0x ym e   15 the journal of engineering research vol 7, no. 1 (2010) 11-21 (2) (3) p1, p2, p3 are load-cell forces; ex and ey are xand ycomponents of the eccentricity, e; and ry is the location of the plate center point with respect to the x-axis. the frictional shear resistance of the asphalt mixture can be calculated using the following relationship: (4) where, fr = the frictional resistance r = resultant force e = eccentricity a = cross-section area h = sample height at any gyration cycle. two specimens per mixture were tested for compactability in both sgc and pda devices. figures 5 and 6 present an example of the behavior of the asphalt mixtures during the compaction process using the data from both sgc and pda devices. the figures show the rate of change in specimen height (from the sgc) and the frictional resistance (from the pda). it is clear that the mixture reaches a condition where applying additional compaction energy results in little or no effect in further densifying the mixture. the point at which the mixture starts exhibiting this level of compaction resistance is termed the locking point of the mixture and is defined as follows: sgc locking point : the sgc locking point is the number of gyrations after which the rate of change in height is equal to or less than 0.05 mm for three consecutive gyrations (fig. 5). pda locking point: it is defined as the number of gyrations at which the rate of change of frictional resistance per gyration is less than 0.01 (fig. 6). the locking point data presented in figs. 7 and 8 for both the sgc and the pda respectively indicate that it takes more energy to densify coarse mixtures compared to the medium and fine mixtures. as the aggregate gradation becomes finer, the compactability of the mixtures generally improves. locking points are much lower than the design number of gyrations recommended by the current superpave design system. the highest locking point is about 70% of the recommended design number of gyrations for the heavy-traffic category (ndes=125). figure 9 presents the good correlation between the locking points determined from the sgc and those determined from the pda. on average, the pda locking points were about 4 gyrations lower than those determined from the sgc data. mixture name ls coarse ls medium ls fine mix type 25.4 25.4 mm 25.4 mm binder type pg 76-22m pg 76-22m pg 76-22m design ac content, volumetric properties, and densif ication % gmm at ni 85.0 88.8 89.1 % gmm at nm 97.7 97.4 97.4 design binder content, % 3.8 3.0 3.3 design air void, % 4.0 4.0 4.0 vma, % 11.1 9.6 10.0 vfa, % 63.5 58.2 60.5 metric (u.s.) sieve gradation, (% passing) 37.5 mm (1½ in) 100 100 100 25 mm (1 in) 92.4 92.6 95.2 19 mm (¾ in) 78.8 79.3 86.5 12.5 mm (½ in) 64.7 66.0 76.9 9.5 mm ( ? in) 52.5 56.1 65.8 4.75 mm (no.4) 36.4 43.2 50.3 2.36 mm (no.8) 24.5 32.7 38.1 1.18 mm (no.16) 15.8 24.3 28.2 0.6 mm (no.30) 10.6 17.6 20.4 0.3 mm (no.50) 7.7 8.7 10 0.15 mm (no.100) 6.1 5.2 5.9 0.075 mm (no.200) 5.1 4.2 4.8 table 2. job mix formula 25.4 mmmixes ndes = 125 gyrtions 0y xm e    22x y ye e r e   re fr ah  3/8 in) 16 the journal of engineering research vol 7, no. 1 (2010) 11-21 6. physical properties of mixtures at their locking point a limited number of mixtures were selected for mixture design using the locking point concept as opposed to the traditional superpave ndes. graphical comparisons of some physical properties from both sets of mixtures are presented in figs. 10 through 12. as anticipated, coma b c 0.00 5.00 10.00 15.00 20.00 25.00 0 50 100 150 200 250 300 350 n o . o f gy r a t i o n s d figure 4. pressure distribution analyzer (a) the pda device (b) analysis of forces (c) inserting the pda in the compaction mold (d) typical results -0.5 0 0.5 1 1.5 2 2.5 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 no. of gyrations r at e of c ha ng e no. of gyrations r at e of c ha ng e figure 5. rate of change of the height sgc compaction 17 the journal of engineering research vol 7, no. 1 (2010) 11-21 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 10 20 30 40 50 60 70 80 90 10 0 11 0 12 0 13 0 14 0 15 0 16 0 17 0 18 0 19 0 20 0 21 0 22 0 23 0 24 0 25 0 26 0 27 0 28 0 29 0 30 0 31 0 32 0 no. of gyrations r at e of c ha ng e of f r no. of gyrations r at e of c ha ng e of f r figure 6. rate of change of frictional resistance sgc compaction s g c lock ing points 70 62 83 73 72 71 68 60 61 878787 0 20 40 60 80 100 1/2" ls c 1/2" ls m 1/2" ls f 1/2" s s t c 1/2" s s t m 1/2" s s t f 1/2" g r c 1/2" g r m 1/2" gr f 1" ls c 1" ls m 1" ls f l oc k in g p oi n figure 7. sgc locking point results pda locking points 64 57 73 69 79 65 67 73 51 49 82 79 0 20 40 60 80 100 1/2" ls c 1/2" ls m 1/2" ls f 1/2" sst c 1/2" sst m 1/2" sst f 1/2" gr c 1/2" gr m 1/2" gr f 1" ls c 1" ls m 1" ls f pd a l oc ki ng p oi nt s figure 8. pda locking point research c ha ng in g po in ts pd a l oc ki ng p oi nt s 18 the journal of engineering research vol 7, no. 1 (2010) 11-21 y = 0.97x 3.96 r2 = 0.85 40 50 60 70 80 90 100 40 50 60 70 80 90 100 sgc locking point pd a l oc ki ng p oi nt figure 9. sgc and pda locking points correlation 4.9 4.1 5.4 4.2 3.9 3.3 3.6 5.1 3.5 4.3 0 1 2 3 4 5 6 grf lsf lsc ssm 1" lsf % ac ndes locking point figure 10. comparison of the design asphalt content 12.2 10.5 13.7 9.2 11.1 10.0 8.4 13.5 9.4 10.9 0 2 4 6 8 10 12 14 16 grf lsf lsc ssm 1" lsf vm a, % ndes locking point figure 11. comparison of the vma results pd a l oc ki ng p oi nt % a c v m a % 19 the journal of engineering research vol 7, no. 1 (2010) 11-21 pacting mixtures to their locking point yielded higher design asphalt contents than those obtained when superpave-design number of gyrations was used. the asphalt content for mixtures designed using the locking point ranged from 3.9% to 5.4% compared to 3.3% to 5.1% for the same mixtures designed using the traditional ndes. it is worth noting that except for the half-inch coarse limestone mixture, there was about 0.6% increase in asphalt content for all other mixtures when they were designed using their locking points at the same level of 4.0% air void. the voids in mineral aggregates (vma) values were 1.1% to 1.2% higher for the mixtures designed at the locking point except for the medium sandstone mixture in which there was 0.8% increase. higher asphalt contents naturally resulted in higher vfa, lower dust/pbeff ratio, and hence higher effective film thickness for the mixtures in consideration. 7. estimating locking point to facilitate the design process, a multiple linear regression model was developed using sas software (sas 2002) to estimate the locking point of the mixture based on certain properties that are thought to influence the performance of the mixture during compaction. the response parameter used was the locking point (lp). since the compaction process is always performed at elevated temperatures, the influence of aggregate structure is thought to be more pronounced than that of the binder although the binder will still maintain some lubrication effect that might contribute to the mixture’s response to the applied compaction energy. several parameters were first introduced in the model including different characteristics of the gradation curves of the designed aggregate structure as well as binder content. a stepwise variable selection procedure was first performed on a general model that contains those variables. the purpose of such a procedure is to remove insignificant variables from the general model. the regression analysis was then conducted on the reduced model using the stepwise variable selection procedure. three parameters were used in the regression analysis which were significant when included in the model as independent variables. these were: the predictive model used is: (5) where lp, vca, p200* ac are: lp = locking point to be estimated vca = volume of coarse aggregate in the aggregate structure p200 * ac = the interaction between the effect of the amount of material passing #200 sieve in the aggregate structure and the estimated asphalt content. the results of the regression procedure are shown in table 3. the fvalue for the model was 45.44 with a pvalue of <0.0001. this indicates that the model is significant in describing the relationship between the response variable and the independent variables. all the parameter estimates for the predictor variables in the model were significant at the 95% significance level selected for the analysis. the model was also checked for any co-linearity between the predictor variables. when there is a perfect linear relationship among the predictors, the estimates for 5.6 4.3 3.3 6.1 5.2 2.5 8.7 3.4 4.5 9.0 0 1 2 3 4 5 6 7 8 9 10 grf lsf lsc ssm 1" lsf te ff, m ic ro ns ndes locking point figure 12. comparison of the effective film thickness  volume of coarse aggregate in the aggregate structure (vca) . the designer can determine the vca in the dry condition of aggregate by performing a unit weight test on the combined material retained on no. 4 sieve for a given blend according to aashto t -19 test procedure (aashto t -19, 2004) , along with determining the combined specific gravity for this material.  percent passing #200 sieve for the aggregate structure in consideration. this parameter is termed as “p 200”.  estimated initial asphalt content (ac).  2001 38 0 62 6 86lp . * vca . * p * ac .   te ff , m ic ro ns 20 the journal of engineering research vol 7, no. 1 (2010) 11-21 a regression model cannot be uniquely computed. the term co-linearity describes two variables that are nearperfect-linearity combinations of one another. when more than two variables are involved, it is often called multi-co linearity, although the two terms are often used interchangeably. the primary concern is that as the degree of multi-co linearity increases, the regression model estimates of the coefficients become unstable and the standard errors for the coefficients can get wildly inflated. the 'vif' option was used to check for multi-co linearity. if stands for variance inflation factor. as a rule of thumb, a variable whose 'vif' value is greater than 10 may merit further investigation. a comparison between the measured and predicted response variable is shown in fig. 13. 8. conclusions the several key findings of this study may be summarized as follows: analysis of variance source df sum of squares mean square f value pr > f model 2 1413.28 706.64 45.44 <.0001 error 11 171.08 15.55 corrected total 13 1584.36 root mse 3.94 r-square 0.89 dependent mean 71.21 adj r-sq 0.87 coeff var 5.54 parameter estimates variable df parameter estimate standard error t value pr > |t| tolerance variance inflation intercept 1 -6.86 8.43 -0.81 0.4329 . 0 vca 1 1.38 0.15 9.04 <.0001 0.99 1.00 dac 1 0.62 0.18 3.45 0.0055 0.99 1.00 table 3. linear regression analysis to estimate locking point 50 55 60 65 70 75 80 85 90 95 100 50 60 70 80 90 100 measured locking point p re d ic te d l o c k in g p o in t r2=0.89 line of equality figure 13. accuracy of the locking point estimation model 21 the journal of engineering research vol 7, no. 1 (2010) 11-21 references alabama department of transportation, 2002, "special provision no. 02-0360(5)-2004 amendment for section 424 ," alabama standard specifications. american association of state highways and transportation officials, 2004," bulk density (unit weight) and voids in aggregate". aashto designation t 19. angelo, d.j., harman, t.p. and paugh, c.w., 2001, "evaluation of volumetric properties and gyratory compaction slope for the quality control of hot mix asphalt production," asphalt paving technology: association of asphalt paving technologistsproceedings of the technical sessions, vol. 70, pp. 729-761. asphalt institute, 2001. "superpave mix design," superpave series no. 2, asphalt institute, lexington, ky. cominsky r j., leahy r b. and harrigan e t., 1994, "level one mix design: materials selection, compaction, and conditioning," strategic highway research program, shrp-a-408. georgia department of transportation, 2003, "special provision-section 828-hot mix asphaltic concrete mixtures,". guler, m., bahia,h.u., bosscher, p, j. and plesha, m. e., 2000, "device for measuring shear resistance of hot-mix asphalt in gyratory compactor," transportation research record no. 1723, transportation research board, washington, dc. sas institute inc, 2002-2003, "sas help and documentation". cary, nc, usa. the online magazine of the asphalt institute, 2007, " www.asphaltmagazine.com". vavrik, w. r. and carpenter, s.h., 1998, "calculating air voids at specified number of gyrations in superpave gyratory compactor," transportation research record 1630, transportation research board, washington, dc, pp. 117-125.  data from the sgc provide valuable information on the compactability of asphalt mixtures.  both sgc and pda results suggest that coarse mixtures are more difficult to compact compared to the medium and fine ones. this emphasizes the importance of relating the level of applied compaction energy to some specific attribute of the asphalt mixtures and not basing it solely on the expected traffic level as currently practiced in the superpave mixture design methodology.  the compaction data also suggest that the current recommended superpave design number of gyrations is too high and subject the mixtures to unnecessary high compaction loads for extended periods of time, which might have an adverse effect on the final mixture volumetrics.  there was a strong correlation between the data from the sgc and pda. this suggests the data from the sgc provides a good indication of mixture compactability.  a statistical estimation model was developed based on parameters that have a significant effect on the compactability of asphalt mixtures. this model can help mix designers in establishing the optimum compaction levels for their mixtures. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 82-90 *corresponding author’s e-mail: raa37@kent.ac.uk doi:10.53540/tjer.vol18iss2pp82-90 a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan rawan allouzi*, nikolaos karydis, and marialena nikolopoulou school of architecture and planning, university of kent, canterbury, united kingdom abstract: the settlement of dana in jordan is undergoing dramatic and rapid changes. the rehabilitation project launched by the royal society for the conservation of nature (rscn) provided new tourist accommodation within the settlement, in a way that preserves its unique vernacular character. this has involved the repair and transformation of existing dwellings as well as the rebuilding of those that have been demolished. currently, the original and rebuilt buildings stand side by side, offering a unique opportunity to compare their thermal performance. this comparison is essential to understand the impact of new construction materials and opening sizes on the thermal performance of vernacular buildings. for this purpose, the thermal performance of rebuilt and original buildings was monitored in august 2019 and february 2020, representing the hot and cold seasons. the recorded data was interpreted statistically, aiming mainly to compare the results from the original rooms with those from the rebuilt rooms. the study showed that the original buildings of dana hotel resist thermal transfer more than the rebuilt ones. both parts provided thermal comfort in summer, with the rebuilt rooms offering slightly better thermal comfort conditions. however, neither original nor rebuilt rooms provided adequate thermal comfort in winter. keywords: dana village; original buildings; rebuilt buildings; thermal comfort; thermal performance; vernacular settlement. دراسة مقارنة بین األداء الحراري لمباني تراثیة أصلیة، وأخرى معاد بناؤھا في قریة ضانا في األردن ، نیكوالس كاریدیس، ماریالینا نیكولوبولو*روان اللوزي الذي أطلقتھ الجمعیة الملكیة -وفر مشروع إعادة التأھیل تمر قریة ضانا في األردن بتغیرات دراماتیكیة وسریعة. :الملخص أماكن إقامة سیاحیة جدیدة في القریة بطریقة تحافظ على طابعھا المحلي الفرید، كما اشتمل على ترمیم –للحفاظ على البیئة ؤه جنبا إلى جنب في الوقت الحالي، وتحویل المساكن القائمة، وإعادة بناء المنازل المتھدمة. تتواجد المباني األصلیة والمعاد بنا مما یوفر فرصة فریدة لمقارنة أدائھا الحراري، وھي مقارنة ضروریة لفھم تأثیر مواد البناء الجدیدة وقیاسات الفتحات على أغسطس األداء الحراري للمباني التقلیدیة، ولھذا الغرض، فقد تمت مراقبة األداء الحراري للمباني األصلیة والمعاد بناؤھا في ، باعتبارھم ممثلین للمواسم الحارة والباردة. على نحو أساسي بھدف مقارنة النتائج من الغرف األصلیة 2020وفبرایر 2019 وتلك المعاد بناؤھا، فقد تمت معالجة البیانات المسجلة احصائیا. أظھرت الدراسة أن المباني األصلیة لفندق ضانا تقاوم انتقال المعاد بناؤھا، كما یوفر كال النوعین من المباني الراحة الحراریة في فصل الصیف، مع مالحظة أن الحرارة أكثر من تلك الغرف المعاد بناؤھا توفر بشكل طفیف مستوى أعلى من األحوال الحراریة المریحة، وبالرغم من ذلك فال توفر الغرف فصل الشتاء. األصلیة أو المعاد بناؤھا مستوى مناسب من الراحة الحراریة خالل .قریة ضانا ؛قریة تقلیدیة ؛الراحة الحراریة ؛المباني المعاد بناؤھا ؛المباني األصلیة ؛األداء الحراري الكلمات المفتاحیة: a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 83 nomenclature r-value: thermal resistance (m2.k/w) tair: air temperature (°c) tout: outdoor temperature (°c) u-value: thermal transmittance (w/m2.k) 1. introduction dana village (30.67° n, 35.61° e) in jordan was built between the 1890s and 1930s. dana was formed of single-story buildings with flat roofs. the organization of contiguous buildings in clusters is one of the main characteristics of the village, looking like one cohesive/compact unit. after the 1960s, the local community of dana started to move out, looking for better services and job opportunities. by 1996, the village was almost abandoned, and most of its buildings turned into ruins (figure 1). therefore, the rscn launched the rehabilitation project of dana village in 2009 to restore vernacular dwellings to accommodate tourists and visitors (al haija, 2011). several restoration strategies were implemented, such as the replacement of the original materials, the transformation of the building envelope, the enlargement of the opening sizes, and the addition of lighting shafts, and bathrooms (table 1). the implementation of these changes affected the indoor thermal conditions that depend mainly on several factors, including the construction materials, opening sizes, thermal mass, and orientation (alzoubi & almalkawi, 2019; fernandes et al., 2015; meier et al., 2004). various previous studies discussed the effect of these factors on an individual basis, comparing the thermal performance of traditional and contemporary buildings. for instance, tawayha et al. (2019) stated that using local and natural materials used in vernacular architecture plays an essential role in improving thermal performance. weber & yannas (2013) and priya et al. (2012) added that the advantages of using local construction materials in improving the thermal performance of vernacular buildings include the low conductivity of thermal transfer and providing stable indoor thermal conditions. furthermore, toe & kubota (2015) and philokyprou & michael (2012) stated that the high thermal mass improves the indoor temperature stability, reduces daily fluctuation, and increases the time lag. indeed, having thicker walls with higher thermal mass adapts vernacular buildings to the hot conditions by reducing exposure to direct solar radiation. whereas in winter, the small size of openings provides thermal comfort as it allows the direct solar gains and protects the indoor spaces from cold and strong winds (tawayha et al., 2019). previously conducted studies presented the impact of transforming the building’s envelope on an individual basis, including the construction materials, thermal mass, and opening sizes and orientation (weber & yannas, 2013; fernandes et al., 2015; alzoubi & almalkawi, 2019). however, there is a shortage in the study of the total impact of the restoration strategies of vernacular buildings as most studies focused on comparing vernacular and contemporary buildings. therefore, this paper seeks to assess the effectiveness of the restoration strategies of dana village by analyzing the thermal performance of both the original and rebuilt vernacular buildings. it aims to identify which type is more effective in providing thermal comfort in summer and winter while highlighting the effect of the restoration on thermal performance. for this purpose, the internal and external conditions of dana hotel were monitored in august 2019 and february 2020 as representative of the hot and cold season. dana has a cold climate in winter and a temperate to warm climate in summer (table 2). figure 1. figure ground plan of dana village (top), dana village before (middle) and after abandonment (bottom) (archive, rscn, 2019). table 1. the original and rebuilt elements of dana village. rawan allouzi, nikolaos karydis, and marialena nikolopoulou 84 table 2. dana monthly air temperature (climate-data.org. retrieved june,21,2020). tair(°c) jan feb mar apr may june july aug sep oct nov dec avg 5.6 6.6 9.6 13.2 16.5 19.6 21.1 21.7 19.4 16.7 11.7 7.4 min 1.5 1.8 4.3 7.4 10.2 12.6 14.6 15.2 12.7 10.4 6.5 2.7 max 9.8 11.4 14.9 19.1 22.9 26.7 27.7 28.3 26.2 23 17 12.2 table 3. thermal properties of the original and new materials (alzoubi & almalkawi, 2019, adapted by the author, 2020) original part rebuilt part m at er ia ls t hi ck ne ss (c m ) t he rm al co nd uc ti vi ty (w /m k ) r -v al ue (m 2 .k /w ) u -v al ue (w /m 2 .k ) m at er ia ls t hi ck ne ss (c m ) t he rm al co nd uc ti vi ty (w /m k ) r -v al ue (m 2 .k /w ) u -v al ue (w /m 2 .k ) e xt er io r w al l stone 70 5 0.14 1. 30 7 stone 20 5 0.040 3. 35 6 a mixture of clay and straw 10 0.16 0.625 concrete 10 1.7 0.059 hollow concrete block 15 0.833 0.180 cement plaster 2.5 1.3 0.019 g ro un d fl oo r earth 50 1.4 0.357 0. 36 7 earth 50 1.4 0.357 1. 09 6 gravel jeer 45 0.38 1.184 base-course stone 15 1.5 0.10 base course stone 10 1.51 0.066 asphalt insulation 5 1.2 0.042 reinforced concrete 10 1.28 0.078 sand 45 0.42 1.071 sand 10 0.40 0.25 cement mortar 2.5 0.931 0.027 cement mortar 2.5 0.39 0.064 terrazzo tile 2.5 1.2 0.021 terrazzo tile 2.5 1.20 0.021 r oo f juniper wood 5 0.17 0.294 0. 48 1 cement mortar 2.5 0.93 0.027 3. 25 7 cane 3 0.056 0.536 concrete 15 1.7 0.088 a mixture of clay and straw 20 0.16 1.25 asphalt insulation 2 1.2 0.017 reinforced concrete 20 1.28 0.156 cement plaster 2.5 1.3 0.019 2. dana hotel dana hotel, established in the late 1980s, was selected as a representative case study of dana village as it consists of both original and rebuilt rooms. the original part of dana hotel is formed of 12 single-story rooms, surrounding the central courtyard from three sides. it is built mainly of stone, juniper wood, cane, and a mixture of clay and straw. on the other hand, the rebuilt part of dana hotel is formed of 6 en-suite hotel rooms composed of a bedroom, bathroom, lighting shaft, and kitchenette (figure 2). it is built mainly of stone, concrete, hollow concrete block, and cement mortar. the use of new construction materials led to having different thermal properties between the original and new building envelope. the thermal properties of both building envelopes are shown in table 3. the calculated u-values highlighted that the building envelope of the original part of dana hotel resists thermal transfer more than the rebuilt part. 3. methodology to practically assess the thermal performance of dana hotel, 10 rooms were monitored for two months in au gust 2019 and january 2020, including 6 original rooms and4 rebuilt rooms (figure 2). figure 2. the original and rebuilt part of dana hotel with logger positions (archive, rscn, 2019). tinytag loggers were used to record air temperature and relative humidity every 15 minutes, placed at 1.5 m above the floor level. the collected data was analyzed and compared between the original and rebuilt rooms, demonstrating the variation of the indoor temperatures throughout the day and providing thermal comfort as the criteria of assessment of this study. a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 85 these criteria of assessment were first analyzed within the original and rebuilt part of dana hotel separately. the statistical method was used to analyze the monitored physical variables to assess the thermal performance. it illustrated the fluctuation of the recorded data (air temperature) throughout the day, from day to day, and from room to room. in addition, thermal comfort was investigated through the adaptive thermal comfort standards bsen15251 (nicol et al., 2012). then, a comparison study was conducted between the final results of both parts, using it as an evaluation method to assess the transformation of dana’s original vernacular buildings within the thermal context. 4. results 4.1. thermal environment 4.1.1. air temperatures in the original rooms the mean indoor temperature of the original spaces was always higher than the mean outdoor temperature in summer and winter (table 4). a typical week of the fluctuation of air temperatures in the different spaces in the different seasons is illustrated in figure 3. in summer, indoor temperatures followed the outdoor temperature profile but with a significant time lag. more specifically, the indoor temperature was low at night, dropping in the early morning to the minimum and then rising rapidly until the afternoon, when reaching the maximum temperature. however, the maximum temperature was always less than the outdoor temperature around noon due to the thermal mass in all spaces except in the kitchen (space #6), which is assumed to be from the high internal heat gains from cooking. the high thermal mass of the original masonry thick walls decreased heat transfer to indoor spaces, preventing heat from reaching indoors in the afternoon, which is the hottest part of the day, reaching 36°c. the high thermal mass demonstrates a thermal lag of 14 hours. therefore, the night-time temperature difference between indoor and outdoor, which sometimes approaches 10°c, was more than during the day. furthermore, the daily variation was noticeable in some spaces more than others due to different functions and occupancy times. small variation and less significant fluctuation over the day were noticed in rooms #1, #3, #4, and #5. this is because the thermal inertia of the thick stone walls and clay roofs offered small fluctuations in the indoor temperature. the high thermal inertia of the original materials increased their capacity to store heat and then emitted it outdoors when the temperature is lower while a small percentage was emitted indoors. however, lower indoor temperatures were noticed in rooms #4 and #5 because their roof is not exposed to the outdoor conditions, having an additional floor at their side. meanwhile, the roof of rooms #1 and #3 is exposed to outdoor conditions and solar radiation, leading to absorbed heat being transferred to indoor spaces. on the other hand, the diurnal swing values at rooms #2 and #6 were more significant and variable, ranging from 4°c to 9.7°c. for room #6, being used as a kitchen explained the excessive indoor temperature that rose above the outdoor temperature most of the time due to cooking’s produced heat. therefore, the kitchen showed the same fluctuation and temperatures of rooms #4 and #5 when it was out of service for maintenance. furthermore, the absence of windows also raised the indoor temperature, storing the heat indoors without efficient airflow and ventilation. whereas for room #2, it was observed that its roof was replaced with a concrete one. having its roof transformed among the other original spaces created a difference in the thermal performance. the new concrete roof failed to resist heat transfer due to its lower capacity to resist heat transfer than the original clay roofs. therefore, the indoor temperature of room #2 showed more fluctuation, reaching higher points compared with the other original hotel rooms. a different pattern of variations was noticed in winter with lower temperatures. the indoor temperature followed the profile of the outdoor temperature in room #3, being the coldest room among the other original rooms. its indoor temperature was either lower than the warm outdoor conditions or close to the cold outdoor conditions. specifically, the indoor temperature did not increase with higher outdoor temperature but decreased with lower outdoor temperature. this is because the sun does not enter this room due to the absence of windows. moreover, it showed less variation in temperatures as it was observed that it was not occupied during the winter monitoring period, and the hotel operators confirmed that. similarly, rooms #6 and #4 were also unoccupied but showed higher temperatures due to reduced heat losses from their roof, which is not exposed to the cold outdoor conditions. room #1 had similar thermal behaviour to room #3 until it started being occupied as it is assumed that higher fluctuation appears during occupation due to the internal heat gains from occupants. it showed warmer conditions in the morning due to the eastern window that brings the sunrays indoors. however, still have small values of diurnal swing that vary from 0.1°c to 3°c. table 4. indoor air temperatures (°c) in the original rooms. tout room #1 room #2 room #3 room #4 room #5 room #6 jan aug jan aug jan aug jan aug jan aug jan aug jan aug mean 6.1 24.7 7.6 26 9.7 26.6 5.6 27 7.9 24.9 6.1 24.8 7.7 27.5 min. -0.9 16.3 4 23.4 5.4 22.6 2.1 24.6 6.7 21.9 2.4 21.7 5.6 23.3 max. 20.4 36.5 11.5 29.2 16.8 33.8 9.7 29.8 9.8 27.2 10.2 27.2 10.3 35 a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 86 figure 3. variation of indoor air temperature in the original rooms in a typical week in summer (top) and winter (bottom). figure 4. variation of indoor air temperature in the rebuilt rooms in a typical week in summer (top) and winter (bottom). on the other hand, the daily variation was more noticeable in room #2, with diurnal swing ranging between 2.4°c and 8°c. the indoor temperature was, most of the time, higher than the outdoor temperature the difference between the outdoor and indoor temperature sometimes approached 12°c when the outdoor temperature was cold, reaching -0.9°c. the indoor temperature reached its maximum in the afternoon and minimum at midnight. this is because when the outdoor temperature reached its maximum near noon, the concrete roof of room #2 transferred the absorbed heat to the indoor space after a short time due to its lower thermal capacity. as expected, the variation of indoor temperature was influenced by the building envelope and opening sizes and orientation, as well as construction materials. the few and small size of windows protected the indoor spaces from direct sun rays during summer days and offered good ventilation and cooling at night. meanwhile, these windows allowed solar gain to indoor spaces during winter days due to their lower elevation angle. 4.1.2. air temperatures in the rebuilt rooms the mean indoor temperature of the rebuilt spaces was close to the average outdoor temperature in summer, while it was always higher in winter (table 5). a typical week of the fluctuation of air temperatures in the different spaces in the different seasons is illustrated in figure 4. in summer, the trend of the indoor and outdoor temperature was different than the original rooms as the changes in the indoor temperature did not follow the changes in the outdoor temperature. specifically, the indoor temperature was higher than the outdoor temperature at night and lower during the day. the fluctuation of the indoor temperature was small, and the temperature was consistent. indeed, the daily variation was not noticeable in all spaces despite different occupancy hours as it was observed that tourists spent their daytime touring the village and its neighbouring areas. they reached the maximum temperature in the afternoon and the minimum in the morning. they provided cooler indoor temperatures when it was the warmest outside and warmer indoor conditions when it was the coldest outside. the enlargement of windows and the addition of lighting shafts helped in providing cooler conditions. in terms of enlarging the external windows, these windows enhanced air movement due to their bigger size. this strategy was efficient as they are northern windows, which do not bring in solar gains. moreover, the addition of lighting shafts provided internal windows that enhanced the air movement and ventilation for cooling. although room #7 does not have a lighting shaft, it provided cool conditions due to having three windows compared with one window in the other rooms. it was also observed that room #8 had slightly colder conditions compared with room #9 and room #10, although they have the same building envelope, opening size, and orientation. this is because room #8 is bigger, increasing the heat transfer area in the room. moreover, room #7 and room #8 have approximately the same area, but room #7 was slightly warmer due to having southwestern windows that bring in solar gains, warming up the indoor spaces in the afternoon. table 5. indoor air temperatures (°c) in the rebuilt rooms. tout room #7 room #8 room #9 room #10 jan aug jan aug jan aug jan aug jan aug mean 6.1 24.7 7.1 25.2 9.4 24.2 8.8 25.8 7.4 26.3 min. -0.9 16.3 4.5 23.6 6.9 22.5 5.9 23 4.2 23.6 max. 20.4 36.5 9.9 27.1 13.7 26 16.5 28.3 17 29.7 a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 87 a different pattern was noticed in winter with wider fluctuation, less affected by the changes of the outdoor temperature. overall, the indoor temperature was higher than the outdoor temperature most of the time. moreover, the daily variation was noticeable sometimes in some spaces more than others. for instance, the diurnal swing varied from 0.1°c to 8.7°c, showing small values at room #7 due to non-occupancy status. meanwhile, the diurnal swing values at rooms #8, #9, and #10 were bigger and more variable according to their occupancy time. 4.1.3. comparison between temperatures in the original and rebuilt rooms the pattern of the indoor temperature variation was different in the original and rebuilt rooms of dana hotel. in the summer, indoor temperatures were higher in the original rooms compared with the rebuilt rooms. the indoor temperature was more steady in the rebuilt rooms, showing less variation. on the other hand, the original spaces performed differently, as some spaces such as #2 and #6 showed more variation compared with the other original spaces. the more extreme values were noticed in the original spaces. however, they were closer to the mean value in the rebuilt spaces. in winter, the indoor temperatures were either similar in both properties and slightly warmer in the rebuilt rooms. moreover, the original spaces showed different performance as some rooms such as #4 and #6 show less variation in the indoor temperature compared with the other original spaces as they were unoccupied at that time. meanwhile, the other original and rebuilt spaces provided almost the same variations in the indoor temperature (figure 5). the analysis highlighted that performance is greatly affected by the materials of the building envelope, opening sizes and orientation. the external walls of the original rooms are thicker with higher thermal storage compared to the walls of the rebuilt rooms, improving the resistance to heat gain in summer and heat loss in winter. moreover, the original rooms of dana hotel have small openings to reduce the climatic influence on the indoor thermal conditions. along with the thick walls, they prevent solar gains in summer, limit the heat gains and losses, and benefit from night ventilation and wind breeze to cool down indoor spaces. meanwhile, it reduces heat loss and allows solar gains in winter figure 6). on the other hand, the addition of lighting shafts and the enlargement of northern windows adapt the rebuilt rooms to the outdoor conditions in summer. bigger and more windows were used in the rebuilt rooms, enhancing air movement while preventing solar gains. the orientation of these openings reduced the impact of their size because the direct solar gain is the main contributor affecting the indoor temperature. therefore, the inward and northern orientation of the windows and lighting shafts of the rebuilt rooms reduce the impact of being more exposed in summer. moreover, passive cooling strategies, such as plants and trees in front of the rebuilt rooms, enhance night cool breezes. the adopted changes in the rebuilt rooms worked together to adapt to dana’s weather as a whole system, including the replacement of the building envelope, the addition of lighting shafts, and the enlargement of northern windows. replacing the building envelope without increasing the opening sizes would disable the whole system, as observed in room #2. the replacement of the original roof while keeping the original opening sizes failed to adapt to the outdoor conditions, providing more variation in the indoor temperatures. figure 5. box plot that shows the mean, extreme, and outlier values of indoor air temperature at each space in the original and rebuilt part of the hotel in summer (top) and winter (bottom). figure 6. the role of the small size of windows in allowing solar gains in winter and preventing summer’s ones. a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 88 4.2. thermal comfort 4.2.1. thermal comfort in the original rooms in summer, the daily mean indoor temperatures were entirely within the calculated ranges of comfort temperatures. the large area of the central courtyard plays a significant role in enhancing the efficiency of the original rooms in providing thermal comfort. the courtyard cools down the indoor temperature in summer as it is shaded by the surrounding walls, preventing the sunrays from warming up the rooms. moreover, the addition of plants, pergolas, and trees also helps provide shading and cooling by evaporation from the plants’ green leaves. overall, 100% of the mean indoor temperature lay within the comfort ranges; most daily mean indoor temperatures lay within the comfort range of category ι, while others lay within the other two categories. thus, desired thermal comfort was achieved in the original rooms in summer. in winter, however, 100% of the daily mean indoor temperatures lay outside the three categories of comfort, showing that the daily mean indoor temperature was significantly below comfort temperatures, suggesting that occupants would be in cold discomfort. thus, all the original rooms showed failure in providing thermal comfort in winter figure 7). 4.2.2. thermal comfort in the rebuilt rooms in summer, 100% of the mean indoor temperature lay within the comfort ranges; most daily mean indoor temperatures lay within the comfort range of category ι, while others lay within category ιι. thus, thermal comfort was achieved in the rebuilt rooms in summer. in winter, 100% of the daily mean indoor temperatures also lay outside the three categories of comfort, suggesting that occupants would be in cold discomfort. thus, all the rebuilt rooms showed a failure to provide thermal comfort in winter figure 8). 4.2.3. comparison between thermal comfort at the original and rebuilt rooms starting with the summer, both the original and rebuilt rooms provided thermal comfort. however, the rebuilt rooms offered slightly better thermal comfort avoiding the higher indoor temperatures. in addition to the passive cooling strategies discussed in both the original and rebuilt rooms, the urban pattern of dana village played an essential role in providing thermal comfort, showing how the original occupants adapted their houses to the climatic conditions (warm summer and cold winter). the original houses were attached, sharing partition walls. this layout is useful in regions with hot and cold climatic conditions, as it minimizes the exposure of the building envelope to the outdoor environment, which consequently reduces heat transfer (bodach et al., 2014; philokyprou et al., 2017). this urban pattern also features open spaces, including streets and private courtyards that are well relatively small and well-enclosed. in this arrangement, the external walls are less exposed to the outdoor climatic conditions, increasing the building envelope’s thermal insulation. figure 7. scatter diagram of the mean indoor temperature within the comfort temperature ranges for the original rooms in summer (top) and winter (bottom). figure 8. scatter diagram of the mean indoor temperature within the comfort temperature ranges for the rebuilt rooms in summer (top) and winter (bottom). a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 89 table 6. the results of comparing the thermal conditions of the original and rebuilt rooms in summer and winter. summer winter original rebuilt original rebuilt temperature higher temperature lower temperature almost the same almost the same diurnal swing higher lower almost the same almost the same decrement value higher lower almost the same almost the same thermal comfort provide thermal comfort provide slightly better thermal comfort do not provide thermal comfort do not provide thermal comfort philokyprou et al. (2013) stated that due to narrow streets and spaces between houses, the neighbouring houses shade each other and reduce the heat gain through walls by radiation. on the other hand, both the original and rebuilt rooms did not provide thermal comfort in winter. indeed, their daily mean indoor temperatures were very low, lying entirely outside the comfort ranges. the analysis of the thermal performance of the original and rebuilt rooms showed that the rscn was more concerned about providing thermal comfort in the summer months because dana is a summer tourist destination. focus on the summer is also because winters in the region are normally shorter than summers. 5. discussion and conclusion the thermal performance of the vernacular buildings of dana demonstrated that the passive and heavyweight buildings, free-running with no heating or cooling systems, can provide thermal comfort during the summer. singh et al. (2010), abdel hadi (2013), philokyprou et al. (2013), weber & yannas (2013), and tawayha et al. (2019) demonstrated that the thermal performance of vernacular buildings is generally better than contemporary buildings due to their construction materials and techniques. therefore, as the rebuilt rooms are built of contemporary materials, the thermal properties of the building envelope were determined by the materials of the external walls, floor, roof, and thickness. the calculated u-values showed that the building envelope of the original rooms of dana hotel resisted heat transfer better than the rebuilt rooms, confirming the results of the previous studies. however, along with the materials, the building design also influences the resulting indoor conditions, including the opening sizes and orientation (fernandes et al., 2015). therefore, monitoring the internal thermal conditions was conducted to show the total impact of the building design. the results showed that the rebuilt rooms provided stable indoor temperatures, although they do not provide the same thermal mass due to the effect of opening sizes and orientation. this highlighted that although the new construction materials provide higher u-values and less thermal resistance, the rebuilt rooms provided lower temperatures with a lower diurnal swing in summer, contradicting previous studies. overall, the analysis from the monitoring conducted showed different behaviour between the original and rebuilt part of dana hotel in summer, while similar behaviour was noticed in winter (table 6). it highlighted that the indoor temperatures were higher in the original rooms, showing more variations in summer. furthermore, both the original and rebuilt rooms provided thermal comfort in summer, with the rebuilt rooms offering slightly better comfort. this research could be expanded by conducting thermal simulation to advance the heat transfer analysis, showing the efficiency of the building’s envelope and its components in resisting heat transfer. moreover, discussing the comfort responses and adaptive behaviour of occupants through a post-occupancy evaluation survey would help identify challenges and suggest potential solutions to the original and rebuilt buildings. conflict of interest the authors declare no conflict of interest. funding no funding was received for this project. references al haija, a. a. (2011). jordan: tourism and conflict with local communities. habitat international, 35(1): 93-100. alzoubi, h. h., & almalkawi, a. t. (2019). a comparative study for the traditional and modern houses in terms of thermal comfort and energy consumption in umm qais city, jordan. journal of ecological engineering, 20(5): 14-22. bodach, s., lang, w., & hamhaber, j. (2014). climate responsive building design strategies of vernacular architecture in nepal. energy and buildings, 81: 227-242. fernandes, j., mateus, r., bragança, l. and correia da silva, j.j., 2015. portuguese vernacular architecture: the contribution of vernacular materials and design approaches for sustainable construction. architectural science review, 58(4): 324-336. meier, i. a., roaf, s. c., gileard, i., runsheng, t., stavi, i., & mackenzie-bennett, j. (2004). the vernacular and the environment towards a comprehensive research methodology. in proc. 21st int. conf. on plea (vol. 2, pp. 719-724). a comparison study between the thermal performance of the original and rebuilt vernacular buildings of dana village in jordan 90 nicol, f., humphreys, m., & roaf, s. (2012). adaptive thermal comfort: principles and practice. routledge. philokyprou, m., michael, a., & thravalou, s. (2013). assessment of the bioclimatic elements of vernacular architecture. the historic centre of nicosia, cyprus. in conference proceedings, le vie dei mercanti xi forum internazionale di studi, aversa, capri (pp. 13-15). philokyprou, m., michael, a., malaktou, e., & savvides, a. (2017). environmentally responsive design in eastern mediterranean. the case of vernacular architecture in the coastal, lowland and mountainous regions of cyprus. building and environment, 111, 91-109. weber, w. and yannas, s. eds., 2013. lessons from vernacular architecture. routledge. https://en.climate-data.org/asia/jordan/dana/dana459275/#temperature-graph. retrieved june,21,2020. روان اللوزي*، نيكولاس كاريديس، ماريالينا نيكولوبولو 1. introduction 2. dana hotel 3. methodology 4. results 4.1. thermal environment 4.1.1. air temperatures in the original rooms 4.1.2. air temperatures in the rebuilt rooms 4.1.3. comparison between temperatures in the original and rebuilt rooms 4.2. thermal comfort 4.2.1. thermal comfort in the original rooms 4.2.2. thermal comfort in the rebuilt rooms 4.2.3. comparison between thermal comfort at the original and rebuilt rooms 5. discussion and conclusion conflict of interest funding no funding was received for this project. references the journal of engineering research (tjer), vol. 16, no. 2 (2019) 142-150 solar pv power intermittency and its impacts on power systems – an overview m.h. albadi * department of electrical and computer engineering, college of engineering, sultan qaboos university, al-khoud, muscat 123, p.o. box 33, oman. abstract: although solar photovoltaic (pv) systems are environmentally friendly, policy makers and power system operators have concerns regarding the high penetration of these systems due to potential impacts of solar power intermittency on power systems. understanding the nature of this intermittency is important to make informed decisions regarding solar power plants, size and location, transmission and distribution systems planning, as well as thermal generation units and electricity markets operations. this article presents a review of solar pv power characteristics and its impacts on power system operation. keywords: solar pv power; intermittency; power system operation; reserve requirements. نظرة عامة الكهربائية على أنظمة الطاقة االطاقة الكهروضوئية وآثارهتذبذب م .ح. البادي * ومشغلي قرارصناع اللدي مخاوف بعض ال ه توجد( صديقة للبيئة ، إال أنpvضوئية )كهرولى الرغم من أن أنظمة الطاقة الشمسية الع :الملخص ولذالك فانه الطاقة الكهربائية. تشغيل نظمأنظمة الطاقة الشمسية الضوئية على عمليات انتاج ذبذبأنظمة الطاقة الكهربائية بشأن اآلثار المحتملة لت تخطيط شبكات النقل ك كذل، و اتخاذ قرارات مستنيرة بشأن حجم محطات توليد الطاقة الشمسية وموقعها من أجل ذبذبلتامن المهم فهم طبيعة هذا لخصائص الطاقة الشمسية الكهروضوئية عرض المقالةوعمليات أسواق الكهرباء. تقدم هذه ةوالتوزيع ، فضالً عن وحدات التوليد الحراري الطاقة الكهربائية. نظموتأثيراتها على تشغيل .األحتياطي الدوار ؛األنتاج تذبذب ؛الطاقة الكهربائية نظمتشغيل ؛ضوئيةكهروالطاقة الشمسية الالكلمات المفتاحية: * corresponding author’s e-mail: mbadi@squ.edu.om doi: 10.24200/tjer.vol16iss2pp142-150 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 142-150 1. introduction solar photovoltaic (pv) systems experienced a tremendous increase in installed capacity in the past decade. according to the 2019 ren21 renewables global status report, the global solar pv installed capacity increased from 17 gw in 2008 to 505 gw in 2018 as shown in fig. 1. about 100 gw solar pv capacity was added in the year 2018. this rapid growth of solar pv system deployment is attributable to several factors such as pv technology improvements, cost reduction, as well as policies and regulations promoting the use of renewable energy resources (ren21 2019). solar pv power is environmentally friendly and can be used to extend fossil fuel reserves’ life. however, the implications of intermittent nature need to be examined. intermittency of solar pv power affects the balance between supply and demand. when supply-demand balance is not maintained, power system frequency deviates from steady state values; consequently, system stability and reliability are jeopardized (kundur, paserba et al. 2004). although it is technically possible to integrate a large amount of intermittent renewable-based facilities in power systems, higher penetration levels result in more challenges to power system stability and reliability. these challenges are quantified using intermittency integration costs (albadi and elsaadany 2010). in general, the impacts of intermittent renewable-based generation facilities on power systems are due to two factors: variability and uncertainty (ummels, gibescu et al. 2007; albadi and el-saadany 2011). understanding solar power variability is necessary to make informed decisions regarding energy policy, solar energy conversion system design, transmission system planning, thermal generation units’ operation, and efficient electricity market operations (sengupta, xie et al. 2018). this paper presents a state-of-the-art review of solar pv power variability and its impacts of power systems. figure 1. global solar pv installed capacity. this introduction is followed by presenting the characteristics of solar pv power output in section 2. section 3 presents the impacts of pv power intermittency on power systems. section 4 discusses the lessons learned from pv integration studies. the main conclusions are summarized in section 5. 2. characteristic of solar pv output to highlight the variability aspect of solar pv power, fig. 2 presents a typical clear sky solar pv system output based on ground measurements and recorded load data of the main interconnected system (mis) of oman on 1 january 2015 (albadi 2017, opwp 2019). as seen from this figure, even with a clear sky and its variation, there is a challenge to power system operator. as seen from fig. 2, solar pv power output is not correlated with the load requirements. this mismatch poses a challenge to power system planner and operators. figure 3 presents the intermittency nature of solar pv power caused by passing clouds. the presented pv power data are based on recorded 1-minute solar irradiation data (opwp 2019). figure 2. manah site solar pv system output (clear sky) and recorded mis system load data on 1 january 2015. figure 3. manah site 1-minute pv power output data for 2 days in 2015 (passing clouds). 16 23 40 71 101 138 177 229 305 405 505 0 100 200 300 400 500 600 2008 2010 2012 2014 2016 2018 s o la r p v g lo b a l c a p a ci ty ( g w ) years 0% 20% 40% 60% 80% 100% 0 1000 2000 3000 0 4 8 12 16 20 24 p v p o w e r (% o f ra te d c a p a ci ty ) s y st e m l o a d ( m w ) time(hours) load pv power 0 0.5 1 0 400 800 1200 p v p o w e r (% o f ra te d c a p a ci ty ) time (minutes) 4-sep-15 4-nov-15 143 m.h. albadi variability is defined as the step changes from one averaging interval to the next one. there are two types of solar irradiance and pv plant output variabilities: deterministic and stochastic. the former variability is related to the position of the sun whereas the latter is related to clouds. the output of different pv plants can be accurately calculated for a clear sky; therefore, the deterministic variability between different plants located in the same geographical area is highly correlated and its smoothing effect is very limited. however, stochastic variabilities caused by passing clouds on a single pv plant are severe and their implications on power system operation can be substantial. the magnitude of stochastic variabilities can be obtained from the difference between the pv system(s) output power during a sunny day and a cloudy day. 2.1 solar pv variability statistical representation mathematically, variability is defined as the step changes from one averaging interval to the next one. to quantify the variability of an intermittent source of power, step changes of its average output is evaluated over different intervals (hoff and perez 2012). for example, considering availability of 1minute data ( ), the 5-minute step changes ( ̅ ) are obtained by calculating the average output for every 5-munite interval ( ̅ ), and then evaluating the changes between the calculated 5-minute data (hoff and perez 2012). ̅ ∑ (1) ̅ ̅ ̅ (2) the standard deviation ( ) of step changes is used to quantify variability over relevant timeframe. for the example of 5-minute data, it is calculated using the following formula (elsinga and van sark 2015): ̅ √ ̅ (3) where is the variance. considering a normal distribution of variations, 3 represent 99.7% confident interval that data (step changes) will be within. however, considering a flat tail distribution, the 99.7 confidence interval requires more than . the variability of the aggregate pv power that can be estimated is affected by correlation between different pv plants and between aggregate pv power and system load. to quantify the trends between time series measurements at two sites, correlation coefficient is used as an index. at value of 1.0, both sites have simultaneously the same trends. when the value is less than 0.25, the correlation is considered to be weak (brouwer, van den broek et al. 2014). for weak and uncorrelated variabilities, the combined variability ( ) of two sites can be found using individual site variabilities ( and ) as described by the following formula. √ (4) 2.2 variability smoothing through aggregation aggregating the output of different pv systems can reduce negative implications. the negative implication of intermittency is highly dependent on smoothing effect due to geographic diversity of these resources (mills 2010). there are different levels of smoothing effect as discussed below: 2.2.1 smoothing effecting of aggregate solar irradiance data although clouds can cause substantial changes in solar irradiance in a single site as shown in fig. 3, aggregating the output of several different solar irradiance meters reduces aggregated sites solar irradiance variability. this effect exists because clouds are diverse in nature as their sizes and shapes are changing and they do not cover different sites at the same time. the authors in jayaraman and maskell (2012) used 1-second data to study temporal and spatial variations of solar irradiance in singapore. the study demonstrated that the global horizontal radiation variability is caused by direct beam radiation variabilities and that diffused radiation variability is very limited. the authors in projects 2015 used 5 second averaged irradiation data from 9 sites located within 0-15 km radius from central location. the authors concluded that combined irradiance is substantially less variable compared with a single site. for example, the step changes of the measured irradiance at a single site exceeded 500 w/m 2 . however, when the measured irradiance at the nine sites were combined, step changes were below 140 w/m 2 and mostly below 100 w/m 2 . in addition, the authors of projects 2015 concluded that both geographical dispersion and number of sites contributes to reducing solar irradiance variability. however, spatial dispersion smoothing effect is less important than increasing the number of sites for the considered timeframe (5 seconds). in murata, yamaguchi et al. (2009), the authors demonstrate that the measured solar irradiation in japan is uncorrelated for 1-minute time-scale when the geographical distance between them is more than 50100km. for the 20-minute time-scale, the correlation coefficient quickly decreases to 0.1 as the distance increases (murata, yamaguchi et al. 2009). similarly, the authors in mills (2010) used 1-min solar irradiance data of 23 sites in the southern great plains in the usa to study aggregation effect. the study reported that single site 1-min to 180-min variability can exceed 60% of the clear sky insolation. moreover, aggregating the solar irradiance of five sites close to each other show that 99.7% of the 15-min and shorter time scales variabilities did not exceed 25% of 144 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 142-150 the expected clear sky output. additionally, the 99.7% of the 15-min and shorter time scales variabilities did not exceed 10% of the clear sky output from 100 sites with 20 km spacing, as shown in fig. 4 (mills 2010). these conclusions are based on measurements of solar irradiance data; therefore, they are considered to be conservative compared to actual aggregated pv power output variability. in perez, lauret et al. (2018), the authors studied the temporal and spatial variabilities of solar radiation in kenya using hourly solar irradiation meteosat satellite derived data of 0.05°×0.05° grids in longitudes and latitudes. the authors reported that ghi measurement in a location can be used to represent an area of 1,225 km 2 in most areas of kenya. moreover, in the central and western highlands, the area is increased to 5,625 km 2 . the authors reported that dni spatial variability is twice that of ghi. regarding inter-annual variability. the authors concluded that obtaining a representative year of solar irradiation in kenya, 5.5 and 7-year data are required for ghi and dni respectively. the authors in luiz, martins et al. (2018) used oneyear of 1-min resolution ground-based irradiance data measured at three sites to analyze the intra-day solar irradiance variability. the surface solar irradiance variability was evaluated using visible satellite images of cloud cover in different brazilian climate zones. the study concluded that generally humid months have more variability than dry months. in addition, dry locations experience more short-timescales variability than more humid locations. willemsen (2016) studied temporal, spatial and the impacts of wind characteristics of short term solar intermittency using 5-seconds data from 200 pv systems in utrecht, the netherlands. to quantify pv power output variability, the author used clearness index and delta clearness index. clearness index is defined as the ratio of the measured global horizontal irradiance to the modeled clear sky irradiance at a horizontal level at a given time. the study concluded that the variation of irradiance between sites becomes independent if the separation distance is 100 meter for 5 seconds data. figure 4. correlation of stochastic solar irradiance variabilities (mills 2010). figure 5. maximum 1-min variability of a 13.2-mw pv plant on a highly variable day (mills 2010). -0.2 0 0.2 0.4 0.6 0.8 1 0 100 200 300 400 500 c o rr e la ti o n c o e ff ic ie n t separation distance (km) 180 min 60 min 30 min 15 min 10 min 5 min 1 min 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 2 4 6 8 10 12 14 16 m a x im u m 1 -m in v a ri a b il it y (% / m in ) pv system size (mw) solar irradiation power output 145 https://www.sciencedirect.com/topics/engineering/solar-radiation https://www.sciencedirect.com/topics/engineering/solar-radiation m.h. albadi 2.2.2. smoothing effect of within a single pv plant in general, solar irradiance variabilities are more severe than power output variability (projects 2015). this smoothing effect is a function of the plant size and the timeframe. in general, larger pv plants have better smoothing effect than smaller plants (lew, brinkman et al. 2013). the authors in mills (2010) conducted measurement analysis of a multi-kw pv plant and showed that the sub-minute power output variabilities are lower than the corresponding solar irradiance variabilities. for a multi-mw power plant, similar smoothing effect is observed in sub 10-minute power output variabilities. figure 5 below presents a linear curve fitting for data presented in (mills 2010). 2.2.3. smoothing effect of aggregate pv plants output the geographic diversity of pv plants decreases the probability of cloud fronts covering different pv plants at the same time. this smoothing effect has been observed (wiemken, beyer et al. 2001). experience from pv plants performance in germany showed that normalized 5-minute output variability at one site may exceed 50%, whereas the normalized variability from 100 pv sites was within 5% of installed capacity (wiemken, beyer et al. 2001). the power output variabilities of pv plants 20 km and 50 km apart are considered uncorrelated for 15-min and 30-minute or shorter time-scales, respectively. to obtain uncorrelated 60-min power output variability, sites should be at least 150 km apart (mills 2010). elsinga and van sark (2015) used the standard deviation of 25 pv systems power output 1-minute data to quantify variability in and around utrecht, the netherlands. the authors concluded that the power output of different pv systems becomes uncorrelated after a certain decorrelation length. for the full year data of the system understudy (100 km 2 ), the decorrelation length was found to be 0.34, 2.6, and 5.0 km, for 1, 5, and 15-minute timeframes, respectively. if the outputs of an n-number of pv power plants are considered uncorrelated, the aggregated output variability is reduced by . it can be concluded that the correlation of output variabilities between different pv plants installed at different locations is a function of both a dispersion area and relevant time-scale. less correlation is achieved over shorter time scales variability and broader geographical dispersion areas. additionally, the speed of the clouds affects this correlation (lave and kleissl 2010; hoff and perez 2012). 2.2.4 smoothing effect of load and other energy resources variabilities the study in (projects 2015) considered the effect of 10 mw additional solar pv capacity on net load variability of a 55 mw islanded system having 4 mw existing pv capacity. the authors concluded that the net load variability caused by additional 10mw dispersed pv plants (9x1.1mw) was very similar to the original load variability. this result is attributed to the fact that aggregated load variabilities are not highly correlated to aggregated pv power variability. in addition, the authors in (energy 2010; halamay, brekken et al. 2011) concluded that aggregating different energy resources reduces the overall variability of the renewable-based power output. in koivisto, das et al. (2019), the authors used time series simulation of wind and solar generation to analyze the variability and uncertainty of intermittent renewable power. the study concluded that increasing intermittent renewable power capacity in future nordic and baltic countries is not expected to cause a significant increase in the hourly ramp rates of net load. the mean value of net load decreased for all studied scenarios. moreover, the standard deviation of the net load is expected to have a slight and notable increase for in 2030 and 2050 scenarios, respectively. the authors concluded that with more geographical dispersion and mix of different renewable energy technologies, the standard deviation of the net load deceases significantly. 3. impacts of solar pv power intermittency on power systems in general, the intermittency of solar pv power can negatively affect power system operation in different aspects. these effects could be either local or system wide depending on installed pv capacity, load profile, flexibility of dispatchable generation units, and system network. 3.1 power quality fluctuation solar irradiation a specific location results in fluctuating solar pv power output. this fluctuation, in turn, can cause power quality distortion such as flicker and overvoltage at that location or feeder (trindade, ferreira et al. 2017). voltage variations caused by variable pv power could be slow (steady) variations during sunny days, in which the voltage at the point of common coupling increases with the power output. in addition, fast (transient) voltage variations occur on partly cloudy days as a result of passing clouds (lew, miller et al. 2010). this voltage variations could affect the tap changers and voltage regulators (trindade, ferreira et al. 2017). in addition, cloudy conditions can increase harmonic distortion levels caused by pv systems as reported in varma, rahman et al. (2016). the degree of power quality distortion caused by pv systems is a function of the characteristics of cloud transients, the size of pv system, the stiffness of the point of common coupling, and the inverter used. 3.2 power flow and system losses the fluctuating output of solar pv systems changes the power flow; therefore, it affects system losses. 146 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 142-150 this effect is a function of the size and location of solar pv facilities. even if pv systems are allocated optimally for certain load and solar conditions, fluctuating output of pv systems will change losses. (albadi, al-hinai et al. 2013, albadi, al-mashaikhi et al. 2015). in general, power system losses are expected to decrease when pv systems are distributed and located in load centers, and vice versa. 3.3 cycling and ramping of thermal generation units high variability of pv power output could result in increased net load variability. this can increase cycling (turning on/off) of thermal-based generation units (albadi and el-saadany 2011). increased cycling results in increased wear-and-tear costs as well as increased emissions. the study in lew, brinkman et al. (2013) concluded that while the additional cycling costs caused by wind and solar power are minimal ($0.14–0.67 /mwh), the fuel cost reductions are substantial ($28–$29/mwh). in addition, the authors concluded that the increased cycling due to renewable energy caused a negligible impact on the co2, nox, and so2 emissions reductions from renewable energy. moreover, the increased ramping requirements of the net-load might result in dispatching generators that have high ramping capabilities regardless of marginal cost. this is of special importance when load ramping up/down requirements are opposite pv power variability. 3.4 operating reserve operating reserve is required to balance short-term demand fluctuations. as the penetration of intermittent pv power in a power system increases, operators would require more operating reserves to maintain supply-demand balance (halamay, brekken et al. 2011, dowell, hawker et al. 2016). additional uncertainty and variability caused by intermittent generation resources over time scales shorter than the starting time of fast-start generation units must be balanced using spinning reserve. however, adding more spinning reserve capacity results in more units operating at sub-optimal operating points with higher marginal costs (ortega-vazquez and kirschen 2007). depending on study assumptions and system operator procedures, different integration studies address reserve requirements in presence of intermittent solar pv power differently (refer to fig. 6). earlier studies considered a constant amount as a reserve requirement for the whole year based on load and/or intermittent renewable power (energy 2010, halamay, brekken et al. 2011). recent studies considered variable hourly requirement to reflect expected load and intermittent power variability (tabone, goebel et al. 2016). two approaches are used to determine the reserve requirements: n-sigma and heuristics. in n-sigma approach, reserves are defined to be equal to a given number of forecasting error standard deviation (halamay, brekken et al. 2011). if the probability distribution function of the forecasting error is gaussian, this method gives a confidence level for reserve requirements. for example, specifying the reserve to be 3-sigma ( ) implies a 99.73% confidence level. as extreme variabilities occur more than what gaussian function predict, some studies use heuristics approaches to quantify reserve requirements. for examples, the authors in lew, brinkman et al. (2013) set regulating reserves to cover 1% of the load and 95% confidence level of the 10minute load and renewable power forecast errors, respectively. load following reserve is set to cover 70% of the 60-minute renewable power forecast errors. in tabone and callaway (2015), tabone, goebel et al. (2016), markov chain modeling is used to predict variability and uncertainty of pv systems for reserve requirement calculation. figure 6. reserve requirements. 4. lessons learned from pv integration studies 4.1 extreme pv power variabilities timings reserve is of special importance during extreme events that include the steepest net load ramps, the minimum net load, and the biggest forecast errors. as pv power variability is dominated by known deterministic (diurnal) changes, its reserve requirements should be based on stochastic (weather) component of solar variability (ibanez, brinkman et al. 2012). extreme solar power variability is attributed to the known daily position of the sun in the sky (sun rise/set) not from fast-moving clouds. 4.2 pv geographical dispersion the authors of tabone, goebel et al. (2016) demonstrated that locations of utility-scale pv systems play an important role on regulation and load following reserve requirement in california’s power system considering additional 12gw pv capacity. when the pv capacity is geographically dispersed, additional reserve requirements are less than 0.05% reserve requirements amount constant variable approaches n-sigma heuristics 147 m.h. albadi and 1.2% of installed pv capacity, for regulation and load following, respectively. these requirements increase to 0.2% and 5.6% for the centralized scenarios. in tabone, goebel et al. (2016), load following reserve is defined as the difference between hourly and 5-min schedules, whereas regulation reserve is the difference between actual net-load and the 5-minute schedule. 4.3 pv power impacts on system stability it is unlikely that pv power will change contingency reserves (brouwer, van den broek et al. 2014). this conclusion stands because the loss of supply resulted from a single contingency at the largest infeed will likely continue to be larger than that resulted from a contingency at single pv plant. the authors in miller, shao et al. (2014) studied transient stability and frequency response of the western interconnection with high wind and solar penetration. the authors concluded that the western interconnection can meet transient stability and frequency regulation objectives with high levels of wind and solar generation. however, in the context of small and islanded grids, high pv power penetration level can change the required contingency reserves. for example, if the solar pv power becomes low due to a passing could, the frequency of the system will drop below frequency ride-through limit of pv inverters and they might disconnect simultaneously. this simultaneous loss of inverters can create a large loss of generation. 4.4 factors affecting intermittency costs to quantify the impact large solar pv power in a specific system, the following aspects need to be considered. ➢ characteristics of solar pv power output: pv power output is a function of both solar irradiation characteristics and geographical dispersion of pv production facilities. ➢ system load and pv power output profiles: highly correlated load and pv power profiles results in lower integration costs. storage and demand response options can reduce net load variability. ➢ conventional generation units and transmission system characteristics: flexible generation units such as natural gas-based power plants can accommodate increased net load variability better than other large units such as coal-fired or nuclear plants. ➢ electricity market design: the market design influences the reserve required to accommodate variable pv power. for example, forecasting errors and reserve requirements are lower when generation scheduling is done more frequently. 5. conclusion the wide-scale integration of solar pv power might bring concerns regarding pv output variability and its challenges to power system operation. although the solar irradiation measured at a specific site experience drastic change during cloudy days, a smoothing effect exists with aggregated measured data from multiple sites. in addition, because solar irradiation data at a specific site is based on one point of very small area, the output of a pv plant that covers hundreds of meters at that site is less variable than the measured solar irradiation. moreover, field experiment shows that the output variability of aggregated pv plants output is much smaller than that of a single pv plant. fortunately, extreme aggregate solar power variability is attributed to the known daily position of the sun in the sky not from fast-moving clouds. to quantify the system-wide impacts of pv power, reserve requirements to address intermittency are to be quantified. studies show that geographical dispersion of pv power facilities is important to reduce reserve requirements. in addition, extreme solar power variability of geographically distributed pv power is attributed to the known daily position of the sun in the sky not from fast-moving clouds. the intermittency integration costs of pv power can be determined by calculating the difference between overall electricity cost with and without pv power. this integration cost is system specific and is affected by several factors such as the characteristics of solar pv power output, correlation between load and pv power output profiles, characteristics existing generation units’ and transmission system, as well as the design of electricity market. conflict of interest the author declares no conflicts of interest. funding no funding was received for this research. references albadi m. a. al-hinai n. al-abri, y. al-busafi and r. al-sadairi (2013), "optimal allocation of solar pv systems in rural areas using genetic algorithms: a case study. international journal of sustainable engineering 6(4): 301-306. albadi m. n. al-mashaikhi, s. al-hinai, r. al-abri, a. al-hinai, q. al-aamri, a. al-mazidi and m. al-gafri (2015), loss reduction in isolated rural area distribution network using photovoltaic system. the journal of engineering research 12(2): 51-59. albadi m. and e. el-saadany (2010), impacts of wind power variability on generation costs-an overview. 148 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 142-150 j. eng. res 7(2): 24-31. albadi m. and e. el-saadany (2011), comparative study on impacts of wind profiles on thermal units scheduling costs. iet renewable power generation 5(1): 26-35. albadi m.h. (2017), electricity sector in oman after 10 years of reform: status, trends, and future perspectives. the electricity journal 30(7): 2330. brouwer a.s.m. van den broek, a. seebregts and a. faaij (2014) impacts of large-scale intermittent renewable energy sources on electricity systems, and how these can be modeled. renewable and sustainable energy reviews 33: 443-466. dowell j.g. hawker k. bell and gill s (2016), a review of probabilistic methods for defining reserve requirements. power and energy society general meeting (pesgm), 2016, ieee. elsinga b. and w. van sark (2015), spatial power fluctuation correlations in urban rooftop photovoltaic systems. progress in photovoltaics: research and applications 23(10): 1390-1397. energy g. (2010), western wind and solar integration study, citeseer. halamay d.a. t.k. brekken, a. simmons and s. mcarthur (2011), reserve requirement impacts of large-scale integration of wind, solar, and ocean wave power generation. ieee transactions on sustainable energy 2(3): 321-328. hoff t.e. and r. perez (2012), modeling pv fleet output variability. solar energy 86(8): 2177-2189. ibanez e.g. brinkman, m. hummon and d. lew (2012), a solar reserve methodology for renewable energy integration studies based on subhourly variability analysis. 2nd international workshop on integration of solar power in power systems proceedings, lisbon, portugal. jayaraman r. and d. l. maskell (2012), temporal and spatial variations of the solar radiation observed in singapore. energy procedia 25: 108117. koivisto m.k. das f. guo p. sørensen e. nuño n. cutululis and p. maule (2019), using time series simulation tool for assessing the effects of variable renewable energy generation on power and energy systems. wiley interdisciplinary reviews: energy and environment: e329. kundur p.j. paserb v. ajjarapu g. andersson a. bose c. canizares n. hatziargyriou d. hill a. stankovic and c. taylor (2004), definition and classification of power system stability ieee/cigre joint task force on stability terms and definitions. ieee transactions on power systems 19(3): 1387-1401. lave m. and j. kleissl (2010). solar variability of four sites across the state of colorado. renewable energy 35(12): 2867-2873. lew d.g. brinkman e. ibanez b. hodge and j. king (2013), the western wind and solar integration study phase 2. contract 303: 275-3000. lew d.n. miller k. clark g. jordan and z. gao (2010), impact of high solar penetration in the western interconnection.. luiz e.w.f.r. martins a.r. gonçalves and e. b. pereira (2018) analysis of intra-day solar irradiance variability in different brazilian climate zones. solar energy 167: 210-219. miller n.m. shao s. pajic and r. d’aquila (2014), western wind and solar integration study phase 3frequency response and transient stability. contract 303: 275-3000. mills a. (2010), implications of wide-area geographic diversity for short-term variability of solar power. lawrence berkeley national laboratory. mills a. (2010), understanding variability and uncertainty of photovoltaics for integration with the electric power system. lawrence berkeley national laboratory. murata a.h. yamaguchi and k. otani (2009), a method of estimating the output fluctuation of many photovoltaic power generation systems dispersed in a wide area. electrical engineering in japan 166(4): 9-19. opwp (2019), oman power and water procurement co. (saoc) website. ortega-vazquez m.a. and d.s. kirschen (2007), optimizing the spinning reserve requirements using a cost/benefit analysis. ieee transactions on power systems 22(1): 24-33. perez r. p. lauret m. perez m. david t. e. hoff and s. kivalov (2018), solar resource variability. wind field and solar radiation characterization and forecasting: a numerical approach for complex terrain: 149-170. projects c (2015), investigating the impact of solar variability on grid stability, australian renewable energy agency ren21 (2019), renewables 2019 global status report. paris, ren21 secretariat. sengupta m.y. xie a. lopez a. habte g. maclaurin and j. shelby (2018), the national solar radiation data base (nsrdb). renewable and sustainable energy reviews 89: 51-60. tabone m. d. and d. s. callaway (2015), modeling variability and uncertainty of photovoltaic generation: a hidden state spatial statistical approach. ieee transactions on power systems 30(6): 2965-2973. tabone m.d. c. goebel and d. s. callaway (2016), the effect of pv siting on power system flexibility needs. solar energy 139: 776-786. trindade f.c.l. t. s. d. ferreira m. g. lopes and w. freitas (2017), mitigation of fast voltage variations during cloud transients in distribution systems with pv solar farms. ieee transactions on power delivery 32(2): 921-932. ummels b.c. m. gibescu e. pelgrum w.l. kling and a. j. brand (2007), impacts of wind power on thermal generation unit commitment and dispatch. ieee transactions on energy conversion 22(1): 149 m.h. albadi 44-51. varma r.k. s. a. rahman t. vanderheide and m. d. n. dang (2016), harmonic impact of a 20-mw pv solar farm on a utility distribution network. ieee power and energy technology systems journal 3(3): 89-98. wiemken e.h. beyer w. heydenreich and k. kiefer (2001), power characteristics of pv ensembles: experiences from the combined power production of 100 grid connected pv systems distributed over the area of germany. solar energy 70(6): 513518. willemsen e. (2016), spatial and temporal variability of solar energy. master thesis, utrecht university, holland. 150 comp030304.qxd the journal of engineering research vol. 2, no. 1 (2005) 77-89 1. introduction with the era of low cost commodity hardware, clusters of workstations (cow) are emerging as platforms for parallel and distributed computing environments. computers connected through lan or wan form an infrastructure grid computing foster and kesselman, (1997). they form a high computing power for massive parallel processing that can be accessed by wide spectrum of application programmers. but the communication cost over the available networking facilities is still very high compared to computing cost. besides, it lacks the necessary reliability found in typical multi-processor interconnection __________________________________________ *corresponding author e-mail: arafeh@squ.edu.om networks. to overcome this deficiency, synchronous message-passing communication may need to be enforced for many parallel applications or middleware software. in general, synchronous communication adds overhead to the already high cost of communication. furthermore, it may develop deadlock problems among the communicating tasks. regardless of any constraints, parallel programs must be efficiently partitioned and scheduled on a cow to achieve any perceivable gain in performance. scheduling tasks, efficiently, on distributed memory architectures is still a challenging problem. a multi-step scheduling approach has been proposed by many researchers, such as (sarkar, 1989; liou and palis, 1997), in order to reduce the complexity of the scheduling problem by using low cost heuristics. in this work, we adopt a a multi-step approach for scheduling tasks with synchronization on clusters of computers b.r. arafeh* department of computer science, college of science, sultan qaboos university, p.o. box 36. al khod 123, muscat, sultanate of oman received 3 march 2004; accepted 10 october 2004 abstract: in this work, a two-step approach is adopted for scheduling tasks with synchronous inter-task communication. to that end, an efficient algorithm, called glb-synch, is introduced for mapping clusters and ordering tasks on processors. the algorithm used the information obtained during the clustering step for selecting a cluster to be mapped on the least loaded processor. a performance study has been conducted on the glb-synch algorithm by simulation. a multi-step scheduling setup has been performed based on a previously developed algorithm for clustering dags with synchronous communication, called nlc-synchcom, and using synthesized dags. we have shown by analysis and experimentation that the glb-synch algorithm retains the same low complexity cost of the first step for clustering. the performance results highlight the drawback of synchronization on speedup scalability. keywords: multi-step scheduling, clustering, mapping and ordering, synchronous message passing, duistributed-memory systems g¢¢üü∏∏îîàà°°ùÿÿº°sg á«∏y ≥∏£j ,∫é©a »erqgƒn êé¡æe ¢vô©f ±~¡dg gòg áe~ÿ .øegõàÿg ∫é°üj’g ägp á÷é©ÿg ä骡eádh~÷ újƒ£n øe ¿ƒµe üƒ∏°scg ≈æñàj åëñdg gòg : glb-synch∫ón é¡«∏y ∫ƒ°ü◊g ” »àdg äéeƒ∏©ÿg á«erqgƒÿg ω~îà°ùj .é¡ñ«jôjh á«hƒ°sé◊g äé÷é©ÿg ≈∏y á÷é©ÿg ä骡e øe äéyƒª› ™jrƒj §«£îj ™°vƒd , á«erqgƒÿg agog ≈∏y á°sgqo âjôlg ~≤d .óªm πb’g èdé©ÿéh é¡£hqh ä骡e áyƒª› qé«àn’ ä骡ÿg ™«ªœ iƒ£nglb-synch~≤dh .iécéùg üƒ∏°scg ωg~îà°séh ájôfgoódgh á¡lƒÿg ä飣ıg ™«ªœ á«∏ª©d é≤hé°s iqƒ£e á«erqgƒn ≈∏y gaéæh ägƒ£ÿg o~©àe ádh~l ωé¶f ôjƒ£j ”(dag)∫é°üj’g qéñày’g ‘ ònéj »àdgh , á«erqgƒÿg ¿céh öjôéàdgh π«∏ëàdg ᣰsgƒh á°sgq~dg äô¡xcg .á÷é©ÿg ä骡e ä飣fl øjƒµàd »ñ«côj üƒ∏°sg á°sgq~dg âe~îà°sg å«m .á÷é©ÿg ä骡e úh øegõàÿg glb-synchá°sgqo èféàf úñj éªc .™«ªéàdg á«∏ª©h á°uéÿg ¤h’g iƒ£ÿg ‘ á«∏y ∫ƒ°ü◊g ” …òdg ∂dòd πké‡ ~«≤©àdg iaéøc øe ê~àe iƒà°ùã ßøà– .á÷é©ÿg ä骡e ´qé°ùj ∫é› ‘ ™°sƒàdg iq~b ™aq á«féµeg ≈∏y á÷é©ÿg ä骡e úh áæegõàe ∫é°üjg äé«∏ªy oƒlƒd á«ñ∏°ùdg öfgƒ÷g ago’g gáá««mmééààøøÿÿgg ääggooôôøøÿÿ.áyrƒÿg iôcgòdg ägp º¶ædg ,áæegõàe á≤jô£h πfé°sôdg πbéæj ,ö«jîdgh º«°sîdg ,™«ªéàdg á«∏ªy ,ägƒ£ÿg io~©àe ádh~l : á«hƒ°sé◊g ä驪éàdg ≈∏y áæegõàÿg á÷é©ÿg ä骡e ádh~÷ ägƒ£ÿg io~©àe á≤jôw 78 the journal of engineering research vol. 2, no. 1 (2005) 77-89 two-step approach for scheduling synchronous parallel programs on distributed-memory architectures. in the first step, tasks of the program are clustered to reduce the communication cost and to avoid deadlocks, assuming unbounded number of processors. in the second step, clusters are mapped and their tasks are ordered for execution on the available number of processors. the main contribution of this work is an algorithm, called guided load balancing with synchronization (glb-synch). it is a basis for mapping and ordering tasks on processors with synchronous communication. the glb-synch algorithm is based on the glb algorithm for cluster mapping that was introduced by radulescu (2001). however, the glb-synch algorithm performs both cluster-mapping and task-ordering, and retains the low complexity of the unbounded number of processors (unc) schedule applied in the first step. the rest of the paper is organized as follows. the next section introduces a background and some definitions related to this paper and the related work. section 3 presents the glb-synch algorithm, while section 4 presents the performance study. finally, section 5 is the conclusion. 2. background and related work 2.1 preliminaries the execution behavior of the program dag is the macro-dataflow model. however, the execution of each task consists of three phases: receive, compute and send. the receive phase includes receiving all messages required by the task for its execution to start. the compute phase is the phase in which the instructions of the task are executed without interruption. we assume a synchronous communication protocol. the send phase includes sending all messages to all dependent tasks in parallel. however, the sender task is blocked, waiting for acknowledgements, until all receiving tasks actually receive the messages. 2.2 multi-step scheduling the multi-step scheduling process can be achieved by two or three steps. first, clustering of tasks without duplication can be performed, assuming unbounded number of processors. second, clusters are mapped on the available processors. third, the tasks of the mapped clusters are ordered for execution on the processors. the clustering problem has been shown to be np-complete (papadimitriou and yannakakis, 1999; sarkar, 1989). polynomial-time heuristic algorithms have been proposed for the clustering problem based on the critical path analysis (sarkar, 1989; wu and gajski, 1990; gerlasoulis and yang, 1993; kwok and ahmad, 1999; kadamuddi and tsai, 2000; shirazi et al. 1990; lee et al. 2003). an overview of cluster-mapping is given next. 2.2.1 mapping clusters to processors cluster-mapping is needed when the number of available processors is less than the number of clusters. there are a number of approaches that are proposed in literature. however, the issue of mapping clusters to processors has not been given enough attention in the literature, and there is much room to explore on this topic. in the next paragraphs, we discuss some of the approaches reported in the literature. sarkar (1989) used a list-scheduling based method to map the clusters to processors, called list cluster assignment (lca). it is an incremental algorithm that performs both cluster-mapping and task-ordering in a single step. kim and browne (1988) proposed a mapping scheme for clusters based on their linear clustering algorithm. the clusters are first merged in order to reduce their number to be at most equal to the number of processors. then the process is followed by heuristics to optimize the mapping step. mainly, the heuristics would choose a processor which has the most appropriate number of channels among currently unallocated processors. wu and gajski (1990) proposed a mapping scheme for clusters based on a dedicated traffic scheduling algorithm that balances the network traffic. the algorithm used generates an initial assignment by a constructive method; then, the assignment is iteratively improved to obtain a better mapping. the heuristic is based on minimizing the total communication traffic. yang and gerasoulis (1994) employed a work profiling method for merging clusters, called wrap cluster merging (wcm) algorithm. first, clusters are sorted in an increasing order of aggregate computational load. then, a load balancing algorithm is invoked to map the clusters to the processors, so that every processor has about the same load. the work of liou and palis (1997) investigated the in this work, a parallel program is modeled as a weighted directed acyclic graph (dag), g = (v, e, ω, λ), where v is the set of task nodes, e is the set of communication edges, ω is the set of task computation weights, and λ is the set of edge communi cation costs. an edge eij = (ni, nj) ∈ e represents a data dependence constraint between the two tasks ni and nj, where the execution of nj must start after receiving all input from ni. the communication cost of message passing along an edge eij is denoted by cij = λ(eij), and the computation weight of a task ni is denoted by ω(ni). we will refer to the source and destination nodes of an edge by the parent node and the child node, respectively. a node that does not have any parent is called an entry node, while a node, which does not have any child, is called an exit node. pred(ni) is the set of immediate predecessors of ni, and succ(ni) is the set of immediate successors of ni. the length of a path is defined as the sum of all computation weights of nodes and all communication costs of edges along the path. the critical path of a dag is the path from an entry node to an exit node that has the maximum length. the computation to communication ratio of a parallel program ( pccr) is defined as its average computation weight divided by its average communication cost. 79 the journal of engineering research vol. 2, no. 1 (2005) 77-89 problem of mapping clusters to processors. they have shown the effectiveness of using the two-phase scheduling approach, in which the task clustering is followed by the cluster-mapping step, over the one-phase scheduling. they proposed a clustering algorithm, called cass-ii (clustering and scheduling system ii), and introduced three algorithms for cluster-mapping schemes, namely, the lb (load-balancing) algorithm, ctm (communication traffic minimizing) algorithm, and the rand (random) algorithm. they applied randomly generated task graphs in an experimental study using their clustering algorithm and cluster-mapping schemes. their work shows that, when task clustering is performed before cluster-mapping, load balancing is the preferred approach for merging clusters. compared to ctm, lb is fast, easy to implement and produces significantly better schedules. radulescu (2001) proposed two algorithms for mapping clusters to processors in a multi-step scheduling approach. both algorithms aim at achieving a better costperformance ratio. the first algorithm, called guided load balancing (glb), exploits knowledge about the task start times that were computed in the clustering step. accordingly, clusters are mapped in the order of their start times to the least loaded processor at that time. the second algorithm, called list load balancing (llb), aims at improving the load balancing throughout the program execution time by performing cluster-mapping and taskordering in one step. there are two benefits reported for this integration. first, it allows dynamic load balancing through the execution of the mapping process, because only the ready tasks are considered in the mapping process. second, it considers communication costs, when selecting tasks for mapping, as opposed to other clustermapping algorithms, such as wcm and glb, which do not. the work by lee et al. (2003), introduced a multi-step scheduling approach using a block dependency dag, that represents the execution behavior of block sparse cholesky factorization operation in a distributed-memory system. the proposed scheduling algorithm consists of two stages. in the first stage, a clustering algorithm, called early-start clustering (esc), is used to cluster tasks while preserving the earliest start time of a task without limiting the potential degree of parallelism, and without considering the number of available processors. in the second stage, a cluster mapping algorithm, called affine cluster mapping (acm), is used to allocate clusters to a given number of processors. the acm algorithm attempts to reduce the communication overhead and balance the workloads among the processors based on two criteria. these are the affinity of a cluster with respect to a processor, in terms of the sum of communication costs required when the cluster is mapped to other processors, and the amount of workload required for a cluster. the work by lee et al. (2003) shows by experiments the effectiveness of applying the proposed scheduling algorithm, compared to other processor mapping methods that are used for parallelizing the sparse cholesky factorization operation. the experiments were conducted on a myrinet cluster system and using benchmark sparse matrices 2.2.2 scheduling with synchronous communication most scheduling algorithms for distributed-memory parallel architectures assume the use of an asynchronous communication protocol for a message-passing system. however, parallel computing on a network of workstations or over the internet is not as reliable as that performed on parallel machines. therefore, the requirement for synchronization at the application level becomes eminent for many software systems. in a synchronous communication, the sender is blocked until an acknowledgement is received from the receiver. this waiting time is called the blocking delay. a deadlock occurs when a sender gets blocked indefinitely, waiting for an acknowledgment from a receiver task in another cluster. at the same time, the receiver task cannot start execution, because one of its predecessor tasks has been blocked indefinitely, waiting for an acknowledgment from a task in some other cluster. a direct deadlock situation between two clusters occurs due to a cyclic dependency relation between them. in general, a deadlock situation may arise due to a chain of dependency relations among a subset of tasks. in this work, we consider direct deadlock situations only. based on the task execution phases, the following definitions of time parameters characterize the scheduling of a task node in a scheduled dag with synchronous communication. the issue of scheduling on distributed-memory parallel architectures with synchronous communication has not been given enough attention in literature. however, the works of kadamuddi and tsia (2000) and arafeh (2003) address this issue, assuming a multi-step scheduling approach. both propose clustering algorithms for tasks with synchronous communication, in which deadlocks are detected and avoided as part of the task clustering step. the work presented in this paper uses the clustering algorithm, called nlc-synchcom, by arafeh (2003) for the task-clustering step. the algorithm proceeds in one pass in the forward direction from entry nodes to exit nodes, one level at a time. the algorithm starts assuming • s_receive(ni): start time for receiving messages by task node ni • e_receive (ni): end time for receiving messages by task node ni • s_compute(ni): start time for computation by task node ni • e_compute(ni): end time for computation by task node ni • s_send(ni): start time for sending messages by task node ni. • e_send(ni): end time for sending messages by task node ni. 80 the journal of engineering research vol. 2, no. 1 (2005) 77-89 each task node is in a cluster by itself. therefore, there are |v| clusters at the beginning of the algorithm. each node in a cluster is designated by its status as a head, tail, regular or singleton. a node in a cluster by itself is given the status of a singleton node. the head task of a cluster is the one that must be scheduled first due to its precedence with respect to all other tasks in the cluster. similarly, the tail task of a cluster is the one that must be scheduled last due to its precedence with respect to all other tasks in the cluster. a regular task in a cluster is one that is not a singleton, head or tail. the selection of a parent node, ni, at level l that will be merged with one of its child nodes is determined using two priority schemes. the first scheme is used to determine the priority of a parent node at level l for merging. it is defined by the parent's completion time, e_send(parent), in descending order. the second scheme is used to determine the priority of merging a child node nj for a parent ni. this priority depends on the maximum remaining time left to the completion of execution from a parent node, ni, to an exit node, excluding ni. since a merging step would zero a (parent, child) edge, then the parent node and all its descendant nodes may have their completion times changed accordingly. thus, the priorities of the parent nodes at each level are found dynamically before the nodes of that level are scanned for merging. on the other hand, the remaining time to the completion of execution for each node is computed at the initialization time only. a priority is given to merge the parent node with a child node which leads to the highest remaining time to completion. a selected child node nj has to pass two tests before merging can be finally applied. these are the deadlock detection and the merging check tests. first, the deadlock detection test ensures that a merging step of a child with its parent's cluster would not cause a deadlock case for the parent's cluster with any other existing cluster in the dag. second, the merging check test ensures that a merging step of a child node with its parent's cluster would not cause an increase in the application's execution time. this execution time is referred to in this paper by the dag parallel time, pt. the (parent, child) edge is zeroed and merging is performed, only, if both tests are passed successfully. the complexity cost of the nlc-synchcom algorithm is o(v(log v + e2)), where v is the number of nodes and e is the number of edges in a dag. for further details, see the paper by arafeh (2003). 3. cluster-mapping and scheduling with synchronous communication 3.1 description of the glb-synch algorithm the glb-synch algorithm is an extension of radulescu glb algorithm for mapping clusters to processors (radulescu, 2001). however, the glb-synch algorithm performs both cluster-mapping and task-ordering in the context of synchronous communication. the algorithm uses the information obtained during the clustering step, based on the nlc-synchcom algorithm, for mapping clusters to a distributed-memory system of unbounded number of homogeneous processors that are fully connected. eventually, the nlc-synchcom algorithm schedules the tasks on the virtual processors, assuming each cluster is allocated to one virtual processor and there is unbounded number of processors or clusters (unc). therefore, we will refer to the dag and schedule generated by the nlc-synchcom algorithm by the clustered dag and the unc schedule, respectively. each node of the clustered dag represents a cluster or virtual processor, and each directed edge between two clusters represents a communication link connecting them. the glbsynch algorithm uses the cluster start time, ts(c), to represent the priority of a cluster, c, for mapping. the start time of a cluster is the start time for a computation phase of the header task in the cluster, and it is given by (1) where t is a task of cluster c. similar to glb, the cluster, which has the earliest start time, is mapped first. in case of a tie, the cluster with the highest workload is mapped first. if there is still a tie, a cluster is selected randomly. based on the execution phases of a task, the unc schedule allocates time slots for all the phases, with no consideration for overlapping computation with communication. at this stage, it is more natural to consider the existence of some overlapping between the computation and the communication phases, as clusters are mapped to physical processors. in this work, we assume that the duration of the receive phase is implicitly handled by a communication processor, and the acknowledgment of a received message is handled by the messagepassing system. each scheduled task should have all its expected messages to be received declared to the message-passing system ahead of the start time of its receive phase. only when all the needed messages have arrived, the task can start the computation phase. however, the duration of the sending phase is still considered explicitly as part of the task's schedule on a processor to enforce synchronization. the glb-synch algorithm maps clusters to a distributed-memory system of bounded number of homogeneous processors that are fully connected. since a complete interconnection network is assumed, there is no consideration for bandwidth contention. furthermore, each processor is assumed to have unlimited number of communication ports and unlimited memory space. since the workload of a task, t, includes the duration starting from s_compute(t) till e_send(t), then the workload of a cluster is defined as (2) because a cluster is not a schedulable unit, the glbsynch algorithm maps a cluster to the least loaded proces , )}(_{)( tcomputesminct ct s ∈ = ))()(()( ∑ ∈ −= ct w ts_computete_sendct 81 the journal of engineering research vol. 2, no. 1 (2005) 77-89 sor as in the glb algorithm. it is expected that clusters mapped to the same processor to be interleaved due to the task-ordering step. the workload of a processor, p, is defined as (3) where, ψ(p) is the subset of clusters mapped to processor p. as a consequence, all inter-cluster communication costs among the mapped subset of clusters must become zero. the algorithm achieves the objective of scheduling tasks on the processors in three steps. in the first step, it assumes that the process of cluster-mapping is generating super-clusters, constructed as aggregates of the mapped clusters to the target processors. in this step, all the time parameters of all tasks are recomputed, due to the zeroing of the inter-cluster communication costs, without performing the task-ordering (i.e. sequentialization) process. it may look as if each aggregate of clusters (i.e. virtual processors) is mapped to a shared-memory multiprocessor. in the second step, the glb-synch algorithm performs task-ordering, based on start times of tasks for the computation phase. all tasks allocated to the same processor are sorted topologically in an increasing order of their start computation time, s_compute. if two tasks have the same s_compute time, the task with the highest blocking delay is scheduled first. the blocking delay of a task, t, is defined as the time it is waiting for acknowledgments for all messages sent by that task, and it is computed by (e_send(t) s_send(t)). if there is still a tie, a task is selected randomly. in the third step, the mapped tasks to the same processor are scheduled using their precedence order. let tir (p) denote the processor ready time on a partial schedule. it is initialized by zero, and it is defined as the end time of the last task, ti, scheduled on that processor. accordingly, a task ti+1, is scheduled for execution at tir (p), if the current start time for the computation phase, s_compute(ti+1) is less than or equal to tir (p). otherwise, the task ti+1 is scheduled to start execution at its designated s_compute time. the glb-synch algorithm is described text. glb-synch algorithm input: 1. a clustered dag 2. the table of time parameters (i.e. unc schedule) 3. number of processors. output: 1. a mapping of the clusters to the processors 2. tasks schedule on the processors. algorithm steps: 1. compute the start time, ts(c), and the workload, tw(c) for each cluster, c. 2. sort the clusters in an increasing order based on ts, breaking ties by choosing the cluster with the highest workload. if there is still a tie, select one randomly. 3. for each cluster, c, do * map c to a processor, p, with the least workload. * zero the inter-cluster communication cost between any currently mapped clusters to p and c. * update the workload of processor p: (4) 4. update the unc schedule due to the mapping step. 5. for each processor, p, do * perform task-ordering for all tasks mapped to p based on s_compute time, breaking ties by choosing the task with the highest blocking delay. if there is still a tie, select one randomly. * for each task ti mapped to a processor (5) else (6) 3.2 complexity analysis the following notations are used to characterize the time complexity of the glb-synch algorithm. c: the number of clusters. e: the number of dag edges, |e|. v: the number of dag vertices, |v|. m: the number of processors. the glb-synch algorithm performs cluster-mapping, task-ordering and scheduling on the processors, based on the clustering step for synchronous communication. the glb-synch algorithm assumes that the formulation of clusters, along with the computation of their workloads, have been performed in the clustering step. besides, it assumes that task clusters are deadlock-free, since all generated clusters by the clustering step had passed the deadlock detection test successfully. theorem 1. the time complexity of the glb-synch algorithm is o(mc + mv logv + ev). proof. the start time of a cluster, ts(c), is the s_compute if ))(_)(( 1 i i r tcomputespt ≤ − then schedule ti to start at s_compute(t i); ∑= ψ∈ )( )()( pc ww ctpt (c) t(p) t(p) t www += );(t t) (t s_compute(p) t iwi i r += )(t t(p) t(p) t iw i r i r += −1 schedule t i to start at )( 1 pt ir − ; 82 the journal of engineering research vol. 2, no. 1 (2005) 77-89 time of the head task of c. then, the cost of the first step is o(c). in step 2, clusters are sorted in o(c logc) time. the determination of the least loaded processor, pj, takes o(m) steps. while, the process of zeroing the inter-cluster communication costs between the already mapped clusters to pj and the current one to be mapped needs o(e). accordingly, step 3 of the algorithm takes o(c(m+e)) time. the updating of the unc schedule at step 4 takes o(ev) steps, due to the need to find the e_receive(ti) and the e_send(ti) time for each task ti. finally, step 5 of the algorithm takes o(mv logv) steps. because, the algorithm orders all tasks mapped to a processor in o(v logv) steps, then schedules tasks on a processor in o(v) steps. since m < c and c < v, the total time complexity of the glbsynch algorithm is o(pc + mv logv + ev). 4. performance study a performance study has been conducted on the multistep approach for scheduling tasks with synchronous inter-task communication. the study is based on the simulation of the multi-step scheduling approach using the nlc-synchcom algorithm for the clustering step, and the glb-synch algorithm for the cluster-mapping and taskordering steps. the performance study had adopted randomly generated dags for experimentation. synthesized random dags are generated so that the results would not be biased towards regular graph structures or certain graph shapes, allowing various dag characteristics to be considered. the objectives of the performance study include assessing the cost of the multi-step scheduling approach in the context of synchronous communication, evaluating the outcome of the multi-step scheduler, and discovering the points of deficiencies and limitations. in this section, the definitions of the chosen performance metrics in the study are given next. then, the simulation set-up for experimentation is described. finally, the performance results are presented and discussed. 4.1 performance metrics the performance metric considered for assessing the cost of each step of the multi-step scheduler is the execution time. the performance metrics considered for evaluating the outcome of the scheduling steps are based on the schedule length (sl). we will refer to slo as the original schedule length of an application dag. it is the parallel time for executing a dag on an unbounded number of processors. slo is equal to the length of the critical path of the dag. the schedule length obtained from executing a dag on a uniprocessor is referred to by sl1, and it is defined as (7) the schedule length obtained by the clustering step (i.e. the unc schedule) is referred to by slc. it is also based on an unbounded number of processors. the schedule length obtained by mapping and scheduling the tasks on bounded number of processors, m, is referred to by slm. the speedup factor, spm, is defined as the ratio of executing a parallel program on a uniprocessor to its execution on m processors, and it is given by (8) definition 1 normalized schedule length nsl(m). in this work, we define the normalized schedule length, nsl(m), as the ratio of the schedule length of a dag on m processors, slm, to its original schedule length, slo. that is, (9) definition 2 utilization u(m). the utilization of a system of m processors, u(m), is defined as the percentage of the m processors' time that is kept busy during the execution of a parallel program due to a certain allocation hwang, (1993). let b(vi , pj) be the blocking time of task vi due to synchronization, when it is executed on pj. hence, (10) where, x(vi , pj) is a 0-1 function defined as follows: (11) and (12) definition 3 efficiency e(m). the computing efficiency, e(m), or the efficiency for short, is defined as the actual percentage of the computing time performed by the m processors during the execution of a parallel program. the efficiency indicates the actual degree of speedup achieved on m processors compared with the maximum value hwang, (1993). it is given by, (13) definition 4 synchronization overhead ratio sor(m). let bm be the average blocking (i.e. synchronization time) for scheduling tasks of a parallel program on bounded number of processors, m. accordingly, the synchronization overhead ratio, sor(m), is defined as the ratio of the average blocking time per processor to the schedule length, slm, where, ∑= ∈viv ivsl )(1 ω m m sl sl sp 1= o m sl sl mnsl =)( m viv pjp jiiji slm pvbvpvx mu × ∑ ∑ + = ∈ ∈ )),()()(,( )( ω otherwise p processor to allocated is vifpvx jiji 0 1),( = = ),(_),(),( jijiji pvsendspve_sendpvb −= m sp slm sl slm v me m mm viv i = × = × ∑ = ∈ 1 )( )( ω 83 the journal of engineering research vol. 2, no. 1 (2005) 77-89 (14) (15) definition 5 load imbalance factor lif(p). let tw(pj) refer to the workload of processor pj. therefore, the load imbalance factor, lif(p), is defined as the percentage of the average processor's time that is kept idle during the execution of a parallel program. actually, it indicates the percentage of non-utilized processor's time, and it is given by (16) from definitions 2, 3 and 4, the utilization is characterized by the following lemma. lemma 1. the utilization u(m) is characterized in terms of the efficiency, e(m), and the synchronization overhead ratio, sor(m), for scheduling a parallel program on m processors as (17) proof. let bave be the average blocking time in a dag, and bm the average blocking per processor for scheduling the dag on m processors. accordingly, the utilization can be rewritten as (18) (19) 4.2 simulation the use of randomly generated directed acyclic graphs (dags) to model parallel applications is a common practice in the evaluation of proposed scheduling heuristics for parallel and distributed computing systems. the use of simulation provides a basis to evaluate the scheduling algorithm independent of the hardware implementation and its organization. many approaches have been proposed in the literature on how to generate synthesized dags randomly (kasahara laboratory, japan, 2004). in this work, a random graph generator is implemented to generate weighted dags, as defined in this paper, with various characteristics based on a method that uses the following factors: 1. the number of tasks, v. 2. the shape factor, α., of a dag: we assume the height ( i.e. number of levels) of a dag is randomly generated from a uniform distribution with a mean value, lmean, equal to v1/2/α. similarly, the width for each level in the dag is ra ndomly generated from a uniform distribution with a mean value, wmean, equal to α v1/2. 3. the maximum out -degree of a node in a dag. 4. the maximum span of an edge in a dag. 5. the computation to communication ratio, ccr: it is taken as the ratio of the average computation weight to the average communication cost. values of ccr in the range 0.1 -0.7 represent fine granularity, values in the range 0.8 -1.4 represent medium granularity, and values greater that 1.4 represent coarse granularity. 6. the mean computation weig ht, ωmean: the computation weight of each node is determined randomly from a uniform distribution with a mean ωmean. 7. the mean communication cost, λmean: the mean communication cost of a dag is equal to ωmean/ccr. each communication cost of an edge is determine d randomly from a uniform distribution with a mean λmean. for the purpose of generating random dags in this study, we have arbitrarily chosen ωmean to be 20. three values of ccr are considered. 0.5, 1.0 and 5.0 , to represent fine, medium and coarse granularity respectively. to contr ol the structure of the dag, we have limited the outdegree of a node to be within the range 1-wmean. also, the span of an edge is limited to be within 1 ⎡0.25 x height⎤. three values for the shape factor, α, are considered. these are 0.5, 1.0, and 2.0. an α < 1.0 represents a da g with a long height and low degree of parallelism; while an α > 1.0 represents a shorter dag with high degree of parallelism. the width of a dag level l indicates the degree of parallelism at that level. therefore, the mean width of a dag, wmean, is adopted as a measure of the potential for the degree of parallelism in the dag. the relationship between the mean width, wmean of a dag and the number of nodes, v, of the dag for constant values of α are shown in fig. 1. the plots shown in fig. 1 indicate the potential speedup expected by scheduling a dag on m processors. accordingly, values of m >> wmean are not expected to have any significant speedup improvements in the pt of a dag. and m pvb b pp vv ji m j i , ),(∑ ∑ ∈ ∈ = m m sl b msor =)( )(1 ))(( )( mu slm ptsl mlif m pjp jwm −= × − = ∑ ∈ )()()( msormemu += m viv pjp jiji m viv pjp iji m viv pjp jiiji slm pvbpvx slm vpvx slm pvbvpvx mu × ∑ ∑ + × ∑ ∑ = × ∑ ∑ + = ∈ ∈ ∈ ∈ ∈ ∈ ),(),( )(),( )),()()(,( )( ω ω )()( )()()( msorme sl b me slm bv memu m m m ave += += × × += 84 the journal of engineering research vol. 2, no. 1 (2005) 77-89 simulation experiments were conducted to measure the cost of each step of the multi-step scheduling techniques used in this work. ten groups of synthesized dag sizes were generated. these range from 50 to 500 nodes with an increment of 50. for each dag size, v, and shape factor, α, we have generated 100 random dags. the nlcsynchcom algorithm is applied on each generated dag for clustering tasks with synchronous communication. the resultant clustered dag is taken as an input to the glb-synch algorithm for mapping and ordering the tasks on the processors. 4.3 performance results the cost of each step in our multi-step scheduling scheme is measured against the number of nodes in a dag. figures 2 and 3 show the average execution time for the clustering step, and the mapping and ordering step, respectively, versus the number of nodes. the average execution time for the mapping and ordering step is taken over all the number of processors considered in the simulation runs. three cases are considered for each scheduling step, based on the shape factor of the dag. for both steps, the execution time increases as the dag size and the degree of parallelism (i.e. α) in the dag increase. from the plots, it can be deduced that the cost of the mapping and ordering step is much less than the cost of clustering, in general. the cost of mapping and ordering does not exceed 25% of the cost of clustering for dag sizes greater than 100 nodes. in order to assess the performance of the glb-synch algorithm, we have to focus a little on the main performance features of the nlc-synchcom algorithm. the main objective of performing the clustering step is to reduce the communication cost in the dag by merging tasks onto clusters, where each cluster can be assigned to a processor. in this way, the clustered dag would have better pccr relative to the initial value of the dag's pccr. the results, shown in figs. 4 and 5, indicate the role of the 0 10 20 30 40 50 0 100 200 300 400 500 600 number of nodes m ea n w id th o f a d a g α=0.5 α=1.0 α=2.0 figure 1. relationship between the mean width and the number of nodes in dag 0 30 60 90 120 150 0 100 200 300 400 500 600 number of nodes ex ec ut io n ti m e of c lu st er in g in m se c α=0.5 α=1.0 α=2.0 figure 2. the average execution time of the clustering step versus the number of nodes 0 2 4 6 8 10 12 0 100 200 300 400 500 600 number of nodes ex ec ut io n ti m e of g lb _s yn ch in m se c α=0.5 α=1.0 α=2.0 figure 3. the average execution time of the mapping and ordering step versus the number of nodes pccr = 1.0 0 20 40 60 80 100 0 100 200 300 400 500 600 number of nodes % o f p c c r im pr ov em en t α=0.5 α=1.0 α=2.0 figure 4. the percentage improvement in pccrdue to the clustering step versus the number of nodes pccr = 1.0 0 10 20 30 40 50 0 100 200 300 400 500 600 number of nodes % r ed uc tio n in c om m c os t α=0.5 α=1.0 α=2.0 figure 5. the percentage reduction in communication cost due to the clustering step versus the number of nodes 85 the journal of engineering research vol. 2, no. 1 (2005) 77-89 pccr = 0.5 0 1 2 3 4 0 5 10 15 20 25 30 35 number of processors sp ee du p v=50 v=250 v=500 (a) pccr = 1.0 0 1 2 3 4 5 6 0 5 10 15 20 25 30 35 number of processors s p ee du p v=50 v=250 v=500 (b) pccr = 5.0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 number of processors s pe ed up v=50 v=250 v=500 (c) figure 7. the average speedup factor versus the number of processors pccr = 0.5 0 4 8 12 16 20 0 5 10 15 20 25 30 35 number of processors n s l v=50 v=250 v=500 (a) pccr = 5.0 0 4 8 12 16 20 0 5 10 15 20 25 30 35 number of processors n s l v=50 v=250 v=500 (c) pccr = 1.0 0 4 8 12 16 20 0 5 10 15 20 25 30 35 number of processors n s l v=50 v=250 v=500 (b) figure 6. the normalized scheduling length of dags versus the number of processors 86 the journal of engineering research vol. 2, no. 1 (2005) 77-89 pccr = 0.5 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors u til iz at io n v=50 v=250 v=500 (a) pccr = 1.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors u til iz at io n v=50 v=250 v=500 (b) pccr = 5.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors u til iz at io n v=50 v=250 v=500 (c) figure 8. the utilization versus the number of processors pccr = 0.5 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors c om pu tin g e ffi nc ie nc y v=50 v=250 v=500 (a) pccr = 1.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors c om pu tin g e ffi ci e nc y v=50 v=250 v=500 (b) pccr = 5.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors c om pu tin g e ffi ci e nc y v=50 v=250 v=500 (c) figure 9. the computing efficiency versus the number of processors 87 the journal of engineering research vol. 2, no. 1 (2005) 77-89 clustering step in improving pccr and reducing the communication cost, respectively, for dags with pccr=1.0. both measures tend to degrade at higher dag sizes and shape factors, as shown in figs. 4 and 5. however, the nlc-synchcom algorithm imposes restrictions on merging nodes to clusters that may cause an increase in the dag's pt. therefore, the unc schedule length, slc, is always less than or equal to slo. for example, the parallel time is reduced by about 9% for dags with a size of 50 nodes, pccr=1 and α = 0.5. the ability of the algorithm for pt reduction degrades, generally, with higher dag sizes. for example, the pt reduction is about 1%, for dags with a size of 500 nodes, pccr = 1.0 and α = 0.5. this brief characterization of the clustering step should support understanding the outcomes of the mapping and ordering step (for more details see arafeh, 2003). the performance results are shown in figs. 6-10. they are taken for dags with α =1.0 and sizes of 50, 250 and 500 nodes. this should provide us with bases of uniformity in our comparisons, analyses and assessments. this does not mean that the effect of the shape factor has been ignored, or that it does not have a role on the type of results generated. on the contrary, all performance results in this work scale proportionally with the value of α. figures 6-10 show the relationships of the normalized scheduling length, nsl, the speedup, sp, the utilization, u, the efficiency, e, and the load imbalance factor, lif, against the number of processors, respectively. in particular, simulation results were collected for number of processors equal to 1, 2, 4, 8, 16 and 32. however, curve fitting techniques were applied in order to obtain smooth curves in those figures. the results shown in fig. 6(a), (b) and (c) depict the relationship between the nsl(m) of dags with pccr valus equal to 0.5, 1.0 and 5.0, respectively, versus the number of processors, m. the nsl results are high with small number of processors, but they approach optimal values at high number of processors. the nsl values are optimal in the cases of dag size of 50 nodes, and close to optimal for higher sizes, when m > 16. the next set of results is for the average speedup factor against the number of processors. they are shown in fig. 7(a), (b) and (c) for pccr values equal to 0.5, 1.0 and 5.0, respectively. before the crossover points, the speedup factor is higher for lower dag sizes. while after the crossover points, the speedup factor becomes higher for higher dag sizes. the crossover points in the speedup curves are expected. because, the potential degree of parallelism (represented by the average dag width) for smaller dag sizes would have higher opportunity to match the available number of processors, before the crossover points, than large dag sizes. accordingly, slm of smaller dag sizes on low number of processors would be close to slo, as shown in fig. 6 by the relationship between the nsl and the number of processors, m. however, the high potential degree of parallelism found in large dag sizes would not have an opportunity for exploitation with a limited number of processors. accordingly, their slm values would be much larger than slo due to processing tasks allocated to the same processor sequentially. but the situation changes after the crossover points, as more processors become available, since they can match the potential degree of parallelism found in high dag sizes. the discrepancies in the crossover points among the three curves is attributed to the average pccr of the dags. it is very clear that the speedup does not scale linearly with the number of processors. definitely, synchronous communication has a serious drawback on limiting the pccr = 0.5 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors lo ad im ba la nc e fa ct or v=50 v=250 v=500 (a) pccr = 1.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors lo ad i m ba la n ce f ac to r v=50 v=250 v=500 (b) pccr = 5.0 0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 number of processors lo ad i m ba la n ce f ac to r v=50 v=250 v=500 (c) figure 10. the load imbalance factor versus the number of processors 88 the journal of engineering research vol. 2, no. 1 (2005) 77-89 degree of speedup that can be achieved. to clarify this point, we compare between the results of the utilization, u(m), and the efficiency, e(m). figure 8(a), (b) and (c) shows the u(m) versus the number of processors, m, for pccr values equal to 0.5, 1.0 and 5.0, respectively, using three dag sizes of 50, 250 and 500 nodes. similarly, fig. 9(a), (b) and (c) shows e(m) versus the number of processors using the same parameters. both measures would have u(m) = e(m) = 1 at m=1. however, at m = 2 the utilization is high, it is greater than 87%. while the efficiency drops sharply to a value as low as 17%. the differences between the values of u(m) and e(m) are attributed to the synchronization overhead ratio, sor(m), as shown previously. this is the blocking time due to synchronous communication, during which a sender is waiting for an acknowledgement from a receiver. the difference between u(m) and e(m) is very wide in the case of fine granularity tasks, but it becomes narrower as the granularity increases. the results of the load imbalance factor, lif(m), are shown in fig. 10(a), (b) and (c). they are taken for α = 1.0 and pccr values equal to 0.5, 1.0 and 5.0, respectively. in general, the behavior of the lif(m) results is the complement of the u(m). 5. conclusions this work has introduced a low cost algorithm, called glb-synch. it is intended to perform task mapping and ordering, in the context of synchronous communication as a part of a multi-step scheduling approach. we have shown by analysis that the complexity of the glb-synch algorithm is o(mc + mv logv + ev). the simulation results show that a multi-step scheduling using the nlc_synchcom and the glb_synch algorithms for clustering, and mapping and ordering, respectively, retain the same low complexity cost for both steps. the performance study has shown limited speedup gain over different dag shape factors. the limitations in the achieved speedups are mainly attributed to the synchronization overhead. further improvements in clustering and mapping techniques are needed to achieve high performance, unless the objective for synchronous parallel and distributed computing is otherwise. acknowledgments i gratefully acknowledge the support provided by sultan qaboos university through research grant no. ig/sci/comp/02/02 in facilitating and encouraging an environment conducive to research and academic excellence in the sultanate of oman. references arafeh, b., 2003, "non-linear clustering algorithm for scheduling parallel programs with synchronous communication on nows," int. j. of computers and their applications, vol. 10(2), pp. 103-114. foster, i. and kesselman, c., 1997, "globus: a metacomputing infrastructure toolkit," the int. j. of supercomputing applications and high performance computing, vol. 11(2), pp. 115-128. gerasoulis, a. and yang, t., 1993, ''on the granularity and clustering of directed acyclic task graphs,'' ieee tpds, vol. 4(6), pp. 686-701. hwang, k., 1993, “advanced computer architecture: parallelism, scalability, programmability”. macgrawhill, inc., new york. kadamuddi, d. and tsai, j., 2000, ''clustering algorithm for parallelizing software systems in multiprocessors,'' ieee transactions on software engineering, vol. 26(4), pp. 340-361. kasahara lab., [http://www.kasahara.elec.waseda.ac.jp/] (viewed in july 2004), japan. kim, s. j. and browne, j.c., 1988, ''a general approach to mapping of parallel computation upon multiprocessor architectures,'' proceedings of int. conf. on parallel processing, vol. ii, pp. 1-8. kowk, y. k. and ahmad, i., 1999, ''static scheduling algorithms for allocating directed task graphs to multiprocessors,'' acm computing surveys, vol. 31(4), pp. 406-471. kowk, y. k. and ahmad, i., 1996, ''dynamic critical-path scheduling: an effective technique for allocating task graphs onto multiprocessors," ieee tpds, vol. 7(5), pp. 506-521. lee, h., kim, j., hong, s. and lee, s., 2003, "task scheduling using a block dependency dag for block-oriented sparse cholesky factorization," parallel computing, vol. 29(1), pp. 135-159. liou, j. c. and palis, m. a., 1997, "a comparison of general approaches to multiprocessor scheduling," proceedings of 11th int. parallel processing symposium, pp. 152-156. papadimitriou, c. and yannakakis, m., 1999, ''towards an architecture-independent analysis of parallel algorithms,'' siam j. on computing, vol. 19(2), pp. 322-328. radulescu, a., 2001, “compile-time scheduling for distributed-memory systems”. ph.d. thesis, faculty of information technology and systems, delft university of technology, delft, the netherlands. sarkar, v., 1989, “partitioning and scheduling programs for execution on multiprocessors”. cambridge, ma:mit press. shirazi, b., wang, m. and pathak, g., 1990, ''analysis and evaluation of heuristic methods for static scheduling,'' j. of parallel and distributed computing, vol. 10, pp. 222-232. wu, m. y. and gajski, d.d., 1990, ''hypertool: a programming aid for message-passing systems,'' ieee tpds, vol. 1(3), pp. 330-343. yang, t. and gerasoulis, a., 1994, ''dsc: scheduling 89 the journal of engineering research vol. 2, no. 1 (2005) 77-89 parallel tasks on unbounded number of processors," ieee tpds, vol. 5(9), pp. 951-967. yang, t. and gerasoulis, a., 1992, "list scheduling with or without communication delays," parallel computing, vol. 19, pp. 1321-1344. ece060627.qxd the journal of engineering research vol. 6, no.1, (2009) 15-20 1. introduction it is widely acknowledged that in order for a removable partial denture (rpd) framework to achieve maximum function, create pleasing aesthetics and remain biocompatible, accuracy of fit must be optimized (sykora, 1997; brudvik, et al. 1992 and stern, et al. 1985). rpd frameworks are fabricated from high-shrinkage metal alloy, leading to difficulties in achieving accurate fitting (fritell, et al. 1985). in order to determine how to hone fabrication processes so that accuracy can be maximised, a careful examination of the size of gaps between frameworks and casts and where these occur is necessary. several researchers have studied such spaces in a research context (sykora, 1997). for example, elastomeric impression material has been allowed to set in gaps and the thickness of the m aterial measured (stern, et al. 1985). again, the framework and cast have been encased in resin then sec_________________________________________ *corresponding author’s e-mail: rjwillaims@uwic.ac.uk tioned to reveal gaps between the framework and cast, accurately maintained by the resin (fritell, et al. 1985). both the latter studies used photographs of sections taken across spaces, which were enlarged by projecting transparencies onto screens to increase accuracy of measurement. other researchers have shown that a framework may appear to fit 'in vivo' but with the use of, for example, custom made feeler gauges under retentive ends of clasps, have shown that there are gaps not seen with the naked eye (murray, et al. 1988). the present study suggests that subjective judgments of fit based on the use of the naked eye or experience may be flawed (calverley, et al. 1987 and rantanen, et al. 1986). the above studies, although very suggestive, are unable to show that the fit of a framework on a cast may be in a distorted state (ali, et al. 1997). whilst acknowledging that a framework may be distorted when fitted to a cast or indeed for that matter the oral cavity, objectively assessing and measuring spaces between a framework and a cast is an important first step in quality assurance. an electronic method for measuring the fit of removable partial denture frameworks to dental casts robert j williams*, tahseen rafik and zeid al-hourani centre for dental technology and school of product and engineering design, university of wales institute, cardiff, llandaff, cardiff, cf5 2yb, united kingdom received 10 january 2007; accepted 9 may 2007 abstract: it is well established that the removable partial denture (rpd) is an effective treatment prosthesis. the objectives of a successful rpd are: to preserve the health of remaining oral structure, restore function and restore esthetics. to achieve these objectives, an rpd framework must fit accurately to the supporting structures. this paper presents a method for measuring the gaps or spaces present between the rpd framework and supporting structures which will enable the dentist and the dental technician to evaluate the accuracy of fitting of the prosthesis before it is delivered to the patient. the method used in this research is based on the principle of electric capacitance and uses a specially designed prototype measurement system. keywords: dental technology, capacitance transducers, small gap measurement ö«cîdg πñb ∑ôëàÿg »fõ÷g ¿éæ°s’g º≤w áecóe ¢sé«≤d á«fhîµdg á≤jôw õª«∏jh äôhhq@êgqƒ¡dg ~jr h ≥«aq ú°ù– , áá°°uuóóÿÿgg¿ƒµj ¿g øe ~h ’ ∞féxƒdg √òg ≥«≤ëàd .√oƒ≤øÿg ¿éæ°s’g áø«xh ioéyg h ‹éª÷g ô¶æÿg ioéyg øª° àj ∑ôëàÿg »fõ÷g ¿éæ°s’g º≤w áø«xh ¿g 髪∏y âhéãdg øe áfeg : áñ°ùf o~– ≈àdg »g º≤£∏d á∏eé◊g aé° y’g h º≤£dg πµ«g hg º°ùl úh √oƒlƒÿg äézgôødg ¿g å«m .º≤£dg ìé‚ øª° j ¬b~dg øe á«déy ¬lq~h ´ƒæ°üe h ºª°üe ¿éæ°s’g º≤w ö«ñ£dg ωƒ≤j ¿g πñb √oé«©dg h èàıg ‘ ¿éæ°s’g º≤w º««≤j øe ¿éæ°s’g »æa h ö«ñ£∏d í«ààd á≤«bo ᫪∏y √qƒ°üh äézgôødg √òg ¢sq~j ¿g øe ~h ’ ∂dòd .¬°vé©à°s’g ìé‚ é°ü«°ün ºª°u (á«féhô¡µdg äé©°sƒÿg ≈∏y ¬f~ñe ‘ ~ªà©j) ré¡l ≈∏y ~ªà©j á«yƒ°vƒe ¬≤jôw ωg~îà°séh äézgôødg √òg ¢sé«b ¤g åëñdg gòg ±~¡j .¢ jôª∏d ¿éæ°s’g º≤w º«∏°ùàh .ájé¨dg √ò¡d ::áá««mmééààøøÿÿgg ääé骪∏∏µµddgg.√ò¨°üdg ägƒéødg ¢sé«b , äé°ùûg ,ácôëàÿg á«fõ÷g ¿éæ°s’g º≤wg ,¿éæ°s’g äé«æ≤j 16 the journal of engineering research vol. 6, no.1, (2009) 15-20 this paper reports on a method of objectively assessing framework fit, which is comparatively fast and efficient and promises to allow accurate measurements of spaces to be made outside the research context on a routine basis. 2. methods for gaps measurements there are three main methods for measuring small gaps. these are x-rays, ultrasonic and capacitive transducers. the theory of operation of each of these methods is presented below. 2.1 x-ray when a material is subjected to x-ray bombardment, some of its electrons will gain energy and leave the atom, creating a void in the vacated shell, thereby releasing a photon of x-ray energy known as x-ray fluorescence. the energy level or wavelength of fluorescent x-rays is proportional to the atomic number and is characteristic for a particular material. the quantity of energy released will be dependent upon the thickness of the material being measured. basically, the x-ray fluorescence unit consists of an xray tube and a proportional counter. emitted photons ionize the gas in the counter tube proportional to their energy, permitting spectrum analysis for determination of the material and thickness. x-ray fluorescence is the most precise measurement method, especially for small-diameter parts, or dual coatings such as gold and nickel over copper. 2.2 ultrasonic ultrasonic thickness gauging is a widely used nondestructive test technique for measuring the thickness of a material from one side. the ultrasonic method uses the pulse reflection principle and is based on the amount of time it takes for ultrasonic waves to pass through and return from inside the object. this is explained below with reference to fig. 1. the measurement process starts by transmitting a pulsed ultrasound energy from the ultrasonic probe through the framework. the ultrasonic pulse will travel through the framework medium until it reaches the interface between the framework and the cast. because of the difference in the medium, part of the ultrasonic energy will be reflected back to the probe (first reflection) and the majority of the energy will continue travelling through the gap. when the wave reached the end of the gap and hits the cast, another reflection occurs (second reflection). the time between the two reflections depends on the depth of the gap and the sound speed in the medium of the gap. gap depth = (t2 t1) * speed of sound in the gap the resolution of the above ultrasonic method depends on the pulse width (the narrower the pulse, the higher the resolution). another constraints is the quality factor of the ultrasonic probe (minimum number of cycles in the pulse). typical ultrasonic transducers have a q-factor of 10. that means we need to have a minimum of 10 cycles per pulse. to achieve a resolution of 0.1 mm, a very narrow pulse is needed with 10 cycles in it. this means a very high frequency ultrasonic wave is needed (around 100mhz). the probes that are capable of working at this frequency are usually expensive and very delicate. a disadvantage of radiation methods is the use of xray or gamma radiation that requires special safety measures for protection of the users against the radiation. the equipments used are also relatively expensive. there were two main limitations for the use of ultrasonic non-destructive testing in this study. these limitations can be summarized as: 1. to achieve a resolution of 0.1 mm with an ultrasound probe it should have a high frequency. this high frequency depends on the speed of the sound traveling in the material (in this case the cobalt chromium rpd framework). the resolution is related to the shortened pulse length that can be produced, which is very difficult to achieve. a demonstration held in the dental technology centre at uwic was promising but proved very expensive. the technique was excluded due to the high expense of instrumentation needed. 2. an x-ray technique can be used to measure a gap between an rpd framework and a cast. however, about 4 images are needed for each framework and these need to be digitalized before the measurement can be taken. this technique is time consuming. ultrasonic probe (transmitter and receiver) cast framework gap (different medium) first reflection second reflection time time difference corresponds to the depth of the gap second reflection amplitude first reflectio n figure 1. ultrasonic method for measuring the gap between the cast and framework 17 the journal of engineering research vol. 6, no.1, (2009) 15-20 because of the above-mentioned difficulties and disadvantages, an alternative method was sought. initially it was proposed that the electronic principle of capacitance might be applicable to the problem in hand. 2.3 capacitive transducers a capacitor in its simplest form consists of two conducting plates separated by an insulating layer called a dielectric. when a capacitor is connected in a circuit across a voltage source, the voltage forces electrons onto the surface of one plate and pulls electrons off the surface of the other plate resulting in a potential difference between the plates. the capacitance of a capacitor is proportional to the quantity of charge that can be stored in it for each volt difference in potential between its plates. mathematically this relationship is written as: c = q/v (1) where c is capacitance in farads, q is the quantity of stored electrical charge in coulombs and v is the difference in potential in volts. the difference in potential or voltage of the capacitor can be calculated using the formula: v = q/c (2) the capacitance of a capacitor is affected by three factors: 1. the area of the plates. 2. the distance between the plates. 3. the dielectric constant of the material between the plates. larger plates provide greater capacity to store electric charge. therefore, as the area of the plates increases, capacitance increases. capacitance is directly proportional to the electrostatic force field between the plates. this field is stronger when the plates are closer together. therefore, as the distance between the plates decreases, capacitance increases. as the distance between the plates increases, capacitance decreases. the ability of the dielectric to support electrostatic forces is directly proportional to the dielectric constant. therefore, as the dielectric constant increases, capacitance increases. taking into account each of the above three factors, the capacitance of a capacitor with two parallel plates can be calculated using the formula: c = (8.855 × 10-12 ka)/d (3) where k is the dielectric constant; a is the area of the plate; d is the distance between the plates and 8.855 × 10-12 is the absolute permittivity of free air space. substituting the value of c described by eq. 3 in eq. 2 will result in: v= d × (q / k × a × 8.855 × 10-1 ) (4) equation 4 shows that if q, k and a are kept constant, then the value of the voltage across the two plates are directly and linearly related to the distance between the two plates. the above principle was used to develop a method of measuring the fit of rpd frameworks to dental casts. the denture is usually made from conductive material and forms one end of a capacitor. the other plate of the capacitor is formed by sticking a small sensor (0.01mm thick) made from conductive foil on the cast. by connecting the ends of these two plates to a high-precision current source and measuring the voltage across the two terminals, since the voltage will be directly related to the gap between the denture and the cast, the distance between the rpd framework and the cast can be determined. 3. experimental set-up and measurements figure 2 show the experimental set-up used in this project. it is based on a high precision current source developed by sensatech research ltd (sensatech research ltd, unit 6 level 3 north, new england house, new england street brighton, uk. bn1 4gh, www.sensatech.com). this device is used to maintain constant current (constant q) in the capacitor, irrespective of its value. the device is powered by 20v dc power supply and the voltage across the capacitor is measured by a digital voltmeter. the voltage should be directly related to the gap between the capacitor's plates. it was necessary to calibrate the system and find the voltages for different known gaps and for different sensor sizes. to do this, a simple setup was built using a micrometer where one plate of the capacitor was mounted on the fixed end of the micrometer while the other plate (the sensor) was fixed on the moving end as shown in fig. 3. the gap between the two plates was varied in 0.1mm steps using the micrometer dial and for each gap value the voltage was measured. the results were tabulated in a look-up table and plotted in fig. 4. the decision was taken to use one sensor size (5x5mm) to carry out the evaluation process. this decision was taken because it has the best straight line and all the rpd components were almost the same size (rest, guide plates, gnd dv high precision current source dc power supply 20v dc dvm figure 2. basic experimental set-up 18 the journal of engineering research vol. 6, no.1, (2009) 15-20 terminus third of clasp retentive arms, and the reciprocal clasp arms). sensors were adapted to a cast in critical areas and a framework fitted to the cast. areas of interest were, for example, under occlusal rests, retentive ends of clasps and guide planes (fig. 5). the sensors were connected to a circuit shown in fig. 6. capacitance readings were taken at each sensor and the size of any gaps determined by using the above graphs. the size of gaps measured by capacitance were then verified using radiography (fig. 6) and by measuring the thickness of acrylic resin usually used for crown and bridge modelling (palavit®g, heraeus kulzer gmbh, philipp-reis-strabe, germany) material which was allowed to flow into gaps. this technique has been used before to evaluate the accuracy of fitting of rpd framework (figs. 7a and 7b). the palavit g resin were placed on the area of concern of the gypsum cast, the framework were then placed to full seating applying sufficient figure pressure allowing the resin to polymerise. the thickness of the resin layer was then measured using digital micrometer calibre. the average of three for each component was then recorded in table 1. all three methods of measurement yielded results that coincided to a large degree in table 1. thus confirmation of the accuracy of the electronic method of measurement described above was provided. dial fixed end moving end adjustable gap micrometer measurement system sensor figure 3. system calibration different sensor sizes 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 gap (mm) vo lta ge (v ) 3x2 3x3mm 5x5mm 10x5mmsensor size= figure 4. results of capacitance for known gaps using sensor size 10x5mm, 5x5mm, 3x3mm 3x2mm high precision current source dc power suppl y dvm gnd rotary switch sensors glued to the cast figure 5. block diagram of the complete fitness measurement system rpd metal framework sensor connected to the measurement system figure 6. sensor connected to the dental cast with rpd framework positioned above figure 7a. x-ray image for rpd framework used to confirm the results of the capacitance 19 the journal of engineering research vol. 6, no.1, (2000) 15-20 4. results the graph represents the measurements of known gaps. it is shown that the voltage increases as the gap size increases. area, as well as gap size, affects the capacitance measurement. thus each line refers to a given area. the results were verified firstly using x-rays. a radiograph of gaps between the model and the alloy framework were measured. secondly a resin (palavit g) was flowed into gaps, allowed to set and the thickness (corresponding to the gap size) measured table 1 shows measurement of gaps using x-ray, palavit g, and capacitance measurements. this was achieved by placing the x-ray on a light table, and using the graticule magnifier (lensel optics pvt ltd. 66/2, d2, midc, chinchwad, pune 411 019, india) with scale range 0.1mm-22mm, touching the x-ray image the gap between the metal framework and the cast was measured. three readings were recorded for each component and the average is shown in table 1. 5. discussions many clasps engage only a 0.25 mm undercut. if the terminus of an active clasp presents a gap between itself and the cast of only 0.1 mm, for example, the retentive function of that clasp will be greatly reduced. similarly, occlusal supports functioning as indirect retainers would not need to move a great distance towards the surface of the underlying tissue to cause lifting of the prosthesis. the objective measurement of such spaces is an important first step in assessing the function of an rpd. the method described above offers this facility conveniently. although there is a thickness of sensor and adhesive used in the method described above, this is only 0.16% of a 0.25 mm gap, for example. further, if the sensor is adapted along the whole length of a clasp, it is likely that the clasp will be distorted so as to maintain a similar gap at the active end which would be evident without the sensor. the same is true of occlusal rests if the foil is adapted over the whole rest seat. the electronic device described above offers a method of simply, accurately and objectively assessing the fit of a framework to a cast in crucial areas, which is readily available and not limited to a research setting. not only can spaces under important areas of the framework be detected but they can be accurately measured. with the use of the device described, frameworks thought to be satisfactory based on subjective assessment may well be deemed to be unsatisfactory, and vice versa references ali, m., narin, r.i., sherriff, m. and waters, n.e., 1997, "the distortion of cast cobalt-chromium alloy partial denture frameworks fitted to a working cast," j. of prosthetic dentistry. vol. 78(4), pp. 419424. brudvik, j.s. and reimers, d., 1992, "the toothremovable partial denture interface," j. of prosthetic dentistry, vol. 68(6), pp. 924-7. calverley, m.j. and moegel, j.r., 1987, "effect on the fit of removable partial denture framework when master casts are treated with cyanoacrylate," j. of prosthetic dentistry, vol. 58(3), pp. 327-329. dunham, d., brudivik, j., morris, j., plummer, k. and cameron, s., 2006, "a clinical investigation of the fit of removable partial dental prosthesis clasp assemblies," j. of prosthetic dentistry, vol. 95(4), pp. 323-6. fritell, d.n., muncheryan, a.m. and green, a.j., 1985, "laboratory accuracy in casting removable partial denture frameworks," j. of prosthetic dentistry, vol. 54(6), pp. 856-862. massad, d., william, a. and lily, t., 2006. "a method to minimize rpd framework deformation during fabrication procedure," dentistry today, vol. 25(2), pp. 122, 124, 126. murray, m.d. and oyson, j.e., 1988, "a study of clinical fit of cast cobalt-chromium clasps," j. of dentistry, vol. 16(3), pp. 135-39. rantanen, t. and eerikainen, e., 1986, "accuracy of palatal plate of removable partial dentures, and influence of laboratory handling of the investment figure 7b. x-ray image for rpd framework used to confirm the results of the capacitance. it is evident that the gaps appear under the major connector, canine, upper right premolar and upper right molar devic e retentive arm g u i d e plane m a j . conn. r e s t – rest seat recip. arm gap on x-ray < 0.1m m 0 . 2 m m 0.3 mm 0 . 1 m m < 0.1mm capac . 0 . 0 m m 0 . 2 m m 0.3 mm 0 . 1 m m 0.0mm palavit®g < 0 . 1 m m 0.15m m 0.27mm 0.95mm <0.1mm table 1. comparison of results of measuring the gap between some components of the rpd framework and the supporting cast 20 the journal of engineering research vol. 6, no.1, (2009) 15-20 on the accuracy," j. of dental materials, vol. 2, pp. 28-31. stern, m.a., brudvik, j.s. and frank, r.p., 1985, "clinical evaluation of removable partial denture rest seat adaptation," j. of prosthetic dentistry, vol. 53(5), pp. 658-62. sykora, o. and sutow, e.j., 1997, "improved fit of maxillary complete dentures processed on high expansion stone casts," j. of prosthetic dentistry, vol. 77(2), pp. 205-8, the journal of engineering research (tjer), vol. 19, no. 1, (2022) 63-72 *corresponding author’s e-mail: supriyakhadka1996@gmail.com doi:10.53540/tjer.vol19iss1pp63-72 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings supriya khadka*, mishan shrestha, and hom b. rijal graduate school of environmental and information studies, tokyo city university, yokohama, japan abstract: this study investigates the overall comfort and productivity of japanese office workers in mixedmode office buildings. the indoor thermal environment is adjusted using the air-conditioning in japanese office buildings to maintain thermal comfort and productivity. thus, it is necessary to research thermal comfort and productivity to understand how occupants prepare themselves to be at a comfortable temperature and perform their daily tasks under mixed-mode (mm) and free-running (fr) modes. environmental parameters such as air temperature, relative humidity, and so on were measured in 17 japanese office buildings with the help of digital instruments, and thermal comfort transverse surveys were conducted for two years in tokyo, yokohama, and odawara of japan. the data were collected every once a month for a day visiting each building with the measurement instruments, together with the questionnaires. almost 3000 votes were collected. this paper evaluates the overall comfort discussions followed by how the occupant could achieve their productivity. the occupants were found to be thermally comfortable and productive in the office. the most suitable comfortable temperature range for mm mode was found to be 22–26 °c and 23–25 °c for fr mode. the workers' productivity range is defined by the globe temperature range of 21–27 °c for mm and 20–27 °c for fr mode. the findings should be useful to suggest that whenever new office buildings are designed, these factors always need to be taken into consideration. keywords: field survey; free running mode; japanese office buildings; mixed mode; productivity; thermal comfort . التحقیق في الراحة الحراریة واإلنتاجیة في المباني المكتبیة الیابانیة المختلطة و میشان شریستا و ھوم ریجال *سوبریا خادكا الدراسة في الراحة العامة واإلنتاجیة للعاملین في المكاتب الیابانیة في مباني المكاتب ذات الوضع المختلط. یتم تبحث ھذه الملخص: ، من ضبط البیئة الحراریة الداخلیة باستخدام تكییف الھواء في مباني المكاتب الیابانیة للحفاظ على الراحة الحراریة واإلنتاجیة. وبالتالي ل الراحة الحراریة واإلنتاجیة لفھم كیفیة إعداد أصحاب المكاتب ألنفسھم لیكونوا في درجة حرارة مریحة الضروري إجراء بحث حو ). تم قیاس معاییر البیئة مثل درجة حرارة الھواء والرطوبة fr) ووضع التشغیل الحر (mmوأداء مھامھم الیومیة في الوضع المختلط ( انیًا بمساعدة األدوات الرقمیة ، وأجریت استطالعات عرضیة للراحة الحراریة لمدة عامین مبنى مكاتب یاب 17النسبیة وما إلى ذلك في افة في طوكیو ویوكوھاما وأوداوارا في الیابان. تم جمع البیانات مرة واحدة شھریًا لمدة یوم لزیارة كل مبنى باستخدام أدوات القیاس باإلض مرة. تقیّم ھذه الورقة مناقشات الراحة الشاملة متبوعة بكیفیة تحقیق الموظف إلنتاجیتھ. 3000إلى االستبیانات التي تم ملؤھا لما یقارب تم العثور على شاغلي المكتب مرتاحین حراریاً ومنتجین في المكتب. تم العثور على أنسب نطاق درجة حرارة مریحة للوضع المختلط )mm الحر ( درجة مئویة لوضع التشغیل 25–23درجة مئویة و 26–22) ھوfr یتم تحدید نطاق إنتاجیة العمال من خالل نطاق .( درجة مئویة لوضع التشغیل الحر. یمكن 27–20درجة مئویة لـوضع التشغیل المختلط و 27إلى 21درجة حرارة الكرة األرضیة من االستفادة من نتائج ھذه الدراسة ألخذھا في عین االعتبار عند تصمیم مباني جدیدة. الراحة ؛اإلنتاجیة ؛النمط المختلط ؛مباني المكاتب الیابانیة ؛التشغیل الحرنمط ؛الدراسة االستقصائیة المیدانیة الكلمات المفتاحیة: .الحراریة 64 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings 1. introduction thermal comfort is a person’s perception of how they feel related to the air temperature, radiant temperature, relative humidity, and air movement of their surroundings. according to ashrae (2017), “thermal comfort is defined as the condition of mind that expresses satisfaction with the thermal environment and is assessed by subjective evaluation.” the fact is that the relationship between the individual and the environment is very complex and active due to the presence of various factors like climate, buildings, social, economic, and some other factors. today, people spend more and more time indoors, where they expect a level of thermal comfort that ensures comfort and wellbeing. even gentle fluctuations can cause discomfort, which may lead to a sudden change in the behaviour or the activity of the occupant. therefore, it is important to provide better working thermal environments, so-called “comfortable environments”. japanese office buildings are well equipped with airconditioning (ac) systems to help in creating thermal comfort. although the japanese office buildings do have ac systems, it seems that the various practices, such as opening doors and windows to allow air movement as much as possible, enable the occupants to be at their required thermal comfort while at work. the japanese government introduced the “cool biz” and “warm biz” programs that recommend an indoor temperature of 28 °c for cooling and 20 °c for heating in the year 2005 (enomoto et al., 2009). productivity is defined as the extent to which activities result in the achievement of the system goals (parsons 2003). individuals always have different thermal expectations, which will differ from individual to individual. people have different thermal expectations, and thus thermal comfort is likely to vary according to the month, season, and mode. according to seppanen et al. (2006), the productivity increases with air temperature up to 21–22 °c. they found that the highest productivity is achieved at an air temperature of around 22 °c. for example, at the air temperature of 30 °c, the productivity is only 91.1% of the maximum, i.e., the reduction in performance is 8.9%. horr et al. (2016) reviewed a broad range of literature and found that many factors affect occupant comfort and productivity. vimalanathan and ramesh babu (2014) investigated the significant effects of indoor room temperature and illumination on the office worker’s performance. it was found that the suitable optimum level for indoor room temperature was 21°c. an experimental study conducted by ismail et al. (2014) found that along with temperature, illuminance and relative humidity also dominated the productivity of workers. based on a review of existing literature by fisk et al. (1997), there is strong evidence that characteristics of buildings and indoor environments significantly influence rates of workers’ productivity. the experimental results in tsay et al.’s (2022) study showed that male and female workers have different optimal temperatures for their productivity at work. rasheed et al. (2021) found that there were significant differences in perceptions of comfort and productivity for those who spent less time and those who spent more time at work. ngarmpornprasert and koetsinchai (2010) found a satisfactory thermal condition for office workers by maintaining the temperature at 26–28 °c for morning periods and 24.5– 26 °c for afternoon and evening periods. based on field surveys, the comfort temperatures in japanese offices have been investigated by several researchers (rijal et al., 2017). however, not much research has been conducted to investigate the relationship between thermal comfort and productivity in japanese office buildings as compared to other countries. comfortable temperatures are important to investigate because the indoor temperatures chosen affect the energy used in the building, and people in thermal comfort are generally more productive. thermal discomfort caused by high or low air temperature had a negative influence on office workers' productivity, and the subjective rating scales were important supplements of neurobehavioral performance measures when evaluating the effects of indoor environmental quality on productivity (lan et al., 2012). however, most of the previous studies are conducted in summer, and thus thermal comfort for other seasons and in mixed-modes office buildings is still unknown. the basic principles are largely universal, but thermal comfort varies from person to person. therefore, long-term data are required to fully describe the occupants’ perceptions and behavioural responses to the thermal environment in their offices (rijal et al., 2017). the more control over the thermal comfort, the better a person can feel and more productive at work they will become. according to tanabe et al. (2007), evaluating the productivity of office workers promotes the effort for energy conservation. the short data collection periods and few samples’ collections are also the major drawbacks of not having significant research on this topic so far. we conducted a transverse survey to record the thermal comfort and productivity responses of the japanese office workers. furthermore, the responses were analyzed to determine the occupants' overall comfort and productivity under mixed-mode (mm) and free-running (fr) mode. this research holds 3000 votes from 17 different mixed-mode office buildings located in tokyo, yokohama, and odawara and the data are collected for two years. 2. methodology 2.1 investigated areas and buildings this field survey was carried out in 17 different office buildings located in tokyo, yokohama, and odawara of japan from august 2014 to october 2015 and from august 2017 to november 2018 (fig. 1). table 1 shows the description of the investigated buildings, supriya khadka, mishan shrestha, and hom b. rijal 65 including locations and investigated floors. the investigated buildings were of change-over mixedmode types. the change-over mixed-mode buildings have openable windows and doors, or can be airconditioning mode depending on the seasons or time of the day (cbe, 2021). the survey was carried out for two years to ensure collecting as much information as possible for different months and different seasons (rijal et al., 2019, 2022). 2.2 thermal measurement survey indoor and outdoor environmental variables were measured, including air temperature, globe temperature, relative humidity, and air movement. they were collected at 1.1 m height above floor level, away from direct sunlight, using digital instruments as shown in fig. 2. climatic data were obtained from the nearest meteorological stations (table 1). table 2 shows the characteristics of the instruments used in the survey. a globe thermometer with a diameter of 75 mm or 40 mm rather than 150 mm is widely used for thermal comfort field surveys (nicol et al., 1994, de dear et al. 1997, nicol et al. 1999, brager et al. 2004, humphreys and nicol 2007, rijal et al., 2019a). the time constant for the globe thermometer with the 150 mm diameter is about 20 minutes (spagnolo and de dear 2003, rijal et al. 2003) but it is less for the globe thermometer with the diameter of 75 mm which would be sufficient to stabilize in the indoor space. furthermore, the response time for temperature measurement in practical life is higher for smaller ones because of the smaller surface (humphreys 1977, nicol et al., 2012). recently, d’ambrosio alfano et al. (2021) found that the globe thermometer with a 50 mm diameter showed lower errors than 38 mm diameter. they also indicated that it is still unclear whether small globes, characterized by low response times, exhibit the same accuracy as the standard 150 mm in predicting the mean radiant temperature. the measured data was recorded 15–20 minutes after the instruments were set to ensure a stable measurement as shown in rijal et al. (2017). figure 1. general view of one of the investigated office buildings in yokohama. figure 2. digital instrument set up (rijal et al. 2017) table 1. description of the investigated buildings building code location* investigated floor** b2 yokohama 1f, 3f~5f b4 yokohama 1f, 2f b5 yokohama 3f~7f b6 yokohama 1f b7 tokyo 1f, 4f b8 tokyo 1f, 2f b13 tokyo 2f~5f b14 tokyo 1f, 3f, 4f b15 tokyo 1f b16 tokyo 1f~3f b17 tokyo 1f b18 tokyo 2f,3f b19 tokyo 1f, 4f b20 tokyo 2f~4f b21 tokyo 4f b22 tokyo 4f b23 odawara 2f *: meteorological station, **: the floor is counted by the american system, f: floor table 2. description of the instruments used. parameter measured trade name range accuracy air temp., humidity tr-76ui 0 to 55 °c, 10% to 95% rh ±0.5 °c, ±5%rh, globe temp. tr-52i −60 to 155 °c ±0.3 °c sibata 08034075 black painted 75 mm diameter globe air movement kanoma x, 654321 0.01 to 5.00 m/s ±0.02 m/s illuminance tr-74ui 0 to 130 klx ±5% 66 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings 2.3 thermal comfort and productivity survey each investigated building was visited for one day each month to collect the instruments’ measurements and subjects’ filled questionnaires. the reading was taken just once for each group on each visit to each office. the survey methods are given in rijal et al. (2017). to collect the data, the instruments were planned and set up on the office table, and the questionnaires were distributed among the occupants seated near the instruments, as shown in fig. 3. when people were filling up the questionnaire, the researcher recorded the common environmental controls and the physical data from them. after collecting the data for that group, the instruments were moved to the next group and, so on. this process was repeated every month. the data includes overall comfort and productivity. table 3 shows the scale used in the survey. we have used a six-point unidirectional scale as it is widely accepted for thermal comfort and productivity surveys (mccartney and nicol, 2002). in particular, it provides more categories for the thermal comfort scale than the scale given in iso (1995). the survey was carried out in the japanese language. we have collected the data from the healthy office workers. 3. results and discussion 3.1 thermal environment during the survey the relationship between the indoor and outdoor thermal environment was investigated by statistical analysis. it was found that the range of indoor globe temperature is similar to the indoor air temperature. globe temperature is highly correlated with the indoor air temperature for mixed-mode (r = 0.76) and freerunning mode (r = 0.71), so the globe temperature was used for further analysis. moreover, the globe temperature measures the combined effects of radiant heat, air temperature, and wind speed. the mean indoor air temperature in mm was maintained at 22.8 °c during winter and 26.5 °c during summer. it is slightly close to the recommendation of the japanese government that the indoor temperature is at 20 °c in winter and 28 °c in summer. the mean outdoor temperatures were 8.5 °c during winter and 28.0 °c during summer for mm conditions. figure 4 shows the relationship between the indoor globe and the outdoor air temperatures for mm and fr modes. the mean globe temperature during the voting was 24.8 °c, 25.0 °c for the mm and fr, respectively. although there is seasonal variation in the monthly outdoor temperature as shown in table 4, the changes in the globe's temperature are quite small. a probable reason is that the workers used heating and cooling during winter and summer to maintain the working thermal environment in mm. figure 3. thermal comfort and productivity survey (rijal et al., 2019). table 3. overall comfort and productivity scale. scale overall comfort productivity 1 very uncomfortable very difficult to work 2 moderately uncomfortable difficult to work 3 slightly uncomfortable slightly difficult to work 4 slightly comfortable slightly easy to work 5 comfortable easy to work 6 very comfortable very easy to work figure 4. relationship between the globe temperature and outdoor air temperature: (a) mm and (b) fr mode. g lo be t em pe ra tu re (° c ) outdoor air temperature (°c ) r2=0.58 outdoor air temperature (°c ) g lo be t em pe ra tu re ( ° c ) r2=0.51 (a) mm (b) fr supriya khadka, mishan shrestha, and hom b. rijal 67 3.2 overall comfort and productivity 3.2.1 distribution of overall comfort to investigate how office workers’ perceive overall comfort under the environmental conditions (i.e. temperature, humidity, and air movement), the overall comfort votes were obtained from the questionnaires. most of the occupants voted for “3. slightly comfortable”, as shown in figure 5. very limited responses were obtained at “1. very uncomfortable”, “2. moderately uncomfortable”, and “6. very comfortable”. the reasons might be that the mm created a comfortable indoor environment and thereby fewer responses at “1. very uncomfortable” and “2. moderately uncomfortable”. as for fr mode, they may use clothing adjustment to feel comfortable. the mean overall comfort for the mm and fr was 3.91 and 3.99, as shown in fig. 5. both of them are very close to “4. slightly comfortable”. this suggests that the occupants are comfortable with their thermal environment at the office building under both modes. 3.2.2 distribution of productivity the productivity responses were obtained by considering the air temperature, humidity, air movement, lighting, indoor air quality, and overall comfort. figure 6 shows the distribution of productivity in mm and fr modes. the mean productivity for the mm and fr were 4.0 and 4.1. most of the responses were at “4. slightly easy to work” and then at “5. easy to work” for both modes. very limited responses were obtained at “6. very easy to work” in both modes. figure 5. distribution of overall comfort: (a) mm and (b) fr mode. figure 6. distribution of productivity: (a) mm and (b) fr mode. table 4. seasonal differences of outdoor temperature and globe temperature. mode description winter spring summer autumn all tout (°c) tg (°c) tout (°c) tg (°c) tout (°c) tg (°c) tout (°c) tg (°c) tout (°c) tg (°c) mm n 680 680 579 579 848 848 839 837 2946 2944 mean 8.5 22.7 19.4 25.0 28.0 26.3 20.1 24.9 19.6 24.8 s.d. 3.4 1.7 4.4 1.7 4.8 1.0 4.9 1.7 8.3 2.0 fr n 45 45 313 313 153 153 438 436 949 947 mean 10.1 22.9 21.0 25.4 22.7 26.1 19.5 24.6 20.1 25.0 s.d. 1.3 2.2 2.5 1.9 3.4 1.0 4.3 1.9 4.3 1.9 tout: outdoor air temperature, tg: globe temperature, n: number of records, s.d.: standard deviation. (b) fr (b) fr (a) mm (a) mm 68 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings 3.2.3 relationship between overall comfort and productivity figure 7 shows the relationship between overall comfort and the productivity of the occupant. the results showed that the higher the overall comfort, the higher the productivity (p<0.001) in both mm and fr modes. a study by leaman and bordass (1999) report that comfort and perceived productivity is greater in buildings where occupants have more control over the environment and in mm buildings that have both natural ventilation and air conditioning. accordingly, further analysis was conducted to confirm the relationship between overall comfort and productivity responses with the globe temperature, which determines the comfortable temperature range for mm and fr modes. 3.3 relation between overall comfort and globe temperature figures 8(a) and 8(b) show the relationship between overall comfort and globe temperature in the mm and fr modes. the majority of responses were within the temperature range of 20–28 °c in both modes. the regression equations obtained are shown in table 5. to find the globe temperature, which corresponds to the peak value of overall comfort, it is necessary to estimate where the curve is horizontal (has a slope of zero). this can be found by equating the equation to zero and differentiating the quadratic equation concerning globe temperature. figure 7. relationship between overall comfort and productivity: (a) mm, and (b) fr mode. figure 8. relationship between overall comfort and globe temperature: (a) mm and (b) fr mode using raw data, and (c) mm and (d) fr mode using binary data. (a) mm (b) fr (c) mm (b) fr (d) fr (a) mm supriya khadka, mishan shrestha, and hom b. rijal 69 the optimum globe temperatures for mm and fr modes are 24.0 °c and 20.3 °c, respectively. beyond the optimum globe temperature, uncomfortable increases because of a decrease or increase in temperature. the overall comfort scale was modified into binary form for the proportion of comfortable pmm(0,1) for mm mode and proportion of comfortable pfr(0,1) for fr mode. the scales of “1. very uncomfortable”, “2. moderately uncomfortable”, and “3. slightly uncomfortable” are classified as uncomfortable and codded by 0. the rest of the scales are classified as comfortable codded by 1. figures 8 (c) and (d) show the quadratic regression analysis conducted between the binary overall comfort data and the globe temperature. the quadratic equations are as shown in table 5. at 0.8 proportion of comfort, the comfortable range is 22–26 °c for mm and 23–25 °c for fr mode. the derivatives of the equations give the optimum globe temperatures of 28.0 °c for mm and 27.0 °c for fr mode. 3.4 relation between productivity and globe temperature to identify the comfortable range for productivity, figs. 9(a) and 9(b) show the regression analysis between productivity and the globe temperature. the quadratic regression equations are as shown in table 6. the optimum globe temperatures for mm and fr modes are 24.3 °c and 23.1 °c. the optimum globe temperature is similar to the overall comfort case. beyond the optimum globe temperature, productivity decreases because of a decrease or increase in temperature. to calculate the temperature for a given proportion, the productivity scale was modified to binary form. the scales of “1. very difficult to work”, “2. difficult to work”, and “3. slightly difficult to work” are classified as nonproductive and codded as 0. the rest of the scales are classified as productive, codded as 1. figures 9(c) and 9(d) show the quadratic regression analysis between the proportion of productivity and the globe temperature. the equations are shown in table 6. at 0.75 proportion of productive responses, the globe temperature ranges of 21–27 °c for mm and 20– 27 °c for fr mode were obtained. above or below the mentioned globe temperature, occupants feel slightly difficult to work. in fr mode, the temperature range is slightly wider than in mm. the derivatives of the equations give the optimum globe temperatures of 23.7 °c for mm and 25.0 °c for fr mode. table 7 compares the optimum temperature with those found in previous studies. in the field study, the globe temperature is considered to be close to the operative temperature (nicol et al., 1999, humphreys et al., 2013). according to our study, the indoor air temperature is close to the globe temperature. based on this evidence, we made some possible comparisons. even though the analyzing index temperatures are different in the previous studies, there are some similarities between them. table 5. quadratic regression equations for overall comfort and globe temperature. mode equation n r2 s.e.1 s.e.2 p mm pmm = -0.02tg2+0.96tg -7.61 2946 0.021 0.003 0.122 <0.001 fr pfr =-0.02tg2+0.81tg -5.37 949 0.024 0.005 0.23 <0.001 mm(0,1) pmm =-0.01tg2+0.56tg -5.95 2946 0.022 0.001 0.07 <0.001 fr(0,1) pfr =-0.01tg2+0.54tg -5.55 949 0.034 0.003 0.13 <0.001 n: number of responses, r2: coefficient of determination, s.e.1 and s.e.2: standard errors of the regression coefficient of tg2 and tg, p: significance level of the regression coefficient. table 6. quadratic regression equations for productivity and globe temperature. mode equation n r2 s.e.1 s.e.2 p mm pmm = -0.015 tg2+0.729tg -4.55 2946 0.013 0.003 0.127 <0.001 fr pfr = -0.015 tg2+0.693tg -3.85 949 0.018 0.005 0.236 0.002 mm(0,1) pmm = -0.0082 tg2+0.39tg -3.86 2946 0.013 0.001 0.07 <0.001 fr(0,1) pfr = -0.009 tg2+0.45tg -4.53 949 0.020 0.002 0.12 <0.001 n: number of responses, r2: coefficient of determination, s.e.1 and s.e.2: standard errors of the regression coefficient of tg2 and tg, p: significance level of regression coefficient. 70 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings table 7. comparison of optimum productivity temperature with previous studies references country building type method index temperature (°c) optimum temperature for productivity (°c) this study japan mm field tg mm: 24 fr: 23 mm range: 22–26 fr range: 22–25 seppänen et al. 2006 literature review ti 21–22 vimalanathan & ramesh babu 2014 india mv field ti 21 tanabe et al. 2013 japan mv field top below 27 ismail et al. 2014 malaysia mv climate chamber twbgt 24 tsay et al. 2022 taiwan (roc) mv climate chamber ti 27(male), 25 (female) tg : globe temperature, ti: indoor air temperature, top: operative temperature, twbgt: wet bulb globe temperature, mv: mechanically ventilated. figure 9. relationship between productivity and globe temperature: (a) mm and (b) fr mode using raw data, and (c) mm and (d) fr mode using binary data. (a) mm (c) mm (b) fr (d) fr supriya khadka, mishan shrestha, and hom b. rijal 71 4. conclusion the thermal comfort survey was conducted for two years in tokyo, yokohama, and odawara of japan. the study evaluated the overall comfort and productivity of the japanese office workers both qualitatively and quantitatively by using the responses and measuring indoor globe temperature. the following results were found. 1. the indoor globe temperature ± standard deviation is maintained at 24.8±2.0 °c in mm. therefore, the fluctuation of the indoor globe temperature was small. 2. the analysis showed that the occupants were highly comfortable and productive in the indoor thermal environment of the office buildings. 3. the productivity increases with increasing the overall comfort and vice-versa. 4. the most suitable comfortable temperature range was found to be 22–26 °c for mm and 23–25 °c for fr mode. the workers’ productivity range is defined by the globe temperature range of 21–27 °c for mm and 20–27 °c for fr mode. this suggests that whenever new office buildings are designed, these factors always needed to be taken into consideration. acknowledgement we are grateful to all those who participated in the survey, and to the students who helped with the data entry. we also thank the following organizations for their cooperation: gotoh educational corporation, hulic co. ltd., nikken sekkei ltd., idea consultants inc., is logistic, panasonic corporation, ps company ltd., tokyo city university, tokyo city university todoroki junior and senior high school, tokyu fudosan next generation engineering center inc. and tsuzuki ward. conflict of interest the authors declare no conflict of interest. funding this research was supported by grant-in-aid for scientific research (b) number 21h01496. references ashrae (2017), ashrae standard 55, thermal environmental conditions for human occupancy, p. 35. brager g s, paliaga g, de dear r (2004) operable windows, personal control, and occupant comfort, ashrae transactions 110 (2): 17-35. cbe (2021), mixed mode, https://cbe. berkeley .edu/mixedmode/index.html. d’ambrosio alfano f r, ficco g, frattolillo a, palella b i, riccio g (2021) mean radiant temperature measurements through small black globes under forced convection conditions, atmosphere 12 (5), 621. de dear r, brager g s, cooper d (1997), developing an adaptive model of thermal comfort and preference. final report ashrae rp884. enomoto h, ikeda k, azuma k, tochihara, y (2009), observation of the thermal conditions of the workers in the ‘cool biz’ implemented office, journal of occupational safety and health japan, 2 (1), 5-10. fisk w j, rosenfeld a h (1997), estimates of improved productivity and health from better indoor environments; 7 (3): 158-172. horr y a, arif m, kaushik a, mazroei a, katafygiotou m, elsarrag e (2016), occupant productivity and office indoor environment quality: a review of the literature, building and environment 105: 369-389. humphreys m a (1977), the optimum diameter for a globe thermometer for use indoors, annals of occupational hygiene, 20 (2), 135–140. humphreys m a, nicol j f (2007), self-assessed productivity and the office environment: monthly surveys in five european countries, ashrae transactions, 113 (1), 606-616. humphreys m a, rijal h b, nicol j f (2013), updating the adaptive relation between climate and comfort indoors; new insights and an extended database, building and environment 63: 40-55. ismail a r, nizam c m, mohd haniff m h, deros b m (2014), the impact of workers productivity under simulated environmental factor by taguchi analysis, apcbee procedia 10: 263 – 268. iso (1995) 10551, ergonomics of the thermal environmentassessment of the influence of the thermal environment using subjective judgement scales, p. 8. lan l, wargochi p, lian z (2012), optimal thermal environment improves performance of office work, rehva journal: 12-17. leaman a, bordass b (1999), productivity in buildings: the ‘killer’ variables, building research & information 27 (1): 4-19. mccartney k j, nicol j f (2002), developing an adaptive control algorithm for europe, energy & buildings, 34 (6): 623-635. ngarmpornprasert s, koetsinchai w (2010), the effect of air-conditioning on worker productivity in office buildings: a case study in thailand, building simulation 3: 165–177. nicol f, humphreys m, roaf s (2012), adaptive thermal comfort: principles and practice, earthscan, london. nicol f, jamy g n, sykes o, humphreys m, roaf s, 72 investigation of the thermal comfort and productivity in japanese mixed-mode office buildings hancock m (1994), a survey of thermal comfort in pakistan toward new indoor temperature standards, oxford brookes university, school of architecture. nicol j f, raja i a, allaudin a, jamy g n (1999), climatic variations in comfort temperatures: the pakistan projects, energy and buildings 30 (3): 261-279. parsons k (2003), human thermal environment: the effects of hot, moderate and cold environments on human health, comfort and performance, p. 328, london: taylor & francis. rasheed e o, khoshbakht m, baird g (2021), time spent in the office and workers’ productivity, comfort and health: a perception study, building and environment 195 (3): 107747. rijal h b, humphreys m a, nicol j f (2017), towards an adaptive model for thermal comfort in japanese offices, building research & information 45(7): 717-729. rijal h b, humphreys m a, nicol j f (2022), chapter 17 adaptive approaches to enhancing resilient thermal comfort in japanese offices, in: nicol f., rijal h.b. and roaf s., eds. the routledge handbook of resilient thermal comfort, edited by, london: routledge, isbn 9781032155975, pp. 279-299. rijal h b, humphreys m a, nicol j f (2019), behavioural adaptation for the thermal comfort and energy saving in japanese offices, journal of the institute of engineering 15 (2), 14-25. rijal h b, humphreys m a, nicol j f (2019a), adaptive model and the adaptive mechanisms for thermal comfort in japanese dwellings, energy & buildings 202, 109371. rijal h b, yoshida h, umemiya n (2003), summer and winter thermal comfort of nepalese in houses, journal of architecture, planning and environmental engineering, 565, 17-24 (in japanese with english abstract). seppänen o, fisk w j, lei q h (2006), effect of temperature on task performance in office environment, berkeley national laboratory. spagnolo j, de dear r (2003), a field study of thermal comfort in outdoor and semi-outdoor environments in subtropical sydney australia, building and environment 38 (5): 721-738. tanabe s, nishihara n, haneda m (2007), indoor temperature, productivity, and fatigue in office tasks hvac&r research, 13 (4): 623-633. tsay y s, chen r, fan c c (2022), study on thermal comfort and energy conservation potential of office buildings in subtropical taiwan, building and environment 208: 108625. vimalanathan k, ramesh babu t (2014), the effect of indoor office environment on the work performance, health and well-being of office workers, journal of environmental health science & engineering 12 (1): 113. 1. introduction 2. methodology 3. results and discussion 4. conclusion c:\users\raoh\desktop\hadj_paper3.xps the journal of engineering research vol. 7 no. 2 (2010) 48-58 _____________________________ *corresponding author’s e-mail: hamid@squ.edu.om a multi-resolution texture image retrieval using fast fourier transform a. abdesselam department of computer science, college of science, sultan qaboos university, p.o. 36, al-khodh, muscat-123, sultanate of oman received 9 december 2009; accepted 6 september 2010 abstract: texture is an important visual property that characterizes a wide range of natural and artificial images which makes it a useful feature for retrieving images. several approaches have been proposed to describe the texture contents of an image. in early research works, such as edge histograms-based techniques and co-occurrence-based approaches, texture descriptors were mainly extracted from the spatial domain. later on, dual spaces (transform of spatial domain) such as frequency space or spaces resulting from gabor or wavelet transforms were explored for texture characterization. recent physiological studies showed that human visual system can be modeled as a set of independent channels of various orientations and scales, this finding motivated the proliferation of multi-resolution methods for describing texture images. most of these methods are either wavelet-based or gabor-based. this paper summarizes our recent study of the use of fourier-based techniques for characterizing image textures. at first, a singleresolution fourier-based technique is proposed and its performance is compared against the performance of some classical fourier-based methods. the proposed technique is then extended into a multi-resolution version. performance of the modified technique is compared against those of the single-resolution approach and some other multi-resolution approaches recently described in literature. two performance indicators were used in this comparison: retrieval accuracy and execution time of the techniques. keywords: fourier transform, texture-based image retrieval, gabor filters, wavelet transform, multiresolution approach 49 the journal of engineering research vol. 7 no. 2 (2010) 48-58 1. introduction content-based image retrieval (cbir) has been an active research topic in the last two decades. as a consequence, several experimental and commercial image retrieval systems have been proposed during this period of time. in a cbir system, image databases are queried using the visual content of an image which is usually represented by low level features such as color, texture, shapes, or a combination of some or all of these features. texture is widely used in cbir systems since it plays a crucial role in characterizing real world images both natural (such as images of clouds, water, trees, remotely sensed data and medical images) and manmade ones (such as images of bricks, fabrics, and buildings ). several approaches have been proposed to describe texture contents of an image. in early works, texture features were mainly extracted from the pixel space itself using edge density, edge histograms, and cooccurrence-based-features to characterize the image texture (haralick et al. 1973; conners and harlow, 1980; amadasun and king, 1989 and fountain and tan, 1998). more recently, various transforms have been used to produce dual spaces from which texture features were extracted. the most common transforms are fourier (tsai and tseng, 1999; weszka et al. 1976 and gibson and gaydecki, 1995), wavelet (smith and s-f, 1994; kokare et al. 2007; huang and dai, 2003; huang and dai, 2004 and huang et al. 2006) and gabor transforms (daugman and kammen, 1987; jain and farrokknia, 1991 and bianconi and fernandez, 2007). moreover, recent studies showed that human visual system can be modeled as a set of independent channels of various orientations and scales (beck et al. 1987), this finding motivated the proliferation of multi-resolution methods for describing texture images. most of these methods are either waveletbased or gabor-based. while the use of gabor filters is supported by physiological evidences (beck et al. 1987), it suffers from serious drawbacks such as the need to tune the parameters of the filter and the complexity of the calculations involved (bianconi and fernandez, 2007). in the other hand, wavelet-based approaches are much simpler and faster. as a consequence, wavelet-based approaches gained much more popularity among the computer vision community. fourier transform has also been widely used in characterizing textures. one of the reasons for its popularity is its suitability for describing periodic functions; and it's known that texture images usually contain quasirepetitive patterns. concentrations of fourier power spectrum values capture dominant orientations of the patterns in the image and their distribution in the frequency space is closely related to coarseness of the texture (tsai and tseng, 1999 and weszka and dyer, 1976). these two features (directionality and coarseness of a texture) are of importance in texture analysis (campbell and robson, 1968). the main drawback of using fourier transform is the poor spatial localization it provides. windowed fourier transform has been introduced to overcome this problem at the cost of a significant increase in computations (yu et al. 2002). this paper proposes a new single-resolution fourierbased technique for characterizing texture images and compares its performance against those of some classical fourier-based techniques. it also describes a multi-resolution version of the technique and evaluates its performance in comparison with those of several other multi-resolution techniques recently described in literature. the two main performance indicators that are used in this comparison are the accuracy and execution time of the techniques. 2. summary of the techniques considered in this research study this study includes key fourier-based techniques, and some recent multi-resolution approaches (mainly gaborand wavelet-based ones) described in literature. 2.1 fourier-based techniques fourier transform has been widely used by image processing research community. it has the very useful property of highlighting the dominant spatial frequencies as well as the dominant orientations of the structures contained in the image. another advantage of using fourier transform is the fact that frequencydomain features are generally less sensitive to noise than spatial domain features (tsai and tseng, 1990). weszka and dyer, 1976 partitioned the spectral domain into ring-shaped and wedge-shaped areas. variance values of the fourier power spectrum of ringshaped regions and wedge-shaped regions were used to describe the coarseness and the directionality of the texture respectively d. tsai and c. tseng, 1999 proposed the use of average energy (power spectrum) of 4 co-centric areas as feature descriptor for roughness of cast surfaces. bayes and neural network classifiers were evaluated. the authors reported 100% classification accuracy rate on cast specimens containing nine roughness classes. d. gibson and p.a. gaydecki, 1995 used the fourier moduli to classify histological images. they conducted an experiment in which they compared their technique with the co-occurrence matrix based approach proposed by (haralick et al. 1973). they concluded that the two approaches worked with 50 the journal of engineering research vol. 7 no. 2 (2010) 48-58 approximately equal success but the fourier based one was much faster. yu et al. 2002 applied a local fourier transform to the image and used first and second moments as texture features. the local fourier transform consisted of applying eight 3x3 templates approximating the fourier transform of the 3x3 window resulting in 8 transformed images for which the first and second moments were calculated to describe the texture content of the image. the authors compared their technique with some other related works (color moments, color correlogram) on an image database of 10000 images using 200 queries. they reported the supremacy of color texture moments-based method. 2.2 wavelet-based techniques interest of computer vision community shifted to wavelet-based approaches when several physiological studies of the visual cortex suggested the use of multiscale analysis of the visual information by visual systems of primates. beck et al.1987 for example, found that the visual cortex can be modeled as a set of independent channels, each with a specific frequency and direction. smith and chang, 1994 proposed a method for classification and discrimination of texture based on energies of image sub-bands. they compared four image decompositions (5-level pyramidal wavelet decomposition, 4x4 uniform sub-band decomposition of the fourier transform, 4x4 dct sub-bands and a simple spatial partition of the image into 4x4 blocks). each sub-band was then represented by its mean and standard deviation. they found that wavelet-based decomposition and fourier techniques produced the best results and the spatial partition had the worst retrieval performance. m. kokare et al. 2007; kokare et al. 2005 and kokare et al. 2006 proposed a set of 2-d rotated filters to which a 5-level discrete wavelet transform was applied. each sub-band was represented by its energy and standard deviation. their approach produced a good characterization of diagonally oriented texture. they reported about 8% improvement of the retrieval rate compared with traditional wavelet transforms. recently, (huang et al. 2003; huang and dai, 2004 and huang et al. 2006) proposed a wavelet based approach that concatenates gradient vectors of the subband images to obtain a single feature vector called composite sub-band gradient (csg) vector. a gradient vector of an image is the histogram that records the total gradient magnitude of the image pixels at different directions. their experiments showed that their approach outperformed the single-resolution technique using the same texture features. some other recent works (wang and yong, 2008 and huang and avlyente 2008) took into account the correlations that exist between different sub-bands when selecting the texture features. wang and yong, 2008 identified the most correlated sub-bands and used a linear regression to estimate the relationship between these sub-bands. their feature vector consisted of the linear regression parameters (ai, bi) of correlated sub-bands as well as the means and standard deviations of the linear approximation errors. their technique produced better performance than classical methods such as pyramid structured wavelet transform, tree structured wavelet transform and gabor, and combination of these methods with glcm (grayscale concurrence matrix) on a dataset made of 40 brodatz images. k. huang and s. aviyente, 2008 used the correlation information to cluster the image sub-bands. the energy of the highest sub-band in each cluster was considered as texture feature. their method performed an effective selection of sub-bands since there was no loss of accuracy compared with the one obtained with the full set of sub-bands. z-he et al. 2009 presented a novel wavelet-based method which used non-separable wavelet filters. they claimed that their approach can capture more directions and edge information which leaded to better retrieval accuracy. recently, complex wavelets transform became a popular tool for texture characterization. this is because of its shift invariance property and its good directional selectivity over the traditional discrete wavelet transforms (kokare et al. 2006; selesnick, 2002, celik and tjahjadi, 2009 and vo and oraintara, 2009). m. kokare et al. 2005 proposed new 2d rotated complex wavelet filters that are non-separable; they are 45 degree rotated version of the complex wavelet filters proposed by selesnick, 2002. the authors conducted experiments on two sets of images (116x16 and 40x16) and reported an improvement in accuracy compared to traditional wavelets and gabor-based method especially when combining the new filters with the dual tree complex wavelet transform since this combination allowed characterization of 12 directions instead of three by traditional real wavelets, and six in most gabor wavelets and complex wavelets techniques celik and tjahjadi, 2009 described a multi-scale texture retrieval classifier that used a gabor-like function and dual-tree complex wavelet transform with 3 scales and 6 directions. the texture feature vector consisted of the variances and entropies of the sub-bands resulting from the transform. their experiments on 24 texture images (six from brodatz album and six from the mit vistex database) showed that the new approach outperforms the traditional discrete wavelet transform and is also robust against noise. 51 the journal of engineering research vol. 7 no. 2 (2010) 48-58 3. proposed methods the proposed single-resolution method is partly inspired from previous energy-based works (tsai and tseng, 1999 and weszka and dyer, 1976). like those methods a discrete fourier-transform is first applied to the original image f(x,y) to obtain the transformed image f(u,v). the discrete 2-d fourier transform f(u,v) of an mxn image f(x,y) is given by the following eq. (1): (1) like those approaches also, the frequency domain is partitioned into several regions. the difference with those works is that sectors are used instead of rings and wedges, where a sector is the intersection of a wedge and a ring; see fig. 1. the advantage of using sectors is the fact that a more accurate description of the power spectrum distribution in the frequency domain can be obtained. each sector characterizes a range of orientations and some levels of coarseness of the texture. moreover, unlike the above mentioned works, the proposed approach takes advantage of the symmetry property of fourier transform and limited the analysis to only half of the frequency space which reduces the overall processing time. let {f(x,y), x=1,n , y=1,n} be the texture image and {f(u,v), u=1,n, v=1,n} be its fourier transform. first, the origin of the transformed image is shifted to the center of the image at position (n/2+1, n/2+1). this will produce a symmetrical fourier image. because of this symmetry, only half of the image is processed to extract texture features. in this implementation the right half of the image is considered; it's defined by eq. (2): (2) the spectral domain is partitioned into half-ringshaped regions and wedge-shaped regions where: a half-ring-shaped region rr1,r2 is defined as follows (3): (3) a wedge-shaped region is defined as follows (4) (4) the intersections of these rings and wedges define sectors sr1,r2 1,2 , as shown in formula (5) in order to reduce the effect of noise, only the significant values of the power spectrum are considered for further processing. the mean and standard deviations of the significant values of power spectrum in these sectors constitute the feature vector. such a feature vector describes three aspects of the texture: its roughness, coarseness and directionality. as for texture roughness, (gibson and gaydecki, 1995) indicated that rough textures tend to contain more energy in the high frequency components than smoother ones. regarding texture coarseness, (tsai and tseng, 1999) indicated that magnitude of power spectra for frequency components away from the origin drop rapidly to approximately zero for coarse textures. finally, it is 0well known that linear structures in the spatial domain at direction produce linear structures in the frequency domain with high energy values at direction +90°. 3.1 feature extraction algorithm (single-resolution) the feature extraction algorithm used in the singleresolution method can be summarized as follows: 1. apply fast fourier transform to original image f(x,y) to get f(u,v) (ie. f(u,v)=fft(f(x,y), where fft = fast fourier transform) 2. consider only pixels with significant power spectrum values (ie. f(u,v)=0 if f(u,v)< threshold) 3. partition the fourier image into n sectors (ie. configure 1. the right half of the frequency domain is partitioned into 12 sectors each representing a specific direction and frequency range. note how “vertical” wedges (1, 5 and 9) are each made of two 22.5o wide wedges (5) 52 the journal of engineering research vol. 7 no. 2 (2010) 48-58 sider only (u,v) satisfying formula (5)) 4. calculate means ( i) and standard deviations ( i) of the n sectors. the feature vector is defined as follows: fv= ( 1, 1 2, 2,… n, n). few local multi-resolution fourier-based methods were proposed for texture analysis. they have mainly applied windowed fourier transform of different sizes (resolutions) to the original image and extracted textural features from those transformed images. the proposed technique adopts a global multi-resolution approach in the sense it applies fourier transform to images of different sizes and the textural features are extracted from the various transformed images. therefore, the new multi-resolution method is just an extension of the single-resolution method with few adjustments introduced to prevent any significant deterioration of the processing time. experiments we have conducted showed that the use of mean value alone instead of mean and standard deviation as a discriminating feature does not deteriorate the retrieval performance and it slightly improves the processing time. they also showed that the use of sectors for multi-resolution approach does not improve significantly the performance as when wedges and rings are used but it slightly increases the processing time. taking into account these two results, the following feature extraction algorithm for our multi-resolution method is adopted. 3.2 feature extraction algorithm (multi-resolution) the feature extraction algorithm can be summarized as follows 1. apply fast fourier transform to original image f0(x,y) to get f0(u,v)(i.e. f0(u,v)=fft(f0(x,y)) 2. set resolution level k to 0; 3. while k < max_level do a. consider only pixels with significant power spectrum values (ie. f(u,v)=0 if f(u,v)< threshold). b. calculate rk, cumulated energies (means) of the right half ring-shaped regions and normalize them (ie. rk= (r1, r2, …rn), where ri= normalized sum of energies for ring-shaped region i). c. calculate wk, cumulated energies (means) of right wedge-shaped regions and normalize them (ie. wk= (w1,w2,..wm), where wi=normalized sum of energies for wedge-shaped region i). the feature vector for level k is therefore defined as fvk=(rk,wk). d. calculate the next resolution image by halving image sizes and applying an interpolation technique on pixel values (nearest, bilinear, or bicubic). e. k=k+1. 4. the feature vector fv is defined as: fv=(fv1, fv2,…fvl), where l is the maximum number of scales considered. 3.3 similarity measurement given two k-dimensional feature vectors f1 and f2, representing two images im1 and im2, the dissimilarity between im1 and im2 can be estimated using various distance metrics. the simplest and most popular one is the euclidean distance. this metric assumes all the components of the feature vector are of the same level of magnitude which is not the case for the feature vectors of the proposed techniques. lower resolution images have low energies. a normalized metric that reduces the bias towards high-magnitude components is needed. few of such metrics were tested; the following two were selected because of the good retrieval results they produced (formula 6 and 7): (6) this metric is known in literature as caneberra metric; it is used to compare feature vectors produced by the single-resolution technique. (7) the metric defined by eq. (7) is used to compare feature vectors produced by the multi-resolution technique because the experiments we have conducted showed that the use of this metric leads to a slightly better results than when caneberra metric is used. 4. experimentation 4.1 parameter setting in our experimentation the fourier domain of an nxn image is partitioned into three half ring-shaped regions rr1,r2 (nrings=3) and four1 wedge-shaped regions w 1, 2 (nwedges=4 ) where {(r1,r2)}={(0,n/6),(n/6,n/3),(n/3,n/2)}and {( 1, 2)}={(-22.5°,22.5°),(22.5°,67.5°),(-22.5°,67.5°), (67.5°, 90°), (-67.5°, -90°)}. this choice is driven by the results of the preliminary testing involving different values for nrings and nwedges. the experiments showed that beyond 6 wedges and four rings results start degrading and the best results are obtained when (nrings= 3 or 4 and nwedges=4 or 6). the setting producing smaller feature vector is chosen in order to ensure a faster execu __________________ 1 the four wedge-shaped regions are wr0 that spans the range (-22.5o, 22.5o), wr1 that spans the range(22.5o, 67.5o), wr2 that spans the range (22,5o, -67.5o), and wr4 that spans the two rabge (67.5o, 90o), and (-67.5o, -90o) 53 the journal of engineering research vol. 7 no. 2 (2010) 48-58 tion time. the three half-ring-shaped regions segment the fourier domain into low, medium and high frequencies; their power spectrums are strongly related to roughness and coarseness of the texture. in the other hand, the power spectrums of the four wedge-shaped regions are closely related to the amount of the linear structures in the four directions: horizontal, vertical and the two diagonals. therefore, the feature vector of the single-resolution approach consists of 24 elements containing the mean and standard deviations of the twelve sectors sr1,r2, 1, 2 shown in fig. 1 and the feature vector of the multi-resolution approach consists of 7xl elements containing the means of the three half-rings and the four wedges of the l images representing l resolutions. see fig. 2. the value of the threshold defining significant power spectrum values is set to the mean of the values of the transformed image by fft. the experiments showed that this value gives good results with all the images that have been tested. 4.2 test dataset the dataset used in the experiments is made of 79 grayscale images selected from the brodatz album we downloaded from: [http://www.ux.uis.no/~tranden/brodatz.html]. images that have uniform textures (ie. similar texture over the whole image) were selected. all the images are of size 640 x 640 pixels. each image is partitioned into 8x8 non-overlapping sub-images from which 4 sub-images are chosen to constitute the image database (ie. database= 316 images) and one to be used as a query image (ie. 79 query images). 4.3 hardware and software environment we have conducted all the experiments on an intel core 2 (2ghz) laptop with 1 gb ram. the software environment consists of ms windows xp professional and matlab7. 4.4 performance evaluation to evaluate the performance of the proposed approaches, we have adopted the well-known efficacy formula (8) introduced by kankahalli et al. 1996. (8) where n is the number of relevant images retrieved by the cbir system, n is the total number of relevant images that are stored in the database, and t is the number of images displayed on the screen as a response to the query. in the experimentation that has been conducted n=4, and t=10 which means efficacy=n/4; the average time needed to perform feature extraction and compare two images is also recorded. it's the average time needed for comparing each of the 79 query images with 316 images in the database. this value is used as a performance indicator of the speed of the methods under investigation. the first set of experiments is designed to compare the results obtained by the proposed single-resolution method with two classical single resolution fourierbased techniques. the purpose of this experiment is to show the improvement obtained by adopting a sectorbased segmentation of the frequency domain over the classical ring/wedge-based segmentation. both of the fourier-based methods partition the fourier space into 4 wedges and 3 rings. the first one (referred to as cf(µ)) uses mean of the power spectrum of the resulting regions as texture feature. the second one (referred to as cf( )) uses standard deviation of the power spectrum of the resulting regions as texture feature. the third one (referred to as cf(µ, )) uses both the mean and standard deviations as texture feature. the second set of experiments was designed to test the performance of the new multi-resolution method. the first experiment is designed to evaluate the effect of the type of interpolation technique (nearest, bilinear and bi-cubic) on both the accuracy and execution time of the proposed method. the second experiment is designed to compare the performance of the multi-reslow frequency ring-shaped area mid-range frequency ring-shaped area high frequency ring horizontal (0o) wedge-shaped region vertical (90o) wedge-shaped region (=2 areas) diagonal (+45, -45) wedge-shaped regions figure 2. the seven regions (3 rings and 4 wedges) considered for each resolution level in the frequency domain and their significance (multi-resolution approach) 54 the journal of engineering research vol. 7 no. 2 (2010) 48-58 olution approach (referred to as mrfft) with the single-resolution one (referred to as srfft) and obtained five other multi-resolution methods. these methods are: dual-tree complex wavelet using means and standard deviations of the sub-bands similar to the one described in (celik and tjahjadi, 2009) (referred to as dtcw(µ, )), rotated wavelet filters proposed by (kokare et al. 2007) (referred to rwf (µ, )), traditional gabor filters using means and standard deviations of the different sub-bands as described in zhang et al. 2000 (referred to as gabor (µ, )), 3-level wavelet-based method using energy of sub-bands as feature descriptor similar to the one proposed by (wouwer et al. 1999) (referred to as mrwedb4) and csg method similar to the one proposed by (huang et al. 2006) (referred to as csgdb10). 5. results and discussion table.1 summarizes the results of the first set of experiments designed to evaluate our single-resolution method. the table shows that the proposed method outperforms the classical fourier methods in terms of accuracy. (there is between 4% and 14% improvement). this can be explained by the fact that partitioning frequency domain into sectors describes more accurately the distribution of the power spectrum in the fourier space than when using rings and wedges. the table also indicates that this gain in accuracy is obtained at the expense of an increase in processing time required to build and compare the twelve-element-vectors of the new method (there is about 43% increase in processing time). but it is worth noting that the extra time is not exorbitant since the total time is still less than many current retrieval techniques (as shown in table 3). cf(µ) = classic fourier method using mean cf( ) = classic fourier method using standard deviation cf((µ, )= classic fourier method using mean and standard deviation srfft = our single-resolution technique using µ and of sectors table. 2 summarizes the results obtained in the experiment designed to identify the most appropriate interpolation technique to be used by our multi-resolution method. as expected the nearest interpolation technique is the fastest among the three approaches; but with 80% accuracy it is far behind the two other techniques. bilinear interpolation technique is the best in terms of accuracy (95%) and it requires less execution time than the bi-cubic technique (about 10% faster). for these reasons bilinear interpolation was used in the proposed multi-resolution technique. table. 3, indicates that the use of a multi-resolution approach improves the retrieval performance by 2% (from 93% to 95%) when compared to the single resolution approach; the price to pay for better performance is an increase in the processing time (about 73% increase), though the new time is still reasonable. table. 3 also shows clearly that, in terms of retrieval accuracy, the proposed multi-resolution method outperforms similar multi-resolution wavelet-based methods described in literature namely pyramidal waveletbased method (mrwedb4), composite sub-band gradient method (csgdb4), and rotated wavelet filters (rwf(µ, )). with 95% accuracy, the new multi-resolution method is as good as the two other methods (ie. dtcw (µ, ) and gabor (µ, )) but with a better execution time (classical gabor(µ, ) is 12 times slower and dtcw(µ, ) 56% slower). the best in terms of processing time is the classical wavelet-based technique (2 times faster than the proposed method) but with lower efficacy rate. figure 3 shows the first 10 retrieved images obtained by our single-resolution method and the three classical fourier methods, for the query sub-image d96-6. figure 4 shows the retrieval results for the same query image obtained by our multi-resolution method and four other techniques (namely dtcw (µ, ), gabor (µ, ), rwf and csgdb10). table.1: comparing our single resolution technique with two classical fourier methods table 2. effect of interpolation technique on accuracy and time of mrft* table 3. comparing performance of our multiresolution method and some other multiresolution techniques 55 the journal of engineering research vol. 7 no. 2 (2010) 48-58 figure 3. retrieval results for our single resolution method and three versions of the classical. retrieved images are sorted by decreasing value of similarity score from left to right and top to bottom 56 the journal of engineering research vol. 7 no. 2 (2010) 48-58 figure 4. retrieval results for our multi-resolution method and four other techniques included in the experimentation. retrieval images are sorted by decreasing value of similarity score from left to right and top to bottom 57 the journal of engineering research vol. 7 no. 2 (2010) 48-58 6. conclusion and future work this paper describes a new single-resolution fourierbased method for retrieving texture images. the experiment that has been conducted shows that the new method outperforms several existing fourier-based methods. it also describes a multi-resolution version of this method that produces better retrieval results than the single resolution approach at the expense of some additional processing time. the multi-resolution method outperforms several existing multi-resolution methods. currently, two possible improvements to the proposed technique are under investigation: a. studying needed modifications for making proposed methods rotation-invariant and b. exploring ways to perform image segmentation based on the fourier features used by the proposed methods. references amadasun, m. and king, r., 1989, "textural features corresponding to textural properties," ieee smc, vol. 19, pp. 1264-1274. beck et al. 1987, " spatial frequency channels and perceptual grouping in texture segregation," comp. vision graphics and image processing, vol. 37, pp. 299-325. bianconi, f. and fernandez, a., 2007, " evaluation of the effects of gabor filter parameters on texture classification,” pattern recognition, vol. 40, pp. 3325-3335. campbell, f.w. and robson, j.g., 1968, " application of fourier analysis to the visibility of gratings," journal of physiology, vol. 197, pp. 551-556. celik, t. and tjahjadi, t., 2009, "multiscale texture classification using dual-tree complex wavelet transform," pattern recognition letters, vol. 30, pp. 331-339. chen, c.c. and chen, c.c., 1999, "filtering methods for texture discrimination," pattern recognition letters, vol. 20, pp. 783-790. conners, r.w. and harlow, c.a., 1980, "a theoretical comparison of texture algorithms," ieee trans. pattern analysis and machine intelligence (pami), vol. 2, pp. 204-222. daugman, j.g. and kammen, d.m., 1987, "image statistics. gases, and visual neural primitives," proc. of ieee icnn 4, pp. 163-175. fountain, s.r. and tan, t.n., 1998, "efficient rotation invariant texture features for contentbased image retrieval," pattern recognition, vol. 31, pp. 1725-1732. gibson, d. and gaydecki, p.a., 1995, " definition and application of a fourier domain texture measure: application to histological image segmentation," comp. biol. vol. 25, pp. 51-557 haralick, r.m., shanmugam, k. and dinstein, j., 1973, "textural features for image classification," ieee trans. systems, man and cybernetics, vol. 3, pp. 610-621. huang, k. and aviyente, s., 2008, "wavelet selection for image classification," ieee trans. on image processing, vol. 17, pp. 1709-1720. huang, p.w. and dai, s.k. 2004, "design of a twostage content-based image retrieval system using texture similarity," information processing and management vol. 40, pp. 81-96. huang, p.w. and dai, s.k., 2003, " image retrieval by texture similarity," pattern recognition, vol. 36, pp. 665-679. huang, p.w., dai, s.k. and lin, p.l., 2006, "texture image retrieval and image segmentation using composite sub-band gradient vectors," j. vis. communication and image representation, vol. 17, pp. 947-957. jain, a.k. and farrokhnia, f., 1991, " unsupervised texture segmentation using gabor filters," pattern recognition, vol. 24, pp. 1167-1186. kankahalli, m., mehtre, b.m. and wu, j.k. 1996, "cluster-based color matching for image retrieval," pattern recognition, vol. 29, pp. 701708. kokare, m., biswas, p.k. and chatterji, b.n., 2007, "texture image retrieval using rotated wavelet filters," pattern recognition letters, vol. 28, pp. 1240-1249. kokare, m., biswas, p.k. and chatterji, b.n., 2005, " texture image retrieval using new rotated complex wavelet filters," ieee trans. on systems, man, and cybernetics b, vol. 35, pp. 1168-1178. kokare, m., biswas, p.k. and chatterji, b.n., 2006, " rotation-invariant texture image retrieval using rotated complex wavelet filters, ieee trans. on systems, man, and cybernetics vol. b 36, pp. 1273-1282. kokare, m., biswas, p.k. and chatterji, b.n., 2005, " texture image retrieval using new rotated complex wavelet filters," ieee trans. on systems, man, and cybernetics, vol. b 35, pp. 1168-1178. selesnick, i.w., 2002, " the design of approximate hilbert transform pairs of wavelet bases," ieee trans. signal processing, vol. 50, pp. 1144-1152. smith, j.r. and s-f., 1994, "transform features for texture classification and discrimination in large image databases," image processing proceedings, icip-94, vol. 3, pp. 407-411. 58 the journal of engineering research vol. 7 no. 2 (2010) 48-58 tsai, d.m. and tseng, c.f., 1999, " surface roughness [6] j.s. classification for castings," pattern recognition vol. 32, pp. 389-405. vo, a. and oraintara, s., 2009, a study of relative phase in complex wavelet domain: property, statistics and applications in texture image retrieval and segmentation," signal processing image communication. wang, z.z. and yong, j.h., 2008, "texture analysis and classification with linear regression model based on wavelet transform," ieee trans. on image processing, vol. 17, pp. 1421-1430. weszka, c.r., dyer, a. and rosenfeld., 1976, " a comparative study of texture measures for terrain classification," ieee trans. system, man and cybernetics, vol. 6. pp. 269-285. wouwer, g.v., scheunders, p. and dyck, d.v., 1999, "statistical texture characterization from discrete wavelet representations," ieee trans. on image processing, vol. 8, pp. 592-598. yu, h., li, m., zhang, h.j. and feng, j., 2002, "color texture moments for content-based image retrieval," international conference on image processing proceedings icip02, pp. 24-28. zhenyu h., xinge y. and yuan y., 2009, "texture image retrieval based on non-tensor product wavelet filter banks," signal processing, vol. 89, pp. 1501-1510. zhang, d., wong, a., indrawan, m. and lu, g., 2000, “content-based image retrieval using gabor texture features”, proc. pacific-rim conference on multimedia, pp. 392-395, sydney, australia. ece061119.qxd the journal of engineering research vol. 5, no.1, (2008) 71-78 1. introduction studies and experience have shown that load model parameters can have a significant effect on the results of dynamic performance and voltage stability of power systems (millanovic, et al. 1995; langevin, et al. 1986; ellithy, et al. 1989; choudhry, et al. 1990; choudhary, 1986; ellithy, et al. 1997 and craven, et al. 1983; xu, et al.1994, vaahedi, el al. 1988; alden, et al. 1976; ellithy, et al. 1997). incorrect parameters of a load model could lead to a power system operating in modes that result in actual system collapse and separation (craven, et al. 1983; xu, et al. 1994). accurate load model parameters are, therefore, necessary to allow more precise calculations of power system control and stability limits which are critical in the planning and operation of a power ___________________________________________ *corresponding author e-mail: mohahelal39@hotmail.com system dynamics has long been recognized, and it has become clear that assumptions regarding load model parameters can impact predicted system performance. several efforts have been devoted to load modeling and evaluation of load parameters through field measurements (xu, et al. 1997; ohyama, et al. 1985). analytical approaches to constructing accurate load models have also been considered (xu, et al. 1997; berg, et al. 1973). voltage-dependent load models for composite load representation are highly recommended by the ieee working group (ieee task force for dynamic performance, 1995) and many utilities (xu, et al. 1997; ohyama, et al. 1985; concordia, et al. 1982). flexible ac transmission system (facts) controllers based on thyristor controlled reactors (tcrs), such as static var compensators (svcs) (ieee special stability control working group, 1994)) are being used by several tuning of svc stabilizers over a wide range of load parameters using pole-placement technique k. ellithy department of electrical engineering, qatar university, doha, qatar received 19 november 2006; accepted 1 april 2007 abstract: this paper investigates the effect of typical load model parameters on the static var compensator (svc) stabilizers tuning. a proportional-integral (pi) type stabilizer is considered and its gain-settings are tuned using the pole-placement technique to improve the damping of power systems. tuning of svc stabilizers (damping controllers) traditionally assumes that the system loads are voltage dependent with fixed parameters. however, the load parameters are generally uncertain. this uncertain behavior of the load parameters can de-tune the gains of the stabilizer; consequently the svc stabilizer with fixed gain-settings can be adequate for some load parameters but contrarily can reduce system damping and contribute to system instability with loads having other parameters. the effect of typical load model parameters on the tuning gains of the svc pi stabilizer is examined and it is found the load parameters have a considerable influence on the tuning gains. the time domain simulations performed on the system show that the svc stabilizer tuned at fixed load parameters reduce the system damping under other load parameters and could lead system instability. keywords: load model parameters, svc stabilizer, pole-placement üé£b’g ™°vh á«æ≤j ωg~îà°séh ∫éªy’g øe áø∏àfl º«≤d ô≤à°ùÿÿg ∫é©ødg òñec’g âdƒødg ¢vƒ©e hp âñãÿg ∞«dƒj »ã«∏dg ~dén áá°°uuóóÿÿgg âñãÿg §ñ°v ≈∏y ∫éªmc’g êpé‰ ‘ áø∏àıg º«≤dg òkéj ádé≤ÿg √òg ¢sq~j :tuning) (stabilizers~bh .»µ«jéà°s’g ádé©ødg ò¨dg iƒ≤dg ¢vƒ©ã ¢uéÿg êpƒªædg üé£bg øcéeg qé«ànéh é¡jóeé©e §ñ°v ºàj å«m πeéµàdg h ö°séæàdg ≈∏y »æñÿg ºµëàdg á«yƒf ωg~îà°sg ”placement)(poleú°ù– ¤g …odƒj »àdg øcée’g ‘ iƒ≤dg º¶f ‘ oéªn’gpower systems) (damping ofqéñày’g ‘ òn’g öéj ¬fg ~«h .~¡÷g ≈∏y i~ªà©e ∫éªm’g ¿g ¢vôøh á°sgq~dg √òg ºàj √oé©dg ‘h …odƒj ~b πh iông ägò¨àe ägp ∫éªmcg oƒlh ™e ¬àhéãdg äóeé©ÿg hp ºµëàÿg agog ∞©°v öñ°ùj ‹éàdéhh §ñ° dg ±gôëfg ¤g …odƒj é‡ .io~fi òz á≤jô£h ±ô°üàj ∫éªm’g ¿g »µ«jéà°s’g ádé©ødg ò¨dg iƒ≤dg ¢vƒ©e ‘ ºµëàdg äóeé©e ᪫b §ñ°v ≈∏y ∫éªm’g êpé‰ ‘ ägò¨àÿg º«b òkéàd á°sgqo ádé≤ÿg ω~≤jh .∫éªm’g ¢ ©h ~æy qgô≤à°sg ω~y ¤g ∫éªm’g ò¨j ~æy idhgog π∏≤j áàhéãdg äóeé©ÿg hp ºµëàÿg ¿g ¿é«h ” å«m âbƒdg ™e ωé¶ædg agog á°sgqo ” ~bh .òñc òkéj é¡d ¿g á°sgq~dg úñjh πeéµàdg h ö°séæàdg ≈∏y »æñÿg .∫éªmc’g ¢ ©h ~æy qgô≤à°sg ω~y öñ°ùj ~b áfg éªc áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.üé£b’g ™°vh -»µ«jéà°s’g ádé©ødg ò¨dg iƒ≤dg ¢vƒ©e äéàñãe ∞«dƒj ,∫éªm’g êpé‰ º«b : 72 the journal of engineering research vol. 5, no.1,( 2008) 71-78 utilities to support the voltage and to increase the capacity of their systems. svcs with additional signals (stabilizing signals) in their voltage control loops have been used to improve the damping of power system electromechanical oscillations and to enhance system stability. the additional signals that are generally used in stabilizers (damping controllers) are rotor speed and bus frequency deviations. in recent years, many designs for svc stabilizers have been proposed to damp out the electromechanical oscillation mode in power systems (cheng, et al. 1992; hsu, et al. 1988; lee, et al. 1994; hamoud, et al. 1987; el-saady, et al. 1998; hammad, et al. 1984; elmetwally, et al. 2003; fang, et al. 2004; yu, et al. 2001; farsangi, et al. 2004). the basic limitation of these designs and/or tuning of svc stabilizers are that the influence of load model parameters has not been taken into account. almost all of these svc stabilizers is based on a constant impedance load (fixed load parameters). the constant impedance load representation is not accurate and is not a good approximation in view of the strong influence of the load voltage sensitivity on the dynamic performance of the power system. the parameters of typical loads vary seasonally, and in some cases change over day. consequently, the svc stabilizers tuned under constant impedance load model may become unacceptable under other load-model parameters. the subject of this paper is to investigate important aspects related to the effect of loads and their parameters uncertainty on tuning of svc stabilizers. the author has initially addressed this problem in (ellithy, et al. 1989; choudhry, et al. 1986). proportional-integral (pi) type stabilizer is considered as the additional stabilizer with the svc. the gain-settings of svc pi stabilizer are determined using pole-placement (eigenvalue-placement) technique to improve the damping of electromechanical oscillations mode in power systems. the interaction between typical load parameters and the tuning gains of the svc stabilizer is investigated. finally, the time domain simulations of the system under disturbance conditions are performed to demonstrate the effect of the load parameters on tuning of the svc pi stabilizer. 2. power system under study the power system under study is a synchronous generator connected to a large power system, of which the single-line diagram is shown in fig. 1. the generator is equipped with ieee type-1 excitation system. full order model (7th order model) of the generator is utilized in the analysis and simulations. the load is connected at a generator terminal and the svc is connected at the mid-point of transmission line. the system parameters and nominal operating point values are given in appendix. 2.1 model of svc with additional stabilizer the thyristor-controlled reactor (tcr) type svc (chang, et al. 1992; lee, et al. 1988; ieee special stability control working group 1994; el-metwally, et al. 2003), shown in fig. 1, is used in the present study. the model of the svc with additional proportional-integral (pi) stabilizer is shown in fig. 2. the stabilizer uses the generator speed deviation (∆ω ) as a feedback signal to generate the auxiliary stabilizing signal ∆vs (a stabilizer output signal) to the svc. the signal ∆vs is added to the main input of the svc to damp out the electromechanical oscillations mode. the signal ∆vs causes fluctuations in the svc suceptance and, hence, in the bus voltage. if the svc stabilizer is tuned correctly the voltage fluctuations act to modulate the power transfer to damp out the electromechanical oscillations mode. the equation of the svc controller (fig. 2) is given by (1) where the auxiliary stabilizing signal (svc stabilizer output signal) ∆vs is given as (2) where kp and ki are the svc stabilizer gain settings. the firing control system of the thyristors is represented by a single time constant tα and gain kα. the wash out circuit is introduced in the stabilizer to assure no permanent effect in the terminal voltage due to a prolonged error in the low frequency that might occur in an overload and to assure that the wash out circuit will not have any effect on the phase shift or gain on the low frequency. the variable inductive susceptance bl of svc is a function of the thyristor firing angle α and is given by (3) vt v xe xe v∞ gen svc load control system b figure 1. power system single-line diagram l s ref k b ( v v v) st α α ∆ = ∆ + − +1 w i s p w st k v k st s ⎛ ⎞⎛ ⎞⎟⎜ ⎟⎜⎟∆ = + ∆ω⎜ ⎟⎜⎟ ⎟⎜ ⎟⎜⎟⎜ ⎝ ⎠+⎝ ⎠1 fig. 2 block di agram of svc wi th p i stabilizer pi stabilizer kα 1 + stα + σ vref v + ∆vs vs,min vs,max blmax blmin + + σ bl0 bl ∆ω + stw (kp ) 1 + stw ki s pi stabilizer figure 2. block diagram of svc with pi stabilizer l s ref k b ( v v v ) st α α ∆ ∆= + − +1 w i s p w st k v k st s ∆ ∆ω ⎛ ⎞⎛ ⎞⎟⎜ ⎟⎜⎟= +⎜ ⎟⎜⎟ ⎟⎜ ⎟⎜⎟⎜ ⎝ ⎠+⎝ ⎠1 ( )− + = − 2 2 2 l s sin b x π α α π ; ≤ ≤2/π α π 73 the journal of engineering research vol. 5, no.1, (2008) 71-78 where xs is the reactance of the svc fixed inductor. 2.2 load model this paper follows the recommendation of the ieee working group (ieee task force on load representation for dynamic performance 1995) and utilities (xu, et al. 1997; ohyama, et al. 1985) in utilizing the voltagedependent load model for composite load representation. utilities normally perform field tests, or in some cases perform regression analysis to establish system load models to be used for power-flow and stability studies. these models are in the form of (4) where pl and ql are the load active and reactive power; vt is the load bus voltage; np and nq are the load parameters.; pl0, ql0, and vt0 are the nominal value of load active power, load reactive power, and bus voltage prior to a disturbance. for small disturbance studies of system damping, the linearized version of eq. (4) is given by (5) where the load representation given in eq. (4) makes it possible the modeling of all typical voltage-dependent load models by selecting appropriate values of the load parameters np and nq. the values of np and nq depend on the nature of the load and can vary between 0 to 3.0 for np and 0 to 6.0 for nq. the load parameters of the composite load (industrial, commercial and residential loads) can be determined by the following equations (ohyama, et al. 1985; berg 1973): (6) where pli and qli are the active and reactive power of ith component and np, nq are their load parameters. the measurement values of the parameters (np, nq) of various kinds of typical power system composite loads are reported in (xu, et al. 1997; ohyama, et al. 1985; concordia, et al. 1982). 3. design of svc pi stabilizer the gain-settings (kp and ki) of the svc pi stabilizer are determined using the pole placement by moving the eignvalues associated with the electromechanical oscillations mode to a prescribed value on the left-half complex plane. it is well known that improving the damping of these oscillations mode can enhance the damping characteristic of a power system. the design procedures and their associated results are given below. 3.1 system without svc stabilizer in the design of the svc stabilizer using the poleplacement technique, the nonlinear equations of the power system are first linearized around an operating point to obtain the state-variables model of the system. in the present study, the state-variables model of the system is obtained using the component connection model (ccm) technique (ellithy, et al. 1989; choudhry, et al. 1986). the equations describe the state-variable model given in (el-metwally, et al. 2003). the state-variables model of the system is expressed as (7) where x = (∆iq ∆id ∆iq ∆ikkq ∆ikkd ∆ifd ∆δ ∆ω ∆vf ∆efd ∆vr ∆bl)t is the state-variables vector. the state variables x1 to x7 correspond to the generator, x8 to x11 correspond to the excitation system and x12 corresponds to the svc. y = ∆ω, the output signal. u = ∆vs, the control signal (stabilizer output signal) to the svc. the system eigenvalues without svc pi stabilizer (open-loop system) for fixed load parameters np = nq = 2 (constant impedance load) are listed in the first column of table 1. the eigenvalues associated with the electromechanical oscillations mode of the synchronous generator are depicted by the complex pair eigenvalues λ6,7. the damping ratio ζ of this poorly damped oscillations mode without the svc stabilizer (λ6,7 = σ ± jβ = −0.67896 ± j11.362) is given as ⎛ ⎞⎟⎜ ⎟= ⎜ ⎟⎜ ⎟⎜⎝ ⎠ 0 0 n p t l l t v p p v ⎛ ⎞⎟⎜ ⎟= ⎜ ⎟⎜ ⎟⎜⎝ ⎠0 0 nq t l l t v q q v = 0 0 l l p t t p p n v v ∆ ∆ = 0 0 l l q t t q q n v v ∆ ∆ t q q d d t v (v v v v ) v ∆ ∆ ∆= +0 0 0 1 ∑ ∑ = = = = 1 1 n nli p pi li i i p n n ( ); p p p ∑ ∑ = = = = 1 1 n nli q qi li i i q n n ( ); q q q = + dx ax bu dt =y cx withou t svc stabilizer with svc stabilizer ( kp=0.0, ki=0.0) (kp=19.892, ki=178.895) λ1,2 -212.210 ± j842.990 -212.700 ± j842.930 λ3 -98.782 -89.620 λ 4 -72.361 -79.480 λ 5 -38.992 -38.820 λ 6,7 -0.679 ± j11.362 -2 ± j11* λ 8,9 -9.893 ± j14.903 -9.070 ± j14.630 λ 10,11 -0.645 ± j0.837 -0.650 ± j 0.900 λ 12 -10.220 table 1. system eigenvalues at load model parameters np = nq = 2 *: exact pole placement 74 the journal of engineering research vol. 5, no.1, (2008) 71-78 the damping ratio ζ = 6% for the electromechanical mode is not good enough. the poor damping of this mode can be also seen from the system time response in figs. 3 and 4. the eigenvalues for this mode should be shifted to more desirable locations by the svc stabilizer (ie. the svc stabilizer is needed to improve the damping of this mode). the tuning gains of the svc stabilizer are described in the following section. 3.2 determination of svc stabilizer gains using pole placement technique the gain-settings (kp and ki) of the svc stabilizer will be determined to improve the damping ratio of the electromechanical mode by shifting the eigenvalues λ6,7 to desired locations. an expression for the gains has been derived by the author in ellithy (1997) and is given by (8) λ6= − σ − jβ is the desired eigenvalue location associated with the electromechanical oscillations mode. d = c [λ6 i a]-1 b where a, b and c system matrices are given in (7). if the eigenvalues λ6,7 = -2 ± j11 (ie. the damping ratio of the electromechanical mode ζ = 18%) are selected at the desired locations, then the gains kp and ki for the svc stabilizer can be computed using (8). the results are given in table 1. the eigenvalues of the closed-loop system (system with the svc stabilizer) are given in the second column of table 1. considerable improvement in the system damping can be expected in view of the closed-loop eigenvalues. the improvement in the damping can also be seen from the system dynamic performance shown in figs. 3 and 4. 4. effect of load parameters on svc stabilizer tuning the tuned gains (kp=19.892and ki=178.895) of the svc stabilizer at the parameters np = nq = 2 are used to check the damping characteristics and stability of the system under different load model parameters. based on these fixed gains, tuned at load parameters np = nq =2, the damping of the electromechanical oscillations mode is reduced under other load parameters as shown in figs. 5, 6, 7 and 9. the tuned gains of the svc stabilizer at other different load parameters are also used to check the system stability under different load parameters. based on these fixed gains, the system damping is reduced and the system may become unstable under other load param − − without svc stabilizer ⎯ with svc stabilizer tuned at np=nq=2 0 1 2 3 4 5 6 7 8 9 10 -8 -6 -4 -2 0 2 4 6 8 10 x 10 -3 tim e (s ec ) g e n e ra to r s p e e d d e v ia ti o n ( p .u ) load param eters np= nq= 2 figure 4. the system dynamic response at load parameters np = nq = 2 for 50% change in tm -without svc stabilizer _ without svc stabilizer tuned at np = nq = 2 x % σ ζ σ β = = +2 2 100 6 0 0.5 1 1. 5 2 2.5 3 3.5 4 4. 5 5 -8 -6 -4 -2 0 2 4 6 8 10 x 10 -4 time (sec ) g e n e ra to r s p e e d d e v ia ti o n ( p .u ) load parameters np= nq=2 -without svc stabilizer _ with svc stabilizer tuned at np = nq = 2 figure 3. the system dynamic response at load parameters np = nq = 2 for 5% change in tm = −2 21 11 22 11 22 b a b k ga ga a ; = −1 12 11 11 i p b k a k a a where: = − 21 12 2 21 1 a a g a ; ( )( ) =1 real part h b s ( )( ) =2 imaginary part h b s = −1m tωσ ; = − 2 11a t mωβ σ ; =12a m ; = − −22a m tωβ αβ ; =22a tωβ ; = 1 h d ( ) = +2 2 2 t s m t ω ωβ 75 the journal of engineering research vol. 5, no.1, (2008) 71-78 eters as shown in figs. 8 and 9. the decreases in the system damping are caused by the eigenvalues λ6,7, which are associated with the systems electromechanical oscillations mode. in order to improve the damping of the oscillations mode (ie. improving the damping ratio of the eigenvalues λ6,7) over a wide range of load model parameters, the svc stabilizer gains kp and ki must be tuned. the computed svc stabilizer gains (kp and ki) for typical load model parameters are given in figs. 10 and 11. these gains have been computed by (8) with the eigenvalues λ6,7 fixed at the desired locations of -2±j11. from these figures, it can be observed that the variations in the load model parameters (np, nq) have a considerable influence on the tuning of the svc stabilizer. while not reported in the paper, the author has also investigated the influence of load parameters when the load is located at the svc bus. the results obtained indicated no significant departure from the results presented here in so far as the influence of the load parameters on the tuning of the svc stabilizers. 5. conclusions this paper has examined the influence of voltagedependent load models on the effectiveness of the svc proportional-integral stabilizer for damping the electromechanical oscillations mode in power systems. -with svc stabilizer tuned at np = nq = 2 with svc stabilizer tuned at np = nq = 0 figure 5. the system dynamic response at load parameters np = nq = 0 for 50% change in tm - with svc stabilizer tuned at np = nq = 2 with svc stabilizer tuned at np = 0.4, nq = 4 figure 6. the system dynamic response at load parameters np = 0, nq = 4 for 5% change in tm - with svc stabilizer tuned at np = nq = 2 with svc stabilizer tuned at np = 1.2, nq = 2 figure 7. the system dynamic response at load parameters np=1.2, nq=2 for 5% change in tm - with svc stabilizer tuned at np = 0, nq = 6 with svc stabilizer tuned at np = nq = 2 figure 8. the system dynamic response at load parameters np = nq = 2 for 5% change in tm g en er at or s pe ed d ev ia tio n (p .u ) time-sec g en er at or s pe ed d ev ia tio n (p .u ) time-sec time-sec g en er at or s pe ed d ev ia tio n (p .u ) g en er at or s pe ed d ev ia tio n (p .u ) time-sec 76 the journal of engineering research vol. 5, no.1, (2008) 71-78 the impact of load model parameters on the svc stabilizer tuning gains obtained via pole-placement technique is investigated and it is shown that load models have remarkable influence on the stabilizer tuning gains. the results have also shown that the svc stabilizer tuned gains under specific load parameters could contribute to the worse damping of electromechanical oscillations and reduce system stability under other load parameters. in particular, simulation with constant active power loads shows system instability when the svc stabilizer is designed assuming constant impedance load. the results presented in this paper reinforce the need for including the load model parameters in the svc stabilizers tuning for damping the oscillations mode of power systems. acknowledgments the author would like to thank qatar university and qatar electricity & water company (qewc) for their support. references alden, r.t.h. and zein el-din, h.m., 1976, "effect of load characteristics on power system dynamic stability," ieee trans on power app. and syst. vol. pas-95, pp. 1762-1768. berg, g.j., 1973, "power-system load representation," proc. iee, vol. 120(3), pp. 344-348. cheng, c.h. and hsu, y.y., 1992, "damping of generator oscillations using an adaptive static var compensator," ieee trans. on power systems, vol. 7(2), pp. 718-725. choudhry, m.a, emarah, a.s., ellithy, k.a. and galanos, g.d., 1986, "stability analysis of a modulated ac/dc system using the eigenvalue sensitivity approach", ieee transactions on power systems, vol. pwrs-1(2), pp.128-137. choudhry, m.a., reza, m.a. and ellithy, k.a., 1990, "design of a robust modulation controller over a wide range of load characteristics for ac/dc systems," ieee trans. on power systems, vol. 5(1), pp. 212-218. concordia, c. and ihara, s., 1982, "load representation in power system stability studies," ieee trans. on power app. and syst, vol. pas-101, pp. 969-977. craven, r.h. and micheal, m.r., 1983, "load representation in the dynamic simulation of the queensland power system", j. of electrical and electronics engineering, vol. 3(1), pp. 1-7. - with svc stabilizer tuned at np = 0, nq = 4 with svc stabilizer tuned at np = nq = 2 figure 9. the system dynamic response at load parameters np = nq = 2 for 5% change in tm figure 10. the computed svc stabilizer gain k1 at different load model parameters figure 11. the computed svc stabilizer gain kp at different load model parameters g en er at or s pe ed d ev ia tio n in p er u ni t time-sec 77 the journal of engineering research vol. 5, no.1, (2008) 71-78 ellithy, k.a., 1997, "effect of voltage dependent load models on power system stabilizers design," fifth int. middle east power system conference (mepcon'05), vol. 1, pp. 370-376. ellithy, k.a. and choudhry, m.a., 1989, "effect of load models on ac/dc system stability and modulation control design," ieee trans. on power systems, vol. 4, pp. 411-418. ellithy, k.a. al-alawi, s.m. and zein el-din, h.m., 1997, "tuning power system stabilizers over a wide range of load models using an artificial neural network," j. of engineering and applied science, college of engineering, cairo university, vol. 44(2), pp. 389-406. el-metwally, k. and ellithy, k., 2003, "a self organizing fuzzy controller for svc stabilizers design in the presence of load characteristics uncertainty", ieee international middle east power conference (mepcon 2003), vol. 1, pp. 31-37. el-saady, g., el-sadek, m. and abo-elsaoud, m., 1998, "fuzzy adaptive model reference approach-based power system static var stabilizer", electric power system research, vol. 45, pp. 1-11. fang, d.z., yang, x., chung, t.s. and wong, k.p., 2004, "adaptive fuzzy-logic svc damping controller using strategy of oscillation energy descent" ieee trans. on power systems, vol. 19(3), pp. 1414-1421. farsangi, m.m., song, y.h. and lee, k.y., 2004, "choice of facts device control inputs for damping interarea oscillations," ieee trans. on power systems, vol. 19(2), pp. 1135-1143. hammad, a.e. and el-sadek, m., 1984, "application of a thyristor controlled var compensator for damping subsynchronous oscillations in power systems," ieee trans. power app. and syst, vol. pas-103, pp. 198-212. hamoud, r.m., irvani, m.r. and hackam, r., 1987, "coordinated static var compensators and power system stabilizers for damping power system oscillations," presented at power winter meeting, paper '87, vol. wm 045(8), pp. 1059-1067. hsu, y.y. and wu, c.j., 1988, "adaptive control of a synchronous machine using the auto-searching method", ieee trans. on power systems, vol. 3(4), pp. 1434-1439. ieee special stability controls working group 1994, "static var compensator models for power flow and dynamic performance simulation," ieee trans. on power systems, vol. 9(1), pp. 229-239. ieee task force on load representation for dynamic performance, 1995, "bibliography on load models for power flow and dynamic performance simulation," ieee trans. on power system, vol. 10(1), pp. 524-538. langevin, m. and auriol, p., 1986, "load response to voltage variations and dynamic stability," ieee trans. on power systems, vol. pwrs-1(4), pp. 112118. lee, s. and liu, c.c., 1994, "an adaptive output feedback static var controller for the damping of generator oscillations," electric power system research, vol. 29, pp. 9-16. mauricio, w. and semlyen, a., 1972, "effect of load characteristics on the dynamic stability of power systems," ieee trans. power app. and syte, vol. pas-91, pp. 2295-2304. milianovic, j.v. and hiskens, i.a., 1995, "effect of load dynamic on power system damping," ieee trans. on power systems, vol. 10(2), pp. 1022-1028. ohyama, t., et al. 1985, "voltage dependence of composite load in power systems," ieee trans. on power app. and syst., vol. pas-104(11), pp. 30643073. xu, w. and mansour, y., 1994, "voltage stability analysis using generic dynamic load models load," ieee trans. on power systems, vol. 9(1), pp. 479-493. xu, w., vaahedi, e. mansour, y. and tamby, j., 1997, "voltage stability load parameter determination from field tests on b.c. hydro's system," ieee trans. on power systems, vol. 12(3), pp. 1290-97. yu, x., khammash, m. and vittal, v., 2001, "robust design of damping controller for static compensator in power systems," ieee trans. on power systems, vol. 16(3), pp. 456-462. zein el-din, h.m. and price, w.w., 1988, "dynamic load modeling in large scale stability studies", ieee trans. on power systems, vol. 3(3), pp. 1039-1045. 78 the journal of engineering research vol. 5, no.1, (2008) 71-78 appendix system parameters and nominal operating values: synchronous generator xd =1.7 xq=1.64 xls=0.15 xkkd=1.605 xkkq=1.526 xffd =1.651 xad =1.55 xaq =1.49 ra=0.001096 rfd=0.000742 rkq=0.054 rkd=0.0131 h=2.37sec ieee type-1 excitation system ka=400 ta=0.05 sec kf = 1.0 tf =1.0 sec ke=-0.17 te= 0.95 sec transmission line x1= 0.15 x2=0.15 static var compensator (svc) kα =20 tα = 0.02 sec xs = 0.8 αo=1400 bl0=-0.164 tw = 0.1 sec vs limit = 0.1 nominal operating values pg =1.0 pfg = 0.85 pl0 =1.25 pfl = 0.5 vt = 0.976 v = 0.987 v∞ =1.0 δ0 = 43.146 0 mech080510b.qxd the journal of engineering research vol. 7, no. 1 (2010) 1-10 1. introduction wear may be defined as the progressive loss of a substance from the operating surface of a body occurring as a result of relative motion of the surface with respect to another body (stan grainger, 1994). the concept embraces metal to metal, metal to other solids and metal to fluid contact. this definition is clearly associated with the surfaces of materials. it must be recognized that wear and friction are not intrinsic properties of a material but are characteristic of the total engineering system and its operating environment any change in stress, temperature, speed can have a significant effect on the type of wear or the wear-rate of a component (askwith, 1980). aluminum-silicon (al-si) castings are considered the most commercially used (bolten, 1998 and mustafa 1995), because of their engineering significance accepted mechanical properties, good corrosion resistance, rela ___________________________________________ * corresponding author’s e-mail: mukeab2005@yahoo.com tively low thermal expansion coefficient and high fluidityduring melting and casting (bolten, 1998). therefore these alloys are often used in many tri-bological and bearing applications, pistons, cylinder heads in combustion chamber zones, camshaft bearings, spark plug bosses and bolt bosses (dana, 2001). there are three families of aluminum-based bearing alloys in common use. they are alpb, al-sn and al-si-cd. cadmium is used in bearing alloys, due to a low coefficient of friction and very good fatigue resistance. abbas and ibrahim (2003) studied the effect of cu and mg adding on wear behavior of (al-8%si) under dry sliding conditions and they concluded that the hardness and wear resistance increase because of formations of hard second-phase particles such as al2cu () and cumgal. israa (2005) studied the effect of lead addition on sliding wear resistance and friction coefficient of al-16%si alloy effect of cd on microstructure and dry sliding wear behavior of (al-12%si) alloy muna k. abbass department of production engineering and metallurgy, university of technology, baghdad, iraq received 10 may 2008; accepted 1 april 2009 abstract: the aim of the present research is to study the effect of cadmium addition on microstructure and wear behavior of the alloy (al-12%si) under dry sliding conditions. wear behavior was studied by using the pin-ondisc technique under different conditions at applied loads 5-20 n, at constant sliding speed and in constant time. the steel disc hardness was 35hrc. all alloys were prepared with different percentages of cadmium (1.0, 2.0, 3.0) wt%. also the base alloy was prepared by melting and pouring the molten metal in a metallic mold. it was found that the cadmium addition to al-si matrix decreases the wear rate and improves the wear properties for alloys containing -cd under loads above 10n. it was also found that the alloy al-12%si containing 3%cd is the best alloy in wear resistance and friction coefficient. this is due to presence of the cd-phase as cuboids or hard particles distributed in a eutectic matrix which reduces the friction coefficient at high loads (20n). keywords: wear resistance, al-si alloy, cadmium, friction coefficient ) áμ«ñ°ùd ±é÷g »b’õf’g ≈∏ñdg ∑ƒ∏°s h …ô¡ûg ö«cîdg ≈∏y ωƒ«eoéμdg òkcéj(al-12%si ¢séñy ò° n ≈æe áá°°uuóóÿÿgg) áμ«ñ°ùd ±é÷g »b’õf’g ≈∏ñdg ∑ƒ∏°sh …ô¡ûg ö«cîdg ≈∏y ωƒ«eoéμdg áaé°vcg òkcéj á°sgqo ” :al-12%si~æyh øjƒ«f 20-5 øe ò¨àe ∫éªmg òkéj â– ∂dp h ( q骰ùÿg ≈∏y ¢uô≤dg ´ƒf øe ≈∏ñdg ré¡l ωg~îà°sgh .úàhék ¥’õfg øerh áyô°spin-on-disc »g …p’ƒødg ¢uô≤dg ioó°u âféch35hrc∂féñ°ùdg ò° – ” ~bh áaé°vg ¿g ßmƒd .á«f~©e ödgƒb ‘ ö°üdg h ô¡°üdg á«∏ª©h ¿ƒμ«∏«°s %12ωƒ«æÿc’g ¢sé°s’g áμ«ñ°ùdg ¤g áaé°v’éh %3 ^ %2 ^ %1 ωƒ«eoéμdg øe áø∏àfl ö°ùf ≈∏y ájhé◊g ∂àμjƒñdg á«°vqg ¤g ωƒ«eoéμdgal-12%si∂dòc .øjƒ«f 10 øe ècg »àdg á£∏°ùÿg ∫éªm’g ~æy h çóãdg ∂féñ°ù∏d h ≈∏ñdg ¢ugƒn ø°ù– h ≈∏ñdg ∫~©e π«∏≤j ¤g …odƒj ∂àμjƒ«dg á«°vqg ‘ áyrƒàe i~∏°u ≥fébo hg ä骫°ùéc ωƒ«eoéμdg qƒw oƒlh ¤g …õ©j gògh .%3 ≈∏y ájhé◊g »g áμ«ñ°s π° ag ¿gh ωƒ«eoéμdg áaé°vg ™e ogoõj ≈∏ñdg áehé≤e ¿g ßmƒd .(øjƒ«f 20) á«dé©dg ∫éªm’g ~æy ∑éμàm’g πeé©e øe π∏≤j »àdgh áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .∑éμàm’g πeé©e ,º«eoéμdg ¿~©e ,áμ«ñ°s ,ωƒ«æÿ’g ,¿ƒμ«∏«°s ,πccéàdg áehé≤e : 2 the journal of engineering research vol. 7, no. 1 (2010) 1-10 under varying loads and sliding speed conditions. she concluded that the wear rate of an alloy increases with speed and applied load. it has been observed that the wear rate and friction coefficient decrease noticeably with increasing wt% of lead. the influence of calcium on the microstructure and properties of an al-7%si-0.3%mgxfe alloy has been studied by sreeja kumarl, et al. (2006), who concluded that the addition of calcium modifies the eutectic microstructure and also reduces the size of intermetallic fe-platelets, causing improved ductility and impact strength. basavakumar, et al. (2007) studied the microstructures and dry sliding wear behavior of al-7si and al-7si-2.5cu cast alloys after various melt treatments like grain refinement and modification. results indicate that the combined grain refined and modified al-7si-2.5cu cast alloys which have microstructures consisting of uniformly distributed -al grains, eutectic al-silicon and fine cual2 particles in the interdendritic region. these alloys exhibited better wear resistance in the cast condition compared with the same alloy subjected to only grain refinement or modification. the aim of present work was to study the effect of cadmium addition on the wear rate and friction coefficient of al-12%si alloy under dry sliding conditions. 2. experimental work 2.1 specimens preparation the al-12%si alloy (base alloy) was melted and different percentages of cadmium (1.0, 2.0, and 3.0) wt% were added to the melt of al-si separately. cadmium metal was enveloped with aluminum foil before being added to the melt. for each experiment about 300gm of prealloyed (al-12%si) in alumina crucible was placed in an electric resistance furnace at a temperature 700ºc which is below the boiling temperature of cadmium metal tp =767ºc. a small amount of flux (2% caf2) was added to the melt to remove the impurities (as slag) from the melt. the melt was poured into a metallic mold which was dried before the casting operation at 120ºc by using the heracus drier type. the castings were left to cool down in the air. the specimens of prepared alloys had 100mm length and 13mm diameter. 2.2 xrd measurement 2.3 wear specimens wear specimens were machined from an ingot and cut according to astm specification d2625-83 to 20mm length and 10mm diameter (astm, 1989). then one surface of each specimen was ground by using emery paper of sic in different grits (220,320,500 and 1000). the polishing was applied to the specimens by using diamond paste of size 1.0µm with a special polishing cloth and lubricant to obtain a clean and smooth surface. the initial surface roughness of the wear pins was ra = 0.20µm. 2.4 wear apparatus a pin-on-disc wear apparatus was used, which was designed according to astm specification f732-82 (astm, 1989) as shown in fig. 1. the pin (specimen) was fixed and the disc (carbon steel) was rotating at a speed of 510 rpm. the tangential force (friction force) was rotating between the pin and the disc through their interface sliding. the main features of the test specimens and loading system are described below. 2.5 test specimen and loading system the test specimen and loading system shown in fig. 1 consist of: figure 1. the pin-on-disc wear appratus the x-ray diffraction measurements have been carried out using shimadzu-labx x-ray diffraction unit model xrd-6000, kv = 40, cu k , and xrf (edxrf) type, twin-x, oxford co. england. these measurements include; base alloy and the al-12%si3%cd alloy to identify and estimate of the phases in an alloy. 3 the journal of engineering research vol. 7, no. 1 (2010) 1-10 1. a squirrel-cage type electrical motor (3-phase, and constant rotational speed of 940 rpm). 2. aluminum pulleys and a v-belt to transmit the speed. 3. a carbon steel rotating disc with a 24 cm diameter and a hardness of 35hrc and a surface finish of 0.8 µm. 4. a mechanism for loading the specimen which is drawn in detail in fig. 1. loading is applied by load weights representing normal force on pin and disc transmitted to the specimen by a lever system and holder. the fulcrum is 36cm and 7cm from disc center 3.7 m/sec relative speed of pin and disc from the point of loading or 34cm and 5cm from disc center 2.7 m/sec relative speed) as the load is applied directly to the holder. 5. four sensitive strain gages fixed on the inner and outer walls of a steel ring taken from tq calibration apparatus (attachment set) of s6072-e-31 techequipment strain meter shown in fig. 2. (the system is fixed on the holder and the force is transmitted directly from the lever to the ring by a 5mm diameter and of 3cm long pin. the whole system was leveled and placed on a heavy steel table to minimize vibration). 2.6 wear test the weight method was used to determine the wear rate of specimens. the specimens were weighted before and after the wear test with a sensitive balance type denver instrument max-210gm with an accuracy of 0.0001 gm. the weight loss w was divided by the sliding distance and the wear rate was obtained by using the following equation (unido, 1990). wear rate = w/sd (1) w = w2 w1 (2) sd = ss . t (3) wear rate (w.r) = w / d.n.t (4) where: w.r = wear rate (gm/ cm) sd = sliding distance (m) ss = linear sliding speed(m/sec.) d = sliding circle diameter (cm) ss = sliding distance (cm) t = sliding time (min) n = steel disc speed (rpm) hardness of = 35 hrc steel disc diameter of = 10 mm specimen length of = 20 mm specimen 2.7 measurement of friction coefficient friction coefficient (μ) is found by dividing the frictional force (f) by the applied load (n) according to the equation as follows: µ = f / n (5) frictional force is found by a specially designed steel ring taken from a calibration apparatus of a tq strain meter mounted on a u-angle block, fixed on the bench and aligned with the left-side of the lever arm. the force is transmitted from the left side of the fulcrum lever to the figure 2. four strain gages mounted on steel ring in addition to strain gage calibration system 4 the journal of engineering research vol. 7, no. 1 (2010) 1-10 ring by a 5mm diameter and 3cm long sliding rod. two half bridges were formed on the inside and outside ring surfaces by means of four strain gauges with a gage factor of 2.065. these strain gages were then connected to a tqs6072 digital bridge strain meter as shown in fig. 3. before running the experiment, the digital strain micrometer display cathode ray screen was calibrated to record the micro-strain reading before and after running. the reading from the strain meter was taken and the calibration curve of load (force) versus strain was plotted in fig. 4 at a constant sliding time (20 min) and sliding speed 2.7 m/sec and the steel disc hardness was 35hrc. 2.8 measurement of microhardness small specimens of base alloy al-12%si (alloy a) and of cd-containing alloys alloys b, c & d were prepared by turning processes in the dimensions of 12mm in length and 10mm in diameter. a wet grinding operation with water was done by using emery paper of sic in the different grits (220, 320, 500, and 1000). polishing was done to the specimens by using diamond paste of size (1µm) and special polishing cloth and lubricant. they were cleaned with water and alcohol and dried with hot air. etching process was done to the specimens by using an etching solution which was composed of (99% h2o+1%hf). figure 3. micro strain meter aparatus used in this study y = 33.473x 0 200 400 600 800 1000 1200 0 5 10 15 20 25 30 35 micro strain μ? figure 4. the calibration curve of load versus strain at a constant sliding time of 20 min, sliding speed of 2.7 m/sec. and steel disc hardness of 35 hrc micro strain  lo ad ( gm ) 5 the journal of engineering research vol. 7, no. 1 (2010) 1-10 then the specimens were washed with water and alcohol and dried. the vickers hardness test was carried out by using vickers hardness tester type einsingenbei u/m, model z323. a diamond indenter was forced into the surface of the specimen being tested under a static load of 300 mg for 10 to 15 sec. measurements of the indentation diagonal were made in 3 to 5 readings and the average hardness hv was found as shown in table 1. 3. results and discussion 3.1 microstructure results primary silicon has a cuboidal form which can be seen in the micrograph (fig. 5a) of an alloy al-12%si. the eutectic mixture, though, is non-lamellar in form and appears, in section, to consist of separate flakes of silicon (grey) and alpha al-phase (white) (basavakumar, et al. 2007). a coarse platelet of the crystals of the al-si eutectic phase was formed in the casting during solidification. these particles are brittle and tend to reduce the mechanical properties of the casting (lidman, 2005 ). during the solidification of the cast al-si alloys in a primary -phase start to separate from the liquid. after nucleation when the temperature decreases the primary phase grows as solid crystals have a dendrite shape. on the polished surface two phases can be distinguished and the secondary arms of the dendrites cut on the sample phase clearly appear zhang, et al. 2003). figure (5b-5d) shows the microstructures of alloys b, c and d containing cadmium 1%, 2% & 3% respectively. it is seen that the particles of cd are distributed as small particles in the matrix and al-si eutectic phase, because it is probable that there is only an extremely small solubility of al in solid cd as indicated in al-cd phase diagram (davis and associates, 1982), see fig. 6. modification and grain refinement in microstructure of al-si alloy have been achieved due to the presence of cadmium particles in the matrix of the eutectic phase. fig. 7 indicates x-ray diffraction analysis results of an alloy (al-12%si-3%cd) and the presence of si and al phases in the microstructure of the alloy. 3.2 wear results 3.2.1 effect of cd on wear rate the addition of cd to the base alloy (al-12 %si) leads to a decrease in wear rate as shown in fig. 8. the wear behavior of base alloy (a) is mild wear (oxidative wear) at low loads 5-10 n, and when the load increases the wear rate increases and transforms to metallic wear at high loads 10-20 n. these results are in agreement with those of other researchers (israa, 2005 and jawdat, 2002). in case of al-12%si containing 1%cd (alloy b) and 2%cd (alloy c) respectively, fig. 8, it can be seen that the wear rate reaches maximum value at load 10 n and then decreases as the applied load increases to 20 n. this is due to work-hardening of the matrix by plastic deformation which helped in reducing the extent of wear of the samples at high loads. during wear at high loads, the temperature increases appreciably, thus lowering the strength of the materials in contact and resulting in an increased contact area and coefficient of friction. the stronger grainrefined and modified alloys recorded the lowest wear rate. this is due to the presence of cd phase as hard particles (the hardness of pure cadmium is 203mpa) distributed in matrix of (al-si), which reduces the contact area between the steel disc and the specimen surface in addition to the role of silicon phase which increases wear resistance. in fig. 8 shows that the wear rate decreases to lower values as cd addition increases to 3%cd in an al-12%si alloy. these results are due to the increase in hardness of the alloy alloy d to 103 hv in comparison with the base alloy (alloy a) which is 70hv, see table 1 which gives the hardness results of the studied alloys (see table 1). 3.2.2 effect of cd on friction coefficient when two surfaces slide together, most of the work done against friction is turned into heat. the resulting rise in temperature may modify the mechanical and metallurgical properties of the sliding surfaces, and it may make them oxidize or even melt. all these things influence the wear rate and friction coefficient. in order to minimize friction forces we must use lubricants. these contaminate the surface preventing adhesive contact and obviously lowering the coefficient of friction (µ) (mikell groover, 1999). in this work, friction coefficient can be reduced, under dry sliding conditions by increasing the surface hardness of the alloy (a1-12%si) through the addition of cadmium to a matrix of eutectic phase (al-si). figure 9 shows the relationship between the coefficient of friction (µ) and sliding time at a constant load of 20n for different alloys (a, b, c & d). in the first sliding time the µ increases with time until it reaches a steady state and then continuing for longer periods (20 minutes or more). this is due to the presence of cuboids or hard cadmium particles distributed in the matrix which resists the large applied forces normal to the surface and hence separate the asperity tips very effectively, while the two surface layers can shear over each other quite easily. this can reduce the coefficient of friction (µ). it was found that as the cd% increases the grains becomes smaller and finer. the stronger grain-refined and modified alloys recorded the lowest coefficient of friction. these results are similar to those of references (basavakumar, et al. 2007 and asm materials, 2002). 3.2.3 worn surface results there are two main types of wear abrasive wear and adhesive wear. the present work studied the former one. specimen no. cd ( wt%) hv (kg/mm 2) a 0% (base alloy) 70 b 1.0 85 c 2.0 97 d 3.0 103 table 1. hardness results of the studied alloys 6 the journal of engineering research vol. 7, no. 1 (2010) 1-10 (al-si) eutectic cd particle 50 μm b 50 μm c cd particle (al-si) eutectic alpha phase a 50 μm 50 μm (al-si) eutectic cd particle d al-12% si alloy al-12% si alloy 1% cd al-12% si alloy 2% cd al-12% si alloy 3% cd a si phase figure 5. the microstructures of the studied alloys (as cast) 7 the journal of engineering research vol. 7, no. 1 (2010) 1-10 adhesive wear is often called frictional wear since it arises from the same basic mechanism of friction that comes from welding and subsequent shearing of minute surface asperities. when two nominally flat surfaces, no matter how smooth are brought into contact, they will only touch at relatively few isolated points. when sliding occurs these welds are sheared. when operating under heavy loads (or high speeds), the wear debris will appear as large mainly metallic particles (severe wear). on the other hand operation under light loads (or low speeds) will produce fine oxide wear debris (mild wear) (ashby and jones, 1980). figures 10a-10d show micrographs of the worn surfaces of specimens (a, b, c & d) at an applied load of 20n during wear test. figure 6. phase diagram of al-cd alloys (davis, j.r. and associates, 1982) 2 è a l s i s i s i a l fig.7 xrd results of an alloy ( al -12%si -3%cd figure 7. xrd results of an alloy (al-12%si 3%cd) in te ns it y (c .p .s ) 2 al-cd 8 the journal of engineering research vol. 7, no. 1 (2010) 1-10 the worn or damaged surface shows continuous grooves and cracking of a long wear track. in some places, some plastic deformation, together with presences of fine oxides debris particles are also observed in case of base alloy al-12%si. while it is seen as smooth and with glassy finish and faint wear lines in the direction of sliding on worn surfaces, this indicates that a mild abrasion wear mode is present in case of al-12%si alloys containing cadmium especially alloy d. 4. conclusions 1. modification and grain refinement in the microstructure of al-si alloy have been achieved by the presence of cadmium particles in the alloy matrix. 2. the wear behavior of base alloy a1-12%si changes from mild wear (oxidative wear) at low loads 5-10 n to metallic wear at high loads 10-20 n. 3. the cadmium added to alloy al-12%si changes the wear behavior at higher loads than 5-10 n. figure 8. the effect of applied load on wear rate of different alloys at a sliding speed of 2.7 m/sec, sliding time of 20 min and steel disc hardness of 35 hrc figure 9. the effect of cadmium adding on friction coefficient () of different alloys at a applied load of 20n and sliding speed of 2.7 m/sec sliding time, min f ri ct io n c oe ff ic ie nt applied load, n w ea r r at e, ( m g/ cm ) *1 08 9 the journal of engineering research vol. 7, no. 1 (2010) 1-10 4. the alloy a1-12 % si containing 3% cd shows the highest wear resistance in comparison with other alloys at a high load 20 n. 5. the cadmium added to alloy al-12 % si reduces the friction coefficient at high loads. 6. the addition of cadmium at different ratios leads to an increase in the hardness of the aluminum silicon based alloy. references abbas, m.k. and ibrahim, e.k., 2003, "effect of copper and magnesium adding on wear behavior of (al-si) alloy," engineering and technology journal, university of technology, baghdad iraq, vol. 22(1), pp. 37-44. ashby, m.f. and jones, d.r.h., 1980, "engineering materials, an introduction to their properties and application," pergamon press ltd, headington hillhall, oxford, england. askwith t.c., 1980, "the basic mechanisms of wear," surfacing journal, vol. 11(4), pp. 2-5. asm materials, 2002, "bismuth added to al-si alloy enables low friction," advanced materials progress. astm., 1989, "metals test method and analytical procedure," vol. 05.02. basavakumar, k.g., mukunda, p.g. and chakraborty, m., 2007, "influence of melt treatments on sliding wear behavior of al-7si and al-7si-2.5cu cast alloys," journal of material science, vol. 42, pp. 7882-7893. bolten, w., 1998, "engineering materials technology," butterworth's, 3rd ed. (a) al-12%si (base alloy ) (b) al-12%si-1%cd ( c) al-12%si-2%cd (d) al-12%si3%cd fig.10 micrographs of the worn surfaces of different alloys at a applied load of 20n , sliding speed of 2.7m/sec and steel disc har dness of 35hrc sliding direction 100ì m 100ì m 100ì m 100ì m figure 10. micrographs of the worn surfaces of different alloys at a applied load of 20 n, sliding speed of 2.7m/sec and steel disc har dness of 35hrc sliding direction (c) sl-12% si-2% cd (d) al-12% si-3% cd (a) al-12% si (base alloy) (b) al-12% si-1% cd 10 the journal of engineering research vol. 7, no. 1 (2010) 1-10 dana, e.c.h., 2001, "information from clevite engine parts division, "engine bearing fundamentals, part 6 bimetals," pp. 1-2. davis, j.r. and associates, 1982, "asm international handbook committee," by a.j. mcalister, p. 545. http://books.geogle.com israa, a.k., 2005, "the effect of lead addition on the wear resistance of al-16%si alloy under dry sliding," journal of engineering and technology university of technology, baghdad, vol.24(1), pp. 56-64. jawdat, a.g., 2002, "effect of some additions on wear resistance and mechanical properties of (al-8% si)," m.sc. thesis, university of technology, baghdad , iraq. lidman, w.g., "aluminum master alloys and additives, the magic ingredients," the aluminum association, association's , website at www.aluminum.org. mikell, p. groover, 1999, "fundamentals of modern manufacturing materials processes and systems," john, wiley & sons, inc. mustafa, s.f., 1995, "wear and wear mechanisms of al22%si -203 composite," wear, vol. 185, p. 189. sreeja kumar, s.s.., pillai, r.m., nogita, k., dahle, a.k. and pai, b.c., 2006, "a influence of calcium on the microstructure and properties of an al-7%si0.3%mgxfe alloy," metallurgical and materials transaction a, vol.37a, p. 2581. stan grainger., 1994, "engineering coatings design and application," woodhead publishing ltd. unido., 1990, "advances in material tech," monitor vienna international center, austria, pp. 9-11. zhang, z., garro, m. and tagliano, c., 2003, "dendrite arm spacing in aluminum alloy cylinder heads produced by gravity semipermanent mold," metallurgical science and technology, a journal published by teksid aluminum twice a year, vol.21(1), pp. 1-9. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 137-145 *corresponding author’s e-mail: madhavi@qu.edu.qa doi:10.53540/tjer.vol18iss2pp137-145 occupant perception of thermal comfort in sleep environments in qatar madhavi indraganti *, 1, farsana kutty 1, reem ali1, lulwa al noaimi1, saaeda al-bader1 and maryam adel al mulla1 department of architecture and urban planning, college of engineering, qatar university, doha, qatar abstract: a thermal comfort field survey in sleep environments was conducted in qatar during winter and spring seasons. a total of 833 sets of objective and subjective thermal and sleep quality responses before going to bed and after getting up were collected. the subjects felt cooler sensations most of the time, preferred warmer sensations and the sleep environments are considered comfortable overall. griffiths comfort temperature (tc) was 24.3 ˚c and 20.2 ˚c in free-running (fr) and air-conditioned (ac) modes respectively. subjects used airconditioner (ac)s adaptively in heating mode in winter. in 82.7% of cases in (ac) mode, the comfort temperature was below the lower limit of the international standard. subjects enjoyed quality sleep in qatar. overall, selfdeclared sleep quality improved with thermal acceptability. depth of sleep was higher in ac mode (mean = 3.86). on the other hand, the mean global pittsburgh sleep quality index (psqi) score was high in general (mean = 10.7), indicating good quality sleep, and significantly so in free-running mode (mean = 11) than in ac mode (mean = 10.4). it increased, as subjects liked their ac systems. this study suggests that overcooling in spring can be avoided by increasing air movement without compromising sleep quality. keywords: bedroom; comfort temperature; occupant behavior; quality of sleep; thermal comfort تصور السكان للراحة الحراریة في بیئات النوم في قطر المالل دة البدر ومریم عاي وساعدمدافي اندراجانتي* و فرسانا كوتي و ریم علي ولولوة النعیم من 833جمع مسح میداني للراحة الحراریة في بیئات النوم في فصلي الشتاء والربیع في قطر ما مجموعھ :الملخص االستجابات الموضوعیة والذاتیة لإلحساس الحراري وجودة النوم قبل النوم وبعد االستیقاظ. على الرغم من شعور األشخاص یئة أكثر دفئا؛ إال أنھم اعتبروا أن بیئات نومھم مریحة بشكل عام. كانت درجة حرارة بالبرودة في أغلب األوقات، وتفضیلھم لب درجة مئویة في وضع تشغیل التكییف. استخدم األشخاص 20,2درجة مئویة في وضع التشغیل الحر و 24,3جریفیث المریحة یف كانت درجة الحرارة المریحة أقل من % من حاالت تشغیل التكی82,7مكیفات الھواء في وضع التدفئة خالل الشتاء، وفي الحد األدنى للمعیار الدولي، وبصفة عامة ارتفعت جودة النوم كما عبر عنھا المشاركون مع ارتفاع تقبلھم لألحوال الحراریة. ل عام في المتوسط، بینما كان مؤشر بیتسبرغ لجودة النوم العالمي مرتفعا بشك 3.86بلغ عمق النوم في وضع تشغیل التكییف ، مما یدل على جودة نوم عالیة بصفة عامة، أما في وضع التشغیل الحر، فقد كانت الجودة مرتفعة بشكل ھام، 10.7بمتوسط في وضع تشغیل المكیف، وتزداد الجودة مع ازدیاد الرضا 10.4، مقارنة بجودة بلغت 11حیث بلغت على مؤشر بیتسبرغ اسة أنھ یمكن تجنب البرودة الزائدة في فصل الربیع بشكل ال یؤثر على جودة النوم عن أجھزة تكییف الھواء. تقترح ھذه الدر عن طریق زیادة حركة الھواء. الراحة الحراریة؛ التھویة؛ جودة النوم؛ التكییف. الكلمات المفتاحیة: occupant perception of thermal comfort in sleep environments in qatar 138 1. introduction sleep is a basic human need and quality sleep is vital for human health and wellbeing. it is known for a long that sleep quality and bedroom temperature are negatively correlated (humphreys, 1970). recent studies have found that room freshness (lower co2 levels) significantly enhanced the occupant’s sleep quality, performance in a test of logical thinking, and reduced the symptoms of the next day reported sleepiness (strøm-tejsen, et al., 2016). that elevated air movement enhances sleep quality and fans could be used to great advantage in heatwave situations are now well established through biophysical modelling and field experiments (morris, et al., 2021; lan, et al., 2019). literature also pointed to subjects sleeping well in japanese homes with the help of adaptive measures such as using cooling devices and fans (imagawa & rijal, 2015). in this context, it is important to note that buildings in qatar are maintained at near-zero indoor airspeeds (indraganti & boussaa, 2018). cheap energy prices coupled with long hot and humid seasons increase the air-conditioner usage and overcooling in buildings is rampant in the gulf cooperation countries (gcc) (elnaklah, et al., 2021; indraganti & boussaa, 2018). further, in qatar, the final electricity consumption in the residential sector is very high (45%) and has increased by 33% between 2015-18. while the country ranked highest in per capita co2 emissions in the world (iea, 2018). moreover, building energy for space cooling and heating contributes substantially to greenhouse gas emissions. gulf nations like qatar have already witnessed average (outdoor) temperature increases of over 2˚c (al-saffar & beeuren, 2020). as there is no custom-developed adaptive thermal comfort standard for the gcc, engineers usually use the predicted mean vote (pmv) approach for indoor environmental design, following international standards (global sustainability assessment system (gsas, 2015; ashrae, 2020; iso, 2005). researchers claimed that this approach was inaccurate in predicting actual thermal comfort and suggested a transition in the air-conditioning practice towards the adaptive framework (humphreys & nicol, 2002; cheung, et al., 2019; parkinson, et al., 2020). thermal comfort in residential spaces and especially sleep environments are under-reported in the middle east. therefore, a field study was conducted to investigate the thermal comfort, quality of sleep and occupant adaptive behaviour in sleep environments of six residences in qatar in the winter and spring seasons. 2. methodology 2.1. survey location and buildings surveyed qatar is a small nation in the arabian peninsula with a hot desert climate (bwh in köppen climate classification). its capital is doha (n25° 17’, e51° 32’, 10 m above the mean sea level). it has five months of harsh summer (may – september) and mild winter (december – february) with very low precipitation. autumn (october, november) and spring (march, april) are short mild seasons of two months each. this is an ongoing longitudinal survey in six randomly selected residences (b1-b6) involving nine healthy participants in doha metropolitan area. this paper includes the data from december 28, 2020 – march 29, 2021, consisting of 833 sets of subjective thermal responses and simultaneous objective environmental measurements. the sars-cov-19 pandemic after march 2020 affected the data collection. table 1 shows the details of the buildings surveyed (figure 1). the buildings are multi-storied low-rise apartment blocks and independent villas. they have a concrete structural frame and concrete block masonry infill walls with sliding and sash windows, a typical style used locally. only b1 had portable fans. buildings b1-b5 have cloth curtains while b6 had roller blinds. no data was collected about the heating set points in the survey. 2.2. data collection set to a 10-minute interval, calibrated sensors of the data loggers were fixed to a hanger or a board at 1.1 m height from the floor in all the bedrooms. they recorded indoor air temperature (ta), globe temperature (tg), relative humidity (rh) and air velocity (va) in all bedrooms. excepting in b2, b4, co2 concentration was also measured in all the other houses. the outdoor daily mean temperature (to) was obtained from the meteorological website (anon., 2021). the mean radiant temperature (tmrt) is calculated following the procedure in the american society of heating, refrigerating and airconditioning engineers (ashrae) handbook (ashrae, 2005) and operative temperature (top) was taken as the average value of tmrt and ta (ashrae, 2020). the instrument details and the set-up used in the surveys are shown in figure 2 and table 2. 2.3. thermal questionnaire and the respondents table 3 lists all the scales used in this survey. nine respondents participated in the survey of which one was a male participant in b4 (age: 60 years, sample size (n): 141). it was learned that men in arab families usually spend winters in tent structures “kashta” (كشتھ) in the desert away from homes, and, therefore, our survey recruited fewer men. consequently, we collected 83.1 % data from females (mean age: 30.2 (standard deviation (sd) (11.1) years). the bedrooms in b4 and b6 had two subjects each, while the rest had one each. male and female samples are treated independently to avoid sampling bias. aged 21-60 years (mean: 35.3 ±15.1 years), the respondents are all healthy residents of qatar living in the surveyed areas for longer than one year and are assumed to be acclimatized to the local climate. no physiological data of the subjects were collected in this study. madhavi indraganti, farsana kutty, reem ali, lulwa al noaimi, saeeda al-bader and maryam adel al-mulla 139 table 1. details of the buildings surveyed. house (sample size) respondents location typology floor surveyed room area (m2) ceiling height (m) ac system thermostat window glazing window area (m2) b1 (110) 1 najma single apartment 2 14.5 3.2 window wallmounted sliding single 1.98 b2 (69) 1 madinat khalifa south single apartment 4 17.4 2.8 split hand-held sliding single 1.44 b3 (79) 1 al-mamoura independent villa 1 20 3 split hand-held sliding single 1.5 b4 (283) 3 binmahmoud south apartment in compound 4 21 3.5 ducted central wallmounted sash double 3.75 b5 (87) 1 al-nasr independent villa 1 20 3 split hand-held sliding double 1.8 b6 (205) 2 al-tarfa apartment in compound 1 80 4.5 ducted central wallmounted sliding double 3 table 2. details of the dataloggers and sensors used in the surveys. description/ manufacturer trade name parameter measured range accuracy b2, b4 thermo-hygro datalogger with probe thermometer with black painted table tennis ball and air velocity sensor; onset, hobo, usa u12-013 (temp/rh/2 ext) air temperature 0 to 50 °c ±0.35 °c (0° to 50°c) humidity 5 to 90% rh ±2.5% (10 to 90%) tmc1-hd probe with globe temperature 0 to 50 °c ±0.35 °c (0° to 50°c) accusense f900o-10-0 hot wire anemometer air velocity 0.15 5 m/s ± 0.05 m/s (0 to 5 m/s) b1, b3, b5, b6 onset hobo bluetooth wireless co2, temp & humidity data logger mx 1102a air temperature 0° to 50°c (32° to 122°f) ±0.21°c from 0° to 50°c humidity 1% to 70% rh ±2% from 20% to 80% co2 concentration 0 to 5,000 ppm ±50 ppm ±5% of reading at 25°c ux120-006m analogue data logger with external temperature probe sensor with sibata 75mm cu globe and degreec air velocity sensor. t-dci-f300-1a3 hot wire anemometer air velocity 0.15 5 m/s ± 0.05 m/s (0 to 5 m/s) tmc1-hd probe with 080340-275 cu globe globe temperature 0 to 50 °c ±0.35 °c (0° to 50°c) figure 1. survey locations and the typical buildings surveyed. figure 2. instrument set-up and the survey environments (a: mx1102a; b: ux120-006m; c: hot-wire anemometer; d: cu globe with thermal probe). occupant perception of thermal comfort in sleep environments in qatar 140 table 3. scales used in the survey. scale value thermal sensation (tsv) thermal preference (tp) overall comfort (oc) depth of sleep sensitivity to warmth (sw) overall sleep quality preference on cooling/ heating thermal acceptability (ta) 6 very comfortable very deep 5 comfortable deep 4 slightly comfortable slightly deep heat and cold 3 hot slightly uncomfortable slightly light cold very good 2 warm much cooler uncomfortable light neither heat nor cold good 1 slightly warm a bit cooler very uncomfortable very light heat bad like unacceptable 0 neutral no change very bad neither like nor dislike acceptable -1 slightly cool a bit warmer dislike -2 cool much warmer -3 cold table 4. descriptive statistics (mean and standard deviation) of outdoor and indoor environmental variables. mode (n) fr (433) ac (400) all (833) outdoor daily mean temperature (°c ) 20.9 (2.4) 21 (2.3) 20.9 (2.3) outdoor relative humidity (% ) 59.7 (12.1) 60.4 (11.6) 60.1 (11.8) operative temperature (°c ) 23.9 (1.5) 19.8 (2.2) 21.9 (2.8) air temperature (°c) 24.1 (1.5) 20.1 (2.2) 22.2 (2.7) relative humidity (%) 61.4 (5.8) 61.3 (6.1) 61.4 (5.9) air velocity (m/s) 0.02 (0.04) 0.04 (0.07) 0.03 (0.06) the respondents were requested to respond to the online google survey twice a day: (1) before going to bed after spending 15 minutes in the bedroom and (2) soon after waking up reflecting on the thermal comfort during sleep. the questionnaire was modelled after imagawa & rijal (2015) in english and translated to arabic. following tsang, et al. (2021), the detailed sleep quality questions were prepared to estimate the pittsburgh sleep quality index (psqi) (buysse, et al., 1989). psqi is a self-scoring subjective measure of sleep quality in all seven dimensions of sleep such as quality, latency, efficiency, disturbances, medications, and daytime disfunction. a higher psqi index refers to better sleep quality. to the psqi, we added the overall sleep quality score and obtained global psqi. the internal review board (irb) at qu approved the questionnaire for ethical compliance. the ashrae 7-point scale was used for evaluating thermal sensation (tsv). the questionnaire had three segments: time and personal identifier/variables followed by branching questions of thermal comfort and the use of controls. a separate one-time background survey collected the building and personal details such as age. the thermal questionnaire included elaborate pictorial checklists for various pieces of garments, three types of bedding, and a list of activities to choose from. it also included a list of environmental controls (ac/fan/heater/ window/curtain/lights) which are coded as 0: not in use, 1: in use and 2: unavailable. the occupants gave 458 (male: 84, female: 374) and 375 (male: 57, female: 318) datasets before going to bed and after getting up respectively. 3. results and discussion 3.1. outdoor and indoor conditions the following sections present the results of the data considered in two groups: fr (free-running – no cooling or heating) and ac (cooling/heating by airconditioning). we used spss ver.27. in fr mode, the mean outdoor temperature was 20.8 ˚c and 20.9 ˚c, before going to bed and after getting up respectively, and similarly, in ac mode it was 21.1 ˚c and 20.9 ˚c respectively. likewise, in fr mode, the mean outdoor relative humidity was 60.1% and 59.4 % before going to bed and after getting up respectively, and similarly, in ac mode it was 59.4% and 61.8 %, respectively. as indoor operative temperature (top) correlated very strongly (r = 0.999-0.996, p < 0.001) with other indoor thermal indices tg, tmrt and ta. the top was selected for madhavi indraganti, farsana kutty, reem ali, lulwa al noaimi, saeeda al-bader and maryam adel al-mulla 141 the analysis to enable easy comparison with literature and standards. the indoor thermal conditions in fr mode were significantly warmer than ac mode (mean difference = ~ 4.6 k, p < 0.05), while the difference in top “before going to bed” and “after getting up” was very small. comparatively, rh was less variable among the four conditions examined (figure 3). this result suggests that people in doha sleep in humidcooler conditions in ac mode and humid-warmer conditions in fr mode. the descriptive statistics of outdoor and indoor variables are presented in table 4. being a winter and spring season study, very low air movement was noted indoors (mean airspeed = 0.03 m/s). the clothing insulation (icl) for all subjects before and after the sleep was estimated. the icl values of western and non-western ensembles were estimated using the standard checklists and reports (ashrae, 2020; iso, 2005; indraganti, et al., 2015; havenith, et al., 2015; mitsuzawa & tanabe, 2001; al-ajmi, et al., 2008). the bedding and mattress insulations were estimated following lin & deng (2008), assuming the subjects to be using similar mattresses and heavy/light duvets and blankets. the metabolic rates were estimated as per (iso, 2005). 3.2. subjective thermal responses overall, the mean thermal sensation was found to be 0.19 (standard deviation: 1.1), indicating that the subjects voted on the cooler side of the 7-point sensation scale. in fr and ac modes it was -0.20 and -0.17 respectively, implying that the subjects were often feeling cooler sensations. in both fr and ac modes, subjects were feeling significantly warmer sensations after getting up (by about ½-sensation vote) (p < 0.05). interestingly, the literature indicates that warmer thermal sensations (tsv > 0.8) are counterproductive to sleep quality improvement (song, et al., 2020). much as expected, the preference (tp) was for warmer environments with the mean tp vote being -0.14 overall (figure 4). in fr mode, tp was significantly different before going to bed and after getting up (p < 0.05), but not in ac mode. the preference for warmer environments was significantly higher in fr mode. this is in part due to the general warmer environment preferences of people noted in the winter season in fr mode, which is not so prominent in ac mode. while the overall comfort (oc) of subjects was the highest in ac mode after getting up, no significant differences before sleep and after sleep were noted in the oc vote. the results also indicated that subjects were comfortable with their sleep environments most of the time, with the proportion of subjects voting on the comfortable side of the oc scale being 63.2% 83.3% and the proportion comfortable (-1 ≤ tsv ≤ 1) also being high (79% 93%) in all the four cases considered as above (figure 4). this result is comparable to the 80% satisfaction rate suggested in ashrae std-55 (ashrae, 2020). 3.3. estimation of comfort temperature comfort temperature refers to the temperature needed indoors for people to feel comfortable. linear regression is one of the methods used to estimate it. in this study, we regressed top with tsv and obtained the following relationship for fr mode: tsv = 0.167 top – 4.185 (r = 0.253, p < 0.001, (1) where n = 433, standard error (se) = 0.031) the equation in ac mode was not significant at 95% confidence interval (ci) and the slopes for “before going to bed” and “after getting up” cases were found to be homogenous at 95% ci. this relationship returned a comfort neutral temperature of 25.1 ̊ c in fr mode and a comfort band (-1 ≤ tsv ≤ 1) of 19.1 ˚c 31.5 ˚c. the mean top when subjects voted neutral in fr mode was 24 ˚c (sd: 1.2), which was slightly lower than the regression natural temperature. in ac mode, it was 20.2 ˚c (sd: 2.2). often, it is noted in field experiments that both subjective sensation and or the thermal stimuli do not fluctuate much, and, consequently, the regression equation may not be significant, as in the case of ac mode in this study (humphreys, et al., 2016, p. 287). therefore, the literature points to the use of the griffiths method to evaluate the comfort temperature (griffiths, 1990). figure 3. distribution of indoor operative temperature and relative humidity in fr and ac modes before going to bed and after getting up as noted in the surveys. occupant perception of thermal comfort in sleep environments in qatar 142 figure 4. distribution of thermal sensation, thermal preference and overall comfort votes in fr and ac modes. further, in field experiments, most subjects express near-neutral sensations even when the temperature varies in fr mode and vice versa in ac mode. this results in a low coefficient of determination between the temperature and thermal sensation. literature also points to errors in neutral temperature attributed to the scattering of the thermal sensation and the lack thereof. therefore, griffiths method is suggested as a more suitable method for small batches of data. it is estimated using the relationship: tc = top + (0 – tsv)/g (2) where tc is the comfort temperature and g is the griffiths coefficient (griffiths, 1990). griffiths method was used to estimate the comfort temperature (tc) testing initially with three different griffiths coefficient (α)s: 0.25, 0.33 and 0.50. significantly, 0.5 as α resulted in the most reliable prediction of tc (with lowest sd value). therefore, 0.5 was used for α as humphreys et al. (2016) (p. 251) also suggested. the following mean comfort temperatures and standard deviations were obtained as shown in the parenthesis: in fr mode 24.7 ˚c (sd = 2.3) and 23.8 ˚c (sd = 1.9) before going to bed and after getting up respectively; and in ac mode 20.8 ˚c (sd = 3) and 19.3 ˚c (sd = 3.2) before going to bed and after getting up respectively. these four values are statistically significantly different from one another (p < 0.05), clearly indicating that the comfort temperature is lower after getting up in both modes. this result is comparable to hong kong ac bedrooms in high-rise residences, where lin and deng (2006) found the preferred temperature to be below 24 ˚c. 3.4. comparison with international standards to compare, the collected comfort data were superimposed on the comité européen de normalisation (cen) and the chartered institution of building services engineers (cibse) guide standards as shown in figure 5 a, b. (european committee for standardization, 2019; cibse, 2015). observable, in fr mode, most of the data (60%) was lying within the ± 2 k comfort band of the adaptive relationship, while in ac mode only 18% was within the respective comfort band. importantly, 34.9 % and 81% of data fell below the lower limit of the cen/cibse guide standards in fr and ac modes respectively. in ac mode, 82.7% (n = 277), and 72.3% (n = 47) cases recorded tc below the cibse guide value in these two seasons (figure 5b). it could mean that subjects were achieving lower winter temperatures by using less heating in winter in ac mode. having habituated to lower temperatures, the subjects continued maintaining a lower temperature in the spring season also as is evident in figure 5b. observable, comfort temperatures were significantly lower in ac mode (p < 0.001) with large effect size (η2 > 0.14, p < 0.05) (figure 5c). a review by lan et al. (2017) suggested that maintaining lower comfort temperature in winter does not compromise sleep quality, much similar to the result obtained in this study. 3.5. evidence of adaptation qatar has salubrious weather in the winter and spring seasons and as a result, subjects used air-conditioning systems adaptively in residences (mean usage = 48%, n = 833). the proportion using acs varied significantly with the indoor operative temperature. as this was majorly a winter study, the acs were used in heating mode and their usage plummeted as the temperature increased as shown in figure 5d. to estimate the probability of ac use at various indoor temperatures, logistic regression of proportion of ac use (pac) with indoor operative temperature was conducted and obtained the following relationship: logit (pac) = -1.248 top + 27.488 (3) where n = 833, r2 = 0.574; p <0.001, and top is the indoor operative temperature (˚c), n is the sample size, p is the level of significance of the regression coefficient and r2 is the cox and snell r2. it can be noted from this relationship that ac usage was 80% at 20.9 ˚c. imagawa and rijal (2015) noted a similar adaptation in cooling device use in japanese bedrooms. 3.6. sleep quality figure 6 (a) shows that a significantly higher proportion of people liked the ac systems when about 60% of the ac systems were in use. this indicated higher proclivity for using acs at bedtime. imagawa and rijal (2015) found slightly lesser (55%) ac usage when subjects liked the cooling systems. results also indicated that subjects liking the ac systems had madhavi indraganti, farsana kutty, reem ali, lulwa al noaimi, saeeda al-bader and maryam adel al-mulla 143 significantly higher global psqi scores. importantly, subjects in fr mode had significantly higher global psqi scores than in ac mode, as shown in figure 6b at all levels of preference for acs. this figure also shows that residents in qatar were generally ranked high in psqi scores, indicating good sleep. we noticed subjects who accepted the thermal environments had significantly higher self-reported overall sleep quality in both modes as shown in figure 6c. tsang, et al. (2021) reported a similar association between overall sleep quality neutral tsv. our results also indicated that subjects in ac mode had deeper sleep than in fr mode (figure 6d). on the other hand, subjects in fr mode had significantly higher comfort temperature than ac mode. however, their comfort temperature at various levels of sensitivity to warmth (sw) did not vary significantly in both modes. it means that the data was not fully explaining the variation in comfort temperature in both modes. on the other hand, results also showed that people sensitive to cold had significantly higher comfort temperature (p < 0.05) than those expressing other sensitivities to warmth as shown in figure 6e. a similar association was noted between top and sensitivity to warmth, indicating that other dimensions of the thermal environment such as air movement could have influenced sw. (a) (b) (c) (d) figure 5. comfort data of this research with markers shown for winter and spring seasons for “before going to bed” and “after getting up” cases super-imposed over the european standard for (a) fr and (b) ac modes; (c) significant differences in mean comfort temperature in fr and ac modes in “before going to bed” and “after getting up” cases (p < 0.05); (d) probability of ac use varying with indoor operative temperature, with markers shown for actual proportion or ac use in 1k bins. bins with fewer than ten samples are omitted. (a) (b) (c) (d) (e) figure 6. association with the three preferences for heating/cooling and (a) mean proportion using acs, (b) mean global pittsburgh sleep quality index score; (c) overall sleep quality varying with thermal acceptability; (d) depth of sleep varying with mode; (f) comfort temperature varying with various levels of sensitivity to warmth. (p < 0.05); (n = 372). occupant perception of thermal comfort in sleep environments in qatar 144 4. conclusion a thermal comfort field survey was done in sleep environments in homes in qatar during the winter and spring seasons of 2020-21. a total of 833 datasets was collected. these contained the thermal comfort data relating to before going to bed and after getting up conditions in six bedrooms. the conclusions are as follows: 1. the subjects expressed cooler sensations and preferred warmer indoors in both fr and ac modes. 2. people are generally comfortable with their sleep environments and the griffiths comfort temperature (tc) was 24.3 ̊ c and 20.2 ̊ c in fr and ac modes respectively. in both modes, tc before going to bed was slightly higher (p < 0.05). 3. subjects used acs adaptively in heating mode in winter and tolerated low bedroom temperatures, and as a result, in ac mode 82.7% of data fell below the lower limit in the cibse guide. 4. the occupants enjoyed quality sleep in qatar. overall, self-declared sleep quality significantly improved with thermal acceptability. 5. the depth of sleep was higher in ac mode (mean = 3.86). on the other hand, the mean global pittsburgh sleep quality index score was high in general, indicating good quality sleep, and significantly so in fr mode (mean = 11) compared to the ac mode (mean = 10.4). it increased, as subjects liked their ac systems. this study suggests that overcooling in spring can be avoided by increasing air movement without compromising sleep quality. conflict of interest the authors declare no conflict of interest. funding qatar national research fund through the undergraduate research experience program (grant numbers: urep26-033-2-010, urep26-034-2-011) funded this research. some instruments used in the survey were funded by the japan society for promotion of science through a post-doctoral fellowship (jsps/ff1/246/p 11372). acknowledgment we thank all the survey respondents and both funding agencies. the views expressed in this paper are sole of the authors. references al-ajmi, f. f., loveday, d. l., bedwell, k. h. & havenith, g., 2008. thermal insulation and clothing area factors of typical arabian gulf clothing ensembles for males and females: measurements using thermal manikins. applied ergonomics, p. 39 (2008) 407–414. al-saffar, a. & beeuren, m. v., 2020. the case for energy transitions in major oiland gas-producing countries. [online] available at: https://www.iea.org/commentaries/ the-case-for-energy-transitions-in-major-oil-and-gasproducing-countries anon., 2021. weather underground. [online] available at: https://www.wunderground.com/ weather/qa/doha [accessed 23 6 2021]. ashrae, 2005. ashrae handbook: fundamentals, si, atlanta, ga, usa: american society of heating, refrigerating and air-conditioning engineers, inc. ashrae, 2020. ansi/ ashrae standard 55-2020 thermal environmental conditions for human occupancy, atlanta: ashrae. buysse, d. et al., 1989. the pittsburgh sleep quality index: a new instrument for psychiatric practice and research. j. psychiatr. res., 28(2), p. 193–213. cheung, t. et al., 2019. analysis of the accuracy on pmv – ppd model using the ashrae global thermal comfort database ii. building and environment, vol.153, pp. 205-217. cibse, 2015. cibse guide a: environmental design, london: the chartered institution of building services engineers. elnaklah, r. et al., 2021. thermal comfort standards in the middle east: current and future challenges. may, p. 107899. european committee for standardization, 2019. en 16798-1:2019 energy performance of buildings ventilation for buildings part 1: indoor environmental input parameters for design and assessment of energy performance of buildings addressing indoor air quality, thermal environment, lighting and acous, brussels: european committee for standardization. global sustainability assessment system (gsas), 2015. gsas building typologies: design guidelines 2015 v2.1, doha: gulf organization for research and development. griffiths, i. d., 1990. thermal comfort in buildings with passive solar features: field studies, s.l.: s.n. havenith, g. et al., 2015. a database of static clothing thermal insulation and vapor permeability values of non-western ensembles for use in ashrae standard 55, iso 7730, and iso 9920. ashrae transactions, 121(1), pp. 197-215. humphreys, m. a., 1970. a simple theoretical derivation of thermal comfort conditions. the journal of the institute of heating and ventilating madhavi indraganti, farsana kutty, reem ali, lulwa al noaimi, saeeda al-bader and maryam adel al-mulla 145 engineers, 38(8), p. 95–98. humphreys, m. a. & nicol, j. f., 2002. the validity of iso-pmv for predicting comfort votes in everyday thermal environments. energy and buildings, vol.34, pp. 667-684. humphreys, m., nicol, f. & roaf, s., 2016. adaptive thermal comfort, foundations and analysis. oxon: routledge. iea, 2018. key energy statistics 2018. [online] available at: https://www.iea.org/countries/qatar [accessed 2 07 2021]. imagawa, h. & rijal, h. b., 2015. field survey of the thermal comfort, quality of sleep and typical occupant behaviour in the bedrooms of japanese houses during the hot and humid season. architectural science review, 58(11), pp. 11-23. indraganti, m. & boussaa, d., 2018. an adaptive relationship of thermal comfort for the gulf cooperation council (gcc) countries: the case of offices in qatar. energy and buildings, vol.159 (2018) , p. 201–212. indraganti, m., lee, j. & zhang, h. a. e. a., 2015. thermal adaptation and insulation opportunities provided by different drapes of indian saris. architectural science review, 58(1), p. 87–92. iso, 2005. iso 7730:2005(e) ergonomics of the thermal environment — analytical determination and interpretation of thermal comfort using calculation of the pmv and ppd indices and local thermal comfort criteria, switzerland: iso. lan, l., tsuzuki, k., liu, y. & lian, z., 2017. thermal environment and sleep quality: a review. energy and buildings, vol.149, pp. 101-113. lan, l., xia, l., tang, j. & wang, z., 2019. elevated airflow can maintain sleep quality and thermal comfort of the elderly in a hot environment. indoor air, vol.29, p. 1040–1049. lin, z. & deng, s., 2006. a questionnaire survey on sleeping thermal environment and bedroom air conditioning in high-rise residences in hong kong. energy and buildings, vol.38, pp. 1302-1307. lin, z. & deng, s., 2008. a study on the thermal comfort in sleeping environments in the subtropics—measuring the total insulation values for the bedding systems commonly used in the subtropics. building and environment, vol.43, p. 905–916. mitsuzawa, s. & tanabe, s.-i., 2001. effect of air movement on thermal comfort under hot and humid conditions while wearing traditional clothing. london, uk, proceedings of moving thermal comfort standards into the 21st century, p. 491500, windsor conference. morris, n. b. et al., 2021. electric fan use for cooling during hot weather: a biophysical modelling study. lancet planet health, vol.5, p. e368–77. parkinson, t., de dear, r. & brager, g., 2020. nudging the adaptive thermal comfort model. energy and buildings, vol.206, p. 109559. song, c., zhao, t., song, z. & liu, y., 2020. effects of phased sleeping thermal environment regulation on human thermal comfort and sleep quality. building and environment, vol.181, p. 107108. strøm-tejsen, p., zukowska, d., wargocki, p. & wyon, d. p., 2016. the effects of bedroom air quality on sleep and next-day performance. indoor air, vol.26, pp. 679-686. tsang, t., mui, k. & wong, l., 2021. investigation of thermal comfort in sleeping environment and its association with sleep quality. building and environment, vol.187, p. 107406. 1. introduction 2. methodology 2.1. survey location and buildings surveyed 2.3. thermal questionnaire and the respondents 3. results and discussion 3.1. outdoor and indoor conditions 3.2. subjective thermal responses 3.3. estimation of comfort temperature 3.4. comparison with international standards 3.5. evidence of adaptation 3.6. sleep quality 4. conclusion conflict of interest funding acknowledgment references chen060329.qxd the journal of engineering research vol. 6, no.1, (2009) 1-7 _________________________________________________ *corresponding author’s e-mail: drkhalidkhazal@yahoo.com selectivity and activity of iron molybdate catalysts in oxidation of methanol khalid khazzal hummadi*1, karim h. hassan2 and phillip c.h. mitchell3 *1al khawarizmi college of engineering, university of baghdad, al. jadiria, baghdad, iraq 2ministry of science and technology, al. jadiria, baghdad, iraq 3school of chemistry, university of reading, reading rg6 6ad, uk received 29 march 2006; accepted 21 april 2007 abstract: the selectivity and activity of iron molybdate catalysts prepared by different methods are compared with those of a commercial catalyst in the oxidation of methanol to formaldehyde in a continuous tubular bed reactor at 200-350 oc (473-623 ok), 10 atm (1013 kpa), with a methanol-oxygen mixture fixed at 5.5% by volume methanol: air ratio. the iron(iii) molybdate catalyst prepared by co-precipitation and filtration had a selectivity towards formaldehyde in methanol oxidation comparable with a commercial catalyst; maximum selectivity (82.3%) was obtained at 573ok when the conversion was 59.7%. catalysts prepared by reacting iron (iii) and molybdate by kneading or precipitation followed by evaporation, omitting a filtration stage, were less active and less selective. the selectivity-activity relationships of these catalysts as a function of temperature were discussed in relation to the method of preparation, surface areas and composition. by combing this catalytic data with data from the patent literature we demonstrate a synergy between iron and molybdenum in regard to methanol oxidation to formaldehyde; the optimum composition corresponded to an iron mole fraction 0.2-0.3. the selectivity to formaldehyde was practically constant up to an iron mole fraction 0.3 and then decreased at higher iron concentrations. the iron component can be regarded as the activity promoter. the iron molybdate catalysts can thus be related to other two-component moo3-based selective oxidation catalysts, e.g. bismuth and cobalt molybdates. the iron oxide functions as a relatively basic oxide abstracting, in the rate-controlling step, a proton from the methyl of a bound methoxy group of chemisorbed methanol. it was proposed that a crucial feature of the sought after iron(iii) molybdate catalyst is the presence of -o-mo-o-fe-o-mo-ogroups as found in the compound fe2(moo4)3 and for fe3+ well dispersed in moo3 generally. at the higher iron(iii) concentrations the loss of selectivity is due to the presence of iron oxide patches or particles which catalyze the total oxidation of methanol, and the loss of activity to blocking of molybdenum sites. keywords: selective oxidation, methanol, formaldehyde, iron molybdate, activity ∫ƒféã«ÿg i~°ùccg ‘ ~j~◊g äg~«ñdƒe øe ~yé°ùÿÿg πeé©dg á«dé©ah á«fé≤àfg …oéªm πyõn ~dén1*ø°ùm ¢ûµ«æg ëôc ,2 π°ûà«e ¢s ö«∏«a h3 áá°°uuóóÿÿgg∫ƒféã«ÿg i~°ùcg ‘ …qééàdg ~yé°ùÿg πeé©déh é¡àfqé≤eh áø∏àfl ¥ô£h iô° ùg ~j~◊g äg~«ñdƒe øe i~yé°ùÿg πegƒ©dg á«dé©a h á«fé≤àfg á°sgqo åëñdg gòg ‘ ” : ¤g ∫ƒféã«ÿg áñ°ùfh ∫éµ°séh ƒ∏«c 1013 ~æy §¨° dg äƒñãh ≥∏£e 623 ¤g 473 úh ìhgîj iqgôm álqo ~æy ƒ°ûùg »hƒñf’g πyéøÿg ωg~îà°séhh ~jé¡j~∏e qƒødg ¤g ¬m≤jô£h ô° ùg ~j~◊g äg~«ñdƒe øe ~yé°ùÿg πeé©dg ¿éh é¡«∏y ∫ƒ°ü◊g ” »àdg èféàædg ä~ccg .éªém %5.5 úé°ùch’g co-precipitationqƒø∏d á«fé≤àfg ≈∏yg ≈£yg éªæ«h …qééàdg ~yé°ùÿg πeé©dg ™e áfqé≤ÿéh ∫ƒféãÿg i~°ùcg ‘ (%59^7) ∫ƒëàdg ¿éc ée~æy ≥∏£e 573 iqgôm álqo ~æy é¡«∏y ∫ƒ°ü◊g øµá »àdgh (%82^3) ~jé¡j~dée ádg~c i~yé°ùÿg πegƒ©dg √ò¡d (á«dé©ødg á«fé≤àf’g) äébóy á°ûbéæe â“ ºk .á«dé©a πbgh á«fé≤àfg πbg â£yg ôîñàdg ºk ö«°sîdgh øé©dg á≤jô£h iô° ëªdg i~yé°ùªdg πegƒ©dg ä’é≤ÿg ™e åëñdg gòg ‘ é¡«∏y ∫ƒ°ü◊g ” »àdg éæéféàf §hôhh .i~yé°ùÿg πegƒ©dg √ò¡d ájh骫µdg ö«cgîdg áñ°ùfh á«ë£°ùdg ámé°ùÿg h ò° ëàdg ¥ô£h é¡àbóyh iqgô◊g álq~d -0^2 oh~ëh ~j~◊g øe ájƒäe áñ°ùf π° ag âféc å«m ~jé¡j~dée qƒa ¤g ∫ƒféã«ÿg i~°ùcg ‘ ~j~◊g h ωƒj~«ñdƒÿg øe πc §hgôj 샰vƒhh éfô¡xcg iqƒ°ûæÿg ´gîn’g ägagôhh á«dé©ød §°ûæªc iqéñàyg øµ‡ ~j~◊g ¿g éªc ~j~ë∏d á«dé©dg áñ°ùædg ~æy ¢véøîf’éh g~ñj ºk 0^3 ~j~ë∏d ájƒäÿg áñ°ùædg ájé¨d é«∏ªy áàhék âféc ~jé¡j~dée qƒø∏d á«fé≤àf’g ¿g h 0^3 ¢sé°sód øjôn’g úfƒµÿg ¤g oƒ©j ¿g øµá i~°ùc’g ‘ ~j~◊g äg~«ñdƒe øe ~yé°ùÿg πeé©dg á«dé©a ¿g ~yé°ùÿg πeé©dg3 moo¬fg å«m .â∏hƒµdgh çƒeõñdg äg~«ñdƒe πãe √~«c’g iqƒ°üdg éæmîbg ~bh ∫ƒfé㫪∏d …h骫µdg rgõàeód »°ùcƒã«ÿg iô°ug áyƒªû π«ãÿg øe ¿ƒjhôh π°üød √~°ùcg i~yé≤c ~j~◊g ~«°ùchg πª©j áyô°ù∏d io~ùg √ƒ£ÿg ‘ ™«eéée oƒlƒh ~yé°ùÿg πeé©∏d o-mo-fe-o-mo-o ¿ƒµªdg ‘ ä~lh éªc3)4 fe(moo~j~ë∏d »dé©dg õ«cîdg ~æy ~yé°ùÿg πeé©∏d á«dé©ødg ‘ ¿g~≤ødg öñ°s ¿gh .ωƒj~«ñdƒÿg ™bgƒe ~°u á«dé©a ¿g~≤ah ∫ƒfé㫪∏d á«∏µdg √~°ùc’g ‘ ~yé°ùj …òdg ~j~◊g ~«°ùchg äéñ«ñm oƒlh ƒg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.∫ƒfé㫪dg i~°ùccgg ,~j~ëdg äg~«ñdƒe ,i~°ùc’g á«fé≤àfg : 2 the journal of engineering research vol. 6, no.1, (2009) 1-7 1. introduction many researches deal with the oxidation of methanol using many different catalysts to produce the formaldehyde which has many important industrial chemical applications in the production of urea, melamine, phenolic resins and ethylene glycol; in the fertilizer and paints industries; and in medicine as a preservative (see, for example, the formaldehyde council, inc. (fci): formaldehyde.org). formaldehyde is manufactured by the selective oxidation of methanol with oxygen or air over a silver (jia-linang li, et al. 2000; wachs, i.e., 2003) or iron molybdate catalysts (farrauto, et al. 1997; soares, et al. 2001; and ivanov, et al. 2000). iron molybdate is a typical two-component selective oxidation catalyst. the two components, iron(iii) oxide and molybdenum(vi) oxide, interact to produce a catalyst which favours the desired product, formaldehyde, rather than the products of total oxidation of methanol, carbon dioxide and water, or other products like carbon monoxide or dimethyl ether. it is the combination of the two oxides which produces the desired active and selective catalyst. iron (iii) oxide by itself is unselective producing carbon dioxide and water; molybdenum trioxide is selective but with low activity (bowker, et al. 2002 and wang, et al. 2001). the overall reaction is ch3oh + 0.5o2 = ch2o + h2o the reaction is exothermic (∆h,= -159 kj mol-1) (diakov, et al. 2002 and dia kov, et al. 2001). the oxidation proceeds through reaction of methanol with the molybdate surface (mccarron iii, et al. 1986). at ambient temperature methoxy groups are formed and water is liberated. as the temperature is raised the methoxy groups begin to decompose forming formaldehyde and more water. the rate-limiting step is the abstraction of methoxy hydrogen via breaking of a c-c bond. the products rapidly desorbed leaving the molybdate phase reduced. re-oxidation is affected by oxygen so completing the catalytic cycle. the active sites are considered to be co-ordinatively unsaturated molybdenum atoms. the overall reaction is in accordance with the mars-van krevelen mechanism, i.e. oxygen is supplied to the reactant as lattice oxygen rather than by direct reaction with oxygen. the lattice oxygen is replenished by reactant oxygen. the technical catalyst composition is 80% moo3 and 20 % fe2o3 equivalent to an iron mole fraction 0.31. iron may be partially replaced by a promoter, eg. chromium. the active catalyst is considered to be fe2 (moo4)3 (soares, et al. 2001). the excess of moo3 is said variously to be required to ensure the stability of the catalyst towards loss of moo3, to maintain the active species and to enhance the surface area (soares, et al. 2001). we shall see that it may be more profitable to describe the catalyst as fe3+ dispersed in moo3. the chemical composition is the major factor determining the catalytic properties. however the catalytic characteristics depend on catalyst preparation method, reflecting changes in the extent of interaction of the iron and molybdenum oxides, the extent of dispersion, the surface area, the pore structure, and the crystal structure (boreskov, 1976). it should be stressed that catalyst structure depends also on other parameter such as metal loading, surface area of the carrier and drying and calcination temperatures (topsoe, et al. 1976). the method of preparation appears to have a significant impact on the activity and selectivity of the catalyst. there have been some studies of different preparation methods, for example sol-gel catalysts by co-precipitated catalysts (wand, et al. 2001; soares, et al. 1997; and soares, 2003). in the present work iron molybdate catalysts were prepared by three methods-kneading or slurry method, precipitation and evaporation, co-precipitation and filtration. the activities and selectivities in the oxidation of methanol were determined at 473-623ok and compared with the activity and selectivity of a commercial catalyst. 2. experimental work 2.1. catalyst preparation ammonium molybdate, (nh4)6mo7o24.4h2o; and iron (iii) nitrate nonahydrate, fe(no3)3 9h2o, analytical grade (>99% purity), were used. 2.1.1. kneading and evaporation: catalyst (1) to prepare 25 g of catalyst, 24.2 g of ammonium molybdate was added to amount of distilled water which is enough so that when added to 25.31 g of iron nitrate powder gradually with continuous stirring to obtain a homogenous paste. the paste was dried in an oven at 383ok for 2 hours to evaporate water and then calcined at 673773ok in a current of air for 4 hour. the calcined catalyst was crushed to small pieces, then to powder using a ball mill. the 100 mesh fraction was mixed with a solution of polyvinyl alcohol (3%) in distilled water. the produced paste was then dried, crushed, and sieved. pellets (or tablets) (7 mm diameter, 4 mm thick) were prepared in a tablet press at 2 atm. the pellets were calcined at 773ok to dissociate the polyvinyl alcohol because it is used as a binder only. 2.1.2. precipitation and concentration: catalyst (2) solutions of ammonium molybdate, 24.52 g in 250 ml of distilled water and iron(iii) nitrate 25.31 g in 150 ml of distilled water were mixed, the ph was ca 2. the precipitate formed was left to settle overnight at room temperature to get rid of excess water and then dried, calcined and formulated as for catalyst 1. 2.1.3. co-precipitation and filtration: catalyst (3) solutions of ammonium molybdate and iron(iii) nitrate were prepared and mixed as in the preparation of catalyst 2. the precipitate was filtered off and washed several times with distilled water until the ph of the filtrate reached 7. the solid was dried, calcined and formulated 3 the journal of engineering research vol. 6, no.1, (2009) 1-7 as before. 2.2. catalyst characterization the analysis of iron and molybdenum were determined by standard atomic absorption spectrophotometer (pye unicam sp9). the x-ray diffraction of powder of prepared and commercial catalysts were carried out by using a phillips xray diffractometer with cukα radiation (1.542 å, 40 kv, 20 ma) scanning over the range of angles of 3 to 70 on 2. peaks were assigned by comparison with the d spacing of typical compounds given in the astm powder diffraction files (powder diffraction file, swarthmore pennsylvania, 1978). surface areas of the catalysts were determined by the bet method (le page, 1987) (carloeraba sorptomic series 1800). pore volumes and densities were determined by the liquid impregnation method (satter filled and charles, 1980). hardnesses were determined with the erweka tbh28 hardness meter (le page, 1987). the bulk density is determined by replacing a weight catalyst in a graduated cylinder then shake it by hand on the table or by vibrator and read the volume of it and then calculate the density, it is the packing or load density of the reactor. the solid density is measured by weighting a symmetrical tablet and calculates its volume. it is carried out for several tablets and takes the average. 2.3. catalyst testing activities and selectivities of the catalysts in the conversion of methanol to formaldehyde were determined in a continuous flow pilot plant (fig. 1). the reactor was stainless steel (316-heat resistant), length 800 mm, inside diameter, 19 mm. heaters were in the form of four separately heat-controlled block shells. the reactor was packed with 110 ml of the catalyst between two layers of inert material such as glass, inert alumina or porcelene balls. test conditions were: reactor temperature 200 to 35°c (473-623ok); pressure, 10 atm (1013 kpa); flow rate, 15.858 cm3/s; methanol, 5.5% by volume in oxygen. in a typical run the reactor was purged with n2 gas and the temperature raised. after establishing steady state conditions, the reaction was started; the methanol-oxygen mixture was pumped upwards into the reactor. the products were passed through three traps in series filled with water cooled to 278ok to absorb formaldehyde, unreacted methanol and side products. analysis of the reaction products was carried out periodically after two hours collection of the samples. formaldehyde analysis was done by treating it with hypodidate solution which was prepared immediately by reacting iodine and naoh. then h2so4 was added to liberate iodine again some reacted with hcho and the excess is backing titrated solution (monti, et al. 1985). unconverted methanol was determined gas chromatographically with a povpak type t column at 420ok (monti, et al. 1985). 3. results and discussion the results of this work has been tabulated in table 1. it can be seen that the catalyst (3) (prepared by co-precipitation) is closest to the commercial catalyst in its composition and physical properties. they are structurally similar as shown by their x-ray diffraction patterns (fig. 2). the most obvious difference between the different preparations is the excess moo3, which is greatest for catalyst (3) table 1. these preparations along with the commercial catalyst enable us to assess the effect of the molybdenum content on the properties of the catalysts. excess molybdenum appears to have little effect on the catalyst density. however, the two catalysts with the highest molybdenum (catalyst (3) and the commercial catalyst) have the greatest surface areas, figure 1. flow diagram of laboratory oxidation of methanol unit 4 the journal of engineering research vol. 6, no.1, (2009) 1-7 pore volumes and hardness. the activity and selectivity in the oxidation of methanol to formaldehyde are calculated. the activity is the percentage of methanol converted to all products and the selectivity as the fraction of formaldehyde in the products, ie. formaldehyde produced divided by methanol converted. the aim in the catalysis is to maximize both the conversion and the selectivity at the lowest reaction temperature. activities and selectivities are shown plotted against the reaction temperature in fig. 3. the behaviors of the catalyst (3) are similar to that of the commercial catalyst. the activities of all catalysts rise with rising temperature and converge to roughly the same conversion at 598ok. the significant distinction between the catalysts is in the selectivity which passes through a maximum at 573ok with the commercial and the co-precipitated catalysts having the highest selectivities. the optimum combination of conversion and selectivity is provided by catalyst (3) and the commercial catalyst. figure 4 shows that the variation of surface area (which for our catalysts is ±1%) is not significant. the prepared catalytic results are consistent with the literature, eg. (soares, et al. 2001); activities at constant surface area tending to the same value independently on the iron (or molybdenum) content of the catalyst; selectivities passing through a maximum with increasing reaction temperature. the implication of this observation is that during use, the catalyst surface stabilizes to a certain composition during the period of 12 h required to attain the higher temcomposition and property catalyst catalyst ( 1) kneading and evaporation catalyst( 2) precipitation and concentration catalyst ( 3) co-precipitation and filtration commercial composition/wt % fe 17.2 15.6 13.8 14.0 mo 50.2 51.7 53.4 53.0 fe2o3 24.6 22.3 19.7 20.0 moo3 75.4 77.7 80.2 80.0 moo3 excess/wt -% c 2.32 4.57 7.12 6.52 mo/fe atomic ratio 1.70 1.93 2.25 2.20 fe/(fe + mo) mole fraction 0.371 0.341 0.308 0.312 colour yellow green yellow green yellow yellow pellet size/cm 0.9×0.9 0.9×0.9 0.9×0.9 0.45×0.4 surface area/m 2 g-1 283 282 287 290 pore volume/ cm 3 g-1 0.28 0.30 0.40 0.35 solid density/g cm -3 b 0.53 0.50 0.52 0.50 bulk density/g cm -3 b 1.05 1.05 1.05 1.10 hardness/10 5 dyne 1.70 1.63 2.1 2.3 table 1. composition and properties of the catalysts a from small industrial formaeldihyde plant in ministry of industry of iraq b excess moo3 over the stoichiometric composition fe2 (moo4)3 (ie. fe2o3 + 3moo3) figure 2. x-ray powder diffraction patterns of the catalysts 5 the journal of engineering research vol. 6, no.1, (2009) 1-7 peratures as shown in fig. 3. the effect of catalyst composition on the activity and selectivity has been discussed with reference to the investigation results and literature data (wachs, et al. 2000). the above discussion is in terms of excess of moo3. however, since moo3 is in excess it would seem logical to express the variation of catalyst composition in terms of iron added (or not) to molybdenum, ie. the fe/mo ratio or the fe/(fe+mo) mole fraction as for other two-component catalysts, for example, the cobalt-promoted molybdenum disulfide based hydrodesulphurization catalyst. it has been observed that in the model of two-component moo3based selective oxidation catalysts; the second component, e.g. bismuth or cobalt as in the selective oxidation of propene, enhances the activity by promoting the slow step, the first c-h bond breaking and the abstraction of h. the first step is catalyzed by the more basic oxide, bi2o3 or coo. we imagine that the greater basicity (or nucleophilicity) of the oxide of the basic oxide favors interaction of the hydrogen of a methyl group with oxide of the catalyst, and subsequent proton transfer to oxide. when we interpret the data correctly we shall see that the iron molybdate catalyst fits into this pattern. it can be seen also that there was a synergy between iron and molybdenum. unfortunately most researchers have not studied a wide range of fe/mo compositions. however, it can fit certain patent data (wachs, et al. 2000) to these data and thereby examine a wider range of compositions. activities and selectivities so obtained are plotted in fig. 5. activities and selectivities have been normalized of the commercial catalyst at 573ok to the values for a similar commercial catalyst reported (wachs, et al. 2000). for the activities it can be seen a typical volcano curve, the activity rising to a maximum value as iron is added to moo3 and then dropping off. this behavior demonstrates synergy between iron and molybdenum. beyond an iron mole fraction of 0.4 the activity begins to increase as fe2o3 takes over. the selectivity to formaldehyde is more or less constant until an iron mole fraction of 0.3 is reached. the selectivity then drops as fe2o3 becomes dominant. this behavior indicates that the selective catalyst is moo3. iron is an activity promoter. we can now understand more clearly how the iron molybdate catalyst works as also reported in the literature (bowker, et al. 2002). it can be suggested here that the crucial feature is to have fe3+ well dispersed in moo3. the formation of patches or particles of fe2o3 is to be avoided if the selectivity to formaldehyde is to be preserved. the loss of selectivity in fig. 5 at higher iron levels is evidently associated with formatted of fe2o3. the loss of activity may be attributed to the blocking of moo3 sites. the dramatic loss of activity and selectivity seen in fig. 5 represents a changeover from the preference for formaldehyde on a moo3 catalyst to total oxidation on fe2o3. the selective catalysis has often been associated with a definite compound, iron (iii) molybdate, fe2(moo4)3. the formation of iron (iii) molybdate seems inevitable in the synthesis of the catalyst. a structural fea35 40 45 50 55 60 65 450 500 550 600 650 t /k co nv er si on /% commercial cat alyst cat alyst 1 kneading cat alyst 2 precipit at ion cat alyst 3 co-precipit at ion 0.35 0.45 0.55 0.65 0.75 0.85 450 500 550 600 650 t /k se le ct iv ity figure 3. conversion (left) and selectivity (right) in conversion of methanol to formaldehyde 130 150 170 190 210 230 250 270 290 120 140 160 180 200 220 240 103conversion/surface area 10 5 s el ec tiv ity /s ur fa ce a re a figure 4. selectivity towards formaldehyde vs. conversion of methanol for iron molybdate catalysts. selectivities and activities are normalized to unit surface area. symbols as in figure 3 6 the journal of engineering research vol. 6, no.1, (2009) 1-7 ture of iron (iii) molybdate is the presence of -o-mo-ofe-o-mo-ogroups. such structures have been regarded as to represent the well dispersed fe3+. the presence of iron(iii) molybdate in our catalysts is indicated from the x-ray diffraction pattern. it had been deducted from the catalytic data that the catalysts with the highest activity and selectivity (commercial catalyst and catalyst 3) have the most favorable fe3+ dispersion. the method of preparation of catalyst (3) (co-precipitation and filtration) is likely lead to a better dispersion of fe3+ and the least likely formation of fe2o3 patches. 4. conclusions it was observed that the iron (iii) molybdate catalyst prepared in the present work by co-precipitation and filtration method has a higher selectivity towards formaldehyde in methanol oxidation comparable with a commercial catalyst. catalysts prepared by reacting iron (iii) and molybdate by kneading or precipitation followed by evaporation, omitting a filtration stage, are less active and less selective. by combining our catalytic data with data from the patent literature, it can be demonstrated that the synergy between iron and molybdenum in regard to methanol oxidation to formaldehyde with an optimum composition corresponding to an iron mole fraction of 0.2-0.3. the selectivity to formaldehyde is practically constant to an iron mole fraction 0.3 and then decreases at higher iron concentrations. if we think of the iron component as the activity promoter then the iron molybdate catalysts can be related to other two-component moo3-based selective oxidation catalysts, eg. bismuth and cobalt molybdate. the iron oxide then functions as a relatively basic oxide abstracting, in the rate-controlling step, a proton from the methyl of a bound methoxy group of chemisorbed methanol. the proposed crucial feature of the iron(iii) molybdate catalyst is the presence of -o-mo-o-fe-o-moogroups as found in the compound fe2(moo4)3 and for fe3+ which is generally the well dispersed in moo3. at the higher iron (iii) concentrations the loss of selectivity is due to the presence of iron oxide patches or particles which catalyze the total oxidation of methanol and the loss of activity to blocking of molybdenum sites. references boreskov, g.k., delmon, b., jacobs, p.a. and poncelet, g., 1976, "in preparation of catalysts," elsevier, amsterdam, pp. 223-227. bowker, m., holroyd, r., elliott, a., morrall, p., alouche, a., entwistle, c. and toerncrona, a., 2002, “catalysis letters,” vol. 83, pp. 165-176. catal, t., 1997, “mars van krevelen,” vol. 34, p. 40. dia kov. v., larfarge, d. and varma, a., 2001, "methanol oxidation dehydrogenation in a catalytic packed bed membrane reactor", catalysis today, vol. 76, pp. 159-167. diakov, v., blackwell, b. and varmay, a., 2002, "methanol oxidative dehydrogenation in a catalytic peaked-ped membrane reactor; experimental and model," chemical engineering science, vol. 57, pp. 1563-1569. farrauto, r.j. and bartholomew, c.h., 1997, “fundamentals of industrial catalytic processes,“ blackie academic and professional, london, p. 489. ivanov, k., mitov, i. and kruster., "selective oxidation of methanol on fe-mo-w catalysts," j. of alloys and compounds, vol. 309(14), pp. 57-60. jia-liang, li., wei-lin dai., ky dong, jingfa domg, 2000, " a new silver-containing ceramics for catalytic oxidation of methanol to formldehyde," materials letters, vol. 44(3-4), pp. 233-236. le pag, j.f. and miquel, j.in., delmon, b., jacobs, p.a. and poncelt, g., 1976, "preparation of catalysts," 0 10 20 30 40 50 60 0 0.2 0.4 0.6 0.8 1 fe mole fract ion co nv er si on /% 40 50 60 70 80 90 100 0 0.2 0.4 0.6 0.8 1 fe mole fract ion se le ct iv ity /% figure 5. conversion and selectivity at 573 k for our iron molybdate catalysts (squares) and other reported catalysts (23) (circles) vs. mole fraction of iron 7 the journal of engineering research vol. 6, no.1, (2009) 1-7 elsevier amsterdam, pp. 39-43. le page, j.f., 1987, "applied heterogeneous catalysis," technip paris. mccarron, e.m. iii., sleight, a.w., mitchell, p.c.h. and sykes. a.g., 1986, “the chemistry and uses of molybdenum,” proceedings of the climax fifth international conference, polyhedron synposia-inprint number 2, pergamon press, oxford, p. 129. monti, d., reller, a., and baiker, a., 1985, "methanol oxidation on k2so4promoted vanadium pentoxide; activity, reducibility and structure of catalysts,” j. of catalysis, vol. 93, pp. 360-367. powder diffraction file, 1978, “alphabetical listing,” swarthmore pennsylvania. satter filled and charles n., 1980, "heterogenous catalysis in practice," mc graw-hill, inc., new york. soares a.p.v, portela m.f.,kiennemann a.,hilaire l. and millet j.m.m., 2001, "iron molybdate catalysts for methanol to formaldehyde oxidation; effect of mo excess on catalystic behaviour," applied catalysis, vol. 206, pp. 221-229. soares a.p.v., portela m.f. and kiennemanu a., 2001, "iron molybdate catalysts for methanol to form aldehyde oxidation; effect of mo excess on the deactivation behaviour," catalysis communications, vol. 2, pp. 159-164. soares, a.p.v., portela, m.f., kiennemann, a. and hilaire, l., 2003, “chemical engineering science,” vol. 58, pp. 1315-1322. soares, a.p.v., portela, m.f. and kiennemann, a., 1997, “3rd world congress on oxidation catalysis,” vol. 110, 807-816. the formaldehyde council, inc. (fci): http://www.formaldehyde.org. topsoe, h., clausen, b.s., burriescin, candia r. and morups, 1976, "in preparation of catalysts,” elsevier, amsterdam, pp. 479-487. wachs i.e.,2003, "extending surface science studies to industrial reaction conditions; mechanism and kinetics of methanol oxidation over silver surface,” surface science, vol. 544, pp. 1-4. wachs, i.e. and briand, l.e., 2000, “us patent 6037290,” lehigh university, bethlehem, pa, usa. wang. c.t. and willey, r.j., 2001, “ journal of catalysis,” vol. 202, pp. 211-219. mech070131.qxd the journal of engineering research vol. 6, no.1, (2009) 51-58 1. introduction the main water supplies in the dhofar region (south of the sultanate of oman) come from a groundwater distribution system. the exploitation of the water system contributes to the underground intrusion of salt water into the coastal plain around salalah (salalah is the administrative capital of dhofar region). hence, as more ground _______________________________________ *corresponding author’s e-mail: sabah1@squ.edu.om water is extracted from the mountains, less water is available to the people living on the plain (alesh, 1998). moreover, the coastal plain (salalah & sea) is an important farming area. the groundwater is used for coconuts, bananas, vegetables and grass fodder cultivation. recently, there has been an increase in grass fodder cultivation using large-scale groundwater irrigation (kobayashi, 2001). in the mountain area (jabal dhofar), there is traditional livestock farming (mostly cows & goats) utilizing wild vegetation as fodder. in the arid area fog and rain water collection from trees in the dhofar region in the sultanate of oman s. a. abdul-wahab*a, h. al-hinaia, k. a. al-najarb, m. s. al-kalbanic amechanical & industrial engineering department, college of engineering, sultan qaboos university, p.o. box 33, p.c. 123, al-khoud, sultaname of oman bcivil aviation and meteorology, meteorology department, salalah airport, muscat, sultanate of oman ccivil & architectural engineering department, college of engineering, sultan qaboos university, p.o. box 33, pc 123, al-khoud, muscat, sultanate of oman received 12 may 2007; accepted 13 november 2007 abstract: this work had two purposes. the first was to measure and investigate the amount of fog and rain water collected by a sample of trees during the summer monsoon season in the mountains of the dhofar region in the south of oman. the second purpose was to assess the potential of trees with different leaf shapes, sizes and cross sectional canopy areas to collect fog and rain water in the area. in order to meet these purposes three different tree species, were selected for experimental investigation and field measurements. they included fig, lemon, and tamarind trees. the experiments were designed and implemented in the monsoon season of 2006, between the 13th of july and the 3rd of september. the daily measurements of water collection showed that the three species of tree have different water collection capacities. it was found that the fig tree collected the least water. the fog water collection of the fig tree over a period of 47 days between the 13th of july and the 3rd of september was 140.5 l/m2, or an average of 2.7 l/m2/d. the fog water collection of the lemon tree was 243.0 l/m2, or an average of 4.4 l/m2/d. in terms of potential collection of the tamarind tree, the results showed that it collected a fog water of 218.9 l/m2, or an average of 4.3 l/m2/d over the same period. the study contributes to the knowledge of how different tree species collect fog and rain water, and concludes with a set of recommendations. keywords: fog water collection by trees, fig, lemon, tamarind monsoon season, dhofar region, oman ¿éªy áæ£∏°s ‘ qéøx á≤£æe øe qéé°t’g øe qé£ec’g h üéñ° dg √é«e ™ªl êéñ∏µdg .¢s ~ªfi h qééædg .cg ~dén ,»féæ¡dg ∫óg ,üégƒdg~ñy .cg ìéñ°u áá°°uuóóÿÿgg¬«ø«°üdg ¬«ª°sƒÿg ìéjôdg º°sƒe ∫ón qéé°t’g øe ¬æ«y é¡©ªœ »àdg qé£e’g h üéñ° dg √é«e ᫪c øe ≥≤ëàdgh ¢sé«≤dg πlg øe ¿éc ∫h’cg .¿é°vôz ¬d ¿éc πª©dg gòg ¿eg : √é«e ™ªl ‘ (¬mé°ùÿgh ºé◊gh πµ°ûdéh áø∏àıg ¥gqh’g ägp) qéé°t’g iq~≤e º««≤j πlg øe ¿éµa êéãdg ¢vô¨dg éecg .(¬dó°u ∞jôn) ¿éªy üƒæl ‘ qéøx á≤£æe ∫éñl ‘ â∏ª°t ~≤a .é¡«∏y ¬«fg~«ÿg äé°sé«≤dg h üqééàdg agôlg ±~¡h áø∏àıg qéé°t’g øe ´gƒfg ákók qé«àng ” ~≤∏a ,ú°vô¨dg øjòg ≥«≤– πlcg øeh .á≤£æÿg ‘ qé£e’g h üéñ° dg øe ådéãdgh ƒ«dƒj øe ô°ûy ådéãdg úh ,2006 ωé©∏d ᫪°sƒÿg ìéjôdg º°sƒe ‘ äòøfh ⪪°u ~b á«fg~«ÿg üqééàdg ¿eg .…~æ¡dg ôªàdgh ,¿ƒª«∏dg h ,úàdg qéé°tg ¬°sgq~dg √òg üéñ° dg √é«e ᫪c ¿eg .√髪∏d é©«ªœ qéé°t’g πbg úàdg iôé°t ¿cg ~lh ~bh .™«ªéàdg ‘ áø∏àfl äé©°s é¡d ákóãdg qéé°t’g ¿éh ᩪûg √髪∏d á«eƒ«dg äé°sé«≤dg âdo ~bh .èªàñ°s 140^5 âféc èªàñ°s øe ådéãdgh ƒ«dƒj øe ô°ûy ådéãdg úh éeƒj 46 iîa i~e ≈∏y é¡©«ªéàh úàdg iôé°t âeéb »àdg2m/l 2^7 ∫oé©j éã hgd/2m/l√é«e ᫪c ¿eg. 243^0 âféc ¿ƒª«∏dg iôé°t ᣰsgƒh ¬©ªûg üéñ° dg2m/l4^4 ∫oé©j éã hcg ,d/2m/l~b é¡féh ~gƒ°t ~≤∏a …~æ¡dg ôªàdg iôé°ûd á«©«ªéàdg iq~≤dg á«méf øe éeg . 218^9 ⩪l2m/l 4^3 ∫oé©j éã éã hg ,äéñ° dg √é«e øed/2m/lüéñ° dg √é«e ™«ªœ ‘ qéé°t’g ±óàng á«ø«c ¬aô©e ‘ á°sgq~dg √òg ºgé°ùjh .iîødg ¢ùøf ∫ón .äé«°uƒàdg øe áyƒªéã ºààîjh gáá««mmééààøøÿÿgg ääggooôôøøÿÿ.¿éªy ,qéøx á≤£æe ,᫪°sƒÿg ìéjôdg º°sƒe ,…~æ¡dg ôªàdg ,¿ƒª«∏dg ,úàdg ,qéé°tc’g ᣰsgƒh üéñ° dg √é«e ™ªl : 52 the journal of engineering research vol. 6, no.1 (2009) 51-58 (nejid) there is traditional nomadic camel farming. these heavy demands on the water supplies stress the need to investigate all possible sources of water which may contribute to the alleviation of this situation (cowiconsult, 1992). non-conventional water supplies, such as fog water collection, offer possible solutions to this problem. fog water collection has the ability to provide an alternative source of water, while at the same time reducing dependence on traditional water sources such as groundwater. the dhofar region, which is located between latitudes 16 and 20°n occupies nearly one third of the sultanate's total area, experiences the annual phenomenon of alkhareef (ie. monsoon season) each summer between june and september. this region of oman is touched by the south west monsoon and experiences by rain and thick fog caused by wet air from the indian ocean. generally, fog occurs on 20% to 25% of days during the year. this fog has the potential to provide an alternative source of water. it can be collected during the monsoon season (from midjune to mid-september) and can provide substantial volumes of water for domestic use and small-scale farming in the mountains. hence, exploitation of this phenomenon and the development of some form of fog harvesting could provide a valuable additional source of water where it is most needed. for example, the collected fog water could be used for the reforestation of the mountains, thus reducing the dependence on groundwater. the first published studies on water collection from the monsoon clouds of the jabal dhofar were developed by fallon (1978). he used wire netting to intercept water and increase the precipitation. stanley price, et al. (1985) started new investigations of fog water collection and the results indicated that up to 50 l/d could be collected by a 1 m2 screen of aluminium wire mesh. a research project into fog collection technology was then undertaken in the dhofar region during the 1989 monsoon season by (barros and whitcombe, 1989). the study investigated the possible applications and benefits of fog water collection techniques. the results of the study in the coastal mountain region showed the feasibility of this technology for collecting water. it was demonstrated that in the upper elevations from 900 to 1000 m, the average collection rates, using 1 m2 standard collectors and much larger collectors, of 30 to 40 l/m2/d were obtained over a 2 month period. this compared well with a study in chile where 6 l/m2/d was collected over a period of 12 months (barros and whitcombe, 1989). however, the pipe cost from collectors to the point of use was one of the major infrastructure costs that made the system uneconomic and hydraulically difficult. for the same area, a second research project was carried out during the monsoon in 1990. a number of fog-collection experiments were undertaken in this project in the coastal mountain region. the results of the experiments indicated the feasibility and applicability of the fog collection technology for collecting water. abdulwahab, et al. (2006a) examined the potential of fog water usage in certain jabal areas (mountains) of the dhofar region when fog collectors were built in the vicinity of the houses in the mountains. the aim of this work was to develop residential-type fog collectors to directly meet the needs of the mountain residents during the monsoon season, and to decrease the abstraction of groundwater. costs were reduced by constructing the collectors close to the point of use. the collected water was used directly in the storage tanks, which were installed in the houses, to irrigate plants or to supply animals and people with water on a daily basis. a detailed examination of several other fog collection studies conducted in the sultanate of oman can be found elsewhere (abdul-wahab, et al. 2006a, 2006b; abdul-wahab and lea, 2006; alesh, 1998, 2003). the collection of water from trees is a long established traditional practice. the experiments carried out in 1989 and 1990 in the dhofar region of the sultanate of oman clearly indicated the ability of trees to collect water in significant amounts (barros and whitcombe, 1989; cowiconsult, 1992). the experiments demonstrated the water collecting ability of trees in the jabal. they also showed the potentially important role of trees in aquifer recharge in mountainous regions. this was based on the fact that much of the water collected by trees during the monsoon exceeded their immediate needs. thus a considerable amount of water is available to seep into the soil and to percolate through into the local groundwater aquifers (barros and whitcombe, 1989; cowiconsult, 1992). cowiconsult (1992) reported that the trees play an important role in the recharge of groundwater. water which percolates through the soil is usually consumed by the vegetation and so benefits the plant growth and thereby grazing on the jabal. this means that water percolating into the groundwater aquifers is likely to benefit groundwater abstraction in either the jabal or the salalah plain. when interpreting the results in terms of potential collection per tree of a given canopy size, many factors required to be considered. among these factors are: the actual size of the canopy, its projection at right angles to the wind direction, the effect of the three dimensional canopy, the amount of collected water reaching the cistern, any evaporation during the monitoring period, the effect of the leaf size, branch and twig size, density, and the effect of the surrounding trees (cowiconsult, 1992). the present study was designed to derive more information on collection of fog and rain water by trees during the summer monsoon season 2006 at dhofar region in the sultanate of oman. in order to meet this objective, a house in the mountains of dhofar region (at ashta) was selected. at this house, fog collection experiments were carried out by using three species of tree including fig, lemon, and tamarind trees. the amount of water collected by each tree was measured in a cistern constructed under each tree. the information derived from this study will be useful in future water harvesting initiatives. 2. methodology a suitable house at ashta in the mountain of the dhofar region was selected for the fog collection experiments 53 the journal of engineering research vol. 6, no.1, (2009) 51-58 (fig. 1) by taking into account four factors, namely, (a) the potential for collecting large volumes of water, (b) the proximity of a water-poor community, (c) accessibility and (d) security. at this house at ashta, three specific trees were selecteda fig, a lemon and a tamarindto harvest the fog and rain water during the monsoon season. characteristics of the three trees used in the study together with the dimensions of their cisterns are summarized in table 1. it can be seen from this table that the approximate cross sectional area of the tamarind tree (42.76 m2) was higher than that of fig (16.18 m2) and lemon tress (20.84 m2). under each tree a cistern was constructed to serve as collecting tank. the purpose of this cistern was to hold fog and rain water harvested by the tree during the monsoon. the cistern was constructed in such away that it was big enough for the size of the tree over it. this in fact ensured that branches of the tree did not extend beyond the rim of the tank and water thus did not drip outside the cistern. it was very important to make sure that the cistern collected all the water that the tree harvested from the fog. it should be noted that the local residents were involved to assist with the construction of the cistern underneath the three trees. the experiments at ashta house on the three trees started on 13th of july and lasted until 3rd september 2006. during the foggy conditions or rain episodes, small fog droplets were deposited on the tree and coalesced to form larger drops which flowed downwards under the force of gravity into the cistern fixed below the tree. the water was then channeled to a storage tank located in the house. moreover, a standard rain gauge was placed in situ at the project site to measure the daily amount of rainfall collected on the site. the purpose of this rain gauge was to determine the contribution of fog relative to the total water yield. it was assumed that whenever rain was recorded by the rain gauge, all the water deposited on the trees originated exclusively from rainfall. on the other hand, if no rain was recorded, all the water collected by the trees was assumed to originate from fog alone. although this technique underestimates the contribution by fog, it does give some indication of the relative contributions of fog and figure 1. map of the sultanate of oman and the project location at ashta house in the mountain of dhofar region 54 the journal of engineering research vol. 6, no.1, (2009) 51-58 rainfall to the total volume of water collected. therefore, rainwater collection was subtracted from the total fog and rainwater collection to yield the fog water collection. 3. experimental results 3.1. fog and rainwater collection data tables 2, 3, and 4 illustrate the total and net fog water collected daily from the three trees during july, august and september 2006, respectively. it should be noted that the net fog water collected was determined by taking the difference between the total water collected (ie. both rain and fog water collected by tree) and rainwater. it can be seen from these tables that the three sample trees yielded considerable difference with respect to total water collected. in general, the amount of total water collected by lemon tree was the highest, followed by tamarind tree, and finally the fig tree. the daily mean total water collected by fig, lemon, and tamarind trees during july was 5.9, 7.2, and 5.4 l/m2/d, respectively (table 2). the monthly total fog and rainwater collected by these three trees during july was 112.0, 136.5, and 103.3 l/m2, respectively. during august, the total water collected daily by fig, lemon, and tamarind trees was 2.9, 5.3, and 5.5 l/m2/d, respectively (table 3). again, the results of water collection showed that the fig tree collected the least water of type of tree characteristics of the studied trees fig lemon tamarind length of the canopy (m) 4.42 4.88 6.10 width of the canopy (m) 3.66 4.27 7.01 cross sectional area of the canopy (m 2) 16.18 20.84 42.76 height of tree from ground to upper branch (m) 6.71 4.08 5.18 height of tree from ground to the low er branch (m) 1.52 0.82 1.16 length of the leaf (cm) 5.0 5.7 3.5 width of the leaf (cm) 3.5 2.7 1.5 approximate area of the leaf (cm 2) 17.50 15.39 5.25 dimension of cistern beneath the tree (cm 3) 310x310x15 310x310x15 400x400x15 table 1. characteristics of trees used in the study total fog and rainwater collected (l/m2/d) net fog water collected (l/m2/d) date fig lemon tamarind rainwater collected (l/m2/d) fig lemon tamarind 13 12.1 13.1 0.0 0.4 11.7 12.7 14 10.2 10.9 0.0 0.6 9.6 10.3 15 10.5 4.4 0.0 1.2 9.3 3.2 16 3.5 4.1 3.3 0.4 3.1 3.7 2.9 17 1.0 0.7 0.9 0.2 0.8 0.5 0.7 18 0.0 0.0 0.0 0.0 0.0 0.0 0.0 19 0.0 0.0 0.0 0.0 0.0 0.0 0.0 20 2.9 3.7 5.2 0.2 2.7 3.5 5.0 21 8.4 8.7 5.2 0.4 8.0 8.3 4.8 22 8.3 10.9 6.5 0.4 7.9 10.5 6.1 23 4.7 10.7 4.4 1.0 3.7 9.7 3.4 24 2.6 3.5 2.6 2.2 0.4 1.3 0.4 25 0.7 0.7 1.2 0.2 0.5 0.5 1.0 26 4.0 5.9 5.5 0.6 3.4 5.3 4.9 27 10.4 22.2 18.4 7.0 3.4 15.2 11.4 28 12.7 18.0 18.8 5.8 6.9 12.2 13.0 29 9.7 11.2 14.2 5.0 4.7 6.2 9.2 30 3.9 1.6 8.5 0.8 3.1 0.8 7.7 31 6.5 6.1 8.4 1.0 5.5 5.1 7.4 average (l/m2/d) 5.9 7.2 5.4 1.44 4.5 5.7 4.9 total (l/m2) 112.0 136.5 103.3 27.4 84.6 109.1 78.1 table 2. total water collected and net fog water collected by the three trees during july 2006 55 the journal of engineering research vol. 6, no.1, (2009) 51-58 the three trees. the total water collected by fig tree was 89.6 l/m2, and that of the lemon tree was 163.6 l/m2, and that of the tamarind tree was 169.5 l/m2. during the three days of september, the amount of fog and rainwater collected by fig was 2.3 l/m2/d, by lemon tree was 3.7 l/m2/d, and by the tamarind tree was 4.0 l/m2/d (table 4). again, the results indicated that the fig tree collected least water of the three trees. the total water collected by the fig tree was only 7.0 l/m2, while that collected by the lemon and tamarind trees was 11.0 l/m2 and 12.0 l/m2, respectively. it should be noted that in september the fog was not so frequent. this explains the lower amounts of water collected. in terms of potential collection per tree, the results indicated that the fig tree collected a total of 208.6 l/m2 over a period of 47 days. the lemon and tamarind trees collected a total of 311.1 l/m2 and 248.8 l/m2, respectively, over the same period. 3.2. fog collection data the daily variations of fog water collected by the three trees during july and august are presented in figs. 2a and 2b, respectively. it is interesting to note that the amount of fog water collected by each tree differed considerably. in general, the amount of fog water collected daily by lemon tree was the highest, followed by the tamarind and fig trees. thus, the lemon and tamarind trees were more efficient than the fig tree in harvesting fog water. looking at figs. 2a and 2b, it can also be seen that the daily amount of fog water collected in july was lower than that collected in august. in certain day of august, the fog coltotal fog and rainwater collected (l/m2/d) net fog water collected (l/m2/d) date fig lemon tamarind rainwater collected (l/m2/d) fig lemon tamarind 13 12.1 13.1 0.0 0.4 11.7 12.7 14 10.2 10.9 0.0 0.6 9.6 10.3 15 10.5 4.4 0.0 1.2 9.3 3.2 16 3.5 4.1 3.3 0.4 3.1 3.7 2.9 17 1.0 0.7 0.9 0.2 0.8 0.5 0.7 18 0.0 0.0 0.0 0.0 0.0 0.0 0.0 19 0.0 0.0 0.0 0.0 0.0 0.0 0.0 20 2.9 3.7 5.2 0.2 2.7 3.5 5.0 21 8.4 8.7 5.2 0.4 8.0 8.3 4.8 22 8.3 10.9 6.5 0.4 7.9 10.5 6.1 23 4.7 10.7 4.4 1.0 3.7 9.7 3.4 24 2.6 3.5 2.6 2.2 0.4 1.3 0.4 25 0.7 0.7 1.2 0.2 0.5 0.5 1.0 26 4.0 5.9 5.5 0.6 3.4 5.3 4.9 27 10.4 22.2 18.4 7.0 3.4 15.2 11.4 28 12.7 18.0 18.8 5.8 6.9 12.2 13.0 29 9.7 11.2 14.2 5.0 4.7 6.2 9.2 30 3.9 1.6 8.5 0.8 3.1 0.8 7.7 31 6.5 6.1 8.4 1.0 5.5 5.1 7.4 average (l/m2/d) 5.9 7.2 5.4 1.44 4.5 5.7 4.9 total (l/m2) 112.0 136.5 103.3 27.4 84.6 109.1 78.1 table 3. total water collected and net fog water collected by the three trees during august 2006 date total fog and rainwater collected (l/m2/d) net fog water collected (l/m2/d) fig lemon tamarind rainwater collected (l/m2/d) fig lemon tamarind 1.0 2.5 3.3 5.4 0.4 2.1 2.9 5.0 2.0 3.5 6.2 5.5 0.2 3.3 6.0 5.3 3.0 1.0 1.5 1.1 0.3 0.7 1.2 0.8 average (l/m2/d) 2.3 3.7 4.0 0.3 2.0 3.4 3.7 total (l/m2) 7.0 11.0 12.0 0.9 6.1 10.1 11.1 table 4. total water collected and net fog water collected by the three trees during september 2006 56 the journal of engineering research vol. 6, no.1, (2009) 51-58 lection by the lemon tree was as high as 19.5 l/m2/d (7 august 2006). the average fog water collected by fig, lemon, and tamarind trees during july was 4.5, 5.7, and 4.9 l/m2/d, respectively (fig. 3a). the average fog collection produced by tamarind tree on august was 4.9 l/m2/d, whereas the fig and lemon trees produced a yield of 4.5 and 5.7 l/m2/d, respectively (fig. 3b). as shown in fig. 4a, the total fog water collected in july by fig, lemon, and tamarind trees in the same order was 84.6, 109.1 and 78.1 l/m2. the total fog collected by these trees during august in the same order was 49.8, 123.8, and 129.7 l/m2 (fig. 4b). over a period of 47 days in the monsoon period, the tamarind tree collected 218.9 l/m2 of fog water, whereas the fig and lemon tress collected 140.5 l/m2 and 243 l/m2, respectively. 4. discussion and conclusion the main aim of this study was to assess the water collection potential of trees and to investigate the fog water potential of the selected area in the mountains of the dhofar region in the sultanate of oman. the water collection efficiency of three different trees was analyzed during the period from 13th july to 3rd september 2006 (ie. 47 days). the experiments were conducted to study the possibilities of harvesting the water of the monsoon fog by different trees for possible use as a supplement to the water sources of the mountains of dhofar region. the trees included in the study were fig, lemon and tamarind trees. figure 2a. amount of fog water collected daily by the three trees during july 2006 figure 2b. amount of fog water collected daily by the three trees during august 2006 57 the journal of engineering research vol. 6, no.1, (2009) 51-58 figure 3a. average fog water collected by the three trees during july 2006 a ve ra ge o f fo g w at er c ol le ct ed ( l /m 2 / d) fig lemon tamarinds 0 1 2 3 4 5 6 fig lemon tamarindsa ve ra ge o f fo g w at er c ol le ct ed ( l /m 2 / d) 0 1 2 3 4 5 6 figure 3b. average fog water collected by the three trees during august 2006 fig lemon tamarinds to ta l a m ou nt o f fo g w at er c ol le ct ed ( l /m 2 ) 0 20 58 the journal of engineering research vol. 6, no.1, (2009) 51-58 the comparisons of the water collection from the three different tress showed that large differences in collection existed as a function of the size of the leaves of the trees and the cross sectional area of their canopies. it was found that the fig tree collected the lowest amount of water. the results indicated that the total water collected during the experimental period of 47 days by the lemon tree was higher (311.1 l/m2) than the amount of water collected by tamarind (284.8 l/m2) and fig trees (208.6 l/m2) over the same period. the findings showed that the fog and rain water collected by the trees during the monsoon season provided significant amounts of water. the study also demonstrated that trees could collect amounts of water which evidently far exceed their consumption during the monsoon. the water collected can be used to supplement or replace the groundwater. the study confirmed that trees play an important role in the recharge of water to soil and groundwater. it can be concluded from this study that trees have the potential to harvest water from fog. there is no doubt that tree planting is important for water conservation in the dhofar region. more researches are required to identify areas where largest collection will take place and where tree planting will have the largest effect on the water resources. therefore, more research is required to determine the optimal locations for a forestation in the dhofar region. acknowledgments the funding provided by sultan qaboos university (project ig/eng/mied/05/02) is gratefully acknowledged. references abdul-wahab, s.a., al-hinai, h., al-najar, k.a. and alkalbani, m.s., 2007a, "feasibility of fog water collection: a case study from oman," j. of water supply: research and technology (aqua), vol. 56(4), pp. 275-280. abdul-wahab, s.a., al-hinai, h., al-najar, k.a. and alkalbani, m.s., 2007b, "fog water harvesting: quality of fog water collected for domestic and agricultural use," environmental engineering science, vol. 24(4), pp. 446-456. abdul-wahab, s.a. and lea, v., 2007, "reviewing fog water collection eorldwide and in oman: between theory and practice," under review. alesh, s. a., 1998, "work done on fog collection (moisture capture) in the sultanate of oman," proceedings of 1st int. conference on fog and fog collection. vancouver, pp. 417-420. alesh, s. a., 2003, "an experimental study of fog water collection in salalah area, sultanate of oman," a project/thesis submitted in partial fulfillment of the requirements for the degree of master of science in hydrology, college of science. sultan qaboos university, sultanate of oman. barros, j. and whitcombe, r. p., 1989, "fog and rain water collection in southern region," planning committee for development and environment in the southern region, sultanate of oman. cowiconsult, consulting engineering and planners as. 1992, "dhofar khareef studies feasibility of fog and rainwater collection and guidelines for pilot projects," final main report. planning committee for development and environment in the dhofar governorate, sultanate of oman. fallon, e.f., 1978, "mist harvesting, jebel qara, dhofar province," al mawared al tabeieiah, vol. 4(8), pp. 4-10. kobayashi, h., 2001, role of material flow analysis in sustainable development: material flow of livestock farming in the dhofar region of oman," technology and development, vol. 14, pp. 21-31. stanley price, m. r., al-harthy, a. h. and whitcombe, r. p., 1988, "fog moisture and its ecological effects in oman," in arid lands: to-day and tomorrow. research and development conference, tucson, arizona, usa, october 1985: pp. 69-88. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 62-71 *corresponding author’s e-mail: luizatdecastro@gmail.com doi:10.53540/tjer.vol18iss2pp62-71 users’ assessment of personal fans in a warm office space in brazil m. andré , c. buonocore , l. de castro*, and r. lamberts laboratory for energy efficiency in buildings, florianopolis, brazil abstract: the use of personal fans allows improving thermal comfort and energy savings in warm office spaces. this is due to individual adjustment and extended indoor temperature acceptability. however, to achieve that, the usability of fans must be assured. therefore, an experiment with 40 people of various age groups was carried out to assess four types of fans, one of which is an evaporative cooling device. the goal was to find out which criteria should be used for selecting a fan to implement in an office space. results show that airflow sensation and speed adjustment are considered the most important, although, noise is also very important, and cost can be an eliminatory criterion. the evaporative device was the best rated even in a space with 70 to 80% relative humidity, as users considered it to have a smooth controllable airflow. the results highlight these aspects should be considered in the selection of a personal fan and could also drive the industry to improve fans design for increasing usability and expanding the use of these systems. keywords: desk fan; thermal comfort; office; warm environment; personal conditioning system. تقییم المستخدمین للمراوح المكتبیة في المكاتب الدافئة في البرازیل و ر. المبرتس *م. أندریھ و ك. بونوكور و ل. دي كاسترو یسمح استخدام المراوح المكتبیة بتحسین الراحة الحراریة وتوفیر الطاقة في المساحات المكتبیة الدافئة؛ نظرا لما :لملخصا إمكانیة التعدیل الفردي لدرجة حرارة الغرفة، وتمدید المدى المقبول من درجات الحرارة الداخلیة. لتحقیق ھذه یوفره ذلك من شخصا من مختلف 40المكاسب، یجب أن یتم التحقق من فعالیة استخدام المراوح المكتبیة؛ ولھذا فقد أجریت دراسة على ضمنھا مروحة التبرید بالبخار. ھدفت ھذه الدراسة لتحدید المعاییر التي األعمار لتقییم أربعة أنواع مختلفة من المراوح من یجب األخذ بھا عند اختیار المراوح المكتبیة، وأظھرت النتائج أنھ وعلى الرغم من أھمیة الضوضاء التي تصدرھا المروحة، أھمیة للمستخدمین، كما أظھرت الدراسة إال أن اإلحساس بتدفق الھواء، وإمكانیة تعدیل سرعة المروحة یعتبران العاملین أكثر كذلك أن معیار التكلفة لم یكن ذا أھمیة كبیرة. حصلت مروحة التبرید بالبخار على إشادة أغلب المستخدمین حتى في المكاتب %؛ وذلك ألنھا تتمتع بتدفق سلس للھواء من وجھة نظر المستخدمین.80% إلى 70ذات الرطوبة النسبیة التي تتراوح بین تؤكد نتائج ھذه الدراسة على المعاییر التي یجب األخذ بھا عند اختیار المراوح المكتبیة، كما یمكن أن تدفع بالمصانع إلى تحسین تصمیم المراوح المكتبیة، مما یؤدي إلى زیادة استخدامھا. التكییف الشخصي.بیئة دافئة؛ نظام ؛مكتب ؛الراحة الحراریة مروحة المكتب؛ الكلمات المفتاحیة: users’ assessment of personal fans in a warm office space in brazil 63 1. introduction in the face of the prospect of global warming, it is important to rethink the way we condition buildings, so lower energy-consuming strategies are expanded. to do that, conditioning design could change the focus from room to microclimate conditioning by applying stimuli close to the body. personal conditioning systems (pcs) allow local adjustment of thermal conditions, enabling a group of people in the same space to control their microclimate according to personal demands (brager, zhang and arens, 2015). in addition, local stimuli can generate alliesthesia, which produces overall thermal comfort (de dear, 2011) with much lower energy consumption than needed for conditioning the total air volume of a room (xu et al., 2017). this approach allows the extension of cooling setpoint temperature, which could produce up to 70% energy savings (hoyt, arens and zhang, 2015). many types of personal devices have been proposed and studied in the last decade (andré, de vecchi and lamberts, 2020b). however, desk fans are considered one of the most efficient devices for warm conditions (m. he et al., 2017; luo et al., 2018; warthmann et al., 2018). they are also easy to implement for being independent of the cooling system infrastructure (boerstra, 2010). previous studies indicate occupants find 30 °c acceptable when they have desk fans (mishra, loomans and hensen, 2016; warthmann et al., 2018) since the increment of air movement reduces the warm sensation up to 3 °c (zhang, arens and zhai, 2015). in shared office spaces, it is common for the occupant to buy the fan when he/she is uncomfortable (boerstra, 2010), but the energy savings are not perceived by him/her, because the building's consumption is paid by his/her employer (y. he et al., 2017). therefore, to achieve the energy savings potential allowed by the extension of setpoint temperature, fans usability and attractiveness are very important to increase occupants’ willingness to use them and meet users’ needs. knecht et al. (2016) indicate usability can be influenced by aesthetics, ease of use of controls and the level of adjustability provided by the device. so, design issues can decrease the device usability, hindering its potential to improve users’ thermal comfort and energy savings. some design issues have already been identified in previous studies. andré et al. (2020a) identified users avoided increasing the fan speed because it also increased the noise level, causing an acoustic nuisance. schiavon et al. (2017) indicate that, in shared office spaces, fan noise might be more annoying to the person who is not using it, as no positive effect is perceived by him/her. therefore, the multiple domains of comfort must be considered as thermal, visual, acoustic and air quality may influence each other (schweiker et al., 2020). another important aspect is fan airspeed adjustment limitation, which may constrain the maximum airspeed and the fine-tuning, as usually, fans have fixed speed levels. in some studies, it was identified that users wanted higher airspeed, but did not increase it to the maximum possible level (zhai et al., 2013), probably because they preferred an intermediate speed level that could not be set by the device. in warmer environments, users indicated a preference for more airspeed even though the maximum speed level was selected (m. he et al., 2017; zhai et al., 2017). this indicates that the maximum speed achieved by the fan was not enough, which may have limited the temperature acceptability. to achieve a higher body cooling effect, the stimuli of the device should target the torso and face (zhang et al., 2010). these are usually affected by desk fans (schiavon and melikov, 2009; simone et al., 2014). however, depending on the fan size and vertical rotation adjustment capability the air jet might hit only the belly and arms (andré, de vecchi and lamberts, 2020a). thus, rotation adjustment in the vertical axis is also important to boost the fan effect. on the other hand, fan size has two implications – restriction of the affected surface area and adaptation to a workstation, where the space available is usually limited for each person. schiavon and melikov (2009) found that increasing the affected body surface area increases heat loss and fan cooling effect. however, in a shared workspace, an individual table area is limited and occupied by paperwork, computer and other supplies that constrain available space, so smaller devices are usually easier to implement. as identified in the literature, design aspects can influence desk fans’ usability. however, few studies were found comparing devices to address these issues. therefore, this study aims to identify the criteria users find most important when choosing and using a personal ventilative device. these criteria could be used for proposing guidelines for the industry and designers to improve this type of device. it could also help researchers and users to select devices with better usability. 2. method to assess users’ acceptability and willingness to use personal fans in shared office spaces, an experiment was set in the laboratory for energy efficiency in buildings (labeee) of the federal university of santa catarina. the building is in florianopolis, a city in the southeast of brazil with a climate classified as subtropical by köppen-geiger (peel, finlayson and mcmahon, 2007) and as 2a by ashrae 169 (2020). ten 2-hour sessions were carried out in february and march 2020, as they are summer months. four people – working on laptops – were included in each section. the experiment room is 17 m² with two external masonry walls and lightweight internal partitions (drywall and plywood with acoustic insulation). windows were shaded externally by fixed shading and internally by blinds, which were controlled by the researchers during the experiment to allow diffused m. andré, c. buonocore, l. de castro and r. lamberts 64 daylight and prevent direct solar radiation. throughout each experiment session, environmental thermal conditions were measured with data loggers (hobo ux-100) at three different points, one inside the room, one in the hallway and another in the building corridor. this way, the different thermal conditions participants would be exposed to during the experiment could be registered. sensors were turned on 30 minutes before the beginning of each session, and they recorded air temperature (from 20 °c to 70 °c ± 0.21 °c) and relative humidity (from 15% to 95% ± 3.5%) every minute. experiment procedure each session followed the procedures summarised in figure 1. the sessions started when participants entered the room, opened the laptops, and filled a personal information questionnaire (infoq). then, participants received instructions about procedures. the experiment was developed so that each occupant could use four personal ventilation devices and could evaluate them comparatively at the end of the experiment. the assessment was done by completing the fan assessment questionnaire (fanq). each device had the same period of usage (15 minutes) and their order was drawn randomly before each session. participants were allowed to freely activate the device, adjust airspeed and position the fan as they please during each usage period (indicated in grey in figure 1). an interval was established to create a gap between the use of each device, lowering the influence of one piece of equipment over the next. continuous use was also avoided as it could reduce the fan cooling effect (parkinson and de dear, 2016). to maintain the use demand, the experiment included variations on personal and environmental conditions during these intervals. the personal variation was based on increasing participants’ metabolic rate. this was achieved with walks through the building and a food break. the walk break consisted of participants walking a 5-minute path outside the experiment room, passing through the laboratory hallway, crossing the building corridor, going down two flights of stairs down to the lower floor, crossing the corridor again, going up two flights of stairs, passing the lab hallway and returning to the experiment room. building corridors were always naturally ventilated, while the lab hallway was not controlled by the researchers during the experiment; therefore, it could be either naturally ventilated or air-conditioned, depending on the day. the food break lasted 10 minutes, and participants were led to a naturally ventilated kitchen in front of the experiment room, where sweet foods, coffee and water were offered. environmental conditions variation was based on changing the conditioning mode. the air conditioning (ac) started completely off, simulating a dead band condition. figure 1. experiment procedure. table 1. tested devices specification. fan label and the main characteristic a) 3-speed ventilative b) 1-speed ventilative c) 2-speed ventilative d) 23-speed evaporative sales image number of speed levels 1 2 3 1 1 2 1 (5 track) 2 (12 track) 3 (23 track) air speed (m/s) 1.25 2.40 2.98 1.17 1.88 2.33 0.81 1.30 1.78 sound power level (dba)a 43.50 48.50 51.90 42.30 43.20 44.40 39.90 48.60 53.90 power (w)b 4.50 3.00 10.00 10.00 cost (usd)c $ 6.12 $ 8.45 $ 11.02 $ 367.33 dimension h x w x d (cm) 10 x 15 x 5 15 x 15 x 12 21 x 20 x 15 17 x 17 x 17 colour orange /green /black black blue & white white or black + 7 light colours rotation adjustment none horizontal vertical and horizontal none other works unplugged w/ rechargeable battery clamp-fixing option water tank for evaporative cooling a measured at 50 cm distance from the centre of the fan b indicated by the supplier c currency of 4.0835 brl to usd on 01/07/2020. reference: http://www.ipeadata.gov.br/ 15 min 15 min 5 15 min 10 15 min 5 15 min 15 min acclimatation fan 1 fan 2 fan 3 fan 4 walk walkcoffee & biscuits hvac off ac fan on ac cooling and fan on 26 °c set point infoq fanq personal information questionnaire (infoq) fans assessment questionnaire (fanq) users’ assessment of personal fans in a warm office space in brazil 65 after the use of the first selected fan – fan 1 in figure 1 – during the first break, the ac fan was turned on at airspeed level 2. and during the second break (after fan 2 usage), the cooling was turned on at the 26 °c setpoint temperature and the ac fan airspeed was reduced to level 1. these variations were not communicated to participants so their decisions on whether to turn on the devices were not influenced by their knowledge of system operation, but mainly by their thermal perception and demand. this strategy was intended to mimic an automatic operation where occupants are not aware of the system status. however, hvac status was informed when requested. it is noteworthy that the room was kept with all windows closed and open blinds before the beginning of the experiment to increase the indoor temperature. therefore, in the first 15 minutes of the experiment participants would have to acclimate to a warm condition. this first phase also aimed to level participants’ initial metabolic rate, so that any activity performed before the beginning of the experiment would be stabilised and would not influence either their thermal perception or demand. selected devices during the experiment sessions, each participant received one device at a time from a group of four devices. therefore, each participant evaluated all four devices shown in table 1. the devices were selected based on availability in the local market and their characteristics, to bring more variety to the experiment. their main differences are different levels of airspeed adjustment; vertical rotation adjustment in just two of them; slightly different sizes; and very different aesthetics. in addition, option d is an evaporative cooling fan, which recirculates air through an internal filter soaked in water. this option has also a much higher purchase cost than the other ones. participant selection to reduce the bias of age and gender, a heterogeneous group of participants was selected. forty people participated in the experiment and each session included two women and two men, from three age groups: 20-30, 31-50 and more than 50 years old. the ethical code in brazil requires that participation in research experiments be voluntary, so the sessions were arranged based on participants’ availability. data analysis the collected personal information (infoq) and assessment questionnaires (fanq) were processed in tables and matrices to analyze the results. also, statistical analysis was performed to verify whether final preferences were dependent on fan assessment selection order or the experiment room operation mode variation. the same analysis was applied to verify if a significant relation could be established when comparing device selection with participants’ anthropometric characteristics (weight, height and gender). to do so, fisher's exact test was conducted considering a confidence level of 95% (p<0.05). this statistical test is the most appropriate for small sample sizes (n=40) and categorical data analysis. the participant's weight and height collected in infoq were used to calculate body mass index (bmi) according to the nutrition ranges of the world health organization (who, no date). the measured environmental variables were also tabulated to analyze air temperature and relative humidity variation in each session. 3. results participants’ anthropometrics the participant selection was successful in building a heterogeneous group. as shown in figure 2, 19 out of 40 participants were women (w) and 21 were men (m). the number of younger people (20 to 30 years old) was a little higher (42%) than other age groups – 30% were 31-50 and 28% were more than 50 years old. regarding body mass index (bmi), most of the participants (62%) are considered to have a normal nutrition rate (who, no date). however, 28% fit into the pre-obesity category and the other 10% into obesity. thus, the sample does not include underweight people and bmi groups are not similar in the set. environmental variables the average indoor air temperature (tair) among all experiment sessions was 28 °c, reaching a maximum of 29.3 °c in one session and a minimum of 26.9 °c in another, while average relative humidity (rh) was 70% ranging from 57% to 82%. room thermal conditions varied throughout the experiment similarly in all sessions. figure 3 illustrates this variation in the second session of the experiment, showing the gradual drop in air temperature (tair) after cooling activation and variation according to occupancy. when occupants left the room, air temperature and relative humidity tended to drop due to the decrease of humidity produced by breathing and transpiration and the reduction of heat exchange between participants and room air. while the air temperature kept decreasing, the relative humidity (rh) increased again quickly when occupants returned to the room. the average variation of tair was 1.5 °c and rh was 15%. figure 2. the proportion of participants per gender (women in blue and men in red) and age groups in years. m. andré, c. buonocore, l. de castro and r. lamberts 66 the highest air temperature variation verified during the same session was 2.1 °c while rh reached 19% variation. air temperature and relative humidity registered in the lab hallway and the building corridor, to which participants were exposed during breaks, were always lower than in the experiment room. on average, the hall was 1.5 °c and 5% below the experiment room; and the corridor, 4.7 °c and 9% below the experiment room. thus, when people left the experiment room during breaks (walks and coffee breaks) they likely felt this difference and a cooling sensation that may have affected their perception of experimental conditions. it has been observed in studies on transient spaces that this change between spaces with different temperatures can generate a sense of relief (yu et al., 2016). and in this case, returning to the experiment room would generate the opposite effect, intensifying thermal discomfort by heat. willingness to use a personal fan in the first questionnaire applied (infoq), participants were asked if they have a fan, and most of them (78%) indicated they have it mainly at home; and 55% have a fan in their workspace. this could indicate that most would have a pre-disposition to use and may already like to use fans. however, the type of fan was not specified, so they could be used to either ceiling or figure 3. air temperature (tair) and relative humidity (rh) registered during the second experiment session. standalone fans instead of small personal devices. as shown in figure 4a, when initially asked (infoq) which operation mode they would usually use in a day with similar conditions to the experiment day, half of the participants indicated natural ventilation (nv) and the other half indicated air conditioning (ac). but fans would only be used with nv (nv+fan). however, figure 4b, shows that 2 of those participants who use ac (5% of total) and 1 of those who use nv+fan (2% of total) preferred to use nv alone. by the end of the experiment, as shown in figure 4c, some participants changed their opinion and most of them (62%) indicated they would prefer either to use a fan associated with natural ventilation (nv+fan) or air conditioning (ac+fan) in a day like an experiment day. the preference ratio between overall ac and nv did not change significantly from 4b to 4c. however, almost half of those who prefer ac seemed to like the idea of using it with a personal fan and most of those who prefer nv thought it would be better to associate it with a personal fan. in the last questionnaire (fanq), 72% of participants indicated they would like to have a personal fan in their workplace. this result indicates that participants probably had no experience using a personal fan in a conditioned room before the experiment and being exposed to the test settings made them consider this possibility. perhaps, turning on the fan while the air conditioning is on is counterintuitive. however, fans were accepted by part of participants as a thermal offset in a simulated situation, so the setpoint could be automatically extended to save energy. the impact on those who prefer natural ventilation was also noticeable. this result might indicate that a possible barrier to spread personal fans is the lack of experiences and opportunities to use them, as they are not usual in office buildings (liu et al., 2018). figure 4. preferred operation mode: natural ventilation (nv), natural ventilation with fan (nv+fan), air conditioning (ac), air conditioning with fan (ac+fan). 50 55 60 65 70 75 27.0 27.5 28.0 28.5 29.0 29.5 9: 30 9: 39 9: 48 9: 57 10 :0 6 10 :1 5 10 :2 4 10 :3 3 10 :4 2 10 :5 1 11 :0 0 11 :0 9 11 :1 8 11 :2 7 r el at iv e hu m id ity (% ) a ir te m pe ra tu re ( °c ) rh tair ac 50% nv 25% nv+fan 25% a) initial: which option would you use in your workplace? ac 45% nv 32% nv+fan 23% b) initial: which option would you prefer in your workplace? ac+fan 25% ac 22% nv 15% nv+fa n 38% c) final: which option would you prefer in your workplace? users’ assessment of personal fans in a warm office space in brazil 67 assessment of fans aspects participants were asked to indicate how important they would consider each of a set of criteria when buying a personal fan. results are shown in figure 5, in which the assessed criteria were grouped based on related aspects. participants rated the most important criteria related to thermal and acoustic aspects. the most important criterion for most participants (73%) was airflow sensation, followed by speed adjustment capability, considered the main aspect by 68%. the only acoustic criterion in this rank – noise produced by the fan – was considered the most important by 63% of people. while the third thermal criterion – vertical rotation adjustment capability – was rated as the most important by only 48% of participants. and the last thermal criterion, the possibility of reaching higher airspeeds (higher maximum speed) was evaluated as the most important by a small number of people (18%), and it seems to be either the second or fourth most important criterion for many participants. this means, adjusting and controlling airspeed is important, but not necessarily by increasing it. financial aspects such as fan energy consumption and cost were also rated as the most important criteria for 40% of people each, and the second most important by most of them, especially the cost issue. the next most important criteria were the functional and practical ones, such as size and usb charge connection availability. by last, aesthetic aspects were rated as the least important, and the possibility to choose the device colour (18%) seemed more important than general aesthetical issues (8%). it was expected that criteria related to controllability, like rotation and speed adjustment, would be highly rated by participants as control is one of the main functions of a personal conditioning device. however, results show people want to be able to control airspeed and direction, but the device must produce a pleasant sensation without noise and must be affordable as well. achieving a higher airspeed seems, on the other hand, to be secondary to the participants. device selection the last outcome of this experiment was to know which device participants would prefer, but before doing that the results were statistically analysed. fisher's statistical analysis indicated that the selection of devices was not significantly influenced by the order in which they were evaluated. the same occurred with the operation mode of the air conditioner, the variation of operation mode did not significantly influence the selection of the fans by participants. thus, it can be concluded that setting a random order of fan evaluation helped reduce the interference of other variables and the bias of device selection. this prevents, for example, that everyone preferred the first fan evaluated or the one that was in operation when the air conditioner was turned off. the fanq questionnaire asked for an initial overall preference, and after asking which device they would consider the best regarding some specific aspect, the overall preference was asked again, but at this time considering a purchase situation, first disregarding the cost, and then considering the cost information presented. as can be noted by the results shown in figure6a, the evaporative device – d – was evaluated as the best in most aspects by most participants. only in the criterion related to the produced noise, device b was considered the best for most of the participants (48%). option c was pointed out as the quietest for only two people (5%) even though options b and c show similar sound power measured levels (42 and 44 dba), as indicated in table 1. option d stands out mainly on the aesthetic criterion, in which 68% of people found it the best option, while only one person pointed out option c as the best in this matter. option c was chosen by more people than b mainly regarding the evaluation of which device provides the better thermal adjustment. the difference between b and c in this matter (8%) was expected to be even greater, considering option b has only one airspeed level while option c has two, and both allow vertical rotation. option d also stands out in this aspect, by having 23 airspeed levels (dial-like button), but no vertical adjustment capability. figure 6a shows the aspects ranking considering the importance of criteria analysed in section 3.4, where the most important is to provide a better sensation and the least important is aesthetics. figure 5. importance of each criterion for purchase decision – fanq. 73% 68% 48% 18% 63% 8% 18% 30% 28% 40% 40% 0% 20% 40% 60% 80% 100% level of importance of each criterion for purchase decision the most 2nd most 3rd most 4th most the least acoustic financialfunctionalaestheticthermal m. andré, c. buonocore, l. de castro and r. lamberts 68 hence, the results of purchase preference disregarding the cost shown in figure 6b is very consistent with the device ranking by aspect and the weight of aspects. by comparing initial and final preferences in figure 6b disregarding cost, slight changes are observed. the evaluation per aspect seems to have influenced participants’ perception and the main impact is the increase in votes to option b and the decrease in votes to option c. option d prevails as the preferred option in both initial and final questions disregarding the cost. however, when the cost was revealed – which is 40 times greater than the cost of option b, the cheapest device –, only one person indicated to be willing to purchase the evaporative device. as option b is cheaper than option c, the difference between them, which was 12% in the nocost question, rises to 18%, and option b is positioned as the preferred option by most participants. this result shows that as the difference in costs becomes greater, this aspect becomes an eliminatory criterion. it is noteworthy that option a was not indicated as the best or preferred option in any question by any participant. on the other hand, the initial preference indicated that one person did not prefer any option and, after the evaluation by aspect, this number grew to two people. in other words, two people would not buy any of these fans, considering the options unsatisfactory. by analysing the final choice of participants disregarding cost in face of their anthropometric characteristics, there was not enough evidence of a statistically significant association between purchase preference and categories of age, gender and bmi. despite that, the evaporative device – d – was preferred mainly by people with normal bmi, women and people aged 31 to 40 years or over 50 years old in this experiment (see figure 7). regarding options b and c, it is noted in figure 7 that men were equally divided in preference while women showed a higher preference for option b. regarding bmi, option b stands out for people of a nutritional level considered normal (who, no date). and more people over 50 years preferred option c while the youngest – between 20 and 30 years – indicated a preference for option b. in a way, as the distribution of age groups and bmi among participants is not equitable, it could be considered that the sample generated a trend that favours option b over option c. in this way, both options could be considered satisfactory and with good cost-benefit by participants. on the other hand, if option d had a more affordable price, it would probably be the preferred option for most people and could achieve a better overall evaluation if it produced less noise. figure 6. device ranking by aspect and overall preference – fanq. figure 7. device purchase preference disregarding cost by participants’ anthropometric characteristic. a) b) c) d) 48% 58% 43% 68% 25% 23% 5% 3% 23% 15% 48% 25% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% better sensation better thermal adjust. quieter aesthetics a) device ranking by aspect 50% 48% 3% 25% 18% 38% 23% 30% 55% initial preference purchase preference unknown cost purchase preference known cost b) overall device ranking a) b) c) d) none 47% 50% 45% 12% 17% 36% 41% 25% 18% 0 5 10 15 20 20-30 31-50 >50n um be r of re sp on se s a) purchase preference by age group (years) unknown cost 40% 58% 30% 11% 30% 32% 0 5 10 15 20 men women b) purchase preference by gender unknown cost 54% 36% 50% 13% 27% 50% 33% 36% 0 5 10 15 20 25 c) purchase preference by bmi unknown cost a) b) c) d) users’ assessment of personal fans in a warm office space in brazil 71 4. conclusion the results presented in this paper allowed us to assess users’ preferences regarding personal ventilative devices in a warm controlled working condition. the study highlights which aspects are considered most important for usability and device selection. as expected for a personal conditioning device, criteria related to controllability, like rotation and speed adjustment, were highly rated in importance by participants (the most important by 48% and 68% of participants, respectively). however, the noise produced by the fan was indicated as the third most important aspect (the most important by 63%). however, two devices with similar measured sound power levels – b with 42 dba and c with 44 dba – may be perceived differently, affecting their assessment regarding noise performance. option b was considered the quieter by 48% of people while option c was rated as the quieter by 5%. as the human sensory system receives information regarding multiple indoor environmental exposures simultaneously, sound effects should not be neglected when designing a personal device. contextual factors such as affordability were also pointed out as an important criterion (the most important by 40% of participants), which was confirmed by the change in the trend of selected devices when purchase cost was revealed. most participants preferred option d (48%) but only one (3%) was willing to pay for it. interestingly, the possibility of increasing airspeed was not deemed as one of the most important like other thermal-related criteria. it was considered the most important criterion by only 18% and the second most important by 33% of the subjects. the most important criterion for most people (73%) is the air movement sensation. therefore, a smooth airflow with good controllability – such as the airflow produced by the evaporative device labelled as d – would be an optimal choice for office occupants in brazilian offices in case this device becomes more affordable and quieter. option d was considered by 48% to produce the better sensation and by 58% to allow better adjustment, which are the most important criteria for most participants. moreover, the conducted experiment may have enlightened participants with the possibility of using both air conditioning and portable fans to achieve thermal comfort in their workspace during warm weather conditions. the cost was considered the second most important criterion for most participants (45%) and had a great impact on a final choice. so, apart from improving design aspects, innovative solutions should be affordable to become popular. also, if office occupants are stimulated to experience the use of desk fans and evaporative cooling devices with air conditioning, this operation mode might become usual. the results discussed in this paper can help researchers to choose better equipment for their studies and help designers and companies to identify ways to improve the characteristics of desk fans. from the perspective of implementation, the use of desk fans in shared office spaces usually occurs in two ways, one in which the occupants purchase their equipment and another in which they receive the fan from someone like their employer. the initial omission of cost in the fanq questionnaire intended to identify which aspects caught the participants' attention more regarding usability. therefore, the results could guide both employers and employees to select a better device by knowing which aspects would be compared to achieve a good usability performance. cost is a high impact factor, but cost-benefit will be different for an employer and an employee considering that employees do not have to deal with office space's energy cost. from the users’ standpoint, the most expensive device (d) was the best for most people, but it was not costbeneficial. they would rather buy a cheaper fan, so the difference among fans regarding usability was evaluated as lower than cost differences. however, the cost-benefit calculation for the employer is more complex as it should consider employee satisfaction, productivity, the purchase cost of multiple devices and the possible energy savings achieved by the extension of setpoint temperature. from this standpoint, if the most expensive device would significantly increase users’ satisfaction and their willingness to accept higher setpoint temperatures, it could be a costeffective option. however, further analyses would be needed to evaluate this long-term thermal comfort and energy-saving potential. participants’ thermal perception responses during the experiment were collected, so a future publication will present the results regarding whether the thermal acceptability was the same using each fan. the impact of environmental conditions on device activation, airspeed and position adjustment will also be addressed. another important issue that should be further investigated is the energy savings potential of using desk fans and extending setpoint temperature in the brazilian context. this analysis will be carried out by computer simulation considering different setpoint temperatures and locations. conflict of interest the authors declare that there are no conflicts of interest regarding this publication. acknowledgment the work reported in this paper was supported by the brazilian agencies coordination for the development of higher education personnel – brazil (capes), funding code number 001 – and the national council for scientific and technological development (cnpq). m. andré, c. buonocore, l. de castro and r. lamberts 70 references andré, m., de vecchi, r. and lamberts, r. (2020a) ‘feasibility of using personal fans for increasing thermal comfort in mixed-mode shared workspaces in brazil: a field study, in nceub (ed.) windsor conference: resilient comfort in a heating world. london, uk: nceub (network for comfort and energy use in buildings), p. 10. andré, m., de vecchi, r. and lamberts, r. (2020b) ‘user-centered environmental control: a review of current findings on personal conditioning systems and personal comfort models’, energy and buildings, 222. doi: 10.1016/j.enbuild.2020.110011. ashrae (2020) ‘standard 169-2020 climatic data for building design standards’, p. 258. boerstra, a. (2010) ‘personal control in future thermal comfort standards?’, in network for comfort and energy use in buildings (ed.) adapting to change: new thinking on comfort. windsor, inglaterra.: network for comfort and energy use in buildings, pp. 9–11. brager, g., zhang, h. and arens, e. (2015) ‘evolving opportunities for providing thermal comfort, building research and information, 43(3), pp. 1– 14. doi: 10.1080/09613218.2015.993536. de dear, r. (2011) ‘revisiting an old hypothesis of human thermal perception : alliesthesia’, building research and information, 39(2), pp. 108–117. doi: 10.1080/09613218.2011.552269. he, m. et al. (2017) ‘the influence of personally controlled desk fan on comfort and energy consumption in hot and humid environments’, building and environment, 123, pp. 378–389. doi: 10.1016/j.buildenv.2017.07.021. he, y. et al. (2017) ‘comfort, energy efficiency and adoption of personal cooling systems in warm environments: a field experimental study’, international journal of environmental research and public health, 14(11), p. 1408. doi: 10.3390/ijerph14111408. hoyt, t., arens, e. and zhang, h. (2015) ‘extending air temperature setpoints : simulated energy savings and design considerations for new and retro fi t buildings’, building and environment, 88, pp. 89–96. doi: 10.1016/j.buildenv.2014.09.010. knecht, k., bryan-kinns, n. and shoop, k. (2016) ‘usability and design of personal wearable and portable devices for thermal comfort in shared work environments’, in proceedings of the 30th international bcs human computer interaction conference, hci 2016, pp. 1–12. doi: 10.14236/ewic/hci2016.41. liu, s. et al. (2018) ‘coordinate control of air movement for optimal thermal comfort coordinate control of air movement for optimal thermal comfort’, 4731. doi: 10.1080/23744731.2018.1452508. luo, m. et al. (2018) ‘thermal comfort evaluated for combinations of energy-e ffi cient personal heating and cooling devices’, building and environment, 143(april), pp. 206–216. doi: 10.1016/j.buildenv.2018.07.008. mishra, a. k., loomans, m. g. l. c. and hensen, j. l. m. (2016) ‘thermal comfort of heterogeneous and dynamic indoor conditions — an overview’, building and environment, 109, pp. 82–100. doi: 10.1016/j.buildenv.2016.09.016. parkinson, t. and de dear, r. (2016) ‘thermal pleasure in built environments: spatial alliesthesia from air movement’. doi: 10.1080/09613218.2016.1140932. peel, m. c., finlayson, b. l. and mcmahon, t. a. (2007) ‘updated world map of the köppen-geiger climate classification’, hydrology and earth system sciences, 11(5), pp. 1633–1644. doi: 10.5194/hess-11-1633-2007. schiavon, s. et al. (2017) ‘thermal comfort, perceived air quality, and cognitive performance when personally controlled air movement is used by tropically acclimatized persons’, indoor air, 27(3), pp. 690–702. doi: 10.1111/ina.12352. schiavon, s. and melikov, a. k. (2009) ‘introduction of a cooling-fan efficiency index’, hvac and r research, 15(6), pp. 1121–1144. doi: 10.1080/10789669.2009.10390882. schweiker, m. et al. (2020) ‘review of multi‐domain approaches to indoor environmental perception and behaviour’, building and environment, 176(december 2019), p. 106804. doi: 10.1016/j.buildenv.2020.106804. simone, a. et al. (2014) ‘thermal comfort assessment of danish occupants exposed to warm environments and preferred local air movement’, indoor air 2014 13th international conference on indoor air quality and climate, pp. 411–418. available at: https://www.scopus.com/inward/record.uri?eid=2s2.084924678156&partnerid=40&md5=c267f839f02c 4ad091d122f8723cacf0. warthmann, a. et al. (2018) ‘personal climatization systems-a review on existing and upcoming concepts’, applied sciences (switzerland), 9(1). doi: 10.3390/app9010035. who, w. h. o. (no date) body mass index bmi. available at: http://www.euro.who.int/en/healthtopics/disease-prevention/nutrition/a-healthylifestyle/body-mass-index-bmi (accessed: 24 january 2019). xu, z. et al. (2017) ‘optimal coordination of air conditioning system and personal fans for building energy efficiency improvement’, energy and buildings, 141, pp. 308–320. doi: 10.1016/j.enbuild.2017.02.051. yu, z. (jerry) et al. (2016) ‘utility of cooling overshoot for energy efficient thermal comfort in temporarily occupied space’, building and environment, 109, pp. 199–207. doi: 10.1016/j.buildenv.2016.09.020. zhai, y. et al. (2013) ‘comfort under personally users’ assessment of personal fans in a warm office space in brazil 71 controlled air movement in warm and humid environments’, building and environment, 65, pp. 109–117. doi: 10.1016/j.buildenv.2013.03.022. zhai, y. et al. (2017) ‘selecting air speeds for cooling at sedentary and non-sedentary office activity levels’, building and environment, 122, pp. 247– 257. doi: 10.1016/j.buildenv.2017.06.027. zhang, h. et al. (2010) ‘thermal sensation and comfort models for non-uniform and transient environments : part i: local sensation of individual body parts’, building and environment, 45(2), pp. 380–388. doi: 10.1016/j.buildenv.2009.06.018. zhang, h., arens, e. and zhai, y. (2015) ‘a review of the corrective power of personal comfort systems in non-neutral ambient environments’, building and environment, 91, pp. 15–41. doi: 10.1016/j.buildenv.2015.03.013. م. أندريه و ك. بونوكور و ل. دي كاسترو* و ر. لامبرتس 1. introduction 2. method experiment procedure figure 1. experiment procedure. selected devices participant selection data analysis 3. results participants’ anthropometrics environmental variables willingness to use a personal fan standalone fans instead of small personal devices. as shown in figure 4a, when initially asked (infoq) which operation mode they would usually use in a day with similar conditions to the experiment day, half of the participants indicated natural venti... assessment of fans aspects device selection 4. conclusion conflict of interest acknowledgment references civl19603_ii.qxd 1. introduction aggregate chemistry plays a key role in asphalt-aggregate adhesion. it was found (curtis et al., 1993) that when cohesive asphalt failures do not occur, aggregate chemistry is much more influential than asphalt composition. active sites on the aggregate surface promote adsorption of polar asphaltic compounds. when these active sites are covered by non-polar compounds or dust that exist naturally on the aggregate surface, the bonding force that ______________________________________________ *corresponding author e-mail: hawahab@kfupm.edu.sa maintains the pavement is weakened. curtis et al. (1991) developed a limestone reactivity test that can determine the number of active sites present on the aggregate surface. after laying the pavement, asphalt-aggregate bonding forces can be weakened by the effects of water. water molecules intrude or diffuse to the aggregate surface and compete with the polar asphaltic compounds for interactions with the active sites. the affinity or compatibility of an asphalt-aggregate pair is very important for minimization of water induced damage. if the affinity is large, only a small percentage of the asphalt-aggregate interaction sites will be lost to water molecules. ring research 1 (2004) 29-38rnal of engineethe jou pavement stripping in saudi arabia: prediction and prevention h.i. al-abdul wahhab*1, i.m. asi1, s.a. ali2, s. al-swailmi3 and a. al-nour4 1 department of civil engineering, king fahd university of petroleum & minerals, dhahran 31261, saudi arabia. 2 department of chemistry, king fahd university of petroleum & minerals, dhahran 31261, saudi arabia. 3 operations & maintenance, riyadh municipality, riyadh, saudi arabia. 4 m&r department, ministry of communications, riyadh, saudi arabia. received 19 june 2002, accepted 11 october, 2003 abstract: pavement weathering or stripping is a major distress in highway networks in arid regions. using the saudi arabian road network as a case study area, seventeen road test sections were selected, out of which eight were stripped and nine were non-stripped. aggregates from quarries used to build these sections were also collected and subjected to detailed physical and chemical tests to evaluate the ability of these tests to distinguish between stripped and non-stripped sections. the modified lottman test was used to distinguish between compacted mixes. in addition, the swedish rolling bottle test, was also found to be effective in being able to distinguish between different asphalt-aggregates for stripping potential. eleven anti-stripping liquid additives, lime and cement, in addition to two polymers, were evaluated for their ability to reduce/eliminate stripping potential of stripping-prone aggregates. it was found that ee-2 polymer, portland cement, and their combination were effective with all aggregate sources. keywords: pavement stripping, roads, polymer, anti-stripping agents, lottman test :¢ü∏îà°ùÿÿg¥ô£dg äéµñ°t é¡d ¢vô©àj »àdg á«°ù«fôdg ôggƒ¶dg øe á«néæÿg ájô©àdg πegƒy öñ°ùh ∞°uôdg äé≤ñw øe ᪰ü◊g ∫é°üøfg á«∏ªy èà©j á©°ùj éªæ«h ᪰ü◊g ∫é°üøfg iôgéx øe êé©j ¥ôw á«féªk é¡æe áµ∏ªÿéh ≥jôw ô°ûyá©ñ°s qé«àng ” iôgé¶dg √òg á°sgq~d .á∏mé≤dg ≥wéæÿg ‘ ¢üfé°üÿg ~j~ëàd é«f髪«ch é«féjõ«a é¡∏«∏–h é¡à°sgqo â“h ¥ô£dg √òg é¡æe â«æh »àdg qoé°üÿg ™«ªl øe ωécôdg ™«ªœ ” .᪫∏°s èà©j é¡æe ‘ á«à∏ø°s’g áfé°sôÿg ≈∏y √é«ÿg á«°sé°ùm qéñàng agôlg ” é° jg .¥ô£dg øe ᪰ü◊g ôjé£jh ájô©àdg á«∏héb ioéjr ‘ qho é¡d ¿ƒµj ~b »àdg úh õ««ªàdg ‘ ¬«∏y oéªày’g øµá ¬fg ~lh ~bh iqgh~dg áj~jƒ°ùdg álélõdg qéñàng agôlg ” éªc .é¡æ«h õ««ªàdg á«féµeg ~j~ëàd ácƒe~ÿg äéæ«©dg äé£∏î∏d äéaé° ÿg øe áyƒª› ωg~îà°séh ájô©à∏d ᪰ü◊g áehé≤e ú°ù– á«féµeg á°sgqo â“ .ájô©à∏d πhéb ò¨dgh ájô©à∏d πhé≤dg ωécôdg äéæ«y ¿éªà°ùjg øfg~d ¿g ~lh .øfg~∏dg øe úyƒæd áaé°v’éh ,…oéy â檰sgh ,»ø£ÿg ò÷g ,»f髪«c öcôe 11 πª°ûj á«à∏ø°s’g(ee-2)â檰s’gh .ájô©àdgh ôjé£àdg ~°v ᪰ü◊g ´gƒfg πµd á«à∏ø°ùdg äé£∏ÿg áeƒ≤e ioéjr ‘ á«dé©a ogƒÿg ìcg âféc …oé©dg :á«méàøÿg äéª∏µdg √é«ÿg òkéj qéñàng ,á«f髪«c äéaé° e ,øfg~∏dg ,∞°uôdg ,᪰ü◊g ôjé£j pairs of low affinity will lose a large percentage of the asphalt-aggregate interaction sites to the more polar and hydrogen bonding water molecules. this leads to stripping. tests were devised (curtis et al., 1993) to determine this important affinity or compatibility for different pairs of asphalt-aggregates. if water is the cause of an asphaltaggregate problem, these tests will allow one to evaluate the possibility of future pavement stripping based on this affinity. when an aggregate absorbs water, the asphalt is “stripped” away. this ultimately leads to pavement failure. moisture degrades the integrity of an asphalt concrete matrix in three areas: loss of cohesive strength in the asphalt film, failure of the adhesive bond between the aggregate and asphalt (stripping), and loss of the chemical bond (integrity) between the asphalt film and the aggregate. other modes of pavement failure due to the presence of water are also possible. water can remove the soluble compounds from the asphalt causing it to fail. failure within the aggregate can also occur. water can promote phase separation within the asphalt, where the more polar molecules form a separate phase with water. a reduction in water induced damage can be achieved by selecting asphalt-aggregate pairs of high affinity, modifying the aggregate surface through silylation, or adding antistripping agents. building roads with low air voids and good drainage reduces water-induced damage by limiting the exposure to water. the ph of the medium can also affect the asphalt-aggregate bond. it was found (curtis et al., 1993) that a high ph (basic or alkaline environment) is detrimental to most asphalt-aggregate bonds. additives such as lime or some liquid anti-stripping agents can improve the performance of some asphalt-aggregate pairs in highly basic environments. among the many factors that contribute to the degradation of asphalt concrete pavements, moisture is a key element in the deterioration of the asphalt mix. since the 1930s, pavement engineers have been working to determine the moisture sensitivity of asphalt concrete mixtures. since that time, numerous tests have been developed to identify moisture-susceptible asphalt concrete mixtures. in general, there are two categories into which the water sensitivity tests can be divided. the first category includes tests which coat “standard” aggregates with asphalt cement. in these tests, the loose mixture is immersed in water, either at room temperature or at boiling temperature, and a visual evaluation is made of the separation of asphalt from the aggregate. the second category includes those tests that use compacted specimens, either laboratory compacted or cores from existing pavement structures (terrel and shute, 1989). these specimens are then water conditioned to simulate the in-service conditions of the pavement structure. the results of these tests are generally evaluated by the ratios of conditioned to unconditioned results using a stiffness or strength test, such as the diametral resilient modulus test. several methods have been developed to determine if an asphalt concrete mix is sensitive to water. the main methods can be summarized as follows (terrel and shute, 1989; curtis et al., 1991; terrel and al-swailmi, 1992; aashto, 1995): 1. nchrp 246 – indirect tensile test and/or modulus test with lottman conditioning. 2. nchrp 247 – indirect tensile test with tunnicliff and root conditioning. 3. aashto t-283 – combines feature of nchrp 246 and 247. 4. boiling water tests. 5. immersion-compression tests (aashto t-165, astm d 1075). 6. freeze-thaw pedestal test. 7. static immersion test (aashto t-182, astm d 1664). 8. conditioning with stability test (aashto t-245). 9. net absorption/desorption test (developed by shrp). 10. environmental conditioning system (ecs) (developed by shrp). pavement weathering or stripping is one of the major distresses in the highway network in the case study area, the kingdom of saudi arabia. pavement stripping severity varies from region to region in the kingdom. the highway network in al-qassim and hail regions is the most affected by pavement stripping. in certain roads where maintenance programs are not efficient, stripping develops potholes that severely affect road performance. the water sensitivity test used in local road departments is the typical water conditioning and evaluation by the marshall stability test. the conditioning phase includes partial saturation of specimens with asphalt and then soaking in a water bath. the specimens are then tested for marshall stability and compared with the results of unconditioned specimens. if the ratio (condition divided by unconditioned) is less than 75%, the mixture is considered sensitive to water. those roads experiencing stripping and weathering have all passed the water sensitivity test. this has indicated that the water sensitivity test using marshall stability is not reliable in determining the sensitivity of asphalt mixtures to water. with the recent developments in the design and evaluation as a result of strategic highway research program (shrp), the ministry of transport (mot) in the case study area has adopted a superpave mix design that utilizes a gyratory compactor. the northern part of the study area is mostly affected by this phenomenon due to the existence of water sensitive aggregates. the overall objective of this paper was to assess stripping problems in arid regions using the highway network of saudi arabia as a case study. secifically, current practices by road agencies were reviewed and the tests used by these agencies were included in the test program. binder aggregate adhesion and susceptibility of that adhesion to water damage were analyzed. comprehensive tests that predict the resistivity of asphalt-aggregate materials 30 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / (individually or as a mix) were adapted or modified. practical treatment methods were suggested to improve stripping resistance of local mixes. 2. methodology the work was carried out in three tasks and extended for 30 months. task 1: literature review: available literature from main research institutions locally and abroad related to the subject of the research were collected, summarized, and utilized to support the knowledge of the researchers in this field. task 2: stripping test selection and evaluation: different tests (physical and chemical) that might be used to detect susceptibility of pavement mixes and/or materials to stripping were evaluated (table 1). construction materials (fresh aggregates, slabs, and cores) from known performance road sections were collected, in consultation with government personnel. these materials were subjected to different tests to evaluate their ability in predicting stripping potential. a total of seventeen test sections were selected: eight stripped and nine unstripped. in selecting the study sections several criteria had to be met adequate drainage had to be present for the pavement surface. construction had to be according to specifications, (i.e. percent air voids (av%) had to be 31 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / serial # crusher road name pavement condition age symbol al-qassim 1 al-swailem ring road (north + east) excellent 5–6 years qn-1 2 burma ring road mid -east flange v. good 5–6 years qn-2 3 debiah 414 west road stat. 45+600 v. good 10 years qn-3 4 artic al-jamal avenue junction road medium stripping 3 years qs-4 5 al-fahd ring road -west flange stat. 14+550 high stripping 5–6 years qs-5 hail 6 al-swailem bagaà road low stripping 3 years hs-1 7 al-hudaires at humairah road (madinah junction) (rd -7771) medium to high stripping 5 years hs-2 8 al-namlah ring road stat. 17+000 v. good 14 years hn-3 eastern province 9 road construction establishment salasel -abqaiq, kp7 low stripping 7–8 years es-1 10 al-hazaà nuaireah -qaysoma road, after nuaireah intersection bridge excellent 6–7 years en-2 riyadh region 11 shibh al-jazira riyadh-dammam expressway stat. 980+000 v. good 8 years rn-1 12 al-awaidah riyadh-taif expressway stat. 511+00 medium stripping 5 years rs-2 taif 13 al-harameen taif-baha road, kp 1220 v. good 5-6 years tn-1 abha 14 ben-jarrallah prince salman sport city road v. good 7-8 years an-1 al-jouf 15 al-swailem sakaka domat al -jandal road v. good 5-6 years jn-1 16 al-harbi tabargel al -quriyat road high stripping 15 years js-2 northern region 17 al-sagaf arar-taif highway, km sign 1523 low stripping 7 years ns-1 table 1. selected test sections from case study area within design limits, compaction temperature within allowable limits, no overheating of asphalt, aggregate gradation within limits). aggregate quarry used for supplying aggregate in the asphaltic concrete layers had to be known and active so that fresh aggregate could be obtained similar to that used in the section. the collected materials were subjected to a number of chemical and physical evaluation tests to evaluate the ability of those tests to predict stripping. these tests included compacted mix evaluation methods such as environmental conditioning system (ecs), modified lottman test and marshall durability test.; asphalt/aggregate blend evaluation methods such as net adsorption in the presence of moisture and the swedish rolling bottle test and aggregate tests such as methylene blue value (mbv), soundness and physical properties. experimental design for the first phase of the project is shown in table 2. the response variables measured on individual materials, mixtures, compacted specimens, pavement cores were used for selection of applicable physical and/or chemical tests that are able to predict stripping and used for evaluation of the different mixes in the second phase of the study. task 3: material collection, mix designs and evaluation: in the second phase, different additives that are known from literature of being useful in preventing stripping were collected and used with the collected aggregate from stripped sections to prepare asphalt mixes using different percentages of additives and combinations. the marshall mix design procedure was used to determine the optimum asphalt content for each aggregate source as shown in table 1. different percentages of anti-stripping agents (as recommended by the manufacturer) were added to each mix. mixes were evaluated using the stripping tests. 3. results and discussion as an initial step in the statistical analysis, the normality assumption of the distribution of the test results was checked by drawing normal probability plots of the data. the statistical data analysis was then carried out in three stages. in stage i, the preliminary analysis, was to confirm that 32 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / number of replicates stripped sections non-stripped sections field samples laboratory samples field samples laboratory samples laboratory test/road section number 1 2 3 . 8 1 2 3 . 8 1 2 3 . 9 1 2 3 . 9 swedish ro lling bottle x x x . x 2 2 2 . 2 x x x . x 2 2 2 . 2 aggregate asphalt blend absorption -desorption x x x . x 2 2 2 . 2 x x x . x 2 2 2 . 2 volumetric properties 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 marshall compaction, vacuum saturation conditioning a nd resilient modulus testing 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 marshall compaction, vacuum saturation conditioning and split tensile testing 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 marshall compaction, marshall conditioning and split tensile testing 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 gyratory compaction, vacuum saturation conditioning and split tensile testing (7% air voids) 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 marshall stability loss 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 3 3 3 . 3 asphalt concrete mix environmental conditioning system (ecs) 2 2 2 . 2 2 2 2 . 2 2 2 2 . 2 2 2 2 . 2 table 2. experimental design x = test is not applicable there was a difference in behavior between stripped and non-stripped sections and that the grouping of the sections was correct. therefore, the statistical evaluation tests for this stage were performed on the test results of the extracted field cores. single factor analysis of variance (anova), using statistica statistical program, was carried out for the test results of the field cores to find if there was a significant difference between the means of the different performed tests of stripped and non-stripped sections. the null hypothesis for the anova test is (lapin, 1997): ho : m-stripped = mnon-stripped vs. where, ho = null hypothesis ha = alternate hypothesis µ = mean value of the test results for the specific test. 33 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / non-stripped sections stripped sections test p-value avg. s.d. avg. s.d. discriminant test limit probability of not being stripped if observation is less than set limit* % loss of resilient modulus 4.3e-05 38.13 10.01 60.84 23.57 45 75% % loss of split tensile strength + (vacuum saturation) 8.6e-07 41.49 7.65 63.63 18.34 48 80% % loss of split tensile strength (no vacuum saturation) 0.00017 32.25 10.03 53.31 24.29 38 73% marshall stability loss, % 9.4e-08 20.05 4.77 52.13 25.87 25 80% % loss of split tensile strength ++ (gyratory compaction) 4.6e-09 38.20 9.26 66.40 17.48 48 86% environmental conditioning system (ecs), after first loading cycle (%) 0.00082 16.67 6.99 32.48 15.80 22 78% environmental conditioning system (ecs), after sec ond loading cycle (%) 0.00455 30.63 10.77 45.94 16.49 37 72% environmental conditioning system (ecs), after third loading cycle (%) 0.01452 42.99 7.40 54.78 17.04 47 71% swedish rolling bottle value after 12 hrs. 0.02715 36.33 17.52 56.92 12.21 48 75% aggregate soundness test 0.0023 7.64 1.65 13.36 4.35 9 83% table 3. summary of analysis of variance results for significant tests * similarly, the probability of being stripped if observation is greater than set limit + conventional lottman test ++ modified lottman test (aashto t-283) anova analysis was performed for the test results of the extracted field cores as shown in table 2. results indicated that at α = 0.05 (the higher the α value, the lower the significance of the difference), for all the performed tests, except the ecs, the means of the stripped and nonstripped sections were significantly different. this implied that the tests were capable of differentiating between stripped and non-stripped sections and therefore indicated a good matching between the classification of the different sections into stripped and non-stripped sections and the test results. stage ii was to find which of the laboratory tests was capable of predicting the stripping potential of the asphalt concrete mixes. therefore, the single factor anova statistical evaluation was performed on the test results of the fresh aggregate and the laboratory prepared mixes, table 2. table 3 shows only tests that were significant and that had a p-value (probability to reject ho when ho is true) less than 0.05 in differentiating between mixes that were prone to stripping from sound mixes (lapin, 1997). the most significant test in predicting stripping (having the smallest p-value) was the loss in split tensile strength when performed on gyratory compacted samples that had 7 ± 1% air voids followed by vacuum saturation then soaking at 60°c for 24 hours (i.e. modified lottman test). the other tests in a decreasing order of significance were: marshall stability test; split tensile strength of marshall compacted specimens soaked @ 60°c for 24 hours after 34 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / s.n. product name physical state recommended dosage (wt%)* stability chemical name flash point 1 lilamin vp 75e liquid 0.2–0.4 heat stable mixture of alkyl and alkylene amines 120°c 2 wetfix ad-4f m.p. 63ºc fatty amine salt >150ºc 3 wetfix® be viscous liquid; b.p. >200ºc 0.2–0.5 heat stable (upto 170ºc) fatty acid + polyamine >100ºc 4 iterlene in/400-s liquid 0.2–0.4 heat stable (170ºc) alkylamido imidazo polyamine >180ºc 5 cecabase® 260 liquid 0.2–0.4 heat stable alkylamido imidazo polyamine >100ºc 6 polyram® l200 liquid heat stable n-alkyl’tallow’ dipropylene triamine >100ºc 7 ec9194a (exxon energy chemicals) liquid 0.2–0.4 heat stable (<250ºc) alkyl imidazoline in aromatic hydrocarbons 8 iterlene in/400 liquid 0.3–0.6 heat stable (upto 170°c) alkylamido imidazo polyamine >180°c 9 iterlene in/400-r liquid 0.2–0.4 heat stable (upto 170°c) fatty alkylamido imidazo polyamine >180°c 10 iterlene in/400-r-1 liquid 0.2–0.4 heat stable (upto 170°c) fatty alkylamido imidazo polyamine >180°c 11 morelife 3300 viscous liquid 0.2–0.5 heat stable (upto 150°c) polycyclo aliphatic polyamines 170°c 12 polybilt granules 2.0–5.0 heat stable >200°c >200°c 13 ee-2 polymer granules 2.0–5.0 heat stable >200°c modified olefin >200°c 14 cement powder 2.0–4.0 − portland cement − 15 lime powder 2.0–4.0 − hydrated lime (calcium hydroxide) − table 4. collected liquid antistripping agents *weight (5) of the antistripping agent added to the bitumen vacuum saturation; resilient modulus of marshall compacted specimens soaked @ 60°c for 24 hours after vacuum saturation; split tensile strength of marshall compacted specimens soaked @ 60°c for 24 hours without vacuum saturation; resilient modulus of gyratory prepared samples having a target air void between 6 and 8% after one cycle in ecs; aggregate soundness test; resilient modulus of gyratory prepared samples having a target air void between 6 and 8% after two cycles in ecs; loss of resilient modulus of gyratory prepared samples having a target air void between 6 and 8% after three cycles in ecs; and swedish rolling bottle value after 12 hrs. rolling. although all tests were statistically significant, it can be seen that they can be divided into three groups: com35 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / antistripping agent code hail al-jouf eastern province riyadh hail (hs1) (js2) (es1) (rs2) (hs2) % loss % loss % loss % loss % loss lilamin vp 75e 36.45% 61.47% 100.0%* 100.0% 48.00% wetfix ad -4f 71.15% 100.00% 38.4% 100.0% 40.30% wetfix® be 32.98% 69.10% 100.0% 100.0% 44.90% iterlene in/400-s 58.40% 79.20% 100.0% 100.0% 51.20% cecabase® 260 40.54% 55.80% 100.0% 47.6% 48.00% polyram® l200 44.30% 72.90% 100.0% 100.0% 53.00% ec9194a 29.24% 64.60% 100.0% 100.0% 41.20% iterlene in/400 45.80% 76.80% 100.0% 100.0% 53.70% iterlene in/400-r 32.59% 57.40% 100.0% 100.0% 44.50% iterlene in/400-r-1 52.80% 86.40% 100.0% 69.0% 32.40% morelife 3300 21.54% 67.80% 100.0% 31.7% 36.60% polybilt 70.46% 80.80% 100.0% 100.0% 39.21% ee-2 polymer 28.75% 32.30% 35.8% 41.9% 42.40% cement 35.85% 40.80% 38.5% 46.9% 46.00% lime 41.80% 49.20% 75.5% 74.9% 48.30% control 49.40% 77.40% 100.0% 100.0% 58.30% table 5. effect of different antistripping agents using modified lottman test % loss of 100 means that conditioned samples have collapsed, resulting in its of zero and 100% its pacted mix, asphalt/aggregate blend and aggregate. from the first group, it can be seen from table 3 that both the modified lottman test (aashto t-283) and the marshall stability loss test (mot-mrdtm 410) had the lowest discriminant probability p-value of 4.6e-09 as compared to a p-value of 9.4e-08 for marshall stability loss. moreover, the government in the case study area is in the process of adopting the superpave mix design. this will eliminate the use of the marshall mix design. lottman, on the other hand, is a simple test that has proven effectiveness and is currently used widely in the united states. the lottman test was therefore selected as the best mix evaluation test to discriminate between compacted mixes. the swedish rolling bottle test was the only one in the second group 36 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / antistripping ave. initial its. ave. final its. average agent code sample* id initial its kg/cm2 final its kg/cm2 kg/cm2 kg/cm2 % loss 12.1 8 11.9 7.9 cecabase® 260 + cement rs2 12.3 8.3 12.1 8.1 33.33% 12.9 9.2 11.2 8.3 morelife 3300 + cement rs2 11.6 9 11.9 8.8 25.77% 11.9 7.6 12.1 7.7 ee-2 polymer + cement rs2 12.4 8.1 12.1 7.8 35.71% 13.7 9.6 12.9 9.3 wetfix ad4f + cement es1 13.5 9.3 13.4 9.4 29.68% 13.9 9.9 13.5 10 ee-2 polymer + cement es1 13.2 9.7 13.5 9.9 27.09% 12.7 7.7 13.4 8 cecabase® 260 + cement js2 13.1 7.6 13.1 7.8 40.56% 12.2 8.8 13.1 9.4 ee-2 polymer + cement js2 12.6 9.1 12.6 9.1 27.97% 12.1 7.6 13 8.2 morelife 3300 + cement hs1 11.6 8.1 12.2 8.0 34.88% 15.5 11.4 15.8 11.1 ee-2 polymer + cement hs1 15.1 10.8 15.5 11.1 28.23% 11.4 7.3 10.9 7.1 morelife 3300 + cement hs2 11.8 8.1 11.4 7.5 34.02% 14.7 9.6 14.2 9.4 ee-2 polymer + cement hs2 15.1 8.9 14.7 9.3 36.59% table 6. effect of combined antistripping agents using modified lottman test * rs: riyadh aggregate js: al-jouf aggregate es: eastern provience aggregate hs: hail aggregate that was significant (p=.02715) in evaluating loose aggregate-asphalt blends stripping tendencies. in the third group, aggregate soundness, which was a significant test (p=0.0023), had the ability to distinguish between aggregates prone to stripping. this test, however, did not have the capability to evaluate additives to asphalt mixes and therefore was eliminated. stage iii was to find the test limits that could be set to screen mixes that were prone to stripping. this was based on the test results of the fresh aggregate and the laboratory prepared mixes. specification limits were calculated for each of the significant tests. these limits could then be used to test if the mix was prone to stripping (i.e. if the result of the evaluation test was higher than the set limit, then the mix would be prone to stripping). assuming a normal distribution of the results, the average and standard deviation values for the samples table 3, were used to find the test limits. the aim was to find the limit that would produce an equal probability of classifying the mix as not being prone to stripping if the test value was lower than the limit, to the probability of classifying the mix as being prone to stripping if the test value was higher than the limit. for example, table 3 shows that for the modified lottman test (aashto t-283) and the marshall stability loss test (mot-mrdtm 410), the discriminating limits were 48% and 25%, respectively. in addition to lime and cement dust, several liquid antistripping agents were purchased and specifications of the products as provided by the companies were also obtained, table 4. the marshall mix design was used to arrive at the optimum asphalt content for the four selected stripped locations. eleven liquid antistripping agents were collected from the original manufacturers and administered at the maximum recommended percentage. moreover, cement, lime and two polymers (polybilt 101 and eastman ee-2) were used at a dosage of 4% as recommended by government guidelines and polymer manufacturers. work was carried out in two stages. the maximum recommended dosages of the antistripping additives were used to quantify the effect of the additive on stripping phenomena and to screen the additives, table 5. in the second stage , additive combinations of promising mixes were evaluated. combinations included dry additives (portland cement and/or lime) and liquid additives or polymers. liquid additives were not mixed with polymers. this was done to avoid adverse chemical reactions, table 6. the modified lottman test and the swedish rolling bottle test were used to evaluate the effectiveness of the different treatments. table 5 shows the typical stage i test results for the hail, al-jouf, eastern province and riyadh region. the results show the loss in the indirect tensile strength values (its), according to the modified lottman test procedure, for all treatment combinations. the effectiveness of the treatments was evaluated based on the level of its improvement for each aggregate as compared to the control mix (no additive). it should be noted that 100% its loss indicated that samples failed during the conditioning phase ,indicating severe stripping. this resulted in a final its of zero and therefore a 100 % its loss. table 6 shows the results of the combined additives. in general, it can be seen that the its of each mix was dependent on the aggregate type. the results of the second phase indicated that for hail aggregate, treatments ec9194a, iterlene in/400, iterlene in/400-r, polyram l200, cecabase 260, wetfix be, lilamin vp 75e, morelife 3300, ee-2 polymer, cement, and lime were effective in eliminating the stripping potential of the aggregate. for al-jouf aggregate, only ee-2 polymer and cement were effective in eliminating the stripping potential of the aggregate. for eastern province aggregate, only wetfix ad-4f, ee-2 polymer, and cement were the effective additives. with the riyadh aggregate, only morelife 3300 and ee-2 polymer were effective additives. the ee-2 polymer was effective in eliminating the stripping potential of all aggregates from all sources. eastman ee-2 polymer and portland cement and their combination proved to be effective with all studied aggregate sources. morelife 3300 antistripping additive combined with cement was the most effective with riyadh and hail aggregates. wetfix ad-4f combined with cement was the most effective with the eastern province aggregate. finally cecabase 260 combined with cement was effective with the al-jouf aggregate. for each type of aggregate, there were specific additives that were effective in eliminating the stripping potential. however, cement and ee-2 polymer combinations were effective in eliminating or effectively reducing the stripping potential of all the tested aggregates. 4. conclusions based on the analysis of variance (anova) results, the modified lottman test, the marshall stability loss test, environmental conditioning system (ecs), and the resilient modulus loss were effective in distinguishing between stripped and non-stripped mixes. ecs had the lowest significance (p = .00082) among these tests while the modified lottman had the highest significance (p = 4.6e–09). the swedish rolling bottle was found to be effective in screening asphalt-aggregate materials for stripping potential. eastman ee-2 polymer and portland cement and their combination proved to be effective with all studied aggregate sources. morelife 3300 anti-stripping additive combined with cement was the most effective with riyadh and hail aggregates. wetfix ad-4f combined with cement was the most effective with the eastern province aggregate. cecabase 260 combined with cement was effective with al-jouf aggregate. acknowledgements the authors would like to acknowledge the support 37 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / of kacst for funded this project, kfupm for provided the research facility, and mot for provided the logistical support in selecting and sampling test sections. references american association of state highway and transportation officials (aashto), 2000, “resistance of computed bituminous mixtures to moisture induced damage,” aashto t-283-89, washington, d.c., usa, pp. 983-991. curtis, c.w., terrel, r.l., perry, l.m., al-swailmi, s., and brannan, c.j., 1991, “importance of asphaltaggregate interactions in adhesion,” proceedings, association of asphalt paving technologists (aapt), vol. 60, p. 476. curtis, c.w., ensley, k., and epps, j., 1993, “fundamental properties of asphalt-aggregate interactions including adhesion and absorption,” strategic highway research program publication, shrp-a-341, washington, d.c., usa. lapin, l.l., 1997, modern engineering statistics, wadswoth publishing company, belmont, ca. terrel, r.l., and shute, j.w., 1989, “summary report on water sensitivity,” strategic highway research program, shrp-a/ir-89-003, washington, d.c., usa. terrel, r.l., and al-swailmi, s., 1992, “final report on water sensitivity of asphalt-aggregate mixtures test development,” strategic highway research program, national research council, washington, d.c., usa. 38 ring research 1 (2004) 29-38rnal of engineethe jouabdul wahhab et al. / corresponding author's e-mail: dr.qahmed50@gmail.com doi:10.53540/tjer.vol19iss2pp152-162 the journal of engineering research (tjer), vol. 19, no. 2, (2022) 152-162 0043 a structured reliability and maintainability assessment model: an application to high voltage motors majdi m. obaid1 and qadeer ahmed2, * 1shedgum gas plant, saudi aramco, shedgum, saudi arabia 2consulting services department, saudi aramco, dhahran, saudi arabia abstract: motors are one of the vital equipment and generally the higher in numbers in oil and gas processing facilities. the primary function is to drive the process equipment such as compressors, fans, pumps etc. unreliability of the motors is a threat to safety but also to production loss and high operating expenditure. motors experience higher failure rates and maintenance costs with age due to lower focus during useful life periods. in order to properly address the long-term reliability and maintainability of the motors and associated subsystems, this paper aims to propose a structured methodology and set of tools to ensure effective assessment. the proposed model mainly consists of data collection, analysis, assessment, financial analysis and later developed actions to properly address the concerns. equipment failure and repair data is a challenge to any reliability assessment; hence, proposed methodology was introduced to collect, verify and validate the data. later, multiple tools such as pareto analysis, failure mode and effect analysis and root cause analysis were used to perform a detailed assessment. weibull analysis was also explored to understand the failure modes, which ultimately helped in improving the availability of the motors. the proposed methodology has been applied to high-voltage motors to observe the effectiveness of the tools and proposed model in addressing reliability and maintainability. the results show significant reliability improvements of 12% (from 58% to 70%) and prove that the structured method can be effectively used in complex process facilities with significant benefits. keywords: reliability; maintainability; data analysis; failure modes & effect analysis; weibull analysis; root cause analysis . هدنموذج تقييم هيكلي للموثوقية والقدرة على االستدامة: تطبيق على المحركات عالية الج مد وقدير أح مجدي محمد العبيد األساسية تهاوتتمثل وظيف ، معالجة النفط والغاز بشكل عام في منشآت تتزايد أعدادها تعد المحركات أحد المعدات الحيوية :الملخص ، للسالمةالمحركات يمثل تهديدًا موثوقيةمثل الضواغط والمراوح والمضخات وما إلى ذلك. إن عدم المعالجة في قيادة معدات أعلى مع تقدم العمر ولكن أيًضا لخسارة اإلنتاج ونفقات التشغيل المرتفعة. تواجه المحركات معدالت أعطال أعلى وتكاليف صيانة بسبب انخفاض التركيز خالل فترات العمر اإلنتاجي. من أجل معالجة الموثوقية طويلة المدى وقابلية الصيانة للمحركات واألنظمة تهدف هذه الورقة إلى اقتراح منهجية منظمة ومجموعة من األدوات لضمان التقييم الفعال. يتكون النموذج المقترح بشكل ،الفرعية سي من جمع البيانات والتحليل والتقييم والتحليل المالي واإلجراءات التي تم تطويرها الحقًا لمعالجة المخاوف بشكل صحيح. أسا تقديم المنهجية المقترحة لجمع البيانات والتحقق منها والتحقق لذلك تم موثوقية،يمثل فشل المعدات وبيانات اإلصالح تحديًا ألي تقييم استخدام أدوات متعددة مثل تحليل باريتو ووضع الفشل وتحليل التأثير وتحليل السبب الجذري إلجراء تقييم ذلك بعد من صحتها. تم تحليل استكشاف تم أوضاع ويبل مفصل. لفهم تطبيق الفشل،أيًضا تم المحركات. توافر تحسين في النهاية في ساعدت والتي راقبة فعالية األدوات والنموذج المقترح في معالجة الموثوقية وقابلية الصيانة. المنهجية المقترحة على المحركات عالية الجهد لم ٪( وتثبت أن الطريقة المنظمة يمكن استخدامها بشكل 70٪ إلى 58٪ )من 12تظهر النتائج تحسينات كبيرة في الموثوقية بنسبة .فعال في مرافق العمليات المعقدة مع فوائد كبيرة لجذري. الفشل وتحليل التأثير؛ تحليل ويبل؛ تحليل السبب ا أوضاع ؛ قابلية الصيانة؛ تحليل البيانات؛ الموثوقية الكلمات المفتاحية: a structured reliability and maintainability assessment model: an application to high voltage motors 153 nomenclature ac alternating current cbm condition-based maintenance cou cost of unreliability dc direct current fmea failure mode and effect analysis fta fault tree analysis hazop hazard and operability hvm high voltage motors mtbf mean time between failures mttr mean time to repair pof probability of failure raam reliability assessment and analysis model rca root cause analysis sem submersible electric motor 1. introduction reliable operations of high voltage motors (hvm) are vital to ensure the overall availability and safe operation of the facility. due to its criticality, it has become an ever-increasingly main topic and equipment in the organizational continuous improvement toolbox. higher plant reliability helps to reduce operating costs as asset failure causes a reduction in production output and high maintenance cost, which in turn limits profitability. additionally, equipment failures increase the probability of an environmental incident and the potential for safety-related accidents. the motor is one of the highest numbers of equipment in any oil and gas facility. based on the design requirements, a proper motor, such as an induction or synchronous motor, is selected to perform its function. in the oil and gas industry, motors are generally used as drivers to many driven equipment such as pumps, compressors, heat exchangers, coolers and other rotating equipment, as shown in figure 1. gas plants are one of the key facilities in the oil and gas industry, and the high availability of these facilities is essential to both revenue generation and safe operations. the purpose of the gas plant is to recover the natural gas liquid (ngl) from a gas such as ethane, butane, propane, and gasoline from gas. the gas plant is equipped with many rotating equipment, such as compressors, turbines, pumps, and gearboxes. one of the critical rotating equipment systems is propane refrigeration. it is a closed-loop system and is critical to gas plants. the refrigeration system has compressors driven by motors. in this system, two three-stage propane compressors driven by a 21,000 hp motor in parallel are fed from three suction scrubbers. these high-voltage motors are used as a case study in this paper. to ensure the high availability of the high voltage motors, a structured methodology is developed and utilized to perform the assessment. this assessment is based on the maintenance downtime and time to failure data and associated maintenance cost for a set of motors. the details are discussed in the later part of the paper. figure 1. general motor appli cations and s etup . the remaining paper is structured as follows: section (2) outlines the literature review to understand the methodologies researcher have used to improve the reliability and availability of high-voltage motors. section (3) presents the methodology to perform this analysis, the analysis and the tools used to address the poor reliability of the motors. section (4) presents the assessment recommendation, and finally, conclusions are discussed in section (5). 2. literature review due to its importance and its wide applications in industry, motor reliability is a topic of interest for both academia and industry. motors are commonly used as a driver of many assets, such as pumps, compressors and fans. being vital equipment in oil and gas processing facilities, the unreliability of the motors is a very critical and significant threat to safety, production loss and operating expenditure. due to its importance, many companies have developed tools, technologies and develop applications to assess motor conditions. proper maintenance is important to ensure the reliability and availability of motors. maintenance of these large motors is proposed using a three-pronged approach edwin and syam (2010). in this work, the motor is considered as a system and split into four subsystems individually, and later the results are combined with observing the overall outcome. later, failure mode and effect analysis (fmea), hazard and operability (hazop) and fault tree analysis (fta) are simultaneously applied to these subsystems to identify their failure modes. in the end, using a condition indicator for each of these failure modes, failures are mitigated by using the warning signals received while monitoring the condition parameters. a statistical analysis of the operation of motors and the current measurement of certain parameters of its operation to online detect the condition of electric motors is developed by zaiets and kondratenko (2019). a genetic algorithm is used in the optimization of the number of incoming parameters with various combinations. a neural network is also introduced for predicting the reliability of an electric motor which allows real-time detection of defects and monitoring of majdi m. obaid and qadeer ahmed 154 the condition of machines. a reliability metric is introduced based on statistical laws superposition of mean time between failures (mtbf) of submersible electric motors against surges ilya et al. (2016). the outcome and results are presented in the form of an electric motor survival probability dynamic pattern as a function of resource consumption to protect submersible electric motors against surges. operational and maintenance cost reduction of induction motors (ims) is discussed in choudhary et al. (2019). costs can be better managed and reduced when the equipment is monitored regularly. monitoring helps early detection of the deterioration of motor health, improvement in proactive response, and minimizing unscheduled downtime. condition-based maintenance (cbm) is an effective method to maintain assets. a cbm strategy for continuously monitored degrading systems with multiple failure modes is presented by liu, xiao, et al. (2013). unlike existing cbm models, this article considers multiple sudden failures that can occur during a system's degradation, and the failure rate corresponding to each failure mode is influenced by either the age of the system. a statistically dependent model is developed considering time-to-maintenance due to system degradation and time-to-failure of different failure modes. an optimum maintenance threshold level is achieved that maximizes the system's availability over its life cycle. an analytical hierarchical process (ahp) accompanied by fuzzy sets theory to determine the most critical component types of distribution power systems to be prioritized in maintenance scheduling is suggested by dehghanian et al. (2011). in the presence of many qualitative and quantitative attributes, fuzzy sets can effectively help to deal with the existent uncertainty and judgment vagueness. a modular approach for a reliability centered asset management process for an electric network with cost and quality is presented by schwan et al. (2007). the goal of the application is the combined control of cost-effectiveness and power quality. application examples of this asset management process for distribution systems present the definition and calculation of both component importance and component condition indices, as well as a detailed simulation of both supply reliability and economic parameters into the future. al-douri et al. (2020) introduced a systematic methodology that accounts for failure early enough during the conceptual design stages. once a base-case design is developed, the methodology starts by identifying the sources of failure that are caused by reliability issues, including equipment, operational procedures, and human errors for a given process system or subsystem. this helps identification of critical process subsystem(s) that are more failure-prone or cause greater downtime than other subsystems. bayesian and monte carlo techniques are utilized to determine the appropriate distributions for the failure and repair scenario(s), respectively. markov analysis is used to determine the system availability. the economic potential of alternative design scenarios is evaluated with the objective of maximizing incremental return on investment (iroi) is utilized to make a design decision. ferreira et al. (2015) presented the key considerations on the reliability and operation of highefficiency motors, offering a comprehensive perspective on the advantages, drawbacks and limitations of high-efficiency industrial motors. a comparison of the life expectancy of ie2-, ie3and ie4class motors under unbalanced and distorted voltage supply is considered. it also demonstrated how moving to a higher efficiency class can avoid investing in a larger motor. hacke et al. (2018) describe the projects and relevant background needed in developing design qualification standards that would serve to establish a minimum level of reliability, along with a review of photovoltaic inverter quality and safety standards, most of which are in their infancy. in this paper, authors have compared stresses and levels for accelerated testing of inverters proposed in the standard drafts and those proposed by manufacturers and purchasers of inverters. design validation testing using realistic operation, environmental, and connection conditions, including under end-use field conditions with feedback for continuous improvement, is recommended for inclusion within a quality standard. 3. reliability assessment and analysis model (raam) a structured reliability assessment and analysis model is highly desired to handle such assessment to address this problem properly. we have introduced a reliability analysis model shown in figure 2. this model can be used for different systems in assetintensive industries to perform reliability assessments. the objective of the assessment is to utilize the existing model to structurally apply to address reliability. this assessment is to analyze the common failures experienced by the subject motors and utilize the data to reduce the average maintenance downtime of the propane compressor motors by enhancing the motor coupling maintenance strategy, such as the rubber blocks replacement and upgrade coupling type. out of the population of motors in operation, a set of eight motors are identified which have experienced failures during the last five years. the pareto analysis was carried out on this set of motors to further narrow down the group of motor failure modes with the worst performance. the primary target for further investigation and weibull reliability analysis is this group of motors. a structured reliability and maintainability assessment model: an application to high voltage motors 155 figure 2. reliability assessment & analysis model. majdi m. obaid and qadeer ahmed 156 table 1. basic data – synchronous motors. specification. value horsepower 21,000 hp power 3450 kw voltage 13,800 v current 184 a frequency 60 hz table 2. average motor operating and downtime data. hp motors average time (hours) downtime (hours) mtbf mttr m-a 33695.50 1368.50 8423.88 342.13 m-b 34438.20 625.80 8609.55 156.45 m-c 33580.90 1483.10 8395.23 370.78 m-d 34800.16 263.84 8700.04 65.96 m-e 33639.40 1424.60 8409.85 356.15 m-f 34121.90 942.10 8530.48 235.53 m-g 34600.07 463.93 8650.02 115.98 m-h 33657.90 1406.10 8414.48 351.53 figure 3. treemap – mean time to repair of motors. figure 4. difference between existing and target mtbf. the analysis started by determining each electric motor failure mode. the highest number of failures are attributed to the failure of motor couplings followed by downtime due to overhaul, rotor, bearing, etc. coupling rubber blocks are essentially non-repairable products and are being replaced upon failure now. most motor failures are found to be due to the coupling of rubber blocks. some of the basic data for the motors are provided in table 1. the motor failure data has been gathered, and the results have shown that for the past four years, the average annual maintenance unavailability for the equipment is around 10.4 days of downtime per year per motor, as shown in table 2. treemap shown in figure 3 exhibits the average repair time of the motor, whereas, figure 4 shows the current mtbf vs target mtbf. 3.1 asset identification the first step is to review the overall reliability and maintenance data to identify the bad actors or assets that require attention. generally, at facilities, there are monthly or weekly data analyses to review and identify such assets with poor reliability. one of the very common techniques to perform such activity is pareto analysis. pareto analysis provides us with an opportunity to identify the few items causing major issues, commonly known as the 80-20 rule. 80% of the problems can be addressed by addressing 20% of the causes. shweta et al. (2018) and taman & tanya (2015). 3.2 pareto analysis identify issues / problems the pareto analysis method is used to define a given problem by viewing the common failures experienced on the motor to determine which failure mode is affecting the machines. it is a decision-making analysis method that separates a statistically limited number of input factors into the largest effect on a desired or undesired result. pareto analysis is based on the idea that 80% of a benefit can be achieved by carrying out 20% of the work. the time of failure, type of failure, and repair cost of each motor is housed in the maintenance management system. the data period of our investigation, from january 2004 to january 2007, is limited to four years. eight motors are those whose four-year history has been reviewed. the common failures experienced by any rotating equipment can be broken down into four major categories: vibration, electrical, instrument, and manufacturing error. the data was collected for all eight motors to determine which failure mode is most common and how that failure contributed to the system downtime. most of the failures experienced were with 70.9% of high vibration and 26.6% of an electrical fault, as can be witnessed in figure 5. the high vibration downtime contributed to a total of 2215 hours; vibration issues with the motors contribute 70.9% of the overall failure modes. a detailed assessment of data vibration failure symptoms, the two most common failures within that realm are coupling and rotor. the coupling failures make up around 73% of the total vibration failures experienced on the eight motors during the analysis, as shown in figure 6. m-a m-b m-c m-d m-e m-f m-g m-h mtbf target a structured reliability and maintainability assessment model: an application to high voltage motors 157 c2 2215 829 78 percent 70.9 26.6 2.5 cum % 70.9 97.5 100.0 c1 otherelecvib 3500 3000 2500 2000 1500 1000 500 0 100 80 60 40 20 0 d o w n t im e p e r c e n t pareto chart of failure mode figure 5. pareto chart of motor failure modes. figure 6. pareto chart of vibration failures. figure 7. basic coupling sub-components. table 3. basic data – coupling failures. no of couplings time to failure (hours) 1 30,000 – 35,000 2 35,000 – 40,000 5 40,000 – 45,000 3 45,000 – 50,000 6 50,000 – 55,000 8 55,000 – 60,000 7 60,000 – 65,000 table 4. 5 why – motor failure (vibration). why cause why did the motor fail? vibration why does vibration occur? coupling misalignment why does coupling misalignment occur? failures of blocks caused by misalignment (70%) why did coupling rubber blocks fail? rubber blocks service lifetime. 3.3 root cause analysis (rca) root cause analysis is a structured method of exploring the root cause of failures. there are many processes to conduct root causes, such as 5-why, fta and proact. each method has a different way of getting to the root cause, which is not part of the discussion. an rca was carried out to understand the causes of the failure of the motor system. motors are equipped with a non-lubricated flexible coupling designed to transmit torque via rubber blocks, as shown in figure 7. as it is used to absorb intense shocks during startup and any abrupt change in load, rubber blocks are an essential part of the coupling design. due to the continuous operation of the motors, the material and type of rubber blocks play a major role in their durability. maintenance data analysis has shown that the coupling rubber blocks for all the motors have been replaced for the first time, as per table 3, following the failure running hours. in this assessment, motor failures were analyzed by some of the common root cause analysis (rca) techniques as follows: 5-why is a method of questioning used to determine the cause and effects of a particular problem or failure? the first why usually starts with a question such as why did the motor fail? the answer forms the foundation of the next question until you get to something fundamental, as shown in table 4. the number five is based on the observation that the general problem is typically resolved by asking 'why' four to five times. the five why the technique is also commonly used during a vibration-related failure symptom as a troubleshooting guide to determine the potential root cause. the impact of coupling misalignment due to defective rubber blocks on vibrations is debated in the field. there are different explanations for why misalignment causes the 2x (2-times speed) vibration frequency problem. fig. 8 shows the cracked rubber blocks, and they have contributed to the problematic frequency. majdi m. obaid and qadeer ahmed 158 figure 8. failure and cracks of rubber blocks. table 5. weibull analysis – dataset. time to failure (hours) cumulative per cent f(t) of couplings failed (tt0) t0 = 20,000h (tt0) t0 = 25,000h (tt0) t0 = 30,000h 30,00035,000 1 15,000 10,000 5,000 35,00040,000 3 20,000 15,000 10,000 40,00045,000 8 25,000 20,000 15,000 45,00050,000 11 30,000 25,000 20,000 50,00055,000 17 35,000 30,000 25,000 55,00060,000 25 40,000 35,000 30,000 60,00065,000 32 45,000 40,000 35,000 3.4 failure modes and effect analysis analysis of failure modes and effects analysis (fmea) is by far the most comprehensive tool available for identifying causes. it is an inductive process that can start at the level of the component (shaft, bearing, coupling, etc.). referring to the table in appendix a, the result of the fmea analysis showed that the motor coupling high vibration failure mode is the highest risk probability among the other failure modes. 3.5 weibull analysis the weibull reliability analysis is found to be very useful in characterizing the time of the equipment to failure data and designing appropriate maintenance strategies using the weibull model as a predictive model black and geitam (1997), kelly (1997), bendaya and dufuaa (2000), and jeong and el-sayed, (2000). the two weibull cumulative distribution function (cdf), f(t) parameters define the fraction failure or the likelihood of failure before time t (or unreliability at time t) and have an explicit equation lewis (1987), abernathy et al. (1983), kapur and lamberson (1977). the weibull cumulative density function for twoparameter is given by abernethy et al. (1983): 𝐹(𝑡) = 1 − 𝑒 −( 𝑡 𝜂 ) 𝛽 1 and the weibull reliability function for two parameters is: 𝑅(𝑡) = 𝑒 −( 𝑡 𝜂 ) 𝛽 2 the shape factor (β) is non-dimensional and reveals the type of failure mode, such as early life failure (β<1), constant life (β=1), or wear-out (β>1) on the bath-tub curve. the other weibull parameter (η) is a characteristic life having the same unit as t, and is a function of the mean time to failure (mttf). the general relationship between η and mttf is given by the following equation: (lewis, 1987)15 𝑀𝑇𝑇𝐹 = 𝐸(𝑇) = 𝜂𝛤 ⌈1 + 1 𝛽 ⌉ 3 where β is the shape factor, 𝜂 is characteristic life, and 𝛤 is the gamma function, coupling rubber block failures data have been extracted from the maintenance records with estimated different guaranteed life t0 and are prepared, as shown in table 5, for plotting on weibull probability plotting paper. the straightest plot is produced, as shown in appendix b (figure 9); the guaranteed lifetime zero (t0) is taken to be 30,000 hours and gives the probability that 63% of the couplings fail at characteristic life (η) of about 51,000 hours, or 5.8 years and a β-value of about 2.2 (indicating a wear-out phase of failure). the gamma function value γ(•) is 0.886; the current mttf calculation is = η * 0.886 = 51,000 * 0.886 = 45,186 hours, with β greater than one (1) and the mttf is 5.1 years of service, the scheduled rubber blocks replacement interval can be read directly from the plot at an acceptable or allowable probability of failure (pof). the cumulative failure probabilities at 35,000 hours are 30% which is the suitable time to schedule rubber blocks replacement of the service life. also, the reliability will be improved to 70% at a lifetime of 35,000 hrs. 3.6 reliability analysis summary the major findings of reliability analysis are summarized as follows: a. the time to failure and downtime data of the high hp motors were collected from the computer maintenance management system, monthly operator's on-site log record in the motor running hours, and from online monitoring of process variable data retrieved from the distributed control system of the gas plant. the period of investigation is based on the data from 2004 to 2007. all of a structured reliability and maintainability assessment model: an application to high voltage motors 159 these motors were thoroughly analyzed as their non-repairable parts were analyzed separately, such as coupling. b. motors are found to have a high average annual maintenance downtime of around 10.4 days (2.84% unavailability) of downtime per year per motor. c. to identify the most critical motor failure mode, the pareto analysis was efficiently used. two main factors depend on the pareto analysis criteria, which are the number of failures and downtime. the high-vibration failure mode is the most critical, as shown in the pareto analysis (around 70.9%). d. to visualize the relative contribution of each failure mode, the failure modes of motors are figured out and plotted in pareto charts. the outcome shows that the high vibration of motors has the highest failure mode with a total downtime of (3,122 hours). e. the high vibration failure mode was analyzed by pareto analysis to identify the most critical sources of this vibration. motor couplings were found to be the main source of the high vibration due to misalignment. f. motor coupling root cause analysis (rca) was conducted to identify the root cause of the high vibration. five why and fmea techniques were used as reliability tools to find the cause of coupling failure. the defective rubber blocks at the coupling were identified as the main root cause of the coupling misalignment. g. the various reliability parameters and indices of these most critical motor couplings are β and η. the weibull reliability analysis is found to be very beneficial to characterize the equipment's main time to failures of the motor couplings and to select an appropriate maintenance strategy for a given coupling. 3.7 cost analysis summary the main objective of this section is to transform the reliability assessment results into financials to justify improvements or to make the right decision. one of the important portions of any reliability analysis is cost analysis. the equipment failure rates and the consequences of engineering practices on life equipment should be defined. to communicate clearly technical outcomes within the organization, it is always proven beneficial to convert the results of equipment life and failures into a financial analysis. assumptions: 1. one motor failure will result in the loss of the feed to the gas plant by 230 mmscfd (million standard cubic feet per day) 2. the natural gas industrial price = is 3.5 usd per thousand cubic feet. the cost of unreliability is equal to 805 thousand usd/day due to the loss of a machine. the total of 10.4 days of downtime per year costs usd 8.372e6 per year. the reliability assessment and data analysis have demonstrated that the rubber blocks of the motor coupling are the main reason for the high value of the cost of unreliability. for wear-out failure modes, the cost of an unplanned failure is much greater than the cost of a planned replacement; there will be an optimum replacement interval for minimum cost. 4. recommendations based on financial and reliability analysis, the following major recommendations were developed to address reliability and maintainability. 1. develop preventive maintenance tasks to schedule replaces the coupling rubber blocks at a fixed 4year interval of service life which will reduce the gas plant downtime from 10.4 days/year to 5 days/ year. this technical proposal will improve reliability figures, increase equipment availability and approximately reduce the cost by 52% of unreliability. 2. as an alternative, upgrading the existing coupling type with dry flexible metal coupling can have the following advantages: • all steel design • maintenance-free and wear-free • angular, radial and axial flexible • torsional rigid and free of play • temperature stability • easy installation • smooth operation the flexible metal coupling costs a maintenancefree option for at least ten years of service life, which will eliminate machine downtime. against the maintenance cost, the new coupling is well justified for improving the reliability and overall productivity of the facility. 5. conclusions this paper proposed a structured model to address long-term maintenance and reliability issues. this paper provides an application of the proposed model and a detailed analysis based on historical maintenance and reliability data. in the paper, some tools have been discussed and applied to ensure the readers can understand the application part of the data and assessment. the outcome is very useful for maintenance management and will allow us to take the correct decision in advance based on this analysis to avoid any operational disturbance and plant downtime. motor failures due to coupling rubber blocks are the highest failure modes, which is 73% of the total number of failures. the reliability analysis shows that an improvement in motor coupling is required to improve the overall cost and gross margins. majdi m. obaid and qadeer ahmed 160 the unreliability cost analysis is an effective and practical reliability tool for converting failure information into the cost for facility management to compare the benefits in financial terms. to improve reliability, the cost analysis demonstrated that an investment in motor coupling is well justified. finally, the proposed reliability assessment model using this paper certainly paves ways to develop new tools, elements, and methods to improve the assessment. new tools such as smart algorithms, data analytics, big data along with machine learning can certainly improve the outcome of later assessments. these new algorithms and technologies can also help predict the failures earlier to have replacements based on the real condition, which can optimize the maintenance cost. the proposed method can be applied to any application within the oil and gas industry as well beyond in any asset-intensive industries. conflict of interest the authors declare that there are no conflicts of interest regarding this article. funding the authors did not receive any funding for this research. acknowledgment the authors would like to acknowledge saudi aramco's management and public relations support. references abernethy, r. b., breneman, j. e., medlin, c. h., & reinman, g. l. (1983). weibull analysis handbook. pratt and whitney west palm beach fl government products division. al-douri, a., kazantzi, v., eljack, f. t., mannan, m. s., & el-halwagi, m. m. (2020). mitigation of operational failures via an economic framework of reliability, availability, and maintainability (ram) during conceptual design. journal of loss prevention in the process industries, 67, 104261. bajaj, s., garg, r., & sethi, m. (2018). total quality management: a critical literature review using pareto analysis. international journal of productivity and performance management, 67(1), 128-154. ben-daya, m., & duffuaa, s. o. (2000). overview of maintenance modeling areas. in maintenance, modeling and optimization (pp. 3-35). springer, boston, ma. bloch, h. p., & geitner, f. k. (1997). practical machinery management for process plants: volume 2: machinery failure analysis and troubleshooting. elsevier. choudhary, a., goyal, d., shimi, s. l., & akula, a. (2019). condition monitoring and fault diagnosis of induction motors: a review. archives of computational methods in engineering, 26(4), 1221-1238. dehghanian, p., fotuhi-firuzabad, m., bagherishouraki, s., & kazemi, a. a. r. (2011). critical component identification in reliability centered asset management of power distribution systems via fuzzy ahp. ieee systems journal, 6(4), 593-602. edwin, v. k. d., & syam, s. a. (2010). reliability assurance of large electric motors in a process plant. in 2010 2nd international conference on reliability, safety and hazard-risk-based technologies and physics-of-failure methods (icresh) (pp. 642647). ieee. ferreira, f. j., baoming, g., & de almeida, a. t. (2015). reliability and operation of high-efficiency induction motors. in 2015 ieee/ias 51st industrial & commercial power systems technical conference (i&cps) (pp. 1-13). ieee. hacke, p., lokanath, s., williams, p., vasan, a., sochor, p., tamizhmani, g., ... & kurtz, s. (2018). a status review of photovoltaic power conversion equipment reliability, safety, and quality assurance protocols. renewable and sustainable energy reviews, 82, 1097-1112. jeong, h. s., & elsayed, e. a. (2000). online surveillance and monitoring. in maintenance, modeling and optimization (pp. 309-343). springer, boston, ma. kelly, a. (1997). maintenance strategy. elsevier. kapur, k. c., & lamberson, l. r. (1977). reliability in engineering design. new york. lewis, e. (1987). introduction to reliability engineering (no. 04; ta169, l4.). liu, x., li, j., al-khalifa, k. n., hamouda, a. s., coit, d. w., & elsayed, e. a. (2013). condition-based maintenance for continuously monitored degrading systems with multiple failure modes. iie transactions, 45(4), 422-435 powell, taman & sammut-bonnici, tanya. (2015). pareto analysis. john wiley & sons schwan, m., schilling, k., & de melo, a. a. (2007). reliability centered asset management in distribution networks-process and application examples in electricity distribution. sukhachev, i. s., & sushkov, v. v. (2016, november). assessment dynamics of reliability and resource consumption of "cable—submersible electric motor" system at west siberia oil fields. in 2016 dynamics of systems, mechanisms and machines (dynamics) (pp. 1-4). ieee. zaiets, n., & kondratenko, i. (2019). development of an intelligent system for predicting the reliability of electric motors. in 2019 ieee 39th international conference on electronics and nanotechnology (elnano) pp. 614-619. a structured reliability and maintainability assessment model: an application to high voltage motors 161 appendix a table 6. example of motor fmea. process steps or product functions potential failure mode potential effects of failure severity (1-10) potential causes of failure occurrence (1-10) current controls detection (1-10) risk priority numbe r (rpn) motor stator short circuit motor grounded 8 1. winding stator service lifetime 2. accumulations of oil and dust on stator winding. 3. damaged stator winding insulation. 4 1. replace stator insulation with double insulation. 2. partial cleaning for stator during pm. 3. re-winding motor stator after 25 years of service 96 motor rotor high vibration rotor coil short imbalanc e rotor 6 1. accumulations of oil and dust on stator winding. 2. rotor mass 3. coil service lifetime 2 1. protection system. 2. rotor balancing. 3. re-winding rotor coil. 5 60 motor bearing s high vibration high temperatur e bearings damage 6 1. misalignment between motor and gearbox. 2. improper bearing fit. 3. lack of lubrication 1 1. monitor the vibration level on both bearings in emp 2. monitor the temperature reading 6 36 motor couplin g high vibration misalign ment between motor and gearbox. 7 1. rubber blocks service lifetime. 9 implement alignment target figures 5 315 majdi m. obaid and qadeer ahmed 162 appendix b figure 9. weibull probability plot. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 101-113 *corresponding author’s e-mail: aa2913@bath.ac.uk doi:10.53540/tjer.vol18iss2pp101-113 extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar abdulla alnuaimi* and sukumar natarajan* department architecture and civil engineering, university of bath, bath, united kingdom abstract: indoor cold discomfort in qatar due to “building overcooling” is increasing, as air-conditioning prevails, and global temperatures rise. overcooling is not dependent only on the inappropriate design of cooling systems, but on “international” thermal comfort standards that are not customized for warm climates. international standards are unintentionally biased towards cooler climates and cultures, the application of which in warm climates result in colder indoor temperatures observed by building occupants and increased cooling energy demand. overcooling is an over-expenditure of energy, resulting in uncomfortably cold indoor thermal conditions, and unnecessary carbon emissions. in this study, the analysis of field data from 6 office buildings in doha in a range of indoor thermal conditions and investigation of overcooling on occupant comfort and building performance is conducted. the analysis uncovers over 35% of occupants state being uncomfortably cold and a consensus across the surveys highlight comfortable temperatures higher by 2°c from current setpoint temperatures. greater occupant comfort and energy efficiency are found by increasing the indoor temperature setpoints investigated through thermal comfort analysis and energy simulation models. around 50% decrease in occupant discomfort and a 15% decrease in cooling energy demand is found. such an adjustment in qatar would improve occupant comfort levels and reduce cooling energy demand throughout the existing office building stock. keywords: building energy; extreme climates; overcooling; thermal comfort. البرد شدید االزعاج في المناخات شدیدة الحرارة: دراسة عن التبرید المفرط في المباني المكتبیة في قطر سوكومار ناتراجان النعیمي،عبدهللا التبرید المفرط بسبب االستخدام المتنامي ألنظمة التكییف، في قطر یزداد االنزعاج من البرد في المباني الناتج عن :الملخص والذي یزداد بدوره بزیادة االرتفاع في درجات الحرارة عالمیا. ال یحدث التبرید المفرط بسبب تصمیم أنظمة التبرید غیر خات الدافئة. بطریقة غیر المناسبة فحسب، ولكن بسبب االعتماد على المقاییس العالمیة للراحة الحراریة غیر المالئمة للمنا مقصودة، تعتبر المقاییس العالمیة منحازة للمناخات الباردة وثقافاتھا، ولذا یتسبب تطبیقھا في المناخات الدافئة في درجات حرارة داخلیة منخفضة تالحظ من قبل شاغلي المباني، وترفع من مقدار الطاقة المستھلكة للتبرید. یعتبر التبرید المفرط طاقة، ویتسبب في أحوال حراریة داخلیة باردة وغیر مریحة، كما یزید بشكل غیر ضروري من انبعاثات الكربون. ھدرا لل مباني مكتبیة في الدوحة، متعلقة باألحوال الحراریة الداخلیة، وباستكشاف أثر التبرید 6حللت ھذه الدراسة بیانات میدانیة من % من المستخدمین عبروا عن عدم ارتیاحھم 35لمباني. كشف التحلیل أن المفرط على راحة المستخدمین، وعلى أداء ا بسبب البرودة، كما أكد عدد ال بأس بھ على أن درجات الحرارة المریحة أعلى بدرجتین مئویتین من درجات الضبط الحالیة، ضبط الحرارة الداخلیة، وارتفع مستوى راحة شاغلي المبنى، وازدادت كفاءة الطاقة المستھلكة، من خالل زیادة درجات الذي تم في تحلیل الراحة الحراریة وفي نماذج محاكاة الطاقة، كما وجد أن عدم ارتیاح شاغلي المباني انخفض بمقدار %. قد تؤدي ھذه التغییرات في قطر إلى تحسین مستویات راحة 15%، وأن استھالك الطاقة للتبرید قد انخفض بمقدار 50 الطاقة المستھلكة للتبرید في المباني المكتبیة القائمة في قطر. شاغلي المباني، وتقلیل .الراحة الحراریة ؛زیادة التبرید؛ أقالیم شدیده الحرارة ؛طاقة المباني الكلمات المفتاحیة: extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 102 1. introduction globally, most of the human population resides in the built environment which is constantly growing at the highest rate ever seen in human history. nearly 40% of the total energy produced, which is generated primarily from non-renewable energy sources, is consumed by the built environment (eia, 2019). this built environment energy consumption is estimated to cause about a third of global carbon emissions (stocker et al., 2013; iea, 2018). in warm climates such as qatar, space cooling for buildings as an end-use is the largest energy consumer, on average 40% of the building energy consumption is used for space cooling, (iea, 2018). building space cooling globally is a substantial energy use sector that accounts for 16% of total energy use in buildings and is expected to triple by the year 2050 (iea, 2018). however, anecdotal evidence suggests that overcooling in warm climates such as qatar is occurring and is a likely cause of energy waste. building overcooling is additionally linked to overall occupant discomfort causing negative health impacts and lower productivity (chen and chang, 2012; brager and arens, 2015). as demand for building cooling is expected to increase, likewise the occurrence of overcooling in warm climate built environments will persist. understanding the thermal environments and the approach to thermal comfort in qatari buildings would uncover, building overcooling, its impacts, and possible solutions. 1.1 qatar thermal comfort qatar’s climate is classified as a hot desert climate (bwh) according to köppen and geiger (köppen, 1918). as the country experiences a hot climate with an annual temperature average of 27.5 °c, maintaining thermal comfort in the qatari built environment necessitates active cooling means. over 55% of the total energy production in qatar is consumed by (ayoub et al., 2014). roughly 60% of the energy consumed in buildings in qatar is a result of heating, ventilation, and air-conditioning (hvac) systems (ayoub et al., 2014). as qatar does not have a specific local thermal comfort standard, international standards are the country’s default guidelines for indoor thermal comfort. qatar’s leading green building rating system the global sustainability assessment system (gsas) (gord, 2019) specifically relies on the american society of heating refrigeration and air conditioning engineers (ashrae) standard 55 (ashrae, 2010) and the combined predicted mean vote (fanger, 1970) as the mean for evaluating thermal comfort in qatar. evermore, the international community is developing greater awareness that a universal interpretation of thermal comfort standards leads to applications in extreme climates that result in occupant discomfort and unnecessary expenditure of energy. the development of several international standards such as the iso (international organization for standardization) 7730 (iso, 2005), the ashrae standard 55 (ashrae, 2010), and the en (european standard) 16798-1:2019 (en, 2019) have been initially focused in the cooler regions of north america and europe. therefore, the associated metrics within these standards for establishing thermal comfort come into question, as they have been developed through data from occupants of these regions (fanger, 1970; toftum and fanger, 2002). the pmv has been questioned by researchers for its applicability in extreme climates as the deviation between the actual sensation vote averages and the pmv have been frequently observed in field studies (de dear and brager, 1998; humphreys and nicol, 2002; indraganti and boussaa, 2017). in thermal comfort research conducted in qatar across 5 private buildings and 4 public buildings, the use of the pmv to assess actual occupant comfort has been found to overestimate the occupant’s response (indraganti and boussaa, 2017). the pmv was found to assume a warmer sensation than what is recorded by the occupant provided tsv across all buildings (indraganti and boussaa, 2017). in addition, within the same study in qatar, the comfort temperature evaluated across the entire study was 24.1°c which is warmer than suggested setpoint temperatures by gsas (gord, 2019) guidelines (indraganti and boussaa, 2017). it is apparent that by using such standards, an overcooled building occupant population is observed. by understanding the thermal comfort behaviours of the local building occupants in qatar, the extent of building overcooling in this climate can be determined. 1.2 building overcooling overcooling is a phrase used to represent the phenomenon of excessive cooling within the built environment that is caused by the purposeful expenditure of energy, which results in occupant cold discomfort. overcooling has become a common occurrence in buildings across warm climates such as qatar (cena and de dear, 2001; mendell and mirer, 2009; yang, wong and zhou, 2014; zhou et al., 2016; paliaga et al., 2019). in the media, overcooling is discussed with anecdotal accounts in the built environment in places such as offices and shopping complexes (murphy, 2015; williams, 2016; green, 2019). overcooling is discussed in terms such as a “summer freeze” and “freezing” which highlight the shared feeling of being “too cold” indoors, while outdoors is much warmer (murphy, 2015; williams, 2016; green, 2019). overcooling is not only associated with an increase in occupant discomfort but likewise, the unnecessary consumption of resources. as building cooling is a significant energy end-use in qatar, any overcooling in buildings directly results in wasteful energy use and unnecessary carbon emissions. researchers in the us have estimated that in 2012 overcooling had an energy impact of 103,929 gwh, an environmental impact of over 50,000 kt co2 emissions, and roughly 10 billion usd (derrible and reeder, 2015; aghniaey and lawrence, 2018). abdulla alnuaimi, sukumar natarajan 103 to assess overcooling in buildings, thermal comfort principles in the literature provide several metrics for evaluating building occupants’ discomfort. subjective metrics like the thermal sensation vote (tsv), thermal preference vote (tpv), thermal acceptance vote (tav), and thermal comfort vote (tcv), are used to directly assess building occupants’ comfort. the tsv and tpv are metrics that indicate the occupant's perception of the thermal environment based upon direct feedback on a thermal scale ranging from “cold” to “hot” votes. the tav and tcv indicate the occupant's acceptance or comfort level of space, based upon occupant feedback on an acceptance/satisfaction scale. likewise, to assess discomfort, temperature metrics define discomfort in relation to predetermined temperature-based criteria based upon idealistic temperature conditions (huh and brandemuehl, 2008; may-ostendorp et al., 2011; hu and karava, 2014; pisello and cotana, 2014; kaam et al., 2017; wang et al., 2017). temperature setpoints and ranges can be used which include seasonal air temperatures ranges and humidity alongside external air temperatures to indicate discomfort (auliciems and de dear, 1986; iso, 2005; cen, 2007; mendell and mirer, 2009; ashrae, 2010; fang, winkler and christensen, 2011; bennett et al., 2012; shi et al., 2015; cibse, 2015; abou-ziyan, alajmi and ghoneim, 2016; landsman, brager and doctor-pingel, 2018; estrella guillén, samuelson and cedeño laurent, 2019). further to assess discomfort, model metrics such as the predicted mean vote (pmv) can be used. in the case of the pmv, overcooling is described as a pmv below -0.5 since the pmv model usually defines thermally comfortable votes to occur between -0.5 to +0.5 for most building types such as offices (fanger, 1970; palme et al., 2007; samiuddin and budaiwi, 2018). establishing the extent of overcooling in buildings in warm climates has been conducted using subjective thermal comfort metrics (alnuaimi and natarajan, 2020). combining the tsv and tpv has found building occupants in warm climates in the global south are on average 16% overcooled (alnuaimi and natarajan, 2020). in addition, the widespread adoption of international standards in warm regions such as the middle east and north africa has been found to be a direct source of occupants overcooling determined through occupant responses (elnaklah et al., 2021). as building overcooling is unwanted based upon building occupants’ feedback. assessing the extend of this overcooling is established through the occupants’ responses. further, the evaluation of overcooling in qatar offices is considered alongside the associated energy consumption. 2. methodology establishing the effect of overcooling on cooling energy demand in typical office buildings in qatar involves (i) an estimation of the extent of cold discomfort, in qatar through occupant responses in surveyed buildings, (ii) evaluating a comfort temperature setpoint, which would eliminate the cold discomfort and not result in hot discomfort, and (iii) evaluating the reduction in cooling energy demand possible through the comfort temperature in contrast to the observed temperature. 2.1 data collection the field data were collected during the summer of 2019, starting in may and ending in august, with a total of 6 visits to 4 private and 2 public office buildings in qatar. buildings with medium to large office working conditions were selected as they represent the typical office conditions in qatar. the summer season was targeted for the data collection as the summer period within warmer climates heavily rely on active cooling systems to offset the heat. an anonymous questionnaire incorporating standardized thermal comfort questions found in iso 7730 was utilized for the data collection. consent, as well as an explanation for the data collection, was given from all building occupants participating in the data collection. the questionnaires were produced in both english and arabic format. the questionnaires were deployed to collect qualitative metrics such as the tsv and the tpv directly from the building occupants on a continuous seven-point thermal scale. the questionnaires were distributed to building occupants that have been in a prolonged seated position to ensure a stable metabolic level corresponding to seating. the environmental parameters in the field data collection were measured using calibrated thermal environment measurement sensors that conform to iso 7730. the air temperature (ta), mean radiant temperature (tmrt), relative humidity (rh), and air velocity (av) were taken as spot readings at the desk of each building occupants’ workplace, which was conducted for every building occupant in the study (figure 1). figure 1. thermal comfort measurement instruments attached to data logging software during the collection of data for the environmental parameters. extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 104 air temperature and relative humidity were measured using the swema hc2a-s air humidity probe, mean radiant temperature was measured using the swema 05 767370 globe temperature sensor, and air velocity was measured using the swema 03 767360 anemometer which is summarised in table 1. the clothing insulations values were established by the ashrae index for non-western clothing ensembles (havenith et al., 2015) which was available for the occupants to select the clothing combination that best represented what they were wearing during their survey participation. the metabolic rate of work was uniformly applied as that for seating as surveying of the occupants was focused on office working occupants that have been seated for twenty minutes or more. the environmental parameters, the clothing insulation value (clo), and the metabolic rate of work (met) were assessed to illustrate each building’s general thermal condition. the parameters were summarized for each building to represent the maximum, mean, standard deviation, and minimum value for each collected parameter recorded during the time of the study. further questions within the survey that ask whether the occupants are “thermally comfortable” and if they feel “too cold” under the current indoor thermal conditions were deployed in the survey. both these questions are plotted for each binned indoor temperature to examine the voting pattern on comfort and feeling too cold in the surveyed office buildings in qatar. 2.2 data analysis analysis of the collected data from the surveyed buildings attempts to demonstrate the extent of discomfort and what constitutes comfort based upon direct occupant responses. the occupant responses are analysed in a spectrum of indoor temperature conditions collected from the survey building within this study. table 1. summary of sensors’ specifications. instrument parameter range accuracy swema hc2a-s air humidity probe ta -50 to +100 ºc ± 0.1°c at 1030°c rh 0 to 100% rh ± 0.8%rh at 10-30°c swema 05 767370 globe temperature sensor tr 0 to +50 ºc ±0.1°c at 050°c swema 03 767360 anemometer av 0.05 3.00 m/s at 15 to 30°c ±0.04 m/s at 0.05 to 1.00 m/s ±4% read value at 1.00 to 3.00 m/s at 15 to 30°c initially, the individual values for tsv and tpv collected from each building occupant are accumulated, and average values are evaluated for both tsv and tpv for each surveyed building. in addition, the pmv and ppd for each individual building occupant are calculated and average values are evaluated for each surveyed building. the calculated average for the tsv, tpv, pmv, and ppd are summarised and serve to contrast observed occupant thermal responses to that of international standard models deployed within qatar for each surveyed building. the tsv and the tpv are classified to correspond to a certain thermal classification depending on the location of the vote on the seven-point scale which is as either a cold discomfort [-3, -1), neutral comfort [-1, +1], or hot discomfort (+1, +3] classification. the hot discomfort percentage (hd) is evaluated by accounting for all the hot discomfort classified votes from the tsv and tpv against the total votes within that building or temperature range. likewise, by accounting for all the cold discomfort classified votes in the tsv and tpv the cold discomfort percentage (cd) is evaluated. the hot and cold discomfort percentages are calculated for the entire study binned by 1℃ indoor operative temperature extracted from the combination of the tsv and tpv. evaluating the hd and cd within a spectrum of indoor operative temperatures represents the occupant’s attitude towards hot and cold discomfort within various temperatures. hotter temperatures are expected to illustrate higher hd values and colder temperatures with higher cd values which would confirm the effectiveness of hd and cd in evaluating discomfort. in addition, the hd and cd are calculated for each building in the study to illustrate occupants’ attitudes towards hot and cold discomfort within each surveyed building. further, a comparison between the hd and cd voting pattern and the voting pattern for the question of feeling “too cold” for each surveyed building is established to illustrate the relationship between cd values and occupants feeling “too cold”. finally, to identify comfort, tsv and tpv are plotted against the indoor operative temperature in both a scatter and density plot. the location of the greatest density of votes for the tsv and tpv corresponding to a 0 on the seven-point scale will identify the temperature range which is most comfortable for the surveyed occupants in the building of qatar. the comfort temperature will be extracted from this analysis and utilized as the comfortable temperature condition for further analysis. the observed temperature is extracted from the mean indoor air temperature summarized later in table 3 2.3 data simulation the well-regarded energyplus as a whole building energy simulator is used, which has been applied in several energy studies in qatar (ayoub et al., 2014; al touma and ouahrani, 2018; andric and al-ghamdi, 2020). as a base energy model, the ansi (american abdulla alnuaimi, sukumar natarajan 105 national standards institute)/ashrae/ies (illuminating engineering society) standard 90.1 medium office commercial prototype building energy model was used. the ansi/ashrae/ies standard 90.1 prototype buildings are intended for use in energy simulations in research and practice and are modelled as a typical building construction covering both commercial and residential building types in numerous climates. these energy models are calibrated in agreement with major institutes such as the us department of energy, the international energy conservation code, the american national standards institute, ashrae, and the pacific northwest national laboratory (pnnl) (doe, 2021) and are summarized in table 2. the energy model results are, in turn, validated to energy use intensities by building type based upon standard 90.1-2019 whole building energy metrics (ashrae, 2019). weather data energyplus weather format (epw) for doha are available from the ashrae international weather for energy calculations (iwec) repository. table 2. summary description of the ansi/ashrae/ies standard 90.1 prototype building model. assumed fixed clothing and metabolic rate are also shown. energy model features value floor area (m2) 4,982 floor to ceiling height(m) 23.8 building glazing fraction 0.33 floors 3 occupancy 18.6 m2/person cooling and air distribution system packaged ac unit multizone vav cooling system sizing auto sized to design day fresh air requirement 9.44 l/s/person setpoint temperature 1°c interval for the range (20°c – 26°c) setback temperature 2°c setback after working hours wall construction (u-value) steel frame (0.704 wm -2k-1) roof construction (u-value) insulation above deck (0.363 wm-2k-1) window overall (u-value) fixed window (4.652 wm -2k-1) occupant features clothing insulation (period / m2kw-1 / clo) 9/30-4/29 / 0.155 / 1.0 4/30-9/29 / 0.78 / 0.5 metabolic rate (wm-2 / met) 70 / 1.2 an energy simulation assessment evaluating the operation of a typical office building across the recorded indoor temperature setpoint in qatar is conducted. the cooling temperature setpoint simulations utilize the recorded indoor temperature setpoint at 1°c interval for the range (20°c – 26°c) with a two-degree (°c) setback after working hours for all the simulations. the seven simulations simulate the building’s operation across the range of recorded indoor setpoint temperatures in a typical annual climate expected for doha to establish the average energy demand for cooling annually. this illustrates the reduction in cooling energy demand that would be possible in the range of selected indoor temperature setpoint conditions within a typical office in qatar. 3. results data were collected from 6 office buildings with occupants adopting similar clothing levels and a metabolic level for office seating was considered. the selected buildings were multi-storey office buildings with open planned offices and multiple occupants in the office areas, who accepted the request for field visits and data collection. a total of 330 persons participated in the questionnaires. environmental parameters, namely air temperature (ta), mean radiant temperature (tmrt), relative humidity (rh), and air velocity (av), were collected for the 330 participants and summarized in table 3. 3.1 collection results the average operative temperature (to) observed during the entire study is 23.5℃, with 24.2℃ being the average for private office buildings and 22.1℃ an average for the public office buildings. a variation in relative humidity between the private and public buildings was observed by about 10% rh. the overall relative humidity (rh) across all buildings was observed at 48.1% with an average of 44.6% for private buildings and 55.3% for public buildings. the air velocity (av) on average didn’t vary significantly between the private and public buildings as the average variation was about 0.01m/s2 and the average across all buildings was 0.16m/s2. the average clothing insulation value (clo) is 1.06 across all buildings. notably, the clothing insulation value is much higher at an average of 1.21 clo for public buildings in contrast to the average in private buildings at 0.98 clo. the difference between the clothing insulation value in the public and private buildings is 0.23 clo and could be a result of an increase in formal attire in the public and governmental settings. as occupants that have assumed a prolonged seated position were surveyed, a metabolic rate (met) of 1.20 was considered for all the buildings surveyed. extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 106 table 3. measurement summary of indoor environmental parameters. building n ta tr to rh av clo met private building a (bpr a) 49 max 24.7 25.0 24.9 47.1 0.44 1.26 mean 23.8 24.0 23.9 43.6 0.11 0.95 1.20 st. dev. 0.9 0.8 0.8 1.9 0.08 0.25 min 22.3 22.3 22.3 40.8 0.01 0.61 private building b (bpr b) 34 max 25.1 25.1 25.1 60.9 0.68 1.54 mean 24.1 24.0 24.1 41.7 0.18 0.93 1.20 st. dev. 0.8 0.8 0.8 11.0 0.16 0.30 min 22.8 22.6 22.7 31.2 0.02 0.61 private building c (bpr c) 75 max 25.8 25.7 25.7 41.8 1.28 1.36 mean 24.8 24.8 24.8 38.7 0.17 1.09 1.20 st. dev. 0.7 0.6 0.7 1.8 0.20 0.14 min 23.6 24.0 23.8 36.1 0.03 0.61 private building d (bpr d) 68 max 26.2 25.2 25.7 64.4 1.63 1.26 mean 24.2 24.3 24.2 54.4 0.20 0.97 1.20 st. dev. 0.8 0.6 0.7 4.7 0.27 0.24 min 22.8 23.3 23.1 48.4 0.02 0.61 public building a (bpu a) 57 max 24.7 25.6 25.2 65.3 0.90 1.54 mean 22.8 22.7 22.7 52.5 0.15 1.16 1.20 st. dev. 1.3 1.4 1.3 5.7 0.16 0.26 min 19.8 19.6 19.7 42.0 0.03 0.61 public building b (bpu b) 47 max 22.7 22.0 22.4 63.5 1.10 1.54 mean 21.7 21.5 21.6 58.2 0.20 1.27 1.20 st. dev. 0.3 0.2 0.3 4.8 0.19 0.30 min 21.2 21.1 21.2 49.5 0.02 0.61 figure 2. occupant responses of “no” and “yes” calculated as a percentage from the entire study binned by 1℃ indoor operative temperature extracted from the questionnaire question stating, do you feel “thermally comfortable” about the thermal condition here and now? exploring the question measuring comfort across all the surveyed buildings, the average percentage of occupants answering “yes” which indicates that the occupant is comfortable was 70% (figure 2). further assessment of the question in the range of binned indoor operative temperatures, the answer of “yes” is on average below 80% in the range of 20℃ to 22℃ and above 80% in the range 23℃ to 26℃ (figure 2). observing the questions within the survey that ask whether the occupants are “thermally comfortable” and if they feel “too cold” under the current indoor thermal conditions highlights the occupant attitude towards comfort and cold discomfort within a typical office building in qatar. likewise, exploring whether an occupant is “too cold” across all buildings through the survey question, the average percentage of occupants answering “yes” indicates that the occupant is experiencing cold the discomfort was roughly 32% (figure 3). in the range of binned indoor operative temperatures, occupants are “too cold” on average above 75% in the range of 20℃ to 22℃ and below 25% in the range 23℃ to 26℃ suggesting potential comfortable indoor temperatures within the 23℃ to 26℃ range (figure 3). 80% 64% 61% 20% 24% 18% 17% 20% 36% 39% 80% 76% 82% 83% 2 0 2 1 2 2 2 3 2 4 2 5 2 6 binned indoor operative temperature d o y o u f e e l t h e r m a l l y c o m f o r t a b l e ? no (%) yes (%) abdulla alnuaimi, sukumar natarajan 107 figure 3. occupant responses of “no” and “yes” calculated as a percentage from the entire study binned by 1℃ indoor operative temperature extracted from the questionnaire question stating, do you feel that the air conditioning here and now is creating a condition that can be called “too cold”? table 4. thermal sensation and preference vote summary. building n tsv tpv pmv ppd private building a (bpr a) 49 max 2.00 2.00 0.79 18.32 mean -0.62 0.08 0.33 9.95 st. dev. 1.09 0.89 0.36 4.58 min -3.00 -2.00 -0.31 5.00 private building b (bpr b) 34 max 2.00 2.00 0.88 21.38 mean -0.26 0.41 0.26 9.71 st. dev. 0.99 1.01 0.40 4.65 min -2.00 -2.00 -0.39 5.02 private building c (bpr c) 75 max 3.00 2.00 0.94 23.93 mean -0.02 0.17 0.66 15.16 st. dev. 1.51 1.06 0.21 4.82 min -3.00 -2.00 -0.02 5.00 private building d (bpr d) 68 max 3.00 2.00 0.98 25.61 mean -0.49 0.11 0.46 11.66 st. dev. 1.15 0.93 0.32 5.85 min -3.00 -2.00 -0.13 5.00 public building a (bpu a) 57 max 3.00 2.40 0.93 29.67 mean -0.34 0.06 0.35 11.34 st. dev. 1.57 1.39 0.42 5.53 min -3.00 -3.00 -1.08 5.11 public building b (bpu b) 47 max 3.00 2.00 0.70 15.37 mean -1.10 -0.09 0.22 9.41 st. dev. 1.40 1.21 0.40 2.80 min -3.00 -3.00 -0.65 5.00 3.2 analysis results analyzing the average tsv across all the buildings that have been surveyed within the study, the average value for tsv is -0.48 in contrast to a higher pmv of +0.38 averaged across all buildings. this illustrates a pmv assumption of +0.86 warmer than tsv on a sevenpoint scale (-3 to +3) (table 4). the higher value of the pmv in contrast to tsv indicates that occupants feel much colder than expected by international standards and that any setpoint derived through such a model will result in an unintentional cold condition which would lead the occupants to feel discomfort. additionally, averaged across private buildings tsv is -0.35 and pmv is +0.46 which is a difference of +0.81 warmer in pmv compared to tsv. in public buildings, tsv is -0.72 and pmv is +0.28 with a difference of +1.00 between pmv and tsv (table 4). ppd through all buildings is on average above 10% with an average of 11.2% which indicates slightly elevated occupant discomfort. examining tpv across all buildings surveyed, the average value of preference is closer to neutral at +0.12 on a seven-point scale (-3 to +3) (table 4). examining tpv for private buildings surveyed there is a tendency to want to be slightly cooler which is observed by the value of +0.19, however, in public buildings surveyed the preference slightly shifts to the 20% 37% 27% 80% 82% 72% 83% 80% 63% 73% 20% 18% 28% 17% 2 0 2 1 2 2 2 3 2 4 2 5 2 6 binned indoor operative temperature d o y o u f e e l t o o c o l d ? no (%) yes (%) extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 108 negative side indicating less desire to be cooler in the public buildings (table 4). in contrast, hot discomfort percentages are less than 4% on average from a range of 20℃ to 23℃ and slightly exceed 10% starting at 24℃ to 26℃ with an average of 8% in that operative temperature range (figure 4). this indicates that there are high cold discomfort percentages in cooler temperature ranges such as from 20℃ to 23℃ which coincides with the elevated discomfort within the public buildings as observed by the “comfort” and “too cold” questions in the survey as well as the tsv and tpv averages. examining hot and cold discomfort percentages for the operative temperature ranges highlights that there is elevated cold discomfort in colder indoor temperatures where the cold discomfort percentage average is 47% between 20℃ to 22℃ and is well above 20% within that temperature range (figure 4). in the warmer temperature ranges starting from 23℃ to 26℃, cold discomfort percentages were on average 13% and never exceeding 20% within the selected binned operative temperature ranges (figure 4). analysing the hot and cold discomfort percentages alongside the answers to the “too cold” question in the survey illustrates commonalities between occupants stating that they feel “too cold” and the cold discomfort percentage evaluated from the occupants tsv and tpv. the highest cold discomfort percentage observed exists in public building b at 30%, likewise, the highest percentage answering yes to the “too cold” question is in public building b at 68% (figure 5). similarly, the lowest occurrence of cold discomfort and a yes answer to the “too cold” question exists in private building c at 12% for cd and 15% as yes for feeling “too cold” (figure 5). the average for cold discomfort throughout the private buildings is lower at 13.5% to that of public buildings being at 25.5% (figure 5). in relation to the “too cold” question, the average across private buildings is 23.3% while the average is much higher in public buildings being at 50% (figure 5). it is noteworthy that the cold discomfort percentage arrives at lesser estimates than what is observed by the yes answers to the “too cold” question as cd takes into account both sensation and preference to determine whether an occupant is experiencing cold discomfort or not. assessing the occupant-voting pattern for both the tsv and tpv across the range of indoor operative temperatures collected during the study indicates the temperatures corresponding to the occupants’ comfort. significant clustering of both the tsv and tpv corresponding with the zero value on the seven-point scale suggesting occupant comfort is highest between 24℃ to 25℃ is observed in figure 6. occupant discomfort was relatively low across the private buildings based on the cold discomfort percentage (cd) only at 14% and 23% as a yes for the question about feeling “too cold”. in contrast, occupant discomfort was much higher across the public buildings as (cd) was observed at 25% and 50% as a yes for feeling “too cold”. the average operative temperature recorded for the private buildings was 24.2℃, opposed to that, public buildings were on average 2℃ colder than the private buildings with an average recorded operative temperature of 22.2℃. 3.3 simulation results as occupant discomfort was analyzed across a range of indoor temperatures, likewise the cooling energy demand simulations are conducted across these temperatures to represent the relationship of the cost of cooling against its impact on occupant cold discomfort. figure 4. hot and cold discomfort calculated as a percentage from the entire study binned by 1℃ indoor operative temperature extracted from the combination of the thermal sensation vote and thermal preference vote. 0% 3 % 6% 5% 12 % 13 % 4% 80 % 31 % 30 % 11 % 14 % 11 % 17 % 2 0 2 1 2 2 2 3 2 4 2 5 2 6 binned indoor operative temperature h o t a n d c o l d d i s c o m f o r t p e r g e n t a g e s f o r b i n n e d i n d o o r o p e r a t i v e t e m p r a t u r e hd cd abdulla alnuaimi, sukumar natarajan 109 figure 5. hot and cold discomfort calculated as a percentage from the entire study for each building extracted from the combination of the thermal sensation vote and thermal preference vote and occupant responses of “no” and “yes” calculated as a percentage from each building in the study extracted from the questionnaire question stating, do you feel that the air conditioning here and now is creating a condition that can be called “too cold figure 6. occupant voting pattern for tsv and tpv on a seven-point scale across the range of operative temperatures collected across all buildings within this study. figure 7. annual cooling energy demand (kwh/m2) evaluated by the simulation of a typical office building alongside cold discomfort percentages and “yes” percentages from the “too cold” question across the range of indoor temperature setpoints recorded in the study. 2% 12 % 12 % 7% 14 % 6% 14 % 15 % 12 % 13 % 21 % 30 % 65 % 68 % 84 % 72 % 63 % 30 % 29 % 24 % 15 % 2 5% 3 2% 68 % b p r a b p r b b p r c b p r d b p u a b p u b buildings in study d i s c o m f o r t a n d f e e l i n g t o o c o l d p e r c e n t a g e s hd cd no (%) yes (%) 72 66 60 55 51 46 42 0 10 20 30 40 50 60 70 80 90 20 21 22 23 24 25 26 d is co m fo rt p er ce nt ag e (% ) binned indoor operative temperature cooling energy demand and occupant discomfort for a range of indoor tempratures cd (%) yes (%) annual cooling electricity demand extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 110 throughout the energy simulations conducted, the energy demand for the entire building considering all end-uses is calculated at, a maximum energy demand of 150.8 kwh/m2 for the 20℃ simulation, a minimum energy demand at 108.8 kwh/m2 for the 26℃ simulation, and an average of 126.6 kwh/m2 across all seven simulations is recorded. the proportion of cooling energy demand as an end-use as part of the entire energy demand for the building is on average 45% across the seven simulations. a reduction in cooling energy demand is observed by an average of 8.5% for every 1℃ increase in the indoor setpoint temperature (figure 7). considering the average temperature in the public buildings of 22℃ and the average temperature in the private buildings of 24℃, a reduction of 16% in cooling energy demand is observed (figure 7). further, considering the possible comfort temperature based upon occupant voting pattern of tsv and tpv in figure 6 of 24℃ and 25℃, a reduction is observed of 16% and 23% respectively (figure 7). 4. discussion with the inexistence of local thermal comfort standards in qatar, international thermal comfort standards, ashrae standard-55 being the most prevalent, are uniformly adopted across the country. public buildings in qatar are more susceptible to adopting international standards as public and government-funded buildings usually adhere to regulated building practices, which would include thermal comfort policies. additionally, as public buildings have their expenses subsidized there is less of an awareness of possible reduction of cooling energy demand as building operational costs are handled by an external entity. in contrast, in private buildings, a frugal interpretation of thermal comfort is observed, as there is a direct incentive to reduce cooling energy demand costs as the expenses are adopted directly by building owners. within the distributed survey across all buildings, a direct question gauging whether occupants feel “too cold” was valuable as it facilitated understanding cold discomfort in these buildings alongside the development of additional metrics such as the tsv and tpv combination to gauge cold discomfort. the cold discomfort percentage (cd) has been found to significantly mirror the number of people claiming that they are uncomfortably cold in actual occupant responses from the “too cold” question in the surveyed buildings. this provided evidence that the cold discomfort percentage can serve as an adequate metric for gauging cold thermal discomfort in buildings in future studies. the pmv in this study has been observed to assume an amplified occupant discomfort due to being too hot in comparison to what is recorded for the tsv. the use of the pmv in international thermal comfort standards in warm climates is the rationale for many discussions of excessive cooling in these climates, and a local alternative would improve the service of these climates. across all buildings, when cold discomfort and hot discomfort percentages were evaluated, a significant reduction in cold discomfort was observed when temperatures increased; however, this was not accompanied by an increase in hot discomfort. this would suggest that a simple intervention would be increasing the indoor setpoint temperatures to reduce both cooling energy demand and cold discomfort without an associated increase in hot discomfort for the building occupants. an overall reduction of about 8.5% in cooling energy demand every increase in 1℃ is significant and can contribute to the overall reduction in energy demand in warm climates as populations continue to grow and the deployment of air conditioning expands. 4.1 limitations and future research this study was limited to the data collected during the summer period of a single year and a longer data collection period would yield a larger collection of data. a greater number of occupant responses in a variety of indoor temperature conditions would yield a further understanding of comfort conditions in different setpoint temperature conditions. additionally, as this study was limited to office buildings in qatar, a more expansive view of several building typologies in qatar could illustrate a clearer understanding of the nature of cooling in qatar as domestic cooling is a significant consumer of energy for cooling in warm climates. the future of active cooling due to global warming and focused urbanized population growth in warm climates will deliver massive demands for cooling across the globe. relying on universal standards for comfort that treat the global population as equals would greatly reduce the efficiency and effectiveness of cooling and heating systems in the built environment. domestic cooling is a possibility for future research where almost all qatar residential units have several air conditioning devices which operate throughout the day for most of the year. likewise, the expansion of available occupant thermal comfort data in qatar will serve the creation of a localized thermal comfort standard for qatar. an established standard for thermal comfort in qatar would serve to reduce temperature setpoint which will reduce discomfort and energy consumption. a simple resolution to how warm climates considers thermal comfort in the built environment such as in this study would aid in reducing the impact of the buildings on the world which continues to face many sustainability challenges. 5. conclusion as global temperatures continue to rise, and the human population continues to grow in warmer climates, the dependency on air condition and its utilization in the built environment is expected to expand considerably. abdulla alnuaimi, sukumar natarajan 111 overcooling in buildings is an unnecessary use of energy causing building occupants to feel cold discomfort, which is a significant drawback to the thermal comfort and energy efficiency of buildings in warm climates. the existence of overcooling in office buildings in qatar and can be minimized through simple interventions such as an increase in the indoor setpoint temperature. looking through the range of indoor temperatures and discomfort levels in the data, cold discomfort can be reduced significantly without an increase in hot discomfort, which, on average, occurs at a setpoint temperature that is about 2℃ warmer from ashrae-55 suggestions. applying energy simulations, it is found that an increase in indoor setpoint would reduce cooling energy demand by around 8.5% for every 1℃. through the appropriate application of localized thermal comfort standards developed based on local data and analysis, applicable indoor setpoint temperatures can be developed for qatar. utilizing the appropriate indoor temperature setpoints, an increase in occupant comfort and a reduction of energy consumption can be achieved. conflict of interest the authors declare no conflict of interest. funding abdulla alnuaimi’s doctoral research is supported by the qatar university scholarships program. references abou-ziyan, h., alajmi, a. and ghoneim, a. (2016) ‘sustainable solution for energy needs of existing building in kuwait’, researchgate.net, (may 2017). available at: https://www.researchgate.net/profile/ali_alajmi2/ publication/316738614_sustainable_solution_for _energy_needs_of_existing_building_in_kuwait /links/59101da14585159781875e85/sustainablesolution-for-energy-needs-of-existing-buildingin-kuwait.pdf. aghniaey, s. and lawrence, t. m. (2018) ‘the impact of increased cooling setpoint temperature during demand response events on occupant thermal comfort in commercial buildings: a review, energy and buildings. elsevier, 173, pp. 19–27. alnuaimi, a. n. and natarajan, s. (2020) ‘the energy cost of cold thermal discomfort in the global south’, buildings. multidisciplinary digital publishing institute, 10(5), p. 93. andric, i. and al-ghamdi, s. g. (2020) ‘climate change implications for environmental performance of residential building energy use: the case of qatar’, energy reports. elsevier, 6, pp. 587–592. ashrae (2010) ‘standard 55-2010, thermal environmental conditions for human occupancy’, american society of heating, refrigerating and air conditioning engineers. ashrae (2019) ‘ansi/ashrae/ies standard 90.12019’, energy standard for buildings except lowrise residential buildings. auliciems, a. and de dear, r. (1986) ‘air conditioning in a tropical climate: impacts upon european residents in darwin, australia’, international journal of biometeorology, 30(3), pp. 259–282. doi: 10.1007/bf02189471. ayoub, n. et al. (2014) ‘energy consumption and conservation practices in qatar a case study of a hotel building’, energy and buildings. elsevier ltd, 84, pp. 55–69. doi: 10.1016/j.enbuild.2014.07.050. bennett, d. h. et al. (2012) ‘ventilation, temperature, and hvac characteristics in small and medium commercial buildings in california’, indoor air, 22(4), pp. 309–320. doi: 10.1111/j.16000668.2012.00767.x. brager, g. and arens, e. (2015) ‘creating high performance buildings: lower energy, better comfort’, in aip conference proceedings. american institute of physics, pp. 58–69. cen (2007) ‘en 15251: indoor environmental input parameters for design and assessment of energy performance of buildings addressing indoor air quality, thermal environment, lighting and acoustics’, thermal environment, lighting and acoustics, cen, brussels. cena, k. and de dear, r. (2001) ‘thermal comfort and behavioural strategies in office buildings located in a hot-arid climate’, journal of thermal biology. elsevier, 26(4–5), pp. 409–414. chen, a. and chang, v. w.-c. (2012) ‘human health and thermal comfort of office workers in singapore’, building and environment. elsevier, 58, pp. 172–178. cibse (2015) environmental design cibse guide a. 8th ed. cibse. de dear, r. and brager, g. s. (1998) ‘developing an adaptive model of thermal comfort and preference’. derrible, s. and reeder, m. (2015) ‘the cost of overcooling commercial buildings in the united states’, energy and buildings. elsevier, 108, pp. 304–306. doi: 10.1016/j.enbuild.2015.09.022. doe (2021) energyplus. available at: https://energyplus.net/ (accessed: 2 december 2019). eia (2019) ‘annual energy outlook 2019’, u.s. energy information administration annual energy outlook. available at: https://www.eia.gov /outlooks/aeo/pdf/aeo2019.pdf. elnaklah, r. et al. (2021) ‘thermal comfort standards in the middle east: current and future challenges’, building and environment. elsevier, 200, p. 107899. extreme cold discomfort in extreme hot climates, a study of building overcooling in office buildings in qatar 112 en, e. u. s. (2019) ‘16798-1: 2019’, energy performance of buildings—ventilation for buildings—part, 1. estrella guillén, e., samuelson, h. w. and cedeño laurent, j. g. (2019) ‘comparing energy and comfort metrics for building benchmarking’, energy and buildings, 205. doi: 10.1016/j.enbuild.2019.109539. fang, x., winkler, j. and christensen, d. (2011) ‘using energyplus to perform dehumidification analysis on building america homes’, hvac and r research, 17(3), pp. 268–283. doi: 10.1080/10789669.2011.564260. fanger, p. (1970) thermal comfort: analysis and applications in environmental engineering. new york, : mcgraw-hill. gord (2019) ‘gsas 2019 design & build guidelines manual’. green, p. (2019) do americans need airconditioning?, new york times. available at: www.nytimes.com/2019/07/03/style/air-condition ing -obsession.html. (accessed: 2 february 2019). havenith, g. et al. (2015) ‘a database of static clothing thermal insulation and vapor permeability values of non-western ensembles for use in ashrae standard 55, iso 7730, and iso 9920’, ashrae transactions, 121, pp. 197– 215. hu, j. and karava, p. (2014) ‘model predictive control strategies for buildings with mixed-mode cooling’, building and environment. elsevier ltd, 71, pp. 233–244. doi: 10.1016/j.buildenv.2013.09.005. huh, j. h. and brandemuehl, m. j. (2008) ‘optimization of air-conditioning system operating strategies for hot and humid climates’, energy and buildings, 40(7), pp. 1202–1213. doi: 10.1016/j.enbuild.2007.10.018. humphreys, m. a. and nicol, j. f. (2002) ‘the validity of iso-pmv for predicting comfort votes in everyday thermal environments’, energy and buildings. elsevier, 34(6), pp. 667–684. iea (2018) ‘the future of cooling. opportunities for energy-efficient air conditioning.’, p. 90. doi: 10.1016/s0181-5512(07)79285-9. indraganti, m. and boussaa, d. (2017) ‘comfort temperature and occupant adaptive behavior in offices in qatar during summer’, energy and buildings. elsevier, 150, pp. 23–36. iso (2005) ‘bs en iso 7730: 2005: ergonomics of the thermal environment. analytical determination and interpretation of thermal comfort using calculation of the pmv and ppd indices and local thermal comfort criteria’. international standards organization geneva. kaam, s. et al. (2017) ‘time-averaged ventilation for optimized control of variable-air-volume systems’, energy and buildings. elsevier b.v., 139, pp. 465– 475. doi: 10.1016/j.enbuild.2016.11.059. köppen, w. (1918) ‘classification of climates according to temperature, precipitation and seasonal cycle’, petermanns geogr. mitt, 64(1918), pp. 193–203. landsman, j., brager, g. and doctor-pingel, m. (2018) ‘performance, prediction, optimization, and user behavior of night ventilation’, energy and buildings. elsevier b.v., 166, pp. 60–72. doi: 10.1016/j.enbuild.2018.01.026. may-ostendorp, p. et al. (2011) ‘model-predictive control of mixed-mode buildings with rule extraction’, building and environment. elsevier ltd, 46(2), pp. 428–437. doi: 10.1016/j.buildenv.2010.08.004. mendell, m. j. and mirer, a. g. (2009) ‘indoor thermal factors and symptoms in office workers: findings from the us epa base study’, indoor air, 19(4), pp. 291–302. doi: 10.1111/j.1600-0668.2009.00592.x. murphy, k. (2015) enduring summer’s deep freeze, new york times. available at: www. nytimes.com/2015/07/05/sunday-review/enduringsummers-deep-freeze.html.%0a (accessed: 2 february 2019). paliaga, g. et al. (2019) ‘eliminating overcooling discomfort while saving energy’, ashrae journal, 61(4), pp. 14–28. palme, m. et al. (2007) ‘dynamic sensation of comfort in buildings: the temperature changes effects’, … 28 the conference and 2 nd …, 2(september), pp. 746–750. pisello, a. l. and cotana, f. (2014) ‘the thermal effect of an innovative cool roof on residential buildings in italy: results from two years of continuous monitoring’, energy and buildings. elsevier b.v., 69, pp. 154–164. doi: 10.1016/j.enbuild.2013.10.031. samiuddin, s. and budaiwi, i. m. (2018) ‘assessment of thermal comfort in high-occupancy spaces with relevance to air distribution schemes: a case study of mosques’, building services engineering research and technology, 39(5), pp. 572–589. doi: 10.1177/0143624418764769. shi, z. et al. (2015) ‘digital campus innovation project: integration of building information modelling with building performance simulation and building diagnostics’, simulation series, 47(7), pp. 51–58. stocker, t. f. et al. (2013) ‘climate change 2013: the physical science basis’. cambridge university press cambridge. toftum, j. and fanger, p. (2002) ‘extension of the pmv model to non-air-conditioned buildings in warm climates’, energy and buildings, 34(6), pp. 533–536. doi: 10.1016/s0378-7788(02)00003-8. al touma, a. and ouahrani, d. (2018) ‘the selection of brise soleil shading optical properties for energy conservation and glare removal: a case study in qatar’, journal of building engineering. elsevier, 20, pp. 510–519. wang, w. et al. (2017) ‘energy conservation through flexible hvac management in large spaces: an ips-based demand-driven control (idc) system’, abdulla alnuaimi, sukumar natarajan 113 automation in construction. elsevier, 83(august), pp. 91–107. doi: 10.1016/j.autcon.2017.08.021. williams, e. (2016) overcooling and overheating buildings emits as much carbon as four million cars., the conversation. available at: www.theconversation.com/overcooling-andoverheating-buildings-emits-as-much-carbon-asfour-million-cars-64950. (accessed: 2 february 2019). yang, w., wong, n. h. and zhou, j. (2014) ‘overcooling and thermal comfort in air conditioned university buildings in singapore’, indoor air 2014 13th international conference on indoor air quality and climate, 554, pp. 546–553. zhou, x. et al. (2016) ‘influence of asynchronous demand behavior on overcooling in multiple zone ac systems’, building and environment. elsevier ltd, 110, pp. 65–75. doi: 10.1016/j. buildenv.2016.10.001. 1. introduction 1.1 qatar thermal comfort 1.2 building overcooling 2. methodology 2.1 data collection 2.2 data analysis 2.3 data simulation 3. results 3.1 collection results 3.2 analysis results 3.3 simulation results 4. discussion 4.1 limitations and future research 5. conclusion conflict of interest funding references comp191202_ii.qxd 1. introduction the complexity of digital logic gates that implement a boolean function is directly related to the complexity of the algebraic expression from which the function is implemented. although the truth table representation of a function is unique, expressed algebraically, it can appear in many different forms. boolean functions may be simplified by various means. we can use the rules of boolean algebra to simplify an expression, but this method has a few problems. one is that there is no algorithm, which can be used to determine whether or not we have obtained the minimum solution. when should we stop looking for a simplification theorem? another problem is that we often have to make the expressions more complicated before attempting a simplification. then we rearrange the terms to obtain advantageous groupings that help to simplify the expressions at a later stage. it is against human nature to climb out of a “local minimum” in the hope of finding a better global solution. but this is exactly what we often have to do. another straightforward and simple way of simplification is the “karnaugh map” method. this method is convenient as long as the number of variables does not exceed five or six. as the number of variables increases, the ____________________________________________ *corresponding author email: ashfak@bttb.net.bd excessive number of squares prevents a reason exceed five or six. as the number of variables increases, the excessive number of squares prevents a reason able selection of adjacent squares. the obvious disadvantage of the map is that it is essentially a trial and error procedure, which relies on the ability of the human user to recognize certain patterns. for functions of six or move variables, it is difficult to be sure that the best selection has been made. the tabulation method overcomes this difficulty. it is a specific step-by-step procedure that is guaranteed to produce a simplified standard-form expression for a function. it can be applied to problems with many variables and has the advantage of being suitable for machine computation. however, it is quit tedious and is prone to mistakes because of its routine, and it is a monotonous process. finally, it is just too cumbersome to manipulate boolean expressions by hand. though computer-based tools have been developed for boolean simplification, we still need knowledge of the basic approach. because, observing the symmetries in a circuit’s function helps to understand its behavior. as cad tools become even more sophisticated, we need a deeper knowledge of algorithms they apply to use the tools effectively. this paper assesses previous methods and describes a new method, which uses the advantages of both the map method and the tabulation method and overcomes their disadvantages. ring research 1 (2004) 39-45rnal of engineethe jou a new approach to simplifying boolean functions a. h. m. ashfak habib*1, md. abdus salam2, zia nadir1, and hemen goswami1 1dept. of computer science & engineering, chittagong university of engineering & technology (cuet), bangladesh 2dept. electrical and computer engineering, college of engineering, sultan qaboos university, muscat, oman. received 19 december 2002; accepted october 27, 2003 abstract: there are many benefits to simplifying boolean functions before they are implemented in hardware. a reduced number of gates decreases considerably the cost of the hardware, reduces the heat generated by the chip and, most importantly, increases the speed. but no method is effective for the simplification of boolean functions, if it involves more than six variables. this paper presents a new manual method of simplification that can be effectively applied to problems with a large number of variables. keywords: tabulation method, map-tabulation, karnaugh map π∏≤j á«fhîµde’g iôfg~dg ‘ äéhgƒñdg o~y π«∏≤j ¿eg .iôfg~dg ‘ é¡£hqh é¡≤«≤– πñb á«fé«dƒñdg ∫gh~dg §«°ùñj ‘ ~fgƒødg øe ~j~©dg ∑éæg :¢ü∏îà°ùÿg ≈∏y ájhé◊g á«fé«dƒñdg ∫gh~dg §«°ùñàh ωƒ≤j á≤jôw ájcg ~lƒj ’ ¬fgh é¡∏ªy áyô°s ~jõjh á∏eéµàÿg iôfg~dg øe i~dƒàÿg iqgô◊g øe π∏≤jh é¡féæh áø∏c øe ägò¨àÿg øe òñc o~y ≈∏y ájhé◊g πfé°ùÿg ‘ é¡≤«ñ£j øµªÿg øeh …h~«dg π◊éh §«°ùñà∏d i~j~l á≤jôw ¢vô©à°ùj åëñdg gòg .ägò¨àe áà°s øe ìcg .iaéøµhh .∞fqéc á£jôn ,á£jôÿg á≤jôw ,ádh~÷g á≤jôw :á«méàøÿg ägoôøÿg 2. description of new map-tabulation method our proposed method, which utilizes both the concepts of the karnaugh map (karnaugh, 1953) and the quinemccluskey’s tabulation techniques (mano, 1979), can be termed as the map-tabulation (m-t) method. with our approach the given boolean function should be expressed as the sum of minterms form (a canonical form), such as: f (a,b,c,d,... .. …,u,v,w,x,y,z) = σ (m0, m2, m3, m10, m17, m18, m50, m52, ..., mk, ml) [1] the input variables and the minterms should be divided in to several groups. each group of input variables will contain maximum of four (4) members, whereas each group of minterms will contain a maximum of sixteen (16) terms. for grouping the input variables we should start from the least significant bit. so the least four bits will form the first group. the 5th to 8th bits will form the second group, and so on. the last group can contain fewer members. now for grouping the minterms it should be confirmed that the minterms are arranged in the ascending order of their subscripts. if there are n numbers of input variables, the minterms will be grouped into 2n / 16 = 2(n 4) groups. the minterms among and including m0 and m15 would form the first group, m16 to m31 forms the second group, m32 to m47 forms third, and so on. it is not nec [2] essary that all these groups must contain elements. some could form empty groups. if such an empty group occurs, an asterisks (*) symbol will be placed in the group. two successive groups are separated by a vertical line ( | ). after grouping the elements the function will become: now this function will be rearranged in a tabular form as shown in table 1. before placing the functions in the table, all minterms are converted to their equivalent binary. the binary information of cell [1,1] (i.e. group no. of minterms, group no. of variables) will be solved by using 40 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45 input variables simplification of group-1 simplification of group-2 n n-1 … 2 1 g ro up a b c d . . . … . . u v w x y z kmap simplified terms kmap simplified terms … selected implicant s 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 0 0 1 1 0 0 0 0 0 1 1 0 3 * * * * * * * * * * * * * * * * * * 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 1 1 1 1 0 0 0 1 1 0 0 0 . . . . . . . . . . . . . . . . . . . . . . . . … … … . . . . . . . . . . . . . . . . . . . . . . . . … … 16 * * * * * * * * * * * * * * * * * * 17 * * * * * * * * * * * * * * * * * * . . . . . . . . . . . . . . . . . . . . . . . . … … … . . . . . . . . . . . . . . . . . . . . . . . . … … 32 * * * * * * * * * * * * * * * * * * 33 * * * * * * * * * * * * * * * * * * . . . . . . . . . . . . . . . . … … . . . . . . . . . . . . . . . . … … n mk ml … … … table 1. general format of simplification table 0 2 3 10 17 18 50 52 , , , ,...,...,..., ...,..., , , , , , ......... 1 2 1 , , , , , , , , ,* ......... 1 2 3 4 k l a b c d u v w x y z f n n m m m m m m m m m m n   = −        ∑ a four variable k-map. simplified terms are placed in the adjacent column. similarly the binary information of other cell’s [2, 1], [3, 1], [4, 1], … , [n,1] are simplified accordingly and the corresponding simplified terms are written. now all the simplified terms are compared and the common terms are marked. for the uncommon terms all the kmaps are compared . if a loop of 1s produces an uncommon term or common terms of groups which are at least 15 groups apart, matches exactly in position and structure with another loop of another map, then those terms are marked with a different symbol, “#”. if no such a match is found for a term, then the term is listed in the last column along with its higher significant bits. the binary information of the second group of variables is simplified for the common terms and the “#” marked uncommon terms. if the second group of variables contains only one variable, no simplification for the higher terms would be needed. each variety of the common term is then placed in the last column without combining their higher term. if the second group of variables contains two to four variables, then a kmap of two to four variables should be formed depending on the number of variables of the second group of vari41 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45 table 2. determination of implicants for example 1 ables. if a term is found common in the first , fourth and 15th group of minterms then the zeroth, third and 14th position of the map is marked with one and being simplified. simplified terms are listed aside. for a “#” marked term the groups of minterms are identified, where the loop of 1s which produces that term is common in position and structure. the higher bits are then simplified by using k-map. if the term exist in group three and its loop of 1s matches with groups one and four then the second position of the k-map would be marked with one and the zeroth and third positions are marked with ×; and are treated as the do not care positions. then the map is solved and the simplified terms are listed aside. all the simplified terms of the second group of variables are then checked and the terms are marked as before. the binary information of the third group of variables are then simplified for the common terms and the “#” marked terms, in the same manner as explained before. similarly, all the groups of variables are simplified sequentially. finally, the simplified boolean expression is found by adding all the terms of the last column. 42 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45 table 3. determination of implicants for example 2 3. examples the map-tabulation method is a combination of the kmap and the tabulation methods. up to four variables can be used and it is exactly the same as the karnaugh map method. when the number of variables increases it uses some techniques of the tabulation method. it can be used for problems involving any number of variables. this method is different for problems involving five or more input variables. now we discuss the case for more than four variables. example 1: 43 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45 (a) (b) (c) (d) 0000010 2 √ 0001000 8 √ 00_1_00 (8,12,24,28) 00_1_00 (8,12,24,28) 0000_10 (2,6) 0001_00 (8,12) √ 00_1000 (8,24) √ 0000110 6 √ 0001100 12 √ 0011000 24 √ 000_110 (6,14) 000110_ (12,13) √ 00011_0 (12,14) √ 00_1100 (12,28) √ 0011_00 (24,28) √ 0_11000 (24,56) √ 00011_ _ (12,13,14,15) √ 00011_ _ (12,13,14,15) √ 00_110_ (12,13,28,29) √ 00_110_ (12,13,28,29) √ 00_11_0 (12,14,28,30) √ 00_11_0 (12,14,28,30) √ 0_11_00 (24,28,56,60) 0_11_00 (24,28,56,60) 0001101 13 √ 0001110 14 √ 0010101 21 √ 0011100 28 √ 0110010 50 √ 0111000 56 √ 0001111 15 √ 0010111 23 √ 0011101 29 √ 0011110 30 √ 0111100 60 √ 1110010 114 √ 00011_1 (13,15) √ 00_1101 (13,29) √ 000111_ (14,15) √ 00_1110 (14,30) √ 00101_1 (21,23) √ 001_101 (21,29) √ 001110_ (28,29) √ 00111_0 (28,30) √ 0_11100 (28,60) √ _110010 (50,114) 0111_00 (56,60) √ 0011111 31 √ 0110111 55 √ 0111101 61 √ 00_1111 (15,31) √ 001_111 (23,31) √ 0_10111 (23,55) 00111_1 (29,31) √ 0_11101 (29,61) √ 001111_ (30,31) √ 011110_ (60,61) √ 00_11_1 (13 ,15,29,31) √ 00_11_1 (13,15,29,31) √ 00_111_ (14,15,30,31) √ 00_111_ (14,15,30,31) √ 001_1_1 (21,23,29,31) 001_1_1 (21,23,29,31) 00111_ _ (28,29,30,31) √ 00111_ _ (28,29,30,31) √ 0_1110_ (28,29,60,61) 0_1110_ (28,29,60,61) 00_11_ _ (12,13,14,15,28,29,30,31) 00_11_ _ (12,13,14,15,28,29,30,31) 00_11_ _ (12,13,14,15 ,28,29,30,31) table 4. determination of prime implicants decimal t u v w x y z terms 2,6 6,14 50,114 23,55 8,12,24,28 24,28,56,60 21,23,29,31 28,29,60,61 12,13,14,15,28,29,30,31 0 0 0 0 _ 1 0 0 0 0 _ 1 1 0 _ 1 1 0 0 1 0 0 _ 1 0 1 1 1 0 0 _ 1 _ 0 0 0 _ 1 1 _ 0 0 0 0 1 _ 1 _ 1 0 _ 1 1 1 0 _ 0 0 _ 1 1 _ _ t'u'v'w'yz' t'u'v'xyz' uvw'x'yz' t'vw'xyz t'u'wy'z' t'vwy'z' t'u'vxz t'vwxy' t'u'wx table 5. prime-implicants f (t, u, v, w, x, y, z) = σ (2,6,8,12,13,14,15,21,23,24,28, 29,30,31,50,55,56,60,61,114) [3] since there are seven input variables they should be grouped into two groups. the first group contains four least significant bits and rest of the variables will remain in the second group. the number of input variables are n=7. therefore, the minterms would be grouped into 2(7 4) = 8 groups. after grouping, the function would be reformed as follows: [4] now all the information of this function should be placed into the simplification table as shown in table 2. at first, the binary information of the first group of minterms and first group of input variables are placed into a 4-variable k-map. combining the 1s according to the usual procedure, the simplified terms w’yz’, wx and wy’z’ are found and are listed aside. similarly, with the rest of the combinations of w, x, y and z, all the other groups of minterms are simplified accordingly and their results are listed. these simplified terms are compared. here the term wx is found at row one and two. therefore check is placed at the right of both terms. similarly, all other common terms are marked. now the k-maps are checked for the uncommon terms. the uncommon terms w’yz’ is generated by combining the 1s of the second and sixth positions of the first k-map.no other k-map is found where ‘1s’ exists in the same locations. therefore w’yz’ along with the higher significant bits of the same group (t’u’v’) are written at the column of selected implicants. similarly, xz along with t’u’v is placed in the last column. the uncommon term w’xyz is slightly different. this term is generated by the one of seventh position of the k-map of fourth row. the k-map of the second row also contains a one at the same location. therefore the symbol “#” is placed on the right of this term. similarly the term wxy’ is generated by the ‘1s’ of eight and ninth position of the k-map of fourth row. the k-maps of first and second rows also contain ‘1s’ at the same location. therefore wxy’ is also marked with # on the right. for the common terms and the terms marked with #, the binary information of the second group of input variables are simplified. since the second group of input variables contains three variables, three-variable k-map is used for the simplification. the first term wx is common at the first and second rows, where the value of tuv is 000 and 001 respectively. that is why, 1s are placed at the zeroth and first positions of the map and is solved accordingly. the simplified higher term found from the map is t’u’. so the ultimate simplified term is t’u’wx, and is listed aside. similarly, all other common terms are listed along with their simplified higher terms. now the higher significant bits are simplified for the uncommon terms marked with # symbol. the first such term is w’xyz, and is generated by the one placed at the seventh position of the k-map of fourth row. the seventh position of the k-map of the second row also contains one. these two 1s of the same location of different maps are marked with a circular boundary. at the k-map of higher significant bits the third position is marked with one and the first position is marked as a do not-care location. for simplification, this location can be treated as one or zero. the result of higher bits simplification is t’v. therefore, finally t’vw’xyz is written in the last column. similarly, simplification of higher bits is performed for the other # marked term, and the result is listed. finally the following simplified terms are found; t’u’v’w’yz’, t’u’vxz, t’u’wx, t’u’wy’z’, t’vwy’z’, uvw’x’yz’, t’vw’xyz and t’vwxy’. by adding these terms we get the ultimate simplified function as follows; f = t’u’v’w’yz’ + t’u’vxz + t’u’wx + t’u’wy’z’ + t’vwy’z’ + uvw’x’yz’ + [5] t’vw’xyz + t’vwxy’ example 2: this example is illustrated in dsa method (rhyne et al., 1977) as, 44 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45      =      ∑ 8 114 7 * 6 * 5 * 4 61,60,56,55,50 3 * 2 ,31,30,29,28,24,23,21 1 ,15,14,13,12,8,6,2 1 ,,, 2 ,,, zyxwvut f implicants 2 6 8 12 13 14 15 21 23 24 28 29 30 31 50 55 56 60 61 114 t'u'v'w'yz' √ × × t'u'v'xyz' × × uvw'x'yz' √ × × t'vw'xyz √ × × t'u'wy'z' √ × × × × t'vwy'z' √ × × × × t'u'vxz √ × × × × t'vwxy' √ × × × × t'u'wx √ × × × × × × × × √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ table 6. selection of implicants f (a, b, c, d, e, f, g, h, i) = σ (6,7,22,23,38,70,102,134,262) [6] solution [7] therefore f (a,b,c,d,e,f,g,h,i) = a’b’c’d’f ‘gh + a’b’e’f ‘gh + a’c’d’e’f ‘gh + b’c’d’e’f ‘ghi’ [ 8] 4. verification of results example 1 outlined above culminates in a simplified boolean expression. the accuracy of this result can be verified by using other methods of simplification. the practical use of the k-map method is limited to up to six variable functions. but the given function contains seven input variables. therefore the k-map method will not be convenient for this particular problem. the quinemccluskey’s tabulation can serve the purpose though it is a monotonous method. the simplification of the given function f (t,u,v,w,x,y,z) = σ (2, 6, 8, 12, 13, 14, 15, 21, 23, 24, 28, 29, 30, 31, 50, 55, 56, 60, 61, 114) , by using the tabulation method is shown in tables 4-6. the simplified expression is: f=t’u’v’w’yz’+uvw’x’yz’+t’vw’xyz +t’u’wy’z’+t’vwy’z’+t’u’vxz+ [9] t’vwxy’+t’u’wx this proves that the result obtained earlier, by using the proposed method is correct. 5. benefits our proposed technique is a manual method of simplification of boolean functions. therefore it should not be compared with computer aided minimization techniques. camp (biswas, 1986) is completely a computer aided minimization technique. modified espresso (malik et al., 1988), espresso (brayton et al., 1985) and mini (hong et al., 1974) are more suitable for computer aided minimization than manual minimization. if we consider the manual minimization technique described by biswas (1971) we will find that the method is almost similar to the quinemccluskey’s tabulation method. in this method huge amounts of bit-by-bit comparisons take place, which is quit monotonous and is prone to mistakes. in our proposed method no such bit-by-bit comparison needs to be performed. now let us consider the directed search algorithm (dsa) (rhyne et al., 1977). this method can be used as a manual method as well as a computer aided minimization method. it uses the red-tree procedure. for manual minimization this red-tree procedure is suitable for functions of five to eight variables. our proposed method can easily handle functions having more than eight variables. 6. conclusions this paper studied the problems of minimizing boolean expressions and proposed an optimal method of boolean function simplification. it also evaluated the performance of a new technique. this method overcomes most of the limitations of the other renowned manual methods of simplification. it can be effectively and easily implemented for problems having large numbers of input variables. complexities arise when the number of minterms for which f(mi) = 1 increase excessively. it is less dependent on the ability of the human user, because it uses a maximum of four variable k-maps. moreover, it is not as monotonous as the other manual methods. further research should be conducted to develop an algorithm that uses this new method. references biswas, n. n., 1986, “computer-aided minimization procedure for boolean functions,” ieee trans. on computer aided design, vol. cad-5, no. 2, pp. 303304. biswas, n. n., 1971, “minimization of boolean functions,” ieee trans. on computers, pp. 925-929. brayton, r. k., hachtel, g. d., mcmullen, c. t., and sangiovanni-vincentelli, a. l., (1985), “logic minimization algorithems for vlsi synthesis,” kluver academic publishers. hong, s. j., gain, r. g., and ostapko, d. l., 1974, “mini: a heuristic approach for logic minimization,” ibm j. res. develop, vol. 18, pp. 443-458. karnaugh, m., 1953, “the map method for synthesis of combinational logic circuits,” trans. aiee. pt i, vol. 72, no. 9, pp. 593-599. malik, a. a., robert, k., brayton, a., newton., r., alberto, l., and vincentelli, s., 1988 “a modified approach to two mano, m. m., 1979, “digital logic and computer designm” prentice-hall: u.s.a, pp. 72-83,102-110. rhyne, v. t., noe, p. s., mckinney, m. h., and pooch, u. w, 1977, “a new technique for the first minimization of switching functions,” ieee trans. on computers, vol. c-26, no. 8, pp. 757-764. 45 ashfak habib et al. / the journal of engineering research 1 (2004) 39-45 17 262 16 * 15 * 14 * 13 * 12 * 11 * 10 * 9 ,134 8 * 7 ,102 6 * 5 ,70 4 * 3 ,38 2 ,23,22 1 7,6 1 ,,, 2 ,,, 3 , ∑   =        ihgfedcba f the journal of engineering research (tjer), vol. 17, no. 2, (2020) 75-88 *corresponding author’s e-mail: cjohnokolie@gmail.com doi:10.24200/tjer.vol17iss2pp75-88 spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos alfred s. alademomi1, chukwuma j. okolie1*, babatunde m. ojegbile1, olagoke e. daramola1, johanson c. onyegbula1, rahmat o. adepo1, and wemimo o. ademeno2 1department of surveying and geoinformatics, faculty of engineering, university of lagos, nigeria. 2department of surveying and geoinformatics, d.s. adegbenro ict polytechnic, ewekoro, ogun state, nigeria. abstract: the increasing student population, developmental, and commercial activities within the university of lagos's main campus have led to an increase in daily noise levels. these have disrupted the serenity and tranquillity of the campus environment. in this study, the noise levels within the campus were assessed at 34 locations to determine compliance with international standards for tolerable noise levels in different environments. for three days, noise levels at the stations were measured using the extech 407730 sound level meter. in the analysis, noise level maps were produced using the inverse distance weighted (idw) interpolation, and the correlation between observed noise levels and noise limits were determined. the noise levels measured in the study area ranged from 41.9 96.6dba. it was observed that the minimum noise levels were associated mostly with residential and conservation areas; while the maximum noise levels were mostly associated with commercial areas, vehicle parks, and transportation corridors. generally, the noise levels exceeded the tolerable limits for academic, commercial, and residential areas set by the world health organisation (who) and the national environmental standards and regulations enforcement agency (nesrea). the pearson’s correlation coefficients (r) between the average 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels and both standards (who and nesrea) were derived as 0.63 and 0.58 respectively, indicating a slightly high positive correlation. these findings serve as a valuable knowledge base to inform the university management on the need to implement abatement measures aimed at maintaining the noise levels within tolerable limits. keywords: geographic information system; inverse distance weighted interpolation; noise pollution; pearson’s correlation coefficient; sound level. تحلیل مكاني وإحصائي لمستویات الضوضاء في الحرم الجامعي و ،رحمت ادیبو ،جوھانسون ك اونیجبوال ، اوالجوكي داراموال, ،باباتوندي م اوجیجیبیلي ،شوكوما ج اوكولي ،ألفرید س االدیمومي ویمیمو ادمینو : أدى تزاید عدد الطالب واألنشطة اإلنمائیة والتجاریة داخل الحرم الجامعي الرئیسي لجامعة الجوس إلى زیادة في مستویات الملخص الضوضاء الیومیة. وقد أدت ھذه االضطرابات الضوضائیة المتزایدة إلى اإلخالل بسكینة بیئة الحرم الجامعي وھدوئھ. في ھذه الدراسة التعرض للضوضاء لتحدید اختبارضاء في أربعة وثالثین موقعًا محددًا داخل حرم جامعة الجوس. وقد أجري یتم تقییم مستویات الضو مدى موافقة بیئة الحرم الجامعي للمعاییر الدولیة فیما یتعلق بمستویات الضوضاء المقبولة في مختلف البیئات. تم قیاس الحد األدنى . extech 407730 المحطات لمدة ثالثة أیام باستخدام مقیاس مستوى الصوتوالحد األقصى ومتوسط مستویات الضوضاء في وتراوحت مستویات الضوضاء التي تم قیاسھا .خرائط مستوى الضوضاء عن طریق االستنباط من معكوس المسافة المرجحة رسمتو دیسیبل. ولوحظ أن الحد األدنى لمستویات الضوضاء یرتبط في الغالب بمناطق سكنیة ومناطق 96.6و 41.9في منطقة الدراسة بین ت النقل. وبشكل محمیة؛ في حین أن الحد األقصى لمستویات الضوضاء یرتبط في الغالب بالمناطق التجاریة ومواقف السیارات وممرا عام، تجاوزت القراءات مستویات الضوضاء الحدود المقبولة للمناطق األكادیمیة والتجاریة والسكنیة التي حددتھا منظمة الصحة ومعیاییر بیرسون بین متوسط مستویات الضوضاء العالمیة والوكالة الوطنیة لتطبیق المعاییر واللوائح البیئیة. تشیر معامالت ارتباط دیسیبل على التوالي، إلى عالقة متبادلة إیجابیة مرتفعة 0.58و 0.63المقدرة بـ المشرعین المذكورین أعاله ضاء الصادرة عن الضو قاعدة معارف قیمة إلبالغ إدارة لتكون نتائجھا ھذه الدراسة وتھدف .قلیالً بین مستویات الضوضاء المقاسة وتلك المسموح بھا دولیا .تدابیر خفض الضوضاء بھدف الحفاظ على مستویاتھا في حدود مقبولة الجامعة بضرورة تنفیذ .الضوضاء؛ موجات صوتیة؛ تلوث سمعي؛ مستوى الصوت؛ معامل ارتباط بیرسون :الكلمات المفتاحیة spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 76 nomenclature i intensity of the sound wave (in w/m2) n number of measurements n number of sampled points used for the idw estimation p power parameter r pearson’s correlation coefficient s’ arithmetic mean of the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels. 𝑆𝑆̅ 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 values according to noise standards s1 standard deviations of s’ s2 standard deviations of s� t calculated t value β sound level 𝜆𝜆𝑖𝑖 influence (weight) of sample points db decibels dba a-weighted sound level in decibels df degree of freedom di distance between the point of interest and the sampled point eqn equation h0 null hypothesis h1 alternative hypothesis io standard reference intensity (= 10-12 w/m2) idw inverse distance weighted lga local government area 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 average sound level 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚 minimum sound level 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 maximum sound level max. maximum min. minimum s/n serial number nesrea national environmental standards and regulations enforcement agency who world health organisation 1. introduction sound waves are sensations perceived by the auditory nerves from the impact of acoustic pulses reaching the ear. these sound waves travel through the air, causing vibrations along the path of propagation which can be detected by the auditory nerves and causes a sensation of hearing to humans and other animals (goshu et al., 2017). the intensity of sound waves, which can be described as the level of its impact on the auditory nerves and environment are generally known to decrease with increasing distance from the source (hansen, 2001). the speed of sound in air is approximately 330m/s and depends on factors such as the temperature, pressure, and density of the medium (paulet et al., 2016; sharma, 2017). noise is a sensation that irritates the auditory nerves (zannin, 2013), and is characterised by irregular vibration of the propagating acoustic media or unpleasant acoustic sensations to the hearer. hence, a sound that may be acceptable to one hearer might be noise to another. prolonged exposure to sound, which initially may be pleasing or acceptable to one, may also be classified as noise since the user may become irritated over time by it. hence, the definition of noise is relative to individuals (mehta et al., 2012; who, 2019). the sound level of an acoustic wave is a measure of the intensity of the sound wave, measured with respect to a threshold level on an accepted scale. noise constitutes a nuisance to many people today as noise pollution has greatly increased in our environment and can make people consider leaving urban areas to regions with less noise (obaidat, 2008). all over the world, urbanity and industrialisation have intensified the problem of environmental noise (gholami et al., 2012). the challenges posed by noise pollution cannot be overemphasised, as individuals, today are hardly aware of not just the negative consequences of longterm exposure to it, but what constitutes it (luqman et al., 2013). generally, the populace tends to overlook several factors as significant contributors to increased sound levels in the environment (obiefuna et al., 2013). however, increased sound levels from sources we might deem justifiable could be sources of noise pollution, as the long-term effects are the same as that caused by repeated exposure to what many would personally term irritating sound. noise emanates from different sources in the environment such as the neighbourhood (niemann et al., 2006; laze, 2017), industrial activities (bublić et al., 2010), and transportation (sotiropoulou et al., 2020). these noise sources are most times unnoticed by individuals due to their active contribution to it, investment of time and concentration in daily activities and personal goals. hence, the negative effects of not acknowledging increased noise pollution as a contributory factor to their health conditions and reduced quality of life are heightened (nwobi-okoye et al., 2015). noise is seen as a normal phenomenon by most urban dwellers and commuters, hence the reason for insufficient studies conducted on assessing its impacts, especially in nigeria (monazzam et al., 2014). industrial workers in particular face this challenge as the nature of their work entails long and intense work hours, which requires high concentration levels despite the noise emanating from the factories they work in (zare et al., 2018). it is known that long term exposure to high sound levels and noise can lead to several medical conditions including increased blood pressure, irritation, hearing impairments to permanent deafness (hatamzadi et al., 2018). increased urbanisation today causes increased sound levels due to traffic congestion and industrialisation, among other sources. several studies have identified a very high correlation between impaired hearing, mental health issues, and annoyance of people and the noise levels they are consistently exposed to (ouis, 2002; babisch, 2011; benocci et al., 2016; hammersen et al., 2016; alimohammadi et al., 2019; sonaviya et al., 2019). high noise levels have also been associated with the changes in genes responsible for vascular functions, infiltration, and remodelling of vascular cells leading a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 77 to acute cardiovascular diseases in people who are exposed to it (munzel et al., 2018). also, the auditory, reading and cognitive prowess of different learners in their learning environments are negatively impacted by high noise levels (shield et al., 2003; diaco, 2014). these and many other adverse effects of noise have led to the formulation of rules and several engineering solutions to manage and/or curtail its negative hazard on human (who, 2018; taufner et al., 2020). several engineering and non–engineering solutions have gained good recognition in the mitigation of noise (science for environment policy, 2017). non–engineering solutions are awareness and regulations aimed at reducing noise levels. engineering solutions include the continuous advancements in vehicular engine technology, construction of roads with surface improvement, low– noise tyre technology, the introduction of hybrid electric vehicles, reduction of frictions along rail lines using acoustic grinding, adequate land-use planning/zoning based on compatible sound levels, and sound-proof technology in building and machinery (oyedepo, 2013; benocci et al., 2016; science for environment policy, 2017; pueh et al., 2019; riboldi et al., 2020). all these engineering solutions are hinged on adequate noise mapping using a variety of approaches (pueh et al., 2019; verma et al., 2019; sonaviya et al., 2019; alam et al., 2020). noise pollution mapping is the determination of the noise level variations in an area to analyse the exposure of regions to unacceptable noise levels which exceed set standards. the literature is replete with research on noise pollution mapping, for example in maisonneuve et al. (2010); can et al. (2011); ruge et al. (2013); monazzam, et al., (2014); halperin (2014); leao et al. (2014); lee et al. (2014); zuo et al. (2014); aguilera et al. (2015); carrier et al. (2016); ragettli et al. (2016), and gloaguen et al. (2019). the mapping generally involves measuring noise levels with a suitable device at various observation stations. although noise generally emanates from defined sources, it varies continuously from the location of the source and gradually thins out in intensity at infinity. hence, measurements of noise levels at every single point in a geographic region is an impracticable task. usually, discrete measurements of the noise levels are made at various observation points (nassiri et al., 2016), well distributed throughout the study area to represent the variations in noise levels for the region. then using a geographic information system (gis), the noise level for all other points is interpolated from the values of the observation points measured in the course of the study. gis can assimilate divergent sources of data making it a very versatile analytical tool for modelling continuous spatial data of environmental variables in natural resource management and biological conservation (nwilo, 1998; akeh and mshelia, 2016). although sound/noise levels data are often collected from point sources, gis spatial interpolation techniques such as inverse distance weighted (idw) interpolation can be used in the estimation of continuous spatial data of phenomena over a region of interest to enhance well-informed decisions (li and heap, 2008; farcas and sivertun, 2009; eason, 2013; taghizadeh-mehrjardi et al., 2013). using gis to map noise pollution in the environment can provide beneficial information on the sources and magnitude of the dangers people are exposed to from these negative acoustic sources. furthermore, it provides sufficient information to aid the implementation of pragmatic measures for curtailing the negative effects of continuous exposure to noise (zannin, 2013; lavandier et al., 2016). the university of lagos is an academic institution with a very high student population and high level of commuter traffic, commercial, and social activities which have led to increased noise levels and disturbance within the campus. as an academic environment, the university has to be conducive to learning and research. therefore, it is required to monitor the contributions of the various noise sources to the overall noise concentration level within the campus. this study aims to map noise levels at various locations within the university of lagos main campus and assess conformity with international standards. the assessment is conducted in line with the guidelines of the world health organisation (who) and national environmental standards and regulations enforcement agency (nesrea) – national environmental (noise standards and control) regulations on acceptable noise levels in the environment. the specific objectives of the study are to determine the minimum, maximum and average noise levels at different environments within the university campus; analyse the noise level variations using idw interpolation; compare the measured noise levels with the who and nesrea noise limits using statistical metrics such as pearson’s correlation coefficient and t-test, and recommend measures for noise abatement and control. 2. materials and methods 2.1 study area the study area is the university of lagos's main campus, akoka in lagos mainland local government area (lga) of lagos state, nigeria. the campus is located between longitudes 3o23’00”e – 3o24’30”e and latitudes 6o30’00”n – 6o31’30”n. it is a lowlying area located at the centre of lagos metropolis; bounded to the east by the lagos lagoon and surrounded by densely populated built-up areas. as a citadel of learning, it has a growing student population, contains faculties and other academic and research infrastructure, including recreational facilities, religious buildings, restaurants, and residential buildings. also, traffic congestion is a common occurrence immediately outside its gates. consequently, the main campus is exposed to increased noise pollution levels. spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 78 2.2 station selection before embarking on a noise mapping study, the choice of observation points and periods of observation must be decided upon. the choice of observation periods is usually influenced by general activity levels of the study area (alam, 2011). the measurements were spread around the campus at thirty-four locations in different environment types. it was ensured that the stations covered areas with different characteristics such as noise activity level, land use type, and presence of external noise sources. hence, the locations were spread across academic, residential, recreational, commercial, religious, and conservation areas, including traffic junctions and vehicle parks. consequently, the measurement stations were categorised into eight environment types as shown in table 1. figure 1 presents a map showing the spatial distribution of the measurement stations in the study area. 2.3 noise level measurement sound/noise level is measured in decibels, denoted as db. sound level, 𝛽𝛽, is defined in walker et al. (2014) using eqn.1. 𝛽𝛽 = (10 𝑑𝑑𝑑𝑑𝐴𝐴) 𝑙𝑙𝑙𝑙𝑙𝑙(𝐼𝐼/𝐼𝐼0) (1) where: i = intensity of the sound wave (in w/m2) io = standard reference intensity (= 10-12 w/m2) the intensity (i) of a sound wave at a surface is the average rate per unit area at which energy is transferred by the wave through or onto the surface. the sound level depends on the intensity of the emitted waves which varies with the square of the distance from the source. for the noise level measurement, this study relied on the extech 407730 digital sound level meter. the meter measures and displays sound/noise pressure levels in decibels from 40 to 130db and it permits choices of “a” and “c” weighting. this device has a basic accuracy of ±2db and a digital display resolution of 0.1db, measuring the minimum and maximum noise levels over time table 1. the distribution of measurement stations by location. s/n environment type *n 1 academic 9 2 commercial/ industrial/ shopping 6 3 conservation area 4 4 hospital outdoor 1 5 public outdoor 1 6 recreational 1 7 residential 3 8 traffic 9 total 34 *n – number of measurement stations with its extended microphone windscreen (extech, 2019). a-weighted continuous equivalent sound levels (𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚) were measured daily at the 34 stations for three days in the morning between 0800 – 1100 h and afternoon between 1300 – 1600 h. 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 represents the average sound level, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 is the minimum sound level and 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 is the maximum sound level. the measurements were made by occupying each measurement station for 2 minutes – with the microphone windscreen of the sound level meter placed in the direction of noise sources – recording instantaneous minimum and maximum noise levels detected over the period as well as the instantaneous noise levels at the slow response time. these were manually recorded, the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 values were computed from eqn. 2 (star–orion south diamond project, 2010) and the arithmetic mean which is the average of 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 measurements per station over the days were also computed for both morning and afternoon. 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 = 10 𝑙𝑙𝑙𝑙𝑙𝑙 �∑𝑓𝑓𝑚𝑚 ∗ 10 𝐿𝐿𝑚𝑚 10⁄ � (2) where: 𝑓𝑓𝑚𝑚 = fraction of total time the constant level 𝐿𝐿𝑚𝑚 is present 𝐿𝐿𝑚𝑚 = sound level in dba 2.4 quantitative analysis following the measurement, the data in the field book was entered into a microsoft excel worksheet. the daily 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels for both morning and afternoon periods were summarised to generate the averages of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels for the entire period. the next stage of the analysis evaluated the compliance of the measured noise levels with the guidelines for noise levels specified by the world health organisation (who) in who (2019) and the national environmental standards and regulations enforcement agency (nesrea) in the national environmental (noise standards and control) regulations (2009). the who noise level guidelines for the environment types in this study as well as the national environmental (noise standards and control) regulations (2009) are summarised in table 2. the environment types include the academic areas (faculty complexes and schools), commercial/ industrial/shopping areas (supermarkets and shopping centres), conservation areas (gardens and parks), hospital outdoor (medical centre of the university), public outdoor (open-air/open space arena), recreational area (sports centre and parks) and the residential areas (staff quarters and students’ hostels). a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 79 table 2. noise limits for the different locations. s/n environment type who (dba) nesrea (dba) 1 academic 55 45 2 commercial/ industrial/ shopping 70 70 3 conservation area 45 45 4 hospital outdoor 45 45 5 public outdoor 70 75 6 recreational 70 75 7 residential 55 50 (source: nesrea, 2009; who, 2019) pearson’s correlation coefficient (r) between the average of the noise levels within the campus and the average of the who and nesrea noise limits was calculated using eqn. 3 (dass, 2013). r = ∑𝑋𝑋𝑋𝑋 ��∑𝑋𝑋2��∑𝑋𝑋2� (3) where x and y are the deviations of the measured average noise level and who/nesrea noise limit from their mean values respectively. next, a twotailed t-test at 0.01 level of significance was carried out to compare the results with the who noise limits figure 1. spatial distribution of measurement stations in the study area. spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 80 using eqns. 4 and 5 for t-test (devore, 2012). t = 𝑆𝑆’ − 𝑆𝑆� �𝑆𝑆1 2+𝑆𝑆2 2 𝑁𝑁 (4) df (degree of freedom) = n – 1 (5) where t is the calculated t value to be compared against the critical value obtained from the t-table. s� is the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 value according to noise standard. s1 and s2 are the standard deviations of the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 values, s’ and s� respectively. for this analysis the sample size (n) is 34 – the number of measurements, the null hypothesis used was h0: s’ = s�, implying no significant difference exists between the measured noise levels and the who and nesrea limits. the alternative hypothesis is h1: s’ ≠ s�. pearson’s correlation analysis and t-test were executed using microsoft excel and the statistical package for the social sciences (spss) software. 2.5 generation of noise level maps the noise data was imported from the microsoft excel worksheet into the arcgis environment where it was represented as point shapefiles. using the inverse distance weighted (idw) tool in the spatial analyst toolbox, noise level surfaces were interpolated for the average morning and afternoon noise levels. in the idw dialog box, the input point shapefiles were selected, the z-value field was set to the mean noise levels and a variable search radius was set for the interpolation. the resulting interpolated surfaces were saved as tiff images. according to li and heap (2008), “the inverse distance weighting or inverse distance weighted (idw) method estimates the values of an attribute at unsampled points using a linear combination of values at sampled points weighted by an inverse function of the distance from the point of interest to the sampled points.” the assumption is that sampled points closer to the unsampled point are more related to it than other sample points further away in their values. with idw, the influence (weight) of sample points diminishes as distance increases, and the resultant spatial interpolation is local (isaaks and srivastava, 1989 in li and heap, 2008). the weights can be expressed as shown in eqn. 6: 𝜆𝜆𝑚𝑚 = 1 𝑑𝑑𝑚𝑚 𝑝𝑝� ∑ 1 𝑑𝑑𝑚𝑚 𝑝𝑝�𝑚𝑚𝑚𝑚=1 (6) where di is the distance between the point of interest and the sampled point, p is a power parameter, and n represents the number of sampled points used for the estimation (li and heap, 2008). idw is very applicable in this study due to the property of sound waves, whose intensity diminishes with increasing distance from the source of sound emission. 3. results and discussion 3.1 assessment of noise levels tables 3 and 4 show the averages of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels at all environment types for the morning and afternoon periods respectively. generally, higher 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels are associated with the afternoon period than during the morning period in the following environments: commercial/ industrial/shopping, conservation, hospital outdoor, public outdoor, and recreational areas. however, the converse is the case in academic, residential, and traffic environments where the higher 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 values occur in the morning period and lower noise levels in the afternoon period. the higher morning noise levels and lower afternoon noise levels at traffic environments can be attributed to the early morning rush experienced by students, university staff, and visitors coming into the school to get to their various destinations. hence, large numbers of people in queues for vehicles and increased vehicular traffic and human activity at those locations are the norm in the morning. figure 2 presents the average of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 noise levels at all measurement locations or stations for the morning and afternoon periods respectively while figure 3 presents the average of the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels at all measurement locations for the morning and afternoon periods, respectively. in the morning, the highest mean noise levels occur at station 1 (university main gate, 77.67dba) and station 12 (centre for information technology and systems bus park, 76.33dba). the lowest mean noise levels occur at station 37 (academic staff quarters at ozolua, 59.88dba) and station 24 (lagoon front close to the university guest house, 60.62dba). in the afternoon, the highest mean noise levels occurred at station 1 (university main gate, 74.55dba) and station 5 (sports centre, 74.23dba). the lowest 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 value occurred at station 15 (in the vicinity of the international school, 56.85dba) and station 6 (in the vicinity of kofo students residential hostel, 58.55dba). in both morning and afternoon periods, some of the highest noise levels were observed at the main gate of the campus. this is because there is a continuous stream of vehicles moving in and out of the school through the main gate. also, outside the main gate is the intersection of the university road and st. finbarr’s road which are very busy roads in lagos mainland. changes in the general activity levels in the school led to slight variations in noise levels at other stations between the morning and afternoon observations. these variations were most evident along non-traffic junctions/roundabouts on roads in the school. figures 4 and 5 show the average noise level maps for the morning and afternoon periods respectively. it is observed that there is a clustering of high noise levels at student residential and academic buildings in the afternoon period. this can be explained by the a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 81 table 3. average of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels at all environment types for the morning period. environment type *n average noise level (dba) 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 academic 9 61.25 74.52 67.41 commercial/ industrial/ shopping 6 64.45 70.68 67.59 conservation area 4 60.62 63.37 62.06 hospital outdoor 1 61.73 61.73 61.73 public outdoor 1 64.38 64.38 64.38 recreational 1 64.17 64.17 64.17 residential 3 59.88 70.42 64.53 traffic 9 61.25 77.67 69.95 total 34 59.88 77.67 66.88 timing of the afternoon observations (1300 – 1600 h), in which activity level including traffic and the early morning buzz is reduced. visitors and students are expected to be at their destinations and lecture halls/classes respectively. this results in an accumulation of students and visitors in academic buildings and/or their respective destinations. in the afternoon, there is a decline in noise levels at the international school area due to a reduction in early morning commuting and commercial activities just outside it. generally, students are expected to be in classes during the afternoon, hence the reduced and relatively stable noise levels for the region compared to its morning noise levels. the faculty of science experiences reduced noise levels in the afternoon. this is attributed to the decline in the mass movement of students to and from lecture halls and laboratories in the afternoon. it can be seen that the lowest noise levels occur at relatively isolated regions including some residential buildings, the lagoon front of the school, and some locations at the outer fringes of the campus boundary. these low noise levels can be attributed to the minimal activity level occurring at such locations. the residential and conservation areas in the university environment are well vegetated and as such could have also contributed to the low noise level due to the dampening effect of green trees on noise dispersion (mansouri et al., 2006). the residential areas with such low noise levels are located some distance away from the centre of the school, and in fact, away from any faculty or major activity hub. conversely, regions with the highest noise levels were located at the major activity hubs of the school. these include hostels; traffic junctions where the noise is largely attributable to vehicles; shopping complexes with a lot of commercial activity; the university’s sports centre and major car parks. this is in line with the submission of sulaiman et al. (2018) that an increase in noise level is significantly caused by an increase in traffic volume. in another study by olayinka and abdullahi (2010), traffic noise was identified as the major source of noise in their study area. table 4. average of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚, 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels at all environment types for the afternoon period. environment type *n average noise level (dba) 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 academic 9 56.85 69.88 63.55 commercial/ industrial/ shopping 6 65.77 73.65 71.22 conservation area 4 59.8 64.95 62.1 hospital outdoor 1 65.7 65.7 65.7 public outdoor 1 65.15 65.15 65.15 recreational 1 74.23 74.23 74.23 residential 3 58.55 68.33 62.43 traffic 9 63.05 74.55 68.38 total 34 56.85 74.55 66.34 3.2 comparison with who and nesrea noise limits figure 6 shows a graph of the average noise levels against the who and nesrea noise limits. it can be seen that average noise levels generally exceeded the environmental standards set by the who for most regions in the study area. the implication is that the who standard that is acceptable for the environmental health of the people living in the university of lagos is generally exceeded and this threatens the sensibilities of the community. a few of the measurement stations had nearly equal noise levels as the who noise levels, and some slightly. a total of 11 stations had average noise levels lower than the who standard while 23 stations had noise levels higher than the who standards. the noise levels at 12 stations exceeded the limits with values as high as 10 – 19dba. these stations were mostly found in the academic and conservation environments. in the same vein, a total of 14 stations had average noise levels lower than the nesrea standard while 20 stations had noise levels higher than the nesrea standard. the noise levels at 15 stations exceeded the limits with values as high as 10 – 22dba. these stations were mostly found in the academic and conservation environments. pearson’s correlation coefficient (r), between the average 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels and the who noise standards was derived as r = 0.63. this value indicates a slightly high positive correlation between the measured noise levels and the who limits. going further with the t-test analysis, the computed value of t was 3.10, whereas the critical value from the t-table was 2.73. since this calculated value of t exceeds the acceptable bounds of the t-distribution at 33df (n=34), there is a significant difference between the measurements and the limits set by who. the computed pearson’s correlation coefficient of 0.63 shows a positive correlation between the average noise levels and the who standards at the measurement stations. the t-test revealed significant differences between the average noise levels and the who standards. pearson’s correlation coefficient between the average spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 82 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels and the nesrea noise standards was derived as r = 0.58. this value indicates a slightly high positive correlation between the measured noise levels and the nesrea limits. the computed value of t was 2.81, whereas the critical value from the ttable was 2.73. since this calculated value of t exceeds the acceptable bounds of the t-distribution at 33df (n=34), there is a significant difference between the measurements and the limits set by nesrea. the computed pearson’s correlation coefficient of 0.58 shows a positive correlation between the average noise levels and nesrea standards at the measurement stations. the t-test revealed significant differences between the average noise levels and nesrea standards. figure 7 presents a map showing a classification of measurement stations based on standards for noise limits. special emphasis is placed on the hostel residential areas and academic buildings (especially faculties of learning) since it is expected that many of the students spend a good part of their daily academic lives within these two environments. it was observed that the average noise levels for these regions of interest are unacceptable to a large extent, particularly in faculties. hence, students are exposed to unacceptable noise levels, which could hamper concentration levels and academic performances. furthermore, outside the health centre, the noise level is significantly greater than the acceptable noise standards. this might pose a problem to patients and health care officers/medical personnel as well. however, it should be noted that the noise level measurements in this study were taken outdoors; and not within the health centre or any other building. figure 2. average of the 𝐿𝐿𝐴𝐴 𝑚𝑚𝑖𝑖𝑚𝑚 and 𝐿𝐿𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 noise levels – morning and afternoon. 40 50 60 70 80 90 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 n oi se l ev el (d b a ) station id min. morning min. afternoon max. morning max. afternoon figure 3. average of the 𝐿𝐿𝐴𝐴 𝑒𝑒𝑒𝑒 noise levels – morning and afternoon. 55 60 65 70 75 80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 n oi se l ev el (d b a ) station id mean. morning mean. afternoon a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 83 figure 5. map of inverse distance weighted interpolation of average noise levels afternoon. figure 4. map of inverse distance weighted interpolation of average noise levels morning. spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 84 figure 7. classification of measurement stations based on standards for noise limits. figure 6. average noise levels against who and nesrea noise standards. 40 45 50 55 60 65 70 75 80 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 n oi se l ev el (d b a ) station id average laeq who nesrea a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 85 4. conclusion environmental noise assessment and analysis from this study has revealed that noise levels obtained from the various environments, in general, failed to conform to acceptable environmental noise standards by who and nesrea. inferring from the analysis of morning and afternoon observations, it was detected that noise level was averagely higher in the morning than the afternoon period and this is usually concentrated at the university gate, the university main road, the various classroom areas, and the student residential areas. this could be as a result of workers and students rushing to resume work and class early in the morning. comparing the noise level for morning and afternoon at the two gates of the university, it was observed that average noise levels between 68.1 – 72 dba were recorded at the second gate. however, the buffer area covered by this noise level range was higher in the morning compared to the afternoon observation. conversely, the main gate, sports centre, new hall, and faculty of social science shopping complex recorded higher noise levels in the afternoon observation divergent to the trend of result from the general outcome of the research. this suggests that the areas need immediate attention to mitigate potential health problems that could affect the performance of the students and workers residing in the campus. furthermore, pearson’s correlation coefficient (r) of 0.63 was derived between the observed average noise levels and the who noise standard. between the average noise levels and nesrea noise standards, the coefficient of correlation was 0.58. these values show a slightly high positive correlation between the measured noise levels and standards. significant variation exists between the measured noise levels, and the who and nesrea noise standards. hence, it could be inferred from this study that the health of the various groups of people living in the environment is in danger. however, a consistent noise modelling investigation should be continued in the region to frequently assess the noise pollution level. a continuous data gathering in this regard can help to establish a very good mathematical model to predict the noise level status of the study area. specific regions of interest, the student inhabited areas (residential and academic buildings), showed much higher noise levels than acceptable noise standards. the university of lagos environment is exposed to significantly higher levels of noise than it is deemed appropriate for healthy living. hence, adequate measures must be taken to curb this menace of noise pollution currently being experienced. immediate and definite measures are required. knowing the implications of noise pollution to the health of a society, proper legislation to regulate human activities concerning noise generation is highly recommended to the local, state, and national legislators. noise is damaging but can be controlled drastically to create a good environment. this study aimed to assess the variations in environmental noise levels within the university of lagos in line with international standards. the interpretation from the predicted noise level maps was limited by the number of observations and measurement stations. another limitation was the length of observation which could have been longer but for logistical and budget constraints. notwithstanding, the findings serve as a knowledge resource to inform a better understanding of noise level variations within a university campus. the awareness of the members of the university community to sources and effects of noise and their perceptions of its impacts is considered in another study. conflict of interest the authors declare that there is no conflict of interest regarding the publication of this manuscript. funding this study did not receive any funding from the public, private or not-for-profit sectors. acknowledgement the authors are thankful to professor alabi soneye of the department of geography, the university of lagos for the provision of the noise level meter used in this study. the authors are also grateful to the editors and reviewers of this paper for their critical feedback that helped in shaping and improving the quality of the output. references aguilera i, foraster m, basagaña x, corradi e, deltell a, morelli x, phuleria hc, ragettli ms, rivera m, thomasson a, slama r, künzli n (2015), application of land use regression modelling to assess the spatial distribution of road traffic noise in three european cities. journal of exposure science & environmental epidemiology, 25(1):97–105. ncbi.nlm.nih.gov/pubmed/252277 31 akeh gi, mshelia ad (2016), the role of geographic information system in urban land administration in nigeria. moj eoc environ sci 1(1): 00004. doi: 10.15406/mojes.2016.01. 00004. alam p, ahmad k, afsar s, akhtar n, (2020), 3d noise mapping for preselected locations of urban area with and without noise barriers: a case study of delhi, india. noise mapping, 7:74–83 https://doi.org/10.1515/noise-2020-0006 alam w (2011), gis based assessment of noise pollution in guwahati city of assam, india. int. j. environ. sci., 2(2): 743 – 751. www. semanticscholar.org/paper/gis-based-assess-men t-of-noise-pollution-in-guwahati-alam/373fc-62 a055d2c210660b7d2e0c5f81bac7dd1c7 https://www.ncbi.nlm.nih.gov/pubmed/?term=phuleria%20hc%5bauthor%5d&cauthor=true&cauthor_uid=25227731 https://www.ncbi.nlm.nih.gov/pubmed/?term=ragettli%20ms%5bauthor%5d&cauthor=true&cauthor_uid=25227731 https://www.ncbi.nlm.nih.gov/pubmed/?term=rivera%20m%5bauthor%5d&cauthor=true&cauthor_uid=25227731 https://www.ncbi.nlm.nih.gov/pubmed/?term=thomasson%20a%5bauthor%5d&cauthor=true&cauthor_uid=25227731 https://www.ncbi.nlm.nih.gov/pubmed/?term=slama%20r%5bauthor%5d&cauthor=true&cauthor_uid=25227731 https://www.ncbi.nlm.nih.gov/pubmed/?term=k%c3%bcnzli%20n%5bauthor%5d&cauthor=true&cauthor_uid=25227731 spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 86 alimohammadi i, fakhradin ak, jamileh a, shahram v, kazem r, mohammad hc, mohammad a (2019), association between noise annoyance and hearing loss caused by chronic exposure to noise among the workers of an automotive industry. journal of trends in medicine, 19:2–5; 2019. doi: 10.15761/tim.1000203 babisch w (2011), cardiovascular effects of noise. noise & health. 13(52): 201-204. benocci r, smiraglia m, zambon g, roman h (2016), predicting hourly traffic noise from traffic flow rate model: underlying concepts for the dynamap project. noise mapping; 3:130–139. doi 10.1515/noise-2016-0010 bublić i, tudor i, skelin d (2010), small scale noise mapping of industrial plants. 1st eaa-euroregio 2010. congress on sound and vibration, 15-18 september 2010, ljubljana, slovenia. can a, van renterghem t, rademaker m, dauwe s, thomas p, de baets b, bot-teldooren d (2011), sampling approaches to predict urban street noise levels using fixed and temporary microphones. j. environ. monit., 13(10): 2710–2719. https://pubs. rsc.org/en/content/articlelanding/2011/em/c1em10 292c carrier m. apparicio p, séguin a. (2016), road traffic noise in montreal and environmental equity: what is the situation for the most vulnerable population groups? j. transp. geogr., 51: 1-8, issn 0966-6923. https://www. sciencedirect.com/science/article/abs/pii/s096669 2315002021 dass hk (1988), advanced engineering mathematics (first edition ed.). https://books. google .com.ng/books?id=ikzlagaaqbaj&printsec=co pyright&redir_esc=y#v=onepage&q&f=false . devore jl (2012), probability and statistics for engineering and the sciences (eighth ed.): richard stratton. https://fac.ksu.edu.sa/sites/default/files/probability _and_statistics_for_engineering_and_the_sciences .pdf diaco sb (2014), effects of noise pollution in the learning environment on cognitive performances. liceo journal of higher education research, 10(1), p. 83+. gale academic onefile, accessed 21 aug. 2020. eason s (2013), strategic noise mapping with gis for the universitat jaume i smart campus: best methodology practices. m.sc. thesis. universitat jaume i, castellon, spain. https://run.unl.pt /bitstream/10362/9194/1/tgeo0102.pdf extech (2019), 407730: digital sound level meter. http://www.extech.com/product/407730 farcaş f, sivertunb a (2009), road traffic noise: gis tools for noise mapping and a case study for skåne region. the international archives of the photogrammetry, remote sensing and spatial information sciences, vol. 34, part xxx. https://www.isprs.org/proceedings/xxxviii/2w11/farcas_sivertun.pdf gholami a, nasiri p, monazzam m, gharagozlou a, monavvari sm, afrous a (2012), evaluation of traffic noise pollution in a central area of tehran through noise mapping in gis. adv. environ. biol., 6(8): 2365-2373. http://www. aensiweb .com/old/ aeb/2012/2365-2371.pdf gloaguen jr, can a, lagrange m, petiot jf (2019), road traffic sound level estimation from realistic urban sound mixtures by non-negative matrix factorization. appl. acoust., 143: 229-238. https://www.sciencedirect.com/science/article/abs/ pii/s0003682x18303475 goshu bs, mamo h, zerihun s (2017), urban noise: a case study in dire-dawa city, ethiopia. eur. j. biophys., 5: 17-26. https://www.semanticscholar. org/paper/urban-noise% 3a-a-case-study-in diredawa-city %2c-goshu-mamo/37d822e3 b5f53e28e-22d5f78fd82f8acac0a37de halperin d (2014), environmental noise and sleep disturbances: a threat to health? sleep. sci., 7(4): 209–212. https://www.sciencedirect.com/science /article/pii/s1984006314000601 hammersen f, niemann h, hoebel j (2016), environmental noise annoyance and mental health in adults: findings from the crosssectional german health update (geda) study 2012. international journal of environmental research and public health, 13, 954; doi:10.3390/ijerph13100954 hansen ch (2001), fundamentals of acoustics. occupational exposure to noise: evaluation, prevention and control. world health organization, 23-52. hatamzadi j, talaiekhozani a, ziaei g (2018), evaluation of noise pollution in ilam city in 2018. j. air pollut. health., 3: 73-80. https://www.researchgate.net/publication/3266841 23_evaluation_of_noise_pollution_in_ilam_city _in_2018 isaaks eh, srivastava rm (1989), applied geostatistics. oxford university press, new york. https://www.questia.com/library/62419760/applie d-geostatistics lavandier c, aumond p, gomez s, dominguès c (2016), urban soundscape maps modelled with geo-referenced data. noise mapping, 3:278–294. doi: 10.1515/noise-2016-0020 laze k (2017). findings from measurements of noise levels in indoor and outdoor environments in an expanding urban area: a case of tirana. noise mapping 4:45–56. doi 10.1515/noise-2017-0003. leao s, ong k, krezel a (2014), 2loud?: community mapping of exposure to traffic noise with mobile phones. environ. monit. assess., 186(10): 6193206. https://www.ncbi.nlm.nih.gov/pubmed/24920095 lee ey, jerrett m, ross z, coogan pf, seto eyw (2014), assessment of traffic-related noise in three cities in the united states. environ. res., 132:182https://fac.ksu.edu.sa/sites/default/files/probability_and_statistics_for_engineering_and_the_sciences.pdf https://fac.ksu.edu.sa/sites/default/files/probability_and_statistics_for_engineering_and_the_sciences.pdf https://fac.ksu.edu.sa/sites/default/files/probability_and_statistics_for_engineering_and_the_sciences.pdf http://www.extech.com/product/407730 https://www.isprs.org/proceedings/xxxviii/2-w11/farcas_sivertun.pdf https://www.isprs.org/proceedings/xxxviii/2-w11/farcas_sivertun.pdf https://www.sciencedirect.com/science/article/abs/pii/s0003682x18303475 https://www.sciencedirect.com/science/article/abs/pii/s0003682x18303475 https://www.questia.com/library/62419760/applied-geostatistics https://www.questia.com/library/62419760/applied-geostatistics https://www.ncbi.nlm.nih.gov/pubmed/24920095 a. alademomi, c. okolie, b. ojegbile, o. daramola, j. onyegbula, r. adepo, and w. ademeno 87 189. https://www.ncbi.nlm.nih.gov/pubmed/24792415 li j, heap ad (2008), a review of spatial interpolation methods for environmental scientists. geoscience australia, record 2008/23. https://d28rz98at9flks.cloudfront.net/68229/rec20 08_023.pdf luqman ya, rowland ag, zhang y, zainab uo (2013), work environment noise levels and risk locations in two selected commercial areas in ibadan, nigeria. global j. med. res., 13(6). https://globaljournals.org/item/2579-workenvironment-noise-levels-and-risk-locations-intwo-selected-commercial-areas-in-ibadan-nigeria maisonneuve n, stevens m, ochab b. (2010), participatory noise pollution monitoring using mobile phones. inf. polity., 15(1,2):51–71. https://dl.acm.org/doi/10.5555/1858974.1858981 mansouri n, pourmahabadian m, ghasemkhani m (2006), road traffic noise in downtown area of tehran, iran. j. environ. health. sci. eng, 3(4): 267-272 mehta r, zhu r, cheema a (2012), is noise always bad? exploring the effects of ambient noise on creative cognition. journal of consumer research, 39(4): 784-799. taufner md, gama ap, slama jg, torres cb (2020), noise metrics analysis in schools near airports: a brazilian case study. noise mapping, 7: 21 – 34. http://doi.org/10.1515/noise 2020 0003. monazzam mr, abbaspour ekm, nassiri p, taghavi l (2014), spatial traffic noise pollution assessment – a case study. int. j. occup. med. environ. health., 28(3):625-634. https://www.ncbi .nlm.nih.gov/pubmed/26190737 munzel t, schmidt f, sebastine s, herzog j. andrea d (2018), environmental noise and cardiosvacular system. journal of american college of cardiology, 71: 688 – 697. nassiri p, elham k, mohammad rm, majid a, lobat t (2016), analytical comparison of traffic noise indices — a case study in district 14 of tehran city. j. low freq. noise. vibr. act. control., 35(3): 221–229. https://journals.sagepub .com/doi/full/10.1177/0263092316660917 nesrea (2009), national environmental (noise standards and control) regulation, national environmental standards and regulations enforcement agency (nesrea) 2009. http://www.nesrea.org/lawsandregulations niemann h, bonnefoy x, braubach m, hecht k, maschke c, rodrigues c, röbbel n. (2006). noise-induced annoyance and morbidity results from the pan-european lares study. noise health. 2006;8(31):63-79. doi:10.4103/14631741.33537. nwilo pc (1998), data acquisition by ground survey methods. in c.u. ezeigbo (ed.) principles and applications of geographic information systems. series in surveying and geoinformatics. isbn 978-027-764-1. nwobi-okoye cc, uyaelumuo ae, duru ca, okoronkwo goo (2015), analysis and modelling of road traffic noise in onitsha metropolis, nigeria. int. j. adv. multidiscip. res., 1(1): 1-20. http://rex.commpan.com/index.php/ijamrr/article/v iew/35 obaidat mt (2008), spatial mapping of traffic noise levels in urban areas. j. transp. res. forum., 47(2): 89-102. https://pdfs.semanticscholar.org/69ca/2b0e8318a0 097090948cfef833b66e3f19d6.pdf obiefuna jn, bisong fe, ettah eb (2013), a gis analysis of noise islands in calabar metropolis, nigeria. j. environ. sci., (3), 12. https://pdfs.semanticscholar.org/3a11/7c93230772 4a169e9cd97b2679431420c1ff.pdf olayinka os, abdullahi sa (2010). comparative analysis of noise descriptors in some selected areas in ilorin metropolis, nigeria. noise control engineering journal, 58(6): 646-657. ouis d (2002), annoyance caused by exposure to road traffic noise: an update. noise health [serial online] available from: http://www. noiseandhealth.org/text.asp?2002/4/15/69/31785 oyedepo so (2013), effective noise control measures and sustainable development in nigeria. world journal of environmental engineering, 1(1): 5-15 available online at http://pubs.sciepub.com/wjee/1/1/2 doi:10.12691/wjee-1-1-2 paulet mv, salceanu a, neacsu om (2016), ultrasonic radar. in 2016 international conference and exposition on electrical and power engineering (epe) (pp. 551-554). ieee. pueh hl, kian ml, garg s (2019), a case study of recording soundwalk of fushimi inari shrine in kyoto, japan using smartphone. noise mapping, 6:94–102 https://doi.org/10.1515/noise-2019-0008 ragettli ms, goudreau s, plante c, fournier m, hatzopoulou m, perron s, smargiassi a (2016), statistical modeling of the spatial variability of environmental noise levels in montreal, canada, using noise measurements and land use characteristics. j. expo. sci. environ. epidemiol., 26(6):597-605. https://www.ncbi.nlm.nih.gov/pubmed/26732373 riboldi ced, lorenzo t, mariani l, rolando a, salucci f. (2020), predicting the effect of electric and hybrid-electric aviation on acoustic pollution. noise mapping, 7:35–56. doi: 10.1515/noise2020-0004 ruge l altakrouri b, schrader a (2013), sound of the city continuous noise monitoring for a healthy city. in proceedings of the 5th international workshop on smart environments and ambient intelligen., 670–675. http://www.csee. usf.edu/~ mlabrador/share/workshops/papers/p670-ruge.pdf science for environment policy (2017), future brief: noise abatement approaches. april 2017 issue 17. https://www.ncbi.nlm.nih.gov/pubmed/24792415 https://d28rz98at9flks.cloudfront.net/68229/rec2008_023.pdf https://d28rz98at9flks.cloudfront.net/68229/rec2008_023.pdf https://dl.acm.org/doi/10.5555/1858974.1858981 https://www.ncbi.nlm.nih.gov/pubmed/26190737 https://www.ncbi.nlm.nih.gov/pubmed/26190737 http://rex.commpan.com/index.php/ijamrr/article/view/35 http://rex.commpan.com/index.php/ijamrr/article/view/35 https://pdfs.semanticscholar.org/69ca/2b0e8318a0097090948cfef833b66e3f19d6.pdf https://pdfs.semanticscholar.org/69ca/2b0e8318a0097090948cfef833b66e3f19d6.pdf https://pdfs.semanticscholar.org/3a11/7c932307724a169e9cd97b2679431420c1ff.pdf https://pdfs.semanticscholar.org/3a11/7c932307724a169e9cd97b2679431420c1ff.pdf https://www.ncbi.nlm.nih.gov/pubmed/26732373 https://www.ncbi.nlm.nih.gov/pubmed/26732373 spatial and statistical analysis of environmental noise levels in the main campus of the university of lagos 88 https://ec.europa.eu/environment/integration/resea rch/newsalert/pdf/noise_abatement_approaches_f b17_en.pdf sharma a (2017), the generalized form of principia’s third law of motion. indian j. sci. technol, 10(34): 1-18. shield bm, dockrell je (2003), the effects of noise on children at school: a review. journal of building acoustics 10(2):97 – 106. sonaviya dr, bhaven n t (2019), 2-d noise maps for tier-2 city urban indian roads. noise mapping, 2019; 6:1-6. https://doi.org/10.1515/noise-20190001 sotiropoulou a, karagiannis i, vougioukas e, ballis a, bouki a (2020), measurements and prediction of road traffic noise along high-rise building façades in athens. noise mapp. 2020; 7:1–13. star–orion south diamond project (2010), environmental impact statement, appendix 5.2.5a: fundamentals of acoustics, sx03733 – section 5.0, version 1.0, 7p. https://www.ceaaacee.gc.ca/050/documents_staticpost/46277/89411 /05-appendix_5-2-5-afundamentals_of_acoustics.pdf sulaiman fs, darus n, mashros n, haron z, yahya k (2018), traffic noise assessment at residential areas in skudai, johor. e3s web of conferences, 34, 02024. doi:10.1051/e3sconf/20183402024 taghizadeh-mehrjardi r, zare m, zare s (2013), mapping of noise pollution by different interpolation methods in recovery section of ghandi telecommunication cables company. johe, winter-spring 2013: 2 (1-2). doi: 10.18869/acadpub.johe.2.1.2.1. verma d, jana a, ramamritham k (2019), classification and mapping of sound sources in local urban streets through audioset data and bayesian optimized neural networks. noise mapping.6:52 -71. https://doi.org/10.1515/noise2019-0005 walker j, resnick r, halliday d (2014), fundamentals of physics (10th ed.). http://www.primeuniversity.edu.bd/160517/vc/eb ook/download/fundamentals-of-physicstextbook.pdf who (2018), environmantal noise guidelines for the europian region, world health organistaion, 2018. (date accessed: 13th august, 2020). https://www.euro.who.int/__data/assets/pdf_file/0 008/383921/noise-guidelines-eng.pdf?ua=1 who (2019), guideline values (date accessed: 18th december, 2019). https://www.who.int/docstore/ peh/noise/comnoise-4.pdf zannin pht (2013), noise mapping of an educational environment. can. acoust., 41(1): 27-34. zare s, hasheminejad n, shirvan he, hasanvand d, hemmatjo r, ahmadi s (2018), assessing individual and environmental sound pressure level and sound mapping in iranian safety shoes factory. romanian j. acoust. vib., 15(1). http://rjav.sra.ro/index.php/rjav/article/view/46/33 zuo f, li y, johnson s, johnson j, varughese s, copes r, liu f, wu hj, hou r, chen h (2014), temporal and spatial variability of traffic-related noise in the city of toronto, canada. sci. total. environ., 15;472:1100-1107. https://www.ncbi. nlm.nih.gov/pubmed/24361745 https://ec.europa.eu/environment/integration/research/newsalert/pdf/noise_abatement_approaches_fb17_en.pdf https://ec.europa.eu/environment/integration/research/newsalert/pdf/noise_abatement_approaches_fb17_en.pdf https://ec.europa.eu/environment/integration/research/newsalert/pdf/noise_abatement_approaches_fb17_en.pdf https://doi.org/10.1515/noise-2019-0001 https://doi.org/10.1515/noise-2019-0001 https://doi.org/10.1515/noise-2019-0005 https://doi.org/10.1515/noise-2019-0005 http://www.primeuniversity.edu.bd/160517/vc/ebook/download/fundamentals-of-physics-textbook.pdf http://www.primeuniversity.edu.bd/160517/vc/ebook/download/fundamentals-of-physics-textbook.pdf http://www.primeuniversity.edu.bd/160517/vc/ebook/download/fundamentals-of-physics-textbook.pdf https://www.euro.who.int/__data/assets/pdf_file/0008/383921/noise-guidelines-eng.pdf?ua=1 https://www.euro.who.int/__data/assets/pdf_file/0008/383921/noise-guidelines-eng.pdf?ua=1 http://rjav.sra.ro/index.php/rjav/article/view/46/33 https://www.ncbi.nlm.nih.gov/pubmed/24361745 https://www.ncbi.nlm.nih.gov/pubmed/24361745 civl070701.qxd the journal of engineering research vol. 6, no. 1 (2009) 66-75 1. introduction international building code (international code council, 2003) has provisions to generate design response spectra for response spectrum analysis of multi-story building structures. to comply with these provisions, response spectrum analysis is performed using a design response spectrum based on parameters specified by the code. internal forces obtained by such a code compliant analysis are in fact predicted internal forces that may occur when the building is subjected to an earthquake represented by the design response spectrum. however, the parameters that control the shape of the design response spectrum are based on soil stiffness and seismic activity of the region and therefore these parameters can only be evaluated approximately. since the structural design of a ______________________________________ corresponding author’s e-mail: tnahhas@uqu.edu.sa building is based on the predicted internal forces in its structural members, the question arises about the effect of error or approximation in the evaluation of the ibc parameters on the predicted internal forces. this question becomes even more important when a number of design alternatives are to be analyzed. this question can only be answered by assessing the sensitivity of the predicted internal forces to these code parameters. such sensitivity data that can show the effect of approximation or error on the predicted internal forces is the focus of this paper. the sensitivity of the predicted internal forces to the code parameters is important for structural designers and engineers as well as for code developers and the authorities trying to ensure safer earthquake resistant buildings. adequacy of building codes is an issue of concern in most countries of the world. such sensitivity data are important to determine the adequacy of seismic design provisions of building codes and to gain confidence in its application to sensitivity of predicted maximum bending moment to ibc design response spectrum parameters tariq m. nahhas department of civil engineering, umm al-qura university, po box 16222, makkah, saudi arabia received 1 july 2007; accepted 5 may 2008 abstract: the parameters used to generate ibc (international building code) compliant design response spectra are specified only approximately for a given site in most cases and values may even be erroneous. so, the question arises about the effect of approximation or error in the values of these parameters on the predicted internal forces in the structural elements of a building calculated on the basis of the provisions of ibc. this paper explores this issue and presents the sensitivities of design response spectra and maximum moments induced in the structure to these ibc parameters. within the critical range of short time periods, the variation in peak acceleration of design response spectra for various site classes is found non-uniform. only peak acceleration is affected by varying ss. thus, for buildings with the fundamental time period of around 0.6 sec and greater, the parameter ss of ibc is immaterial. also, for a given site class, the design response spectra are almost identical in the peak acceleration region for all values of s1. the maximum moment is insensitive to ss except for the site class e. this implies that any error in the evaluation of this parameter will not affect the design of the structure. keywords: structural modeling, ibc, multi-story building, response spectrum, moment capacity ≈°üb’g aéæëf’g ωõ©h dƒñæàdg ≈∏y ‹h~dg aéæñdg oƒµd ‹gõdõdg º«ª°üàdg äé«æëæe ägò¨àe òkcéj ¢séëf ~ªfi øh ¥qéw /o áá°°uuóóÿÿgg) ‹h~dg aéæñdg oƒc ™e á≤agƒàÿg ‹gõdõdg º«ª°üàdg äé«æëæe ~«dƒj êéàëj :ibcøe òãc ‘h ¬fg ’eg ,¢vô¨dg gò¡d ω~îà°ùj ägò¨àe º«≤d ôj~≤j ¤g ú©e ™bƒÿ ( áhƒ°ùùg äéä°ûæÿg ô°uéæ©d á«∏ng~dg iƒ≤dg º«b ‘ dƒñàdg ábo ≈∏y ôj~≤àdg gòg aƒ°s òkcéj i~ã ≥∏©àj á°ùøf ¢vôøj …òdg ∫gdƒ°ùdg ¿eéa ,gòd .áb~∏d gô≤àøe ôj~≤àdg gòg ¿ƒµj ¿é«m’g åëñdg èféàf øe ßmƒd ~bh .äéä°ûæª∏d ≈°übc’g aéæëf’g ωõy h ‹gõdõdg º«ª°üàdg äé«æëæe ≈∏y ‹h~dg aéæñdg oƒc ägò¨àe òkcéj ábqƒdg √òg ∞°ûµà°ùjh .‹h~dg aéæñdg oƒc ≈∏y aéæh ᪫b ‘ ò¨àdg ¿éc éªc .iò°ü≤dg á«æeõdg ägîø∏d êô◊g i~ÿg øª°vh äélq~dg áø∏àfl ™bgƒÿ ‹gõdõdg º«ª°üàdg äé«æëæÿ ≈°übc’g ´qé°ùàdg º«b ‘ ò¨àdg ‘ ωé¶àf’g ω~yss ò¨àÿg ᪫≤d òkéj …cg ∂déæg ¢ù«d ¬féa ,∂dòd .´qé°ùà∏d iƒ°ü≤dg ᪫≤dg ≈∏y iòkéj ‘ gô°üà≤essá«fék 0^6 ᪫≤h »°sé°sc’g øeõdg ägp êéñÿg ≈∏y ‹h~dg aéæñdg oƒc øe º«b ™«ª÷ h ™bƒÿg ¢ùøæd ‹gõdõdg º«ª°üàdg äé«æëæÿ ≈°übc’g ´qé°ùàdg á≤£æe ‘ éñjô≤j ≥hé£j ~lh ~bh .èccg hg1s ᪫≤d òkcéj …cg oƒlh ω~y é° jcg úñj ~bh .ss≈∏y álq~dg ™bgƒe aéæãà°scéh aéæëf’g ωõ©d ≈°übc’g ~◊ge .icé°ûæÿg º«ª°üj ≈∏y ôkdƒj ’ ò¨àÿg gòg º««≤j ‘ cé£ÿg ¿cg »æ©j gògh . áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.≈°übc’g aéæëf’g ωõy,‹gõdõdg º«ª°üàdg äé«æëæe ,qghoc’g io~©àÿg êéñÿg ,á«fé°ûfe’g álòªædg : 67 the journal of engineering research vol. 6, no. 1 (2009) 66-75 real-world problems. despite the importance of this issue and despite the widespread publications in the area of seismic analysis of multi-story buildings, like (hooper, 2006), (anagnostopoulos and demopoulos, 2003), (wilkinson and thambiratnam, 2001) and (chopra and goel, 2001) and many others, no published work seems to exist on this issue. this issue, therefore, has been investigated in this paper to determine the sensitivity of the predicted structural response of multi-story buildings to design response spectra parameters specified in the code. ibc 2003 was selected for this study because this code has replaced ubc 1997 (international conference of building officials, 1997), has become popular in usa and many developing countries where it has been used as a base code. to establish the sensitivities as described above, the effects of ibc parameters on the shape of generated design response spectra were determined. it was important to do so because if the design response spectrum is insensitive to the ibc parameters, the predicted internal moments based on the design response spectrum will also be insensitive. therefore, this paper first presents the curves showing the variations in the shapes of design response spectra due to the variations in the ibc parameters. next, the sensitivity of internal forces is generated. only the maximum bending moment is considered. shear forces in the structural members were found to have similar sensitivity and are not included in this paper. the maximum predicted bending moments are evaluated for the case of a simple structure of academic nature and then for three cases of different low rise buildings that are more common for residential purposes. for all the four cases, a series of design response spectra are generated by varying ibc parameters and the sensitivity curves are presented. to determine if a given structural element in a building will be able to resist the bending moment produced by the code-compliant design response spectrum, it is necessary to calculate the moment capacity of the element. since the hand calculations of the moment capacity of a large number of structural members were not practical for obtaining results from the analysis of a number of buildings, an excel spreadsheet program was developed to calculate the moment capacity for any beam or column. the assumptions made to calculate the full ultimate moment capacity were based on the ultimate strength design of reinforced concrete sections in bending. results have been presented graphically to show the sensitivity of the predicted bending moment induced due to the seismic load to the variations in design response spectra. these graphs clearly show similar trend for all cases. these data are helpful to the structural designers in estimating how the design of the structural elements of a building will be affected by changes to the input design response spectra. also, the data will be helpful in improving the code provisions in the future. 2. 3-d structural modeling to produce the sensitivity data as described above, a repeated modeling of three-dimensional structure of multi-story buildings for varying values of ibc parameters is required. the common modeling approach of the available structural analysis software packages requires visualization of the three-dimensional wire frame structural model of a given building before preparing the input data. due to the huge amount of data involved with threedimensional structural models of multi-story buildings, the modeling process using the available software packages becomes laborious, time-consuming, and errorprone. to generate the structural models of multi-story buildings with confidence, easy and error-free modeling is required. for this reason, a specialized pre-processor software package for multi-story buildings called plan23d (engineering optimization software, 2007) was used in conjunction with sap2000 (computers & structures inc., 2005) was used for structural analysis. the philosophy of plan23d is different from all software packages available for structural analysis and modeling. this software offers a conceptually different new approach without a prior visualization of the three-dimensional structural model. it makes all required input data related to the floor plan of the building. therefore, the user's role in identification of elements, joints, bays or columns and labeling them for the sake of preparing the input data is completely eliminated. instead, all input will solely be related to the floor plan. all the required data are input through a few clicks of the mouse pointer. it calculates all the distributed loads acting on the beams due to the slab dead and live loads including the wall loads and generates ibc design response spectra. section specification is quite easy and it also relieves the user from worrying about the local axes for specifying the correct orientation of the columns. in addition, it produces ready-to-go data files for several structural analysis software packages. in this work, using plan23d, ready-to-go input data files were generated for sap2000. the choice of sap2000 was mainly due to its availability. any other structural analysis software package will do the job equally well in conjunction with plan23d. the effectiveness of this approach was validated by performing the structural analysis of multi-story buildings using sap2000 without using plan23d and then repeating the analysis for the same building using sap2000 in conjunction with plan23d. comparing the time, effort and accuracy of results, it was validated and verified that the time and effort required to perform the analysis reduce drastically by orders of magnitudes (nahhas and imam, 2001). 3. ibc design response spectra in this section, ibc provisions for the design response spectra are briefly described because most of the discussion in this paper refers to them. extracting the information presented in this section merely from the reference to the code may not be obvious to all readers. the information in this section however is totally based on ibc 2003 code (international code council, 2003). 68 the journal of engineering research vol. 6, no. 1 (2009) 66-75 3.1 ibc site class a "site class" specified by letter a, b, c, d, e or f is determined on the basis of the shear wave velocity of soil (vs) as described in table 1615.1.1 of ibc 2003 and reproduced here in table 1. standard penetration resistance (n) or the drained shear strength of soil (su) may be used to determine the site class where vs is not known. the information about these tests can be found in the code. 3.2 mapped spectral accelerations maximum considered spectral response acceleration for 5% damping at structural periods of 0.2 and 1.0 seconds are referred to as ss and s1, respectively. figures 1515(1) through 1615(10) of ibc 2003 show these mapped spectral response accelerations for any given geographical location. the maps show the contours of constant values of ss and s1. each contour line is labeled as percentage of g. for example, a value of 150 from the maps will be interpreted as 1.5g. the maps of spectral accelerations given in the code are for usa only. for other countries of the world, agencies and organizations responsible for seismic classification will provide the data. 3.3 site coefficients ibc "site coefficients" fa and fv are given in table 2 and table 3 respectively. for any value of ss that is in between the values of ss for any two columns (shown in the column headings), fa is obtained by linear interpolation. similarly for any value of s1 in between the values given in the column headings, linear interpolation will be required to obtain fv. 3.4 design response spectrum ibc 2003 design response spectrum is shown in fig. 1. instructions for generating this design response spectrum are given in the code. it is re-written below as a computer code procedure: 4. sensitivity of response spectra sensitivity of response spectra to the ibc parameters is assessed by generating three different sets of plots. first for a given value of s1, ss is varied. for each value of ss, response spectra are generated for all five site classes. next, for a given value of ss, s1 is varied. again the response spectra are generated for all the five site classes. these curves show the effect of ibc site class. in a similar manner, the effect of ss is shown for a several values of s1 and the effect of s1 is shown for a several values of ss. a total of eighteen plots, each containing a set of response spectra, were generated for obtaining the sensitivities of response spectra shapes to the ibc parameters. the sensitivities, as observed from these eighteen sets of response spectra, are discussed in the following sections. however, to cut down the size of the paper, only four plots are given here and other similar ones are not included in table 1. ibc site class definitions table 2. ibc site coefficients fa table 3. ibc site coefficients fv 0 2 time period, t (sec) sp ec tra l a cc el ea ra tio n (g ) s ds s d1 s 1tt 0 figure 1. ibc 2003 design response spectrum 69 the journal of engineering research vol. 6, no. 1 (2009) 66-75 this paper. these plots can however be easily generated for any given values of ibc site class, and the mapped accelerations ss and s1 by using the software pplan23d. 4.1 effect of site class the effect of ibc site class on the shape of the design response spectra was investigated by plotting design response spectra for all five site classes for a set of values of ss and s1. only, two of these plots are shown in figs. 2 and 3. the plots were generated in two groups. one group consisted of plots with s1 = 0.8 and ss = 2.5, 2.0, 1.9 and 1.7 and the other group consisted of plots with ss = 1.5 and s1 = 0.7, 0.8, 0.9 and 1.0. from these plots, it is observed that within the critical range of short time periods, design response spectra for class b, c and d have the same peak acceleration and class a design spectrum has always smaller peak acceleration as compared to all other site classes. other than the insensitivity of the design response spectra in the short acceleration region for class b, c and d, the design response spectrum is sensitive to the site class for all sets of values of ss and s1. 4.2 effect of short period ss the effect of ss on the shape of the design response spectra was assessed by plots of sets of design response spectra for a given site class and a given value of s1 with different values of ss. a typical plot of this type, for ibc site class b and s1 = 0.8, is shown in fig. 4. from plots of this type, for all site classes, it becomes obvious that only the peak acceleration is affected by varying the value of ss. other than the peak acceleration, the design response spectra coincide for all values of ss and show insensitivity to the variations in ss. this insensitivity to ss is rather unexpected. thus, for buildings with the fundamental time period of around 0.6 sec, ss is immaterial. also, it was observed from observing all the sets of curves for all site classes that the design response spectra peak acceleration area widens as we move from site class a to site class e. this implies that softer soil will result in larger internal moments and buildings of a wider range of fundamental time period will be affected. 4.3 effect of long period s1 the third type of plots included sets of design response spectra for a given site class and a given value of ss with varying values of s1. sets of such curves were generated for all the five site classes. these plots show the sensitivities of response spectra to s1. a typical plot of this type (for ibc site class d, ss = 2.0) is shown in fig. 5. 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 0 0.5 1 1.5 2 2.5 3 3.5 4 time period, t (sec) sa (g ) class a class b class c class d class e figure 2. sensitivity of ibc site class ( si = 0.8g, ss = 2.5g) 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 0 0.5 1 1.5 2 2.5 3 3.5 4 time period, t (sec) sa (g ) class a class b class c class d class e figure 3. sensitivity of ibc site class (s1 = 0.8g, ss = 1.5g) 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 0 0.5 1 1.5 2 2.5 3 3.5 4 time period, t (sec) sa (g ) ss =1.7 g ss =1.9 g ss = 2.0 g ss =2.5 g figure 4. sensitivity of ss (ibc site class b, s1 = 0.8g) 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 0 1 2 3 4 5 time period, t (sec) sa (g ) s = 1.6 g s = 1.2 g s = 1.0 g s = 0.8 g s = 0.7 g 1 1 1 1 1 figure 5. sensitivity of s1 (ibc site class d, ss = 2.0g) 70 the journal of engineering research vol. 6, no. 1 (2009) 66-75 from plots of this type, for all site classes, it was concluded that for a given site class, the design response spectra always have the same peak acceleration for all values of s1. in other words the value of peak acceleration is insensitive to s1 for a given site class. for time periods beyond the peak accelerations the spectral accelerations become sensitive to variations in s1 but the effect is not so obvious and seems linear. however, it is not possible to make a judgment in this regard from these curves. its effect will become more obvious from the plots showing the sensitivity of maximum moment to s1 presented in the following sections. 5. building models although a number of reinforced concrete buildings were analyzed, only four cases are presented in this paper. the first case is not a real building but was designed as an academic exercise. the three other cases selected from the existing buildings of the third world. the data for these three cases were obtained from the maps of the buildings. these four cases of low rise buildings are studied to evaluate the sensitivity of internal forces to the ibc parameters. they are described in the following paragraphs: 5.1 case1: building a the three dimensional structural model of building a is shown in fig. 6. obviously, this is not a real-world building but simply an academic problem. it was studied and the sensitivities were obtained because some of the characteristics may be easily understood from such a simple building. it is a four story building with two 4x6 meters rooms on each floor. each floor has a height of four meters. all columns and beams have the same section of 20x40 cm. columns oriented so that the larger dimension of the section is along the direction of earthquake excitation which is along y-axis. in addition to the mass of structural members, a total of 1177 kg/m2 of mass were assumed to represent the slab dead and live loads. its fundamental mode of vibration was found to have the time period of 2.48 seconds. table 4 shows time periods and modal participating mass ratios for each of the 10 modes. 5.2 case 2: building b the floor plan for the building b of case 2 is shown in shown in fig. 7. it is a four story concrete building. its data are given in table 5 and its 3-d structural model is shown in fig. 8. table 6 shows the time periods and modal participating mass ratios for each of the first ten modes. 4 m 6 m 6 m 4 m 4 m 4 m 4 m x y z figure 6. structural model for building a mass participation (%) mode period (sec) x-axis y-axis vertical 1 2.481839 0.0000 84.2003 0.0000 2 1.981634 79.5797 0.0000 0.0000 3 1.918419 0.0000 0.0000 0.0000 4 0.787776 0.0000 10.7570 0.0000 5 0.608784 0.0000 0.0000 0.0000 6 0.554694 13.1163 0.0000 0.0000 7 0.449446 0.0000 3.7494 0.0000 8 0.353105 0.0000 0.0049 0.0000 9 0.348743 0.0000 0.0000 0.0000 10 0.329210 0.0000 0.9871 0.0000 sums ---------- 92.6960 99.6986 0.0000 table 4. participating mass ratios for building a figure 7. building b floor plan (dimensions in meters) figure 8. structural model of building b 71 the journal of engineering research vol. 6, no. 1 (2009) 66-75 5.3 case 3: building c the building floor plan for this three-story concrete building is shown in fig. 9 and its 3-d structural model is shown in fig. 10 and the data related to the building is shown in table 7. table 8 shows the time periods and the modal participating mass ratios for the first ten frequencies. 5.4 case 4: building d the building floor plan for this four-story building is shown in fig. 11 and its 3-d structural model is shown in fig. 12. the data related to the building is shown in table 9. table 10 shows modal participating mass ratios for the first 10 frequencies. story height 4m (same of all floor) beam section s b1: 20cm x 90 cm b2: 20cm x 70 cm b3: 20cm x 40 cm b4: 12cm x 60 cm column section s c1: 20cm x 90 cm c2: 20cm x 80 cm c3: 20cm x 70 cm c4: 20cm x 60 cm dead load 977 kg / m 2 live load 200 kg / m 2 excitation direction horizontal (along x) table 5. structural data for building b mass participation (%) mode period (sec) x-axis y-axis vertical 1 2.452523 0.0000 81.3266 0.0000 2 1.511711 0.0004 1.5401 0.0000 3 1.282251 74.9151 0.0000 0.0004 4 0.763149 0.0000 10.9075 0.0000 5 0.498438 0.0001 0.2364 0.0000 6 0.412277 0.0000 4.2719 0.0000 7 0.397150 0.0001 0.0020 0.0000 8 0.316919 0.0010 0.1535 0.0000 9 0.307761 16.5303 0.0000 0.0028 10 0.303859 0.0000 0.0051 0.0000 sums ---------- 91.4471 98.4431 0.0033 table 6. participating mass ratios for building b figure 9. building c floor plan (dimensions in meters) figure 10. structural model of building c story height floor 1: 5m, floors 2 &3: 4m beams section b1: 20x60 column section c1: 20x70, c2: 20x60, c3:20x30 dead load 977 kg / m 2 live load 200 kg / m 2 excitation direction horizontal (along x) table 7. structural data for building c mass participation (%) mode period (sec) x-axis y-axis z-axis 1 1.931639 0.0000 92.6450 0.0000 2 1.288993 23.6038 0.0000 0.0000 3 1.085399 60.7300 0.0000 0.0000 4 0.601114 0.0000 6.4904 0.0000 5 0.386256 0.4550 0.0000 0.0000 6 0.342578 0.0000 0.8182 0.0000 7 0.324441 0.0010 0.0000 0.0000 8 0.304040 11.2648 0.0000 0.0000 9 0.268213 0.0000 0.0022 0.0000 10 0.249552 0.0783 0.0000 0.0000 sums ----------96.1329 99.9558 0.0000 table 8. participating mass ratios for building c figure 11. building d floor plan (dimensions in metres figure 12. structural model of building d 72 the journal of engineering research vol. 6, no. 1 (2009) 66-75 6. discussion of results the sensitivity data was generated for the four buildings as described in the previous section. the results obtained for these four building cases are presented here. similar results, for other building cases, are not presented here due to limitation of number of pages. for each of these four cases, the sensitivity curve was generated by keeping two of the parameters constant and varying the third one. for example, to obtain the site class sensitivity, certain typical values of ss and s1 were chosen and site class was varied. the design response spectrum was obtained for each of the values of the site class and the maximum moment in the structure was obtained from structural analysis based on the ibc compliant design response spectrum using the computer programs plan23d and sap2000. plan23d generated the structural models as well as the ibc compliant design response spectra and produced error-free input data files for the software sap2000 which was used to obtain the internal forces in structural members. for all the four building cases, similar exercises were repeated for obtaining the sensitivity of maximum moment to ss and s1. in the case of ss, a typical value of s1 was selected and curves were generated for maximum moment versus ss for all the five site classes for all four cases. similarly, to obtain the sensitivity of maximum moment to s1, a typical value of ss was selected arbitrarily and again curves were generated for maximum moment versus ss for all the five site classes. the results are discussed in the following sections. 6.1 site class sensitivity sensitivity of maximum moment to the site class is shown in fig. 13 for all the four cases. the figure also shows, for each building, the moment capacity of the structural member that was found to have the maximum predicted bending moment. this makes the data more meaningful and useful for practicing engineers. it is obvious from these figures that ibc site class parameter affects the predicted moment significantly and as mentioned earlier, the predicted maximum bending moments increases for softer soils since ibc site class a represents "hard rock" and class e represents soft soil. thus, due to the high sensitivity of predicted maximum moment to the site class, this parameter must be determined as exactly as possible. here a weakness in the code provisions becomes apparent. the code provides discrete values of the site class and the code has no provision to allow site class as a continuous variable. with the high sensitivity of the predicted moment to this parameter, the discrete values of ibc site class may lead to errors. for example, we see in fig. 12 that the maximum moment will for building a as 71825 kg-m even though the shear wave velocity of the soil is only marginally inside the range for site class e where as if there were a continuous function for the site class, a value slightly higher than 46765 kg-m might have been the predicted moment. this will make the maximum moment grossly over-estimated. and conversely, a wrong estimate of the site class may cause the site class to be d instead of e, resulting in a dangerously low estimate of the maximum moment. 6.2 short period acceleration sensitivity sensitivity of the maximum moment to the short period acceleration ss (ibc parameter) is shown in figs. 14 to 17 for the four cases. it becomes obvious from these figures showing the sensitivity of moments to ss that the maximum moment is insensitive to this important ibc parameter except for the site class e. this implies that any errors in the evaluation of this parameter will not affect the design of the structure. however, the insensitivity of this important parameter may represent an error in the ibc code. story height floor 1:5m, floors 2 -4: 4m beams section b1:20 x 60 column section c1:20 x 70, c2:20 x 60 dead load 977 kg / m 2 live load 200 kg / m 2 excitation direction xdirection (horizontal 1) table 9. structural data for building d mass participation (%) mode period (sec) x-axis y-axis z-axis 1 2.686017 0.0000 90.1560 0.0000 2 1.696465 18.5077 0.0021 0.0000 3 1.536623 63.0913 0.0004 0.0000 4 0.848023 0.0000 7.8133 0.0001 5 0.553572 0.0653 0.0005 0.0000 6 0.503251 0.0012 0.0000 0.0000 7 0.472579 0.0000 1.6927 0.0001 8 0.424656 13.1580 0.0000 0.0000 9 0.362340 0.0359 0.0004 0.0000 10 0.344050 0.1251 0.0000 0.0000 sums ---------- 94.9846 99.6654 0.0002 table 10. participating mass ratios for building d 0 10000 20000 30000 40000 50000 60000 70000 80000 a b c d e moment capacityibc site class m ax im um m om en t ( kg -m ) building a building b building c building d figure 13. ibc site class sensitivity 73 the journal of engineering research vol. 6, no. 1 (2009) 66-75 since the plots show the moment capacity of the structural member subjected to the maximum bending moment by a dotted line, it can be easily seen if the buildings will be safe during an earthquake or not. the plot for building a which does not represent a real building but was designed as an academic exercise clearly indicates that it is safe if built on soil of type ibc site class a, b or c but will be unsafe if built on softer soil that fall in the category of ibc site class d and e. on the other hand all the three real buildings namely buildings b, c and d have predicted bending moments much larger than the moment capacity. this may mean that the selected buildings are not well designed. in fact, it is true that most residential buildings are not designed for seismic loads specially in the third world countries. this conclusion that is eaily seen from the presented curves is important for authorities to take a look and evaluate the safety of residential low rise buildings for earthquakes. another minor observation relates to the sensitivity for site class e as compared to other site classes. it exhibits differently from other classes and indicates that maximum moment has some sensitivity to ss become steeper for higher values of ss. 6.3 long period acceleration sensitivity for the four building cases, sensitivity of the maximum moment to the ibc parameter, namely long period acceleration s1 is shown in figs. 18 to 21. the maximum moment is not insensitive to this ibc parameter. it has quite significant sensitivity. this implies that any errors in the evaluation of this parameter will definitely affect the design of the structure. 10000 20000 30000 40000 50000 60000 70000 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity s building a, s1= 0.8 g figure 14. sensitivity of moment to ss (building a) 10000 60000 110000 160000 210000 260000 310000 360000 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity s building b, s = 0.8 g1 figure 15. sensitivity of moment to ss (building b) 10000 60000 110000 160000 210000 260000 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity s building c, s = 0.8 g1 figure 16. sensitivity of moment to ss (building c) 10000 60000 110000 160000 210000 260000 310000 360000 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity s building d, s = 0.8 g1 figure 17. sensitivity of moment to ss (building d) 10000 20000 30000 40000 50000 60000 70000 80000 90000 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity 1 building a, s = 1.5 g s figure 18. sensitivity of moments to s1 (building a) 74 the journal of engineering research vol. 6, no. 1 (2009) 66-75 it is interesting to note from these sensitivity curves that the relationship between maximum moment and long period acceleration is very close to being linear. this suggests that ibc code may be improved by providing a linear relationship, so that the error in the estimate of the maximum moment due to variation in the value of s1 may be easily estimated. this will help the practicing engineers in exploring the possibilities of designing similar buildings at different geographical locations that have different seismic activity. 7. conclusions the paper presents data giving the sensitivity of structural response to the variations in the ibc-compliant response spectra. from the data presented in this paper, the following conclusions are made: the design spectra for the ibc site class b, c and d have approximately the same peak acceleration for short time periods but class a has a clearly visible lower value. within the critical range of short time periods, class a design spectrum has always smaller peak acceleration as compared to that of class b, c and d. the variation is not uniform. only the peak acceleration is affected by varying the value of ss. thus, for buildings with the fundamental time period of around 0.8 sec, ss is immaterial. the design response spectra are almost identical in the range of short periods for all values of s1. the site class affects the computed moment significantly and thus this parameter must be determined as exactly as possible. however, the code provides discrete values of the site class and the code has no provision that allows site class as a continuous variable. the maximum moment is insensitive to ss except for the site class e. this implies that any error in the evaluation of this parameter will not affect the design of the structure. the maximum moment has a quite significant sensitivity to s1. this implies that any error in the evaluation of this parameter will definitely affect the design of the structure. the relationship between the maximum moment and the long period acceleration is very close to being linear. the data and the results presented in the paper are useful in understanding the impact of ibc code provisions and the output of response spectrum analysis. it also serves as an example and points to a lack of such data in the published literature. the presented technique used in generating the data is also useful from academic point of view and will help the researchers in accomplishing their computational experiments required to understand structural response of multi-story buildings in a manner that is fast, easy and error-free. references anagnostopoulos, s., demopoulos, a., 2003, "preliminary assessment of a simplified code method for earthquake design of asymmetric buildings", proc. concrete structures in seismic regions: fib 2003 symposium. chopra, a. k., goel, r. k, 2001, "a modal pushover analysis procedure for estimating seismic demands for buildings", earthquake engineering & structural dynamics, vol. 31(3), pp. 561 582. computer & structures inc, 2005, "sap2000 software package", http://www.csiberkeley.com, california, usa. 10000 60000 110000 160000 210000 260000 310000 360000 0.6 0.7 0.8 0.9 1 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity 1 building b, s = 1.5 g s figure 19. sensitivity of moment to s1 (building b) 10000 60000 110000 160000 210000 260000 0.6 0.7 0.8 0.9 1 s (g) m ax im um m om en t ( kg -m ) class a class b class c class d class e moment capacity 1 building c, s = 1.5 g s figure 20. sensitivity of moment to s1 (building c) 10000 60000 110000 160000 210000 260000 310000 360000 410000 0.6 0.7 0.8 0.9 1 s (g) maximum moment (kg -m) class a class b class c class d class e moment capacity 1 building d, s = 1.5 g s figure 21. sensitivity of moment to s1 (building d) 75 the journal of engineering research vol. 6, no. 1 (2009) 66-75 engineering optimization software, 2007, "plan23d user's manual", www.planopt.com, georgia, usa. hooper, j. d., 2006, "general seismic design requirements for buildings", proc. 100th anniversary earthquake conference. international code council inc., 2003, “international building code 2003”, falls church, virginia. international conference of building officials, 1997, "uniform building code", whittier, california. nahhas, t. m., imam m. h. 2001, "three-dimensional structural modeling of multi-storey buildings for obtaining moment envelopes", proceedings of eighth international conference on civil and structural engineering computing. wilkinson s, thambiratnam d., 2001, "simplified procedure for seismic analysis of asymmetric buildings", computers and structures, 79:2833-2845. ece060422.qxd the journal of engineering research vol.4, no.1 (2007) 75-81 1. introduction photovoltaic (pv) systems have been used worldwide for the last three decades. their early applications were mainly concentrated in remote areas and in uses were other types of energy are either very expensive or not feasible. however, with the reduction in their fabrication costs, pv systems have seen a tremendous increase in different applications. the maximum efficiency of utilization of solar panels is obtained at maximum power operating point, which is a function of panel physical characteristics and fabrication parameters, solar irradiation, and operating temperature. to design a pv-based system, a model of the panel to be used in simulation prior to implementation is often required. some pv panel equivalent circuit approximations based on a single-diode model _____________________________________ *corresponding author’s e-mail: hadj@squ.edu.om have been studied (ouennoughi, z and cheggar, m., 1999; lee, j.i., brini, j. and dimitriadis, c.a., 1998; araujo, g.l., sanchez, e. and marti, m., 1982; gottschalg, r. et al. 1999; kaminski, a. et al. 1997). whereas, the simple models presented give acceptable results only for single crystalline cells. however, for polycrystalline cells, which are cost effective, the models presented are not accurate enough and hence the extended model of two-diode gives better results (araujo, g.l., sanchez, e. and marti, m., 1982). in addition to this, a precise determination of the internal physical parameters of cells and panels is not always possible and the different errors introduced on these parameters during parameter extraction process induce large errors in the models of solar panels. in this paper, analytical model for equivalent circuit parameters of solar panels and arrays are derived from basic cell models used to build them. the effects of equivalent series and shunt resistances on the panel and array analytical modelling and simulation of photovoltaic panels and arrays h. bourdoucen* and a. gastli electrical and computer engineering department, college of engineering, sultan qaboos university, p.o. box 33, p.c. 123, al-khod, muscat, oman received 22 april 2006; accepted 11 september, 2006 abstract: in this paper, an analytical model for pv panels and arrays based on extracted physical parameters of solar cells is developed. the proposed model has the advantage of simplifying mathematical modelling for different configurations of cells and panels without losing efficiency of pv system operation. the effects of external parameters, mainly temperature and solar irradiance have been considered in the modelling. due to their critical effects on the operation of the panel, effects of series and shunt resistances were also studied. the developed analytical model has been easily implemented, simulated and validated using both spice and matlab packages for different series and parallel configurations of cells and panels. the results obtained with these two programs are in total agreement, which make the proposed model very useful for researchers and designers for quick and accurate sizing of pv panels and arrays. keywords: photovoltaic cell, solar cell, pv system, pv panel, analytical modelling á«fƒ° àdƒødg ±ƒø°üdgh ìgƒdód á«∏«∏ëàdg álòªædgh iécéùg »∏£°ü≤dg ∫oéy ,ø°shoqƒh êém áá°°uuóóÿÿgg§°ùñe »°véjq êpƒ‰ ω~≤j ¬fg ìî≤ÿg êpƒªædg éjgõe øeh .á«°ùª°ûdg éjóî∏d á«©«ñ£dg ⁄é©ÿg ≈∏y é«æñe á«fƒ° àdƒødg ±ƒø°üdgh ìgƒdód é«∏«∏– élpƒ‰ åëñdg gòg ω~≤j : á°sgqo ” éªc .»°ùª°ûdg ´é©°t’gh iqgô◊g á°uénh á«lqéÿg πegƒ©dg òkéj qéñày’g ‘ êpƒªædg ònéjh .»fƒ° àdƒødg ωé¶æ∏d 𫨰ûàdg iaéøc ~≤a ¿h~h áø∏àıg ìgƒ∏dgh éjóÿg ∫éµ°t’ ∫éµ°tg iécéëã ∂dph èféàædg ábo øe ≥≤ëàdgh ádƒ¡°ùh ìî≤ÿg êpƒªædg ≥«ñ£j ”h .á«fƒ° àdƒødg ìgƒd’g πªy ≈∏y álôm ägòkéj øe é¡d éÿ …rgƒàdgh ‹gƒàdg äéehé≤e òkéj g~l g~«øe ìî≤ÿg êpƒªædg π©éj é‡ éeé“ á≤agƒàe ú›éfèdg èféàf âféch .üójéeh ¢ù«ñ°s èegôh ωõm ωg~îà°séh …rgƒàdgh ‹gƒàdg ≈∏y á∏°üàe ìgƒd’gh éjóÿg øe áø∏àfl .á«fƒ° àdƒødg ±ƒø°üdgh ìgƒdód ≥«boh ™jô°s º«ª°üj ≈∏y ∫ƒ°üë∏d úªª°üÿgh úãméñ∏d áá««mmééààøøÿÿgg ääggooôôøøÿÿgggg.á«∏«∏ëàdg álòªædg ,á«fƒ° àdƒødg ìgƒd’g ,á«fƒ° àdƒødg º¶ædg ,á«°ùª°ûdg éjóÿg : 76 the journal of engineering research vol.4, no.1 (2007) 75-81 characteristics are studied as their effects are important for the operation of pv systems. spice and matlab are used for model validation and sensitivity evaluation of the series and shunt resistances. the approach presented is valid for single-diode model as well as for two-diode model provided the specific parameters for a given model are considered, as it will be shown in subsequent section of the paper. 2. some models for solar cells and panels 2.1 analytical model for panel and array solar panels and arrays may be described in terms of a set of electric and optical parameters that represent solar cell properties. this is in addition to the number of cells and panels connected in series and/or in parallel. most of solar cell models available in the literature represent the solar cell by a single diode in parallel with a current source as shown in 1-a (ouennoughi, z. and cheggar, m., 1999; lee, j.i., brini, j. and dimitriadis, c.a., 1998). however, in order to have a more generalized and accurate model, a two-diode equivalent circuit model shown in 1-b has been developed (araujo, g.l., sanchez, e. and marti, m., 1982; gottschalg, r. et al. 1999; kaminski, a. et al. 1997). this model consists of an ideal current source, which represents the optical irradiation connected in parallel with two different diodes d1 and d2 and a shunt resistance rsh. all these elements are connected to a series resistance rs. note that diode d1 models the generated photocurrent in the space charge region, which dominates the total current at low diode voltages whereas diode d2 models the recombination photocurrent outside the space charge region. this photocurrent is more dominant at high diode voltages. the shunt and series resistances are two important parameters in the operation of solar cells. this is because the shunt resistance affects mainly the panel power output and the series resistance affects the efficiency as well as the fill factor (lee, j.i. et al.; mcmahon, t.j. et al.). therefore, the accuracy in determining these two parameters and the knowledge of the different errors involved in their determination is a key point for a better simulation of the panel and array characteristics. note also that the absolute values of rsh are very important in cells qualification testing, module performance testing and failure analysis (mcmahon, et al.). with reference to fig. 1-b, for a generalized panel equivalent circuit, the i-v characteristics of solar cells can be expressed in terms of physical and electrical parameters as, (1) where, iph is the total photogenerated current, il the load current, id1 and id2 the equivalent diode currents and ish the net current through the shunt resistances rsh. for a single solar cell or a single panel that can be also considered as a cell, the currents id1, id2 and ish may be expressed by the following equations (gottschalg, r. et al. 1999). (2) (3) (4) (5) in the above equations, rs and rsh are the series and shunt resistances respectively, isd1 and isd2 are diffusion and saturation currents respectively, n1 and n2 are the diffusion and recombination diode ideality factors, k is the boltzman's constant, q is the electronic charge, t is temperature in kelvin, c0 and c1 are empirical constants modeling the temperature and the irradiation dependence, and g is the irradiation in w/m2. note that in case of a single-diode model shown in fig. 1-a id2 must be removed from eq. (1). a typical connection configuration of cells and panels that form a pv array used for power system applications to feed a resistive load is shown in fig. 2. for this typical array configuration with m horizontal units and n vertical units, the cells or panels connected in series are numbered as p/ci1 to p/cim for a row i, where i varies from 1 to n, whereas, the panels or cells connected in parallel placed in a given column j are numbered as p/c1j to p/cnj where j varies from 1 to m. note that p/c represents a cell or panel unit. rsh id i vish load d l lrsiph a) single-diode model. rshid1 id2 i vish load d1 d2 l l iph rs b) two-diode model figure 1. solar cell models shi2di1diphili −−−= ( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ += 1riv ktn q expii sll 1 1sd1d ( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ += 1riv ktn q expii sll 2 2sd2d sh sll sh r riv i + = ( ) gtcci 10ph ×+= a) single-diode model b) tw-diode model 77 the journal of engineering research vol.4, no.1 (2007) 75-81 as indicated above and has been reported in the literature, the extracted values of both the series resistance rs and the shunt resistance rsh have significant errors (jervase, j., bourdoucen, h. and al-lawati, 2001). it is therefore important to take this in consideration and study the effects of these two parameters on the panel and array characteristics. based on the models developed elsewhere (araujo, g.l. et al.; gottschalg, r. et al.; kaminski, a. et al.) and by taking into account typical changes of rs and rsh from one cell to another, an improved and useful analytical panel model is proposed. the expressions to be used for the model together with eq. (1) are given in subsequent section. typical figures for errors on the rs and rsh have been suggested and their effects are studied by simulation using matlab/simulink and spice software packages. for demonstrating the validity of the analytical model developed, panels having different number of cells and different sizes have been considered. however, typical results will be shown for panel of 72 cells (6x12) with six cells connected in series and twelve in parallel. assuming the basic cell parameters similar except for rs and rsh, and making appropriate change of variables on saturation currents and ideality factors, then expressions of total currents, ideality factors, equivalent resistances rs and rsh can then be formulated using analytical expressions. this is supported by practical considerations where values of these parameters are affected by the way external connections of cells and panels are done. based on models shown in fig. 3 for a set of two ideal cells d1 and d2 connected in series and in parallel, one can write a set of equations for each type of configuration: these are parallel, series and combination of parallel-series connections. parallel connection: with reference to model shown in fig. 3a, one can write i1 = is1 [exp (vpq / (n1kt)) 1] for d1 and d2. assuming d1 and d2 having equivalent characteristics, the parallel equivalent circuit composed of the two diodes can be represented by current source it = 2i1 and diode d as shown on the figure. the i-v expression of this circuit can be expressed as: (6) for m cells connected in parallel, saturation current of the equivalent diode is to be multiplied by m and hence the total current becomes: (7) similar analysis can be done for finding equivalent series and shunt resistances rs and rsh. for the series resistance of a parallel connection of two cells, the current through each resistance is half the total current it , while the voltage drop stays unchanged. thus, equivalent series resistance rse of two cells in parallel circuit equals half the single-cell series resistance rs. hence, for m cells connected in parallel rse = rs / m. the same approach applies for shunt resistances and hence, the equivalent shunt resistance for m cells is rshe = rsh / m. series connection: with reference to circuit models of fig. 3-b, one can write: (8) for d1 and d2. the voltage vs for a single cell can be expressed as: (9) since the two cells connected in series are assumed to have equivalent characteristics, the voltage across the equivalent cell d is 2vs. one needs to multiply the ideality factor n1 in eq. (8) and eq. (9) by 2. for the case of n equivalent cells connected in series, the value of n1 is to be multiplied by n. thus, the equivalent voltage and current for a panel of n cells can be expressed as (10) (11) similar analysis can also be done for finding equivalent series and shunt resistances rs and rsh. the equivalent series resistance of the two cells equals two times the single cell series resistance. hence, for n cells connected in series the total series resistance rse = nrs. for shunt resistance, the same approach applies, and the equivalent shunt resistance rsh for n cells is rshe = nrsh. note that the above approach has also been used to deduce the equivalent panel parameters for a two-cell model. combined series-parallel connections: in case of panel or array with a combined series and parallel connections of many solar cells, a change of variables can be done on the physical parameters of a single cell described by eqs. (1-5) to determine the model of the panel. hence, expressions of the equivalent parameters can be derived as follows. p/c1 1 p/c1 2 p/c1 3 1 mp/c p/c 2 1 p/c2 2 p/c2 3 2 mp/c p/cn 1 p/cn 2 p/cn 3 p/c resitive load il n m figure 2. typical connection of panels and cells forming a pv array that feeds a resistive load. [ ]1))ktn/(qvexp(mimii 1p1s1t −== [ ]1))ktn/(qvexp(ii 1s1s1 −= ) i ii ln( q kt nv 1s 1s1 1s + = ) i ii ln( q kt n.nv 1s 1s1 1t + = [ ]1))ktnn/(qvexp(ii 1t1st −= [ ]1))ktn/(qvexp(i2i 1p1st −= p/c11 p/c12 p/c13 p/c1m p/c21 p/c22 p/c23 p/c2m p/cn1 p/cn2 p/cn3 p/cnm 78 the journal of engineering research vol.4, no.1 (2007) 75-81 (12) where, n1,2e and isd1,2e are respectively the ideality factors and saturation currents for the panel equivalent two diode circuit. thus, eqs. (2-4) can be extended to panels and arrays and expressed as follows, (13) (14) (15) note that since rs and rsh vary also with optical irradiance and ambient temperature, the equivalent array series and shunt resistances rse and rshe will also vary with these two parameters. these are considered in simulations as shown in the following section. a formulation of this dependency for one single cell can be found in (veissid, n., de-andrade, a.m., 1991). to validate the proposed analytical model of panels and arrays to be used for designing pv systems, simulations were done using matlab and spice. the following sections describe how the above models were implemented using these two packages. 2.2 matlab/simulink simulation a matlab/simulink program was used for to validate the developed solar panel model. figure 4 shows the solar panel model as it is implemented with simulink based on previous equations. the inputs to the model are the cell temperature t in ok and the solar irradiation g in w/m2. the output is a voltage that drives a resistive load rl. note that the solar panel current can be varied by changing the load resistance and both voltage and current at the output of the solar panel can be tracked and measured. the model of the panel was based on the implementation of eqs. (1), (5) and (12-15) with the assumption that all cells' parameters are interrelated according to expressions (12). 2.3 spice simulation circuit models for cells' and panels' configurations shown in fig. 2 have been implemented by generating models using the model editor in spice. this has been done for basic solar cells having actual parameters as well as for panels and arrays modeled using the developed analytical approach. the results obtained were compared with the ones obtained using matlab/simulink for similar configurations. these are presented in a subsequent section. it is worth noting that the analytical model presented above is very useful for simulation and modeling of pvbased systems. it simplifies the mathematical computations and design of solar panels without altering physical parameters that are very well known for solar cells. hence, the parameters derived even though purely mathematical give accurate physical behavior of the pv panel and array. 3. results typical parameters of solar cells used to evaluate the developed analytical model are given in (gottschalg, r. et al. 1999). these are c0=2.19x10-3 a.m2/w, c1=0 a.m2/w/k, g=1000w/m2, t=55oc, n1=0.99, n2=1.9, isd1=2.4x10-9 a, isd2=5.5x10-5 a. the series and shunt resistances are kept variable to see their effect on the output power of the simulated arrays. however, their values for the basic cell units are rsh = 200 ω and rs = 2.5x10-2 ω. i1 vp d it d2 vp d1 vp i1 a) vs d1 i1 it d2 i1 vs d 2vs b) figure 3. parallel (a) and series (b) connections of cells, and corresponding equivalent circuits used to build analytical model of panels and arrays ∑ = − ∑ = −∑ = − ∑ = − ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ =⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ = n 1j 1m 1i 1 ijp she n 1j 1m 1i 1 ijs se rr,rr 2e21e1 n.nn,n.nn == 2sdi.me2sdi,1sdi.me1sdi == ( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ += 1riv ktn q expii sell e1 e1sd2d ( ) ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ −⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ += 1riv ktn q expii sell e2 e2sd2d she sell sh r riv i + = 79 the journal of engineering research vol.4, no.1 (2007) 75-81 variations of the output power as a function of the load voltage, for different values of cell and panel equivalent series resistance rs and shunt resistance rsh for the panel of 72 cell configuration are shown in figs. 5 and 6 respectively. the percentage changes in the values of rs (relative to typical value of rs=2.5x10-2 ω) used to obtain these characteristics are 50, 100, 150, and 200%. the values of shunt resistance rsh in percentage relative to typical value of rsh = 200 ω used to obtain these characteristics are also 50, 100, 150, and 200%. note the significant effect of series resistance fluctuations on the output power (refer to fig. 5). however, the shunt resistance has practically an insignificant effect. theses significant fluctuations of the output power versus the output voltage occur at low load resistances as they are more affected by rs rather than rsh which is dominated by the two diodes connect to it in parallel. on the other hand, the main parameters that affect the output current and voltage of the modeled panel have been considered. these are the solar irradiation g of the site and the temperature t of the panel. to illustrate their effects, the circuit model of fig. 4 was simulated with different temperatures and solar irradiations using a 72 cell panel. figure 7 shows the simulation results. notice that, if the temperature of the panel decreases, the output voltage increases. for instance, when the temperature of the panel is 20oc and the irradiation is 1000 w/m2 the open circuit voltage (voc) is about 6.3 v. however, if the temperature is decreased to 0oc and the irradiation is kept constant, the open circuit voltage (voc) becomes equal to about 6.6 v. notice also that reducing the irradiation will result in decreasing the output voltage of the panel. note also that, if the irradiation is decreased from 1000 w/m2 to 500 w/m2, the open circuit voltage is decreased from about 6.3 v to 5.8 v at a constant temperature of 20oc. (a) solar panel simulation block diagram (b) details of solar panel given in a) above 2 m 1 v+ + v + v iph isd rse rshe u[1]-is1*(exp(q/n1/k/u[2]*u[3])-1)-is2*(exp(q/n2/k/u[2]*u[3])-1) isd np*(c0+c1*u[2])*u[1] iph + i+ isignal 2 t 1 g 27+273.15 t g t v+ m solar panel r l 1000 g 74.77 18.69 figure 4. solar panel model implemented with matlab/simulink output voltage, ( ) 0 2.0 4.0 6.0 8.0 o ut pu t p ow er , ( w ) 0 20 40 60 figure 5. output power as a function of load voltage for different values of panel equivalent series resistance rs (upper curve corresponding to lower % change in rs and lowest curve to highest change). the values of rs in % of typical value 2.5x10-2 ω are: 50, 100, 150 and 200 output voltage, (v) 80 the journal of engineering research vol.4, no.1 (2007) 75-81 the results obtained using spice and matlab were in very good agreement. simulation results shown in figs. 8a) and 8-b) for il-vl and pl-vl characteristics confirm this fact. 4. conclusions an analytical model for pv panels and arrays based on extracted physical parameters of solar cells has been presented in this paper. the proposed approach has the advantage of simplifying mathematical modeling of different cells' and panels' configurations without losing necessary accuracy of system operation. the effects of temperature and solar irradiance have been considered in the modeling. the developed analytical model has been simulated and validated using both matlab and spice packages for different cells and panels connected in series and parallel. this makes the proposed model very useful for researchers and systems designers as it allows a quick and accurate sizing of pv panels and arrays. figure 6. output power as a function of load voltage for different values of panel equivalent shunt resistsance rsh. the values of rsh in % of typical value 200 ω are: 50, 100, 150, and 200 figure 7. simulated solar panel i-v characteristics for different temperatures and solar irraditions output voltage (b) p-v characteristics figure 8. comparison of i-v and p-v characteristics obtained by matlab for t = 25oc and g = 1000 w/m2 (72 cells panel) output voltage (v) (a) i-v characteristics a sensitivity analysis of the output power with changing rse and rshe for a panel of 72 cells has been also carried out. errors on rse and rshe of up t o ±40 % of the nominal values of rse and rshe which are respectively 0.05 ω and 400 ω . the results of simulations have shown that deviations of output power due to 40% change in rse and rshe were less that 7% and 1%, respectively. output voltage, (v) output voltage, (v) 81 the journal of engineering research vol.4, no.1 (2007) 75-81 references araújo, g. l., sanchez, e. and m. marti, 1982, "determination of the two-exponential solar cell equation parameters from empirical data," solar cells, 5, pp. 199-204. gottschalg, r., rommel, m., infield, d.g. and. kearney, m.j., 1999, "the influence of the measurement environment on the accuracy of the extraction of the physical parameters of solar cells," journal meas. sci. technol., 10. pp. 797-804. jervase, j., bourdoucen, h. and al-lawati, a., 2001, "solar cell parameter extraction using genetic algorithms," meas. sci. technol., 12, pp. 1922-1925. kaminski, a., marchand, j. j., fave, a. and laugier, a., 1997, "new method of parameters extraction from dark i-v curve," 26th pvsc, pp. 203-205. lee, j. i., brini, j. and dimitriadis, c. a., 1998, "simple parameter extraction method for non-ideal schottky barrier diodes," electronics letters, 34(12), pp. 1268-1269. mcmahon, t. j., bosso, t. s. and rummel, s. r., 1996, "cell shunt resistance and photovoltaic module performance," 25th pvsc, pp. 1291-1294. ouennoughi, z. and cheggar, m., 1999, "a simpler method for extracting solar cell parameters using the conductance method," solid-state electronics, 43, pp. 1985-1988. veissid, n. and de-andrade, a. m., 1991, "the i-v silicon solar cell characteristics parameters temperature dependence, an experimental study using the standard deviation method," 10th psec, portugal, pp. 43-47. comp090520.qxd the journal of engineering research vol. 7, no. 1, (2010) 31-41 _________________________________________________ *corresponding author’s email: manojmaharana@gmail.com a corrective strategy to alleviate overloading in transmission lines based on particle swarm optimization method manoj kumar maharana* and k. shanti swarup department of electrical engineering, indian institute of technology, madras, chennai 600 036, india received 20 may 2009; accepted 23 august 2009 abstract : this paper presents novel corrective control actions to alleviate overloads in transmission lines by the particle swarm optimization (pso) method. generator rescheduling and/or load shedding is performed locally, to restore the system from abnormal to normal operating state. the appropriate identification of generators and load buses to perform the corrective control action is an important task for the operators. a new direct acyclic graph (dag) technique for selection of participating generators and buses with respect to a contingency is presented. the effectiveness of the proposed approach is demonstrated with the help of the ieee 30 bus system. the result shows that the proposed approach is computationally fast, reliable and efficient, in restoring the system to normal state after a contingency with minimal control actions. keywords: direct acyclic graph, participating generators, corrective control strategy, particle swarm optimization, generator rescheduling, load shedding nomenclatures äéäjõ÷g ~°ûb á«∏yéa á≤jôw ≈∏y i~ªà©ÿg π«°uƒàdg •ƒ£n ‘ π≤ãdg ioéjr ∞«øîàd á«ë«ë°üj á«é«jgî°seg ühqgƒ°s »àfé°t .∑ ,éfqé¡e qéeƒc êƒféecg ::áá°°uuóóÿÿgg) äéäjõ÷g ~°ûm á«∏yéa á≤jôw ᣰsgƒh π«°uƒàdg •ƒ£n ‘ ¬° féødg ∫é≤k’g ∞«øîàd »ë«ë°üj ~j~l π©a åëñdg gòg ω~≤jpso∫gõfeg hcg ~dƒÿg â«bƒj √oéyeg â“ ( ᪡e »g √ô£«°ù∏d á«ë«ë°üj ägagôlg ré‚e’ äéméñdg ∫é≤kcg h äg~dƒª∏d ö°séæÿg ¢ü«î°ûàdg .á«©«ñ£dg ádé◊g ¤g á«©«ñw òz πª©dg ádém øe ωé¶ædg ioéy’ é«∏fi õ‚g ~b π≤ãdg ) »μ«∏jôcc’g êé«ñdg §£îª∏d i~j~l ¥ôw .äó¨°ûª∏d g~ldagìî≤ÿg üƒ∏°s’g ih~l .é¡á~≤j ” ~b áfqé£dg á«méædg øe äéméñdg h áªgé°ùÿg äg~dƒÿg qé«àn’ ä~ªàyg ~b ( ) ωé¶f i~yé°ùã é¡ë«°vƒj ” ~b(ieee 30πbcéh áfqé£dg ádé◊g ~©h á«©«ñ£dg ádéë∏d ωé¶ædg √oéy’ aƒøch ¥ƒkƒe h ,é«hé°ùm ™jô°s ƒg ìî≤ÿg üƒ∏°s’g ¿éh èféàædg úñj . .iô£«°s ägagôlg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .π≤ãdg õjõ©j ,√oƒÿg 𫨰ûj ioé©z ,äéäjõ÷g ~°ûm á«∏yéa ,á«ë«ë°üj iô£«°s á«é«jgî°seg ,áªgé°ùe äg~dƒe ,ô°téñe »μ«∏jôccg êé«h §£fl : n : number of particles in the swarm. n : number of elements in a particle. w : initial weight of the particle. t : generation number. rand () : uniform random value in the range [0, 1]. c1, c2 : acceleration constant. vi : voltage of the generator buses. pt i : best position of particle i so far. sij : the mva flow on the line. pi & qi : real and reactive power of bus ‘i’. sf : factor of safety (generally 0.9 to 0.95). ai, bi, ci : are the cost coefficient of generators. ng : number of participating generators. nb : number of participating buses in the group. 32 the journal of engineering research vol. 7, no. 1, (2010) 31-41 ing generators and load buses for generator rescheduling/load shedding. with respect to the contingency, the participating generators are classified into two groups based on the power flow directions. generation in one group of generators is increased while in the other group it is decreased. generators which are contributing to the contigency line (generator flows contributing to the contingency line) to the contingency line identified as the generator decrease (gd) group and the generators which are not contributing to the contingency line are categorized as the generator increase (gi) group. the corrective control strategy is modeled as an optimization problem. from the literature it has been found that the particle swarm optimization (pso) method proposed by kennedy and eberhart (1995) is robust and fast in solving non-linear, non-differentiable problems. the corrective control action for overload-alleviation is a pso-based-generatorrescheduling and/or load-shedding method applied to the gi and gd groups. the proposed corrective control action provides an optimal solution of generator-rescheduling and load-shedding which would bring back the system to normal state. 2. graph theoretic approach in a power system, all generators do not supply power to all loads. the generators which supplying power to a particular load can be identified easily by graph theory. graph theory converts the entire power system into a unidirectional hierarchical structure, based on the power flow contribution from the generators to the loads. graph theory organizes the buses and lines of the network into a homogeneous group according to the concept of 'reach of a generator', 'generator area' and 'links'. the homogeneous group is called the direct acyclic graph (dag) and it is unidirectional in nature. 2.1 directed acyclic graph (dag) a graph is a set of nodes and a set of edges. a cycle is a path with the same node at the beginning and the end. an acyclic graph is a graph with no path that starts and ends at the same node. a directed acyclic graph (dag) contains no cycles; this means that if there is a route from node 'a' to node 'b', then there is no way back. 2.1.1 reach of a generator (rog): the reach of a generator is defined as the set of buses which are reached by power produced by that generator. 1. introduction when a major power system disturbance occurs, protection and control actions are required to stop the power system’s degradation. the aim of corrective control strategy is to avoid cascading outages or system collapse, and to maintain system reliability. the selection of generators and load buses for corrective control action is a crucial task for the system operator. a fast identification of the participating generators, load buses and a proper control action are essential for secure and reliable operation of a power system. under this condition, a minimum number of control actions like rescheduling of generators/load shedding for the participating generators and loads are efficient for the affected power system. in literature many methods for transmission line overload alleviation via corrective strategies have been report ed. alleviation of line overloads by generator rescheduling/load shedding based on redial basic function (rbf) neural network is reported by ram et al. (2007). conjugate gradient technique to minimize the line overloads in conjunction with the local optimization is given by shandilya et al. (1993). there are several publications available that describe direct methods of line overload alleviation using generation rescheduling and load shedding (talukdar et al. (2005); medicheral et al. (1981); chan et al. (1979) and christie et al. (2000). in these methods the system operator has no choice over the selection of the generators or tagged buses to alleviate the overloads. relative-electrical-distance-based (red-based) real power rescheduling for the participating generators to alleviate overload of lines is proposed by yesuratnam et al. (2007), whereas multi-objective-pso-based generator rescheduling/load shedding for alleviation of overload in a transmission network is proposed by hazra et al.; (2007). sensitivity-optimized-based participating generator selection and alleviating overloads by rescheduling generators is proposed by dutta et al.; (2008). in this type of control, all the generators in a system may be divided into two groups, but in practical cases some generators do not supply power to the congested lines. in such cases, all the generators are handled unnecessarily, increasing the complexity of the control strategy. the optimal power flow (opf) technique is the most accurate method for overload alleviation. however, opf calculation is computationally expensive and time-consuming. the main intent of the paper is to propose a technique to identify the participating generators for corrective control actions. in this dag is used to identify the participat nl : number of lines in the group. pgi : real power generated by the generator ‘i’. pd & pl : are the total demand and loss of the system. gij &bij : conductance and susceptance of the line ‘ i to j’. pt i , g : global best position of particle in the population. k1, k2, k3 : the penalty factors (normally large positive real value). 33 the journal of engineering research vol. 7, no. 1, (2010) 31-41 branches of a link are all in the same direction. furthermore, this flow in a link is always from a generator areas of rank n to a generator area of rank m where m is always strictly greater than n. the construction of dag with the help of rog, ga and links will be understood easily by an example given below. for example consider a 6-bus 11-line system. figure 3(a) shows the real power flows from generator to load. by definition the rog of generator g1 encompasses all the buses generator g2 includes buses 2, 3, 4, 5, 6 and the rog of generator g3 includes buses 3, 5 and 6. bus 1, which is supplied by generator g1 only comes under generator area ga1 and the rank is one as only g1 power to it. similarly buses 2 and 4 come under generator area ga2 and the rank is 2 as power is supplied by both generator g1 and g2. buses 3, 5 and 6 come under generator area ga3 and these buses are supplied power from all the three generators, so their rank is 3. lines 1-2 and 1-4 are links between generator areas ga1 and ga2, whereas lines 2-3, 2-5, 2-6 and 4-5 are links between generator area ga2 and ga3. the line 1-5 is the link between generator areas ga1 and ga3. the dag for a six-bus system is shown in fig. 3(b). 3. particle swarm optimization (pso) pso is a simple and efficient population-based optimization method proposed by kennedy and eberhart (1995). pso is motivated by social behavior of organisms such as fish-schooling and bird-flocking. in pso, potential solutions called particles fly around in a multidimensional problem space. a population of particles is called a swarm. each particle in a swarm flies in the search space towards the optimum or a quasi-optimum solution based on its own experience, experience of nearby particles, and power from a generator reaches a particular bus if it is possible to find a path through the network from the generator to the bus for which the direction of travel is always consistent with the direction of the flow as computed by a power flow program or a state estimator proposed by bialek; (1996) and kirschen et al. (1997). for large systems, the reach of a generator (rog) can be determined using the algorithm, explained in the flowchart shown in fig. 1. 2.1.2 generator area (ga): the generator area is defined as a set of contiguous buses supplied by the same generator. unconnected sets of buses supplied by the same generator are treated as separate generator area. a bus therefore belongs to one and only one generator area. the rank of generator area is defined as the number of generators supplying power to the buses. it can never be lower than one or higher than the number of generators in the system. for networks of a more realistic size, the generator area can be determined using the algorithm which is explained in the flowchart shown in fig. 2. 2.1.3 links: having divided the buses into generator areas, each branch is either internal (ie. it connects two buses which are part of the same generator area) or external (ie. it connects two buses which are part of different generator area) to a generator area. one or more external branches connecting the different generator area will be called a link. it is very important to note that the actual flows in all the start count all the generator bus in the system (set k = 1) check for all the connected branches add to rog end no check for last bus is the bus in rog no yes print rog and stop select a generator bus and add to rog no yes check for the flows from generator bus yes no yes no yes check for all the generator bus (k = generator) k = k +1 figure 1. flow chart for reach of a generator (rog) start construct reach of generator (rog) find common bus add the bus to ga supply by the generator as rog end is the bus in ga print ga and stop compare buses of rog (n) with rog (n -i) no yes yes no i = i +1 n= n -1 & n  i figure 2. flowchart for generator generator area (ga) 34 the journal of engineering research vol. 7, no. 1, (2010) 31-41 is very low, the particle may not explore sufficiently, and if vmax is very high, it may oscillate about optimal solution. velocity clamping effect has been introduced to avoid the phenomenon of "swarm explosion". in the proposed method, velocity is controlled within a band as (4) where vmax,t is maximum velocity at generation t, and vmax and vmin are initial and final velocity, respectively. 4. mathematical formulation for corrective control strategy the corrective control strategy by generator rescheduling/load shedding has been divided into two groups (gd and gi) of optimization problems below. 4.1 modeling for generator decrease (gd) group in the generator decrease group, the goal is to reduce the generation with respect to load such that the bus voltage constraints are within the limits. this problem can be solved by classical economic load dispatch with lineflow and voltage limits as constraints. the objective of the constrained economic dispatch problem (ie. voltage and line flow constraints) is to determine the most economic loading of the generators such that the load demand in the gd group are within their limits. the objective is to determine the optimal set of generation pgi (i=1,2, . .ng) so as to minimize the total cost of generation "ft" given by (5) subject to equality constraints (6) (1) (2) the acceleration constant c1 called cognitive parameter pulls each particle towards local best position whereas constant c2 (called social parameter) pulls the particle towards a global best position. usually c1 and c2 range from 0 to 4 given in pso tutorial. [online]. the inertia weight w is an important factor for the pso's convergence. it is used to control the impact of the previous history of velocities on the current velocity. a large inertia weight factor facilitates global exploration (ie. searching of a new area) while a small weight factor facilitates local exploration. therefore, it is wise to choose a large weight factor for initial iterations and gradually reduce the weight factor in successive iterations given by bhaskar et al. (2008). this can be done by using (3) where wmax and wmin are maximum and minimum weight, respectively, iter is iteration number, and itermax is maximum iteration allowed. with no restriction on the maximum velocity (vmax) of the particles, velocity may move towards infinity. if vmax 1 2 3 4 5 6 g1 g2 g3 figure 3. (a) flow diagram of six bus system (b) dag of six bus system global best position among particles in the swarm. let us define a search space s as n-dimension while the swarm consists of n particles. at time t, each particle i has its position defined by 1 2{ , ,...., } i i i i t nx x x x and a velocity defined by 1 2{ , ,...., } i i i i t nv v v v in variable space s. the position and velocity of each particle change with time. the velocity and position of each particle in the next generation (time step) can be calculated as 1 1 , 2 v rand() ( ) rand() ( ) i i i i t t t t i g i t t v w c p x c p x            1, 2,...., 1 1 i i ix x v i n tt t      max min max max w w w w iter iter     max min maxmax, max v v v v iter t iter     2minimize ( ) ( ) 1 ng f x f a p b p ct i gi i gi ii      ( ) 0 ( ) 0 1 ng g x p p pgi ldi       35 the journal of engineering research vol. 7, no. 1, (2010) 31-41 5. algorithm for proposed corrective control strategy to alleviate line overloads due to contingency of one or more lines in the system, the following sequence of control actions is expected from the operator. 1. decrease the bus power injections at the sending end bus of the contingency line. this is incorporated by decreasing the generation at this bus and/or at the buses feeding power to it. 2. maintain the bus power injections constant at the receiving end bus of the contingency line by increasing the generation at this bus and/or at the buses feeding power to it. 3. if the load demand is not met satisfying all the line constraints, then curtail the load at the receiving-end participating loads to which power is being fed from this bus. based on powerflow/state estimation results the pre and post-contingency dag are constructed, and gd and gi groups are identified, by comparing the pre-and-postcontingency generator areas. the generator rescheduling and/or load shedding optimization problems for the gd and gi group are solved by a pso technique. adjustment of generation and loads for the participating generators and load buses obtained from the pso technique are the corrective control actions for the alleviation of overloads. in the gd group, adjust the generation to a load within minimum generation cost, in the gi group, adjust the generation and load such that there are no overloads in any lines in this group. the implementation procedure of the proposed algorithm is given below. step 1: initialize randomly the individual of the population according to the limits of each generating unit (except a slack bus) including individual dimensions, searching points and velocities. the new velocity strategy equation is formulated and the maximum and minimum velocity limits of each individual are calculated using (13) and (14) (13) (14) step 2: compute the slack bus generator vector, losses and line flows using newton-raphson load-flow method for the above generators. inequality constraints (7) fitness function f*t used in pso for this group is formulated including all the constraints as follows. (8) 4.2 modeling for generator increase (gi) group the aim of the generator increase group is to increase the generation within the generator limits so as to meet the demand, and if not possible, switch to load-shedding. as generation increases in this group, there may be an overload in some of the lines. alleviation of overloads in the gi group can be formulated as an optimization problem as follows: the objective function (9) subject to equality constraints: (10) inequality constraints: (11) the fitness function f*t used in pso for this group is formulated as follows: (12) min max min max max ( ) 0 gi gi gi i i i ij ij p p p h x v v v s s           1 2 3 * 2 max 2( ) ( ) 1 1 2( ) nb nllimf f k v v k s st t ij ijli lii j limk p pslack slack           2max( ) 0 ( ) 0 f x s s sfij ijij all      2 [ cos( ) all sin( )] 0 ( ) 0 2 [ cos( ) all sin( )] 0 p g v v v gi ij i j ij i ji j bij i j g x q b v v v bi ij i j ij i ji j gij i j                                   min max min max( ) 0 min max p p pi i i h x q q qi i i v v vi i i              * max 2 2( ) ( )11 1 nl nb limf s s sf k v vt ij ij li lij i        max min max 2 p pd dvd             max min min 2 p pd dvd              where max max 1 n p pid i    and min min 1 n p pid i    i= 1, 2,. . . n (number of generators) and â = 0.01 a smaller value for smooth convergence. 36 the journal of engineering research vol. 7, no. 1, (2010) 31-41 load reduction factor (lrf) = present modified load = (1.lrf x initial mva (18) load at the bus step 13: the latest gbest value generated by the individual is the optimal generation for each unit, which is obtained by satisfying the reduced loads and all constraints in gi the group. 6. simulation and results 6.1 case 1: outage of line 4-12 of ieee 30 bus the buses occupied by the generator areas for the base case and the changes of buses due to outage of lines 4-12 are given in table-1. the base-case power-flow diagram indicating different generator areas is given in fig. 4. the base case direct acyclic graph (dag) acquired from table 1 is shown in fig. 5(a). the bus numbers occupied by the generator areas ga11, ga13 and gacom are altered, there is no change of buses in the generator areas ga1, ga2, ga5, and ga8 after the outage of line 412 as given in table 1. before outage the flows in the lines 4-12 are supplied from ga1 and ga2. the generator areas ga1 and ga2 are considered as gd group where as generator areas ga8, ga11, ga13 and gacom are considered as the gi group as shown in fig. 5(b). the nature of pso convergence characteristics for the gd and gi groups is shown in fig. 6. the convergence time for the gd group varies from 1.93 sec to 2.2 sec and for the gi group it varies from 9.1 sec to 9.8 sec respectively. step 3: to account for slack unit limit violation, and voltage limit violation, the total operating cost is augmented by non-negative penalty terms k1, k2 and k3. calculate augmented cost f*t using (8) for gd group and (12) for gi group. step 4: among the population, the minimum augmented fuel cost value is taken as the best value. the best-augmented fuel cost value in the population is denoted as the gbest. remaining individuals are assigned as the pbest. step 5: modify the velocity v of each individual real power-generating unit pgi using (1). step 6: check the limits on velocity using (15) (15) step 7: modify the member position of each individual pgi using (16). (16) (17) step 9: modified member positions in step 8 are taken as the initial value for the n-r load-flow method. compute slack bus power and line-flows using the n-r load flow method. step 10: calculate the augmented fuel cost using (8) for the gd group and (12) for the gi group. gbest and pbest values are assigned. if the gbest value is better than gbest value in step 4 then current value is set to gbest. if the present pbest value is better than the pbest value in step 4, current value is set to pbest. step 11: in the gd group if the iteration reaches the maximum go to step 13, otherwise go to step 4 and the gbest and pbest values obtained in step4 are replaced by the latest gbest and pbest values acquired in step 10. in the gi group if the iteration reaches the maximum and the solution does not converge, then go to step 12. step 12: reduce the load using the load-reduction factor given in eq. (18) and jump to step 4 after replacing gbest and pbest values with the latest values obtained in step 10. ( 1) ( 1)max maxif , then ( 1) ( 1)min minif , then t tv v v vid d id d t tv v v vid d id d         ( 1) ( ) ( 1)t t tpg pg vid id id     step 8: ( 1)tidpg  must satisfy the capacity limits of the generators and are given by (17) ( 1) ( 1)max maxif pg , then pg ( 1) ( 1)min minif pg , then pg t tpg pgid id id id t tpg pgid id id id         net load at overload bus allowable po wer to the bus total mva load to verify the effectiveness of the proposed corrective control strategy, simulation was carried out on the ieee 30 bus power system. the simulation was done in a 2.66 ghz pentium iv, 512 mb ram personal computer. cost coefficients and mw limits of the generators are given in appendix a. the selection of contingency cases was considered randomly. the upper and lower limits of load bus voltages were taken as 1.05 p.u. and 0.95 p.u. respectively. the generator bus voltages were fixed to their specified values. line loading (mva) limits of 125% of base case were considered. in the pso-based optimization method, a population size of 10 with number of iterations limited to a maximum of 50 was taken. an increases in population size provides better optimal results sacrificing more computation time. pso parameters c1= 2.0, c2= 2.1, wmax = 0.9, wmin= 0.4 were selected from hazra et al.; (2007) and bhaskar et al.; (2008). for each test case, 50 independent trials were carried out. an acceptable variation from 1% to 10% is observed in few worst cases among the 50 trials. 37 the journal of engineering research vol. 7, no. 1, (2010) 31-41 table-2 shows the results of the corrective control strategy of generator rescheduling/load shedding for the outage of line 4-12. it can be observed from table-2 that the line flows after the occurrence of contingency (b) exceeds the mva limits (a). the line overloading is removed by rescheduling generators 1, 2, 8, 11, 13 and a 30.13 mva of load shedding, shared by load buses 12, 14-19, 23 and 24 respectively. the post-contingency flows (c) are within the mva limits (a) after the control strategy as seen from table-2. the pre-contingency and post-contingency voltage magnitudes after corrective control action are given in table 3. from table 3 it is observed that all load bus voltages are with in the voltage limits. sl. no. generator area bus numbers occupied (base case) bus numbers occupied (after outage of line 4-12) 1 ga1 1, 3 1, 3 2 ga2 2, 4 2, 4 3 ga3 6, 7, 8, 27-30 6, 7, 8, 27-30 4 ga4 12-16, 18, 23 13 5 ga5 5 5 6 ga6 9, 10, 11, 2022 9, 10, 11, 1626 7 ga7 17, 19, 24-26 12, 14, 15 table 1. buses occupied by generator areas for ieee 30 bus system 5 16 17 g8 8 g1 1 g5 g2 g1 1 2 3 4 6 7 9 10 11 12 13 14 15 26 27 18 19 20 21 22 23 24 25 28 29 30 g1 3 figure 4. base case flow diagram indicating generator area of ieee 30 bus system figure 5. (a) base case dag for ieee 30 bus system g d g r ou p g i g r o u p g a 1 l ink 1 g a c o m g a 2 g a 5 g a 8 g a 1 3 g a 1 1 l in k 2 l ink 3 l ink 5 l ink 6 l ink 8 l in k 9 figure 5. (b) dag after outage of line 4-12 (a) gd group (b) gi group figure 6. pso convergence characteristics for outage of line 4-12 38 the journal of engineering research vol. 7, no. 1, (2010) 31-41 1. overload condition 2. corrective control strategies generation load line 4-12 out (a) max cap. (b) contin -gency flows pre contingency control action pre contingency control action (c) post cont. flow lines mva mva bus mw mw bus mva mva mva remarks 4-6 6-9 6-10 9-10 12-13 16-17 18-19 19-20 10-20 10-17 22-24 23-24 24-25 25-27 28-27 6-28 48.23 20.86 14.79 41.15 23.38 5.11 3.57 9.09 12.19 8.56 8.98 3.19 1.88 4.32 21.84 16.33 62.68 30.13 20.34 47.67 28.34 9.56 4.49 13.77 16.44 18.92 9.85 5.91 5.46 8.28 22.17 17.03 1* 2* 5 8* 11* 13* 138.69 57.56 24.56 35.00 17.93 16.91 105.58 59.17 24.56 37.12 22.41 18.24 2 3 4 10 12* 14* 15* 16* 17* 18* 19* 20 21 23* 24* 26 25.14 2.68 7.77 6.14 13.48 6.40 8.57 3.94 10.71 3.33 10.09 2.31 20.77 3.57 10.98 4.18 25.14 2.68 7.77 6.14 7.81 3.71 4.69 2.28 6.21 1.93 5.85 2.31 20.77 2.07 6.36 4.18 46.35 14.64 12.09 35.22 20.73 2.17 0.78 6.61 9.05 8.28 8.16 1.12 1.18 4.16 18.17 12.88 30.13 mva load shedding and there are no overload in any lines table 2. simulation results for outage of line 4-12 of ieee 30 bus system the ‘*’ indicates the alteration of generation and loads as corrective control action at that bus pre contingency bus voltage post contingency voltage (after control action) sl. no. voltage (pu) sl. no. voltage (pu) sl. no. voltage (pu) sl. no. voltage (pu) 1 1.050 16 1.047 1 1.050 16 1.046 2 1.045 17 1.043 2 1.045 17 1.048 3 1.023 18 1.031 3 1.039 18 1.046 4 1.017 19 1.028 4 1.036 19 1.047 5 1.010 20 1.033 5 1.010 20 1.050 6 1.014 21 1.036 6 1.032 21 1.049 7 1.004 22 1.036 7 1.015 22 1.049 8 1.010 23 1.030 8 1.035 23 1.042 9 1.053 24 1.024 9 1.045 24 1.039 10 1.048 25 1.019 10 1.044 25 1.037 11 1.082 26 1.002 11 1.082 26 1.019 12 1.060 27 1.025 12 1.048 27 1.043 13 1.071 28 1.009 13 1.071 28 1.029 14 1.045 29 1.005 14 1.049 29 1.024 15 1.041 30 0.993 15 1.048 30 1.013 table 3. pre and post contingency voltage magnitude generator rescheduling (mw) p generator number result reported on dutta et al. (2008) proposed approach 1 -59.0 -55.59 2 19.9 18.70 5 13.0 18.30 8 6.0 4.60 11 6.5 9.20 13 7.0 idle total (mw) 111.4 106.39 table 4. comparison result for congestion management of lines 1-2 and 2-6 39 the journal of engineering research vol. 7, no. 1, (2010) 31-41 participating generators for outage of line 4-12 generator number max (mw) min (mw) best (mw) 1 (gd) 109.56 104.43 107.58 2 (gd) 61.71 57.55 59.77 8 (gi) 39.06 30.78 34.43 11 (gi) 24.89 19.86 22.33 13 (gi) 21.44 16.39 18.69 table 5. generators statistical performance for 50 trials trial number trial number g en er at io n (m w ) g en er at io n (m w ) a) gd group (g1, g2) b) gi group (g8, g11, g13) figure 7. trial by trial statistical performance for outage of line 4-12 trial number trial number a) gd group b) gi group figure 8. statistical performance of convergence time for outage of line 4-12 t im e in s ec t im e in s ec 40 the journal of engineering research vol. 7, no. 1, (2010) 31-41 references abou el-ela, a.a., bishr, m., allam, s. and el-sehiemy, r., 2005, "optimal preventive control actions using multi-objective fuzzy linear programming technique," electric power systems research, vol. 74, pp. 147-155. bhaskar, g. and mohan, m.r., 2008, "security constrained economic load dispatch using improved particle swarm optimization suitable for utility," electric power and energy system, vol. 30, pp. 609-613. bialek, j., 1996, "tracing the flow of electricity," iee proceedings-generation transmission distribution, vol. 143(4), pp. 313 320. chan, s.m. and schweppe, f.c., 1997, "ageneration reallocation and load shedding algorithm," ieee transaction on power apparatus and systems, vol. pas-98(1), pp. 26-34. christie,r.d., wollenberg, b. and wangensteen, i., 2000, "transmission management in the deregulated environment," proceedings of the ieee, vol. 88(2), pp.170-195. dutta, s. and singh, s.p., 2008 , "optimal rescheduling of generators for congestion management based on particle swarm optimization", ieee transactions on power systems, vol. 23(4), pp 1560-1569. hazra, j. and sinha, a. k., 2007, "congestion management using multi-objective particle swarm optimization," ieee transaction on power systems, vol. 22(4), pp. 1726 1734. kennedy, j. and eberhart, r., 1995, "particle swarm optimization," proceeding of ieee, international conference in neural network, perth, austrlia, pp. 1942-1948. kirschen, d., allan, r. and strbac, g., 1997, "contributions of individual generators to loads and flows," ieee transaction on power systems, vol. 12(1), pp. 52-60. medicheral, t.k.p., billinto, r. and sachdev, m.s., 1981, "generation rescheduling and load shedding to alleviate line overloads system studies," ieee transaction on power apparatus and systems, vol. pas-100(1), pp. 36-42. pso tutorial. [online]. available: http://www.swarmintelligence.org/tutorials.php ram, d., srivastava, l., pandit, m. and sharma, j., 2007, "corrective action planning using rbf neural network," applied soft computing, vol. 7, pp. 10551063. shandilya, a., gupta, h. and sharma, j., 1993, "method for generation rescheduling and load shedding to alleviate line overloads using local optimization," iee proceedings generation, transmission and distribution, vol. 140(5), pp. 337-342. talukdar, b.k., sinha, a.k., mukhopadhyay, s. and bose, a., 2005, "a computationally simple method for 6.2 comparison study for overload of line 1-2 and 2-6 in the ieee 30 bus systems the lines 1-2 and 2-6 are overloaded. the generator area for the base case is given in table-1, and the dag is shown in fig. 5a. the overload in the lines 2-6 is due to the load demand in the generator areas 5, 8 and 11. the lines 2-6 receive power from generator areas 1 and 2, so the ga1 and ga2 are declared as the gd group. the power flows through the lines 2-6 to supply the loads in the generator areas 5, 8, 11 and gacom. generator areas 5, 8, 11 and gacom are declared as gi group. the lines 1-2 and 2-6 overloads are removed by rescheduling the generators 1, 2, 5, 8 and 11. the comparison results for overload alleviation by generator rescheduling are given in table 4. before rescheduling lines 1-2 and 2-6 are loaded to 170.04 mva and 66.79 mva with respect to their line limit 130 mva and 65 mva respectively. after rescheduling the power flows on the lines 1-2 and 2-6 are 129 mva and 60 mva reported in dutta et al. (2008) whereas in the proposed approach they are found to be 129.35 mva and 61.37 mva respectively. 6.3 statistical performance analysis the gd and gi group optimizations are carried out for 50 independent trials for the ieee 30 bus system. the maximum and minimum generation output, best output obtained, and the maximum number of times for the gd and gi groups, are given in table 5. the trail-by-trail statistical performance in respect of generation output and convergence times for the gd and gi groups for outage of lines 4-12 of the ieee 30 bus system are shown graphically in figs. 7 and 8 respectively. from the above results, we observe that the proposed method can alleviate the line overloads due to contingency, in any system within minimal control actions, thereby preventing the cascading of outages, leading to blackout or system collapse. 7. conclusions a novel approach to corrective control strategy of generation-rescheduling and/or load shedding with subject to contingencies is presented. identification of effective generator and/or load buses due to a contingency is achieved using a direct acyclic graph (dag). the concept of local optimization is utilized, wherein the implementation of control action becomes easy and effective. this enables the operator to quickly select the appropriate number of buses for a good sub-optimal solution. this task is achieved by means of a particle swarm optimization (pso) method, which provides the best solution with less control decision and actions corresponding to generation and/or load increase/decrease respectively. the solution was sufficient for initiating control actions during an emergency as it protects the system from cascading outages. 41 the journal of engineering research vol. 7, no. 1, (2010) 31-41 cost-efficient generation rescheduling and load shedding for congestion management," electrical power and energy systems, vol. 27, pp. 379-388. yesuratnam, g. and thukaram, d., 2007, "congestion management in open access based on relative electrical distance using voltage stability criteria," electric power system research, vol. 77, pp.1608 1618. mech070529.qxd the journal of enginering research vol. 6, no. 1, (2009) 21-36 1. introduction establishing the design environmental conditions (like design wave height, wind speed, current speed) for different types of marine structures (seawater intake structures, breakwaters, port and harbor structures, shore protection structures, submarine pipelines, open sea loading/unloading terminals, oil terminals and offshore platforms etc.) is ________________________________________ *corresponding author’s e-mail:nsubram@kisr.edu.kw very essential from safety and economic point of view. a lack of correct information on environmental design condition will result either in an unsafe structure or an over designed (and hence uneconomical) structure. therefore it is very essential to predict the design wave heights for different return periods correctly. as on date, in the arabian gulf, most of the coastal structures appear to be over designed, since no systematic extreme wave height prediction is available. recently neelamani, et al. (2006) has carried out the extreme wave analysis, especially for extreme gravity waves in the arabian gulf s. neelamani*, k. al-salem, and k. rakha * coastal and air pollution department, environment and urban development division, kuwait institute for scientific research, p.o. box . 24885, 13109 safat, kuwait received 29 may 2007; accepted 8 september 2007 abstract: the extreme significant wave heights and the corresponding mean wave periods were predicted for return periods of 12, 25, 50, 100 and 200 years for 38 different locations in the territorial and offshore locations of countries surrounding the arabian gulf. the input wave data for the study is hindcast waves obtained using a wam model for a total period of 12 years, (1993 to 2004). the peak over threshold method (with 1.0 m as threshold value), is used for selecting the data for the extreme wave analysis. in general, a weibull distribution is found to fit the data well compared to the gumbel distribution for all these locations. from the joint probability of wave height and wave period, a simple polynomial relationship (tmean = c3 (hs)c4) is used to obtain the relationship between the significant wave height and mean wave period for all the 38 locations. the value of c3 is found to vary from 3.8 to 4.8 and the value of c4 is found to vary from 0.19 to 0.32. the mean wave period was found to be more sensitive to change in locations within the gulf and it is less sensitive to change in return periods from 12 years to 200 years. the significant wave heights for 100 year return period varied from 3.0 to 4.5 for water depths of 9 to 16 m, whereas in the offshore sites (depths from 30 to 60 m) it varied from 5.0 to 7.0 m. a large number of coastal projects are in progress in the arabian gulf and many new projects are being planned in this region for the future. the results of the present study will be highly useful for optimal design of the ocean structures for these projects. keywords: extreme waves, arabian gulf, gumbel and weibull distribution, significant wave height, mean wave period, hindcasted wave data »hô©dg è««∏ÿéh iƒ°ü≤dg êggƒe’g äéyéøjqg énq ωqéc ,⁄é°ùdg ~dén ,êéeó«f ωé«féeg ôhƒ°s áá°°uuóóÿÿggéeéy 12 ωg~îà°sg ” å«m »hô©dg è«∏ÿéh á£fi 38 ~æy ∂dph (éeéy 200 h 100,50 ,25 120 πc √ôe) çh~◊g ä’éªàm’ iƒ°ü≤dg êgƒe’g äéyéøjqg üé°ùm ” : øy é¡yéøjqg ~jõj »àdg êgƒecód iƒ°ü≤dg äéfé«ñdg π«∏ëàd ∂dph »hô©dg è«∏î∏d oé©h’g »féæk »°véjq êpƒ‰ ωg~îà°séh ìéjôdg øe á£ñæà°ùÿg(2004-1993) êgƒe’g äéfé«h øe .~mgh gîe ä’éªàmg ≈∏y √oéªàyg øe ìccg ™bƒÿg ≈∏y oéªày’éh §°sƒàÿg álƒÿg øer úh ábóy •éñæà°sg ” ~bh üé°ùë∏d πñªl á≤jôw øe π° acg πñjh á≤jôw ¿cg á°sgq~dg âæ«h ~bh áñ°ùædéh éecg gîe 16 ¤g 9 øe ≥ª©h √髪∏d gîe 4^5 ¤g 3 øe ìhgîj éeéy 100 πc √ôe çh~m ∫éªàm’ ᫪«ª°üàdg êgƒe’g äéyéøjqg ¿g é° jg á°sgq~dg âæ«h ~bh .çh~◊g .gîe 7 ¤g 5 øe ìhgîàa (gîe60 ¤g 30 øe) á≤«ª©dg ≥wéæª∏d .»hô©dg è«∏î∏d á«∏mé°ùdg á≤£æÿg ‘ é¡d §«£îàdg hg égaé°ûfeg …qél √òãc ™jqé°ûÿ …oé°üàb’g º«ª°üàdg ~j~– ‘ i~«øe ¿ƒµà°s åëñdg gòg èféàf áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.ìéjôdg øe á£ñæà°ùÿg êgƒe’g äéfé«h ,álƒÿg øer ,êgƒe’g äéyéøjqg ,π«∏ëà∏d πñjh h πñªl á≤jôw,»hô©dg è«∏ÿg ,iƒ°ü≤dg êgƒec’g äéyéøjqg : 22 the journal of enginering research vol. 6, no. 1, (2009) 21-36 kuwaiti territorial waters for a total of 19 locations. this study provides an extension of previous study to cover the whole arabian gulf for the benefit of the gcc countries for the safe and cost effective design of coastal and offshore structures. the wave conditions in the arabian gulf are relatively low, since it is basically fetch limited. the arabian gulf, a marginal sea in a typical arid zone, is an arm of the indian ocean. it lies between the latitude of 24-30 degree n. the gulf covers an area of 226,000 square km. it is 990 km long and its width ranges from 56 to 338 km. it has a total volume of 7000 to 8400 km3 of seawater (emery, k.o., 2006; purser, b.h., seibold, e., 1973; elgindy, a., hegazi, m., 1996; and al-yamani, f.y., et al. 2004). the entire basin lies upon the continental shelf. the average water depth of the arabian gulf is about 35.0 m. but depths more than 107 m occur in some places. the gulf's water depth increases in the south east direction. the gulf is connected to the gulf of oman and the arabian sea through the strait of hormuz, which is 56 km wide and with an average water depth of 107 m and allows water exchange between the arabian gulf and arabian sea. in the arabian gulf, in general the dominant wind direction is northwesterly (elshorbagy, et al. 2006). arabian gulf is one of the very active marine areas on the earth. most of the oil produced in the gulf countries is transported through the arabian gulf waters. it is also strategically important area. most of the countries around the arabian gulf rely on the seawater for desalination and for cooling purposes in power plants. a large number of coastal and offshore project activities are going on/being planned in the arabian gulf waters like artificial coastal developmental projects such as palm and world shaped water fronts in dubai, durrat al-bahrain a jewelry shaped water front development in bahrain and similar projects in qatar, ultra modern ports in kuwait, a number of submarine pipeline and offshore oil and gas platforms, projects for development of tourism industries etc. design of all these varieties of marine structures require estimate of design wave height for different return periods, which is not available at present. an attempt is made in the present paper to report the extreme waves in the arabian gulf waters for different return periods. caires and sterl, (2005) have estimated the 100 year return value for significant wave height from the era-40 data for the whole oceans of the earth. the wind data used in this study is obtained from grid of 1.5o x 1.5o. unfortunately this course grid can not provide much information for the countries surrounding the arabian gulf, since the width of the arabian gulf itself is of the order of 1.5o only. in this study wind data on a finer grid size of 0. 5o x 0. 5o was used and the wind speeds are linearly interpolated to the wam model grid size of 0.1o x 0.1o. 2. literature review large number of works were done by many scientists around the world on extreme value prediction of winds and waves. gumbel (1958) is the first who has developed a statistical method for predicting the extreme values of natural random events like wind speed. recorded annual maximum wind speed for as many years as possible, is the input for this method. gumbel's extreme value distribution is widely used by the wind engineering community around the world, since the method is simple and robust. st. denis (1969 and 1973) has discussed gumbel distribution in the context of extreme wave prediction. information related to the collection of data samples for extreme value analysis can be found from nolte (1973), cardone, et al. (1976), petrauskas and aagaard (1971) and jahns and wheeler (1973). details regarding the plotting formula used for the extreme wave predictions are available in kimball (1960), gringorten (1963) and petrauskas and aagaard (1971). the procedure to extreme wave height predictions are explained in sarpkaya and isaacson (1981) and in kamphuis (2000). extreme value analysis for waves is discussed in detail in mathiesen, et al. (1994) goda, et al. (1993) and goda (1992). coles (2001) has provided the statistical details of extreme value prediction based on the annual maximum data points and peak over threshold (pot) method. additional information on pot and its application is provided in ferreira and guedes soares (1998) and leadbetter (1991). all these literatures provide the information and knowledge for carrying out detailed extreme value analysis and are used for the present work. 3. input data generation in this study, the wave data is hindcasted using a wam model for a total period of 12 years, starting from 1st jan 1993 to 31st december 2004. the wam model is a thirdgeneration wind driven wave model. the wam model is a two dimensional model that uses non-stationary and non-homogeneous wind fields in predicting wind waves. the physical processes included in the model are: windwave interaction (generation), wave-wave interaction (quadruplet), dissipation (white capping) and shallowwater bottom dissipation (bottom friction). the output from the wam model is the significant wave height and the mean wave period for every one hour. the data is hindcasted for the whole arabian gulf waters with a grid size of 0.1o x 0.1o. the model was validated using measured data as provided in (al-salem, et al. 2005). the measurements covered a period of about four years where the wave height and period were recoded at a wave buoy and two towers. the model results showed some under prediction in the wave heights for storm events. this was due to the under prediction in the wind speed for such events. using regression analysis, a suitable correction was proposed for waves exceeding 1.0 m rakha, et al. 2007). figure 1 provides a sample of the corrected model results for the significant wave height hs and the mean 23 the journal of enginering research vol. 6, no. 1, (2009) 21-36 wave period tm at the wave buoy. it can be seen that the model predicts hs and tm very well. more results on the model validation can be found in al-salem, et al. (2005) and rakha, et al. (2007). the extreme wave analysis is carried out for a total of 38 different locations in the arabian gulf as shown in fig. 2. each location has a total of 105, 192 hourly wave conditions. the longitude, latitude and the water depth of each location is given in table 1. 0 0.5 1 1.5 2 0 200 400 600 800 1000 1200 1400 1600 1800 h s (m ) buoy wam 0 0.5 1 1.5 2 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 time from jan . 1 1994 (hr) h s (m ) figure 1. comparison between measured and predicted wave conditions (year 1994, at wave buoy form) figure 2. locations in the arabian gulf waters for extreme wave analysis gmt may 23 10:35:42 2006 24 the journal of enginering research vol. 6, no. 1, (2009) 21-36 the maximum and average significant wave heights for these 38 locations based on the 12 year hindcasted data is provided in fig. 3. the highest maximum significant wave height is hindcasted at location 28 (hs = 5.33 m) and the lowest maximum significant wave height is hindcasted at location 8 (hs = 1.82 m). similarly the maximum average wave height for 12 year has occurred at location 27 with hs = 0.77 m and the minimum average wave height has occurred at location 5 with hs = 0.21 m. 4. methodology the gumbel and weibull distribution is used for the extreme value prediction. the input data point selection is done carefully. the statistics of long term prediction of wave requires that the individual data points used in the statistical analysis be statistically independent. hence any hourly wave height depends very much on the wave height of the previous hours and hence the theoretical condition of statistical independence is not met. hence, in order to produce independent data points, only storms should be considered. the commonly used method to separate wave heights into storms is called peak over threshold (pot) analysis (coles 2001). mathiesen, et al. (1994) recommends that the minimum time interval between local maxima be somewhat longer than the time lag for which the auto-correlation function is 0.3 to 0.5. they also recommend that generally a time interval of two to four days suffices. for the present work with 12 years location longitude (oe) latitude (on) water depth (m) remarks/nearest country 1 49.2 29.9 15 iran 2 49.1 28.1 15 saudi arabia 3 49.6 27.5 15 saudi arabia 4 53.4 26.7 61 iran 5 54.9 26.5 11 iran 6 55.9 26.6 31 iran 7 54.0 26.4 55 iran 8 48.7 28.2 9 saudi arabia 9 49.9 26.9 16 saudi arabia 10 50.8 26.4 12 bahrain 11 51.8 25.7 19 qatar 12 51.9 24.4 10 uae 13 52.9 25.2 16 uae 14 54.4 24.7 10 uae 15 55.2 25.3 16 uae 16 55.9 25.9 20 uae 17 50.9 26.2 9 north west of qatar 18 50.6 25.4 17 south west of qatar 19 50.3 25.9 20 in between saudi arabia and bahrain 20 49.3 27.6 9 saudi arabia 21 48.8 29.8 10 in between kuwait -iran 22 49.2 29.2 33 in between saudi arabia -iran 23 49.7 28.6 45 in between saudi arabia -iran 24 50.2 27.9 48 in between saudi arabia -iran 25 50.8 27.3 62 in between bahrain -iran 26 51.5 26.7 39 in between qatar -iran 27 52.2 26.2 44 in between qatar -iran 28 53.2 25.8 54 in between uae -iran 29 54.4 25.8 59 in between uae -iran 30 55.5 26.3 57 in between uae -iran 31 52.4 27.2 79 iran 32 51.6 27.6 22 iran 33 50.9 28.4 42 iran 34 50.4 28.9 44 iran 35 49.9 29.7 24 iran 36 48.7 29.1 19 kuwait 37 52.0 25.3 15 east of qatar 38 53.4 24.9 20 uae table 1. longitude, latitude and local water depth at 38 different locations in the arabian gulf waters 25 the journal of enginering research vol. 6, no. 1, (2009) 21-36 of data, one can hence expect about 1095 to 2190 data points for each location, if the time interval between the data is considered as 4 days and 2 days respectively. it means that for each location, one can expect about 91 and 182 data points per year if the time interval between data is 4 days and 2 days respectively. for the present paper the number of storm events/year (fig. 4) is 100 or less than 100, which means that the average time interval between the data for 1.0 m threshold value is 3 days or more. hence the selection of 1.0 m threshold follows the recommendation of mathiesen , et al. (1994). it is seen that there are 9 locations which has more than 80 no. of storm events/year with threshold significant wave height of 1.0 m. it is also seen that there are 14 locations which have 60 to 80 storm events/year with threshold significant wave heights of 1.0 m. there are six locations amongst the selected 38 locations with less than 40 storm events/year with threshold significant wave height of 1.0 m. these important information are vital for marine operations around these locations. the data points used in the pot analysis are the peaks occurring during each storm with threshold wave height of 1.0 m. the total number of data points used for the extreme wave analysis is hence 12 times the no. of storm events/year with threshold hs value of 1.0 m as provided in the above figure. the data points for each location are arranged in descending order. the probability of exceedence, q is calculated using the formula. maximum and average value of the hindcasted hs for the arabian gulf during 1.1.1993 to 31.12.2004 0 1 2 3 4 5 6 0 4 8 12 16 20 24 28 32 36 40 locations h s (m ) maximum hs average hs figure 3. maximum and average value of the hindcasted significant wave height for arabian gulf waters based on the hindcasted waves from 1.1.1993 to 31.12.2004 no of storm events /year with threshold significant wave height of 1.0 m in the arabian gulf 0 20 40 60 80 100 120 0 4 8 12 16 20 24 28 32 36 40 locations in the arabian gulf n o. o f s to rm e ve nt s/ ye ar w ith th re sh ol d h s of 1 .0 m figure 4. number of storm events/year with threshold significant wave height of 1.0 m in the arabian gulf waters maximum and average value of the hindcasted hs for the arabian gulf during 1.1.1993 to 31.12.2004 no of storm events/year with threshold significant wave height of 1.0 m in the arabian gulf locations in the arabian gulf n o. o f st or m e ve nt s/ ye ar w it h th re sh ol d h s of 1 .0 m 26 the journal of enginering research vol. 6, no. 1, (2009) 21-36 q = (i-c1)/(n+c2) (1) where i : rank n : total number of data points c1=0.44 and c2 = 0.12 for gumbel distribution and c1=0.20+ (0.27/α) and c2 = 0.20 + (0.23/α) for weibull distribution, where α is the shape parameter. the value of α is varied from 0.8 to 1.3 with an increment of 0.05 and the value of α, which gives best fit for the data set is selected. the detailed description of gumbel and weibull distribution can be found from many sources (kamphuis, et al. 2000). for the purpose of quick reference and use for the readers, the salient details of these distributions are provided in annex. a. 4.1 prediction of the wave height for the selected return period the return period, tr and the probability of exceedence are linked by the following expression q = 1 / (l tr) (2) where l is the number of event/year. for the present problem, we know the total number of storm events exceeding threshold value of hs = 1.0 m for each location in the arabian gulf waters. since the data is for a total duration of 12 years, the value of 'l' can be calculated immediately (refer fig. 4). now according to gumbel distribution, the wave height expected for a selected return period htr can be estimated as follows (kamphuis 2000): htr = g -b in[ in (1/p) ] (3) ie. htr = g -b in [ in{(ltr)/(ltr-1)1/α }] (4) according to the weibull distribution, the wave height expected for a selected return period htr can be estimated from the following formula (kamphuis 2000): htr = g + b [ in (1/q)]1/α (5) ie. htr = g + b [ in (ltr) ]1/α ] (6) now it is possible to obtain the extreme wave height for any selected return period. 4.2 joint probability of mean wave periods and significant wave heights wave height and wave periods are independent parameters. however we can see that as wave height increases, it is likely that wave period also increase. on the other hand, the probability of occurrence of high waves and long periods are more pronounced than the probability of occurrence of high waves and short periods. joint probability of wave height and wave period is used for predicting the wave period for a wave height of any desired return periods kamphuis 2000. an example of joint wave period-wave height distribution for location 23 in the arabian gulf waters is given in table. 2. the joint distribution is simplified by relating wave period to wave height via the combinations of greatest frequency. for example, in table 2, interpolation gives tmean = 5.627 sec corresponding to hs = 2.625 m. now the significant wave height and mean wave period is related by the following equation: tmean = c3 (hs)c4 (7) the value of c3 and c4 and the corresponding coefficient of correlation r2 is obtained for all the 38 locations in the arabian gulf waters and is given in figs. 5, 6 and 7 respectively. the figures show that the average coefficient of correlation is 0.948, which is of acceptable value for using c3 and c4 values to obtain the mean wave period for a selected hs value. it is recommended to use the respective c3 and c4 values for the chosen locations for the estimation of tmean. the average value of c3 and c4 is 4.398 and 0.2648 respectively and may be used for finding out the approximate value of the mean wave period for the arabian gulf waters for the selected significant wave heights corresponding to a desired return period of the event. 5. results and discussions in the following are the steps used for the long-term prediction of waves in the arabian gulf: a. the data set for each location is obtained based on peak over threshold value of significant wave height of 1.0 m for all 38 locations for the hindcasted data for the period from 1.1.1993 to 31.12.2004. b. the wave heights obtained at each location are arranged in descending order. c. the plotting formula, as discussed in eq. 1 is used to reduce the wave height data to a set of points describing the probability of exceedence of wave height, q. d. the wave height is then plotted against the reduced variate of gumbel distribution (-in [in (1/p)]) and weibull distribution ([in (1/q)]1/α ). e. a straight line is fitted by using least square techniques through the points to represent a trend. the slope and intercept is obtained. from this, the parameters of the probability distribution are obtained. f. eqs. 4 and 6 are used for predicting wave heights for chosen return period (12 year, 25 year, 50 year, 100 year, 200 year etc.) for gumbel and weibull distribution respectively. a typical gumbel distribution plot for location 23 is provided in fig. 8. the equation of the best line fit and 27 the journal of enginering research vol. 6, no. 1, (2009) 21-36 mean wave period, t mean (sec) hs (m) 0-1.0 1.012.0 2.013.0 3.014.0 4.015.0 5.016.0 6.017.0 7.018.0 8.019.0 total no. of occurr ence mean value (sec) 0-0.25 0 2519 21308 6202 350 35 0 0 0 30414 2.648 0.251-0.5 0 91 7930 21796 1308 102 0 0 0 31227 3.289 0.51-0.75 0 0 42 13349 2028 214 2 0 0 15635 3.654 0.751-1.0 0 0 0 3225 2360 144 1 0 0 5730 3.963 1.01-1.25 0 0 0 1391 3153 135 2 0 0 4681 4.233 1.251-1.5 0 0 0 276 6775 287 13 1 0 7352 4.505 1.51-1.75 0 0 0 6 4465 322 36 4 0 4833 4.583 1.751-2.0 0 0 0 0 1563 1145 26 4 0 2738 4.942 2.01-2.25 0 0 0 0 105 1288 31 2 0 1426 5.451 2.251-2.5 0 0 0 0 3 542 22 5 0 572 5.551 2.51-2.75 0 0 0 0 0 217 25 3 0 245 5.627 2.751-3.0 0 0 0 0 0 72 67 1 0 140 5.993 3.01-3.25 0 0 0 0 0 7 96 1 0 104 6.442 3,251-3.5 0 0 0 0 0 0 28 1 0 29 6.534 3.51-3.75 0 0 0 0 0 0 48 0 0 48 6.5 3.751-4.0 0 0 0 0 0 0 12 1 0 13 6.577 4.01-4.25 0 0 0 0 0 0 5 0 0 5 6.5 4.251-4.5 0 0 0 0 0 0 0 0 0 0 table 2. joint distribution of significant wave height and mean wave period for location 23 in the arabian gulf waters (number of occurrences over 12 years) the value of c 3 at different locations in the arabian gulf 0 1 2 3 4 5 6 0 4 8 12 16 20 24 28 32 36 40 locations c3 figure 5. the value of c3 for different locations in the arabian gulf waters the value of c 4 at different locations in the arabian gulf 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 4 8 12 16 20 24 28 32 36 40 locations c4 figure 6. the value of c4 for different locations in the arabian gulf waters the value of c3 at different locations in the arabian gulf locations the value of the c4 at different locations in the arabian gulf locations c 3 c 4 28 the journal of enginering research vol. 6, no. 1, (2009) 21-36 the correlation coefficient are provided. for the same location, the weibull distribution plot is given in fig. 9 and the equation for the best line fit and the value of correlation coefficient is also given. similar plots are prepared for all the 38 locations. it is found that weibull distribution is better than gumbel distribution for all the 38 locations. the location parameter, scale parameter and the coefficient of regression obtained for all the 38 locations based on the gumbel distribution are given in figs. 10, 11 and 12 respectively. these parameters can be used in eq. 4 in order to obtain the wave heights corresponding to the required return period. the location parameter varies from 1.18 to 1.59 and the scale parameter varies from 0.14 to 0.64. the coefficient of regression for most of the locations is closer to 0.95 (except location 8 where the total number of data points with peak over threshold significant wave height of 1.0 is small compared to other locations) and hence one can have confidence in the best line fit for the data. similarly, the location parameter, scale parameter, shape parameter and the coefficient of regression obtained for all the 38 locations based on the weibull distribution are given in figs. 10, 14 , 15 and 16. these parameters can be used in eqn. 6 to get wave heights corresponding to any required return period. it is seen that the location parameter varies from 0.904 to 1.044, scale parameter from 0.24 to 1.126. the shape parameter for the best line fit is also varies from 0.95 to 1.3. the coefficient of regression for the best line fit for most of the locations fit is closer to 1.0 and is better than the corresponding gumbel distribution fits for different locations. hence it is recommended to use weibull distribution for extreme wave height prediction in the arabian gulf waters. the predicted wave heights for the different locations based on gumbel distribution for return periods of 12 years, 25 years, 50 years, 100 years and 200 years are provided in fig. 17 for quick reference. similar plot based on the weibull distribution is provided in fig. 18. also a plot showing the predicted extreme significant wave height for 100 year return period in the arabian gulf at different locations is given in fig. 19 for quick reference. in general the extreme waves on the territorial waters of kuwait, saudi arabia, bahrain, qatar and uae are smaller compared to the iran's territorial waters and in the arabian gulf midway between the longitudinal boundaries of both the sides. if we consider the whole arabian gulf, then the predicted 100 year significant wave varies from 2.2 m (in the saudi arabian territorial waters) to 7.0 m (midway between uae and iran). even on the longitudinal direction of the arabian gulf along its midway, the 100 year return period ways are of the order of 5 m in the northern part of the gulf and is about 6.0 to 7.0 m in the southern part of the gulf. this could be due to the higher water depths and longer fetch length available for the southern part of the gulf for the north west winds. design of any marine structures in these locations need to consider this points for safe and economic designs. the territorial waters off uae coast, where large number of artificial coastal development projects are going on, the 100 year return period significant waves are of the order of 5.0 to 5.5 m. the complete picture of the predicted extreme waves for different return periods in the arabian gulf can be used for economic and safe design of the projects proposed for the near future and also for assessing the reserve strengths of different ocean structures functioning at present in these waters. collection of more number of reliable past data is very important for prediction of the extreme values with different return periods. simiu, et al. (1978) found that the sampling error in estimating a wind speed with a 50 year return period from 25 years of data, with a 68% confident level is about ±7%. the error in estimating the 1000 year return period value from 25 years of data is calculated to be ±9%. on similar tone, the error in estimating the 200 year and 100 year return period value from 12 years of data would be ±6% and ±4%. the users should consider this uncertainty while using the predicted extreme wave height data for the design purpose. the value of r^^ 2 at different locations in the arabian gulf 0 0.2 0.4 0.6 0.8 1 0 4 8 12 16 20 24 28 32 36 40 locations r^ ^2 figure 7. the value of r2 for different locations in the arabian gulf waters for eqn. 7 the value of r ∧∧ 2 at different locations in the arabian gulf locations 29 the journal of enginering research vol. 6, no. 1, (2009) 21-36 gumbel distribution plot for location 23 in the arabian gulf y = 2.5763x 3.6396 r2 = 0.9788 -4 -2 0 2 4 6 8 10 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 hs (m) -in [i n( 1/ p) ] figure 8. gumbel distribution plot for location 23 in the arabian gulf weibull distribution plot for location 23 in the arabian gulf y = 1.5126x 1.5443 r2 = 0.9875 0 1 2 3 4 5 6 7 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 hs (m) [in (1 /q )]^ ^( 1/ al ph a) figure 9. weibull distribution plot for location 23 in the arabian gulf location parameter for gumbel distribution for different locations in the arabian gulf 0 0.5 1 1.5 2 0 5 10 15 20 25 30 35 40 locations lo ca tio n pa ra m et er figure 10. location parameter based on gumbel distribution for 38 locations in the arabian gulf gumbel distribution plot for location 23 in the arabian gulf hs (m) ln [ ln ( 1/ p )] weibull distribution plot for location 23 in the arabian gulf hs (m) location parameter for gumbel distribution for different locations in the arabian gulf locations l oc at io n pa ra m et er 30 the journal of enginering research vol. 5, no. 1 (2008) 21-36 scale parameter (beta) for gumbel distribution for different locations in the arabian gulf 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 5 10 15 20 25 30 35 40 locations sc al e pa ra m et er figure 11. scale parameter based on gumbel distribution for 38 locations in the arabian gulf coefficient of regression for the best line fit of gumbel distribution for different locations in the arabian gulf 0.86 0.88 0.9 0.92 0.94 0.96 0.98 1 0 10 20 30 40 locations co ef fic ie nt o f re gr es si on figure 12. coefficient of regression for the best line fit for gumbel distribution for 38 locations in the arabian gulf location parameter (gamma) for weibull distribution for different locations in the arabian gulf 0.85 0.9 0.95 1 1.05 1.1 0 4 8 12 16 20 24 28 32 36 40 locations lo ca tio n pa ra m et er figure 13. location parameter based on weibull distribution for 38 locations in the arabian gulf scale parameter (beta) for gumbel distribution for different locations in the arabian gulf locations sc al e pa ra m et er coefficient of regression for the best line fit of gumbel distribution for different locations in the arabian gulf locations c oe ff ic ie nt o f r eg re ss io n location parameter (gamma) for weibull distribution for different locations in the arabian gulf locations l oc at io n pa ra m et er 31 the journal of enginering research vol. 6, no. 1, (2009) 21-36 scale parameter (beta) for weibull distribution for different locations in the arabian gulf 0 0.2 0.4 0.6 0.8 1 1.2 0 4 8 12 16 20 24 28 32 36 40 locations s ca le p ar am et er figure 14. scale parameter based on weibull distribution for 38 locations in the arabian gulf shape parameter (alfa) for weibull distribution for different locations in the arabian gulf 0.8 0.9 1 1.1 1.2 1.3 1.4 0 4 8 12 16 20 24 28 32 36 40 locations s ha pe p ar am et er figure 15. shape parameter based on weibull distribution for 38 locations in the arabian gulf coefficient of regression for the best line fit of weibull distribution for different locations in the arabian gulf 0.88 0.9 0.92 0.94 0.96 0.98 1 0 4 8 12 16 20 24 28 32 36 40 locations c oe ffi ci en t o f r eg re ss io n figure 16. coefficient of regression for the best line fit for weibull distribution for 38 locations in the arabian gulf scale parameter (beta) for weibull distribution for different locations in the arabian gulf locations sc al e pa ra m et er shape parameter (alfa) for weibull distribution for different locations in the arabian gulf locations sh ap e pa ra m et er coefficient of regression for the best line fit of weibull distribution for different locations in the arabian gulf locations c oe ff ic ie nt o f r eg re ss io n 32 the journal of enginering research vol. 6, no. 1, (2009) 21-36 predicted extreme significant w ave heights at different locations in the arabian gulf for different return periods based on gumbel distribution 1 2 3 4 5 6 7 8 0 4 8 12 16 20 24 28 32 36 40 locations in the arabian gulf hs (m ) tr=12 years tr=25 years tr=50 years tr=100 y ears tr=200 y ears figure 17. predicted extreme significant wave heights in the arabian gulf waters for different return periods based on gumbel distribution predicted extreme significant wave heights at different locations in the arabian gulf for different return periods based on weibull distribution 1 2 3 4 5 6 7 8 0 4 8 12 16 20 24 28 32 36 40 locations in the arabian gulf hs (m ) tr=12 years tr=25 years tr=50 years tr=100 years tr=200 years figure 18. predicted extreme significant wave heights in the arabian gulf waters for different return periods based on weibull distribution figure 19. predicted extreme significant wave heights for 100 year return periods in the arabian gulf waters for different return periods based on weibull distribution gmt may 23 10:43:50 2006 predicted extreme significant wave heights at different locations in the arabian gulf for different return periods based on gumbel distribution locations in the arabian gulf h s (m ) predicted extreme significant wave heights in the arabian gulf waters for different return periods based on gumbel distribution locations in the arabian gulf h s (m ) 33 the journal of enginering research vol. 6, no. 1, (2009) 21-36 a typical plot showing the relationship between the significant wave height and mean wave period for location 23 is provided in fig. 20. the best fit polynomial equation and the coefficient of regression is given in the plot. for all the 38 locations in the arabian gulf, similar exercise is carried out to obtain the mean wave periods for different return periods, viz. 12, 25, 50, 100 and 200 years and the mean wave period obtained is provided in fig. 21. it can be seen that the mean wave period range from 4.5 s to 8.1 sec when all the locations and all the return period range of 12 to 200 years are considered together. in fact, one can see that for a selected location, the difference between the mean wave period for 12 year return period and 200 year return period is only of the order of 0.5 sec, where as the location has very significant effect on change of mean wave period. for example, for location 8 (saudi relationship between the mean wave period and significant wave height for location 23 in the arabian gulf tmean = 4.3368(hs)0.2844 r2 = 0.965 0 1 2 3 4 5 6 7 8 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 hs (m) tm ea n (s ) figure 20. relationship between the mean wave period and significant wave height for location 23 in the arabian gulf mean wave periods for the predicted hs values based on weibull distribution for locations in the arabian gulf 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 0 4 8 12 16 20 24 28 32 36 40 locations in the arabian gulf m ea n w av e pe ri od ( s ec ) tr=12 years tr=25 years tr=50 years tr=100 years tr=200 years figure 21. the predicted mean wave period for significant wave height obtained based on weibull for 12, 25, 50, 100 and 200 year return periods relationship between the mean wave period and significant wave height for location 23 in the arabian gulf hs (m) t m ea n (s ) mean wave periods for the predicted hs values based on weibull distribution for locations in the arabian gulf locations in the arabian gulf m ea n w av e pe ri od ( se c) 34 the journal of enginering research vol. 6, no. 1 (2009) 21-36 arabian territorial waters), the mean wave period is in between 4.5 and 5 sec for return periods in the range of 12 to 200 years, whereas for location 28 (offshore in between uae and iran), the mean wave period is in between 7.5 and 8 sec for return periods in the range of 12 to 200 years. this aspect is very important in the design of ocean structures which are very sensitive for wave periods (like the wave transmission characteristics of floating breakwater, which is very sensitive for wave period). for clarity and quick reference, fig. 22 is provided, which shows the mean wave period at different locations in the arabian gulf for 100 year return period event. 6. conclusions gumbel and weibull extreme value distributions are used in order to obtain significant wave heights at 38 different marine locations in the arabian gulf waters. data obtained based on the wam model for 12 years are used. peak over threshold of 1.0 m is used for synthesizing the raw data. based on joint probability analysis, the mean wave period for the significant wave heights of intended return periods were obtained based on polynomial fit between the mean period and significant wave period. in the following are the conclusions obtained out of this investigation. 1. the weibull distribution is very suitable for extreme wave height prediction in the arabian gulf waters. 2. though the arabian gulf covers an area of about 226,000 square km, the extreme significant wave height vary from 2.2 m to 7.0 m for 100 year return period, among these 38 locations. this spatial variation of the wave height must be considered for design of marine structure in these locations. 3. in general the value of extreme significant wave heights are smaller in the territorial waters of kuwait, saudi arabia, bahrain, qatar compared to the territorial waters of uae and iran. 4. the maximum value of the 100 year return period significant wave height is 7.0 m and it is expected to occur in deeper waters in between uae and iran (around location 28). 5. it is found that the mean wave period in the arabian gulf waters is ranging from 4.5 sec to 8.1 sec for the significant wave heights corresponding to 100 year return period. 6. the mean wave periods are more sensitive for special variations rather than return periods. 7. variation of return period from 12 years to 200 years has increased the mean wave period by 0.5 sec only, whereas variation of space from coastal waters of saudi arabia to the offshore waters of uae has changed the 100 year return period mean wave period from 4.7 sec to 8.0 sec. a large number of coastal and offshore projects are in progress and many new projects are planned for the near future in the arabian gulf waters. the results of the present study will be highly useful for optimal design of different types of sea structures in the arabian gulf. acknowledgments the authors wish to express their gratitude to kuwait foundation for advancement of science for partially sponsoring the project, and thanks to kuwait institute for scientific research, kuwait for providing all the infrastructure facilities to carry out this work. figure 22. mean wave period in the arabian gulf for 100 year return periods 35 the journal of enginering research vol. 6, no. 1 (2009) 21-36 references al-salem, k., rakha, k., sulisz, w. and al-nassar, w., 2005, "verification of a wam model for the arabian gulf", arabian coast conf., dubai. uae. al-yamani, f.y., bishop, j., ramadhan, e., al-husaini, m. and a. al-ghadban, 2004, " oceanographic atlas of kuwait's waters," kuwait institute for scientific research, kuwait. p. 203. caires, s. and sterl, a., 2005, "100-year return value estimates for ocean wind speed and significant wave height from the era-40 data", j. of climate, vol. 18, pp. 1032-1048. cardone, v.j., pierson, w.j. and ward, e.g., 1976, "hindcasting the directional spectra of hurricane generated waves", j. of petrol. tech., vol. 28, pp. 385-394. coles, s., 2001, “an introduction to statistical modeling of extreme values”, springer-verlag, p. 208. denis, m. st., 1969, "on wind generated waves", in topics in ocean engineering, ed. c.l. bretschneider, gulf publishing co., texas, vol. i, pp. 37-41. denis, m. st., 1973, "some cautions on the employment of the spectral technique to describe waves of the sea and the response thereto of oceanic systems", proc. offshore technology conference, houston, paper no. otc 1819, pp. 827-837. el-gindy, a. and hegazi, m., 1996, “atlas on hydrographic conditions in the arabian gulf and the upper layer of the gulf of oman”, university of qatar, p. 170. elshorbagy, w., azam, m.h. and taguchi, k., 2006, "hydrodynamic characterization and modeling of the arabian gulf", j. of waterway, port, coastal and ocean engineering, asce, vol. 132(1), pp. 47-56. emery, k.o., 1956, "sediments and water of the persian gulf", bull. amer. ass. petrol. geol., vol. 40(10), pp. 2354-2383. ferreira, j.a. and guedes soares, c., 1998, "an application of the peaks over threshold method to predict extremes of significant wave height", j. offshore mech. arct. eng., vol.120, pp. 165-176. goda, y., hawkes, p., mansard, e., martin, m.j., mathiesen, e., peltier, e., thompson, e. and van vledder, g., 1993, "intercomparison of extremal wave analysis methods using numerically simulated data", proc. 2nd int. symp. on ocean wave measurement and analysis, asce, new orleans, pp. 963-977. goda, y., 1992, "uncertainty of design parameters from viewpoint of extreme statistics", j. of offshore mechanics and arctic engineering., asme, vol. 114, pp. 76-82. gringorten, i.i., 1963, "a plotting rule for extreme probability paper", j. of geophys. res., vol. 68, pp. 813-814. gumbel, e.j., 1958, “statistics of extremes”, columbia university press, new york. jahns, h.o. and wheeler, j.d., 1973, "long-term wave probabilities based on hindcasting of severe storms", j. of petrol. tech., vol.25, pp. 473-486. kamphuis, j.w., 2000, "introduction to coastal engineering and management," advanced series on ocean engineering, vol. 16, world scientific, singapore, chapter. 4. long term wave analysis. pp. 81-102. kimball, b.f., 1960, "on the choice of plotting positions on probability paper", j. of amer. stat. assn., vol. 55, pp. 546-560. leadbetter, m.r., 1991, "on a basis for “peak over threshold" modeling", stat. prob. lett., vol. 12, pp. 357-362. mathiesen, m., hawkes, p., martin, m.j., thompson, e., goda, y., mansard, e., peltier, e. and van vledder, g., 1994, "recommended practice for extreme wave analysis", j. hydraulic research., iahr, vol. 32, pp. 803-814. neelamani, s., al-salem, k. and rakha, k., 2006, "extreme waves for kuwaiti territorial waters", communicated to ocean engineering, pergaman press, uk. nolte, k.g., 1973, "statistical methods for determining extreme sea states", proc. 2nd international conference on port and ocean engineering under arctic conditions, univ. of iceland, pp. 705-742. petrauskas, c. and aagaard, p., 1971, "extrapolation of historical storm data for estimating design wave height", j. of soc. petrol. eng., vol. 11, pp. 23-37. purser, b.h. and seibold, e., 1973, "the principal environmental factors influencing holocene sedimentation and diagenesis in the persian gulf", in: "persian gulf", purser, b.h. (ed), berlin, pp. 1-9. rakha, k.a., al-salem, k. and neelamani, s., 2007, "hydrodynamic atlas for kuwaiti territorial waters", kuwait journal of science and engineering. vol. 34 (1a), pp. 143-156 sarpkaya, t., de st. m. and isaacson, q., 1981, "mechanics of wave forces on offshore structures”, van nostrand reinhold company, new york, usa. simiu, e., bietry, j. and filliben, j.j., 1978, "sampling errors in estimation of extreme winds," j. of the structural division, asce 104, pp. 491-501. 36 the journal of enginering research vol. 6, no. 1, (2009) 21-36 annexure a gumbel distribution the gumbel distribution is given as p = exp [ -exp{-(h-g)/b}] (1) where p = probability of non exceedence (p=1-q) g = location parameter and b = scale parameter. eqn. 1 is linearised by taking logs on both sides twice and it becomes -in[in(1/p)] = (1/ b) h –(g/b) (2) -in[in(1/p)] is called as the reduced variate of gumbel distribution. for the data set described, a least square best line fit is made and the value of b and g is obtained from the slope and intercept of the best line fit. weibull distribution the weibull distribution is a three parameter di stribution and is given as p = 1 – exp[-{(h-g)/b}á] (3) which may also be expressed as q = exp[ -{(h-g)/b}á] (4) where á is the shape parameter, which can be varied from 0.8 to 1.3. for the purpose of linear transformation, take logarithm on both sides and after reorganization, we get [in(1/q)] 1/á = (1/b) h –(g/b) (5) [in(1/q)] 1/á is called as the reduced variate of weibull distribution. again, for the data set described, a least square best line fit is made and the value of b and g is obtained from the slope and intercept of t he best line fit. as discussed, the shape parameter á is varied from 0.8 to 1.3 in step of 0.05 and the best line fit is selected based on the highest value of correlation coefficient. appendix c:\users\raoh\desktop\paper_3.xps the journal of engineering research vol 8 no 2 (2011) 19-31 1. introduction building transient thermal simulation software programs such as trnsys require hourly weather parameters in standard typical meteorological year format (tmy2, for example). each line of the tmy2 record has a total of 70 parameters and indices that need to be specified based on measured data or calculated from measured data or estimated. some of the important elements that normally are not measured are _________________________________________ *correseponding author’s e-mail: alrawahi@squ.edu.om the hourly terrestrial solar radiation fluxes on horizontal surfaces. terrestrial solar radiation; direct beam, diffuse and global hourly solar fluxes are modelled and calculated based on average daily measured data. other data such as dew point and wet bulb temperatures are also calculated based on the measured dry bulb and relative humidity data. the design of building conventional hvac systems requires cooling load calculations which are based on hourly weather data. also, the design of solar cooling systems requires prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman n.z. al-rawahi*, y.h. zurigat and n.a. al-azri department of mechanical and industrial engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, al khoud, muscat, oman received 7 april 2011; accepted 20 september 2011 abstract: in this paper, hourly terrestrial radiation: direct beam, diffuse and global solar radiation are modelled and calculated based on daily measured data for a horizontal surface. in addition, the same parameters were modelled for inclined surfaces. most of the parameters modelled in this work represent a part of the input data required by building thermal simulation and solar energy systems software. important trends of the solar radiation on tilted surfaces as a function of time and direction are being presented and discussed. the comparison of some of the results with measured data from other sources shows good agreement. the effect of tilt angle and orientation on the incident solar radiation fluxes are presented along with optimum surface tilt angles and directions for maximum solar radiation collection in muscat area. the results presented in this paper are quite useful for quick estimation of solar radiation for calculations of cooling load and solar collector performance. also, the models and the computer code developed in this work form the backbone of any computer-aided building thermal design and solar systems design calculations. keywords: solar radiation, solar cooling, solar collectors 20 n.z. al-rawahi, y.h. zurigat and n.a. al-azri cooling load calculations, collector and thermal energy storage sizing, and performance evaluation under transient conditions. thus, the need for accurate predictions of hourly solar flux data for a given location is quite obvious. solar radiation models are used to predict direct, diffused, and global radiation based on certain known variables. the models can be classified into either parametric or decomposition models. the models in the former category require the knowledge of atmospheric conditions (ashrae handbook, 1999; iqbal 1978; gueymard 1993). the decomposition models, on the other hand, use the known solar global radiation data (liu and jordan 1960; lam and li 1996; badescu 2002). a good comparison between the two types of models can be found in (wong and chow 2008) where comparison between representative models from both categories with measured data was presented. decomposition models use solar global radiation data on horizontal surface to predict the direct and diffuse radiation components. the use of these models is supported by the availability of a large number of locations that keep records of the global radiation on a horizontal surface. the decomposition models can be further classified according to the used method as empirical, analytical, numerical and statistical, and neural networks models (liu and jordan 1962; lam and li 1996; tymvios et al. 2005; zervas et al. 2008; ulgen and hepbasli 2009; mellit et al. 2010). a number of models were shown to predict the global solar radiation. however, modelling the diffuse solar radiation has been a challenge for researchers. this difficulty in modelling the diffuse radiation component reflects on the ability to have accurate models for the solar radiation on tilted surfaces which is generally calculated using total, direct, and diffuse solar radiation fluxes. the accuracy of models for solar radiation on tilted surfaces is highly affected by the accuracy of the diffuse solar radiation models. the isotropic model of liu and jordan (1962) and the anisotropic model of perez et al. (1990) are among the well known models in this field and will be used in this work. in their study, (padovan and del col 2010) compared the predictions of a number of models used to predict diffuse and total radiation on tilted surfaces including the models of (liu and jordan 1962; perez et al. 1990). they found that for southern tilted surfaces, the models have similar accuracy but for the western tilted surfaces the anisotropic models, like that of perez et al. (1990), perform better. this is because, in the northern hemisphere unlike the west-tilted surfaces, the south-tilted surfaces receive more direct solar radiation and less diffuse radiation. they found that the error in the diffuse solar radiation increases the error in estimating radiation on tilted surfaces significantly. similar conclusions were reported by noorian et al. (2008). in the presented work, decomposition models for the prediction of hourly solar fluxes on horizontal and inclined surfaces have been implemented with detailed description presented in the next sections. these models and the computer code developed in this work form the backbone of any computer-aided building thermal design and solar systems design calculations. 2. estimation of hourly solar flux compo -nents the total hourly solar radiation falling on a horizontal surface, ith in w.h/m2, could be estimated from the daily global radiation, h (in kj/m2 per day) in accordance with (duffie and beckman 1984) as follows: (1) here 3.6 is a conversion factor (3.6 w.h/m2 = kj/m2), r is the ratio ith / h and is given by (collarespereira and rable 1979): (2) (3) (4) w is the hour angle of the sun (in degrees) given by: (5) h is the time of the day (solar time) in hours and ws is the sunset hour angle (in degrees), given by: (6) is the location latitude angle (in degrees) and is the declination angle of the sun (in degrees) given by: (7) day is the day of the year under consideration (1 to 365). it should be noted that while the equations involving solar radiation terms use solar time, the different climate data obtained from meteorological stations are normally given in standard time (local time in the 21 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman weather station). in addition, the output from the solar radiation calculations needs to be expressed in standard time as it is used in solar thermal systems simulations. it is noted, however, that many published work do not take this time difference into consideration. this could be due to the fact that the difference between the standard time and the solar time is sometimes small. in this work, the difference between solar and standard timings was taken into account. the well established relation between the solar time tsol and standard time tstd is given by (duffie and beckman, 1984): (8) where lst is the standard meridian for the local time zone, lloc is the longitude of the weather station, and the parameter t is the equation of time (in minutes) defined as: (9) where b is given by: (10) the difference between solar time and standard time over the year for seeb/muscat area is shown in fig. 1. although the difference is small reaching a maximum of 24 minutes, the data reported by seeb meteorological center based on local standard time the total solar radiation consists of the direct or beam radiation coming directly from the solar disc and the diffuse component scattered to the ground from the sky dome. the latter depends on the clarity of the sky and could be estimated from the correlation of (collares-pereira and rabl 1979) which gives the daily average diffuse radiation, hd, as: (11) where kt is the clearness index for the day, defined as the ratio of the daily radiation on a horizontal surface to the daily extraterrestrial radiation (ho) on that surface, that is: (12) and (13) isc is the solar constant = 1367 w/m2. the hourly values of the diffuse solar radiation can be estimated from the equation by (liu and jordan 1960) which gives: (14) the direct (beam) component of the solar radiation on a horizontal surface is then obtained from: (15) the hourly value of the direct solar radiation on a ibn (16) where s is the solar zenith angle calculated by (17) alawi (1995) for developing typical solar radiation data for oman. 3. estimation of extraterrestrial radiation components the extraterrestrial radiation on a plane normal to the sun rays is a function of the distance between the day 100 150 200 250 300 350 40050 figure 1. difference between solar time and stand ard time over the year for seeb/muscat area 22 n.z. al-rawahi, y.h. zurigat and n.a. al-azri earth and the sun and is given by duffie and beckman (1984): (18) using the solar zenith angle, the extraterrestrial radiation on a horizontal plane can be calculated from: (19) combining eqs. (17) and (19), we get the extraterrestrial radiation on a horizontal plane at any hour angle, w as: (20) if the total extraterrestrial solar radiation energy incident on a horizontal plane, ieh, over a certain period of time is required, then eq. (20) is to be integrated from hour angles w1 to w2. this results in the following relation: (21) here, ieh is in j/m2. to have ieh in w.h/m2 eq. (21) is to be divided by 3600. on the other hand, the total extraterrestrial radiation on a plane normal to the sun rays over a certain period of time within a certain day is given simply by multiplying eq. (18) by that time period in seconds. this is because the normal total extraterrestrial radiation is constant over the day. 4. estimation of hourly solar flux on inclined surfaces the total radiation received by an inclined surface consists of beam, diffuse, and reflected radiation. the reflected radiation here is the radiation reflected from the surrounding ground. different models for estimating the total solar radiation on inclined surfaces had been proposed and evaluated. such studies include the work of noorian et al. (2008) who evaluated twelve models, the study by nijmeh and mamlook (2000), and that by notton et al. (1996). most of the presented models in these reviews proved to give good results. the majority of these models use similar terms for the beam and reflected radiation while differ in the method of calculating the diffuse radiation portion. the general form of total hourly solar radiation energy incident on an inclined surface iti (in kj/m2), is given by: (22) where itd is the diffuse radiation on the tilted surface, is the angle the surface makes with the horizontal, r is the surrounding diffuse reflectance for the total solar radiation, and rb is the hourly geometric factor given by: (23) here is the incidence angle, the angle between the beam radiation on a surface and the normal to that surface, and it is given by: (24) where is the surface azimuth angle defined as the deviation of the projection on a horizontal plane of the normal to the surface from the local meridian, with zero due south, east negative, and west positive. among the methods available for prediction of radiation fluxes on tilted surfaces is the method of (liu and jordan 1962). this is a simple isotropic model that had been tested by many investigators and found to give good results. in this method the diffuse radiation on the tilted surface is given by: (25) another method for the estimation of the diffuse solar radiation is the anisotropic model of (perez et al. 1990). most of the reviewers agree that this model gives better results (noorian et al. 2008; padovan and del col 2010). the diffuse radiation on the tilted surface in the perez et al. (1990) method is given by: (26) where the coefficients f1 and f2 are given by: (27) where is the sky brightness and f11, f12, f13, f21, f22, and f23 are given in table 1 as functions of the sceh 23 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman sky clearness index, , and the terms a and b are given by: (28) (29) the sky's brightness, gives an indication about the thickness of the clouds and is given by: (30) where m is the relative optical air mass (kasten 1966). the sky clearness index, , represents the transition from totally overcast sky to a low turbidity clear sky. it is given by: (31) it is used to specify the sky clearness categories which are used in table 1. the upper and lower bound for each category can be found in perez et al. (1990). for example, a sky clearness index of = 1 represents an overcast sky, whereas a sky clearness index of = 8 represents a clear sky. 5. comparison with measured data to validate the model presented in this work, a comparison with hourly measured data from different sources has been conducted. figures 2 and 3 show comparisons of predictions of the model developed in this work with data from phoenix, arizona for june 21st and with those from seeb meteorological center, table 1. coefficients used in eq. (27) for different sky clearness, , categories figure 2. comparison between predicted and measured hourly global radiation data of figure 3. comparison between predicted and measured hourly global solar radiation data of seeb/muscat/oman 1200 1000 800 600 400 200 0 5 7 9 11 13 15 17 19 21 hours 1000 900 800 700 800 600 500 400 300 200 100 0 24 n.z. al-rawahi, y.h. zurigat and n.a. al-azri respectively. clearly, the predictions are in good agreement with the measured data. the maximum difference between predicted and measured data does not exceed 7%. this shows that the model is a reliable tool in predicting hourly solar flux data from measured daily averaged data. 6. results based on the above formulations a computer program has been developed to calculate the hourly solar radiation parameters on horizontal and inclined surfaces. the daily average solar radiation flux data of typical meteorological year for seeb/muscat (zurigat et al. 2003) were used as input. sample results that show the effect of tilt angle and orientation on the incident solar radiation are presented. in addition, the optimum surface tilt angles and directions for maximum solar radiation collection in muscat area is deduced. this information is useful when a simple solar collector is to be installed without a tracking system. figure 4 shows the hourly extraterrestrial, global, and diffuse radiation on a horizontal surface on january 21st and june 21st. the extraterrestrial and the global radiation increase from january to june by about 30% and 80%, respectively. the increase in the figure 4. terrestrial and extraterrestrial solar radiation in seeb/muscat area during january 21st and june 21st 25 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman figure 5. sun path diagram for seeb / muscat area figure 6. solar radiation in seeb/muscat area for planes with different tilt angles, , towards south (a), east (b), north (c), and west (d) directions for januatry 21st 26 n.z. al-rawahi, y.h. zurigat and n.a. al-azri diffuse solar radiation is about 95%. these significant increases in the global and diffuse solar radiation indicate the matching between the availability of solar energy and the need for cooling during hot summer months. thus, the potential use of solar thermal cooling where it is needed most is obvious. a number of simulations were carried out to investigate the effect of the tilt angle and the tilt direction. to make sense of the results, the sun path diagram for seeb/muscat area is presented in fig. 5. the diagram shows that the sun is in the southern half of the sky most of the year, especially during the midday period when the solar radiation is at its daily peak. however, the sun is in the northern hemisphere during most of june which represents the hottest month of the year. it can be concluded from the diagram that, for most part of the year, tilting solar collection surfaces towards the south will increase the incident solar radiation on these surfaces. this is clearly illustrated by the results presented in this study. figures 6 and 7 show the hourly solar radiation in seeb/muscat area for a flat surface at different tilt angles towards south, east, north, and west directions during january 21st and june 21st, respectively. with the exception of surfaces tilted towards the south, the solar radiation on horizontal and tilted surface during june 21st (see fig. 7) is greater than that during january 21st. the general trend of the incident solar radiation for surfaces tilted towards the east and the west is almost identical for the two days with higher values for june 21st. on the other hand, as the tilt angle towards the north increases, a higher effect on the hourly solar radiation during january 21st is noticed. it is also noticed that the general trend of the incident solar radiation is almost the same with north-tilted surfaces during january 21st, i.e. the solar radiation increases until it reaches a maximum at noon time then it decreases. while tilting surfaces towards the north has less effect on the solar radiation magnitude, the general trend of the solar changes as the tilt angle increases beyond 60o. for larger tilt angles, the increase of the solar radiation levels off before noon and starts to decrease to have a local minimum point at noon. this is clearly understood by looking at the solar path diagram (see fig. 5). in june, before 12 noon, the sun moves from the northern half to the southern half of the sky which means that the north tilted surfaces will receive less direct solar radiation. the solar radiation on the inclined surface during this period consists of the diffuse solar radiation and the reflected solar radiation only. this shows the importance of having good models for these two components. figure 7. solar radiation in seeb/muscat area for planes with different tilting angles, , towards south (a), east (b), north (c), and west (d) directions for june 21st 27 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman for south-tilted surfaces, figs. 6a and 7a show different trends. compared with june 21st, the solar radiation during january 21st is less sensitive to changes in the tilt angle. the maximum radiation is reached at a tilt angle of 40o at which the received solar radiation increases 40% above that received by a horizontal surface. in june 21st, the received solar radiation decreases significantly as the tilt angle increases. the trends shown in figs. 6 and 7 indicate that strategies to control the amount of received solar radiation on a surface depend on the surface orientation. the combined effect of tilt angle and surface orientation on the total incident solar radiation during the months of january and june is shown in fig. 8. orientation angles, , of 0o, 90o, 180o, and 270o correspond to south, west, north, and east directions, respectively. this figure also shows whether orientations other than the main four (e, w, s, n) have desirable features in terms of received solar radiation. while the figure shows different trends for the two months, there are no special features for other orientations regarding the received solar radiation. during january the minimum received solar radiation is for surfaces tilted towards the north while the maximum received solar radiation is for surfaces tilted towards the south. during the month of june this behaviour is reversed. in january, the maximum received solar radiation is achieved with a tilt angle of about 40o towards the south. during june, the horizontal surface collects more solar radiation than other tilted surfaces figure 8. solar radiation in seeb/muscat area for surfaces of different inclination angles, , as a function of the oriengation angle, , for january (a) and june (b). oriengation angle is measured with 0o due south and positive towards east 28 n.z. al-rawahi, y.h. zurigat and n.a. al-azri figure 9. average monthly incident solar radiation in seeb/muscat area for planes with different tilt angles angles, , toards south (a), east (b), north (c), and west (d) directions, respectively figure 10. average monthly solar radiation in seeb/muscat area for vertical planes with different orientations. orientation angle is measured with 0o due south and positive towards east 29 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman except for surfaces with a tilt angle of 20o towards the north where the received solar radiation is slightly higher. it should be noted that the results presented here assume that the surrounding diffuse reflectance is isotropic, i.e. the same in all directions (dry bare ground of r = 0.2). looking at the broader picture, fig. 9 shows the incident solar radiation on tilted surfaces for the whole year. as noticed previously, surfaces tilted towards the east and the west have similar trends and the trend is almost the same for other tilt angles. for surfaces tilted towards the south, in the period between october to march, the effect of tilt angle on the received solar radiation is relatively small. however, the tilt angle has more significant effect between april and september. for surfaces tilted towards the north the received solar radiation is sensitive to the tilt angle throughout the year. vertical surfaces ( = 90o) require further attention as they represent the major part of the building envelop which contributes significantly to the cooling/heating loads. figure 10 shows the received solar radiation for vertical surfaces at different compass orientations. during summer the westand east-facing vertical surfaces receive higher solar radiation than the southand north-facing ones. these results explain why the orientation of buildings in muscat area should be eastwest with the longest vertical surface facing south. during summer time, this orientation minimizes the solar gain and consequently, the cooling load. it is noticed also that south-facing vertical surfaces receive more solar radiation than north-facing ones for most part of the year and the difference is considerably large. tables 2a and 2b present quite useful information regarding solar radiation in seeb/muscat area. for each month it shows the average solar radiation on horizontal surfaces, the tilt angle that results in maximum received solar radiation and the corresponding solar radiation, the range of tilt angles that will maintain the received solar radiation within 1% from the maximum value, and the solar radiation on vertical surfaces in the four main directions. the information about the received solar radiation on horizontal and vertical surfaces can be quite useful in estimating the building heat gain. also, the information about the maximum received solar radiation and the tilt angle are useful for sun tracking solar collectors that are manually adjusted every month or so. one of the general rules of thumb in solar radiation is that during table 2a. summary of incident solar energy for tilted surfaces for the months of january through june using the model of (perez et al. 1990) table 2b. summary of incident solar energy for tilted surfaces for the months july through december using the model of (perez et al. 1990) 30 n.z. al-rawahi, y.h. zurigat and n.a. al-azri summer the tilt angle that will give the maximum received solar radiation equals the location latitude minus 15o. the presented data is in agreement with this rule since the latitude of seeb/muscat area is 23.35o. another rule of thumb is that during winter time the tilt angle that will give the maximum received solar radiation equals the location latitude plus 15o. this again is in agreement with the presented range of tilt angle for maximum received solar radiation. figure 11 shows the average received solar radiation if the tilt angle is kept the same throughout the year. when the tilt angle of a solar collector is fixed, a tilt angle of about 25o will receive the maximum solar radiation. figure 11 also shows the resulting reduction in the incident solar radiation, in percentage, as a function of tilt angle. for example, using a fixed horizontal surface ( = 90o) will result in about 10% reduction in total solar radiation incident on that surface. conclusions in this paper, prediction of hourly terrestrial solar radiation on a horizontal surface: direct beam, diffuse and global from measured daily averaged global solar radiation on the same surface has been conducted. the same parameters were also modelled for inclined surfaces. thus, an important tool for building and solar thermal systems design, simulations, and performance assessments has been developed. the predicted hourly solar radiation incident on a horizontal surface was compared with hourly data measured locally at seeb meteorological centre as well as elsewhere outside oman. the effect of tilt angle and orientation on the incident solar radiation fluxes are presented along with optimum surface tilt angles and directions for maximum solar radiation collection in muscat area. this information is useful when a simple solar collector is to be installed without a tracking system. in general, surfaces tilted towards the east and the west have similar trends and the trend is almost the same for other tilt angles. for surfaces tilted towards the south, in the period between october to march, the effect of tilt angle on the received solar radiation is relatively small. however, the tilt angle has more significant effect between april and september. for surfaces tilted towards the north the received solar radiation is sensitive to the tilt angle throughout the year. during summer the westand east-facing vertical surfaces receive higher solar radiation than the southand north-facing ones. these results explain why the orientation of buildings in muscat area should be eastwest with the longest vertical surface facing south. in addition, a table that presents important monthly solar radiation information for horizontal, inclined and vertical surfaces has been presented. assuming isotropic reflection of diffuse radiation, the results show that for january the maximum received solar radiation is figure 11. average daily incident solar radiation energy in seeb / muscat area for different tilt angles 31 prediction of hourly solar radiation on horizontal and inclined surfaces for muscat/oman achieved with tilt angle of about 40o towards the south. for june, the horizontal surface collects more solar radiation than other tilted surfaces except for surfaces with a tilt angle of 20o towards the north where the received solar radiation is slightly higher. it is believed that the results presented in tables 2a and 2b are quite useful for quick estimation of solar radiation for calculations of cooling load and solar collector performance. acknowledgments the research leading to these results has received research project funding from the research council of the sultanate of oman, research agreement no. cr squ 010 001. references al-hinai ah, al-alawi sm (1995), typical solar radiation data for oman. applied energy 52:153163 ashrae handbook (1999), hvac applications, atlanta (ga): ashrae. collares-pereira m, rabl a (1979), the average distribution of solar radiation correlations between diffuse and hemispherical and between daily and hourly insolation values. solar energy 22:155164. duffie ja, beckman wa (1980), solar engineering of thermal processes. john wiley, new york. gueymard c (1993), critical analysis and performance assessment of clear-sky solar irradiance models using theoretical and measured data. solar energy 51(2):121-138. iqbal m (1978), estimation of the monthly average of the diffuse component of total insolation on a horizontal surface. solar energy 20(1):101-105. kasten a (1966), a new table and approximate formula for relative optical air mass. arch. meteorol geophys. bioklimatol ser. b 14:206-223. lam jc, li dhw (1996), correlation between global solar radiation and its direct and diffuse components. building and environment 31(6):527-535. liu byh, jordan rc (1962), daily insolation on surfaces tilted towards the equator. trans ashrae 67:526-541. mellit a, eleuch h, benghanem m, elaoun c, massi pavan a (2010), an adaptive model for predicting of global, direct and diffuse hourly solar irradiance. energy conversion and management 51:771-782. nijmeh s, mamlook r (2000), testing of two models for computing global solar radiation on tilted surfaces. renewable energy 20(1):75-81. noorian am, moradi m, kamali ga (2008), evaluation of 12 models to estimate hourly diffuse irradiation on inclined surfaces. renewable energy 33(6):1406-1412. notton g, muselli m, louche a (1996), two estimation methods for monthly mean hourly total irradiation on tilted surfaces from monthly mean daily horizontal irradiation from solar radiation data of ajaccio, corsica. solar energy 57(2):141-153. padovan a, del col d (2010), measurement and modeling of solar irradiance components on horizontal and tilted planes. solar energy 84:20682084. perez r, ineichen p, seals r, michalsky j, stewart r (1990), modelling daylight availability and irradiance components from direct and global irradiance. solar energy 44:271-289. tymvios fs, jacovides cp, michaelides sc, scouteli c (2005), comparative study of angstrom's and artificial neural networks methodologies in estimating global solar radiation. solar energy 78:752-762. ulgen k, hepbasli a (2009), diffuse solar radiation estimation models for turkey's big cities. energy conversion and management 50:149-156. zervas pl, sarimveis h, palyvos ja, markatos ncg (2008), prediction of daily global solar irradiance on horizontal surfaces based on neural-network techniques. renewable energy 33:1796-1803. zurigat yh, sawaqed n, al-hinai h, jubran b (2003), development of typical meteorological years for different climatic regions in oman, final report, petroleum development oman. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 17, no. 2, (2020) 59-74 *corresponding author’s e-mail: bkareem@futa.edu.ng doi:10.24200/tjer.vol17iss2pp59-74 development of oee error-proof (oee-ep) model for production process improvement b. kareem a, * , a.s. alabi a , t.i. ogedengbea, b.o. akinnuli a , o.a. aderoba b , and m.o. idris c amechanical engineering, federal university of technology, akure, nigeria b mechatronics engineering, federal university of oye-ekiti, ikole campus, nigeria c mechanical engineering, osun state university, osogbo campus, nigeria abstract: the global demand for effective utilization of both humans and machinery is increasing due to wastage incurred during product manufacturing. excessive waste generation has made entrepreneurs find it difficult to breakeven. the development of dynamic error-proof overall equipment effectiveness (oee) model for optimizing a complex production process is targeted at minimizing/eradicating operational wastes/losses. in this study, the error-proof sigma metric was integrated into the extended traditional oee factors (availability, performance, quality) to include losses due to waste and man-machine relationships. error-proof sigma statistics enabled continuous corrective measures on unsatisfactory or low-level oee resulted from process output variations (quantity delivered or expected), which were mapped into sigma statistical standards (oneto sixsigma). application of the model in a processing company showed that errors of the process were reduced by 78% and 42% respectively for traditional oee and the new error-proof oee (oee-ep). the results revealed that the oee-ep model is better than the other existing schemes in terms of losses elimination in the production process. keywords: oee dynamism; sigma metric; process integration; productivity. تطویر نموذج تجنب األخطاء الدینامیكیة المتعلقة بالفعالیة العامة للمعدات لتحسین عملیة اإلنتاج دیروبا و م. إدریسأو. ،كینوليأب. ،ت. أوجینجبي ،البيأ. أ ،ب. كریم یتزاید الطلب العالمي على االستخدام الفعال للقوى العاملة واآلالت تجنباً للھدر الذي یحدث أثناء عملیة التصنیع حیث :الملخص تجنب األخطاء الدینامیكیة یستھدف تطویر نموذج .یصعب التولید المفرط للنفایات من عملیة تحقیق نتائج تصنیعیة مرضیة المتعلقة بالفعالیة العامة للمعدات تحقیق االستفادة المثلى من عملیة اإلنتاج المعقدة عن طریق تقلیل النفایات و الخسائر التشغیلیة العامة للمعدات استخدم في ھذه الدراسة قیاس سیجما المقاوم للخطأ الختبار العوامل المؤثرة في نظام الفعالیة .إلى أدنى حد ممكن (التوفر واألداء والجودة) لتضمین الخسائر الناجمة عن النفایات والعالقة بین اإلنسان واآللة. أتاحت إحصائیات سیجما المتعلقة بتجنب األخطاء اتخاذ إجراءات تصحیحیة مستمرة على مستوى غیر ُمرٍض أو منخفض من نظام الفعالیة العامة للمعدات نتیجة سیجما). 6-1ي ناتج العملیة (الكمیة التي تم تسلیمھا أو المتوقعة)، والتي تم تعیینھا إلى معاییر إحصائیة للسیجما (لالختالفات ف % على التوالي بالنسبة 42% و78أن أخطاء العملیة انخفضت بنسبة المعالجة وقد أظھر تطبیق النموذج في إحدى الشركات .و النموذج الجدید لنظام تجنب األخطاء الدینامیكیة المتعلقة بالفعالیة العامة للمعداتلنظام الفعالیة العامة للمعدات التقلیدي أفضل من المخططات األخرى من حیث وكشفت النتائج أن نموذج تجنب األخطاء الدینامیكیة المتعلقة بالفعالیة العامة للمعدات .التخلص من الخسائر في عملیة اإلنتاج .اإلنتاجیة؛ تكامل العملیة الفعالیة العامة للمعدات؛ مقیاس سیجما؛امیكیة دین :المفتاحیة الكلمات development of oee error-proof (oee-ep) model for production process improvement 60 nomenclature g quality products delivered per unit time (year) pg total quantity produced per unit time (year) wg actual waste generated (%) i counter for overall equipment effectiveness factor j counter for sigma value n traditional scheme ( n =3), new scheme ( n =5) oee overall equipment effectiveness coee effectiveness improvement factor (%) wp planned (expected) waste (%) 1t actual production volume per unit time (year) 2t planned production volume per unit time (year) 0t actual system performance per unit time (year) nt system performance expected per unit time (year) at actual human productivity per unit time (year) st expected human productivity per unit time (year) x time (year) iy equipment effectiveness factor at a given year, x 'yi contribution of equip. effectiveness factors (%) β performance efficiency of equipment (%) σ improvement (error-proof) factor α availability efficiency of production equip. (%) μ quality rate (efficiency) of products (output) (%) ω waste generation rate (efficiency) of equip. (%) γ human/ergonomics-equipment efficiency (%) ep error-proof mse mean square error 1. introduction over the decades, manufacturing industries and organizations concentrated mostly on mass production of goods without paying attention to how best the overall equipment effectiveness (oee) measures can be integrated into the system to enhance productivity (muchiri and pintelon, 2008; dilworth, 2013). in the recent past, there had been an attitudinal change of corporate managers towards integrating oee measures into manufacturing systems (dilworth, 2013; martand, 2014; prinz, 2017). the high cost of operations maintenance was among the reasons responsible for the change (gharbi and kennen, 2000). maintenance in this context is defined as a combination of all technical and associated administrative activities required to keep equipment, installations, and other physical assets in good working condition or restore them to their original condition (mwanza, 2017). industrial managers realized that there is a possibility of achieving significant savings in operation costs under proper oee practice (godfrey, 2002; bruce, 2006). oee measured under resource availability (kareem and jewo, 2015) can play a vital role in the performance improvement of industrial operations (kadiri, 2000; butlewski et al., 2018). the productivity improvements achieved so far at the industrial operations level from the past studies have been significant (munoz-villamizar et al., 2018), but insufficient because of emerging challenges of the working environment and waste generation that the organizations need to address. for the enhancement of sustainable productivity, factors of effectiveness measures (equipment availability, quality of products and plant performance, etc.) should be considered at the design stage of the production process. overall equipment effectiveness can be sustainable under regular assessment of and improvement on those effectiveness factors (joshi and gupta, 1986). industrial sectors, in compliance with the twenty-firstcentury development goal, are now moving from the traditional method of measuring productivity to the modern method where conducive working environments and waste elimination strategies are being considered important to maximize profit. wastes are products of material, man, machine, methods, among other production process factors (ghazali et al. 2013). on this basis, value stream mapping (vsm) of processes is very important to find out production processes that required improvement or motivation (meyer and stewart, 2018). therefore, the traditional oee model needs to be improved upon by considering other emerging factors that influence productivity in modern industrial operations to survive the competition. other challenges such as productivity monitoring, continuous improvement measures, customers’ satisfaction, and environmental dynamism faced by the production companies need to be addressed. overall equipment effectiveness measures can be extended to provide a basis for plant materials or overproduction wastes measurement and control. overproduction is a condition where a company manufactures goods above the planned (expected) quantity. in a smart industrial environment, overproduction is a waste from the use of production resources such as material, manpower, and machines b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 61 (muraa, 2016; munoz-villamizar et al., 2018). poorly planned and worn-out equipment can lead to wastage in terms of worker’s time, material, and other components of the production line. different techniques for eliminating wastes in production lines have been on the ground (lennon, 2016). none of them have integrated the error-proof sigma metric tool into oee, as done in this study, as a continuous improvement method of productivity monitoring, which targets customers’ satisfaction by minimizing process variation. this study aims to develop an error-proof oee (oee-ep) model that integrates sigma matric into the modified traditional oee measures. in the past, oee was measured based on only three principal factors availability, performance, and quality efficiencies (afsc, 2010; alexanda, 2012; adams, 2014). this study expands the scope of measuring oee by including human/ergonomics and waste generation into the effectiveness measure. besides, the static nature of traditional oee measure was replaced by a tractable, dynamic oee which utilized an error-proof sigma metric as a continuous improvement tool. the application of error-proof or process variation (sigma metric) tool has not been popular in manufacturing industries as an improvement tool until recent times. the use of the sigma metric in oee as an error-proof parameter can be hardly found. therefore, the integration of the error-proof parameters into the oee measure has widened its scope for sustainable application in modern and complex industrial systems. the rest of the paper is presented as follows: related works are in section 2; modelling and integration of oee factors are in section 3; results and discussion are in section 4; and conclusions are given in section 5. 2. literature review literature was reviewed in line with overall equipment effectiveness measure, sigma metric, and improvement tools, and areas of contribution to oee research. 2.1. overall equipment effectiveness measure traditional oee factors (availability, performance, and quality) performances were optimized using response surface methodology (kunsch et al., 2012). the model introduced no new input to the traditional oee factors. a multidimensional view of technology (amt) in the oee measure provided by swamidassa and kothab (1998) had a direct impact on the performance of large scale firms only. a good oee measure should be versatile and applicable to small-, mediumand largescale industrial systems. the lean bundle oee approach proposed by shah and ward (shah and ward, 2003), considered plant size only while plant integration and aging were neglected. waste generation and workspace (ergonomic) condition were neglected in many other studies related to traditional oee measures (wilson, 2010; ljungberg, 1998; madhavan et al., 2011; nachiappan and anantharaman, 2006; muchiri and pintelon, 2008; wang and pan, 2011; andersson and bellgran, 2015; binti aminuddin et al., 2016). the overall equipment effectiveness of manufacturing line (oeeml) scheme was established by braglia et al., (2009) in an attempt to overcome a limitation of individual equipment over jointly operated machines. the model, however, failed to explain to which extent the effectiveness was supportive to the in-process inventory of spares and workplace management. the application of the traditional oee metric in a steel company by almeanazel (2010) led to the realization of 99% in quality factor, 76% in availability factor, and 72% in plant performance without practical implementation. he suggested the application of lean tools such as single minutes exchange of dies (smed), computer maintenance management system (cmms), and integrated production planning (ipp) to the production system to validate the outcome. the relationship between oee and process capability (pc) measures was established which showed that a cutoff point of 1.33 capability indices (ci) instead of popular 1.0 was possible (garza-reyes et al., 2010). therefore, the oee cutoff point can be shifted beyond 1.0 when measured based on process capability. however, oee greater than 1.0 can be defined as over capability utilisation in a lean manufacturing environment where wastages are not allowed. the efficacy of the oee measure as a productivity tool was established by hansen (2002). productivity can only lead to profitability in an organization if workplace conditions and waste generation rates are considered during oee evaluation. the challenge of variation of oee across firms was addressed by creating a system of a dynamic process for oee evaluation and control over time (zuashkiani et al., 2011). traditional actions taken at fixing oee challenges (reactive maintenance, poor morale) were not sustainable in the long run, due to the emergence of lower oee. under good management, risks and challenges of productivity (reactive maintenance, poor worker morale, and hazard control) are transferrable to relevant experts in the production cycle to sustain oee. 2.2. sigma metric and improvement tools the improvements by reducing losses can be estimated (represented) in terms of sigma statistical control metric based on a normal distribution. on this metric, six sigma (0.0000034) process variation can perform more efficiently than one (0.69), two (0.31), three (0.067), four (0.0062), or five (0.00023) sigma process variations because it provides the highest and practicable improvement probability (99.99966%) which is close to 100% (marselli, 2004). the sixsigma metric allows only 3.4 defects in one million products. six-sigma tool was introduced into a lean development of oee error-proof (oee-ep) model for production process improvement 62 manufacturing strategy recently, and it was classified as blackbelt and greenbelt, based on productivity enhancement (domingo and aguado, 2015). six-sigma focuses on reducing process variation and enhancing process control, while lean manufacturing seeks to eliminate or reduce wastes (non-value added activities) using teamwork, clean, organized, and well-marked workspaces (michael, 2015). lean and six-sigma have the same general purpose of providing the customer with the best possible quality, cost, delivery, and a newer attribute. lean achieves its goal by using philosophical tools such as kaizen, just-intime (jit), workplace organization (5s) and visual controls, single minute exchange of dies (smed), 100% sampling (jidoka), while six-sigma is based on statistical analysis, design of experiment and hypothesis test. there is a need to economize and simplify the oee improvement process by replacing the conglomerate of lean technologies with a singular sigma metric tool for process variation measure. 2.3. areas of contribution to oee research traditional or basic oee model as an important metric in total productive maintenance (tpm) has been successfully applied to packaging, chemical, automobile, production, foundry, and pulp product (rayon fiber) industries in the past (michiri and pintelon, 2008; munteanu et al. 2010; hossain and sarker, 2016, bhattacharjee et al. 2019; sayuti et al., 2019). the traditional oee has given birth to new oee models which are either static/deterministic or stochastic in nature. the new static/deterministic oee measures include: overall asset effectiveness (oae), overall plant effectiveness (ope), total equipment effectiveness performance (teep), production equipment effectiveness (pee) and overall factory effectiveness (ofe) applied in packaging and chemical processing (muchiri and pintelon, 2008); doubly weighted grouping efficiency (dwge) applied in cellular manufacturing systems (sarker, 2001); overall line effectiveness (ole) for automobile industries (nachiappan and anantharaman, 2006); overall equipment effectiveness of a manufacturing line (oeeml) for automobile firm (braglia et al., 2009); global production effectiveness (gpe) for global manufacturing system (lanza et al., 2013); overall throughput effectiveness (ote) for wafer fab and glass firm (muthiah et al., 2008); overall equipment effectiveness market-based (oeemb) for iron and steel industry (anvari et al., 2010); equipment performance and reliability (epr) model for semiconductor production system (samat et al., 2012); rank-order centroid (roc) method in overall weighting equipment effectiveness (owee) for fiber cement roof production system (wudhikarn, 2010); overall equipment and quality cost loss (oeqcl) for fiber cement manufacturing system (wudhikarn, 2012); oee and productivity measure (oeep) in automobile industry (andersson and bellgran, 2015); and oeetotal productive maintenance (tpm) and lean manufacturing (lm) measures in manufacturing systems (binti aminuddin et al., 2016). the stochastic oee models evolved include: probability density function (normal and beta distributions) of oee applied to waterproofing coatings firm (zammori et al., 2010); and the simulation-based taguchi method in weighted oee for crimping manufacturing line (yuniawan et al., 2013). automation of oee measures has been carried out through: integration of a communication system and manufacturing execution system (mes) into the automated data collection (adc) system to enhance accurate data collection in the manufacturing line to enable accurate estimation of throughput, unit per hour (uph) machine rate for semiconductor assembly firm (wang and pan, 2011); and development of software package to identify losses associated with equipment effectiveness (singh et al., 2013). apart from the established drawbacks from the stated studies which include equally weighted oee parameters, subjective weights determination, unexplained lowest oee–loss relationship, insufficient data to obtain the weights, weights probability approximation, inadequate general and quality cost accounting (hossain and sarker, 2016), and effectiveness limitation to machine/equipment rate only, it is inferable from the past studies that the traditional and the evolved oee models have been applied separately to either machine(s)/equipment, plant, production/manufacturing line, or cellular manufacturing system without considering productivity of production line in terms of supply output (delivery) and demand (expectation). besides, the emerging models only built their measures around the basic oee three factors-availability, performance and quality. the stated models cannot work in a fastmoving product (beverage) production line where material needed in the production line is supplied as and when due in batches based on planned (expected) output delivery. therefore, production waste can come from either surplus material supply or overproduction (higher product volume). they failed to also consider issues of: dynamic nature of process variation in a single fast-moving product production line over the years that will enable the manager to plan ahead accurately to meet customers’ delivery as and when required. this will call for a new oee model with the following peculiarities: the ability to moderate/control yearly oee variations to enable attainment of improved and balanced effectiveness of a single fast-moving product production line; inclusion of waste (surplus in materials/ overproduction) and human/ergonomic (production floor environment) factors which are critical to the production line; and integration of regression model and error-proof sigma metric parameters to serve as yearly predictive/dynamic oee improvement tool for management to make the right choice of action. once b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 63 the quality rate is measured based on percentage defective/rework items of output quantity, then it is sensible to consider waste rate as an independent oee factor which is measurable using percentage leftover material (products) of the total input material (output product), excluding defective material (items). it is inferable from the literature that the past oee models are deficient, and hence require improvement to enhance a robust effectiveness measure. new or emerging oee models should take care of a dynamic industrial environment. replacement of cumbersome and costly to implement improvement tools (kaizen, jit, smed, vsm, 5s, automated system, weighting system, lean bundle, lean blackbelt & green belt, etc.) with singular error-proof sigma metric will simplify its application in industries. in this study, the oee model will be made robust to cope with modern production line realities of enhancing competitiveness by integrating workplace conditions and waste generation factors into it. oee measure will be made dynamic through the introduction of a continuous improvement tool, error-proof sigma metric, and integration of time-dependent predictable oee parameters. a simplified, tractable oee error-proof (oee-ep) model will be provided to serve as a good replacement to the cumbersome and costly implementable past models. 3. integration and modelling of oee parameters the methodology employed entails integration of overall equipment effectiveness (oee) factors, formulation of the oee model, and analysis of oee model parameters. 3.1. integration of overall equipment effectiveness parameters for a robust oee evaluation in a balanced production line system for a fast-moving product, a dynamic oee error-proof model was developed through the integration of the traditional oee factors (availability, performance, and quality measures) with other emerging critical-factors/parameters (workplace condition, waste generation, and error-proof sigma metric). figure 1 shows the proposed relationship of the overall equipment effectiveness-oee factors; quality, performance, availability, waste generation, and human ergonomic/workplace condition (temperature). human/ergonomic elements as related to work environments such as temperature, illumination, and workers' wellbeing, have effects on productivity, and its magnitude depends on the degree of deviation from the established standards. wastes can also hinder smooth workflow by creating unnecessary shortage/surplus in outputs of material, machine, manpower, etc. through excessive inventory operation and customers’ satisfaction. the duo, ergonomic and waste, was very important oee factors neglected many years past. the effective production process is attainable under the condition of excellent equipment availability, line performance, product quality, working environment, and waste reduction. this condition was hard to meet simultaneously in the manufacturing system. a sustainable (acceptable) industrial plant operation should have oee greater than 0.5 (50%) (ljungberg, 1998). this shows that the oee of 0.5 is the minimum effectiveness measure required of a manufacturing system to succeed. an oee below 0.5 indicates danger or poor process performance, while an oee≥1 signifies a highly productive process. however, a world-class oee for a production line should be more than 0.85 (sayuti et al., 2019). in a single product balanced production line, an oee factor can be evaluated, in terms of productivity, based on a simple ratio of output delivered to the expected output. variations that existed between the expected and delivered outputs are known as process errors. the multi-factor oee measure led to the emergence of multiple errors due to its series relationship. fivefactor oee can be represented as a series control system with errors measured as a difference between input (expected) oee and output (delivered) oee (fig. 2). the idea is to reduce the error (variation) to make the process effectively satisfactory. this needs to be done gradually on the production line. every error reduction attained periodically is translatable to the oee errorproof sigma metric and evaluable using a productivity improvement index, which can be monitored by the management for decision making (interface). the integration of the error-proof (ep) sigma metric as an improvement tool (𝜎𝜎) in line with the oee control system in fig. 2 has resulted in continuous improvements (eqns. (1)-(3)). figure. 1. overall equipment effectiveness factors integration. b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, m.o. idris 64 figure. 2. system control and improvement framework. x)σ-1(oee=epoee (1) if, ,1≥epoee stop, improvement is satisfactory (2) otherwise, ,1<epoee≤0 continue (3) the improvement )epoee( will continue over time, x (eqns. (1) and (3)) in the presence of error-proof )ep( sigma metric, σ until it attains a satisfactory value (eqn.(2)). power x in eqn. (1) can also be described as improvement (fading) time. system performance was evaluated based on, if; oee ≤ 0.5, then productivity is poor; but for oee > 0.5 means productivity is good, while productivity is described as excellent if oee ≥ 1. with reference to fig. 2, the model relating the traditional oee factors (availability, performance, and quality) was identified to fall into the family of control model on the basis at which a modified version was developed by integrating emerging oee factors (waste and human) into it. the arrow shows a series relationship between the oee factors indicating that a deficiency (error) realized in one factor has multiplying effects on the other oee factors. this reflection is expressed in eqns. (4) and (5). the interface served as an oee outcome error review platform motivated by a sigma metric continuous improvement window that enabled the gradual reduction of process variation to a satisfactory level (eqns. (1)-(3)). error-proof (ep) is process variation measured at a given time over actual and expected outputs. the target of any production firm is to make process variation close to zero to sustain productivity. error-proof (ep) criterion is a strategy of gradual elimination of process variation in a single product production line by the adoption of the sigma metric standards to enhance continuous improvement of the firm’s oee towards the attainment of oee-ep 3.2. overall equipment effectiveness model formulation productivity as a measure of plant effectiveness relating the expected and delivered outputs can be analyzed based on the control diagram given in fig. 2. equation (4) in the earlier studies (bruce, 2006; dilworth, 2013) for measuring the traditional (old) overall equipment effectiveness (oee) was conformed to the established system. αβμ=oee (4) where: α is the availability efficiency of a production equipment β is the performance efficiency of the equipment μ is the quality rate (efficiency) of products (output) equation (4), however, did not consider human ergonomic/workplace condition, γ and waste generation, ω factors (shown in figs 1 and 2), which are also important in determining the overall production line’s effectiveness. hence, eqn. (4) was enhanced to include ergonomic and waste (eqn. (5)), αβμγω='oee (5) equation (5) is a modification of the traditional oee model (bruce, 2006; dilworth, 2013) in eqn. (4). the relationship, as stated before, is a resemblance of the control system represented by fig. 2. the productivity variations in individual or combined oee factors (in series) are measurable as a ratio of output to input resources. dynamic sub-models for evaluating productivity contributions 'iy over the year x of each of the oee factors iy ( )ω.&γ,μ,β,α were then formulated and evaluated as expressed in eqns. (6) and (7), respectively. =)x('yi iy ( )ω.&γ,μ,β,α (6) if [ ,1≥'iy ,1<'iy≤0 (7) system availability, )α(yi was determined as the ratio of actual (delivered) production volume per unit time, )x(t1 and the planned (expected) production volume per unit time, )x(t 2 . stop improvement satisfactory continue improvement b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 65 )x(2t )x(1t=)α(iy (8) system performance, )β(yi was measured as the system performance delivered (in %) per unit time )x(t 0 over the system performance expected (in %) per unit time, )x(t n as shown in (9). )x(nt )x(0t=)β(iy (9) system quality, )(µiy was measured by the quality products delivered, that is, % difference between total quantity produced and total defective/rework items )(xg (in %) over the total quantity produced per unit time, )(xg p (in %). )x(pg )x(g =)μ(iy (10) system human factor/ ergonomics, )γ(yi was estimated using the expression based on the actual (delivered) human working environment contribution to productivity, )x(ts over the expected contribution, ).x(ta eqn. (11) was evaluated based on the production environment (temperature) per unit time, this was found to affected workers’ productivity. the environmental (temperature) change, )x(ta was a planned (expected) production floor conditioning temperature per unit time while, )x(ts was the actual (delivered) workplace temperature. )x(st )x(at=)γ(iy (11) ambient temperature was considered a major human element because it was found that the temperature of the working environment had a critical effect on the machine and the operator’s productivity. a balanced working temperature is needed to enhance simultaneous satisfactory performance in machinery and operator. unstable environmental conditions and climate change will make this very difficult to achieve. therefore, the temperature is a critical ergonomic factor to consider. system waste )ω(yi was measured based on system percentage waste reduction (% allowable material/product surplus) performance, as the ratio of the planned (expected) waste (in %), )x(pw and the actual waste generated (in %), )x(g w )x(wg )x(wp=)ω(iy (12) for clarity, eqns. (8), (10), and (12) may look similar, but they are not the same. eqn. (8) is a measure based on planned (expected) production volume before the actual production took place. eqn. (10) is a measure based on the actual (delivered) quantity of production output before deduction of bad quality or rework product, while eqn. (12) is a measure based on planned (material/product) waste before production over the actual (material/product) waste generated in the production line excluding rejected/rework product. therefore the measures are clearly different because waste was measured in relation to raw material surplus/overproduction, while quality was measured in relation to defective/rework product output. measured effectiveness of individual factors iy ( )ω&γ,μ,β,α were substituted into eqns. (4) and (5) to determine, annually, traditional oee and new 'oee , respectively. evaluation of the outcomes was carried out by following the similar procedure stated before (eqns. (1)-(3)). unsatisfactory effectiveness was dynamically improved upon by utilizing error-proof sigma metric (𝜎𝜎) continuous improvement integrated into the 'oee , that is )oee( c in eqn. (14). this was evaluated over minimum acceptable effectiveness, .min ' i )x(y obtained from each factor, .min *)x('iy (eqn. (13)), x)jσ-1(.min)x( ' iy=.min *)x('iy (13) x ) j -σ1(∏ n i .min(x) ' iy= c oee (14) time(fading)timprovemenisx,power valuesigmaforcounter6,1=j scheme,newor(old)ltraditionafor3or5=n it is inferable from eqn. (14) that jσ = 0 signifies full improvement on the existing performance, while jσ = 1 indicates no improvement at all. therefore, the improvement factor on process variation (sigma errorproof) achievable in the production system was ranged between 0 and 1. this can be reasonably fitted into the statistical standard variations, jσ where 6-1=j . on this premise, only 0.067 and 0.0000034 errors (variations) were allowed under three-sigma )σ( 3 and six-sigma )σ( 6 metrics, respectively. six-sigma can be a good replacement for a costly 100% inspection strategy (jidoka) (marselli, 2004; wilson, 2010). the highest possible improvement attainable, maxcoee in the process (eqn. (14)) was selected as optimum value, *epoee as depicted by eqn. (15). development of oee error-proof (oee-ep) model for production process improvement 66 max c*c* oee=oee=epoee (15) the model as stated is a constrained linear/ nonlinear programming model that can be solved via an analytical approach, and regression analysis using excel tool. 3.3. analysis of model parameters the production firm where data was collected was established in a popular city in nigeria more than two decades ago. the company was into mass production of a single but highly patronized beverage product. the company is a flow process, balanced production line firm, producing a fast-moving product to everincreasing customers. the data collected were primary, and were gotten from a production line of the company based on the available records and presented in tabular format after analysis on the yearly bases to align with the data need of the model. the name of the company was concealed to protect her confidentiality and integrity. there have been reported cases that emanated from data analysis of unstable line availability, system performance, product quality, material waste, and floor condition during the production process which have affected customers patronage due to loss of goodwill, which in turn served as a hindrance to stable and sustainable productivity of the firm. this has defeated the objective of adopting a balanced line production system to enable a continuous flow of product for satisfactory performance at meeting customers’ delivery. based on the foregoing, there is the need to minimize process variation such that the production line is utilized effectively with the target of meeting customers’ demand (delivery) as preplanned with little or no surplus or shortage. on this basis, there is the need to introduce error-proofing (ep) to the traditional and emerging oee metrics as a solution to this challenge of productivity instability. a small data set on expectations and deliveries of oee factors spanning eight (8) years was made available from the company for the analysis (tables 15). the performance factors iy ( )ω.&γ,μ,β,α were estimated based on the expected and delivered production output parameters on a yearly basis ),x(t1 ),x(t 2 ),x(t 0 ),x(t n ),x(g ),x(g p ),x(ta ),x(ts ),x(pw ),x(g w using eqns. (8)-(12). for example, system availability, )α(yi was estimated from the yearly, x , expected production volume (in million), ),x(2t (10,011.00), and delivered production volume (in million), ),x(1t (11,211.17) using eqn. (8) as 12.1= 00.011,10 17.211,11 =)α(iy the other performance ratios were obtained using a similar computation method, and the results are shown in the last columns of tables 1-5. to prevent static performance measures and to allow data updating of oee factors, a dynamic approach was proposed via the regression model by utilizing yearly expected (delivered) output as the dependent variable and the time (in the year, x) as the independent variable. in the analysis, different regression options (exponential, linear, logarithmic, polynomial and power, etc.) were analyzed using excel-tool to find equations that best-fit the data based on the highest coefficient of determination 2r criterion. actual and predicted output outcomes were validated using the mean square error (mse) statistic. the error between the actual and predicted results should not exceed 10% (peng and huang, 2012; ryan et al., 2013, hayes, 2018). the best polynomial regression equations )1≈r( 2 obtained are presented as eqns. (16-25). the yearly predicted results on oee factors by substituting year (x =1, 2,…, 8) into eqns. (16-25) are presented in table 6. predicted oee for individual factor (availability as a sample) )p(i )α(y was measured using similar approach as that of actual oee measure, )a(i )α(y (table 7). the mean square error mse for availability as a sample is: ∑ n 1=x=i n 2])p()α(iy)a()α(iy[ = 00055.0=mse mses for other oee factors were obtained using a similar method. then, oees based on 3-factor (old method) and 5-factor were estimated using eqns. (4 and 5) and the results are shown in table 7. the minimum (lowest) acceptable oee outcomes, .min ' i )x(y corresponding to year(s), min,*x for individual factors (table 7) were selected for continuous improvement analysis on yearly basis (eqn. (13)). for actual (a) availability, 97.0=)x(y .min ' i with error-proof, 3-sigma metric, 067.0=σ 3=j at year, 1=x , the improvement on the minimum oee, .min *' i )x(y was computed by (eqn. (13)), 90.0=)067.01(97.0=)x(y 1.min*'i other elements were computed by following the same process, and their corresponding improvements on 3-factor and 5-factor bases were estimated using b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 67 eqn. (14). for 3-factor, then, 62.0≈)78.0)(89.0)(90.0(=oee c the obtained results are presented in table 8. the same procedure was used for analysis based on the error-proof, six-sigma metric, 0000034.0=σ 6=j (table 9). the best improvement strategy (eqn. (15)) to apply was selected based on maximum improvement, *coee achieved so far (tables 8 and 9). 4. results and discussion results are discussed based on; overview of the plant effectiveness trends establishment for oee factors, oee performance prediction and evaluation, equipment effectiveness evaluation, and equipment effectiveness improvement. 4.1. overview of the plant effectiveness computation results (eqns. 8-11,13) for the plant’s effectiveness (productivity) factors in terms of expected and delivered system availability, system performance, system quality, ergonomic (temperature) and waste generation are presented in tables 1-5. the analysis indicated that the plant was not working to expectation in terms of satisfying the expected deliveries. the plant was expected to satisfy the requirements of meeting customers’ demand (delivery) in terms of equipment availability, product quality, plant performance, working environment (temperature), and waste generation. the company was able to satisfy world-class productivity delivery requirements in a few years. excellent (world-class) productivity (100%) was attained in equipment availability in the years 1, 2, 4, and 5; quality in the year 2; plant performance in the years 1 and 2; working condition (temperature) in the years 1, 2, 6, 7 and 8; and material waste generation in the years 7 and 8. there was evidence of productivity overshot (greater than 100%) due to higher delivery (in quantity) than expected (years 1, 2, 4, and 5). effective deliveries of process factors were below expectations in many years under review (tables 1-5). the worst, system waste productivity of 0.32 was found in the second year of the plant’s operation. this result was odd and inconsistent with other outcomes, and hence not reliable. the plant probably required alignment or corrective maintenance during this period. the results generally showed the management the need to improve on and moderate some effectiveness factors to attain the goal of meeting satisfactory productivity (100%) every year. in this case, the establishment of regression models had table 1. system availability data and analysis. year (x) expected production volume (million) t2(x) production volume delivered (million) t1(x) ratio yi(α) 1 10,011.00 11,211.17 1.12 2 13,541.26 13,541.26 1.0 3 15,096.49 14,593.49 0.97 4 14,281.36 16,290.13 1.14 5 16,035.19 18,058.25 1.13 6 18,996.82 18,499.26 0.97 7 19,736.88 19,298.22 0.98 8 20,808.32 20,341.56 0.98 table 2. system performance data and analysis. year (x) performance expected (%)tn(x) performance delivered (%) to(x) ratio yi(β) 1 90.00 90.00 1.00 2 91.00 91.00 1.00 3 91.50 80.00 0.87 4 91.50 79.00 0.86 5 92.00 82.00 0.89 6 92.00 77.00 0.84 7 92.50 80.00 0.86 8 92.50 82.00 0.89 table 3: system quality data and analysis. year (x) quality product delivered (%)g(x) quality product expected (%) gp(x) ratio yi(μ) 1 75.00 77.00 0.97 2 76.00 70.00 1.09 3 77.00 79.00 0.98 4 77.00 79.50 0.97 5 78.00 81.00 0.96 6 78.50 81.00 0.97 7 79.00 81.20 0.97 8 80.00 81.21 0.99 table 4. system waste data and analysis. year (x) actual (delivered) manufacturing waste (%)gw(x) planned (expected) manufacturing waste (%) pw(x) ratio yi(ω) 1 3.50 3.18 0.91 2 3.10 1.00 0.32 3 3.10 2.16 0.70 4 2.50 1.86 0.74 5 2.00 1.87 0.94 6 2.00 1.86 0.93 7 1.85 1.85 1.00 8 1.85 1.85 1.00 development of oee error-proof (oee-ep) model for production process improvement 68 table 5. production floor ergonomics temperature data. year (x) expected production floor temperature(± 2°c) ta(x) delivered production floor conditioning temperature(± 2°c) ts(x) ratio yi(γ) 1 22.00 22.00 1.00 2 22.00 22.00 1.00 3 23.00 28.50 0.77 4 22.00 22.00 1.00 5 22.00 22.20 0.99 6 23.00 23.00 1.00 7 22.00 22.00 1.00 8 22.00 22.00 1.00 motivated dynamism in evaluating plant effectiveness, while the introduction of error-proofing parameters had played a prominent role in providing continuous improvement by eliminating process variation. 4.2. trends establishment for oee factors regression analysis using excel tool on the company’s production line data (tables 1-5), yielded polynomial equations as the best models for the prediction of the expected, t2(x) and delivered, t1(x) production volume data at a given time (x) as given in eqns. (16) and (17), respectively; t2(x) = 2.4673x6 63.552x5 + 587.96x4 2231.9x3 + 2195.6x2 + 5815.7x + 3668.3; r² = 0.99 (16) t1(x) = -2.7512x6 + 82.94x5 975.96x4 + 5655.7x316801x2 + 25373x 2119.2; r² = 0.9992 (17) ,1≈r 2 is an indication that the models had predicted the production volume data accurately. the regression models obtained for the prediction of the expected and delivered: system performance (tn(x), t0(x)); system quality (gp(x), g(x)); system waste (pw(x), gw(x)); and system ergonomic (temperature) (ta(x), ts(x)) are respectively given as follows: tn(x) = -0.0007x6 + 0.0184x5 0.1976x4 + 1.109x3 3.4913x2 + 6.1828x + 86.375; r² = 0.9865 (18) t0(x) = -0.0521x6 + 1.4675x5 16.284x4 + 89.924x3 255.37x2 + 339.91x 69.625; r² = 0.9845 (19) gp(x) = 0.0031x6 0.0823x5 + 0.8438x4 4.2353x3 + 10.682x2 11.654x + 79.437; r² = 0.9941 (20) g(x) = 0.0276x6 0.7996x5 + 9.2103x4 53.372x3 + 161.19x2 231.88x + 192.6; r² = 0.9796 (21) pw(x) = 0.0028x6 0.0779x5 + 0.8562x4 4.6474x3 + 12.86x2 17.056x + 11.563; r² = 0.9998 (22) gw(x)) = 0.0058x6 0.1668x5 + 1.9113x4 11.011x3 + 33.156x2 48.399x + 27.68; r² = 0.9774 (23) ta(x) = 0.0125x6 0.3375x5 + 3.5625x4 18.563x3 + 49.425x2 62.1x + 50; r² = 1 (24) ts(x) = 0.0302x6 0.7738x5 + 7.7099x4 37.775x3 + 94.537x2 112.67x + 72.938; r² = 0.9846 (25) high coefficients of determination ,1≈r 2 in all cases had indicated that the models were accurate in predicting the expected and delivered outputs. 4.3. performance prediction and evaluation the actual data and predicted results of the production process using regression models (16-25) for the expected and delivered outputs (production volume, product quality, plant performance, floor temperature, and manufacturing waste) are presented in table 6. it can be clearly observed from the table that the predicted results and actual data were in close resemblance (tables 1-5). this showed that the established regression models were adequate for expected and delivered performance prediction (table 6). a sudden rise in waste output prediction was noticed in the fifth year, but as a whole, the actual system effectiveness measures were not largely affected as shown in table 7. irregular (unstable) productivity outcome of the factors for the years under review signaled the need for the management to improve on the plant’s effectiveness. 4.4. evaluation of equipment effectiveness the results of the oee prediction and evaluation of the production line are presented in table 7. it was revealed from the results that the predicted oees values were very close to the actual values (tables 15). the highest mean square error (mse) of 0.0089 was estimated between the outcomes of the actual and predicted parameters for all oee factors (table 7). this was eminently within the acceptable statistical error limit of 0 – 10%. similarly, the 5-factor (new oee) and the 3-factor (old oee) prediction outcomes were very close to that of actual results. the maximum mse of 0.0047 was computed between actual and predicted oee results. the errors were far less than the 10% limit. hence, the model can be effectively applied to the company’s production process to predict and evaluate her equipment effectiveness within an acceptable error margin. the least actual and predicted factor effectiveness (yi’.min) were (0.97,0.95); (0.96,0.94); (0.84,0.78); (0.77,0.84); and (0.70,0.68), respectively for the years (3,6,7); (5,6); (6,7); (3,3); and (3,3). the waste effectiveness (0.32*, 0.33*) were neglected because of its wide gap to the next higher value (0.70, 0.68). the least actual and predicted oee results for the old and the new models were (0.79, 0.68) and (0.45, 0.48). these results were obtained at years (3, 3) and (6,7), respectively (table 7). the irregularity and low oee outcomes had indicated the need for the company to improve the process to survive amid of present and b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 69 future competitors. it was generally revealed that oee from the old model were in most cases higher than those obtained from the new model. the implication of utilizing the old model in the firm was for the management to have the impression that the plant’s effectiveness was good whereas in real sense it was not from the new model’s results. this indicated that the company manager should be vigilant at finding other underline factors which might have critically influenced the oee of the plant. production floor conditions (temperature) and material waste were good examples of such factors identified in and integrated into this system. the oees from the old model seemed unrealistic because of its greediness in estimating the plant’s effectiveness as compared to the realistic new model which was more encompassing and robust in accommodating emerging critical factors that affect productivity. the new model oee outcomes enabled the firm’s manager to know the true condition of the plant's effectiveness to take decisive action at improving the system to prepare it for present and future competitions. it was noticed that oees prediction or evaluation results from old and new methods were not stable through the years under review and were unlikely to be stable in the future. therefore, the enhancement of improvement through the closing of process variation (gap) between expected and delivered effectiveness will go a long way to bring the plant’s performance stability and effectiveness in operations. 4.5. equipment effectiveness improvement the improvement results on plant’s effectiveness by minimizing process variations via application of statistical sigma metrics under 3-sigma (j=3) and 6sigma (j=6) using the oee error-proofing (oee-ep) model (eqns. (13), (14) and (15)) are shown in tables 8 and 9, respectively for 3-factor and 5-factor oee measures. in both cases, the improvement was noticed in varying proportions over the minimum (lowest) oee measured for the single (isolated) and combined factors. the improvement results under the 3-sigma metric were not constant (table 8) over the minimum benchmark (table 7) for the years under review. maximum improvements, actual (0.62, 0.29) and predicted (0.56, 0.28), were estimated (using eqn. (15)) in the first year using 3-factor and 5-factor effectiveness measures, respectively, while the improvements were reducing (fading) steadily in the subsequent years (table 8). the implication of these unstable improvements was to notify the firm’s manager of the need to put in place a sustainable plan towards meeting present and future delivery requests. the results (table 8) further showed the greediness of 3-factor, 3-sigma error-proofing in providing higher improvement outcomes than the 5-factor, 3-sigma model. the improvements for the 9th year showed a similar trend. it was further noticed that the improvements became smaller in future years for both scenarios. this indicated that the improvements will continue to reduce in future years, and then converge to a point. on this basis, the application of this model in measuring the plant’s effectiveness improvement was sustainable due to the integration of regression and error-proof windows that enabled data updating to enhance improvement. however, improvements measured based on error-proof, 6-sigma metric were better and more stable than 3-sigma metric across singular and combined oee factors (table 9). it was revealed from the table that 6-sigma errorproofing produced stable, satisfactory, and sustainable oee improvements of actual and predicted values (0.78, 0.70) and (0.42, 0.40) over the lowest acceptable oee for the 3-factor and 5-factor models, respectively. the improvement excesses of the 3factor model were reflected in these results. though the improvement attained using the 5-factor model was lesser than that of the 3-factor, but both were world-class compliant, sustainable, satisfactory, and significant to survive any emerging competitiveness. in comparison with past similar studies, attainment of 42% improvement based on the 5-factor, 6-sigma model was considered better than 20% and 23% obtained using lean & green (pampanelli et al., 2014) and lean bundle (shah and ward, 2003) models, respectively. table 6. oee factors-annual outcome predictions. oee factors year(x) 1 2 3 4 5 6 7 8 availability expected, t2(x) 9,974.58 13,758.50 14,594.79 14,765.67 16,075.86 18,443.86 20,269.77 20,579.62 availability delivered, t1(x) 11,212.73 13,531.04 14,590.74 16,337.48 17,908.30 18,651.29 19,164.59 20,354.39 quality expected, gp(x) 76.98 70.17 78.48 80.41 80.39 82.37 83.32 86.53 quality delivered, g(x) 74.99 76.04 76.87 77.11 77.43 77.47 76.03 73.42 performance expected, tn(x) 89.995 91.030 91.400 91.61 91.477 91.66 91.466 90.263 performance delivered, t0(x) 89.991 91.066 79.995 81.975 84.925 80.375 71.375 77.375 ergonomics expected, ta(x) 21.997 21.996 22.987 21.968 21.938 22.892 21.829 21.744 ergonomics delivered, ts(x) 23.996 24.076 24.320 22.312 23.576 24.823 18.949 21.767 waste expected, pw(x) 3.18 1.03 2.10 2.12 2.15 3.14 4.75 8.26 waste delivered, gw(x) 3.50 3.10 3.12 2.55 2.30 2.87 4.44 7.06 development of oee error-proof (oee-ep) model for production process improvement 70 table 7. prediction and evaluation of oee for the production process. oee measure, yi /year (x) 1 2 3 4 5 6 7 8 mse.10-3 yi’.min. x*min availability yi(α) actual (a) predicted (p) 1.12 1.12 1.00 0.98 0.97 0.99 1.14 1.11 1.12 1.11 0.97 1.01 0.98 0.95 0.98 0.99 0.55 0.97 0.95 3,6 7 quality yi(μ) actual predicted 0.97 0.97 1.09 1.08 0.98 0.98 0.97 0.96 0.96 0.96 0.97 0.94 0.97 0.91 0.99 0.95 3.00 0.96 0.94 5 6 performance yi(β) actual predicted 1.00 1.00 1.00 1.00 0.87 0.87 0.86 0.89 0.89 0.92 0.84 0.87 0.86 0.78 0.89 0.86 1.25 0.84 0.78 6 7 human/ergonomics, yi(γ) actual predicted 1.00 0.91 1.00 0.91 0.77 0.84 1.00 0.99 0.99 0.93 1.00 0.92 1.00 1.15 1.00 1.00 6.71 0.77 0.84 3 3 waste yi(ω ) actual *unaccepted predicted, bold, mini. acceptable 0.91 0.91 0.32* 0.33* 0.70 0.68 0.74 0.83 0.94 0.94 0.93 1.09 1.00 1.09 1.00 1.17 8.90 0.70 0.68 3 3 5-factor oee’-new-method (a) (p) 0.98 0.90 0.34 0.32 0.45 0.48 0.70 0.78 0.89 0.86 0.74 0.82 0.82 0.84 0.86 0.85 2.73 0.45 0.48 3 3 3-factor oeeold method (a) (p) 1.08 1.08 1.06 1.06 0.83 0.84 0.95 0.95 0.96 0.98 0.79 0.82 0.82 0.68 0.86 0.73 4.70 0.79 0.68 6 7 table 8. improvement dynamism on the oees using three-sigma metrics. improvement dynamism oee-ep (oeec) at σj=3 = 0.067 measured at yi’min./year (x) (eqn. (14)), yi’*(x)min 1 2 3 4 5 6 7 8 9 availability (α) actual (a) predicted (p) 0.90 0.87 0.84 0.98 0.79 0.77 0.73 0.72 0.68 0.67 0.64 0.63 0.59 0.59 0.56 0.55 0.52 0.51 quality (μ) actual predicted 0.89 0.88 0.84 0.82 0.78 0.76 0.73 0.71 0.68 0.67 0.63 0.62 0.59 0.58 0.55 0.54 0.51 0.50 performance (β) actual predicted 0.78 0.73 0.73 0.68 0.68 0.63 0.64 0.59 0.59 0.55 0.55 0.52 0.52 0.48 0.48 0.45 0.45 0.42 human/ergonomics, (γ) actual predicted 0.72 0.78 0.67 0.73 0.63 0.68 0.58 0.64 0.54 0.59 0.51 0.55 0.47 0.52 0.44 0.48 0.41 0.45 waste (ω ) actual predicted 0.65 0.63 0.61 0.59 0.57 0.55 0.53 0.52 0.49 0.48 0.46 0.45 0.43 0.42 0.40 0.39 0.38 0.36 max. improvement over 3-factor, 3sigma: oeec* (oee-ep*) (eqn. (15)) (a) (p) 0.62 0.56 0.52 0.46 0.42 0.37 0.34 0.30 0.27 0.25 0.22 0.20 0.18 0.16 0.15 0.13 0.12 0.11 max. improvement over 5-factor, 3sigma: oeec* (oee-ep*) (eqn. (15)) (a) (p) 0.29 0.28 0.21 0.20 0.15 0.14 0.10 0.10 0.07 0.07 0.06 0.06 0.04 0.04 0.03 0.03 0.02 0.02 table 9. improvement dynamism on the oees using six-sigma metrics. improvement dynamism oee-ep (oeec) at σj=6 = 0.0000034 measured at (yi)min. /year (x) (eqn. (14)), yi’*(x)min. 1 2 3 4 5 6 7 8 9 availability (α) actual (a) predicted (p) 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 0.97 0.95 quality (μ) actual predicted 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 0.96 0.94 performance (β) actual predicted 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 0.84 0.78 human/ergonomics, (γ) actual predicted 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 0.77 0.84 waste (ω ) actual predicted 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 0.70 0.68 max. improvement over 3-factor, 6sigma: oeec* (oee-ep*) (eqn. (15)) (a) (p) 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 0.78 0.70 max. improvement over 5-factor, 6sigma: oeec* (oee-ep*) (eqn. (15)) (a) (p) 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 0.42 0.40 the 3-factor based performance (65%) attained by hedman et al. (2016) using automatic measurement approach can only compete with the 62% improveement of this new 3-factor, 3-sigma error-proof model (table 8), but far below 78% attainable from the 3factor, 6-sigma error-proof model (table 9). b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 71 5. conclusion the oee model has been effectively formulated to plan production, labor utilization, maintenance, and working environment as reflected in the attainment of waste reduction, availability, quality, and performance improvements. the integration of human/ergonomic workplace environment (temperature) and system waste generation factors into the oee measures has broken new ground in the field of overall equipment effectiveness research. attainment of dynamism in the system effectiveness measures through the incorporation of best regression models and integration of encompassing error-proof, sigma metric parameter has made the system unique in terms of providing sustainable performance as compared to other popular techniques. the six-sigma has performed well in system errorproofing and at the same time providing better productivity improvement platforms for the fastmoving goods production process. the models were applied successfully to measure critical equipment effectiveness factors which are availability, performance, quality, workplace condition (temperature), and material waste of a production plant. on this basis, challenges that hindered the effective operation of the plant were mitigated and improved system was sustained. results obtained using different scenarios of the model application to a production line showed a low performance which was improved upon using a better improvement program, the new oee-ep model. the application of the oee-ep model has made a landmark achievement by providing the highest level of improvements (0.42, 0.78), which are far better than those obtainable from past studies (0.20, 0.23). this landmark achievement can be attributed to excellent accuracy provided by error-proof, six-sigma metric in minimizing process variation. specific conclusions drawn from this study are enumerated as follows: i. unstable overall equipment effectiveness measures were obtained for the company in the years under review. this revealed the need for the company to be evaluating her productivity/overall equipment effectiveness annually for early correction of any ailing process factor before it gets out of control to stabilize the changing firm’s productivity over time. ii. the continuous improvement of the production process has been achieved for the company using overall equipment effectiveness error-proof strategy without an increase in input resources. this outcome has shown the management of the company the undesirability of increasing input resources before getting the productivity of the production process improved. proper choice of resources management strategy can sustain and improve productivity. this has been shown by the outcome of this study. iii. there is a wide gap between traditional 3-factor oee and new scheme 5-factor oee results. therefore, in the presence of uncertainty, it is advisable that the company base its productivity measure on 5-factor which is more robust. iv. the integration of error-proof sigma metric as a means of continuous improvement in the production process of the company has resulted in sustainable and stable productivity improvement. this lean metric has performed better than other improvement metrics such as lean bundle, lean green, and automation. v. the data set available for productivity/oee prediction was small. this can affect prediction accuracy. to improve prediction accuracy, it is advisable that the company has a good record of annual oee data for timely data updates. further study is required in real-time implementation of the process through the application of computer programs to handle large data, to serve as a source of artificial intelligence at enhancing robust decision making in the future as regards the performance of the oee-ep model. in such a study the null hypothesis shall be; the new oee-ep model is robust enough to withstand emerging (future) wastes and environmental challenges, while the alternative hypothesis shall be the model cannot withstand future challenges. proper identification of processes that required improvement, choice of appropriate improvement methods (tools) required and adequate representation of the process outcomes for good decision making are other areas of further research. the implementation of this scheme will require good experience and deep knowledge of processes identification and analysis with criticality and necessity consideration. conflict of interest the authors declare that there are no conflicts of interest as regards this article. funding managements of federal university of technology, akure, and fast moving products industry, ibadan, nigeria are acknowledged for providing funds and facilities for the successful completion of this project. acknowledgment b. kareem thanks the management and staff of fast moving products industry for providing relevant information used in this study. the contribution of industrial and production engineering research group leader, federal university of technology akure (futa), nigeria is also acknowledged. development of oee error-proof (oee-ep) model for production process improvement 72 references adams ja (2014), human factors engineering. new york: mcmillan publishing. afsc-air force systems command, (2010),air force systems command design handbook, human factors engineering. united kingdom: mcmillan publishing. alexander dc (2012), the practice and management of industrial ergonomics. new york: englewood cliffs, nj: prentice-hall. almeanazel otr (2010), total productive maintenance review and overall equipment effectiveness measurement. jordan journal of mechanical and industrial engineering (jjmie) 4(4): 517 – 522. andersson c, bellgran m (2015), on the complexity of using performance measures: enhancing sustained production improvement capability by combining oee and productivity. journal of manufacturing systems, 35: 144-154. anvari f, edwards r, starr a (2010), evaluation of overall equipment effectiveness based on market. journal of quality in maintenance engineering 16(3): 256-270. bhattachcharjee a, roy s, kundu s, tiwary m, chakraborty r (2019), an analytical approach to measure oee for blast furnaces. journal ironmaking & steelmaking; processes, products and applications, published online: 31jan 2019. binti aminuddin na, garza-reyes ja, kumar v, antony j, rocha-lona l (2016), an analysis of managerial factors affecting the implementation and use of overall equipment effectiveness. international journal of production research 54(15): 4430-4447. braglia m, frosolini m, zammori f (2009), overall equipment effectiveness of a manufacturing line (oeeml): an integrated approach to assess systems performance. journal of manufacturing technology management 20(1) 8-29. bruce ch (2006), best practices in maintenance, http://www.tpmonline.com/articles_on_total_ productivemaintenance/management.htm, accessed 26:09:2011. butlewski m, dahlke g, drzewiecka-dahlke m, gorny a, pacholski l (2018), implementation of tpm methodology in worker fatigue management a macroergonomic approach. advanced intelligence systems and computing 605:32-41. dilworth jb (2013), production and operations management. new york: mcgraw-hill. domingo r, aguado s (2015), overall environmental equipment effectiveness as a metric of a lean and green manufacturing system. sustainability 7(7): 9031-9047. garza‐reyes ja, eldridge s, barber kd, soriano‐meier h (2010), overall equipment effectiveness (oee) and process capability (pc) measures: a relationship analysis. international journal of quality & reliability management 27(1): 48-62. gharbi a, kennen jp (2000), production and preventive maintenance rates control for a manufacturing system: an experimental design approach, international journals of production economics 65: 275-287. ghazali a, adegbola aa, olaiya ka, yusuff on, kareem b (2013): evaluation of workers’ perception on safety in selected industries within ibadan metropolis. seem research and development journal 2(1): 82-88. godfrey p (2002), how the return on investment and cash flow can be improved by using the overall equipment effectiveness measure. manufacturing engineer 81(2): 109-112. hansen rc (2002), overall equipment effectiveness: a powerful production/maintenance tool for increased profits. new york: industrial press inc. hayes af (2018), introduction to mediation, moderation, and conditional process analysis: a regression-based approach. new york: the guilford press. hedman r, subramaniyan m, almström p (2016), analysis of critical factors for automatic measurement of oee. procedia cirp 57: 128-133 hossain msj, sarker br (2016), overall equipment effectiveness measures of engineering production systems, conference paper: 2016 annual meeting of the decision science institute, austin, texas, usa, accessed at www.researchgate.net/publication/315666207 joshi s, gupta r (1986), scheduling of routine maintenance using production schedule and equipment failure history. international journal of computers and industrial engineering 10(1) 11-20. kadiri m. a. (2000), “scheduling of preventive maintenance in a manufacturing company: a computer model approach”, unpublished m.s. thesis, dept. of ind. and prod. eng., university of ibadan, ng. kareem b, jewo ao (2015), development of a model for failure prediction on critical equipment in the petrochemical industry. engineering failure analysis 56: 338–347. künsch hr, stefanski la, carroll rj (2012), conditionally unbiased bounded-influence estimation in general regression models, with applications to generalized linear models. journal of the american statistical association 84 (406): 23-37 lanza g, stoll j, stricker n, peters s, lorenz c (2013), measuring global production effectiveness. forty sixth cirp conference on manufacturing systems 2013, procedia cirp, 7 (2013), 31-36. lennon p (2016), root-cause analysis underscores the https://www.emeraldinsight.com/author/anvari%2c+farhad https://www.emeraldinsight.com/author/edwards%2c+rodger https://www.emeraldinsight.com/author/starr%2c+andrew https://www.emeraldinsight.com/author/braglia%2c+marcello https://www.emeraldinsight.com/author/frosolini%2c+marco https://www.emeraldinsight.com/author/zammori%2c+francesco http://www.tpmonline.com/articles_on_total_%20productivemaintenance/management.htm http://www.tpmonline.com/articles_on_total_%20productivemaintenance/management.htm https://www.emeraldinsight.com/author/arturo+garza-reyes%2c+jose https://www.emeraldinsight.com/author/eldridge%2c+steve https://www.emeraldinsight.com/author/barber%2c+kevin+d https://www.emeraldinsight.com/author/soriano-meier%2c+horacio https://ieeexplore.ieee.org/xpl/recentissue.jsp?punumber=2189 https://ieeexplore.ieee.org/xpl/recentissue.jsp?punumber=2189 https://www.sciencedirect.com/science/article/pii/s2212827116311763#! https://www.sciencedirect.com/science/article/pii/s2212827116311763#! https://www.sciencedirect.com/science/article/pii/s2212827116311763#! https://www.sciencedirect.com/science/journal/22128271 https://www.sciencedirect.com/science/journal/22128271/57/supp/c b. kareem, a.s. alabi, t.i. ogedengbe, b.o. akinnuli, o.a. aderoba, and m.o. idris 73 importance of understanding, addressing and communicating cold chain equipment failures to improve equipment performance. technical report, los angeles, usa. ljungberg o (1998), measurement of overall equipment effectiveness as a basis for tpm activities. international journal of operations & production management 18(5): 495-507. madhavan e, paul r, rajmohan m (2011), study on the influence of human factors. european journals, 23: 179-192. marselli m (2004), lean manufacturing and sixsigma, wire journal international, http://www.aemconsulting.com/pdf/lean.pdf. accessed 12-10-2018 martand t (2014), industrial engineering and production management. new delhi: s.chand & company ltd. meyer fe, stewart jr (2018), motion and time study for lean manufacturing, 3rd edition, js-89867 us, https://rzo0o4khw01.storage.googleapis.com/md ezmddmxnjcwoq==01.pdf accessed 19-052018. michael jk (2015), six sigma green belt training manual, michael jk (uk based), lagos, ng. muchiri p, pintelon l (2008), performance measurement using overall equipment effectiveness (oee): literature review and practical application discussion. international journal of production research 46(13): 3517-3535. muñoz-villamizar a, santos j, montoya-torres j, jaca c (2018), using oee to evaluate the effectiveness of urban freight transportation systems: a case study. international journal of production economics 197: 232-242. munteanu d, gabor c, munteanu i, schreiner a (2010), possibilities for increasing efficiency of industrial equipment. bulletin of the transilvania university of brasov, series i: engineering sciences 3(52):199204. muraa md (2016), worker skills and equipment optimization in assembly line balancing by a genetic approach. technical report, dept. of civil and ind. eng., univ. of pisa, italy. muthiah kmn, huang sh, mahadevan s (2008), automating factory performance diagnostics using overall throughput effectiveness (ote) metric. international journal of advanced manufacturing technology 36(8): 11– 824. mwanza bg (2017), an assessment of the effectiveness of equipment maintenance, practices in public hospitals. technical report, univ. of johannesburg, south africa. nachiappan rm, anantharaman n (2006), evaluation of overall line effectiveness (ole) in a continuous product line manufacturing system. journal of manufacturing technology management 17(7): 987-1008 pampanelli ab, found p, bernardes am (2014), a lean & green model for a production cell. journal of cleaner production 85: 19-30 peng l, huang y (2012), survival analysis with quartile regression models. journal of the american statistical association 103 (482): 637649 prinz c (2017): implementation of a learning environment for an industrie 4.0 assistance system to improve the overall equipment effectiveness. technical report, a ruhr-universität bochum, chair of production systems, bochum, germany. ryan ag, montegomery dc, peck ea, vining gg, ryan ag (2013), solution manual to accompany introduction to linear regression analysis. new jersey: john wiley & son inc. samat ha, kamaruddin s, azid ia (2012), integration of overall equipment effectiveness (oee) and reliability method for measuring machine effectiveness. south african journal of industrial engineering 23(1): 92-113. sarker br (2001), measures of grouping efficiency in cellular manufacturing systems. european journal of operational research 130(3): 588-611. sayuti m, juliananda, syarifuddin f (2019), analysis of the overall equipment effectiveness (oee) to minimize six big losses of pulp machine: a case study in pulp and paper industries. international conference on science and innovated engineering (i-cosine), iop conf. series: materials science and engineering 536 (2019): 012136 doi:10.1088/1757-899x/536/1/012136. shah r, ward pt (2003), lean manufacturing: context, practice bundles, and performance. journal of operations management 21(2): 129149 singh r, shah db, gohil am, shah mh (2013), overall equipment effectiveness (oee) calculation automation through hardware & software development. procedia engineering 51: 579-584 swamidassa pm, kothab s (1998), explaining manufacturing technology use, firm size and performance using a multidimensional view of technology. journal of operations management 17(1): 23-37 wang ty, pan hc (2011), improving the oee and uph data quality by automated data collection for the semiconductor assembly industry. expert systems with applications 38(5): 5764-5773. wilson l (2010), how to implement lean manufacturing. new york: mcgraw-hill inc. wudhikarn r (2010), overall weighting equipment effectiveness. proceedings of the 2010 ieee ieem, 23-27. wudhikarn r (2012), improving overall equipment cost loss adding cost of quality. international journal of production research https://www.emeraldinsight.com/author/ljungberg%2c+%c3%95rjan http://www.aemconsulting.com/pdf/lean.pdf https://rzo0o4khw01.storage.googleapis.com/mdezmddmxnjcwoq==01.pdf https://rzo0o4khw01.storage.googleapis.com/mdezmddmxnjcwoq==01.pdf https://www.sciencedirect.com/science/article/pii/s0959652613003958#! https://www.sciencedirect.com/science/article/pii/s0959652613003958#! https://www.sciencedirect.com/science/article/pii/s0959652613003958#! https://www.sciencedirect.com/science/journal/09596526 https://www.sciencedirect.com/science/journal/09596526 https://www.sciencedirect.com/science/journal/09596526/85/supp/c https://www.sciencedirect.com/science/journal/02726963 https://www.sciencedirect.com/science/article/pii/s1877705813000830#! https://www.sciencedirect.com/science/article/pii/s1877705813000830#! https://www.sciencedirect.com/science/article/pii/s1877705813000830#! https://www.sciencedirect.com/science/article/pii/s1877705813000830#! https://www.sciencedirect.com/science/journal/18777058 https://www.sciencedirect.com/science/article/abs/pii/s0272696398000163#! https://www.sciencedirect.com/science/journal/02726963 development of oee error-proof (oee-ep) model for production process improvement 74 50(12): 3434-3449. yuniawan d, ito t, bin me (2013), calculation of overall equipment effectiveness weight by taguchi method with simulation. concurrent engineering 21(4): 296-306. zammori f, braglia m, frosolini m (2011), stochastic overall equipment effectiveness. international journal of production research 49(21): 6469-6490. zuashkiani a, rahmandad h, jardine aks (2011), mapping the dynamics of overall equipment effectiveness to enhance asset management practices. journal of quality in maintenance engineering 17(1): 74-92. https://www.emeraldinsight.com/author/zuashkiani%2c+ali https://www.emeraldinsight.com/author/rahmandad%2c+hazhir https://www.emeraldinsight.com/author/jardine%2c+andrew+ks 1. introduction 2. literature review 3. integration and modelling of oee parameters the best improvement strategy (eqn. (15)) to apply was selected based on maximum improvement, achieved so far (tables 8 and 9). 4. results and discussion 4.1. overview of the plant effectiveness 4.2. trends establishment for oee factors 4.3. performance prediction and evaluation 4.4. evaluation of equipment effectiveness 4.5. equipment effectiveness improvement conflict of interest funding acknowledgment references comp27103.qxd 1. introduction multidimensional signal processing is of paramount importance in application areas such as biomedicine, computer vision, multimedia, industrial inspection and remote sensing. in all these areas end-users and system developers have to work with multidimensional data sets (haralick and shapiro, 1992; plataniotis and venetsanopoulos, 2000). noise filtering is an essential part of any image processing based system, whether the final information is used for human inspection or for an _________________________________________ *corresponding author e-mail: lazhar@squ.edu.om automatic analysis. a number of sophisticated multichannel filters have been developed to date for image filtering (pitas and venetsanopoulos, 1990; astola and kuosmanen, 1997). nonlinear filters applied to images are required to suppress the noise while preserving the integrity of edge and other detail information. to this end, vector processing of multichannel images is more appropriate compared to traditional approaches that use instead component-wise operators machuca and phillips, (1983). for instance, the vector median filter (vmf) minimizes the distance in the vector space between the image vectors as an appropriate error criterion astola et al. (1990). it inherently utilizes the correlation of the channels and keeps the the journal of engineering research vol. 2, no. 1 (2005) 1-11 vector directional distance rational hybrid filters for color image restoration l. khriji* department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, al-khod 123, sultanate of oman. received 27 january 2003; accepted 24 december 2003 abstract: a new class of nonlinear filters, called vector-directional distance rational hybrid filters (vddrhf) for multispectral image processing, is introduced and applied to color image-filtering problems. these filters are based on rational functions (rf). the vddrhf filter is a two-stage filter, which exploits the features of the vector directional distance filter (vddf), the center weighted vector directional distance filter (cwvddf) and those of the rational operator. the filter output is a result of vector rational function (vrf) operating on the output of three sub-functions. two vector directional distance (vddf) filters and one center weighted vector directional distance filter (cwvddf) are proposed to be used in the first stage due to their desirable properties, such as, noise attenuation, chromaticity retention, and edges and details preservation. experimental results show that the new vddrhf outperforms a number of widely known nonlinear filters for multi-spectral image processing such as the vector median filter (vmf), the generalized vector directional filters (gvdf) and distance directional filters (ddf) with respect to all criteria used. keywords: rational functions, vector rational filters, vector median filters, vector directional distance filters ¢¢üü∏∏îîàà°°ùÿÿgg” ~bh .±é«w’g io~©àe qƒ°üdg á÷é©ÿ ''á«g霒g áaé°ùª∏d øé¡ÿg »≤£æÿg äé¡éàÿg ≈ø°üe '' ≈ª°ùj á«£ÿg òz äé«ø°üÿg øe ~j~l ´ƒf åëñdg gòg ω~≤j : √oéøà°s’g ºàj å«m á«ø°üà∏d úà∏môe øe ¿ƒµàjh á«≤£æÿg ∫gh~dg ≈∏y äé«ø°üÿg øe ´gƒf’g iòg ~ªà©jh .áfƒ∏ÿg qƒ°üdg á«ø°üj äóµ°ûe π◊ äé«ø°üÿg øe ´ƒædg gòg ≥«ñ£j ádgo ≥«ñ£j œéf ìî≤ÿg ≈ø°üÿg êôn ¿ƒµjh .»≤£æÿg πeé©dg éjgõe ™e õcôÿg ¿hrƒe á«g霒g áaé°ùª∏d äé¡éàÿg ≈ø°üeh á«g霒g áaé°ùª∏d äé¡éàÿg ≈ø°üe éjgõe øe áø∏àıg ºgéjgõe øe ioéøà°sód ∂dph õcôÿg ¿hrƒe á«g霒g áaé°ùª∏d äé¡éàÿg ≈ø°üÿ ådéãdgh »g霒g áaé°ùÿg ¬éàe é«ø°üe ¿éækg äélôfl çók ≈∏y »≤£æÿg ¬éàÿg .π«°uéøàdgh ±gƒ◊éh ®éøàm’gh êƒ∏dg ¢séñàm’gh aé°vƒ° dg ∞«øîj πãe ≈£°sƒdg ¬éàÿg ≈ø°üe πãe ±é«w’g o~©àe qƒ°üdg èdé©ÿ áahô©ÿg á«£ÿg òz äé«ø°üÿg øe o~y øy iqéñy ìî≤ÿg ≈ø°üÿg äélôfl áä«g ¿g üqééàdg èféàf âæ«h ~bh .áe~îà°ùÿg òjé©ÿg áaéµd áñ°ùædéh á«g霒g áaé°ùª∏d äé¡éàÿg ≈ø°üeh ωé©dg á«g霒g áaé°ùª∏d äé¡éàÿg ≈ø°üeh g::áá««mmééààøøÿÿgg ääggooôôøøÿÿ.á«géœc’g áaé°ùª∏d äé¡éàÿg ≈ø°üe ,≈£°sƒdg ¬éàÿg ≈ø°üe ,»≤£æÿg ¬éàÿg ≈ø°üe ,á«≤£æÿg ∫gh~dg áfƒ∏ÿg qƒ°üdg á÷é©ÿ á«géœc’g áaé°ùª∏d øé¡ÿg »≤£æÿg äé¡éàÿg ≈ø°üe desirable properties of the scalar median; the zero impulse response, and the preservation of the signal edges. vmfs are derived as maximum likelihood estimators for an exponential distribution. they perform accurately when the noise follows a long-tailed distribution (e.g. exponential or impulsive); moreover any outliers in the image data are easily detected and eliminated by vmf's. a second class of filters, called vector directional filters (vdf), uses the angle between the image vectors as an ordering criterion trahanias et al. (1996). the vdf's are optimal directional estimators and consequently are very effective in preserving the chromaticity of the image vectors. a drawback of vdf lies in the fact that they do not consider the magnitude of the image vectors, separating in this way the processing of vector data into directional processing and magnitude processing. however, the resulting filter structures are complex and the corresponding implementations may be slow since they operate in two steps. a third class of filters uses rational functions in its input/output relation, and hence the name ‘’vector rational filter’’ (vrf) (khriji et al. 1999; khriji and gabbouj, 2001). there are several advantages to the use of this function, similar to a polynomial function, a rational function is a universal approximator (it can approximate any continuous function arbitrarily well); however, it can achieve a desired level of accuracy with a lower complexity, and possesses better extrapolation capabilities. moreover, it has been demonstrated that a linear adaptive algorithm can be devised for determining the parameters of this structure leung and haykin, (1994). in this paper, a novel nonlinear vector filter class is proposed: the class of vector directional distance-rational hybrid filters (vddrhfs). the vddrhf is formed by three sub-filters (two vector directional distance filters and one center weighted vector directional distance filter) and one vector rational operation. vddrhfs are very useful in color (and generally multichannel) image processing, since they inherit the properties of their ancestors. they constitute very accurate estimators in longand shorttailed noise distributions and, at the same time, preserve the chromaticity of the color image. moreover, they act in small window and require a low number of operations, resulting in simple and fast filter structures. this paper is organized as follows. section 2 briefly reviews rational functions and vector rational function filters. the weighted vector directional distance filters are presented in section 3. in section 4, we define the vector directional distance-rational hybrid filter (vddrhf) and point out some of its important properties; in addition, the proposed filter structures have been considered. section 5 includes simulation results and discussion of the improvement achieved by the new vddrhf. in order to incorporate perceptual criteria in the comparison, the error is measured in the uniform l*a*b* color space, where equal color differences result in equal distances (pratt, 1991). section 6 concludes the paper. 2. rational and vector rational function filters a rational function is the ratio of two polynomials. to be used as a filter, it can be expressed as: (1) where x1,x2,...,xm are the scalar inputs to the filter and y is the filter output, ao,bo, aij and bij (i = 1,...,m, j = 1,...,m) are filter parameters. the representation described in eq. 1 is unique up to common factors in the numerator and denominator polynomials. the rational function (rf) must clearly have a finite order to be useful in solving practical problems. like polynomial functions, a rational function is a universal approximator leung and haykin, (1994). moreover, it is able to achieve substantially higher accuracy with lower complexity and possesses better extrapolation capabilities than polynomial functions. straight forward application of the rational functions to multichannel image processing would be based on processing the image channels separately. this however, fails to utilize the inherent correlation that is usually present in multichannel images. consequently, vector processing of multichannel images is desirable machuca and phillips, (1983). the generalization of the scalar rational filter definition to vector and scalar signals alike is given by the following definition: definition 2.1 let x1, x2,...,xn be the n input vectors to the filter, where xi = [x1i, x2i,...,xmi ]t and xki € {0,1,...,m}, m is an integer. the vrf output is given by (2) where p is a vector-valued polynomial and q is a scalar polynomial. both are functions of the input vectors. the ith component of the vrf output is written as 2 the journal of engineering research vol. 2, no. 1 (2005) 1-11 ∑ ∑∑ ∑ ∑∑ = == = == +++ +++ = m j m k kjjk m j jj m j m k kjjk m j jj xxbxbb xxaxaa y 1 1 2 1 10 1 1 2 1 10 ... ... [ ] [ ]tm n n n rfrfrf xxxq xxxp xxxrfvrf ,,, ),,,( ),,,( ,,, 21 21 21 21 l l l l = = = (3) where (4) and (5) (6) when the vector dimension is 1, the vrf reduces to a special case of the scalar rf. 3. weighted vector directional distance filters (7) (8) the power parameter p is a design parameter ranged from 0 to 1. it controls the importance of the angle criterion versus the distance criterion in the overall filter process. at the two extremes, p=0 or p=1, the operator behaves as either magnitude processing or directional processing, respectively. the case of p=0.5 gives equal importance to both criteria. we have adopted a constant operational value p=0.25 as explained by karakos and trahanias (1997). this represents a compromise between the different values implied by the different noise models. moreover, since the performance measures remain practically unchanged for a range of p values, which includes the value p=0.25, this is “safe” value independent of the noise distribution. if all weight coefficients are set to the same value, then all angular distances will have the same importance and the wvddf operation will be equivalent to the vddf. if only the center weight is varied, whereas other weights remain unchanged, the wvddfs perform the center weighted vector directional distance filtering (cwvddf). 3 the journal of engineering research vol. 2, no. 1 (2005) 1-11 { }m xxxq xxxp rf n n i i ,,1,0 ),,,( ),,,( 21 21 l l l ∈ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ = ( ) l l ++ += ∑ ∑ ∑ = = = n k n k i k i kkk n k i kkn i xxa xaaxxxp 11 12 2121 1 021 ,,, ( ) ∑ ∑ = = −+= n j n k p kjjkn xxbbxxxq 1 1 021 ,,, l p . is the l p-norm, and the squar e bracket notation used in eq. (3) above, [ ]α refers to the integer part of α , +ℜ∈α . jkbb ,00 > are constant, and iriia ,,2,1 l used in eq. (4) is a function of the input ( )nirii xxxfa ,,, 21,,2,1 ll = let ml zzx →: represent a multichannel image, where l is an image dimension and m characterizes a number of channels. in the case of standard color images, parameters l and m are equal to 2 and 3 respectively. let where, ( )ji xx ,θ represents the angle between two mdimensional vectors imiii xxxx ,,, 21 l= and jmjjj xxxx ,,, 21 l= and ( ) πθ ≤≤ ji xx ,0 . ( ) ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ = − ji t ji ji xx xx xx . . cos, 1θ if ordered distances give the ordering scheme )()2()1( nddd ≤≤≤ l and the same ordering is implied to the input vector -valued samples )(),(),( 2211 nn dxdxdx l , it results in the ordered input set )()2()1( nxxx ≤≤≤ l . the output of the wvddf is the sample { }nxxxx ≤≤≤∈ l21)1( associated with the minimum weighted angular magnitude distance { }ndddd ≤≤≤∈ l21)1( . thus, the wvddfs are outputting the sample from the input set, so that the local distortion is mini mized and no color artifacts are produced (lukac, 2002). ( ) [ ] nip xxwxxwd pn j jij p n j jiji ,,2,1 ,1,0 ., 1 11 l=∈ ⎥⎦ ⎤ ⎢⎣ ⎡ −⎥⎦ ⎤ ⎢⎣ ⎡= − ∑ = ∑ = θ { }nizxw li ,,2,1; l=∈= represent a filter window of a finite length n, where nxxx ,,, 21 l is a set of noisy samples. note, that the position of the filter window is determined by the central sample 2/)1( +nx . let us assume that nwww ,,, 21 l represent a set of positive real weights, where each weight jw , for nj ,,2,1 l= is associated with the input sample jx . introducing the aggregated weighted angu lar-magnitude distance associated with input sample ix gives 4. vector directional distance-rational hybrid filters (vddrhf) 4.1. design procedure the vmf effectively removes impulsive noise and the vector directional filter operates on the directional domain of color images. the filtering schemes based on directional processing of color images may achieve better performance than vmf based approaches in terms of color chromaticity (direction of color data) preservation. in fact, the combination of the direction and the magnitude process (vector directional distance filter) is suitable for the human visual system and can give a better-balanced result between noise reduction and chromaticity retention. moreover, as pointed out in (khriji, et al. 1999; khriji and gabbouj, 2001), a vector rational filter performs well for relatively high snr gaussian contaminated environments. when both impulsive and gaussian noises are present, neither the vector rational filter nor vddf perform well. thus, it is necessary to use a hybrid structure filter. this structure is made of two filtering stages, as shown in fig.1. they combine in the first stage the lp-norm criteria and angular distance criteria to produce three output vectors in which two vector directional distance filter outputs and one center weighted vector directional distance filter output to eliminate impulsive noise, preserve edges and color chromaticity. in the second stage a vector rational operation acts on the above three output vectors to produce the final output vector. the aim of the final stage, in addition to its detail preserving capability, is to remove gaussian noise and small magnitude impulsive noise. the vddrhf is defined as follows: (9) (10) where, the coefficients c1, c2 and c3 are some constants. (11) therefore, the vddrhf operates as a linear low-pass filter between three nonlinear sub-operators, the coefficients of which are modulated by the edge-sensitive component. the proposed structures of the vddrhf are shown in figure 1. 4.2. edge sensor the proposed edge sensor is written as, (12) depending on the value of the parameter p, the edge sensor behaves as follows: 1. p=0, the edge sensor is based on the magnitude difference between the vectors in the l2-norm sense, (13) 2. p=1, the angles between the directions of the color vectors is now used as an edge sensitivity measure. the goal is to sustain the sharpness of the filtered image by 4 the journal of engineering research vol. 2, no. 1 (2005) 1-11 definition 4.1 the output vector )( ixy of the vddrhf, is the result of a vector rational function taking into account three input sub -functions which form an input functions set { }321 ,, φφφ where the “central one ” ( 2φ ) is fixed as a center weighted vector directional distance sub -filter, [ ])(),(. )( )()( 31 3 1 2 ii j ijj ii xxdkh x xxy φφ+ φ +φ= ∑ = β where, d[.] is a scalar output function, which plays an important role in rational function as an edge sensing term, [ ]321 ,, ββββ = characterizes the constant vector coefficient of the input sub -functions. in this approach, we have chosen a very s imple prototype fil ter coefficients which satisfy the unbiased condition: 0 3 1 =∑ =i iβ . in our study, [ ]t1,2,1 −=β and h and k are some positive constants. the parameter k is used to control the amount of the nonlinear effect. the sub-filters 1φ and 3φ are chosen so that an acceptable compromise between noise reduction, edge and chromaticity preservation is achieved. it is easy to observe that this vddrhf differs from a linear low pass filter mainly for the scaling, which is introduced on the 1φ and 3φ terms. indeed, such terms are divided by a factor proportional to the output of an edge-sensing term characterized by the function [ ]31 , φφd . the weight of the vector directional distance -operation output term is accordingly modified, in order to keep the gain constant. the behavior of the proposed vddrhf structure for different positive values of parameter k is the following: 1. 0≈k , the form of the filter is given as a linear low pass combination of the three nonlinear sub functions: )(.)(.)(.)( 332211 iiii xcxcxcxy φ+φ+φ= 2. ∞→k , the output of the filter is identical to the central sub -filter output and the vector rational function has no effect: )()( 2 ii xxy φ= 3. for intermediate values of k the [ ]31 , φφd term perceives the presence of a detail and accordingly reduces the smoothing effect of the operator. [ ] ( ) ( )[ ]21313131 .,, ppd −φ−φφφ=φφ θ [ ] 23131 , φ−φ=φφd where, 2 . denotes the l 2-norm. preserving the transitions detected in the color space, where transactions are represented by the angles between the color vectors. at a fixed luminance, small angles between color-vectors denote “color” homogeneous regions; whereas, large angles indicate edges as given below: (14) for intermediate value of p (0<p<1) both criteria (distance and angle) are used, and in turn they contribute to the filtering process. 4.3. the proposed filter structures the vector directional distance-rational hybrid filters (vddrhfs) are promising detail preserving filtering structures since it was shown that every subfilter is able to preserve signal details within their subwindows karakos and trahanias, (1997). vddrhfs are grouped into two classes: unidirectional and bidirectional vddrhfs. the structures for a 3x3 window unidirectional and a bidirectional vddrhfs are shown in figs. 1(a) and 1(b), respectively. only the points indicated in black in each window mask are used in the corresponding operation. unidirectional vddrhfs are designed to preserve 5 the journal of engineering research vol. 2, no. 1 (2005) 1-11 [ ] ( )23131 ,, φφ=φφ θd figure 1. structures of vddrhf. (a) the unidirectional structure, (b) the bidirectional structure (a) (b) image details along the vertical, horizontal and the two diagonal directions. therefore, the samples of the same value neighborhood must be located along those directions in order to preserve the center sample by unidirectional vddrhfs. also, bidirectional vddrhfs can preserve details within the two corresponding directions in one operation. the central subfilter is a center weighted vector directional distance filter characterized by its high detail preservation capability. one of the following three sets of weights can be used depending on the noise properties and the image details gabbouj et al. (1990). mask m1 emphasizes details in the horizontal and vertical directions, while m2 the two diagonal directions. on the other hand, mask m3 seeks details in all of these directions simultaneously. 5. simulation results vddrhf have been evaluated, and their performance has been compared against those of some widely known vector nonlinear filters: the vector median filter (vmf), the distance directional filter (ddf), and the generalized vector directional filter (gvdf) (karakos and trahanias, 1997), using rgb color images. the noise attenuation properties of the different filters are examined by utilizing two color images: (1) part of pepper image (256x256 pixels); unit (2) the rose image (240x150 pixels). the test images have been contaminated using various noise source models in order to assess the performance of the filters under different scenarios: * gaussian noise: n(0,σ 2) * impulsive noise: each image channel is corrupted independently using salt and pepper noise. we assume that both salt and pepper are equally likely to occur. * mixed gaussian-impulsive noise: the impulsive noise is fixed (salt and pepper 2% in each image channel), while the variance of the gaussian noise is varied. the original images, as well as its noisy versions, are represented in the rgb color space. this color coordinate system is considered to be objective, since it is based on the physical measurements of the color attributes. the filters operate on the images in the rgb color space. a number of different objective measures can be utilized for quantitative comparison of the performance of the different filters. these criteria provide some measure of closeness between two digital images by exploiting the differences in the statistical distributions of the pixel values (eskicioglo et al. 1995). the most widely used measures are the mean absolute error (mae), and the mean square error (mse) defined as: (15) (16) notwithstanding the rgb is the most popular color space used conventionally to store, process, display and analyze color images, the human perception of color cannot be described using the rgb model (pratt, 1991). consequently, measures such as the mean square error defined in the rgb color space is not appropriate to quantify the perceptual error between images. it is therefore important to use color spaces, which are closely related to the human perceptual characteristics and suitable for defining appropriate measures of perceptual errors between color vectors. a number of such color spaces are used in areas such as multimedia, video communications (e.g., high definition television), motion picture production, printing industry, and graphic arts. among these, perceptually uniform color spaces are the most appropriate to define simple yet precise measures of perceptual errors. the commission internationale de l'eclairage (cie) standardized two color spaces, l*u*v* and l*a*b*, as perceptually uniform (gonzales and woods, 2002). conversion from rgb to l*a*b* color space is explained in detail in (gonzales and woods, 2002). rgb values of both the original noise free and the filtered image are converted to corresponding l*a*b* values for each of the filtering methods under consideration. in the l*a*b* space, the l* component defines the lightness, and the a* and b* components together define the chromaticity. in l*a*b* color space, we computed the normalized color difference (ncd) (plataniotis and androutsos, 1998) which is estimated according to the following expression: (17) where ∆elab is the perceptual color error between two color vectors and defined as the euclidean distance between them, given by: 6 the journal of engineering research vol. 2, no. 1 (2005) 1-11 ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ ⎟ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎜ ⎝ ⎛ 111 131 111 : 101 030 101 : 010 131 010 : 321 mmm ∑ ∑ = = −= m i n j jiji dymn mae 1 1 1,, 1 ∑ ∑ = = −= m i n j jiji dymn mse 1 1 2 2,, 1 ∑ ∑ ∑ ∑ ∆ = = = = = m i n j lab m i n j lab e e ncd 1 1 * 1 1 where m, n are the image dimensions, jiy , is the vector value of pixel (i, j) of the filtered image, jid , is the corresponding pixel in the original noise free image, and ||.|| 1, ||.||2 are the −1l and −2l vector norm s, respectively. (18) where ∆l* , ∆a* , and ∆b* are the differences in the l*, a*, and b* components, respectively. e*lab is the magnitude of the original image pixel vector in the l*a*b* space and given by: the results obtained are shown in the form of plots in figs. 2-4 for the three noise models: gaussian, impulsive, and gaussian mixed with impulsive, respectively. the simulation results of two vddrhf structures are very close to each other (slight differences), we hence reported only those provided by the bidirectional structure given by fig. 1(b). as can be verified from the plots, the vddrhf filters provide better results than those obtained by any other filter under consideration. recall that vddrhf filter uses no information about the type and the degree of noise corruption. moreover, consistent results have been obtained when using a variety of other color images and the same evaluation procedure. 7 the journal of engineering research vol. 2, no. 1 (2005) 1-11 ( ) ( ) ( ) 5.0222 ]***[ balelab ∆+∆+∆=∆ ( ) ( ) ( ) 5.0222 ]***[* bale lab ++= figure 2. comparative results for the color test images contaminated by gaussian noise, (a) pepper image, (b) rose image σ2 (a) (b) 8 the journal of engineering research vol.2, no. 1 (2005) 1-11 figure 3. comparative results for the color test images contaminated by impulsive noise (salt and papper). (a) pepper image (b) rose image) the filtered images are presented for visual assessment, since in many cases they are, ultimately, the best subjective measure of the efficiency of image processing techniques. figures 5(a) and 5(b) show the noise free and the corrupted rose image by mixed noise (impulsive 2% in each channel and gaussian n (0,50), respectively. figures 5(c) and 5(f) represent the filtered images using ddf, gvdf, vmf and vddrhf, respectively. all the filters considered operate using a 3x3 square processing window. an additional sample processing results of pepper image are presented in figs. 6(a) and 6(f). figure 6(a) shows a part (128x128) of original color pepper image and fig. 6(b) shows the corrupted version of fig. 6(a) with additive (4% in each channel) impulsive noise. figures 6(c) and 6(f) show the results using ddf, gvdf, vmf and vddrhf, respectively. a comparison of the images clearly favors the proposed vddrhf over its counterparts (vmf, gvdf, ddf). the proposed vddrhf can effectively remove impulses, smooth out nominal noise and keep edges, details and color uniformity unchanged as we can see from the related error measures summarized in the plots. 9 the journal of engineering research vol. 2, no. 1 (2005) 1-11 (a) figure 4. comparative results for the color test images contaminated by impulsive mixed noise (salt and pepper 2% in each component, and gaussian with zero mean and variable variance), (a) papper image, (b) rose image (b) 10 the journal of engineering research vol. 2, no. 1 (2005) 1-11 figure 6. (a) and (b) are parts of the original noise free and the contaminated pepper image with additive (4% in each channel) impulsive noise, respectively. images (c) (f) are the results using ddf, gvdf, and vddrhf, respectively figure 5. (a) and (b) are the original noise free and the contaminated rose image by mixed noise (impulsive 2% in each channel and gaussian with zero mean and variance 50), respectively. images in (c)-(f) are the results using ddf, gvdf, and vddrhf, respectively • (a) (f) (e) (d) (c) (b) (a) (b) (c) (d) (e) (f) 6. conclusions this paper introduced a new class of nonlinear filters for multichannel image restoration. the vddrhf filters are two-stage filters. they exhibit very desirable filtering properties and utilize in an effective way the performance of the vector rational function filters and the features of vector directional distance filters. simulation results and subjective evaluation of the filtered images demonstrated the robustness of the vddrhf under different noise distributions and have indicated that the vddrhfs outperform all other filters under consideration. moreover, the results have shown that vddrhfs have achieved three main objectives: noise attenuation, chromaticity retention and, edge and detail preservation. acknowledgments the author wishes to thank sultan qaboos university (squ) for supporting the project and providing the software development (from squ project grant # ig/eng/eced/03/05). references astola, j., haavisto, p. and neuvo, y., 1990, "vector median filter," proceedings of the ieee, vol. 78, pp. 678-689. astola, j. and kuosmanen, p., 1997, " fundamentals of nonlinear digital filtering," crc press, boca raton, new york. eskicioglo, a.m., fisher, p.s. and chen, s., 1995, "image quality measures and their performance," ieee transactions on communication, vol. 43, pp. 29592965. gabbouj, m., haavisto, p. and neuvo, y., 1990 "recent advances in median filtering," communications, control and signal processing, e. arikan, ed. ankara, turkey: elsevier science publishers b.v., vol. ii, pp. 1080-1094. gonzales, r.c. and woods, r.e., 2002, "digital image processing," second edition, prentice hall. haralick, r. m. and shapiro, l.g., 1992, "computer and robot vision," vol.1, addison-wesley, reading, ma. karakos, d.g. and trahanias, p.e., 1997, "generalized multichannel image-filtering structures," ieee transactions on image processing, vol. 6(7), pp. 1038-1045. khriji, l. and gabbouj, m., 2001, "adaptive order statistics rational hybrid filters for multichannel image processing," ieice transactions on fundamentals of electronics, communications and computer sciences, (special section: noise reduction and cancellation techniques), vol. e84-a (2), pp. 422-431. khriji, l., cheikh, f. a. and gabbouj, m., 1999, "high resolution digital resampling using vector rational filters," j. of optical engineering, spie, special issue on sampled imaging systems, vol. 38 (5), pp. 893-901. leung, h. and haykin, s., 1994, "detection and estimation using an adaptive rational function filter," ieee transactions on signal processing, vol. 42(12), pp. 3366-3376. lukac, r., 2002, "adaptive impulse noise filtering by using center-weighted directional information," proceedings of cgiv'2002, pp. 86-89. machuca, r. and phillips, k., 1983, "applications of vector fields to image processing," ieee transactions on pattern analysis and machine intelligence, vol. pami-5, pp. 316-329. pitas, i. and venetsanopoulos, a.n., 1990, "nonlinear digital filters: principles and applications," kluwer academic, norwell, ma. plataniotis, k.n. and venetsanopoulos, a.n., 2000, "color image processing and applications," springer verlag, germany. plataniotis, k.n., androutsos, d. and venetsanopoulos, a.n., 1998, "adaptive multichannel filters for colour image processing," signal processing: image communication, vol. 11, pp. 171-177. pratt, w.k., 1991, "digital image processing," john wiley and sons, new york. trahanias, p.e., karakos, d.and venetsanopoulos, a.n., 1996,"directional processing of color images: theory and experimental results," ieee transactions on image processing, vol. 5(6), pp. 868-880. 11 the journal of engineering research vol. 2, no. 1 (2005) 1-11 furthermore, it is worth mentioning that the proposed filter has comparable or less computational complexity to those used in the comparison, particularly the vdf. the computationally intensive part of the algorithms is the distance calculation part. however, this step is common in all multichannel algorithms considered here. the vector rational operation in the second stage does not introduce significant additional computational cost. in the absence of any fancy or fast algorithms, the number of compa rators used in the median filter with a window of size n is 2 )1( − = nn nc . according to f igs. 1(a)-(b), the first stage of the vddrhf requires 41 comparators: 10 comparators for ( )511 =φ n , 10 comparators for ( )533 =φ n and 21 comparators for ( )722 =φ n with the weights indicated by m1). the second stage requires for the denominator, one multiplication, three additions and one division per output sample. in addition, the different subfil ters can be run in parallel reducing the execution time and making the new filters suitable for real -time implementation with digital signal processors. therefore, our design is simple, does not increase the numerical complexity of the multichannel algorit hm and delivers excellent results for complicated multichannel signals, such as color images. the journal of engineering research ii (2005) 00-00 mech010203_current.qxd 1. introduction the study of projectile impact against targets or structures and structural components that undergo large plastic deformation have become main concerns for military and civilian applications (goldsmith, 1972). weights dropped, during maintenance or operation, on pipelines that transport dangerous materials, or on offshore platforms, and falling rocks in mountainous regions against shields or bridges, constitute examples of impact loading. these cases, as well as the design of shields themselves, against bullets or missiles, with much more complex behavior with reference to deformation modes, have been reported in the literature (goldsmith, 1972; davies, 1984a; davies, 1984b; reid, 1985; jones, 1989; zhao;2000). extensive theoretical and experimental research has been conducted in order to establish the basis for the design of selected structures (awerbush, and bonder, 1974; beckman and goldsmith, 1978; corbett et al., 1997; _____________________________________________ *corresponding author e-mail:azhari_n@yahoo.com corbett and reid, 1993; corran et al., 1983; jones and shen, 1993; langsen, and larson, 1990; shen, and jones, 1993; wen and jones, 1993; wen and jones, 1994;). however, more experimental, theoretical as well as numerical research is still needed for the study of plastic collapse mechanisms of targets against projectiles. results were reported (wen and jones, 1993; wen and jones, 1994; corbett and reid, 1993) for a series of tests conducted on the quasi-static and dynamic local loading of circular plates resting on a ring support. a criterion for the inelastic rupture of ductile metal beams subjected to large dynamic loads was formulated and presented in (jones and shen, 1993). the results of experiments on pipelines with their ends fully clamped were presented in (jones et al., 1992), whereas the same type of experimental investigation was reported in (wen and jones, 1993; wen and jones, 1994) for clamped circular plates loaded by transverse impact in both concentric and eccentric loading positions. the finite element method (fem) has been used extensively to simulate many applications in structural dynamring research 1 (2004) 59-74rnal of engineethe jou plastic deformation of thin plates subjected to quasi-static and dynamic loadings a. a.n. aljawi* king abdulaziz university, college of engineering, p.o. box 80204, jeddah 21589, saudi arabia received 1 february 2003; accepted 11 november 2003 abstract: deformation and failure of thin plates of mild steel were studied under quasi-static and dynamic impact loadings. particular emphasis was placed on responses of simply supported circular plates subjected to centric orthogonal loadings. the latter comprised loadings due to relatively massive rigid cylindrical strikers with a hemispherical-end as well as a flat-end. the projectile motions featured variable and low impact velocities. generally, good agreement was found between experimental results and those predicted by finite-element techniques for displacement-time curves and for force histories of the striker. it was concluded that the abaqus-based study (both the implicit and the explicit versions) revealed beneficial insights into the impact mechanics of plates by rigid projectiles. keywords: thin plate, fem, orthogonal impact, quasi-static gg:¢ü∏îà°ùÿÿòkéj â– π«∏– ôn’gh âhék ¬ñ°t π«∏– éªg~mg ú∏«∏– ωg~îà°séh é¡à°sgqo ” …ô£dg ö∏°üdg øe áyƒæ°üÿg á≤«bôdg íféø°üdg π°ûah √ƒ°ûj ájõcôÿg i~eé©àÿg ∫éªm’g .±gƒ◊g øe ioƒæ°ùÿg ájôfg~dg íféø°üdg ≈∏y ájõcôÿg i~eé©àÿg ∫éªm’g òkéj ≈∏y ¢uén õ«côj ∑éæg ¿éc .᫵«eéæj~dg äée~°üdg iò¨àe ωg~£°ueg äéyô°s á«ahò≤ÿg äécô◊g ärôhg .á룰ùe ±gƒm ägp iôngh ájhôc ∞°üf ±gƒm ägp á∏«≤k á«fgƒ£°seg ¥qé£e øy áä°téædg ∂∏j πª°ûj qé°ùÿ iôncg h âbƒdg ™e iò¨àÿg ámgre’g äé«æëæÿ ò¨°üdg a…õ÷g á«æ≤j ≥jôw øy á©bƒàÿg ∂∏jh á«ñjôéàdg èféàædg úh ~«l ≥agƒj ∑éæg ¿éc .á° øîæeh ≈∏y á«æñÿg á°sgq~dg ¿cg úñj .¥qé£ÿg øy áªléædg iƒ≤dgabaqusáñ∏°üdg äéahò≤ÿg ωg~£°ueg 鵫f鵫e ‘ áj~› í°vgƒdg h »æª° dg ¬«≤°ûh .íféø°üdéh :á«méàøÿg ägoôøÿg .âhék ¬ñ°t ,…~eé©àdg ωoé°üàdg ,ò¨°üdg a…õ÷g á«æ≤j ,á≤«bq áë«ø°u ics (aljawi, 1999; bammann et al.,1993; hks, 1998; kormi et al., 1994; kormi et al., 1993, kormi and webb, 1993; sun et al., 2000). to this end, a number of algorithms were developed and implemented in fem codes. preliminary studies were conducted by the present author on a system comprising a simply supported circular plate subjected to centric orthogonal loading, prior to the commencement of full-scale analysis, whereby responses from an fem code were compared with published data and other experimental results (shen, and jones, 1993). experimentally measured data (wen and jones, 1993; wen and jones, 1994; corbett and reid, 1993) were utilized for material properties in the fem analysis. encouraging results were obtained from these exploratory studies. following these preliminary analyses, the present study was initiated, where abaqus code (both the implicit and the explicit versions, ver. 5.8) were employed to investigate the deformation of a ductile material under dynamic loading. the developed model was used to study the response of a circular mild steel plate subjected to centric loading. in an effort to limit the rather wide scope of the study, results of work concerning only two cases are presented here. these are quasi-static loading and loading with a relatively large mass, traveling at several low initial velocities. both cases employed one striker with a hemispherical end and another striker with a flat face. the material properties used in the model were linear in the elastic range. they featured both linear and non-linear isotropic work-hardening characteristics in the plastic regime, being strain dependent for the dynamic case. a cowper-symonds type of power law (symonds, 1965) was utilized to model the strain-rate dependence of the material. comparisons of results obtained from both the implicit and the explicit versions of abaqus on the nonlinear quasi-static and dynamic cases were also reported. 2. quasi-static loading with hemisperical ended and flat-ended strikers 2.1. finite element model a two-dimensional axi-symmetric discretized model, shown in figure 1a was considered. the model, which represents a circular plate of constant thickness t, loaded at its center, as shown in figure 1b, consisted of four parts. these are a) the plate, b) the rigid surface representing the striker, c) the mass element associated with the striker, and d) a contact link that allows the energy to be transferred from the striker to the plate using the surface interaction. for the quasi-static case no mass was considered for the striker. figure 1a shows half of the plate to be considered for mesh generation. a 4-noded axi-symmetric continuum element suitable for large deformation plasticity was used (cax4) for this purpose, while the rigid body (striker) was modeled by 2-noded rigid elements (rax2). the number of elements was 70 along the radius of the plate, while 14 elements were selected across the thickness. the particular number of elements across the radius and thickness were decided as dictated by the size of the mesh analysis. the boundary condition imposed on the model constrained all nodes on the left half of the model (plane of symmetry in figure 1a) to only move perpendicular to the model. additional boundary conditions were needed that (a) (b) figure 1. (a) discretized 2-d axi-symmetric model, (b) uniform simply supported plate and striker 60 aljawi /the journal of engineering research 1 (2004) 59-74 are consistent with the model used. additionally, constraints were imposed on the reference node located at the tip of the rigid body, the striker. it is to be noted that this rigid striker, shown in figure 1, can carry the relatively large mass element of the striker, and generates the impact loading of the plate. 2.2. quasi-static loading of simply-supported steel plate in this investigation mild steel was used for modeling (a) (b) plate thickness figure 2. (a) load-displacement curves for the hemisperical-faced striker, (b) thinned deformed shapes, for plate thickness of 3,5,6,8 and 10 mm plate thickness, t (mm) d/t yield stress, yσ (n/mm 2 ) ultimate stress, uσ (n/mm 2 ) young’s modulus, e (n/mm 2 ) 3 100 273 440 228 5 60 352 472 220 6 50 305 446 235 8 37.5 342 472 211 10 30 251 490 230 table 1. material properties for quasi-static case (corbett and reid, 1993) 61 aljawi /the journal of engineering research 1 (2004) 59-74 (a) (b) figure 3. perforated, (a) abaqus, and (b) experimental (corbett and reid, 1993), deformed shapes, for plate thickness of 3, 5, 6, 8 and 10 mm (a) (b) plate thickness figure 4. (a) load-displacement curves for flat-end striker, (b) deformed shapes, for plate thickness of 3, 5, 6, 8 and 10 mm 62 aljawi /the journal of engineering research 1 (2004) 59-74 figure 5. comparison of the load-displacement curves between the experimental (corbett and reid, 1993) and abaqus analysis, for hemispherical and flat-ended striker of plate thickness 10 mm plate striker contact zone figure 6. 3-dimensional discretised model 63 aljawi /the journal of engineering research 1 (2004) 59-74 the plate of 300 mm in diameter, with 3, 5, 6, 8, and 10mm in thickness t. material properties and dimensions of the plate were taken from the experimental investigation reported in (corbett and reid, 1993 ) and given in table 1. in this case, for all calculations, a piecewise elastic-plastic material model was used, with poisson’s ratio, v=0.3, and mass density, p =7830 kg/m3. note that the plates are simply-supported at a diameter of 248 mm, and loaded by a striker of 12.7 mm diameter (figure 1b). the numerical simulation was carried out using the implicit standard version of abaqus 5.8. the computational comparison between both implicit and explicit algorithms will be discussed later. figure 2a shows the load-displacement curves when the hemispherical-end striker was used. as may be expected for quasi-static loading, the displacements were highly dependent on the diameter to thickness ratio (d/t). the results shown in figures 2(a) and 2(b) are qualitatively in good agreement with those reported in (corbett and reid, 1993 ). figure 2(b) shows the sectioned profiles of the plate. both figures indicate that the process involved continuous loading up of the punch load with displacement to perforation with remarkable thinning of the plate in the vicinity of the loaded area. for a plate thickness of 3 mm, figures 2(a) and (b) show that a membrane-dominated mode of collapse took place. this indicated a continuous indentation of the plate leading to thinning. this effect became less pronounced as plate thickness increased. consequently, a knee or change of slope can clearly be observed in figure 2(a) for the 6, 8 and 10 mm plates, indicating that bending-dominated plastic collapse mode took place. once more, membrane effects also took place after plastic collapse. it is worth noting that the plate material under the area of the punch thinned to about 40-60% of the original plate thickness. the explicit method in finite element analysis code enables a failure criterion to be incorporated into the solution procedure using the effective plastic strain, i.e. the plastic strain generated by the deviatoric stress state. this implies that failure is dependent on the von-mises stress and hence the failure occurrence is unaffected by the prevailing spherical stress state. the algorithm employed allows removal of elements from the mesh as a result of failure due to tearing and/or ripping of the structure. abaqus explicit version allows for the automatic removal of elements from the mesh when the equivalent plastic strain exceeds the specified value. the material behaves elastically before initial yield is reached. plastic strain then occurs following the conventional von-mises plasticity theory. if strain continues to increase, damage will remain, as long as the plastic strain is less than or equal to the offset plastic strain. however, if the plastic strain exceeds the offset plastic strain, damage will increase linearly, until the equivalent failure plastic strain is reached. at this strain level the material is considered to have failed. figure 3(a), where all elements which have an equivalent plastic strain less than one have been removed, clearly indicates the continuous loading of the plate up to perforation, as well as the good agreement with the experimental findings as shown in figure 3(b). load-displacement curves as well as sectioned profiles of the plates that are defined in table 1, when loaded with the flat-end striker, are shown in figures 4(a) and (b). it may be verified from inspection of figures 4(a) and 4(b) that global deformation behavior of the plates are similar to the plates which are loaded with the hemispherical-end striker. what is observed is that, there is a continuous loading up of the punch load with increasing displacement, until perforation occurs; furthermore, the governing trend seems to be membrane-dominated mode of collapse for thin plates, and bending-dominated plastic collapse for thicker plates. however, load-displacement curves indicated that lower loads as well as lower deflections are required for perforation using a flat-end striker than their counterparts with the hemispherical-end striker. figure 5 shows the comparison of the load-displacement curves between the experimental results reported in (corbett and reid, 1993 ) and the current finite element model analysis, for both the hemispherical-ended and the flat-ended strikers when these are acting on a plate of thickness 10mm. it can be verified from figure 5 that good agreement exists between the finite element predictions and the experimental findings. material mild steel specimen identification * stiii-26 yield stress, óy (n/mm 2) 261.7 ultimate stress, óu (n/mm 2) 417.7 åf 0.2769 a (n/mm2) 541.5 b 0.04 n 0.2259 table 2. material properties calculated from given measured values (wen and jones, 1993, 1994) effective stress, σ (n/mm 2 ) effective strain, ( ε %) 261.7 314.3 347.3 372.1 392.3 409.4 424.4 429.6 437.0 449.8 460.9 9.9 5.0 10.0 15.0 20.0 25.0 30.0 31.9 35.0 40.0 45.0 table 3. material isotropic work hardening characteristics (wen and jones, 1993) 64 aljawi /the journal of engineering research 1 (2004) 59-74 3. dynamic loading of clamped steel plates with flat-end striker an experimental investigation was reported in (wen and jones, 1993; wen and jones, 1994 ) for clamped circular plates loaded (in both concentric and eccentric loading positions) by cylindrical strikers of large masses with blunt ends, traveling at speeds up to 5.0 m/s. the case taken up in the present study was for a striker mass of 81 kg, traveling at a speed of 4.43 m/s. the plate was 6mm in thickness, 228.6mm in diameter, clamped at a diameter of 152.4 mm, and loaded by a flat-end striker 17.85 mm in diameter. in the current investigation only mild steel was used for modeling the plate. 3.1. material properties material properties of the plate in the elastic range were taken as modulus of elasticity, e = 2.01 n/mm2, poisson’s ratio, v = 0.3, and mass density of the material, p =7830 kg/m3 the material was considered to be linear in the elastic range, and the nonlinear stress-strain behavior in the plastic range obeyed an expression of type: [1] where values of a, b, and n were based on experimental data (wen and jones, 1994), and are listed in table 2. thus, nonlinear work hardening characteristics of the material, based on the data given in table 2, are presented in table 3 that the first row of stresses correspond to the first row of strains, column by column, respectively. experimental findings (wen and jones, 1993; wen and jones, 1994) indicated that a plate structure that is subjected to an in-plane mass impact, even within the range of low velocity impacts, is highly sensitive to strain rate and inertia effects. it is both velocity and mass sensitive. the strain rate can be expressed using cowpersymonds (symonds, 1965) empirical power law: [2] where, ε p, is the equivalent plastic strain, εs (ε p) is the quasi-staic yield stress, σ (ε p) denotes the dynamic yield stress, ε is thestrain rate, and d and p are material constants. for mild steel d=40.5s-1 and p=5. however, it was pointed out in (symonds, 1965) that dynamic stress-strain curves become flatter with an increase in strain-rate, with reduced material strain hardening. therefore, it was suggested that for mild steel, d=1300s-1 and p=5, or d=300s1 and p=2.5 can also be used (jones, 1983). 3.2. modeling both 2-d and 3-d models were used for the dynamic loading conditions. the 2-d model was the same as in the quasi-static case, figure 1. the 3-d, discretized model, shown in figure 6, consisted of four parts. these are a) the plate, b) the rigid surface representing the striker, c) the mass element associated with the striker, and d) a contact link that allows the energy to be transferred from the striker to the plate. figure 6 shows half of the symmetrical plate as discretized. due to symmetry, only one-quarter of the plate is modeled. the half-plate model however, would enable comparisons for the case of eccentric impact loading. the mesh shown in figure 6 is symmetrical. it is uniform except in the impact zone, where the mesh is much denser than elsewhere. three and four node shell elements (s3, s4r) were used in the discretized model, such that internal angles of the 4-node shell elements were in the range of 45o 135o to get better results. the striker is visible in figure 6 at the impact zone. the rigid striker was modeled as a cylindrical body with either a hemispherical-end, or a flat-end. for the hemisphericalend striker, the outermost segment of the end, containing the point of contact of the tip of the striker with the plate, had a radius of 8.925 mm. the radius of the second circular arc defining the hemispherical tip was also 8.925 mm. for the 3-d model, as shown in figure 6, a line drawn parallel to the surface axis from the end of the second arc, and revolved 360 degrees about the central axis of the striker, generated the cylindrical body of the striker. an auxiliary node was attached to the tip of the striker. the utility of the auxiliary node was to represent the mass of the striker. 3.3. boundary conditions and constraints for the 3-d model, the boundary conditions imposed on the model were consistent with the existence of a diametrical plane of symmetry and a fully clamped outer circumference, figure 6. linear multi-point constraints were also imposed along the boundary lines joining the unrefined and refined meshes. the latter action allows the nodes on both sides of the line of symmetry to respond equally. additionally, constraints were imposed on the reference node located at the tip of the rigid body, the striker. it is to be noted that this rigid striker carried the relatively large mass element of the striker, and generated the impact loading on the plate. 3.4. loading for the 3-d model, impact loading was applied on the plate by energizing a 40.5 kg mass element, representing a total striker mass of 81 kg for the whole plate, with an initial velocity of 4.43 m/s. a mass of 81 kg was applied in the 2-d model. the motion of the striker was in a direction orthogonal to the plane of the plate. gravity forces were ignored, since these were considered to be relatively 65 aljawi /the journal of engineering research 1 (2004) 59-74 small. in order to allow the energy to be transferred from striker to the plate, an appropriate contact zone was established between the contacting surfaces. no friction was introduced in the contact zone, and also no failure criterion was incorporated into the solution procedure. 3.5. adaptive meshing adaptive meshing is one of the tools that is available in abaqus/explicit version 5.8. it enables the model to maintain a high quality mesh automatically by allowing the mesh to move independent of the underlying material even when it is subjected to large deformations. with adaptive meshing, node and element numbering and configure 7. comparison of the displacement time response of flat-end striker between the experimental (wen and jones, 1993) and abaqus 2-d model for different values of d and p figure 8. comparison of velocity v/s time response of flat-end striker between the experimental (wen and jones, 1993) and abaqus 2-d model for different values of d and p 66 aljawi /the journal of engineering research 1 (2004) 59-74 nectivity remain constant throughout the analysis, despite the fact that nodal locations change with respect to the material. however, the same elements adopt to remain well shaped even as the underlying material changes shape. although adaptive meshing is only available for first-order, reduced–integration continuum elements, it was imposed throughout the 2-d model. 3.6. discussion of results the response of the flat-end striker to orthogonal impact with the plate is described in figure 7 by global deformations, with the possibility of a local failure at the region of impact. the results show the predictions due to the 2-d model, with no strain rate (ind), as well as the experimentally (exp) measured values (wen and jones, 1993; wen and jones, 1994). figure 8 displays the variation of velocity of the auxiliary node attached to the rigid striker surface with time (ind) for a striker velocity of 4.43 m/s. the same figure also displays the experimentally (exp) measured values (wen and jones, 1993; wen and jones, 1994). note that the fem predictions (ind) were obtained when strain rate dependence was not included in the analysis. the curves identified by d=1300, d=300 ( su et al.,1995) and d=40.5 in figures 7 and 8 were obtained when strain rate dependence was included in the analysis for d=1300s-1 and p=5, d=300s-1 and p=2.5, and d=40.5s-1 and p=5, respectively. with reference to figure 7, it was observed that the experimentally measured maximum forward displacement of the striker (and hence of the plate) is reached after 5.8 msec. at this time the striker has already ruptured the plate, but it does not have the potential to go through. thereafter the plate motion is stopped, and then reversed, as may be observed by the falling slope in figure 7. this reversal may be explained in terms of the stored recoverable strain energy within the plate, causing the plate to oscillate, such that the striker perhaps separates from the plate, and bounces back in the reverse direction. it is interesting to note that finite element analysis predicts this phenomenon correctly. thus referring to figure 7, when strain rate dependence was not included in the analysis (ind), the striker penetrates and passes through the plate, causing local damages by rupture. it does not rebound. however, when strain rate dependence was considered in the analysis, trends that are similar to the experimental results can be clearly observed. thus in the case of dcurves in figure 7 the striker does not penetrate through the plate, signaling that the material is strongly strain rate dependent. this is particularly true for the case when d=300s-1 and p=2.5. this observation agrees with the study reported in ( su et al.,1995). more interestingly, lower values of the strain rate constant d tends to lower the contact time of the striker with the plate, indicating that strain hardening properties of the material also influence the development of local plastic deformation, especially during the early stage of deformation. the phenomenon revealed by figure 7 can be also witnessed from a study of the velocity-time curve of figure 8. experimentally, the velocity of the striker (exp) is seen to decrease monotonically with time until it becomes zero at time = 5.8 msec. the striker then rebounds, and bounces back at a constant velocity of about 1 m/s. the d-curves figure 9. comparison of load-displacement curves of flat-end curves of flat-end striker between experimental (wen and jones, 1993) and abaqus 2-d mode for different values of d and p 67 aljawi /the journal of engineering research 1 (2004) 59-74 [x10-3] (a) (b) (c) (d) figure 10. equivalent plastic strain contour, peeq, for elements in vicinity of impacted zone for (a) strain-rate independent, (b) d = 1300s-1 and p = 5, (c) d = 300s-1 and p = 2.5, (d) d = 40.5s-1 and p = 5 figure 11. model kinetic energy curves, ke, and total energy curves due to plastic deformation, pd, of flat-end striker for different values of d and p 68 aljawi /the journal of engineering research 1 (2004) 59-74 figure 12. comparison of load-displacement curves for the flat-end striker, between 2-d and 3-d models, at v = 4.43 m/s, for d = 300s-1 and p = 2.5 (a) (b) (c ) (d) figure 13. contact pressure contour at time: 1.04, (b) 4.16, (c) 6.16 and (d) 7.28 msec. (d = 300s-1 and p = 2.5) 69 aljawi /the journal of engineering research 1 (2004) 59-74 (a) (b) (c ) (d) figure 14. vos mises stress contour at time: (a) 1.04, (b) 4.16, (c) 6.16, and (d) 7.28 msec. (d = 300s-1 and p = 2.5 figure 15. load-displacement curves for quasi-static case with flat-end striker using implicit and explicit (adaptive and non-adaptive meshing) methods 70 aljawi /the journal of engineering research 1 (2004) 59-74 in figure 8 indicate that the striker actually rebounds with a constant velocity when the strain rate effect is taken into consideration. it may be noticed in the same figure that when strain rate is not taken into account (ind), the striker decelerates upon impact, and then continues to travel in the same direction. this is a clear indication that the striker actually goes through the plate, according to this prediction. the variation of the striker load with displacement is shown in figure 9 for the same cases indicated as in figures 7 and 8. here the load is computed as the product of mass times the acceleration of the reference node attached to the rigid surface. figure 9 shows that, as the striker comes into contact with the plate, strong oscillations are induced in the plate. furthermore, lower loads were observed when no strain rate was taken into account during the analysis. it is also evident from figure 9 that, for the case when d=300s-1 and p=2.5, the fem analysis results closely matches with the experimental findings. closer observation in figure 9 shows that lower values of strain rate hardening modulus (d) tend to increase the load required for deformation, and also tend to decrease the contact time of the striker with the plate. figure 10 shows the contours of equivalent plastic strain within the impact zone at time t=8.0 msec for the cases treated in figures 79. as it would be expected, with no strain rate effect (a), the striker has actually gone through the plate, and the plate is perforated. cases (b) to (d) depict the cases with strain rate effects, where all elements, which have equivalent plastic strain greater than one, have been removed. it may be pointed out that cracking may occur in the plate. plugging of the plate is also possible, especially with d=300s-1 and p=2.5. although not shown here, it was found that, with d=300s-1 and p=2.5, a small increase in the initial speed to 4.98 m/s tended to increase the punch load as well as the contact time, and consequently causing more energy to be dissipated due to plastic deformation. the time variation in kinetic energy (ke), and energy dissipated due to plastic deformation (pd), are shown in figure 11 for the cases of strain-rate dependent (d-curves) and strain-rate independent (ind) materials. as expected, all of the curves indicated that, as kinetic energy decreases with time, most of the energy is being transferred to the plate, and consumed for rupturing and plastic work dissipation. then kinetic energy tends to increase where considerably less energy is available for the striker. furthermore, with strain rate effect, more energy is dissipated due to plastic deformation for d=300s-1 and p=2.5 than for other values. this is the case in which the contact time of the striker with the plate is longer than for d=1300s-1 and p=5, and for d=40.5s-1 and p=5. figure 12 shows comparisons of load-displacement curves between the 2-d and 3-d models, using the explicit method. note that adaptive meshing was not included in the analysis since it is not available for shell elements. both models were observed to yield similar trends. the 3d model however, depicted a higher peak load and lower contact time than the 2-d model. it must be emphasized here that the 2-d model is rather limited in the amount of insight it provides. the 3-d model was therefore utilized in an effort to acquire more precise information regarding changes on the surface of the plate due to impact loading within the impact zone; figure 16. dynamic load-displacement curves for the flat-end striker, using implicit and explicit (adaptive and non-adaptive meshing) methods, at v = 4.43 m/s 71 aljawi /the journal of engineering research 1 (2004) 59-74 contours of the contact pressure. the results presented in figure 13 for four different stages of impact, i.e., at the time when the load at the beginning of its monotonic increase, at its peak, at the instant when it becomes zero, and during the rebound. inspection of figure 13 shows that the contour profiles do exhibit the symmetry of the plate. it can be observed that, during the impact, the contours of the constant pressure between the striker and the plate is localized, and that they are concentrated in the vicinity of the outer radius of the striker. the contact pressure disappears, however, during the period of rebound. contours of the von-mises equivalent stresses shown in figure 14 indicate that the growth of the plastic zone commences from the right side to the vicinity of the impact, and grows gradually outward in a monotonic fashion. 4. explicit versus implicit since it is possible to study quasi-static nonlinear problems as well as linear dynamic problems by the use of both the implicit and the explicit versions of abaqus (su et figure 17. comparison of the acceleration time responses of flat-end striker, using implicit and explicit (adaptive and non-adaptive meshing) methods, at v = 50.0 m/s figure 18. comparison of the velocity time responses of flat-end strike, using implicit and explicit (adaptive and non-adaptive meshing) methods, at v = 50.0 m/s 72 aljawi /the journal of engineering research 1 (2004) 59-74 al., 2000; hks, 1994), it would be of interest to compare the results of the two. furthermore the comparisons can be extended to quasi-static and dynamic nonlinear problems in which adaptive meshing can be taken into consideration. normally the explicit method is limited to short transient problems. in order to be able to run the explicit method for quasi-static problems, it would be necessary to make the inertia effect negligibly small by allowing the striker to move at a very low constant speed. below both the explicit and the implicit methods were used for the analysis of clamped circular plate of section 3 loaded by a flat-end striker. for the quasi-static case, and using the explicit method (with and without adaptive meshing), the striker was allowed to move at a uniform speed of 0.05mm/s while it traveled a distance of 16mm. a fixed mass scale factor of 50 was used in order to increase the computational speed of explicit analysis. using the implicit method however, no speed controls were applied on the striker while traversing the same distance. it took only 177.19 s of cpu time for the implicit solution, while it required 3 h,13 min, and 26 s of cpu time for the explicit non-adaptive, and 3 h, 18 min, and 4 s of cpu time for the explicit method with adaptive meshing. it was found that cpu time is inversely related to striker velocity for computation by the explicit version, whereas cpu time for the implicit solution is not affected by striker velocity. figure 15 displays the load-displacement curves for quasi-static loading using the implicit and explicit (with adaptive and non-adaptive meshing) methods. there seems to be good agreement between the implicit and the explicit non-adaptive meshing solutions. the large discrepancy of explicit solution with adaptive meshing can clearly be observed for higher loads, that is, when plastic deformation on the plate becomes severe, especially in the vicinity of the edge of the striker. adaptive meshing does not seem to adopt itself to large deformation effects on the mesh. although not shown here, similar large discrepancies were observed for adaptive meshing when comparing stresses and strain contours in the vicinity of the striker edge. figure 16 shows the variation of load-displacement curves obtained by the implicit and explicit (with adaptive and non-adaptive meshing) methods when the striker was given an initial velocity of 4.43 m/s and a transient period of 8.0 s. it may be verified from figure 16 that the implicit method results in more oscillations than those obtained using the explicit method at the beginning of impact. it took 1041.2 s of cpu time for the implicit method, using a haftol (half tolerance) value of 120000 (higher than the reaction force). for more accurate results, a decrease in haftol, to a value of 10% of the reaction force, would significantly increase the cpu time. nevertheless, the cpu time of the implicit solution is still small when compared to 2228 s and 2142 s required by the explicit method, with and without adaptive meshing, respectively. increasing the initial striker velocity to 50 m/s and considering a transient period of 0.5s, results in a cpu time of 580 s for the implicit method using a haftol value of 120000, while for the explicit method it takes 381 s and 240 s, with and without adaptive meshing, respectively. variations of striker acceleration and of velocity with time for this case are displayed in figures 17 and 18, respectively. it may be verified from these figures that, according to the implicit analysis, the striker remains in contact with the plate, while it passes through the plate according to the explicit analysis. in summary, the implicit method seems to be best suited for quasi-static and slow contact problems. the explicit method excels in the speedy analysis of impact problems with high speed and low transient periods. the latter method has the added capability to utilize adaptive meshing and the possibility to incorporate a failure criterion. 5. conclusions standard implicit and explicit versions of the abaqus (ver.5.8) code were used to investigate the behavior of simply supported and clamped circular plates when subjected to centric orthogonal impact loadings. strikers having hemispherical-end and flat-end rigid cylinders induced the loadings. one case reported here dealt with quasi-static loadings. dynamic loadings involved a relatively high striker mass, and a number of low impact striker velocities. material properties that were utilized in the fem modeling of the plate were obtained from experiments conducted under dynamic load conditions. material properties were linear in the elastic range, and exhibited nonlinear isotropic work-hardening characteristics in the plastic regime. they were strain dependent. the agreement between the experimental and computed values was note worthy. the implicit methods gave better results for quasistatic and slow contact problems. acknowledgments the author would like to express his appreciation to professor m. akyurt for his useful comments and help during the preparation of this manuscript. references aljawi, a. a., 1999, “penetration and failure of thin plates subjected to impact loading,” third international conf. on shock & impact loads on structures 99”, 24-26 november, 42-50, 1999, singapore. awerbush, j., and bonder, s. r., 1974, “analysis of the mechanics of perforation of projectiles in metallic plates,” int. j. solid structures. vol. 10, 671-684. bammann, d. j., chjiesa, m. l., horstemeeyer, m. f., and 73 aljawi /the journal of engineering research 1 (2004) 59-74 weigaten, l. t, 1993æ “failure in ductile materials using finite element methods, structural crashworthiness and failure”, ed n. jones, and wierzbick, elsevier applied science, london and new york, 1-54. beckman, m. e., and goldsmith, w., 1978,“the mechanics of penetration of projectiles into targets,” int. j. engng. sci, vol. 16, 1-99. corbett, g. g., and reid, s. r.,1993, ”quasi-static and dynamic local loading of monolithic simplysupported steel plate,” int. j. impact. engng. vol. 13, no. 3. pp. 423-441. corbett, g. g., reid, s. r., and johnson, w., 1997, “impact loading of plates and shells by free-flying projectiles: a review,” int. j. impact engng, vol. 18, no. 2, 141230. corran, r. s. j., shadbolt, p. j., and ruiz, c., 1993, “impact loading of plates an experimental investigation,” int. impact engng, vol. 1, no. 1, 3-22. davies, g. a. o., 1984a, ”structural impact and crashworthiness,” volume 1, elsevier applied science publishers, london and new york. davies, g. a. o., 1984b, ”structural impact and crashworthiness,” volume 2, elsevier applied science publishers, london and new york. goldsmith, w., 1972, “impacts,” edward arnold, london. hks, inc., 1994, “application of implicit and explicit finite element techniques to metal forming,” j. mater. process. techno., 45, pp. 649-656. hks, inc, 1998, “abaqus/explicit user’s manual, theory and examples manuals and post manual, version 5.8, explicit,”. jones, n., 1989, “structural impact,” cambridge university press. jones. n, 1983, “structural aspect of ship collisions,” in structural crashworthiness, chapter 11, pp. 308-337 (edited by n. jones and wierzbicki). butterworth. jones. n., birch., s,e., birch., c, s., xhu, l., and brown, m., 1992, “an experimental study on the lateral impact of fully clamped mild pipes,” proc. inst. mech. engrs., vol. 206 (e), 111-127. jones, n., and shen, w. q., 1993, “criteria for the inelastic rupture of ductile metal beams subjected to large dynamic load,” structural crashworthiness and failure, ed. n jones and t. wierzbicki, elsevier applied science publishers, london and new york, 95-130. kormi, k., shaghoueie., and duddell, d.a., 1994, “finite element examination of dynamic response of clamped beam grillages impacted transversely at their center by a rigid mass,” int. j., impact engng, vol. 15, no. 5, 687-697. kormi, k., webb, d. c., and etheridge, r. a.e., 1993, “the use of fem to evaluate the response of damaged pipespart 1-static loading ‘, pvp-vol. 264, piping, support and structural dynamics, 147-158, asme. kormi, k., and webb, d. c., 1993, “the use of fem to evaluate the response of damaged pipespart 2dynamic loading ‘, pvp-vol. 264, piping, support and structural dynamics, 159-167, asme. langsen, m., and larson, p. k., 1990, “dropped object plugging capacity of steel plates: an experimental investigation,’ int. j. impact engng, vol. 9, no. 3, 289316. reid, s. r., 1985, “metal forming and impact mechanics,” pergamon press ltd. shen, w. q., and jones, n., 1993,“dynamic response of a grillage under mass impact,” int. j. impact engng, vol. 123, 555-565. su, x. y., yu, t. x., and reid, s. r., 1995, “inertiasensitivity impact energy-absorbing structures. part ii: effect of strain rate,” int. j. impact. engng, vol. 16, no. 4, pp. 673-689. sun, j. s., lee, k. h., and lee, h.p., 2000, “ comparison of implicit and explicit finite element methods for dynamic problems,” j. mater. process. technology, 105, pp. 110-118. symonds, p. s., 1965, “viscoplastic behavior in response of structures to dynamic loading,” behavior of materials under dynamic loading (ed by n. j., huffington), 106-124, sme, new york. wen, h.m, and jones, n., 1993, “experimental investigation of the scaling laws for metal plates struck by large masses,” int. j. impact. engng, vol. 13, no. 3, pp. 485-505. wen, h. m., and jones, n., 1994, “experimental investigation into the dynamic plastic response and perforation of a clamped circular plate struck transversely by a mass,” proc. ints. mech. engrs, i. mech. e., vol. 208, 113-137. zhao, x. l., and grzebieta, 2000, ”structural failure and plasticity,” proceedings of the 7th int. sym. on structual failure and plasticity, implast 2000, pergamon. 74 aljawi /the journal of engineering research 1 (2004) 59-74 ring research 1 (2004) 59-73rnal of engineethe jou c:\users\raoh\desktop\paper_1.xps the journal of engineering research vol. 8 no. 2 (2011) 1-9 1. introduction nowadays, industrial organizations complexity exists as an immense international interest and knowledge for the scientific basis. studying the importance of reducing complexity into industrial organizations was recommended as one of several solutions to __________________________________________ *corresponding author’s e-mail:garbie@squ.edu.om recovery the existing financial recession (garbie 2009 and 2010; garbie and shikdar 2009, 2010 and 2011). it was defined as systemic characteristics which integrate several key dimensions of the industrial environment including size, variety, information, uncertainty, control, cost, and value (kamrani and adat 2008). flexibility and agility are considered as the most desircomplexity analysis of industrial organizations based on a perspective of systems engineering analysts i. h. garbie* and a. a. shikdar department of mechanical and industrial engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, al khoud, muscat, oman received 4 april 2011; accepted 18 september 2011 abstract: complexity in industrial organizations became more difficult and complex to be solved and it needs more attention from academicians and technicians. for these reasons, complexity in industrial organizations represents a new challenge in the next decades. until now, analysis of industrial organizations complexity is still remaining a research topic of immense international interest and they require reduction in their complexity. in this paper, analysis of complexity in industrial organizations is shown based on the perspective of systems engineering analyst. in this perspective, analysis of complexity was divided into different levels and these levels were defined as complexity levels. a framework of analyzing these levels was proposed and suggested based on the complexity in industrial organizations. this analysis was divided into four main issues: industrial system vision, industrial system structure, industrial system operating, and industrial system evaluating. this analysis shows that the complexity of industrial organizations is still an ill-structured and a multi-dimensional problem. keywords: complexity, industrial organization, systems engineering 2 i.h. garbie and a.a. shikdar ability of certain system properties for the industrial organizations. these properties will give industrial enterprises more ability to cope with increased environmental uncertainty and adapting to the faster pace of change of today's markets (giabchetti et al. 2003). the industrial organizations are sometimes viewed as an intrinsic structural property of the system. the structural property is defined as how individual system components relate to each other and how the relationship determines overall system behavior (arteta and giabchetti, 2004). sometimes, manufacturing systems complexity considers the product variety. this means increasing in the product variety increases in the complexity in the manufacturing systems (kuzgunkaya and elmaraghy 2006). also there is another classification of industrial organizations complexity: time-independent complexity and time-dependent complexity. time -independent complexity is used to add the complexity arising from the designer's perception while time-dependent complexity is either combinational or periodic (kuzgunkaya and elmaraghy 2006). there are several concepts such as: product, process, and operation complexity can be used as a component in industrial enterprises. the product complexity focuses on product features and specifications while the process complexity analysis focuses on the tools, equipment and operations used to manufacture it (hu et al. 2008). supply chain management complexity such as: upstream complexity, internal manufacturing complexity, and downstream complexity are considered as complexity issues regarding manufacturing systems (bazarth et al. 2009). manufacturing strategy also plays an important role in complexity in industrial enterprise such as just-in-time manufacturing (lean manufacturing), flexible manufacturing, cellular manufacturing, agile manufacturing, concurrent engineering, etc. the complexity measures are defined as not only intrinsic to the system being studied but also depend on extrinsic properties of the observer. although most measurements were concentrated on operational measures, both structural and operational characteristics are important to the performance of the system as a whole. operational complexity measures the uncertainty associated with the material and information flows of the system (arteta and giabchetti 2004). the complexity levels in industrial firms are estimated through several case studies based on general framework which includes a questionnaire focusing on each issue in a firm (garbie and shikdar 2010). this paper is organized into several sections. section 1 presents the importance of complexity concept. section 2 reviews previous research work about the complexity. analysis issues are explained in section 3. section 4 presents a hypothetical example. conclusions and recommendations for further work will be introduced in section 5. 2. literature review several research works have been published in this area but none of work mentioned the concept of design for complexity. a methodology based on a simulation model to analyze the complexity in mixed-model assembly production systems was suggested by (kamrani and adat 2008). a measurement framework to analyze the structural properties of the enterprise system was presented (giabchetti et al. 2003). a complexity measure was developed for the business process level inside the organization on one product (e.g. the prepaid phone card) (arteta and giachetti, 2004). a new metrics for assessing the structural complexity of system configurations was suggested (kuzgunkaya and elmaraghy 2006). they based on machine complexity, buffer type complexity, and material handling system complexity as a structural complexity although they can be considered as an operational or dynamical complexity. product, process, and operational complexity in modeling and assessment of manufacturing complexity were introduced (elmaraghy and urbanic 2003, 2004; elmaraghy et al. 2005). each assessment was evaluated independently. three different types of complexity were suggested (bazarth et al. 2009) to represent and model supply chain complexity such as: upstream complexity, internal manufacturing complexity, and downstream complexity. they used these complexities to study the impact effect on a manufacturing plant performance. a coding system of machines, buffers, and material handling equipment to measure complexity based on time-independent complexity (static or structural complexity) of those major components was classified and designed by (elmaraghy et al. 2005). the structural (static) complexity measure to evaluate the complexity in mixed assembly lines was used (hu et al. 2008). several case studies involving 14 italian companies were conducted (perona and miragliotta 2004) to investigate how complexity can affect a manufacturing company's performance. an analytical model for measuring system complexity was presented based on information entropy and probability distribution of resource allocations (cho et al. 2009). operational complexity was measured as a function of cost through supply chain management systems (wu et al. 2007). they indicated that inventory costs are associated with operational complexity. an entropic related to complexity measures to quantify the complexity associated with information content of schedules and variations between schedules was used (huatuco et al. 2009). yang (2010) presented a computational approach to investigate effecting of scheduling with processing 3 complexity analysis of industrial organizations based on a perspective of systems engineering analysts time on two stage hybrid flow shop systems. a mathematical model was suggested to identify functional requirements and the associated design parameters regarding complexity level (tomiyama et al. 2007). kashyap and sinha (2011) studied the complexity related to the mental fatigue required of a person for doing a specific job. while they used how to manage stress, they developed a general model to estimate overall complexity of a profession although they tried to evaluate the job complexity of an engineer. mazur and chen (2011) presented organizing work, communication and managing conflict as the most important issues among team members to complete the project. they considered multifunctional knowledge, teamwork capabilities and working relationships with organizing work as a complex problem. 3. analysis issues to identify the main issues of industrial organizations complexity (ioc), there are five important questions to be asked such as the following to describe how the complexity of industrial enterprises can be studied. these questions were presented by (garbie and shikdar 2011) in analyzing and estimating the complexity levels in industrial organizations. 1. how the complexity issues of an industrial organization is identified and analyzed? 2. how the complexity level of an industrial organization is estimated? 3. how can an industrial organization reduce its complexity? 4. which issues are more important than others? 5. how can industrial organizations identify the adverse factors for reducing complexity? based on these questions, full analysis of complexity regarding these issues will be analyzed and explained through the recommended issues from perspective of systems engineering analysts. these can be represented into four main phases as follows (fig. 1). figure 1 shows a briefly industrial organizations complexity procedure consisting of four phases. each phase will be discussed with its associated issues. it can be noticed from fig. 1 that the procedure of analyzing industrial organizations complexity follows the four phases parallel. based on these concepts and issues mentioned in the previous sections, it can be noticed that the industrial organization complexity (ioc) consists of major issues. these issues are: industrial organization vision complexity (iovc), industrial organization design complexity (iodc), industrial organization operating complexity (iooc), and industrial organization evaluating complexity (ioec) (as shown in fig. 1). the mathematical model of industrial organization components and the corresponding complexity relationships between them in order to emphasize on particular vision, design, operating, and evaluating is presented in the following equations (1 and 2). as each component or element in these systems is a potential source of uncertainty (due to its state), the measuring of complexity for each one is highly valuable. then, iocl is clearly modeled as the following eq. (1) as a function of previous sub-complexities. eq. (1) can be rewritten as eq. (2). each term represents sub-complexity measure of complexity measure of industrial organization (iocl). adding these terms with relative weights is considered. these weights can be used as a reason existing to differentiate between major issues of complexity. iocl = f (iovc, iodc, iooc, ioec) (1) iocl = wiovc [iovc] + wiodc + wiooc [iooc] + wioec [ioec] (2) figure 1. four phases for design for industrial enterprises complexity 4 i.h. garbie and a.a. shikdar where: iocl = industrial organization complexity level, iovc = industrial organization vision complexity, iodc = industrial organization design complexity, iooc = industrial organization operating complexity, ioec = industrial organization evaluation complexity. wiovc, wiodc, wiooc, and wioec are relative weights of organization vision, organization structure, organization operating, and organization evaluation, respectively. because the trade-offs frequently exist between these objectives, a comprehensive analysis for each individual measure is needed. the value of these relative weights may reflect the system analyst's subjective preferences based on his/her experience or can be estimated using tools such as analytical hierarchy process (ahp). in this paper, the relative weights using the ahp are estimated and changed frequently according to the new circumstances by decision maker or a group of decision makers (abdi 2006). these groups are represented in senior management level, manufacturing and/or production engineers, plant managers, operators, and suppliers. these relative weights can be estimated using ahp according to the next matrix. for example, suppose , then this means that industrial organization design complexity (iodc) is four times more important than operating complexity (iooc). 3.1 phase 1: industrial organizations vision complexity the industrial organizations vision complexity is the first step in the design for industrial enterprises complexity. the major issue of this phase is how to collect the main components (elements) of industrial organizations vision. the organization vision usually specifies what supply chain management (scm) representing in number of suppliers (nos), demand variability (dv) representing in number of customers (noc), introducing s new product (np), product life cycle (plc) representing also in product development (pd), and time to market (ttm) requires and how they are affected and effecting on the complexity of industrial organizations. the industrial organizations vision complexity (ievc) will be represented mathematically as a function of these issues as shown in the following eq. (3). ievc = f (scm (nos), dv (noc), np, plc (pd), ttm) (3) 3.2 phase 2: industrial organizations design complexity the second phase, industrial organizations design complexity (iodc) procedure is the designing for system complexity itself. it is mainly concerned with different elements to represent the complexity of it. these elements are: product structure and design (psd), system design (sd), and manufacturing strategies (ms). for each main element, there are several sub-main elements which play an important role in the value of complexity. for example, the psd has four different types to represent the complexity in the product design such as number of parts per product (nnp), number of operations per part (nop), processing or manufacturing time per operation (pt), product size and weight (psw). all of them have a significant effect on the complexity of manufacturing/production process. system design (sd) is playing a major role in complexity in industrial organizations. it can be observed how complexity is the analysis of the system design (sd). the sd divides the complexity analysis into three major issues: production system size (pss), material handling system (mhs), and plant layout system (pls). for the pss, there are three different classifications of production system: small-sized production system (ssps), medium-sized production system (msps), and large-sized production system (lsps). each classification type represents or introduces a significant effect on complexity. also, the material handling system (mhs) and plant layout system (pls) play an important role in identifying the complexity in industrial organizations. the mhs consists of material handling equipments (mhe) with different types of equipments (e.g. conveyor, trucks, forklifts, crane, etc.), material handling storage system (mhss) (e.g. manual or automated storage and retrieval), and identification systems (is). how it can be thought about all previous components and degree of complexity related to each one. facility planning or plant layout system (pls) has a significant effect on complexity by different types of configuration. it can be seem that functional layout (fl) or process layout is more complex than product layout (pl) and/or cellular layout (cl). complexity with respect to manufacturing strategies (ms) is totally different than (sd) and (psd) because it was looking for which strategy can be applied. complexity in lean manufacturing (lms) for example is affected by increased product variety if it is comparing with product layout (pl) in mass production system. but in general, complexity in lean manufacturing system (lms) is low comparing it with agile manufacturing system (ams), flexible man5 complexity analysis of industrial organizations based on a perspective of systems engineering analysts ufacturing system (fms), and reconfigurable manufacturing system (rms). regarding complexity in ams, fms, and rms, they can deal easily with any changes in product design (modifications), unpredictable demand, etc. but infrastructure of these systems itself is more sophisticated and complicated to lead these systems to be more and more complex. the mathematical expression of industrial organizations design complexity (iodc) can be modeled as the following eqs. (4 and 5) in different facets. iodc = f (psd, sd, ms) (4-1) iodc = wpsd (psd) + wsd (sd) + wms (ms) (4-2) iodc = f (npp, nop, pt, psw, mhs, pss, pls, lms, fms, ams, rms) (5) the wpsd wsd and wms are relative weights of product design, system design and manufacturing strategies, respectively. eq. (5) can be rewritten as a very board general representing the lowest level of information in industrial organization design complexity as the following eq. (6). (6) 3.3 phase 3: industrial organizations operating complexity once the industrial organization system vision and design complexities become available to the perspective of industrial systems analysts and designers, the design for complexity related to the system operating becomes urgent to be analyzed and evaluated. this phase may involve further activities in data collection and processing. design for industrial organizations operating or dynamic complexity (iooc) is different than previous ones (system vision, and system design). in this analysis, it can be noticed that there are three major items of complexity: resource status of operating complexity (rsoc), work in progress complexity (wipc), and business operations complexity (boc). resources mean equipment (e.g. machining equipment, forming equipment, material handling equipment, etc.) and human. in this analysis, it will be concentrated on the resource reliability (rr), resource capability or flexibility (rc), resource utilization (ru), resource scheduling/ rescheduling (rs/r), and human scheduling/rescheduling (hs/r). for example, maintenance level plays a vital role in resource reliability. this means that the lower the maintenance level is the lower in machine capacity (reliability). also, work in progress complexity (wipc) representing in buffer between workstations or departments is considered one of measuring degree of complexity inside the production plant (factory). there are several important issues that can be used to evaluate the business operations complexity (boc). these issues are: organization plans (op), organizing work (ow), structure of management levels (sml), staffing developing and motivation (sdm), decision making (dm), communication between and within management levels (cml), managing conflict, change, culture and stress (mcs), and finally leadership roles in management (lr). then, industrial organizations operating complexity (iooc) can be modeled to measure or evaluate the complexity level as the following eqs. (7 and 8). iooc = f (rsoc, wipc, boc) (7-1) iooc = wrso (rso)+wwip (wip)+wboc (boc) (7-2) iooc = f (rr, rc, ru, rs/r, hs/r, wip (bs), op, ow, sml, sdm, dm, cml, mcs, lr) (8) 3.4 phase 4: industrial organizations evaluation complexity the fourth phase in the design for industrial organizations complexity (ioec) procedure is the complexity regarding the system evaluation. as industrial organizations have a great impact on the performance measurements, they still have a problem in measuring these complexities especially regarding selection of the objectives. in this paper, there are five different objectives that can be used to evaluate the complexity. they are: product cost (pc), response (r) representing in manufacturing lead time, system productivity (sp) representing in system utilization, product quality (pq) representing in number of scrap (defect rate), and appraising and rewarding performance (arp). they also can be modeled mathematically as the following eq. (9). ioec = f (pc, r, sp, pq, arp) (9) 4. a hypothetical example this numerical example is used to estimate the level of complexity of industrial organizations. this can be implemented through three sequential steps. the first step is used to estimate the relative weights of industrial organizations based on the four major issues, iovc, iodc, iooc, and ioec. the relative weights between these issues are estimated by using the ahp as the following matrix based on the pair wise comparisons of the four major issues. it can be noticed that the relative weights of iovc is estimated to be equivalent to the iodc, twice as important as the iooc and four times more important than the ioec. the iodc is estimated to be three times more important than the ieoc and four times more important than ioec. the 6 i.h. garbie and a.a. shikdar iooc is also estimated to be twice as important as the ieec. as a result, the values of relative weights are estimated at 0.35, 0.40, 0.16, and 0.09 for iovc, iodc, iooc, and ioec, respectively. eq. (2) is used to calculate the industrial enterprises complexity level (iocl) incorporating of the four major issues as the following eq. (10). iocl= 0.35 [iovc] + 0.40 [iodc] + 0.16 [iooc] +0.09 [ioec] (10) in the second step, the multiple regression models are used to estimate and formulate the complexity levels for the major and sub-major issues. the values were assumed to follow a uniformly distributed random variable with known parameters as shown as in table 1 for 50 generating values for each issue. the multiple regression models for industrial organizations vision complexity, industrial organizations design complexity (product structure and design, system design and manufacturing strategies), industrial organizations operating complexity (resource status and business), and industrial organizations evaluation complexity are presented as the following eqs. (1117), respectively. this was done by using minitab statistical software package to generate a random variable and formulating a multiple regression model. ievc = -0.03560.0374 (nos) 0.000013 (noc) + 0.0065 (pd) + 0.0842 (ttm) (11) psdc = 0.359-0.00436 (npp) + 0.0353 (nop) + 0.0123 (pt) + 0.0759 (ind1) 0.0420 (ind 2) + 0.0241 (ind 3) 0.0039 (ind 4) ind 1, ind 2, ind 3, and ind 4 are used for the product size and weight. (12) sdc = 0.2440 + 0.00479 (ssps) -0.00795 (msps) + 0.00018 (lsps) + 0.0848 (fl) 0.0681 (cl) + 0.0609 (pl) + 0.092 (mhss) + 0.0290 (mhe) + 0.189 (is) (13) msc = 0.441 -0.059 (lms) + 0.245 (fms) 0.028 (ams) 0.024 (rms) (14) rosc = 0.5980.379 (rr) + 0.509 (rc) + 0.113 (ru) 0.50 (rs/r) + 0.233 (hs/r) (15) boc = 0.980 0.0029 (op) 0.0076 (ow) + 0.0015 (sml) + 0.171 (s) +0.234 (d) + 0.0050 (m) 0.0266 (dm) 0.0642 (cml) 0.134 (mcs) 0.289 (lr) (16) ioec = 0.457 + 0.00210 (pc) + 0.0088 (r) 0.172 (sp) 0.40 (pq) + 0.117 (arp) (17) the third step is used to determine the complexity level of iovc, psdc, sdc, rosc, and ioec, respectively by optimizing eqs. (11-17). the constraints are identified based on the range values of each sub-issue which are listed in table 1. to determine the relative weights between psdc, sdc, and msc and between rsoc, wipc, and boc, the following matrices are used to estimate these values and the pair wise comparison between product structure and design, system design and manufacturing strategies is illustrated in these matrices. with respect to iodc, it can be noticed that a psdc is estimated to be equivalent to the sdc and twice as important as a msc. the same estimation is done relate to the sdc with psdc and msc. regarding the iooc, it can be noticed that rsoc is estimated to be four times more important than the wipc and two times as important as the boc. the boc is also estimated to be four times more important than wipc. as a result, the relative weights for iodc between psdc, sdc, and msc are estimated at 0.40, 0.40, and 0.20, respectively. also, the relative weights for iooc between rsoc, wipc, and boc are estimated at 0.544, 0.110, and 0.345, respectively. the results of complexity level in each sub-issue and major issue are illustrated in table 2. then, eq. (10) is used again to calculate the global complexity level based on the information taken from table 2 and the estimated relative weights as the following. iocl = 0.35 [0.4582] + 0.40 [0.3099] + 0.16 [0.3023] + 0.09 [0.4570] = 0.3740 it can be noticed from the results given from the previous equation and table 3 that the level of complexity in this enterprise equals to 37.40% and this value seems ranked in a medium range. it seems that vision complexity represents more important (0.16/0.3740 = 42.78%) than design complexity (33.15%). the percentage values of operating and evaluating complexity are 13.36% and 10.70%, 7 complexity analysis of industrial organizations based on a perspective of systems engineering analysts table 1. data for complexity issues based on uniform distribution, u [a, b] u represents the uniform distribution with [a] the lower limit and [b] is the upper limit, *represents the product size (small, small-medium, medium, medium-large, large), *'represents the product weight (light, light-medium, medium, medium-heavy, heavy), **represents the degree of variety in functional layout, **'represents the number of manufacturing cells, **''represents the variety of models (single, batch, and mixed), ***represents the appraising and rewarding performance as a percentage. 8 i.h. garbie and a.a. shikdar respectively. these values can be totally different from an industrial organization to another one based on the sub-major issues and the relative weights between submajor and major issues. regarding the complexity reduction, it can be noticed that if the values of submajor issues reduced, the complexity level will be reduced too. it also seems that the vision complexity and design complexity represent the adverse factors for reducing complexity. with respect to iovc, it is not easily to reduce the number of suppliers, for example, and it should increase this number. also, for iodc, the number of parts per product, number of operations per part, processing time per operation, product size and weight, production system size are representing adverse factors for reducing complexity because they cannot be changed. 5. conclusions it can be noticed from this analysis that complexity issues are not simple. it required emphasize on each of the main issues and the sub-main. hence, industrial organizations complexity (ioc) will involve four major issues: vision complexity, design complexity, operating complexity, and evaluation complexity. analyzing complexity in industrial enterprises is based on the four main issues and it can be mathematically expressed as a total global function as the following eq. (18). ioc = f (scm (nos), dv (noc), np, plc (pd), ttm, nnp, nop, pt, psw, mhe, mhss, is, ssps, mdps, slps, fl, cl, pl, lms, fms, ams, rms, rr (ml), rc < ru, rs / r, hs, / r, wip (bs), op, ow, sml, dm, cml, lr, sdm, mcs, pc, r, sr, pq, arp) (18) the ioc issues should be dynamic and they should evolve with and adapt to the changing internal and external environment. until now, the ioc remains a research topic of immense international interest. this will represent the degree of freedom of industrial organizations designers to identify which issue is more significant than others. the main contribution in this paper is how to identify and model the components of industrial enterprises complexity in any industrial firms (organizations) at any time considering these components. the authors intend to extend this research to apply this analysis and formulation the suggested model to estimate and optimize the degree of complexity in any industrial enterprises towards full validation of the complexity theory which will be discussed and presented in the future research. acknowledgments the authors would like to acknowledge the financial support provided by the sultan qaboos university (grant no. ig/eng/mied/10/01) to carry out this research work. table 2. complexity levels in the four phases of industrial enterprises table 3. values of major issues regarding the complexity level 9 complexity analysis of industrial organizations based on a perspective of systems engineering analysts references abdi mr (2006), performance evaluation of reconfigurable manufacturing. the proceedings of the 36th cie conference on computer and industrial engineering, taipei, taiwan 2479-2491. arteta bm, giachetti re (2004), a measure of agility as the complexity of the enterprise system. robotics and computer-integrated manufacturing 20:495-503. bazarth cc, warsing dp, flynm bb, flynn ej (2009), the impact of supply chain complexity on manufacturing plant performance j. of operations management 27:78-93. cho s, alamoudi r, asfour s (2009), interactionbased complexity measures of manufacturing systems using information entropy. int. j. of computer integrated manufacturing 22(10): 909922. elmaraghy wh, urbanic rj (2003), modeling of manufacturing systems complexity. annals of the cirp 53(1):363-366 elmaraghy wh, urbanic rj (2004), assessment of manufacturing complexity. annals of the cirp, 53(1):401-406. elmaraghy ha, kuzgunkaga o, urbanic rj (2005), manufacturing systems configuration complexity. annals of the cirp 54:445-448. garbie ih, shikdar a (2009), determination of complexity levels in industrial firms. proceedings of the int. conference of industrial engineering research (ierc), miami, florida, usa, may 30june 3:1423-1428. garbie ih (2009), a vision for reconfiguring industrial organization due to the global recession. proceedings of the 39th of int. conference on computers and industrial engineering, france, 658-663, july 6-8, 2009. garbie ih, shikdar a (2010), complexity level in industrial firms: case studies and implementation. proceedings of the 2010 international conference on industrial engineering and operations management (ieom 2010), dhaka, bangladesh, january 9-10, 2010. garbie ih (2010), a roadmap for reconfiguring industrial enterprises as a consequence of global economic crisis (gec). j. of service science and management 3(4):419-428. garbie ih, shikdar a (2011), analysis and estimation of complexity level in industrial firms. int. j. of industrial and systems engineering 8(3):1-23. giabchetti re, martinez ld, saenz oa, chen cs (2003), analysis of the structural measures of flexibility and agility using a measurement theoretical ramework. int. j. of production economics 86: 47-62. hu sj, zhu x, koren wy (2008), product variety and manufacturing complexity in assembly systems and supply chains. annals of the cirp 57: 45-48. huatuco lh, efstathiou j, calinescu a, sivadasan s, kariuki s (2009), comparing the impact of different rescheduling strategies on the entropic-related complexity of manufacturing systems. int. j. of production research 47(1):4305-4325. kamrani ak, adat a (2008), manufacturing complexity analysis: a simulation-based methodology-collaboration engineering. theory and practice edited by karmrani ak springer, 227-248. kashyap n, sinha s (2011), estimation of job complexity in an industrial organization. int. j. of industrial and systems engineering vol. 7(1):2644. kuzgunkaya o, elmaraghy ha (2006), assessing the structural complexity of manufacturing systems configurations. int. j. of flexible manufacturing systems 18:145-171. mazur lm, chen sj (2011), a task-member assignment model for complex engineering. int. j. of industrial and systems engineering 7(1):1-25. perona m, miragliotta g (2004), complexity management and supply chain performance assessment. int. j. of production economics 90:03-115. tomiyama t, amelio vd, urbanic j, elmaraghy hw (2007), complexity of multi-disciplinary design. annals of the cirp 56(1):85-188. wu y, frizelle g, efstathiou j (2007), a study on the cost of operational complexity in customer-supplier systems. int. j. of production economics 106:217-229. yang j (2010), a new complexity proof for the twostage hybrid flow shop scheduling problem with dedicated machines. int. j. of production research 48(5):531-1538. c:\users\raoh\desktop\paper_7.xps the journal of engineering research vol. 8 no. 1 (2011) 61-73 1. introduction the reliability of any structure is the probability of its survival under possible extreme loadings. the if structure is a concrete barrier that lies at a certain depth in the soil, and a missile (a rigid projectile) impacts the top of the soil cover normally, and subsequently after penetrating the soil cover completely hits the barrier with certain striking velocity then the reliability of the barrier can be described as having the probability of not getting penetrated the the missle. in a deterministic sense, if the estimated depth of penetration is small compared to the thickness of the con_______________________________________ *corresponding author’s e-mail: nadeem@ksu.edu.sa crete barrier, the barrier may be considered safe. however, this safety is not absolute safety but "probable safety ". it is "probable safety" because all material and geometric properties have some inherent variability, which makes determination of absolute safety almost impossible. an effort can only be made to design the barrier with a desired reliability for a given range and type of missile impact. in the recent past, (choudhary et al. 2002; siddqui et al. 2002, 2003 and siddiqui 2003) carried out reliability assessments of different types of target under missiles, projectiles or jet aircraft impacts. a simplified procedure for reliability estimation of underground concrete barriers against normal missile impact n. a. siddiqui department of civil engineering, king saud university, riyadh, saudi arabia received 6 jul 2008; accepted 27 june 2010 abstract. underground concrete barriers are frequently used to protect strategic structures like nuclear power plants (npp), deep under the soil against any possible high velocity missile impact. for a given range and type of missile (or projectile) it is of paramount importance to examine the reliability of underground concrete barriers under expected uncertainties involved in the missile, concrete, and soil parameters. in this paper, a simple procedure for the reliability assessment of underground concrete barriers against normal missile impact has been presented using the first order reliability method (form). the presented procedure is illustrated by applying it to a concrete barrier that lies at a certain depth in the soil. some parametric studies are also conducted to obtain the design values which make the barrier as reliable as desired. keywords: missile, soil-penetration, projectiles, concrete barriers, structural reliability; form 62 the journal of engineering research vol. 8 no. 1 (2011) 61-73 determination of the reliability of underground concrete targets very much depends on estimation of an damages (eg. penetration) caused in these targets due to missile impact. in the past, (forrstal and luk 1992; siddiqui and abbas 2002 and siddiqui et al. 2006) studied mechanics of missile penetration in soil targets and (forrestal et al. 1994, 1996, 2003; frew et al. 1998, 2006 and khan et al. 2003) studied the penetration in concrete targets. in all instances, they determined expressions for depths of penetration and verified the analytical results with experimental findings. a careful review of the literature reveals that although considerable material is available on the mechanics of missile penetration into soil and concrete targets, studies on the reliability assessment of underground concrete targets are very scanty. further, the relevant work, eg. (chaoudhury et al. 2002 and siddiqui et al. 2002) is on concrete targets which are buried under rock. there appears to be no work, however, on the reliability assessment of concrete barriers that are buried under the soil and hit by a missile after completely penetrating the soil cover. with this in mind, the present study, a procedure has been proposed for the reliability assessment of concrete barriers that lie at a certain depth in the soil, and a missile (rigid projectile) impacts the top of the soil cover normally, and, subsequently after completely penetrating the soil cover hits the concrete barrier with a certain striking velocity. for this purpose, using expressions available in the literature, the striking velocity of a missile at any depth of soil has been derived and then expressions for the depths of penetration in the crater and tunnel regions of a concrete barrier have been deduced. these depths of penetration have been employed for the derivation of limit state functions. using the derived limit state functions, the reliability assessment of an underground concrete barrier has then been carried out through the first order reliability method (nowak and collins 2000). the presented procedure was illustrated by applying it to a concrete barrier that lay at a certain depth in the soil. some parametric studies were also conducted to obtain the design values which make the barrier as reliable as desired. 2. problem formulation in the present study, an underground concrete barrier is assumed as failed if a missile completely penetrates the concrete thickness. this is due to the fact that things of importance are kept under these barriers, and a missile can cause considerable damage to them only if it penetrates the concrete barrier completely. to derive limit state functions under the above criterion of failure, first we require expressions for depths of penetration in a concrete barrier due to missile impact. in the present study, these depths of penetration have been derived under the following assumptions and idealizations: 1. the missile is rigid, ie. its deformation is negligible, and only soil and concrete deformations have been considered. 2. the missile is considered to be of ogival nose shape. 3. the impact of the missile is normal and axi-symmetric. 4. the missile does not carry any warhead and so no explosion has been considered. 5. the rear face scabbing of the concrete target, due to missile penetration, is negligible. in the present study we are concerned with the reliability analysis of a buried concrete target which is considered failed if the missile penetrates it completely. hence, conditions that lead to a higher depth of barrier penetration will be a conservative estimate for reliability calculations. assumption (1) and (3) provide higher depths of penetration than if the missile deforms or impacts the target at a certain angle other than normal (ie. oblique impact). for most targets, an ogival-nose-shaped missile penetrates more than conical nose shape missile; hence, assumption (2) is also on the conservative side (siddiqui and abbas 2002). a missile carrying a warhead destroys the things behind the concrete barrier in two stages: first it fully penetrates the concrete barrier and then it explodes its warhead to destroy everything behind the barrier. if the missile fails to penetrate the barrier, the destruction behind the concrete barrier may not be significant. for this reason, in assumption (4), the explosion of the warhead is neglected and the penetration of the concrete barrier has been considered as the failure of the concrete barrier. as the subject of the present study is related to the concrete barrier's penetration, rear face scabbing is ignored in assumption (5). under the above assumptions and idealizations, when a missile impacts a uniform soil target at normal incidence (fig. 1) with an initial velocity vo, and proceeds through the soil cover with velocity vz at any depth z, its nose experiences an axial force fz. forrestal and luk (1992) have developed the following expression for this axial force fz as (1) where, s and s are defined in eq. (a-1) (appendix). from newton's second law the equation for the rigid body motion of a missile with a mass m is given 63 the journal of engineering research vol. 8 no. 1 (2011) 61-73 as (2) when a missile hits a concrete barrier, lying at a depth of z in the soil with a striking velocity vz it creates a conical-shaped crater region with a depth of about two missile shank diameter ie. 4a, followed by a circular cylinder-shaped tunnel region with a diameter nearly equal to the shank diameter ie. 2a, (forrestal et al. 1994, 1996). the depths of penetration in the crater region zsc and tunnel region zft from the concrete surface can be expressed as (3) (4) (5) where, s = a parameter (frew et al. 1998) fc = unconfined compressive strength of concrete; n = dimensionless constant that depends on missile caliber radius head (fig. 3) (6) c = density of concrete target; and vc = rigid body missile velocity at zsc = 4a, it is given by (7) substituting expression of vz from eq. (a-13) in the above equation we get (8) putting this value of vc of eq. (8) in eq. (5), we get c as (9) putting this value of c from eq. (9) in eq. (4) we get (10) substituting eq. (10) for and eq. (a-11) for t into eq. (3) we have (12) if the depth of concrete target is dc, then the limit state function (or safety margin) g(x) can be written as (13) substituting zsc from eq. (11) we have (14) where (x) denotes the vector of random variables given by (15) where, dg = z = depth of soil above the concrete surface (fig. 1). the final depth of penetration in the crater region is where the missile velocity becomes zero subject to a maximum depth of 4a. it is due to the fact that beyond 4a the tunnel region begins and the depth of penetration in this tunnel region can be expressed as (forrestal et al. 1994, choudhury et al. 2002, siddqui 2003). (16) putting the value of vc from eq. (8) in the above equation we get (11) 64 the journal of engineering research vol. 8 no. 1 (2011) 61-73 (17) the limit state function g (x) for this region, measured from the concrete surface, therefore, can be expressed as (18) substituting the expression of zft from eq. (17) in eq. (18) we get the limit state function for the tunnel region as (19) where, (x) denotes the same vector of random variable as given by eq. (15). since we have two limits state functions one for the crater region and other for tunnel region to assign an appropriate limit state function for reliability calculations, first we assume that the missile lies in the crater region, and with this assumption we compute crater depth zsc. if this crater depth is less than the 4a, the limit state function given by eq. (14) is used. otherwise eq. (19) is employed for reliability calculations. 3. reliability estimation having established the limit state function, the next step is the reliability assessment of the buried concrete target against normal missile impact. for this purpose, a reliability technique known as first order reliability method (form) (melchers 1999) has been employed. a brief description of form is provided herein. in brief, in this approach of reliability estimation, the basic random variables xi (i = 1,2,..,n) are first transformed to 'reduced uncorrelated variables' with zero mean and unit variance as (20) w h e r e i n µx and x are the mean and standard deviations of the variables xi. the reliability is measured in terms of a reliability index, , which is defined as the shortest distance to the failure surface from the origin in a zcoordinate system, and it is related to the probability of failure or probability of limit state violation for any limit state as (21) wherein pf is the failure probability and -1(.) is the inverse of standard normal distribution function. the reliability index is found from the solution of the constrained optimization problem: minimize (z) = (zt z)1/2 (22) subject to g (z) = 0 (23) wherein z is a vector of the basic random variables in the standard normal space and g (z) is the limit state function in the standard normal space. the relevant relations as obtained through the lagrange multiplier approach for optimization may be summarized as follows: the most probable failure point/design point/checking point is expressed as (24) wherein i* are the direction cosines and given as (25) missle soil concrete barrier figure 1. problem formulation 65 (26) (27) the reliability index and design points zi* are evaluated by solving the eqs. (24) through (27) iteratively for normally distributed variables. if the design variables do not follow normal distribution, the nonnormal probability distributions may be incorporated in the analysis by transforming the non-normal variables into equivalent normal variables. the parameters of the equivalent normal are estimated using the following relations: (28) (29) wherein fi and fi =nonnormal distribution and density functions of xi; and and = cumulative distribution and probability density functions for standard normal variates respectively. 3.1 numerical algorithm a simple numerical algorithm for form, as presented by (nowak and collins 2000), is provided below. (i) formulate the limit state function. determine the probability distribution and appropriate parameters for all random variables xi (i =1, 2…n) (ii) obtain an initial design point (xi*) by assuming values for n-1 of the random variables xi. mean values are often a reasonable choice. solve the limit state equation g (x) = 0 for the remaining random variables. this ensures that the design point is on the failure boundary. xi* corresponding to nonnormal distribution, determine the equivalent normal standard deviation and mean using eqs. (28) and (29). if one and more xi* values corresponds to a normal distribution, then the equivalent normal parameters are simply the actual parameters. zi*} corresponding to the design point xi* using eq. (20). function with respect to the reduced variates using the equation. (30) define a column vector {g} as the vector whose elements are these partial derivatives multiplied by -1 (31) (vi) calculate an estimate of using t he following relation (32) (vii) calculate a column vector containing the sensitivity factors using (33) figure 2. idealized shear strength-pressure relaslope = p o r l crh = r`/(2r) figure 3. definition of caliber radius head (crh) 66 the journal of engineering research vol. 8 no. 1 (2011) 61-73 (viii) determine a new design point in reduced variates for n-1 of the variables using 34) (ix) determine the corresponding design point values in original coordinates for the n-1 values using the relation (35) (x) find out the value of the remaining variable by solving the limit state function g (x) = 0. (xi) repeat steps (iii) to (x) until and the design point {xi*} converge. 4. numerical study to illustrate the application of the present formulation, a solid missile of (siddiqui et al. 2002) was chosen. this missile had a mass of 182 kg, a shank radius of 0.0825 m and an impact velocity of 411-m/s. other statistical data that are needed for reliability assessment are presented in table 1. with these statistical data and using limit state functions (eqs. 14, 19) the reliability of an underground concrete barrier was evaluated. for this purpose, the algorithm presented above was employed. 5. discussion of results table 2 presents the result of the reliability analysis. the analysis gives the reliability index ( ) of an underground concrete barrier as 1.66 and its corresponding probability of failure (pf) as 4.90 x 10-2. these values show that the underground concrete barrier is not as reliable as generally desired for structures of importance. this is due to the fact that in a probabilistic design of important strategic structures, a reliability index ( ) of less than 3 is never chosen (siddiqui et al. 2003). there is a number of ways to improve the above from 1.66 to a desirable range, eg. by increasing concrete thickness, concrete compressive strength, the shear strength of the soil (by specifying compaction), or by constructing the concrete barrier deeper in the soil. table 1. statistical data and random variables table 2. results of the analysis 67 it is to be noted that the desirable range shown in table 2 was chosen to cover a range which is generally desired to ensure the safety of most important civil engineering structures (siddiqui 2003, siddiqui et al. 2003, khan et al. 2006). however, in practice these values are to be decided for any strucgture on the basis of extensive calibration studies as well as on a consideration of cost issues and the consequences of failure. 5.1 sensitivity analysis a sensitivity analysis was carried out to study the influence of various random variables on the reliability of an underground concrete barrier. this influence is measured in terms of sensitivity factor ( j ), defined above in eqn. (33). the values of sensitivity factors are presented in a bar chart in fig. 4. this figure shows that the sensitivity factor for missile mass, caliber radius head (crh) and impact velocity are positive, indicating that an increase in the magnitude of these variables will decrease the concrete barrier reliability. other random variables are negative, indicating that anincrease in their magnitude will increase the reliability of the underground concrete barrier. further, figure 4. sensitivity diagram 68 the journal of engineering research vol. 8 no. 1 (2011) 61-73 this is due to the fact that as the velocity of a missile increases its penetrating energy increases rapidly whereas with the shank radius of missile resistance offered by soil (or concrete) to missile penetration also increases sharply. 5.2 parametric study as the desired reliability for any important structure should range from 3 to 4, in the present study, some parametric studies were also carried out to illustrate how to achieve the desired reliability by altering the design values of the parameters. for this purpose, a reliability index equals to 3.2 ( t= 3.2 ) was selected and the following parametric studies were conducted to obtain the desired value of the chosen parameter to achieve this reliability. also, for this purpose, the variation of ( t )2 was studied for different nominal values of the selected parameter. the value of the parameter for which ( t )2 is close to zero, yields the value of the parameter which makes the barrier as reliable as desired. 5.2.1 effect of depth of soil cover in the present study, the concrete barrier was placed at 10 m under the soil cover. this overlying soil acted as energy dissipater to the missile. if a concrete barrier is placed in deeper soil, the missile will lose more of its energy before hitting the concrete barrier and thus the reliability of the concrete barrier will improve. figure 5 shows that when depth of the soil cover nomical option then this option may be selected in order to achieve the desired reliability. otherwise, designers may consider some other parameters, discussed below, along with this one to achieve the desired reliability. 5.2.2 effect of unconfined compressive strength of concrete since the unconfined compressive strength of concrete offers resistance to missile penetration, an increase in its value will correspondingly increase the resistance to missile penetration. based on the principle “ the less the penetration the more the reliability", one can say that the barrier reliability will increase with an increase in the compressive strength of concrete. figure 6 shows that to achieve the desired reliability (keeping other parameters same as given in table 1) the compressive strength of the concrete should be at least equals to 70 mpa. if to construct a concrete barrier with compressive strength above 70 mpa is very expensive then to achieve a desirable range of reliability index ( t = 3.2) it is better to increase the strength of the concrete along with other feasible parameters such as the thickness of concrete barrier, depth of the soil cover, shear strength (by specifying compaction) etc. 5.2.3 effect of concrete thickness in the present study an underground concrete barrier is said to have failed if a given missile completely figure 5. effect of depth of soil-cover on the reliability of buried concrete target 69 the journal of engineering research vol. 8 no. 1 (2011) 61-73 desired reliability of 3.2 ( t = 3.2), the concrete thickness should not be less than 3.0 m. as increasing the thickness of concrete barrier is quite an expensive option, it is better to also consider other parameters along with the thickness to reach the desired reliability. 5.2.4 effect of missile velocity the velocity of a missile is a measure of its kinetic increases the reliability of concrete barrier decreases. figure 8 shows that the present concrete barrier is reliable enough if the impacting velocity of the missile is less than 300 m/s. if concrete barrier is expected to receive missiles with higher velocities, its reliability should be improved by altering its design parameters, as discussed above. figure 6. effect of compressive strength of concrete on the reliability of buried concrete target figure 7. effect of concrete thickness on the reliability of buried concrete target 70 the journal of engineering research vol. 8 no. 1 (2011) 61-73 estimation of underground concrete barriers against the normal impact of a missile has been presented. the presented procedure was applied to a numerical example to estimate its reliability. some parametric studies were also conducted to illustrate how to achieve the desired reliability by altering the design values of the parameters. for this purpose, a reliability index equals to 3.2 ( t = 3.2 ) was selected and analysis was done to obtain the design value of the chosen parameter to achieve this reliability. the variation of ( t )2 was studied for different nominal values of the selected parameter. the value of the parameter for which ( t )2 is close to zero, yields the value of the parameter which makes the barrier as reliable as desired. to demonstrate the relative influence of various random variables (design parameters) on barrier reliability a sensitivity diagram was also shown. acknowledgment the work presented in this paper was supported by the deanship of scientific research, research centre, college of engineering, king saud university, riyadh, saudi arabia. references choudhury, m.a., siddiqui, n.a. and abbas, h., 2002, "reliability analysis of a buried concrete soil targets," int. j. of impact engineering, vol. 12, pp. 427-444. forrestal, m. j., altman, b. s., cargile, j. d. and hanchak, s. j., 1994, "an empirical equation for penetration depth of ogive-nose projectiles into concrete target." int. j. of impact engineering, vol. 15(4), pp. 395-405. forrestal, m. j., frew, d. j., hanchak, s. j. and bar, n. s., 1996, "penetration of grout and concrete target with ogive-nose steel projectiles," int. j. of impact engineering, vol.18(5), pp. 465-476. forrestal, m. j., frew, d. j., hickerson, j. p. and rohwer, t. a., 2003, "penetration of concrete targets with deceleration-time measurements." int. j. of impact engineering, vol. 28, pp. 479497. frew, d. j., hanchak, s. j., green, m. l. and forrestal, m. j., 1998, "penetration of concrete target with ogive-nose steel rods," int. j. of impact engineering, vol. 21(6), pp. 489-497. frew, d. j., forrestal, m. j. and cargile, j. d., 2006, "the effect of concrete target diameter on projectile deceleration and penetration depth," int. j. of impact engineering, vol. 32, pp.15841594. khan, u., siddiqui, n.a., umar, a. and abbas, h., 2003, "local damage in concrete targets under impact load," defence science journal, drdo, india, vol. 53(1), pp. 67-73. khan, r.a., siddiqui, n. a., naqvi, s.q.a. and ahmad, s., 2006, "reliability analysis of tlp figure 8. effect of missile velocity on the reliability of buried concrete target 71 the journal of engineering research vol. 8 no. 1 (2011) 61-73 tethers under impulsive loading." reliability engineering and system safety, vol. 91(1), pp. 73-83. melchers, r.e., 1999, "structural reliability analysis and prediction," john wiley and sons, second edition, 1999, usa. nowak, a.s. and collins, k.r., 2000, "reliability of structures," mc graw hill, singapore. siddiqui, n.a., choudhury, m.a. and abbas, h., 2002, "reliability analysis of projectile penetration into geological targets," reliability engineering and system safety, vol. 78, pp. 1319. siddiqui, n.a. and abbas, h., 2002, "mechanics of missile penetration into geo-materials," structural engineering and mechanics, an international journal, vol. 13(2), pp. 639-652. siddiqui, n.a., 2003, "reliability analysis of concrete barriers under normal missile impact," proceedings of national conference on advances in civil engineering: perspectives of developing countries (acedec 2003), hbti kanpur, india, 15-16 january, vol. 1, pp. 20-29. siddiqui, n.a., iqbal, m.a., abbas, h. and paul, d.k., 2003, "reliability analysis of nuclear containment without metallic liners against jet aircraft crash." nuclear engineering and design, vol. 224, pp. 11-21. siddiqui, n. a., kumar, s., khan, m. a. and abbas, h., 2006, "a simplified procedure to incorporate soil nonlinearity in missile penetration problems," structural engineering and mechanics, an international journal, vol. 23(3), pp. 249-262. 72 the journal of engineering research vol. 8 no. 1 (2011) 61-73 73 the journal of engineering research vol. 8 no. 1 (2011) 61-73 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 parametric investigation of organic rankine cycle evaporator for low temperature applications u. suhas* and g. veershetty department of mechanical engineering, national institute of technology karnataka, mangalore, india. abstract: the present work deals with the development of thermodynamic model of low temperature basic organic rankine cycle (orc) system and a chevron plate heat exchanger evaporator sub-model using engineering equation solver (ees). work output is evaluated using the orc thermodynamic model, while the evaporator sub-model calculates the total surface area of the heat exchanger. using these mathematical models, the effect of evaporation pressure, expander inlet temperature and pinch point temperature difference (pptd) on the network output and evaporator cost are studied. in addition to this, the effect of plate spacing and plate width of chevron plate heat exchanger on pressure drop and evaporator cost are analyzed in detail. finally, thermodynamic and geometric optimization is carried out using genetic algorithm to identify the optimum parameters at which the network output is maximized and pressure drop in the evaporator is minimized. sensitivity analysis showed that optimum evaporator pressure existed at which network output is maximum. thermodynamic optimization showed that work output was maximum (5.03 kw) at evaporator pressure of 5.77 bar. no improvement in the work output was seen with increase in pptd and expander inlet temperature. increase in plate width and plate spacing led to increase in evaporator cost and decrease in pressure drop. keywords: evaporator; genetic algorithm; plate heat exchanger; optimization; sensitivity analysis; thermodynamic model. الحرارة درجة ذات للتطبيقات العضوية رانكين دورة لمبخر البارمتري التحقيق المنخفضة سوهاسي . * فيرشيتي ج . ، ذو (orc) العضوية رانكين األساسي لدورة للنظام الحراري الديناميكي النموذج تطوير الحالي العمل يتناول: الملخص الهندسية. المعادالت محلل باستخدام شيفرون أللواح الحراري المبادل لمبخر الفرعي والنموذج المنخفضة الحرارة درجة (ees) الحراري الديناميكي النموذج باستخدام العمل ناتج بتقييم قمناorc ، للمبخر الفرعي النموذج استخدمنا بينما ودرجة التبخر ضغط تأثير بدراسة قمنا الرياضية، النماذج هذه . وباستخدام الحرارة لمبادل الكلية الحتساب المساحة ذلك، إلى باإلضافة. المبخر وتكلفة ناتج صافي على (pptd) القرص نقطة حرارة درجة وفرق الموسع مدخل حرارة قمنا وأخيًرا،. بالتفصيل المبخر وتكلفة الضغط انخفاض على شيفرون حراري مبادل لوح وعرض اللوح تباعد تأثير حللنا زيادة عندها يتم التي المثلى مالتاالمع لتحديد الجينية الخوارزمية باستخدام والهندسي الحراري الديناميكي التحسين بإجراء يكون عند وصول االمثل للمبخر الضغط أن الحساسية تحليل و قد أظهر. المبخر في الضغط انخفاض وتقليل صافي الناتج ضغط عند( واط كيلو 5..3) كان إنتاج أقصى أن الحراري الديناميكي التحسين كما أظهر. األقصى لصافي الناتج الحد مدخل حرارة ودرجة القرص نقطة حرارة درجة زيادة وفرق مع الناتج في تحسن أي يلحظ هذا ولم. بار 3.55المبخر .الضغط انخفاض في ونقص المبخر تكلفة في زيادة إلى األلواح وتباعد اللوح عرض في الزيادة كما أدت. الموسع الديناميكا نموذج الحساسية؛ تحليل التحسسينات؛ الحرارة؛ مبادل لوح الجينية؛ الخوارزمية المبخر؛: المفتاحية الكلمات .الحرارية؛ صافي الناتج * corresponding author’s e-mail: suhas.upa@gmail.com doi: 10.24200/tjer.vol16iss2pp130-141 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 nomenclature a area of the heat exchanger, m 2 b plate spacing, m bo boiling number dh hydraulic diameter, m dt_sup degree of superheat, °c dp pressure drop, bar f friction factor g mass velocity, kg/m 2 s ge1,ge2,ge3,ge4 non dimensional geometric parameter h enthalpy, j/kg hfg enthalpy of vaporization, j/kg k thermal conductivity, w/m k l length of the plate, m m1 mass flow rate of the working in kg/s m2 mass flow rate of the heat transfer fluid in kg/s n no of segments in evaporation zone nu nusselt number p pressure, bar pco corrugation pitch, mm pr prandtl number q ’’ average heat flux (w/m 2 ) q heat transfer rate, kw re reynolds number t plate thickness, m t temperature, °c v specific volume, m 3 /kg w plate width, m ̇ work output, kw x vapor quality α convective heat transfer coefficient (w/m 2 k) β chevron angle (degree) ρ density (kg/m 3 ) µ viscosity (kg/m s) ƞ efficiency ƞt isentropic efficiency of the expander 1,2,3,4 thermodynamic states in input cd condenser ev evaporator f fluid phase g vapor phase hf hot fluid p plate pp pump r refrigerant sp single phase th thermal tp two phase w water side wf working fluid orc organic rankine cycle pptd pinch point temperature difference, °c lmtd log mean temperature difference, °c 131 u. suhas and g. veershetty 1. introduction extraction of heat from low temperature sources has become inevitable along with other alternative sources of energy such as solar, wind etc. this is due to rapid urbanization and industrialization. climate change and global warming has forced the policy makers in various countries to adopt green technologies (arnaud et al. 2017). in recent years, a significant amount of research work has been done on organic rankine cycle (orc) technology. low grade heat can be effectively utilized using orc. heat is absorbed by the organic fluid with low boiling point, in the evaporator. the high-pressure vapor is then passed through the expander which produces power (usman et al. 2015). man et al. (2013) analyzed a solar flat plate collector driven regenerative orc and concluded that system performance improves at higher expander inlet temperature and lower condensation pressure. wang et al. (2010) examined the performance of a solar operated orc system by using evacuated tube and flat plate collectors as a source of heat for the orc power block. they concluded that overall efficiencies of 4.2% and 3.2% were achieved with the evacuated solar collector and flat plate collector respectively. pei et al. (2010) performed a numerical simulation of a low temperature solar thermal regenerative orc system by using small concentration ratio compound parabolic collectors were used. they found that regeneration had a positive effect on the orc efficiency. however, collector efficiency declined due to increase in the collector operating temperature. francesco et al. (2015) simulated a 6 kwe solar power plant using trnsys software by assessing the system performance. they demonstrated that the efficiency throughout the year remained almost constant despite seasonal fluctuations. some researchers have also studied orc systems using other heat sources such as waste heat (evangelos and christos, 2018; filiz et al. 2015; imran et al. 2014; jian et al. 2015; jiangfeng et al. 2013; and seyedali et al. 2017), biomass combustion (riffat 2012; roshaan et al. 2017) and geothermal (hossein et al. 2017; xinghua et al. 2017). orc performance mainly depends on heat exchangers, expander and choice of working fluid. moreover, the effect of operating conditions is significant. in the recent past, researchers have focused on these aspects. dai et al. (2009) carried out parametric optimization of orc for low grade waste heat recovery. their studies showed that, the presence of recuperator within the orc power block did not contribute to the betterment of orc efficiency, under the given waste heat conditions. onder kaska (2014) analyzed a waste heat recovery orc in steel industry. in his/her study, exergy losses in each orc component were assessed by using actual plant data. maximum rate of irreversibility was observed in the evaporator followed by expander, condenser and pump. higher evaporator pressures yielded greater thermal and exergy efficiencies. sahar safarian and fereshteh aramoun (2015) conducted a detailed energy and exergy analysis of 4 different configurations of orc; basic, regenerative, turbine bled and both regenerative and turbine bleeding. they concluded that regenerative orc with turbine bleeding outperformed the rest in terms of energy and exergy efficiency. significant work has been carried out on parametric optimization of various orc configurations and applications, using different working fluids. imran et al. (2015) developed a thermal and hydraulic design model of a plate heat exchanger, for a low temperature orc system. using the model, they optimized its geometrical parameters with genetic algorithm. evaporator cost and pressure drop were chosen as objective functions. plate length, width and spacing were chosen as decision variables. they concluded that the plate length had a significant effect on the evaporator pressure drop and cost. thermodynamic and cost assessment of plate heat exchanger for low temperature orc applications is very limited in the open literature. this work consists of three parts. in the first part of this study, a parametric investigation is carried out to study the effect of thermodynamic parameters on the cost and performance of the plate heat exchanger. the effect of geometrical parameters of the heat exchanger on cost and pressure drop are analyzed in the second part. finally, thermodynamic and geometric optimization is carried out for maximizing work output and minimizing pressure drop respectively using genetic algorithm. the main objectives of the study are: a) to develop a thermodynamic model to evaluate the network output of the system; b) to develop a plate heat exchanger evaporator sub-model to estimate the surface area of the evaporator and hence, the cost of the heat exchanger; c) to examine the effect of thermodynamic and geometrical parameters of the plate heat exchanger on its cost and work output of the orc system; d) to optimize thermodynamic parameters to maximize network output of orc system using genetic algorithm; and e) to minimize pressure drop by optimizing the geometrical parameters of the evaporator, using genetic algorithm. 2. system description in this study, a low-temperature organic rankine cycle power generation system is assessed. r245fa is used as the working fluid because of its lower operating pressure compared to other orc working fluids and is preferred for small-scale power generation. thermophysical properties and other important characteristics of r245fa are listed in table 1. the orc power block consists of a working fluid pump to pump the fluid to the desired pressure as shown in fig. 1. this pressurized fluid is then passed on to the evaporator where heat addition takes place. the 132 132 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 pressurized vapor passes through the expander, where the actual expansion of the working fluid takes place and the pressure drops. finally, the vapor condenses in the condenser to complete the cycle. 3. thermodynamic analysis the governing equations used in the model are explained in section 3.1 and 3.2. plate type heat exchangers are used as evaporators in small scale orc power plants, because they are compact and have better heat transfers coefficients when compared to shell and tube heat exchangers. the plate heat exchanger is divided into three zones i.e. preheating zone, 2 phase zone and superheating zone as shown in fig. 2. under the considered operating conditions, each section is designed separately using logarithmic mean temperature difference method. the following assumptions were made to simplify the analysis. a) the system is in a steady state. b) r245fa is in a saturated condition at the condenser outlet. c) isentropic efficiency of pump and expander are constant at 70%. d) the fouling effects of the heat exchanger are negligible. the program was written in function format using engineering equation solver (ees). the design process is iterative and the iterations were carried out until pressure drop on the cold side of the evaporator converged. the evaporator was redesigned for each iteration and its cost was evaluated based on the surface area of the heat exchanger. table 1. thermophysical properties of r245fa. sl property value unit 1 molecular mass 134.05 kg/k-mol 2 critical temperature 154.01 °c 3 critical pressure 3.64 mpa 4 boiling point 15.14 °c 5 ozone depletion potential 0 6 global warming potential 950 /100 year figure 1. schematic demonstration of a low temperature organic rankine cycle system. figure 2. temperature profile in the evaporator. 133 u. suhas and g. veershetty 3.1 thermodynamic model of orc system the main aim of the model was to estimate the surface area of the chevron plate heat exchanger for the given inputs as mentioned in table 2. sensitivity analysis was carried out using this model to analyze the effect of thermodynamic and geometric parameters on the work output and the cost of the system. heat balance equations used in the orc model are as follows: preheat zone ( ) (1) superheating zone ( ) (2) based on inlet and exit ( ) (3) pinch point is the location in a heat exchanger where the difference in temperature of hot and cold fluids is minimum. this is represented as, (4) calculation of network output pump work is given by, ̇ (5) the heat addition in the evaporator, ̇ (6) isentropic efficiency of the expander is calculated as the ratio of actual work done by the expander to that of the isentropic work done by the expander. (7) actual work done by the expander, ̇ (8) for a given isentropic efficiency of the expander, the actual work done by the expander can be calculated as, ̇ (9) network output is evaluated as, ̇ ̇ ̇ (10) thermal efficiency is evaluated as, ̇ ̇ (11) 3.2 evaporator sub-model 3.2.1 single phase the heat transfer is calculated as, (12) log mean temperature difference, ( ) (13) the overall heat transfer coefficient of single phase (imran et al. 2014), table 2. inputs to the model. sl no symbol parameter value unit 1 dh hydraulic diameter 0.0035 m 2 β chevron angle 45 degree 3 pco corrugation pitch 0.007 m 4 kp thermal conductivity of the plate 13.5 w/m-k 5 n number of segments in evaporation zone 20 6 tp thickness of the plate 0.0005 m 7 ƞp isentropic efficiency of pump 70 % 8 ƞt isentropic efficiency of expander 70 % 9 m2 mass flow rate 0.7 kg/s 134 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 (14) the convective heat transfer coefficient for r245fa in plate heat exchanger is calculated as (imran et al. 2015), ( ) ( ) (15) 3.2.2 two phases in this region, the correlations used in single phase, based on constant fluid properties cannot be used. the fluid properties tend to vary as the quality of the fluid changes. therefore, a modified lmtd method is employed. the evaporation region is discretised into ‘n’ smaller sections so that there are incremental changes in fluid properties at each section. therefore, constant fluid properties are assumed in each section. in this analysis, the value of n was restricted to 20 as the processing time of the model increased when the value of ‘n’ was increased and negligible change in the evaporator area was observed. the heat transfer rate for i th section is evaluated as (16) log mean temperature difference, ( ) (17) the two-phase overall heat transfer coefficient (imran et al. 2014), (18) the nusselt no. correlation for r245fa evaporation in plate heat exchanger (han et al. 2003) (19) where, ( ) ( ) (20) ( ) ( ) (21) the equivalent reynolds number and boiling number are given by, (22) (23) [ ( ) ] (24) where, g denotes the mass velocity and dh represents the hydraulic diameter of the flow channel. (25) where, w is the plate width and b is the plate spacing (jiangfeng et al. 2013). frictional pressure drop is calculated as (imran et al. 2014), (26) a single phase frictional pressure drop factor is given by (imran et al. 2015), (27) two phase frictional factor is expressed as (han et al. 2003), (28) ( ) ( ) (29) ( ) ( ) (30) cost of evaporator evaporator cost is linearly related to evaporator heat transfer area (imran et al. 2015) and is given by, ), (31) where, a is the heat transfer area in m 2 4. results and discussion parametric investigation is carried out to examine the effect of each thermodynamic parameter on the performance of the orc system. the effect of heat exchanger geometry on the cost and pressure drop is also analyzed in detail. in this study, the parameter to be investigated is varied whereas others are kept constant. the effect of thermodynamic parameters on the orc system performance is discussed in section 4.1. section 4.2 highlights the effect of geometrical parameters of the heat exchanger on the evaporator cost and pressure drop. 135 u. suhas and g. veershetty 4.1 effect of thermodynamic parameters on system performance 4.1.1 evaporator pressure the heat source temperature and the condensation temperature were set at 100 °c and 40 °c respectively. evaporator pinch point temperature and degree of superheat were kept constant at 5 °c. evaporator pressure was varied from 4 to 10 bar. figure 3 shows the effect of evaporator pressure on the network output and thermal efficiency when r245fa is used as the working fluid. it can be observed that the work output increases initially and then declines. hence, for a given working fluid, there exists an optimum evaporator pressure at which, work output is maximum. mass flow rate of the working fluid decreases with an increase in evaporator pressure as shown in fig. 4. the initial increase in the work output is due to the higher enthalpy difference across the expander. but, when the pressure is increased beyond 6 bar, the effect of reduced mass flow rate outweighs the increase in enthalpy difference across the expander. this results in reduction of network output at higher evaporator pressure. in case of thermal efficiency, at higher pressures, in addition to the network output being higher, the heat input to the system decreases. latent heat of vaporization decreases as the evaporation pressure increases. therefore, the thermal efficiency increases with an increase in evaporator pressure. figure 5 shows the effect of evaporator pressure on the evaporator area. the surface area decreases with an increase in evaporator pressure. evaporation temperature increases as a result of increasing evaporator pressure. as the pinch point temperature difference (pptd) remains constant, the exit temperature of the heat transfer fluid increases correspondingly. as a consequence of this, heat transfer rate decreases in the evaporator. hence, reduction in the evaporator surface area is observed with an increase in evaporator pressure. figure 6 depicts the variation of the evaporator with evaporator pressure cost. since, the evaporator area decreases with the increase in evaporator pressure, the cost also decreases. however, in practical applications, the evaporator surface area cannot be varied. therefore, this theoretical study only helps the designer to identify the operating condition of orc where the evaporator area (and hence, the cost) is minimum. figure 3. effect of evaporator pressure on network output and thermal efficiency. 4.1.2 pinch point temperature difference the heat source temperature and the condensation temperature were set at 100 °c and 40 °c respectively. evaporator pressure was fixed at 8 bar. superheat degree was maintained constant at 5 °c. the pinch point temperature difference was varied from 2 °c to 14 °c. when the pinch point temperature is increased while maintaining all other parameters constant, hot fluid exit temperature, thf, 3, increases. however, since, expander inlet and exit conditions remain constant, by energy balance, mass flow rate of the working fluid decreases as shown in fig. 8. this also explains the reduction in the network output as indicated in fig. 7. moreover, change in pptd does not have any impact on the thermal efficiency. as pptd is varied, both network output and heat added to the cycle decreases proportionately. therefore, the ratio of work output to heat added to the cycle remains unaffected. figure 4. effect of evaporator pressure on mass flow rate. figure 5. effect of evaporator pressure on evaporator area. figure 6. effect of evaporator pressure on evaporator cost. 136 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 figure. 9 shows the effect of pptd on the evaporator area. in addition to the reduction in mass flow rate, log mean temperature difference also increases. as a result of this, figure 7. effect of pinch point temperature difference on network output and thermal efficiency. figure 8. effect of pinch point temperature difference on mass flow rate. figure 9. effect of pinch point temperature difference on evaporator area. figure 10. effect of pinch point temperature difference on evaporator cost. the heat exchanger area decreases with an increase in pptd. the cost of the evaporator decreases because of the reduction in the evaporator surface area as indicated in fig. 10. 4.1.3 expander inlet temperature the heat source temperature and the condensation temperature were set at 100 °c and 40 °c respectively. the pinch point temperature difference was set at 5 °c. the evaporator pressure was kept constant at 6.96 bar (corresponding to saturation temperature of 75 °c). the degree of superheat was then increased up to 20 °c. with the increase in expander inlet temperature, the enthalpy difference across the expander increases. hence, according to energy balance, the mass flow rate of the working fluid decreases as depicted in fig. 12. figure 11 indicates the reduction in net power output and thermal efficiency. this is because, the marginal gain in enthalpy drop across the expander is lesser than the rate of decrease of mass flow rate. figure 13 shows the variation of evaporator surface area with expander inlet temperature. with an increase in expander inlet temperature, the area reduces initially, then increases again. due to the constraint of pptd, the mass flow rate decreases by energy balance. hence, there is a significant reduction in heat transfer rate in the evaporation section. this effect is predominant at lower expander inlet temperatures. as the inlet temperature is increased further, the log mean figure 11. effect of expander inlet temperature on network output and thermal efficiency. figure 12. effect of expander inlet temperature on mass flow rate. 137 u. suhas and g. veershetty temperature difference in the superheat zone also decreases. this results in increase in the superheated zone area. this phenomenon explains the increase in the evaporator area at higher expander inlet temperatures. the evaporator cost also exhibits the same trend as that of the evaporator area as shown in fig. 14. it can be seen that, for dry fluids such as r245fa, increasing expander inlet temperature by more than 5 °c, will not yield any positive results as the surface area requirement of the evaporator and cost increases. moreover, work output also reduces due to lower mass flow rates. 4.2 effect of geometrical parameters on pressure drop and evaporator cost plate spacing was kept constant at 0.003 m and the width of the plate was varied from 0.005 m to 0.009 m. hydraulic diameter was calculated based on eq. 24. figure 15 shows the effect of plate width on the cost of the evaporator and frictional pressure drop. as the width of the plate increases, the area increases slightly. this results in marginal increase in the evaporator cost. increase in plate width also causes reynolds number to increase. the frictional pressure drop is related to reynolds number as shown in eq 15. therefore, total pressure drop in the plate heat exchanger reduces. the variation of evaporator cost and frictional pressure drop with plate spacing is presented in fig. 16. in this case, the plate width was fixed at 0.5 m and plate spacing was varied from 0.002 m to 0.005 figure 13. effect of expander inlet temperature on evaporator area. figure 14. effect of expander inlet temperature on evaporator cost. m. the increase in plate spacing creates more space for indirect contact between the heat source and the working fluid. this leads to reduced heat transfer. to maintain constant heat transfer, additional plates have to be added, which leads to increase in the evaporator area, therefore the evaporator cost increases. when plate spacing is increased, the hydraulic diameter and reynolds number increases. this directly impacts the frictional pressure drop. frictional pressure drop reduces with increase in plate spacing. 4.3 genetic algorithm optimization as per the preceding analysis, parametric optimization is carried out using genetic algorithm (ga). the purpose of this study is: a) to maximize the work output by optimizing evaporator pressure, expander inlet temperature and pinch point temperature difference; and b) to minimize the pressure drop by optimizing the plate spacing and plate width in chevron plate heat exchanger. genetic algorithm begins from a population of possible solutions. these are called individuals. this prevents the convergence of the algorithm to suboptimal solutions. the algorithm moves towards the optimal one by applying the darwinian principle of ‘survival of the fittest’. fitness function is the criteria for selecting the individuals. the initial population is generated randomly. a new population is created by applying crossover and mutation operators to the selected individuals (parents). this technique differs from conventional optimization techniques as it involves search from a group of solutions. this figure 15. effect of plate width on evaporator cost and frictional pressure drop. figure 16. effect of plate spacing on evaporator cost and frictional pressure drop. 138 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 prevents convergence to suboptimal solutions. the procedure is explained in the form a flowchart as shown in fig. 17. 4.3.1 maximize network output the objective function for the basic orc cycle is given by, maximize ( ̇ subjected to constraints, 4 bar ≤ pev ≤ 10 bar ; 5 ≤ superheat ≤ 20 ; 5 ≤ pptdev ≤ 20 ; thf,1 = 100 °c & tcd = 40 °c input data for parametric optimization is presented in table 3 and table 4. maximum evaporator pressure was fixed at 10 bar as this is the limiting pressure for scroll expanders. heat source temperature and condensation temperature were kept constant at 100 °c and 40 °c respectively. network output was chosen as the objective function. fig. 3 showed that there exists an optimum evaporator pressure for which the work output is maximum. the optimization results are tabulated in table 5 and table 6. maximum power output of 5.03 kwe is attained at evaporator pressure of 5.77 bar. the increase in degree of superheat and pinch point temperature difference affects the network output of the system negatively, as seen from the results. figure 17. simplified flowchart of genetic algorithm. table 3. data of thermodynamic optimization for maxi mizeing work output. working fluid r245fa population size 64 stop generation 128 mutation probability 0.18 range of evaporator pressure (bar) 4-10 range of degree of superheat (ºc) 5-20 range of pinch point temperature difference (˚c) 5-20 heat source temperature (˚c) 100 condensation temperature (˚c) 40 table 4. data of geometric optimization for minimizing pressure drop. working fluid r245fa population size 64 stop generation 128 mutation probability 0.175 degree of superheat (˚c) 5 evaporator pressure (bar) 5.77 pinch point temperature difference 5.0 range for plate width (m) 0.1-0.5 range for plate spacing (m) 0.002 to 0.007 table 5. thermodynamic optimization results for maximum network output. parameter value unit p 5.77 bar dt_sup 5 ˚c pptd 5.0 ˚c wnet 5.03 kwe table 6. geometric optimization results for minimum pressure drop. parameter value unit plate spacing 0.007 m plate width 0.5 m pressure drop 0.03076 kpa 139 u. suhas and g. veershetty 4.3.2 minimize pressure drop the objective function is given by, minimize (dp), subjected to constraints, 0.1 ≤ w ≤ 0.5; 0.002 ≤ b ≤ 0.007 pev = 5.77 bar; dt_sup = 5 °c; pptdev = 5 °c (optimized performance parameters for maximum network output) in case of geometric optimization, only plate width and plate spacing were considered. the evaporator pressure was set at 5.77 bar (optimized condition), degree of superheat and pinch point temperature difference was fixed at 5 °c. the optimum value of plate width and plate spacing was evaluated so that pressure drop was minimum. the optimization results showed that plate width and plate spacing were at their upper limit set before the start of the optimization process. however, the increase in plate spacing and plate width results in increase in the evaporator area and the cost as indicated in table 7. therefore, a trade-off has to be made by the designer between cost and pressure drop requirements. 5. conclusion this study dealt with the effects of thermodynamic and geometric parameters on the performance and cost of the orc system. parametric optimization was carried out using genetic algorithm. r245fa was used as the working fluid. the main conclusions are as follows: 1. evaporator area decreases with increase in evaporator pressure resulting in the reduction of evaporator cost. optimum evaporator pressure was found to be at 5.77 bar which resulted in maximum network output of 5.03 kw. 2. network output decreased with increase in expander inlet temperature. higher inlet temperature ie. degree of superheat greater than 5°c led to increase in evaporator cost. table 7. variation of heat exchanger area with plate spacing and plate width. parameter value unit plate spacing 0.002 0.007 m plate width 0.1 0.5 m evaporator area 8.32 10 m 2 3. increase in pinch point temperature difference resulted in decrease of network output and evaporator cost (because of smaller surface area). 4. increase in plate width and plate spacing led to increase in evaporator cost and reduction in pressure drop. 5. a tradeoff has to be made between pressure drop and heat exchanger cost depending on the application. conflict of interest the authors declare no conflicts of interest. funding national institute of technology, karnataka surathkal, india. acknowledgment this project needed intensive literature review and latest research data which was provided by the national institute of technology, karnataka surathkal, india. references arnaud l, nicolas t, philippe h, rémi r, and stéphane c (2017), organic rankine cycle design and performance comparison based on experimental database, applied energy, 204: 1172-1187. dai y, jiangfeng w, lin g (2009), parametric optimization and comparative study of organic rankine cycle (orc) for low grade waste heat recovery, energy conversion and management, 50: 576-582. evangelos b, christos t (2018), investigation of a hybrid orc driven by waste heat and solar energy, energy conversion and management,156: 427-439 filiz t.o. n, rıdvan s, atakan t (2015), thermodynamic analysis of an organic rankine cycle (orc) based on industrial data, applied thermal engineering, 91: 43-52. francesco c, massimo d d’accadia, maria v, marco s (2015), design and simulation of a prototype of a small-scale solar chp system based on evacuated flat-plate solar collectors and organic rankine cycle, energy conversion and management, 90: 347–363. han dong-hyouck, lee kyu-jung, kim yoon-ho, (2003), experiments on the characteristics of evaporation of r410a in brazed plate heat exchangers with different geometric configurations, applied thermal engineering, 23: 1209-1225 140 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 130-141 hossein n, arash n, farshad j.f. (2017), conventional and advanced exergy analyses of a geothermal driven dual fluid organic rankine cycle (orc), applied thermal engineering, 122: 59-70. imran m, byung s.p. hyouck j.k. dong h.l, muhammad u (2014), thermo-economic optimization of regenerative organic rankine cycle for waste heat recovery applications, energy conversion and management, 87: 107– 118. imran m, muhammad u, byung-sik park, hyouck-ju kim, dong-hyun l (2015), multi-objective optimization of evaporator of organic rankine cycle (orc) for low temperature geothermal heat source, applied thermal engineering, 80: 1-9. jian s, yin s, chun-wei gu (2015), thermodynamic analysis and performance optimization of an organic rankine cycle (orc) waste heat recovery system for marine diesel engines, energy, 82: 976-985. jiangfeng w, zhequan y, man w, shaolin m, yiping d (2013), thermodynamic analysis and optimization of an (organic rankine cycle) orc using low grade heat source, energy, 49: 356-365. man w, jiangfeng w, yuzhu z, pan z, yiping d (2013), thermodynamic analysis and optimization of a solar-driven regenerative organic rankine cycle (orc) based on flat-plate solar collectors, applied thermal engineering, 50: 816-825. onder k (2014), energy and exergy analysis of an organic rankine for power generation from waste heat recovery in steel industry, energy conversion and management, 77: 108-117. pei g, li j, ji j (2010), analysis of low temperature solar thermal electric generation using regenerative organic rankine cycle, applied thermal engineering, 30: 998-1004. riffat (2012), experimental investigation of a biomass-fired orc-based micro-chp for domestic applications, fuel, 96: 374-382. roshaan m, faraz a, man-hoe kim (2017), thermodynamic analysis of organic rankine cycle used for flue gases from biogas combustion, energy conversion and management, 153: 627640. sahar s, fereshteh a (2015), energy and exergy assessments of modified organicrankine cycles (orcs), energy reports, 1: 1-7. seyedali s, saeed j, krishna k (2017), exergy-based optimization of an organic rankine cycle (orc) for waste heat recovery from an internal combustion engine (ice), applied thermal engineering, 126: 447-457. usman m, muhammad i, dong h.l, byung s.p (2015), design and experimental investigation of a 1 kw organic rankine cycle system using r245fa as working fluid for low-grade waste heat recovery from steam, energy conversion and management,103:1089-1100. wang x.d, zhao l, wang j.l. zhang w.z, zhao x.z, wu w (2010), performance evaluation of a low-temperature solar rankine cycle system utilizing r245fa, solar energy, 84: 353-364. xinghua l, yufeng z, jiang s (2017), system performance optimization of orc-based geo-plant with r245fa under different geothermal water inlet temperatures, geothermics, 66: 134-142. yiping d, jiangfeng w, lin g (2009), parametric optimization and comparative study of organic rankine cycle (orc) for low grade waste heat recovery. energy conversion and management, 50: 576-582. 141 mech080707.qxd the journal of engineering research vol. 6, no. 2 (2009) 20-32 1. introduction cylinder made of monolithic material such as metal and concrete (eg. asphalt concrete), is a common component employed in numerous applications such as pressure vessels (eg. hydraulic cylinders, gun barrels, pipes, boilers and fuel tanks), accumulator shells, emergency breathing cylinders, cylinders for aerospace industries, nuclear reactors, military applications and civil structures etc. (arya, bhatnagar, 1976; bhatnagar, et al. 1980; becht, et al. 2000; gupta, et al. 2001; perry and aboudi, 2003; buttlar, et al. 2004 and you and buttlar, 2005). in some of these applications such as pressure vessel for industrial gases or a media transportation of high-pressurized fluids and piping of nuclear reactors, the cylinder has to operate under severe mechanical and thermal loads, causing significant creep hence reduced service life (gupta and pathak, 2001; tachibana and iyoku, 2004 and hagihara and miyazaki, 2008). as an example, in the high temperature engineering test reactor, the temperature reaches of the order of 900 0c (tachibana and iyoku, 2004). the piping of reactor cooling system is subjected to high _______________________________________________ *corresponding author’s e-mail: guptavk_70@yahoo.co.in temperature and pressure and may be damaged due to high heat generated from the reactor core (hagihara and miyazaki, 2008). creep analysis of thick-walled cylinder made of isotropic monolithic material and subjected to internal pressure has been presented by (weir, 1957; king and mackie 1967 and pai 1967) has solved the problem for orthotropic cylinders. in all these analyses it was assumed that the strains are infinitesimal and the deformation is referred with respect to original dimensions of the cylinder. rimrott, 1959, used generally accepted assumptions of constant density, zero axial strain and distortion energy theory to derive equations for creep rate, creep strains and creep stresses in a, closed end, thick-walled hollow cylinder subjected to internal pressure. bhatnagar and gupta, 1969 obtained the solution for thick walled cylinder made of an orthotropic material and subjected to internal pressure. in recent years, the problem of creep in composite cylinders made of functionally graded materials (fgms) operating at high pressure and temperature has attracted the interest of many researchers. fukui and yamanaka, 1992 investigated the effect of gradation of components on strength and deformation of thick-walled functionally graded (fg) tubes subjected to internal pressure under effect of material parameters on steady state creep in a thick composite cylinder subjected to internal pressure tejeet singha and v.k.gupta*b adepartment of mechanical engineering, s. b.s. college of engineering and technology, ferozepur, 152001, india *b department of mechanical engineering, university college of engineering, punjabi university, patiala-147002, india received 25 april 2008; accepted 26 october 2008 abstract: the steady state creep in alsicp composite cylinder subjected to internal pressure was investigated. the creep behavior of the material were described by threshold stress based creep law by assuming a stress exponent of 5. the effect of size and content of the reinforcement (sicp), and operating temperature on the stresses and strain rates in the composite cylinder were investigated. the stresses in the cylinder did not have significant variation with varying size and content of the reinforcement, and operating temperature. however, the tangential as well as radial strain rates in the cylinder could be reduced to a significant extent by decreasing size of sicp, increasing the content of sicp and decreasing operating temperature. keywords: cylinder, creep, composite, internal pressure, modeling »∏ngo §¨° d á°vô©e ᵫª°s áñcôe áfgƒ£°se’ º¶àæÿg ∞mõdg ≈∏y ioéÿg πegƒy òkcéj *géàhƒl .∑ h èæ°s äéé«j áá°°uuóóÿÿgg¢scg ¢vgîaeéh ∞mõdg ¿ƒféb ≈∏y á«æñÿg oé¡l’g iƒb ¢sé°scg ≈∏y é¡ø°uh ” ~b ioéÿg ∞mr ∑ƒ∏°s ¿eg .»∏ngo §¨° d á°vô©e áñcôe áfgƒ£°sc’ º¶àæÿg ∞mõdg á°sgqo ” : á°sgq~dg äqé°tcg ~bh .áñcôÿg áfgƒ£°s’éh ádé£à°s’g h oé¡l’g ä’~©e ≈∏y 𫨰ûàdg iqgôm álqoh ájƒ≤àdg iƒàfih ºé◊g πeéy òkcéj á°sgqo ” ~bh gòg .á°ùªn éjhé°ùe oé¡le’g π≤j áfgƒ£°seód ájô£≤dg h á«°séªÿg oé¡lc’g ä’~©e ¿cg ~lh ~≤a ∂dp ™eh .𫨰ûàdg iqgôm álqoh ájƒ≤àdg iƒàfih ägò¨j ™e rõ©e πµ°ûh ~«≤àj ’ áfgƒ£°s’éh ägoé¡le’g ¿g ¤g .𫨰ûàdg iqgôm álqo π«∏≤j h iƒàùg ioéjr h ºé◊g π«∏≤j ~æy ®ƒë∏e ~m ¤g áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.álò‰ ,≈∏ngo §¨°v ,öcôe ,∞mr ,áfgƒ£°sg : 21 the journal of engineering research vol. 6, no. 2 (2009) 20-32 plain strain conditions. fukui et al. 1993 extended the work to investigate the effect of graded components on residual stresses in a thick walled fg tube under uniform thermal loading. chen et al. 2007 studied the creep behavior of a thick walled fgm cylinder subjected to internal and external pressures. the asymptotic solutions were obtained on the basis of taylor expansion series and compared with the results obtained by finite element analysis. you et al. 2007 investigated steady state creep in a thick walled fgm cylinder subjected to internal pressure. the effect of variation in material parameters on the stresses induced in cylinder was investigated. abrinia et al. 2008 obtained the analytical solution for radial and circumferential stresses in a thick fgm cylindrical vessel under the influence of internal pressure and temperature. the study indicated that stresses in fgm cylinder could be lowered by tailoring the material properties along radial direction of the cylinder. under severe thermo-mechanical loads cylinder made of monolithic materials may not perform well. the metal matrix composites (mmcs) such as aluminium/aluminium alloy matrix reinforced with silicon carbide offer excellent mechanical properties like high specific strength and stiffness along with high temperature stability. therefore, these are suitable for cylinder applications exposed to high pressure and high operating temperature (nieh, 1984; roy and tsai, 1988; fukui et al. 1993; salzar et al. 1996 and gupta et al. 2004. with these forethoughts, it is decided to investigate the steady state creep in a cylinder made of alsicp composite and subjected to high pressure and high temperature. a mathematical model has been developed to describe the steady creep behavior of the composite cylinder. the model developed is used to investigate the effect of material parameters viz particle size and particle content, and operating temperatures on the steady state creep response of the composite cylinder. 2. selection of creep law (1) where the symbols a', n, q, e, r and t denote respectively structure dependent parameter, true stress exponent, true activation energy, temperature-dependent young's modulus, gas constant and operating temperature. the true stress exponent n appearing in eq. (1) is usually selected as 3, 5 and 8 which correspond to three welldocumented creep cases for metals and alloys: (i) n = 3 for creep controlled by viscous glide processes of dislocation, (ii) n = 5 for creep controlled by high temperature dislocation climb (lattice diffusion), and (iii) n = 8 for lattice diffusion-controlled creep with a constant structure (tjong and ma, 2000). though, some of the research groups (mishra and pandey, 1990; pandey et al. 1992; gonzalex-doncel and sherby, 1993 and pandey et al. 1994) have used a true stress exponent of 8 to describe steady state creep in al-sicp,w (subscript 'p' for particle and 'w' for whisker) composites but a number of other research groups (park et al. 1990; mohamed et al. 1992; park and mohamed, 1995; cadek et al. 1995; yoshioka et al. 1998; li and mohamed, 1997; li and langdon, 1997 and li and langdon, 1999) have observed that a stress exponent of either ~3 or ~5, rather than 8, provides a better description of steady state creep data observed for discontinuously reinforced al-sic composites. keeping this in view, a stress exponent of 5 is used to describe steady state creep behavior of the composite cylinders in this study. the justification regarding choice of stress exponent is further elaborated in the next section. 3. estimation of creep parameters the creep law given by eq. (1) may alternatively be written as: (2) in aluminium based composites, undergoing steady state creep, the effective creep rate, eε , is related to the effective stress, eσ , through well documented threshold stress, 0σ , based cree p law given by (mishra and pandey, 1990; pandey et al. 1992; gonzalez doncel and sherby, 1993; pandey et al. 1994; park et al. 1990; mohamed et al. 1992; part and mohamed, 1995; cadek et al. 1995; yoshioka et al. 1998; li and mohamed, 1997; li and langdon , 1997; li and langdon, 1997; li and langdon, 1999; tjong and ma 2000 and ma and tjong, 2001). 0 n q' ea expe e rt σ σ ε − −⎛ ⎞ ⎛ ⎞ = ⎜ ⎟ ⎜ ⎟ ⎝ ⎠⎝ ⎠ ( )0 nme eε σ σ⎡ ⎤= −⎣ ⎦ where 11 / nq'm a exp e rt −⎛ ⎞ = ⎜ ⎟ ⎝ ⎠ and the stress exponent n is chosen as 5. the creep parameters m and so appearing in eq . (2) are dependent on the type of material and are also affected by the temperature (t) of application . in a composite , the dispersoid size (p) and the content of dispersoid (v) are the primary material variables determining these parameters . in the present study , the values of m and σ0 have been extracted from the experimental creep results reported for al -sicp composite under uniaxial loading (pandey et al. 1992). though, pandey et al. 1992 suggested a stress exponent of 8 to describe steady state creep in these composites . but due to the objections pointed by several research groups (park et al. 1990; mohamed et al. 1992; park and mohamed, 1995; cedek et al. 1995; yoshioka et al. 1998; li and mohamed , 1997; li and longdon , 1997 and li and langdon , 1999), we have used a stress exponent of 5 to describe steady state creep in composite (al-sicp) taken in the study of (pandey et al. 1992). the individual set of creep data of (pandey et al. 1992) have been plotted as 1 5/ε versus σ on linear 22 the journal of engineering research vol. 6, no. 2 (2009) 20-32 scales as shown in figs. 1(a)-(c). from the slope and intercepts of these graphs, the values of creep parameters m and σο have been obtained and are reported in table-1. this approach of determining the threshold stress σο, is known as linear extrapolation technique (lagneborg and bergman, 1976). to avoid variation due to systematic error, if any, in experimental results, the creep results from a single source have been used. the ε1/5 versus σ plots corresponding to the observed experimental data points of alsicp composites (pandey et al. 1992) for various combinations of particle size, particle content and operating temperature exhibit an excellent linearity as evident from figs.1(a)-(c). the coefficient of correlation for these plots has been reported in excess to 0.916 as given in table-1. in the light of these results, the choice of stress exponent n = 5, to describe the steady state creep behavior of al-sicp composite, is justified. the accuracy of the creep response of the composite cylinder, to be estimated in subsequent sections, will depend on the accuracy associated with prediction of creep parameters m and σο for various combinations of material parameters and operating temperature. to accomplish this task, the creep parameters given in table-1 have been substituted in the constitutive creep model, eq. (2), to estimate the strain rates corresponding to the experimental stress values reported by (pandey et al. 1992) for al-sicp composite for the various combinations of material parameters and temperature as given in table-1. the estimated strain rates have been compared with the strain rates observed experimentally by (pandey et al. 1992). figs. 2(a)-(c) show an excellent agreement between the strain rates estimated from eq. (2) and those observed experimentally, to inspire confidence in the creep parameters estimated in this study. 16 20 24 28 32 10o 10-1 10-2 ε1 /5 (s 1 /5 ) σ (mpa) (1a) 20 30 40 50 60 70 10-1 10-2 ε1 /5 (s -1 /5 ) σ (mpa) (1b) 30 35 40 45 50 55 10-1 10-2 ε1 /5 (s 1 /5 ) σσ (mpa) (1c) figure 1. variation of ε1/5 versus σ in al-sicp composite for different (a) particle sizes of sic, (b) vol% of sic and (c) temperature σ (mpa) (2a) σ (mpa) (2b) 23 the journal of engineering research vol. 6, no. 2 (2009) 20-32 4. mathematical formulation consider a long, closed end, thick-walled hollow cylinder made of al-sicp composite having inner and outer radii of a and b respectively and subjected to internal pressure p. the coordinates axes r, θ and z are taken respectively along radial, tangential and axial directions of the cylinder. the present analysis is based on the following assumptions: i. material of the cylinder is incompressible, isotropic and has uniform distribution of sicp in aluminium matrix. ii. pressure is applied gradually and held constant during the loading history. iii. stresses at any point in the cylinder remain constant with time ie. steady state condition of stress is assumed. iv. elastic deformations are small and are neglected as compared to creep deformations. (3) (4) (5) considering the equilibrium of forces on an element of the cylinder in the radial direction, we may write, (gupta and pathak, 2001) (6) where σθ is the tangential stress. since material of the cylinder is assumed to be incompressible, therefore, (7) the generalized constitutive equations for creep in an isotropic composite (gupta et al. 2005), when reference frame is along the principal directions of r, θ and z are given by, (8) (9) (10) where σr, σθ, σz are respectively the radial, tangential and axial stresses. following von-mises yield criterion (dieter, 1988), the effective stress is given by, (11) σ (mpa) (2c) figure 2. comparison of experimental and estimated strain rates in al-sicp composite for different (a) particle sizes of sic, (b) vol% of sic and (c) temperature p (µm) t ( 0c) v (vol %) m (s-1/5/mpa) óo (mpa) coefficie nt of correlati on 1.7 14.5 45.9 350 10 4.35e-03 8.72e-03 9.39e-03 19.83 16.50 16.29 0.945 0.999 0.998 1.7 350 10 20 30 4.35e-03 2.63e-03 2.27e-03 19.83 32.02 42.56 0.945 0.995 0.945 1.7 350 400 450 20 2.63e-03 4.14e-03 5.92e-03 32.02 29.79 29.18 0.995 0.974 0.916 table 1. creep parameters used for al-sicp composites in the present study the radial ( rε ) and tangenti al ( θε ) strain rates in the cylinder are respectively given by: r r d u dr ε • = ru rθ ε • = where r du u dt • = is the radial displacement rate and u is the radial displacement. r d r dr θ θ ε ε ε= − r r d r dr θ σ σ σ= − 0r zθε ε ε+ + = where zε is the strain rate in the axial ( z) direction. 2 2 e r r z e [ ]θ ε ε σ σ σ σ = − − 2 2 e z r e [ ]θ θ ε ε σ σ σ σ = − − 2 2 e z z r e [ ]θ ε ε σ σ σ σ = − − 1 2 2 2 1 2 2 /[( ) ( ) ( ) ]e z z r rσ σ σ σ σ σ σθ θ= − + − + − in a cylinder made of incompressible material with closed end, the plane strain condition exist i.e. the axial strain rate )( zε is zero (popov 2001). therefore, eqs. (3), (4) and (7) on simplifying yields, 24 the journal of engineering research vol. 6, no. 2 (2009) 20-32 (12) where c is the constant of integration. substituting eq. (12) into eqs. (3) and (4), we get (13) (14) (15) using eq. (15) into eq. (11), one gets (16) substituting eqs. (13) and (15) into eq. (8) we get, (17) (18) where, and, on substituting eq. (18) into equilibrium eq. (6) and integrating, we get, (19) where c1 is another constant of integration. the boundary conditions for a cylinder subjected to only internal pressure are given as, (i) at r = a, σr = p (negative sign implies compressive stress) (20) (ii) at r = b, σr = 0 (21) applying boundary conditions stated above, in eq. (19), we get, the values c1 and i1, thus obtained, are substituted in eq. (19) to get the radial stress, σr, (22) where, using eq. (22) in eq. (18), the tangential stress, σθ, is obtained, (23) substituting eqs. (22) and (23) in eq. (15), we get the axial stress, σz, (24) based on the analysis presented, a computer program has been developed to calculate the steady state creep response of the composite cylinder for various combinations of size and content of the reinforcement (sicp), and operating temperature. for the purpose of numerical computation, the inner and outer radii of the cylinder are taken 25.4 mm and 50.8 mm respectively, and the internal pressure is assumed to be 85.25 mpa. the dimensions of cylinder and the operating pressure chosen in this study are similar to those used in earlier work (johnson et al. 1961) on thick walled aluminum alloy (rr59) cylinder. the radial, tangential and axial stresses at different radial locations of the cylinder are calculated respectively from eqs. (22), (23) and (24). the distributions of radial and tangential strain rates are computed from eqs. (8) and (9) respectively. the creep parameters m and σo required during the computation process are taken from table-1 for the desired combination of particle size, particle content and operating temperature. 5. results and discussions before discussing the results obtained in the present study, it is necessary to check the accuracy of analysis carried out and the computer program developed. to accomplish this task, following present analysis, the tangential, radial and axial stresses have been computed for a copper cylinder, for which the results are available in the literature (johnson et al. 1961). the dimensions of the cylinr c u r • = 2r c r ε = − 2 c r θε = under the assumption of plane strain condition ie. 0=zε , the eq. (10) becomes, 2 r z σ σθσ + = ( )3 2e r σ σ σθ= − 4 23 ce r re σ σ σθ ε ⎛ ⎞ ⎜ ⎟− = ⎜ ⎟ ⎝ ⎠ putting eε and eσ respectively from eqs. (2) and (16) in the above equation and simplifying, one gets, 1 22 i ir / nr σ σθ − = + 1 14 2 1 3 n / ncni m + ⎛ ⎞⎡ ⎤ ⎜ ⎟= ⋅⎢ ⎥ ⎜ ⎟⎣ ⎦ ⎝ ⎠ 2 2 3 i oσ= 1 2 122 in i ln r cr / nr σ = − ⋅ + + 1 1 222 ni c i ln b / nb = − 2 2 1 2 2 p i ln( a / b ) i / n / nn ( a b ) +⎡ ⎤⎣ ⎦= − −− 2 2 2 / n / nx b r i ln( r / b )rσ − −⎡ ⎤= − +⎢ ⎥⎣ ⎦ 2 2 2 p i ln( a / b ) x / n / n( a b ) +⎡ ⎤⎣ ⎦= − −− 2 2 21 12 r/ n / nx r b i ln n b σθ ⎡ ⎤⎛ ⎞ ⎛ ⎞− −= − ⋅ + + ⋅ +⎢ ⎥⎜ ⎟ ⎜ ⎟ ⎝ ⎠ ⎝ ⎠⎣ ⎦ 1 12 2 2 2 2 r/ n / n / nx b r r i . lnz n b σ ⎡ ⎤ ⎡ ⎤− − −= − + ⋅ + +⎢ ⎥ ⎢ ⎥ ⎣ ⎦ ⎣ ⎦ 25 the journal of engineering research vol. 6, no. 2 (2009) 20-32 der, operating pressure and temperature, and the values of creep parameters used for the purpose of validation are summarized in table-2. to estimate the parameters m and σo for copper cylinder, firstly the values of σe have been calculated at inner and outer radii of cylinder by substituting in eq. (11) the values of σe , σr, σθ and σz at these locations as reported in the study of (johnson et al. 1961). the values of stresses σr, σθ and σz and the tangential strain rates (εθ) reported in reference (johnson et al. (1961) at inner and outer radii are substituted in eqn. (9) to estimate the effective strain rates (εe) at the corresponding radial locations. the effective stresses and effective strain rates estimated at the inner (εe = 189.83 mpa and εe = 2.168 x 10-8 s-1 ) and the outer (εe = 116 mpa and εe = 1.128 x 10-9 s-1 ) radii of the copper cylinder are substituted in creep law, eq. (2), to estimate the creep parameters m and σo for copper, table-2. the creep parameters have been used in the developed software to compute the distribution of tangential strain rate in the copper cylinder. the tangential strain rates thus obtained have been compared with those reported by (johnson et al. 1961). a nice agreement is observed in fig. 3, therefore verifying the accuracy of analysis presented and software developed in the current study. 5.1 effect of particle size and particle content figs. 4(a)-(d) show the variation of creep stresses in composite cylinder for varying sicp size from 1.7 µm to 45.9 µm. the trend of stresses obtained is similar to those reported by (bhatnagar and arya 1974). the radial stress, fig. 4(a), remains compressive over the entire cylinder radius with maximum at the inner radius and zero at the outer radius, due to the imposed boundary conditions given in eqs. (20) and (21). the tangential stress shown in, fig. 4(b), remains always tensile and increases on moving from the inner to the outer radius of cylinder. unlike radial and tangential stresses, the axial stress, figs. 4(c), decreases from maximum compressive, observed at the inner radius, to reach a maximum tensile value at the outer radius of the cylinder. it is clearly evident from figs. 4(a)-(c) that the variation in particle size does not have a sizable effect on the stresses. the maximum variation observed in radial and tangential stresses are less than 1% and for axial stress it is around 5% at a radius of 30.5 mm, in cylinders having relatively coarser sicp of size 14.5 and 45.9 µm. when compared to cylinder having finer sicp of 1.7 µm size. the variation of effective stress,σe, given by eq. (11), for varying particle size of sic is shown in fig. 4(d). the effective stress shown does not change on varying the particle size near the inner radius; however, towards the outer radius it decreases cylinder material : copper cylinder dimensions: a = 25.4 mm, b =50.8 mm. internal pressure = 23.25 mpa, external pressure = 0 operating temperature = 250 0c creep parame ters estimated: m = 3.271×10 -4 s-1/5/mpa, óo = 11.32 mpa table 2. summary of data used for validation figure 3. comparison of tangential strain rates in copper cylinder estimated from current analysis and reported by johnson et al. r (mm) (4a) r (mm) (4b) 26 the journal of engineering research vol. 6, no. 2 (2009) 20-32 marginally with increasing particle size. the maximum decrease of about 1% is observed at a radius 40.6 mm in cylinders containing relatively coarser sicp (14.5µm and 45.9 µm) when compared to cylinder having finer sicp of 1.7µm size. the strain rates given by eqs. (8) and (9) are related to the effective strain rate, εe, which ultimately depends upon (σe σo), as evident from creep law given by eq. (2). therefore, to investigate the effect of particle (sicp) size on the creep rates, the distribution of (σe σo), is plotted in fig. 5. though, the effective stress varies a little with particle size but the threshold, σo, decreases from 19.83 mpa to 16.29 mpa with the increase in particle size from 1.7µm to 45.9 µm, table 1. under external load, as the deformation of material reaches to threshold value, the interfacial debonding occurs between particles (sicp) and the matrix (al) (chen et al. 2008). due to which the stress triaxiality in the matrix near the reinforcement will decrease and the debonded particles are no longer be able to transfer the stress. the damage dissipation caused by interfacial debonding increases with increase in particle (sicp) size (chen et al. 2007), thereby decreasing the threshold stress, table 1. as a result of this, the stress difference (σe σo) exhibits sizable variation throughout the cylinder with decreasing sicp size from 45.9 µm to 1.7µm. the decrease observed towards the inner radius are relatively more than those observed near the outer radius. as a consequence, the effective strain rate (εe ) also decreases significantly with decreasing sicp size as shown in fig. 6(a). the decrease observed is about two orders of magnitude on decreasing sicp size from 45.9µm to 1.7 µm. the decrease in effective strain rate may be attributed to decrease in creep parameter m and increase in threshold stress σo with decreasing size of sicp reinforcement (table-1) as evident from creep law given in eq. (2). the radial (compressive) and tangential (tensile) strain rates, fig. 6(b), decrease on moving from the inner to the outer radius of the cylinder. throughout the cylinder, the radial and tangential strain rates remain equal in magnitude but have opposite nature due to the condition of incompressibility (eq. 7) and of plane strain condition (εz = 0). the effect of particle size on both the strain rates is similar to those observed for effective strain rate in fig. 6(a). the tangential and radial creep rates in the cylinder may be reduced to a significant extent by reinforcing finer sicp particles. the smaller size particles will be larger in number for the same volume fraction of reinforcement, thereby, leading to: (i) more transfer of load to the reinforcement with a corresponding reduction in the level of effective stress acting on the matrix, and (ii) enhancement of substructure strength, which help in restraining the r (mm) (4c) r (mm) (4d) figure 4. variation of creep stresses in composite cylinder for varying particle size of sic (v = 10 vol%, t = 350oc) r (mm) figure 5. variation of stress difference in composite cylinder for varying particle size of sic (v = 10 vol%, t = 350oc) 27 the journal of engineering research vol. 6, no. 2 (2009) 20-32 creep flow (li and langdon, 1998; peng et al. 1998; peng et al. 1999; han and langdon, 2002). further, the threshold stress observed for finer particles is higher than those observed for coarser one (table 1) due to lesser degree of debonding between particles and matrix (chen et al. 2007). as a result, the load is effectively transferred from the matrix to reinforcement hence reducing the creep rates in cylinder having finer sized sicp. figures 7(a) to 7(c) show respectively the variation of radial, tangential and axial stresses in composite cylinder containing different sicp content ie. 10%, 20% and 30% by volume. the radial stress as shown in fig. 7(a) does not vary on changing the content of sicp except for a small variation noticed in the middle of cylinder. unlike particle size, the increase in particle content leads to some sizable r (mm) (6a) r (mm) (6b) figure 6. variation of strain rates in composite cylinder for varying particle size of sic (v = 10 vol%, t = 350oc). r (mm) (7a) r (mm) (7b) r (mm) (7c) 28 the journal of engineering research vol. 6, no. 2 (2009) 20-32 variation in tangential and axial stresses as shown respectively in figs. 7(b) and 7(c). on decreasing the content of sicp from 30% to 10%, the tangential stress increases near the inner radius but decreases towards the outer radius of the cylinder. the maximum increase and decrease observed in tangential stress are about 8% and 3.5% respectively at the inner and outer radii of the cylinder. similar to tangential stress, the tensile values of axial stress as observed in fig. 7(c), increases near the inner radius but decreases towards the outer radius of the cylinder on decreasing the content of sicp from 30% to 10%. however, at the inner radius, the compressive axial stress decreases by about 10% on decreasing the particle content from 30% to 10%. therefore, by incorporating more amounts of reinforcement (sicp), the tangential and axial (only tensile values) stresses increases near the inner radius but decreases near the outer radius. the effect of particle content on effective stress, fig. 7(d), is similar to those observed for tangential stress in fig. 7(b). the stress difference, (σe σo), shown in fig. (8), decreases significantly over the entire radii with increasing sicp content. the decrease observed is relatively more prominent towards the inner radius compared to those observed around the outer radius. as expected, the effective strain rate decreases significantly with increasing content of sicp, fig. 9 (a). the effective strain rate reduces by about two orders of magnitude on increasing the content of sicp from 10% to 30%. the reduction in effective strain rate, given by eq. (2), may be attributed to decrease in creep parameter m and increase in threshold stress σo with increasing content of sicp (table-1). the impact of particle content on tangential and radial creep rates, fig. 9(b), is similar to those observed for effective strain in fig. 9(a). by increasing the content of sicp in the composite cylinder, the inter-particle spacing decreases which cause the increase in threshold stress (li and langdon, 1999) but decrease in creep parameter m (table-1), both these factors contribute in reducing the strain rates significantly. mishra and pandey (1990), in their review of uniaxial creep data of nieh (1984), have also observed that the creep rate in sic (whisker) reinforced aluminum alloy (6061al) composite could be reduced significantly by increasing the content of reinforcement. similar effect of increasing sic (particle) content on strain rate has been observed by (pandey et al. 1992) for alsicp composite under uniaxial test. r (mm) (7d) figure 7. variation of creep stresses in composite cylinder for varying particle content of sic (p = 1.7 µm, t = 350o c) r (mm) figure 8. variation of stress difference in composite cylinder for varying particle content of sic (p = 1.7 µm, t = 350 oc) r (mm) (9a) 29 the journal of engineering research vol. 6, no. 2 (2009) 20-32 5.2 effect of temperature the creep in any material is significantly affected by operating temperature. therefore, this section discusses the effect of varying operating temperature on the stresses and strain rates in a thick cylinder made of aluminium matrix composite containing 20 vol% of sicp. figs. 10(a)(c) show respectively the variation of radial, tangential, axial stresses in a composite cylinder for three different operating temperatures ie. 350 0c, 400 0c and 450 0c. the effect of temperature on the stresses is not significant, except for a slight variation observed in tangential stress, fig. 10(b). on increasing the operating temperature from 350 0c to 450 0c, the tangential stress increases a little (around 1%) near the inner radius but towards the outer radius of the cylinder it does not change. the effective stress, fig. 10(d), at the inner radius does not change with varying temperature; however, it exhibits a marginal decrease at the outer radius with increase in temperature. the stress difference, (σe σo), decreases throughout the cylinder with decreasing temperature, fig. 11. the effective as well as radial and tangential strain rates observed in figs. 12(a) and 12(b) respectively increases by about two orders of magnitude on increasing the temperature from 350 0c to 450 0c. with increase in operating temperature, the threshold stress σo decreases (gonzalez and sherby, 1993; cadek et al. 1998) and the creep parameter m increases (table-1), as a result of which the strain rates in the composite cylinder increase to a significant extent. the effect of temperature on the creep rate observed in this study are similar to those reported by (pandey et al. 1992) for alsicp composites under uniaxial creep test. r (mm) (9b) figure 9. variation of strain rates in composite cylinder for varying particle content of sic (p = 1.7 µm, t = 350oc) r (mm) (10a) r (mm) (10b) r (mm) (10c) 30 the journal of engineering research vol. 6, no. 2 (2009) 20-32 6. conclusions the present study has led to the following conclusions: . the steady state radial, tangential and axial stresses increases on moving from the inner to the outer radius of a thick walled composite cylinder. the stress distributions do not vary significantly for various combinations of particle size, particle content and operating temperature, except for some sizable variation observed in tangential and axial stresses with varying content of sicp. . the tangential as well as radial strain rates in an isotropic thick-wall internally pressurized alsicp cylinder decreases on moving from the inner to the outer radius. the strain rates induced in the composite cylinder could be reduced significantly by incorporating finer size of reinforcement (sicp), increasing the content of reinforcement and decreasing the operating temperature. references abrinia, k., naee, h., sadeghi, f. and djavanroodi, f., 2008, "new analysis for the fgm thick cylinders under combined pressure and temperature loading," american journal of applied sciences, vol. 5(7), pp. 852-859. arya, v.k., bhatnagar, n.s., 1976 "creep of thick walled orthotropic cylinders subjected to combined internal and external pressures," j mech engng sci, vol. 18(1), pp. 1-5. becht iv, c. and chen, y., 2000, "span limits for elevated temperature piping," asme j pressure r (mm) (10d) figure 10. variation of creep stresses in composite cylinder for varying operating temperature (p = 1.7 µm, v = 20 vol%) r (mm) figure 11. variation of stress difference in composite cylinder for varying operating temperature (p = 1.7 µm, v = 20 vol%) r (mm) (12a) r (mm) (12b) figure 12. variation of strain rates in composite cylinder for varying operating temperature (p = 1.7 µm, v = 20 vol%) 31 the journal of engineering research vol. 6, no. 2 (2009) 20-32 vessel technol, vol. 122(2), pp. 121-124. bhatnagar, n.s. and arya, v.k., 1974, "large strain creep analysis of thick walled cylinder," int j non linear mechanics, vol. 9(2), pp. 127-140. bhatnagar, n.s. and gupta, s.k., 1969, " analysis of thick-walled orthotropic cylinder in the theory of creep," j physical soc japan vol. 27(6), pp. 16551662. bhatnagar, n.s., arya, v.k. and debnath, k.k., 1980 "creep analysis of orthotropic rotating cylinder," asme j pressure vessel technol, vol. 102, pp. 371377. buttlar, w.g, wagoner, m, you z. and brovold, s.t., 2004, "simplifying the hollow cylinders tensile test procedure through volume-based strain," j. of association of asphalt paving technologies (aapt), vol. 73, pp 367-400. cadek, j., oikawa, h. and sustek, v., 1995, "threshold creep behavior of discontinuous aluminium and aluminium alloy matrix composites: an overview," mater sci engng, vol. a190(1), pp. 9-23. cadek, j., pahutova, m. and sustek, v., 1998, "creep behavior of a 2124 al alloy reinforced by 20 vol% silicon carbide particulate," mater sci engng, vol. a246(1), pp. 252-264. chen, j.j., tu., s.t., xuan, f.z. and wang, z.d., 2007, "creep analysis for a functionally graded cylinder subjected to internal and external pressure," the journal of strain analysis for engng. design, vol. 42(2), pp. 69-77. chen, j.k, huang, z.p. and yuan, m.a, 2008, "constitutive theory of particulate-reinforced viscoelastic materials with partially debonded microvoids," computational material science, vol. 41, pp. 334-343. chen, j.k, huang, z.p and zhu, j., 2007, "size effect of particles on the damage dissipation in nanocomposites.," composite science and technology, vol. 67, pp. 2990-2996. dieter, g.e., 1988, " mechanical metallurgy," london: mcgraw-hill. fukui. y., yamanaka, n. and wakashima, k., 1993, "the stresses and strains in a thick-walled tube for functionally graded material under uniform thermal loading," jsme , vol. 36a(2), pp. 156-162. fukui, y. and yamanaka, n., 1992, "elastic analysis for thick-walled tubes of functionally graded material subjected to internal pressure," jsme int. j. series i, vol. 35(4), pp. 379-385. gonzalez-doncel, g. and sherby, o.d., 1993, "high temperature creep behavior of metal matrix aluminium-sic composites," acta metall mater, vol. 4(10), pp. 2797-2805. gupta, vk., singh, s.b., chandrawat, h.n. and ray, s., 2004, "creep behavior of a rotating functionally graded composite disc operating under thermal gradients," metall mater trans, vol. 35a(4), pp. 1381-1391. gupta, v.k., singh, s.b., chandrawat, h.n. and ray, s., 2005, "modeling of creep behavior of a rotating disc in presence of both composition and thermal gradients," asme j engng mater technol, vol. 127(1), pp. 97-105. gupta, s.k. and pathak, s., 2001, "thermo creep transition in a thick walled circular cylinder under internal pressure," indian j pure appl math, vol. 32(2), pp. 237-253. hagihara, s. and miyazaki, n., 2008, "finite element analysis for creep failure of coolant pipe in light water reactor due to local heating under severe accident condition," nuclear engng design, vol. 238(1), pp. 33-40. han, b.q. and langdon, t.g., 2002, "factors contributing to creep strengthening in discontinuouslyreinforced materials," mater sci engng, vol. a322(1), pp. 73-78. johnson, a.e., henderson, j. and khan, b., 1961, "behavior of metallic thick-walled cylindrical vessels or tubes subjected to high internal or external pressures at elevated temperatures," proc instn mech engrs, vol. 175(25), pp. 1043-1069. king, r.h and mackie w.w., 1967, "creep of thickwalled cylinders," asme j basic engng, vol. 89(4), pp. 877-884. lagneborg, r. and bergman, b., 1976, "the stress/creep behavior of precipitation-hardened alloys," metal sci vol. 10(1), pp. 20-28. li, y. and langdon, t.g., 1998, "a comparison of the creep properties of an al-6092 composite and the unreinforced matrix alloy," metall mater trans, vol. 29a(10), pp. 2523-2531. li, y. and langdon, t.g., 1999, "an examination of a substructure-invariant model for the creep of metal matrix composites," mater sci engng, vol. a265(1), pp. 276-284. li, y. and langdon, t.g., 1997, "creep behavior of an al6061 metal matrix composite reinforced with alumina particulates," acta mater, vol. 45(11), pp. 4797-4806. li, y. and langdon, t.g., 1999, "fundamental aspects of creep in metal matrix composites," metall mater trans, vol. 30a(2), pp. 315-324. li, y. and mohamed, f.a., 1997, "an investigation of creep behavior in an sic-2124 al composite," acta mater, 1997;45(11):4775-4785. ma, z.y. and tjong, s.c., 2001, "creep deformation characteristics of discontinuously reinforced aluminium-matrix composites," composites sci technol, vol. 61(5), pp. 771-786. mishra, r.s. and pandey, a.b., 1990, "some observations on the high-temperature creep behavior of 6061 alsic composites," metall trans, vol. 21a(7), pp. 2089-2090. mohamed, f.a., park, k.t. and lavernia, e.j., 1992, "creep behavior of discontinuous sic-al composites," mater sci engng , vol. a150(1), pp. 21-35. 32 the journal of engineering research vol. 6, no. 2 (2009) 20-32 nieh, t.g., 1984, "creep rupture of a silicon carbide reinforced aluminium composite," metall trans, vol. 15a(1), pp. 139-145. pai, d.h., 1967, "steady state creep analysis of thick walled orthotropic cylinders," int j mech sci, vol. 9(6), pp. 335-348. pandey, a.b., mishra, r.s. and mahajan, y.r., 1994, "high-temperature creep of al-tib2 particulate composites," mater sci engng, vol. a189(1-2), pp. 95-104. pandey, a.b., mishra, r.s. and mahajan, y.r., 1992, "steady state creep behavior of silicon carbide particulate reinforced aluminium composites," acta metall mater, vol. 40(8), pp. 2045-2052. park, k.t., lavernia, e.j. and mohamed, f.a., 1990, "high temperature creep of silicon carbide particulate reinforced aluminum," acta metall mater, vol. 38(11), pp. 2149-2159. park, k.t. and mohamed, f.a., 1995, "creep strengthening in a discontinuous sic-al composite," metall trans, vol. 26a(12), pp. 3119-3129. peng, l.m., zhu, s.j., ma, z.y., bi, j., chen, h.r. and wang, f.g., 1998, " creep behavior in an al-fe-v-si alloy and sic whisker-reinforced al-fe-v-si composite," j mater sci, vol. 33(23), pp. 5643-5652. peng, l.m., zhu, s.j., ma, z.y., bi, j., wang, f.g., chen, h.r. and northwood, d.o., 1999, "high temperature creep deformation of al18b4o33 whiskerreinforced 8009al composite," mater sci engng, vol. a265(1), pp. 63-70. perry, j. and aboudi, j., 2003, "elasto-plastic stresses in thick walled cylinders," asme j pressure vessel technol , vol. 125(3), pp. 248-252. popov, e.p., 2001, "engineering mechanics of solids," singapore: pearson education. rimrott, f.p.j., 1959, "creep of thick-walled tubes under internal pressure considering large strains, j appl mech , vol. 26, pp. 271-274. roy, a.k. and tsai, s.w., 1988, "design of thick composite cylinders," asme j pressure vessel technol, vol. 110(3), pp. 255-261. salzar, r.s., pindera, m.j. and barton, f.w., 1996, "elastoplastic analysis of layered metal matrix composite cylinders-part 1: theory.," asme j pressure vessel technol, vol. 118(1), pp. 13-20. tachibana, y. and iyoku, t., 2004, "structural design of high temperature metallic components," nuclear engng design, vol. 233(1-3), pp. 261-272. tjong, s.c. and ma, z.y., 2000, " microstructural and mechanical characteristics of in situ metal matrix composites," mater sci engng, vol. r29(3-4), pp. 49-113. weir, c.d., 1957, "the creep of thick-walled tube under internal pressure," j appl mech, vol. 24, pp. 464-466. yoshioka, h., suzumura, y., cadek, j., zhu, s.j. and milicka, k., 1998, "creep behavior of ods aluminium reinforced by silicon carbide particulates: ods al-30 sicp composite," mater sci engng , vol. a248(1), pp. 65-72. you, l.h., ou h. and zheng, z.y., 2007, "creep deformations and stresses in thick-walled cylindrical vessels of functionally graded materials subjected to internal pressure," composite structures, vol. 78, pp. 285-291. you, z. and buttlar, w.g., 2005, "application of discrete element modeling techniques to predict complex modulus of asphalt-aggregate hollow cylinders subjected to internal pressure," j. of transportation research board, vol. 1929, pp. 218-226. chen071231.qxd the journal of engineering research vol. 6, no. 2 (2009) 1-11 ___________________________________________ *corresponding author’s e-mail: abuyaghi@ju.edu.jo uptake of cd(ii) using natural zeolite: batch and continuous fixed-bed studies luna m. marashdeha and ahmad m. al-haj-ali*b a department of chemical engineering, university of jordan, amman 11942, jordan *b department of petroleum and chemical engineering, sultan qaboos university, p.o. box 33, al-khod 123, muscat, sultanate of oman received 31 december 2007; accepted 7 october 2008 abstract: uptake of cd(ii) ions by natural phillipsite tuff was investigated both in shake-flask and fixed-bed columns. equilibrium uptake, qe, was found to best fit langmuir adsorption isotherm with a maximum value of 25.78 mg/g. percent removal of cd ions was close to 100% from initial metal ion concentrations in the range 50 75 mg/l at 5.0 g zeolite/l. also, qe was found to vary exponentially with zeolite dose. break points as high as 350 minutes were obtained from bed treatment at favorable conditions of a low solution flow rate and high bed depth. in batch experiments, equilibrium ph increased to < 8.0 excluding chemical precipitation as part of the removal while in fixed-beds the final ph exceeded 9.0. it is suggested that a sieve action of zeolite porous structure plays a role as an uptake mechanism in addition to the ion exchange. keywords: metal uptake, ion exchange, cadmium, adsorption isotherms, fixed-bed column, natural zeolite, jordan iôªà°ùÿÿg áàhéãdg á≤ñ£dg h áñlƒdg »à≤jô£d á°sgqo /»©«ñ£dg âj’ƒjgõdg ᣰsgƒh ωƒ«eoéµdg ¢ué°üàeeg »∏y êé◊g ~ªmg h i~°tgôÿg éfƒd ::áá°°uuóóÿÿgg≠∏ñj ¢uéà°üeeód ≈°übc’g ~◊g ¿g ~lh ~bh .áàhék äé≤ñw h álélq ¥qgho ‘ é«ñjôœ âjé°ùñ«∏«a ´ƒf øe âj’ƒjr ᣰsgƒh ωƒ«eoéµdg äéfƒjcg ¢ué°üàeeg á°sgqo â“ øe %100 ‹gƒm ≠∏ñj ¥qgh~dg ‘ ádgreód ájƒäÿg áñ°ùædg ¿cg é° jcg ~lh ~bh .òª‚’ êpƒ‰ ™ñàj ¿gõj’g ¿g éªc âj’ƒjõdg áyôl ™e á«°scg ádg~h ¢übéæàj h ºz /º¨∏e 25^78 𫨰ûj ±hôx ~æy ∂dp h âj’ƒjõdg iq~b péøæà°seg πñb á≤«bo 350 √qg~≤e øeõd ¢ué°üeoe’g ôªà°seg ~≤a áàhéãdg á≤ñ£dg ‘ éecg .îd/ωƒ«eoéc º¨∏e 75-50 õ«cîdg ägp π«déùg 8^0 øe πbg ¤g »æ«lhq~«¡dg ¢sc’g ™øjqg ¿gõj’g á°sgq~d álélôdg ¥qgh~dg üqéœ ‘ .âj’ƒjõdg øe á≤«ªy á≤ñwh ωƒ«eoéµdg ∫ƒ∏ù ¢ øîæe ¿éjôl ∫~©e ‘ πãªàj áªfóe ¢ué°üeo’g á«dbg ¿cg ≈∏y á°sgq~dg èféàf ∫~j .9^0 ¤g »æ«lhq~«¡dg ¢sc’g ™øjqg ~≤a iôªà°ùÿg áàhéãdg á≤ñ£dg üqéœ ‘ éecg .ωƒ«eoéµ∏d »f髪«c ö«°sôj çh~m »øæj gògh .»lqéÿg í£°ùdg ≈∏y êƒj’g ∫oéñàdg ¤g áaé°veg ¿~©ÿg ägƒéa ∫ón égqé°ûàfeg ~©h ωƒ«eoéµdg äéfƒjc’ »féjõ«ødg ¢ué°üàee’g ≈∏y πª°ûj áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .¿oqc’g ,»©«ñ£dg âjójõdg ,áàhéãdg á≤ñ£dg êôh ,êgõje’g rgõàee’g äé«æëæe ,ωƒ«eoéµdg ,êƒjc’g ∫oéñàdg ,ägõ∏ødg ádgreg : notation b langmuir isotherm parameter related to adsorption energy (l/mg) co initial cd ion concentration (mg/l) ce equilibrium cd ion concentration (mg/l) d zeolite dose in equilibrium experiments (g /l solution). iupac international union of pure and applied chemistry k equilibrium constant of cd hydrolysis reaction kf freundlich isotherm parameter related to adsorption capacity (l/g) n freundlich isotherm parameter related to adsorption intensity (dimensionless) qe equilibrium uptake (mg cd/g zeolite) qm experimental parameter indicating maximum uptake (mg/g) ql langmuir isotherm parameter related to adsorption capacity (mg/g) r dimensionless separation factor to describe the type of adsorption s experimental coefficient (l/g) z symbol representing active sites on the zeolite surface. 2 the journal of engineering research vol. 6, no. 2 (2009) 1-11 1. introduction heavy metal ions are major pollutants in the wastewaters of many industrial activities. cadmium is introduced into the aquatic environment by discharge of liquid effluents with relatively low metal concentration from smelting, metal plating, cadmium-nickel batteries, phosphate fertilizer, mining, paints, metal alloys and other chemical and engineering industries (alloway, 1995; stumm and morgan, 1996). such metal ions are soluble in water and could penetrate through soil layers thus polluting soil, public water supplies and accumulating in agricultural crops. cd is the most toxic element readily taken up by plants (dabrowski et al. 2004). therefore, the removal of cd and other metal ions from waste effluents before their discharge is a health and environmental priority required by international and local standards. the harmful effects of cadmium on human health are well documented including a number of acute and chronic disorders such as renal damage, anemia and hypertension (friberg et al. 1974). the discharge limits from industrial effluents vary according to fate or application of wastewater. for example, the maximum allowable concentration of cd in the jordanian wastewater (both industrial and treated domestic sewage) standards is 0.01 mg/l for reuse in irrigation and discharge to streams and lakes; 0.02 mg/l for underground water recharge and 0.07 mg/l for discharge to the sea. drinking water guideline value for cd recommended by world health organization (who) is 0.005 mg /l (who, 2004), which is adopted by the jordanian standard for drinking water. two engineering techniques for heavy metal removal are usually applied, namely chemical precipitation and ion exchange. these methods suffer from disadvantages such as non-selectivity, high cost and waste disposal problems. natural zeolites are crystalline microporous aluminosilicates with well-defined structures that consist of a framework formed by tetrahedra of sio4 and alo4 connected through oxygen atoms. the frame-work contains channels and cavities with molecular dimensions from 0.3 to 1 nm occupied by alkaline and alkaline earth metal cations as well as water molecules. these cations possess the ability to move freely and can be exchanged by more selective cations from aqueous solutions (sand and mumpton, 1978). in several parts of the world, natural zeolites have been used as ion exchangers for the removal of ammonium ions from municipal wastewater (hedstrom, 2001); cesium and strontium ions from radioactive wastewater and heavy metals from industrial wastewater (ouki et al. 1994). fixed beds of zeolite may be used as packing materials in subsurface barriers to control ground water pollution (cincotti et al. 2006). these applications are stimulated by the non-toxic nature of zeolites as well as their availability in the many countries. application of natural zeolites for the removal of heavy metals such as cd(ii), pb(ii), ni(ii), etc. has been reported in several studies considering both in equilibrium batch experiments (berber-mendoza, et al. 2006; alvarez-ayuso, 2003; lee and moon, 2001) as well as dynamic fixed-bed column experiments (inglezakis and grigoropoulou, 2004; alhaj-ali and al-hunaidi, 2004) or both (vaca-mier, et al. 2001). results of the above studies indicated that natural zeolites possess good ion exchange properties and thus can be used to substitute or assist ion exchange systems employing commercial synthetic resins in heavy metal removal from wastewater. this recommendation is based on relatively high removal efficiency as well as low cost of the mineral. in the northeastern part of jordan, large deposits of natural zeolite produced by alteration of basaltic glass are present. currently this mineral is being mined and made commercially available for agricultural applications where it is used as a soil conditioner. environmental applications of this mineral, particularly as an adsorbent / ion exchange) has been the subject of investigation for more than 15 years. published studies on this mineral have shown that it is of a promising potential in heavy metal removal from wastewater (dwairi, 1992; al-haj-ali and el-bishtawi, 1997, rashdan, 2002). this zeolite has become a commercial commodity in the local market available at low prices. the fact that natural zeolite can be regenerated after heavy metal removal makes it even more attractive as an ion exchanger. according to rashdan (2002), reactivation of zeolite loaded with pb, cd, ni, cu and zn ions with cacl2, kcl or nano3 was successful up to 80-100% of the pre-adsorbed amounts of the indicated metals depending on the metal as well as the type and concentration of eluant solution. the main objectives of this study are to evaluate the ability of natural jordanian zeolite for cd ions removal under certain conditions both in batch and continuous fixed-bed contact modes. 2. experimental 2.1 materials the zeolite used in this work was phillipsite-rich tuff mined from the aritain mountain area in the northeastern part of jordan and supplied by the natural resources authority of jordan. the zeolite was crushed using a jaw crusher and sieved into three size cuts, namely, 355-500, 500-710 and 710-850 µm. particles were washed with tap water until the dust and fines are removed. the mineral was immersed in 25g.dm-3 solution of nacl (analytical reagent, analar co.) in distilled, deionized water at room temperature for 24 h. the solution was replaced by a fresh one and the mineral continued to be immersed for another 24 h. the sodium homo-ionic zeolite was then filtered, washed gently with distilled, deionized water and ovendried at 35 oc. the chemical composition of raw zeolite (as received) was obtained by x-ray fluorescence (xrf) analysis and is given in table 1, which also presents related physical properties of the zeolite. the langmuir and bet surface areas, total pore volume and average pore 3 the journal of engineering research vol. 6, no. 2 (2009) 1-11 diameter were determined for two zeolite sieve cuts (355500 and 500-710 µm) using a quantachrome autosorb automated gas sorption system. data are calculated and reported using autosorb 1 program for nitrogen at 77k. cadmium nitrate was used as sources for cd ions. stock solutions of 1000 mg/l were prepared by dissolving the exact weight of cd(no3)2.4h2o in distilled-deionized water and used to provide the desired cd ion concentration by proper dilution. glassware was cleaned by detergent and double rinsed thoroughly by tap water and distilled-deionized water. 2.2 equipment, procedures and sample analysis 2.2.1 batch experiments equilibrium uptake was determined by shake-flask batch experiments. a volume 150 ml of solution containing 50 or 75 mg/l was placed in a 250 ml erlenmeyer flask. the initial ph was adjusted to a specified initial value (4.0 or 5.5) while the final ph was not controlled. an accurately-weighed zeolite mass (in the range 0.12-1.2 g) was added to the solution thus making slurry concentration in the range 1.08.0 g/l. a series of such flasks was then agitated at a constant speed in a laboratory shaker at room temperature (2022°c). after 120 hours of contact to ensure that equilibrium is attained, the zeolite particles were separated by filtration. a sample of the solid-free solution was analyzed for cd remaining concentration using a using flame atomic absorption spectrophotometer (pye-unicam model). before sample analysis, the aas was calibrated using certified standard cadmium solutions. meanwhile, the final ph of the solution was measured using mettler toledo ph meter calibrated with three standard buffer solutions of ph 4.01, 7.00 and 9.21. in another group of shake-flask experiments, blank runs involving zeolite in distilled-deionized water (no cadmium present) were conducted to investigate the effect of adsorbent on solution ph. in these experiments, the same solution volume, slurry concentration and initial ph were maintained and the same procedure above was followed. a third group of shake-flask experiments were conducted to investigate the effect of ph on chemical precipitation of cadmium in the absence of zeolite adsorbent. solutions at initial concentrations of 24, 48, 70 and 91 mg cd/l were shaked in 250-ml flasks at a specified ph value ranging from 4.0 to 10.0 then samples of precipitate-free solution were analyzed for remaining cd concentration using aas technique as indicated above. 2.2.2 continuous experiments these experiments were conducted using a laboratory scale glass column, 1.5cm inside diameter and 90 cm height. the column is gently packed with different masses of zeolite which provide different bed depths. the feed solution is pumped from a glass storage tank to an overhead constant-level tank made of plexi-glass and fed to the column through a glass flow meter capable of reading in the range 10-80 cm3.min-1. flow control is achieved by valves fitted at the top and at the bottom of column. a filter is located at both ends of the column and flexible connections permit column to be charged, sampled, emptied and washed. samples at zero time (from solutions before contact with zeolite) as well samples from effluent solutions are withdrawn at predetermined times in the range of 3 to 450 min). the concentrations of cd in these samples were measured using aas as discussed above. experimental column variables and their tested ranges are given in table 2. 3. results and discussion 3.1 equilibrium uptake equilibrium uptake (qe, mg cd/g zeolite) was calculated from the following mass balance equation: qe = (co ce)/d (1) weight percent 34.20 10.57 7.02 11.35 7.28 2.00 1.65 1.23 0.33 zeolite composition * sio2 al2o3 feo cao mgo k2o na2o tio2 other * un-sieved 2.380 g/cm 3 1.035 g/cm 3 yellowish brown 124 m2/g 92 m2/g 4.02 x 10 -2 cm3/g 1.75 å 120 m2/g 90 m2/g 3.90 x 10 -2 cm3/g 1.75 å (å =0.1ì m) for un-sieved zeolite true density bulk density color for size cut 355 -500 µm langmuir surface area bet surface area total pore volume avg. pore diameter for size cut 710 -850 µm langmuir surface area bet surface are a total pore volume avg. pore diameter table 1. physical and chemical properties of zeolite tuff 4 the journal of engineering research vol. 6, no. 2 (2009) 1-11 experiments were carried out to determine the equilibrium uptake using different zeolite doses. the results obtained are plotted as percent cadmium ion removal versus d (fig. 1a) and as qe versus d (fig. 1b) for two initial cd ii ion concentrations (co= 50 and 75 mg/l). figure 1 (a) shows that almost 100% removal was possible using a minimum 5 g/l of zeolite at both initial concentrations. at lower d values, removal percent was higher for the lower co value as expected with a difference of ~ 17% at 1 g zeolite/l. as a general trend and for all cases, it is clear that as the amount of zeolite per unit volume of solution is increased, metal uptake is significantly reduced. this observation is explained by the competition between metal ions for active (ion exchange) sites on the zeolite surface. higher co resulted in higher uptake per unit mass of zeolite (qe). using non-linear regression technique, the data plotted in fig. 1 (b) are found to best fit an exponential decay function in the form: qe = qm. exp (s.d) (2) parameter qm (mg/g) is the upper limit of this decay function at minimum d indicating maximum uptake. the correlation coefficients for the two curves shown are 0.996 and 0.989 indicating good fit. parameter s indicates how sharp is the change of metal uptake, qe, with zeolite dose, d, that is, the slope of this relationship. the quantity qm increases as co increases. 3.2 equilibrium ph in fig. 2, the equilibrium solution ph is plotted as a function of zeolite dose for the two initial concentration values considered. the solution ph has increased sharply from 4.0 for zeolite-free solution to 6.48 after addition of 0.8 g zeolite/l. final ph continued to increase with increasing zeolite dose up to 7.72 at d = 8.0 g zeolite/l for co = 50 mg/l. similarly, ph increased between 6.37 and 7.34 when d increased from 0.8 to 8.0 g zeolite/l for co = 75 mg/l. at any zeolite dose, equilibrium ph is found to be slightly lower for the higher initial metal ion concentration. at d = 8.0 g zeolite/l, the final ph is 7.34 for co = 75 mg/l while it is 7.72 for 50 mg/l. similar observations have been reported by many authors investigating zeolite / metal systems. al-haj-ali and el-bishtawi (1997) investigated pb(ii) and ni(ii) sorption by a fine sieve cut from the same zeolite mineral and found that solution ph increased from an initial value of 4.0 to an equilibrium value of 5.7-7.9 for pb(ii) and to 6.9-7.3 for ni(ii) ions. moderate changes in ph were also reported by genc-fuhrman, et al. (2007) for cd(ii) adsorption on natural clinoptilolite, alumina and sand. the solution ph increased from 6.5 initially to 7.7, 7.4 and 7.2, respectively at a zeolite dose of 20 g/l at equilibrium. the results obtained in this study are also in agreement with those of by alvarez-ayuso et al. (2003) who investigated cd ions uptake by clinoptilolite natural zeolite from greece. they have found that starting with initial cd ion concentrations of 50, 60 and 75 mg/l, the equilibrum ph values were 7.1, 7.0 and 6.9, respectively. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 zeolite dose, d (g/l) 10 20 30 40 50 60 70 80 90 100 110 p e rce n t r e m o v a l co = 75 mg cd/l co = 50 mg cd/l particle size: 355-500 m initial ph = 4.0 µµ (1a) 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 zeolite dose, d (g/l) 0 5 10 15 20 25 30 u p ta k e , q e (m g c d / g ) co = 50 mg cd/l co = 75 mg cd/l particle size: 355-500 m initial ph = 4.0 µ (1b) figure 1. effect of zeolite dose on (a) percent removal (b) equilibrium uptake 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 zeolite dose, d (g/l) 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 s o lu tio n p h particle size: 355-500 m µµ uptake: zeolite in cd solution (c o=50 mg/l) at ph 4.0 uptake: zeolite in cd solution (c o=75 mg/l) at ph 4.0 blank: zeolite in water at ph 4.0 blank: zeolite in water at ph 5.5 figure 2. equilibrium ph for cadmium uptake and blank systems pe rc en t r em ov al zeolite dose, d (g/l) u pt ak e, q e (m g c d/ g) zeolite dose, d (g/l) so lu tio n ph zeolite dose, d (g/l) 5 the journal of engineering research vol. 6, no. 2 (2009) 1-11 the final solution ph is the resultant of two effects; firstly the initial ph of the cadmium nitrate solution, and secondly the addition of zeolite particles to that solution. it is known that the ph of cd(no3)2.4h2o solution (and other similar metal salts) is below 7.0, the ph of pure water, since this is a weak salt obtained from a strong acid (hno3) and a weak base [cd(oh)2]. as an example, the ph of a solution whose co = 70 mg cd /l is 5.7; the lower the cd initial concentration, the higher is the solution ph. the effect of the second parameter, that is, addition of zeolite was experimentally investigated for two different initial ph values (4.0 and 5.5) using a zeolite particle size cut of 355-500 µm in distilled-deionized water (blank runs). the results are shown also in fig. 2 along with equilibrium ph from uptake experiments. in general, the final ph of zeolite-extract is always higher than that of the water. it is seen that as the zeolite dose added to water increased from 0.8 to 8.0 g/l, the final ph significantly increased from 4.0 to 8.9 and from 5.5 to 9.6. this is due to release of alkaline species from the active sites on the zeolite surfaces in contact with water which has been demonstrated by similar studies involving natural zeolite (vaca-mier, et al. 2001, alvarez-ayuso, et al. 2003; camilo, et al. 2005; kocaoba, et al. 2007). raw natural zeolite possesses a significant content of alkaline metals (ca, mg, k and na). sodium is the major species contributing to ph rise in the case of the particular natural zeolite used in the study since it has been added to the mineral preparing it as na+ ion exchanger. calcium carbonate, present as an impurity in the zeolite tuff is expected to take part in the ph rising process. it is noteworthy that the higher the initial ph of water, the higher the final equilibrium ph of zeolite-water slurry; the difference narrows down at high d. the decrease in solution ph due to the removal of cd ions from solution by zeolite will counter the rise in ph due to zeolite dissolution thus maintaining ph level below 8.0 compared to blank experimets resulting in high ph levels (9.19.6). the fact that equilibrium ph increases during metal uptake (pb, ni, cd, cu, etc.) raises a concern about possible chemical precipitation of metal ions as hydroxide thus no. flow rate, f (ml/min) bed depth, h (cm)* initial cd(ii) conc., c o (mg/l) particle size cut, (µm) initial solution ph break time, tb (min) h/f (min /cm2) 1 22 14.4 50 355-500 4.0 95 0.65 2 15 14.4 50 355-500 4.0 215 0.96 3 10 14.4 50 355-500 4.0 350 1.44 4 10 9.6 50 355-500 4.0 190 0.96 5 10 4.8 50 355-500 4.0 75 0.48 6 15 14.4 100 355-500 5.5 105 0.96 7 15 14.4 75 355-500 5.5 130 0.96 8 15 14.4 50 355-500 5.5 230 0.96 9 15 14.4 75 355-500 4.0 130 0.96 10 15 14.4 75 710-850 4.0 110 0.96 11 15 14.4 50 500-710 4.0 130 0.96 12 15 14.4 50 500-710 5.5 140 0.96 table 2. break time in zeolite beds at different experimental conditions *corresponding to 10, 20 and 30 g of zeolite mass table 3. specification of cadmium due to aqueous dissolution and hydrolysis reactions 6 the journal of engineering research vol. 6, no. 2 (2009) 1-11 masking true uptake or ion exchange. kocaoba, et al. (2007) declared that most of the heavy metal ions tend to precipitate at ph > 6 and thus making it difficult to quantify true adsorption. cadmium may be an exception from this general statement as indicated by cadmium specification data in table 3 which is compiled by referring to different aqueous chemistry and metal removal publications. the main relevant conclusion from this table is that chemical precipitation does nor occur at ph < 8.0. other published studies give more insight about the possibility of precipitation during the course of sorption. deliyanni and matiz (2002) conducted their experiments on cd sorption by synthetic akaganeite [β-feo(oh)] up to ph 9.5 from a solution of 10 mg cd/l considering that cadmium removal at this ph is due only to sorption. they based their conclusion on the thermodynamic aqueous speciation of cd which indicates that chemical precipitation (as hydroxide) is not taking place below ph 9.5. srivastava, et al. (2004) in their study on cd adsorption on kaolinite considered cd2+ as the only species dominant up to ph 10. using solutions containing 15 mg cd/l, they excluded chemical precipitation as a cd removal mechanism in the studied range. however, rangel-mendez and streat (2002) reported that cd exists predominantly as cd(h2o)62+ ions and precipitates as cd(oh)2 just below ph 8. the results of shake-flask experiments conducted to determine the ph at which cd precipitates from solutions at 24, 48, 70 and 91 mg/l in absence of zeolite are shown in fig. 3 as percent cadmium reduction against ph. this figure indicates that precipitation of cd ions as cd(oh)2 is rather slight (less than 10%) below ph 8.6 for both initial concentration of 24 and 48 mg/l. for the two higher initial concentrations of 70 and 91 mg/l, precipitation is significant only beyond ph 8.3. these results are in agreement with the above-mentioned studies (deliyanni and matiz, 2002; srivastava, et al. 2004; berber-mendoza, et al. 2006). since the maximum ph reached in batch uptake experiments in this study is less than 8.0 (fig. 2), therefore chemical precipitation of cadmium at ph < 8.0 is ruled out and cd removal is confirmd as a totally sorption process. the uptake mechanism is discussed further in section 3.5. 3.3 modeling uptake with adsorption isotherms the adsorptive capacity of a solid is concluded from equilibrium sets of data that need to be fitted to mathematical models to facilitate design of adsorption systems and broaden their application. equilibrium uptake data were fitted to popular models of adsorption isotherms, namely; langmuir, freundlich and sips models. langmuir isotherm has the following non-linear form: qe = qlb.ce / (1 + b.ce) (3) based on parameter "b", the dimensionless separation factor, r, is defined as follows: r = 1/ (1+b.co) (4) this factor is used to decide whether the adsorption process is favorable (0 < r < 1) or not. in the cd(ii) / phillipsite tuff system investigated, "r" has the values of 0.02 and 0.03 for initial concentrations of 75 and 50 mg/l, respectively, indicates a very favorite adsorption process. freundlich isotherm has the following form: qe = kf. ce 1/n (5) the predictions of the two models are shown in fig. 4 compared to experimental data. regression analysis has shown that langmuir isotherm represents the data better than freundlich isotherm (correlation coefficients, r2, are 0.995 and 0.977, respectively. this is an indication of mono-layer coverage of metal ions binded to specific sites rather than accumulation of ions on particle surfaces. langmuir and freundlich isotherms parameters obtained in this study by fitting experimental equilibrium data to these models are given in table 4 compared to those reported in the literature for cadmium uptake by other adsorbents and ion exchangers. the langmuir parameter, ql, represents the maximum sorption capacity which is 25.78 mg cd/g for the system being studied. this is a relatively high sorption capacity compared with the performance of other sorbents from the same nature, ie. soils and minerals. 3.4 breakthrough curves the effects of solution flow-rate, bed depth, initial concentration of cd ions in the feed, zeolite particle size and initial solution ph on the sorption behavior of cd(ii) ions are presented and discussed. the breakthrough time (tb) is taken as an indicator of zeolite bed efficiency and defined as the time required to reach the breakpoint assumed to be at 5% breakthrough (that is, the time at which c/co = 0.05). the higher the break time, the higher is the metal uptake and the more efficient is the bed. the tb values obtained from various breakthrough curves describing cd(ii) uptake by zeolite bed are given in table 2 along with the respective column parameters. figure 3. percent cadmium reduction versus solution ph pe rc en t c ad m iu m r ed uc tio n solution ph 7 the journal of engineering research vol. 6, no. 2 (2009) 1-11 3.4.1. effect of solution flow rate breakthrough curves obtained using three solution flow-rates keeping other parameters constant (zeolite mass = 30 g equivalent to 14.4 cm of bed depth) are shown in fig. 5. break times of 350 and 215 min obtained at the low f values of 10 and 15 ml/min are considered high and denotes an efficient bed. however, the bed performance would be lower at higher solution flow rates. this is inferred from the significantly shorter break time at 22 ml/min. 3.4.2. effect of bed depth in fig. 6, the bed depth (h) was varied keeping all other parameters constant. as expected, the highest value of tb was obtained using the highest bed depth (14.4 cm) corresponding to the highest zeolite mass available. decreasing h while keeping f constant at 10 ml/min) resulted in shorter break time. it is concluded that at the conditions tested, a bed depth less than 9.6 cm (zeolite mass less than 20 g) will not give a satisfactory performance. the effect of bed depth, h, and solution flow rate, f, reference sorbent material and country langmuir parameters ql (mg/g) b (l/mg) freundlich parameters kf (l/g) n garcia-sanchez et al. (1999) sepiolite, spain 8.26 1.67 6.17 3.80 singh et al. (2001) phosphatic clay; florida, us 24.5 0.111 -- -- alvarez-ayuso et al. (2003) clinoptilolite, greece 4.60 0.162 -- -- berber-mendoza et al. (2006) clinoptilolite, mexico 29.11 0.022 20.57 2.22 cincotti et al. (2006) clinoptilolite; sardinia, italy 10.40 1.08 -- -- sprynskyy et al. (2006) hcl-treated clinoptilolite, ukraine 4.88 0.078 0.636 2.20 genc-fuhrman et al. (2007) clinoptilolite, uk -- --0.21 1.07 genc-fuhrman et. al. (2007) sand, denmark -- --0.42 3.57 genc-fuhrman et al. (2007) alumina, denmark -- --6.35 0.86 genc-fuhrman et al. (2007) fly ash, denmark -- --5.01 0 .64 this work phillipsite tuff, jordan 25.78 0.636 10.29 3.78 table 4. cd(ii) adsorption parameters of jordanian zeolite compared to other zeolites and mineral adsorbents 0 5 10 15 20 25 30 35 40 45 50 55 ce (mg cd/l) 0 5 10 15 20 25 30 35 q e (m g c d / g ) experimental data freundlich isotherm langmuir isotherm figure 4. modeling experimental data with a adsorption isotherms figure 5. breakthrough curves at variable solution flow rate c / c o time (min)ce (mg cd/l) q e (m g c d/ g) 8 the journal of engineering research vol. 6, no. 2 (2009) 1-11 can be combined in one parameter, the h/f ratio which has the units of min/cm2. the break time, tb is plotted versus h/f for the data given in table (2) obtained at constant co (50 mg/l), constant initial ph of solution (4.0) and using one size cut of particles (355-500 µm). the point of original (0, 0) is also included in the data. linear regression supported a straight line (fig. 7) with 0.957 correlation coefficient, 246.7 as the slope and -30.4 as the intercept. the basic idea behind the h/f ratio is the complementing roles of sorbent bed depth and flow rate of feed solution. furthermore, it is an interaction of two effects, namely, the availability of active surface sites from the sorbent side and the adequacy of residence time from the solution side. both factors are crucial for a successful sorption process which usually involves mass transfer of ions from the aqueous phase to solid surface and the subsequent binding to that surface. using high bed depth, which provides easily accessible ion exchange sites, compensates for lack of sufficient residence time of ions in contact with the bed. to simplify the idea, it can be said that for a given removal efficiency, using high bed depth with high flow rate is the same as using a low bed depth with a low flow rate. 3.4.3. effect of initial cd ion concentration figure 8 shows the effect of cd(ii) ion concentration in feed solution, co, on break time is shown in maintaining constant f = 15 ml/min and h = 14.4 cm. the slopes of breakthrough curves decreased as co decreased indicating better sorption behavior at lower co, where break time reached 230 min compared to 150 and 105 min at the higher co values of 75 and 100 mg/l, respectively. this observation is in agreement with the findings of takac and calik (1998) for ion exchange. 3.4.4. effect of zeolite particle size the effect of zeolite particle size on breakthrough curves is shown in fig. 9. the effect is not large where break time decreased only by 15% when coarse particles (710-850 µm) are used compared to the case of the finer particles (355-500 µm). a look at table (1) reveals that these two size cuts have similar physical properties in terms of specific surface area (9192 m2/g) and average pore diameter (1.75å). therefore, it is not surprising that the two zeolite fractions exhibit comparable sorption performances. figure 6. breakthrough curves at variable bed depth figure 7. effect of bed-depth to solution flow-rate ratio, h/f 0 50 100 150 200 250 300 350 400 450 time (min) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 c / c o 100 mg/l 75 mg/l 50 mg/l initial cd ion concentration fixed conditions : bed depth: 14.4 cm flow rate: 15 ml/min particle size: 355500 m initial ph: 5.5 µµ figure 8. breakthrough curves at variable initial cd ion concentration 0 50 100 150 200 250 300 350 400 450 500 time (min) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 c / c o 710-850 m 355-500 m zeolite particle size µµ µµ fixe d conditions : bed de pth: 14.4 cm flow rate : 15 ml/m in co = 75 mg/l initial ph: 4.0 figure 9. breakthrough curves at different zeolite particle sizes c / c o time (min) c / c o time (min) bed depth / solution flow rate, h/f (min/cm2) b re ak t im e, t b (m in ) c / c o time (min) 9 the journal of engineering research vol. 6, no. 2 (2009) 1-11 3.4.5. effect of initial solution ph and change in ph figure 10 involves two breakthrough curves obtained using solutions at two initial ph values keeping other parameters constant. little effect is observed (about 10% more) in break time as ph increased from 4.0 to 5.5. this is a well-known observation in metal uptake studies which is explained by the fact that the concentration of hydrogen ions (h+), which competes with cd ions for attachment on zeolite sites, decreases with ph increase thus enhancing cd ions removal. hence, the bed reaches break point earlier at ph 5.5 than in the case of ph 4.0. the changes of solution ph during cadmium ion uptake in a typical experiment are shown in fig. 11. starting with feed solutions initially adjusted to ph 4.0 or 5.5; the solution ph during sorption has risen sharply in the first 15 minutes to 8.49.0 then declined slowly and leveled off at 6.26.5. based on the discussion in section 3.2, chemical precipitation in the form of cd(oh)2 will occur at when ph 8.0-8.5 is exceeded thus masking the sorption process. the reason for this sharp rise in ph during the initial sorption stage is explained in the light of the kinetics of ion exchange occurring in the bed. both the sodium release and cadmium uptake occur initially at high rates in the upper fresh zone of the bed where the driving forces for these processes are high. at later stages down the bed, the driving forces of these processes will decrease and solution ph is levelled off. it is recommended, therefore, that ph inside zeolite beds should also be maintained below 8.0 to avoid deposition of precipitate on the surface of particles. such deposition will make it difficult and impractical to regenerate the zeolite bed for further use. 3.5 mechanism of cd uptake based on the unique structure of zeolite discussed earlier which is characterized by being both crystalline and porous, the removal of metal ions can be described as ion exchange, sieve action or both depending on the relative ionic sizes and pore diameter. the average pore diameter of the investigated size cuts of zeolite is 1.75 å (table 1) which indicates a macroporous structure. according to the iupac standard, pores larger than 50 nm (0.5 å) are considered as macropores. the radii of the crystal, cd2+, and hydrated, cd(oh)+, ions are 0.97 and 5.0 å (pauling, 1970). the following sequence compares ionic sizes to average zeolite pore size: ionic radius of cd2+ < zeolite average pore size < ionic radius of cd(oh)+ this comparison reveals that it is possible for cd2+ ions, the major cadmium species present at the experimental conditions (ph < 8.0) to diffuse inside the pores of zeolite in addition to their sorption on the external surface. however, the hydrated ion, cd(oh)+, is too large to penetrate into the pores and thus its removal is restricted to ion exchange at the external surface. this means that two mechanisms of uptake are plausible: (a) ion exchange process which may occur at the external surface or at the active sites inside the large pores which may be represented by the following chemical equation: cd2+(aq) + 2 na+ (z) cd2+ (z) + 2 na+(aq) (6) (b) sieve action followed by physical adsorption on the internal surface of zeolite pores. there are similar observations of these phenomena reported in the literature (lee and moon, 2001; sprynskyy, et al. 2006). 4. conclusions jordanian phillipsite tuff is a cost-effective material for the removal of cadmium ions from aqueous solutions. the study has shown that zeolite dose (slurry concentration) and initial cadmium ion concentration are key factors influencing equilibrium uptake.slurry ph increased slightly with sorption progress. equilibrium uptake is best modeled by langmuir model implying monolayer cover0 50 100 150 200 250 300 350 400 450 time (min) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 c / c o 4.0 5.5 initial solution ph: fix e d conditions : be d de pth: 14.4 cm flow ra te : 15 m l/m in co = 50 m g/l pa rticle size : 500-710 mµ figure 10. breakthrough curves at different initial solution ph values 0 50 100 150 200 250 300 350 400 450 time (min) 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 e fflu e n t p h fixed conditions : bed depth: 14.4 cm flow rate: 15 ml/min co = 50 mg/l particle size: 500710 mµ initial solution ph: 4.0 5.5 figure 11. changes in effluent solution ph with time in zeolite fixed-bed c / c o time (min) e ff ic ie nt p h time (min) 10 the journal of engineering research vol. 6, no. 2 (2009) 1-11 age at specific charged sites. breakthrough curves obtained demonstrated the applicability of the tested zeolite mineral in continuous fixed bed column operation. the combination of bed height and solution flow rate are the two parameters to adjust for enhanced bed efficiency at a given cadmium ion concentration in the feed. solution chemistry at ph < 8.0 as well as pore diameter of zeolite supports cadmium uptake mechanism by both ions exchange and sieve action. references al-haj-ali, a. and el-bishtawi, r., 1997, "removal of metal ions using jordanian zeolite tuff", j. of chem. technol. biotechnol., vol. 69, pp. 27-34. al-haj-ali, a. and al-hunaidi, t., 2004, "breakthrough curves and column design parameters for sorption of lead ions by natural zeolite", environmental technology, vol. 25, pp. 1009-1019. alloway, b.j., 1995, "heavy metals in soil", wiley, new york. alvarez-ayuso, e.; garcia-sanchez, a. and querol, x., 2003, "purification of metal electroplating wastewaters using zeolites", water research vol. 37, pp. 4855-4862. berber-mendoza, m.s., leyva-ramos, r., alonsodavilla, p., mendoza-barron, j. and diaz-flores, p.e., 2006, "effect of ph and temperature on the ionexchange isotherm of cd(ii) and pb(ii) on clinoptilolite", j. of chem. technol. biotechnol., vol. 81, pp. 966-973. burgess, j., 1988, "ions in solution: basic principles of chemical interactions", ellis horwood, london, uk. camilo, c.; carmen, g. and paula, m., 2005, "sorption characteristics of heavy metal ions by a natural zeolite", j. of chem. technol. biotechnol. vol. 80(4), pp. 477-481 cincotti, a., mameli, a., locci, a. m., orru, r. and cao, g., 2006, "heavy metal uptake by sardinian natural zeolites: experiment and modeling", industrial & eng. chem. research vol. 45, pp. 1074-1084. dabrowski, a., hubicki, z., podkoscielny, p. and robens, e., 2004, "selective removal of the heavy metal ions from waters and industrial wastewaters by ionexchange method", chemosphere, vol. 56, pp. 91106. dwairi, i. m., 1992, "evaluation of jordanian zeolite tuffs for decontamination and immobilization of cs", dirasat, vol. 19b, pp. 61-73. friberg, l.m., piscaor, g.f. and nordberg, t., 1974, "cadmium in the environment", second edition, crc press, cleveland, us. garcia-sanchez, a., alvarez-ayuso, e. and jimenez de blas, o., 1999, "sorption of heavy metals from industrial wastewater by low-cost mineral silicates", clay minerals, vol. 34, pp. 469-477. genc-fuhrman, h., mikkelson, p. and ledin, a., 2007, "simultaneous removal of as, cd, cr, cu, ni and zn from stormwater: experimental comparison of 11different sorbents", water research, vol. 41, pp. 591-602. hedstrom, a., 2001, "ion exchange of ammonium in zeolites: a literature review", j. of environ. eng., vol. 127(8), pp. 673-681. inglezakis, v.j. and grigoropoulou, h., 2004,"effects of operating conditions on the removal of heavy metals by zeolite in fixed bed reactors", j. of hazardous materials, vol. 112(1-2), pp. 37-43. kocaoba, s.; orhan, y. and akyuz, t., 2007, "kinetics and equilibrium studies of heavy metal ions removal by use of natural zeolite", desalination, vol. 214, pp. 110. lee, d.h. and moon, h., 2001, "adsorption equilibrium of heavy metals on natural zeolites", korean j. of chem. eng., vol. 18(2), pp. 247-256. ouki, k.s., cheeseman, c.r. and perry, r., 1994, "natural zeolite utilization in pollution control: a review of applications to metals' effluents", j. of chem. technol. biotechnol., vol. 59(2), pp. 121-126. pauling, l., 1970,"general chemistry", w.h. freeman & co., new york. rangel-mendez, j.r. and streat, m., 2002, "adsorption of cadmium by activated carbon cloth: influence of surface oxidation and solution ph", water researech, vol. 36, pp. 1244-1252. rashdan, z f., 2000, "investigation of natural zeolitic tuffs on their ability for sewage cleaning purposes", phd. thesis, dept. of chemistry, oldenburg university, oldenburg, germany, pp. 102107. sand, l.b. and mumpton, f.a., 1978, "natural zeolites: occurance, properties, use", pergamon press, oxford, uk. singh, s.p., ma, l.q. and harris, w.g., 2001," heavy metal interactions with phosphatic caly: sorption and desorption behavior", j. of environmental quality, vol. 30, pp. 1968-1975. sips, r., 1948, "on the structure of a catalyst surface", j. of chemical physics, vol.16, p.490. sprynskyy, m., buszewski, b., terzyk, a. and namiesnik, j., 2006, "study of the selection mechanism of heavy metal (pb2+, cu2+, ni2+ and cd2+) adsorption on clinoptilolite", j. of colloid & interface sci., vol. 304, pp. 21-28. srivastava, p., singh, b. and angrove, m.j., 2004, "competitive adsorption of cadmium (ii) onto kaolinite as affected by ph", 3rd australian new zealand soils conference, university of sydney, australia, 5-9 december, 2004. stumm, w. and morgan, j.j., 1996, "aquatic chemistry: chemical equilibria and rates in natural waters", third edition, wiley, new york. takac, s., calik, g., aytar, m. and t. ozdamar, t., 1998, "separation kinetics of 1-phenyl-amine by ionexchange process", biochem. eng. j., vol. 2, pp. 101-112. 11 the journal of engineering research vol. 6, no. 2 (2009) 1-11 vaca-mier, m., callejas, r.l., gehr, r., cisneros, e.e.j. and alvarez, p.j., 2001, "heavy metal removal with mexican clinoptilolite: multi-component ionic exchange", water research, vol. 35(2), pp. 373-378. who (world health organization), 2004, "guidelines for drinking water quality", third edition, who, geneva. the journal of engineering research (tjer), vol. 17, no. 1, (2020) 1-10 *corresponding author’s e-mail: c.alalouch@squ.edu.om doi:10.24200/tjer.vol17iss1pp1-10  sustainable construction in oman: the potential role of the industrialized building systems m. s. saleh, and c. alalouch * civil and architectural engineering department, college of engineering, sultan qaboos university, p.o. box 33 al-khoudh, muscat 123, oman abstract: the omani construction industry has been recently witnessing a boosted expansion associated with the notable growth in population and economy. in comparison with its neighboring countries, the construction industry in oman is criticized for the lack of sustainability practices. recently, the government and professional bodies have taken proactive actions to promote sustainable construction practices. the industrialized building systems (ibs) could play an important role in the omani construction industry in the future due to its environmental, economic, and social benefits. however, the lack of evidence on the contribution of ibs to the omani construction industry suggests that this approach is not yet at the top of the stakeholder agenda. this paper focuses on the potential contribution of ibs to sustainable construction practices in oman. through a literature review, a comprehensive list of sustainability factors affecting ibs applications in the omani construction industry was developed. a stakeholder survey of omani construction industry including clients, engineers, and contractors was conducted to understand their perception on the importance of these factors. statistical analysis revealed that the environment-related and the implementation-related factors are more important than the social-related factors. the economic-related factors were not found to be significantly more important than any other group of factors. moreover, construction-time was found to be the most important and the most influential factor among the 62 factors considered in this study. the paper concludes with recommendations to promote and improve applications of industrialized building system in the omani construction industry. keywords: construction industry; industrialized building system (ibs); oman; sustainable construction; sustainability. المصنعة اإلنشاء أنظمةاإلنشاء المستدام في عمان: دور * ش. العــلــوش وح ص. صــالـم. في قتصاديالياد أعداد السكان وتحسن الوضع افي عمان نهضة كبيرة بسبب ازد التعميرو نشاءقطاع اإليشهد :الملخص ما قورن بالدول إذاستدامة مبادئ اإل التطبيق الكافي ل انعداميتعرض للنقد بسبب . إال أن هذا القطاع مازال العقود األخيرة اإلنشاء أنظمةتشكل ن كل من الحكومة والهيئات المهنية تسعى التخاذ إجراءات بهذا الخصوص. إة. ونتيجة لهذا فالمجاور جتماعية. اقتصادية واالعماني لما لها من فوائد بيئية و والتعمير نشاءفرصة لتحسين وضع االستدامة في قطاع اإل المصنعة . قطاعهذا الب ات المعنيةئيس من أولويات الفيبدو انه لعمان واألنظمة ال يزال خجوال في استخدام هذه هذا فإن بالرغم منو اعتمدت في دفع عجلة االستدامة في عمان. المصنعة اإلنشاء أنظمةن هدف هذه الدراسة هو تحديد مدى مساهمة إف ،ولذلك اعتمدت الدراسة على مراجعة المصادر العلمية المختلفة وذلك ,أوال :ين بحثيتين لتحقيق الهدف المرجوتالدراسة على طريق في عمان. ثانيا، استخدمت الدراسة المصنعة اإلنشاء أنظمةتطبيق على لعوامل التي ممكن أن تأثر باالستنباط قائمة متكاملة عوامل وأولوياتهم في المهندسين والمقاولين وأصحاب المشاريع بخصوص أهمية هذه الهم رأي كل من لفاستبيانا شامال حصائيا. أظهرت النتائج أن العوامل البيئية والعوامل التطبيقية أكثر أهمية من إتم تحليل نتائج االستبيان التعامل معها. باإلضافة الى ذلك فقد تبين أن تكن العوامل االقتصادية أكثر أهمية من أي من العوامل األخرى. مالعوامل االجتماعية. بينما ل ص وقت اإلنشاء" هو العامل األهم على االطالق من بين االثنين وستين عامال. بناء على هذه النتائج فإن الدراسة تستخل" اإلنشاء أنظمةلتسهيل استخدام والتعمير العماني نشاء أصحاب القرار في قطاع اإلمجموعة من التوصيات للمختصين و تشجيعها في هذا القطاع.كمساهمة في تحسين االستدامة و المصنعة اإلنشاء ؛عمان ؛أنظمة اإلنشاء المصنعة ؛أنظمة اإلنشاء مسبقة الصنع ؛ التعميرنشاء وقطاع اإل الكلمات المفتاحية: االستدامة. ؛المستدام sustainable construction in oman: the potential role of the industrialized building systems 2 1. introduction in spite of the cutbacks of public investments and the broad slowdown in the economy due to reduction in oil prices, the construction sector in oman is still growing at a steady rate. in real terms, the construction projects registered in oman were reported to grow by 9.4% annually during 2012–2016 (timetric, 2017). the construction sector is the largest segment of the non-oil economy in oman and as such, it remains a pillar in the government endeavor for diversification of the country’s economy toward its long-term vision 2040 (oxford business group, 2018). achieving and maintaining sustainable development has become increasingly more prevalent for many countries over the last decades. in line with its neighboring countries, oman has made plans for economic diversification to reduce its dependency on the fast depleting oil reserves, through building competencies across other sectors of the economy (the world bank). subsequently, the government had allocated huge investments on infrastructure, housing projects, and industrialization which had led to a tremendous boom across the construction industries (oxford business group, 2013). this huge development is accompanied by the raising demand to maximize the efficient use of natural resources. the current global economic slowdown had put pressure on the omani government to take steps to address issues related to quality, sustainability, and standardization in the construction sector to meet its needs and to bring its construction industry in par with international standards. the government of oman has taken several actions to promote the development of sustainable construction practices. in spite of these efforts, many recent studies showed that the construction industry in oman still lacks evidence of sustainability practices. saleh and alalouch (2015) have addressed several major challenges to the application of sustainable construction in oman and grouped them into four major clusters: economic, capacity/professional, societal, and technological challenges. similarly, powmya and abidin (2014) found that most of the stakeholders in oman believe that applying sustainable practices in oman would result in a dramatic raise in construction cost without any quantifying benefits, a view that is considered a challenge by itself. moreover, alalouch et al. (2016), in their comparative study of two energy-efficient houses in oman and qatar, concluded that applying sustainability practices in oman still at early stages which require vast shifts in the construction industry. in a more recent study, the energy-saving potential in a residential building in oman was explored. the result showed the potential for great saving in energy when energy passive measures are put in place (alalouch et al., 2019). in fact, several recent research efforts have been focusing on the urban expansion in oman (e.g. alalouch et al. 2019; alkamali et al. 2017). on the other hand, the industrialized building systems can respond to many challenges in the construction sector in oman. zhang sees the ibs to be the future of the construction industry worldwide due to its ability for capitalizing on the strength of the industrial production (zhang, skitmore, and peng, 2014). industrialization from the construction point of view is considered part of the modernization process through developing methods of production and technology systems, where production operations are centralized, mechanized, and mass productionoriented (lessing, 2006). warszawski (1999) explained that an industrialization process is an investment in technology facilities and equipment, aiming at maximizing production output, minimizing demand on labor, and improving quality (nawi et al., 2014). therefore, the objective of this study is to identify the sustainability criteria for ibs and assess their importance as perceived by the construction industry stakeholders in oman. recommendations that are proposed in this study could be used by different stakeholders to help to promote and improving the application of ibs in the omani construction industry. 2. sustainability aspects of ibs though prefabrication and standardization terminologies are still in use, industrialized building system (ibs) has become the most common term used to represent those terminologies due to the research context of the malaysian construction industry (nawi, et al., 2014). industrialized building construction is defined as a generic process of rationalization and standardization of work processes in the construction industry to reach cost efficiency, higher productivity and quality (cib, 2010). a more elaborate definition for industrialized building construction is a change of thinking and practices to improve the production of construction and to produce high quality, custom-built, and environmentfriendly buildings through an integrated process, optimizing standardization, organization, cost, value, mechanization and automation (cib, 2010). the industrialized building systems (ibs) is considered by several researchers to be the cornerstone rising for improving the construction industry, e.g. (cook, 2005; hampson and brandon, 2004). this is due to the manufacturing of construction components in a controlled environment, either at-site or off-site, then installed or assembled into construction works. although many studies have focused on the implementation of the ibs, there is a debate on its potential effect on the construction industry. for example, some studies considered that ibs is more expensive than conventional construction systems (birkbeck and scoones, 2005); while other studies found that project cost had been reduced remarkably due to ibs use over time (goodier and m. s. saleh and c. alalouch 3 table 1. ibs features that promote sustainability (richard, 2006). 2007; gibb and isack, 2003). on the other hand, ibs has been linked to green construction (jaillon and poon, 2009; zhang and skitmore, 2012). the sustainability aspects of implementing ibs practices can encourage the coordination between the built environment and construction processes, thus creating urban settlements that affirm healthy environments, social interaction, and encourage economic development. richard identified eight key sustainability benefits of applying prefabrication in construction (richard, 2006); as listed in table 1. the following highlights the environmental, economic and social sustainability aspects of ibs in construction. 2.1. environmental aspects the environmental benefits derived from implementing ibs practices will vary according to the specific system implemented. with no doubt, typical traditional construction could have a better environmental performance over some prefabrication systems, but still the potential exists for ibs practices to have better environmental performance if they were well implemented and managed. as reported by waskett (2001), one specific program being developed with the european-commission backing has the potential anticipated benefits associated with ibs of a 50% reduction in the amount of water used for the construction of a typical house, and a 50% reduction in the use of quarried materials used in the construction. another environmental benefit of using ibs elements is that conventional formworks could be eliminated or greatly reduced, which will reduce the issue of construction waste and its subsequent environmental problems. prefabrication can also promote a safer working environment for construction workers. the need of a large number of workers and raw materials are also noticeably reduced at the sites. in addition, as products are manufactured in a controlled environment and standard in sizes, wastage will be minimized at both factory and construction sites (bari, et al., 2012). equally important, ibs has the potential to accommodate modern and smart building energy systems (mechanical, electrical, fire safety, etc.) that are known for their high environmental performance as ibs has been seen as the future of the construction industry (zhang, skitmore, and peng, 2014) and linked to green construction (jaillon and poon, 2009). 2.2. economic aspects there are many economic benefits for implementing ibs practices in construction; among them are cost savings, quality, and speed of construction. operation and maintenance costs are reduced due to the highquality characteristics of prefabricated components. yunus and yang (2011) mentioned that the cost of labor and materials will be also reduced significantly by adopting ibs. on the other hand, kamaruddin et al. (2013) revealed that a considerable number of contractors are unwilling to take part in projects that include ibs. this is due to the higher costs associated with the procurement of materials and hiring welltrained labour despite the overall cost saving it offers. jaillon and poon (2009) concluded that adopting ibs practices in a project could lead to an overall reduction of 16% of labour costs, as well as 15% of total construction time. in addition, goodier and gibb (2007) and ding (2008) emphasized that due to its standardization process; ibs will result in a significant time saving, improve quality control, and reduce material. to achieve these benefits all of the construction team must realize and understand the sustainability potential of ibs. this review demonstrates the sustainability potential of ibs as one of the modern construction systems that could achieve better economic viability if it is properly utilized and implemented. figure 1. potential reduction in cost when industrialized construction is used (cib, 2010) . criteria 1.  reproduction, which ensures higher productivity and quality 3.  working conditions to avoid losing time through severe weather 4.  waste reduction due to modular coordination, bulk purchasing and factory  applied finishes 5.  factory conditions, which avoid later repairs 7.  flexible components, which allow for planning changes 6.  precision  in production keeping the construction site clean and free of debris adaptability 7.  demountable  components,  which allow  for  a  major  reconfiguration and relocation without demolition waste perspective economy 2.  simplified  processes, which reduces the total energy involved factory  production sustainable construction in oman: the potential role of the industrialized building systems 4 2.3. social aspects adopting prefabrication in construction will have a remarkable effect on the labour market. investing in automation will help in reducing the dependence on unskilled workers and foreign labours (el-abid and ghazalia, 2015). prefabrication can also promote healthier working environments as a result of cleaner and safer construction sites. in addition, it improves the health of the building’s occupants by controlling moisture and volatile organic compounds levels in a factory setting. controlling the quality of building components will lead to minimizing risks of chronic health issues of occupants (ghaffarianhoseini, et al. 2018). in urban areas, traffic disruption to congested roadways due to construction activities is very crucial. traditional construction practices usually result in pollution, noise, dust, and require more trips to the construction site. on the contrary, ibs can reduce onsite construction activities and time, which inevitably reduce the disturbance associated with construction. in addition, prefabricated units are normally transported in large sizes leading to a limited number of trips to the site, and thus creating less disruption to the congested city traffic and the local community. 3. development of a comprehensive list of sustainability factors for ibs application a literature review was conducted aiming at developing a comprehensive list of typical sustainability factors affecting the quality of ibs implementation. although the implementation of ibs practices vary from a country to another, some applications reflect the adoption of ibs in developing countries like oman. among these applications are ‘standardization’ (gann, 1996; lessing, stehn, and ekholm, 2005; pasquire, gibb, and blismas, 2004), ‘prefabrication’ (dawood, 1996; song, 2005) and ‘system building’ (finnimore, 1989; luo, riley, horman, and kremer, 2008). sustainability factors related to these applications were reviewed, summarized and categorized according to the main pillars of sustainability, which are economy, environmental, and social, in addition to implementation-related factors. it is worth mentioning that each factor could have been broken further into “indictors”, but this would have resulted in an unmanageable number of variables. equally important, this study focuses on identifying and evaluating the overarching factors that might facilitate or hinder the implementation of ibs in oman. hence, the literature review was limited to the factors only as summarized in table 2. table 2. sustainability factors for ibs application. economic-related factors criteria code factors cost ec-01 maintenance and operation costs ec-02 disposal costs ec-03 life cycle costs ec-04 initial construction costs ec-05 material costs ec-06 labour cost time ec-07 construction time ec-08 lead-times ec-09 speed of return on investment ec-10 transportation and lifting ec-11 production ec-12 design stage adoption social-related factors criteria code factors social issues sc-01 local economy sc-02 participation and control sc-03 community disturbance sc-04 traffic congestion sc-05 public awareness sc-06 public participation sc-07 principles and values labour market sc-08 influence on job market sc-09 knowledge and skills sc-10 labor availability safety and health sc-11 workers’ health and safety sc-12 working conditions sc-13 disaster preparedness design and arch issues sc-14 site attributes sc-15 aesthetic options environment-related factors criteria code factors waste en-01 waste generation en-02 waste disposal energy consumption en-03 embodied energy en-04 design and construction en-05 operational energy recycle en-06 recyclable / renewable contents en-07 reusable / recyclable elements pollution en-08 site disruption en-09 pollution generation en-10 environment administration en-11 ecology preservation en-12 health of occupants (indoor air quality) en-13 inclusive environment resource consumption en-14 water consumption en-15 land use en-16 material consumption implementation-related factors criteria code factors technical quality tc-01 durability tc-02 defects and damages tc-03 loading capacity tc-04 integration of building services tc-05 integration of supply chains tc-06 constructability tc-07 usage efficiency tc-08 adaptability and flexibility tc-09 technology enforcement and regulations im-01 standardization im-02 governance im-03 legislation im-04 policy and strategy match im-05 building capacity im-06 design standard and project function im-07 project control guidelines im-08 integrated environmental and economic program im-09 procurement system m. s. saleh and c. alalouch 5 4. data collection and analysis 4.1. survey design and sample characteristics the survey aimed to investigate the perspective of the omani construction industry’s stakeholders on the relative importance of sustainability factors for ibs applications in oman. a pilot study was conducted to ensure the effectiveness of the survey; accordingly, it was developed and distributed to the target groups. the questionnaire survey consisted of three parts. the first part focused on collecting background information about the respondent including years of experience in the construction industry, profession, organization, and the number of ibs projects. in the second part, respondents were asked to rate the importance of each sustainability factor; presented in table 2; with respect to the ibs applications. a fivestep likert-scale was used, in which (1) refers to “least significant” and (5) “most significant”. to ensure consistency in understanding the factors across participants, the questionnaire included a clarifying definition for each factor. in the third part, respondents were given the opportunity to provide supplementary factors in an open-ended question. a total of 80 questionnaires were distributed and only 54 responses were received, which gives an overall response rate of 67.5%. among those 54 respondents, 19% were from clients/developer sub-group, 43% from consultants/designers, 19% from governmental institutes, and 13% from contractors. furthermore, 50% of the respondents were practicing as engineers, 26% architects, and 15% project managers. all of the survey participants had professional experience with industrialized building systems. about 52% had an experience of less than 5 years, while 48% had more than 5 years of ibs experience. fig. 2 shows a detailed distribution of respondents among different sub-groups. 4.2. data analysis and discussion in order to identify the most important factors as perceived by the respondents, the data was analyzed statistically. for the purpose of this study, descriptive analysis and tests of differences of the important data of the factors considered are reported. 4.2.1. the importance of the categories of factors the first question in this study was to identify the most important group of factors as perceived by the respondents. hence, the mean score for the factors in each one of the four categories was calculated across all responses. fig. 3 shows that the environmentrelated factors are seen as the most important group of factors followed by the implementation-related factors. the least important group was the socialrelated factors whereas the economic-related factors came in the third place. figure 3. mean score of each group of factors. table 3. normality of the variables. e co n o m ic re la te d e n v ir o n m en tre la te d s o ci a lre la te d im p le m en ta ti o n -r el a te d kolmogorovsmirnov 0.96 0.11 0.109 0.109 sig. 0.2 0.1 0.16 0.16 figure 2. sub-groups of respondents. sustainable construction in oman: the potential role of the industrialized building systems 6 table 4. differences between the variables t (sig.). paired-sample t-test economic-related environment-related social-related implementation-related economic-related 0 environment-related -1.47 (0.148) 0 social-related 0.63 (0.53) 3.1 (0.00)* 0 implementation-related -1.1(0.28) 0.43 (0.67) -2.39 (0.02)** 0 * significant at 0.01, ** significant at 0.05 figure 4. overall importance of the ibs sustainability factors based on the ranking of these factors. figure 5. the influence of ibs sustainabililty factors. each data point represents one factor and the associate number represents the factor's overall rank. m. s. saleh and c. alalouch 7 table 5. statistical analysis and ranking of sustainability factors for ibs applications in oman. sustainability performance factors mean value* standard deviation overall importance** overall influence *** economic-related factors ec-07 construction time 4.11 1.09 h q1 ec-02 disposal costs 3.78 1.16 m-h q2 ec-11 production 3.57 1.25 m q4 ec-08 lead-times 3.56 1.00 m q3 ec-09 speed of return on investment 3.56 1.18 m q4 ec-03 life cycle costs 3.50 1.11 m q3 ec-12 design stage adoption 3.50 1.00 m q3 ec-01 maintenance and operation costs 3.37 1.22 m-l q4 ec-04 initial construction costs 3.37 1.26 m-l q4 ec-05 material costs 3.37 1.19 m-l q4 ec-06 labour cost 3.35 1.26 m-l q4 ec-10 transportation and lifting 3. 1.00 l q3 environment-related factors en-01 waste generation 3.87 1.05 m-h q1 en-09 pollution generation 3.85 1.02 m-h q1 en-02 waste disposal 3.83 1.00 m-h q1 en-10 environment administration 3.80 1.05 m-h q1 en-07 reusable / recyclable elements 3.76 1.10 m-h q1 en-12 health of occupants (indoor air 3.76 1.26 m-h q2 en-14 water consumption 3.67 1.30 m q2 en-04 energy consumption in design and 3.65 1.08 m q1 en-15 land use 3.61 1.14 m q2 en-06 recyclable / renewable contents 3.57 1.13 m q4 en-11 ecology preservation 3.57 1.14 m q4 en-08 site disruption 3.48 1.00 m q3 en-05 operational energy 3.46 1.18 m q4 en-16 material consumption 3.46 1.14 m q4 en-03 embodied energy 3.39 1.19 m-l q4 en-13 inclusive environment 3.37 1.15 m-l q4 social-related factors sc-11 workers’ health and safety 3.89 1.19 m-h q2 sc-03 community disturbance 3.80 1.16 m-h q2 sc-12 working conditions 3.80 1.05 m-h q2 sc-09 knowledge and skills 3.72 1.11 m q1 sc-13 disaster preparedness 3.70 1.09 m q1 sc-05 public awareness 3.59 1.04 m q3 sc-10 labor availability 3.59 1.17 m q4 sc-14 site attributes 3.54 1.06 m q3 sc-01 local economy 3.43 1.13 m-l q4 sc-04 traffic congestion 3.35 1.03 m-l q3 sc-06 public participation 3.31 1.16 m-l q4 sc-16 physical space 3.28 1.12 m-l q4 sc-15 aesthetic options 3.20 1.14 l q4 sc-07 principles and values 3.15 1.16 l q4 sc-08 influence on job market 3.15 1.22 l q4 sc-02 participation and control 3.11 1.11 l q3 implementation-related factors tc-02 defects and damages 4.00 1.13 h q2 tc-05 integration of supply chains 3.83 1.07 m-h q1 tc-06 constructability 3.83 1.15 m-h q2 im-01 standardization 3.76 1.23 m-h q2 tc-01 durability 3.72 1.05 m q1 tc-03 loading capacity 3.72 0.93 m q1 tc-07 usage efficiency 3.63 1.16 m q2 im-05 building capacity 3.61 1.08 m q3 im-06 design standard and project function 3.59 0.97 m q3 tc-04 integration of building services 3.57 1.00 m q3 im-07 project control guidelines 3.57 1.00 m q3 im-03 legislation 3.54 1.07 m q3 im-08 integrated environmental and 3.54 1.11 m q3 im-04 policy and strategy match 3.50 1.04 m q3 im-09 procurement system 3.46 1.23 m q4 im-02 governance 3.37 0.94 m-l q3 tc-08 adaptability and flexibility 3.31 1.24 m-l q4 tc-09 technology 3.30 1.19 m-l q4 * the order of the factors reposes the level of importance within the group. ** based on fig. 4. h= high, m-h= medium-high, m= medium, m-l= medium-low, l= low. *** based on fig. 5. q1= high importance, high consensus, q2= high importance, low consensus, q3= low importance, high consensus, q4= low importance, low consensus. sustainable construction in oman: the potential role of the industrialized building 8  however, the figure shows that the differences between the groups are marginal. it was necessary therefore to test the statistical significance between the differences in the mean scores of the groups. therefore, the normality of the data was examined using the one-sample kolmogorov-smirnov test (k-s test). the significant value of this test (p < 0.05) indicates a deviation from normality. the results of this test indicated that all variables are normally distributed as shown in table 3. hence, the pairedsample t-test was used to check if the differences between the importance of the variables are statically significant. the results of the test of differences is reported in table 4, which shows that the environment-related factors were seen as significantly more important than the social-related factors (t=3.1, df= 53, p < 0.01). whereas, the participants valued the implementationrelated factors significantly more than the socialrelated factors (t= -2.39, df= 53, p < 0.05). this suggests that the environment-related and the implementation-related factors are more important than the social-related factors. in contrast, no significant differences were found between the economic-related factors and any other variable. 4.2.2. the overall importance of the factors the next step was to identify the most important factors among the 62 factors considered in this study. therefore, the average score for each factor was calculated as shown in fig. 4. a close examination of the trend line revealed a clear break in the mean score at four points, which suggests that the factors can be split into five levels in terms of importance. the level with the highest important factors contains only two factors namely, construction time and defects and damages. on the other hand, the low importance level has five factors i.e. transportation and lifting, aesthetic options, principles and values, influence on job market, participation and control. although fig. 4 provides an insight into the structure of the importance of each factor based on a simple ranking of the mean score, it does not take into account information regarding the consensus among the participants’ answers. in order to explore the relationship between the level of importance and the level of consensus, the mean score and the standard deviation were normalized to render them comparable. fig. 5 shows a scatter plot between the two normalized variables. in this figure, the average of each axis was used to split the factors into four quarters based on their level of influence (i.e. importance and consensus). the results show that there is a general consensus about twelve important factors out of the 62 factors included in this study which are shown in the figure in the q1 quarter (q1= high mean value with low st. deviation). these factors are: 1-construction time, 4-waste generation, 5-pollution generation, 6-waste disposal, 7integration of supply chains, 9-environment administration, 13-reusable/recyclable elements, 16knowledge and skills, 17-durability, 18-loading capacity, 19-disaster preparedness, 21-energy consumption in design and construction. this suggests that these factors have the highest potential to influence the adoption of ibs in the omani market. table 5 shows the mean score, standard deviation, importance level and influence level for each factor. construction time was considered the most important and the most influential factor across all factors. in the environment-related factors attention should be given to waste generation, pollution generation, waste disposal, environment administration, and reusable/recyclable elements. as for the socialrelated factors, workers’ health and safety, community disturbance and working conditions were seen as the most important and most influential factors in this group. the results for the implementation-related factors are not inclusive. 5. conclusion this study presented the sustainability factors affecting the implementation of ibs in the context of oman construction industry. the research identified 62 factors based on the main pillars of sustainable development, consisting of 12 economy-related factors, 16 environment-related factors, 16 socialrelated factors, and 18 implementation-related factors. a questionnaire was conducted to investigate the perspective of the omani construction industry stakeholders on the relative importance of these factors. analysis of the collected data revealed that environment-related factors were generally considered the most important factors related to ibs, followed by implementation-related factors and social-related factors, while economic-related factors were the least important. the top three environment-related factors included en-01 (waste generation), en-09 (pollution generation), and en-02 (waste disposal). the top three implementation-related factors include tc-02 (defects and damages), tc-05 (integration of supply chains), and tc-06 (constructability). the top three social-related factors include sc-11 (workers’ health and safety), sc-03 (community disturbance), and sc12 (working conditions). finally, the top three economic-related factors include ec-07 (construction time), ec-02 (disposal costs), and ec-11 (production). equally important, construction time was found to be the most important factor and the one that has the highest consistency level concerning its importance. the omani government and construction professionals can use the findings of this study as guidelines to focus their effort in promoting the application of ibs in oman by emphasizing the factors that matter the most to the stakeholders while fitting within the government’s shortand long-term strategy and targets. on the other hand, the study provides essential information to the decision-makers in the construction industry to understand the key m. s. saleh and c. alalouch 9 areas where the implementation of ibs practices needs improvement, leading to implement proper actions and management strategies to maximize the benefits of ibs practices. in addition, the findings of this study lay the groundwork for future investigation of effective actions to manage these sustainability factors in the omani construction industry. future research in this area should focus on developing measurable “indicators” for each factor that are meaningful for the omani construction industry and explore the differences in the importance of these factors across building types. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. acknowledgment the authors would like to acknowledge the civil and architectural engineering department and sultan qaboos university for their support. the authors would like also to thank everyone who was involved in the data collection for this research. references alalouch c, al-hajri s, naser a, al hinai a (2019), the impact of space syntax spatial attributes on urban land use in muscat: implications for urban sustainability. sustainable cities and society, 46: 101417. https://doi.org/10.1016/j.scs.2019.01.002 alalouch c, alsaadi s, alwaer h, al-khaled k (2019), energy saving potential for residential buildings in hot climates: the case of oman. sustainable cities and society, 46:101442. https://doi.org/10.1016/j.scs.2019.101442 alalouch c, saleh m s, al-saadi s (2016), energyefficient house in the gcc region. procedia social and behavioral sciences, 216:736-743. alkamali n, alhadhrami n, alalouch c (2017), muscat city expansion and accessibility to the historical core: space syntax analysis. energy procedia, 115:480-486. https://doi.org/ 10.1016/j.egypro.2017.05.044 bari n, abdullah n, yusuff r, ismail n, jaapar a (2012), environmental awareness and benefits of industrialized building systems (ibs). procedia social and behavioral sciences 50: 392-404. birkbeck d, scoones a (2005), prefabulous homes: the new housebuilding agenda. constructing excellence: london. cib. (2010). industrialised construction: state of the art report, tg 57 publication. international council for research and innovation in buildings and construction: zurich. cook b (2005), an assessment of the potential contribution of prefabrication to improve the quality of housing: a scottish perspective. information quarterly 7(2): 50-55. dawood n (1996), “an integrated intelligent planning approach for modular construction,” in proc. of third congress on computers in civil engineering. anaheim, usa. ding d (2008), sustainable construction: the role of environmental assessment tools. journal of environmental management 86(3): 451. el-abid k m, ghazalia f m (2015), motivations and limitations of prefabricated building: an overview. applied mechanics and materials 802: 668-675. finnimore b (1989), houses from the factory. unwin hyman: london. gann d m (1996), construction as a manufacturing process? similarities and differences between industrialized housing and car production in japan. construction management and economics 14(5): 437450. ghaffarianhoseini a, alwaer h, omrany h, alalouch c, clements-croome d, tookey j (2018), sick building syndrome: are we doing enough? architectural science review 61(3): 99-121. https://doi.org/10.1080/00038628.2018.1461060 gibb a, isack f (2003), re-engineering through preassembly: client expectations and drivers. building research and information 31(2): 146-160. goodier c, gibb a (2007), future opportunities for offsite in the uk. construction management and economics 25(6): 585-595. hampson k, brandon p (2004), construction 2020: a vision for australia's property and construction industry. cooperative research centre for construction innovation for icon: australia. jaillon l, poon c (2009), sustainable construction aspects of using prefabrication in dense urban environment. journal of construction management and economics 26(9): 953-966. kamaruddin s, mohammad m, mahbub r (2013), enhancing the quality of life by adopting ibs: an economic perspective on mechanisation and automation. procedia: social and behavioral sciences 101: 71-80. lessing j (2006), industrialised house-building concept and processes, phd thesis, department of construction sciences, lund institute of technology: sweden. lessing j, stehn l, ekholm a (2005), “industrialised housing: definition and categorization of the concept,” in proc. of the in the 13th annual conference in the international group for lean construction, sydney, australia. luo y, riley r, horman m j, kremer g o (2008), decision support methodology for prefabrication decisions on green building projects. in p. mcdermott and m. m. a. khalfan (eds.), symposium on sustainability and value through construction procurement. university of salford: uk. sustainable construction in oman: the potential role of the industrialized building 10  nawi m n, abu hanifa f a, kamar k a, lee a, azman m n (2014), modern method of construction: an experience from uk construction industry. australian journal of basic and applied sciences 8(5): 527-532. oxford business group (2013), going green: sustainable building practices should help address housing shortages. in the report: oman 2013. oxford business group (2018), omani construction sector weathers external forces to continue posting solid growth. in فhe report: oman 2018. pasquire c, gibb a, blismas n (2004), off-site production: evaluating the drivers and constraints. paper presented at international group for lean construction: denmark. powmya, a., and abidin, n. z. (2014). the challenges of green construction in oman. international journal of sustainable construction engineering 5(1): 33-41. richard r b (2006), industrialised, flexible and demountable building systems: quality, economy and sustainability,” in proc. criocm2006: international research symposium on advancement of construction management and real estate, beijing. saleh m s, alalouch c (2015), towards sustainable construction in oman: challenges and opportunities. procedia engineering 118: 177-184. song j (2005), considering prework on industrial projects. journal of construction engineering and management 131(6): 723-733. the world bank (n.d.), oman country data. retrieved december 15, 2014, from http://data.worldbank.org/country/oman timetric (2017), construction in oman, key trends and opportunities to 2021. timetric: usa. waskett p (2001), current practice and potential uses of prefabrication. building research establishment ltd (bre): watford yunus r, yang j (2011), sustainability criteria for industrialised building systems (ibs) in malaysia. procedia engineering 14: 1590-1598. zhang x, skitmore m (2012), industrialized housing in china: a coin with two sides. international journal of strategic property management 16(2): 143-157. zhang x, skitmore m, peng y (2014), exploring the challenges to industrialized residential building in china. habitat international 41: 176-184. comp081011.qxd the journal of engineering research vol. 7, no. 1 (2010) 62-69 1. introduction the most popular techniques used for numerical simulations of various electronic devices are the finite difference and the finite element methods (fdm and fem). although, some other techniques, such as the boundary integral and monte carlo methods have been used in some cases, the two former methods are the most widely used despite of some of their disadvantages. the main drawbacks of these methods are long running time, large memory space requirements and conver____________________________________ corresponding author’s e-mail: hadj@squ.edu.om gence problems. in the present work, an alternative seminumerical approach is used for the simulation of electronic planar devices. the approach is based on a mathematical method known as the method of lines (mol) (ames, 1977, sadiku, et al. 2000; vietzorreck, et al. 2000; barcz, et al. 2003; vietzorreck, et al. 2004; pascher, et al. 2005; yan, et al. 2005; gonzalo, et al. 2006; plaza, et al. 2006 and chen, et al. 2007) which is a differential-difference technique. it is a versatile technique developed by determination of potential profile in planar electronic structures using a semi-analytical technique hadj bourdoucen**a, mokrane dehmasb and el-bachir yallaouic *a department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, pc 123, al-khoud, sultanate of oman **a communication and information research centre, sultan qaboos university, muscat, sultanate of oman b university of boumerdes, department of electrical engineering, 35000, algeria c department of mathematics and statistics, college of science, sultan qaboos university, po box 36, pc 123, al-khoud, sultanate of oman received 11 october 2008; accepted 13 september 2009 abstract: in this paper, a semi-analytical technique known as the method of lines (mol) with uniform and non-uniform discretization schemes is developed. the aim is to determine static potential profile in planar electronic structures. even though this method has been known for some time, there has been reported work on its application to planar semiconductor device analysis for voltage profile determination. since most current electronic devices are manufactured using planar and quasi-planar technology, the proposed algorithm is well suited for device analysis prior to fabrication. compared with known popular methods such as finite difference and finite elements methods, the proposed technique is relatively simple, more accurate and unlike other methods, has no convergence problem. in addition to this, its semi-analytical nature, which consists of reducing one computing dimension, allows saving significant memory and computation time. typical planar electronic structures are considered to demonstrate their suitability for these devices, and the obtained results are presented and discussed. keywords: planar structure, method of lines, potential profile, mol á«∏«∏ëàdg ¬ñ°t á«æ≤àdg ωg~îà°seéh ájƒà°ùe á«fhîμdeg πcé«¡d ~¡÷g íeóe ~j~– …hó©j ò°ûñdg , ¢séªmo ¿gô≤e ,ø°shoqƒh êém ::áá°°uuóóÿÿgg) •ƒ£ÿg á≤jôw º°seéh áahô©ÿg h á«∏«∏ëàdg ¬ñ°t á«æ≤àdg ôjƒ£j ” åëñdg gòg ‘mol~¡÷g íeóe ~j~– ¢vô¨h ∂dph ᪶àæe òzh ᪶àæe á∏°üøæe ä飣fl ™e ( ~©jh .ájƒà°ùe á∏°uƒe áñ°t §féñæd ~¡÷g íeóe ~j~– h π«∏ëàd iqƒ°ûæe çéëhcg ~lƒj ’ áfcéa ≥hé°s âbh òæe áahô©e á≤jô£dg √òg ¿cg ºzôhh . ájƒà°ùe á«fhîμdeg πcé«¡d øcé°ùdg ¥ô£déh á≤jô£dg √òg áfqé≤ã h .ájƒà°ùe áñ°t hcg ájƒà°ùe äé«æ≤j ωg~îà°séh ™æ°üj §féñædg ∂∏j ö∏zg ¿g å«m é¡©«æ°üj πñb á«fhîμdc’g §féñædg á°sgq~d éñ°séæe ìî≤ÿg ωrqgƒÿg .iônc’g ¥ô£dg πãe üqé≤àdg á∏μ°ûe øe ≈fé©j ’ h ábo ìccg é¡fg éªc é«ñ°ùf §°ùhcg ámî≤ÿg á«æ≤àdg ¿cg ~‚ ioh~ùg ô°uéæ©dg á≤jôw h ioh~ùg ¥qgƒødg á≤jôw πãe á©fé°ûdg ámî≤ÿg á≤jô£dg ≥«ñ£j ”h .üé°ù◊g øerh iôcgp øe πμd ≠déh ¢ øîh íª°ùj é‡ ~mgh ≈ñ°sƒm ~©h ¢ «øîj πª°ûj á«∏«∏– áñ°t á©«ñw é¡d ámî≤ÿg á«æ≤àdg ¿eéa ≥ñ°séÿ áaé°vc’éh .é¡«∏y ∫ƒ°ü◊g ” ≈àdg íféàædg á°ûbéæeh í«°vƒj ”h »°sé«b iƒà°ùe ≈fhîμdcg πμ«g áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.~¡÷g íeóe ,•ƒ£ÿg á≤jôw ,ájƒà°ùÿg πcé«¡dg : 63 the journal of engineering research vol. 7, no. 1 (2010) 62-72 r. pregla and co-workers for the analysis of planar microwave structures (schultz, 1980, pregla and pascher, 1989, pregla, 2002). for a given partial differential equation subject to some boundary conditions, all but one of the independent variables are discretized. this allows finding a one-dimensional analytical solution. the semianalytical nature of the mol makes the computational efforts much less intensive than the above-cited methods applied so far to microwave structures. this method has been used recently for the necessity of reducing long calculation times and therefore to significantly reduce the computation time (zhou, et al. 2000; pascher, et al. 2005; yan, et al. 2005; gonzalo, et al. 2006; plaza, et al. 2006 and chen, et al. 2007). the method of lines has been used by several authors for the analysis of microwave structures in both quasi-static and full-wave modes (schultz, 1980; pregla and pascher, 1989; pregla, 2002; barcz, et al. 2003; vietzorreck and pascher, 2004 and yan, et al. 2005). however, the method has not been extended to the voltage profile determination of planar semiconductor devices. in the following sections, the application of the mol to the analysis of planar structures will be demonstrated based on a two-dimensional poisson's equation. since the discretization schemes are important parameters in the efficiency of the algorithm, the uniform and nonuniform schemes will both be presented. finally, some of the obtained results for basic semiconductor planar structures will be depicted. 2. technique of analysis let us introduce the approach of the method of lines based on solving the very popular poisson's equation given by eq. (1) for the domain shown in fig. 1. (1) if we uniformly discretize the variable x, then the functions (x ,y) and f(x ,y) in eq. (1) are transformed into the sets (xi,y) and f(xi,y) along the lines x=i.h where i=1,2,3…n. n is the total number of lines within the structure and h is the discretizing size interval as shown in fig. 1. this operation transforms eq. (1) into a system of n differential equations of the form: (2) if we let (xi,y) = i(y) and f(xi ,y) = fi(y) then, the ith difference approximation on the discretized variable may be written as: (3.a) and its second derivative may be written as: (3.b) if we express the vectors of functions  and f as: (4.a) and (4.b) then, eq. (2) can be written in the following matrix form: (5) where p is a second order differential operator, which is an nxn tri-diagonal matrix of the form: x i  x i1    2 2 1 x i     2 2x i  x i   x i 1    2 2 1 x i i1 h h/2. . . y i1 . . . figure 1. equidistant discretization pattern using the method of lines       2 2 2 2 x, y x, y f x, y x y               2 2 2 2 i i i x , y x , y f x , y x y         1i i ix h      2 1 1 1 2 2 2i i i i i i x x hx h                  1 2 t n, ,...,     1 2 t nf f , f ,..., f 2 2 2 1d p f dy h    64 the journal of engineering research vol. 7, no. 1 (2010) 62-69 (6) where p1 and p2 are determined by the left and right boundaries of the structure of fig. 1. since p is a real-symmetric and tri-diagonal matrix, the following two possible conditions on 0 and n+1 may be met: there exists a nonsingular nxn matrix t such that the matrix  given by: (7) is diagonal and where the elements i are the eigenvalues of p. t is a matrix of the corresponding eigenvectors and tt is its transpose. for a non-uniform discretization scheme, the approach presented above has to be slightly modified. hence, the second derivative expressed by eq. (3.b) for the uniform discretization will now be expressed as given below for the non-uniform case: (8) where ei = (h1 + hi-t) / 2 denotes the ith interval size between the dotted lines shown in fig. 2 and hi is the non-uniform discretization interval. (9) where h is an assumed normalized discretization interval. considering the dirichlet-dirichlet lateral boundary conditions of the analyzed structure, which means    , eq. (9) takes the following matrix form (10) where (11) (12) and (13) whereas the matrix d of dimension (n+1xn) is a first order difference operator given by: (14) (15) which takes the matrix form: (16) where (17) and (18) 1 2 1 0 1 2 1 1 2 1 0 1 p p ... ... ... p            1. either 0 and/or 01n  where in this case 1p and/or 2p2  (dirichlet condition): 2. or 0x  and/or 0 x n    where 1p and/or 1p2  (neumann condition). tt pt  2 1 2 1 2i i ii x x , i , , .... n . ex           ei ei+1 hi-1 hi  i i 1  x i1  x i1  x i i1 x y figure 2. non-equidistant discretization pattern with the method of lines the difference eq. (3.a) for the non-uniform case is firstly normalized by multiplying both of its sides by hhh i and to get the relation:  1i i i ii h h h h x h               1 h x hhr r d    0 1h i h r diag ; i , ,... n . h          1 2 t n, , k ,     0 1 t x n , , ..., x x x                1 0 1 1 1 1 0 1 d ... ...             similarly, the second order operator given by eq. (8) is normalized by multiplying both of its sides by the factor ih e / h to get the following expression: 2 2 1 1 2i i i ii e h h ; i , , ..., n . h e x xx                       1 t e xx e xhr r d     1e i h r diag , i , 2, ...n . e         2 2 2 2 2 2 1 2 t xx n , , ..., x x x                   65 the journal of engineering research vol. 7, no. 1 (2010) 62-69 and dt is the transpose of d. inserting the expression of x given by eq. (10) into eq. (16) gives: (19) if we let (20) and (21) then eq. (19) may be written as: (22) inserting the expression of xx from (19) into the system given by eq. (2) gives the following matrix form (23) where f is a vector of functions obtained from discretizing f (x,y). (24) where (25) and (26) since dxx is a tri-diagonal matrix, then the system given by eq. (24) needs to be decoupled before being solved. furthermore, since dxx is real-symmetric, there exists an orthogonal matrix t such that the elements of the diagonal matrix  are the eigenvalues of p. t in this case, is also the matrix of the corresponding eigenvectors and tt its transpose. using the above result, the system given by eq. (24) is decoupled by pre-multiplying it by tt and obtaining the following system: (27) or simply (28) where (29) and (30) (31) the general solution of each equation of the system (31) takes the form: (32) (33) where the coefficients ai are determined by substituting vi of eq. (31) by the above expression. in the end, the vector of potentials  is obtained from v using the following expression: (34) which expresses analytically the solutions along the discretization lines as a function of y. based on the above algorithm, a computer program has been developed for the simulation of semiconductor devices to determine the potential distribution within planar structures. it includes the following steps: 1. enter the structure geometry and the dimension of each layer, the type of boundary conditions, the electric charge profile as well as the discretization interval h. 2. perform discretization by finding the total number of lines and construct the second order operator p. 3. determine the eigenvalues of the matrix p, the elements of  and the corresponding matrix of eigenvectors t. 4. find in the transformed domain the space charge density and homogeneous horizontal surface boundary vectors by pre-multiplying by tt. 5. compute the constants ai , bi by considering the field and potential interface continuity conditions as well as the surface boundary conditions. 6. perform the inverse transformation to find the analytic solution in the original domain along each discretization line.    2 1 1te xx e h h e eh r r d r r dr r    x h ed r dr t xx x xd d d   2 1 1e xx xx eh r d r     2 1 2 2 1 e xx e d r d r f dy h    multiplying both sides of eq. (23) by re 1 yields: 2 2 2 1 xx d d g dy h    1 er   1 eg r f       2 2 2 1 t t t t d t t pt t t g dy h    2 2 2 1d v v f dy h   tf t g tv t  assuming  i i  2 then eq. (28) is further transformed into a set of n ordinary differential equations as 22 2 1i i i i d v v f i ,..., n hdy         i i i i i piv a cosh y b sinh y vh h                where ai and bi are constants depending on the horizontal side boundaries. the specific solution vpi is expressed as a linear combination of the functions fi and its linearly independent derivatives  n if as      1 2 0 1 2 m pi i i i i i i i imv a f a f a f ... a f ...      e er r t v   66 the journal of engineering research vol. 7, no. 1 (2010) 62-69 3. case studies in this section, the developed algorithm is applied to a set of typical planar structures used in current electronics technology. the objective of the simulation is to determine the potential profile throughout the structures having different physical, geometrical and boundary conditions. using a uniform discretization scheme with an interval size h = 1 m results in a total of 30 intervals, the following lateral boundary conditions have been considered: a) neumann-dirichlet b) dirichlet-neumann c) dirichlet-dirichlet d) neumann-neumann the results showing the obtained potential profile contours are depicted in fig. 4. with reference to this figure, one can observe that the change in the potential profile is strongly related to the boundary conditions. this is in accordance with most practical situations where a boundary of a semiconductor h2 h1 l2 l va (applied voltage) y x (n) layer 1 1 layer 2 2 l1 (n) figure 3. cross sectional view of a two-layer planar semiconductor structure. (dimensions are given in text) 0 5 10 15 20 25 30 0 5 10 15 x (µm) y (µm) a) n-d 19 v 16 v 13 v 10 v 7 v 4 v 1 v 0 5 10 15 20 25 30 0 5 10 15 x (µm) y (µm) b) d-n 19 v 16 v 13 v 10 v 7 v 4 v 1 v 0 5 10 15 20 25 30 0 5 10 15 x (µm) y (µm) c) d-d 19 v 16 v 13 v 10 v 7 v 4 v 1 v 0 5 10 15 20 25 30 0 5 10 15 x (µm) y (µm) d) n-n 19 v 16 v 13 v 10 v 7 v 4 v 1 v figure 4. potential profiles in the structure in figure 3 with neumann-dirichlet, b) dirichlet-neumann, c) dirichlet-dirichlet and d) neumann-neumann lateral boundary conditions the effect of boundary conditions on the potential profile has been first considered by finding the potential profile throughout a two-layer planar structure as shown in fig. 3. the electrical, physical and geometrical constants of this structure are: 1 = 10 -11 c/cm3; 2 = -2 x 10 -11 c/cm3; va = 20 v; 1 = 2 = 119 0; l = 30 m; 11 = 8 m; 12 = 10 m; h1 = 10 m; h2 = 5 m and the value of 0 is 8.85x10-12 f/m. 67 the journal of engineering research vol. 7, no. 1 (2010) 62-69 device is either forced to have a specified voltage or left free. after this, two-layer symmetrical structures shown in fig. 5 have been simulated. note that in both structures the layers near to the groundsides have a constant charge density , whereas the charge profile throughout the second layers varies linearly with y as   a(y-h1) + b. the dimensions and the physical parameters for both structures are: 1 = 3.5 x 1011 c / cm3; a = -1013 c / cm4 b = -1011 c / cm3; = 11.0  ; l = 30 m mh2 = 15 m; and va = 20 v. for the structure 5-a, 11 = 12 = 10 m and the boundary conditions are dirichletdirichlet; whereas for the structure 5-b, 11 = 12 = 6 m and the boundary conditions are neumann-neumann. the analysis is carried out with 63 uniform intervals for the device 5-a and 65 intervals for the device 5-b. figure 6 shows the simulation results of the obtained potential profiles for both structures 5-a and 5-b. due to symmetry, only half of these structures need to be simulated by considering the neumann condition at the axis of symmetry. it has been verified that the analytical results are the same for all symmetric lines with respect to the axis of symmetry. hence, for symmetrical devices the numerical effort can be significantly reduced by reducing the simulated space and therefore, reducing the total number of lines by a factor of two. a two-dimensional metal oxide semiconductor (mos) capacitor has been also simulated using the developed algorithm. the structure consists of a metallic gate, a ptype silicon substrate and an oxide film layer as shown in fig. 7. x (a) h2 h1 l1 l (n) 1 2 l2 x y h2 h1 l1 l va 1 2 l2 (b) figure 5. cross sectional view of two-layer semiconductor symmetrical structures 0 5 10 15 20 25 30 0 5.0 10 15 x (µm ) y (µm) (a) 16 v 13 v 9 v 5 v 1 v 0 5 10 15 20 25 30 0 5 10 15 x (µm) y (µm) (b) 16 v 16 v 12 8 v 4 v 1 v (a) (b) figure 6. potential profiles for the structures of figures 5-a and 5-b respectively 68 the journal of engineering research vol. 7, no. 1 (2010) 62-69 because of the presence of a layer of silicon dioxide layer (which is a dielectric material) between the gate and the substrate, this device exhibits the properties of a capacitor. it is assumed in this analysis that the metal and the semiconductor work functions are equal, and that there is no electric field either within the oxide or at the oxidesilicon interface. hence, the electric field is supposed to be zero whenever the applied voltage is zero. since the two-dimensional model of the considered device is symmetric, the investigation is limited to the left part of the device as illustrated in fig. 7. when a positive voltage is applied at the electrode, free majority carriers (holes) are repelled by the induced electric field. equilibrium condition requires that the potential is zero at the self-adjusting boundaries of the resulting fully-depleted layer and elsewhere in the neutral region. the lateral boundary conditions of the silicon and oxide layers are of neumann type. the discretization pattern is formed by 10 concentrated equidistant lines crossing the curved area and 31 largely spaced lines elsewhere. the space charge layer is subdivided into 10 sub-layers to perform a stair-step approximation on the curved boundary. the dimensions used are: l = 35 m; l = 17 m; 0.1 < xo < 5.0 m; 30 < xs < 50 m; 1014 < na < 1018 cm-3; 10 < vg < 100 (vg < vbreakdown ). the program has been executed for different xo and na and the results obtained are shown in fig. 8. the results shown in this figure agree with both the previously stated boundaries and, the physical requirements. an important observation is that considering only a part of the neutral region can still save a lot of numerical efforts. this is because the potential within the whole neutral region is zero. this last point is particularly important when the left lateral side is far from the space charge region boundary, and consequently, the number of discretization lines can also be reduced. vg x y xo xs xd psio2 depletion neutral l l figure 7. two-dimensional model of the mos capacitor used in the present analysis 0 10 20 30 40 0 5 10 15 -50 0 50 100 150 x (μm) y (μm) v figure 8. three-dimensional representation of the potential distribution in the device of figure 7 69 the journal of engineering research vol. 7, no. 1 (2010) 62-69 4. conclusions the method of lines (mol) is an efficient numerical tool for solving partial differential equations. its strength lies in its assured convergence and its semi-analytical nature for both uniform and non-uniform discretization schemes. as it appears from the analysis done in this work, the mol can be applied to electronic devices to find static potential profiles for different structures. a direct application of the mol is to determine the breakdown voltages of power electronic devices needed for design prior to manufacturing. the developed algorithm allows significant saving in memory storage and computation time with assured convergence. this facilitates its extension to more sophisticated electronic structures as well as its integration into current computer-aided design tools. references ames, w.f., 1977, "numerical methods for partial differential equations," academic press, new york. barcz, a., helfert, s.f. and pregla, r., 2003, "the method of lines applied to numerical simulation of 2d and 3d bandgap structures," proceedings of 5th international conference on transparent optical networks, 1, pp. 126-129. chen, r. s., fang, d.g. and li, x.g., 2007, "analysis of open microstrip lines by mol," international journal of microwave and millimeter-wave computer-aided engineering, vol. 3(2), pp. 109 113. gonzalo, p., ricardo, m. and francisco, m., 2006, "quasitm mol/mom approach for computing the transmission-line parameters of lossy lines," ieee trans. microw. theory technology, vol. 54(1), pp. 198-209. pascher, w., pregla, r. and vietzorreck, l., 2005, "fast full-wave analysis of distributed mems transmission lines by the mol," international conference on wireless communications and applied computational electromagnetics, ieee/aces, 3-7 , pp. 763 766. plaza, g., marques, r. and medina, f., 2006, "quasi-tm mol/mom approach for computing the transmission-line parameters of lossy lines," ieee transmission microw. theory tech.nology, vol. 54(1), pp. 198-209. pregla, r. and pascher, w., 1989, " numerical technique for microwave and millimeter passive structures," ed. john willey and sons. pregla, r., 2002, "efficient and accurate modeling of planar anisotropic microwave structures by the method of lines," ieee transmission microwave theory and techniques, vol. 50, pp. 1469-1479. sadiku, m.n.o. and garcia, r.c., 2000, "method of lines solution of axisymmetric problems," southeastcon 2000. proceedings of the ieee, pp. 527 530. schultz, u., 1980, "the method of lines a new technique for the analysis of planar microwave structures," phd thesis, fern univ. hagen, germany. vietzorreck, l., coccetti, f., chtchekatourov, v. and russer, p., 2000, "numerical methods for the highfrequency analysis of mems capacitive switches," proceedings of the ii. topical meeting on silicon monolithic integrated circuits in rf systems 26 28. april 2000 garmisch germany, pp. 123-124. vietzorreck, l. and pascher, w., 2004, "efficiency enhancement by reduction of modal complexity in the analysis of cascaded planar circuits by the mol," microwave symposium digest, ieee mtt-s international, 3, pp1423-1426. yan, l., hong, w., wu, k. and cui, t.j., 2005, "investigations on the propagation characteristics of the substrate integrated waveguide based on the method of lines," microwaves, antennas and propagation, iee proceedings, vol. 152(1), pp. 35 42. zhou, g.r., fang, d.g. and feng, n.n., 2000, "entire domain based diakoptic method of lines," ceem 2000 proceedings, asia-pacific conference on environmental electromagnetics, pp. 270-273. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 91-100 *corresponding author’s e-mail: khan@squ.edu.om doi:10.53540/tjer.vol18iss2pp91-100 numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example hayder. m. khan*, saleh. al-saadi, naima. benkari, and aliya al-hashim department of civil and architectural engineering, sultan qaboos university, muscat, oman abstract: the use of cooling systems has increased because of drastic changes introduced after the oil revolution in the local architecture by using modern building designs and practices that may not suit oman's climate. the traditional architecture in this country is characterized by its passive cooling design features, which the locals fine-tuned through trial and error. these sustainable passive design practices developed over the ages have the ability to provide healthy and comfortable spaces for their occupants while preserving environmental and natural resources. these strategies are still evident in many traditional settlements and buildings in oman. the present research evaluates the effect of natural ventilation in achieving thermal comfort in traditional omani houses. for this purpose, a traditional house in a hot-arid climate zone with vernacular architecture typology was chosen as a case study. computational fluid dynamics (cfd) was used to model the house and analyse the airflow and temperature levels in its spaces. the results show the efficiency of such a design to enhance natural ventilation and thermal comfort in similar geo-climatic regions. keywords: cfd; traditional architecture; natural ventilation; thermal comfort. ي: بیت الخندق نموذجا محاكاة رقمیة إلمكانیات التھویة في المنزل العماني التقلید عالیة الھاشم قاري،نعیمة بن الساعدي،صالح خان،حیدر مراد أدت التغییرات الجذریة التي حدثت بسبب ثورة النفط إلى ازدیاد استخدام أنظمة التبرید في العمارة المحلیة عن :الملخص مناخ عمان. تزخر العمارة التقلیدیة العمانیة بمیزات طریق توظیف تصامیم وممارسات البناء الحدیثة، والتي ربما ال تناسب تصمیم التبرید السلبي التي صقلھا السكان المحلیون من خالل محاوالت التجربة والخطأ، وتتمتع ھذه الممیزات السلبیة یئة والموارد بالقدرة على توفیر فراغات صحیة ومریحة لشاغلیھا مع الحفاظ على الب -والتي طورت خالل السنین-المستدامة الطبیعیة، وال تزال ھذه االستراتیجیات ماثلة في العدید من المستوطنات والمباني التقلیدیة في عمان. یقیم البحث الحالي تأثیر التھویة الطبیعیة في تحقیق الراحة الحراریة في البیوت العمانیة التقلیدیة. ولھذا الغرض، فقد تم اختیار منزل تقلیدي في منطقة یة حارة جافة یتمتع بخصائص العمارة التقلیدیة لدراسة حالة، كما تم توظیف النمذجة الحسابیة للموائع لعمل نموذج رقمي مناخ للمنزل ولتحلیل مستویات تدفق الھواء ودرجة الحرارة في فراغاتھ، وقد أظھرت النتائج كفاءة مثل ھذا التصمیم في تعزیز ي مناطق مناخیة وجغرافیة مماثلة.التھویة الطبیعیة والراحة الحراریة ف .الراحة الحراریة ؛التھویة الطبیعیة؛ البناء التقلیدي؛ الحسابیةالموائع اداینمیك :الكلمات المفتاحیة numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example 92 nomenclature e expectancy factor ℎ𝑐𝑐 heat transfer coefficient (wm-2. k-1) fi shape factor 𝑓𝑓𝑐𝑐𝑐𝑐 garment insulation factor (clo) mrt mean radiation temperature (k) met metabolic rate (met) pmv predicted mean vote ppd predicted percentage dissatisfied 𝑃𝑃𝑣𝑣 vapor pressure of water vapor (mb) t air temperature (oc) twall wall temperature (oc) tcl cloth temperature (oc) 𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 active work (w m-2) 1. introduction civilization in oman goes back a thousand years, evident in many built settlements around the country. these settlements had desert architecture consisting of courtyard houses with thick walls and several vents, known locally as '''farkha' used to provide ventilation and thermal comfort by increasing airflow in the inner rooms. however, in the last fifty years, the oil sector has enriched the economy in oman, leading to drastic population growth, hence the increased need for buildings. these buildings were built using modern designs that are not sustainable nor responsive to the local climate, and are dependent mainly on mechanical systems to provide thermal comfort. according to an interview conducted by (gharibi 2014), most of the houses in oman do not include thermal insulation, and air conditioning units work for an average of 17.5 hours during the day. the field survey from (abdul majid et al. 2014) shows that 100% of the bedrooms in the houses are equipped with air-conditioning units set to run at a relatively lower temperature equal to 21 ℃. this relatively low-temperature thermostat setting, alongside other factors, has increased energy consumption in oman by fifteen times during the thirty years from 1980-2010. around 77% of this energy goes to the building sector, with 48-55% used in the housing sector, and 80% of this share is dedicated to running air conditioning systems (gharibi 2014; krarti and dubey 2017; alalouch et al. 2019; al-saadi and al-jabri 2020; al-saadi 2021). in order to bring these very high values to more reasonable and sustainable levels, there is a need to focus on reducing the use of air conditioning in houses to reduce energy consumption, especially with the prediction that energy consumption will increase in the coming years due to the lack of local building energy code, 'oman's population growth and the weak consumption consciousness due to subsidized energy tariff. this extravagant energy consumption affects the environment, dilapidating the non-renewable energy resources, while constituting a heavy load on the national economy, where the energy subsidies are equal to 8.9% of the omani gross domestic product (gdp) (krarti and dubey 2017). researchers are increasingly advocating for the study of vernacular architecture to generate a more sustainable built environment (almatawa, elmualim, and essah 2012). however, in oman, a few works have examined the environmental properties of vernacular architecture. the research of (al-hinai, batty, and probert 1993) is among the earliest of these studies, where four climatic regions were identified in oman, along with a unique house typology for each region. these climatic regions vary from hot and humid along the north coast to bitterly cold on the high mountains. most of the populated areas in oman are in hot and arid regions, such as in muscat and its surroundings, where the courtyard house is a popular housing typology for the hot arid climate (egan 1975). the traditional building materials and technology relied on thick/dense walls and roofs to absorb the heat during the day and release it when the air temperatures drop during the night while losing part of the heat to the sky during the night through longwave radiation. these houses use vents on both inner and outer walls and along the courtyard to facilitate the convection in rooms. furthermore, in the muscat region, there are times when sea or land breezes will be active during the day and night and in these times the air temperature will be relatively lower than the daily average. during these times, the convection inside the building will be essential in providing thermal comfort. the vents, windows, and courtyard will also enhance the ventilation rate (damlūji 1998; scudo 1988). the present paper aims to reveal the mechanisms through which the architectural typology and components generate indoor thermal comfort. thermal comfort can be defined as "the condition of mind that expresses stratifications with the thermal environment, it requires subjective evaluation" (szokolay 2012). the thermal comfort can be calculated using the predicted mean vote (pmv) method (fanger 1970). the value of pmv represents the reaction of a building's occupants to the thermal environment on a scale of seven points (-3: cold, -2: cool, -1: slightly cool, 0: neutral, 1: slightly warm, 2: warm, 3: hot). the values calculated by equations consider various physical properties like activity level, clothes, temperature, humidity, etc. the values depend on physiological variables and show how people adapt to their climate with their physical properties. an example of the adaptation is how the people of hot arid areas feel comfortable even in relatively high temperatures (28 ℃) (givoni 1998). this was confirmed by the thermal comfort survey done by nicol in baghdad (nicol 1974), where he found out that some locals can feel comfortable at a relatively high temperature equal to 32 ℃, which can be achieved by increasing the convection (air movement) in the room. the adaption effect made fanger update his method to simulate the thermal comfort in non-airconditioned buildings in a warm area by using the expectancy factor (p. ole fanger and toftum 2002). hayder m khan, saleh al-saadi, naima benkari, and aliya al-hashim 93 the present research will use the pmv method to judge the role of the openings in the vernacular house in enhancing the thermal comfort inside a traditional omani house (bait khandaq) in the al-batinah region during a day in april with a temperature daily average of 29 ℃ and daily range 12 ℃. the house was built following a traditional typology and received several additions and transformations during different times in the last two hundred years. the computational fluid dynamics (cfd) technique simulates the thermal condition and airflow in and around the house. this technique was used to study the natural ventilation in vernacular architecture in similar or close climatic conditions, such as in the south of iran for instance (heidari, sahebzadeh, and dalvand 2017). this technique to examine natural ventilation is preferred over other methods due to its flexibility (van hooff and blocken 2010). the present paper will begin by describing the house’s architectural properties and the numerical techniques used; then, it will present the results by focusing on the indoor thermal comfort, before addressing the main conclusion of the research. 2. the house the house known as bait al-khandaq is part of the traditional settlement of hujrat musalmat in wadi al ma'awil, located south of muscat. the house was initially built as a one-level building in the early nineteenth century, then it was extended and a second level was added. in recent years, this house was surveyed by researchers and students from the civil and architectural engineering department in sultan qaboos university (naima benkari 2015). they documented the house architecturally and photographically, recorded its building structure and materials, and measured its indoor thermal conditions (figure 1). many of the typical architectural element of traditional houses in the al-batinah region is shown in this house, such as the courtyard, thick walls, and air vents. the house has a bulky structure with walls made of earthen brick and plaster, with a thickness that can reach over 70 cm on some walls. this thickness was necessary to provide thermal inertia to keep the indoor cool enough until midday. the thickness usually decreases at the first level to give extra space for supporting the ceiling beams and lighten the weight to be supported by the main walls on the ground floor. furthermore, the external walls are usually painted with a light colour. the house has a flat roof to increase radiant cooling (al-hinai, batty, and probert 1993; damlūji 1998) and is surrounded by a parapet high enough to cover most of a standing person, mainly for privacy. the upper floors are constructed using a structure made of date palm trunks, cut in half or four, covered with palm leaves matting, supporting the massive layer of mud and small rocks covered with locally made lime-based plaster (sarooj) (meddah et al. 2020). the house's floor is paved with natural earthen materials as well. the house is rectangular, with a rectangular courtyard located at the centre of the house (table 1). a semi-open gallery surrounds the courtyard and constitutes a transition area distributing the circulation to the rooms. the rooms, with variable lengths and limited width, are organized around the courtyard. their width is limited to less than 3 metres due to the size of the palm tree poles. figure 1. bait al khandaq perspective (n. benkari 2017). numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example 94 table 1. building parameters. building elements value house area 510 m2 courtyard area 25 m2 windows surface area 1-1.6 m2 vent surface area 0.006 to 0.06 m2 wall structure sarooj + mud brick + sarooj (thermal conductivity 0.75 w/mk) roof structure rammed earth + grass + wood derivatives (thermal conductivity 0.3 w/mk) floor hardstone + soil earth (thermal conductivity 1.2 w/mk) room width 2.8 to 3 m circular vent diameter 20 cm number of rooms 8 number of gallery or corridor 2 (external+ internal) for the same reason, similar widths can be found in the houses around the arabian gulf (kay 1991). the rooms have a high ceiling to provide enough space for the hot air to stagnate at the top, without too much accommodating of the room's users before it is exhausted through the vents. the openings in the rooms are located at two different levels in the external walls. the lower windows and small slit vents are near the floor (30-50 cm above the floor). they are the source of airflow in the room and control the thermal sensation, as the local people usually sit close to the floor. the windows are elongated with no glass. wooden shutters are used to reduce the room's solar irradiation. circular vents are usually pierced higher in the external walls and those adjacent to the gallery. they are used to exhaust the hot air, stagnating at the top of the room, and facilitate cross ventilation. the walls overlooking the courtyard of the house may contain doors and vents only. there are small rectangular slit openings (30 x 10 cm) in the roof's parapet that provide an air jet to cool the roof surface during the daytime. the roof is commonly used for sleeping during summer nights, and these slit openings may increase the airflow, which improves thermal comfort. the daily weather in the region is characterized by sea and land breezes during the day and night, where the air temperature drops below the average. to take advantage of these breezes, the rooms in the upper level are equipped with more openings to provide higher airflow for the residents. the courtyard has several functions and properties. it enhances the natural ventilation and stores the cool air generated from longwave heat loss from the roof to the sky during the night. the cool air can flow from the courtyard to the rooms, through the vents, then leave the rooms through the openings in the external walls. however, most of the time, the house depends on sea and land breezes. the house does not include a wind catcher or basement, similar to other houses in other parts of albatinah. this may happen due to tropical storms that are common in this area. furthermore, the house is located in a village with a building fabric that is not dense, and it is exposed to the sun from four directions. 3. methodology this paper aims to examine how the openings can affect the thermal comfort in different room parts. several software and practical methods can be used to predict the thermal comfort in the room; however, only the cfd can show the spatial variations of thermal comfort resulting from the small vents that are distributed all around the room. the cfd divided the space into several cells (grid/mesh), from which the cfd solver constructs a control volume (finite volume). in every control volume, the software will solve some unsteady second-order discretized momentum, mass and energy conservation equations. these equations are closed by using turbulence models, reynolds-averaged navier-stokes equations (k-ε). this model comes in different versions. for this paper, realizable k-ε turbulence models by shih et al.(shih et al. 1995) were used. the turbulence in the boundary layer volume near walls was modelled using all y+ wall treatment models keeping the value of y+ around one in regions inside the house due to the relatively low air velocity. one of the natural ventilation features is the existence of the temperature differences between the air and the surfaces; however, they are limited. for this, the boussinesq approximation was used to calculate the changes in the air density. this method is more efficient than others because it requires less computational power. figure 2 shows the house modelled inside an atmospheric boundary layer wind tunnel with computational domain dimensions taken from the benchmark of franke et al. (franke et al. 2011), where a simulation for atmospheric boundary layer for flow around a courtyard flow is used in a number of papers like (shao, walker, and woolliscroft 1993; rajapaksha, nagai, and okumiya 2003). the boundary conditions (velocity, etc.) sets are according to the suggestions of (richards and hoxey 1993) to simulate the atmospheric boundary layer. table 2 presents the initial and boundary conditions. due to the significant role of solar radiation in thermal comfort, the solar irradiation on the building surfaces was modelled using a solar load model. this model calculates the solar hayder m khan, saleh al-saadi, naima benkari, and aliya al-hashim 95 intensity according to time, date, and geographic location. the heat transfer by radiation between surfaces can be found by applying a surface-to-surface transfer model (s2s). also, the grey thermal radiation model is used to make radiation properties independent from the wavelength. the inlet air conditions as velocity, turbulent kinetic energy, and turbulent dissipation rate are defined using a user-defined function and based on the equations from (richards and hoxey 1993). the simulation was done in wellvalidated software star ccm+ version 2021.1. the model was validated against the experimental works of other researchers for similar cases of mixed convection flow as shown in figure 3 as the comparison between experimental and cfd for vertical temperature distribution in enclosed space under mixed convection move from inlet to outlet on different levels. the grid distribution is set using a polyhedral mesh (arbitrary polyhedral). this type of mesh (i.e., a group of cells) is non-uniform unstructured, and it was selected because of its computational efficiency and faster solution convergence. a prism layer is constructed next to the surfaces to prevent cell deformation and enhance the solution by making the gridline normal to the surface with high gradients for the velocity. local refinement for the mesh was used in places in the house with a small dimension like an air vent. the grid sizes range from 3 m in the sides of the domain far from the house to 0.05 m in the air vents and windows. the grid numbers are approximately 19 million grids, with 10 million in the house and another 9 million in the outer region. it takes several days to complete one simulation using multi-core xenon dual-cpu equipped in a workstation pc. figure 2. the computational model. figure 3. comparison for the temperature distribution along the vertical axis between experimental work from (blay 1992) and cfd simulation for numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example 96 table 2. assumptions related to the computational model. the condition value initial conditions the temperature at room 36 oc inlet condition velocity inlet wind speed 5m/s, temperature 28 oc sky temperature temperature =0.0552 (tair) 1.5 outlet pressure outlet boundary conditions building wall surfaces to act as no-slip wall, domain surface slip wall wall and roof radiation properties brown scratch brick (emissivity=0.8, reflectivity= 0.2, transmissivity = 0) time and date date: 01/04/2020, time 12:00 pm geographic location lat 23.4o, long 57.8o garment insulation 0.57 expectancy factor 0.7 metabolic rate 1 met relative humidity (rh%) 20% (set as constant) time step and duration 0.1 s for 20 s the simulation outputs are air temperature values, air velocity, and mean radiant temperature of the building surfaces (mrt), which are necessary to predict thermal comfort. mrt accounts for the average radiation from the room surfaces to the centre of the room, and it can be calculated using the following equation, 𝑇𝑇𝑅𝑅 = ∑ 𝑇𝑇𝑤𝑤𝑤𝑤𝑐𝑐𝑐𝑐 𝑖𝑖𝐹𝐹𝑖𝑖 4 𝑖𝑖=1 (1) where 𝐹𝐹𝑖𝑖 is the shape factor and twall is the walls surface temperatures; the value of tr (mrt) is used to find the lt, which is the thermal load on the human body as: 𝐿𝐿𝑡𝑡 = 𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 − [3.96 ∗ 10−8 ∗ 𝑓𝑓𝑐𝑐𝑐𝑐(𝑇𝑇𝑐𝑐𝑐𝑐 + 273)4 − (𝑇𝑇𝑅𝑅 + 273)4 + 𝑓𝑓𝑐𝑐𝑐𝑐 ∗ ℎ𝑐𝑐(𝑇𝑇𝑐𝑐𝑐𝑐 − 𝑇𝑇) + 3.05 ∗ 10−3(5733 − 6.99(𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜) − 𝑃𝑃𝑣𝑣) + 0.42(𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 − 58.15) + 1.7 ∗ 10−5 ∗ 𝑀𝑀𝑀𝑀𝑀𝑀 (5867 − 𝑃𝑃𝑣𝑣) + 0.0014 ∗ 𝑀𝑀𝑀𝑀𝑀𝑀(34 − 𝑇𝑇)] (2) where (met) is the metabolic rate, (work) is the active work, 𝑓𝑓𝑐𝑐𝑐𝑐 (clo) is the garment insulation factor, and hc represents the heat transfer coefficient between the clothes and the air. the equations and information to find these variables can be found from standards like (iso 7730) (standardization 2005). tcl stands for the cloth temperature and it is calculated as: 𝑇𝑇𝑐𝑐𝑐𝑐 = 35.7 − 0.028(𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜) − 𝑙𝑙𝑐𝑐𝑐𝑐[3.96 ∗ 10−8 ∗ 𝑓𝑓𝑐𝑐𝑐𝑐[(𝑇𝑇𝑐𝑐𝑐𝑐 + 273)4 − (𝑇𝑇𝑅𝑅 + 273)4] + 𝑓𝑓𝑐𝑐𝑐𝑐 ∗ ℎ𝑐𝑐(𝑇𝑇𝑐𝑐𝑐𝑐 − 𝑇𝑇)] (3) both lt and met are used to find the pmv with the following equation: 𝑃𝑃𝑀𝑀𝑃𝑃 = (0.303 exp(−0.036𝑀𝑀𝑀𝑀𝑀𝑀) + 0.028)𝐿𝐿𝑡𝑡 (4) the value of pmv can indicate the thermal comfort according to the following index (-3: cold, -2: cool, -1: slightly cool, 0: neutral, 1: slightly warm, 2: warm, 3: hot). fanger and toftum updated the above equations to show the adaption effect for indigenous people who live in houses without an air conditioning system by using the expectancy factor (e), which works as a correction factor (p. ole fanger and toftum 2002). therefore, equation (4) became: 𝑃𝑃𝑀𝑀𝑃𝑃𝑜𝑜𝑟𝑟𝑟𝑟 = 𝑀𝑀[(0.303 exp(−0.036𝑀𝑀𝑀𝑀𝑀𝑀𝑜𝑜𝑟𝑟𝑟𝑟) + 0.028)]𝐿𝐿𝑡𝑡 (5) 𝑀𝑀𝑀𝑀𝑀𝑀𝑜𝑜𝑟𝑟𝑟𝑟 is the reduced value of met and it is used to represent the fact that humans living in hot climates tend to be less active. stavrakakis et al (stavrakakis et al. 2008) used the following set of equations to find the value of 𝑀𝑀𝑀𝑀𝑀𝑀𝑜𝑜𝑟𝑟𝑟𝑟 , 𝑀𝑀𝑀𝑀𝑀𝑀𝑜𝑜𝑟𝑟𝑟𝑟 = � −0.067 ∗ 𝑃𝑃𝑀𝑀𝑃𝑃 + 𝑀𝑀𝑀𝑀𝑀𝑀 𝑓𝑓𝑓𝑓𝑓𝑓 𝑃𝑃𝑀𝑀𝑃𝑃 ∈ (0,1) −0.067 ∗ 𝑃𝑃𝑀𝑀𝑃𝑃 + 1.004 ∗ 𝑀𝑀𝑀𝑀𝑀𝑀 𝑓𝑓𝑓𝑓𝑓𝑓 𝑃𝑃𝑀𝑀𝑃𝑃 ∈ (1,2) −0.067 ∗ 𝑃𝑃𝑀𝑀𝑃𝑃 + 1.013 ∗ 𝑀𝑀𝑀𝑀𝑀𝑀 𝑓𝑓𝑓𝑓𝑓𝑓 𝑃𝑃𝑀𝑀𝑃𝑃 ∈ (2,3) (6) the predicted percentage of dissatisfied (ppd) is widely used with pmv to show the percentage of people that may be feeling discomfort. however, ppd was not considered in this study due to the limited available space and the quantitative nature of this equation. 4. results during the initial survey for the house, it was noticed that the design of the house is different from what had been found in the north of the arabian peninsula. for example, the ventilation in traditional houses in baghdad and iran mainly depend on the windcatcher and courtyard. in these countries, the houses are closed from three directions by adjacent houses and overlook into a narrow alleyway. this is not the case in the chosen house in al-batinah region. furthermore, the courtyard is covered with a wooden structure, which suggests that it may be covered during hot weather days. accordingly, the main source for the airflow is openings, as shown in figure 4. here, the air begins hayder m khan, saleh al-saadi, naima benkari, and aliya al-hashim 97 flowing from the front windows, looping in the rooms, and goes out through the vents to the gallery and the courtyard, and these effects are extended from the first level to the ground level. another feature of this building is the design of the roof's parapet wall. in the house, many slit openings were used and were located just above the roof to increase the airflow near the upper surface of the roof. this is important because the roof is usually hotter than the rest of the building and the ambient air due to solar radiation. this extra flow of air can cool it back to a slightly higher temperature than the ambient temperature. furthermore, at night, the roof will lose heat to the sky, making it cooler and cool the air above; this makes the air heavier and can sink easily through the courtyard to the rest of the house. the parapet's edge has a zigzag shape and induces air movement (i.e., the bernoulli effect), as shown in figure 5. this effect is important at night when the residents use the roof to sleep during the summer. figure 6 shows the airflow in the room (fr1), which is the front room on the first level, where the air enters from the windows between the front gallery and the room. the air circulates then leaves the room through the door and vents toward the courtyard or through external vents and windows to the outside. figure 6 shows that the values of thermal comfort and air velocity changed locally and that the lowest values are near the windows and in the path of the air, which shows the importance of having as many windows as possible near the places of the residents’ setting. figure 4. the airflow from the room to the house's core. numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example 98 figure 5. the air velocity is distributed above the roof by using a streamline. figure 6. (left) the component of air velocity in the normal direction, (right) the value of adaptive pmv. figure 7. the predicated volume average values for pmv in various rooms of the house. figure 8. the predicated volume average values for pmv in various rooms of the house with different cases. hayder m khan, saleh al-saadi, naima benkari, and aliya al-hashim 99 the prediction for thermal comfort is shown in figure 7 for the case of keeping all the windows and doors open (allopen case), where the letter f refers to the first level and g refers to the ground level and the right side of the figure shows the distribution of the rooms. the outdoor air temperature is relatively lower than the initial indoor air in the rooms to simulate the conditions during the breezes. the air flows from rooms r2 and r1 toward the gallery and courtyard to back rooms like r7, r8, and r4. the air in its path toward the back of the house will gain heat, and its temperature will rise, making the values of pmv higher in the back rooms. so, it is better to have a limited air connection between the core of the house (gallery and courtyard) and the surrounding spaces, which can explain why there are no windows between the rooms and the courtyard. the role of external windows is shown in room gr3, where no windows or vents on exterior walls and has only a door and vents on the courtyard side. this is not enough to introduce enough air to cool the room and enhance the values of pmv. similar conditions happened in lateral rooms like r5 and r7, but with external windows; however, the air passes parallel to the outer opening without penetrating the rooms. the figure also shows the difference between the normal and adaptive values of pmv. the adaptive comfort shows that indigenous residents feel between neutral to slightly warm toward the conditions in the house without using ac units. however, it is still needed to examine the pmv under more extreme conditions like in mid-summer in future research and the passive techniques used to adapt to such conditions. the present study compared different situations like keeping all the doors and windows opened (allopen), opening only the windows (close (door)), and closing the windows and doors (close (door and windows)). figure 8 shows the lowest values for pmv are when all the openings are open (allopen). however, the rooms in the back of the house present a slight difference compared to rooms in the front. again, this demonstrates the need to have many windows on the house's external walls with fewer or no windows on the internal walls. 5. conclusion this research attempted to address the question of thermal comfort and how it is affected in spaces of traditional architecture in al-batinah region, and how can these architectural solutions be employed to achieve more sustainable designs for contemporary houses. there are few papers on the numerical simulation of traditional houses in oman despite its unique design compared to other traditional houses in the region. the study concluded that there is a need for airflow to provide indoor thermal comfort. it has been shown that this airflow is better generated when more windows exist on the external perimeters than on the interior ones. the thermal comfort can vary with the room location. additionally, the thermal conditions in the house's core are more dependent on the rooms than the outdoor air. there are many other building elements in other houses that could be studied in future research. similarly, other aspects could also be examined, such as the role of the airspeed, temperature, direction, and relative humidity. conflict of interest the authors declare no conflicts of interest funding the research described herein was funded under an internal grant programme (ig/eng/caed/20/02) from sultan qaboos university, oman. the authors gratefully acknowledge the continuous support of the university and the department of civil and architectural engineering. references abdul majid, noor, nozomi takagi, shuichi hokoi, sri ekasiwi, and tomoko uno. 2014. “field survey of air conditioner temperature settings in a hot, dry climate (oman).” hvac&r research 20. https://doi.org/10.1080/10789669.2014.953845. alalouch, chaham, saleh al-saadi, husam alwaer, and kamel al-khaled. 2019. “energy saving potential for residential buildings in hot climates: the case of oman.” sustainable cities and society 46 (april): 101442. https://doi.org/10.1016 /j.scs.2019.101442. al-hinai, h., w. j. batty, and s. d. probert. 1993. “vernacular architecture of oman: features that enhance thermal comfort achieved within buildings.” applied energy 44 (3): 233–58. https://doi.org/10.1016/0306-2619(93)90019-l. almatawa, m. s., a. a. elmualim, and e. a. essah. 2012. “passive and active hybrid approach to building design in saudi arabia.” in , 163–74. kos, greece. https://doi.org/10.2495/arc120151. al-saadi, saleh n. 2021. “pragmatic retrofitting strategies for improving thermal, energy, and economic performance of an institutional building in a cooling-dominated climate.” journal of building engineering 44: 103326. https://doi.org/10.1016/j.jobe.2021.103326. al-saadi, saleh n., and khalifa s. al-jabri. 2020. “optimization of envelope design for housing in hot climates using a genetic algorithm (ga) computational approach.” journal of building engineering 32: 101712. https://doi.org/10.1016/j.jobe.2020.101712. benkari, n. 2017. “the defensive vernacular settlements in oman, a contextual study.” international journal of heritage architecture: studies, repairs and maintence 1 (2): 175–84. numerical simulation of ventilation potential in traditional omani house: bait al khandaq as an example 100 https://doi.org/10.2495/ha-v1-n2-175-184. benkari, naima. 2015. “wadi al-ma’awil: hujrat msalmat documentation & heritage management plan.” muscat. blay, d. 1992. “confined turbulent mixed convection in the presence of horizontal buoyant wall jet.” htd vol.213, fundamentals of mixed convection. https://ci.nii.ac.jp/naid/10005732529/en. damlūji, s. s. 1998. the architecture of oman. garnet. https://books.google.com.om/books?id=t7jgqga acaaj. egan, m.d. 1975. concepts in thermal comfort. prentice hall ptr. https://books.google.com .om/books?id=gdpsaaaamaaj. fanger. 1970. thermal comfort: analysis and applications in environmental engineering. danish technical press. https://books.google.com. om/books?id=s0fsaaaamaaj. fanger, p. ole, and jørn toftum. 2002. “extension of the pmv model to non-air-conditioned buildings in warm climates.” energy and buildings 34 (6): 533–36. https://doi.org/10.1016/s0378-7788(02) 00003-8. franke, jörg, antti hellsten, heinke schlünzen, and bertrand carissimo. 2011. “the cost 732 best practice guideline for cfd simulation of flows in the urban environment: a summary.” international journal of environment and pollution 44: 419–27. https://doi.org/10.1504 /ijep.2011.038443. gharibi, hamad al. 2014. “urban growth from patchwork to sustainability. case study: muscat,” 323. givoni, b. 1998. climate considerations in building and urban design. wiley. https://books.google. com.om/books?id=mgkarz_berac. heidari, abolfazl, sadra sahebzadeh, and zahra dalvand. 2017. “natural ventilation in vernacular architecture of sistan, iran; classification and cfd study of compound rooms.” sustainability 9 (6). https://doi.org/10.3390/su9061048. hooff, t. van, and b. blocken. 2010. “coupled urban wind flow and indoor natural ventilation modelling on a high-resolution grid: a case study for the amsterdam arena stadium.” environmental modelling & software 25 (1): 51– 65. https://doi.org/10.1016/j.envsoft.2009.07.008. kay, shirley. 1991. architectural heritage of the gulf. motivate pub. krarti, moncef, and kankana dubey. 2017. “energy productivity evaluation of large scale building energy efficiency programs for oman.” sustainable cities and society 29 (february): 12– 22. https://doi.org/10.1016/j.scs.2016.11.009. meddah, m. s., n. benkari, s. n. al-saadi, and y. al maktoumi. 2020. “sarooj mortar: from a traditional building material to an engineered pozzolan -mechanical and thermal properties study.” journal of building engineering 32 (november): 101754. https://doi.org/10.1016 /j.jobe.2020.101754. nicol, fergus. 1974. “an analysis of some observations of thermal comfort in roorkee, india and baghdad, iraq.” annals of human biology 1 (november): 411–26. https://doi.org/ 10.1080/03014467400000441. rajapaksha, i., h. nagai, and m. okumiya. 2003. “a ventilated courtyard as a passive cooling strategy in the warm humid tropics.” renewable energy 28 (11): 1755–78. richards, p. j, and r. p hoxey. 1993. “appropriate boundary conditions for computational wind engineering models using the k-ϵ turbulence model.” journal of wind engineering and industrial aerodynamics, proceedings of the 1st international on computational wind engineering, 46–47 (august): 145–53. https://doi.org/10.1016 /0167-6105(93)90124-7. scudo, gianni. 1988. “climatic design in the arab courtyard house.” environmental design: journal of the islamic environmental design research centre 12. shao, l, r walker, and m woolliscroft. 1993. “natural ventilation via courtyards: the application of cfd.” in aivc. copenhagen. https://www.aivc.org /resource/natural-ventilation-courtyardsapplication-cfd. shih, tsan-hsing, william w. liou, aamir shabbir, zhigang yang, and jiang zhu. 1995. “a new k-ϵ eddy viscosity model for high reynolds number turbulent flows.” computers & fluids 24 (3): 227–38. https://doi.org/10.1016/00457930(94)00032-t. standardization, international organization for. 2005. iso 7730 2005-11-15 ergonomics of the thermal environment: analytical determination and interpretation of thermal comfort using calculation of the pmv and ppd indices and local thermal comfort criteria. iso. https://books.google.com.om/books?id=p3ycoae acaaj. stavrakakis, g.m., maria koukou, vrachopoulos michail, and n.c. markatos. 2008. “natural crossventilation in buildings: building-scale experiments, numerical simulations and thermal comfort evaluation.” energy and buildings 40 (january): 1666–81. szokolay, s. 2012. introduction to architectural science. taylor & francis. https://books.google. com.om/books?id=ruft 1. introduction 2. the house 3. methodology 4. results 5. conclusion civl23403_current.qxd 1. introduction stone mastic asphalt (sma) is a gap-graded mix with a skeletal stone-to-stone arrangement that requires a higher viscosity asphalt binder to keep the interlocked aggregates bound and intact. the interlocking nature of the mix is expected to increase the stability and to minimize the lateral displacement of the aggregates that tend to reduce __________________________________________ *corresponding author e-mail: ratnas@eng.upm.edu.my permanent deformation or rutting of the mix. the concept was first developed in germany in the early sixties. it was further developed in the united states in the early 1990s, (selim et al., 1994). whenever a conventional dense graded aggregate gra dation is altered to a gap graded matrix , the voids in mineral aggregates (vma) increases considerably. such a void content must be filled with a mastic that has excellent shear properties so as to hold the gap graded coarse aggregate matrix over a long period of time. in order to achieve this, recycled ground tire rubber of #40 mesh ring research 1 (2004) 53-58rnal of engineethe jou laboratory evaluation of ground tire rubber in stone mastic asphalt r. muniandy*1, a. a. selim2, s. hassim1, and h. omar1 1department of civil engineering, university putra, malaysia 2 department of civil & environmental engineering, south dakota state university, brookings, sd, usa received 23 april 2003; accepted 5 november 2003 abstract: stone mastic asphalt (sma) is a gap-graded mix whereby stiffer asphalt cement is required to bind the stone matrix or arrangement of stones together. although various asphalt additives are traditionally available, the use of rubber crumbs in sma is still a new rresearch endeavor. many countries around the world are facing serious problems on what to do with reject or discarded tires. in the present study, commercial truck tires, containing 70% natural rubber, were ground and pre-blended in 80-100 penetration asphalt for use in sma mixtures. an assessment was made of the laboratory performance of rubberized sma in terms of stability, resilent modulus, dynamic creep and tensile strength ratio. it was observed that the performance of sma with ground tire rubber was for superior as compared to sma mix with unmodified asphalt. sulfur and styrene butadeline rubber (sbr) were used in rubberized sma mixes as additives to test the sensitivity of sma mixtures. as standard practice a 0.3% newly developed cellulose oil palm fiber was used in sma to minimize the asphalt drain-down effects. keywords: stone mastic asphalt, rubber, tire, recycling, modified, roads :¢ü∏îà°ùÿÿg) aéµ£°üÿg …ôî°üdg â∏ø°s’gsma≈∏y iqoéb iƒbg ᫨ª°u ¢ugƒn hp ¿ƒµj é¡«a ω~îà°ùÿg â∏ø°s’g ¿g å«m èjq~àdg ámƒàøe á£∏n ƒg ( äéàa ωg~îà°sg ¿g ’g ,â∏ø°s’g ™e ioéy ω~îà°ùj äéaé°v’g øe ´gƒfg i~y ¿g øe ºzôdg ≈∏y .áe~îà°ùÿg qƒî°ü∏d áø∏àıg ωéém’g úh ∂°sé“ øjƒµj ) ∫g äé£∏n ‘ •é£ÿgsmaøe ¢ü∏îàdg á«ø«c »gh ’g iò£n á∏µ°ûe ¬lgƒj ⁄é©dg ∫ƒm ¿g~∏ñdg øe ~j~©dg .¤h’g á«ãëñdg ¬∏mgôe ‘ ∫gõj ’ ( hp â∏ø°sg ¤g é¡àaé°vgh é¡≤ë°s ~©h áäÿéh 70 áñ°ùæh »©«ñ£dg •é£ÿg øe ¿ƒµÿg äéæmé°ûdg ägqéwg äéàa ω~îà°sg áãj~m á°sgqo ‘ .áødéàdg ägqéw’g ) á£∏n êéàf ’ 100-80 ¥gîng πeé©esma,áfg~∏dg qg~≤eh ,á«jéñãdg ¢sé«≤h πª©ÿg ‘ äéæ«©dg agog qéñàng ≥jôw øy á£∏ÿg √òg º««≤j ” ~bh .( ) ∫g á£∏n agog ¿g áhôéàdg ∫ón øe ~lh ~bh .~°ûdg iƒb áñ°ùfh ,»µ«eéæj~dg ∞mõdghsma~«©h ~m ¤g ¥éa ~b ägqéw’g ¥ƒë°ùe é¡«dg ±é° ÿg ( ) øjoéjƒ«h øjôjéà°ùdgh âjèµdg •é£e ω~îà°sg éªc .áaéx’g ∂∏j ¿h~h égagog sbr) ∫g äé£∏n ‘ (smaá«°sé°ùm ¢sé«≤d ±é° e πeé©c á«wé£ÿg ( ) ∫g äé£∏n ‘ áäÿéh 0^3 .áñ°ùæh π«îædg ±é«dg rƒdƒ«∏°s âjr ωg~îà°sg ” »°sé«b agôléch .äé£∏ÿg √ògsma»∏ø°ùdg ±gõæà°s’g ägòkéj øe ~ë∏d ( .â∏ø°sód :á«méàøÿg ägoôøÿg.¥ô£dg ,(qƒ£e) ∫~©e ,ôjh~àdg ioéyg ,qéwcg ,•é£e ,»¨ª°üdg ôé◊g øe â∏ø°scg from shredded tires was used in the pre-blending of unmodified 80-100 asphalt, (lundy et al., 1987) (sma in general in prone to asphalt drain-down during mix storage and transportation to the construction site). even after the placement of the mix, the softening of the mastic during hot days tends to slowly drain down the asphalt. this slow drain down effect of the binder is further accelerated by the contact pressure from wheel loading which may result in the ultimate loss of bitumen from the top part of the asphalt layer. to minimize this problem cellulose fibers were used. 2. material characterization sma is a high strength mix that requires good quality aggregates. inferior quality aggregates may be crushed upon repeated loading that may in turn alter the stone matrix posture entirely. therefore, aggregate quality must be controlled to ensure the superior performance of sma mixtures, (bukowski, 1991). since granite is abundantly available in malaysia, it was identified as a prime candidate for use in sma mix. in our research, the public works department specifications (pwd), (public works department, 1985), for aggregate properties were adopted in for mulating stone mastic asphalt, table 1. a typical gradation was formulated that would give higher stability and reliability with a maximum size of 14 mm. approximately, 80% of the aggregates were larger than 2 mm, and more than 70 percent larger than 8mm, figure 1. traditional 80/100 penetration asphalt is considered to be too soft for use in sma. it has to be modified to increase the viscosity. polymers have been traditionally used to modify asphalts for specific applications. *uma = unmodified asphalt tra2 = tire rubber asphalt with 2% rubbe content by weight of asphalt tra3 = tire rubber asphalt with 3% rubber content by weight asphalt tra4 = tire rubber asphalt with 4% rubber content by weight asphalt tra4-2.2s = tire rubber asphalt with 4% rubber content and 2.2% butanol by weight of asphalt tra4-5b = tire rubber asphalt with 4% rubber content and 5% sulfur by weight of asphalt 0 10 20 30 40 50 60 70 80 90 100 0.0000 0.1000 0.2000 0.3000 0.4000 0.5000 0.6000 0.7000 0.8000 0.9000 1.0000 sieve number p er ce n t p as si ng lb db ub md actb linear (md) # 20 0 # 10 0 # 80 # 50 # 40 # 30 # 20 # 16 # 10 # 8 # 4 1/ 4" 3/ 8" 1/ 2" 3/ 4" 1" aggregate more costly to produce. surface easy to finish critical mixture readily unstable with slight excess of asphalt or water porous mixture lack tensile strength when mixed with fuel oil best result with heavy bitumen harsh mixture, inclined to segregate not critical porous mixture lack tensile strength when mixed with fuel oil best result with heavy bitumen figure 1. stone mastic asphalt (sma) and conventional gradations no. type of test result % pwd requirement 1 l.a abrasion 19.70 ≤ 30 % 2 crushing value 26.20 ≤ 30 % 3 impact value 12.6 ≤ 15 % 4 soundness test (sodium sulfate) 1.76 ≤ 15% 5 polishing stone value 50.9 ≥ 49 6 flakiness and elongation test ( 3:1 ratio) 17.1 ≤ 20% 7 specific gravity 2.62 ≥ 2.60 8 water absorption 0.469 ≤ 2% table 1. granite aggregate properties used in sma mix asphalt blend* penetration (mm ) softening point (°c) thin film oven uma 85.7 52.3 0.003 tra2 79.3 63.3 0.009 tra3 71.5 68.2 0.010 tra4 62.0 70.3 0.011 tra4-2.2s 61.7 70.3 0.009 tra4-5b 60.3 71.0 0.007 table 2. physical properties of rubberized asphalt 54 ring research 1 (2004) 53-58rnal of engineethe joumuniandy et al. / however, the use of polymers increases the overall cost of asphalt mix production. recent experiences in polymer modified asphalt (pma), (bukowski, 1991) showed that it costs at least three times more compared to conventional asphalt. tire rubber from recycled shredded tires were pulverized to #40, which is about 450 micron size. the tire rubber powder was checked for deleterious material like particles from the steel belting that could have been *mixed with the tire powder during the shredding, grinding, and separation process. this was done to ensure the purity of the selected material. the unmodified asphalt was blended with 2%, 3% and 4% ground tire rubber. the physical properties of the tire rubber modified asphalt are shown in table 2. sulfur and styrene butadiene random (sbr-butanol) were also used in our study to check the sensitivity of sma mixes. 3. marshall mix design marshal mix design was carried out in accordance with the american standard for testing and materials (astm) d1559. fifteen 100 mm diameter marshall specimens were prepared for each tire rubber blended asphalt. each speciment required approximately 1200 grams of granite aggregates. a total of 60 batches were used in the preparation of sma samples. four sets of asphalt blended with 2%, 3%, and 4% tire rubber and one unmodified set were prepared for the mix design. the aggregate samples were heated in an oven to 170°c while the asphalt blends were heated to 165°c based on the viscosity-temperature of the asphalt. for each asphalt type, five sets of three specimens, each with an asphalt content of 4%, 5%, 6%, 7% and 8%, were prepared. a 0.3% cellulose fiber by weight of aggregates were used in all sma samples. the compaction of the sma mixtures was done using an automatic compactor with a standard 50 blow on each side. it is a common practice in malaysia to use 50 blow marshall compaction for hot mix asphalt mixtures. 4. optimum asphalt content the optimum asphalt content for each rubber blend was determined at maximum bulk density, maximum stability, and at 4% air voids using the asphalt institute method, table 3. the average of four optimums was considered in the preparation of specimens for performance tests. additional specimens were made with the optimums of each asphalt blend. marshall properties tests were carried out in accordance with astm d1559. the performance of the rubberized and control samples are shown in figures 2 to 7 and in table 3. the resilient modulus test was carried out in accordance with astm d4123, figure 4. samples with more than 4 percent tire rubber blended asphalt were found to have non-desirable properties. contents higher than 4% tire rubber were tried. however, it was found to be very difficult to blend the pulverized tire beyond 4% due to clumping and dispersion. hence the study focused on a maximum tire rubber content of 4% (by weight of asphalt) only. marshall tests with various proportions of tire blend showed that sma with 4% ground tire rubber (tra4) displayed better properties compared to other sma mixtures. however, sma with a tire rubber blend of more than 4% did not display good marshall properties. the workability was greatly affected. hence, the results of sma with greater than 4% tire rubber are not dismix mr (mpa) stability (kn) flow bulk density vma (%) vtm (%) vfa (%) optimum ac uma 3436.3 12.65 3.43 2.28 18.16 4.95 72.72 5.86 tra2 3612.7 14.31 4.50 2.311 17.11 3.62 78.86 5.9 tra3 3764.7 15.01 4.76 2.313 16.96 3.44 79.70 6.0 tra4 3942.0 17.30 4.66 2.321 16.76 5.60 66.58 6.1 tra4-2.2s 4275 17.88 4.60 2.35 17.82 4.95 72.31 tra4-5b 4885 18.31 4.69 2.35 17.79 5.04 69.87 table 3. results of sma mix performance load cycle sample 1% strain 3% strain 1 2 3 ave 1 2 3 ave uma 52 92 61 68.3 1711 1755 2100 1855.3 tra4 79 71 90 80.0 2782 2712 2854 2782.7 tra4-2.2s 180 105 158 147.7 3244 3004 3328 3192.0 tra4-5b 125 100 150 125.0 8360 4641 4885 5962.0 table 4. permanent deformation of rubberized sma 55 ring research 1 (2004) 53-58rnal of engineethe joumuniandy et al. / 8 10 12 14 16 18 3 4 5 6 7 8 9 percent asphalt st ab ili ty (k n) uma tra2 tra3 tra4 1 2 3 4 5 6 7 4 5 6 7 8 % asphalt tire rubber blend fl ow (m m ) uma tra2 tra3 tra4 2.230 2.250 2.270 2.290 2.310 2.330 4 5 6 7 8 % asphalt tire rubber blend b ul k d en si ty uma tra2 tra3 tra4 15 16 17 18 19 20 4 5 6 7 8 % asphalt tire rubber blend % v m a uma tra2 tra3 tra4 1800 2300 2800 3300 3800 4300 4800 3 4 5 6 7 8 9 percent asphalt r es il ei n t m o d u lu s (m p a) uma tra2 tra3 tra4 0 1 2 3 4 5 6 7 8 9 10 4 5 6 7 8 % asphalt tire rubber blend vt m (% ) uma tra2 tra3 tra4 figure 2. marshall stability vs asphalt blend figure 5. flow vs asphalt blend content figure 6. vma vs asphalt blend figure 7. vtm vs asphalt blend content figure 4. resilient modules vs % asphalt blend content figure 3. bulk density vs asphalt blend content 56 ring research 1 (2004) 53-58rnal of engineethe joumuniandy et al. / cussed in this paper. a sensitivity analysis was carried out on sma-4 using sulfur and a styrene butadiene rubber product (butanol). a total of fifteen samples were prepared and tested for marshall properties, table 3. the optimum sulfur and butanol contents were determined by preparing and testing 15 sma specimens with sulfur proportions of 1,2,3,4, and 5 % and 15 specimens with 2,4,6, and 8 % butanol by weight of asphalt tire rubber blend. figure 8 shows the optimum sulfur and butanol content determined from the resilient modulus test. 5. dynamic creep performance sma specimens were tested for permanent deformation in the uniaxial repeated compressive load. sets of three specimens were made for each asphalt tire rubber blend at their individual optimum asphalt content and trimmed to 50mm with a 2:1 ratio of the diameter. the specimens were then tested in accordance with astm d3497 to determine the number of load cycles to reach 1% strain , 3% strain and ultimate failure, table 4 and figure 9. an australian ipc matta machine was used in the testing of the sma specimens. the selected test parameters were as follows: test temperature : 50°c pulse width : 500ms pulse period : 2000 test load stress : 500kpa 6. moisture induced damage analysis roads in tropical countries like malaysia are exposed to extreme moisture conditions. as such, it was deemed appropriate to carry out a moisture induced damage test on sma samples. the samples were prepared and tested using the modified lottman test in accordance with aashto t283. two sets of three specimens for each test were prepared. one set of samples was used as a control while the other set was vacuum saturated at 28 in hg with water for 25 minutes. both sets (control and conditioned) were then tested for indirect tensile strength at 25°c using a loading rate of 51 mm/minute. the tensile strength ratio (tsr) of the samples was determined as a ratio of conditioned strength over control strength, figure 10. a minimum of 0.7 is generally specified. the higher the tsr , the higher the durability of the mix. 7. discussion the marshall mix properties of sma with tire rubber blends of 2%, 3%, and 4% generally increased with a gradual increment in optimum asphalt content. the optimum asphalt contents ranged between 5.86 to 6.10%. 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 0 1 2 3 4 5 6 7 8 9 10 11 % sulfur/butanol in 4% tire rubber blend r es ili en t m od ul us , m p a sulfur butanol figure 8. resilient modulus vs sulfur and butanol content 0 5000 10000 15000 20000 1% 3% rupture % strain n um be r o f l oa d cy cl es sma0 sma4 sma4+2.2 sma+5 figure 9. strain vs number of load cycles 84 86 88 90 92 94 96 98 100 0% 2% 3% 4% 4+ 2.2 % 4+ 5% type of asphalt blend ts r (% ) figure 10. tensile strength ratio of sma mix 75 ring research 1 (2004) 53-58rnal of engineethe joumuniandy et al. / sma with 4% tire rubber showed higher values in terms of stability and resilient modulus. a sensitivity analysis on sma-4% tire rubber with 2.2% sulfur and 5 % sbr additive greatly improved the sma mixture’s permanent deformation or rut potential. it was found that sma with 4% blend took 80 load cycles to reach 1% strain and 2782 load cycles to reach 3% strain while the control sma mix displayed a much lower value of 68.3 cycles for 1% and 1855 cycles for 3% strain levels. the addition of a small amount of sulfur and butonal, separately, resulted in a marked increase in the dynamic creep performance of the mix. the addition of sulfur 2.2% by weight of rubberasphalt blend increased the load cycles to 148 for 1% strain and to 3192 for 3% strain. however, the addition of 5 % butanol by weight of rubber-asphalt blend had the highest value. it took 5962 cycles to reach the 3 % strain level and almost 20,000 load cycles to failure. the control mix on the average failed at 5181 cycles. the indirect tensile test also showed that all sma mixes displayed a tensile strength ratio of more than 70%. sma with 4% tire rubber without any additives showed a remarkable 98% tsr. however, sma with 4% tire rubber and additives did not improve the moisture damage of sma. instead lower tsr values of 95% and 96%, respectively, were displayed. 8. conclusions in conclusion, the rubberized stone mastic asphalt mix has shown great promise. this mix may be able to supersede the conventional mix. since sma requires a stiff binder the traditional 80/100 penetration asphalt can be blended with recycled ground tire rubber. this will give good binding properties for sma performance. a cost effective sma with reject or discarded tire rubber will give excellent pavement performance not only in tropical regions with high rainfall, but also in regions with high temperatures. the laboratory study has shown that there is a great potential for ground tire rubber for use in sma, up to 4% with 0.3% cellulose oil palm fiber. however, a field trial needs to be carried out to evaluate its full potential. references bukowski, j., 1991, “sma comes to usa,” journal of the national asphalt pavement association vol.6, no.2, p.5. lundy, j.r., hicks, r.g., richardson, and e., 1987, “evaluation of rubber modified asphalt performance mt. st. helens project,” asphalt pavemen technology, proceedings associations of technologists, vol.56, pp.573-598. public works department, 1985, “manual of quartzite based stone matrix asphalt mixtures (smam)”, proceedings of the third materials engineering conference 804, asce, p.635-642. 85 ring research 1 (2004) 53-58rnal of engineethe joumuniandy et al. / mech090414.qxd the journal of engineering research vol. 7, no. 1, (2010) 42-52 1. introduction submerged arc welding (saw) is a multi-factor, multiobjective metal joining technology in which several __________________________________________ *corresponding author’s e-mail: sdatta@nitrkl.ac.in process control parameters interact in a complicated manner and influence differently on quality of the prepared multi-objective optimization of submerged arc welding process saurav datta* and siba sankar mahapatra department of mechanical engineering, national institute of technology, rourkela, orissa-769008, india received 14 april 2009; accepted 5 september 2009 abstract: submerged arc welding (saw) is an important metal fabrication technology specially applied to join metals of large thickness in a single pass. in order to obtain an efficient joint, several process parameters of saw need to be studied and precisely selected to improve weld quality. many methodologies were proposed in the past research to address this issue. however, a good number of past work seeks to optimize saw process parameters with a single response only. in practical situations, not only is the influence of process parameters and their interactive effects on output responses are to be critically examined but also an attempt is to be made to optimize more than one response, simultaneously. to this end, the present study considers four process control parameters viz. voltage (ocv), wire feed rate, traverse speed and electrode stick-out. the selected weld quality characteristics related to features of bead geometry are depth of penetration, reinforcement and bead width. in the present reporting, an integrated approach capable of solving the simultaneous optimization of multi-quality responses in saw was suggested. in the proposed approach, the responses were transformed into their individual desirability values by selecting appropriate desirability function. assuming equal importance for all responses, these individual desirability values were aggregated to calculate the overall desirability values. quadratic response surface methodology (rsm) was applied to establish a mathematical model representing overall desirability as a function involving linear, quadratic and interaction effect of process control parameters. this model was optimized finally within the experimental domain using pso (particle swarm optimization) algorithm. a confirmatory test showed a satisfactory result. a detailed methodology of rsm, desirability function (df) and a pso-based optimization approach was illustrated in the paper. keywords: submerged arc welding (saw), multi-objective optimization, response surface methodology (rsm), desirability function (df), particle swarm optimization (pso) ¢ùwé¨dg ωéë∏dg á≤jô£d á«yƒ°vƒÿg √o~©àÿg á«∏yéødg gôjéñgée qéμfé°s éñ«°s , éjgo ∞jqgƒ°s ::áá°°uuóóÿÿggωéë∏d πegƒy i~y ,aƒøc π°üøe ≈∏y ∫ƒ°ü◊g πlg øe .i~mgh iôe ‘ iòñc ∑骰sg ägp ¿oé©e §hôd â∏ª©à°sg gpeg á°uénh ᪡e ™«æ°üj é«lƒ∏æμj ƒg ¢ùwé¨dg ωéë∏dg âféc á≤hé°ùdg ∫éªy’g øe º¡e o~y ∂dp ™eh .´ƒ°vƒÿg gòg ™e πeé©à∏d á≤hé°s çƒëh ‘ é¡mgîbeg ” ~b ¿oé©e º¶f øe ~j~©dg .á«yƒædg ú°ùëàd áb~h ¢sq~j ¿g öéj ¢ùwé¨dg ∑éæg øμdh áb~h é¡°üëa ” ~b äélôıg ≈∏y á∏yéøàÿg é¡jgòkéj h á«∏ª©dg ägoôøe òkéj §≤a ¢ù«d ,á«∏ª©dg ´é°vh’g ‘ .i~mgh áhééà°séh ájgoôøe h ¢ùwé¨dg ωéë∏dg π«©øàd πecéj áyô°s ,á«àdƒødg :»gh iô£«°ùdg ägoôøÿ äé«∏ªy á©hqg qéñày’g ú©h ònéj á«dé◊g á°sgq~dg ¿éa ∂dp πlg øeh .áæegõàe iqƒ°üh √~mgh áhééà°sg øe ìcéh á«∏yéa ägp é¡∏©÷ ádhéfi üƒ∏°sg ìgîbg ” ‹é◊g ôjô≤àdg ‘ .rôÿg ¢vôyh í«∏°ùàdg h ájpéøædg ≥ªy »g rôÿg á°s~æg éjgõã §ñjôj qéàıg ∂∏°ùdg á«yƒf .ohîc’g rhôh ,á«ñfé÷g áyô°ùdg ,∂∏°ùdg ájò¨j áaéμd ájhé°ùàe ᫪gg ¢vgîaéh .áñ°séæÿg áhƒzôÿg ádg~dg qé«àneg ᣰsgƒh ájoôødg á≤jô£dg ‘ .¢ùwé¨dg ωéë∏d äé«yƒædg √o~©àÿg áhééà°s’g ™e áæegõàe á«∏yéa hp πc πhéb èe~e áhééà°s’g »yéhôdg í£°ùdg á≤jôw .á∏eé°ûdg áhƒzôÿg º«≤dg üé°ù◊ ⩪l ~b áhƒzôÿg ájoôødg º«≤dg √òg ¿éa ,äéhééà°s’g rsmπãá »°véjq ωé¶f ¢ù«°séàd é¡≤«ñ£j ” ~b äé«erqgƒn ∫骩à°séh »ñjôéàdg π≤◊g ∫ón gòng á∏«©øj ” ~b êpƒªædg gòg .ägoôøÿg h á«∏ª©∏d »∏yéøj »yéhq »£n òkéj ≈∏y πª°ûj ¬dg~c á∏eé°ûdg áñzôdgpsoäé°uƒëa . ¤g á∏°üøe á≤jôw .á«°vôe èféàf âæ«h áj~«ccéjrsm áñzôdg ádg~dgdf á«∏yéødg üƒ∏°sg ≈∏y i~ªà©ÿg á«∏yéødg ∂dòchpso.åëñdg ‘ áfé«ñj ” ~b ««mmééààøøÿÿgg ääggooôôøøÿÿgg ¢ùwéz ωé◊ : áwsa ö«éà°ùÿg í£°ùdg á≤jôw ,±g~g’g √o~©àe á«∏yéa ,rsm ¬ñzôdg ádgo ,df äéäjõ÷g ~°ûm ¬«∏yéa ,pso . 43 the journal of engineering research vol. 7, no. 1, (2010) 42-52 weld. weld quality depends on various features of bead geometry, mechanical-metallurgical characteristics of the weld as well as on weld chemistry. moreover, the cumulative effect of combined aforesaid quality features determines the extent of joint strength that determines functional aspects if the weld is subjected to practical field of application. therefore, preparation of a satisfactory good quality weld seems to be a challenging job. complete knowledge regarding the mode of influence of the process control parameters and their interactions are to be exactly known prior to select an optimal process environment capable of producing desired quality weld. however, saw optimization is a difficult task due to simultaneous fulfillment of multi-quality features which should be close to the desired target value at the optimal setting. in practice, it may happen that an improvement of one response may cause severe loss to another quality feature for a particular parametric combination. tay and butler (1996) proposed an application of an integrated method using experimental designs and neural network technologies for modeling and optimizing a metal inert gas (mig) welding process. correia et al. (2004) used genetic algorithm (ga) to decide near-optimal settings of a gmaw welding process. dongcheol et al. (2002) suggested the use of genetic algorithm and response surface methodology (rsm) for determining optimal welding conditions. hsien-yu tseng (2006) proposed an integrated approach to address the welding economic design problem. the integrated approach applied general regression neural network (nn) to approximate the relationship between welding parameters (welding current, electrode force, welding time, and sheet thickness) and the failure load. an analytical formula was generated from the trained general regression neural network, and the mathematical model for the economic welding design was constructed. ga was then applied to resolve the mathematical model and to select the optimum welding parameters. these parameters were recommended for use to obtain the preferred welding quality at the least possible cost. zhao et al. (2006) focused on the performance -predicting problems in the spot welding of the body-galvanized steel sheets. artificial neural networks (anns) were used to describe the mapping relationship between welding parameters and welding quality. after analyzing the limitation that existed in standard back propagation (bp) networks, the original model was optimized based on a lots of experiments. a lot of experimental data about welding parameters and corresponding spot-weld quality were provided to the ann for study. the results showed that the improved bp model can predict the influence of welding currents on nugget diameters, weld indentation and the shear loads ratio of spot welds. the forecasting precision was quite high satisfying the practical application value. pasandideh and niaki (2006) presented a new methodology for solving multi-response statistical optimization problems. this methodology integrates desirability function and simulation approach with a genetic algorithm. the desirability function was used for modeling the multi-response statistical problem whereas the simulation approach generated required input data and finally the genetic algorithm was implemented to optimize the model. praga-alejo et al. (2008) highlighted that the neural network (nn) with ga as a complement are good optimization tools. the authors compared its performance with the rsm that is generally used in the optimization of the process, particularly in welding. many designed experiments require the simultaneous optimization of multiple responses. the common trend to tackle such an optimization problem is to develop mathematical models of the responses. these indicate the entire process behavior. the effect of process parameters on different responses can be analyzed from the developed models. multiple linear regression and response surface methodology are two common tools available for developing the mathematical models of the responses as a function of process parameters. depending on the requirement, each quality features/responses are optimized (maximized or minimized) to determine the optimal setting of the parameters. however, this method is applicable for the optimization of a single objective function. in a multiobjective case, it is essential to convert these multiple objectives to an equivalent single objective function which has to be optimized finally. a common approach is to use a desirability function combined with an optimization algorithm to find the most desirable settings. in the desirability function approach, individual response desirability values are calculated depending on the target as well as prescribed tolerance limit of the response variables. individual desirability values are then aggregated to calculate the overall desirability value. the optimal setting is one which can maximize the overall desirability. in doing so, a mathematical model is required for overall desirability. the model is then optimized finally. however, as the number of factors that affect the complexity of a multiple response problem increases, conventional optimization algorithms can fail to find the global optimum. for these situations, a common approach is to implement a heuristic search procedure like the ga and ann or other optimization algorithms like controlled random search (crs) price, w. l. (1977). however, it has been found that ga was adapted many times by previous researchers; less effort was made on application of crs and even pso in optimizing features of submerged arc weld. in consideration of the above, the present study aims at evaluating a near optimal parameter setting for the optimization of bead geometry parameters of a submerged arc weld. the study proposes integrating rsm-based desirability function approach and a pso algorithm for multi-response optimization of saw. bead geometry parameters of submerged arc weld on mild steel were selected as multi-objective responses and they were optimized to select the optimal process environment. finally, the study concludes the effectiveness and application feasibility of the proposed integrated approach. 44 the journal of engineering research vol. 7, no. 1, (2010) 42-52 2. desirability function (df) approach individual desirability values related to each of the quality parameters are calculated using the formula proposed by derringer and suich in1980. there are three types of desirability function: lowerthe-better (lb), higher-the-better (hb) and nominalthe-best (nb). in the present investigation, for reinforcement and bead width lb criteria; and for penetration depth hb criteria have been selected. this is because, the objective of the work was to minimize reinforcement and bead width (to reduce weld metal consumption) and to maximize penetration depth as strength of the welded joint directly depends on penetration depth. the nb criterion is generally selected in cases where responses have their fixed target value. (1) (2) (3) (4) (5) (6) the individual response desirability values were accumulated to calculate the overall desirability, using the following eq. (7). here, d is the overall desirability value, di is the individual desirability value of ith quality characteristic and n is the total number of responses. wi, is the individual response weightage. (7) 3. response surface methodology (rsm) the response function that represents any of the output features of the weldment can be expressed as an individual desirability value using the lower-thebetter (lb) criterion is shown in fig. 1. the value of ŷ is expected to be the lower the better. when ŷ is less than a particular criteria value, a desirability value id equal to 1; if ŷ exceeds a particular criteria value, the desirability value equals to 0. id varies within the range 0 to 1. the desirability function of the lower-thebetter (lb) criterion can be written as below (eqs. 1 to 3). here, miny denotes the lower tolerance limit of ŷ , the maxy represents the upper tolerance limit of ŷ and r represents the desirability function index, which is to be assigned previously according to the consideration of the optimization solver. if the corresponding response is expected to be closer to the target, the index can be set to the larger value, otherwise a smaller value. figure 1. desirability function (lb) if ,ˆ minyy  1id if ,ˆ maxmin yyy  r i yy yy d          maxmin maxˆ if maxˆ yy  , 0id an individual desirability value using the higherthe-better (hb) criterion is shown in fig. 2. the value of ŷ is expected to be the higher the better. when ŷ is exceeds a particular criterion value, according to the requirement, the desirability value id is equals to 1; if ŷ is less than a particular criteria value, ie. less than the acceptable limit, the desirability value is equals to 0. the desirability function of the higher-the-better (hb) criterion can be written as below (eqs. 4 to 5). here, miny denotes the lower tolerance limit of ŷ , the maxy represents the upper tolerance limit of ŷ and r represents the desirability function index, which must have been previously according to the consideration of the optimization solver. if the corresponding response is expected to be closer to the target, the index can be set to the larger value, otherwise to a smaller value. figure 2. desirability function (hb) if ,ˆ minyy  0id if ,ˆ maxmin yyy  r i yy yy d          minmax minˆ if maxˆ yy  , 1id 1 2 1 1 2( ............ )n www w nd d d d  45 the journal of engineering research vol. 7, no. 1, (2010) 42-52 (8) here, y is the response. v = voltage (ocv), wf = wire feed rate, tr = traverse speed and n = electrode stickout. the selected relationship is a second-degree response surface, which is expressed as follows: (9) the response surface methodology (rsm) is an efficient tool, which is widely applied for modeling the output response(s) of a process in terms of the important controllable variables and then finding the operating conditions that optimize the response. the eq. (9) can be written as a multiple linear regression model as follows: (10) the method of least squares can be used to estimate the regression coefficients in eq. (10). in this study regression coefficients were computed by statistical software package minitab (release 14). 4. particle swarm optimization (pso) algorithm particle swarm optimization (pso) is a populationbased stochastic optimization technique developed by dr. eberhart and dr. kennedy in 1995, inspired by the social behavior of bird-flocking or fish-schooling. pso shares many similarities with evolutionary computation techniques such as genetic algorithms (ga). the system is initialized with a population of random solutions and searches for optima by updating generations. however, unlike ga, pso has no evolution operators such as crossover and mutation. in pso, the potential solutions, called particles, fly through the problem space by following the current optimum particles. each particle keeps track of its coordinates in the problem space which is associated with the best solution (fitness) it has achieved so far. (the fitness value is also stored.) this value is called pbest. another "best" value that is tracked by the particle swarm optimizer is the best value, obtained so far by any particle in the neighbors of the particle. this location is called lbest. when a particle takes all the population as its topological neighbors, the best value is a global best and is called gbest. the particle swarm optimization concept consists of, at each time step, changing the velocity of (accelerating) each particle toward its pbest and lbest locations (local version of pso). acceleration is weighted by a random term, with separate random numbers being generated for acceleration toward pbest and lbest locations. in the past several years, pso has been successfully applied in many research and application areas. it has been demonstrated that pso gets better results in a faster, cheaper way than other methods. another reason that pso is attractive is that there are few parameters to adjust. one version, with slight variations, works well in a wide variety of applications. particle swarm optimization was used for approaches that can be used across a wide range of applications, as well as for specific applications focused on a specific requirement. 5. experimentation bead-on-plate submerged arc welding (on mild steel plates of thickness 10 mm) was carried out following 34 full factorial design which consists of 81 combinations of voltage (ocv), wire-feed rate, traverse speed and electrode stick-out. each process control parameters was varied in three different levels during experiments. interaction effects of process parameters were assumed negligible in the present study. three responses related to features of bead geometry viz. bead width, reinforcement and depth of penetration were selected in the present study. a copper coated electrode wire of diameter 3.16 mm (aws a/s 5.17:eh14) was used during the experiments. welding was performed with flux (aws a5.17/sfa 5.17) with grain size 0.2 to 1.6 mm with basicity index 1.6 (al2o3+mno2 35%, cao+mgo 25% and sio2+tio2 20% and caf2 15%). the experiments were performed on a submerged arc welding machineindarc autoweld major (maker: iol ltd., india). while the weld was being made, the specimens were prepared for metallographic test. features of bead geometry (macrostructure) were observed in optical trinocular metallurgical microscope (make: leica, ger( , , , )y f v wf tr n 2 0 1 2 3 4 11 2 2 2 22 33 44 12 13 14 23 24 34 . . . . . . y b v b wf b tr b n b v b wf b tr b n b v wf b v tr b v n b wf tr b wf n b tr n                0 , 4 1i i b  , 4 1ii i b  and all ijb (interaction term coefficients) are constant. the dimensions (units) of the constants should be such that they must take care of the dimensional similarity on both sides of eq. (9). 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 y x x x x x x x x x x x x x x                               here, 14 1 ( )i iy f x  and 1 1 2 2 3 3 4 4 5 11 6 22 7 33 8 44 9 12 10 13 11 14 12 23 13 24 14 34 , , , , , , , , , , , , , b b b b b b b b b b b b b b                             1 2 3 4 5 2 2 2 6 7 2 8 9 10 11 12 13 14 , , , , , , , , . , . , . , . , . , . x v x wf x tr x n x v x wf x tr x n x v wf x v tr x v n x wf tr x wf n x tr n               46 the journal of engineering research vol. 7, no. 1, (2010) 42-52 parameters units notation level -1 level 0 level +1 voltage (ocv) volts v 27 28 29 wire feed rate cm/min wf 655 970 1285 traverse speed cm/min tr 72 98 124 stick-out mm n 27 29 31 table 1. process control parameters and their limits design of experiment (factorial combination) response data (bead geometry) sl. no. v wf tr n p r w 1 -1 -1 -1 -1 3.849 1.761 10.061 2 -1 -1 -1 0 3.748 1.725 10.520 3 -1 -1 -1 1 3.627 1.709 11.219 4 -1 -1 0 -1 3.472 1.374 9.151 5 -1 -1 0 0 3.451 1.368 9.320 6 -1 -1 0 1 3.410 1.382 9.729 7 -1 -1 1 -1 3.155 1.287 8.821 8 -1 -1 1 0 3.214 1.311 8.700 9 -1 -1 1 1 3.253 1.355 8.819 10 -1 0 -1 -1 4.149 1.836 10.980 11 -1 0 -1 0 4.038 1.780 11.530 12 -1 0 -1 1 3.907 1.744 12.320 13 -1 0 0 -1 3.762 1.446 9.720 14 -1 0 0 0 3.731 1.420 9.980 15 -1 0 0 1 3.680 1.414 10.480 16 -1 0 1 -1 3.435 1.356 9.040 17 -1 0 1 0 3.484 1.360 9.010 18 -1 0 1 1 3.513 1.384 9.220 19 -1 1 -1 -1 4.649 2.067 11.559 20 -1 1 -1 0 4.528 1.991 12.200 21 -1 1 -1 1 4.387 1.935 13.081 22 -1 1 0 -1 4.252 1.674 9.949 23 -1 1 0 0 4.211 1.628 10.300 24 -1 1 0 1 4.150 1.602 10.891 25 -1 1 1 -1 3.915 1.581 8.919 26 -1 1 1 0 3.954 1.565 8.980 27 -1 1 1 1 3.973 1.569 9.281 28 0 -1 -1 -1 3.638 1.565 11.671 29 0 -1 -1 0 3.577 1.515 11.980 30 0 -1 -1 1 3.496 1.485 12.529 31 0 -1 0 -1 3.321 1.208 10.121 32 0 -1 0 0 3.340 1.188 10.140 33 0 -1 0 1 3.339 1.188 10.399 34 0 -1 1 -1 3.064 1.151 9.151 35 0 -1 1 0 3.163 1.161 8.880 36 0 -1 1 1 3.242 1.191 8.849 37 0 0 -1 -1 3.888 1.670 12.550 38 0 0 -1 0 3.817 1.600 12.950 39 0 0 -1 1 3.726 1.550 13.590 40 0 0 0 -1 3.561 1.310 10.650 41 0 0 0 0 3.570 1.270 10.760 42 0 0 0 1 3.559 1.250 11.110 43 0 0 1 -1 3.294 1.250 9.330 44 0 0 1 0 3.383 1.240 9.150 45 0 0 1 1 3.452 1.250 9.210 table 2. design of experiment and data related to bead geometry parameters 47 the journal of engineering research vol. 7, no. 1, (2010) 42-52 many, model no. dmlm, s6d & dfc320 and q win software). the domain of the experiment is shown in appendix (table 1). the design of experiment (doe) and collected experimental data, related to individual quality indicators of bead geometry are listed in appendix (table 2). these data were utilized in proposed integrated optimization approach, to be discussed later. 6. results and discussions of proposed optimization approach 6.1 calculation of individual desirability values and overall desirability function the flow chart of the approach is furnished below in appendix (fig. 3). response data were transformed to their individual desirability values using a desirability function approach (fuller, d. and scherer, w., 1998). these are shown in table 3. for depth of penetration hb (higher-the-better) and for reinforcement as well as bead width lb (lower-the-better) criteria were selected. the index of desirability function was the selected one. in this computation the minimum and maximum values of each response (table 2) were denoted as ymin and ymax respectively. individual desirability values of the responses were clustered to calculate the overall desirability value (table 3). it was assumed that all responses are equally important. the same weight was assigned to all responses. 6.2 development of response surface model of overall desirability rsm was applied to derive a mathematical model of overall desirability. overall desirability was expressed as a function of four process control parameters. the model consists of linear, square (quadratic) and interaction terms of the process parameters affecting the overall desirability value. the constant term and coefficients of the 46 0 1 -1 -1 4.338 1.931 13.089 47 0 1 -1 0 4.257 1.841 13.580 48 0 1 -1 1 4.156 1.771 14.311 49 0 1 0 -1 4.001 1.568 10.839 50 0 1 0 0 4.000 1.508 11.040 51 0 1 0 1 3.979 1.468 11.481 52 0 1 1 -1 3.724 1.505 9.169 53 0 1 1 0 3.803 1.475 9.080 54 0 1 1 1 3.862 1.465 9.231 55 1 -1 -1 -1 3.523 1.509 14.101 56 1 -1 -1 0 3.502 1.445 14.260 57 1 -1 -1 1 3.461 1.401 14.659 58 1 -1 0 -1 3.266 1.182 11.911 59 1 -1 0 0 3.325 1.148 11.780 60 1 -1 0 1 3.364 1.134 11.889 61 1 -1 1 -1 3.069 1.155 10.301 62 1 -1 1 0 3.208 1.151 9.880 63 1 -1 1 1 3.327 1.167 9.699 64 1 0 -1 -1 3.723 1.644 14.940 65 1 0 -1 0 3.692 1.560 15.190 66 1 0 -1 1 3.641 1.496 15.680 67 1 0 0 -1 3.456 1.314 12.400 68 1 0 0 0 3.505 1.260 12.360 69 1 0 0 1 3.534 1.226 12.560 70 1 0 1 -1 3.249 1.284 10.440 71 1 0 1 0 3.378 1.260 10.110 72 1 0 1 1 3.487 1.256 10.020 73 1 1 -1 -1 4.123 1.935 15.439 74 1 1 -1 0 4.082 1.831 15.780 75 1 1 -1 1 4.021 1.747 16.361 76 1 1 0 -1 3.846 1.602 12.549 77 1 1 0 0 3.885 1.528 12.600 78 1 1 0 1 3.904 1.474 12.891 79 1 1 1 -1 3.629 1.569 10.239 80 1 1 1 0 3.748 1.525 10.000 81 1 1 1 1 3.847 1.501 10.001 p (penetration), r (reinforacement) and w (bead width) 48 the journal of engineering research vol. 7, no. 1, (2010) 42-52 factors/interaction of factors were evaluated with the statistical-software package minitab (release 14). minitab's multiple linear regression approach was used to derive this model (eq. 11). (11) figure 3. proposed optimization approach 2 2 2 2 0.625 0.0505 0.0317 0.0677 0.0244 0.0343 0.0445 0.112 0.0315 0.0003 . 0.0166 . 0.0049 . 0.114 . 0.0131 . 0.0411 . d v wf tr n v wf tr n v wf v tr v n wf tr wf n tr n                step 1: selection of appropriate desirability function step 2: transformation of response data into invidual desirability values step 3: calculation of overall desirability step 4: mathematical modeling of overall desirability using rsm step 5: check for model adequacy step 6: optimization pso algorithm step 7: validation and recommendation 49 the journal of engineering research vol. 7, no. 1, (2010) 42-52 sl. no. dp dr dw d 1 0.4953 0.3280 0.8223 0.5112 2 0.4315 0.3666 0.7624 0.4941 3 0.3552 0.3837 0.6712 0.4506 4 0.2574 0.7428 0.9411 0.5646 5 0.2442 0.7492 0.9191 0.5519 6 0.2183 0.7342 0.8657 0.5177 7 0.0574 0.8360 0.9842 0.3615 8 0.0946 0.8103 1.0000 0.4248 9 0.1192 0.7631 0.9845 0.4474 10 0.6845 0.2476 0.7024 0.4919 11 0.6145 0.3076 0.6306 0.4921 12 0.5319 0.3462 0.5275 0.4597 13 0.4404 0.6656 0.8669 0.6334 14 0.4208 0.6935 0.8329 0.6241 15 0.3886 0.6999 0.7677 0.5933 16 0.2341 0.7621 0.9556 0.5545 17 0.2650 0.7578 0.9595 0.5776 18 0.2833 0.7320 0.9321 0.5782 19 1.0000 0.0000 0.6268 0.0000 20 0.9237 0.0815 0.5431 0.3444 21 0.8347 0.1415 0.4281 0.3698 22 0.7495 0.4212 0.8370 0.6417 23 0.7237 0.4705 0.7911 0.6458 24 0.6852 0.4984 0.7140 0.6247 25 0.5369 0.5209 0.9714 0.6477 26 0.5615 0.5380 0.9635 0.6627 27 0.5735 0.5338 0.9242 0.6565 28 0.3621 0.5380 0.6122 0.4923 29 0.3237 0.5916 0.5719 0.4784 30 0.2726 0.6238 0.5002 0.4398 31 0.1621 0.9207 0.8145 0.4954 32 0.1741 0.9421 0.8120 0.5107 33 0.1735 0.9421 0.7782 0.5029 34 0.0000 0.9818 0.9411 0.0000 35 0.0625 0.9711 0.9765 0.3898 36 0.1123 0.9389 0.9806 0.4693 37 0.5199 0.4255 0.4975 0.4792 38 0.4751 0.5005 0.4452 0.4731 39 0.4177 0.5541 0.3617 0.4375 40 0.3136 0.8114 0.7455 0.5745 41 0.3192 0.8542 0.7311 0.5842 42 0.3123 0.8757 0.6854 0.5723 43 0.1451 0.8757 0.9178 0.4886 44 0.2013 0.8864 0.9413 0.5517 45 0.2448 0.8757 0.9334 0.5849 46 0.8038 0.1458 0.4271 0.3685 47 0.7527 0.2422 0.3630 0.4045 48 0.6890 0.3173 0.2676 0.3882 49 0.5912 0.5348 0.7208 0.6108 50 0.5905 0.5991 0.6946 0.6264 51 0.5773 0.6420 0.6370 0.6181 52 0.4164 0.6024 0.9388 0.6175 table 3. individual desirability values and calculated overall desirability 50 the journal of engineering research vol. 7, no. 1, (2010) 42-52 the extent of the significance of presence of factors (and interaction of factors) within the model was checked with the analysis of variance method (anova) (table 4). based on the calculated p-value (probability of significance) of the terms (from table 4) under considerations, insignificant terms (p-value less than 0.05) were excluded and the final reduced model consisting of significant terms was derived (eq. 12). this model was optimized (maximized) finally using pso a algorithm. (12) 53 0.4662 0.6345 0.9504 0.6551 54 0.5035 0.6452 0.9307 0.6712 55 0.2896 0.5981 0.2950 0.3711 56 0.2763 0.6667 0.2742 0.3697 57 0.2505 0.7138 0.2222 0.3412 58 0.1274 0.9486 0.5809 0.4126 59 0.1647 0.9850 0.5980 0.4595 60 0.1893 1.0000 0.5837 0.4798 61 0.0032 0.9775 0.7910 0.1346 62 0.0909 0.9818 0.8460 0.4226 63 0.1659 0.9646 0.8696 0.5182 64 0.4158 0.4534 0.1855 0.3270 65 0.3962 0.5434 0.1529 0.3205 66 0.3640 0.6120 0.0889 0.2706 67 0.2473 0.8071 0.5170 0.4691 68 0.2782 0.8650 0.5223 0.5009 69 0.2965 0.9014 0.4961 0.5100 70 0.1167 0.8392 0.7729 0.4230 71 0.1981 0.8650 0.8160 0.5190 72 0.2669 0.8692 0.8277 0.5769 73 0.6681 0.1415 0.1203 0.2249 74 0.6423 0.2529 0.0758 0.2310 75 0.6038 0.3430 0.0000 0.0000 76 0.4934 0.4984 0.4976 0.4964 77 0.5180 0.5777 0.4909 0.5277 78 0.5300 0.6356 0.4529 0.5343 79 0.3565 0.5338 0.7991 0.5337 80 0.4315 0.5809 0.8303 0.5926 81 0.4940 0.6066 0.8302 0.6289 dp (desirability of penetration), dr (desirability of reinforcement), dw (desirability of bead width), od (overall desirability) predictor coefficient p-value comment constant 0.6250 0.000 significant v -0.0505 0.000 significant wf 0.0317 0.003 significant tr 0.0677 0.000 significant n 0.0244 0.019 significant v2 -0.0343 0.054 insignificant wf*wf -0.0445 0.013 significant tr*tr -0.1120 0.000 significant n*n -0.0315 0.076 insignificant v*wf -0.0003 0.983 insignificant v*tr 0.0166 0.185 insignificant v*n 0.0049 0.694 insignificant wf*tr 0.1140 0.000 significant wf*n -0.0131 0.292 insignificant tr*n 0.0411 0.001 significant s = 0.0742512 r-sq = 78.4% r-sq (adj) = 73.8% table 4. check for significance of the constant and coefficients in the model 2 2 0 581 0 0505 0 0317 0 0677 0 0244 0 0445 0 112 0 114 0 0411 d . . v . wf . tr . n . wf . tr . wf .tr . tr.n          p value (probability of significance), s, r-sq (determine to what extent the model can predict well) 51 the journal of engineering research vol. 7, no. 1, (2010) 42-52 6.3 particle swarm optimization (pso) in the present study, the reduced mathematical model for overall desirability (eq. 12) was optimized using a pso algorithm. it is a constrained optimization problem since the experimental domain was defined by the bounds on the process variables (v, wf, tr and n). the objective is to maximize (eq. 12) subject to the bounds on the process variables. the values of the parameters of the pso algorithm used here as follows: population size= 50, range of velocity variation vmax = +4, vmin = -4, maximum number of iteration = 100, weighting factor = 0.8, decrement factor (alpha) = 0.9 and social parameters c1 = 2.0 and c2 = 2.0. by trial and error the values of the aforesaid parameters were chosen so as to improve an objective function value (overall desirability) at the optimal setting. after optimization the optimal setting becomes: (optimal value of overall desirability becomes 0.707). figure 4 shows the convergence curve in pso. due to non-availability of optimal factors value within equipment's provision, a compromise has to be made. the optimal setting should be modified and set to: after evaluating the optimal parameter settings, the next step is to predict and verify the enhancement of quality characteristics using the optimal parametric combination. table 5 reflects the satisfactory result of confirmatory experiment. it indicates that the quality of the weld has improved. 7. conclusions weld quality in saw depends on features of bead geometry, mechanical-metallurgical characteristics of the weld as well as on weld chemistry. the weld quality improvement is treated as a multi-factor, multi-objective optimization problem. the practical application of saw requires efficient optimization methodology because process parameters are expected to interact in a complex manner. therefore, any optimization algorithm must seek to identify interaction effects of input factors and be incorporated in the course of an optimization procedure in a convenient way for developing an efficient methodology. the developed methodology based on rsm, desirability function and pso algorithm can be applied in practice for continuous quality improvement and off-line quality control. the desirability function approach converts each of the responses (objectives) into their individual desirability value. corresponding to each objective, these individual desirability values are then accumulated to compute the overall/composite desirability function, which is to be optimized (maximized) finally. rsm has been applied to derive a mathematical model of overall desirability represented as a function of process control parameters. this mathematical model has been optimized within an experimental domain. although the paper considers saw, the procedure is quite generic and can be applied to any process where complex relations among input and output parameters are difficult to predict. references correia, d.s., goccalves, c.., sebastiao, s.c.junior and ferraresi, v.s., 2004, " gmaw welding optimization using genetic algorithms", journal of brazilian society of mechanical science and engineering, 26(1) , doi: 10.1590/s1678-58782004000100005. dongcheol, k., sehun, r. and hyunsung, p., 2002, "modeling and optimization of a gma welding process by genetic algorithm and response surface methodology", international journal of production research, vol. 40(7), pp. 1699-1711. fuller, d. and scherer, w., 1998, "the desirability function: underlying assumptions and application implications, systems, man, and cybernetics", 1998 ieee international conference, vol. 4, pp. 40164021. 0 384 0 916 0 982 0 936 .v .wf tr . n .                 0 1 1 1 v wf tr n                 figure 4. pso convergence curve optimal setting prediction experiment level of factors 0 1 1 1v wf tr n 0 1 1 1v wf tr n overall desirability 0.769 0.789 table 5. results of confirmatory experiment n.b. subscripts on factors notation represents factor levels 52 the journal of engineering research vol. 7, no. 1, (2010) 42-52 hsien-yu tseng, 2006, "welding parameters optimization for economic design using neural approximation and genetic algorithm", international journal of advanced manufacturing technology, vol. 27(9-10), pp. 897-901. leticia, c., cagnina and susana, c. esquivel., 2008, "solving engineering optimization problems with the simple constrained particle swarm optimizer", informatica, vol. 32, pp. 319-326. price, w.l., 1977, "a controlled random search procedure for global optimization", the computer journal, vol. 20(4), pp. 367-370. pasandideh s.h.r. and niaki, s.t.a., 2006, "multiresponse simulation optimization using genetic algorithm within desirability function framework", applied mathematics and computation, vol. 175(1), pp. 366-382. praga-alejor, r.j., torres-trevino, l.m. and pinamonarrez, m.r., 2008, "optimization welding process parameters through response surface, neural network and genetic algorithms", proceedings of the 2008 electronics, robotics and automotive mechanics conference, vol. 00, pp. 393399. tay, k.m. and butler, c., 1996, "modeling and optimizing of a mig welding process a case study using experimental designs and neural networks", quality and reliability engineering international, vol. 13(2), pp. 61 70. zhao, x., zhang, y. and chen g., 2006, model optimization of artificial neural networks for performance predicting in spot welding of the body galvanized dp steel sheets, advances in natural computation, vol. 4221/2006, doi 10.1007/11881070. civl070127.qxd the journal of engineering research vol. 6, no.1, (2009) 8-14 1. introduction inorganic pollutants from natural and anthropogenic sources are often found in drinking water. these pollutants include cadmium, lead, mercury, zinc, aluminum, chromium, cobalt, copper, nickel, and iron. the removal in order to of these pollutants should result in extremely low levels to comply with regulations. the conventional treatment sequence of coagulation-flocculation, sedimentation, filtration usually results limited removals of these compounds, which makes advanced treatment processes necessary. activated carbon is used almost exclusively because of its high overall adsorption capacity when the removal of these compounds is desired (najm, et al. 1991). carbon adsorption is expensive, nonselective, adsorptive interference in the presence of naturally organic matter (nom), competition effects of nom are exacerbated with time, and it interacts with treatment chemicals. surfactant modified clays can provide selectivity, and are produced from inexpensive base material and are chemically regenerable. the absorption capacity of clay _____________________________________ *corresponding author’s e-mail: ahmemoh@iit.edu minerals have been shown to improve significantly due to the modification with quaternary ammonium compounds. the molecular structure of the modifying cations was also shown to play an important role in controlling the preference absorption and it could be concluded that modification of a specific clay mineral with a quaternary ammonium salt can produce a sorbent that is capable of sorbing inorganics from aqueous solutions (barrer, et al. 1989; barrer, et al. 1955). the sorption capacity of this modified clay can be maximized by the choice of appropriate surfactant. despite the fact montmorillonite have a large internal surface area, which can be as large as 800 m2/g (jamrah, et al. 1993; ahmed, 1996) and is comparable to 1000 m2/g for activated carbon (weber, 1972), very few studies were carried out to evaluate the possibility of using these minerals for the treatment of water and wastewater from inorganic contaminants (heavy metals). examples of these studies include the use of hexadeceltrimethylammonium (hdtma) modified ziolite as an adsorbent for the removal of chromate and other inorganic anions (haggerty, et al. 1994; katsuhiko, et al. 2006). most of cadmium adsorption on hdtma modified montmorillionite mohd. elmuntasir i. ahmed department of civil engineering, faculty of engineering & architecture, university of khartoum, sudan received 27 january 2007; accepted 21 april 2007 abstract: in this paper the possibility of cadmium removal from aqueous solutions by adsorption onto modified montmorillonite clay is investigated. batch adsorption experiments performed revealed an enhanced removal of cadmium using hdtma modified montmorillonite to 100% of its exchange capacity. modified montmorillonite adsorption capacity increases at higher phs suggesting adsorption occurs as a result of surface precipitation and hdtma complex formation due to the fact that the original negatively charged montmorillonite is now covered by a cationic layer of hdtma. adsorption isotherms generated followed a langmuir isotherm equation possibly indicating a monolayer coverage. adsorption capacities of up to 49 mg/g and removals greater than 90% were achieved. anionic selectivity of the hdtma modified monmorillonite is particularly advantageous in water treatment applications where high concentrations of less adsorbable species are present, and the lack of organoclay affinity for these species may allow the available capacity to be utilized selectively by the targeted species. keywords: environmental quality, adsorption, montmorillonite, cadmium removal ᣰsgƒh ¬d~©ÿg âjéfƒ∏jqƒªàæÿg ∫骩à°séh √é«ÿg øe ωƒ«eoéµdg ¢ué°üàeghdtma ~ªmg ô°üàæÿg ~ªfi áá°°uuóóÿÿgg ∫éh ¬d~©ÿg âjéfƒ∏jqƒªàæÿg áhôj ∫骩à°sg á«féµeg åëh ” ábqƒdg √òg ‘ :hdtma” »àdg ¬àhéãdg üqééàdg .¢ué°üàe’g ≥jôw øy √é«ÿg øe ωƒ«eoéµdg ádgr’ ö«°sîd áé«àf ∂dp h »æ«lhq~«¡dg ºbôdg ioéjõh ogoõj ¬dgr’g ¿g h á«doéñàdg âjéfƒ∏jqƒªàæÿg ábéw øe %100 πj~©àh ωƒ«eoéµdg øe ádgõÿg ᫪µdg ioéjr øµá ¬jg âø°ûc égagôlg ∫g áæë°ûh áñdé°ùdg âjéfƒ∏jqƒªàæÿg áæë°t ∫oé©àd ¬é«àf ƒg é° «g ¬dgr’g oéjorg .äéñcôe øjƒµj ºk ∫~©ÿg âjéfƒ∏jqƒªàæÿg í£°s ≈∏y ωƒ«eoéµdg hdtma~lh .áñlƒÿg §£fl ƒg ö°ùf’g iqgô◊g §£ıg ¿g é° jglangmuir.%90 ¤g ádgrg ö°ùfh ºz/º¨e 49 ¢ué°üàeg ábéw â≤≤m üqééàdg .~mgƒdg í£°ùdg ägp á«£¨àdg ≈∏y ∫~j é‡ áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .ωƒ«eoéµdg ádgrg ,âjéfƒ∏jqƒªàæÿg ,¢ué°üàe’g ,áä«ñdg ioƒl : 9 the journal of engineering research vol. 6, no.1, (2009) 8-14 the adsorption studies that have been performed with clays have utilized the clay in its natural state, with no effort to improve the adsorptive capacity of the clay by pretreatment with other chemicals. the chemical treatment of the clay may be needed in some applications (jamrah, et al. 1993; ahmed, 1996). the major goal of this paper is to chemically modify naturally occurring wyoming montmorillonite with quaternary amine salt (hdtma). the objective of the chemical treatment was to improve the adsorptive capacity of montmorillonite. 2. base materials organo-clays are typically smectite clays in which the natural exchange cations are replaced by alkylammonium or quaternized cationic surfactant (barrer, 1978). this substitution of organic cations changes the surface properties of the clay from strongly hydrophilic to organophilic and decreases the hydration of the clay as the aluminosilicate surface area of the mineral not covered by the organic exchange ions decreases. further, these exchange ions act as pillars that hold the aluminosilicate sheets apart (jamrah, et al. 1993; ahmed, 1996). organoclay complexes have been utilized for many years (brindley, et al. 1969). the clay used in this work, wyoming montmorillonite, has been utilized in previous studies ((jamrah, et al. 1993; ahmed, 1996; haggerty, et al. 1994; katsuhiko, et al. 2006). it is also used as base materials in some commercially available organo-clays (eg. bentones b27, b34, b38). 2.1 montmorillionite montmorillonite is an aluminum dioctahedral smectite clay mineral. based on the data presented in table 1, it would appear that the different types of montmorillonite can be distinguished by their structural formulae. wyoming montmorillonite, prior to use, was purified to achieve a size of < 2 µm by the process described elsewhere (jamrah, et al. 1993). the cation exchange capacity (cec) of the clay was measured using the method of calcium-magnesium exchange (ahmed, 1996), and the cec per unit mass of the clay was found to be 93.0 meq/100 grams of the dry clay. 2.2 surfactant the surfactant used in this study is hexadecyltrimethylammonium bromide, table 2, which is a quaternary ammonium compound, which is a cationic surfactant that can be conveniently produced by reacting a suitable tertiary amine with an alkylating reagent. the reagent is usually an organic halide or organic sulfate. the hydrophobic group in the quaternary ammonium salts has a positive charge when dissolved, and thus adsorbs strongly onto negatively charged surfaces. furthermore, this positive charge is not affected by the ph of the media, indicating that quaternary ammonium salts are stable, especially in acidic solutions 5. 3. experimental methods and analysis 3.1 batch adsorption experiments all adsorption experiments were carried out using a standard batch procedure at room temperature (25oc). first, the adsorbent was prepared in 50 ml borosilicate glass centrifuge tubes with teflon-lined screw caps. clay samples of 50 mg were weighed in each of the tubes (as determined from preliminary experiments described later), then the surfactant solutions were added in amounts sufficient to replace the desired percentage of the cec. although preliminary kinetic studies on hdtma showed that the reaction was complete in 4-6 hours (jamrah, et al. 1993), the tubes were then shaken on a rotary shaker (3 rpm) for a period of 16-24 hours. the clay suspension in the tubes was allowed to settle, and the separation of the phases was achieved by evaporation at room temperature. adsorption isotherms for the organo-clays were then generated, with cadmium using 200 ml flasks. stock solution of cadmium was prepared, and added in the desired initial concentrations. cadmium concentrations were measured before and after each ph adjustment. in the case of controlled ph experiments, the ph was periodformula source ( si7 .790 +4 al0.210 +3 )( al 3.070 +3 fe0.400 +3 mg0. 490 +2 )o20( oh)4 foster (1953) 11 ( si7 .680 +4 al0. 320 +3 )( al 3.050 +3 fe0.420 +3 mg0. 520 +2 )o20( oh)4 earley et al. (1953) 12 ( si7 .710 +4 al0.290 +3 )( al 3.010 +3 fe0.380 +3 fe0 .040 +2 mg0.520 +2 )o20(oh)4 grim (1968) 13 table 1. reported structural formulae of wyoming montmorilonite surfactant abbreviation molecular weight (g) structure hexadecylt ri-methylammonium bromide hdtm a 364.46 ch 3 n (ch 2 ) 15 ch 3 ch 3 ch 3 + table 2. the cationic surfactant used in clay modification 10 the journal of engineering research vol. 6, no.1, (2009) 8-14 ically checked during the adsorption process and adjusted using buffer solutions; a minimal headspace was allowed to insure good mixing throughout the adsorption experiment. the flasks were shaken in a gyrotory shaker model g33-b positive rotary motion shaker(250 rpm). a reaction period of 18-24 hours was allowed, although preliminary kinetic experiments on cadmium showed that the uptake reaction was complete in 14 hours (ahmed, 1996). following the reaction period, the flasks were allowed to settle and samples were taken for analysis of residual aqueous concentrations. phase separation was achieved by filtration using 0.45 µm membrane filters. sorption isotherms were generated using initial cadmium concentrations ranging from 5 to 30 mg/l. a total of 50 mg of modified montmorillionite and 150 ml of cation solution were placed in a 200-ml flasks. sorption at 100% hdtma loading (fully exchanged clay, (haggerty, et al. 1994 ) was examined for variable ph ranges and the effect of ph variation on cation sorption was determined and will be discussed later. ph was measured before and after equilibrium. 3.2 analysis samples were analyzed for cadmium concentration using a multiple ion meter (px meter) model 925 (fisher scientific), which measures ph, voltage, temperature and ion concentration. the electrode used was an orion combination electrode (sure-flow combination electrode model 9648), ph changes were accounted for in regard to the electrode potential behavior versus ph and limits of detection. prior to measurement of concentrations, samples were prepared in a 50 ml beakers after being filtered and adjusted for ionic strength. the ionic strength adjustment was done after the ph measurement, this is for the fact that the electrode performance depends solely on the use of ionc strength adjuster (5m nano3, model 940011). both the standard electrode and ion-specific electrodes were immersed in sample beaker solution after connection to the px meter. the px meter was then allowed to stabilize its reading (5 to ten minutes) and give signal before reading is recorded. 4. results and discussions 4.1 adsorption isothrems a study of the adsorption of cadmium onto modified montmorillonite as a function of the amount of adsorbent was first conducted to determine the optimum amount of clay (least ratio of amount of used clay to amount of removed cadmium) to be used in the following experiments. figure 1 indicates that the final concentration decreases with increasing amount of clay added but the clay loading capacity increases to a certain extent after which clay loading decreases indicating that the clay used is much more than needed for the limited amount of cadmium available. this experiment yielded 50 mg of clay as the optimum adsorbent amount to be used for 150 ml solution, 30 mg/l initial cadmium concentration, and ph 6. results in fig. 1 are normalized to the maximum obtained adsorption capacity of clay and a maximum initial concentration of 30 mg/l. figure 2 shows cadmium adsorption isotherms at different phs. the results shown are averages of two experiments and errors in these isotherms were in the 5% range. as indicated by fig. 2 modified montmorillonite has a little affinity for cadmium ion cd2+ (low ph) compared to other species cd(oh)+, cd(oh)2, hcdo2-, cdo2 consecutively in order (higher phs), however the amounts of each, sorbed, cannot be determined exactly but can be arranged in the order of which is favorable. also according to a study conducted by (haggerty, et al. 1994) on the sorption of anions by organoclays he concluded that sorption of some anions that are present in the solution is favorable to some degree, and that might even be greater at medium ph ranges (5.8 to 7.5). hence we can conclude figure 1. normalized final cadmium concentration (cf) and normalized solid phase concentration (q) as a function of the amount of clay mass used in 150 ml solution with 30 mg/l cadium 0 50 100 150 200 250 0 0. 2 0. 4 0 .6 0. 8 1 11 the journal of engineering research vol. 6, no.1, (2009) 8-14 for sure that at lower ph ranges sorption of cadmium cationic species is strong relatively (as we will see later that sorption of cadmium is generally weak specially below ph 6 where cadmium is present entirely as cd+), and at higher ph ranges sorption of anionic species is strong. also at ph's greater than 2 the isotherms have a steeper slope that means the adsorptive capacity increases at higher equilibrium solute concentrations at these ph values, over that at lower concentration. that means, at higher ph values, in general, the modified clay is efficient for column application than at lower ph values. however, it must be emphasized that the performance shown by the isotherm is indicative only of the performance of the modified clay under the static isotherm test conditions. additional column tests would have to be conducted to evaluate the performance of the modified clay in a continuous flow system. the isotherm data of fig. 3 can be described by a linear isotherm, a freundlich isotherm, or a lamgmuir isotherm equations as displayed by equations 1, 2, and 3 respectively (weber, 1972): (1) a freundlich isotherm, (2) or a langmuir isotherm [r] (3) figure 2. cadmium adsorption isotherms at different ph’s. initial cadmium concentrations ranging from 5 to 30 mg/l. a total of 50 mg of modified montmorillionite and 150 ml of cation solution were placed in a 200-ml flasks figure 3. adsorption of cadmium onto modified montmorillionite as a function of ph, (cf) is the normalized final cadium concentration and (q) is the normalized solid phase concentration 12 the journal of engineering research vol. 6, no.1, (2009) 8-14 where q is the sorbed mass per unit weight of adsorbent, cf is the final equilibrium solution concentration, and k is the adsorption capacity, a is the langmuir adsorption constant. the adsorption isotherms fitting at different phs are given in table 3. the linear equation is valid ideal solution conditions, which implies low solute concentration at which the solute is completely ionized and at equilibrium conditions. the nonlinearity observed on the adsorption isotherms at higher solute concentrations (>20 mg/l) is mainly due to the violation of the assumption of monolayer adsorption of the linear isotherms. the nonlinearity observed on the adsorption isotherms at higher solute concentrations (>20 mg/l) is mainly due to the violation of the assumption of monolayer adsorption of the linear isotherms. langmuir's isotherm is based on the assumption that points of valency exist on the surface of the adsorbent and that each of these sites is capable of adsorbing one molecule; thus, the adsorbed layer will be one molecule thick. furthermore, it is assumed that all the adsorption sites have equal affinities for molecules of the adsorbate and that the presence of adsorbed molecules at one site will not affect the adsorption of molecules at an adjacent site. table 3 suggests a strong relationships is provided by all isotherms, however, the langmuir is the best representation of the adsorption of cadmium onto hdtma modified montmorillionite. this result conforms to the monolayer adsorption assumption underlying the langmuir isotherm. 4.2 ph effects the ph of a solution from which adsorption occurs may, for many for one or more a number of reasons, influences the extent of adsorption. because hydrogen and hydroxide ions are adsorbed quite strongly, the adsorption of other ions is influenced by the ph of the solution. furthermore, to the extent to which the ionization of an acidic or basic compound affects its adsorption, ph affects adsorption in that it governs the degree of ionization. cadmium (cd2+) is present totally as the divalent species up to ph 8, in the absence of any precipitating anions such as phosphate and sulfide. cadmium begins to hydrolyse at ph 9, forming cd(oh)+ species (pourbaix, 1974). higher hydroxy species of cadmium are not relevant at the ph values commonly found in the environment. dominance of cd2+ species in contaminated lakes is reported (snoeyink, et al. 1980). under the conditions of these studies, ph dependence was observed and the major portion of the cadmium was in the form of cd2+ at low phs and cdoh+ at high phs (>8) (pourbaix, 1974). generally complexes of metal tends to sorb more strongly at clay surfaces than do the free metal ions (ahmed, 1996). at ph values below 6, where the cadmium is present entirely as cd2+, there is very little adsorption, relatively. as the ph increases, the removal of cadmium from solution by adsorption on modified clay particles increases gradually. in the ph range 6 to 8.5 the cd2+ ion decreases, and the predominant cadmium species become the adsorbable hydroxy-cadmium complexes, ph isotherm type 2=ph 4=ph 6=ph 8=ph 10=ph 12=ph linear k 1.5771 2.0519 2.5759 2.7459 3.2704 3.7675 fkcq = 2r 0.963 0.9523 0.9827 0.9641 0.9865 0.9752 freundlich k 1.7638 2.833 2.9761 3.2223 3.7622 4.7909 n 1 0.9524 0.8702 0.9373 0.9346 0.9354 0.8925 nfkcq 1 = 2r 0.9806 0.9753 0.9875 0.9853 0.9907 0.9896 langmuir a 0.011681 0.033632 0.019097 0.013501 0.021121 0.026598 b 151.5152 83.33333 158.7302 232.5581 181.8182 175.4386 f f ac abc q + = 1 2r 0.9934 0.9842 0.9918 0.995 0.9945 0.9968 table 3. adsorption isotherm fitting at different ph’s obtained from figure 2 13 the journal of engineering research vol. 6, no.1, (2009) 8-14 cdoh+, cd(oh)o2. presentation in fig. 3 is to summarizes the adsorption of cd(ii) onto hdtma modified wyoming montmorillionite, using ph as the master variable. a geochemical equilibrium model (minteqa2 (allison geoscience consultants, inc, 2003)) was used to study the speciation of cadmium (30 mg/l) solution. the equilibrium ph was about 6.8 and mostly cadmium exists as a free ion. when the ph was fixed at 12 the speciation contained only hydroxyl species (59% cdoho2 and 40% cdoh3-) and no solids precipitated and therefore at such low cadmium concentrations precipitation is not a removal mechanism. also minteqa2 geochemical equilibrium model was used to study the speciation of cadmium in solutions containing the modified clay (50 mg in 150 ml solution and 30 mg/l cadmium). the adsorption data were taken from table 3 and the ph was fixed at the experimental values (2, 4, 6, 8, 10, and 12). the results of this modeling are shown in table 4. the adsorption was assumed to be solely dependent on free cadmium ion, and therefore table 4 can be used to infer upon the quantities of hydroxocomplexes adsorbed and adsorption nature. at phs up to 8 cadmium exist solely as a free ion and the increase in adsorption capacity is attributed to clay behavior, while at higher phs (10, 12) the formation of cadmium hydroxocomplexes increases significantly and surface reaction with clay is completely responsible for adsorption as in the case of ph 12. there is a suggestion that the almost total metal removal at the alkaline ph region may be caused by chemical precipitation. however, results from adsorption experiments using metal concentrations well below the precipitation limit have produced the same effect (haggerty, et al. 1994), and is included in this study. therefore, it seems clear that even at moderately high concentrations ie. 10-5 to 10-4 m, an adsorption of the metal ions by the modified clay can prevent metal precipitaion from occuring. moreover, adsorptive removal can only result in a less-than-monolayer coverage. this was observed in most metal adsorption experiments. both the rate and extent of adsorption by particles of a fixed size should vary approximately linearly with the dosage of adsorbent over a range of doses that do not result in great differences in the concentration of solute remaining in bulk solution phase 6. large differences in the concentration of residual solute introduce a second variable for both rate and capacity for adsorption. however the particle size of the modified montmorillonite is very small in the order of 10-6 m4, and this indicate the large surface area it has and that almost every pore is available for adsorption. 4.3 nature of the adsorbate in any consideration of adsorption from solution one fact is immediately evident; the solubility of the solute is, to a large extent, a controlling factor for adsorption equilibria. the solubility of cadmium increases at higher phs and therefore the adsorption capacity was higher. haggerty, et al. 1994, suggested three mechanisms for adsorption of oxyanions on modified clays which are; 1) sorption by hdtma admicelles; 2) chemical reduction; and 3) surface precipitation and formation of hdtma complexes. the first mechanism is excluded since the clay is rinsed with water after modification to remove excess surfactant. the second and third mechanisms are possible and confirm the increase of adsorption at higher phs but the third is the likely mechanism since oxyanions can attach to the cationic end of the hdtma on the montmorillonite surface. only these qualitative data can be inferred from this discussion, however more detailed work is needed to confirm this conclusion. 5. conclusions in summary, this study has investigated an alternative sorbent (montmorillionite modified with hdtma) that can be used in water treatment applications. however, although results presented herein are promising, more ph percent of total cadmium bound in species +2cd +cdoh ocdoh 2 − 3cdoh adsorbedcd +2 2 91.9 0 0 0 8.1 4 87.3 0 0 0 12.7 6 85.8 0 0 0 14.2 8 84.4 0.6 0 0 15.0 10 41.0 30.9 19.2 0 8.8 12 0 1.0 59.3 39.6 0 table 4. minteqa2 speciation results of 150 ml of 30 mg/l cd and 50 mg modified clay 14 the journal of engineering research vol. 6, no.1, (2009) 8-14 experimental work needs to be done especially to investigate leaching of hdtma and bromide into the treated water, and reaction with disinfectants (chlorine and ozone) to form potentially harmful byproducts. the conclusions of this paper are: 1. the ability of modified montmorillonite to remove cadmium hydroxo-compounds from solution increased with ph due to the fact that organoclays removed the anionic species first, the modified clays were able to selectively remove these sorbates from water with no significant competition from other inorganic components of surface waters. 2. adsorption isotherms at different ph values indicated that the hydroxo forms of cadmium were preferentially adsorbed on the modified montmorillionite, which suggested that surface precipitation and hdtma complex formation might be the uptake mechanism. 3. the adsorption of cadmium ions onto modified montmorillionite was found to follow a langmuir isotherm suggesting a monolayer coverage. aknowledgments the author would like to acknowledge dr. ahmad jamrah (university of arbid, jordan) who supplied the wayoming montmorillionite and provided much needed help with materials. references ahmed, m.i., 1996, modified clays for the removal of cadmium from aqueous media. msce thesis, bradley universiy, peoria, il, usa. allison geoscience consultants, inc. and hydrogeologic, inc., 2003, minteqa2 for windowsequilibrium speciation model version 1.50, flowery branch, ga, usa. barrer, r.m. and macleod, d.m., 1955, "activation of montmorillionite by ion exchange and sorption complexes of tetra-alkylammonium montmorillionites," transactions of faraday society, vol. 51, pp. 1290-1300. barrer, r.m., 1978, "ziolites and clay minerals as sorbents and molecular sieves," academic press, london. barrer, r.m., 1989, "clay minerals as selective and shape selective sorbents," pure and applied chemistry, vol. 61, pp. 1903-1912. brindley, g.w., wiewiora, k. and wiewiora, a., 1969, "intercrystalline swelling of monmorillonite in some water organic mixtures," clay-organic studies, part xvii." american mineralogist, vol. 43, pp. 106-115. earley, j.w., osthaus, b. and milne, i.h., 1953, "purification and properties of montmorillonite," american mineralogist, vol. 7-8, pp. 102-108. foster, m.d., (1953). " geochernical studies of clay minerals (ii): relation between ionic substitution and swelling in montmorillonite." american mineralogist, 11-12, 143-154. grim, r.e., (1968), clay mineralogy, (ed. 2) mcgraw hill book company, new york. haggerty, g.m., bowman, r.s., (1994). "sorption of chromate and other inorganic anions by organo-ziolite." environmental science and technology, 28, 3, 452458. jamrah, a.i. and dentel, s.k. (1993). "application of organoclays to water treatment: selective removal of trichlorophenol." proc. 1993 awwa ann. conf., san antonio. katsuhiko, i. and junta, y., (2006). " sorption and desorption properties of cadmium and copper on soil clays in relation to charge characteristics." soil science and plant nutrition, 52, 5-18. najm, i.n., snoeeyink, v.l., lykins b.w., and adams, j.q, (1991). "using powdered activated carbon: a critical review." journal american water works association, 83,1,65. pourbaix, m., (1974). atlas of electrochemical equilibria in aqueous solutions, national association of corrosion engineers, houston, tx, usa. snoeyink, v.l, jenkins, d., (1980). water chemistry, 1st ed., john wiley & sons inc., nj. weber, w.j. (1972). adsorption, in physicochemical processes for water quality control. 1st ed., ny:john wiley & sons. pdf compressor c:\users\raoh\desktop\albadi1.xps the journal of engineering research vol. 7 no. 2 (2010) 1-9 a general model for describing the performance of brushless doubly-fed induction machines s. m. allam*a, a. m. azmyb, and m. a. el-khazendarc a electrical power and machines department, faculty of engineering, tanta university, tanta, egypt b electrical power and machines department, faculty of engineering, tanta university, tanta, egypt c electrical power and machines department, faculty of engineering, tanta university, tanta, egypt received 27 may 2009; accepted 19 june 2010 abstract: this paper presents a generalized model, by which the dynamic and steady-state behaviour of the brushless doubly-fed induction machine (bdfim) can be precisely predicted. the investigated doubly-fed machine has two sets of three-phase stator windings with different pole numbers. the rotor is a squirrel-cage type with a simple modification in order to support the two air-gap rotating fields that are produced by the stator windings and have different pole numbers. the machine model is derived in the qdo-axis variables. the qdoaxes are attached to rotor and hence, it rotates at the rotor speed ( m). the electromagnetic torque expression is also obtained based on the presented model. the winding function method is applied as a convenient approach that can be used for machine-winding inductances calculations. sample case studies are introduced to examine the performance of the proposed model in both steady-state and dynamic conditions. keywords: brushless doubly-fed machine, dynamic modelling, steady-state modelling, generalized theory, induction machine, self-cascaded, speed control, variable speed drives ___________________________________ *corresponding author’s e-mail: sm_allam@hotmail.com * 2 the journal of engineering research vol. 7 no. 2 (2010) 1-9 1. introduction squirrel-cage induction motors are the most widely used motors in ac-drive applications. this is attributed to their ruggedness, low cost, and few maintenance requirements. the number of industrial applications, which utilize ac adjustable speed drives (asds), is increasing rapidly. the primarily reason for this development is the significant advent of power electronics. the main disadvantage of using singly-fed machines for asds is the need for large and expensive power conditioning units that should withstand the full motor power. this increases both the capital cost and the harmonic current pollution of the overall system (li et al. 1991). the brushless doubly-fed induction machines (bdfims) shows a great promise in reducing the rating of power electronic converters needed for asds application (boger et al. 1995). the bdfim represents a development of self-cascaded machines with two three-phase stator windings having different number of poles accommodated in a single machine frame. in addition, it is required to introduce a modification to the cage rotor structure (boger et al. 1996; boger and wallace, 1995; zaiping et al. 1998 and wing et al. 2002). in a bdfim, the rotor structure determines the magnetic coupling between the two stator windings, which in turn determines the machine behaviour (wang et al. 2002). bdfim is a special form of slip recovery machines that reduce the capacity of the required inverter to be used if the required speed-control range is limited ((li et al. 1991; boger et al. 1995; boger et al. 1996; boger et al. 1996; boger and wallace, 1995; zaiping et al. 1998; wing et al. 2002 and wuilliamson et al. 1997). this will lead to a significant reduction in the drive cost. owing to its brushless property, the bdfim is mechanically robust and reliable. it is found that the bdfim has the ability to provide precise speed control and good power factor with reduced converter capacity, as low as 25% of the machine rating (boger et al. and boger and wallace, 1995). a further advantage of the bdfim is derived from the fact that the power electronic converters do not process the bulk power entering or leaving the machine directly. the dynamic model of brushless doubly-fed machines (bdfms) was first presented in (li et al. 1991). however, the presented model is based on a special configuration of both stator and rotor winding connections. the configurations of both stator and rotor appear to be complex compared to conventional induction machine. the dynamic model introduced in (li et al. 1991), has addressed the 3/1 (power winding to control winding pole-pair numbers) exclusively. an extended model, which is appropriate for any number of pole-pairs, was introduced in (boger et al. 1995). however, the structure of the bdfm used is similar to that used in (li et al. 1991). the main problem of these models is the complicated structure of both stator and rotor windings that these models based upon. nomenclature b, j : viscous coefficient and rotor inertia respectively. irk : rotor loop current. nk : turns per phase of either stator windings. p : the differential operator "d/dt". pk : pole-pairs of either stator windings. r, l : rotor radius and axial length respectively. rsk : per-phase stator resistance. tl,tem : load and electromagnetic developed torque respectively. isk, vsk : stator current and voltage respectively. o : permeability for vacuum. m : rotor mechanical angular position with respect to phase (a) reference axis of the two sets of stator windings. : angular position along the stator inner surface. g-1( m : the inverse air-gap function. n1 ( m : the winding function of winding (i). l lk : stator leakage inductance. lmk : stator magnetizing inductance. mr : mutual inductance between two rotor loops. lmr : rotor loop magnetizing inductance. m : mechanical rotor speed in rad./sec. ek : stator electrical frequency in rad./sec. rb, lb : rotor bar resistance and leakage inductance respectively. re, le : end ring segment resistance and leakage inductance respectively. 3 the journal of engineering research vol. 7 no. 2 (2010) 1-9 a new model based on a dual stator winding induction machine is presented in (munoz and lipo, 2000). the two-stator windings are simultaneously fed from two independent variable frequency power supplies. this results in increasing the capital cost of the overall system. the aim of this paper is to present a generalized model valid for predicting the dynamic and steady-state performance of bdfims under different loading conditions. in addition, the paper aims at investigating the possibility of simpler structure of the bdfim than that used before. the investigated bdfim consists of a simply modified squirrel-cage rotor and a stator with two separate windings wound for a dissimilar number of poles. each stator winding is fed from an independent power supply. only one of the two stator windings is fed from a variable-frequency variable-voltage inverter. 2. system description figure 1 shows the schematic diagram of the investigated bdfim. there are two three-phase windings with different number of poles, in order to avoid magnetic mutual coupling, placed in the stator slots (li et al. 1991). these windings are simultaneously fed from two independent power supplies and each one produces a separate torque component. one set of stator windings is called the power windings; through which full machine power must pass, and is fed directly from a constant voltage constant frequency power supply. the other set is called the control windings, through which partial controlled power passes, and is fed from a variable voltage variable frequency power supply. the rotor is a squirrel-cage type with a simple modification in its structure to support the two air-gap rotating fields that have different pole numbers produced by the two-sets of stator windings (wang et al. 2002). in (williamson et al. 1997), it is stated that the essential requirement for brushless doubly-fed machine action is that the frequency and distribution of currents induced in the rotor by the first air-gap field match that induced by the second. the two stator windings will be then coupled via the rotor, even though they have different pole numbers and may be excited at different frequencies. this implies that the number of rotor bars should be equal to the summation of pole-pair numbers of both stator windings. thus, the number of rotor bars will be small for any realistic combination of pole-pair numbers which results in a very high-referred rotor leakage reactance (willaimson et al. 1997) therefore, more work has been done to increase the number of rotor slots and distribute each bar between several slots. however, all ways used in rotor bars arrangements are not easily fabricated. furthermore, the cost of this type of machines is high since both stator and rotor have to be redesigned. in this paper, a very simple modification in the rotor cage is introduced. the rotor bars are divided into number of groups that equal the required number of rotor bars (summation of pole-pairs of the two-stator windings). separating each group from the others can be easily done at one side only of the two end rings. this means that one end ring is divided into a number of sections; each connecting a number of bars, while the other end ring connects all rotor bars. for 6-pole and 2-pole stator windings, the number of required rotor bars or groups equals four. therefore, one side of the end rings should be divided into four groups or sections. a schematic diagram of the bdfim is shown in fig. 1. (1) 3. inductance calculations the machine-winding inductances can be calculated by a particular convenient approach called winding functions method (liang et al. and xu et al. 1991). the winding function is defined as the spatial distribution of mmf due to a unit current flowing in the winding. according to this method, the machine inductances are calculated by an integral expression representing the placement of winding turns along the air-gap periphery. this method is particularly convenient for the analysis of unusual machines since it assumes no symmetry in the placement of any motor coil in the slots (xu et al. 1991). according to the winding function theory, the mutual inductance between any two windings (i) and (j) in any machine can be easily computed from the following expression: each stator phase is assumed to be sinusoidally wound along the air gap and the squirrel-cage rotor is modelled as "n" identical magnetically coupled circuits or loops (munoz and lipo, 2000). each loop encompasses two adjacent rotor bars and the connection portions of the end rings between them. a detailed analysis for the inductance calculations is given in a previous paper (allam et al. 2008). 4. dynamic model in qdo-axesrotor reference frame in order to eliminate the time-varying inductances in 6 6 6 2 2 2 figure 1. schematic diagram of the bdfim 6-pole stator winding 2-pole stator winding cage rotor main supply inventor 4 the journal of engineering research vol. 7 no. 2 (2010) 1-9 voltage equations, the machine was represented in the qdoaxes reference frame. the qdo-axes are fixed in the rotor, which rotates at the rotor speed ( m). figure 2 shows the phase axes relationship for phase "a" of any of the two stator windings axes and the nth-rotor loops axes. it should be noted that all equations used for the twostator windings are similar except for their different number of poles, turns per phase, resistance, and inductance. so, instead of writing theses equations twice, the subscript "k" is used denoting either the power or the control windings. hence, the same equation can be used to describe the two stator windings just by defining "k" with suitable number related to each one of the two sets of stator windings. the transformation matrix that transforms the abck-axis variables to either stator winding qdok-axis variables is given by: (2) where, r = m + , = r/2 and r is the angle between two adjacent rotor bars. the transformation matrix of either rotor circuits of nrotor loops axes to the corresponding qdork-axis variables can be written based on fig. 2 as: (3) it should be noted that, for a modified cage rotor, which is presented in this paper and due to the discontinuity of one of the two end rings, some rotor loops represent open loops. for the open loops at which the end ring is separated, the rotor loop current is set to zero and thus, the corresponding columns or elements are eliminated from the two matrices. the number of these columns equals the number of end ring sections (ie. summation of pole-pairs of both stator windings). this reduces the order of the matrices to be '3*m' instead of be '3*n' , where 'm' is the total number of closed rotor loops (ie. difference between total number of rotor bars or loops and open loops at which the end ring is separated). mathematically "m" can be given by: m = n pk, where pk is the summation of the pole-pairs of the two stator windings. 4.1 qdo-axes voltage equations the final qdok-axes stator voltage equation can be written in a matrix form as: (4) where: it should be noted that the rotor can be modelled as two independent circuits. each one has "n"-rotor loops and is related to one of the two stator windings with the same number of poles of the corresponding one. so, the subscript 'k' may be also used to define the two different rotor circuits related to the corresponding two stator windings. this is attributed to the mutual decoupling of the two stator windings since they have different number of poles. figure 3 shows a schematic diagram of n-rotor loops winding distribution for any of the corresponding two-stator windings just after defining 'k'. an open circuit rotor loop from one side of end rings is also shown in fig. 3. the final qdork-axes rotor voltage equation can be written in a matrix form as: figure 2. phase axes relationship of the mdfim figure 3. rotor loops winding distribution 5 the journal of engineering research vol. 7 no. 2 (2010) 1-9 4.2 qdo-axes flux linkage relations since the two stator windings are mutually decoupled, it should be noted that, in the flux linkage calculations, there is no component of flux linkage of any of the two stator windings or rotor circuits due to the current of the other one. hence, the flux linkage in each of the stator windings or rotor circuits is established due to only two components. the first is due to the current in the same winding or circuit, and the second is due to the current of the corresponding one, which has the same number of poles. the total qdok-axes equation of the stator flux linkage can be written as: the qdok-axes stator flux linkage due to the corresponding qdok-axes stator winding current is given by: (6) where: the qdok-axes stator flux linkage due to the corresponding qdork-axes rotor current is given by: (7) the total qdork-axes rotor flux linkage equation can be written as: the qdork-axes rotor flux linkage due to the corresponding qdork-axes rotor current is given by: (8) the qdork-axes rotor flux linkage due to the corresponding qdok-axes stator winding current is given by: the final form of qdo-axes system model in the rotor reference frame is given in the appendix. 4.3 electromagnetic developed torque equation the electromagnetic developed torque of a bdfim can be divided into two components. each one is due to the mutual coupling between one of the two-stator windings and the corresponding rotor circuit. thus, the total electromagnetic developed torque is given by the summation of these two components and can be written in terms of qdoaxis currents as: (10) where: (11) (12) 4.4 the electromechanical equation the electromechanical equation of a bdfim can be easily written as: (13) equations (4) through (13) provide a useful dynamic representation of the machine. thus, this model can be used to describe the dynamic performance of the machine in the qdo rotating axes. 5. steady-state model 5.1 steady-state voltage equations it is well known that the presented qdo-axes dynamic model is obtained in the rotor reference frame. hence, in order to obtain the steady-state voltage equations, the differential operator "p" should be replaced by the rotor frequency, ie. p = j ( ek pk m). the final steady-state voltage equations for both the two stator windings and the corresponding two rotor circuits can be rewritten in a matrix form as: (14) where 'sk' is the slip related to both stator windings and can be given by (15) equation (14) has a similar form to the corresponding steady-state voltage equations of the conventional inducλ l i where: where: (9) 6 the journal of engineering research vol. 7 no. 2 (2010) 1-9 tion machine. however, it can be observed that the term representing the mutual inductance between any of the two stator windings and the corresponding rotor circuits, lmkr, is not the same as the mutual-inductance term between any of the two rotor circuits and the corresponding stator windings, lmrk. this case is not observed in the conventional induction machine. the reason is attributed to the representation of the rotor in the proposed bdfim by an equally m-rotor loops or circuits, while the stator is represented by two equivalent three-phase circuits. however, in the conventional induction machine, equivalent three-phase circuits are used to represent both the stator and rotor. 5.2 steady-state electromagnetic torque express ion the steady-state electromagnetic torque developed by each one of the two stator windings can be expressed as: (16) or, it can be derived in another form as follows: (17) where, the acronym "re" stands for "real part". 6. sample case studies sample case studies are presented using the proposed model. the presented samples of the simulation results are obtained using matlab/simulink software with the calculated parameters of an actual set-up system. the stator and the rotor of a conventional three-phase squirrel-cage induction motor, rated at 1.1 kw, 460 v and 4-poles, are modified to fit the nature of bdfim. one set of the two stator windings, the power windings, has six poles, while the other, the control windings, has two-poles. all mutual and magnetizing inductances are calculated using winding function method (allam et al. 2008). stator resistance and leakage inductance are measured using standard techniques that used for the conventional induction machines. dimensions of both rotor bar and end-ring are used to calculate their resistance and leakage inductance. experimentally, in studying the performance of the bdfim under different operating conditions, the six-poles power windings are connected to a constant voltage constant frequency supply. on the other hand, the two-poles control windings are connected to an inverter operating in a constant voltage per hertz (v/f) mode (ie. doubly-fed operation) or left open (ie. singly-fed operation). 6.1 singly-fed operation figure v4 shows the run-up response of a singly-fed mode of the bdfim. in this mode of operation, the power windings are connected directly to the mains supply, while the control windings are left open under no-load condition. the power windings are connected in a star configuration, where the mains supply is set to 100-volt (line-line voltage) with a frequency of 50-hz. to simulate the case of open circuit in the control windings, a very high resistor is connected to its terminals. it can be noted that, the machine gives a response similar to that of the conventional 6-pole induction machine, where the motor runs up to a steady-state speed that is close to the synchronous speed, ie. 1000-rpm. the predicted 6-pole electromagnetic torque during free acceleration is also shown in the figure. the similarity between simulated and experimental results confirms that the proposed model is valid to describe the transient behaviour of the conventional induction machine operating only with one set of three-phase stator windings. 6.2 doubly-fed operation for more understanding of the machine behavior, the bdfim is examined in a doubly-fed mode under different modes of operations in a sequential manner. figure v.5 shows the run-up response of a doubly-fed mode of the bdfim followed by two different operating conditions. in the simulation process, the power winding is firstly connected directly to the mains supply while the control winding is left open under no-load condition up to 4.0 seconds, when a 0.25 n.m load torque is applied. then, the control winding is connected to the mains supply via an inverter after 4.0 seconds from the loading time (ie. 8.0 seconds from starting). the power windings and the control windings are connected in star configuration. experimentally, the voltage of the mains power supply is set to 100-volt (line-line voltage) with a frequency of 50-hz. on the other hand, the control-winding inverter is figure v.4. run-up response of a singly-fed mode of the bdfim under no-load conditions 7 the journal of engineering research vol. 7 no. 2 (2010) 1-9 operated in a constant voltage per hertz mode (220/50 v/hz) and the output control frequency is set to 10 hz. the control winding is excited with a reversed sequence with respect to the power-winding sequence. the predicted electromagnetic torque is also shown in fig. v.5. it can be observed from the figure that, the effect of inserting the control winding with a reversed sequence excitation is to reduce the steady-state operating speed. 6.3 steady-state characteristics to highlight the steady-state performance, fig. 6 represents the simulated and experimental speed as a function of the control winding frequency at constant load torque. where the mains supply is set to 150-volt (line-line voltage) with a frequency of 50-hz. the control winding is excited with the reverse sequence of the power winding. it is clear from the figure that the speed control through the frequency of the control winding is possible. the range of the speed variation can be increased with proper voltage and frequency control of the control winding. conclusions this paper has presented an accurate analysis that is valid for predicting the dynamic and steady-state performance of brushless doubly-fed induction machines. a general model in the qdo-axes has been derived from the basic relations of the machine. the winding function method is used as a viable method for calculating the machine inductances. in addition, a corresponding steady state model is derived using appropriate relations. in order to confirm the validity of the proposed model, sample case studies have been introduced using actual experimental set-up parameters. the agreement between simulated and experimental results ensures the validity and effectiveness of the proposed model. references allam, s.m., azmy, a.m., khater, m.m. and elkhazendar, m.a., 2008, "a generalized dynamic model for brushless doubly-fed induction machines," engineering research journal (erj), vol. 31(4), pp. 34554, minoufiya university, faculty of engineering, shebin el-kom, egypt, issn 1110-118. boger, m.s. and wallace, a.k., 1995, "performance capability analysis of the brushless doubly-fed machine as a wind generator," seventh international conference on electrical machines and drives, (conf. publ. no. 412), pp. 458-461. boger, m.s., wallace, a.k., spee, r. and li, r., 1996, "investigation of appropriate pole number combinations for brushless doubly-fed machines applied to pump drives," ieee transactions on industry applications, vol. 32(1), pp. 189-194. boger, m.s., wallace, a.k., spee, r. and li, r., 1995, "general pole number model of the brushless doubly-fed machine," ieee transactions on industry applications, vol. 31(5), pp. 1022-1028. li, r., wallace, a.k., spee, r. and wang, y., 1991, "two-axis model development of cage-rotor brushless doubly-fed machines," ieee transactions on energy conversion, vol. 6(3), pp. 453-460. liang, f., xu, l. and lipo, t.a., 1991, "d-q analysis of a variable speed doubly ac excited reluctance motor," electrical machines and power systems, vol. 19(2), pp. 125-138. munoz, a.r. and lipo, t.a. 2000, "dual stator winding induction machine drive," ieee transactions on industry applications," vol. 36(5), pp. 1369-1379. wang, f., zhang, f. and xu, l., 2002, "parameter and performance comparison of doubly-fed brushless machines with cage and reluctance rotors," ieee figure v.5. run-up response of a doubly-fed mode of a loaded mdfim with a control frequcy of 10 hz figure 6. speed/frequency variation of a bdfim reversed sequence) 8 the journal of engineering research vol. 7 no. 2 (2010) 1-9 transactions on industry applications, vol. 38(5), pp. 1237-1243. williamson, s., ferreira, a.c. and wallace, a.k., 1997, "generalized theory of the brushless doubly-fed machine. part 1: analysis," iee proc.-electr. power appl., vol. 144(2), pp. 111-122. xu, l., liang, f. and lipo, t.a., 1991, "transient model of a doubly excited reluctance motor," ieee transactions on energy conversion, vol. 6(1), pp. 126-133. zaiping, p., yanfeng, z. and nagrial, m.h., 1998, "review and future application of brushless doubly-fed machine," the 1998 international conference on communication, computer & power, pp. 306-311. 9 the journal of engineering research vol. 7 no. 2 (2010) 1-9 appendix a. final form of qdo-axes system model in rotor reference frame the journal of engineering research (tjer), vol. 19, no. 2, (2022) 180-189 0043 corresponding author's e-mail: honovo@unilag.edu.ng doi:10.53540/tjer.vol19iss2pp180-189 dynamic impact response of nanosized precipitates bearing waspaloy harrison okechukwu onovo*, david ehigie esezobor, and muideen adebayo bodude department of metallurgical and materials engineering, university of lagos, nigeria. abstract: dynamic response of nanosized precipitates bearing (npb) waspaloy, a class of superalloy, is crucial to design structural components for critical rotating applications and other severe dynamic impact predisposed applications. the focus of this study is, therefore, to investigate the compressive dynamic behaviour of npb waspaloy at a high strain rate under various deformation conditions. the technique of modified split hopkinson pressure (mshp) bar was implemented through impact deformation of waspaloy under wide-ranging deformation temperatures (-180 750 oc) and strain rates of (4*103 7.5*103 s-1). the outcomes of the experiments divulge how to flow stress relates directly to strain rate and is inversely proportional to deformation temperatures. work hardening rate affects npb waspaloy maximally at the iciest deformation temperature (-180 oc) and the highest strain rate (7.5*103 s-1) considered. under high straining and high rate of straining deformation conditions, the rate with which work-hardening occurs is destabilized due to the thermal softening effect during deformation. it is established that when strain is constant, the flow stress dependency on the rate of straining depicts linear relation. the thermal softening effect on npb waspaloy is concerted at around extreme strain rate (7.5*103 s-1) and iciest deformation temperatures within -180 ~ 25 oc. mapping of strain, strain rate and deformation temperatures representing input parameters with the resultant flow stress provide an unambiguous analytical view of the effects of dynamic impact deformation. flow stresses at elevated temperatures are correlated directly with grain growth, mainly influenced by adopted deformation temperatures. keywords: deformation, stress-strain, superalloy, strain rate, turbines, mapping. لمترسبات النانومترية في سبيكة الواسبالويلتأثير الديناميكي الاستجابة بودود وميدين أ.إيسيزوبور وديفيد إ. هاريسون أ. أونوفو أمر بالغ وهي فئة من السبائك الفائقة مترسبات النانومترية المحملة في سبيكة الواسبالويل الديناميكية لستجابة الا :الملخص لتأثيرات الديناميكية الشديدة. تركز لاألهمية لتصميم المكونات الهيكلية للتطبيقات الدوارة الحرجة وغيرها من التطبيقات المعرضة على استكشا معدالت هذه الدراسة، بالتالي، تحت ف سلوك الضغط الديناميكي لسبيكة الواسبالوي المحملة بمترسبات نانومترية تشوه عالية وظروف تشوه متنوعة. تم تنفيذ تقنية القضيب الضاغط المعدلة من هوبكنسون المقسم عبر تشوه التأثير للواسبالوي (. 1-ثانية 310× 7.5إلى 310× 4ت تشوه )من درجة مئوية( ومعدال 750إلى 180-تحت درجات حرارة تشوه شاسعة )من تكشف نتائج التجارب كيف يرتبط إجهاد التدفق مباشرة بمعدل التشوه ويتناسب عكسيًا مع درجات الحرارة التشوهية. يؤثر معدل التشوه حرارة درجة عند أقصى بشكل النانومترية بمترسبات المحملة الواسبالوي سبيكة في العمل درجة 180-) األبردتقسية معدل تشوه وأعلى معدل (.1-ثانية 310× 7.5) معتبر مئوية( العالي، يتم تقويض التشوه ومعدل العالية التشوه ظروف تحت حدوث التقسية بسبب تأثير التراجع الحراري أثناء التشوه. يتم تأسيس أنه عندما يكون التشوه ثابتًا، فإن تبعية إجهاد التدفق على التشوه تظهر عالقة خطية. يتركز تأثير التراجع الحراري على سبيكة الواسبالوي المحملة بمترسبات النانومترية عند معدل معدل درجة مئوية. توفر رسم الخرائط 25~ 180-( ودرجات حرارة التشوه األكثر برودة ضمن 1-ثانية 310× 7.5تشوه شديد ) ال مع إجهاد التدفق الناتج عرًضا تحليليًا واضًحا آلثار التشوه الناتج عن التأثير كمعلمات إدخ تهودرجات حراره للتشوه ومعدل الديناميكي. يرتبط إجهاد التدفق عند درجات حرارة مرتفعة مباشرة بنمو الحبوب، والذي يتأثر بشكل رئيسي بدرجات الحرارة .التشوهية المعتمدة .التشوه؛ سبيكة فائقة؛ معدل التشوه؛ توربينات؛ رسم الخرائط-اإلجهاد ه؛ و تش الكلمات المفتاحية: dynamic impact response of nanosized precipitates bearing waspaloy 181 nomenclature dhpb direct hopkins pressure bar shpb split hopkins pressure bar mshpb modified split hopkins pressure bar npb nano precipitate bearing fcc face-centred cubic crystal edm electric discharge machine lom light optical microscope sem scanning electron microscope tem transmission electron microscope 1. introduction crucial materials applications susceptible to highimpact deformations necessitate the adoption of highperformance materials that can withstand impact deformations at high loading rates and avert components failure and/or catastrophic breakdown of materials while in service. the choice of such materials will not elude superalloys synonymous with good mechanical properties at elevated temperatures, which include retained high strength, outstanding corrosion resistance, good oxidation resistance together with the amazing blend of formability and toughness (pollock and sammy 2006, liu et al. 2005). superalloy waspaloy essentially contains above 18% chromium next to the base metal, nickel, and is basically strengthened by precipitation hardening processes. the hardening procedure involves about three staggered heat treatment sequences, which commence with solution treatment, stabilizing by reheating, followed by age hardening treatments (xue, 2018). waspaloy is a speciality alloy extensively used in critical applications like turbine combustion components and critical mechanisms of nuclear reactors, to mention a few, that are prone to dynamic impact deformations. mostly, such applications require strength retentive materials with an additional high strength-to-weight ratio. waspaloy's strength-toweight ratio can be enhanced through the introduction of nanotechnology by tailoring the precipitation hardening phases to the nanoscale by means of controlled heat treatment (veerappan et al. 2018). consequently, adequate knowledge and documentation of nano-scaled precipitation hardened waspaloy high strain rates dynamic impacts response is indispensable to design structural components of waspaloy for critical rotating applications and other severe dynamic impact predisposed applications. well established and modified technique to study dynamic impact deformation (rate of material straining) of materials is the direct hopkins pressure bar (dhpb) (hopkinson, 1914), later altered to split hopkins pressure bar (shpb) by kolsky, as well known as kolsky bar (kolsky, 1949), and later improved to the modified split hopkins pressure bar (mshpb) (granier and grunenwald, 2006), to uncover behavioural characteristics of materials subjected to high strain rate in the form of either compressive, tensile or torsional mechanical loading. the role of materials' rate of straining in understanding materials' behaviour under dynamic impact loadings is very significant. some pieces of literature have reported strain rates of soft and hard materials, which fall into either low or high loading rates. materials strain rate sensitivity has been observed to increase precipitously beyond the critical region in a strain rate of about 1000 s-1 (lee and sun, 2004). more so, the flow stress, generally and predominantly in metals with fcc, has been experientially reported to intensify progressively with a rising strain rate in areas beneath crucial points (veerappan et al., 2018, lee and sun, 2004). flow stress subsequently increases more swiftly as the deformation rate progresses beyond the critical region of the strain rate. interestingly, the changeover of flow stress around the critical region has been ascribed to frequent alteration in the motion of dislocation (regazzoni et al. 1987). another school of thought opined that the transition could be regarded as enhanced rates of defeats formation, such as twin and dislocation. attention has been drawn to the reactional mirror of the flow stress-strain relation, which reveals the condition of innumerable microstructural arrangements throughout the deformation progression (lee and sun, 2004). as deformation progresses, the domineering stimulus of plastic behaviour of materials undergoing high impact (high strain rate) deformations are fundamentally the slip (dislocation) and twining defective mechanisms. basically, the initial stage of the twinning mechanism requires a larger amount of stress as compared to the implementation of a dislocation slip. in addition, the multiplication of dislocation slip takes place very quickly at high strain rates deformations, from which the materials strengthening effect evolves (lee et al. 2000, gilbert et al. 2006). however, both twinning defects, as well as multiplicative dislocation improve the flow properties of metals, especially the strength (li et al. 2010). scholarly investigations have reported some characteristic properties of waspaloy. investigations have documented the effects of aluminium and titanium content, together with the consequence of solution treatments, including microstructure, strength, and hardness (chang and liu 2001). the hot ductility has been studied (andersson et al. 2013), as well as its tensile strength and creep behaviour (wilshire and scharning 2008), while the density has been estimated to be high, more than 95 % dense (kamran et al. 2008). further studies on waspaloy focused on cyclic fatigue loading (yeom et al. 2002), hardness and grain size (stefan et al. 2010), the effect of gamma prime (pike, 2008), and the electrical and mechanical properties (whelchel et al. 2009). nevertheless, there is a paucity of experiential research reports on the dynamic impact behaviour of waspaloy at extreme deformation conditions. the need to characterize high loadings dynamic impact response harrison okechukwu onovo, david ehigie esezobor, and muideen adebayo bodude 182 and flow behaviour of waspaloy is sensible and implementable through the modified compressive split hopkinson pressure bar technique, which aids in establishing the correlation between the flow properties and the resultant microstructures of waspaloy. the present study, therefore, implements the latter technique to unfold comprehensive dynamic impact deformation and microstructural characteristics of waspaloy at low and elevated temperatures for everyday service applications. this would put forward a wholistic investigational report of scarce waspaloy's dynamic impact response at high strain rate deformation. 2. experimental design 2.1. materials preparation the major elemental composition of waspaloy used in this study includes chromium, cobalt, molybdenum, titanium, iron, aluminium, and carbon, in mass per cent (%) as 19.5, 13.5, 4.3, 3, 2, 1.4, and 0.07, respectively, and was sourced in plate form from prof. ao ojo of the mechanical engineering department, university of manitoba, canada. the cylindrical specimens' dimension in mm is 9.5 by 10.5 (diameter by height) and were smoothly sectioned out with an electric discharge machine (edm) (asala et al. 2019). 2.2. development of npb waspaloy prior to dynamic impact testing, specimens were developed into npb particles through lengthy thermal treatment processes and controlled cooling rate, as illustrated in fig. 1. the key hardening features of waspaloy are the precipitated gamma prime particles, which has its sizes beyond nanoscale in normal wrought condition. the three major conditions that encourage the formation of particles in waspaloy are temperature, soaking period and cooling rate. customarily, exceptional properties of metallic alloys are achieved at the nanoscale, and to tailor npb waspaloy, lengthy thermal treatments (solution treatment, stabilization and age-harden), different soaking time, and various thermal cooling processes ranging from 0.25 500 oc/s was implemented. however, it was found that a cooling rate of 28 30 oc/s effectively yielded the envisaged npb waspaloy. this implies that any deviation in the aforementioned procedures and conditions will set undesirable results in motion. the surfaces of the specimens were smoothened by polishing them with emery paper of 1600 grits in order to accomplish npb waspaloy, while al2o3 liquid (0.6 µm particle size) was used to buff-polished the surfaces to minimize shear and frictional forces during impact deformation. molybdenum disulfide grease was used at the ends of the specimen to lubricate its surfaces and hence reduce effects from frictional forces. post-surface smoothening is the specimen particle size range estimation and resolution during synthesis, which proved difficult with a light optical microscope (lom) together with the scanning electron microscope (sem), as the resulting images were imperceptible. hence, a high-resolution transmission electron microscope (tem) aids in resolving the synthesized nanoparticle sizes. the nanoparticle structure was inspected with the jeol 2010f transmission electron microscope domesticated at the university of manitoba. the microscope has an auxiliary edax ultrathin window detector. the tem samples were ground to thin foils and polished using emery paper from 800 to 1500 grits to about 100 μm thick. coupons 3 mm in diameter were then punched out from the polished thin foils, followed by dimpling and electropolishing processes. twin-jet electro-polisher was adopted for coupons electropolishing in a solution containing alcohol of 90 % methyl and acid of 10% perchloric at the following approximate conditions: temperature (40 oc), voltage (1.5 v) and current (1.5 dc) (asala et al. 2016). 2.3. high strain rate deformation of npb waspaloy high-strained impact tests were conducted at high strain laboratory, mechanical engineering, university of manitoba, using a compressive shpb at wide-ranging high strain rates varied from 4*103 7.5*103 s-1 with deformation temperatures between -180 oc to 750 oc. 3. results and discussion 3.1. evolution of flow stress dynamic impact response of npb waspaloy deformed under wide-ranging strain rates: 4*103 to 7.5*103 (s-1) at diverse deformation temperatures from -180 to 750 oc are presented in figs. 2a through d. the mechanical and flow behaviour of the deformed specimens is, to a large extent, reliant on the deformation parameters such as the rate of straining and temperature of npb waspaloy specimens. the flow stress rapidly upsurges from the inception of dynamic deformation or low strains for each dynamic loading condition. it increases sluggishly at larger strains. the observed surge in flow stress from the onset of 1080 oc solutionization 4 hr 845 oc stabilization 24 hr 760 oc aging 16 hr time (hour) t e m p e ra tu re ( o c ) figure 1. thermal treatment conditions and cooling processes adopted for synthesis of npb waspaloy. dynamic impact response of nanosized precipitates bearing waspaloy 183 impact deformation could be linked to the formation and intensification of dislocation density together with the growth of sub-grain boundaries, which emanate from dynamic recovery and working hardening of materials (wang et al. 2008). beyond the peak stress region, flow stress declines as strain rises in magnitude. flow stress declination occurs when the strain rate hardening effect, in addition to strain, prevails through the thermal softening effect introduced by heat generated during dynamic impact deformation of materials (chiou et al. 2009). fig. 2(a to d) shows noticeable flow stress declination as the strain rate remains steady and the deformation temperature rises. besides, under constant deformation temperature, strain rates energies rise changes in flow stress, which rises, which rises significantly as the strain rate also rises from 4*103 to 7.5*103 (s-1) on the impacted npb waspaloy. these results are in good trend with earlier statements (srinivasan et al. 1993, garcia et al. 1994, lee and sun 2004, zhou and baker 1994, lee et al. 2005, yuan and liu 2005, thomas et al. 2006, wang et al. 2008, lee et al. 2011, lee and kao 2014) that reported the relation of high dynamic impact deformations to different loading conditions of other superalloys. 3.2. 3d mapping of dynamic impact response of npb waspaloy nascent innovative techniques will not elude the implementation of data mapping for unabridged descriptive analysis of the continuous relationship of deformation parameters which include strain rate and deformation temperature. earliest in data mapping of deformation behaviour considered contourable 2d data mapping of deformation parameters focusing on strain rate and deformation temperature, which confirms that strain rate sensitivity parameter varies directly with deformation temperature but inversely to strain rate (garcia et al. 1994). additionally, the first attempt at 3d mapping is a rhombus data matrix conducted at a steady strain rate for deformation temperature, strain and stress (quan et al., 2013). this data mapping is faulted (quan et al. 2016), as stress is analyzed as an input effect of deformation rather than a resultant expressive effect. further, the first apt 3d data mapping was applied to ensuing stress data with varied temperatures, strain rate and strain (quan et al. 2016). this study, however, implements the most suitable 3d figure 2. comparison of dynamic impact response of npb waspaloy as flow stress at constant strain rate: (a) 4*103 s-1, (b) 5*103 s-1, (c) 6*103 s-1, (d) 7.5*103 s-1 under varied deformation temperatures (-180 – 750 oc). (c) (d) (b) (a) harrison okechukwu onovo, david ehigie esezobor, and muideen adebayo bodude 184 mapping of experiential data of waspaloy dynamically deformed to further investigate the unceasing relationship existing among deformation parameters of temperature, flow stress, strain rate, and strain. dissimilar to stress ~ strain relation in 2d, the maps of experiential response data in 3d are demonstrated by diverse colour shades. the 3d data mapping is unceasing and avails the choice of interpreting response data (flow stress) as a block of data (figure 3a) in three fundamental orientations (x-, y-, and z-axis). it is also applicable to any constant deformation parameters that include strain, deformation temperature, and strain rate, as presented in figure 3b. figure 3c presents the deformation behaviour of waspaloy at constant strain but varied deformation temperature together with strain rate, which further validates the region of ultimate flow stress to fall between the strain value of 0.3 and 0.4. similarly, figure 3d represents the deformation characteristics of waspaloy at a fixed strain rate, which obviously reveals that flow stress progresses as the strain rate increases. likewise, at fixed deformation temperature, figure 3e shows that flow stress is maximum at the iciest temperature, denoting a decline in flow stress as deformation temperature upsurges, while figure f attempts to expose the inner portions of the mapped data by showing the flow stress distributions at the initial and middle points of each of the deformation parameters under scrutiny. evidently, the deformation response at different conditions considered is more visually appreciated in comparison to the conventional 2d stress ~ strain relation. (a) (b) (c) (d) figure 3. 3d map of npb waspaloy flow stress distribution at constant (a) no constant parameter (block), (b) strain rate, strain and deformation temperature, (c) strain, (d) strain rate, (e) deformation temperature, and (f) initial and middle point of each deformation parameter. (f) (e) dynamic impact response of nanosized precipitates bearing waspaloy 185 from the foregoing, the interpolative and extrapolative practicability of 3d mappings are not exhaustive yet and cannot be exaggerated as the resultant data for any deformation condition (deformation temperature, strain rate and strain) could be domiciled in the research database to enable and enhance analysis on impact deformation response of waspaloy. the enriched populated data can be converted via different programming languages into finite element modelling (fem) accepted formats and introduced into fem for further simulative analytical approach in different areas of study and hence, curb the rigorous and expensive investigational processes. 3.3. flow stress reliance on strain rate reliance of flow stress on strain rate at the specific strain, in addition to deformation temperatures, is portrayed in figures 4(a, b and c) and 5(a, b and c), which further illustrate the existing relationship amid waspaloy's deformation parameters: deformation temperature, and strain rate sensitivity. evidently, a steady strain rate and strain, increasing the deformation temperature, suggest an attendant decrease in flow stress. invariably, flow stress increases under steady strain and deformation temperature as the strain rate increases, which confirms waspaloy's strain rate sensitivity. from figures 4 and 5, the flow stress at a fixed temperature (-180 oc) and strain rate (4*103 s-1) increased from 1051 to 1221 to 1613 mpa. this corresponds to plastic strains: 0.1, 0.2 and 0.3, respectively. flow stress increase is observed to be consistent as seen across other corresponding deformation temperatures with regard to strain rates under various compressive impact deformation conditions considered. flow stress resulting from strain rate of (4*103 s-1) and strain (0.1) (figure 4) increased from 326 to 1051 mpa as deformation temperature decreased from 750 to -180 oc, respectively. the behaviour is evident in another strain rate (5*103 to 7.5*103 s-1) and strain 0.2 to 0.3 under discourse. consequently, npb waspaloy exposed to high dynamic impact deformation depends to a large extent on the deformation temperature. 3.4. effect of thermal softening the specimen's property, such as thermal conductivity and the rate at which the material is deformed, accounts for the different influences apparent on the plastic behaviour, which is attributed to the adiabatic softening effect on materials. incidentally, the heat generated during low strain rates deformation typically has sufficient time to dissipate or escape into the environment as deformation progresses, thereby causing little or no effect on the specimen's temperature being tested, resulting in isothermal deformation testing. however, high strain rate deformation occurs so fast (within a split second) that there is no ample time for the heat generated during deformation to disperse away, and this paves the way for heat accumulation within the specimen, which invariably affects the behavioural dispositions of the specimen undergoing dynamic impact deformation. the phenomenal testing situation is fundamentally denoted as an adiabatic heating condition. strikingly, the resultant effect of adiabatic heating conditions grossly influences the thermal softening of materials (shi et al. 2014). essentially, the effect of a swift rise in temperature during plastic deformation is generally expressed by equation (1): ∆𝑇 = 𝛽 𝜌 ∗ 𝐶𝑝 ∫ 𝜎 ∗ 𝑑𝜀 𝜀 0 (1) where ∆𝑇, 𝜀, 𝐶𝑝, 𝜌, 𝑎𝑛𝑑 𝜎 are the temperature rise in waspaloy while undergoing plastic deformation, true axial stress, the axial plastic strain, density (g/cm3), and specific heat (j/g.k) respectively. density, ρ is taken to be 8.19 (g/cm3), heat capacity cp is 435 (j/g.k), stress 𝜎 is the flow stress and strain interval 𝑑𝜀, which corresponds to true strain. as the integral function in equation (1) denotes the total plastic work done on the specimen, 𝛽 is a fractional representation of total work from plastic deformation, transformed into heat, with a conventional value of 0.9. the thermal softening effect accounts for the reduction or outright loss of the material's strain or workhardening capabilities (ranc and chrysochoos, 2013). figures 6(a to d) show the disparity of the rise in temperature during deformation regarding the true strain at fixed deformation temperatures and strain rate. temperature rise, ∆t and true strain increase proportionately at a constant strain rate. however, with temperature rise, ∆t has a reverse trend with respect to deformation temperature. in all the loading conditions for direct impact deformation waspaloy in discourse, the maximum thermal softening or temperature rise effect takes place at the iciest temperature (-180 oc) and maximum strain rate (7.5*103 s-1) in the impacted specimen. 3.5. rate of work-hardening flow stress-strain graphs at various impact deformation conditions (figure 2a to d) demonstrate work hardening tendencies. figures (12 to 15) provide the disparity of the work-hardening rate (𝜕𝜎 𝜕𝜀⁄ ) with respect to different deformation temperatures from -180 to 750 (oc) and true strains from 0.1 to 0.5 as well as various strain rates from 4*103 to 7.5*103 (s-1). thus, the rate at which work-hardening occurs is subject to prime agents of plastic deformation mechanisms, which include strain, deformation temperature, and strain (loading) rate. manifestly, the maximum work-hardening effects recorded in this study occur at the iciest deformation temperature (-180 oc), the primary stage of true strain (approximately 0.2) and the maximum strain rate (7.5*103 s-1). the work-hardening rate of waspaloy reduces simultaneously as strain and deformation temperatures rise, as contained in figure 7 (a to d). deformation temperature rise in compressive high harrison okechukwu onovo, david ehigie esezobor, and muideen adebayo bodude 186 strain rate deformation is attributable to adiabatic heating, which occurs when the impact energy utilized in deforming metallic materials is wholly transformed into heat energy, but then, the heat energy generated is somewhat lockup in the metal, as it is not swiftly dissipated to the metal's surroundings, thereby increasing the overall temperature level of the metal. figure 5. dependency of flow stress on deformation temperature at fixed strain rate and varied strain: (a) 0.1, (b) 0.2 and (c) 0.3. figure 4. dependency of flow stress on strain rate at fixed deformation temperature and varied strain: (a) 0.1, (b) 0.2 and (c) 0.3. figure 6. temperature rise, ∆t during plastic deformation of npb waspaloy at constant strain, varied strain rate: (a) 4*103 s-1, (b) 5*103 s-1, (c) 6*103 s-1, (d) 7.5*103 s-1 and different temperature (-180 – 750 oc). (a) (b) (c) (d) dynamic impact response of nanosized precipitates bearing waspaloy 187 conversely, for fixed deformation temperature, and strain, the work-hardening rate perpetually rises concurrently with the strain rate. hence, the thermal softening consequence is a resultant outcome of a rise in temperature during dynamic impact deformation. convincingly, the rate with which work hardening of waspaloy occurs, demonstrates consistent positive effects for every value of true strain (0.1 to 0.5), strain rate (4*103 to 7.5*103 s-1) and deformation temperature (-180 to 750 oc). however, the dynamic work hardening effect is established to dominate thermal softening results on waspaloy specimens studied under stable dynamic plastic deformation at dissimilar conditions: deformation temperatures (low and high) and strain rate (high) loading situations (lee et al. 2011). 3.6. microstructural examination the light optical microscopy and scanning electron microscopy of dynamically impacted npb waspaloy at a high strain rate (6*103 s-1), room (25 oc) temperature is singled out for reasons of conspicuous presence of deformation shear bands which exposed the occurrence of dense dislocation and straight annealing twins' (lamella-like) substructures (plates 1 and 2). the annealing twins present depicts the viability of systematic changes in crystals as a result of cooling effects. moreso, the shear bands being a softening mechanism, leads to an abrupt decrease in flow stress due to inhomogeneous plastic deformation. the during transmission electron microscopic (tem) examination (plate 3 and 4), the diffraction patterns were investigated, which reveal that the waspaloy gamma prime diffraction pattern is on the 001 zone, while the micrographs disclose the range of particle sizes within the nanometer scale. conclusion 1. the cooling rate of 28-30 oc/s effectively plate 1. light optical image of waspaloy impacted at 𝜀ሶ 6 ∗ 103 (s-1) and 25 (oc). annealing twins plate 2. scanning electron image of waspaloy impacted at 𝜀ሶ 6 ∗ 103 (s-1) and 25 (oc). impact deformation shear bands plate 3. tem images of npb particles waspaloy. plate 4. tem images of npb waspaloy with sadp indexing. harrison okechukwu onovo, david ehigie esezobor, and muideen adebayo bodude 188 produced npb waspaloy. 2. nanoscale precipitation hardened waspaloy flow stress improvement at strain rate recommends its suitability for applications where high dynamic impact competence is a design consideration. 3. the declination of flow stress values at elevated temperatures is traceable to grain growth grossly encouraged by the deformation temperature. 4. dynamic work hardening effect influences the outcome of thermal softening under stable dynamic plastic deformation acknowledgements the superalloy samples, technical know-how, laboratories, reagents, equipment and pieces of machinery used in this study were provided by professor olanrewaju a. ojo of the mechanical engineering department, university of manitoba, canada. the authors are grateful to him and his research team. references andersson, j., sjoberg, g. p., viskari, l., & chaturvedi, m. (2013). effects of different solution heat treatments on the hot ductility of superalloys. journal of materials science and technology, 29, 43-53. asala, g., andersson, j., & ojo, o. a. (2016). precipitation behavior of γ′ precipitates in the fusion zone of tig welded ati 718plus®. international journal of advanced manufacturing technology, 87, 2721-2729. asala, g., andersson, j., & ojo, o. a. (2019). a study of the dynamic impact behaviour of in 718 and ati 718plus® superalloys. philosophical magazine, 99(4), 419-437. chang, k-m., & liu, x. (2001). the effect of γ′ content on the mechanical behaviour of the waspaloy alloy system. journal of materials science and engineering a, 308, 1-8. chiou, s-t., tsai, h-l., & lee, w-s. (2009). impact mechanical response and microstructural evolution of ti alloy under various temperatures. journal of materials processing technology, 209(5), 22822294. gilbert, a., wilcox, b. a., & hahn, g. t. (2006). the effect of strain rate on dislocation multiplication in polycrystalline molybdenum. the philosophical magazine: a journal of theoretical experimental and applied physics, 12(117), 649-653. granier, n., & grunenwald, t. (2006). a modified split hopkinson pressure bar for toughness tests. journal of physics iv france, 134, 813-818. hopkinson, b. (1914). a method of measuring the pressure produced in the detonation of high explosives or by the impact of bullets. philosophical transactions of the royal society of london. series a, containing papers of a mathematical or physical character, 213(437), 497-508. kamran, a. m., poonjolai, e., & neil, h. (2008). highdensity selective laser melting of waspaloy. journal of materials processing technology, 195(1-3), 77-87. kolsky, h. (1949). an investigation of the mechanical properties of materials at very high rates of loading. proceedings of the physical society b, 62(11), 676700. lee, w. s., shyu, j. c., & chiou, s. t. (2000). the effect of strain rate on impact response and dislocation substructure of 6061-t6 aluminium alloy. scripta materialia, 42(1), 51-56. lee, w-s., & kao, h-c. (2014). impact deformation and fracture behaviour of cobalt-based haynes 188 superalloy. world academy of science, engineering and technology. international journal of chemical, molecular, nuclear, materials and metallurgical engineering, 8(8), 774-778. lee, w-s., & sun, t-n. (2004). plastic flow behaviour of inconel 690 super alloy under compressive impact loading. materials transactions, 45(7), 2339-2345. lee, w-s., & sun, t-n. (2004). plastic flow behaviour of inconel 690 super alloy under compressive impact loading. materials transactions, 45(7), 2339-2345. lee, w-s., lin, c-f., chen, t-h., & chen, h-w. (2011). dynamic impact response of inconel 718 alloy under low and high temperatures. materials transactions, 52(9), 1734-1740. lee, w-s., liu, c-y., & sun, t-n. (2005). dynamic impact response and microstructural evolution of inconel 690 superalloy at elevated temperatures. international journal of impact engineering, 32, 210-223. li, x., wei, y., lu, l., lu, k., & gao, h. (2010). dislocation nucleation governed softening and maximum strength in nano-twinned metals. nature, 464, 877-880. liu, h-s., yan, b-h., huang, f-y., & qiu, k-h. (2005). a study on the characterization of high nickel alloy micro-holes using micro-edm and their applications. journal of materials processing technology, 169, 418-426. pike, l. m. (2008). development of a fabricable gamma-prime (γ') strengthened superalloy. tms. https://www.tms.org/superalloys/10.7449/2008/su peralloys_2008_191_200.pdf. pollock, t. m., & sammy, t. (2006). nickel-based superalloys for advanced turbine engines: chemistry, microstructure and properties. journal of propulsion and power, 22, 361-374. quan, g-z., lv, w-q., mao, y-p., zhang, y-w., & zhou, j. (2013). prediction of flow stress in a wide temperature range involving phase transformation for as-cast ti–6al–2zr–1mo–1v alloy by artificial neural network. materials and design, 50, 51-61. quan, g-z., zhang, z-h., zhou, y., wang, t., & xia, y-e. (2016). numerical description of hot flow dynamic impact response of nanosized precipitates bearing waspaloy 189 behaviors at ti-6al-2zr-1mo-1v alloy by gasvr and relative applications. materials research, 19(6), 1-17. ranc, n., & chrysochoos, a. (2013). calorimetric consequences of thermal softening in johnsoncook's model. mechanics of materials, 65, 44-55. regazzoni, g., kocks, u. f., & follansbee, p. s. (1987). dislocation kinetics at high strain rates. acta metallurgica, 35(12), 2865-2875. shi, c., lai, j., & chen, x. g. (2014). microstructural evolution and dynamic softening mechanisms of al-zn-mg-cu alloy during hot compressive deformation. materials, 7(1), 244-264. srinivasan, r., ramnarayan, v., deshpande, u., jain, v., & weiss, i. (1993). computer simulation of the forging of fine grain in-718 alloy. metallurgical transactions a, 24(9), 2061-2069. stefan, o., anders, w., & goran, s. (2010). the effect of grain size and hardness of waspaloy on the wear of cemented carbide tools. international journal of advanced manufacturing technology, 50, 907-915. thomas, a., el-wahabi, m., cabrera, j. m., & prado, j. (2006). high temperature deformation of inconel 718. materials processing technology, 177(1-3), 469-472. veerappan, g., ravichandran, m., & marichamy, s. (2018). mechanical properties and machinability of waspaloy for aerospace applications – a review. iop conference series: materials science and engineering, 402, 012039. wang, y., shao, w. z., zhen, l., yang, l., & zhang, x. m. (2008). flow behaviour and microstructures of superalloy 718 during high temperature deformation. materials science and engineering a, 497, 479-486. whelchel, r. l., kelekanjeri, v. s. k. g., & gerhardt, r. a. (2009). mechanical and electrical characterization in age-hardened waspaloy microstructures. international heat treatment and surface engineering, 3(1-2), 35-39. wilshire, b., & scharning, p. j. (2008). theoretical and practical approaches to creep of waspaloy. journal of materials science and technology, 25(2), 242248. xue, l. (2018). chapter 16 laser consolidation—a rapid manufacturing process for making net-shape functional components. in j. lawrence (ed.), advances in laser materials processing (pp. 461505). woodhead publishing. yeom, j. t., williams, s. j., & park, n. k. (2002). low-cycle fatigue life prediction for waspaloy. journal of materials at high temperatures, 19(3), 153-161. yuan, h., & liu, w. c. (2005). effect of the δ phase on the hot deformation behaviour of inconel 718. materials science and engineering a, 408(1-2), 281-289. zhou, l. x., & baker, t. n. (1994). effects of strain rate and temperature on deformation behaviour of in 718 during high temperature deformation. materials science and engineering a, 177(1-2), 19. mech050306.qxd the journal of engineering research vol. 5, no.1, (2008) 55-61 1. introduction the traditional method that the omanis has used in the past for producing the omani sarooj can be summarized in the following steps (hago and al-rawas, 1999): 1. the soil used in producing sarooj is usually brought from farms and lands used for agricultural purposes, and perhaps other specific bare lands. however, it should be noted that the soil used differs in its clay content, and chemical and mineralogical composition from one area to another. this soil is sieved in order to remove gravels and other undesirable materials. __________________________________________ *corresponding author’s e-mail: ahago@squ.edu.om 2. the sieved soil is mixed thoroughly with water to form "ghailah" (soil paste) which should be left for one day. this step is repeated for two consecutive days and finally circular disks of soil are made and left to dry. the process of drying the circular disks may take a week. 3. dry logs of date trees are arranged in three layers. small openings are left between the logs of trees to allow for air circulation. dry leaves are placed between logs of dates trees to help in the burning process. this arrangement, shown in fig. 1, is known as "mahabba". 4. the prepared soil disks are packed on the top of the third layer of the dates trees. they can reach a height of one meter. design of a rotary kiln for production of sarooj a. w. hago* and a. a. al-rawas * department of civil and architectural engineering, college of engineering, sultan qaboos university, p.o. box 33, al-khod 123, sultanate of oman received 3 march 2006; accepted 1 april 2007 abstract: sarooj is a local term used for pozzolana. it has been used extensively as a hydraulic binder in buildings, hydraulic installations, forts and castles, and in the renovation of historical monuments. it is produced by burning specific clay soil that possesses adequate quantities of silica, alumina and iron oxides. in the past, the material was produced using a large amount of energy most of which is wasted by blowing winds. the burning system was usually thermally inefficient, and the output is rarely checked for quality. test on products produced by using the traditional method may yield poor quality sarooj due to inadequate burning or over burning or even insufficient grinding in the end. in this research, extensive experimentation has been performed on the method of production of sarooj to determine the factors influencing sarooj quality. based on these factors, a new method of production is proposed. in this method, a fully mobile kiln was designed to produce sarooj on the site. the kiln can produce 200 kg/h of sarooj, and can be installed within one hour of arriving on the site. it can be towed by a tractor or a normal vehicle to the desired location, connected to the natural gas supply and the electric power. the construction is light, compact, and easy to start and shut down. in this paper, the plant is described in detail, together with some preliminary results of testing done on sarooj produced by the kiln. full details of tests on the product of the kiln will be given in a subsequent paper. keywords: sarooj, pozzolana, hydraulic binder, mobile plant, building materials êhhqé°üdg êééàf’ qgho ¿ôa º«ª°üj ¢sghôdg »∏y ôeéy h ƒég ~mgƒdg ~ñy áá°°uuóóÿÿgg.á«îjqéàdg qék’g º«eôj ‘h ´ó≤dg h ¿ƒ°ü◊g h ᫵«dhq~«¡dg ägõ«¡éàdg h äéjéæñdg ‘ »µdhq~«g â檰scéc ™°sgh ¥é£f ≈∏y πª©à°ùj ¿éc ¿’hrƒñ∏d »∏fi º°sg êhqé°üdg : ωg~îà°séh êhqé°üdg èàæj ¿éc »°véÿg ‘ .ωƒ«æÿc’g ~«°ùchcgh ~j~◊g ~«°ùchcg h 鵫∏«°ùdg øe á«aéc ä髪c ≈∏y …ƒà– »àdg á«æ«£dg áhîdg øe áæ«©e ´gƒfcg ¥gômeéh êhqé°üdg èàæjh äé°uƒëa …c’ œéædg ¢vô©j ée gqoéfh ,ioéy éjqgôm a∞c òz πª©à°ùÿg ¥ô◊g ωé¶f ¿éc éªc .≥∏£dg agƒ¡dg ‘ áhîdg ¥ôm áé«àf ìéjôdg π©øh é¡ñ∏zcg ™«° j ábé£dg øe iòñc ä髪c øµj ⁄ å«m ,álé◊g øy ~fgõdg ¥gîmc’g öñ°ùh hg ‘éµdg òz ¥gîm’g öñ°ùh áäjoq á«yƒf gp ¿ƒµj ~b áj~«∏≤àdg á≤jô£dg ∫骩à°séh èàæÿg êhqé°üdg ¿g éªc .á«yƒædg øe ~céà∏d √òg ≈∏y oéæà°se’éhh .êhqé°üdg á«yƒf ≈∏y ôkdƒj »àdg πegƒ©dg ôjô≤àd êhqé°üdg êéàfeg á≤jôw ≈∏y á∏eé°t á°sgqo agôleg ” ,åëñdg gòg ‘h .¥ô◊g álqo øe ~céà∏d á∏«°sh ∑éæg èàæj ¿cg øµá ¿ôødgh .™bƒÿg ‘ ¬dƒ°uh øe i~mgh áyé°s ∫ón ¬ñ«côj øµá êhqé°üdg êéàfe’ π≤æàe qgho ¿ôa º«ª°üàh ∂dph ,êhqé°üdg êéàf’ i~j~l á≤jôw ìgîbg ” ,πegƒ©dg .á«féhô¡µdg ábé£dg h »©«ñ£dg ré¨dg ägõ«¡éàh π°uƒj ºk ,üƒ∏£ÿg ™bƒÿg ¤g öë°s áhôy hcg qgôl ᣰsgƒh ¬ñë°s øµá éªc ,áyé°ùdg ‘ êhqé°üdg øe éegôz gƒ∏«c »àfée ‹gƒm ägqéñàneód áj~«¡ªàdg èféàædg ¢ ©ñd é°vôy º° j éªc ,π«°üøàdéh ¿ôødg º«ª°üj ¢vôy ” ,ábqƒdg √òg ‘h .á∏«¨°ûj ádƒ¡°sh ¬jéfƒµe áwé°ùhh ¿rƒdg áøîh ìî≤ÿg ¿ôødg réàáh .á≤m’ ábqh ‘ ¿ôødg èàæe ≈∏y ägqéñàneód á∏eéµdg π«°uéøàdg ô°ûf ºà«°sh ,ìî≤ÿg ¿ôødg ᣰsgƒh èàæÿg êhqé°üdg ≈∏y âjôlg »àdg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .á«fé°ûfeg oggƒe ,∫é≤f ¿ôa ,»µ«dhq~«g â檰sg ,¿’hrƒh ,êhqé°u : 56 the journal of engineering research vol. 5, no.1, (2008) 55-61 5. a layer of white wadi limestones is placed between the soil disks for the production of "al-nourah" which will be mixed with the soil during the burning process and gives it its white colour. 6. after completing all the preparation of the mahabba, the burning process starts. copious fumes are produced in the process (fig. 2) and with time they disappear. the mahabba is to be visited after four days from burning, no traces of fire could be seen, but the burned soil disks will still be hot. the logs of the date trees will be completely destroyed by the fire, turning them into ashes. water is then sprayed over the mahabba, in order to extinguish any fire that could not be detected. 7. the sarooj must be left at the site for a period of about two months, before it can used. it is believed by sarooj experts who worked with the traditional method that this period would give the sarooj more strength. after that, sarooj is ground into a powder form. in this method of sarooj production, temperature recording is not possible and the period of calcination cannot be observed or controlled. the calcined product is pulverised through two operations: primary and secondary. the primary operation is done manually, using tamping rods and shovels which involve a considerable amount of effort and time, while the secondary operation is done mechanically using compacting rollers which are hand driven, see fig. 3. the end product is screened to get the required fineness. although the capital cost of production is not usually high, high thermal efficiency cannot be obtained due to the small size of the work and the intermittent operation. the proposed design should consider these aspects. the calcination process involves baking the raw material in a heat transfer apparatus like a kiln. the feed for a kiln is required to be of a particular size, shape and density according to the likely breaking down of the feed during its travel through the kiln. a variety of fuels can be used to obtain the required calories of heat per kilogram of charge to be calcined. however, in the process of the design of the kiln, the following aspects should be considered: 1. feed size, shape, density and moisture content; 2. hoisting and charging arrangement; 3. thermal efficiency in terms of tons of sarooj produced per ton of fuel; 4. control of temperature and duration of burning; 5. operational efficiency. in order to meet a local small demand, small capacity plants have been developed, using different heaters as vertical shafts furnaces or rotary kilns of fluidized beds (thate and patel 1978; gupta and rao, 1978). in all of these kilns, the installations are fixed in position, which increases the cost of production for a small intermittent plant operation. to reduce the cost, a semi mobile flash/dryer calciner unit was proposed by (salvador and pons 2000). still the size of their plant was too big for small demand production. the overall size of their kiln was 14.5 m long, 3 m wide and 4.5 m high weighing 23 tons. clearly such a plant cannot be fully mobile. 2. a better design the review of the existing method given above indicated a need for a new design which will incorporate the following improvements: figure 1. sarooj mahabba figure 2. sarooj burning figure 3. grinding of sarooj 57 the journal of engineering research vol. 5, no.1, (2008) 55-61 1. hoisting and emptying of the feed into the kiln can be mechanized; 2. easy firing and efficient use of energy will be provided by the use of natural gas burners; 3. kiln shell should be structurally strong and be designed for temperature and hoop stresses; 4. in the shell, a non-conductor lining should be designed to withstand the high temperature of the kiln; 6. reduction of heat losses so as to reduce cost of production; 7. good utilization of space of the kiln as reusable. the traditional method abandons the area where the mahabba is established until sarooj is removed. this could take more than two months to allow reuse of the place occupied by one firing operation; 8. the continuous calcination temperature recording through thermocouples can help control the movement of the calcination zone; 9. easy handling and movement of the kiln. specifically, the designed kiln should meet the following specifications: 1. operating temperature = 750oc; 2. maximum temperature not to exceed 1000oc; 3. temperature on the outside surface of the kiln not to exceed 50 oc; 4. rate of feed to the kiln = 200 kg/h; 5. natural gas pressure = 3 bars; 6. volume of air for combustion = 16 m3/h; 7. speed of rotation of kiln = 30 rpm. 3. raw material processing and handling the excavated material from the clay bed comes in the form of fine soil varying in size from 20 mm to fine powder. the material can be hard desiccated clays and shale, or it can be granular soft clays. two methods can be used for preparing the material for burning in the kiln, depending on the hardness of the raw material: 3.1 hard desiccated clays and shales desiccated clays and shales of sizes ranging between 5 mm to 10 mm can be fed directly to the kiln without processing. these materials are usually stiff and unsolvable in water. they can withstand rolling in the kiln without disintegration. particles larger than 10 mm have to be separated by sieving and broken to a size roughly less than 5 mm or can be excluded all-together if it was in small quantities. 3.2 soft granular clays soluble in water such soils cannot withstand the rolling that takes place inside the kiln and can easily disintegrate into powder. once disintegrated, they can stick to the sides of the kiln and hinder the easy flow of the material. they can also create lumps in the kiln which require more time for complete burning. in order to overcome these effects, such soils have to be mixed with about 12% water and molded in a form of balls or pellets before putting in the kiln, to facilitate their movement inside the kiln. the pellets are obtained by using a pelletizer designed specifically for the purpose, see fig. 4. the pelletizer is composed of a cylindrical drum 600 mm in diameter and 2 m long, rotated about its longitudinal axis which is supported at its ends by roller bearings. it is rotated at a speed of 20 rpm by an electric motor. the moistened soil is introduced into the pelletizer drum through a side opening. after the soil is loaded the opening is closed with a cover, and the motor is started to set the drum in rotation. in three minutes, pellets start to form. the size of the pellets is controlled by changing the moisture content of the soil. the moisture content of the soil can be increased by spraying water through end openings in the drum. when all of the soil has changed into pellets, the motor is stopped, and the palletized soil is extracted from the drum through the side opening. the pellets are then spread in the sun and left to dry. after drying, they are ready to be loaded in the kiln. 3.3 loading of the material into the kiln the feed to the kiln, whether processed or unprocessed material can be loaded into the kiln by using a manual procedure, or by using a hoist. the hoist, fig. 5, consists of a steel bucket 2 mm thick having dimensions 30x30x40 cm and can hold 40 kg of soil. the bucket rolls over a pair of mild steel channels 160 x 80 mm, each made by welding 2 angles 80x80 mm. the supporting rails have an inclination of 80o with the horizontal, and are tied to the frame supporting the kiln. a pair of guide wheels attached to the bucket roll inside the channels. the bucket is pulled along the rails by means of a 5 mm diameter wire rope tied near the bottom on either side. the wire rope is passed over a series of pulleys and is pulled by means of a manually operated winch. the bucket rolls over the rails till they curve downwards at the top of the kiln. the guide rollers stop at the end of the rails and the winch winds up some more length of the wire rope to haul up the bottom of the bucket and turn, it upside down in the chute. after emptying the bucket, the winch pulls out the rope which puts the bucket down on rails again right side up. the empty bucket slides along the rails to the bottom of the lifting rails where it is filled again. figure 4. pelletizer 58 the journal of engineering research vol. 5, no.1, (2008) 55-61 4. the kiln depending upon the heat transfer unit selected the cost of calcinations forms 40 to 60 % of the total cost of manufacture of pozzolana. a mobile cylindrical double-coned rotary kiln was proposed for this design. various alternative shapes could be have been adopted for the kiln, but the selected shape was favored for several reasons viz. minimum wear and tear of lining due to sliding of charge; minimum choking and clinkering; more uniform burning and a better distribution and flow of the charge through the kiln. the kiln was designed to give the required optimum temperature (750oc) and a period of calcination of two hours. the designed kiln was built from 6mm steel plates of, and consisted of three parts: a central cylindrical drum 1200 mm external diameter with a length of 15 mm, and two truncated cones at the ends, see fig. 6. each cone was tapered from a diameter of 1200 mm at the mating end with a central cylindrical portion to 300 mm at the free end. the overall length of the kiln was 3170 mm. the fire brick used in the lining of the kiln was manufactured in the laboratories of the college of engineering at sultan qaboos university. a fire clay containing a high percentage of alumina was used for this purpose. special molds constructed from steel plates were manufactured for casting the fire bricks. fire clay for making the bricks was imported from pakistan and was supplied by skylark company for building materials in sharjah. a hydraulic compression machine was used for pressing the clay in the moulds. the mortar used in building the fire brick inside the kiln was made of high alumina cement with sand. the mortar courses were only about 5 mm thick, so as to reduce thermal volume changes in the mortar when subjected to a high heat input/discharge. the thickness of the brick lining in the kiln was 170 mm in the cylindrical portion where the fire was concentrated, and 120 mm in the conical portion, fig. 7. the kiln was built over a trailer made of steel rolled sections with four wheels to facilitate movement both during manufacturing and operation stages. it can be dismantled, moved over land (road network) and sea, and assembled within few hours. it rests on four roller supports fixed over the trailer. the four rollers allow free rotation of the kiln. the kiln rotates about its longitudinal axis with a speed of 30 rpm. the rotation is provided by a 10 hp electric motor through a steel shaft, two pulleys and three belts, fig. 8. when dragged in place, the trailer was supported by four jacks on the ground to anchor the wheels against rotation. because the front two jacks were shorter than the rear, the overall body of the kiln became tilted about 1 to 2 degrees with the horizontal to facilitate withdrawal of the product at the end. figure 5. the hoist figure 6. parts of the rotary kiln figure 7. lining the kiln with fire bricks figure 8. kiln mounted on a trailer 59 the journal of engineering research vol. 5, no.1, (2008) 55-61 one gas burner was used to provide the required heat energy. the burner was bp7 premix type manufactured by australian combustion company. the burner uses natural gas at a pressure of 3 bars and with a rate of 16 m3/hr. it includes an automatic ignition of the flame and flame safety detection. the detection system utilizes the flame rectification principle. for complete combustion, an air blower was needed to supply the burner with air. air and gas are mixed in the burner at a well calculated ratio by the mixer itself, depending on the temperature of the burner. a small processor unit controls the operation of the burner, and the whole process is automatic. assuming a calorific value of natural gas of 38 mj/m3, this means that heat is input to the system at a rate of 608 mj/hr. attempts to estimate the necessary heat requirement to produce sarooj from clayey soils by field experimentation gave an estimate of 5579 mj/ton of soil (hago, et al. 1995 and hago and al-rawas, 1997). thus for a given size of feed, the duration of burning can approximately be determined. a batch of 0.2 ton, needs 2 hours calcination time. figure 9 shows the burner and blower mounted in the kiln. the operation of filling the kiln will be done in batches of five buckets each. the volume of the calcination zone was designed to hold one batch's production. temperature in the kiln was monitored by means of thermocouple inserted through the kiln drum. the arrangement enables the monitor to read the temperature on the display of the burner control unit. the shell of the kiln is designed to serve three purposes. viz., withstand the calcination temperature and structural stresses as well as minimize heat losses. inside, a 170 mm thick fire brick lining made of fire claim was used. the desired calcination temperature was 750oc. however, structural components were designed to take up a maximum temperature of 1000oc which may happen due to accidental rise in temperature. the temperature on the outside of the outer shell will be slightly higher than the ambient temperature. the thickness of each component of the kiln was selected to minimize the heat loss through conduction as much as possible. most fire bricks withstand a temperature of 1500oc. 5. structural design of the kiln and its supports the design was carried out on lines of that adopted for steel structures. it was checked for combined stresses due to internal load and temperature. tensile stress in steel (hoop tension) was assumed at 460 n/mm2; compressive stress in brickwork was assumed at 14 n/mm2, with cement-lime mortar grade iv according to bs 5628-1, 1992. unit weight of charge of loose clay soil was taken as 14 kn/m3. the coefficient of thermal expansion of brickwork and steel was assumed to be 6x10-6 and 11x10-6 per degree centigrade respectively. the total downward load includes the dead weight of the kiln shell plus the estimated soil charge plus the weight of the supporting frame itself. due to the small size of the installation, wind load was neglected in the design of the supporting structure. safety factors on all loads were taken as 1.4. full detailed set of drawings for the construction the kiln is given in figs. 10(a) and (b). 6. pulverising after allowed to cool in natural air, the calcined product drawn out from the kiln can be carried in 2-wheel trolleys straight to the pulverizer. the crusher used for this purpose is a disc mill used for batch or continuous fine grinding of soft to hard material (hardness grades up to 8 mohs), having a capacity of 150 kg/hr. it can be adjusted to any desirable size of the final product. 7. sample results of tests on the kiln product clay samples were produced from 3 sites in oman and were burnt in the kiln to produce sarooj. mixes were produced from sarooj and lime and were designated as fanja (k), al-fulaij (k) and bahla (k). for comparison, the soil from the same sites was used to produce sarooj in the laboratory and mixes produced in this way were designated as fanja (l), al-fulaij (l) and bahla (l). results of the initial setting, slump, consistency, density, compressive strength tests on the two series are compared in table 1. the initial setting times of the laboratory-burnt and kilnburnt samples for each site show differences of 1.5, 3.0 and 5.5.0 hours for fanja, bahla, and al-fulaij samples, respectively. the initial setting time was maximum (11 hours) for fanja (k) and al-fulaij (k) samples, and minimum (3.5 hours) for bahla (l) sample. the slump values for the laboratory-bunrt and kiln-burnt samples of all sites are similar and consistent. bahla (k) sample has the highest slump value (55 mm) whereas fanja (l) sample has the lowest value (14 mm). al-fulaij samples show intermediate slump values (25-28 mm). table 1 also shows that the consistency values of fanja figure 9. burner and blower mounted in the kiln 60 the journal of engineering research vol. 5, no.1, (2008) 55-61 and al-fulaij samples are comparable. the density values showed slight differences between the laboratory-burnt and kiln-burnt samples for all three clays. al-fulaij and bahla showed a difference in density of only 1.8% while fanja samples showed a difference in density of only 1.1%. results of the compressive strength showed an agreement between the laboratory-burnt and kiln-burnt samples for fanja and al-fulaij clays. however, notable differences were recorded for bahla samples. fanja (l) sample attained the highest strength of 26.50 mpa at 90 days whereas bahla (k) sample showed the lowest strength of 4.82 mpa for the same duration. al-fulaij (l) and al-fulaij (k) samples exhibited strength of 11.98 and 11.82 mpa, respectivley. all samples showed development of strength with time. compressive strength and consistency results showed a similar trend. 8. conclusions based on the research findings, a new method of production of sarooj was proposed. a fully mobile kiln was designed to produce sarooj on the site. the kiln can produce 200 kg/h of sarooj, and can be installed within one hour of arriving to the site. it can be towed by a tractor or a normal vehicle to the desired location, connected to the natural gas supply and the electric power. the construction is light, compact and easy to start and shut down. preliminary tests showed that the designed kiln could compressive strength, mpa sample designation temp./ duration, °c/hour initial setting time, hours slump, mm consistency, % density, kg/m3 7 days 14 days 28 days 90 days fanja (l) 740/2 9.5 14 31 1985 5.04 11.11 17.63 26.50 fanja (k) 740/2 11 15 30 1964 4.39 10.37 18.36 23.80 al-fulaij (l) 740/2 5.5 28 35 1989 5.52 6.64 8.63 11.98 al-fulaij (k) 740/2 11 25 33 1952 4.07 5.46 7.58 11.82 bahla (l) 740/2 3.5 53 37 1990 5.84 6.29 7.33 8.79 bahla (k) 740/2 6.5 55 46 1953 1.83 2.54 3.05 4.82 table 1. results of the engineering tests of sarooj fig.10a: details of the mobile rotary kiln for sarooj production, drawing 1/2 elevation left view figure 10a. details of the mobile rotary kiln for sarooj production, drawing 1/2 61 the journal of engineering research vol. 5, no.1, (2008) 55-61 produce sarooj with properties comparable to those of sarooj produced under controlled conditions in the laboratory. references bs5628-1, 1992, code of practice for use of masonry, british standards institution. gupta, j.s. and rao, a.v.r., 1978, "production of clay pozzolana by fluidized bed technique,". trans indian ceramic society, vol. 37, pp. 165-171. hago, a.w., al-rawas, a.a. and al-harthy, a.s., 1995, "the omani sarooj," department of civil engineering, sultan qaboos university, sultanate of oman, unpublished report, vol. 74, pp. 1-74. hago, a.w. and al-rawas, a.a., 1997, "properties of the omani sarooj," engineering journal of the university of qatar, qatar, vol. 10, pp. 81-91. hago, a.w. and al-rawas, a.a., 1999, "an investigation into the traditional method of production of omani sarooj," j. of science and technology, sultan qaboos university, sultanate of oman, vol. 4, pp. 3543. salvador, s. and pons o., 2000, "a semi-mobile flash dryer/calciner unit to manufacture pozzolana from raw clay soils-application to soil stabilization," construction and building materials, vol. 14, pp. 109-117. thatte, c.d. and patel, j.k. 1978, "a new design of plant for the manufacturing of pozzolana," indian highways journal, pp. 14-24. fig. 10b details of the mobile rotary kiln for sarooj production, drawing no. 2/2figure 10b. details of the mobile rotary kiln for sarooj production, drawing 2/2 ece070512.qxd the journal of engineering research vol. 6, no.1, (2009) 46-50 1. introduction the attractive features of fiber optic sensors such as lack of sparking and immunity to electromagnetic fields has given it a great advantage over electronic sensors in many applications, including civil engineering, oil, gas and power generation industries (abad, et al. 2003). in addition to its usefulness in embedded sensors applications and the inherited attractive features of optical fiber sensors, fiber bragg gratings (fbgs) are driven by their performance accuracy and versatility of the utilized technology (agarwal, 2001 and bourdoucen and al-lawati, 2006). fiber optics communication signal carriers comprise about 95 percent of worldwide transmission capacity (chan, et al. 1999). the introduction of the well known wavelength division multiplexing (wdm) and later the dense wdm techniques gave the fiber optic communica______________________________________ *corresponding authors e-mail: lawati@squ.edu.om tions such a leading edge in providing astronomical bandwidth capabilities of more than 10 terabits per second on a single optical fiber. in fact, the japanese nippon telegraph and telephones corporation has recently announced on its website reaching a data transmission rate of 14 tbps over a 160 km optical fiber. this attractive feature of wdm makes it popular for applications other than communications such as multiplexing many fbg sensors simultaneously on a single fiber. such applications were demonstrated in numerous published articles, eg. (abad, et al. 2003 and fan, et al. 2004). other multiplexing techniques to interrogate fbg sensors were also reported eg. (gebremichael, et al. 2005 and o-eland.com/sensorproducts/sensor_strain.html). in this paper, a study of the effect of integrating an fbg sensor was conducted in a four wavelength wdm communication system operating at 1550 nm region with 1 nm spacing. the operating wavelengths are shown in fig. 1 below. four wavelengths are used to study the mutual integration of fbg strain sensors in wdm networks, effects on quality factor ali al-lawati* and hadj bourdoucen department of electrical and computer engineering, college of engineering, sultan qaboos university, p.o. box 33, p.c. 123, al-khodh, muscat, sultanate of oman received 12 may 2007; accepted 3 october 2007 abstract: a study of the effect of integrating an fbg sensor in a four wavelength wdm communications system operating at 1550 nm is presented. the simulations considered focus on the mutual effects of both the sensing and the communications systems. the effect of power levels of the interrogating optical source on the performance of the two systems is also investigated under excitation levels of up to 10 dbm. the network layout used in the simulations is based on an actual optical link in oman having a variety of spans. the results obtained at data rates of 2.5 and 10 gbps with variable strains up to ±600 µs show a good tolerance in terms of quality of transmission for the two systems. however, the greater the strain values, the more noticeable are the degradations of transmission quality parameters of the communications system. keywords: fbg, fiber optic sensor, wdm, quality factor ägô©°ûàe èeo fbg á«fƒ° dg ä’é°üj’g äéµñ°t ™ewdmioƒ÷g πeé©e ≈∏y òkéàdg h ø°shoqƒh êém h *»jgƒ∏dg ≈∏y ¬¬°°uuóóÿÿgg~æy πª©jh ¬«lƒe ∫gƒwg á©hqcg ω~îà°ùj »àdgh ≈lƒÿg ∫ƒ£dg io~©àe ä’é°üj’g º¶f ™e áµhé°ûàÿg ¬jô°üñdg ±é«dg ägô©°ûà°ùe èeo òkéàd á°sgqo åëñdg gòg ω~≤j : ≈àm úeé¶ædg agog ≈∏y »fƒ° dg q~°üÿg iq~b iƒà°ùe òkéj øe ≥≤ëàdg ” ∂dòc .ä’é°üj’g º¶fh qé©°ûà°s’g øe πµd ¬doéñàÿg ägòkéàdg ≈∏y iécéùg õcôjh .îe ƒféf 1550 äô¡xg ~bh .i~ÿg ¬yƒæàe ¬yƒª› ≈∏y …ƒà– »àdgh ¿éªy ‘ ájô°üh ±é«dg áµñ°t ≈∏y iécéùg ‘ áe~îà°ùÿg áµñ°ûdg §«£îj ~æà°ùjh . äg~mƒdg 10 ¤g π°üj iqékeg iƒà°ùe ∂dp ™eh .úeé¶ædg óµd ∫é°sq’g ¬«yƒf å«m øe g~«l áfhôe øjgî°s hôµj ée 600 ¤g π°üj ∫é©øfg ò¨àe ™e 10 h 2^5 øe äéfé«ñdg ä’~©e ‘ â≤≤– »àdg èféàædg .ä’é°üj’g ωé¶æd ∫é°sq’g ioƒl ‘ ¢ü≤f ¤g …odƒj ∫é©øf’g ᪫b ioéjr ¿g ßmó«a áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .√oƒ÷g πeé©e ,»lƒÿg ∫ƒ£dg io~©àe ,á«fƒ° dg ±é«d’g ägô©°ûà°ùe ,áµhé°ûàÿg ájô°üñdg êgôh ±é«dg ägô©°ûà°ùe : 47 the journal of engineering research vol. 6, no.1, (2009) 46-50 effects of sensing and communications systems. these four wavelengths or channels are referred to hereafter as λ1, λ2, λ3 and λ4 corresponding to wavelength of 1548.5, 1549.5, 1550.5, 1551.5 nm respectively. the simulations considered focus on the mutual effects on both the sensing and the communications systems. the effect of power levels of the interrogating optical source on the performance of the systems is also investigated under excitation levels of up to 10 dbm. the network layout used in the simulations is based on an actual optical link in oman having a variety of spans operating at 2.5 and 10 gbps. the strain sensor is located 35 km away from the nearest optical network access point that consists of an optical add and drop multiplexer (oadm). 2. theory and system description the fbg sensor exhibits a shift of ∆λg in the reflected bragg wavelength from the central wavelength λg due to change in temperature and strain. ∆λg can be found using the following relation (kang, et al. 1998): (1) where, δl,δn, δ∆ and δt are the changes in the fiber, length, index of refraction, optical grating period and temperature respectively. assuming that the temperature effect is negligible, the bragg wavelength shift due to a homogenous and isotropic strain on the grating can be expressed as (lim, et al. 2001; madhav and asokan, 2004): (2) where, pe is the effective elastooptic coefficient of the fiber and ε is the applied strain. pe can be found using the following equation: (3) where µ is the poisson's ratio and pij are fbg fiber strain optic tensors. the applied strain will cause a shift in the bragg wavelength, λg or λ2 (with 99% reflectivity), towards the neighboring wdm channels λ1 and λ3. this in turn, will cause degradation in the communications system mainly to the immediate channels surrounding λ2. since the fiber dispersion curve is not perfectly flat and is a function of the wavelength, the effect on the neighboring channels will not be identical and therefore both surrounding channels need to be examined. in order to study this effect, the optical network shown in fig. 2 was used. the sensing fbg wavelength is detected 35 km away from samail city. an optical adm is implemented to add or drop λ2 to the wdm link. the interrogating signal of the fbg is sent from muscat and the response measured at muscat as well. the fbg sensor's dynamic range assumed in the simulations is typical and is also commercially available (peng, et al. 2002). each link on the network is constructed from an smf and a dcf fibers to compensate the chromatic dispersion at 1550 nm. the signal attenuation in both fibers is compensated using optical amplifiers. 3. results and discussion the system layout shown in fig. 2, was simulated based on the well known schrödinger. the equation was solved using the split step fourier method. the principle of the split-step fourier method (ssfm) technique is based on the fact that the propagation of a pulse over the full length of optical fiber is considered by dividing the total length of the fiber into small segments in such a way that changes in the envelopes of optical signals can be considered sufficiently small. within these segments, the linear and the nonlinear operators in the schrödinger equation can be considered to act independently of each other. hence, the effect of propagation along the fiber segment is determined by first using nonlinear operator followed by the linear operator. in this way, the accuracy and efficiency of operator splitting techniques depend on the way discretization is done in time and spatial domains (sinkin, et al. 2003). as for the parameters of simulation and simplifications used, these are based on the fact that the strain is homogenous and isotopic. the factor pe in eq. (2), which is a function of neff, pij and π, is having a numerical value of approximately 0.22 and hence this equation simplifies to (1-pe) ε = 0.78 ε. this approximation, well known in the literature was used in all simulations. the q-factor is then derived from the solution of schrodinger equation. for a non-linear dispersive and lossy single mode and dispersion compensated fibers used in the simulations, the evolution of the profile of a slowly varying field is obtained by solving the schrödinger equation given by: 1547 1548 1549 1550 1551 1552 1553 -40 -35 -30 -25 -20 -15 -10 -5 0 w a ve length (nm) p o w e r (d b m ) fbg wavelength changes figure 1. spectra of the four dm sources used in simulation g n n 2 n l 2 n t l l t t λ λ ∆λ λ ∆ λ ∆ ∂ ∂ ∂ ∂⎛ ⎞ ⎛ ⎞ = + + +⎜ ⎟ ⎜ ⎟∂ ∂ ∂ ∂⎝ ⎠ ⎝ ⎠ ∆λg = λg.(1-pe)ε pe= (n 2 eff/2)[ p12 µ (p11 + p12)] 48 the journal of engineering research vol. 6, no.1, (2009) 46-50 (4) where, a is the amplitude of the varying electric field envelope, β1 the inverse group velocity, β2 and β3 are the first and the second order group velocity dispersion, a is the absorption coefficient, and γ is the nonlinear coefficient of the considered fiber. these values are taken for standard smf and dcf fibers at a center wavelength of 1550 nm. note that the q factor is approximated using mean values and standard deviations of the signal samples as indicated by relation: (5) where m1, m0, σ1, σ0 are the mean values and standard deviations of the signal samples when a "1" or a "0" is received respectively." in addition to these parameters, the simulations have been performed for wdm sources having constant power of 0 dbm, whereas the interrogating source used for the sensor was varied from 0 dbm to 10 dbm. two different bit rates were used to have a more realistic scenario transmitting at 2.5 gbps and 10 gbps. figures 3 to 6 illustrate the quality factor profiles as a function of strain for each wavelength. figure 3, shows a flat profile of the q-factor of the fbg sensor for the full range of strains considered (ie. -600 µs to +600 µs) which is an indication of a good performance. this flattened behavior of the q-factor seen for different power levels of the interrogating source indicates also a good tolerance to power changes. these two observations imply that integrating the sensing system for the above range of strains within a wdm communication system is viable. however, the effects on the communications system need to be observed by examining figs. 4, 5 and 6 corresponding to λ1, λ3 and λ4 respectively. the upper curve of fig. 4, quasi insensitive to strain adm bahla muscat izki samail ghala fbg sensing device 25 km 84 km 37 km 48 km 35 km figure 2. physical communication system layout connecting muscat to bahla used in simulation to study the effect of fbg sensors on communication system 2 3 23 1 2 2 3 a a i a a a a i a a z t 2 6 2t t β β β γ ∂ ∂ ∂ ∂ = − × − × × + × − + × × ∂ ∂ ∂ ∂ 1 0 1 0 m m q σ σ − = + -600 -400 -200 0 200 400 600 5 6 7 8 9 10 11 strain (us) q u a lit y f a ct or ( d b ) figure 3. quality factor q, in db, for tran smission performance of fbg ( channel 2) operating at a data rate of 10 g bps for different input source power ,, €€: 0 dbm, o: 3 dbm, ◊◊: 6 dbm, ∇∇ : 10 dbm, -600 -400 -200 0 200 400 600 6 8 10 12 14 16 18 20 22 24 26 28 strain (us) q u la tit y f a ct o r q ( d b ) figure 4. quality factor q, in db, for transmission performance of channel 1 (shortest wavelength on fig. 1) operating at a data rate of 2.5 gbps. οο: system not coupled to sensing system, €€: 0 dbm, ××: 3 dbm, ◊◊: 6 dbm, ∇ : 10 dbm, 49 the journal of engineering research vol. 6, no.1, (2009) 46-50 changes, corresponds to the q-factor of λ1 prior to systems integration. this figure shows that integrating the sensor system causes a minor drop of the q-factor. however, the overall flat response is still maintained. at strains approaching -600 µs the curves drop slightly due to the effect of shifted λg approaching λ1. the same observations are valid for fig. 5 representing λ3. the faster drop of q-factor for λ3 at the neighborhood of +600 µs is most probably due to non-uniform dispersion characteristics of the fiber and other nonlinear effects. figure 6 corresponds to quality of transmission of λ4. as expected, due to not having λg adjacent to this wavelength, the effect on the performance of this channel is very minor. a flat response throughout the dynamic range of the sensor can be clearly observed. eye diagrams for the four wavelengths at an applied strain of -200 µs and a power of 6 dbm for the interrogating source wavelength are shown in fig 7. the diagrams exhibit good eye opening values for all wavelengths. the relatively wider eye opening for λ1 despite having the bragg shift towards it is mainly due to the fact that it is operating at a data rate of 2.5 gbps, whereas the remaining three wavelengths operate at 10gbps. 4. conclusions a study of the effect of integrating an fbg strain sensor in a four wavelength wdm communication system operating at 1550 nm region is presented. the simulations considered focused on mutual effects of both systems. the effect of power levels of the interrogating optical source on the quality factor of the two systems has also been investigated considering a network layout in oman having variety of spans. the simulations performed on the systems at data rates of 10 gbps and 2.5 gbps for strains up to ±600 µs, exhibit a good tolerance for the integration operation. the obtained quality factors and eye diagrams for all wavelengths used and at different interrogating sensor source values, suggest that the integration has a minor effect on -600 -400 -200 0 200 400 600 6 8 10 12 14 16 18 20 22 strain (us) q u a lit y f a ct o r (d b ) figure 5. quality factor q, in db, for transmission performance of channel 3 operating at a data rate of 10 gbps. οο: system not coupled to sensing system, €€: 0 dbm, ××: 3 dbm, ◊◊: 6 dbm, ∇ : 10 dbm, -600 -400 -200 0 200 400 600 6 8 10 12 14 16 18 20 22 strain (us) q u a lit y f a ct o r (d b ) figure 6. quality factor q, in db, for transmission performance of channel 4 operating at a data rate of 10 gbps. οο: system not coupled to sensing system, €€: 0 dbm, ××: 3 dbm, ◊◊: 6 dbm, ∇ : 10 dbm, 0.8ns 0.2ns 0.2ns 0.2ns (c) (d) figure 7. eye diagrams for the four wavelengths for an applied strain of -200 µµs and a power of 6 dbm for the interrogating source. eye diagrams for wavelengths 1 to 4 are respectively sho wn in (a through d) . note that the fbg response is shown in fig. b) with a value of q of about 10.3 db (a) (b) 50 the journal of engineering research vol. 6, no.1, (2009) 46-50 both systems. this suggests the possibility of utilizing existing communications systems for fbg sensing applications. references abad, s., araújo f. m., ferreira, l. a., santos, j. l. and lópez-amo, m., 2003, "interrogation of wavelength multiplexed fiber bragg gratings using spectral filtering and amplitude-to-phase optical conversion", j. lightwave technol., vol. 21(1), pp 127-131. agrawal, g., 2001, "applications of non linear fiber optics", academic press. bourdoucen, h. and al lawati, a., 2006, "effect of dispersion compensation on fbg sensing systems performance", j. of instruments and experimental techniques, (springer), vol. 49(3), pp. 403-407. chan, p., gong, w.j.m. and demokan, m., 1999, "multiplexing of fiber bragg grating sensors using an fmcw technique", ieee photonics technology letters, vol. 11(11), pp. 1470-1472. fan, y. and kahrizi, m., 2004, "applicability of an optimized fiber optic smart structure", proceedings of the international conference on mems, nanos and smart structures (icmens'04), pp. 386-392. gebremichael, y. m., li w., meggitt, b. t., boyle, w. j. o., grattan, k. t. v., mckinley, b., boswell, l. f., aarnes, k. a., aasen, s. e., tynes, b., fonjallaz, y. and triantafillou, t., 2005, "a field deployable, multiplexed bragg gratings sensor system used in an extensive highway bridge monitoring tests", ieee sensors journal, vol. 5(3), pp. 510-519. kang, s., kim, s., lee, s., kwon, s., choi, s. and lee, b., 1998, "temperature-independent strain sensor system using a tilted fiber bragg grating demodulator ", ieee photonics technology letters, vol. 10(10), p. 146. lim, j., yang, q., jones, b. and jackson, p., 2001, "strain and temperature sensors using multimode optical fiber bragg gratings and correlation signal processing", technology conference, budapest, hungary, pp. 1463-1466. madhav, k. and asokan, s., 2004, "spectrum estimation by wavelength shift time-stamping in a fiber bragg grating sensor", ieee photonics technology letters, vol. 16(5), p. 1355. peng, p., tseng, h. and chi, s., 2002, "self-healing fibre grating sensor system using tunable multiport fibre laser scheme for intensity and wavelength division multiplexing", electronics leters, vol. 38(24), pp. 1510-1512. sinkin, v., holzlohner, r., zweck, j. and menyuk, c.r., 2003, "optimization of the split-step fourier method in modeling optical-fiber communication systems," j. lightwave technol., vol. 21, pp. 61-68. warf, b., 2006, “the professional geographer", vol. 58(1), pp. 1-11. www.o-eland.com/sensorproducts/sensor_strain.html. the journal of engineering research (tjer), vol. 19, no. 2, (2022) 140-151 0043 corresponding author's e-mail: wiseman_1988@yahoo.com doi:10.53540/tjer.vol19iss2pp140-151 yearly improvement of grid-connected solar pv system parameters by planar concentrators alaa n. abed1,*, naseer k. kasim2, and hazim h. hussain1 1 department of atmospheric science, college of science, mustansiriyah university, baghdad, iraq 2 ministry of electricity/training and energy research office, baghdad, iraq abstract: planar concentrators are used in the current manuscript to improve the solar pv system parameters (electrical energy, array yield, and solar irradiation). additionally, study the temperature (both the ambient temperature and the temperature of the pv modules), performance ratio, and efficiency. the current pv system is situated at al-taji town in baghdad. these improvements are achieved by using planar concentrators to increase solar radiation (made of aluminium metal). the results demonstrated a 21% increase in the yearly average energy output for improved solar pv modules. the improved solar pv modules' average yearly array yield increased by 20.6%. compared to the reference pv modules, the improved solar pv modules received 24% more solar irradiation yearly on average. the monthly average of the performance ratio (pr) and efficiency to the improved solar pv modules and reference solar pv modules are 89.3% & 13.61%, and 91.2% & 13.89%, respectively. the yearly average temperatures of the reference pv solar modules and improved pv solar modules are 48.8oc and 46.0oc, respectively, at an average ambient temperature of 29.2oc. the originality of this work is the successful improvement of the electrical energy of the grid-tied pv system, in addition to studying the performance of the second generation of photovoltaic solar modules (cigs), where cigs is the pv module technology that is used in this manuscript. keywords: planar concentrators; performance; efficiency; pv modules; grid-connected. العامة نظام الطاقة الشمسية الكهروضوئية المتصلة بالشبكة لمعامالتالتحسن السنوي بواسطة المركزات المستوية عالء ن. عبد*, نصير ك. قاسم, حازم ح. حسين تُستخدم المركزات المستوية في البحث الحالي لتحسين معلمات نظام الطاقة الشمسية الكهروضوئية )الطاقة الكهربائية، الملخص: وا المصفوفة، حرارة وعائد ودرجة المحيطة الحرارة )درجة الحرارة درجة دراسة إلى باإلضافة الشمسي(. االلواح إلشعاع الكهروضوئية( ونسبة األداء والكفاءة. يقع النظام الكهروضوئي الحالي في بلدة التاجي ببغداد. يتم تحقيق هذه التحسينات باستخدام سنوي متوسط ك٪ 21المستوية لزيادة اإلشعاع الشمسي )المصنوع من معدن األلمنيوم(. أظهرت النتائج زيادة بنسبة المركزات الطاقة من متوسط لأللواحإنتاج زاد المحسنة. الكهروضوئية المصفوفةالشمسية الشمسية لأللواحالسنوي عائد الكهروضوئية الكهروضوئية المرجعية، تلقت الوحدات الشمسية الكهروضوئية المحّسنة إشعاًعا شمسيًا باأللواح٪. مقارنةً 20.6المحسنة بنسبة الطاقة الشمسية الكهروضوئية المحسنة أللواح( والكفاءة prالشهري لنسبة األداء ). المتوسط كمتوسط سنوي٪ 24أكثر بنسبة درجات الحرارة ٪ على التوالي. يبلغ متوسط 13.89 &٪ 91.2٪ و 13.61 &٪ 89.3الكهروضوئية المرجعية عند وااللواح متوسط عند درجة مئوية على التوالي، 46.0و درجة مئوية 48.8 والُمحسَّنةالشمسية الكهروضوئية المرجعية لأللواحالسنوية يبلغ محيطة حرارة الكهروضوئية 29.2درجة لنظام الكهربائية للطاقة الناجح التحسين هو العمل هذا أصالة مئوية. درجة من الثاني الجيل أداء دراسة إلى باإلضافة بالشبكة. ) االلواحالمرتبطة الكهروضوئية حيث cigsالشمسية هي cigs أن (. .هذا البحث خدمة في تقنية الوحدة الكهروضوئية المست .المنظومة المتصلة بالشبكة ؛الكهروضوئية األلواح؛ الكفاءة؛ األداء ؛المركزات المستوية الكلمات المفتاحية: yearly improvement of grid-connected solar pv system parameters by planar concentrators 141 nomenclatures ac: alternating current. am: array area. cigs: copper indium gallium selenide (solar module). dc: direct current. eac: alternating energy production. eincp: energy increment percentage. hr: reference solar irradiance. ht: in-plane solar irradiation. mppt: maximum power point tracking. pr: performance ratio. pv: photovoltaic. s.irr: solar irradiation. tm: pv module actual temperature. tref : pv module reference temperature. β: temperature coefficient. ηref : rated (nominal) efficiency of pv modules pota : planar concentrators optimum tilt angles. mota : modules optimum tilt angles. pv/t: photovoltaic/thermal systems 1. introduction it is necessary to develop and employ renewable energy sources in view of the global warming issues facing the globe in general and iraq in particular, as well as the severe imbalance between the supply and demand for energy. solar photovoltaic technology is mainly used to reduce the demand for fossil fuels used in the production of electric power and to reduce the gap between consumption and production to achieve a stable state in the electric grid. employing renewable energy sources reduces the large number of local diesel generators that are used to produce electricity when the grid shuts down, which has a negative impact on both human health and the environment. solar energy resources are a clean energy source that can be exploited to meet global energy demands (abedin and rosen, (2011). the introduction of clean, renewable energy technology is a great achievement because iraq suffers from a shortage in the supply of electrical energy to institutions and homes and also suffers from an increase in pollutants as a result of the widespread use of local diesel generators to make up the shortfall in electrical power outages (alaa, 2020). employing solar power reduces emissions because it doesn't produce any greenhouse gases like co2 (alshamani et al.,2016). in contrast to conventional power systems like diesel generators, solar technology like photovoltaic solar modules is used even on rooftops. it can be directly harnessed at the place of generation without specifically requiring a transmission grid (naveen et al., 2017). from sunrise to night, solar energy is abundant everywhere, giving us enough time to harness it. there are no negative effects of solar energy on the environment. in contrast to traditional energy sources like gas and oil, which are concentrated in a certain region of the world, clean energy sources are present everywhere. the rapid adoption of green energy and energy efficiency is reducing climate change, providing significant energy security, and having an economic benefit. (seitel, 1975). ronald et al.(2000) tested the effect of a pv solar module with a v-trough and a fixed tilt angle in a swedish climate. they demonstrated that the flatplate fixed reflector increases the pv module's annual power production by 20% to 25%. according to a study by pavlov et al. (2015), using planar concentrators increased daily power by 35% during specified times of the year when there were no clouds. for poly-si and amorphous-si modules, the monthly power increases are calculated at 18% and 26%, respectively. tabaei and ameri. (2015) found that the use of booster reflectors made of aluminium foil increased the power output of polycrystalline silicon by 14%. the maximum power output when using aluminium foil reflectors with water film is 50.4%. in 2009, planar concentrators made of various materials were attached to the high and low sides of the solar pv module to examine its performance. in order to determine the most effective type of material for planar concentrators that harvests the most electrical energy with the fewest extra heat, experiments are conducted with stainless, aluminium, chrome, and steel planar concentrators. it has been found that planar concentrators made of chrome material provide 27.65% greater energy output than planer concentrators made of aluminium foil and 34.1% more energy output than planer concentrators made of stainless steel (rizk and nagarial, 2009). according to some studies, horizontal planar concentrators with vertical modules are better situated at high latitudes (duffie et al ., 2020). under the climatic circumstances of al-dhahran (ksa), bahaidarah et al. (2013) developed a solar system that is cooled by the flow of water on the backside of the pv module. the outcomes showed that the operating temperature of the pv solar modules decreased to about 20%, and the electrical efficiency increased to 9%. the pv/t system has planar concentrators that boost solar radiation to 950w/m2 and a water mass flow rate of 0.042kg/sec. alaa n. abed, naseer k. kasim, and hazim h. hussain 142 figure 1. present pv solar system. figure 2. diagram of pv system block circuit. table 1. specifications of the solar pv system and solar pv module. module features value system features value module model ts-165c2 cigs inverter model sma sb-5000t-21 max power (pmax) 165 w number of modules 30 open-circuit voltage (voc) 88.7 v inverter size 5.30 kwp short-circuit current (isc) 2.66 a inverter efficiency 97% max power voltage (vmpp) 68.5 v system size 5 kwp max power current (impp) 2.41 a pv modules tilt angle 30o max reverse current (ir) 6.5 a planar concentrators tilt angle operating temperature -40°c to 85°c temperature coefficient of pmax -0.30% /oc the results of the combined electrical and thermal efficiencies are reported at 71.40, where the pv electrical efficiency is 12.40% and the thermal efficiency is 59% (palaskar et al., 2015). this study aims to raise energy output because doing so will increase revenue, which will lead to good economic feasibility. 2. pv system description the present pv system has been mounted in baghdad at latitude 33.3ᴼn and longitude 44.4ᴼe, as exposed in figure 1. the present solar pv system includes 30 modules, which are designed in 5 strings with six series-connected modules, as exposed in figure 2. the specifications of solar pv systems and solar pv modules are shown in table 1. 3. solar pv modules and planar concentrators the current solar pv system in this work is divided into two groups: the first group, which is classified as an improved group that contains 12 pv modules, and the second group, which is classified as a reference group that contains 18 pv modules. the increment percentage (gain) in the performance characteristics of the solar pv system is then calculated by comparing the improved group to the reference group. the improved group receives solar radiation of a higher intensity than the reference group (one of which is from the planar concentrators and the other from the sun). in contrast to the second group (18 modules), the improved group's planar concentrators are attached in front of it, as indicated in fig. 3 (added to 12 modules). the inverter in the present solar pv system has two inputs: the improved group input, which contains 12 pv modules (1980 w), and the reference group input, which has 18 pv modules (2970 w). the data is gained by connecting an inverter to a computer with a speed wire. as seen in fig. 4, an inverter exhibits data in two clusters (a and b). the reference group data are represented by cluster (b), while the improved group data are represented by cluster (a). the inverter in the current system involves two inverters with the maximum point tracker (mppt) technology (yilmaz et al., 2019). since the reference group contains 18 pv modules and the improved group contains 12 pv modules, before all the calculations, the power of the reference yearly improvement of grid-connected solar pv system parameters by planar concentrators 143 group is divided by 18 and multiplied by 12 to accomplish an equalization in power (number of pv modules) between the improved and reference groups so that each group contains 12 pv modules. the planar concentrators are manually attached to the solar modules and oriented to reflect the sun's radiation directly onto the solar modules in order to attain the optimum angle. then it is monitored since the sun's path varies throughout the months and prevents planar concentrators from reflecting all of the solar radiation that reaches them. as a consequence, the concentrators' angle must be adjusted to coincide with the sun's elevation angle for each month. six of the optimum angles throughout the course of a year are determined after all these observations and adjustments, and they are shown in table 2. throughout the year, the pv solar system modules' tilt angles are fixed at 30° in accordance with baghdad's latitude. table 2. planer concentrators and pv solar modules have optimum tilt angles. pota mota month 17o 30o december 20 o 30o november and january 25o 30o october and february 30o 30o march and september 35o 30o april, may and august 37o 30o june and july where: pota and mota are abbreviations of planar concentrators' optimum tilt angles and modules' optimum tilt angles, respectively. figure 3. reference and improved solar pv modules (groups). figure 4. screen capture of inverter data presentation in the laptop. alaa n. abed, naseer k. kasim, and hazim h. hussain 144 4. preparing the study the current experiment is carried out at the almansour company in baghdad (longitude 44.4 degrees east and latitude 33.3 degrees north). performance parameters that have been improved include electrical energy, array yield and solar radiation. the temperature of the improved pv modules, reference pv modules and ambient air are all measured. all of the parameters above-mentioned are recorded for both the reference and improved groups. each group's data is collected from the inverter, which exhibits the solar pv system data in the individual (a and b) manner as described above. the study is conducted from 7:30 am to 6:00 am on each day. this work is a sample study, so the days in the middle of the month (13th, 14th, 15th, 17th) or even 19 in the case of cloudy months are those in which performance improvement is being studied because it is impossible to calculate electrical energy improvement on cloudy days. the purpose of doing this study in the middle of the month is that this is because solar radiation and ambient temperature are on average for each month. the current study was carried out in 2020 for a whole 12 months. 5. performance parameters the performance parameters (electrical energy, performance ratio (pr), efficiency, and array yield in addition to solar irradiation, pv modules, and ambient temperatures) for the reference and improved (modules) groups are studied. speedwire directly obtains energy and power from an inverter, whereas the following equations are used to calculate efficiency, pr, and solar irradiance. a digital thermometer is used to measure both the temperature of the solar pv modules and the surrounding environment. 5.1. efficiency efficiency is classified into an array, system, and inverter efficiencies. the array efficiency (ɳ𝑃𝑉 ) is based on the dc power output, while the system efficiency (ɳ𝑆𝑌𝑆 ) is based on the ac energy output (de lima et al.,17; abed et al.,2020). the array efficiency is the ratio of the dc energy output to the in-plane solar irradiation multiplied by the area of pv modules (pv array) (kumar et al., 2017; attari and asselman, 2016). the array efficiency is calculated as follows: ɳ𝑃𝑉 = 100∗𝐸𝐷𝐶 𝐻𝑇∗𝐴𝑚 % (1) the system efficiency is given in equation 2: ɳ𝑆𝑌𝑆 = 100∗𝐸𝐴𝐶 𝐻𝑡∗𝐴𝑚 % (2) where 𝐻𝑡 is the in-plane solar insolation, 𝐴𝑚 is the solar pv array area, and edc is the dc energy. inverter efficiency is given in equation 3: ɳ𝐼𝑁𝑉 = 100∗𝐸𝐴𝐶 𝐸𝐷𝐶 % (3) an inverter's efficiency ranges from 97% to 96% because it is indoors (kasim et al., 2020). 5.2. performance ratio (pr) pr is an essential indicator since it reveals all the negative impacts of the solar photovoltaic system. the pr allows for the comparison of pv solar systems regardless of solar radiation resources, tilt angle, orientation angle, their nominal, and power capacity. it shows how near the actual pv system is to reaching ideal performance during actual working time. (ozden et al., 2017; obaid et al 2020). pr is given in equation 4. 𝑃𝑅 = 𝑌𝐹 𝑌𝑅 % (4) where: yf is the final yield estimated by equation (5), (yr) is the reference yield estimated by equation (6). the final yield (yf) is the yearly, monthly, or even daily output ac energy of the pv system divided by the rated power of the pv system at standard test conditions (1 kw/m2 and 25°c cell temperature) (sharma et al.,2013; obaid et al., 2019). the yf denotes the number of hours per day that the solar pv system is working at its rated capacity. equation 5 calculates yf. 𝑌𝐹 = 𝐸𝐴𝐶 𝑃𝑟𝑎𝑡𝑒𝑑 (𝑘𝑊ℎ/𝑘𝑊) (5) where eac represents ac energy production (kwh), the reference yield (yr) is calculated by dividing the reference irradiance, which is equal to 1 kw/m2, by the in-plane global insolation. yr is given by equation 6 (adaramola et al., 2015; rezk et al., 2019). 𝑌𝑅 = 𝐻𝑇 𝐻𝑅 (𝑘𝑊ℎ/𝑘𝑊𝑃) (6) where 𝐻𝑅 and 𝐻𝑇 are the reference irradiance and inplane global insolation (irradiation), respectively. when the equations. (5) and (6) are substituted into equation (4), and equation (7) is gotten (kasim et al., 2019). yearly improvement of grid-connected solar pv system parameters by planar concentrators 145 𝑃𝑅 = 𝐸𝐴𝐶∗ 𝐻𝑅 𝑃𝑃𝑉,𝑟𝑎𝑡𝑒𝑑∗𝐻𝑇 % (7) the performance ratio (pr) can also be estimated by equation 8: 𝑃𝑅 = 𝜂𝐴𝑐𝑡𝑢𝑎𝑙 𝜂𝑟𝑒𝑓 (8) equation 8 is utilized to calculate the (pr). the array yield (ya) is the dc energy output for a specific period of time divided by the nominal power of the pv solar system. the ya represents the number of hours per day that the solar pv system is operating at its nominal power (abed et al., 2020). the array yield is given in equation 9: ya = edc ppv.rated (kwh/kwp) (9) where edc represents the direct energy output (dc) (kwh) of the pv solar array. the actual (pac) and rated (prated) power are estimated as follows: pac = hr*ηactual*am (10) prated = hr * ηref*am (11) the actual efficiency (ηactua) is estimated via equation 12. it can also be estimated by equation 13. ηactual = ηref [1-β(tm-tref)] (12) where ηref is the rated efficiency (15.2%), am is the area of reference and improved pv modules (13.04 m2 each), β is the temperature coefficient that equals -0.3%/°c, tref is the reference modules' temperature (25°c), and tm is the actual pv modules' temperature (kasim et al., 2019). ηactual = electrical power solar irradiance∗area (13) from equation 10, equation 14 is obtained as follows: hr= pac am∗ηactual (14) where pac, am and ηactual are the actual power, improved group area and actual efficiency, respectively. equations 14 and 8 are used to calculate solar irradiance (hr) and actual efficiency (ηactual), respectively. the formula used to calculate energy increment percentage (eincp) is given as follows: eincp= (eim-eref)/ eref *100% (15) where eim, and eref are the energy of the improved and reference groups, respectively. eincp is an abbreviation of the energy increment percentage or the energy incp (energy gain). equation 14 is used to calculate the energy increment percentage. 6. results and discussion figure 5 shows the energy incp, improved, and reference pv modules (groups). the highest energy output values coincide with the peak solar radiation and the optimum tilt angle for solar pv modules. in may, the maximum monthly daily energy values of the reference and improved modules were 12.99 kwh and 16 kwh, respectively, whereas the energy incp for this month was 23.3%. while the minimum values are 8, 28 kwh, and 10 kwh, respectively, in december, the energy incp for this month is 18%. utilizing planar concentrators results in a 21% monthly daily average increase in energy over the course of the year (34.0 kwh). the physical interpretation of may having the highest energy production is that it gets the most solar radiation and has the optimal tilt angle for pv solar modules, which is 30o. tests are conducted on one day of each month that lies in the middle of the month, and this day is considered the monthly daily average. in fig. 5, the energy incp curve has two crests and three bottoms. the crests show up during the spring and autumn seasons as energy incp increases. while the bottoms show up during the cold and hot seasons as energy incp decreases. the daily behaviour of the power produced by improved and reference pv modules is shown in figure 6. the difference between the two groups is clearly shown in the figure. at 12:00 pm, the maximum power values for the reference and improved pv modules are 1.712 kw and 2.308 kw, respectively, with an incp (the gain) of 34.4%. at 5:00 pm, incp has the lowest value of 4.9%, while on average, throughout the day, it is 24.7%. the data for all parameters studied in this research) electrical energy, array yield, solar radiation, efficiency, and pr) are included in the appendix. figure 7 shows the performance ratio of improved and reference pv modules. the maximum pr values of reference solar pv modules and improved solar pv modules are 94.5% and 95.5%, respectively, in december and january. in comparison, the minimum performance ratios were 84.3% and 88.1% in july, respectively. the performance ratio for the improved reference pv modules is 89.2% and 91.1%, respectively, on a monthly daily average over the course of the year. alaa n. abed, naseer k. kasim, and hazim h. hussain 146 because the planar concentrators attached to the pv modules in the improved group increase the temperature of these pv modules, the pr of the improved pv modules is lower than the pr of the reference pv modules. as demonstrated in fig 7, the pr of the improved pv modules and the pr reference pv modules converge in the cold and mild months while diverging in the hotter seasons. despite the fact that planar concentrators increase solar radiation, which in turn raises the temperature of solar pv modules, the pr of the improved modules remains to be excellent throughout the winter. figure 5. energy incp, improved and reference pv modules (groups) energy produced. figure 6. the power produced for improved and reference modules by using planar concentrators yearly improvement of grid-connected solar pv system parameters by planar concentrators 147 figure 7. pr of the improved and reference groups. figure 8. efficiency of improved and reference pv modules (groups). figure 9. solar irradiance of the s.irr incp, improved and reference pv modules (groups). alaa n. abed, naseer k. kasim, and hazim h. hussain 148 the pr of a pv system indicates how close it approaches ideal performance during real operation and allows comparison of pv systems independent of location, tilt angle, orientation, and their nominal rated power capacity (khalid et al.,2016; kasim et al., 209). figure 7 shows that the pr increases and decreases when the temperature of the pv modules decreases (during the cold months) and increases (during the hot months). the improved and reference efficiency of pv modules is shown in fig. 8. the two coldest months, january and december, had maximum efficiency values for the improved pv modules and reference pv modules of 14.4% and 14.56%, respectively. in contrast, efficiency drops to its lowest levels in july and september, at 12.8% and 13.3%, respectively. fig 8 shows that the largest efficiency of pv modules does not exceed the efficiency at stc (15.2%) because the temperature of solar pv modules is greater than 25°c. planar concentrators slightly lower the efficiency of pv solar modules by increasing solar radiation, which raises the temperature of the pv solar modules. the improved and reference groups' monthly daily average efficiency is 13.6% and 13.85%, respectively. this shows a slight difference in efficiency between the improved pv modules and the reference pv modules. figure 9 displays the solar irradiation for the improved group and the control group. while the reference group solar irradiation reaches its maximum in may and june at 610 wh/m2, the improved group solar irradiation reaches a maximum in march at 780 wh/m2. the minimum solar irradiation for the improved and reference groups was 475.45 wh/m2 and 400.91 wh/m2, respectively, in december. this is due to the fact that both the tilt angles of solar pv modules and planar concentrators are optimum in march, but they are not optimal in june. therefore, the solar irradiation (s.irr) incp reaches its highest in march (30%) and its minimum in december (18%). these incp values correspond to 181 wh/m2 and 74.5 wh/m2, respectively. the concept of array yield and its equation are described above. the array yield is shown in figure 10. in may, the array yield reaches its maximum value. the maximal array yields (monthly daily average) for the reference group and improved group are 6.56 kwh and 8.1 kwh, respectively. where the array yield incp is 23.2%, this value of energy incp corresponds to 2.0 kwh. in december, because there is the lowest amount of solar energy, the array yield is at its lowest value. the minimum array yields for the reference group and improved group are 4.18 kwh and 4.94 kwh, respectively. the monthly daily average (over the course of the year) of the array yield incp as a consequence of the usage of planar concentrators is 22.7%, and this value of energy incp corresponds to 17.8 kwh. figure 11 illustrates the ambient air, improved, and reference pv module temperatures. figure 10 shows that the maximum temperatures for the reference pv modules, improved pv modules, and ambient in july are 57.5°c, 61°c, and 41.7°c, respectively. the temperature of the improved pv modules is about 3.5°c higher than that of the reference pv modules. in january and december, the minimum temperatures for the reference pv modules, improved pv modules, and ambient air were 34.15 °c, 36.5 °c, and 16.3 °c, respectively. the average monthly daily temperatures of reference pv solar modules, improved pv solar modules, and ambient are 45.94 degrees celsius, 48.8 ° c, and 29.15 ° c, respectively. in cold months, the temperature difference between reference and improved solar pv modules decreases because the temperature of the improved pv modules decreases. figure 11 shows that the difference between improved and reference group temperatures is small, indicating that the planar concentrators add just a small amount of heat. conclusion the current study reaches the following conclusions: • the use of planar concentrators results in significant gains (incp) for the pv solar system parameters (energy output, array yield and solar irradiation). • the maximum energy increases in the summer months while it begins to decrease in the winter, based on the atmospheric elements, which include the ambient temperature and solar radiation. the energy enhancement varies throughout the year, dependent on these atmospheric elements. • the improved pv modules get a small amount of heat from the planar concentrators. yearly improvement of grid-connected solar pv system parameters by planar concentrators 149 figure 10. array yields of improved reference groups. figure 11. pv solar modules and ambient temperatures. • because the improved pv modules are becoming hotter than the reference pv modules, there performance and efficiency are lower than those of the reference pv modules, but the difference between them is very small. • the daily average monthly (throughout the year) energy gained (eincp) is about 21%, which is equivalent to 34.0 kwh. • the daily average monthly solar irradiation incp (s.iir incp) is about 25%, which is equivalent to 127wh/m2. • march and may have the highest incp for all performance parameters due to the mild ambient temperature and sufficient solar radiation. • due to the highest solar radiation intensity and ambient temperature, solar pv modules reach their highest temperatures in july, august, and september acknowledgement all the researchers are grateful to mustansiriyah university/college of science, al-zawraa company / al-mansour company and training & energy research office /ministry of electricity for their great efforts in supporting the accomplishment of this work. conflict of interest the authors declare that there are no conflicts of interest regarding this article. funding the authors did not receive any funding for this research. alaa n. abed, naseer k. kasim, and hazim h. hussain 150 references abed, a. n., hussain, h. h., & kasim, n. k. (2020). efficiency and performance improvement via using optical reflectors of on-grid cigs pv solar system. karbala international journal of modern science, 6(1), 5. adaramola, m. s., & vågnes, e. e. (2015). preliminary assessment of a small-scale rooftop pv-grid tied in norwegian climatic conditions. energy conversion and management, 90, 458465. alaa n. abed (2020), "performance improvement of cigs pv solar grid tied system using planer concentrators, case study: baghdad," ph.d. dissertation, atmospheric. dept., mustansiriyah. univ., baghdad. iraq. al-shamani, a.n., sopian, k., mat, s., hasan, h.a., abed, a.m., and ruslan, m.h., (2016). experimental studies of rectangular tube absorber photovoltaic thermal collector with various types of nanofluids under the tropical climate conditions. energy conversion and management, 124,528-542. attari, k., elyaakoubi, a., & asselman, a. (2016). performance analysis and investigation of a gridconnected photovoltaic installation in morocco. energy reports, 2, 261-266. bahaidarah, h. m., rehman, s., gandhidasan, p., & tanweer, b. (2013, june). experimental evaluation of the performance of a photovoltaic panel with water cooling. in 2013 ieee 39th photovoltaic specialists conference (pvsc) (pp. 2987-2991). ieee. de lima, l. c., de araújo ferreira, l., & de lima morais, f. h. b. (2017). performance analysis of a grid connected photovoltaic system in northeastern brazil. energy for sustainable development, 37, 79-85. duffie, j. a., beckman, w. a., & blair, n. (2020). solar engineering of thermal processes, photovoltaics and wind. john wiley & sons. ellabban, o., abu-rub, h., & blaabjerg, f. (2014). renewable energy resources: current status, future prospects and their enabling technology. renewable and sustainable energy reviews, 39, 748-764. h abedin, a., & a rosen, m. (2011). a critical review of thermochemical energy storage systems. the open renewable energy journal, 4(1). kasim, n. k., hussain, h. h., & abed, a. n. (2019). performance analysis of grid-connected cigs pv solar system and comparison with pvsyst simulation program. international journal of smart grä±d, 3, 172-179. kasim, n. k., hussain, h. h., & abed, a. n. (2020). studying the performance of second-generation pv solar technology under baghdad climate. research journal in advanced sciences, 1(2). [available online https://royalliteglobal.com/rjas/article/view/389]. khalid, a. m., mitra, i., warmuth, w., & schacht, v. (2016). performance ratio–crucial parameter for grid connected pv plants. renewable and sustainable energy reviews, 65, 1139-1158. kumar, n. m., kumar, m. r., rejoice, p. r., & mathew, m. (2017). performance analysis of 100 kwp grid connected si-poly photovoltaic system using pvsyst simulation tool. energy procedia, 117, 180-189. naveen, s., mufassireen, a., shashikiran, n., and naganagouda, h. (2017). solar photovoltaic systems–applications & configurations. international research journal of engineering and technology (irjet), 4(08),1851-1855 obaid, n. m., hashim, e. t., & kasim, n. k. (2020). performance analyses of 15 kw grid-tied photo voltaic solar system type under baghdad city climate. journal of engineering, 26(4), 21-32. obaid, n. m., kasim, n. k., & hashim, e. t. (2019). performance assessment of first grid–tied pv solar system under baghdad city climate condition. iraqi journal of science and technology, 10(1), 63-71. ozden, t., akinoglu, b. g., & turan, r. (2017). long term outdoor performances of three different ongrid pv arrays in central anatolia–an extended analysis. renewable energy, 101, 182-195. palaskar, v. n., & deshmukh, s. p. (2015). waste heat recovery study of spiral flow heat exchanger used in hybrid solar system with reflectors. science, 5, 476-482. pavlov, m., migan-dubois, a., bourdin, v., pons, m., haeffelin, m. and badosa, j.(2015). experimental and numerical study of the influence of string mismatch on the yield of pv modules augmented by static planar reflectors. ieee journal of photovoltaics, 5(6), 1686-1691. rezk, h., gomaa, m. r., & mohamed, m. a. (2019). energy performance analysis of on-grid solar photovoltaic system-a practical case study. international journal of renewable energy research (ijrer), 9(3), 1292-1301. rizk, j., & nagarial, m. h. (2009). impart of reflectors on solar energy systems. international journal of electrical and electronics engineering, 3-9. rönnelid, m., karlsson, b., krohn, p., & wennerberg, j. (2000). booster reflectors for pv modules in sweden. progress in photovoltaics: research and applications, 8(3), 279-291. seitel, s.c., 1975. collector performance enhancement with flat reflectors. solar energy, 17(5), 291-295. sharma, v., & chandel, s. s. (2013). performance analysis of a 190 kwp grid interactive solar photovoltaic power plant in india. energy, 55, 476-485. tabaei, h., & ameri, m. (2015). improving the effectiveness of a photovoltaic water pumping https://royalliteglobal.com/rjas/article/view/389 yearly improvement of grid-connected solar pv system parameters by planar concentrators 151 system by using booster reflector and cooling array surface by a film of water. iranian journal of science and technology. transactions of mechanical engineering, 39(m1), 51. yilmaz, u., turksoy, o., & teke, a. (2019). improved mppt method to increase accuracy and speed in photovoltaic systems under variable atmospheric conditions. international journal of electrical power & energy systems, 113, 634-651. appendix table 3. energy, array yield and s.iir for improved and reference pv solar modules. s.irr incp (%) reference pv modules s.irr wh/m2 improved pv modules s.irr wh/m2 reference yield incp (%) reference modules array yield kwh/kwp improved modules array yield kwh/kwp energy incp (%) reference pv modules energy kwh improved pv modules energy kwh month 20.6% 411.8182 496.8182 19.0% 4.34 5.16 19.0% 8.59 10.22 january 22.9% 437.2727 536.3636 21.1% 4.45 5.39 21.1% 8.81 10.67 february 30.4% 598.0244 780 24.8% 6.36 7.94 24.8% 12.6 15.731 march 28.1% 573.0924 733.3333 24.1% 6.27 7.78 24.1% 12.41 15.4 april 26.0% 608.3333 766.6667 23.2% 6.56 8.08 23.2% 12.99 16 may 23.0% 610.8333 748.3333 20.0% 6.47 7.77 20.0% 12.82 15.38 june 22.8% 573.3333 704.1667 17.6% 6.06 7.13 17.6% 12 14.11 july 23.8% 570.8333 706.6667 20.2% 6.11 7.35 20.2% 12.1 14.55 august 23.8% 595.8333 737.5 21.3% 6.23 7.56 21.3% 12.34 14.965 september 22.8% 536 658.3333 22.7% 5.76 7.07 22.7% 11.41 14 october 21.3% 427.2727 518.1818 20.2% 4.42 5.31 20.2% 8.75 10.52 november 18.6% 400.9091 475.4545 18.0% 4.18 4.94 18.0% 8.28 9.772 december table 4. efficiency and pr for improved and reference pv solar modules. pr of reference pv modules pr of improved pv modules efficiency of improved pv modules (%) efficiency of reference pv modules (%) month 95.8% 94.4% 0.143561877 0.145570495 january 92.7% 91.3% 0.138832392 0.140900673 february 92.5% 88.6% 0.134630136 0.140653298 march 91.2% 88.4% 0.134343434 0.138651433 april 89.9% 87.8% 0.133509003 0.1366044 may 88.5% 86.3% 0.131187576 0.134448308 june 88.1% 84.3% 0.128188393 0.133897111 july 89.2% 86.7% 0.131718125 0.135604217 august 87.2% 85.4% 0.129811105 0.132491347 september 89.6% 89.5% 0.136043984 0.136181466 october 94.0% 93.2% 0.141683502 0.142918547 november 94.8% 94.4% 0.143436661 0.1441352 december mech071006.qxd the journal of engineering research vol. 6, no. 1 (2009) 59-65 1. introduction filled polymer composites are attracting ever growing interest. therefore, a lot of research work has been carried out on reinforced polymers composites (fang, z. and hu, q., 1999; mitsuishi, k., 1997; lee, k.m. and han, c.d., 2003; hu, x., et al. 2003; mcginness, g.b. and bradaigh, c.m.o., 1997; kim, j.i., ryu, s.h. and chang, y.w., 2000; tavman, j.h., 1996; solomon, m.j., et al. 2001; marchant, d. and jayaraman, k., 2002; ray, s.s., maiti, p., okamoto, m., yamada, k. and ueda, k., 2002; zhang, q. and archer, l.a., 2002; wang, z., wu, q., dong, j., hu, y. and qi, z., 2002; roy, d., bhowmick, a.k. and de, s.k., 1993; nam, j.d. and sefferis, j.c.,1999). fiberreinforced plastic composites began with cellulose fiber in phenolics in 1908, later extending to urea and melamine, and reaching commodity status in the 1940s with glass ________________________________________ corresponding author’s e-mail: mahmoud@squ.edu.om fiber in unsaturated polyesters (mohanty, a.k. misra, m. and hinrichsen, g., 2000). nowadays mineral fillers and fibers are extensively used in the plastics industry to achieve desired properties or to reduce the price of the finished article (angles, m.n., salvado, j. and dufresne, a., 1999). the importance of fiber-filled composites arises largely from the fact that such materials can have unusually high strength and stiffness for a given weight of material. in addition, there is an increasing interest in environmental concerns. it is incorporated to maximize the use of renewable resources and also to minimize the wastes. so the valorization of a lignocellulosic residual material and its use as a lightweight and economical source of reinforcement in thermoplastic composites has received substantial attention (angles, m.n., salvado, j. and dufresne, a., 1999). lignocellulosic materials are the most abundant renewable biomaterial of photosynthesis on earth (mohanty, a.k. misra, m. and hinrichsen, g., 2000). compared to inorganic fillers, the main advantages of lignocellulosics are their renewable nature, wide variety of dynamic mechanical moduli of residual softwood-filled polystyrene mahmoud abdel-goad department of petroleum and chemical engineering, college of engineering, sultan qaboos university, p.o. box 33, pc 123, al-khoud, muscat, sultanate of oman received 6 september 2007; accepted 10 march 2008 abstract: residual softwood sawdust (rss) was added to polystyrene (ps) that has molecular weight 280000 g/mol. the ps composite was prepared by melt mixing technique. the dynamic mechanical tests are carried out using ares-rheometer (rheometric scientific, usa) in the dynamic mode and parallel plate geometry with diameter 25 mm. the measurements were performed at temperatures ranging from 140 to 240°c and frequencies varied from 0.1 to 100 rad/s at strain 1% and gap setting 2 mm. the dynamic mechanical properties in terms of complex modulus, g*, torque, compliance moduli, loss tangent and complex viscosity has been studied for fiber-filled ps composite. the viscoelastic properties of the filled and unfilled systems have been compared. these properties are found to be improved by the addition of rss. the dynamic mechanical moduli and viscosity were found to rise with fiber loading. keywords: residual softwood sawdust, composite, dynamic mechanical analyzer, shear creep, stress relaxation, melt viscosity áæ«∏dg ö°ûÿg éjé≤ñh ƒ°ûùg øjòà°sg ‹ƒñ∏d ᫵«eéæj~dg 鵫f鵫ÿg ¢üfé°ün ogƒ÷g ~ñy oƒªfi áá°°uuóóÿÿgg¢üfé°üÿg .qƒ¡°üÿg §∏n á≤jô£h §∏ıg øjòà°sg ‹ƒñdg ò° – ”.(∫ƒe/ºl28000) ‹é©dg åjõ÷g ¿rƒdg …p øjòà°sg ‹ƒñdg ¤g ö°ûÿg iqé°ûf éjé≤h ¬aé°veg â“ : ¢üfé°üÿg ¿g úñj åëñdg øeh .á«fék/áød 100 ¤g 1 úh ée ooôj h ájƒäe álqo 240 ¤g 140 úh ée iqgôm äélqo ~y î«eƒáôdg ré¡l ᣰsgƒh ég~j~– ” á«lƒdƒjôdg .᫵«à°sóh √oéªc ™°shg ¥é£f ≈∏y é¡eg~îà°sg øµá é‡ øjòà°sg ‹ƒñdg ¤g á«ñ°ûÿg äéø∏ıg √òg áaé°véh gòãc âæ°ù– ~b á«lƒdƒjôdg ääggooôôøøÿÿggáá««mmééààøøÿÿgg.álhõ∏dg ,»cô◊g »µ«f鵫ÿg π«∏ëàdg ,§«∏ÿg ,ú∏dg ö°ûÿg iqé°ûf : 60 the journal of engineering research vol. 6, no. 1 (2009) 59-65 fillers available throughout the world, nonfood agricultural-based economy, low energy consumption, cost and density, high specific strength and modulus, high sound attenuation of lignocellulosic-based composites, comparatively easy processability due to their nonabrasive nature, which allows high filling levels, resulting in significant cost savings, and relatively reactive surface, which can be used for grafting specific groups. moreover, the recycling by combustion of lignocellulosic-filled composites is easier compared to inorganic fillers systems. therefore, the possibility of using lignocellulosic fillers in the plastic industry has received considerable interest and the study of cellulosic materials reinforced polymer composites that contain cellulosic materials has been recognized as an important area of research for over a decade (bledzki, a.k., reihmane, s. and gassan, j., 1996; george, j., bhagawan, s.s., prabhakaran, n. and thomas, s., 1995; lowys, m.p., desbrieres, j. and rinaudo, m., 2000; felix, j.m. and gatenholm, p., 1991; swerin, a., 1998; gatenholm, p., bertilsson, h. and mathaisson, a., 1993). environmental concerns generated by plastic materials are generating an increasing interest toward the development of ecological products. the use of disposable plastics increases the undegradable waste portion, for this reason, it is necessary to develop more recyclable and/or biodegradable plastics to reduce the amount of landfill. where that method of land-filling will prevent the plant roots to grow and as a result the agriculture will negatively be affected. in addition plastics waste has considerable volume since the weight of it in the waste stream is only 8 percent, but it takes up nearly 20 percent of the volume in landfills. the composites of natural fibers and non-biodegradable synthetic polymers may offer a new class of materials, but are not completely biodegradable. the designing materials compatible with the environment becomes the target from the government regulations and growing environmental awareness throughout the world. dynamic mechanical tests give more information about a composite material than other tests. the tests, over a wide range of temperatures and frequencies, are especially sensitive to all kinds of relaxation process of matrix resin and also to the morphology of the composites. dma is a sensitive and versatile thermal analysis technique, which measures the modulus (stiffness) and damping properties (energy dissipation) of materials as the material are deformed under periodic stress (saha, a.k., das, s., bhatta, d. and mitra, b.c., 1999). since polymer melt flow behavior is strongly affected by the nature of the filler type, including its morphology, surface chemistry and concentration, rheological studies can also assist in the development of formulations designed to facilitate industrial processability. the motivation of this work is to prepare biodegradable polystyrene composite by adding waste agricultural residues (residual softwood sawdust (rss)) to commercial polystyrene and analysis this composite mechanically compared to the original material of neat polystyrene. 2. experimental part 2.1 materials and preparation in this study ps composite was prepared by adding commercial ps to residual softwood sawdust. the ps used has high molecular weight of 280000 g/mol (280k). saw dust used in this study is waste agricultural residues (residual softwood sawdust (rss)) which was provided from egypt. the residual softwood sawdust was prepared by drying and shredding in a particle size of about 1mm. 7 wt% fibers were added to the ps. the size of the fiber is expected to be lower than millimeter scale after mixing with molten ps. composite samples were prepared by dry mixing homogeneously the fiber with the polymer followed by melt mixing at about 300°c for 2 hours because the glass transition temperature of neat polystyrene is 100°c, fillers shift the glass transition temperatures that means 300°c was chosen as the lowest temperature needed for 2 hours to incorporate the fiber into polystyrene with molecular weight 280000 g/mol. the samples were prepared under that condition (300°c for 2 hours) to ensure getting homogeneous distribution of the fibers throughout the polymer sample. molten-sate sample (at 300°c ) was molded at 5 bars for 30 minutes. the samples were shaped in the disc form with diameter around 25 mm and thickness about 2 mm. the samples of neat ps were moulded also with the same dimensions of 25 mm diameter and 2 mm thickness at about 250°c under 5 bars for about 30 minutes. 2.2 instrument and measurements the melt rheological properties of the material were determined using ares-rheometer (rheometric scientifics, usa). the ares instrument is a modular and extendable rheometer works under nitrogen atmosphere in a wide range of temperatures and frequencies. during testing at or above ambient temperature, gaseous nitrogen is used. the sample to be tested is positioned between two plates in the test station. upper plate is connected to the transducer and lower plated is connected to motor. before placing the sample, the normal force of the system should be kept around zero. the sample is positioned on the lower plate followed by lowering the stage until the sample just contacts the upper plate completely. the actual gap size is read electronically and this allows absolute moduli to be determined. in this work the measurements were performed in the dynamic mode and 25 mm parallel plates geometry with gap settings about 2 millimeters. the soak time at the measuring temperature was 4 minutes. that soaking time is required to transfer the heat homogeneously throughout the sandwiched samples between the plates and stabilizing the temperature before starting the measurement. the strain amplitude was kept to be 1% in the whole frequency range to ensure linearity and 8 points per frequency-decade were obtained. the samples were measured in a range of temperature from 140 to 240°c with interval temperature 10°c as a function of frequency. 61 the journal of engineering research vol. 6, no. 1 (2009) 59-65 the frequency ω, varied from 10-1 and 102 rad/s. horizontal and vertical shift factors were obtained from a two dimensional shifting. 3. results and discussion the log-log curves of the data measured at different temperatures were superposed into master curve at a reference temperature, t0, by shifting in the horizontal and vertical directions as shown in fig. 1. in this figure the shear complex modulus (g*) is plotted for unshifted and shifted data as a function of frequency. g* decreases with increase the temperature as seen in the case of unshifted data. the experimental data were shifted into a single curve (master curve) by using of the time-temperature superposition principle which described by (ferry, j.d., 1980) as: where at is the empirically derived shift factor and constants c1 and c2 are material specific and their values are listed in table 1. t0 is chosen in this study to be 160°c. table 1 shows variation in the shifting accuracy from sample to sample. that indicates the difficulties in shifting such materials (composites) using wlf principle unlike linear homopolymeric materials. the at shifts the data obtained at different temperatures along the log frequency, ω axis as shown in fig.1 and in vertical direction is given purely by bt (bt = ρt / ρtt0 pearson, d., fetters, l. and grassley, w., 1994 ). where ρ is the material density and the shift factors for these samples are plotted with respect to temperature in fig. 2. at decreases by increasing the temperatures up to a certain limit then becomes somewhat independent on the temperature. bt is nearly independent on the temperatures up to around 200°c then decreases by increasing the temperature as shown in fig. 2. the master curves of the shear complex modulus, g* for the fiber-filled and unfilled ps at 160°c is logarithmically plotted against the frequency (angular) in fig. 3. this quantity of g* can be resolved into real (g') and imaginary (g'') components such that g*= g' + i g'' (heinrich, g. and kluppel, m., 2002). where g' is the ability of the material to store energy in the cycle of the deformation, for that it is called storage modulus and g'' is defined as the ability of the material to dissipate the energy per cycle of the deformation and it is called loss energy. the dynamic spectrum of shear storage modulus (g') and shear loss modulus (g'') with respect to frequency contains information regarding the manner in which a sample responds to small magnitude deformation applied over varying temperatures and time scale. the master curve of g* displays three distinct regions of behavior and that is typically for flexible-chain monodisperse homopolymer as the glass transition, rubbery plateau, and terminal zones when going from high frequency to low frequency. the glass transition zone is the regime between the glass and rubber-like zones at which the deformation frequency is high compared to the average relaxation time of the polymer chain. since some molecular chain segments can not move while some are free to move. if the stress is initially applied, the segment will move in such a manner as to reduce the stress on it. after moving, the segment has less stored energy because of the reduction in stress, so the excess energy was dissipated as heat and the observed modulus is high as shown in fig. 3. as the frequency of deformation decreases, g* decreases until the intermediate at which the rubbery plateau is observed. this rubbery plateau region is independent on the frequency (the flatness of g* (ω )) over several decades of frequency as seen in fig. 3. it reflects the rubber-like ( ) ( ) ( ) ( ) 1 0 0 2 0 t c t tt log a log t c t t τ τ − − = = + − 10-3 10-2 10-1 100 101 102 103 102 103 104 105 106 107 at bt at temp. ps280k: unshifted data shifted data at 160°c g * p a ω rad/s polymer c1 c2 r 2 ps280k 6.16 94.92 0.98 ps280k/7%cf 3.46 67.77 0.72 ps280k/9%cf 6.85 172.15 0.88 figure 1. the complex modulus of ps280k as a function of frequency 120 140 160 180 200 220 240 10-5 10-4 10-3 10-2 10-1 100 101 102 ps280k at bt a t & b t temp., °c figure 2. at and bt vs. temperature table 1. wlf constants of ps and ps composites at t0 = 160oc 62 the journal of engineering research vol. 6, no. 1 (2009) 59-65 properties of the polymer and indicating that polymer chains have sufficient time to relax locally but at adequately high molecular weights are entangled and exhibit behavior consistent with rubbery network (network like structure reflects in rubbery regime). figure 3 shows that g* modulus for the filled ps with 7% rss is higher than this of neat ps. numerically at around 50 radians/s the value of g* increased from 0.15 to 0.23 mpa by the incorporation of 7% filler. generally speaking, the incorporation of fibers as fillers into polymeric material matrix results in high specific strength and modulus for polymer composites based on those fillers. moreover, these fillers have easy processability due to their nonabrasive nature, which allows high filling levels therefore resulting in enhancing the cross links density hence the physical network structure. their nonabrasive nature also leads to relatively reactive surface, which can be used for grafting specific groups of polymers, accordingly, the density of chemical cross links increases. as shown in the results the complex modulus increases by introducing of filler in polystyrene. the main reason, which can be summarized from the explanation above, is because of the formation of the physical polymer-fiber network, which leads to increase the strength of the ps composite. here, the contribution of the chemical cross links can not be completely ignored since it can be formed even as non-significant contribution compared to the physical network like structure. similar results are shown in figs. 4 and 5 for the shear compliance (the storage and loss compliance). the storage, j'(ω) and loss, j''(ω) compliance are plotted with logarithmic scale in figs. 4-5. j'(ω) is a measure of the energy stored and recovered per deformation cycle, therefore it is called the storage compliance. j''(ω) is a measure of the energy dissipated as heat per cycle of the sinusoidal deformation, for that is called the loss compliance. the plot of j'(ω) has roughly the appearance of mirror image of g* plot reflected in the frequency axis. figures 4 and 5 represent three distinct regions of behavior as explained above for g* modulus. however, the intermediate regime (plateau regime in fig. 4 and rubbery in fig. 5) can not be clearly seen in these figures for insufficient high molecular weights of polystyrene. additionally, the effect of thermal degradation can not be ignored that may be occurred under the preparation conditions of ps composites (at 300°c for 2 hours). the same effect of the addition of rss is found in the shear compliance where the decrease in the j'(ω) (fig. 4) and j''(ω) (fig.5) of filled ps compared to the unfilled ps. as an example in fig. 4 at equals to100 radian/s the value of j' changed from 8.7 10-6 to 4 10-6 pa-1 by the addition of 7% rss. this is caused by the formation of the network (physical network-like structure), hence the increase in the strength of the ps/rss composite. 10-3 10-2 10-1 100 101 102 103 103 104 105 ps280k+7% filler pure ps280k g * p a ω rad/s figure 3. master curve of g* for filled and unfilled ps280k vs ω at t0 = 160oc 100 101 102 103 10-6 10-5 10-4 pure ps280k ps280k+7% filler j' p a -1 ω rad/s figure 4. master curve of j' for filled and unfilled ps280k vs ω at t0 = 160oc 102 103 10-6 10-5 pure ps280k ps280k+7% filler j' ' p a1 ω rad/s figure 5. master curve of j'' for filled and unfilled ps280k vs ω at t0 = 160oc 63 the journal of engineering research vol. 6, no. 1 (2009) 59-65 tan δ is plotted in fig. 6 as a function of frequency for filled and unfilled ps. tan δ measures the imperfection in the elasticity and it indicates the relative degree of viscous to elastic dissipation of the material. therefore, tan δ is called the loss factor. figure 6 shows a decrease in tan δ with the frequency followed by increasing again and the minimum of tan δ is in the range of about 25 to 80 radian/s for both of filled and unfilled ps. the minimum in tan δ reflects the minimum in loss energy per cycle of the sinusoidal deformation, which is related to the rubbery regime where the perfection of the elasticity due to the entanglements of the polymer chains. tan δ is shown to increase by the addition of the fillers as seen in fig. 6. 10-1 100 101 102 103 100 ps280k+7% filler pure ps280k ta n δ ω rad/s figure 6. master curve of tan δ for filled and unfilled ps280k vs ω at t0 = 160oc 10-3 10-2 10-1 100 101 102 103 102 103 104 105 106 ps280k+7% filler pure ps280k η * p a .s ω rad/s figure 7. master curve of η* for filled and unfilled ps280k vs ω at t0 = 160oc the complex viscosity η* is plotted versus ω in fig. 7 and it is related to the dynamic viscosity, η' as, η* = η' iη''. η* decreases monotonically with increasing ω and falls by many orders of magnitude as shown in fig. 7. at very low frequencies, particularly in the case o f unfilled ps, η* approaches the zero-shear viscosity, ç0 (newtonian viscosity branch) where the viscosity is independent on the frequency and ç0 can be obtained from g'' and ω as, η0 = ( ) 0 g lim ω ω ω→ ′′ (mahmoud halim abdel-goad, a., 2000) . figure 7 shows an increase in the viscosity for 7% fiber -filled ps compared to unfilled ps, particularly in the newtonian viscosity branch. since the values of η0 are about 2.8 ×10 5 and 1.1×105 pa.s for ps composite and neat ps, respectively. this increase in η* by the incorporation of 7% rss may be because the formation of a layer of immobilized polymer chains around the fiber. that is generally as described by einstein equation, which predicts the viscosity of a newtonian fluid containing a very dilute suspension of rigid spheres as, η =η1 (1+ ke × φ) where η is the viscosity of the mixture, η1 is the viscosity of the suspending liquid and φ is the volume fraction of fillers and ke is the einstein coefficient which for spherical particles is 2.5 and vary according to the particle shape and orie ntation (hornsby, p.r., 1999) shows in fig. 7, the presence of fibers in viscous ps melts not only increases its viscosity but also influences its shear rate dependency which alters the beginning of frequencies -independent . where the beginning of frequencies -independent is shifted by the incorporation of 7% rss filler from 0.032 (in the case for neat ps) to 0.057 radian/s (in the case of ps composite) as shown in fig. 7. 10-3 10-2 10-1 100 101 102 103 10-4 10-3 10-2 ps280k+7% filler pure ps280k t o rq ue , n m ω rad/s figure 8. master curve of the torque for filled and unfilled ps280k vs ω at t0 = 160oc 64 the journal of engineering research vol. 6, no. 1 (2009) 59-65 the torque is investigated for the samples of neat ps and ps composite and logarithmically plotted as a function of the frequency as shown in fig. 8. in this figure the torque increases monotonically by many orders of magnitude with increasing up to about 4 radian/s then increases gradually. and the value of the torque for ps composite is higher than this of neat ps. as shown in fig. 8, at 100 radian/s the torque increased from 4.7 10-3 to 7 10-3 nm. this because of the increase in the strength and stiffness of the ps by the addition of the rss since the formation of a polymer matrix-fibers network consisted of polymer chains and fibers. the shear re laxation modulus g(t) for fiber -filled and unfilled ps is plotted against time with logarithmic scales in fig. 9. the modulus g(t) defined as the relaxation stress/strain ratio at constant deformation. at intermediate times g(t) flattens somewhat at a le vel which is associated with the average spacing between entanglement coupling points. the height of g(t) at the intermediate time (rubbery zone) is the plateau modulus, ogn . in this study ogn was evaluated from the rubbery zone in g(t) to be about 0.23 and 0.18 mpa for 7% fiber filled -ps and unfilled ps, respectively as shown in fig. 9. and generally fig. 9 shows higher g(t) for fiber -filled samples than this in the cases of unfilled sample. this again evidence f or the high stiffness and strength of composites compared to neat ps. the width of the rubber -like regime for ps composite is observed to be longer than this of neat ps. it may be because of the formation of the temporary entanglements between the cellulos ic fiber and the polymer chains and that is proved too at the longest relaxation time. at long times g(t) falls sharply and approaches the longest relaxation time, τd .τd is defined as the time required to leave the chain completely its tube (the constrain t of their surrounding chains) by reptation or snake -like motion. τd is enlarged for fiber -filled ps, since it increases from around 1.1 second to 4 seconds by the addition of 7% filler as shown in fig . 9. that is due to the resistance of the fiber for t he mobility of the chain due to the formation of the temporary entanglements. therefore , the time needed for the chains to move within the constraint of their surrounding molecules is long. the observations in fig. 9 are found for the shear creep stress, j(t) in fig. 10 , as the logarithmic plots of j(t) have roughly the appearance of mirror image of g(t) plot reflected in the time axis. in fig. 10 j(t) for neat ps and 7% fiber -filled ps are plotted as a function of time. this figure indicates that j(t) increases with time and the effect of the rss addition is clear since values of j(t) at 0.1 second are 1.14 ×10-5 and 7.94 ×10-6 pa-1 for filled and unfilled ps, respectively , as explained above due to the networks formation. the relaxation spectrum h (fig. 11) and retardation spectrum l (fig. 12) are logarithmically plotted versus time. they refer to the deformation in shear. besides, they are useful qualitatively in gauging the distribution of relaxation and retardation mechanisms, respectively in different regions of the time scale. in polymers a broad spectrum of relaxation times exists due to the large freedom of polymer chain configurations. these figures indicate the effect of the addition of fiber to the ps, since the increase in the strength and stiffness of ps by the addition of rss reflects in relaxation and retardat ion spectrum as shown in fig. 11 and 12. since at 0.01 second h(t) increases from about 1.6 ×104 to 3.7×104 pa and l(t) changes from 1 ×10-6 to 2×10-6 pa-1 by the incorporation of 7% filler to the sufficiently high molecular weight ps. 10-3 10-2 10-1 100 101 102 103 103 104 105 106 τd ~τd ps280k+7% filler pure ps280k g (t) p a time, s figure 9. master curve of g(t) for filled and unfilled ps280k vs time at to = 160oc 10-4 10-3 10-2 10-1 100 101 102 101 102 103 104 105 ps280k+7iller pure ps280k h (t) p a time, s figure 10. master curve of j(t) for filled and unfilled ps280k vs time at to = 160oc 10-2 10-1 100 101 10-6 10-5 10-4 pure ps280k ps280k+7% filler j( t) p a -1 time, s figure 11. master curve of h(t) for filled and unfilled ps280k vs time at to = 160oc 65 the journal of engineering research vol. 6, no. 1 (2009) 59-65 4. conclusions the viscoelastic properties are improved by the addition of residual softwood sawdust (rss) and their values rise with filler loading. therefore, fiber-filled composites have much potential for applications of environmental friendly plastics owing to their strength and stiffness. acknowledgments the financial support by the international bureau in germany, helpful discussions of dr. w. pyckhouthintzen (iff,fzj, research center juelich, germany) are greatly acknowledged. references angles, m.n., salvado, j. and dufresne, a., 1999, " j. of applied polymer science," vol. 74, pp. 1962-1977. bledzki, a.k., reihmane, s. and gassan, j., 1996, "j. of applied polymer science," vol. 59, pp. 1329-1336. fang, z. and hu, q., 1999, "die angewandte makromolekulare chemie," vol. 265(nr.4474), pp. 1-4. felix, j.m. and gatenholm, p., 1991, "j. of applied polymer science," vol. 42, pp. 609-620. ferry, j.d., 1980, "viscoelastic properties of polymers", 3rd ed. wiley, new york. gatenholm, p., bertilsson, h. and mathiasson, a., 1993, " j. of applied polymer science," vol. 49, pp. 197-208. george, j., bhagawan, s.s., prabhakaran, n. and thomas, s., 1995, "j. of applied polymer science," vol. .57, pp. 843-854. heinrich, g. and klüppel, m., 2002, "advances in polymer science," vol. 160, pp. 1-29. hornsby, p.r., 1999, " advances in polymer science," vol. 139, pp. 1-213. hu, x., et al. 2003, "macromolecules," vol. 36, pp. 823829. kim, j.i., ryu, s.h. and y.w.chang., 2000. "j.of applied polymer science," vol. 77, pp. 2595-2602. lee, k.m. and han, c.d., 2003, " polymer," vp;. 44, pp. 4573-4588. lowys, m.p., desbrieres, j. and rinaudo, m., 2000, " food hydrocolloids," vol. 15, pp. 25-32. mahmoud a.-halim abdel-goad, 2000, phd thesis, muenster university, germany. marchant, d. and jayaraman, k., 2002, ind.eng.chem.res., vol. 41, pp. 6402-6408. mcginness, g.b. and bradaigh, c.m.o., 1997, "j.nonnewtonian fluid mech," vol. 73, pp. 1-28. mitsuishi, k., 1997, "die angewandte makromolekulare chemie," vol. 248, pp. 73-83. mohanty, a.k., misra, m. and hinrichsen, h., 2000, " macromol.mater.eng," vol. 276/277, pp. 1-24. nam, j.d. and sefferis, j.c., 1999, " j.of polymer science:part b:polymer physics," vol. 37, pp. 907918. pearson, d., fetters, l. and grassley, w., 1994, " macromolecules," vol. 27, pp. 711-719. ray, s.s., maiti, p., okamoto, m., yamada, k. and ueda, k., 2002, " macromolecules," vol. 35, pp. 3104-3110. roy, d., bhowmick, a.k. and de, s.k., 1993, " j. of applied polymer science," vol.49, pp. 263-273. saha, a.k., das, s., bhatta, d. and mitra, b.c., 1999, " j. of applied polymer science," vol. 71, pp. 1505-1513. solomon, m.j., et al. 2001, "maromolecules," vol. 34, pp. 1864-1872. swerin, c., 1998, "colloids and surfaces a: physicochemical and engineering aspects," vol. 133, pp. 279-294. tavman, j.h., 1996, " j.of applied polymer science," vol. 62, pp. 2161-2167. wang, x., wu, o., dong, j., hu, y. and qi, z., 2002, " j.of applied polymer science," vol. 85, pp. .29132921. zhang. q. and archer, l.a., 2002, "langmuir," vol. 18, pp. 10435-10442. 10-4 10-3 10-2 10-1 100 101 102 10-9 10-8 10-7 10-6 10-5 pure ps280k ps280k+7% filler l( t) p a -1 time, s figure 11. master curve of h(t) for filled and unfilled ps280k vs time at to = 160oc mech061224.qxd vol. 5, no.1 (2008) 62-70ring researchrnal of engineethe jou 1. introduction contact phenomena involving deformable bodies thrive in industry and everyday life. many authors have considered the mechanical system vibration due to surface interaction and examine the surface theories with the presence of friction (brockley, et al. 1970; soom and kim, 1983; ibrahim, and rivin, 1994; ibrahim, 1994; aronov et al. ______________________________________ corresponding author’s e-mail: jdabdo@squ.edu.om 1984a; aronov, et al. 1984b; aronov, et al. 1984c; tan and roger, 1998; tworzydlo, et al. 1992; tzou, et al. 1998). in contrast to the earlier published research, the classical notion of friction coefficient is entirely abandoned in this work. instead, friction is obtained as a result of system's dynamic response that includes time-dependent forces at the contact. numerous works have been devoted to study the friction-induced vibration. for ease of setup and interpretation an idealized physical system consisting of a mass sliding on a moving belt has been considered very often identification of friction/vibration interaction between solids j. a. abdo*1 and n. al-rawahi *1 department of mechanical & industrial engineering, college of engineering, sultan qaboos university, p.o. box 33, al-khoud 123, oman received 24 december 2006; accepted 1 april 2007 abstract: dry-friction forces have been shown to depend not only on the characteristics of the surface in contact but also on the dynamic interaction of the contacting bodies. a viscoelastic mathematical model that accounts for the interaction at micro-scale of rough surfaces is developed. the mathematical formulation relates the tribological events at microscopic and macroscopic scales vibration response of a "mass on moving belt". the viscoelastic properties are presented by combining loss modulus with young's modulus to obtain a differential operator on the interference, reminiscent of the kelvin-voigt model. the analysis of the system establishes the relation between friction force and speed and supports observed behavior of many systems with friction. the derivations do not rely on a phenomenological account of friction, which requires a presumed friction coefficient. instead the friction force is accounted for as a result of interaction of the rough surfaces. this has led to a set of nonlinear ordinary differential equations that directly relate the vibration of the system to the surface parameters. it is shown that, as a result of coupling of the macrosystem's dynamics and contact, there are combinations of parameters at microand macroscale that yield negative slope in friction force/sliding speed relation, a well known source of dynamic instability. keywords: friction/vibration interaction, dynamic interaction of surfaces áñ∏°üdg ogƒÿg úh rgõàge’g h ∑éµàme’g úh πyéøà∏d …~«∏≤j òz ∞°uh h~ñy π«ªl1*»mghôdg π«ñf h áá°°uuóóÿÿgg…ƒàëj »°véjq êpƒ‰ qƒw .á°ùeóàÿg ωé°ùl’g úh »µ«eéæj~dg πyéøàdg ≈∏y é° jg øµdh á°ùeóàÿg í£°s’g ¢üfé°ün ≈∏y §≤a ~ªà©j ’ ±é÷g ∑éµàm ’g iƒb ¿g ~lh : úh »∏yéøàdg rgõàg’g §hôj »°véjôdg êpƒªædg .ájô¡ûg oé©h’g äéjƒà°ùe ‘ áæ°ûÿg í£°s’g πyéøj qéñày’g úy ‘ ònéj å«ëh á°ùeóàÿg í£°s’g áfƒ«dh álhõd òkéj ≈∏y ωgõm ≈∏y áyƒ°vƒe á∏àµd 'á≤«b~dg oé©h’gh ájô¡ûg oé©h’g äéjƒà°ùe ‘ (â«jõàdg h πcéàdg h ∑éµàm’g) -»lƒdƒñjgôj é«ñ°ùf ácôëàÿg í£°s’g πyéøj º∏©h á≤∏©àÿg πegƒ©dg ) 'èæj' πeé©eh ~≤ødg πeé©e êéeoeg ≥jôw øy éæg ¬∏ª©à°ùe áfƒ«∏dgh álhõ∏dg ¢üfé°ün .'∑ôëàeyoung's modulusωg~îà°séh πng~à∏d ¥é≤à°tg πeé©e ≈∏y ∫ƒ°üë∏d ( ) äƒéjéa øø∏c êpƒ‰voigtkelvin-’ êpƒªædg ôjƒ£j . ∑éµàm’g ™e ᪶fcg i~©d ßmóÿg ∑ƒ∏°ùdg ºy~jh áyô°ùdgh ∑éµàm’g iƒb úh ábó©dg ¢ù°sdƒj ωé¶ædg π«∏– .( ßmóÿg ∑ƒ∏°ùdg ºy~jh áyô°ùdg h ∑éµàm’g iƒb ™e πeé©àdg ºàj áféa ∂dp øy é°vƒy.áæ°ûÿg í£°s’g úh πyéøà∏d áé«àæc ∑éµàm’g iƒb ™e πeé©à∏d áj~«∏≤àdg á≤jô£dg ≈∏y ~ªà©j áé«àf ¬fg qé¡xeg ” .í£°s’g ¢üfé°ünh ωé¶ædg rgõàgg úh ô°téñe πµ°ûh §hôj »àdgh á«£nódg ájoé©dg á«∏°véøàdg ä’oé©ÿg øe áyƒª› gòg øy èàf ~bh ∑éµàm’g ™e ᪶fg i~©d ödé°s π«e é¡æy èàæj »àdg ^á≤«b~dg h ájô¡ûg oé©h’g äéjƒà°ùe ‘ ¢üfé°üÿg øe áyƒª› ~lƒj áféa ,ωé¶æ∏d á≤«b~dg oé©h’g iƒà°ùe ‘ ᫵«eéæj~dg h ¢ùeóàdg úh ¿gîbód .»µ«eéæj~dg qgô≤à°s’g ω~©d ±hô©e q~°üe ƒgh ,¥’õf’g áyô°sh ∑éµàm’g iƒb úh ábó©dg ‘ ««mmééààøøÿÿgg ääggooôôøøÿÿgg .á°ùeóàÿg ωé°ùl’g úh »µ«eéæj~dg πyéøàdg rgõàg’g h ∑éµàm’g úh πyéøàdg :á 63 vol. 5, no.1 (2008) 62-70ring researchrnal of engineethe jou (panovko, et al. 1965; nayfeh and mook, 1979; mitropolskii and nguyen, 1997; popp, 1991; tondl, 1991; mcmillan, 1997 and thomsen and fidlin, 2002), and it will be in this present study. the physical contact model and governing equations with the inclusion of the viscoelastic properties are presented in this paper to study the dynamic effects of the surface irregularities as a result of mechanical interactions. several analytical or computational methods have been applied to solve the dry friction problems to increase the understanding of friction and vibration phenomenon in mechanical systems. some of the elastic, elastic-plastic, and plastic contact models are reformulated to estimate a normal and tangential contact force, contact area and contact stiffness. the elastic and the elastic-plastic models developed by abdo and farhang (2005) and by abdo and shamseldin (2005) in addition to the research results pertinent to the estimation of contact stiffness, abdo (2006) and abdo (2005) are used in this study along with the addition of the viscoelastic effects of contact. in this work, the viscoelastic properties are presented by combining loss modulus with young's modulus to obtain a differential operator on the interference, reminiscent of the kelvin-voigt model. there is no assumption of friction/velocity relation made in the formulation of the governing differential equations. the elastic and ratedependent contact force components between rough surfaces are presented to obtain the equations governing the vibration response of mass-spring-damper system. 2. the dynamic model the basic assumptions of the elastic model along with the shoulder-to-shoulder contact model (abdo and farhang 2005), and abdo 2005 and (abdo and shamseldin, 2005) are used in this study to account for the elastic contribution of rough surfaces. elastic contact model has viewed the surface irregularities as identical asperities with spherical summits differing only in their heights that can be measured from the mean plane separation of two rough surfaces. the misaligned asperities permit shoulder-to-shoulder contact. when two solid bodies are brought into contact, the real contact area will only be a fraction of the apparent macroscopic contact area. the introduction of contact slope allowed the consideration of normal and tangential components of the force. referring to the free body diagram shown in fig. (1), the governing equation of the normal motion can be expressed as (1) the governing equation of the tangential motion can be expressed as (2) no where fna is the normal force applied to mass m. fnc is the normal contact force between the two contacting rough surfaces. it is the scalar sum of the resisting and assisting viscoelastic contact components (3) in this section, we are interested in developing a mathematical f ormulation relating the tribological events at microscopic and the macroscopic scales vibration response of a mass on a moving belt. an example of the physical system is shown in fig. 1 which consists of a mass m on a belt that moves at constant speed, y . the mass is a rigid body, at time t positioned at x(t) in a fixed frame of coordinates. it is subjected to a normal applied static load fna, linear spring-loading force kx, plus damping force cdx / dt. the h0 indicates the ini tial mean plane separation of the two contacting rough surfaces corresponding to the normal applied static load, fna. the model doesn’t assume a friction coefficient; therefore, friction is instead obtained as a result of system’s dynamic response that in cludes time -dependent forces at the contact. the tangential -to-normal contact load ratio, the classical definition of the friction coefficient, is shown to depend on system response that in turn is dependent on both the structural and contact characteristics. fk fc fnc fnc ft ft y k fna x m h0 c y m fna figure 1. schematics of block on a moving platform and its free-body diagram ho nc namh f f− = − tmx k x c x f+ + = nc ne ne nv nvf f f f f − + − += + + + the terms nef − and nef + denote, respectively, the resisting and assisting elastic normal contacts at the asperity interference slopes (bengisu et al. 1997; bengisu et al. 1999), as shown in fig. 2. it remains a challenge to determine whether there is a net elastic influence on slidi ng surfaces. as shown in fig. 2 , when in static equilibrium, without the presence of an applied tangential force, the contact force due to negative (4) 2.1 approximation of elastic contact forces the elastic normal contact force per unit area is the sum of the resisting and assisting elastic normal contacts at the asperity interference slopes (5) the analytical solution of the force in eq. (5) may be approximated as the product of the two terms as follows: fne = dne fne (6) where dne and fne, are a function of surface properties and a function of dimensionless mean plan separation between two rough surfaces, respectively. two integrations are involved in the contact force calculations. the function dne is defined in abdo (2006) as (7) (8) where h is the normalized mean plane separation. the fne function consists of double integral function over r and s domains. the integration of the contact function occurs contact slope will be equal to that due to the positive slope; therefore, the net tangential force on a surface is zero. in the presence of an applied tangential force, the equilibrium condition dictates that the net tangential force to be the equilibrating force. bengisu and akay (bengisu et al. 1997) for example assume that the net elastic tangential force, ie. those due to positive slopes minus those due to negative slopes, is negligible. we have chosen instead to include the factor of the friction due to the positive or resisting slopes. in the presence of an applied tangential load that is within the bounds of the static friction force, the statistical summation of asperity contacts corresponding to positive slopes is no longer equal to that for contacts with negative slopes. according to (bengisu and akay 1997) the assisting and resisting contacts balance in the case of sliding, with the expected conclusion that friction force is a result of mainly adhesive forces. it is expected that as tangential applied load is increased an d the tangential contact loads due to positive -slope contacts continue to rise while those due to negative -slope contacts decrease. in the limiting case, or the so -called maximum static friction, it is postulated that the load is completely supported by p ositive -slope contacts. the views forwarded by bengisu and akay and postulated herein perhaps represent two extremes of what may actually take place in a frictional contact. in the former, the friction is assumed to be mainly due to the adhesive forces whereas in the latter the friction force is assumed to be borne solely by mechanical interaction of roughness. in this way we can obtain a generic formulation that can be easily adapted to any assumption. the scenario for contact of asperities when a maximum force occurs is shown in fig. 3 . likewise, nvf − and nvf + are those fo rces due to viscous effects, ie. the rate dependent effect of interference at positive and negative asperity slopes. the elastic and viscous terms ar e a consequence of a viscoelastic contact, but they are defined separately here for the purpose of implementing the approximate analytical description of each term derived later. the tangential force, ft, in eq. (2) exerted through the frictional contact o n mass m. it is the scalar sum of the elastic interactions of the collection through both resisting and assisting viscoelastic contact components . t te te tv tvf f f f f − + − += + + + ne ne nef f f − += + ae tef retef ae nef re nef aef ref slopepositive slopenegative directionsliding figure 2. asperity shoulder contact at positive and negative interference slopes ae tef a e nef a ef d ir ectio nslid in g figure 3. maximum possible static friction of asperity ç ð â ó'ned e .= 2 48 05 3 â , ç , 'e , ó are the normalized radius of curvature of asperity mean summit, asperity density, composite modulus of the material and standard deviation of asperity height distribution, respectively. the integral part of the normal force per unit area is ( ) β −∞ ∫ ∫ ⎛ ⎞⎟⎜ ⎟⎛ ⎞⎜ ⎟⎜ ⎟⎜ ⎟⎟⎜ ⎜ ⎟⎟⎜ ⎟⎜⎛ ⎞ ⎟⎜ ⎟⎜ ⎟⎟ ⎟⎜ ⎜ ⎜ ⎟⎟ ⎟⎜ ⎜ ⎜= − − φ⎟⎟ ⎟⎜ ⎜ ⎟⎜⎟ ⎟⎜ ⎜ ⎟⎜β⎟⎜ ⎟⎝ ⎠⎜ ⎟⎜ ⎟⎟⎜ +⎜ ⎟⎟⎜ ⎟⎜ ⎟⎜ ⎟⎟⎜ β⎝ ⎠ ⎟⎜ ⎟⎜ ⎟⎜⎝ ⎠ 5 43 22 2 2 0 2 1 4 1 4 ( s h ) ne h r f s h rdr s ds r 64 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 over the range of asperity-offset misalignment, referred to as r-integration. this integration is to account for the total expected interference. the height sum of the rough surface, s, is represented by a gaussian probability density function, φ(s). hence, in relating micron-scale events to the resulting macro-scale expectations, two integrations must be performed: firstly, the integration over possible asperity offset and secondly the statistical integration over all possible asperity height sums, s-integration, by implementing a gaussian probability function. the resulting integral forms describing an elastic contact of rough surfaces are analytically non-integrable. hence, their solution must be obtained numerically. this fact imposes serious limitations on their use in a dynamic analysis of a mechanical system. to circumvent this shortcoming, we developed approximate analytical functions for viscoelastic contact of rough surfaces. using an interactive search involving dynamic plotting of the function, fne is approximated using a considerably simpler function. the result is an approximate function that depends on two independent variables, β, h. using the dynamic plotting technique, fne can be approximated by a product of exponential function of h and a polynomial function of β. the piecewise analytical solution of fne is obtained at three different intervals as shown in the following equations: (9) (10) (11) the analytical solution of elastic tangential contact per unit area is developed. the analytical approximation can be described as the product of two functions: a function of surface properties dte and a function of normalized mean plane separation, h, between two rough surfaces, fte, as follows: fte = dte fte (12) note that the constants dne and dte are identical. fte is expressed as (13) using dynamic fitting technique, the piecewise analytical solution of the function fte is approximated in closedequations for two intervals of h as (14) (15) 2.2 approximation of viscoelastic contact forces therefore, a modified form of hertz contact results in which the modulus operator is used instead of the usual elastic modulus (16) where ω the interference between the two asperities and is derived in (abdo 2005; and abdo et al. 2005) as (17) substitution of modulus operator in the hertz equation of contact and subsequent expansion revealed two forces. one is an elastic force having the form encountered previously, and a second is viscous force defined by (18) the viscoelastic properties are presented by combining loss modulus with young's modulus to obtain a differential operator on the interference, reminiscent of the kelvin-voigt model. to develop an expression for the normal and tangential components of the viscous forces, the formulations of low to moderate contact force between interacting asperities developed in abdo (2005) and verified in abdo (2006) are considered. the interacting asperities are not assumed to occur only at the asperity peaks, thus allowing the possibility of oblique contacts wherein the local contact surfaces are no longer parallel to the mean planes of the mating surfaces. the normal component of the viscous force can be written so as to include the oblique contact ( ) ( ) hnef . . h . . e h . −⎡ ⎤= − + β + − + β⎢ ⎥⎣ ⎦ ≤ ≤ 3 1 0399 18 0389 118 0 15 ( ) . hnef . . h e . h . −⎡ ⎤= − + β⎢ ⎥⎣ ⎦ ≤ ≤ 38 2 199 9538 15 30 ( ) . hnef . . h e . h . −⎡ ⎤= − + β⎢ ⎥⎣ ⎦ ≤ ≤ 49 3 11496 32001 30 45 ( ) ( s h ) te h r r f s h dr s ds r β −∞ ∫ ∫ ⎛ ⎞⎟⎜ ⎟⎛ ⎞⎜ ⎟⎜ ⎟⎜ ⎟⎟⎜ ⎜ ⎟⎟⎜ ⎟⎜⎛ ⎞ ⎟⎜ ⎟⎜ ⎟⎟ ⎟⎜ ⎜ ⎜ ⎟⎟ ⎟⎜ ⎜ ⎜= − − φ⎟⎟ ⎟⎜ ⎜ ⎟⎜⎟ ⎟⎜ ⎜ ⎟⎜β β⎟⎜ ⎟⎝ ⎠⎜ ⎟⎜ ⎟⎟⎜ +⎜ ⎟⎟⎜ ⎟⎜ ⎟⎜ ⎟⎟⎜ β⎝ ⎠ ⎟⎜ ⎟⎜ ⎟⎜⎝ ⎠ 5 43 2 22 2 2 0 2 1 4 2 1 4 ( ) ( ) . h te . . . h f e . . . h . − ⎡ ⎤+ β− β⎢ ⎥ ⎢ ⎥= ⎢ ⎥+ + β− β⎢ ⎥⎣ ⎦ ≤ ≤ 2 35 1 2 53001 02396 0000598 1694 00847 000029 0 27 ( ) . htef . . . h e . h . −⎡ ⎤= + β− β⎢ ⎥ ⎣ ⎦ < ≤ 2 48 2 19279 91601 002498 27 45 the viscoelastic model is developed using the kelvin-voigt’s constitutive equation relating stress to strain. in kelvin -voigt model, the composite modulus of the material ′e and the loss modulus ev are used to introduce the viscoelastic modulus differential operator e e tυ ∂ ′ + ∂ . υ ⎛ ⎞∂ ⎟⎜ ′= β + ω⎟⎜ ⎟⎟⎜⎝ ⎠∂ 3 24 3 f e e t r s h r ⎛ ⎞⎟⎜ ⎟⎜ω = − − ⎟⎜ ⎟⎜ β⎟⎜⎝ ⎠ + β 2 2 2 1 2 1 4 υ υ ⎛ ⎞⎟⎜∂ ⎟⎜ ⎟= β ω⎜ ⎟⎜ ⎟∂ ⎜ ⎟⎟⎜⎝ ⎠ 3 24 3 f e t 65 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 (19) (20) since the interference, ω, depends on the normalized mean plane separation, h, and the tangential offset between two asperities, r, its derivative with respect to time may be written as (21) utilizing eq. (18) then (22) the total expected contact force per unit area introduced in the (greenwood and trip 1971) is employed to develop the total expected viscous normal contact force per unit area as follows (23) utilizing eq. (19) through eq. (22), then eq. (23) may be written as (24) (25) (26) fnvh and fnvr are the analytical functions for the normalized viscous normal force. using the dynamic plotting technique, the piecewise analytical solutions for fnvh and fnvr can be approximated by a product of exponential function of h and a polynomial function of β as follows (27) (28) in a similar manner, the solution of viscous tangential contact force is developed. the viscoelastic tangential force is (29) the total expected viscous tangential contact force may be written as (30) utilizing eqs. (20) through (22) and eq. (29), then eq. (30) may be written as (31) + − υ α ∂ = + = β ω ∂+ 1 3 2 2 2 4 1 3 1 nv nv vf f f e ts where αs is the interference slope defined by two points of interactions with respect to the mean plane. to simplify eq . (19), we co nsider first the derivation of the rate of change of the interference function, ω ∂ ω = ωω ∂ 3 1 2 23 2 ( ) t ∂ ω ∂ ω ω = + ∂ ∂ h r h r ( )s hr r r h r r ⎛ ⎞⎟⎜ ⎟⎜⎛ ⎞ ⎛ ⎞ ⎟⎜ − ⎟ ⎟ ⎟⎜ ⎜⎜ ⎟ ⎟ ⎟⎜ ⎜⎜ω = − − + + +⎟ ⎟ ⎟⎜ ⎜⎟ ⎟⎜ ⎟⎜ ⎜⎟ ⎟ ⎟β⎜β ⎟ ⎟⎜ ⎜β β⎝ ⎠ ⎝ ⎠ ⎟⎜β + ⎟⎜ ⎟⎜⎝ ⎠ β 3 2 2 2 2 22 2 2 1 2 1 4 4 44 r e n v n v s r f fπη ∫ ∫= 22 (ω, r) φ (s) rdrds ( )β −∞ ∫ ∫ − π ′= η β η σ π ⎛ ⎞⎟⎜ ⎟⎜ + ⎟⎜ ⎟⎜ ⎟⎜ β⎝ ⎠ ⎛ ⎞⎟⎜ ⎟⎜ − − ω⎟⎜ ⎟⎜ β⎟⎜⎝ ⎠ 2 2 4 3 4 20 2 1 2 2 2 2 4 1 2 1 4 4 s h nv v h s f e r r s h e rdrds where η = ′ v v e e is the loss factor for the friction film or material and σ is the standard deviation of the surface heights. the solution for the viscous contact normal force cons ists of two parts, involving the rate change of the normalized mean plane separation, h , and the rate of change of the normalized tangential misalignment, ie. the sliding speed, = −r x y . the total expected viscous contact nor mal force, due to the interaction on both shoulders, is ( ) nv nv nv nv nvh nvr f f f d f h f r − += + = + where nvd is a constant term associated with h and r that depends on the surface parameters and material properties. it is define as follows ′= − π βη η σ2 44 05nv vd . e ( )( ) ( ) . h nvh nvh . h nvh f . . h . . e , f h . f . . h e . h . − − ⎧⎪ == − + β + + β⎪⎪⎪= ≤ ≤⎨⎪⎪ = − + β ≤ ≤⎪⎪⎩ 32 1 45 2 3759 5056 0585 1289 0 27 878 20969 27 45 ( ) ( ) . h nvr nvr . h nvr . . . f e , h . . . f h . f . . . h e , . h . − − ⎧ ⎛ ⎞⎪ ⎟⎪ ⎜ + β− β ⎟⎪ ⎜ ⎟=⎪ ⎜ ⎟⎪ ⎜ ⎟⎜⎪ ⎟⎟⎜ + + β− β⎪ ⎝ ⎠⎪⎪= ≤ ≤⎨⎪⎪⎪ = + β− β ≤ ≤⎪⎪⎪⎪⎪⎪⎩ 2 32 1 2 2 46 2 5567 02879 0000804 2154 0095 0000206 0 27 285197 1401 00389 27 45 − + α α ∂ = = β ω ∂+ 1 3 2 2 2 4 3 1 tv tv v s f f e ts tv tvf f −= π η22 (ω, r) φ (s) rdrds ( )s h tv tv v h s f . a e r r r s h e rdrds β −∞ ∫ ∫ − ′= πη β η σ ⎛ ⎞⎟⎜ ⎟⎜ + ⎟⎜ ⎟⎜ ⎟⎜ β⎝ ⎠ ⎛ ⎞ ⎛ ⎞⎟ ⎟⎜ ⎜⎟ ⎟⎜ ⎜− − ω⎟ ⎟⎜ ⎜⎟ ⎟⎜ ⎜β β⎟ ⎟⎜ ⎜⎝ ⎠ ⎝ ⎠ 2 2 4 3 4 20 2 1 2 2 2 2 1 4 05 1 4 4 2 as before, the solution for the viscous contact tangential force per unit area consists of two parts involving h and r . therefore, eq . (31) may be written as 66 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 (32) where dtv is identical to dnv. ftvh and ftvr are the analytical functions for the normalized viscous tangential force. using the dynamic plotting technique, the piecewise analytical solutions for ftvh and ftvr can be approximated by a product of exponential function of h and a polynomial function of β as follows (33) and (34) 2.3 the model analysis and simulation according to the dynamic model described in fig. (1), the dimensionless mean plane separation between two surfaces is (35) (36) (37) eqs. (37) and (38) govern the motion of mass m in the normal and tangential directions. the equations tie the vibration behavior of the system to both structural and contact properties of the friction interface. the dynamic model was developed based on several assumptions to simplify the study. the elastic normal and tangential contact components are due to both positive and negative interference slopes of an asperity as shown in fig. 2 and are added algebraically to provide their combined effect. there may or may not be a net resisting elastic tangential force when the two surfaces in relative sliding contact. when this force exists, it is described as a fraction of the maximum static friction force that is due to only positive slopes. the viscoelastic tangential forces are only produced by the resisting contacts. 3. results ( )tv tv tvh tvrf d f h f r= +2 ( ) ( ) . . . . hf e , h .tvh h . . . ftvh . hf . . . h e , . h .tvh ⎧ ⎛ ⎞⎪ ⎟⎪ ⎜ + β− β ⎟⎪ ⎜ −⎟⎪ ⎜= ≤ ≤⎟⎪ ⎜ ⎟⎪ ⎜ ⎟⎜⎪ ⎟⎟⎜ + + β− β⎪ ⎝ ⎠⎪⎪= ⎨ ⎪⎪ −⎪ = + β− β ≤ ≤⎪⎪⎪⎪ 26879 03188 0000764 33 0 271 2179 011 0000333 2 4627995 13598 00348 27 452 ( ) . htvr tvr . h tvr f . . h e h .f f . h e . h . − − ⎧⎪ = + ≤ ≤⎪= ⎨ ⎪ = ≤ ≤⎪⎩ 35 1 46 2 0708 3195 0 27 9798 27 45 = −oh h h where oh is the initial mean plane separation of the contacting rough surfaces due to an applied normal force and ( )ˆh h σ is the dimensionless displacement in the normal direction. ( )na ne ne nv nvh nvrmh f d f d f h f r= − − + ( )te te tv tvh tvrmx kx cx d f d f h f r= − − + + + when h is negative, all viscoelastic components exist and are additive for normal motion. the responses of the dynamic model are studied at different belt speeds and a dimensionless average radius of a summit curvature, ,β to investigate the roughness effect. the equation of motion was solved using a matlab simulink program. unlike in other publications the friction force is accounted for as a result of interaction of the rough surfaces and response of the system. therefore, the tangential -to-normal contact ratio is a time -dependent variable based on structure and contact characteristics of the system. table 1 summarizes the parameters used. the simulations include moderate applied load condition corresponding to a high initial separation, h0 = 3.5. the total range of dimensionless mean plane separation is from 0 to 4.5 for contact force functions. the β values considered in the simulatio n are 100 and 140 to ensure low plasticity index so that the asperity deformation will remain predominantly elastic. for cases performed but in order to reduce the length of the article only the results of case 1 will be presented. ( )mµσ 15 kn (n/m) 3.5x 108 ( )2−mmη 400 c (n.s/m) 21166 vη 0.001 0h 3.5 e′ (gpa) 113.1 β 100 m (kg) 5 y (m/s) 0.015 case 1 fna=0.002 fc table 1. parameters from a steel sample the analytical solutions of normal to tangential contact forces presented in this work are applicable to dimensionless mean plane separation, h = 0 to 4.6 and the dimensionless average radius of asperity summit curvature, β = 10 to 140. t he approximate analytical solutions developed included the normal and tangential contact forces due to viscoelastic effects. the purely elastic term or rate dependent term is presented as a product of two functions. the first corresponds to a constant te rm for a given surface and the second depends on the separation of mean surfaces as well as tangential asperity offset. it suffices to evaluate the approximate equations with respect to these functions. therefore the evaluations will only involve ne te nvh tvh nvrf , f , f , f , f and tvrf , not the constant terms dne, dte, dnv, and dtv. the comparison between the original integral function of the elastic normal and tangential contact forces, nef , tef as depicte d in eqs. 8 and 13, respectively, and their piecewise analytical solutions ne nef , f1 2 and nef 3 and te tef , f1 2 and tef 3 for the ranges of h = 0 to 1.5, h = 1.5 to 3.0 and h = 3.0 to 4.5 at diff erent values of β are performed. the comparisons at β = 10, 40, 75, 100, 125, and 140 67 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 the results for case (1) are depicted in fig. 7 through fig. 10. figure 7a shows the trajectory of tangential motion, tangential speed versus tangential displacement of the mass m. the zoom-in view (fig. 7b) shows decreasing spiral, which represents a mass response of decaying oscillation corresponding to a stable behavior. the trajectories in fig. 7 and the time history in fig. 8 indicate an initial decay followed by a growth in vibration up to a level at which the vibration is sustained. in case of the normal motion of the mass the structural decay is absent consistent with the system's depiction in fig. 1 in which no structural stiffness or damping exists for the vibration of mass m along the normal to the platform. as shown in fig. 9(a,b) the vibration is grown and sustained at a value of 0.005 mm. the friction coefficient defined as the tangential to normal friction shown in fig. 10 depicts a steady state value of 0.026 for case 1. 4. conclusions this paper has addressed the dynamic effects of the surface irregularities (roughness) as a result of mechanical interactions. we have exhibited the cross-influence of tribological interactions that occur at the micron-scale and the dynamics of a mechanical system involving macroscale events. the utility of the contact theories in this work was demonstrated through the dynamic analysis of a simple mechanical system. we have considered a mechanical system comprising a spring-damper-mass and a platform in relative tangential and normal movements while maintaining frictional contact between mass and platform rough surfaces. the results pertaining to the prediction of contact forces were used along with the addition of the viscoelastic effects of contact. the approximate analytical solutions were obtained for normal and tangential contact components in viscoelastic contact. the term viscoelastic was separated into components of elastic and rate-dependent interactions. it was shown that the formulation of contact forces allow the inclusion of frictional contact without the necessity of including the friction/velocity relation phenomenologically. no presumption of a friction coefficient has been made. the friction force is accounted for as a result of interaction of the rough surfaces. rather the tangential (friction) force has been a result of the consideration of contact at the micron-scale level between two asperities on the surfaces and the statistical summation of this interaction to obtain the macrolevel expectation functions. this led to a set of nonlinear ordinary differential show good agreements between the original and the approximation functions throughout the ranges of .β the relative percentage errors are between 9% and 10%. the comparisons of the original and viscous functions for the same values of β are also performed. the estimation of viscous the functions shows a relative percentage error between the original and the approximation functions is less than 20%. this is below the inherent uncertainty involved in surface measurements that can climb to as much as 50%. comparisons between the original integral function of the elastic normal contact force, nef , as depicted in eq. 8 and its piecewise analytical solutions, ne nef , f1 2 and nef 3 for the ranges of h = 0 to 1.5, h = 1.5 to 3.0 and h = 3.0 to 4.5, respectively, at β = 100 and 140 are shown in fig. 4. as seen in the figures, the comparisons show good agreement between the original and the approximate functions for β = 100 and 140. the relative percentage error between nef and ne nef , f1 2 and nef 3 is 9% at β = 100 and 10% at β =140. the comparison between the approximate functions tef and its analytical solutions and tvhf and its analytical solutions for the ranges of h = 0 to 1.5, h = 1.5 to 3.0 and h = 3.0 to 4.5, respectively, at β = 100 and 140 are shown in figures 5 and 6. the relative percentage error between the original and the approximation functions for both functions are 17% and 19%, respectively. elastic normal force, fne, at beta= 100 & 140 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 h fn e analytical at beta= 100 original at beta= 100 analytical at beta= 140 original at beta= 140 0 2 4 6 8 1.5 2 2.5 3 h fn e 2 0 0.01 0.02 0.03 0.04 0.05 3 3.5 4 4.5 h fn e 3 figure 4. elastic normal force at β = 100 & 140 elastic tangential force, fte, at beta = 100 & 140 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 h fte analytical at beta= 100 original at beta= 100 analytical at beta= 140 original at beta= 140 0 0.0005 0.001 0.0015 0.002 0.0025 3 3.5 4 4.5 h fte 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 h fte figure 5. elastic tangential force at β = 100 & 140 viscoelastic tangential force, ftvh, with beta = 100 & 140 0 2 4 6 8 10 12 14 0 1 2 3 4 h ftv h analytical at beta= 100 original at beta= 100 analytical at beta= 140 original at beta= 140 0 0.001 0.002 0.003 0.004 0.005 0.006 3 3.5 4 4.5 h fe vh 0 0.1 0.2 0.3 0.4 0.5 0.6 1.5 2 2.5 3 h ftv h figure 6. viscoelastic tangental force at β = 100 & 140 68 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 (a) (b) figure 7 (a, b). trajectory of tangential motion and the zoom-in view, case 1 (a) (b) figure 8 (a, b). tangential speed of the mass and the zoom-in view, case 1 (a) ( b) figure 9 (a, b). normal speed of the mass and the zoom-in view, case 1 figure 10 (a, b). load ratio (friction function), case 1 69 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 equations that directly relate the vibration of the system to the surface parameters, mechanical system parameters and physical parameters. the study has shown the following: references abdo, j. a. and farhang, k., 2005, "elastic-plastic contact model for rough surfaces based on plastic asperity concept," int. j. of non-linear mechanics, vol. 40, pp 495-506. abdo, j. a. and shamseldin, e., 2005, "modeling of contact area, contact force, and contact stiffness of mechanical systems with friction," asme mechanical engineering congress paper no. imece2005-82980, orlando, florida. abdo, j. a., 2005, "experimental technique to study tangential-to-normal contact load ratio," tribology transactions, vol. 48, pp 389403. abdo, j. a., 2006, "modeling of normal and tangential contact stiffness of rough surfaces," int. j. of modeling and simulation, vol. 26(4), pp 295-302. aronov, v., d'souza, a. f., kalpakjian, s. and sharper, i., 1984a, "interaction among friction, wear and system stiffness-part 1: effect of normal load on system stiffness," j. tribol., vol. 106, pp. 54-58. aronov, v., d'souza, a. f., kalpakjian, s. and sharper, i., 1984b "interaction among friction, wear and system stiffness-part 2: vibrations induced by dry friction," j. tribology, v. 106, pp. 59-64. aronov, v., d'souza, a. f., kalpakjian, s. and sharper, i., 1984c "interaction among friction, wear and system stiffness-part 3: wear model," j. tribology, vol. 106, pp. 59-64. bengisu, m. t. and akay, a., 1997, "relation of dry friction to surface roughness," asme j. of trib, vol. 119, pp. 18-25. bengisu, m. t. and akay, a., 1999, "stick-slip oscillations: dynamics of friction and surface roughness," j. of acoust. soc. am., vol. 105, pp. 194-205. brockley, c. a. and ko, p. l., 1970, "the measurement of friction and friction-induced vibration trans. asme pp. 543-549. greenwood, j. a. and tripp, j. h., 1971, "the contact of two rough nominally flat rough surfaces," proc. instn. mech. engrs., vol. 185, pp. 625-633. ibrahim, r. a., 1994, "friction induced vibration, chatter, squeal and chaos, part ii: dynamic and modeling applied mechanics reviews," vol. 47, pp. 227-253. ibrahim, r. a. and rivin, e., 1994, "friction-induced vibration, part i: mechanics of contact and friction applied mechanics reviews," vol. 47, pp. 209-226. mcmillan, j., 1997, "a non-linear model for self-excited vibration," j. sound vibration, vol. 205(3), pp 323335. mitropolskii, y.a. and nguyen, v. d., 1997, "applied sympotic methods in nonlinear oscillation," kluwer, dordrecht. nayfeh, a. h. and mook, d. t., 1979, "nonlinear oscillations," wiley, new york. panovko, y. g. and gubanova, i. i., 1965, "stability and oscillation of elastic systems; paradoxes, fallacies and new concepts," consultants burea, new york. popp, k., 1992, "some model problems showing stickslip motion and chaos, friction-induced vibration, chatter, squeal, and chaos," asme de, vol. 49, asme design engineering division, new york, pp. 1-12. soom, a. and kim, c., 1983, "interactions between dynamic normal and frictional forces during unlubricated sliding," transactions of the asme, j. of lubrication technology, vol. 105, pp. 221-229. tan, x. and roger, r., 1998, "simulation of friction in multi-degree-of-freedom vibration system," asme j. dyn. syst., control, vol. 120, pp. 144-146. thomsen, j. j. and fidlin, a., 2002, "analytical approximations for stick-slip vibration amplitudes," int. j. of non-linear mechanics, vol. 38, pp 389-403. tondl, a., 1991, "quenching of self-excited vibration," elsevier, amsterdam. tworzydlo, w., becker, e. and oden, j., 1992, "numerical modeling of friction-induced vibration and dynamic instabilities," app.l mech. rev. vol. 45(7), pp. 255274. tzou, k., wickert, j. and akay, a., 1998, "in-plane vibration modes of arbitrarily thick disks," asme j. vibr. acout., vol. 120, pp. 384-391. • the tangential trajectories of the four cases showed initial decay dominated by structural damping followed by a growth in vibration up to a level at which vibration is sustained that is due to the interaction of surface asperities of the moving platform and those of the structure mass. • the vibration of the normal direction of the four cases grow to a level at which it is sustained that is due to frictional interaction between the platform surface roughness and that of mass m and the absence of the structural decay. • the low tangential to normal contact load ratio of the four cases is due in part to the relatively low surface roughness in the present study in order to ensure the predominance of elastic interaction at the contact. 70 the journal of engineeeing research vol. 5, no.1 (2008) 62-70 the journal of engineering research (tjer), vol. 19, no. 2, (2022) 163-179 corresponding author's e-mail: ammarm.odeh@gmail.com doi:10.53540/tjer.vol19iss2pp163-179 comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes ammar m. odeh qc auditor, nokia networks, rafiq al hariri ave, amman, jordan abstract: one of the major factors that limit the performance of an optical fiber communication system is dispersion. in order to get a high transmission range with high data rates, techniques must be in place to compensate for the dispersion caused by fiber nonlinearity. dispersion compensating fiber (dcf) and fiber bragg grating (fbg) are the trending dispersion compensation techniques in optical fiber communication. the use of dcf and fbg as a method of dispersion compensation can notably enhance the overall performance of the system. broadening is a function of distance as well as the dispersion parameter (d). the dispersion parameter is given in ps/nm/km and changes from fiber to fiber and also is a function of wavelength. in this paper, we investigate pre-, post-, and symmetrical compensating schemes in three different compensating models: dcf, fbg, and dcf cascaded to fbg. the system performance was evaluated in terms of q-factor and bit error rate (ber) for one optical channel communication system at 120 gbps return to zero (rz) signal launched over single-mode fiber (smf) of 100 km by using optisystem 7.0 software. keywords: bit-error rate, dispersion compensating fibers, dispersion compensation, fiber bragg gratings, qfactor الثانية باستخدام /جيجابت 120قل بصري ذو قناة واحدة بسرعة الن مقارنة أداء طرق تعويض التشتت ة، ومتناظرمؤخرة، ةنظم مسبق عمار عودة على نطاق :الملخص يعد التشتت أحد العوامل الرئيسية التي تحد من أداء نظام اتصاالت األلياف الضوئية. من أجل الحصول تقنيات للتعويض عن التشتت الناجم عن األلياف الالخطية. تعد األلياف المعوضة إرسال عالي بمعدالت بيانات عالية، يجب وضع من تقنيات تعويض التشتت الشائعة في اتصاالت األلياف الضوئية. يمكن أن (fbg) وشبكة األلياف الضوئية (dcf) للتشتت التوسيع دالة للمسافة اعتبر شكل ملحوظ. كطريقة لتعويض التشتت إلى تحسين األداء العام للنظام ب fbg و dcf يؤدي استخدام . وهي أيًضا دالة لطول الموجة أخرىلياف إلى أوتتغير من ps / nm / km يتم إعطاء معلمة التشتت بوحدة .(d) لتشتت لمعامل و و fbg و dcfبدراسة مخططات التعويض السابقة والالحقة والمتناظرة في ثالثة نماذج تعويضية مختلفة: ي هذا البحث قمنا ف dcf المتتالية إلىfbg عامل تم تقييم أداء النظام من حيث .q ( ومعدل خطأ البتber واحد اتصال قناة بصرية ( لنظام ( لمسافة smf( التي تم إطالقها عبر األلياف أحادية الوضع )rzجيجابت في الثانية إلشارة العودة إلى الصفر ) 120بسرعة برمجة. optisystem 7.0كم باستخدام 100 .q؛ معامل معدل الخطأ في البت؛ ألياف تعويض التشتت؛ تعويض التشتت؛ حواجز األلياف الزجاجية الكلمات المفتاحية: comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 164 nomenclature 𝛽 fiber propagation constant, in (radians/meter). γ attenuation between two amplifiers, in (db).  effective nonlinear coefficient, in (rad/km/w) ∆f the bandwidth that measures the nf, in (nm or hz). ʌ fbg grating period, in (μm). λ optical wavelength, in (nm). λb the central wavelength of the reflected spectrum satisfies the bragg condition, in (μm). π math constant = 3.14159265 𝜏 time delay, often in (ns) 𝜔 angular frequency, in (radians/second or hertz). bt system bit rate, in (gbps). c speed of light, in (m/s). ddcf chromatic dispersion coefficient for dcf, in (ps/nm/km). dsmf chromatic dispersion coefficient for smf, in (ps/nm/km). f signal center frequency, in (hz). 𝑘0 free-space wavenumber. ldcf dispersion compensating fiber length, in (km). lsmf single mode fiber length, in (km). n number of amplifiers in the fiber link. nf noise figure of amplifier, in (db). ne the grating effective refractive index. 𝑃𝑟 optical received power by the photodetector, in (dbm). 𝑃s average source signal power into the first span. 𝑃𝑡 optical transmitted power by cw laser, in (dbm). q q-factor. 1. introduction optical fiber is the key component in various types of telecommunications networks, due to the intensive demands for high bandwidth, transmission speed and distance. such characteristics make an optical fiber a preferred means of transmission. when informationcarrying light pulses propagate through an optical fiber, they suffer from attenuation, temporal broadening, and even interact with each other through nonlinear effects in the fiber. (ramaswami, r., sivarajan, k., & sasaki, g, 2009), (thyagarajan, k., & ghatak, a., 2006). correct functioning of an optical data link depends on modulated light reaching the receiver with enough power to be demodulated correctly. attenuation is the reduction in power of the light signal as it is transmitted (junipernetworks, 2022). with the recent development of practical erbium-doped fiber amplifiers (edfas), attenuation has become no longer a capacity-limiting issue. but dispersion is still a major concern that corrupts the signal quality. dispersion is one of the critical barriers to an increase in the transmission capacity of an optical fiber, particularly in a single-mode fiber, causes an optical pulse to broaden as it travels along the fiber cable. the broadening is so severe that adjacent pulses begin to overlap to the degree that a photo-detector can no longer distinguish clear boundaries between pulses. as a result, reduces the effective bandwidth which further escalates the error rate due to an increasing inter-symbol interference (isi) (sharma, a., singh, i., bhattacharya, s., & sharma, s., 2019). in order to widen bandwidth of optical fiber transmission systems, techniques for dispersion compensation over transmission bandwidth for the accumulated chromatic dispersion of standard smf must be in place for mitigating signal deterioration caused by dispersion and nonlinearities. (kaler, r. s., sharma, a. k., & kamal, t. s, 2002). there are three basic limitations in modern optical transmission systems: the optical signal-to-noise ratio (osnr), dispersion, and nonlinear effects (grunernielsen, l., wandel, m., kristensen, p., jorgensen, c., jorgensen, l. v., edvold, b., ... & jakobsen, d., 2005). this paper will discuss how the limitations due to dispersion can be dealt with by the use of dispersion compensating fiber and fiber bragg grating. another proposed model is applied by combining dcf with fbg in a manner that contributes to improving performance at a certain level of input power and makes the comparison with other methods. our analysis is based on the chromatic dispersion which is dominant in the single mode fiber. this raises the question of how well the dcf and fbg outperform one another is compared in a practical approach by varying the input optical power intensity with constant modulation scheme while keeping tab on the q-factor for measuring how dispersion poses a serious problem in optical communication link. 2. basic theory: 2.1 dispersion when the optical signal strength is low, the dispersion can be seen in the linear range, but if the input light intensity is high enough, it may exhibit nonlinear properties (ahmed, 2017). fiber dispersion plays an essential part in the propagation of short pulses as all the pulses with different spectral components travel at different speeds. this is the essence of chromatic dispersion (cd). chromatic dispersion is a phenomenon in an optical fiber that results in the temporal broadening of optical pulses as they pass through the fiber due to the dependency of group index on wavelength (kaler, r. s., sharma, a. k., & kamal, t. s, 2002). inside smf, chromatic dispersion can be classified into two parts: waveguide dispersion and material dispersion. material dispersion is a function of the source spectral width, which contains components, ammar m. odeh 165 namely modes, at different wavelengths. these wavelengths propagate at different velocities depending on the refractive index of the silica, the material used to make optical fiber. the core of the fiber is not the only place where the modes can exist; they also extend slightly into the cladding material (sharma, a., singh, i., bhattacharya, s., & sharma, s., 2019). the power distribution of a mode between the fiber's core and cladding is a function of wavelength; the longer the wavelength, the more power in the cladding. this power distribution alters with a change in wavelength, affecting the effective index or propagation constant of the mode. this phenomenon is classified as waveguide dispersion. (ramaswami, r., sivarajan, k., & sasaki, g., 2009). the dispersion is related to the second derivative of the propagation constant (β): 𝐷 = − 2𝜋𝑐 λ2 𝑑2𝛽 𝑑𝜔2 (1) where c is the light velocity in the vacuum, l is the wavelength and 𝜔 is the frequency. as a result of dispersion, different delay times will be generated between different spectral components of the optical signals. for a single-mode fiber of length (l), the specified wavelength components at the angular frequency (𝜔) would arrive at the receiver terminal with a time delay (𝜏): τ = l/𝑣𝑔 (2) where 𝑣𝑔 is the group velocity, calculated as follows: 𝑣𝑔 = 𝑑𝜔/𝑑𝛽 (3) where β is the propagation constant. the propagation constant is a measure of changes in the amplitude and phase of the electromagnetic wave as they travel through a medium. the propagation constant can be written in terms of the free space wave number and the effective index: 𝛽 = 𝑘0 𝑛𝑒 = 𝜔 𝑐 𝑛𝑒 (4) and the effective index as 𝑛𝑒 = 𝑛0 + ∆𝑛𝑒 (5) where 𝑛0 is the refractive index of the cladding and ∆𝑛𝑒 is the effective index difference. 2.2 dispersion compensation techniques the data rate of the fiber optic communication link cannot be increased beyond a certain limit due to intersymbol interference. to achieve high data rates, dispersion compensator is the most crucial part to sustain link robustness against the impairments imposed on the light-wave optical system. dispersion compensating fiber, fiber bragg gratings, electronic dispersion compensation (edc), and digital filters are the most available techniques for dispersion compensation. (hossain, m. b., adhikary, a., & khan, t. z., 2020). this study is limited to dcf and fbg. the alternative techniques are outside the scope of this paper. (a) dispersion compensation fiber dispersion compensation fiber is a sort of optical fiber that has dispersion characteristics opposite those of transmission link fibers. the dispersion of the fiber can be controlled by varying the refractive index profile and the relative index value. the transmission fiber is a standard smf (ssmf) having zero dispersion at a wavelength of 1310 nm and a dispersion of 17 ps/nm/km at a wavelength of 1550 nm (chen, w., li, s., lu, p., wang, d., & luo, w., 2010). for narrowband externally modulated lasers, the bandwidth of the signal is governed by the modulation. increased modulation speed leads to increased bandwidth. on the other hand, increasing the optical bit rate leads to an increased signal distortion effect. a popularly used criterion for the maximum allowed fiber distance limited by fiber dispersion (without an amplifier) is given by: 𝐿𝑚𝑎𝑥 = 𝐾 𝐵2|d| (6) where d is the dispersion coefficient of the transmission fiber, b is the system bit rate and k is a constant depending on many details, such as laser spectral width, modulation method, and the wavelength of the optical signal. therefore, as b increases, l decreases with the square root of b (gruner-nielsen, l. and wandel, m. and kristensen, p. and jorgensen, c. and jorgensen, l.v. and edvold, b. and palsdottir, b. and jakobsen, d., 2005) (mapyourtech, 2014). to cancel out the accumulated dispersion of the link road, the following principle will be obeyed: ldcf ddcf + lsmf dsmf = 0 (7) where dsmf is the dispersion coefficient of the transmission fiber, lsmf is the length of the transmission optical fiber, ddcf is the dispersion coefficient of the dcf, and ldcf is the length of the dcf. the higher the dispersion coefficient of the compensating fiber, the smaller the required length of the of the compensating fiber. the effective core area of the dcf is much smaller than that of transmission fiber, 20 μ𝑚2, compared to the 85 μ𝑚2 of smf, and hence exhibits higher nonlinearities. the wavelength of operation is an comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 166 important consideration, in which small increases in wavelength leading to relatively large changes in mode size, where the propagation in the less dense media (the cladding) provides an increase in propagation than of the dense media (the core). this, in the nutshell, creates a large negative dispersion value. (willner, a. e., song, y-w, mcgeehan, j., z pan, & hoanca, b. , 2005). by reducing the optical power, the influence of non-linear optical effects (noe) can be reduced. the dcf takes a long length of fiber to compensate for a span, about one-fifth of the length of the transmission fiber, entailing a large amount of space and a relatively large insertion loss. additionally, the dcf can only accept a limited signal power because the small-mode field of the fiber and the long propagation length result in nonlinear pulse distortions at high signal intensity (kaminow, i. p., & li, t. (eds.), 2002). again, a dcf is a fiber that causes dispersion opposite to conventional single-mode fibers. dcf has negative dispersion over the same range of wavelengths in which smf has positive dispersion. multiple wavelengths are used to transmit data in a single fiber using wavelength division multiplexing (wdm). wdm is a technique that enables the use of different data types over fiber networks in the form of light. a single virtual fiber network is produced by allowing many light channels, each with a unique wavelength, to be sent simultaneously over an optical fiber network; hence, a single fiber can be shared for several services. dcf provides fixed negative dispersion for all channels in the wdm system and considered as a good compensating device for its reference wavelength where all optical channels are not compensated to the same degree, which results in residual dispersion. this leads to dispersion slope mismatch, the term used to describe how residual dispersion goes up and down (t. duthel, s. jansen, p. krummrich, m. otto, and c. schäffer, 2005). on the other hand, fiber gratings, the other technique used in this paper, offer an alternative approach that overcomes the problems. (b) fiber bragg grating it is one of the most flexible and evolved techniques for dispersion compensation. the principle behind fbg is to reflect wavelengths that satisfy the bragg condition and transmit the desired wavelengths. the unwanted wavelengths are selected by changing the grating period. the simple model of this fiber-based device and the possibility that a single parameter can be changed to alter dispersion are two of its primary advantages (t. duthel, s. jansen, p. krummrich, m. otto, and c. schäffer, 2005). fbg consists of a periodic modulation of the refractive index in the core of a single-mode optical fiber. the bragg wavelength also varies along the grating length, thus different frequency components of an incident optical pulse are reflected at different points depending on where the bragg condition is satisfied. this can be easily expressed as follows: λb = 2neffʌ (8) where λb is called the bragg wavelength, neff is the grating effective refractive index and ʌ is the grating period. compared to dcf, fbg-based dispersion compensators can be made to be very compact, with potentially lower insertion loss and reduced optical nonlinearity (p.s. westbrook, b.j. eggleton, 2005). the low insertion loss of the fbg allows a significant reduction in the amplification requirements. this feature lowers the system cost and can improve the optical signal-to-noise ratio (osnr). most applications need a dispersion tuning range that includes both positive and negative values. this is the key benefit of fbg technology that can't be met by dcf. changing the physical parameters of the fiber grating allows us to obtain a narrow or wide bandwidth or other special characteristics (sharma, a., singh, i., bhattacharya, s., & sharma, s., 2019). 2.3 osnr and ber: the number of amplifiers affects the quality of the transmission line. hence, the performance of the transmission line must be determined using a measurement technique called the optical signal-tonoise ratio. osnr is used to identify the signal quality in the optical domain. it is the amount of difference between the signal power and noise power in the unit (db) somewhere within the modeled system. while edfa amplifies the optical signal, the osnr is reduced because such an amplifier adds unwanted amplified spontaneous emission (ase) noise into the optical signal (sajjan, s. m., seshasai, v., & sadashivappa, g, 2015), (ramaswami, r., sivarajan, k., & sasaki, g., 2009). osnr is one of the most diagnostic metrics for monitoring the optical network, as the ber is directly linked to the osnr signal. the osnr value depends on many factors, including data rate for a specific modulation format, location in the network and the target ber level. a low osnr value indicates that the receiver will probably not detect or recover the signal (tassé, 2012). the osnr value can be calculated as: osnrf,db = 158.9 + 𝑃s,dbm − γdb − nf − 10 log(∆f) − 10 log(n) (9) where osnrf is the final osnr value that faced the receiver (in db), 𝑃s is the average input source signal power (in dbm), nf is the amplifier noise figure, (in db), ∆f is the bandwidth measures nf (in hz), n is the number of amplifiers used in the transmission system, γ is the span loss (in db). therefore, osnr, ber, and q-factor are major indicators of the quality of the overall system in digital communication systems. the q-factor, the criterion for measuring the quality of the received signal investigated at the receiver, limits the ber to be ammar m. odeh 167 below10−9(one error in 109 bits), and is approximated as follows: ber = 1 2 erf [ q √2 ] ≈ e−q 2 /2 q√2π (10) whereby erf(x) is the error function (special function) widely used in statistical computations. 3. simulation setup in this study, three dispersion compensation methods (dcf, fbg, and the combining of both in one model) with various configurations (i.e., pre, post, and symmetrical scheme) are considered for dispersion compensation using optisystem 7.0 software for single-channel optical communication. the simulation setups of the proposed optical system are shown in figures from fig.1 to fig.9. at the transmitter end, the simulation setups consist of a continuous wave (cw) laser used as a source of light, a return to zero modulation format, and a mach– zehnder (mz) modulator to externally modulate the input rz signal with an extinction ratio of 30db. all the optical pulses are sent at a transmission rate of 120 gbps. the laser source is a cw type at a frequency of 193.1 thz and the output power is varied from -5 dbm to 20 dbm. the loop control system has two loops for all models except for dcf and dcf-fbg models in symmetrical mode. the reason behind it is to equalize the transmission length. to adequately compensate for the dispersion slope and cumulative dispersion in the transmission fiber (smf), each span is made up of 100 km of transmission fiber and 20 km of dcf. however, since the symmetrical design only has one loop, the overall length of the fiber channel is still 240 km. at the receiver terminals, an avalanche photodiode (apd) detector is used to detect the optical pulses and then convert them to electrical signals. simulation parameters, optical fiber properties, and fbg fiber parameters are tabulated in tables 1, 2, and 3, respectively. 4. results and discussion single-channel nonlinear effects are caused mainly by self-phase modulation (spm), it causes optical pulses to have a wider spectral range by interacting with the medium that passing through and imposing a phase modulation on itself. the rz modulation format is used for three simulation models, which is preferred in high bit rate systems at a higher input of power compared to nrz, as it can help in providing great results in reducing dispersion distortion and inter-symbol interference. we have analyzed the effects of spm for three cases: dcf, fbg, and dcf cascaded to fbg, using performance metrics: q-factor, ber, and eye diagrams. the comparative value of the q-factor for different models is given in tables 4, 5, and 6. according to the simulation results for the three compensation models, dcf in a symmetrical scheme provides the best q-factor of 23.2824 and ber of 2.41855e-120 at an input power of 15 dbm compared to all other model configurations in this study. figure 16 depicts the case. generally, dcf performs well in a mid-power range between 10 and 15 dbm, among all other dcf schemes. on the other hand, the cascaded system performs better in the low-power region of less than 10 dbm and achieves the highest q-factor of 10.8361 and ber 9.15197e-028 in pre-compensation mode at an input power of 5 dbm, as shown in fig. 17. at an input power of 15 dbm and above, fbg performs better compared to other models and achieved satisfactory results, and gives almost identical values, 7.508, 7.65518, and 7.63105 for the three compensation modes, pre-, post-, and symmetrical, respectively. the reason behind this is that the system involved with dcf exhibits nonlinearities at high power intensity. which is considered one of the suggested solutions to use the fbg, so that it does not require additional cables to compensate for dispersion and thus fewer amplifiers, which achieves the highest possible osnr value along the link. table 1. simulation parameters parameters values reference wavelength [thz] 193.1 channel type single-channel transmission channel data rate [gbit/s] 120 sequence length [bit] 128 samples per bit 128 signal format unipolar (single polarity) rz optical external modulator. er [db] 30 receiver bandwidth [hz] 0.75 * rb table 2. optical fiber parameters parameters values smf dcf length [km] 100 20 attenuation [db/km] 0.2 0.5 dispersion [ps/(nm.km)] 17 -85 dispersion slope [ps/(nm^2 . km)] 0.075 -0.375 effective area [micro m^2] 85 20 kerr coefficient, n2 [m^2/w] 2.6e-020 2.6e-020 table 3. fbg parameters parameters values fbg dcf-fbg bandwidth [thz] 1 1 frequency [thz] 193.1 193.1 dispersion [ps/(nm.km)] -1720 -860 insertion loss [db] 2 2 comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 168 figure1. pre-compensation using dcf simulation model. ammar m. odeh 169 figure 2. post-compensation using dcf simulation model. comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 170 figure 3. symmetrical compensation using dcf simulation model. ammar m. odeh 171 figure 4. pre-compensation using idcfbg simulation model. comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 172 figure 5. post-compensation using idcfbg simulation model. ammar m. odeh 173 figure 6. symmetrical compensation using idcfbg simulation model. comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 174 figure 7. pre-compensation using dcf-fbg simulation model. ammar m. odeh 175 figure 8. post-compensation using dcf-fbg simulation model. comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 176 figure 9. symmetrical compensation using dcf-fbg simulation model. ammar m. odeh 177 figure 10. comparison of transmission power vs qfactor influence of fbg simulation model. figure 11. comparison of transmission power vs q-factor influence of dcf-fbg simulation model. figure 12. comparison of transmission power vs q-factor influence of dcf simulation model. figure 13. comparison of transmission power vs qfactor influence in pre-compensation mode. figure 14. comparison of transmission power vs q-factor influence in post-compensation mode. figure 15. comparison of transmission power vs q-factor influence in symmetrical mode. figure 16. eye diagram for dcf symmetrical compensation simulation result at -15 dbm input power. 0 2 4 6 8 10 5 0 5 1 0 1 5 2 0 pre post symmetrical 0 5 10 15 5 0 5 1 0 1 5 2 0 pre post symmetrical 0 5 10 15 20 25 5 0 5 1 0 1 5 2 0 pre post symmetrical 0 5 10 15 20 5 0 5 1 0 1 5 2 0 fbg dcf-fbg dcf 0 5 10 15 20 25 5 0 5 1 0 1 5 2 0 fbg dcf-fbg dcf 0 5 10 15 20 25 5 0 5 1 0 1 5 2 0 fbg dcf-fbg dcf comparing dispersion compensation methods for 120 gb/s optical transmission: pre, post, and symmetrical schemes 178 figure 17. eye diagram for dcf-fbg pre-compensation simulation result at 5 dbm input power table 4. simulation results of the fbg model transmitted signalpower q-factor pre post symm. -5 0 0 0 0 0 2.87013 0 5 3.33472 4.80579 3.59826 10 5.84047 7.02279 6.53683 15 8.005 9.01556 8.79573 20 7.508 7.65518 7.63105 table 5. simulation results of the dcf-fbg model transmitted signalpower q-factor pre post symm. -5 3.9294 2.65284 3.15258 0 7.43716 4.98011 5.57083 5 10.8361 8.162 8.50435 10 12.4673 11.8314 11.0831 15 7.79289 12.8964 10.2684 20 2.2428 5.74072 3.26283 table 6. simulation results of the dcf model transmitted signalpower q-factor pre post symm. -5 0 0 0 0 3.31692 3.09557 3.05317 5 7.43216 7.22034 7.43353 10 14.574 15.0505 16.0879 15 15.3546 19.4849 23.2824 20 4.28873 7.05111 7.19432 5. conclusion: the performance of each setup was reported in terms of q-factor and ber. the evaluations of the systems depend on the best value of the q-factor, which corresponds to ber. this paper focuses on the singlechannel optical transmission system. after analyzing the results and graphs, it is clear that fbg compensation has the lowest q-factor when compared to other simulation models in the lowand mid-power region (less than 20 dbm). dcf experiences the best performance at an input power of 10 and 15 dbm. to prevent nonlinear impairments, in which dcf is limited in optical input power and has a fairly large insertion loss, fbgs might take over from dcf as the preferred method for in-line dispersion compensation. the proposed dcf-fbg system shows its advantage in the low power region of 5 dbm and lower, and surpasses the fbg system by up to 15 dbm. this approach combines the benefits of both dcf and fbg as an effective way for dispersion compensation and loss reduction, respectively. this, in comparison to other models, has a significant contribution in providing the best q factor in the low-power region. there is a key demand for longer fibers, which leads to higher costs associated with the positioning of amplifiers along the fiber routes. also, dispersion increase with higher bandwidth, and the emphasis is strong on dispersion compensation points. therefore, fbg can be focused at a single location and tends to have fewer compensation points and amplifiers. so, for overall consideration, dcf and the cascaded system for their best q-factor are the promising approaches, but a significant negative dispersion coefficient from dcf is required for the reduction of positive dispersion in smf. this leads to an increase in the non-linear effect and overall cost. in addition, since it only corrects the center wavelength of the pulse, in which shorter wavelengths will be overcompensated and longer wavelengths will be undercompensated. unlike dcf, the fbg device can compensate for chromatic dispersion at multiple wavelength variations. as a result, fbg is the preferred option for compensating for chromatic dispersion. conflict of interest the authors declare that there are no conflicts of interest regarding this article. funding the authors did not receive any funding for this research. references ahmed, r. k. (2017). performance evaluation of high data rate optical communication system utilizing fbg compensated dispersion schemes under different modulation techniques. diyala journal of engineering sciences, 10(2), pp. 94-106. chen, w., li, s., lu, p., wang, d., & luo, w. (2010). dispersion compensation optical fiber modules for 40 gbps wdm communication systems. frontiers of optoelectronics in china, 3(4), pp. 333-338. gruner-nielsen, l. and wandel, m. and kristensen, p. and jorgensen, c. and jorgensen, l.v. andedvold, b. and palsdottir, b. and jakobsen, d.(2005). dispersion-compensating fibers. journal of lightwave technology, 23(11), pp. ammar m. odeh 179 3566-3579. hossain, m. b., adhikary, a., & khan, t. z. (2020). performance investigation of different dispersion compensation methods in optical fiber communication. asian journal of research in computer science, 5(2), pp. 36-44. junipernetworks. (2022, aug 2). fiber-optic cable signal loss, attenuation, and dispersion. retrieved from juniper networks: https://www. juniper.net/documentation/us/en/hardware/mx96 0/acx1000/topics/concept/fiber-optic-cable-signal -loss-attenuation-dispersion-understanding.html kaler, r. s., sharma, a. k., & kamal, t. s. (2002). comparison of pre-, post-and symmetricaldispersion compensation schemes for 10 gb/s nrz links using standard and dispersion compensated fibers. optics communications, 209(1-3), 107-123. kaminow, i. p., & li, t. (eds.). (2002). optical fiber telecommunications iv-a: components (vol. 1). elsevier. mapyourtech. (2014, sep 30). factors that contribute to chromatic dispersion and how to combat it? retrieved 2022, from map your tech: p.s. westbrook, b.j. eggleton. (2005). fiber gratings. encyclopedia of modern optics, pp. 501-513. ramaswami, r., sivarajan, k., &amp; sasaki, g. (2009). optical networks: a practical perspective. morgan kaufmann. sajjan, s. m., seshasai, v., & sadashivappa, g. (2015). dwdm link design and power budget calculation. international journal of advanced research in electrical, electronics and instrumentation engineering, 4(4), 2538-2548. sharma, a., singh, i., bhattacharya, s., & sharma, s. (2019). performance comparison of dcf and fbg as dispersion compensation techniques at 100 gbps over 120 km using smf. nanoelectronics, circuits and communication systems (pp. 435-449). singapore: springer. t. duthel, s. jansen, p. krummrich, m. otto, and c. schäffer. (2005). multi-channel residual dispersion compensation in a 40 gb/s wdm system utilizing a single all-fiber delay line filter. optical fiber communication conference and exposition and the national fiber optic engineers conference, p. jwa20. tassé, j. s. (2012, dec 6). osnr values in dwdm networks: blog. retrieved from exfo thyagarajan, k., & ghatak, a. (2006). some important nonlinear effects in optical fibers. guided wave optical components and devices, 91-100. willner, a. e., song, y-w, mcgeehan, j., z pan, & hoanca, b. . (2005). dispersion management. in encyclopedia of modern optics (vol. 4, pp. 353365). elsevier. pdf compressor civl071006.qxd the journal of engineering research vol. 6, no.1 (2009) 76-81 1. introduction bitumen, which remains after vacuum distillation of crude oil, is a complex mixture of organic and inorganic compounds. the physical, mechanical, and rheological behavior of bitumen in road and building construction is governed by its chemical composition. for this reason, considerable effort goes toward widening the knowledge of this composition and relating it to performance (anhorn and badakhshan, 1994; yui, 1997; bonemazzi and giavarini, 1999; garcia-perez, et al. 2001; zhao, et al. 2001; michaelian, et al. 2001; el akrami, et al. 1999; parker and mcfarlane, 2000). the chemical composition of bitumen is very complex, although its components can be broadly categorized as maltenes and asphaltenes (raki, et al. 2000; yarranton, et al. 2000; oh and deo, 2002; zhao, et al. 2001; proctor, et al. 2001; japanwala, et al. 2002). asphaltenes are defined as the black-coloured fraction of bitumen that is insoluble in n-heptane. maltenes, composed of saturated compounds, aromatic compounds and resins are soluble in n-heptane. the ratio of asphaltenes to the other constituents, and the maltenic ________________________________________ corresponding author’s e-mail: mahmoud@squ.edu.om fraction composition have a significant effect on the viscoelastic properties of bitumens, and hence on their performance as road paving binders (garcia-morales, et al. 2004). the high temperature sensitivity of bitumen makes it difficult to optimize both highand low-temperature properties. the binders are too brittle at low temperatures and lack cohesion at high temperatures. the temperature sensitivity of straight-run bitumens can be reduced by adding polymer herrington, et al. 1999). the polymer addition allows an increase in the resistance of the binder to permanent deformation at high temperature. besides, the fracture properties including critical stress intensity factor (k1c) at low temperature of polymer modified bitumen were shown to be higher than those of the bitumen base (champion-lapalu, et al. 2002). therefore, polymer-bitumen blends find considerable interest from different authors (fawcett and lor, 1992; fawcett and mcnally, 2000; lu, 2000; airey, 2003). the application of polymeric materials in agriculture and horticulture has increased considerably in recent years, not only as replacement for traditional materials but also as a mean of effecting improvement. thus, the application of plastics has led to a significant improvement in waste plastic-modified bitumen: rheological study mahmoud abdel-goad department of petroleum and chemical engineering, college of engineering, sultan qaboos university, al-khod, p.o. box. 33, p.c. 123, muscat, sultanate of oman received 6 october 2007; accepted 29 june 2008 abstract: bitumen blends were prepared for road applications by the introduction of 9% of waste polyethylene by weight (9 wt%). the relaxation stress, relaxation and retardation spectrum and viscosity of bitumen blends were studied at different temperatures and compared to those of the original pure bitumen. these properties were tested using an aresrheometer (rheometric scientific, co.) equipment. the measurements were performed in the dynamic mode, plate-plate geometry of 8 mm diameter over the temperature range from -10 to 60°c and angular frequency (ω) varied from 10-1 to 102 radian/s. the relaxation stress and viscosity were modified by the addition of waste pe. the results also indicate that the incorporation of the waste pe enhances stability of the bitumen blends. keywords: bitumen, waste plastics, relaxation stress, relaxation spectrum, viscosity á«lƒdƒjq á°sgqo :᫵«à°sóñdg ogƒÿg äéø∏îã qé≤dg ú°ù–– ogƒ÷g ~ñy oƒªfi áá°°uuóóÿÿggäélqo ~æy î«áôdg ré¡l ᣰsgƒh ég~j~– ” áfhôÿg h álhõ∏dg ¢üfé°ün.¥ô£dg ∞°uq ¢vô¨d qé≤dg ¤g ú∏jéãjg ‹ƒñdg äéø∏fl áaé°véh iò° – ” •ƒ∏ıg qé≤dg : áaé°véh gòãc âæ°ù– ~b áeéy áø°üh á«lƒdƒjôdg ¢üfé°üÿg ¿g úñj åëñdg øeh .»cô◊g i霒g ‘ á«fék/áød 100 ¤g 1 úh ée ooôj h ájƒäe álqo 60 10 úh ée iqgôm é¡eg~îà°sg ø°ùëj é‡ ájƒäe álqo 60 ‹g π°üj iqgôm äélqo ~æy aéæëfg ¿h~h é¡jéñk ø°ùëj é‡ â©øjqg ~b iqgôë∏d é¡∏ª– álqo ¢ün’éhh qé≤dg ¤g ᫵«à°sóñdg äéø∏ıg √òg .iqé◊g oóñdg ‘ ¢ün’éh ™°sgh ¥é£f ‘ áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.álhõ∏dg ,aénî°s’g ~¡l, »lƒdƒjq,᫵«à°sóh äéø∏fl ,qéb : 77 the journal of engineering research vol. 6, no.1 (2009) 76-81 technological processes in the growing and storing of agricultural crops. as a result, a large amount of waste plastic films, containing ldpe, eva and a minor proportion of stabilizers, are generated every year all over the world (garcia-morales, 2004). therefore, the possibility of disposing of troublesome waste plastics within road bitumens provides an environmental and economical solution. the rheological characterization of the bituminous materials is important to better understand its microstructure. therefore, the rheological study has a considerable interest by different authors (mehrotra, 1991; eastick and mehrotra, 1990, airey, 2003; mastrofini and scarsella, 2001; perez-lepe and martinez-boza, 2003). it is well established that the linear behavior of bitumens is described by the general linear visco-elastic model. the rheological analysis is a sensitive and versatile thermal analysis technique, which measures the modulus (stiffness) and damping properties (energy dissipation) of materials as the material are deformed under periodic stress (seha, et al. 1999). since the modulus of bitumen can vary by 3 to 4 orders in magnitude over the in-service temperature range, it was the objective of this work to modify the bitumen at inservice temperatures in hot countries by adding waste pe. modified bitumen and neat bitumen will be characterized rheologically. 2. experimental work 2.1 materials and preparation bitumen blends were prepared by adding waste plastics to commercial bitumen (60/70). the waste pe bags were sorted from the garbage and shredded into coarse particles. the waste pe bags and the bitumen were weighed and heated individually in an oven until melting. waste pe in molten state was poured into the molten bitumen, mechanically stirred to give a homogenous sample and then stored at 25°c. the bitumen blends and the original neat bitumen samples were compression-molded at 70°c and 5 bars for 1 hour and shaped for the rheology measurements in a disc form with 8 mm diameter and 1 mm thickness. 2.2 measurements procedure the rheological characterization was studied for neat bitumen and bitumen blends by using dynamic an aresrheometer (rheometric scientific) equipment in plateplate geometry of 25 mm diameter. the sample to be measured was placed on the lower plate and the upper plate was lowered until just contact the sample. the measurements were performed in the temperature range from 10 to 60°c and at an angular frequency (ω) in the range 102 to10-1 radian/s in the dynamic mode and the strain amplitude was kept at 1% in the linear viscoelastic regime. six points per decade in frequency were obtained. the actual gap size is recorded electronically and allows absolute moduli to be determined. 3. results and discussions the master curve is obtained by shifting the measured data at different temperatures (t) into a single curve at a reference temperature t0 as shown in fig. 1. by using the time-temperature superposition principle which is described by williams, landel and ferry (ferry, 1980) as, log at = -c1 (t-t0) / (c2 + (t-t0)) where at is the horizontal shift factor, t0 is a reference temperature, t is measured temperature and constants c1 and c2 are material specific. c1 and c2 are roughly determined as listed in table 1. at of bitumen blends is plotted as a function of temperature in fig. 2. in this figure at decreases with the increase in the temperature up to 25°c (at ~ t-0.14 ) then becomes nearly independent on the temperatures. the shift factor (at) shifts the data obtained at different temperatures in the horizontal direction. in the vertical direction the experimental data are shifted by bt (bt = ρt / ρ0τ0 [30] ), where ρ is the material density. the effect of the temperature on bt is shown in fig. 3 for bitumen blends. the vertical shift (bt) is nearly constant at low temperatures but after 25°c it falls rapidly ( bt ~ t-0.17 ) with the increase of the temperature as shown in fig. 3. table 1. material constants 10-910-810-710-610-510-410-310-210-1 100 101 102 103 104 105 106 103 104 105 106 107 bt at non-shifted exp.data master curve at t 0=10°c master curve at t 0=60°c g t p a time, s figure 1. non shifted experiments data of g(t) at different temperatures are shifted into master curves at t0 = 25oc by using at and bt material c1 c2 bitumen 24.9 196 bitumen with 9%pe 1.8 30 78 the journal of engineering research vol. 6, no.1 (2009) 76-81 the shear relaxation stresses g(t) is plotted against time with logarithmic scales at t0 = 10, 25 and 60°c in figs. 4-5 for neat bitumen and bitumen/pe blend. g(t) is the relaxation stress at constant deformation. at long times of each t0, the master curve of each g(t) shows flow regime which is related to the complete relaxation of the macromolecules and released from the restrictions of the entanglements. the time taken in the relaxation processes is called the longest relaxation time. in this regime, the viscous component is higher than the elastic component because of the release of the entanglements which are the responsible for the elasticity. therefore, g(t) decreases sharply and falls by many order of magnitude in this regime as shown in fig. 4 and 5. at intermediate time of each master curve is located the rubbery zone that is the signature of the elasticity. as a result, g(t) is independent on the time in this regime since the loss in the energy in every cycle of the deformation is minimum. at short times the third regime of the viscoelastic behaviour is found only in the case of pure bitumen which is called dynamic glass transition because of the possibility to measure the neat bitumen samples at very low temperatures (-10°c). the master curves at reference temperature (10°c) are shifted toward the long time, those at 25°c appear to be in the intermediate and the master curves at reference temperatures 60°c are shifted toward the short time as shown in fig. 4. this is because the time of relaxation of the entanglement chains is inversely proportional to temperature, since at low temperatures the chains take a long time in the relaxation processes, unlike at high temperatures. g(t) modulus at 60°c is lower than at 25°c and g(t) at 25°c is lower than at 10°c as seen in figs. 4 and 5. figures 4 and 5 are evidence that the incorporation of the waste plastics into the bitumen increases the relaxation stress over all the temperatures and time ranges. this increase in g(t) moduli is more noticeable at long times and high temperatures as shown in these figures. where the values of g(t) modulus at 5 seconds and 10°c is 2 times higher than this of neat bitumen by the addition of 9% pe as shown in fig. 4. this factor reaches up to 7 at 50 seconds (the same temperature) as shown in fig. 4. the ratio g(t)bit.-9%pe/ g(t)bitumen at the end of flow regime in the cases of t0 = 10, 25 and 60°c are 3, 6 and 10. this means that the bitumen/pe blend is more stable than neat bitumen at high temperatures for long times. this is due to the formation of the bitumen-polymer network which -20 0 20 40 60 10-1 100 101 102 103 104 105 106 a t temp., °c figure 2. at of bitumen/pe blend as a f unction of temperature at t0 = 25oc -20 -10 0 10 20 30 40 50 60 70 1 b t temp., °c figure 3. bt of bitumen/pe blend as a function of temperature at t0 = 25oc 10-810-710-610-510-410-310-210-1100101102103104 102 103 104 105 106 107 neat bitumen bit.-9% pe g t p a time, s figure 4. g(t) of bitumen/pe and neat bitumen at t0 = 25oc 10-1210-1110-1010-910-810-710-610-510-410-310-210-1100101102103104105106 102 103 104 105 106 107 t0= 60°c t0= 10°c neat bitumen bit.-9% pe g t p a time, s figure 5. g(t) of bitumen/pe and neat bitumen at t0 = 10 and 60oc 79 the journal of engineering research vol. 6, no.1 (2009) 76-81 increases the stiffness and strength of the bitumen blends. that is confirmed too in figs. 6 and 7. in these figures, the shear equilibrium creep (je) is logarithamically plotted versus time for bitumen and bitumen/pe blend. the addition of pe to bitumen enhances (je) in particular at long times and the differences among (je) values of bitumen/pe blend is more observable at long times due to the stability of the bitumen-pe network. the relaxation (h(t)) and retardation (l(t)) spectrum of neat bitumen and bitumen/pe blend are plotted as a function of time at 25°c in figs. 8 and 9. h(t) and l(t) indicate the distribution of relaxation and retardation mechanisms, respectively in different regions of the time scale. they show a broad spectrum of relaxation and retardation times. a broad mechanical relaxation, observed at 25°c, is associated with the collapse of a compact structure, constituted by asphaltene particles surrounded by solid resin. the addition of pe to bitumen modifies this relaxation. the values of h(t) increase by the addition of waste pe to bitumen, in particular, at long times as shown in fig. 8. this is clear at 40 seconds where h(t) of bitumen blend is higher than that of neat bitumen by a factor 2 with the addition of 9% pe as shown in fig. 8. this factor becomes 6 at 100 seconds as shown in fig. 8. the same scenario is observed in l(t) (fig. 9) since at 40 seconds the addition of 9% pe to bitumen results in changing the values of l(t) for neat bitumen from 5.0 x 10-5 to 4.3 x 10-1 pa-1 as shown in fig. 9. however, at 1000 seconds the values of l(t) are 5.0 x 10-3 and 5.7 x 10-4 pa-1 for pure bitumen and bitumen/9%pe blend, respectively as shown in fig. 9. this as explained earlier due to the increase in the stiffness and strength of the bitumen by the incorporation of the pe. the complex viscosity (η*) as a function of temperature for neat bitumen and bitumen blends at ω = 0.1 radian/s is plotted in fig. 10. as shown in this figure, the incorporation of the waste pe into the bitumen increases the viscosity over all the temperatures range. this is because the viscosity of the bitumen matrix is higher than neat bitumen. the complex viscosity (η*) of the blends at high temperatures are higher than at low temperatures. as shown in fig. 10 the values of η* at 15°c increase from 5.0 x 104 (neat bitumen) to 5.7 x 105 pa.s by the addition 9% pe, respectively but at 60°c the differences among the values of η* are high. these values are 2.7 x 105 and 5.7 x 105 pa.s for neat bitumen and bitumen/9%pe blend, respectively as shown in fig. 10. this difference is due to the change in bitumen compo10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101 102 103 104 10-8 10-7 10-6 10-5 10-4 10-3 pure bitumen bit.-9% pe je p a1 time, s figure 6. je of bitumen/pe and neat bitumen at t0 = 25oc 10-1110-1010-910-810-710-610-510-410-310-210-1100101102103104105106 10-7 10-6 10-5 10-4 10-3 t0 =60°c t0 =10°c neat bitumen bit.-9 % pe je (t ) p a -1 time, s figure 7. je of bitumen/pe and neat bitumen at t0 = 10 and 60oc 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101 102 103 104 101 102 103 104 105 106 107 neat bitumen bit.-9% pe h (t) p a time, s figure 8. h(t) of bitumen/pe and neat bitumen at t0 = 25oc 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101 102 103 104 10-9 10-8 10-7 10-6 10-5 10-4 10-3 neat bitumen bit.-9e l( t) p a -1 time, s figure 9. l(t) of bitumen/pe and neat bitumen at t0 = 25oc 80 the journal of engineering research vol. 6, no.1 (2009) 76-81 nents considerably at ~ 40°c, particularly the ratio of the asphaltenes to the maltenes (the main components of bitumen) which has a significant effect on the viscoelastic properties of bitumen. 4. conclusions bitumen blends were prepared as a road paving binders by the introduction of 9% waste pe in a molten state followed by mechanical mixing. the materials were rheologically characterized using an aresrheometer (rheometric scientific, co.) equipment. the measurements were performed in the dynamic mode, plate-plate geometry of 25 mm diameter over the temperature range from -10 to 60°c and angular frequency (ω) varied from 102 to 10-1 radian/s. the relaxation stress, relaxation and retardation spectrum and viscosity of bitumen blends were studied at different temperatures and compared to those of the original neat bitumen. the results showed that the relaxation stress and viscosity were developed by the incorporation of 9% of waste plastics into neat bitumen. the results also showed that the incorporation of the waste plastics enhances stability of the bitumen blends. an increase in these properties resulted from the incorporation of waste plastics into bitumen, this increase is significant particularly at high temperatures for a long time. as a consequence, the use of waste pe bags can be considered a suitable alternative from both environmental and economical points of view. acknowledgments the financial support by the international bureau in germany, helpful discussions of dr. w. pyckhout and dr. s. khale at fzj, germany are greatly acknowledged. references abdel-goad, m., 2000, "phd thesis," muenster university, germany. airey, g.d., 2003, "fuel," vol. 82, pp. 1709-1719. airey, g.d., 2003, "fuel," vol. 82, pp. 1709-1719. anhorn j. l. and badakhshan a., 1994, " fuel," vol. 73, pp. 1499-1503. bonemazzi f. and giavarini c., 1999, "journal of petroleum science and engineering," vol. 22, pp. 1724. champion-lapalu, l., wilson, a., fuchs, g., scramoncin, c., martin, d., planche, j.p., 2002, "energy & fuels," pp. 143-147. eastick, r. r. and mehrotra, a. k., 1990, "fuel processing technology," vol. 26, pp. 25-37. el akrami, h.a., yardim, m.f., ekinci, e., 1999, "energy & fuels," vol. 13, pp. 1030-1036. fawcett, a. h. and mcnally, t., 2000, "polymer," vol. 41, pp. 5315-5326. fawcett, a.h. and lor, s.k., 1992, "polymer," vol. 33, pp. 2003-2006. ferry, j.d., 1980 "viscoelastic properties of polymers", 3rd ed. ( wiley, new york). garcía-morales, m., partal, p., navarro, f. j., martínezboza, f., gallegos, c., gonzález, n., gonzález, o. and muñoz, m.e., 2004, "fuel," vol. 83, pp. 31-38. garcía-morales, m., partal, p., navarro, f.j., martínezboza, f., gallegos, c., gonzález, gonzález, o., and muñoz, m.e., 2004, "fuel," vol. 83, pp. 31-38. garcìa-pèrez m., chaala a., yang j. and roy c., 2001, "fuel," vol. 80, pp. 1245-1258. herrington, p.r., wu, y. and forbes, m.c., 1999, "fuel," vol. 78, pp. 101-110. japanwala, s., chung, k.h., dettman, h.d., gray, m.r., 2002, "energy & fuels," vol. 16, pp. 477484. lu, x. and isacsson, u., 2000, "polymer testing," vol. 20, pp. 77-86. mastrofini, d. and scarsella, m., 2001, "fuel," vol. 42, p. 24. mehrotra, a. k., 1991, "journal of petroleum science and engineering," vol. 5, pp. 261-272. michaelian k.h., zhang s.l., hall r.h. and bulmer j.t., 2001, "spectrochimica acta part a: molecular and biomolecular spectroscopy," vol. 57, pp. 73-81, 2001. oh, k., deo, m.d., 2002, "energy & fuels," vol. 16, pp. 694-699. parker, r.j., mcfarlane, r.a., 2000, "energy & fuels," vol. 14, pp. 11-13. pérez-lepe, a., martínez-boza, f.j., gallegos, c., gonzález, o., muñoz, m.e. and santamaría, a., 2003, "fuel," vol. 82, pp. 1339-1348. proctor, l.m., toy, e., lapham, l., cherrier, j., chanton, j.p., 2001, "environ. sci. technol," vol. 35, pp. 14201424. raki, l., masson, j.f., collins, p., 2000, "energy & fuels," vol. 14, pp. 160-163. saha, a.k., das, s., bhatta, d. and mitra, b.c., 1999, "j. of applied polymer science," vol. .71, pp. 15051513. yarranton, h.w., lboudwarej, h.a., jakher, r., 2000, -20 -10 0 10 20 30 40 50 60 70 101 102 103 104 105 106 107 108 neat bitumen bit.-9% pe η* p a. s temp., °c figure 10. η* of bitumen/pe and neat bitumen as a function of temperature at ω = 0.1 rad/s 81 the journal of engineering research vol. 6, no.1 (2009) 76-81 "ind. eng. chem. res," vol. 39, pp. 2916-2924. yui s., 1997, "fuel and energy abstracts," vol. 38, pp. 78. zhao, l. s. kotlyar, j. r. woods, b. d. sparks, k. hardacre and k. h. chung, 2001, "fuel," vol. 80, pp. 1155-1163. zhao, y., gray, m.r., chung, k.h., 2001, "energy & fuels," vol. 15, pp. 751-755. chen060503.qxd the journal of engineering research vol. 4, no.1 (2007) 103-119 _______________________________________ *corresponding authors e-mail: hfa101@psu.edu numerical modeling of foam drilling hydraulics ozcan baris2, luis ayala,*1, and w. watson robert1 *1 petroleum and natural gas engineering program, the pennsylvania state university,122 hosler building, university park, pa 16802, usa 2 turkish petroleum company, thrace district management 39750 luleburgaz/kyrklareli, turkey received 3 may 2006; accepted 4 october 2006 abstract: the use of foam as a drilling fluid was developed to meet a special set of conditions under which other common drilling fluids had failed. foam drilling is defined as the process of making boreholes by utilizing foam as the circulating fluid. when compared with conventional drilling, underbalanced or foam drilling has several advantages. these advantages include: avoidance of lost circulation problems, minimizing damage to pay zones, higher penetration rates and bit life. foams are usually characterized by the quality, the ratio of the volume of gas, and the total foam volume. obtaining dependable pressure profiles for aerated (gasified) fluids and foam is more difficult than for single phase fluids, since in the former ones the drilling mud contains a gas phase that is entrained within the fluid system. the primary goal of this study is to expand the knowledge-base of the hydrodynamic phenomena that occur in a foam drilling operation. in order to gain a better understanding of foam drilling operations, a hydrodynamic model is developed and run at different operating conditions. for this purpose, the flow of foam through the drilling system is modeled by invoking the basic principles of continuum mechanics and thermodynamics. the model was designed to allow gas and liquid flow at desired volumetric flow rates through the drillstring and annulus. parametric studies are conducted in order to identify the most influential variables in the hydrodynamic modeling of foam flow. keywords: drilling, foam flow, modeling iƒzôdg ωg~îà°séh »µ«eéæjohq~«¡dg ôøë∏d »ªbq êppƒ‰ äôhhq ¿ƒ°ùjcgh ,’éjcg ¢ùjƒd ,¢ùjqéh ¿é°srhcg üƒ≤k çg~mg á«∏ª©h iƒzôdéh ôø◊g ±ô©jh .iôn’g ájoé©dg ôø◊g πfgƒ°s é¡∏x ‘ â∏°ûa »àdg ±hô¶dg øe áyƒª› ™e ≥agƒà«d qƒw ôøë∏d πfé°ùc iƒzôdg ωg~îà°sg ¿g :á°uóÿg ,πfgƒ°ùdg ¿gqho ¿g~≤a á∏µ°ûe öæœ :éjgõÿg √òg øe .¿rgƒà∏dg ᣰsgƒh ôø◊g hg …~«∏≤àdg ôø◊g á≤jô£c iôn’g ôø◊g ¥ôw ™e áfqé≤ÿéh éjgõe i~y ád iƒzôdg ωg~îà°séh ôø◊g ¤g π°uƒàdg èà©j .iƒzô∏d »∏µdg ºé◊gh ,ré¨dg ºém áñ°ùf ,ioƒ÷g ≈∏y iƒzôdg ∞«æ°üj ~ªà©j .ôø◊g ö≤ãe ié«m ioéjrh ¥gîn’g ä’~©e ´éøjqg ,êéàf’g á≤£æÿ qô° dg π«∏≤j úw ¿’ ∂dph i~mgh á∏môã ádƒëàÿg πfgƒ°ù∏d º«≤dg ¢ùøf ≈∏y ∫ƒ°ü◊g øe áhƒ©°u ìcg á«∏ªy iƒzôdgh (ré¨dg ¤g ádƒëàÿg) igƒ¡ÿg πfgƒ°ù∏d é¡«∏y πjƒ©àdg øµ‡ §¨° ∏d º«b ᫵«eéæjohq~«¡dg ôggƒ¶∏d á«aô©ÿg i~yé≤dg ™«°sƒàh á°sgq~dg √òg øe »°sé°s’g ±~¡dg πãªàj .πfgƒ°ùdg ≥a~j ™e ≥agîj réz ≥a~j á∏môe ≈∏y ¤h’g ádé◊g ‘ πªà°ûj ôø◊g ” ,ájé¨dg √ò¡dh .áø∏àfl á«∏«¨°ûj ±hôx ‘ á≤«ñ£j ” ºk øeh »µ«eéæjohq~«g êpƒ‰ ôjƒ£j ” ~≤d ìcg iƒzôdg ᣰsgƒh ôø◊g á«∏ªy º¡ødh .iƒzôdéh ôø◊g á«∏ª©d áñmé°üÿg ≥a~àh íª°ù«d ºª°u êpƒªædg gòg ¿g ôcòdéh ôj~lh .ájqgô◊g ä髵eéæj~dgh ᫵«f鵫ÿg á«∏°ù∏°ùà∏d á«°sé°s’g çoéñÿg ¤g oéæà°s’g ∫ón øe ôø◊g ωé¶f ‘ iƒzôdg ≥ao π«µ°ûj ∫g êpƒªædg ≈∏y gòkéj ìc’g ägò¨àÿg ~j~– ᫨h ájqé«©e äé°sgqo i~y âjôlcg ~bh .¥ƒ£dg á≤∏mh ö≤ãdg ö«° b ‘ ié¨àñÿg ᫪é◊g ≥a~dg ä’~©e ö°ùëh πfgƒ°ùdgh ré¨dg .iƒzôdg ≥a~d »µ«eéæjohq~«g .êpƒ‰ π«µ°ûj ,iƒzôdg ≥ao ,ôøm :á«méàøÿg ägoôøÿg notation a = cross-sectional area (l2) a = okpobiri and ikoku parameter aw = area of fluid in contact with the wall per unit volume (l 2/l3) at = total nozzle area, (l 2) [a] = left-hand side matrix of the hydrodynamic formulation b = okpobiri and ikoku parameter [b] = right-hand side vector of the hydrodynamic formulation f ia = fehlberg constant coefficient of runge-kutta method d = pipe diameter (l) dh = hydraulic diameter (l) dij f = fehlberg constant coefficient of runge-kutta method 104 the journal of engineering research vol. 4, no.1 (2007) 103-119 e = okpobiri and ikoku parameter fw = wall shear force per unit volume (m/l 2-t2) fg = gravitational force per unit volume (m/l 2-t2) f if = fehlberg variable coefficient of runge-kutta method fw = fanning friction factor (dimensionless) g = acceleration of gravity (l/t2) gc = unit conversion factor (32.174 lbm-ft/lbf-s 2) id = pipe inner diameter (l) od = pipe outer diameter (l) k = consistency index (m/l-t) m& = mass flow rate (m/t) mw = molecular weight of the fluid (lbm/lbmol) n = power-law index (dimensionless) p = pressure of the system (m/l-t2) po = pressure immediately above the bit, or bit upstream pressure (m/l-t 2) q = volumetric flow rate (l3/t) r = pipe radius (l) or universal gas constant (10.7315 psia-ft3/lbmol-°r) [r] = vector containing the derivatives of the main unknowns of the hydrodynamic formulation re = reynolds number (dimensionless) rop = rate of penetration (l/t) t = temperature of the system (t) ti = temperature at gridblock “i” (t) ti+1 = temperature at gridblock “i+1” (t) [u] = vector containing the main unknowns of the hydrodynamic formulation v = volume (l3) vf = foam velocity (l/t) vn = nozzle velocity (l/t) z = compressibility factor (dimensionless) x = x-direction, coordinate parallel to flow (l) subscripts 1 = annulus’ outer pipe 2 = annulus’ inner pipe dp = drillpipe dc = drillcollar f = foam g = gas i = running index j = running index in = initial m = mixture s = solid o = oil, or reference point w = water l = liquid greek symbols â = volumetric thermal expansivity of the liquid (1/t) å = pipe roughness (l) φ = porosity (dimensionless) è = angle of inclination with respect to the vertical (rad) ê = isothermal compressibility of liquid (l-t2/m) ì f = foam effective viscosity (m/l-t) ð = ratio of the circumference of a circle to its diameter = 3.14159265(…) ñ = density (m/l3) äpb = pressure drop across the bit (m/l-t 2) äx = incremental segment length or block size or step size (l) ã = foam quality (dimensionless) 105 the journal of engineering research vol. 4, no.1 (2007) 103-119 1. introduction the conventional method of drilling is to use either water-based or oil-based muds. however, the use of these drilling muds may result in increased formation damage or a significant reduced rate of penetration. compared to conventional drilling, underbalanced or foam drilling has several advantages. these advantages include: avoidance of lost circulation problems, minimizing damage to pay zones, higher penetration rates and bit life. while air or mist drilling depends on high volumetric flow rate, foam drilling relies on bubble strength to remove the cuttings. there are also potential disadvantages to the use of foam for drilling. these include: corrosion, downhole fires, waste water disposal, and cost of consumables. foam is composed of a continuous liquid phase that surrounds and traps the gaseous phase and the most common way to categorize foams is with respect to their qualities. foam quality is the ratio of the gas volume to the total volume and is expressed as: (1) raza and marsden (1967) performed an empirical study of foam flow through tubes. they concluded that foams behave like a pseudoplastic fluid. david and marsden (1969) took into consideration the fluid slippage at the tube wall and the compressibility of foam. they described the rheology of foam in terms of these two factors and described foam as a pseudoplastic fluid with low gel strength. in terms of foam rheology, one of the earliest studies was presented by millhone, et al. (1972). they used a mathematical model to predict trends such as optimum gas and liquid flow rates, pressures, circulation time and solids-lifting capability of the drilling foam. in terms of investigating foam rheology, they plotted viscosity of foam versus quality of foam which indicates that the viscosity increases as the quality increases. mitchell (1970) studied the viscosity of foam by performing experiments using capillary tubes. according to him, foam behaves similarly to a bingham plastic fluid. his observations led him to ignore the wall slippage effect during foam flow. beyer, et al. (1972) study was one of the first studies which included a mathematical model to describe the flow of foam in vertical pipes and annuli. the model was a 1d, steady-state model that uses a bingham plastic model for all rheological calculations. using the real gas law, lord (1981) derived an equation of state for foams. he presented an equation of state that accounted for the presence of solids such as propants (utilized in hydraulic fracturing) or cuttings (in drilling operation). sanghani and ikoku (1982) used concentric annular pipe viscometer in their experiments performed to investigate the rheology of the foam. according to them, flowing foam behaves as a pseudoplastic fluid without a yield value for shear rates in the range 150 sec-1 to 1000 sec-1 and that when the quality is kept constant, effective viscosity becomes proportional to shear rate. okpobiri and ikoku (1986) proposed a semi-empirical method for predicting frictional pressure losses for the contiguous flow of foam and cuttings. a model that predicts the pressure drop across the bit nozzles for foams was also presented. additionally, they came up with a technique to predict minimum volumetric requirements for foam and mist drilling operations. ozbayoglu, et al. (2000) not only conducted experiments on foam rheology but also compared six different rheological models with the data obtained from their experiments. they observed that foam behaves like a power law fluid at lower qualities and behaves like a bingham plastic fluid for qualities above 90%. ozbayoglu, et al. (2003) analyzed cutting transport phenomena with foam in horizontal and highly-inclined wells. they performed experiments to verify the results they obtained from the computer model. they developed a one-dimensional, three-layer model and calculated the fluid properties and pressure profile along the wellbore using the principles of mass and momentum balance equations for steady, isothermal conditions. kuru, et al. (2004) proposed a new methodology for hydraulic optimization of foam drilling for maximizing drilling rate. the procedure includes the determination of optimum combination of gas/liquid ratio, back pressure and total flow area. they adopted okpobiri and ikoku's approach for bit pressure drop calculations. they proposed a linear relationship between the optimum back pressure and the depth of the well. lourenço, et al. (2004) carried out an empiricallybased study which explored the effects of foam quality, foam texture, pressure, temperature, and geometry of the conduit on the rheology of the foam. they concluded that texture of the foams also has a significant influence on viscosity. the authors also stated that wall slippage is another important parameter in characterizing foam flow. they developed empirical correlations for the slippage coefficient that are independent of foam quality. ozbayoglu, et al. (2005) investigated the effect of bubble size and texture of the foam on its rheological properties. according to the authors, even if foam qualities and flow conditions are kept constant, the rheological properties also change with changes in surfactant. li and kuru (2005) recently presented a 1-d, unsteady-state, twophase mechanistic model of cuttings transport with foam in vertical wells. the model predicts optimum foam flow rate and rheological properties to maximize the cuttings transport efficiency. 2. description of the model a single-phase, one-dimensional, steady-state hydraulic model is developed to describe the flow of foam inside the drillstring and annuli. foam flow is considered to be that of a pseudo-single phase (homogeneous model) and it is assumed that the flow can be fully characterized by solving the equations of conservation simultaneously. the continuity and momentum equations that describe the system of one-dimensional, area-averaged governing equaviv v g g + =γ 106 the journal of engineering research vol. 4, no.1 (2007) 103-119 tions written in their non-conservative forms, written for the foam mixture, are given below: continuity: (2) momentum balance: (3) these two equations are used in the proposed hydraulic model expressed in customary field units. for simplicity, the effect of temperature is superimposed in the model by implementing prescribed temperature gradients. as a consequence of this, an energy balance equation was not required since it is assumed that the wellbore temperature gradient is known and can be inputted. the assumption here is that injection of foam has little impact on the wellbore temperature, which is dominated by the thermal condition of the surroundings (ie. geothermal gradients that are particular to the region). once the temperature gradient is known, the following equation is used to calculate the temperature at each pipe increment: (4) 3. closure relationships the closure relationships that define the various parameters of the mass and momentum balance equations are presented and discussed below. these parameters are the wall shear force (fw) the gravitational force (fg), equation of state, and rheological model. field units, as incorporated in the hydraulic model, are used throughout this section. wall shear force (fw): wall shear force per unit volume (lbf/ft3) between the wall and the fluid is calculated a function of the fanning friction factor: (5) where the wetted area per unit volume (aw) is defined as: (6) inside the drillpipe, which is a single pipe system, the wetted area becomes: (7) inside the annulus, where the flow occurs between two pipes, the wetted area becomes: (8) where r1 and r2 are the radii of the outer and inner pipe in the annulus respectively and the "d" represents the diameter of the corresponding pipe. the fanning friction factor (fw) is calculated as a function of reynolds number (re). reynolds number is a dimensionless factor that incorporates fluid's density, velocity, effective viscosity and hydraulic diameter (dh) and it is expressed as: (9) for circular tubes, the hydraulic diameter is calculated as the wetted equivalent diameter, which is computed as shown below: (10) for a single circular tube, this definition becomes: (11) while for two concentric tubes, one writes: (12) once the reynolds number and the equivalent diameter are calculated, the friction factor can be estimated. to determine the friction factor, the flow must be determined to be either laminar or turbulent. by convention, when the reynolds number is less than 2100, laminar flow is present. the analytical expression from the moody chart can then be used: (13) for turbulent flow, an explicit form of the colebrook correlation derived by chen (1979) is used in this study. this correlation is written in terms of reynolds number and pipe roughness (ε) takes the form: (14) 0 x v v dx dt t v dx dp p v ff p f f t f f =∂ ∂ +⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ +⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ ρρ wg f ff ffdx dv v dx dp −−=+ ρ x dx dt tt i1i ∆×⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +=+ where ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ dx dt is the temperature gradient (°f/ft) defined by the user and äx (ft) is the incremental segment length. c fff www g2 vv faf ρ = volumetotal fluidbywettedareawalltotal aw = d 4 r 2 xr xr2 a 2w === ∆π ∆π 21 2 2 2 1 21 2 2 2 1 21 w dd 4 )rr( )rr(2 x)rr( x)rr(2 a − = − + = − + = ∆π ∆π f hff dvre µ ρ = fluidbywettedareawalltotal fluidofvolume 4d h = dr2 xr2 xr 4d 2 h === ∆π ∆π 21 21 2 2 2 1 h ddx)r2r2( x)rr( 4d −= + − = ∆ππ ∆ππ re 16 f w = ( )1.1098 0.8981 1 5.0452 5.8506 4.0 log log 3.7065 re 2.8257 re εε⎡ ⎤⎛ ⎞⎢ ⎥⎜ ⎟= − − + ⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦ h hw d df 107 the journal of engineering research vol. 4, no.1 (2007) 103-119 with the use of wetted area and friction factor, the wall shear force (fw) can be determined after calculation of density and velocity. density is obtained from the thermodynamic model which is based on the equations of state for liquids and gases presented below. velocity is obtained by solving the mass and momentum balance equations. gravitational force (fg): the gravitational force term takes into consideration the density change with depth and the inclination of the well. for a homogeneous fluid, the gravitational force (fg) can be expressed as: (15) equation of state: foam is composed of a gas and a liquid phase. in this study, the thermodynamic properties of these two phases are computed using the peng-robinson (1976) equation of state and the slightly compressible model for the gas and liquid phases, respectively. the equation of state for the liquid phase is written as: (16) (17) where the cubic polynomial form of the peng-robinson eos is solved for the z factor using newton-raphson technique. the equation of state for the foam is written in terms of quality, as shown below: (18) foam viscosity: sanghani and ikoku (1982) studied foam rheology and correlated the power law index (n) and consistency index (k) in terms of foam quality. li and kuru (2005) modified these indices by performing regressional analysis on the data generated in sanghani and ikoku's work and proposed the following equations: for 0.915 < γ < 0.98: k = -2.1474.γ + 2.1569 n = 2.5742.γ − 2.1649 (19a) for γ < 0.915: k = 0.0074.e3.5163.γ n = 1.2085.e-1.9897.γ (19b) once the k and n indices are obtained, the effective viscosity of the foam is calculated from the equation developed by sanghani and ikoku (1982). for flow within the drillpipe, the effective foam viscosity is calculated as: (20a) and for flow through the annulus, (20b) 4. pressure drop across the bit okpobiri and ikoku (1982) proposed the following implicit model for estimating bit pressure drop (∆pb) for foam flow: (21a) where: (21b) (21c) (21d) (21e) 5. handling cuttings based on the values of rate of penetration (rop), porosity, densities of the formation fluids and formation rock at bottomhole conditions, and the saturation distribution of the formation fluids, the proposed model calculates actual mixture density values within the annulus by taking the weighted average of the foam and cuttings densities in terms of mass flow rate. cuttings are introduced to the system after the bit, and they are composed of solids, water, oil, and gas coming out of formation. the mass flow rate of cuttings is calculated as: (22) ρ θ=g f c g f cos g ( )( ) ( )( ) 0 0 0 1 1 1ρ ρ β κ⎡ ⎤= + − − + −⎣ ⎦l l t t p p where â is the volumetric thermal expansivity and ê is the isothermal compressibility. the equation of state for the gas phase is written in terms of the compressibility factor as follows: ρ = gg pmw zrt ( )1ρ ρ ρ= γ + − γf g l 1n f f d v8 n4 1n3 k − ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +=µ , 1n 12 f f dd v12 n3 1n2 k − ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ − ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ +=µ . 2ln⎡ ⎤− ∆∆ = +⎢ ⎥ ⎣ ⎦ o b b n o p pa p ev b p ff g mmw zrtm a & & = , fl l m m b & & ρ = , cga2 1 e = , po = pressure immediately above the bit or upstream pressure, in psia, and vn = nozzle velocity, in ft/s, calculated as a function of foam mass flow rate ( fm& ), foam density ( fρ ), and total nozzle area (at), as follows: ρ = & f n f t m v a ; lgf mmm &&& += gowsolidscuttings mmmmm &&&&& +++= 108 the journal of engineering research vol. 4, no.1 (2007) 103-119 where: the subscripts "w", "o", and "g" stand for formation water, oil, and gas, respectively. the mixture density calculation that includes the effect of both cuttings and foam is: (23) where the subscript "f" stands for the value of cuttingsfree foam density obtained through eq. (18). mixture density (ρm) replaces cuttings-free foam density during hydraulic calculations (eqs. 2 and 3) taking place within the annulus. 6. numerical procedure equations (2) and (3) represent a system of first-order ordinary differential equations (odes) that describe the hydrodynamics of foam flow. pressure and velocity are the two principal unknowns of the system, while temperature is defined by the geothermal gradient, as described by eq. (4). in order to solve the set of ordinary differential equations simultaneously, the runge-kutta method is used in this study. this study implements the method proposed in the hydrodynamic model of ayala and adewumi (2003), based on the cash & karp embedded rungekutta procedure, described by press et al. (1994). in this method, the differential equations are written in a matrix form as follows: (24) where [a] is a 2x2 square matrix and [b] is a 2x1 column matrix. the derivative matrix [u] is called the solution vector, which includes both pressure and velocity values. one must write eq. (24) explicitly in terms of the derivative matrix [u] in order to make it suitable for rungekutta procedure as shown below: (25) thus, [r] can be written in terms of the matrix elements of [a] and [b] as follows: (26) where: (27) therefore, the two differential equations that compose the model (eqs. 2 and 3) can thus be written in a form analogous, in field units, as shown below: (28) runge-kutta method calculates the solution vector u(x + ∆x) as a function of uo. as stated in the above section, u(x) contains pressure and velocity of the fluid for this study. the method gives the fifth estimate as: (29) (30) (31) 7. results and discussion the influence of different input parameters on foam flow can be evaluated separately. the parameters that are considered to be influential on foam flow calculations through a drillstring are: geometry of the system (ie. pipe diameters), geothermal gradient of the field, quality of the foam at the surface, total flow rate at standard conditions, and rate of penetration (rop). a base case was constructed so that the effect of each parameter is analyzed with respect to a certain reference point. for the base case, foam consists of nitrogen and water with flow rates of 2 mscf/min and 5 gal/min respectively at 5,000 psia and 65°f. table 1 summarizes these conditions, where it is assumed that the hole is cased for most of the total depth and that the influence of the openhole section on the hydrodynamics of the flow is minimal. in the subsequent sections, this case will be used as the base-line for comparison purposes. ssolids )1(aropm ρφ ⋅−⋅⋅=& iii saropm ⋅⋅⋅⋅= ρφ& i = w, o, g cuttings cuttings f f m cuttings f m m m m ρ ρ ρ × + × = + & & & & ]b[]u[ dx d ]a[ = [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] 1 1 where d u a b dx d u r r a b dx − − = = = ⎥⎦ ⎤ ⎢⎣ ⎡ +− − − = 2113 2214 3241 baba baba aaaa 1 ]r[ ⎥⎦ ⎤ ⎢⎣ ⎡=⎥⎦ ⎤ ⎢⎣ ⎡= 2 1 43 21 b b]b[ ;aa aa]a[ ( ) ( ) 2 2 1 144 144 f f f f w g c p f f f f f f f f f w g c t p t v dt f f g t dxpd v vdx dt v v f f g p t dx p ρ ρ ρ ρ ρ ρ ρρ ⎡ ⎤∂⎛ ⎞− + +⎢ ⎥⎜ ⎟∂⎡ ⎤ ⎝ ⎠⎢ ⎥=⎢ ⎥ ⎢ ⎥∂⎛ ⎞ ∂ ∂⎛ ⎞ ⎛ ⎞⎣ ⎦ − − +⎢ ⎥⎜ ⎟ ⎜ ⎟ ⎜ ⎟∂ ∂ ∂⎝ ⎠ ⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦ ( ) 6 5 1= + ∆ = + ∑ f fth o i i i u x x u a f where fia values are constant coefficients proposed by cash & karp (1990). fif values are obtained by using the [r] matrix calculated from eq. (26) as: 6i1,fdurxf fj f ij 1i 1j o f i ≤≤⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ += ∑ − = ∆ where fijd values are also proposed by cash & karp. then the final solution fo r u(x+äx) can be calculated by expanding eq. (29) as: ( ) f 6 f 6 f 5 f 5 f 4 f 4 f 3 f 3 f 2 f 2 f 1 f 1o fafa fafafafauxxu ++ ++++=+ ∆ 109 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 1 shows the pressure and temperature profiles in the drillstring and annulus. pressure increases as depth increases towards the bottom of the well and starts to decrease as the flow goes through the annulus. this is an indication that the gravitational force is dominant as compared to frictional force. the steeper slope in the annulus section reflects the fact that both the gravitational and frictional forces are formed in the opposite direction of the flow. figure 2 demonstrates that, although there is a significant change in pressure throughout the wellbore, the velocity of the foam does not change significantly except for the sections where the geometry of the system changes. the two sharp changes in velocity indicate the presence of a diameter change. the reason for the small slope in the velocity profile in fig. 2 (drillstring and annulus) can be explained by fig. 3, which shows the density profiles. as it is seen from fig. 3, foam density changes slightly except for the section where the transition from the drillstring to the annulus takes place. that section is the place where the cuttings enter the foam and make it heavier mostly because of the more dense rock particles. since the velocity calculations are mostly driven by the density, for the case of drillstring and annulus, no dramatic change is seen in the velocity distribution. in fig. 4, a different slope in the quality profiles of the drillstring and annulus can be easily recognized. as given in eq. (1), foam quality depends on both gas and liquid volumes. however, since the liquid is assumed to be slightly compressible and there is no condensation of gas, the change in the volume of the liquid phase is negligible. therefore, there is an increase in foam quality inside the drillstring because of the increase in gas volume (decrease in gas density). the decrease in density is the result of the fact that rather than pressure effect, temperature effect dominates the gas density calculations. throughout the annulus, pressure and temperature work together to make the gas density decrease with steeper slope and that makes the quality more affected. 7.1 effect of quality the quality of foam at the surface can be changed either by changing the gas and liquid injection flow rates specified at standard conditions or by altering injection conditions (pi, ti). the effect of different values of initial quality on the hydrodynamics of the system is investigated by changing injection rate specification at standard conditions. to understand the effect of quality, the program was tested with a constant gas flow rate of 2-mscf/min and with three different liquid flow rates of 12.5, 8.85, and 2.64-gal/min forming a foam with qualities of 80%, 85% and 95%. figure 5 shows the pressure profiles for different quality values at surface injection conditions. in this plot, one can see that as the quality at the surface decreases, the slope of the pressure profile increases because of the higher density of low quality foams. the opposite effect on velocity profile is demonstrated in fig. 6. lower qualities imply higher density values because of the higher liquid fraction. this fact is seen in fig. 7. changes in surface quality become unimportant after the bit because of the contribution of the higher density rock particles to foam density. the surface quality affects only the magnitudes of quality profiles. the trends are all the same for four different surface qualities (fig. 8). 7.2 effect of temperature gradient vs. isothermal assumption injection pressure, psia 5,000 injection temperature, °f 65 total depth, ft 10,000 qg at the injection point, mscf/min 2 ql at the injection point, gal/min 5 quality at the injection point, ã 90.9% total flow rate, scf/s 33.3 iddp, in 5 oddp, in 5.5 iddc, in 4.5 oddc, in 7 idcasing, in 9.25 drillcollar depth, ft 9,500 nozzle diameter, in 3 x 13/32 pipe roughness, in 0.00015 temperature gradient, °f/ft 0.015 rop, ft/min 0.5 porosity, ? 0.25 rock density, lbm/ft3 170 formation water density, lbm/ft3 64 formation oil density, lbm/ft3 45 formation oil saturation, so 0.3 formation water saturation, sw 0.4 formation gas saturation, sg 0.3 liquid compressibility factor, 1/psi 1.00e-06 liquid volume expansivity, 1/°f 1.00e-06 foam water + nitrogen table 1. base case conditions two different temperature gradients other than the base case were tested using the proposed model. one of them is a quite large value (0.03 °f/ft) whereas the other one is taken as 0 °f/ft which represents isothermal conditions. the effect of temperature gradient can be clearly seen from all of the plots given in figs. 9 to 12. figure 9 shows that as the temperature increases more rapidly, the slope of the pressure profile decreases. this is because of the direct influence of temperature gradient on the pressure derivative calculations in eq. (28). one can easily see that the temperature gradient has a positive contribution to dx dv and a negative contribution to dx dp . this fact reflects itself in fig. 10 by affecting the velocity profile trend. the velocity profile changes its trend into the opposite 110 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 1. pressure and temperature profiles inside the drillstring and annulus (base case) figure 2. velocity profile inside the drillstring and annulus (base case) figure 3. density profiles inside the drillstring and annulus (base case) 111 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 4. quality profile inside the drillstring and annulus (base case) figure 5. pressure profiles inside the drillstring and annulus (effect of quality) figure 6. velocity profiles inside the drillstring and annulus (effect of quality) 112 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 7. density profiles inside the drillstring and annulus (effect of quality) figure 8. quality profiles inside the drillstring and annulus (effect of quality) figure 9. pressure profiles inside the drillstring and annulus (effect of temperature gradient) 113 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 10. velocity profiles inside the drillstring and annulus (effect of temperature gradient) figure 11. foam density profiles inside the drillstring and annulus (effect of temperalture gradient) figure 12. quality profiles inside the drillstring and annulus (effect of temperature gradient) 114 the journal of engineering research vol. 4, no.1 (2007) 103-119 the main reason for most of the changes in the hydrodynamics of the system lies on fig. 11 which shows the foam density profiles with changing temperature gradients. since the gas density is highly dependent on temperature, when there is a change in temperature, it is reflected first in the density, then in the quality, and finally in pressure and velocity. there is a clear difference in foam density trends in fig. 11. in fact the density profile changes its trend from a negative slope to a positive slope. this result can be explained by real gas law. since temperature is inversely proportional to gas density, when the geothermal gradient changes, an effect in density trend is expected. the gradient selected for the base case makes the density profile almost flat. that is why when the gradient is changed; the slope changes its sign. as a consequence of this, quality trends are similar to the ones in the density profiles (fig. 12). 7.3 effect of total flow rate in a foam system, the quality can be kept constant even when the flow rates of gas and liquid phases are changed if the properties between gas and liquid phases are maintained. therefore one can increase or decrease the total flow rate while keeping the surface quality constant. in this section, the analysis of the total flow rate was done by using three different combinations of gas and liquid rate. table 2 summarizes these combinations. as it can be seen from figs. 13 through 15, the effect of total flow rate is not significant in the drillstring. in fact, it only changes the initial velocity of the foam which makes the velocity profiles different from each other (fig. 14). since the model uses the ratio of the gas and liquid flow rates rather than the total flow rate as a parameter used in the calculations, changing only the total flow rate without a change in quality does not play an important role. the real effect can be seen after the bit where the new density of the system is calculated using the total mass flow rate which is calculated from the initial gas and liquid flow rates. since the rate of penetration was kept constant for this analysis, the mass flow rate after the bit is directly proportional to the initial flow rates. this fact is reflected in all profiles. as the total flow rate increases the slope of the pressure (fig. 13) and quality profiles (fig. 15) also increase. this is mostly because of the change in gravitational force. 7.4 effect of wellbore dimensions this section discusses the effect of the wellbore geometry which consists of the inner and outer diameters of drillpipes, drillcollars, and inner diameter of the casing. bit diameter and casing setting depth were kept constant since the changes in diameters of pipes are considered to be sufficient to see the effect of the geometry. the diameters of different pipes are shown in table 3. most standard drillpipe, drillcollar, and casing sizes are bounded by the minimum and maximum reference values indicated in table 3. in this section, the effect of broad changes in the dimensions of casing, drill pipe, and drill collars on the hydrodynamic profiles is demonstrated. for instance, the effect of smaller diameter can be easily recognized in figs. 16 to 18. the effect of larger diameter is not significant because its cross-sectional area is relatively close to the one of the base case. on the other hand, the cross-sectional area of the 2.5-in diameter pipe is almost one of tenth of the pipe with 7.5 in diameter. since the area directly affects the velocity, there is significant difference of the velocity profiles inside the drillpipe (fig. 17). the difference in velocity profiles affects the frictional force calculations, although the gravitational force is not affected at all. the influence of this difference can be clearly seen in the pressure and quality profiles clearly (figs. 16 and 18). 7.5 effect of rate of penetration (rop) cuttings mass flow rate is controlled by the rate of penetration and the bit diameter. however, if the bit diameter is changed to analyze the effect of cuttings, then other pipe diameters should also be changed in order to maintain reasonable results. that would cause a change in well geometry whose effect is discussed separately. therefore, changing the rop is used to analyze the effect of cuttings entering the fluid system. four different rop values were considered: 0 ft/min (no penetration, only circulation), 0.25 ft/min, 0.5 ft/min (base case) and 1 ft/min. as with the nozzle size effect and since the cuttings are introduced into the system after the bit, only the plots in the annulus were included. figures 19 to 21 indicate that the rop total flow rate (scf/s) gas flow rate (scf/min) liquid flow rate (gal/min) quality (%) 8.33 1.25 0.5 90.9 16.67 2.5 1 90.9 33.3 (base case) 5 2 90.9 50.02 7.5 3 90.9 table 2. combination of gas and liquid flow rates (effect of total flow rate) iddp, in oddp, in iddc, in oddc, in idcasing, in 2.5 3 2 4.5 6 5 (base case) 5.5 4.5 7 9.25 7.5 8 7 9.5 10 table 3. pipe diameters used in wellbore geometry analysis direction as the temperature gradient increases inside the drillstring. for the base case, the effects of pressure change and temperatur e gradients were compensating each other so that the change in velocity was gradual. since pressure and temperature changes have an opposite effect on fluid density, accounting for both effects is critical in hydrodynamic studies. 115 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 13. pressure profiles inside the annulus (effect of total flow rate) figure 14. velocity profiles inside the drillstring and annulus (effect of total flow rate) figure 15. quality profiles inside the drillstring and annulus (effect of total flow rate) 116 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 16. pressure profiles inside the drillstring and annulus (effect of wellbore geometry) figure 17. velocity profiles inside the drillstring and annulus (effect of wellbore geometry) figure 18. quality profiles inside the drillstring and annulus (effect of wellbore geometry) 117 the journal of engineering research vol. 4, no.1 (2007) 103-119 figure 19. pressure profiles inside the annulus (effect of rop) figure 20. velocity profiles inside the annulus (effect of rop) figure 21. quality profiles inside the annulus (effect of rop) 118 the journal of engineering research vol. 4, no.1 (2007) 103-119 mostly affects the density of the fluid after the bit which is in a direct relationship with gravitational force values inside the annulus. as a consequence, the pressure profiles are very different from each other. the change in density also triggers differences in quality and velocity profiles since they are both directly dependent on density. 8. conclusions this study has presented and developed a one-dimensional, steady-state hydraulic model of underbalanced drilling operations using foam. the model was tested for various injection conditions, well depths, foam qualities, bit and wellbore sizes, and temperature gradients to gain a better understanding of their influence on the overall drilling operation. conclusions derived from this work are described below: acknowledgments the authors wish to thank the turkish petroleum corporation (tpao) and the petroleum and natural gas engineering program at penn state university for the support provided to this study. references ayala, l.f. and adewumi, m., 2003, "low-liquid loading multiphase flow in natural gas pipelines," j. of energy resources and technology, trans. asme, vol. 125, pp. 284-293. beyer, a.h., millhone, r.s. and foote, r.w., 1972, "flow behavior of foam as a well circulating fluid," spe paper 3986 presented at the 47th spe fall meeting in san antonio, texas, usa. cash, j. r., karp a. h., 1990, "variable order rungekutta method for initial value problems with rapidly varying right-hand sides," acm transactions on mathematical software, vol. 16, pp. 201-222. chen, n.h., 1979, "an explicit equation for friction factor in pipe," industrial & engineering chemistry fundamentals," vol. 18(3), pp. 296-297. david, a. and marsden, s.s., 1969, "the rheology of foam," spe paper 2544 presented in the spe fall meeting of the society in denver, colorado. kuru, e., okunsebur, o. m. and li, y., 2004, "hydraulic optimization of foam drilling for maximum drilling rate," spe paper 91610 presented at the spe/iadc underbalanced technology conference and exhibition held in houston, texas, usa. li, y. and kuru, e., 2005, "numerical modelling of cuttings transport with foam in vertical wells," j. of canadian petroleum technology, vol. 44(3). lord, d. l., 1981, "analysis of dynamic and static foam behavior," j. of petroleum technology, pp. 39-45. lourenço, a. m. f., miska, s. z., reed, t. d., pickell, m. b. and takach, n. e., 2004, "study of the effects of pressure and temperature on the viscosity of drilling foams and frictional pressure losses," spe paper 84175, spe drilling & completion, vol. 19(3), pp. 139-146. millhone, r. s., haskin, c. a. and beyer, a. h., 1972, "factors affecting foam circulation in oil wells," spe paper 4001 presented in the spe fall meeting in san antonio, texas, usa. mitchell, b. j., 1970, "viscosity of foam," phd thesis, the university of oklahoma, oklahoma, 1970. okpobiri, g. a. and ikoku, c. u., 1982, "experimental determination of solids friction factors and minimum volumetric requirements in foam and mist drilling and well completion operations," final report, fossil energy, us department of energy, university of tulsa, oklahoma, usa. okpobiri, g. a. and ikoku, c. u., 1986, "volumetric requirements for foam and mist drilling operations," spe paper 11723, spe drilling engineering (february). • the quality of foam or the volume fraction of gas is directly affected by the density of the gas, which is impacted by variations in system pressure and temperature. moreover, when the change in temperature is larger or smaller than the change in pressure, a change in density and a corresponding change in the quality of foam occurs. • the assumption of isothermal conditions is not a reasonable assumption for the system under consideration given the impact of temperature on gas density and the corresponding impact of gas density on the quality of foam. • the quality of foam as realized at the wellhead can be predetermined by varying the volumetric flow rate of the gas and liquid, which in turn changes the mixture’s density at the surface. the changes in foam quality at the surface impact the pressure trends and velocity profiles with depth in the wellbore. these effects become negligible after the cuttings are mixed into the foam fluid because the change in fluid density is negligible compared to that of the cuttings density. • if the proportion of gas and liquid is kept constant, variations in total volumetric flow rate do not change the quality of the foam. the impact is seen however in the difference in total mass flow rate. the change in total mass flow rate does not affect the other variables in drillstring except for the velocity. • the gravitational force is almost always dominant when compared to that of the frictional force, and results in an increase in pressure when moving downhole. for the upward flow in the annulus, a decrease in pressure is experienced given that both the gravitational and frictional forces act together against fluid pressure. • the magnitude of velocity of the fluid flowing inside a wellbore is highly dependent on the density of the fluid. when the density is kept constant, it is seen that the change in velocity profile is negligible. however, if the density change is large enough, a significant change in velocity is observed. 119 the journal of engineering research vol. 4, no.1 (2007) 103-119 ozbayoglu, m. e., akin, s. and eren, t., 2005, "foam characterization using image processing techniques", spe paper 93860 presented at the spe western regional meeting held in irvine, california, usa. ozbayoglu, m. e., kuru, e., miska, s. and takach, n., 2000, "a comparative study of hydraulic models for foam drilling," spe paper 65489 presented at the spe/petroleum society of cim international conference on horizontal well technology held in calgary, alberta, canada. ozbayoglu, m. e., miska, s. z., reed, t. and takach, n., 2003, "cuttings transport with foam in horizontal & highly-inclined wellbores", spe paper 79856 presented at the spe/iadc drilling conference held in amsterdam, the netherlands. peng, d. and robinson, d.b., 1976, "a new twoconstant equation of state", industrial engineering chemical fundamentals, vol. 15(1), pp. 59-64. press, w. h., teukolsky, s.a., vetterling, w. t. and flannery, b. p., 1994, “numerical recipes in fortran: the art of scientific computing,” cambridge university press, second edition, pp. 701-716. raza, s. h. and marsden, s.s., 1967, "the streaming potential and rheology of foam," spe paper 1748, society of petroleum engineers journal, pp. 359-368. sanghani, v. and ikoku, c. u., 1982, "rheology of foam and its implications in drilling and cleanout operations, " topical report, the university of tulsa, oklahoma, usa. c:\users\raoh\desktop\paper_4.xps tjer 2012, vol. 9, no 1, 31-36 1. introduction water generally contains suspended and colloidal solids from land erosion, decaying, vegetation, microorganisms, and color-producing compounds. coarser materials such as sand and silt can be eliminated to a considerable extent by plain sedimentation, but finer particles, such as those between 1-100 mm, must be chemically coagulated to produce larger flocs which are removable in subsequent settling and filtration (mark 1975). in considering the aggregation of particles in a colloidal dispersion, it is useful to distinguish between two distinct steps: _________________________________________ *corresponding author’s e-mail: salam@unizwa.edu.om (1) particle transport to affect particle contacts, and (2) particle destabilization to permit attachment when contact occurs. particle transport in an aqueous systems is essentially a physical process, which is accomplished by brownian diffusion, fluid motion, and sedimentation. it can be controlled by physical parameters such as temperature, velocity gradient, and particle size. particle destabilization is clearly a colloidal chemical process which can be controlled by both chemical and physical parameters (stumm, charles 1968). there are many substances which react suitably with water to preparation and usage of polyaluminum chloride as a coagulating agent sk al-dawery*a and oh al-jouborib *a department of chemical engineering, university of nizwa, sultanate of oman b department of chemical engineering, university of baghdad, iraq received 5 january 2010; accepted 13 december 2010 abstract: three different coagulating agents, iraqi-produced alum, swedish-produced alum, and laboratory-produced polyaluminum chloride (pac) were used to clarify samples taken from iraqi river water. a laboratory rig was built to produce pac from the reaction between hydrochloric acid and aluminum powder. the results after raw water treatment showed that the dosage of pac required for treatment was 98% less than that of alum when it was applied to water at a low turbidity and almost 90% less in the case of much higher turbidity. the application of pac produced rapidly forming flocs and more compact sludge compared to that of alum. however, there were no significant changes in the ph of the solution. from the results obtained, the pac agent performed better compared to alum and might be considered a good alternative to alum for raw water and waste water treatment. keywords: alum, polyaluminum chloride, water treatment 32 sk al-dawery and oh al-joubori produce such an effect, known collectively as coagulants. the precipitate so formed in the water is called the floc. the larger and heavier the floc is, the quicker the rate of settlement (george 1988). commonly used coagulants in water treatment are: 1. coagulats based on aluminum, such as aluminum sulfate (al2(so4)3.18h2o), sodium aluminates (na3alo3), poly aluminum chloride [al2(oh)xcl6-x]n, potash aluminum (alk(so4)2 . 12h2o), and ammonia aluminum (alnh4(so4)2 . 12h2o), 2. coagulats based on iron, such as ferric sulfate (feso4.7h2o), chlorinated copperas, and ferric chloride (fecl3), and 3. polyelectrolytes, which are long-chain synthetic polymers with a high molecular weight. these organic chemicals are commercially available under a wide variety of trade names (mark 1975). generally, the type of coagulants and aids available are defined by the plant process scheme and of course, dosages of these substances can be regulated to meet changes in raw water quality. also, mechanical mixing can be adjusted by varying the speed of the flocculator paddles. jar tests are widely used to determine optimum chemical dosages for treatment. this laboratory test attempts to simulate the full-scale coagulation-flocculation process and can be conducted for a wide range of conditions. the interpretation of test results involves visual and chemical testing of the clarified water. pac has been found by others to be an acceptable alternative flocculating and coagulating agent for drinking water, waste water, and industrial water treatment (cleartech inc. 2008; alhadidi et al. 2009). halvadiya et al. (2008) used pac to clarify turbid lake water, in their work, and found that the required dosage of pac was 60% less than that of alum. many companies, such as prakash chemical agencies private of india and the qikang company of china manufacture pac for water treatment using different types of raw materials. the purposes of this work are as follows: 1experimental testing: since locally produced alum is highly impure and many operating problems can occur in its production (al-dawery, butris 2002), pac is considered an alternative coagulant and has to be tested experimentally. 2product comparison: the intent of this work was also perform a comparison between pac and other types of coagulants such as iraqi and swedish produced alums. 3waste aluminum use: to test the efficacy of makings use of waste aluminum powder from an aluminum plant to produce pac on a laboratory scale based on the reaction between aluminum powder and hydrochloric acid. 2. experimental work 2.1 material preparation 2.1.1 alum preparation as stated above, two types of alums, swedish produced alum (consisting of 9.63% al2o3) and iraqi produced alum (consisting of 9.28% al2o3) were used for comparison. for raw water treatment, an alum solution was required. for such, a 1% alum solution was prepared by dissolving 10 grams of crystal alum in 1 liter of distilled water. 2.1.2 polyaluminum chloride production poly aluminum chloride was produced at the laboratory by dissolving 2.7 grams of aluminum powder (99.9%) in 100 ml of boiled hydrochloric acid solution of 0.1n. the total normality of the reactant solution would be around 1 n (kohlschutter et al. 1942; alcoa world chemicals 2001). the hydrochloric acid solution and aluminum powder were placed in a two-neck 250 ml flask of as shown in fig. (1). a temperature thermometer was placed in one neck while the other neck was connected to a glass condenser in order to reduce loss of hcl through vaporization. the flask was heated via hot metal equipped with a temperature regulator in order to bring the reaction to a suitable operating heat level, and the temperature was then adjusted to 80 °c. the reaction time was considered to be the time at which the sample’s dark color disappeared, the solution become transparent, and all hcl vapor had been removed. the determined time was 5 hours. after the reaction, the sample was drawn and analyzed. the results showed that the specific gravity of the product was 1.078 and, therefore, the net weight of the sample was 97 g (7 grams of salt and 90 grams of water). the laboratory analysis of the produced pac is presented in table 1. 2.2 turbidity test turbidity level was considered the criterion index for water clarity. thus, conventional jar testing equipment was used for this purpose. before testing, samples of untreated raw water were collected from different sites along the tigres river and yielded different levels of turbidity such as 30, 150, 550 and 800 floriz33 preparation and usage of poly aluminum chloride as a coagulating agent ing turbidity units (ftu). each water sample was treated separately with different types of coagulating agents. after treatment, residual turbidity of clear water layer was tested at a chosen sedimentation time such as at 0, 10, 20 and 30 minutes. 2.3 ph test testing of the river water’s ph was undertaken before and after water treatment with all three types of coagulants, but there were no significant changes in the resultant ph. thus, the effect of the coagulating agents, especially the pac, on the ph of the treated water was tabulated. however, ph of the raw water was 8.4. 3. results and discussion this section presents the results of testing the performance of different types of coagulants on the clarity of river water samples at different levels of turbidity. the effects of increasing doses of coagulants (iraqi produced alum, swedish produced alum and pac) on water samples were determined and shown in figs. 2-5. these figures show the increased removal of water impurities with an increase in the dose of all types of agents. however, the pac produced the lowest water impurities. coupled with a low range of dosing, the pac produced better results than either of the alum coagulats. by considering a turbidity level of 16 ftu as an index for clear water, the required dosage concentrations of coagulating agents for achieving this level of turbidity were determined and presented in table 2. these data show that the dosage of pac required for raw water treatment was 98% less than alum at low water impurities (ftu=30), and required 90% less agent at higher levels of impurities (ftu=800). however, other researchers have stated that the pac required for water treatment is 60% less as compared to alum (halvadiya et al. 2008). during the experiment, it was noted that during the pac treatment flocs formed rapidly and the sludge produced was more compact than that of the alum. this could be due to the great ease of pac hydrolysis as compared to that of alum. pac emits polyhydroxides with long molecular chains and great electrical charges in the solution, thus maximizing the physical action of flocculation. the coagulation can then be carried out by neutralizing the negative charges on colloids by the ionic sites and then causing a decrease in zeta potential without changing the ph or alkalinity of water. also, it was noted that the alum concentration needed for water treatment at turbidity (i.e., 89 mg/liter at ftu=800) is much higher than that required for treatment at a lower turbidity (i.e., 16 mg/litre at ftu=30). this can be explained by the additional charge from colloidal suspended solids, and therefore, the high ionic heaviness of the from alum figure 1. laboratory apparatus for pac production table 1. laboratory analysis of the pac product 34 sk al-dawery and oh al-joubori coagulant was required to overcome these forces, enhance van der waals force and decrease zeta potential. however, lower dosages of pac were required for all turbidity levels. this could be due to the fact that pac has a higher hydrolysis characteristic than other alums, which may lead to the enhancement of adsorption of the present coordinated hydroxide group. simple hydroxide ions may bind strongly to many solid surfaces. alternatively, the replacement of an figure 3. relation between residual turbidity and coagulat conentration at a 150 ftu level concentration mg/l concentration mg/l 35 preparation and usage of poly aluminum chloride as a coagulant agent figure 4. relation between residual turbidity and coagulant concentration at 550 ftu level figure 5. relation betwen residual turbidity and coagulant concentration at 800 ftu level concentration mg/l 36 sk al-dawery and oh al-joubori aqua group by a hydroxyl group in the coordination sheath of a metal atom (as is present in alum) may render the complex more hydrophobic by reducing the interaction between the central metal atom and the remaining aqua groups, as can be observed in the following equations: this reduction in solvent-coagulant interaction might then, in turn, enhance the formation of covalent bonds between the metal atom and specific sites on the surface of the colloidal particle by reducing the energy necessary to displace water molecules from the coordination sheath. finally, adsorption becomes especially pronounced in the pac species because more than one hydroxide group per "molecule" can become attached at the interface (stumm, charles 1968). 4. conclusions the following conclusions can be drawn: based on the reaction between waste powder and hydrochloric acid, the laboratory-produced sample of pac performed well in river water treatment. as compared to alum, an application of pac yielded lower residual turbidity of treated water. experimental results showed that the dosage of pac required for river water treatment was 98% less as compared to alum at low turbidity, and required 90% less at higher levels of turbidity. operation treatment using pac gave excellent results as measured by rapid formation of flocs and compact sludge, and a shorter time for sedimentation. pac also, has a wide range of dosage, and it may be well-suited to a wide range of turbidity. thus, as compared to alum, it has better coagulation effects. references alcoa world chemicals (2001), polyaluminum chloride (pac). [http://www.alumina.alcoa.com]. al-dawery s, butris l (2002), quality improvement of the locally produced alum. iraqi j. of chemical and petroleum engineering 3:50-58. alhadidi a, kennedy m, diepeveen a, prummel h (2009), scaling potential calculations using different methods. desalination and water treatment 6:138-143. cleartech industries inc (2008), drinking water treatment. clear pac-180 1327-41. george s (1988), basic water treatment for application world-wide. 2nd ed, thomas telford, london 3751. halvadiya h, gangadharan pk, anand p (2008), deionization of coagulated, clarified, turbid gauri shankar lake waters by using ion-exchange technology. separation science and technology 43(8):2183-2195. kohlschutter hw, hantelmann p, diener k, schilling h (1942), basic aluminum chlorides. zeitschr anorg, und allgemeine chem. 248(4):319-344. mark jh (1975), water and waste-water technology. john wiley & sons inc. 270-455. stumm w, charles ro (1968), stoichiometry of coagulation. j. awwa 60:514-539. table 2. concentration of coagulats to produce water at ftu = 16 c:\users\raoh\desktop\paper_2.xps the journal of engineering research vol. 8 no. 1 (2011) 12-18 1. introduction ni-cr-mo alloyed cc steels are widely used for transmission gears, as it provides combination of soft ductile core, high fatigue strength and surface hardness, thereby wear resistance of contacting surfaces (wilks, et al. 1994). overly fluctuating ni and mo cost in the recent times, alarmed automotive industries worldwide, as it affect the final component cost. vehicle cost plays significant role, when automotive manufacturers compete locally and globally. hence, automotive industries demand cost effective steels without compromise in mechanical properties, machinability, performance and durability. melloy et al. (1973) found optimum combination of __________________________________________ *corresponding author’s e-mail: cengg@yahoo.com hardenability and toughness by adding 15 to 25 ppm of boron. irvine and pickering (1957) showed that, 20 ppm of boron results in similar tensile strength, achieved by alloying 1.0% cr and 0.7% mn. stumpf and banks (2006) found 10 to 30 ppm of boron enhances hardenability of steel through segregation at austenite grain boundaries and hence delays the nucleation of ferrite and pearlite. boron addition did not affect the hot and cold working properties of as rolled steels. it behaves as plain carbon steel but with higher hardenablity, reports szuch and delve (1967). gear tooth bending strength and loading capacity were ensured by impact strength of steel. lower impact strength results in gear tooth failure. impact strength of conventional low carbon boron steel (cbs) boron steel: an alternative for costlier nickel and molybdenum alloyed steel for transmission gears a. verma*a, k. gopinatha and s.b. sarkarb *amds, mechanical engineering department, indian institute of technology, madras, india (t.n.) 600 036 bmahindra ugine steel company ltd., jagdish nagar, khopoli (m.h.) india 410216 received 17 may 2009; accepted 4 october 2009 abstract: case carburized (cc) low carbon steels containing ni, cr and mo alloying elements are widely used for transmission gears in automobile, as it possesses desired mechanical properties. in order to cut cost and save scarce materials like ni and mo for strategic applications, steel alloyed with boron has been developed, which gives properties comparable to ni-cr-mo alloyed steel. in the process of steel development, care was taken to ensure precipitation of boron which results in precipitation hardening. the characterization of the developed boron steel had exhibited properties comparable to ni-cr-mo alloyed steel and superior to conventional boron steel. keywords: gear steel, boron steel, carburization distortion, impact strength, cost 13 the journal of engineering research vol. 8 no. 1 (2011) 12-18 (20mncr5b steel) was inferior to ni-cr-mo alloyed steel (en353 of bs970 standard). in cbs, titanium (ti) was added to control the nitrogen because of its higher chemical affinity towards nitrogen than boron. kapadia et al. (1968) reported that, ti forms tin and boron goes free in the solid solution, resulting in higher hardenability of steel. however, in absence of ti, boron forms precipitates of bn. hence, reduces effective boron, which was responsible for increasing hardenability of steel. boron hardenability potential was inversely proportional to the carbon content of the steel (rahrer and armstrong, 1947; irvine and pickering, 1963). rahrer and armstrong concluded that, al deoxidized steel forms aln with nitrogen, which was thermodynamically more stable than bn. but aln forms more slowly than bn, in austenite. considering the above facts, development of low carbon boron steel (dbs) and its merits over cbs and ni-crmo alloyed steel for potential gear applications were reported in this paper. the scope restricts to comparison of impact strength, tensile strength and case carburization distortion. 2. boron steel development the process followed to develop low carbon boron steel was eaf lf vd ccp (eaf-electric arc furnace, lfladle furnace, vd-vacuum degassing and ccpcontinuous casting process). the same process was followed for development of cbs and ni-cr-mo alloyed steel. the required charge was carefully prepared to control tramp elements such as cu, sn and p etc. eaf with ebt (eccentric bottom tilting) facility was used for steel melting, as it had superior slag control. the slag was tapped in a preheated ladle where molten steel was deoxidized with deoxidizers and fresh lime was added for making fresh slag. the melt was later transferred to the lf station for addition of ferroalloys in order to achieve the desired chemistry and temperature. a reducing slag was created as it aids in melt refining. the ladle was subsequently transferred to the vd station where melt degassing was carried out below 10-3 bar. vd with continuous argon purging through porous plug ensured removal of gases like hydrogen, oxygen and nitrogen. boron in the melt was added after vd. prior to boron addition, the melt was adequately deoxidized with al and other deoxidizers. the melt inclusions were controlled by proper deoxidation, slag physio-chemical characteristics, vacuum degassing, soft rinsing, argon shrouding and tundish metallurgy. the molten steel was cast in a 250 x 200 mm cast iron mould by continuous casting process (ccp) setup. the ccp setup had mould ems (electro magnetic stirrer), auto mould level control (amlc) device, well designed tundish and sub entry nozzle (sen) system. the rectangular billets were rolled to achieve reduction ratio of 1:6 or higher on the steel. the typical chemical composition of the ni-cr-mo alloyed steel, conventional low carbon boron steel (cbs) and developed low carbon boron steel (dbs) were shown in table 1. the compositions of three steels were measured by spectromax, ccd spectrometer from spectro ai gmbh. 3. results 3.1 metallurgical studies the test specimen for metallurgical studies were prepared from the forged and annealed (8800c for an hour) steel bars. grain sizes of steel were measured by comparison chart method as mentioned in the astm e112 standard. mixed grain sizes at 500x were observed for cbs in the leica make metallurgical microscope. grain sizes of 5 and finer (9 to 10) were observed for dbs under similar microscope at 500x. the grain size distribution for three steels was shown in fig. 1. grains of similar size and uniform distribution were observed for dbs where as cbs showed grains of mixed size and non-uniform distribution. the etching solution was prepared by mixing 100 ml of distilled water with four grams of picric acid and 0.5% of soap solution. the etching was performed at room temperature for over thirty minutes. the cleanliness study of dbs steel was conducted as per astm e 45 standard. the microscopic analyses of inclusion ratings were summarized in table 2. microstructure of boron steel shows lamellar pearlite and ferrite only. table 1. chemical compositions of cbs dbs and ni-cr-mo alloyed steel table 2. inclusion ratings of the developed boron steel as per astm e 45 standard 14 the journal of engineering research vol. 8 no. 1 (2011) 12-18 3.2 mechanical properties evaluation impact strength qualification of any steel is desired for consideration as gear steel. instrumented dynamic impact tests were carried out by brugger test method (tikhonov and palagin, 1994). brugger test specimens (fig. 2) were prepared from the forged bar of cbs, dbs and ni-cr-mo alloyed steel. the specimens were treated under identical carburizing, hardening, tempering parameters and cycle durations as used for existing ni-cr-mo alloyed steel gears. specimens from all three steels were carburized in the same batch, so any differences in mechanical properties due to heat treatment parameters and cycles were eliminated. impact tests were carried out on zwick / roell gmbh make, instrumented dynamic impact testing machine (model rkp 450) at room temperature. comparison of impact load results of cbs, dbs and nicr-mo alloyed steels were shown in fig. 3. impact load figure 1. microscopic image of grain size distrubution in boron steel at 500 x a) conventional boron steel (cbs), b) developed boron steel (dbs), c) ni-cr-mo alloyed steel 15 the journal of engineering research vol. 8 no. 1 (2011) 12-18 for dbs and ni-cr-mo alloyed steel were comparable (52 kn) where as cbs showed lower impact load (40kn). mechanical properties of steels (soft condition) were measured on zwick / roell gmbh make tensile testing machine (model z250) and tentative steel costs were compared and tabulated in table 3. load vs deflection curve obtained during tensile testing for all the steels were shown in fig. 4. dbs showed highest tensile strength of 789 mpa where as cbs showed maximum elongation of 11.73 %. tensile strength and elongation of ni-cr-mo alloyed steel were in between the two boron steels. 3.3 comparison of carburizing distortion gears from dbs and cbs were forged and machined to similar design and geometrical parameters (weight 5.6 kg, outer diameter (od) 180 mm and inner diameter (id) conforms to 76g5 mm of iso 286-1: 1988) as used for existing ni-cr-mo alloyed steel gears. in order to avoid geometrical / profile variation due to carburizing and tempering cycles, gears from three steels, were loaded to same batch in a furnace. twenty five gears from each of the three steels were checked for gear dimensions (id and id width) before and after carburization to evaluate distortion / growth pattern. mitutoyo, japan make gauges were used for measurement of gear dimensions. the carburizing cycle of duration eight hours, comprises of heating gears to temperature of 930 oc for carburization and hardened at 830 oc, followed by oil quenching at 110 oc for twenty minutes. the gears were then tempered at 150 oc for more than three hours. all the gears were metallurgically characterized and results were reported in table 4. the grinding / finishing stocks on gear id post carburizing were plotted in fig. 5. id distortion signifies the deviation of circularity measured at three locations over gear id width and average result reported. id distortion trend of dbs gears, were comparable (average 0.225 mm) to that of existing ni-cr-mo alloyed steel. the average distortion of cbs gears were 0.400 mm, higher compared to dbs and ni-cr-mo alloyed steel by about 0.175 mm. gears id width (measured along the gear axis) distortion trend for all three steels were plotted in fig. 6. dbs gears width growth trend were comparable to ni-cr-mo alloyed steel with average growth of 0.05mm. cbs gears width grows by average of 0.09 mm, higher by 0.04 mm than dbs and existing ni-cr-mo alloyed steel. gears id grinding stock trend of all three steels were more consistent than gears width distortion trend. all other gear dimensions (like lead, profile, crowning and span length) of dbs were comparable to ni-cr-mo alloyed steel and with in the specifications (k templates) when tested on klingelnberg, gmbh make lead and profile testing machine. 4. discussions higher impact strength of ni-cr-mo alloyed steel was attributed to the alloying of ni and mo. these elements results in the formation of low carbon martensite and thereby imparts sufficient strength and toughness to the steel (bepari and shorowordi, 2004). cbs had lower impact strength compared to dbs. in cbs, titanium fixes nitrogen by forming tin precipitates and free boron in solution segregated to austenite grain boundaries, reduces the cohesive force and thereby impact toughness (kapadia, et al. 1968; azarkevich, et al. 1995). however, in dbs titanium presence was in traces, boron reacts with nitrogen and carbon to form precipitates of boron nitride and boron carbide respectively. these precipitates removed from solution at the grain boundaries and thereby improve the impact properties. this is in line with the findings of irvine and pickering (1963). also, treppschuh et al. (1967) showed that, toughness of case hardened and cold worked steels can be improved by boron addition, provided boron should combine with nitrogen. tensile strength of dbs steel was 789 mpa, much higher compared to cbs steel (383 mpa) and better than existing nicr-mo alloyed steel (679 mpa). boron nitride and boron carbide precipitates at the grain boundaries resulted in precipitation hardening of dbs steel. these precipitates acts as barrier for movement of dislocations and thereby figure 2. brugger impact test specimen figure 3. impact test results of ni-cr-mo alloyed steel and boron steels by brugger method 16 the journal of engineering research vol. 8 no. 1 (2011) 12-18 imparts higher strength to the dbs. however, in cbs, boron segregation at the grain boundaries resulted in suppression of ferrite and pearlite formation (brownrigg, 1973). this results in lower strength of cbs steel. ferrite strengthening effects of ni and mo results in higher strength of the ni-cr-mo alloyed steel (davis, 2000). thermal stresses introduced during case carburization were responsible for dimensional changes or component distortion (thelning, 1974). these dimensional changes needs to be incorporated in machining tolerance and removed by grinding and finishing. the amount of retained austenite exhibits significant effects on dimentable 3. mechanical properties and cost comparison of cbs, dbs and ni-cr-mo alloyed steel figure 4. load deflection curve for ni-cr-mo alloyed steel and boron steels number of gears figure 5. grinding stock of gears made of ni-cr-mo alloyed and boron steels after cc 17 the journal of engineering research vol. 8 no. 1 (2011) 12-18 sional stability. after quenching, steels contain some retained austenite along with martensite which increases with amount of alloying elements dissolved during austenitization. during carburizing, the amount of carbon in the case increases with applied carbon potential, but its increase leads to higher austenite retention (bensely, 2008). higher the alloying elements results in higher retained austenite and thereby lower distortion (stumpf and banks, 2006; mohanty, 1995). in addition, presence of ni suppresses ms temperature which further enhances retained austenite content. hence, ni-cr-mo alloyed steel had lower id and width distortion. higher retained austenite reduces strength and residual stresses but subsequent tempering reduces retained austenite by transforming to martensite (parrish, 1999). in cbs, gear id and width distortion values were higher compared to ni-cr-mo alloyed steel. this was obviously due to lower retained austenite and high carbon martensite in cbs (boron has no effect on ms temperature). lower distortion in dbs may be due to controlled and uniform grain size distribution, in contrast to the mixed grain size distribution observed in cbs. 5. conclusions 1. dbs was comparable to the existing ni-cr-mo alloyed gear steel (en353 of bs 970 standard) and superior to cbs (20mncr5b). dbs specimens failed at an average impact load of 53.1 kn, comparable to 54 kn of ni-cr-mo alloyed steel and higher than 38.68 kn of cbs. 2. tensile strength of dbs was highest (789 mpa) where as elongation of cbs was highest (11.73%). tensile strength (679 mpa) and elongation (9.18%) of ni-cr-mo alloyed steels was in between the two table 4. metallurgical characterization of gears after cc and tempering number of gears figure 6. width distortion trend of gears made of ni-cr-mo alloyed and boron steels after cc 18 the journal of engineering research vol. 8 no. 1 (2011) 12-18 boron steels. 3. distortion of gears id and width due to carburization of dbs and ni-cr-mo alloyed steel gears were comparable to average of 0.225 mm and 0.05 mm respectively. however, distortion of id and width of cbs gears were higher to 0.400 mm and 0.09 mm respectively. 4. lower distortions of ni-cr-mo alloyed steel were associated with higher retained austenite compared to cbs. however, lower distortion of dbs gears could be due to controlled and uniform grain size distribution. 5. lower distortion in the dbs against cbs will save finished component cost in terms of tool repair or replacement period, thereby increasing the productivity. 6. with this huge potential of enhanced mechanical properties, lower raw material cost, reduced machining requirements and higher cost savings, dbs can be considered as alternative steel for transmission gears against costly ni-cr-mo alloyed steels and inferior cbs. acknowledgment the authors are grateful to mr. k. parthasarathy, retd. dy. general manager, product development, materials engineering department, ashok leyland limited chennai, india, for their support and technical help. the authors express sincere gratitude to the management of musco steel plant, india for their guidance and co-ordination. references azarkevich, a.a., kovalenko, l.v. and krasnopolskii, v.m., 1995, "the optimum content of boron in steel," j. of metal science and heat treatment, vol. 37(1-2), pp. 22-24. bensely, a., venkatesh, s., lal, d.m., nagarajan, g., rajadurai, a. and junik, k., 2008, "effect of cryogenic treatment on distribution of residual stress in case carburized en353 steel," materials science and engineering a, vol. 479, pp. 229-235. bepari, m.m.a. and shorowordi, k.m, 2004, "effects of molybdenum and nickel additions on the structure and properties of carburized and hardened low carbon steels," j. of materials processing technology, vol. 155-156, pp. 1972-1979. brownrigg, a., 1973, "boron in steela literature review 1956-1972," j. of australian institute of metals, vol. 18(3), pp. 124-136. davis, j.r., 2000, asm specialty handbook, asm international. irvine, k.j. and pickering, f.b., 1957, j. of iron and steel institute, vol.187, pp. 292-302. irvine, k.j. and pickering, f.b., 1963, j. of iron and steel institute, vol. 201, pp. 518-526. kapadia, b.m., brown, r.m. and murphy, w.j., 1968, "the influence of nitrogen, titanium and zirconium on the boron hardenability effect in constructional alloy steels," trans. of the metallurgical society of aime, vol. 242, pp. 1689-1694. melloy, g.f., slimmon, p.r. and podgursky, p.p., 1973, "optimizing the boron effect," j. of metallurgical and materials transaction b, vol. 4(10), pp. 22792289. mohanty, o.n., 1995, "on the stabilization of retained austenite: mechanism and kinetics," materials science and engineering b, vol. 32, pp. 267-278. parrish, g., 1999, "carburizing: microstructure and properties," asm international. rahrer, g.d. and armstrong, c.d., 1947, "the effect of carbon content on the hardenability of boron steels," trans. of the asm, vol. 40, pp. 1099-1112. stumpf, w. and banks, k., 2006, "the hot working characteristics of a boron bearing and a conventional low carbon steel," materials science and engineering a, vol. 418, pp. 86-94. szuch, r.l. and delve, f.d., 1967, "hardenability of boron treated carbon steels," j. of blast furnace and steel plant, pp. 930-936. thelning, k.e., 1974, "steel and its heat treatment," bofors handbook, butterworths, london, pp. 466 471. tikhonov, a.k. and palagin, y.m., 1994, "method of testing gear wheels in impact bending, technical information," metal science and heat treatment, vol. 36(12), pp. 655 -657. treppschuh, h., randak, a., domalski, h.h. and kurzeja, j., 1967, “influence of boron in the properties of structure steels and tool steel,” vol. 87, pp. 1355-1366. wilks, t.p., cavallaro, g.p., subramanian, c., strafford, k.n., french, p. and allison, j. e., 1994, "conditions prevailing in the carburizing process and their effect on the fatigue properties of carburized gears," j. of materials processing technology, vol. 40, pp. 111125. civl060808.qxd the journal of engineering research vol. 5, no.1, (2008) 79-96 ________________________________ corresponding author e-mail:alnuaimi@squ.edu.om comparison between experimental and 3d finite element analysis of reinforced and partially pre-stressed concrete solid beams subjected to combined load of bending, torsion and shear a. s. alnuaimi department of civil and architectural engineering, college of engineering, po box 33, pc 123, al-khoud, muscat, oman received 8 august 2006; accepted 8 april 2007 abstract: this paper presents a non-linear analysis of three reinforced and two partially prestressed concrete solid beams based on a 20 node isoparametric element using an in-house 3d finite element program. a non linear elastic isotropic model, proposed by kotsovos, was used to model concrete behaviour, while steel was modelled as an embedded element exhibiting elastic-perfectly plastic response. allowance was made for shear retention and for tension stiffening in concrete after cracking. only in a fixed direction, smeared cracking modelling was adopted. the beams dimensions were 300x300 mm cross section, 3800 mm length and were subjected to combined bending, torsion and shear. experimental results were compared with the non-linear predictions. the comparison was judged by load displacement relationship, steel strain, angle of twist, failure load, crack pattern and mode of failure. good agreement was observed between the predicted ultimate load and the experimentally measured loads. it was concluded that the present program can confidently be used to predict the behaviour and failure load of reinforced and partially prestressed concrete solid beams subjected to a combined load of bending, torsion and shear. keywords: beam, solid beam, bending, shear, torsion, direct design, concrete, reinforced concrete, stress analysis, combined loading á«fõl á«fé°sôn õfgƒlh áë∏°ùÿÿg áfé°sôÿg øe ¬àª°üe õfgƒ÷ io~ùg ô°uéæ©dg ωg~îà°séh »zgôødg π«∏ëàdgh á«∏ª©ÿg èfféàædg úh áfqé≤e ¢ü≤dg h agƒàd’gh aéæëf’g øe á©ªà› ∫éªmc’ â©° ng oé¡l’g »ª«©ædg »∏y áá°°uuóóÿÿggá≤jôw ω~îà°ùj »hƒ°sém èeéfôh ωg~îà°séh é«fõl oé¡l’g á≤hé°s áfé°sôÿg øe øjõfélh áë∏°ùÿg áfé°sôÿg øe ¬àª°üe õfgƒl ákóãd é«£n ’ ó«∏– ábqƒdg √òg ¢vô©j : π«ã“ ” éªæ«h ,¢sƒaƒ°ùjƒc ¤g iõ©j »£n ’ ¿ôe êpƒ‰ ωg~îà°séh èeéfèdg ‘ áfé°sôÿg ∑ƒ∏°s πãe ~bh .á£≤f øjô°ûy gp oé©h’g »kók gô°üæy πª©à°ùj …òdg io~ùg ô°uéæ©dg éªc .≥≤°ûàdg ~©h ~°ûdg áehé≤e á«°uénh ¢ü≤dg áehé≤e aé≤hg ‘ áfé°sôÿg á«°uén ¿éñ°ù◊g ‘ òng ~bh .¢üdén »wé£eh ¿ôe ∑ƒ∏°s ägp áfƒa~e ô°uéæ©c í«∏°ùàdg ~j~m ∑ƒ∏°s ¤g â©° ngh ∏e3800 õfé÷g ∫ƒwh ∏e 300*300 õféé∏d »°vô©dg ™£≤ÿg oé©hg âféc ~bh .§≤a âhék √éœg ‘ ≥°ûdg aé≤h ¢vgîag ™e ¬àà°ûÿg ¥ƒ≤°ûdg êpƒ‰ ω~îà°sg qg~≤eh ,πª◊gh ámgre’g úh ábó©dg øª° àj áfqé≤ÿg ô°uéæy âféch .á«£nódg π«déëàdg èféàæh á«∏ª©ÿg èféàædg áfqé≤e â“ ~bh .¢ü≤dgh agƒàd’gh aéæëf’g øe á©ªà› ∫éªmg øe é¡«∏y π°üm »àdg á«fé¡ædg ∫éªm’g úh √~«l âféc á≤agƒÿg ¿cg ßmƒd ~bh .qé«¡f’g ´ƒfh ≥≤°ûàdg •é‰gh ,qé«¡fód áññ°ùÿg ∫éªm’gh ,πàødg ájhgrh ,~j~◊g ‘ √ƒ°ûàdg õfgƒé∏d qé«¡f’g ∫éªmgh ∑ƒ∏°ùh dƒñæà∏d á≤k πµh áeg~îà°sg øµá π«∏ëàdg ‘ ω~îà°ùÿg üƒ°sé◊g èeéfôh ¿g ∂dp øe èàæà°ùjh .πª©ÿg ‘ â°ù«b »àdg ∫éªm’gh »£nódg π«∏ëàdg .¢ü≤dgh agƒàd’gh aéæëf’g øe á©ªà› ∫éªm’ á°vô©ÿgh oé¡l’g á«fõl á«fé°sôÿg õfgƒ÷gh áë∏°ùÿg á«fé°sôÿg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.oé¡lc’g ,áë∏°ùÿg áfé°sôÿg ,¢üb ,agƒàdeg ,aéæëfeg ,ô°téñÿg º«ª°üàdg á≤jôw ,áરüÿg ¿gõ«÷g ,¿gõ«÷g : notation md, td, vd = design bending moment, torsion and shear force respectively le/lc = experimental to computed failure load ratio l.f. = load factor (percentage of applied load to design load ) l.f.= (t/td+m/md)/2 exp = experimentally measured values fcu = concrete cube compressive strength 80 the journal of engineering research vol. 5, no.1, (2008) 79-96 1. introduction the behaviour of solid beams when subjected to combined loading is very complex. a detailed analysis would normally require a three-dimensional finite element model. preston and austin (1992) cited that "the design and analysis of rc beams is so complex that, in general, it is impossible for engineers to consider all aspects of a problem at once". ojha, et al. (1974) have studied the behaviour of reinforced concrete rectangular beams under combined torsion, bending and shear and found a sharp reduction in the torsional stiffness compared to flexural stiffness after cracking of the beam. the reduction depends on a number of factors such as the loading combination, the strength and distribution of the steel reinforcement and the form of the cross section. tests conducted by thurlimann (1979), revealed that the torsional strength of beams relies on the outer concrete shell of about 1/6 the diameter of the largest circle inscribed into the perimeter connecting the corner longitudinal bars. mitchell and collins (1974), described the torsional shear stress as circulating in the periphery of the section. its intensity is distributed in a parabolic shape with the maximum stress at the outside fibre and zero at some distance from the surface. rahal and collins (1995) developed a three-dimensional analytical model capable of analysing rectangular sections subjected to combined loading of biaxial bending, biaxial shear, torsion and axial load. the model takes into account the shear-torsion interaction and concrete spalling. it idealizes the rectangular cross-section resisting shear and torsion as made of four transversely reinforced walls with a varying thickness and varying angle of a principal compressive strains. the vertical shear stress due to the shear force is uniformly resisted by the vertical walls and the lateral shear stress is resisted by the horizontal walls. they tested their model and concluded that the model predicts very close results to experimental behavioural and ultimate load results. rahal (2000) developed an equation relating the ultimate torsional moment and ultimate shearing stress in the walls of the equivalent tube. the walls of the section resisting the shear stresses were idealized as a reinforced concrete membrane element subjected to pure in-plane shearing stress. macgregor and ghoneim (1995) stated that "in a solid section, the shear stresses due to direct shear are assumed to be distributed uniformly across the width of the section, while the torsional shears only exist in the walls of the assumed thin-walled tube. for this reason, the direct summation of the two terms tends to be conservative and a root square summation is used". they proposed design equations for torsional resistance in which the outer skin alone contributes to the torsional resistance of a solid beam. they claimed that the thickness of the wall resisting torsion in a solid member is on the order of onesixth to one-quarter of the minimum width of a rectangular member. ibell, et al. (1998) used an upper-bound plasticity analysis in a 2d model for the assessment of shear f’c = concrete cylinder compressive strength f t ' = concrete tensile strength fy = yield stress of the longitudinal steel fyv = yield stress of the transverse steel fpu = ultimate strength of the pre -stressing wire tshr = shear stress in concrete due to shear force ttor = shear stress due to torsion εy = longitudinal steel yield strain ( εy = fy/e) εpy = prestressed wires yield strai n (εpy = fpy/e) εyv = transverse steel yield strain ( εyv = fyv/e) g = the elastic shear modulus of the un -cracked concrete β = shear retention factor åcr = cracking strain ( ' t cr c f å e = ) ån = average of the three principal strains at any cracked point ( å å å ån + + = 1 2 3 3 ) èi = angle of twist at a section èi = (df+dr)/lh (see fig. 9) ø = rate of twist ø = ( è2 – è1)/a 81 the journal of engineering research vol. 5, no.1, (2008) 79-96 in reinforced concrete beams. the results from this model were compared with experimental results. it was recommended that the 2d analysis to be extended to more general 3d collapse analysis. rabczuk and eibl (2004), presented a model using a mesh free 2d galerkin/ finite element approach. the concrete was modelled with particles and reinforcement with beam elements. for steel, an elastoplastic constitutive law with isotropic hardening and tension cut-off was used. the concrete was modelled via a continuum damage model, where an anisotropic tensile damage variable was used to capture the behaviour of concrete in tension. they used a displacement controlled approach for testing their model. they compared the results from their model with experimental results from one rectangular and one i-section pre-stressed beams. they concluded that a full 3d simulation would be more appropriate. in this research a in-house 3-d finite element program was used for non-linear analysis of this computational study. the program was developed by el-nuonu (1985) using kotsovs' concrete model. this model was based on experimental data obtained at imperial college london from tests on the behaviour of concrete under complex stress states (kotsovos and newman, 1979; and kotsovos, 1979). the testing techniques used to obtain this data were validated by comparing them with those obtained in an international co-operative programme of research into the effect of different test methods on the behaviour of concrete. this model is capable of describing the behaviour of concrete under uniaxial, biaxial and triaxial stress conditions. it requires only the concrete cube compressive strength fcu to define the behaviour of concrete under different stress states. more information about this model is given in (kotsovos and pavlovic, 1995). 2. research significance predictions from an in-house 3d fe program are compared with experimental results of solid reinforced and partially pre-stressed concrete beams subjected to combined load of bending, shear and torsion. the beams were designed using the direct design approach with plastic stress field for the reinforced beams and elastic stress field for the partially pre-stressed beams. 3. a 3d finite element program a standard incremental-iterative procedure was adopted for solution. the load increments were equal to 10% of the design load for the first two increments and 5% for the remaining increments. the maximum number of increments was 50 and the maximum number of iterations in each increment was 200. the convergence being deemed satisfactory if the ratio of the square roots of the sum of the squares of the residual forces to that of the applied loads did not exceed 5%. the stresses in the cross-section nearest to mid-span were analysed. the stress distribution at the last converged increment was used for the analysis. in deciding on the predicted mode of failure, the load-deflection relationship, steel strain and ultimate load were taken into consideration. the program was extensively used by bhatt and lim (1999)(a,b) for the analysis of slabs, internal columnflat slab junctions and punching shear failure of flat slabs lim and bhatt (1998). good agreement between predicted and experimental results was found. before cracking or crushing, the concrete behaviour is assumed to be non -linear elastic isotropic. crushing occurs at a point when all the three principal stresses are compressive and th e state of stress is on the ‘yield’ surface. in the case of concrete crushing, complet e loss of strength is assumed ie. no compression softening is allowed for. when the concrete cracks in any direction, concrete ceases to be isotropic and crushing can occur if the minimum principal strain (compressive) reaches a value taken as equal to 0.003. after cracking, smeared crack approach with simple tension stiffening and shear retention equations are employed to represent the post cracking behaviour of concrete ( fig. 1). cracks are assumed to be orthogonal and once formed remain in their direction. the stress -strain relationship in tension was assumed to be linear up to 'tf and immediately after cracking the tensile stress tf is reduced to 0.8 ' tf . thereafter, tf decreases linearly with strain and is zero at the maximum strain of 0.003 which roughly corresponds to yield strain of steel of 0.0025. transfer of shear stresses across cracks is mode lled by means of the 'shear retention' factor β which defines the shear modulus of cracked concrete as âg, where g is the elastic shear modulus of the un -cracked concrete. the shear retention factor â = 1.0 if εn ≤ åcr and â = 0.25 åcr/ ån if ån > åcr, where εcr = cracking strain ( å ' t cr c f e = ) and ån = average of the three principal strains at any cracked point ( 1 2 3 å å å å 3n + + = ). the reinforcement is modelled as one dimensional element embedded in the solid concr ete elements. elast ic-plastic stress -strain behaviour without strain hardening was used in this research. only uniaxial resistance is considered with no provision for kinking or dowel action of bars. in the cells were the pre -stressing wires are present th e prestressing wires are assumed to act as unstressed steel with a yield stress equal to the difference between the yield stress fpy and the stress at service fpe (fig. 2). in the 3d program, a 150x150x150 mm isoparametric solid element with twenty node and twenty seven gauss points was used. the concrete cylinder compressive strength 'cf is taken as f’c= 0.8fcu n/mm 2, the young’s modulus ec = 5000 ' cf n/mm 2, the split cylinder tensile strength ' 't cf . f= 054 n/mm 2 and the poisson’s ratio was set at a constant value of 0.15. 82 the journal of engineering research vol. 5, no.1, (2008) 79-96 4. tested beams three reinforced and two partially pre-stressed concrete beams were analysed. the reinforced beams were tested by alnuaimi and bhatt (2006) while the partially prestressed beams were tested by alnuaimi (2007). all beams were 300x300 mm cross section and 3.8 m length. they were subjected to a combined load of bending, torsion and shear (table 1). the main variables studied were the ratio of the shear stress due to torsion to shear stress due to shear force τtor /τshr which varied between 0.69 and 3.04 and the ratio of the torsion to bending moment td /md which varied between 0.26 and 1.19. the concrete mix consisted of cement, uncrushed 10mm gravel and sand with water/ cement ratio of 0.55. three cubes, 100x100x100 mm, and six cylinders, 150x300 mm, for each beam were cast from the same concrete used for casting each beam. the specimen and the samples were kept under damp hessian for about four days and then under room condition. the samples were tested on the day the beam was tested to determine the cube and cylinder compressive strengths and split cylinder tensile strength of concrete. the pre-stressing wires were tensioned using a simple arrangement of two nuts with ball bearings such that the wire could be stressed with a pair of spanners by tightening the nuts. the force in the wires was measured using a simple load cell developed for this purpose. table 2 shows the average yield strengths of reinforcement and compressive and tensile strengths of concrete. the concrete cube and cylinder compressive strengths shown for each beam in table 2 are the measured average strengths of the three cubes and three cylinders respectively and the concrete tensile strength shown is the measured average strength of three cylinders tested for split test. all results were obtained from samples cured along side each beam. the concrete cube compressive strengths used were ranging between 37 n/mm2 and 61 n/mm2 (mostly normal strength concrete). only high yield deformed bars, 8, 10 and 12 mm diameter for longitudinal and 8 mm diameter for transverse were used as reinforcement. the longitudinal steel yield strength fy for each beam is the measured average of the average of three samples of each bar type and the transverse yield fyv strength is the average measured strength of three samples. for pre-stressing only 5mm diameter wires with yield stress fpy of 1570 n/mm2 were used. the calculated reinforcement was used in the test span (middle 1200 mm). between the test span and the beam ends, more longitudinal and transverse steel was used to resist negative moment at the supports and to ensure failure occurred in the test span. figure 3 shows typical arrangement of reinforcement and fig. 4 shows the provided reinforcement and arrangement of longitudinal bars for each beam. the solid circles in fig. 4 represent the longitudinal bars or pre-stressing wires in which strain was measured nearest to mid-span. strains in the stirrups nearest to mid-span on the front and rear faces were also reported. t e n sio n stiffe n in g 0 0 .2 0 .4 0 .6 0 .8 1 0 0 .0 0 2 s t r a in f t/ f' t s h e a r re te n tio n εε nε c r β m in 1 .0 figure 1. tension stiffening and shear retention curves fpy fpe fpy fpe st re ss strain figure 2. transfer and service stresses in a pre-stressing wire β 83 the journal of engineering research vol. 5, no.1, (2008) 79-96 5. test setup and instrumentation figure 5 shows typical load and support arrangement and fig. 6 shows a testing rig with typical beam installed. the test rig is a three-dimensional frame designed to allow application of torsion, bending moment and shear force. the model was mounted on two steel stools fixed to the concrete floor at a distance of 1.8 m a part. the test span was 1.2 m long centred at mid-span. the beam was simply supported by a set of two perpendicular rollers at each support and a system of pin-and-roller at the mid-span of the top face. at the support, the lower roller allows axial displacement and the upper one allows rotation about a horizontal axis at the soffit level of the beam. the diameter of each roller was 100 mm and the length was 300 mm. the rollers were separated by 300x300x20 mm steel plates and a similar plate was put between the upper roller and the soffit of the beam. at the mid-span of the top face, a 300x100x30 mm steel plate was placed using cement plaster and the pin-and-roller system was installed between the steel plate and the load cell. the pin prevented rigid body motion and the roller allowed rotation about horizontal beam axis. torsion was applied by means of a torsion arm fixed to each end of the beam (fig. 7) while bending moment and shear force were a result of applied load at mid-span across the beam width at the top face. td md vd ôtor ôshr td/md ôtor / ôshr beam no. knm knm kn n/mm2 n/mm2 ratio ratio reinforced beams btv13 26 50.89 61.08 4.16 3.00 0.51 1.39 btv14 13 50.89 61.08 2.08 3.00 0.26 0.69 btv15 39 32.89 41.08 6.24 2.05 1.19 3.04 partially pre -stressed beams btv16 13 50.89 61.08 2.08 3.00 0.26 0.69 btv17 39 32.89 41.08 6.24 2.05 1.19 3.04 table 1. load combination fcu f'c f’t fy fyv fpy beam no. n/mm2 n/mm2 n/mm2 n/mm2 n/mm2 n/mm2 reinforced beams btv13 40 28.5 3.45 500 500 btv14 37 25.7 2.92 500 500 btv15 61 38.2 4.38 500 500 partially pre -stressed beams btv16 52 36 3.42 500 500 1570 btv17 53 36 3.44 500 500 1570 table 2. average material properties see x-sect ion 2y8 @ 60 mm c/c y8 @ 60 mm c/c 1300 mm 1200 mm 1300 mm 3800 mm 2 t est span 1 figure 3. typical arrangement of reinforcement see x-section 2 of fig. 4 84 the journal of engineering research vol. 5 no.1, (2008) 79-100 this support and loading arrangement allowed full rotation (no torsion is resisted by the rollers) about the centre line of the beam soffit and displacement in the beam axial direction. it produced constant torsional shear stress over the entire length of the beam and maximum normal stress due to bending and shear stress due to shear force occurred near the mid-span. torsion, bending and shear loads were applied using hydraulic pumps and load cells were used to measure the load at each jack location. the load cells were connected to a data logger for data acquisition. linear voltage displacement transducers (lvdt) were used to measure the vertical displacement at the bottom face of the beam fig. 8(a) while for the measurement of twist, three transducers were located on the centreline of the front and rear faces as shown in figs. 8(b-c). rotation at any of the vertical sections was obtained by dividing the vertical difference in displacement between directly opposite transducers by the distance between these points as shown in fig. 9(a). using the notations in this figure the angle of twist is equal to (dr+df)/lh. in the tested beams, the relative twist is the difference between the angle of twist θ2 caused by the displacements in transducers 6 and 9 (fig. 8) and the angle of twist θ1 caused by the displacements in transducers 4 and 7. the rate of twist ψ = (θ2 θ1)/α, where α is the distance between the two sections (fig. 9(b). section 1 -1 4y8 2y10 4y12 btv14 2y8 2y10 4y12 y8@170 mm c/c section 2 -2 btv15 4y8 2y8 2y10 4y12 section 1 -1 2y8 2y8 2y10 4y12 y8@82 mm c/c section 2 -2 section 1 -1 4y8 2y8 2y10 4y12 2y8 2y8 2y10 4y12 btv13 y8@120 mm c/c section 2 -2 btv16 2y8 2y8 2y8 2y8 +2wires y8@170 mm c/c section 2 -2 2wires 2y8 +2wires section1 -1 4y8 2y8 2y8 2wires btv17 2y8 2y8 2y8 2y10 +2wires y8@80 mm c/c section 2 -2 2wires 2y10 +2wires section1 -1 4y8 2y8 2y8 2wires 2y8 figure 4. provided reinforcement in the test span and outside the test span 85 the journal of engineering research vol. 5, no.1, (2008) 79-96 pin and roller 1000mm 900mm 900mm 1000mm 1200mm test span perpendicular roller s perpendicular rollers torsional arm torsional arm figure 5. typical load and support arrangement figure 6. test rig with a typical beam installation figure 7. torsional arm while testing 86 the journal of engineering research vol. 5, no.1, (2008) 79-96 to measure strain in the bars, a pair of strain gauges, 6mm long, was fixed on directly opposite faces of the bar and connected to a data logger. accordingly, the axial strain recorded at each load stage was taken as the average reading of both gauges. crack width and crack development were measured by means of a crack width measuring microscope. for each experiment, the design load was divided into load increments. the value of each of the first three increments was 10% of the design load while for each of the rest increments it was 5% of the design load until failure. the point load applied at the mid-span and that applied at the trosional arms were exerted simultaneously until failure. the first step was to zero all load cells and record instrument readings with minimum possible loads on the model. the beam was considered to have collapsed when it could resist no more loads. this usually happened after a major crack spiralled around the beam cross-section near the mid-span dividing the beam into two parts connected by the longitudinal reinforcement. 6. comparison between the experimental and computational results in order to obtain a finite element solution for the tested beams, the 3-d program was used for the non-linear n.a mid span 2 test span 1200mm 1 3 (a) bottom face n.a 4 5 6 n.a 7 8 9 1100mm (b) front face (c) rear face figure 8. locations of lvdt df dr lh a = 1100mm è1 è2 (a): angle of twist at a section èi = (df+dr)/lh (b): rate of twist ø = ( è2 – è1)/a figure 9. deformation of beam section due to torsion 87 the journal of engineering research vol. 5, no.1, (2008) 79-96 analysis. measured values of a concrete cube compressive strength and steel tensile strength were used. the concrete cylinder compressive strength, young's modulus, poisson's ratio, concrete tensile strength, shear retention and tension stiffening were used as explained in section 3. 6.1 load displacement relationship figure 10 shows the vertical measured and computed displacements at the centre of the bottom face of each beam. it is clear from this figure that a good agreement was achieved between experimental and computational results for beam btv14. in the case of beams btv13 and btv16 the program predicted stiffer behaviour than the measured. the measured values from beams btv15 and btv17 were small and difficult to compare with the predicted values. with exception to the computed displacement of btv16, measured and computed displacements of all beams with bending dominant (td /md<1) (btv13, btv14 and btv16) reached the displacement limit of span/250. the beams with torsion dominant (td /md>1) (btv15 and btv17) experienced measured and computed relatively smaller displacements and did not reach the span/250 limit. 6.2 strain in the longitudinal steel figure 11 shows a good agreement between the experimental and computational results in the reinforced concrete beams for the longitudinal steel in the front face. figure 12 shows an acceptable agreement between the measured and predicted strains in the pre-stressing wires of beams btv16 and btv17. in the beams with td / md<1, the longitudinal steel or prestressing wires in the front face yielded or reached yield strain while slightly less strain was recorded in the ones with td /md>1. figure 13 shows a very good agreement in the strain ratios between the measured and predicted values of the reinforced beams in the rear face. figure 14 shows good agreement in the case of rear face partially pre-stressed beams. in most cases, measured and predicted results show longitudinal bars or pre-stressing wires in the front faces, yielded or reached near yield strain. in the rear face, only the prestressing wires in btv16 yielded while in the rest of the beams strain was below the yield strain. 6.3 strain in transverse steel figure 15 shows strain ratios in the front face transverse steel; in general, a very good agreement between experimental and computational results was achieved. with exception to btv16, the transverse steel in the front face yielded or reached near yield strain. figure 16 shows strain ratios in the rear face transverse steel; a good agreement was achieved between the measured and predicted values in the case of beams btv13, btv14 and btv17. in the case of btv15, the program predicted larger values of strain for the same load than the measured ones. it should be noted that both measured and predicted values of strain in the rear face were negligible and the steel did not reach yield strain. this due to the fact that for practical reasons, that stirrup areas required for the front face where shear stresses are additives were used in the rear face subtractive as well. so, the beams are over reinforced in the transverse direction of the rear face. 6.4 relative angle of twist figure 17 shows the experimental and computational relative angles of twist. these angles were calculated as explained in section 5. it can be seen that a very good agreement has been achieved between experimental and computational results in most cases. however, in beam btv17 the observed angle was larger than the computed one. in beams where torsion was dominant (td /md>1) the relative angle of twist was relatively larger than when the bending was dominant (td /md<1). for the same load combination, the partially prestressed beam btv16 experienced a smaller twist than the reinforced beam btv14. however, the difference in the case of beams btv15 and btv17 was negligible. 6.5 failure load column 4 of table 3 shows a very good agreement between the measured le and computed lc failure loads. all beams failed near their design loads. 6.6 crack pattern and mode of failure both the computed and measured results showed that in the case of beams in which bending was dominant (td /md<1) almost vertical cracks started in the bottom face and at the lower half of the front and rear sides. these cracks were followed by inclined cracks in succeeding load increments until they first appear and in the top flange at about 80% of a failure load. in the beams where torsion was dominant (td /md>1), inclined cracks extended into the bottom face one increment after they were formed in the front and rear sides. in both groups, the smaller the ratio td /md, the closer is the angle of crack to vertical. figures 18-20 show typical computed and observed crack development of the tested beams. the computed and observed cracks shown are taken at the load increment just before failure. in beams btv13, btv14 and btv16 the mode of failure was mostly flexural where the beam experienced relatively large displacement and the flexural steel yielded. a small number of large cracks caused failure at the time of flexural steel yielding. beams btv15 and btv17 failed by diagonal cracking due to high torsional shear stress and the failure mode was less ductile with small displacement, less longitudinal steel strain and larger transverse steel strain than the bending dominant beams. it is clear that a good agreement between the observed and computed cracks was achieved on the intensity and direction of cracks. 88 the journal of engineering research vol. 5, no.1, (2008) 79-96 btv13 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0.0 2.0 4.0 6.0 8.0 disp.(m m ) l.f . exp comput btv14 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.0 2.0 4.0 6.0 8.0 disp.(m m ) l.f . exp comput. btv15 -0.2 0 0.2 0.4 0.6 0.8 1 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 disp.(m m ) l.f . exp comput btv16 0 0.2 0.4 0.6 0.8 1 1.2 1.4 -2 0 2 4 6 8 10 12 14 16 18 20 disp.(m m ) l.f . comput exp btv17 0 0.2 0.4 0.6 0.8 1 1.2 -1 0 1 2 3 disp.(m m ) l.f . comput exp figure 10. vertical displacement at mid-span 89 the journal of engineering research vol. 5, no.1, (2008) 79-96 front face longitudinal steel btv13 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 ◊◊ y l. f. exp comput front face longitudinal steel btv14 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 1.2 ◊◊ y l. f. exp comput front face longitudinal steel btv15 -0.2 0 0.2 0.4 0.6 0.8 1 -0.2 0 0.2 0.4 0.6 0.8 1 ◊◊ y l.f . exp comput figure 11. strain ratios in the front face longitudinal steel front face wire btv16 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.4 0.6 0.8 1 1.2 1.4 ηηψψ⎤⎤y l. f. comput exp front face wire btv17 0 0.2 0.4 0.6 0.8 1 0.6 0.8 1 1.2 ηηψψ⎤⎤ y l. f. comput exp figure 12. strain ratios in the front face pre-stressing wires ε/εy ε/εy ε/εy ε/εy ε/εy 90 the journal of engineering research vol. 5, no.1, (2008) 79-96 rear face longitudinal steel btv13 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 ηηψψ⎤⎤y l. f. comput exp rear face longitudinal steel btv14 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 ηηψψ⎤⎤y l.f . exp comput rear face longitudinal steel btv15 -0.2 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 ηηψψ⎤⎤y l. f. exp comput figure 13. strain ratios in the rear face longitudinal bars rear face wire btv16 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.4 0.6 0.8 1 1.2 1.4 ηηψψ⎤⎤ y l. f. comput exp rear face wire btv17 0 0.2 0.4 0.6 0.8 1 0.4 0.6 0.8 1 ηηψψ⎤⎤y l. f. comput exp figure 14. strain ratios in the rear face pre-stressing wires ε/εy ε/εy ε/εy ε/εy ε/εy 91 the journal of engineering research vol. 5, no.1, (2008) 79-96 front face stirrup btv13 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0 0.1 0.2 0.3 0.4 y l. f. exp comput front face stirrup btv14 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 -0.1 0 0.1 0.2 0.3 0.4 0.5 y l. f. exp comput front face stirrups btv15 -0.2 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 y l.f . exp comput front face stirrups btv16 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 -0.1 0 0.1 0.2 0.3 0.4 y l.f . comput exp front face stirrups btv17 -0.2 0 0.2 0.4 0.6 0.8 1 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 y l.f . comput exp figure 15. strain ratios in the front face stirrups ε/εy ε/εy ε/εy ε/εy ε/εy 92 the journal of engineering research vol. 5, no.1, (2008) 79-96 rear face stirrup btv13 0 0.2 0.4 0.6 0.8 1 1.2 -0.1 0 0.1 0.2 0.3 ≅≅&& y l.f . exp comput rear face stirrups btv14 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 -0.05 0 0.05 0.1 ≅≅&& y l. f. exp comput rear face stirrups btv15 0 0.2 0.4 0.6 0.8 1 -0.1 0 0.1 0.2 0.3 0.4 0.5 ≅≅&& y l. f. exp comput rear face stirrups btv16 0 0.2 0.4 0.6 0.8 1 1.2 1.4 -0.05 0 0.05 0.1 0.15 0.2 0.25 ≅≅ y l. f. comput exp rear face stirrups btv17 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 ≅≅ y l. f. comput exp figure 16. strain ratios in the rear face stirrups ε/εy ε/εy ε/εy ε/εy ε/εy 93 the journal of engineering research vol. 5, no.1, (2008) 79-96 relative angle of twist btv13 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 5 10 15 20 25 30 35 40 45 angle(rad)x10 -3 l.f . exp comput relative angle of twist btv14 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0 3 6 9 12 15 18 angle(rad)x10 -3 l.f . exp comput relative angle of twist btv15 0.0 0.2 0.4 0.6 0.8 1.0 0 10 20 30 40 50 angle(rad)x10 -3 l.f . exp comput relative angle of twist btv16 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0 2 4 6 8 10 12 angle(rad)x10 -3 l.f . exp comput relative angle of twist btv17 0.0 0.2 0.4 0.6 0.8 1.0 0 10 20 30 40 50 angle(rad)x10 -3 l.f . exp comput figure 17. relative angle of twist 94 the journal of engineering research vol. 5, no.1, (2008) 79-96 7. conclusions a comparison was conducted between experimental results and an in-house 3d finite element analysis of three reinforced and two partially prestressed concrete solid beams. the beams were designed using the direct design method and subjected to different load combinations of bending, shear and torsion. the comparison was judged by load displacement relationship, steel strain, angle of twist, failure load, crack pattern and mode of failure. very good agreements were obtained in the cases of steel strain ratios, angle of twist, failure load and mode of failure. acceptable agreement was obtained for displacement values. overall, it can be concluded that the predictions from the 3d finite element program was shown to be in a good agreement with the experimental results and therefore, 1 2 3 4 td/md ôtor/ôshr le/lc beam no. ratio ratio ratio reinforced beams btv13 0.51 1.39 0.98 btv14 0.26 0.69 0.96 btv15 1.19 3.04 0.93 partially pre -stressed beams btv16 0.26 0.69 1.00 btv17 1.19 3.04 1.02 mean 0.98 table 3. ratios of measured and predicted failure loads figure 18. computed and observed cracks near failure in the front face (btv14) figure 19. computed and observed cracks near failure in the front face (btv15) 95 the journal of engineering research vol. 5, no.1, (2008) 79-96 proven to be a good tool for the prediction of beam behaviour and ultimate load of solid reinforced and partially prestressed concrete beams subjected to combined load of bending, torsion and shear. large differences in load combinations (td/md 0.26 1.19 and τtor/ τshr 0.69 3.04) did not result in large discrepancies between measured and predicted results. references alnuaimi a.s. and bhatt, p., 2006, "design of reinforced concrete solid beams, structures and buildings journal," thomas telford limited, vol. 159(4), pp. 197-216. alnuaimi, a.s., 2007, "direct design of partially prestressed concrete solid beams structural engineering and mechanics,” techno-press ltd, vol. 27(6), pp. 741-771. bhatt, p. and lim, b.t., 1999a, "flat slab-column junctions with shear and moment transfer: a comparison between finite element predictions and experiments," proc. of 7th acme conference, university of durham, (ed. bettes, p.) pp. 11 14. bhatt, p. and lim, b.t., 1999b, "punching shear capacity of internal column-flat slab junction with in-plane restraint: a comparison between finite element predictions and experiments. developments in analysis and design using finite element methods," civil-comp press, (ed. b.h.v. topping), pp. 141 147. el-nuonu, g. f. r., 1985, "design of shear wall-floor slab connections. ph.d. thesis, university of glasgow. ibell, t.j., morley, c.t. and middleton, c.r., 1998, "an upper-bound plastic analysis for shear," magazine of concrete research, vol. 50(1), pp. 67-73. kotsovos, m.d. and newman, j. b., 1979, "a mathematical description of the deformation behaviour of concrete under complex loading," magazine of concrete research, vol. 31(107), pp. 77-90. kotsovos, m.d. and pavlovic, m.n., 1995, structural concrete, finite element analysis for limit-state design," thomas telford publications, 1 heron quey, london e14 4jd. kotsovos, m.d., 1979, "a mathematical description of the strength properties of concrete under generalized stress," magazine of concrete research, vol. 31(108), pp. 151-158. lim, b.t. and bhatt, p., 1998, "punching shear failure of flat slabs: a comparison between finite element predictions and experiments," advances in civil and structural engineering computing for practice, (ed. b. topping), civil comp-press, pp. 163 173. macgregor, j.g. and ghoneim, m.g., 1995, "design for torsion. aci structural journal," vol. 92-s20, pp. 211-218. mitchell, d. and collins, m.p., 1974, "behaviour of structural concrete beams in pure torsion," publication no. 74-06, department of civil engineering, university of toronto, toronto, ontario, canada. ojha surendra, k., 1974, "deformation of reinforced concrete rectangular beams under combined torsion, bending and shear.," aci journal, vol. 7126, pp. 383-391. preston, j.i. and austin, m.a., 1992, "solid modelling of rc beams:2, computational environment," journal of computing in civil engineering, vol. 6(4), pp. 404-416. rabczuk, t. and eibl, j., 2004, "numerical analysis of prestressed concrete beams using a coupled element free galerkin/finite approach," int. j. of solid and structures, vol. 41(3-4), pp 1061-1080. rahal, k.n., 2000, "torsional strength of reinforced concrete beams," canadian journal of civil engineering, vol. 27, pp. 445 453. figure 20. computed and observed cracks near failure in the front face (btv16) 96 the journal of engineering research vol. 5 no.1, (2008) 79-96 rahal, k.n. and collins, m.p., 1995, "analysis of sections subjected to combined shear and torsion ," a theoretical model. aci structural journal, vol. 92(4), pp. 459-469. thurlimann, b., 1979, "torsional strength of reinforced and prestressed concrete beams-ceb approach," institut fur baustatik und konstruktion, eth. zurich., vol. 92, pp 117-143. also can be found in: american concrete institute, detroit, sp59. civl070618.qxd the journal of engineering research vol. 6, no.1, (2009) 37-45 1. introduction reclaimed asphalt pavement (rap) is the result of removing old asphalt pavement material. rap consists of high quality well-graded aggregate coated with aged asphalt cement. the removal of asphalt concrete is done for reconstruction purposes, resurfacing, or to obtain access to buried utilities. the disposal of rap represents a large loss of valuable source of high quality aggregate. supplies of natural high quality aggregate are depleting in some areas in the world, or can be costly to transport to the construction site. existing portland cement concrete and asphalt concrete pavements provide a source of high quality aggregate that can be recycled. recycling can contribute to the waste disposal and to the conservation of natural resources (yrjanson, 1989 and kenai, et al. 2002). kenai, et al. 2002, conducted a study on the use of recycled concrete and bricks as an aggregate in concrete. ________________________________________ *corresponding author’s e-mail: aloraimi@squ.edu.om the study used either fine aggregate replacement or coarse aggregate replacement or both. percentages of replacement were 25, 50, 75, and 100% of the aggregate. the study recommended limiting the amount of recycled aggregate to 75% and 50% for the coarse and fine aggregate, respectively. a reduction in compressive strength was reported with the increase in recycled aggregate replacement. the study found that the relationships between tensile and compressive strength for natural aggregate concrete can be used for the recycled aggregate concrete. murshed, et al. 1997, investigated the use of combinations of coarse and fine rap aggregate in normal concrete mixes and compared the results of compressive strength to conventional mixes with 0.4 and 0.5 water cement ratios. compressive strength values were found to decrease with the increase in rap content. the study concluded that the concrete mixes containing rap can qualify for concrete applications such as sidewalks, driveways, curbs, and gutters. recycling of reclaimed asphalt pavement in portland cement concrete salim al-oraimi*, hossam f. hassan and abdulwahid hago department of civil and architectural engineering, p.o. box 33, al-khoud, p.code 123, sultan qaboos university, muscat, sultanate of oman received 18 june 2007; accepted 14 september 2007 abstract: reclaimed asphalt pavement (rap) is the result of removing old asphalt pavement material. rap consists of high quality well-graded aggregate coated with asphalt cement. the removal of asphalt concrete is done for reconstruction purposes, resurfacing, or to obtain access to buried utilities. the disposal of rap represents a large loss of valuable source of high quality aggregate. this research investigates the properties of concrete utilizing recycled reclaimed asphalt pavement (rap). two control mixes with normal aggregate were designed with water cement ratios of 0.45 and 0.5. the control mixes resulted in compressive strengths of 50 and 33 mpa after 28 days of curing. the coarse fraction of rap was used to replace the coarse aggregate with 25, 50, 75, and 100% for both mixtures. in addition to the control mix (0%), the mixes containing rap were evaluated for slump, compressive strength, flexural strength, and modulus of elasticity. durability was evaluated using surface absorption test. keywords: reclaimed asphalt pavement, concrete, compressive strength, elastic moduli á«à檰s’g áfé°sôÿg ‘ ™lî°ùÿÿg »à∏ø°s’g ∞°uôdg ôjh~j ioéyeg ƒég ~mgƒdg~ñy ,ø°ùm ωé°ùm ,»áô©dg ⁄é°s ::áá°°uuóóÿÿgg≥agôª∏d ∫ƒ°uƒ∏d hcg í£°ùdg á«£¨j ioéye’ hcg aé°ûfe’g ioéyeg ¢vgôzc’ ádgre’g á«∏ªy ºàjh .¥ô£dg øe â∏ø°sc’g äé≤ñw ádgreg á«∏ªy øe ™lî°ùÿg »à∏ø°se’g ∞°uôdg èàæj ωécôdg qoé°üe øe q~°üÿ gq~g ogƒÿg √òg øe ¢ü∏îàdg πãáh ,â∏ø°sc’éh ≈£¨e ioƒ÷g ¤éy ωécq ≈∏y ™lî°ùÿg »à∏ø°se’g ∞°uôdg …ƒàëj ée éñdézh .¥ô£dg â– áfƒa~ÿg á«e~ÿg á«fé°sôÿg äé£∏ÿg øe úàyƒª› º«ª°üàh ∂dph ,™lî°ùÿg »à∏ø°se’g ∞°uôdg ωécq øe áyƒæ°üÿg áfé°sôÿg ¢üfé°ün á°sgqo åëñdg gòg ∫héæàj ^ioƒ÷g ‹é©dg¤g aéÿg áñ°ùf ¿ƒµj %50 h %25 h , áféÿg ‘ ôø°u âféc é¡æe πc ‘ …oé©dg ωécôdg¤g ™lî°ùÿg »à∏ø°s’g ∞°uôdg ωécq áñ°ùf ¿g éªc ,%50 á«féãdg áyƒªûg ‘h %45 é¡æe ¤h’g ‘ â檰s’gg 50 éeƒj 28 ~©h é¡h •é¨¶f’g oé¡leg ≠∏h ~b (áféÿg ‘ ôø°u) ™lî°ùÿg »à∏ø°s’g ∞°uôdg ωécq ≈∏y …ƒà– ’ »àdg äé£∏ÿg ¿cg ~lh , äé£∏ÿg qéñàneéhh . %100 h %75 h ,áfhôÿg πeé©eh , aéæëf’g oé¡lgh , •é¨¶f’g oé¡lgh , •ƒñ¡dg qéñàng â檰 j ô°û©dg äé£∏ÿg ≈∏y ägqéñàng i~y âjôlg ~bh . á«féã∏d ∫éµ°séh éé«e 33h ¤hód π°ùcéh éé«e .»ë£°ùdg ¢ué°üàe’g ᣰùjƒh äé£∏ÿg √òg áeƒáo º««≤j ” éªc áá««mmééààøøÿÿgg ääggooôôøøÿÿgggg.áfhôÿg πeé©e ,•é¨° fe’g iƒb ,áfé°sôÿg ,™lî°ùÿg »à∏ø°s’cg ∞°uôdg : 38 the journal of engineering research vol. 6, no.1, (2009) 37-45 limbachiya, et al. 2000, used recycled concrete as an aggregate in high strength concrete. results indicated that up to 30% of recycled concrete aggregate had no effect on strength. at higher percentages, there was a gradual reduction in strength. the study presented a method to adjust the water cement ratio to overcome this reduction in strength. the study concluded that the high strength concrete made with recycled concrete aggregate can have equivalent engineering and durability performance to normal high strength concrete. jankovic, 2002, in his study, compared the effect of polymer admixture with a percentage of 0, 4, and 8 % on concrete made with combinations of recycled brick and river sand. the study concluded that there is no effect of polymer on compressive and flexural strength. however, the polymer provided some improvement in water resistance and frost resistance. the study recommended using the concrete made from recycled blocks in thermal insulators, and in bearing walls for buildings. this paper presents the results of a study conducted on the evaluation of using rap in concrete mixes. two mix grades are designed. coarse aggregate is replaced with the coarse fraction of the rap aggregate with various percentages including 0, 25, 50, and 75%. mix properties including: slump, compressive and flexural strength, elastic modulus and surface absorption are presented. 2. experimental program aggregates used in the concrete mix consisted of 20 mm coarse aggregate (ca), fine aggregate (fa) and recycled asphalt concrete pavement (rap). as a result of the cold milling operation, the rap is in the form of loose particles coated with aged asphalt cement. rap was separated by sieving on the 5 mm sieve size into coarse and fine rap. normal portland cement type i was used. the aggregate and rap gradation are shown in fig. 1. the physical properties of aggregate and rap are shown in table 1. two normal portland cement concrete control mixes (with no rap aggregate) were designed with ratios of 1: 1.9 : 2.9 : 0.5 and 1 : 1.7 : 2.5 : 0.45 for cement to fine aggregate to coarse aggregate to water. the cube compressive strength after 28 days of water curing resulted in 33 and 50 mpa for the two mixes, respectively. the mixes were referred to as mix 30 and mix 50. the coarse aggregate was selected to be replaced with aggregate coarse agg. fine agg. coarse rap fine rap bulk sg 2.78 2.57 2.35 2.40 bulk sg (ssd) 2.81 2.65 2.40 2.45 apparent sg 2.84 2.78 2.5 2.5 absorption (%) 1.8 1.5 1.8 1.6 la abrasion (%) 19.5 26.4 table 1. aggregate and rap physical properties 0 20 40 60 80 100 0.0 0.1 1.0 10.0 100.0 sieve size (mm) pe rc en ta ge p as si ng (% ) . rap 20 mm coarse agg. fine agg. coarse rap fine rap figure 1. grain size distribution for aggregate and rap 39 the journal of engineering research vol. 6, no.1, (2009) 37-45 coarse rap aggregate as it constitutes a higher percentage in the mix. the percentages of replacement were 0 (control), 25, 50, 75, and 100 %, by weight of the coarse aggregate. table 2 shows the mix quantities for the two mixes. the aggregate weights are based on saturated surface dry (ssd) condition. the fresh concrete mixes were tested for slump (astm c143-98) and unit weight (astm c138). twelve 100 mm cubes, three 150 mm cubes, three 150 by 300 mm cylinders, and three 100 by 100 by 500 mm prisms were cast for each mix. all specimens were subjected to water curing. the 100 mm cube specimens were tested for compressive strength according to british standards (bs) (bs 1881-116) after 7, 14, 28, and 90 days of curing. the cylinders were tested for both modulus of elasticity (astm c469-94) and compressive strength (astm c873) after 28 days of curing. the prisms were tested for flexural strength (astm c78) after 28 days of curing. the 150 mm cubes were used to evaluate the durability of the mixes by the surface absorption test (bs 1881-208) after 56 days of curing. 3. results and discussion 3.1. fresh concrete properties table 3 shows the slump and unit weight for the two mixes for different percentages of rap replacement. the table indicates a reduction in the slump value from 163 to 20 mm for mix 30 and from 55 to 5 mm for mix 50 with the increase in the percentage of rap replacement from 0 to 100%. in general, the unit weight shows the same trend for both mixes as it decreases with the increase in percentage of rap content. 3.2. compressive strength figure 2 shows the results of the cube compressive strength (fcu) test after 7, 14, 28, and 90 days of curing for the different percentages of rap replacement for mix 30. the figure indicates the expected gain in strength with age. the figure also shows the reduction in strength with the increase in rap content. figure 3 shows the cube compressive strength (fcu) results for mix 50. the figure also shows the gain in strength with curing and the reduction in strength with the addition of rap for all mixes. quantity (kg/m 3) mix rap % cement fine agg. coarse agg. coarse rap water 30 0 380.0 730.0 1100.0 0.0 190.0 25 380.0 730.0 825.0 275.0 190.0 50 380.0 730.0 550.0 550.0 190.0 75 380.0 730.0 275.0 825.0 190.0 100 380.0 730.0 0.0 1100.0 190.0 50 0 425.0 714.3 1070.0 0.0 191.4 25 425.0 714.3 802.9 267.1 191.4 50 425.0 714.3 535.7 534.3 191.4 75 425.0 714.3 267.1 802.9 191.4 100 425.0 714.3 0.0 1070.0 191.4 table 2. mix quantities rap percentage mix parameter 0 25 50 75 100 30 slump, mm 163 95 90 85 20 unit weight, kg/m 3 2458 2405 2392 2357 2323 50 slump, mm 55 43 20 12 5 unit weight, kg/m 3 2442 2458 2435 2389 2377 table 3. slump and unit weight for different rap content mixes 40 the journal of engineering research vol. 6, no.1, (2009) 37-45 for the 28 days compressive strength, the reduction in strength is indicated in fig. 4 for both mixes. the figure indicates approximately 10% higher reduction in strength for mix 50 compared with mix 30. at 100% rap replacement, the reduction is approximately 58 % for both mixes. figures 5 and 6 show the development of cube compressive strength at different curing periods for both mixes and different rap percentage. the figures show the ratio of the compressive strength at different curing periods to the compressive strength at 28 days of curing (development ratio). the development of strength was generally the same for both mixes. the results are similar to the reported typical values for the gain in strength for normal concrete (mehta and monteiro, 1993; neville, 1987). the compressive strength for cylinders (fcyl) after 28 days of curing as well as the ratio of fcyl to fcu are shown in table 4. the results indicate the reduction in strength with the increase in rap content, which is consistent with the decrease in the case of cube specimens. the ratio of fcyl to fcu ranged from 0.77 to 0.89 for all specimens. 3.3. flexural strength table 5 shows the flexural strength (modulus of rupture) (fr) results for the prisms after 28 days of curing. predicted values based on the aci code equations (mehta and monterio, 1993) and the cylinder compressive strength (fcyl) are also shown, in addition to the ratio of (fr/fcyl). a general trend of reduction in strength with the increase in rap content can be seen. the modulus of rup 10 15 20 25 30 35 40 45 50 55 60 0 25 50 75 100 percent of rap replacement c om pr es si ve s tre ng th (m pa ) . 7 days 14 days 28 days 90 days figure 2. compressive strength for mix 30 10 15 20 25 30 35 40 45 50 55 60 0 25 50 75 100 percent of rap replacement c om pr es si ve s tre ng th (m pa ) . 7 days 14 days 28 days 90 days figure 3. compressive strength for mix 50 41 the journal of engineering research vol. 6, no.1, (2009) 37-45 0 10 20 30 40 50 60 70 0 20 40 60 80 100 120 percent of rap replacement c om p. s tre ng th r ed uc tio n (% ) . mix 30 mix 50 figure 4. percentage reduction in compressive strength 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 0 14 28 42 56 70 84 98 curing period (days) st re ng th d ev el op m en t ( f cu/ f cu 28 ) . 0% rap 25% rap 50% rap 75% rap 100% rap figure 5. strength development ratio for cube compressive strength (mix 30) 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 0 14 28 42 56 70 84 98 curing period (days) st re ng th d ev el op m en t ( f cu/ f cu 28 ) . 0% rap 25% rap 50% rap 75% rap 100% rap figure 6. strength development ratiofor cube compressive strength (mix 50) 42 the journal of engineering research vol. 6, no.1, (2009) 37-45 ture decreased from 4.0 to 2.7 mpa for an increase in rap replacement of 100% for mix 30, about 33% reduction in strength. for the higher strength mix (mix 50), the modulus of rupture decreased from 5.5 to 3.9 mpa for the 100% rap replacement, which amounts to 29% reduction. the modulus of rupture results obtained from the laboratory tests are shown to be in agreement with the range given by the aci equations. moreover, the ratio of (fr/fcyl) for both mixes and for the different percentages of rap replacement agrees with typical reported values for normal concrete (mehta and monterio, 1993). 3.4. modulus of elasticity the modulus of elasticity was determined according to astm c469-94 on the cylinder specimens before crushing them. the results are shown in figs. 7 and 8 for mix 30 and 50, respectively. for comparison, the aci building code 318-83 gives the following expression for the static modulus of for normal weight concrete (neville, 1987). (1) where, ec = the modulus of elasticity in gpa and fcyl = the 28 days cylinder strength in mpa. the british standards for the structural use of concrete bs 8110: part 2: 1985 tabulates typical values of the static modulus of elasticity based on the 28 days cube strength. an expression is proposed by neville, 1987 based on the bs standards as follows: (2) where, ec = the modulus of elasticity in gpa and fcu = the 28 days cube strength in mpa. rap percentage mix parameter 0 25 50 75 100 30 fcyl a 29.4 23.8 20.9 15.9 12.4 fcyl/fcu b 0.89 0.81 0.80 0.81 0.87 50 fcyl a 39.5 30.3 24.0 19.8 16.9 fcyl/fcu b 0.79 0.79 0.77 0.83 0.81 table 4. cylinder compressive strength afcyl = cylinder compressive strength in mpa, bfcyl/fcu = ratio of cube t o cylinder compressive strength rap percentage mix modulus of rupture (fr), mpa 0 25 50 75 100 30 laboratory 4.0 4.30 3.3 3.1 2.7 aci code a 3.6 3.2 3.0 2.6 2.3 aci code b 5.4 4.9 4.6 4.0 3.5 aci code c 3.4 3.0 2.8 2.5 2.2 fr/fcyl 12 15 13 16 19 50 laboratory 5.5 4.5 3.8 4.5 3.9 aci code a 4.1 3.6 3.2 2.9 2.7 aci code b 6.3 5.5 4.9 4.4 4.1 aci code c 3.9 3.4 3.0 2.8 2.5 fr/fcyl 11 12 12 19 19 table 5. prism flexural strength (modulus of rupture) alower range = cylf66.0 , bupper range = cylf0.1 , and crecommended value = cylf62.0 . c cyle . f=470 . c cue . f= 0 339 1 43 the journal of engineering research vol. 6, no.1, (2009) 37-45 both expressions 1 and 2 were used as shown in figs. 7 and 8. the results indicate a decrease in the modulus as rap percentage is increased. the results also indicate that the obtained results fall between the values predicted from both equations up to 50% rap replacement. for higher percentages of rap, the modulus is lower than that given by both equations. a regression analysis was performed on the ten mixes to obtain equations similar to eqs. 1 and 2, the resulting equations were as follows: (3) (4) where, ec, fcyl and fcu are as defined before; and r2 and r2adj are the coefficient and adjusted coefficient of determination, respectively. 3.5. durability the initial surface absorption test (bs 1881-208) was performed as an indicator for the durability of the mixes. the test gives the water flow (in ml/m2/sec) into the surface of a dry cube specimen subjected to a head of 200 mm. water is allowed to penetrate the surface for periods of 10, 30, 60 and 120 minutes. at the end of each period, flow measurements were made. figure 9 shows the results for mix 30. the figure indicates a reduction in surface absorption with time. the same observations apply to fig. 10 (mix 50). the flow at 120 minutes for mix 30 was in the range of 0.057 to 0.093 ml/m2/sec. lower flow values were obtained for mix 50 with values in the range of 0.043 to 0.063 ml/m2/sec. the results did not indicate a significant difference in the absorption with the increase in rap content for both mixes. however, a lower flow was obtained for the stronger mix (mix 50) which should be anticipated. 5 10 15 20 25 30 35 40 0 25 50 75 100 percentage of rap ec (g pa ) . laboratory aci bs figure 7. modulus of elasticity for mix 30 5 10 15 20 25 30 35 40 0 25 50 75 100 percentage of rap ec (g pa ) . laboratory aci bs figure 8. modulus of elasticity for mix 50 . c cyle . f= 11065 (r2 = 0.85, r 2adj = 0.87) . c cue . f= 1050 61 (r2 = 0.85, r 2adj = 0.87) 44 the journal of engineering research vol. 6, no.1, (2009) 37-45 4. conclusions reclaimed asphalt pavement was used as a coarse aggregate substitute in two different normal concrete mixes having 28 days cube compressive strengths of 33 and 50 mpa. rap was used with 25, 50, 75, 100% replacement of coarse aggregate. the slump decreased with the increase in rap content. the compressive and flexural strength decreased as well with the increase in rap content. the general trend of strength development, as well as the relations between flexural strength, elastic modulus and compressive strength for the rap mixes agreed well with that for normal concrete. the surface absorption was not significantly affected by the addition of rap. the results indicated the viability of rap as an aggregate in non-structural concrete applications. the percentage of rap should be limited according to the application. low slump should also be considered when utilizing rap in the mixes. references astm c138-2000, standard test method for unit weight, yield, and air content (gravimetric) of concrete. astm c143-98, standard test method for slump of 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450 0.500 0 10 20 30 40 50 60 70 80 90 100 110 120 130 time, min. fl ow , m l/m 2 / se c. 0% 25% 50% 75% 100% figure 9. initial surface absorption for mix 30 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450 0.500 0 10 20 30 40 50 60 70 80 90 100 110 120 130 time, min. fl ow , m l/m 2 / se c. 0% 25% 50% 75% 100% figure 10. initial surface absorption for mix 50 45 the journal of engineering research vol. 6, no.1, (2009) 37-45 hydraulic-cement concrete. astm c469-94, standard test method for static modulus of elasticity and poisson's ratio of concrete in compression. astm c78-94, standard test method for flexural strength of concrete (using simple beam with third point loading). astm c873-99, standard test method for compressive strength of concrete cylinders cast in place in cylindrical molds. bs 1881: part 166: 1983, method for determination of compressive strength of concrete cubes. bs 1881: part 208: 1996, recommendations for the determination of the initial surface absorption of concrete. jankovic, k., 2002, "using recycled brick as concrete aggregate," proceedings of the international conference: sustainable concrete construction, dundee, uk, pp. 232-240. kenai, s., debieb, f. and azzouz, l., 2002, mechanical properties and durability of concrete made with coarse and fine recycled aggregates," proceedings of the international conference: sustainable concrete construction, dundee, uk, pp. 383-392. limbachiya, m.c., leelawat, t. and dhir, r.k. 2000, "use of recycled concrete aggregate in highstrength concrete," materials and structures/materieux et constructions, vol. 33, pp. 574-580. mehta, p.k., monteiro, p.j.m., 1993, concrete: microstructure, properties, and materials, mcgraw hill, new york. murshed, d., fahmy, m. and taha, r., 1997, "use of reclaimed asphalt pavement as an aggregate in portland cement concrete," aci materials journal, vol. 94(3), pp. 251-256. neville, a.m. and brooks, j.j., 1987, "concrete technology," longman scientific & technical, uk. yrjanson, w.a., 1989, "recycling of portland cement concrete pavements, synthesis of highway research practice 154," national cooperative highway research program, transportation research board, national research council, washington, pp. 3-36. c:\users\raoh\desktop\paper_8.xps the journal of engineering research vol. 8 no. 2 (2011) 66-73 1. introduction friction occurs in all mechanical systems, eg. bearings, transmission, hydraulic and pneumatic cylinders, valves, brakes and wheels. in many engineering applications, frictional contact occurs between machine parts and the characterization of contact behavior becomes an important subject in solving tribological problems such as friction induced vibration wear. issues related to mechanical sealing, performance and life of machine elements, and thermal are few examples. the pioneering work of greenwood and williamson (1966) has been utilized by many researchers (ibrahim and rivin 1994; mulhearn and samuels 1962; abdo and farhang 2005, seabra and berthe 1987) as a basic for further extension to obtain contact models for general or specific contact problems for mainly elastic contact. on the other hand, the work of pullan and williamson (1972) utilized as a basic model for pure plastic contact. in an attempt to bridge the gap between the pure elastic and pure _______________________________________ *corresponding author’s e-mail: jdabdo@squ.edu.om plastic contact, researchers halling and nuri (1972), greenwood and tripp (1971), and abdo and alyahmadi (2004) developed a wide intermediate range of interest where elastic-plastic contact triumph. in many cases the classical friction model cannot capture the characteristics such as stribeck effect, stick-slip, pre-sliding hysteretic motion, break-away force, which play a significant role in application on high precision positioning control (abdo et al 2010, abdo and al-yahmadi 2009). the role of friction modeling can be categorized according to whether or not the friction compensation is model-based. examples of non-model-based compensators include high-gain feedback, and impulsive control. this paper contains four sections including the introduction. section 2 reviews the single state elastoplastic friction model given by dupont et al. (2002). section 3 presents an adaptive compensator to achieve velocity tracking in the presence of the friction force described by the single state elastoplastic friction model and a proof of it is also given in this section. adaptive compensator of single state elastoplastic friction model a. a. abouelsouda, j. abdo*b and r. zaierb adepartment of electronics and communications engineering, faculty of engineering, cairo university, egypt *bdepartment of mechanical and industerial engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, al-khoud, muscat, oman received 19 april 2011; accepted 25 september 2011 abstract: a nonlinear friction is an unavoidable phenomenon frequently experienced in mechanical system between two contact surfaces. an adaptive compensator is designed to achieve tracking of a desired velocity trajectory in the presence of friction force described by a single state elastoplastic friction model. the adaptive compensator includes an adaptive observer and a computed force controller. the closed loop system is shown to be stable using lyapunov second method. simulation results show the effectiveness of the proposed compensator. keywords: adaptive compensator, elastoplastic friction model, lyapunov second method 67 a. a. abouelsoud, j. abdo and r. zaier since the single state elastoplastic friction model contains a bounded function to describe different friction phases and depends on the immeasurable elastic state, the proposed compensator uses an adaptive observer and a computed force controller. stability analysis of the proposed adaptive compensator is carried out using lyapunov second method. simulation results are given in section 4 and conclusion is drawn in section 5. 2. single state elastoplastic friction model in this section we review the single state elastoplastic friction model given by dupont et al. (2002), based on which we design an adaptive velocity compensator to follow a desired velocity trajectory. the rigid body displacement x is composed of elastic (z) and plastic (w) components as, x = z + w (1) friction models define the elastic dynamics explicitly, while the plastic displacement w is defined implicitly. following dupont et al. (2002) model, the friction force is given by: (2) (3) where zss is defined as (4) and fss (x) is the steady state friction force also called the stribeck function which is shown in fig. 1. the stribeck function fss (x) is bounded from below and above as (5) and (6) (7) figure 2 shows a typical shape of m (z, m) as given in (dupont et al. 2002). 3. adaptive compensator in this section an adaptive velocity compensator is designed to achieve tracking of a desired velocity trajectory vd (t) in the presence of friction force that was described by eqs. (2 and 3). a similar approach was proposed by canudas et al. (1995) but with a simple friction model and constant m (z, x). consider the equation of motion of a single mass m subject to driving force u and friction force fl (8) where v is the velocity of the mass (v = x). figure 1. stribeck curve of steady state friction force fss (x) versus rigid body velocity 68 adaptive compensator of single state elastoplastic friction model (9) and in matrix form eq. (9) can be given as (10) the only available signal for measurement is the velocity v of the rigid body, hence, expressing y = v in terms of the state x leads to (11) where (12) definition 1 positive real (pr) function a strictly proper rational function g(s) is positive real (pr) if g(s) is analytic for re [s] 0 and re [g(j ) > 0 for al < . definition 2 strictly positive real (pr) function a strictly proper rational function g(s) is strictly positive real (spr) if g (s) is positive real for some real > 0. the kaman-yakubovich (ky) lemma (narendera and annaswamy 1989): a strictly proper rational function g(s) with state space realization (a,b,c) g(s) = c (si a)-1 is strictly positive real (spr) if there exists positive definite symmetric matrices p=pt>0 and q=qt>0, such that pa + at p = -q (13) 3.1 adaptive observer the adaptive observer is given as (14) figure 3. adaptive compensator adaptation law state observer controller 1/ms friction model 69 a. a. abouelsoud, j. abdo and r. zaier l = (l1 l2)t is the observer gain and l1 and l2 are chosen to make the transfer function go(s) = c (si a + lc)-1 b strictly positive real (spr). since the pair (c, a) is observable (seabra and berthe 1987) for o 0, we can place the eigenvalues of the matrix (alc) at any desired position. we notice also that the transfer function g(s) = c(si a)-1 b has a zero at (o / 1 ) (transfer function equal to zero), hence by placing the eignevalues of the matrix a-lc at the location of the zero (o / 1 ) the remaining eigenvalue is in the left half plane and thus we obtain a strictly positive real function (spr). a zero of a transfer function is a real or complex number. depending on kaman-yakubovich (ky) lemma there exist a positive definite symmetric matrices p = pt > 0 and q = qt > 0, such that p (a lc) + (a = lc)t p = -q (15) 3.2 controller the controller u is chosen as (16) (17) > 0 is an adaptation gain. 3.3 stability analysis the close loop system given by eqs. (9), (14),(16) and (17) is analyzed in this section utilizing the lyapunov second method. consider hence (16) from the selected gain l discussed in section 3.1 to make the transfer function c(si a + lc)-1 b spr there exist a positive definite symmetric p and q which satisfies eq. (13). substituting for the control u into (3.1) we obtain (17) consider the lyapunov function (18) time derivative of w along-with the solution of eqs. (16 and 17) is given as (19) where (20) where, 70 adaptive compensator of single state elastoplastic friction model 71 a. a. abouelsoud, j. abdo and r. zaier 0 0.5 1 1.5 2 2.5 3 3.5 4 v vd t sec 1.5 1 0.5 0 -0.5 -1 figure 6. rigid body velocity v and desired square wave velocity vd 0 0.5 1 1.5 2 2.5 3 3.5 4 time sec alphaestimate alpha 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -0.1 72 adaptive compensator of single state elastoplastic friction model this leads to (21) (22) or (23) to regulate the velocity v to follow vd we assume |v| < vm . by completing the squares the following inequalities are reached 4. simulation results the desired velocity function is chosen as vd (t) = 10 sin (10t) this is a standard test signal. the elastoplastic friction model parameters are given in table 1. the stribeck friction function is given by: fss = [ fm + (fm-fm)e-(1000v)2 ] sign(v) (23) zba = 0.7169 zss (24) the gains of the observer and controller are given in table 2. it is obvious from figures 4-7 that the proposed adaptive compensator provides an accurate tracking capability to the desired trajectory while keeping all other signals bounded. conclusions in this paper a single state elastoplastic friction model is reviewed and an adaptive compensator is proposed to track a desired velocity trajectory in the presence of friction force model. the adaptive compensator includes an adaptive observer which consists of a state estimator and an adaptation law to one of the unknown functions of the friction model. the controller cancels the nonlinearities in the friction model using the estimated state and parameter. the model stability is determined by using lyapunov analysis. simulation of the proposed adaptive compensator to track both sinusoidal and square wave signals show almost perfect tracking and signals are bounded in the closed loop system which indicate the proposed compensator is effective. references abdo j, al-yhmadi a (2009), the effect of controlled frequency and amplitude of vibration on friction. solid state phenomena 147:380-385. abdo j, al-yahmadi a (2004), a wear model for rough surfaces based on the ultimate-stress asperity concept. int. j. of applied mechanics and engineering 9:11-19. abdo j, farhang k (2005), elastic-plastic contact model for rough surfaces based on plastic asperity concept. int. j. of non-linear mechanics 40/4: 495-506. abdo j, tahat m, abualsoud a, danish m (2010), the effect of frequency of vibration and humidity on the stick-slip amplitude. int. j. of mechanics and materials in design 6(1):45-51. canudas c de wit, olsson h, astrom kj, lischinsky p (1995), a new model for control of system with table 1. system parameters table 2. compensator gains 73 a. a. abouelsoud, j. abdo and r. zaier friction. ieee transaction in automatic control 40(3):419-425. dupont p, hayward v, armstrong b, altpeter f (2002), single state elastoplastic friction models. ieee transaction in automatic control 47(5):783-787. greenwood ja, williamson jbp (1966), contact of nominally flat surfaces. proc. r. soc. 295:300. greenwood ja, tripp jh (1971), the contact of two rough nominally flat rough surfaces. proc. instn. mech. engrs. 185:625-633. halling j, nuri ka (1972), contact of rough surfaces of working-hardening materials. in the mechanics of contact between deformable bodies, eds. de pater and kalker. delft: university press 412. ibrahim ra, rivin e (1994), friction-induced vibration, part i: mechanics of contact and friction. applied mechanics reviews 47:209-226. mulhearn to, samuels le (1962), the abrasion of metal: a model of the process, wear 5:478-498. narendera ks, annaswamy am (1989), stable adaptive systems prentice hall. pullen j, williamson jbp (1972), on the plastic contact of rough surfaces. proc. roy. soc (london) a327:159-173. seabra j, berthe d (1987), influence of surface waviness and roughness on the normal pressure distribution in the hertzian contact. asme j. of tribology 109:462-470. the journal of engineering research (tjer), vol. 18, no. 2, (2021) 72-81 *corresponding author’s e-mail: niko.knebel@berlin.de doi:10.53540/tjer.vol18iss2pp72-81 adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman nikolaus knebel* and harindren paramahamsa department of urban planning and architecture, german university of technology in oman, muscat, oman abstract: the discourse on energy-efficient buildings tends to shift out of the realm of architecture and into engineering and programming. in the latter disciplines, sophisticated solutions are always being developed and introduced to the green building industry. however, such solutions require financial and technical capacities that are only found in technologically highly developed countries. in developing countries and their ever-growing metropolises, a much more robust approach seems appropriate. this paper explores how the energy demand for creating comfortable interior environments in very hot climates can be significantly reduced by promoting straightforward architectural means. it seeks to relocate the discourse back into the realm of design, where less costly, more feasible, and long-lasting solutions can be found. the paper presents two case studies of residential architecture in muscat, oman. through simulations of solar insolation as well as daylight studies, it shows how architectural form can respond to two solar phenomena heat and light – in such a way that insolation on glazed facades is minimized, while illuminance of interiors is optimized. it claims that this can lead to a new regional architectural idiom that puts the famous modernist credo on its head by stating that energy-efficient buildings in a hot climate zone require “light and air, but no sun”. keywords: ventilation; insolation; illuminance, energy-efficiency, visual comfort; oman الحارة: دراستا حالة من مسقط، عمانتعدیل الشكل المعماري للمباني السكنیة في المناخات نیكوالس نیبل* و ھاریندرین باراماھمسا یمیل الحدیث عن المباني عالیة الكفاءة إلى التحول من مجال العمارة إلى مجاالت الھندسة والبرمجة، التي تقدم :الملخص قدرات مالیة وتقنیة ال تتوفر إال في البلدان باستمرار حلوال متطورة لصناعة المباني الخضراء، غیر أن ھذه الحلول تتطلب المتقدمة تقنیا، بینما یبدو اتباع نھج أكثر رصانة مناسبا في البدان النامیة وعواصمھا المتنامیة باستمرار. تستكشف ھذه الورقة یر عن طریق وسائل كیف یمكن أن تقلل الطاقة المستھلكة لخلق بیئات داخلیة مریحة في المناخات شدیدة الحرارة إلى حد كب معماریة بسیطة ومباشرة، كما أنھا تسعى إلعادة الحدیث عن المباني عالیة الكفاءة إلى مجال التصمیم المعماري، حیث توجد حلول طویلة المدى أقل تكلفة، وأكثر جدوى. تستعرض الورقة دراستي حالة للعمارة السكنیة في مسقط في عمان، وتظھر من –الشمسي الساقط، وضوء النھار، كیف یمكن للشكل المعماري أن یستجیب لظاھرتین شمسیتین خالل محاكاة االشعاع بطریقة تقلل االشعاع الشمسي الساقط على الواجھات الزجاجیة، في الوقت نفسھ الذي یتم فیھ تحسین –الحرارة والضوء إقلیمي جدید یقلب التوجھ الحداثي، ویعبر عن اإلضاءة الداخلیة. تزعم الورقة أن ھذه الطریقة ستؤدي إلى أسلوب معماري عمارة عالیة الكفاءة في المناخ الحار ترفع شعار "ضوء وھواء، لكن ال شمس!". .الضوء ؛العمارة، المباني الموفرة للطاقة ؛الراحة الحراریةاإلضاءة؛ ؛التشمس التھویة؛ الكلمات المفتاحیة: mailto:author@email.com adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman 73 1. introduction the development of energy-efficiency of buildings however late it may be in the race against global warming is progressing. yet, it can be observed that this progress tends to take place in the development of more and more technically sophisticated building components and increasingly automated operations. the field of innovation seems to be in the fields of civil, mechanical, and electrical engineering as well as in programming rather than in architectural design. a poignant example where this tendency is manifested is the prototypical zero-energy buildings that were developed in the framework of the so-called solar decathlon competitions over the past decade in europe, the united states, china, the middle east, and lately also in africa (solar decathlon, 2020). while these buildings become more and more equipped with smart materials and smart technologies, their architectural designs show limited progress. this development bears the risk of raising the planning, construction, and operations of energyefficient buildings to a level of complexity at a time when simplicity is called for. today, the world is facing a “global air-conditioning surge” (cox, 2012). this phenomenon describes that the current rapid urban growth that is taking place in the metropolises of the global south, which are mainly in hot or tropical climates, is coupled with the rapid emergence of middle-class societies and thus growing demand for building cooling systems. however, for the time being, this demand can often only be met by installing outdated and/or oversized mechanical cooling systems and technologies in buildings that are designed and constructed inefficiently. current practice in muscat, oman, for example, is to construct single-leaf concrete blocks without insulation. the problem is that the progress in making buildings more energy-efficient, which is mainly happening in the disciplines of engineering and programming, is outstripping the financial capacities and technological expertise of the markets. the question raised here is, how much progress towards energy-efficiency of buildings can be achieved through the architecture itself, which may be a technically much simpler discipline than engineering or programming but a very effective means for saving energy through design concepts? all too often, this potential of the discipline of architecture is not used to its full extent. this can be seen on a global scale in the contradiction between the general similarity of contemporary architectural forms versus the variety of climatic conditions. this paper makes a case for architectural design. it aims to explore how in the early stage of a design process, the shape of a building can be formed with simple means as arranging functions in floor plans and orienting openings in facades so that the basis for energy-efficiency is already laid before the planning process enters into the potentially more cost-intense and high-tech realms of engineering and programming in later stages. this approach to energy efficiency through architectural design is demonstrated in this paper with two case studies of residential buildings in muscat, oman. the city is growing rapidly (statista, 2018), and the already hot climate is bound to increase even more (pal & eltahir, 2015). the construction industry operates mainly with unskilled labourers, and the market for innovative technologies seems too small to attract state-of-the-art products and service providers. in this situation, which is representative of many metropolises of the global south, architectural design can make a significant difference. when designing a building, the most basic climatic setting is its relation to the sun. the sun is a source of energy, and it radiates heat and light. this dual effect from the same cause poses a particular challenge for the design of an energy-efficient building in a hot climate because the form of a building needs to keep the heat out but let the light in. or, to formulate it as a design paradox: buildings should be “closed but open” at the same time. respectively, the building form evolves from a negotiation of the two parameters of insolation and illuminance; both of which can be simulated through local weather data and a digital model of the building’s preliminary design and be optimized at an early stage of the planning process, as is done in this study. this paper connects the issues of energy efficiency and comfort by looking at buildings as artefacts in relation to their urban location and solar geometry and also by conceiving buildings as environments that are perceived by their users. the binding element is the building form. the research questions are thus: 1) insolation: what are the consequences for building form when under the conditions of intense insolation in hot climates, the heat gains in buildings are minimized through measures on the primary level of architectural design in the early stages of a design process (before a secondary level of material science or civil engineering is even entered)? 2) illuminance: what are the consequences for building form when under the climatic conditions of a predominantly clear sky with a very high luminance as well as a very high annual amount of sunlight hours, visually comfortable conditions for the indoor environment are pursued? 3) ventilation: what are the consequences for building form when the periods of natural ventilation through cross-ventilation of all major rooms are enabled through design for the period of passive cooling throughout the year to be prolonged? nikolaus knebel and harindren paramahamsa 74 figure 1. aerial view of a typical suburb in the western outskirts of muscat, oman, showing singlefamily villas on spacious plots; wide streets, no trees. (photo by harindren paramahamsa). figure 2. windtower. figure-ground plan; exploded axonometry 1; exploded axonometry 2. (illustrations by authors). figure 3. windtower. rendering of the view from the west, view into the shaded, ventilated outdoor space within the building volume. (illustrations by authors). this paper shows how adaption to climatic conditions by avoiding insolation on glazed openings, providing comfortable illuminance, and enabling natural ventilation can lead to designs that are different from the standard practice and can be an inspiration for a new vernacular architectural idiom. the paper first introduces the two designs. it then explains the method of developing these designs with regard to simulations of insolation, solar access as well as illuminance. finally, it discusses the findings with an outlook on climate-adapted building typologies for this region. 2. cases the two case studies of this paper are designed for residential architecture in muscat, oman; by coincidence, they are located in the same suburb, al khoud. one is a multi-unit tower, and the other is a single-family villa, both of which stand for the predominant building typologies of the urban fabric (figure 1). hence, beyond the individual cases, the design solutions as well as the research findings from this paper will be transferrable to other design tasks and will have a general value for the discourse on energyefficient buildings in this city and climate. case one – windtower the windtower project is located on a 500m2 plot with a side length of 25m by 20m and is part of a linear row of several such units. building regulations as defined in the title deed (“krooki”) allow for full coverage of the plot (built area ratio = 1,0), a height of nine stories (“g+8”), and a mix of commercial and residential use. the plot is oriented such that the longer sides face open spaces; towards a very broad street in front-facing south-west and towards a wide-open field at the back facing northeast. the northwest and southeast sides face adjacent buildings of the same volume, which stand at a distance of only 5m (figs 2, 3). the design (figs. 2, 3) aims at negotiating insolation, illuminance, and ventilation through architectural form. the first decision was to avoid insolation on glazed facades exposed to the sun, which led to the radical closure of the elevations facing the southwest and northeast facades. to compensate for this, the second decision was to open the centre of the tower and orient all apartments towards this void space to provide illuminance and ventilation. furthermore, all service spaces are arranged towards the rear, facing adjacent buildings with gaps between them to allow for a second source of daylight as well as cross-ventilation. the vertical circulation is through two cores that are connected by an outdoor bridge in the middle of the tower. on the entrance level to each unit, there is a large terrace that provides a shaded and ventilated outdoor living space. all rooms are simple and spacious rectangles, and through this form, they are functionally neutral. there are two rooms per unit on each floor, and the units can be combined as singlefloor, duplex, or triplex apartments. the plans are based on a modular grid, and the construction could easily be implemented with prefabricated construction elements. in this design, simplicity is key. while the basic layout is clear, simple, and similar in all four quadrants of the tower, the two main facades are designed differently in order to adjust precisely to the solar orientation. however, this is achieved by modifying only one architectural element. this element is a segment of the enclosing wall that is moved to the ideal position to keep sunlight from directly hitting the windows for most of the day while letting the breeze adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman 75 come through. on the ground floor towards the street side, these wall segments are turned outwards to create an entrance gesture and tension by introducing an irregularity in the otherwise relatively regular composition of the facade. case one – umbrellahouse the umbrellahouse is located on a rectangular 23m by 50m plot and has an area of 1,150 sqm. it is within an area of similar such plots on which free-standing, twostorey villas are built. the building regulations, as stated in the title deed (“krooki”), define setbacks from the plot boundary of 8m in front, 2m on the sides, and 5m at the back and allow a maximum building height of 8m. the rectangular plot is oriented at a 30° angle along its longer dimension to the northeast (figs. 4 and 5). the design’s (figs. 4 and 5) goal is to balance between light, air, and sun through architectural form. the first decision that was taken was to put all private living areas of the villa on one level and to elevate this main volume of the building to create a shaded garden. the ‘public’ area of the house – mainly the arab reception room called majlis – remains in a smaller volume on the ground floor and has a view to both sides of the garden. the second decision was to keep all outer facades closed to ensure privacy and avoid insolation on unprotected glazed facades. the third decision was to punctuate the upper volume with light yards, which are also the main elements that structure the spaces. on the one side, a large circular yard is set within the square volume of the living room, creating a continuum of different zones – an open kitchen, play areas, dining, and living room (see floor plan in figure 11). on the other side, two narrow yards divide each pair of the four bedrooms and provide light along the long sides of these spaces. in both yards, there are small terraces that enhance the connection between the garden and the living spaces above it. to underline the aim of simplicity, only one architectural element of this design is used for negotiating the issues of insolation versus illuminance. in this case, it is the openings of the roof slab over the yards and their dimensions and positionings. for the long and narrow yards, the openings are set at a third of the yard’s width, which will keep almost all direct sunlight off the glazed facades and yet provide ample daylight. over the large round courtyard, the opening is an oval that is turned with its long axis into exact northsouth orientation and the short axis in the eastwest direction. this creates wider roof cantilevers to keep out the flat sun in the mornings and afternoons, and a shorter one to prevent the higher sun at noon from hitting the glazed facades (figs 4, 5). 3. methods the projects are in an early stage of the design process and therefore this research was carried out through simulations with a digital model. since the designs call for simplicity, the same approach was applied for the simulation and evaluation method. it was decided to use only one cad programme, autodesk revit 2019 (v19.0.1.1) with its corresponding autodesk insight add-in 2019 (v4043), a building information modelling (bim) software, through which an integrated design process is enabled, which means that direct feedback between design iterations and simulation results is possible. revit is widely used in the professional as well as the educational realm of architecture, and thus the steps demonstrated here can be replicated by students and practitioners. figure. 4. umbrellahouse. figure-ground plan; exploded axonometry 1 facing the street side towards the east; exploded axonometry 2; rendering. (illustrations by authors). figure 5. umbrellahouse. rendering of the view from the east, view into the shaded and ventilated outdoor space under the building volume. (illustrations by authors). nikolaus knebel and harindren paramahamsa 76 input data simulations of insolation, as well as illuminance, require local weather data. in revit, this data is available through a link to the databank of the world meteorology organization (wmo). for this study, the data from a weather station located within a range of less than 10km from the sites of the two case studies were selected (autodesk revit weather data, 2020). digital model the digital model was built with only the minimal required specifications to emphasize the desire for simplicity at this early stage of the design process. to study insolation as well as illuminance, the important question to answer is: which building elements are solid/opaque and which are transparent/glazed. for the opaque building elements, the structural strength or thermal capacity of the material is not yet relevant at this stage and is thus not defined further. however, what needs to be defined is the light reflectance value (lrv) of the surfaces of the building elements. to limit the specifications to this one factor, the authors defined a new material in the revit materials library, which is actually a “non-material” because it has no other specifications than a light reflectance value that is expressed through equivalent colour values (rgb) an option that is offered in the revit settings (autodesk revit reflectivity settings for materials, 2020). further, to deliberately stay away from any detailed design decisions, the reflectance for the different surfaces within the rooms, the standard settings from leed analyses were carried over (leed v4, 2019). thus, the specifications for the light reflectance value throughout both projects are walls 50%, floors 20%, and ceilings 80%, respectively. for the glazed building elements, the specifications are kept equally simple. since neither the frame and glass qualities nor other factors, e.g., the solar heat gain coefficient (shgc), is important at this early stage of the design process, the windows were modelled simply as 100% transparent walls. further, it needs to be noted that in the digital models of the two case studies, the urban contexts of the plots were treated differently. while for the windtower, the mutual shading effect from the adjacent buildings was taken into consideration, this was not done for the umbrellahouse, which was modelled without neighbouring buildings. their effect on the result was deemed negligible because it would affect only a small portion of the facade, which is without glazed openings. simulated parameters different simulations were needed to find an optimal balance between minimal heat gains through glazed openings as a precondition for energy-efficiency and sufficient daylighting as a condition for visual comfort. they were carried out and set in relation to each other: accumulated insolation on the opaque and glazed facades and solar access in addition to illuminance levels in the rooms of the buildings. accumulated insolation the parameter insolation describes the energy per unit area that is received from the sun on a particular surface and is cumulated over a time interval. the unit of insolation is kwh/m2. for the analysis of the insolation, the relevant input data is the global horizontal irradiance (ghi) that sums up the direct and diffuse horizontal irradiance from the sun on a horizontal surface. in this study, the given ghi data for the specific location that is pre-set in revit was used. since the focus of the analysis in this paper is to compare the insolation on opaque as well as on glazed facades only vertical building elements walls and windows and not horizontal elements roof and floors were selected for the evaluation. and since the aim is to link the analyses of insolation and illuminance, a decision about whether to measure at a specific point in time or over a period of time had to be taken. since the illuminance analysis is often, e.g., in leed, set to three representative days for significantly different solar situations (jun 21, sep 21, dec 21) as well as a morning and an afternoon time as a representative of one day (9:00 a.m. and 3:00 p.m.). it was decided to simulate the insolation on the same specific days, however, not as a peak or average insolation, but as accumulated insolation for the time from sunrise to sunset. in this way, the strategies for reducing heat gains to create energy efficiency and the strategies for mediating daylight to create visual comfort can be assessed in relation to each other. solar access a solar access study is a means to analyze which areas in a building receive direct sunlight during a defined period of time and within defined time intervals. the latter issue, the accumulation of measurements over a period of time, makes it different from a shading study that only features a particular point in time. in this study, we look at the duration of one day (6:00 a.m. – 6:00 p.m.) on the specific days that were defined to be representative for each season (jun 21, sep 21, dec 21). while in temperate climate zones solar access studies are used to ensure that rooms in buildings have a minimum of direct sunlight as a means of comfort, this is different in hot climate zones like muscat. here, the aim is to avoid direct sunlight in rooms of buildings, and thus in this case the defined threshold within a solar access study is to be set at a maximum time value (e.g. one or two hours) in order to achieve thermal and visual comfort as well as energyefficiency. for the design development of the case studies presented in this paper, the solar access study was used to fine-tune the shape of the architectural elements that block the sun from the glazed facades. illuminance the parameter illuminance describes the total adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman 77 luminous flux incident on a surface per unit area (ref) at a defined point in time. the unit of illuminance is lux (lx). the analysis of the illuminance requires setting the input data about the luminance of the sky appropriately for the location. the important factors are the daylight hours and the cloudiness of the sky. daylight hours are the period between sunrise and sunset at the horizon during which natural illumination of the environment is experienced. in muscat, the length of daylight hours per day varies moderately over a year, with 10:41 hours of daylight on the 21st of december and 13:35 hours on the 21st of june (weatherspark, 2020). cloudiness describes the percentage of cloud coverage of the sky when viewed from a specific location. muscat experiences two significantly different periods of cloudiness. for nine months, from september through may, the sky is clear for 70% of the day and overcast for only 15% of the time, while for three months, in june, july, and august, the sky is either clear or overcast or in an inbetween state during 30% of the time for each of these conditions (weatherspark, 2020). given these conditions, it seems justified to set the luminance distribution of the sky for the simulation to “cie clear sky” as defined by the international commission on illumination (cie) and offered in revit as a default setting (autodesk revit sky models settings, 2020). the illuminance in the rooms is then measured in the digital model through grid points that are set 80cm (32 inches) above ground and at a distance of 30cm (12 inches) from each other. other parameters in addition to the parameters that can be quantified through the above-mentioned simulations, there are other parameters that contribute to the success of a design at this stage of planning that can be qualified through an analogue assessment, e.g., natural ventilation. natural ventilation natural ventilation of buildings is an important factor in energy-efficient design because it not only contributes to a healthy air exchange but also enhances thermal comfort through air movement, which enables the user to perceive air temperature as slightly lower than it actually is. a well-ventilated building saves energy because the periods of time in which a building can be cooled by passive mode without technology can be prolonged. 4. results the results from the analyses of insolation, solar access as well as illuminance are first viewed separately for each case and then discussed in comparison. results for case one – windtower insolation the insolation on the opaque as well as the glazed facades is analyzed by dividing the building into four quadrants in accordance with the floor plans that have four separate units per floor. for each quadrant, two building elements are measured, and the results are compared: one is the opaque, outer facade facing the open spaces in front or behind the building, and the other is the glazed, inner facade that is oriented towards the central void space. as can be seen in each series of simulations for the four quadrants for each of the three dates (figures 6, 7, 8, 9), the difference between the insolation on the opaque facade as compared to the one on the glazed facade is significant. while throughout the year, the opaque facades receive between 1,700 wh/m2 and 4,000 wh/m2, the glazed facades receive between 350 wh/m2 and 700 wh/m2. in summary, through the windtower’s design approach, the accumulated insolation on the glazed facades can be reduced by around 80% as compared to the opaque facades, which are exposed to direct sunlight. this is a significant difference and an important precondition for an energy-efficient building. figure 6. simulation of insolation (in wh/m2) on opaque as well as glazed facades of northwest quadrant for the duration of one day on 21 jun, 21 sep, 21 dec. figure 7. simulation of insolation (in wh/m2) on opaque as well as glazed facades of south-west quadrant for the duration of one day on 21 jun, 21 sep, 21 dec. figure 8. simulation of insolation (in wh/m2) on opaque as well as glazed facades of south-east quadrant for the duration of one day on 21 jun, 21 sep, 21 dec. nikolaus knebel and harindren paramahamsa 78 solar access the simulation shows that on the typical floor plan of the windtower, the solar access never rises above two hours per day at any time of the year. the areas where this period of direct sunlight exposure is experienced are very small in comparison to the vast majority of the area where there is none at all (figure 10). this result proves the effectiveness of the design approach that is based on a simple floor plan with four similar units oriented to different directions but which are differentiated through only one simple, robust and permanently-installed architectural element a wall on the terrace that serves as a “sun-blocker”. this one wall is fine-tuned for each of the four directions by adjusting its length and position on the terrace. illuminance the simulations of the illuminance levels on a typical floor plan of the windtower in the morning (9:00 a.m.) and afternoon (3:00 p.m.) on each of the three selected days, show that the illuminance levels are rarely below 50lx nor above 1000lx (figs 11, 12, 13). there are some minor exceptions; for example, during the morning hours in the southeast apartment, some areas close to the window show illuminance levels higher than 1000lx (figs. 11, 12, 13). during the afternoon hours in the northeast apartment, some areas have illuminance levels lower than 50lx (figs 11, 12, 13). however, these are temporary situations that only affect small areas of the overall floor plan. overall, it can be seen that the illuminance levels within the apartments are generally in a range that provides visual comfort for users. furthermore, it is remarkable that the outdoor areas of each apartment show only slightly higher illuminance levels than the indoors, which makes them comfortable areas into which the living spaces can be extended. it also means that views from inside to outside have little contrast and thus glare is avoided, which again raises the visual comfort for users. results for case two – umbrellahouse insolation the simulation of the insolation on the outer, opaque facades and on the inner, glazed facades of the umbrellahouse shows that for one day of each of the three days that are representative for the seasons, there is always a stark contrast between the insolation on the two different facades (figure 14). while the cumulative daily insolation on the inward-facing facades of the glazed courtyards is between 350-700 wh/m2, it is mostly in the range of 3-4,500 wh/m2 on the fully closed outward-facing facades in this case, the south and east facades; a difference of around 85-90%. some minor areas of the glazed facades show slightly higher insolation due to temporary exposure to direct sunlight, but this is deemed rather negligible due to the limitation of the area and the period of exposure. this can be seen in the round courtyard facade in winter and also the south-facing long facades of the rectangular courtyards in spring and autumn as well as its shorter ends in summer (figure 14). further exceptions on the outer facade are in summer when the south-facing facade receives less energy due to the seasonally high altitude of the sun, as well as in winter when the east (and respectively the west) facade receives less energy due to the higher degree of the azimuth of the sun. figure 9. simulation of insolation (in wh/m2) on opaque as well as glazed facades of northeast quadrant for the duration of one day on 21 jun, 21 sep, 21 dec figure 10. simulation of solar access on typical floor plan between 6 am and 6 pm on 21 jun, 21 sep, 21 dec figure 11. simulation of illuminance levels (lx) of the typical floor plan on 21 jun 9 am (left), 3 pm (right) figure 12. simulation of illuminance levels (lx) of the typical floor plan on 21 sep 9 am (left), 3 pm (right) figure 13. simulation of illuminance levels (lx) of the typical floor plan on 21 dec 9 am (left), 3 pm (right) adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman 79 to complete the assessment of the consequences of the umbrellahouse’s design concept that aims at providing comfortable living spaces with minimal insolation, not only the accumulated insolation on the large glazed facades that envelope the interior living spaces were simulated but also the insolation on the exterior living spaces in the garden. the results show how the elevated volume of the villa shields off solar radiation from the ground floor and reduces the accumulated insolation significantly for around half the area, and in many of these areas up to 95% (figure 15). solar access the simulation of the solar access parameter shows that no area of the rooms on the main floor of the umbrellahouse receives more than 2 hours of direct sunlight at any time of the year, while most of the area received none at all (figure 16). the least solar access is in summer, while the most are in winter. however, this situation needs to be seen in combination with the overall weather during this period when the outdoor temperatures are pleasant, and potential heat gains can be compensated by natural ventilation. illuminance the illuminance levels for the rooms of the umbrellahouse are simulated for a typical floor plan, which in this case is the elevated first floor at the three set days of the year at 9:00 a.m. as well as at 3:00 p.m. (figs 17, 18, 19). what can be seen from these simulations is that the illuminance levels of the rooms on the first floor of the house are always between a minimum of 50lx and a maximum of 1000lx. in the bedrooms, there are some minor incidents of illuminance levels of 2000lx or above; in the morning hours of spring and fall and very scarcely in the afternoon hours of summer as well as in the afternoon hours of winter. these exceptions are too small in area and duration as to matter for the overall assessment of the situation. the large room that wraps around the circular courtyard and contains living, dining, playing, and kitchen zones receives a remarkably even illuminance throughout the year. only in winter does the south-facing zone receive direct sunlight and show illuminance levels above 2000lx. as mentioned in the discussion of the cases above, such a situation at this time of the year is not ideal but occurs only for a short period of time. 5. discussion the results of the simulations of the two case studies can be discussed on two levels. first, what they mean for the research questions of this paper. and second, what do they mean for an outlook onto the bigger picture of the architecture of energy-efficient buildings in oman as well as in emerging cities in hot climates at large? figure 14. simulation of insolation (in wh/m2) on opaque as well as glazed facades of northwest quadrant for the duration of one day on 21 jun, 21 sep, 21 dec. figure 15. simulation of insolation (in wh/m2) on ground level (garden) for the duration of one day on 21 jun, 21 sep, 21 dec. figure 16. simulation of solar access on typical floor plan between 6 am and 6 pm on 21 jun, 21 sep, 21 dec. figure 17. simulation of illuminance levels (lx) of the typical floor plan on 21 jun 9 am (left), 3 pm (right). figure 18. simulation of illuminance levels (lx) of a typical floor plan on 21 sep 9 am (left), 3 pm (right). figure 19. simulation of illuminance levels (lx) of a typical floor plan on 21 dec 9 am (left), 3 pm (right). nikolaus knebel and harindren paramahamsa 80 reflections on the research questions on the first research question: the designs of the windtower, as well as the umbrellahouse and their successful performance, is proven through the simulations presented in this paper. it reaffirms that the consistent avoidance of heat gains through architectural elements leads to specific idioms of architecture in a hot climate zone: a) the outer envelope becomes radically closed; b) to compensate for this, the building volume needs to incorporate void spaces; c) the openings then orient towards to these inner voids. such climate-responsive idioms are found in the vernacular architecture of the region but are so far rarely applied in contemporary buildings and urban typologies, at least not in oman. on the second research question: the designs show that a focused pursuit of visual comfort requirements leads to a re-balancing of open and closed facades that is different from the vernacular approach. while in the traditional architecture of hot climates, e.g., in oman, windows are deliberately kept very small and embedded in the plane of largely closed facades, the contemporary designs presented in this paper have very large glass facades. however, these facades are always protected from direct sunlight. the difference between these two approaches is that the former creates visual discomfort due to the stark contrast between the large, dark wall surfaces and the small, very bright light spots from the windows, while the latter provides a comfortable and ample distribution of diffuse and indirect daylight deep into the rooms. further, the designs demonstrate the advantages of providing two sources of daylight from different directions. in the process of designing these projects, it also became apparent that in hot climates with predominantly clear skies, the provision of daylight needs to be approached with a maximum rather than a minimum illuminance level. on the third research question: the two designs show that effective cross-ventilation can only be provided if the room configurations are such that corridors are avoided, and rooms have multiple openings, if possible, in different directions. such an inner logic of space arrangements leads to an outer form that is not compact but spread out or folded. this logic is counterintuitive to the often-formulated aim of keeping the surface-to-volume ratio at a minimum in order to save the facade area and reduce exposure to the hot environment. however, compact volumes tend to become quite deep and thus create insufficiently lit and unventilated interiors. therefore, designs should avoid compact forms that appear energy-efficient only at first sight and instead prioritize more expansive forms that provide passive cooling through ventilation and visual comfort through evenly distributed indirect daylight. at this point, it should be added that while the designs have addressed the articulation of the floor plans and facades to adapt to the climatic conditions, they have not addressed the design of the roof yet. the reason for this is that given the effective application of photovoltaic panels in the region – it is assumed that all buildings should be equipped with a flat roof-top array of the largest possible size to generate electricity and operate as a double roof (knebel, n & wassmer, m., 2019). the findings presented here are drawn from only two case studies, but the conclusions could be transferred beyond these individual cases and could become strategies for designing energy-efficient and comfortable buildings in hot climates. a starting point for further exploration would begin with employing architectural means before adding mechanical solutions. other disciplines that contribute to the same goals may pick up the planning process from here but not jump over these fundamental issues. architecture that consistently translates local climatic conditions into built form does create a local idiom and thus an identity for a place a quality that should be conceptualized from real requirements rather than be construed from an artificial image as is so often the case (knebel, 2019). 6. conclusion while this paper makes a case for architectural design to be emphasized in the quest for energy-efficient buildings, the results have consequences for urban design, too. the paper shows that when considering the climate-adaptiveness of buildings in a hot climate like in oman, quite unusual forms evolve. for example, the windtower is a bulky volume with closed exterior facades and inward-looking apartments towards a central void space. which typology would this be in the known terms of urban design? or, the umbrellahouse is a flat volume with closed exterior facades, and inwardoriented rooms elevated over a garden. as which typology would this be categorized in urban design? in an attempt to describe these typologies for the two cases, one could say that the windtower is a ‘flipped block’, in which the courtyard moves from a horizontal to a vertical position. similarly, the typology of the umbrellahouse could be seen as a ‘flipped slab’, in which the thin flat volume is no longer in a vertical but in a horizontally elevated position (figure 20). figure 20. transformation of existing typologies as a consequence of climate adaption. adapting architectural form of residential buildings in hot climates: two case studies from muscat, oman 81 seen from this perspective, the results from this paper open up a necessary and more general discussion about appropriate building typologies for this region and climate. these two potentially new typologies are certainly not the only ones that are appropriate for the context of residential buildings in the hot climate of oman. instead, they are only the beginning of a necessary search for appropriate and innovative typologies. finally, the fact that major large-scale urban design projects in oman (such as madinat al irfan, the new downtown of muscat, or al duqm, the new harbour city on the arabian sea) neglect specific climate implications and instead propose urban typologies that originate from the temperate climate zone and are not necessarily appropriate for a location within the hot and arid climate zone, show the urgency for this discussion to be carried on. it can be summarized that for climate-adaptiveness on both scales, architectural and urban, the first step is to exploit the potential of form with a self-imposed limitation to pursue simplicity before moving on into the potentially much more complex realms of material sciences and systems automation. this paper shows that the rule of thumb for designing for energy efficiency and comfort in hot climates through negotiating illuminance, ventilation, and insolation can be subsumed as an inversion of the modernist architect’s credo and says: “light, air, and no sun”. conflict of interest the authors declare no conflict of interest. funding this study is part of the ecohaus project of the german university of technology that was led by the author from 2011-2021 and received several research grants from the research council of the sultanate of oman. acknowledgment the two case projects are based on designs commissioned by private clients, which were then developed further in the process of this study. the authors would like to thank the clients for their trust and enthusiasm in finding new solutions for buildings in oman. thanks also go to christina cernovsky for supporting the early design phase of the windtower as well as to karsten schlesier, carmen neuhaus, felix huebbers and mariam abdou as well as markus oetzel, wilfried hofmann of bauplus for contributions to the planning of the umbrellahouse in its early phase. references autodesk revit reflectivity settings for materials (2020): https://knowledge.autodesk.com/searchresult/caas/cloudhelp/cloudhelp/enu/bpa-perfor mancestudies/files/guid-01ae68a3-ed41-44d3 -8809-58c4d492e6c4-htm.html, accessed on 25/03/20 autodesk revit sky models settings (2020): https://knowledge.autodesk.com/searchresult/caas/cloudhelp/cloudhelp/enu/bpahelp/files/guid-a93dc4a2-7017-4191-bfe0915ccec4d3ad-htm.html, accessed 25/03/20 autodesk revit weather data (2020): https://knowledge.autodesk.com/support/revitproducts/learnexplore/caas/cloudhelp/cloudhelp/2019/enu/rev it-analyze/files/guid-ce607f67-d169-4e6d9b7f-35bee9101bec-htm.html, accessed 25/03/20 cox, s., 2012. cooling a warming planet: a global air conditioning surge. retrieved from https://e360.yale.edu/features/cooling_a_warming _planet_a_global_air_conditioning_surge. accessed 24/04/2020 knebel, n. & wassmer, m. 2019. monitoring, evaluating and optimizing the energy supply of a photovoltaic system of a zero-energy-building in oman: case study of the gutech ecohaus. in proceedings of cate19 (comfort at the extremes), hariot-watts university, dubai, uae, 10-11 april 2019 knebel, n., 2019. identity through efficiency. (re)discovering passive cooling strategies as an architectural idiom for the gulf region. in proceedings of cate19 (comfort at the extremes), hariot-watts university, dubai, uae, 10-11 april 2019 leed, 2019. reference guide for building design and construction, v4. retrieved from https://www.usgbc.org/resources/leed-reference-guidebuilding-design-and-construction. accessed 05/05/20 pal, j., eltahir, e., 2015. future temperature in southwest asia projected to exceed a threshold for human adaptability. nature clim change 6, 197– 200 (2016). https://doi.org/10.1038/nclimate2833, accessed 15/05/20 solar decathlon, 2020. www.solardecathlon.gov statista, 2018. oman: population growth from 2007 to 2017. available from: https://www.statista.com/statistics/455525/populati on-growth-in-oman/, accessed 12/10/18 weatherspark, 2020. https://weatherspark.com/y/105797/averageweather-in-muscat-oman-year-round, accessed 29/03/20 نيكولاس نيبل* و هاريندرين باراماهمسا 1. introduction 2. cases 3. methods simulated parameters 4. results insolation 5. discussion 6. conclusion conflict of interest funding acknowledgment the journal of engineering research (tjer), vol. 18, no. 2, (2021) 124-136 *corresponding author’s e-mail: sneha.asrani.96@gmail.com doi:10.53540/tjer.vol18iss2pp124-136 thermally comfortable affordable housing: a study on residential building code in india sneha s. asrani 1*, rajan u. rawal 1, yash d. shukla 1, peter graham 2, priyanka p. bhanushali 1, and arjun r. desai 1 1 centre for advanced research in building science and energy (carbse), cept university, k. l. campus, ahmedabad, india 2 global buildings performance network (gbpn), 24 avenue hoche, paris, france abstract: on average, india has more than 3000 cooling degree days (cdd). the multifamily public housing being constructed under india’s prime minister awas yojana (pmay) scheme is aimed at providing formal housing to the society’s economic weaker section (ews). it is essential that this housing delivers thermally comfortable indoors to the occupants. this study mapped the design and construction practices followed under pmay urban (pmay-u) against india’s residential energy building code, eco niwas samhita (ens). the metric prescribed in ens is residential envelope transmittance value (retv). for 80 pmay-u projects, information related to spatial design, and walling assemblies were collated. 30 projects were short-listed for detailed analysis. the retv for selected projects was calculated. the analysis demonstrated that walling assemblies and technologies having a lower thermal transmittance value (u-value) resulted in meeting the ens code-compliant retv number. the study further extended to assess the energy performance of the housing unit by exploring the change in walling assemblies. the dwelling unit’s thermally comfortable hours and energy performance index (epi) were reported to vary between 4145 and 6034, and 64 kwh/m2year and 68 kwh/m2year respectively. keywords: adaptive thermal comfort; affordable housing; eco niwas samhita, ecbc – r, pmay – u; residential building code. سكن مریح حراریا وبتكلفة معقولة: دراسة حول كود البناء السكني في الھند و بیتر جراھام و بریانكا بانیوشالي و أرجون دیساي كالروال و یاس شسنیھا ارساني*، و راجان درجة تبرید یومي في المتوسط. یھدف نظام اإلسكان العام متعدد العائالت الذي 3000یوجد في الھند أكثر من :الملخص وبالضرورة یجب أن تكون ھذه أسسھ رئیس الوزراء أواس یوجانا إلى توفیر مساكن للطبقات األقل دخال من السكان، المساكن ذات بیئات حراریة مریحة لقاطنیھا. قارنت ھذه الدراسة بین ممارسات التصمیم والبناء المتبعة في تنفیذ نظام اإلسكان العام مع تلك المتبعة في كود البناء والطاقة للمباني السكنیة في الھند، الجدید ذكره أن المعیار المتبع في كود البناء مشروعا من 80والطاقة ھو نفاذیة الغالف السكني. تم تجمیع معلومات متعلقة بتصمیم الفراغات وتكوینات الجدران من مشروعا لتحلیلھا تحلیال دقیقا، كما تم حساب نفاذیة الغالف السكني لھذه المشاریع. 30مشاریع اإلسكان العام، ثم تم اختیار لتقنیات ذات النفاذیة الحراریة المتدنیة حققت نفاذیة الغالف السكني المطلوبة في كود أظھر التحلیل أن تكوینات الجدران وا البناء والطاقة. وباإلضافة إلى ھذا، فقد امتدت الدراسة لتقیم أداء الطاقة للوحدة السكنیة من خالل استكشاف أثر تغییر ساعة 4145ریة للوحدة السكنیة تتراوح بین تكوینات الجدران، األمر الذي اتضح على أثره أن ساعات الراحة الحرا كیلو واط ساعة للمتر 68كیلو واط ساعة للمتر المربع و 64ساعة، وأن مؤشر أداء الطاقة یتراوح بین 6034و المربع. .البناء السكني كود السكن بأسعار معقولة؛ الراحة الحراریة التكیفیة؛ الكلمات المفتاحیة: thermally comfortable affordable housing: a study on residential building code in india 125 nomenclature retv residential envelope transmittance value of the building envelope (w/m2). shgc solar heat gain coefficient. u-wall thermal transmittance value of the wall (w/m2 k). u-roof thermal transmittance value of the roof (w/m2 k). wfr window to floor ratio. 𝜔𝜔𝑖𝑖 orientation factor of opaque/non-opaque building envelope components. top operative temperature (°c). 1. introduction the building envelope’s thermal characteristics play a significant role in achieving thermal comfort, reducing the energy demand and consumption. reliance on mechanical systems to achieve thermal comfort impacts the environment negatively. the electricity consumption owing to the use of air conditioning systems and fans accounts for approximately 20% of the total electricity used in buildings, across the globe (international energy agency, 2018). the combined negative environmental impact of the increasing built-up area and frivolously energy-consuming buildings necessitates a building envelope that provides comfortable indoors throughout harsh outdoor conditions, without much use of electricity. india has more than 3000 cooling degree days (cdd). metropolitan cities, especially the ones having hotdry or warm-humid climate (figure 1 [a]), are the worst affected; e.g., ahmedabad having hot-dry climate has 3514 cdd, mumbai and kolkata having warm-humid climate have 3386 and 3211 cdd, respectively, and delhi and hyderabad having composite climate have 2881 and 3221 cdd, respectively (sivak, 2009). such high prevalence of cdd expands the growth potential of mechanical cooling and ventilation – primarily room air conditioners (ac), to achieve a comfortable indoor environment (somvanshi, 2019). this would lead to a rise in residential energy consumption, estimated to multiply by more than eight times by 2030. however, this estimated increase can be moderated with policy and market strategies prioritizing energy efficiency (global buildings performance network, 2014). the indian government launched its ambitious ‘housing for all’ mission in 2015. under this mission, the prime minister awas yojana urban (pmay-u) scheme aimed to address the urban housing shortage among economically weaker section (ews), low-income group (lig), and middle-income group (mig). as part of the program, approximately 11.30 million houses have been approved, and 5.1 million have been constructed so far. this has attracted an investment of usd 100.26 billion, out of which usd 14.86 billion have already been released as government assistance (pmay (u), 2021). india’s bureau of energy efficiency (bee) launched the voluntary energy conservation building code-residential (ecbc-r), also called eco niwas samhita (ens), in 2018. it adopts a variation of over all thermal transfer value (ottv) (rawal, et al., 2018; saidur, et al., 2009) called, residential envelope transmittance value (retv). retv determines a building envelope’s performance based on the thermal transmittance of the walling (u-value) and roofing assembly (u-roof), solar heat gain coefficient (shgc), and solar reflectance index. this paper presents a study aimed at understanding the feasibility of applying ens to pmay-u projects, bridging the gap between the ‘what is’ – the prescribed measures for code compliance, to the ‘how to’ – steps involved in assessing code compliance. moreover, the study was specifically aimed at perceiving the implication of various walling technologies in achieving ens compliance. the study focused on the affordable housing projects (ahp) component of the pmay-u scheme developed under a partnership between government and private real estate developers. while other studies did attempt to evaluate a residential building’s thermal performance using retv, but these efforts were limited to either one building or a maximum of two climate zones (agarwal & samuelson, 2021; kumar et al., 2021); walling assemblies used in affordable housing construction and their impact on the indoor environment has also been studied separately (bardhan & debnath, 2018; sen et al., 2014). this study simultaneously encompasses several building envelopes across all climate zones, variations in walling assembly, and energy simulations. 2. methodology national building code (nbc, 2016) and energy conservation building code (bee, 2018) have divided india into five climate zones namely hot-dry, warm-humid, temperate, composite, and cold. as a first step, information of various pmay-u projects was retrieved from secondary sources – i.e., public domain government websites (government of rajasthan, local self government department 2018; govt. of himachal pradesh, directorate of urban development 2014; govt. of india, bmtpc n.d.; govt. of india, ministry of housing and urban affairs 2018; govt. of india, ministry of housing and urban affairs 2018; govt. of jharkhand, urban development & housing department 2019; govt. of karnataka, karnataka slum development board 2016; govt. of maharashtra 2016; govt. of maharashtra 2017; govt. of maharashtra 2018; govt. of maharashtra 2019a; govt. of maharashtra 2019b; govt. of punjab, punjab urban development authority n.d.; govt. of tamil nadu, tamil nadu slum clearance board 2017; mhada n.d.; mhada n.d.; mohua n.d.; union sneha asrani, rajan rawal, yash shah, peter graham, priyanka bhanushali, and arjun desai 126 territory of daman, daman municipal council 2017; upavp n.d.). an initial list of 80 projects was compiled with details related to [a] availability of the building’s technical drawings, [b] construction material, [c] climate zone, and [d] status of completion. figure 1[a] shows the distribution of those 80 projects based on their climate zone. the initial list was filtered based on the availability of the building’s technical drawings. the filtered projects, i.e., the ones for which technical drawings were available, were sampled considering the following parameters: • walling assembly: this type and magnitude of residential construction needs to be economical, time-bound, and have a long service life. the building materials & technology promotion council (bmtpc) identified and elaborated innovative walling assemblies and technologies which would not only be resource-efficient but also optimize construction time (mohua, 2020). separate model projects called the light house projects (lhp) are aimed at understanding the intricacies of transferring such walling technologies to the field. thus, an effort was made to capture the variety of walling assemblies and technologies. • climate zone: figure 1[b] shows the distribution of sampled projects (n=30) based on their climate zone. this sampling criterion ensured that there was at least one project from each climate zone. • stage of construction: projects were categorized into either of the following – [a] proposal/planning stage, [b] under construction, or [c] completely constructed. this helped in mapping out the change in building/construction material over time. the sampling resulted in a finalized sub-list of 30 projects (hereafter known as selected projects). each selected project consisted of one or more apartment building blocks, having an average of 6 floors (minimum 2, maximum 36). the dwelling units had a 1bhk (bedroom hall kitchen) configuration, and carpet area in the range of 22.8 m2 to 55.1 m2. figure 1[c] shows the distribution of buildings (n=92), belonging to selected projects (n=30), based on their climate zone. figure 2 mentions various walling assemblies and technologies used in the construction of selected projects. the majority of projects (n=19) used conventional walling assemblies like burnt-clay and fly-ash bricks, aac, or concrete blocks. however, the remaining projects (n=11) employed innovative technologies/assemblies like reinforced eps core panel or glass fiber reinforced gypsum system. the technical drawings of all selected projects were digitized to *.dwg format. the retv of all buildings (n=92) from the selected projects was calculated using the ens compliance check tool. the retv formula (bee, 2018) is as mentioned in equation. 1: figure 1. distribution of – [a] 80 projects for which information was gathered, [b] selected projects (n=30), [c] buildings from selected projects (n=92) – as per climate zone. thermally comfortable affordable housing: a study on residential building code in india 127 figure 2. distribution of selected projects based on walling assembly. 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 = 1 𝐴𝐴𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 × ��𝑎𝑎 × ∑ �𝐴𝐴𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 × 𝑈𝑈𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 × 𝜔𝜔𝑖𝑖𝜔𝜔𝑖𝑖� 𝑛𝑛 𝑖𝑖=1 � + �𝑏𝑏 × ∑ �𝐴𝐴𝑛𝑛𝑜𝑜𝑛𝑛−𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 × 𝑛𝑛 𝑖𝑖=1 𝑈𝑈𝑛𝑛𝑜𝑜𝑛𝑛−𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 × 𝜔𝜔𝑖𝑖�� + �𝑐𝑐 × ∑ �𝐴𝐴𝑛𝑛𝑜𝑜𝑛𝑛−𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 × 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑜𝑜𝑜𝑜𝑖𝑖 ×� 𝑛𝑛 𝑖𝑖=1 �� ( 1) • 𝐴𝐴𝑜𝑜𝑛𝑛𝑒𝑒𝑜𝑜𝑒𝑒𝑜𝑜𝑜𝑜𝑜𝑜: envelope area (excluding the roof) of dwelling units (m2). it is the gross external wall area including the area of the walls and the openings such as windows and doors. • 𝐴𝐴𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 and 𝐴𝐴𝑛𝑛𝑜𝑜𝑛𝑛−𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 : areas of different opaque and non-opaque building envelope components, respectively (m2) • 𝑈𝑈𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖 and 𝑈𝑈𝑛𝑛𝑜𝑜𝑛𝑛−𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑖𝑖: thermal transmittance of different opaque and non-opaque building envelope components (w/m2k) • 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑜𝑜𝑜𝑜𝑖𝑖 : equivalent solar heat gain coefficient values of different non-opaque building envelope components • 𝜔𝜔𝑖𝑖: orientation factor of respective opaque and non-opaque building envelope components; it is a measure of the amount of direct and diffused solar radiation that is received on the vertical surface in a specific orientation for each of the buildings, the dimensions and uvalue of the walling and roofing assembly, and dimensions of windows, doors, ventilators, and their overhangs were input manually into the tool, which in turn returned retv, window to floor ratio (wfr), and u-roof as output parameters. table 1 give the compliance values for output parameters. note that only the walling assembly’s u-value was taken into account for retv calculation, while the u-roof was returned as an independent output parameter. previous studies mention that for low-rise buildings the roofing assembly’s impact on heat transfer is more significant than that of the walling assembly (al-tamimi, 2021). however, this study chose to only study the influence of the walling assembly’s heat gain on the building envelope, owing to the following reasons: • a typical pmay-u building was mid-rise, having 6 floors. • innovation and variation in construction materials and assemblies across the nation. • the ens code only takes the walling material’s u-value into account for retv calculation. the u-roof has been prescribed to be limited to 1.2 w/m2k, to reduce heat gain from the roof. this study assessed the change in a case-study building’s energy performance corresponding to the change in its walling assembly. figures 3, 4, and 5 illustrate the unit, and floor plan, and section of the building situated in ahmedabad’s hot-dry climate; the building had 14 floors, having 16 dwelling units (carpet area: 27.2 m2) on each floor, and using fly-ash bricks as walling material. figures 6, and 7 establish various conventionally (‘a’ to ‘f’) and nonconventionally (‘g’ to ‘r’) used walling assemblies, and their nomenclature, i.e., w_id. the u-values of these walling assemblies (except ‘a’ and ‘b’) were referred from rawal, et al.’s study (2020) and can be seen in fig. 11. figure 3. case-study dwelling unit plan. 0 5 10 15 20 25 30 aac blocks burnt clay bricks eps core with gypsum plaster fly ash bricks glass fiber reinforced gypsum (gfrg) monolithic rcc precast wall and slab reinforced eps core panel system solid concrete block structural stay in place formwork system no. of projects w al lin g m at er ia l hot & dry warm & humid composite temperate cold all dimensions are in mm sneha asrani, rajan rawal, yash shah, peter graham, priyanka bhanushali, and arjun desai 128 table 1. ens compliance limits. parameter\climate hot-dry warm-humid composite temperate cold min wfr (%) 10 16.6 12.5 12.5 8.3 max u-roof (w/m2 k) 1.2 1.2 1.2 1.2 1.2 max retv (w/m2) 15 15 15 15 1.8 table 2. electrical appliance inputs for simulation. location equipment nos. wattage (w) location equipment nos. wattage (w) bathroom led light 1 9 toilet incandescent bulb 1 60 drawing room led tv 1 37 kitchen led light 1 9 cfl bulb 1 20 bedroom led light 2 9 tube light 1 28 fan 1 75 fan 1 75 figure 4. floor plan of the case study. figure 5. section a-a of the case study. thermally comfortable affordable housing: a study on residential building code in india 129 the building’s walling assembly was sequentially changed from ‘a’ to ‘b’, ‘c’, ‘d’, till ‘r’, keeping other input parameters constant. the retv corresponding to each change was calculated. furthermore, the implication of changing the walling assembly on the building’s energy efficiency was explored by conducting energy simulations, using energyplus version 9.3. here, only the walling assembly was varied, keeping all other parameters, such as roofing assembly, door, window details, etc., unchanged. the floor-to-floor height was taken to be 3 m; the roofing assembly comprised of 4 mm thick ceramic glazed tiles, 50 mm thick cement screed, 150 mm thick rcc slab, and 15 mm thick internal plaster, having a resultant u-value of 3.89 w/m2k. windows were taken to be 6 mm thick single glazed clear glass having a u-value of 5.8 w/m2k, and shgc 0.8. the building was taken to be operated in mixed mode, with the room split air conditioner (ac) being run at 28°c between 10 pm and 6 am only in the bedroom, while the rest of the house was in naturally ventilated mode. the average household size was 4 – comprising 2 adults and 2 children. the electrical equipment load input can be found in table 2. the energy performance index (epi) and thermally comfortable hours were the output of simulations. please note that the simulation results could not be validated owing to the lack of a baseline. 3. results figure 8 shows the retv variation across various climate zones. referring to table 1, fig. 2, and fig. 8 simultaneously, it can be observed that the retv of most of the projects from hot-dry and composite climate was non-compliant, while that of the projects from warm-humid, temperate, and cold climate was compliant. this could be attributed to the use of regionally preferred walling assemblies; assemblies used in hot-dry and composite climate were found to have higher u-values, while those used in warmhumid, temperate, and cold climate had relatively lower u-values (refer fig. 11 and fig. 8). the correlation between walling material’s u-value and the envelope’s retv is seen to be established in fig. 10. the retv variation across various spatial orientations can be seen in fig. 9. the mean retv was found to be the lowest for the east-west (e-w) orientation, i.e. when the building’s longer side would be parallel to the e-w axis. the remaining output parameters of the ens compliance tool were wfr and u-roof. the wfr for the majority of the buildings (63 out of 92) was found to be compliant with ens; no statistically significant correlation between retv and wfr was found. the u-roof for the majority of the buildings (89 out of 92) was found to be non-compliant with the ens value. figure 6. walling assemblies and corresponding id for ‘a’ to ‘f’. sneha asrani, rajan rawal, yash shah, peter graham, priyanka bhanushali, and arjun desai 130 figure 7. walling assemblies and corresponding id for ‘j’ to ‘r’. figure 11 depicts the u-value and corresponding retv range for various walling assemblies. it can be gathered that extruded poly styrene (eps) based walling assemblies had lower u-value, hence, the retv range was on the lower side. aac blocks were found to have the lowest u-value and retv amongst the conventional walling assemblies. new and innovative technologies like monolithic rcc or precast wall and slab had a relatively higher u-value and retv. further on, the implication of varying the walling assembly on the building’s energy performance was thermally comfortable affordable housing: a study on residential building code in india 131 explored. the walling assembly of the case study was changed to the w_ids mentioned in figs. 6, and 7; the retv corresponding to each change was calculated. the energy performance index (epi) and thermally comfortable hours (for 80% satisfaction as per imac 2014 natural ventilation model) were also derived, for each change. figure 12 shows the u-value and retv for the respective alternatives. assemblies with a lower uvalue resulted in lower retv. this correlation was corroborated by fig. 10. hi-tech walling assemblies ’d’ to ‘h’ yielded retv close to compliance value, non-conventional assemblies ‘k’ to ‘r’ performed similarly. interestingly, both ‘b’ and ‘r’ had the same component – aac blocks, yet the difference between their u-values was 0.54 w/m2 k. the ens specified u-value of aac blocks was used for ‘b’; it was the u-value of just the block. the u-value of ‘r’ was determined by a guarded hot box experiment (rawal. et. al., 2020); it was the resultant u-value of the aac blocks + heat-insulating, self-curing mortar assembly. the lowest u-value and retv were achieved by ‘r’. figure 13 illustrates the epi and thermally comfortable hours across the walling assemblies. the non-conventional assemblies ‘k’ to ‘q’ had relatively higher epi than the conventional ones. the thermally comfortable hours across the assemblies ranged between 5000 and 5500 (out of 8760 hours in a year). the innovative walling assemblies ‘d’ and ‘f’ yielded the lowest retv and epi while providing relatively more thermally comfortable hours. amongst conventional assemblies, aac blocks (‘b’) were found to have relatively lower retv and epi, and more thermally comfortable hours. figure 14 illustrates the range of indoor operative temperatures (top) of one dwelling unit on the 2nd floor during the day (6:00 am to 6:00 pm) and night, across all walling assemblies. studying figs. 12, 13, and14 simultaneously, the top range across ‘a’ to ‘j’ is similar, with the average day top being lower than that in the night. while ‘m’ to ‘r’ follow the opposite trend – average night top lower than that in the day; likely due to relatively high u-values, thus, resulting in lesser comfortable hours. also, ‘k’ and ‘l’, having comparatively higher retv, have a wider gap between the average day and night top. figure 15 shows the difference in epi and energy consumption concerning that of the baseline (‘a’); figure 16 shows the difference in equivalent co2 emitted and its savings (%) for all walling assemblies. non-conventional walling assemblies (‘k’ to ‘r’ in table 2) provided retv close to ens compliant value, but, resulted in higher energy consumption and co2 release relatively. walling assemblies having insulation (‘d’ to ‘i’) had the potential to reduce operational energy consumption and co2 release by 2%. figure 8. retv variation across different climate zones. figure 9. retv variation across different orientations. figure 10. correlation between u-value and retv. sneha asrani, rajan rawal, yash shah, peter graham, priyanka bhanushali, and arjun desai 132 figure 17 depicts the compounded impact of operational energy consumption and equivalent co2 emitted by respective assemblies, at the city (ahmedabad, india) level. the walling assemblies with insulation (eps or glass fibre) – ‘d’ to ‘i’, required much less operational energy, thus, emitted less equivalent co2. amongst the conventional assemblies, aac blocks (‘b’ and ‘r’) were deemed to be the best performer. figure 11. variation of u-value and retv across walling assemblies. figure 12. u-value and retv across various walling assemblies. figure 13. epi and comfort hours across various walling assemblies. 0 1 2 3 4 5 0 5 10 15 20 25 30 aac blocks burnt clay bricks eps core with gypsum plaster fly ash bricks glass fibre reinforced gypsum monolithic rcc precast wall and slab reinforced eps core panel solid concrete block structural stay in place formwork thermal transmittance-u-wall (w/m2 k) retv (w/m2) w al lin g m at er ia l retv u-value 0.0 0.5 1.0 1.5 2.0 2.5 0 2 4 6 8 10 12 14 16 18 20 a b c d e f g h i j k l m n o p q r t he rm al t ra ns m it ta nc eu -w al l (w /m 2 k ) r e t v ( w /m 2 ) walling assembly retv ens compliant retv u-value 63 64 64 65 65 66 66 67 67 68 68 3500 4000 4500 5000 5500 6000 6500 a b c d e f g h i j k l m n o p q r e p i (k w h/ m 2 / ye ar ) n o. o f h ou rs in t he y ea r walling assembly comfort hours epi thermally comfortable affordable housing: a study on residential building code in india 133 figure 14. range of indoor operative temperature across various assemblies. figure 15. the difference in epi and energy consumption across walling assemblies. figure 16. the difference in equivalent co2 emitted and savings (%) across various walling assemblies. figure 17. energy consumption and co2 emitted based on walling assembly – on the city level. -3% -2% -1% 0% 1% 2% -2.0 -1.0 0.0 1.0 2.0 a b c d e f g h i j k l m n o p q r δ e ne rg y co ns um pt io n (% ) δ e p i (k w h/ m 2 / ye ar ) walling assembly epi energy consumption -3% -2% -1% 0% 1% 2% -0.6 -0.4 -0.2 0.0 0.2 0.4 a b c d e f g h i j k l m n o p q r δ c o 2 em it te d (% ) δ c o 2 em it te d (m t ) m ill io ns walling material δ co2 emitted (mt) δ co2 emitted (%) thermally comfortable affordable housing: a study on residential building code in india 134 4. discussion the analysis showed a strong correlation between the u-value of the walling assembly and the envelope’s retv. moreover, the mean retv was lowest for buildings with an e-w orientation. this study brings out the following design-level simplistic, changes which could be easily implemented for significant improvement in the building’s thermal performance: • selection of appropriate walling material or assembly: the walling material is often selected by prioritizing the speed and cost of construction, with a reduced cognizance of its thermal characteristics. selection of walling assemblies with an insulating layer is advisable, also suggested in bardhan & debnath’s study. for example, constructing monolithic or precast rcc walls in a hot-dry climate is likely to be cost and time-efficient, but can prove to be unadvisable from the perspective of thermal performance. rcc has a higher u-value and will facilitate the inflow of excess heat into the building; however, walling assemblies with insulation, such as eps or glass fibre, with a relatively lower u-value, would allow a significantly lower amount of heat into the building. it is also advisable to utilize thermally insulating mortars/plasters when the usage of insulation-based walling assemblies is not feasible. for example, as mentioned in fig. 12, the walling assembly consisting of aac blocks and a heat-insulating mortar has a u-value of 0.45 w/m2k and the lowest retv (9.35 w/m2), but the assembly yielded relatively higher epi and fewer comfortable hours in comparison to the other assemblies. therefore, it is pertinent to balance the tradeoffs between retv, epi, and thermal comfort. another study highlights how optimizing the window material thickness could reduce the shgc, thus, improve the thermal and energy performance of a building (kumar et al., 2021). • optimizing building orientation: as a part of the initial building design process, assessing various building orientations and calculating their corresponding retvs can be helpful. • maintaining wfr: a statistically significant correlation between retv and wfr was not found; however, the building’s wfr must comply with ens due to its influence on the ventilation rate. • providing appropriate roofing insulation: as mentioned in section 3, the u-roof for most of the projects was non-compliant with the ensspecified value. even though the u-roof is not considered for the retv calculation, a higher uroof can lead to uncomfortable thermal conditions on the top floor, also suggested in indraganti's study. another study (agarwal & samuelson, 2021) underlined how the ens does not provide insights into the building envelope performance based on regional climatic context. 5. conclusion this study provides insight into the pmay-u housing in context to the residential energy building code ens. the data collection started with compiling information on 80 pmay-u projects from secondary sources. using filtering and sampling criteria, such as availability of technical drawings, construction material, climate zone, etc., a sub-list of 30 projects was finalized. these selected projects had one or more apartment building blocks. the retv of all buildings was calculated using the ens compliance tool. the analysis showed that the walling assembly’s u-value and the building’s orientation had a significant impact on the building’s thermal performance. aac blocks were found to have the lowest u-value and retv. new and innovative construction technologies like precast wall and slab, and monolithic rcc had relatively higher u-value and retv. furthermore, the implication of a change in walling assembly on a building’s thermal and energy performance was assessed by calculating the energy performance index (epi) and thermally comfortable hours. for a case study, the retv, epi, and thermally comfortable hours corresponding to various walling assemblies, conventional and nonconventional, were calculated. the new and innovative walling assemblies with insulation had a lower u-value, retv, epi, and provided relatively more comfortable hours. these assemblies had the potential to reduce the building’s operational energy demand by up to 2%. this 2% reduction on a building level can be amplified multifold on a city level. hence, this study reinforces the potential of ens code in making better-informed decisions for the design and construction of efficient buildings. the authors recommend that the thermal characteristics of the walling, roofing, and window materials be paid more consideration. moreover, by evaluating the retv for various orientational permutations, the thermal performance of the envelope could be optimized., in addition to improving energy efficiency, resource usage, and occupant comfort. conflict of interest the authors declare that there is no conflict of interest. funding global buildings performance network (gbpn) and shakti sustainable energy foundation (ssef). thermally comfortable affordable housing: a study on residential building code in india 135 acknowledgment the authors gratefully acknowledge the grant from global buildings performance network (gbpn). references agarwal, a., & samuelson, h, 2021, too hot to stay at home: residential heat vulnerability in urban india. journal of physics: conference series, 2069(1), 012166. https://doi.org/10.1088/17426596/2069/1/012166 al-tamimi n, 2021, an optimum thermal insulation type and thickness for residential buildings in three different climatic regions of saudi arabia. civil engineering and architecture, 9(2), 317 327. doi: 10.13189/cea.2021.090205. bardhan, r., & debnath, r, 2018, evaluating building material based thermal comfort of a typical lowcost modular house in india. materials today: proceedings, 5(1), 311–317. https://doi.org/ 10.1016/j.matpr.2017.11.087 bee, 2018, energy conservation building code – residential. government of rajasthan, local self government department. 2018, presentation of rajasthan state for sanction by central sanctioning & monitoring committee under housing for all [pdf], retrieved: 14 september, 2021, from https://pmay-urban.gov.in/uploads/ presentations /rajasthan_csmc.pdf govt. of himachal pradesh, directorate of urban development. 2014, invitation of expression of interest [pdf], retrieved: 14 september, 2021 from http://bmtpc.org/datafiles/cms/file /pdf_files/03_tender_eoi_solan_hpud.pdf govt. of india, bmtpc. n.d., demonstration housing project using emerging technologies [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/files/1bm tpc-demo-housing.pdf govt. of india, ministry of housing and urban affairs. 2018, annual report, 2017-18 [pdf], retrieved: 14 september, 2021, from http://164.100.161.188/upload/uploadfiles/files/ne w_ar-2017-18%20(eng)-website.pdf govt. of india, ministry of housing and urban affairs. 2018, csmc_minutes_24-07-2018 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/files/csm c_minutes_24-07-2018.pdf govt. of jharkhand, urban development & housing department. 2019, proposal for 34 projects under vertical iii (ahp) [pdf], retrieved: 14 september, 2021, from: http://mohua.gov.in /upload/uploadfiles/files/jharkhand(1).pdf govt. of karnataka, karnataka slum development board. 2016, csmc meeting, 11th august 2016 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/files/2csm c11ksdbppt.pdf govt. of maharashtra. 2016, csmc meeting, 22nd july 2016 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/ files /6csmc010maharashtra.pdf govt. of maharashtra. 2017, csmc meeting 30th september 2016 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles /files/4csmc013maharashtra.pdf govt. of maharashtra. 2018, csmc meeting, 28th november 2019 [pdf], retrieved: 14 september, 2021, from https://pmay-urban.gov.in /uploads /presentations/maharashtra(5).pdf govt. of maharashtra. 2019a, csmc meeting, 29th august 2019 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/ uploadfiles/files/maharashtra_46_csmc.pdf govt. of maharashtra. 2019b, csmc meeting, 28th june 2019 [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles /files/maharashtra(7).pdf govt. of punjab, punjab urban development authority. n.d., development of ews dwelling units at patiala-sangrur road, patiala city [pdf], retrieved: 14 september, 2021, from https://pmayurban.gov.in/uploads/presentations/punjab_csmc .pdf govt. of tamil nadu, tamil nadu slum clearance board. 2017, proposal for 456 projects under vertical iii & iv [pdf], retrieved: 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/files/csmc 021tamilnadu.pdf global buildings performance network. 2014, residential buildings in india : energy use projections and savings potentials. viewed -14 september 2021, <https://www.gbpn.org/wpcontent/uploads/2021/06/infographie_ppt_1. pdf> indraganti, m, 2010, using the adaptive model of thermal comfort for obtaining indoor neutral temperature: findings from a field study in hyderabad, india. building and environment, 45(3), 519–536. https://doi.org/10.1016/j.buildenv. 2009.07.006 international energy agency. 2018, the future of cooling opportunities for energyefficient air conditioning. viewed – 14 september 2021, < https://www.iea.org/reports/the-future-of-cooling> kumar, g., thakur, b., & de, s, 2021, energy performance of typical large residential apartments in kolkata: implementing new energy conservation building codes of india. clean technologies and environmental policy, 23(4), 1251–1271. https://doi.org/10.1007/s10098-020http://mohua.gov.in/upload/uploadfiles/files/1bmtpc-demo-housing.pdf http://mohua.gov.in/upload/uploadfiles/files/1bmtpc-demo-housing.pdf http://164.100.161.188/upload/uploadfiles/files/new_ar-2017-18%20(eng)-website.pdf http://164.100.161.188/upload/uploadfiles/files/new_ar-2017-18%20(eng)-website.pdf http://mohua.gov.in/upload/uploadfiles/files/csmc_minutes_24-07-2018.pdf http://mohua.gov.in/upload/uploadfiles/files/csmc_minutes_24-07-2018.pdf https://pmay-urban.gov.in/uploads/presentations/punjab_csmc.pdf https://pmay-urban.gov.in/uploads/presentations/punjab_csmc.pdf https://pmay-urban.gov.in/uploads/presentations/punjab_csmc.pdf http://mohua.gov.in/upload/uploadfiles/files/csmc021tamilnadu.pdf http://mohua.gov.in/upload/uploadfiles/files/csmc021tamilnadu.pdf thermally comfortable affordable housing: a study on residential building code in india 136 02022-7 mhada. n.d., mouje-khoni, thane. retrieved on 14 september, 2021, from https://mhada.gov.in/en/ content/mouje-khoni-thane mhada. n.d., scheme-wanjara. retrieved on 14 september, 2021, from https://mhada.gov.in /sites/default/files/scheme-wanjara.pdf mohua. n.d., dpr no.1 ray nagar federation, solapur [pdf] retrieved on 14 september, 2021, from http://mohua.gov.in /upload/ uploadfiles/ files/5maharashtra02-09csmc.pdf mohua, 2020, compendium of innovative emerging technologies shortlisted under global housing technology challenge india. national building code of india (nbc), 2016, <http://www.bis.org.in/sf/nbc.htm> pathak a, garg t, kumar s, 2020. a policy strategy for decarbonising the building sector: facilitating ens implementation in affordable housing. technical report, retrieved from https://aeee.in/ pmay (u). [online], retrieved: 29 november, 2021, from https://pmay-urban.gov.in/. rawal r, kumar d, pandya h, airan k, 2018, residential characterization. ahmedabad: centre for advanced research in building science and energy. technical report, retrieved from: https://shaktifoundation.in/ rawal r, shukla y, shah a, gowri g, 2020, thermal performance of walling material and wall technology, part-2. technical report, retrieved from http://carbse.org/, https://www.beepindia.org rawal r, shukla y, bhanushali p, desai a, asrani s, 2021, a policy strategy for decarbonizing the building sector: project reportjanuary 2021. technical report, retrieved from http://carbse.org/ roychowdhury a, sareen r, singh m, 2020, beyond the four walls of pmay: resource efficiency, thermal comfort and liveability in the affordable housing sector. technical report, retrieved from http://.cseindia.org/ sen, r., kandra, s., & chattopadhyay, s, 2014, investigation on the performance of alternative walling materials in an affordable housing unit situated in warm humid climate. 30th international plea conference: sustainable habitat for developing societies: choosing the way forward proceedings, 1(december 2014), 524–531. sivak m, 2009, potential energy demand for cooling in the 50 largest metropolitan areas of the world: implications for developing countries. energy policy, 37(4), 1382–1384. https://doi.org/ 10.1016/j.enpol.2008.11.031 somvanshi a, 2019, a midsummer nightmare: decoding the link between comfort, space cooling and energy consumption in a climate-stressed world. technical report, retrieved from http://.cseindia.org/ upavp. n.d., majhola yojana, moradabad. retrieved: on 14 september, 2021, from: https://upavp.in/pmay/majhola-yojna/ union territory of daman, daman municipal council. 2017, pradhan mantri awas yojanahousing for all (urban). retrieved: on 14 september, 2021, from http://mohua.gov.in/upload/uploadfiles/files/dama n%26diu_27_csmc.pdf http://www.bis.org.in/sf/nbc.htm https://aeee.in/ https://shaktifoundation.in/ http://carbse.org/ http://.cseindia.org/ http://.cseindia.org/ https://upavp.in/pmay/majhola-yojna/ http://mohua.gov.in/upload/uploadfiles/files/daman%26diu_27_csmc.pdf http://mohua.gov.in/upload/uploadfiles/files/daman%26diu_27_csmc.pdf سكن مريح حراريا وبتكلفة معقولة: دراسة حول كود البناء السكني في الهند سنيها ارساني*، و راجان روال و ياس شكلا و بيتر جراهام و بريانكا بانيوشالي و أرجون ديساي 1. introduction 2. methodology 4. discussion 5. conclusion conflict of interest funding acknowledgment references the journal of engineering research (tjer), vol. 18, no. 2, (2021) 114-123 *corresponding author’s e-mail: jdiazborrego@us.es doi:10.53540/tjer.vol18iss2pp114-123 urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate julia díaz-borrego*, ignacio acosta, miguel ángel campano, and samuel domínguez-amarillo department of building construction, universidad de sevilla, seville, spain abstract: the urban configuration in the mediterranean area, which has noticeable solar radiation and significant thermal oscillation, is one of the factors that affect the microclimate and the urban heat island phenomenon of the cities located in that zone. to contribute to the improvement of urban comfort, this work presents an experimental methodological proposal for the analysis and multicriteria optimization of performance indicators using a parametric workflow through grasshopper. this workflow can lead to the development of a set of environmental design criteria as support tools in the design of the urban configuration. the methodology is based on the quantification of two parameters on the analysis and management of daylight in interior spaces by means of dynamic metrics and, on the other hand, on the evaluation of hygrothermal conditions and the urban heat island effect, which affect both the daytime comfort of passersby and the nighttime comfort of dwellings. this study considers the most typical variables of urban layout, both morphological and material. the case study analyzed is located in the city of seville (spain), which has a significant percentage of its contemporary buildings without a suitable adaptation to local climate conditions. the results obtained from the case study prove the potential of the proposed methodology to evaluate and optimize lighting and thermal performance at the neighbourhood scale. thus, it provides urban planners with a powerful decision support tool for both urban expansion and urban renewal. keywords: thermal comfort; daylighting; urban configuration; multi-criteria analysis and optimization; urban heal island; outdoor microclimate. الراحة الخارجیة: استراتیجیة الحي التخطیط الحضري لتحسین االشعاع الشمسي الساقط على المباني وتحقیق التوازن في في مناخ البحر األبیض المتوسط جولیا دیاز بوریغو* و ایجناثیو اكوستا و میجیل كامبانو و سامویل دومینجیز اماریلو المعروف بإشعاع شمسي ملحوظ وتذبذب حراري -یعد التكوین الحضري في منطقة البحر األبیض المتوسط :الملخص مل التي تؤثر على المناخ المحلي وظاھرة الجزر الحراریة الحضریة للمدن في المنطقة. بھدف تحسین أحد العوا -كبیر الراحة على المستوى الحضري، یستعرض ھذا البحث اقتراحا منھجیا تجریبیا للتحلیل والتحسین متعدد المعاییر لمؤشرات سھوبر؛ حیث یمكن أن تؤدي الطریقة المقترحة إلى األداء باستخدام طریقة معتمدة على متغیرات ومصممة ببرنامج جرا تطویر مجموعة من معاییر التصمیم البیئي التي یمكن اعتبارھا أدوات دعم في تصمیم التكوین الحضري، تعتمد المنھجیة میكیة، على القیاس الكمي لمعاملین مؤثرین في تحلیل وإدارة ضوء النھار في الفراغات الداخلیة عن طریق المقاییس الدینا وكذلك على تقییم ظروف الحرارة والرطوبة وتأثیر جزیرة الحرارة الحضریة، التي تؤثر بدورھا على الراحة الضوئیة للمشاة، والتبرید السلبي للمباني لیال. تتناول ھذه الدراسة متغیرات التخطیط الحضري الشكلیة والمادیة األكثر شیوًعا. تقع في مدینة إشبیلیة في إسبانیا، والتي تحتوي على نسبة كبیرة من المباني المعاصرة دون التكیف دراسة الحالة التي تم تحلیلھا المناسب مع الظروف المناخیة المحلیة. أثبتت نتائج الدراسة قدرة المنھجیة المقترحة على تقییم وتحسین االضاءة واألداء أداة قویة لدعم قرار التمدد والتوسع الحضري.الحراري على مستوى الحي؛ ولھذا فھي توفر للمخططین الحضریین البناء السكني. ؛الراحة الحراریة ؛جزر حراریة ؛التكوین الحضري الكلمات المفتاحیة: urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 115 1. introduction energy efficiency in buildings is nowadays a must. however, a deeper understanding of the urban environment impact on building performance and on the inhabitants’ well-being is required. solar obstruction caused by the urban environment is the main physical factor affecting lighting comfort (xue et al., 2014), as well as the proper chrono regulation due to the lack of daylight access. in addition, urban form influences the building energy performance, with an increase of up to 10% of the energy consumption (ratti et al., 2005). the urban configuration also affects the climatic and microclimatic variables, playing a decisive role in the urban heat island (uhi) phenomenon, which is mainly responsible for the nocturnal heat stress. likewise, outdoor temperature, wind speed, and solar radiation to which a building is exposed are not those represented by the synoptic regional climate, but the local microclimate modified by the structure of the city (givoni, 1989). thus, the urban configuration determines the outdoor environment quality, modifying the regional climate and affecting the comfort and health of the inhabitants. therefore, the correlation of multidisciplinary analyses can act as an active tool in decision-making (naboni et al., 2019). for this purpose, urban elements that influence the building performance -urban climate, urban morphology and density, and building form must be integrated with a single calculation. in addition, to assess the environmental efficiency of urban design, it is necessary to perform a comparison between solar accessibility inside buildings and outdoor thermal comfort, which will be affected by the solar potential and the uhi effect, respectively, determining the energy performance of buildings. however, due to its complexity, the coupling of simulation engines for multivariate analysis of urban environmental behaviour has been explored by a few studies (naboni et al., 2019; natanian & auer, 2020). 1.1. background urban environment performance is an essential factor to take into account in building efficiency and comfort analysis; as it influences user thermal sensation (outdoor thermal comfort), as well as the solar accessibility, lighting, and energy consumption of buildings. therefore, public space configuration affects both the external microclimate and the interior of the building (ali-toudert & mayer, 2006). daylight access is essential to avoid excessive energy consumption and improve visual comfort, as well as to help improve the health of the occupant through proper circadian stimulation (acosta et al., 2016). furthermore, circadian disturbance caused by lack of access to natural light can promote depression and other serious medical conditions (bellia et al., 2020). thus, compared to electric light, daylight offers better visual comfort by constantly varying its intensity, direction, and spectral composition (wong, 2017). however, excessive radiation and exposure to daylight can also lead to glare, as well as an excess of solar radiation can lead to overheating and thus a further increase in cooling energy consumption. therefore, it is necessary to perform an analysis to improve urban layout optimizing the use of daylight, both in terms of thermal radiation management and illuminance. in this regard, only a small number of studies analyze indoor daylight behaviour using dynamic metrics. most of them focused on the analysis of opening or shading devices. in addition, most research analyzing urban morphological parameters focuses only on the analysis of the building envelope solar potential. for its part, the urban configuration also affects the climatic and microclimatic variables, playing a decisive role in the so-called uhi phenomenon. uhi is an effect related to urbanization that causes significant climatic changes. it affects both air temperature and relative humidity, as well as airflow and wind speed in the built environment (núñez peiró et al., 2017), raising air temperature values in urban areas compared to rural areas (street et al., 2013). this phenomenon was first studied by oke (oke, 1988) at the end of the nineteenth century, who analyzed its links with different urban morphological parameters. urban densification contributes to the intensity of uhi, thus affecting the microclimatic conditions of the environment and, consequently, the indoor and outdoor conditions of comfort. it also aggravates the effects of climate change and heatwaves, posing a serious threat to the health and comfort of citizens. therefore, the uhi is the most appropriate indicator to quantify the environmental modification produced by the effect of urbanization on a specific territory. concerning the evaluation parameters, the main parameter that allows evaluating the urban morphology influence on microclimate and daylight accessibility is the urban canyon proportion. this urban parameter, coined by oke, specifies the ratio between the height of the building and the width of the road (h/w). thus, orientation and urban canyon proportion influence solar access potential, airflow, and other microclimatic data, such as the urban heat island effect. nowadays, the swift computational tools evolution and development allows more complex analysis by reducing the computational cost. in this sense, several studies have explored the correlation between urban form and different environmental performance parameters; however, few have developed a multicriteria combination of contradictory or conflicting aspects, a complex but essential challenge. 1.2. aims and objectives this research aims to reach a methodological proposal that allows determining the optimal urban layout to optimize the use of daylight in the mediterranean area, both in terms of thermal radiation management and illuminance. julia díaz-borrego*, ignacio acosta, miguel ángel campano, and samuel domínguez-amarillo 116 in addition, the proposed methodology also aims to provide a useful evaluation tool for urban planners, allowing predicting with a single workflow both the daylight and the thermal behaviour of the designed urban configuration. 2. methodology to promote energy savings and wellness in buildings, it is essential to quantify the urban configuration influence both on the interior daylight access and on the outdoor hygrothermal conditions. however, due to the complex calculations required, there is a lack of these studies. nevertheless, today, parametric tools development allows the environmental analysis at the urban scale to search for more efficient configurations. the parametric design method could be defined as a fast way to generate multiple design solutions and configurations by varying the values of the input parameters of the initial design model (taleb & musleh, 2015). thus, the integration in a parametric and dynamic workflow of urban data sets and the different tools used allows carrying out a multicriteria analysis and optimization. therefore, the method proposed in this study brings the integration of several parameters, factors, and metrics in a single parametric workflow through the rhinoceros 3d software tool with the aid of the grasshopper application. grasshopper is a visual programming language that allows the assessment of complex urban systems through a parametric workflow. such workflow can perform a multi-objective analysis and an optimization process because of the coupling of the simulation engines. the following specific light and environmental simulation programs are used in this workflow: radiance, energyplus, openstudio and urban weather generator (uwg). all of them couple to grasshopper through ladybug tools −ladybug, honeybee and dragonfly−, which acts as a parametric environmental plugin for grasshopper by means of a free and open-source development (roudsari & pak, 2013). thus, this workflow allows a multicriteria analysis and optimization of the initial results, to enhance the lighting efficiency of buildings in an urban fabric, as well as to establish design guidelines for cities for the study of climatic and morphological conditions at the neighbourhood scale. figure 1 shows the digital workflow developed for this study, as well as the input data, lighting and thermal metrics and plugins that materialize it. the following sections show the methods used to build this workflow. 2.1. input parameters the simulation parameters combine both static and dynamic parameters. static parameters are those that establish the climatic conditions and define the morphological parameters that are intrinsic to the urban fabric to be analyzed. dynamic parameters are those that determine the purpose of the evaluation, being related to light optimization and thermal comfort. thus, these values will vary to quantify their impact. the geometric model has been performed using rhinoceros 3d based on the morphological parameters that define the urban fabric to be analyzed. once the urban geometry has been obtained, it will be parameterized using grasshopper for its subsequent analysis and optimization. figure 1. methodological diagram of the multi-criteria environmental analysis of the parametric workflow in grasshopper. urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 117 the building understudy has been modelled in greater detail, introducing its floor plans, as well as the facade openings. however, the urban environment has been developed volumetrically, in a much more simplified way since the “influence of volumetric and relative building layout largely overweigh the importance or relevance of geometric detailing on building envelopes” (compagnon, 2004). in turn, the weather data file from the location under study was imported in energyplus (*.epw) format. both direct normal and diffuse horizontal irradiance, as well as the sky model built up by perez et al. (perez et al., 1993) and validated by cie (l’eclairage, 2003), were used. 2.2. microclimatic evaluation method uhi is the most evident and quantitative indicator of the environmental change produced by the effect of urbanization on a specific territory. its clearest consequence is the increase in air temperature during the night concerning the surrounding rural environment. the uhi intensity is defined as the thermal difference between the urban sector to be analyzed and the nearby rural or non-urban area at a given time (observatori metropolità del canvi climàtic, 2015). formally, ∆𝑇𝑇𝑢𝑢−𝑟𝑟 = 𝑇𝑇𝑢𝑢 − 𝑇𝑇𝑟𝑟 (1) where ∆𝑇𝑇𝑢𝑢−𝑟𝑟 is the uhi intensity, 𝑇𝑇𝑢𝑢 the temperature of the urban area, and 𝑇𝑇𝑟𝑟 the temperature of the rural or non-urban area. for this equality to be satisfied, the geographical conditions of both points must be similar. this expression is based on lowry's additive model (lowry, 1977) whose simplified expression states: the average temperature of a point in the urban area, 𝑇𝑇𝑢𝑢, is equal to the sum of the regional climate, c, the effect of local geographic factors in the urban area, 𝐺𝐺𝑢𝑢, and the influence of urbanization, u; 𝑇𝑇𝑢𝑢 = 𝐶𝐶 + 𝐺𝐺𝑢𝑢 + 𝑈𝑈 (2) the average temperature of a point in the rural area, 𝑇𝑇𝑟𝑟, is equal to the sum of the regional climate, c, plus the effect of local geographic factors in the rural area., 𝐺𝐺𝑟𝑟: 𝑇𝑇𝑟𝑟 = 𝐶𝐶 + 𝐺𝐺𝑟𝑟 (3) thus, it is necessary to evaluate the hygrothermal and microclimatic conditions of the urban context correctly to reduce their environmental impact (salvati et al., 2016). until now, this analysis has generally been carried out using typical climate samples based on historical records from the nearest meteorological station, which is usually rural; therefore, they were performed without taking into account the effect produced on the climate by the urban environment. thus, the present study starts with a limitation in the collection of the specific climatological data of the urban context; that prevents an adequate prediction of environmental optimization (alexander et al., 2015; nakano et al., 2015). recent studies have focused on how the urban environment affects climatology (bueno et al., 2015; salvati et al., 2016, 2019; street et al., 2013), as well as the effect of urban microclimate on building performance, resulting in the creation of the urban weather generator (uwg) software. uwg produces a modified weather file, generated from the rural environment file, which reflects the real microclimatic conditions of the urban setting under study. therefore, the “urban weather conditions” are introduced in the workflow of this study using uwg (through ladybug and dragonfly plugins). in this way, ladybug imports the energyplus weather files (*.epw) -rural weather datainto grasshopper. for its part, the dragonfly plugin allows uwg software coupling, which transforms the rural weather data, generating a modified weather file as shown in figure 2. finally, the ladybug plugin performed the detailed analysis of climate data, once transformed into 'urban weather data' and its subsequent visualization, as well as the interaction of said processed data with other plugins. once the modified urban weather file is obtained, the real urban thermal radiation is analysed by evaluating the mean radiant temperature (mrt) in the outdoor environment in july, the warmest month of the year. mrt is defined as “the uniform temperature of an imaginary enclosure in which the radiant heat transfer from the human body equals the radiant heat transfer in the actual nonuniform enclosure” (fundamentals, 2001). the evaluation was carried out using ladybug tools, by means of the tools and components shown in figures 3 and 4, on the outside, to assess which urban configuration is more favourable in this sense. 2.3. daylight evaluation method daylight potential is quantified using the metric, both spatial and dynamic, spatial daylight autonomy (sda). this one was introduced by the illuminating engineering society of north american (iesna) in 2012 in the lm-83 lighting protocol for the evaluation of lighting behaviour in architectural spaces (committee, 2012). the sda metric indicates daylight sufficiency, defined as the percentage of a surface area that exceeds 300 lux of illuminance during 50% of the occupancy time. figure 2. comparison of the rural and the simulated urban weather. julia díaz-borrego*, ignacio acosta, miguel ángel campano, and samuel domínguez-amarillo 118 figure 4. grasshopper components flowchart for the multicriteria environmental and experimental analysis. table 1. illuminance thresholds and definition of the sda metric. metric minimum illuminance threshold minimum temporal threshold analysis result acceptable threshold preferred threshold sda 300 lux 50% % surface with sufficient illumination >55% of surface area >75% of surface area the illuminance thresholds for the definition of this metric correspond to those stipulated by iesna (ies lm-83-12) (table 1). the adopted occupancy time for daylight evaluation is between 8:00-18:00 every day throughout the year, as also recommended by iesna. since the research is based on the comparative analysis of light behaviour resulting from the different variations of the urban fabric, the solar accessibility analysis was carried out without considering blinds. likewise, other important parameters to be considered, such as the material's reflectance of the urban context or glazing ratios, have been considered as fixed values and, therefore, they were not of relevance in the comparative analysis. this is acceptable as this paper, as mentioned above, focuses on the impact of the urban environment on indoor daylighting quality. the evaluation was carried out using the plugin of ladybug tools, honeybee. honeybee allows the evaluation of indoor daylighting by connecting the grasshopper visual programming environment with the validated light simulation engines for ray tracing radiance and daysim (acosta et al., 2015; reinhart & breton, 2009; reinhart & walkenhorst, 2001). for such evaluation, the sensors were set up in a grid on the first floor of the building to be analyzed; the analysis area is reduced only to the six-meter depth concerning facade, being the greatest depth that daylight can reach inside a building the "active zone". a sensitivity analysis is also performed for different values of the radiance daylight calculation parameters to balance the optimization of the calculation time and the accuracy of the light calculation. this evaluation was carried out on the lower floor of the building as it is the most unfavourable floor in terms of the illuminance evaluation of the sda metric. 2.4. multi-criteria analysis and optimization the results obtained from the performance indicators have been exported to a spreadsheet application for the processing and multicriteria analysis of the output data, figure 3. evaluation workflow to analyze the urban thermal radiation through ladybug tools. urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 119 through the colibri plugin of grasshopper, as well as to the software design explorer for their visualization. to achieve a set of environmental design criteria, optimization was carried out using grasshopper's octopus component. octopus allows programming evolutionary algorithms based on predefined parameters to optimize multiple object functions; as well as defining the optimal design parameter values to improve environmental performance. the use of this component is crucial to relate the parameters and performance indicators (indoor light accessibility and outdoor microclimatic conditions). the results obtained from the multicriteria optimization have been exported, finally, once again for the processing and analysis of the output data. 3. methodological application in a case study the application of the proposed methodology has been carried out using a small sample of the study. this case of study is located in the city of seville (spain), with a latitude of 37.42º and a longitude of 5.40º. seville has a mediterranean climate, with mainly clear skies and warm and dry summers. in addition, these high heat periods are prolonged, extending from may to september, raising the level of discomfort and also acting as a risk factor for the health of the inhabitants. furthermore, seville has a significant percentage of its contemporary buildings without a suitable adaptation to the local climate conditions. for the analysis, a typical urban expansion configuration has been chosen (figure 5), since it can be considered as the most frequent urban configuration in newly constructed neighbourhoods. 3.1. input parameters as mentioned above, the weather conditions correspond to the city of seville, and they are imported to grasshopper as energyplus weather files (*.epw) by ladybug. in this case, the static building and urban input parameters used are those shown in table 2. for its part, default daylighting radiance settings have been entered as shown in table 3. as dynamic parameters, to assess the proper use of daylight and improve microclimatic conditions, two orientations have been considered in the urban analysis, n-s/e-w (0.0) and ne-sw/nw-se (45.0) orientations, in each iteration. in addition, regarding the urban canyon, different street widths (5, 10, 15, 20, 25, 30, 35, and 40 m) were considered, generating different urban canyon proportions. 3.2. case study results the results show a direct relationship between urban canyon ratio and daylight accessibility; the lower h/w ratio, the higher the sda metric. for its part, orientation does not significantly affect solar accessibility, according to this location (figures 6 and 7). likewise, the proportion of urban canyon h/w equal to or greater than one had practically no effect on mrt; however, for lower ratios, mrt increases (figure 7). orientation also did not have a significant effect on the average mrt outdoors. however, if the detailed behaviour of each orientation is analyzed, a significant temperature difference between each orientation in ns/e-w urban distribution is observed. nevertheless, in ne-sw/nw-se urban distribution, similar temperature values on all facades of the block are achieved. 3.3. analysis of results from the results of this limited-size case study, it can be concluded that a lower proportion of urban canyons in the urban fabric will significantly improve the solar accessibility of buildings in urban configurations, without significantly varying the outside mrt if this proportion is equal to or higher than 4/5 ratio. furthermore, for proportions equal to or lower than a ratio of 4/5, the ne-sw / nw-se orientation appears to obtain more favourable values, as it shows a more uniform thermal behaviour on all the building facades. such lower ratios, in turn, favour street ventilation and reduce uhi in the urban canyon. furthermore, ratios equal to or greater than do not allow sufficient solar access, leading to increased energy demand and, therefore, to lower buildings energy savings. therefore, an inappropriate ratio can affect not only the economy of the occupants but also their health and visual comfort. however, thermal variations in proportions greater than 4/5 are imperceptible. it is necessary to reduce this ratio below 4/5 so that outdoor mrt begins to increase significantly. table 2. static simulation parameters. static input parameters values window-to-wall ratio (wwr) 25% glazing transmittance 0.55 interior building reflectance 0.50 building typologies courtyard building use residential average urban height 20 m maximum sensible anthropogenic heat flux of the urban area 20 w/m2 context reflectance 0.35 trees and grass cover 0 table 3. daylighting radiance settings. static input parameters values ambient bounces 2 ambient division 5000 limit weight 2e-5 julia díaz-borrego*, ignacio acosta, miguel ángel campano, and samuel domínguez-amarillo 120 figure 5. urban geometry and daylight autonomy map for the lower floor of the building (case study). figure 6. results for the case study iterations. figure 7. urban canyon proportion and spatial daylight autonomy correlation. urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 121 thus, the 4/5 urban canyon ratio allows obtaining adequate indoor solar access while the outdoor thermal radiation is also controlled. in addition, for this proportion, orientation is not a significant parameter. finally, it should be noted that, when the urban canyon ratio is equal to or lower than 4/5, this ratio becomes a more important parameter than orientation for the daylight access and outdoor temperature optimization process. however, orientation begins to be more relevant when the h/w ratio is lower than 4/5. thus, it must be taken into account that, with lower h/w ratios, the n-s orientation can cause a thermal imbalance in the different facades of the building, with the spaces located on the south facade suffering higher temperatures. in these cases, the use of vegetation on the street (depending on the type of vegetation, size, and density, as well deciduous or evergreen) can be another possible strategy to reduce overheating and direct solar radiation on buildings. it should be noted that these conclusions are only valid for the hypotheses under study, which are based on a specific case of a typical urban configuration located in the city of seville. in addition, these hypotheses have the initial limitation, or premise, of having all the urban blocks the same, with the same shape, height, orientation and distance between them. in addition, this case study requires a more in-depth analysis introducing other urban parameters, as the reflectance of the materials, shadow cast by vegetation, window-to-wall ratio, buildings coverage ratio and urban density. this is because this case study has been used to show the efficacy and functionality of the methodological proposal presented in this paper. 4. discussion nowadays, the promotion of energy savings and wellness in buildings is common, and the urban scale plays a fundamental role in achieving it. nevertheless, a larger-scale analysis involves a much more complex calculation. therefore, the purpose of this research was to propose an experimental methodology capable of performing a multi-objective analysis while reducing computational costs. thus, this study presents an experimental methodological proposal based on a parametric workflow through grasshopper and ladybug tools. this experimental method allows the integration of urban data sets with the different tools used in a parametric and dynamic workflow, allowing multicriteria consideration, for better design options. the novelty of this research also lies in: the use of sda daylight metric to indicate the daylight sufficiency of the spaces. the multi-objective analysis of contradictory or conflicting aspects (indoor daylight access and outdoor thermal comfort). the integration of uwg in the workflow, taking into account the climate effect produced by the urban environment (uhi). the use of grasshopper's octopus component, which allows programming evolutionary algorithms to optimize multiple object functions. its use is crucial to relate both urban parameters and performance indicators. in this way, this new methodological approach can contribute to performing more complex analyses (at the urban scale, with the possibility of introducing numerous parameters and variables) to achieve more solid studies. this is done in a faster and easier way due to the multi-objective optimization and the coupling of several simulation engines in a single workflow. the results obtained in the case study presented in this paper prove the method viability of such workflow. for this reason, the importance of this document is based not on the case study results, but the new methodological proposal potential. finally, it is worth highlighting the workflow power to find differences and similarities between the optimal values of design parameters in different urban configurations and climates; as well as to establish, locally or globally, design guidelines as support tools in urban design. 5. conclusion this research presents a methodological proposal based on a parametric workflow through grasshopper and ladybug tools. its purpose is to define the optimal values of urban parameters to improve both daylight access and outdoor thermal comfort. the results obtained in the case study presented in this paper prove the method viability of such a workflow to determine a set of environmental design criteria at an urban scale from a multi-criteria analysis. this can be further validated with real-world on-site tests. from the results of the case study, located in seville (spain), the 4/5 ratio of the urban canyon seems to be the optimal h/w proportion analyzed to achieve the greater thermal and daylight comfort values for the typical urban expansion configurations. on the other hand, placing buildings with a ne-sw/nw-se orientation with urban canyon ratios equal or lower to 4/5 shows a better average thermal performance in all the facades. the novelty of this research is based on the analysis of the proper use of daylight, simultaneously considering the two aspects in an integrated way. on the one hand, the effect of the urban environment on solar access by using sda dynamic lighting metric in an urban context; on the other, the analysis of the modification of the urban morphology on outdoor thermal radiation and uhi phenomenon. as a limitation of the research, it is necessary to have a deeper analysis introducing other urban urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 122 parameters such as material reflectance, the shadow cast by vegetation, window-to-wall ratio, building coverage ratio, or urban density. in addition, the integration of other performance indicators, such as other dynamic daylight metrics (ase,cda, dam, …), the dynamic outdoor thermal comfort metric otca (that assess the links between the outdoor environment and human well‐being), operational energy should be implemented. finally, the proposed methodology should be tested in other climate locations and urban configurations. conflict of interest the authors declare no conflict of interest. funding the results presented were funded by the agencia estatal de investigación (aei) and the ministerio de ciencia e innovación (mcin) of the government of spain through the research project dynalight: efficient design for biodynamic lighting to promote the circadian rhythm in shift work centres (ref bia201786997-r). the authors wish to express their thanks for all the technical and financial support provided. acknowledgment the authors thank the estatal de investigación (aei) and the ministerio de ciencia e innovación (mcin) of the government of spain. the authors are also grateful to all developers of the open-source engine: ladybug tools, colibri, octopus, including the teams of energyplus, openstudio, radiance and urban weather generator. references acosta, i., campano, m. á., & molina, j. f. (2016), window design in architecture: analysis of energy savings for lighting and visual comfort in residential spaces. applied energy, 168, 493–506. acosta, i., munoz, c., esquivias, p., moreno, d., & navarro, j. (2015), analysis of the accuracy of the sky component calculation in daylighting simulation programs. solar energy, 119, 54–67. alexander, p. j., mills, g., & fealy, r. (2015), using lcz data to run an urban energy balance model. urban climate, 13, 14–37. ali-toudert, f., & mayer, h. (2006), numerical study on the effects of aspect ratio and orientation of an urban street canyon on outdoor thermal comfort in hot and dry climate. building and environment, 41(2), 94–108. bellia, l., acosta, i., campano, m. á., & fragliasso, f. (2020), impact of daylight saving time on lighting energy consumption and on the biological clock for occupants in office buildings. solar energy, 211(september), 1347–1364. bueno, b., nakano, a., & norford, l. (2015), “urban weather generator: a method to predict neighborhood-specific urban temperatures for use in building energy simulations” in icuc9 9th international conference on urban climate jointly with 12th symposium on the urban environment, i. committee, i. e. s. d. m. (2012), ies spatial daylight autonomy (sda) and annual sunlight exposure (ase), daylight metrics committee. approved method ies lm-83-12. illuminating engineering society of north america. compagnon, r. (2004), solar and daylight availability in the urban fabric. energy and buildings, 36(4), 321–328. fundamentals, a. h. (2001). si edition, atlanta: american society of heating. refrigeration and air-conditioning engineers, inc. givoni, b. (1989), urban design in different climates. l’eclairage, c. i. de. (2003), spatial distribution of daylight–cie standard general sky. iso, 15469. lowry, w. p. (1977), empirical estimation of urban effects on climate: a problem analysis. journal of applied meteorology, 16, 129–135. naboni, e., natanian, j., brizzi, g., florio, p., chokhachian, a., galanos, t., & rastogi, p. (2019), a digital workflow to quantify regenerative urban design in the context of a changing climate. renewable and sustainable energy reviews, 113(june), 109255. nakano, a., bueno, b., norford, l., & reinhart, c. f. (2015), “urban weather generator – a novel workflow for integrating urban heat island effect within urban design process”. building simulation conference, 1901–1908. natanian, j., & auer, t. (2020), beyond nearly zero energy urban design: a holistic microclimatic energy and environmental quality evaluation workflow. sustainable cities and society, 56(february), 102094. núñez peiró, m., sánchez-guevara sánchez, c., & neila gonzález, f. j. (2017), update of the urban heat island of madrid and its influence on the building’s energy simulation. in sustainable development and renovation in architecture, urbanism and engineering (pp. 339–350). springer international publishing. observatori metropolità del canvi climàtic. (2015), la isla de calor en el área metropolitana de barcelona y la adaptación al cambio climático. 101. oke, t. r. (1988), street design and urban canopy layer climate. energy build, 11, perez, r., seals, r., michalsky, j., & sciences, a. (1993), all-weather model for sky luminance distribution--preliminary configuration and validation. 50, 235–245. ratti, c., baker, n., & steemers, k. (2005), energy urban layout for building solar optimization and outdoor comfort balance: neighborhood strategy in mediterranean climate 123 consumption and urban texture. energy and buildings, 37(7), 762–776. reinhart, c. f., & breton, p.-f. (2009), experimental validation of autodesk® 3ds max® design 2009 and daysim 3.0. leukos journal of illuminating engineering society of north america, 6(1), 7–35. reinhart, c. f., & walkenhorst, o. (2001), validation of dynamic radiance-based daylight simulations for a test office with external blinds. energy and buildings, 33(7), 683–697. roudsari, m. s., & pak, m. (2013), “ladybug: a parametric environmental plugin for grasshopper to help designers create an environmentallyconscious design”. proceedings of bs 2013: 13th conference of the international building performance simulation association, 3128–3135. salvati, a., coch roura, h., & cecere, c. (2016), urban heat island prediction in the mediterranean context: an evaluation as the urban weather generator model. architecture, city and environment, 11(32), 135–156. salvati, a., monti, p., coch roura, h., & cecere, c. (2019), climatic performance of urban textures: analysis tools for a mediterranean urban context. energy and buildings, 185, 162–179. street, m., reinhart, c., norford, l., & ochsendorf, j. (2013), “urban heat island in boston an evaluation of urban airtemperature models for predicting building energy use”. proceedings of bs 2013: 13th conference of the international building performance simulation association, 1022–1029. taleb, h., & musleh, m. a. (2015), applying urban parametric design optimisation processes to a hot climate: case study of the uae. sustainable cities and society, 14(1), 236–253. wong, i. l. (2017), a review of daylighting design and implementation in buildings. renewable and sustainable energy reviews, 74(july 2016), 959– 968. xue, p., mak, c. m., & cheung, h. d. (2014), the effects of daylighting and human behavior on luminous comfort in residential buildings: a questionnaire survey. building and environment, 81, 51–59. 1. introduction 1.1. background 1.2. aims and objectives 2. methodology 2.1. input parameters 2.2. microclimatic evaluation method 2.3. daylight evaluation method 2.4. multi-criteria analysis and optimization 3. methodological application in a case study 3.1. input parameters 3.2. case study results 3.3. analysis of results 4. discussion 5. conclusion conflict of interest funding acknowledgment references c:\users\raoh\desktop\paper_2.xps the journal of engineering research vol. 8 no 2 (2011) 10-18 1. introduction the study of solitary-wave dates from 1845 when john scott russell (svendssen 2006) reported his experiments. the discovery of the solitary wave have lead to many mathematical investigations that try to understand its properties. in particular, korteweg and ______________________________________ *corresponding author’s e-mail: ksmida@ksu.edu.sa de vries derived their well-known kdv equation for water waves propagating in one direction on shallow water (dean et al. 1991). they also found an exact solution of the kdv equation for a single wave propagating with a uniform velocity without changing formthe solitary wave. cfd analysis of water solitary wave reflection k. smida*a, h. lamloumib, z. hafsiab and k. maalelb adepartment of medical equipments technology, applied medical sciences college, majmaah university, p.o. box 66, al majmaah 11952, kingdom of saudi arabia blaboratoire de modélisation en hydraulique et environnement (lmhe). enit, b.p 37 le belvédère, 1002 tunis, tunisia received 4 april 2011; accepted 19 september 2011 abstract: a new numerical wave generation method is used to investigate the head-on collision of two solitary waves. the reflection at vertical wall of a solitary wave is also presented. the originality of this model, based on the navier-stokes equations, is the specification of an internal inlet velocity, defined as a source line within the computational domain for the generation of these non linear waves. this model was successfully implemented in the phoenics (parabolic hyperbolic or elliptic numerical integration code series) code. the collision of two counter-propagating solitary waves is similar to the interaction of a soliton with a vertical wall. this wave generation method allows the saving of considerable time for this collision process since the counter-propagating wave is generated directly without reflection at vertical wall. for the collision of two solitary waves, numerical results show that the run-up phenomenon can be well explained, the solution of the maximum wave run-up is almost equal to experimental measurement. the simulated wave profiles during the collision are in good agreement with experimental results. for the reflection at vertical wall, the spatial profiles of the wave at fixed instants show that this problem is equivalent to the collision process. keywords: wave generation, head-on collision, solitary waves, source line, wave run-up 11 k. smida, h. lamloumi, z. hafsia and k. maalel the problem of head-on collision of two solitary waves takes its origin from the study of the run-up of these waves on a vertical wall (tsunami case). the tsunamis are the most known for the several types of extreme waves (neelamani et al. 2009) and to simulate a tsunami wave, the best and simplest model is the solitary wave. the study of the dynamics of these waves helps greatly with the protection of coastal upland from erosion which is one of the challenging problems. different types of shore protection structures are in use around the world. indeed, in the absence of viscosity and in the symmetric case, both problems become equivalent. many studies have been conducted on this problem, analytically, numerically and experimentally, including the calculation of the maximum run-up (or superlinear amplitude increase) and the phase shifts. maxworthy (1976) conducted experiments on the collision of two solitary waves; where he showed that the maximum run-up reached a value higher than the sum of waves amplitudes before collision. su et al. (1980) and craig et al. (2006) presented analytical and numerical studies on the head-on collision of two solitary waves and they found that the collision was not elastic, and a small amount of energy was lost by the waves to form secondary waves. wu (1998) presented an analytical study for head-on and following collisions of solitary waves of unequal amplitudes. he distinguished three regimes: the single peak, the double peak and the critical regime. according to lubin et al. (2005), there is a lack of numerical simulation models of the head-on collision waves based on navier-stokes equations. the most known method for linear and non linear wave generation is the internal source region proposed by lin et al. (1999). however, the existing cfd code does not integrate any numerical wave method. power et al. (1984) studied the reflection of a solitary wave on a vertical wall by solving the boussinesq equations analytically as well as numerically. the analytical solution is obtained through an asymptotic matching technique, while the numerical solution is based on a finite difference scheme. they calculated the amplitude of the maximum run-up and the time at which it reached. they found that the incident wave is not reflected immediately as predicted by linear theory. in addition, they showed that the wave undergoes a phase delay during the collision. this phase was found to be inversely proportional to the square root of the initial wave amplitude. cooker et al. (1997) also studied the reflection of a solitary wave on a vertical wall. their numerical approach based on a boundary integral method is used to calculate the fluid potential flow described by the euler equations. they calculated the wall residence time, the time the wave crest remains attached to the wall, a concept introduced for the first time by (temperville 1979). they showed that the wall residence time provides an unambiguous characterization of the wave. they compared their results on the time of attachment and detachment of the wave crest with the asymptotic formulas of (su et al. 1980). further results on the flow were obtained, including the maximum run-up and the instantaneous forces exerted on the wall. their numerical results on the residence time are in accordance with measurements taken from a film on the reflection of a solitary wave on a vertical wall from the experiences of (maxworthy 1976). the proposed numerical wave method is used in the simulation of two-dimensional head-on collision of two solitary waves and the reflection of a solitary wave at vertical wall. it is based on the navier-stokes equations with additional transport scalar equation to describe the free surface evolution (the volume of fluid (vof) equation). the non linear solitary waves are generated by two source lines in internal flow region. at each source line, an internal time dependent inlet velocity is imposed according to the wave characteristics (hafsia et al. 2009). the numerical results are compared to experimental results reported by ming et al. (2003) and non linear wave analytical solution. numerical results given by power et al. (1984) are used to validate the reflection of solitary wave at vertical wall. 2. governing equations and boundary conditions 2.1 transport equations for unsteady flow and incompressible fluid, the mass conservation equation is written in cartesian coordinates: (1) with u and w are the velocity components respectively in the x and z directions. in order to avoid velocity damping in uniform horizontal flow, the damping force in horizontal direction is not considered. the momentum transport equation describing this velocity component is: (2) a dissipation zone is introduced in order to damp the wave at the open boundaries. within this region, a friction source term is added to the momentum trans12 cfd analysis of water solitary wave reflection port equation in the vertical direction: (3) with, p, the pressure; v, kinematic viscosity; , fluid density; g, gravitational acceleration and (x) is a damping function equals to zero except for the added dissipation zone. a linear damping law is adopted: (4) 2.2 free surface treatment the free surface displacements may be treated with a convective transport equation describing the fraction of flow (vof) in each cell of the computation domain: (5) the free surface profile is considered as a twophase flow involving water phase and air phase. we assume that the sliding between the two phases is negligible and that there is no mass exchange across the interface. hence, the velocity field at the free surface is continued. 2.3 initial and boundary conditions the initial condition considered is still water with no wave or current motion. the following boundary conditions are considered when solving the above mentioned transport equations: * for the free surface boundary condition, the normal stress is imposed by setting the pressure p equal to the atmospheric pressure patm (p = patm). * for open boundary condition, a dissipation zone is added in order to avoid wave reflection at each end. within such zone it is advantageous to consider, in addition to the damping friction force, a numerical dissipation by applying coarse grids in the dissipation zones. the neumann boundary condition is specified at the end of each dissipation zone. to generate numerically a given wave, based on the 2-d navier-stokes equations, the source region proposed by (lin et al. 1999) is transformed to source line method (fig. 1). since source region is reduced to a source line, the mass flux per unit time is imposed in terms of vertical velocity wi (t) at the bottom of the internal inlet. the mass source term in the transport equation is then considered as a time dependent inlet boundary condition. 2.4 the internal inlet velocity following to (hafsia et al. 2009), the wave is generated on horizontal source line with vertical pulsating velocity wi (t). the expression of wi (t) at the source line is given by the following time dependent boundary condition: (6) where ls is the length of the internal source line. using this transformation, the mass source region is modeled as an internal inlet imposed at the bottom of the control cell. the velocity wi (t) is upward or downward following that the mass is added wi (t) > 0) or subtracted (wi (t) < 0). the specified internal inlet velocity depends on the desired free surface wave profiles such as linear monochromatic wave or nonlinear solitary wave. assuming that the source line is located at xs = 0, the free surface elevation is function only of the time t. for solitary wave, we have: (7) (a) source region (b) source line wi ls figure 1. internal source for wave generation; (a) the source region method proposed by (lin et al. 1999); (b) the horizontal source line method proposed by (hafsia et al. 2009) 13 k. smida, h. lamloumi, z. hafsia and k. maalel h: wave height; and c the wave celerity given by: (8) the parameter k is given by: (9) the wavelength of a solitary wave is theoretically infinitely long. however, for practical purposes we can define an arbitrarily wavelength as: (10) the apparent wave period is defined by the following ratio: (11) the distance xn is introduced to make the free surface elevation negligible at the initial time and is given by: (12) 3. results and discussions 3.1 introduction two source lines were introduced within the computational domain to generate solitary waves propagating in constant depth (fig. 2). several analytical solutions are suggested in the literature for the head-on according to ming et al. (2003) experiments, two solitary waves were considered with different amplitudes of h1 = 5.70 10-3 m (left wave) and h2 = 5.43 10-3 m (right wave). the still water depth is d = 5.00 10-2 m. the horizontal distance between the two source lines for generating these waves is s12 = 2.00 m. in horizontal direction, the left source line is located at s1 = 6.03 m and the right source at s2 = 8.03 m. these two sources were located at the same elevation from the bottom ds = 0.6 d. the computational domain is discretized into uniform mesh in horizontal direction with x = 2.50 10-2 m. the minimum grid size in vertical direction are chosen near the free surface z = 4.03 10-4 m. the time step is t = 1.33 10-2 s. 3.2 waves generation in order to explain the solitary waves generation, fig. 3 illustrate the free surface profiles for different times. at t = 0 the free surface is horizontal due to that the internal inlet velocity is equal to zero. at t = 0.4 s (fig. 3a), a small free surface elevation is noted. indeed, the internal source perturbation became more influential (figs. 3b and 3c). at t = 0.71 s (fig. 3d), the two waves have remarkable shapes. after t = 0.52 s, the free surface reaches the maximum of amplitude for the two source lines, the potential energy is maximum (fig. 3e). after this instant, the solitary waves separation can be observed (figs. 3f to 3h). at t = 1.400 s, each source line will generate two solitary waves propagating in opposite directions denoted: w1l and w1r for the source s1 and w2l and w2r for the source s2. the generated four solitary waves (fig. 4) are in accordance with the analytical solutions (svendssen 2006). 3.3 head-on collision a sequence of spatial profiles during collision of the two counter-propagating solitary waves is shown in fig. 5. the free surface profiles between the times of t = 1.505 s and t = 1.806 s (figs 5a to 5d) represent the waves propagation before collision. the numerical and experimental results are in a good agreement except some observed discrepancies. for the same head-on collision configuration, based on euler equations, the differences between the predicted results given by (hammack et al. 2004) and experimental data are attributed to the viscous effects and the precision of experimental measurements. at t = 2.307 s, the two solitary waves propagating in opposite directions (w1r and w2l) are merged into a unique wave (fig. 5e). at this instant, the potential energy is maximum (the vertical velocity is negligible), the merging wave seems to be at rest and has a maximum amplitude higher than the sum of both incoming waves. numerical simulation show that the figure 2. computational domain and positions of the two source lines introduced to study the collision of two solitary waves at constant water depth 14 cfd analysis of water solitary wave reflection figure 3. free surface profiles of generated waves at different times: a) t = 0.40 s; b) t = 0.47 s; c) t = 0.52 s; d) t = 0.71 s; e) t = 0.92 s; f) t = 1.00 s; g) t = 1.08 s; h) t = 1.28 s 15 figure 4. comparison between analytical (ooo) and numerical free surface profiles (--) for the two solitary waves w1r and w2l before collision (t = 1.40 s) figure 5. comparison between experimental (ooo) during the head-on collision at different times: a) t = 1.505 s; b) t = 1.603 s; c) t = 1.705 s; d) t = 1.806 s; e) t = 2.307 s (maximum of run-up); f) t = 2.605 s; g) t = 3.004 s; h) t = 3.505 s 16 tary wave at vertical wall is investigated by the source line method. the wall is located at x = 0, the incident solitary wave located far away the wall, is assumed to propathe solitary wave considered has an amplitude of h = 2 10-2 m. the still water depth is h = 2 10-1 m. in horizontal direction, the source line is located at la = 5 m from the vertical wall. the source was located at an elevation from the bottom hs = 0.5h. the computational domain is discretized into uniform mesh in horizontal direction with x = 3.125 10-2 m. the minimum grid size in vertical direction are chosen near the free surface z = 1.1 10-3 m. the time step is t = 1.19 10-2 s. figure 8 represents the reflected wave profile at t = 6.88 s, the reflected wave retrieve its initial shape with small loss of amplitude. in fact, after reflection the solitary wave loses energy to a dispersive wave train and loses height, so ultimately the speed of the reflected wave is smaller than before collision. spatial profiles during reflection of the solitary wave at a vertical wall are shown in fig. 9. from t = 4.74 s to t = 5.17 s (figs. 9a to 9d), the wave runup due to the kinetic energy reduction is observed. once the wave crest reaches the wall, the reflection does not occur immediately as predicted by the classical linear wave theory. the wave crest remains at the wall for a certain time, called the phase delay, to complete the reflection process. during this time delay, the wave amplitude continues to increase. the maximum run-up which is twice greater than the initial wave amplitude is reached at t = 5.31 s (fig. 9e). from this instant to t = 6.02 s (figs. 9f to 9j), the wave amplitude starts to decrease and the wave begins to move away from the wall and propagates in the positive x direction. the numerical profiles agree fairly well with numerical free surfaces profiles given by (power et al. 1984). conclusions we have studied the head collision of two solitary waves with almost equal amplitudes by two internal mass source lines. each source line will generate two solitary waves. the comparison with analytical non linear wave solution shows that the solitary waves are accurately generated. between the two source lines, the two counter-propagating solitary waves are merged into a unique wave with an amplitude higher than the sum of both incoming waves. the navierstokes solution of the maximum wave run-up is almost equal to experimental measurements. after merging, the two separated solitary waves show small differences between their initial amplitude and the simulated one. these differences are attributed to secondary waves. since the system of two head-on colliding waves is equivalent to a single solitary wave hiths figure 6. computational domain and position of the source line introduced to study the figure 7. comparison between numerical results given by (power et al. 1984) (ooo) and free surface profiles ( ) before the reflection of a solitary wave at vertical wall; t = 4.02 s figure 8. comparison between numerical results given by (power et al. 1984) (ooo) and numerical free s urface profiles ( ) fter the reflection of a solitary wave at vertical wall; t = 6.88 s 17 k. smida, h. lamloumi, z. hafsia and k. maalel figure 9. comparison between numerical results given by (power et al. 1984) (ooo) and numerical free surface profiles ( during the reflection of a solitary wave at vertical wall at different times; a) t = 4.74 s; b) t = 4.88 s; c) t = 5.02 s; d) t = 5.17 s; e) t = 5.31 s; f) t = 5.45 s; g) t = 5.60 s; h) t = 5.74 s; i) t = 5.88 s; j) t = 6.02 s 18 cfd analysis of water solitary wave reflection ting a vertical wall, this problem is also treated in this study; using one source line for wave generation. compared to numerical results given by (power et al. 1984), the simulated results are considered satisfactory. as a perspective of this study, we can investigate the head-on and co-propagating collision of solitary waves with unequal amplitudes. also the superposition of other wave types can be treated by this source line generation method. references cooker mj, weidman pd, bale ds (1997), reflection of a high amplitude solitary wave at a vertical wall. j. fluid mech. 342:141-158. craig w, guyenne p, hammack j, hendersen d, sulem c (2006), solitary wave interactions. phys. fluids 18:1-25. dean rg, dalrymple ra (1991), water wave for engineers and scientists. world scientific. hafsia z, ben haj m, lamloumi h, maalel k (2009), internal inlet for wave generation and absorption treatment. j. of coastal engineering 56:951-959. hammack j, henderson d, guyenne p, yi m (2004), solitary-wave collisions. proceedings of the 23rd asme offshore mechanics and arctic engineering (a symposium to honor theodore yao-tsu wu), vancouver, canada. lin p, liu plf (1999), internal wave-maker for navier-stokes equations models. j. of waterway, port, coastal, and ocean engineering 125(4): 207-215. lo dc, young dl (2004), arbitrary lagrangian-eulerian finite element analysis of free surface flow using a velocity-vorticity formulation. j. of computational physics 195:175-201. lubin p, chanson h, glockner s (2010), large eddy simulation of turbulence generated by a weak breaking tidal bore. environ fluid mech. 10:587602. lubin p, vincent s, caltagirone jp (2005), on the navier-stokes equations simulation of the head-on collision between two surface solitary waves. c. r. mecanique 333:351-357. maxworthy t (1976), experiments on the collisions between solitary waves. j. fluid mech. 76:177185. ming yi, haink tu (2003), head-on collision of kdv solitons. http:math.psu.edu/dmh/solitons/. neelamani s, al-salem k, rakha k (2009), extreme gravity waves in the arabian gulf. j. of engineering research 6(1):21-36. power h, chwang at (1984), on reflection of a planar solitary wave at a vertical wall. wave motion 6:183-195 north holland. su cs, mirie rm (1980), on head-on collisions between two solitary waves. j. fluid mech. 98: 509-525. svendsen ia (2006), introduction to nearshore hydrodynamics. world scientific. temperville a (1979), interactions of solitary waves in shallow water theory. arch. mech. 31:177-184. wu ty (1998), nonlinear waves and solitons in water. physica d. 123:48-63. c:\users\raoh\desktop\mohamed_paper1.xps the journal of engineering research vol. 7 no. 2 (2010) 69-79 design of a realistic test simulator for a built-in self test environment a. ahmad* and d. al-abri * department of electrical and computer engineering, college of engineering, sultan qaboos university p. o. box 33, postal code 123; muscat, sultanate of oman received 14 may 2009; accepted 16 march 2010 abstract: this paper presents a realistic test approach suitable to design for testability (dft) and builtin self test (bist) environments. the approach is culminated in the form of a test simulator which is capable of providing a required goal of test for the system under test (sut). the simulator uses the approach of fault diagnostics with fault grading procedure to provide the tests. the tool is developed on a common pc platform and hence no special software is required. thereby, it is a low cost tool and hence economical. the tool is very much suitable for determining realistic test sequences for a targeted goal of testing for any sut. the developed tool incorporates a flexible graphical user interface (gui) procedure and can be operated without any special programming skill. the tool is debugged and tested with the results of many bench mark circuits. further, this developed tool can be utilized for educational purposes for many courses such as fault-tolerant computing, fault diagnosis, digital electronics, and safe reliable testable digital logic designs. keywords: digital system testing, built-in self test, design for testability, test vector, fault diagnosis fault collapsing, realistic test, fault cover, iteration _____________________________________ *corresponding author’s e-mail: afaq@squ.edu.om * 70 the journal of engineering research vol. 7 no. 2 (2010) 69-79 1. introduction over the years there has been a remarkable growth in very large scale integration (vlsi) techniques. tremendous advances in fabrication technologies have led to increased integrated circuit (ic) densities. the overall impact has been the decrease in cost per function of digital processing hardware and greater system reliability (ahmad, 2005). however, as a consequence of higher integration densities, circuits have become very complex. to verify their correct functioning, ics are required to be thoroughly tested. the cost involved in testing of even large scale integration (lsi) chips is a substantial portion of the total manufacturing cost (ahmad, 1989). further, advances in ic technology are occurring at a rate faster than those in test technology. a direct consequence of this is that testing methods, which are inadequate even for lsi circuits, are grossly unable to cope with the increased circuit sizes because of higher chip densities (abdi, et al. 2008; ahmad, 1989 and ahmad, 2005). formal design-for-testability (dft) techniques are concerned with providing access points for testing (ie. enhancing controllability and observability in ics) (al-lawati et al. 2004 and ahmad, et al. 2004). currently, dft and bist have become effective and widely acceptable tools for tackling test problems for vlsi chips and systems (zorian et al. 2000). this is because of the inherent advantages of dft and bist techniques (zorian et al. 2000). however, any technique chosen must be incorporated within the framework of a powerful computer aided designed (cad) system providing semi-automatic analysis and feed back, such that the rule of ten can be kept under control: if one does not find a failure at a particular stage, then detection at the next stage will cost 10 times as much! (ahmad 2005). testability is a measure to determine a desired degree of accuracy with which the functionality of any system or circuit or component can operate. a very high cost of test, and an uncertain level of delivered system quality are the indirect costs of non dft design testability. further, adding the time spent trying to diagnose the fault makes a non dft design very expensive. when testability is introduced at the design stage, it dramatically lowers the cost of test and the time spent at test. properly managed, testability enhances the assurance of product quality and smoothes production scheduling assuring high availability of service, and less maintenance cost. since testability is not a technological innovation, the preparedness of a mindset that motivates and creates a constant awareness of the importance of ease-oftesting at all levels of developments and use of the systems is essential. in real sense it can be said that testability is a very critical parameter to the manufacturing process of any system and a system that cannot be readily tested is not really to be considered as manufacturable engineering. further elongating the problem of testing is at present, in general most system designers and electronics engineers have little knowledge about testing, and thus the companies frequently hire test technology experts to guide their designers on test problems and the comnomenclature ate automatic test equipment bist built-in self test cad computer aided design cut circuit under test dft design for testability fc fault coverage fm fault matrix gnd ground ic integrated circuit pi primary inputs po primary outputs s-a-0 stuck at 0 s-a-1 stuck at 1 soc system on chip sut system under test gui graphical user interface lsi large scale integration vlsi very large scale integration 71 the journal of engineering research vol. 7 no. 2 (2010) 69-79 panies are even bound to pay higher remunerations to the test experts than to the total gross salaries of their designers. this reflects also today's university education and their designed curriculum. due to this weakness of designed curriculum, everyone learns about design, but only truly dedicated students learn about test. the next generation of engineers involved with design technology should be made aware of the importance of test and trained in test technology to enable them to produce high quality and defect-free products. this paper is an alternate effort for presenting an orientation of a conceptual teaching methodology to bridge the gap between the designers and test engineers. many researchers have put effort to this direction (ahmad and al-abri, 2005; ahmad et al. 2006 and ali et al. 2005), (blyzniuk et al. 1998; dariusz, 1998 and devadze et al. 2002), ivask et al. 1998; jutman et al. 2002; liu et al. 2008; su et al. 2008; ubar 1998; ubar et al. 2002; ubar and orasson, 2003 and ubar and wuttke, 2001). since automatic test equipments (ates) are costly and not affordable to many organizations, as viable option the process of simulation is usually adapted. this gap in the market has called for this prototype testers and simulated tools (ahmad and al-abri, 2009; ahmad et al. 2006 and ali et al. 2005), (liu et al. 2008 and su et al. 2008). however, such developed tools cannot cope with exhaustive tests because the application of 2n test vectors to a device with n inputs is not effective if n is large. as the number of tests, 2n grows exponentially with n, the number of tests required increases rapidly. for n inputs the truth table contains 2n rows. if a test application is at 1 test/ s, this would take 18 min for n = 30, 13 days for n = 40, and 36 years for n = 50. thus it is quite obvious that the method of exhaustive testing is unacceptable for testing. we developed a test tool which applies an alternative approach based on the observation that in a sut, any particular input test vector will usually cover a significant number of faults (ahmad and al-abri, 2009). any random selection of tests of reasonable size therefore can be expected to achieve reasonable fault coverage. hence this paper advocating for creating specific tests for faults most likely to occur. also, none of the researched test simulator has been used the most commonly available platform of excel. we chose excel due to portability, economy, wide storage space and its easy interface with other programming languages like c++, matlab, visual basic and java. we have already demonstrated its webct applicability through our research work (ahmad et al. 2006). this paper is further extension to our work by incorporating test pattern generation and fault simulation algorithms in cost effective manner. the features included in this tool are described in section 3 of this paper. hence, our developed test tool through this paper is very much helpful for both the electronic embedded industries along with the educational institutions. the presented paper considers how to improve the skills of students to be educated for hardware, embedded and system on chip (soc) design in test related topics. although, here in this paper, our discussion will be restricted on only one component of the tool but the tool has multifaceted usage in different courses with different usages like digital system simulation, fault insertion, test sequence generation, fault matrix generation, fault coverage, fault diagnostic tree generation, predicting targeted goal of test, to find hard faults, fault collapsing sets, fault equivalence sets, test generation and many more topics of interest in various subjects. besides the section of introduction, the structure of the paper consists of terminologies, basic concepts of testing and assumptions in section 2. section 3 is devoted towards the development of the test tool, its functionality and software architecture of the tool. simulation experiment and analysis are embodied in sections 4 and 5 respectively. the conclusions and future research highlights are presented in the section 6. 2. basic concepts and acquitance of terminologies definition 1: fault is defined as a physical failure mechanism due to some defects of the circuit. definition 2: error is defined as the condition (or state) of a system containing a fault (ie. that the system deviation from the correct state). in general, the physical defects are modeled by using faults. in particular, logic-level models, such as stuck-at, are adapted to measure circuit testability. functional verification based on simulation aims at detecting faults of circuits. definition 3: circuit simulation is a process of defining the digital circuit into its primitive gate level structure. definition 4: fault insertion is a process where every possible fault is simulated into the sut. a simulated fault is said to be injected into the circuit of the system. definition 5: fault detection is a process of comparing the response of a known good version of the circuit to that of the actual circuit existed in an environment of fault, 72 the journal of engineering research vol. 7 no. 2 (2010) 69-79 for a given stimulus set (test). a fault is detectable by the test if there exist any difference between fault free and faulty response. the process is repeated for each stimulus set and the fault matrix (fm) can be generated. definition 6: the stuck-at (s-a) fault model is used in functional testing. this fault model uses the single s-a model, the most common fault model used in fault simulation because of its effectiveness in finding many common types defects. stuck-at fault model behavior occurs if the terminals of a gate are stuck at either a high [logic 1 or vcc voltage] or a low [logic 0 or gnd voltage]. if it is s-a at one, it is denoted by s-a-1 whereas, if it is s-a zero, it is denoted by s-a-0. the fault sites for this fault model include the pins of primitive instances. example 1: figure 1 shows the possible s-a faults that could occur on a single and gate. there are total 6 faults. the 'a' and 'b' are known as primary inputs (pi), whereas, 'c' is known as primary outputs (po). since pi = 2, po = 1 thus, the size of fm = 4x7. in general, the size of fm can be defined as in eq. (1) below. size of fm = 2n * (2 * l + 1) * m (1) where n is the number of pi, m is the number of po and l is the total connecting lines in the sut. the fm for fig. 1 can be given as: definition 7: fault coverage (fc) is defined as the percentage of faults detected from among all faults that the test pattern set tests. the table 1 below shows the fc capability of each test. it is also to be observed that the different faults can be tested via different tests. table 2 shows that how faults are covered by different tests. figure 1. 2 inputs and gate table 1. fc capability table 2. faults detected by different tests 73 the journal of engineering research vol. 7 no. 2 (2010) 69-79 definition 8: an un-testable or redundant fault is a fault for which no test can exist. usually, un-testable faults cannot cause functional failures, so the testing tools exclude them when calculating test coverage. example 2: as an example consider the sut of fig. 2. in this circuit, signal g always has the value of 1, no matter what the values of a, b, and c. if d is s-a-1, this fault is undetectable because the value of g can never 2.1 assumptions some of the assumptions which are considered by the test techniques to test the systems are: a) stuck-at fault model is considered in the system. b) in the systems the gate delays are assumed to be zero. c) systems to be tested are limited to combinational circuits only. d) un-testable / redundant faults are isolated from fault cover of the system. e) binary up counter is used as test sequences to test the system. f) single fault model is considered in the system. 3. the developed tool the developed tool named as "digital tester1 v7 (2008)" is an updated version of the software introduced through the reported research work (ahmad and al-abri, 2009 and ahmad et al. 2006). the concept of fault equivalence and collapsing is adapted in the development of the tool. as the concept says that two faults of a sut circuit are called equivalent if they have exactly the same set of tests and transform the circuit such that the two faulty circuits have identical output functions. as an example consider a two input and gate with input lines a, b and output c. the stuck-at-0 faults on any input line (line a or line b) of this and gate can lead to a zero output which is the same faulty result of the stuck-at-zero fault on this gate's output line c, and these three faults have the same test vector (1, 1) therefore, for a two input and gate, the two input stuck-at0 and one output stuck-at-0 faults are equivalent. because of the in-distinguish-ability of equivalent faults, only one of them needs to be tested. using the conception of fault equivalence, we can collapse or delete most of equivalent faults. this procedure is called equivalence fault collapsing, which partitions a circuit into disjoined equivalence sets, chooses one fault from each equivalent sets, and forms an equivalent collapsed fault set. for example, because of the equivalence of the three stuck-at-0 faults in case of 2-input and gate, two of them can be collapsed. these concepts speeds-up the process of testing. the current tool is now equipped with many more features like: ==> tracing fault equivalence set, ==> finding fault collapsing set, ==> incorporation of a faster fault insertion mechanism, ==> describing structure models in easy manner, ==> best search for the targeted test goal, and ==> realistic test search. 3.1 description of basic functions figure 3 shows the first working window which is dedicated to receive general information about the numbers of inputs; structure levels and number of iterations. as soon as entries are complete it will generate the levels and will ask about the number of the gates in each of the levels which can be viewed in the pictorial example (example 3, fig. 4). according to the information entries the structure model can be entered by clicking the "generate structure". this activated window "structure" will only ask the nature of gate logic and the inputs to this gate in prescribed levels of the structure model. it also provides the provision of entering only a one start test sequence. as the description of the sut structure model is complete, the fm can be generated by clicking the "generate fault matrix". a quick fault insertion procedure is adapted in generating fm. all tests or partial tests can be opted during the generation of the fm procedure. the window can be visualized by activating "fault matrix" from the task bar. the fm also provides about the testability of each test along with the information of fault collapsing and equivalence. figure 2. an example of redundant fault 74 the journal of engineering research vol. 7 no. 2 (2010) 69-79 75 the journal of engineering research vol. 7 no. 2 (2010) 69-79 viewed by activating "iterations" from the task bar. finally, the testability results can be outputted by clicking the "opt testing" from the task bar. example 3 a 3-input and 3-level structure of digital system is considered with targeted goal of maximum 6 iterations. the overall architecture of the developed software tool is described in the ensuing section (section 3.2). 3.2 the architecture of developed software model the architecture of the software is based upon two basic principles; user input and backend computation. each user input triggers the actions like generation of circuit structure and generation of fault matrix. figure 5 depicts these actions and user interface data. the portion of architectural design from fault matrix onwards is shown in fig. 6 where the programmed algorithmic test generation is adapted to generate each iteration output. based on user requirement the iteration output presents the statistical summary of the test. each of the steps requires programs based on knowledge of system simulation, fault insertion, incorporation of test generation algorithms, fault collapsing, fault equivalence, optimization and iteration process. a sample of the backend computation for fault matrix generation is briefly given in fig. 7. 4. simulation experiments there are many standard test bench circuits of iscas85 (hansen et al. 1999) are simulated using this developed software. it is complex to present the logical diagram of those circuits. here we present below a case study to make it more readable and beneficial for researchers for the usefulness of the tool for their research works. 4.1 simulation run to demonstrate the simulation run, a circuit shown in fig. 8 is considered which has 3 inputs, 1 output, 3 levels of model structure, 4 logic gates, and 16 possible faults. after defining the structure model (via "generate structure"), a fm is generated (via "generate fault matrix") which is shown in fig. 9. by activating the "generate test sequence" with required goal of maximum six iterations, the result is outputted as shown in fig. 10. the results of each of the iterations can also be seen just by clicking iterations from the task bar. 5. analysis of results table 3 shows the maximum possible test vectors {ti} for i = 1 to 8 with respect to primary inputs of circuit shown in fig. 8. figure 11 demonstrate the fault detect-ability (fault coverage) of each tests obtained via the fault matrix result shown in fig. 9. as it can be seen from the results of fig. 10 that there requires only 3 tests to achieve the 100% of the fault coverage. the first test selected t7 {x3=1, x2=1, x1=0} which provides 43.75% of the fc. this is the first iteration on the basis of the maximum element, conflict resolution and the best intersection of the sets. on the same way of search of algorithmic procedure t5 {x3=1, x2=0, x1=0} is selected which enhances the fc to 81.25%. finally, t2 {x3=0, x2=0, x1=1} is selected to reach the target of 100% fc. as it can be visualized from figure 12 that although the test time and cost was allocated for a maximum of six iterations (user's set test cost) however, due to the algorithmic design it ends at 3rd iteration process (optimum test cost) and achieves the target of 100% fault coverage. this means even saving the 50% of the provision allocated for the test. further, it can be visualized that only 37.25% of test requirement is needed to achieve the goal. thereby, the designed algorithmic procedure saves the 62.75% of the test cost as well as the test time in comparison to the exhaustive test approach of 100% test cost (maximum test cost). conclusions the developed tool considers all effective points of test procedures, like target, fault collapse, given constraints, fault equivalence, best intersection sets and enumerating the data for each steps and hence very useful. the developed tool is dynamic in nature since it keeps in mind about the pace of the electronic industries. and, therefore, we made a provision of applying 14 inputs; accommodating 63 levels of structured layers of digital systems where as in each of the levels 10000 gates can be defined. the nature of these gates may be of any kinds logic gates such as: not, or, and, nor, nand, exclusive or, and exclusive nor. the provisions of the latches and flip-flops are also incorporated in the developed tool. the developed tool is user friendly. these powerful attributes of this developed tool and as well its enormous applications makes this an excellent learning tool for the courses in the area of digital system testing in an elearning environment along with the usage of electronic industries for incorporating the test procedures in the embedded systems. since the dft and bist environments incorporate the test circuitry in its embedded design therefore, this developed tool will be very much cost effective measure in designing the digital circuits in dft and bist environments. we developed the software tool in excel 2007 which is limited to 2 gigabytes of memory for the excel process under windows xp. this 2 gigabyte limit is a 76 the journal of engineering research vol. 7 no. 2 (2010) 69-79 figure 6. the program outputs figure 7. the beckend computation algorithm for fault matrix generation 77 the journal of engineering research vol. 7 no. 2 (2010) 69-79 limit on the virtual memory address space. virtual memory used by a process is larger than the working set memory reported by windows task manager, so the amount of useable memory under excel 2007 is considerably a little less than twice that of excel 2003. increased spreadsheet row and column capacity of one million rows by 16,000 columns enables the user to import and work with massive amounts of data and figure 8. a model for demonstration figure 9. the fm for the model of figure 5 figure 10. test results 78 the journal of engineering research vol. 7 no. 2 (2010) 69-79 achieve faster calculation performance. therefore, the developed tool can cope with large complex sut in efficient manner. our further orientation of the research work is to extend the capability of the tool to compute the reliability, controllability, observability, availability and maintenance scheduling of digital systems. acknowledgments the acknowledgments are due to sultan qaboos university (sultanate of oman) for providing research support grant (squ-dvc/ psr/raid/2010/23). references abdi, a., tahoori m. b. and emamian, e. s., 2008, "fault diagnosis engineering of digital circuits can identify vulnerable molecules in complex cellular pathways", sci. signal., vol. 1(42), pp 1 10. ahmad, a., 1989, "on a design approach for figure 11. % of fault coverage of each test figure 12. test grading table 3. possible test vectors 79 the journal of engineering research vol. 7 no. 2 (2010) 69-79 reducing aliasing errors and achieving higher testability goals in combinational circuits", indian institute of technology, roorkee ph. d. thesis. ahmad, a., 2005, "testing of complex integrated circuits (ics) the bottlenecks and solutions", asian journal of information technology, vol. 4(9), pp. 816 822. ahmad, a., al-abri, d., 2009, "design of dynamic test tool in the area of digital system testing", presented and published in the proceedings of international conference on computer, communication and power (icccp'09) held at oman, pp. 1-4 (computer and network systems). ahmad, a., al-abri, d. and al-ramhi, m. m., 2006, "design of an e-learning process in the area of digital system testing", presented and published in the proceedings of international conference on distance education (icode2006), held at muscat, sultanate of oman, march 27 31, 2006, pp 1-8. ali, l. sidek, r., ishak, a. alauddin, m. a., and bambang, s. s., 2005, "design of a low cost ic tester", american journal of applied science, vo. 2(4), pp. 824 827. al-lawati, a. m. j. and ahmad, a., 2004, "realization of a simplified controllability computation procedure a matlab-simulink based tool", sultan qaboos university journal for scientific research science and technology, oman, vol. 8, pp. 131 143. ahmad, a., al-lawati, a. m. j. and al-naamany, a. m., 2004, "identification of test point insertion location via comprehensive knowledge of digital system's nodal controllability through a simulated tool", asian journal of information technology, vol. 3(3), pp. 142 147. blyzniuk, m., cibakova, t., gramatova, e., kuzmicz, w., lobur, m., pleskacz, w., raik, j. and ubar, r., 2000, "hierarchical defect-oriented fault simulation for digital circuits", ieee european test workshop, cascais, portugal, pp. 151-156. dariusz, b., 1998, "how faults can be simulated in self-testable vlsi digital circuits?", proceedings euromicro'98, vol. 1, pp.10180. devadze, s., jutman, a., sudnitsyn, a., ubar, r., wuttke, h. d., 2002, "teaching digital rt-level self-test using a java applet", 20th ieee conference norchip'2002, copenhagen, denmark, pp. 322-328. hansen, m. c., yalcin, h. and hayes, j. p., 1999, "unveiling the iscas-85 benchmarks: a case study in reverse engineering", ieee design & test, vol. 16(3), pp. 72 80. ivask, e., raik, j., ubar, r., 1998, "comparison of genetic and random techniques for test pattern generation", proceedings of the 6th baltic electronics conference, tallinn, pp. 163-16. jutman, a., raik, j., ubar, r., 2002, "ssbdds: advantageous model and efficient algorithms for digital circuit modeling, simulation & test", proceedings of 5th international workshop on boolean problems (iwsbp'02), freiberg, germany, pp. 157-166. liu, y., su, w. and fan, t., 2008, "the digital circuit fault diagnosis interface design and realization based on vxi", proceedings isecs international colloquium on computing, communication, control, and management 2008 (cccm '08), guangzhou, vol. 2, pp. 215 219. su, w., zhang, s., xue, l., 2008, "research and realization of digital circuit fault probe location process", proceedings international conference on intelligent computation technology and automation 2008 (icicta'08), hunan, vo. 2, pp. 897 900. ubar, r., 1998, "dynamic analysis of digital circuits with multi-valued simulation", microelectronics journal, elsevier science ltd., vol. 29(11), pp. 821-826. ubar, r., jutman, a., orasson, e., raik, j., wuttke, h.d., 2002, "internet-based software for teaching test of digital circuits", microelectronics education, marcombo u boixareu ed., pp. 317 320. ubar, r., orasson, e., 2003, "e-learning tool and exercises for teaching digital test", proceedings of 2nd ieee conference on signals, systems, decision and information technology, sousse, tunisia, vol. cit-6, pp.1 6. ubar, r., wuttke, h.d., 2001, "the dildis-project using applets for more demonstrative lectures in digital systems design and test", proceedings of 31st asee/ieee frontiers in education conference, reno, nv, usa, abstracts, p. 83. zorian, y., mourad, s., 2000, "principles of testing electronic systems", j. wiley & sons, inc. new york, 2000, p. 420. c:\users\raoh\desktop\paper_7.xps tjer 2012, vol. 9, no 1, 55-63 1. introduction recently, the oil and gas industry has witnessed a steady increase in the use of short-radius horizontal drilling operations to enhance well productivity. the trend is expected to continue in the future as oil and gas industries become increasingly aware of the bene fits associated with it while targeting thin reservoirs and oil reserves which are pocketed and scattered. cost savings and simplicity are the major driveing _________________________________________ *corresponding author’s e-mail: tasneem@squ.edu.om factors in developing the technology and tools to drill sidetracks of short or ultra-short radius. furthermore, the rapid decrease in active reserves, the ever increasing difficulty of exploratory development of the matured fields and the transfer of exploration targets from conventional reservoirs to unconventional ones with low permeability, low pressure and low productivity pose new challenges to the drilling technique dynamic analysis and vibration of beam inside annulus for ultra short-radius water jet drilling t pervez*a, sa al-hiddabia, a al-yahmadia and ac seibib *a department of mechanical and industrial engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, al-khoud, muscat, oman b petroleum institute, abu-dhabi, uae received 3 october 2010; accepted 17 may 2011 abstract: conventional water-jet nozzle systems have been developed and partially used in the oil and gas industry to drill horizontal sidetracks. however, this technique still presents a few shortcomings associated with tube buckling and water jet sagging. due to these problems, the drilled hole deviates from the desired path and does not reach the target reservoir. the issue becomes more complex due to the continuously moving boundaries representing the borehole profile, which is, in turn, governed by the nozzle dynamics. a mathematical model representing the dynamics of water jet drilling confined in a borehole along with drilling mud is developed to predict the sagging phenomenon during the drilling process. the closed form solution of the governing equation is obtained for horizontal drilling in shallow formation layers. the solution shows the strong influence of nozzle vibration and the magnitude of thrust force at the nozzle tip on the profile and the diameter of drilled hole. for sidetrack drilling of greater than 400 m length, the magnitude of sagging is large enough to miss the target reservoir. furthermore, the drill string buckles at certain magnitudes of thrust forces and penetration lengths. keywords: ultra short radius, water-jet drilling, dynamics and control 56 t pervez, sa al-hiddabi, a al-yahmadi and ac seibi and technology (zhu 2009). it is often much cheaper to initiate a lateral or sidetrack from an existing wellbore to reach small pockets of reserves rather than drilling a new well. recent use of short-radius drilling has resulted in an enhanced production of depleted fields in south asia. the operator of a mature offshore oil field has significant production increase, when over 100 short-radius wells have been drilled as a re-development project to access bypassed reserves (halliburton 2009). horizontal and multilateral wells are the direct results of the advances in directional drilling which have allowed operators to select the most appropriate well profile, geometry and entry points. these wells have an edge over the conventional vertical wells in terms of ultimate recovery in many marginal wells. the productivity of horizontal and multilateral wells can be estimated using the semi-analytical model developed by (tabatabaei, ghalambor 2009). it is believed that these multilaterals will perform better in thin reservoirs, double porosity and discreetly fractured reservoirs, reservoirs with anisotropic vertical permeability and in various enhanced oil recovery concepts. however, it is more expensive to drill and complete such a well. hence, there is a need to develop an economical, reliable and simple technology to drill sidetracks. current technology for drilling provides a number of systems to generate short-radius drilling such as the "conventional" coil-tubing drilling system (kolle, marvin 2000), the flexible pipe rotary steerable system (laprade 2000), the water jet drilling system (dickinson et al. 1990) and the coiled tubing conveyed drilling technique (buset et al. 2001). the goal is to improve the production profile around the mother well by penetrating the thin reservoirs and hydrocarbon pockets behind or surrounding the actual reservoir. for horizontal sidetracks of ultra short-radius ie., a radius of curvature smaller than 4 meters or less, the water jet drilling system seems to be a viable option (kolle 1998). in the last two decades, the conventional water-jet nozzle system has been developed and used to cut, machine, and drill or fracture several target materials such as metals, ceramics, and rocks (fair 1998). the system worked satisfactorily for metals and ceramics as it has received the proper attention of researchers and technology developers. however, a few shortcomings need to be overcome before the oil and gas industry will accept this as an alternative in directional drilling of sort and ultra short-radius sidetracks. the two major challenges associated with the water jet drilling system are tube buckling and the sagging of the water jet which results in a deviation from the desired path. it has been observed that within few tens of meters, the tubing moves downward through the formation due to sagging (stoner 1999). due to the tube sagging, drilling in a thin oil-sand layer becomes problematic resulting in missing the desired target. therefore, it is important to predict the path a of drilled hole in order to achieve effective and efficient use of the water jet drilling system before undertaking any drilling activities. a better understanding of the sagging phenomenon is crucial for designing such a drilling system with an embedded directional control to minimize the deviation between the drilled hole and the desired target. the current work focuses on studying the dynamics of a water jet drilling system during horizontal drilling. the developed model is able to predict the profile of a drilled hole and identifies the important parameters which contribute to the sagging phenomenon. 2. problem description the water jet drilling system under study consists of a tube with a nozzle attached at its end. the tube is pushed axially to set it initially at a prescribed location to start drilling the hole. the water jet shooting out of the nozzle at a very high velocity does the drilling. this results in a hole with a diameter greater than the tube diameter which makes it easy for the tube to penetrate through the formation until the desired length of the drilled hole is attained. then the tube is clamped at the kick-off point from where it can be pushed forward and can drill the hole, as shown in fig. 1 while the nozzle end is free to move. this allows the tube and nozzle system to be modeled as a cantilever beam. the tube geometry is uniform and is made of homogenous and isotropic material. the borehole geometry will depend on the lateral displacement and velocity of the nozzle. it is assumed that, during drilling operation, there is some clearance between the hole and the tube diameter which will restrict the deformation of the drilling tube within elastic region. in addition, the lateral vibration of the tube will be constrained by the borehole. the axial dynamics of the system was modeled as a quasi-static problem. due to these effects, the developed mathematical model is not only complex but also nonlinear in nature. in addition the system is modeled as a continuous medium, resulting in partial differential equations describing the lateral dynamics of the jetting system. 2.1 governing equation the governing equations of the tube are derived by considering an infinitesimal section of the tube, as 57 dynamic analysis and vibration of beam inside annulus for ultra short-radius water jet drilling shown in fig. 1, with the surrounding hole at a distance 'x' from the left end, which represents the starting point of drilling. using the force equilibrium and moment equilibrium, the equation of motions for the infinitesimal tube element can be written as: (1) (2) where m(x,t) is the bending moment, v(x,t) is the shear force, f(x t) is the external lateral force per unit tube length, p(x,t) is the axial force, is the inclination angle of tube center line, x is the axial displacement and w is the lateral deflection of the tube. furthermore, = t t + f f where and are the mass density and cross-sectional area with subscripts t and f refers to the tube and fluid respectively. small deflections, written as: (3) using the elementary beam bending theory, in combination with eqs. (1) through (3), yields a single partial differential equation of motion for the system: (4) where, e is the young's modulus of elasticity and i is the moment of inertia. due to very low velocity of the fluid in the tube, its effect is considered negligible. the external forces acting on the tube can be expressed as: ƒ (w, x, t) = fg + fd + fi (5) fg, fd and fi represent gravity, damping and impact forces, respectively. the impact force originates from the contact of the tube with the borehole. the buoyancy force is the same magnitude as the fluid weight inside the tube, hence the gravity force is equal to the weight of the tube. the damping force is approximated using the following equation (jansen 1993): fi = k | w cu | if w > cu or = k | w cl | if w < cl or = 0 otherwise (7) figure 1. free body diagram of an infinitesimal element of drilling tube (6) 58 t pervez, sa al-hiddabi, a al-yahmadi and ac seibi where, k, cu and cl denote the formation stiffness (constant), of the upper and lower hole clearances, respectively. equation (4) in conjunction with eqs. (5) to (7) becomes a nonlinear partial differential equation. the solution of the governing equation requires initial and boundary conditions, which are given below: 3. numerical solution the governing equation describing the dynamics of tube sagging along with the boundary and initial conditions is solved numerically using method of lines (schiesser 1991) with a uniform x-grid and finite difference approximations for x-derivatives. the spatial discretization reduces the partial differential equation into a set of time dependent nonlinear ordinary differential equations. the number of equations increases as the tube penetrates into the formation. the resulting set of differential equations is stiff in nature and a special solver is used to obtain the solution. let xo,x1,x2,...,xn represent the nodal points along x-axis, where xo and xn are the nodal points at the clamped and free end of the tube. using backward and central difference schemes to approximate derivatives at the fixed end and all other nodal points in spatial discretization, the following can be written: where subscripts '0' and '-1' refer to the node at the clamped end and the node preceding the clamped end. use of the other two boundary conditions of eq. (8) gives: which yields wn+1 = 2wn wn-1 and wn+2 = 4wn wn-1 + wn-2. similarly, for node 'i', one can be written substituting the above equations in eq. (4) gives: letting 'y' represents thevector containing the deflection and its time derivative at 'xi', where i = 1,2, ...n, and 'n' is the total number of discretization points ie. (10) using eq. (10), a matrix representation of the resulting set of nonlinear ordinary differential equations is used to obtain the solution: (11) due to the fact that the problem is “stiff”, an implicit method is used to obtain the solution. the code is written in matlab software, which starts with one node point representing the free and as it penetrates the formation. the simulation starts with zero slope and deflection at the free end as an initial condition, ie., y(1) (t = 0) = [ 0 0 ]t. the solution progresses in time to determine the deflection and slope vector at node ‘i’: (12) assuming a constant rate of penetration, the tube penetrates distance x into the formation in a quasistatic mode. to proceed further with drilling, another node point is added, resulting in an increase in the number of equation by two (one each for deflection and slope). a new initial 'y' vector is generated using previous initial condition and the current solution at node '1'. the algorithm continues until required penetration is reached. 4. results and discussions the mathematical model developed for water jet drilling systems is simulated in matlab software based on the approach described in the previous section. in particular, the study focuses on predicting the tube sagging which results when the water jet deviates from its desired horizontal trajectory. the drilling (8) and and which yields (9) 59 dynamic analysis and vibration of beam inside annulus for ultra short-radius water jet drilling fluid is water with a density of 1000 kg/m3. the outer diameter and thickness of the tube are 31.75 mm and 3mm, respectively. the tube material is steel with e=200 gpa and = 7830 kg/m3. the initial clearance between tube and borehole is 5 mm. the clearance is a function of the formation properties and nozzle design parameters such as the water jet diameter and the twist that is used to spin the drilling fluid inside the nozzle. figure 2 shows the profile(with upper and lower boundaries) of the drilled hole obtained from simulation in absence of thrust force (p=0 kn). it is obvious from the figure that drilling starts horizontally but later sinks down due to the sagging phenomenon. as soon as the tube centerline deflects from the horizontal reference and transients are diminished, and the drilled hole significantly deviates from the desired borehole trajectory. it is also clear from the figure that the drilled hole profile is not smooth due to cavitation of the nozzle. the lateral vibration of the nozzle is the major of cause of this behavior. to ascertain this, the lateral movement of tube centerline at three specific nodes is plotted in fig. 3. these nodes are chosen near the clamped, middle and nozzle ends. the mean position of clamped end node remains horizontal indicating negligible lateral motion. contrary to this, the middle and end nodes have significant lateral motion exceeding tube-hole clearance resulting in frequent impact with lower boundary of the drilled hole. however due to sagging phenomenon, the lateral movement of the drilling tube is considerably reduced until it stays in constant contact with lower boundary of the drilled hole. for systems, where the drilling tube is kept under tension due to the pressure difference across the nozzle, the thrust force is not negligible. in order to see the effect of thrust force on the profile of the drilled hole, simulations were carried out for 10 kn and 20 kn of thrust force as shown in figs. 4(a) and 4(b) respectively. as the magnitude of thrust force increased, the profile become more irregular. in case of 20 kn thrust force (fig. 4(b)), the borehole profile was highly irregular, which is unacceptable from an application point of view. in order to alleviate this problem, a closed loop controller must be designed to assure that the drilled borehole is straight and does not sag or have a highly irregular profile. the following paragraphs discuss the design and implementation of a control system for the drilling system. in order to control the system, nonlinear inverse control method is used for eq. (4) to design a controller to minimize the sagging problem. for this purpose, a control force u(t), is applied at the nozzle end. the addition of the control force modifies the boundary condition i.e., the shearing force must be equal to the control force u(t). also, the equation of motion for the last node 'n' modifies to: (13) the control force, u(t), is chosen to cancel the undesired terms in the above equation and a new control variable v(t) is introduced to have the desired dynamics for node 'n' i.e.,: where a = [a1 a2]t. matrix a2 is given below a1 = [ (o)nxn ] with 'o' and 'i' 12(a) 12(b) 60 t pervez, sa al-hiddabi, a al-yahmadi and ac seibi horizontal water jet drilling, path prediction horizontal reference upper boundary of the hole sagging deflection tube centerline lower boundary of the hole penetration (m) x 103 oscillation of three different points on the drillstring showing impact with hole upper boundary of the hole start of the tube (0.2 m) middle of the tube (5 m) near the nozzle (10 m) lower boundary of the hole time (sec) figure 3. impact of the drilling tube (start, middle and near nozzle end) with the hole boundary 61 dynamic analysis and vibration of beam inside annulus for ultra short-radius water jet drilling (14) where constants k1 and k2, termed as controller gains, must be selected so that the characteristic equation has roots with negative real parts. moreover, the values of k1 and k2 can be chosen to have the desired dynamics for the last node. for full compensation, u(t) is selected as: figure 5 shows the profile of drilled hole using full compensation control having gains of k1 = 871 and k2 = 1742.5. the settling time of the water-jet drilling water jet drilling: 10 kn thrust force horizontal reference upper boundary of the hole tube centerline under buckling lower boundary of the hole penetrawtion (m) water jet drilling 20 kn horizontal reference upper boundary of the hole tube centerline under buckling lower boundary of the hole penetration (m) figure 4. effect of thrust force on the profile of drilled hole and tube buckling at (a) p = 10 kn, (b) p = 20 kn (15) 62 t pervez, sa al-hiddabi, a al-yahmadi and ac seibi system with full compensation controller is 0.2 seconds. it is clear from the figure that the deviation in the drilled hole from the desired hole trajectory is negligible. the front end of the tube is maintained along x 103 upper boundary of the hole horizontal references tube centerline lower boundary of the hole penetration (m) figure 5. drilled hole profile with full compensation controller with gains k1 = 871 and k2 = 1742.5 x 103 upper boundary of the hole horizontal reference tube centerline lower boundary of the hole penetration (m) figure 6. drilled hole profile with partial compensation controller with gains k1 = 871 and k2 = 1742.5 63 dynamic analysis and vibration of beam inside annulus for ultra short-radius water jet drilling the desired trajectory by the controller while the portion of the tube between the clamped and nozzle ends rests on the lower boundary of the drilled hole. the above control law requires the function f ( wn, xn, t ) which might be difficult to obtain in actual applications. hence, a partial compensation control which does not require this function is also used to determine the profile of the drilled hole as shown in fig. 6. the same values of k1 and k2 are used. this controller does not require the position information of the two nodes preceding the last node. however, the drilled hole profile is not regular and partially deviates from desired trajectory as shown in fig. 6. 5. conclusions a closed-form solution is obtained to predict the path of the drilled hole and associated sagging in orizontal water-jet drilling process incorporating gravity, damping and impact forces. it is found that in absence of a controller the drilled hole has irregular profile and sags more than 20 mm for 3 m of penetration. the variation in thrust from 10 kn to 20 kn results in unacceptable hole profile and trajectory for field applications. in addition, the amount of sagging reaches to more than 4 m for 400 m length of penetration, which results in missing the thickness of oil-sands in shallow oil fields. therefore, the use of uncontrolled water jetting system for drilling long sidetracks is not possible. however, adding full compensation controller with gains of k1=871 and k2=1742.5, the drilled hole trajectory remains within the set clearance of 5 mm. the drilling tube-nozzle systems settles within 0.2 seconds and desired trajectory and profile of drilled can be attained by adjusting the controller gains. these gains can be set to different values depending on nozzle dynamics, clearance between tube and hole, and material properties of tube and formation. references buset p, riiber m, eek a (2001), jet drilling tool: cost-effective lateral drilling technology for enhanced oil recovery. spe 68504, spe coiled tubing roundtable, usa. dickinson w, pesavento mj, dickinson rw (1990), data acquisition, analysis and control while drilling with horizontal water jet drilling systems. cim/spe # 90-127, canada. fair c (1998), development of high pressure abrasive jet drilling. spe of aime, usa. halliburton case history (2009), short-radius drilling technology enhances production in depleted field. case history # ho1770. jansen jd (1993), nonlinear dynamics of oilfield drillstrings. ph.d thesis, delft university. kolle jj (1998), water and abrasive jetting, and mechanical techniques expedite hard rock drilling. oil and gas journal 96(16):90-94. kolle jj, marvin mh (2000), jet-assisted coiled tubing drilling with supercritical carbon dioxide. proceedings of etce/omae2000 joint energy conference, new orleans, usa. laprade jd (2000), slim hole rotary steerable shortradius horizontal drilling system. 8th international williston basin horizontal well workshop, north dakota, usa. stoner ms (1999), hole deviation is defined. denver spe drilling study group meeting, usa. schiesser we (1991), the numerical method of lines integration of partial differential equations. academic press, san diego, usa. tabatabaei m, ghalambor a (2009), a new method to predict performances of horizontal and multilateral wells. iptc paper # 13122, doha, qatar. zhu h (2009), development and trend of underbalanced drilling technology. petroleum geology and engineering 23(2):80-82. civl50827.qxd vol. 3, no. 1 (2006) 75-78ring researchrnal of engineethe jou 1. introduction the main function of a highway pavement is to transport people and goods in a safe, comfortable, and economic manner. this implies that the pavement structure should provide two basic types of services. first, the pavement should provide a functional service by giving the users a safe and comfortable ride for a specific range of speed. second, the pavement must provide a structural service by supporting traffic loading and withstanding environmental influences. one of the important factors that determine a pavement functional service is the level of skid resistance. safe driving depends on an adequate surface friction for vehicle maneuvering, turning and braking. surface friction is generally given by the following equation: f = fw (1) where, f = tractive force (horizontal force applied to the test tire at the tire-pavement contact), (pound-force, lbf); f = friction factor or friction coefficient; and w = dynamic vertical load on test wheel (lbf). ________________________________________ *corresponding author’s e-mail: amansour@ksu.edu.sa the value of the coefficient f depends upon several factors including tire pressure; tire wear and inflation pressure; vehicle speed; environmental conditions (wet and dry); pavement temperature; aggregate angularity; and asphalt content. the effect of these factors will be discussed later. a standard test procedure would apply the above test under preset values of the above mentioned factors. an example of such standard procedures is that recommended by astm e274, from which a skid number, sn is calculated as: sn = 100 f = 100f/w (2) where f is obtained in a precisely defined manner (for example, by sliding a locked standard sized tire from a constant speed, typically 40 mph, along an artificially wetted pavement surface). it is important at this point, and for subsequent discussions, to indicate that the higher is the sn value the more skidding resistant is the pavement surface. the importance of pavement friction is basically related to driver safety during wet weather. to ensure a safe highway travel in a wet weather, a pavement must have sufficient skid resistance to enable drivers to perform driving tasks without the risk of skidding and loss of vehicle control. pavement surfaces without sufficient skid resistance would endanger the safety of users in two ways. effects of pavement skid resistance on traffic accidents abdullah i. al-mansour* civil engineering department, king saud university, p.o. box 800, riyadh, 11421, saudi arabia received 27 august 2005; accepted 27 november 2005 abstract: the ministry of transport (mot) in the kingdom of saudi arabia had collected a massive amount of friction measurements using a mu-meter covering most of the major highway network in the kingdom. traffic accident data of 89 high accident rate locations from four main different highway classes were extracted from the mot accident records. pavement skid resistance for the selected locations was determined from the pavement skid resistance records. the objective of this paper is to utilize these data to investigate the effects of pavement skid resistance on traffic accidents. the analysis included establishing relationships between skid resistance and accident number, accident significance and accident density. it was determined that a decreasing skid resistance leads to an increase in traffic accidents. a critical value of skid resistance was also established based on number, significance and density of accidents. keywords: pavement, skid resistance, traffic accidents, density, significance degree, critical value ájqhôÿg çogƒ◊g ≈∏y ∞°uôdg ¥’õfg áehé≤e òkéj qƒ°üæÿg ¬∏dg~ñy@ ::áá°°uuóóÿÿggádé°sôdg √òg ±~¡j .ájqhôÿg çogƒ◊g äéfé«h ™ªlh á«°ù«fôdg ¥ô£dg áµñ°t ≈∏y ¥’õf’g áehé≤ÿ ájqho äé°sé«b πª©h ájoƒ©°ùdg á«hô©dg áµ∏ªÿéh π≤ædg iqgrh ωƒ≤àe øe ±éæ°ug á©hq’ ™bƒe 89 ∫ ájqhôÿg çogƒ◊éh ¥’õf’g áehé≤e §hq á°sgq~dg â∏ª°t .ájqhôÿg çogƒ◊g ≈∏y ¥’õf’g áehé≤e òkcéj åëñd äéfé«ñdg √òg øe ioéøà°s’g ¤g çogƒ◊g ioéjr ¤g …odƒj ¥’õf’g áehé≤e πeé©e ¢véøîfg ¿g á°sgq~dg âæ«h ~≤d .çogƒ◊g áaéãch iqƒ£nh o~y øe πch ¥’õf’g áehé≤e úh äébóy ôjƒ£j á°sgq~dg â∏ª°t .¥ô£dg .¥ô£dg øe ∞æ°u πµd çogƒ◊g áaéãch iqƒ£nh o~y ≈∏y aéæh ¥’õf’g áehé≤ÿ álôm º«b á°sgq~dg äo~m éªc ájqhôÿg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.álô◊g º«≤dg ,iqƒ£ÿg álqo ,çogƒ◊g áaéãc ,ájqhôe çogƒm ,¥’õf’g áehé≤e ,∞°uq : 76 vol. 3, no. 1 (2006) 75-78ring researchrnal of engineethe jou first, it increases the stopping distance, which is a direct function of the coefficient of friction, provided by the pavement surface. secondly, it increases the risk of hydroplaning. when the water film on a pavement is of a certain thickness, vehicles may hydroplane; i.e. the tires may be separated from the pavement by the water wedge formed between the tire surface and the pavement surface. total hydroplaning occurs when fluid pressure forces (generated as a result of change in momentum of the fluid particles) in the water-wedged region exceed the total downward load on the tire (agrawal et al. 1977). the direct consequence of these two events is the high probability of the driver being involved in an accident. the ultimate goal of skid resistance measurements is to identify the level of friction provided by the pavement surface in the field. this is typically done either directly by measuring the skid level of the pavement surface, or indirectly by measuring the surface texture. furthermore, to assess the required skid resistance levels, laboratory tests should be performed to measure texture and relate those measurements to observed levels of skid-resistance on the road. there are many factors that affect pavement surface friction. the study by al-mansour et al. (2002) indicated that traffic level, highway class, pavement age and percent of air voids in the asphalt mix have significant effects on pavement friction. percent of asphalt content in the mix has a marginal effect. pavement skid resistance is reduced as a pavement gets older. skid number increases as the percent of air voids in the mix increases and decreases by increasing asphalt content. several studies have been conducted over the last three decades with solid conclusions indicating that the rate of accident occurrence during a wet weather is much higher than that during a dry weather. the study by ryell et al. (1979) reported that the rate of wet accidents (number of wet-weather accidents per 100 million vehicle-miles) increased from about 27 at an average skid number of about 55, to about 75 at an average skid number of about 25. the results of another study (zipkes, 1976) reported the significant reduction in the percent of wet-pavement accidents from 78% to 30% due to increasing the pavement coefficient of friction from 0.20 to 0.45 by applying surface grooving. on the other hand, other studies analyzed specific sites during wet and dry weather periods and concluded that the risk of being involved in an accident during wet periods is significantly higher than that during dry periods. salt (1976) concluded that whilst a surface with a friction coefficient of 0.60 and above may by chance be a scene of an accident in which a vehicle skids in wet weather, the risk that it will be the scene of repeated skidding accidents is extremely small. this risk first becomes measurable with a coefficient of 0.55 to 0.60 and increases sharply by more than 20 times as the coefficient falls to values of 0.40 to 0.45 and by about 300 times when the coefficient is 0.30 to 0.35. there are no standards agreed upon which define minimum acceptable skid resistance levels. several studies have been conducted to develop criteria for critical skid resistance levels. in general such studies were either based on skid resistance and traffic accident analysis or based on calculating the required stopping sight distance. a study conducted on the streets of muscat indicated minimum skid number of 0.45 on normal sites (ali et al. 1999). another study (rizenbergs et al. 1977) concluded that pavement with skid number below 26 are very slippery and must be corrected. 2. objectives the ministry of transport (mot) had collected a massive amount of friction measurements using a mu-meter covering most of the major highway network in the kingdom of saudi arabia. the main objective of this study is to utilize this data to investigate the effect of skid resistance level on traffic accidents. 3. data collection in order to achieve the objective, wet pavement skid resistance data and traffic accidents data for ten main highways with different classes and traffic levels were extracted from the mot skid resistance and traffic accidents files (al-tamime, 2002). four roadway classes were included: dual roads (a) roads that have double lanes in one-way direction; expressway (b) roads that have three lanes or more in one-way direction; expressway with frontage roads (c); and undivided roads (d). the exact locations of the traffic accidents were matched with pavement skid resistance measurements. this process resulted in determining the skid number of the pavement for accident location. the traffic accident report does not specify clearly the cause of accident. therefore, all accidents reported in the year 1999 for the selected locations were incorporated in the analysis, including accidents in both wet and dry weather conditions. in addition to the number of accidents, other data such as types of accidents (death, injury and damage), degree of accident significance, and density of accident were also recorded. types of accident included accident that resulted in death, injury or damage. an accident degree of significance is defined as a recurring accident site factor. density of accidents is the number of traffic accidents annually per kilometer of road. 4. analysis and results 4.1 basic statistics a total of ten highways with 340 accidents occurred at 89 locations were used in the analysis. the number of accidents and accident locations for each road are shown in fig. 1. road number 15, in the south western region of the kingdom of saudi arabia, had 145 accidents which was the highest number of accidents of the roads included in the analysis. roads number 500 and 517 had the lowest number of accidents. similarly, road number 15 has the highest number of accident locations and roads numbers 500 and 517 had the lowest number of accident 77 vol. 3, no. 1 (2006) 75-78ring researchrnal of engineethe jou location. 4.2 effects of skid resistance on number of accidents although accident reports did not specify the main cause of accidents, an attempt was made to relate skid resistance measurements to the total number of accidents for the analysis period. the analysis indicated that the number of accidents increased as skid resistance decreased. a low level of skid resistance resulted in an increase in stopping distance and a loss of vehicle control at a high speed. the relationship between number of accidents and pavement friction measured as a skid number is shown in fig. 2. this data were best fitted in the following form: na = a snb (3) where; na = number of accidents; sn = skid number; and a, b = regression parameters the values of a and b were estimated to be 2.799 and 0.408, respectively. the coefficient of correlation (r2) of the regression equation was 0.425. this relationship indicated that the number of accidents approach almost a constant number at a skid number greater than 0.45. to further investigate this observation, the number of accidents were plotted as a bar chart against the ranges of skid resistance. the plot is presented in fig. 3. pavement with skid resistance less than 0.35 have the highest number of accidents the data also indicate the same observations, that the number of traffic accidents are almost the same for skid number greater than 0.45. these results lead to the conclusion that the critical skid number required at high speed roads is 0.45. a similar analysis to the one just presented was conducted for each highway class. as indicated earlier four classes of highway are included. these classes are a, b, c, and d. this analysis was conducted on highway class a, b and d. highway class c was excluded because of the lack accident data. a summary of the results is presented in table 1. it is clear that as the highway class gets higher the critical skid number value increases. these results were expected since the operating speed on a higher highway class is more than that of a less class highway. 4.3 effect of skid resistance on accident density the density of accidents is defined as the number of traffic accidents made annually for each road kilometer. accident density is calculated as number of accidents divided by the analysis period and multiplied by section length. the relationship between skid number and accident density is presented in fig. 4. it was found that as skid number decreases, accident density increased. the maximum accident density of 2.66 occurs at a skid number less than 0.35. at a range of 0.75 0.80, accident density was found to be 1.76. accident density seems to be constant at about 1.85 for a skid number greater than 0.45 (critical value). the statistical regression procedure (neter et al. 1985) was used to model the accident density based on skid resistance number. the best model was found to be in the following form: ad = a snb (4) the values of the parameters a and b were estimated to be 1.59 and 0.278, respectively. the p-value for the model is 0.0023 and the adjusted coefficient of determination (r2) is 0.36. 0 20 40 60 80 100 120 140 160 r15 r40 r50 r60 r65 r80 r85 r279 r500 r517 roadway number n um be r o f a cc id en t a nd a cc id en t l oc at io n number of accidents number of accident locations figure 1. number of accidents and accident locations 0 1 2 3 4 5 6 7 8 9 10 11 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 skid number a cc id en t n um be r/l oc at io n figure 2. relationship between skid number and traffic accidents 0 20 40 60 80 100 < 0.35 0.35 0.45 0.45 0.55 0.55 0.65 0.65 0.75 0.75 0.80 skid number n um be r o f a cc id en ts figure 3. effect of skid number of traffic accidents highway class critical sn model adj r2 a 0.48 na = 2.79 sn-0.41 0.425 b 0.42 na = 2.76 sn-0.38 0.265 d 0.38 na = 2.73 sn-0.36 0.588 tble 1. critical skid number values 78 vol. 3, no. 1 (2006) 75-78ring researchrnal of engineethe jou 4.4 effect of skid resistance on accident degree of significance the accident degree of significance is defined as recurring accident site factor. this factor was determined by multiplying number of accidents by accident seriousness divided by analysis period multiplied by average annual traffic and section length. the seriousness of accident is calculated as: [1 + (number of death accident + number of injury accidents + number of damage accident)/ total number of accidents]. the accident degree of significance is strongly bounded with the number of accidents. whenever skid resistance level is low, accident degree of significance increases. the effect of skid resistance on accident degree of significance is shown in fig. 5. pavement sections with skid number less than 0.35 has an accident degree of significance of 2.51. the results also indicated that the accident degree of significance is almost the same for all sections with skid resistance number greater than 0.45. this is another indication that the critical skid resistance number on high speed road can be considered as 0.45. at this critical skid resistance level, the degree of accident significance for all highway classes is 1.7; highway class "a" is 1.7; highway class "b" is 2.1; and for highway class "d" is 2.2. 5. conclusions within the scope of this investigation and based on the results obtained, the following may be concluded: 1. the most common method of evaluating pavement skid resistance was the locked-wheel skid test using astm e274 test method. 2. there are no standards defined for minimum acceptable skid resistance levels. 3. road number 15, in the southwestern region of the kingdom has the highest number of accidents and accident locations of all the roads included in the analysis. 4. pavements with a skid resistance number less than 0.35 have the highest number of traffic accidents. the number of traffic accidents seems to be constant for all sections with a skid resistance number equal to or greater than 0.45. 5. the maximum accident density of 2.66 occurred at pavement sections with a skid resistance less than 0.35. accident density remained constant on all sections with a skid resistance number equal to or greater than 0.45. 6. whenever skid resistance number is low, accident degree of significance increases. pavement sections with skid resistance number less than 0.35 have 2.51 accident degree of significance. accident degree of significance remains constant for all sections with skid resistance number equal to or greater than 0.45. references agrawal, s.k. and henry, j.j, 1977, "technique for evaluating hydroplaning potential of pavements," trrl report lr633. ali, g., al-mahrooq, r. and tahqa, r., 1999, "measurement, analysis, evaluation, and restoration of skid resistance on streets of muscat," j. of the transportation research board, transportation research record no. 1655, washington, d. c. , usa al-mansour, a., al-suhaibani, a. and al-syyari, s., 2002, "pavement skid resistance analysis and evaluation of highway network in saudi arabia," first gulf conference on roads, kuwait, march 2002. al-tamime, m., 2002, "pavement skid resistance and traffic accidents," senior project, college of engineering, king saud university, saudi arabia. neter, j., wasserman, w. and kutner, m., 1985, applied linear statistical models, irwin inc., homewood, illinois, usa. rizenbergs, r. l., burcheit, j.c. and warren, l., 1977, "relation of accidents and pavement friction on rural two-lane roads," trrl report lr633. ryell, j., corkill, t. and musgrove, c., 1979, "skid resistance of bituminous pavement test sections: toronto by-pass project," trrl report lr712. salt, g.f., 1976, "research on skid resistance at the transport and road research laboratory," trrl report lr622. zipkes, e., 1976, "the influence of grooving of road pavement on accident frequency," trrl report lr623. 0 0.5 1 1.5 2 2.5 3 < 0.35 0.35 0.45 0.45 0.55 0.55 0.65 0.65 0.75 0.75 0.80 skid number a cc id en t d en si ty figure 4. effect of skid number level on accident density 0 0.5 1 1.5 2 2.5 3 < 0.35 0.35 0.45 0.45 0.55 0.55 0.65 0.65 0.75 0.75 0.80 skid number d eg re e of s gn ifi ca nc e figure 5. effect of skid nuber level on accident degree of significance comp080425.qxd the journal of engineering research vol. 6, no. 2 (2009) 12-19 1. introduction although many codes for example binary coded decimal (bcd), excess-3 code, hamming code, cyclic redundancy code (crc), check sum and many others exist and are in use. but the gray codes which are named for frank gray who patented the use of them in shaft encoders (gray, 1953) due to its attribute of single distance only, which avoids ambiguous switching situations, is particularly used to handle safely and conveniently the control problems. the term gray code is sometimes used to refer to any single-distance code, that is, one in which adjacent code words differ by 1 in one digit position only, this property can be seen in table 1 which shows the gray codes for size n = 1 to 4 bits. unlimited applications are accounted for the gray code. some of them are mentioned here as follows. however, more can be learned through the references pro vided as in the references (sundberg, 1975; ludman, 1981; er, 1984, and proskurowski and ruskey, 1985; _____________________________________ *corresponding author’s e-mail: afaq@squ.edu.om lee, 1986; conway, et al. 1989; skiena, 1990; press, et al. 1992; etzion and peterson, 1992; hiltgen, et al. 1996; savage, 1997; ruskey, 1997; guan and dah-jyu, 1998, moshe and tuvi, 1999; black paul 2004; alan and alessandro mei, 2004; jywe-fei and lai, 2005; bitner, et al. 2005). gray codes were applied to mathematical puzzles before they became known to engineers. the gray code arises naturally in many situations. gray's interest in the code was related to what we would now call analog to digital conversion. the goal was to convert an integer value, represented as a voltage, into a series of pulses representing the same number in digital form. the technique, as described in gray's patent, was to use the voltage being converted to displace vertically an electron beam that is being swept horizontally across the screen of a cathode ray tube. the screen has a mask etched on it that only allows the passage of the beam in certain places; a current is generated only when the beam passes through the mask. the passage of the beam will then give rise to a series of on/off conditions corresponding to the pattern of mask holes that it passes. a less complex algorithmic procedure for computing gray codes afaq ahmad*a and mohammed m. bait suwailama *a department of electrical and computer engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, muscat, sultanate of oman received 25 april 2008; accepted 26 october 2008 abstract: the purpose of this paper is to present a new and faster algorithmic procedure for generating the nbit gray codes. thereby, through this paper we have presented the derivation, design and implementation of a newly developed algorithm for the generation of an n-bit binary reflected gray code sequences. the developed algorithm is stemmed from the fact of generating and properly placing the min-terms from the universal set of all the possible min-terms [m0 m1 m2 …. mn] of boolean function of n variables, where, 0 < n < 2 n-1. the resulting algorithm is in concise form and trivial to implement. furthermore, the developed algorithm is equipped with added attributes of optimizing of time and space while executed. keywords: gray code, min-terms, boolean function, algorithm, processing time, memory space, binary …ôl rƒeq üé°ù◊◊ g~«≤©j πbg á«erqgƒn ägƒ£n º∏jƒ°s â«h ~ªfi h ~ªmg ¥éacg ::áá°°uuóóÿÿgg~«dƒàd i~j~÷g á«erqgƒÿg ωg~îà°sg h º«ª°üj ägƒ£nh ¥é≤à°tg ë~≤j ” ~≤d .»féæk ºbq -¿ ºéëh …ôl rƒeq ~«dƒàd á©jô°s i~j~l á≤jôw ë~≤j ¤g åëñdg gòg ±~¡j á«æ«dƒñdg ádg~∏d áeé©dg oh~◊g áyƒªûg iô¨°üdg oh~ë∏d ™bgƒÿg qé«àng h ~«dƒj á≤«≤m øe á°ùñà≤e iqƒ£ÿg á«erqgƒÿg ¿eg .»féæk ºbq -¿ ºém ägph á°ùµ©æÿg …ôl õeq á∏°ù∏°s πbcg ¬e~îà°ùe iôcgòh h π° acg ò«øæj øer ägp ’ƒ∏m èàæj iqƒ£ÿg á≤jô£dg ¿cg ¤eg áaé°veóh gòg .≥«ñ£à∏d á∏héb ᨫ°üh óm áœéædg á≤jô£dg »£©j .ò¨àe ¿ øe áfƒµàÿg ádg~∏d .éªém áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.»féæãdg ωé¶ædg ,iôcgòdg ºém ,ò«øæàdg øer ,äé«erqgƒÿg ,á«æ«dƒñdg ádg~dg ,iô¨°üdg oh~◊g ,…ôl õeq : 13 the journal of engineering research vol. 6, no. 2 (2009) 12-19 mechanical position sensors use gray code to convert the angular position (angle-measuring devices) of a shaft to digital form. gray codes were used in telegraphy. the gray code also forms a hamiltonian cycle on a hypercube, where each bit is seen as one dimension. in data transmission, gray codes play an important role in error detection and correction. solving puzzles such as the tower of hanoi and the brain, the study of bell-ringing, analog-digital signal conversion, classifying of venn diagrams, continuous space-filling curves, enhancing the resolution of spectrometer for a satellite application, labeling the axes of karnaugh maps are the processes where gray codes are used due to its uniqueness. gray codes are also beneficial in genetic algorithms due to its incremental change property. using gray codes for addressing the memory results in saving of the power because a few address lines change as the program counter advances to the next location. also, gray codes are extensively used by digital system designers for passing multi-bit count information between synchronous logic that operates at different clock frequencies. in some numerical problems, gray codes can be useful in situations of looping over many values of a bit. furthermore, due to its attribute gray code could be a good choice for the search of the optimal test-sequences in digital system testing. hence it can be said that the gray codes which were originally designed to prevent spurious output from electromechanical switches. today they are widely used to facilitate error correction in digital communications such as digital terrestrial television and some cable tv systems. since the gray code has enormous applications as mentioned above has many facet researches as is evident in surveying the literature (gray, 1953; sundberg, 1975; ludman, 1981; lee, 1986; conway, et al. 1989; skiena, 1990; press, et al. 1992; etzion peterson, 1992; hiltgen, et al. 1996; savage, 1997; ruskey, 1997; guan dah-jyu, 1998, moshe and tuvi, 1999; black paul 2004; alan and alessandro mei, 2004; jywe-fei and lai, 2005; bitner, et al. 1976; er, 1984; proskurowski and ruskey, 1985; ruskey, 1993; dominique, 2000; lassing, et al. 2003, goddyn and gvozdjak 2003 and vajnovszki and walsh 2006). it is clear through the literature survey that there had been much discovered and written about the gray code; it is associated with many elegant circuits and algorithms. however, the algorithms generating the gray code was still done with the crude techniques (ruskey, 1993; dominique, 2000; lassing, et al. 2003, goddyn, gvozdjak 2003 and vajnovszki and walsh 2006). researches are available in the literature only to script the faster codes but not much deviated from the existing crude algorithms for generating the gray codes. this paper presents a new concept of generating the gray code of n-bit size. the developed algorithm is stemmed from the fact of generating and properly placing the min-terms from the universal set of all the possible min-terms (m0 m1 m2 …. mn) of boolean function of n variables, where, 0 < n < 2n-1. the resulting algorithm is in concise form and trivial to implement. we designed an efficient algorithm to write the codes which reduces the processing time and memory space requirements. 2. gray code conversion conventional approaches no doubt, a simple recursive equation in mod (2) operation can convert a simple binary -…-8-4-2-1 codes to the gray one. the hardware is also simple which is based on the bank of exclusive-or gates. to make the content of this paper more readable to audience the description of the generation procedure of the gray code is given below: to convert a binary number [bn-1 bn-2 …. b1 b0] to its corresponding gray code (gn-1 gn-2 …. g1 g0), start at the left with the bit bn-1 (the nth, most significant bit) and use the following recursive equations. gn-1 = bn-1 (1) (2) or, in general for ith bit where i varies as, 2 < i < n, the following eq. (3) can be used. (3) the above eqs. (1) and (3) are illustrated for a 3-bit gray code encoder in an example below. example 1: let a binary code [b3 b2 b1] = [1 0 0]. base 10 binary code gray code 4-bit gray code 3-bit gray code 2-bit gray code 1-bit 0 0000 0000 000 00 0 1 0001 0001 001 01 1 2 0010 0011 011 11 3 0011 0010 010 4 0100 0110 110 5 0101 0111 111 6 0110 0101 101 7 0111 0100 100 8 1000 1100 9 1001 1101 10 1010 1111 11 1011 1110 12 1100 1010 13 1101 1011 14 1110 1001 15 1111 1000 table 1. gray code patterns of size n = 1 to 4 bits gn-2 = bn-1 ⊕ bn-2 gn-i = bn-i+1 ⊕ bn-i so, g3 = b3 = 1; g2 = b3 ⊕ b2 = (1 + 0) mod 2 =1; g1 = b2 ⊕ b1 = (0 + 0) mod 2 = 0. therefore, binary code [b3 b2 b1] = [1 0 0] when encoded in gray code gives [g3 g2 g1] = [1 1 0]. 14 the journal of engineering research vol. 6, no. 2 (2009) 12-19 the derivation of the generalized eqs. (1) (3) can be computed as below. from table 1 above, a set of the following boolean functions expressed in the form of min-terms σ mi where i varies from 0 to 7, can be derived as: g3 (b3, b2, b1) = σ (m4, m5, m6, m7) (4) g2 (b3, b2, b1) =σ (m2, m3, m4, m5) (5) g1 (b3, b2, b1) = σ (m1, m2, m5, m6) (6) the minimized version of above boolean functions using the k-maps and hence the implementations of those minimized functions are as given in figs. 1 and 2 respectively. similarly, the reversal of gray code bits [gn gn-1 … g2. g1] again into -…-8-4-2-1 weighted binary code bits [bn bn-1 … b2. b1] can be performed by using the equations as given below: bn = g n (7) (8) . . (9) or, in general, to compute the binary code bits bn-1, … , b2, and. b1 the following recursion equation can be used where, i varies from 1 to n-1. (10) example 2: let a gray code [g3 g2 g1] = [1 1 0]. the derivation of the generalized eqs. (7) (10) can be easily computed as: from table 1 above, a set of the following boolean functions expressed in the form of min-terms σ mi where i varies from 0 to 7, can be derived as: b3 (g3, g2, g1) = σ (m4, m5, m6, m7) (11) b2 (g3, g2, g1) = σ (m2, m3, m4, m5) (12) b1 (g3, g2, g1) = σ (m1, m2, m4, m7) (13) the above boolean functions can be minimized using the k-maps as demonstrated below (see fig. 3). whereas, the implementation of those minimized boolean functions is shown in fig. 4. since the gray code encoder needs first to obtain the figure 1. k-maps and minimized boolean functions for g3, g2 and g1 g3 g2 g1 b3 b2 b1 msb lsb figure 2. a 3-bit binary to gray converter circuit b1 = b2 ⊕ g1 bn-i = bn-i+1 ⊕ gn-i so, b3 = g3 = 1; b2 = b3 ⊕ g2 = (1 + 1) mod 2 =0 ; b1 = b2 ⊕ g1 = (0 + 0) mod 2 = 0. therefore, gray code [g 3 g2 g1] = [1 0 0] when encoded in binary code [b3 b2 b1] gives = [1 1 0]. bn-1 = bn ⊕ gn-1 15 the journal of engineering research vol. 6, no. 2 (2009) 12-19 binary data (through a binary counter) before generating the gray code. this process requires 2-stages as shown in fig. 5. thus, it is imperative to derive a mechanism to avoid this situation which needs much time and space to implement. the ensuing section is a consequence to it. 3. computing gray codes proposed methodology if we look to the table 2 and 3 which lists the gray codes, respective min-terms and equivalent decimals for n = 1 to 6 forced us to derive the following conclusions. 1. gray code of size n has a specific pattern relationship between min-terms of the gray code of its predecessor code of size n-1. 2. the gray code of size n can be directly scripted using the n-bit k-map where the min-terms cells are to be read clock-wise and down to the row as explained in the fig. 6. the example of fig. 6 is a 4-variable k-map used to figure 3. k-maps and minimized boolean functions for b3, b2 and b1 b3 b2 b1 g3 g2 g1 msb lsb figure 4. a 3-bit gray to binary converter circuit clk binary counter binary to gray converter figure 5. black model of binary to gray code conversion table 2. decimal, gray code, and (min-terms): for, n = 1 to 5 16 the journal of engineering research vol. 6, no. 2 (2009) 12-19 generate gray code which reads as [m0 m1 m3 m2 m6 m7 m5 m4 m12 m13 m15 m14 m10 m11 m9 m8]/. the derivation 2 is not feasible since generating the k-map for higher variables are difficult to manipulate. therefore, the derivation 1 is the point of our work. 4. proposed algorithm the following eqs. (14) to (18) describe the gray codes for bit size 1 to 5 respectively. a specific relation between the min-terms patterns is visible and summarized in the form of a theorem below: g(1) = [m0 | m1]/ (14) g(2) = [m0 m1 | m3 m2]/ = [g(1); [m3 m2]/] (15) g(3) = [m0 m1 m3 m2 | m6 m7 m5 m4]/ = [g(2); [m6 m7 m5 m4]/] (16) g(4) = [m0 m1 m3 m2 m6 m7 m5 m4 | m12 m13 m15 m14 m10 m11 m9 m8]/ = [g(3); [m12 m13 m15 m14 m10 m11 m9 m8]/ (17) g(5)seti = [m0 m1 m3 m2 m6 m7 m5 m4 m12 m13 m15 m14 m10 m11 m9 m8]/ = g(4) g(5)set ii = [m24 m25 m27 m26 m30 m31 m29 m28 m20 m21 m23 m22 m18 m19 m17 m16]/ g(5) = [g(4); [m24 m25 m27 m26 m30 m31 m29 m28 m20 m21 m23 m22 m18 m19 m17 m16]/] (18) theorem 1: g (n) is a matrix of order 2n x n in binary format of nbit forming 2n min-terms of n variables. by analyzing the patterns we reach to the conclusion that g(n) can be obtained first by writing the min-terms of g(n-1) then appending the min-terms by advancing each of the minterms starting from the last to the first by a value of 2n-1. proof: proof is as visible through the eqs. from (14) to (18). an algorithm is designed on the basis of the study of theorem 1 and is as given below. algorithm step 0: start by inputting the bit size (n) of the gray code to be generated; step 1: initialize a vector v = [0 1]; step 2: count a loop for k = 1 to n; step 3: check that k = n or not, if yes go to step 8; step 4: increment the counter k by 1 i.e. k = k +1; step 5: deci mal gray code (m) deci mal gray code (m) 0 000000 (m 0) 48 110000 (m 48) 1 00001 (m 1) 49 110001 (m 49) 3 000011 (m 3) 51 110011 (m 51) 2 000010 (m 2) 50 110010 (m 50) 6 000110 (m 6) 54 110110 (m 54) 7 000101 (m 7) 55 110111 (m 55) 5 000101 (m 5) 53 110101 (m 53) 4 000100 (m 4) 52 110100 (m 52) 12 001100 (m 12) 60 111100 (m 60) 13 001101 (m 13) 61 111101 (m 61) 15 001111 (m 15) 63 111111 (m 63) 14 001110 (m 14) 62 111110 (m 62) 10 001010 (m 10) 58 111010 (m 58) 11 001011 (m 11) 59 111011 (m 59) 9 001001 (m 9) 57 111001 (m 57) 8 001000 (m 8) 56 111000 (m 56) 24 011000 (m 24) 40 101000 (m 40) 25 011001 (m 25) 41 101001 (m 41) 27 011011 (m 27) 43 101011 (m 43) 26 011010 (m 26) 42 101010 (m 42) 30 011110 (m 30) 46 101110 (m 46) 31 011111 (m 31) 47 101111 (m 47) 29 011101 (m 29) 45 101101 (m 45) 28 011100 (m 28) 44 101100 (m 44) 20 010100 (m 20) 36 100100 (m 36) 21 010101 (m 21) 37 100101 (m 37) 23 010111 (m 23) 39 100111 (m 39) 22 010110 (m 22) 38 100110 (m 38) 18 010010 (m 18) 34 100010 (m 34) 19 010011 (m 19) 35 100011 (m 35) 17 010001 (m 17) 33 100001 (m 33) 16 010000 (m 16) 32 100000 (m 32) table 3. decimal, gray code, and (min-terms): for, n = 6 figure 6. a 4-variable k-map 17 the journal of engineering research vol. 6, no. 2 (2009) 12-19 for i = 0 to 2n-1-1; vi = [v(2n-1-i) + 2n-1]; step 6: modify vector v as v = [v, vi]; step 7: go to step 3; step 8: transpose v i.e. v/; step 9: step 10: output g(n); step 11: stop a debug example: example 3: to elaborate the computing of variables in the above mentioned steps a debug test of the above algorithm is carried out for n = 3 and is presented below in table 4. 5. implementation of algorithm the above designed algorithm is implemented using the matlab code. the out put of the m-file with the name "gray_generator_proposed" can be visualized as shown in fig. 7. a sample output of the program shown in the figure is only for n = 4. this is provided just to make it more readable, otherwise the output for the higher values of n will require much space to present. similarly, we also encoded the conventional method of gray code conversion into matlab script. to compare the efficiency amongst these two approaches we run both of the programs for n = 2 to 10 and some of the results are presented and made available through figs. 8 to 9 in the ensuing section below. >>gray_generator_proposed please enter a valid positive integer ( > 1) : 4 gran output (s) =================== 0 0 0 0 0 0 0 1 0 0 1 1 0 0 1 0 0 1 1 0 0 1 1 1 0 1 0 1 0 1 0 0 1 1 0 0 1 1 0 1 1 1 1 1 1 1 1 0 1 0 1 0 1 0 1 1 1 0 0 1 1 0 0 1 convert v / into binary format i.e. (v /)10 → (v/)2 = gray code of n -bit size i.e. a matrix g (n) of size (n+1) by n; i vi v (v /)base 2 [0 1] 0 to 1 [3 2] [0 1 3 2] 0 to 3 [6 7 5 4] [0 1 2 3 6 7 5 4] 000 001 011 010 110 111 101 100 table 4. a debug test of algorithm for, n = 3 2 3 4 5 6 7 8 9 10 0 0.5 1 1.5 2 2.5 x 10 5 n-bits m e m o ry s p a c e our propos ed method conventional method figure 8. memory requirements comparison 2 3 4 5 6 7 8 9 10 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 n-bits r e la ti v e e rr o r figure 9. relative error judgment 18 the journal of engineering research vol. 6, no. 2 (2009) 12-19 elapsed time is 0.000000 seconds. figure 7. matlab command window output 6. results matlab 7.0 on a p 4 cpu 1.5 ghz, 512 mb of ram is used as the bench marks for testing the codes for both of the computing methodologies. the processing time and memory space required to implement both of the programs (the conventional and the proposed one) are recorded while running the programs. the results for a subset of the study, for n = 2 to 10 where the memory requirement comparison of the conventional approach of gray code converter and the proposed approach is presented in fig. 8. the memory requirement is mapped in bytes. further, to judge the efficiency of the proposed algorithm, a relative error plot is shown in fig. 9. with respect to the memory requirements, the behavior of the relative error shown in fig. 9, demonstrates almost an exponential characteristic except for n = 2. and, this exception for (n = 2) is because of that the implementing the gray code generator using conventional approach requires significant more memory space than the proposed approach of generating the gray codes for n = 2. the run time requirements comparison is demonstrated through fig. 10. 7. conclusions a time space optimal algorithmic procedure to generate gray code-words of any bit length n is presented through this paper. the comparative study reveals that the proposed approach is not only faster but also, requires about 25% less memory space on average while compared with the conventional method of gray code-word generation technique. since gray code is widely used for on line system monitoring hooked with sensors and with on board systems and hence these two parameters (space and time requirements) are very critical in these applications. therefore this proposed algorithmic procedure is more advantageous. 8. references alan, a., bertossi, alessandro, mei., 2004, "time and work optimal simulation of basic reconfigurable meshes on hypercubes," j. of parallel and distributed computing, vol. 64(1), pp.173 180. bitner, j., ehrlich, g. and reingold, e., 1976, 'efficient generation of the binary reflected gray code and its applications," communications of the acm, vol. 19(9), pp. 517-521. black paul, e., (2004), "gray code," nist national institute of standards and technology. conway, j., sloane, n. and wilks, a., 1989, "gray codes and reflection groups," graphs and combinatorial, vol. 5, pp. 315-325. dominique roelants van baronaigien., 2000, "a loopless gray-code algorithm for listing k-array trees," j. of algorithms, vol. 35(1), pp.100-107. er, m.c., 1984, "on generating the n-array reflected gray codes," ieee transactions on computers, vol. 33, pp. 739-741. etzion, t., paterson, k.g., 1996, "near optimal singletrack gray codes," ieee trans. on information theory, vol. 42(3), pp. 779-789. goddyn, l., gvozdjak, p., 2003, "binary gray codes with long bit runs," electron. j. combin. vol. 10, pp. 110. gray, f., 1953, "pulse code communication," united states patent number 2, 632, 058. guan dah-jyu., 1998, "generalized gray codes with applications," proceeding national science council of republic of china, vol. a(22), pp. 841 848. hiltgen, p., paterson, k. g. and brandestini, m., 1996, "single-track gray codes," ieee trans. on information theory, vol. 42(5), pp. 1555-1561. jywe-fei fang , kuan-chou lai, 2005, "embedding the incomplete hypercube in books," information processing letters, vol. 96(1), pp. 1-6. lassing, e. g. strom, t., ottosson. and agrell, e., (2003), "computation of the exact bit error rate of coherent m-array psk with gray code bit mapping," ieee transactions on communications, vol. 51(11), pp. 1758-1760. lee, p. j., 1986, "computation of the bit error rate of coherent m-array psk with gray code bit mapping," ieee transactions on communications, vol. com-34(5), pp. 488-491. ludman, j. e., 1981, "gray code generation for mpsk signals," ieee transactions on communications, vol. com-29(10), pp. 1519-1522. moshe schwartz and tuvi etzion, 1999, "the structure of single-track gray codes," ieee transactions on information theory, vol. 45(7), pp. 2383 2396. figure 10. run time requirements comparison 19 the journal of engineering research vol. 6, no. 2 (2009) 12-19 press, w.h., flannery, b.p., teukolsky, s. a. and vetterling, w.t., 1992, "gray codes, numerical recipes in fortran," cambridge, england: cambridge university press, pp. 886-888. proskurowski, a. and ruskey, f., '1985, "binary tree gray codes," j. algorithms 6, pp. 225-238. ruskey, f., 1997, "a survey of venn diagrams," elec. j. of comb, a special issue. ruskey, f., 1993, "simple combinatorial gray codes constructed by reversing sub-lists," proceedings of the 4th international symposium on algorithms and computation, p. 201-208, december 15-17. savage, carla., 1997, "a survey of combinatorial gray codes," society of industrial and applied mathematics review, vol. 39, pp. 605-629. skiena, s., 1990, "gray code, computational discrete mathematics: combinatorics and graph theory with mathematica," reading, ma: addison-wesley, pp. 42-43 and 149. sundberg, c. e., 1975, "bit error probability properties of gray-coded m.p.s.k signals," iee electronics letters, vol. 11(22), pp. 542-544. vajnovszki, v. and walsh, t.r., 2006, "a loop-free twoc:\users\raoh\desktop\albadi3.xps the journal of engineering research vol. 7 no. 2 (2010) 24-31 1. introduction there is a strong growth in wind-installed capacity world-wide due to many reasons. on one hand, there is a public awareness and concerns about emissions, climate change and environmental issues about other competing source of energy. this encourages governments in many parts of the world to promote renewable sources of energy, including wind power, to provide part of electricity needs. therefore, incentives and tax credits programs have been implemented in many countries to reach approved national and local renewable portfolio standards (rps). on the other hand, improved wind turbine technology results in lower costs of wind power. in the year 2007 alone, 19.696 gw of wind power capacity were added. by the end of december 2007, _________________________________________ *e-mail: mbadi@squ.edu.om currently working with department of electrical and computer engineering, college of engineering, sultan qaboos university, po box 33, pc 123, al-khoud, muscat, sultanate of oman in the year 2007 alone, 19.696 gw of wind power capacity were added. by the end of december 2007, the global installed capacity reaches 93.849 gw. figure 1 shows the 13 countries that have 1gw or more installed capacity of wind generation in 2007 (world wind energy association, 2008). the fluctuation of wind power output is directly attributed to wind speed variations. these variations depend on changing climatic conditions and result in an intermittent and highly variable source of power. the intermittent nature of wind power represents a challenging constraint on power system operators as supply and demand need to be continuously balanced in real time. therefore, the presence of any intermittent source of power, such as wind power facilities, increases balancing requirement and associated costs (gil and joss, 2006). impacts of wind power variability on generation costs an overview m. h. albadi* and e. f. el-saadany department of electrical and computer engineering, university of waterloo, 200 university ave. w, waterloo, on, n2l3g1, canada received 23 may 2009; accepted 5 april 2010 abstract: although wind power is sustainable, environmental friendly and relatively inexpensive source of electricity, the effects of its intermittent nature on power systems need to be carefully investigated. this paper presents an up-to-date overview of the impacts of wind power variability on overall generation cost. recent case studies from different utilities around the globe demonstrated that wind integration costs are much lower than anticipated by earlier studies. keywords: wind power variability, wind impacts on power systems, balancing requirements, wind integration costs * 25 the journal of engineering research vol. 7 no. 2 (2010) 24-31 is technically possible, higher integration costs can be incurred when wind penetration level increases. examples of areas currently having relatively high wind power penetration levels are west denmark, north of germany and galicia in spain. it is very decisive to accurately estimate the impacts and costs of wind power on the system operation when planning for high wind power penetration levels (hilttinen, 2006). after this introduction, wind power variability and its general impacts on power systems are highlighted. in section 3, the paper discusses the effects of wind power on conventional generation costs. experiences with high wind power penetration levels are summarized in section 4, followed by specific conclusions. 2. wind power variability and its impacts 2.1 reasons for variability wind speed is characterized by its high variability, both spatially and temporally. on a global scale, spatial variability is attributed to the fact that there are different climate regions on the world affected by the altitude and solar insulation. on a regional scale, wind speed varies according to the geographical location depending on the sizes of land and sea, and the presence of mountains and plain areas. in a local perspective, the type of vegetation and the local topography have a major role on wind speed (ackermann, 2005). at a given location, long-term temporal wind variability describes the fact that the amount of wind may vary annually. however, many studies estimated that the variation of mean wind power output from one 20year period to the next has a maximum standard deviation of 10% (ackermann, 2005). thus, the uncertainty of wind power production is not large for the seasonal variations are more predictable than annual ones. wind speed synoptic variations associated with the passage of weather systems are not very predictable more than few days ahead. wind variations with the time of day, which is called diurnal variations, are quite predictable. turbulences are variations of wind speed over minutes and seconds (ackermann, 2005). this type of variations can affect power quality depending on both the network impedance at the point of common coupling and the type of wind turbine. variable-speed turbines has a smoother power output compared to fixedspeed ones. diurnal and synoptic variations can affect power balancing requirements (ackermann, 2005). clear peaks corresponding to the synoptic, diurnal and turbulent effects have been presented in (hoven, 1957) as shown in fig. 2. 2.2 aggregation effect aggregation of wind turbine outputs reduces the temporal volatility of wind power by two aspects: increased number of turbines within a wind farm and spatial distribution of wind generation resources (ackermann, 2005). on one hand, as the number of turbines within a wind farm increases, turbulent wind effect is dampen because wind gusts do not hit individual turbines at the same time. in fact, a small number of turbines (n) are needed to achieve a significant smoothing effect, as the percentage variation of power output is reduced to n-1/2, ideally. on the other hand, a wider geographical dispersion of wind farm reduces the impact of diurnal and synoptic variations. spatial distribution of wind farms with certain aggregated capacity has a much lower up and down ramping rate requirements compared to a large single wind farm figure 1. countries having more than 1gw wind power installed capacity installed capacity by end of 2007 26 the journal of engineering research vol. 7 no. 2 (2010) 24-31 having the same capacity (ackermann, 2005). for example, it has been reported that the combination of wind power outputs from 17 geographically dispersed sites in ontario is estimated to reduce the variability of the aggregated wind power output about 60-70% compared to the output from one site, for 10-minute and 1hour data (truewind, 2005). table 1 presents a summary of the sources of wind power temporal variation and relevant impacts on power system operation. 2.3 general impacts of wind power the impacts of wind power integration on any power system depend on two main factors: penetration level and systems flexibility (ackermann, 2005). in general, increased wind penetration will increase the impacts perceived by the system. on the other hand, systems which are inherently more flexible than others will be able to accommodate more wind power without perceiving unwanted impacts from wind generation facilities. smith et al. reported that the existing case studies have explored wind capacity penetrations of up to 20% to 30% of system peak and have found that the primary considerations are economic, not physical (smith et al. 2007). this is summarized by how to deal with wind power uncertainty and variability. the first step to increase wind power penetration level is to understand and to quantify its impacts on utility systems. in addition, wind power plants design and operation, power systems design and operation, and market rules have great influence in accommodating wind power. there are two types of wind power impacts on system operation: short-term and long-term impacts. short-term impacts deals with operational time scale whereas long-term impacts involve planning for peak load periods. short-term wind integration impacts involve system balance related issues, which are represented by requirements and costs related to the fluctuating wind power. the main impacts of wind power on power systems are listed below. 2.3.1 generation efficiency fluctuating wind power affects other conventional dispatchable generation units. intermittent performance of wind power leads to the fact that conventional units might operate in a suboptimal unit commitment. this problem can be reduced by accurate wind producfigure 2. wind spectrum at farm brookhaven national laboratory table 1. wind speed variability summary 27 the journal of engineering research vol. 7 no. 2 (2010) 24-31 tion forecasting and prediction. however, even with good forecasting and prediction tools, fluctuation of wind power causes less efficient operation of thermal units (ackermann, 2005). the costs of high wind penetration on thermal units have been studied in (smith et al. 2007). the cost of reduced generation efficiency on the system depends on the initial unit commitment of conventional units, wind power uncertainties, and system flexibility. 2.3.2 reserves a back up reserves for regulation (1-minute), load following (5-minute) and operating reserve (10minute) are needed to account for wind power fluctuations. the main factors that affect the amount of required reserves are initial load variations, the size of the balancing region, and the geographical dispersion of wind power resources. ontario case study (energy, 2006) presents a good example of reserve requirements as a function of wind penetration. normally, the 1-munite regulation requirement is not of concern, even with high wind penetration level, due to aggregation and spatial distribution of effect. 2.3.3 curtailed energy in some cases wind turbines' output needs to be curtailed to preserve balanced, stable and secured operation of the grid. in low load periods, system operator may consider wind power curtailment to ensure that enough dispatchable resources are online to guarantee enough reserves and ramping capabilities in order to accommodate demand's fluctuation (energy, 2006). in other cases, the operator will order wind power curtailment when wind power is high, local demand is low and transmission lines to other areas are approaching their thermal limits (ackermann, 2005). the need to curtail some wind power output depends not only on penetration level, but also on system flexibility. this flexibility can be enhanced by introducing more demand side management (dsm) or demand response (dr) programs. 2.3.4. reliability variable sources can be relied on in supplying peak demands if their patterns are correlated with that of the demand, an example would be solar power and air conditioning demand. normally, wind power has no correlation with the load in many parts of the world. for example, ontario, which is a summer peaking utility, has more wind power in winter than in summer (inc 2006). even with winter peaking utilities, there is no guarantee that wind power will be available when needed most (peak periods). policy makers should be careful in estimating capacity value of wind power. large wind power penetration will need large amounts of conventional plant to ensure supply adequacy and security (ilex and umist, 2002). additionally, the reliability of power systems that have high wind power penetration levels is affected by original load demand, the geographical dispersion of wind power resources, the available transmission capacity, and the system flexibility. 2.3.5 transmission and distribution losses wind power can increase or decrease network losses (transmission and distribution) depending on the wind power penetration level, the correlation between wind production and load profile, and the location of wind energy resources relative to load centers (quezada et al. 2006). moreover, high wind penetration levels may result in higher transmission capacity requirements; consequently, higher transmission losses. transmission congestions can occur when wind generation is away from loads (ilex and umist, 2002). 2.3.6. voltage and reactive power induction generators used in wind turbines inherently absorb reactive power from the grid. fluctuations in power output results in voltage fluctuations when using fixed speed turbines. however, modern variable speed turbines, such as doubly fed induction generators (dfigs), can provide reactive power support using appropriate interfacing (ackermann, 2005). 3. the effect of wind power on generation costs although it is technically possible to integrate a large amount of wind power, higher wind penetration levels normally results in higher integration costs. west denmark, north of germany and galicia in spain are examples of areas that have a relatively high wind penetration levels. in west denmark, which has a wind power penetration level (maximum wind production/peak demand) of 58%, about 24% of consumed energy comes from wind power generation facilities (soder et al. 2007). the general approach of evaluating wind power impacts on the system is by evaluating the physical impacts of wind on the system and calculating the consequent costs. this process involves obtaining wind data that represents the actual performance of wind power plants. each potential wind power plant location is represented by sets of virtual anemometers that are used for power production calculations (corp and inc 2004). most wind integration studies divide wind impacts into the four time frames corresponding to that of system operation: namely, regulation, load following, scheduling and unit commitment; fig. 3. the net load, which is the load minus wind production, should be 28 the journal of engineering research vol. 7 no. 2 (2010) 24-31 served to maintain system balance for all time frames. in regulation, short term reserve capacity is needed to maintain demand-generation balance, and usually performed by automatic generation control (agc) of some designated generating units within seconds to few minutes. load following is in the range of 10 minutes up to several hours, and includes morning ramp-up and evening ramp-down capabilities. in scheduling and unit-commitment, the operator objective is to ensure that sufficient generation is available for hours and days ahead, respectively, of the real-time operation. the system operator should make sure that operating reserves are sufficient to maintain system balance according to local reliability standards eg. north american electric reliability council (nerc) standards (nrec, nerc.com). initially, it was thought that operational impacts of wind generation will cost utilities about $15 to $20 per mwh of wind energy. however, recent studies concluded that additional reserve capacity needed to integrate wind generation is at most 10% of the total wind plant nameplate capacity and in most cases, between 3% and 5% (gil and joss, 2006). the results of many case studies developed in europe and north america estimated that the resulting operating costs are between $2 and $4 per mwh for low wind power penetration levels and increase up to $6 per mwh for high penetration levels. these amounts, which represent the cost balancing operations, should be reflected on integration fees designed by utilities for wind projects. regulation impacts of spatially distributed wind generation facilities are expected to be modest due to smoothing effect of uncorrelated short term wind production. for example, it has been reported that the addition of 1500 mw (10% of peak demand) and 3300 mw (15% of peak demand) of wind in the us increased the regulation requirements by 8 mw (zavadil et al. 2004) and 36 mw (energy, 2005), respectively. the load-following impacts of wind are attributed to increased net load variability compared to original load variability. all wind integration impacts studies concluded that the distribution of the net load changes flatten and expand with large scale wind integration. this means an increase in ramping requirements for longer hours of the year. as seen in table 2 (smith et al. 2004), the dominant balancing cost component is related to the unit commitment time frame in which decisions must be made about which units to start and stop and when to do so to maintain system reliability at minimum cost (holttinen, 2006). wind integration introduces uncerfigure 3. operation time frames table 2. down of reserve costs attributed to wind power 29 the journal of engineering research vol. 7 no. 2 (2010) 24-31 tainty into the day-ahead unit-commitment process. this impact results in an increase in the operating cost up to $5 to $6 per mwh of wind power production at wind power level of 20 to 30%. this increase in cost varies depending on flexibility of generating units, fuel cost, market regulations, and both load and windgeneration resources characteristics. wind power forecasting can significantly reduce costs associated with day-ahead uncertainty. smith et al. reported that the state-of-the-art commercially available forecasting tools can provide 80% of the benefits that would have been achieved from perfect forecasting (smith et al. 2004). moreover, the authors of (ummels et al. 2007) concluded that forecasting error does not have a significant impact on thermal units operating costs. 4. wind integration cost summary the results of many studies of wind integration costs that are related to balancing requirements presented in (world wind energy association, 2008; gil and joss, 2006) and (smith et al. 2007) are compiled in fig. 4. as demonstrated by the figure, these results are scattered and quite difficult to compare. the scattering is attributed to different methodologies used, data and tools used, terminology and metrics in representing the results, as well as system characteristics (holttinen, 2006). moreover, different time scales are used by different studies to estimate additional reserve requirements. additionally, some studies consider only operating costs whereas others take into consideration investments for new reserves. another reason for these discrepancies is the fact that some studies consider transmission possibilities to maintain the supplydemand balance while others do not. in spite of the scattered results, one can draw the following conclusions: * the incremental cost of balancing requirement is low at low wind penetration levels and increases with higher wind penetration levels (holttinen, 2006). at high penetration levels the cost of required reserves is significantly less when considering the net load variations (load wind) compared with wind power variability. figure 4. wind power balancing requirements related costs. (currency conversion rates 1 pound = 1.95 us$ and 1 euro = 1.45 us $) nordel finland ireland germany-greennet denmark-greennet norway-greennet sweden-greennet uk xcel colorado xcel colorado xcel minnesota ca rps wuig/xcel pacificcorp bpa we energies great river 30 the journal of engineering research vol. 7 no. 2 (2010) 24-31 * the flexibility of conventional (dispatchable) generation plays a crucial role in reducing wind power integration costs. therefore, hydro dominated countries, eg. norway, have lower costs than thermal stations dominated countries, eg. germany. the coordination of the operation of hydro generating systems with the electricity produced from wind farms opens up new possibilities to the system operators and utilities. the output of wind facilities can be stored in water reservoirs and delivered later at times when needed most (gil and joss, 2006). * available transmission capacities to neighboring systems, eg. denmark, facilitate the increase in wind penetration levels without increasing balancing requirements to a certain extent. for example, the greennet study expected balancing costs of $0.65 per mwh wind production at 67% penetration level (peak wind production / peak demand) (meibom, 2006). nevertheless, wind power integration costs increases when a neighboring country gets more wind power because exporting possibilities decrease. * to minimize integration costs, wind power production facilities should be geographically dispersed especially when transmission capacity is limited. germany has a relatively high integration costs compared with denmark, norway and sweden because of concentration of wind power production in north-west region (holttinen, 2006). conclusions due to improved wind turbines technologies and environmental concerns about other competing source of energy, there is a strong growth of wind power capacity around the globe. however, wind power is of intermittent nature. therefore, to maintain high reliability in presence of wind power, system operator should consider wind generation capacity during high and low wind conditions, system robustness to survive a fault, reserve and ramping capabilities, transmission capacity, control or down regulation of available wind power. recent studies show that the dominant wind integration cost component occurs in the unit commitment time frame due to the uncertainty and variability of wind production. however, available forecasting tools can provide about 80% of benefits that would have been gained from perfect forecasting. although wind power integration cost is highly system dependent, one can conclude that these costs increase as wind penetration level increases. the flexibility of existing dispatchable generation units and the available transmission capacity to neighboring areas play a key role in reducing integration costs. additionally, a wider geographical dispersion of wind power turbines results in a lower integration costs. acknowledgment m. h. al-badi gratefully acknowledges the financial support received from sultan qaboos university, muscat, oman, for supporting his phd program at the university of waterloo, on, canada. references ackermann, t., 2005, "wind power in power systems," chichester, john wiley & sons. corp, e. and inc, w., 2004, " characterization of the wind resource in the upper midwest," wind integration study-task 1. energy, g., 2006, ontario wind integration study a report prepared for ontario power authority (opa) independent electricity system operator (ieso) and the canadian wind energy association (canwea). energy, g., 2005, " the effects of integrating wind power on transmission system planning," reliability, and operations: report on phase 2, prepared for the new york state energy research and development authority. gil, h.a. and joss, g., 2006, "integration of wind generation with power systems in canada: overview of technical and economical impacts," canmet energy technology centre varennes cetc 2006-016. holttinen, h., 2006, " design and operation of power systems with large amounts of wind power," first results of iea collaboration, in global wind power conference adelaide, austrlia. hoven, v.d., 1957, "power spectrum of horizontal wind speed in the frequency range from 0.0007 to 900 cycles per hour," j. metrology, vol. 14, pp. 160-164. inc, h.e., 2006, analysis of future wind farm development in ontario a report prepared for ontario power authority. ilex and umist, 2002, quantifying the system costs of additional renewables in 2020. meibom, p., weber, c., barth, r. and brand, h., 2006, " operational costs induced by fluctuating wind power production in germany and scandinavia," deliverable d5b disaggregated system operation cost and grid extension cost caused by intermittent rese grid integration, greenneteu27. 31 the journal of engineering research vol. 7 no. 2 (2010) 24-31 quezada, v. and e. al. 2006, "assessment of energy distribution losses for increasing penetration of distributed generation," ieee t power syst, vol. 21, pp. 533-540. smith, j.c., milligan, m.r., demeo, e.a. and parsons, b., 2007, " utility wind integration and operating impact state of the art, ieee t power syst, vol. 22, pp. 900-908. smith, j.c., milligan, m.r., demeo, e.a. and parsons, b., 2004, "wind power impacts on electric power system operating costs: summary and perspective on work to data," in the american wind energy association global wind power conference chicago, illinois. soder, l., hofmann, l., orths, a., holttinen, h.a., wan, waya and tuohy, a.a.t.a., 2007, " experience from wind integration in some high penetration areas," ieee t energy conver, vol. 22, pp. 4-12. truewind, a., 2005, "an analysis of the impacts of large-scale wind generation on the ontario electricity system. the north american electric reliability corporation (nrec) (nrec), reliability standards, http://www.nerc.com. ummels, b.c., gibescu, m., pelgrum, e., kling, w.l. and brand, a.j., 2007, "impacts of wind power on thermal generation unit commitment and dispatch," ieee t on energy conver, vol. 22, pp. 44-51. world wind energy association, wind turbines generate more than 1 % of the global electricity, bonn 21 february 2008. zavadil, r., king, j., xiadong, l., ahlstrom, m., lee, b., moon, d., finley, c., alnes, l., jones, l., hudry, f., monstream, m. and lai, s., 2004, " wind integration studyfinal report," j. smith xcel energy, minnesota department of commerce, enernex corporation, and wind logics inc. d:\desktop\mech100102(3).xps tjer 2012, vol. 9, no. 2, 27-35 1. introduction the knowledge of the availability of solar radiation at any given location is important for the design and performance evaluation of solar drying equipment. the data on hourly, daily and monthly average solar radiation strike on horizontal and inclined surfaces are essential for designing solar drying equipment. actual measurements of insolation are generally made in only a few places because of the cost of the instruments required, and the care needed for maintaining it. simulations can be made on the basis of theoretical and statistical relationships developed for estimating ________________________________________ *corresponding author’s e-mail: basunia@squ.edu.om insolation from various weather observations, most of which are usually made at many locations. the temperature and relative humidity have great influence on the open air natural sun drying of farm crops as it is commonly practiced in bangladesh and in other developing countries. in most tropical countries, including bangladesh, the climate is characterized by hot and humid air during harvesting. this unfavorable climatic condition dictates the need for a more effective method of drying, grain. because of the difficulties with open air sun drying particularly in wet weather, an alternative and efficient drying method is needed in tropical countries. simulation of solar radiation incident on horizontal and inclined surfaces ma basunia*a, h yoshiob, and t abec adepartment of soils, water & agricultural engineering, college of agriculture and marine sciences, sultan qaboos university, p.o. box 34, al-khod 123, muscat, sultanate of oman bdepartment of biomechanical systems, ehime university, matsuyama, 790-8566, japan cdepartment of biomechanical systems, ehime university, matsuyama, 790-8566, japan received 2 january 2010; accepted 25 december 2011 abstract: a computer model was developed to simulate the hourly, daily and monthly average of daily solar radiation on horizontal and inclined surfaces. the measured hourly and daily solar radiation was compared with simulated radiation, and favourable agreement was observed for the measured and predicted values on clear days. the measured and simulated monthly averages of total (diffuse and beam) daily solar radiation were compared and a reasonable agreement was observed for a number of stations in japan. the simulation showed that during the rice harvesting season, september to october, there is a daily average of 14.7 mj/m2 of solar irradiation on a horizontal surface in matsuyama, japan. there is a similar amount of solar radiation on a horizontal surface during the major rice harvesting season, november to december, in bangladesh. this radiation can be effectively utilized for drying rough rice and other farm crops. keywords: solar energy, simulation, geometric factor, daily and hourly solar radiation 28 ms basunia, h yoshio and t abe bangladesh receives abundant solar radiation throughout the year, so a solar heated dryer can be used to solve the drying problem. most of the farmers in bangladesh operate on a small scale and cannot afford mechanically powered drying systems. an intermediate solution takes advantage of the ready availability of solar energy to utilize it for drying. it was reported (bala and zia-uddin 1990) that solar dryers can be designed to enhance the effectiveness of solar drying in bangladesh, and that there is potential for their adoption and application by small farmers. this would reduce grain loss and help to maintain the quality of dried grain. solar drying systems must be properly designed in order to meet the particular drying requirements of specific crops and to give a satisfactory performance with respect to energy requirements (steinfeld and segal 1986). the designer should investigates both theoretically and practically, the physical needs of solar crop dryers and evaluate the performance of the solar dryers based on available solar energy. this paper deals with a simulation model to be used for the simulation of hourly, daily, and monthly average radiation incidents, both on horizontal and tilted surfaces. 2. theoretical considerations 2.1 radiation geometry the geometric relationships between a plane of any particular orientation relative to the earth at any time and the incoming beam solar radiation, that is, the position of the sun relative to that plane, can be described in terms of several angles (benford and bock 1939). simplification of the geometric relationships relating to the angle of incidence of beam radiation on a surface to the other angles leads to the following equation. (1) where is the angle of incidence the angle between the beam radiation on a surface and the normal to that surface; z is the angle of incidence; is the angle between the plane of the surface considered and the horizontal, 0 180º, s is the solar azimuth angle; is the surface azimuth angle, the deviation of the projection on a horizontal plane of the normal to the surface from the local meridian, with zero due south, -180º 180º. for the horizontal surface, the angle of incidence is the zenith angle of the sun z. for this condition, = 0, and eqn. (1) becomes (2) where is the latitude, the angular location north or south of the equator; is the angular position of the sun at solar noon, called declination; is the hour angle, the angular displacement of the sun east or west of the local meridian due to rotation of the earth on its axis at 15º per hour, with morning negative and afternoon positive;. the relationship for the angle of incidence of surfaces sloped due north or due south can be derived from the fact that surfaces with slope to the north or south have the same angular relationship to beam radiation as a horizontal surface at an artificial latitude of ( ). modifying eqn. (2) yields (3) the declination can be computed from the following equation (duffie and beckman 1991): (4) where n is the day of the year. recommended average days for months and values of n by month are given by klein (1977) and are shown in table 1. equation (2) can be solved for the sunset hour angle, s, when z = 90º (5) the number of day light hours is given by following equation: (6) 2.2 geometric factor (rb) the ratio of direct (beam) radiation on the tilted surface to that on a horizontal surface is called geometric factor and it is usually expressed as: (7) cos z and cos are both determined from eqns. (2) and (3). the optimum azimuth angle for the flatplate collectors is usually zero in the northern hemisphere and 180º in the southern hemisphere. in this case eqs. (2) and (3) can be used to determine cos z and cos , respectively, leading in the northern hemisphere, ( = 0), to (8) 2.3 the theoretical (extraterrestrial) radiation on a horizontal surface the amount of extraterrestrial radiation or radiation above earth's atmosphere at any time of year is given by the following equation (duffie and beckman 1991). 29 simulation of solar radiation incident on horizontal and inclined surfaces (9) where en is the extraterrestrial radiation measured on the plane and normalized to the radiation on the nth day of the year, w/m2; and sc is the solar constant (1367 w/m2). at any given time, the extraterrestrial radiation incident on a horizontal plane outside of the atmosphere is the normal incident of solar radiation as given by eqn. (9) and divided by rb. (10) substituting the value of cos z from eqn. (2) in eqn. (10) yields e for a horizontal surface at any time between sunrise and sunset as: (11) (12) (13) the hourly extraterrestrial radiation strikes on a horizontal surface can be estimated from eqn. (12) by integrating it between hour angles 1 and 2. the integration for an hour period results in the following equation, where io is the hourly extraterrestrial radition on horizontal surface, j/m2. (14) the hourly extraterrestrial radiation can also be determined from the following approximate equation, evaluating at the mid point of the hour. (15) the daily theoretical solar radiation, h0 striking on a horizontal surface is obtained by integrating eqn. (12) over the period from sunrise to sunset. the integration yields where h0 is the daily extraterrestrial radiation, j/m2, and s is the sun set hour angle in degrees. (16) 2.4 estimation of average solar radiation on a horizontal surface the original angstrom-type regression equation relates monthly average daily radiation to radiation on a clear day, at a given location. the verage fraction of possible sunshine hours is given by (17) table 1. recommended average days for months and value of n (the day of the year) by months source: klein (1977) m d c m 30 ms basunia, h yoshio and t abe where hm is the monthly average daily radiation on a horizontal surface hc is the average clear sky daily radiation for the location and month considered a, and b are empirical constants, nd is the monthly average daily hours of bright sunshine nm is the monthly average of the maximum possible daily hours of bright sunshine i.e., the day length represented by n of the average day of the month in eqn. (6). page (1964) and others have modified the method to base it on extraterrestrial radiation on a horizontal surface rather than on clear day radiation, using equation: (18) where hom is the extraterrestrial radiation for the given location, averaged over the period considered, and a1 and b1 are constants depending on location. lof et al. (1966) developed sets of constants, a1 and b1 for various climate types and locations based on radiation data then available. for poona, india, a1 = 0.30 and b1 = 0.51 through eqn. (18). the ratio (hm/hom) is termed as the monthly average clearness index. 2.5 estimation of hourly solar radiation on a horizontal surface from daily data when hour-by-hour performance estimations for a solar system are needed, it is necessary to start with daily data and then to estimate hourly values from daily data. the ratio of hourly total radiation i to daily total radiation h on a horizontal surface is expressed as a function of day length. the ratio i/h is represented by the following equation of (collares-pereira, rabl 1979): (19a) where the parameters c1 and c2 are given by: (19b) (19c) the ratio of hourly diffuse radiation (id ) to the daily diffuse radiation (hd) on a horizontal surface is a function of time and day length. liu and jordan (1962) reported that id/hd is the same as i0/h0, and is estimated from the following equation. (20) 2.6 beam and diffuse components of hourly global radiation on a horizontal surface an hourly clearness index (kh) can be defined as (21) where i is the hourly global radiation on a horizontal surface, j/m2. the usual approach is to correlate id/i, which expresses the fraction of the hourly radiation on a horizontal plane which is diffuse, with kh, the hourly clearness index. id is the hourly diffuse component of global radiation (i). this correlation may not represent a particular hours very closely, but over a number of hour it adequately represents the diffuse fraction (duffie and beckman, 1991). orgill and hollands (1977) used data from canadian stations (erbs et al. 1982) have used data from united states and one australian station, and (reindl et al. 1990) used independent data from the united states and europe. the three correlations are essentially identical, although they were derived from three separate data sources. the correlation between hourly diffuse and global radiation developed by (erbs et al. 1982) is as follows: (22a) (22b) (22c) 2.7 beam and diffuse components of daily global radiation on a horizontal surface the daily clearness index, kd is defined as the ratio of a particular day’s radiation h on a horizontal surface, to the extraterrestrial radiation h0 for that day. hd is the daily diffuse component of the total global radiation, h on that particular day, j/m2. mathematically it is expressed as (23) the following statistical relationship was given by collares-pereira and rabl (1979): (24a) 31 simulation of solar radiation incident on horizontal and inclined surfaces (24b) (24c) (24d) 2.8 beam and diffuse components of monthly global radiation on a horizontal surface the monthly average clearness index (km) is defined as the ratio of monthly average daily radiation on a horizontal surface, hm to the monthly average daily extraterrestrial radiation h0m. hdm is the diffuse component of the total global monthly average daily radiation hm, j/m2. mathematically it can be expressed as (25) the (erbs et al. 1982) correlation between the fraction of the monthly average daily radiation hdm/hm which is diffuse, and the monthly average clearness index km is represented by the following equations: (26a) (26b) 2.9 hourly radiation on a tilted surface the radiation on the tilted surface includes three components: beam, isotropic diffuse, and solar radiation diffusely reflected from the ground (reindl et al. 1990). considering all diffusion is isotropic, the result is an equation that gives the hourly radiation on a tilted surface, it , in terms of parameters that can be determined either theoretically or empirically as is seen in eqn. 27. (27) where ib is the beam component of i, rb is the geometric factor, fvs is the view factor to the sky for a surface sloped at an angle from the horizontal, fvg is the view factor of the surface to the ground, and r is the surrounding diffuse reflectance for the total solar radiation. equation (27) can be written as (28) where it is total hourly radiation incident on a tilted surface and is the slope of the surface considered in degrees. 2.10 daily total radiation on a tilted surface if the diffuse and ground reflected radiations are each assumed to be isotropic, then in a manner similar in eqn. (27), for a given day, the radiation on tilted surface can be written as (29) where rbm is the monthly average daily geometric factor, or the ratio between the monthly average daily beam radiation on the tilted surface and that on a horizontal surface. for a surfaces with = 0, rbm is expressed as (30) where m is the sunset hour angle for the sloped surface for the mean day of the month, which is given by (31) where "min" means the smaller value of the two items in the brackets. if we divide by the monthly average daily extraterrestrial radiation h0m and substitute k = h/h0m, then eqn. (30) becomes (32) the ratio hd/h is the daily fraction of the diffuse radiation and can be found from eqn. (24) as a function of kd. therefore eqn. (32) can be used to calculate the daily total radiation on a tilted surface. for 0.17 kd < 0.75 32 ms basunia, h yoshio and t abe 2.11 monthly average daily radiation on a tilted surface for use in solar process design, the monthly average daily radiation on a tilted surface is needed. the method described by liu and jordan (1963) which is extended by klein (1977) has been widely used. if the diffuse and ground reflected radiations are each assumed to be isotropic, then in a manner similar to eqn. (28), the monthly mean solar radiation on an unshaded tilted surface can be expressed as (33) where htm is the monthly average daily total radiation on a sloped surface and hdm/hm is a function of km. 3. simulation procedure for hourly and daily radiation on horizontal and tilted surfaces the basic equations to be used are eqs. (28) and (29). the first steps are to obtain id/i, rb, hd/h and rbm. the ratios id/i and hd/h are functions of kh and kd, and can be obtained from eqns. (22) and (24), respectively. for a particular day of the year and location where n and are known, can be calculated from eqn. (4). for known values of and , sunset hour angle s and day light hours nm or n are calculated from eqns. (5) and (6), respectively. for the calculation of hourly radiation on a tilted surface, rb is needed and can be computed from eqn. (8). the calculation of rbm requires sunset hour angle of the sloped collector. equation (31) is used to calculate the sunset hour angle m. with known n and s, from eqns. (15) and (16), the hourly extraterrestrial (theoretical) radiation i0 and daily extraterrestrial (theoretical) radiation, ho, can be estimated. then with assumed daily hours of bright sunshine (nd), the nd/nm ratio is calculated to be used in eqns. (18) and to find the value of daily radiation h on a horizontal surface. from the known value of h, the value of hourly radiation i on a horizontal surface is determined from eqn. (19). so both kh and kd are known from eqns. (21) and (23), respectively. from known kh and kd ratios, id/i and hd/h ratios can be calculated from eqs. (22) and (24), respectively. id value also can be calculated from a known hd using eqn. (20). ib is the difference between i and id. is considered equal to . then eqns. (28) and (29) are used to calculate the hourly and daily radiation incident on a tilted surface. 4. simulation procedure for monthly average daily radiation on horizontal and tilted surfaces the basic equation to be used is eqn. (33). the first steps are to obtain hdm/hm and rbm. the ratio hdm/hm is a function of km and it can be obtained from eqn. (26). for the mean day of any month n is known (table 1) and can be calculated from eqn. (4). for known values and , sunset hour angle s and monthly average of the maximum possible daily hours nm or n are calculated from eqns. (5) and (6), respectively. then with assumed monthly average daily hours of bright sunshine (nd), the nd/nm ratio is calculated to be used in eqn. (18). with known n and s, the extraterrestrial (theoretical) radiation h0m is calculated from eqn. (16). the monthly average daily radiation hm on a horizontal surface is calculated from eqn. (18). then km is obtained from eqn. (25). equation (26) is used to calculate the hdm/hm ratio from km. the calculation of rbm requires sunset hour angle m. equation (31) is used to calculate sunset hour angle s. finally, global radiation on an inclined surface is calculated using eqn. (33). in this way monthly average daily radiation incidences on horizontal and tilted surfaces are calculated for the 12 months of the year. a computer program was written in basic programming language. 5. measured and simulated solar radiation data a pyranometer we used to measure the solar radiation incident on a horizontal surface for several days in october 2004 from morning to evening. the pyranometer was connected to a personal computer via a data logger (green klt 77) to collect the radiation data at one minute intervals using basic programming language. sensitivity of the pyranometer was 10.5 mv/(cal.cm2.min) (moll-goregiatic type ms-61). the model described was applied to generate total global hourly, daily, and monthly average radiation on a horizontal and tilted surfaces. the comparison between the measured total (beam and diffuse) and the simulated total (beam and diffuse) solar radiation on a horizontal surface versus time for clear days are shown in figures 1 and 2. there was favourable agreement between the measured and predicted values on a day with clear skies. for example in october 5, 2004 day, actual measured insulation on a horizontal surface was 8.04 33 simulation of solar radiation incident on horizontal and inclined surfaces mj/m2 and simulated one was 9.5 mj/m2. similarly in october 24, 2004 day, actual measured insulation was 8.6 mj/m2 and simulated one was 8.9 mj/m2. so the hourly and daily radiations incident on horizontal surfaces can be simulated from the theoretical relationships. the measured monthly average daily radiation incidents on a horizontal surface of a flat plate collector were compared for a number of stations in japan. there was favorable agreement between the measured and simulated values (figure 3). it is expected that simulated values will show the same level of accuracy with the measured data of other places. simulations show that during the paddy harvesting period from mid-september to mid-october, about 14.7 mj/(m2.day) of solar energy is available on figure 1. total (beam and diffuse) measured and simulated solar radiation a horizontal surface versus time for a clear day on october 5, 2004 figure 2. total (beam and diffuse) measured and simulated solar radiation on a horizontal surface versus time for a clear day on october 24, 2004 34 ms basunia, h yoshio and t abe a tilted surface at matsuyama, japan, and that an almost similar amount (14.1 mj/(m2.day)) of solar energy is also available during the major rice harvesting season in bangladesh from november-december. this energy can be utilized for drying rough rice and other farm crops. the model has been developed on the basis of available relations for hourly, daily, and monthly average radiation on horizontal and tilted surfigure 3. variations of measured and simulated monthly average daily global solar radiation with time of the year at three stations of japan j f m a m j j a s o n d month 25 20 15 10 5 0 table 2. simulated monthly average daily global solar radiation incident on horizontal and inclined surfaces at matsuyama, japan and bangladesh note: tilted angle is considered as equal to the latitude of respective places 35 simulation of solar radiation incident on horizontal and inclined surfaces faces. bangladesh lies in a narrow range of latitude (20º 26º n) and, therefore, the same level of solar radiation is expected to be incident throughout the country. 6. conclusions a simulation model was developed on the basis of best available relations for daily, hourly and monthly average radiation on horizontal and tilted surfaces. the simulated hourly and daily solar radiation values were compared with the measured values and a favourable agreement was observed between them. the measured monthly average daily solar radiation was compared with the simulated values and a reasonable agreement was observed for a number of stations in japan. considering the narrow range of latitude and similar terrain and weather conditions, the same level of solar radiation is expected throughout bangladesh. the amount of solar energy available during major rice harvesting season in matsuyama, japan, and bangladesh are almost the same and can be used for drying rough rice and other farm crops. references bala, bk, zia-uddin atm (1990), simulation of solar drying of rough rice. proclamation of the 1st world renewable energy congress, reading, uk 938-942. benford f, bock je (1939), transaction of the illumination engineering society 34:200-203. collares-pereira m, rabl a (1979), the average distribution of solar radiation-correlation between diffuse and hemispherical and between day and hourly insolation values. j. of solar energy 22:155-162. duffie ja, beckman wa (1991), solar engineering of thermal process. a wiley-international publication. john wiley and sons, new york. erbs dg, klein sa, duffie ja (1982), estimation of the diffuse radiation fraction for hourly, daily and monthly-average global radiation. j. of solar energy 28:293-300. klein sa (1977), calculation of monthly average insolation on tilted surfaces. j. of solar energy 19: 325-330. liu, byh, jordan rc (1962), daily insolation on surfaces tilted toward the equator. ashrae journal 3(10):53-58. liu, byh, jordan rc (1963), the long-term average performance of flat-plate solar energy collectors. j. of solar energy 7:53-59. lof gog, duffie ja, smith co (1966), world distribution of solar radiation. engineering experimentation station report 21, university of wisconsin, madison. orgill jf, hollands kgt (1977), correlation equation for hourly diffuse radiation on a horizontal surface. j. of solar energy 19:357-363. page jk (1964), the estimations of monthly mean values of daily total short-wave radiation of vertical and inclined surfaces from sunshine records for latitudes 400n-400s. proc. of the un conference on new sources of energy 4:378-385. reindl dt, beckman wa, duffie ja (1990), diffuse fraction correlation. j. of solar energy 45:1-6. steinfeld a, segal i (1986), a simulation model for solar thin layer drying process. j. of drying technology 4:536-554. mech080830.qxd the journal of engineering research vol. 6, no. 2 (2009) 33-39 1. introduction the scarcity of fossil fuels, increasing their costs and increasing the environmental pollution from fossil fuel combustion will make renewable energy sources as more attractive one. agricultural residue is one form of biomass, which is readily available but is largely not utilized in energy recovery schemes. pyrolysis is one of thermo chemical process which converts the solid biomass in to liquid (bio-oil), gas and solid. the liquid product, pyrolytic oil, approximates to biomass in elemental composition, and is composed of a very complex mixture of oxygenated hydrocarbons. it is useful as a fuel, may be added to petroleum refinery feed stocks or upgraded by catalysts to __________________________________________________ *corresponding author’s e-mail: ganapathy_sundar@yahoo.com produce premium grade refined fuels, or may have a potential to be used as a chemical feed stocks. bio-oils are generally preferred products because of their high calorific value, the gross calorific value of the bio oil obtained from pyrolysis of rapeseed oil cake (suat ucar and ahmet, 2008), empty fruit bunches (abdullah and gerhauser, 2008) and cashew nut shell (das et al. 2004) were 33, 36 and 40 mj/kg respectively, their ease of transportation and storage, their low nitrogen and sulphur content and their opportunity to be converted into chemicals. coconut shell as an agricultural residue is available abundant in india with an annual production of more than 0.94 million tons in the year of 1994 and projected with the production of more than 1.50 million tones for the year of 2010. coconut shell is more suitable for pyrolysis process, since they contain less amount of ash, more amount of pyrolysis of coconut shell: an experimental investigation e. ganapathy sundaram*a and e. natarajanb *adepartment of mechanical engineering, velammal engineering college, chennai 600 066, tamilnadu, india binstitute for energy studies, college of engineering, anna university, guindy, chennai 600 025, tamilnadu, india received 30 august 2008; accepted 20 january 2009 abstract: fixed-bed slow pyrolysis experiments of coconut shell have been conducted to determine the effect of pyrolysis temperature, heating rate and particle size on the pyrolysis product yields. the effect of vapour residence time on the pyrolysis yield was also investigated by varying the reactor length. pyrolysis experiments were performed at pyrolysis temperature between 400 and 600°c with a constant heating rate of 60°c/min and particle sizes of 1.18-1.80 mm. the optimum process conditions for maximizing the liquid yield from the coconut shell pyrolysis in a fixed bed reactor were also identified. the highest liquid yield was obtained at a pyrolysis temperature of 550 °c, particle size of 1.18-1.80 mm, with a heating rate of 60 °c/min in a 200 mm length reactor. the yield of obtained char, liquid and gas was 22-31 wt%, 38-44 wt% and 30-33 wt% respectively at different pyrolysis conditions. the results indicate that the effects of pyrolysis temperature and particle size on the pyrolysis yield are more significant than that of heating rate and residence time. the various characteristics of pyrolysis oil obtained under the optimum conditions for maximum liquid yield were identified on the basis of standard test methods. keywords: slow pyrolysis; coconut shell; recycling; biomass »ñjôœ åëh :~æ¡dg rƒl iô°û≤d …qgô◊g ∫óëfe’g ¿élgîf .eg h ωq~æ°s »kéhéféz .eg áá°°uuóóÿÿgg≈∏y ä骫°ù÷g ºémh ,iqgô◊g ∫~©eh ,∫óëf’g iqgôm álqo òkcéj ~j~– πlcg øe ¢uén πyéøe i~mh ‘ ~æ¡dg rƒl iô°û≤d å£ñdg ∫óëf’g ≈∏y üqééàdg agôlcg ” : iqgô◊g álqo ‘ …qgô◊g ∫óëf’g üqéœ agôlg ” ~bh .πyéøÿg ∫ƒw ò«¨j ≥jôw øy ∂dph ∫óëf’g gòg œéf ≈∏y qéîñ∏d áeébe’g âbh òkéj á°sgqo ” ∂dòch . áœéædg á∏¨dg ±hô¶dg ~j~– ” ~bh .º∏e 1^8 1^18 oh~m ‘ ä骫°ù÷g ωéémcg ™eh ,á≤«bo πµd ájƒäe álqo 60 …hé°ùj âhék úî°ùj ∫~©ãh ,ájƒäe álqo 600 ¤g 400 úh ée ¬«∏y ∫ƒ°ü◊g ” ~æ¡dg rƒl øe ~féy πfé°s ≈°übg ¿g ~lh ~bh .~æ¡dg rƒl iô°û≤d …qgô◊g ∫óëf’g aéækcg ~æ¡dg rƒl øe ~fé©dg πfé°ùdg øe q~b ≈°übg ≥«≤– πlg øe ¬«déãÿg ¿eg .º∏e 200 πyéøÿg ∫ƒw h ,á≤«bo πµd ájƒäe álqo 60 úî°ùàdg ∫~©eh .º∏e 1^8 1^18 oh~m ‘ ä骫°ù÷g ºémh ,ájƒäe álqo 550∫óëf’g álqo âféc ée~æy ¤g èféàædg ò°ûjh .áø∏àıg …qgô◊g ∫óëf’g ±hôx â– ‹gƒàdg ≈∏y %30 33 h %38 44 ,%22 31 éjhé°ùe ¿éc ré¨dg h ,πfé°ùdg h ,∫óëf’g øe π°üëàÿg ~fé©dg âjõ∏d ¢üfé°üÿg ∞∏àfl ¿g éªc .áeéb’g âbhh úî°ùàdg ∫~©e øe òãµh ᫪gg ìcg »g ∫óëfc’g á«∏ªy ~féy ≈∏y ä骫°ù÷g ºémh …qgô◊g ∫óëf’g iqgôm álqo qékbg ¿g .qéñàn’g ¥ôw qé«©e ¢sé°sg ≈∏y ég~j~– ~b ” πfé°ùdg øe q~b ≈°übcg ≥«≤ëàd ±hô¶dg π° ag πx ‘ ∫óëf’g øy áœéædg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .ájƒ«◊g á∏àµdg ,ôjh~àdg ioéyg ,~æ¡dg rƒl iô°ûb ,å£ñdg …qgô◊g ∫óëf’g : 34 the journal of engineering research vol. 6, no. 2 (2009) 33-39 volatile matter and available with lower cost in rural areas during all the sessions of the year. its higher fixed carbon content leads to the usage of solid obtained from the pyrolysis process as an activated carbon material for water treatment purpose. therefore, a practical method should be applied to pyrolyse coconut shell for providing the valuable product. ozlem onay et al. (2006) pyrolyzed rapeseed in a free fall reactor under nitrogen flow to determine the role of pyrolysis temperature and sweeping gas flow rate on the bio-oil yield at atmospheric pressure. esin apaydin-varol et al. [(2007) investigated the effects of temperature on the product yields and its composition of pistachio shell in a slow pyrolysis with fixed bed reactor. ozlem onay (2007) performed fast pyrolysis of safflower seed to identify optimum process conditions for maximizing the biooil yield in a fixed bed reactor, the optimum process conditions were: final pyrolysis temperature 600°c, sweeping gas flow rate 100 cm3/min and heating rate 300 °c/min. fixed-bed slow and fast pyrolysis experiments have been conducted on a sample of rapeseed to investigate the effects of heating rate, pyrolysis temperature, particle size and sweep gas velocity on the pyrolysis product yields and its chemical compositions by ozlem onay et al. (2004). sevgi sensoz et al (2008) pyrolyzed safflower seed press cake in a fixed-bed reactor to determine the effects of pyrolysis temperature, heating rate and sweep gas flow rates on the yields of the products using pyrolysis temperatures between 400 and 600 °c with a heating rates of 10, 30 and 50 °c/min. ayse e. putun et al (2002) conducted the slow pyrolysis of soybean cake in a fixed-bed reactor under three different atmospheres: static, for determining the effects of pyrolysis temperature and particle size, nitrogen and steam for determining the product yields. pyrolysis of corn stalks with batch wise in a laboratory captive sample reactor wire mesh at atmospheric pressure was carried out by zabaniotou et al (2007) by varying the temperature (470-710 °c) with an average heating rate of 60 °c/s and a reaction time of 0.2 s. the effect of the process conditions such as heating rate, temperature and particle size on the product distribution, gas composition and char reactivity of olive waste and straw at high temperature (800 -1000°c) in a free-fall reactor at pilot scale was carried over by rolando zanzi et al (2002). olive bagasse was pyrolyzed in a fixed bed reactor (sensoz et al. 2006). alberto j. tsamba et al (2006) studied the pyrolysis characteristics and global kinetics of coconut and cashew nut shells. for the coconut shell, the grater mass loss was reported at the temperature interval of 280-415 °c at the heating rate of 10°c/min. the literature study shows, the yield and composition of pyrolysis products greatly depends on the reaction parameters: temperature, particle size of the fuel, heating rate and residence time. presently, there are limited researches that reported specifically about the effect of temperature, heating rate, particle size and residence time on the product yields of coconut shell pyrolysis. the optimum process conditions to produce highest liquid yield from coconut shell are also not available in the literature. yun ju hwang et al (2008) extracted high capacity disordered carbons from coconut shells as anode materials for lithium batteries. afrane et al. (2008) and gratuito et al. (2008) produced activated carbons from the coconut shells under the physical methods. dinesh mohan et al (2008) studied the wastewater treatment by the low cost activated carbons derived from coconut shells and coconut shell fibers. amudaa et al. (2007) produced a highly effective adsorbent material from the coconut shell combined with aquatic waste for the removal of heavy metal from industrial wastewater. wei su et al (2006) prepared micro porous activated carbon from raw coconut shell. furthermore, the former pyrolysis systems, the pyrolysis process were carried out under the flow of inert gas medium and the residence time was varied by means of inert gas flow rate. hence, some experiments of pyrolysis of coconut shell with the effect of different process parameters on the yield of liquid, gas and solid yield are described in this paper. the experiments were conducted with the aim of determining the optimum process conditions to maximize the liquid yield from the slow pyrolysis of coconut shell in a fixed bed reactor. 2. materials and methods 2.1. raw material-coconut shell samples of coconut shell in the present study were originated from pollachi, coimbatore district, located in the tamilnadu, india. immediately after getting, the coconut shells were sun dried for few days to remove the moisture content, after which they were ground in a high speed rotary cutting mill and screened by standard sieve (is designation 460-1962) separator to give fractions of less than 0.15 mm, 0.15-0.30 mm, 0.30-0.60 mm, 0.601.18 mm and 1.18-1.80 mm. the fixed carbon, volatiles and ash present in the sample were determined by astm standards. the elements carbon, hydrogen, nitrogen, oxygen and sulphur content on the sample were also identified by the astm standards. the calorific value and moisture content of the sample were measured with astm d 5865:2007 and astm d 3173:2003 standards respectively. the properties of sample, components and elements are given in table 1. components volatiles 72.93% fixed carbon 19.48% ash 0.61% elemental analysis a (%) carbon 53.73% hydrogen 6.15% oxygen b 38.45% nitrogen 0.86% sulphur 0.02% calorific value (mj /kg) 20.88 moisture content (%) 6.98% table 1. components and elemental analysis of coconut shell a weight percentage on dry basis b by difference 35 the journal of engineering research vol. 6, no. 2 (2009) 33-39 2.2. experimental apparatus and procedure as shown in fig. 1, the 316 stainless steel tubular reactors with a length of 200 mm and 300 mm, and an internal and external diameter of 27 mm and 33 mm were used to conduct the experiments, respectively. in each run, a 15 g of the sample biomass was placed inside the reactor then the reactor was placed inside the electric furnace with the capacity of 2 kwh, which was specifically designed and manufactured for this study. the reactor was heated externally by an electric furnace. the temperature inside the reactor was measured by a cr al: k type thermocouple. the temperature and heating rate was controlled by an external pid (proportional-integral-derivative) controller. a 6.5 mm id and 9 mm od stainless steel pipe was used to connect the reactor and glass condenser. the connecting pipe between the reactor and the condenser system was maintained at 200°c to avoid condensation. the condensable liquid products (bio-oil + water) were collected in a 300 mm length coil type glass condenser. the temperature of the condenser was maintained at 25°c by circulating the water in the condenser. after each experiment, the condensed liquid was collected into the sample bottle and the liquid weight was calculated by the weight difference of sample bottle before and after the liquid was collected. after pyrolysis, the solid residue was removed and weighed. then the gaseous phase was calculated from the material balance. the biomass sample input, liquid and solid char were measured by the electro balance weighing machine with an accuracy of +/ 0.01 g. all the yields were at an average yield of at least three with an experimental measurement error of less than ± 1.5%. the experiments performed in the fixed-bed reactor were carried out in four groups. in the first, to determine the effect of the pyrolysis temperature on the coconut shell pyrolysis yields, a 15 g of the sample with the particle size of 1.18-1.80 mm was placed in the 200 mm length reactor then the reactor was placed inside the furnace and the temperature was raised at 60 °c/min to a final temperature of either 400, 450, 500, 550, or 600 °c and held for until no further significant release of gas was observed. the second group of experiments was performed in order to determine the effect of particle size on the pyrolysis yields. the experiments were conducted at particle sizes less than 0.15 mm, 0.15-0.30 mm, 0.30-0.60 mm, 0.60-1.18 mm or 1.18-1.80 mm. for all these experiments, the heating rate, the final pyrolysis temperature (based on the results of first group of experiments) and the reactor lengths were 60°c /min, 550 °c and 200 mm respectively. the third group of experiments was performed in a 200 mm length reactor to determine the effect of heating rate on the pyrolysis yield. at first, 15 g of sample was placed in the reactor then the reactor was placed inside the furnace and the temperature was raised at either 20 °c/min, 40°c/min or 60 °c/min by pid controller. for all these experiments, the final pyrolysis temperature and particle size were 550 °c and 1.18-1.80 mm respectively, based on the results of the first and second group of experiments. finally, the effect of residence time on the pyrolysis yield was determined by varying the reactor lengths. the experiments was performed with the reactor lengths of 200 mm and 300 mm. based on the first three groups of experiments the final pyrolysis temperature, particle size and heating rate were maintained at 550 °c, 1.18-1.80 mm and 60 °c/min respectively. the pyrolysis oil obtained at optimum condition was tested for its properties and functional groups. fourier transform infrared spectroscopy (ft-ir) analysis was performed on a perkin elmer spectrum device with a resolution of 1.0 cm-1 for finding the functional groups present in the bio oil. 3. results and discussion 3.1 effect of temperature on the pyrolysis yield figure 2 presents the product distribution from slow pyrolysis of coconut shell at a pyrolysis temperature of 400 600 °c. the liquid and gas yield were increased from 38 to 43 wt% and 30 to 33 wt% respectively, when the pyrolysis temperature was increased from 400 to thermocouple reactor furnace condenser thermocouple oil figure 1. pyrolyser experimental setup 36 the journal of engineering research vol. 6, no. 2 (2009) 33-39 600°c, whereas the solid yield significantly decreased from 32 to 22 wt%. the decrease of liquid yield and increase of gas yield were observed at higher temperature. the decrease in liquid and gas yield at lower temperature is due to incomplete decomposition of the shell (pyrolysis is not complete). decrease in liquid and char yield and increase in gas yield at higher temperature is due to not only the secondary cracking of pyrolysis vapor but also solid char. similar results were observed in the fixed bed slow pyrolysis of rapeseed at 30°c/min in which the liquid yield increased from 41 to 47 wt% when the pyrolysis temperature increased from 400°c to 550 °c (ozlem and kocker, 2004), but the liquid yield of safflower seed press cake and soybean cake were found to be 30 to 33 wt% and 26 to 30 wt% respectively at the same temperature range (sensoz and angin, 2008; putin et al. 2002). this similarity and variation of the liquid yield could be due to the variation of the components present with individual biomass. 3.2. effect of particle size on pyrolysis yield from fig. 3, the solid and liquid yields are significantly increased as the particle size is increased from less than 0.15 mm to1.80 mm and gas yield is found to decrease. the study found that solid and liquid yield increases from 23 to 26 wt% and 37 to 43 wt% respectively and gas yield decreases from 39 to 30 wt% when the particle size is increased from less than 0.15 to 1.18-1.80 mm. the increases in the solid yield with the increasing particle size for the sample biomass could be due to greater temperature gradient inside the particles. so that at a given time the core temperature is lower than that of the surface, which possibly gives rise to an increase in the solid yield (zanzi et al. 2002). decrease in particle size leads to decreases of liquid yield because the residence time of the volatiles in the reactor is longer, favor the cracking of hydrocarbons, when smaller particles are used (sensoz et al. 2000). more residence time of volatiles inside the reactor leads to cracking of heavier molecules (tar) in to lower molecules at lower particle size ranges and it results in increase of gaseous product. there is a significant effect of particle size on the char, oil and gas yield in the pyrolysis of olive bagasse and pine sawdust (sensoz et al. 2006; wei et al. 2006). 3.3. effect of heating rate on pyrolysis yields figure 4 shows the pyrolysis product yields of coconut shell at different heating rates of 20, 40 and 60 °c/min with a constant pyrolysis temperature of 550 °c and particle size of 1.18-1.80 mm in a 200 mm length reactor. the liquid yield was low at lower heating rate and increased with increase of heating rate; the liquid yield was 40 wt% at the heating rate of 20 °c/min and increased to 43 wt% when the heating rate increased to 60°c/min. the gas yield also increased with increase of heating rate, but the char yield decreased with the increase of heating rate. the char yield was decreased from 29% to 25 wt% when the heating rate increased from 20 °c/min to 60 °c/min (fig. 4). the increase of the oil yield with the increase of heating rate may be due to higher heating rates that break the heat and mass transfer barriers in the particles (haykiri et al. 2006). the cracking of the pyrolysis vapors at higher heating rates leads to increase of gas yield. the same trend was predicted by other researchers (sensoz 2003; sensoz and angin, 2007; onay 2007 and karaosmanoglu et al. 1999) on pine bark, safflower seed press cake, pistacia khinjuk seed and straw and stalk of the rapeseed plant. 3.4. effect of residence time on pyrolysis yield the effect of residence time on the pyrolysis yield was investigated by varying the reactor length. figure 5 shows the influence of residence time on the pyrolysis yield. the gas yield from 300 mm reactor was higher than that from the 200 mm length reactor. the liquid yield from 200 mm reactor was higher than that from the 300 mm reactor. this figure 2. yield of pyrolysis products at various pyrolysis temperatures 15 20 25 30 35 40 45 <.15 .15-.30 0.300.60 0.601.18 1.181.80 particle size (mm) p yr ol ys is y ie ld (% ) solid liquid gas figure 3. yield of pyrolysis products with different particle size 20 25 30 35 40 45 20 40 60 heating rate (°c/min) p yr ol ys is y ie ld (% ) solid liquid gas figure 4. effect of heating rate on pyrolysis yield 37 the journal of engineering research vol. 6, no. 2 (2009) 33-39 shows the opposite trends in the liquid yield. this could be due to the secondary cracking of the liquid yield in the 300 mm length reactor due to prolonged residence time of the vapours inside the reactor. the solid residue yield in 300 mm reactor was higher than that in the 200 mm reactor. 4. properties of pyrolysis oil physical properties were determined for the bio-oil using the american standards for testing and materials (astm). the physical, chemical properties and calorific value of the bio-oil obtained under optimum conditions are given in table 2. the ir spectrum of the bio-oil at optimum condition is given in fig. 6. the o-h stretching vibrations between 3200 and 3400 cm-1 indicate the presence of phenols and alcohols. the c-h stretching vibrations between 3200 and 3400 cm-1 and c-h deformation vibrations between 1350 and 1475 cm-1 indicate the presence of alkanes. the c=o stretching vibrations with absorbance between 1650 and 1750 cm-1 indicate the presence of ketones or aldehydes. the absorbance peaks between 1575 and 1675 cm-1 represent c=c stretching vibrations indicative of alkenes and aromatics. the c-h stretching and bending vibrations between 1380 and 1465 cm-1 indicates the presence of alkane groups in pyrolysis oils derived from biomass. absorptions between 1300 and 900 cm-1 indicates the carbonyl components (ie. alcohols, esters, carboxylic acids or ethers). the absorbance peaks between 900 and 690 cm-1 represent o-h stretching vibrations indicative of aromatic groups. properties bio-oil elemental (wt %) carbon 75.4 hydrogen 11.7 nitrogen 2.4 oxygen 10.5 density (kg/m 3) 1090 viscosity, 50°c (cst) 36 flash point (°c) 80 higher heating value (mj/kg) 38.6 table 2. properties of bio-oil under optimum conditions 3 5 0 0 3 0 0 0 2 5 0 0 2 0 0 0 1 5 0 0 1 0 0 0 5 0 0 0 2 0 4 0 6 0 8 0 1 0 0 1 3 6 8 1 0 2 4 7 5 7 1 1 1 5 1 464 12661 713 285 5t ra n sm itt a n ce % w a ven um b er (cm -1 ) 3418 2933 1 515 figure 6. ft ir spectra of pyrolysis oil comparison of liquid yield in 200 mm and 300 mm sectors l iq ui d y ie ld ( % ) particle size (mm) particle size (mm) so lid y ie ld ( % ) particle size (mm) g as y ie ld ( % ) figure 5. effect of residence time on pyrolysis yield 38 the journal of engineering research vol. 6, no. 2 (2009) 33-39 5. conclusions in this study, pyrolysis experiments of the coconut shell are carried out in a fixed bed reactor with different pyrolysis temperature, particle size, heating rates and reactor length. the bio oil obtained at the optimum parameters are tested for finding its properties. the following are the findings from pyrolysis experiments of coconut shell in the fixed bed reactor: . the optimum process conditions for maximizing the liquid yield of slow pyrolysis of coconut shell in a fixed bed are: pyrolysis temperature 550 °c, particle size 1.18-1.80 mm, heating rate 60 °c/min and the reactor length 200 mm. the maximum liquid yield of 45 wt% was obtained at the optimum condition. . employing higher particle size results in more amounts of liquid and solid yield and less amount of gas yield compared to lower particle size. the gas yield decreases from 40 to 30 wt % when particle size is increased from less than 0.15 to 1.8 mm. . the liquid yield increased from 40 to 43 wt% when the heating rate was increased from 20 °c/min to 60°c/min. the effect of heating rate on the pyrolysis yield is not significant compared to other parameters in this study. . the rapid devolatilization of cellulose and hemi cellulose increased the pyrolysis conversion from 69 to 88 wt% when the temperature was increased from 400 to 550°c. . the maximum liquid yield of 43 wt% and 35 wt% were obtained at reactor length of 200 mm and 300 mm respectively. . the elemental composition of bio-oil is found to be better than that of feedstock. the heating value of the bio-oil is found to be similar to that of diesel fuel. ftir analysis showed that the bio-oil composition was dominated by oxygenated species. the high oxygen content is reflected by the presence of mostly oxygenated fractions such as carboxyl and carbonyl groups produced by pyrolysis of the cellulose and phenolic and methoxy groups produced by pyrolysis of the lignin. references abdullah, a. and gerhauser, h., 2008, "bio-oil derived from empty fruit bunches fuel," vol. 87, pp. 26062613. afrane, g. and osei-wusu achaw., 2008, "effect of the concentration of inherent mineral elements on the adsorption capacity of coconut shell-based activated carbon," bioresource technology; vol. 99, pp. 6678-6682. alberto, j., tsamba., weihong yang. and wlodzimierz blasia., 2006, "pyrolysis characteristics and global kinetics of coconut and cashew nut shells," fuel processing technology, vol. 87, pp. 523-550. amuda, o.s., giwa, a.a. and bello, i.a., 2007, "removal of heavy metal from industrial wastewater using modified activated coconut shell carbon," biochemical engineering journal, vol. 36, pp. 174181. das, p., sreelatha, t. and anuradda, g., 2004, "bio oil from pyrolysis of cashew nut shell-characterization and related properties," biomass and bioenergy, vol. 27, pp. 265-275. dinesh mohan., kunwar, p. singh and vinod, k. singh., 2008, "wastewater treatment using low cost activated carbons derived from agricultural by products a case study," journal of hazardous materials, vol. 152, pp. 1045-1053. gratuito, m.k.b., panyathanmaporn, t. and chumnanklang, r.a., 2008, "production of activated carbon from coconut shell: optimization using response surface methodology," bioresource technology; vol. 99, pp. 4887-4894. haykiri-acman, h., yaman, s. and kucukbayrak, s., 2006, "effect of heating rate on the pyrolysis yields of rapeseed," renewable energy, vol. 31, pp. 803810. karaosmanoglu, f., tetik, e. and gollu, e., 1999, "bio fuel production using slow pyrolysis of the straw and stalk of the rapeseed plant," fuel processing technology, vol. 59, pp. 1-12. onay, o., 2007, "fast and catalytic pyrolysis of pistacia khinjuk seed in a well-swept fixed bed reactor," fuel, vol. 86, pp. 1452-1460. onay, o., 2007, "influence of pyrolysis temperature and heating tate on the production of bio-oil and char from safflower seed by pyrolysis, using a wellswept fixed-bed reactor," fuel processing technology, vol. 88, pp. 523-531. onay, o. and kochar, o.m., 2006, "pyrolysis of rapeseed in a free fall reactor for production of bio oil," fuel, vol. 85, pp. 1921-1928. ozlem, o. and kockar, o.m., 2004, "fixed-bed pyrolysis of rapeseed," biomass and bioenergy , vol. 26, pp. 289-299. putun, a.e., apaydin, e. and putun, e., 2002, "bio-oil production from pyrolysis and steam pyrolysis of soybean-cake," product yields and composition, energy, vol. 27, pp. 703-713. sensoz, .s. and angin, d., 2008, "pyrolysis of safflower seed press cake: part 1. the effects of pyrolysis parameters on the product yields," bio resource technology, vol. 99(13), pp. 5492-5497. sensoz, s., 2003, "slow pyrolysis of wood barks from pinus brutia ten.and product compositions," bio resource technology, vol. 89, pp. 307-311. sensoz, s., angin, d. and yorgun, s., 2000, "influence of particle size on the pyrolysis of rapeseed (brassica napus l.): fuel properties of bio-oil," biomass and bioenergy, vol. 19, pp 271-279. 39 the journal of engineering research vol. 6, no. 2 (2009) 33-39 sensoz, s., demiral, i. and gercel, h.f., 2006, "olive bagase pyrolysis," bio resource technology, vol. 97, pp. 429-436. sensoz,s. and angin, d., 2007, "pyrolysis of safflower seed press cake: part 1. the effects of pyrolysis parameters on the product yields," bioresource technology, vol. 99(13), pp. 5492-5497. suat ucar., ahmet., r. and ozkan., 2008, "characterization of products from the pyrolysis of rapeseed oil cakk," bioresource technology, vol. 99, pp. 8771-8776. varol, e.a., putun, e. and putun, a.e., (2007), slow pyrolysis of pistachio shell," fuel, vol. 86, pp. 18921899. wei, l., shaoping, l., zhang, l. and zhang, h., 2006, "characteristic of fast pyrolysis of biomass in a free fall reactor," fuel processing technology, vol. 87, pp. 863-871. wel su, li z. and yapping, z., 2006, "preparation of micro porous activated carbon from raw coconut shell by two-step procedure," chinese journal of chemical engineering, vol. 14, pp. 266-269. yun ju hwang, jeong, s.k. and shin, j.s., 2008, "high capacity disordered carbons obtained from coconut shells as anode materials for lithium batteries," journal of alloys and compounds, vol. 448, pp. 141147. zabaniotou, a., ioannidou, o., 2007, "evaluation of utilization of corn stalks for energy and carbon material production by using rapid pyrolysis at high temperature," fuel, vol. 87(6), pp. 834-843. zanzi, r., sjostrom, k. and bjornbom, e., 2002, "rapid pyrolysis of agricultural residues at high temperature," biomass and bio energy, vol. 23, pp. 357-366. comp090615.qxd the journal of engineering research vol. 7, no. 1 (2010) 70-77 1 . introduction z source inverter overcomes the problems in the traditional voltage source and current source inverters. it employs a unique impedance network coupled with the inverter main circuit to the power source. it consists of voltage source from the rectifier supply, impedance network, three phase inverter and three phase load. the ac voltage is rectified to dc voltage by the front end rectifier. the output voltage of the rectifier is 1.35 times the dc input. the rectifier output dc voltage is fed to the impedance network, which consist of two equal inductors (l1 = l2 = l) and two equal capacitors (c1 = c2 = c). the network inductors are connected in series arms and capacitors are connected in diagonal arms as shown in fig. 1. ______________________________________________ *corresponding authors’ e-mail: s_thangaprakash@rediffmail.com the impedance network is used to buck or boost the input voltage depending on the boosting factor derived by applying the shoot through which is to be forbidden in conventional vsis. this network also acts as a second order filter and it should require less inductance and less capacitance. the inverter main circuit consists of six switches (igbts). z source inverters use a unique impedance network, coupled between the power source and converter circuit, to provide both voltage buck and boost properties, which cannot be achieved with conventional voltage source and current source inverters. the unique feature of the z source inverter is that the output ac voltage can be any value between zero and infinity regardless of dc voltage. that is, the z source inverter is a buckboost inverter that has a wide range of voltage control. modified space vector modulated z source inverter with effective dc boost and lowest switching stress s. thangaprakash*a and a. krishnanb * department of electrical engineering, sri shakthi institute of engineering and technology, tamil nadu, india b k.s.r. college of engineering, tiruchengode, tamil nadu, india received 15 june 2009; accepted 3 october 2009 abstract: this paper presents a modified control algorithm for space vector modulated (svm) z-source inverters. in traditional control strategies, the z-source capacitor voltage is controlled by the shoot through duty ratio and the output voltage is controlled by the modulation index respectively. proposed algorithm provides a modified voltage vector with single stage controller having one degree of freedom wherein traditional controllers have two degrees of freedom. through this method of control, the full utilization of the dc link input voltage and keeping the lowest voltage stress across the switches with variable input voltage could be achieved. further it offers ability of buck-boost operation, low distorted output waveforms, sustainability during voltage sags and reduced line harmonics. the svm control algorithm presented in this paper is implemented through matlab/simulink tool and experimentally verified with z-source inverter prototype in the laboratory. keywords: z source inverter, space vector pulse width modulation (svpwm), switching stress .»méàøe §¨°v céwhg ‘ (¢so) ‘ irõ©ÿgh »fé° ødg ¬éàÿéh ∫~©ÿgh (r) q~°üe ägp ácé◊g ¿éæ°tôc .cg ,¢téc gôhéμfék.¢s ¢¢üüîî∏∏ÿÿgggg ¬dƒfi q~°üe øe ∫~©ÿg »fé° ødg ¬éàª∏d io~fi iô£«°s ᫪kqézƒd åëñdg ω~≤j :(svm-z))q~°üe ´ƒf ∞ãμe ¿g áj~«∏≤àdg iô£«°ùdg äé«é«jgî°sg ‘ .zºàj ( …ƒàëjh i~mgƒdg á∏môÿg ô£«°ùe ‘ ∫~©e á«£dƒa ¬éàe ôaƒj ámî≤ÿg ᫪kqézƒ∏dg .πj~©àdg ô°tdƒe ᣰsgƒh ¬«∏y iô£«°ùdg ºààa äélôıg éeg ölgƒdg áñ°ùf ¬hô°v ‘ ¬«∏y iô£«°ùdg ®éø◊g ºàjh (¢so)§hq á«£dƒa äón~e ≈∏y ∫ƒ°ü◊g ºàj iô£«°ùdg øe á≤jô£dg √òg ‘ .ájô◊g øe úàlqo ≈∏y …ƒà– áj~«∏≤àdg ägô£«°ùÿg éªæ«h i~mgh álqo øe ájôm ≈∏y •ƒñg aéækg √rgdƒe ^álôıg äélƒÿg πμ°t ‘ π«∏b √ƒ°ûj ^ õjõ©àdg á«∏ªy ≈∏y iq~≤dg ≈£©j ∂dp ¤g áaé°v’éh.iò¨àe á«£dƒa äón~e ™e í«jéøÿg èy »£dƒa §¨°v céwhcg ≈∏y iô£«°ùdg á«erqgƒn ¿g .§ÿg ωéé°ùfg ‘ ∞©° dgh á«£dƒødgsvm ∫ón øe égò«øæj ” ~b åëñdg gòg ‘ áe~≤ÿg matlab/simulink∫ón é«ñjôœ ég~«ccéj ”h )´ƒf q~°üã á°ùcéyz.èàıg ‘ ( áá««mmééààøøee ääé骪∏∏cc)´ƒf q~°üã ¬°ùcéy :z ∫~©e ¢vôy hp »fé° a ¬éàe ´ƒf ¢ ñf, ( (svpwm)»méàøe §¨°v 71 the journal of engineering research vol. 7, no. 1 (2010) 70-77 several control methods for the z source inverter have been developed since the z source inverter was proposed in 2002. in these control strategies, the capacitor voltage is controlled by the shoot-through duty ratio and the output voltage is controlled by the modulation index, respectively. both of the two controllers are designed separately, thus the whole system stability is not guaranteed. furthermore, it cannot make full use of the dc input voltage, which results in low control voltage margin and quite high voltage stress across the switches. the voltage boost is inversely related to the shoot through duty ratio, therefore the ripple in shoot through duty ratio will result in ripple in the current through the inductor as well as the voltage in the capacitor. when the output frequency is low, the inductor current ripple becomes significant. all the above problems are addressed by different control techniques with various control algorithms. to address the different requirements mentioned above simultaneously, this paper presents a modified voltage vector for the z source inverters based on space vector pwm techniques, which is composed of the traditional voltage vector and the unique boost factor of the z source inverter. proposed modified control algorithm provides full utilization of the dc link input voltage and lowest voltage stress across the switches when the variable input voltage is applied. 2. operation of z source inverter the operating principle and conventional control methods of the z source inverter are discussed in (fang zheng peng, 2003). the traditional three-phase voltage source inverter has six active vectors in which the dc voltage is impressed across the load and two zero vectors when the load terminals are shorted through either the lower or upper three devices, respectively. however, three phase z source inverter bridge has one extra zero state called shoot through, when the load terminals are shorted through both the upper and lower devices of any one phase leg, any two phase legs, or all three phase legs. this shoot-through zero state is forbidden in the traditional voltage source inverter, because it would cause a shoot-through. the z source network makes the shoot-through zero state efficiently utilized throughout the operation. equivalent circuit of the z source inverter could be drawn as shown in fig. 2. diode is inserted in the equivalent circuit to represent the unidirectional power flow. it has three operating modes, namely traditional zero mode, shoot through mode and nonshoot through mode. assuming that the inductors l1 and l2 and capacitors c1 and c2 have the same inductance (l) and capacitance (c) respectively, from the symmetry of the z source network and equivalent circuit we have, (1) mode i the inverter bridge is under shoot through state for an interval t0 , during a sampling period ts. the diode in the equivalent circuit will be reverse biased in this case. the voltage across the impedance elements could be related as, (2) where vd the dc voltage before the impedance is network and is the dc link voltage of the inverter. mode ii the circuit is in any one of the active vectors thus operating in one of the non shoot through state for an interval of t1, during the switching cycle ts. the diode in the equivalent circuit will be forward biased in this case. from the symmetry of the z source network and equivalent circuit we have, (3) where v0 is dc supply voltage and t = t0 + t1. mode iii the inverter bridge is in any one of the traditional zero vectors (000 or 111) thus operating in one of the traditional zero vector mode. the average voltage of the inductors over one switching period (ts) should be zero in steady state, from (2) and (3) we have, or dcd-voltage source z source 3-phase inverter to ac load or motor figure 1. z-source inverter system figure 2. equivalent circuit of z-source inverter 1 2c c cv v v  ; 1 2l l lv v v  l cv v ; 2d cv v ; 0iv  0l cv v v  ; 0dv v 0i c l cv v v v v    0 1 0( ) 0cl l s t t v v v v t      1 0 1 0 cv t v t t   72 the journal of engineering research vol. 7, no. 1 (2010) 70-77 similarly the average dc link voltage across the inverter bridge can be found as follows, the peak dc link voltage across the inverter bridge is expressed in (3) and it can be written as, (4) the basic idea of control is to turn traditional zero state into shoot-through zero state, while keeping the active vectors unchanged, thus we can maintain the sinusoidal output and at the same time achieve voltage boost from the shoot through of the dc link. to achieve good performance for both the dc boost and the ac output voltage in the z source inverter, several close-loop control methods were proposed. in these control methods, the capacitor voltage is controlled by the shoot-through duty ratio do and the output voltage is controlled by the modulation index ma respectively. the two degree-of-freedom controllers are designed separately. it has some inherent drawbacks. 3. modified modulation technique 3.1. space vector modulation svpwm method is an advanced, computation-intensive pwm method and is possibly the best among all the pwm techniques for variable speed applications. because of its superior performance characteristics, it has been finding widespread application in recent years. all the existing pwm methods have only considered implementation on a half bridge of a three phase bridge inverter. if the load neutral is connected to the center tap of the dc supply, all three bridges operate independently, giving satisfactory pwm performance. with the machine load, the load neutral is normally isolated, which causes interaction among the phases. this interaction was not considered in other pwm techniques. svpwm considers this interaction of the phases and optimizes the harmonic content of the three phase isolated neutral load. there are not separate modulation signals in each of the three phases in svpwm techniques. instead, a voltage vector is processed as a whole. therefore, it is very suitable to control the shoot-through time. figure 3 the discontinuous inner circle shows the trajectory of voltage space vectors (v) for the traditional three-phase two-level pwm inverters. (5) (6) (7) (8) 0 1 0 1 0 1 0 * 0 * (2 ) .ci i c s t t v v t v v v v t t t        0 0 0 0 1ˆ 2 1 2i c l c v v v v v v bv d        where, d0 = 1 1 0 t t t =shoot through duty ratio and b is boost factor. the voltage stress vs across the switch is equal to the peak dc-link voltage 0îv bv . therefore, to minimize the voltage stress for any given voltage gain (g = b.ma), we need to minimize b and maximize modulation index (ma) with the restriction of that their product is the desired value. on the other hand, we should maximize b for any given modulation index ma, to achieve the maximum voltage gain. figure 3. voltage space vectors with shoot through states for z source inverter the output voltage of the inverter is determined by the different voltages between each inverter arm and the time duration in which the different voltage is maintained. eight voltage vectors 0v  , 1v  , 2v  …….. 7v  corresponding to the switching states 0s  = [0 0 0], 1s  = [1 0 0]........ 7s  =[1 1 1]respectively. 1v  , 2v  …..…. 6v  are called active vectors 0v  and 7v  are called traditional zero vectors. the length of the active vectors is unity and length of the zero vectors is zero. in one sampling interval ts, the output voltage vector of the traditional inverter v  is split into the two nearest adjacent voltage vectors. these two nearest active vectors and the traditional zero vectors are used to synthesize the output voltage vector. nv  and 1nv   (where n = 0……6) vectors are applied at times t1 and t2 respectively, and zero vectors are applied at tz times. for example in sector i, vector v  can be synthesized as, 1 2 1 2 0 7( )z s s s t t t v v v v orv t t t         1 2 cos( ) 63 st v t     2 2 3 cos( ) 23 st v t     1 2 2 (1 cos( ) ) 33 z s st t t t v t         73 the journal of engineering research vol. 7, no. 1 (2010) 70-77 3.2. modified space vector modulation for zsi for a three-phase-leg two level vsi, both continuous switching (eg. centered svm) and discontinuous switching (eg. 60 -discontinuous pwm) are possible with each having its own unique null placement at the start and end of a switching cycle and characteristic harmonic spectrum. the same strategies with proper insertion of shoot through modes could be applied to the three-phase-leg z source inverter with each having the same characteristic spectrum as its conventional counterpart. there are fifteen switching states of a three-phase-leg z-source inverter. in addition to the six active and two null states associated with a conventional vsi, the z source inverter has seven shoot-through states representing the short-circuiting of a phase-leg (e1), two phase-legs (e2) or all three phase-legs (e3). these shoot-through states again boost the dc link capacitor voltages and can partially supplement the null states within a fixed switching cycle without altering the normalized volt-sec average, since both states similarly short-circuit the inverter three-phase output terminals, producing zero voltage across the ac load. shoot-through states can therefore be inserted to existing pwm state patterns of a conventional vsi to derive different modulation strategies for controlling a three-phase-leg z-source inverter. preferably, the shoot-through states should be inserted such that equal null intervals are again maintained at the start and end of the switching cycle to achieve the same optimal harmonic performance. the middle shoot-through state is symmetrically placed about the original switching instant. the traditional switching pattern for sector-i is shown in fig.4 and modified switching pattern for sectori is shown in fig. 5. the active states {100} and {110} are left/right shifted accordingly by to/3 with their time intervals kept constant, and the remaining two shootthrough states are lastly inserted within the null intervals, immediately adjacent to the left of the first state transition and to the right of the second transition. this way of sequencing inverter states also ensures a single device switching at all transitions, and allows the use of only shoot-through states e1, e2, and e3. the other shootthrough states cannot be used since they require the switching of at least two phase-legs at every transition. the modulating signal for the modified svm strategy could be derived from the following equations, where 2 2 3 n wt n        the trajectory of voltage vector v  should be a circular while maintaining pure sinusoidal output line– to-line voltages. the boundary of the linear modulation and over modulation is the hexagon. the time duration for the active vectors are kept constant throughout the operation and the zero vector time is conveniently placed depending upon the angle of the space vector (tz is decreased when voltage vector is increased). the maximum output line-to-line voltage is obtained when the voltage vector trajectory becomes the inscribed circle of the hexagon and v  becomes 3 2 . this limitation of the length of the active vector affects the smooth operation of loads like motor drives where overdrive is desired. the continuous centered svm state sequence of a conventional three-phase-leg vsi, where three state transitions occur (e.g., null (000) active(100) active (110) null(111) ) and the null states at the start and end of a switching cycle st span equal time intervals to achieve optimal harmonic performance. with threestate transitions, three equal-interval ( 0 3 t ) shootthrough states can be added immediately adjacent to the active states per switching cycle for modulating a z source inverter where 0t is the shoot-through time period in one switching cycle. figure 4. traditional switching pattern for z source inverter figure 5. modified switching pattern for sector 1 to z source inverter 74 the journal of engineering research vol. 7, no. 1 (2010) 70-77 (9) (10) (11) (12) (13) (14) 3.3. modified voltage vector (15) further d0 could be related with the modulation index ma as, (16) (17) in boost mode of operation, shoot-through periods for shoot-through vector are acquired from the traditional zero vector and are calculated by (16). in non-boost mode of operation, z source inverter operates as a traditional vsi and the boost factor (b) constantly equals one. so the new vector, which is accomplished by both the operating modes non-boost as well as boost, could be defined as, (18) (19) for non-boost mode (20) 3.4. modified control algorithm (21) (22) max( ) maxsp offv v v t   max( ) maxsn offv v v  m ( ) mid sp id offv v v  m ( ) mid sn id offv v v t   min( ) minsp offv v v t   min( ) min 2sn offv v v t   { , } {1, 4},{3, 6},{5, 2}sp sn  where 0 3 t t  shoot through duty ratio . as aforementioned, the modulation signal produced by the conventional vsi method cannot produce voltage vector beyond 3 2 and z source inverters need two stage controllers separately for boost mode and non-boost mode. a modified control technique is presented in this section to overcome the above limitation and allow the voltage vector to be operated beyond 3 2 with single stage controller block. normally the shoot-through duty ratio is defined as follows, 0 min( ) z s t d t  ; for 0 2   2(1 ) sin( ) 33 sv t      0 1 ad m  3 2 ; 2 3 a av m m v  then 0 2 1 3 d v  ; 0 1 1 41 2 1 3 b d v     ' .v b v   and and 1 ' . 4 1 3 v v v     ; for boost mode for boost mode 'v v   ; the length of the modified voltage vector 'v  could be extended beyond 3 2 with proper placement of shoot-through time as in expressions (9) – (14). trajectory of the modified space vector v’ (continuous outer circle) and the switching pa ttern for z source inverter are shown in fig. 3 and fig. 5, respectively. as mentioned earlier, the operation of z source inverter can be divided into two modes: one is nonboost mode and the other is boost mode. in the nonboost mode, the z source inverter operates like the traditional three-phase voltage source inverter, the output voltage vector is limited in the range of ( 30 2 ), while operation in the boost mode, with the help of shoot-through zero state (vector) to boost the voltage, the z source inverter can overcome the voltage limitation, any voltage vector beyond 3 2 can be easily implemented with proper shoot-through time given in (15). in order to implement the modified control technique to have single-stage boost and non-boost controllers the algorithm could be developed as follows. during non-boost mode of operation ( ie. if the length of the modified voltage vector 'v , is less than or equal to 3 2 ) the length of the voltage vector is equal to the modified voltage vector and shoot -through time period is zero. 'v v   ; 0 0d  in the other case, d uring boost mode, ( ie. if the length of the modified voltage vector 'v  , is beyond 3 2 ) the length of the voltage vector is as follows, 1 . ' 4 ' 1 3 v v v    75 the journal of engineering research vol. 7, no. 1 (2010) 70-77 and (23) in the z source inverter system with the modified control method, only the output voltage needs to be sensed, unlike the traditional control methods that use both the output voltage and the capacitor voltage. the capacitor voltage vc and the dc-link voltage vi are dynamically regulated with the input dc voltage to assure the desired output voltage. when the dc input voltage vo is too low to output the desired voltage vref directly, the source inverter will operate in the boost mode to step up the voltage. if, v0 is high enough to produce the desired output voltage, the z source inverter will operate in the non-boost mode, just like the traditional three phase v-source inverter. by using the modified control method, output voltage can faithfully follow the reference voltage, and the voltage stress across the switches can maintain minimum regardless of the input voltage. 4. results and discussions 4.1. simulation results the traditional modulation concepts and modified control algorithm based implementation have been verified through simulation by matlab/simulink for three bridgetwo level z source inverters. the z source network has been constructed using existing laboratory components of l1 = l2 = l = 1mh and c1 = c2 = c = 1000 f. with the input dc voltage of 150v and fs = 5khz. the simulation results are shown in figs. 7-9. modulation signal is generated through the modified svpwm technique with modulation index, ma = 0.7; mf = 100; shoot-through duty ratio d0 = 0.3 and boost factor b = 2.5. figure 7. shows the modulation signal generated with the presented algorithm through modified svpwm technique. output phase voltage phase current, voltage across the z source capacitors and current through the z source inductors of the traditional control method are depicted in fig. 8 and the modified control algorithm based results are depicted in fig. 9. the output phase voltage and current of the modified control algorithm is improved significantly and the capacitor voltage (dc link voltage) is efficiently boosted and the voltage ripples across the capacitor and current ripples through the z source inductor are reduced considerably. 0 2 '2 1 3 4 ' 3 v d v v     the implementation of the proposed control algorithm is shown in fig. 6. it is clearly exposed that, the modified voltage vector 'v  is the output of the pi controller and it determines the system operating mode and corresponding d0 and v  according to the algorithm. figure 6. implementation of the control algorithm 0 0.5 1 1.5 2 2.5 x 10 4 -1 -0.5 0 0.5 1 ma gn itu de (v ) time (s) figure 7. modulation signal of proposed msvm algorithm 0 0.05 0.1 0.15 0.2 -200 0 200 time (s) o ut pu t vo lta ge ( v ) (a) 0 0.05 0.1 0.15 0.2 -20 -10 0 10 20 time (s) o ut pu t cu rr en t (a ) (b) 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0 100 200 300 time (s) c ap ac ito r v ol ta ge (v ) (c) 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 -20 0 20 time (s) cu rre nt (a ) figure 8. simulation results of the z source inverter with traditional svpwm a) output voltage b) output current c) z source capacitor voltage d) z source inductor current 76 the journal of engineering research vol. 7, no. 1 (2010) 70-77 the output phase voltage and current of the modified control algorithm is improved significantly and the capacitor voltage (dc link voltage) is efficiently boosted (shown in fig. 10) and the voltage ripples across the capacitor and current ripples through the z source inductor are reduced considerably. 4.2. experimental validation experimental verification using a three bridge-two level z source inverter prototype has been performed to validate the simulation results (shown in fig. 11). the hardware inverter is controlled digitally using a texas instruments tms320f240 digital signal processor (dsp) with composed c codes for generating the required references and a general purpose timer in the dsp for generating the common 5-khz triangular carrier needed for reference comparison. the dsp generated pwm pulses were then sent out through six independent pwm channels to gate the six switches (igbt modules) of the implemented inverter. the experimental switching line voltage and current waveform of the inverter with modulation index ma = 0.7; shoot-through duration do = 0.3; boost factor b = 2.5 for the 150 vdc with the low-pass (fc = 1khz) lc filter network is shown in fig. 12. 5. conclusions this paper presents a novel control algorithm to control the capacitor voltage (boost factor) of the z source 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 -200 0 200 time (s) o u tp u t v o lt a g e ( v ) (a) 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 -20 0 20 time (s) o ut pu t cu rr en t (a ) (b) 0 0.05 0.1 0.15 0.2 0 100 200 300 time (s) c ap . v ol ta ge ( v ) (c) 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 -20 0 20 time (s) c u rr e n t (a ) (d) figure 9. simulation results of the z source i nverter with modified svpwm a) output voltage b) output current c) z source capacitor voltage d) z source inductor current d c l i nk v ol t v s shoot t hr ough 0 100 200 300 400 500 0 0. 05 0. 1 0. 15 0. 2 0. 25 0. 3 0. 35 shoot t hr ough sv m m sv m figure 10. dc link voltage vs shoot through for traditional svm and modified svm output terminals fiber optic connnectors gate drive board power supply connector input dc voltage terminals figure 11. prototype of z source inverter figure 12. experimental switching line voltage (upper) and current waveform (lower) of the inverter d c l in k v ol ta ge ( v ) dc link volt vs shoot through shoot through 77 the journal of engineering research vol. 7, no. 1 (2010) 70-77 1) linear control of the output voltage is possible. 2) automatic changeover of operations from boost mode to buck mode and vice versa according to the variable input voltage. 3) economical and easy to implement since, the output alone is sensed to control the inverter. 4) single stage control provides one degree of freedom of controlling the output voltage. 5) it has reduced harmonics, low switching stress power and low common mode noise. references bimal k. bose, 2005, "modern power electronics and ac drives," prentice hall of india. donald grahame holmes, 1996, "the significance of zero space vector placement for carrier based pwm schemes," ieee transactions on industry applications, vol. 32, no. 5, pp. 1122-1129. dorin o neacsu, 2006, "power switching converters: medium and high power," tayler and francis group, crc press, boca raton, fl. fang zheng peng, 2003, "z source inverter," ieee trans. ind.ustry applications., vol 39(2), pp. 504-510. fang. z., peng, m., shen, a., joseph, l. m., tolbert and d. j.adams, 2004, "maximum constant boost control of the z-source inverter," in proeedings. eee ias'04, pp. 142-147. fang zheng peng, xiamoing yuvan, xupeng fang and zhaoiming qian, 2005, "z-source inverter for motor drives," ieee transactions on power electronics, vol. 20(4), pp. 857-863. holtz, j., 1992, "pulse width modulation-a survey," ieee transactions on industrial electronics, vol. 39, pp. 410-420. jin-woo and ali keyhani, 2007, "control of a fuel cell based z source converter," ieee transactions on energy conversion, vol. 22(2), pp. 467-476. keliang zhou and danwei wang, 2002, "relationship between space vector modulation and three phase carrier based pwm: a comprehensive analysis," ieee transactions on industrial electronics, vol. 49(1), pp. 186-196. poh chiang loh, mahintha vilathgamuwa, yue sen lai, geok tin chua and yun wei li, 2005, "pulse width modulation of z-source inverters," ieee transactions on power electronics, vol 20(6), pp. 1346-1355. quang-vinh tran, tae-won chun, jung-ryol ahn and hong-hee lee, 2007, "algorithms for controlling both the dc boost and ac output voltage of zsource inverter," ieee transactions on industrial electronics, vol. 54(5), pp. 2745-2750. shen, m. and f. z. peng, 2008, "operation modes and characteristics of the z-source inverter with small inductance or low power factor," ieee transactions on industrial electronics, vol. 55(1), pp. 89-96. sidney r. bowes and yen-shin lai, 1997, "the relationship between space vector modulation and regular sampled pwm," ieee transactions on industrial electronics, vol. 44(5), pp. 670-679. inverter to be operated beyond the modulation 3 2 with one degree of freedom. simulations as well as experimental results are given to validate the algorithm. this presented model greatly reduces the switching stress across the employed power switches (igbts). it advantageously utilizes the dc link voltage more effectively than the traditional svpwm methods. moreover it has the following advantages c:\users\raoh\desktop\paper_6.xps the journal of engineering research vol. 8 no 2 (2011) 49-58 deformable interfaces requires various characteristdistributions called flow regimes or flow patterns (hewitt 1982). understanding the transition from stratified flow to slug flow is important in hydrocarbon transportation systems, and has constantly stimulated the research in this direction (valluri et al. 2008). slug flow is a commonly observed pattern in horizontal and 1. introduction liquid-liquid flows appear in various industrial processes and in the petroleum industry in particular, where mixture of gas with associated liquids (oil, condensate and/or water) are produced and transported together. during their co-current flow in a pipe the _______________________________________ *corresponding author’s e-mail: lakehal@ascomp.ch transition of gas-liquid stratified flow in oil transport pipes d. lakehal*a, m. laboisa, d. caviezela, and b. belhouachib aascomp gmbh, technoparkstr. 1, zurich, ch 8005, switzerland bimperial college london, prince consort road, london sw7 2by, uk received 17 april 2011; accepted 23 september 2011 abstract: large-scale simulation results of the transition of a gas-liquid stratified flow to slug flow regime in circular 3d oil transport pipes under turbulent flow conditions expressed. free surface flow in the pipe is treated using the level set method. turbulence is approached via the les and vles methodologies extended to interfacial two-phase flows. it is shown that only with the level set method the flow transition can be accurately predicted, better than with the two-fluid phase-average model. the transition from stratified to slug flow is found to be subsequent to the merging of the secondary wave modes created by the action of gas shear (short waves) with the first wave mode (high amplitude long wave). the model is capable of predicting global flow features like the onset of slugging and slug speed. in the second test case, the model predicts different kinds of slugs, the so-called operating slugs formed upstream that fill entirely the pipe with water slugs of length scales of the order of 2-4 d, and lower size (1-1.5 d) disturbance slugs, featuring lower hold-up (0.8-0.9). the model predicts well the frequency of slugs. the simulations revealed important parameter effects on the results, such as two-dimensionality, pipe length, and water holdup. keywords: stratified flow, two-phase flow, level-set method slug 50 d. lakehal, m. labois, d. caviezel and b. belhouachi near horizontal gas liquid flows. it is the regime with large coherent disturbances, causing large pressure fluctuations and variations in the flow rates that can affect process equipment. it is also characterized by intermittent appearance of aerated liquid masses that entirely fill the pipe cross-section and travel downstream at a large velocity. earlier techniques for the prediction of slugs were based on various linear stability theories. the transition from stratified to slug flow regime has often been associated with the sudden growth of interfacial waves due to the kelvinhelmholtz instability. the onset of this instability for long waves has early been determined analytically assuming a continuous growth of a small-amplitude long wave into a slug, driven by kelvin-helmholtz instabilities (taitel and dukler 1976). wall shear and interfacial stress were later accounted for, but the longwave assumption was retained, facilitating an integral momentum balance. although the criteria for linear instability obtained in these early theoretical studies show good agreement with experimental conditions at the onset of slug formation, the underlying assumptions have been increasingly undermined by subsequent work and now seem unlikely to be justifiable. other authors (valluri et al. 2008) report observations invalidating the assumption that a single long wave develops continuously into a slug (lin and hanratty 1986). this slug formation subject is addressed here by means of large scale simulation using the cmfd code (transat of ascomp gmbh, 2010), combining the level set method for interface tracking and les/vles for turbulence. this model combination of itm and les/vles is referred to as leis, short for large eddy and interface simulation (lakehal 2010). we discuss the 3d results of flow transition to slug in two different pipe configurations under turbulent flow conditions. turbulence is approached via the les and v-les methodologies extended recently (liovic and lakehal 2007) to interfacial two-phase flows. the simulations were conducted 3d. 2. the mathematical model in transat, the incompressible two-fluid flows are represented by a single equation by a technique known as the interface tracking method (itm). for more details, the reader can refer to (lakehal et al. 2002). the equations are of the form: where u stands for the fluid velocity and p for the pressure, is the density, and is the viscosity. the source terms in the rhs of the momentum equation represents the body force, fb, the wall shear, fw, and the surface tension, fs. material properties are updated locally based on a phase marker field, denoted here by the level-set function (sussman et al. 1994). is a distance function, whose value represents the distance to the interface. the interface itself has a value of zero, is positive on one side and negative on the other. other material properties like viscosity, thermal conductivity and heat capacity are also updated in the same way. transat uses the immersed surfaces technology (ist), whereby the wall shear (fw) appears explicitly in the equations based on the solid level-set function s that defines solid obstacles (in addition to the gas-liquid function ). to track the interface and update material properties, a topology equation is solved for the level-set function : (2) conceptually, itms are capable of capturing the topology of interfaces and resolving accurately the interfacial boundary layers, independently from the reynolds number. but since full dns resolving all turbulence and interface motions is practically elusive, one is forced to solve the flow using statistical models like rans, or les, solving directly for super-grid scale and modelling unresolved or sub-grid scales (sgs). for turbulent interfacial flows, use should be made of the filtered form of the equations above (lakehal 2010; liovic and lakehal 2007). this is now known as the leis, short for large eddy & interface simulation, in which turbulent scales and interface deformations larger than the grid size are directly solved, whereas sub-scales are modelled (lakehal 2010; liovic and lakehal 2007). because statistically steady-state flow conditions are difficult to attain in 3d, recourse is made here of the v-les (instead of les), where the flow-dependent cut-off filter is larger and independent from the grid. although contradictory in terms of scale separation, one actually could resort to rans closure models to deal with turbulence as well, at the expense of affecting the degree of interface topology resolution; higher eddy viscosity levels at the interface could hamper the highfrequency surface motions, i.e. wrinkling. be it as it may, advancing the rans form of the above system of equation in time is referred to as urans. 3. the numerical approach the cmfd code (transat 2010) developed at (1) 51 transition of gas-liquid stratified flow in oil transport pipes ascomp is a multi-physics, finite-volume code based on solving multi-fluid navier-stokes equations. the code uses structured meshes, though allowing for multiple blocks to be set together. mpi and openmp parallel based algorithms are used in connection with multi-blocking. the grid arrangement is collocated and can thus handle more easily curvilinear skewed grids. the solver is pressure based (projection type), corrected using the karki-patankar technique for subsonic to supersonic compressible flows. high-order time marching and convection schemes can be employed; up to third order monotone schemes in space. multiphase flows can be tackled using (i) interface tracking techniques for both laminar and turbulent flows (level set, vof with interface reconstruction, and phase field), (ii) n-phase, phase-averaged mixture model with algebraic slip, and (iii) lagrangian particle tracking (one-to-four way coupling). as to the level set, use is made of the 3rd order quick scheme for convection, and 3rd order weno for re-distancing. mass conservation is enforced using global and local mass-conserving schemes (lakehal et al 2002). to mesh complex geometries, use is made of the immersed surfaces technology (ist) developed by implemented in the code (labois and lakehal 2011). 4. slug formation in condensation-induced waterhammer 4.1 problem description the experiment of martin et al. (2005) was designed for the purpose of investigating the phenomenon of condensation-induced waterhammer in an ammonia refrigeration system. waterhammer was initiated by introducing warm ammonia gas over static subcooled ammonia liquid placed in a horizontal carbon steel pipe 6.0 m in length. the only data used for comparison with the simulation include the isothermal subsonic case, with no condensation. the apparatus was designed to simulate an industrial environment whereby ammonia liquid is standing in a partiallyfilled horizontal pipe in thermal equilibrium with ammonia gas above it. the essential elements of the test setup consist of a horizontal pipe and a high pressure tank containing warm ammonia gas, as shown in fig. 1. the test pipe has a nominal diameter of 150 mm and length of 6 m, made of 80 carbon steel, having internal diameter of 146.3 mm, and wall thickness 11.0 mm. the pressure tank contains warm ammonia gas on top of liquid in thermal equilibrium at ambient conditions inasmuch as the entire test facility was outdoors. between the pressure tank and the test pipe were three valves and a metering orifice. the angle valve remained fully open, while flow was initiated by a solenoid valve for a given position of the throttle valve. the ammonia in the insulated test pipe was introduced from an ancillary system containing a compressor, an auxiliary tank, and another tank for purging non-condensable gases. for each test, care was exercised to transfer ammonia liquid to or from the test pipe to establish the desired depth and equilibrium temperature. instrumentation consisted of an orifice meter to determine the transient mass flow rate utilizing both upstream p0 and differential pressure transducers p . in order to calculate the mass flow rate through the orifice, which was calibrated with water, the upstream temperature was monitored by rtd1. another pressure transducer (pd) was located in the downcomer. the initial liquid temperature within the test pipe was recorded by rtd2, mounted on the bottom of the pipe. for the measurement of waterhammer or shock pressures four piezoelectric pressure transducers pcb1, pcb2, pcb3, and pcb4 were mounted as shown. the first three piezoelectric transducers were on the bottom of the pipe, while pcb4 was mounted on the pipe centerline at the end cap. in order to determine the gas pressure within the test section during the transient event four diaphragm differential pressure transducers labeled pace1, pace2, pace3, and pace4 were mounted on the top of the pipe, as shown in fig. 1. by maintaining one side of the diaphragm open to the atmosphere and the other side connected to the top of the pipe gage pressures were measured during the test. the test procedure consisted of reducing the liquid and gas temperature within the test pipe to the desired value by means of a compressor. 4.2 simulation setup a full 3d computational domain is considered in these simulations. a portion of the grid is shown in fig. 2. the pipe length is 6.3m, and the diameter is 0.14 m. the multi-block grid strategy is used to cover the domain with adjacent sub-domains (coloured differently). here boundary fitted grids were used rather than the ist. the blocks are distributed between 12 processors for mpi parallel execution. the results presented here were obtained for a grid composed by 360,000 curvilinear cells distributed over 12 blocks. the leis approach was employed here, relying on the level-set technique for interface tracking. sub-grid scale (sgs) modelling was achieved using the miles approach (lakehal 2010), where the diffusive effects of unresolved turbulence motion is left to the scheme. the inflow boundary conditions involve fixing the superficial gas and liquid velocities and the void fraction as specified in the experiments. the right-end of the cap was left open now since we deal with the nonphase case; pressure boundary conditions were used, in combination with a special scheme for the void fraction to control global and phase-specific volume conservation. specifically at the inflow, we have set the 52 d. lakehal, m. labois, d. caviezel and b. belhouachi 53 transition of gas-liquid stratified flow in oil transport pipes that is the formation of vortex shedding past the slug, immediately after sealing. this observation is corroborated by all three instantaneous velocitycomponent contours, in particular w'. the mechanism is similar to what may be expected in flows past fixed blunt bodies; the slug plays somehow the same role, as it travels with a lower velocity than the gas after sealing. the breaking of the free surface after slug sealing is also perfectly illustrated in fig. 4. again, the panels combine free-surface and velocity iso-contours. the figure shows the vigorous plunging of the breaker after sealing, pretty much similar to what we observe in plunging breakers. the third panel shows the impact of the breaker over the stratified liquid surface, a region characterized by a very high level of turbulence production. vortex shedding is again visible as in the previous panels, though the breaker seems to have affected the coherence of their motion. the slug tail and centre speeds are discussed in fig. 5 below, displaying the position of the slug versus time. a linear dependency is revealed, which is in agreement with the measurements of martin et al. (2005), who obtained an average slug speed of us = 9.4 m/s, under these conditions. the slug speed is rather constant, as it has been found in the earlier 2d simulations (results not shown here). in case of the hydrodynamic slugging, the slug velocity can be calculated from gas and liquid flow rates if the void fraction is known, in horizontal lines the mean velocity of the liquid in the body of the slug is approximately equal to the mixture velocity, or can be estimated analytically collins et al. (1978) using: (3) where um stands for the mixture velocity. the result shown in fig. 5 reveal that our leis simulation predicts the slug speed (tail and centre) in accord with the figure 3. surface displacements at various time instants figure 4. vortex shedding past the slug coloured by velocity fluctuations, highlighting wave breaking 1.201 0.532 / 1 1.2 1.4 t (s) 6 5 4 3 2 slug center (9.8t 8.33) slug tail (8.46t 6.82 collins 1978 (vs = 9.33 m/s) linear fits figure 5. slug speed (tail and centre): leis vs. analytical solution 54 d. lakehal, m. labois, d. caviezel and b. belhouachi theory (3), which gives us=9.33 m/s, and experiment of (martin et al. 2005) (us=9.4 m/s). this is an interesting result for practical applications, which shows that although a coarse grid has been used, the leis concept is capable to predict one of the most important flow features of pipe flows. 5. slug formation at the wasp facility 5.1 problem description the experiments were performed at the imperial college wasp facility with the test section mounted horizontally. gas and water were fed from two different entries perpendicular to the main pipe (fig. 6). slugs were monitored from close to the point where they were first initiated until they decayed or exited the pipe. twin-wire holdup probes were used to monitor the liquid level at a series of locations along the pipe. slugs were discriminated from large waves by measuring the velocity using cross correlation of the outputs of successive probes (the waves travel at a lesser velocity that that of the mixture and slugs travel at a velocity higher than that of the mixture). the length of the stainless steel test section is 37 m and its diameter is 77.92 mm, the pressure at the outlet is 1 atm, and the temperature is 25oc. the liquid water is introduced below a stratification plate at the bottom of the test-line and the gas is introduced above it. the superficial velocities of the two phases (air and water) are: usl = 0.611 m/s and usg = 4.64 m/s, respectively. 5.2 simulation setup use was made here of the ist technique to mesh the pipe. the pipe cad file was created using rhinoceros software, and immersed into a cartesian grid, as shown in fig. 6. the 2d simulations were performed in a pipe of length 17 m. the 3d simulations were performed in a shorter domain of 8m, consisting of 715.0 cells, then in a longer one of 16 m, consisting of 1.200.000 cells. the simulation time for the 8 m pipe simulation was 10 days on a low bandwidth dell pc (2 nodes x 4 cores; open mp parallel protocol) for 20s real time, and 53h on a high bandwidth 18 nodes ibm multicore computer (openmp protocol). the leis approach was employed here, with a filter width of 0.1d, combined with the level-set technique for interface tracking. subscale modelling of turbulence was achieved with the kmodel with filter width set equal to 0.1d (labois, lakehal 2011). the inflow boundary conditions involve fixing the superficial velocities and the void fraction, as specified in the experiment. specifically at the inflow, we have set the following values for the turbulent flow conditions: gas superficial velocity usg = 4.64 m/s; liquid superficial velocity usl = 0.611 m/s; void fraction = 50%. 5.3 simulation results (2d) figure 7 shows the measured liquid hold-up at different probe locations along the axis: 5.01 m, 5.69 m, 6.99 m and 13.32 m. the signals display distinct largewave structures developing along the pipe that could in fact be considered as slug-structures (ujang 2003). while a traditional slug is a structure blocking the cross section of the pipe completely, large-wave structures with a length scale larger than the pipe diameter can also be termed as 'slugs'. the 3d simulations discussed next will help make the distinction between the different structures. slugs or large-wave structures are captured around location x = 3 m and beyond (results not shown here). the periodicity of slug occurrence is clearly visible from these locations (x > 5 m). figure 8 depicts the calculated water holdup in 2d at different probe locations along the axis, from 5.65 to 15 m. while the signal is qualitatively similar to the measured one in terms of slug or large-wave structures intermittency, it is unclear whether slugs were indeed captured; various locations exhibit water holdup of about hl/d = 0.8-0.9. be it as it may, large surface perturbations were already captured upstream close to the inflow, at x = 0.76m, while the experiment there shows liquid hold-up not exceeding hl/d = 0.2. for the locations considered, the data and cfd provide a similar picture as to wave frequency. figure 6. computational ist grid. the cad file is immersed in a cartesian grid 55 transition of gas-liquid stratified flow in oil transport pipes 5.4 simulation results (3d) to address the effect of pipe length the flow was reproduced in pipes with different lengths: 8 m and 16 m, using the same flow conditions. figure 9 shows the development structures at different probe locations along the axis, at 5.01m, 5.695m and 6.995m. these were obtained from 3d simulations in the short pipe (l = 8 m). the distinct patterns at different locations show the variations in the slug frequencies. slug frequency decays as the location of the probe is moved further downstream. slugs or large-wave structures are predicted at downstream locations close to the pipe end: x = 7 m, in contrast to the experiment and longerpipe simulations, both indicating a shorter position for the early slugs. further, in contrast to the 2d results discussed earlier, slugs or large disturbances of the surfaces are not predicted upstream close to the inflow, but downstream. these results are interpreted later on in terms of slug frequency, and compared to the longer pipe results. turning now to the 16m long pipe, the formation of the different types of slugs is well illustrated in fig. 10. the first panel exhibits a 'large-scale slug', which, in the experiments is often referred to as 'operation slug'. this slug is formed upstream (x < 3 m) and fills entirely the pipe (hl/d = 1) with an average size of the order of 2-4 d. although the lower panels do not show a 100% water holdup filling the pipe as in the first one, the liquid structures are travelling at a higher speed than the mean flow, which makes them 'slugs'. here one observes that gas bubbles are caught inside the slug, which explains that the measured liquid hold up hl/d is less than unity (usually hl/d is between 0.80 and 0.95). figure 7. measured liquid holdup for usl = 0.61 m/s and usg = 4.64 m/s figure 8. simulated liquid holdup for usl = 0.611 m/s and usg = 4.64 m/s 56 d. lakehal, m. labois, d. caviezel and b. belhouachi the slugor large-wave structures frequency results shown in fig. 11 are qualitatively similar to the structures observed in the experiment. the lines in green correspond to the 16m pipe simulations; the red ones to the 8m simulations; both in 3d. the shift in the frequency peak observed for the two simulations is clearly due to the difference in pipe length, as the outflow boundary condition has an important impact on the flow. the frequency of the slugs is measured as a function of the abscissa. in the 16 m case, better figure 9. water holdup for the 3d simulation (short pipe: 8 m) figure 10. the formation of different kinds of slugs (long pipe: 16 m) 57 transition of gas-liquid stratified flow in oil transport pipes results are obtained as a peak frequency around 3.5 m can be seen, which is almost equal to the measured value. there is however a difference in terms of interpretation, when the frequency of slugs is evaluated based on hl/d of 0.8 or 0.85. it is clear the simulation and measurement agree best for hl/d = 0.85. moreover, the evolution of the slug frequency along the axis of the pipe is in good agreement with the data, although the result suggests that the simulation time was not enough to acquire all the slugs with lower frequency (0.3 slug/s). we thus conclude that the data. conclusions the development of large-wave structures in gasliquid 3d pipe flows has been examined using detailed cmfd simulations combining the level-set approach for interface tracking and vles for turbulence modelling. the investigation complements earlier ones (lakehal 2008) which had shown that the approach predicts more details of the flow as compared to the eulerian two-fluid model, even if it is computationally expensive. our predictions show that the itm approach can be used for a one mile pipeline, using five million cells, which require a week of simulation on a 128 cpu cluster. the 3d turbulent flows simulations clearly provide a better picture of the flow, in particular regarding the onset of slug formation and frequency, the latter being predicted in accord with the experiment. from a practical standpoint, the results show that oil transport in a portion of pipelines can be approached more rigorously. the information from detailed 3d simulations could actually be employed to calibrate coarse-grained 1d lumped parameter models, which fail under other conditions than straight pipes. these detailed simulations should be ultimately coupled with a 1d code like olga; it could then be activated in fixed critical regions (e.g. over a hill). acknowledgments this research has been undertaken within the joint project on transient multiphase flows and flow assurance. the author(s) wish to acknowledge the contributions made by the uk engineering and physical sciences research council (epsrc) and the following: ascomp, gl noble denton; bp exploration; cd-adapco; chevron; conocophillips; eni; exxonmobil; feesa; ifp energies nouvelles; institutt for energiteknikk; pdvsa (intevep); petrobras; petronas; spt group; shell; sintef; statoil and total. references collins r, moraes de, davidson jf, harrison d (1978), the motion of a large gas bubble rising through liquid flowing in a tube. j. fluid mech. 89(3):497-514. hewitt gf (1982), handbook of multiphase systems. ed. g. hestroni, hemisphere /mcgraw-hill. labois m, lakehal d (2011), very-large eddy simulation (v-les) of the flow across a tube bundle. nucl. eng. design 241:2075-2085. lakehal d (2010), leis for the prediction of urbulent multifluid flows for thermal hydraulics applications. nucl. eng. design 240:2096-2106. lakehal d (2008), large-scale simulation of stratified gas-liquid flow transition and slug in oil pipes. 6th int. conf. cfd in oil & gas, metallurgical & process industries, sintef/ntnu, trondheim norway. lakehal d, meier m, fulgosi m (2002), itm for the prediction of interfacial dynamics and mass transfer in multiphase flows. int. j. heat & fluid flow 23:242-255. figure 11. comparison of experiment and simulations of slug frequency for 2 pipe lengths; 8m and 16m. left panel: threshold hl / d = 0.8; right panel: threshold hl / d = 0.85 58 d. lakehal, m. labois, d. caviezel and b. belhouachi lin py, hanratty tj (1986), prediction of the initiation of slugs with linear stability theory. ijmf 12:79-98. liovic p, lakehal d (2007), interface-turbulence interactions in large-scale bubbling processes. int. j. heat & fluid flow 28:127-144. martin cs, brown r, brown j (2005), condensationinduced hydraulic shock laboratory study. final report, (ashrae). sussman m, smereka p, osher s (1994), a level set approach for incompressible two-phase flow. jcp 114:146-158. taitel y, dukler ae (1976), a model for flow regime transitions in horizontal and near horizontal gasliquid flow. alche j 22:47-55. transat user manual. ascomp gmbh, 2010. www.ascomp.ch/transat. ujang pm (2003), studies of slug initiation and development in two-phase gas-liquid pipeline flow. phd thesis, university of london, london, uk. valluri p, spelt pdm, lawrence cj, hewitt gf (2008), numerical simulation of the onset of slug initiation in laminar horizontal channel flow. ijmf 34:206-225. c:\users\raoh\desktop\paper_6.xps tjer 2012, vol. 9, no. 1, 46-54 1. introduction propene is an important link between relatively cheap and abundant natural gas and many petrochemicals and intermediates such as propylene oxide, acrylonitrile and isopropyl alcohol. this importance has spurred a seires of recent research efforts towards _________________________________________ *corresponding author’s e-mail: baba@squ.edu.om developing catalysts for oxidative dehydrogenation of propane as an alternative route to creating propene (mamedov, cortes 1995). the alternative route has the potential to lower reaction temperatures and pressure due to the absence of a thermodynamic equilibrium, which is a major limitation of the non-oxidative route. however, thus far the catalysts proposed for oxydehydrogenation of propane over vanadium oxide supported on kieselguhr or mcm-41 by jibril*a, a attaa, sa al-dressb, mc al-kinanyb and ha al-megrenb apetroleum and chemical engineering department, sultan qaboos university, p.o. box 33, postal code 123, al-khoud, muscat, oman bpetroleum & petrochemicals research institute, king abdulaziz city for science and technology, p.o. box 6086, riyadh 11442, saudi arabia received 7 july 2009; accepted 15 may 2011 abstract: supported vanadium oxide (5 wt%) on either kieselguhr or mesoporous mcm-41 was prepared using impregnation method and tested as a catalyst in propane oxidative dehydrogenation (pod). the catalyst samples were characterized using x-ray elemental analysis, brunauer-emmett-teller (bet) physisorption, and z-ray photoelectron spectroscopy (xps). after impregnation, the catalyst surface area decreased compared with that of the support. more drastic decrease was observed in the case of mcm-41 (77%) than the kieselguhr supported sample (48%). there are also different degrees of vanadium oxide-support interaction as reflected by the xps result. si-o binding energy of 531.5 ev was observed on mcm-41-supported sample compared with 529.5 ev for the kieselguhr-supported sample. the catalyst tests were conducted at atmospheric pressure, with a propane to oxygen ratio of 0.7 3.6 and a reaction temperature of 400 700 °c. oxidative dehydrogenation and combustion products were observed. minor cracking reaction products (methane, ethane, and ethene) were also produced above 550 °c. the highest propene yield of 14% was obtained from the kieselguhr-supported sample at 700 °c and with a c3h8/o2 ratio of 1.5. keywords: oxidative dehydrogenation, propane, propene, vanadium oxide, kieselguhr, mcm-41 "kieselguhr"(mcm-41) -0 (si 47 by jibril, a atta, sa al-dress, mc al-kinany and ha al-megren oxydehydrogenation of propane to propene have exhibited low yields. as of publication, there is no economic motivation for developing the process on a commercial scale due to low yields (cavani et al. 2007). because of this low yield, the current research effort is toward developing active catalysts to conduct the reaction at temperatures lower than 600 oc with low selectivity to undesired products. to this end, a large number of supported oxides and mixed-oxides have been reported as summarized in recent reviews (mamedov, cortes 1995; blasco, lopez-nieto 1997; centi et al. 2001). most of the catalysts reported, such as v-mg-o (mamedov, cortes c 1995; centi et al. 2001), mg-v-sb-o (michaels et al. 1996a; michaels et al. 1996b) and v-nb-o (smits et al. 1993) are based on vanadium. others, such as those based on molybdenum, have also been shown to be active for the reaction (stern et al. 1997). for the vanadiumbased system, high selectivity to propene has been associated with the presence of monomeric or oligometric tetrahedral v5+ species. the bridging oxygen between vanadium and the support activates propane (christodoulakis et al. 2004; routray et al. 2004). in the case of v-mg-o, three different phases mg3v2o8, mg2v2o7 and mgv2o6 have been identified. two of them orthovanadate (chaar et al. 1987) and pyrovanadate (siew et al. 1990; ruiz-guerrero et al. 1992) are associated with the activation of alkane. in both phases, the catalysts reactivities depend on the difference in the nature and surface density of v-o bonds (mamedov, cortes 1995; blasco, lopez-nieto 1997). furthermore, the nature of the support influences the surface density and reducibility of the v-o bond (mamedov, cortes 1995; tsilomelekis et al. 2007; xie et al. 2000). silica was identified as exhibiting higher selectivity to propene than more acidic supports such as alumina and titania (kondratenko et al. 2006; singh et al. 2005; grabowski 2004). high vanadia dispersion was realized on high surface mesoporous silica such as m41s (pena et al. 2001; buyevskaya et al. 2001), hms (knotek et al. 2007) and mcm-41 (karakoulia et al. 2008; jibril, ahmed 2006; shanahan, holder 1999). our previous study demonstrated the effect of intra and extra-framework positions of the co, ni or mo in mcm-41 on the selectivity to propene (jibril, ahmed 2006). the aim of this work was to further explore the effects of silica-based supports (i.e., kieselguhr and mcm-41 on the vanadium species and their catalytic performances in the oxidative hydrogenation of propane. kieselguhr is diatomaceous earth (primarily silica dioxide based) that has been demonstrated to be effective as a support in oxidative dehydrogenation of propane (jibril et al. 2006) and other reactions (wells 1991). in addition to the catalytic test, the properties of a sample of the catalysts were studied using brunauer-emmett-teller (bet), surface area, (x-ray fluorescene (xrf), and x-ray photoelectron spectroscopy (xps). the catalysts were found to be active for the reaction. 2. experimental 2.1 preparation of catalysts the materials used in the experiment were: cetyltrimethyl ammonium chloride (c16tmacl) (sigma) tetraethoxysilane (teos) (sigma, 98%), sodium aluminate (fluka, 98%), ammonium vanadate [(nh4)vo3] (aldrich, 99%) and commercial kieselguhr (bdh). high purity propane (99.95%), oxygen (99.995%), and helium (99.999%) were obtained from a local supplier and used as received. the mesoporous mcm-41 support was prepared according to a procedure reported earlier (berndt et al. 2000): 2 g of c16tmacl was dissolved in 500 ml of distilled and deionized water containing 50 ml of 0.125 m of naoh at room temperature. the mixture was stirred for 2 hours. to the above solution, 8 grams of sodium aluminate in 50 ml water and 5 grams of teos were added. the mixture was continuously stirred for 48 hours at ambient temperature. a precipitate was formed, filtered, washed with distilled and deionized water and dried for 10 hours at 100 ºc in a vacuum oven. to remove the organic surfactant templates, the synthesized sample was refluxed in 200 ml of ethanol for 12 h. the extracted material was filtered, washed with ethanol and dried for 10 hours at 100 °c in a vacuum oven. mcm-41 and kieselguhr supported vanadium oxide catalysts were prepared by wet impregnation of mcm41 and commercial kieselguhr with an aqueous solution containing ammonium vanadate with a concentration calculated to give a total loading of 5 w% of vanadium as an oxide. the prepared catalysts were dried at 100 °c in a vacuum oven for 16 hours before calcination in the air at 700 °c for 3 hours. in the subsequent discussion, the catalysts v-oxide/mcm-41 and voxide/kieselguhr are represented as v/m and v/k respectively. 2.2 catalyst characterization the nitrogen adsorption-desorption isotherm of mcm-41, commercial kieselguhr, or supported vanadium oxide catalysts samples were measured with the micromeritics asap-2010 system using nitrogen sorption at 77 k. prior to the experiments, approximately 0.2 g of sample was degassed at 523 k for 10 hours. the surface areas and porosities were obtained. the surface area was calculated using a multipoint 48 oxydehydrogenation of propane over vanadium oxide supported on kieselguhr or mcm-41 bet model. the pore sizes were obtained using berrett-joyner-halenda (bjh) and bet models. the total pore volumes were estimated at a relative pressure of 0.99 atm, assuming full surface saturation with nitrogen. xps studies were recorded with a jeol jps 9010mc photoelectron spectrometer using mgk (1253.6ev) radiation from an x-ray source operated at 10kv and 20 ma. the measurements were performed at room temperature and the working pressure was lower than 3.5x10-7 pa. x-ray elemental analysis studies were taken at 30 kv at room temperature using a jeol element analyzer jax-3201. 2.3 catalyst evaluation the catalyst was tested in a laboratory bench scale pilot plant fitted with mass flow meters and a fixed bed down pass flow stainless steel reactor (inside diameter = 8 mm, and length = 300 mm) placed in a cylindrical furnace equipped with a coaxial thermocouple. the catalyst zone (about 30 mm) in the middle of the reactor was filled with 1.0 g of catalyst diluted with an equal amount of quartz particles (100-600 mesh) in order to minimize temperature gradients. in the center of the catalyst bed, a thermocouple was installed so that it was in contact with the catalyst particles were measure the reaction temperature. the heating zones at the inlet and the outlet of the reactor was measured in a thermocouple located inside the furnace and were controlled by a temperature controller (cole parmer, digi-sense). quartz wool was placed in the empty space in the reactor to minimize gas-phase reactions. a feed flow of 83 ml/minute with -propane to oxygen ratios ranging from 0.7 3.6 at atmospheric pressure was used with helium employed as a diluent. the catalyst zone of 30 mm mades the contact time 4.4 seconds. the activities of the catalysts and product distribution were studied at 400, 500, 600 and 700 °c. blank reactor runs were conducted as described above in the absence of catalysts. no significant conversions of c3h8 or o2 were observed. carbon mass balance obtained was within ± 8%. the reactants and reaction products were passed through a heated-line to an online rga varian 3800 series gas chromatography equipped with a flame ionization detector (fid) to analyze hydrocarbon products (ch4, c2h6, c2h4, c3h8 and c3h6) and thermal conductivity detectors (tcd) to analyze o2, co and co2. the analyses were based on external standard calibration. conversions, selectivity, and carbon balance are defined below: conversion = [propane (in) propane (out)] x 100/propane (in) selectivity = [moles of product / moles of products] x 100 carbon balance = [moles of c atoms (out) / moles of c atoms (in)] x 100 3. results and discussion 3.1 catalyst characterization bet surface area, pore volume, and average pore size based on the bet and bjh methods for the catalysts and bare mcm-41 and kieselguhr are shown in table 1. the impregnation of the active component and calcinations causes reductions in the bet surface areas and pore volumes of the supports. for the less porous kieselguhr; the decrease in the area (48%) is much lower than that of mesoporous mcm-41 (77%). there are corresponding decreases in pore volumes and increases in pore diameters as was found in an earlier report (liu et al. 2008). this suggests a blockage of narrow pores by vanadium oxides dispersed on the surface of the supports. x-ray elemental analysis studies are recorded using x-ray source operating at 30 kv. the analysis of suptable 1. bet surface areas, pore volumes and average pore diameters of the supports and catalysts (a) bet surface area (b) pore volume (c) average pore size determined based on the bet method (d) average pore size determined based on the bjh method 49 by jibril, a atta, sa al-dress, mc al-kinany and ha al-megren ported vanadium oxide on kieselguhr showed two major peaks in sik attributed to kieselguhr (sio2) and vk attributed to metal oxide as shown in fig. 1a. also the analysis of supported vanadium oxide on mcm-41 showed three major peaks sik and alk attributed to mcm-41 support and vk attributed to 0.00 1.00 2.00 3.00 4.00 5.00 6.00 figure 1a. x-ray elemental analysis of v-oxide/kieselguhr support 0.00 1.00 2.00 3.00 4.00 5.00 6.00 figure 1b. x-ray elemental analysis of v-oxide/mcm-41 support 50 oxydehydrogenation of propane over vanadium oxide supported on kieselguhr or mcm-41 metal oxide as shown in fig. 1b. the quantitative analysis as summarized in table 2 shows that the composition of vanadium oxide on each support is about 5 wt% after the calcinations. the results in table 3 shows the results of xps for the catalysts and supports. figures 2a and 2b show two of the xps spectra, as samples of wide and narrow scan spectra (o1s, si2p3/2 and v 2p3/2) of v2o5/kieselguhr. it can be speculated that the v2p3/2 binding energy of 519.1 +1ev for the v-line is due to the presence of vanadium species in an oxidation state v5+ in the vanadium samples. its variation in the catalysts indicates the effect of v-o-support interactions. the xps of the si2p3/2 energy region shows the presence of a well defined line at 102.1ev characteristics of sio2 state of kieselguhr and mcm-41 supports. the al2p photoelectron spectra exhibit al 2p1/2 at 74.5 ev, which shares characteristics of al2o3. a strong o 1s spectral line recorded at 532 ev is attributed to oxygen from the supports and supported vanadium oxides. the o 1s line for mcm-41 support could be divided into two bands (in the curve fitting) at 531.5 ev (si-o) and 532.2 ev (al-o). in the case of voxide/kieselguhr the two bonds are identified at 529.5 ev (si-o) and 531.5 ev (v-o), while in the case of voxide/mcm-41 the three bonds are identified at 531.5 ev (v-o), 531.2 ev (si-o) and 532.7 ev (al-o). the variations in oxygen binding energies reflect the effects of the supports on the catalysts surface properties. 3.2 tests of catalysts the performances of the catalysts in propane oxidative dehydrogenation were studied. figure 3 summarizes the effects of temperature and feed compositions on propane conversions and products distributions. propene is the desired product. in the oxidative dehydrogenation reaction, we expect carbon poisoning effects to be minimized. this is because the oxygen in the feed stream reacts with the carbon or its precursor to produce co/co2. the catalysts exhibit low conversions at 400 °c. both propene and co2 appear to be primary products on v/k while only propene is a primary product on v/m. the occurrence of the two reactions suggests the presence of sites of different reducibilities on the catalysts (klisinska 2006). on v/k, the conversion increases exponentially to 60% at 700 °c, with a corresponding linear decrease in propene selectivity. co was produced at 500 °c and higher temperatures with selectivities of 5 15%. it is perhaps a secondary product of propene further reactions. when the conversion increases to 5%, selectivity to co2 increases about six fold that of co. the high selectivity to co2 is in line with the presence of highly reducible sites. these attack the electron-rich region of propyl or propane species to produce co2. the exponential increase in propane conversion is due to cracking of propane as evidenced by the high selectivities to methane and ethane. however, the cracking seems to be catalyst surface mediated due to table 2. x-ray elemental analysis table 3 catalysts (a) curve fitting of the 0 1s binding energy 51 by jibril, a atta, sa al-dress, mc al-kinany and ha al-megren lower levels of conversion and selectivity to co2 observed on v/m (ji et al. 2008). on the v/m, propene is the only primary product. its further reaction led to co instead of the co2 observed in the case of v/k at 500 °c. at 600 °c, selectivity to co2 of 66% was obtained. thus, there was co to co2 switch between the temperatures of 500 and 600 °c. this suggests the dynamic change in reducibilities of the catalyst’s surface. it may also be associated with an increase in oxygen mobility at higher temperature, binding energy (ev) figure 2b. narrow scan xps spectrum of 0 1s, si 2p/32 and v 2p3/2 of 5% v2o5 / kiesel o1s si 2p v 2p 0 900 800 700 600 500 400 300 200 100 binding energy (ev) 18000 14000 10000 6000 2000 figure 2a. wide scan xps spectrum of supported 5% v205 on kieselguhr 540 535 530 525 105 100 95 90 520 515 510 505 oxygen silicon vanadium 1000 800 600 400 200 52 2 rather than co, depending on the nature of the catalysts (owen et al. 2008). similar to v/k, a substantial amount of cracking products with selectivity of about 35% was observed. in order to gain further insight into the effects of catalysts on the cracking reaction, the propane/oxygen ratio was varied at 700 °c (fig. 4). as the ratio increases (0.7 3.6), the conversion and selectivities to cox decrease with corresponding increase in selectivities to propene. the relative amount of oxygen in the feed determines the selectivity to propene. for the oxygen-rich feed; v/k exhibits selectivity to propene of 3%. this improves to 23% when the amount of propane increases five times at a constant oxygen flow. the ratio of co/co2 also changes (2.8-0.7) from low to high propane flow. the distribution is different on v/m catalysts. the dominant product is co2 at all feed compositions. the differences in cox distributions by the catalysts are in line with differences in oxygen binding energy in v/k and v/m http://(srdata.nist.gov/xps/ (nist x-ray photoelectron spectroscopy database). 4. conclusions vanadium-oxide supported on kieselguhr is more active than that supported on mcm-41 in propane figure 4. effect of propane / oxygen ratio on propane conversion and products selectivities for (a) v2o5 / kieselguhr and (b) v2o5 / mcm-41 c3h8/02 ratio reaction temperature, oc figure 3. effect of temperature on propane conversion and products selectivities for (a) v2o3/kieselguhr and v2o3/mcm-41 53 by jibril, a atta, sa al-dress, mc al-kinany and ha al-megren oxidative dehydrogenation at the conditions explored. the latter's surface area was more drastically decreased when it was impregnated with the vanadium oxide. both propene and co2 appeared to be primary products on vanadium-oxide/kiesulguhr while only propene was the primary product on vanadiumoxide/mcm-41, suggestive of sites of different reducibilities on the catalysts. the superior performance was maintained when the catalysts were further tested in both oxygen-rich and oxygen-deficient feedstock. the higher activity and different products distributions of kieselguhr supported samples is associated with its lower si-o binding energy as evidenced by xps. the different extent of vanadium oxide interaction with supports affects both its surface properties and catalytic activities. acknowledgments for assistance in catalysts characterization, the authors are grateful to engr. fahad h. al-malki of petroleum & petrochemicals research institute, king abdulaziz city for science and technology, riyadh, saudi arabia. references berndt h, martin a, bruckere a, schreier e, miiller d, kosslick h, wolf gu, liicke b (2000), structure and catalytic properties of vox/mcm materials for the partial oxidation of methane to formaldehyde. j. catal. 191(2):384-400. blasco t, lopez-nieto jm (1997), oxidative dehydrogenation of short chain alkanes on supported vanadium oxide catalysts. appl. catal. a 157(12):117-142. buyevskaya ov, brückner a, kondratenko ev, wolf d, baerns m (2001), fundamental and combinatorial approaches in search for and optimisation of catalytic materials for oxidative dehydrogenation of propane to propene. catal. today 67(4):369378. cavani f, ballarini n, cericola a (2007), oxidative dehydrogenation of ethane and propane: how far from commercialization? catal. today 127 (14):113-131. centi g, cavani f, trifirò f (2001), selective oxidation by heterogeneous catalysis, fundamental and applied catalysis, kluwer academic, new york. chaar ma, patel d, kung mc, kung hh (1987), selective oxidative dehydrogenation of butane over vmgo catalysts j. catal. 105(2):483-498. christodoulakis a, machli m, lemonidou aa, boghosian s (2004), molecular structure and reactivity of vanadium based catalysts for propane oxydehydrogenation studied by in-situ roman spectroscopy and catalytic activity measurements. j. catal. 222(2):293-306. grabowski r (2004), kinetic oxidative dehydrogenation of propane on vanadia/titania catalysts, pure and duped rubidium. appl. catal. a: general 270(1-2):37-47. jibril by, ahmed s (2006), oxidative dehydrogenation of propane over co, ni and mo mixed oxides/mcm-41catalysts: effects intra and extraframework locations of metals on product distribution. catal. comm. 7(12):990-996. jibril by, al-kinany mc, al-khowaiter sh, al-drees sa, al-megren ha, al-dosari ma, al-rasheed rh, al-zahrani sm, abasaeed ae (2006) performances of new kieselguhr-supported transition metal oxide in propane oxydehydrogenation. catal. comm. 7(2):79-85. jung jic, lee h, kim h, park s, chung y-min, kim jt, lee js, oh s-h, kim ys, song ik (2008) effect of calcinations temperature on the catalytic performance of co9fe3bi1mo12o51 in the oxidative dehydrogenation of n-butene to1,3-butadiene. catal. comm. 9(10):2059-2062. karakoulia sa, triantafyllidis ks, lemonidou aa (2008) preparation nd characterization of vanadia catalysts supported on non-porous, microporous and mesoporous silicates for oxidative dehydrogenation of propane (odp). mesopor. mater 110(1):157-166. klisinska a, loridant s, grzybowska b, stoch j, gressel i (2006), effect of additives on properties of v2o5/sio2 and v2o5/mgo catalysts: ii. structure and physiochemical properties of catalysts and their correlation with with oxidative dehydrogenation of propane and ethane. appl. catal. a: general 309(1):17-24. kondratenko ev, cherian m, baerns m (2006), oxidative dehydrogenation of propane over differently structured vanadia-based catalysts in the presence of o2 and n2o. catal. today 112(14):60-63. knotek p, capek l, bulánek r, adam j (2007), vanadium supported on hexagonal mesoporous silica: active and stable catalysts in the oxidative dehydrogenation of alkanes. top. catal. 45(14):51-55. liu w, yin ls, dai h, wang s, sun h, tong auc (2008), mgo-modified vox/sba-15 as catalysts for oxidative dehydrogenation of n-butane. catal. today 131(1-4):450-456. mamedov ea, cortes cv (1995), oxidative dehydrogenation of lower alkanes using vanadium oxidebased catalysts. the present state of art and outlooks. appl. catal. a 127(1-2):1-40. michaels jn, stern dl, grasselli rk (1996a), oxydehydrogenation of propane over mg-v-sb 54 oxydehydrogenation of propane over vanadium oxide supported on kieselguhr or mcm-41 oxide catalysts. i. reaction network. catal. lett. 42(3-4):135-137. michaels jn, stern dl, grasselli rk (1996b), oxydehydrogenation of propane over mg-v-sb oxide catalysts. ii. reaction kinetics and mechanism. catal. lett. 42(3-4):139-148. nist x-ray photoelectron spectroscopy database. http://srdata.nist.gov/xps/. owen os, kung hh (1993), effect of cation reducibility on oxidative dehydrogenation of butane on orthovanadates. j. mol. catal. 79(13):265-284. pena ml, dejoz a, formés v, rey f, vásquez mi, lópez-nieto jm (2001), v-containg mcm-41 and mcm-48 for selective oxygenation of propane in gas phase. appl. catal. a 209 (1-2):155-64. ruiz-guerrero a, rodriguez ri, fierro jlg, sonnen v, herrmann jm, volta jc (1992), new developments in selective oxidative by heterogeneous catalysis. stud. surf. sci. catal. 72:203-212. routray k, reddy krsk, deo g (2004), oxydetive dehydrogenation of propane on v2o5/al2o3 and v2o5/tio5 catalysts: understanding the effect of support by parameter estimation appl. catal. a 265 (1):103-113. shanahan kl, holder js (1999), tritium aging effects of palladium on kieselguhr. j. alloys, compd, 293-295:62-66. smits rhh, seshan k, leemreize h, ross jrh (1993), influence of preparation method on the performance of vanadia-nobia catalysts for the oxidative dehydrogenation of propane. catal. today 16(3-4):513-523. siew hsd, soenen v, volta jc (1990), oxidative dehydrogenation of propane over v-mg-o catalysts. j. catal. 123(2):417-435. singh rp, bañares ma, deo g (2005), effect of phosphorous modifier on v2o5/tio2 catalyst: odh of propane. j. catal. 233(2):388-398. stern dl, grasselli rk (1997), reaction network and kinetics of propane oxydehydrogenation over nickel cobalt molybdate. j. catal. 167(2):560569. stern dl, michaels jn, decaul l, grasselli rk (1997), oxydehydrogenation of n-butane over promoted mg v-oxide based catalysts. appl. catal. a 153(1-2):21-30. tsilomelekis g, christodoulakis a, boghosian s (2007), support effects on structure and activity of molybdenum oxide catalysts for the oxidative dehydrogenation of ethane. catal. today 127 (14):139-147. wells mg (1991), handbook of petrochemicals and processes, gower publishing co. ltd. xie s, chen k, bell at, iglesia e (2000), structural characterization of molybdenum oxide supported on zirconia. j. phys. chem. b: 104(43):1005910068. the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 modelling of electrical submersible pumps for petroleum multiphase fluids, an intelligent approach supported by a critical review and experimental results m. mohammadzaheria, * , r. tafreshi b , z. khan c , h. ziaiefar a , m. ghodsi a , m. franchek c , and k. grigoriadis c a sultan qaboos university, muscat, oman b texas a&m university at qatar, education city, doha, qatar c university of houston, 4800 calhoun road, houston, usa. abstract: this paper initially reviews existing empirical models which predict head or pressure increase of two-phase petroleum fluids in electrical submersible pumps (esps), then, proposes an alternative model, a fully connected cascade (fcc in short) artificial neural network to serve the same purpose. empirical models of esp are extensively in use; while analytical models are yet to be vastly employed in practice due to their complexity, reliance on over-simplified assumptions or lack of accuracy. the proposed fcc is trained and cross-validated with the same data used in developing a number of empirical models; however, the developed model presents higher accuracy than the aforementioned empirical models. the mean of absolute prediction error of the fcc for the experimental data not used in its training, is 68% less than the most accurate existing empirical model. keywords: electrical submersible pumps; empirical models; cascade artificial neural networks; multiphase petroleum fluid. الغاطسة الكهربائية للسوائل البترولية المتعددة ، نهج نمذجة المضخات ذكي مدعوم بمراجعة نقدية ونتائج تجريبية ري اهظمرتضى دمحم أ * رضا تفريشي ، ب ، زوروا خان جـ ، حميد رضا زيايفر أ ، موجاتبا غودسي أ ، ماثيو فرانشيتك جـ ، كارولوس غريغورياديس جـ الورقة مبدئيًا النماذج التجريبية الحالية التي تتنبأ بزيادة في ارتفاع منسوب أو زيادة ضغط تستعرض هذه : الملخص السوائل البترولية ثنائية الطور في المضخات الغاطسة الكهربائية ؛ ثم تقترح نموذًجا بديًًل ، وهو عبارة عن شبكة عصبية أن النماذج التجريبية للمضخات الغاطسة الكهربائية صناعية متتالية متصلة اتصال كامل لخدمة نفس الغرض. علما ب مستخدمة على نطاق واسع ؛ في حين أن النماذج التحليلية لم تستخدم بعد على نطاق واسع في الممارسة العملية بسبب من مدى تعقيدها أو اعتمادها على افتراضات مبسطة للغاية أو غير دقيقة. ويتم تجريب الشبكة المتصلة بالكامل و التحقق صحتها باستعمال نفس البيانات المستخدمة في تطوير عدد من النماذج التجريبية ؛ ومع ذلك ، فإن النماذج المطورة تعطي دقة أعلى من النماذج التجريبية المذكورة أعًله. فمتوسط خطأ التنبؤ المطلق للمتتالية المتصلة بالكامل ، و تبين أن متوسط ٪ من النموذج الحالي التجريبي 86المتصلة اتصال كامل غير المستخدمة في التجربة ، أقل بنسبة البيانات للشبكة التجريبية األكثر دقة . سوائل ؛الشبكات العصبية االصطناعية المتتالية ؛النماذج التجريبية ؛لمضخات الغاطسة الكهربائية : االمفتاحية الكلمات .البترول متعددة األطوار * corresponding author’s e-mail: mmzahery@gmail.com doi: 10.24200/tjer.vol16iss2pp77-86 mailto:mmzahery@gmail.com the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 nomenclature a fixed value of a variable b,b bias in the fcc c weight matrix of fcc cascade connection e training error ei model parameter f activation function g gravity acceleration g error gradient vector h head (m) h hamiltonian matrix j model parameter n number n rotational speed (rpm) p pressure (mpa) q flow rate (lpm, litres per minute) t,w fcc weight matrices v variable x unknown parameter greek letters α gas void ratio β flow parameter of the affinity law δ head parameter of the affinity law η a coefficient in fcc training θ a vector of fcc parameters λ a coefficient in fcc training φ an indicator of surging indices d data g gas in input l liquid m mixture of gas and liquid p parameters abbreviations ann artificial neural network esp electrical submersible pump fcc fully connected cascade ann lse least square of errors 1. introduction electrical submersible pumps (esps), invented in 1927, are the second prevalent artificial lift methods in the petroleum industry (machado et al. 2019) and are employed for downhole drilling operations at various flow rates (mohammadzaheri et al. 2016; zhu et al. 2018). selection of esp size is a crucial task, because overor under-sizing leads to premature equipment failure or inadequate oil recovery, respectively. esp manufacturers provide curves to be used in choosing the right size of esps. however, these curves are valid when only liquid is pumped. in some reservoirs, esps should pump two-phase flow with high gas content. in this case, instead of the curves, the designer needs models which predict the head of esps with fluids comprising of gas and liquid. these models are known as ‘head predicting’ models, which may accompany supplementary models (widely known as ‘critical’ models) providing flow information, e.g. border of surging, within the esp ( barrios and prado 2011). critical models are largely used to define the validity area of a head-predicting model. head-predicting models with no accompanying critical model are considered to be valid for the entire operating area. except for head-predicting and critical models, some other models have also been developed for side purposes such as estimation of gas bubble size by zhu and zhang (2017), or in-situ gas volume fraction by zhu and zhang ( 2016), in esps pumping two-phase flows; these models are outside the scope of this paper. analytical and empirical methods have been employed to develop head predicting models for two-phase flows in esps. analytical head predicting models have been derived based on mass and momentum balances (sachdeva 1988; sun and prado 2005). however, they are based on unrealistic assumptions and/or oversimplification of complex physics of two-phase flows within the esp. therefore, analytical models are not considered an ideal option to model two-phase flow in esps, and empirical models are widely trusted alternatively (zhou and sachdeva 2010; mohammadzaheri et al. 2018). section 2 of this article reviews prevailing empirical models. section 3 reports the development of a cascade artificial neural network to predict esp head for two-phase fluids. the proposed model is shown to be valid for a range of operating conditions. 2. review of esp empirical models in the literature, six empirical models have been reported to estimate esps head or pressure for twophase fluids in steady state situation. some have their own critical models. two major sources have supplied the experimental data for development of empirical models (lea and bearden 1982; cirilo 1998). in this section, the models are classified on the basis of data source. group 1: the models developed with the data collected from diesel fuel/carbon dioxide mixtures. group 2: the models developed with the data of experiments carried out on water/air mixture. group 1 models are more impor tant for the oil industry, as the diesel fuel/ carbon dioxide mixture is more similar to petroleum fluids. 2.1 group 1 group 1 models are based on the data presented in lea and bearden ( 1982). these data cover an extensive range of gas void fractions (0 to 50%) and intake pressures (0.345 to 2.758 mpa). these data 78 m. mohammadzaheri, r. tafreshi, z. khan, h. ziaiefar, m. ghodsi, m. franchek, and k. grigoriadis are the result of 12 series of experiments; each with a particular pressure and flow rate. at each series, gas void fraction has changed and head generated by a number of esp pumps (based on esp type) has been recorded. both the mixture of air and water and the mixture of carbon dioxide and diesel fuel have been used in these experiments. in this group, models 1 and 3 directly predict head; whereas, model 2 predicts pressure increase and needs (1) to estimate head, hm, out of pressure increase,  p: where ρ and g represent fluid density and gravity acceleration, respectively. index m refers to mixture of gas and liquid. model 1 this model was developed by turpin et al in 1986 (turpin et al. 1986) and is claimed to be valid both for air/water and co2/diesel fuel mixtures: where ql and qg are liquid and gas flow rates, pin is input pressure in mpa, and hl is the head from manufacturer’s catalogue for a liquid flow rate of ql + qg. to find unknown parameters (x1 and x2, identified as 16.47 and -2.84), (2) can be re-written as or with use of head, flow rate and pressure information from experiments, (3) would be a linear equation with two unknowns. least square of error (lse) method can straightforwardly solve such equations (mohammadzaheri et al. 2009). all the empirical models of group1 are convertible to a linear equation similarly. model 1 is valid where the pump is stable or ϕ<1, where model 2 sachdeva et al. proposed this model in 1992 sachdeva et al. (1992): where is pressure increases in mpa for 6or 8stage pumps (depending on the pump type) (zhou and sachdeva 2010). α is gas void fraction, ie. the ratio of free gas volumetric rate (qg) to the total volumetric rate (qm): similar to model 1, (5) can be converted to a linear equation solvable by lse method. parameters of (5) are listed in zhou and sachdeva ( 2010) for multiple stages of esps. model 2 has no accompanying critical model; thus, it is assumed to be valid for the whole operating area. model 3 this model was offered by zhou and sachdeva in 2010 (zhou and sachdeva 2010) where c is pressure unit coefficient, eg. 1, 1000, 0.145 or 145 for psi, ksi, kpa or mpa. hmax and qmax are nominal maximum head and flow rate of the pump. advantageously, use of hmax and qmax can make head dimensionless in empirical models of esps, as originally proposed in romero (1999). moreover, use of aforementioned parameters allows the model to utilize catalogue information. model 3 is indeed a revised version of model 2, which estimates head instead of pressure, alike other empirical models developed in 1999 onwards. the output of (7) is evidently the maximum head when flow rate is zero. parameters of (7) are listed in zhou and sachdeva (2010). k2 and k3 in models 2 and 3 merely depend on esp type as to zhou and sachdeva (2010); however, can be considered as the gain of model 3 as demonstrated in (8): this gain ( ) would depend on the pressure unit. as an alternative to remove this dependency, dimensionless pressure could be used. model 3 has been validated only for stable operating areas before surging. (9) defines maximum stable flow rate for a given pair of pressure and gas void ratio as to zhou and sachdeva (2010): for 8 stages of i-42b radial esp, k3c=3.7453, e7=0.07244, e8=0.318544.the unit for flow is lpm (litres per minute). 2.2 group 2 the data (used to develop these models are presented in cirilo (1998), where air is the only gas and water is the sole liquid. in this paper, cirilio 79 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 claimed that the gas void fraction at the border of surging solely depends on intake pressure: two of group 2 models are valid at particular flow regimes: bubbly regime and elongated bubble regime (estevam 2002; gamboa and prado 2010). model 4 romero(1999) analyzed the data of cirilo (1998), and suggested a new model to predict head (model 4) in which its validity area is defined by (12): oddly, gas void ratio, gas volumetric flow rate or intake pressure play no role in this model. critical model of (12) determines both maximum gas void ratio and minimum liquid flow rate: model 5 duran and prado (2003) suggested long-known homogenous model of (13) or (14) for bubbly regime: it is worth mentioning that (14) is twin of (13). the only difficulty to use (13) is the need for density. to address this weakness and ease calculations, duran and prado (2003) presented the following closure correlation: (6), (14) and (15) can find any of ql, qg, qm, ρl , ρg , ρm and α based on others. (16) is the critical model of the homogenous model: the set of (16) and (6) determines maximum gas flow rate in bubbly regime; that is, at any higher flow rate, the homogenous model is invalid. model 6 duran and prado suggested a head-predicting model peculiar to the elongated bubbly regime (duran and prado 2003): both pressure and density are absent in this model. lse technique has been employed to find constants -0.0224 and -0.0103. model 6 is accompanied by a critical model (18). the set of (18) and (6) provides with the minimum gas flow rate or gas void fraction in elongated bubble regime: ( ( ) ) ( ( ) ) ( ) 2.3 unconsidered inputs to empirical models the first overlooked input of models 1-6 is rotational speed. all these models have been developed with the data collected at a single rotational speed of 3500 rpm. hence, these empirical models are valid only at this rotational speed. affinity laws of (19) are used to extend the validity area of empirical models. these laws are developed with the data collected at different rotational speeds (zhou and sachdeva 2010): where n is rotational speed in rpm, β=1 and 0.8 have been suggested for liquid and two-phase fluids respectively, and δ=2 has been suggested for both (zhou and sachdeva 2010). temperature is the other overlooked input to models 1-6. the reason is nonexistence of experimental data which reflect temperature effect. experiments on esps pumping liquid have shown that that temperature has a significant effect only if numerous (>100) pump stages are used (kirvelis and davies 2003). 2.4 review summary two groups of empirical models (models 1-3 and models 4-6) for two-phase fluids in esps were reviewed in this section. the parameters of these models have been identified using experimental data of air/water (both groups) and carbon dioxide/diesel fuel (group 1) mixtures. these models estimate head (models 1, 3-6) or pressure increase (model 2) generated by esp systems, and are supplemented by five critical models defining their validity area. the validity area of an empirical model without a critical model is the operating area where its associated data have been gathered. head/pressure predicting models, except for model 80 m. mohammadzaheri, r. tafreshi, z. khan, h. ziaiefar, m. ghodsi, m. franchek, and k. grigoriadis 2, have parameters which can be found in the catalogue. majority of models have three inputs: the first is either intake pressure or density and the second and third are two of ql, qg,, qm or α. pump rotational speed and temperature have been overlooked in empirical modelling so far. all empirical models, except for model 5, are linear equations with few unknowns or can be converted to such linear equations through logarithm. as a result, straightforward technique of lse can be employed to identify handful unknown parameters of models. however, this research shows that the cost of this simplicity is loss of accuracy. in this research, a complex mathematical architecture, namely fully connected cascade (fcc) architecture, is used, which is a type of artificial neural networks (anns) with the most powerful architecture for system identification (hunter et al. 2012). 3. fcc development an fcc was designed and developed to model mixtures of carbon dioxide and diesel fuel pumped by eight stages of an i-42b radial esp up to head of 16.76 m, detailed in lea and bearden (1982). the fcc was developed with the same data as models 1, 2 and 3 and is also comparable with 5. motivated by a comparative review of section 2, the fcc has one output, hm , and three inputs: pin, qm and α. as depicted in fig.1, the fcc has a hidden layer with sigmoid activation functions of f, presented in (20), and weight matrix of w: five neurons have been considered for the hidden layer. this choice is based on a multilayer perceptron developed for a similar purpose (mohammadzaheri et al. 2015) and the fact that the number of hidden layer neurons in fccs should be less than this number in multilayer perceptron’s (hunter et al. 2012). fig. 1 and eq. (21) illustrate the proposed fcc: where w and t are the matrices of the first and the second layer weights, c is the matrix of cascade connection weights. b and b represent biases of the first and the second layers, respectively. fcc model of (21) has 29 parameters in total. 111 data sets have been used to develop and cross-validate the fcc model: 69 data sets for initialization and identification of parameters, 25 sets for overfitting prevention (also known as validation) and 17 sets for cross-validation or test. the process of parameter identification (known as training) is iterative as detailed in appendix a. at each iteration of fcc training, the real and model outputs for 69 sets of training data and 25 sets of validation data were calculated; mean of square of their discrepancies were considered as the training and validation error. inconsistency in trend of these dual errors is as a sign of over-fitting and should trigger to end training. over-fitting diminishes the generality of anns (mohammadzaheri et al. 2007). afterwards, the error was similarly calculated for 17 sets of test data, used neither in training nor in validation. a reasonably small test error ensured cross-validation. 4. experimental assessment of models this section compares head prediction error in different models and operating areas. mean and maximum absolute test error of the fcc are 2.73% and 7.46% of maximum fluid head (16.76 m), respectively. figure 2 presents head prediction error of different models/operating areas for test data, table 1 presents these information for the whole experimental data. models 4 and 6 are not presented because their parameters have been identified using experimental data of air-water mixture; thus, these models are not comparable with the developed fcc. models 1 and 3 have critical models defining their validity or stability area, where surging absents. surprisingly, stable areas of these models do not match. models 2 and 5 have no validity area in terms of input pressure and gas void fraction. the background of cells in table 1 reflects models validity, cells with white, light grey and dark grey background show the operating areas where the model is valid, partly valid (in some flow rates) or invalid, respectively. fcc is considered valid for the whole operating area due to cross-validation. figures 3-5 show the average head prediction error for test data at each operation condition or «intake pressure, gas void fraction» pair. figures 3-5 exhibit the following results:  in 9 out of 11 operating conditions, fcc provides the most accurate head prediction. in two other operating conditions, fcc stands second after models 3 and 1, respectively. parameter identification of an fcc (alike any other empirical model) is an off-line processes and should be performed only once, unless mechanical/geometrical properties of the system considerably change. eqs. (20 and 21) can be solved using any programming languages used by computer or even engineering calculators.  the homogenous model of 5 performs better at higher intake pressures.  fig. 3-5 (presenting the test data) are consistent with table 1 (presenting the whole data) in model 2 has the weakest performance even weaker than homogenous model of 5. 81 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 figure 1. a schematic of the employed cascade artificial neural network. figure 2. mean of absolute test error of different models. table 1. absolute prediction error in m for different models/operating areas, with use of all experimental data. pin α model 1 model 2 model 3 model 5 fcc 0.34 10% 4.4794 7.3481 1.5249 1.4201 0.3301 0.34 15% 3.6546 3.5022 2.2303 2.6305 3.6848 0.2761 0.34 20% 2.6291 2.0887 4.8159 0.2893 0.34 30% 2.0193 1.5408 1.3218 5.0083 0.4958 0.34 40% 0.7112 0.3429 0.9351 5.3848 0.1243 0.69 10% 0.9770 6.8847 1.2913 1.7243 0.2952 0.69 15% 1.4056 4.0781 1.8101 1.9155 0.4395 0.69 20% 1.5108 3.1795 2.0356 2.5146 0.2699 0.69 30% 3.1539 2.3640 1.4171 3.0655 0.7127 0.69 40% 1.9507 1.1623 0.8797 3.5723 0.4273 2.76 30% 1.1361 3.0234 1.7786 1.6668 0.3354 2.76 40% 0.8983 2.5197 1.3117 1.3563 0.2370 2.76 50% 2.7563 2.2885 1.7650 1.7355 0.3938 average 2.0986 3.1018 1.6101 2.9127 0.3559 5. conclusion head estimation of two-phase petroleum fluids in esps is widely performed using empirical models, and analytical and numerical models have yet to be trusted for this task. analytical models are based on idealistic assumptions and demand hardly accessible information of fluid. this article first critically reviewed the existing popular empirical models, then proposed a fully connected cascade artificial neural network as a substitute of the reviewed empirical models. the developed fcc, provided a superior estimation accuracy. 82 m. mohammadzaheri, r. tafreshi, z. khan, h. ziaiefar, m. ghodsi, m. franchek, and k. grigoriadis figure 3. the trend of average head prediction error respect to gas void fraction for different models at intake pressure of 0.34 mpa, for test data. figure 4. the trend of average head prediction error respect to gas void fraction for different models at intake pressure of 0.69 mpa, for test data. figure 5. the trend of average head prediction error respect to gas void fraction for different models at intake pressure of 2.78 mpa, for test data. conflict of interest the authors declare no conflicts of interest. funding this work was supported by nprp grant from the qatar national research fund (a member of qatar foundation), grant number is 7-1114-2-415. references barrios l. prado m . g . (2011), modeling two phase flow inside an electrical submersible pump stage. journal of energy resources technology, 133(4): 1-10. cirilo, r. (1998), air-water flow through electric submersible pumps, a master dissertation at the department of petroleum engineering, university of tulsa, usa. duran j. m. prado m.g. (2003), esp stages air-water two-phase performance-modeling and experimenttal data. spe report, no. 87626. estevam v. (2002), a phenomenological analysis about centrifugal pump in two-phase flow operation. a phd thesis at faculdade de engenharia mecânica, universidade estadual de campinas, brazil. gambo j. prado m.g. (2010), visualization study of performance breakdown in two-phase performance of an electrical submersible pump, international pump users symposium, houston, usa. hunter d.h. yu, m.s. pukish iii, j. kolbusz and b. m. wilamowski (2012), selection of proper neural network sizes and architectures a comparative study. ieee transactions on industrial informatics, 8(2): 228-240. kirvelis r. davies d. (2003), enthalpy balance model leads to more accurate modelling of heavy oil production with an electric submersible pump. chemical engineering research and design, 81(3): 342-351. lea j. f. bearden j. (1982), effect of gaseous fluids on submersible pump performance. journal of petroleum technology 34(12): 922-930. machado, a. p. f. resende c.z. cavalieri d.c. (2019), estimation and prediction of motor load torque applied to electrical submersible pumps. control engineering practice 84: 284-296. mohammadzaheri m.a. alqallaf ghodsi m. ziaiefar h. (2018), development of a fuzzy model to estimate the head of gaseous petroleum fluids driven by electrical submersible pumps. fuzzy information and engineering 10(1): 99-106. mohammadzaheri m. chen l. (2008), intelligent modeling of mimo nonlinear dynamic process plants for predictive control purposes. the 17th world congress of the international federation of 83 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 automatic control, seoul, korea. mohammadzaheri m. chen l. (2010), intelligent predictive control of model helicopters' yaw angle. asian journal of control. 12(6): 1-13. mohammadzaheri m. chen l. a. ghaffari and j. willison (2009), a combination of linear and nonlinear activation functions in neural networks for modeling a de-superheater. simulation modelling practice and theory. 17(2): 398-407. mohammadzaheri m. mirsepahi a. asef-afshar o. and koohi h. (2007), neuro-fuzzy modeling of superheating system of a steam power plant. applied mathematical sciences. 1: 2091-2099. mohammadzaheri m. tafreshi r. khan z, franchek m. and grigoriadis k. (2015), modeling of petroleum multiphase fluids in esps, an intelligent approach. offshore mediterranean conference, ravenna, italy. mohammadzaheri, m., r. tafreshi, z. khan, m. franchek and k. grigoriadis (2016), an intelligent approach to optimize multiphase subsea oil fields lifted by electrical submersible pumps. journal of computational science 15: 50-59. moré j.j. (1978), the levenberg-marquardt algorithm: implementation and theory. lecture notes in mathematics, vol. 630, springer, berlin, heidelberg: 105-116. nguyen d. widrow b. (1990), improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights. international joint conference on neural networks. san diego, usa, 17-21 june, 1990. romero m (1999), an evaluation of an electrical submersible pumping system for high gor wells, a master dissertation at university of tulsa, usa. sachdeva r (1988), two-phase flow through electric submersible pumps, a phd thesis at university of tulsa, usa. sachdeva r. doty d. and schmidt z. (1992), performance of axial electric submersible pumps in a gassy well, spe rocky mountain regional meeting. casper, usa. sun d. prado m. (2005), modeling gas-liquid head performance of electrical sub mersible pump s. journal of pressure vessel technology, 127(1): 3138. turpin j.l. lea j.f. bearden j.l. (1986), gas-liquid flow through centrifugal pumps—correlation of data, the third international pump symposium, college station, texas, usa. wilamowski (2012), selection of proper neural network sizes and architectures—a comparative study. ieee transactions on industrial informatics. 8(2): 228-240. zhou d. sachdeva r . (2010), simple model of electric submersible pump in gassy well. journal of petroleum science and engineering. 70(3): 204213. zhu j. zhang h.q. (2016), mechanistic modeling and numerical simulation of in-situ gas void fraction inside esp impeller. journal of natural gas science and engineering. 36: 144-154. zhu j. zhang h.q. (2017), numerical study on electrical-submersible-pump two-phase performance and bubble-size modeling. spe production & operations. 32(3): 267-278. zhu j. zhu h, wang z.j. zhang, r. cuamatzimelendez, j. a. m. farfan and h.q. zhang (2018), surfactant effect on air/water flow in a multistage electrical submersible pump (esp). experimental thermal and fluid science 98: 95111. 84 84 m. mohammadzaheri, r. tafreshi, z. khan, h. ziaiefar, m. ghodsi, m. franchek, and k. grigoriadis appendix a training of the fcc was composed of four major tasks: 1. data normalization 2. error function definition 3. parameters initialization 4. parameter tuning a.1 data normalization the difference of magnitude scale among input or output columns of data negatively affects the parameter identification (training) process and should be avoided through normalization, a process to assure input or output data columns have almost same scale of magnitude (mohammadzaheri and chen 2008). in this research, for the purpose of normalization, the input and output data columns were mapped into the range of [-1 1]. in practice, the input(s)/outputs(s) of an fcc, trained with the mapped data, should be mapped/demapped into the real range. a.2 error function definition the error function represents the discrepancy between the fcc and real system outputs for identical inputs. mean of squared errors was used as the error function in this research, presented as (a.1): where hat (^) shows the estimated value by the fcc and hm is the measured head of the mixture. n is the number of data sets. a.3 parameters initialization the fcc shown in fig. 1 has a sigmoid activation function in the hidden layer. as depicted in fig.a.1, a sigmoid function is nearly linear for an interval; outside this interval, the output is nearly fixed or saturated (mohammadzaheri et al. 2016). figure a.1. the sigmoid function presented by (6) and its linear interval. in practice, any input to a sigmoid activation function, with a value outside the linear interval (shown in fig. a.1), has trivial influence on the output. within the structure of an fcc, it means the input has trivial impact on the error and training process. nguyen-widrow algorithm suggests initial values for fcc parameters so that inputs to the sigmoid functions lie within the linear interval. random functions are also used in this algorithm to avoid repeated initial weights (nguyen and widrow, 1990). a.4 parameter tuning after initialization, the parameters are tuned iteratively to minimize the error (detailed in section a.2). the error is influenced by 29 parameters of (21). hence, e can be presented as e(θ), where θ is a vector of 85 85 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 77-86 all parameters of the fcc. the first step is to analytically derive the first and second derivatives of e(θ), error back propagation. in this research, levenberg-marquardt method, eq.(a.2), detailed in (moré 1978, mohammadzaheri and chen 2010), was used to minimize e(θ), calculated with 69 sets of the training data and (a.1). levenberg-marquardt method is based on approximation of the error function using taylor series up to the second order derivatives: where g and j are the vector and matrix of partial derivatives of error respect to θ elements. algorithms to find η and λ have been detailed in(jang et al. 2006, mohammadzaheri and chen 2010). appendix b figures 2-5 and table 1 are only concerned with the models’ error. they provide no information about real values of head. figures b.1-2 tend to compensate this shortcoming through depicting measured/predicted head versus flow rate for two operating areas. figure b.1. experimentally measured and predicted head by different models versus flow rate at intake pressure of 2.76 mpa and gas void fraction of 50%. figure b.2. experimentally measured and predicted head by different models versus flow rate at intake pressure of 0.69 mpa and gas void fraction of 15%. 86 pdf compressor mech090603.qxd the journal of engineering research vol. 7, no. 1, (2010) 53-61 1. introduction in recent years, the development of metal matrix composites (mmcs) has been receiving worldwide attention on account of their superior strength and stiffness in addition to high wear resistance and creep resistance compared to their corresponding wrought alloys. the ductile matrix permits the blunting of cracks and stress concentrations by plastic deformation and provides a material with improved fracture toughness (aigbodion. v.s, 2007). the present trend, therefore, seems to be towards the development of discontinuously reinforced metal matrix composites which are gaining widespread acceptance primarily because they have recently become available at a __________________________________________ *corresponding author’s e-mail: aigbodionv@yahoo.com relatively low cost compared to uni-and multi-directional continuous-fiber reinforced mmcs and the availability of standard or near-standard metal working methods which can be utilized to form these mmcs (aigbodion, v.s, 2007, and clyne, t. w, 2000). nowadays, research all over the globe is focusing mainly on aluminium (aigbodion, v.s., 2007, and yaro, s.a., et al. 2006) because of its unique combination of good corrosion resistance, low density and excellent mechanical properties. the unique thermal properties of aluminium composites such as metallic conductivity with a coefficient of expansion that can be tailored down to zero, add to their prospects in aerospace and avionics (sagail, a. and leisk, g., 1992). effect of thermal ageing characteristics of al-si-fe/sic particulate composite synthesized by double stir casting v.s. aigbodion* and s.b. hassan * department of metallurgical and materials engineering, ahmadu bello university, samaru, zaria, nigeria received 3 june 2009; accepted 6 september 2009 abstract :the effect of thermal ageing on the microstructure and properties of 10wt% and 20wt%sic particulate reinforced al-si-fe matrix composite, produced by double stir casting route, have been studied. the composite samples were solution heat-treated at 500oc for 3 hrs and aged at 100, 200, and 300oc with ageing time between 60 and 660 minutes. the ageing characteristics of these grades of composite were evaluated using hardness values, impact energy, tensile properties and microstructure. the tensile strength, yield strength, hardness values increased as the percentage of silicon carbide increased in the alloy with decreased impact energy in both the as-cast and thermally age-hardened samples. the increases in hardness values and strength during thermal ageing are attributed to the formation of coherent and uniform precipitation in the metal lattice. it was found that both grades of composites showed acceleration in thermal ageing compared to the monolithic alloy. however, the 20wt%sic reinforced composite showed more acceleration compared to 10wt%sic reinforced composite. keywords: al-si-fe/sic, thermal ageing, composite, mechanical properties, microstructure and precipitation äéäjõl ö«cîd …qgô◊g ωoé≤àdg ¢ugƒn òkcéjal-si-fe/sic∞yé° ÿg ö°üdg ᣰsgƒh á©æ°üÿg ¿é°ùm.ü .¢s , ¿ƒjoƒñμjcg .¢s . ± áá°°uuóóÿÿgg äéø°ugƒe h ájhôμ«ÿg ägaé°ûf’g ≈∏y …qgô◊g ωoé≤àdg òkcéj :10wt% and 20wt%sic ¬ë∏°ùe äéäjõlal-si-feá≤jôw ᣰsgƒh , èàæe ,¬ñcôe áaƒø°üe ¬jƒäe álqo 300 ,200,100 ≈∏y ¬eoé≤j ”h äéyé°s çók i~ÿ ájƒäe 500 álqo ≈∏y éjqgôm πeé©ÿg §«∏ÿéh â∏eƒy ~b áñcôÿg êpéªædg .¬à°sgqo â“ ~b »féæãdg ö°üdg º«b .…hôμjéÿg ö«cîdg ∂dòch ,~°ûdg ¢ugƒn ,ωg~£°u’g ábéw ,√oó°üdg º«b ᣰsgƒh 顪««≤j ” ~b äéñcôÿg √ò¡d ωoé≤àdg ¢ugƒn .á≤«bo 660 ¤g 60 √~ÿ âbƒdg ωoé≤j ™e ioó°üdg º«b ioéjr ¿eg .éjqgôm áeoé≤àÿg êpéªædg ö∏°üj h ö°üdg øe πc ‘ ωg~£°u’g ábéw ioéjõh áμ«ñ°ùdg ‘ ¿ƒμ∏°ùdg ~«hqécg áñ°ùf ioéjõh ogoõj ioó°ü ≈∏y π°ü– ~°ûdg iƒb áfqé≤e …qgô◊g ωoé≤àdg ‘ π«é©j âæ«h ~b ¬øæ°üÿg äéñcôÿg øe πc ¿g ±é°ûàcg ” .¿~©ÿg áμñ°t ‘ á∏°üàe h ᪶àæe äéñ°sôj øjƒμj ¤g iõ©j …qgô◊g ωoé≤àdg aéækc’ iƒ≤dg h ¿eéa ∂dp ™eh .¬éàæÿg äéμ«ñ°ùdéh20wt% sic :¤g áfqé≤déh ìccg π«é©j úñj öcôÿg í∏°ùÿg 10wt% sic .öcôÿg í∏°ùÿg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg :al-si-fe/sic .ö«°sôj øjhôμj ée ö«côj ,á«μ«féμ«e ¢ugƒn ,öcôe ,…qgôm ωoé≤j , 54 the journal of engineering research vol. 7, no. 1, (2010) 53-61 the age-hardening characteristics of an alloy are generally modified by the introduction of reinforcement. these modifications are due to the manufacturing process, the reactivity between the reinforcement and the matrix, the size, the morphology and volume fraction of the reinforcement (sagail, a., leisk, g., 1992; rajan, t.v. and sharma, c.p, 1988). in contrast to this extensive data available on the heat treatment of al-si-fe alloys with convectional alloying elements (yaro, s.a., et al. 2006). relatively little is demanded for their ageing hardening characteristics particularly in silicon carbide reinforcement (aigbodion, v.s., 2007). hence, there is need for research to be carried out in this very important area. the aims of the present research are to determine the microstructure and thermal ageing characteristics of al-si-fe/sic particulate composites with a view to obtaining the optimum thermal age hardening procedure that would enable the achievement of the desired mechanical properties. 2. experimental procedure 2.1 materials the experimental materials used in this study included: silicon carbide with an average particle size of 10µm, ferrosilicon, high purity aluminium electrical wire obtained from northern cable company nocaco (kaduna), moulding boxes, silica sand, and bentonite obtained from national metallurgical development centre, jos, nigeria. 2.2 equipment the equipment used in this study included: pyrometer, mechanical stirrer, a crucible, an electrical resistance furnace, a rockwell hardness tester, a charpy impact machine, a tinus olsen tensile machine and a metallurgical microscope with a built-in camera. 3. methods of sample production the synthesis of the metal matrix composite that was used in this research was achieved using a double stircasting method at the foundry shop of the national metallurgical development center, jos, nigeria. the samples were produced by keeping the percentage of iron and silicon constant, reinforced with silicon carbide particles of 10wt% and 20wt%. high-purity aluminium electrical wires was charged in a graphite crucible kept in an electric resistance furnace and 0.01%nacl-kcl powder were used as a cover for melting the alloy in order to minimize oxidation of aluminium by excluding oxygen and creating a protective atmosphere inside the furnace. after the melting of the pure aluminium, the temperature of the furnace was raised to 720oc for the purpose of superheating the aluminium melt. the required quantity of 2.8% silicon and 0.8%iron was added to the melt using ferrosilicon, and the melt was thoroughly stirred (aigbodion, v.s., 2007; aigbodion,v.s. and hassan, s.b., 2007). with progressive melting, the furnace temperature was raised to 780oc and the melt was held at this temperature for 10 minutes. the melt was then skimmed to remove the oxides and impurities. the molten metal was continuously stirred in order to ensure a near-uniform distribution of alloying elements and prevent the elements from settling at the bottom due to their higher density. for each melting, 200g of charge materials was used to produce the alloy. the silicon carbide particles were preheated to 1000oc, to make their surface oxidized (aigbodion, v.s. and hassan, s.b., 2007). the alloy was then cooled down to a temperature just below the liquidus point (580oc) to keep the slurry in a semi-solid state. at this stage the preheated sic particles were added and mixed manually. manual mixing was used because it was very difficult to mix using automatic device when the alloy was in a semi-solid state. after manual mixing was done, the composite slurry was reheated to a temperature of 720oc and then automatic mechanical mixing was carried out for about 20 minutes at an average stirring speed of 150 rpm. in the final mixing processes, the furnace temperature was controlled between 730 and 740oc and the pouring temperature was controlled to about 720oc (aigbodion, v.s., 2007; aigbodio, v.s and hassan, s.b., 2007). a preheated sand mould with diameter 18 mm and 300mm length was used to produce cast bars. after casting, the samples were machined into tensile, impact and hardness test specimens for the purpose of determining the mechanical properties. 4. heat treatment of the samples the test samples were solution heat-treated at temperature of 500oc in an electrically heated furnace, soaked for 3 hours at this temperature and then rapidly quenched in warm water at 65oc. thermal ageing of the test samples was carried out at temperatures of 100, 200 and 300oc, for various ageing times of 60 to 660 minutes, and then cooled in air. the thermal ageing characteristic of these grades of composites was evaluated using hardness values obtained from solution heat-treated samples of the investigated composites subjected to the aforementioned temperature conditions. the tensile and impact tests were conducted at the peak ageing time of the various ageing temperatures (aigbodion, v.s., 2007; rajan, t.v. and sharma , c.p., 1988). 5. determination of hardness values the hardness values of the samples were determined according to the provisions in astm e18-79 using the rockwell hardness tester on "b" scale (frank well test rockwell hardness tester, model 38506) with a 1.56 mm steel ball indenter, minor load of 10kg, major load of 100kg and hardness value of 101.2 hrb as the standard block. before the test, the mating surfaces of the indenter, plunger rod and test samples were thoroughly cleaned by removing dirt, scratches and oil and the testing machine 55 the journal of engineering research vol. 7, no. 1, (2010) 53-61 was calibrated using the standard block. the samples were placed on anvils, which acted as support for the test samples. a minor load of 10kg was applied to the sample in a controlled manner without inducing impact or vibration and the zero datum position was established. the major load of 100 kg was then applied, and the reading was taken when the large pointer came to rest or had slowed appreciably and dwelled for up to 2 seconds. the load was then removed by returning the crank handle to the latched position and the hardness value read directly from the semi automatic digital scale. three indentations were taken and the average represents the hardness values (aigbodion, v.s., 2007 and metal handbook., 1985). 6. determination of the tensile properties the tensile properties of the as-cast and peaking thermal ageing samples were conducted on tinus-olsen tensile testing machine with a strain rate of .002s-1. the test pieces were machined to the standard shape and dimensions as specified by the american society for testing and materials (annual books of astm standards, 1990). the sample was locked securely in the grips of the upper and lower crossbeams of the testing machine. a small load was initially applied to seat the sample in the grips and then the load was increased until failure occurred. 7. impact strength determination the impact test of the as-cast and peaking thermal ageing samples was conducted using a fully instrumented avery denison test machine. the mass of the hammer was 22 7 kg and the striking velocity was 3.5 m/sec. charpy impact tests were conducted on notched samples. standard square impact test sample measuring 75 x 10 x 10 mm with notch depth of 2 mm and a notch tip radius of 0. 02 mm at an angle of 45o was used (annual books of astm standards, 1990). before the test sample was mounted on the machine, the pendulum was released to calibrate the machine. the test samples were then gripped horizontally in a vice and the force required to break the bar was released from the freely swinging pendulum. the value of the angle through which the pendulum had swung before the test sample was broken corresponded with the value of the energy absorbed in breaking the sample and this was read from the calibrated scale on the machine (annual books of astm standards, 1990). 8. microstructural examination metallographic samples were cut from the produced samples. the cut samples were then mounted in bakelite, and mechanically ground progressively on grades of sicimpregnated emery paper (80-600 grits) sizes using water as the coolant. the ground samples were then polished using one-micron size alumina polishing powder suspended in distilled water. final polishing was done using 0.5 micron alumina polishing powder suspended in distilled water. following the polishing operation, etching of the polished specimen was done using keller's reagent. the structure obtained was photographically recorded using an optical microscope with built-in camera (aigbodion, v.s., 2007; aigbodion, v.s. and hassan. s.b., 2007). 9. results and discussion 9.1 results the various microstructures developed for the two grades of composites in the as-cast and thermally agewt% sic as-cast peak aged at 100oc peak aged at 200oc peak aged at 300oc 0 31.1 60.1 61.5 62.0 10 57.2 75.0 74.5 84.0 20 67.0 86.0 87.0 89.0 table 1. hardness (hrb) values for as-cast and thermal age-hardened with 10wt% and 20wt% of silicon carbide wt% sic as-cast (n/mm2) peak aged at 100oc (n/mm2) peak aged at 200oc (n/mm2) peak aged at 300oc (n/mm2) 0 45.60 59.90 56.40 52.40 10 68.46 73.35 75.00 73.00 20 79.98 80.75 82.50 80.34 table 2. yield strength of al-si/sic composites with 10wt% and 20wt% of silicon carbide wt % sic as-cast (n/mm2) peak aged at 100oc (n/mm2) peak aged at 200oc (n/mm2) peak aged at 300oc (n/mm2) 0 58.69 62.50 67.18 64.20 10 96.06 98.50 98.50 100.20 20 106.12 120.43 125.00 121.85 table 3. tensile strength of al-si-fe/sic composites with 10wt% and 20wt% of silicon carbide 56 the journal of engineering research vol. 7, no. 1, (2010) 53-61 hardened condition are shown in micrographs 1-12. the thermal ageing response measured by the variation of hardness with time for two grades of composites is shown in figs. 1-3. table 1shows the hardness values of both the thermally as-cast and age-hardened al-si-fe/sic composites. the results of yield strength and ultimate tensile strength with weight fraction of silicon carbide at various ageing temperatures are shown in tables 2-3. the results of impact energy with weight fraction of silicon carbide (sic) at various ageing temperatures are shown in table 4. 10. discussions 10.1 thermal ageing characteristics of al-si-fe/ sic composites from figs. 1-3, it can be seen that there is a steep rise in the hardness values of each grade of the composite at initial stages for all ageing temperatures and then a fall after reaching the various peak ageing times are reached, corresponding to over-ageing. however, at higher ageing temperature the materials developed peak hardness at shorter ageing time, because the rate of precipitation of the second phase materials is faster and hence increases in hardness values. the time to obtain peak hardness is shorter according to the sequence: 100oc > 200oc > 300oc (see figs. 1-3). the thermal age-hardening behavior of the al-sife/sic particulate composites are similar to al-si-mg/sic particulates as reported by cottu et al. (1992) ie. hardness continuously increases to a maximum during thermal ageing and then decreases later due to over ageing. it is interesting to note that in the reinforced aluminium alloy metal-matrix, as the volume fraction of sic increase to 20wt% in the aluminium alloy there is a monotonic reduction in the time required to reach peak hardness (see fig. 3). the 20wt%sic addition, yielded the highest hardness value. as far as hardening behavior of the composites is concerned, particle addition in the matrix alloy increases the strain energy in the periphery of the particles in the matrix and these tendencies may be due to the formation of the dislocation at the boundary of the ceramic particles by the difference in the thermo-expansion coefficient between the matrix and ceramic particles during solution treatment and quenching since a lot of dislocations generate in the main matrix/particle interface (bedir, f., 2006). thus, dislocations cause the hardness increase in composite as well as residual stress increase because it acts as non-uniform nucleation sites in the interface following the age treatment. it is thought that the higher the amount of the ceramic particles in the matrix, the higher the density of the dislocation, and as a result, the higher the hardness of the composite. 10.2 microstructural analysis microstructures of the as-cast and heat-treated composites at various ageing temperatures and peaking ageing time were examined by the use of a metallurgical microscope. they were observed to contain primarily -al, silicon eutectic and sic particle as shown in micrographs 112. in the microstructure of the composites si phase was seen apparently in the eutectic regions, and intermetallic compounds such as fesial3, iron-containing phases due to the secondary alloying elements (aigbodion, v.s., 2007; rajan, t.v. and sharma, c.p, 1988). the microstructure of the as-cast unreinforced al-sife alloy is shown in micrograph 1, while micrographs 23 show the microstructure of the as-cast reinforced alloys with sic particles. the microstructure reveals that there are small discontinuities and a reasonably uniform distribution of sic particles. the ceramic phase is shown as dark phase, while the metal phase is white. these structures are in agreement with the co-continuous interlaced phases studied by other researchers (rohatgi, p.k., et al. 1988 and whitehouse, a.f., 1990). micrographs 4-12, show the 0 2 0 4 0 6 0 8 0 1 0 0 0 2 0 0 4 0 0 6 0 0 a g e in g t im e ( m in u t e s ) h a r d n e s s v a lu e s ( h 0 w t% s i c 1 0 w t% s i c 2 0 w t% s i c ageing time (minutes) h ar dn es s v al ue s (h r b ) figure 1. variation of hardness values with ageing time at aageing temperature of 100oc ageing time (minutes) h ar dn es s v al ue s (h r b ) figure 2. variation of hardness values with ageing time at ageing temperature of 200oc ageing time (minutes) h ar dn es s v al ue s (h r b ) figure 3. variation of hardness values with ageing time at ageing temperature of 300oc 0 200 400 600 100 80 60 40 20 0 0wt% sic 10wt% sic 20wt% sic 0 200 400 600 100 80 60 40 20 0 0wt% sic 10wt% sic 20wt% sic 0 200 400 600 100 80 60 40 20 0 0wt% sic 10wt% sic 20wt% sic 57 the journal of engineering research vol. 7, no. 1, (2010) 53-61 microstructure of the thermally age-hardened al-sife/sic composites. the microstructure reveals the dissolution and distribution of the sic particles in the metal matrix and the presence of precipitates at the particles matrix interfaces with precipitation and dissolution of the sic particles and silicon eutectic phase (aigbodio, v.s., 2007; aigbodion, v.s. and hassan, s.b., 2007). the formation and presence of precipitates at the particlesmatrix interfaces may be appreciated by comparing micrographs of the composite in the as-cast state micrographs 1-3 and in the peak age-hardened state micrographs 4-12. the micrographs in the peak age-hardened state reveal precipitates covering the surface at the particles-matrix interfaces. cottu et al. (1992) reported a similar behavior of increased precipitation at interfaces for aluminum alloy 6061 reinforced with whiskers of silicon carbide. 10.3 hardness value from table 1, the hardness values for thermally as-cast and age-hardened composite increases as the percentage of silicon carbide increases from 10wt% to 20wt% in the alloy. this is due to an increase in the percentage of the hard and brittle phase of the ceramics body in the alloy. the hardness values of the thermally age-hardened alloy are almost equal at various ageing temperatures with percentage sic additions (see table 1), in line with the earlier observation of (cottu et al. 1992 and ikechukwuka 1997). 10.4 yield and tensile strength from tables 2-3, the yield strength and ultimate tensile strength of both the as-cast and age-hardened composites increased with the increased percentage silicon carbide to 20 wt% sic. the as-cast samples have a value of 79.98 and 106.12n/mm2 for yield and tensile strength at 20% sic addition. for the thermally age-hardened samples, similar trends were also observed as for the case of the as-cast samples, in that the yield strength and ultimate tensile strength increased to 20% wt of sic. the ultimate tensile strengths of the thermally age-hardened samples at 20 wt% sic are 120.43, 125.00 and 121.85n/mm2 at peak ageing of 100, 200 and 300oc respectively. (see table 3). also the yield strengths of age-hardened samples at 20 wt% sic addition are 80.75, 82.50 and 80.34n/mm2 at peak ageing of 100, 200 and 300oc respectively (see table 2). the higher values of both ultimate tensile strength and yield strength obtained at 20 wt% sic are attributed to the uniform distribution of the sic particles in the microstructure of the 20 wt% sic (see micrographs 3, 10-12). it can be seen that the inter-particle distance is smaller. the contribution of inter-particle distance to strain hardening arises from the fact that the space permitted a dislocation to maneuver round obstacles limits the yield stress given by the formula (dieter, g.e., 1988): the smaller the values of  the greater the yield stress as a result of strain hardening. it therefore supports the fact that (micrographs 3,10-12) with smaller . a substantial increase in both the yield strength and ultimate tensile strength was obtained when compared with the results of as-cast samples. for example the yield strength of al-si-fe/10sic particles increased from 68.46 to 75.00n/mm2, while the ultimate tensile strength increased from 96.06 to 98.50n/mm2 in the as-cast and peak age-hardened at 200oc respectively. the yield and tensile strength of al-si-fe/20sic particles also increased from 79.98 to 82.50 n/mm2 and 106.12 to 125.00 n/mm2 in the as-cast and thermally age-hardened at 200oc temperature (see tables 2-3). these results show that the al-si-fe/10sic particles samples has 9.55% and 2.54% increases in yield strength and ultimate tensile strength respectively over those of the as-cast samples at peak ageing at 200oc, while for the alsi-fe/20sic particles the respective increments were 3.15% and 17.79% for yield strength and ultimate tensile strength. the increase in strength of the thermally agehardened samples over those of the as-cast are attributed to the uniform distribution of sic in the aged samples than the as-cast samples respectively (see micrographs 1 and 3 and micrographs 4-12). sagail and leisk (1992) and whitehouse et al. (1991) attributed this to an increase in dislocation density at the interfaces. an increase in dislocation density strain hardens the metal-matrix locally and provides heterogeneous nucleation sites for precipitation, thereby accelerating the ageing response. the modulated structures formed during ageing in these grades of composites enhanced these properties and this is in agreement with the earlier works on al-si-mg/sic (cottu, j.p. et al. 1992 and ikechukwuka. n.a, 1997). the addition of sic particles affects not only the precipitation kinetics but also the relative amounts of the various phases present in the microstructure. besides during quenching from the solution heat treatment temperature, the sic particles cool more slowly than the matrix (since the particles have a lower thermal conductivity). this causes the matrix around the particles to be wt% sic as-cast peak aged at 100oc peak aged at 200oc peak aged at 300oc 0 17.0j 28.0j 24.0j 22.0j 10 14.5j 20.0j 19.0j 19.0j 20 12.0j 17.0j 15.5j 16.0j table 4. impact energy for as-cast and age-hardened with 10wt# and 20wt# of silicon carbide ô = gb  . ô = yield stress, ë = inter -particle distance, g = shear energy of dislocation and b = burger vector. 58 the journal of engineering research vol. 7, no. 1, (2010) 53-61 micrograph 1. microstructure of the al-si-fe alloy, showing silicon eutectic (black) containing feal3si phases and al (white) (x200). the structure reveals the eutectic silicon containing fesial3 phase in aluminium matrix. micrograph 2. microstructure of the al-si-fe alloy with 10 wt% of sic the the structure reveals the dissolution of the eutectic silicon phase and uniform distribution of sic (black) in al matrix (white) (x200). micrograph 3. microstructure of al-si alloy reinforced with 20 wt% of sic. the structure reveals and dissolution of the eutectic silicon phase and uniform distribution of sic (black) in -al matrix (white) (x200). micrograph 4. microstructure of the al-si-fe alloy after peak aged at 100oc (x200). the structurre reveals dissolution and precipitation of the silicon eutectic (black) and the feal3si phases in  al matrix (white). 59 the journal of engineering research vol. 7, no. 1, (2010) 53-61 micrograph 5. microstructure of the al-si-fe alloy after peak aged at 200oc (x200). the structure reveals uniform dissolution and precipitation of the silicon eutectic (black) and the feal3sui phases in -al matrix (white). micrograph 6. microstructure of the unreinforced alloy after peak aged at 300oc (x200). the structure reveals dissolution and precipitation of the silicon eutectic (black) and the feal3si phases with some segregation of silicon eutectic phases in -al matrix (white). micrograph 7. microstructure of the reinforced alloy (10 wt% of sic) after peak aged at 100oc (x200). the structure reveals the dissolution of the eutectic silicon phase and uniform distribution of sic (black) in -al matrix (white). micrograph 8. microstructure of the reinforced alloy (10 wt% after peak aged at 200oc (x200). the structure reveals the dissolution of the eutectic silicon phase and distribution of sic (black) in  -al matrix (white). 60 the journal of engineering research vol. 7, no. 1, (2010) 53-61 micrograph 9. microstructure of the reinforced alloy (10 wt% of sic after peak aged at 300oc (x200). the structure reveals the dissolution of the eutectic silicon phase and uniform distribution of sic (black) with precipitation of the sic particles. micrograph 10. microstructure of the reinforced alloy (20 wt% of sic) after peak aged at 100oc (x200). the structure reveals the dissolution of the eutectic silicon phase and uniform distribution of sic (black) with precipitation covering. micrograph 11. microstructure of the reinforced alloy (20 wt% of sic) after peak aged at 200oc (x200). the structure reveals the dissolution of the eutectic silicon phase and well uniformly distribution of sic (black) with precipitation covering the aluminium phase (white). micrograph 12. microstructure of the reinforced alloy (20 wt% of sic) after peak aged at 300oc (x200). the structure reveals the dissolution of the eutectic silicon phase and uniform distribution of sic (black) with precipitation covering. 61 the journal of engineering research vol. 7, no. 1, (2010) 53-61 warmer than the bulk matrix. the high dislocation density and the high solubility centers (the warm particlesmatrix interfaces) in the composite are favorable conditions for precipitation formation. this mechanism is a possible explanation for the higher volume fraction of precipitates obtained in the composite of the age hardening condition (aigbodion, v.s., 2007; aigbodion, v.s. and hassan, s.b., 2007). 10.5 impact energy from the table 4, the impact energy of both the as-cast and age-hardened samples, decreased as the percent sic addition increases in the alloy. the brittle nature of the reinforcing materials (sic) plays a significant role in degrading the impact energy of the composite, since the unreinforced alloy and the alloy with 10wt% sic particles have the highest impact energy, indicating that they are the toughest of them all (aigbodion, v.s., 2007). 11. conclusions from the results of the investigations, the following conclusions have been made: 1) the uniform distribution of the sic particles in the microstructure of both the as-cast and thermally agehardened al-si-fe/sic composites is the major factor responsible for the improvement in the mechanical properties. 2) the addition of sic particles affects not only the precipitation kinetics but also the relative amounts of the various phases present in the microstructure, and that during quenching from the solution heat treatment temperature, the sic particles cool more slowly than the matrix (since the particles have a lower thermal conductivity). this causes the matrix around the particles to be warmer than the bulk matrix. the high dislocation density and the high solubility centers (the warm particles-matrix interfaces) in the composite are favorable conditions for precipitation formation. this mechanism is a possible explanation for the higher volume fraction of precipitates obtained in the composite of the age-hardening condition. 3) this research therefore, has established that these grade of composites respond to precipitation-hardening heat treatment. references aigbodion, v.s., 2007, “development of al-si-fe/sic particulate composite as advanced materials for engineering applications,” m.sc thesis, department of metallurgical engineering, ahmadu bello university, zaria, nigeria, pp. 45-50. aigbodion,v.s. and hassan, s.b., 2007, “ effect of silicon carbide reinforcement on microstructure and properties of cast al-si-fe/sic particulate composites,” materials science and engineering a: 447, pp. 355-360. annual books of astm standards, 1990, section 1. iron and steel products, vol. 261 01.04, pp. 13-63. bedir, f., 2006, "fabrication of two layered al-b4c composites by conventional hot pressing under nitrogen atmosphere and their characterization,” j. of mech.sci. and tech. ksme, vol. 20(7), pp. 10021011 clyne, t.w., (ed.), 2000, “comprehensive composite materials,” vol. 3, metal matrix composites (ser. eds), pp. 1-8. cottu, j.p., coudere, j.j., viguier, b. and bernard, l., 1992, “ j. mater sci 27,” pp. 3068-3074. dieter.g.e., 1988, “ mechanical metallurgy,” mcgrawhill company, pp. 184-223. fishman, s.g. and dhingra, a.k., 1988, “ reinforced metal composites,” materials park, oh: asm international publication, p. 249. ikechukwuka, n.a., 1997, “characterization of aluminium alloy 2618 and its composites containing alumina particles,” unpublished phd thesis department mechanical engineering, university of askatchewan, saskatoon, pp. 87-200. metals handbook, mechanical testing, 1985, 9th edition, vol. 8, american society for metals, the united states of america, pp. 250-257 rajan, t.v. and sharma, c.p., 1988, “heat treatment principles and techniques,” published by prenticehall of india private limited, pp. 35-159. rohatgi, p.k., yarandi, f.m. and liu, y., proc. of inter. symposium on advances in cast. sagail, a., leisk, g., 1992, “heat treatment optimization of alumina/aluminium metal matrix composites using the taguchi approach,” scripta metallurgica et materials, pp. 871-876. yaro, s.a., aigbodion, v.s. and mohammed, o.y., “effects of copper addition on the mechanical properties of al-si-fe alloy,” international research journal in engineering & technology (irejest) futo, owerri, nigeria, pp. 32-39. whitehouse, a.f., shahani, r. a. and clyne, t.w., 1991, metal matrix composites: processing, microstructure and properties,” 12, international symposium roskilde, n.hansen et al. eds, ris national laboratory, denmark, pp. 74 1-748. chen060214.qxd the journal of engineering research vol. 5, no.1 (2008) 47-54 1. introduction activated carbon is used in many applications due to its adsorption capacity and large internal surface area. these applications are partly facilitated by the fact that the carbon precursor or the preparation conditions could be chosen to achieve a unique set of desirable properties (luehrs, et al. 1998; wartelle, et al. 2000 and byrne, et al. 1997). both naturally occurring and synthetic carbonaceous solid have been used (gergova, et al. 1997; jagtoyen, et al. 1992; benaddi, et al. 2005; yalcm, et al. 2000 and nagano, et al. 2000). though new synthetic precursors such as polymeric fibers and phenolic-resin have been reported to produce high surface area products (cai, et al. 2004), agricultural wastes are increasingly being investigated as precursors (wartlle, et al. 2004; byrne, et al. 1997; gergova, et al. 1996; yalcm, 2000; liou, 2004 and carvalho, et al. 2003). wood is the most favored source of activated carbon production (35%) followed by coal (28%), lignite (14%), coconut shell (10%), peat (10%) and others (3%) (byrne, et al. 1977). _____________________________________ *corresponding author’s e-mail: baba@squ.edu.om other lignocellulosic materials are also used which include nutshell (wartelle, et al. 2000), apricot stones (gergova, et al. 1996), rice husks (yalcm, et al. 2000), cork waste (carvalho, et al. 2003) and dates pits (caturla, et al. 1991; molina-sabio, et al. 1995; wigmans, et al. 1989; pelekani, et al. 2000 and daifullah, et al. 2003). the activating agents for preparing activated carbons can be used before or after carbonizing the raw materials. some of the activating chemicals used in this process are boric acid, phosphoric acid, nitric acid and zinc chloride (girgis, et al. 2002 and el-hendawy, 2006). once carbonization is achieved, the agents most widely used to activate the char are: steam, carbon dioxide, air or their combinations. for example, when phosphoric acid is used as an activation chemical, the acid react with the lignocellulosics thermal decomposition products. this leads to bond weakening and formation of cross-linked structure. this reduces the release of volatile materials, restricts tar formation and loss of porosity (teng, et al. 1998). in addition, the effect of the chemical agent leads to lower activation temperature and higher carbon yields (el-akkad, et al. 19978). study of date palm stem as raw material in preparation of activated carbon o. houache1, r. al-maamari2, b. al-rashidi2 and b. jibril*2 1oman polypropylene llc, sohar industrial port complex, p.o. box 277, pc 322, falaj, al-qabail, sohar, oman *2department of petroleum and chemical engineering, sultan qaboos university, p.o. box 33, al khod, pc 123, muscat, oman received 14 february 2006; accepted 10 march 2007 abstract: activated carbon adsorbent was prepared using omani date palm tree stem as a precursor. precursor samples were subjected to thermal treatment (at 400, 500 and 600 oc) before or after impregnation with either h3po4 (85 wt %) or koh (3 wt %). the activated carbon obtained was characterized by bet (surface area and porosity), gas pycnometry (true density) and sem (texture). sample subjected to carbonization, without chemical activation, exhibited low surface areas ~ 1.0 m2/g at 400 and 500 oc and 124 m2/g at 600 oc. further treatment of such samples with either the acid or the base did not show improvement in surface area or other properties. impregnations of the precursor with acid before carbonization significantly improved the surface area to as high as 1,100 m2/g at a carbonization temperature of 500 oc. thus, activated carbon with a moderate surface area could be produced from date palm stem using low carbonization temperature. keywords: activated carbon, chemical activation, bet surface area, porosity, sem, date palm stem §°ûæÿg ¿ƒhôµdg ò° – ‘ ωén ioéªc πîædg ´hòl ωg~îà°sg á°sgqo ¢tgƒm ôªy1…ôª©ÿg ~°tgq ,2…~°tgôdg q~h ,22*πjèl éhéh , áá°°uuóóÿÿgg,500 ,400 iqgôm äélqo ≈∏y ájqgô◊g á÷驪∏d â©° ncg πîædg ´hòl äéæ«y ,§°ûæÿg ¿ƒhôµdg êéàfe’ ωén ioéªc »∏ùg πîædg ´hòl ωg~îà°sg ” á°sgq~dg √òg ‘ : á«ë£°ùdg ámé°ùÿg ¢sé«b ≥jôw øy §°ûæÿg ¿ƒhôµdg ioƒl º««≤j ” .…~yéb hg »° ªm ∫ƒ∏fi ‘ π«ªëàdg h ôª¨dg ∫ón øe »f髪«µdg §«°ûæàdg ~©h hcg πñb ájƒäe álqo 600 h â©° ncg »àdg iônc’g äéæ«©déh áfqé≤e á«ë£°ùdg ámé°ùÿg å«m øe i~«l èféàf â£ycg ájqgô◊g á÷é©ÿg πñb ¢ ª◊éh é¡à÷é©e ” »àdg äéæ«©dg .á«æñdg h áaéãµdg h á«eé°ùÿg h .ájqgô◊g á÷é©ÿg ~©h é«f髪«c é¡£«°ûæj ” »àdg ∂∏j hg ≥ñ°ùe »f髪«c §«°ûæj ¿h~h ájqgô◊g á÷驪∏d áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.πîædg ´hòl ,êhîµd’g ô¡ûéh í°ùÿg ,á«eé°ùÿg ,á«ë£°ùdg ámé°ùÿg ,»f髪«µdg §«°ûæàdg ,§°ûæÿg ¿ƒhôµdg : 48 the journal of engineering research vol. 5, no.1 (2008) 47-54 based on the foregoing, we found it of interest to study different lignocellulosics materials as sources of activated carbon. here, we report phosphoric acid and potassium hydroxide treatment of an omani date palm stem for preparing activated carbon. oman and neighboring areas have millions of date-palm-trees. every year tons of date palm tree trunk and stem are disposed off. this gives a free or low-cost, abundantly supplied precursor. as stated above, many workers have reported different aspects of production of activated carbon from date pits, but to our knowledge, there are no reports on the use of date-palmtree stem for such endeavor. the aim of this work was to prepare and characterize activated carbons from date palm stems using chemical activation method which will contribute to further understanding the effects of activating chemical on the precursor's surface and other properties of the final products. 2. experimental setup activated carbon was prepared and characterized at the petroleum and chemical engineering laboratory, sultan qaboos university, oman. 2.1 sample preparation date palm stems were obtained from nizwa region of oman. they consist of a major outer part having softmesh and channel-like structure and inner hard part having longitudinal fibrous structure, as shown in fig. 1. samples were prepared by cutting slices of 1 cm thick from different parts of the stem before drying them in a muffle heater at 105 oc for 24 hrs. longer drying time did not show observable change in the samples weight. the dried samples were then used as precursors to obtain char and activated carbon. before or after activation (impregnation) with either h3po4 (85 wt %) or potassium hydroxide koh (3 wt %) the weighed samples were carbonized in a muffle furnace under nitrogen atmosphere. the chemical activation was accomplished by having 30g of each dried sample immersed in 100g of activating agents [either h3po4 (85 wt%) or koh (3 wt%)] at room temperature then increasing the heating temperature to 85 oc at a heating rate of 15 oc/min. the sample was exposed to this temperature for 3 hrs before cooling to ambient temperature and left at rest for 16 hrs. these treated samples were then dried at 105oc. in all cases, carbonization was done under flow of nitrogen (250 ml/min). it took place by heating the samples at 100 oc for 20 min before raising the temperature in steps of 50 oc up to the desired temperature (400, 500 or 600 oc). the final temperature was held constant for two hours before cooling the sample to ambient temperature. density and surface analysis took place after grounding and sieving the samples to sizes between 212 and 250 microns. in the test, , the samples are represented based on the activation procedure as follows: s-untreated = date palm stem; s-c = carbonized date palm stem; s-c-h3po4 = carbonization of date palm stem followed by acid activation; s-c-koh = carbonization of date palm stem followed by base activation; s-h3po4-c = acid activation followed by carbonization of date palm stem and s-kohc = base activation followed by carbonization of date palm stem. 2.2 characterization of samples samples of the activated carbon were characterized by physical adsorptions of n2 at 77.4 k, using a quantachrome autosorb 1c analyzer. the properties obtained are: (i) bet surface area (sbet, m2/g), (ii) total pore volume from nitrogen held as liquid at p/po = 0.95 (vp, cm3/g) and (iii) average pore diameter. prior to the analysis, samples were cleaned by heating them, under flow of helium, at 300 oc for 3 hrs. densities of the samples were measured by displacement of helium using a quantachrome ultrapycnometer 1000. the textures of samples of the products were observed using sem scanning electron microscope (jeol jsm 840a). it was equipped with an energy dispersive x-ray analyzer that allowed semi-quantitative analysis of solids. 3. results and discussion the precursor was subjected to different carbonization times and activations and the decreases in sample weight were recorded. the char yield of date palm stem was compared with those obtained from other cellulose materials such as almond shell and coconut shell, in fig. 2. it could be observed that the char yielded from date palm stem is substantially higher than that from other lignocelluloic sources (al-rashidi 2004). carbonization without chemical impregnation of the precursor yielded char that exhibited surface areas of about 1.0 m2/g at 400 and 500 oc. the area increased to 124 m2/g at 600 oc. the yields were improved by impregnations of the precursor with either h3po4 or koh before carbonization. the surface area also increased. the base treatment increased the area figure 1. sample of the date palm stem 49 the journal of engineering research vol. 5, no.1 (2008) 47-54 by one order of magnitude at 400 or 500 oc, but led to an inferior product at 600 oc, as shown in table 1. on the other hand, acid treatment improved the surface areas at all temperatures: 690.9, 1101.0 and 632.2 m2/g at 400, 500 and 600 oc respectively. this suggests that by chemical treatment of date palm stem, the chemistry of the carbonization was changed. due to dehydration, different macromolecular networks were formed which resulted in promotion of cross linking of structure and suppression of tar formation during the carbonization (attia, 1997). furthermore, the increase and decrease in surface areas with the chemical treatment, indicate the complex and interacting effects of different changes in the sample. due to different degrees of interaction, some volatiles components may react and link to carbon structure, or other compounds could be formed thereby changing the product porosity and the extent of exposed surface area. the characteristic of the product was further explored by measuring the true densities. figure 3 shows increase in densities with increase in carbonization temperature. when the char formed was subjected to post-treatment with either acid or base, different behaviors were observed. for the char produced at 400 oc, impregnation with koh exhibited a little decrease in density. however, at higher temperatures, there was an increase of the density by 21.2 and 5.8% at 500 and 600 oc respectively. on the other hand, samples treated with acid, show a decrease in the densities; 33.5, 33.3 and 8.8% at 400, 500 and 600 oc respectively. in all cases, the samples were treated to adjust the ph of the water in which the char was suspended. depending on the surface morphology of the char, the koh may be binded with surface functional groups (shih, et al. 2005). such interactions would lead to corsslinking reactions with some of the volatiles thereby retaining them in the solid matrix. this would result into increase in the net mass of the sample, thereby increasing the density. addition of h3po4 on the other hand, may lead to dehydrogenation and release of gases, thereby decreasing the net mass and density of the char. impregnation of the precursor prior to carbonization lead to partly physical diffusion of the chemical to the internal structures of the samples and partly chemical reaction with compound in the structure. carbonization resulted into chars of different characteristics. their differences decrease with an increase in the temperature (fig. 3). all samples appear to converge to constant similar densities. this implies that the final products obtained may have similar true densities (shih, et al. 2005). the sem micrographs of (a) date palm stem and chars produced at (b) 400, (c) 500 and (d) 600 oc are shown in fig. 4. the micrograph of the precursor exhibited a dense 0 10 20 30 40 50 400 500 600 carbonization temperature, oc c ha r y ie ld (% ) date stem coconut shell regenerated cellulose figure 2. comparison of char yields for different precursors at different carbonization temperature temperature ( oc) sample 400 500 600 s-c 0.0936 0.7700 119.0 s-c-h3po4 1.2600 0.3610 31.50 s-h3po4-c 691.0 1100.0 632.0 (sbet, m2/g) s-koh-c 27.80 29.40 94.70 s-c 539.0 295.0 28.20 s-c-h3po4 104.0 67.00 41.40 s-h3po4-c 36.90 41.60 37.30 avg. pore diameter (å) s-koh-c 45.90 50.70 36.00 s-c 0.0012 6 0.00568 0.0839 s-c-h3po4 0.0033 0.00061 0.0326 s-h3po4-c 0.6370 1.1500 0.5890 (vp, cc/g) s-koh-c 0.0320 0.0372 0.0852 table 1. surface area, acverage pore diameter, and total pore volume of samples prepared at different heat treatment 1 1.5 2 2.5 350 450 550 650 carbonization temperature, oc d en si ty , ( cm 3 / g) s-c s-h3po4-c s-koh-c s-c-h3po4 s-c-koh figure 3. density of different samples after carbonization 50 the journal of engineering research vol. 5, no.1 (2008) 47-54 structure and smooth surface. the chars show irregular and flaky texture. interconnecting large pores that increase with the temperature could be observed. this suggests the increase in the release of the volatile compounds from the samples, in line with the increase in the true density. the surface textures were changed when the chars were subjected to h3po4 and koh (fig. 5). the texture became denser (higher for acid treated fig. 5(a-c), than base-treated samples fig. 5(d-f)) and exhibited pores of different sizes and shapes. this indicates the chemical effects on the samples or physical washing of organic matter from the surface. the base-treated samples exhibited grey particles covering the surface. acid and base impregnation before carbonization resulted into more observable porous surface (fig. 6 a-c (acid) and d-f (base)). again, pores of different sizes and shapes could be observed. the higher surface area of samples (fig. 6b) than that of (fig. 6c) could be related to the fact that heating above 500 oc destroyes the tiny pores structure and therefore reduces the available surface area. the smaller pore perimeters of samples in fig. 6d-f than that of fig. 6a-c suggest that the chemical washing done by the acid is superior to the one done by the base which might have led to higher surface area. the acid-impregnated samples exhibited surface area an order of magnitude higher than the base treated ones. the base treatment resulted into interaction that produced porous surface, with shallow pores, as indicated by the sem images. surface areas and porosities for the treated carbons are summarized in table 1. the surface area increases and the average pore diameter decreases with temperature. carbonized date palm stems are characterized by poorly developed porous structures of low surface area (0 119 m2/g), low pore volumes (0.00126 0.0839 cm3/g) and medium to large average pore diameter (28.2 539.0 å). subsequent acid treatment of chars reduced porosity development with a decreased surface area (0.36-31.5 m2/g), decrease in the pore volume (0.0006-0.0326 cc/g) and consequent increase in the mean pore dimension (41.4-104 å) compared to carbonization. perhaps inorganic materials started to react at lower temperature but on heat treatment to higher temperature when the pores were being reorganized; inorganic materials might have filled or blocked some portion of the macropores (lua, et al. 2004). activation with phosphoric acid and then carbonization at 500 oc led to the best porous product with surface area of 1100 m2/g, internal porosity of 1.15 cm3/g and a mean pore dimension of 41.6 å. it can be observed a b c d figure 4. sem micrographs of date palm stem chars produced at different temperatures: (a) precursor (as is), (b) 400oc (c) 500oc, (d) 600oc 51 the journal of engineering research vol. 5, no.1 (2008) 47-54 a b c d e f figure 5. sem micrographs of chars produced at different temperature and subjected for either h3po4 or koh: (a) s-400-acid, (b) s-500-acid, (c) s-600-acid (d) s-400 base, (e) s-500-base, (f) s-600-base 52 the journal of engineering research vol. 5, no.1 (2008) 47-54 a b c d e f figure 6. sem micrographs of chars produced by subjecting the precursor to either h3po4 or koh and then carbonized at different temperature: (a) s-acid-400, (b) s-acid-500, (c) s-acid-600, (d) s-base-400, (e) s-base-500, (f) s-base-600 53 the journal of engineering research vol. 5, no.1 (2008) 47-54 that activation with phosphoric acid and then carbonization showed superiority under all treatments. the improvement in surface characteristics when comparing phosphoric acid activation to carbonization were very highly significant; also phosphoric acid activation resulted in a highly significant improvement over potassium hydroxide activation. the results showed that the date palm stem chars consist mainly of macropores and mesopores at low treatment temperature. this result may be further discussed based on the fact that the proportion of a particular porosity generated in the active carbon depends on the preparation conditions, natural and chemical composition of the precursor (plante, et al. 1988). the conversion of cellulose material to solid carbon yields porous carbon. the porosity depends on the original anatomy of the precursor. the activation treatment given; either by heat treatment or by chemicals results in the formation of surface oxygen complexes. this also results in the increase in the aromaticity of carbon surface due to cross-linking and polymerization of organic molecules, removal of volatiles and opening of blocked pores as the heat treatment temperature is raised (kidena, et al. 2004). during the process of activation, the spaces between the elementary crystallites become cleared of various carbonaceous compounds. this complex and different degrees of physical and chemical changes in the sample led to similarly irregular surface area, porosity and texture of the samples. 4. conclusions porous activated carbon adsorbent has been prepared from date palm trees. high carbon yield was observed. the products exhibited dominance of meso and macropores, especially at low carbonization temperature (400 500oc). generally, phosphoric acid treated precursor yielded higher porosity. a surface area of 1100 m2/g was obtained from the precursor impregnated with h3po4 and carbonized at 500 oc. the study has shown that date palm trees stem could be used as raw material for the preparation of activated carbon with good surface properties. more investigations are necessary to give further insight into the nature of the surface functional groups and other characteristics of the product. acknowledgments the financial assistant of sultan qaboos university is gratefully acknowledged. we would like to thank mr. saif almaamari of the college of science, sultan qaboos university, for the sem analysis. references al-rashdi, b., 2004, preparation of activated carbon from date palm stem,” final year student project, department of petroleum & chemical engineering, sultan qaboos university, oman. attia, a.a., 1997, "some surface properties of activated carbons prepared by gasification with different gases," adsorpt. sci. technol. vol. 15, pp. 707-715. benaddi, h., bandosz, t.j., jagiello, j., schwarz, j.a., rouzaud, j.n., legras, d. and béguin, f., 2005, "surface functionality and porosity of activated carbons obtained from chemical activation of wood," carbon, vol. 38, pp. 669-674. byrne, e. and nagle, d.c., 1997, "carbonization of wood for advanced materials applications," carbon, vol. 35, pp. 259-266. cai, q., huang, z.h., kang, f. and yang, j.b., 2004, "preparation of activated carbon microspheres from phenolic-resin by supercritical water activation," carbon, vol. 42, pp. 775-783. carvalho, a.p., cardoso, b., pires, j. and brotas de carvalho, m., 2003, "preparation of activated carbons from cork waste by chemical activation with koh," carbon, vol. 41, pp. 2873-2876. caturla, f., molina sabio, m. and rodríguez-reinoso, f., 1991, "preparation of activated carbon by chemical activation with zncl2 ," carbon, vol. 29, pp. 9991007. daifullah, a.a.m. and girgis, b.s., 2003, “impact of surface characteristics of activated carbon on adsorption of btex,” physicochem eng. asp. vol. 214, pp. 181-193. gergova, k. and eser, s., 1996, "effects of activation method on the pore structure of activated carbons from apricot stones," carbon, vol. 34, pp. 879-888. girgis, b.s. and el-hendawy, a.a., 2002, "porosity development in activated carbons obtained from date pits under chemical activation with phosphoric acid," micro. meso. mater. vol. 52, pp. 105-117. jagtoyen, m., thwaites, m., stencel, j., mcenaney, b. and derbyshire, f., 1992, "adsorbent carbon synthesis from coals by phosphoric acid activation," carbon, vol. 30, pp. 1089-1096. kidena, k., matsumoto, m., katsuyama, s., murata, s. and nomura, m., 2004, “development of aromatic ring size in bituminous coals during heat treatment in the plastic temperature range,” fuel proc. techn. vol. 85, pp. 827-835. liou, t.h., 2004, "evolution of chemistry and morphology during the carbonization and combustion of rice husk," carbon, vol. 42, pp. 785794. lua, a.c. and yang, t., 2004, "effects of vacuum pyrolysis conditions on the characteristics of activated carbons derived from pistachio-nut shells," j. coll. interface sci. vol. 276, pp. 364-372. luehrs, d.c., rogers, t.n. and godbole, k.a., 1998 "partition coefficients of organic solutes between water and activated carbon," encyclopedia of environmental analysis and remediation, meyers ra, (ed): john wiley & sons. pp 3404-3415. molina-sabio, m., rodríguez-reinoso, f., caturla, f. and sellés, m.j., 1995, "porosity in granular carbons activated with phosphoric acid," carbon, vol. 33, pp. 1105-1113. 54 the journal of engineering research vol. 5, no.1 (2008) 47-54 nagano, s., tamon, h., adzumi, t., nakagawa, k. and suzuki, t., 2000, "activated carbon from municipal waste," carbon, vol. 38, pp. 915-920. pelekani, c. and snoeyink, v.l., 2000, "competitive adsorption between atrazine and methylene blue on activated carbon: the importance of pore size distribution", carbon, vol. 38, pp. 1423-1436. plante, p., roy, c. and chornet, e., 1988, "co2 gasification of wood charcoals derived from vacuum and atmospheric pyrolysis," can. j. chem. eng. vol. 66, pp. 307-312. shih, c.c. and chang, j.r., 2005, "genesis and growth of platinum subnano-particles on activated-carbon characterized by x-ray absorption spectroscopy: effects of preparation conditions," mater. chem. phy, vol. 92, pp. 89-97. teng, h., sheng yeh and li-y hsu, 1998, “preparation of activated carbon from bituminous coal with phosphoric acid activation,” carbon, pp. 1387-1395. wartelle, l., marshall, w., toles, c. and johns, m., 2000, "comparison of nutshell granular activated carbons to commercial adsorbents for the purgeand-trap gas chromatographic analysis of volatile organic compounds," j. chrom. a, vol. 879, pp. 169-175. wigmans, t., 1989, "industrial aspects of production and use of activated carbons," carbon, vol. 27, pp. 1322. yalçιn, n. and sevinç, v., 2000, "studies of the surface area and porosity of activated carbons prepared from rice husks," carbon, vol. 38, pp. 1943-1945. copy (2) of comp50813.qxd the journal of engineering research vol. 3, no. 1 (2006) 55-62 1. introduction spectral efficiency is one of the most important issues in wireless systems. because of limited available frequency spectrum, current cellular radio systems adopt the concept of frequency reuse to utilize the same frequency repeatedly at different locations. however, while frequency reuse provides a more efficient use of the limited available spectrum, it also introduces unavoidable co-channel interference. a large frequency reuse distance can enhance the channel quality by reducing co-channel interference, but will decrease the system spectral efficiency. one challenge for cell engineering is to optimize the tradeoff among channel quality, system efficiency, and the costs of infrastructure and user terminals. the main aim of this paper is to build a generalized mathematical model that can be used to study effect of different parameters on the spectral efficiency of microcellular and macrocellular systems. the used model investigates also, the effect of using cell sectorization on the _________________________________________ *corresponding author’s e-mail: abbosh@itee.uq.edu.au spectral efficiency. cell sectorization is an economic way to increase spectral efficiency. every cell is divided into multiple sectors, and each sector is served by a dedicated antenna (fig. 1). the co-channel interference in such a system will be decreased as a single omni-directional generalized model for spectral efficiency of cellular systems a.m. abbosh* school of itee, university of queensland, brisbane, qld4072, australia received 13 august 2005; accepted 18 october 2005 abstract: a general mathematical model was used to study the spectral efficiency of cellular systems. the model took into consideration effects of first and second tier interference. the random distribution of users across the whole area covered by the cellular system as well as the shadowing and multipath effects were taken into account. the influence of sectorization on value of the spectral efficiency was also studied. the model investigated the effects of the cell radius and cluster size on the spectral efficiency. the influence of different propagation conditions on the spectral efficiency was also considered. results of simulation showed that an improvement of up to 70% can be achieved in the spectral efficiency, when using a six-sector system in comparison with the omni-directional system. also, the spectral efficiency was shown to decay by a rate proportional to the square of the radius. it was also shown that spectral efficiency improved in severe propagation conditions. using a higher value for the cluster size decreased the spectral efficiency by a rate proportional to the square root of the cluster size. keywords: spectral efficiency, sectorization, cellular systems, cluster size, path loss models sector 3 sector 2 sector 1 3 sectors ant figure 1. diagram showing principles of cell sectorization bs ájƒ∏ÿg ᪶f’g ‘ ∞«£dg iaéøµd ωéy êpƒ‰ ¢tƒñy .ω.cg@ »fgƒ°û©dg ™jrƒàdg ,êéãdgh ∫h’g ∞°ü∏d πng~àdg ägòkéj qéñày’g ô¶æh êpƒªædg gòg ònéj .ájƒ∏ÿg ᪶f’g ‘ ∞«£dg iaéøc á°sgq~d ωéy »°véjq êpƒ‰ åëñdg ω~îà°ùj :á°uón …ƒ∏ÿg ,ωé¶æ∏d á≤£æe áfõœ òkéj êpƒªædg ¢sq~j éªc .äélƒª∏d ¥ô£dg o~©j öñ°ùh πng~àdg ägòkéj ∂dòch ,…ƒ∏ÿg ωé¶ædg πñb øe πeéµdéh i飨ÿg á≤£æÿg ≈∏y úe~îà°ùª∏d èféàf ¿g .iaéøµdg √òg ≈∏y äélƒÿg qé°ûàfg ±hô¶d áø∏àıg ä’é◊g òkéjh ,∞«£dg iaéøc ≈∏y áyƒªûg ºémh á«∏ÿg ô£b ∞°üf òkcéj åëîñdg èdé©j éªc .∞«£dg iaéøc ᪫b ≈∏y éªc r~mgƒdg √霒g ωé¶f ™e áfqé≤ª©wé≤e áà°ùh ωé¶f ωg~îà°sg ~æy %70 π°üj qg~≤ã ∞«£dg iaéøc ‘ ø°ù– ≈∏y ∫ƒ°ü◊g ¿éµe’éh ¬fg í°vƒj åëñdg gòg ‘ áe~≤ÿg iéµùg .áyƒªûg ºé◊ »©«hîdg qò÷g ™e ö°séæàj ∫~©ã ∞«£dg iaéøc ‘ ¢übéæj ¤g …odƒj áyƒªûg ºém oéjorg ¬fg í°vƒj áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.qé°ûàf’g ≥jôw ¿g~≤a ᪶fg ,áyƒªûg ºém ,ájƒ∏ÿg ᪶f’g ,™«£≤àdg ,∞«£dg iaéøc : 56 the journal of engineering research vol. 3, no. 1 (2006) 55-62 antenna at the base station is replaced by several directional antennas, each radiating within a specified sector. a given cell will receive interference and transmit with only a fraction of the available co-channel cells. the proposed model in this paper is an extension to the model proposed by (hammerschmidt et al. 1997) and modified by (walke 2002; alouni and goldsmith, 1999). it is based on: (a) the two-slope path loss model that considers the effect of distance on value of received signals and interferers, (b) suzuki model (shankar, 2002) that considers the effect of shadowing and multipath, (c) a uniform distribution of users in the whole area under consideration, (d) interference from first and second tiers so that effect of all considerable interferences are included, (e) adaptive simpson quadrature (gander and gautschi, 2000) to solve the model equation. 2. channel and system models the mechanisms which govern radio propagation in cellular systems are complex and diverse, but they can be attributed to three basic propagation mechanisms; reflection, diffraction and scattering. the total signal received at the receiver is, therefore, the phasor superposition of a number of reflected, diffracted and scattered signal components. as a result of the above three propagation mechanisms, the received signal can be characterized by three nearly independent phenomena including path loss variation with distance, slow lognormal shadowing, and fast multipath fading (andersen et al. 1995). these phenomena are illustrated in fig. 2. 2.1 path loss model the difference (in db) between the transmitted signal level and the signal in the general area of the receiver (area mean power) is referred to as the pathloss (see fig. 2c). path loss is due to the decay of the intensity of a propagating radio-wave. in the following analyses and simulations, the two-slope path-loss model derived from erceg et al. (1992) will be used to obtain the average received power as a function of distance. according to this model, the average path loss is given by: (1) (2) (3) r ec ei ve d po w er ( d8 m ) r ec ei ve d po w er ( d8 m ) r ec ei ve d po w er ( d8 m ) figure 2. variation of received power with distance in a cellular system. part a is a zoom from b which is a zone from c ⎪ ⎪ ⎩ ⎪ ⎪ ⎨ ⎧ >⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ β+⎟ ⎠ ⎞ ⎜ ⎝ ⎛ λ π α ≤⎟ ⎠ ⎞ ⎜ ⎝ ⎛ λ π α = b b b l rr r r log10 r4 log10 rr r4 log10 p )p/p(log10p rt10l = λ = rtb hh4 r where; r is the transmitter -receiver distance; br is the break point distance; th and rh are transmitter and receiver antenna heights, respectively; λ is the wavelength; α is the path loss exponent when the propagation distance r ≤ br while ( α+β ) is the path loss exponent when r > br ; tp is the transmitted power, and rp is the received power. the simulation introduced in this paper uses the following values for the above parameters. the path loss exponent α ranges from 2 to 3, while β ranges from 1 to 2. transmitte r antenna height th is taken as 10 m in microcellular system and as 30 m in macrocellular system, while rh is always assumed 2 m. these values meet the practical limits of cellular systems. distance (m) (c) long term variation 57 the journal of engineering research vol. 3, no. 1 (2006) 55-62 2.2 lognormal shadowing (4) where; m and σ are the mean and standard deviation of the normal distribution respectively, and are measured in db. the m value is normally approximated by the area mean power, p, although strictly speaking, they are only identical when σ = 0 db (linnartz, 1993). if the local mean signal is expressed in watts, then the corresponding distribution is a lognormal distribution of the form (parsons, 2001): (5) note that the terms m and σ in eq. (5) are the mean and standard deviation of the normal distribution presented in eq. (4) and they retain their logarithmic units. the shadowing standard deviation, σ , of the normal distribution tends to be dependent on frequency and the nature of the environment in the vicinity of the terminal. in most cases, σ is relatively independent of transmission path length and antenna height (rowe and williamson, 1986). generally, the variability of the shadowing is greater in more urbanized areas. mogensen, et al. (1991) have reported σ values to be 6.5 to 8.2 db in urban areas. in this paper, σ was assumed to be 7 db. 2.3 rayleigh fading in addition to distance-dependent path loss and environmental shadowing, the signal received at a terminal receiver will undergo significant envelope fading over distances of a few tens of wavelengths (see fig. 2a). this fluctuation is attributable to multipath propagation, whereby the signal propagates not by one, but by many different paths to the receiver. several multipath components of the transmitted signal arriving at the receiver at the same time will each have different time delays, and therefore, different phase shifts. these add either constructively or destructively to produce a random phase shift. this creates rapid changes in the signal strength, an effect known as rayleigh fading (parsons, 2001). it is important to realize that the rayleigh distribution has been found to model the variability of the received signal envelope when the envelope is expressed in volts. if the signal envelope is expressed in watts, the appropriate statistical distribution is no longer the rayleigh distribution, but rather the exponential distribution. hence, if p is the total momentary power received, it will have an exponential probability density function of the form (parsons, 2001): (6) if the momentary received power, p, is expressed in db, then the probability distribution function is given by (shepherd & milnarich, 1973): (7) 2.4 lognormal shadowing and rayleigh fading in the previous subsections, it was shown that over relatively small distances, the signal is well described by rayleigh statistics, with the local mean over a somewhat larger area being lognormally distributed. it is of interest, therefore, to examine the overall distribution of the received signal in these larger areas. it might be reasonable to expect the distribution to be a mixture of rayleigh and lognormal, which has been shown by investigators to be the case by parsons and ibrahim (1983) examined the nakagami-m and weibull distributions, which both contain the rayleigh distribution as a special case, but came to the conclusion that the statistics of the radio signal can be represented by a mixture of rayleigh and lognormal statistics in the form of a rayleigh distribution with a lognormally varying mean. this combined distribution is normally referred to as the suzuki distribution. the probability density function of the momentary received signal power, p, (in watts) in this case is given by (shanker, 2002); (8) 3. carrier-to-interference ratio to simplify the analyses, the following assumptions have been made in the co-channel interference model. first, the system is considered to be interference-limited in which the thermal noise power is negligible relative to the co-channel interference power. thus, the ratio of carrier power-to-noise plus interference power reduces to the carrier-to-interference power ratio (cir). therefore, the ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ σ −− σπ = 2 2 2 )mp( exp 2 1 )p(f the variability associated with large scale environmental obst acles leads to the local mean power fluctuating about a constant area mean power over medium distances (~100 m) (see fig. 1). this phenomenon is known as shadowing and arises due to obstacles in the propagation path such as buildings, hills and foliage. ex periments and analysis indicated that this variability can be approximated by a lognormal distribution (parsons, 2001 ). by lognormal it is meant that the local -mean signal, ,p expressed in logarithmic values has a normal probability di stribution function of the form (parsons, 2001 ); [ ]α− σπ = exp 2p)10(log 10 )p(f e where 2 2 10 2 )m)p(log10( σ − =α ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ − = p p exp p 1 )p(f ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ −= p 10 10 p)10(log exp p10 )10(log ])db[p(f 10/p ee ( ) pdexp 2p)10(log p p exp10 )p(f 0 2 e α− σπ ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ − = ∫ ∞ 58 the journal of engineering research vol. 3, no. 1 (2006) 55-62 desired user cir, γd , can be written as: (9) where pd is the received power level (in w) from the desired mobile at a distance rd from its bs; pi is the total interfering power (in w). pi (ri) and pj (rj) are the received power levels (in w) from the i-th and j-th interfering mobile located within the first and second tiers, respectively and at distances of ri and rj from the desired mobiles' bs. also, n and 2n are numbers of active cochannel cells in the first and second tiers respectively. it is clear from eq. (9) that effects of interferers from the first tier as well as from the second tier are included in the calculations. this is one of the main differences between the model presented in this paper and other models which neglected effect of the second tier interferes. refering to fig. 3b, it is clear that n is equal to 6 but in case of sectorization with s sectors, it is modified to: n = 6/s (10) assume that r and θ define the polar position of a mobile unit relative to its base station, then values of ri and rj can be calculated from: (11) (12) where: d1 is the reuse distance (see fig. 3a) and it is equal to (walke, 2002): (13) where: r is radius of the cell, and n is the cluster size. d2 is the distance to the second tier (see fig. 3b) and it can be proven to be equal to; (14) throughout this paper, it will be assumed that the cochannel interfering signals add up incoherently since this leads to a more realistic assessment of the co-channel interference in cellular systems (prasad & kegel, 1991). also, since the signal powers of both the desired and interfering mobiles experience fluctuations due to multipath fading, shadowing, and the random location of users in their respective cells, γd is also a random variable which depends on the distribution of pd and p1. 4. users location model for analytical convenience, all the mobiles (desired and interfering users) are assumed to be mutually independent and uniformly distributed in their respective cells. thus, the probability density function (pdf) of the mobile’s polar coordinates (r, θ) relative to their bs's are (alouni & goldsmith, 1999): (15) (16) where: ( )∑∑ == + ==γ n2 1j jj n 1i ii dd i d d rp)r(p )r(p p p )sin(rd2rdr 1 22 1i ϕ++= )sin(rd2rdr 2 22 2j ϕ++= (b) figure 3. a 7-cell reuse system which shows: (a) the cell radius r and the distance ri from a random ms to the desired bs; and (b) the first and second tier cochannels at distances d1 and d2 (thick arrow in the diagram), respectively from the desired bs. notice the 6 and 12 cochannels that exist in the first and second tiers, respectively. rn3d1 = rn*3d 2 = ( ) ( )2o o rr rr2 )r(f − − =τ π =θθ 2 1 )(f rrr o ≤≤ and π≤θ≤ 20 second tier first tier 6 7 2 3 7 d1 r 5 4 6 ri r rd (a) 59 the journal of engineering research vol. 3, no. 1 (2006) 55-62 ro corresponds to the closest distance the mobile can be from the bs antenna, and is taken to be equal to 20 m for microcellular systems and 100 m for macrocellular systems. 5. spectral efficiency (17) where: ns is the total number of active serviced channels per cell; ck [b/s] is the maximum data rate of the k-th user; and w [hz] is the total allocated bandwidth per cell. ck is taken to be the shannon capacity of the k-th user in the cell which depends on γk, the received cir of that user, and wk, the bandwidth allocated to that user. se quantifies the tradeoff between the increased system efficiency induced by a small frequency reuse and the decreased capacity of each user resulting from the corresponding increase in co-channel interference for a constant γk , ck is given by shannon's formula: (18) however, γk is not constant in our system since both the interference and signal power of the k-th user will vary with mobile’s locations and propagation conditions. when γk varies with time, ck equals to the average channel capacity of the k-th user (alouni & goldsmith, 1999), and it is given by: (19) (20) (21) where it has been assumed for tdma or fdma systems. (22) 6. results and discussion the models discussed in previous sections were used to study effect of different parameters on the spectral efficiency of microcellular and macrocellular systems. the simulations assumed different types of sectorization. the investigations were made at the two frequencies of 900 mhz and 2 ghz bands that are typically used in modern cellular systems such as the global system for mobile communication (gsm) and the universal mobile telecommunication system (umts). also, the effect of different propagation conditions on spectral efficiency wass considered. in the model presented in this paper, the effect of the second tier interferes is included alongwith the effect of the first tier interferes. to show the error that may occur when neglecting effect of the second tier interferers, the simulation was repeated for one case that is of a microcellular system with six sectors. the results are shown in fig. 4a. it is obvious that the error could be as high as 30%. this justifies the general model presented in this paper. in this paper, the concept of spectral efficiency (se) for fully loaded systems will be considered. in the full loaded system, the cell’s resource (serviced channels) is fully and the number of interferers is constant and equal to n, as seen in eq. (9). se of a cell is defined as the sum of the maximum bit rates/hz/unit area supported by a cell’s bs (alouni & goldsmith, 1999 ). since frequencies are reused at a distance d1, which covers the area of a single cluster of size n. the area covered by one cluster is ( 2/nr33 2 ). se [b/s/hz/m 2] is therefore given by; )2/nr33(w c se 2 sn 1k k∑ = = )1(logwc k2kk γ+= ( ) ( )∫ ∞+ += 0 kkk2kk df1logwc γγγ where )(f kγ is the pdf of the k th user’s cir. it is possible to define the average spectral efficiency se [b/s/hz/m 2] as the sum of the maximum average data rates/hz/unit area for the system, given by eqn.17, with ck replaced by .c k assuming that all users are assigned the same bandwidth then kc becomes the same for all users and will be replaced by .c se can therefore be written as; 2 s rnw33 cn2 se = substituting for c from eqn.19 in to eqn.20 yields; ( ) ( )∫ ∞+ γγ γγ+= 0 22 df1log rn33 2 se ks wnw = figure 4 shows variation s of se with cell radius in microcellular a nd macro-cellular systems at a frequency of 900 mhz. in fig. 4, the results of usi ng different number of sectors we re compared to each other and to the case of using an omni-directional antenna , which is shown in this figure as the case of s=1. it was clear that increasing n umber of sectors per cell cause s an increase in se by up to 70% in comparison with the system that uses an omni-directional antenna (see figs. 4a and 4c). se decreased rapidly with increase in the cell radius. when using a curve fitting, it was possible to show that the decay in se was proportional to the square value of the cell r adius. when the cell radius increased , the cochannel interference decreased. this effect enhanced the se value. on the other hand, incr easing the cell radius increased the area covered by a constant number of channels , which reduce d se . it see ms that the second parameter had the higher influence on se and the overall effect wa s towards decreasing se . 60 the journal of engineering research vol. 3, no. 1 (2006) 55-62 sp ec tr al e ff ic ie nc y (b ps /h z/ km 2 ) 40 35 30 25 20 15 10 5 0 100 200 500 1000 cell radius (m) (a) sp ec tr al e ff ic ie nc y (b ps /h z/ km 2 ) 70 60 50 40 30 20 10 0 100 200 500 1000 cell radius (m) (b) sp ec tr al e ff ic ie nc y (b ps /h z/ km 2 ) 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 1000 2000 5000 cell radius (m) (c) sp ec tr al e ff ic ie nc y (b ps /h z/ km 2 ) 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1000 2000 5000 cell radius (m) (d) figure 4: variation of spectral efficiency with cel l radius for different sectorization values at f=900 mhz: (a) microcellular system with α =2, β =1; (b) microcellular system with α =3, β =2; (c) macrocellular system with α =2, β =1; and (d) macrocellular system with α =3, β =2. results of simulation by a model that neglects effect of the second tier interference is shown in (a) for the case s=6. 61 the journal of engineering research vol. 3, no. 1 (2006) 55-62 7. conclusions in this paper, a general mathematical model was presented to study the effect of cell sectorization on the spectral efficiency of cellular systems. the presented model took into consideration the effect of all valuable interference, i.e. in the first as well as in the second tier regions. the model considered also the random distribution of users over the whole area. the combined effects of the lognormal fading due to shadowing and rayleigh fading due to the multipath propagation were included through the use of suzuki model. the effect of using different values of cluster size on the spectral efficiency was also investigated. results of simulation indicated that an improvement of up to 70% in the spectral efficiency value can be obtained through the use of a six-sector system in comparison with the omni-directional system. it was shown that the spectral efficiency variation was inversely proportional to the square value of the cell radius. it improved as the path loss exponent increased. an increasing cluster size would cause a decay in the spectral efficiency by a rate proportional to the square root of the cluster size. references alouni, m. and goldsmith, a., 1999, "area spectral efficiency of cellular mobile radio systems," ieee trans. on vehicular technology, vol. 48(4), pp. 1047-1066. andersen, j., rappaport, t. and yoshida, s., 1995, "propagation measurements and models for wireless communications channels," ieee communications magazine, vol. 33(1), pp. 42-49. erceg, v., ghassemzadel, s., taylor, m. and schilling, d., 1992, "urban/suburban out-of-sight propagation modeling," ieee communications magazine, pp. 5661. gander, w. and gautschi, g., 2000, "adaptive quadrature; revisited," bit, vol. 40, pp. 84-101. hammerschmidt, j., hasholzner, r. and drewes, c., 1997, "shannon-based capacity estimation of future sp ec tr al e ff ic ie nc y (b ps /h z/ km 2) 35 30 25 20 15 10 5 100 200 500 1000 cell radius (m) (a) sp ec tr al e ff ic ie nc y (b ps /h z/ km 2) 70 60 50 40 30 20 10 0 100 200 500 1000 cell radius (m) (b) figure 5. variation of spectral efficiency with cell radius for different cluster size values; (a) f=900 mhz, α =2, β =1, s=1; and (b) f=900 mhz, α =2, β =1, s=6. figure 5 shows variation s of se with the cell radius for different values of cluster size n. it is shown that when n increases se decreases by a rate which can be proven using curve fitting to be proportional to the square root of n. the lower number of cells per cluster, the greater the number of channels that can be used per cell. however, frequencies reused within a short distance increase the co -channel interference. as shown in fig. 5, the overall effect is that reducing the cluster size enhances the spectral efficiency. hence, from a design viewpoint, the smallest possible value of n is desirable i n order to maximize se . 62 the journal of engineering research vol. 3, no. 1 (2006) 55-62 micro-cellular wireless communication systems," int. symposium on information theory, isit97, germany. linnartz, j., 1993, narrowband land-mobile radio networks, artech house, usa. mogensen, p., eggers, p., jensen, c. and andersen, j., 1991, "urban area radio propagation at 955 and 1845 mhz for small and microcells," proc. ieee global telecommunications conference, usa, pp. 1297-1302. parsons, j. and ibrahim, m., 1983, "signal strength prediction in built-up areas; part 2, signal variability," proc. ieee part f, vol. 130(5), pp. 385391. parsons, j., 2001, the mobile radio propagation channel, 2nd edition, john wiley & sons, usa. prasad, r. and kegel, a., 1991, "improved assessment of interference limits in cellular radio performance," ieee transactions on vehicular technology, vol. 40, pp. 412-419. rowe, g. and williamson, a., 1986, "mobile radio propagation in auckland at 851 mhz," electronics letters, vol. 21(22), pp. 1154-1155. shankar, p., 2002, "introduction to wireless systems," john wiley & sons, usa. shepherd, w. and milnarich, j., 1973, "basic relations between a rayleigh distributed randomly varying voltage and a decibel record of the voltage," proc. ieee, pp. 1765-1766. walke, b., 2002, mobile radio networks; networking, protocols and traffic performance, 2nd edition, john wiley & sons, usa. civl050925.qxd the journal of engineering research vol. 4, no.1 (2007) 82-94 1. introduction occurrence of past earthquakes has shown (arya et al. 1977; monge, 1969; steinbrugge, 1963) that masonry buildings have suffered the maximum damage compared to all other building systems. strengthening measures for masonry buildings have been recommended (arya, 1967; krishna and chandra, 1969; moinfar, 1972; plummer and __________________________________________ *corresponding author’s e-mail:mqamar@lycos.com blume, 1953; yorkdale, 1970) for achieving non-collapse masonry constructions. in-spite of poor seismic performance and mainly due to low cost of masonry structures, they are being used increasingly in all countries, except perhaps in the relatively more developed ones. a number of base isolation schemes have been developed and implemented (beck and skinner, 1974; blakeley et al. 1979; buckle et al. 1990; caspe, 1970; chopra et al. 1973; fintel and khan, 1969; kelly, 1984; megget, 1978; skinner et al. 1975; skinner and mcverry, 1975) in the last forty years in which a superstructure is connected to a seismic response of pure-friction base isolated masonry building with restricted base-sliding m. qamaruddin*1 and s. ahmad2 1 department of civil engineering, jamia millia islamia, new delhi 110025, india 2 department of civil engineering, aligarh muslim university, aligarh 202002, india received 25 september 2005; accepted 19 september 2006 abstract: the earthquake response of pure-friction base isolated masonry building with restricted base sliding is presented in this paper. a clear smoothened surface is created at the plinth level of the building above the damp-proof course, and the superstructure rests at this level to slide freely, except for the frictional resistance and the rigid stopper. as the superstructure is free to slide at the plinth level, there will be a feeling to the occupants that it might slide more than the permissible relative displacement limit provided at the top of the substructure. this may create a fear among the occupants that the building may overturn. in view of this, a mathematical model was developed for masonry building with restricted base sliding system using rigid stopper. feasible position of the stopper has been determined such that the seismic response of the structure is reduced considerably in comparison with that of corresponding fixed base structure. investigation was also made to determine the seismic response of the buildings with varying time period, mass ratio, coefficient of friction and damping coefficient subjected to koyna and five other pseudo earthquakes. the pseudo earthquakes were generated either by increasing or decreasing the ground acceleration and duration of the koyna accelerogram. it turns out from the present study that the pure-friction isolated restricted base sliding system is effective in reducing the seismic force acting on the masonry building with low value of coefficient of friction. keywords: pure-friction, restricted base sliding system, koyna accelerogram, mass ratio, pseudo earthquakes, rigid stopper, sliding force ¥’õf’g ioh~fi á«céµàmg i~yéb …p »bƒhéw ≈æñÿ á«dgõdõdg áhééàà°s’g øj~dg ôªb .ω 1@ 2~ªmcg .¢s h áá°°uuóóÿÿggájôé◊g i~yé≤dg iƒà°ùe ‘ ¢ù∏eg í£°s ~dƒàj å«m .¥’õf’g ioh~fi á«céµàmg i~yéb …p »bƒhéw ≈æñÿ á«dgõdõdg áhééà°sód á°sgqo ¢vôy ºà«°s ábqƒdg √òg ‘: ¥’õf’g ájôm …ƒ∏©dg πµ«¡∏d ¿g éãh .áñ∏°üdg áàñãÿgh á«céµàm’g áehé≤ÿg g~y ée ,iƒà°ùÿg gòg ‘ ¥’õf’g ájôm …ƒ∏¨dg πµ«¡∏d í«àjh,áhƒwôdg ™fée ∫ée~e ¥ƒa ájéæñ∏d ≥∏îj ~b é‡ ,≈æñÿg øe ≈∏y’g öféé∏d ¬h 샪°ùÿg »ñ°ùædg ámgr’g ~m øe ìcg ∑ôëàj ~b áféh ≈æñÿg »∏zé°t ~æy qƒ©°t ∑éæg ¿ƒµ«°ùa ,ájôé◊g i~yé≤dg iƒà°ùe ‘ ¥’õfód ~j~– ” ~bh .áñ∏°u áàñãe ωg~îà°séh ¥’õf’g ioh~fi i~yéb ägp á«bƒhéw ájéæh π«ãªàd »°véjq êpƒ‰ ôjƒ£j ” ~≤a ,gò¡d ô¶fh .ö∏≤æj ~b ájéæñdg ¿céh ú∏zé°ûdg úh éaƒn á«dgõdõdg áñ∏éà°s’g ~j~– é° jcg åëñdg πª°t éªc .áàhéãdg ~ygƒ≤dg äghp êéñÿg ∂∏j ™e áfqé≤ÿéh òãµh πbcg ≈æñª∏d á«dgõdõdg áhééà°s’g ¿ƒµj å«ëh áñ∏°üdg áàñãª∏d »∏ª©dg ™bƒÿg ∫r’õdg ä~dh ~bh .iông áøjõe ∫õd’r á°ùªnh éæjƒc ∫gõdr ¤g é¡yé° ng ~æy …ooîdg §«ñãàdg πeé©eh ∑éµàm’g πeé©e ,á∏àµdg áñ°ùf ,ooîdg øer é¡«a ∞∏àîj ≈àdg äéjéæñ∏d ‘ ∫é©a ôkcg ägp ,¥’õf’g ioh~fi ádõ©æÿg á«céµàm’g ~ygƒ≤dg ωé¶f ¿g ¤g á«dé◊g á°sgq~dg â°ü∏n ~bh .éæjƒc ∫gr’r i~eh »°vq’g π«é©àdg ¿é°ü≤æh hcg ioéjõh éeg áøjõÿg .¢ øîæÿg ∑éµàm’g πeé©e äghp á«bƒhé£dg äéjéæñdg ≈∏y iôkdƒÿg á«dgõdõdg iƒ≤dg ¢ué≤fg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .¥’õf’g iƒb ,áñ∏°u áàñãe ,áøjõe ∫gõdr ,á∏àµdg áñ°ùf ,éæjƒc ∫gõdr ,¥’õf’g ioh~fi ~ygƒ≤dg ωé¶f ,ωéàdg ∑éµàm’g : 83 the journal of engineering research vol. 4, no.1 (2007) 82-94 substructure through flexible structural elements and /or energy dissipation devices. most isolation systems are too advanced, as well too expensive for the application to masonry buildings in developing countries where about 80% of the houses are single or double storey and made of masonry. an alternative is to provide a base isolation system in which the isolation mechanism is purely sliding friction. pure-friction seismic isolation (pfsi) is introduced between the superstructure and the substructure to provide lateral flexibility and the energy dissipation capacity. the main idea for solving the seismic strength problem of masonry buildings has come from the past history of earthquakes. after the dhubri earthquake, the damage study (gee, 1934) showed that those buildings in which the possibility of movement existed between the superstructure and substructure suffered less damage than those buildings in which no such freedom existed. based on such encouraging seismic behavior of small structures, a simple mathematical model was first introduced by qamaruddin (1978) to compute the seismic response of masonry building with pure-friction seismic isolation (pfsi) system. this concept was established through analytical and experimental studies (qamaruddin, 1978) made for masonry buildings. feasibility study of the pfsi concept was made by testing (qamaruddin,1978 and qamaruddin et al. 1984) pilot house models by inserting different sliding materials between the house models and their base. it was observed by steady state testing of these models that there was no amplification of accelerations, whereas such acceleration amplification was observed in similar fixed-base structures. thus, these pilot tests had strengthened the pfsi concept that by introducing a discontinuity at the plinth level of the superstructure, the effective seismic force can be reduced as compared to that of the conventional models. the concept of such system was further strengthened by the damage studies made (li li, 1984) after the xintai (1966), bohai (1969) and tang shan (1976) earthquakes in which it was found that adobe buildings which were free to slide on their foundations (by accident) survived with little or no damage whereas others which were tied on their foundations collapsed. many researchers have carried out experimental investigations to study the dynamic behavior and to compare the performance of the pfsi house models with that of the conventional ones. half-scale models of brick house were tested (qamaruddin, 1978) under shock loading. out of these eight models, there were six conventional ones, and two models with a sliding substructure. tests results showed that the house models with sliding base had a significant reduction in response and exhibited adequate behavior up to very high base accelerations as compared with the similar models with fixed base. similar pfsi concept has been proposed and tested in china (li li, 1984; fu, 1988) with encouraging results. model studies were also carried on aluminum and masonry specimens with pfsi scheme (zongjin et al. 1989). there was reasonably good agreement between the experimental and theoretical results for the sliding structure. several researchers (chandrasekaran, 1970; lin et al. 1986; malushte et al. 1989; mittal, 1971; mostaghel et al. 1983a and 1983b; newmark, 1965; qamaruddin, 1978; qamaruddin et al. 1986a and 1986b; younis et al. 1984; westermo et al. 1983) have investigated behavior of sliding systems subject to harmonic and earthquake-type excitations. tehrani and hasani (1996) presented the seismic behavior of conventional real structures in iran. they have considered the mathematical model for low-rise masonry buildings as a rigid body system for earthquake behavior and response study. they have obtained the well-established results that as the friction coefficient (0.10 to 0.30) of the sliding layer is increased, maximum acceleration of the rigid structure is increased linearly up to a maximum value of base acceleration. in the case of base-isolated structure, with the addition of flexible layer at substructure level the peak base displacement increase significantly during earthquakes and the structure can collide upon adjacent structures like boundary retaining walls, entrance ramps, etc. further, in the case of long buildings, expansion gaps are invariably provided to accommodate the displacements taking place due to temperature variations. for such a long base-isolated buildings, there are likely chances of impact to occur at the expansion gaps when the two buildings vibrate out of phase. such incidence of impact in such buildings has been reported during 1994 northridge earthquake resulting in higher accelerations in the superstructure than the predicted accelerations (nagarajaiah et al. 2001). nagarajaiah et al. (2001) have evaluated seismic performance of the base isolated fire control and command (fcc) building in los angeles during the 1994 northridge earthquake and the effect of impact. it is shown that the seismic performance of the fcc building in the northridge earthquake was satisfactory, except for increased shear and drift due to the impact. this building had no pfsi system. malhotra (1997) has investigated the effects of seismic impacts between the base of an isolated building and the surrounding retaining wall. no pfsi system has been provided in the building studied by the researcher. tsai (1997) investigated the effect of bumping of the base isolated building against its surroundings providing a space for the deformation of the isolation system. the results indicate that the impact wave induced by the bumping can create an extremely high acceleration response in the structure. pfsi system has not been provided in the building studied by tsai. dimova (2000) has presented a study on modeling of collision in sliding systems subjected to seismic excitations. matsagar and jangid (2003) have investigated the seismic response of multistory building supported on various base isolated systems during impact with adjacent structures. in view of the above investigations, it turns out that the seismic response of single storey masonry building with pfsi system and restricted base sliding has not been studied. even if proper plinth width has been provided in the free sliding of the superstructure at plinth level, there will be a feeling to the occupants that such building might slide more than the permissible limits provided. this may cre84 the journal of engineering research vol. 4, no.1 (2007) 82-94 ate a psychological fear regarding over-turning or tilting of the building. further increasing the damping does reduce the displacement, but at the expense of increasing acceleration and the storey drift. keeping this in view, the investigations has been carried out with the following objectives to: 1. develop the mathematical model for a single storey masonry building with restricted base sliding system (provision of rigid stopper). 2. determine the most feasible position of the stopper such as to reduce the seismic response of the building with restricted base sliding in comparison with that of the fixed base. 3. study the acceleration response of masonry building subjected to koyna earthquake and five pseudo earthquakes with different time period, mass ratio, coefficient of friction and damping coefficient. 2. mathematical idealization 2.1 mathematical model it is assumed that a layer of a suitable material with known frictional coefficient is laid between the contact surface of bond beam of superstructure and plinth band of substructure of a single-storey masonry building with restricted base sliding. the superstructure is allowed to slide at certain limit, which is restricted by the rigid stoppers provided around the building. the restricted sliding type building is idealized as a discrete mass model with two degree of freedom for computing the earthquake response (fig. 1). the spring action in the system is assumed to be provided by the shear walls, which resists shear force parallel to the direction of earthquake shock. internal damping is represented by a dashpot that is parallel with spring. the mass of roof slab and one-half of the height of walls is lumped at the roof level, other half height of the walls is lumped at the level of bond beam, the lower mass assumed to rest on a plane with dry friction damping to permit sliding of the system at certain limit. the coefficient of friction (assumed as static) between the sliding surfaces is considered to be constant throughout the motion of the system. materials used for building construction are linearly elastic within the limit of proportionality, thus the idealized spring is linear elastic. its stiffness is computed by considering bending as well as shear deformations in the wall element. sliding displacement at contact surface between the bond beam and the plinth bond can occur without overturning or tilting. the building is assumed to be subjected to only one horizontal component of ground motion at a time. the effect of vertical ground motion is not considered here. the stoppers, which restrict the sliding displacements, are considered as rigid. 2.2 equations of motion phase 1: initially, bottom mass moves with the base so long as sliding force does not overcome the frictional resistance. so the building behaves as a single degree of freedom system and therefore equation of motion is: (1) (2) figure 1. mathematical model for masonry building with restricted base sliding stopper stopper µ gmt ∆ y xb xt mt mbzb∆ zt 0)zz(k)zz(cxm btsbtstt =−+−+ &&&& or, ( )ty)zz()zz(2z bt2btt &&&&&& −=−ω+−ξω+ where, cs= coefficient of the viscous damper; sk = spring constant; mt = top mass; mb = bottom mass; tt z,x &&&& = absolute and relative acceleration of the top mass respectively; y(t) and ( )ty&& = ground displacement and acceleration at time t respectively; zb, zt = lateral relative displacement of masses mb and mt respectively; tb z,z && = relative velocity of masses mb and mt, respectively; ω=natural frequency (=1/t, where t is time period), ξ = damping ratio, xt = zt + y and xb = zb + y. ( )ty&& 85 the journal of engineering research vol. 4, no.1 (2007) 82-94 phase 2: the sliding of the bottom mass begins when the sliding force overcomes the frictional resistance at the plinth level. the force to cause sliding sf, is given by (3) (4) where, g = acceleration due to gravity; total mass (mt) = mb + mt and µ = coefficient of friction. the building now acts as a two degree of freedom system for which the equations of motion can be written in a simplified form as: (5) (6) phase 3: the stopper is assumed as rigid such that the bottom mass rebounds with the same force as it strikes the stopper. it is also assumed that the initial velocity of the mass is zero at the rebound phase. the stopper is positioned in such a way that at some instant t1 when the displacement of base mass zb(t1) reaches ∆, the system strikes the stopper. therefore, the motion of the system will be reversed if the frictional resistance at the plinth level is overcome by the sliding force (s/f ), given below: (7) (8) (9) where, β=1/(2+ θ). when the bottom mass moves backward the equations of motion will remain the same as given in phase 2. phase 4: 2.3 solution of equations of motion modal superposition technique cannot be employed in the solution of equations of motion as (a) the force-deformation behaviour of the sliding systems is non-linear and (b) the system is non-classically damped because of difference in the damping in the isolation system in comparison with the damping in the building superstructure. therefore, the equations of motion are solved numerically for different phases using runga-kutta fourth order method (bennet et al. 1956) for obtaining the complete seismic response time-history. this method was used for numerical integration, since it is self starting and the solutions are stable and accurate to a definite precision. a computer programme has been developed to compute the time-wise earthquake response of masonry building with restricted sliding-base system. the solution methodology and the implementation of the programme are presented in detail elsewhere (hoda, 2002) 3. parametric study the parameters considered in the present study are: time period (t, tp*), mass ratio (θ, mr*), dry coefficient of friction (µ, cof*) and damping ratio (ξ, zeeta*) for estimating forces and displacements of single-storey masonry building with restricted sliding. the response has been computed for koyna (longitudinal component) earthquake of december 11, 1967 and five pseudo earthquakes, which have been generated either by increasing or decreasing the duration and ground acceleration of koyna earthquake as shown in table 1. these pseudo earthquakes cover a variety of different type of earthquakes. the seismic response of interest are: (1) absolute acceleration (2) the maximum relative displacement of the superstructure and (3) the residual relative displacement of the superstructure at top of the substructure at the end of the ground motion. the range of different parameters' value that has been estimated to cover a wide variety of singlestorey masonry buildings are as shown in table 2. (* marked notations are used in figs. 3 to 7 for convenience) it is assumed that a coefficient of friction less than 0.10 in sliding will be difficult to obtain in actual building construction, and for a µ value greater than 0.25, no sliding motion may occur in most real earthquakes and the system may act like a fixed base structure. depending upon the time period, mass ratio plays an important role in the masonry buildings. the mass of the roof slab and one half of the height of the walls are lumped at the roof level and ( ) bbbtsbtsf xmzzk)zz(cs &&&& −−+−= sliding of bottom mass occurs if fs > µ m t g ( )tyf)zz()zz(2z bt2btb &&&&&& −=+−θω−−θξω− ( )ty)zz()zz(2z bt2btt &&&&&& −=−ω+−ωξ+ where, f = )z(sign)1(g b&θ+µ ; bz&& = relative acceleration of bottom mass; )z(sign b& = +1 if )z( b& is positive; )z(sign b& = –1 if )z( b& is negative and b t m m =θ = mass ratio . ( ) btbtsbtsbbf xmzzk)zz(cxms &&&&&& −−+−+=′ the sliding of the bottom mass occurs backward if fs′ > µ m t g and the equations of motion are g iven as follows: 0)z(signgmxm)zz(k)zz(cxm btbtbtsbtsbb =µ++−−−− &&&&&&& or, )t(yf)zz()zz(2z bt 2 btb && &&&& −=β+−βθω−−βθξω− and, ( )ty)zz()zz(2z bt2btt &&&&&& −=−ω+−ξω+ at any time during the motion of the system if fs or fs′ < µ m t g, then the sliding of the bottom mass is stopped but the top mass continues to vibrate. therefore, the system will behave as a single degree of freedom and its equatio n of motion is the same as given in phase 1. 86 the journal of engineering research vol. 4, no.1 (2007) 82-94 the other half of the height of the walls is lumped at the top of the substructure. as the height of the single-storey building increases, the mass ratio increases. the lower mass assumed to rest on a plane with dry friction damping to permit sliding of the system at the base of the structure. the distance of the stopper to be provided at the top of the substructure measured from the superstructure for a combination of t, θ, ξ and µ under different earthquakes is important for the design of the structure. 4. discussion of results response of free sliding (fs), restricted base sliding (rs) and fixed base (fb) of single-storey masonry buildings with various parameter combinations subjected to koyna and other five-pseudo earthquakes (fig. 2) are presented in this investigation. the parametric study showed the effect of several parameters on the seismic response of the structure, however, some of the parameters depicted little changes if any. in view of this, the upper limit and lower limit effects of the parameter are shown in the figures. the seismic response of single-storey building with restricted base sliding has been compared with free base sliding and fixed base. the results of this study are discussed in the following sections. 4.1 effect of viscous damping the representative acceleration response curves (figs. 3 and 4) show that at a particular mass ratio, time period and coefficient of friction in the case of restricted base sliding buildings, an increase in damping ratio (from 0.05 to 0.15) decreases the acceleration response. 4.2 time period effect it is seen from figs. 3 to 7 that for a particular time period and damping ratio, as the coefficient of friction increases, the peak displacement decreases in the case of building with free sliding base. similar trend is generally true in the buildings with restricted base sliding system. the slight departure in this trend observed in few cases is due to the fact that positions of stopper are different in different building parametric cases. it is also observed from the figs. 3 to 7 that for koyna and all five pseudo earthquakes (only representative results for koyna and pseudo earthquake-3 are shown in figs. 3 to 6), there is not much variation in the absolute acceleration for structure with restricted base sliding with increase in time period at a particular coefficient of friction, mass ratio and damping ratio. it is also observed that the acceleration response of the free sliding system is independent of its time period. 4.3 influence of stopper position table 2 shows the position of the stopper for different time periods of the system ranging from 0.06 to 0.10 seconds subjected to koyna and five-pseudo earthquake shocks. influence of stopper's position is seen from figs. 3 to 7. if the stopper is placed close to the building (before its sliding), then its acceleration approaches to the corresponding values as obtained in the case of fixed base. but, if the stopper is placed at a distance approximately equal to the peak displacement of the free sliding system, far away from the building, then its acceleration is very near to that of the corresponding values of the free sliding system. in between, acceleration increases by decreasing the stopper distance from the building. this trend has been observed in all the cases subjected to koyna and other five-pseudo earthquakes (only representative graphs for koyna earthquake and pseudo-earthquakes-3 have been presented in this paper). these are the most feasible ranges of stopper positions for different mass ratios. this study shows that if the stopper is placed anywhere in this range, then the absolute acceleration of the top mass of the system is much less than the corresponding fixed base system. so, the occupants of the buildings will have psychological comfort by providing stopper in the sliding base system. type of earthquake koyna earth quake pseudo earth quake-1 pseudo earth quake-2 pseudo earth quake-3 pseudo earth quake-4 pseudo earth quake-5 time factor (tf) 1.00 1.50 1.50 1.00 1.00 0.50 acceleration factor (af) 1.00 0.50 1.00 1.50 0.50 1.00 table 1. scale factors for pseudo earthquakes time period (t) (tp) sec. mass ratio (θθ) (mr) damping ratio (ξξ) (zeeta) coefficient of friction (µµ) (cof) restricted base sliding range (∆∆),mm (delta) 0.06 2.0,2.5,3.0 0.05,0.10,0.15 0.15 2 to 26 0.07 2.0,2.5,3.0 0.05,0.10,0.15 0.15 2 to 26 0.08 2.0,2.5,3.0 0.05,0.10,0.15 0.15 2 to 37 0.09 2.0,2.5,3.0 0.05,0.10,0.15 0.15 2 to 37 0.10 2.0,2.5,3.0 0.05,0.10,0.15 0.15 2 to 37 table 2. data for computing seismic response 87 the journal of engineering research vol. 4, no.1 (2007) 82-94 figure 2(a). accelerogram (m m /s/s) of koyna earthquake -7000 -2000 3000 8000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 time in seconds a cc el er a tio n figure 2(b). accelerogram (m m /s/s) of pseudo earthquake-1 -8000 -3000 2000 7000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 time in seconds a cc el er a tio n figure 2(c) accelerogram (m m /s/s) of pseudo earthquake-2 -7000 -2000 3000 8000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 time in seconds a cc el er at io n figure 2(a). accelerogram (mm/s/s) of koyna earthquake figure 2(b). accelerogram (mm/s/s) of pseudo earthquake-1 figure 2(c). accelerogram (mm/s/s) of pseudo earthquake-2 88 the journal of engineering research vol. 4, no.1 (2007) 82-94 figure 2(d). accelerogram (m m /s/s) of pseudo earthquake-3 -8000 -3000 2000 7000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 time in seconds a cc el er a tio n figure 2(e). accelerogram (m m /s/s) of pseudo earthquake-4 -8000 -3000 2000 7000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 tme in seconds a cc el er at io n figure 2(f). accelerogram (m m /s/s) of pseudo earthquake-5 -8000 -3000 2000 7000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 time in seconds a cc el er a tio n figure 2(d). accelerogram (mm/s/s) of pseudo earthquake-3 figure 2(e). accelerogram (mm/s/s) of pseudo earthquake-4 figure 2(f). accelerogram (mm/s/s) of pseudo earthquake-5 89 the journal of engineering research vol. 4, no.1 (2007) 82-94 figure 3(a). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.06, cof = 0.15, mr = 3.0) 0.10 0.30 0.50 0.70 0.90 1.10 1.30 8 9 10 11 12 13 14 stopper position (mm) a cc el er at io n zeeta=.05(fb) zeeta=.05(rs) zeeta=..05(fs) zeeta=.10(fb) zeeta=.10(rs) zeeta=.10(fs) zeeta=.15(fb) zeeta=.15(rs) zeeta=.15(fs) figure 3(b). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.10, cof = 0.15, mr = 3.0) 0.10 0.30 0.50 0.70 0.90 1.10 1.30 8 9 10 11 12 13 14 stopper position (mm) a cc el er at io n zeeta=.05(fb) zeeta=.05(rs) zeeta=..05(fs) zeeta=.10(fb) zeeta=.10(rs) zeeta=.10(fs) zeeta=.15(fb) zeeta=.15(rs) zeeta=.15(fs) figure 4(a). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.06, cof = 0.15, mr = 3.0) 0.10 0.40 0.70 1.00 1.30 1.60 1.90 2.20 22 24 26 28 30 32 34 36 stopper position (mm) a cc el er at io n zeeta=.05(fb) zeeta=.05(rs) zeeta=..05(fs) zeeta=.10(fb) zeeta=.10(rs) zeeta=.10(fs) zeeta=.15(fb) zeeta=.15(rs) zeeta=.15(fs) figure 3(a). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.06, cof = 0.15, mr = 3.0) figure 3(b). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.10, cof = 0.15, mr = 3.0) figure 4(a). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.06, cof = 0.15, mr = 3.0) 90 the journal of engineering research vol. 4, no.1 (2007) 82-94 figure 4(b). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.10, cof = 0.15, mr = 3.0) 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50 1.70 1.90 22 24 26 28 30 32 34 36 38 stopper position (mm) a cc el er at io n zeeta=.05(fb) zeeta=.05(rs) zeeta=..05(fs) zeeta=.10(fb) zeeta=.10(rs) zeeta=.10(fs) zeeta=.15(fb) zeeta=.15(rs) zeeta=.15(fs) figure 5(a). variation of absolute acceleration (g) w ith stopper position for koyna earthquake (tp = 0.06, cof = 0.15, zeeta = 0.05) 0.10 0.40 0.70 1.00 1.30 8 9 10 11 12 13 14 stopper position (mm) a cc el er a tio n fixed b a se m r=2.0(fs) m r=2.0(rs) m r=2.5(fs) m r=2.5(rs) m r=3.0(fs) m r=3.0(rs) figure 5(b). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.10, cof = 0.15, zeeta = 0.05) 0.10 0.40 0.70 1.00 1.30 7 8 9 10 11 12 13 14 stopper position (mm) a cc el er at io n fixed base mr=2.0(fs) mr=2.0(rs) mr=2.5(fs) mr=2.5(rs) mr=3.0(fs) mr=3.0(rs) figure 4(b). variation of absolute acceleration (g) with stopper position for pesudo earthquake-3 (tp = 0.10, cof = 0.15, mr = 3.0) figure 5(a). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.06, cof = 0.15, zeeta = 0.5) figure 5(b). variation of absolute acceleration (g) with stopper position for koyna earthquake (tp = 0.10, cof = 0.15, zeeta = 0.05) 91 the journal of engineering research vol. 4, no.1 (2007) 82-94 figure 6(a). variation of absolute acceleration (g) w ith stopper position for pseudo earthquake-3 (tp = 0.06, cof = 0.15, zeeta = 0.05) 0.20 0.50 0.80 1.10 1.40 1.70 2.00 22 25 28 31 34 37 stopper position (mm) a cc el er a tio n fixed b a se m r=2.0(fs) m r=2.0(rs) m r=2.5(fs) m r=2.5(rs) m r=3.0(fs) m r=3.0(rs) figure 6(b). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.10, cof = 0.15, zeeta = 0.05) 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 19 22 25 28 31 34 37 40 stopper position (mm) a cc el er at io n fixed base mr=2.0(fs) mr=2.0(rs) mr=2.5(fs) mr=2.5(rs) mr=3.0(fs) mr=3.0(rs) figure 7(a). variation of absolute acceleration (g) w ith coefficient of friction (tp= 0.06, mr = 3.0, zeeta = 0.05) 0.10 0.20 0.30 0.40 0.50 0.60 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 cof f icient of f riction a cc el er a tio n ko yna earthquake(rs) p seudo earthquake-1(rs) p seudo earthquake-2(rs) p seudo earthquake-3(rs) p seudo earthquake-4(rs) p seudo earthquake-5(rs) figure 6(a). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.06, cof = 0.15, zeeta = 0.5) figure 6(b). variation of absolute acceleration (g) with stopper position for pseudo earthquake-3 (tp = 0.10, cof = 0.15, zeeta = 0.05) figure 7(a). variation of absolute acceleration (g) with coefficient of friction (tp = 0.06, mr = 3.0, zeeta = 0.05) 92 the journal of engineering research vol. 4, no.1 (2007) 82-94 4.4 mass ratio effect effect of mass ratio is seen from figs. 5 and 6 and it is observed that as the mass ratio increases, generally, the acceleration response decreases in all the cases of parametric combinations for a particular time period and damping ratio in the case of free sliding base. the possible reason for decrease in the acceleration response due to increase of mass ratio is that for a system, as the mass ratio increases for a given time period and damping, the values of bottom mass and total mass decreases. this implies that the input dynamic energy is decreased and thus the acceleration response decreases. but, generally, no definite pattern of acceleration response variation has been observed in the case of buildings with restricted base sliding with varying values of the mass ratio. in case of fixed base, acceleration response is independent of mass ratio. 4.5 effect of different earthquake for the koyna and the other five-pseudo earthquakes, it is observed that acceleration of restricted base sliding structure is less than the corresponding fixed base structure for all the parametric combinations. it can be seen from figs. 3 to 7 that for a particular time period, mass ratio and damping ratio, acceleration responses are similar for low coefficient of friction for all given earthquakes. but, with the increase of friction, acceleration response is high for pseudo earthquake-3 due to its maximum ground acceleration in comparison to other earthquakes. similarly, increase in acceleration response is very small for pseudo earthquake-4 due to minimum ground acceleration. among all the five cases of earthquakes under study, pseudo earthquake-4 has least acceleration response, whereas pseudo earthquake-3 has maximum response for each coefficient of friction at time period 0.06 sec. as the time period is increasing pseudo earthquake-2 has least value whereas pseudo earthquake-3 still has maximum response. 5. conclusions the following conclusions have been drawn from the seismic response investigation of single-storey masonry building with restricted base sliding system: 1. the restricted sliding base system is effective in reducing the effective seismic force acting on the building with low value of coefficient of friction. 2. the acceleration response decreases for the parametric combinations of time period, coefficient of friction and damping ratio as the distance of stopper from the restricted sliding base structure increases. 3. there is not much variation in the acceleration of the structure with restricted base sliding with increase in time period at a particular value of the coefficient of friction, mass ratio and damping ratio. 4. in the restricted base sliding system, the acceleration of the superstructure is much less than the corresponding fixed base system for the most feasible ranges of stopper positions for different mass ratios. 5. generally, no definite pattern of acceleration response variation has been observed in the restricted base sliding structure with varying values of the mass ratio, whereas, the acceleration response is independent of mass ratio in the case of fixed base system. 6. the acceleration of the restricted base sliding structure is less than the corresponding fixed base structure but more than the corresponding free sliding system for all the parametric combinations in the case of. koyna and the other five-pseudo earthquakes. 7. the restricted sliding base system will minimize the fear regarding overturning, over sliding or tilting of structure among the residents and provide psychological comfort to the occupants. 8. the effect of vertical component of an earthquake ground motion and the flexibility of the stopper on the seismic response of the system are recommended for future investigation. acknowledgments the authors wish to acknowledge the financial support given by the all india council for technical education (aicte), new delhi, under the grant nos. f.no.: 8020/rid/taptec/38-2001-2002 and f. no.: 151/cd/ef (041)/2003-2004 for the research investigation. figure 7(b). variation of absolute acceleration (g) with coefficient of friction (tp = 0.10, mr = 3.0, zeeta = 0.05) 0.10 0.20 0.30 0.40 0.50 0.60 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 cofficient of friction a cc el er at io n koyna earthquake(rs) pseudo earthquake-1(rs) pseudo earthquake-2(rs) pseudo earthquake-3(rs) pseudo earthquake-4(rs) pseudo earthquake-5(rs) figure 7(b). variation of absolute acceleration (g) with coefficient of friction (tp = 0.10, mr = 3.0, zeeta = 0.05) 93 the journal of engineering research vol. 4, no.1 (2007) 82-94 references arya, a. s., 1967, "design and construction of masonry buildings in seismic areas," bulletin of indian society of earthquake technology,vol. 4 (2). pp. 2537. arya, a. s., chandra, b. and rani, p., 1977, "influence of natural disasters (earthquakes) on educational facilities, part i collection and analysis of available information annotated bibliography," report, department of civil engineering, iit, roorkee (formerly, school of research and training in earthquake engineering, university of roorkee). arya, a. s., chandra, b. and qamaruddin, m., 1978, "a new building system for improved earthquake performance," proceedings of the sixth symposium on earthquake engineering, iit roorkee (formerly,university of roorkee), vol. i, pp. 499504. arya, a. s., chandra, b. and qamaruddin, m., 1981, "a new concept for resistance of masonry buildings in severe earthquake shocks," j. of the institution of engineers (i), vol. 61(c16), pp. 302-308. arya, a.s., 1984, "sliding concept for mitigation of earthquake disaster to masonry buildings," proceedings of the 8th world conference on earthquake engineering, san francisco, vol. 5, pp. 951-958. beck, j. l. and skinner, r. i., 1974, "the seismic response of reinforced concrete bridge pier," j. of earthquake engineering and structural dynamics, vol. 2, pp. 343-358. bennet, a. a., milne, w. e. and bateman, h., !956, "numerical integration of differential equations", dover publications, inc., new york. blakeley, r. w. g. , 1979, "recommendations for the design and construction of base isolated structures," bulletin of the new zealand national society for earthquake engineering, vol. 12 (2). buckle, i.g. and mayes, r. l., 1990, "seismic isolation: history, application and performance a worldview," j. of earthquake spectra, vol. 6(2), pp. 161-201. caspe, m. s., 1970, "earthquake isolation of multistory concrete structures," j. of american concrete institute, vol. 67, pp. 923-933. chandrasekaran, a. r., 1970, "earthquake response of friction mounted masses," bulletin of indian society of earthquake technology, vol. 7 (1), pp. 47-53. chopra, a. k., clough, d. p. and clough, r. w., 1973, "earthquake resistance of building with a soft first story," j. of earthquake engineering and structural dynamics, vol. (1), pp. 347-355. dimova, s. l., 2000, "numerical problems in modeling of collision in sliding systems subjected to seismic excitations," advances in engineering software. vol. 31, pp. 467-471. fintel, m. and khan, f. r., 1969, "shock-absorbing softstory concept for multistory earthquake structures," j. of american concrete institute, vol. 66, pp. 381390. fu, yu-an, 1988, "sliding-vibrating response of elastic structures," bulletin of the new zealand national society for earthquake engineering, vol. 21, pp. 190197. gee, e. r., 1934, "dhubri earthquake of 1930," memories of geological survey of india, lxv, pp. 1-106. hoda, m. n., 2002, "seismic response of single-storey structure with restricted base sliding," m. tech. dissertation, department of civil engineering, z. h. college of engineering and technology, aligarh muslim university, india. kelly, j. m., 1984, "aseismic base isolation: review and bibliography," j. of soil dynamics and earthquake engineering, vol. 5, pp. 202-216. krishna, j. and chandra, b., 1969, "srengthening of brick buildings in seismic zones," proceedings of the 4th world conference on earthquake engineering, chile. li, li., 1984, "base isolation measures for seismic buildings in china," proceedings of the eight world conference on earthquake engineering, san francisco, vol. iv, pp. 791-798. lin, b. c. and tadjbakhsh, i. g., 1986, "effect of vertical motion on friction-driven isolation systems," j. of earthquake engineering and structural dynamics, vol. 14, pp. 609-622. malhotra, p. k., 1997, "dynamics of seismic impacts in base isolated buildings," j. of earthquake engineering & structural dynamics, vol. 26, pp. 797813. matsagar, v. a. and jangid, r. s., 2003, "seismic response of base isolated structures during impact with adjacent structures," j. of engineering structures, vol. 25, pp. 1311-1323. malushte, s. r. and singh, m. p., 1989, "a study of seismic response characteristics of structures with friction damping," j. of earthquake engineering and structural dynamics, vol. 18, pp. 767-785. megget, l. m., 1978, "analysis and design of a baseisolated reinforced concrete frame building," bulletin of the new zealand national society for earthquake engineering, vol. 11 (4). mittal, k. c., 1971, "sliding and overturning of objects during earthquakes," m.e. thesis, iit, roorkee india (formerly, university of roorkee). moinfar, a. a., 1972, "earthquake resistant design of brick masonry buildings", proceedings of european symposium on earthquake engineering, london, england. monge, j. e., 1969, "seismic behavior and design of small buildings in chile," proceedings of the 4th world conference on earthquake engineering, chile. mostaghel, n., hejazi, m. and tanbakuchi, j., 1983a, "response of sliding structures to harmonic motion," j. of earthquake engineering and structural dynamics, vol. 11, pp. 355-366. mostaghel, n. and tanbakuchi, j., 1983b, "response of 94 the journal of engineering research vol. 4, no.1 (2007) 82-94 sliding structures to earthquake support motion", j. of earthquake engineering and structural dynamics, vol. 11, pp. 729-748. nagarajaiah, s. and sun, x., 2001, "base isolated fcc building: impact response in northridge earthquake," j. of structural engineering, american society of civil engineers, vol. 127(9), pp. 10631075. newmark, n. m., 1965, "effect of earthquakes on dams and embankments," geotechnique, the institution of civil engineers, vol. 15, pp. 139-160. plummer, h. c. and blume, j. a., 1953, reinforced brick masonry lateral force design, structural clay products institute, washington. paulson, t. j., abrams, d. p. and meyes, r. l., 1991, "shaking-table study of base isolation for masonry buildings," j. of structural engineering, american society of civil engineers, vol. 117, pp. 3315-3336. qamaruddin, m., 1978, "development of brick building systems for improved earthquake performance," ph.d. thesis, iit roorkee india (formerly university of roorkee). qamaruddin, m., arya, a. s. and chandra, b., 1978, "experimental evaluation of seismic strengthening methods of brick buildings," proceedings of the sixth symposium on earthquake engineering, iit roorkee india (formerly university of roorkee), vol. i, pp. 353-359. qamaruddin, m., chandra, b. and arya, a. s., 1984, "dynamic testing of brick building models," proceedings of the institution of civil engineers, london, vol. 77(2), pp. 353-365. qamaruddin, m., arya, a. s, and chandra, b., 1986a, seismic response of brick buildings with sliding substructure," j. of structural engineering, american society of civil engineers, vol. 122(3), pp. 558-572. qamaruddin, m., rasheeduzzafar, arya, a. s. and chandra, b., 1986b, "seismic response of masonry buildings with sliding substructure," j. of structural engineering, american society of civil engineers, vol. 122(9), pp. 2001-2011. skinner, r. i., beck, j. l. and bycraft, g. n., 1975, "a practical system for isolating structures from earthquake attack," j. of earthquake engineering and structural dynamics, vol. 3, pp. 297-309. skinner, r. i. and mcverry, g. h., 1975, "base isolation for increased earthquake resistance of buildings," bulletin of the new zealand national society for earthquake engineering, vol. 8(2). steibrugge, k. v. and flores, r., 1963, "the chilean earthquake of may 1960 a structural engineering view point," bulletin of the seismological society of america, vol. 53(2). tehrani, f. m. and hasani a., 1996, "behavior of iranian low rise buildings on sliding base to earthquake excitation", proceedings of the 11th world conference on earthquake engineering, mexico city, paper no. 1433. tsai, h. c., 1997, "dynamic analysis of base isolated shear beams bumping against stops," j. of earthquake engineering & structural dynamics, vol. 26, pp. 515-528. westermo, b. and udwadia, f., 1984, "periodic response of a sliding oscillator system to harmonic excitation," j. of earthquake engineering & structural dynamics, vol. 11, pp. 353-366. yorkdale, a. h., 1970, "masonry building systems," state of art report no. 9, technical committee no. 3. int. conference on planning and design of tall buildings, lehigh university. younis, c. j. and tadjbakhsh, i. g., 1984, "response of sliding structures to base excitation," j. of engineering mechanics, american society of civil engineers, vol. 110, pp. 417-432. zongjin, l., rossow, e. c. and shah, s. p., 1989, "sinusoidal forced vibration of sliding masonry system," j. of structural engineering, american society of civil engineers, vol. 115(7), pp. 17411755. the journal of engineering research (tjer), vol. 19, no. 1, (2022) 73-84 *corresponding author’s e-mail: kksalam@lautech.edu.ng doi:10.53540/tjer.vol19iss1pp73-84 bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender a. o. arinkoola, s.o alagbe, k.k. salam*, and b.m. ajagbe department of chemical engineering, p.m.b. 4000, ladoke akintola university of technology (lautech), ogbomoso. oyo state. nigeria abstract: bentonite used during drilling and cementing operations in nigeria are imported into the country despite the availability of this raw material, which will increase the gross domestic products (gdp) of this country if properly processed and utilized instead of imported bentonite. this investigation evaluated the performance of three selected nigerian bentonites sourced from the south-western part of nigeria for oil well-cementing operation. the raw and beneficiated samples from each location were prepared and characterized using x-ray fluorescence (ivt-20 xrf) to determine the elemental composition of each sample pre and post beneficiation. the beneficiated sample with properties close to that of wyoming bentonite was chosen for the oil well cement slurry analysis. the measured parameters used in the analysis of the suitability of local samples used in oil wellcementing operation were compressive strength and rheological properties. box-behnken design (bbd) was used for the study of the optimum quantity of bentonite required for cementing specification and bi-optimization values for both the comprehensive and rheological properties. the results of this investigation showed that all local clay samples in raw form failed the specification of the american petroleum institute (api). after beneficiation, ibeshe bentonite (ib) had the highest sodium content and the ratio of the exchangeable cations, 190% increment in sodium-ion was recorded in ib after beneficiation and the slurry yield of ib (2.7 ft3 per sack) was greater than that of wyoming bentonite 1.9 ft3 per sack) when both are subjected to the same experimental condition. bioptimization models predicted optimum experimental conditions with an accuracy of between 0.35 and, 1.26 % for both comprehensive and rheological parameters keywords: rheology; cement slurry; extender; compressive strength; beneficiation, clay. األسمنت النفطي باستخدام موسع من مصادر محلیة لمالطالتحسین الثنائي للمتانة واألداء اإلنسیابي أ. ارینكوال و س. االجبي و ك. سالم* و ب. اجابي مما یزید من یتم استیراد البنتونیت المستخدم أثناء عملیات الحفر والتدعیم في نیجیریا على الرغم من توفر المادة الخام ، :الملخص الناتج المحلي اإلجمالي لھذا البلد إذا تمت معالجتھ واستخدامھ بشكل صحیح بدالً من البنتونیت المستورد. قیم ھذا التحقیق أداء ثالثة العینات بنتونایت نیجیریة مختارة تم الحصول علیھا من الجزء الجنوبي الغربي من نیجیریا لعملیة تدعیم آبار النفط باالسمنت. تم تحضیر ) لتحدید التركیب األولي لكل عینة قبل وبعد ivt-20 xrfالخام والمستفیدة من كل موقع وتمییزھا باستخدام األشعة السینیة الفلوریة ( أسمنت آبار النفط. كانت مالطلتحلیل wyomingالتخصیب. تم اختیار العینة المستفیدة مع خصائص قریبة من تلك الخاصة ببنتونایت ر المقاسة المستخدمة في تحلیل مدى مالءمة العینات المحلیة المستخدمة في عملیة تدعیم آبار الزیت ھي مقاومة االنضغاط المعایی لدراسة الكمیة المثلى من البنتونایت المطلوب لتدعیم box-behnken design (bbd)والخصائص االنسیابیة. تم استخدام الخصائص الشاملة والریولوجیة. أظھرت نتائج ھذا التحقیق أن جمیع عینات الطین المحلیة في المواصفات وقیم التحسین الثنائي لكل من أعلى محتوى ibeshe bentonite (ib)). بعد التخصیب ، كان لدى apiشكلھا الخام فشلت في مواصفة معھد البترول األمریكي ( المالطبعد التخصیب وإنتاج ib٪ في أیون الصودیوم في 190ة بنسبة من الصودیوم ونسبة الكاتیونات القابلة للتبادل ، وتم تسجیل زیاد كیس) عندما یخضع كالھما لنفس 44قدم مكعب لكل 1.9( wyoming bentoniteكیس) أكبر من 33قدم مكعب لكل ib )2.7لـ ٪ لكل من المعاییر الشاملة 1.26و 0.35الحالة التجریبیة. تنبأت نماذج التحسین الثنائي بالظروف التجریبیة المثلى بدقة تتراوح بین واالنسیابیة مالط. سمنت؛ موسع؛ قوة الضغط؛ إثراء؛األ مالطاالنسیابیة؛ الكلمات المفتاحیة: bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 74 nomenclature a extender, % api american petroleum institute b accelerator, % bbd box behken design bhct bottom hole circulating temperature, °f bhst bottom hole static temperature, °f bib beneficiated ibeshe bentonite bwoc by weight of cement bwow by weight of water be,i, l beneficiated samples from ewuobi, ibeshe and lukosi c antifoam, % cs comprehensive strength, psi d disoersant, % fl fluid loss, ml owcs oil well cement slurry re,i, l raw samples from ewuobi, ibeshe and lukosi pv plastic viscosity, cp uca ultrasonic cement analyser xrf x-ray fluorescence yp yield point, lb/100ft2 1. introduction cementing operation is a key operation in both exploration and exploitation activities due to the support it gives during the isolation of zones. during this operation, the annular space between the casing and the drilled well is filled with cement slurry which is allowed to set and solidify thereby joining the casing to the formation(falode et al., 2013). oil well cement slurry is primarily designed to pass some tests before it can be pumped downhole and these tests vary depending on the type of designed formulation (salam et al., 2013, 2015). depending on the operational design to be implemented, suitable additives are added to cement slurry to enhance its properties in order to achieve the desired performance and some of these additives serve as a retardant, viscosifier, extender, antifoam among others. bentonite is one of the additives used to improve cement slurry performance, it is capable of performing more than one function such as the increase in slurry yield, density, rheology, thickening time and compressive strength (boniface and appah, 2014). optimizing the use of bentonite by locally sourcing it instead of importing will reduce the cost of cement operation and the possibility of replacing cement with locally sourced bentonite without compromising comprehensive strength was demonstrated in the experimental investigation conducted by ahmad et al. (2011) where the quantity of bentonite in cement slurry formulation was varied. the research proved that only 30% of the cement can be replaced with bentonite without compromising the comprehensive strength while the cement formulation also resisted sulphate attack. oriji et al. (2014) tested the suitability of nigerian bentonite as an additive in both fresh and saltwater and the results show that in freshwater nigerian bentonite promising results are obtained. the performance of nigerian bentonite cement slurry produced using freshwater can be improved provided cement slurry designs are tailored towards specific operational requirements while in saltwater (xantham is a good substitute for bentonite in seawater), the properties of cement slurry are not compatible with the standard set by api. cement extenders are routine additives used for reducing slurry density and increasing the yield of cement slurry. bentonite is a common material used as an extender and the one used for cement in nigeria is imported despite the under-utilization of deposits of bentonitic clay in nigeria. the estimates of bentonitic clay deposits discovered in several towns in nigeria was placed above 700 million metric tons (afolabi et al., 2017). for just this reason and due to the vast resources employed in the importation of foreign bentonite, the federal government placed an embargo on the importation of foreign bentonite in 2003 and mandated that locally sourced bentonite be used. as a result, there has been increased research into the possibility of using nigerian local bentonite as a substitute for the imported clays in oil well operations while still maintaining the standards of the oil industry(abdullahi et al., 2011; nweke et al., 2015; oriji et al., 2014). bentonitic clay obtained from deposits around the country has been tested in an attempt to find one that meets api specifications. most of the local clays tested generally have proven to be calcium based and did not meet with api rheological values. however, their properties are improved with beneficiation (boniface and appah, 2014; dewu et al., 2011; mesubi et al., 2008). apart from bentonite, the suitability of other cheaper materials that can be produced locally and compete with replaced commercially sourced additives without compromising slurry produced have been subjected to experimental investigation. performance of sodium silicate extracted from coal fly ash from south africa and imported sodium silicate was compared by kaduku et al. (2015) and the comparison showed that cement slurry produced using locally sourced sodium silicate have a slightly lower density, is easier to pump and have high compressive strength. similar research on the use of local raw materials instead of imported was conducted by (boniface and appah, 2014) in nigeria and algeria (oualit et al., 2018). both investigations targeted locally cement for the formulation of oil well cement slurry. these two investigations agreed that their respective sourced cement can be deplored for cementing operation provided appropriate additives are added. the optimization of various additives are not investigated (boniface and appah, 2014) while the influence of a. o. arinkoola, s.o alagbe, k.k. salam, and b.m. ajagbe 75 various additives on the performance of cement slurry was not investigated (oualit et al., 2018). statistical optimization strategies have been applied for optimal use of variables and influence of each of the variables considered(adeleye et al., 2009; salam et al., 2010, 2015). bentonite is one of the crucial materials needed during cementing operation due to the different characteristics it exhibited (thixotropic, swelling and absorption properties). wyoming bentonite is the most commonly available commercial bentonite for cementing and drilling fluid applications (dewu et al., 2011). the importation of wyoming bentonite for oil and gas well-cementing operation has continued to divert some large amount of foreign exchange while nigeria is having a large reserve of bentonitic clay deposit which can be beneficiated to upgrade the properties of the clay to that of the standard commercial bentonite for oil well cementing and other industrial uses (abdullahi et al., 2011). it is therefore of interest to investigate the suitability of nigeria bentonite as an extender in the design of oil wellcementing slurry and its influence concerning other cement additives on cement slurry performance. 2. materials and methods 2.1 materials acquisition the selected nigerian bentonites that were used for this study were sourced from lukosi (7° 7' 50.952'' n, 3° 22' 2.1'' e) in ogun state, ibeshe (6° 33' 54.504'' n, 3° 28' 56.82'' e) and ewu ebi (6° 32' 51.54'' n, 3° 30' 54.828'' e) in lagos state, all in south-west, nigeria. the class g cement, accelerator, antifoam and mix water that were used were sourced from the cement laboratory of sowsco well services (nig.) ltd, plot 212, trans-amadi industrial layout, port-harcourt, river state, nigeria. 2.2 beneficiation of clay samples from each of the three clay samples, ten sets of 100 g of each sample were set aside for beneficiation using sodium carbonate (na2co3) as shown in table 1. to each of the samples (sample with na2co3), 50 ml of distilled water was added, kneaded and were allowed to stay for 24 h for ion exchange to take place. after allowing the kneaded samples to stay for 24 h each of the samples were oven-dried at 75° for 8 h. the beneficiated samples were re-grounded and sieved again to achieve the 63μm before characterization. the chemical composition of the beneficiated samples was analyzed by using ivt-20 computer-automated x-ray fluorescence (xrf) spectrometer. 2.3 cement slurry formulation and design of experiment class g cement (moderate sulphate resistant type) was the primary cement used while the accelerator, antifoam, and extender are added as additives. all the additives used are in house additives used by sowsco nigeria limited except the extender, which was replaced by bib. cement slurry was prepared based on the formulation tabulated in table 2. table 1. composition of beneficiated samples s/n quantity of each sample of the nigeria bentonite (g) quantity of beneficiating agent (na2co3(aq)) (g) total weight of each sample (g) 01 100 1.0 101.0 02 100 2.0 102.0 03 100 3.0 103.0 04 100 4.0 104.0 05 100 5.0 105.0 06 100 6.0 106.0 07 100 7.0 107.0 08 100 8.0 108.0 09 100 9.0 109.0 10 100 10.0 110.0 table 2. formulation of class g cement slurry. lab. qty unit dry components material per cent pounds absolute volume volume (gallons) specific gravity 482 g #1 cement class g 100 94 0.0382 3.5908 3.14 slurry yield (ft3/sack) 1.9 77 g #2 bentonite imported bentonite 16 15.04 0.0454 0.6828 2.65 slurry density (lb/gal) 13.8 36 g #3 cacl2 cacl2 7.5 7.05 0.0685 0.4829 1.75 total fluid, % 82.35% total fluid, gals 9.3 liquid components 4.2 ml #1 antifoam 0.9 0.821 0.099 0.996 4 ml #2 dispersant 1 0.787 0.0801 1.18 389 ml #3 fresh water 80.7 75.803 9.1 1 total 193.502 14.0356 bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 76 the design of the slurry was carried out according to api standard and desired density, thickening time, fluid loss, and rheological behaviour at the time of pumping, appropriate transition time and enough strength at the time of set were achieved. the concentrations of most solid cement additives were expressed as a percentage by weight of cement (bwoc) while liquid additive concentrations were expressed in gallons per sack of cement as tabulated in table 2. the calculations for the mass of the components was made according to the recommendations of the api practices and also the slurry density and the slurry yield of the design of cement slurry for each of the runs of the experiments was calculated using the eqns. (1) to (5). slurry density � lb gal � = total weight (lb) total volume (gal) (1) slurry yield (ft3) = total volume (gal) 7.48 (gal/ft3) (2) 𝑉𝑉olume (gal) = weight of the component in gallon (lb) ∗ the absolute volume of component (3) total weight (lb) = ∑ weight of all the slurry components (4) total volume(gal) = ∑ volume of all the slurry components (5) where 7.48(𝑔𝑔𝑔𝑔𝑔𝑔 𝑓𝑓𝑓𝑓3⁄ ) is the conversion factor for slurry yield. beneficiating agent = na2co3 (aq) the density of each of the samples of bib was first determined using mud balance. the cement and solid additives were weighed and uniformly dry blended before mixing with fresh water in the stainless-steel container of the constant speed mixer (4000 rpm). the mixer motor was turned on and maintained at 4000 rpm at a uniform rate, for 25 s and then liquid additives were added. thereafter, the mixing speed was increased to 12000 for 35 s. the slurries were prepared and kept at room temperature under atmospheric pressure. the process of slurry preparation was adopted for mixing of (17) different formulations based on the change in values of variables tabulated in table 3 using bbd. the individual and interaction effect of the variables that formed owc were evaluated, r-square was used to determine how close the model developed values are to the measured data used in model development and validation. 2.4 rheological properties measurement a 12-speed rotational viscometer was used according to api specifications (api 10a, 2010) to evaluate the rheological properties of owc slurry using the locally sourced beneficiated clay as an extender for each run of the experiment. the slurry was first conditioned to bottom hole circulating temperature (bhct) of 100o f using an atmospheric consistometer. the plastic viscosity (pv) and yield point (yp) of the cement slurry were computed at both ambient and test temperatures using eqns. (6) and (7), respectively: 𝑃𝑃𝑉𝑉(𝑐𝑐𝑐𝑐) = 1.5 ∗ (𝛾𝛾300 − 𝛾𝛾100) (6) 𝑌𝑌𝑃𝑃 ( 𝐼𝐼𝐼𝐼𝐼𝐼 100𝑓𝑓𝑓𝑓3 ) = 𝑃𝑃𝑉𝑉 − 𝛾𝛾100 (7) where 𝛾𝛾300 and 𝛾𝛾100 are viscometer dial readings at 300 and 100 rpm, respectively. 2.5 compressive strength test determination the method that was adopted in this study was nondestructive. in this method, ultrasonic cement analyzer (uca) (model 4265) was used to analyze the compressive strength of the cement slurry for each run of the experiment at 12 and 24 h. for this research work, the curing condition was at the downhole pressure of 3,500 psi, bhst of 115o f, and bhct of 100o f. 3. results and discussion 3.1 characterization of beneficiated nigerian bentonite the xrf results obtained before and after the beneficiation of the selected nigerian bentonites using 4.0 g of na2co3(aq) were presented in table 4. the effect of each dose of na2co3(aq) on each clay sample was monitored and it was observed that after the beneficiation using 4.0 g of na2co3(aq), ibeshe clay sample was selected for the cement slurry analysis because it has the highest value of the sodium ion content in order to meet api specifications of bentonite. the results of other concentrations (1.0, 2.0, 3.0, 5.0, 6.0, 7.0, 8.0, 9.0, and 10.0 g of sodium carbonate) are not shown because of space constraint. the ratio of si/al of the raw sample was less than the 3.1 recommended by api (api 10b) which justified the need for beneficiation as shown in table 4. table 3. the levels of the variables. independent variable coded variable low (wt %) high (wt %) 1 extender (bib) a 16 31 2 accelerator b 5 10 3 antifoam c 0.5 1.4 4 dispersant d 0.7 1.1 a. o. arinkoola, s.o alagbe, k.k. salam, and b.m. ajagbe 77 the highest sodium content which is responsible for swelling in the cement slurry was 1.74 wt% which corresponded to that of ibeshe clay. this value is in a close range to the value reported for the imported bentonite (2.19 wt %) (james et al., 2008). all the clay samples are known to contain alumina (al2o3) and silica (sio2); and this is also true of feldspars which decompose to form clays by hydrothermal activities (mark, 2010). therefore, after beneficiation, the alumina content and other associated oxides were changed and met the requirements for oil well cementing. ibeshe clay showed the highest promising potential. the ratio of the exchangeable cations, calcium / magnesium for the three selected bentonite [x: (1.30/0.86 = 1.51), y(0.67/0.43 = 1.55), z (3.37/2.12 = 1.58)] is greater than 1.0, which shows the three samples are stable as the minimum ratio of calcium /magnesium required by the standard is 0.5 (dontsova and norton, 2001). 3.2 slurry density and yield slurry density is significant to wellbore hydrostatic pressure. however, deviations from design density may cause changes in all other slurry properties which may cause well problems. the reason why extenders are needed is to maintain optimal cement properties when decreasing the density of the cement slurry to accommodate formation limitations during the cement placement process. the slurry density and yield obtained from the various formulations were presented in table 5 using the selected beneficiated bentonitic clay sample from the ibeshe deposit. it was observed that the density ranged from 12.1 – 14 lb/gal while the slurry yield ranged from 1.9 – 2.9 ft3/sk. the basis of this experiment was the results obtained using the imported wyoming bentonite as presented in table 5 which revealed a slurry density and yield of 13.8 lb/gal and 1.9ft3/sk, respectively. according to api recommendation, the value of density ranged from 11.9 15.8 lb/gal and yield ranges from 1.14 – 2.85 ft3/sk are acceptable for cement slurry design (nelson, 1990). the results obtained showed that the beneficiated ibeshe clay (bic) can be used as an alternative for the imported wyoming bentonite. 3.3 bib and cs of owcs the results obtained for the cs at 12 and 24 h for 17 different runs for owcss oil well cement slurries formulated using bbd were presented in table 5. the values of cs obtained at 12 and 24 h using wyoming bentonite was 650 and 730 psi, respectively (table 5). the slurry formulation for run 13 in this study exhibited higher cs of 670 and 770 psi at 12 and 24 h, respectively when compared with that of imported bentonite. careful observation showed that dispersant, slurry stability and inter fluid chemical reactivity have a tremendous influence on compressive strength. generally, there is no universal agreement on strength requirement for oil well cement except the minimum values; however, values of compressive strength after 24 h are often requested to ascertain cement slurry set under the borehole condition and for the reason that at 24 h the slurry will have a better strength development. the one-factor effect of the four variables bib (a), accelerator (b), antifoam (c) and dispersant (d) with the experimental range set in table 3 was used to the response cs at 12 and 24 h to change in each of the variables was presented in fig. 1. in fig. 1(a), the influence of each of the variables on cs at 12 h was presented. the result showed all the variables individually decreased the value of cs when their values were increased from minimum to maximum as set in table 5. similar behaviour can observed later in fig. 1(b) for cs at 24 h, all the variables decreased the cs with an increase in their values except for d. in the case of d, cs increased with an increase in its value from 0.7 to 1.1. table 4. xrf results of the analysis of beneficiated nigerian bentonites with 4.0 g of na2co3. oxides elements (james et al., 2008) lukosi (wt%) ibeshe (wt%) ewuobi (wt%) rl bl ri bi re be sio2 si 48.8 42.95 55.18 45.5 56.72 40 57.8 al2o3 al 15.54 36.25 20.19 29.85 20.11 27.55 14.84 fe2o3 fe 6.44 0.15 0.07 1.5 1.96 3.5 1.43 mno mn 0.07 0.02 0.02 0.001 0.01 0.1 0.09 mgo mg 3.5 1.4 0.86 0.95 0.43 2.5 2.12 cao ca 5.22 0.9 1.3 0.45 0.67 3.12 3.37 na2o na 2.19 0.4 1.7 0.6 1.74 0.15 0.7 k2o k 0.75 0.65 0.41 0.35 0.32 0.12 0.36 tio2 ti 0.49 0.01 0.01 0.15 0.09 0.25 0.08 p2o5 p 0.13 0.01 0.01 0.03 0.04 0.13 0.01 loi loi 15.73 17.25 20.15 20.15 17.83 22.5 19.5 where rl, ri and re are raw results for lukosi, ibeshe and ewuobi clay samples while bl, bi and be are beneficiated results for lukosi, ibeshe and ewuobi clay samples. bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 78 figure 1. effect of individual factors (a, b, c & d) on cs at 12 h (a) and at 24 h (b), respectively. the interactive effect of two variables at a time while the two other variables are kept at their mid-point was used to study the influence of variables on cs. figs. 2(a) and 2(b) are interactive effects for cs at 12 h while figs. 2(c) and 2(d) are the interactive effects for cs at 24 h. fig. 2(a) shows the effect of a simultaneous increase in a and d while b and c are kept constant at their midpoint (7.5 and 0.95, respectively). the figure showed that at a minimum value of d (0.7), an increase in a from 16 to 31 lead to a decrease in cs from 1362 to 460. 39. at the maximum value of d (1.1), an increase in a led to an increase in cs at 12 h from 428.5 to 607.48. at a minimum value of a (16), cs decreased with an increase in d from 0.7 to 1.1 while at a maximum value of a (31), cs increased with an increase in d from 0.7 to 1.1 respectively. other behaviours of simultaneous change in variables values on cs at both 12 and 24 h are presented in figs. 2(b), 2(c), and 2(d). based on the behaviour of variables both at individual and interactive levels, an empirical model was developed for cs at 12 and 24 hours as presented in eqns. (6) and (7), respectively. 𝐶𝐶𝐶𝐶12 = 714.67 − 180.73𝐴𝐴 − 75𝐵𝐵 − 17.21𝐶𝐶 − 196.58𝐷𝐷 + 270.13𝐴𝐴𝐷𝐷 + 69.45𝐵𝐵𝐷𝐷 (8) 𝐶𝐶𝐶𝐶24 = 637.32 − 39.72𝐴𝐴 − 43.45𝐵𝐵 − 15.40𝐶𝐶 + 110.77𝐷𝐷 + 10.11𝐶𝐶𝐷𝐷 + 40.21𝐴𝐴𝐶𝐶2 (9) the r-square values of the empirical models presented cs at 12 and 24 h are 0. 956 and 0.994, respectively. the analysis of variance (anova) of the two models for cs at 12 and 24 h was presented in table 6. the respective model f-value of 18.08 and 82.37 implied that the selected modified model for cs at 12 and 24 h are significant with p-value of 0.3% and 0.2% respectively. the result of the f-values shows that there is a small likelihood that the fvalue obtained was due to noise in the data used for this study. it was discovered that for cs at 12 h, four out of the six model terms are significant to model development (model terms are significant to model development provided their prob>f value is less than 0.05). they are a, b, d, ad, and bd respectively. the lower the prob>f value, the more influential the model term to the model developed. ad has the most influential effect on the developed model and that was why its coefficient value in eqn. (1) has the most numerical strength out of all the coefficient values in that equation. for cs at 24 h, five out of the six model terms are significant to the model developed and d has the most influential effect on the prediction of cs at 24 h. 3.4 effects of bib and other additives on rheological properties of owcs the laboratory results of replacing imported extenders with bib in determining pv and yield points for different experimental runs were presented in table 5 while the result of the experiment conducted with wyoming bentonite (imported) is presented in table 7 (control experiment). it was observed from the table the pv of all the owcs mixed in the laboratory were below 100 (cp) and with the equivalent slurry yield which can be considered as a desirable result to keep cement slurry pumpable within the pre-determined thickening time according to (api rp 10b-2, 2009). no premature gelation of cement slurry is observed for all the cement samples and the values of the yp calculated show that all the slurries are pumpable (abbas et al., 2014). the individual effect of the change in additives values on pv and yp of the owcs is presented in fig. 3. pv increased with an increase in the values of both b (accelerator) and c (antifoam), respectively, while pv decreased with an increase in a (bib) and d (dispersant), respectively. in the case of yp, an increase in values of additives a and d (dispersant) decreased the yp of owcs in this research while the increase in the values of additives b (accelerator) and c (antifoam) increased the value of yp. while c and d have influential effects on yp, both a and b have a mild influence on yp. d ev iation f rom r ef erence point c s a t 1 2 h -1.00 -0.50 0.00 0.50 1.00 440.00 557.81 675.63 793.44 911.25 a a b b c c d d d ev iation f rom r ef erence point c s a t 2 4 h -1.00 -0.50 0.00 0.50 1.00 500 562.5 625 687.5 750 a a b b c c d d (a) (b) a. o. arinkoola, s.o alagbe, k.k. salam, and b.m. ajagbe 79 figure 2. interaction effect of cs at 12 (a) and 24 h (b), respectively figure 3. one factor plot of variables against pv: (a), and yp (b), respectively. 428.70 662.05 895.40 1128.76 1362.11 c s a t 1 2 h 16.00 19.75 23.50 27.25 31.00 0.70 0.80 0.90 1.00 1.10 a d 512.54 648.33 784.12 919.91 1055.70 c s a t 1 2 h 5.00 6.25 7.50 8.75 10.00 0.70 0.80 0.90 1.00 1.10 b d 458.48 541.06 623.64 706.22 788.80 c s a t 2 4 h 16.00 19.75 23.50 27.25 31.00 0.50 0.72 0.95 1.17 1.40 a c 458.48 541.06 623.64 706.22 788.80 c s a t 2 4 h 0.50 0.72 0.95 1.17 1.40 0.70 0.80 0.90 1.00 1.10 c d: disp d ev iation f rom r ef erence point p v -1.00 -0.50 0.00 0.50 1.00 13.50 17.43 21.36 25.29 29.22 a a b b c c d d d ev iation f rom r ef erence point y p -1.000 -0.500 0.000 0.500 1.000 13.51 22.46 31.42 40.37 49.33 a ab b c c d d (a) (b) (c) (d) (a) (b) bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 80 table 5. cement slurry properties based on an experimental design using bib as an extender. a b c d slurry density slurry yield cs at 12 h cs at 24 h plastic viscosity yield point st order run (%) (%) (%) (%) (lb/gal) (ft3/sk) (psi) (psi) (cp) (lbs/100ft2 ) 9 1 31 5 0.9 0.9 12.5 2.7 600 650 18 29 8 2 16 7.5 0.5 1 13.8 1.9 670 700 15 37 2 3 23.5 7.5 0.9 1 13.6 2 580 600 15 23 7 4 16 5 0.9 0.9 13.8 1.9 620 660 21 33 3 5 31 7.5 1.4 0.7 12.7 2.6 440 500 22.5 27.7 13 6 23.5 7.5 0.9 1.1 14 1.9 520 580 16.5 20.5 15 7 23.5 7.5 0.9 1 14 1.9 550 600 22.5 29.5 10 8 23.5 5 1.4 1 13.8 2 650 720 24 22 4 9 23.5 10 1.4 1 14 1.9 630 700 13.5 25.5 11 10 16 7.5 1.4 1 13.6 1.9 600 680 22.5 22.5 14 11 16 10 0.9 1 13.8 1.9 660 700 21 32 6 12 23.5 5 0.5 1 13.6 2 670 750 16.5 24.5 12 13 23.5 7.5 0.9 1 13.8 2 640 770 19.5 22.5 5 14 23.5 7.5 0.9 1 13.8 2 640 770 19.5 22.5 17 15 31 10 0.9 0.9 12.7 2.6 450 560 16.5 29.5 16 16 31 7.5 0.5 0.8 12.1 3 530 600 16.5 21.5 1 17 23.5 10 0.5 1 13.7 2.1 600 650 22.5 26.5 table 6. anova table of cs at 12 and 24 h. cs at 12 h cs at 24 h terms f-value prob > f terms f-value prob > f model 18.08 0.003 model 82.37 0.002 a 32.4 0.0023 a 30.41 0.0117 b 17.83 0.0083 b 92.08 0.0024 c 1.97 0.219 c 11.04 0.0449 d 13.19 0.015 d 67.93 0.0037 ad 18.39 0.0078 cd 1.8 0.2722 bd 2.12 0.2054 ac2 17.55 0.0248 table 7. cement slurry properties using wyoming bentonite as an extender. thickening time (mins) rheology (rpm) pv (cp) yp (lb/100ft2) compressive strength (psi) fluid loss (ml/30 mins) 50 bc 70 bc 100bc 600 300 200 100 6 3 12 h 24 h 290 325 345 50 45 40 35 10 10 15 30 650 730 20 a. o. arinkoola, s.o alagbe, k.k. salam, and b.m. ajagbe 81 table 8. anova table of pv and yp. pv yp f-value prob > f f-value prob > f model 275.51 0.0003 model 16.26 0.0008 a 452.26 0.0002 a 1.5 0.2608 b 73.13 0.0034 b 1.84 0.2173 c 413.02 0.0003 c 11.97 0.0106 d 299.93 0.0004 d 46.49 0.0002 ab 230.55 0.0006 ac 19.51 0.0031 ac 464.19 0.0002 ad 6.63 0.0368 ad 246.5 0.0006 cd 12.87 0.0089 bc 1383.54 < 0.0001 bd 123.82 0.0016 cd 434.18 0.0002 the interaction plots of the combinations of the four additives and their corresponding influence on both the pv and yp were presented in figs. 4(a), 4(b), and 4(c) describe the interactions of additives and their response on the pv calculated while fig. 4 (d), 4e, and 4(f) described the interactions of additives on yp. the simultaneous effect of change in values of a and b on pv was presented in fig. 4(a). in this plot, pv values were reduced with an increase in a from 16 to 31 at both minimum and maximum values of b (5 and 10). however, the values of pv increased from 22.41 to 33.28 when b was varied from 5 to 10 for a minimum value of a (16). for interaction between a variety of values of c and d shown in fig. 4(b), pv significantly increased with an increase in d from 15 to 43.45 at a minimum value of c (0.7). at maximum value of c (1.1), pv values decreased from 21.17 to 10.65 when d was increased from 0.5 to 1.4. the response of pv to an interaction between b and c was presented in fig.4 (c). figures 4(d), 4(e), and 4(f) describe the interactions of additives and corresponding response on the yp calculated from the experimental data obtained from the laboratory. the simultaneous effect of change in values of a and c on yp was presented in fig. 4(d). in this 3-d plot, the values of yp values decreased with an increase in a from 16 to 31 at both minimum and maximum values of c (0.7 and 1.1). however, at a maximum value of a (31), yp increased with an increase in c from 17.38 to 42.82. in the case of simultaneous variation of a and d shown in fig. 4(e), yp increased from 38.33 to 52.62 for an increase in a from 16 to 31 at a minimum value of d (0.7) while at a maximum value of d (1.1), yp decreased with increase in a. for interaction between variation in values of c and d shown in fig. 4(f), a minimum value of d favour increase in yp while yp decreased with an increase in c at maximum d. based on the behaviour of variables both at the individual and interactive level, empirical models were developed for pv and yp as presented in eqns. (10) and (11), respectively. 𝑃𝑃𝑉𝑉 = 22.57 − 5.28𝐴𝐴 + 2.11𝐵𝐵 + 4.48𝐶𝐶 − 6.66𝐷𝐷 − 3.32𝐴𝐴𝐵𝐵 − 4.14𝐴𝐴𝐶𝐶 + 11.53𝐴𝐴𝐷𝐷 − 4.13𝐵𝐵𝐶𝐶 − 6.47𝐵𝐵𝐷𝐷 − 9.74𝐶𝐶𝐷𝐷 (10) 𝑌𝑌𝑃𝑃 = 31.42 − 1.32𝐴𝐴 + 0.88𝐵𝐵 + 6.30𝐶𝐶 − 14.06𝐷𝐷 + 6.42𝐴𝐴𝐶𝐶 − 8.47𝐴𝐴𝐷𝐷 − 14.29𝐶𝐶𝐷𝐷 (11) the analysis of the empirical models developed for both pv and yp shows that their r-square values are 0.99 and 0.94, respectively. the adequate precision of both pv and yp are 53.37 and 15.48, values that show that the values are adequate for the developed models since the values are greater than four (4). the anova of pv and yp was tabulated in table 8. the respective model f-values of 275.51 and 16.26 implied that the selected modified models for pv and yp are significant to the developed models with 0.3% and 0.8% likelihood of the fvalues reported was due to noise. the model terms. a, b, c, d, ab, ac, ad, bc, bd and cd are significant model terms for the empirical model developed for the prediction of pv while terms c, d, ac, ad, and cd (five out of seven model terms) are significant model terms used for the empirical model developed for the prediction of yp, respectively. 3.5 bi-optimization studies a numerical optimization technique based on desirability function was carried out to determine the workable optimum conditions for the design of oil well cement slurry using bib as an extender in owcs formulation. in order to provide an ideal case for oil well cementing, the additives were set in range based upon the requirements of the design. the values for the lower and upper limits of the additives were stated in table 3. the desirability value was 1. the simulated (predicted), experimental and control experimental results are tabulated in the simulated result was achieved using the specified values of the variables: a (30.17), b (5.4), c (1.17) and d (0.9). the experimental and the simulated values for the properties tested in this study are close which suggested that the models developed are predictive and can determine to a high degree of accuracy the rheological and mechanical properties of class g owc. bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 82 figure 4. 3d surface plots of pv (a, b, and c), and yp (d, e, and f), respectively. 16.07 20.37 24.68 28.98 33.28 p v 16.00 19.75 23.50 27.25 31.00 5.00 6.25 7.50 8.75 10.00 a b 10.65 18.85 27.05 35.25 43.45 p v 0.50 0.72 0.95 1.17 1.40 0.70 0.80 0.90 1.00 1.10 c d 11.85 16.15 20.45 24.76 29.06 p v 5.00 6.25 7.50 8.75 10.00 0.50 0.72 0.95 1.17 1.40 b c 17.38 23.73 30.09 36.44 42.80 y p 16.00 19.75 23.50 27.25 31.00 0.50 0.72 0.95 1.17 1.40 a c 7.58 18.84 30.10 41.36 52.62 y p 16.00 19.75 23.50 27.25 31.00 0.70 0.80 0.90 1.00 1.10 a d 9.37 23.54 37.72 51.89 66.06 y p 0.50 0.72 0.95 1.17 1.40 0.70 0.80 0.90 1.00 1.10 c d (d) (c) (a) (b) (f) (e) a. o. arinkoola, s.o alagbe, k.k. salam, and b.m. ajagbe 83 table 9. validation of developed models. cs (psi) pv (cp) yp (lb/100ft2) at 12 h at 24 h simulation value 592 640 16 29 measured value 601±1.44 637±1.26 15±1.01 30±0.35 control (100% cement) 650 730 15 30 4. conclusion for the utilization of locally sourced bentonite as an extender during oil well-cementing operation, the following deductions were drawn: • the sodium ion concentration of bib increased from 0.6 to 1.74 wt% when beneficiated with 4 g of sodium carbonate and it was the highest for the three clay samples beneficiated. • all the seventeen (17) formulated cement slurries designed box-behnken design proved to have required rheological properties according to api standard using bib as an extender. • the thickening time and fluid loss values obtained for all experimental runs are within the acceptable range according to the api standard. • all the values obtained from laboratory analysis for compressive strength after 24 h periods can structurally hold casings at bhst of 115f and bhct 100 f. • the slurry yield of 2.7 ft3 per sack was recorded using bib and was greater than that of wyoming imported bentonite (1.9 ft3 per sack) when both formulations are subjected to the same experimental condition (run 1). • the model developed from the bi-optimization studies predicted both the comprehensive and rheological parameters within 1.44%, 1.26%, 1.01% and 0.35% for cs at 12 h, cs at 24 h, pv and yp, respectively. • all the submissions are laboratory-scale deductions; therefore, it is recommended that this research be subjected to practical placement modelling for confirmation of the results presented. conflict of interest the authors declare that there is no conflict of the interaction of any form on this manuscript. funding no funding was received for this research. acknowledgment the author wishes to acknowledge the management of sowsco well services (nig.) ltd, plot 212, transamadi industrial layout, port-harcourt, river state, nigeria, for providing class g cement, accelerator, antifoam and, mix water and for allowing alagbe b to use their laboratory for data gathering. references abdullahi, a. s., ibrahim, a. a., muhammad, m. a., kwaya, m. y., and mustapha, s. (2011). comparative evaluation of rheological properties of standard commercial bentonite and a locally beneficiated bentonitic clay from a marine deposit in upper benue basin, nigeria. british journal of applied science and technology, 1(4), 211–221. adeleye, j. o., salam, k. k., and adetunde, i. a. (2009). analysis of rheological properties of treated nigerian clay using factorial design. european journal of scientific research, 37(3). afolabi, r. o., orodu, o. d., and efeovbokhan, v. e. (2017). applied clay science properties and application of nigerian bentonite clay deposits for drilling mud formulation : recent advances and future prospects. applied clay science, 143, 39–49. https://doi.org/10.1016/j.clay.2017.03.009 ahmad, s., barbhuiya, s. a., and elahi, a. (2011). effect of pakistani bentonite on properties of mortar and concrete. february 2019. https://doi.org/10.1180/claymin.2011.046.1.85 boniface, o. a., and appah, d. (2014). analysis of nigerian local cement for slurry design in oil and gas well cementation. 5, 176–181. dewu, b. b. m., arabi, s. a., oladipo, m. o. a., funtua, i. i., mohammed-dabo, i. . a., and muhammad, m. a. (2011). improvement of rheological properties of bentonitic clays using sodium carbonate and a synthetic viscosifier. international archive of applied sciences and technology, 2, 43–52. dontsova, k., and norton, l. d. (2001). effects of exchangeable ca : mg ratio on soil clay flocculation, infiltration and erosion. sustaining the global farm. selected papers from the10th international soil conservation organization meeting at purdue university, pp 571–578. falode, o. a., salam, k. k., arinkoola, a. o., and ajagbe, b. m. (2013). prediction of compressive strength of oil field class g cement slurry using factorial design. journal of petroleum exploration and production technology, 3(4). https://doi.org bi-optimization of durability and rheological performance of oil well cement slurry using locally sourced extender 84 /10.1007/s13202-013-0071-0 james, o. o., adediran, m. m., adekola, f. a., odebunmi, e. o., and adekeye, j. i. d. (2008). beneficiation and characterisation of a bentonite from north-eastern nigeria. journal of the north carolina academy of science, 124(4), 154–158. kaduku, t., daramola, m. o., obazu, f. o., and iyuke, s. e. (2015). synthesis of sodium silicate from south african coal fly ash and its use as an extender in oil well cement applications. june 1175–1182. mark, u. (2010). characterization of ibere and oboro clay deposits in abia state, nigeria for refractory applications. international journal of natural and applied sciences, 6(3), 296–305. mesubi, m. a., adekola, f. a., odebunmi, e. o., adekeye, j. i. d., state, o., and science, m. (2008). beneficiation and characterisation of a bentonite from north-eastern nigeria. journal of the north carolina academy of science, 124(4), 154–158. nweke, o. m., igwe, e. o., and nnabo, p. n. (2015). comparative evaluation of clays from abakaliki formation with commercial bentonite clays for use as drilling mud. african journal of environmental science and technology, 9 (june), 508– 518. https://doi.org/10.5897/ajest2015.1904 oriji, a. b., lawal, f. s., bala, z., and engineering, g. (2014). evaluating the efficiency of nigerian local bentonite as an extender in oil well cementation. 28, 78–88. oualit, m., irekti, a., and hami, b. (2018). evaluation of the performance of local cement for oil well-cementing operations in algeria. 5, 5–13. salam, k. k., adeleye, o. j., and arinkoola, a. o. (2010). evaluation of rheological properties of beneficiated locally sourced mud using factorial design. international journal of oil, gas and coal technology, 3(2). https://doi.org/10.1504/ijogct.2010.033561 salam, k. k., arinkoola, a. o., ajagbe, b. m., and sanni, o. (2015). modeling of rheological properties of class g cement slurry. petroleum and coal, 57(5). salam, k. k., arinkoola, a. o., ajagbe, b., and sanni, o. (2013). evaluation of thickening time of oil field class g cement slurry at high temperature and pressure using experimental design. 2, 361–367. 1. introduction 2. materials and methods 2.1 materials acquisition 2.2 beneficiation of clay samples 2.3 cement slurry formulation and design of experiment 2.4 rheological properties measurement 2.5 compressive strength test determination 3. results and discussion 3.1 characterization of beneficiated nigerian bentonite 3.2 slurry density and yield 3.3 bib and cs of owcs 3.4 effects of bib and other additives on rheological properties of owcs 3.5 bi-optimization studies 4. conclusion conflict of interest funding acknowledgment references c:\users\raoh\desktop\paper_5.xps tjer 2012, vol. 9, no. 1, 37-45 1.introduction optimization of parameters in machining is a non linear model with constraints, so it is difficult to conducted this optimization using conventional approaches. as an alterna-tive, non-conventional approaches ________________________________________ *correponding author’s e-mail: abhanglb@yahoo.co.in have become useful approaches to solve machining parameter optimization problems. the lingo solver program is global optimization software; it gives minimum values of surface roughness and their respective optimal conditions. in many real machining applicaoptimal machining parameters for achieving the desired surface roughness in turning of steel lb abhang* and m hameedullah department of mechanical engineering, aligarh muslim university, aligarh, india received 26 july 2010; accepted 1 january 2011 abstract: due to the widespread use of highly automated machine tools in the metal cutting industry, manufacturing requires highly reliable models and methods for the prediction of output performance in the machining process. the prediction of optimal manufacturing conditions for good surface finish and dimensional accuracy plays a very important role in process planning. in the steel turning process the tool geometry and cutting conditions determine the time and cost of production which ultimately affect the quality of the final product. in the present work, experimental investigations have been conducted to determine the effect of the tool geometry (effective tool nose radius) and metal cutting conditions (cutting speed, feed rate and depth of cut) on surface finish during the turning of en-31 steel. first and second order mathematical models are developed in terms of machining parameters by using the response surface methodology on the basis of the experimental results. the surface roughness prediction model has been optimized to obtain the surface roughness values by using lingo solver programs. lingo is a mathematical modeling language which is used in linear and nonlinear optimization to formulate large problems concisely, solve them, and analyze the solution in engineering sciences, operation research etc. the lingo solver program is global optimization software. it gives minimum values of surface roughness and their respective optimal conditions. keywords: metal cutting, response surface methodology, surface roughness, optimization 38 lb abhang and m hameedullah tions, three conflicting objectives are often considered. these are the maximum production rate, minimum operational cost, and quality of machining. therefore, three cutting parameters and tool geometries need to be determined in a turning operation: cutting velocity feed rate, depth of cut, and tool nose radius. the purpose of the turning operation is to produce low surface roughness. proper selection of cutting parameters and tools can produce longer tool life and lower surface roughness. hence, the design of experiments by factorial input, with response surface methodology (r.s.m). applied to cutting parameters, was adopted to study the surface roughness (abhang, hameedullah 2010). hard turning eliminates the series of operation required to produce the component and thereby reduces the metal cutting cycle time. this results in improvement of productivity (konig et al. 1984). the proper selection of the cutting tool inserts and cutting conditions plays an important role in process planning for steel turning. this task is traditionally completed by the process planner by taking the data from the hand books and cutting tool catalogues. since different cutting tool inserts are available with different geometries, coatings and materials, it becomes difficult for the process planner to decide the required inserts on the basis of available data (wang et al. 2002). it is necessary that reliable mathematical models be formulated so that optimum selection of tool geometry and cutting parameters can be achieved. to date very few models have been developed for the case of en-31 alloy steel turning incorporating tool geometry (tool nose radius) and the cutting conditions. researchers have studied the effect of the metal cutting parameters on surface roughness, but the combined effect of tool nose radius and cutting parameters (cutting speed, feed rate, and depth of cut) has not been studied properly. therefore, in this paper, an effort has been made to develop the mathematical models incorporating tool geometry (tool nose radius) and cutting conditions by performing the experiments on the en-31 alloy steel with a tungsten carbide tool and then optimizing the turning process by lingo-solver programmer. suresh et al. (2002) have developed a surface roughness prediction model for turning mild steel using a response surface methodology. as surface roughness prediction model has also been optimized by using genetic algorithms. feng, wang (2002) have developed empirical models for surface roughness prediction in finish turning. they have included work piece hardness, feed rate, and tool point angle, depth of cut, cutting speed and cutting time. ozel, karpat (2005) studied the predictive modeling of surface roughness and tool wear in hard turning using regression and neural networks. the cutting tool used was made of cubic boron nitride. in this study, effects of cutting edge radius, work-piece hardness, cubic-baron-nitride (cbn) context, cutting length, cutting speed and feed rate on surface roughness and tool wear were experimentally investigated. effective rake angle was kept constant in this study. dilbag (2007) studied the optimization of tool geometry and cutting parameters for hard turning. the cutting tool used was made of mixed ceramic inserts. in this study, effects of cutting speed, feed rate, and negative rake angle and tool nose radius on surface roughness were experimentally investigated. however, effective depth of cut has been kept constant through out the study. yang, tarng (1998) employed the taguchi method to investigate the cutting characteristics of s45c steel bars using tungsten carbide tools. the optimal cutting parameters namely, cutting speed, feed rate and depth of cut for turning operations with regard to performance indexes such as tool life and surface roughness are considered. kopac et al. (2002) investigated the optimal machining parameters for achieving good surface roughness in fine turning of cold pre-formed steel c15 e4 (iso). manna, bhattacharyya (2004) took the significant cutting parameters into consideration and used multiple linear regression mathematical models relating the surface roughness height ra and rt to the cutting parameters for turning al/sic-mmc. aslan et al. (2007) used an orthogonal array and analysis of variance to optimize cutting parameters in turning hardened aisi4140 steel with a al2o3 ceramic tool coated with ticn. the flank wear and surface roughness had been selected as investigated quality objectives and the authors tried to determine optimal values of cutting parameters, such as cutting speed, feed rate and depth of cut. nalbant et al. (2007) used the taguchi method to find good surface roughness in turning of aisi 1030 steel bars using tin coated tools in terms of three cutting parameters, namely insert radius, feed rate and depth of cut. it is evident from the above-mentioned literature that all the models studied only cutting parameters (i.e., cutting conditions) for the turning process. additionally, the optimization of the tool geometry (considering effective tool nose radius) and the metal cutting conditions for the en-31 alloy steel turninghave never been reported. therefore, an effort has been made in this paper to obtain the optimum values of the tool geometry and the cutting conditions for minimizing the surface roughness by lingo-solver software. 2. methodology in this work, experimental results were used for modeling using response surface roughness methodology (rsm). rsm is a collection of mathematical and statistical techniques that are useful for the modeling 39 optimal machining parameters for achieving the desired surface roughness in turning of steel and analysis of problems in which response of interest is influenced by several variables and the objective is to optimize the response. rsm is practical, economical and relatively easy for use. additionally it has been widely researched in the modeling metal cutting process (birdie 1993, hasegawa et al. 1976; montgomery 1991). rsm was also successfully used for application in tool life testing (abhang et al. 2010; mehrban et al. 2008) surface analysis and tool wear rate in metal cutting. the experimental data was utilized to build a mathematical model (first order, second order, and an exponential model) by the regression method. this mathematical model was taken as objective function and was optimized using a lingosolver programmer to obtain the machining conditions for the required surface finish the following linear relationship is commonly used for representing the mathematical models in metal cutting: y = ø (v, f, d, r) + (1) where, v, f, d, and r are the speed, feed, depth of cut and tool nose radius respectively of the metal cutting processes, and is the error, which is normally distributed with mean = 0 according to observed response y and ø is the response function. the relationship between surface roughness and other independent variables is modeled as shown below: ra = c va fb dc rd (2) where c is constant, b, c, and d are the exponents. equation (2) can be represented in linear mathematical from as shown: ln ra = ln (c) + alnv + b ln (f) + c lnd + dlnr (3) the constants and exponents a, b, c and d can be obtained by the method of least squares. the firstorder linear model developed from the equation, can be represented as follows. y1 = y= boxo + b1x1 + b2x2 + b3x3 + b4x4 (4) where, y1 is the estimated response based on firstorder equation on the logarithmic scale, y is the measured surface roughness xo = 1(dummy variable), x1, x2, x3 and x4 are logarithmic transformations of cutting speed, feed rate, depth of cut, and tool nose radius, respectively. the experimental error is and b-values are the estimates of corresponding parameters. if this model is not sufficient to represent the process, then the second-order model will be developed. the general second order model is as given below: (5) where y2 is the estimated response based on second order equation, the parameters bo, b1, b2, b3, b4, b11, b12, b13 and b44 are to be estimated by the method of least squares. optimization of surface finish by the lingosolver approach lingo is a mathematical modeling language which is used in linear and non-linear optimization in engineering sciences, operation research, etc.the simplicity of operation and computational efficiency are the two main attractions of the lingo solver approach. optimization helps find the answer that yields the best result, or attains the highest profit, output, or happiness. it also has the potential to find the answer that achieves the lowest cost, minimize surface roughness, minimizes waste, or prevents discomfort. through the lingo solver optimization we can obtain the global optimum values (i.e., minimization or maximization characteristics). 3. experimental details a detailed survey has been carried out to find out how metal cutting parameters, namely cutting speed, feed rate, depth of cut and tool nose radius of the single point cutting tool were selected for experimentation. the range of each parameter is set at three different levels, namely low, middle and high based on industrial practices as shown in table 1. the factorial design with eight added centre points (24 + 8) used in this work is a composite design. the complete design consists of 24 experiments as shown in table 2 (all factors are in coded form). the coded number for variables used in tables 1 and 2 are obtained from the following transformation equations as suggested by (birdie 1993; montgomery 1991): x1 = (ln v ln 112) / (ln 112ln 39) (6) x2 = (ln f ln 0.10) / (ln 0.10 ln 0.06) (7) x3 = (ln d ln 0.4) / (ln 0.4ln 0.2) (8) x4 = (ln r ln 0.8) / (ln 0.8ln 0.4) (9) where x1 is the coded value of cutting speed v, x2 is the coded value of feed rate f, x3 is the coded value of depth of cut d, and x4 is the coded value of tool nose radius r. 40 lb abhang and m hameedullah in this investigation, a commercial alloy steel work piece (en-31 steel alloy) is machined on heavy duty lathe machine (ltm-20). the chemical composition of the material is shown in table 3. this material is suitable for a wide variety of automotive type applications the construction of axles, roller bearings, ball bearings, shear blades, spindles, mandrels, forming and molding dies, rollers, blanking and forming tools, knurling tools and spline shafts. these are all examples of automotive components produced using materials where turning is the prominent machining process used. an optical surface roughness measuring microscope was used to measure surface roughness (ra) of the machined components. the surface roughness was measured at three equally spaced locations around the circumference of the workpieces to obtain statistically significant data for each test. the cutting tools used for experimentation were cnma 120404, cnma 120408, cnma 120412 and diamond shape carbide (make: widia india limited, bangalore). the tool holder used for experimentation was widax sclcr 1212, fo9 (iso designation). 4. results and discussion the experimental study was conducted to see the effect of cutting parameters and tool geometry (i.e., nose radius) on the outcome of the steel turning process. the variation of surface roughness with respect to the variables are shown in figs. 1 through 5. it can be observed that cutting speed (v) and nose radius (r) have a negative influence, while feed rate (f) and depth of cut (d) have a positive influence on the surface roughness (ra). the surface roughness (ra) of en-31 steel decreased with increased cutting speed (v) and tool nose radius (r) whereas it increased with an increasing feed rate (f) and depth of cut (d). it can also be seen that the feed rate influences the surface roughness more predominantly than the other factors. large values of tool nose radii also yield low surface roughness as compared to the smaller tool nose radii of the cutting tool (fig. 5). hence smaller values of feed rate and depth of cut must be selected in order to achieve better surface finish during the steel turning process. in order to under stand the relationship between the machining response and parameters, the experimental results were used to develop the mathematical models table 1. process variabls and their levels table 2. design matrix with experimental results table 3. chemical composition of an alloy steel [en-31] work piece table 4. anova for the first-order model (first block) note: df = degree of freedom, s = significant table 5. anova for the second-order model (whole block) note: df = degree of freedom, s = significant 41 optimal machining parameters for achieving the desired surface roughness in turning of steel using rsm. in this work, a commercially available statistical minitab software package was used for the computation of the regression constants and exponents. the developed equations clearly show that the feed rate is the most influencial parameter on surface roughness followed by tool nose radius and depth of cut. this is in agreement with the work of (birdie 1993; sundaram, lambert 1981). the increase in feed rate increases surface roughness, but surface roughness decreases with increasing cutting velocity and tool nose radius. during machining, if the feed rate is increased, the normal load on the tool also increases figure 1. surface roughness and feed rate relationship (depth of cut 0.2mm and nose radius 0.44 mm) figure 2. surface roughness and cutting speed relationship (depth of cut 0.20mm and nose radius 0.4mm) figure 3. surface roughness and depth of cut relationship (tool nose radius 0.4 mm and feed rate 0.6 mm/rev.) 42 lb abhang and m hameedullah and it will generate heat which in turn increases the surface roughness. this is anticipated as it is well known that for a given tool nose radius, the theoretical surface roughness is generally (ra = f2 /32r) (sundaram, lambert 1981). thus, with an increase in depth of cut, the surface roughness value increases, because with an increase in depth of cut chatter may result causing degradation of the workpiece surface (chen 2000), and a larger tool nose radius reduces surface roughness. the surface roughness values obtained by using an insert radius of 1.2 mm were less than the surface roughness values obtained by using the insert radii of 0.8 mm and 0.4 mm. the reason for obtaining better surface quality with in insert radius of 1.2 mm than with the other two inserts may be ascribed to the form of better roundness of this insert than the other two. 4.1 the roughness models the proposed first order model developed from the above functional relationship using the rsm method is as follows: ra = 8.6 0.00017v + 28.2f + 3.74d 0.688r (10) the transformed equation of surface roughness prediction is as follows: ra = 26.049 (v-0.0265 f0.224 d0.114 r-0.038) (11) equation (11) is derived from the eq. (10) by substituting the coded values of x1, x2, x3 and x4 in terms of lnv, lnf, lnd and lnr. the analysis of variance and the f-ratio test have been performed to justify the fitness of the mathematical model. since the calculated value of the f-ratio is more than the standard tabulated value of the f-ratio for surface roughness as shown in table 4, the model is adequate. their p-values smaller than 5% or equal to zero, at a 95% confidence level to represent the definite relationship between the machining parameters and machining response for the turning process. the multiple regression coefficient of the first-order model was found to be 0.8671.this shows that the first order can explain the variation to the extent of 86.71%. in order to see whether a secondorder model can represent more accurately than the first order, a second order model was developed. it is to be noted from the second-order equations that some of the coefficients are not considered. only significant figure 4. surface roughness and tool nose radius relationship (depth of cut 0.2 mm and feed rate 0.06 mm/rev.) figure 5. surface roughness and tool nose radius relationship (depth of cut 0.4 mm and feed rate 0.1 mm/rev.) 43 optimal machining parameters for achieving the desired surface roughness in turning of steel parameters and their coefficients are included in the second-order equation. the remaining insignificant parameters are omitted. the student's t-test was applied to determine the significance and non-significance of these parameters and their coefficients. the second-order mathematical model is as follows: y2 = 5.13 + 0.0155v + 55.0f + 9.51d + 0.60r 0.0000 51v2 + 0.0187f2 0.0110d2 0.00102r2 48.6vf10.5 vd 0.38dr (12) where, y2 is the estimated response of surface roughness based on second-order equation, v is the cutting speed in meters/minute, f is the feed rate in mm/revolution, d is the depth of cut in mm and r is the tool nose radius in mm. it is observed that feed rate has positive influence followed by depth of cut, cutting speed, and tool nose radius on the surface roughness (ra). the surface roughness (ra) of en-31 steel decreased with increasing cutting speed (v) and tool nose radius, whereas it increased with an increase in feed rate and depth of cut. the analysis of variance for the second order model is shown in table 5. the model is adequate since their p-values are smaller than 5% or equal to zero at a 95% confidence level to represent the relationship between the machining parameters and machining response for the turning process. the multiple regression coefficient of the second order was found to be 0.8983. this means that the secondorder can explain the variation to the extent of 89.83%. since the difference of multiple regression coefficients between the first order and the second order is only 3.12%, it can be concluded that the first order model is adequate to represent the steel turning process under consideration. the first order mathematical model acts as an objective function in order to minimize the output surface roughness factor. ramin = 8.68 0.00017v + 28.2f + 3.74d 0.688r (13) the constrained optimization problem is stated the objective function of minimum ra using the above model. the constraints are subjected to, 39 < v < 189, 0.06 < f < 0.15, 0.2< d < 0.6 and 0.4 < r < 1.2, xi1 < xi < xi4. in this case xi1 and xi4 are the upper and the lower bounds of process variables xi, and x1, x2, x3, and x4 are the logarithmic transformation of cutting speed, feed rate, depth of cut and tool nose radius. table 6 shows a global optional solution found at step: 6 with optimum machining conditions. the objective value obtain by lingo-solver is 10.26223 µm. 4.2 optimization the objective of the optimization was to find cutting parameters within the speed range of 39 miters/minute to 189 meters/minute; the feed rate range 0.06 mm/revolution to 0.15 mm/revolution, a depth of cut range from 0.2 mm to 0.6 mm and a tool nose radius range of 0.4 to 1.2 mm. cutting parameters all should be carried out so that the surface roughness (ra) is minimized. the best result consisted of speed of 189 meters/minute, a feed rate of 0.06 mm/revolution, a depth of cut of 0.2 mm and a tool nose radius of 1.2 mm. this gave an optimum surface roughness (ra) of 10.26 µm corresponding to the dezirable 95% confidence interval. hence in order to reduce machining time and to achieve a better surface finish and metal removal rate, a combination of a high speed and a low feed rate, with a lower depth of cut and high tool nose radians must be selected for the machining process. this optimization approach is quite advantageous in order to have the range of the surface roughness values and their corresponding optimum machining conditions for certain ranges of input machining parameters. it would be helpful for a manufacturing engineer to select the machining conditions for the desired machining performance of the product. this lingo-solver approach provides global optimum machining conditions for corresponding minimum values of surface roughness. the lingo-solver table 6. global optimum solution table 7. confirmation test results note: v cutting speed m/min. f feed rate mm/rev, d depth of cut mm and r nose radius mm 44 lb abhang and m hameedullah approach, used to optimize the mathematical model, was found to be the most useful technique for research. with the known boundaries of surface roughness and machining conditions, machining could be performed with a relatively high rate of success, with selected machining conditions. 4.3 verification test of optimal result after identifying the most effective parameters, the final step is to verify the optimal values of parameters and the surface roughness (i.e., response) by conducting confirmation experiments and comparing the results of these validation runs with respects to the values obtained by the lingo-solver optimization model. the validation experiments were conducted according to the optimal process parameter levels (i.e., high cutting speed, low feed rate, low depth of cut and high tool nose radius). three trials were conducted and the corresponding surface roughness values were measured. the average experimental values and software predicted value are 10.29 µm and 10.26 respectively, at the 95% confidence levels. the table. 7 shows experimental values and optimal values of surface roughness. the experimental values were compared with the predicted values from the lingosolver and the software found that the experimental values were very close to the predicted values. 5. conclusions a reliable surface roughness model for steel turning was developed using rsm and incorporated cutting speed, feed rate, depth of cut, and the tool nose radius. the study was optimized by the lingo-solver approach, which is a global optimization technique. this has resulted in a fairly useful method of obtaining process parameters in order to attain the required surface quality. the optimal parameter combination of the turning process corresponded to a cutting speed of 189 mitres/minute, a feed rate of 0.06 mm/revolution, a depth of cut of 0.2 mm, and a tool nose radius of 1.2 mm by the lingo-solver approach.this has validated the trends available in the literature and extended the data range to the present operating conditions, apart from improving the accuracy and modeling by involving the most recent modeling method. the application of lingo-solver optimization to obtain optimal machining conditions will be quite useful at the computational planning stage in the production of high quality goods with tight tolerances by a variety of machining operations, and in the adaptive control of automated machine tools. acknowledgment the authors would like to express their deep gratitude to the department of mechanical engineering of aligarh muslim university for providing laboratory facilities and financial support. references abhang lb, hameedullah m (2010), analysis of surface roughness by turning process using statistical method. proceedings of the int. conf. on advances in industrial engineering application chennai, india 138-146. abhang lb, hameedullah m (2010), a predictive model for tool wear rate in metal cutting using response surface methodology. proceedings of the int. conf. on advances in industrial engineering applications chennai, india 171-180. aslan e, camuscu n, birgoren b (2007), design optimization of cutting parameters when turning hardened aisi4140 steel with al2o3 + ticn mixed ceramic tool. materials and design 28:1618-1622. birdie ma (1993), surface roughness model for turning grey c.i. (154bhn). proceedings of institution of mechanical engineering, part b, j. of engineering manufacture 207:43-54. chen w (2000), cutting forces and surface finish when machining medium hardness steel using cbn tools. int. j. of machine tools and manufacture 4(1):455-466. dilbag r (2007), optimization of tool geometry and cutting parameters for hard turning. j. of materials and manufacturing process 22:15-21 feng cx, wang x (2002), development of empirical models for surface roughness prediction in finish turning. the int. j. of advanced manufacturing technology 20:348-356. hasegawa ma, seireg ra, lindberg (1976), surface roughness model for turning. int. j. of tribology 8(2):285-289. konig w, komanduri r, tonshoff hk, ackerson hg (1984), machining of hard materials. annals of cirp 33(2):417-427. kopac j, bahor m, sokovice m (2002), optimal machining parameters for achieving the desired surface roughness in fine turning of cold preformed steel work piece. int. j. of machine tools and manufacture 42:707-716. manna a, bhattacharyya b (2004), investigation for optimal parametric combination for achieving better surface finish during turning of al/sic-mmc. int. j. of advanced manufacturing technology 23:658-665. montgomery dc (1991), design and analysis of experiment's. third edition, john wiley and sons, new york, usa 521-568. mehrban d, naderi v, panahizadeh h, moslemi n 45 optimal machining parameters for achieving the desired surface roughness in turning of steel (2008), modeling of tool life in turning process using experimental method. int. j. of material forming springer/esa form 1-4. nalbant m, gokkaya h, sur g (2007), application of taguchi method in the optimization of cutting parameters for surface roughness in turning. materials and design 28:1379-1385. ozel t, karpat y (2005), predictive modeling of surface roughness and tool wear in hard turning using regression and neural networks. int. j. of machine tools and manufacture 45:467-479. suresh r, deshmukh s (2002), a genetic algorithmic approach for optimization of surface roughness prediction model. int. j. of machine-tools and manufactures 42:675-680. sundaram rm, lambert bk (1981), mathematical models to predict surface finish in fine turning of steel. part 1, int. j. of production research 19:547-556. wang x, balaji ak, jawahir is (2002), performance based optimal selection of cutting condition and cutting tools in multi-pass turning operations using genetic algorithms. int. j. of production research 40(9):2050-2065. yang wh, tarng ys (1998), design optimization of cutting parameters for turning operations based on the taguchi method. j. of materials processing technology 84:122-129. the journal of engineering research (tjer), vol. 18, no. 1, (2021) 36-43 *corresponding author’s e-mail: hadeeradeeb@uomosul.edu.iq doi: 10.53540/tjer.vol18iss1pp36-43 determination of rock slope stability using stereographic projection between sulav and amadiyah resorts – a study from northern iraq hadeer gh. m. adeeb 1,*, and ibrahim s. i. al-jumaily 2 1 dams and water resources research center, mosul university, mosul, iraq 2 petroleum and mining engineering college, mosul university, mosul, iraq abstract: geological discontinuities play a significant role in the assessment of rock slope stability. rock slope stability has been studied on the main road between sulav and amadiyah resorts in duhok governorate on the southern limb of mateen anticline, to determine the expected rockslides on this road. five stations were chosen to study these rockslides that may occur on these steep slopes. all these stations were within the pila spi formation that consists of hard dolomitic limestone and covering the areas from sulav resort towards amadiyah district with a length of up to 2.5 km. the stereographic analysis was used to study and classify the stability of these slopes. the analysis showed in all stations the possibility of plane sliding to happen on the bedding plane, and the wedge sliding between the bedding plane and planes of all joint sets, as well as the occurrence of rockfall on some stations. keywords: rock slope stability; stereographic analysis; amadiyah; sulav. والعمادیة سوالف مصیفي بین الستیریوغرافي سقاطاإل باستخدام الصخریة المنحدرات استقراریة حدیدت العراق شمال من دراسة ابراھیم الجمیلي ابراھیم سعد و *ھدیر غازي محمد ادیب تم دراسة استقراریة المنحدرات الصخریة تلعب االنقطاعات الجیولوجیة دوًرا مھًما في تقییم استقرار المنحدرات الصخریة. :الملخص یة على الطریق العام بین مصیفي سوالف والعمادیة في محافظة دھوك وعلى الجناح الجنوبي لطیة متین المحدبة، وذلك لتحدید االنزالقات الصخر الطریق. اختیرت خمسة محطات لدراسة االنزالقات الصخریة التي قد تحدث على تلك المنحدرات؛ والمتكونة من تكوین بیالسبي المتوقعة على ھذا استخدم التحلیل .كم 2.5الجیري الدولومایتي الصلدة والشدیدة االنحدار، وابتداءاً من مصیف سوالف باتجاه مصیف العمادیة وبطول یصل الى حدود ستیریوغرافي في دراسة وتصنیف استقراریة تلك المنحدرات الصخریة. تبین من التحلیل احتمالیة حصول االنزالقین المستوي، على اسطح ال التطبق واالسفیني بین مستویات التطبق ومستویات مجامیع الفواصل في كل محطات الدراسة، فضالً عن حدوث الساقط الصخري في بعض المحطات. .سوالف؛ العمادیة ؛التحلیل الستیریوغرافي ؛استقراریة المنحدرات الصخریة :المفتاحیةالكلمات determination of rock slope stability using stereographic projection between sulav and amadiyah resorts 37 1. introduction most road network, in the study area, is located in mountainous terrain. it is a very steep topography and slope failures are often caused when road construction is undertaken. large-scale instability phenomena in cut slopes frequently occur in structurally and geologically complicated regions. rockslides or so-called slope failure or landslides are considered natural phenomena with geological hazards that may occur suddenly or over time. a simple landslide definition according to working party on world landslide inventory (wp/wli) (1990) and cruden (1991) is "the movement of a mass of rock, earth or debris down a slope". landslides refer to outstanding materials and environmental challenges to the societies living in areas exposed to various risks of landslides (unisdr, 2015). geological discontinuities play an important role in assessing rock slope stability. particularly, the orientation of geological discontinuities i.e. dip direction and dip angle with respect to the slope orientation represents a significant input parameter in the assessment of rock slope stability. the risk of the rockslides occurs when parts of the rocks move, slide, and then collapse due to the overpowering of the catalytic forces against resistant forces for sliding. the catalytic forces are attributed either to natural or to human activity factors. the natural factors include tectonic and gravity activities, exposition of the region to differential weathering, intense rainfalls, rapid snow melting, and ground vibration by earthquakes. the human activities involve vegetation removal, excavation operations, manmade activities represented by the removal of rock masses under the roads during road cut operations, construction of buildings, and residential houses in the mountainous area. all of these factors lead to unbalancing the rock masses and soils and consequently become weak, and their strength represented by the coherence and friction forces are reduced against the movement. the daylighting slope refers to a slope where the discontinuities dip at an angle less than the exposure slope angle and in the same direction of slope inclination. the areas below the southern limb of mateen anticline on the main road linking between sulav and amadiyah resorts are considered one of the most vulnerable areas where the daylighting rock layers are highlighted (markland, 1972). this is due to the cutting road construction in the region, which is considered an important tourist place. the main failure mechanisms shown in this study can be summarized as planar failure which is governed by the main discontinuity dipping in the direction of the slope. the wedge failure mechanism is governed by two intersection discontinuities with their intersection lines dipping towards the slope. finally, the rockfall consists of loose blocks or slabs due to slipping, rolling, or toppling on the slope. the study area is considered one of the areas exposed to the risk of landslides where the occurrence of rock cuts prevails (daylighting). this is due to human activities represented by the road cutting operations in these areas that lead to landslides. one of the reasons that help to slide is the street design line, which in some parts is at a high angle with the direction of the bedding strike. in this case, the layers form a weak zone, which allows the catalytic forces for sliding to overcome the anti-sliding forces and triggered the landslides. this is due to weak cohesion and friction forces with increasing the weight coefficient of the materials. the higher the layers dip, the greater their instability and the failure start to move downward or remain in an unstable situation. for this reason, this area was considered one of the areas where a high risk of landslides is expected, as well as the rockfall, in which the rock masses generally fall free from the highest slopes, particularly in the areas where rocks are daylighted. 2. location of the study area the study area is located on the main road between sulav and amadiyah resorts in duhok governorate – the kurdistan region of iraq. the road is estimated for about 2-3 km long, and the area is about 65 km ne of duhok city, between the longitudes (43° 28‵ 25‶) and (43° 29‵ 55‶) e, and latitudes (37° 06‵ 13‶) and (37° 06‵ 18.72‶) n. (fig. 1). 3. geology of the study area the study area is situated geologically on the southern limb of the mateen anticline within the high folded zone of iraq, (fig. 2). the exposed rocks are of the pila spi formation of the middle-upper eocene. it consists of well-bedded, medium to thick layers, the pale white of hard dolomitic limestone. the formation is exposed, with a steep slope ranged from 40° to 70°, as a continuous belt surrounding the anticline in the form of continuous anticlinal ridges. these outcrops have a high density of joint sets are exposed on the northern edge of the main road between sulav and amadiyah resorts. 4. seismic hazards in the region of the study area strong earth trembling triggers landslides in many different topographic and geologic settings. rockfall, soil slides, and rockslides from steep slopes, with relatively thin disaggregated soil or rock, or both have been the most abundant type of landslide triggered by historical earthquakes (wieczorek, 1996). the seismic hazards map of the investigated area indicates its occurrence in a zone classified as a minor damage zone, with seismic intensity between 4-5, (al-sinawi and al-qasrani, 2003), (fig. 3). this means that the region is under the influence of seismic activity. hadeer gh. m. adeeb, and ibrahim s. i. al-jumaily 38 study area figure 2. geological map of the studied area (sissakian and fouad, 2012), imposed over the dtm (sissakian et al., 2020). figure 1. google earth image shows the location of the studied area. figure 3. seismic isointensity map of iraq 1900-1988 (al-sinawi and al-qasrani, 2003). determination of rock slope stability using stereographic projection between sulav and amadiyah resorts 39 5. objective the main objective of this study is to assess the stability of these rock slopes by conducting geological analyses, to help the authorities establish appropriate engineering solutions to mitigate these landslides and to reduce their risks and severity on road users and buildings. some of these analyses are to determine the type of landslide that occurs and is likely to occur on the side of the road, as well as the factors that affect the landslide. 6. methodology the method of the study includes both field and laboratory work. in the fieldwork, a detailed field survey of these landslides was carried out across five (5) stations on the main road linking sulav and amadiyah resorts, as evidence exists in these stations that indicate that landslides have occurred. it is therefore considered study stations. at each station, layer and joint attitudes were measured in terms of the dip angle and dip direction, as well as taking photographs of each station. the laboratory work included analysing the obtained measurements using the schmidt net for stereographic projection, by representing the attitude of the layers and joints by great circles using software concerning stereographic projection. both discontinuities and slope faces were represented as planes using the abovementioned net (wyllie and mah, 2004). the stability or instability of the rock mass depends on its weight and on the external forces that affect it, such as the dynamic loads, the shear forces, and the pore water pressure on the sliding surface. this leads to define the safety factor, which is the ratio of the actual shear strength to the shear resistance required for equilibrium. it is accounted for by the angle of internal friction, (department of the army, us army corps of engineer, 2003). the internal friction angle of the dolomitic limestone layers of the pila spi formation was represented by a circle of 25º (al jawadi, 2013). the following symbols and colours are used below in the stereographic figures: ø: circle of cohesion angle for discontinuities surfaces. s0=gs: general slope face and bedding attitude as great circles. s1: first discontinuity attitude as a great circle. s2: second discontinuity attitude as a great circle. s3: third discontinuity attitude as a great circle. s4: fourth discontinuity attitude as a great circle. s5: fifth discontinuity attitude as a great circle. plane sliding wedge sliding rockfall 7. analysis of the rock slope stability the rock slopes were analysed to examine their stability through selected stations in the pila spi formation that is exposed in the study area as given in table 1. the following is an analytical description of the landslide and rockfall of such rock slopes at these stations according to the hunt classification (hunt, 2006): table 1. geological and engineering data of the study stations in the investigated area. geological and engineering data stations formation bedding attitude (s0) = general slope face (gs) first discontinuity attitude (s1) second discontinuity attitude (s2) third discontinuity attitude (s3) fourth discontinuity attitude (s4) fifth discontinuity attitude (s5) cohesion (φ) angle 1 pila spi formation 80/180 88/250 56/293 60/066 77/089 --25 ° 2 62/178 80/074 51/300 58/044 84/096 -- 3 66/179 51/279 54/308 50/067 80/082 -- 4 61/179 80/260 44/291 30/058 83/075 79/289 5 63/178 --58/297 64/036 81/080 48/352 figure 4. the stereographic diagram of the station (1) shows the relationship between the slope face (bedding plane) and the joints, and plane and wedge sliding on the limestone beds of the pila spi formation. figure 5. showing the plane sliding (p), and the wedge sliding (w) at station 1. w p hadeer gh. m. adeeb, and ibrahim s. i. al-jumaily 40 station no. 1 the slopes of the dolomitic limestone of pila spi formation were exposed on the left side of the main road overlooking the entrance of sulav resort. this station is characterized by the possibility of the plane sliding on the bedding planes, which also represents the slope face, and the wedge sliding between the bedding planes and the discontinuities represented by the joint sets (s1, s2, s3, and s4) because the friction angle of the sliding surfaces is less than their dip angle (fig. 4 and table 1). such joints also act as lateral release surfaces for sliding (fig. 5). station no. 2 the first part of this station is the extension of the first station. the landslides are of two types: the plane sliding on the bedding plane, and the wedge sliding between the bedding plane and the same joint sets referred to in station 1, (figs., 6(a), and 7). whereas rockfall, plane, and wedge sliding occur in the second part of the station, because of changes in the direction of the main road and its intersection with the bedding strike (figs., 6(b), and (8)). the presence of the joints helps to cause rockfall in this part of the station. rockfall occurs on steep slopes; therefore the unstable rocks are difficult to stay in their position due to the effect of the joints so that they fall to the bottom of the slope. some of the rockfalls are gathered in very narrow longitudinal ditches that have been prepared for this purpose. it is noted that most of these blocks settle down on the main road due to the narrow area of the shoulder of the road. such blocks are nearly a few meters away from the buildings in sulav resort, which threatens to destroy them and cause endanger to their inhabitants, as well as to cause significant environmental hazards (fig. 8). station no. 3 this station is characterized by the occurrence of the rockfall as well as both plane and wedge sliding due to the high slope of the outcrops overlooking the side of the main road. the plane sliding occurs on the bedding plane, while the wedge sliding occurs between the bedding plane and the joint sets mentioned in the previous two stations, (figs. 9, 10, and 11). station no. 4 this station consists of two parts, in each of which there is the plane sliding on the bedding planes, and wedge sliding between the bedding planes and the same joint sets described in the previous stations, as well as the rockfall in the other parts of this station, (figs., 12 (a and b), 13, and 14). (a) (b) figure 6. the stereographic diagrams of the station 2 ((a) and (b)) show the relationship between the slope face (bedding plane) and the joints, and the different types of sliding on the dolomitic limestone of the pila spi formation. figure 7. showing the plane sliding (p), and the wedge sliding (w) at station 2a. person p w determination of rock slope stability using stereographic projection between sulav and amadiyah resorts 41 figure 10. showing the plane sliding (p), and the wedge sliding (w) at station 3. w p 0 0.5 m figure 8. showing the rockfall (f), and the plane sliding (p), and the wedge sliding (w) at station 2b. w f p figure 12. the stereographic diagrams of station (4a, b) show the relationship between the slope face (bedding plane) and the joints, and the different types of sliding on the dolomitic limestone of the pila spi formation. (b) (a) figure 11. showing the rockfall (f), and the plane sliding (p), and the wedge sliding (w) at station 3. p w f figure 9. stereographic projection of the station (3) shows the relationship between the slope face (bedding plane) and the joints, and the different types of sliding on the dolomitic limestone of the pila spi formation. hadeer gh. m. adeeb, and ibrahim s. i. al-jumaily 42 station no. 5 the rock slopes of this station are also exposed on the left side of the main road leading to amadiyah district centre. the two types of sliding are repeated; plane sliding on the bedding planes, and wedge sliding between the bedding planes and the same joint sets for the above-mentioned stations, (figs, 15 and 16). 8. conclusion investigation of rock slopes along the road joining sulav and amadiyah resorts has achieved the following conclusions: 1. the rock slopes consist of limestone beds of pila spi formation (middle-upper eocene) at the southern limb of mateen anticline trending eastwest. these beds are dipping southward, with their dip slope ranges (40°-70°). 2. there are five sets of discontinuities (joints) disrupting the limestone beds at high angles and with various angular relations with the bedding strike. 3. stereographic representation of bedding and discontinuity planes, as well as the angle of internal friction for limestone beds, have characterized the following modes of rock failure in the investigated area: a. in all study stations where the road trend coincides with the bedding trend (or at low angles), the rock sliding has been categorized into plane and wedge sliding. the plane sliding occurs down-dip of bedding planes where the joint sets accommodate as back and lateral release for sliding of rock blocks. wedge sliding occurs down-dip of intersection line between conjugated joint sets that are with low to moderate angles with bedding planes. meanwhile, wedge sliding occurs between vertical joints transversal to bedding strike and bedding planes as well. b. the rockfall occurs at some stations where the road trend is at a high angle with the bedding p w w figure 14. showing the rockfall (f), and the plane sliding (p), and the wedge sliding (w) at station 4b. w p f figure 13. showing the two types of the plane sliding (p), and the wedge sliding (w) at station 4a. p w figure 15. the stereographic diagram of station (5) shows the relationship between the slope face (bedding plane) and the joints, and plane and wedge sliding occurred on dolomitic limestone of pila spi formation. s2 s5 s3 ø s4 s0 figure 16. showing the plane sliding (p), and the wedge sliding (w) at station 5. determination of rock slope stability using stereographic projection between sulav and amadiyah resorts 43 strike. 4. in all study stations, it has been found that the longitudinal ditches, designed to accommodate falling and sliding rock masses, appear to be very close to the slopes where potential landslides and rockfall are likely to happen. this contributes to the falling of the rock masses on the main road immediately away from these ditches. 9. recommendation based on the findings obtained from the analysis of the information relating to the types of rock sliding in the study area, some recommendations were proposed as follows: 1establish good drainage networks for rainwater and springs whose routes intersect with roads and maintain good monitoring of them, particularly during the rainy season, and ensure that they are not blocked and diverted. 2one of the significant recommendations, and as an aimed solution to preserve the tourist area from sliding and falling rock blocks, the researchers suggest that a wire mesh barrier can be placed between the longitudinal ditches and the main road. this barrier is placed at an appropriate height with the inclination of the slopes, to prevent the rolling rock blocks from reaching the main road. conflict of interest the authors declare no conflict of interest. funding no funding was received for this project. acknowledgment the authors would like to express their gratitude to the mosul university, mosul, iraq for their support. also, we would like to thank the dams and water resources research center, and petroleum and mining engineering college, the university of mosul for their facilities, which have helped to improve the quality of this work. references al-jawadi, a. s. h., 2013. evaluation of the rock mass at the bekhme dam site, north east iraq. unpublished phd. thesis, college of science, university of mosul. 183p. al-sinawi, s. a., and al qasrani, z., 2003. earthquake hazards considerations for iraq, 4th international conference of earthquake engineering and seismology, may 2003, tehran, islamic republic of iran. cruden, d.m., 1991. a simple definition of landslide. bulletin international association of engineering geology, 43, pp. 27-29. department of the army u.s. army corps of engineers, 2003. engineering and design slope stability., appendix c. hunt, r. e., 2006. geologic hazards: a field guide for geotechnical engineers, 1st edition, crc press taylor and francis group, an informa business, bocca raton, london, and new york.323p. markland, j. t., 1972. a useful technique for estimating the stability of rock slopes when the rigid wedge sliding type of failure is expected. imperial college of rock mechanics research, report no. 19, 10pp. sissakian, v. k., and fouad, s.f., 2012. geological map of iraq, scale 1: 1000 000, 4th edition. iraq geological survey publications, baghdad, iraq. sissakian, v. k., ghafur, a. a., abdulhaq, h. a., omer, h. o., 2020. determining the tectonic origin of the gara and mateen anticlines using geomorphological and structural forms, iraqi kurdistan region. ukh journal of science and engineering· 4(2), pp84-100. unisdr, 2015. sendai framework for disaster risk reduction 2015-2030 functions, united nations office for disaster risk reduction (unisdr), 2009 unisdr. terminology on disaster risk reduction, geneva, may 2009. wieczorek, g. f., 1996. landslide triggering mechanisms. in a special report 247: rock strength properties and their measurement. in turner a. k. and schuster r. l., transportation research board, national research council, special report, washington dc, chapter 4, pp.76-90. wp/wli (international geotechnical societies' unesco working party on world landslide inventory) 1990. a suggested method for reporting a landslide. bulletin international association of engineering geology, 41: 5-12. wyllie, d.c. and mah, c.w., 2004. rock slope engineering: civil and mining, 4th edition. based on the third edition by e. hoek and j. bray, spon press, taylor and francis group, london and new york. 431p. 1. introduction 2. location of the study area 3. geology of the study area 4. seismic hazards in the region of the study area the following symbols and colours are used below in the stereographic figures: ø: circle of cohesion angle for discontinuities surfaces. s0=gs: general slope face and bedding attitude as great circles. s1: first discontinuity attitude as a great circle. s2: second discontinuity attitude as a great circle. s3: third discontinuity attitude as a great circle. s4: fourth discontinuity attitude as a great circle. 7. analysis of the rock slope stability station no. 1 the first part of this station is the extension of the first station. the landslides are of two types: the plane sliding on the bedding plane, and the wedge sliding between the bedding plane and the same joint sets referred to in station 1, (figs., 6(... this station is characterized by the occurrence of the rockfall as well as both plane and wedge sliding due to the high slope of the outcrops overlooking the side of the main road. the plane sliding occurs on the bedding plane, while the wedge sliding... 8. conclusion 9. recommendation conflict of interest funding no funding was received for this project. acknowledgment references the journal of engineering research (tjer), vol. 19, no. 1, (2022) 1-12 *corresponding author’s e-mail: waris@squ.edu.om doi:10.53540/tjer.vol19iss1pp1-12 comparison of oman seismic code with international counterparts based on regional seismic hazard muhammad bilal waris1*, khalifa al-jabri1, and issa el-hussain2 1 department of civil and architectural engineering, sultan qaboos university, muscat, sultanate of oman. 2 earthquake monitoring center, sultan qaboos university, muscat, sultanate of oman. abstract: this study compares the oman seismic code (osc) developed in 2013 with the uniform building code 1997 (ubc) and the international building code 2006 (ibc) that are frequently used for seismic analysis and design in oman. base shear, story shear force distribution, and story drift ratios are compared using the equivalent static load and the response spectrum analysis methods. the discussion considers reinforced concrete buildings having a footprint of 25 m × 40 m with heights of 10 m, 19 m, 40 m, and 58 m having 3-, 6-, 13and 19 stories, respectively. the buildings are considered as supported on very dense soil (type c) and have normal occupancy conditions. the zone-1 of osc is compared with the zone-1 of ubc and ibc considering the low and the normal ductility class definitions of osc. for the low ductility class, the values of design base shear for different building heights from ibc are 82% for the equivalent static method and ranged between 83%-85% for the response spectrum method compared to osc. the base shear values by ubc are 73%-81% and 75%-78% for the equivalent static and the response spectrum methods, respectively, in comparison to osc. for normal ductility class, ibc and ubc estimates are higher than osc for all building heights except for the 3-story building. the story shear forces and story drift ratios from osc also compared well with the international counterparts. the comparison indicates that zone-1 in osc is well matched to zone-1 of ubc and ibc. keywords: oman seismic code; equivalent static load method; response spectrum analysis; international building code (ibc); uniform building code (ubc). المخاطر على بناءً الدولیة نظیراتھامقارنة بین الكود العماني لتصمیم المباني ضد مخاطر الزالزل مع الزلزالیة اإلقلیمیة محمد بالل وارث، خلیفة الجابري، عیسى الحسین 2013الذي تم تطویره في عام و -) osc 2013( لتصمیم المباني ضد الزالزل كود العمانيبین ال تقارن ھذه الدراسة :الملخص في تصمیم المباني لمقاومة شائعیستخدمان بشكل واللذین) ibc 2006) وكود البناء الدولي (ubc 1997كود البناء الموحد ( مع الحمل الثابت المكافئولتحقیق اھداف الدراسة، تم تحلیل مباٍن مشیدة من الخرسانة المسلحة باستخدام طرقتي .سلطنة عمان في الزالزل 19و 13و 6و 3م وتتكون من 58م و40م و19م و10م بارتفاع 40× م 25. تبلغ مساحة المباني تحت الدراسة تحلیل طیف االستجابةو مقارنة المنطقة و ) وفي ظروف إشغال عادیةcعلى تربة شدیدة الكثافة (نوع مشیدةباني المكذلك تم افتراض أن على التوالي. ،طابقًا مع األخذ في االعتبار تعریفات فئة اللدونة المنخفضة ،كود البناء الموحد وكود البناء الدولي في كٍل من 1بالمنطقة الكود العمانيمن 1 االزاحة ، ونسببین الطوابق المختلفة القص ىالقاعدة، وتوزیع قو وة القص عندمقارنة قشملت الدراسة .في الكود العمانيوالعادیة قاعدة الرتفاعات المباني ال عند قصقوى القیم فإنبالنسبة لفئة اللیونة المنخفضة، للطوابق المختلفة. وقد بینت نتائج الدراسة، أنھ ٪ لطریقة طیف االستجابة 85-٪83بین تالمكافئة وتراوح ثابتةة ال٪ للطریق82 بلغتكود البناء الدولي المحسوبة باستخدام المختلفة كود ، عند استخدام ٪ للطریقة المكافئة الثابتة وطیف االستجابة، على التوالي78-٪75٪ و81-٪73 وقد بلغت .الكود العمانيمقارنةً بـ أعلى من كود البناء الدولي وكود البناء الموحد نتائجفئة اللیونة العادیة، فإن وفیما یخص. الكود العمانيالبناء الموحد بالمقارنة مع ونسب االزاحة بین الطوابق قوى القص نتائج مقارنة وعندطوابق. 3لجمیع ارتفاعات المباني باستثناء المبنى المكون من الكود العماني الكود العمانيفي 1أن المنطقة وقد خلصت الدراسة الىتھا الدولیة. ابشكل جید مع نظیر الكود العماني فإنھا تتطابق المختلفة من من كود البناء الموحد وكود البناء الدولي. 1لمنطقة بشكل جید مع ا مطابقة .البناء الموحدكود ؛كود البناء الدولي ؛تحلیل طیف االستجابة ؛طریقة الحمل الثابت المكافئ ؛الكود الزلزالي العمانيالكلمات المفتاحیة: comparison of oman seismic code with international counterparts based on regional seismic hazard 2 nomenclature ca acceleration coefficient for calculation of base shear in ubc cd deflection amplification factor in ibc cs coefficient for calculation of base shear in ibc ct coefficient for calculation of natural period ‘t’ in osc cu limiting factor for the natural period in ubc cv velocity coefficient for calculation of base shear in ubc fa acceleration factor based on the type of soil and seismic zonation in ibc. fv velocity factor based on the type of soil and seismic zonation in ibc. ft additional force at the top story to account for higher modes (kn) hi height of the ith-story above ground (m) hn total building height (m) i importance factor n number of stories in the building r response modification factor in ibc or overstrength and ductility factor in ubc q behavior factor qr(t) seismic load reduction factor for period ‘t’ s1d 1.0s elastic spectral acceleration (g) sae elastic spectral acceleration (g) sar design (reduced) spectral acceleration (g) sd1 1.0s design spectral acceleration in ibc (g) sds 0.20s design spectral acceleration in ibc (g) ssd short period elastic spectral acceleration (g) t prominent natural period of the building (s) t0 initial spectral period (s) tl long spectral period (s) ts short spectral period (s) vi shear force at the ith-story v base shear (kn) wi weight of the ith-story (kn) w total seismic weight of the structure (kn) δfn additional force at the top story to account for higher modes in osc (kn) ω0 system over-strength factor in ibc ωo seismic force amplification factor in ubc 1. introduction lateral loads due to earthquakes pose a challenge to the structural design of buildings that principally carry gravity loads. seismic events induce forces that are proportional to the structure's mass, act over a short duration, and must be resisted by the building without collapse. further, minor excitation should not cause any damage to minimize maintenance costs. the lateral forces due to earthquakes have a major effect on the structural integrity that must be ensured under all circumstances. experience and research have provided technical solutions that can minimize loss of life and damage associated with earthquakes. to ensure desired behaviour, seismic codes lay down the essential requirements and procedures for seismic analysis and structural design. due to a wide spectrum of civil infrastructure, specialized regulations are available for each category. the highest emphasis is laid on the seismic design of buildings to avoid catastrophic loss of life. international and national guidelines are developed and enforced to ensure life safety and minimize damage to the building stock. seismic loads are very sensitive to regional seismology and geography, but in the absence of a national code, designers rely on international guidelines. once a national seismic design code is developed, a comparative study with international standards is essential to identify the shortcomings in past designs that are based on international counterparts while also calibrating the national code. many countries have established their seismic specifications and compared those with international counterparts to investigate and quantify the differences (nahhas 2011, marino et al. 2005, noor et al. 1997, fenwick and davidson 2002, mcintosh and pezeshk 1997). the sultanate of oman is part of the arabian plate, which includes the continent of arabia and the oceanic regions of the red sea, the arabian sea, the sea of oman, and the gulf of aden. no major earthquakes have been instrumentally recorded in oman. however, global databases list an earthquake of magnitude 5.1 (richter scale) in the al-kamil area during 1971 (qamaruddin and al-harthy 2000). further, there have been numerous felt earthquakes that were reported by residents (qamaruddin and al-harthy 2000). most of these reports are related to strong earthquakes occurring at a teleseismic distance (more than 1000 km). since 2003, the earthquake monitoring center (emc) at sultan qaboos university has been monitoring the seismic activity in oman using the national seismological network. a summary of the available data from 2003 to 2017 indicates that on average 1292 earthquakes were recorded by the network annually, out of which, 398 were identified as regional (proximity to oman). further, annually an average of 66 earthquakes was recorded in the oman mountains and/or the gulf of oman. among the earthquakes that originated in the oman mountains, the maximum magnitude recorded was 5.3 in the year 2015. using this recorded data and the global catalogues of seismic records, emc carried out extensive seismic hazard assessment studies for oman (deif et al. 2013, el-hussain et al. 2012, el-hussain et al. 2014). these studies were utilized to develop the first seismic code for the design of buildings in oman to help achieve optimized design and performance of buildings with the desired level of performance and safety. this research compares the oman seismic code (osc) with the international counterparts using the equivalent static method and the response spectrum analysis method. based on the current national practice for seismic design of buildings, the international building code 2006 (ibc) and the uniform building code 1997 (ubc) are considered for comparison. the muhammad bilal waris, khalifa al-jabri, and issa el-hussain 3 zone equivalence among the codes is based on the level of seismic hazard. reinforced concrete momentresisting frame buildings with heights of 10 m, 19 m, 40 m, and 58 m are considered with 3-, 6-, 13and 19 stories, respectively. the study also compares the codes in terms of the two ductility classes defined in osc. 2. oman seismic code for buildings the oman seismic code (osc) divides the sultanate of oman into two seismic zones, referred to as zone-1 and zone-2. the zones are defined based on the design peak ground acceleration (pga), where zone-1 is the high seismic hazard zone (pga ≥ 50 gals), which includes muscat, sohar, diba, and khasab. zone-2 is the low seismic hazard zone (pga < 50 gals), which includes nizwa, sur, and salalah. osc defines elastic spectral accelerations ssd and s1d for the two zones depending on the type of soil as shown in table 1. these parameters are the basis for the estimation of seismic action using the elastic response spectrum. osc provides detailed procedures for the seismic design of buildings based on equivalent static load, response spectrum, and response history analysis methods. the code divides buildings into the following three categories based on the total building height 'hn': ● low-rise buildings: hn ≤ 20.0 m. ● med-rise buildings: 20.0 m < hn ≤ 40.0 m. ● high-rise buildings: 40.0 m < hn ≤60.0 m. where the building height 'hn' is to be measured from ground floor level and buildings with heights more than 60.0 m are not covered by the code. the code further provides guidelines for the detailing and design of reinforced concrete, steel, composite, and masonry structures that complement the respective design provision of the eurocodes. 3. methodology this study considers zone equivalence among the codes based on seismic hazard, which aligns the zone1 of osc with zone-1 of both ubc and ibc. though ubc identifies muscat as a zone-2a, waris et al. (2017) reported that this definition greatly overestimates the seismic actions using ubc and ibc compared to osc for the equivalent static load method. as zone-1 is the lowest seismic zone in the respective international codes, zone-2 of osc falls below their minimum seismic hazard level definition. the study uses ordinary reinforced concrete moment resisting frame buildings with normal occupancy conditions with four building heights, assumed to be resting on very dense soil (type-c). a single soil type is considered in the study as the goal is a comparison of the codes and the building height and type-c is the most frequently encountered soil in muscat (elhussain et al. 2013). for reinforced concrete buildings, osc identifies two ductility classes, a low ductility class (dcl) that represents buildings designed based on eurocode (ec2) only without incorporating provisions of osc, and a normal ductility class (dcn) where the provisions of osc are incorporated in the design and detailing to ensure proper energy dissipation and ductile behaviour. 3.1 structural system a simple building layout with a footprint of 25 m × 40 m is considered. figure 1 shows the typical structural plan with a grid spacing of 5 m in either direction. the first story has a height of 4.0 m and all subsequent stories have a height of 3.0 m. the following four building heights are considered in this study: ● 03-story building: g + 02 floors, hn= 10 m. ● 06-story building: g + 05 floors, hn = 19 m. ● 13-story building: g + 12 floors, hn = 40 m. ● 19-story building: g + 18 floors, hn = 58 m. all beams have dimensions of 600 mm × 200 mm, while all columns are 400 mm × 400 mm. concrete compressive strength and reinforcement yield strength are 35 mpa and 460 mpa, respectively. these dimensions and strength parameters are based on the norms in the construction industry in oman. the structural elements are considered uncracked in all cases. typical values of permanent and variable actions are considered as summarized in table 2. etabs software is used for the numerical modeling of the buildings. the floors are modeled as rigid diaphragms and p-∆ effects are considered in the analysis. as the buildings are symmetric and the principal goal of this study is to compare the three codes, the calculated story shear force is applied at the centre of mass for the respective story level and cases for ±5 % eccentricity are not included in the discussion. table 1. elastic spectral accelerations and seismic zonation (osc, 2013). local soil type seismic zones zone-1 zone-2 ssd (g) s1d(g) ssd (g) s1d(g) a 0.16 0.064 0.08 0.032 b 0.20 0.08 0.10 0.04 c 0.24 0.136 0.12 0.068 d 0.32 0.192 0.16 0.096 e 0.50 0.28 0.25 0.14 f site-specific geotechnical investigation and dynamic site response analysis required ssd short period elastic spectral acceleration. s1d – 1.0-second elastic spectral acceleration. comparison of oman seismic code with international counterparts based on regional seismic hazard 4 figure 1. typical floor plan for all buildings with the definition of xand y-direction. table 2. gravity loads acting on building. description typical floor (kn/m2) roof (kn/m 2) -self-weight of 150 mm thick slab -finishes and services total permanent action 3.75 2.50 6.25 3.75 3.60 7.35 variable action 3.0 3.0 3.2 equivalent static load method osc, ibc, and ubc each have their own set of specifications for the calculation of base shear ‘v’ and distribution of the lateral storey shear force along the building height. this section briefly reviews the equations used in the equivalent static load method by the three codes. the values of the different seismic parameters adopted in this study are indicated in their respective context. a. omani seismic code (osc) the total equivalent seismic load (base shear) – ‘v’ in the direction of the earthquake is calculated by eqn. (1). 𝑉𝑉 = 𝑊𝑊 𝑔𝑔 𝑆𝑆𝐴𝐴𝐴𝐴(𝑇𝑇) ≥ 0.11 𝑊𝑊 𝑔𝑔 𝑆𝑆𝑆𝑆𝑆𝑆𝐼𝐼 (1) where ‘w’ is the total seismic weight of the structure, ‘sar’ represents design (reduced) spectral acceleration that is calculated as: 𝑆𝑆𝐴𝐴𝐴𝐴(𝑇𝑇) = 𝑆𝑆𝐴𝐴𝐴𝐴(𝑇𝑇) 𝑞𝑞𝐴𝐴(𝑇𝑇) (2) qr(t) is the seismic load reduction factor given as: 𝑞𝑞𝐴𝐴(𝑇𝑇) = � 1 + � 𝑞𝑞 𝐼𝐼 − 1� 𝑇𝑇 𝑇𝑇𝑠𝑠 0 < 𝑇𝑇 < 𝑇𝑇𝑠𝑠 𝑞𝑞 𝐼𝐼 𝑇𝑇 ≥ 𝑇𝑇𝑠𝑠 � (3) where ‘q’ is the behaviour factor that depends upon the ductility class of the reinforced concrete building. the importance factor ‘i’ is taken as 1.0 for normal occupancy. the code provides eqn. (4) for calculation of the prominent natural period but recommends verification using the rayleigh quotient method. the value of ‘ct’ is 0.075 for reinforced concrete momentresisting frames. ‘sae’ is the elastic spectral acceleration based on the prominent natural period ‘t’ that is calculated using eqn. (5). 𝑇𝑇 = 𝐶𝐶𝑡𝑡(𝐻𝐻𝑛𝑛) 3 4 (4) 𝑆𝑆𝐴𝐴𝐴𝐴(𝑇𝑇) = ⎩ ⎪ ⎪ ⎨ ⎪ ⎪ ⎧0.4𝑆𝑆𝑆𝑆𝑆𝑆 + 0.6𝑆𝑆𝑆𝑆𝑆𝑆 𝑇𝑇 𝑇𝑇0 𝑇𝑇 ≤ 𝑇𝑇0 𝑆𝑆𝑆𝑆𝑆𝑆 𝑇𝑇0 < 𝑇𝑇 ≤ 𝑇𝑇𝑆𝑆 𝑆𝑆1𝑆𝑆 𝑇𝑇 𝑇𝑇𝑆𝑆 < 𝑇𝑇 ≤ 𝑇𝑇𝐿𝐿 𝑆𝑆1𝑆𝑆𝑇𝑇𝐿𝐿 𝑇𝑇 2 𝑇𝑇 > 𝑇𝑇𝐿𝐿 ⎭ ⎪ ⎪ ⎬ ⎪ ⎪ ⎫ (5) 𝑇𝑇𝑆𝑆 = 𝑆𝑆1𝐷𝐷 𝑆𝑆𝑆𝑆𝐷𝐷 ; 𝑇𝑇0 = 0.20𝑇𝑇𝑆𝑆 (6) where ‘ssd’ and ‘s1d’ are elastic spectral accelerations associated with short period ‘ts’ and one-second elastic spectral acceleration, respectively. ‘ssd’ and muhammad bilal waris, khalifa al-jabri, and issa el-hussain 5 ‘s1d’ are based on the geographic location of the building (seismic zone) and the soil conditions at the site as shown in table 1. ‘tl’ is the transition limit of the response spectrum to the long-period range that is taken as 8s in osc. ‘t0’ defines the start of the plateau in the response spectrum and ‘ts’ defines its end, calculated using eqn. (6) based on the zone and the soil type. a graphical representation of the parameters in the spectrum and its distinct regions are presented in fig. 2. the lateral force ‘vi’ on the ith-story of the building is calculated using eqn. (7) and (8). 𝑣𝑣𝑖𝑖 = (𝑉𝑉 − 𝛥𝛥𝐹𝐹𝑁𝑁) 𝑤𝑤𝑖𝑖ℎ𝑖𝑖 ∑ 𝑤𝑤𝑖𝑖ℎ𝑖𝑖𝑁𝑁𝑖𝑖=1 (7) 𝛥𝛥𝐹𝐹𝑁𝑁 = 0.0075𝑁𝑁𝑉𝑉 (8) ‘wi’, ‘hi’ are the seismic weight and overall height of the ith-story, respectively. the seismic mass typically includes the total weight of the floor, plus half of the weight of the vertical elements (walls; columns) located immediately above and below that floor. an additional seismic force ‘∆fn’ should be considered to act at the top floor (roof-level) of the building to account for the contribution of higher modes of vibration. ‘∆fn’ is estimated using eqn. (8) and is based on the total number of stories in the building ‘n’ and the total base shear ‘v’. b. international building code 2006 (ibc) according to ibc, the base shear is calculated based on the natural period of the building ‘t’ and is given as: 𝑉𝑉 = 𝐶𝐶𝑠𝑠𝑊𝑊 (9) 𝐶𝐶𝑠𝑠 = ⎩ ⎨ ⎧𝑚𝑚𝑚𝑚𝑚𝑚 � 𝑆𝑆𝑆𝑆𝑆𝑆 𝑅𝑅 𝐼𝐼⁄ , 𝑆𝑆𝑆𝑆1 (𝑅𝑅 𝐼𝐼⁄ )𝑇𝑇 , 0.01� 𝑇𝑇 ≤ 𝑇𝑇𝐿𝐿 𝑆𝑆𝑆𝑆1 (𝑅𝑅 𝐼𝐼⁄ )𝑇𝑇2 𝑇𝑇𝐿𝐿 < 𝑇𝑇⎭ ⎬ ⎫ (10) figure 2. elastic response spectrum in omani seismic code (osc, 2013). where ‘w’ is the building’s seismic weight. ‘sds’ and ‘sd1’ represent design spectral response accelerations for short period (0.2-seconds) and a longer period (1.0second), respectively. these are based on spectral response accelerations ‘ss’ and ‘s1’ associated with the geographic location of the building and the soil conditions at the site using eqn. (11) and (12). 𝑆𝑆𝑆𝑆𝑆𝑆 = 2 3 𝐹𝐹𝑎𝑎𝑆𝑆𝑆𝑆 (11) 𝑆𝑆𝑆𝑆1 = 2 3 𝐹𝐹𝑣𝑣𝑆𝑆1 (12) ‘fa’ and ‘fv’ are based on the type of soil and seismic zonation. for this study, as per ibc, ss=0.25, s1=0.10 for zone-i. this gives a value of fa=1.20 and fv=1.70. in ibc, the natural period of a building can be estimated by: 𝑇𝑇 = 𝐶𝐶𝑡𝑡(𝐻𝐻𝑛𝑛)𝑥𝑥 (13) where ct=0.0466 and x=0.9 for moment resisting reinforced concrete structures, ‘hn’ represents the total height of the building. the response modification factor ‘r’ depends upon the ductility characteristics of the building and the importance factor ‘i’ depends upon the type of occupancy. i=1.0 for normal occupancy. the lateral force ‘vi’ for the ith-story is determined using eqn. (14): 𝑣𝑣𝑖𝑖 = 𝑉𝑉. 𝑤𝑤𝑖𝑖ℎ𝑖𝑖 𝑘𝑘 ∑ 𝑤𝑤𝑖𝑖ℎ𝑖𝑖 𝑘𝑘𝑛𝑛 𝑖𝑖=1 (14) ‘wi’, ‘hi’ are the seismic weight and overall height of the ith-story, in eqn. (14), the coefficient k=1.0 for structures with a fundamental period t ≤ 0.50 s, and k=2.0 for a fundamental period t ≥ 2.50 s. for structures having a period between 0.5 and 2.5 s, coefficient k shall be 2.0 or can be determined by linear interpolation between 1 and 2. c. uniform building code 1997 (ubc) in ubc, the calculation of base shear ‘v’ on a building is done using eqn. (15), which also defines upper and lower bound for the value: � 𝑉𝑉 = 𝐶𝐶𝑣𝑣𝐼𝐼𝑊𝑊 𝑅𝑅𝑇𝑇 δ𝑦𝑦 0.11𝐶𝐶𝑎𝑎𝐼𝐼𝑊𝑊 ≤ 𝑉𝑉 ≤ 2.5𝐶𝐶𝑎𝑎𝐼𝐼𝑊𝑊 𝑅𝑅 � (15) where ‘w’ is the seismic weight of the structure, ‘i’ is the importance factor that depends on occupancy and usage of the building, ‘r’ is ductility and overstrength factor that depends on the basic structural system, and the lateral-force resisting system of the building. ‘cv’ and ‘ca’ are seismic coefficients associated with the structure’s sensitivity to the velocity and acceleration comparison of oman seismic code with international counterparts based on regional seismic hazard 6 of ground motion, respectively. these are based on the geographic location of the structure (seismic zone) and soil conditions at the site. cv=0.13 and ca=0.09 are used for zone-1 and soil type-c. the over-strength factor ‘r’ depends upon the structural ductility and the importance factor i=1.0 for normal occupancy. in eqn. (15), the upper bound for base shear tends to govern for stiff structures while the lower bound is applicable for flexible structures. in ubc, the prominent natural period of a building with height ‘hn’ can be calculated as: 𝑇𝑇 = 𝐶𝐶𝑡𝑡(𝐻𝐻𝑛𝑛) 3 4 (16) ct=0.0731 is used to reinforce concrete momentresisting frames. the lateral force ‘vi’ for the ith-story of the building is estimated using eqn. (17): 𝑣𝑣𝑖𝑖 = (𝑉𝑉 − 𝐹𝐹𝑡𝑡) 𝑤𝑤𝑖𝑖ℎ𝑖𝑖 ∑ 𝑤𝑤𝑖𝑖ℎ𝑖𝑖𝑛𝑛𝑖𝑖=1 (17) ‘wi’, ‘hi’ are the seismic weight and overall height of the ith-story, respectively. ‘ft’ is an additional lateral force assumed to act at the top of the structure. this force is used to approximate the effect of higher modes of vibration. the magnitude of ‘ft’ is determined based on the natural period of the building ‘t’ and the base shear ‘v’, as given in eqn. (18). 𝐹𝐹𝑡𝑡 = � 0 𝑇𝑇 < 0.70𝑠𝑠𝑠𝑠𝑠𝑠 0.07𝑇𝑇𝑉𝑉 0.70 𝑠𝑠𝑠𝑠𝑠𝑠 ≤𝑇𝑇 ≤ 3.6𝑠𝑠𝑠𝑠𝑠𝑠 0.25𝑉𝑉 𝑇𝑇 > 3.6𝑠𝑠𝑠𝑠𝑠𝑠 � (18) 3.3 response spectrum method each code defines an elastic response spectrum for use in the response spectrum analysis. the spectrum is based on the relationships defined in eqn. (5), (10), and (15) for osc, ibc, and ubc, respectively. fig. 3 shows the comparison of the elastic response spectrum for the three codes considering the zone and soil type considered in the study. the comparison shows that the selection of zone definitions for international codes (ubc and ibc) are closely matched with zone-1 of osc. the spectrums as defined in fig. 3 for the respective code are employed in the response spectrum analysis considering the first 30 modes of vibration to ensure more than 99% modal participation for all building heights. direct spectral input using the parameters of this study is made in etabs for ubc and ibc codes, while the spectrum of osc is generated using eqn. (5). the base shear, story shear forces, and displacements are obtained employing modal response spectrum analysis in etabs. 3.4 structural ductility osc defines two ductility classes; the low ductility class (dcl) and the normal ductility class (dcn) based on the energy dissipation ability of the structure. if the structural design incorporates the detailing and structural requirements outlined in the osc code, the structure is classified as normal ductility class (dcn), otherwise, it is considered as the low ductility class (dcl). the behaviour factor ‘q’ in eqn. (3) is used to account for the ductility class. the ubc and ibc define ordinary moment resisting frames (omrf) and special moment resisting frames (smrf) using the same criterion. ubc uses over-strength and ductility factor ‘r’ and seismic force amplification factor ‘ωo’, while ibc uses response modification coefficient ‘r’, system over-strength factor ‘ω0’ and deflection amplification factor ‘cd’. the factor ‘r’ is employed to distinguish structural behaviour as given in eqn. (10) and (15) for ibc and ubc, respectively. the values of all these factors for the two classes are summarized in table 3. the subsequent discussion uses low ductility class to refer to dcl and omrf, and normal ductility class to indicate dcn and smrf. 4. results and discussion this section first compares the buildings considered in the study in terms of their prominent period and how this may affect their behaviour. the results for the equivalent static and the response spectrum analysis methods are then discussed in terms of base shear, lateral force distribution, and story drift ratios. figure 3. comparison of elastic response spectrum. table 3. behaviour/ductility factors for ductility classes. ductility class behaviour / ductility factor ibc ubc osc r ω0 cd r ωο q low ductility class (omrf/ dcl) 3 3 2.5 3.5 2.8 1.0 normal ductility class (smrf/ dcn) 8 3 5.5 8.5 2.8 3.5 muhammad bilal waris, khalifa al-jabri, and issa el-hussain 7 4.1. natural period and spectral acceleration table 4 lists the estimated natural periods for the four building heights based on osc, ibc, and ubc using eqn. (4), (13), and (16), respectively. rayleigh’s quotient method is used in etabs for the estimation of the natural period. the natural period in the two directions is nearly the same with the x-direction being marginally higher. figure 3 shows these values on the response spectrum, making it clear that the selected building heights cover a broad range of the spectrum. periods estimated using the codes are very close to the values using etabs for small building heights (10 m and 19 m) but underestimate the period for tall buildings (40 m and 58 m). it is because this study has not considered any shear walls that are common in tall buildings. as the natural period in the two directions is very close, the subsequent discussion only considers results for the x-direction. table 4 also lists the spectral acceleration to be used in the equivalent static load method for the respective building height and code. the spectral accelerations obtained from ubc are 94%-96% compared to osc, while values from ibc are 83% of osc for all the building heights. this is because both ibc and osc have ts=0.567s while for ubc, the value of ts=0.58s. therefore, the response spectrum of osc is quantitatively closer to ubc while qualitatively it resembles ibc. 4.2 base shear due to the difference in definition, the base shear calculated using the equivalent static load method or response spectrum analysis method cannot be directly compared among the codes. in osc, eqn. (1) or the spectrum is used to directly estimate the design seismic actions, while in ubc and ibc, the values are calculated using eqn. (15) and (10) must be amplified using ‘ωo’ and ‘ω0’, respectively. therefore, the comparison is presented in terms of the design base shear. a. equivalent static load method the comparison of the design base shear in x-direction for the two ductility classes using the equivalent static load method is summarized in fig. 4. since the natural period obtained from the detailed numerical model differs from the code provided estimate, additional code provisions must be considered. ubc recommends that the period should not exceed 1.4 times the ‘t’ as defined by eqn. (16), while ibc recommends a limit of 𝐶𝐶𝑢𝑢𝑇𝑇 for ‘t’ as defined in eqn. (13). the factor ‘cu’ depends upon the ‘sd1’ value and is estimated as 1.67 for this study. osc does not define any such limit and recommends using the value estimated by rayleigh’s quotient method. in this study, the base shear calculation using ubc for the 19-story building is the only case governed by this provision. for the low ductility class, the design base shear is nearly constant for 6-, 13and 19story buildings, while the value for the 3-story building is almost 2/3 of these values. this is due to the small value of ‘r’ (ubc, ibc) and ‘q’ (osc). the base shear value is governed by ‘cv’ for ubc, ‘cs’ for ibc, and ‘sar(t)’ in the case of osc. for the normal ductility class, a continuous increase in base shear with building height is observed. the base shear for the 3-story building is still nearly 2/3 (~67%) for ubc and ibc, but for osc it is 83% of the value for the 6-story building. the base shear increased by 15%, 5%, and 10% for 6-story to 13-story buildings in ibc, ubc, and osc respectively. for, 13-story to 19-story buildings, all the codes showed an increase of 47%. due to a large ‘r’ and ‘q’ factor, the base shear for 13and 19-story buildings are governed by the minimum limits in the respective codes, which causes the jump in base shear for the 19-story building due to larger seismic weight. (a) low ductility class. (b) normal ductility class. figure 4. design base shear using equivalent static method. table 4. prominent natural period of the buildings and spectral acceleration. model hn (m) natural period (s) spectral acceleration (g) ibc eqn. (13) ubc eqn. (16) osc eqn. (4) etabs ibc ubc osc x y 3-story 10.0 0.37 0.41 0.42 0.41 0.41 0.200 0.225 0.240 6-story 19.0 0.66 0.67 0.68 0.76 0.74 0.149 0.171 0.179 13-story 40.0 1.29 1.16 1.19 1.56 1.51 0.073 0.083 0.087 19-story 58.0 1.80 1.54 1.58 2.30 2.21 0.049 0.057 0.059 comparison of oman seismic code with international counterparts based on regional seismic hazard 8 as the spectral accelerations for the osc are higher than the other codes, its base shear values should be higher than ubc and ibc. this is true in fig. 4(a) for the low ductility class, where values of base shear from osc are higher for all the building heights. the values from ibc are nearly 82% of osc for the respective building height, which is the same as the ratio of spectral acceleration among the two codes. for ubc this value varies between 73%-76% for the 3to 13story buildings and is 81% for the 19-story building, which are lower compared to the ratio of spectral accelerations discussed earlier. for the normal ductility class in fig. 4(b), osc provides marginally smaller base shear except for the 3-story building, where values of ubc and ibc are nearly 85% of osc. for the 6story, both codes provide 10% higher estimates, while for the 13and 19-story, ibc and ubc estimates are 13% and 4% higher than osc, respectively. this is due to provisions in osc that relate ‘q’ to the member design through local ductility requirements thus ensuring better energy dissipation and higher reduction in base shear for longer periods. b. response spectrum method the base shear using modal response spectrum analysis for the three codes is shown in fig. 5. the values are lower compared to the equivalent static method for both the ductility classes. in the case of the low ductility class, the values range between 90%-94% for ibc, 84%-94% for ubc, and 88%-92% for osc compared to the equivalent static method. ubc shows the largest drop of 84% for the 19-story building because the base shear in the equivalent static method is controlled by the limit of 1.4t as defined by eqn. (16). the base shear using the response spectrum analysis gets smaller due to the increased contribution of higher modes that is overestimated by the equivalent static load method. for the normal ductility class, the ratio for the 3and 6-story values are the same as observed in the low ductility class. the ratios reduced to 81%, 88%, and 85% for ibc, ubc, and osc, respectively, for 13-story buildings, and 55%, 60%, and 57% for ibc, ubc, and osc, respectively, for the 19-story building. this is because base shear in the equivalent static method is governed by the minimum limit rather than the response spectrum values for the 19-story building. all the codes (osc, ibc, ubc) recommend that the values of base shear calculated using modal response spectrum method (figure 5) should not be less than a certain fraction of the value calculated using equivalent static method (figure 4). the fraction is 0.85 for osc and ibc, and 0.90 for ubc. as per osc, if the ratio is not satisfied, the response quantities (displacement and internal forces) should be scaled up corresponding to this factor. a lighter shade is used in fig. 5 to extend the bars to this limit as required. this limit has a significant influence on the 19-story building for the normal ductility class only. for the low ductility class, the factor is 1.07 with ubc for the 19-story building. in the normal ductility class, for 13-story building factors are 1.05, 1.02, and 1.003, and for the 19-story building, the factors are 1.53, 1.49, and 1.49 for ibc, ubc, and osc, respectively. 4.3. lateral force distribution figure 6 presents the story shear force distribution calculated using eqn. (7), (14), and (17) for osc, ibc, and ubc, respectively. the values from eqn. (14) and (17) are scaled using ‘ω0’ and ‘ωo’ factors, respectively, as listed in table 3. to facilitate the comparison among building heights, the vertical axis is normalized using the total building height ‘hn’, which is 10 m, 19 m, 40 m, and 58 m for the 3-, 6-, 13and 19-story buildings, respectively. since the story force distribution for the three codes does not consider ductility classes (eqn. (7), (14), and (17)), the distributions only differ in terms of values. for the 3story building, the three codes have almost linear force distribution, which indicates that the contribution of higher modes is negligible. the lateral force from ubc and ibc for intermediate stories has the same constant ratio with the osc counterpart as observed for base shear, while the ratio is different for the top story (roof). this is due to the similar story shear distribution model in osc (eqn. (7)] and ubc [eqn. (17)), while for ibc (eqn. (14)), the value of ‘k’ is 1.0 for the 3story building. for the 6-story building, only a slight deviation is observed for the top story in the case of osc and ubc, which indicates that for low-rise buildings ℎ𝑛𝑛 ≤ 20𝑚𝑚, the additional force at the top story is insignificant in both the codes. for ibc, the ratios of individual story shear force relative to osc increase with story height reflecting the nonlinear distribution (𝑘𝑘 > 1.0) being utilized in ibc (eqn. (14)) in contrast to the linear pattern of osc and ubc. for the 13-story building (fig. 6), an additional force on the top floor to account for higher modes is observed in ubc and osc, while the distribution for ibc becomes nonlinear to account for higher modes. compared to osc, the intermediate story force in ubc is 74% and 103%, while for the roof level force is 79% and 110% in low and normal ductility classes, respectively. for ibc, the shear force for intermediate floors increases with an increase in building height but shows a drop at the top due to the additional force (eqn. (8) considered in osc. the trend for the 19-story building is the same as the 13-story but the ratio of the lateral force from ubc is 84% and 108% of the osc values for the roof and 81% and 103% for other floors in low and normal ductility classes, respectively. this indicates that both osc and ubc account for the higher modes using a similar approach, but osc has a smaller allowance for this additional force (eqn. (8)) compared to ubc (eqn. (18)). the allowance in osc further does not follow the same trend as ubc with an increase in building height. for the response spectrum method, the values as obtained from etabs are presented in fig. 7, after scaling the values for ibc and ubc using ‘ω0’ and ‘ωo’ factors as listed in table 3. muhammad bilal waris, khalifa al-jabri, and issa el-hussain 9 (a) low ductility class (b) normal ductility class figure 5. design base shear using response spectrum method. (a) low ductility class (b) normal ductility class figure 6. shear force distribution along building height using equivalent static method. the values are not corrected for the limit discussed previously, since it only applies to internal forces. for the 3and 6-story buildings, the distribution is almost linear as observed in the case of the equivalent static method. this confirms the assumption of minimal contribution from higher modes in all the codes. osc, therefore, rightly recommends only using the equivalent static method for low rise buildings (ℎ𝑛𝑛 ≤ 20𝑚𝑚) that satisfy the structural regularity criterion outlined in the code. for 13and 19story buildings, the distribution is similar for the three codes having a clear nonlinear distribution, therefore using the response spectrum method is more appropriate and is correctly made mandatory by osc for medium and high-rise buildings (20 𝑚𝑚 < ℎ𝑛𝑛 ≤ 60𝑚𝑚). 4.4 story drift ratio as discussed in the previous section, ibc, ubc, and osc use different factors to account for ductility and inelastic behaviour of the buildings as listed in table 3. this study compares the story drift ratio defined as the relative displacement of a story level/floor compared to the floor immediately below divided by the inter-story height as given in eqn. (19). 𝛿𝛿𝑖𝑖 = 𝛥𝛥𝑖𝑖 − 𝛥𝛥𝑖𝑖−1 ℎ𝑖𝑖 (19) where ‘𝛥𝛥𝑖𝑖’, ‘𝛥𝛥𝑖𝑖−1’ are displacements of the ith story and the story immediately below it and ‘ℎ𝑖𝑖’ is the height of the ith-story. the story drift ratios need to be modified using ‘cd/i’, ‘0.7r’, and ‘q/i’ for ibc, ubc, and osc, respectively. for the low ductility class, these factors are 2.5, 2.45, and 1.0, and for the normal ductility class, these are 5.5, 5.95, and 3.5 for ibc, ubc, and osc, respectively. the distribution of story drift ratios for the equivalent static method is shown in fig. 8. the permissible limit set by the three codes is 2.0%, which is satisfied for all building heights. it can be observed that ibc and ubc provide very similar results in all the cases, while osc provides higher values. comparing the maximum drift ratio, for the low ductility class, the ratio of ibc/osc is a constant 69% for all building heights while ubc/osc ratio is 66% for 3-,6and 13-story and 71% for the 19-story building. for normal ductility class, ibc/osc ratio is 45%, 57%, 60% and 60%, and ubc/osc ratio is 52%, 67%, 63% and 63% for 3-,6-,13and 19-story buildings, respectively. it is further observed that the drift ratios for the two ductility classes (low/normal) have a constant ratio for all stories in a building for a given code. these ratios are summarized in table 5 for the two analysis methods. for the equivalent static method, this ratio stays constant for low-rise buildings in ibc and ubc at 1.21 and 1.00, respectively, and then decreases with the building height. this drop is because the design base shear is governed by minimum base shear criteria for these heights in the normal ductility class. comparison of oman seismic code with international counterparts based on regional seismic hazard 10 (a) low ductility class (b) normal ductility class figure 7. shear force distribution along building height using response spectrum method. (a) low ductility class (c) normal ductility class figure 8. story drift distribution along building height using equivalent static method. (a) low ductility class (b) normal ductility class figure 9. story drift distribution along building height using response spectrum method table 5. comparison of story drift ratio among ductility classes (low/normal) building equivalent load method response spectrum method ibc ubc osc ibc ubc osc story-03 1.21 1.00 0.80 1.21 1.00 0.80 story-06 1.21 1.00 1.00 1.00 story-13 1.09 0.98 0.95 story-19 0.73 0.72 0.64 in osc, the ratio first increases from 0.8 to 1.0 from 3-story to 6-story building and then start to decrease like ibc and ubc but at a higher rate. this is because osc uses a variable value of 𝑞𝑞𝐴𝐴(𝑇𝑇) for 𝑇𝑇 < 𝑇𝑇𝑆𝑆 as listed in eqn. (3) and for the 3-story building 𝑞𝑞𝐴𝐴(𝑇𝑇) = 0.80𝑞𝑞. figure 9 shows the drift ratios using the response spectrum method. the response spectrum method provides similar drift values for the 3and 6-story buildings, while it has smaller values for 13and 19story buildings, due to the smaller shear forces. the comparison of the two ductility classes remains the same as discussed for the equivalent static method, which further extends to 13and 19-story buildings, as listed in table 5. this verifies that the difference observed among the ductility classes in the equivalent static method for the 19-story building is due to the muhammad bilal waris, khalifa al-jabri, and issa el-hussain 11 minimum base shear criteria. the drift ratios in osc are higher because the same factor ‘q’ is used for response modification and deflection amplification (𝑞𝑞/𝑞𝑞 =1). on the other hand, in ibc the 𝐶𝐶𝑑𝑑/𝑅𝑅 ratio is 0.833 and 0.688 for low and normal ductility classes, respectively, whereas in ubc the factor is 0.70 for both ductility classes. therefore, the drift ratio estimated using osc will always be higher compared to its counterparts. since osc uses the same allowable limit of 2%, it will require a stiffer design compared to ibc and ubc to satisfy this limit. a comparison of fig. 8 and fig. 9 shows that the equivalent static load method provides good agreement of drift ratios with the response spectrum method for 3 and 6-story buildings but higher estimates for 13and 19-story buildings. this means that using an equivalent static method for medium and high-rise buildings will lead to stiffer designs in all the codes. 5. conclusion the study presented a comparison of the oman seismic code (osc) with international codes, uniform building code-1997 (ubc) and international building code-2006 (ibc), based on the equivalence of seismic hazard. base shear, lateral force distribution, and storydrift ratios are compared using the equivalent static method and the response spectrum method. four building heights of 10 m, 19 m, 40 m, and 58 m are considered with 3-, 6-,13and 19stories, respectively. both low and normal ductility classes identified in osc are compared and discussed. the following observations are made: • the high seismic zone of osc (zone-1) is wellmatched with zone-1 of ubc and ibc in terms of seismic hazard. • for the low ductility class, the base shear values for ibc are 83%-85% compared to osc, using both the equivalent static load and response spectrum methods. • for the low ductility class, the base shear values for ubc are 73%-81% and 75%-78% compared to osc using the equivalent static load and the response spectrum methods, respectively. • for the normal ductility class using the equivalent static method, the base shear value for ibc is 86%,109%, 113%, and 113% of osc for 3-, 6-,13, and 19story buildings, respectively. the base shear value for ubc is 85%,110%, 104%, and 104% of osc for 3-, 6-,13and 19story buildings, respectively. • for the normal ductility class using the response spectrum method, the base shear value for ibc and ubc is 87% of osc for the 3-story building and 110% of osc for the other building heights. • the story shear force distribution in the equivalent static method for osc is similar to ubc, while it considers a marginally smaller allowance for the additional force at the top. • the story shear force distribution in the 13and 19story buildings using the response spectrum method showed the clear contribution of higher modes. • story drift ratios from the response spectrum method indicate that the two ductility classes in all codes provide the same deformation of the structure while the normal ductility class targets better energy dissipation through proper detailing. • the use of the response spectrum method for medium and high-rise buildings suggested by osc will lead to a better economy in design due to lower values of loads and smaller drift ratio estimates that control the stiffness of the structure. the study concludes that the zone-1 of osc is closely matched to zone-1 of ubc and ibc contrary to the recommendation of ubc that defines muscat (zone-1 of osc) as zone-2a. acknowledgement this research is not funded by any organization; however, the authors appreciate the use of computation facilities at the sultan qaboos university (squ) and thank computer and structures inc. (csi) for allowing the use of etabs for non-funded research. conflict of interest the authors declare no conflict of interest. funding no funding was received for this research references deif a, el-hussain i, al-jabri k, toksoz n, el-hady s, al-hashmi s, al-toubi k, al-shijbi y, alsaifi m (2013), deterministic seismic hazard assessment for sultanate of oman. arabian journal of geosciences 6(12): 4947-4960. el-hussain i, deif a, al-jabri k, toksoz n, el-hady s, al-hashmi s, al-toubi k, al-shijbi y, alsaifi m, kuleli s (2012), probabilistic seismic hazard maps for sultanate of oman. natural hazard 64: 173–210. el-hussain i, deif a, al-jabri k, al-hashmi s, altoubi k, al-shijbi y, al-saifi m (2014), seismic hazard assessment for sultanate of oman”, technical report, project # 22409017, earthquake monitoring center, sultan qaboos university. el-hussain i, deif a, al-jabri k, mohamed a. m. e, al-rawas g, toksöz m. n, sundararajan n, elhady s, al-hashmi s, al-toubi k, al-saifi m, al-habsi z (2013), seismic micro-zonation for muscat region, sultanate of oman, natural hazards 69: 1919–1950. emc (2003-2017), archives of seismological bulletin of oman’s seismological network (2003-2017), earthquake monitoring center, comparison of oman seismic code with international counterparts based on regional seismic hazard 12 sultan qaboos university. http://www .squ.edu.om/emc/bulletins. en 1992-1-1 (2004): eurocode 2 – design of concrete structures – part 1-1: general rules and rules for buildings, european committee for standardization. en 1993-1-1 (2005): eurocode 3 – design of steel structures – part 1-1: general rules and rules for buildings, european committee for standardization. en 1994-1-1 (2004): eurocode 4 – design of composite steel and concrete structures – part 11: general rules and rules for buildings, european committee for standardization. en 1996-1-1 (2005): eurocode 6 – design of masonry structures – part 1-1: general rules for reinforced and unreinforced masonry structures, european committee for standardization. etabs-16, structural analysis software, computer and structures inc, berkeley, california. fenwick r, lau d, davidson b (2002), comparison of the seismic design requirements in the new zealand loadings standard with other major design codes. bulletin of the new zealand society for earthquake engineering 35(3): 190203.international building code (2006), international code council inc, falls church, va, usa. marino em, nakashima m, mosalam km (2005), comparison of european and japanese seismic design of steel building structure, engineering structures 27(6): 827-840. mcintosh rd, pezeshk s (1997), comparison of recent u.s. seismic code. journal of structural engineering 1997-8: 993-1000. nahhas tm (2011), a comparison of ibc with 1997 ubc for modal response spectrum analysis in standard-occupancy buildings. earthquake engineering and engineering vibration 10: 99113. noor ma, ansari ma, seraj sm (1997), critical evaluation and comparison of different seismic code provisions, journal of civil engineering ( institution of engineers, bangladesh) 25(1): 1-20. oman seismic design code for buildings (2013), earthquake monitoring center, sultan qaboos university. qamaruddin m, al-harthy a (2000), “earthquake hazard potential in oman” in proc. 12wcee, auckland, new zealand. uniform building code (1997), international conference of building officials, whittier, california, usa. waris m.b, saeed h, al-jabri k, el-hussain i (2017), “comparison of omani seismic code for buildings with international counterparts” in proc. 16wcee, santiago, chile. 1. introduction 2. oman seismic code for buildings 3. methodology 3.1 structural system 3.2 equivalent static load method a. omani seismic code (osc) b. international building code 2006 (ibc) c. uniform building code 1997 (ubc) 3.3 response spectrum method 3.4 structural ductility 4. results and discussion 4.1. natural period and spectral acceleration 4.2 base shear a. equivalent static load method b. response spectrum method 4.3. lateral force distribution 4.4 story drift ratio 5. conclusion acknowledgement conflict of interest funding references the journal of engineering research (tjer), vol. 17, no. 1, (2020) 24-33 doi:10.24200/tjer.vol17iss1pp24-33  *corresponding author’s e-mail: abhossen@squ.edu.om identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition of the hrv signal sarah qasim ali and abdulnasir hossen * department of electrical and computer engineering, college of engineering, sultan qaboos university oman, p. o. box 33, al-khoudh, 123 muscat, oman. abstract: the advancement of telecommunication technologies has provided us with new promising alternatives for remote diagnosis and possible treatment suggestions for patients of diverse health disorders, among which is the ability to identify obstructive sleep apnea (osa) syndrome by means of electrocardiograph (ecg) signal analysis. in this paper, the standard spectral bands’ powers and statistical interval-based parameters of the heart rate variability (hrv) signal were considered as a form of features for classifying the sultan qaboos university hospital (squh) database for osa syndrome into 4 different levels. wavelet packet analysis was applied to obtain and estimate the standard frequency bands of the hrv signal. further, the single perceptron neural network, the feedforward with back-propagation neural network and the probabilistic neural network have been implemented in the classification task. the classification between normal subjects versus severe osa patients achieved 95% accuracy with the probabilistic neural network. while the classification between normal subjects versus mild osa subjects reached accuracy of 95% also. when grouping mild, moderate and severe osa subjects in one group compared to normal subjects as a second group, the classification with the feedforward network achieved an accuracy of 87.5%. finally, when classifying subjects directly into one of the four classes (normal or mild or moderate or severe), a 77.5% accuracy was achieved with the feedforward network. keywords: sleep apnea; identification; classification; hrv; wavelet packet decomposition; neural networks. تحديد انقطاع التنفس االنسدادي أثناء النوم باستخدام الشبكات العصبية االصطناعية hrv إلشارةوتحليل حزمة األطوال الموجية *وعبدالناصر حسين سارة قاسم علي أتاح لنا التقدم العلمي في مجال تقنيات االتصال بدائال جديدة واعدة للتشخيص عن بعد واقتراحات عالجية الملخص: محتملة لمرضى مختلف االضطرابات الصحية. ومن ضمن هذه التقنيات القدرة على كشف مرضى متالزمة انقطاع التنفس النطاقات ). اعتبرت في هذه الدراسة قوىecgادي أثناء النوم عن طريق تحليل إشارة التخطيط الكهربائي للقلب (االنسد ) سمة من سمات hrvالطيفية القياسية واإلحصائيات المقسمة على فترات زمنية إلشارة تغير معدل نبضات القلب ( تصنيف وتقسيم قواعد بيانات مستشفى جامعة السلطان قابوس المتعلقة بمتالزمة انقطاع التنفس االنسدادي الى اربع اللتي بنيت عليها الدراسة. وقد hrvتقدير نطاقات التردد القياسية إلشارة مستويات. تم إستخدام حزمة األطوال الموجية ل تم تصميم كل من الشبكة العصبية الحسية الوحيدة، و الشبكة العصبية ذات التغذية األمامية واالنتشار الخلفي، والشبكة حاء والمرضى من ذوي األعراض % في التمييز بين األص95العصبية اإلحتمالية في مهمة التصنيف. حقق اإلختبار دقة % في التفرقة بين األشخاص األصحاء واألشخاص المصابين 95الحادة باستخدام نموذج الشبكة العصبية اإلحتمالية, ودقة بالمتالزمة ويعانون من أعراض بسيطة. أما عند تشكيل عينة تشمل مرضى يعانون من أعراض مختلفة الشدة (بسيطة أخرى من أشخاص أصحاء فإن نموذج الشبكة العصبية ذات التغذية األمامية واإلنتشار الخلفي ومتوسطة وشديدة) وعينة واإلنتشار األمامية التغذية ذات %, وتمكن نموذج الشبكة العصبية87.5تمكنت من التفريق بين العينتين بدقة بلغت .%77.5من التمييز بين األنواع األربعة بدقة بلغت الخلفي نقطاع التنفس أثناء النوم؛ الكشف؛ التصنيف؛ معدل تغير نبضات القلب؛ تحليل حزمة امتالزمة مفتاحية:الت كلماال األطوال الموجية؛ الشبكات العصبية اإلصطناعية. identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition 25  1. introduction a sleep disorder occurs when the pattern of sleep is interrupted repeatedly during sleep (kumar v 2008). lack of sleep results in abnormalities in functions of the brain leading to cognitive impairment, changes in mood, low productivity, daytime sleepiness and abnormal hormonal rhythms (wilson s 2016). sleep apnea is a chronic disease that affects the health and productivity of individuals (national heart, lung and blood institute 2016) since it causes abnormal sleep pattern. obstructive sleep apnea (osa) is the most common type of sleep apnea followed by central sleep apnea (csa) and mixed sleep apnea (msa). osa affects 3~4% and 2% of middle-aged men and women respectively (lee w et al. 2008). unlike csa which results from heart failure or brain disorders, where the brain fails to control breathing leading to cessation of all respiratory airflow and movements, osa results from a repeated process of complete or partial collapse in the upper airways of the respiratory system ranging from few seconds (minimum 10 sec) to minutes despite the ongoing brain efforts for the body to breath. the osa events may occur more than 30 times and up to 100 times per hour. msa on the other hand, is a mixture of csa and osa in the same individual. (american academy of sleep medicine 2001; national sleep foundation 2016). osa has been related to some serious co-morbidity such as cardiovascular diseases, arrhythmia, strokes, obesity, depression, certain types of hypertension and type 2 diabetes mellitus (global leaders in sleep and respiratory medicine 2013; xie w et al. 2014). there are several screening methods used for osa detection to find evidence of its presence in patients for further evaluation. these methods depend on psychometric and physical evaluations during the routine health check-ups. polysomnography (psg) sleep study is the gold standard test for sleep apnea diagnosis. this test requires the patient to sleep in a sleep laboratory while attached using several electrodes to many devices for different biometric measures carried out by qualified sleep physicians overnight. the severity of sleep apnea is commonly determined by an apneahypopnea index (ahi) which represents the number of obstructive, central, mixed and hypopnea episodes occurring during an hour of sleep (american academy of sleep medicine 2001). if the ahi ranges between 0-5 apneic episodes during an hour of study time or sleep then the condition is considered normal. an index of 5-15 is considered mild while an index of 15 – 30 is considered moderate and if the index is 30 or above, the subject is considered to have a severe degree of sleep apnea (global leaders in sleep and respiratory medicine 2013). electrocardiography (ecg) is a method used to measures the electrical activity of the heart by placing electrodes on different parts of the body (webmd 2016). a normal sinus rhythm reflects the normal activity of the heart while pumping blood to perform the sympathetic and parasympathetic activities (ucdavis health system 2016). a typical ecg signal is produced when the heart chambers contract and expand to pump oxygenated-blood throughout the body and circulate the desaturated blood to the lungs. fig. 1 (a) shows a typical ecg signal (sharma s. et al. 2019). heart rate (hr) is a simple measurement that indicates the average number of heart beats during a certain time period (usually, a minute). a low hr reflects resting status while high hr indicates stress or exertion (moore j 2016). heart rate variability (hrv) on the other hand is a measure of the time variability in milliseconds between consecutive beats or correspondingly in the instantaneous hr. in other words, variation analysis of instantaneous hr versus time axis. hrv is sometimes called the r-r interval (rri) analysis, where r is the peak point of the qrs complex in the ecg wave, or the inter-beat-interval (ibi) analysis. when the individual is at rest, high hrv is favorable while low hrv is observed at an active or stressed state. hrv has been used as a measurement to assess overall cardiac health and reflect the state of the autonomic nervous system (ans) activities (hamilton g. et al. 2019). the ans is the involuntary division of the nervous system and consists of autonomic neurons that conduct impulses from the central nervous system (brain and/or spinal cord) to glands, smooth muscles and cardiac muscles (dantest clinicians team 2016). the role of the ans is to continuously fine-tune the functions of organs and organs systems to maintain internal stability and balance. ans has two main components called the sympathetic and parasympathetic nervous systems (sns and psns respectively). the sns triggers the fight or flight response leading to increased heart rate, blood pressure and sweating, and pupil dilation etc. on the other hand, the pns complements the operations (a) (b) figure 1: (a) normal ecg signal. (b) ecg signal at apnea episode. sarah qasim ali and abdulnasir hossen 26 performed by the sns and triggers the rest and digest response where the opposite behavior occurs. when the airway is partially or completely obstructed; the heart rate changes and hence ecg signal alters. when the oxygen level decreases in the body during sleep apnea event the heart cells receives less oxygen and hence the heart rate is reduced and the r-r interval increases between consecutive beats as shown in fig.1 (b) (sharma s. et al. 2019). this alters the brain's sleep and wakes the brain for immediate action. the brain responds by sending strong tones to the respiratory system to increase breathing speed. the later increases the heart rate suddenly and hence increases the blood pressure in order to pump more blood to compensate for the lack of oxygen. in frequency-domain analysis, signals can be represented in a graph that shows how much (energy) of the signal lies within given frequency bands over a range of frequencies. the well-known fourier methods such as the fast fourier transform (fft) implementation of the discrete fourier transform (dft) are usually used for identifying the available spectral content for both stationary and non-stationary signals (polikar r 2011). however, for non-stationary signals, the frequency content varies with time and hence dft based methods fail to provide the timerelated information at which those frequencies occur. moreover, it can only reflect the frequencies that are present in the signal but not when they were present. since most of the physiological signals like the ecg, etc. are non-stationary signals; time-frequency analysis such as the short-time fourier transform (stft) and discrete wavelet transform (dwt) are used as alternatives to the fourier analysis when estimating the available psd content (polikar r 2011). the classification features used in this research depend on the power spectral density (psd) at different frequency levels estimated by implementing the discrete wavelet packet decomposition (dwpd) method. this is to overcome the resolution related problems of the stft. the discrete wavelet decomposition utilizes various mother wavelets of different scales to be able to adapt to fast and slow changes in the analyzed signal (polikar r 2011). the wavelet decomposition method is implemented using filter banks (misiti m et al. 1996). a set of high pass and low pass filters allow the signal to be decomposed reaching a certain decomposition level in which the signal can be further analyzed, de-noised or compressed (misiti m et al. 1996). the psd calculation is done by mathematical modulation to the filters output coefficients (sysel p et al. 2008). the dwt allows only the low-frequency components of the low pass filter to be analyzed to further levels as they are thought to be the ones that carry important information (misiti m et al. 1996). however, dwpd allows both outcomes of the filters (low and highfrequency) to be further decomposed. the later emphasizes the outliers, edges and transient signals which are crucial to tackle the osa episodes. hence, the dwpd leads to finding more desirable features for classification applications. in time domain analysis, a signal instance’s real values are visualized. a time domain graph shows how a signal changes with time. osa episodes can be analyzed by observing the cyclic length variability of the hrv (a.k.a. rri) signal. the term nn is used sometimes in place of the rr to emphasize that normal beats are being processed. there exist multiple well-established features that are normally used to analyze the beat-to-beat intervals (mietus j et al. 2002). these features include: the root mean square of successive differences (rmssd), the standard deviation of successive differences (sdsd), the standard deviation (std.) of entire rri signal, the mean of the entire rri signal, the nnx family measures which include (nn of x≤50: nn50, nn30, nn20…etc.), and finally the pnnx family measure. 2. ecg database it has been suggested by some researchers, that the uniqueness of the data sets affects the classification results of the different proposed methods, hence similar results cannot be obtained using different datasets (lado m et al. 2011). in this research the ecg signals were collected from the sleep laboratory of the physiology department of the sultan qaboos university hospital (squh) while performing psg studies for 80 subjects. these records were obtained from 20 normal subjects (0<ahi<5), 20 mild subjects (5<ahi<15), 20 moderate subjects (15<ahi<30) and 20 severe subjects (ahi≥30). the records were divided into two groups: a training set and a test set; where both of the sets comprise of 10 normal ecg signals and 10 apneic ecg signals from each of the mentioned groups (total of 40 signals per set). features were extracted from both sets and the training set is used to train the neural networks for classification purpose while the test set was used to check the classification performance. 3. rri extraction the hrv signal is generated by finding the r-to-r intervals (rri) from the original ecg signals as declared in fig. 1 (a). the proposed method in (al ghunaimi b 2003) was used to generate the rri data. in order to accurately identify those r peaks, qrs detection is to be carried. the qrs detector that was used in (al ghunaimi b 2003) is a part of the physionet tools available in the physionet website and is based on the pan-tompkins algorithm. intervals corresponding to normal-to-normal peaks are extracted. the generated rri data could contain false intervals, missed intervals and/or ectopic intervals. identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition 27  false rr intervals were removed by setting the lower and upper limits of its values to ones found in normal subjects which are typically around 400-2000 milliseconds. removing of outliers was achieved using a 41-points moving average filter (maf). resampling at 1 hz and substituting of missed peaks were then achieved by simple linear interpolation implemented by matlab (al ghunaimi b 2003). re-sampling and estimation of missed value are intended to generate an equally spaced rri data and preserve the temporal sequence that is necessary for the frequency domain analysis. at this point it can be assumed that a clean rri data sampled at 1 hz containing no missed or outlier values was generated. 4. spectral features in this research, the rri signals were decomposed using discrete wavelet packet decomposition up to 9 levels in order to define the vlf band precisely according to the standard definition (vlf starts at 0.0033 hz). at the ninth level, the signal would have been decomposed into 512 frequency bands including the low-frequency and high-frequency bands.  the psd summation of these bands provides the total power spectrum of the signal, while the psd summation by grouping according to certain frequency bands would provide us the desired features for the analysis. the discrete rri signals are sampled at fs=1 hz and the maximum spectral frequency of rri is found by fs/2=0.5 hz. using dwpd, the 0.5 hz is decomposed into 512 bands (9 levels = 29), while each band covers 0.5/512= 0.0009765625 hz. therefore, the standard spectral bands of the decomposed rri are covered as shown in fig. 2. in addition, the spectral values near zero hz have the most energy content that dominates other spectral values. therefore, we intend to exclude those spectral values below 0.0033 hz; hence, the first three bands of the decomposed rri signal were ignored and the psd of the vlf was estimated starting from the 4thband at 0.00390625 hz up to 0.040039063 hz.  fig. 3 shows the spectral powers summation of the different bands defining each feature. furthermore, three other features extracted from the power ratios of the vlf, lf and hf features are calculated to form a total of six spectral features. these ratios are described as in fig. 4. in fig. 5, the power of the hf band is sketched for both normal and severe subjects for the training set. the values of the hf band at normal subjects are higher than those of severe subjects. however, some severe subjects have powers that overlap with normal subjects resulting in difficulties to classify and hence a combination of features are to be investigated (the university of nottingham 2017). 5. statistical features in this work, multiple features were calculated from the rri signals in time domain using matlab software. these features include the root mean square of successive differences (rmssd) which is calculated by finding the square root of the mean of the successive differences between adjacent nns. the standard deviation of successive differences (sdsd) where the standard deviation of average nn intervals (sdann) is calculated over a short period of 5-minutes or so to measures the change in heart rate due to cycles longer than five minutes. the standard deviation (std.) of entire rri signal, the mean of the entire rri signal, the nnx family measures which include (nn of x≤50: nn50, nn30, nn20 etc.), where it represents the number of pairs of successive nn’s that differ by more than x=50, 30, 20…etc. milliseconds and finally the pnnx family measure which includes the nnx measure divided by the total number of nns of the signal. to sum up, the features are: the rmssd, pnn15, pnn20, pnn30, pnn50, nn15, nn20, nn30, nn50, sdann and standard deviation of the entire signal which equate to eleven features. in order to observe a relation between time domain and frequency domain features, the values of rmssd, nn50, and pnn50 features were normalized by dividing them to the total power of the rri signal as shown in fig. 6. similar behavior of fig. 5 was observed where the power of the highfrequency band for severe apnea level subjects decrease and the same features’ values increase for normal subjects. this allows us to use these features in a similar analogy as spectral features for representing the parasympathetic activity of the autonomous nervous system for instance. of course, variations are also present in some of the osa patients where their spectral or time domain features exhibit similar behavior as normal subjects making it more difficult to differentiate between the cases using a single feature. 6. neural networks the artificial neural networks (anns) were designed based on the rudimentary understanding of the biological nervous system back in the 1950s to help solving and computing any arithmetic or logical statement (hagan m et al. 1996). the anns are a collection of computational units called neurons composed of inputs with weights, biases and transfer functions to perform the thresholding act and produce an output. the three networks used in this work are of supervised learning type; this is because we already have information about the subject’s original condition (normal, mild, moderate, and severe). in supervised learning, the training set consists of sarah qasim ali and abdulnasir hossen 28  figure 2. frequency bands of rri packet decomposition into 9 levels. figure 3. calculation of the three main features. figure 4. main three spectral ratio features. figure 5. power spectral destiny (hf band) between normal and sever subjects. figure 6. normalized pnn50 % values for normal and severe subjects. identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition 29  100.(%) fntnfptp tntp accuracy    100.(%) fntp tp ysensitivit   100.(%) fptn tn yspecificit   multiple training examples each is a pair of input (feature) and the desired output (target). the supervised learning algorithm then tries to process and analyze the training examples to produce an inferred function (classification boundary region), which can be used to map new examples. if the network training and the learning algorithm are performed well, an optimal scenario would be generated in which the network will generalize and be able to determine the class label (target) of unseen instances correctly (mohri m et al. 2012). this section introduces the three artificial neural networks’ models used in this research and describes their learning algorithms and transfer functions briefly. the perceptron network refers to multiple-inputs single-neuron network. it is considered the building block of a more complicated network called the feed forward neural network. usually the single-layer perceptron is used for binary and linear classification. the feed forward network is one of the most commonly used artificial neural networks and is typically composed of multiple perceptrons aligned in layers. it is considered the first type of ann’s that was used to solve non-linear problems where the perceptron has failed and is considered powerful networks that can almost approximate any function (hagan m et al. 1996). probabilistic neural networks (pnn’s) are widely used classification and pattern recognition. when employed for classification problems, the class probability of input is estimated and the class with the highest probability is selected as the output class. this means that the input belongs to the class that provides the highest probability when introduced to that input. the design of a pnn is straightforward and extremely less complicated than other multilayer networks. this is because it does not depend on weights learning and hence does not require training. a pnn network generalizes well and is guaranteed to converge to a bayesian classifier (simple well-studied conditional probabilistic classifier) providing the correct probability when presented with enough training data samples. a pnn consists of several sub-networks, each of which is a parzen window pdf estimator for each of the classes. the inputs are the set of measurements/features and are used as centers for the radial basis (gaussian functions) of the second layer. the third layer performs an average operation of the outputs from the second layer for each class and a final voting is performed by the third layer selecting the largest value to determine the associated class label. in this research, the perceptron network is composed of one neuron of multiple inputs for one output classification, and two neurons for four outputs classification embedded with hard limit transfer function. the feedforward network is composed of two hidden layers each of five neurons (total 10) of the tangent-sigmoid transfer function and an output layer of one or two neurons according to the number of outputs (similar to perceptron layer) with pure linear transfer function. the probabilistic network is composed of input neurons changing according to the number of inputs with radial basis transfer function. figs. 7 and 8 show the connections of the layers for the feed forward and probabilistic neural networks. the performance of the training step and testing step of the neural networks are investigated by calculating well-known performance metrics such as the specificity, sensitivity and accuracy. the specificity reflects the number of accurately diagnosed healthy subjects while the sensitivity reflects the number of the accurately identified patients. the accuracy is a measure of both of the correctly classified patients and normal among the total experiment set. in fig. 9, the actual meaning for each performance metric can be observed while the following equations reveal how they are calculated. figure 7. feedforward neural network layer connections. figure 8. probabilistic neural network layer connections. (1) (2) (3) sarah qasim ali and abdulnasir hossen 30  figure 9. networks performance metrics. 7. implementation in this work, four different versions of classification have been carried out using the perceptron network, the feedforward with back-propagation network and the probabilistic neural network. these classification versions include: • version 1: normal vs. severe classification • version 2: normal vs. mild classification • version 3: normal vs. patient (mild, moderate, severe) classification • version 4: normal vs. mild vs. moderate vs. severe classification. the discrete wavelet packet decomposition was implemented to extract power estimations of the rri signals of squh ecg database at the classical frequency bands with bi-orthogonal mother wavelet at 9 levels decomposition. four different schemes of training and test data are selected:  scheme 1: original first half data for trial: original second half data for the test.  scheme 2: original second half data for trial: original first half data for the test.  scheme 3: large set simulated from first-half data for trial: original second half data for test.  scheme 4: large set simulated from second-half data for trial: original first half data for test. the networks in scheme 1 and 2 are trained by the features of original data sets containing 10 subjects for each of the normal, mild, moderate and severe osa states. while in scheme 3 and 4, the networks are trained using a large simulated set generated from the original data sets. the large training set was acquired by generating randomly, 1000 uniformly distributed feature sets between the maximum and minimum values of each feature from the original extracted spectral features. the networks testing and performance computation were carried by the remaining equivalent-size original data set of 10 subjects in each case. the statistical time-domain features are used with classification version number 4 only as it is the most complicated classification version and hence investigation with different features may increase the chance of enhancing the performance. 8. results and discussion figures 10 and 11 display the actual classification accuracies of every network performing version-1 classification of normal vs. severe osa conditions and version-3 normal vs. patient osa conditions respectively. the spectral features used are numbered in figs. 11 and 12 as: (1. vlf, 2. lf, 3. hf, 4. vlf/lf, 5. vlf/hf, 6. lf/hf). the patient-class refers to a set including all the apneic levels of mild, moderate and severe subjects where the ahi index is exceeding 6 and up to 100 or more. the features were selected based on the highest training results they provided before testing the networks as well as for their lowest discrepancy among other schemes. figures 12 and 13 summarize the results of the networks at versions-2 and version-4 classification by emphasizing the spectral features generating the highest performances. it can be noticed that version-4 classification was the most difficult and the results are extremely poor. herein, it can be concluded that spectral features alone are not enough for this complicated task. however, when using statistical time-domain features with the feedforward network; the result is extremely improved reaching an accuracy of 77.5% using two features only. the features are of pnn20 & pnn30 respectively. this shows that the statistical features extracted from the rri signal at the time domain used with the feedforward network were able to generate a non-linear decision surface that helped classifying those subjects unlike the probability estimates or linear classification boundaries generated by the probabilistic and perceptron networks respectively. it can be observed that the feedforward with back-propagation neural network achieved high accuracies at different classification versions demonstrating the power and consistency of the network. further, the specificity and sensitivity percentages are very close to each other at the different classification versions of the different networks which implies that only few unique cases were miss classified. table 1, summarizes the results of the networks at the different classification versions by emphasizing the features generating the highest performances using scheme 1 of training. the classification performance between normal individuals and severe osa patients was highest using the probabilistic network where the network was able to classify the subjects into either normal or severe conditions at 95% accuracy. this is an interesting result and shows how powerful the probabilistic network can be even when trained with a few sample points; it was still able to generalize well. moreover, this accuracy was achieved when using a single spectral feature which is the very low frequency power feature. the test set included 20 subjects (10 normal and 10 severe) meaning that the network only missed the correct classification of one severe subject at a sensitivity of 90%. identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition 31  figure 10. networks classification results for normal vs. severe. figure 11. networks classification results for normal vs. patient. figure 12. networks classification results for normal vs. mild. figure 13. networks classification results for normal vs. mild vs. moderate vs. severe. figure 14. neural networks topology for the four classification versions. the second classification version between normal and mild osa patients witnessed its best performance at 95% accuracy by the feedforward network among all training schemes (original and large training sets). hence, it can be considered a robust network algorithm, since many examples were introduced to the network that included different types of apneic episodes exhibiting different cessation occurrences and durations. moreover, when comparing this result to other researches available in the literature, it is noticed that many of them had their highest results when dealing with small sets (~ 30 subjects where sarah qasim ali and abdulnasir hossen 32  ~10 are normal and ~20 are severe). the latter makes our results acceptable and the employed database was efficient when used for classification. the third classification version between normal and osa patients on the other hand, achieved performance accuracy of 87.5% using three different features (vlf+vlf/hf +lf/hf) by the feedforward network and achieved 85% accuracy by the perceptron network with the three different features (lf+ hf+ vlf/hf). finally, it can be observed that the highest achieved test set accuracy of 77.5% in the fourth classification version between normal, mild, moderate and severe osa subjects was accomplished by the feedforward network using a combination of two statistical features of pnn20 and pnn30 respectively. the representations in fig. 14 demonstrate each type of network topology with different features as inputs for the classification versions corresponding to the best accuracies listed in table 1. for the third version, the perceptron network is sketched oxygenated blood although its performance accuracy was 85% and less than that of the feedforward network 87.5%, in order to show the topology of the perceptron network. 9. conclusion obstructive sleep apnea (osa) is a common breathing-related sleep disorder affecting individuals of different age groups, genders and origins. it is characterized by short-duration cessations in breathing during sleep due to the collapse of the upper airway. osa is associated with major co-morbidities such as cardiovascular diseases, arrhythmias, strokes, obesity, depression, certain types of hypertension and diabetes. the golden and reliable standard test for the detection of osa is a polysomnographic sleep study. however, this test is time/labor-consuming, expensive and cumbersome. analysis of a heart rate variability (hrv) signal that is obtained from an electrocardiograph (ecg) signal in time or frequency domain is an effective, non-invasive and promising method for the detection of osa. in this research, single perceptron, feedforward with back propagation and probabilistic artificial neural networks are investigated for their performance in classifying squh database subject’s severity degree against four classification versions. the highest achieved accuracy of 95% was obtained when using vlf feature with the probabilistic neural network for normal vs. severe classification (version1). the feedforward neural network achieved an accuracy of 95% as well when classifying normal versus mild osa patients at a combination of lf and vlf/hf ratio features. in version 3, the feedforward network achieves 87.5% accuracy using three features vlf and vlf/hf and lf/hf for normal vs. patient (including: mild, moderate and severe subjects in one group) classification. in the same version, the perceptron network achieved the highest performance accuracy of 85% using lf along with hf and vlf/hf ratio combination of features. finally, for osa severity degree classification (verion-4) statistical time-domain features provided the highest accuracy of 77.5% when using a combination of pnn20 and pnn30 features with the feedforward neural network. the results are considered promising since the networks only used a maximum of three features to provide such results. some of the limitations of this work include the ecg database size, neural networks training processing-time especially the feedforward network, feature dimensions (combinations of inputs). future research recommendations are to be on investigating deep learning neural networks. network type classification version features specificity sensitivity accuracy probabilistic normal vs. severe vlf 100% 90% 95% feedforward normal vs. mild lf+ vlf/hf 90% 100% 95% perceptron feedforward normal vs. patient normal vs. patients lf+ hf+vlf/hf vlf+vlf/hf +lf/hf 80% 90% 86.67% 86.67% 85% 87.5% feedforward normal vs. mild vs. moderate vs. severe pnn20+pnn30 80% 76.67% 77.5% table 1. highest classification results of the networks at training scheme 1. identification of obstructive sleep apnea using artificial neural networks and wavelet packet decomposition 33  conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references al ghunaimi b (2003), statistical signal characterization and sub-band decomposition for heart rate variability analysis in patients with obstructive sleep apnea (masters dissertation, sultan qaboos university). american academy of sleep medicine (2001), international classification of sleep disorders, revised: diagnostic and coding manual. chicago, illinois: american academy of sleep medicine. dantest clinicians team, “ans balance assessments”. (2016). sited from: https://www.dantest.com/dtr_ans_overview.htm. global leaders in sleep and respiratory medicine. (2013), sleep apnea facts and figures. resmed corp. san diego, ca, usa. hagan m. t. demuth, h.b. beale m.h., and de jesús o (1996), neural network design (vol. 20). boston: pws publishing company. hamilton g., chai-coetzer c., (2019), “update on the assessment and investigation of adult obstructive sleep apnea” australian journal of general practice, 48 (4): 176-181. kumar v.m. (2008), sleep and sleep disorders. indian journal of chest diseases and allied sciences 50(1): 129. lado m.j. vila x.a. rodríguez-liñares l. méndez a. j. olivieri d. n. and félix p (2011), detecting sleep apnea by heart rate variability analysis: assessing the validity of databases and algorithms. journal of medical systems 35(4): 473-481. lee, w., nagubadi, s., kryger, m. h., & mokhlesi, b. (2008). epidemiology of obstructive sleep apnea: a population-based perspective. expert review of respiratory medicine, 2(3), 349–364. mietus, j e; peng, c.k.; henry, i.; goldsmith, r.l.; goldberger, a.l. (2002). “the pnnx files: re– examining a widely used heart rate variability measure”. heart. 88: 378–380. misiti, m., misiti, y., oppenheim, g., & poggi, j. m. (1996). matlab wavelet toolbox user’s guide. the mathworks. inc. mass. mohri m., rostamizadeh a., talwalkar a., (2012) foundations of machine learning. the mit press isbn 9780262018258 moore j. heart rate variability course (2016), “heart rate variability vs. heart rate”, sited from: https://hrvcourse.com. national heart, lung and blood institute. what is sleep apnea?, (2012). sited from: https://www.nhlbi.nih.gov/health/health-topics/ topics/sleepapnea national sleep foundation. (2016). “sleep apnea”. sited from: https://sleepfoundation.org/sleepdisorders-problems/sleep-apnea polikar, r. (2011). the wavelet tutorial parts 1-3. second edition. sited from: https://cseweb.ucsd.edu/baden/doc/wavelets/polik ar_wavelets.pdf. sharma s., raval m., acharya u.r., 2019, a new approach to identify obstructive sleep apnea using an optimal orthogonal wavelet filter bank with ecg signals’ informatics in medicine unlocked, 16, 100170: 1-13. sysel, p., & misurec, j. (2008). estimation of power spectral density using wavelet thresholding. in s. kartalopoulos, a. buikis, n. mastorakis, & l. vladareanu (eds.), wseas international conference proceedings mathematics and computers in science and engineering (no. 7). world scientific and engineering academy and society the university of nottingham, practice learning resources: cardiology teaching package. (2017). http://www.nottingham.ac.uk/nursing/practice/res ources/cardiology/function/normal_duration.pp ucdavis health system. ‘heart rate’, (2016). sited from:www.ucdmc.ucdavis.edu/sportsmedicine/res ources/heart_rate_description.html webmd, ‘electrocardiogram’, (2013). sited from: webmd.com/heartdisease/electrocardiogram/wilso n, s. (2013). sleep disorders (oxford psychiatry library). 2nd ed. oxford, united kingdom: oxford university press, usa, p.1 xie, w., zheng, f., & song, x. (2014). obstructive sleep apnea and serious adverse outcomes in patients with cardiovascular or cerebrovascular disease: a prisma-compliant systematic review and meta-analysis. medicine, 93(29): 1-8. civl080703.qxd vol. 3, no. 1 (2006) 79-83ring researchrnal of engineethe jou 1. introduction the kingdom faces problems similar to those of arid zones having vast areas. the kingdom of saudi arabia is a large country (over 2.2 million square kilometers), occupying an area approximately the size of western europe or one-fourth the size of the united states. it has a population of approximately 20 million scattered throughout the country with major concentrations in several distinct regions (ministry of planning, saudi arabia, 2001). generally, the weather in saudi arabia is arid and hot, with some of the earth's harshest terrain. _________________________________________ *corresponding author's e-mail: ahmadi@kfupm.edu.sa furthermore, the kingdom includes features from mountains in the southwest and near the red sea, to sand dunes in the central and eastern parts of the country. female students that attend colleges in the kingdom are about 95,000. the media has raised some concern about the difficulties faced by girls traveling to the cities for their studies. in order to reduce travel hazards experienced by female students, some newspapers have suggested that new colleges to be opened near villages. the purpose of this study is to examine difficulties associated with female students traveling from villages into cities. the main goal is to investigate the characteristics of their trips and offer some recommendations in order to improve the conditions in which they travel. travel characteristics of female students to colleges in the kingdom of saudi arabia h.m. al-ahmadi* civil engineering department, college of engineering, king fahd university of petroleum and minerals, p.o. box 31261, dhahran, saudi arabia received 8 july 2003; accepted 26 march 2005 abstract: this study is focused on the problems faced by female students residing in the kingdom of saudi arabia. many of these female students live far from their colleges and they have to travel from villages to cities in order to obtain a university degree. in so doing they travel long distances everyday using various modes of transport. as they are not permitted to drive, these students depend on male drivers to take them to their colleges. the aim of this study was to investigate and identify problems associated with such modes of transport. results show that most students experience some problems and encounter disproportionate levels of hardship. the findings were based on a survey carried out in the kingdom and was distributed to representative sample of female students. this paper includes a detailed analysis of the data from which a number of conclusions and recommendations were made. the conclusions suggested that the travel situation encountered by female students who travel by passenger car could be improved in terms of time, cost and convenience. thirteen percent of female students rated the vehicles they used to be in unacceptable condition for traveling while 6% marked the drivers as not violating traffic rules, and the majority (53%) was captives to their current mode of transportation. keywords: saudi arabia, travel characteristics, college female students, driver performance, mode of transportation ájoƒ©°ùdg á«hô©dg áµ∏ªÿg ‘ äé«∏µ∏d äéñdé£dg π≤æj ¢üfé°ün ……~~ªªmm’’gg ..ωω..ìì @@ ::áá°°uuóóÿÿgg√gôª∏d ì骰ùdg ω~©d ngô¶fh .á«©eé÷g ioé¡°ûdg ≈∏y ∫ƒ°üë∏d ¿~ÿéh äé©eé÷gh äé«∏µ∏d ôé¡dg h iô≤déh ä骫≤ÿg äéñdé£dg π≤æj ¢üfé°ün áaô©e ≈∏y á°sgq~dg √òg õcôj √ògh .π≤æàdg á«∏ªy ‘ äéhƒ©°üdg ¢ ©h ø¡lgƒj äéñdé£dg ö∏zg ¿g á°sgq~dg èféàf øeh .äé«∏µ∏d π≤æàdg ‘ πlôdg ≈∏y äéñdé£dg ~ªà©j ájoƒ©°ùdg á«hô©dg áµ∏ªÿg ‘ ioé«≤déh π°üøe π«∏– ≈∏y ábqƒdcg √òg …ƒà–h .áµ∏ªÿg ≥wéæe ∞∏àfl ‘ äé«∏µ∏d äó≤æàÿg äéñdé£dg øe á«fgƒ°ûy äéæ«y ≈∏y âyrh ,¢vô¨dg gò¡d ä~ycg äéféñà°sg ≈∏y ä~ªàyg èféàædg ø¡∏«°uƒàh ¿ƒeƒ≤j øjòdg ú≤fé°ùdg øª«b äéñdé£dg øe %6 h g~l áä«°sh áä«°s úh ø¡∏≤æj πfé°sƒd áeé©dg ádé◊g øª«b äéñdé£dg øe %13 ¿g á°sgq~dg âæ«h å«m ᩪûg äéfé«ñ∏d .é«dém áe~îà°ùÿg á∏«°sƒdg òz …ôncg π≤æj πfg~h ø¡d ôaƒàj ’ äéñdé£dg øe %53 ¿gh áæeb’g ioé«≤dg ~ygƒ≤h úeõà∏e òz º¡fcg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.π≤ædg á∏«°sh ,ú≤fé°ùdg qéªycg ,äé«∏µ∏d äó≤æàÿg äéñdé£dg ¢üfé°ün ,áæe’g ioé«≤dg ,ájoƒ©°ùdg á«hô©dg áµ∏ªÿg : 80 vol. 3, no. 1 (2006) 79-83ring researchrnal of engineethe jou this paper specifically addresses objectives to define characteristics of female student travel from the village to the city, which include but not limited to the investigation of the proportion of 'captive' female students in saudi arabia and investigation of the characteristics of the students' daily trips to and from college. 2. sampling as a homogeneous group, female students seeking their university degree were chosen as the subjects of this study, because they form a special group who experiences a unique set of problems relating to their travel requirements. they start their journey at dawn, some of them travel hundreds of kilometers before reaching their destination. this is, therefore, an exhausting ordeal. these students have access to a variety of road transportation including passenger cars, suvs (sports utility vehicle) called 'suburban' in saudi arabia, taxis, buses, or minivans. more than 2000 questionnaires were sent to the deans of colleges to distribute to female students coming from villages to study. three follow-up request letters were sent out before an acceptable number of questionnaires were received. this number was 1360 of the total questionnaires distributed. of these, 428 were rejected due to the fact that the female students were living in the city. finally, 892 questionnaires were deemed acceptable for analysis. 3. data collection the characteristics were obtained mainly from a data set which was collected as part of this study. the data of this paper were generated from a comprehensive study (al-ahmadi, et al. 2002). the first step involved reviewing a number of sources of information appropriate to the subject. these included a literature survey, expert opinion, focus group meetings and common sense. this process was difficult and required several iterations before a satisfactory set of variables was obtained from which a pool of variables was prepared. these included: time, distance and cost of trips; driver nationality, age and performance; mode of transportation chosen and reason for the choice; vehicle owner, manufacturer and condition. the second step was the development of the survey instrument, which took the form of a self-response type questionnaire. it was first sent to the department of girls' education where it was extensively reviewed and tested to assure that no ambiguities were contained in the list. the questionnaire was then sent through the ministry of education to selected colleges. selection was based on criteria associated with typographical features throughout saudi arabia (ministry of education, saudi arabia). table 1 shows the list of colleges where the questionnaire was distributed. 4. data analysis primary analysis revealed a number of interesting factors. first, it was found that 14% of female students traveling from village to city were married. second, female students spent, on average, 393 riyals monthly on their travel. third, vehicle occupancy was on average 20 students per vehicle. fourth, analysis of driver ownership showed that 65% of the drivers owned the vehicles. finally, statistics for the time and distance of travel revealed that, on average, each one-way trip was 55 km long and took 55 minutes. detailed analyses of the other characteristics are as follows: 4.1 driver age figure 1 shows that 69% of the drivers were aged 3050. drivers of this age group generally do not suffer from health issues associated with older drivers, such as poor eyesight, which may negatively impact their driving (forbes, 1972). also, these drivers were mature and possess experience, expertise and knowledge of the road rules which is importance during th driving. drivers under the age of 30 (15%), on the other hand, may possess poorlydeveloped attitudes towards driving. accordingly, there are many concerns with this age group regarding the ignoring of road rules, dangerous driving and incidents involving anger and immaturity (al-shara, alsaef, 1988; shipp, 1988). college city intermediate college al-nimas intermediate college al-ula intermediate college qwaieh intermediate college qariyat college of education science department makkah al mokarrama h college of education science department hail college of education science department majmaah college of education science department jizan college of education science department tabuk college of education science department al-ahsa college of education science department al-kharj college of education science department abha college of science dammam college of education unayzah college of education science department al-baha college of science taif table 1. the colleges that the questionnaires were distributed 18 vol. 3, no. 1 (2006) 79-83ring researchrnal of engineethe jou 4.2 driver performance figure 2 shows that only 5% of the respondents felt that driver performance was unacceptable. in fact, 79% rated the driving as either excellent or good. however, it must be pointed that this data is "softscience" in that it relates to the students' subjective interpretation about their drivers' adherence to traffic rules. 4.3 vehicle condition statistics on vehicle condition mirrored those of driver performance. figure 3 shows that only a small percentage (13%) felt that the condition of the vehicles in which they travel was unacceptable. in fact, 41% of the respondents regarded the vehicles as good while 28% felt they were in excellent condition. once again, this was subjective data based on the passengers' interpretation of the vehicles; not reflecting the actual mechanical condition of the vehicles. 4.4 owner of mode of transportation figure 4 relates to the owner of the vehicle. spouses (46%) were those who owned and drove the passenger cars. this fact further disadvantaged students who used passenger cars for a number of reasons. first, the spouses' own study schedules must be factored into the travel plans of the students. second, the full cost of maintaining and operating the vehicle must be borne by the spouses. twenty three percent of the students traveling used company vehicles while 17% relied on government vehicles. 4.5 driver nationality figure 5 shows that about 58% of the drivers transporting female students to their colleges were saudis. however, due to the large number of non-responses to this question (33%), these statistics required interpretation. therefore, it must be assumed that saudis made up a substantial proportion of drivers. it was observed that asians and arabs made up 6% and 2% of the driver population respectively. 4.6 mode of transportation figure 6 indicates the mode of transportation used by students. the chart shows that buses were the most common mode of transport (39%). the second form was pasdriver age > 60 5% 50 60 21% 40 50 31% < 30 15% no answer 4% 30 40 24% figure 1. pie chart for driver age figure 2. pie chart for driver performance figure 3. pie chart for vehicle condition figure 4. pie chart for owner of mode of transportation poor 4% bad 9% acceptable 15% excellent 28% no answer 3% good 41% no answer 2% excellent 38% poor 3% acceptable 14% good 41% no answer 5% spouse 46% company 23% government 17% others 8% bad 2% 28 vol. 3, no. 1 (2006) 79-83ring researchrnal of engineethe jou senger cars (22%), while this type of vehicle had a maximum carrying capacity of 4 passengers; the study revealed that auto occupancy ranged from 1 to 2. the third most common vehicle category was shared by the 'suburban' (16%) which had a capacity of 8 passengers and the minivan (19%) which could carry up to 13 passengers. cost was an important factor, in that passenger cars were the least cost efficient form of transport for the students due to the low vehicle occupancy. 4.7 reason for choice of mode there were a number of reasons for choosing a particular form of transport. from fig. 7 it could be seen that the majority (53%) of the respondents had to rely on their chosen form of transport because they had no other option. however, members of this group would choose another option if one were made available. 4.8 vehicle manufacturer figure 8 shows that japanese and korean vehicles made up 35% of the total. generally, japanese and korean vehicles were small passenger cars with a maximum carrying capacity of 4 passengers. american vehicles (23%), on the other hand, were larger vehicles such as the 'suburban' with a capacity of about 8 passengers. 5. conclusions and recommendations having carefully examined the data, several conclusions and recommendations could be suggested. these are as follows: female students spent, on average, 392 riyals monthly traveling to and from college. as this is quite high, measures should be taken to reduce this cost. special lowcost transport services could be provided or transport costs could be fully or partially subsidized by the government. in addition, travel costs could be reduced if each passenger car carried the maximum number of passengers. this could be achieved by publishing and posting special bulletins showing information of routes and costs of travel at the relevant colleges so that students could arrange a form of car pooling. on average, students traveled for about 55 minutes and covered a distance of approximately 55 kms one-way each driver nationality arab 2% no answer 33% saudi 58% asian 6% far east 1% others 0% figure 5. pie chart for driver nationality mode of transportation taxi 1% no answer 3% passenger car 22 % bus 39% sub urban 16% van 19% figure 6. pie chart for mode of transportation mode choice reason no answer 10% safety 3% others 1% security 9% cost 2% no other option 53% personal comfort 22% figure 7. pie chart for reason for choice of mode vehicle manufacturer korean 2% european 2% others 10% japanese 34% no answer 29% american 23% figure 8. pie chart for vehicle manufacturer 83 vol. 3, no. 1 (2006) 79-83ring researchrnal of engineethe jou day. this could be alleviated by providing residences for students near colleges. as it might be impractical for spouses who work in the village to live in the city, female students could stay in the accommodation during the week near the college and travel home to the village at the weekend. while the majority of the drivers (69%) fell into the least problematic age group, there were concerns regarding drivers at the two ends of the age spectrum. there need to be regular medical checks for drivers over the age of 50 and driver education courses for those under the age of 30. the majority of female students (93%) thought that driver performance was acceptable, while 5% thought it was unacceptable. this 5% might pose danger in causing accidents. therefore, drivers are recommended to take special tests and permits for those driving female students to colleges. female students (84%) responded that the vehicle condition was acceptable. on the other hand, 13% believed that it was unacceptable. therefore, it is recommended that special vehicle inspection should be done regularly for the vehicle in which female students travel. the spouses (46%) made substantial portion of the female student drivers. the majority of the female students (53%) traveling from village to city for studying did not have any other mode of choice. this made them captive to the mode they were using. fourteen percent of the students were married and were disproportionately disadvantaged with regards to their travel conditions. these students had associated family responsibilities. solutions are required with respect to scheduling and travel costs. special attention should be paid to studying schedules, and allowances should be made with respect to start/finish times to enable these students to strike an acceptable balance between their responsibilities to their families and studies. effort should be made for encouraging female students in villages to get higher education in universities by opening colleges in the middle distance between villages. acceptable government buses should be provided with experienced drivers adhering to traffic laws for transporting the female students from villages to cities. references al-ahmadi, h.m., ratrout n., al-ofi k. and alsughaiyer m., 2002, "accident for female teachers and students traveling in rural areas in saudi arabia," final report submitted to kacst. al-shara, a. and alsaef, k., 1988, "analysis of the social and psychology factors of the drivers in saudi arabia,” final report submitted to kacst. forbes, t.w., 1972, human factors in highway traffic safety research, new york: wiley. kingdom of saudi arabia, ministry of planning, 2001, statistical year book, central department of statistics, issue 36 , 1420 a.h. ministry of education, 2001, statistics records, statistics department, saudi arabia 1420 a.h. shipp, m., 1988, "vision testing policies for driver license renewal: benefit or barrier,” phd thesis, the university of michigan. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 19, no. 2, (2022) 129-139 0043 corresponding author's e-mail: zahidmushtaq69@gmail.com doi:10.53540/tjer.vol19iss2pp129-139 exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil zahid mushtaq*, m. hanief mechanical engineering department, national institute of technology, srinagar, jammu and kashmir, india. abstract: it is imperative to emphasize the wear-reducing and load-sustaining abilities of the lubricating oils, as they are an essential part of their operation. in this study, the anti-wear and extreme pressure (ep) performance of jatropha oil was explored. the investigation was executed on a 4-ball tester machine according to astm-d4172 and d-2783 standards. nano-sized molybdenum disulfide and graphene were added to the base oil, whose addition reduced the wear on the test balls but failed to improve the ep properties. however, the addition of garlic oil elevated the ep properties of the base oil. the last non-seizure load and weld point of the jatropha oil was elevated by up to 3 levels, while the load wear index was increased by 91.88%. the flash temperature parameter of jatropha oil was increased by 658% and 250% at 100kgf and 126kgf load, respectively, after the addition of 5% garlic oil. further, the wear scars were examined under an optical microscope and fesem, which proved that the addition of nano-additives reduced the wear of the balls during anti-wear tests at lower loads, and garlic oil reduced the wear in ep conditions and produced smoother surfaces. these improvements were attributed to the formation of tribo-films on the surfaces as characterized by the eds analysis. it was concluded that garlic oil could be a substantial ep additive to vegetable oils (especially jatropha oil) to lift their load-carrying capacity without hampering their environmentally friendly nature. keywords: jatropha oil; wear; garlic oil; molybdenum disulfide; graphene. والضغط الشديد لزيت الجاتروفا المعدل بمضافات النانو وزيت الثوم للتآكل استكشاف أداء التشحيم المضاد زاهد مشتاق و م. حنيف من الضروري التأكيد على قدرة زيوت التشحيم على تقليل التآكل والحفاظ على الحمل، حيث إنها جزء أساسي من عملها. الملخص 4( لزيت الجاتروفا. تم إجراء التحقيق على آلة اختبار epوالضغط الشديد )في هذه الدراسة، تم استكشاف األداء المقاوم للتآكل . تم إضافة ثاني كبريتيد الموليبدينوم والجرافين بحجم النانو إلى الزيت األساسي، d-2783و astm-d4172كرات وفقًا لمعايير مع ذلك، أدت إضافة زيت الثوم إلى زيادة . و epوالذي قللت إضافته من تآكل كرات االختبار لكنها فشلت في تحسين خصائص مستويات، في 3للزيت األساسي. تم رفع الحمل غير المضبوط األخير ونقطة اللحام لزيت الجاتروفا بما يصل إلى epخصائص د ٪ عن250٪ و 658٪. تم زيادة معامل درجة حرارة الوميض لزيت الجاتروفا بنسبة 91.88حين تم زيادة مؤشر الحمولة بنسبة ٪. عالوة على ذلك، تم فحص ندوب التآكل تحت المجهر 5كجم على التوالي بعد إضافة زيت الثوم بنسبة 126كجم و 100حمل ، والتي أثبتت أن إضافة إضافات النانو قللت من تآكل الكرات أثناء اختبارات مقاومة التآكل عند األحمال fesemالبصري و وقلل زيت الثوم من التآك على epل في ظروف المنخفضة، وأنتج أنعم األسطح. تُعزى هذه التحسينات إلى تكوين أفالم تريبو للزيوت النباتية )خاصة زيت الجاتروفا( أََساِسيًّا . تم استنتاج أن زيت الثوم يمكن أن يكون مضافًا edsاألسطح كما تتميز بتحليل الصديقة للبيئة لرفع قدرتها على تحمل األحمال دون إعاقة طبيعتها .الجرافين موليبيديوم داسلفايد؛ ؛زيت الثوم؛ التآكل ؛ زيت الجاتروفا الكلمات المفتاحية: exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil 130 nomenclature fesem field emission scanning electron microscope. ftp flash temperature parameter. gar garlic oil. gr graphene. jo jatropha oil. lnsl last non-seizure load. lwi load-wear index. m molybdenum disulfide. ra roughness average. rq root mean square roughness. w.p weld point. wsd wear scar diameter. 1. introduction the consumption and scarcity of energy are on the alltime rise. the condition of shortage of fossils is inevitable in the future due to its present rapid exhaustion (wang et al., 2020). tribological losses are a major source of energy losses, and there is a dire need to minimize their level to save energy in various industries (tahir et al., 2018). lubrication plays a major role in reducing friction and wear by forming a film between the mating parts and preventing their direct contact (yazawa et al., 2014). a majority of the lubricants available in the market are petroleum-based products. they provide good lubrication but are not biodegradable and eternal. they contribute to pollution, and their demand needs to be reduced by introducing environmentally friendly lubricants like vegetable oils (shafi & charoo, 2021). vegetable oils have been under the spotlight of research for the last couple of decades for their impeccable lubricating properties. we can say that they are a sure-shot-cure to mitigate the universal problem of swift crude oil depletion (owuna et al., 2020; woma et al., 2019). very vast research has concluded that vegetable oils can make as good lubricants, and their properties can be ameliorated by mixing desirable types of additives like anti-friction and anti-wear additives, nanoparticles, extreme pressure additives, etc. (chan et al., 2018; hernández-sierra et al., 2019; panchal et al., 2017; zainal et al., 2018). jatropha oil has staggering tribological potential, and it has been proved to be a viable option to replace mineral oils by numerous studies. the lubricating properties of jatropha oil have been improved by adding various types of additives (mushtaq & hanief, 2021b, 2021a, 2021c; talib et al., 2019, 2021). there are a number of different types of additives available which can enhance the performance of base lubricants, and care needs to be taken in selecting the best and most suitable one. if the nanosized additives are added to a lubricant, then it is called a nano-lubricant. the friction and wear-reducing capacities of the nano-lubricants are much better as compared to the base oils (kotia et al., 2018). recent research has suggested that additives like graphene and mos2 are good options for lubricant additives to reduce friction and wear (berman et al., 2014; he et al., 2020; lin et al., 2011; meng et al., 2020; wu et al., 2018). the addition of graphene nanoparticles was reported to boost the tribological properties of vegetable oils and decrease the coefficient of friction and wear scar diameter (kiu et al., 2017). it has been proved that mos2 and graphene nanoparticles, when added in the mixed form, are very effective in reducing friction and wear even better than when added individually (gong et al., 2017; xu et al., 2015). garlic oil is a natural plant oil that is also used nowadays as a liquid additive to enhance the properties of lubricants. it has been reported to enhance the extreme pressure properties of base oils by a considerable amount and elevate their weld point (li et al., 2014). garlic oil was added to palm oil and coconut oil, and the extreme pressure properties of the mixed lubricants were examined on a four-ball tester. it was observed that the mixed lubricants gave better performance than the mineral oil (chikezie & ossia, 2018). garlic oil improved the oxidative stability and pour point of the coconut oil and converted it into a much better performing lubricant without harming its nonhazardous nature (kumar et al., 2021). the impact of the combination of graphene and mos2 additives on the anti-wear behaviour of jatropha oil has not been tested yet. also, the effect of garlic oil on the ep properties of jatropha oil still remains to be investigated. hence, the current study emphasizes the anti-wear and extreme pressure properties of jatropha oil. nano-sized molybdenum disulfide (m) and graphene (gr) were added in the assorted form in the ratio of 1:1 in weight percentages of 0.1 to 0.4%. the above nano-lubricants were reported to reduce the wear scar diameter of the test balls during astmd4172 tests. however, they did not elevate the weld points of the test lubricants during astm-d2783 tests. hence, pure garlic oil was added in 1% to 5% by volume to the jatropha oil because of its biodegradable nature. it showed promising results and improved the extreme pressure properties of the jatropha oil, and elevated the weld points. hence, it can be summarized that the addition of nano-additives improves the antiwear properties of the jatropha oil, and the garlic oil boosts the extreme pressure properties and makes it resist more load. 2. materials and methodology 2.1 equipment and the tribo-pairs a four-ball tester (as shown in figure 1) was employed to conduct the anti-wear and extreme pressure experiments with 52100 chromium steel balls of 12.7mm diameter as tribo-pairs. both the equipment and the test balls were delivered by ducom instruments zahid mushtaq and m. hanief 131 pvt. ltd. the physical properties of the test balls are given in table 1. a four-ball tester consists of a steel cup in which three test balls are accommodated in a triangular shape, with all three making contact with each other, as shown in figure 2. then the pot is filled with the test lubricant so that all three balls are completely submerged in it. the 4th ball is gripped in the upper part, which is rotated with the help of a motor. the 4th ball is then pressed against the rest of the three balls in the form of an equilateral tetrahedron such that it makes point contact with all three balls simultaneously. the steel cup is also connected to an electrical heater and a temperature sensor to maintain a constant temperature according to the requirement of the experiment. the operating parameters are fed through computer software equipped with the machine, and after attaining the desired parameters like load and temperature, the test is started. as the upper ball is pressed and rotated against the three lower balls, it causes a scar on each of them. the diameter of the three scars on the lower balls is measured and then averaged, and this value signifies the lubricating capabilities of the test lubricant used in the steel cup. 2.2 synthesis of bio-lubricants the base oil used in this study was vegetable oilbased jatropha oil, whose fatty acid combination is shown in table 2, while its physicochemical properties are presented in table 3. the nano-additives of molybdenum disulfide and graphene were procured from a reliable supplier, and their properties are given in table 4. the lubricants were synthesized by adding different proportions of additives as percentages of weights. both the nano-additives were mixed in the ratio of 1:1 and added by weight from 0.1% to 0.4% after weighing on a highly accurate weight balance. the nano-lubricants then produced were subjected to 6 hours of ultrasonication to ascertain the homogeneous dispensation of nanoparticles before starting the experiments. figure 3 shows the prepared nanolubricant samples of jatropha oil mixed with nanoadditives at different percentages. however, the pure garlic oil was used as an extreme pressure additive and was added to the pure jatropha oil by volume percentages from 1% to 5%. the magnetic beads were submerged in the test tube containing the mixtures and were stirred for 2-3 hours on a magnetic stirrer. all the lubricant samples were placed in separate test tubes. figure 2 schematic diagram. table 1. physical properties of the tribo-pairs. s. no physical properties 52100 cr steel ball 1. hardness (hrc) 63 66 2. elastic modulus (gpa) 190 210 3. density (g/cm3) 7.81 4. poisson’s ratio 0.29 5. melting point (oc) 1424 6. tensile strength (mpa) 2240 7. yield strength (mpa) 2030 table 2. fatty acid combination of jatropha oil s.no. fatty acids composition (%) 1. myristic acid c14:0 1.2 1.6 2. palmitic acid c16:0 15 16 3. stearic acid c18:0 9 10 4. oleic acid c18:1 40 42 5. linoleic acid c18:2 31 34 6. arachidic acid c20:0 0.3 – 0.6 figure 1. four-ball tester. exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil 132 2.3 experimental procedure the anti-wear (a/w) tests were conducted on the four-ball tester according to astm-d4172 standards. the operating conditions associated with it are: load = 392 n, temperature = 75oc, rpm = 1200 and time = 3600 seconds (1 hour). four 52100 chromium steel balls were used for each test, and all of them were washed thoroughly with acetone before the tests to remove any contamination. the three balls were placed in the steel cup, and it was filled with the test lubricant. the upper ball was pressed against these lower three balls by applying the load and other parameters. as soon as the desired level of parameters was reached, the test was started. the test stopped spontaneously after the completion of the given time. each sample of nano-lubricants was tested for its anti-wear properties by filling the steel cup with them one by one. after completing the test, the test lubricant was displaced, and the balls were removed and washed again with acetone. then the wear scars on the lower three balls were examined and measured under an optical microscope and fesem. the size of the scars gave the anti-wear performance of the test lubricant. the extreme pressure (ep) tests determine the load sustaining capacity of the test lubricant and give the maximum load it can withstand before causing failure. the ep tests were conducted on the same machine under the astm-d2783 standard. the operating conditions associated with it are: load = varying from 6kgf to 800kgf, temperature = 27oc ±2, rpm = 1760 ±40, and time = 10 seconds. starting from the lower loads, the load was increased step by step and continued until the seizure of the four balls took place. the balls and the test lubricant are not reusable and can be used only in one test. hence they need to be replaced after every test. the balls were cleaned with acetone before each test to remove any unwanted particles from the surface. the lubricant samples with mixed nanoadditives (0.1 to 0.4 wt.%) and garlic oil (1 to 5 vol.%) were tested alternately for their extreme pressure capabilities under different loads until their seizure occurred. their weld points were noted down, and the scars produced on the balls were examined and measured under an optical microscope and fesem. in case of any discrepancy, the ambiguous tests were repeated three times to bring more transparency. 3. results and discussions 3.1 anti-wear tests the one-hour anti-wear tests were performed according to astm d-4172 standards using a four-ball tester. the balls were lubricated with jatropha oil with and without additives. table 3. physicochemical properties of jatropha oil. s.no. properties jatropha oil 1. density (g/cm3) 0.918 2. kinematic viscosity at 40oc 47 54 cst 3. kinematic viscosity at 100oc 9 -10 cst 4. viscosity index (vi) 180 186 5. flash point (oc) 186oc 7. pour point (oc) 8oc table 4. properties of nano-additives. specifications graphene molybdenum disulphide formula c mos2 appearance grayish black powder black powder purity >99% 99.9% odour odorless odorless thickness/aps 5 – 10 nm 80 – 100 nm length 5 – 10 micron density 4.2 g/cm3 5.06 g/cm3 melting point 3650oc 1185oc molecular weight 12.01 g/mole 160.07 g/mole figure 3: nano-lubricants prepared from jatropha oil. figure 4: average wear scar diameter (wsd) of test balls. zahid mushtaq and m. hanief 133 the additives were mos2 nanoparticles and graphene nanoflakes mixed in the ratio of 1:1 in the different weight percentages of 0.1 to 0.4%. an optical microscope and fesem were used to examine the worn-out lower balls and measure the diameters of their respective wear scars produced after the tests. the three values of diameters were averaged to get the average wear scar diameter (wsd). figure 4 displays the average wsd of the tests recorded on the test balls lubricated with different mixtures. the average wsd of test balls, when lubricated with pure jatropha oil, is 833.413 µm. the diameter recorded is within the range of wsd (510 µm – 870 µm) allowed for commodity oils (cermak et al., 2013). however, as the nanoadditives were added to the oil, the anti-wear properties reportedly got improved, and the average wsd was reduced. the minimum average wsd was reported at 0.1% concentration of nano-additives. the average wsd of jatropha oil with 0.1%m+gr was 633.291 µm, as shown in figure 4. it is clearly evident that the introduction of nano-additives resulted in lowering the size of the wear scar. the jatropha oil reported an improvement of 24.01% in reducing the wear at 0.1% of activation. hence, 0.1% was found to be the optimum percentage of additives for minimum wsd. this improvement in the anti-wear properties of jatropha oil can be related to the intrusion of nanoadditives inside the peaks and valleys and decrease the surface roughness of the surface. moreover, the additives are more prone to chemically react with the metallic surface, lubricant and the environment. this reaction gives rise to a protective layer on the surface and safeguards the tribo-pairs from being worn out rapidly. the combination of the two nano-additives was used in this research because it has been reported in the literature that the combination of mos2 and graphene additives performs more effectively as antiwear additives than when they are used separately. the graphene delays the oxidation of mos2 and helps it stay along the surface for a much longer time during a sliding operation. while as, the mos2 helps the graphene stay along and not let break up into fine and weak particles (gong et al., 2017; xu et al., 2015, 2018). figure 5 shows the low-magnification (10x) optical and high-magnification (3000x) fesem images of the wear scars on the test balls lubricated with different mixtures. figures 5(a, b) show the wear scar on the ball at low and high magnifications when it was lubricated with pure jatropha oil. a large number of abrasive furrows can be seen spread all over the scar in figure 5(a), indicating abrasive ploughing as the dominant wear mechanism. figure 5(b) shows the material removal and deposition caused by the abrasive action. after the intrusion of the nano-additives by 0.1 wt%, the scar diameter is reduced to a great extent, and less intense grooves can be seen in figure 5(c). the grooves are also less deep and not widespread, giving somewhat less-ploughed area at the top and bottom, which reduces the total wear. figure 5(d) shows a comparatively smoother surface with few fine lines of abrasion. the deposits of nano-additives on the surface can also be seen as black marks in the above image. this clearly indicates that the jatropha oil performed better after the intrusion of mos2 and graphene nanoadditives and reduced the wear of the test balls to a large extent. this improvement is attributed to the formation of a protective film due to the tribochemical reaction between the nanoparticles and the metallic surface. this film prevents direct and intense rubbing between the balls and restricts the wear. figure 5: optical and fesem images of wear scars on balls lubricated with (a, b) pure jo (c, d) jo + 0.1% m+gr. exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil 134 3.2 extreme pressure (ep) tests the ep tests were conducted according to the astmd2783 standards on a four-ball tester. the steel cup was lubricated with different mixtures of jatropha oil and additives. after the tests, the diameter of the scars produced on the lower three balls was measured on an optical microscope. the average wsd recorded at different loads and lubricant mixtures as compared with the compensation scar diameters (csd) are compiled in table 5. it is observed that the wsd for jatropha oil increases with the increase in load. after adding the garlic oil, the wsd increased with the increasing load but at a slow speed as compared to the base oil. the wsd of jatropha oil was reduced by 76.47% after adding 5% garlic oil at a load of 100kgf. when the wsd is within the 5% limits of the csd, then it is called the last non-seizure load (lnsl). the lnsl for pure jatropha oil was found to be 50kgf, and the load at which the four balls get welded (weld point) was recorded as 160kgf. however, after adding garlic oil from 1 to 5 vol.%, both lnsl and weld points got elevated, as presented in table 5. the lnsl of jatropha oil got elevated to 63kgf and 100kgf at 1% and 5% addition of garlic oil, respectively. further, the weld point of jatropha oil increased to 250kgf and 315kgf at 1% and 5% garlic oil, respectively. the load-wear index (lwi) of the lubricant mixtures was also calculated following the astm-d2783 standards and is plotted in figure 6. the lwi of pure jatropha oil was recorded as 26.671. however, it was observed that the lwi increased after the addition of garlic oil. the maximum improvement in lwi was 51.177, as recorded with the addition of 5% garlic oil. hence, it was concluded that the garlic oil elevates lnsl, weld point as well as lwi of jatropha oil. however, it is also reported that the addition of nano-additives had no effect on any of the extreme pressure properties of the jatropha oil. 3.3 wear analysis the high magnification (3000x) fesem image of the test ball lubricated with pure jatropha oil at 126kgf is shown in figure 7(a). it is observed from this figure that the main wear mechanism has shifted to adhesion at a higher load. large cracks and heavy material removal by adhesion can be seen in figure 7(a). it specifies that the lubricant film produced without any additive is not durable and is easily breached, especially at higher loads. this allows severe engagement between the asperities of the two surfaces and results in material removal and the formation of cracks on a large scale. figure 7(b) shows the fesem image of the test ball lubricated with jatropha oil + 5% garlic oil at 200kgf. the addition of garlic oil to the jatropha oil minimizes the adhesive wear between the contacting surfaces and gives a much smoother surface, as shown in figure 7(b). the surface has very small cracks and adhesive marks. it proves the effectiveness of garlic oil in minimizing the wear between the sliding surfaces. after the addition of garlic oil to the jatropha oil, a much smoother surface is produced even at a higher load of 200kgf. this study agrees with the results experienced by several researchers that garlic oil is a very effective extreme pressure additive that improves the lubricity of the lubricant and reduces the wear between the sliding surfaces. it has the capability to form protective tribofilms from the oxides, sulphates, and sulphides of iron. the low-molecular-weight sulphides present in the garlic oil react much readily with the metallic surfaces to form a protective layer. the unsaturated double bond enhances the adsorption of the film on the metallic surface, thereby making it more effective in restricting the direct metal-to-metal contact and minimizing the wear (gasni et al., 2019; jesbin & mahipal, 2021; li et al., 2014). table 5. load versus wsd for various lubricant mixtures. load (kgf) csd (mm) pure jo jo+1% gar jo+3% gar jo+5% gar 50 0.36 0.37 (lnsl) 63 0.39 0.51 0.37 (lnsl) 80 0.42 1.73 0.98 0.42 (lnsl) 100 0.46 2.04 1.77 1.03 0.48 (lnsl) 126 0.5 2.23 2.46 1.66 0.91 160 0.54 w.p 2.55 2.12 2.08 200 0.59 2.93 2.29 2.12 250 0.63 w.p 3.68 3.14 315 w.p w.p figure 6: load wear index for different lubricant mixtures. zahid mushtaq and m. hanief 135 figure 7. fesem of wear scars on balls lubricated with (a) pure jo at 126kgf (b) jo + 5% gar at 200kgf. 3.4 3d topographic analysis the surface topographic analysis of wear scars was conducted using a 3d profilometer with a non-contact lens from rtec instruments, usa. the 3d view, texture pattern, and the roughness parameters (ra, rq) of the wear scars were also revealed. the roughness parameters were calculated at three different places of a wear scar, and then the three values were averaged. figure 8 shows the 3d view, texture pattern, and the roughness parameters of the wear scar on the test ball lubricated with pure jatropha oil at 126kgf load. figures 8(a) and 8(b) clearly show the deep crests and troughs depicting large wear and rough surface when the balls are lubricated with jatropha oil without additives. it means the lubricating film is breached, causing severe wear to the tribo-pairs. figure 8(c) gives the values of roughness parameters recorded at three different places of the wear scar. the averaged values of ra and rq are 7.663 and 10.7733, respectively. the high values of ra and rq signify the high roughness of the wear scar. figure 9 corresponds to the wear scar lubricated with jatropha oil + 5% garlic at 200kgf. after the intrusion of the garlic oil, the wear on the tribo-pairs is restricted up to a large extent, as evident from figures 7 (a, b). figure 9(a) displays a less rugged surface as compared to figure 8(a). also, the peaks and valleys observed in figure 9(b) are of lesser depth as compared to figure 8(b). figure 8: 3d topographic analysis of the wear scar lubricated with pure jo at 126kgf load (a) 3d image (b) texture pattern (c) roughness parameters. the corresponding average values of ra and rq are 4.966 and 6.274. after the addition of 5% garlic oil to jatropha oil, there was a reduction of 35.2% and 41.7% in ra and rq, respectively, representing a significant improvement in the surface roughness of the wear scar. the results of the 3d profilometer analysis depict that the addition of garlic oil to jatropha oil is beneficial in terms of reducing wear and the surface roughness of the wear scar. after the addition of garlic oil, a much smoother surface with fewer and smaller peaks, less ra and rq, is obtained. these results are in full agreement with the wear analysis on fesem, as shown in figure 7. exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil 136 figure 9: 3d topographic analysis of the wear scar lubricated with jo + 5% gar at 200kgf load (a) 3d image (b) texture pattern (c) roughness parameters. figure 10: eds analysis of wear scars lubricated with (a) pure jo (b) jo + 0.1% m+gr (c) jo + 5% gar. table 6. eds elemental composition of wear scars lubricated with different mixtures (weight %). element pure jo jo + 0.1% m+gr jo + 5% gar molybdenum 12.71 sulfur 0.24 2.6 5.18 carbon 3.2 8.15 1.3 oxygen 15.2 9.23 25.3 iron 79.57 66.4 63.88 silicon 0.321 0.298 0.28 chromium 1.22 1.12 1.15 phosphorus 0.02 0.32 2.79 3.5 eds analysis figure 10 shows the eds analysis of the wear scars lubricated with the jatropha oil with and without additives. the elemental composition present in the wear scars is given in table 6. it is observed from figure 10(a) that molybdenum is absent in the scar when it is lubricated with pure jatropha oil. table 6 zahid mushtaq and m. hanief 137 shows that the major components are iron and oxygen with small amounts of sulfur, silicon and carbon. figure 10(b) represents the eds of wear scar lubricated with jatropha oil and nano-additives. in figure 10(b), molybdenum is present, and the content of carbon is also increased due to the presence of graphene, as shown in table 6. this depicts the presence of molybdenum disulfide and graphene nano-additives necessary for the tribo-film formation. the eds of wear scar lubricated with jatropha oil mixed with garlic oil is presented in figure 10(c). molybdenum can again be found absent in this figure. while as, table 6 shows the reduced content of carbon and the presence of higher contents of sulfur, oxygen and phosphorus. this is a clear indication of the formation of oxides and sulfides on the surface of the wear scar due to the high content of sulfur in the garlic oil, which promotes the creation of the protective tribo-film and restricts the wear between the sliding surfaces. 3.6 flash temperature parameter (ftp) ftp is a criterion used to determine the critical temperature exceeding which the lubricating film suffers a breakdown at given conditions. it specifies the effectiveness of the lubricant. the higher the value of ftp, the stronger the lubricating film and the better the lubricant. it depends on the wsd and the load and can be calculated following the ip-239 standard, 1986, (lane, 1957) from the equation below ftp = applied load / (wsd) 1.4 the ftp values observed for different lubricants at different loads are compiled in table 7. it is observed that the ftp of the additized oils is greater than the base oil. the ftp of jatropha oil increases from 36.85 to 279.42 after the addition of 5% garlic oil at 100 kgf. hence, it can be outlined that the addition of garlic oil elevates the ftp of the jatropha oil and boost the lubricity. table 7. flash temperature parameter calculations. load (kgf) pure jo jo+1% gar jo+3% gar jo+5% gar 50 201.13 209.00 63 161.71 253.43 235.42 80 37.13 82.29 269.48 100 36.85 44.96 95.94 279.42 126 40.99 35.73 61.97 143.78 160 43.14 55.87 57.38 200 44.40 62.69 69.84 250 40.34 50.37 4. conclusion the following conclusions can be drawn from the current study: 1. graphene and mos2 nano-additives reduce the wear scar diameter of the test balls but do not improve the extreme pressure properties of the jatropha oil. the lowest wear scar is observed with 0.1% of the nano-additive concentration. 2. garlic oil improves the extreme pressure properties of jatropha oil by elevating the lnsl and weld point, increasing the lwi, and reducing the wear of balls at high loads. the lnsl of the jatropha oil was increased by 1 to 3 levels, while the weld point was elevated by 2 to 3 levels after the addition of garlic oil. the lwi of jatropha oil increased from 26.671 to 51.177. 3. the wear analysis on fesem revealed that adding garlic oil to jatropha oil minimized the wear and produced smoother surfaces at 200kgf load as compared to the surfaces lubricated with pure oil at 126kgf load. the 3d topographic analysis showed that the inclusion of garlic oil in the jatropha oil lowered the wear and surface roughness. it produced much smoother surfaces and decreased the values of ra and rq. 4. eds analysis of the wear scars confirmed the presence of nano-additives and the formation of the tribo-film on the surface. the presence of sulfide and oxide films was detected on the wear scars lubricated with jatropha oil mixed with garlic oil. 5. the ftp of jatropha oil was elevated to a large extent after the addition of garlic oil. conflict of interest the authors declare that there are no conflicts of interest regarding this article. funding the authors did not receive any funding for this research. acknowledgment the authors would like to acknowledge the help and support of tribology lab, mechanical engineering department, national institute of technology, srinagar, j&k, india. exploring the anti-wear and extreme pressure lubricating performance of jatropha oil modified with nano-additives and garlic oil 138 references berman d, erdemir a, sumant av (2014), graphene: a new emerging lubricant. materials today 17(1): 31–42. cermak sc, biresaw g, isbell ta, evangelista rl, vaughn sf, murray r (2013), new crop oilsproperties as potential lubricants. industrial crops and products 44: 232–239. chan ch, tang sw, mohd nk, lim wh, yeong sk, idris z (2018), tribological behavior of biolubricant base stocks and additives. renewable and sustainable energy reviews 93: 145–157. chikezie aw, ossia cv (2018), enhancing the tribological properties of biodegradable oils by organic additives under extreme conditions. jurnal tribologi 19: 1–18. gasni d, chandra d, putra aa, fajri r (2019), effect of garlic oil as lubricant additive into coconut and palm oils on the physical and tribological properties. iop conference series: materials science and engineering 602(1): 12068. gong k, wu x, zhao g, wang x (2017), nanosized mos2 deposited on graphene as lubricant additive in polyalkylene glycol for steel/steel contact at elevated temperature. tribology international 110: 1–7. he j, sun j, meng y, pei y (2020), superior lubrication performance of mos2-al2o3 composite nanofluid in strips hot rolling. journal of manufacturing processes 57: 312–323. hernández-sierra mt, bravo-sánchez mg, báez je, aguilera-camacho ld, garcía-miranda js, moreno kj (2019), improvement effect of green lubricants on the tribological and mechanical performance of 4140 steel. applied sciences (switzerland) 9(22): 4896. jesbin ks, mahipal d (2021), evaluation of tribological characteristics of natural garlic oil as an additive in rubber seed oil. iop conference series: materials science and engineering 1114(1): 12052. kiu ssk, yusup s, soon cv, arpin t, samion s, kamil rnm (2017), tribological investigation of graphene as lubricant additive in vegetable oil. journal of physical science 28: 257–267. kotia a, rajkhowa p, rao gs, ghosh sk (2018), thermophysical and tribological properties of nanolubricants: a review. heat and mass transfer 54(11): 3493–3508. kumar sa, suresh g, hariprasad v, deepak gs, akhil ps (2021), enhancement of oxidative stability and cold flow properties of coconut oil using natural antioxidant additives for development of biolubricant. proceedings of the institution of mechanical engineers, part j: journal of engineering tribology 235(4): 758–764. lane tb (1957), the flash temperature parameter, a criterion for assessing ep performance on the fourball machine. journal of the institute of petroleum 43: 181–183. li w, jiang c, chao m, wang x (2014), natural garlic oil as a high-performance, environmentally friendly, extreme pressure additive in lubricating oils. acs sustainable chemistry and engineering 2(4): 798–803. lin j, wang l, chen g (2011), modification of graphene platelets and their tribological properties as a lubricant additive. tribology letters 41(1): 209–215. meng y, xu j, jin z, prakash b, hu y (2020), a review of recent advances in tribology. friction 8(2): 221–300. mushtaq z, hanief m (2021a), enhancing the tribological characteristics of jatropha oil using graphene nanoflakes. jurnal tribologi 28: 129– 143. mushtaq, z., & hanief, m. (2021b). evaluation of tribological performance of jatropha oil modified with molybdenum disulphide micro-particles for steel-steel contacts. journal of tribology, 143(2), 1–13. mushtaq z, hanief m (2021c), friction and wear performance of jatropha oil added with molybdenum disulphide nanoparticles 715-722. owuna fj, dabai mu, sokoto ma, dangoggo sm, bagudo bu, birnin-yauri ua, hassan lg, sada i, abubakar al, jibrin ms (2020), chemical modification of vegetable oils for the production of biolubricants using trimethylolpropane: a review. egyptian journal of petroleum 29(1): 75–82. panchal tm, patel a, chauhan dd, thomas m, patel jv (2017), a methodological review on biolubricants from vegetable oil based resources. renewable and sustainable energy reviews 70: 65–70. shafi wk, charoo ms (2021) an overall review on the tribological, thermal and rheological properties of nanolubricants. tribology materials, surfaces and interfaces 15(1): 20–54. tahir nam, abdollah mf bin, tamaldin n, amiruddin h, mohamad zin mr bin (2018), a brief review on the wear mechanisms and interfaces of carbon based materials. composite interfaces 25(5–7): 491–513. talib n, jamaluddin na, sheng tk, kiow lw, abdullah h, ahmad s, saleh a (2021), tribological study of activated carbon nanoparticle in nonedible nanofluid for machining application. evergreen 8 (2) 454-460. talib n, sani asa, hamzah na (2019), modified jatropha nano-lubricant as metalworking fluid for machining process. jurnal tribologi 23: 90–96. wang h, wang g, qi j, schandl h, li y, feng c, yang x, wang y, wang x, liang s (2020), scarcityweighted fossil fuel footprint of china at the provincial level. applied energy 258: 114081. woma ty, lawal sa, abdulrahman as, olutoye ma, ojapah mm (2019), vegetable oil based lubricants: challenges and prospects. tribology online 14(2): zahid mushtaq and m. hanief 139 60–70. wu h, wang l, johnson b, yang s, zhang j, dong g (2018), investigation on the lubrication advantages of mos2 nanosheets compared with zddp using block-on-ring tests. wear 394–395: 40–49. xu y, peng y, dearn kd, zheng x, yao l, hu x (2015), synergistic lubricating behaviors of graphene and mos2 dispersed in esterified bio-oil for steel/steel contact. wear 342–343: 297–309. xu y, peng y, you t, yao l, geng j, dearn kd, hu x (2018) nano-mos2 and graphene additives in oil for tribological applications. nanotechnology in oil and gas industries 151–191 yazawa s, minami i, prakash b (2014), reducing friction and wear of tribological systems through hybrid tribofilm consisting of coating and lubricants. lubricants 2(2): 90–112. zainal na, zulkifli nwm, gulzar m, masjuki hh (2018), a review on the chemistry, production, and technological potential of bio-based lubricants. renewable and sustainable energy reviews 82: 80–102. the journal of engineering research (tjer), vol. 18, no. 1, (2021) 52-61 *corresponding author’s e-mail: satishgudey5@gvpce.ac.in doi: 10.53540/tjer.vol18iss1pp52-61 enhanced exponential reaching law-based sliding mode control of shunt active power filter in an electrical distribution system vinay kumar naguboina1 and satish kumar gudey2,* 1 vignan’s institute of information technology, andhra pradesh, india, 2 gayatri vidyaparishad college of engineering (autonomous), andhra pradesh, india abstract: in this work, a three-phase shunt active power filter (shapf) is proposed to address the current related issues in a three-phase electrical distribution system (eds). a sliding mode controller (smc) and an enhanced exponential reaching law-based smc (eerl-smc) are proposed for a shapf to compensate for the load current. the controller’s performance is tested by injecting the current harmonics into the system. a non-linear load along with different loads on the distribution side is connected in parallel in a distribution network at the point of common coupling (pcc). modelling of the system is done using state-space analysis. the stability of the system is analyzed using the state feedback approach. the reference source currents are generated using the instantaneous pq theory. for variations in the load, the thd in the source current is realized. it is found that eerl-smc is more effective for a shapf in reducing the high-frequency oscillations and settling time for convergence. the source voltage and current waveforms are observed to be sinusoidal. both the controllers are effective in reducing the thd levels in the source current as per the ieee standards. a comparison between the controllers is presented in terms of settling time, thd in source current. pscad v4.6 is used for simulation works. keywords: electrical distribution system (eds); shunt active power filter (shapf); point of common coupling (pcc); total harmonic distortion (thd); sliding mode controller (smc); enhancement exponential reaching law (eer-law). القدرة الفعالة المتوازي في لمرشحوحدة التحكم باإلنزالق المبنیة على قانون الوصول األسي المحسن شبكة التوزیع الكھربائي ,*2و ساتیش كومار جودي 1فیناي كومار ناغوبوینا تصفیة لحل مشاكل شبكات توزیع الكھرباء ذات الثالثة ة أوجھمقترحا الستخدام مرشح للقدرة الفعالة ذو ثالث بحثقدم ھذه الی الملخص: قترح استعمال وحدة التحكم بوضع االنزالق بنسختیھا العادیة والمبنیة على قانون الوصول األسي المحسن لتعویض تیار یأوجھ، كما ن النظام بتوافقیات التیار. تم توصیل الحمل في مرشح للقدرة الفعالة المتوازي ذو ثالثة أوجھ، حیث یتم اختبار اداء وحدة التحكم بحق حمل خطي وأحمال اخرى مختلفة بالتوازي في شبكة توزیع الكھرباء عند نقاط الربط المشتركة، ووضع نموذج النظام عن طریق اء تیارات التمثیل المصفوفي للمعادالت التفاضلیة، كما تم تحلیل استقرار النظام عن طریق نھج التغذیة الراجعة للحالة، ویتم انش المصدر المرجعي باستخدام النظریة المعممة للقوة التفاعلیة اللحظیة في الدوائر ثالثیة الطور، كما تحقق تشوه توافقي كلي في التیار المصدري نتیجة الختالف الحمل. اشارت نتیجة الدراسة الى ان وحدة التحكم باالنزالق المبنیة على قانون الوصول األسي المحسن أثیرا اكبر في تقلیل ذبذبات التردد العالي وزمن السكون في مرشح القدرة الفعالة ذو ثالثة أوجھ، كما لوحظ أن جھد المصدر تعطي ت وأشكال موجة التیار ھي جیبیة بطبیعتھا. نجح كال من وحدتي التحكم باالنزالق في تقلیل التشوه التوافقي الكلي في مصدر التیار كما مھندسي الكھرباء واإللكترونیات. كما تم في ھذه الورقة عرض مقارنة بین نوعي وحدات التحكم باالنزالق ورد في معاییر معھد ُوضح من خاللھا زمن السكون والتشوه التوافقي الكلي في مصدر التیار. .تشوه توافقي؛ وحدة التحكم باإلنزالق ؛نقاط الربط المشتركة؛ ذو ثالثة أوجھ قدرة الفعالةتوازي للال مرشح؛ شبكة توزیع الكھرباءالكلمات المفتاحیة: vinay kumar naguboina and satish kumar gudey 53 1. introduction the electrical distribution system (eds) consists of different types of loads like linear loads, non-linear loads and sensitive loads. at the point of common coupling, all these loads are connected in parallel. among these loads, non-linear loads inject current harmonics into the electrical distribution system (eds). this increases the total harmonic distortion (thd) percentage in current at the pcc. this will cause a huge amount of impact on the performance of both sensitive loads and linear loads and may also damage these loads. the usage of non-linear loads like computers, televisions, printers etc. has been increasing day by day and hence, increases nonlinearity in source current. these non-linear currents will adversely affect the system voltage. if such voltages are supplied to the loads then the system performance will be drastically affected (thentral, t. m. t. et al., 2021). the problem of non-linearity is mainly because of the non-linear loads connected at pcc. these loads have to be isolated from the linear and sensitive loads to prevent the voltage and current distortions, which is not practically possible. hence, proper preventive measures have to be taken to protect these linear and sensitive loads from the source voltage and current distortions. this has led the researchers to design compensating devices that will protect these loads. passive filters are conventional filters that can be designed to protect the loads from the harmonics. the development of power semiconductor devices and signal processing devices and availability at reduced cost have attracted researchers to work with active filters (af). af can perform numerous functions like harmonic filtering, damping, voltage regulation, load balancing and other power quality issues arising in a distribution system. af’s are further categorized as pure active power filters (paf) and hybrid active power filters (haf). paf’s mainly consists of only one single voltage source converter fed with a dc capacitor. haf includes multiple or single voltage source pwm converters with passive filters like inductor and capacitor and/or resistors. generally, for high power applications, haf is more commonly used for harmonic mitigation in terms of performance and cost. in this work, a shunt active power filter for harmonic mitigation is presented in a distribution system with the mathematical model and stability analysis. the main contributions of this work are as follows (i) a shapf has been designed and simulated to protect the sensitive loads from current distortions. (ii) enhancement exponential reaching law (eerlaw) is added to the conventional smc controller to improve the performance of the system. (iii) a comparative analysis is performed in thd’s of the source current with different loads. section ii discusses the passive filters used for harmonic mitigation. 2. passive and active power filters there are different kinds of passive filters like low pass filters, high pass filters, single tuned filters, double-tuned filters etc. these filters consist of passive elements like resistors, inductors and capacitors (s. m. mozayan et al. 2016). as shown in fig. 1, r, l, c values should be calculated based upon the formulas given in (1). 𝐶𝐶 = 𝑄𝑄𝑐𝑐/(6.28 𝑉𝑉𝑉𝑉2) 𝑋𝑋 = 1/(6.28 𝑉𝑉ℎ𝐶𝐶) 𝐿𝐿 = 𝑋𝑋/ 2 ℎ𝑉𝑉 (1) 𝑄𝑄 = 6.28𝑉𝑉𝐿𝐿/𝑅𝑅 𝑅𝑅 = 1/6.28𝑉𝑉𝐶𝐶 where qc= reactive power of filter (mvar), v= supply voltage (v), q = quality factor, h = tuning harmonic order of the filter, f is the power frequency. the tuned low pass filter and high pass filters will eliminate the harmonics based upon their cut-off frequency. the designed single tuned filter will eliminate the particular harmonic for which it is tuned. the dominant frequency harmonic component has to be identified in supply voltage or current. this can be done by using fast fourier transforms (fft) analysis. the filter has to be designed in such a way as to eliminate that harmonic component in voltage or current. the designed filter has to be connected in parallel to the load so that it acts as a conductance path for that harmonic component. similarly, it should also act as a high impedance path for the remaining harmonic components. thus, the harmonic components are deviated through passive filters without reaching the loads. thus, the percentage of thd in load voltage and current will be decreased. thus, by using the single tuned filters the thd can be reduced. similarly, if the two frequency components are found to be dominant then double-tuned filters are preferred. these filters are designed to eliminate the two harmonic frequency components. the main drawback of passive filters is the presence of resonance between the line and the filters. if the impedance of the designed filter and system impedance is equal then the system is said to be in resonance condition. this will inject the abnormal disturbances line noise into the system. the basic drawback of these passive filters is that they can eliminate only the single or double frequency components. for any variations in the non-linear loads enhanced exponential reaching law-based sliding mode control of shapf in an eds 54 that are in one condition, the harmonics occurring in the system will change dynamically. this made the researchers shift their focus towards the devices, which can provide dynamic and effective solutions (alirezajavadi et al. 2017). among the available different active power filter (apf), shunt active power filter (shapf) and series active power filters (seapf) will function dynamically. these filters can provide effective solutions. the shunt active power filter can mitigate the current harmonics and the series active power filter acts as a voltage regulator. unlike the traditional passive filters, apf will have the flexibility to provide multiple functions like eliminating multiple harmonic frequency components, injection of reactive power, correction of power factor, regulation of voltage and voltage flicker reduction etc. also due to the decrement in the manufacturing cost of power semiconductor devices and signal processing devices manufacturers have shown more interest in apf’s (alireza javadi et.al 2016). however, the manufacturing cost of apf’s is quite high when compared with the conventional passive filters. apf’s are mainly classified into two types based upon their application. they are single-phase apf and 3-φ apf. however, the usage of single-phase apf is only restricted to low-power applications. this made the researchers search for the device, which can be used in high power applications i.e. three-phase apf. in this paper, the performance of the three-phase shunt apf is analyzed through simulation works. 3. shunt active power filter these type of filters consists of active power filter (apf) connected in shunt across the load. in this active power filter, the voltage source currentcontrolled converter is used to inject the compensating current to suppress the distortions present in load voltage and load current (sharma, s. et al. 2020). the distortions in load voltage and current are calculated separately and then the resultant total error is calculated by summing the voltage error and current error. the compensation is done by injecting the compensating current, which will suppress the distortions in voltage and current. figure 2 shows the schematic representation of the shapf. it consists of an ac power supply feeding a non-linear load. a shunt apf consisting of a voltage source inverter (vsi) connected in shunt at pcc to inject a shunt current to provide compensation against current harmonics. ls and ll are the source side and load side inductances, which play a key role in the compensation. along with the sensitive loads, nonlinear loads are always present in the distribution network lse is the series winding inductance to the vsi and vdc is the input dc voltage supply. in the circuit shown in figure 3, a three-phase shunt active power filter (shapf) is presented with a sliding mode controller (smc), which consists of a current-controlled vsc. this converter should insert the compensating current if needed. the compensating current will be with a phase shift of exactly 180 degrees to the distortion that occurred in the source current. now to inject this compensating current the current-controlled voltage source converter should function dynamically and should be turned on instantly (javadi, a. et.al 2017), (javadi, a et.al 2016). for this purpose, a controller should be designed in such a way that it should generate the firing pulses during the need for compensation. (a) (b) figure 1. (a) single tuned passive filter, and (b) double tuned passive filter. figure 2. schematic representation of shapf. figure 3. block diagram of three-phase shunt active power filter. vinay kumar naguboina and satish kumar gudey 55 4. smc with pq theory figure 4(a) shows the smc controller with a linear sliding surface. smc is one of the robust non-linear controllers. it was used in many practical applications because of its simple construction and its accuracy. this smc can vary the reference signals and can easily track the control signals dynamically when compared with other controllers. this smc design mainly consists of two parts. one is choosing the sliding surface and the other is fine-tuning the sliding coefficients. in this proposed work, an smc with a linear sliding surface is chosen and the sliding coefficient k1value is chosen as 100. the equation for a linear sliding surface is presented in (4). this smc is used to generate the firing pulses to the converter, which will inject the required compensating current. this compensating current is exactly a 180˚ phase shift to harmonics in the source current. the instantaneous pq theory is used to calculate the error in the source current. the actual and the reference values of the three-phase voltages and currents are converted into stationary reference frames. then these voltages and currents are multiplied together to generate the actual and reference values of active and reactive power. now the resultant error in powers is calculated by analyzing the real values with the reference values. the obtained perror and qerror are converted into their respective phase current errors by using the inverse instantaneous pq theory, (ma, h. et al., 2017). thus, these obtained error signals are given to the pulse width modulator to generate the firing pulses at the desired instant. the switching frequency chosen is 10 khz. fig. 4 (b) shows the controller block diagram. 𝑃𝑃 = 𝑉𝑉𝛼𝛼𝐼𝐼𝛼𝛼 + 𝑉𝑉𝛽𝛽𝐼𝐼𝛽𝛽 (2) 𝑄𝑄 = 𝑉𝑉𝛽𝛽𝐼𝐼𝛼𝛼 − 𝑉𝑉𝛼𝛼𝐼𝐼𝛽𝛽 (3) 𝜎𝜎𝑠𝑠 = 𝑘𝑘1(𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟 − 𝐼𝐼𝑎𝑎𝑐𝑐𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎) (4) (a) (b) figure 4. (a) linear sliding surface σs., and (b) schematic diagram of the controller for shapf. enhanced exponential reaching law-based sliding mode control of shapf in an eds 56 5. modeling of shunt active power filter using statespace analysis to determine the stability of shapf, the state-space model is derived using state-space analysis. figure 5 shows the equivalent circuit of shapf. the voltage across the load (vl), current injected by the shapf (ise) and current through the power factor correction capacitor (ic) have been considered as state variables. the controller can be designed by considering any one of the state variables as a controlling parameter. in this proposed controller the current through the power factor correction capacitor is considered as the controlling parameter (m. h. rashid, 2011). the frequency response characteristics of the designed system are obtained and its stability margins are analyzed. �̇�𝑥 = 𝐴𝐴𝑥𝑥 + 𝑏𝑏1𝑢𝑢 + 𝑏𝑏2𝑉𝑉𝑠𝑠 + 𝑏𝑏3𝑖𝑖𝐿𝐿 (5) 𝑦𝑦 = 𝐶𝐶𝑥𝑥 (6) the state feedback approach is used to obtain the frequency response characteristics. the state model is represented in (7). the transfer function obtained for the system considered in the open-loop condition is represented in (8). � 𝑣𝑣�̇�𝐿 𝚤𝚤̇̇𝑐𝑐 𝚤𝚤𝑠𝑠�̇�𝑟 � = ⎣ ⎢ ⎢ ⎢ ⎡0 1 𝐶𝐶 0 0 − 1 𝐿𝐿𝑠𝑠 − 1 𝐿𝐿𝐿𝐿 − 𝑅𝑅𝑠𝑠 𝐿𝐿𝑠𝑠 1 0 − 𝑅𝑅𝑠𝑠𝑠𝑠 𝐿𝐿𝑠𝑠𝑠𝑠⎦ ⎥ ⎥ ⎥ ⎤ � 𝑣𝑣𝐿𝐿 𝑖𝑖𝐶𝐶 𝑖𝑖𝑠𝑠𝑟𝑟 � + � 0 0 𝑉𝑉𝑑𝑑𝑑𝑑 𝐿𝐿𝑠𝑠𝑠𝑠 �𝑢𝑢 + � 0 1 𝐿𝐿𝑠𝑠 0 �𝑉𝑉𝑠𝑠 + � 0 − 𝑅𝑅𝑠𝑠 𝐿𝐿𝑠𝑠 + 𝑅𝑅𝐿𝐿 𝐿𝐿𝐿𝐿 0 �𝑖𝑖𝐿𝐿 (7) 𝐺𝐺(𝑠𝑠) = 𝐶𝐶(𝑠𝑠𝐼𝐼 − 𝐴𝐴)−1𝑏𝑏1 (8) where a is the state matrix of order 3x3, b1 is the input matrix of order 3x1, b2 is the source voltage matrix of order 3x1, b3 is the load current matrix of order 3x1, c is the output matrix i.e. c= [0 1 0]. from the gain margin and phase margins acquired from the bode plots, it can be concluded that the system is stable. the transfer function obtained is given in (9). figure 5. equivalent circuit for shunt apf. 𝐺𝐺(𝑠𝑠) = −1𝑟𝑟08𝑠𝑠 𝑠𝑠3+5.002𝑟𝑟05𝑠𝑠2+3.535𝑟𝑟−09𝑠𝑠+2.193𝑟𝑟06 (9) it is found that a gm of infinity and a pm of 90 degrees is obtained for the designed system, which can be considered as a stable system with high stability margins. the controller works effectively with the chosen state variable. table 1 represents the configuration parameters. the rating of the shapf is 1.431 kva for a threephase distribution system. 6. simulation results the work considers different types of loads (like linear, non-linear and sensitive loads) in the electrical distribution network for simulation analysis. the current drawn from the source is shown in fig. 7. the non-linear load representation is shown in fig. 8(a). all three phases of the source current are said to be non-linear. the nonlinear load considered is a threephase thyristor bridge rectifier with different combinations of loads on the dc side (ouchen, s et al. 2021), (satish kumar gudey et al. 2014). the magnitude of the load current is 10 a (peak/phase). table 1. system parameters. symbol definition value vs supply voltage (l-n) 141.4 (max) f supply frequency 50 hz ls line inductance 10 µh rs line resistance 0.5 ω rnon,lnon non-linear load 10 ω, 16 mh rl, ll load resistance, inductance 20 ω, 0.05 mh lse switching ripple filter inductance 5 mh filter power factor capacitor capacitance 228 µf pf load power factor 0.8 lag figure 6. frequency response of single-phase shapf. vinay kumar naguboina and satish kumar gudey 57 a high nonlinearity is observed in source current in three phases (a, b and c). thd percentage of 16.7753 is observed in the source current waveform in phase a. according to the ieee standards, it is not safe to operate the linear loads and sensitive loads supplying this non-linear current (satish kumar gudey et.al 2014), (satish kumar gudey et.al 2015). hence, an appropriate filter should be tuned to inject the compensating current with exactly 180 degrees phase shift to nullify this harmonic current. fig. 8 (b) shows the harmonic spectrum of the non-linear source current. 6.1 rectifier with r-load assume that linear and non-linear load (assume that there will be a three-phase bridge rectifier connected to a resistive load) is connected at the pcc. the load voltages and load currents at the linear load are as shown in fig. 9. the magnitudes of the load voltage and current are 100 v (rms), 15 a (rms). figure 10 shows the current through the non-linear load, the compensated current, current drawn from the source. it is observed that the source current thd is 0.75%, which is compensated by the shapf. thus, the polluted thd content is reduced by using the shapf and is kept under the regulations imposed by the ieee standards. 6.2 rectifier with rl-load now with the rectifier rl load, the thd of the source current is observed to be 22.44%. now the shapf is connected to the circuit and simulated. the threephase voltage at load and phase currents through the linear load is shown in fig. 12. figure 12 shows the waveforms for load voltage and current when the systems are fed with a rectifier with rl-load. the magnitudes of the load voltage are 100v (rms) and the magnitude of load current at the pcc is 10.25a (rms). figure 7. the waveform of the source current is in phases a, b and c. (a) (b) figure 8. (a) three-phase thyristor bridge rectifier as non-linear load, and (b) thd spectrum in the source current. figure 9. three-phase waveforms of (a) voltage at load, and (b) linear load current. figure 10. the waveform of current drawn by the nonlinear load inl, shunt apf current iaf, source current is. figure 11. thd in load voltages and load currents. figure 12. three-phase waveforms of voltage at load and phase currents drawn by the linear load. enhanced exponential reaching law-based sliding mode control of shapf in an eds 58 figure 13. waveform of current drawn by the non-linear load inl, shunt apf current iaf, source current is. (a) (b) figure 14. (a) thd in three-phase voltages and three-phase currents through the linear load, and (b) phase plane projection of sliding surface σsand its derivative through simulation. table 2. source current thd with a non-linear load connected firing angle of 36˚. s. no. resistance (ω) inductance (mh) % of thd in source current 1 16 16 1.21 2 16 20 0.88 3 20 16 1.06 4 20 24 0.75 5 24 24 0.75 from fig. 14(a) it is clear that the designed shunt apf is injecting the compensating current to maintain the thd in the load current within the permissible limits.smc along with the shapf is said to work well in compensating the current harmonics. figure 14 (b) shows the locus of the sliding surface σwith its derivative. the system reaches origin within a finite period of 3 ms with high-frequency oscillations i.e. chattering is present in smc (vinay kumar naguboina et.al 2018). the chattering and hence, the settling time can be reduced by using an enhanced exponential sliding mode controller (eerl-smc) as discussed in section vii. 7. eer-law-based sliding mode control for shapf the chattering effect is the main drawback of the conventional smc’s. this had led the researchers to concentrate on higher-order smc, which can avoid this chattering effect. in this paper the enhanced exponential reaching law is added to the existing conventional smc to decrease the settling time, steady-state error, reducing response time (quoc-nam trinh, et al. 2014): hence, eerl based smc can be used as an alternative instead of going for the higherorder smc, which is presented in this proposed work. here in this case the sliding surface s chosen is as shown in (10). exponential reaching will have a high adaptive function when it is compared with the traditional controllers. the error signal obtained at the sliding surface will be given to the eerl as shown in fig. 15. the equation of eerl-smc is as shown in equation (10). 𝑠𝑠 • = −𝜆𝜆𝑠𝑠 − � 𝜉𝜉 𝐷𝐷(𝑆𝑆) |𝑠𝑠|𝛾𝛾. 𝑠𝑠𝑠𝑠𝑠𝑠( 𝑠𝑠)� (10) here 𝐷𝐷(𝑠𝑠) = 𝛼𝛼 + (1 − 𝛼𝛼)𝑒𝑒−𝛽𝛽𝑥𝑥|𝑠𝑠| > 0 the stability of the system will not depend on the value of d(s). the values of λ , ξ and xβ are always positive integers. their values are in the range of 0 and 1. the time for reaching the system depends on s . as the value s decreases, the chattering effect can be minimized. the sαβ obtained is passed through (10) and converted to sabc, which is then compared with a repetitive waveform operating at 10 khz to create the triggering pulses to the vsi circuit. the simulation results were obtained by adding the exponential reaching law (vinay kumar naguboina et.al 2018), (mozayan et.al 2016), (nayak v et.al 2020) to the presented smc is given in sections 7.1 and 7.2. 7.1 non-linear load (r-load) in this simulation, only the non-linear load of 15a (max. value) is connected at the pcc. then the resultant current drawn from the source is observed. vinay kumar naguboina and satish kumar gudey 59 figure 15. waveforms of load, current injected by the shunt apf and source current. figure 16. three-phase voltages at the load. figure 17. thd in load voltage and source current. figure 18. the waveform of the current is drawn by the non-linear load for a firing angle of 90 degrees, current injected by the shunt apf, source current. figure 19. waveforms of load voltage. from simulation results shown in figs. 15,16, and 17, it is clear that the source current is free of harmonics. the harmonic content in the source is reduced to 2.27%, which is kept under regulation as per ieee standard. also, the thd in voltage at the pcc is within the allowable limits as per the standards. 7.2 non-linear load (rl-load) the performance of the designed controller with eerl-smc is analyzed by connecting the nonlinear load with rl-load (zobaa et.al 2014). in this simulation, a controlled rectifier is operated at a firing angle of 90 degrees. the resultant waveforms obtained is represented in figure 18. the designed filter is tested by changing the load parameters. the non-linear load (controlled rectifier fed rl load) is connected at the pcc. the controlled rectifier is operated at a firing angle of α=36˚. the variation in the percentage of thd in the source current is observed and is listed in table 4. (a) (b) figure 20. (a) switching pulses generated by the controller, and (b) thd in load voltage and source current. table 3. percentage of thd in load phase voltages va,vb,vc, source current. table 4. source current thd with non-linear load connected to rl-load operated at α =36˚. s.no. resistance (ω) inductance (mh) % of thd in source current 1 16 16 3.12 2 16 20 3.62 3 20 16 2.42 4 20 24 2.89 5 24 24 2.22 type of load va vb vc is r-load 0.0215 0.0293 0.02214 2.274 rl-load 0.0425 0.0428 0.0423 4.573 enhanced exponential reaching law-based sliding mode control of shapf in an eds 60 thus, by analyzing the simulation results it can be concluded that the designed shapf is having the capability to inject the compensating current whenever it is needed. thus, it can be concluded that the designed shapf can act dynamically and mitigate the current distortions, which occur in the system. the convergence plot of the controller is drawn by considering the sliding surface on the x-axis and its derivative on the y-axis. then the resultant plot obtained is as shown in fig. 21. figure 21. phase plane projection of sliding surfaces and their derivative through simulation. figure 22. load voltage and load current with only nonlinear load at pcc. figure. 23 source voltage and current. table 5. comparison of smc and eerl-smc for a shapf. s.no. parameters smc eerl-smc 1 % of thd in load current 0.053 0.056 2 % of thd in load voltage 0.21 0.22 3 settling time 3 ms 0.1 ms the resultant is converging towards the origin, which indicates that the system is stable. the convergence time taken by the controller with eerlsmc is quite less i.e. 0.1 ms when the phase plots of the controller with and without eerl are compared. thus, it can be concluded that the controller with eerlshows less response time of 0.1 ms when compared with smc, which converges in 3 ms. the chattering effect is also reduced as indicated in the waveforms obtained through simulation. figure 22 shows the load phase voltage of 100 v (rms) and load phase current of 3.97 a (rms) when a non-linear load is placed at pcc. the controller effectively improves the source current waveform to be sinusoidal as shown in fig. 23. table 5 shows the performance comparison of smc and eerl-smc. both of them work effectively to maintain the thd within permissible limits. the eerl-smc is better with less settling time during convergence. 8. conclusion in this article, a shunt active power filter in an eds performance is simulated by connecting the non-linear load at the pcc. both the conventional smc and eerl-smc are used to compensate for the effects of connected non-linear load at the pcc. it was identified that the proposed filter is functioning effectively to mitigate the polluted content in the source current. the current references are generated using the instantaneous p-q theory. the proposed controller for shapf is tested in three-phase systems by connecting r and rl elements on the dc side to the non-linear load. for the proposed system, the polluted content in voltage and current is reduced and is regulated by the controller. the thd obtained using both the controllers are kept under the limitations imposed by the ieee standards. it highlights the need for a reliable power supply for critical loads and with an increase in the number of non-linear loads. the stability of the system is analyzed using frequency response characteristics. it is found that the system is stable. the finite-time convergence is indicated through a phase plot, which realizes a faster settling time of 0.1 ms using eerlsmc compared to 3 ms in a conventional smc. the eerl-smc is an alternative to the higher-order smc for chattering reduction and finite-time convergence. the main applications of shapf are in pv systems, smart grid networks, distributed generation and machine control. conflict of interest the authors declare no conflict of interest. funding no funding was received for this study. vinay kumar naguboina and satish kumar gudey 61 references alireza javadi, lyne woodward, and kamal alhaddad (2017),: real-time implementation of a three-phase thseaf based on vsc and p+r controller to improve power quality of weak distribution systems”, ieee trans. on power electronics., 33(3): pp. 2073-2082. alireza javadi, handy fortin blanchette, and kamal al-haddad (2012): a novel transformerless hybrid series active filter, ieee trans. on power electronics,99, 5312-5317. alireza javadi, and kamal al-haddad, kouro, s (2015), a single-phase active device for power quality improvement of electrified transportation, ieee transactions on industrial electronics, 62, 3033-3041. ghosh, a, jindal, a. k. and joshi, a (2004), “design of a capacitor supported dynamic voltage restorer (dvr) for unbalanced and distorted loads,” power delivery, ieee transactions on, 19, pp. 405-413. gupta, r, ghosh, a., and joshi, a (2011), “performance comparison of vsc based shunt and series compensators for load voltage control of distribution systems”, in ieee transactions on power delivery, 26(1). thentral, t. m. t. et al., (2021) “development of control techniques using modified fuzzy based sapf for power quality enhancement,” in ieee access, 9, 68396-68413. javadi, a., hamadi, a, ndtoungou, a and al-haddad, k (2017): power quality enhancement of smart households using a multilevel-thseaf with a pr controller, ieee transactions on smart grid., 8, 465-474. javadi, a, lyne woodward and al-haddad, k, (2016) “experimental investigation on a hybrid series active power compensator to improve power quality of typical households”, ieee transactions on industrial electronics, vol. pp. mozayan, s. m., saad, m., vahedi, h., fortinblanchette, h. and soltani, m., (2016) “sliding mode control of pmsg wind turbine based on enhanced exponential reaching law,” in ieee transactions on industrial electronics, vol. 63, no. 10, pp. 6148-6159. ma, h., li, y., xiong, z. (2019) “discrete-time sliding-mode control with enhanced power reaching law," in ieee transactions on industrial electronics, 66(6), 4629-4638. ma, h., wu, j., xiong, z. (2017)“a novel exponential reaching law of discrete-time slidingmode control”, ieee transactions on industrial electronics, 64(5), 3840-3850. nayak v and satish kumar gudey (2020) “an enhanced exponential reaching law-based smc for a ups system in industrial applications,” in serbian journal of electrical engineering, vol. 17, no. 3, october 2020, 313-336. ouchen, s., benbouzid, m. blaabjerg, f. betka, a. and steinhart, h. (2021) “direct power control of shunt active power filter using space vector modulation based on supertwisting sliding mode control,” in ieee journal of emerging and selected topics in power electronics, 9(3), 32433253 quoc-nam trinh, hong-hee lee (2014): improvement of unified power quality conditioner performance with enhanced resonant control strategy, iet gener transm. distrib., 8(12), 21142123. rashid m. h. (2011) power electronic converters handbook 3rd edition, elsevier, chapter-41, pp.1215-1216. satish kumar gudey and rajesh gupta (2014), “sliding mode control in voltage source inverterbased high-order circuits, international journal of electronics, vol. 102, no. 4, pp. 668-689. satish kumar gudey and rajesh gupta (2014), “sliding mode control of dvr for mitigation of source side voltage disturbances”, international conference on standards for smart grid ecosystem, 6-7 march bangalore. satish kumar gudey and rajesh gupta (2015) “reduced state feedback sliding-mode current control for voltage source inverter-based higherorder circuit”, iet power electron., 8 (8), 13671376. sharma, s, verma, v and behera, r. k. (2020) “realtime implementation of shunt active power filter with reduced sensors,” in ieee transactions on industry applications, 56(2), 1850-1861. vinay kumar naguboina, satish kumar gudey (2018), “a compensation strategy using thseaf in a distribution system based on smc for source side and load side control” – ieee international conference on computing, power and communication technologies – 27th & 28th september, galgotias university, vol.pp.433-438. vinay kumar naguboina, satish kumar gudey, (2018), “enhanced exponential reaching law smc for analyzing pq issues in a distribution system using thseaf” – indian council international conference (indicon) – 15thto 18th december, iit madras, vol.pp.1-6. zobaa, a. f. (2014), “optimal multiobjective design of hybrid active power filters considering a distorted environment”, ieee trans. ind. electron., 61, pp. 107-114. 1. introduction 2. passive and active power filters 3. shunt active power filter 4. smc with pq theory 5. modeling of shunt active power filter using state-space analysis 6.1 rectifier with r-load 7. eer-law-based sliding mode control for shapf 7.1 non-linear load (r-load) 7.2 non-linear load (rl-load) the performance of the designed controller with eerl-smc is analyzed by connecting the non-linear load with rl-load (zobaa et.al 2014). in this simulation, a controlled rectifier is operated at a firing angle of 90 degrees. the resultant waveforms obt... 8. conclusion conflict of interest funding c:\users\raoh\desktop\paper_3.xps the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 1. introduction a large number of environmentally related health studies has been conducted in different countries during the past several decades, which reveals that poor air quality contributes to increasing numbers of illnesses and deaths among adults and children (brunekreef and holgate, 2002). there has been growing public concern in kuwait in recent years, like other countries in the world, about health problems associated with deteriorating air quality. this is atleast partly because, kuwait has been __________________________________________ *corresponding author’s e-mail: nahla@alaswar.net exposed to various types of fine particles that may have originated from oil fires blamed for the ecological destruction caused by the iraqi invasion. it may also be because of and other particulates due to dry/wet deposition such as sulfate, nitrate and pops (persistent organic pollutants), which are derived chiefly from the combustion of fossil fuels in oil production, transportation, refining and power generation plants. fine particles are of the greatest health concern, especially when they are saturatexploring possible seasonal trends in mortality rate in relation to age, gender and environmental conditions in kuwait: a preliminary analysis n. al-rifaia*, a. elmi and a.r. khan college for women, department of environmental technology management, university of kuwait, p.o. box 5969, safat 13060, kuwait received 7 october 2009; accepted 6 january 2010 abstract: inherently, kuwait a high frequency of dusty events through out the year due to its arid climate. particulate air pollution has been linked to increased mortality due to various illnesses. for the purpose of improving health services, it is important to develop/establish an empirical basis for the possible link between mortality and air quality. in this paper, we examined mortality rates in kuwait in relation to age, sex and dust episodes during a five-year period. each year was divided into four seasons to detect any potential episodes in the death rate for a particular season. the results showed that the mortality rates did not correlate significantly (p = 0.67) with the occurrence of total dust events. contrary to the popular perception in kuwait that illnesses and, consequently, deaths increased in recent years as a result of deteriorating air quality due to particulate matter, the collected data did not show a consistent temporal trend. while this work represents an initial effort on an important issue upon which further investigations can be elaborated upon, a more detailed data have to be collected and analyzed. keywords: air quality, dust episodes, mortality, pollution 20 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 ed with hydrocarbons, because they can be inhaled deep into the lungs. respiratory and cardiovascular mortality were found to be related to pm10 (miller et al. 1979). coarse particles that are transported long distance with high wind speeds are also a health risk (ostro et al. 1999; schwartz et al. 1999; kwon et al. 2002). adverse effects of air pollution include an increase in cardiovascular and respiratory diseases leading to a high rate of deaths among elderly people (delucchi et al. 2002 and morris, 2001). a brief review of these studies is provided by chen et al. (2004). although there is strong evidence that particulate matter adversely affects human health, the precise proportion of mortality attributable to deteriorated air quality is difficult to quantify, due to complex interactions among many factors. for example, chen et al. (2004) studied the effect of asian dust storm events during spring on daily mortality in taipei, taiwan, during the period 1995-2000. they found no statistically significant relationship between dust storms and mortality rate. several other studies have also addressed the relationship between particulate air pollution and daily mortality at low levels of exposure (schwartz and marcus, 1990). schwartz (1991) analyzed the relationship between air pollution and daily mortality in detroit, for 1973 to 1982. mortality in philadelphia due to tsp and so2 was studied over the years 1973 to 1980 (schwartz and dockery, 1992). these workers reanalyzed mortality data from london for the winters of 1958 to 1972 and they found no evidence of a threshold in the relationship between black smoke and total daily mortality. therefore, it remains unclear the lowest level of black smoke above which significant relationships between air pollution and daily mortality could be found. to our knowledge, no specific published study has yet explored the link between deteriorating air quality and health problems in kuwait. furthermore, air pollution levels differ dramatically with respect to seasons. consequently, the estimated effect of air pollution on daily mortality might be different in different seasons. the primary objective of this study is to explore possible relationship between the four seasons and the mortality rates of different age groups during the period 1999-2005 in kuwait. although mortality rate studies data show the largest effects among elderly people (dominci et al. 2003), in the current study, all ages have been included to build up a general view of a preliminary investigation into whether there are any specific season(s) in which there is an increases in overall mortality in kuwait for both males and females. 2. data collection and analysis the data were collected from the publications of the department of statistics and medical records of the kuwait ministry of health, from1998-to-2005, as well as from the annual reports of population statistics issues (1998-2005) from the kuwait ministry of planning. particulate air pollution data were obtained from environmental public authority (epa) station located at the polyclinic in ali sabah al-salem residential area, about 47km south kuwait city. the data collected were presented as total number of death per month for eight age group sectors for kuwaiti male and female. metrological data were obtained from kuwait international airport for the entire period of concern. each year was divided into four seasons: winter (the three adjoining months of december, january, and february), spring (march, april, and may), summer (june, july, and august) and fall (september, october, and november). the seasonal monthly average was calculated and divided by the total kuwaiti population of males or females for each specific year to normalize the data. these normalized data are presented as number of deaths per million people. analysis of variance (anova) was performed separately on each season within a year. data were analyzed statistically as a factorial design with three main factors: age of 8 groups (<1-9, 10-19, 20-29, 30-39, 40-49, 50-59, 60-69, and 70->70), sex (male and female) and seasons (winter, spring, summer, and fall). when interactions of the main factors were significant, interactions rather than individual factors are discussed. similarly, when a main factor's effect was significant without any interaction, fisher's f test statistic was used to determine statistical significance within each main factor at a probability level of 0.05 ( = 0.05). once the data has been tested, preplanned lsd (least significant difference) multiple comparisons were used on the factors which were found to be significant to quantify the differences. all statistical analyses were conducted using the statistical analysis system, sas (2004) for windows. 3. results and discussion a summary of the statistical analysis of the results of the main and interaction effects is presented in table 1. the results show that the effects of all the main factors' significant except for seasons in 2002 and 2003. we found that interaction between age and gender are significant (p<0.05) for all the years investigated except 2005, with 2001 being statistically on the border (p = 0.06). also, another striking aspect of these findings is that genderand-season interaction was not significant in any year except 2004 (table 1). this is an indication that male and female mortality rates behaved similarly in relation to season. the result also shows that age-and-season interaction was significant for only 2000. this is consistent with the observations that elderly people are affected the most, especially during spring and winter seasons as discussed in detail in the following section. figures 1 to 4 present normalized death rates in kuwait for female and males from 1999 to 2005. the normalization was based on dividing the number of deaths for a particular age group in each year by the total population of 21 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 table 1. summary of the analysis of variance (p-value) statistical from 1999 2003 using glm procedure figure 1. death rates in kuwait for (a) female and (b) males during winter season from 1999 to 2005 age 22 figure 3. death rates in kuwait for (a) female and (b) males during fall season season from 1999 to 2005 age 23 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 that year. (for back calculation, we multiplied by a million). the effect of the growing population on the number of deaths has thus been eliminated. in general, it was observed that there was a higher death rate for males of age of 15-35 in all seasons than for females, the fall season of 2003, when the female death rate matched that of males. one of the most plausible explanations is the high number of deadly traffic accidents among the teenage male population. however, to verify this hypothesis, there is a need for further investigation. it was also obvious from our data that the death rate dropped for the older age group (beyond 70) for both male and female due to dwindling numbers of both sexes as expected. it appears that most people die at the age group 65-70 and a few died in the older age group as explained later. this study included all reasons of death, which were in some years high for fatal diseases, accidents, circulatory system, cancer, respiratory and heart diseases over other years. the number of deaths for the age group of 20-25 years was found to be higher than the average for males in the winter of 1999, 2002 and 2004 (figure 1b) in the spring of 2000, 2002 and 2004 (figure 2b), in the summer of 1999 and 2000 (figure 3b) and in the fall of 2002 (figure 4b). however, the death rate for tthe age group of 30-35 was higher than the mean only for males in the winter and summer of 2004 and fall of 2003 (figure 4b), and for females in the summer of 2003 (figure 3a). mortality occurred mostly among the elderly; it was noticed that total number of deaths for both males and females increased with age, the maximum being in the range of 60-70 years. this trend is true for all the seasons in each year, which is consistent with the literature (eg. delucchi et al. 2002). this may have been due to the generally poorer health condition of elderly people, which was exacerbated by deteriorating air quality. average deaths for this particular age group was higher for spring and winter of 2000 and then it declined in the following years. this decline in the mortality rate may be due to improved health care services, since we did not observe any noticeable decrease in air pollutants. however, the somewhat good news (decline in the mortality rate) did not last long as there was an increasing trend for total death towards 2002-2003 in most of the seasons, especially in winter. the pollution levels in winter are always high due to lower temperatures, high humidity and the low planetary boundary layer resulting in a build-up due to inadequate dispersion (khan and al-salem, 2007). based on the age groups the highest seasonal death rate figure 4. death rates in kuwait for (a) female and (b) males during summer season from 1999 to 2005 age 24 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 for males, is 44.62 in year 2000 for age group 70>age>60 while the second highest was 40.28 in year 2004 for the age group 70>age>60. for females the highest seasonal death rate was 29.44 in the year 1999 for the age group <1 year while the second highest is 29.23 in year 2001 for the age group <1 year, (fig. 1). in spring the average seasonal death rate for all ages was 260.9 for males (about 61.7% of the total reported deaths) and 161.7 for females (38.6%) based on age groups, the highest seasonal death rate for males, was 41.87 in year 2001 for the 75>age>65 age group while the second highest was 41.21 in year 1999 for the 65>age>55 age group. for females the highest seasonal death rate was 28.52 in year 2000 for the 70>age>60 age group while the second highest was 27.49 in year 2002 for the <1 year age group, (fig. 2). the average seasonal death rate for all ages was at a minimum in summer for both sexes: 250.8 males (about 60.7% of the total reported deaths) and 162.2 females (39.3%). for males, the highest seasonal death rate was 33.18 in year 2001 for the 65>age>55 age group while the second highest was 31.12 in year 1999 for the 75>age>65 age group. for females the highest seasonal death rate was 31.11 in year 2003 for the 75>age>65 age group while the second highest was 23.84 in year 2001 for the 75>age>65 age group, (fig. 3). in summer the frequency of dust storms is high. strong winds and high temperatures facilitate dispersion of pollutants. the demand for electrical power raises the power load to peak, forcing the power stations to operate in full capacity and thus emitting more pollutants. in fall the average seasonal death rate for all ages was 242.5 for males (about 61.3% of the total reported deaths) and 154.2 (38.9%) for females. for males, the highest seasonal death rate was 36.71 in year 2002 for the 75>age>65 age group while the second highest was 34.89 in year 2000 for the 70>age>60 age group. for females the highest seasonal death rate was 29.20 in year 2004 for the 75>age>65 age group while the second highest was 28.01 in 2003 for the <1yr age group, (fig. 4). overall, from the analysis of 5 years of data, it was noted that the average seasonal death rate for all ages was maximum in winter for both sexes: 297.6 males, (about 59.5%) and 202.9 females (about 40.5%) of the total reported deaths. the effects of poor air quality on daily mortality could be delayed for an unexpected period, depending on the initial health condition of individuals. since the number of days in which there were measurable dusty events (rising dust, suspended dust, dust storms, and haze) increased table 2. total number of haze, suspended dust, raising dust and dust storms days in kuwait for the period 1999 2005 per season table 3. annual total number of unstated age of death as compared to total number of deaths 25 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 gradually over time with 2003 having the highest (table 2), it was plausible to expect the death rate to continue to rise in subsequent years. figure 5 shows the frequency of dusty events as a function of season variation for the seven-year period 1999 to 2005. a maximum concentration of particulate matter was observed in the summer season of year 2005 and minimum concentration was in winter season of year 2004. in the summer season, the dust storms are more frequent as compared to any other season while in winter the frequency of dusty events is minimum, based on the data of these 7 years. by comparing table 2 with figs. 1 to 4, one would observe that the total number of days with haze, suspended dust, rising dust and dust storms was highest in summer and the lowest in winter, while the mortality rate was highest in winter and lowest in summer. this could be because the victims, especially the elderly died after a period of illness. in kuwait, pollutants concentrations, as reported by khan and salem (2007), and al-azmi et al. (2008), are generally higher in the winter season as compared to summer due to prevalent meteorological conditions such as low temperatures, slow winds and a lower planetary boundary layer (pbl) in winter. middleton et al. 2008 observed an increased risk of hospitalization at elevated levels of particulate matter and ozone on dusty storm days. similarly, perez et al. (2008) found convincing associations of mortality with course saharan dust (but not with local coarse dust). such hypothesis which relates the effect of dusty seasons or air quality to specific causes of death may require daily monitoring of hospital admissions and patient history. however, this is beyond the scope of this research and requires further investigation. the time series models require daily death rate data, which are not available. the death rate p>0.67 did not correlate significantly with the pooled environmental air quality data for all types of dusts for either females or males or for the total, in any of the investigated seasons. the lack of any significant association between particulate type of pollutants reported herein and mortality rates limited our ability to address the winter spring summer fall max 147.52 (2002) mean 105.32 man 25.13 (2004) max 329.5 (2005) mean 196.39 min 93.5 (2002) max 2429.7 (2005) mean 654.24 min 106.0 (2002) max 657.7 (2005) mean 212.22 min 73.7 (2004) dec jan feb mar apr may months jun jul aug sep oct nov figure 5. monthly dust measurement at um-alheaman station 26 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 precise effects of particulate air pollution on daily mortality, in relation to gender, season, and age. complex interactions among these factors might have masked any significant individual effects. the computation of correlation coefficients among age, season, gender, and air quality was complicated by the fact that there was a large number of deaths with unstated age at the time of death over all the years (table 3), indicating that full demographic information, such as date of death, were not recorded accurately. therefore, data on the age at the time of death are excluded from the analysis. more accurate recordkeeping of data such as date of patient sickness before death, reason of death, place of death etc. so that more detailed data analysis can be done. several studies conducted in the usa suggest that current levels of particulate pollution in urban air are associated not only with short-term but also with long-term increases in respiratory morbidity, mortality as well as lung cancer. for example, nevalainen and pekkanen (1998) evaluated the change in life expectancy assuming the long-term increase in mortality. they found the effect of particulate air pollution on life expectancy to be notable in countries with high cardiovascular mortality like usa. barceló et al. (2009) have reported deprivation to be associated in a statistically significant way with the geographical variation in mortality in the census tracts of the metropolitan area of barcelona, in the period 1994 to 2003. in our study however, there were no significant (p>0.05) temporal trends of increased/decreased death rates for either females or males for all the seasons investigated during five consecutive years (figs. 1-4), except in the 6570 age group. it should be noted however, that all death data presented in this paper are reported both inside and outside hospitals. schwartz, 2001 reported that airborne particles had greater impacts on deaths outside the hospital in chicago, suggesting that exposure to particles increases the risk of sudden deaths. vaneckova et al. (2009) reported that o3 and pm10 were found to be nonsignificant factors in the regions where air pollutants were studied. socio-economic status and the proportion of vegetation or developed land in each statistical local area (sla) were also not a significant factor explaining the increased mortality. trigo el al. 2009 have presented the impact of the european heat wave in 2003 on mortality in portugal. the mortality of women increased 79% while men had a 41% increases in all 18 districts of the country. prediction of death rates for various seasons is quite difficult because death can occur for any reason, particularly through unexpected accidents, a leading cause of death among the youth. it should be noted however, that the poor correlation between dusty days and mortality rates does not mean that air quality has no effect on mortality. the effect of other pollutants and the combination of particulate matter with other pollutants on all types of death due to respiratory illnesses will be the focus of our next research. conclusions the significance of air pollution in terms of public health has been recognized worldwide, and kuwait is no exception. the primary aim of this paper was to analyze available data to see whether or not there is an increasing trend in mortality in relation to season, gender and age sector. we found that this type of mortality occurred mostly among the elderly. a correlation analysis showed no significant (p>0.05) association between death rates and air quality indicators. given the inherent limitations of risk assessments, further studies are needed to support and quantify the relationship between mortality rate and air pollution under kuwait living conditions so that appropriate public health policies can be developed. we plan to carry out a more in-depth investigation of deaths that are directly related to pollution, such as specific death cases as a result of respiratory and lung cancer. acknowledgment we thank the department of statistics and medical records at kuwait ministry of health, and statistics and census sector at kuwait ministry of planning for providing us with the data. we also thank mr. kasem el-fadli of the directorate general of civil aviation, the meteorological department, the climatologically division and the environmental public authority (epa) for providing us with dust measurement data. finally, we are thankful to miss muneera al-qassar for her assistance in the data management. reference al-azmi, b. n., nassehi v. and khan, a. r., 2008, "comparative assessment of ambient air quality in rabia for year 2001 and 2004 in the state of kuwait," american journal environmental sciences, vol. 4 (1), pp. 50-62. barceló, m. a., saez, m. and saurina, c., 2009, "spatial variability in mortality inequalities, socioeconomic deprivation, and air pollution in small areas of the barcelona metropolitan region, spain," science of the total environment, vol. 407, pp. 5501-5523. brunekreef, b. and holgate, s.t., 2002, "air pollution and health," lancet vol. 360, pp. 1233-1242. chen, y.s., sheen, p.c., chen, e.r., liu, y.k., wu, t.n. and yang, c.y., 2004, "effects of asian dust storm events on daily mortality in taipei, taiwan," environmental research, vol. 95, pp. 151-155. delucchi, m.a., murphy, j.j. and mccubbin, d.r., 2002, "the health and visibility cost of air pollution: a comparison of estimation methods," journal of environmental management, vol. 64, pp. 139-152 dominci, f., aidan, m.c., zeger, s. and same, j., 2003, "airborne particulate matter and mortality: time 27 the journal of engineering research (2010) vol. 8 no. 1 (2011) 19-27 effects in four us cities," american journal of epidemiology, vol. 157, pp. 1055-065. khan, a. and al-salem, s., 2007 "seasonal variation effect on airborne pollutants in an urban area of the state of kuwait," journal of environmental research and development, vol. 1(3), pp. 215-218. kozlov, m.v., 2004, "retrospective analysis of the age at death in two heavily polluted and unpolluted russian towns," chemosphere vol. 56, pp. 405-410. kwon, h.j., cho, s.h., chun, y., lagarde, f. and pershagen, g., 2002, "effects of the asian dist events on daily mortality in seoul, korea," environmental research, vol. 90, pp. 1-5. middleton n., yiallouros p., kleanthous s., kolokotroni o., schwartz j., dockery d.w., demokritou p. and koutrakis p., 2008, "a 10-year time series analysis of respiratory and cardiovascular morbidity in nicosia, cyprus: the effect of short-term changes in air pollution and dust storms," environmental health, vol. 7 (39), pp. 1-16. miller, f.j., gardner, d.e., graham, j.a., lee, r.e. and bachmann, j.d., 1979, "size considerations for establishing a standard for exhalable particles," journal of air pollution control association, vol. 29, pp. 610-615. morris, r.d., 2001, "airborne particulates and hospital admissions for cardiovascular disease: a quantitative review of the evidence," environmental health perspectives, vol. 109, pp. 495-500. nevalainen j. and pekkanen, j., 1998, "the effect of particulate air pollution on life expectancy," the science of the total environment, vol. 217, pp. 137141. ostro, b.d., hurley, s. and lipsett, m.j., 1999, "air pollution and daily mortality in the coachella valley, california: a study of pm10 dominated by coarse particles," environmental research, vol. 81, pp. 231238. perez l., tobias, a., querol x., kunzli n., pey j., alastuey a., viana m., valero n., gonzalez-cabre m. and sunyer j., 2008, "coarse particles from saharan dust and daily mortality," epidemiology, vol. 19(6), pp. 800-807. pope, c.a, schwartz, j. and ransom, m.r., 1992, "daily mortality and pm10 pollution in utah valley," archives in environmental health, vol. 47, pp. 211217. sas institute inc., 2004. statistical analysis system (sas) software version 9.1.2, cary, nc: sas institute inc. schwartz, j. and marcus, a., 1990, "mortality and air pollution in london: a time series analysis," american journal of epidemiology, vol. 131, pp. 185-194. schwartz, j., 1991, "particulate air pollution and daily mortality in detroit," environmental research, vol. 56, pp. 204-213. schwartz, j., 2001, "is there harvesting in the association of airborne particles with daily deaths and hospital admissions," epidemiology, vol. 12, pp. 55-61. schwartz, j. and dockery, d.w., 1992, "increased mortality in philadelphia associated with daily air pollution concentrations," american review on respiratory diseases, vol. 4, pp. 600-604. schwartz, j., norris, g., larson, t., sheppard, l., claiborne, c. and koenig, j., 1999, "episodes of high coarse particle concentrations are not associated with increased mortality," environmental health perspectives, vol. 107, pp. 339-342. trigo, r. m., ramos, a. m., nagueira, p. j., santos, f. d., garcia-herrera, r., gouveina, c. and santo, f., 2009, "evaluation of impact of extreme temperature based indices in the 2003 heat wave excessive mortality in portugal," environmental. environmental science and policy, vol. 12, pp. 844854. vaneckova, p., beggs, p. j. and jacobson, c. r., 2010, "spatial analysis of heat-related mortality among the elderly between 1993 and 2004 in sidney, australia," social science and medicine, (in press). mech060513.qxd the journal of engineering research vol. 5, no.1 (2008) 7-19 __________________________________________ *corresponding author e-mail: asiri.net@gmail.com broadband vibration attenuation using hybrid periodic rods s. asiri department of mechanical engineering, king abdulaziz university, p.o. box 80204, jeddah 21589, saudi arabia received 13 may 2006; accepted 17 december 2006 abstract: this paper presents both theoretically and experimentally a new kind of a broadband vibration isolator. it is a table-like system formed by four parallel hybrid periodic rods connected between two plates. the rods consist of an assembly of periodic cells, each cell being composed of a short rod and piezoelectric inserts. by actively controlling the piezoelectric elements, it is shown that the periodic rods can efficiently attenuate the propagation of vibration from the upper plate to the lower one within critical frequency bands and consequently minimize the effects of transmission of undesirable vibration and sound radiation. in such a system, longitudinal waves can propagate from the vibration source in the upper plate to the lower one along the rods only within specific frequency bands called the "pass bands" and wave propagation is efficiently attenuated within other frequency bands called the "stop bands". the spectral width of these bands can be tuned according to the nature of the external excitation. the theory governing the operation of this class of vibration isolator is presented and their tunable filtering characteristics are demonstrated experimentally as functions of their design parameters. this concept can be employed in many applications to control the wave propagation and the force transmission of longitudinal vibrations both in the spectral and spatial domains in an attempt to stop/attenuate the propagation of undesirable disturbances. keywords: pass band, stop band, broadband vibration isolator, piezoelectricity, vibration attenuation, wave propagation, tunable filtering notation a cross-section area b width of piezo-insert cd elastic modulus dp electrical displacement of the piezo-insert e young's modulus ep electrical field intensity of the piezo-insert f force fi total interface force fn excitation harmonic force at the end of strut fp piezo-force hp piezo-coupling constant ájqh~dg úé¡dg ¿éñ° b ∫骩à°séh ¢ jôy ¥é£f ≈∏y ägrgõàg’g ∞«øîj …ò°ùy .¢s áá°°uuóóÿÿggá©hqg øe πµ°ûàj ƒ¡a ¬dhé£déh ¬«ñ°ûj øµá ωé¶ædg gòg .¢ jôy ¥é£f ≈∏y ägrgõàg’g ∫rgƒy øe g~j~l éyƒf á«ñjôéàdg h ájô¶ædg úà«méædg øe ω~≤j ¬bqƒdg √òg : hõ«ñdg ôkg §«°ûæàhh .»féhô¡c qõh ≥ë∏eh ò°üb ê~©e aõl øe á«∏n πc ¿ƒµàj h á∏kéªàe éjón áyƒª› øe ¿ƒµàj ájqh~dg ¿éñ° ≤dg .úmƒd úh π°üj ájrgƒàe áæé¡e ¿éñ° b é¡©æeh rgõàg’g øe ∞«øîàdg é¡æµá ájqh~dg ¿éñ° ≤dg ¿g åëñdg gòg í°vhcg.é©e á«dé©ÿg h ¢ øîæÿg ooîdg ägp ägrgõàg’g ∫õy ‘ ôkdƒe πµ°ûh ωé¶ædg πª©j ¿g øµá »féhô¡µdg ∫ón πø°sc’g ¤g ≈∏yc’g ìƒ∏dg ‘ rgõàg’g q~°üe øe …ô°ùj ¿cg øµá á«dƒ£dg äélƒÿg ωé¶ædg gòg ‘ .áhƒzôÿg òz qékb’g π«∏≤j ‹éàdéhh ooîdg äébé£f øª°v ¿éjô°ùdg øe ò«¨j ™e é«∏ª©e ájô¶ædg áægôh ” ~bh .á«lqéÿg ájrgõàg’g ägôkdƒÿg á©«ñ£d é≤ah ¬ª«ª°üj øµá äébé£ædg √òg ¢vôy .'∞bƒàdg ¥é£f ' ägooîdg äébé£f øª°v §≤a ¿éñ° ≤dg ‘ êéµÿgh »ø«£dg údéûg ‘ ägrgõàg’g qé°ûàfg ∫õ©d io~©àe äé≤«ñ£j ‘ ω~îà°ùj ¿g øµá ωƒ¡øÿg gòg .≥«ñ£àdgh ájô¶ædg úh ≥agƒj ∑éæg ¿cg ~lhh iôkdƒÿg πegƒ©dg .áhƒzôÿg òz äéhgô£°v’g qé°ûàfg ∞«øîj hcg ±é≤j’g ádhéfi ääggooôôøøÿÿggáá««mmééààøøÿÿgg .iôj驪∏d πhé≤dg í«°tîdg ,äélƒÿg ¿éjô°s ,ägrgõàg’g ¢ øn ,á«féhô¡c hõ«ñdg ,i~ÿg ¢ jôy rgõàg’g ∫réy ,∞bƒàdg ¥é£f ,qhôÿg ¥é£f : 8 the journal of engineering research vol. 5, no.1 (2008) 7-19 1. introduction the periodic rods in the proposed mechanical filter act as the transmission paths of the vibration from the upper plate where the source of vibrations is located to the lower plate. therefore, a proper design of these rods is essential to the attenuation of the vibratory energy and the noise radiated into any structure connected to the lower plate. a periodic rod consists of an assembly of identical cells connected in a repeating array which together form a 1-d periodic structure. the study of periodic structures has a long history. wave propagation in periodic systems related to crystals and optics has been investigated for approximately 300 years. brillouin (1953) developed the theory of periodic structures for solid state applications and then, in the early seventies, the theory was extended to the design of mechanical structures (mead, 1970; cremer, et al. 1973). since then, the theory has been extensively applied to a wide variety of structures such as spring-mass systems faulkne and hong, 1985, periodic beams (mead, 1970; faulkner and hong, 1985; mead, 1971; mead, 1971; mead and markus, 1983; roy, 1986; gupta, 1970; mead, 1986; richards and pines, 2003; mead, 1996), stiffened plates (gupta, 1970; mead, 1986; mead and yaman, 1991), ribbed shells (mead, 1987) and space structures. examples of such structures are found in many engineering applications. these include bulkheads, helicopter drive shafts (richads and pines, 2003), airplane fuselages, vehicle engine mounting systems (asiri, 2005), and helicopter gearbox supporting systems (asiri, et al. 2002). each such structure has a repeating set of stiffeners which are placed at regular intervals. sackman, et al. sackman, et al. (1999) presented a layered notch filter device that is limited only to high-frequency vibrations. such a filter which was developed theoretically based on the floquet theory is a periodically layered stack of two alternating materials with widely different densities and stiffnesses. the work presented here is an experimental implementation of the work of baz (2001) in which an active periodic spring mass system is employed theoretically, in a quasi-static manner, to control the wave propagation of longitudinal vibration. periodic rods in passive mode of operation exhibit unique dynamic characteristics that make them act as mechanical filters for wave propagation. as a result, waves can propagate along the periodic rods only within specific frequency bands called the "pass bands" and wave propagation is attenuated within other frequency bands called the "stop bands". the spectral width and location of these bands are fixed for a 1-d passive periodic structure, but are tunable in response to the structural vibration for active periodic structures (baz, kpc active piezo-stiffness due to the control gain k kps structural piezo-stiffness kd dynamic stiffness matrix kg control gain kij appropriately partitioned matrices of the stiffness kp total stiffness of the piezo-insert l element length la length of cell a lb length of cell b mij appropriately partitioned matrices of the mass n number of cells qp electrical charge sp strain of the piezo-insert tp thickness of piezo-insert t transfer matrix of the unit cell tk transfer matrix of the kth cell tp stress of the piezo-insert u longitudinal deflection v applied voltage y state vector = {ul fl}t yk eigenvetor of transfer matrix of the unit cell yk eigenvetor of the transpose of transfer matrix of the unit cell α logarithmic decay of amplitude of state vector β phase difference between the adjacent cells εs electrical permittivity λ eigenvalue of transfer matrix of the unit cell ρ density µ propagation constant ω excitation frequency 9 the journal of engineering research vol. 5, no.1 (2008) 7-19 2001; asiri, et al. 2004). the shape memory alloy has been used as a source of irregularities to attenuate the wave propagation in periodic rods (asiri, et al. 2004). the spectral finite element analysis and transfer matrix method (baz, 2001; asiri, et al. 2004; ruzzene and baz, 2000; baz, 2000) will be used to analyze the hybrid periodic rod and determine the propagation parameter, which indicates the regions of stop bands and pass bands. this paper is organized in four sections. in section 1, a brief introduction is given. section 2 presents the theoretical background of the hybrid periodic rod and section 3 demonstrates the performance characteristics of the tunable mechanical filter. comparisons between the theoretical and experimental characteristics are also presented in section 3. section 4 summarizes the findings and the conclusions of the present study and also outlines the direction for future research. 2. system modeling 2.1 overview in this section, the emphasis is placed on studying the dynamics of the tunable mechanical filter in order to demonstrate its unique filtering capabilities. the dynamics of one-dimensional hybrid periodic rods in their active and passive modes of operation are determined using the transfer matrix method. the basic characteristics of the transfer matrices of periodic rods are presented and related to the physics of wave propagation along these rods. the methodologies for determining the pass and stop bands as well as the propagation parameters are presented. in this paper, the focus is placed on hybrid periodic rods consisting of a straight rod with periodically placed piezoelectric inserts as shown in fig. 1. 2.2 spectral finite element method of the rod the first development of the spectral finite element (sfe) method occurred in the early eighties (patera, 1984). the main distinct between the sfe and finite element (fe) stems from the type of the shape functions used to approximate the model equations. in fe method, the shape function is only a function of a spatial variable, whereas in sfe the shape function has two independent variables: the spatial variable (x) and the spectral variable (ω). as a result, the sfe method is much more accurate than fe method. for example, one can show that the exact natural frequencies for a rod can be found by one element using sfe method, whereas a large number of elements needs to be used to get the same accuracy using the fe method. consider the rod shown in fig. 2 the equation of motion of the rod is given by: (1) where u is the longitudinal deflection, ρ is the density and e is young's modulus. then, assuming a solution u(x,t) = u (x) eiωt, reduces the equation of motion to: (2) using the following spectral shape function: u(x) = ae-ikx + beiks, which is also a solution of eq. (1), yields the spectral finite element description of the dynamics of the rod. this results in the following dynamic stiffness matrix of the rod: (3) the corresponding transfer matrix [t] takes the following form: (4) where (5) 2.3 dynamics of the hybrid rod consider now the dynamics of the hybrid periodic rods which consists of a passive sub-cell and an active piezoelectric sub-cell as shown in fig. 3. piezoelectric inserts cell figure 1. typical example of hybrid periodic rod u al x l r figure 2. rod undergoing longitudinal vibrations ( / ) 0xx ttu e uρ− = 2( ) ( ) 0xxu x k u x+ = where k = wave number = / eρ ω . [ ] 4ikl 3ikl 2ikla a a a a 3ikl 2ikl 4ikl )ikla a a a a a ( 1 e )ikl 2e ( 1 e )iklea k l 2e ( 1 e )ikl ( 1 e − − − − ⎡ − − − + ⎢= ⎢ − − + −⎢⎣ 1 1k k kd lld dlr lr [ t ] k 1d k k k krr d d 1ll rld d klr rr dlr − −⎡ ⎤−⎢ ⎥ ⎢ ⎥= − − −⎢ ⎥− ⎢ ⎥⎣ ⎦ 4(1 ) ea iklk e ikld ll l −= − , ( )3 22 ( 1 ) ,ea ikl iklk e e ikld lr l −= − − + ( )3 22 ( 1 )−= − − +ikl ikldrl eak e e ikll , and 4(1 )−= − ikldrr ea k e ikl l . 10 the journal of engineering research vol. 5, no.1 (2008) 7-19 2.3.1 passive sub-cell as shown in eq. (4), the dynamic characteristics of the passive sub-cell (a) can be described as follows: (6) in a more compact form, eq. (6) can be rewritten as: (7) where is the transfer matrix of the passive sub-cell a. one can show that the transfer matrix is a simplectic matrix (asiri, et al. 2002). 2.3.2 active sub-cell the constitutive equations of the active piezoelectric insert are given by baz (2001): (8) where ep, dp, tp and sp are the electrical field intensity, electrical displacement, stress and strain of the piezoinsert. moreover, εs, hp and cd define the electrical permittivity, piezo-coupling constant and elastic modulus. eq. (8) can be rewritten in terms of applied voltage vp, interface piezo-force fi, electrical charge qp and net deflection (ur-ui) as follows: (9) where tp and b are the diameter and the length of the piezo insert. eliminating the charge qp from eq. (9) gives: (10) let the piezo-voltage vp be generated according to the following control law: vp = -kg (ur-ui) (11) where kg is the control gain which can be complex if phase shift is allowed. appropriate phase shift means that the piezo-insert can act as a damper. then, eq (10) reduces to: (12) eq. (12) can be used to generate the force vector {f1 fr}t acting on the piezo-insert rewritten as: (13) with kp = kpc+kps is the total stiffness of the piezo-insert which is complex if the gain is complex by virtue of the phase shift. hence, the dynamic equation of the active sub-cell is given by: (14) sub-cells rod a piezo-insert b rod a piezo-insert b fl fi fi fr la lb ul ui ui ur left (l) interface (i) right (r) (a) complete cell (b) passive sub-cell (c) – active sub-cell figure 3. unit cell of the hybrid periodic rod − − − ⎡ ⎤− ⎢ ⎥⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪⎢ ⎥=⎨ ⎬ ⎨ ⎬ ⎢ ⎥⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭− +⎢ ⎥ ⎣ ⎦ 1 1 dd dilil ilaa al i 1l id d d dd dll ii li llil ila a a aa a k k k u u f fk k k k k k where u and f define the deflection and force vectors with subscripts l and i denoting the left and interface sides of the passive sub -cell and d ija k defines the elements o f dynamic stiffness matrix of the sub -cells a which can be calculated from eq. (5). ⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪=⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ l i a l i u u t f f s p p p dp pp e 1/ h d = t s-h c ε⎡ ⎤⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪⎢ ⎥⎨ ⎬ ⎨ ⎬⎢ ⎥⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭⎣ ⎦ ( )2 / 4 sp p p p b b d r i bi b p v /t 1/ h q /d l = (u u )/lf / d -h c ε π ⎧ ⎫ ⎡ ⎤ ⎧ ⎫⎪ ⎪ ⎪ ⎪⎢ ⎥⎨ ⎬ ⎨ ⎬⎢ ⎥ ⎪ ⎪⎪ ⎪ ⎩ ⎭⎣ ⎦⎩ ⎭ 2[( / 4)( ) ]( )s d 2 si p b p b p b r if = h d v + d c h /l u uε π ε { [( ( ) ]}( )s d 2 si p g b p p r if = h b k + t b c h /l u uε ε ( ) { 1 1} ipc ps r u k k u ⎧ ⎫ = + − ⎨ ⎬ ⎩ ⎭ where 2[( / 4)( ) ]s d 2 sk = h d k and k = d c h /lpc p b g ps p bbε π ε with kpc and kps denoting the active piezo -stiffness due to the control gain kg and the structural piezo -stiffness , respectively. ( ) { } ⎧ ⎫ ⎧ ⎫⎧ ⎫ +⎨ ⎬ ⎨ ⎬ ⎨ ⎬ ⎩ ⎭⎩ ⎭ ⎩ ⎭ i i pc ps r r f u-1 = k k -1 1 f 1 u ⎡ ⎤ ⎧ ⎫ = ⎢ ⎥ ⎨ ⎬ ⎢ ⎥ ⎩ ⎭⎣ ⎦ p p i p p r k k u -k k u ⎡ ⎤ ⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪⎢ ⎥ =⎨ ⎬ ⎨ ⎬⎢ ⎥ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭⎢ ⎥⎣ ⎦ d dii irb b i i d d r rri rrb b k k u f k k u f 11 the journal of engineering research vol. 5, no.1 (2008) 7-19 now, the state vectors at ends r and i of the passive sub-cell b are related through the transfer matrix derived from eq. (14) as follows: (15) in a more compact form, eq. (15) can be rewritten as: 2.3.3 dynamics of entire cell the dynamics of the entire cell can be determined by the assembly of the dynamic equations of the passive and active sub-cells which are given by eq. (3) and eq. (15) respectively. this yields the following dynamic equations: (16) in a more compact form, eq. (16) can be rewritten as: (17) where y and [tk] denote the state vector = {ul fl}t and the transfer matrix of the k cells. for exactly periodic rods, [tk] = [t] and the eigenvalue problem of [t ] can be written as: (18) where λ is the eigenvalue of [t] combining eq. (17) and (18) gives: (19) indicating that λ of the matrix [t] is the ratio between the elements of the state vectors at two consecutive cells. hence, the magnitude of λ determine the nature of wave dynamics in the periodic rod as follows: a further explanation of the physical meaning of the eigenvalue λ can be extracted by rewriting it as: (20) where µ is defined as the "propagation factor" which is a complex number whose real part (α) represents the logarithmic decay of the state vector and its imaginary part (β) defines the phase difference between the adjacent cells. one can rewrite eq. (19) as: (21) (22) from eq. (21) and eq.(22), we get: (23) eq. (19) indicates that: therefore, the equivalent conditions for the pass and the stop bands can be written in terms of the propagation constant parameters (α and β) as follows: the propagation factor, µ, can be calculated directly from the diagonal elements of overall transfer matrix as following: (24) in case of a rod element, the transfer function has two eigenvalues µ and 1/µ and according to the properties of where d ijb k defines the elements of dynamic stiffness matrix of the sub -cell b which can be calculated from eq. (13). − − − − ⎡ ⎤− ⎧ ⎫⎧ ⎫ ⎢ ⎥⎪ ⎪ ⎪ ⎪= ⎢ ⎥⎨ ⎬ ⎨ ⎬ ⎢ ⎥⎪ ⎪ ⎪ ⎪⎩ ⎭ − + ⎩ ⎭ ⎢ ⎥⎣ ⎦ 1 1 dd diiir irbb b ir 1 1 ir d d dd drr ii rrir irb b bb b k k k uu ff k k k k k ⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪=⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ r i b r i u u t f f where bt is the transfer matrix of the active sub-cell b. + ⎧ ⎫ ⎧ ⎫⎪ ⎪ ⎪ ⎪= ⎡ ⎤⎨ ⎬ ⎨ ⎬⎣ ⎦ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ l l k l lk 1 k u u t f f where ⎡ ⎤ ⎢ ⎥= = ⎢ ⎥⎣ ⎦ 11 12 k b a 21 22 t t t t . t t t + = ⎡ ⎤⎣ ⎦k 1 k ky t y [ ] k kt y yλ= k 1 ky yλ+ = if 1λ = , the wave propagates along the rod without any change of amplitude, indicating a pass band and if not then the wave will be attenuated, indicating a stop band . ie eµ α βλ += = 1 l li l lk k u u e f f α β+ + ⎧ ⎫ ⎧ ⎫ =⎨ ⎬ ⎨ ⎬ ⎩ ⎭ ⎩ ⎭ and considering only the jth components l ju of the deflection vector ul, at cells k and k+1, which can be written as : i jk l l ljk 1 jk 1 jk u u and u e φ + + = = where l jnu and jnφ denote the amplitude and phase shift of the jth component l ju at the n th cell. ( ) ( ) ( ) 1 1 1 l l l ljk jk jk jk jk jk ln u / u ln u / u i iφ φ α β + + + = + = = + α = 1l lj jk k ln(u / u ) + = logarithmic decay of amplitude, and ( )1jk jkβ φ φ+= − = phase difference between the adjacent cells a. if α =0 (ie. µ is imaginary), then we have “pass band ” as there is no amplitude attenuation. b. if 0α ≠ (ie. µ is real or complex), then we have “stop band ” ” as there is amplitude attenuation defined by the value of α . 1 1 2 2aco sh 2 +⎛ ⎞= ⎜ ⎟ ⎝ ⎠ t t µ 12 the journal of engineering research vol. 5, no.1 (2008) 7-19 the simplectic matrix (or transfer matrix), the trace is equal to the summation of the eigenvalues. so one can show that: (25) a better insight into the physical meaning of the eigenvalues µ and µ-1 can be gained by considering the following transformation of the cell dynamics into the wave mode component domain: (26) substituting eq. (26) into eq. (25), it reduces to: (27) (28) but, because of the particular nature of the eigenvalues of [t] as they appear in pairs (µ, µ-1), then the above equation reduces to: (29) eq. (29) can be expanded to give: (30) for the jth component of w, we have: (31) 3. system performance 3.1 overview in order to demonstrate the feasibility of the theoretical concepts presented, experimental investigations are conducted. these investigations were carried out on two main steps. in the first step, the vibration attenuation characteristics of the hybrid periodic rod with two actuators on the shaker was studied and evaluated. in the second step, the tunable mechanical filter is used to evaluate its performance for attenuating the vibration transmission from the vibration source, the motor, to the lower plate. in the present study, the piezo actuators model 712a02 (see the index table 1) from pcb piezotronics, inc. (depew, ny) are used. the actuator shown in fig. 5 has 1 ac o s h 2 ⎛ ⎞+ ⎜ ⎟= ⎜ ⎟ ⎝ ⎠ λ λµ { } 1 1 1 1 l r k k l k l k r l l k k ll k l k u y w w f u w and y w f w φ φ φ φ+ + + + ⎧ ⎫⎪ ⎪= = =⎨ ⎬ ⎪ ⎪⎩ ⎭ ⎧ ⎫⎧ ⎫⎪ ⎪ ⎪ ⎪= = =⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ where φ is the eigenvector matrix of the transfer matrix [ t]. also, wk is the wave mode component vector which has the right -going wave component wr and left -going wave component wl. [ ] ( ) [ ] 1 1 1 1 r r l l k k l l l lk k r r l l l l l lk k w w y t y t , w w w w or t w w φ φ φ φ + + − + ⎧ ⎫ ⎧ ⎫ ⎪ ⎪ ⎪ ⎪ = = =⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ ⎧ ⎫ ⎧ ⎫ ⎪ ⎪ ⎪ ⎪+⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ note that the matrix ( )1 t−φ φ reduces to the matrix of eigenvalues of the matrix [ t], ie.: ( )1 2 2 1 2 1 r r l l n n l l l lk k w w diag , ,..., , w w λ λ λ λ− + ⎧ ⎫ ⎧ ⎫ ⎪ ⎪ ⎪ ⎪=⎨ ⎬ ⎨ ⎬ ⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ 1 1 0 0 r r l l l l l lk k w w w w λ λ− + ⎧ ⎫ ⎧ ⎫ ⎪ ⎪ ⎡ ⎤ ⎪ ⎪=⎨ ⎬ ⎨ ⎬⎢ ⎥⎣ ⎦⎪ ⎪ ⎪ ⎪⎩ ⎭ ⎩ ⎭ where [ ] 1 1 11 2 1 2diag , ,.. , and diag , ,.. .− − −⎡ ⎤λ = λ λ λ = λ λ⎣ ⎦ 1 1 1, and k k k k r r l l l l l lw w w w+ + −= λ = λ it is clear from eq. (31) that the eigenv alue jλ is the ratio between the amplitude of the right -going waves whereas 1jλ − defines the ratio between the amplitude of the left -going waves. hence, if ( ) ( ) ( )1 11 1j j j j,then and the pair of ,λ λ λ λ− −< > and the pair of ( 1j j,λ λ − ) denote attenuation of the amplitude of wave propagation from cell k to cell k+1. this can be clearly understood by considering fig. 5. note that the right -going wave ( rl jk w ) is attenuated by jλ as it prop agates from cell k to k+1 and the resulting wave ( 1+ r l jk w ) has a lower amplitude as 1 r r jl ljk jk w wλ + = from the first part of eq. (31). the left-going wave ( 1+ l l jk w ) is amplified by 1jλ − as it propagate s from cell k+1 to cell k. hence, the left going wave ( 1+ l l jk w ), at cell k+1, has a lower amplitude than that at k. accordingly, the amplitudes of both the right and left -going waves ( rl jk w and l l jk w ) at the left end of cell k are higher than the right and left -going waves ( 1+ r l jk w and 1j k l lw + ) at left end of cell k+1. therefore, if ( )j 1,λ < then the pair ( jλ , 1−jλ ) define attenuation (or stop bands) along the wave propagation direction from cell k to cell k+1. also, if ( ) 1j ,λ = then the pair ( jλ , 1− jλ ) define pro pagation without attenuation ( ie. pass bands). 1 1 1, and j j j jk k k k r r l l l j l l j lw w w wλ λ + + −= = and or drag 13 the journal of engineering research vol. 5, no.1 (2008) 7-19 two built-in side bolts so that the short aluminum rods can be bolted from both sides. figure 6 shows the frequency response characteristics of the actuator authority between 150-5000 hz with a sensitivity of 0.015 pound/volt and a peak excitation voltage of 100 volts. 3.2 experimental facilities figure (7a) shows the dimensions of the hybrid periodic rod used to build the tunable mechanical filter. it consists of three short aluminum rods shown (see for properties table 2 in the appendix) bolted together by the actuator as shown in fig. (7b). the fundamental longitudinal natural frequency of the periodic rod is 640 hz. the frequency response of such a rod is obtained both in its passive (open-loop) and active (closed-loop) modes of operation. two experimental test rigs have been employed in this study to evaluate the performance of the tunable mechanical filter. the first test rig aims at monitoring the vibration transmission characteristics of the hybrid rod alone as influenced by geometrical and material discontinuities. figure 8 shows the details of the employed test facility. the objective is to measure the transfer function between the input excitation, the shaker, and the output response of the tip of the rod. first, for the passive plain rod (ie. no periodicity and open circuit), second, for the passive periodic rod (ie. with periodicity and open circuit), and third, for the hybrid periodic rod (ie. with periodicity and short circuit). the phase shifter has been used to introduce a feedback signal to the actuator to produce the destructive interference of waves. the experimental setup shown in fig. 9 was used to demonstrate the feasibility of using the tunable mechanical filter to isolate the vibrations induced by a motor. the experiment has been done for the three cases described in first test rig. two phase shifters have been used to produce the negative feedback signals to the two actuators. attenuation rl jk w jλ 1 r l jk w + propagation l cell k r l cell k+1 r ll jk w 1jλ − 1 l l jk w + direction amplification high amplitude low amplitude figure 4. attenuation of right and left-going waves in the propagation direction figure 5. piezoelectric actuator pcb piezotronic model 712a02 frequency range, hz 150 to 5000 dynamic performance broadband force (min), ib/volt 0.004 clamped force, ib/volt 0.015 static performance free displacement, ì in/volt 2.56 environmental temperature, of -10 to +150 capacitance nf 65 resistance ohm (min) 1 x 107 dc excitation -125 to 500 ac excitation (off resonance) ± 100 electrical input voltage (max), vpk ac excitation (on resonance) ± 80 weight, oz 1.27 height 0.4 size, in diameter 2.0 electrical connector bnc plug mounting thread 10-32 male housing material titanium mechanical sealing type welded hermetic table 1. the main geometrical and performance parameters of the actuators material aluminum density 2700 kg/m 2 modulus of elasticity 71 gpa wave speed 5128 m/s table 2. material properties of the aluminum rods 14 the journal of engineering research vol. 5, no.1 (2008) 7-19 figure 6. the frequency response characteristics of the actuator authority 3/4 in p iezo electric a ctuator s 1/2 in 1.6 in 1 9 in 6 in (a) (b) figure 7. geometrical parameters of the hybrid rod signal conditioner phase shifter power amplifier fft analyzer hybrid rod actuator accelerometers actuator shaker figure 8. test setup for evaluating the vibration transmission characteristics of the rod piezoelectric actuators m ag ni tu de ( lb s/ v ) ph as e (d eg .) frequency (hz.) frequency (hz.) 15 the journal of engineering research vol. 5, no.1 (2008) 7-19 3.3 experimental results 3.3.1 single hybrid rod in all the reported results, the hybrid periodic rod has provided a viable means for extending the width of the stop band between 0-4000 hz. also, attenuations of more than 20 db are obtained with control voltages less than 5 volts as shown in fig. 11 for the two control actuators. note that frequency range higher than 4000 hz was not considered because of the limitation imposed by the actuator bandwidth. figure 12 displays the lateral vibration distribution over plain, passive periodic, and hybrid periodic rods at four different frequencies using the psv-200 scanning laser vibrometer. figure 12 shows that the lateral vibration transmission is attenuated also in addition to the longitudinal waves whenever the frequencies lie inside the stop band with the hybrid periodic rod out-performing the plain and the passive periodic rods. 3.3.2 the table-like system figure 13 shows the transfer function of the tunable mechanical filter with plain, passive periodic, and hybrid periodic rods when it is subjected to broad band excitations from vibration source, the motor. the excitation from the motor can be considered as a random excitation and the geometrical information of the configuration of rods is the same as shown in fig. 7. the periodic rods in active mode are very efficient in the low frequency range while in the high frequency range they have the same effifigure 9. experimental setup used to demonstrate the feasibility of using the table-like to isolate the vibrations induced by the motor figure 10a shows the magnitude of the transfer functions of plain, passive periodic, and hybrid periodic rods when they are subjected to random excitation in the axial direction. these transfer functions quantify the response of the free end of the ro ds to input excitations at the other end of the rods. figure 10a indicates clearly that the passive periodic rod exhibits experimentally a broad stop band between 650 -4000 hz whereby the vibration transmission through the rod is almo st eliminated. further more, fig. 10a indicates that the hybrid rod provides an effective means for attenuating the vibration transmission over a very broad frequency range. it is particularly effective for stopping the low frequency vibration in the range below 600 hz where the passive periodic rod has been ineffective. this result is particularly important in using a hybrid rod that combines both the passive and active strategies to stop high as well as low frequency wave propagation. figure 10b shows that the widths of the st op bands can be clearly predicted theoretically by plotting the real part α of the propagation parameter µ . for values of α ≠ 0, the stop bands can be cle arly identified and match closely the experimental results for both the passive and hybrid rods. 16 the journal of engineering research vol. 5, no.1 (2008) 7-19 ciency of the passive mode. that is why the blue and green lines in fig. 13 have the same trend above 1500 hz. the corresponding control voltage of the upper control actuators is shown in fig. 14. 4. conclusions this paper has presented the tunable mechanical filter as a new tool for isolating vibrations. it consists of four figure 10. the transfer functions of the plain, passive periodic, and hybrid periodic rods with the propagation factor in both passive and active modes 1 1 10-2 10-4 10-1 10-2 10-3 10-4 figure 11. the control voltage of (a) the lower actuator, and (b) the upper actuator (a) (b) 17 the journal of engineering research vol. 5, no.1 (2008) 7-19 (a) (b) (c) figure 12. the vibration of plain, passive periodic and hybrid periodic rods at: (a) 160 hz (in pass band), (b) 800 hz (in stop band), and (c) 1520 hz (in stop band) at 160 hz at 800 hz at 1520 hz 18 the journal of engineering research vol. 5, no.1 (2008) 7-19 hybrid periodic rods installed between two plates. the theory governing the operation of this class of rods has been presented. the factors governing the design of effective periodic rods have been identified. the performance characteristics of passive and hybrid periodic rods alone have been measured experimentally and compared with the theoretical predictions of the real part of the propagation factor. close agreement between theoretical predictions and experimental results has been achieved. the performance of the tunable mechanical filter with a motor assembly has also been monitored experimentally. the predictions of the stop bands have also been found to be in close agreement with the experimental results. the periodic rods in active mode are very efficient in the low frequency range while in the high frequency range they have the same efficiency of the passive mode. as a result, the mechanical filter with hybrid periodic rods is found to be more effective in attenuating the transmission of vibrations from the source of vibrations to the lower plate over a broader frequency range extending from 150-4000 hz with control voltages not exceeding 5 volts. this concept of the tunable mechanical filter can be implemented efficiently in many applications. for instant, the gearbox support system of helicopter and automotive vehicle engine mounting systems. with such unique filtering characteristics, it would be possible to control the wave propagation both in the spectral/spatial domains in an attempt to stop/confine the propagation of undesirable disturbances. figure 13. the transfer function in case of collocated sensor/actuator arrangement with broadband excitation from the motor excitation frequency, hz tr an sf er f un ct io n, d b 0 500 1000 1500 2000 2500 3000 excitation frequency, hz figure 14. the control voltage applied to the upper control actuator 101 10o 10-1 10-2 10-3 19 the journal of engineering research vol. 5, no.1 (2008) 7-19 references asiri, s., 2005, "vibration isolation of automotive vehicle engine using periodic mounting systems," spie 2005, san diego, california, usa. asiri, s., baz, a. and pines, d., 2004, "active periodic struts for gearbox support system," proc. of spie, smart structures and materials 2004: damping and isolation, vol. 5386, pp. 347-358. asiri, s., baz, a. and pines, d., 2002, "periodic struts for gearbox support system," proc. of the 2002 international congress and exposition on noise control engineering, inter noise 2002, paper number in02-644, dearborn, mi, usa. baz, a., 2001, "active control of periodic structures," asme journal of vibration and acoustics, vol. 123, pp. 472-479. baz, a., 2002, "vibration damping," class notes, university of maryland at college park. baz, a., 2000, "spectral finite element modeling of longitudinal wave propagation in rods with active constrained layer damping," smart materials and structures, vol. 9(3), pp. 372-377. brillouin, l., 1953, "wave propagation in periodic structures," 2nd ed. dover. cremer, l., heckel, m. and ungar, e., 1973, "structureborne sound," springer-verlag, new york. doyle, j., 1997, "wave propagation in structures," 2nd ed., springer-verlag, new york. faulkner, m. and hong, d., 1985, "free vibration of a mono-coupled periodic system," j. of sound and vibrations, vol. 99, pp. 29-42. gupta, s., 1970, "natural flexural waves and the normal modes of periodically-supported beams and plates," j. of sound and vibration, vol. 13, pp. 89-111. mead, d. j. , 1986, "a new method of analyzing wave propagation in periodic structures; applications to periodic timoshenko beams and stiffened plates," j. of sound and vibration, vol. 114, pp. 9-27. mead, d.j., 1970, "free wave propagation in periodically supported, infinite beams," j. of sound and vibration, vol. 11, pp. 181-197. mead, d. j., 1971, "vibration response and wave propagation in periodic structures," j. of engineering for industry, vol. 21, pp. 783-792. mead, d. j., 1975, "wave propagation and natural modes in periodic systems: i. mono-coupled systems," j. of sound and vibration, vol. 40, pp. 1-18. mead, d. j., 1996, "wave propagation in continuous periodic structures: research contributions from southampton," j. of sound and vibration, vol. 190, pp. 495-524. mead, d. j. and bardell, n. s., 1987, "free vibration of a thin cylindrical shell with periodic circumferential stiffeners," j. of sound and vibration, vol. 115, pp. 499-521. mead, d. j. and markus, s., 1983, "coupled flexurallongitudinal wave motion in a periodic beam," j. of sound and vibration, vol. 90, pp. 1-24. mead, d. j. and yaman, y., 1991, "the harmonic response of rectangular sandwich plates with multiple stiffening: a flexural wave analysis," j. of sound and vibration, vol. 145, pp. 409-428. orris, r. and petyt, m., 1974, "a finite element study of harmonic wave propagation in periodic structures," j. of sound and vibration, vol. 33(2), pp. 223-237. patera, a., 1984, "a spectral element method for fluid dynamics: laminar flow in channel expansion," j. of comput. physics, vol. 54, pp. 468-488. pierre, c., 1988, "mode localization and eigenvalue loci veering phenomena in disordered structures," j. of sound and vibration, vol. 126, pp. 485-502. ravindra, b. and mallik, k., 1992, "harmonic vibration isolation characteristics of periodic systems," j. of sound and vibration, vol. 154(2), pp. 249-259. richards, d. and pines, d.j., 2003, "passive reduction of gear mesh vibration using a periodic drive shaft," j. of sound and vibration, vol. 264(2), pp. 317-342. roy, a. and plunkett, r., 1986, "wave attenuation in periodic structures," j. of sound and vibration, vol. 114, pp. 395-411. ruzzene, g. and baz, a., 2000, "control of wave propagation in periodic composite rods using shape memory inserts," j. of vibration and acoustics, vol. 122, pp. 151-159. sackman, s., kelly, j. and javid, a. a., 1999, "layered notch filter for high-frequency dynamic isolation," j. of pressure vessel technology, vol. 111, pp. 17-24. singh, a., pines, d. and baz, a., 2004, "active/passive reduction of vibration of periodic one-dimensional structures using piezoelectric actuators," smart materials and structures, vol. 13, pp. 698-711. snowden, j.c., 1979, "vibration isolation: use and characterization," nbs handbook, vol. 128, us national bureau of standards. szefi, j., smith, e. and lesieutre, g., 2004, "design and testing of a compact layered isolator for highfrequency helicopter gearbox isolation," prec. of 45th aiaa structures, structural dynamics and materials conference, palm spring, ca, usa. c:\users\raoh\desktop\paper_1.xps tjer 2012, vol. 9, no. 1, 1-10 ________________________________________ *corresponding author’s e-mail:drkarimaa@yahoo.com thermal performance of solar hot water systems using a flat plate collector of accelerated risers ke amori* and ns jabouri department of mechanical engineering, nahrain university, university of baghdad, baghdad, iraq received 3 may 2010; accepted 27 september 2010 abstract: this study focuses on a comparison of the performance of two similar locally-fabricated solar water heaters. one of the collectors features a new design for accelerated absorber; its risers are made of converging ducts whose exit area is half that of the entrance. the other collector is a conventional absorber, with risers of the same cross sectional area along its length. each collector is the primary part of an indirect thermosyphon circulation solar hot water system. both collectors face south with a fixed tilt angle of 33.3 from the horizontal. a side-by-side experiment was conducted on the two solar water heaters from january to april of 2009 for different water withdrawal profiles, continuous, interrupted and no load, as well as for horizontal and vertical storage tank orientations. two types of storage tanks were investigated, those with two concentric cylinders, and those with helically-coiled tubes in the cylinder. results show that a considerable enhancement of thermal performance (approximately 60%) of absorbed heat (useful gain) at solar noon was obtained for the new design, in comparison with the conventional type. the instantaneous efficiency was 31.5% for the accelerated absorbed flat plate at solar noon, while that of conventional absorber was (16.5%). the longitudinal water temperature variations in the risers of accelerated absorber were larger than that of the conventional absorber. the stratification in the storage tank was significantly improved for the solar hot water system, with the new absorber design in which the maximum temperature measured was 50 c (vertical storage tank), while that for conventional absorber was 37 c. the stratification obtained for a coiled tube in a cylinder storage tank indicated a good thermal performance with less space requirements. also, the circulation rate in the accelerated absorber was higher when it was connected to a coiled tube than when it was connected to the concentric cylinders. keywords: solar, accelerated flat plate collector, storage tank, coiled tube, stratification 2 ke amori and ns jabouri 1. introduction solar hot water systems function as heat exchangers. they receive solar radiant energy and transfer it to the flowing fluid. the performance of solar systems varies as the design variables change, so it is therefore necessary to identify the parameters affecting this design and the operational variables. keltt et al. (1984) studied through experimentation the thermal performance of submerged coil heat exchangers for single wall coil and double wall coil for different tank sizes namely (300 l and 450 l) for different load flow rates. khalifa (1999) investigated a thermosyphon domestic hot water system to show the important variables that affect the performance of the solar system such as the temperature variation along the absorber fins, tubes and in the flow direction as well as the thermosyphonic mass flow rate. the design of an efficient storage tank heat exchanger has been investigated by (shokouhmand et al. 2008) with different coil pitches and curvature ratios. an enhancement in the heat transfer rate is obtained due to the centrifugal force due to the curvature of the tube, resulting in the secondary flow development. the objective of the present work is to compare experimentally thermal performance of a locally-made new solar flat plate collector (named here as accelerated solar collector) to a flat plate collector of straight risers the conventional type hot water system for a different water withdrawal pattern. additionally, the study will show the stratification in the storage tank. in order to carry out a fair comparison, the collectors are made as identically possible. 2. theory thermal analysis has covered in many solar thermal engineering texts (duffie, beckman 1974; lunde 1980). therefore, only equations which describe the thermal performance of the system will be described in this paper. actual collector and supply useful energy gain for the collector closedloop cycle, the hourly useful energy gain can be calculated by: (1) the heat obtained by the water withdrawal can be calculated as: (2) where qcoll heat is transferred in w, m mass flow rate in kg/s, cp water specific heat in j/kg.k, t is the measured temperature at the collector inlet or outlet (oc), and t20 and t18 are supply water temperatures at inlet and outlet from the storage tank, as shown in fig. (1b) and fig. (3). the useful energy enhancement is calculated as: (3) the collector instantaneous efficiency can be determined according to: (4) ig is the global solar radiation on tilted collector at 33.3 from horizontal calculated from that measured on a horizontal surface. 3. experimental setup two identical solar water heaters are manufactured in this experiment; one is supplied with flat plate collectors of accelerated absorber, the other is supplied with flat plate collectors of conventional absorber. figure 1 shows the configuration of solar water heater tested when its storage tank is composed of two concentric cylinders, and when it is a helical coiled tube in cylindrical shell. 3.1 accelerated flat plate solar collector a flat plate collector with an accelerator plate consists of nine equally-spaced parallel converging rectangular cross-sectional area copper risers (fig. 2) with an inlet hydraulic diameter of 2 cm. the inlet dimensions are 20x20 mm, and the outlet hydraulic diameter 13.3 mm, while the outlet dimensions are 10x20 mm and tube length of 1.2 m. the riser to riser distance is 0.1 m, and each tube has an equal right and left fin length of 40 mm. these tubes are connected with two headers of rectangular cross sectional area one at each end to collect or distribute water from and into the risers, and from collector's outlet and inlet. the joints between the headers and risers ends are welded by using brass alloy. a mat painted copper sheet is used as an absorbing plate of 1.20 x 0.9 m and of 1.9 mm thickness. this sheet is welded to the risers by brass alloy, the structure is good insulated with a fire wool coating of 70 mm thickness. 3 thermal performance of solar hot water system using a flat plate collector of accelerated risers 3.2 conventional flat plate solar collector a conventional flat plate collector without accelerator plate is designed identical to an accelerated flat plate collector, and it consists of nine equal spaced parallel copper rectangular cross sectional area pipes of inlet and outlet with equal hydraulic diameters of 20 mm. in addition, the inlet and outlet cross sections dimensions are 20x20 mm and 1.20 m length, and the riser to riser distance is 0.1 m. each tube has an equal fin length to right and left of 40 mm, and these tubes are connected with two headers of rectangular cross sectional area. the risers are welded to the header using brass alloy. a copper sheet is used as an absorbing plate of 1.20 x 0.9 m and of 1.9 mm thickness. this sheet is welded to the risers using a brass alloy, the pipes are supported from the bottom by a galvanized plate of 2 mm thickness. the structure is wellinsulated with a glass wall coating of 70 mm thickness. the collector box frame is manufactured from a 2 mm thickness galvanized sheet formed as a box, and the collectors are connected to their containers by screws. the collectors are well-insulated by using glass wool insulation of 70 mm thickness from the back of the collectors and 50 mm thickness from the perimeter of the collectors. a glass panel of 4 mm thickness is used as a transparent cover for the collector with area of 1.25x0.95 m². the glass cover lies on the collector frame in a small channel 25 mm wide, and is fixed by using a black silicon. a piece of rubber tape is fixed between the glass and channel to prevent air leakages. the cover-absorbing plate spacing is 70 mm. 3.3 storage tanks 3.3.1 shell and tube storage tank two concentric horizontal cylinders form the storfigure 1. close loop themosyphon solar water heater, the storage tank is a) two concentric cylinders b) helical coiled tube in cylindrical tank (b) (a) storage tank 4 ke amori and ns jabouri age tank made from a galvanized plate of 2 mm thickness. the internal cylinder has an inner diameter of 375 mm, 1 m length and capacity of 110 l. the outer cylinder has an internal diameter of 470 mm and a length of 1.2 m, forming an annular space of 90 l. the internal cylinder is used for the closed-loop water (energy carrier) while the annular space between internal and external cylinders is used for loading water (domestic use water). each cylinder has two holes to form inlet and outlet ports. the tank is well insulated by a glass wool insulator of 70 mm thickness, and the storage tank can be oriented horizontally or vertically as required. 3.3.2 coiled tube in shell another storage tank manufactured by (karima et al. 2011) is used in the present work. this tank (made from a galvanized sheet) is of 125 l volume, 1m height and 0.4 m diameter. a coiled tube is inserted inside in the tank which in turn is well insulated with glass wool of 10mm thickness. a 0.6 m high coiled tube is made by winding a copper tube of 9.5 mm diameter around a cylinder to form twelve turns of 0.25 m outside coil diameter. a spacer was placed between each two consecutive coil turns to ensure a uniform pitch along the coil, which measured 0.05 m. 3.4 temperature measurement one hundred and five (105) calibrated thermocouples (type t copper and constantan) are used to measure the temperature at various points of water tubes, water storage tanks, inlet and outlet of collectors, inlet and outlet of the withdrawal water, ambient and glass cover, as shown in fig. 3. the entire number of thermocouples are joined to digital thermometer reader (autonics-t4wm/ k(ca) 0-1200). the ambient temperature is measured by a using a mercury in glassthermometer. 3.5 flow rate measurement the mass flow rate of the thermosyphon circulation flat plate collectors is so small tht only the laser flow meter can detect the flow which in not available, so a transparent tube of 250 mm length and 10mm diameter is connected between collector outlet and storage tank inlet. an ink injection is used to calculate the circulation rate by injecting ink and measuring the time required to pass the 250 mm distance using stopwatch to compute fluid velocity. the water load flow rate is measured by using a flow meter of 2-22 l/min. range. 3.6 test procedure the thermosyphon circulation solar water flat plate collectors were connected as a closed loop (indirect) system. the experiments were carried out from january to march 2009. before each test, the following preparations were made, the closed collector loop figure 2. configuration of manufactured flat plate collector collector risers container frame insulation a) collector cross sectional view b) accelerated riser (front view) c) conventional riser (front view) 5 thermal performance of solar hot water system using a flat plate collector of accelerated risers was filled with water, the glass cover is cleaned thoroughly and the measurements and apparatus were checked. then the storage tanks were filled with water and readings were taken each half hour from sunrise to sunset. the newly-designed system was tested side by side with the conventional type. the test was conducted from sunrise to sunset for five different mass flow rates of load water withdrawal profiles. these profiles were: continuous load of 2.6 l/min, the profile (ii) shown in fig. 7 which is equivalent to daily consumption of single storage tank volfigure 3. locations of temperature measurements figure 4. solar radiation on horizontal surface and ambient temperature (5/3/2009 at 33.3 n. baghdad) 6 ke amori and ns jabouri ume, emptying the storage tank twice a day and finally emptying it once a day. this propose is to test the effect of storage tank orientation (horizontally and vertically) on the thermal performance and study the effect of using a coiled tube for collector closed-loop instead of cylindrical tank. in all the above cases all temperature measurements, flow rate (circulation rate and loop flow rate), windspeed are recorded. 4. results and discussions figure 4 shows the ambient temperature (measured (every 15 minutes) and solar radiation for the 24 hours of 27-1-2009 on the solar collector, which is obtained from the ministry of science and technology solar center in baghdad. it is clear that the peak solar radiation was between hours of 11 to 13. the peak ambient temperature was 19 c. figure 5 shows the hourly water temperature rise along a selected riser c-c for the accelerated conventional type. it is clear that the temperature rises with time as water proceed along the risers length, but lesser temperature values are indicated for the conventional type since the accelerated riser occupies less water volume at the cross section especially at the riser on the upper part. figure 6 (a & b) shows a comparison figure 5. water temperature distribution along risers local time (hr) acclerated riser c-c local time (hr) acclerated riser c-c 7 thermal performance of solar hot water system using a flat plate collector of accelerated risers figure 6. the thermosyphon circulate rate in the collector (33.3 n. baghdad) a) profile i, b) profile ii figure 7. water withdrawal profile (case ii) load case ii local time (a) (b) 8 ke amori and ns jabouri of the circulation rate between the accelerated and conventional flat plate for load profile i and load profile ii shown in fig. 7 respectively. a significant increase in the circulation rate is recorded from the first hour of day to the last hour, due to the new design criteria of the risers (converging cross sectional area). the maximum increase was at solar noon, and was due to exposing the same area to the same incident solar radiation. however, the water content in the converging risers was less than that in the straight riser so the temperatures increased and its density was lighter, resulting in it rising faster. figure 8 shows the percentage enhancement in absorbed heat when using an accelerated collector compared to a conventional collector (horizontal tank case ii). the useful energy enhancement was calculated using eq. (3). the maxifigure 8. percentage enhancement in absorbed heat when using accelerated collector compared with conventional collector (horizontal tank case ii) figure 9. the instantaneous efficiency of the conventional and (accelerated collector for load profile ii) figure 10. effect of tank orientation and type on thermosyphon circulation rate (profile ii) figure 11. stratification in shell and tube horizontal storage tank (accelerated collector load profile i) 9 thermal performance of solar hot water system using a flat plate collector of accelerated risers mum enhancement was 60% at solar noon and the minimum was measured at the day’s run hour. thus the new design recorded an instantaneous efficiency of 32.8% , while the conventional type indicated 17% as a maximum value, as shown in fig. 9. figure 10 shows that the vertical orientation of the storage tank enhances the circulation rate, and the coiled tube in cylinder has a much more effective design than the two concentric cylinders. figure 11 shows the horizontal stratification in the concentric cylinders storage tank when connected to the accelerated collector. the upper right corner shows the hottest region (inlet of hot water from collector and exit of hot load water t=24.5 c), while the lower left corner is the coldest region exit of cold water to the collector and the inlet of cold load water t=16.5 c). figure 12 shows the vertical stratification in the two concentric cylinders for no load where the maximum temperature indicated at the upper surface (inlet of collector hot water and exit of load hot water tmax = 51 c). figure 13 shows the vertical stratification in the coiled tube in the cylindrical storage tank for load profile ii , where the maximum temperature indicated at the upper surface (inlet of collector hot water) and exit of load hot water tmax = 37.1 c. 5. conclusions this research shows that the system thermal behavior is sensitive to any component and can be enhanced further by modifying the design parameters and material selections. the following concluding remarks can be made, based upon this work. * the stratification obtained in the storage tank is significantly affected by the collector inlet temperature. * the stratification obtained in the storage tank relates proportionately to the type of load rate (or load withdrawal pattern). * increasing the flow rate of the load decreases the stratification in the storage tank * the effectiveness of the storage tank (shell and coiled tube) reaches approximately 78% for (450 ml/min) load flow rate. * using a car radiator instead of the coiled tube enhances the thermal performance of the storage tank. * using an accelerated (converging) riser will enhance the thermal performance of the thermosyphonic solar collector because it will increase the circulation rate. references duffie ja, beckman wa (1974), solar energy thermal process. john wiley & sons inc. new york. karima ae, mustafa ff, sahar m (2011), solar 0.1 0.2 0.3 0.4 x 0.8 0.7 0.6 0.5 0.4 figure 12. stratification in two concentric cylinders vertical orientation (accelerated collector for no load) -0.2 0 0.2 radius (m) figure 13. stratification in coiled tube in cylinder vertical storage tank (accelerated plate load ii) 0.8 0.6 0.4 0.2 10 ke amori and ns jabouri water heater with shell and helical coiled tube heat exchanger as a storage tank. j. of engineering 17(5):1213-1222. khalifa an (1999), thermal performance of locally made flat plate solar collectors used as part of a domestic hot water system. energy conversion & management 40:1825-1833. klett de, goswami dy, saad mt (1984), thermal performance of submerged coil heat exchangers used in solar energy storage tanks. j. of solar energy engineering 106:373-375. lunde p (1980), solar thermal engineering, space heating and hot water systems. john wiley & sons inc. new york. myrna d (1997), high performance in low flow solar domestic hot water systems. msc. thesis mechanical engineering university of wisconsinmadison. shokouhmand h, salimpour mr, akhavanbehabadi ma (2008), experimental investigation of shell and coiled tube heat exchangers using wilson plots. int. communications in heat and mass transfer 35:84-92. soteris k (2009), thermal performance, economic and environmental life cycle analysis of thermosyphon solar water heaters. solar energy 83:39-48. c:\users\raoh\desktop\paper_1.xps the journal of engineering research vol. 8 no. 2 (2011) 1-9 1. introduction nowadays, industrial organizations complexity exists as an immense international interest and knowledge for the scientific basis. studying the importance of reducing complexity into industrial organizations was recommended as one of several solutions to __________________________________________ *corresponding author’s e-mail:garbie@squ.edu.om recovery the existing financial recession (garbie 2009 and 2010; garbie and shikdar 2009, 2010 and 2011). it was defined as systemic characteristics which integrate several key dimensions of the industrial environment including size, variety, information, uncertainty, control, cost, and value (kamrani and adat 2008). flexibility and agility are considered as the most desircomplexity analysis of industrial organizations based on a perspective of systems engineering analysts i. h. garbie* and a. a. shikdar department of mechanical and industrial engineering, college of engineering, sultan qaboos university, p.o. box 33, postal code 123, al khoud, muscat, oman received 4 april 2011; accepted 18 september 2011 abstract: complexity in industrial organizations became more difficult and complex to be solved and it needs more attention from academicians and technicians. for these reasons, complexity in industrial organizations represents a new challenge in the next decades. until now, analysis of industrial organizations complexity is still remaining a research topic of immense international interest and they require reduction in their complexity. in this paper, analysis of complexity in industrial organizations is shown based on the perspective of systems engineering analyst. in this perspective, analysis of complexity was divided into different levels and these levels were defined as complexity levels. a framework of analyzing these levels was proposed and suggested based on the complexity in industrial organizations. this analysis was divided into four main issues: industrial system vision, industrial system structure, industrial system operating, and industrial system evaluating. this analysis shows that the complexity of industrial organizations is still an ill-structured and a multi-dimensional problem. keywords: complexity, industrial organization, systems engineering 2 i.h. garbie and a.a. shikdar ability of certain system properties for the industrial organizations. these properties will give industrial enterprises more ability to cope with increased environmental uncertainty and adapting to the faster pace of change of today's markets (giabchetti et al. 2003). the industrial organizations are sometimes viewed as an intrinsic structural property of the system. the structural property is defined as how individual system components relate to each other and how the relationship determines overall system behavior (arteta and giabchetti, 2004). sometimes, manufacturing systems complexity considers the product variety. this means increasing in the product variety increases in the complexity in the manufacturing systems (kuzgunkaya and elmaraghy 2006). also there is another classification of industrial organizations complexity: time-independent complexity and time-dependent complexity. time -independent complexity is used to add the complexity arising from the designer's perception while time-dependent complexity is either combinational or periodic (kuzgunkaya and elmaraghy 2006). there are several concepts such as: product, process, and operation complexity can be used as a component in industrial enterprises. the product complexity focuses on product features and specifications while the process complexity analysis focuses on the tools, equipment and operations used to manufacture it (hu et al. 2008). supply chain management complexity such as: upstream complexity, internal manufacturing complexity, and downstream complexity are considered as complexity issues regarding manufacturing systems (bazarth et al. 2009). manufacturing strategy also plays an important role in complexity in industrial enterprise such as just-in-time manufacturing (lean manufacturing), flexible manufacturing, cellular manufacturing, agile manufacturing, concurrent engineering, etc. the complexity measures are defined as not only intrinsic to the system being studied but also depend on extrinsic properties of the observer. although most measurements were concentrated on operational measures, both structural and operational characteristics are important to the performance of the system as a whole. operational complexity measures the uncertainty associated with the material and information flows of the system (arteta and giabchetti 2004). the complexity levels in industrial firms are estimated through several case studies based on general framework which includes a questionnaire focusing on each issue in a firm (garbie and shikdar 2010). this paper is organized into several sections. section 1 presents the importance of complexity concept. section 2 reviews previous research work about the complexity. analysis issues are explained in section 3. section 4 presents a hypothetical example. conclusions and recommendations for further work will be introduced in section 5. 2. literature review several research works have been published in this area but none of work mentioned the concept of design for complexity. a methodology based on a simulation model to analyze the complexity in mixed-model assembly production systems was suggested by (kamrani and adat 2008). a measurement framework to analyze the structural properties of the enterprise system was presented (giabchetti et al. 2003). a complexity measure was developed for the business process level inside the organization on one product (e.g. the prepaid phone card) (arteta and giachetti, 2004). a new metrics for assessing the structural complexity of system configurations was suggested (kuzgunkaya and elmaraghy 2006). they based on machine complexity, buffer type complexity, and material handling system complexity as a structural complexity although they can be considered as an operational or dynamical complexity. product, process, and operational complexity in modeling and assessment of manufacturing complexity were introduced (elmaraghy and urbanic 2003, 2004; elmaraghy et al. 2005). each assessment was evaluated independently. three different types of complexity were suggested (bazarth et al. 2009) to represent and model supply chain complexity such as: upstream complexity, internal manufacturing complexity, and downstream complexity. they used these complexities to study the impact effect on a manufacturing plant performance. a coding system of machines, buffers, and material handling equipment to measure complexity based on time-independent complexity (static or structural complexity) of those major components was classified and designed by (elmaraghy et al. 2005). the structural (static) complexity measure to evaluate the complexity in mixed assembly lines was used (hu et al. 2008). several case studies involving 14 italian companies were conducted (perona and miragliotta 2004) to investigate how complexity can affect a manufacturing company's performance. an analytical model for measuring system complexity was presented based on information entropy and probability distribution of resource allocations (cho et al. 2009). operational complexity was measured as a function of cost through supply chain management systems (wu et al. 2007). they indicated that inventory costs are associated with operational complexity. an entropic related to complexity measures to quantify the complexity associated with information content of schedules and variations between schedules was used (huatuco et al. 2009). yang (2010) presented a computational approach to investigate effecting of scheduling with processing 3 complexity analysis of industrial organizations based on a perspective of systems engineering analysts time on two stage hybrid flow shop systems. a mathematical model was suggested to identify functional requirements and the associated design parameters regarding complexity level (tomiyama et al. 2007). kashyap and sinha (2011) studied the complexity related to the mental fatigue required of a person for doing a specific job. while they used how to manage stress, they developed a general model to estimate overall complexity of a profession although they tried to evaluate the job complexity of an engineer. mazur and chen (2011) presented organizing work, communication and managing conflict as the most important issues among team members to complete the project. they considered multifunctional knowledge, teamwork capabilities and working relationships with organizing work as a complex problem. 3. analysis issues to identify the main issues of industrial organizations complexity (ioc), there are five important questions to be asked such as the following to describe how the complexity of industrial enterprises can be studied. these questions were presented by (garbie and shikdar 2011) in analyzing and estimating the complexity levels in industrial organizations. 1. how the complexity issues of an industrial organization is identified and analyzed? 2. how the complexity level of an industrial organization is estimated? 3. how can an industrial organization reduce its complexity? 4. which issues are more important than others? 5. how can industrial organizations identify the adverse factors for reducing complexity? based on these questions, full analysis of complexity regarding these issues will be analyzed and explained through the recommended issues from perspective of systems engineering analysts. these can be represented into four main phases as follows (fig. 1). figure 1 shows a briefly industrial organizations complexity procedure consisting of four phases. each phase will be discussed with its associated issues. it can be noticed from fig. 1 that the procedure of analyzing industrial organizations complexity follows the four phases parallel. based on these concepts and issues mentioned in the previous sections, it can be noticed that the industrial organization complexity (ioc) consists of major issues. these issues are: industrial organization vision complexity (iovc), industrial organization design complexity (iodc), industrial organization operating complexity (iooc), and industrial organization evaluating complexity (ioec) (as shown in fig. 1). the mathematical model of industrial organization components and the corresponding complexity relationships between them in order to emphasize on particular vision, design, operating, and evaluating is presented in the following equations (1 and 2). as each component or element in these systems is a potential source of uncertainty (due to its state), the measuring of complexity for each one is highly valuable. then, iocl is clearly modeled as the following eq. (1) as a function of previous sub-complexities. eq. (1) can be rewritten as eq. (2). each term represents sub-complexity measure of complexity measure of industrial organization (iocl). adding these terms with relative weights is considered. these weights can be used as a reason existing to differentiate between major issues of complexity. iocl = f (iovc, iodc, iooc, ioec) (1) iocl = wiovc [iovc] + wiodc + wiooc [iooc] + wioec [ioec] (2) figure 1. four phases for design for industrial enterprises complexity 4 i.h. garbie and a.a. shikdar where: iocl = industrial organization complexity level, iovc = industrial organization vision complexity, iodc = industrial organization design complexity, iooc = industrial organization operating complexity, ioec = industrial organization evaluation complexity. wiovc, wiodc, wiooc, and wioec are relative weights of organization vision, organization structure, organization operating, and organization evaluation, respectively. because the trade-offs frequently exist between these objectives, a comprehensive analysis for each individual measure is needed. the value of these relative weights may reflect the system analyst's subjective preferences based on his/her experience or can be estimated using tools such as analytical hierarchy process (ahp). in this paper, the relative weights using the ahp are estimated and changed frequently according to the new circumstances by decision maker or a group of decision makers (abdi 2006). these groups are represented in senior management level, manufacturing and/or production engineers, plant managers, operators, and suppliers. these relative weights can be estimated using ahp according to the next matrix. for example, suppose , then this means that industrial organization design complexity (iodc) is four times more important than operating complexity (iooc). 3.1 phase 1: industrial organizations vision complexity the industrial organizations vision complexity is the first step in the design for industrial enterprises complexity. the major issue of this phase is how to collect the main components (elements) of industrial organizations vision. the organization vision usually specifies what supply chain management (scm) representing in number of suppliers (nos), demand variability (dv) representing in number of customers (noc), introducing s new product (np), product life cycle (plc) representing also in product development (pd), and time to market (ttm) requires and how they are affected and effecting on the complexity of industrial organizations. the industrial organizations vision complexity (ievc) will be represented mathematically as a function of these issues as shown in the following eq. (3). ievc = f (scm (nos), dv (noc), np, plc (pd), ttm) (3) 3.2 phase 2: industrial organizations design complexity the second phase, industrial organizations design complexity (iodc) procedure is the designing for system complexity itself. it is mainly concerned with different elements to represent the complexity of it. these elements are: product structure and design (psd), system design (sd), and manufacturing strategies (ms). for each main element, there are several sub-main elements which play an important role in the value of complexity. for example, the psd has four different types to represent the complexity in the product design such as number of parts per product (nnp), number of operations per part (nop), processing or manufacturing time per operation (pt), product size and weight (psw). all of them have a significant effect on the complexity of manufacturing/production process. system design (sd) is playing a major role in complexity in industrial organizations. it can be observed how complexity is the analysis of the system design (sd). the sd divides the complexity analysis into three major issues: production system size (pss), material handling system (mhs), and plant layout system (pls). for the pss, there are three different classifications of production system: small-sized production system (ssps), medium-sized production system (msps), and large-sized production system (lsps). each classification type represents or introduces a significant effect on complexity. also, the material handling system (mhs) and plant layout system (pls) play an important role in identifying the complexity in industrial organizations. the mhs consists of material handling equipments (mhe) with different types of equipments (e.g. conveyor, trucks, forklifts, crane, etc.), material handling storage system (mhss) (e.g. manual or automated storage and retrieval), and identification systems (is). how it can be thought about all previous components and degree of complexity related to each one. facility planning or plant layout system (pls) has a significant effect on complexity by different types of configuration. it can be seem that functional layout (fl) or process layout is more complex than product layout (pl) and/or cellular layout (cl). complexity with respect to manufacturing strategies (ms) is totally different than (sd) and (psd) because it was looking for which strategy can be applied. complexity in lean manufacturing (lms) for example is affected by increased product variety if it is comparing with product layout (pl) in mass production system. but in general, complexity in lean manufacturing system (lms) is low comparing it with agile manufacturing system (ams), flexible man5 complexity analysis of industrial organizations based on a perspective of systems engineering analysts ufacturing system (fms), and reconfigurable manufacturing system (rms). regarding complexity in ams, fms, and rms, they can deal easily with any changes in product design (modifications), unpredictable demand, etc. but infrastructure of these systems itself is more sophisticated and complicated to lead these systems to be more and more complex. the mathematical expression of industrial organizations design complexity (iodc) can be modeled as the following eqs. (4 and 5) in different facets. iodc = f (psd, sd, ms) (4-1) iodc = wpsd (psd) + wsd (sd) + wms (ms) (4-2) iodc = f (npp, nop, pt, psw, mhs, pss, pls, lms, fms, ams, rms) (5) the wpsd wsd and wms are relative weights of product design, system design and manufacturing strategies, respectively. eq. (5) can be rewritten as a very board general representing the lowest level of information in industrial organization design complexity as the following eq. (6). (6) 3.3 phase 3: industrial organizations operating complexity once the industrial organization system vision and design complexities become available to the perspective of industrial systems analysts and designers, the design for complexity related to the system operating becomes urgent to be analyzed and evaluated. this phase may involve further activities in data collection and processing. design for industrial organizations operating or dynamic complexity (iooc) is different than previous ones (system vision, and system design). in this analysis, it can be noticed that there are three major items of complexity: resource status of operating complexity (rsoc), work in progress complexity (wipc), and business operations complexity (boc). resources mean equipment (e.g. machining equipment, forming equipment, material handling equipment, etc.) and human. in this analysis, it will be concentrated on the resource reliability (rr), resource capability or flexibility (rc), resource utilization (ru), resource scheduling/ rescheduling (rs/r), and human scheduling/rescheduling (hs/r). for example, maintenance level plays a vital role in resource reliability. this means that the lower the maintenance level is the lower in machine capacity (reliability). also, work in progress complexity (wipc) representing in buffer between workstations or departments is considered one of measuring degree of complexity inside the production plant (factory). there are several important issues that can be used to evaluate the business operations complexity (boc). these issues are: organization plans (op), organizing work (ow), structure of management levels (sml), staffing developing and motivation (sdm), decision making (dm), communication between and within management levels (cml), managing conflict, change, culture and stress (mcs), and finally leadership roles in management (lr). then, industrial organizations operating complexity (iooc) can be modeled to measure or evaluate the complexity level as the following eqs. (7 and 8). iooc = f (rsoc, wipc, boc) (7-1) iooc = wrso (rso)+wwip (wip)+wboc (boc) (7-2) iooc = f (rr, rc, ru, rs/r, hs/r, wip (bs), op, ow, sml, sdm, dm, cml, mcs, lr) (8) 3.4 phase 4: industrial organizations evaluation complexity the fourth phase in the design for industrial organizations complexity (ioec) procedure is the complexity regarding the system evaluation. as industrial organizations have a great impact on the performance measurements, they still have a problem in measuring these complexities especially regarding selection of the objectives. in this paper, there are five different objectives that can be used to evaluate the complexity. they are: product cost (pc), response (r) representing in manufacturing lead time, system productivity (sp) representing in system utilization, product quality (pq) representing in number of scrap (defect rate), and appraising and rewarding performance (arp). they also can be modeled mathematically as the following eq. (9). ioec = f (pc, r, sp, pq, arp) (9) 4. a hypothetical example this numerical example is used to estimate the level of complexity of industrial organizations. this can be implemented through three sequential steps. the first step is used to estimate the relative weights of industrial organizations based on the four major issues, iovc, iodc, iooc, and ioec. the relative weights between these issues are estimated by using the ahp as the following matrix based on the pair wise comparisons of the four major issues. it can be noticed that the relative weights of iovc is estimated to be equivalent to the iodc, twice as important as the iooc and four times more important than the ioec. the iodc is estimated to be three times more important than the ieoc and four times more important than ioec. the 6 i.h. garbie and a.a. shikdar iooc is also estimated to be twice as important as the ieec. as a result, the values of relative weights are estimated at 0.35, 0.40, 0.16, and 0.09 for iovc, iodc, iooc, and ioec, respectively. eq. (2) is used to calculate the industrial enterprises complexity level (iocl) incorporating of the four major issues as the following eq. (10). iocl= 0.35 [iovc] + 0.40 [iodc] + 0.16 [iooc] +0.09 [ioec] (10) in the second step, the multiple regression models are used to estimate and formulate the complexity levels for the major and sub-major issues. the values were assumed to follow a uniformly distributed random variable with known parameters as shown as in table 1 for 50 generating values for each issue. the multiple regression models for industrial organizations vision complexity, industrial organizations design complexity (product structure and design, system design and manufacturing strategies), industrial organizations operating complexity (resource status and business), and industrial organizations evaluation complexity are presented as the following eqs. (1117), respectively. this was done by using minitab statistical software package to generate a random variable and formulating a multiple regression model. ievc = -0.03560.0374 (nos) 0.000013 (noc) + 0.0065 (pd) + 0.0842 (ttm) (11) psdc = 0.359-0.00436 (npp) + 0.0353 (nop) + 0.0123 (pt) + 0.0759 (ind1) 0.0420 (ind 2) + 0.0241 (ind 3) 0.0039 (ind 4) ind 1, ind 2, ind 3, and ind 4 are used for the product size and weight. (12) sdc = 0.2440 + 0.00479 (ssps) -0.00795 (msps) + 0.00018 (lsps) + 0.0848 (fl) 0.0681 (cl) + 0.0609 (pl) + 0.092 (mhss) + 0.0290 (mhe) + 0.189 (is) (13) msc = 0.441 -0.059 (lms) + 0.245 (fms) 0.028 (ams) 0.024 (rms) (14) rosc = 0.5980.379 (rr) + 0.509 (rc) + 0.113 (ru) 0.50 (rs/r) + 0.233 (hs/r) (15) boc = 0.980 0.0029 (op) 0.0076 (ow) + 0.0015 (sml) + 0.171 (s) +0.234 (d) + 0.0050 (m) 0.0266 (dm) 0.0642 (cml) 0.134 (mcs) 0.289 (lr) (16) ioec = 0.457 + 0.00210 (pc) + 0.0088 (r) 0.172 (sp) 0.40 (pq) + 0.117 (arp) (17) the third step is used to determine the complexity level of iovc, psdc, sdc, rosc, and ioec, respectively by optimizing eqs. (11-17). the constraints are identified based on the range values of each sub-issue which are listed in table 1. to determine the relative weights between psdc, sdc, and msc and between rsoc, wipc, and boc, the following matrices are used to estimate these values and the pair wise comparison between product structure and design, system design and manufacturing strategies is illustrated in these matrices. with respect to iodc, it can be noticed that a psdc is estimated to be equivalent to the sdc and twice as important as a msc. the same estimation is done relate to the sdc with psdc and msc. regarding the iooc, it can be noticed that rsoc is estimated to be four times more important than the wipc and two times as important as the boc. the boc is also estimated to be four times more important than wipc. as a result, the relative weights for iodc between psdc, sdc, and msc are estimated at 0.40, 0.40, and 0.20, respectively. also, the relative weights for iooc between rsoc, wipc, and boc are estimated at 0.544, 0.110, and 0.345, respectively. the results of complexity level in each sub-issue and major issue are illustrated in table 2. then, eq. (10) is used again to calculate the global complexity level based on the information taken from table 2 and the estimated relative weights as the following. iocl = 0.35 [0.4582] + 0.40 [0.3099] + 0.16 [0.3023] + 0.09 [0.4570] = 0.3740 it can be noticed from the results given from the previous equation and table 3 that the level of complexity in this enterprise equals to 37.40% and this value seems ranked in a medium range. it seems that vision complexity represents more important (0.16/0.3740 = 42.78%) than design complexity (33.15%). the percentage values of operating and evaluating complexity are 13.36% and 10.70%, 7 complexity analysis of industrial organizations based on a perspective of systems engineering analysts table 1. data for complexity issues based on uniform distribution, u [a, b] u represents the uniform distribution with [a] the lower limit and [b] is the upper limit, *represents the product size (small, small-medium, medium, medium-large, large), *'represents the product weight (light, light-medium, medium, medium-heavy, heavy), **represents the degree of variety in functional layout, **'represents the number of manufacturing cells, **''represents the variety of models (single, batch, and mixed), ***represents the appraising and rewarding performance as a percentage. 8 i.h. garbie and a.a. shikdar respectively. these values can be totally different from an industrial organization to another one based on the sub-major issues and the relative weights between submajor and major issues. regarding the complexity reduction, it can be noticed that if the values of submajor issues reduced, the complexity level will be reduced too. it also seems that the vision complexity and design complexity represent the adverse factors for reducing complexity. with respect to iovc, it is not easily to reduce the number of suppliers, for example, and it should increase this number. also, for iodc, the number of parts per product, number of operations per part, processing time per operation, product size and weight, production system size are representing adverse factors for reducing complexity because they cannot be changed. 5. conclusions it can be noticed from this analysis that complexity issues are not simple. it required emphasize on each of the main issues and the sub-main. hence, industrial organizations complexity (ioc) will involve four major issues: vision complexity, design complexity, operating complexity, and evaluation complexity. analyzing complexity in industrial enterprises is based on the four main issues and it can be mathematically expressed as a total global function as the following eq. (18). ioc = f (scm (nos), dv (noc), np, plc (pd), ttm, nnp, nop, pt, psw, mhe, mhss, is, ssps, mdps, slps, fl, cl, pl, lms, fms, ams, rms, rr (ml), rc < ru, rs / r, hs, / r, wip (bs), op, ow, sml, dm, cml, lr, sdm, mcs, pc, r, sr, pq, arp) (18) the ioc issues should be dynamic and they should evolve with and adapt to the changing internal and external environment. until now, the ioc remains a research topic of immense international interest. this will represent the degree of freedom of industrial organizations designers to identify which issue is more significant than others. the main contribution in this paper is how to identify and model the components of industrial enterprises complexity in any industrial firms (organizations) at any time considering these components. the authors intend to extend this research to apply this analysis and formulation the suggested model to estimate and optimize the degree of complexity in any industrial enterprises towards full validation of the complexity theory which will be discussed and presented in the future research. acknowledgments the authors would like to acknowledge the financial support provided by the sultan qaboos university (grant no. ig/eng/mied/10/01) to carry out this research work. table 2. complexity levels in the four phases of industrial enterprises table 3. values of major issues regarding the complexity level 9 complexity analysis of industrial organizations based on a perspective of systems engineering analysts references abdi mr (2006), performance evaluation of reconfigurable manufacturing. the proceedings of the 36th cie conference on computer and industrial engineering, taipei, taiwan 2479-2491. arteta bm, giachetti re (2004), a measure of agility as the complexity of the enterprise system. robotics and computer-integrated manufacturing 20:495-503. bazarth cc, warsing dp, flynm bb, flynn ej (2009), the impact of supply chain complexity on manufacturing plant performance j. of operations management 27:78-93. cho s, alamoudi r, asfour s (2009), interactionbased complexity measures of manufacturing systems using information entropy. int. j. of computer integrated manufacturing 22(10): 909922. elmaraghy wh, urbanic rj (2003), modeling of manufacturing systems complexity. annals of the cirp 53(1):363-366 elmaraghy wh, urbanic rj (2004), assessment of manufacturing complexity. annals of the cirp, 53(1):401-406. elmaraghy ha, kuzgunkaga o, urbanic rj (2005), manufacturing systems configuration complexity. annals of the cirp 54:445-448. garbie ih, shikdar a (2009), determination of complexity levels in industrial firms. proceedings of the int. conference of industrial engineering research (ierc), miami, florida, usa, may 30june 3:1423-1428. garbie ih (2009), a vision for reconfiguring industrial organization due to the global recession. proceedings of the 39th of int. conference on computers and industrial engineering, france, 658-663, july 6-8, 2009. garbie ih, shikdar a (2010), complexity level in industrial firms: case studies and implementation. proceedings of the 2010 international conference on industrial engineering and operations management (ieom 2010), dhaka, bangladesh, january 9-10, 2010. garbie ih (2010), a roadmap for reconfiguring industrial enterprises as a consequence of global economic crisis (gec). j. of service science and management 3(4):419-428. garbie ih, shikdar a (2011), analysis and estimation of complexity level in industrial firms. int. j. of industrial and systems engineering 8(3):1-23. giabchetti re, martinez ld, saenz oa, chen cs (2003), analysis of the structural measures of flexibility and agility using a measurement theoretical ramework. int. j. of production economics 86: 47-62. hu sj, zhu x, koren wy (2008), product variety and manufacturing complexity in assembly systems and supply chains. annals of the cirp 57: 45-48. huatuco lh, efstathiou j, calinescu a, sivadasan s, kariuki s (2009), comparing the impact of different rescheduling strategies on the entropic-related complexity of manufacturing systems. int. j. of production research 47(1):4305-4325. kamrani ak, adat a (2008), manufacturing complexity analysis: a simulation-based methodology-collaboration engineering. theory and practice edited by karmrani ak springer, 227-248. kashyap n, sinha s (2011), estimation of job complexity in an industrial organization. int. j. of industrial and systems engineering vol. 7(1):2644. kuzgunkaya o, elmaraghy ha (2006), assessing the structural complexity of manufacturing systems configurations. int. j. of flexible manufacturing systems 18:145-171. mazur lm, chen sj (2011), a task-member assignment model for complex engineering. int. j. of industrial and systems engineering 7(1):1-25. perona m, miragliotta g (2004), complexity management and supply chain performance assessment. int. j. of production economics 90:03-115. tomiyama t, amelio vd, urbanic j, elmaraghy hw (2007), complexity of multi-disciplinary design. annals of the cirp 56(1):85-188. wu y, frizelle g, efstathiou j (2007), a study on the cost of operational complexity in customer-supplier systems. int. j. of production economics 106:217-229. yang j (2010), a new complexity proof for the twostage hybrid flow shop scheduling problem with dedicated machines. int. j. of production research 48(5):531-1538. the journal of engineering research (tjer), vol. 17, no. 2, (2020) 89-99 *corresponding author’s e-mail: alawim@squ.edu.om doi:10.24200/tjer.vol17iss2pp89-99 modeling, investigating, and quantification of the hot weather effects on construction projects in oman hajar al balushi 1, mubarak al-alawi 2, *, and mohammed al shahri2 1 civil engineer, petrofac e&c, muscat, oman 2 dept. of civil and architectural engineering, sultan qaboos university, muscat, oman abstract: the construction industry is recognized as one of the industries most exposed to climatic conditions. construction projects are mainly executed in an outdoor environment and the activities are considered weather-sensitive. severe weather conditions can directly affect the productivity and efficient operation of construction projects. in addition, the weather is known to be one of the main factors that decrease labour productivity in construction projects causing project delays, cost overruns, and contractual claims between contractors and clients. oman has a hot climate with very high temperatures in summer, warm winters, and low annual rainfall. during extremely hot weather, labour productivity significantly decreases, as construction work may stop partially or completely, therefore, this paper investigates and quantifies the hot and humid weather effects in construction projects in oman. a construction productivity model was developed using the work/rest schedule proposed by the national institute for occupational safety and health (niosh), usa. the daily weather temperature and relative humidity of muscat were inputs into the model and the expected productivity in terms of working hours were the output of the model. the model was applied to case studies, which involved three completed construction projects under different testing scenarios in muscat. results indicate that implementing the influence of hot and humid weather can lead to an extension of 3–38% longer project duration compared to the planned duration. keywords: construction; heat stress; productivity; project duration; modeling. نمذجة وبحث تأثیر الطقس الحار على مشاریع البناء في عمان ومحمد الشحري ، *مبارك العلوي، ھاجر البلوشي بیئة یتم تنفیذ مشاریع البناء بشكل أساسي في و تعتبر صناعة البناء من أكثر الصناعات المعرضة للظروف المناخیة: الملخص خارجیة وتعتبر أنشطتھا حساسة للطقس. یمكن أن تؤثر الظروف الجویة القاسیة بشكل مباشر على إنتاجیة وكفاءة عملیات یة التي تقلل من باإلضافة إلى ذلك ، من المعروف أن تأثیرات الطقس على صناعة البناء ھي أحد العوامل الرئیس ،مشاریع البناء ، مما یتسبب في تأخیر المشروع ، وتجاوز التكالیف ، والمطالبات التعاقدیة بین المقاولین وأصحاب المشروع. إنتاجیة العمالة أثناء الظواھر الجویة شدیدة الحرارة، وشتاءدافئ درجات حرارة عالیة جدا في الصیف وب ویمتاز مناخ عمان جاف وحاریعتبر ھذه الورقة البحثیة إلى نمذجة ف . تھدتوقف أعمال البناء جزئیًا أو كلیًاؤدي إلى مما یقد تنخفض إنتاجیة عمال البناء بشكل كبیر نسبة تأثیرات الطقس الحار والرطب في مشاریع البناء في عمان. تم تطویر نموذج إنتاجیة البناء باستخدام جدول قیاسبحث وو درجة حرارة الطقس الیومیة تم استخدام). nioshالمعھد الوطني للسالمة والصحة المھنیة ( لمقترح منالراحة اإلى العمل مدخالت في النموذج وكانت اإلنتاجیة المتوقعة من حیث ساعات العمل ھي ناتج النموذج. تم كمسقط لمحافظةوالرطوبة النسبیة ر الطقس الحار النتائج إلى أن تأثی أشارتسیناریوھات اختبار مختلفة. عدة في ظل و في ثالث مشاریع انشائیةتطبیق النموذج ٪ مقارنة بالمدة المخطط لھا.38إلى 3والرطب یمكن أن یؤدي إلى إطالة مدة المشروع بنسبة .النمذجة ؛مدة المشروع ؛اإلنتاجیة ؛التشیید؛ اإلجھاد الحراري :الكلمات المفتاحیة modeling, investigating, and quantification of the hot weather effects on construction projects in oman 90 1. introduction construction projects are executed in an outdoor environment, most of its activities being conducted by workers outdoors (al shebani and wedawattab, 2014) and therefore, are affected by various weather conditions (el-rayes and moselhi, 2001). weather events such as extreme cold, heat, wind, snowfall or precipitation are recognized among the factors causing noticeable project delays, cost overruns, and contractual claims. it can significantly affect a project’s schedule and produce significant deviations from the baseline schedule (shahin, abourizk, and mohamed, 2011; ballesteros-péreza, et al., 2017). the construction industry needs to take note of different weather conditions and improve their working environment to make it safer. weather simulation can assist in optimizing the project’s schedule to make it more robust in terms of weather impacts. weather simulation models can be seen as extremely complicated random number generators that outputs resemble weather circumstances at a certain place (kerkhove and vanhoucke, 2017). the simulation models follow a pattern to construct the necessary models and quantify the impact of adverse weather conditions on actual operation projects: 1) study of construction processes, 2) understanding the effect of weather on those processes, 3) identify the weather variables that influence the construction processes, 4) search for weather information sources, 5) choosing a modelling method, 6) develop an appropriate tool that will generate weather variables, 7) apply the model to real projects to validate the developed model (moselhi, gong, and el-rayes, 1997). although the processes of constructing a model to allow integrating weather effect in construction projects planning and operations seem to be well addressed, however, it is a challenging process in terms of finding the correct data, and generalize the model to all construction projects. for example, productivity data for different construction trades must be available to model the construction operation and not all construction companies keep track of such data. the national institute for occupational safety and health (niosh) issues regulations and work guidelines regarding occupational exposure to high temperatures and hot environment. such regulations can be used to construct a model to help predict weather variables that affect the performance of the construction operation and will allow predicting its negative impact on the duration of the construction project. thus, this research will use the threshold limits values published by niosh in hot weather region like experienced in oman. the climate of oman can be described as dry and hot with low annual rainfall, very high temperatures in summer and warm winters. during extremely hot weather conditions, labour efficiency may tremendously decrease, because construction work may stop partially or entirely. moreover, workers have reduced working hours due to the ministry of manpower has issued a directive to all companies, especially construction firms in the country, to stop all outdoors activities between 12.30 pm and 3.30 pm, from june to august (ministry of manpower, 2008), when the weather is extremely hot and humid. the effect of such low productivity and shorter working hours during the hot summer months in oman may cause construction delays and therefore, additional costs of the project. the consequence is a financial loss that must be borne by either the contractor, the client or both (ballesteros-péreza, et al., 2017). therefore, this research aims to model, investigate, and quantify the effect of hot and humid weather on construction projects duration in oman. 2. background extreme weather conditions can cause productivity to fall and delays in the project, especially in the construction industry, because most of its activities being managed by labours outdoors (al shebani and wedawattab, 2014). hot weather affects construction worker physiologically as well as psychologically. physiologically, individuals may suffer heat stress or stroke in the hot weather. all of this may affect a worker psychologically; to the point where he or she wishes not to be exposed to the unpleasant working conditions and become demoralized and less productive (ibbs and sun, 2017; koehn and brown, 1985). based on the findings of yildirim et al. (2009), grimm and wagner (2009), and thomas and yiakoumis (1987), there is an inverse relationship between the increase in temperature and labour productivity. previous studies showed that at a temperature between 100° 110° f the quality of the work declined, also working at this range of high temperatures resulted in serious health hazards and low productivity (koehn and brown, 1985). palmer and creagh (2013), grimm and wagner (1974), and thomas and yiakoumis (1987) noted that an increase in humidity could also be adversely affecting labour productivity. ibbs and sun (2017) concluded that humidity has a critical effect on productivity, but its impact is smaller compared to that of temperature. studies also indicate that temperatures above 110° f and below -10° f with humidity above 50% are intolerable, and it is difficult to achieve construction operations (kohen and brown, 1985). though extreme weather condition in construction projects is recognized as one of the factors causing the productivity to fall, producing noticeable project delays, cost overruns, waste of resources and contractual claims (apipattanavis, et al., 2010; chan and au, 2008; el-rayes and moselhi, 2001; lee et al., 2017), it is also reported to be one of the main factors that influence financial performance and business hajar al balushi, mubarak al-alawi, and mohammed al shahri 91 continuity (moselhi, gong, and el-rayes, 1997; al shebani and wedawattab, 2014; guo, chen, and chiu, 2017). weather can affect a construction process in various ways; it can slow down the works by lowering the performance of construction labours and materials and can cause a temporary work stoppage, which therefore affects the plan and schedule of the project (al shebani and wedawattab, 2014). thus, understanding inclement weather influence on any construction project can reduce claims and arguments caused by delays (apipattanavis, et al., 2010; chan and au, 2008; dytczak et al., 2013). previous studies investigated the impact of inclement weather on construction activities and productivity. moselhi et al. (1997) quantified the influence of weather conditions on daily construction activity; they developed an automated decision support system called weather to study the impact of adverse weather conditions on the labour productivity and workstoppage of construction operations. the developed weather model estimates construction productivity as well as the duration of construction activities and weather patterns in different modes to improve the accuracy of the planning and scheduling. a mathematical study of weather by koehn and brown (1985) employed some non-linear equations to examine the effect of weather changes on the productivity rate. their investigation suggests a clear relationship between overall construction performance and weather-related factors such as temperature and humidity it shows that at a temperature between 100° and 110° f the quality of the work declined, also working at this range of high temperatures resulted in serious health hazards and low productivity. koehn and brown (1985) found that construction productivity decreases rapidly at elevated temperatures, and there is an adverse relationship between construction productivity and humidity. the quantification of weather impacts on productivity was reported by the national electrical contractors association (neca) (hanna, 2004). the research engaged two travelling electricians installing electrical boxes and duplex outlets while operating in an environment chamber that monitored the temperature and humidity. the study found that productivity levels differ depending on humidity and temperature. grimm and wagner (1974) studied the productivity in masonry construction over a period of nine months during the construction of 283 test walls and published a diagram showing the impact of temperature and humidity on the productivity. the research found a decline in productivity as temperature and relative humidity deviated. they reported that relative humidity had a much greater effect on productivity rate. ahuja and nandakumar (1985) and kavanagah (1985) measured the impact of weather as a percentage in their construction modelling and analyzed how frequently weather resulted in decreased activity. abourizk and wales (1997) research used a general regression neural network to study the relationship between weather variables and earthwork productivity. this study investigated three weather variables: precipitation, daily maximum temperature, and daily minimum temperature. it demonstrated the impact of these weather variables has on a project schedule by calculating the duration for the same project according to different start dates. the south dakota department of transportation (kenner, et al., 1998) adopted a pragmatic solution by combining their construction records with weather records to calculate contract time and determine time extensions for adverse weather. yildirim, koyuncu, and koyuncu (2009) findings showed a negative association between temperature increase and labour productivity level. palmer and creagh (2013) state that the rise in humidity due to climate change reduces labour productivity. studies also indicate that temperatures above 110° f with humidity above 50% are intolerable and all useful work stops, therefore, it is difficult to achieve efficient construction operations (arditi, 1985). thomas et al. (1990) described the factor model for evaluating the productivity of labourintensive construction activities. the validity of the factor model is demonstrated by considering the effect of temperature and relative humidity on productivity. thomas et al. (1990) also noted that the increase in humidity could be adversely affecting labour productivity. ibbs and sun (2017) concluded that humidity has a critical effect on productivity, but its impact is smaller compared to that of temperature. shahin et al. (2011) proposed a simulation model that quantify the effects of extreme weather events on construction projects and to assist in project planning and decision support. miroslaw et al. (2013) presented a numerical procedure to identify the efficient construction project structure and a corresponding schedule. it addressed the impact of inclement weather conditions on technological operations. marzoughi et al. (2018) developed a model using multivariate statistical techniques and an analytical network (anp) method to assess the duration of project operations, taking into consideration the effect of weather. 3. heat stress this section covers a brief introduction about heat stress in hot environments and present recommendations for workers exposed to heat and hot environments derived from the occupational exposure to heat and hot environments published by national institute for occupational safety and health (niosh) (niosh, 2018). total heat stress is controlled by the three factors, the heat generated in the body, the heat gained from the environment, and the heat lost to the environment from the body. environmental and/ or metabolic heat modeling, investigating, and quantification of the hot weather effects on construction projects in oman 92 stress results in physiological reactions (heat stress) to promote heat transfer from the body back to the surroundings to maintain core body temperature. many of the heat exposure responses in the body are desirable and useful. however, at some level of heat stress, the compensatory mechanisms are no longer capable of keeping body temperature at the rate needed for normal body tasks. consequently, there is an increase in the risk of heat-related illnesses, disorders, and other hazards. workers who are subjected to extreme heat or work in indoor or outdoor hot environments or even those involved in exhausting physical activity may be at risk for heat stress. those at risk of heat stress include outdoor workers and workers in hot environments, for example, firefighters, bakery workers, farmers, construction labours, miners, boiler room workers, and factory workers. outdoor labours are subjected to a great deal of exertional and environmental heat stress. the recommendations are intended to provide limits of heat stress so that heat-related illnesses and disorders are reduced. niosh proposes that employers implement measures to protect the health of labours exposed to heat and hot environments. employers need to monitor environmental heat and determine the metabolic heat produced by workers (e.g., light, moderate, or heavy work). based on the increase in risk, additional modifications (e.g., worker health interventions, clothing, and personal protective equipment) may be required to shield workers from heat stress. in hot conditions, physiological monitoring and medical screening are suggested. employers, supervisors and workers need to be trained to recognize symptoms of heat-related illness, adequate hydration, care and use of heat-protective garments and equipment, the impact of multiple heattolerance risk variables (e.g., drugs, alcohol, obesity, etc.), the significance of acclimatization, the significance of reporting symptoms and suitable first aid. employers should have a plan to acclimatize a new employee because it has been shown that the lack of acclimatization is a significant factor associated with worker heat-related illness and death. niosh recommends that employers provide the means for adequate hydration and that their employees be encouraged to hydrate themselves with drinking water below 15 ° c (59 ° f) close to the work zone. workers in heat less than 2 hours and involved in moderate work activities should drink 1 cup (8 oz.) of water every 15–20 minutes. furthermore, employers should establish a work/rest schedule and provide a cool environment for employees to rest and recover (e.g., air-conditioned or shaded). these essentials are intended to protect the health of workers from heat stress or injuries in hot environments. 4. work-rest hours prediction model this section describes the research methods that have been used for assessing the relationship between the high temperature (t) and relative humidity (rh) and the productivity of the labours. two primary data were collected for this research. first, historical weather data for daily temperature and relative humidity in muscat city were collected from the public authority for civil aviation (paca) for the years from 2015 to 2018. the data received from paca were not clean, with missing records and limited years; only from 2015 until 2018 while the data needed should be at least from 2008. therefore, actual weather data from the national oceanic and atmospheric administration (noaa) was used. the data from noaa covered daily temperature in muscat from 2001 until 2018. to cover the daily relative humidity for the years from 2001 to 2018; daily relative humidity received from paca was repeated every two years, with an assumption that the relative humidity will not differ much within two years and it was observed that almost same relative humidity was repeating every two years. the database was established using the daily temperature (mean, maximum and minimum) from noaa and daily relative humidity (mean, maximum and minimum) from paca. the temperature ranged from a low of 12.4 °c to a high of 47.5 °c, and the relative humidity was in the range of 5%-100%. it was observed that the temperature and relative humidity were extreme in the months from april to september while it is normal during the remaining months of the year. therefore, weather data from april to september was considered in this research. to maintain a healthy and safe working environment, the national institute for occupational safety and health (niosh) developed a work/rest schedule for workers. table 1 shows the work and rests time per hour for a certain temperature and humidity to minimize the effect of heat on the labours. the workload category is expressed as worker’s metabolic heat production: light work = <180 kcal/hr; moderate work = 180–300 kcal/hr; heavy work = 300– 415 kcal/hr; and very heavy work = >520 kcal/hr (niosh, 2018). table 1 is used for the quantification of weather effects on labourers' productivity in table 2, which illustrates the percentage of rest per hour due to high temperature. hajar al balushi, mubarak al-alawi, and mohammed al shahri 93 more tables that are comprehensive were developed for the considered months – april to september. data were summarized for each month, and most-likely case and pessimistic cases were defined for each workload category for the data analysis. the most likely case reflected the percentage of rest per hour for the mean adjusted temperature based on mean relative humidity while the pessimistic case is the percentage of rest per hour for the maximum adjusted temperature based on maximum relative humidity. these two cases were defined because they reflect the worst-case scenarios compared to other cases (e.g. minimum adjusted temperature based on minimum relative humidity). the data was analyzed using easyfit software to find out the best-fit probability distribution function that represents the probability of percentage of rest per hour the work may experience for each month. for each month; six probability distributions were found as follows: • light work: most likely and pessimistic • moderate work: most likely and pessimistic • heavy work: most likely and pessimistic randomly generating percentages for rest per hour in for a month will generate a full month affected by the weather condition, which is not logical and not valid. therefore, the probability of occurrence of negative weather effect was calculated from the weather data set. table 3 below illustrates the probability of occurrence of the affected days for each month; that was found using the following equation: the probability of occurrence of an event is defined according to eqn. (1): 𝑃𝑃(𝐸𝐸) = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑜𝑜𝑓𝑓𝑓𝑓𝑜𝑜𝑛𝑛𝑓𝑓𝑛𝑛𝑓𝑓𝑛𝑛 𝑜𝑜𝑛𝑛𝑜𝑜𝑜𝑜𝑜𝑜𝑛𝑛𝑛𝑛𝑜𝑜 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑜𝑜𝑜𝑜𝑜𝑜𝑝𝑝𝑛𝑛𝑓𝑓𝑛𝑛 𝑜𝑜𝑛𝑛𝑜𝑜𝑜𝑜𝑜𝑜𝑛𝑛𝑛𝑛𝑜𝑜 (1) where the number of favourable outcomes is the percentage of rest per hour which is more than 0%, and the number of possible outcomes is the total number of days considered in the data. the probability of occurrence along with the probability of the percentage of expected rest to work were jointly used to generate the expected rest percentage for construction labours. 5. model validation the data needs to be statistically validated. a statistical comparison of the generated output for each month was carried out to ensure that the similarity between the historical data and the generated data is statistically adequate. the rest percentage prediction model was used to generate data for the expected percentages of rest construction labour may experience in months from april to september. table 1. work/rest schedule for workers wearing normal clothing (niosh, 2018). adjusted temperature (c) † light work (minutes work/rest) moderate work (minutes work/rest) heavy work (minutes work/rest) 32.22 normal normal normal 32.78 normal normal normal 33.33 normal normal normal 33.89 normal normal normal 34.44 normal normal normal 35 normal normal 45/15 35.56 normal normal 45/15 36.11 normal normal 40/20 36.67 normal normal 35/25 37.22 normal normal 35/25 37.78 normal 45/15 30/30 38.33 normal 40/20 30/30 38.89 normal 35/25 25/35 39.44 normal 30/30 20/40 40 normal 30/30 20/40 40.56 normal 25/35 15/45 41.11 45/15 20/40 caution 41.67 40/20 15/45 caution 42.22 35/25 caution caution 42.78 30/30 caution caution 43.33 15/45 caution caution 43.89 caution caution caution 44.44 caution caution caution † note: adjust the temperature reading as follows before going to the temperature column in the table: full sun (no clouds): add 13°f, partly cloudy/overcast: add 7° f no shadows visible/work is in the shade or at night: no adjustment per relative humidity: 10%: subtract 8° f, 20%: subtract 4° f, 30%: no adjustment, 40%: add 3° f, 50%: add 6° f, 60%: add 9° f ‡high levels of heat stress; consider rescheduling activities. table 2. percentage of rest per hour for workers wearing normal work clothes. adjusted temperature (c) % rest per hour light work moderate work heavy work 32.22 0 0 0 32.78 0 0 0 33.33 0 0 0 33.89 0 0 0 34.44 0 0 0 35 0 0 25 35.56 0 0 25 36.11 0 0 33.33 36.67 0 0 41.67 37.22 0 0 41.67 37.78 0 25 50 38.33 0 33.33 50 38.89 0 41.67 58.33 39.44 0 50 66.67 40 0 50 66.67 40.56 0 58.33 75 41.11 25 66.67 100 41.67 33.33 75 100 42.22 41.67 100 100 42.78 50 100 100 43.33 75 100 100 43.89 100 100 100 44.44 100 100 100 modeling, investigating, and quantification of the hot weather effects on construction projects in oman 94 the number of generated instances was equal to the original data and statistical measures such as the mean, the variance, and the standard deviation were used to compare the generated data with the original data. table 4 shows sample results for april. it can be seen that the generated data is reasonably replicating the real data except in the pessimistic scenario for moderate work, the data shows a large difference in the variance. the summary of the comparison between the generated data and the real data for other months is as follows: may: there is a big difference in the mean, variance and standard deviation between the generated and original data in the light work category for both scenarios (a most likely and pessimistic). in moderate work, for a pessimistic scenario, and in heavy work, a most likely scenario. while the pessimistic cases in both moderate work and heavy work, the data shows a large difference in the variance. • june: in all work categories, for the most likely scenario; the data shows a noticeable difference in mean, variance and standard deviation. also, a contrast was found in the mean and variance of the compared data in the pessimistic case in all work categories. • july: the generated data is reasonably duplicating the real data in the heavy work, both scenarios and moderate work most likely cause. however, in the moderate work, scenario pessimistic; and both scenarios in light work, the data shows a big difference in the mean, variance and standard deviation. • august: the mean, variance and standard deviation for the real and generated data were not consistent for the most likely scenario, in light and moderate work category. also, the difference in variance was found to be large in the pessimistic scenario for light work. however, for the remaining scenarios, the data are in good agreement. • september: the generated data replicates the real data fairly except in the moderate work, most likely scenario; which indicates a significant difference in the mean, variance and standard deviation. this discrepancy found in the analyzed data can be because of the assumption that was made earlier. the collected data has two years of relative humidity records and an assumption that the relative humidity will not have a huge annual variance and therefore the two years relative humidity data can be replicated to cover the missing data. 6. application of the model in real construction projects as described in the previous section, the construction productivity model will be applied in three real construction projects. the projects schedules were fed with four different weather effects testing scenarios. the testing scenarios are as follow: • the first scenario (sc-1): the project activities are categorized as a moderate workload, and the weather conditions are most likely. • the second scenario (sc-2): the project activities are categorized as moderate workload, and the weather conditions are pessimistic. • the third scenario (sc-3): the project activities are categorized as heavy workload and the weather conditions are most likely. • the fourth scenario (sc-4): the project activities are categorized as heavy workload, and the weather conditions are pessimistic. these four testing scenarios were used because this research work is motivated to investigate and quantify the effect weather on construction project schedules and such objective can be reached if the projects were tested against the moderate and heavy type of construction work activities. activities performed during the hot and humid season are targeted in this part of the study. in addition, it is hypothetically assumed that the original project schedule has no weather consideration in estimating the project activities durations. however, the usual practice is to extend the activity duration to count for unforeseen extreme weather conditions. table 3. probabilities of occurrence of the affected days for each month. month percentage of days affected light work moderate work heavy work most likely pessimistic most likely pessimistic most likely pessimistic april 0 10 6 38 26 72 may 11 47 41 31 72 93 june 16 44 48 76 84 92 july 7 20 21 47 66 86 august 1 14 10 35 44 76 september 0 11 7 37 37 76 hajar al balushi, mubarak al-alawi, and mohammed al shahri 95 table 4. comparing april generated data against the real data. workload category light work moderate work heavy work scenario most-likely pessimistic most-likely pessimistic most-likely pessimistic best distribut ion function constant =0 gamma frechet beta gamma general extreme value type of data real data generated data real data generated data real data generated data real data generated data real data generated data mean 4.42 4.42 1.92 1.53 19.96 12.18 8.96 8.26 37.00 24.68 variance 213.33 225.34 67.60 52.95 859.53 363.09 265.70 267.42 997.6 2 936.88 std. deviation 14.61 15.01 8.22 7.28 29.32 19.06 16.30 16.35 31.59 30.61 6.1 project a the construction of the project took 365 days. in this research, the construction part only was analyzed, which included; sub-structure works, super-structure works, and masonry works. the critical path of this part was identified, and its duration is 250 days. figure. 1 (a) shows the affected activities by the first testing scenario (sc1) applied to the construction schedule for project a. the increase in activity duration is shown in red font. for the remaining activities, the effect was negligible this is because of a lower percentage of rest per hour in a particular activity and because some of the activities are in october, november, december and january which are not considered in this research. the critical path duration for project a was increased by 41.5 days to reach a total of 291.5 days. figure. 1 (b) shows the construction activities affected by the second scenario. the affected activities are more than that of the first scenario. this increase in affected activities is due to the high percentage of rest per hour for the pessimistic scenario compared to the same moderate load work category. it was observed that the critical path duration reaches a total of 307.5 days with an increase of 57.5 days. the construction activities affected by the third testing scenario (sc3) for the project is shown in fig. 1(c). it was observed that the critical path duration was increased by 60.5 days to reach a total of 310.5 days. the high duration of critical path compared to previous scenarios is because of the high percentage of rest per hour for heavy load category. the construction activities affected by the fourth scenario is shown in fig. 1(d). it was seen that the critical path duration is 344 days, with an increase of 94 days. this is the worst-case scenario in which the load category is heavy, and weather conditions are pessimistic. the percentage of rest per hours for this scenario is high. thus the labour will take more rest in an hour during his work. as a result, the new duration for the affected activities is high compared to other scenarios. in conclusion, for project a, it was found that the percentage increase in critical path duration for first, second, third and fourth scenarios are 17%, 23%, 24% and 38% respectively. the effect of hot weather was highest for the heavy load work and pessimistic weather conditions, and least for moderate workload category and most likely weather condition. 6.2 project b the project duration for project b is 423 days, started on 3rd april 2010 and completed on 31st may 2011, and the critical path duration of the construction part was 199 days. the construction activities affected after applying the first scenario (sc1) is shown in fig. 2(a). it was observed that the first four activities were affected by the hot weather, resulting in a shift in the start date of the remaining activities. some of the activities were not affected by the weather conditions this is due to a lower percentage of rest per hour inactivity and because some of the activities are in october, november and december. the critical path duration was increased by eight days to reach a total of 207 days. figure 2(b) shows the construction activities affected by the second testing scenario (sc2); which is the moderate workload category and pessimistic weather condition. the affected activities more than that of the first scenario. this is a result of the high percentage of rest per hour for the pessimistic case compared to the same load work category, which is moderate. consequently, the duration of the critical path was affected due to this increase in the duration of activities. it reached a total of 221 days with an escalation of 22 days. for the third scenario (sc3), the output of the construction productivity model is shown in fig. 2(c). it is shown that the hot weather affected the first seven activities, which increased in the duration of the critical path. the critical path duration for the construction part was increased by 18 days to reach a total of 217 days. modeling, investigating, and quantification of the hot weather effects on construction projects in oman 96 figure. 1. project a results after tested under four testing scenarios (sc1, sc2, sc3, and sc4); (a) affected activities and their durations for sc1, (b) affected activities and their durations for sc2, (c) affected activities and their durations for sc3, (d) affected activities and their durations for sc4. figure. 2. project b results after tested under four testing scenarios (sc1, sc2, sc3, and sc4); (a) affected activities and their durations for sc1, (b) affected activities and their durations for sc2, (c) affected activities and their durations for sc3, (d) affected activities and their durations for sc4. hajar al balushi, mubarak al-alawi, and mohammed al shahri 97 affected by the fourth testing scenario (sc4). it is obvious from the graph that more activities are by the figure. 2(d) shows the construction activities affected hot weather conditions. these observations are expected; since this scenario covering the heavy load work category and pessimistic weather conditions; in which the percentage of rest per hour is the highest. the labour will take more time resting per hour during his work, resulting in an extended duration to finish the work of the activity. it was seen that the critical path duration is 242 days, with an increase of 43 days from the duration of the original critical path. in conclusion, the percentage increase in the critical path duration for the four scenarios was calculated. it was found that a 4% increase for the first scenario, 11% for the second scenario, 9% for the third scenario and 22% for the fourth scenario. the effect of hot weather was highest for the heavy load work and pessimistic weather conditions, and least for moderate workload category and most likely weather condition. 6.3 project c the construction duration of project c is 239 days; it was started on 24th june 2018 and accomplished on 28th february 2019. as followed in previous projects, the construction part was studied for this research which includes; sub-structure works and superstructure work. the duration of the critical path for the construction part is 150 days. the construction activities affected after applying the first scenario is shown in fig. 3(a). the graph shows that the hot weather affected three activities. for some of the activities the weather conditions had no impact, this is due to the low percentage of rest per hour for the moderate and most likely scenarios and because some of the activities are in october, november, december and january, which are not considered in this analysis. the critical path duration was calculated and found to be 154 days; it was increased by four days compared to the original critical path duration. figure. 3 (b) shows the results for the second scenario; which is the moderate workload category and pessimistic weather condition. the hot weather conditions affected more activities compared to the first scenario. this is because of the high percentage of rest per hour for the pessimistic weather condition. subsequently, the critical path duration was increased by five days. it reached a total of 155 days. for the third scenario, the output of the construction productivity model is presented in fig. 3(c). it shows that the critical path duration for the construction part was increased by eight days to reach a total of 158 days. figure. 3. project c results after tested under four testing scenarios (sc1, sc2, sc3, and sc4); (a) affected activities and their durations for sc1, (b) affected activities and their durations for sc2, (c) affected activities and their durations for sc3, (d) affected activities and their durations for sc4. modeling, investigating, and quantification of the hot weather effects on construction projects in oman 98 the output of the construction productivity model for the fourth scenario is shown in fig.3(d). more activities were impacted by the hot weather conditions. the graph shows that the first eight activities increased in their working duration, resulting in an extension of the critical path duration. this is predicted because this scenario is the worst scenario; where the percentage of rest per hour is high. it was seen that the new critical path duration for the construction part is 167 days with a rise of 17 days from the duration of the original critical path. in conclusion, the percentage increase in critical path duration was calculated for the four scenarios. for the first and second scenarios, the increase was found to be equal to 3%, 5% for the third scenario and an increase of 11% was calculated for the fourth scenario. it was observed that the effect of extreme weather conditions was highest for the heavy load work and pessimistic weather conditions, and least for moderate workload category and most likely weather condition conclusion in this study, a construction productivity model was developed using the daily temperature and relative humidity of muscat, oman from the years 2001 to 2018. the daily temperature and relative humidity were used to calculate the percentage of rest per hour rule presented by niosh. the generated data were validated statistically against the original data. the rest per hour model was able to satisfactory replicate the real data with some discrepancies. this result may be driven from the fact that the collected data had a simplification due to lack of data and only two years relative humidity records were used in the study. this is a limitation to the model and a future extension of this model will incorporate a large range of relative humidity and temperature records. the construction productivity model was used to investigate and quantify the weather effects on project duration. it was applied to three completed construction projects. for each project, four weather testing scenarios were applied to investigate the impact of hot weather under different workloads on the project duration. the critical path duration was calculated for each scenario in each project and compared with the original critical path durations. the results showed an increase in the critical path duration for all three projects in all scenarios. it was found that 17%, 23%, 24% and 38% increase in critical path duration for project a, for sc-1, sc-2, sc-3 and sc-4, respectively. for project b the percentages were 4%, 11%, 9% and 22% for each scenario respectively. in addition, 3%, 3%, 5% and 11% increase in critical path duration for project c were found for the four presented scenarios. conflict of interest authors declare no conflicts of interest. funding this study did not receive any funding from the public, private or not-for-profit sectors. references abourizk, s. m., & wales, r. j. (1997). combined discrete-event/continuous simulation for project planning. journal of construction engineering and management, 123(1), 11-20. ahuja, h. n., & nandakumar, v. (1985). simulation model to forecast project completion time. journal of construction engineering and management, 111(4), 325-342. al shebani, m., & wedawattab, g. (2014). making the construction industry resilient to extreme weather: lessons from construction in hot weather conditions. procedia economics and finance, 635-642. apipattanavis, s., sabol, k., molenaar, k., rajagopalan, b., xi, y., blakard, b., & patil, s. (2010). integrated framework for quantifying and predicting weather-related highway construction delays. journal of construction engineering and management, 136(11), 11601168. arditi, d. (1985). construction productivity improvement. journal of construction engineering and management, 111(1), 1-14. ballesteros-péreza, p., rojas-céspedesb, y. a., hughesa, w., kabiria, s., pellicer, e., moramelià, d., & campo-hitschfeld, m. l. (2017). weather-wise: a weather-aware planning tool for improving construction productivity and dealing with claims. automation in construction, 81-95. chan, e. h., & au, m. c. (2008). relationship between organizational sizes and constractor's risk pricing behaviors for weather risk under different project values and durations. journal of construction engineering and management, 134(9), 673-680. dytczak, m., ginda, g., szklennik, n., & wojtkiewicz, t. (2013). weather influence-aware robust construction project structure. 11th international conference on modern building materials, structures and techniques. vilnius. el-rayes, k., & moselhi, o. (2001). impact of rainfall on the production of highway construction. journal of construction engineering and management, 125-131. grimm, c., & wagner, n. (1974). weather effects on mason productivity. journal of the construction division, 19-35. guo, s.-j., chen, j.h., & chiu, c.h. (2017). fuzzy duration forecast model for wind turbine construction project subject to the impact of wind hajar al balushi, mubarak al-alawi, and mohammed al shahri 99 uncertainity. automation in construction, 81, 401410. hanna, a. s. (2004). the effect of temperature on productivity. neca, national electrical contractors association. ibbs, w., & sun, x. (2017). weather's effect on construction labor productivity. journal of legal affairs and dispute resolution in engineering and construction, 04517002-2 04517002-7. kavanagh, d. p. (1985). siren: a repetitive construction simulation model. journal of construction engineering and management, 111(3), 308-323. kenner, s., johnson, r. l., miller, j. r., salmen, j. a., & matt, s. a. (1998). development of working day weather charts for transportation construction in south dakota. south dakota department of transportation, office of research. kerkhove, l. p., & vanhoucke, m. (2017). optimised scheduling for weather sensitive offshore construction projects. omega, 66(a), 58-78. koehn, e., m. , a., & brown, g. (1985). climatic effects on construction. journal of construction engineering and management, 129-137. lee, j., park, y.-j., choi, c.-h., & han, c.-h. (2017). bim-assisted labor productivity measurement method for structural formwork. automation in construction, 84, 121-132. marzoughi, f., arthanari, t., & askarany, d. (2018). a decision support framework for estimating project duration under the impact of weather. automation in construction, 87, 287-296. ministry of manpower. (2008). regulation of occupational safety and health measures in establishments subject to labor law. muscat, oman: ministry of manpower. moselhi, o., gong, d., & el-rayes, k. (1997). estimating weather impact on the duration of construction activities. canadian journal of civil engineering, 24, 359-366. niosh. (2018, 5 25). the national institute for occupational safety and health (niosh). (centers for disease control and prevention) retrieved 8 1, 2019, from https://www.cdc.gov/niosh/index.htm palmer, c., & creagh, s. (2013, february 25). climate change linked to declines in labour productivity. retrieved from the conversation: http://theconversation.com/climate-change-linkedto-declines-in-labour-productivity-12407 shahin, a, s. m. abourizk, & y. mohamed. (2011). modelling weather-sensitive construction activity using simulation. journal of construction engineering and management, 238-246. thomas, h., & yiakoumis, i. (1987). factor model of construction productivity. journal of construction engineering and management, 623-639. thomas, r. h., maloney, w. f., horner, r. w., smith, g. r., honda, v. k., & sanders, s. r. (1990). modeling construction labor productivity. journal of construction engineering and management, 116(4), 705-726. yildirim, k., koyuncu, c., & koyuncu, j. (2009). does temperature affect labor productivity: crosscountry evidence. applied economics and international development, 9(1), 29-38. 1. introduction 2. background 3. heat stress 4. work-rest hours prediction model 5. model validation 6. application of the model in real construction projects 6.1 project a 6.2 project b 6.3 project c in conclusion, the percentage increase in critical path duration was calculated for the four scenarios. for the first and second scenarios, the increase was found to be equal to 3%, 5% for the third scenario and an increase of 11% was calculated for t... conclusion conflict of interest c:\users\raoh\desktop\paper_3.xps tjer 2012, vol. 9, no. 1, 21-30 ___________________________________________ *corresponding author’s e-mail: kal_yani_79@yahoo.co.in classification of static security status using multi-class support vector machines s kalyani*a and ks swarupb *a department of electrical and electronics engineering, kamaraj college of engineering & technology, tamilnadu, india b department of electrical engineering, indian institute of technology madras, chennai 600036, india received 23 february 2010; accepted 1 december 2010 abstract: this paper presents a multi-class support vector machine (svm) based pattern recognition (pr) approach for static security assessment in power systems. the multi-class svm classifier design is based on the calculation of a numeric index called the static security index. the proposed multi-class svm based pattern recognition approach is tested on ieee 57 bus, 118 bus and 300 bus benchmark systems. the simulation results of the svm classifier are compared to a multilayer perceptron (mlp) network and the method of least squares (mls). the svm classifier was found to give high classification accuracy and a smaller misclassification rate compared to the other classifier techniques. keywords: static security, classifier, multi-class svm, pattern recognition nomenclature skm complex power flow in branch k-m in mega-volt ampere (mva) mva limitkm thermal limit of branch k-m in mva |vkmin| minimum allowable voltage limit of kth bus in p.u. (taken as 0.90 p.u.) |vkmax| maximum allowable voltage limit of kth bus in p.u. (taken as 1.10 p.u.) |vk| bus voltage magnitude of kth bus in p.u. |pgimin| minimum generation limit of ith generator bus in mega watts (mw) |pgimax| maximum generation limit of ith generator bus in mega watts (mw) |pgi| real power generation of ith generator bus in mega watts (mw) k voltage angle at ith bus in radians sgi complex power generation at ith bus in mva sli complex power load at ith bus in mva nl number of branches (includes both transmission lines and transformers) in the system nb number of buses in the power system network ng number of generators in the power system network 22 s kalyani and ks swarup 1. introduction the power system security is an important concern in the planning and operational studies of power systems. the primary aim of an electric power system is to provide an adequate uninterrupted supply of electrical power to customer premises within the set limits of frequency and voltage levels. this task has to be solved in real time and in a safe, reliable and economical manner. security assessment is the analysis performed to determine whether, and to what extent, the system is reasonably safe from serious interference during its operation. occurrence of certain severe disturbances may cause the system to shift go to an undesirable emergency state, if the system security level is not previously well defined. hence, effective control of power systems demands a quick security evaluation of their operating states (arora, surana 1996). power system security enables a system to remain secure without serious consequences to any credible contingencies on pre-selected list. security analysis may be broadly classified as static security assessment (ssa) and transient security assessment (tsa). static security analysis evaluates the post contingency steady state condition of the system neglecting transient behavior and other timedependent variations. transient security analysis evaluates system performance in terms of rotor angle stability, as it progresses after a disturbance (shahidehpour 2003). the traditional method used for static security analysis involves full ac load flow for each contingency scenario. this procedure is highly time-consuming and infeasible for real time applications (pang et al. 1973; pang et al. 1974). a method is, therefore, required to access security using realtime data. this leads to the application of the pattern recognition (pr) approach. in recent years, many artificial intelligence (ai) techniques have been proposed to overcome the pitfalls of the traditional method of security evaluation. ai techniques like the self-organizing feature map (swarup, corthis 2006), and the multilayer feed forward with a back propagation algorithm (saeh, khairuddin 2008) have been applied for the problem of static security assessment. various literatures has also reported the use of an ann -based pattern recognition approach (boudour, hellal 2006; luan et al. 2000) a genetic-based neural network (azah, maniruzzaman 2001), a fuzzy logic combined with a neural network (haghifam, zebarjadi 1996), and a query-based learning approach in neural networks (huang 2001) for the static security evaluation process. the performance of all these existing techniques are highly problem dependent and hence their suitability cannot be underestimated/minimized. moreover, because of the nature of the input features used, these methods are found to be incapable of quickly predicting the future insecure operation. nowadays, pattern recognition (pr) techniques have demonstrated great importance in security evaluation of large electric power systems (sa, munro 1984). in the pattern recognition approach, the main bulk of simulation is done off-line to generate sufficient data for training set (pang et al. 1974). using the training data set, the classification function is designed, from which system security can be judged in a short period of time. this paper presents the application of pattern recognition (pr) approach to static security assessment of large scale power systems in multi-class mode. the classifier in the pr system is designed by support vector machines (svm). svm is a new and promising tool for learning separating functions in a pr system with the capability of handling non-linear separability. the svm classifier is designed for multi-class classification based on the calculation of a term called 'static security index' (ssi), for each specified contingency. in this paper, four class logic is used for the definition of system security viz., secure, critically secure, insecure, highly insecure. an operator needs to know the exact severity level of disturbances for a given system operating condition. on-line security evaluation allows the operator to know the security status and determine the corrective actions. 2. static security assessment (ssa) static security is the ability of the system to reach a state within the specified secure region following a contingency. a set of the most probable contingencies is first specified for security evaluation. this set may include single line outage, a loss of a generator, or a sudden increase in load. the violations of thermal limits of transmission lines and bus voltage limits are the main concerns for static security analysis. in conventional practice, security assessment is obtained by analytically modeling the network and solving the load flow equations repeatedly for all of the prescribed outages, one contingency at a time (lo, peng 1997). this traditional approach is not entirely satisfactory because the huge number of simulations needed to be carried out. a given system operating condition is said to be 'static secure', if the bus voltage magnitudes and real power generation of generator buses are well within their limits, without any occurrence of line overloads. in this paper, the term, statis security index (ssi) indicates the system security level for a given system operating condition and a specified contingency. the ssi is defined by calculating the line overload index (loi), voltage deviation index (vdi) and generation 23 classification of static security status using multi-class support vector machines deviation index (gdi) as given by eqs. (1), (2) and (3) respectively. (4) (5) where skm and mva limitkm represent the mva flow and the mva limit of branch k-m, |vkmin|, |vkmax| and |vk| the minimum voltage limit, maximum voltage limit and bus voltage magnitude of the kth bus respectively. in addition, pgimin, pgimax and pgi represent the minimum generation limit, maximum generation limit and real power generation of the ith generator bus respectively, and nl, nb and ng represent the number of lines, buses and generators respectively. 3. pattern recognition (pr) approach a pattern is a pair comprised of an observation and a meaning. a pattern recognition infers meaning from observation. pattern recognition is defined as 'the act of taking in raw data and taking an action based on the category of data'. it aims to classify the data or patterns based on either priori knowledge or on statistical information extracted from the patterns (pecas et al. 1988). a complete pattern recognition system, as shown in fig. 1, consists of a sensor that gathers observations to be classified; a feature extraction mechanism that computes numeric or symbolic information from observations and a classification scheme that classifies the observations, relying on extracted or selected features. 4. application of pattern recognition to static security assessment the classification of power system state is the primary stage of security assessment in large scale real power system networks. from a pattern recognition perspective, the ssa problem is considered to be a classification problem whereas the pre-contingency system attributes are used to predict the post contingency system security status. pattern recognition can be seen as a classification process which is normally employed to reduce the online computational requirements. in designing a pattern recognition system, a considerable amount of work is done off-line. this work is used for the generation of a set of characteristic operating points necessary to design a classification function called the 'security function'. once the classifier is derived, an actual assessment of any new data sample located on-line can be made by evaluating the security function and can be classified as belonging to any one of the multi-classes described. this makes the pr system (1) (2) (3) 24 s kalyani and ks swarup much faster than any other method of security assessment. the sequence of steps performed in the off-line and on-line stages in applying a pattern recognition approach to static security evaluation is shown in fig. 2. 4.1 pattern generation the success of pattern recognition relies on a good training set. the training set patterns may be obtained on-line from real-time occurrences or can be synthesized from off-line simulations (laveen 1974). in this paper, the train set and test set patterns are generated by off-line simulations. different operating conditions were considered by varying the system load from 50% to 200% of base load. for each operating scenario, single line outages (one at a time) are simulated and a load flow solution is obtained. evaluating the static security index (ssi) as given by eq. (5), each pattern is labeled as belonging to one of four classes secure, critically secure, insecure and highly insecure. all variables describing an operating condition constitute the components of pattern vector represented as x = {x1, x2, xn}. these variables include loads, generation's power flows in lines, voltage magnitude and angle at buses. the primary variables called steady state variables, forming the components of pattern vector are given below: x = {|v |k, k, sgi, sli, skm} where |v|k voltage magnitude at bus i figure 1. block diagram of pattern recognition (pr) approach figure 2. svm based pattern recognition for static security assessment 25 classification of static security status using multi-class support vector machines k voltage angle at bus i sgi complex power generation at generator bus i sli complex power load at load bus i skm mva power flow in branch k-m 4.2 feature selection the number of variables describing the power system state in the pattern vector is significantly large. this phase involves selecting from a large set of variables that will give more useful information to build the classification function. these variables are termed features and the process of obtaining them is called feature selection (siri et al. 1992). the features form the components of a vector called feature vector represented as z = {z1, z2, …., zm}. the importance of feature selection is to reduce the unclear of dimensionality and the search space for learning algorithm. input features may be selected by engineering judgment. but such selections will be subjective with the possibility of important variables that are rejected. a common method of feature selection is sequential feature selection, consisting of two components an objective function called criterion and a sequential search algorithm. in this paper, a 'sequential forward selection' (sfs) method for feature selection process is used. the criterion which this method seeks to minimize over all feasible feature subsets is the misclassification rate for classification models. the sfs method starts with an empty candidate set and adds feature variables sequentially until addition of further variables does not decrease the criterion (minimization of misclassification). 4.3 multi-class svm based classifier design after extracting the desired features by using sequential forward feature selection method, the final step is to design a classification function. the classifier represents the boundary between the separating classes. the design of the classifier is based on the design (training) set. there are many training algorithms like least squares, linear programming, etc. although these existing algorithms, are less time consuming to use, they were found to have poor classification accuracy. the main requirements for a security function are high classification accuracy and less misclassification rate. hence, a need arises to devise a more suitable learning algorithm. this led to the idea of applying a recently introduced machine learning tool called the support vector machine (svm) in the classification phase of the pr system. the following section gives a brief introduction to the multi-class svm and the procedure to use for classification. overview of svm sims is a learning systems designed to automatically trade off accuracy and complexity by minimizing an upper bound on generalization error. svm is more suitable for classification problems, particularly those involving more than two classes. the svm paradigm, originally designed for the binary classification problem, has a nice geometrical interpretation of discriminating one class from another by a hyper plane with the maximum margin (abhisek 2007). svm performs the job of classification by implementing a nonlinear mapping of input vectors to a high dimensional feature space, where a linear decision surface is constructed. this paper focuses on a static security evaluation problem, which in also this paper, is a multi-class pr problem. due to various complexities, a direct solution of a multi-class problem using a single svm formulation is usually avoided. the better approach is to use a combination of several binary svm classifiers to solve a given multi-class problem. popular methods are: one-versus-all method using a winnertakes-all strategy; one-versus-one method implemented by max-wins voting (kai-bo, sathiya 2005). in this paper, the latter technique was used,presisely, the one-versus-one method for designing the multi-class svm classifier. each of the two methods is briefly discussed as follows: a) one-versus-all (ova) this is conceptually the simplest multi-class svm classification method. it constructs k svm models, class 1 (positive) versus all other classes (negative), class 2 versus all other classes, ... , class k versus all other classes, where k is number of classes (chih, chih-jen 2003), a comparison of methods for multiclass support vector machines. taiwan). given a set of training data samples of length £ (x1 , y1) , (x2, y2 ) , . . . , (xe, ye) ; xi rm , where each training sample xi has m features describing a particular signature and belongs to one of the k classes, ie., y {y1, . . . , yk }. the basic concept behind svm is to search for a balance between the regularization term ½ (wi)t wi and the training errors. classification of new instances in the one-versus-all method is done by a winner-takes-all strategy, in which the classifier with the highest output function assigns the class. for a multi-class problem defining k classes, one needs to solve k quadratic programming (qp) optimization problems of size £. hence, this approach is computationally expensive and not commonly preferred. b) one-versus-one (ovo) this method constructs k (k 1)/2 binary classifiers, where each one is trained on data from two classes. in other words, for every pair of class, a binary svm problem is solved. the classification in one-versus-one method is done by a max-wins voting strategy 26 s kalyani and ks swarup (mwv). after each of the k (k 1)/2 binary classifiers make its vote, the decision function assigns an instance x to a class having largest number of votes (kai-bo, sathiya 2005; chih-wei, chin 2006). in this case, a tie occurs with two classes having identical votes, the one with smallest index is selected. one of the benefits of this approach is that for every pair of classes, a smaller optimization problem is dealt with, thereby resulting in saving of total computation time. hence, if in an average, each class has £/k data points, only the k (k 1)/2 quadratic programming (qp) problems of size less than £ have to be solved. this is found to be quite less compared to the ova approach, where k number of qp problems, each of size £ needs to solved. svm training algorithm for classification task the procedure or steps involved in applying svm for the problem of multi-class classification is as follows: 1. data scaling/preprocessing the input features in train set and test set needs to be scaled properly before applying svm. this is important as the kernel values depend on the inner products of the feature vector. scaling prevents the domination of any feature over the other because of higher numeric values involved and also avoids numerical difficulties during calculation. it is recommended to linearly scale each attribute to the range of [0, 1]. 2. design of svm model choice of kernel the radial basis function (rbf) kernel is the first choice because of its widely-known accuracy. further, it is capable of handling non-linear relations existing between the class labels and input attributes. the second reason is that rbf kernel, unlike other kernels, has only one kernel parameter, thereby reducing the complexity of the model. adjusting the kernel parameters there are two parameters associated with rbf kernels the penalty parameter, c the and rbf kernel parameter, . the goal is to identify optimal (c, ) for the classifier to accurately predict the unknown data (test data). this can be achieved by a technique called 'cross-validation'. in a v-fold cross validation, the whole training set is divided into v subsets of equal size. sequentially one subset is tested using the classifier trained on the remaining (v-1) subsets. thus, each instance of the train set is predicted once and the cross-validation accuracy is the percentage of data samples that are correctly classified (min et al. 2005). in this study, a grid search is used on c and using 5-fold cross validation. all pairs of (c, ) were tried and the one with highest crossvalidation accuracy was selected. it was also realized that using exponentially growing sequences of c and is a practical method to identify optimal parameters. these sequence was used c = {2-5, 2-3, . . . , 215} and = {2-15, 2-13 , . . . , 25 } in the svm experiment. 3. training and testing the svm model after designing the svm model with the chosen kernel and optimal parameters, it is trained with the scaled input-output train set data samples. once the performance of the svm classifier is found satisfactory in the training phase, the model is validated with test data samples to access its overall performance. 5. performance evaluation of the classfier the performance of the svm classifier is validated by calculating the following performance measures for train set, test set and combined set (combination of train and test sets). (1) mean squared error (mse) n no. of data set samples dok desired output obtained from off-line simulation aok actual output obtained from svm classifier model (2) classification accuracy (ca) (7) (3) misclassification (mc) rate (8) in power system security evaluation, it is important to ensure that the misclassification rate is kept at minimal. in particular, the chances of a highly insecure state being wrongly predicted as secure, needs to be reduced, as it may be lead to failure of control actions and hence a severe 'blackout'. thus, the classification system for security evaluation must be efficiently designed to have high classification accuracy and a low misclassification rate. (6) 27 classification of static security status using multi-class support vector machines 6. simulation results and discussion the proposed multi-class svm based pattern recognition approach to static security evaluation is implemented on 57 bus, 118 bus and 300 bus ieee standard systems. the multi-class svm classifier is designed and tested using libsvm software package (cin-chung, chih-jen 2004). normally in power system networks, transmission lines are permitted to carry power to a maximum limit of 125% to 130% of scheduled value so as to meet the small increase in demand. this will not pose serious problems as the transmission line is normally designed to meet this requirement. however, the system losses will increase due to large the amount of heat dissipation involved. based on this, the mva limit of the lines and transformers is taken as 130% of the base case mva flow in the simulation work. the security limit of voltage magnitude at load buses is imposed in the range of 0.90 pu to 1.10 pu. 6.1 data generation different scenarios have been considered by varying the system real power generation and load from 50% to 200% of base case values. the variation in real power generation is limited to its minimum and maximum values. contingencies of single-line outages are simulated for each operating condition. for a given operating condition and specified contingency, a load flow solution by fast decoupled method is obtained and the static security index as given by eq. (5) is calculated. in the calculation of ssi, the weighting factors for loi, vdi and gdi are assumed as w1 = 3, w2 = 2 and w3 = 1.5 respectively. these weighting factors are fixed based on the order of priority in requirement of system security. ssi is a percentage measure of system security level, taking value in the range between 0 and 100. based on the computed value of ssi, data patterns in each operating condition are categorized in one of the following four classes. 6.2 design of multi-class svm model the steady state variables obtained from load flow solution are recorded as pattern variables. the variables included in the pattern vector are bus voltage magnitude, bus voltage angle, complex power generation at generator buses, complex power load at load buses and mva flow in all branches. the large size pattern vector is reduced by sequential forward approach of the feature selection. this selects the pattern variables having higher discriminating power and hence determines the feature vector, which is an optimal subset of pattern vector. the data samples in the feature vector are randomly split into two parts train set and test set. the multi-class svm classifier model is designed using the training data samples. the performance of svm classifier depends on the type of kernel function and svm parameters. the radial basis function (rbf) was chosen as kernel function in our svm experiment. the parameters of svm model to be tuned are the penalty parameter (c) and the rbf kernel parameter ( ). the optimal values of these parameters are obtained by a '5-fold cross validation' procedure for a cross-validation accuracy of 100%. the svm model designed is validated by rating its performance, when subjected to randomly generated test set samples, whose class labels are unknown. 6.3 test case results table 1 shows the results of data generation and feature selection of the pr system. the total number of operating scenarios simulated and number of cases belonging to each class is shown for the test cases studied. the number of feature variables extracted from the pattern vector, as seen from table 1 is significantly less, as is evident from the figure of dimensionality reduction (defined as the ratio of number of components in the feature vector to that in the pattern vector). the svm classifier is trained using the normalized values of selected features. the data samples in the feature vector are normalized to a range of [0, 1] using the min-max normalization method. it is the one of the widely used techniques for the data scaling process. it performs linear transformation on original data by scaling the data in each attribute to fit in a specific range. table 2 shows the results of classification in training, testing and combined phases for ieee 57 bus, 118 bus and 300 bus test systems. the results of svm classifier is compared with multilayer perceptron (mlp) and method of least squares (mls) classifiers in terms of performance measures. the mlp network, designed and trained using the neural network toolbox in matlab 7.6, consists of a hidden layer with the 30 neurons of 'tansig' transfer function. the network is trained with the levenberg marquardt algorithm (learning rate = 0.05; performance goal = 0.001; epochs = 600). it can be easily seen from table 2 that the svm classifier gives a fairly high classification accuracy and a less misclassification rate, in particular for class k=4, compared to mlp and mls classifiers. furthermore, the time taken by the svm algorithm is much compared to mlp network, even for a large size system. figure 3(a) and 3(b) shows the cross validation plot of the 28 s kalyani and ks swarup svm classifier trained with the rbf kernel using 5fold cross validation technique for the ieee 57 bus and ieee 300 bus systems respectively. the optimal values of svm parameters obtained for a cross validation accuracy of 80%, as seen from fig. 3(a) for ieee 57 bus, are penalty parameter, c = 211 = table 1. results of pattern generation and feature selection stages of pr system table 2. classification results on train, test and combined sets for multi-class ssa * as there are no samples belonging to the class of static highly insecure (class k=4) case studies 29 classification of static security status using multi-class support vector machines 2048 and rbf kernel parameter, = 2-2 = 0.25. the svm parameters selected for the ieee 300 bus for a cross validation accuracy of 80%, seen in fig. 3(b), are penalty parameter, c = 22 = 4 and rbf kernel parameter, = 20 = 1.00. 7. conclusions in this paper, the svm based pattern recognition (svm-pr) approach for static security evaluation in multi-class mode is presented. the proposed multiclass svm-pr model was tested on ieee standard test systems. simulation results showed that high accuracy security functions are realizable with an svm classifier. the svm classifier also proved to give a lower misclassification rate compared to mlp or any equivalent methods. a good classification system should to reduce the chances of highly insecure states being erroneously predicted to almost zero (misclassification rate corresponding to class k=4). this is achieved by the svm classifier, thereby reducing the posibility of failure of the control actions. future work will focus on the improvement of performance of the svm classifier by adopting different methods for feature selection and svm parameter tuning. references abhisek u (2007), intelligent systems and signal processing in power engineering. pringerverlag, switzerland. arora cm, surana sl (1996), transient security evaluation and preventive control of power systems using pr techniques. ie (india) 76:199-203. azah ms, maniruzzaman ha (2001), static security assessment of a power system using geneticbased neural networks. electric power components and systems 29:1111-1121. boudour m, hellal a (2006), combined use of unsupervised and supervised learning for large scale power system static security assessment. int. j. of power & energy systems 26(2):157-163. chih-chung c, chih-jen l (2001), libsvm: a library for support vector machines. software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm. chih-wei h, chih-jen l (2005), a comparison of methods for multi-class support vector machines. taiwan (cjlin@csie.ntu.edu.tw). haghifam mr, zebarjadi v (1996), fuzzy logic and neural network approach to static security assessment for electric power systems. proceedings of 4th european congress on intelligent techniques and soft computing 3:2009-2013. http://www.ee.washington.edu/research/pstca (power system test case archive). http://www.pserc.cornell.edu/matpower (matpower 3.2). huang sj (2001), static security assessment of a power system using query-based learning approaches with genetic enhancement. iee proceedings-generation, transmission & distribution 148(4):319-325. kai-bo d, sathiya ks (2005), which is the best multi-class svm method? an empirical study. springer-verlag, berlin heidelberg 278-285. laveen k (1974), patterns in pattern recognition. -5 -1 3 8 11 15 -5 -2 2 5 9 14 penalty paarameter, c penalty parameter, c (a) ieee 57 bus systesm (b) ieee 300 bus system 4 0 -4 -6 -11 -15 figure 3. selection of svm parameters for multi-class ssa problem 5 1 -2 -7 -11 -15 30 s kalyani and ks swarup ieee transactions on information theory it20(6):697-722. lo kl, peng lj (1997), design of artificial neural networks for on-line static security assessment problems. proc. of the 4th int. conf. on apscom-97, hong kong 288-293. luan wp, lo kl, yu yx (2000), ann based pattern recognition technique for power system security assessment. ieee int. conf. on electric utility deregulation, restructuring and power technologies, london 197-202. min jhy, chan l (2005), bankruptcy prediction using support vector machine with optimal choice of kernel function parameters. expert systems with applications 28(5):603-614. pang ck, kovio aj, el-abiad, ah (1973), application of pattern recognition to steady state security evaluation in a power system. ieee transactions on systems, mans and cybernetics smc-3(6):622-631. pang ck, prabhakara fs, el-abiad ah, koivo aj (1974), security evaluation in power systems using pattern recognition. ieee transactions on power apparatus & systems pas-93:969-976. pecas lja, machiel bfp, marques dsjp (1988), on-line transient stability assessment and enhancement by pattern recognition techniques. electrical machines and power systems 25:293310. sa dcjmg, munro n (1984), pattern recognition in power system security. int. j. of electrical power & energy systems 6(1):31-36. saeh is, khairuddin a (2008), static security assessment using artificial neural network. ieee 2nd int. power & energy conference (pecon'08) 11721178. shahidehpour sm (2003), communication and control in electric power systems. wiley interscience, john wiley & sons, third edition. siri w, sharkawi mael (1992), feature selection for static security assessment using neural networks. ieee int. symposium on circuits & systems, san diego, california 10-13:1693-1696. swarup ks, corthis bp (2006), power system static security assessment using self-organizing neural network. j. of indian institute of science 86(4):327-342. chen060319.qxd the journal of engineering research vol. 5, no.1 (2008) 20-29 1. introduction in 1978, the united states environmental protection agency (usepa) prepared a list of 129 organic and inorganic pollutants found in wastewater that constitute serious health hazards. this list, known as the priority pollutants list, includes the following thirteen metals: antimony, arsenic, beryllium, cadmium, chromium, copper, lead, mercury, nickel, selenium, silver, thallium, and zinc. unlike organic compounds, metals are non_______________________________________ *corresponding author’s e-mail: ashour@squ.edu.om biodegradable and, therefore, must be removed from wastewater. zinc is present in the air, soil, water, and in almost all food. zinc is naturally released into the environment, although industrial activities are mostly responsible for zinc pollution. elevated levels of zinc may come from a variety of sources like mining and foundry activities, zinc, lead, and cadmium refining, steel production, carbon combustion, and solid waste incineration. zinc is commonly used to coat iron and other metals for the prevention of oxidation. various zinc salts are industrially used in wood preservatives, catalysts, photographic paper, and accelerators for rubber vulcanization, ceramics, textiles, biosorption of zinc on immobilized green algae: equilibrium and dynamics studies d. sheikha1, i. ashour*2 and f.a. abu al-rub3 1department of chemical and petroleum, uae university, p.o. box 17 555 al-ain, uae 2department of petroleum and chemical engineering, sultan qaboos university, p.o. box 33, al-khoud, muscat 123, sultanate of oman 3chemical engineering department, just, p.o. box 30, irbid, jordan received 29 march 2006; accepted 24 january 2007 abstract: the efficacy of using blank alginate beads and immobilized dead algal cells for the removal of zinc ions from aqueous solutions was investigated. it was found that the sorption capacities were significantly affected by solution ph; with higher ph favoring higher zinc ion uptake. dynamics and isotherm experiments were carried out at the optimal ph 5.0. zinc uptake on either sorbent was found to be rapid where approximately 90% of the maximum zinc uptake occurred within the first 30 min in both cases of blank alginate and immobilized algal cells. the equilibrium data for the biosorption of zinc ions onto both sorbents were fitted to the langmuir, freundlich, and dubinin-radushkevich (d-r) isotherm equations. the presence of copper and nickel in aqueous solutions was found to suppress the sorption process. the results of the dynamics studies revealed that the biosorption of zinc on immobilized dead algal cells followed pseudo-second order kinetics with little intraparticle diffusion mechanism contribution. keywords: biosorption, heavy metals, immobilized green algae, pseudo second order model, alginate, equilibrium isotherms á«cô◊gh ¿gõj’g á°sgqo :agô° ÿg ödéë£dg ≈∏y ∂fõdg ¢ué°üeoeg áî«°t .o1qƒ°téy ,2*üôdg ƒhg , áá°°uuóóÿÿgg.√é«ÿg ‘ áñfgòdg ú°uqéÿg äéfƒjcg øe ¢ü∏îàdg ‘ agô° ÿg ödéë£dg éjón é¡«∏y âñãe äéñ«ñm h ioôøæe äéñ«ñm ωg~îà°seg á«dé©a øe ≥≤ëà∏d á°sgq~dg √òg ±~¡j : ~bh .»æ«lhq~«¡dg ºbôdg ioéjõh ogoõj é¡à«dé©a h √õàªÿg ioéÿg á©°s ¿g å«m ,»æ«lhq~«¡dg ºbôdéh gòãc ôkéàj ¢ué°üeo’g iq~b ¿g âæ«h ¢ué°üeo’g üqééàd ájèıg èféàædg á°ùaéæe äéfƒjg oƒlh ¿cg üqééàdg ∂∏j âæ«h éªc .∂«d~fôa h òª‚’ »lpƒ‰ πãe á°uéÿg á«°véjôdg êpéªædg ωg~îà°séh 5 »æ«lhq~«¡dg ºbôdg ~æy ¿gõj’g äé°sgqo èféàf π«∏– ” ogƒÿg ≈∏y 5 »æ«lhq~«¡dg ºbôdg ~æy ¢ué°üeo’g á«∏ªy ¿g á«cô◊g äé°sgq~dg èféàf âæ«h éªc .é¡àféøc øe π∏≤jh ¢ué°üeo’g á«∏ªy §«ñãj ¤g …odƒj πµ«ædg h ¢séëædg πãe ¬«ñ°ûh ¬ø°u h øµá ú°uqéÿg ¢ué°üeog ¿g èféàædg ∂∏j âæ«h éªc ,¤h’g ¬yé°s ∞°üædg ∫ón â∏°üm ∫ƒ∏ùg øe iõàªÿg ú°uqéÿg ᫪c øe %90 ¿g å«m á©jô°s â°sqo »àdg .á«féãdg ¬lq~dg øe »cô◊g §ªædg áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.áàhék iqgôm äélqo ~æy ¿gõj’g ,á«féãdg álq~dg øe á«cô◊g á°sgq~dg §‰ ,äéæ«÷’g ,áàñãÿg agô° ÿg ödéë£dg ,á∏«≤ãdg ¿oé©ÿg ,…ƒ«◊g ¢ué°üeo’g : 21 the journal of engineering research vol. 5, no.1 (2008) 20-29 fertilizers, pigments, and batteries. water reservoirs are contaminated by the run-off from these industries. other sources of metallic zinc traces in drinking water are water treatment processes and pick-up of metallic ions during storage/distribution. these toxic metals cause accumulative poisoning, cancer, brain damage, etc., when they are found above the tolerance levels. according to few surveys from the public health services of different countries, significant number of people has been exposed to the hazards of excess metals in the municipal water supplies (agrawal and pandey, 2004). traditional treatment methods for removing zinc from wastewater include chemical solvent extraction (kongolo, et al. 2003; preston and du preez, 2000), chemical precipitation (veeken, et al. 2003), membrane filtration (srisuwan and thongchai, 2002), ion exchange (kurama and catalsarik, 2000), and adsorption (ramos, et al. 2002; mohan and singh, 2002; peric, et al. 2000; galiatsatou, et al. 2002). for dilute metal concentration, ion exchange, reverse osmosis and adsorption can be applied. however ion exchange and reverse osmosis have high operating cost, which makes adsorption a better alternative for heavy metals removal (abu al-rub, et al. 2004 ). one of the promising techniques for the removal of metals is the use of living or nonliving organisms and their derivatives. in general, biosorption can be defined as the passive sequestering of metal ions by metabolically inactive biomass (volesky, 2003). biosorbents, which are sorbents of biological origin, have proved to be good sorbents for many different pollutants. research in the area of biosorption suggests it to be an ideal alternative for decontamination of metal containing effluents (gupta, et al. 2000). indeed, a wide variety of microorganisms (both living and nonviable) have been found to be capable of sequestering trace levels of metal ions from dilute aqueous solutions. the nonviable forms have been proposed as potential sorbents, since these are essentially dead materials, which require no nutrition to maintain the biomass. problems associated with metal toxicity in living biomass and the need to provide suitable growth condition also do not arise. early studies have shown that nonliving biomass may be even more effective than living cells in sequestering metallic elements (sheng, et al. 2004). three major sources of biomass can be readily identified: (1) waste biomass (aksu and yener, 2001; aksu and gonen 2003; aksu and akpinar, 2001), (2) microorganisms (rao and viraraghavan, 2002; denzili, et al. 2004; feng and aldrich, 2004; arica, et al. 2004; abu al-rub, et al. 2004; pagnanelli, et al. 2001; ibanez and umetsu, 2004). (3) agricultural wastes bark (mckay, et al. 1999), sugar industry mud (magdy and daifullah, 1998), peat (ho, et al. 2002), tree fern (ho et al. 2005), olive pomace ( pagnanelli, et al. 2003), palm tree leaves (abu al-rub, 2006). cell immobilization is an attractive technique to fix and retain biomass on suitable natural or synthetic materials support for a range of physical and biochemical unit operations (abu al. rub, et al. 2004). the main advantages of this technique include: improvement of biomass performance, biosorption capacity and facilitate separation of biomass from pollutant bearing solution. the objective of this study is to investigate the efficacy of using blank alginate beads and immobilized algal biomass for the removal of zinc ions from aqueous solutions. the effects of different parameters, such as solution ph, shaking time, and zinc ions concentrations on the sorption capacity were investigated. comparison between sorption on blank alginate and immobilized dead algal biomass will also be investigated. equilibrium modeling will be carried out using the langmuir, freundlich, and dubinin-radushkevich isotherm equations. 2. experimental 2.1 chemicals the stock solution of zinc used in this study was prepared using analytical reagent grade of znso4.7h2o (bdh, uk) in deionized water. 8.8 g of znso4.7h2o was weighed and added into 1000 ml volumetric cylinder and completed with deinionized water, then stored. 2.2 preparation of biosorbent immobilized algal cells were prepared by entrapping powdered chlorella pyenoidosa (watershed, usa) in an alginate matrix produced by ionic polymerization in calcium chloride solution, according to the following procedures (abu al rub, et al. 2004): the powdered algal cells were suspended in a 2% sodium alginate (bdh, uk) solution kept at a temperature of 60oc. the mixture was then dropped into a 2% calcium chloride (bdh, uk) solution using a peristaltic pump. the drops of na-alginate solution gelled into 3.5±0.1 mm diameter beads upon contact with calcium chloride solution. the beads were washed well and then rinsed in deionized water and stored at 4oc. for blank alginate beads, similar procedures were followed, but without algae. 2.3 determination of functional groups the functional acidic groups on the prepared algal cells were determined using boehm's titration method (abdulkarim and abu al-rub, 2004; strelko, et al. 2002): 1 g of the powdered algal cells was dispersed in 50 ml deionized water. the suspension was mixed with 0.1n solutions of sodium bicarbonate, sodium carbonate, and sodium hydroxide, and then shaken for 48 h at room temperature. after this time, the sample was left for 6 h so that particles can settle. the sample was then filtered and 10 ml of filtrate were titrated with 0.1 n volumetric hcl standards using a methyl red as the indicator. according to boehm's titration method, sodium bicarbonate can neutralize carboxyl groups, sodium carbonate can neutralize carboxyl, lactones and lactols groups, and sodium hydroxide can neutralize carboxyl, lactones, lactols and phenols 22 the journal of engineering research vol. 5, no.1 (2008) 20-29 groups. table 1 lists the different functional groups available on algal cells. 3. procedure 3.1 equilibrium adsorption isotherm all experiments were conducted by adding a specific amount of beads into 100 ml reagent bottles containing 50 ml of the zinc solution. different initial concentrations were used: for zinc (20-350 ppm). the mixtures in these bottles were agitated for 1 hour in a shaker at 25oc. the zinc solution, then, was separated from sorbent and the concentration of zinc ions was determined using a varian atomic absorption spectrophotometer (spectra aa, 880). the uptake, which represents the amount sorbate sorbed per units mass of sorbent is calculated using the following equation: (1) where qe is the uptake (mg/g) at equilibrium, co the initial concentration sorbate (mg/ml), ce the concentration at equilibrium (mg/ml), v the initial volume of solutions, and w is the mass of sorbent (g). 3.2 adsorption dynamics the dynamics studies were carried out by conducting batch biosorption experiments with 100 ppm of zinc at ph 5.0. samples were taken at different time periods and analyzed for their zinc concentrations. 3.3 competitive biosorption the competitive biosorption of zinc with nickel and copper were investigated. the studies involved experiments with different metals concentrations. all the experiments were conducted at 25oc using the same procedures used in the single adsorption experiments, and were carried out in triplicate. 4. results and discussion 4.1 effect of ph on zinc biosorption heavy metals biosorption is highly ph dependent (abu al rub, et al. 2004). the effect of ph on zinc sorption capacity of the immobilized dead algal cells was studied at 100 ppm zinc initial concentration and 0.15 g of algal cell. figures 1 and 2 demonstrate the variation of the uptake of zinc with equilibrium ph. the uptake of zinc increases from 1.83 to 7.26 mg/g over a narrow ph range (3.0-5.0). no significant increase in zinc uptake at ph values above 5.0 was observed. similar trend was obtained in previous studies (martins, et al. 2004; mohapatra and gupta, 2005; ramos, et al. 2002). solution ph affects both cell surface metal binding sites and metal chemistry in water (abu al-rub, et al. 2006). at low ph values, ion exchange reaction involving metals are in competition with the high concentrations of h+ in the solution. with increasing ph, more ligands, such as amino and carboxyl groups, on sorbent are exposed and thus negative charges result and attraction between these negative charges and the metals increases the biosorption capacity on the cell surface. another reason for increasing zinc removal is that the isoelectric point for algal biomass is at ph 3.0 (abu al-rub et al. 2006). at ph values above zero-point charge, the algal cells would have a negative net charge functional group meq h+/g algae carboxyl 0.02 carboxyls, lactones and lactols 0.01 carboxyls, lactones, lactols and phenols 0.035 table 1. the functional groups on the algal cells w v)cc( q eoe − = ph 1 2 3 4 5 6 7 e 0 2 4 6 8 im m obilized dead algal cell figure 1. effect of ph on zinc uptake (initial zinc ions concentration = 100 ppm mass of algal cell = 1.15 g) 0 2 4 6 8 1 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 1 2 3 4 5 6 q e (m g /g ) c e (m g/l) p h figure 2. effect of ph on zinc removal (mass of algal cells = 0.30 g) q e (m g/ g) ph q e (m g/ g) 23 the journal of engineering research vol. 5, no.1 (2008) 20-29 and the ionic state of ligands on algal cell surface is such as to endorse reaction with zinc ions. thus, electrostatic attraction between zinc and the negatively charged algal cells surface occurs which enhances the biosorption above ph 3.0 (abu al-rub, et al. 2006). 4.2 dynamics of zinc biosorption the variation in uptake of zinc with shaking time was studied using solutions of zinc with initial concentration of 100 ppm at ph 5.0 and 0.3 g of algal cell. the shaking time was varied from 3 to 180 min. figure 3 reveals the variation of the uptake of zinc versus time for immobilized inactive algae and blank alginate. the maximum zinc uptake with either sorbent was reached after 60 min. figure 3 also indicates that the sorption of zinc on these sorbents involves two stages: in the first stage, sorption is rapid where approximately 90% of the maximum zinc uptake occurs within the first 30 min in both cases of blank alginate and immobilized algal cells. this rapid sorption indicates that passive surface sorption occurs on the algal cells or beads surface. the second stage is slow and may involve other adsorption mechanisms such as intraparticle diffusion. the advantage of such rapid sorption in practical applications is that smaller reactor volumes can be used (abu al rub, et al. 2004). the pseudo-second order kinetic, suggested by (ho and mckay, 1999 and ho, 2004), is based on the sorption capacity of the sorbent and is given by: (2) an integrated pseudo-second order rate law can be obtained from eq. (2) for the boundary conditions t = 0 to t = t and qt = 0 to qt = qt, and is given by: (3) eq. (3) can be rearranged to obtain a linear form: (4) where qe is the amount of sorbate sorbed at equilibrium (mg g-1);); qt is the amount of sorbate sorbed at time t (mg g-1); and k is the equilibrium rate constant of pseudo-second order sorption (g mg-1 min-1). the linear form see eq. (4) is obtained by plotting t/qt versus t, with the slope of 1/qe and intercept of 1/kqe2, as shown in fig. 4. the values of sorption rate constant and the equilibrium uptake with the correlation coefficient are listed in table 2. figure 4 and the results listed in table 2 indicate that the kinetics for the sorption of zinc on blank alginate beads and immobilized inactive algae follow pseudo-second order kinetics. the weber and morris equation, 1963, given by eq. (5), can be used to test for the contribution of intraparticle diffusion. tim e (m in) 0 20 40 60 80 100 120 140 160 180 200 t 0 2 4 6 8 im m obilized dead algal cell b lank alginate beads figure 3. effect of shaking time (initial zinc ions concentration = 100 ppm, ph = 5.0, mass of algal cell = 0.3 g) 2 te t )qq(k dt dq −= kt q 1 )qq( 1 ete += − kq 1 q t q t 2 eet += pseudo-second order sorbents k1,ads (l/min) k2,ads (g/mg.min) qe (mg/g) r 2 immobilized dead algal cells 9*10-5 0.012 11.02 0.9985 blank alginate beads 0.0031 0.011 10.24 0.9981 table 2. kinetic parameters for the biosorption of zinc ions on immobilized dead algal cells and blank alginate beads time (min) 0 5 10 15 20 25 30 35 t/q t (m in g m g 1 ) 1 2 3 4 5 6 immobilized dead algal cell blank alginate beads figure 4. kinetics of biosorption of zinc: pseudosecond order kinetics (mass of algal cell = 0.3 g, initial zinc concentration = 100 ppm, t = 25oc, ph = 5.0) time (min) t/q t( m in .g .m g1) time (min) q t( m g/ g) 24 the journal of engineering research vol. 5, no.1 (2008) 20-29 (5) where kd is the rate constant of intraparticle diffusion. intraparticle diffusion mechanism would be involved if plotting qt vs. t0.5 resulted in a straight line. figure 5 shows the variation of qt versus t0.5. for immobilized inactive algae and blank alginate beads the relationship is a straight line but doesn't pass through the origin which indicates that the intraparticle diffusion is not the only mechanism involved in the biosorption of zinc on immobilized dead algal cell. 4.3 biosorption isotherm zinc ions biosorption capacities of blank alginate beads and immobilized dead algal cells are presented as a function of the equilibrium concentration of zinc ions and the results are shown in fig. 6. figure 6 displays that the amount of zn+2 ions sorbed per unit mass of the biosorbent increased with increasing equilibrium concentration of zinc ions in the biosorption medium. higher equilibrium concentration enhances the mass transfer driving force, thus increasing the uptake. in addition, increasing equilibrium metal ion concentrations increases the number of collisions between metal ions and sorbent, which enhances the sorption process (abu al rub, et al. 2004). from this figure, the maximum biosorption capacity on the blank alginate beads was 8.58 mg/g and on immobilized dead algal cells was 9.38 mg/g. table 3 compares the maximum sorption capacities obtained in this study with some other values reported in the literature. the sorption capacity of zinc using the immobilized dead algal cells chlorella vulgaris is greater than that has been found using similar biosorbents. it is known that the constituents of the cells wall of algae provide an array of ligands with a mosaic of functional groups capable of binding various metallic ions. indeed, it has been shown that many metal-binding mechanisms are involved in the biosorption process; these include ion exchange, complexation, coordination, and microprecipitation (sheng, et al. 2004). in addition to these, immobilization enhances the contribution of physical sorption. the sorption data were analyzed according to the linear form of the langmuir isotherm (6) 5.0d tkq = t0 .5 ( m in ) 0 .5 1 2 3 4 5 6 t 0 1 2 3 4 5 6 7 im m o b iliz e d d e a d a lg a l c e ll b la n k a lg in a te b e a d s figure 5. variation of qt versus time (weber-morris equation 5.6) ce (mg/l) 0 50 100 150 200 250 300 350 0 2 4 6 8 10 immobilized dead algal cell blank alginate beads figure 6. experimental isotherms of zinc ions sorbed on different sorbents (ph = 5.0 mass of algal cell = 0.3 g) adsorbent adsorption capacity (mg/g) of zn+2 reference waste tea leaves 11.77 tee and khan, 1988 moss (mixture) 9.87 al-asheh et al. 1997 hazelnut shells 1.78 cimino and toscano, 2000 peat 9.3 mckay et al. 1998 fungi 9.81 puranik et al. 1999 aquatic moss fontinalis antipyretica 15 martins et al. 2004 immobilized dead algal cells chlorella vulgaris 9.38 this study table 3. adsorption capacity for zn+2 using different low cost adsorbents e e max bc1 bc qq + = t0.5 (min)0.5 q t( m g. g1 ) ce (mg/l) q e (m g/ g) 25 the journal of engineering research vol. 5, no.1 (2008) 20-29 where qmax is the maximum sorbate uptake under the given conditions, b is a coefficient related to the affinity between the sorbent and sorbate. the plots of specific sorption (ce/qe) against the equilibrium concentration (ce) for zinc are shown in fig. 7, and the isotherm constants qmax, b, along with the correlation coefficient, r2, are presented in table 4. the r2 values suggest that the langmuir isotherm provides a good model of the sorption system. the sorption equilibrium constant of the langmuir model b provides a measure for the adsorption efficiency since it indicates the sorbent affinity at low concentrations, and hence it measures the initial gradient of the adsorption isotherm. higher values of b indicate higher affinity and thus higher sorption efficiency. the sum of the squares of the errors (errsq) was used as a non-linear error function to determine the langmuir parameters as used by (ho, et al. 2002): (7) where p is the number of experimental data. the results of using the non-linear error analysis are shown in table 5, where they show that the langmuir isotherm constants b, and qmax are very consistent among all methods and they were very close to those obtained using linear analysis. the freundlich equation is one of the earliest empirical equations used to describe sorption equilibrium data. it does not establish a finite uptake capacity of the sorbent and can thus be reasonably applied in the low to intermediate concentrations ranges. the freundlich isotherm model is given by the equation (ho, et al. 2002): (8) where k and n are the freundlich constants which represent sorption capacity and sorption intensity, respectively. the linear freundlich isotherm plots for the sorption of the zinc onto both blank alginate beads and immobilized dead algal cells are presented in fig. 8. table 4 shows the linear freundlich sorption isotherm constants and the coefficients of determination; r2. examination of the plot and the r2 values suggests that the freundlich isotherm could describe the sorption of zinc on both sorbents. this is supported by the fact that algal cells have active sites with different energies, a key assumption in freundlich isotherm model. the values of the freundlich model parameters can be used to predict the affinity between the c e (m g .l -1 ) 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 c e /q e ( g .l 1 ) 0 5 1 0 1 5 2 0 2 5 3 0 3 5 im m o b iliz e d d e a d a lg a l c e ll b la n k a lg in a te b e a d s c e/ q e (g .l1 ) ce (mg.l -1) figure 7. langmuir isotherms of zinc ions sorbed on different sorbents of algal cell = 0.3 g, ph = 5.0) n/1ee kcq = model parameter blank alginate beads immobilized algal cells k (l/mg)1/n (mg/g) 2.85 3.26 n 5.01 5.15 freundlich r2 0.95 0.96 qmax (mg/g) 9.25 9.67 b (l/mg) 0.04 0.06 langmuir r2 0.99 0.99 qd (mmol/g) 8.83 9.44 bd (l/j2.mol2) 2.01*109 2*10-9 e (kj/mole) 15.772 15.81 d-r r2 0.97 0.98 table 4. adsorption linear isotherms parameters for the sorption of zinc ions by blank alginate beads and immobilized algal cells model blank alginate beads immobilized algal cells freundlich: k (l/mg)1/n (mg/g) 2.94 3.58 freundlich n 5.27 5.86 marquadt's psd 0.031 0.12 langmuir qmax (mg/g) 8.37 8.83 langmuir b (l/mg) 0.087 0.18 errsq 0.96 0.95 table 5. adsorption non-linear isotherms parameters for the sorption of zinc ions by blank laginate beads and immobilized algal cells errsq = ( )∑ = − 2 1 p e,exp e,cal i i q q 26 the journal of engineering research vol. 5, no.1 (2008) 20-29 sorbate and sorbent. high values of these parameters indicate high adsorptive capacity. from table 4 the magnitude of k and n for both sorbents suggests easy uptake with high sorptive capacity. the freundlich isotherm constants were also determined by the non-linear regression of marquardt's percent standard deviation (mpsd) (marquardt, 1963): (9) and the results are tabulated in table 5. these results demonstrate that the values of the k and n obtained by non linear regression are remarkably consistent and quite similar to the linear transform values shown in table 4. another less commonly used model to describe sorption of zinc on immobilized dead algal cells is the dubinin-radushkevich (d-r) isotherm. this isotherm is generally expressed as follows (dubinin, 1960): (10) radushkevich (1949) and dubinin (1965) have reported that the characteristic sorption curve is related to the porous structure of the sorbent. the constant, bd, is related to the mean free energy of sorption per mole of the sorbate as it is transferred to the surface of the solid from infinite distance in the solution and this energy can be computed using the following relationship (ho, et al. 2002): (11) figure 9 is a plot of the linear form of eq. (10), and the parameters found from the slope and intercept are listed in table 4. the values of e calculated are 15.77 kj/mole and 15.81 kj/mole for blank alginate beads and immobilized dead algal, respectively. these values are within the range of ion-exchange mechanisms (8-16 kj/mole) (abu al-rub, 2004), confirming that the ion exchange mechanism plays a significant rule in the biosorption mechanism. this finding of the ion exchange mechanism contribution agrees with the suggested mechanism of biosorption proposed by (volesky, 2003). the qd values are consistent with the linear qmax values previously determined for the langmuir isotherm. 4.4 sorption/desorption of zinc sorption and desorption of zinc on immobilized dead algal cells and blank alginate beads were investigated by conducting three repeated cycles sorption/desorption experiments. desorption experiments were carried out by shaking the sorbents with 20 ml 0.1 m hcl for two hours. the sorbents were then rinsed with deionized water to remove any residual acidity. the regenerated sorbents were then used in the removal of zinc ions and the results are demonstrated in fig. 10. figure 10 indicates that zinc uptake was improved after the first cycle. this improvement may be attributed to the fact that the acid used in desorption could remove some contaminants that might have been bound previously by the algal cells. ln (c e) 0 1 2 3 4 5 6 e 0.0 0.5 1.0 1.5 2.0 2.5 im m obilized dead algal cell b lank alginate beads ln (ce) ln ( q e) figure 8. freundlich isotherms of zinc ions sorbed on different sorbents (mass of algal cell = 0.3 g, ph = 5.0) ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜ ⎝ ⎛ ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ +−= 2 e dde )c 1 1ln(rtbexpqq db2 1 e = (ln (1+1/ce)) 2 20 40 60 80 100 120 (q e ) 0.0 0.5 1.0 1.5 2.0 2.5 immobilized dead algal cell ln (1+1/ce), blank vs ln (qe), blank (ln (1+1/ce)) 2 ln ( q e) figure 9. dubinin-radushkevich equation isotherm of zinc ion sorbed on different sorbents (mass of algal cell = 0.3 g, ph = 5.0) mpsd = ∑ = ⎛ ⎞− ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜⎝ ⎠ 2 1 p e,exp e,cal i e,exp i q q q 27 the journal of engineering research vol. 5, no.1 (2008) 20-29 4.5 effect of impurities on biosorption of zinc the biosorption of zn+2 + cu+2 and zn+2 + ni+2 at equilibrium ph 5.0 on immobilized dead algal cells are shown in fig. 11. these figures show the equilibrium uptake of zinc decreased regularly with increasing equilibrium metals concentrations. the decrease in the uptake of zinc in the presence of ni+2 ions and cu+2 ions is simple to explain. the presence of the other component develops a competition for the adsorption sites on the surface and some sites are occupied by the second component. as a consequence, the first component has a smaller "parking space" and its uptake is decreased (aksu and akpinar, 2001). these results were in agreement with other studies using different adsorbent and/or different heavy metals (abu al rub, 2004; juang, et al. 1999). the electronegativity values of zn+2, cu+2, and ni+2 metal ions are 1.65, 1.75, and 1.8 respectively. the more electronegative metal ions will be more strongly attracted to the immobilized dead algal cells surface (arica, et al. 2004). cu+2 has the highest affinity for immobilized dead algal sites and it has a greater negativity than both zn+2 and ni+2. this can explain the significant suppression of zn+2 uptake in the presence of cu+2. 5. conclusions this study proved the practical feasibility of using blank alginate beads and immobilized dead algal cells for the removal of zinc from wastewaters. the solution ph plays a very important role for zinc adsorption onto both sorbents. at ph < 2, zinc did not adsorb, the adsorption increased drastically as the ph was increased from 2.0 to 6.0, and the maximum adsorption uptake was achieved at ph 5.0. in addition, the uptake of zinc increased as the equilibrium zinc ion concentration increased. sorption of zinc on these sorbents was found to follow pseudo-second order kinetics. moreover, alginate beads and immobilized chlorella vulgaris cells could be used in successive sorption/desorption cycles to remove zinc ions from aqueous solutions, which suggest that immobilization can provide an efficient and convenient method for repetitive use of algal cells. sorption of zinc on alginate beads and immobilized chlorella vulgaris has been found to follow langmuir, freundlich, or d-r isotherm models. the presence of cu+2and ni+2 in the system led to a decrease in the removal of zinc ions by immobilized dead algal cells. acknowledgments the contribution of dr. aly dowdair in the analysis procedures is gratefully acknowledged. references abdulkarim, m. and abu al-rub, f., 2004, "adsorption of lead ions from aqueous solutions on activated carbon and chemically modified activated carbon prepared from date pits," adsorption science and technology, vol. 22 pp. 119-134. abu al-rub, f. a., 2006, "the biosorption of zinc on palm tree leaves: equilibrium, kinetics, and thermodynamics studies," sep. sci. technol., vol. 41, pp. 3499-3515. abu al-rub, f. a., 2004, "adsorption of lead on jordanian low-grade phosphate," ads. sci. technol., vol. 22, pp. 65-179. abu al-rub, f., el-naas, m., benyahia f. and ashour, i., 2004, "biosorption of nickel on blank alginate beads, free and immobilized algal cells," process biochemistry, vol. 39, pp. 1767-1773. c y c le n u m b e r 0 1 2 3 4 0 2 4 6 8 1 0 im m o b ilize d d e a d a lg a l c e ll b la n k a lg in a te b e a d s cycle number q e (m g/ g) figure 10. repeated uptake removal of zinc (initial zinc ions concentration = 100 ppm, ph = 5.0, mass of algal cell = 0.3 g) ce zn +2 (mg/l) 0 50 100 150 200 250 300 350 q ez n + 2 (m g /g ) 0 2 4 6 8 10 in the presence of cu+2 in the presence of ni+2 single zn+2 ce zn+2 (mg/l) q e z n+ 2 (m g/ g) figure 11. effect of the presence of other metals on zinc biosorption (ph 5, mass of algal cells = 0.3 g) 28 the journal of engineering research vol. 5, no.1 (2008) 20-29 abu al-rub, f., el-naas, m., ashour, i. and al-marzouqi, m., 2006, "biosorption of copper on chlorella vulgaris from single, binary and ternary metal aqueous solutions," process biochemistry, vol. 41, pp. 457-464. agrawal, a., sahu, k.k. and pandey, b.d., 2004, "removal of zinc from aqueous solutions using sea nodule residue," physicochem. eng. aspects, vol. 237, pp. 133-140. aksu, z. and akpinar, d., 2001, "competitive biosorption of phenol and chromium (vi) from binary mixtures onto dried anaerobic activated sludge," biochemical engineering journal, vol. 7, pp. 183193. aksu, z. and gonen, f., 2004, "biosorption of phenol by immobilized activated sludge in a continuous packed bed: prediction of breakthrough curves," process biochemistry, vol. 39, pp. 599-613. aksu, z. and yener, j., 2001, "a comparative adsorption / biosorption study of mono-chlorinated phenols onto various sorbents," waste management, vol. 21, pp. 695-702. arica, m.y., bayramoglu, g., yýlmaz, m., bekta, s. and genç, ö., 2004, 'biosorption of hg2+, cd2+, and zn2+ by ca-alginate and immobilized wood-rotting fungus funaliattrogii," j. of hazardous materials, vol. b109, pp. 191-199. denzili, a., cihanger n., rad, a.y., taner, m. and alsancak g., 2004, "removal of chlorophenols from synthetic solutions phanerochaete chrysosporium," process biochemistry, vol. 39, pp. 2025-2030. dubinin, m.m, 1960, "the potential theory of adsorption of gases and vapors for adsorbents with energetically non-uniform surface," chem. rev, vol. 60, pp. 235-266. feng, d. and aldrich c., 2004, "adsorption of heavy metals by biomaterials derived from the marine alga ecklonia maxima," hydrmetallurgy, vol. 73, pp. 110. galiatsatou polymnia, metaxas michail. and kasselouririgopoulou vasilia, 2002, adsorption of zinc by activated carbons prepared from solvent extracted olive pulp," j. of hazardous materials, vol. 91, pp. 187-203. gupta r., ahuja p., khan s. and saxena, r.k., 2000, "microbial biosorbents: meeting challenges of heavy metal pollution in aqueous solutions," current science, vol. 78, pp. 967-973. ho, y.s., 2004, "pseudo-isotherms using a second order kinetic expression constant adsorption," vol. 10, pp. 151-158. ho, y.s., chiang t.h. and hsueh, y.m., 2005, "removal of basic dye from aqueous solution using tree fern as a biosorbent," process biochemistry, vol. 40, pp. 119-124. ho, y.s., porter, j.f. and mckay, g., 2002, "equilibrium isotherm studies for the sorption of divalent metal ions onto peat: copper, nickel, and lead single component system," water, air, and soil pollution, vol. 141, pp. 1-33. ho, y.s. and mckay, g., 1999, "the sorption of lead (ii) ions on peat," water research, vol. 33, pp. 578-584. ibanez, j.p. and umetsu y., 2004, "uptake of trivalent chromium from aqueous solutions using protonated dry alginate beads," hydrometallurgy, vol. 72, vol. 327-334. juang, r.s., wu, f.c. and tseng, r.l., 1999, "adsorption removal of copper (ii) using chitosan from simulated rinse solutions containing chelating agents," water research, vol. 33, pp. 2403-2409. kongolo, k., mwema, m.d., banza, a.n. and gock, e., 2003, "cobalt and zinc recovery from copper sulphate solution by solvent extraction," minerals engineering, vol. 16, pp. 1371-1374. kurama, h. and catalsarik, t., 2000, "removal of zinc cyanide from a leach solution by an anionic ionexchange resin," desalination, vol. 129, pp. l-5. magdy, y.h. and daifullah, a.a.m., 1998, "adsorption of a basic dye from aqueous solutions onto sugarindustry-mud in two modes of operations," waste management, vol. 18, pp. 219-226. marquardt, d.w., 1963, "an algorithm for least-squares estimation of non-linear parameters," j. soc. (ind.) appl. math, vol. 11, pp. 431-441. martins, r., pardo, r. and boaventura, r., 2004, "cadmium(ii) and zinc(ii) adsorption by the aquatic moss fontinalis antipyretica: effect of temperature, ph and water hardness," water research, vol. 38, pp. 693-699. mckay, g., porter, j.f. and prasad, g.r., 1999, "the removal of dye colours from aqueous solutions by adsorption on low-cost materials," water air soil pollution, vol. 114, pp. 423-438. mohan, d. and singh k., 2002, "singleand multicomponent adsorption of cadmium and zinc using activated carbon derived from bagasse an agricultural waste," water research, vol. 36, pp. 2304-2318. mohapatra, h. and gupta, r., 2005, "concurrent sorption of zn(ii), cu(ii) and co(ii) by oscillatoria angustissima as a function of ph in binary and ternary metal solutions," bioresource technology, vol. 96, pp. 1387-1398. pagnanelli, f., trifoni, m., beolchini, f., esposito, a., toro, l. and veglio, f., 2001, "equilibrium biosorption studies in single and multi-metal systems," process biochemistry, vol. 37, pp. 115124. pagnanellia francesca., mainellia sara., vegli francesco. and toroa luigi., 2003, "heavy metal removal by olive pomace: biosorbent characterization and equilibrium modelling," chemical engineering science, vol. 58, pp. 4709-4717 peric, j., trgo, m. and medvidovic, v., 2004, "removal of zinc, copper and lead by natural zeolite a comparison of adsorption isotherms," water research, vol. 38, pp. 1893-1899. preston, j.s. and du preez, a.c., 2000, "separation of 29 the journal of engineering research vol. 5, no.1 (2008) 20-29 nickel and calcium by solvent extraction using mixtures of carboxylic acids and alkylpyridines," hydrometallurgy, vol. 58, pp. 239-250. ramos leyva., jacome bernal., barron mendoza., rubio fuentes. and coronado guerrero., 2002, "adsorption of zinc(ii) from an aqueous solution onto activated carbon," j. of hazardous materials, vol. b90, pp. 2738. rao, j.r. and viraraghavan, t., 2002, "biosorption of phenol from an aqueous solution by aspergillus niger biomass," bioresource technology, vol. 85, pp. 165-171. sheng, p.x., ting, y.p., chen, j.p. and hong, l., 2004, "sorption of lead, copper, cadmium, zinc, and nickel by marine algal biomass: characterization of biosorptive capacity and investigation of mechanisms," j. of colloid and interface science, vol. 27, pp. 131-141. srisuwan, g. and thongchai, p., 2004, "removal of heavy metals from electroplating wastewater by membrane," science technology, vol. 24, pp. 965976. strelko, v., malik, d.j., streat, m., 2002, "characterization of the surface of oxidized carbon adsorbents, carbon," vol. 40, pp. 95-104. veeken, h.m., akoto, l., hulshoff, w. and weijma, j., 2003, control of the sulfide (s-2) concentration for optimal zinc removal by sulfide precipitation in a continuously stirred tank reactor," water research, vol. 37, pp. 3709-3717. volesky, b., 2003, "sorption and biosorption. (isbn 09732983-0-8),” bv sorbex, inc. monterial st. lambert, quebec, canada. weber, w.j. and morris j.c., 1963, "kinetics of adsorption on carbon from solution," j. sanitary engineering division, american society of chemical engineering, vol. 89, pp. 31-59. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 16, no. 2 (2019) 87-95 power system stabilizer design using notch-filter – barka ii (oman) case study h.m. soliman, m.h. albadi * , a.al-maamari, and a.al-kasbi department of electrical and computer engineering, sultan qaboos university, muscat 123, po box 33, oman abstract: the use of power system stabilizers (psss) to damp power system swing mode oscillations is of extreme practical importance. this manuscript presents an approach to the stabilization of a single machine infinite bus system (smib). the proposed control is based on the notch-filter approach to cancel the poles near to the imaginary axes. the approach is based on the root locus method. application to barka ii power station connected to the main interconnected system of oman is presented. the peak load at summer is considered as the system is near to instability. the barka smib is modeled as a fourth order non-linear system. a linearized model is then obtained using matlab/simulink. the simulation results validate the proposed design for the pss. keywords: notch-filter; power system stabilizers; root locus. عمان()ii بركاء دراسة حالة -تصميم مثبت نظام الطاقة باستخدام مرشح الشق الكاسبي ، أ. المعمري ، أ. *م. ح. البادي ،هـ. م. سليمان عملية ذو أهمية كبرى. وتقوم هذه الورقة البحثية قدرهلتقليل تأرجح نظام ال ةقدر : يعد استخدام مثبتات نظام الالملخص يستند عنصر التحكم المقترح إلى نهج و . ( )مولد واحد متصل بشبكه كبيرهة قدربتقديم طريقة لتحقيق الثبات في نظام . وتم ورالمحل الهندسي للجز. ويعتمد النهج على طريقة ) التخيلي( الرأسيالقريبة من المحور الجذورمرشح الشق إللغاء يعتبر الِحمل االعلى في الصيف بمثابة و في سلطنة عمان. هالرئيسي الشبكهبالمتصلة ii التطبيق على محطة كهرباء بركاء على أنه عند اتصالها بشبكه عمانبركة توليد ةلمحط ةمثبت نظام القدرالنظام من عدم االستقرار. وتم تصميم اقتراب و اثبتت نتائج .matlab / simulink نموذج خطي باستخدام استنتاجبعة. ثم تم نظام غير خطي من الدرجة الرا .المحاكاة صحة التصميم المقترح لمثبت نظام الطاقة __________________________________________________________________________________________ .المحل الهندسي للجزور؛ ةمثبتات نظم القدر؛ مرشح حزمه ضيقه للتردد: الكلمات المفتاحية *corresponding author’s e-mail: mbadi@squ.edu.om doi:10.24200 /tjer.vol.16iss2pp87-95 mailto:mbadi@squ.edu.om the journal of engineering research (tjer), vol. 16, no. 2 (2019) 87-95 nomenclature cpss conventional pss facts flexible ac transmission system mis main interconnected system pss power system stabilizer pid proportional–integral–derivative smib single machine infinite bus system mechanical input power of the generator in p.u. electrical output power of the generator in p.u generator direct-axis transient reactance in p.u. direct and quadrature-axis synchronous reactances, respectively, in p.u. transmission line and transformer reactance in p.u. d-axis open circuit field time constant, sec. m inertia constant in seconds. δ rotor angle in rad. ω rotor speed in rad/s. quadrature-axis transient voltage in p.u. field voltage in p.u. exciter-avr gain and time constant, respectively. v infinite bus voltage in p.u. p, q machine real and reactive power loading at the infinite bus in p.u, respectively. s laplace operator. 1. introduction power systems are frequently subjected to undesirable disturbances due to several reasons such as continuous load variations, set-point changes, lightning strikes, and faults. consequently, they exhibit low frequency oscillations that may either decay gradually, or continues to grow, causing system separation. these low frequency oscillations are due to the lack of electromechanical system damping (kundur, balu et al. 1994, sauer and pai 1998, pal and chaudhuri 2006, machowski, bialek et al. 2008). the desired additional damping can be provided by a supplementary excitation control through a power system stabilizer (pss). the main problem encountered in conventional pss design is that power systems constantly experience changes in operating conditions due to variations in generation and load. therefore, a conventionally designed pss may fail to maintain stability over a wide range of operating points. further, the performance of conventional pss is degraded once the deviation from the nominal point becomes significant. coping with uncertainties imposed by continuous variation in operating points, has become the priority of the pss designers. to make the performance of a pss robust, the design algorithm must take into consideration power system uncertainty due to load variation. power system uncertainty can be modeled in different approximations: interval plant (soliman, elshafei et al.2000; soliman 2014), polytopic form (rao and sen 2000), norm-bounded form (soliman and shafiq 2015), μ-synthesis (castellanos, messina et al. 2008), and linear fractional transformation (werner, korba et al. 2003). there exist different pss designs. the pss that keeps system stability in the face of system uncertainty is termed robust stabilizer; whereas if it can retain the closed loop poles in a desired region so as to achieve good dynamic performance, it is called pole placer. the design can be carried out in the s-domain, complex frequency, or in the time domain. in the sdomain, a phase-lead or pid-robust pss using kharitonov theorem is presented in soliman, elshafei et al. (2000) and soliman 2014), respectively. in the time domain, the powerful linear matrix inequality (lmi) optimization is used for the synthesis of pole placer pss (rao and sen 2000). power systems that are subjected to a series of lightning strikes with the associated auto reclosure of circuit breakers are represented as a markov chain and the pole placer synthesis is presented in soliman and shafiq (2015). when a power system is equipped with a facts controller in addition to the pss, using both controllers provides tighter control grip on the system. in case of failure of either one, a design called reliable control is presented in soliman, dabroum et al. (2011) to preserve system stability. fuzzy logic has recently emerged as a potential technique for pss design. besides its ability to accommodate the heuristic knowledge of a human expert, the advantage of a fuzzy pss is that it represents a nonlinear mapping that can cope with the nonlinear nature of power systems. an adaptive pss that uses on-line self-learning fuzzy systems is discussed in elshafei, el-metwally et al. (2005). although the performance of a well-designed modelfree fuzzy pss is acceptable, it lacks systematic stability analysis and controller synthesis. many attempts to overcome this drawback have been made by providing a model-based fuzzy pss that guarantees stability and performance of power systems. in the past ten years, research efforts on fuzzy logic control have been devoted to model-based fuzzy control systems (feng 2006). stability and performance limits of model-based fuzzy control systems can be achieved via lmi techniques (tanaka and wang 2004). an indirect adaptive fuzzy controller as a power system stabilizer also, used to damp inter-area modes of oscillation is presented in hussein, saad et al. (2010). most power system controls are subject to saturation due to physical limitations of actuators (kundur, balu et al. 1994; sauer and pai 1998). none of the above references tackled the problem of control signal limits practically imposed. when the controller saturation is not considered in the design phase, the performance of the designed control system seriously deteriorates. the design of a robust pss pole placer taking into consideration the control constraints is reported recently in soliman and shafiq (2015) and soliman and el metwally (2017). the designs achieve regional pole placement, to obtain good dynamic behavior against load variation; subject to the control limit constraint. metaheuristics in the 88 h.m. soliman, m.h. albadi, a.al-maamari, and a.al-kasbi design of robust pss is given b y peres, júnior et al. (2018). metaheuristics are also used to tackle power system nonlinearities as presented in rahmatian and seyedtabaii (2019). although the above progress in pss design tackles difficult problems of power systems dynamics, it uses state or output feedback which is not the standard form of the conventional pss (cpss). in this paper, three pss designs are presented: 2 conventional (single lead, double lead) and the new proposed approach (notch-filter based double lead). the paper demonstrates that the conventional approaches fail to stabilize the system whereas the newly proposed notch-filter based double lead provides a very effective method to stabilize the system. this result is due to the fact that the system has slightly damped complex poles. in addition, the paper uses a notch-filter based pss to stabilize a practical power system in oman (barka ii power plant). the paper is organized as follows. the mathematical model of smib power system and the problem statement are presented in section 2. section 3 presents the problem solution using the root locus method and the stability analysis of the closedloop system under the proposed control scheme. simulation results for the smib power system are provided in section 4 to validate the effectiveness of the proposed controller. the paper is concluded in section 5. 2. single-machine infinite-bus model in this section, the mathematical model for the non-linear dynamics of a singlemachine infinitebus system (smib) is presented. the smib will be used in the design procedure, which will be addressed in the next section. the machine delivers the electrical power to the infinite bus. the voltage regulator controls the input to a solidstate rectifier excitation system, which provides the field voltage to maintain the generator terminal voltage at a desired value, fig. 1 (kundur, balu et al. 1994, sauer and pai 1998). the states of the machine are its rotor angle , its speed , its quadrature-axis transient voltage , and the field voltage . the exciter-voltage regulator is modeled as a first order transfer function. all of the variables are normalized on a per-unit (p.u.) basis, except the time constant is in seconds. the mathematical nonlinear model describing the dynamics of smib is given by the following equation the states of the machine are its rotor angle, its speed, its quadrature-axis transient voltage, and the field voltage. the exciter-voltage regulator is modeled as a first order transfer function. all of the variables are normalized on a per-unit (p.u.) basis, except the time constant is in seconds. the mathematical nonlinear model describing the dynamics of smib is given by the following equation (kundur, balu et al. 1994; sauer and pai 1998): ̇ (1) ̇ [ ( ) ] ̇ [ ] (3) ̇ ( ) (4) where , , and are given by the following equation: (5) ( ) (6) √( ) ( ) (7) the power system model is linearized at a particular equilibrium point to obtain the linearized system model given in the state-space form. ̇ & (8) the components of the state vector are defined as [ ] where δ denotes the perturbation of the states, input, and outputs from their equilibrium values. figure 1. single-machine infinite-bus system. 89 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 87-95 3. design of pss in this section, three pss designs will be presented: single lead, double lead and notch-filter based double lead. it will be shown that the conventional single and double lead fail to stabilize the system. whereas the notch-filter based double lead provides a very effective method to stabilize the system. this is due to the fact that the system has slightly damped complex poles. 3.1 open loop response: the system response without pss for a cleared three phase fault at the generator terminal causing δδ=0.1 rad is shown in fig. 2. it is evident that the open loop response is very oscillatory. so, a pss is needed to damp the oscillations. 3.2 pss design the conventional pss (cpss) is a lead controller. it has two forms: single lead, or double lead; which have the forms: , , respectively. design trials using the root locus method will be presented. 1) design trial#1: single and double lead pss the linearized model for barka ii working during the extreme summer condition is obtained from (11). selecting typical values for the poles and zeros of a single lead controller, the pss is given as follows: the zoomed root locus for the single lead pss is shown in fig. 3. the zoomed root locus for the double lead pss is shown in fig. 4. figure 2. system response without power system stabilizer. figure 3. zoomed root locus design for single lead power system stabilizer. -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 -15 -10 -5 0 5 10 15 root locus real axis (seconds -1 ) im a g in a ry a x is ( s e c o n d s -1 ) 90 h.m. soliman, m.h. albadi, a.al-maamari, and a.al-kasbi the pss designed by the root locus method was unsuccessful because it contains pairs of complex-conjugate poles (-0.124 ± j13.2) that lie close to the imaginary axis in the s-plane as shown in fig. 3 and fig. 4. these poles are lightly damped. this will result in an unstable system even for small controller gain. to solve this problem, a notch filter is used to cancel those poles (pole-zero cancellation). 2) design trial#2: notch-filter based double lead pss there are many transfer functions of a controlled process that contains pairs of complex-conjugate poles located close to the imaginary axis in the s-plane. this will result in a lightly damped performance. one way to control this system is to design a controller with zeros near the undesirable lightlydamped poles of the plant. these zeros can attenuate the effect of the lightly-damped poles. the dominant closed-loop poles of the system can then be placed in a more desirable position. such a controller is called a notch filter. before getting into the specifics of a notch filter, it should be noted that due to the nature of most systems, exact pole/zero cancellation cannot be obtained; nor should it be attempted. approximate cancellation will give us many of the desirable characteristics without the pitfalls. unfortunately, this method is unreliable because when an added zero does not exactly cancel the corresponding unstable pole (which is always the case in real life), a part of the root locus will be trapped in the right-half plane. this causes the closed-loop response to be unstable. therefore, the pole-zero cancellation method can be carried only in lhs of the complex plane (stable region). pole-zero cancellation is a straightforward search through the poles and zeros looking for matches that are within tolerance. the transfer functions are first converted to zero-pole-gain form. the notch filter is used to cancel complex-conjugate poles as seen in fig. 5. figure 4. zoomed root locus design for double lead power system stabilizer. figure 5. root locus design for notch-filter based double lead power system stabilizer. -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 -20 -15 -10 -5 0 5 10 15 root locus real axis (seconds -1 ) im a g in a ry a x is ( s e c o n d s -1 ) real axis (seconds -1 ) im a g in a ry a xi s (s e co n d s1 ) 91 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 87-95 to cancel those poles, we elapse on them complex-conjugate zeros. this method is known as pole-zero cancellation. as seen, the roots of the numerator of the controller are the same as the complex poles of the denominator of the plant; two poles at -1, -20 are selected for the denominator of the controller. using the matlab command “rlocus”, one can generate a root locus plot as shown in fig. 5. the complex poles near the imaginary axis have been canceled and more of the root locus is now in the left half plane. this means that a larger controller gain, k, can be employed, while maintaining stability. the best gain to dampen oscillation and to be far away from unstable region is shown in fig. 6. the resulting double lead pss that cancels the poles near to the imaginary axes with best gain, pushing the poles to the left as far as possible is given below. (9) the proposed design will be tested through an actual system in oman. 4. simulation results power stability can be classified into three types: angle, frequency, and voltage. this section tackles the angle-stability problem. the simulation will demonstrate the effectiveness of the proposed design. the proposed design is applied to the barka ii power station in mis. more details about power plants in mis are found in albadi ( 2017) and albadi, el-rayani et al. 2018). barka ii has 5 synchronous generators (3 gas turbines and 2 steam turbines) connected to the grid as shown in fig. 7. real axis figure 6. double lead (notch filter) power system stabilizer with the best gain. oman mis (v∞) p&q transformer 1 figure 7. barka ii power station. transformer 5 synch. generator 5 synch. generator 1 92 h.m. soliman, m.h. albadi, a.al-maamari, and a.al-kasbi table 1. data of the steam generator in barka ii. parameter value xd 2.38 pu x𝑑′ 0.257 pu xq 2.26 pu o 314 rad/s t𝑑0′ 13.22 s h 1.6 s ke 25 te 0.05 s xe 0.15 p 0.6 pu q 0.3 pu v 1 pu manufacturer data sheets, obtained from the barka ii power station, for the steam-turbine generator equipped with a pss are used to get data required for the simulation. the rest of the four machines in barka ii station and mis are represented by their thevenin equivalent. the peak load of 120mw, 60 mvar, occurs in summer, 2016 is used in the simulation. considering base values of 200 mva and 11 kv, the above machine and transformer data are represented in p.u in table 1. the linearization of (1-7) around the operating point (p=0.6, q=0.3) can be done analytically or numerically. the analytical method is given as follows (kundur, balu et al. 1994): to calculate the heffron-phillips model parameters, the fourthorder model should be linearized. the linearized form of the model (1-7) is as follows: [ ] [ ] (10) in (10), k1, k2, k3, k4, k5, k6, are constants depending on system parameters and the operating conditions. analytical expressions for calculating the k’s are given in soliman, elshafei et al. (2000). this model, known as the heffron-phillips model, is shown in fig. 8. a much easier method for linearization is using the matlab/simulink command “linmod”. for the generator peak load (120 mw, 60 mvar), the corresponding operating condition is [0.6359, 0, 1.2349, 1.9999]. the state-space model is given below: the response to the same disturbance mentioned before, 0.1 rad, is shown in fig. 9. as seen from fig, 9, the system without pss takes a long time to dampen the oscillation, which is more than 20s. on the other hand, the system with pss takes less than 6s to damp the oscillations. it is worth noting that persistent oscillations are very harmful to the system as they decrease the lifetime of generators and can damage the rotor (shaft fatigue). [ ] [ ] , [ ] & [ ] (11) 5. conclusion power system stabilizers have been considered to improve power system damping. the pss is an auxiliary control system that is applied as part of an excitation control system. this paper proposes a linearized block diagram of a power system with a pss and the execution of the pss controller oscillations damping in a smib using a small signal model. the system model equations were generated from the phasor diagram of the smib and then were linearized using matlab/simulink. the root locus method was used to design the pss in two stages. in the first stage, a single lead was considered. a much better damping response was achieved by double lead in s stage. a pss was designed using double lead notch-filter method for actual system in oman using data of barka ii power station. simulation results reveal that the designed pss reduces the system oscillation from 20s to less than 6s. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this research. acknowledgment the authors would like to thank smn barka power plant (barka ii) for their support and collaboration. 93 the journal of engineering research (tjer), vol. 16, no. 2 (2019) 87-95 figure 8. heffron-phillips model. figure 9. system response without and with proposed notch-filter based power system stabilizer. references albadi, m. h. (2017), electricity sector in oman after 10 years of reform: status, trends, and future perspectives. the electricity journal 30(7): 23-30. albadi m.h. el-rayani y.m. el-saadany e.f. and al-riyami h.a. (2018), impact of solar power projects on lmp and transmission losses in oman. sustainable energy technologies and assessments 27: 141-149. castellanos r. a. messina and sarmiento h (2008), a μ-analysis approach to power system stability robustness evaluation. electric power systems research 78(2): 192-201. elshafei a.l. k. el-metwally and a. shaltout (2005), a variable-structure adaptive fuzzy-logic stabilizer for single and multi-machine power systems. control engineering practice 13(4): 413-423. feng g (2006), a survey on analysis and design of model-based fuzzy control systems. ieee transactions on fuzzy systems 14(5): 676-697. hussein t. m. saad a. elshafei and a. bahgat (2010), damping inter-area modes of oscillation using an adaptive fuzzy power system stabilizer. electric power systems research 80(12): 1428-1436. kundur p. n. j. balu and m. g. lauby (1994), power system stability and control, mcgraw-hill, new york. machowski j. j. bialek and j. bumby (2008), power system dynamics stability and control. john wiley and sons, chichester. pal b. and chaudhuri b (2006), robust control in power systems, springer science and business media. peres w. i. c. s. júnior and j. a. passos filho (2018), gradient based hybrid metaheuristics for robust tuning of power system stabilizers. international journal of electrical power & energy systems 95: 47-72. rahmatian m. and s. seyedtabaii (2019). multimachine optimal power system stabilizers design based on system stability and nonlinearity indices using hyper-spherical search method. international journal of electrical power and energy systems 105: 729-740. rao p. s. and i. sen (2000) robust pole placement stabilizer design using linear matrix inequalities. 94 h.m. soliman, m.h. albadi, a.al-maamari, and a.al-kasbi ieee transactions on power systems 15(1): 313319. sauer p.w. and pai m.a. (1998). power system dynamics and stability. upper saddle river, new jersey: prentice hall, usa. soliman h. a. elshafei a. shaltout and m. morsi (2000), robust power system stabiliser. iee proceedings-generation, transmission and distribution 147(5): 285-291. soliman h. and m. shafiq (2015), robust stabilisation of power systems with random abrupt changes. iet generation, transmission and distribution 9(15): 2159-2166. soliman h.m.a. dabroum m. mahmoud and m. soliman (2011), guaranteed-cost reliable control with regional pole placement of a power system. journal of the franklin institute 348(5): 884-898. soliman h.m. and k.a. el metwally (2017), robust pole placement for power systems using twodimensional membership fuzzy constrained controllers. iet generation, transmission and distribution 11(16): 3966-3973. soliman m. (2014), parameterization of robust threeterm power system stabilizers. electric power systems research 117: 172-184. tanaka k and h. o. wang (2004), fuzzy control systems design and analysis: a linear matrix inequality approach, john wiley and sons. werner h.p. korba and t. c. yang (2003), robust tuning of power system stabilizers using lmitechniques. ieee transactions on control systems technology 11(1): 147-152. 95 template for electronic submission to acs journals this paper template will be used only for accepted papers to encourage authors to submit to tjer template for submission of manuscripts to the journal of engineering research this document is to be used as a template for preparing manuscripts for submission. authors are requested to consult the “author guidelines” of the journal for detailed guidelines and procedures for submission. starting the next page, all parts of the manuscript are included (text, tables, and graphics). tables and figures need to be inserted as close as possible to the point at which they are discussed in the text of the manuscript, and must be either at the beginning or the end of the page. using the template a standard tjer publication includes different components (i.e., title, abstract, main text, figure captions, etc.). all of these components are included in the template. when filling the template, authors need to keep the original format of the template as is. this includes the fonts, spacing, page margins, figures and tables titles styles, headers and footnotes. see the guide, notes, notice, or instructions for authors on the journal’s homepage to determine which parts should be included for the manuscript that you are preparing. 1. the best way for filling any part of the template is to highlight the text of the template that you want to replace and begin typing your manuscript. 2. otherwise, if the document is already prepared, you will need to copy the paragraph from your original document and paste it in the proper place of the template using the paste option “keep text only”. this will reserve the original template formatting style and change the text only. 3. to insert graphics within the text or as a figure, chart, scheme, or table, create a new line and insert the graphic where desired. if your graphic is not visible, ensure that the word style is “normal” with an automatic height adjustment. if the size of the artwork needs to be adjusted, re-size the artwork in your graphics program and re-paste the artwork into the template (maximum width for single-column artwork, 7.7 cm; maximum width for double-column artwork, 16 cm. 4. make sure that page numbers are present on all pages before submitting your manuscript. 5. after completing the editing of your manuscript, please delete these instructions and any sections that are not needed. 6. save the file with the graphics and tables in place using the save as command with a .doc or .docx ms-word file type. 7. make sure that the text in your manuscripts are free of any typos and paragraphs have no widows and orphans (these are lines at the beginning or end of a paragraph, which are left dangling at the top or bottom of a column, separated from the rest of the paragraph) 8. replace this line with the main title of the manuscript submitted to the journal of engineering research (tjer) first a. author 1, second b. author 2*, third c. author 3 1 first department, first university, address, city, country name 2 second company department, company address, city, country name 3 third department, third university, address, country name abstract: these instructions give you guidelines for preparing papers for tjer. use this document as a template if you are using microsoft word 6.0 or later. otherwise, use this document as an instruction set. paper titles should be written in uppercase and lowercase letters, not all uppercase. do not write “(invited)” in the title. full names of authors are preferred in the author field. put a space between authors’ initials. the abstract must be a concise yet comprehensive reflection of what is in your article. in particular, the abstract must be self-contained, without abbreviations, footnotes, or references. it should be a microcosm of the full article. the abstract must be limited to 250 words. be sure that you adhere to these limits; otherwise, you will need to edit your abstract accordingly. the abstract must be written as one paragraph, and should not contain displayed mathematical equations or tabular material. the abstract should include three or four different keywords or phrases, as this will help readers to find it. it is important to avoid over-repetition of such phrases as this can result in a page being rejected by search engines. ensure that your abstract reads well and is grammatically correct. keywords: enter keywords or phrases in alphabetical order, separated by semicolons. عنـــوان البحـــــث المؤلفون فقط الملخص: space reserved for arabic translation. for arabic native speaking authors, please provide the arabic translation for: · paper title · author names · abstract (less than 250 words) · keywords الكلمات المفتاحية: أدخل الكلمات المفتاحية حسب الترتيب األبجدي بينها فواصل. corresponding author’s e-mail: author@email.com doi:10.24200/tjer.vol.xxissxppxxx-xxx ( the journal of engineering research (tjer), vol. xx, no. x, (20xx) xx-xx ) ( first a. author, second b. author and third c. author ) nomenclature a separate nomenclature section should list in detail the symbols used in the manuscript, their definitions, and their units in si system. the nomenclature list should be in an alphabetical order with greek symbols following the alphabetical listing. subscripts and superscripts should follow greek symbols and should be identified with a heading. 1. introduction this document is a template for microsoft word versions 6.0 or later. if you are reading a paper or pdf version of this document, please download the electronic file, tjer_template.docx, from the tjer website at https://journals.squ.edu.om/ index.php/tjer/about/submissions#author guidelines so you can use it to prepare your manuscript. if your paper is intended for a conference, please contact your conference editor concerning acceptable word processor formats for your particular conference. 2. guidelines for manuscript preparation when you open tjer_template.docx, select “page layout” from the “view” menu in the menu bar (view | page layout), (these instructions assume ms 6.0. some versions may have alternate ways to access the same functionalities noted here). then, type over sections of tjer_template.docx or cut and paste from another document and use markup styles. the pull-down style menu is at the left of the formatting toolbar at the top of your word window (for example, the style at this point in the document is “text”). highlight a section that you want to designate with a certain style, and then select the appropriate name on the style menu. the style will adjust your fonts and line spacing. do not change the font sizes or line spacing to squeeze more text into a limited number of pages. use italics for emphasis; do not underline. 3. tjer publishing policy the general tjer policy requires that authors should only submit original work that has neither appeared elsewhere for publication nor is under review for another refereed publication. the submitting author must disclose all prior publication(s) and current submissions when submitting a manuscript. do not publish “preliminary” data or results. the submitting author is responsible for obtaining the agreement of all coauthors and any consent required from employers or sponsors before submitting an article. the tjer strongly discourages courtesy authorship; it is the obligation of the authors to cite only relevant prior work. the tjer does not publish conference records or proceedings but can publish articles related to conferences that have undergone rigorous peer review. minimally, two reviews are required for every article submitted for peer review. 4. manuscript preparation length of manuscript papers submitted for publication in tjer could be up to ten printed pages. it is advisable that authors use the manuscript perpetration guideline in order to reduce the amount of time spent on typesetting the manuscript. the font sizes and the manuscript specifications are outlined as follows. typing and font specification manuscripts should be typed single-spaced on one side of a sheet only, with margins of about 2.5 cm on each side of each page. the font should be times new roman for all the components of the manuscript. font size should be as follows: manuscript title: 16pt, authors names: 11pt, affiliations:10pt, body text 10pt, section title: 12pt, bold, subsection title: 11pt, bold, italic. elements of the manuscript full-length papers generally consist of the title, author(s) affiliation(s), abstract, keywords, nomenclature if applicable, introduction, body, conclusion, references, list of figures and table captions, and original figures and tables for reproduction. a paper may also include appendices and acknowledgments. author names and affiliation(s) author names on technical papers should include the name of each author followed by his/her company affiliation. if two or more authors have the same company affiliation, it should not be repeated after each name. do not use titles such as mr., mrs., ms., miss, prof. or dr. keywords leave two blank lines after the abstract and type 6-10 key words and phrases (separated by semicolons) that can be used to index the paper, listing the keywords in order of importance for indexing. do not repeat words or strings of words from the title. capitalize the first letter of each word. mathematical notations and equations to avoid errors in editing and typesetting, authors should clearly identify subscripts, superscripts, greek letters, and other symbols. equations should be numbered using arabic numerals placed in parentheses. break equations to fit within the page. avoid ambiguities in equations and fractions in the text through the careful use of parentheses, brackets, solids (slants), etc. note that in-text, fractions are usually “broken down” to fit in one line and confusion can result if terms are not properly labeled. the conventional order of brackets is {[0]}. units and abbreviations the international system of units (si units) is required. unit symbols should be used with measured quantities, i.e. 1 mm, but not when unit names are used in the text without quantities, i.e. “a few millimeters.” acronyms and abbreviations should be defined the first time they are used in the text. tables all tables must be cited in the body of the paper. tables should be numbered in arabic numerals in the order they are cited in the paper. tables should be the smallest size possible without loss of clarity. table headings and column headings should be as concise as possible. table 1. this is a sample data table. temp.a densityb viscosityc surface tensiond 2 949.87 4 949.51 1230 53.06 6 949.99 999 53.02 8 949.75 876 54.39 10 949.07 795 55.29 14 948.02 689 55.89 18 946.69 623 56.02 22 945.12 573 56.99 26 943.18 536 57.01 a change in temperature above the melting point in oc. bdensity in kg/m3 cviscosity in cp. d surface tension in dyn/cm if the table or figure is wider than one column, you can switch to one column and insert it. the procedure is as follows: place the insertion point at the beginning of the text that will appear in the columns. display the page layout tab of the ribbon and click on the breaks tool. word displays a list of break types. click on continuous. word inserts a continuous section break in your document. place the insertion point at the end of the text that will appear in the columns. repeat steps 2 through 4 to insert another continuous section break. place the insertion point anywhere within the text that will appear in the columns. click the columns tool in the page setup group. word displays a number of column options. choose the option that indicates how many columns you desire. table 2 below is a copy of table 1 but reformatted with one column layout using the steps listed above. please include a brief summary of the main points of the paper, do not replicate the abstract as the conclusion, and consider elaborating on the translational importance of the work or suggest applications and extensions. figures all figures must be cited in the body of the paper. figures should be numbered in arabic numerals in the order in which they are cited. related figures or figure sections may be labeled with a shared numeral and consecutive lowercase letters (eg. figs. 1a through 1d). figures should be the smallest size possible without loss of clarity, preferably designed to occupy a single column (8.2 cm of width) of a printed journal page. once the paper is accepted for publication, a separate file for each figure in .tif or .tiff format is to be submitted to the technical editor for final printing. the proper resolution of your figures will depend on the type of figure it is as defined in the “types of figures” section. author photographs, color, and grayscale figures should be at least 300dpi. line art, including tables, should be a minimum of 600dpi. in order to preserve the figures’ integrity across multiple computer platforms, we accept files in the following formats: .eps/.pdf/.ps. all fonts must be embedded or text converted to outlines in order to achieve the best-quality results. to insert images in word, position the cursor at the insertion point and either use insert | picture | from file or copy the image to the windows clipboard and then edit | paste special | picture (with “float over text” unchecked). figures can be sized between column and page widths if the author chooses, however it is recommended that figures are not sized less than column width unless when necessary. a sample figure is given in figure 1 : figure. 1. figure caption is here. note that “fig.” is abbreviated. there is a period after the figure number, followed by two spaces. it is good practice to explain the significance of the figure in the caption. table 2. this is a sample data table. ( first a. author, second b. author and third c. author ) ( 4 ) temperaturea 2 4 6 8 10 14 18 22 densityb 949.87 949.51 949.99 949.75 949.07 948.02 946.69 945.12 viscosityc 1230 999 876 795 689 623 573 surface tension d 53.06 53.02 54.39 55.29 55.89 56.02 56.99 5. conclusion please include a brief summary of the main point s of the paper, do not replicate the abstract as the conclusion, and consider elaborating on the translational importance of the work or suggest applications and extensions. conflict of interest authors need to declare any possible conflict of interest in this section. funding funding body needs to be reported with details of the funded projects. acknowledgment the preferred spelling of the word “acknowledgment” in american english is without an “e” after the “g.” use the singular heading even if you have many acknowledgments. avoid expressions such as “one of us (s.b.a.) would like to thank ...” instead, write “f. s. author thanks ...”. in most cases, sponsor and financial support acknowledgments are placed in the unnumbered footnote on the first page, not here. references the references should be included at the end of the manuscript in a separate section named references without a number. all references must be listed alphabetically by the last name of the first author. ensure that the reference information is complete and accurate, including as necessary and in the following order: first names and initials of all authors; year of publication; title of paper, report, or book chapter (in quotes); title of book or periodical (in italics); volume and issue numbers; name and location of publisher (for books), name and location of publisher or sponsor (for proceedings), or city of publication and inclusive page numbers. below is a list of sample references: article: al-harthy m, begg s, bratvold rb (2007), copulas: a new technique to model dependence in petroleum decision making.journal of petroleum science and engineering 57(1): 195-208. book: jaworski p, durante f, hardle wk, rychlik t (2010), copula theory and its applications.. new york: springer. conference: payne d. b. and stern j. r. (1985), “wavelengthswitched passively coupled single-mode optical network,” in proc. iooc-ecoc, boston, ma, usa, pp. 585–590. technical report: pradier e (2011), copula theory: an application to risk modeling. technical report, grenoble inp-ensimag. thesis and dissertation: author j. k. (year), “title of thesis,” m.s. thesis, abbrev. dept., abbrev. univ., city of univ., abbrev. state. civl180603_current.qxd 1. introduction the use of polymer-modified asphalt (pma) to achieve better asphalt pavement performance has been observed for a long time (al-abdul wahhab and al-amri, 1991; bouldin et al., 1991; sartori et al., 1996; baig and alabdul wahhab, 1998). the properties of pma are depend ent on the polymer characteristics and content, and bitu_____________________________________________ *corresponding author e-mail: hawahab@kfupm.edu.sa men nature as well as the blending process. despite the large number of polymeric products, there are relatively few types that are suitable for bitumen modification. when used as modifiers, polymers should be compatible with bitumen, be capable of being processed by conventional mixing and laying equipment, and be able to maintain their premium properties during mixing storage and application in service. the use of a modifier should also be cost-effective. two classes of polymers are typically used in asphalt modification; elastomers and plastomers. the ring research 1 (2004) 19-28rnal of engineethe jou performance modification of asphalt binders using thermoplastic polymers h. i. al-abdul wahhab* and i. abaker civil engineering department, kfupm, dhahran 31261,saudi arabia received 18 june 2003; accepted 11 october 2003 abstract: there is a need to improve the performance of asphalt binders to minimize stress cracking that occurs at low temperatures and plastic deformation at high temperatures. importation of used asphalt-polymers from abroad, leads to an increase in the total construction cost as compared to the cost if the used polymers were of local origin. the main objective of this research was to modify locally produced asphalt. ten polymers were identified as potential asphalt modifiers based on their physical properties and chemical composition. after preliminary laboratory evaluation for the melting point of these polymers, five polymers were selected for local asphalt modification. in the initial stage, required mixing time was decided based on the relation between shear loss modulus and mixing time. the optimum polymer content was selected based on superpave binder performance grade specifications. the suitability of improvement was verified through the evaluation of permanent deformation and fatigue behavior of laboratory prepared asphalt concrete mixes. the results indicated that the rheological properties of the modified binders improved significantly with sufficient polymer content (3%). the aging properties of the modified binders were found to be dependent on the type of polymer. the fatigue life and resistance to permanent deformation were significantly improved due to enhanced binder rheological properties. thus, local asphalts can be modified using thermoplastic polymers. keywords: polymers, modified asphalt, fatigue cracking, rutting, roads :¢ü∏îà°ùÿÿg.á©øjôÿg iqgô◊g äélqo ‘ ¿~∏dg o~îàdg π«∏≤jh á° øîæÿg iqgô◊g álqo ~æy äé≤≤°ûàdg π«∏≤àd â∏ø°s’g ¢ugƒn ú°ùëàd álé◊g â©ñf √ò¡d »°ù«fôdg ±~¡dg .¥ô£dg aé°ûf’ á«∏µdg áø∏µàdg ~jõj é‡ êqéÿg øe oqƒà°ùj »gh â∏ø°s’g ¢ugƒn ú°ù– ‘ ω~îà°ùj »àdg ogƒÿg øe øfg~∏dg èà©j é¡°ugƒn ≈∏y aéæh égqé«àng ” é«∏fi á©æ°üÿg øfg~∏dg øe ´gƒfg iô°ûy á°sgq~dg â∏ª°t .â∏ø°s’g ¢ugƒn ú°ùëàd é«∏fi á©æ°üÿg øfg~∏dg ωg~îà°sg ƒg ábqƒdg ” á°sgq~dg øe ¤h’g á∏môÿg ‘ .é¡eg~îà°sg øµá øfg~∏dg √òg øe ´gƒfg á°ùªn ¿g qé¡°üf’g álqo álqo ~j~– qéñàneg ô¡xg .á«f髪«µdgh á«féjõ«ødg ago’g á£∏n äéø°ugƒe ≈∏y aéæh ≈∏ãÿg øj~∏dg ᫪c qé«àng ” .¢ü≤dg oé¡lg πeé©eh §∏ÿg øer úh ábó©dg ≈∏y goéªàyg §∏î∏d üƒ∏£ÿg øeõdg ~j~– äé£∏ÿg áehé≤e ¿g èféàædg äô¡xg .ájèàfl äéæ«y ≈∏y o~îàdg h π∏µdg …qéñàng agôlg ᣰsgƒh äé£∏ÿg agog ‘ ø°ùëàdg álqo º««≤j ” .õ«ªàÿg áäÿéh 3 áaé°vg ~æy âæ°ù– ~b álhõ∏dgh ≥a~àdg ¢ugƒn ¿g èféàædg äô¡xg éªc .√~«l iqƒ°üh âæ°ù– ~b o~îàdgh π∏µdg äé≤≤°ûj ~°v øfg~∏déh áæ°ùùg .ájoƒ©°ùdg á«hô©dg áµ∏ªÿéh â∏ø°s’g ¢ugƒn ú°ùëàd ∂hé°s øfg~d ωg~îà°seg á«féµeeg á°sgq~dg äô¡xcg ,áeéy iqƒ°üh .øfg~∏dg øe :á«méàøÿg ägoôøÿg.¥ô£dg ,¿~∏dg o~îàdg ,äé≤≤°ûj ,õ«ªàÿg â∏ø°s’g ,øfg~∏dg relative usefulness of these polymers can vary widely. the concept of modifying asphalt binders and mixtures is not new. many researchers have studied the effects of mineral fillers on the behavior of asphalt (al-abdul wahhab and al-amri, 1991; bouldin et al., 1991; sartori et al., 1996; baig and al-abdul wahhab, 1998). another early method of asphalt modification consisted of mixing two or more asphalt binders of different paving grades or sources. this technique has continued through the years and often delivers a satisfactory product. one major problem, however, is that sometimes the two asphalts are not chemically compatible with each other. compatibility cannot be predicted effectively. this leads to premature pavement distress (anderton, 1991). for the past two decades, significant research has been conducted on pma mixtures. polymers can successfully improve the performance of asphalt pavements at low, moderate and high temperatures by increasing mixture resistance to fatigue cracking, thermal cracking and permanent deformation. it was found that the rheological and engineering properties of pma mixtures largely depend on the polymer type and content. the improvement in fatigue life and resistance to permanent deformation is mainly due to the improvement in the rheological properties of the binders (khattak and baladi, 2001). when an asphalt binder is subjected to cyclic load or stress, the material response to tension and compression consists of three major components: elastic, viscoelastic and plastic. the compression plastic behavior, i.e. permanent strain, is responsible for permanent deformation and consequently results in pavement rutting. in most cases, it has been found that the structure of the pure modified binder is completely different from the structure of the polymer in the asphalt-aggregate mixture (wegan and brule, 1999). when a continuous network of the polymer phase could be detected in the binder, no continuous network of the polymer phase could be detected in the asphalt-aggregate mixture. this is the case with 7% ethylene vinyl acetate (eva) or styrene butadiene styrene (sbs) modified binders. however, for low polymer content (e.g.3%), the same structure of the polymer phase in the binder and in the asphalt-aggregate mixtures was reported. eva seems to have an affinity for the aggregate surface. if the eva phase, swollen by a part of the maltenes, has a higher solubility parameter compared to the asphalt phase, then the eva phase will, at equilibrium, surround the aggregate with its many polar groups. the compatibility and storage stability of polymermodified bitumen (pmb) were shown to be largely dependent on polymer content and were influenced by the characteristics of base bitumens and polymers. polymer modification increased the elastic responses and dynamic modulus of bitumens at intermediate and high temperatures. it also reduced the temperature susceptibility and glass transition, and limiting stiffness temperatures of the bitumens. the degree of the improvement generally increases with polymer content, but varies with bitumen source/grade and polymer type (isacsson and lu, 1999). the strategic highway research program (shrp) conducted a seven-year $50-million asphalt research and developed study on performance-based specifications for asphalt binder that were intended to replace the present penetration and viscosity graded asphalt specifications (harrigan et al., 1994). the program yielded a set of tests to measure asphalt binder properties at high, intermediate and low temperatures, and for fresh and aged conditions. these tests were intended to represent eight to ten years of pavement service. similar work was carried out to adapt shrp binder performance specifications for the arabian gulf countries (gc) (al-abdul wahhab et al., 1996). in this work, the gc region was divided into different temperature zones and the required asphalt performance grade (pg) for each zone was specified. they concluded that the asphalt cement as used locally is only suitable for about 40 percent of the gc area. they found that the required asphalt binder ranged between pg 64-10 and pg 76-10. it was shown that different modification procedures (ranging from air blowing to polymer modification) need to be performed for different asphalt binders. imported polymers are used to improve local asphalt concrete mixes in many countries. off-spec polymers, which are considered industrial wastes, are sold at minimal cost. these polymers, if proved effective for asphalt modification, could lead to a significant reduction in construction cost. the main objective of the present research was to characterize and test the properties of modified bitumens containing thermoplastic polymers. this was achieved through the evaluation of permanent deformation and fatigue behavior of laboratory prepared asphalt mixes. the factors affecting modification were also evaluated. moreover, this study aims at improving asphalt mixes by modifying locally produced asphalts with polymer. 2. experimental set-up and procedure this study consisted of two phases; the first phase dealt with the collection of asphalt from riyadh (ry) and ras tanura (rt) refineries, and polymers from saudi basic industries company (sabic)-the use of local polymers to modify local asphalts was evaluated based on superpave binder performance specifications. in the second phase, polymer-modified asphalts were used in a hot asphalt concrete mixture (hacm) and tested for resistance to fatigue, to permanent deformation and to environmental effects, figure 1. 2.1 materials ten different grade polymers were used in this study. these were placed into six main groups; linear low density polyethylene, high density polyethylene, low density polyethylene, medium density polyethylene, polypropylene and styrenic heavy waste, table 1.the coarse and fine 20 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and aggregates used in the asphalt mixture were obtained from abu-hadriyah, eastern province of saudi arabia. 2.2 identification of potential polymers the first phase of the modification work focused on the identification of potential polymers. the cited literature coupled with manufacturer recommendations and melting temperatures were used to identify the tentative polymer concentrations, blending times and blending temperature. based on the melting temperatures, p5, p6, p7, p8, and p9 were found to be suitable for blending with asphalt. polymers that had a melting temperature of more than 180°c were rejected. polymers were blended with the asphalt from the refineries. the blending temperature was determined based on the melting temperature of the polymers. the selected blending temperature was made equal to the melting temperature plus 5oc in order to ensure good dispersion of polymers. a special blender was assembled, consisting of a high shear blade, a heated oil bath, and a direct current (dc) motor capable of producing up to 2500 revolutions per minute (rpm), figure 2. each polymer type was blended with 800 gm of asphalt binder to produce the polymer-modified asphalt binder. the quantities of each polymer were fixed at 3, 4, and 5% by weight of the binder. the optimum polymer content was selected based on the evaluation of different percentages of the polymerasphalt binder blends through their improvements in the performance grade. the third parameter determined was the storage stability of the blends. two major factors were evaluated in this test: the effect of extended storage at high temperature in the range of 160 to 180°c, and the effect of mechanical agitation of the modified binders. the test procedure was conducted at 160oc for 72 hours. the agitation speed was fixed at 500 rpm. the criterion for determining the storage stability was based on conducting dynamic shear rheometer (dsr) at high temperature of pavement design as recommended by bahia et al. (2001). in addition, rotational viscosity (rv) was added as a new measurement for storage stability to ensure adequate pump-ability of the binder. the ratio of separation under agitation conditions was calculated. the binder was 21 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and figure 1. schematic flow chart for work procedure heater heating oil asphalt binder, 800 gr. electric motor 2500 rpm. the blender head handle figure 2. schematic diagram for blending machine sn commercial name grade code 1 high density polyethylene (hdpe) p1 2 linear low density polyethylene (lldpe) p2 3 polypropylene (pp) p3 4 low density polyethylene (ldpe) p4 5 linear low density polyethylene (lldpe) p5 6 medium density polyethylene (mdpe) p6 7 linear low density polyethylene (lldpe) p7 8 high density polyethy lene (hdpe) p8 9 high density polyethylene (hdpe -ii) p9 10 styrenic heavy waste (shw) p10 table 1. list of selected polymers considered stable if the percent loss in rv and g*/sin δ was less than 20%. 2.3 sample testing optimum blending time was determined by means of the parameter, g*/sin δ at 76°c (g* = shear modulus; δ = phase angle). the g*/sin δ was evaluated during the blending process at an interval of 30 minutes. the maximum blending time was assumed to be two hours. the storage stability of the pma for various polymers was evaluated by simulating the actual storage conditions in the field. the storage temperature was 160°c and stirred at a speed of 200 rpm. the storage time was fixed to be three days. samples were taken from the top and bottom portion of the storage container at an interval of 24 hours. g*/sin δ at 76°c were measured to identify the storage stability and the maximum possible storage time. based on the shrp binder performance specifications, the final performance grade was determined for the ry and rt asphalt modified with selected polymers. the polymer content was fixed at three levels (2%, 3%, and 4%). asphalt concrete mix was designed following marshall mix design in accordance with the moc specifications. the optimum asphalt content obtained for wearing course was used to prepare control mixes and polymer-modified asphalt concrete specimens. the standard marshall specimens of 100 mm × 62.5 mm were prepared for each polymer and control mix. 22 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and r2 = 0.9638 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 50 100 150 blending time (minutes) g */ si n ä, k p a figure 3. determination of optimum blending time of ry with p5-4% asphalt type ry rt g*/sin ä kpa g*/sin ä kpa polymer type time (hr) top bottom loss % top bottom loss % 0 1.635 1.635 0.0 1.425 1.425 0.0 24 1.896 2.016 6.0 1.623 1.803 10.0 48 2.543 2.896 12.2 2.171 2.236 2.9 72 3.177 3.786 16.1 2.526 2.674 5.5 p5-3% rv after 72 h 2436 2596 6.2 2037 2062 1.2 0 1.241 1.241 0.0 1.108 1.108 1.1 24 1.203 1.256 4.2 1.101 1.139 3.4 48 1.288 1.378 6.5 1.212 1.292 6.7 72 1.639 1.811 9.5 1.547 1.635 5.4 p6-3% rv after 72 h 2768 2808 1.4 2237 2250 0.6 0 1.381 1.381 0.0 1.337 1.337 0.0 24 1.750 2.011 14.8 2.942 3.024 2.7 48 4.490 7.626 69.8 3.154 19.229 83.6 72 8.109 13.776 69.9 5.648 21.331 73.5 p7-3% rv after 72 h 3202 5732 79.0 3737 6062 38.4 0 1.219 1.219 0.0 1.081 1.081 0.0 24 1.55 1.568 1.1 1.123 1.155 2.8 48 2.455 2.517 2.5 2.235 2.271 1.6 72 3.678 3.987 7.8 4.477 4.742 5.6 p8-3% rv after 72 h 2635 2799 5.9 2485 2490 0.2 0 1.191 1.191 0.0 1.056 1.056 0.0 24 1 .973 2.011 1.9 1.602 1.861 13.9 48 2.513 2.713 7.4 2.713 2.927 7.3 72 3.149 3.5 10.0 3.164 3.497 9.5 p9-3% rv after 72 h 3210 3350 4.2 2795 2850 1.9 table 2. measured stored stability rv = rotational viscosity the following tests were carried out in order to fully evaluate the polymer-modified mixes and to compare with the control mix: lottman test for moisture susceptibility (aashto t-283-89), resilient modulus test (astm d 4123), fatigue test and permanent deformation (rutting test) 3. results and discussion 3.1 blending time, storage stability and performance grade figure 3 shows the variation of g*/sin δ modified by 4% p5 polymer with blending time. table 2 shows the final results for different types of polymers. the stable blends that passed storage condition were p5, p6, p8, and p9. the stable state was determined based on the difference between the value of g*/sin δ at the top and the bottom of the storage container. it was assumed that the sample was stable if the difference was less than or equal to 20%. from this assumption it was clear that p7 had a higher separation potential when blended with two type of asphalt. therefore, it was rejected. in general, all asphalt polymer blends after storage exhibited an increase in g* (top or bottom) in comparison to the freshly tested blend (unstored). this may be attributed partially to an increase in the molecular association between the polymer and the asphalt binder and partially to the aging of asphalt due to prolonged storing. the rotational viscosity (rv) increased with storage time. this increase in viscosity may limit the storage time so as to reduce the possibility of getting very stiff asphalt. the rotational viscosity should not be more than 3,000 cp. it was found that the rv of p5, p6, and p8 blends were still within the limit after 72-hour storage. the rv of p9, however, exceeded 3000 cp. this indicated that a blend with p9 cannot be stored for more than 48 hours. the results of pg in table 3 showed that polymer types p5, p6, p8 and p9 satisfied the pg binders performance requirements with percentage of 3%, 2%, 2% and 2% respectively, when blended with ry asphalt. they also satisfied the pg binder specification with percentage of 3% when blended with rt asphalt. it was noticeable that an increase in polymer content caused an increase in creep stiffness and a decrease in m-value. it appeared that a 23 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and asphalt type polymer type polymer content (%) ry rt zeropolymer 0 pg 70-10 pg 64-22 2 pg 70-22 pg 70-22 3 pg 76-16 pg 76-10 p5 4 pg 82-10 pg 76-10 2 pg 76-22 pg 70-22 3 pg 76-16 pg 76-16 p6 4 pg 76-10 pg 76-10 2 pg 76-22 pg 70-22 3 pg 76-22 pg 76-10 p8 4 pg 82-16 pg 76-10 2 pg 76-16 pg 70-22 3 pg 76-10 pg 76-10 p9 4 na pg 76-10 table 3. performance grade of different polymers 0 10 20 30 40 50 60 70 control p5-3% p6-3% p8-3% p9-3% m ix type it s p er ce nt ag e l os s, % ry rt figure 4. lottman test results figure 5. comparison of mr (ksi) at 50oc and 25oc for ry 0 200 400 600 800 1000 1200 1400 control p5-3% p6-3% p7-3% p8-3% p9-3% m ix type m r , k si 50, dgrees centigrade 25, dgrees centigrade figure 6. comparison of mr (ksi) at 50oc and 25oc for rt higher polymer content reduced the resistance against low temperature cracking. 3.2 asphalt concrete mix evaluation water sensitivity test (lottman test aashto t-283-89): this test was carried out in order to find the moisture susceptibility (stripping resistance) of modified and control mixes using indirect tensile strength (its). table 4 and figure 4 show the percent loss in the indirect tensile strength after lottman conditioning of the samples. the results obtained indicate that the average percent loss in strength due to moisture damage is effectively reduced in p8 modified asphalt concrete mix. the percent loss in p5, p6, and p9 are less than the percent loss in the control mix but still they are higher than the recommended limit based on superpave specification (20%). however, significant improvement of stripping resistance of local mixes was observed when p8 was used to modify the asphalt. 24 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and sample no. asphalt type polymer type condition (i.t.s.) 1 2 3 average loss% initial 11.10 10.42 9.23 10.25 control final 5.51 4.98 5.26 5.25 48.80 initial 15.51 16.20 15.28 15.66 p5-3% final 9.75 10.35 13.45 11.18 28.60 initial 16.00 15.60 14.20 15.27 p6-3% final 11.20 10.45 9.16 10.27 32.70 initial 13.52 12.65 12.68 12.95 p8-3% final 10.31 11.42 10.80 10.84 16.30 initial 13.56 14.17 13.21 13.65 ry p9-3% final 8.52 7.77 8.32 8.20 39.90 initial 10.00 9.12 8.37 9.16 control final 3.23 4.05 4.77 4.02 56.17 initial 18.51 18.01 15.28 17.27 p5-3% final 8.72 9.35 13.45 10.51 39.15 initial 14.99 14.76 13.16 14.30 p6-3% final 8.26 10.45 8.19 8.97 37.31 initial 13.28 11.65 12.31 12.41 p8-3% final 8.31 11.42 7.87 9.20 25.89 initial 13.56 15.17 14.16 14.30 rt p9-3% final 5.81 4.87 4.28 4.99 65.12 table 4. i.t.s. test results ry rt temperature temperature mix type 50°c 25°c 50°c 25°c control 85.0 547.3 76.5 497.3 p5 – 3% 171.9 576.4 335.2 850.1 p6 – 3% 293.6 654.2 258.5 780.8 p8 – 3% 369.3 799.8 344.6 1005.3 p9 – 3% 305.6 671.7 305.6 447.5 table 5. resilient modulus (ksi) asphalt rt ry regression factors @ 25 oc regression factors @ 50 oc regression factors @ 25 oc regression factors @ 50 oc mix type i s i s i s i s control 9946.5 -0.524 7924.9 -0.626 806.3 -0.261 920.8 -0.247 p5 6890.9 -0.464 3623.5 -0.411 4032.8 -0.362 1566.3 -0.286 p6 7961.3 -0.462 3514.0 -0.398 36590.0 -0.549 9489.1 -0.451 p8 4384.3 -0.365 1801.8 -0.295 7012.5 -0.429 5026.0 -0.403 p9 9812.7 -0.462 4079.9 -0.393 562.5 -0.202 920.7 -0.286 table 6. regression factors for fatigue test 25 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and 1000 10000 100000 1000000 no. of load repetitions, n 10 100 1000 in iti al t en si le s tr ai n, m ic ro s tr ai n r y c o n t r o l r y p 5 3 % r y p 6 3 % r y p 8 3 % r y p 9 3 % r t c o n t r o l r t p 5 3 % r t p 6 3 % r t p 8 3 % r t p 9 3 % figure 7. fatigue curves for modified wearing course mixes at 25oc 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 n o . o f l o a d r e p e titio n s, n 1 0 1 0 0 1 0 0 0 in iti al t en si le s tra in , m ic ro s tr ai n r y c o n t r o l r y p 5 3 % r y p 6 3 % r y p 8 3 % r y p 9 3 % r t c o n t r o l r t p 5 3 % r t p 6 3 % r t p 8 3 % r t p 9 3 % figure 8. fatigue curves for modified wearing course mixes at 50oc 26 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and 100 1000 10000 100000 1000000 no. of load repetitions, n 100 1000 10000 100000 p er m an en t s tra in , m ic ro s tra in r y c o n t r o l r y p 5 3 % r y p 6 3 % r y p 8 3 % r y p 9 3 % r t c o n t r o l r t p 5 3 % r t p 6 3 % r t p 8 3 % r t p 9 3 % figure 9. rutting curves at 100-initial micro-strain @ 25oc 100 1000 10000 100000 1000000 n o. of load r epetitions, n 100 1000 10000 100000 1000000 10000000 p er m an en t s tra in , m ic ro s tra in r y c o n t r o l r y p 5 3 % r y p 6 3 % r y p 8 3 % r y p 9 3 % r t c o n t r o l r t p 5 3 % r t p 6 3 % r t p 8 3 % r t p 9 3 % figure 10. rutting curves at 100-initial miro-strain @ 50oc generally, the its results of both ry and rt asphalt seem to be the same except when polymer type p9 is used. the percent loss in its of p9-ry is almost 50% of the percent loss of p9-rt. this indicates that p9 is somehow more effective when used with the ry asphalt. resilient modulus (mr test astm d 4123): resilience is a measure of pavement response in terms of repeated stresses and corresponding strains. polymer-modified mixes showed an improved diametral resilient modulus compared to the control mix. the average resilient modulus at 25°c was found to be 547 ksi (3782 mpa) for the control ry mix. this value increased to 800 ksi for the p8 mix. the average resilient modulus at 50°c increased from 85 ksi (587 mpa) for the control ry mix to 369 ksi (2552 mpa) for p8 (table 5, figures 5 and 6). although control rt mix had a low mr compared to the control ry mix, its blends with p5, p6, and p8 showed higher mr values compared to ry blends. this phenomenon may be linked to the way polymers are dispersed in the asphalt and the cross-linking between polymer molecules and asphalt molecules. fatigue performance: diametral fatigue test results for wearing course mixes of the four polymer types (p5, p6, p8 and p9 in addition to the control mix) at testing temperatures of 25°c and 50°c are shown in figures 7 and 8 for each polymer. nine samples were tested for fatigue (three at each strain level). linear regression analyses were done at each strain level. the results showed a normal linear relationship between the logarithm of applied initial tensile strain and the logarithm of fatigue life (number of applied load repetitions until failure). the fatigue data were analyzed by regression analysis to determine the fatigue relationship parameters in the following form: [1] where: εt = initial tensile strain, nf = number of load repetitions to failure; i = anti-log of the intercept of the logarithmic relationship; and s = slope of the logarithmic relationship. regression parameters for equation (1) are shown in table 6. the results obtained show significant improvement in fatigue life of the control specimen when modified with polymer, particularly at low strain and high temperature levels. p8 showed higher fatigue life (about 50 times higher than control mix) followed by p9, p6, and p5, respectively at both high and low temperatures. permanent deformation: permanent deformation was simultaneously recorded during fatigue tests at 25°c and 50°c, figures 9 and 10. the results indicated that a straight-line relationship exists between the logarithm of number of load repetitions and the logarithm of permanent strain. the permanent deformation properties were determined using : [2] εp = accumulated permanent strain; n = number of load repetitions; i = anti-log of the intercept of the logarithmic relationship; and s = slope of the straight line of the log-log plot. parameters i and s were obtained using permanent deformation experimental data in a regression analysis. results indicated that polymer modification reduced the permanent deformation potential of asphalt mixes when compared to the control mix. the improved rutting resistance was clear at low and high temperatures. p5 showed similar behavior to the control mix at 25°c. the behavior became more elastic at 50oc, p8 showed the best rutting performance followed by p9, p6, and p5, respectively. 4. conclusions the tested polymers were very effective in improving the rheological properties of the neat asphalt binders to meet and satisfy the performance requirements. mixtures with mdpe, hdpe and hdpe (ii) showed a significant increase in resilient modulus at both high and low temperatures compared to the control mixture. modified asphalt mixtures showed a significant variation in their water susceptibility. some mixtures were more sensitive than others. the average percentage loss in strength due to mois ture damage was reduced effectively by hdpe. the results showed significant improvement in fatigue behavior of all the polymer type mixes compared with the control mix. mixes with hdpe showed higher fatigue life (about 50 times higher than the control mix) followed by hdpe-ii, mdpe and lldpe, respectively, at high and low temperatures for both types of asphalt. the results indicated that polymer modification resulted in a 20% reduction in permanent deformation compared to the control mix. the improved rutting resistance was clear at both low and high temperatures. mixing with lldpe showed similar behavior to the control mix at 25°c, though the behavior was more elastic at 50°c. mixing with hdpe showed the best rutting performance followed by hdpeii, mdpe, and lldpe, respectively. acknowledgements the authors would like to acknowledge the support provided by king fahd university of petroleum & minerals (kfupm) and saudi basic industries company (sabic). 27 ring research 1 (2004) 19-28rnal of engineethe jour /abakeabdul wahhab and references a1-abdul wahhab, h., and al-amri, g., 1991, “laboratory evaluation of reclaimed rubber asphaltic concrete mixes,” asce journal of materials, 3 (3), pp. 189-203. a1-abdul wahhab, h., ali, f., asi, i., and al-dhubeeb, i., 1996, “adaptation of shrp performance based asphalt specifications to the gulf countries,” final report, king abdulaziz city for science and technology (kacst), riyadh. anderton, g.l., 1991, “modified asphalt binders for military airfield pavement,” proceedings, aapt, vol. 60. bahia, h., hanson, d.i., zeng, m., zhai, h., khatri, m., and anderson, r. (2001), “characterization of modified asphalt binder in superpave mix design,” nchrp report 459, national academy press, washington, d.c. baig, m..,, and al-abdul wahhab, h., 1998, “mechanistic evaluation of hedmanite and lime modified asphalt concrete mixes,” asce journal of materials, 10 (3), pp. 153-160. bouldin, m.g. et al., 1991, “rheology and microstructure of polymer/asphalt blends,” journal of rubber chemistry and technology, 67 (4). harrigan, e.t. et al. (editor), 1994, “the superpave mix design system manual of specifications, test methods, and practices,” shrp-a-379, national research council, washington, d.c. isacsson, u.,. and lu, x., 1999, “modified asphalt pavement materials the european experience,” proceedings, aapt, vol. 68 p. 35. khattak, m.j., and baladi, g.y., 2001, “paving asphalt polymer blends: relationships between composition, structure and properties,” trb 10, no. 01-2813, 80th transportation research board annual meeting, washington, d.c. sartori, f., bonemazzi, f., bbraga, v., and corrieri, r., 1996, “characteristic of polymers and polymer modified binders,” trr no. 1535, transportation research record , washington, d.c. wegan, v., and brule, b., 1999, “polymer modified asphalt cements used in the road construction industry,” proceedings, aapt, vol. 68, p. 64. 28 abdul wahhab and abaker / the journal of engineering research (2004) 19-28 the journal of engineering research (tjer), vol. 17, no. 1, (2020) 47-58 *corresponding author’s e-mail: hayomchm@yahoo.com doi:10.24200/tjer.vol17iss1pp47-58  probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency of synthesized carbon-based nanomaterials haiyam mohammed alayan1,*, mustafa mohammed aljumaily2, mohammed abdulhakim alsaadi3, and mohd ali hashim4 1 chemical engineering department, university of technology, baghdad, iraq 2 department of civil engineering, al-maaref university college, ramadi, iraq 3 national chair of materials sciences and metallurgy, university of nizwa, nizwa, sultanate of oman 4 chemical engineering department, university of malaya, kuala lumpur, malaysia abstract: the present work investigates the effect of the type of carbon precursor on the adsorptive proficiency of as-prepared carbon nanomaterials (cnms) for the removal of methylene blue dye (mb) from aqueous media. a comparison study was applied to assess the growth of cnms from the decomposition of methane (cnmy1) and acetylene (cnmy2) using response surface methodology with central composite design (rsm/ccd). the produced nanomaterials were characterized using fesem, edx, tem, bet surface area, raman, tga, ftir, and zeta potential. the as-prepared adsorbent displayed different morphologies and under the experimental conditions, 10 mg of cnmy1 and cnmy2 was responsible for 97.7 % and 96.80% removal of dye. the maximum adsorptive uptake predicted by langmuir isotherm was about 250 and 174 mg/g for cnmy1 and cnmy2, respectively. the as-synthesized carbon nanomaterial in this study could be explored as a great potential candidate for dye-bearing wastewater treatment. keywords: adsorption; carbon nanomaterials; chemical vapor deposition; dyes; response surface methodology. السالئف الهيدروكربونية الغازية على الكفاءة اإلدمصاصية للمواد تأثير منالتحقق النانوية المستخلصة من الكربون محمد علي هاشم و هيام محمد عليان *، مصطفى محمد الجميلي، محمد عبدالحكيم السعدي من نوعين مختلفين من الغازات الهيدركربونية في هذه الورقة البحثية، تم تصنيع انابيب الكربون النانوية :الملخص عامل مساعد ثابت. وقد استخدم البرنامج الحاسوبي كيميائي تحليلي تبخيري و واالستيلين) بواسطة مفاعل (الميثان design experiment قيمة لكل واحدة من عوامل التفاعل أفضلللحصول على المختبرية مخطط التجارب لتصميم كمية أفضلخلط الغازات المستعملة على ان تتحقق في المنتوج بةوالتي تتضمن درجة حرارة التفاعل، زمن التفاعل، ونس ، feesmواص المادة الناتجة باستخدام خافضل قدرة على ازالة صبغة الميثيل االزرق من الماء. لقد فحصت نمو مع edx ، tem ،betواجريت فحوصات االشعة تحت الحمراء ب ،ftir والسلوك الحراري باستخدام ،tga ، ملغرام من المادة الممتزة الناتجة من تفكك الميثان واالستلين لها 10 ان . بينت النتائجzeta potentialوجهد الزيتا خير ما يمثل langmuirجد ان االيزوثرم حسب % على الترتيب. وو 96.8و 79.7القدرة على ازالة الملوث بنسبة ل الترتيب. بينت ملغم⧵غم للمادة المنتجة من تفكك الميثان واالستلين ع 174و 250النظام وبامتصاصية قصوى تصل الى ألغراض واعدة الستخدامهامادة يجعلهاالدراسة ان المادة اانانوية التي تم تصنيعها قد ابدت أداءا متميزا في االمتزاز مما االمتزاز المتنوعة. .االصباغ الملوثة ؛تبخيري ذو عامل مساعد مفاعل ؛انابيب كاربونية نانوية ؛االمتزاز :حيةالمفتا تلماالك probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  48  1. introduction carbon nanomaterials (cnms) have been intensively investigated for environmental applications due to their outstanding properties. they have enlightened different prospects as attractive sorbents for the removal of undesirable pollutants from water. their high surface area, hybridized sp2 carbon cluster and optimal binding efficiency enable strong adsorption capacity for organic contaminants (hu et al. 2019; sun et al. 2016). among the various techniques being used to synthesize cnms, the chemical vapor deposition (cvd) is considered as the most promising technique, due to its versatility and suitability for large scale production. in cvd process and under tunable growth conditions, a statistical study should be conducted to achieve an adequate evaluation of the optimal growth conditions for cnms, such as the utilized catalyst, carbon precursor type and flow rate, growth temperature and reaction time (alayan et al. 2019; gromov et al. 2016). the agglomeration and ineffective recovery of cnms are considered significant limitations which will confine their application in environmental remediation. thus, it is essential to develop innovative adsorbents that can overcome the current challenges. possible techniques proposed to resolve these drawbacks include centrifugation and attachment of magnetic nanoparticles. nevertheless, both options will crucially add cost and complexity to the adsorption process. accordingly, it is believed that growing carbon nanostructures on micro-scaled carbon support will enlighten the production of promising alternatives to the conventional adsorbents (meshot et al. 2017). industrial effluents accept worrying quantities of colored pigments which impose harmful effects on the human health and surrounding ecosystems. methylene blue (mb) contaminant has a potential danger of bioaccumulation due to its stability and low biodegradability, hence, efficient removal techniques of practical significance is imperative to remove it from industrial wastewaters before it is released to the environment. of all treatment methods, adsorption based process has been widely implemented since it is simple, bear large volume of wastewater and devoid of utilizing organic solvents (elsagh et al. 2017; siddiqui et al. 2019). in this work, we aim to assess the effect of hydrocarbon source being supplied in cvd reactor on the adsorption capacity of the as-prepared cnms. unlike previous studies that usually employ metallic substrates in the growth process of cnms we fabricated our carbon nanomaterial by incorporating powder activated carbon (pac) as analogues of noncarbon supports to boost the synergetic effect of hybridized cnms to enhance their adsorptive affinity (alayan et al. 2017). in this context, pac was employed as a catalyst substrate to motivate the growth of cnms from methane and acetylene decomposition. response surface methodology with central composite design (rsm-ccd) technique was applied to optimize the preparation parameters followed by characterizing the as-prepared nanomaterial using edx, bet, fesem, tem, raman spectroscopy, tga, ftir, zeta potential and goniometer. later, the adsorption performance of the fabricated hierarchical cnms was investigated for the removal of mb dye from water samples. 2. materials and methods 2.1 materials methylene blue, powder activated carbon, nickel (ii) nitrate hexahydrate ni (no3)2·6h2o, were purchased from sigma-aldrich. methane ch4, acetylene c2h2, hydrogen h2, nitrogen (n2), were supplied in analytical grade and used without further purification. 2.2 synthesis of cnm/pac using cvd a mixture of 2 g of dried pac and 5 ml catalyst solution of nickel (ii) nitrate hexahydrate (1.0 wt %) was sonicated at 60oc for 1 h. then the niimpregnated pac was dried at 100oc overnight. the dried and grinded ni/pac samples were exposed to two sequenced stages of thermal treatment in cvd reactor including calcination and reduction at 350 and 600 ºc for 2 and 1 h under 200 ml/min of n2 and h2. the growth of hierarchal cnm/pac was achieved by shuffling a ceramic boat of the ni/pac into the reaction tube of cvd reactor. the growth process was conducted with the following conditions: ch4 and h2 were introduced to the cvd reactor at a ratio range of 1-4 within a temperature range of 750-1000 ºc for 20 -60 min. then the reactor was cooled and the produced cnms were collected. the second hierarchal cnm/pac was synthesized from the decomposition of c2h2 following the same experimental procedure. however, in this case the growth reaction was conducted within the temperature range 550-750oc. the yield of cnms (cnmy) was calculated using equation (1): yield (1) where m2 and m1 are the weight of the sample after and before reaction, respectively. 2.3 experimental design and optimization a design of experiment using ccd was employed to investigate how to obtain optimal growth parameters namely, hydrothermal temperature, reaction time and input gas ratio. the suggested regression model and the relevant statistical parameters were evaluated by the analysis of variance haiyam alayan, mustafa aljumaily, mohammed alsaadi, and mohd hashim 49  (anova). additionally, for optimization sake, maximizing the growth of cnms was our response criterion. accordingly, tridimensional plots were constructed to demonstrate the interaction between the underworking growth parameters. 2.4 characterization the pac, ni/pac, and cnms were dispersed onto copper grids to examine their surface morphologies using sem equipped with energy dispersive x-ray spectroscopy and transmission electron microscopy. raman spectra and thermogravimetric analysis of cnms were developed. the bet method was carried out to determine the surface area of the prepared nanomaterials. the ftir spectra over the range 4000 to 400 cm-1 was used to analyze the surface changes. surface charge analysis was investigated using the zeta potential measurements of a 0.01 wt % suspensions by zetasizer. finally, to debate the hydrophobicity issue, the goniometer was used to assess the surface hydrophobicity in terms of its contact angle. 2.5 measurements of dye performance the removal of mb on the as-prepared cnms produced from the decomposition of ch4 and c2h2 were carried in a batch mode as follow: a fixed amount of each adsorbent (10 mg) was agitated for 30 min at room temperature into the dye solution (50 ml) at initial concentration of 50 mg/l and ph 11. at the end of the adsorption experiments, the dye solution was centrifuged and the persisted concentration was determined using uv-vis spectroscopic at maximum wavelength (665 nm). the removal efficiency and the adsorbent uptake (mg/g) were calculated by the following relationships: removal efficiency (%) = 100 (2) uptake = v (3) where co and ce (mg /l) are the initial and equilibrium dye concentrations, respectively, v (l) is the volume the dye solution and w (g) is the amount of the adsorbent. the adsorption efficiency can be demonstrated via the analysis of experimental data to estimate the best fitted kinetic models and adsorption isotherms (chen and bai 2013). kinetics of adsorption of mb on as-prepared cnm were investigated using both the pseudo-first-order and pseudo-second-order models. in addition, langmuir and freundlich isotherms were applied to study the sorbent-sorbate interactions. 3. results and discussion 3.1 model establishment and statistical analysis the growth temperature (a), growth time (b) and input gases flow ratio; (h2/ch4) or (h2/c2h2) (c) were optimized for cnm/pac synthesis using 2-level ccd with one central point by the rsm/ccd method. the doe software suggested conditions covered the growth temperature of 750 -1000 ºc, 550 750 ºc for ch4 and c2h2 decomposition, respectively. also, for both hydrocarbon precursor, the reaction time and gases ratio were varied under the range of 20 60 min, 1.0 4.0. the optimizing criterions were devoted to maximizing the yield percentage of cnmy1 and cnmy2 for ch4 and c2h2, respectively, while the independent parameters were kept in prescribed ranges.  the anova summarized in table 1 shows that the major determinants for cnmy1 obtained from ch4 decomposition are the growth temperature (a), the interaction of the growth temperature and the gas ratio (ac) and the second-order effect of growth temperature (b2). the high model f-value and the low pvalues < 0.05 of these confirm their significance effect. the high correlation coefficient r2 value for cnmy1 and cnmy2 (> 0.98) suggest that the table 1. anova results for (cnmy1). *source sum of square df mean square f-value p-value model 2.71 10 0.27 210.54 0.0005 a 1.18 1 1.18 914.87 < 0.0001 b 0.074 1 0.074 57.22 0.0048 c 0.068 1 0.068 52.94 0.0054 ab 0.11 1 0.11 84.02 0.0027 ac 0.23 1 0.23 178.35 0.0009 bc 0.0087 1 0.0087 6.76 0.0804 a2 0.009 1 0.009 69.48 < 0.0036 b2 0.043 1 0.043 33.05 0.0105 c2 0.0011 1 0.0011 0.89 0.4158 a2c 0.048 1 0.048 37.40 0.0088 rsquared 0.998 std.dev. 0.0359 adj. r-squared 0.994 c.v. % 1.35 pred. r-squared 0.938 precision 46.61 *a (ºc): temperature, b (min): time, c (feed gas flow ratio): h2/ch4 table 2. anova results for cnmy2. source sum of square df mean squar e f-value p-value model 0.692 8 0.0865 101 0.00986 a 0.0176 1 0.0176 20.4 0.0456 b 0.0522 1 0.0522 60.8 0.0160 c 0.0334 1 0.0334 38.9 0.0247 ab 0.0654 1 0.0654 76.1 0.0129 ac 0.0921 1 0.0921 107.0 0.00920 bc 0.231 1 0.231 269.0. 0.00370 a2 0.0197 1 0.0197 23.0 0.0409 a2 b 0.0428 1 0.0428 49.9 0.0195 rsquared 0.998 std. dev.0.0293 adj. r-squared 0.988 c.v. % 0.604 pred. r-squared 0.929 precision 26.4 *a (ºc): temperature, b (min): time, c (feed gas flow ratio): h2/c2h2 probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  50  figure 1. predicted vs. actual values; (a) cnmy1, (b) cnmy2. table 3. the optimum conditions suggested by doe for cnmy1. no. a b c cnmy1% desirability 1 950 60.0 1.00 36.7 0.989 ✓ 2 950 59.8 1.00 36.5 0.986 3 950 60.0 1.00 36.5 0.986 4 950 60.0 1.03 36.4 0.984 5 950 59.7 1.00 36.3 0.982 6 950 60.0 1.13 35.7 0.970 7 950 60.0 1.22 35.1 0.959 8 950 60.0 1.00 33.4 0.927 9 950 60.0 1.52 33.0 0.917 10 950 20.0 1.00 23.0 0.674 table 4. the optimum conditions suggested by doe for cnmy2. no. a b c cnmy2 % desirability 1 560 38.0 1.00 154.47 0.919 ✓ 2 561 38.0 1.00 152.9 0.916 3 560 38.5 1.00 154.47 0.914 4 560 38.0 1.12 152.9 0.913 5 561 38.4 1.00 154.47 0.912 obtained model offers a successful correlation between the growth variables and good estimation of the responses which is confirmed by the plots of their predicted values against the actual values presented in fig. 1a and 1b. furthermore, the statistical results in table 2 support the significance of the suggested model for cnmy2 from c2h2 pyrolysis for the linear terms (a), (b), and (c). the interaction effect of (ac) and (bc) displayed p-values of 0.009 and 0.0036, respectively indicating their strong impact over the studied range as well. the regression equation for cnmy1 and cnmy2 in terms of their actual values are given by equations (4) and (5): 𝐿𝑛 𝐶𝑁𝑀𝑌1 11.4 0.0304 𝐴 0.0789 𝐵 11.0 𝐶 5.79 10 𝐴𝐵 0.0251 𝐴𝐶 0.0011 𝐵𝐶 1.45 10 𝐴 0.00046 𝐵 0.0128 𝐶 1.41 10 𝐴 𝐶 4 𝐿𝑛 𝐶𝑁𝑀𝑌2 26.0 0.0637𝐴 0.385 𝐵 0.28 0.00121 𝐴𝐵 0.000715 𝐴𝐶 0.00566𝐵𝐶 4.59 10 𝐴 8.96 10 𝐴 𝐵 (5) 3.1.1 response surface methodology the effect of the substantial factors and the feature of the obtained response surface in the experiment were examined by using rsm/ccd. the interaction effects of the process parameters on cnmy1 and cnmy2 based on the empirical model were illustrated by rsm plots. fig. 2 displayed an outstanding increase in cnmy1 with increasing the reaction temperature from 750 to 950°c, implying the favor of high reaction temperature for the growth of cnms fig. 2a. there was no significant improvement for cnmy1with the increase in (h2/ch4), since high concentration of h2 will motivate the backward reaction thus reducing the yield of cnmy1 fig. 2b. also, the cnmy1 improved with increasing growth time and then dropped as the reaction proceeds fig. 2c. this decrease in the yield of the cnms at long reaction period is attributed to the catalyst deactivation due to the formation of amorphous carbon. in fig. 3a, increasing reaction temperature from 550 to 650°c has a distinguished improvement on the throughput of cnmy2 at fixed h2/c2h2 (1.0) which is in agreement with the role of pac substrate in minimizing the deactivation of impregnated catalyst even at high growth temperatures (tao and crozier 2016). however, an increase in cnmy2 for low reaction temperatures and h2/c2h2 at fixed reaction time is observed in fig. 3b, and the high temperature accompanying with high gas flow ratio was unfavorable for cnmy2. this might be since the high h2 content will accelerate the reduction of metal oxides which in turn encourages the deposition of pyrolytic carbons on the catalytic sites. moreover, it can be noticed from fig. 3c that at short reaction time the yield of cnnmy2 was high and almost the same at h2/c2h2 of 1.0 and 4.0 which indicates the insignificance effect of the gas ratio. this observation motivates the production of cnms at small reaction periods to get high yield and prevent catalyst deactivation as well (shen et al. 2016). a  b haiyam alayan, mustafa aljumaily, mohammed alsaadi, and mohd hashim 51  figure 2. response surface illustration of the interaction effects on cnmy1: (a) growth temperature and time, (b) growth temperature and gas ratio and (c) time and gas ratio. figure 3. response surface illustration of interaction effects on cnmy2: (a) growth temperature and time, (b) growth temperature and gas ratio and (c) time and gas ratio. probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  52 3.1.2 optimization study numerical optimization was applied to determine the optimal growth conditions for ch4 and c2h2 decomposition. thus, the predefined criterions for the optimal working conditions were applied by keeping the synthesis parameters within the prescribed upper and lower limits of the growth conditions in order to achieve the maximum yield and the highest desirability value which will confirm the adequacy of the predicted models. table 3 indicated that the predicted optimum conditions for cnmy1 were 950°c, 60 min, and 1.0 corresponding to growth temperature, growth time and h2/ch4, respectively leading to yield ~ 37.0 %. a yield of ~ 155 % for cnmy2 was obtained at the optimal conditions of a 560oc, 38 min and 1.0 h2/c2h2 (table 4). 3.2 characterization 3.2.1 morphology the fesem and tem images of the nickel-doped activated carbon (ni/pac) show a successful imbedding for the catalyst particles in the pores of pac substrate (fig. 4a and 4b). the elemental analysis attained from edx (fig. 4c) shows that the loaded substrate is comprised of 88.58 wt% carbon, 7.48 wt% oxygen, wt% 3.26 wt% nickel and 0.67 wt% silicon. the growth of various types of cnms figure 4. (a) fesem, (b) tem and (c) edx images for ni/pac before the growth reaction. b a figure 5. fesem images of (a) cnmy1 and (b) cnmy2 obtained at optimal conditions. a b c haiyam alayan, mustafa aljumaily, mohammed alsaadi, and mohd hashim 53  from this low catalyst amount validates the catalyzing role of pac and its ability to provide good media for catalyst dispersion (li et al. 2014). fig. 5 shows the morphologies of hierarchical nanomaterial shaped from the methane (cnmy1ch4) and acetylene (cnmy1-c2h2) decomposition at the optimal growth conditions. the hybrid cnmy1ch4 displayed in fig. 5a has groove-like form and resemble the pac porous structures. cnmy1-ch4 hybrid is comprised from short interposed carbon nanotubes (cnts) which are randomly oriented in the substrate matrix and embedded on the solid walls of the pores the activates carbon substrate. the hybrid cnmy1-c2h2 in fig. 5b and 5c, which was obtained at the optimal synthesis conditions, show bushy and graphitized tubular structures of cnts with 10–40 nm, and an intercalated catalytic nickel nanoparticle (cnp) which was separated from the substrate and confined at the tip of the grown cnt. this observation suggests a tip growth mechanism for the cnt which triggered by the precipitation of the generated carbon fragments on the ni surface followed by the growth of the tube and terminated finally due to the catalyst deactivation (jeong et al. 2016). 3.2.2 tga analysis the tga analysis in fig. 7 assesses the quality and thermal stability of cnmy1 and cnmy2 obtained from ch4 and c2h2. the tga profile can be highlighted by the initial gradual loss due to moisture evaporation of 2.0 % and 5.0 % followed by constant drops to the onset combustion temperature at 507 and 455 ºc for cnmy1 and cnmy2, respectively. the third region was featured by a steep drop region and ended with the oxidization section of the remaining amorphous carbon at temperature higher than 500 ºc. additionally, a well-graphitized structure for cnmy1 is obtained and confirmed by its high onset temperature. the tga analysis indicates that the produced carbon nanostructures from both hydrocarbon precursor sustain a temperature as high as 800 ºc (zhou et al. 2014). 3.2.3 bet surface area the surface area of the pac, ni/pac, cnmy1 and cnmy2 was determined by nitrogen adsorption– desorption measurements. the results illustrated in table 5 show a reduction in the surface area of ni/pac due to the catalyst impregnation. furthermore, an enhancement in the surface area of the obtained hybrids cnmy1 and cnmy2 because the added structure on the substrate. the total pore volume of cnmy1 and cnmy2 reached 0.29 cm3/g and 0.44, respectively. this logical result can be ascribed to the development of new micropores resulting from the growth of different nanostructures on the substrate which will be beneficial for improving the adsorbent sorption capacity (wang et al. 2017). 3.2.4 raman analysis raman analysis displayed in fig. 7 for the cnmy1 and cnmy2 detected the main d and g characteristic band. the d peak at around 1316 cm−1 suggests the presence of disordered graphite structure while the c-c stretching vibration is characterized by the g band at around 1594 cm−1. the high intensity of d band confirms the defective nature of the obtained structure due to the thermal treatment step. also, the grown hybrids involved multiwall structure confirmed by the disappearance of the radial breathing mode (rbm) at wavelength below 400 cm-1(ahmad et al. 2018). the organizational degree of the obtained hybrids can be represented by id/ig and it is noticed that synthesized nanocarbon structure ratio was 0.9 and 0.82 from the decomposition of ch4 and c2h2, respectively. this indicates that the decomposition of c2h2 enhanced the growth of highly graphitized nanostructure which was also observed in the tga analysis. also, the cnms synthesized from acetylene pyrolysis at750°c demonstrated id/ig ratio of 0.74 which suggests a lesser graphitic structure than that obtained at 560°c. 3.2.5 ftir analysis the functional group and surface chemistry of cnmy1 and cnmy2 synthesized at the optimal growth conditions were presented in the ftir spectroscopy in fig. 8. absorption bands at 3450– 3500 cm-1 assigned to the –oh group. the ascribed peaks from the ni catalyst carbonyl were found at 1979 and 2138 cm−1 (chavan et al. 2012), whereas the peak emerged at ~ 2350 cm-1 is attributed to the presence of aromatic sp2 c–h stretching vibration. the aromatic rings and c=c stretch were appeared at 1447 and 1563 cm−1, respectively. it is observed that the cnmy2 hybrids has fewer functional groups than that produced from ch4 decomposition, accordingly higher hydrophobicity is expected for cnmy2 (ezzeddine et al. 2016) 3.2.6 zeta potential and contact angle the adsorption mechanism is greatly affected by the surface charge and the hydrophobicity of the adsorbent (aljumaily et al. 2018). ni/pac is not a hydrophobic material and its cast film a contact angle of 65° as presented in table 6. the cnmy1 has lower zeta potential than that obtained for cnmy2.the surface positive charge and basicity of cnmy1 could be due to the oxygen-free carbon sites (πelectron-rich regions) located in the basal planes. furthermore, the cnmy1 hybrids show enhanced hydrophobicity associated with high contact angle of 138° but lower than that obtained for cnmy2. the super hydrophobic cnms synthesized from acetylene decomposition demonstrated high contact angle of 160° with high value of zeta potential due to the mutual repulsion of cnms surface charges. probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  54 figure 6. tga curves for cnmy1 and cnmy2. table 5. bet results of pac, ni/pac, cnmy1 and cnmy2. property pac ni/pac cnm y1 cnm y2 bet (m2/g) 101.1 97.2 164.6 333.83 total pore volume (cm3/g) 0.09 0.07 0.29 0.44 average pore diameter (ºa) 34.89 21.29 96.19 65.09 figure 7. raman spectrum of cnmy1 and cnmy2. figure 8. ftir spectrums for free mb, cnmy1 and cnmy2 after adsorption. 3.2.7 kinetics and isotherm studies two models were applied to analyze the kinetic data obtained from the batch experiments of mb adsorption on cnmy1 and cnmy2: pseudo-first order, and pseudo-second models (fig. 9). the relevant parameters derived from the corresponding equations are provided in table 7, imply that the adsorption process has not entirely followed the pseudo first-order rate expression for both adsorbents (fig. 9a and 9b). in contrast, the applicability of the pseudo-second-order model better to describe the sorption kinetics of mb dye is clearly observed in fig. 9c and 9d confirmed by the high correlation coefficient of the linear plots. this observation suggests the possibility of chemisorption adsorption mechanism through exchange or sharing electrons between the adsorbent and adsorbate (subramani and thinakaran 2017). for the isotherm study, the favorability of the langmuir isotherm can be conveniently represented by the equilibrium parameter (rl); (0<rl <1) which is defined by equation (6). 𝐑𝐋 𝟏 𝟏 𝐊𝐋 𝑪𝒐 (6) the values summarized in table 8 and the linear plots presented in fig. 10 agreed well with the langmuir isotherm (model 1) for cnmy1 and cnmy2 signifying monolayer adsorption with heterogeneous surface binding, however, sorption capacity values of cnmy2 was conveniently described by freundlich model (model 2) as well. this kind of duality has been stated in many publications (wang et al. 2014). comparison of the maximum adsorption uptake of mb on several adsorbents are presented in table 9. the as-prepared hybrid carbon nanomaterials in the present work have demonstrated noteworthy adsorption uptake of 250 and 174 mg/g for cnmy1 and cnmy2, respectively. table 6. zeta potential results for ni/pac and the obtained carbon structures. sample zeta potential (mv) contact angle (º) ni/pac +1.67 65 cnmy1 (950 ºc) +9.4 138 cnmy2 (560 ºc) 34.3 160 haiyam alayan, mustafa aljumaily, mohammed alsaadi, and mohd hashim 55 table 7. experimental values of kinetics model’s constants. model equation parameters values cnmy1 cnmy2 pseudo-first-order ln 𝑞 q 𝑙n𝑞 𝐾 t r2 k1 qe 0.83 0.89 0.02 0.03 38.9 27.0 pseudo-second-order 𝑡 𝑞 1 𝐾 𝑞 1 𝑞 𝑡 r2 k1 qe 0.999 0.99 0.002 0.0015 122.7 145.0 table 8. linearized equations of studied isotherm models. model equation parameters values cnmy1 cnmy2 langmuir 𝐶 𝑞 1 𝐾 𝑞 1 𝑞 𝐶 qm 250.0 174.0 kl 0.65 1.5 r2 0.98 0.99 rl 0.03 0.01 freundlich 𝑙𝑛𝑞 𝑙𝑛𝐾 1 𝑛 𝑙𝑛𝐶 r2 0.86 kf 85.0 n 2.8 0.93 83.7 2.6 table 9. maximum adsorption uptake (qm) comparison with previous studies. adsorbent qm (mg/g) reference cnmy1 250 the present work cnmy2 174 the present work mwcnts 109.31 (zohre et al. 2010) activated carbon 123 (suresh et al. 2011) mwcnts by cvd (acetylene, fe/si) 50.25 (liu et al. 2014) cotton stalk 147.06 (deng et al. 2011) graphene/magnetite 43.08 (ai et al. 2011) oxidized-cnts 99.83 (norzilah et al. 2011) calcined titanate nt 133.33 (xiong et al. 2010) mwcnts 59.7 (wang et al. 2012) fe3o4 – mwcnts (hno3) 48.06 (ai et al. 2011) probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  56 figure 9. fittings of pseudo-first order and pseudo-second order kinetics models for mb adsorption on cnmy1 (a, b) and cnmy2 (c, d), respectively. figure 10. langmuir and freundlich model plots for mb adsorption on cnmy1 (a,b) and cnmy2 (c,d), respectively. haiyam alayan, mustafa aljumaily, mohammed alsaadi, and mohd hashim 57 4. conclusion the growth of cnms from different precursors and optimizing their growth parameters were investigated. the optimal growth conditions for the hybrid carbon nanomaterials obtained from methane and acetylene were found at the temperature of 950, 60, h2/ch4 of 1.0 and 560 ºc, time of 38 min, and h2/c2h2 of 1.0, respectively. undoubtedly, the carbon source played a significant role to affect the morphology of cnms deposits due to discrepancies in their pyrolytic behaviors. methane has the simplest chemical structure and comparatively stable at hightemperature than acetylene. also, the gibbs free energies of the c2h2 has lower decomposition temperature than methane which gives rise to lower growth temperature of cnms. the mb removal process on each adsorbent was fitted well to a pseudosecond order kinetics model and the adsorption system was excellently presented by langmuir isotherm model with a maximum adsorption capacity of 250 and 174 mg/g for cnm/pac-ch4 and cnm/pac-c2h2, respectively. conflict of interest the authors declare no conflicts of interest. acknowledgment the authors would like to acknowledge the university of malaya research grant umrg (rp034d-15aet) and university of technology iraq for funding this research. references ahmad, a., razali, m. h., kassim, k., and amin, k. a. m. (2018). "synthesis of multiwalled carbon nanotubes supported on m/mcm-41 (m= ni, co and fe) mesoporous catalyst by chemical vapour deposition method." journal of porous materials, 25(2), 433-441. ai, l., zhang, c., and chen, z. (2011). "removal of methylene blue from aqueous solution by a solvothermal-synthesized graphene/magnetite composite." journal of hazardous materials, 192(3), 1515-1524. ai, l., zhang, c., liao, f., wang, y., li, m., meng, l., and jiang, j. (2011). "removal of methylene blue from aqueous solution with magnetite loaded multi-wall carbon nanotube: kinetic, isotherm and mechanism analysis." journal of hazardous materials, 198, 282-290. alayan, h. m., alsaadi, m. a., abo-hamad, a., alomar, m. k., aljumaily, m. m., das, r., and hashim, m. a. (2017). "hybridizing carbon nanomaterial with powder activated carbon for an efficient removal of bisphenol a from water: the optimum growth and adsorption conditions." desalination and water treatment, 95, 128-143. alayan, h. m., alsaadi, m. a., alomar, m. k., and hashim, m. a. (2019). "growth and optimization of carbon nanotubes in powder activated carbon for an efficient removal of methylene blue from aqueous solution." environmental technology, 40(18), 2400-2415. aljumaily, m. m., alsaadi, m. a., das, r., hamid, s. b. a., hashim, n. a., alomar, m. k., alayan, h. m., novikov, m., alsalhy, q. f., and hashim, m. a. (2018). "optimization of the synthesis of superhydrophobic carbon nanomaterials by chemical vapor deposition." scientific reports, 8(1), 2778. chavan, s., vitillo, j. g., larabi, c., quadrelli, e. a., dietzel, p. d., and bordiga, s. (2012). "functionalization of cpo-27-ni through metal hexacarbonyls: the role of open ni2+ sites."microporous and mesoporous materials, 157, 56-61. chen, l., and bai, b. (2013). "equilibrium, kinetic, thermodynamic, and in situ regeneration studies about methylene blue adsorption by the raspberrylike tio2@ yeast microspheres." industrial & engineering chemistry research, 52(44), 1556815577. deng, h., lu, j., li, g., zhang, g., and wang, x. (2011). "adsorption of methylene blue on adsorbent materials produced from cotton stalk." chemical engineering journal, 172(1), 326-334. elsagh, a., moradi, o., fakhri, a., najafi, f., alizadeh, r., and haddadi, v. (2017). "evaluation of the potential cationic dye removal using adsorption by graphene and carbon nanotubes as adsorbents surfaces." arabian journal of chemistry, 10, s2862-s2869. ezzeddine, z., batonneau-gener, i., pouilloux, y., and hamad, h. (2016). "removal of methylene blue by mesoporous cmk-3: kinetics, isotherms and thermodynamics." journal of molecular liquids, 223(supplement c), 763-770. gromov, d. g., bulyarskii, s., pavlov, a., scorik, s., shulyat'ev, a., and trifonov, a. y. (2016). "catalytic cvd-growth of array of multiwall carbon nanotubes on initially amorphous film co– zr–n–o." diamond and related materials, 64(supplement c), 97-102. hu, c., lin, y., connell, j. w., cheng, h. m., gogotsi, y., titirici, m. m., and dai, l. (2019). "carbon‐based metal‐free catalysts for energy storage and environmental remediation." advanced materials, 31(13), 1806128. jeong, s., lee, j., kim, h.-c., hwang, j. y., ku, b.c., zakharov, d. n., maruyama, b., stach, e. a., and kim, s. m. (2016). "direct observation of morphological evolution of a catalyst during carbon nanotube forest growth: new insights into growth and growth termination." nanoscale, 8(4), probing the effect of gaseous hydrocarbon precursors on the adsorptive efficiency  58 2055-2062. li, y.-h., gao, h.-q., yang, j.-h., gao, w.-l., xiang, j., and li, q.-y. (2014). "multi-wall carbon nanotubes supported on carbon fiber paper synthesized by simple chemical vapor deposition." materials science and engineering: b, 187, 113119. liu, y., kang, y., mu, b., and wang, a. (2014). "attapulgite/bentonite interactions for methylene blue adsorption characteristics from aqueous solution." chemical engineering journal, 237, 403-410. meshot, e. r., zwissler, d. w., bui, n., kuykendall, t. r., wang, c., hexemer, a., wu, k. j. j., and fornasiero, f. (2017). "quantifying the hierarchical order in self-aligned carbon nanotubes from atomic to micrometer scale." acs nano. norzilah, a., fakhru'l-razi, a., choong, t. s., and chuah, a. l. (2011). "surface modification effects on cnts adsorption of methylene blue and phenol." journal of nanomaterials, 2011, 55. shen, k., chen, x., chen, j., and li, y. (2016). "development of mof-derived carbon-based nanomaterials for efficient catalysis." acs catalysis, 6(9), 5887-5903. siddiqui, s. i., fatima, b., tara, n., rathi, g., and chaudhry, s. a. (2019). "recent advances in remediation of synthetic dyes from wastewaters using sustainable and low-cost adsorbents." the impact and prospects of green chemistry for textile technology, elsevier, 471-507. subramani, s., and thinakaran, n. (2017). "isotherm, kinetic and thermodynamic studies on the adsorption behaviour of textile dyes onto chitosan." process safety and environmental protection, 106, 1-10. sun, d. l., hong, r. y., liu, j. y., wang, f., and wang, y. f. (2016). "preparation of carbon nanomaterials using two-group arc discharge plasma." chemical engineering journal, 303, 217230. suresh, s., sugumar, r. w., and maiyalagan, t. (2011). "equilibrium and kinetic studies on the adsorption of methylene blue from aqueous solution onto activated carbon prepared from murraya koenigii (curry tree) stems." asian journal of chemistry, 23(10), 4486. tao, f., and crozier, p. a. (2016). "atomic-scale observations of catalyst structures under reaction conditions and during catalysis." chemical reviews, 116(6), 3487-3539. wang, c., yang, s., ma, q., jia, x., and ma, p.-c. (2017). "preparation of carbon nanotubes/graphene hybrid aerogel and its application for the adsorption of organic compounds." carbon, 118, 765-771. wang, j., chen, z., and chen, b. (2014). "adsorption of polycyclic aromatic hydrocarbons by graphene and graphene oxide nanosheets." environmental science & technology, 48(9), 4817-4825. wang, s., ng, c. w., wang, w., li, q., and hao, z. (2012). "synergistic and competitive adsorption of organic dyes on multiwalled carbon nanotubes." chemical engineering journal, 197, 34-40. xiong, l., yang, y., mai, j., sun, w., zhang, c., wei, d., chen, q., and ni, j. (2010). "adsorption behavior of methylene blue onto titanate nanotubes." chemical engineering journal, 156(2), 313-320. zhou, l., ji, l., ma, p.-c., shao, y., zhang, h., gao, w., and li, y. (2014). "development of carbon nanotubes/cofe 2 o 4 magnetic hybrid material for removal of tetrabromobisphenol a and pb (ii)." journal of hazardous materials, 265, 104114. zohre, s., ataallah, s. g., and mehdi, a. (2010). "experimental study of methylene blue adsorption from aqueous solutions onto carbon nano tubes." international journal of water resources and environmental engineering, 2 (2), 016 -028. comp041122.qxd the journal of engineering research vol. 5, no.1 (2008) 1-6 1. introduction given the constraints of the high cost and limited availability of radio spectrum, efficient spectrum usage is a cornerstone in the economic success of third generation cellular systems. one novel approach to using available resources more efficiently is the combination of existing radio systems into coordinated, hybrid systems, in order to combine the strength and capabilities of the individual systems. the combination of systems adds a degree of freedom in the sense that at any time the most spectrally efficient transport system can be chosen depending for example on quality-of-service requirements and traffic characteristics. most of the third generation systems use cdma as the multiple access method. furthermore some systems use tdma or fdma or both as the multiple access method. cdma systems use spreading methods to spread a narrow band signal into a wideband signal. since the spreading process does not change the power of the signal, the power spectral density of the spread signal is lower than the power density of the narrowband signal. figure 1 shows the principle of the spreading and de-spreading process. the narrow band useful signal is spread and the resulting signal is transmitted over a wideband channel _________________________________________ *corresponding author’s e-mail: abbosh@itte.uq.edu.au and de-spread in the receiver. as can be seen in fig. 1, the de-spreading process in the cdma receiver operates as a spreading process for a narrowband interfering signals. in terms of coexistence of tdma/fdma and cdma systems, the narrow band tdma (or fdma) channels act as narrowband interferer for the cdma system if they exist in the same frequency band as the cdma carrier. from the tdma (or fdma) systems point of view, the cdma system acts like a noise-like interferer with a low power spectral density. the receiver filter in the tdma (or fdma) receiver will make sure that the noise-like interfering signal will only be received in the narrow bandwidth of the tdma system. thus, only a small part of the interference power originating in the cdma system will be received, see fig. 2. the need for future multiple access schemes employing an overlay is stressed in many papers (schilling, lomp and garodnick, 1993; grieco and schilling, 1994; himimy and gupta, 1996; widdowson, 1997; koorevaar and ruprecht, 1999). in all these papers simulation algorithms were used to study the effects of overlaying different schemes on the performance of the hybrid systems. in addition to simulation, some authors tried to study the overlay process from a theoretical point of view (singh, 1997; latchman and chitamu, 2003). the inaccuracy of the model used due to oversimplification of the assumptions leads to incorrect expectations. theoretical investigation into spectral coexistence of cdma and tdma systems a.m. abbosh*1 and m. a. shehathah2 1school of itee, the university of queensland, brisbane q4072, australia 2department of electronics engineering, university of mosul, iraq received 22 november 2004; accepted 11 october 2006 abstract: the scarcity of available radio spectrum presently limits the extension of modern multimedia systems. this paper presents a theoretical investigation into the possibility of using a frequency overlay of a narrowband code division multiple access (cdma) system and a time division multiple access (tdma) system to provide a greater spectral efficiency. this paper shows that under certain conditions the two systems can operate in the same frequency band and in the same area with a considerable improvement in the overall capacity of the whole system. keywords: capacity, cdma, hybrid system, tdma »æeõ º«°ù≤j hp ông ™e …õeôdg º«°ù≤àdg hp o~©àÿg õ«øëàdg ωé¶æd »ø«£dg ¢ûjé©àdg ‘ …ô¶f »≤«≤– åëh ¢tƒñy .ω.cg1*¬jéë°t .cg .ω h2* áá°°uuóóÿÿggωg~îà°sg á«féµeg ‘ éjô¶f é«≤«≤– éãëh ω~≤j ábqƒdg √òg .áãj~◊g §fé°sƒdg √o~©àÿg ᪶f’g ‘ ™°sƒàdg øe o~– ~lgƒàÿg …ƒjogôdg ∞«£dg ‘ á«dé◊g √q~ædg ¿g : πª©dg úeé¶æ∏d øµá áæ«©e ±hôx â– ¬fg åëñdg í°vƒjh .ècg ∞«w iaéøc òaƒàd »æer º«°ù≤j hp ôngh á≤«°v áeõëh …õeq º«°ù≤j hp o~©àe õ«ø– ωé¶æd …ooôj ™wé≤j .ωé¶æ∏d á«∏µdg ¬©°ùdg ‘ ®ƒë∏e ø°ù– ™e á≤£æÿg ¢ùøf ‘h ooîdg áeõm ¢ùøæh gáá««mmééààøøÿÿgg ääggooôôøøÿÿ .»æeõdg º«°ù≤àdg h o~©àÿg õ«øëàdg ,øé¡e ωé¶f ,…õeôdg º«°ù≤àdg hp o~©àÿg õ«øëàdg,¬©°ùdg : 2 the journal of engineering research vol. 5, no.1 (2008) 1-6 in this paper, a fourier transform is used to investigate the benefits of spectral coexistence of cdma and tdma systems. also the model used in this paper considers the case of complete overlapping between the two systems to explore the worst case condition. effects of the voice activity ratio, the tdma duty cycle and the code used in cdma system besides the lack of synchronization between cdma and tdma are included in the analysis. 2. analysis the analysis considers the mutual effect of cdma and tdma systems operating in the same area and using the same frequency band. the overall capacity of the hybrid system is investigated to see whether the performance is degraded or enhanced. in the following analysis the capacity of the system is assumed to be the total number of users working at the same time with a signal to noise ratio above the threshold level. 2.1 effect of the cdma signal on the tdma receiver assume a cdma signal of the form: (1) where pc is the cdma received power at the base station, m(t) is the message signal, c(t) is the chip sequence, wc is the radian carrier frequency. f spread data f spread data interfering signal × × channel code code m(t) f interfering signal f message signal message signal spread interfer ence f fig.1: e ffect of tdma signal on cdma receiver figure 1. effect of the tdma signal on the cmda receiver channel f cdma interfering f bpf f cdma signal tdma signal f figure 2. effect of the cmda signal on the tdma receiver )twcos()t(m)t(cp2)t(r cc= 3 the journal of engineering research vol. 5, no.1 (2008) 1-6 let us assume that: tb is the bit duration, tc is the chip duration, k1 = tb / tc is the number of chips per bit, c(f) is the fourier transform (f.t.) of c(t), m(f) is f.t. of m(t) and r(f) is f.t. of r(t). as k1 >>1 according to definition of cdma then m(f) appears as a dirac delta function to the relatively wideband c(f) spectrum. for a single bit of duration tb then r(f) is: (2) where ctb(f) is the f.t of the k-chips in a single bit duration tb which is given by: (3) where ui = ±1 and sinc (x) = sin (µx) / µx. the amount of the cdma interference energy experienced in a bandwidth equal to b by a tdma signal working at carrier frequency ft , is given as: (4) substituting from (2) into (4); (5) since f varies from ft b/2 to ft + b/2 (as maximum), and assuming ft = fc (full overlaying of cdma & tdma); hence: f ft = (fc ± b/2) ft = ± b/2. as ± b/2 is very small relative to fc, we can assume that sin c (± b/2 fc) = 1. hence (5) can be written as: (6) therefore, the interference power from the cdma source in that small bandwidth (b ) can be expressed as: (7) substituting (6) into (7); (8) assuming that the result of the double integral of (8) is: (9) where g1 is a constant that depends on the bandwidth (b ) and code type and length. the values of g1 has been calculated numerically using matlab program and walsh code. depending on the parameter values shown in table 1, the value of g1 is shown to be equal to 0.0021. the signal to noise ratio at the tdma receiver (st ) can be expressed as: (10) 2.2 effect of the tdma signal on the cdma receiver let the tdma signal be represented by: (11) where b(t) is the information signal (message). at the cdma receiver, the tdma signal appears as: (x) = g(t) c(t) (12) substituting (11) into (12); (13) let td be the bit duration for the tdma signal, kd = integer (td / tc ) = number of chips per bit. for a single bit of duration , td, f.t of x(t) is: (14) ctd (f) is f.t of k2 chips in a single bit duration which is given by: (15) the amount of the tdma energy (et) actually lying within the passband of the cdma receiver working at carrier frequency fc can be written as in (6): (16) the value of the tdma power interfered with the cdma signal is: )]ff(c)ff(c[ 2 p )f(r ctbctb c −++= ∑ = −−= 1 1 )12()sinc()( k i tcifj icctb eufttfc π df)f(r)f(e f 2/btf 2 i ∫ − = df 2f 2 b tf 1k 1i ct)1i2)(cff(jeiu) cf cffinc(sct2 cp2)f(ie ∫ − ∑ = −−−−= π dfeu f p )f(e 2f 2 b tf 1k 1i ct)1i2)(cff(j i2 c c i ∫ − ∑ = −−−= π df)f(ep 2 b tf 2 b tf ici ∫ + − = dfdfeu f p p 2f 2 b tf 1k 1i ct)1i2)(cff(j i 2 b tf 2 b tf 2 c c ci ∫ − ∑ = −−− ∫ + − = π c1ci pgp = oc11 t t npvgn p s + = where tp is tdma received power, 1n =number of cdma users, v = voice activity ratio ( ),5.0≤ on = thermal noise = ktb ; where k = boltzman n’s constant = 1.38*10 -23 joules/kelvin, t is the temperature in kelvin. )twcos()t(bp2)t(g tt= )twcos()t(c)t(bp2)t(x tt= )]ff(c)ff(c[ 2 p )f(x ttdttd t −++= ∑ = −−= 2 1 )12()(sinc)( k i tcifj icctd eufttfc π dfeu f p )f(e 2f 2 b cf 2k 1i ct)1i2)(tff(j i2 c t i ∫ − ∑ = −−−= π 4 the journal of engineering research vol. 5, no.1 (2008) 1-6 (17) as before, the integral above can be assumed to be equal to: (18) the value of g2 can be calculated for all the possible walsh codes. figure 3 shows this result. it is to be concluded here that due to lack of synchronization cdma and tdma systems, as they work separately, then all possible values of g2 are shown in fig. 3 can be expected. to analyze the worst case condition we shall take g2 =0.00386, see fig. 3. the signal to noise ratio at the cdma receiver ( sc ) after using (18) can be expressed as: (19) 3. capacity of the hybrid system in order to see if the spectral coexistence of cdma and tdma systems is useful and possible we have to investigate the variation of the system capacity (cdma+tdma) as a function of number of cdma users and tdma users for different cdma received power (or different tdma received power), this study is developed for constant values of sc and st . solving (10) for pt : (20) substituting in (19) we get: (21) solving for n2 , we get: (22) adding n1 to both sides of the above equation we get the overall capacity nt : (23) solving (21) for n1 : (24) by adding n2 to both sides of (24): (25) 4. results of analysis depending on the results of the theoretical analysis discussed in the last section, variations of the overall capacity in the hybrid system with different parameters is investigated for the worst case (that is, maximum overlapping; fc = ft ). values of the needed parameters are selected as shown in table (1) to comply with practical systems. figure 4 shows variations of the overall capacity with number of cdma users (n1) for different values of the received cdma power (pc). it is shown that for a constant value of pc, the overall capacity increases with increasing n1 till a certain limit after which any increase in n1 causes a sudden drop in the overall capacity. this is because increasing the number of users causes degra df)f(ep 2 b cf 2 b cf iti ∫ + − = parameter value parameter value btb /1= 19200 hz ct 1/1.2288*10 6 s ts 8 cs 0.04 table 1. values of different parameters used in the paper t2ti pgp = figure 3. variation of g2 with walsh code index and possible synchronizations. maximum value for g2 (the worst case scenario) is shown ot22c1 c c npvgnvp)1n( p s ++− = θ otc11tt nspvgnsp += oot1c1t22c1 c c n)nsvgpns(vgnvp)1n( p s +++− = θ )nsvgpns(vgs )nsvp)1n(s(p n ot1c1t2c occ1cc 2 + +−− = θ 1 ot1c1t2c occ1cc t n)nsvgpns(vgs )nsvp)1n(s(p n + + +−− = θ )]vgpsvgnsvps[ ]vpsnsnsvgns[p n 1ct22ccc ccoccott22cc 1 θ θ + −+− = 2 1ct22ccc ccocot22cc t n)]vgpsvgnsvps[ ]vpsnsnsvgns[p n + + −+− = θ θ 5 the journal of engineering research vol. 5, no.1 (2008) 1-6 dation in the signal to noise ratio. if the amount of increase in the number of users cause this to be lower than the threshold then all users will be assumed in the off state and hence a sudden decrease in the capacity occurs. this suggests that it is of great importance to choose carefully the number of cdma users in the hybrid system. the case is different when changing the number of tdma users. figure 5 shows variations of the overall capacity with the number of tdma users (n2) for different values of pc. it is obvious that the overall capacity increases with increasing n2. depending on the above results, it is possible to say that under certain limits on values of n1, n2, pc and pt , the hybrid system (cdma+tdma) can have a total capacity which is higher than the cdma (or tdma) system working alone by about 100%. this conclusion can be explained with the help of figs. 4 and 5. maximum capacity with a tdma system working alone, ie. n1 = 0, or with a cdma system working alone, ie. n2 = 0, is 48 users for the parameter values assumed in this paper. while with the hybrid system maximum capacity can reach up to 96 users. this means that the spectral coexisno. of cmdamusers (n1) o ve ra ll ca pa ci ty figure 4. variation of overall capacity with n1 for the different values of pc o ve ra ll ca pa ci ty no. of tdma users (n2) figure 5. variation of overall capacity with n2 for different values of pc 6 the journal of engineering research vol. 5, no.1 (2008) 1-6 tence of cdma and tdma systems is possible and has real benefits under certain conditions. 5. conclusions in this paper, theoretical investigations on the spectral coexistence of narrow band cdma and tdma systems has been presented. results of analysis have proved the possibility of the spectral coexistence of the two systems under certain conditions which shows that the available frequency band can be used in an efficient way. it has been shown that an increase of about 100% in the overall capacity can be obtained for a certain range of values for the number of cdma users, number of tdma users, received cdma power, and received tdma power. references grieco, d. and schilling, d., 1994, "the capacity of broadband cdma overlaying a gsm cellular systems," ieee 44th vehicular technology conference, vol. 1, pp. 31-35. himimy, h.h. and gupta, s.c., 1996, "overlay of cellular cdma or amps forward and reverse link," ieee transactions on vehicular technology, vol. 45(1), pp. 51-56. koorevaar, p. and ruprecht, j., 1999, "frequency overlay of gsm and cellular b-cdma," ieee transaction on vehicular technology, vol. 48(3), pp. 696-707. latchman, p. and chitamu, p., 2003, " the study of band sharing for gsm 900 and cdma 2000 1x coexistence within the 900 mhz cellular band," www.ee.wits.ac.za / ~ comms/. pabst, r. and seidenberg, p., 2001, "coexistence of umts and egprs to enhance network capacity," proc. of the 10th aachen symposium on signal theory, germany, pp. 43-48. schilling, d., lomp, g.r. and j. garodnick, 1993, " broadband cdma overlay," ieee 43rd vehicular technology conference, pp. 452-455. sinha, r., 1997, "spread spectrum interference issues in the 900 mhz ism band," m.sc. thesis, the state university of new jersey, usa. widdowson, t., 1997, "a cdma data overlay of the gsm network", ieee pimrc, pp. 160-163. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 17, no. 2, (2020) 112-125 *corresponding author’s e-mail: nhinai@squ.edu.om doi:10.24200/tjer.vol17iss2pp112-125 analysis of factors affecting motivation in projects: a case study in oil and gas industry in the sultanate of oman nasr al-hinai 1, *, sujan piya 1, and khalid al-wardi 2 1 department of mechanical and industrial engineering, sultan qaboos university, muscat, oman 2 department of petroleum and chemical engineering, sultan qaboos university, muscat, oman abstract: motivation is a key element for successful project execution. however, different people are motivated by various means in different sectors. hence, this work aims to study, analyze and define the main motivational factors in project execution in the oil and gas industry. to achieve this, a dedicated survey is prepared and distributed among employees working in the oil and gas industry in oman. to ensure diversifications and minimize biases, the survey is distributed randomly to people working in projects in different organizations in the oil and gas industry. altogether, 86 respondents completed all the survey questions. the study revealed that, in general, external (extrinsic) factors have a significant motivational influence on employee performance in which money is considered as a powerful motivator. moreover, task achievement is found to be the major intrinsic motivational factor influencing the employee’s performance. the study also reveals that irrespective of years of experience, organizational type and the level at which employee works, lack of management support will have a significant influence in lowering team motivation in the oil and gas project. the study can serve as a guideline for the oil and gas industry to target the specific factor which helps enhance the level of motivation of a specific segment of employees. keywords: motivation; project management; oil and gas industry; extrinsic factors; intrinsic factors. تحلیل العوامل المؤثرة على الدافع في تنفیذ المشاریع: دراسة حالة في صناعة النفط والغاز في سلطنة عمان خالد الورديو، سوجان بیا، *نصر الھنائي التحفیز ھو عنصر أساسي لنجاح تنفیذ المشروع. بید أن مختلف الناس تحركھم وسائل مختلفة في مختلف القطاعات. :الملخص یھدف ھذا العمل إلى دراسة وتحلیل وتحدید العوامل المحفزة الرئیسیة في تنفیذ المشاریع في صناعة النفط والغاز. وفي سبیل ئیة وتوزیعھا على الموظفین العاملین في صناعة النفط والغاز في سلطنة عمان. ولضمان تحقیق ذلك تم إعداد دراسة استقصا تنویع عینة االختبار وتقلیل التحیزات الشخصیة إلى أدنى حد، ُوزعت الدراسة االستقصائیة عشوائیا على عاملین في مشاریع كشفت الدراسة أن العوامل الخارجیة وسئلة االستبیان. جمیع أ موظفاً 86منظمات مختلفة في صناعة النفط والغاز. أكمل ل تابعة تؤثر بشكل عام في أداء الموظفین بشكل تحفیزي كبیر حیث تعتبر األموال حافزاً قویاً. وعالوة على ذلك، تبین أن إنجاز المھمة بصرف النظر عن سنوات یشكل العامل التحفیزي األساسي الرئیسي الذي یؤثر على أداء الموظف. وتكشف الدراسة أیضاً أنھ ، والمستوى الذي یعمل بھ الموظف، فإن نقص الدعم اإلداري لھ تأثیر كبیر في خفض المنفذة للمشروع الخبرة، ونوع المنظمة ي تالوامل ع النفط والغاز. ویمكن أن تكون الدراسة بمثابة دلیل لصناعة النفط والغاز الستھداف العیراالحافز الجماعي في مش تعزیز الحافز لدى شریحة معینة من الموظفین.ساعد على ت النفط والغاز؛ تحفیز الموظفین؛ إدارة المشاریع؛ عوامل داخلیة؛ عوامل خارجیة. :المفتاحیة الكلمات analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 113 1. introduction every project manager would like to run a project that finishes with the desired quality standards, in the given time limit and within the predetermined budget such that it stands up to the end-users’ expectations. despite the much attention given to projects, it is claimed that most of the projects do not reach their objective. the majority of projects do fail for several reasons. these reasons can be due, but are not limited, to poor planning, poor execution, poor coordination, unclear scope of work, unqualified project manager, conflicts between team members, lack of motivation, etc. therefore, the success of a project depends to a large extent on how to overcome and minimize the causes of failure. the responsibilities put on the shoulders of project managers, as well as on the organization handling the project, is thus huge and require them to develop special managerial skills, frameworks, policies, and procedures (al-hinai et al., 2020). in many projects, the failure rate is considered to be high. for example, at the end of the last century, the standish group found that 44% of information system projects are partially successful and about 40% were cancelled (field, 1997). this leaves only 16% of the projects to be successful. one would expect that due to the increase in the use and the adaptation of project management tools and practices, the success rate will drastically improve. however, later figures released by the standish group is not much different. according to the 2014 chaos report, 31.1% of projects are cancelled before completion and 52.7% of projects face 189% cost overrun. the same scenario of higher failure rates in projects is not limited to the it projects, but it is encountered by other types of projects (chaos report, 2015). hence, many practitioners started to accept failure as unavoidable (mcmanus and wood-harper, 2003, mahaney and lederer, 2006). nevertheless, such an attitude may have a further negative consequence on projects causing them to further exceed in-time, onbudget, or not finishing them according to the scope. the proper approach is, therefore, to strive and explore every possible way to increase the probability of project success rate. since running and conducting projects is at the end the result of individuals working together, personal motivation becomes one of the major pillars in the overall project success. hence, acquiring the proper motivation process and rewards system for team members influences the project manager and his team to execute the project activities effectively and efficiently to meet project desires (meredith, 2016). demarco and lister (1999) conducted extensive surveys and game exercises to understand the reason for the project’s failure. they concluded that the absence of motivation is a frequent reason. thus, while project managers may possess the required technical skills to manage projects, lack of motivation may lead to poor performance (germann, 2004). further, sansone and harackiewicz (2000) defined motivation as the inner driver that stimulates and guides behaviour. while this definition may be debated as it limits it to inner drivers and ignores the outer drivers, yet one may conclude that the motivation process can be broken into three essentials (nel et al., 2001). these essentials are the motive, the behaviour and the goal. the motivation process starts with the motive that a person has, which in turn initiates certain actions or a series of activities. these activities or actions are what defines the behaviour of the individual and his means of achieving a certain desired output or in other words, the goal. hence, the goal can be thought of as the end result of something initiated by a motive. therefore, to increase the probability of success when conducting projects the project manager, as well as the organization, needs to ensure that the team members work harmoniously and every member of the team is able to fulfil his personal needs, interest, desires, and aspirations along with the objective of the project. a number of researchers pointed that at the absence of motivation, gifted individuals perform far less than their actual capabilities and they may be outperformed by ordinary motivated individuals (germann, 2004; bateman and snell, 1999). an environment that inspires cooperation, mutual trust, honesty, and gives a sense of belonging will be positively reflected in the team members’ performance as they will tend to do their best when they are in an environment that makes them feel valued. to achieve this, it would require “motivation’’. therefore, the definition of motivation can be generalized to be the essential means for stimulating the energy and the enthusiasm of the team in order to accomplish project goals by providing satisfactions to the team needs, interest and aspirations (khoshnevis and tahmasebi, 2016; hall et al., 2009). it is worth emphasizing here that a highly motivated person would be encouraged to contribute his extreme efforts individually and as a team towards the effective performance of the project. therefore, organizations must have the ability to identify and understand the motivational requirements for their employees as it is demonstrated that the higher the satisfaction of project team members’ needs is, the greater the chances for successful project execution will be. oil and gas development projects usually require a huge investment and the risk in the project is often quite high (kim and choi, 2019). similar to other parts of the world, the oil and gas industry in sultanate of oman is divided into three major sectors: upstream, midstream and downstream. the upstream sector includes exploration and productions of hydrocarbon fields by bringing the crude oil and natural gas from underground to the surface. companies like petroleum development oman (pdo), occidental oman (oxy), oman oil nasr al-hinai, sujan piya, and khalid al-wardi 114 exploration and production and daleel petroleum are an example of companies dealing with upstream industry. the midstream sector involves the transportation by either pipelines or other means of transportation of crude oil and gas from production sites (upstream) to refineries (downstream), an example of companies working in midstream companies is oman gas company (ogc). finally, the downstream sector deals with the refinement of petroleum crude oil and the processing and purification of raw natural gas, which is done by companies like orpic, oman oil refineries and petroleum industries company. the oil and gas industry is still considered to be the major contributor to the gdp in many countries, especially in the middle east and the sultanate of oman depends heavily on it. this is reflected by the huge ongoing investments in this industry. these huge investments are usually associated with many ageing fields or newly discovered ones. this results in continuous project execution to ensure the sustainability of oil and gas production. this, on the other hand, increases the overall cost of production. moreover, the falling oil price is putting further pressure on the shoulders of oil and gas companies as it lowers their margin of profit in contrast to the increase in cost. while projects in the oil and gas industry have to continue, attention is now more serious than ever to ensure that these projects are successfully implemented and executed to meet the desired outcomes while meeting the on-budget and withintime constraint or at least minimize the deviations. thus, the main objective of this study is to tackle a major pillar that affects project success. specifically, the study is aiming to address two main research questions as shown below: 1. to determine the motivational factors that mainly affect employee’s performance in projects related to the oil and gas industry in the sultanate of oman. 2. to examine the effect of motivational categories on team motivation during various project phases. the focus of this study is to look at both, the intrinsic factors that can be monitored and analyzed by the organization and the extrinsic factors that can be influenced by it through the support and enforcements of the senior management. while one may argue that organizations have less degree of influence over the personal variables, however, understanding how these personal variables (intrinsic motivators) and how they emerge with time will enable senior management to better understand and organize the work-related motivational factors and accordingly their reward systems. 2. the need for motivation effective management of projects is crucial to the sustainability of organizations (kanagarajoo et al., 2019) and employee motivation plays a crucial role in managing project effectively (van der kolk et al., 2019). motivation is an essential element for stimulating the energy and enthusiasm of employee’s performance towards the achievement of the project’s goals. according to lawal and okhankhuele (2014), motivation is a method that starts with a physiological deficit or need that activates a behaviour or a drive aiming at a goal encouragement. hence, they divide motivation into three interlinked elements, needs, drives and incentives. in practice, motivation helps to energize employees to work. therefore, finding appropriate techniques to fulfil employees’ desires and needs is considered the key to a successful motivation process. the main challenge lays in the fact that each individual has different needs, wants and desires (tabassi et al., 2011). in this sense, there may be many motivation sources that can affect the employee’s performance. yet, these sources cannot be generalized as what motivate workers differs accordingly. taylor (1903) who is voted the most influential in the field of management science, long stated that what motivates an individual employee today, may not motivate him tomorrow because human needs are limitless. further to that, according to maslow (1943) “a person tries to satisfy the most important need first. when the need is satisfied, it will stop being a motivator and the person will then try to satisfy the next most important need” (kotler et al., 2018). motivation is, therefore, an essential key element to enhance productivity in the workplace (asproni, 2004). a highly motivated employee puts more effort and energy towards the achievement of the project and the organizational goals by providing better performance. well-motivated employees are usually results-oriented people in which they look forward to alternative methods to perform the assigned tasks while improving their performance. therefore, employees should be motivated to bring high-quality service in an efficient way. achieving this will have a better effect on project progress and success, which in turn can be reflected on the employee’s performance and achievement (nchorbuno, 2011). it is of utmost importance to understand that the motivation sources do not rely only on external sources but also can be internal sources. based on psychological theories that concentrate mainly on human’s/employee’s needs, motivation source can be classified into two types. intrinsic sources are in-built in the employee himself and as well between the employee and the job. these make the employee creative, enthusiastic, find work pleasurable, want to face challenges, and achieve something. therefore, intrinsic motivation happens when an individual finds, or more precisely perceives, the activity itself interesting, enthusiastic, pleasant, and hence it derives spontaneous satisfaction (herzberg, 1987). different theories related to motivation discussed intrinsic motivation through content theories that may be found in maslow (1943), alderfer’s (1972), herzberg et al. analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 115 (1959), etc. the second sources are extrinsic sources that come externally from outside the job itself. for example, the employer, the working environment, and other encouragements and incentives including salary, promotion, awards, and benefits, which in return increases employers’ level of commitment to work (lawal and okhankhuele, 2014) or even punishments that may lead to having a certain behaviour not to be repeated (skinner, 1969). theories discussing such extrinsic factors are categorized under process theories and such theories that can be found in adams (1963), locke (1968), vroom (1964), etc. nevertheless, the two factors, intrinsic and extrinsic, cannot be separated as a number of studies showed intrinsic factors have an influence on the extrinsic factors (dolfi and andrews, 2007; lee-kelly and leong, 2003; mueller and turner, 2009, kooij et al., 2010). there are indeed numerous studies and literature over the past seven decades on how can team members be motivated at their jobs such as dwivedula and bredillet (2010), sharp et al. (2009), hall et al. (2009), nicholson (2003), pinder (1998), heimovics and brown (1976), and labor relations institute (1946). the early survey study conducted by labor relations institute (1946) revealed 10 motivators for employees working in the industry. these factors are appreciation over achieved work goals, sympathy over personal problems, good salaries, interesting work, loyalty to employees, job security, promotion, working conditions, feeling in control of things and tactful disciplining. later studies indicated that salaries and appreciation over achieved work goals are ranked to be the top motivators (keller, 1978; kovach, 1987, 1995; wiley, 1997; fisher and xue ya yuan, 1998; baddoo et al., 2006). however, the study conducted by ferratt and short (1986) who studied the needs of an insurance company employees identified five central needs out of which none of these needs is related to good salaries. similarly, beecham et al. (2008) reviewed 92 papers that studied the motivational factors for software engineers. accordingly, they were able to identify the seven most common motivators for these software engineers and none of these motivators is related to good salaries. the high uncertainty and high demand nature of the projects’ environment resulted in a highly pressurized environment for project managers and project team members to work in. there exist at least three areas of pressure in the project environment, the uncertainty in projects, the extreme need for integration, and the frequent occurring urgencies (turner and mueller, 2003). this made it very clear that people working in projects must possess special characteristics that may not necessarily be required in people working in, for example, classical line departments. while some may argue that the project environment itself cause project team members to be self-motivated (verma, 1996), others argue that this is an oversimplification as it underlines the focus to how project team members can be motivated rather than what motivates them (seiler et al., 2012). this may partly explain why there are fewer literature addressing the factors that motivate project team members (tampoe and thurloway, 1993; sharp et al., 2009). tampoe and thurloway (1993) conducted a study to identify the motives for project managers and team members working in it and r&d organizations. accordingly, they identified five motives, recognition for achievement, mutuality, authority and control, ability to use their own creativity, and belonging. verma (1996) advocated eight motivational factors for project managers including the project culture, reward system, work itself, environment, previous success, feeling confident about the work, and competition. the list for possible motivators for project managers was later increased to contain other factors such as career development, receiving positive feedback, contribution to the organization, etc. (linberg, 1999; mak and sockel, 2001). mahaney and lederer (2006) investigated the effect of intrinsic and extrinsic rewards on information system projects success. their study revealed that intrinsic rewards leads to client satisfaction and maintain quality whereas extrinsic rewards affect the implementation success (on-time and within budget). further, seiler et al. (2012) developed an integrated model aimed at measuring the factors influencing project managers’ motivation. in their model, they defined motivational dimensions containing 47 items. an empirical study showed that project managers are motivated by a mixture of instinct and extrinsic factors along with specified needed support from the organization. khan and jalbani (2009) studied the effect of the cash and non-cash reward compensation system for employees working at pakistan petroleum limited company. their analyses were done based on a survey and a number of interviews with 60 employees. the employees are from different management level at the company. their finding showed that lower-level employees are motivated more by cash rewards and less by non-cash rewards. however, the effect is viceversa for the upper management level employees, who give much less attention to the cash rewards. as for the middle management level, the results indicated that cash and non-cash rewards are equally important. in this work, cash rewards are considered as an extrinsic motivator while non-cash rewards are considered as intrinsic motivators. in short, while herzberg (1987) advocated that workers could be motivated by using intrinsic and extrinsic rewards, careful study of the literature suggests that the effect of intrinsic and extrinsic rewards is not as conclusive. many studies, as the above-discussed ones, showed that different workers will be motivated by different means in different organizations. they further revealed that the needs and motivators might differ according to the nature of the work and type of organization. actually, the effect may not only depend on the organization, but it may change as the years of experience and/or the held nasr al-hinai, sujan piya, and khalid al-wardi 116 position changes. nevertheless, as a failure in projects can have a catastrophic effect, all possible stimuli in project success must be considered. therefore, understanding how intrinsic and extrinsic rewards affect project success is very crucial. one may not emphasize enough that satisfying the needs, wants and desires of workers running projects are of great importance to any successful organization. giving the employees and workers what makes them feel more connected and appreciated within the organization, either physiological or psychological, increases the chances of team members working in more harmony as well as having better understanding and appreciation of the organization’s goals and policies. it can also enhance the overall productivity and may lead to a reduction in for example absenteeism and turnovers. 3. research methodology and data collection in light of the above literature, studies followed different approaches and each has focused on a different scope. the results from these studies were collected using different means. some were based on conducting surveys, and others used structured and/or unstructured feedbacks via interviews or written responses. some studies focused on motivational factors related to a specific group of professionals, while others were more general. nevertheless, these researches share many common motivators. therefore, this research aims to identify the motivational factors and study their effects on employees working in the oil and gas industry in oman and the possible impact on project execution. to achieve this aim, experts from industry and academia held a number of meetings and brainstorm sessions discussing this matter. the reason for using brainstorming techniques is due to the fact that it is one of the most effective techniques for creative problem solving and resolve biases among participants to reach consensus (piya et al., 2019). further, related motivational factors collected from the above studies as well as other cited ones were summarized in table 1. thereafter, the table with an initial suggestion from the authors to cluster the collected factors into motivational categories was presented to the experts for revision and further discussions. the final decision was made to divide the motivational factors into intrinsic and extrinsic factors and then cluster them into subgroups or categories under each motivational type, as shown in table 2. during the allocation of factors to categories, factors with similar meaning were merged together and others were reformulated such that a consistent formulation is achieved under each category. subsequently, five motivational categories were defined under intrinsic motivational type and six categories under extrinsic motivational type as shown in table 2. by following this approach, we were able to identify relevant motivational factors and reduce redundancies while forming more balanced and proper category clusters for them. accordingly, four research questions were identified for this study. these questions are: 1how do the intrinsic and extrinsic motivations affect performance? 2how do different intrinsic motivational categories affect performance? 3how do different extrinsic motivational categories affect performance? 4how do different motivational categories affect team motivation during various project phases? furthermore, it was agreed that the design variables for this study are: 1gender 2organization type; upstream, midstream or downstream 3years of experience 4position as a result, it was concluded that this study will seek to understand the following: • is there a difference between intrinsic and extrinsic motivation in the performance of the employees? • which extrinsic motivation reward has more effect on the performance of the employees? • which intrinsic motivation reward has more effect on the performance of the employees? • how do various design variables affect motivation? • how does the motivation influence the team members’ performance level at the various project phases? in order to conduct this study and find out the level of motivation and its impact on employees and their organization, an initial online questionnaire survey with mainly multiple-choice format was developed. in designing the questionnaire, an approach similar to hwang et al. (2015), which is the widely used approach by project management researches, is followed. once the questionnaire was designed, the second set of meetings with several experts on project execution from oil industrial companies were conducted. the experts, in accordance with people from academia, were consulted to review the questionnaire and provide feedback. consequently, the comments provided by them were used to redesign and modify the questions and to improve the overall format of the questionnaire. the final questionnaire consists of two sections, a and b. section a targets the demographic information about the respondent including gender, years of experience, workplace, work area, a position held and the organization type they are working for. section a is carefully designed to address and fulfil a number of aims. it is used at the beginning as a qualifying part to determine whether the respondent is a qualified participant or not. in addition, the different questions in this section include the design variables of this analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 117 study and accordingly they help in categorizing the collected data from section b and understand how, for example, gender, years of experience affects motivation. further, this demographic information helps us to avoid making wrong or unrealistic interpretations from the collected data, and hence, avoid making inaccurate causal statements that may be associated with descriptive research. section b consists of mainly multi-choice questions. these multi-choice questions aim to measure the effect of intrinsic and extrinsic motivation in project execution in the oil and gas industry in oman and answer the research questions. the questionnaire was made available and distributed via survey monkey website and app (an online survey-distribution service). accordingly, email invitations with the link for the questionnaire were sent to employees working in different oil and gas projects in oman. furthermore, to enhance and increase the respondent rate, the link was made available and sent through social communication media to random practitioners working in the oil and gas industry in oman. the later medium was found to be the most convenient to reach the targeted participants and have them participate. from a total of 304 invitations that were sent via emails and social media, 71 were not valid emails and contact numbers. 4. discussion of results and findings 4.1. analysis and validations of collected data the total number of respondents who completed the survey and cross the qualifying check part is 86 respondents. thus, the response rate is 36.9%. according to shih and fan (2007), in the metaanalysis, the average response rate in web-surveys is 19%. hence, the achieved response rate in this study is above that average which reduces the probability for systematic biases due to non-response. of the 86 respondents, 69 are male and 16 are female with various years of experience, a likely representation of gender working in projects in the oil and gas industry in oman. the majority of respondents, forming about (48%) of the sample size, have 5-15 years of experience followed by (18%) of the sample size having 0-5 years of experience. the remaining 20 % of the respondents have more than 15 years’ experience. out of 86 respondents, around 73% hold non-managerial position, and the rest hold a managerial position in various oil and gas companies in oman. this finding supports the actual environment of projects, as projects need more technical staff than managerial staff. table 1: overview of motivational factors cited from literature. reference considered motivational factors ferratt and short (1986) need for guidance; social needs; esteem needs; achievement needs; power needs herzberg (1987) company policy and administration; supervision; relation with supervisor/relationship with peers/personal life (interpersonal relationships); work condition; salary; relationship with subordinates; status; security; achievement; recognition for achievement; work itself; responsibility; advancement; growth keller (1978); kovach (1987) kovach (1995); fisher and xue ya yuan(1998) a full appreciation of work done; tactful disciplining; personal loyalty to employees; good wages; project reward system; promotional; growth in the organization; sympathetic understanding on personal problems; job security; good working conditions; feeling of being on things; interesting work; competition verma (1996) supervision; project culture/environment; reward system competition; work content; previous success; self-belief linberg (1999); mak and sockel (2001) receiving positive feedback; autonomy; belonging and making a contribution to the entire system; career development wiley (1997); baddoo et al. (2006) wages and rewards mahaney and lederer (2006) favourable annual performance appraisals; project completion celebration; job security; technical training; flexible work schedule; job promotion; financial bonus; newer technology (pc or laptop); time off; choice of future assignment; opportunity to work at home; private office space; pride; sense of contribution to organization; public praise beecham et al. (2008) need to identify with the task; clear goals; working on identifiable pieces of work; tasks variety; a clear career path; personal interest and understanding the task; job satisfaction tampoe and thurloway (1993); sharp et al. (2009); recognition; bounded power mutuality/belongings; creative autonomy amin et al. (2010) individual rewards/ group rewards/ tangible extrinsic rewards/ intangible extrinsic rewards; altruism; courtesy; conscientiousness; civic virtue; sportsmanship rony and suki (2017) employee extrinsic reward (promotions and benefits, work location, and working conditions); internal communication; training and development; leadership; employee intrinsic reward (receiving positive values for doing meaningful works) nasr al-hinai, sujan piya, and khalid al-wardi 118 table 2: summary of intrinsic and extrinsic motivational categories and sub-motivational factors. motivational type motivational category motivational factor in tr in si c freedom to initiate being able to initiate important decisions being able to initiate control measure when needed having the ability to use their own creativity being able to take a time off when needed pleasure/enjoyment/fu n having a good relationship with team members and superiors being able to have fun when implementing given tasks and enjoy the moment having a fair competitive atmosphere having a personal interest in the assigned task feeling satisfied when performing the task task challenges assigned to challenging tasks/projects being able to identify and understand given/assigned tasks having clear task/project goals/scope assigned to a variety of tasks task achievement being recognized for achievements being involved during project completion celebration feeling proud when completing tasks being recognized for previous achievement and success being treated as a professional as a result of personal achievements others receiving non-materialistic rewards having good relationships with superiors, peers and subordinates other non-mentioned intrinsic means of motivation e xt ri ns ic engagement to company policy being able to engage in making policy for the company being able to participate in making important decisions being able to suggest the future assignment being able to influence the roles of team members being able to contribute to the organization administration rewards receiving materialistic rewards receiving rewards based on performance having a clear promotional system based on performance having a group incentive reward system having an individual incentive reward system receiving rewards as recognition for achievement training and development having a clear career and promotional paths receiving technical and nontechnical training having the opportunity for advancement and growth in the company receiving proper guidance and supervision work environment having an appreciation of work done having job security having necessary support and understanding of personal problems working with coherent team members being part of a team who works in an act of altruism being treated with courtesy and good humour having good working conditions availability of needed space and new technologies (office, pc) having flexible work schedules having the opportunity to work from home receiving necessary resources (financial, personnel, material, etc.) and support from management to complete the work salary/wages receiving a good salary receiving good monthly/quarterly/yearly bonuses other having a good internal communication system receiving a favourable annual performance appraisals receiving public praise other non-mentioned extrinsic means of motivation oman, being an oil producer country, depends heavily on oil production as the main contributor to its gdp. especially, stakes are high in projects related to the exploration and extraction of oil and gas. when validating the sample size, we find that (85%) of respondents are from employees working in an upstream organization as compared to 6% and 9% of responses from midstream and downstream organizations, respectively. therefore, this higher response rate from upstream organizations’ employees is expected due to the bigger size of this sector compared with the other two sectors. further, analysis of collected responses revealed that the majority of respondents, about (60%), are working in main offices analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 119 while the remaining 40% are based in sites or desert areas. 4.2. effects of intrinsic and extrinsic motivation as stated in previous sections, there are two main means of motivation. intrinsic, which are related to the work itself such that it provides a sense of pleasure, inventive, enthusiasm, challenge and/or achievement. the second one is extrinsic means that comes from outside the actual conducted work like salary, promotion, benefits, and increased commitment. in general, the collected statistics, as shown in fig. 1, revealed that 60% of respondents believe that extrinsic factors or means have the major influence on their performance as compared to 40% who are motivated intrinsically. furthermore, as shown in fig. 1, employees’ response coincides with the identified intrinsic and extrinsic factors’ categories that affect employee’s motivation on the project execution. very few respondents identify other factors, apart from the four intrinsic and five extrinsic categories that motivate them. out of 86 respondents, 46% of respondent selected “task achievement” suggesting that it is the major intrinsic factors that affect employee motivation in the oil and gas industry. surprisingly, “freedom to initiate” received the least preferred choice (13%) as factors of motivation among the four identified categories. this may be due to the reason that the nature of the project in the oil and gas industry is generally complex. therefore, the employee may hesitate to take new initiative due to the fear of backlash if their initiative fails. task achievement is followed by “task challenges” and “work pleasure” accounting 22% and 15%, respectively. similarly, in terms of extrinsic factor, “salary” stands out with 44% of respondent suggesting it as a major influencer of motivation. the finding is consistent with monteiro et al., (2015), which has shown that the individuals’ motivational orientation is associated with their attitude towards money. “salary” is then followed by “work environment”, “training and development”, and “participation in policy making” accounting 18%, 16%, 12% and 7%, respectively. this confirms that salary is still the most powerful extrinsic factor to get employees to do more of what they do. further, the respondents were requested to identify the influential factor that lowers the motivation of the project team. the results obtained are as shown in fig. 2. all together five factors were identified, out of which 51% respondent pointed out that missing management supports is the most influential factor leading to a decrease in team motivation. 20% of respondents selected conflict between team members followed by lack of rewarding system (16%). failing to achieve projects goals and work nature has the lowest rating among identified factors with 6% and 7%, respectively. from the observation, it is clear that extrinsic factors belonging to “work environment” and “rewards” categories have more effect in lowing team motivation. this highlights the importance of extrinsic factors when it comes to motivation. results also suggest that while people are motivated more by salary as suggested in fig. 1, however, the absence of this motivational factor does not necessarily lower the team motivation. nonetheless, definitely, it will not increase the level of motivation as well. 4.3. effects of design variables on motivation results in section 4.2 do not take into account the effect of various design variables in the analysis of motivation. certainly, the design variables of for example years of employee experience on work will influence the factor that affects motivation. it is agreed that the motivator is a dynamic factor, which changes with the increase in the employee experience and position (shekhar et al., 2013). according to maslow hierarchy of need, a human being first tries to satisfy a physiological need. once this need is satisfied, other higher-level needs in the hierarchy dominate as a motivator until it reaches self-actualization. therefore, to analyze such effect, next, the survey results are segregated based on the various design variables considered in this study and then analyzed to draw conclusions. the effect of years of experience on intrinsic and extrinsic categories that affect motivation is presented in the form of a stacked bar chart in fig. 3. the chart clearly shows that employees were attracted more by extrinsic motivators at the beginning of their career. the percentage of employee motivated by extrinsic motivator decreases as he/she gains experience and progresses in their career. as the employee matures (> 15 years of experience), the extrinsic motivator is replaced with the intrinsic motivator to motivate them in the project. as shown in the figure, there is a sudden jump in the percentage of respondents who are in favour of an intrinsic motivator in this category of employee. tables 3 and 4 presents the overall survey results obtained by segregating the data based on all the design variables considered except “gender”. this is due to the fact that the number of females who took part in the survey was very less (less than one third) as compared to their male counterparts. note that the %age figure in the tables is rounded up or rounded down depending on the decimal value. the following information can be extracted from table 3. • task achievement is of major importance as an intrinsic motivator in the initial year of working in the oil and gas project. as the employee gets more experience, pleasure achieved from the job and challenges involved in the task will be the main motivator. freedom to take initiatives is more important to the people who are working in a non-managerial position as compared to the manager. this may be due to the reason that as a manager, the position itself entitle them with the ability to take initiatives. nasr al-hinai, sujan piya, and khalid al-wardi 120 figure 1. main intrinsic and extrinsic motivational categories and their importance allotment as preferred by oil and gas industries employees. figure 2. the most influential factors that lower the team motivation in projects. figure 3. effect of experience on motivation. • in the initial years, salary dominates the factor as an extrinsic motivator. however, rather than salary, it is a reward system and training, and self-development which is important as they mature. further, a pleasant work environment dominates the extrinsic motivation landscape for those who are working in the midstream section of the oil and gas industry. in this case, the percentage of the participants who are motivated by “work environment” related factors is three times higher than those who are motivated by “salary” related factors. • the result shows that employee enjoys engaging themselves in the administrative position (for eg. developing company policy and procedure) when they were somewhere in the middle of their years of experience. however, even in this case, very few employees prefer this role to motivate them. rather than upstream and downstream, this role is lucrative for the employee who works in the midstream section of the oil and gas industry and is more preferred by the employee working in the managerial position. the above results are in consistence with maslow theory in which employees try to satisfy the most important needs first and these can be accomplished by having a satisfactory monetary value and rewards. when the monetary need is satisfied, then the employee will try to satisfy the next need located in the maslow hierarchy such as self-esteem and selfdevelopment. both intrinsic and extrinsic motivators are important to help the managers to motivate their team to put all their effort toward the achievement of projects goal. similarly, table 4 shows the effect of design variables on other factors related to motivation. the following information can be extracted from the table. • interestingly, for all design variables, missing management support is found to be the major factor that will significantly lower employee motivation except when the employee is highly matured (i.e., who has a work experience of more than 15 years). in this case, the employee will be less motivated towards work if it is not challenging and the work is of routine nature. • the survey results show that employees themselves or their counterpart working at the same level is the person who can have the main influence on motivation, irrespective of the design variables. this is then followed by the manager. this is due to the reason that the employee will have interaction mainly with their counterpart or with the manager. stakeholders have hardly any role to play to motivate employees towards work. analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 121 • the results show that motivation is important at the initial phase and the buildup phase of a project. however, it does not mean that motivation is not needed at the final closing phase. in general, out of three phases, most of the participants felt that it is mainly needed at the buildup phase of the project. moreover, it was found from the survey result that the level of motivation of most of the employees working in the oil and gas sector in oman is medium. around 46% of the survey participant agreed to this view. the combined percentages of high and very high levels indicate that 32 % of employees are well motivated by their organizations while 21 % of employees are demotivated. a very low level of motivation has been found in employees that have 0-5 years of experience. on the other hand, a very high level of motivation is found with highly experienced employees. table 3: effect of design variables on intrinsic and extrinsic motivation. d es ig n v ar ia bl e intrinsic motivation (%) extrinsic motivation (%) task achievement freedom to initiate task challenge pleasure salary reward training and development engagement in company policy work environment years of experience 0-5 50% 11% 28% 11% 55% 22% 6% 0% 17% 5-10 48% 17% 20% 15% 42% 10% 21% 8% 19% 10-15 49% 11% 26% 14% 51% 14% 9% 6% 20% 15+ 0% 0% 33% 67% 0% 34% 66% 0% 0% position m an ag er ia l 43% 4% 30% 22% 39% 9% 22% 13% 17% n on -m an ag er ia l 47% 18% 21% 14% 45% 15% 16% 5% 19% organization type u ps tr ea m 49% 15% 23% 12% 45% 15% 19% 7% 14% m id st re am 0% 20% 40% 40% 20% 0% 0% 20% 60% d ow ns tr ea m 50% 0% 12.5% 12.5% 50% 0% 12.5% 0% 37.5% nasr al-hinai, sujan piya, and khalid al-wardi 122 table 4: effect of design variables on other factors of motivation. d es ig n v ar ia bl e factor that lowers motivation (%) people who have a major influence on motivation (%) phase at which motivation has a major influence (%) m is si ng m an ag em en t s up po rt c on fl ic t be tw ee n m em be rs l ac k of re w ar di ng sy st em fa ili ng to ac hi ev e pr oj ec t g oa l w or k na tu re e m pl oy ee m an ag er se ni or m an ag er st ak eh ol de rs in iti al p ha se b ui ld up ph as e c lo si ng p ha se years of experience 0-5 50% 33% 11% 6% 0% 50% 44% 0% 6% 50% 33% 17% 5-10 51% 15% 21% 9% 4% 37% 40% 23% 0% 33% 54% 13% 10-15 49% 29% 11% 3% 9% 46% 34% 14% 6% 49% 37% 14% 15+ 33% 0% 0% 0% 67% 67% 33% 0% 0% 33% 33% 34% position m an ag er ia l 65% 13% 17% 4% 0% 43% 35% 17% 4% 39% 52% 9% n on m an ag er ia l 46% 23% 15% 6% 10% 42% 37% 19% 2% 40% 44% 16% organization type u ps tr ea m 47% 19% 18% 7% 8% 44% 35% 18% 3% 41% 45% 14% m id st re am 60% 20% 20% 0% 0% 20% 60% 20% 0% 40% 40% 20% d ow ns tr ea m 75% 25% 0% 0% 0% 37.5% 37.5% 25% 0% 25% 62.5% 12.5% 5. managerial implications and limitations the study identified the motivational factors and then clustered these factors into categories that belong to intrinsic and extrinsic motivational types. the study shows the there are various extrinsic and intrinsic factors that affect the employee’s level of motivation at different phases, working experiences and positions. therefore, the result of this study can be used by the concerned authority and project managers to tailor the motivational factor that suits the need of the individual employee. this will have a significant positive impact on employee’s satisfaction, as well as organizational goals. the rate at which employees being motivated in their organization can be improved by focusing more to mitigate the factors that lower team motivation like improving management support, managing the conflict and apply the rewarding system in the organization. therefore, every organization should incorporate motivational packages and incentives in its corporate policies and plans in order to enable its implementation without interruptions to their business. it is recommended that a consistent study and assessment of individual organization needs to be carried out in order to know what motivates their employees and how motivation can affect their analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 123 performance. according to the results, the manager can develop the necessary action plan and then implemented it to ensure a comfortable environment that encourages employees to get their maximum efforts towards high performances and ensure efficient and effective project execution. in this research work, the data were collected only from employees working in the oil and gas industries in oman. therefore, the analysis and the interpretation of the data may not be applicable to the oil and gas industries in other parts of the world. more research is needed and data needs to be collected from various countries with significant project activities in oil and gas industries is necessary to generalize the research outcome. 6. conclusion and future research directions the main objective of this study is to define and evaluate the effect of motivation on project execution for oil and gas employees in the sultanate of oman. for the purpose, the study was carried out mainly based on the literature review and questionnaire survey. altogether 86 participants took part in the survey. these participants are working at various oil and gas companies and at different levels in the sultanate of oman. it can be figured out from the result that there is a significant relationship between motivation and employees’ performance. the survey results revealed that extrinsic factors are the most influential factor to motivate the employee in their work. salary is found to be the powerful extrinsic motivator for most of the employees. on the other hand, task achievement and challenge in the work are found to be the major intrinsic motivation factors influencing the employee’s performance. the study also analyzed the effect of various design variables on motivation. the results show that irrespective of years of experience, position and organizational type, missing support from the management is found to have a major impact in lowering employees motivation. the result also revealed that motivation is mainly needed during the initial and buildup phase of project execution. as a future research direction, this finding can be considered as a hypothesis that can be expanded to verify up to what extent this is valid and answer why motivation is more important at these phases. furthermore, the employee himself and their counterpart are found to be the main person who can increase the level of employee’s motivation at the workplace. most of the workers working in the oil and gas project in oman seems to have a medium level of motivation at work. this is especially true for those employees who have very few years of working experience. further to this, it will be an interesting extension to anchor the motivational factors and examine how they are related to achieving the project scope and meeting the project schedule while maintaining the project budget. many research articles can be found in literature dedicated to measuring the maturity level of an organization to adopt and use project management tools. however, the scientific literature is still lacking the development of a solid framework capable of measuring the maturity level of the motivation system of the organization. hence, a possible extension of studies concerned about the motivation of employees should consider this direction. such frameworks can be more comprehensive by targeting the different needs of employees during the different phases of the project execution as well as their years of experience and containing a number of related key performance indicators (kpi). moreover, recently piya et al., (2020) have identified various factors and their interrelationships that affect agility level in the oil and gas supply chain. some of the identified factors of agility will have an impact on employee motivation. therefore, the research can be extended to identify the factors of agility and their level of impact on motivating employees in the oil and gas project. conflict of interest the authors declare that there are no conflicts of interest as regards this article. funding this study did not receive any funding from the public, private or not-for-profit sectors. acknowledgment the main author would like to acknowledge the financial support he received from sultan qaboos university during his sabbatical leave when this study was conducted. moreover, the work of the main author has been partially carried out during his sabbatical research visit to the university of nottingham (uk). the author would like to thank dr. ender özcan and the university of nottingham for such an opportunity. references adams, j.s. (1963). toward an understanding of equity. journal of abnormal and social psychology, 67, 422–436. alderfer, c.p. (1972). existence, relatedness, and growth; human needs in organizational settings. free press, new york. al-hinai, n., abusharkh, e., al-shamli, a., & summad, e. (2020). an innovative framework to measure the maturity level of core soft skills and competencies (sscm) needed by project managers in oil and gas industry, 10th annual international conference on industrial engineering and operations management, dubai, uae, pp. 2597-2604 . nasr al-hinai, sujan piya, and khalid al-wardi 124 amin, a., hassan, m. f. b., ariffin, m. b. m. (2010). framework of intrinsic and extrinsic motivators of knowledge sharing: a case of training institutes of an oil and gas company in malaysia, 2010 international symposium on information technology, kuala lumpur, 1428-1432. asproni, g. (2004). motivation, teamwork, and agile development. agile times, 4(1), 8-15. baddoo, n., hall, t., jagielska, d., (2006). software developer motivation in a high maturity company: a case study. journal of software process improvement and practice, 11(3), 219–228. bateman, t.s., snell, s., (1999). management— building competitive advantage 4th ed. mcgraw-hill, boston. beecham, s., baddoo, n., hall, t., robinson, h., sharp, h., (2008). motivation in software engineering: a systematic literature review. information and software technology, 50(9–10), 860–878. chaos report (2014). the standish group report. project smart, https://www.projectsmart.co.uk /whitepapers/chaos-report.pdf (last accessed on nov. 13, 2019). chaos report (2015), the standish group international, https://standishgroup.com/sample_ research_files/chaosreport2015-final.pdf (last accessed on nov. 13, 2019). demarco, t., lister, t., (1999). peopleware: productive projects and teams 2nd ed. dorset house, new york. dolfi, j., andrews, e.j., (2007). the subliminal characteristics of project managers: an exploratory study of optimism overcoming challenge in the project management work environment. international journal of project management, 25, 674–682. dwivedula, r., bredillet, c.n., (2010). profiling work motivation of project workers. international journal of project management, 28, 158–165. ferratt, t.w., short, l.e., (1986). are information systems people different: an investigation of motivational differences. management information systems quarterly, 10(4), 377–387. field, t. (1997). when bad things happen to good projects. cio, 11(2), 54–62. fisher, c.d., xue ya yuan, a., (1998). what motivates employees? a comparison of us and chinese responses. the international journal of human resource management, 9(3), 516–528. germann, m., (2004). influence of project managers on the motivational factors of a project team. paper presented at the 13th international conference on management of technology iamot, washington, dc. april. hall, t., baddoo, n., beecham, s., robinson, h., sharp, h., (2009). a systematic review of theory use in studies investigating the motivations of software engineers. acm transactions on software engineering and methodology, 18(3), 48–77. heimovics, r., brown, f.g., (1976). municipal employee behavior as an exchange process. midwest review of public administration, 10(4), 201–215. herzberg, f. (1987). one more time: how do you motivate employees?. harvard business review 65(5), 109–120. herzberg, f., mausner, b., snyderman, b.b., (1959). the motivation to work. john wiley, new york. hwang, b. g., zhao, x., & ong, s. y. (2015). value management in singaporean building projects: implementation status, critical success factors, and risk factors. journal of management in engineering, 31(6), 1-10. kanagarajoo, m. v., fulford, r., & standing, c. (2019). the contribution of social media to project management. international journal of productivity and performance management. keller, a.e., (1978). personnel positioning. infosystems, 25(6), 50. khan, b. a., jalbani, a. a., (2009). effect of compensation system on employee motivation: a case study of pakistan petroleum limited. journal of independent studies and research – msse, 7(2), 121 – 130. khoshnevis, h, tahmasebi, a. (2016). the motivation system in a governmental organization. procedia social and behavioral sciences, 230, 212 – 218. kim, s. t., & choi, b. (2019). price risk management and capital structure of oil and gas project companies: difference between upstream and downstream industries. energy economics, 83, 361-374. kooij, d., delange, a.h., jansen, p.g.w., kanfer, r., dikkers, j.s., (2010). age and work-related motives: results of a meta-analysis. journal of organizational behavior, 31(8), 1111–1136. kotler, p., armstrong, g., opresnik, m. o. (2018). principles of marketing, 17th edition. pearson education. kovach, k.a., (1987). what motivates employees? workers and supervisors give different answers. business horizons, 30(5), 58–65. kovach, k.a., (1995). employee motivation. addressing a crucial factor in your organization's performance. employment relations today, 22(2), 93–105. labor relations institute, (1946). do you know your workers' wants?. foremen facts ix, 21, 1–3. lawal, a. l., okhankhuele, o. t. (2014). the effect of motivation on project execution in nigeria: a case study of bayelsa plastic industry, yenagoa, baylesa state, nigeria. british journal of applied science & technology, 4(21), 2985-3002. lee-kelly, l., leong, k.l., (2003). turner's five functions of project-based management and situational leadership in it services projects. international journal of project management, 21, analysis of factors affecting motivation in projects: a case study in oil and gas industry in oman 125 583–591. linberg, k.r., (1999). software developer perceptions about software project failure: a case study. the journal of systems and software, 49(2), 177–192. locke, e.a., (1968). towards a theory of task motivation and incentives. organizational behavior and human performance, 3, 157–189. mahaney, r. c., lederer, a. l., (2006). the effect of intrinsic and extrinsic rewards for developers on information systems project success. project management journal, 37(4), 42-54. mak, b.l., sockel, h., (2001). a confirmatory factor analysis of is employee motivation and retention. information & management, 38(5), 265–276. maslow, a.h., (1943). a theory of human motivation. psychological review, 50, 370–396. mcmanus, j., wood-harper, t. (2003). information systems project management: the price of failure. management services, 47(5), 16–19. meredith, m. (2016). project management -a managerial approach. asia: wiley plus learning space, john wiley & sons (asia) pte ltd. monteiro, d. l. c., peñaloza, v., pinto, f. r., coria, m. d. c. d., & calderón, l. m. o. (2015). attitudes towards money and motivational orientation to work in brazilian young workers. contaduría y administración, 60(1), 11-30. mueller, r., turner, r., (2009). leadership competency profiles of successful project managers. international journal of project management, 28(5), 437–448. nchorbuno, d. a. (2011). the role of motivation on employee performance in the public sector: a case study of the university for development studies-wa campus, (doctoral dissertation). nel, p.s., gerber, p.d., van dyk, p.s., haasbroek, g.d., schultz, h.b., sono, t., werner, a., (2001). human resources management 5th ed. oxford university press, cape town, south africa. nicholson, n., (2003). how to motivate your problem people. harvard business review, 81, 56–67. pinder, c.c., (1998). work motivation and organizational behavior. prentice hall, upper saddle river, nj. piya, s., shamsuzzoha, a., & khadem, m. (2019). an approach for analysing supply chain complexity drivers through interpretive structural modelling. international journal of logistics research and applications, 1-26. piya, s., shamsuzzoha, a., khadem, m., & al-hinai, n. (2020). identification of critical factors and their interrelationships to design agile supply chain: special focus to oil and gas industries. global journal of flexible systems management, 21(3), 263-281. rony, n. i., suki, n. m., (2017). modelling the relationships between internal marketing factors and employee job satisfaction in oil and gas industry. asian social science, 13(3), 135-144. sansone, c., harackiewicz, j., (2000). intrinsic and extrinsic motivation — the search for optimal motivation and performance. academic press, san diego, ca. seiler, s., lent, b., pinkowska, m., & pinazza, m. (2012). an integrated model of factors influencing project managers' motivation—findings from a swiss survey. international journal of project management, 30(1), 60-72. sekhar, c., patwardhan, m., & singh, r. k. (2013). a literature review on motivation. global business perspectives, 1(4), 471-487. sharp, h., baddoo, n., beecham, s., hall, t., robinson, h., (2009). models of motivation in software engineering. information and software technology, 51(1), 219–233. shih, t.-h., fan, x., (2007). response rates and mode preferences in web-mail mixed-mode surveys: a meta-analysis. international journal of internet science, 2(1), 59–82. skinner, b.f., (1969). contingencies of reinforcement: a theoretical analysis. prenticehall, englewood cliffs, nj. tabassi, a. a., ramli, m., abu bakar a. h. (2011). training, motivation and teamwork improvement: the case of construction firms. african journal of business management, 5(14), 5627-5636. tampoe, m., thurloway, l., (1993). project management: the use and abuse of techniques and teams (reflections from a motivation and environment study). international journal of project management, 11(4), 245–250. taylor, f.w. (1903), shop management (new york, ny: american society of mechanical engineers), oclc 2365572. project gutenberg. turner, j.r., mueller, r., (2003). on the nature of the project as a temporary organization. international journal of project management, 21, 1–8. van der kolk, b., van veen-dirks, p. m., & ter bogt, h. j. (2019). the impact of management control on employee motivation and performance in the public sector. european accounting review, 28(5), 901-928. verma, v.k., (1996). human resource skills for the project manager: the human aspects of project management, volume two. project management institute, upper darby, pa. vroom, v.h., (1964). work and motivation. john wiley & sons, new york. wiley, c., (1997). what motivates employees according to over 40 years of motivation surveys. international journal of manpower, 18(3), 263– 288 1. introduction 2. the need for motivation 3. research methodology and data collection 4.1. analysis and validations of collected data 4.2. effects of intrinsic and extrinsic motivation 4.3. effects of design variables on motivation 6. conclusion and future research directions conflict of interest funding references mech060826.qxd the journal of engineering research vol. 5, no.1 (2008) 30-36 notation 1. introduction in industrial operations, various augmentation techniques are in practice, which reduce equipment sizes and increase out-puts without impairing the efficiency of pro ___________________________________________ *corresponding author’s e-mail: feroz@caledonian.edu.om duction. impinging jet systems have been well established as an effective technique for heating, cooling, or drying a target surface and have a variety of industrial applications, including the tampering of glass, the drying of film and textile, and the cooling of hot steel plate and gas turbine cascade. recently, the jet impingement has also been used for the cooling of microelectronic components. literature reveals several investigations (gardon and cobonpue, 1962; gardon and akfirat, 1965; korger and krizek, 1966; subba rao et al. 1973; coeuret, 1975; chin and tsang, 1978; chang et al. 1995, kendoush, 1998; lee et al. 2002; travnicek and tesar, 2003) on heat/mass transfer using jets in open and closed containers. however, studies on the effect of submerged impinging jets in closed cylindrical cell in the presence of solids are meager. the presence of solids provides a large surface of contact and causes rapid removal of heat in electro-chemical processes such as electrochemical surface treatment of selective area, annealing of metals and plastic sheets. the present study is directed towards the investigations on mass transfer with impinging submerged jets in closed cylindrical cell in the presence of solids and correlation of data in impingement and decreasing coefficient regions. mass transfer studies with submerged impinging jets in closed cylindrical electrolytic cell in the presence of solids s. feroz *1, b.v.l. rao2, c. bhaskarasarma3 and v.s.r.k. prasad3 1 department of m&ie, caledonian college of engineering, sultanate of oman 2 department of chemical engineering, govt. polytechnic college, visakhapatanam 530 012, india 3 department of chemical engineering, andhra university, visakhapatanam 530 003, india received 26 august 2006; accepted 11 february 2007 abstract: an experimental study of mass transfer in forced convective flow of fluid electrolyte through submerged jets impinging normal to the target surface in a closed cylindrical cell in the presence of solids (porcelain beads) is reported. the pertinent dynamic and geometric variables of this study are flow rate, diameter of the nozzle, height of the nozzle from the target surface and solids fraction. the mass transfer measurements, made by the electrochemical method propose empirical correlations in the impingement and decreasing coefficient regions. keywords: impingement region, decreasing coefficient region, mass transfer coefficient, target surface cf correlation factor dc diameter of the cell, m dp diameter of solids, m dj diameter of the nozzle, m e1 toe7 electrodes on target surface h height of the nozzle from target surface, m jd mass transfer factor kl mass transfer coefficient, m/s nre reynolds number vd jρ µ χ solids fraction x average radial distance of electrode from the target surface, m v velocity of fluid electrolyte from the nozzle, m/s .áñ∏°üdg ogƒÿg oƒlh ‘ á≤∏¨ÿg á«fgƒ£°se’g á«∏ÿg ‘ iqƒª¨ÿg ωg~£°u’g òagƒf ™e ioéÿg ∫é≤àfeg äé°sgqo rhòa.¢s1*hgq é檰ûµd .± .ü ,2qé°sôh .∑ .q .¢s .± h 3éeô°ugôµ°ùh .¢s,3 áá°°uuóóÿÿggoƒlh ‘ á≤∏¨e á«fgƒ£°sg á«∏ÿ iôfgo í£°ùh áe~£°üÿg iqƒª¨ÿg iqƒaéædg ∫ón øe …ô°ù≤dg »∏ª◊g πfé°ùdg ≥a~j â«dhîµdg ‘ ioéÿg ∫é≤àfg ∫héæàj ájèàfl á°sgqo : áñ°ùfh ,±~¡dg í£°s øe üƒñfe’g ágƒa ´éøjqg ,üƒñfc’g ágƒa ô£b ,≥a~àdg ∫~©e πª°ûj á°sgq~dg √ò¡h á∏°üdg ägp á«°s~æ¡dgh ádé©ødg ägò¨àÿg ¿g .(rôÿg ±õn) áñ∏°üdg ogƒÿg .πeé©ÿg ¢ü≤fh ωg~£°ug ≥wéæã á«ñjôœ äébóy ìî≤j ,á«f髪«chô¡µdg á≤jô£déh âjôlcg »àdg ,ioéÿg ∫é≤àfg äé°sé«b ¿g .áñ∏°üdg ogƒÿg ääggooôôøøÿÿggáá««mmééààøøÿÿgg.±~¡à°ùÿg ≈≤∏◊g í£°ùdg ,ioéÿg ∫é≤àfg πeé©e ,πeé©ÿg ¢übéæj á≤£æe ,ωg~£°u’g á≤£æe : 31 the journal of engineering research vol. 5, no.1 (2008) 30-36 2. experimental set-up and procedure an electrolytic cell of 0.1475m diameter and 0.173m height fabricated with pvc has been employed for studying the effects of various parameters. a schematic diagram of the apparatus is shown in fig. 1. the set-up mainly consists of a storage tank (t), a pump (p), a rotameter r , an electrolytic cell (c), and a selector switch (s). the storage tank is made of a copper sheet with a capacity of about 0.150 m3 and is provided with a drain valve (v1) at the bottom. the outlet line at the bottom of the tank is connected to the suction side of the pump and in between a valve (v2) was provided to isolate the pump from the storage tank whenever necessary. a 0.746 kw motor is used to drive the pump, which circulates the electrolyte solution. the outlet from the pump is divided into two lines one for direct entry of electrolyte into the cell through a rotameter and the other to serve as a by-pass which returned to the storage tank, the by-pass being controlled by valve (v3). three convergent nozzles of diameters 0.002112 m, 0.005623 m and 0.007831 m were fabricated using copper tube. nozzle is provided with a screw threading to fit with the inlet tube. a rotameter of fischer and porter make, with a metering range of 0 3.33x10-4 m3/s has been used to measure the electrolyte flow rate. the cell is fixed between two circular hylam plates of the same dimensions. the top plate is provided with two holes, one for the insertion of nozzle and the other for exiting the electrolyte and returning to the electrolyte storage tank. copper tubes are provided as inlet and outlet for the electrolyte. the bottom plate served as "target surface". concentric ring electrodes (e2 to e7) are cut out of copper sheet of 0.00125 m thick and fixed at equidistant ie. 0.01 m apart, flush with the inside surface. the central electrode (e1) is a small disc cut out of a copper rod of 0.02 m diameter, fixed at center point of the target plate where the jet impinges. all electrodes are provided with copper terminals for making electrical connections. the inlet copper tube served as the counter electrode. the electrolytic cell is charged with varying quantities of solids (sf) (0.5 % to 3% of total volume), which are a mixture of spherical and cylindrical shaped porcelain beads of 0.005 m diameter and 0.008 m length with a bulk density of 1200 kg/m3 and true density of 1267 kg/m3. nozzles of three different sizes 0.007831 m, 0.005623 m and 0.002112 m are used to generate the impinging jet flow. limiting currents are measured for reduction of ferri-cyanide ion, once the flow rate and temperature was stabilized. the electrochemical procedure adopted for the measurement is same as described in the earlier studies (lin et al. 1951; sakakihara et al. 1994; subba rao et al. 1973). an electric potential was applied in steps across the test and counter electrode and the corresponding currents were noted. the attainment of limiting current was observed by a sharp rise in potential for a small increase in current. the electrolyte consisted of equi-molal solutions of (0.01m) of potassium ferricyanide and potassium ferrocyanide with an excess indifferent electrolyte 0.5n sodium hydroxide. measurements of limiting currents are obtained at different flow rates. for any individual run the v3 e t v1 p v2 r c sf s figure 1. schematic diagram of experimental set-up 32 the journal of engineering research vol. 5, no.1 (2008) 30-36 3. results and discussion the hydrodynamics of the impinging jet flow may be visualized (gardon and cobonpue, 1962; gardon and akfirat, 1965; korger and krizek, 1966; subba et al. 1973; coeuret, 1975; chin and tsang, 1978) as the jet issuing from the nozzle onto the target surface, gradually flows away radially towards the outer edge of the target surface or the confining wall of the electrolyte cell depending upon whether it is free jet or submerged jet. in the presence of solids in closed cells with submerged impinging jets, the flow patterns are complex for any mathematical treatment and hence an experimental modeling has been attempted by the empirical method. the plot of mass transfer coefficients (kl) (calculated from the limiting current density, similar to earlier studies) (lin et al. 1951; sakakihara et al. 1994; subba et al. 1973) for the reduction of ferricyanide ion is drawn against the dimensionless distance 'x/dc', where 'x' is the average radial distance of the ring electrodes from the center of the cell and is shown in fig. 2. a close inspection of the trends in these plots shows the existence of the following regions due to different flow patterns: (1) impingement region wherein the limiting current densities are comparatively very high which can also be referred to as stagnation zone. (2) transition region wherein the limiting current densities decrease with the radial distance of the electrode. (3) wall jet region wherein the limiting current densities decrease relatively very slowly as compared to those in the transition region. the change from region 2 to region 3 is not abrupt but smooth, hence these two region ie. transition and wall jet region are therefore combined to form a "decreasing coefficient region". it is also observed that only central electrode (x/dc = 0) falls into the impingement region. 3.1 effect of flow rate on mass transfer coefficients the mass transfer coefficient data of e1, e3 and e7 are plotted against the flow rates as shown in the fig. 3. it has been found that the values of kl increase gradually with an increase in the flow rate, due to high turbulence on the target surface both in the case of impingement (e1) and decreasing coefficient regions (e3 and e7). temperature remained constant within ± 0.1 0c and if the temperature varied more than ± 0.1 0c, the run was discarded. the electrolyte was analyzed for each and every run and the reproducibility of data was tested from time to time by repeating one of the previous runs under identical conditions. the diameters of nozzles, height of the nozzle from the target surface and solid fraction ar e varied and experiments are repeated for each case. the ranges of varia bles cove red are given in table 1. s.no variable minimum maximum 1. flow rate (qx10 5), m3/s 3.4 14.0 2. velocity (v), m/s 0.705 40.0 3. height of the nozzle from the target surface, (hx102), m 1.0 8.0 4. diameter of the nozzle, (dj x102), m 0.2112 0.7831 5. average radial distance of the electrode on the target surface from the center, (x*102), m 0.0 6.328 8. reynolds number, nre 6707 102405 9. schmidt number, sc 698.6 931.9 table 1. ranges of variables covered in the present day 0 0.5 1 1.5 0 0.1 0.2 0.3 0.4 0.5 x/dc k l x1 05 m /s qx105, m3/s 3.4 14.0 figure 2. variation of mass transfer coefficients with x/dc (dp: 5x10-3 m, λ: 0.01, dj: 0.7831x10-2 m, h: 2.5x10-2 m) 0.01 0.51 1.01 1.51 2.01 1 6 11 16 qx105, m3/s k l x1 05 , m /s figure 3. variation of mass transfer coefficients with flow rates (dp: 5x10-3 m, λ: 0.01, dj: 0.5623 x 10-2 m, h: 2.5x10-2 m) 33 the journal of engineering research vol. 5, no.1 (2008) 30-36 3.2 the effect of height of nozzle from target surface on mass transfer coefficients the effect of nozzle height from the target surface on the mass transfer coefficients is shown in figs. 4a and 4b for three different heights covered in the present study ie. 0.01 m, 0.04 m and 0.08 m. it has been found that with an increase in the height of the nozzle from the target surface, the values of the mass transfer coefficient decreases. the variation of mass transfer coefficients with nozzle height is found to be marginal in both impingement (e1) and decreasing coefficient regions (e7), as shown in figs. 4a and 4b, respectively. as the height increases the liquid jets that emerge out from the nozzle are loosing the impact of their momentum and not fully reach the target surface where the mass transfer occurs. similar observations were reported in earlier studies on single, open free and submerged jets (gardon and cobonpue, 1962; gardon and akfirat, 1965; korger and krizek, 1966; subba et al. 1973; coeuret, 1975). 3.3 the effect of nozzle diameter on mass transfer coefficients the values of mass transfer coefficients of e1 and e7 are plotted against velocity for three different nozzle diameters keeping other parameters constant as shown in figs. 5a and 5b. it has been found that in both regions kl increases with an increase in the nozzle diameter for a given velocity due to large scale eddy turbulence on the target surface. 3.4 the effect of solids on mass transfer coefficients the variation in the mass transfer coefficients of e1 and e7 with the percentage of solids presence is shown in figs. 6a and 6b. the percentage of solids varied from 0.5% to 3% of volume of the cell. mass transfer coefficients are found to increase marginally with the increase in solids up to 1% and then decreases in both impingement 0.01 1.01 2.01 0.1 0.6 1.1 1.6 2.1 2.6 3.1 v, m/s k l x1 05 , m /s figure 4a. variation of mass transfer coefficients with velocity for different heights of nozzle from target surface in impingement region ( e1, dp: 5x10 -3 m, χχ :0.01, dj: 0.21120x10 -2 m) 0.00 0.05 0.10 0.01 1.01 2.01 3.01 4.01 5.01 v, m/s k l x1 05 , m /s figure 4b. variation of mass transfer coefficients with velocity for different heights of nozzle from target surface in decreasing coefficient region ( e7, dp: 5x10 -3 m, χχ: 0.01, dj: 0.21120x10 -2 m) 0.50 1.00 1.50 2.00 0.1 10.1 20.1 30.1 40.1 v, m/s k l x1 05 , m /s figure 5a. variation of mass transfer coefficients with velocity for different nozzle diameters in impingement region ( e1, dp: 5x10 -3 m, χχ: 0.01, h: 2.5x10-2 m). 0.01 0.06 0.1 10.1 20.1 30.1 v, m/s k l x1 05 , m /s figure 5b. variation of mass transfer coefficient s with velocity for different nozzle diameters in decreasing coefficient region (e7, dp: 5x10 -3 m, χχ: 0.01, h: 2.5x10-2 m) 34 the journal of engineering research vol. 5, no.1 (2008) 30-36 and decreasing coefficient regions. it has been observed that at a lower solid fraction the scouring action together with the hydrodynamic turbulence increased the mass transfer coefficients marginally. at a higher solid fraction, the effect of solids is found to decrease gradually resulting in lower coefficients. 4. comparison figure 7 shows the variation of mass transfer coefficients with and without the presence of solids. plot 'a' showed the data of the present study (density: 1200 kg/m3). plot 'b' gave the data on mass transfer coefficients in the absence of solids while plot 'c' gave the data in the presence of high-density solids (density: 2320 kg/m3). the plots showed that low-density solids (density: 1200 kg/m3) as in the case of the present study gave improvements up to 3 folds over the coefficients obtained in the absence of solids. the plots also revealed that the coefficients with low density solids (density: 1200 kg/m3) were found to be 200 times greater than those obtained with high density solids (density: 2320 kg/m3). solid-fluid mixing appears to be more significant and severe with low-density solids, favoring a scouring action as well as hydrodynamic turbulence on the transfer surface resulting in higher coefficients. figure 8 shows the comparison of the coefficients obtained from the study on heat transfer data on turbulent submerged liquid jets of (chang et al. 1995), single submerged jets in open containers of (rao et al. 1971), free multi-jets data of venkateswarlu and (raju, 1979) with those obtained from the present data in the impingement region. the coefficient data of the present study (coeuret, 1975) however are found to be higher than those obtained with turbulent submerged liquid jets (heat transfer), single submerged jets and free multi-jets. 0.50 1.00 1.50 0.1 0.6 1.1 1.6 2.1 2.6 3.1 3.6 4.1 4.6 5.1 v, m/s k l x1 0 5 , m /s figure 6a. variation of mass transfer coefficients with velocity for different solids fraction in impingement region ( e1 dp: 5x10 -3 m, dj: 0.7831x10 -2 m, h: 2.5x10 -2 m) 0.01 0.06 0.11 0.16 0.1 0.6 1.1 1.6 2.1 2.6 3.1 3.6 4.1 4.6 5.1 v, m/s k l x 1 0 5 , m /s figure 6b. variation of mass transfer coefficients with velocity for different solids in decreasing coefficient region ( e7 dp: 5x10-3 m, dj: 0.7831x10 -2 m, h: 2.5x10 -2 m) 0.00 0.10 0.20 0.30 0.40 0 0.1 0.2 0.3 0.4 0.5 x/dc k l x 1 05 ,m /s k lx 1 0 5 , m / s x/dc figure 7. comparison plots of present study (pl ot ‘a’ density: 12 00 kg/m3) with the higher density solids (plot ‘c’ density: 232 0 kg/m3) and without the presence of solids (plot ‘b’) (dp: 5x10 -3 m, χχ: 0.01, h: 2.5x10 2 m, dj: 0.7831x10 -2 m, q: 8.75x10 -5, m) 0.00 0.50 1.00 1.50 1.00 11.00 qx105, m3/s k l x 1 0 5 , m /s figure 8. comparison plots of present study with chang et al. 1995 [turbulent subme rged single jets, heat transfer] , subba rao et al. (1973) [single s ubmerged jets in open container] , venkateswarlu and raju (1979) [free multi-jets in open container] 35 the journal of engineering research vol. 5, no.1 (2008) 30-36 5. correlations developed earlier investigations (gardon and cobonpue, 1962; gardon and akfirat, 1965; korger and krizek, 1966; subba rao, et al. 1973; chang et al. 1995) on heat and mass transfer in free and submerged impinging jets used conventional nu/jh-re or sh/jd-re type of correlations. similarly, the correlations are developed for mass transfer coefficient data, taking into account the effect of design and operating variables in the present study. the following format of the equations has been used to develop generalized correlations by the regression analysis. (1) the data on reduction of ferri-cyanide ion in the impingement region yielded the following equations of correlation: for the impingement region: (2) (3) for the decreasing coefficient region: (4) (5) the above equations are valid for the ranges of variables covered in the present study. the plots of the data correlated in accordance with the eqs. 2 to 5 (correlation factor versus flow reynolds number) are shown in figs. 9-12. (6) jd = c (nre) n1 (φ1)n2 (φ2)n3 (χ)n4 jd = 203.43 ( nre) –0.92 (h/dc) -0.038 (dj/dc) 1.06 (χ) 0.151 for χ ≤ 0.01 jd = 28.41 ( nre) –0.93 (h/dc) -0.051 (dj/dc) 1.07 (χ)-0.312 for χ ≥ 0.01 jd =7.10 (nre) –0.96 (h/dc) -0.042 (dj/dc) 1.08 (χ)0.057 (x/dc) -0.804 for χ ≤ 0.01 jd =2.43 (nre) –0.95 (h/dc) -0.018 (dj/dc) 1.08 (χ)-0.18 (x/dc) -0.814 for χ ≥ 0.01 correlation factor (cf)=j d / ( (φ1 ) n 1 (φ2 ) n 2 (φ2 ) n 3 (χ) n4) 0.00 0.01 0.10 1.00 1000 10000 100000 1000000 nre cf figure 9. correlation plot for impingement region for χ ≤ 0.01 0.00 0.00 0.00 0.01 0.10 1.00 1000 10000 100000 1000000 nre cf figure 10. correlation plot for impingement region for χ ≥ 0.01 0.00 0.00 0.00 0.01 0.10 1000 10000 100000 1000000 nre cf figure 11. correlation plot for decreasing coefficient region for χ ≤ 0.01 36 the journal of engineering research vol. 5, no.1 (2008) 30-36 6. conclusions the correlations developed can be employed in the design of the electrochemical cell with submerged impinging jets in the presence of solids. the following inferences have been drawn from the results of the study: (i) the mass transfer coefficients is found to increase with an increase in the flow rate both in impingement and decreasing coefficient regions as a result of an increase in turbulence at the target surface (ii) the mass transfer coefficients is found to increase with an increase in the nozzle diameter at a given velocity in both the regions, due to large scale turbulence at a given velocity (iii) the mass transfer coefficients showed a decrease in their value with the increase in the height of the nozzle from the target surface as the impact of jets did not fully reach the target surface due to local flow interactions. the decrease is found to be marginal in both the regions, and (iv) the mass transfer coefficients initially showed an increase in value with the increase in solids fraction up to 1.0% and then decreased in both regions. it has been observed that at a lower solid fraction the scouring action together with the hydrodynamic turbulence increased the mass transfer coefficients marginally. at higher solid fraction, the effect of solids is found to decrease gradually resulting in lower coefficients. references coeuret, f., 1975, "transfert de matierelors del'impact normal de jets liquides circulaires immerges," j. of chemical engineering science, vol. 30, pp. 12571263. chin, d.t. and tsang, c.h., 1978, "mass transfer to an impinging jet electrode," j. of electrochemical society, vol. 125(9), pp. 1461-1469. chang, c.t., kojasoy, g. and landis, f., 1995, "confined singleand multiple-jet impingement heat transferi. turbulent submerged liquid jets," int. j. of heat and mass transfer, vol. 38(5), pp. 833-842. gardon, r. and cobonpue, j., 1962, "heat transfer between a flat plate and jets of air impinging on it," int. developments in heat transfer, asme, new york, pp. 454-460. gardon, r. and akfirat, j.c., 1965, "the role of turbulence in determining the heat transfer characteristics of impinging jets," int. j. of heat and mass transfer, vol. 8, pp. 1261-1272. korger, m. and krizek, f., 1966, "mass transfer coefficient in impingement flow from slotted nozzles," int. j. of heat and mass transfer, vol. 91, pp. 337-344. kendoush, a.a., 1998, "theory of stagnation region heat and mass transfer to fluid jets impinging normally on solid surfaces," chemical engineering and processing, vol. 37, pp. 223-238. lin, c.s., denton, e.b., gaskill, h.s. and putnam, g.l., 1951, "diffusion controlled electrode reactions," industrial and engineering chemistry, vol. 43(9), pp. 2136-2143. lee, d.h., lee, y.m., kim, y.t., won, s.y. and chung, y.s., 2002, "heat transfer enhancement by perforated plate installed between an impinging jet and the target plate," int. j. of heat and mass transfer, vol. 45, pp. 213-217. sakakihara, y., flora, j.r.v., suidan, m.t. and kuroda, m., 1994, "measurement of mass transfer coefficients with an electrochemical method using dilute electrolyte solutions," j. of water research, vol. 28(1), pp. 9-16. subba rao, b., krishna, m.s. and jagannadha raju, g.j.v., 1973, "ionic mass transfer with submerged jets-i," periodica polytecnica chemical engineering, vol. 17(2), pp. 185-195. travnicek, z. and tesar, v., 2003, "annular synthetic jet used for impinging flow mass transfer," int. j. of heat and mass transfer, vol. 46, pp.3291-3297. venkateshwarlu, p. and jagannadha raju, g.j.v., 1979, “ionic mass transfer with imping multi-jets, indian j. technol, vol. 17(1), pp. 1-4. 0.00 0.00 0.00 0.00 0.01 0.10 1000 10000 100000 1000000 nre cf figure 12. correlation plot for decreasing coefficient region for χ ≥ 0.01 the journal of engineering research (tjer), vol. 18, no. 1, (2021) 26-35 *corresponding author’s e-mail: madawy@su.edu.sa doi: 10.53540/tjer.vol18iss1pp26-35 stacked delta design of three-phase permanent-magnet fault current limiters mohamed eladawy 1, 3*, and ibrahim a. metwally 2, 3 1 electrical engineering department, college of engineering, shaqra university, riyadh, saudi arabia 2 department of electrical and computer engineering, college of engineering, sultan qaboos university, muscat, oman 3 electrical engineering department, faculty of engineering, mansoura university, mansoura, egypt abstract: this paper proposes an improvement for the dynamic performance of pre-saturated stacked permanent magnet biased three-phase fault current limiter (pmfcl) through comsol finite element simulation. the nonlinear demagnetization behaviour of the permanent magnet, especially in the upper part of the b-h curve with negative magnetic field intensity, has been modelled through the jiles-atherton method. this enables a realistic representation of the pmfcl dynamic behaviour throughout its entire operations of pre-fault, fault and fault removal, respectively. the experimental measurements were considered to validate the trends of the simulation outcomes during the entire operation of pmfcl. extensive finite element simulation shows that the stacked design of pmfcl can increase the capability of fault current limiting with proper selection of the number and arrangement of the ac coils around the iron core (soft magnet). results reveal that the division of ac coils into series differential connected sub coils, with an even number, can increase the limiting capability by increasing the ac coil number of turns, without exceeding the permissible tolerances of voltage drop and power losses. moreover, this stacked design is subjected to parametric investigation for different fault types, either symmetrical or unsymmetrical, or even when changing the fault current peak value. keywords: comsol multiphysics; iron core, neodymium; permanent magnet; pre-saturated core; fault current limiter. ات الدائمةیلمحددات تیار الخطأ تعتمد على المغناطیس ثالثى األوجھ مكدس دلتاتصمیم يإبراھیم متول و *يمحمد العدو یقترح ھذا البحث تحسینًا لألداء الدینامیكي لمحدد التیار الخطأ ثالثي األوجھ سابق التشبع بواسطة المغناطیسیات الملخص: تم نمذجة سلوك إزالة المغناطیسیة الالخطیة .comsol من خالل محاكاة العناصر المحدودة (pmfcl) الدائمة فى الحالة السالبة لشدة المجال المغناطیسي، من خالل طریقة bh للمغناطیس الدائم ، خاصة في الجزء العلوي من منحنى jiles-atherton والتى تتیح تمثیًال واقعیًا للسلوك الدینامیكي pmfcl خالل عملھ. تم مقارنة القیاسات العملیة مع نتائج اسعة أن التصمیم وذلك للتحقق من صحة النتائج. تُظھر محاكاة العناصر المحدودة الو pmfcl المحاكاة خالل عمل یمكن أن یزید من عملیة تحدید تیار الخطأ وذلك باالختیار المناسب لعدد وترتیب ملفات التیار المتردد pmfcl المكدس لـ حول القلب الحدیدي (المغناطیس الناعم). تكشف النتائج أن تقسیم ملفات التیار المتردد إلى ملفات فرعیة متصلة بطریقة رقم زوجي ، یمكن أن یزید قدرة تحدید التیار مع زیادة عدد لفات ملف التیار المتردد ، دون تجاوز الحد تفاضلیة متسلسلة ، ب المسموح بھ للھبوط فى الجھد والقدرة المفقودة. عالوة على ذلك ، یخضع ھذا التصمیم للتحقق من تحدید تیار الخطأ لكل .تماثلة ، أو حتى عند تغییر القیمة العظمى لتیار الخطأأنواع األخطاء المختلفة ، سواء كانت متماثلة أو غیر م .أمحدد تیار الخط ؛قلب سابق التشبع ؛مغناطیس دائم ؛نیودنیوم ؛القلب الحدیدي ؛كومسولحزمة برامج الكلمات المفتاحیة: mailto:author@email.com stacked delta design of three-phase permanent-magnet fault current limiters 27 1. introduction the fault current levels may be increased due to the progressive penetration of distributed generation into electrical power networks. consequently, the power system reliability may be threatened due to these increased levels of fault current forthcoming the switchgear approaching limits. therefore, upgrading infrastructure, circuit breakers, and/or survival switchgear represent a great challenge for utilities to conserve the reliability and full protection of power system against high levels of such fault current. this process of upgrading and/or bus splitting procedure can be considered a costly solution and bulky time consuming (duggan 2006). therefore, limiting fault current within the switchgear operating limits represents a leading candidate solution to avoid this upgrading process, which raises the importance of fault current limiter (fcl) technology (eladawy et al. 2018; moscrop 2013). fcls can be generally classified, according to their constructive technology, into solid-state fcls (ssfcls) (radmanesh et al. 2016), quench-type superconducting fcls (sfcls) (jia et al. 2016), and pre-saturated iron core fcls (pcfcls) (eladawy et al. 2018; moscrop 2013; eladawy et al. 2019; eladawy et al. 2020; pannu et al. 2013; yuan et al. 2017; xiao et al. 2013, zhao et al. 2007). ssfcls utilize the power semiconductor switches for controlling the current flow through the circuit. an external controlling circuit, in general, is required for the detection of both fault and post-fault recovery events. however, the associated high power losses and unavoidable delays in fault detection and clearing process limit the reliability of this type of fcls. in the quench-type sfcl, fault current limiting takes place due to the changeover from the initial state of very low impedance associated with superconductivity to very high impedance during fault occurrence. however, the sfcls have operational difficulties, which limit their application, due to the required accompanying cryogenics, for preventing burnout of superconductor, and unavoidable delays in both fault reaction and post-fault recovery events (radmanesh et al. 2016). pcfcls offer a self-governing performance of triggering for both fault occurrence and post-fault recovery events. the soft magnet (iron core), in such pcfcls, is deeply saturated, which represents negligible impedance as the relative permeability µr ≈ 1 during the healthy operating condition. this deeper saturation state can be offered by either a dc coil or permanent magnet (pm) (eladawy et al. 2019, eladawy et al. 2020). during fault condition, the highly counter-magnetic flux created by the fault current can in turn desaturate the soft magnet towards higher relative permeability values, which corresponds to a sharp increase of the soft magnet impedance and hence limiting the fault current. however, the dc coil can be either a simple nonsuperconducting one (i.e. copper coil) or a superconductor that permanently conserves its superconductivity state during the entire operation of the device. consequently, this dc coil generates a highly magnetomotive force (mmf), which can be considered as the main drawback of such dc-biased pcfcls due to the high power losses and high induced back electromotive force (emf) across the dc coil terminals during normal and fault conditions, respectively (moscrop 2013). permanent magnet biased fcls (pmfcls) can replace and overcome the above-mentioned drawbacks of the traditional dc-biased pcfcl. in the three-phase pmfcl, the dc coils are replaced with pm for effective saturation of the associated soft magnet. moreover, the simple construction of pm and its adequate behaviour for transition between healthy and faulty conditions encourage the spreading out of such pmfcls, especially for increasing their operational capabilities with stacked designs. however, the truthful representation of pm with its nonlinear demagnetizing behaviour, especially in the upper part of its constructive material b-h loop with negative magnetic field intensity, is a challenge for the design process of such pmfcls (eladawy et al. 2019). consequently, finite element (fe) simulation provides an attractive tool, especially when validated with empirical results, for profoundly realizing the dynamic behaviour and tendencies of such pmfcls. this paper proposes a time-domain, threedimensional (3d) fe simulation of comsol multiphysics package (comsol 2013) of a stacked design of three-phase pmfcl with performance investigation through, electric-circuit magnetic-field coupled model. however, the nonlinear demagnetizing representation of pm material, especially in the upper part of its b-h hysteresis loop, is based on the well-known jiles-atherton (ja) modelling method. the aftermaths of ja model have been coupled with comsol to comprise the constitutive b-h characteristics of pm exhibits nonlinear demagnetizing behaviour. predominantly, the fe modelling technique has been verified and validated through comparison with experimental results of a given pmfcl. furthermore, the proposed stacked design of pmfcl is devoted to the parametric investigation, especially the governing ones, to increase the operational capability in limiting fault current for distribution networks at different voltage levels. mohamed eladawy and ibrahim a. metwally 28 2. design topology and operating principles 2.1 design topology since the first conception of pcfcl, numerous system geometries have been inspected. regardless of the different design topologies, the principle of operation remains the same. fig. 1 shows the pmfcl included in the simplified test circuit (per phase), where the fault is represented by a switch (eladawy et al. 2018). this switch is used to simulate a solid short circuit across the load terminal. some of the common electrical parameters for the proposed delta-shaped pmfcls are shown in table 1. the design arrangement of four ac coils and five ndfeb-n42 pm segments (4ac-5pm) three-phase delta-shaped pmfcl (eladawy et al. 2020) is shown in fig. 2(a). in this design, the ac coil of each phase has been divided into an even number of sub coils, conserving the total number of turns for each phase, as shown in table 1. these sub coils are equally distributed and wound on the periphery of soft magnet, which is represented by 1000 mm of side length equilateral triangle, forming the delta-shaped core, as shown in fig. 2(a). the subdivided coils of each phase are connected in series, with differential connection (eladawy et al. 2020), to carry the line current of the protected power line. the uniformly distributed pm segments are sandwiched and arranged between the two soft magnet triangles throughout the delta-shaped design. the magnetization of each two adjacent pms are in opposite directions, see parallel arrows in fig. 2(a), in such a way to ensure forcing the soft magnet into a deeper magnetic saturation state. this leads to the soft magnet with very low impedance throughout the system normal operating condition. additionally, this distribution of pms in the magnetic circuit permits their force sources to bridge the distance between the soft magnets, hence resulting in a flux path with low reluctance. this supports effective saturation and pm protection during the healthy condition. furthermore, this ensures effective fault current limiting in fault condition due to the high level of flux linkage crossing the magnetic circuit “demagnetization”. accordingly, the coils flux path is perpendicular to the pm magnetization direction, and along the soft magnet circumference track. moreover, each two adjacent pm pieces have opposite magnetization directions, as shown in fig. 2(a). this allows flux to return through the upper and lower triangles of the soft magnet. to prevent pm demagnetization, it is recommended to locate both the static mmf source and the ac mmf source in orthogonal positions, and include the pm into fcl magnetic circuit (yuan et al. 2018). the dimensions of pm and soft magnet are selected such that the soft magnet is deeply saturated, which results in ac coils per phase with negligible inductance value during the healthy operating conditions. in addition, a substantial increase in the inductance per phase for the ac coil is obtained as a result of the significant rise of the soft magnet relative permeability, and hence its inductance, at fault occurs due to the demagnetization effect. however, the ac field induces eddy currents in the rare-earth neodymium pm (ndfeb-n42). such an eddy current can be minimized by laminating the two equilateral triangles of the soft magnet core and shielding the electrically conductive pm. this will in turn minimize or prevent the pm demagnetization due to heat buildup. the operational capability of pmfcl can be increased using stacked arrangements (one top of the other) of the soft magnet triangles. this stacked arrangement results in doubling the core thickness, as shown in figs. 2(b) and 2(c). this will in turn increase the voltage drop value, due to the increase of ac coil inductance, during the healthy condition of operation. however, this may exceed the maximum permissible value of the voltage drop (≈ 1%) stipulated by the utilities (eladawy et al. 2020). another multi-storey design can be obtained by aligning the magnetization direction in all storeys in the same direction, as shown in fig. 2(d). in this design, the number of pm segments has been doubled and each head-to-head pm couples have the same direction of magnetization and opposite to the adjacent ones in opposite direction, as shown in fig. 2(d). each ac coil around the soft magnet is subdivided to eliminate the leakage figure 1. simplified test circut (per phase) comprising the pmfcl with fau it representation switch. table 1. common electrical parameters for delta shaped pmfcls. symbol quantity value us the rms value of the source line voltage 220√3 v rline line resistance 0.1 ω rload load resistance 3 ω nac total number of ac coil turns of each phase 48 acoil copper wire cross-sectional area 60 mm2 σ coil wire conductivity 6×107 / f power frequency 50 hz stacked delta design of three-phase permanent-magnet fault current limiters 29 flux through the non-magnetic separation between the two soft magnets, forming the 12ac-10pm multistorey delta design. therefore, the number of pms, the height of the ac coil and its sub coils number strictly control the dynamic behaviour of the pmfcl. 2.2 operating principles the operating principle of pmfcl remains the same as that of the dc-biased pcfcl. the rare-earth pm drives the soft magnet into deeper saturation with negligible inductance. therefore, the pmfcl exhibits a non-limiting effect during the healthy condition of the system. figure 2 illustrates that the pm is divided into smaller parts and distributed over the periphery of the soft magnet. this distribution provides effective saturation and reduces the heat build-up inside the pm, hence avoiding the decrease of its magnetism at higher temperature (coey 1996). in general, for each phase, two identical ac coils are required for effective limiting the two half cycles of the fault current. however, the even number of subdivided coils are connected in series, for each phase, with the differential connection. this differential connection decreases the voltage drop during the non-limiting condition of the pmfcl (eladawy et al. 2018; moscrop 2013; eladawy et al. 2019, eladawy et al. 2020). the flux produced by the normal line current, during the normal operation, is sufficiently small to develop satisfactory demagnetizing mmf, which is able to eliminate the strong magnetization of the soft magnet. thus, the pmfcl approximately displays non-limiting behaviour. during fault occurrence, the highly generated mmf, by the fault current, can demagnetize the soft magnet towards higher values of µr in the b-h curve linear region. a one-half cycle of fault current initiates half of the ac sub coils to come out of their initial strong saturation. therefore, they are sharply rushed into a higher value of impedance. the subsequent fault current second half cycle drives the other part of the ac sub coils into a deeper saturation state, and vice versa, as shown in fig. 3. at the fault removal instant, the pmfcl promptly recovers the initial state of strong saturation with negligible impedance and no limiting effect in less than 2 ms (eladawy et al. 2020, yuan et al. 2017). (a) (b) (c) (d) figure 2. three-dimensional view of three-phase pmfcls with arrow directions of pms magnetizations (a) 4ac-5pm reference delta-shaped design (eladawy et al. 2020), (b) 4ac-5pm stacked delta, (c) 8ac-5pm stacked delta, and (d) 12ac-10pm multi-storey delta design. all dimensions are in mm. mohamed eladawy and ibrahim a. metwally 30 nevertheless, the delta-shaped topology of the soft magnet provides magnetic symmetry configuration, and easier in manufacturing and assembly. this will, in turn, result in an approximately similar clipping ratio of fault current for any type of fault, and a similar voltage drop for the three phases in the healthy condition as the device is used in series without disturbing the balance condition of the power system. the nonlinear b-h hysteresis loop is represented through ja method (prigmore et al. 2013, pop et al. 2011), and is considered as a base for further modelling of rare-earth neodymium pm (ndfebn42) behaviour. more details about the characterizing b-h hysteresis loop of ndfeb-n42 and the methodology of ja model can be found in (eladawy et al. 2019; eladawy et al. 2020), (prigmore et al. 2013, pop et al. 2011). ja method can effectively be used for modelling the anisotropic magnets exhibit remnant magnetic flux density. accordingly, ja model is integrated with the fe simulation of comsol, in the time domain, for representing the nonlinear behaviour of pm, especially in the upper part of its b-h hysteresis loop with negative h (eladawy et al. 2019, eladawy et al. 2020). definitely, the constitutive b-h relation represents the pm material using ja hysteresis model has been integrated into comsol (comsol 2013). consequently, the foremost equations that have been modified to characterize a pm material exhibit a prescribed direction of remnant magnetic flux density. this enables updating the location of the intersecting point between both the characteristics of pm and soft magnet at each time step, and hence shows a realistic representation of the pmfcl behaviour during its entire operation. such coupled fe-ja model, which encompasses the nonlinear representation of pm through ja method, is predominantly validated with the experimental results of (zou et al. 2008, li et al. 2012), as shown in fig. 4. however, the experimental setup in (li et al. 2012) has been modelled through the coupled fe-ja model, and the simulation results of this coupled model are compared with the experimental measurements as shown in fig. 4. figure 4 shows that the experimental results are in good agreement with simulation results with an error difference of ±3% due to the missing ac coil dimensions (li et al. 2012). therefore, the actual representation of pm demagnetizing behaviour through ja model in the upper part of the b-h curve with negative magnetic field intensity offers truthful tendencies for the pmfcl dynamic behaviour in limiting fault current during normal and fault conditions. 3. performance investigation of stacked pmfcls according to the above-mentioned operating principles of pmfcl, a 3d time-varying, comsol multiphysics (comsol 2013) electric-circuit magnetic-field coupled fe model is reproduced. however, the three-phase stacked pmfcls designs, figs. 2(b), (c), and (d), are compared with the figure 4. comparison between the pmfcl experimental results of (li et al. 2012) and the comsol fe simulation comprises the nonlinear behavior of pm. (a) (b) figure 3. surface plots at fault condition for (a) magnetic flux density b in (t) with cone directions of b, and (b) magnetic field intensity h in (a/m), both at t = 0.09 s, for the 8ac-5pm pmfcl stacked delta design. nac = 48. stacked delta design of three-phase permanent-magnet fault current limiters 31 reference design of 4ac-5pm (eladawy et al. 2020) shown in fig. 2(a). however, the summarized parameters in table 1 are common for all pmfcls in fig. 2. for the same total number of ac turns of each phase, the changes are through the thickness of the soft magnet, number of pm segments, and/or the number of ac sub coils. the line current reaches a peak value of 100 a and 3.1 ka, as shown in fig. 2, (without any fcl), during the normal non-limiting and the fault conditions, respectively. the total simulation time is considered to be 0.14 s, at which the fault has been started and removed at t = 0.04 s and at t = 0.1 s, respectively. in general, the clipping ratio (k) of fault current, during the fault condition, is calculated as follows (eladawy et al. 2019): ( ) flf iiik ˆ/ˆˆ −= (1) where îf and îl are the prospective and limited peak values of fault current, respectively. this clipping ratio is termed to be calculated at the last peak of fault current, which represents the minimum value of the clipping ratio. furthermore, the pmfcl dynamic performance is a strictly multi-parameter function in the physical and constructive parameters. it severely depends on the relative cross-sectional areas of the geometrical design topology of the soft magnet, the type of pm biasing, and the ac coil number of turns. comprehensive fe simulations have been executed to identify the influence of such parameters on the dynamic performance of pmfcl (eladawy et al. 2018; moscrop 2013; eladawy et al. 2019, eladawy et al. 2020). subsequently, the dynamic performance can be characterized by the clipping ratio (k) of fault current, and the total power loss (ploss) and the voltage drop across pmfcl terminals (∆vfcl) throughout the normal operation of the system. 4. performance analysis 8ac-5pm stacked design 4.1 number of ac coil turns for increasing the current clipping ratio of the 8ac5pm design, the number of turns of ac coils per phase is increased (nac = 96), with conserving the other parameters in table 1. figures 5 (a), (b), and (c) show the three-phase currents of the 8ac-5pm stacked design, fig. 2(c), during the entire simulation interval due to symmetrical three-phase, and unsymmetrical line-to-line and single line-to-ground faults, respectively. figure 2(d) shows the voltage drop, across each phase of the ac coils, during the steady-state condition after fault removal, for the 8ac-5pm stacked delta pmfcl. it can be observed from figs. 5 (a), (b), and (c) that the pmfcl promptly recuperates its initial saturation state after fault removal with a negligible impedance of the pmfcl due to the strong saturation state offered by the distributed pm segments. however, the 8ac-5pm stacked delta pmfcl results in the minimum value of fault clipping ratio k ≈ 84.4 %, k ≈ 87.1 % and k ≈ 81.8 %, for symmetrical three-phase, and unsymmetrical line-to-line and single line-to-ground faults, respectively. ∆vfcl ≈ 0.89% and ploss ≈ 0.66% during the normal operating condition, without any observed significance on non-faulted phases (r) and (y and b) in case of line-to-line and single lineto-ground fault, as shown from fig. 6 (b) and (c), respectively. irrespective of the duplication of the soft magnet volume, this design offers the enriched capability to limit fault current without exceeding both the power losses and voltage drop permissible tolerances. significant reduction in the voltage drop can be achieved by changing the arrangement of ac coils around the soft magnet with doubled soft magnet volume for the 8ac-5pm stacked design, as shown in fig. 7, due to the effective saturation of the soft magnet and consequently reduces the current clipping ratio. the outcomes of fe simulations for the 4ac5pm pmfcl reference delta-shaped design (eladawy et al. 2020), shown in fig. 2(a) in accordance with table 1, have been reproduced and the soft magnet volume has been taken as a reference for all other designs. this design will result in k ≈ 68 %, ∆vfcl ≈ 0.8% and ploss ≈ 0.39%, as shown in fig. 8. however, any further increase in soft magnet volume, as shown in fig. 2(b), or nac will in turn increase the steadystate voltage drop and power losses. this can be referred to the increase in the ac coil impedance due to the increase in the coil area or its number or turns. therefore, the 4ac-5pm stacked delta design exhibits some limiting effect throughout the normal operating condition. conserving the same number of ac turns, the multi-storey delta design of 12ac-10pm gives the lowest fault current clipping ratio, voltage drop, and power losses, as shown in fig. 8. this referred to the deeper saturation state of the soft magnet due to the increase of pm segments. these lower values of ∆vfcl encourages increasing the ac coil turns for increasing the clipping ratio of fault current without exceeding the permissible limits of voltage drop and power losses. figure 8 shows that the 8ac-5pm stacked delta design offers adequate characteristics in limiting fault current while conserving the same number of ac coil turns for each phase, the number of pm segments and increasing the volume of the soft magnet. figure 7 shows the dynamic performance of 8ac5pm pmfcl with changing the ac coil number of turns per phase. however, increasing nac leads to an increase in the ac coil inductance. consequently, the pmfcl offers a higher clipping ratio of fault current with increasing nac due to the increase of fault mohamed eladawy and ibrahim a. metwally 32 impedance, as shown in fig. 8. accordingly, this increase of nac leads to an increase in both the voltage drop and power losses throughout the normal operation of the system, due to the increase of steadystate impedance (eladawy et al. 2018, moscrop 2013). therefore, the ac coil number of turns can be considered as a crucial parameter, which controls the pmfcl entire operating conditions. (a) (b) (c) (d) figure 6. clipping effect of fault current and voltage drop waveforms for the 8ac-5pm stacked delta pmfcl due to symmetrical three-phase fault, line-to-line fault and single-line-to-ground fault. nac = 96. (a) three-phase fault currents, (b) line-to-line fault, (c) single-line-to ground fault currents, and (d) voltage drop. figure 8. comparison between the different proposed stacked designs of pmfcls. volume is referred to the reference design of 4ac-5pm delta design (eladawy et al. 2020). nac = 48. figure 7. variation of the dynamic performance of 8ac5pm pmfcl stacked design with changing the ac coil number of turns. figure 5. three-phase currents of the test circuit without any pmfcl. stacked delta design of three-phase permanent-magnet fault current limiters 33 4.2 fault type and fault current peak value for ensuring the reliability of the three-phase 8ac5pm pmfcl, different unsymmetrical faults have been tested through extensive fe simulations. these unsymmetrical faults that comprise line-to-ground (lg), double line-to-ground (llg), and line-to-line (ll) are compared with those for the symmetrical three line-to-ground (lllg) and three-line (lll) faults. figure 9 shows the satisfactory capability of the 8ac-5pm pmfcl in limiting the fault current (k ≥ 81.8%) for different fault types, when nac = 96, without exceeding the permissible tolerances of both the voltage drop and power losses. the small difference in the current clipping ratio for different phases (for the same fault type) can be referred to as the point on wave switching. moreover, extensive fe simulations show that the healthy phases are tremendously unaffected by the faulted phase in the case of unsymmetrical faults. consequently, decreasing the line resistance (rline) will in turn increase the fault current peak value. referring to table 1 parameters and for nac = 96, the fault current has been changed up to 500% of its reference value (îf = 3.1 ka), as shown in fig. 10. it can be observed from fig. 10 that the increase of fault current will in turn lead to higher demagnetization of the soft magnet, which leads the operating point towards the higher relative permeability in the b-h curve linear zone. thus, this leads to higher fault impedance and higher fault current clipping ratio as shown in fig. 10. furthermore, decreasing the line resistance leads to an increase in the steady-state line current, which will in turn slightly increase the voltage drop and the power losses. however, neither the voltage drop nor the power losses exceed their permissible tolerances due to the effective saturation of the soft magnet throughout the healthy condition of the system. subsequently, the 8ac-5pm stacked design of pmfcl offers improved capability in limiting any fault type even when the fault current peak value is increased. accordingly, the connection between the ac sub coils is a crucial design parameter that significantly affects the dynamic performance of the pmfcl. however, the cumulative connections of sub coils increase the overall ac coil inductance during the non-limiting operation. this leads to an increase in the voltage drop across the pmfcl terminals. moreover, the ac coil wire’s cross-sectional area significantly affects the power losses, especially during the non-limiting condition. therefore, it should be regulated to obtain power losses percentage with the specified tolerance by the utility. however, similar tendencies with different values are obtained with the 12ac-10pm pmfcl. this ensures that the rearrangement of the ac coils around the soft magnet can be very useful in extending the range of operation of the stacked or multi-storey pmfcl designs. figure 10. variation of the dynamic performance of 8ac5pm pmfcl stacked design with changing the fault current peak value. nac = 96. figure 9. fault current clipping ratio of three-phase 8ac-5pm pmfcl for different fault types. nac = 96. mohamed eladawy and ibrahim a. metwally 34 5. conclusion the capability of fault current limiting of the threephase delta-shaped pmfcl design is enhanced by introducing the stacked and multi-storey delta-shaped designs. extensive 3d, time domain, electric-circuit magnetic-field coupled model, fe simulations of comsol multiphysics have been carried out taking into consideration the realistic representation of the pm demagnetization behaviour in the upper part of the b-h hysteresis loop with negative magnetic field intensity. ndfeb-n42 pm is only considered to bias the soft magnet. the pm demagnetizing behaviour is represented through the b-h loop of ja model. the pm segments are sandwiched with a uniform distribution between two equilateral triangles. for effective saturation, the magnetization directions are kept opposite for each two adjacent pms. the ac sub coils of each phase are placed at the air gaps between each two adjacent pms, and wound around the soft magnet with a differential series connection. increasing the soft magnet cross-sectional area leads to an increase in both the clipping ratio of fault current and the voltage drop across the terminals of ac coils. this limits the capability of the pmfcl in limiting the fault current to be within the specified tolerance of the voltage drop. the division of the ac coils, into an even number, of sub coils and locating them at the air gaps between the soft magnets leads to the reduction of the inductance of such coils during the non-limiting performance of the pmfcl. this has the effect of reducing the voltage drop. therefore, this gives the opportunity to increase the ac coils number of turns. this will in turn increase the capability of fault current limiting using the stacked or multi-storey configurations. the three-phase 8ac-5pm stacked delta design shows adequate dynamic performance in limiting the fault current without duplicating the pm segments number. results reveal that this design has enriched the capability to limit all fault types, either symmetrical or unsymmetrical, with increased peaks of fault currents. conflict of interest the authors declare no conflict of interest. funding no funding was received for this study. references coey jmd (1996), rare-earth iron permanent magnets, clarendon press, isbn: 9780198517924. comsol group ltd., stockholm, sweden (2013), comsol multiphysics software package v.4.4. duggan pm (2006), integration issues for fault current limiters and other new technologies-a utility perspective. in proc. ieee power eng. soc. gen. meeting, montreal, qc, canada: 1–3. eladawy m, metwally ia (2018), design and performance analysis of single-phase presaturated core fault current limiters, ieee trans. power del., 33(6): 2603-2612. eladawy m, metwally ia (2018), a novel five-leg design for performance improvement of threephase presaturated core fault current limiter, ieee trans. magn., 54(7), sequence no. 8001410: 1-10. eladawy m, metwally ia (2020), compact designs of permanent-magnet biased fault current limiters, iet electr. power appl., (3): 471 – 479. eladawy m, metwally ia (2018), design and performance analysis of single-phase presaturated core fault current limiters, ieee trans. power del., 33(6): 2603-2612. eladawy m, tarhuni nj, metwally ia (2019), transient dynamic analyses of presaturated core fault current limiters through flux and inductance versus current modelling, iet electr. power appl., 13 (9): 1263-1272. jia y, yuan j, shi z, zhu h, geng y, zou j (2016), simulation method for current-limiting effect of saturated-core superconducting fault current limiter, ieee trans. appl. supercond., 26(4), art. no. 5601404. li q, xu j, zou l, et al. (2012), modelling methodology and experimental verification of the permanent-magnet-biased saturation-based fault current limiter, iet electr. power appl., 6(8): 504–512. moscrop jw (2013), experimental analysis of the magnetic flux characteristics of saturated core fault current limiters, ieee trans. magn., 49(2): 874-882. pannu m, valent y, garbi u (2013), pre-saturated core fault current limiter, australasian universities power engineering conference, aupec, hobart, tas, australia, 29 sept. 3 oct. pop nc, caltun of (2011), jiles-atherton magnetic hysteresis parameters identification, acta phys. polonica a, 120(3): 491–496. prigmore j, mendoza ja, karady gg (2013), comparison of four different types of ferromagnetic materials for fault current limiter applications, ieee trans. power del., 28(3): 1491-1498. radmanesh h, fathi sh, gharehpetian gb, heidary a (2016), bridge-type solid-state fault current limiter based on ac/dc reactor, ieee trans. power del., 31(1): 200–209. xiao h et al. (2011), analysis of transient overvoltage in 220 kv saturated core hts fcl, ieee trans. magn., 47(10): 2620–2623. yuan j, et al. (2017), a novel hybrid saturated core fault current limiter topology considering permanent magnet stability and performance,” stacked delta design of three-phase permanent-magnet fault current limiters 35 ieee trans. magn., 53(6), art. no. 8400304. yuan j, zhou h, gan p, et al. (2018), a novel concept of fault current limiter based on saturable core in high voltage dc transmission system, aip advances, 8(5), paper 056636: 1-6. zhao c, wang s, qiu j, et al. (2007), transient simulation and analysis for saturated core high temperature superconducting fault current limiter, ieee trans. magn., 44(4): 1813-1816. zou l, liu h, li q, siew wh (2008), analysis and simulation of the pmfcl based on coupled field-circuit modeling methodology, in proc. int. conf. high voltage eng. appl., chongqing, china, 9-13: 498-502 1. introduction 2. design topology and operating principles 2.1 design topology 2.2 operating principles 3. performance investigation of stacked pmfcls 4. performance analysis 8ac-5pm stacked design 4.1 number of ac coil turns 4.2 fault type and fault current peak value 5. conclusion conflict of interest funding references the journal of engineering research (tjer), vol. 19, no. 1, (2022) 33-40 *corresponding author’s e-mail: alawim@squ.edu.om doi:10.53540/tjer.vol19iss1pp33-40 productivity factors in construction smes in oman exploration study amwaj alabdali1, mubarak al alawi1*, mohammed al shahri1, and shadi altawil2 1 department of civil and architectural engineering, sultan qaboos university, al khod, oman 2department of civil and infrastructure engineering, american university of ras al khaimah, united arab emirates abstract: construction projects are unique and influenced by many external and internal factors hindering the efficiencies of operations. different construction companies may have different influencing productivity factors depending on the company’s scale. large construction companies have better resource management systems than the construction small and medium enterprises (sme). the construction smes work in small scale projects, and the amount and type of the required resources are different from the large construction companies. therefore, identifying the factors affecting construction productivity is essential to determine the best action to avoid any unforeseen productivity reduction. this study aims to rank the most influencing productivity factors in small and medium construction (sme) companies. a questionnaire survey was used to collect data from small and medium construction companies registered in oman tender board. the relative importance index (rii) was used to evaluate the responses. sixty-five companies out of the 207 targeted population participated in the survey. it was found that incomplete drawings and drawing omissions, poor site condition and management, delay in payment, unavailability of proper tools and equipment, and workers absenteeism and turnover were found as the top factors that influence construction productivity. keywords: construction; productivity; rii; oman; smes دراسة استكشافیة -عوامل اإلنتاج في المقاوالت الصغرى والمتوسطة في عمان الطویلأمواج العبدلي مبارك العلوي* محمد الشحري وشادي وقد یكون لشركات البناء المختلفة التي تعوق كفاءة عملیات البناء. الداخلیة والخارجیة تتأثر مشاریع البناء بالعدید من العوامل :الملخص كمیة ونوع لكون شركات البناء الكبرى لدیھا أنظمة إلدارة الموارد أفضل من المؤسسات الصغیرة والمتوسطة، كما أن تأثیر مختلف یعد تحدید ھذه العوامل ضروریًا لتحدید أفضل الممارسات لتجنب أي انخفاض غیر الموارد المطلوبة تختلف بحسب إمكانات الشركة. متوقع في إنتاجیة البناء. ھدفت ھذه الدراسة إلى تصنیف عوامل اإلنتاجیة األكثر تأثیراً على شركات المقاوالت الصغیرة والمتوسطة. تم لجمع البیانات من شركات المقاوالت الصغیرة والمتوسطة المسجلة في مجلس المناقصات العماني. تم استخدام مؤشر استخدام استبیان مجموعة مستھدفة. تبین أن الرسومات غیر المكتملة 207شركة من أصل 65) لتقییم الردود. شارك في المسح riiاألھمیة النسبیة ( إدارتھ، والتأخیر في الدفع، ونقص األدوات والمعدات، وتغیب العمال واستقاالتھم، من أكثر واألخطاء في الرسم، وسوء حالة الموقع و العوامل التي تؤثر على إنتاجیة البناء في مشاریع الشركات الصغیرة والمتوسطة العاملة في قطاع التشیید. ؛ سلطنة عمان؛ الشركات الصغیرة والمتوسطة riiبناء؛ إنتاجیة؛ الكلمات المفتاحیة: productivity factors in construction smes in omanexploration study 34 1. introduction the construction industry has a substantial impact on the international and local economies. in australia, the construction industry contributes about $360 billion in revenue and grows around 9% of the gross domestic product (gdp). it has a projected annual growth rate of 2.4% from 2020 until 2025 (australian industry and skills committee, 2020). in malaysia, the industry has been contributing between 3 to 5 % of the gdp (ali khan, shahir liew, and bin ghazali, 2014). also, india's construction industry (real estate and urban development) produced about $ 5 billion in revenues, and its contribution to the indian gdp was 9% (padia, 2020). while the omani construction market is expected to generate 6% of the compound annual growth rate (cagr) over the period between 2019 and 2024 (intelligence, 2020). this forecast is translated by the 2,410 active construction projects in oman, with a value of over usd 190 billion (intelligence, 2020). omani construction industry market varies in enterprises standards and contributing to the industry. mainly, the dominant enterprises in the omani market are the small and medium (smes). the prominent appearance of smes in the construction industry is due to their low investment head compared to other enterprises (bevan and yung, 2015). smes presents a pivotal role in the economic growth of the country, and it also contributes toward improving the quality of life and living standards. the construction industry smes face many challenges. the challenges may be related to different project characteristics and requirements such as the project location, project complexity, required resources, project time requirements etc. (azis, memon, abdul rahman, nagapan, and qadir, 2012). also, high competition opposes some challenges, such as the shrinkage of the profit margins. some other challenges can be related to some groups with different controlling factors such as labour and human-related factors, financial, administrative/managerial, and planning groups (almathami et al., 2017). all of these groups have a direct relationship with construction productivity. it may negatively affect construction productivity and may negatively influence the construction project cost, time, and quality. occupational health and safety practices can also affect construction productivity (al mawil et al., 2021). thus, smes are characterized by a high failure rate caused by many factors, such as low productivity (adebowale and agumba, 2021). therefore, it is essential to identify productivity controlling factors. the construction productivity term is defined as the ratio of production output to what is required to produce, such as resources used to generate the production output (intergraph, 2012). in addition, productivity is defined as the total output ratio to the entire input (whiteside, 2006). the production rates help to conduct a relative level of productivity determination. a production rate in the construction industry is when workers are expected to complete a defined task given a set of resources (guntuk and koehn, 2009). the production rate depends on the time at which workers are expected to finish the assigned task according to the work specification. thus, the evaluation and analysis of the productivity factors help develop the construction industry by averting low productivity project environments. the construction industry in many different countries is experiencing growth and challenges at the same time. some of the challenges are related to maintaining a production rate suitable to achieve the project's goals. however, construction productivity performance is controlled by different factors, and it is essential to identify these factors to allow for better development in project planning and control. therefore, this study aims to investigate the construction productivity factors in omani construction smes. 2. literature review many researchers agree that productivity is one of the essential variables at which production activities are evaluated and monitored. several researchers have also declared that to maintain a competitive production level. this section discusses different previous researches in the area of construction productivity. the resources collected are semi-structured according to the geographical location of the study. hughes and thorpe (2014) investigated the construction productivity enabling factors in the australian construction industry. a structured questionnaire survey was used in the study and was distributed to construction project managers. the study concluded that the rework was the most construction productivity-hindering factor followed by an incompetent supervisor, incomplete drawings. the investigation was conducted by a structured questionnaire survey sent to several random construction project managers. they concluded that rework, incompetent supervisor, incomplete drawing, labours work overload and unavailability of materials are the most influencing productivity factor. factors relating the labour productivity to the project schedule performance in indonesia were studied by soekiman (2011), who investigated the labour productivity factors that affect the project schedules in the indonesian construction industry. the study was made through a questionnaire survey distributed to respondents who manage various types of construction projects within large, medium and small companies. the results showed that incompetent supervisors issuing unclear instructions to labours is the most influencing productivity factor. the unavailability of materials, the financial status of the client, and the unavailability of definite schedules are also major productivity factors. jaafar (2005) investigated the factors affecting amwaj alabdali, mubarak al alawi, mohammed al shahri, and shadi altawil 35 construction labour productivity in malaysian residential projects. the data were obtained through a questionnaire survey, and a total of 100 respondents were collected from contractors, developers, and consultants. it was found that the unavailability of materials was the most critical productivity factor followed by the delay in payment, late issuing of shop drawings, and poor construction management. pornthepkasemsant and charoenpornpattana (2019) identified the factors affecting construction productivity in thailand’s construction industry. they identified the following factors as the most influential productivity factors: (1) lack of labour, (2) lack of experienced labours, (3) incomplete drawings, (4) inefficient communication with the engineer, and (5) financial difficulties. on the other hand, tuan hai (2019) analysed the affecting factors on construction productivity in vietnam. the data were obtained through a survey and recognized the following vital determinants: construction workers’ motivations, working tools, the organization and management of the onsite production and labour safety. dixit (2017, 2018) investigated the construction productivity factors in the indian construction industry. the data were collected by a structured questionnaire survey distributed to a group of professionals working in the indian construction industry. it was found the company’s management lack of decision-making skills was a crucial productivity determinant. also, poor planning and controlling, inefficient logistics and supply chaining management systems, and unavailability of labours have a substantial effect on construction. golchin and kim (2018) studied factors affecting the construction labour productivity in the iranian construction industry. the study found that rework was the most influencing productivity factor. delay in payment, unavailability of equipment and tools and incomplete drawings were also major construction productivity factors. similarly, ghoddousi and poorafshar (2015) investigated labour productivity in iranian construction projects. the study targeted sixty chief executive officers. it was found that the salary and incentives, delay in payment, and the ethical behaviour of managers influence construction productivity. al-rubaye and mahjoob (2020) identified the main factors affecting labour productivity in different iraqi construction companies. it was found that construction site management, site accessibility, terrorist attacks, and the financial challenges of the contractor are major construction productivity factors. khaleel and nassar (2018) also studied and analysed the factors affecting labour productivity in the iraqi construction industry. data were collected from different sources such as site surveys, interviews with engineers and experts, and previous research in the arab region. the unavailability of materials was found the most influencing productivity factor followed by weather conditions, political issues, and stability and security. in the jordanian construction industry, the construction productivity factors were investigated by hiyassat et al. (2016). a questionnaire survey was sent to 200 engineers/foremen who work in the construction industry—the most influencing productivity factor labours experience. financial challenges, inefficient communication between engineers and labours, and poor work scheduling were also considered major construction productivity factors. shurrab (2018) conducted a questionnaire survey to evaluate the effect of a motivational factor on the construction project managers in jordan. shurrab (2018) found that the wage rate, personal development opportunities, and rewards have a direct relationship with construction productivity. al-abbadi and agyekum-mensah (2019), on the other hand, also evaluated the effects of motivational factors on the construction professional’s productivity in the jordanian construction industry. a questionnaire survey and semi-structured interviews were used to collect the data. it was found that personal/career development and on-time payment and rewards and incentives, wage rate, and respect contributes positively toward improving construction productivity. el-gohary and aziz (2014) explored the influencing factors to construction labour productivity in the egyptian construction industry, and it was found that the labour skill and experience, competence of construction manager, incentive programs and the availability of materials are major construction productivity factors. factors influencing productivity in the construction industry in saudi arabia's kingdom were studied by almathami et al. (2017). the results showed that the financial situation of the client and the contractor, project site-related factors and the availability of materials and equipment are influencing construction productivity factors. mahamid (2013) investigated the contractor’s perspective toward factors affecting labour productivity in building construction in palestine. a structured questionnaire survey was used to collect data from contractors working in building construction. the results showed rework as the most influencing construction productivity factor followed by the communication between the project's parties, the client's financial status, labour experience, and lack of materials. enshassi (2014) investigated the productivity factors of the palestinian construction craftsmen. a structured questionnaire survey was distributed to many of the construction craftsmen. the results indicated that the highest-ranking productivity factors are the availability of materials, safety, delay in payment, inspection delays, low wages, and lack of craftsmen experience. in the context of the yemeni construction industry, alaghbari et al. (2018) studied labours construction productivity. it was found that the experiences level of productivity factors in construction smes in omanexploration study 36 labour, materials availability, site management, and political stability were considered the most influencing construction productivity factors. jarkas et al. (2015), on the other hand, explored the labours productivity factors in bahrain's construction industry. a structured questionnaire survey was used to collect the data. it was found that the coordination between different design disciplines, incompetent site supervisors, design errors and omissions, delay in responding to contractor’s requests for information and labour skills and experience are the most influencing construction productivity factors. similarly, jarkas and bitar (2012) studied the construction labour productivity factors in kuwait and used a structured questionnaire survey to collect the data from a group of contractors. the analyses of the data recognized that the clarity of specifications, coordination among different design disciplines, the proportion of work subcontracted, change orders, and lack of labour supervision were the most influencing construction productivity factors. jarkas et al. (2015) investigated the construction labour productivity in oman. a questionnaire survey was distributed to the top contractors registered in oman’s tender board. it was found that change orders, design errors and omission, incompetent labour supervision, delay in responding to requests for information contribute to construction productivity. jarkas et al. (2012) studied the same in the context of the qatari construction industry. a structured questionnaire survey was also used to collect data and concluded that the communication between site management and labour, labours experience and skills, availability of materials are the productivity influencing factors. factors affecting employee productivity in the united arab emirates (uae) construction industry were investigated by ailabouni et al. (2009). the study was carried out by distributing a questionnaire survey. the results identified the following factors as most significant: proper work timings, balancing work and recreation, time with family, leadership skills of supervisors, payment, and technical competency. based on the above literature review, the construction productivity factor area of research was well established globally and regionally. however, the study didn’t take into consideration small and medium construction companies. the reason is maybe because of the challenge related to the communication with these companies. 3. problem statement construction is a crucial activity within any economy, given its contribution to any economic prospects' gross domestic product. fulfilling good quality construction projects within the required period is accomplished by maintaining a high productivity working environment. productivity is a significant determinant in construction industry projects. productivity level and efficiency are affected by various factors such as rework, incompetent supervisor and drawing, work overload, and lack of material, increasing the construction project's financial risks. hence, improving construction productivity can be performed through the identification of the productivity retaining factors. besides, it is essential to identify these factors and rank them according to their effectiveness on construction productivity. therefore, this study aims to identify factors affecting the productivity of the omani construction industry. this study aims to identify the acquirer construction companies' productivity factors and contractors who contribute locally in the small and medium construction projects. analysing the productivity factors, challenges and consequences for these companies and contractors will produce a clear insight into the omani construction market. 4. methodology the relevant data for this investigation were collected based on related previous studies on construction productivity. see table 1; twenty-one factors were collected from those studies. the selection of these factors was based on their frequency of appearance in the literature. this fact is considered a limitation, and the future direction of the study is to identify the predominant productivities factors affecting the construction smes. a questionnaire survey was prepared using “google form”, and it contains two parts. the first part covers the participant's demographic information, and the second part captures the participant's perception of the construction productivity factors surveyed. table 1. productivity factors surveyed. no. productivity factors 1 delay in payment 2 construction method 3 unrealistic time and cost estimate 4 work supervision 5 effectiveness of communication protocol (i.e. clarity of instructions, delay in responding to request for information (rfi) ) 6 incomplete drawings or drawings omission 7 overcrowding 8 lack of materials 9 lack of proper material handling and storage 10 lack of tools and equipment 11 accidents 12 change orders 13 inspection delay 14 workers absenteeism and turnover 15 poor site conditions and management 16 site accessibility 17 planning and workflow 18 labours skills and experiences 19 weather effect 20 project specifications 21 rework amwaj alabdali, mubarak al alawi, mohammed al shahri, and shadi altawil 37 five points likert scale is used in the questionnaire survey. these points are (5) extreme effect, (4) substantial effect, (3) moderate effect, (2) minor effect, and (1) no effect. the data collected will be analysed using the relative importance index (rii) technique. the relative importance index is calculated using the following formula: rii = 5 (𝑛𝑛5)+4(𝑛𝑛4)+3(𝑛𝑛3)+2(𝑛𝑛2)+𝑛𝑛1 5(𝑛𝑛1+𝑛𝑛2+𝑛𝑛3+𝑛𝑛4+𝑛𝑛5) × 100 (1) where: n1, n2, n3, n4, and n5 = the number of respondents who selected: 1, for no effect; 2, for minor effect; 3, for moderate effect; 4, for substantial effect; and 5, for extreme effect, respectively. the target population of the study are the construction companies registered in oman tender board. the other criterion used is that the registered construction company should be classified as small and medium-sized enterprises (sme). accordingly, the number of construction smes registered in the oman tender board were 207(2020). to obtain a statistically representative sample of the population, eqn. (2) was used as follows: 𝑛𝑛 = 𝑚𝑚 1+( 𝑚𝑚−1 𝑁𝑁 ) (2) where, n= the sample size of the limited population, m= the sample size of the unlimited population n = the sample size of the available population, respectively. m is estimated by eqn. (3): 𝑚𝑚 = 𝑧𝑧 2×𝜌𝜌×(1−𝜌𝜌) 𝜀𝜀2 (3) where, z = the statistic value for the confidence level p = the value of the population proportion that is being estimated, obtained to be 0.50. ε = the sampling error of the point estimate. m, is approximated as follows: 𝑚𝑚 = (1.96)2 × 0.5 × (1 − 0.5) (0.05)2 ≈ 385 accordingly, the representative sample size of the population required is determined as shown below: 𝑛𝑛 = 385 1 + (385 − 1207 ) ≈ 135 5. discussion of results the questionnaire survey was distributed to 207 small and medium construction companies registered in the oman tender board. the data collection processes extended for three months, and sixty-five responses were received, representing 48% of the representative sample. table 2 below shows the demographic information of the respondents. the majority were omanis, and males represent 86% of the total respondents. fifty-one per cent of the respondents are within the 25-35 age range, and the majority hold a bsc degree. also, 57% of the respondents represent the job title “manager”. however, some of the respondents have written their job title as “company owner”, and it is known that it is not a job title. the majority of the small and medium construction companies’ owners run their businesses. therefore, it is assumed that the company owners are considered managers. regarding the working experiences, the majority of the respondent’s work experience is under 15 years. figure 1 below shows the relative importance indexes (rii) for the surveyed construction productivity factors. the higher the value of the rii, the more the influence of the construction productivity factor is considered by the small and medium construction companies. incomplete drawings or drawings omissions were found the most productive factor influencing the construction work, and work supervision were found the least influencing productivity factor. incomplete drawings or drawing omissions are managerial aspects and are supposed to be resolved during the cost estimating. the contractor is expected to inquire about any missing information; however, the small and medium construction companies use a unit rate cost estimating technique to price the project without even looking into the drawing’s specifications. that is why the incomplete drawings and drawings omissions scored high, from the authors' point of view. table 2. demographic information. no demographic information % 1 nationality omani 97 non-omani 3 2 gender male 86 female 14 3 age 25-35 51 36-46 29 46-55 18 >55 2 msc/meng 8 bsc 49 diploma 37 other (primary education and high school) 6 5 job title manager 57 engineer 25 site superintendent 5 other (company owner) 14 6 work experience (years) < 5 23 5-10 29 11-15 20 16-20 15 >20 12 productivity factors in construction smes in omanexploration study 38 figure 1. relative importance indexes (rii) for the surveyed productivity factors. poor site conditions and management, lack tools and equipment, workers’ absenteeism and turnover, and planning and workflow are among the top influencing construction productivity factors, which can be correlated to the delay in payment. small and medium construction companies rely on subcontractors to do some specialized work such as repairing, mechanical, electrical and plumbing (mep), and finishing works. suppose poor management of resources and delay in payment is experienced in the project. in that case, it is expected these factors will exert a dominos effect, which means that all be affected and may express the same results. many different factors usually cause rework. for example, drawing omissions, lack of proper quality control, and lack of supervision may cause rework. the rework scored high (above 80%). this result is driven by the fact that incomplete drawings and drawings omission were ranked first. the lack of materials, unrealistic time and cost estimate, and overcrowding 8th, 9th, and 10th, respectively. these factors are related to resources management, and it can be seen that the small and medium construction companies may have management issues. the weather effect, construction method, site accessibility, project specification, accidents were found among the least influencing construction productivity factors. this result is expected since the small and medium construction companies usually involve building small manageable projects. also, they are sometimes managed by a primary large construction company in case they work as subcontractors. 6. conclusion maintaining planned construction productivity levels is essential to achieve the projects’ milestones. however, construction productivity is influenced by many factors, including managerial, technological, or labour factors. it is necessary to identify the most influencing construction productivity factors to better plan construction activities. many previous studies identified the most influencing construction productivity factors affecting large construction companies, and this study focused on small and medium construction companies working in the omani construction industry. the study found that incomplete drawings and drawing omissions most influence construction productivity, followed by poor site conditions and management. the delay in payment, the financial aspect of the construction productivity factor, was ranked third. the technical problem in drawings omissions is the dominant productivity hindering factor for construction smes; however, most of the highly ranked factors belong to the management aspect. this finding highlights the need for construction smes to improve their organizational management efficiency. the future extension of this study is to cover more small and medium construction companies to allow the findings of this study to be generalized in the oman construction industry. conflict of interest the authors declare no conflict of interest. 60.00 65.00 70.00 75.00 80.00 85.00 incomplete drawings or drawings omission poor site conditions and management delay in payment lack of tools and equipment workers absenteeism and turnover planning and work flow rework lack of materials unrealistic time and cost estimate overcrowding inspection delay lack of proper material handling and storage labours skills and experiences effectiveness of communication protocol (i.e. clarity… change orders weather effect construction method site accessibility project specifications accidents work supervision relative importance index (rii) p ro du ct iv it y fa ct or s amwaj alabdali, mubarak al alawi, mohammed al shahri, and shadi altawil 39 funding no funding was received for this research. references adebowale, o. j., and agumba, j. n. (2021). a metaanalysis of factors affecting labour productivity of construction smes in developing countries. journal of engineering, design and technology. doi:https://doi.org/10.1108/jedt-05-2021-0277 ailabouni, n., gidado, k., and painting, n. (2009). factors affecting employee productivity in the uae construction industry. annual conference; 25th, association of researchers in construction management (pp. 555-564). nottingham: association of researchers in construction management. al mawil, b., al alawi, m., elazouni, a., and almamun, a. (2021). construction smes safety challenges in water sector in oman. safety science, 105156. al-abbadi, g. m., and agyekum-mensah, g. (2019). the effects of motivational factors on construction professionals productivity in jordan. international journal of construction management, 1-13. alaghbari, w., al-sakkaf, a., and sultan, b. (2018). factors affecting construction labour productivity in yemen. international journal of construction management, 19(1), 79–91. ali khan, r., shahir liew, m., and bin ghazali, z. (2014). malaysian construction sector and malaysia vision 2020: developed nation status. retrieved from sciencedirectassets. almathami, k. y., trigunarsyah, b., and coffey, v. (2017). factors influencing productivity in construction. resilient structures and sustainable construction, 1-6. al-rubaye, z., and mahjoob, a. (2020). identify the main factors affecting labor productivity within different organizational structures in the iraqi construction sector. iop conf. series: materials science and engineering, 1-15. australian industry and skills committee. (2020, 4 3). national industry insights report. retrieved from australian industry and skills committee: https://nationalindustryinsights.aisc.net.au/industri es/construction azis, a. a., memon, a. h., abdul rahman, i., nagapan, s., and qadir, l. (2012). challenges faced by construction industry in accomplishing sustainablity goals. 2012 ieee symposium on business, engineering and industrial applications, (pp. 628-633). bevan, e. a., and yung, p. (2015). implementation of corporate social responsibility in australian construction smes. engineering, construction and architectural management, 22(3), 295-311. cherian, a. (2020). the construction industry in the perspective of an economic boost of the. retrieved from irjet: https://www.irjet.net/ archives/v7/i8/irjet-v7i846.pdf dixit, s. (2017). international journal of civil engineering and technology, 8(6). dixit, s. (2018). analysing enabling factors affecting the on-site productivity in indian construction industry. periodica polytechnica architecture, 49(2), 185–193. dixit, s., and saurabh, k. (2019). impact of construction productivity attributes over construction project performance in indian construction projects. periodica polytechnica architecture, 50(1), 89–96. el-gohary, k. m., and fayek aziz, r. (2014). factors influencing construction labor productivity in egypt. journal of management in engineering, 30(1), 1-9. enshassi, a. (2014). constraints to palestinian construction craftmen productivity. proceedings iglc, 1507-1517. ghoddousi, p., and poorafshar, o. (2015). labour productivity in iranian construction projects. international journal of productivity and performance management, 64(6), 811-830. golchin rad, k., and kim, s.-y. (2018). factors affecting construction labor productivity: iran case study. iranian journal of science and technology transactions of civil engineering, 42, 165–180. guntuk, c., and koehn, e. (2009). construction productivity and production rates. hiyassat, m., hiyari, m., and j. sweis, g. (2016). factors affecting construction labour productivity: a case study of jordan. international journal of construction management, 16(2), 138-149. hughes, r., and thorpe, d. (2014). a review of enabling factors in construction industry productivity in an australian environment. construction innovation, 14(2), 210-228. intelligence, m. (2020). oman construction market growth, trends, and forecast (2020 2025). retrieved from mordor intelligence: https://www.mordorintelligence.com/industryreports/oman-construction-market intergraph. (2012). factors affecting construction labor productivity. retrieved from intergraph: https://www.intergraph.com/assets/global/docume nts/spc_laborfactors_whitepaper.pdf jaafar, a. k. (2005). factors affecting construction labour productivity for malaysian residential projects. structural survey, 23(1), 42-54. jarkas, a. (2015). factors influencing labour productivity in bahrain's construction industry. international journal of construction management, 15(1), 94-108. jarkas, a., and bitar, c. (2012). factors affecting construction labor productivity in kuwait. journal of construction engineering and management, 138(7), 811-820. jarkas, a., al balushi, r., and raveendranath, p. productivity factors in construction smes in omanexploration study 40 (2015). determinants of construction labour productivity in oman. international journal of construction management, 15(4), 332-344. jarkas, a., kadri, c., and younes, j. (2012). a survey of factors influencing the productivity of construction operatives in the state of qatar. international journal of construction management, 12(3), 1-23. khaleel, t., and nassar, y. (2018). identification and analysis of factors affecting labour productivity in iraq. matec web of conferences, 1-11. mahamid, i. (2013). contractors perspective toward factors affecting labor productivity in building construction. engineering, construction and architectural management, 20(5), 446-460. padia, s. (2020). construction 100 smart cities by 2020. retrieved from invest india: https://www.investindia.gov.in/sector/construction pornthepkasemsant, p., and charoenpornpattana, s. (2019). identification of factors affecting productivity in thailand’s construction industry and proposed maturity model for improvement of the productivity. journal of engineering, design and technology, 17(5), 849-861. shurrab, m. (2018). evaluating the effect of motivational dimensions on the construction project managers in jordan. engineering, construction and architectural management, 25(3), 412-424. soekiman, p. (2011). factors relating to labor productivity affecting the project schedule performance in indonesia. procedia engineering, procedia engineering, 865–873. tribune, f. (2016). construction market in iran to see strong growth. retrieved from financial tribune: https://financialtribune.com/articles/economybusiness-and-markets/52917/construction-marketin-iran-to-see-strong-growth tuan hai, d. (2019). analysis of affected factors on construction productivity in vietnam. international journal of civil engineering and technology, 10(12), 854–864. whiteside, j. (2006). construction productivity. retrieved from http://www.icoste.org/: http://www.icoste.org/aace2006%20papers/est0 8.pdf 1. introduction 2. literature review 3. problem statement 4. methodology 5. discussion of results conflict of interest funding references the journal of engineering research (tjer), vol. 17, no. 1, (2020) 34-40 *corresponding author’s e-mail: saabta75@yahoo.com doi:10.24200/tjer.vol17iss1pp34-40  a c o m p a r a t i v e s t u d y o f p h y s i c a l p r o p e r t i e s u s i n g v a r i o u s gr a d e s a s p h a l t b i n d e r w i t h d i f f e r e n t t y p e o f f i b e r s sady a. tayh a,*, rana a. yousif b, and qais s. banyhussana a highway and transportation engineering department, faculty of engineering, mustansiriyah university, baghdad, iraq b highway and transportation engineering department, faculty of engineering, mustansiriyah university, baghdad, iraq abstract: for a long time, bitumen has been utilized as the essential material for asphalt pavement construction. the factors of increasing axle loads, increasing traffic movement, critical climate conditions and many forms failures in construction have steered many researchers to seek some methods to enhance the asphalt binder properties. even though various types of modifiers have been utilized in strengthening asphalt concrete, fibers have attracted the most attention due to their high and desirable characteristics. it is realized that the good distribution of the modifier in asphalt binder can generate a strong network in the interior structure of the blend, causing bitumen mastic to be more coherent. in this study, a laboratory investigation of the rheological and physical properties of various grades of bitumen modified by two types of fibers was conducted. three grades of asphalt were used in this study (60-70 penetration grade, 80-100 penetration grade and pg-76 grade) with two types of fibers with different percentagescellulose oil palm fiber (copf) (0.15, 0.3, 0.45, 0.6, and 0.75%) by weight of asphalt and carbon fiber (0.75, 1.25, 1.75, 2.25, and 2.75%) by weight of asphalt. the results showed enhancement in physical performance of the modified bitumen in terms of the decrease in penetration values, as well as a rise in the softening point and viscosity values. the fibers’ modified asphalt binders showed improved rheological properties and can raise the grade of asphalt depending on the base asphalt type. keywords: carbon fiber; cellulose oil palm fiber; asphalt grade; modified bitumen; rheology. االسفلتية المثبتاتمختلفة من درجاتلتأثير أنواع مختلفة من األلياف على الخصائص الفيزيائية قيس صاحب بني حسن و رنا عامر يوسف ، * سعدي عبد تايه المحور على حمولة ثقل العوامل مثل و لقد دفعت عدة سفلت. األ في الرصف ب كمادة أساسية يستخدم القار بعيد كان وال يزال منذ زمن :الملخص قوية بعض الطرق لتلسعي إليجاد في البناء العديد من الباحثين ل قصوروالعديد من أشكال ال قاسيةوزيادة حركة المرور والظروف المناخية ال ، إال أن األلياف جذبت سفلت األ مواد تماسك في تقوية مواد المحسنة على الرغم من استخدام أنواع مختلفة من الو. سفلت األ تماسكمواد خصائص شكل ت سفلت يمكن أن األ تماسك ل للمادة المضافة المتساوي توزيع الأن الحظنا قد ووالمرغوبة. عالية أكبر قدر من االهتمام بسبب خصائصها ال حول مخبريهفحوصات إجراء ب ألدراسةو قمنا في هذه . كبيرا تماسًكا عجينة االسفلت تماسكمما يؤدي إلى للمزيجشبكة قوية في البنية الداخلية درجة )7060ثالث درجات من األسفلت ( منا استخد حيث نوعين من األلياف. باستخدام والفيزيائية لمختلف أنواع القار المعدل حركية الخواص ال ، copf) (0.15(الزيت نخيل السليلوز من ) مع نوعين من األلياف بنسب مختلفة. ألياف (pg-76درجة إختراق و درجة 10080إختراق ، سفلت. أظهرت النتائج وزن األ من )٪ 2.75، 2.25، 1.75، 1.25، 0.75وزن األسفلت وألياف الكربون ()٪ من 0.75، 0.6، 0.45، 0.3 مواد تماسك المعدل من حيث انخفاض قيم االختراق ، وكذلك ارتفاع في درجة التليين وقيم اللزوجة. وأظهرت لألسفلت الفيزيائي اء تحسن في األد سفلت حسب نوع األسفلت األساسي.األ جودة ولوجية محسنة يمكن أن ترفع درجة ب أللياف خصائص ي با سفلت المعدلة األ .لخواص الحركية ا ؛ القار المعدل ؛ االسفلت درجة ؛ نخيل الزيت ؛ الياف السليلوز ؛ الكربون ألياف الكلمات المفتاحية: a comparative study of physical properties using various grades asphalt binder with different type of fibers  35  1. introduction the continuous increase in axle loads, increase in traffic movement, critical environment circumstances, and failures in pavement construction as well as the an enormous increase in asphalt binder prices have justified endeavors of many researchers to seek some methods to enhance bitumen properties by producing higher viscosity low-priced asphalt for pavement construction. there is an extensive range of applications of bitumen modifiers in road construction (muniandy et al., 2008). regarding the rise in the consumption of bitumen all over the globe, several experiments were conducted to boost the characteristics of such product. besides the developments in the field of designing concept and applying asphalt pavement, these trials have led to using more suitable materials required for the construction of asphalt pavement. such materials may be used straight as modifiers to strengthen or to enhance the fundamental components of bitumen (airey, 2004). 1.1 carbon fiber modifier many types of additives were utilized to maintain the phase composition and advance the engineering properties and supply more strength to asphalt pavement (ahmedzade, 2013). lately, the appliance of fibers, as a method for asphalt modification, has concerned many researchers. this is due to their intrinsic compatibility with bitumen and exceptional mechanical characteristics, as carbon fibers could supply a good promise for asphalt modification. modifying asphalt mastic with carbon fiber is expected to increase stiffness and permanent deformation resistance as well as improving fatigue characteristics of the mixture. the cold service temperature behavior of carbon fiber-modified asphalt mixtures was expected to be boosted as well due to the high tensile strength of the carbon fiber (cleven, 2000). nejad et al. (2014) conducted research utilizing carbon fiber to strengthen asphalt concrete. the outcomes showed that the use of carbon fibers in the asphalt concrete can noticeably improve the mechanical properties that benefit all the related fields involved issues like construction and maintenance. physical properties of various types of fibers and their enhancing and strengthening influence on bitumen mastic properties were studied by many researchers (ling et al., 2014; wu et al., 2015). their results revealed that fibers can shape a three-dimensional space composition in bitumen that may transfer and buffer the applied loads. this network can offer a decrease in drain-down of asphalt binder from the mixture throughout forming a thicker mastic coating for aggregate particles. fibers can efficiently enhance the rutting resistance of asphalt mastic. 1.2 cellulose fiber modifier the introduction of cellulose fibers, from annually renewable resources, is now a well-liked prevalence in the reinforcement of asphalt mastic. these offer advantages to the environment in relation to the degradability as well as utilization of materials from natural resources (rout et al., 2001). cellulose is respected as the main framework element of the fiber structure. the benefits of natural lingo-cellulosic fibers over conventional strengthening materials are the satisfactory mechanical properties, low price, low density, suitable thermal properties, non-abrasivity, and improved energy salvage (doan et al., 2006; abiola et al., 2014). there are some researches that has attempted to investigate the potential of cellulose fibers to bolster asphalt binder. kumar et al. (2004) studied the modification of sma with two types of fibers, jute and synthetic. jute fiber sma mixtures showed better permanent deformation and higher creep modulus than synthetic fibers sma mixtures. yang et al. (2006) investigated the resistance to deformation resistance of cellulose and polyester fibers modified bitumen. the achieved results displayed the significant improvement of the rut resistance. however, polyester fibers were relatively better regarding the rutting resistance than cellulose fibers samples do. some researchers evaluated the influence of date palm and textile fibers on open graded course mixture characteristics. the results have shown that the utilization of fibers in the asphalt mixture led to a considerable decrease in drain-down contrasted to other types of modified mixtures (hassan and aljabri, 2005; sharma and goyal, 2006). muniandy and huat (2010) used cellulose fiber extracted from oil palm bunch to study the rheological performance of asphalt binders. the tests have shown the base binder (pg64-22) could be upgraded to pg70-22 grade as well as improving the fatigue performance of sma mixture. the optimum cellulose fiber to be introduced to the base bitumen was found to be 0.6%. another research conducted by muniandy et al. (2008) dealt with the advantage of the date empty fruit bunch (defb) and oil palm trees to generate cellulose fiber to be used as additives in the bitumen. the outcomes showed that the used fibers improved the rheological properties of the asphalt mastic by increasing the pg grading two steps from pg58 to be pg70 at 0.3% oil palm cellulose fiber concentration. other researchers have attempted to study the influence of several cellulose fiber types and percentages on rheological and physical rheological properties of bitumen. the cellulose fiber modified bitumen displayed an increase in viscosity and softening point values as well as a reduction in penetration with increasing cellulose fiber content. the rheological tests showed an improvement in rutting resistance and fatigue cracking (maniruzzaman et al., 2015a; maniruzzaman et al., 2015b; al-otaibi et al., 2016; bonica et al., 2016). according to the literature stated above, the sady a. tayh, rana a. yousif, and qais s. banyhussan 36  application of fibers and tire crumb rubber in asphalt may enhance the mechanical behavior represented by fatigue performance, strength, and marshall stability. however, for comparative purposes, the extent of the effect of each type of modifiers and their optimum proportion of adding them to the base asphalt cement have a big influence on the implementation of bitumen. so, the major objective of this study is to realize the best additive amongst them and the best percentage for strengthening bitumen mortar throughout a series of tests on the rheological and physical properties of the modified and unmodified bitumens. to this end, cellulose oil palm fiber, carbon fibers, and tire crumb rubber with different proportions by weight of the binder are added to three types of binders (80/100, 60/70, and pg75) to be tested and evaluated. 2. materials and methods 2.1 asphalt binders traditionally the 80/100 binder is incapable of opposing the heavy weights from different trucks because it does not have a suitable fatigue and rutting performance. the use of modifiers such as crumb rubber and fibers may improve the service properties of the asphalt mastic by different mechanisms. therefore, for the purpose of investigating the effect different additives on different grade asphalt binders, three grades of asphalt were used as base asphalt binders (80-100 and 60-70 penetration grade asphalts as well as pg-76 performance grade asphalt). the laboratory assessments were conducted in order to estimate the binders’ properties for penetration, softening point, viscosity, specific gravity, and ductility. table 1 shows the physical properties of the three base binders used in this study. 2.2 additives two types of additives were used in this study: 2.2.1 carbon fiber: the physical properties of carbon fiber employed in this research are summarized in table 2. 2.2.2. cellulose oil palm fiber (copf): the oil palm bunches were collected from a plantation area then cut into small pieces. the cellulose oil palm fibers were made by chemical-r of pulping method. the oil palm fibers used are with a moisture content of about 5% and a cellulose content of at least 75% and a ph of 7±5%. the ash content is 3.5% and the density is about 1.5 g/cm3. the cellulose oil palm fiber size distributions are presented in table 3. 2.3 preparation of modifier-asphalt mastic modified asphalt binder specimens were made according to standard techniques employed and published in several journal researches. all the prepared samples were derived from the same neat asphalt binder sources. two kinds of fibers (cellulose oil palm fiber, and carbon fibers) were used as modifier. a high shear mixer was used to achieve an excellent distribution of fibers into the base binders with a revolution of about 1000 rpm. the fibers were discretely placed into the oven at 105°c for 24 h to make sure that the least moisture, and a 500 g sample of bitumen were put into a container and heated at 160°c for 30 min to soften it prior to blending. to explore the effect of fibers on bitumen mortar, various fiber proportions in the asphalt mortars were added. the cellulose fibers (copf) contents used were (0.15, 0.3, 0.45, 0.6 and 0.75%) by weight of asphalt, while the carbon fiber contents used were (0.75, 1.25, 1.75, 2.25 & 2.75%) by weight of asphalt. blending then continued at 160°c for 30 min to provide a homogeneous blend. in order to prevent the contrary influence of extreme heating, the temperature was cautiously observed using a thermometer with a thermocouple probe. after completion, the modified binder was extracted from the metal tins and separated into small containers enclosed by aluminum foil then kept for further testing at room temperature. 3. results and discussion 3.1. conventional tests the traditional tests, like penetration (astm d5), softening point (astm d36) and viscosity (astm d4402), were first conducted for the neat binders and then for modified binders according to standards. 3.1.1 penetration test: figs. 1 and 2 show penetration test results for neat and modified asphalt binders. from this figure, it can be noticed that, with the increase in the percentage of modifiers, the value of penetration decreases. on the other hand, the figures show that cellulose fiber modified binders have higher values of penetration as compared with carbon fiber asphalt binder samples. the penetration value of the modified high-quality binder samples (pg76) did not change significantly. table 1. physical properties of asphalt cement. binder properties specification 80-100 pg-76 60-70 penetration (25 °c, 100gm, 5 sec.) (0.1mm) astm d5 84 33.5 63.5 softening point (ring & ball) °c astm d36 46.5 73.5 49 viscosity at 135°c (cp) astm d4402 297.45 1541.8 382.3 ductility (25°c, 5cm/min). astm d113 >100 >100 >100 specific gravity at 25°c (g/cm3 ) astm d70 1.036 1.03 1.04 a comparative study of physical properties using various grades asphalt binder with different type of fibers  37  table 2. physical properties of carbon fiber. source 100% carbon fiber from japan tensile strength (gpa) 3.53 0 yarn type 12k carbon fiber tensile modulus (gpa) 230 chopped length (µm) 150 carbon content (%) ≥ 95 filament diameter (µm) 7.0 density (g/cm3) 1.76 table 3. cellulose oil palm fiber size distribution. no. sieve size (µm) % passing 1 580 90 2 425 80 3 250 65 4 180 55 5 150 35 6 106 20 table 4. the maximum allowable percentage of additives to each base binder type. additive type binder type max. additive % carbone fiber 80-100 1.75 60-70 1.35 pg76 1.1 cellulose fiber 80-100 0.28 60-70 0.23 pg76 0.1 figure 1. effect of carbon fiber content on penetration for asphalt binders. figure 2. effect of cellulose fiber content on penetration for asphalt binders. figure 3. effect of carbon fiber content on softening point for asphalt binders. figure 4. effect of cellulose fiber content on softening point for asphalt binders. figure 5. effect of carbon fiber content on viscosity for asphalt binders. figure 6. effect of cellulose fiber content on viscosity for asphalt binders. 3.1.2. softening point test: the outcomes of the softening point test for the three types of binders and two types of fibers are shown in figures 3 and 4. it can be seen, as opposed to the penetration measurements, with increasing the fiber percentage, the softening point increased. there were somewhat significant changes in softening point values liable on the sort of the base binder. however, the softening point values for the high-quality asphalt binder (pg76) modified by sady a. tayh, rana a. yousif, and qais s. banyhussan 38  the two fiber types did not change significantly. 3.1.3. viscosity tests: figures 5 and 6 show the viscosity values of the control and modified bitumen. the viscosity values of the fiber modified bitumen are greater than the base bitumen. the more modifiers are introduced into the base bitumen, the greater viscosity values are gained. some viscosity values of the fiber modified asphalt binders were higher than 3 pa.s, and so didn’t fulfill the requirement standard of superpave™. based on that, the recommended proportions to add each additive to each of the three base binders used in this study were estimated and presented in table 4. the viscosity value is a crucial issue for transportation of bitumen. excessively high value of viscosity is unfavorable when transporting bitumen using pipelines. thus, over-viscous bitumens are inappropriate for use in pavement industry. based on the previous outcomes, the introduction of the two fibers into the base asphalt binders enhances the physical properties of the modified bitumen binders as indicated by the reduction in penetration value and increase in softening point, and an increase in the viscosity values, thus enhancing modified binder stiffness and increasing its capability to resist rutting deformation. 3.2 rheological properties for modified asphalt binder the dynamic rheological characteristics of the cellulose modified asphalt mastic were determined utilizing a dynamic shear rheometer (dsr, haake, rheo stress rs1, phoenix) throughout an extensive extent of temperatures to illustrate both viscous and elastic behaviors of unaged modified binders. two 25mm diameter parallel plates with a gap of 1mm were determined to do the dsr measurements. a constant sinusoidal stress of 0.12 kpa was directed at low strain to keep all the tests performed within the linear viscoelastic limit. the real strain and rotation were measured for determining viscoelastic factors, counting complex modulus (g*) and phase angle (δ). furthermore, the rutting factor (g*/sinδ) was utilized to primarily measure the rutting resistance and stiffness of the asphalt mastics (yi-qiu et al., 2010). 3.2.1. temperature sweep test at high temperature: the fundamental rheological properties of the neat asphalt and modified bitumens, cellulose fibers (copf) and carbon fiber by temperature sweep test at un-aged state using dynamic shear rheometers (dsr). a dsr oscillation temperature sweep test was performed at intermediate and high temperatures that ranged from 40 to 82°c with a 6-degree step to determine the change in binder performance as a function of temperature and 10 rad/s (1.59 hz) frequency. a minimum value of 1.0 kpa for g*/sin δ is recommended to ensure that asphalt binder could resist well against permanent deformation at the designed performance grade temperature. the minimum limit for g*/sin δ is 1000 pa for an unaged asphalt binder. figs. 7 and 12 present the relation between the rutting factor and temperature during temperature sweep test showing the failure temperatures for the two types of modified binders. the cellulose modified bitumens have greater values of rutting factor than the control bitumens. consistent with the specification, these modified bitumens have a reduced rutting susceptibility, especially at high service temperature contrasted to the base binders. from table 5, it can be seen that the failure temperatures of the modified asphalt binders are greater than the base asphalt binders. as a result, the high-temperature grade (pg) for the base binders has been increased. the addition of the two fiber types into the base bitumens caused them to be stiffer. greater g* and g*/sinδ of these modified bitumens show that the bitumen could sustain additional energy when loading is directed, and similarly, they reveal that the modified binders have better resistance to rutting and fatigue cracking under intermediate and high service temperatures. the asphalt binders modified with carbon fiber mostly have similar resistance to rutting, which are clearly better than the control bitumens. at the same time, the base binder pg76 was the highest in terms of rutting resistance and stiffness followed by 60/70 and 80/100 binders respectively. so, the control binder source has a great impact on the rutting parameter values of various binders in this research. it can be evidently seen that the unmodified binders are categorized as pg58, pg64, and pg76, respectively. coming back to the maximum failure percent of each additive to each control asphalt type according to the viscosity test results, the addition of 1.75 and 1.35% carbon fiber to 80-100 and 60-70 asphalt binders, respectively, had enhanced the pg grading two steps for 80-100 binder from pg58 to pg70, and enhanced pg grading one step from pg64 to pg70, while the effect of adding 1.1% of carbon fiber to pg76 binder did not change the pg grading significantly. the same trend is occurring with the addition of cellulose fiber to the three base binders, with slight differences. the addition of 0.28 and 0.23% cellulose fiber to 80-100 and 60-70 asphalt binders, respectively, had enhanced the pg grading two steps for 80-100 binder from pg58 to pg70 and enhanced pg grading one step from pg64 to pg70, while the effect of adding 0.1% of cellulose fiber to pg76 binder did not change the pg grading significantly. 4. conclusion in this paper, the effects of adding cellulose oil palm fiber, and carbon fiber to three asphalt binder types were studied. there are several conclusions derived from the results:  the two modifiers are well dispersed in the asphalt, due to the fact that no blocks or agglomerates can be seen after the a comparative study of physical properties using various grades asphalt binder with different type of fibers  39 blending process.  figure 7. temperature sweep test for 80-100 asphalt binder modified with carbon fiber. figure 8. temperature sweep test for 60-70 asphalt binder modified with carbon fiber. figure 9. temperature sweep test for pg76 asphalt binder modified with carbon fiber. figure 10. temperature sweep test for 80-100 asphalt binder modified with cellulose fiber. figure 11. temperature sweep test for 60-70 asphalt binder modified with cellulose fiber. figure 12. temperature sweep test for pg76 asphalt binder modified with cellulose fiber. table 5. failure temperature of the modified binders.  the addition of carbon fiber and cellulose fibers to the asphalt binder improves the physical properties of the base asphalt binders as indicated by the decrease in penetration value and rise in softening point temperatures.  the introduction of the two modifiers into base asphalt binders can considerably improve the viscosity of the bitumen, which is helpful to enhance the high temperature performance of asphalt binders. cellulose and carbon fibers displayed very high viscosity values when added at high percentages to the base asphalt binders that exceed the specification requirement of 3 pa.s at 135°c.  cellulose fibers have the most significant enhancement to viscosity and rutting factor of asphalt binders as compared with carbon fibers.  asphalt cement source may influence the behavior sady a. tayh, rana a. yousif, and qais s. banyhussan 40  of the modified mastic in different ways.  the quality of the base asphalt binder is a crucial factor. higher quality base asphalt binder showed little enhancements (pg76), whereas the lowest quality binders exhibit considerable enhancement in the conducted tests (80-100 and 60-70 binders).  the dsr tests presented that fiber modified bitumen mastics have higher rutting parameters (g*/sinδ) contrasted by with the base bitumen mastics employed in this research. thus, fibers can significantly enhance asphalt mastic flow. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. acknowledgment we would like to express our thanks and gratitude to all those who contributed to this modest work, especially the technical staff of the asphalt technology laboratory at the university of putra, malaysia, as well as the central transport laboratory at the faculty of engineering, mustansiriya university. references abiola os, kupolati wk, sadiku er, and ndambuki jm (2014), utilisation of natural fibre as modifier in bituminous mixes: a review. construction and building materials 54: 305-312. ahmedzade p (2013), the investigation and comparison effects of sbs and sbs with new reactive terpolymer. construction and building materials 38:285-291. airey gd (2004), fundamental binder and practical mixture evaluation of polymer modified bituminous materials. international journal of pavement engineering 5(3): 137-151. al-otaibi hm, al-suhaibani as, and alsoliman ha (2016), physical and rheological properties of asphalt modified with cellulose date palm fibers. world academy of science, engineering and technology, international journal of civil, environmental, structural, construction and architectural engineering 10(5): 583-587. bonica c, toraldo e, andena l, marano c, and mariani e (2016), the effects of fibers on the performance of bituminous mastics for road pavements. composites part b: engineering 95: 76-81. cleven m.a. (2000), “investigation of the properties of carbon fiber modified asphalt mixtures,” master thesis, michigan technological university, mishigan. doan ttl, gao sl, and mäder e (2006), jute/polypropylene composites i. effect of matrix modification. composites science and technology 66(7-8): 952-963. hassan hf, and al-jabri ks (2005), effect of organic fibers on open-graded friction course mixture properties. international journal of pavement engineering 6(1): 67-75. kumar p, sikdar pk, bose s, and chandra s (2004), use of jute fibre in stone matrix asphalt. road materials and pavement design 5(2): 239-249. ling pa, peng wa, bo li, pan pa, and shaopeng wu (2014), investigation of rheological characteristics of carbon fiber modified asphalt binder. key engineering materials 599. maniruzzaman am, hamad aw, maleka am, and elsergany m (2015a), rheological properties of cellulose oil palm fiber (copf) modified 80100 asphalt binder. journal of advanced research in materials science 5(1): 10-20. maniruzzaman am, hamad aw, maleka am, jakarni fm, and bahru uj (2015b), effect of viscoelastic behavior of cellulose oil palm fiber (copf) modified 60-70 asphalt binder for deterioration for roads and highways. jurnal teknologi 75(11): 17-23. muniandy r, and huat bb (2006), laboratory diameteral fatigue performance of stone matrix asphalt with cellulose oil palm fiber. american journal of applied sciences 3(9): 2005-2010. muniandy r, jafariahangari h, yunus r, and hassim s (2008), determination of rheological properties of bio mastic asphalt. american journal of engineering and applied sciences 1(3): 204-209. nejad fm, vadood m, and baeetabar s (2014), investigating the mechanical properties of carbon fibre-reinforced asphalt concrete. road materials and pavement design 15(2): 465-475. rout j, misra m, tripathy ss, nayak sk, and mohanty ak (2001), the influence of fibre treatment on the performance of coir-polyester composites. composites science and technology 61(9): 1303-1310. sharma v, and goyal s (2006), comparative study of performance of natural fibres and crumb rubber modified stone matrix asphalt mixtures. canadian journal of civil engineering 33(2): 134-139. wu mm, li r, zhang yz, fan l, lv yc, and wei jm (2015), stabilizing and reinforcing effects of different fibers on asphalt mortar performance. petroleum science 12(1): 189-196. yang j, shi x, wan j, qian g, pan w, and yang y (2006), evaluation of rutting resistance of doublelayered asphalt mixes. road materials and pavement design 7: 533–542. yi-qiu t, li x, and zhou x (2010), interactions of granite and asphalt based on the rheological characteristics. journal of materials in civil engineering 22(8): 820-825. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 19, no. 2, (2022) 95-105 0043 corresponding author's e-mail: mbadi@squ.edu.om doi:10.53540/tjer.vol19iss2pp95-105 submitted on 2021-08-26 accepted on 2022-05-26 exploring the impacts of covid-19 pandemic on oman's electricity sector m. albadi1,3, *, a. malik1, a. al badi1, s. al rawahi2, m. al housni2, m. al shabibi2, 1department of electrical and computer engineering, sultan qaboos university, muscat, oman 2 load dispatch centre, oman electricity transmission company saoc, muscat, oman 3 arab open university – oman, muscat, oman abstract: this article reviews the recent trends of oman's electricity sector before the covid-19 pandemic outbreak. the impacts of the pandemic on the main interconnected system (mis) of oman were analyzed using hourly load data. the analysis shows that the mis demand declined as a result of the decrease in economic activities during the lockdown. in addition, the mis demand experienced temporal and geographical variations: the former is demonstrated by a shift in peak demand hours, while the latter is represented by a reduction in muscat's urban areas' load compared with those of other areas. keywords: oman; covid-19 pandemic; electricity sector; main interconnected system; lockdown, load change. عمان على قطاع الكهرباء في سلطنة 19-آثار جائحة كوفيد م الشبيبي ،م الحوسني ،ص الرواحي ،ع البادي ، ع مالك ،م البادي . تم تحليل covid-19يستعرض هذا المقال االتجاهات الحديثة لقطاع الكهرباء في سلطنة عمان قبل تفشي جائحة :الملخص باستخدام بيانات األحمال الكهربائية المسجلة. ( في سلطنة عمان mis) ةالمترابط ةالرئيسي الشبكة الكهربائيةتأثيرات الوباء على يظهر التحليل أن الطلب على الكهرباء قد انخفض نتيجة النخفاض األنشطة االقتصادية أثناء اإلغالق. باإلضافة إلى ذلك، شهد نما تجلت التغيرات الطلب على الكهرباء تغيرات زمانية ومكانية. تجلت التغيرات الزمانية من خالل تحول ساعات ذروة الطلب، بي المكانية في انخفاض الطلب على الكهرباء في المناطق الحضرية في محافظة مسقط مقارنة بتلك الموجودة في المحافظات األخرى. عمان :الكلمات المفتاحية الكهرباء ؛covid-19جائحة ؛سلطنة المترابطة ؛قطاع الرئيسية الكهربائية ؛ اإلغالق ؛الشبكة . االحمالتغير impacts of covid-19 on oman's electricity sector 96 nomenclature aer: authority for electricity regulation apsr: authority for public services regulation bst: bulk supply tariff covid-19: coronavirus disease 2019 crt: cost reflective tariff ipps: independent power producers iea: international energy agency iwps: independent power and water producers medc: muscat electricity distribution company mis: main interconnected system mjec: majan electricity company msm: muscat securities market mzec: mazoon electricity company oetc: oman electricity transmission company opwp: oman power and water procurement company raec: rural areas electricity company tou: time-of-use tariff. 1. introduction the outbreak of covid-19 was reported in wuhan, china, in december 2019 and continues worldwide. as of october 8, 2020, the covid-19 pandemic has had over 36.3 million registered cases and claimed over 1 million lives (worldometers.info 2020). apart from human suffering and death, the epidemic caused a significant global economic downturn. the largest economies in the world have been among the most affected by the pandemic (elgin, basbug et al., 2020). the covid-19 pandemic has had direct adverse economic effects in various ways. to mention a few, isolated or hospitalized employees cannot join the workforce with numerous demandand supply-side consequences. to lower covid-19 transmission rates and reduce the burden on healthcare systems, governments have adopted a wide range of restrictive public health measures, including school and factory closures, travel restrictions, and community lockdowns (atkeson 2020). as seen in china, singapore, and hong kong, these steps have helped limit the increase in new infections (anderson, heesterbeek et al., 2020). however, such measures have slowed down economic activity by restricting human mobility and industry. the covid-19 pandemic and associated public health regulations specifically disrupted supply chains and decreased activity in the manufacturing and service sectors, leading to increased layoffs. stock markets have plummeted around the world, and unemployment claims have risen to unprecedented levels. millions of employees have lost their jobs and cannot pay their credit card bills. restaurants and shops are only slowly reopening; many cannot pay their rent. industrial companies cannot make payments on their equipment leases. landlords have less income and cannot keep up with their mortgages. suddenly, the world is awash in credit risk (nerc 2018). lockdown initiatives have significantly lowered the need for power in both the consumer and manufacturing sectors over the last few months. the international energy agency (iea) (iea 2020, iea 2020) reported that global electricity demand declined by 2.5% in the first quarter of 2020. as a result of the shutdown steps, electricity demand drastically decreased, with knock-on effects on the power mix. during periods of complete lockdown in many countries, electricity demand has been reduced by 20% or more as decreases in commercial and industrial operations greatly outweigh residential demand increases. these demand reductions have raised the share of renewables in electricity supply as such demand generally does not affect their production. demand for all other electricity sources— including coal, gas, and nuclear power—dropped. electricity demand showed the first signs of growth as containment eased in italy and germany in april. this pattern was confirmed in may as lockdown steps were weakened by more countries (e.g., india, france, spain, and the uk). in june, in most countries except india, where the recovery was more pronounced, the electricity demands, weather corrected, remained at 10% below the same month's 2019 level. the demand for electricity in china fell below that during the lockdown in january and more dramatically in february (13% less than in february 2019, the leap year being corrected). part of the discrepancy was also attributed to a relatively colder winter in china in 2019 than in 2020. electricity demand in china fully recovered in june 2020 and was even higher than that last year (iea 2020). the iea report (iea 2020) predicted that all types of fuel demand would be affected. according to the report, oil demand could drop by 9% (around 9 million barrels/day) annually. in large part, coal demand could decline by 8% because electricity demand will be nearly 5% lower throughout the year. gas and nuclear demand will also fall in response to lower demand for power and industrial applications. however, renewables demand is expected to increase because of dispatch priority and recent growth in capacity caused by new projects. the coronavirus situation has proven that renewable energy is more appropriate for the situation because of its decentralized existence and can be a stable supply during the worst unpredictable scenarios (norouzi, de rubens et al. 2020). global co2 emissions are predicted to fall by 8% to levels seen ten years ago, according to the iea (iea 2020). the report also warns that the rebound in emissions may be larger than the decline unless the wave of investment to restart the economy is dedicated to cleaner and more resilient energy infrastructure. the international finance corporation reported (bakovic, kroese et al.) a 15% decrease in demand in several countries where it does business on average in march and april 2020. slower growth in demand as a m. albadi, a. malik, a. al badi, s. al rawah, m. al housni, and m. al shabibi 97 consequence of declining covid-19–led economic activity could hold oil prices down, yet uncertainty is to be anticipated. low oil prices would benefit oilimporting nations, especially those where oil prices are linked to natural gas. given the economic downturn in china and the rising shale gas output in the united states, natural gas prices were still at historic lows before covid-19. spot rates in countries with a significant share of renewables hit values near zero marginal costs (e.g., at some hours in brazil, mexico, peru, and turkey). according to an international study published in the journal nature climate change (le quéré, jackson et al. 2020) examining daily co2 emissions from 69 countries, all 50 u.s. states, and 30 chinese provinces, reflecting a total of 85% of the world population as well as 97% of global co2 emissions, carbon emissions dropped colossally by 17% in april 2020 compared with those last year. emissions from some nations decreased by as much as 26% on average during the peak of april confinement measures. according to the study, global emissions could decline by 4.4% to 8% by the end of this year. decreasing factory production, fewer cars on the road, and less power generation contributed to this decline. moreover, global renewable energy production increased by 3%, mainly powered by new online solar and wind projects over the past year. in one study (buechler, powell et al., 2020), the authors developed a unified modelling framework to quantify and compare electricity usage changes across countries and regions. this study was applied to demand data for electricity from 58 countries and regions around the world, representing about 60% of the world's population and 75% of global electricity demand. utilizing the 53 countries' daily data, the authors found that the total electricity demand declined by as much as 10% in april 2020 compared with the modelled demand, adjusting for weather, seasonal, and temporal effects. this decrease in demand is characterized by four groups, ranging from a mild 2% shift to a severe 26% decline. the authors explored how government restrictions and mobility patterns were related to electricity demand changes and found that these linkages were most pronounced in the extreme and severe impact groups and less so in the moderate and mild ones. this, in turn, can provide essential information in the future, helping utilities, companies that purchase power directly, and public officials respond more effectively to the impacts of future viral outbreaks and ongoing efforts to build more resilient, sustainable infrastructure systems. a global reduction in electrical energy demand and price was also observed in some countries. the reported drops in consumption reached 25% in italy, 20% in france, and 12% in the uk (mylenka and novyk 2020). one study (haxhimusa and liebensteiner 2020) reported that in 16 european countries, when the covid-19 outbreak was at its peak, electricity demand was reduced by 19%, which resulted in a significant drop in carbon emissions by 34%. the delay or nonpayment of utility bills by end consumers will have a detrimental effect on distribution, transmission, and generation companies. thus, the lockdowns have caused delays in power projects because of supply chain disruptions, workforce unavailability, and project financing issues. these developments confirm the dependence of the global economy on the energy system. all countries should consider energy as a basic human right, and governments should seek their supplies from more than one country to avoid delays in energy projects during future global crises. the renewable energy sector was negatively impacted by covid-19, mainly related to delays in the power plants' equipment delivery or construction. the spread of the virus has caused several industries' shutdowns and issues along the supply chain across multiple industry sectors. contractors relying on an international workforce are also impacted as travel restrictions, or quarantine measures are imposed, leading to labour shortages. project developers may be facing penalties or, in some cases, losing tax incentives, tariffs, or other revenue sources. for example, byd, a producer of rechargeable batteries, was unable to complete the required testing for some batteries, which led to delays in delivering the needed quantities of these batteries to the european market (mylenka and novyk 2020). furthermore, while governments in developing countries are big buyers of renewable energy products, the available funds are diverted for medical purposes because of the virus, which further hurts renewable energy projects. the renewable energy sector's primary global concern is supply chain issues, with solar and wind projects already witnessing logistical delays. the effects of covid-19 on energy systems have been studied in terms of system operation (elavarasan, shafiullah et al. 2020, noorazar, srivastava et al. 2020, safari, price et al. 2020), demand and load forecasting (agdas and barooah 2020, chen, yang et al. 2020), energy policy (graff and carley 2020), and energy markets (graf, quaglia et al. 2020, wang, yang et al. 2020, ruan, wu et al. 2021). the pandemic has raised many unanswered questions about the long-term operation and planning of power systems (navon, machlev et al., 2021), such as the incorporation of renewable energy sources (hosseini, 2020) and their impact on frequency stability (milano, dörfler et al. 2018), power system expansion planning in the light of the covid-19 pandemic (gorenstin, campodonico et al. 1993), and the need for the availability of new tools to system operators in times of global health crisis (wang, yang et al. 2020). most papers in the literature that estimate changes in demand during the covid-19 pandemic account for temperature and temporal variables (e.g., seasonality, day of the week). some papers do this by using a predictive model to estimate what demand impacts of covid-19 on oman's electricity sector 98 would have been during 2020 without a pandemic, incorporating temperature and temporal influencing factors into regression models (prol and sungmin 2020, ruan, wu et al. 2020). others take an econometric approach, estimating changes via fixed effects in the model controlling for key electricity demand factors such as time of day, day of the week, week of the year, and temperature, essentially comparing electricity demand for an hour during covid-19 to that for a similar hour in a similar week of the year after controlling for temperature (leach, rivers et al. 2020). another approach uses a direct comparison to historical data, adjusting the results using the temperature sensitivity of the load (mcwilliams and zachmann 2020). as the economic activities of modern societies are heavily dependent on the use of electricity, monitoring electricity consumption patterns can provide some insight into the impact of lockdowns that are imposed due to pandemics such as covid-19 (janzen and radulescu 2020). this paper analyzes the demand profile of the mis system after the pandemic compared to the previous year for the same date and time. the comparison is also made at the three distribution company levels. although the covid-19 pandemic affected the global economy adversely, at the same time, it had some positive impacts. these include; increasing global renewable energy production, low oil prices that benefit oil-importing nations, reduction in global co2 emissions, reduction of electrical energy demand, reducing the flight services and providing many services online. following this introduction, the article will review recent trends in oman's electricity sector before the covid-19 pandemic in section 2. afterwards, the article presents the direct implications of covid-19 on the main interconnected system in terms of spatial and temporal changes. in addition, section 3 presents some indirect implications of covid-19 on the electricity sector in oman. section 4 discusses the recommendations related to the operation and planning of the omani power sector. section 5 presents the main conclusions of the article. 2. recent trends in oman's electricity sector the demand in mis represents about 88% of the consumption in oman's electricity sector (apsr 2020). this study focuses on the demand change patterns of the main interconnected system (mis) of oman. the recent trends of oman's electricity sector are summarized in the below sections. 2.1 overview of oman's electricity sector before may 2005, oman's electricity sector was a vertically integrated system owned and operated by the ministry of housing, electricity, and water (albadi 2017, albadi, al-badi et al. 2020). the electricity sector law, which was issued on july 7, 2004, through royal decree 78/2004, established a new market structure. generation, transmission and distribution activities were unbundled through the establishment of different government-owned companies. in addition, the authority of electricity regulation (aer) was established to regulate the new market structure (apsr 2020). in august 2020, the aer was renamed the authority for public services regulation (apsr), as its mandate was extended to include electricity, water, and gas public services. currently, there are three distinct power systems in oman: the main interconnected system (mis) in the northern part of the country (fig. 1), the dhofar power system (dps) in the south, and small isolated systems in the rural areas electricity company (raec). in 2019, the number of registered customer accounts reached 1,114,502, 121,190, and 41,586 in the mis, dps, and raec, respectively. the supplied electricity reached 29.618769 twh, 3.075685 twh, and 1.101589 twh in the mis, dps, and raec, respectively (aer 2019). 2.2 mis demand trends since restructuring the recent mis trends are summarized in the figures below. the supplied energy of the mis increased consciously from 8.4 twh in 2005 to 29.62 in 2019 (fig. 2) but did not increase afterwards. this might be attributed to the application of the cost-reflective tariff (crt) and the use of energy efficiency measures. a similar trend can be found in the mis's peak demand. apart from the drastic increase in supplied energy, another significant achievement of electricity sector restructuring in oman is the continuous improvement of the mis's efficiency. its transmission and distribution losses decreased from almost 25% to about 8% (fig. 3). however, the subsidies increased from omr 66.3 million in 2005 to omr 441 million in 2019 because of the increase in power-subsidized demand and capacity investments (fig. 4). this increase is attributed to the increase in supplied energy and the fact that permitted tariffs are lower than the economic cost (albadi 2017). figure. 1. the structure of oman's main interconnected system (mis). m. albadi, a. malik, a. al badi, s. al rawah, m. al housni, and m. al shabibi 99 figure. 2. evolvement of mis-supplied energy (twh) and peak demand (gw). 2.3 impacts of cost-reflective tariff to encourage economic efficiency, the oman power and water procurement company (opwp) purchases energy from the producer based on a bulk supply tariff (bst) that reflects the marginal cost. the bst energy charge has a time-of-use (tou) tariff structure, and it is applied to distribution companies (medc, mzec, mjec) but not end-use consumers. to reduce subsidies, large industrial, commercial, and government customers who consume more than 150 mwh per year started paying the crts from january 2017 onward. the crt has several components, including the bst energy charge (omr/mwh), a transmission charge (omr/mw), a distribution charge (omr/mwh), and a fixed service charge (albadi 2017). these charges are subject to the approval of the apsr. the transmission charge is linked to the system's peak demand. therefore, the crt's application on large customers helped reduce both the mis's peak demand and supplied energy. crt customers started energy efficiency programs as well as rooftop pv projects. 2.4 recent investments and projects according to the opwp, the first solar power plant in ibri (500 mw) will be commissioned in 2021 (opwp 2020). three more solar independent power producers (ipps) of a similar size will be commissioned in 2022, 2023, and 2024 as well as a wind ipp of 100 mw in the sharqiyah region for completion in 2023 (opwp 2020). this will enable the opwp to reach its target of 16% electricity production from new solar, wind, and waste-to-energy projects by 2025. apart from renewable energy projects, other major developments in the mis include activating the electricity spot market in 2020, connecting the mis to duqm via a 400 kv transmission line in 2023, and creating demand response programs (opwp 2020). on december 15, 2019, authorities announced that 49% of the oman electricity transmission company (oetc) shares would be sold to china state grid international development (oetc 2020). this $1 billion deal is the start of partial privatization projects that will include other nama holding group government-owned companies, such as medc (nama 2020). figure. 3. evolvement of mis losses since electricity sector restructuring. figure. 4. evolvement of mis subsidies since electricity sector restructuring. 3. implications of covid-19 on mis 3.1 covid-19 in oman on february 24, 2020, the first two cases of covid19 were confirmed, and the first death was recorded on april 1 of the same year. since then, the country has restricted movement among the governorates by installing checkpoints on entry and exit points on the roads. on april 10, the muscat governorate was put under lockdown, which was lifted on may 29. the purpose of the lockdown was to reduce confirmed cases and family gatherings during the month of ramadhan and the eid al-fitr seasons. a second lockdown was established among all governorates, including the main roads and inter-governorate roads, from july 25 to august 8, 2020 (moh 2020). the second lockdown was to restrict social activities and family gatherings during the eid al-adha season. these restrictions had a direct impact on oman's economic activities and reduced the demand for electricity. 3.2 capacity investment charge since the beginning of 2020, oil prices have been falling, primarily caused by the decline in oil demand because of global lockdown measures to contain covid-19. the decline in prices is compounded by the failure of opec+ to compromise on supply cuts (haxhimusa and liebensteiner 2020). oil revenues are now set to plummet further in 2020, following a 0 2 4 6 8 0 10 20 30 40 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 m is n e t s u p p li e d ( t w h ) m is p e a k d e m a d ( g w ) year mis net supplied twh 24.6 22 21.7 18.719.1 16.2 13.614.412.611.6 10 9.2 8.8 8.4 8.4 0 5 10 15 20 25 30 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 m is l o ss e s (% ) year 66.3 88 60.763.1 109.5 111.8 105 184.6 226.4 214.1 344.2 365.3 362 418 441 0 50 100 150 200 250 300 350 400 450 500 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 m is s u b si d ie s (m il li o n o m r ) year impacts of covid-19 on oman's electricity sector 100 massive wave of declining global demand set off by the pandemic. oman's economy is expected to shrink by 2.8% in 2020. despite significant diversification efforts in recent years, the country's fiscal position is still highly exposed to oil price fluctuation as oil contributes nearly 35% of the gdp, 74% of revenue (with gas), and 66% of export. oman's budget breaks even at almost $87 per barrel, so sub-$30 oil, coupled with a lower production quota under the april opec+ deal, will imply a nearly 17% gaping budget deficit in 2020 (cornell 2020). the mis's electricity demand declined as a result of the lockdown and the decrease in commercial and industrial activities. because of the financial implications of covid-19 on oil prices, on may 17, 2020, the opwp informed eight ipps and independent water producers (iwps) listed on the muscat securities market (msm) about its intention to temporarily suspend investment fees from energy and water capacity fees. however, on may 27, 2020, the opwp retracted this letter following a statement confirming the ministry of finance's commitment to continue providing the necessary financing for the electricity sector, which was announced on may 21, 2020 (opwp 2020). 3.3 load changes caused by covid-19 omani authorities have been taking precautionary measures against the spread of covid-19 since the first covid-19 case was detected in oman on february 24, 2020. by mid-march, schools, colleges, and universities were closed. the supreme committee of oman, under the chairmanship of the minister of interior, was tasked with handling the developments resulting from the spread of covid-19. the committee made more tough decisions to prevent the spread of the virus from march 18, 2020, onward. these included the closures of restaurants, shopping malls, religious places, shops and marketplaces, sports and cultural clubs, sports halls, courts, health clubs, barbershops and beauty salons, etc. the committee ordered the suspension of all gatherings, activities, and conferences, closed all tourist sites, restricted entry to the sultanate from all land, sea, and air entry points with the exception of omani citizens, and stopped omanis from leaving the sultanate. on wednesday, april 8, 2020, omani authorities ordered a 12-day lockdown of the muscat province to prevent the further spread of covid-19. according to their directive, all movement into and out of the province was prohibited between friday, april 10, and wednesday, april 22. this lockdown was further extended until may 29, 2020. for comparison purposes, we have selected two periods from march 18 to april 9 and then from april 10 to may 29 in the years 2019 and 2020 to observe the pattern of the mis's load data and the three distribution companies' load data during the mentioned periods. the second lockdown was much stricter than the first, which included banning people from entering or leaving the muscat governorate, which was implemented on april 10. the hourly data used in this analysis were obtained from the load dispatch centre of oetc. a. march 18 to april 9 figure 5 and table 1 compare the mis's load and weather temperature from march 1 to march 17 of 2019 and 2020, a period before the lockdown occurred in 2020. before the lockdown, the overall peak during this period was 3,280 mw in 2019 and 3,359 mw in 2020, with a positive difference of 80 mw. the average load during the same period was 2,635 mw in 2019 and 2,822 mw in 2020, with a positive difference of 187 mw. energy consumption during this period was 1,075,212 mwh in 2019 and 1,151,388 mwh in 2020, with a positive difference of 76,175 mwh. the weather temperature during this period was also, on average, lower in 2019 compared with that 2020. in 2019, the average temperature for the period was 23.2°c, and in 2020, it was 24.2°c. figure 6 and table 1 compare the mis's load and temperature from march 18 to april 9 in 2019 and 2020, during a lockdown period in 2020. the overall peak during this period was 4,757 mw in 2019 and 4,006 mw in 2020, with a negative difference of 752 mw. the average load during the same period was 3,246 mw in 2019 and 2,833 mw in 2020, with a negative difference of 413 mw. energy consumption during this period was 1,792,046 mwh in 2019 and 1,563,914 mwh in 2020, with a negative difference of 228,132 mwh. the weather temperature during this period was also, on average, higher in 2019 compared with that 2020. in 2019, the average temperature was 28.0°c, and in 2020, it was 26.8°c. some of the higher loads in 2019 can be attributed to the weather during this period, but overall, a significant load reduction can be observed in 2020 after the lockdown. it is also possible that weather temperature is also affected by the lockdown as the previous trend of 2019 from march 1-17 is a bit on the higher side. per the load growth forecast, a growth of 2%–3% was expected in 2020. however, the load growth appears to be limited and balanced by the decrease in demand caused by the lockdown. thus, the overall load trend seems to match last year's trend if the weather is corrected. the closures of shopping malls, mosques, schools, industrial areas, and commercial activities have resulted in a significant drop in system load. an increase in residential load may have been caused by the lockdown, partly balancing the reduction in commercial and industrial loads. m. albadi, a. malik, a. al badi, s. al rawah, m. al housni, and m. al shabibi 101 table 1. comparison of system load, energy, and weather temperature for the period 1-17 march and march 18–april 10 in 2019 and 2020 for mis. march 1 to march 17 march 18 to april 10 2019 2020 diff. 2019 2020 diff. max load (mw) 3280 3359 80 max load (mw) 4757 4006 -752 avg. load (mw) 2635 2822 187 avg. load (mw) 3246 2833 -413 energy (mwh) 1075213 1151388 76175 energy (mwh) 1792046 1563915 -228132 max temp (ºc) 31.9 31.4 -0.5 max temp (ºc) 40.9 40.0 avg. temp (ºc) 23.2 24.2 1 avg. temp (ºc) 28.0 26.8 table 2. comparison of peak load and energy for the period april 10–may 29 in 2019 and 2020 for distribution companies and mis companies load/energy 2019 2020 difference medc max. load (mw) 2243 2012 -231 energy (mwh) 1882871 1794286 -88584 load factor (%) 70% 74.3% 4.4% mzec max load (mw) 1936 2007 71 energy (mwh) 1522088 1643013 120925 load factor (%) 65.5% 68.2% 2.7% mjec max. load (mw) 1528 1470 -58 energy (mwh) 1348166 1359617 11450 load factor (%) 73.5% 77.1% 3.5% mis (total) max. load (mw) 5536 5415 -121 energy (mwh) 4753125 4796916 43791 load factor (%) 71.6% 73.8% 2.3% figure. 5. comparison of mis load and temperature for the period march 1–17 in 2019 and 2020. 18.0 23.0 28.0 33.0 38.0 43.0 48.0 0 500 1000 1500 2000 2500 3000 3500 t e m p ( ºc ) m is l o a d ( m w ) load 2019 load 2020 temperature 2019 temperature 2020 impacts of covid-19 on oman's electricity sector 102 figure. 6. comparison of mis load and temperature for the period march 18–april 9 in 2019 and 2020. b. april 10 to may 29 figures 7–10 show the daily peak trend of load for the distribution companies and the mis for the period 2019–2020. table 2 compares the maximum load and energy experienced by distribution companies and the mis from april 10 to may 29 in 2019 and 2020. in the case of medc, both peak demand and energy were reduced in 2020, whereas for mzec, energy consumption and maximum demand in 2020 were higher than those in the previous year. the reason is the composition of industrial and commercial, and residential sectors in medc and mzec. since most of the commercial activities take place in the muscat area, the medc load was reduced during the lockdown period. another factor is that most of the people who work in the muscat area come from different parts of oman, and when the lockdown was announced, they moved back to their respective places. overall, the mis's peak load was reduced by 2.1% (121 mw); however, this year's energy consumption was higher than that of the last year for this period. lower peak demand and higher energy consumption are suitable for electric utilities as this suggests a better utilization of assets. there is a yardstick used called the load factor for this purpose. the load factor (lf) is defined as follows: 𝐿𝐹 = 𝐸𝑛𝑒𝑟𝑔𝑦 𝐶𝑜𝑛𝑢𝑚𝑝𝑡𝑖𝑜𝑛(𝑀𝑊ℎ) 𝑃𝑒𝑎𝑘 𝑑𝑒𝑚𝑎𝑛𝑑 (𝑀𝑊)×ℎ𝑜𝑢𝑟𝑠 (ℎ𝑟𝑠) × 100% (1) the load factor could be annual, seasonal, etc., depending on the selected period. the higher the load factor, the more efficient use of existing resources and enhanced social welfare. therefore, we observe that the system load factor improved from 71.6% to 73.8% during the lockdown period. it may be noted that system peak is not simply adding a maximum load of distribution companies. this is because the maximum demand of these distribution companies is not coinciding with the system peak. the load factors of distribution companies have also improved during the lockdown period. it may be noted that although the peak load of mzec increased in 2020, the load factor of mzec still improved by 2.7%. one of the highlight and lesson learned from lockdowns are that the system load factor and distribution companies' load factors improved, reflecting greater efficiency and enhancing social welfare as consumption during non-peak hours increased, making better use of utility plants and lowering the average cost of electricity. figure. 7. comparison of medc load for the period april 10–may 29 in 2019 and 2020. figure. 8. comparison of mzec load for the period april 10–may 29 in 2019 and 2020. 18.0 23.0 28.0 33.0 38.0 43.0 48.0 53.0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 t e m p ( ºc ) m is l o a d ( m w ) load 2019 load 2020 temperature 2019 temperature 2020 0 500 1000 1500 2000 2500 d e m a n d ( m w ) medc 2019 medc 2020 0 500 1000 1500 2000 2500 d e m a n d ( m w ) mzec 2019 mzec 2020 m. albadi, a. malik, a. al badi, s. al rawah, m. al housni, and m. al shabibi 103 figure. 9. comparison of mjec load for the period april 10–may 29 in 2019 and 2020. figure. 10. comparison of mis load for the period april 10–may 29 in 2019 and 2020. 4. recommendations for power sector's operation and planning we can learn from the recommendations and lessons learned in different parts of the world regarding power sector operation and planning during the pandemic time in the following references (elavarasan, shafiullah et al. 2020, noorazar, srivastava et al. 2020, safari, price et al. 2020) (navon, machlev et al. 2021). here we will discuss some technical and nontechnical recommendations that are more relevant to the omani power sector. 1. power companies need to ensure the safety and security of their employees. 2. because of the immobility, consumers may not pay bills during this time. as a result, utilities should take precautions to avoid disconnecting customers due to late payments or insufficient funds in their accounts. 3. the government may compensate for the financial loss to power companies that may occur because of the nonpayments of electricity bills or due to suppressed demand. 4. with the changing situation of covid-19 infection cases and fresh waves of variants of the virus, the supreme committee's decisions also change, sometimes putting more restrictions on lockdown and travel ban and sometimes relaxing the restrictions, which lead to the unpredictability of electricity demand. therefore, there is a need to have more accurate short-term load forecasting for economic dispatch and to minimize the risk of electricity interruptions. 5. opwp should keep its renewable target as planned earlier, and there is no need to delay the procurement of renewable energy plants because of the suppressed demand. similarly, power distribution companies should keep investing in renewable energy plants to support the government policy on renewables. however, mis should account for the technical difficulties faced by system operator when the share of renewable become large. the high share of renewables may negatively impact the grid stability resulting in frequency fluctuations and high ramp rates. for that ramp rate, control algorithms with an energy storage system should be developed. in this regard, the interconnection with gcc countries will also help to stabilize the grid. 6. in order to fulfil peak load demand, demand-side load management strategies decrease investment costs in the power generation and transmission sectors. oman should not only look into demandside management programs as a policy but also promote energy efficiency measures in industrial and commercial sectors, such as energy-efficient green buildings. 7. since there will be more restrictions in coming years on countries regarding greenhouse gas emissions, it is, therefore, recommended that opwp consider a generation mix with nuclear as an option in their long-term plans. 8. the modern trend is moving from the conventional grid to the smart grid (al abri, malik et al. 2020). oman should consider modernizing its grid toward a smart grid. the smart grid will give a lot of flexibility in meeting and matching the demand with supply under uncertain conditions like a pandemic. an epri (electric power research institute) report outlines an initial approach to quantifying and monetizing about 22 smart grid benefits, which include: optimized generation operation; deferred generation, transmission, and distribution capacity; reduced ancillary service costs; reduced congestion costs; reduced equipment failure; reduced maintenance costs; reduced meter reading costs; reduced electricity theft; reduced emissions; reduced wide-scale blackouts, and so on (wakefield 2010, malik and bouzguenda 2013). as a case study in oman, a research paper has evaluated the demand-side management (dsm), energy efficiency measures and distributed generation benefits of the smart grid in the mis system of oman. grid enhancement, customer contribution to the grid, and both options at the same time are among the scenarios developed. the scenarios are evaluated for peak reduction, and their benefits are calculated in terms of avoided generation, transmission, and distribution costs, as well as environmental costs. 0 500 1000 1500 2000 d e m a n d ( m w ) mjec 2019 mjec 2020 0 1000 2000 3000 4000 5000 6000 d e m a n d ( m w ) mis 2019 mis 2020 impacts of covid-19 on oman's electricity sector 104 5. conclusions the covid-19 pandemic has had adverse effects on different sectors that are essential for people's livelihoods, such as health systems, different industrial and commercial sectors, the oil market, the international markets, the labour market, and the electricity market. oman's energy sector, like those of the rest of the world, has been gravely affected by covid-19. energy demand decreased after the reduction of commercial and industrial activities and the closure of schools. the reduction of the industrial load has also had a significant impact on the country. the government and society have taken steps to support those who cannot pay their energy bills after losing their jobs. although oman has not faced severe human casualties, its economic activities have been affected. in oman, oil and gas contributed nearly 74% of the revenue in 2018, and global demand has declined this year (group 2020). thus, the overall economy shrunk by 2.8% in 2020. oman's government has announced relief measures to mitigate this economic impact and help individuals and businesses navigate these unprecedented times. some examples of these measures are the decreasing interest rate and tax on some activities and the postponement of loan instalments for small and medium establishments for several months. the positive impact of covid-19 is that the world has witnessed changes consistent with forecasted developments, such as digital solutions, low emissions, and clean energy. thus, this new investment should be geared toward building a healthier, more resilient, and netzero-emissions economy. innovation in both technologies and business models is the primary driver of economic growth. acknowledgment the authors acknowledge the support received from oetc and squ. conflict of interest the authors declare that there are no conflicts of interest regarding this article. funding the authors did not receive any funding for this research. references aer. (2019). "authority of electricity regulation annual report." retrieved august 2019, retreieved from https://www.aer.om/en/about. agdas, d. and p. barooah (2020). "impact of the covid-19 pandemic on the us electricity demand and supply: an early view from data." ieee access 8: 151523-151534. al abri, d., a. malik, s. al-saadi, m. albadi, y. charabi and n. hosseinzadeh (2020). smart grids and smart buildings. handbook of climate change mitigation and adaptation. m. lackner, b. sajjadi and w.-y. chen. new york, ny, springer new york: 1-56. albadi, m., a. al-badi, r. ghorbani, a. al-hinai and r. al-abri (2020). "enhancing electricity supply mix in oman with energy storage systems: a case study." international journal of sustainable engineering: 110. albadi, m. h. (2017). "electricity sector in oman after 10 years of reform: status, trends, and future perspectives." the electricity journal 30(7): 23-30. anderson, r. m., h. heesterbeek, d. klinkenberg and t. d. hollingsworth (2020). "how will countrybased mitigation measures influence the course of the covid-19 epidemic?" the lancet 395(10228): 931934. apsr. (2020). "authority for public services regulation website." retreieved from https://www.apsr.om/index.php/en/home. atkeson, a. (2020). what will be the economic impact of covid-19 in the us? rough estimates of disease scenarios, national bureau of economic research. bakovic, t., r. kroese, n. lama and e. akcura. (june 2020). "the impact of covid-19 on the power sector." retrieved august 10 2020, from https://www.ifc.org/wps/wcm/connect/f73f9cf33abd-4378-b5b6-c8eb8c4c1b45/ifc-covid19powersector-final_web_rev.pdf?mod=ajperes &cvid=n9.o4sq. buechler, e., s. powell, t. sun, c. zanocco, n. astier, j. bolorinos, j. flora, h. boudet and r. rajagopal (2020). "power and the pandemic: exploring global changes in electricity demand during covid-19." arxiv preprint arxiv:2008.06988. chen, y., w. yang and b. zhang (2020). "using mobility for electrical load forecasting during the covid-19 pandemic." arxiv preprint arxiv:2006.08826. cornell, p. (2020). "after the coronavirus and low oil prices, oman's future is more global." retreieved from https://www.atlanticcouncil.org/blogs/ mena source/after-the-coronavirus-and-low-oil-pricesoman s-future-is-more-global/ on september 3 2020. elavarasan, r. m., g. shafiullah, k. raju, v. mudgal, m. t. arif, t. jamal, s. subramanian, v. s. balaguru, k. reddy and u. subramaniam (2020). "covid-19: impact analysis and recommendations for power sector operation." applied energy 279: 115739. elgin, c., g. basbug and a. yalaman (2020). "economic policy responses to a pandemic: developing the covid-19 economic stimulus index." covid economics 1(3): 40-53. https://www.aer.om/en/about https://www.apsr.om/index.php/en/home m. albadi, a. malik, a. al badi, s. al rawah, m. al housni, and m. al shabibi 105 gorenstin, b., n. campodonico, j. costa and m. pereira (1993). "power system expansion planning under uncertainty." ieee transactions on power systems 8(1): 129-136. graf, c., f. quaglia and f. a. wolak (2020). "learning about electricity market performance with a large share of renewables from the covid-19 lockdown." program energy sustain. dev: 1-21. graff, m. and s. carley (2020). "covid-19 assistance needs to target energy insecurity." nature energy 5(5): 352-354. group, o. b. (2020). "diversification and value addition eyed as large-scale energy projects come online in oman." retreieved from https://oxfordbusinessgroup.com/overview/adaptevolve-diversification-and-value-addition-remainlong-term-goals-raft-large-scale-projects. haxhimusa, a. and m. liebensteiner (2020). "effect of covid-19 on power sector emissions." hosseini, s. e. (2020). "an outlook on the global development of renewable and sustainable energy at the time of covid-19." energy research & social science 68: 101633. iea (2020). covid-19 impact on electricity. paris, france, iea. iea (2020). global energy review 2020. paris. janzen, b. and d. radulescu (2020). "electricity use as a real-time indicator of the economic burden of the covid-19-related lockdown: evidence from switzerland." cesifo economic studies 66(4): 303321. le quéré, c., r. b. jackson, m. w. jones, a. j. p. smith, s. abernethy, r. m. andrew, a. j. de-gol, d. r. willis, y. shan, j. g. canadell, p. friedlingstein, f. creutzig and g. p. peters (2020). "temporary reduction in daily global co2 emissions during the covid-19 forced confinement." nature climate change 10(7): 647-653. leach, a., n. rivers and b. shaffer (2020). "canadian electricity markets during the covid-19 pandemic: an initial assessment." canadian public policy 46(s2): s145-s159. malik, a. s. and m. bouzguenda (2013). "effects of smart grid technologies on capacity and energy savings–a case study of oman." energy 54: 365-371. mcwilliams, b. and g. zachmann (2020). electricity consumption as a near real-time indicator of covid 19 economic effects. iaee energy forum/covid-19 issue 2020. milano, f., f. dörfler, g. hug, d. j. hill and g. verbič (2018). foundations and challenges of low-inertia systems. 2018 power systems computation conference (pscc), ieee. moh. (2020). "omanvscovid19." retreieved from covid19.moh.gov.om. mylenka, t. and b. novyk (2020). "impact of covid19 on the global energy sector." pv magazine 24. nama. (2020). "nama group website." retreieved from https://www.nama.om/. navon, a., r. machlev, d. carmon, a. e. onile, j. belikov and y. levron (2021). "effects of the covid-19 pandemic on energy systems and electric power grids—a review of the challenges ahead." energies 14(4): 1056. nerc (2018). probabilistic adequacy and measures. atlanta, ga, usa, , north american electric reliability corporation (nerc). noorazar, h., a. srivastava, k. s. sajan and s. pannala (2020). "data-driven operation of the resilient electric grid: a case of covid-19." arxiv preprint arxiv:2010.01746. norouzi, n., g. z. de rubens, s. choubanpishehzafar and p. enevoldsen (2020). "when pandemics impact economies and climate change: exploring the impacts of covid-19 on oil and electricity demand in china." energy research & social science 68: 101654. oetc. (2020). "oman electricity transmission company s.a.o.c website." from https://www.omangrid.com/en/pages/home.aspx. opwp. (2020). "oman power and water procurement company website." retreieved from www.omanpwp.om. opwp (2020). opwp's 7-year statement (2019 – 2025) prol, j. l. and o. sungmin (2020). "impact of covid19 measures on short-term electricity consumption in the most affected eu countries and usa states." iscience 23(10): 101639. ruan, g., d. wu, x. zheng, h. zhong, c. kang, m. a. dahleh, s. sivaranjani and l. xie (2020). "a crossdomain approach to analyzing the short-run impact of covid-19 on the us electricity sector." joule 4(11): 2322-2337. ruan, g., j. wu, h. zhong, q. xia and l. xie (2021). "quantitative assessment of us bulk power systems and market operations during the covid-19 pandemic." applied energy 286: 116354. safari, n., g. price and c. chung (2020). "comprehensive assessment of covid‐19 impact on saskatchewan power system operations." iet generation, transmission & distribution. wakefield, m. (2010). "methodological approach for estimating the benefits and costs of smart grid demonstration projects." electric power research institute: palo alto, ca, usa. wang, b., z. yang, j. xuan and k. jiao (2020). "crises and opportunities in terms of energy and ai technologies during the covid-19 pandemic." energy and ai 1: 100013. worldometers.info. (2020). "covid-19 coronavirus pandemic." retreieved from https://www.worldometers.info/coronavirus/. https://www.nama.om/ https://www.omangrid.com/en/pages/home.aspx file:///c:/users/farouk/appdata/local/temp/rar$dia10100.35607/www.omanpwp.om https://www.worldometers.info/coronavirus/ elec170803.qxd the journal of engineering research vol. 2, no. 1 (2005) 12-24 1. introduction 1.1 takagi-sugeno models developing mathematical models of real systems is a central topic in many disciplines of engineering and science. models can be used for simulations, analysis of the system's behavior, better understanding of the underlying mechanisms in the system, design of new processes, or design of controllers. takagi-sugeno (t-s) modeling plays an essential role in deriving local linear models of the nonlinear dynamic system under concern (lo and chen, 1999; takagi and sugeno, 1985). through the use of the heuristic rules inherent in the fuzzy systems, t-s fuzzy models, then make it possible to have a transparent like system which is governed by the fuzzy inference system and rules. fuzzy modeling concerns the methods of describing the characteristics of a system using fuzzy inference rules. fuzzy modeling methods have a distin _______________________________________ *corresponding author e-mail: ebrgallaf@eng.uob.bh guishing feature in that they can express complex nonlinear systems linguistically. in a similar way fuzzy clustering has been utilized as well in classifying data-driven fuzzy modeling, since it draws a methodology for assigning label to similar data. such assignment does give quantitative directions for shaping the fuzzy membership functions. model validation and verification is also an important task in the modeling paradigm. this is due to the choice of the right model from a number of models that might present similar characteristics. statistically validated models in addition to probabilistic validation are used sometimes to make the suitable choice of a system model. in general, fuzzy control systems can be classified as linguistic (lo and chen, 1999; mamdani and assilian, 1975). the linguistic type fuzzy control system is well recognized and received by the control society. the t-s type fuzzy system, which will be used in this article mainly focuses on the modeling aspect. it has been reported that a t-s fuzzy system can exactly model any nonlinear system wang, et al. (2000). on the other hand there is a main drawback takagi-sugeno neuro-fuzzy modeling of a multivariable nonlinear antenna system e. a. al-gallaf* deptartment of electrical and electronics engineering, college of engineering university of bahrain, p.o. box 13184, kingdom of bahrain received 17 august 2003; accepted 7 april 2004 abstract: this article investigates the use of a clustered based neuro-fuzzy system to nonlinear dynamic system modeling. it is focused on the modeling via takagi-sugeno (t-s) modeling procedure and the employment of fuzzy clustering to generate suitable initial membership functions. the t-s fuzzy modeling has been applied to model a nonlinear antenna dynamic system with two coupled inputs and outputs. compared to other well-known approximation techniques such as artificial neural networks, the employed neuro-fuzzy system has provided a more transparent representation of the nonlinear antenna system under study, mainly due to the possible linguistic interpretation in the form of rules. created initial memberships are then employed to construct suitable t-s models. furthermore, the t-s fuzzy models have been validated and checked through the use of some standard model validation techniques (like the correlation functions). this intelligent modeling scheme is very useful once making complicated systems linguistically transparent in terms of the fuzzy if-then rules. keywords : neuro-fuzzy systems, fuzzy clustering, takagi-sugeno modeling, nonlinear systems ::¢¢üü∏∏îîàà°°ùÿÿggò¨dg ᫵«eéæj~dg äé«∏ª©∏d êpé‰ og~yg πlg øe ∂dph ºµëàdg º¶æd (á«fhîµd’g á«ñ°ü©dg äéµñ°ûdg πãe) á«còdg º«∏©àdg πfé°sh ωg~îà°sg ¤g åëñdg gòg ±~¡j )ωg~îà°séh á∏«ã“ ogôÿg ωé¶ædg áhééà°sg ™jrƒj ” ~≤a qéw’g gòg ‘h .ájoé«ày’g π«ãªàdg πfé°sƒh é¡∏«ã“ ö©°üj »àdgh √~©dg ägò¨àÿg ägp á«£ntechnique clustering)ωé¶f aéæh h π«ãªàd äé≤ñw á°ùªn øe áfƒµÿg á«fhîµd’g á«ñ°ü©dg äéµñ°ûdg ωg~îà°sg ” ºk øe h (logicfuzzyêpƒ‰ aéæñd ω~îà°ùj ±ƒ°s …òdgh ( ) ä’é°üjg ωé¶æd ƒg h »µ«eéæjo »£n òz ωé¶ædantenna systemmulti-variableqoéb ìî≤ÿg êpƒªædg ¿g äéñk’ á«fé°üm’g ¥ô£dg ωg~îà°sg ” gòngh .( .ºµëàdg ωé¶f aéæh ºk øeh i~≤©ÿg äé«∏ª©dg π«ã“ ≈∏y áá««mmééààøøÿÿgg ääggooôôøøÿÿgg .á«£nódg ᪶fc’g ,ƒæ«cƒ°s »cécéj ᪶fcg öñ° ÿg …oƒ≤æ©dg ™ªéàdg áññ° ÿg á«ñ°ü©dg ᪶fc’g : ägò¨àÿg o~©àeh »£n òz »fgƒg ωé¶æd öñ° ÿg ≥£æÿgh á«ñ°ü©dg éjóÿg ≈∏y ~ªà©e ƒæ«cƒ°s »lécéj êpƒ‰ of the linguistic model compared with the t-s model in that there is a difficulty in dealing with a multidimensional system since a large number of fuzzy rules have to be used. gorzalczany et al. (2000), has briefly presented and compared four neuro-fuzzy systems used for rule-based modeling of dynamic processes (chaotic mackey-glass time series). the following systems have been considered: nfmod the proposed system, the well-known anfis and nfident systems, and an alternative neurofuzzy system already reported in literature. the main criterion of comparison of all systems is their performance (modeling accuracy) versus interpretability (the transparency and the ability to explain generated decisions; it also includes an analysis and pruning of obtained fuzzyrule bases). on the other hand, zhang and knoll (1995). have proposed an approach for solving multivariate modeling problems with neuro-fuzzy systems. instead of using selected input variables, statistical indices are extracted to feed a fuzzy controller. the original input space was transformed into an eigen-space. if a sequence of training data are sampled in a local context, a small number of eigenvectors which possess larger eigen-values provide a good summary of all the original variables. fuzzy controllers can be trained for mapping the input projection in the eigen-space to the outputs. implementations with the prediction of time series was used to validate the concept. the article of ikonen and kortela (2000) is concerned with a process modeling using fuzzy neural networks. in distributed logic processors (dlp) the rule base is parameterized. the dlp derivatives required by gradientbased training methods are given, and the recursive prediction error method is used to adjust the model parameters. the power of the approach is illustrated with a modeling example where nox-emission data from a full-scale fluidized-bed combustion district heating plant are used. the method presented in their paper was general, and can be applied to other complex processes as well. bologna (2001) has presented a new neuro-fuzzy model denoted as fuzzy discretized interpretable multi-layer perceptron (fdimlp). fuzzy rules were extracted in polynomial time with respect to the size of the problem and the size of the network. he applied our model to three classification problems of the public domain. it turned out that fdimlp networks compared favorably with respect to efunn and anfis neuro-fuzzy systems. ning et al. (2001) presented a fuzzy satisfactory clustering algorithm in their paper. it started with two cluster centers and adds new center if necessary. a system data set was quickly divided into several satisfactory fuzzy clusters by the algorithm. a t-s type fuzzy model was then, identified. chen and linkens (1998) introduced a three-layered rbf (redial basis function) network to implement a fuzzy model. differing from existing clustering-based methods, in their approach the structure identification of the fuzzy model, including input selecting and partition validating, was implemented on the basis of a class of sub-clusters created by a self-organizing network instead of raw data. the important input variables which independently and significantly influence the system output can be extracted by a fuzzy neural network. on the other hand, the optimal number of fuzzy rules can be determined separately via the fuzzy c-means clustering algorithm with a modified fuzzy entropy measure as the criterion of cluster validation. akkizidis and roberts (2001) proposed an algorithmic methodology for identifying and modeling non-linear control strategies. the methodology presented was based on choices of different fuzzy clustering algorithms, projection of clusters and merging techniques. the best features of well-known clustering methods such as the gustafsonkessel and mountain method were combined. the latter was used to determine and define the number and the approximate positions of the cluster prototypes; whereas the former was used to define the shapes of the clusters according to the data distribution. the projection of the prototypes and variables of clusters was a recognized approach to extracting the information included in the data clusters into fuzzy sets. merging these fuzzy sets, based on proposed guidelines, can minimize the number of rules and make the identifying control strategy more transparent. bossley (1997) has looked into the problem of antenna modeling via neuro-fuzzy systems, however, getting an optimized five layers neural network was not easily achieved due to the large number of generated fuzzy rules. 1.2 article contribution the system under study is typical of the type used for oceanary satellite communication systems and has a high nonlinear coupling among its two outputs. hence, it is required to have transparent sub-models. this class of multivariable system has been modeled via a classical neuro-fuzzy system as in bossley (1997). however, it did result in a large number of rules, and large number of training patterns were required. in this respect, this research frame work is investigating the use of clustered fuzzy rules, that makes it easy for the training mechanism to be achieved in less time with fewer number of rules. fuzzy sets in the antecedent of the rules are obtained from the partition matrix by projection onto certain antecedent variables. the obtained point-wise fuzzy sets are then approximated by some suitable parametric functions. the transparency of the antenna model obtained using the above approach may be hindered by the redundancy present in the form of many overlapping (compatible) membership functions. certain similarity measures were used in order to assess the compatibility (pair-wise similarity) of fuzzy sets in the rule base, in order to detect sets that can be merged. fuzzy sets estimated from antenna training data can also be similar to the universal set, thus adding no information to the model. sets of such nature were removed from the antecedent of the rules, thus reducing the number of the fuzzy rules. 13 the journal of engineering research vol. 2, no. 1 (2005) 12-24 2. intelligent dynamic systems modeling 2.1 intelligent modeling fuzzy modeling and control are typical examples of techniques that make use of human knowledge and deductive processes. various alternative approaches have been proposed, fuzzy logic and set theory being one of them. artificial neural networks and fuzzy models belong to the most popular model structures used. from the input-output view, fuzzy systems are flexible mathematical functions, which can approximate other functions or just data measurements with a desired accuracy. compared to wellknown approximation techniques such as neural networks, fuzzy systems provide a more transparent representation of the system under study, which is mainly due to the possible linguistic interpretation in the form of rules. the logical structure of the rules facilitates the understanding and analysis of the model in a semi-qualitative manner, close to the way human reason about the real world. given the state of a system with a given input, the next state x(k + 1) can be determined. in the sense of discrete-time setting, it can be written as: x(k + 1)= f (x(k), u(k)) (1) where x(k) and u(k) are the state and the input at time k, respectively, and f is a static function. fuzzy models of different types can be used to approximate the state-transition function. as the state of a system is often not measured, input-output modeling is usually applied. the most common is the narx (nonlinear auto-regressive with exogenous input) model, as defined by y(k+1) = f ( y(k), y(k-1),…, y(k ny +1), u(k), u(k-1),…, u(k nu +1) ) (2) where y(k) ,…y(k ny + 1) , and u(k) ,…, u(k ny + 1) denote the past model outputs and inputs respectively and ny and nu are integers related to the model order (usually selected by the designer). for instance in eq. (3), a linguistic fuzzy model of a dynamic system may consist of rules of the following form : ri : if y(k) is ai1 and y(k-1) is ai2 and,…, y(kn+1) is ain and u(k) is bi1 and u(k-1) is bi2 and,…, u(k-m+1) is bim then y(k+1) is ci (3) in eq. (3), the input dynamic filter is a simple generator of the lagged inputs and outputs, and no output filter is used. since the fuzzy models can approximate any smooth function to any degree of accuracy, models of the type in eq. (3) can approximate any observable and controllable modes of a large class of discrete-time nonlinear systems. to facilitate data-driven optimization of fuzzy models (learning), differentiable operators (product, sum) are often preferred to the standard min and max operators. once the structure is fixed, the performance of a fuzzy model can be fine-tuned by adjusting its parameters. tunable parameters of linguistic models are the parameters of antecedent and consequent membership functions (determine their shape and position) and the rules (determine the mapping between the antecedent and consequent fuzzy regions). x = (x1…………xn)t (4) y = (y1…………..yn)t (5) 3. neuro-fuzzy modeling figure 1 shows typical five layers of a neuro-fuzzy system that can be employed to accomplish a rule network. typically, such rules are if x1 is a11 and x2 is a21 then y=b1 (6) if x1 is a12 and x2 is a22 then y=b2 (7) nodes in the first layer compute the membership degree of the inputs in the antecedent fuzzy sets. the product nodes π in the second layer represent the antecedent conjunction operator. the normalization node ν and the summation node σ realize the fuzzy-mean operator. using smooth antecedent membership functions, such as a gaussian function, as given below (8) in which cij and τij parameters are adjusted by gradientdescent learning algorithms, such as back-propagation. this allows a fine-tuning of the fuzzy model to the available data in order to optimize its prediction accuracy. there may be a lot of structure/parameter combinations which make the fuzzy model behave in a satisfactory way. the problem can be formulated as that of finding the structure complexity which will give the best performance in generalization. in our approach we choose the number of rules as the measure of complexity to be properly tuned on the basis of available data. we adopt an incremental approach where different architectures having different complexity (i.e. number of rules) are first assessed in cross-validation and then compared in order to select the best one. 14 the journal of engineering research vol. 2, no. 1 (2005) 12-24 it is assumed that a set of n input-output data pairs ( ){ }niy ii ,...,2,1, =x is available. recalling that inpf i ℜ∈x are input vectors and yi are output scalars. denote inpfn ×ℜ∈x a matrix having the vectors tkx in its rows, and nℜ∈y a vector containing the outputs y k: ( ) ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜ ⎝ ⎛ ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ − −= 2 2 exp,, ij ijj ijijjaij cx cx τ τµ 15 the journal of engineering research vol. 2, no. 1 (2005) 12-24 layer no.1 layer no.2 layer no.3 layer no.4 layer no.5 figure 1. a five layer neurofuzzy network architecture local linear model curves of equidistance clusters center v1 v2 data x x y µ (x) projected clusters data if x is a1 then y is b1 if x is a2 then y is b2 a1 a2 b1 b2 µ (y) figure 2. hyper ellipsoidal fuzzy clusters the initialization of the architecture is provided by a hyper-ellipsoid fuzzy clustering procedure inspired by babuska and verburggen (1995). this procedure is clustering the data in the input-output domain obtaining a set of hyper-ellipsoids which are a preliminary rough representation of the input/output mapping. methods for initializing the parameters of a fuzzy inference system form the outcome of the fuzzy clustering procedure. here we use the axes of the ellipsoids (eigenvectors of the scatter matrix) to initialize the parameters of the consequent functions. we project the cluster on the input domain to initialize the centers of the antecedents and we adopt the scatter matrix to compute the width of the membership functions. once the initialization is done, the learning procedure begins. in the case of linear t-s models this minimization procedure can be decomposed in a least-squares problem to estimate the linear parameter of the consequent models and a nonlinear minimization to find the parameters of the membership functions. the structural identification loop (the outer one) searches for the best structure, in terms of optimal number of rules, by increasing gradually the number of local models. 4. fuzzy pattern clustering 4.1 fuzzy clustering identification methods based on fuzzy clustering originate from data analysis and pattern recognition, where the concept of graded membership is employed to represent the degree to which a given object, represented as a vector of features, is similar to some prototypical object. based on that similarity, feature vectors can be clustered such that vectors within a cluster are as similar as possible, and vectors from different clusters are as dissimilar as possible. this thought of fuzzy clustering is depicted in fig. 2. data is clustered into two groups with prototypes v1 and v2, using the euclidean distance measure. the partitioning of the data is expressed in the fuzzy partition matrix whose elements µij are degrees of membership of data points (xi, yi) in a fuzzy cluster with prototypes vj. the concept of similarity of data to a given prototype leaves enough space for the choice of an appropriate distance measure and of the character of the prototype itself. prototypes can be defined as linear subspaces, or the clusters can be ellipsoids with adaptively determined shape akkizidis and roberts, (2001). from these clusters, the antecedent membership functions and the consequent parameters of the t-s model can be extracted as follows, (bossley, 1997). if x is a1 then y = a1x + b1 if x is a2 then y = a2x + b2 (9) each obtained cluster is represented by one rule in the t-s model. membership functions for fuzzy sets a1 and a2 are generated by pointwise projection of the partition matrix onto the antecedent variables. such pointwise defined fuzzy sets are then approximated by a suitable parametric function. 4.2 fuzzy clustering algorithm u = [uij]i i = 1,...,c, j = 1,...,n (10) (11) second, every constructed cluster is nonempty and different from the entire set, that is, (12) the general form of the objective function used in fuzzy clustering is (13) where w(xi) is a prior weight for each xi and d(xj, vk) is the degree of dissimilarity between the data xi and the supplemental element vk, which can be considered as the central vector of the kth cluster. degree of dissimilarity is defined as a measure that satisfies two assumptions given by (14) (15) based on the above background, fuzzy clustering can be precisely formulated as an optimization problem: minimize 16 the journal of engineering research vol. 2, no. 1 (2005) 12-24 consider a finite set of elements { }nxxx ,...,, 21=x as being elements of the finp dimensional euclidean space inpfℜ , that is, .,...,2,1, njx inpfj =ℜ∈ the issue is to perform a partition of such collection of elements into c fuzzy sets wi th respect to a given criterion, where c is a given number of clusters. the criterion is usually to optimize an objective function that acts as a performance index of clustering. the end result of fuzzy clustering can be expressed by a parti tion matrix u such that : in eq . (10), iju is a numerical value in [0,1] and expresses the degree to which an element jx belongs to the ith cluster. however, there are two additional constraints on the value of uij. first, a total mem bership of the element xx j ∈ in all classes is equal to unity; that is : ∑ = = c i iju 1 1 for all j=1,2,…,n ∑ << = n j ij nu 1 0 for all i=1,2,…, c ( ) ( )[ ] ( )∑ = ∑ = ∑ = = c i kj n j c k ijikij ,vxduxwgvuj 1 1 1 ,, ( ) ,0≥kj ,vxd ( ) ( )jkkj xvd,vxd = (16) subject to one of the widely employed clustering methods based on eq. (16) is the fuzzy c-means (fcm) algorithm. the objective function of the fcm algorithm is expressed in the form of (17) where m is called exponential weight that influences the degree of fuzziness of the membership (partition) matrix. to solve this minimization problem, the objective function j(uij,vk) in eq. (17) is differentiated with respect to vk ( for fixed uij, i=1,…,c, j=1,…,n ) and to uij ( for fixed vk , i=1,…,c ) and the conditions of eq. (11), are applied obtaining (18) (19) the system described by the eqs. (18) and (19) cannot be solved analytically. however, the fcm algorithm provides an iterative approach to approximating the minimum of the objective function starting from a given position. 5. linear state space models extraction 5.1 t-s fuzzy space model at each sample time k, given an operating point condition (for example u(k 1) and y (k 1), a local linear fuzzy state-space model can be constructed via calculating the degree of fulfillment µi (x(k)) of the antecedents, using product as the fuzzy logic and operator. the inference of the entire structure (hierarchy) due to rule i results in a sub-model (1) which can be expressed as: (20) (21) defining ζ1 , η1 and θ1 as follows : (22) (23) (24) where x(k), u(k) and y(k) for the state-space description are defined as: (25) (26) (27) in order to employ quadratic programming for systems which depend on current as well as on the previous inputs, it is necessary to construct a state-space representation, such that the state vector x(k) to accommodate not only the state variables, appearing in y(k), but also the previous inputs and the offset as last element. this results in a system with only current inputs, but leads to a more complex a-matrix. 17 the journal of engineering research vol. 2, no. 1 (2005) 12-24 ∑ = = c i iju 1 1 for al l j =1,2,…, n and ∑ << = n j ij nu 1 0 for all i=1,2,…, c ( ) ( )[ ] ( )( )[ ]kx kykx ky l r i li lil r i li l ∑ = ∑ = +=+ 1 1 1.1 µ µ( ) ( )[ ] ( ) njcki ,vxd,uxwg,vuj c i kj n j c k ijikij ,...,2,1;,...,2,1, , 1 1 1 == = ∑ = ∑ = ∑ = ( ) 1 1 2 1 >−= ∑ = ∑ = mvxu,vuj c i ij n j m ijkij ( ) ( )∑ =∑ = == nj j m ij n j m ij i c,ixuu v 1 1 ,...,2,1 1 ( ) ( ) njci vx vx u c k m kj m ij ij ,...,2,1;,...,2,1 , 1 1 1 1 1 2 1 1 2 == − − = ∑ = − − ( ) ( ) ( )[ ]lililili kukyky θηζ ++=+ 1 ( )[ ] ( )[ ]kx kx l r i li lil r i li l ∑ = ∑ == 1 1 . µ ζµ ζ ( )[ ] ( )[ ]kx kx l r i li lil r i li l ∑ = ∑ == 1 1 . µ ηµ η ( )( ) ( )( )kx kx l r i li lil r i li l ∑ = ∑ == 1 1 . µ θµ θ ( ) ( ) ( ) ( ) ( ) ( ) ( ) t udin oynon ony nnku kunkx kxnkxkx k ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −− −− − = ...1... ...... 1 111 x ( ) ( ) ( ) ( )[ ]tni kukukuk ...21=u ( ) ( ) ( ) ( )[ ]tni kxkxkxk ...21=y the latter contains also η s, corresponding to the previous inputs. if the maximal delay in the input i, i=1,…,n i is ui,dmax, then the number of the additional columns is ( )∑ −=ini diu1 max, 0,1max . in the last column of a are st ored the offsets θ s. the columns with η s correspond to the previous inputs, stored in the state vector; these columns are not included in b. the ones in a correspond to the delayed values of a certain variable. the local linear system matrices are derived as follows: a is αα × square matrix, where (28) (29) (30) the ones in c are positioned such that y1(k) = x1(k). at any time index k , initially the control signal u(k 1) is used. however, after the optimization, u(k) is available and could be used in next iterations. 6. takagi-sugeno fuzzy model validation 6.1 correlation tests traditionally more rigorous statistical validation tests are employed in which model residuals are examined, and if found to be sufficiently correlated with a function of the data then the model is inadequate. this is achieved by defining a matrix z, where z(xt) is (31) in which xt is the observational vector of inputs, outputs and errors seen up to time step t, and m(t) represents the degree of dependency of the two training signals y(t) and u(t), i.e. (32) and m(t 1) is a monomial of the vector xt given by (33) the following two hypotheses are defined : where the purpose of validation is to use the data to decide if h0 holds. two different test statistics have been proposed in the literature, the most common being the standard sample correlation measure, ρ(k), is defined as, (34) (35) where ho hold is asymptotically a x2(s) distribution where s is the number of delays, td. for a given acceptance level (typically 95%) a critical point is found, and if elements of d are outside this acceptance region, ho is rejected. 7. modeling of a nonlinear system : a case study 7.1 antenna system (input-output training pattern) to test these proposed neuro-fuzzy methodologies further, they are applied to model a realistic nonlinear dynamical system. the system considered is a nonlinear (mimo) dynamics of an antenna system with two coupled inputs 18 the journal of engineering research vol. 2, no. 1 (2005) 12-24 ( ) ( ),1max ,10,1max 11 1 max,1 ∑ −= +∑ −+= = = oin j yj in i di n u α αα b is an in×α and c is a α×on matrix: ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = moom k moom k kk mm kk k inononon in in ,2,1, ,22,21,2 ,12,11,1 00 00 ηηη ηηη ηηη b ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = 10000000 00001000 00000100 0 0 000010 0000001 1,,,1,2,1, 1,2,,21,22,21,2 1,1,1,11,13,12,11,1 kkk kkk mmommmommmm kkk kkkkk mmommoom kkkkk mmommook kom kk kkkk αα α α θηηζζζ θηηζζζ θηηζζζζ onjonionononon jonio ji a . 0010 001 ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ = kkk moom kkkk c ( ) ( ) ( ) ( )[ ]tdt ttmtmtmxz −−= ,...,1, ( )ttttt eyux 111 ,, −−−= ( ) ( ) ( )21 2 −−= tutytm oh : ( )te is uncorrelated with ( ) ( ) ()( ) 0., =zteexz t and h 1 : ( )te is correlated with ( ) ( ) ()( ) 0., ≠zteexz t ( ) ( ) ( )[ ] ( ) ( ) ][ ( ) ( )[ ]∑ +−=∑ +−= ∑ +−= +−+− +− = 1 1 1 1 1 1 11 11 kn t kn t kn t tetektmktm tektmn kρ in which k =1,…,td and ( ) [ ].1,1 +−∈kρ if h o holds, this statistic asymptotically approaches a normal distribution, and with 95% confidence limits , h o is accepted if ( ) ]/96.1,/96.1[ nnk −∈ρ . an alternative stati stic is given by: ( )[ ][ ] ( ) ( )[ ] ( ) ( )[ ][ ] ( ) ( )[ ]( )[ ]∑ −=−∑ −= ∑ −= − ∑ −= ⎥⎦ ⎤ ⎢⎣ ⎡= dtntd tn t dtntdtnt tetztztz tetztend 1 1 1 1 1 1 2 and outputs. a data set containing 500 samples of training patterns were produced by applying random torques to the different channels, with suitable sampling rate and an amplitude drawn from uniform random distribution in the range (-1.5 , + 1.5) n/m. antenna system a coupled two degree of freedom satellite dish, typical of the type used for oceanary satellite communication systems, is presented. the behavior of the antenna is described by the following nonlinear idealized time invariant state space equations: (36) (37) and (38) where ϕ is the azimuth angle, ψ is the elevation angle, bϕ and bψ are the associated friction coefficients, and tϕ and tψ are the torques applied to the axes. to produce a more realistic simulation, the outputs are corrupted by additive gaussian noise, [eϕ(t) eψ(t)]t, representing a crude approximation to measurement noise. the azimuth is permitted to turn through a complete revolution, while end stops restrict the elevation to the interval [0,π]. in this antenna there are essentially two sources of nonlinearity: that produced from the end stops on the elevation and the other as a results of the non-isotropic moment of inertia tensor. indeed when isotropy is present the state-space equations (above) are linear. the strength of this non-linearity depends on the degree of anti-isotropy and the angular velocities of the antenna. these torques are chosen to emulate typical operating conditions. such block diagram used to produce the identification data was simulated through simulink/matlab, using a set of nonlinear differential equations that describe the antenna system. half of the training pattern was used in the modeling of the dynamic system whereas the other half was used to validate the fuzzy models resulting from the modeling. for a typical sequence of training data, such responses of the antenna inputs-outputs is shown in fig. 3. 7.2 training pattern and clustering 7.3 neuro-fuzzy modeling neuro-fuzzy modeling is applied to the problem of identifying a discrete model of the antenna. a fuzzy model can be constructed from data by using the output of the clustering algorithm and by constructing regressors to form inputs to the neuro-fuzzy network. hence a conventional linear difference model with regressors is constructed containing previous inputs and outputs, i.e. (39) (40) fuzzy if-then rules can be extracted by projecting the clusters onto the axes and the membership functions of the fuzzy sets generated by pointwise projection of the partition matrix onto the antecedent variables. then consequent parameters for each rule are obtained as least squares estimates. when an initial structure is obtained through clustering, the membership functions and the consequent parameters are tuned to satisfy certain cost function through the learning procedure of the neuro-fuzzy. 7.4 membership functions and associated fuzzy rules as a result, the membership functions of all the inputs (regressors) and outputs are shown in fig. 5 for azimuth angle. the antenna system has six inputs (in terms of fuzzy model) and two outputs, hence two groups of seven sets of mfs are shown. each universe of discourse (set) has three 19 the journal of engineering research vol. 2, no. 1 (2005) 12-24 [ ]tψψϕϕ &&=z ( ) ( ) ( ) ( ) ( ) ( ) ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −′−− +′ −′−− = i iibt ii iibt ψψψ ψ ψψ ψψϕϕ ϕ ψψ ϕϕ 2sin cossin 2sin 2 2 1 22 && & &&& & &z ( ) ( )⎥⎦ ⎤ ⎢ ⎣ ⎡ +⎥⎦ ⎤ ⎢⎣ ⎡= te te ψ ϕzy 0100 0001 as discussed in section (3), fuzzy modeling of any dynamical system could be achieved through clustering the training data. in this respect, fig. 3 shows the employed input -output data training pattern. for this simulation example, clustering ha s been applied to th e antenna training pattern. in fig. 4 it is shown the training pattern following appl ying the clustering algorithm, where it illustrates clearly the clusters and their three associated centers is shown in fig 4 . for instant, the fi gure shows the training pattern which has been clustered into three . to reduce the fuzzy rules while preserving the model accuracy, the number of clusters were chosen to be three clusters . the fuzziness parameter m was kept at 2.2 with a termination cr iterion ∈ =0.01. the result of the clustering algorithm is the fuzzy partition matrix and the cluster centers matrix, which will be used to construct the fuzzy model for the antenna system. ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) t kkk kk ktktkt k ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −∆−− −− −−− = 1,2,1 ,2,1 ,1,2,1 ϕψψ ϕϕϕ ψϕϕ ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) t kkk kk ktktkt k ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −∆−− −− −−− = 1,2,1 ,2,1 ,2,1,1 ψψψ ϕϕψ ψψϕ 20 the journal of engineering research vol. 2, no. 1 (2005) 12-24 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 -2 0 2 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 -2 0 2 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 -5 0 5 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 -5 0 5 tim e (s ) input azimuth torque (nm) input elevation torque (nm) output elevation torque (rad) output azimuth torque (rad) figure 3. input-output data training pattern -1.5 -1 -0.5 0 0.5 1 1.5 -3 -2 -1 0 1 2 3 4 center 1 center 2 center 3 torque 1 rad figure 4. extracted clusters functions associated azimuth angle figure 5. extracted membership functions of all inputs mfs representing the assigned three clusters. such memberships are representing the range of the inputs. 7.5 fuzzy sub-models: the t-s fuzzy model presented has been used to identify the nonlinear antenna system. as was mentioned before, the number of rules in the t-s fuzzy model equals the number of clusters in the product space. the consequent of each rule is a local model that approximates the output of the real function for the range of x for which the rule is applicable. as a result of the modeling development, the following rules are obtained for azimuth and elevation angles : here the c and d matrices are common for all of the three fuzzy sub-models, and the d matrix is equal to zero. furthermore, the elevation angle dynamics is of the same above structure. the antenna simulation system incorporating the three models are shown in fig. 6. consequently, fig. 6 shows the actual antenna output superimposed over the evaluated fuzzy model output. from the figure, it is apparent how the fuzzy model output resembles the actual system output. 7.6 fuzzy sub-models validation figure 7 displays the cross-correlation function of the error signal with the first input signal of the antenna. the correlation in the figure is within the confidence interval, which indicates that the two signals are not correlated. to further investigate the constructed local linear sub-models of the antenna, figure 8 shows the attained linearized sub-models over the antenna time response. in terms of antenna nonlinear behavior, it is obvious that the entire operating region has been sub-divided into a number of local models which could be employed for further control synthesis. from the shown antenna response, fuzzy models are useful for describing the antenna dynamics where the underlying physical mechanisms are not completely known and the antenna behavior is understood in qualitative terms. consequently, an important property of fuzzy models is their capability to represent nonlinear dynamic systems. therefore, the obtained fuzzy sub-models can also be applied to systems that are well understood but due to the nonlinearities untraceable with standard linear methods. rule-based structure of fuzzy models allows for integrating heuristic knowledge with information obtained 21 the journal of engineering research vol. 2, no. 1 (2005) 12-24 rule 1: if ( )1−ktϕ is 11f and ( )2−ktϕ is 12f and ( )1−ktψ is 13f and ( )1−kϕ is 14f and ( )2−kϕ is 15f ( )1−kψ is 16f and ( )2−kψ is 17f then ( ) ( ) ( )ttt ubxax 11 +=& and ( ) ( )tt xcy 1= . rule 2: if ( )1−ktϕ is 21f and ( )2−ktϕ is 22f and ( )1−ktψ is 23f and ( )1−kϕ is 24f and ( )2−kϕ is 25f ( )1−kψ is 26f and ( )2−kψ is 27f then ( ) ( ) ( )ttt ubxax 22 +=& and ( ) ( )tt xcy 2= . rule 3: if ( )1−ktϕ is 31f and ( )2−ktϕ is 32f and ( )1−ktψ is 33f and ( )1−kϕ is 34f and ( )2−kϕ is 35f ( )1−kψ is 36f and ( )2−kψ is 37f then ( ) ( ) ( )ttt ubxax 33 +=& and ( ) ( )tt xcy 3= ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −− −− = 000000 000000 000100 0013.003315.08022.00273.00348.0 000001 00116.02998.00300.04265.009510.0 1a ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −−− −− = 000000 000000 000100 0002.003715.00844.00384.00373.0 000001 00119.00475.00525.04667.09354.0 2a ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ −− −− = 000000 000000 000100 0002.002987.07887.01098.01294.0 000001 00108.00157.00142.04099.09020.0 3a ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = 10 01 00 0240.00 00 00128.0 1b , ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = 10 01 00 0137.00 00 00128.0 2b , ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = 10 01 00 0028.00 00 00118.0 3b ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ === 000100 000001 321 ccc ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ === 00 00 321 ddd where, 22 the journal of engineering research vol. 2, no. 1 (2005) 12-24 0 100 200 300 400 500 600 -4 -2 0 2 4 6 time (s) 0 100 200 300 400 500 600-4 -2 0 2 4 time (s) azimuth angle (rad) elevation angle (rda) figure 6. fuzzy model responses (azimuth and elevation angels) compared to the antenna outputs -500 -400 -300 -200 -100 0 100 200 300 400 500 -3 -2 -1 0 1 2 lag cross correlation of torque 1 with error -500 -400 -300 -200 -100 0 100 200 300 400 500-1 0 1 2 3 4 lag cross correlation of torque 2 with its associated error cross correlation of torque 1 with error figure 7. cross correlation of system first input with its associate error of the antenna system from antenna measurements. the global operation of the antenna nonlinear process is divided into several local operating regions. within each region ri, a reduced order linear model in armax form, is used to represent the local antenna behavior. that was not restrictive, and any appropriate model forms can also be used. 8. conclusions this article has concentrated on the modeling of nonlinear dynamic systems via the utilization of the well known fuzzy modeling paradigm, the takagi-sugeno (t-s) technique. t-s models depend heavily on some initial membership centers of the universe of discourse of used fuzzy variables, such centers have been obtained by employing clustering algorithm. once such centers are computed, a fuzzy system can establish initial membership centers through which they are updated via a neural network learning mechanism. one of the advantages of t-s modeling is that systems can be modeled by few rules, and consequently fewer linear sub-models. this advantage has overcome the problem of the large number of rules in the fuzzy modeling. fuzzy models have also been verified and validated through some standard validation techniques, where they have shown clearly the successful ability of t-s techniques to model nonlinear systems with a good degree of accuracy. references akkizidis, s. and roberts, n., 2001, "fuzzy clustering methods for identifying and modeling of non-linear control strategies": proceedings of the institution of mechanical engineers. part i: journal of systems and control engineering, vol. l(215), pp. 437-452. babuska, r. and verburggen, h., 1995, "identification of composite linear models via fuzzy clustering," proceedings of the european control conference, italy, pp. 1207-1212. bologna, g., 2001, "a new neuro-fuzzy model," proceedings of the international joint conference on neural networks (ijcnn'01), vol. 2, washington-dc (usa), pp. 1328-1333. bossley, k., 1997, "neuro-fuzzy modeling approaches in system identification," ph.d. thesis, university of southampton, u.k. gorzalczany, e., marian, b. and gluszek, a., 2000, "neuro-fuzzy networks in time series modeling," proceedings of the international conference on knowledge-based intelligent electronic systems, vol. 1, brighton, (uk), pp. 450-453. ikonen, e., najim, k. and kortela, u., 2000, "neurofuzzy modelling of power plant flue-gas emissions," j. of engineering applications of artificial intelligence, vol. 3(6), pp. 705-717. lo, j. and chen, y., 1999, "stability issues on takagisugeno fuzzy model, parametric approach", ieee transaction of fuzzy systems, vol. 7(5), pp. 597-607. mamdani, e. and assilian, s., 1975, "an experiment in linguistic synthesis with a fuzzy logic controller," international j. of man-machine studies, vol. 3(5), pp. 1-13. min-you, chen. and linkens, a., 1998, "fast fuzzy modeling approach using clustering neural networks," ieee international conference on fuzzy systems, vol. l(2), ieee world congress on computational intelligence, pp. 1088-1093. 23 the journal of engineering research vol. 2, no. 1 (2005) 12-24 figure 8. obtained linearized antenna sub-models ning, l., shaoyuan, l. and yugeng, x., 2001, "modeling ph neutralization processes using fuzzy satisfactory clustering," ieee international conference on fuzzy systems, vol. l(1), pp. 308-311. takagi, t. and sugeno, m., 1985, "fuzzy identification of systems and its application to modeling and control," ieee transaction systems, man and cybernetics, vol. 15(1), pp. 116-132. wang, h., li, j., niemann, d. and tanaka, k., 2000, "ts fuzzy model with linear rule consequence and pdc controller: a universal framework for nonlinear control systems," proceedings of the 9th ieee international conference on fuzzy systems. zhang, j. and knoll, a., 1999, "modelling multivariate data by neuro-fuzzy systems", proceedings of the 1999 ieee/iae conference on computational intelligence for financial engineering (cifer), new york, (usa), pp. 267-270. 24 the journal of engineering research vol. 2, no. 1 (2005) 12-24 the journal of engineering research 11 (2005) 00-00 the journal of engineering research 11 (2005) 00-00 the journal of engineering research 11 (2005) 00-00 template for electronic submission to acs journals the journal of engineering research (tjer), vol. 19, no. 2, (2022) 106-128 0043 corresponding author's e-mail: afzal19@squ.edu.om doi:10.53540/tjer.vol19iss2pp106-128 design and mixing performance of passive micromixers: a critical review afzal husain1*, asharul islam khan1, wasim raza2, nabeel al-rawahi1, nasser al-azri1, and abdus samad2 1mechanical and industrial engineering department, sultan qaboos university, muscat, oman 2department of ocean engineering, indian institute of technology madras, tamil nadu, india. abstract: this study extracts and reports notable findings on passive micromixers by conducting an exhaustive review of designs, their features, and mixing performance. the study has covered the relevant articles on passive micromixers published from 2010 to 2020. the analysis of filtered and selected articles sums up passive micromixers into four categories: designed inlets, designed mixing-channel, lamination-based, and flow obstacles-based. the prominent mixing channel categories identified in the study are split-and-recombine (sar), convergent-divergent (c-d), and mixed (sar, c-d, and others). moreover, differences in mixing channel designs, number of inlets, and evaluation methods have been used in comparing the mixing performance of passive micromixers. the sar and the obstacles-based micromixers were found to outperform the others. the designs covered in the present review show significant improvements in the mixing index. however, these studies were conducted in an isolated environment, and most of the time, their fabrication and device integration issues were ignored. the assortment and critical analysis of micromixers based on their design features and flow parameters will be helpful to researchers interested in designing new passive micromixers for microfluidic applications. keywords: microfluidics, passive micromixers; mixing index; split-and-recombine; convergent-divergent; lamination; off-set inlets. : مراجعة نقديةغير النشطةالدقيقة للخالطات األداء تصميم وفعالية عبد الصمدوناصر العزري و نبيل الرواحي و وسيم رضاو أشار اإلسالم خانو أفضال حسين تصاميم :الملخص عن شاملة مراجعة بعد السلبية الدقيقة الخالطات عن البيانات أهم ونشر استخالص على الدراسة هذه تقوم م. 0202و 2010الدقيقة السلبية بين العامين الخالطات المنشور عن ومميزات وأداء هذه الخالطات حيث استهدفت هذه الدراسة تصميم المداخل وتصميم أن الخالطات الدقيقة السلبية يمكن تقسيمها بناء على أربعة محاور: نجدوتحليلها دراسات وبعد فرز هذه ال وخلصت هذه الدراسة إلى أن أبرز الفئات المبنية على قناة الخلط هي: الفصل وإعادة .وعوائق الدفققناة الخلط وخاصية التغليف والمتقاربة والمتعددة األنواع. بمتال-الجمع، على اإلضافة إلى ذلك باعدة القطر فإن مقارنة أداء الخلط في هذه الدراسة تم بناؤها وطرق التقييم. تبين االختالف وعدد المداخل على كذلكفي تصميم قنوات الخلط وتلك المبنية الجمع وإعادة أن خالطا الفصل يم التي شملتها هذه المراجعة قد أثبتت تطورات التصامكما أن .الدقيقة السلبية األخرى الخالطات عن أدائها عوائق الدفق تتفوق في إال أن ما يجدر ذكره أن هذه الدراسات قد أجريت في بيئات معزولة كما يغلب عليها إهمال جانب طبيعة .هامة في مؤشرات الخلط الدقي الخالطات وتحليل تصنيف إن األخرى. األجهزة مع المتعلقة بتكامليتها والقضايا الخالطات هذه مزايا تصنيع على بناء قة تصميمها ومعامالت الدفق فيها سيساعد الباحثين المهتمين بتصميم خالطات دقيقة جديدة في تطبيقات الموائع الدقيقة. الخلط ؛الخالطات الدقيقة السلبية ؛الموائع الدقيقة الكلمات المفتاحية: الجمع ؛ مؤشر وإعادة والتباعد ؛الفصل ؛ التغليف ؛التقارب . المداخل المزاحة mailto:afzal19@squ.edu.om design and mixing performance of passive micromixers: a critical review 107 1. introduction a passive micromixer is an important component of lab-on-chip devices. they are one of the most frequently used add-ons in microfluidics for rapid mixing. micromixers are getting extensive attention in numerous fields of science and engineering disciplines such as biochemical analysis, drug discovery, sensing, microelectromechanical systems (mems), micro total analysis systems (μ-tas), bio-diagnostics, lab-on-achip (johnson and locascio, 2002), and chemical processing. it can be used standalone as well as an integrated mixing unit in microfluidic systems. micromixers are used to mix two or more fluids rapidly within a short channel length and short time. the advantages associated with the micro-dimension of channels, such as reduced sample volume and reagent requirement along with the lesser amount of wasteproduct produced, lead to low operating cost, high surface area to volume ratio providing better heat dissipation, faster analysis results, portability, and high-throughput through parallelization. these advantages have enhanced their prospects of usage in a wide range of applications such as dna purification (kastania et al., 2016), protein folding (jang et al., 2019), enzyme catalysis (zhang et al., 2021), nanoparticle synthesis (westerhausen et al., 2016), and water quality monitoring. an exhaustive review of the applications of micromixers can be found elsewhere (jaywant and arif, 2019; jeong et al., 2010; lee and fu, 2018). the micromixer design, passive structures, and the flow reynolds number (re) significantly affect the mixing performance and need further research to meet the contemporary developments in the microsystems. the reynolds number indicates the degree of dominance of inertial force over viscous force. turbulent flow, characterized by a high reynolds number (re > 2300), has greater inertial force causing flow instabilities in the form of vortices and eddies, which can promote mixing in the channel. however, the flow reynolds number in a micromixer (re < 100) is much lower than the limiting reynolds number for a turbulent flow (re > 2300). hence, the adequate mixing of two fluids is difficult to achieve in the absence of flow instability under laminar flow conditions inside the micromixers (re<<100) as compared to the turbulent flow (bhopte et al., 2010). mixing by diffusion is a slow process, which has little relevance to most fluid mixing applications, where fast mixing is a prerequisite. thus, microfluidic mixing is a challenging problem for low diffusivity fluids under laminar flow conditions and hence requires design improvements to enhance mixing. the micromixers are of the types: active, passive, and hybrid. in addition, each of them has many subclassifications based on the channel designs and mixing mechanism. the mixing in active micromixers needs external forces like the electric field, acoustic field, magnetic field, and thermal field to disturb the flow. however, mixing in passive micromixers occurs due to molecular diffusion and chaotic advection strengthened through channel structural modifications. passive micromixers are low-cost, reliable, robust, efficient, and easy to use. passive micromixers have been used in mixing both miscible and immiscible fluids. the hybrid micromixers use important characteristics of both active and passive micromixers to enhance the mixing capabilities (bazaz et al., 2018). the passive micromixers are of yand t-shape depending on the inlet design. they also differ in terms of mixing-channel configurations, such as obstacles in the main channel or attached to the walls, and designed mixing-channel geometry profiles (e.g., zigzag, splitand-recombine (sar), and a teardrop). there is a proportional relationship between flow path length and mixing performance based on the localized mixing and global mixing principles (su et al., 2019). the mixing efficiency/performance of the passive micromixer is largely dependent on the geometric layouts and reynolds number. thus, there is a striking possibility to achieve different levels of mixing by tweaking and tuning the geometric parameters. for example, in the case of imitate cantor structure passive micromixers, reducing space between adjacent fractal obstacles in the mixing channel and increasing the mixing chamber's height improve the mixing performance (with optimum mixing performance at 600 µm height) (wu and chen, 2019a). in chaotic advection-type passive micromixers having dual opposing structures on microchannel walls, the alignment of structure strips affects the mixing performance (chen et al., 2015). the mixing performance is enhanced by interdigital inlets (cook et al., 2011). additionally, the geometric-parameters tolerances associated with any fabrication process slightly influence the performance of the passive micromixers. further, in a study, the authors calculated the cumulative distribution function for three cases to obtain the probability of having mixing efficiency lower than 95% for a tesla micromixer (stanciu, 2015). for instance, a small change (up to 5%) in geometric parameters with six and seven elements yields respective probabilities of 11.1% and 2.08% for getting mixing performance lower than 95% (stanciu, 2015). although there has been a large array of improvements achieved in the last few years, intense research for more efficient passive micromixers for microfluidic systems is continued to date. most of the research has focused on the design of new passive micromixers and the estimation of their mixing performance through experiments and numerical simulations. furthermore, several studies have reviewed the literature on micromixers, such as a comparative review on passive micromixers (raza et al., 2020), a review on droplets based micromixers (chen et al., 2019a), a review on active and passive micromixers (bayareh et al., 2019), a short review on micromixers classification (mukhopadhyay, 2018), afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 108 and a review on micromixers (gaozhe et al., 2017). these studies are mainly focused on passive micromixers, their types, and to some extent, fabrication techniques, with some emphasis on micromixers having both passive and active mixing features. there is a significant void in the literature to get a comprehensive spectrum of passive micromixers, their capabilities, applications, and feasibility of integration. nevertheless, this review investigates the dominant microchannel structures adopted for mixing enhancement in the last decade. additionally, it sets the base to propose favorable channel structures to be further explored in the future. the study conducts an extensive analysis of the published findings on passive micromixers from the year 2010 to 2020. there are six sections in the article. section 1 briefly introduces the concept and challenges of passive micromixing, section 2 illustrates the research method, section 3 corresponds to the background study, section 4 aggregates the literature from 2010 to 2020 and classifies passive micromixers, section 5 presents the analysis and discussion, and the section 6, the last section, summarizes the findings and discusses future research. the study has followed a systematic literature review approach. several protocols were set, including the databases to search for articles, inclusion and exclusion criteria for articles, and extraction of relevant information from the filtered articles. this approach helps in getting articles objectively rather than the exploration of databases randomly. the systematic literature review has become a common method of conducting a literature review in different fields of science and engineering. the online databases referred to for collecting the articles include science direct, google scholar, springer, and ieee xplore. articles published only in english and published in the years between 2010 and 2020 were considered. 2. background the fluid flow in microchannels is mostly laminar when the channel is of small dimensions, and fluid velocity is low (bayareh et al., 2019). the efficiency and sensitivity of microfluidic devices are influenced by the passive structures in the micromixers and their designs (gaozhe et al., 2017). the micromixers found in the literature can be broadly categorized into four principles such as lamination-based, injection-based, droplets-based, and chaotic advection-based (figure 1). the lamination-based micromixers can be further categorized into parallel and serial lamination-based passive micromixers. in addition, they are subclassified into; channels with obstacles (or baffles), curved-channel, convergence-divergence, and unsymmetrical (bayareh et al., 2019). the laminationbased micromixers have a layered structure, and they achieve excellent mixing in a short time (milliseconds) (gaozhe et al., 2017). the micromixers based on injection have nozzles through which fluid passes in the form of many sub-streams into the mainstream. the droplet-based passive micromixers inject droplets for mixing (yang et al., 2018) into the main channel. the micromixers based on chaotic advection work on the principle of stretching, folding, and breaking of fluid interfaces for the laminar flow. mixing in passive micromixers is mainly due to the phenomena of molecular diffusion and chaotic advection. the chaotic advection is favorable when the fluids have large size molecules and small diffusion coefficients. this study classifies passive micromixers based on inlet designs (designed inlets), mixing-channel layout/design, lamination-based, and obstacle-based (figure 2). 2.1 designed inlets the mixing performance of passive micromixers varies with the type of inlets. the designed inlets include (1) an angle between the two inlets (t-, y-, and configuration with other angles), (2) swirl-induced inlet, and (3) numbers of the inlet (e.g. 2, 3, 4, and 5). the mixing performance of passive micromixers increases with the angle between inlets. the inlet angle has been varied from 60° to 300°; however, the commonly designed inlets are tand y-shape (figure 3). figure 3 shows the designed inlets and their arrangements at different angles. figure 1. working principles of passive micromixers. figure 2. classification of micromixers based on geometric features. figure 3. micromixers based on inlet channels. micromixer principles lamination parallel lamination serial lamination injection droplets chaotic advection passive passive micromixers designed inlets designed mixingchannels laminationbased flow obstacles design and mixing performance of passive micromixers: a critical review 109 figure 4. micromixer with swirl inlets. it was suggested that the optimal inlet mixing angles are 90°-180° for high mixing and low energy dissipation (zhang et al., 2016). the maximum and minimum mixing performances were obtained with an intersection of 300° between the inlets at re values of 60 and 80 (zadeh and marahel, 2011), respectively. in addition to various configurations of the inlets, the swirl inlets (figure 4) have shown a considerable impact on the mixing performance. for example, the t-shape micromixers with swirl inlets showed a mixing enhancement by a factor of more than two compared to a t-mixer at re = 20–150. additionally, a t-shaped micromixer with swirl inlets achieved a mixing index of 40 times that of the simple tmicromixer at re = 100. in the three-dimensional serpentine micromixers, the impact of the antisymmetric inlets was significant at the early stage; mixing index enhancement of 2–10 times was observed at re = 100 (matsunaga and nishino, 2014). the swirl in the inlets improves the mixing performance at intermediate re values (66 < re < 180). the reverse orientation swirl (r-o-swirl) inlets micromixers performed better than the same orientation swirl (s-o-swirl) inlets micromixers. dundi et al. (2019) found that the mixing performance of a micromixer with r-o-swirl inlets increases 300% to 400% at 160 < re < 180 and with s-o-swirl inlets increases 30% to 70% at 266 < re < 372, as compared to a t-shape passive micromixer without swirl. although the simplest tand y-shape passive micromixers have two inlets, there are passive micromixers with more than two inlets. in y-shape passive micromixers with three inlets, the pressure drop is uniform, and fluid flows with constant velocity, resulting in fast diffusion of fluids (adam and hashim, 2012). the four-inlet passive micromixers can have 50% less pressure drop and improved mixing performance of 300% to 500% as compared to the normal passive micromixers (bhopte et al., 2010). the swirl-induced two-inlets passive micromixers have higher numerical diffusion than the swirl-induced fourinlet passive micromixers at 120 < re < 240 (okuducu and aral, 2019a). in general, swirl inlets and reverse-oriented swirl inlets have shown significant improvement especially advection-driven mixing at higher reynolds numbers. these inlet designs can be looked at for integrating with other 3d designs of the main channel to achieve higher mixing in a shorter length. 2.2 mixing-channel layout/design-based the channel layout plays an important role in enhancing the mixing performance. passive micromixers with hydrophobic microchannels have better performance than hydrophilic microchannels. the spiral microchannels produce higher mixing performance as compared to straight microchannels (zhang et al., 2016). the prominent two-dimensional (2d) and three-dimensional (3d) microchannel layouts are spiral and wave (e.g., single spiral, double spiral, helical, sinusoidal, rectangular, square, and zigzag), sar, and convergent-divergent (cd), and mixed (sar, cd, and others). 2.2.1 wave/spiral shaped microchannel this section dealt with the wave and spiral-shaped mixing-channel designs proposed in the last decades. chen et al. (2016a) reported the shape optimization of a micromixer with six different channel layouts. the authors found that the square wave microchannel produced the highest mixing performance among the tested six layouts. the order of decreasing mixing performance was square wave > multi-wave > zigzag > t > mouth > loop micromixers (chen et al., 2016a). scherr et al. (scherr et al., 2012) proposed a micromixer based on logarithmic spirals and measured the mixing performance at the reynolds number varying between 1 and 70. they noticed a decline in mixing as the re values increased from 1.0 to 15 with a transition at re = 15 (mixing efficiency = 53%); however, mixing increased with a further increase in re and reached 86% at re = 67. the new design showed higher mixing as compared to the archimedes spiral and meandering-s mixers. papadopoulos et al. (2014) investigated the effects of inlet velocity in a zigzag microchannel geometry with a t-inlet. it was reported that the number of zigzags affects the mixing at a fixed inlet velocity. a higher velocity requires a larger number of zigzags for efficient mixing. a total of 150 zigzag units were recommended to achieve efficient mixing performance (98%) at a velocity of 2 mm/s and at a diffusion coefficient of 10-10 m2/s. wang et al. (2018) designed a modified s-shaped microchannel having asymmetric lateral wall structures and conducted numerical simulations for both the modified microchannel and simple s-shaped microchannel at re values ranging from 0.1 to 100. they found that as re values increase from 0.1 to 100, modified s-shaped microchannel exhibits high mixing performance than simple s-shaped. it is due to the formation of secondary flow and higher inertial effects. rafeie et al. (2017) proposed a 3d fine-threaded lemniscate-shaped micromixer to overcome the limitations of other 3d micromixers. they reported a mixing performance of over 90% at reynolds numbers ranging from 1.0 to 1000, both numerically and experimentally. they observed that the chaotic advection (due to dean flow) and diffusive mixing (due afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 110 to grooves) are the main cause of high mixing. zhang et al. (2012) proposed a 3d-twisted compression–expansion microchannel for passive micromixers. they applied simulations to compare the new design with a t-shape having the same crosssectional perimeters. the new design was reported to have higher mixing efficiency compared to t-mixer. moreover, clamping angles at 90 degrees, 45−90 degrees, and 45 degrees have different efficiency outputs, with maximum efficiency reported at 45 degrees causing the least pressure drop. yang et al. (2013) proposed a micromixer with overlapping dual spirals. the micromixer showed enhanced performance due to sustained dean vortices. there are several studies on mixing-channel design with a helical structure and its variations. the helical structure can be single or double. liu et al. (2015) studied 3d crosslinked double-helical microchannel for fast mixing at low-re value. they compared the new channel with a helical and a straight channel of the same dimensions. the numerical simulation revealed higher mixing of the 3d cross-linked double-helical microchannel with full mixing within three cycles at re = 0.0015~600. the helical microchannel in 3d micromixers introduces chaotic advection, and the double helix has an agitation effect. viktorov and nimafar (2013) analyzed numerical and experimentally mixing performance and pressure drop in two micromixers (chain one and chain 2) consisting of 3d sar with a special arrangement of microstructures called 'chain mixer'. they found that the chain and teardrop micromixer, due to the sar mechanism, have higher mixing performance (reached up to 98%) compared to a simple t-shape and oshaped mixer for reynolds numbers ranging from 0.083 to 4.166. furthermore, chain mixers showed lower pressure drop as compared to teardrop mixers. al-halhouli et al. (2015) designed ilsc and ω channels to enhance mixing efficiency by introducing the dean vortices and molecular diffusion. they compared the performance of new designs with a spiral design using numerical simulation and experiment. the ilsc and ω configurations have complete mixing when the reynolds number lies between 0.01 and 50, and on the other hand, the complete mixing in the semicircle shapes arranged in a spiral occurred at re > 50. afzal and kim (2015b) numerically carried out multi-objective optimization of the sigma micromixer using a multi-objective genetic algorithm. the author observed that the geometric parameters could be effectively fixed to maximize the mixer performance. the square-wave channel exhibited significant improvement over other wavy channels and is simpler in fabrication. it is found that in spiral micromixers, repetitive alteration of the flow direction has improved the mixing in a wide range of reynolds numbers due to changes in dean vortices' rotational direction. furthermore, a design with a smaller radius of curvature is better for mixing. 2.2.2 sar-based microchannel the microchannel based on the sar principle has higher mixing performance as compared to non–sar (ansari and kim, 2010). the sar-based microchannel can be with a chamber and without a chamber. the chambered sar design has several variations as circular, elliptical, and rhombus. the splits in the chambered sar can be two or more. moreover, gidde et al. (2019b) numerically investigated the impact of balanced and unbalanced splits on the mixing performance and pressure drop in the circular split and recombine (csar) and elliptical split-and–recombine (esar) (cortes-quiroz et al.) micromixers at reynolds numbers ranging from 0.1 to 75. they observed that at 0.1 ≤ re ≤ 5, mixing is due to diffusion, while for re ˃ 5, mixing is caused by the secondary flow, separation vortices, and sar effect. the esar micromixer (with balanced and unbalanced split) has higher mixing performance and less pressure drop than the csar micromixer over the entire range of re and showed good mixing performance at reynolds numbers ranging from 20 to 75. xia et al. (2011) investigated experimentally and numerically mixing performance in a planar asymmetric split-and-recombine micromixer with a fan-shaped cavity at re = 1–80. they observed higher mixing due to the combination of expansion vortices (caused by the converging-diverging structure of the fan-shaped cavity) and dean vortices, along with unbalanced inertial collisions. the proposed micromixer having a fan-shaped cavity of width equal to three times the width of the major sub-channel, achieved a mixing index of around 75% at re ≥ 60. hossain and kim (2014) designed a passive micromixer with unbalanced three-split rhombic subchannels and evaluated performance against a two-split micromixer at several reynolds numbers ranging from 0.1 to 120. they observed that the mixing performance of the three-split rhombic microchannel is 1.49, 1.67, 1.56, and 1.44 times higher than the corresponding two splits at re = 30, 40, 50, and 60, respectively. further, hossain and kim (2015) investigated the mixing performance of a 3d serpentine split-and-recombine microchannel consisting of a series of "oh"-shaped segments in a reynolds number range of 0.1–120. they observed that continuous split-and-recombine of the fluid streams due to the oand h-structures generates chaotic advection and thereby improves the mixing. the inclusion of sar design with the serpentine structure improved the mixing performance compared to a previously proposed 3d serpentine design in a reynolds number range of 0.1–70. gidde et al. (2019a) numerically investigated the flow features, mixing performance, and pressure drop in a proposed eye-shaped split and collision (es-sac) passive micromixer for re ranging from 0.1 to 45. they reported that mixing performance is enhanced due to the unbalanced collision of the two fluid streams in the sub-channels of unequal widths. among three different ratios of the sub-channel width, i.e., 1, 1.4, design and mixing performance of passive micromixers: a critical review 111 and 2, the least mixing was observed for the width ratio of 1, while the highest mixing was observed for the width ratio of 2. furthermore, the pressure drop in the micromixer with a sub-channel width ratio of 2 was the least among the three different width ratios. nimafar et al. (2012) proposed a microchannel of an h-shape in which the fluid split and repeatedly recombined to maximize the diffusive mixing. the experimental investigation demonstrated that h-micromixer was more efficient than the other two tested devices, i.e., t and o-micromixer, in the reynolds number range of 0.08–4.16. the new h-micromixer indicated a mixing performance of 98% at re = 0.083. viktorov et al. (2015) proposed two modified designs of chain mixers, i.e., the y-y and the h-c microchannel based on the sar principle, and investigated mixing and flow numerically and experimentally for reynolds numbers ranging from 1 to 100. the mixing efficiency of these two passive micromixers showed flat characteristics with a value higher than 90% over the entire range of the reynolds number investigated, while the mixing efficiency of the teardrop mixer showed a decrease in the mid-range of the reynolds number. moreover, the pressure drop was also less in the newly proposed designs as compared to the teardrop micromixer. ansari et al. (2010) evaluated the mixing performance of a sar micromixer having subchannels of unequal widths using numerical simulations and experiments at different reynolds numbers between 10 and 80. the sub-channels of unequal widths caused unequal mass flux splitting and unbalanced collision on the recombination of mixing species. the combined effect of sar, unbalanced collisions of the fluid streams, and dean vortices in the curved channels caused mixing enhancements. when the major sub-channel was twice as wide as the minor sub-channel, maximum mixing performance was achieved at a re value greater than 40. ruijin et al. (2017) proposed a passive micromixer that generates continuous lamination layers using baker transformation for mixing enhancement. they investigated the mixing performance numerically for reynolds numbers ranging from 0.01 to 10 and compared its performance against other splittingmerging micromixers such as helical-mixer and smale-mixer. they found that the baker mixer achieved higher mixer efficiency due to a better stratification effect than those in the helical-mixer and smale-mixer at low reynolds numbers. however, due to the convection–dominant mixing at the higher reynolds numbers, the baker mixer showed lower mixing compared to the other two mixers. furthermore, due to the presence of contracting channels in the baker mixer, the pressure drop through it was higher than that in the other two mixers. pennella et al. (2012) reported planar sar micromixers with variable curvature, i.e., clothoidshaped mixing units. they investigated the mixing performance numerically and experimentally for a reynolds number range of 1 to 110. due to the secondary flows in the repetitive mixing units and recirculation effects at the junctions of the unit, the proposed mixer achieved a mixing efficiency of 80% for the reynolds number greater than 70. yang et al. (2015) proposed a passive micromixer with 3d tesla structures and investigated the mixing performance experimentally and numerically for reynolds numbers ranging from 0.1 to 100. the proposed micromixer showed a maximum mixing efficiency of 94% and a pressure drop of less than 1054 pa at re = 100. le et al. (2014) designed a microchannel that utilizes the impact of stretching-folding in both vertical and horizontal directions through vortices and transverse flow to enhance the performance even for low re values. they evaluated the mixing performance numerically and reported high efficiency (more than 80%) for reynolds numbers ranging from 0.5 to 60. the new design indicated 220-240% higher mixing efficiency than a rhombic mixer with branch channels and pure rhombic at a low-re value. ta et al. (2015) proposed a trapezoidal channel arranged in a zigzag pattern (tzm) for mixing based on the sar principle at low re. the numerical simulation results indicated a mixing efficiency of more than 81% for re values in the range of 0.1 to 80. for re ≤ 0.9 and re ≥ 20, the mixing efficiency was more than 90%. additionally, the mixing efficiency of the proposed design was 4.07 and 5.58 times as compared to three-split rhombic and t-shape micromixers, respectively, at re = 20. the high mixing was due to the transverse flows and distortion of fluid streams. sheu et al. (2012) performed numerical simulations and experiments for the mixing analysis of the proposed micromixers consisting of several staggered three-quarter ring-shaped channels and a semi-circular channel. they found that the staggered curved-channel mixer with a tapered channel achieved 20% higher mixing efficiency with 50% more pressure drop compared to the staggered curved-channel mixer with or without sudden-contracted channels. the dean vortices formed in the curved channels due to uneven split-and-recombine structures and impingement effects enhanced the mixing in the channel. tran-minh et al. (2014) proposed a microchannel based on the sar principle with ellipse-like micro-pillars and determined the performance using numerical simulations. the proposed design attained a mixing efficiency of 80% for re values in the range of 1 to 80. kefala et al. (2014) proposed a sar mixer design with multiple labyrinthine spirals and conducted numerical simulations to compare its performance with the zigzag, spiral, and linear designs. they observed that the labyrinthine microchannel has higher mixing performance (63%) compared to the spiral (36.5%) and zigzag (35.5%) designs at re = 0.5. the microchannel with spiral, zigzag, and labyrinthine designs showed higher mixing by 8%, 11%, and 92%, respectively, compared to the t-shape mixer. chen et al. (2018) modified a previously existing afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 112 f-shaped mixer to propose three designs, namely eshape, stacking e-shape micromixer (sesm), and folding e-shape micromixer (fesm). they evaluated their performance experimentally for re values in the range of 0.5 to 100. the new designs work on the chaotic advection and sar mechanism. the mixing efficiency of the f-mixer and e-mixer started to decrease after reaching a maximum value of 92% and 94%, respectively, at re = 80. however, the mixing efficiency of fesm and sesm remained close to 100%. for 25 ≤ re ≤ 80, all three modified designs performed better than the f-shaped mixer. at re ˃ 80, the mixing performances of two-layer mixers were better than single-layer mixers. the sar, in general, increases mixing performance due to bifurcation and collision. however, its effects are prominent in the advection regime. the unbalanced sar shows improved performance over balanced sar designs. further, the square and elliptical-shaped mixing channels exhibited higher mixing performance than circular channels. the contraction and expansion of dean vortices and distortion of flow streams increase fluid chaos resulting in enhanced mixing. 2.2.3 convergent-divergent microchannels the passive micromixers with microchannel based on the convergent-divergent (c-d) principles have mixed first by the convergence of fluids and then by divergence repeatedly. the convergence of the fluid streams reduces the diffusion path length, which enhances the mixing. furthermore, the fluid layers get stretched on entering the divergent section, which increases the surface area across which diffusion occurs. in addition, due to the sudden expansion of the flow area at high reynolds numbers, expansion vortices are generated that contribute to mixing. gidde et al. (2018) compared the mixing performance of passive planar micromixers having circularand square-shaped mixing chambers numerically, at reynolds numbers ranging from 0.1−75. at re = 0.1, 1, and 5, both designs exhibited similar mixing performance; however, the micromixer with square chambers showed higher mixing performance (99% when re < 1) compared to the circular chambered micromixer. moreover, both designs with a constriction width of 200 µm attained a mixing index over 95% in a reynolds number range of 15–75. afzal and kim (2015a) designed a c-d channel with sinusoidal channel walls. the c-d channel with sinusoidal walls was found to have 19% higher mixing efficiency as compared to straight, square-wave, zigzag, and sinusoidal microchannels with uniform cross-sections for a fixed mixing length. further, the new design achieved 92% mixing performance (independent of re in a range of reynolds number, 0.25 ≤ re ≤ 4) when sinusoidal walls have two periods. the design parameters, i.e., the aspect ratio of the cross-section, the ratio of amplitude to wavelength, and the re values, affect the mixing performance of passive micromixers based on the c-d principle. parsa et al. (2014) proposed a sinusoidal microchannel with c-d and evaluated the mixing performance by applying numerical simulations and experiments. they observed that the increase in the aspect ratio of the microchannel cross-section improves the mixing performance and reduces the pressure drop due to dean vortices formation for re-value ranging from 0.2 to 50. khan and tandon (2017) modified a c-d channel as a wingshaped and serpentine square wave and evaluated mixing performance via numerical simulations for re values from 0.01 to 100. they found that a wingshaped microchannel produced a mixing performance of 87% with a pressure drop of 72 k pa at re = 60. additionally, a wing size of 100-μm gives better mixing efficiency than the 75-μm wings. rampalli et al. (2020) modified serpentine square wave micromixers by replacing square wave with convergent-divergent portion and evaluated the mixing efficiency for low reynolds numbers (re < 100). they found that for re-value less than ten, diffusion mechanism-controlled mixing in square wave convergent-divergent (sqw-cd) mixer, stretching and folding due to c-d caused mixing when re value was greater than 10. the sqw-cd mixer performed better than the conventional square wave mixer when re-value was greater than 10. hong et al. (2019) designed a gourd-shaped (contract and expand) microchannel and performed a numerical simulation to evaluate the mixing performance (re values from 1 to 100). the enhancement of mixing was due to vortex generation in the proposed design, and it achieved a mixing efficiency of 83% with a pressure drop of 4860 pa at re = 100. the aspect ratio, amplitude, and pitch of the converging-diverging units play a significant role in mixing enhancement along with the contraction and expansion of the fluid streams. further, a spiral channel with sinusoidal wave-shaped mixing units, which inherently consists of convergent and divergent portions, shows better mixing due to the synergetic effects of dean vortices, folding, and stretching of fluid interfaces. 2.2.4 hybrid design based on combined structures of sar, c-d, and others there is a number of passive micromixers based on sar, c-d, and other principles combined to achieve higher mixing at different values of re. afzal and kim (2012) introduced a hybrid design using principles of both sar and c-d. the new design has convergent-divergent channel walls with sinusoidal variations, besides splitting the main channels into two and recombining. they measured the impact of the reynolds number (between 10 and 70), sinusoidal wall amplitude, and channel aspect ratio on the mixing efficiency. they observed a symmetric double vortex pair at the throat of the convergentdivergent channel and secondary motions in the subchannels, which enhanced the mixing performance. design and mixing performance of passive micromixers: a critical review 113 numerical simulation results revealed that the new design has a higher efficiency compared to the t-mixer and unbalanced split and collision micromixer. gidde and pawar (2020) studied numerical flow features and mixing performance of rectangular bafflebased triple split-and-recombine (rb-tsar) and elliptical baffle-based triple split-and-recombine (ebtsar) micromixers. baffles of rectangular and elliptical shapes were placed in diffuser-shaped mixing chambers to generate flow in the transverse direction through the sar mechanism. baffles increase the contact area between the layers of mixing species by deflection and folding, thereby contributing to the mixing. the micromixer with elliptical baffles showed better mixing and less pressure drop compared to the micromixer with rectangular baffles due to strengthened secondary flow, separation vortices, and less flow resistance caused by the curved shape. furthermore, the increase in the divergence angle of the diffuser caused mixing to improve and a reduction in pressure drop. fan and hassan (2010) studied experimentally, and numerically a scaled-up micromixer having units called cross and omega at 1 ≤ re ≤ 50. in addition to the dean vortices at high reynolds numbers, split-andrecombine in cross units and focusing/diverging in the omega units were the main mixing mechanisms. the proposed design with five cells achieved a mixing efficiency of 70% at re = 50, which indicates that this design can be used at high flow rates. xie et al. (2010) introduced a new design of micromixers with eight semicircles in the microchannel and measured mixing performance both numerically and experimentally for re values in the range of 0.08 to 40. they observed higher mixing performance (i.e., 90%, for re < 0.1 and re > 10) compared to a simple t-shape mixer. moghimi and jalali (2020) proposed a micromixer in which mixing occurred through multi-lamination and flow-resistance reduction caused due to combined usage of injection and recombination in a zigzag layout. experimental and numerical results indicated that the proposed mixer attained a mixing efficiency of 98.02% in a short length of 1858 microns. cheri et al. (2013) proposed eight planar micromixer designs using four different shapes of obstacles (straight (s), arc (a), chevron (ch), and check mark (cm)) in two different chambers (round corner rectangular (rcr) and hexagon (h)). using numerical simulations for the mixing analysis in the reynolds number range of 0.1– 40, it was found that the rcr-s mixer showed optimum performance among the eight mixers with a maximum value of mixing efficiency to pressure drop ratio. the experimental data verified the numerical results that rcr-s micromixer achieved a mixing efficiency of 0.89 and 0.99 in a length of 1.18 mm at re = 0.1 and 40, respectively. the rcr and h chambers with different obstacles showed minimum pressure drop and maximum efficiency, respectively. the sinusoidal form of sub-channels, which inherently have contraction and expansion passages, exhibited significant mixing improvement. the divergence angle plays a critical role in a diverging mixing chamber. further, a round-cornered rectangular mixing chamber is preferred over a sharp-cornered rectangle. the baffles in a mixing chamber increase fluid chaos leading to higher mixing. 2.3 lamination-based designs the lamination-based passive micromixers implement parallel (sadabadi et al., 2013) and sequential/serial (lim et al., 2011) lamination techniques to enhance mixing performance. in parallel lamination, each mixing fluid stream is divided into two or more streams which are joined together to form a single stream comprising alternate layers of each mixing fluid side by side. while in sequential/serial lamination mixer designs, the fluid stream (from the inlet) is split and recombined serially (taheri et al., 2019). here two or more sub-streams emerge from the inlet stream, and at the recombining point, they are joined again as a single stream. as the number of sub-streams increases, the mixing performances are enhanced. lim et al. (2011) compared the mixing performance of a 3d (cmm) with a conventional passive micromixer applying numerical simulations and experiments (lim et al. 2011). they modified cmm by adding six layers of horizontally and vertically crossing manifold micromixer (h/v-cmm) and achieved a mixing efficiency of over 90% in a channel of length (250 mm) less than five times the channel width. the split-andrecombine and the momentum of the fluids were the main mixing mechanisms in theses designs. moreover, in a lamination based passive micromixers, there can be crossing manifold structure with horizontally crossing manifold micromixer (h-cmm), and vertically crossing manifold micromixer (v-cmm). buchegger et al. (2011) proposed a horizontal multi-lamination micromixer having wedge-shaped vertical fluid inlets for enhanced mixing performance in a short time. the wedge-shaped inlets produced uniform lamination layers, which were not observed in the micromixer with a straight inlet. nakahara et al. (2011) proposed a 3d passive lamination micromixer comprised of a multilayered flow generator. by using the proposed technique of the flow generator, multilayer flows were combined in a channel to generate fluid layers with reduced diffusion length. the experimental investigation showed that the proposed design achieved 1.64 times faster mixing compared to a micromixer with y-inlet. generally, the mixing increases with the increase in the number of sub-streams. the wedge-shaped inlets produce uniform diffusion layers hence, improved mixing efficiency. further, crossing manifolds, although they increased the width of the mixing channel, improved the mixing efficiency significantly. 2.4. based on flow obstacles the obstacles in the flow streams enhance the mixing performance through the generation of vortices and afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 114 chaotic advection, thereby shortening the diffusion path length. depending upon the positioning of the obstacles, the passive micromixers can be categorized as micromixers with obstacles furrowed/attached to the sidewalls, micromixers with obstacles inside the channels, i.e., attached to the top/bottom wall, and micromixers with herringbone grooves/ribs as obstacles. 2.4.1. obstacles furrowed/attached/embedded on the channel sidewalls under this category, the designs with obstacles of different shapes attached to the sidewalls of the straight/curved microchannels are discussed. nason et al. (2014) numerically studied flow and mixing in a 3d channel with rectangularand triangular-shaped obstacles on the walls for reynolds numbers in the range of 10 to 100. the effect of phase shift of obstacles, channel aspect ratio, and obstacle height on mixing and pressure drop were studied. the greater phase shift of the obstacles produced enhanced mixing because of increased recirculation in the cross-section along with a lower pressure drop compared to a no-phase shift of the obstacles. the design having triangular-shaped obstacles achieved higher mixing performance and lower pressure drop than the design with rectangularshaped obstacles. the effect of the channel aspect ratio on the mixing was insignificant, but the pressure drop varied significantly and showed the highest pressure drop for the highest aspect ratio. furthermore, the design with the largest obstruction height showed maximum mixing at higher reynolds numbers. sarma and patowari (2016) numerically investigated the mixing and pressure drop in passive micromixers with rectangular-, triangularand semicircular-shaped obstructions attached to the wall in a flow of ultra-low reynolds number of 0.053334. the authors investigated the effects of different parameters, i.e., intersection angle between the inlets (eθ), obstacle arrangement, obstacle depth, channel aspect ratio (ar), and obstacle packing factor (opf) on mixing efficiency, and found that the design with rectangular-shaped obstacles achieved the highest mixing with the highest pressure drop, while triangular and semi-circular obstacles had a similar mixing performance. the design with triangular obstacles showed the least pressure drop among these designs. the design with a staggered arrangement showed the highest mixing without significant change in pressure drop. moreover, the greater obstacle depth and obstacle packing factor increased both mixing and pressure drop, whereas a larger entrance angle enhanced mixing with a negligible effect on the pressure drop. wang et al. (2015) investigated numerically mixing in a channel with symmetrical cylindrical furrows /grooves for re = 1–500. they reported that the addition of grooves in the channel increased the mixing performance by 2.22 and 2.18 times compared to conventional y-micromixer at re = 10 and 100, respectively. vijayanandh et al. (2019) studied the effect of different shapes (square, curved, and triangular) of ridges attached to the wall of a meandering channel. they reported that the channel with triangular-shaped ridges has higher mixing performance compared to the geometries with square and curved-shaped ridges. tsai and wu (2011) designed a planar passive micromixer consisting of radial baffles in a curved channel. multidirectional vortices, i.e., dean vortices due to the curved channel and longitudinal vortices due to the radial obstacles, produced significant mixing in this design. milotin and lelea (2016) investigated mixing in three passive micromixers ( design with no baffle in microtube, design with four baffles covering the half of the cross-section, and the design with four baffles covering the quarter of the cross-section) for re = 0.2 – 91. the baffles covering the quarter of the cross-section showed higher mixing for all re values due to rotational flow, increased interface area and higher residence time caused by 90°– delayed orifice. lin et al. (2011) designed a chaotic micromixer consisting of a square-wave structure and periodic cubic grooves. the experimental comparison indicated that the overall performance was better than the performances of accoris, slit interdigital, caterpillar, and t-mixer for re values from 30 to 220. the higher mixing was due to the combined effect of flow stretching (caused by cubic grooves) and laminar recirculation (caused by square-wave structure). fractal structures such as cantor, minkowski, and koch have also been used as obstacles by many researchers to enhance mixing. wu and chen (2019b) proposed a rectangular fractal-based micromixer named imitate cantor structure (ics). the effects of fractal obstacle number, obstacle height, fractal series, and obstacle spacing on mixing were studied using numerical simulations. the better performance of a quadratic fractal series with four obstacles as compared to the primary fractal series with the same number of obstacles indicated that the increase in the number of obstacles and fractal series has a positive impact on the mixing. furthermore, a staggered arrangement with zero spacing and the largest height showed efficient performance due to vortex generation. wu and chen (2019a) introduced a 3d micromixer consisting of a cantor structure termed as imitate cantor structure micromixer (icsm). the cantorbased obstacle improved mixing with the mechanisms of chaotic advection and folding of fluid interfaces. they investigated the impact of channel height, obstacle height, the spacing between the obstacles, and fractal obstacles arrangement on mixing performance for reynolds numbers ranging from 0.01 to 100. different configurations were obtained through variations in the above-mentioned parameters, which showed mixing performance of over 90% for 0.1 > re > 50. the icsm with a channel height of 600 μm showed the best performance with a minimum efficiency of 85% over the entire range of re. the design and mixing performance of passive micromixers: a critical review 115 increase in the channel height increased an increase in the effective area for fluid folding and, thus, mixing. chen and chen (2020) examined the effect of minkowski fractal obstacles of both primary and secondary types on the mixing and pressure drop at re = 0.01, 0.1, 1, 10, and 100. the micromixer with secondary minkowski fractal obstacles showed better mixing than the micromixer with primary type obstacles with a value of more than 98% at re = 0.01, 0.1, and 100, but with a higher pressure drop. mixing was enhanced due to vortex formation and chaotic convection with an increase in the obstacle height and a reduction in the obstacle spacing. the vortex due to fractal obstruction was more obvious in the micromixer with minimum obstacle spacing, maximum obstacle height, and the highest reynolds number. zhang and chen (2018) examined the impact of height, spacing, numbers, and configurations of the koch fractal obstacles attached to the sidewalls on mixing performance. they concluded that greater height, more number, and less spacing of the obstacles have a positive effect on mixing performance. different configurations formed with four obstacles have a negligible effect on the mixing. the vortex areas produced due to the obstacle helped the mixer to achieve an efficiency of more than 90% at re = 0.03 and 240. in a similar study with koch fractal obstacle, x. chen and tian (2020) studied the effects of depth, spacing, number, angular inclination, and distribution of baffles on the mixing. the configuration with the bilateral arrangement showed better mixing than the single-sided baffle structure at re ≥ 10. furthermore, the design with a lesser spacing, more numbers, and higher inclination of baffles showed better mixing performance. chen et al. (2019b) studied the effects of rounding the corners of obstacles and examined mixing in the primary fractal micromixer (pfm), secondary fractal micromixer (sfm), and rounding the corners of the secondary fractal micromixer (rcsm). they noticed a higher mixing efficiency of sfm (over 90 % at 0.1, 1, 10, and 100) and rcsm compared to pfm. the rounding of the corner reduced the pressure drop significantly with a slight decrease in mixing compared to sfm. chen and chen (2019) proposed several topological micromixers with circular structures (tmcs) with different ratios of an obstacle to channel height using the topological optimization method. the obstacles reversed the flow and promoted mixing through an increased flow path at low-re and enhanced chaotic advection at high-re. the mixing was obtained in the configuration tmc0.75 with an efficiency of more than 90% over the entire re-range. the chaotic advection in the circular mixer was seen at re = 50, while in the tmc mixer, it was observed at re = 10. cortes-quiroz et al. (2010) performed numerical simulation and optimization of a tmicromixer having curved baffles attached to the wall over a reynolds number range of 1–250. they studied the effects of baffle radius, baffle pitch, and height of the channel on the mixing and pressure drop. the mixing was enhanced due to recirculation and transversal flows caused by the radial baffles. besides, the curved shape reduces dead volumes and the issue of clogging in the mixer. the proposed design achieved a mixing of over 85% for re = 1 and ≥ 50. however, due to the larger pressure drop for re > 100, a reynolds number range of 50–100 was recommended for efficient operation. xu and chen (2019) proposed four passive micromixers, i.e., common micromixer (cm), common micromixer with scaling elements (cmwse), snake-like micromixer (sm), and snakelike micromixer with scaling elements (smwse), and evaluated their mixing performance and pressure drop using numerical simulations. they observed that scaling elements enhance chaotic advection and thus positively affect the mixing performance. the order of mixing performance exhibited was as smwse (over 92 %) > sm > smwse > cm. however, the pressure drop in the mixers with scaling elements was found to be higher than in the designs without it. the recirculation and transverse flow induced by the baffle, obstacles, and groves on the channel walls, in general, increased flow mixing, which is further increased with the increase in the height of the obstacles/baffles. further, the packing factor and spread of these obstacles significantly affect the mixing and pressure drop in the channel. the secondary fractals add to the mixing phenomena; moreover, rounding of fractal corners reduces pressure drop significantly with a small decrease in mixing performance. 2.4.2. obstacles attached to the top/bottom walls several researchers have performed the mixing enhancement by placing obstacles inside the channel by attaching them to the top/bottom walls. monaco et al. (2010) studied the effects of different shapes and arrangements of obstacles on the mixing in a t-mixer using the lattice-boltzmann method. they reported that the layout of obstacles is a more dominant factor for mixing enhancement than using many obstacles. furthermore, rectangular-shaped obstacles showed the best mixing. chen and zhao (2017) performed the optimization of obstacles-layout in a tmixer through an orthogonal experiment and investigated the decreasing order of sensitive parameters, i.e., obstacles height > geometric shape > symmetric layout = a number of obstacles. the optimized multi-unit obstacle micromixer showed more than 90% mixing over a wide range of peclet numbers. the blockage due to the obstacles changed the velocity field and promoted chaotic advection to enhance the mixing. alam et al. (2014) studied the impact of cylindrical obstacles in a curved microchannel on the mixing performance, applying numerical simulations for re values in the range of 0.1 to 60. they noticed higher afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 116 mixing efficiency for the design with cylindrical obstacles in a curved microchannel (88 % at re =0.1 and re > 15) compared to the designs with circular obstacles in a t-shape channel and a simple curved channel. the circular and hexagonal obstructions showed similar mixing performances at all re values; on the other hand, diamond obstructions had less mixing performance except at re = 50. wang et al. (2014) proposed a passive mixer with triangular baffles embedded in the main channel and evaluated mixing both experimentally and numerically at different re values ranging from 0.1 to 500. the proposed design showed higher mixing (2.48, 4.75, and 8.32 times at re = 1.0, 100, and 500, respectively) than a conventional y-shape mixer. the mixing increased with the increase in the apical angles of the triangle and the number of triangles. out of cross, triangle, pentagon, star, hexagon, and i-shape obstacles in the y-shape passive micromixers, obstacles with sharp edges showed higher mixing performance. the mixing performance of cross, star-, and i-shaped obstacles was 77%, 75%, and 72%, respectively (pravinraj and patrikar, 2019). the addition of a screw-shaped propeller blade in a ymixer generated additional flow disturbances in the zdirection and, thus, more mixing compared to a flatshaped blade (liu et al., 2014). the design, even with a single screw-shaped propeller blade, achieved a mixing of over 80% for a flow rate of 12 µl/min. the introduction of a porous twisted tape (permeability = 10-10 m2) in a t-channel improved the mixing at low to medium re, while the solid twisted tape improved the mixing at high-re (kurnia and sasmito, 2019). the twisted tape increased the secondary flow and convective mass transfer with an additional pressure drop. pendharkar and patrikar (2014) designed a hexagonal passive micromixer and studied the effects of microchannel fabrication on mixing performance for re values in the range of 0.1 to 0.5. the addition of roughness to the passive mixer geometry resulted in an improvement in mixing (up to 10% at re = 0.1 to 0.5). shi et al. (2019) numerically studied mixing in a koch fractal-based obstructions micromixer and reported that the obstruction pattern present on the top wall (same side) of the channel wall produced more mixing (approx. 98.2%, at re = 0.1). cook et al. (2012) designed a planar scaled-up micromixer consisting of an uneven interdigital inlet, staggered teardrop-shaped obstructions, and wall protrusions. the experimental investigation at re = 1˗100 showed that the proposed design achieved the maximum efficiency of 68.5% at re = 1 because of multi-lamination caused by the uneven interdigital inlet. the vortices formed due to the obstructions caused mixing at high reynolds numbers. the layout of the obstacles is a more significant parameter than the number of obstacles, although mixing efficiency increases with an increase in the number of obstacles, which directly contributes to pressure drop in the channel. the sharp-edged obstacles are more effective for mixing enhancement. furthermore, the packing of obstacles may have deteriorating effects due to the trapping of fluid on the mixing performance. 2.4.3. ribs and grooves as obstacles many researchers have implemented obstacles in the form of ribs and grooves to enhance the chaotic motion of the fluids and mixing. several findings have revealed that the more the number of internal ribs, the higher the mixing performance of passive micromixers. the height of a channel affects the flow, and the most effective ratio of the height of the main channel to the herringbone structure was found to be 2:1 (whulanza et al., 2018). the closely spaced blocks in the microchannel can result in the trapping of fluid and can affect the mixing throughput. kim et al. (2011) studied the effects of four geometric parameters of ribs, such as rib angle, rib height, rib width, and rib spacing numerically, to optimize the mixing performance using the response surface method. the increase in the rib height reduced the area between the wall and rib, which caused the transversal velocity of the vortices to increase and, thus, the mixing. they observed maximum mixing efficiency (over 90 % with channel length 1344 µm) when rib angle, rib height/channel depth, rib width/rib height, and rib spacing/rib height were 35.6°, 0.7, 0.127, and 1.10, respectively. du et al. (2010) analyzed the mixing performances of slanted groove micromixers (sgms) numerically and staggered herringbone micromixers (shms) (pendharkar, 2014). single helical flow in sgm and two helical flows, i.e., large vortex for long arm and small vortex for shorter arm in shm, were observed. furthermore, a large peak-to-peak distance in the concentration profile and slow convergence were observed in sgm. hence, faster and more efficient mixing in shm, while slower and coarser mixing in sgm was noticed. guo et al. (2010) studied numerically the impact of rib width, pitch, and channel aspect ratio on mixing performance. the mixing performance of the proposed design decreased with the increase of h/w (channel height/channel width) up to a value of 0.6, and beyond that, the influence was negligible. the ribs caused enlargement of the interfacial area, and thus higher mixing was seen with the addition of more ribs and larger width of rib. in an experiment conducted on a passive scaled-up micromixer with four meandering elements and 36 slanted grooved structures (horizontally stacked), cook and hasan (2011) reported higher performance with the increment in the value of re from 5 to 100 due to the formation of dean vortices. additionally, cook et al. (2013) designed a scaled-up passive micromixer consisting of a main mixing channel of 155.8 mm long, 3 mm wide, and 0.75 mm deep, with nine slanted grooves, and measured the mixing performance experimentally and numerically for re values ranging from 0.5 to 100. they reported mixing performance design and mixing performance of passive micromixers: a critical review 117 from 53% at re = 5 up to 90% at re = 100. the dean vortices and helical flows were formed due to the curved channel structure and grooves, respectively. chen et al. (2015) designed a passive micromixer having dual opposing strips on the top and bottom walls of a microchannel and varying cross-sectional area to enhance mixing performance by increasing contact surface area, two-layer vortices generation, and enhanced diffusion flux. the numerical simulation revealed mixing performance of 93.97%, 92.53%, and 92.49% at re = 0.56, 2.8, and 5.6, respectively. however, misalignment between the top and bottom grooves reduced the mixing. fan and hassan (2010) proposed a curved micromixer with grooves (cmg) and compared its mixing performance with a slanted grooved micromixer (sgm) and curved micromixer without grooves (cm) for re values in the range of 1 to 50. cmg showed better performance than sgm and cm. in addition, both cmg and sgm achieved better mixing than cm at re = 1.0. the new design showed 60% mixing efficiency at re = 50 with a 1.8 kpa pressure drop. at high re values, the curved channel created the dean vortices, while slanted grooves generated helical flow at low-re values. the elements optimized for the new design included the slanted angle (θ), the width of the grooves angle (ω), and the height ratio of grooves to channel height (hg/h). okuducu and aral (2019b) proposed a passive micromixer with semi-circular ridges in the microchannel. the new design with the convex alignment of semi-circular elements was found to have a mixing index and mixing performance more than the classical t-shape mixer by the factors of 8.7 (at re = 1 to 10) and 3.3 (at re = 20), respectively. they reported the enhanced mixing (over 80 %) due to the helicoidalshaped fluid stream generated by the convex alignment of semi-circular elements for re values from 0.1 and 40. the staggered herringbone design has v-shaped ridges attached to the channel walls, and these arrangements enhance mixing at low re values by creating transverse flow patterns. hama et al. (2018) proposed a micromixer (reverse-staggered herringbone) based on staggered herringbone design geometry and evaluated the mixing performance both numerically and experimentally. they found that the flow rate ratio of fluids positively affects the mixing; however, the re values impact was not significant in the first three and six cycles. the number of ribs creates more chaos. hence more mixing of fluid in the channel, the rib height to channel height ratio is a critical parameter in micromixer design; the larger the ratio greater the mixing. a staggered herringbone groove micromixer, which induces dual helical flows, performs better than a mixer with inline helical grooves. 3. analysis and discussion the reynolds number and peclet number are the two significant parameters for the design of passive micromixers. hence, for an accurate comparison of the performance of different micromixer designs and configurations, it is essential to define these numbers in a similar manner and at the same location. however, after a thorough literature review, it was noticed that most of the studies had defined these numbers either at the inlets or at the main channel. therefore, carrying out performance comparisons between different designs based on these numbers is highly cumbersome. furthermore, the review of studies reveals that most of the studies performed have reported only mixing index without attributing proper significance to the pressure drop encountered, device fabrication methods, design complexity, wettability effects, and surface roughness. hence, for proper advancement of the micromixer designs and their adoption to real applications, the issues mentioned above should be addressed through a proper design criterion prepared by the research community for future studies. the research studies on micromixers have highlighted that many variables, such as geometric parameters, mixing-channel designs, and flow reynolds numbers, affect the mixing performance. the high reynolds number value reduces the contact time between the two fluids and decreases mass diffusion leading to lower efficiency in diffusive mixing-based designs. the fluid mixing at a very low re value is dominated by the residence time and the total path of the flow (nason et al., 2014). this study presents a classification of micromixers into several categories, such as designed inlets, designed mixing channels, lamination phenomenon, and flow obstacles. additionally, charts and graphs present the study outcomes. analysis of the research studies has shown important outcomes, which will assist in understanding various aspects of micromixers. figure 5 shows the published articles and evaluation methods used in the studies. the year 2019 reported the highest number of studies, followed by 2014, 2015, 2011, and 2018. the performance evaluation methods used in these studies are mainly numerical simulations or both numerical simulations and experiments, while very few have carried out the only experimental study. the maximum number of studies correspond to experiments and numerical simulations together, followed by numerical simulations alone. most of the studies revealed that numerical simulations show relatively higher mixing than the actual value achieved through experiments. but a great deal of work has been done on mixing performance via numerical simulations. however, the reliability of the results is always doubtful due to external factors of the real environment unless they are substantiated with experiments. hence, there is a need to conduct more studies with experiments for measuring the micromixer performance rather than doing only numerical simulations. afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 118 figure 5. (a) the number of published articles on passive micromixers, and (b) evaluation methods used in passive micromixers. figure 6. (a) design type as reported in different studies, and (b) mixing-channel designs as reported in different studies. (a) (b) (a) (b) design and mixing performance of passive micromixers: a critical review 119 (a) (b) figure 7. (a) number of articles with different inlet shape, and (b) number of inlets as reported in the existing studies. (a) (b) figure 8. (a) the mixing performance of passive micromixers, and (b) re values considered in the existing studies. figure 6 shows the distribution of the studies carried out under different microchannel categories and mixing-channel designs. it shows that the maximum number of obstacle-based designs have been proposed, followed by sar, mixed category, designed inlet, c-d, and wave. the microchannel has various designs as sinusoidal, square, trapezoidal, etc. the passive micromixers have inlet shapes as y-, t-, f-, e-, h-c, tapered, and hybrid. the number of studies on micromixers is plotted against inlet shape. the maximum number of passive micromixers reported have inlet shape as t-, followed by y-, and hybrid. moreover, the most reported number of inlets is two, followed by three. most of the researchers who reported high performance have used two inlets. figure 7 shows the distribution of studies reported under different inlet shapes and their numbers. most of the studies have reported achieving maximum mixing performance between 80 and 100%. very few of them have shown mixing performance between 60 to 75 %. figure 8 shows the mixing performance of passive micromixers and the range of re values reported in the existing studies. although the results of the analysis cannot firmly establish the superiority of one method over the others, it can qualify for demonstrating the trends of passive micromixer performance concerning different factors. appendix a presents the reported studies with above 90% mixing performance, including re values, designs, and other details. 4. conclusion the number of articles on passive micromixers has increased significantly due to proposed wider applications; however, the research on the mixing performance and micromixer designs is highly fragmented. this review of designs and mixing performance of micromixers have classified passive micromixers based on the designed inlets, mixingchannel design, fluid lamination, and obstacles in the channel. based on this review study, the following points can be concluded: 1. most of the researchers have used numerical simulations for evaluating micromixer designs, although the simulation results are not always comparable to physical experiments. 2. micromixers with t-shape inlets and two inlets are the most preferred choice of micromixers researchers. the obstacles and sar-based micromixer designs have the edge over other types. moreover, new mixing-channel designs, such as the koch fractal and minkowski fractal, are under investigation for higher mixing performance. 3. although several designs have demonstrated higher mixing performance using water-dye experimental and numerical studies, there is a scarcity of work that can demonstrate their capabilities in real applications. furthermore, in most of these works, little consideration has been afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 120 given to the required device material and fabrication process. 4. the open literature lacks micromixer classification based on microfluidic applications. 5. the adoption of micromixers to the real application is highly dependent upon the ease of fabrication and device integration. hence, future research should consider the factors like material, fabrication method, and design integration along with capabilities in proposed applications. 6. the micromixer designs demonstrating good merit in these aspects, in addition to better mixing, will help integrate them into microsystems for actual applications and accelerate the market growth of microfluidic systems. conflict of interest the authors have no conflict of interest. funding deanship of research fund, grant no. rf/eng/meid/19/01. acknowledgment the authors acknowledge the support of sultan qaboos university, oman, for conducting this research. references adam t, hashim u (2012), simulation of passive fluid driven micromixer for fast reaction assays in nano lab-on-chip domain. procedia engineering(50): 416-425. afzal a, kim k-y (2012), passive split and recombination micromixer with convergent– divergent walls. chemical engineering journal 203: 182-192. afzal a, kim k-y (2015a), convergent–divergent micromixer coupled with pulsatile flow. sensors and actuators b: chemical 211: 198-205. afzal a, kim k-y (2015b), multi-objective optimization of a passive micromixer based on periodic variation of velocity profile. chemical engineering communications 202(3): 322-331. al-halhouli aa, alshare a, mohsen m, matar m, dietzel a, büttgenbach s (2015), passive micromixers with interlocking semicircle and omega-shaped modules: experiments and simulations. micromachines 6(7): 953-968. alam a, afzal a, kim k-y (2014), mixing performance of a planar micromixer with circular obstructions in a curved microchannel. chemical engineering research and design 92(3): 423-434. ansari ma, kim k-y (2010), mixing performance of unbalanced split and recombine micomixers with circular and rhombic sub-channels. chemical engineering journal 162(2): 760-767. ansari ma, kim k-y, anwar k, kim sm (2010), a novel passive micromixer based on unbalanced splits and collisions of fluid streams. journal of micromechanics and microengineering 20(5): 055007. ansari ma, kim k-y, kim sm (2018), numerical and experimental study on mixing performances of simple and vortex micro t-mixers. micromachines 9(5): 204. bayareh m, ashani mn, usefian a (2019), active and passive micromixers: a comprehensive review. chemical engineering and processing-process intensification 147: 107771. bazaz sr, mehrizi aa, ghorbani s, vasilescu s, asadnia m, warkiani me (2018), a hybrid micromixer with planar mixing units. rsc advances 8(58): 33103-33120. bhopte s, sammakia b, murray b (2010), "numerical study of a novel passive micromixer design," in proc. 2010 12th ieee intersociety conference on thermal and thermomechanical phenomena in electronic systems, las vegas, nv, usa, pp. 110. buchegger w, wagner c, lendl b, kraft m, vellekoop mj (2011), a highly uniform lamination micromixer with wedge shaped inlet channels for time resolved infrared spectroscopy. microfluidics and nanofluidics 10(4): 889-897. chen p-c, pan c-w, kuo y-l (2015), performance characterization of passive micromixer with dual opposing strips on microchannel walls. chemical engineering and processing: process intensification 93: 27-33. chen x, li t (2017), a novel passive micromixer designed by applying an optimization algorithm to the zigzag microchannel. chemical engineering journal 313: 1406-1414. chen x, li t, li x (2016a), numerical research on shape optimization of microchannels of passive micromixers. ieee sensors journal 16(17): 65276532. chen x, li t, zeng h, hu z, fu b (2016b), numerical and experimental investigation on micromixers with serpentine microchannels. international journal of heat and mass transfer 98: 131-140. chen x, liu s, chen y, wang s (2019a), a review on species mixing in droplets using passive and active micromixers. international journal of environmental analytical chemistry 101:422-432. chen x, shen j, hu z (2018), fabrication and performance evaluation of two multilayer passive micromixers. sensor review 38(3): 321-325. chen x, tian y (2020), passive micromixer with baffles distributed on both sides of microchannels based on the koch fractal principle. journal of chemical technology & biotechnology 95(3): 806-812. chen x, zhang s, wu z, zheng y (2019b), a novel design and mixing performance of passive micromixers: a critical review 121 koch fractal micromixer with rounding corners structure. microsystem technologies 25(7): 27512758. chen x, zhao z (2017), numerical investigation on layout optimization of obstacles in a threedimensional passive micromixer. analytica chimica acta 964: 142-149. chen y, chen x (2019), an improved design for passive micromixer based on topology optimization method. chemical physics letters 734: 136706. chen y, chen x (2020), numerical investigations of a passive micromixer based on minkowski fractal principle. international journal of chemical reactor engineering 18(2): 20190105. cheri ms, latifi h, moghaddam ms, shahraki h (2013), simulation and experimental investigation of planar micromixers with short-mixing-length. chemical engineering journal 234: 247-255. chung c-k, shih t, wu b, chang c (2010), design and mixing efficiency of rhombic micromixer with flat angles. microsystem technologies 16(8-9): 1595-1600. cook k, fan y, hassan i (2011), "experimental investigation of a scaled-up passive interdigital micromixer with circular-sector mixing elements," in proc. asme/jsme 2011 8th thermal engineering joint conference, honolulu, hawaii, usa, pp. t10118 cook kj, fan y, hassan i (2012), experimental investigation of a scaled-up passive micromixer with uneven interdigital inlet and teardrop obstruction elements. experiments in fluids 52(5): 1261-1275. cook kj, fan y, hassan i (2013), mixing evaluation of a passive scaled-up serpentine micromixer with slanted grooves. journal of fluids engineering 135(8). cook kj, hassan ig (2011), "experimental investigation of a scaled-up passive curved channel micromixer with slanted grooves," in proc. asme 2011 9th international conference on nanochannels, microchannels, and minichannels, edmonton, alberta, canada, pp. 111-117. cortes-quiroz ca, azarbadegan a, moeendarbary e, zangeneh m (2010), "analysis and optimization of a passive micromixer with curved-shaped baffles for efficient mixing with low pressure loss in continuous flow," in proc. asme 2010 8th international conference on nanochannels, microchannels, and minichannels collocated with 3rd joint us-european fluids engineering summer meeting, montreal, canada, pp. 13471355. du y, zhang z, yim c, lin m, cao x (2010), evaluation of floor-grooved micromixers using concentration-channel length profiles. micromachines 1(1): 19-33. dundi tm, raju v, chandramohan v (2019), numerical evaluation of swirl effect on liquid mixing in a passive t-micromixer. australian journal of mechanical engineering: 1-15. fan y, hassan i (2010a), experimental and numerical investigation of a scaled-up passive micromixer using fluorescence technique. experiments in fluids 49(3): 733-747. fan y, hassan i (2010b), "numerical simulation of a novel passive micromixer with curved channel and slanted grooves," in proc. asme 2010 8th international conference on nanochannels, microchannels, and minichannels collocated with 3rd joint us-european fluids engineering summer meeting, montreal, canada, pp. 12031212. gaozhe c, li x, huilin z, jianhan l (2017), a review on micromixers. micromachines 8(274): 27. gidde rr, pawar pm (2020), flow feature and mixing performance analysis of rb-tsar and eb-tsar micromixers. microsystem technologies 26(2): 517-530. gidde rr, pawar pm, gavali sr, salunkhe sy (2019a), flow feature analysis of an eye shaped split and collision (es-sac) element based micromixer for lab-on-a-chip application. microsystem technologies 25(8): 2963-2973. gidde rr, pawar pm, ronge bp, misal nd, kapurkar rb, parkhe ak (2018), evaluation of the mixing performance in a planar passive micromixer with circular and square mixing chambers. microsystem technologies 24(6): 2599-2610. gidde rr, pawar pm, ronge bp, shinde ab, misal nd, wangikar ss (2019b), flow field analysis of a passive wavy micromixer with csar and esar elements. microsystem technologies 25(3): 10171030. guo l, zhang s, han k (2010), "numerical simulation of mixing performances of a passive micromixer with internal ribs," in proc. aip conference proceedings 1207(1), xi’an, china, pp. 1024-1029. hama b, mahajan g, fodor ps, kaufman m, kothapalli cr (2018), evolution of mixing in a microfluidic reverse-staggered herringbone micromixer. microfluidics and nanofluidics 22(5): 54. hong w, shi h, huang z, long m, xu h, liu z (2019), design and simulation of a passive micromixer with gourd-shaped channel. journal of nanoscience and nanotechnology 19(1): 206-212. hossain s, kim k-y (2014), mixing analysis of passive micromixer with unbalanced three-split rhombic sub-channels. micromachines 5(4): 913928. hossain s, kim k-y (2015), mixing analysis in a three-dimensional serpentine split-and-recombine micromixer. chemical engineering research and design 100: 95-103. hossain s, lee i, kim sm, kim k-y (2017), a micromixer with two-layer serpentine crossing channels having excellent mixing performance at low reynolds numbers. chemical engineering afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 122 journal 327: 268-277. hsieh s-s, lin j-w, chen j-h (2013), mixing efficiency of y-type micromixers with different angles. international journal of heat and fluid flow 44: 130-139. jang h, pawate as, bhargava r, kenis pja (2019), polymeric microfluidic continuous flow mixer combined with hyperspectral ft-ir imaging for studying rapid biomolecular events. lab chip 19(15): 2598-2609. jaywant sa, arif km (2019), a comprehensive review of microfluidic water quality monitoring sensors. sensors (basel) 19(21). jeong gs, chung s, kim cb, lee sh (2010), applications of micromixing technology. analyst 135(3): 460-473. johnson tj, locascio le (2002), characterization and optimization of slanted well designs for microfluidic mixing under electroosmotic flow. lab on a chip 2: 135. kastania as, tsougeni k, papadakis g, gizeli e, kokkoris g, tserepi a, gogolides e (2016), plasma micro-nanotextured polymeric micromixer for dna purification with high efficiency and dynamic range. anal chim acta 942: 58-67. kefala i, papadopoulos ve, kokkoris g, karpou g, moschou d, papadakis g, tserepi a (2014), a passive micromixer for bioanalytical applications, in proc 4th micro and nano flows conference, ucl, london, uk, pp. 1-8. khan aa, tandon u (2017), simulation of wingshaped passive micromixers using comsol. international journal of computer science engineering (ijcse) 6(1): 10-17. kim bs, kwak bs, shin s, lee s, kim km, jung h-i, cho hh (2011), optimization of microscale vortex generators in a microchannel using advanced response surface method. international journal of heat and mass transfer 54(1-3): 118-125. kurnia jc, sasmito ap (2019), performance evaluation of liquid mixing in a t-junction passive micromixer with a twisted tape insert. industrial & engineering chemistry research 59(9): 3904-3915. le the h, le-thanh h, tran-minh n, karlsen f (2014), "a novel passive micromixer with trapezoidal blades for high mixing efficiency at low reynolds number flow," in proc. 2nd middle east conference on biomedical engineering, pp. 25-28. le the h, le thanh h, dong t, ta bq, tran-minh n, karlsen f (2015), an effective passive micromixer with shifted trapezoidal blades using wide reynolds number range. chemical engineering research and design 93: 1-11. lee c-y, fu l-m (2018), recent advances and applications of micromixers. sensors and actuators b: chemical 259: 677-702. li l, chen qd, tsai c (2014), three dimensional triangle chaotic micromixer. advanced materials research 875-877: 1189-1193. li x, chang h, liu x, ye f, yuan w (2015), a 3-d overbridge-shaped micromixer for fast mixing over a wide range of reynolds numbers. journal of microelectromechanical systems 24(5): 13911399. lim tw, son y, jeong yj, yang d-y, kong h-j, lee k-s, kim d-p (2011), three-dimensionally crossing manifold micro-mixer for fast mixing in a short channel length. lab on a chip 11(1): 100-103. lin y, yu x, wang z, tu s-t, wang z (2011), design and evaluation of an easily fabricated micromixer with three-dimensional periodic perturbation. chemical engineering journal 171(1): 291-300. liu k, yang q, chen f, zhao y, meng x, shan c, li y (2015), design and analysis of the cross-linked dual helical micromixer for rapid mixing at low reynolds numbers. microfluidics and nanofluidics 19(1): 169-180. liu y-j, chen p-y, yang j-y, tsou c, lee y-h, baldeck p, lin c-l (2014), three-dimensional passive micromixer fabricated by two-photon polymerization for microfluidic mixing. sensors and materials 26(2): 39-44. matsunaga t, nishino k (2014), swirl-inducing inlet for passive micromixers. rsc advances 4(2): 824829. mehrdel p, karimi s, farré-lladós j, casals-terré j (2018), novel variable radius spiral–shaped micromixer: from numerical analysis to experimental validation. micromachines 9(11): 552. milotin r, lelea d (2016), the passive mixing phenomena in microtubes with baffle configuration. procedia technology 22: 243-250. moghimi m, jalali n (2020), design and fabrication of an effective micromixer through passive method. journal of computational & applied research in mechanical engineering (jcarme) 9(2): 371383. monaco e, luo kh, brenner g (2010), "numerical investigation on the efficiency of a passive micromixer with the lattice boltzmann method," in proc. v european conference on computational fluid dynamics eccomas cfd, lisbon, portugal, pp. 1-20. mondal b, mehta sk, patowari pk, pati s (2019), numerical study of mixing in wavy micromixers: comparison between raccoon and serpentine mixer. chemical engineering and processing-process intensification 136: 44-61. mukhopadhyay s (2018), short review on the classification of micromixers. international journal of composite materials and matrices 4(1): 1-3. nakahara t, ootani n, asanuma t, hagio y, hiramaru d, terao k, okonogi a, oohira f, kotera h, suzuki t (2011), "development of a three dimensional passive lamination micromixer," in proc. asme-jsme-ksme 2011 joint fluids engineering conference, hamamatsu, japan, pp. design and mixing performance of passive micromixers: a critical review 123 371-373. nason f, pennati g, dubini g (2014), computational modeling of passive furrowed channel micromixers for lab-on-a-chip applications. journal of applied biomaterials & functional materials 12(3): 278-285. nimafar m, viktorov v, martinelli m (2012), experimental comparative mixing performance of passive micromixers with h-shaped sub-channels. chemical engineering science 76: 37-44. okuducu mb, aral mm (2019a), computational evaluation of mixing performance in 3-d swirlgenerating passive micromixers. processes 7(3): 121. okuducu mb, aral mm (2019b), novel 3-d t-shaped passive micromixer design with helicoidal flows. processes 7(9): 637. papadopoulos v, kefala i, kaprou g, kokkoris g, moschou d, papadakis g, gizeli e, tserepi a (2014), a passive micromixer for enzymatic digestion of dna. microelectronic engineering 124: 42-46. parsa mk, hormozi f, jafari d (2014), mixing enhancement in a passive micromixer with convergent–divergent sinusoidal microchannels and different ratio of amplitude to wave length. computers & fluids 105: 82-90. pendharkar g, deshmukh r, patrikar r (2014), investigation of surface roughness effects on fluid flow in passive micromixer. microsystem technologies 20(12): 2261-2269. pennella f, rossi m, ripandelli s, rasponi m, mastrangelo f, deriu ma, ridolfi l, kähler cj, morbiducci u (2012), numerical and experimental characterization of a novel modular passive micromixer. biomedical microdevices 14(5): 849862. pravinraj t, patrikar r (2019), "modeling, fabrication and investigation of mixing in low-cost passive pdms micromixers," in proc. 2019 32nd international conference on vlsi design and 2019 18th international conference on embedded systems (vlsid), delhi, india, pp. 185-190. rafeie m, welleweerd m, hassanzadeh-barforoushi a, asadnia m, olthuis w, ebrahimi warkiani m (2017), an easily fabricated three-dimensional threaded lemniscate-shaped micromixer for a wide range of flow rates. biomicrofluidics 11(1): 014108. rampalli s, dundi tm, chandrasekhar s, raju v, chandramohan v (2020), numerical evaluation of liquid mixing in a serpentine square convergentdivergent passive micromixer. chemical product and process modeling: 11. raza w, hossain s, kim k-y (2020), a review of passive micromixers with a comparative analysis. micromachines 11(5): 455. ruijin w, beiqi l, dongdong s, zefei z (2017), investigation on the splitting-merging passive micromixer based on baker's transformation. sensors and actuators b: chemical 249: 395-404. sadabadi h, packirisamy m, wüthrich r (2013), high performance cascaded pdms micromixer based on split-and-recombination flows for lab-on-a-chip applications. rsc advances 3(20): 7296-7305. sarma p, patowari pk (2016), design and analysis of passive y-type micromixers for enhanced mixing performance for biomedical and microreactor application. journal of advanced manufacturing systems 15(03): 161-172. scherr t, quitadamo c, tesvich p, park ds-w, tiersch t, hayes d, choi j-w, nandakumar k, monroe wt (2012), a planar microfluidic mixer based on logarithmic spirals. journal of micromechanics and microengineering 22(5): 055019. sheu ts, chen sj, chen jj (2012), mixing of a split and recombine micromixer with tapered curved microchannels. chemical engineering science 71: 321-332. shi x, wang l, huang s, li f (2019), a novel passive micromixer with array of koch fractal obstacles in microchannel. journal of dispersion science and technology 42(2): 236-247. stanciu i (2015), uncertainty analysis of mixing efficiency variation in passive micromixers due to geometric tolerances. modelling and simulation in engineering 2015: 343087. su t, cheng k, wang j, xu z, dai w (2019), a fast design method for passive micromixer with angled bend. microsystem technologies 25(11): 43914397. ta bq, lê thanh h, dong t, thoi tn, karlsen f (2015), geometric effects on mixing performance in a novel passive micromixer with trapezoidalzigzag channels. journal of micromechanics and microengineering 25(9): 094004. taheri ra, goodarzi v, allahverdi a (2019), mixing performance of a cost-effective split-andrecombine 3d micromixer fabricated by xurographic method. micromachines 10(11): 786. tran-minh n, dong t, karlsen f (2014), an efficient passive planar micromixer with ellipse-like micropillars for continuous mixing of human blood. computer methods and programs in biomedicine 117(1): 20-29. tsai r-t, wu c-y (2011), an efficient micromixer based on multidirectional vortices due to baffles and channel curvature. biomicrofluidics 5(1): 014103. tseng l-y, yang a-s, lee c-y, hsieh c-y (2011), cfd-based optimization of a diamond-obstacles inserted micromixer with boundary protrusions. engineering applications of computational fluid mechanics 5(2): 210-222. vijayanandh v, pradeep a, suneesh p, babu ts (2019), "design and simulation of passive micromixers with ridges for enhanced efficiency," in proc. iop conference series: materials science and engineering, bengaluru, india, pp. 012106. viktorov v, mahmud mr, visconte c (2015), afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 124 comparative analysis of passive micromixers at a wide range of reynolds numbers. micromachines 6(8): 1166-1179. viktorov v, mahmud mr, visconte c (2016), design and characterization of a new hc passive micromixer up to reynolds number 100. chemical engineering research and design 108: 152-163. viktorov v, nimafar m (2013), a novel generation of 3d sar-based passive micromixer: efficient mixing and low pressure drop at a low reynolds number. journal of micromechanics and microengineering 23(5): 055023. wang h, shi l, zhou t, xu c, deng y (2018), a novel passive micromixer with modified asymmetric lateral wall structures. asia‐pacific journal of chemical engineering 13(3): e2202. wang l, ma s, han x (2015), micromixing enhancement in a novel passive mixer with symmetrical cylindrical grooves. asia‐pacific journal of chemical engineering 10(2): 201-209. wang l, ma s, wang x, bi h, han x (2014), mixing enhancement of a passive microfluidic mixer containing triangle baffles. asia‐pacific journal of chemical engineering 9(6): 877-885. westerhausen c, schnitzler lg, wendel d, krzyszton r, lachelt u, wagner e, radler jo, wixforth a (2016), controllable acoustic mixing of fluids in microchannels for the fabrication of therapeutic nanoparticles. micromachines 7(9): 150. whulanza y, utomo ms, hilman a (2018), "realization of a passive micromixer using herringbone structure," in aip conference proceedings 1933(1), bali, indonesia, 040003. wu z, chen x (2019a), a novel design for 3d passive micromixer based on cantor fractal structure. microsystem technologies 25(1): 225-236. wu z, chen x (2019b), a novel design for passive micromixer based on cantor fractal structure. microsystem technologies 25(3): 985-996. xia g, li j, wu h, zhou m, wang h (2011), "a novel passive micromixer based on asymmetric splitand-recombine with fan-shaped cavity," in proc. asme 2011 9th international conference on nanochannels, microchannels, and minichannels, edmonton, alberta, canada, pp. 135-141. xie h, zhao x, yang h (2010), "experimental and numerical study on a planar passive micromixer with semicircle mixing elements," in proc. ieee/asme 2010 international conference on advanced intelligent mechatronics, montreal, qc, canada, pp. 1013-1016. xu j, chen x (2019), numerical study on mixing performance of 3d passive micromixer with scaling elements. journal of the brazilian society of mechanical sciences and engineering 41(10): 453. yang a-s, chuang f-c, chen c-k, lee m-h, chen sw, su t-l, yang y-c (2015), a high-performance micromixer using three-dimensional tesla structures for bio-applications. chemical engineering journal 263: 444-451. yang j, qi l, chen y, ma h (2013), design and fabrication of a three dimensional spiral micromixer. chinese journal of chemistry 31(2): 209-214. yang l, li s, liu j, cheng j (2018), fluid mixing in droplet‐based microfluidics with t junction and convergent–divergent sinusoidal microchannels. electrophoresis 39(3): 512-520. zadeh hf, marahel a (2011), "simulations of flow and mass transfer in a passive micromixer," in proc. asme 2011 9th international conference on nanochannels, microchannels, and minichannels, edmonton, alberta, canada pp. 179-185. zhang h, bai y, zhu n, xu j (2021), microfluidic reactor with immobilized enzyme-from construction to applications: a review. chinese journal of chemical engineering 30: 136-145. zhang s, chen x (2018), a novel passive micromixer based on koch fractal principle. journal of the brazilian society of mechanical sciences and engineering 40(10): 487. zhang w, wang x, feng x, yang c, mao z-s (2016), investigation of mixing performance in passive micromixers. industrial & engineering chemistry research 55(38): 10036-10043. zhang y, hu y, wu h (2012), design and simulation of passive micromixers based on capillary. microfluidics and nanofluidics 13(5): 809-818. design and mixing performance of passive micromixers: a critical review 125 appendix a: reference year inlet shape inlet no design type mixing-channel design mixing performance (max %) re value evaluation method (moghimi and jalali, 2020) 2020 y-shape 2 mixed injection, sar, and zigzag 98.02 100 both numerical and experimental (chen and chen, 2020) 2020 t-shape 2 obstacle furrowed secondary minkowski fractal 98 0.01 numerical simulation 98 1 98 100 (chen and tian, 2020) 2020 t-shape 2 obstacle furrowed-koch fractal (30 angle) 99.48 100 numerical simulation (rampalli et al., 2020) 2020 t-shape 2 c-d serpentine square 100 160 < or <180 numerical simulation (wu and chen, 2019a) 2019 t-shape 2 obstacle furrowed imitate cantor structure 90 >=50 or <=0. 1 numerical simulation (shi et al., 2019) 2019 t-shape 2 obstacle inside-koch fractal 98 0.1 numerical simulation (chen and chen, 2019) 2019 t-shape 2 obstacle furrowedminkowski fractal 95 >=5 numerical simulation (chen et al., 2019b) 2019 t-shape 2 obstacle furrowedsecondary koch fractal 95 0.05100 numerical simulation 90 0.1100 (mondal et al., 2019) 2019 t-shape 2 c-d raccoon 100 0.1100 numerical simulation serpentine 100 0.1100 (dundi et al., 2019) 2019 t-shape 2 designed inlet swirl induced 95 106 numerical simulation (gidde et al., 2019b) 2019 t-shape 2 sar elliptic 99 2075 numerical simulation (xu and chen, 2019) 2019 y-shape 2 obstacle furrowedsmwse 92 22 both numerical and experimental (shi et al., 2019) 2019 t-shape 2 obstacle insidekoch obstacle 98.2 0.1 numerical simulation (gidde et al., 2018) 2018 t-shape 2 sar circular and square chamber 99 <1.0 numerical simulation (zhang and chen, 2018) 2018 t-shape 2 obstacle furrowed-koch fractal 99 0.03 numerical simulation 90 240 (chen et al., 2018) 2018 f-shape 2 sar stacked 92 80 both numerical and experimental e-shape 94 80 (ansari et al., 2018) 2018 t-shape 2 c-d serpentine 91 2922 1 both numerical and experimental (mehrdel et al., 2018) 2018 t-shape 3 mixed double spiral 98.5 0.110 both numerical and experimental afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 126 (hama et al., 2018) 2018 t-shape 2 obstacle ribs-reverse staggered herringbone 98.5 100 both numerical and experimental (hossain et al., 2017) 2017 y-shape 2 sar serpentine 99 0.210 both numerical and experimental (rafeie et al., 2017) 2017 y-shape 2 wave spiral 90 11000 numerical simulation (chen and li, 2017) 2017 t-shape 2 obstacle furrowed trapezoidal zigzag 95 <=0. 5 or >=5 both numerical and experimental (chen et al., 2016a) 2016 t-shape 2 obstacle furrowedtrapezoidal zigzag 93 <=0. 5 or >=5 numerical simulation (viktorov et al., 2016) 2016 h-c shape 2 sar h-c 90 >=1 or <=10 0 both numerical and experimental (chen et al., 2016b) 2016 t-shape 2 c-d serpentine 95 0.1100 both numerical and experimental (milotin and lelea, 2016) 2016 t-shape 2 obstacle furrowedcircular 100 0.291 numerical simulation (afzal and kim, 2015a) 2015 t-shape 2 c-d sinusoidal 92 0.5 numerical simulation (chen et al., 2015) 2015 y-shape 2 obstacle ribs-two strips 93.97 0.56 both numerical and experimental (le the et al., 2015) 2015 hybrid 3 sar trapezoidal 90 >= 20, or <=0. 9 both numerical and experimental (liu et al., 2015) 2015 y-shape 2 wave double helical 99 0.003 -30 both numerical and experimental (al-halhouli et al., 2015) 2015 t-shape 2 wave ilsc 100 0.01 50 both numerical and experimental omega 100 0.01 50 semi-circular 100 > 50 (viktorov et al., 2015) 2015 y-shape 2 sar chain (y-y and h-c) 90 1100 both numerical and experimental (ta et al., 2015) 2015 hybrid 3 sar trapezoidal zigzag 94 >=20 both numerical and experimental 90 <=0. 9 (li et al., 2015) 2015 t-shape 2 sar overbridge 90 0.0150 both numerical and experimental (yang et al., 2015) 2015 t-shape 2 sar tesla 90 0.1100 both numerical and experimental (le the et al., 2014) 2014 hybrid 3 sar trapezoidal 95 40 both numerical and design and mixing performance of passive micromixers: a critical review 127 experimental (parsa et al., 2014) 2014 t-shape 2 c-d sinusoidal 99 0.2 both numerical and experimental 99 >30 (li et al., 2014) 2014 y-shape 2 sar triangular 90 0.01 numerical simulation (nason et al., 2014) 2014 y-shape 2 obstacle furrowed triangular 90 50 numerical simulation (wang et al., 2014) 2014 y-shape 2 obstacle inside-triangular 91.2 0.1 both numerical and experimental (cheri et al., 2013) 2013 hybrid 3 mixed hexagonal chamber 99 40 both numerical and experimental (hsieh et al., 2013) 2013 hybrid 2 designed inlet at -60° 99 0.027 experiment (cook et al., 2013) 2013 y-shape 2 obstacle ribs 90 100 both numerical and experimental (yang et al., 2013) 2013 y-shape 2 wave double spiral 90 40 both numerical and experimental (viktorov and nimafar, 2013) 2013 t-shape 2 sar chain 98 0.416 both numerical and experimental (afzal and kim, 2012) 2012 t-shape 2 mixed sinusoidal wall 90 70 numerical simulation (nimafar et al., 2012) 2012 h-shape 2 sar h-shape 98 0.083 both numerical and experimental (sheu et al., 2012) 2012 tapered 2 sar staggered curve 90 50 both numerical and experimental (tsai and wu, 2011) 2011 y-shape 2 obstacle furrowed-radial 93 81 both numerical and experimental (lim et al., 2011) 2011 multilaminati on mul tiple laminatio n 3d-cmm 90 1 experiment (lin et al., 2011) 2011 y-shape 2 obstacle square wave 90 >=40 both numerical and experimental (kim et al., 2011) 2011 t-shape 2 obstacle rib 95 1.69 both numerical and experimental (tseng et al., 2011) 2011 y-shape 2 obstacle diamond 96 >10 or <0.1 numerical simulation (zadeh and marahel, 2011) 2011 hybrid 4 designed inlet 300° 100 60 numerical simulation (chung et al., 2010) 2010 hybrid 3 sar rhombic 95 >180 both numerical and experimental afzal husain, asharul islam khan, wasim raza, nabeel al-rawahi, nasser al-azri, and abdus samad 128 (du et al., 2010) 2010 t-shape 2 obstacle staggered herringbone 95 0.3 numerical simulation (xie et al., 2010) 2010 t-shape 2 mixed semi-circular 90 <0.1 or >10 both numerical and experimental (ansari et al., 2010) 2010 t-shape 2 mixed unbalanced circles 90 >40 both numerical and experimental (cortes-quiroz et al., 2010) 2010 t-shape 2 obstacle furrowedcurved 95 50100 both numerical and experimental (bhopte et al., 2010) 2010 hybrid 4 designed inlet split 92 1 numerical simulation evaluation of properties of non-stoichiometric alumina magnesia spinel using thiourea as fuel by varying soaking time evaluation of properties of non-stoichiometric alumina magnesia spinel using thiourea as fuel by varying soaking time sathi banerjee 1, soumya mukherjee 2*, and srinath ranjan ghosh 1 1 department of metallurgical & materials engineering, jadavpur university, india 2* department of metallurgical engineering, kazi nazrul university, india abstract: the present article reports a simple and cost-effective process to prepare the crystalline mgal2o4 spinel using a non-stoichiometric amount of magnesium nitrate, aluminum nitrate by solution combustion route. thiourea was used as a fuel and reducing agent while soaking was carried at 1000ºc with different soaking periods. after slow drying of mixed solutions at 80ºc for 4-5 h, a gel was formed and got characterized by dta/tga (differential thermal analysis and thermal gravimetric analysis) to observe the effect of temperature variation and identify the range of temperature where crystalline nature of the powder was noted. a powder sample was prepared from the gel after annealing at 1000ºc followed by soaking for 4 h, 5 h, and 6 h to compare the variation of particle size with respect to time. the calcined powders were characterized by xrd (x-ray powder diffraction) to determine the phases and crystal planes present in the sample. ft-ir (fourier-transform infrared spectroscopy) was used to study the types of metal oxide or metal-metal bond present in the sample along with m-o coordination studies. fesem (field emission scanning electron microscopy) was used to observe the morphological structure of the sample. edax (energy dispersive x-ray analysis) was used to observe the percentage of each element present in the sample. bulk densities were estimated from 2.4156 g/cm3 to 2.8571 g/cm3 and the rapid increase in apparent porosity of samples 7.4289%, 10.3630% and 32.51% for 4 hours 5 hours and 6 hours respectively were also noted. the average crystal size of spinel particles was noted to be about 48nm, 36 nm and 47 nm respectively. it had been observed that the average crystal size of spinel particles was about 48 nm, 36 nm and 47 nm respectively. finally, the hardness of spinel was evaluated by vicker hardness test and evaluated to be10.52 gpa (1073 hv), 4.087 gpa (416.7hv) and 5.079 gpa (517.9hv). تقييم خصائص ألياف الألمنيوم المغنيسي غير المتكافئ باستخدام مادة الثيوريا كوقود عن طريق تغيير وقت النقع ساثي بنجري, سوميا مخرجي*, و سريناث رانجان جوش الملخص: تناقش هذه الورقة العلمية عملية بسيطة وفعالة من حيث التكلفة لإعداد الاسبنيل mgal2o4 البلوري باستخدام كمية غير متكافئة من نترات المغنيسيوم ونترات الألومنيوم بطريقة احتراق المحلول. واستخدمت الثيوريا كوقود وعامل خفض في حين نقعت في درجة حرارة 1000سيليزية مع فترات مختلفة من النقع. بعد التجفيف البطيء للمحاليل المختلطة في درجة حرارة 80 سيليزية لمدة 4-5 ساعة ، تشكل هلام بواسطة dta/tga (التحليل الحراري التفاضلي والتحليل الحراري لقياس الحرارة) لمراقبة تأثير التباين في درجات الحرارة وتحديد مدى درجة الحرارة حيث لوحظ الطبيعة البلورية للمسحوق. وأعدت عينة من المسحوق من الهلام بعد تلدينها في درجة حرارة 1000 سيليزية ثم نقعها لمدة 4 و 5 و 6 ساعات لمقارنة اختلاف حجم الجسيمات بمرور الزمن. وقد تميزت المساحيق المحسوبة بإشعاع xrd (حيود مسحوق الأشعة السينية). لتحديد المراحل والطائرات البلورية الموجودة في العينة ، ft-ir (جهاز "فورييه" لتحويل طيف الأشعة تحت الحمراء) لدراسة أنواع أكسيد المعادن أو رابطة المعادن الموجودة في العينة بالإضافة الى دراسة رابطةm-o و fesem (الفحص المجهري للإلكترون بمسح الانبعاثات الميدانية) لمراقبة البنية المورفولوجية للعينة و edax (تحليل الأشعة السينية المشتتة للطاقة) لمراقبة النسبة المئوية لكل عنصر موجود في العينة. وقُدِّرت الكثافات السائبة من 2.4156 غرام لكل سنتي ميتر مكعب إلى 2.8571 غرام لكل سنتي ميتر مكعب، ولوحظ أيضاً الزيادة السريعة في المسامية الظاهرة للعينات 7.4289 و10.3630 في المائة و32.51 في المائة لمدة 4 و 5 و6 ساعات على التوالي. وقد لوحظ أن متوسط الحجم البلوري لجسيمات الاسبنيل حوالي 48 و 36 و 47 نانومتر على التوالي. كما لوحظ أن متوسط الحجم البلوري لجسيمات الاسبنيل حوالي 48 و 36 و 47 نانومتر على التوالي. وأخيرا ، تم تقييم مدى صلابة الاسبنيل من خلال اختبار فيكر هاردنج ، وتم تقييمه إلى 10.52 gpa (1073 hv) ، 4.087gpa (416.7 hv) ، و 5.079 gpa (517.9 hv). الكلمات المفتاحية: يوريا؛ اسبنيل؛ مغنيسيوم؛ روابط كيميائية؛ تحليل الطور. 1. introduction spinel magnesium aluminate synthesis from industrial waste is a simple cost-effective, and energy-efficient environment-friendly process. magnesium aluminate spinel offers a unique combination of high-temperature properties including high melting point, excellent resistance against chemical attack, potentially high strength even at high temperatures and very good thermal characteristics. the material possesses high strength at both normal and high temperature with a stable phase of magnesium aluminate (mgal2o4) without any phase transition till it reaches melting point (2135ºc) (s.r ghosh et al. 2018, soumya mukherjee 2020). high-temperature materials from mgo-al2o3 system are mainly focused on applications in the steel, cement and glass industries. pv marakar kutty et al. 2013 studied soft chemical route synthesis of magnesium aluminate spinel using oxalic acid as an organic template and nitric acid as an oxidizing agent. variable synthesis conditions cause a difference in particle sizes, chemical homogeneity, reactivity and microstructural features due to change in process parameters. (saberi ali et al. 2008) the stoichiometric spinel is applicable for transparent ceramics and possesses excellent optical properties. the combination of such an enriched approach of transparency plus excellent optical properties gives stoichiometric spinel a promising candidature for the development of transparent armour, visible-infrared window material. other important applications are catalyst support, membrane material, dye absorbent for chemical sensing, ceramic paints meant for high-temperature protection, which could be possible by engineering purity of compound, particle size, surface area and porosity size and its distribution within a material matrix. the use of mgal2o4 has shown superiority as armour materials with respect to sapphire, alon, soda-lime silicate glass, and mgf2 in terms of its excellent performance and affordability (narges habibi et al. 2017). the material in stoichiometric form is a promising material for humidification sensors, photocatalyst as well to decompose organic pollutant like red methylene discharged from textile industries/industrial operations. (mostafa y nassar et al. 2014) stoichiometric spinel magnesium aluminate is processed by various methods like: sol-gel (debsikdar j.c 1985; naskar m.k. et al. 2005) precipitation (li j-g et al. 2000), aerosol method (yang n et al. 1992), co-precipitation (guo j et al. 2004), combustion synthesis, freeze-drying, decomposition of an organometallic complex in supercritical fluids, hydrothermal route, plasma spray decomposition of powders, (bickmore r clint et al. 1996; bratton r.j. 1969; barj m. et al. 1992; pommier c. et al. 1990; yang ning et al. 1992; varnier olivier et al. 1994) microwave-assisted combustion route (torkian leila et al. 2011), polymerized complex method (lee p.y et al. 2006; du xuelian et al. 2014), mechanochemical route (domanski d. et al. 2004), self-propagating high-temperature synthesis (gorshov v. a. et al. 2017) and others. sol-gel technique is extensively used among all chemical routes because it has the advantage of producing pure, ultrafine powders at low temperatures with high surface area and pore size distribution (ewais e.m.m et al. 2015). among the wet chemical routes, the solution combustion technique has been regarded as one of the effective and economic methods due to its convenient processing, simple experimental setup and significant time-saving and high purity products (patil k.c. et al., 2002 lazau i et al., 2008). some recent articles also focus on spinel for photocatalytic activity under solar irradiation for the treatment of methylene blue discharged by textile and chemical industries (salem shiva et al., 2020), the degradation mechanism of periclase-spinel refractory bricks used in the upper transition zones of a cement rotary kiln (zhou wenying et al., 2021). the present investigation is undertaken to develop non-stoichiometric magnesium aluminate spinel powders from sol-gel magnesium and aluminium nitrate using thiourea as fuel. the majority of the study is carried for stoichiometric spinel while limited research is available for the non-stoichiometric case. the study has been carried to observe the effect of different soaking time temperatures on the densification and properties of mgal2o4 spinel. microhardness of spinel also has been studied after phase analysis, bonding and morphological characterization of a synthesized sample. 2. materials and methods mixed solutions of non-stoichiometric magnesium and aluminium nitrate with a ratio of approximately 1:1 were prepared by dissolving the corresponding amount of magnesium and aluminium nitrate in 80ºc distilled water. thiourea, as a fuel and reducing agent, is added to the mixture. after mixing the corresponding nitrates, these solutions were kept for 3-4 h to evaporate excess moisture. the gelatinous precipitate was cooled to room temperature. the gel was calcined at 1000ºc at a rate of 5ºc/min for a different soaking time of 4, 5, 6 h respectively and kept inside the heating chamber at the furnace for cooling. the transition to crystallization was noted in the dsc-tga analyzer (model no.: pyris diamond tg/dta make: perkinelmer (singapore) in nitrogen atmosphere (150ml/min). phase analysis was carried by xrd (riaku, ultima iii, cu kα, 1.54åλ) having scan range 10-80° with a scan rate of 5°c/min. bonding analysis was confirmed from ftir spectra (shimadzu, ir prestige-21) using kbr for pellet formation. morphological features were studied by sem followed by edx (jeol, jax 840a) analysis to observe the elements and composition of the sample. microhardness was evaluated using vickers hardness testing machine (model vm, capacity 50 kgf, make fie). 3. results and discussion differential thermal analysis / thermal gravimetric analysis (dta/tga) of the gel was performed using a dta-tga analyzer (model no.: pyris diamond tg/dta make: perkinelmer (singapore) in nitrogen atmosphere (150ml/min). nitrogen purging was carried to maintain a neutral atmosphere for reaction and crystallization and to note transitions from amorphous to crystalline one close to theoretical results. a platinum crucible was used with alpha alumina powder as a reference concerning the sample. due to the low reactivity of platinum, there was a negligible chance of reaction of the crucible with the powder. in the dta/tga curve, there were two endothermic peaks that suggest physically absorbed water loss and it was responsible for the first weight loss from about 80°c to 120°c regions. then with 2nd weight loss at 240°c, the exothermic peak suggests oxidation of organic compounds which consists of carbonaceous and sulphur elements. the formation of γ-alumina was suggested by an endothermic peak at 320°c. fernandes macedo et al. noted such formation of γ-alumina at about 350 during the kinetic study of γ-α alumina transitions. a similar observation was noted by (maria macêdo et al. 2007). a minor endothermic peak at about onset of 400°c was possibly due to decomposition of nitrate precursor salts accompanied with minor weight change. the small exothermic peak at 500°c suggests the decomposition of nitrates. from there on the crystallization process starts and that’s why there was a minor endothermic peak till 590°c and satisfies the third weight loss in the tg curve. the above findings were noted by the same author while preparing stoichiometric spinel sample using thiourea, urea and citric acid as fuels (soumya mukherjee, 2020 s.r. ghosh et al. 2018). figure 2 shows the xrd graph of the calcined sample prepared at 1000°c. from the xrd plot, we can observe that synthesized powder has almost the same “d” (inter planner spacing) value as compared to standard data from jcpds (card number #01-075-1800). major growths of peaks were noted along with those crystallographic directions which leads to the thermodynamic stability of the compound formed. from the experiment major planes indexed were noted mostly along (111), (220), (311), (400), (511) and (440) and those planes have corresponding 2θ values of 20.20, 30.85, 36.90, 42.95, 59.40 and 65.20 degrees respectively. this observation indicates that the above mentioned sol-gel process was an excellent efficient and economic process for spinel synthesis. using sherrer’s equation we have calculated the crystallite size in the range of 36 nm to 48 nm. crystallite size was about 48 nm, 36 nm, and 47 nm after 4 h, 5h and 6h of soaking period respectively. deconvolution on prominent peaks was not noted which may indicate phase separation or other phase formation. none of the annealed samples exhibits the presence of intermediate oxides or precursors. peak broadening was not noted for none, which indicates the presence of crystalline peaks instead of amorphous ones. xrd study confirms successful annealing at requisite conditions for all cases forming crystalline spinel magnesium aluminate. so, we can say that almost all peaks correspond to the magnesium aluminate spinel which was a good sign of getting the new economical route of synthesis of spinel. figure 3 shows the ftir spectra of the samples after calcination at a different soaking time are. after calcination at 1000ºc thiourea disappeared and removed completely. scanning of spectral peaks was carried within 4500-450 cm-1 while the major spectral peak for compound formation was within in and around 1000 cm-1. for spinel magnesium figure 1. tga/dta analysis of the prepared gel samples prepared in a nitrogen atmosphere with a heating rate of 5°c/min till 900°c. figure 2. xrd analysis of the spinel powder prepared and calcined at 1000 °c for 6 hours, 5 hours and 4 hours respectively with 1:1 mg nitrate, al nitrate precursors. aluminate synthesized using urea as fuel and reducing agent, octahedral m-o (al-o) stretching was noted at about 539 cm-1 while mg-o-al vibration is observed at about 677 cm-1 . the majority of m-o coordination was found within 1000 cm-1 as noted from the present experiment. similar to the present experimental observations majority of m-o coordination were found within 1000 cm-1. the above bonding analysis noted from ftir spectra were noted to be in correspondence with research findings by the same author in other research articles (s.r. ghosh et al. 2018; mukherjee 2020). the band at 524 cm-1, 671 cm-1, and 1111 cm-1 shows the stretching vibration of al-o-mg. it was also seen from the graph that 1485 cm-1, 1651 cm-1 , 3344 cm-1 reflects h2o stretching vibration. if we increase the soaking time, it was found that the peak at 1485 cm-1 got eliminated from the graph which indicates that at a higher temperature the concentration of h2o was less. the presence of adsorbed moisture was noted since ftir analysis was carried in the air atmosphere. if the analysis has been carried in a controlled atmosphere with neutral gas purging (nitrogen/argon) presence of adsorbed moisture on the powder surface have been eliminated. figures 4, 5, and 6 show the edx analysis of the mgal2o4 spinel nanoparticle synthesized by the sol-gel method at 1000°c and soaked for 4,5,6 h respectively. it was deduced from morphology that the grain size of the ma particle increases with increasing temperature with no change in phase indicating stabilization of crystal structure. the edx spectra of the spinel powder calcined at 1000°c temperatures for different soaking times were taken at different positions. it is indicated that the cubic grain structures of spinel with a ratio of 1:1 magnesium, al precursors added were studied to observe the required elements present in the matrix as shown in morphology. edx of different soaking time has been shown in figs. 3, 4 and 5 respectively. all edx spectra indicate the presence of mg, al, o elemental peaks in the sample. the intense stacking of the powder grains on each other was dominant in the figure where edx spot analysis has been carried. xrd, ftir and edx analyses correlate with each other indicating successful phase formation of spinel using non-stoichiometric precursors. figure 7 shows the sem of the ma spinel calcined at different soaking periods at a fixed temperature. the surface morphology of the powders magnesium aluminate spinel was investigated by scanning electron microscopy (sem). the morphological structure of the mgal2o4 powder calcined at 1000ºc temperatures for different soaking time 4 h, 5 h, 6 h were exhibited by sem. it can be seen from the figure that the spinel had become much more agglomerated in 5 h and 4 h soaking period in comparison to 6 h soaking. minute observation reveals that some minor intergranular cracks were noted among agglomerated chunk for sample undergoing 5 h soaking while nil figure 3. ft-ir spectra of mgal2o4 product calcined at 1000°c prepared after annealing for 4 hours, 5 hours and 6 hours respectively with 1:1 non-stoichiometric ratio of mg-al nitrate precursors using thiourea as fuel. (a) (b) (c) figure 4. edx analysis of the mgal2o4 spinel nanoparticle synthesized by the sol-gel method at 1000°c for 4 hours. figure 5. edx analysis of the mgal2o4 spinel nanoparticle synthesized by the sol-gel method at 1000°c for 5 hours. figure 6. edx analysis of the mgal2o4 spinel nanoparticle synthesized by the sol-gel method at 1000°c for 6 hours. figure 7. sem morphology of magnesium aluminate (ma) spinel calcined at the different soaking period of 4hours (a), 5 hours (b) and 6 hours (c) for the fixed temperature at 1000°c using 1:1 non-stoichiometric ratio of mg:al nitrates using thiourea as fuel figure 8. variation of bulk density (bd), material weight and soak time of spinel grains by sintering at 1000°c function of the compact weight. figure 9. variation of apparent porosity (ap), and soak time of spinel grains on sintering at 1000°c at the function of the material weight. table 1. dry and wet bulk density, apparent porosity, material weight of spinel soaked for different periods at fixed temperaturetable 1: dry and wet butable 1: dry and wet bulk density, apparent porosity, material weight of spinel soaked for different periods at fixed temperature. hrs dry bulk density (g/cm3) wet bulk density (g/cm3) apparent porosity material weight (g) 4 2.8571 2.9325 0.0742 0.97 5 2.7609 2.8646 0.10363 0.73 6 2.4156 2.7407 0.3251 0.78 cracks or interconnected porosity was noted for sample undergoing annealing for 4 h. nevertheless the sample prepared after soaking for 6 h exhibits less agglomeration. from this observation, we can conclude that soaking time at around 5 h gives us the desired spinel structure. the above morphological features were found to be tallying with density and apparent porosity variation noted for spinel samples heated at 1000°c for 4 h, 5h and 6 h soaking period. such possible result may be due to the difference of thermal expansion behaviour of al-based precursors and magnesium-based precursors and for maintaining non-stoichiometric ratio in comparison to stoichiometric ratio used for spinel formation. variation of bulk density (bd), material weight and soak time of spinel grains by sintering at 1000°c as a function of the compact weight was depicted in fig. 8. densities were calculated using the archimedes principle. here, the densities were changed from 2.4156 g/cm3 to 2.8571 g/cm3 with sintering temperature 1000°c. according to the literature [5], the values of the bulk densities were obtained around 3 g/cm3 for the mgal2o4 spinel. the density of the sample is maximum for 4 hours sample while least for 6 hours soaking period. the density variation commensurate with the morphological findings from sem analysis. figure 9 shows the apparent porosity (%) values of sintered samples soaked at different temperatures. the apparent porosities of samples were decreased with increasing sintering temperature. these results have shown that the samples were sintered better with increasing temperature. sinhamahapatra et al. [6] have produced the magnesia-rich magnesium aluminate spinel with the natural magnesite and the synthetic caustic magnesia. however, spinel and periclase and forsterite were found in the samples due to the presence of silica as impurities. in our study, magnesium aluminate spinel was formed at 1000°c. the decrease in apparent porosity of samples from 4 h to 6 h indicates that the densification occurred through liquid phase sintering (fig. 6). apparent porosity and bulk density were two vital parameters to assess the extent of densification of ceramic compacts. the variation of apparent porosity and bulk density of spinel samples with sintering temperatures were shown in figs. 7 and 8. it reveals that for ss composition highest bulk density of 2.8571 was achieved at1000°c with a soaking time of 4 h and a slight decrease in bulk density value of 2.7609 and 2.4156 at 5 h and 6 h soaking respectively was observed. the rapid increase in apparent porosity of samples 7.4289%, 10.3630% and 32.51% for 4h, 5 h and 6 h respectively also indicates that high soaking time affects apparent porosity. it was also noted that variation of apparent porosity with soaking time was in correspondence with morphological findings from sem analysis. finally, the hardness of pellet samples of spinel was carried by vicker’s hardness testing. at first, the powder material is properly mixed with 1.5% pva (polyvinyl alcohol). the mixture is then dried and ultrasonicated to loosen its agglomerated particles. then the mixture is pressed with 4-ton pressure to make a pellet. after that, the pellet is sintered at 1200°c for 30 minutes to remove the pva present inside the pellet. at high temperature, various particles present in the material rearranged their bonding which made the material denser. corresponding hardness values were evaluated as 10.52 gpa (1073 hv), 4.087 gpa (416.7 hv) and 5.079 gpa (517.9 hv). sample synthesized after soaking at 4 h show more hardness than the powder soaked at 5 h and 6 h. this will suggest that crystal size will increase as well as hardness will decrease if we increase the soaking time. 4. conclusion non stoichiometric1:1 ratio of magnesium nitrate, al nitrate along with thiourea as fuel and the reducing agent was utilized to synthesis spinel magnesium aluminate. crystallization for spinel onset was in the range 500-700°c with a minor endothermic peak at 590°c. proper phase for spinel was developed even with non-stoichiometric precursors at 1000°c after soaking for 4 h, 5 h and 6 has confirmed from xrd. the crystallite size of corresponding spinel particles was about 48 nm, 36 nm and 47 nm respectively. ftir analysis confirmed the m-o coordinations and it was noted that al-o coordination was about 539 cm-1 while mg-o-al vibration was observed at about 677 cm-1. morphology analysis by sem reveals that spinel becomes more agglomerated in 5 h and 4 h soaking period in comparison to 6 h soaking. minute analysis reveals minor intergranular cracks for samples undergoing 5 h soaking. edx confirms the elemental composition analysis. bulk densities were estimated from 2.4156 g/cm3 to 2.8571 g/cm3. the rapid increase in apparent porosity of samples were 7.4289%, 10.3630% and 32.51% for 4 h 5 h and 6 h, respectively. bulk density and apparent porosity results were found to be in tallying with microstructural features obtained after synthesis using non-stoichiometric ratio and thiourea as fuel. hardness values were evaluated as 10.52 gpa (1073 hv), 4.087 gpa (416.7 hv) and 5.079 gpa (517.9 hv) in the form of a pellet sample by vicker hardness tester. conflict of interest the authors declare no conflict of interest. funding no funding was received for this study. acknowledgment the authors would like to thank the department of metallurgical & material engineering for providing technical support to carry out dsc-tga analysis, xrd, ftir and morphology studies by sem. references barj m, bocquet j.f, chhor k, pommier c, (1992), submicronic mgal2o4 powder synthesis in supercritical ethanol. journal of materials science 27: 2187-2192. bickmore r clint, waldner f. kurt, treadwell r. david, laine m. richard, (1996), ultrafine spinel powders by flame pyrolysis of a magnesium aluminium double hydroxide. journal of the american ceramic society 79(5): 1419-1423. bratton r.j, (1969), coprecipitates yielding mgal2o4 powders. american ceramic society bulletin 83: 759-762. domanski daniela, urretavizcaya guillermina, castro j. facundo, gennari c. fabiana, (2004), mechanochemical synthesis of magnesium aluminate spinel powder at room temperature. journal of the american ceramic society 87(11): 2020-2024. du xuelian, liu yaqiang, li liqiang, chen wencong, (2014), synthesis of mgal2o4 spinel nanoparticles via polymer-gel and isolation –medium-assisted calcination. journal of materials research 29(24): 2921-2927. ghosh chandrima, ghosh arup, halder manas kamal (2015), studies on densification, mechanical, micro-structural and structure properties relationship of magnesium aluminate spinel refractory aggregates prepared from indian magnesite. materials characterization 99: 84-91. ghosh s.r, mukherjee s, banerjee s (2018), solution combustion synthesis of alumina spinel and its characterization. interceram 67: 34-41. gorshkov v.a, miloserdov p.a, yukhvid v. i, sachkova n.v, kovalev i.d, (2017), preparation of magnesium aluminate spinel by self-propagating high temperature synthesis metallurgy methods. inorganic materials 53: 1046-1052. habibi narges, wang yuan, arandiyan hamidreza, rezaei mehran (2017), low temperature synthesis of mesoporous nanocrystalline magnesium aluminate (mgal2o4) spinel with high surface area using a novel modified sol-gel method. advanced powder technology 28: 1249-1257. lazau ioan, ianos robert, pacurariu cornelia, savii cecilia, (2008), spinel preparation by combustion synthesis: the influence of synthesis conditions on the features of the resulted spinel mgal2o4 powder. romanian journal of materials 3: 224-232. lee p.y, suematsu h, yano t, yatsui k, (2006), synthesis and characterization of nanocrystalline mgal2o4 spinel by polymerized complex method. journal of nanoparticle research 8: 911-917. macêdo maria laponeide fernandes, bertran celso aparecido, osawa carla cristiane, (2007), kinetics of the γ → α-alumina phase transformation by quantitative x-ray diffraction.  journal of materials science 42: 2830–2836. https://doi.org/10.1007/s10853-006-1364-1 marakkar kutty p v, dasgupta subrata (2013), low temperature synthesis of nanocrystalline magnesium aluminate spinel by a soft chemical method. ceramics international 39(7): 7891-7894. mukherjee soumya, (2020), evolution of spinel magnesium aluminate by combustion route using glycine as fuel and its characterization. international journal of emerging trends in science and technology 6(1): 24-27. nassar y. mostafa, ahmed s. ibrahim, samir ihab (2014), a novel synthetic route for magnesium aluminate (mgal2o4) nanoparticles using sol-gel auto combustion method and their photocatalytic properties. spectrochimica acta part a: molecular and biomolecular spectroscopy 131: 329-334. păcurariu c, lazău i, ecsedi z, lazău r, barvinschi p, mărginean g, (2007), new synthesis methods of mgal2o4 spinel. journal of the european ceramic society 27: 707-710. patil kashinath c, aruna s. t, mimani tanu, (2002), combustion synthesis: an update. current opinion in solid state and materials science 6: 507-512. pommier c, chhor k, bocquet j.f, barj m, (1990), reactions in supercritical fluids, a new route for oxide ceramic powder elaboration, synthesis of the spinel mgal2o4. materials research bulletin 25(2): 213-221. saberi ali, golestani-fard farhad, sarpoolaky hosein, willert porada monika, gerdes thorsten, simon reinhard (2008), chemical synthesis of nanocrystalline magnesium aluminate spinel via nitrate-citrate combustion route. journal of alloys and compounds 462: 142-146. salem shiva (2015) application of autoignition technique for synthesis of magnesium aluminate in nano scale: influence of starting solution ph on physico-chemical characteristics of particles. materials chemistry and physics 155(1): 59-66. salem shiva, nouri babak, ghadiri mohammad, (2020), photoactivity of magnesium aluminate under solar irradiation for treatment of wastewater contaminated by methylene blue: effect of self-combustion factors on spinel characteristics. solar energy materials and solar cells 218: 110773. torkian leila, amini m mostafa, bahrami zohreh, (2011), synthesis of nano crystalline mgal2o4 powder by microwave assisted combustion. journal of inorganic materials 26(5): 550-554. varnier olivier, hovnanian nadine, larbot andré, bergez pierre, cot louis, charpin jean, (1994), sol-gel synthesis of magnesium aluminum spinel from a heterometallic alkoxide. materials research bulletin 29(5): 479-488. yang ning, chang li, (1992), structural inhomogenity and crystallization behavior of aerosol reacted mgal2o4 powders. materials letters 15(1-2): 84-88. zhou wenying, yan wen, ma sanbao, schafföner stefan, dai yajie, li yawei, (2021), degradation mechanisms of periclase-magnesium aluminate spinel refractory bricks used in the upper transition zone of a cement rotary kiln. construction and building materials 272: february 2021: 121617. a b c a b c a b c *corresponding author’s e-mail: smmukherjee3@gmail.com doi: 10.53540/tjer.vol18iss1pp44-51 46 45 the journal of engineering research (tjer), vol. 18, no. 1, (2021) 12-19 *corresponding author’s e-mail: sherif@squ.edu.om doi: 10.53540/tjer.vol18iss1pp12-19 flexural behavior of frp bars after exposure to elevated temperatures sherif e. el gamal1, 2 *, abdulrahman m. al-fahdi1, mohammed meddah1, abdullah al-saidy1, and kazi md abu sohel1 1civil and architectural engineering, sultan qaboos university, muscat, oman 2menoufia university, shebin el-kom, egypt abstract: this study investigates the flexural behavior of fibre-reinforced polymer (frp) bars after being subjected to different levels of elevated temperatures (100, 200 and 300°c). three types of glass frp bars (ribbed, sand coated, and helically wrapped) and one type of carbon frp bars (sand coated) were used in this study. two testing scenarios were used: (a) testing specimens immediately after heating and (b) keeping specimens to cool down to room temperature before testing. test results showed that as the temperature increased the flexural strength and modulus of the tested frp bars decreased. at temperatures higher than the glass transition temperature (tg), significant flexural strength and modulus losses were recorded. smaller diameter bars showed better residual flexural strength and modulus than the larger diameter bars. the immediately tested bars showed significant strength and modulus losses compared to bars tested after cooling. different types of gfrp bars showed comparable results. however, the helically wrapped bars showed the highest flexural strength losses (37 and 60%) while the sand coated bars showed the lowest losses (29 and 39%) after exposure to 200 and 300℃, respectively. the carbon frp bars showed residual flexural strengths comparable to those recorded for the gfrp bars; however, they showed lower residual flexural modulus after being subjected to 200 and 300℃. keywords: fibre-reinforced polymer bars; frp; glass frp; carbon frp; flexural strength; flexural modulus; elevated temperatures; bar diameters. سلوك اإلنثناء لقضبان البولیمر المقوى باأللیاف بعد تعرضھا لدرجات حرارة مرتفعة شریف الجمل*, عبدالرحمن الفھدي, محمد مداح, عبدهللا السعیدي, و كازي ابو سھیل بعد تعرضھا لمستویات مختلفة من درجات تبحث ھذه الدراسة في سلوك االنثناء لقضبان البولیمر المقوى باأللیاف :الملخص البولیمر المقوى من قضباندرجة مئویة). استخدمت في ھذه الدراسة ثالثة أنواع 300، 200، 100الحرارة المرتفعة ( المقوى باأللیاف الكربونیة البولیمر الزجاجیة (مضلع ، مغلف بالرمل ، وملفوف حلزونیًا) ونوع واحد من قضبان باأللیاف (مغلفة بالرمل). تم استخدام سیناریوھین لإلختبار: أ) اختبار العینات مباشرة بعد التسخین و ب) االحتفاظ بالعینات لتبرد إلى جة حرارة الغرفة قبل اإلختبار. أظھرت نتائج اإلختبار أنھ كلما زادت درجة الحرارة انخفضت قوة و جساءة القضبان ضد در ، تم تسجیل ھبوط كبیر في القوة والجساءة ضد االنثناء. )gt (االنثناء. عند درجات حرارة أعلى من درجة حرارة التزجج و جساءة إنثناء أفضل من القضبان ذات األقطار الكبیرة. أظھرت القضبان التي أظھرت القضبان ذات األقطار الصغیرة قوة تم اختبارھا على الفور نقص كبیر في القوة والجساءة مقارنة بالقضبان التي تم اختبارھا بعد انخفاض حرارتھا. أظھرت مع ذلك ، أظھرت القضبان الملفوفة حلزونیًا نتائج متقاربة. و الزجاجیة البولیمر المقوى باأللیاف مختلفة من قضبانالنواع األ ٪) بعد التعرض 39و 29٪) بینما أظھرت القضبان المغطاة بالرمل أقل خسائر (60و 37أعلى خسائر في مقاومة االنثناء ( المسجلة الكربونیة قوة انثناء مماثلة لتلك على التوالي. أظھرت قضبان البولیمر المقوى باأللیاف درجة مئویة300و 200إلى 300و 200؛ ومع ذلك ، فقد أظھرت معامل جساءة منخفض بعد تعرضھا لـ الزجاجیة البولیمر المقوى باأللیاف لقضبان درجة مئویة. .قطار القضبانأ ؛ درجات الحرارة المرتفعة؛ قوة اإلنثناء ؛األلیاف الكربونیة ؛الزجاجیة األلیاف ؛قضبان البولیمر المقوى باأللیاف :تاحیةالمف لماتالك flexural behavior of frp bars after exposure to elevated temperatures 13 1. introduction in recent years, fiber reinforced polymers (frp) have become one of the promising reinforcing materials in concrete structures. frp reinforcing bars can be used instead of steel bars in reinforced concrete (rc) structures because of their excellent properties, which include corrosion resistance, high strength-to-weight ratio, appropriate fatigue performance, electric insulation and easy cutting and handling (el-gamal et al. 2014). these advantages increased their use to repair existing structures (al-saidy et al. 2015, 2017; el-gamal et al. 2019). they also increased their use instead of steel bars in several new reinforced concrete (rc) structures in corrosive environments such as bridge deck slabs, parking garage slabs, rc pavements, and rc columns (benmokrane et al. 2007, 2008; el-gamal et al. 2009; thébeau et al. 2010; elgamal et al. 2010; bouguerra et al, 2011; dulude et al. 2011; el-gamal and alshareedah 2020 a, b). on the other hand, they have some disadvantages compared to steel including no ductility, lower elastic modulus and shear strength. in addition, they are very sensitive to elevated temperatures (ashrafi et al. 2017; bazli and abolfazli 2020; jafarzadeh and nematzadeh 2020) the wide use of frp bars in concrete structures led to new challenges for engineers. one of these challenges is the performance of frp bars when subjected to elevated temperatures. frp bars performance is poor under elevated temperatures. as the temperature increases, the matrix softens at the glass transition temperature level (tg). above tg, the matrix elastic modulus significantly reduces due to a change in its molecular structure. the thermal properties of fibres, however, are better than the matrix and can continue sustaining the load in a longitudinal direction until reaching the temperature level of the fibres. engineers when designing a structure must consider the time that the structure can withstand high temperatures. therefore, it is important to investigate the behavior of frp bars subjected to elevated temperatures. robert and benmokrane (2010) investigated the tensile strength, shear strength and flexural strength of one type of glass frp (gfrp) bar subjected to elevated temperatures. the elevated temperature levels ranged between 100 and 325°c. the results showed that the shear and the flexural strengths were much more sensitive to high temperature than the tensile strength. the flexural strength of the gfrp bars increased when the temperature decreased. at 120°c, the mechanical strength and flexural modulus dropped because of the change of the state of the polymer. alsayed et al. (2012) investigated the tensile properties of gfrp bars subjected to elevated temperatures for different periods. the study included two groups of specimens. the first group was 60 specimens without concrete cover (bare bars) with 1 m length. while the second group consisted of 60 specimens covered with concrete to represent the actual case when using the gfrp bars in concrete. the elevated temperatures used were 100, 200 and 300°c. the bars were tested in tension after exposure to the elevated temperatures. the authors concluded that the tensile strength decreased as the level of temperature or exposure period increased. the losses of tensile strength ranged between 3.1 to 35.1% and 9.7 to 41.9% for concrete-covered and bare gfrp bars, respectively. maranan et al. (2014) studied the flexural behavior of sand-coated gfrp bars of different diameters subjected to elevated temperatures up to 150°c. as the temperature increased, the flexural strength and modulus of the gfrp bars decreased. when approaches the tg of the bars, the polymer changed from glassy (hard) material to rubbery (soft) material. the polymer begins to lose the ability to hold the fibre together and transfer stresses from one fibre to another. for all temperatures from 21 to 80°c, the load increased linearly with deflection up to failure. in addition, temperatures from 100 to 150°c exhibited a non-linear behavior and modulus degradation before reaching the maximum load because the tg of the bar was about 117°c. the research recommended additional studies to provide further information that can be used to establish a relationship that can predict the tensile response of the gfrp bars from the bending response at elevated temperatures. ashrafi et al. (2017) investigated the physical and thermal properties of frp bars subjected to elevated temperatures. the test specimens used were four types of frp bars; sand-coated gfrp, helically wrapped cfrp, and grooved cfrp bars with two different resins (epoxy and vinyl ester). the research concluded that, as the temperature increased the tensile strength of the frp bars decreased. also, the bigger the diameter the greater tensile strength. the results of cfrp bars showed linear deformation until reaching the ultimate load. the critical temperatures of the sandcoated gfrp bars when losing 50% of its tensile strength were 300, 375, 377 and 450°c for 4, 6, 8, and 10 mm respectively. the critical temperatures of the three types of cfrp bars tested were 330, 360 and 450°c, respectively. hamad et al. (2017) studied the effect of elevated temperatures on the mechanical properties of frp bars and the bond behavior between frp bars and concrete. the test specimens used were four types; sand-coated cfrp, helically wrapped basalt frp (bfrp), helically wrapped gfrp, and steel bars. the elevated temperatures used ranged between 125 to 450°c. after exposure to elevated temperatures, the specimens were left to cool in the air after taken from the electrical furnace. the tensile strength reduction of frp bars was almost linear up to a critical temperature of 325°c. at a temperature of 450°c, the gfrp and bfrp bars melted and totally lost their tensile strength capacity. steel bars had a minor reduction in tensile strength compared with the frp bars. among the frp bars, the sherif e. el gamal, abdulrahman m. al-fahdi, mohammed meddah, abdullah al-saidy, and kazi abu sohel 14 cfrp showed the highest bond strength with concrete because of their better surface characteristics. however, steel bars attained the highest bond strength and modulus. the study recommended additional experimental work using other types of commercially available frp bars. the literature shows that few studies investigated the flexural behavior of the frp bars subjected to elevated temperatures and recommended conducting more research on this topic. in addition, these studies are limited to specific types and diameters of the frp bars. therefore, to enrich the knowledge about this topic, it was decided to conduct this research study to investigate the flexural behavior of different types and diameters of frp bars exposed to elevated temperatures. test parameters include the four types of frp bars (sand-coated gfrp, helically wrapped gfrp, grooved gfrp, and sand-coated cfrp), three diameters of the grooved gfrp bars (10, 16 and 20 mm), three target temperature levels (100, 200 and 300°c), and three exposure periods (1, 2 and 3 hours). in addition, two testing scenarios were investigated: a) testing specimens immediately after heating; and b) keeping specimens to cool down to room temperature before testing. 2. experimental work 2.1. frp bars three different types of gfrp bars and one type of cfrp bars were used in this study. the gfrp bars included; a) grooved (g1) with nominal diameters of 10, 16 and 20 mm; b) sand-coated (g2) with a nominal diameter of 10 mm, and c) helically wrapped (g3) with a nominal diameter of 10 mm. while sand-coated cfrp (c1) bar with a nominal diameter of 10 mm was used. table 1 presents the list of frp types, manufacturers and their properties. figure 1 shows photos of the bars used in this investigation. 2.2. test specimens and parameters frp bars were prepared for flexure tests. the length of all specimens was 240 mm. the total number of specimens tested in this study was 99 specimens. three specimens of each diameter were tested at room temperature as control specimens. the remaining specimens were exposed to temperatures of 100, 200, and 300°c for one hour. an electric furnace was used to heat up the specimens. the temperature was increased at a rate of 3°c degrees per minute until reaching the required temperature then was kept constant at the required temperature level. all the g1 specimens were tested using two testing scenarios. half of them were tested immediately after taken out from the furnace. the other half was kept in the lab for one hour to cool down before testing. all other types of frp bars were tested immediately after taken out from the furnace. 2.3. flexural test procedure all specimens were tested under a three-point bending test set-up. the flexural tests were conducted in accordance with the astm d790 standard (astm d790-03, 2003). the clear span between supports was 180 mm and the overhangs were 30 mm. the tests were carried out using a 600 kn mts machine under a loading rate of 4 mm/min. the applied load and the mid-span deflection were recorded using a data logger. figure 2 shows a typical specimen during the testing. (a) grooved gfrp bars (g1). (b) sand-coated gfrp bar (g2). (c) helically wrapped gfrp bar (g3). (d) sand-coated cfrp bar (c1). figure 1. frp bars used in this study. figure 2. frp specimen during flexure testing. table 1. properties of frp bars used in this study. type glass carbon g1a g2b g3c c1b nominal diameter 10 16 20 10 10 10 tensile strength (mpa) 1150 1102 1094 1185 849 1596 tensile modulus (gpa) 60.5 61.2 66.4 52.3 48.8 120 ultimate strain (%) 1.9 1.8 1.65 2.26 1.81 1.33 area (mm2) 66.5 199 284 71.3 71.3 71.3 a made in uae; b made in canada; c made in the usa. flexural behavior of frp bars after exposure to elevated temperatures 15 3. test results and discussions 3.1. summary of results table 2 summarizes the experimental test results of all tested bars. the values of the flexural strength and modulus presented are the average values from three specimens. the flexural strength (fb) and the flexural modulus (eb) were calculated using eqn. (1) and eqn. (2), respectively. more details about these equations can be found in appendix a and appendix b, respectively. 𝑓𝑓𝑏𝑏 = 8𝐹𝐹𝐹𝐹 𝜋𝜋𝑑𝑑𝑏𝑏 3 (1) 𝐸𝐸𝑏𝑏 = (𝐹𝐹50−𝐹𝐹20) (δ50−δ20) 4𝐹𝐹3 3𝜋𝜋𝑑𝑑𝑏𝑏 4 (2) where f, l, and db, are the maximum concentrated load, clear span, and nominal diameter of the bars, respectively; f50 and f20 are the concentrated load at 50 and 20% of the maximum load, respectively. δ50 and δ20 are the mid-span deflection at f50 and f20, respectively. the values of f and δ in eqn. (2) were obtained from the load-deflection curve of each bar. specimens exposed to 100 and 200ºc showed a linear relationship between the load and mid-span deflection until failure for all frp types. some of the specimens exposed to 300ºc show nonlinear behavior at higher loads levels close to failure. 3.2. failure modes of tested bars the general failure mode of most frp bars after testing is defined by crushing of the bars in the compression zone and rupture of fibres in the tension side as shown in fig. 3(a). some of the 20 mm diameter bars showed an interlaminar shear failure at high temperatures (200 and 300℃) as shown in fig. 3(b). it occurs due to high horizontal shear force that affected the weak matrix at high temperatures. 3.3 effect of temperature level and bar diameter (type g1 bars) figures 4(a) and 4(b) represent the residual flexural strength of type g1 bars as a function of temperature level and bar diameter for both immediate and cooling testing scenarios, respectively. figure 4(a) shows that, in the immediate scenario, the residual flexural strength decreased as the temperature level increased. after exposure for one hour to 100℃, the residual flexural strengths ranged between 78 to 84% of the control samples. the residual flexure strength was significantly affected after exposure to 200 and 300℃. at 200℃, the residual flexural strengths ranged between 19 and 67% and then reduced at 300℃ to range between 11 and 48%. this decrease in the residual flexural strength at higher temperature levels could be related to the matrix softening. as the temperature exceeds the tg temperature, the matrix elastic modulus significantly reduces due to a change in its molecular structure and results in lower efficiency in transferring shear forces between fibres, consequently, results in lower flexural strength of the frp bars. fig. 4(a) also shows that, in the immediate scenario, the flexural strength losses at 100℃ was not affected by the bar diameter where the matrix at this temperature level was not significantly affected and the failure was dominant by the fibres. at this temperature level, the residual flexure strengths were 78, 84, and 84% for the 10, 16, and 20 mm diameter bars, respectively. at 200 and 300℃ (higher than the tg), it can be noticed that the residual strength decreased as the bar diameter increased. at 200℃, the residual flexural strengths were 67, 32, and 19% for the 10, 16, and 20 mm diameter bars, respectively. similar observations were recorded at 300℃ where the residual flexural strengths were 48, 31 and 11%, respectively. at these temperature levels, matrices become weak and the shear strength of the matrix dominants the failure. in smaller diameter bars, there are lower horizontal shear stresses in the matrix, which delays the matrix failure and results in higher residual strength. these shear stresses increase as the bar diameters increase and result in earlier failure of the matrix and lower flexural strength. this can be seen from the interlaminar shear failure observed in the 20 mm bars at high temperatures as shown in fig. 3(b). this agrees with previous research studies (robert and benmokrane 2010; maranan et al. 2014). in the cooling scenario (fig. 4(b)), it can be noticed that the residual flexural strength was not significantly affected by increasing the temperature level or bar diameter. the flexural strength losses ranged between 5 and 8%, 4 and 14%, and 3 and 12% for bars exposed to 100, 200 and 300℃ temperatures, respectively. this indicates that allowing the temperature of the specimens to cool down before testing enhanced their strength and resulted in residual flexural strengths close to those of the reference specimens. this shows that the matrix gained strength after cooling and were able to transfer the shear forces between the fibres, which resulted in the high flexural strength values compared to the immediate scenario. figures 5(a) and 5(b) show the residual flexural modulus of all g1 bars for both immediate and cooling specimens, respectively. the modulus results were consistent with the strongest results. it can be noticed that, for the immediate scenario, the modulus decreased as the temperature level increased. the residual flexural modulus was only about 18% at 200 and 300℃ for the 16 and 20 mm bars (losses of about 82%). this again could be related to the weakness of the resin due to high temperatures especially for bigger bars that did not lose heat during immediate testing. for cooling specimens, the used temperature levels have a slight effect on the flexural modulus of the tested bars. for example, the residual flexural moduli of the 10, 16 and 20 mm was 99, 98 and 92% at 200℃ and 100, 98 and 84% at 300℃, respectively. sherif e. el gamal, abdulrahman m. al-fahdi, mohammed meddah, abdullah al-saidy, and kazi abu sohel 16 table 2. flexural strength and modulus for test specimens. types g1 g2 g3 c1 diameter 10 mm 16 mm 20 mm 10 mm 10 mm 10 mm specimen 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) 𝑓𝑓𝑏𝑏 (mpa) eb (gpa) room temperature 1055 46.1 773 45.9 942 54.9 1131 37.7 814 34.1 830 80.0 100-1h-i 823 45.4 652 44.2 793 45.4 1080 37.3 732 33.6 700 79.8 200-1h-i 710 32.9 246 8.4 176 10.1 806 26.1 516 26.2 538 37.9 300-1h-i 512 29.6 242 8.6 108 10.0 689 26.0 326 20.2 441 26.6 100-1h-c 1005 45.0 735 46.0 865 51.8 200-1h-c 909 45.8 738 45.2 849 50.5 300-1h-c 933 46.0 724 44.8 918 45.9 g1 g2 g1 g1 (a) 10 mm diameter (b) 20 mm diameter figure 3. the general typical failure mode of frp bars after testing. (a) (b) figure 4. residual flexural strength of g1 at different temperatures: (a) immediate; (b) after cooling. 78 67 48 84 32 31 84 19 11 0 20 40 60 80 100 120 140 100-1h-i 200-1h-i 300-1h-i r es id ua l f le xu ra l s tr en gt h (% ) temperature level 10mm 16mm 20mm 95 86 8 89 5 96 9492 90 97 0 20 40 60 80 100 120 140 100-1h-c 200-1h-c 300-1h-c r es id ua l f le xu ra l s tr en gt h (% ) temperature level 10mm 16mm 20mm flexural behavior of frp bars after exposure to elevated temperatures 17 (a) (b) figure 5. residual flexural modulus of g1 at different temperatures: (a) immediate; and (b) after cooling. (a) (b) figure 6. the behavior of different types of frp bars exposed to different temperature levels: (a) residual flexural strength; (b) residual flexural modulus. 3.4. effect of frp type figure (6a) shows the residual flexural strength of all types of frp bars with a 10 mm diameter. similar to type g1, the flexural strength of all frp types decreased as the temperature level increased due to the resin decomposition at higher temperatures above the tg (around 120℃). therefore, a significant decrease of the flexural strength occurred because the polymer lost part of its ability to hold the fibre and to transfer the stresses between fibres. all types showed close residual flexure strengths at different temperatures. the residual flexure strength values ranged from 78 to 95%, 63 to 71, and 40 to 61% at 100, 200, and 300℃, respectively. however, at all temperature levels, type g2 (sand coated gfrp bars) showed the highest residual flexural strength compared to other types. the relationship between the flexural modulus and temperature is shown in fig. 6(b). it can be seen that the flexural modulus was not affected at 100℃. for both 200 and 300℃, the flexural modulus was significantly decreased for all frp types. the residual flexural modulus of the gfrp bars at 200 and 300℃ ranged from 69 to 77% and 59 to 69, respectively. the cfrp bars (c1), however, showed the lowest residual flexural modulus among all types with losses of about 53 and 67% at 200 and 300℃, respectively. this may point to a weaker contact between the matrix and the carbon fibres at a temperature higher than the tg. it may be also due to the type of matrix used in manufacturing the carbon bars. additional studies are required in the future to clarify this point. 4. conclusion based on the test results of this research, the following conclusions can be drawn: 98 71 64 96 18 19 83 18 18 0 20 40 60 80 100 120 140 100-1h-i 200-1h-i 300-1h-i r es id ua l f le xu ra l m od ul us (% ) temperature level 10mm 16mm 20mm 98 9 9 10 0 10 0 98 9894 92 84 0 20 40 60 80 100 120 140 100-1h-c 200-1h-c 300-1h-c r es id ua l f le xu ra l m od ul us (% ) temperature level 10mm 16mm 20mm 78 67 48 95 71 61 90 63 40 84 65 53 0 20 40 60 80 100 120 140 100-1h-i 200-1h-i 300-1h-i r es id ua l f le xu ra l s tr en gt h (% ) temperature level g1 g2 g3 c1 98 71 64 99 69 69 99 77 59 10 0 47 33 0 20 40 60 80 100 120 140 100-1h-i 200-1h-i 300-1h-i r es id ua l f le xu ra l m od ul us (% ) temperature level g1 g2 g3 c1 sherif e. el gamal, abdulrahman m. al-fahdi, mohammed meddah, abdullah al-saidy, and kazi abu sohel 18 • the flexural strengths and moduli of all immediately tested specimens decreased as the temperature level increased. for all 10mm diameter bars, the residual flexural strengths ranged between 40 and 61%, while the residual flexure moduli ranged between 33 and 69%, after exposure to 300℃. • bars with larger diameters showed lower residual flexural strengths and moduli compared to smaller diameter bars. the bars with 10, 16, and 20 mm diameters tested immediately after exposure to 300℃ showed residual flexural strengths of 48, 31, and 11%, respectively. • at 200 and 300℃, the immediately tested g1 specimens showed significant flexural strength and moduli losses (29 to 89%) compared to the losses in the g1 bars tested after cooling to room temperature (0 to 16%). • all tested gfrp bars showed comparable behavior at different temperatures. however, type g3 showed the highest flexural strength losses (37 and 60%) and type g2 bars showed the lowest losses (29 and 39%) after exposure to 200 and 300℃, respectively. • the carbon frp bars showed residual flexural strength ratios comparable to those of the gfrp bars; however, they showed lower residual flexural modulus ratios at 200 and 300℃. conflict of interest the authors declare no conflict of interest. funding this study was funded by the department of civil and architectural engineering, college of engineering, sultan qaboos university, oman. acknowledgment the authors would like to acknowledge sultan qaboos university, the civil and architectural engineering department, and all the technicians at the structural laboratory for their help and support. references alsayed, s., al-salloum, y., almusallam, t., elgamal, s., aqel, m. (2012). performance of glass fiber reinforced polymer bars under elevated temperatures. composites part b, vol. 43, pp. 2265-2271. ashrafi, h., bazli, m., najafabadi, m., oskouei, a. (2017). the effect of mechanical and thermal properties of frp bars on their tensile performance under elevated temperatures. construction and building materials, vol. 157, pp. 1001–1010. astm d790-03 (2003). standard test methods for flexural properties of unreinforced and reinforced plastics and electrical insulating materials. astm international, west conshohocken, pa, usa. bazli m, abolfazli m (2020), mechanical properties of fibre reinforced polymers under elevated temperatures: an overview. polymers 12, 2600; doi:10.3390/polym12112600. benmokrane, b., eisa, m., el-gamel, s., denis, t., elsalakawy, e. (2008). pavement system suiting local conditions. aci concrete international magazine. november 2008: pp.34-39. benmokrane, b., el-salakawy, e., el-ragaby, a., and el-gamal, s.e. (2007). performance evaluation of innovative concrete bridge deck slabs reinforced with fibre-reinforced-polymer bars. canadian journal of civil engineering, 34(3), 298-310. bouguerra, k., ahmed e.a., el-gamal, s.e., and benmokrane, b. (2011). testing of full-scale concrete bridge deck slabs reinforced with fiber reinforced polymer (frp) bars. construction and building materials journal, vol. 25, 3956–3965. dulude, c., ahmed, e., el-gamal, s., benmokrane, b. (2011). testing of large-scale two-way concrete slabs reinforced with gfrp bars. advances in frp composites in civil engineering, 287-291. el-gamal, s.e., abdul rahman, b., benmokrane, b (2010). deflection behavior of concrete beams reinforced with different types of gfrp bars. the 5th international conference on frp composites in civil engineering, beijing, china, sept. 27-29 el-gamal, s.e., al-nuaimi, a., al-saidy, a., allawati, a. (2014). flexural strengthening of rc beams using near-surface mounted fibre reinforced polymers. the brunei international conference on engineering and technology, institut teknologi brunei, brunei darussalam, november 1-3, 10p. el-gamal, s.e., alshareedah, o. (2020a). the behavior of axially loaded low strength concrete columns reinforced with gfrp bars and spirals. engineering structures, 216, 110732 el-gamal, s.e., alshareedah, o. (2020b). experimental study on the performance of circular concrete columns reinforced with gfrp under axial load. international conference on civil infrastructure and construction (cic 2020), february 2-5, 2020, doha, qatar. el-gamal, s.e., benmokrane, b., and el-salakawy, e.f. (2009). cracking and deflection behavior of one-way parking garage slabs reinforced with cfrp bars. aci special publication, sp-264-3, 3352. hamad, r., johari, m., haddad, r. (2017). mechanical properties and bond characteristics of different fiber reinforced polymer rebars at elevated temperatures. construction and building materials, vol. 142, pp. 521-535. hibbeler, r.c. (2018). structural analysis, 10th edition in si units, pearson, prentice hall, pp. 700. about:blank about:blank https://link.springer.com/chapter/10.1007/978-3-642-17487-2_59 https://link.springer.com/chapter/10.1007/978-3-642-17487-2_59 about:blank about:blank about:blank about:blank about:blank about:blank flexural behavior of frp bars after exposure to elevated temperatures 19 jafarzadeh h, nematzadeh m (2020), evaluation of post-heating flexural behavior of steel fiberreinforced high-strength concrete beams reinforced with frp bars: experimental and analytical results. engineering structures 225: 111292. maranan, g., manalo, a., karunasena, b., benmokrane, b., lutze, d. (2014). flexural behavior of glass fibre reinforced polymer (gfrp) bars subjected to elevated temperature. the 23rd australasian conference on the mechanics of structures and materials (acmsm23), vol. i, byron bay. robert, m., benmokrane, b. (2010). behavior of gfrp reinforcing bars subjected to extreme temperatures. journal of composites for construction, vol. 14, no. 1, pp. 353-360. thébeau, d., benmokrane, b., and el-gamal s.e. (2010). three-year performance of continuously reinforced concrete pavement with gfrp bars. the 11th international symposium on concrete roads, seville, spain, october 13-15, 11p. appendix a: eqn. (1): flexure strength (fb) the flexural strength was derived using the following equations (hibbler r.c. 2018): 𝑓𝑓𝑏𝑏 = 𝑀𝑀 𝐼𝐼 𝑦𝑦 (a1) where 𝑓𝑓𝑏𝑏 is the flexural strength of the frp bar (n/mm2); m is maximum bending moment under the load (n.mm); i is the moment of inertia of the frp bar about its neutral axis (mm4), y is the perpendicular distance from the neutral axis of the frp bar to its outer surface (mm). 𝑀𝑀 = 𝐹𝐹 2 �𝐹𝐹 2 � = 𝐹𝐹𝐹𝐹 4 (a2) 𝑦𝑦 = 𝑑𝑑 2 (a3) 𝐼𝐼 = 0.25 𝜋𝜋 (𝑑𝑑 2 )4 (a4) where, d is the diameter of the gfrp bar (mm) substituting eqns. (a2), (a3) and (a4) into eqn. (a1) results in: 𝑓𝑓𝑏𝑏 = 𝐹𝐹𝐹𝐹 4× 0.25 𝜋𝜋 � 𝑑𝑑 2� 4 𝑑𝑑 2 (a5) 𝑓𝑓𝑏𝑏 = 8𝐹𝐹𝐹𝐹 𝜋𝜋 𝑑𝑑3 (a6) appendix b: eqn. (2): flexure modulus, (eb) the calculations of the flexural modulus were derived using the double integration method given in (hibbler r.c. 2018). 𝑑𝑑2𝑣𝑣 𝑑𝑑𝑥𝑥2 = 𝑀𝑀 𝐸𝐸𝐼𝐼 = 𝐹𝐹𝑥𝑥 2𝐸𝐸𝐼𝐼 (b1) 𝐸𝐸𝐼𝐼 𝑑𝑑𝑣𝑣 𝑑𝑑𝑥𝑥 = ∫ 𝐹𝐹𝑥𝑥 2 𝑥𝑥 0 𝑑𝑑𝑑𝑑 = 𝐹𝐹𝑥𝑥2 4 + 𝐶𝐶1 (b2) 𝐸𝐸𝐼𝐼𝐸𝐸 = ∫ ( 𝐹𝐹𝑥𝑥2 4 + 𝐶𝐶1) 𝑥𝑥 0 𝑑𝑑𝑑𝑑 = 𝐹𝐹𝑥𝑥3 12 + 𝐶𝐶1𝑑𝑑 + 𝐶𝐶2 (b3) at 𝑑𝑑 = 𝐹𝐹 2 → 𝑑𝑑𝑣𝑣 𝑑𝑑𝑥𝑥 = 0 from 1, 0 = 𝐹𝐹( 𝐿𝐿 2) 2 4 + 𝐶𝐶1 → 𝐶𝐶1 = − 𝐹𝐹𝐹𝐹 2 16 (b4) at 𝑑𝑑 = 0 → 𝐸𝐸 = 0 from 2, 0 = 0 + 0 + 𝐶𝐶2 → 𝐶𝐶2 = 0 (b5) by substituting the values of c1 and c2 into eqn. (b3), 𝐸𝐸𝐼𝐼𝐸𝐸 = 𝐹𝐹𝑥𝑥 3 12 − 𝐹𝐹𝐹𝐹 2 16 𝑑𝑑 (b6) at x = l/2, δ = δmax 𝐸𝐸𝐼𝐼δ𝑚𝑚𝑚𝑚𝑥𝑥 = 𝐹𝐹( 𝐿𝐿 2) 3 12 − 𝐹𝐹𝐹𝐹 2 12 �𝐹𝐹 2 � = −2𝐹𝐹𝐹𝐹 3 96 (b7) 𝐸𝐸𝑏𝑏 = −2𝐹𝐹𝐹𝐹3 96 𝐼𝐼δ𝑚𝑚𝑚𝑚𝑚𝑚 = −𝐹𝐹 ∆𝑚𝑚𝑚𝑚𝑚𝑚 2𝐹𝐹 3 96 �0.25 𝜋𝜋 � 𝑑𝑑 2� 4 � (b8) 𝐸𝐸𝑏𝑏 = −𝐹𝐹 δ 4𝐹𝐹 3 3 𝜋𝜋 𝑑𝑑4 (b9) l (mm) f f/2 f/2 x l f f/2 f/2 https://www.x-mol.com/paperredirect/1309584629869809664 https://www.x-mol.com/paperredirect/1309584629869809664 https://www.x-mol.com/paperredirect/1309584629869809664 https://www.x-mol.com/paperredirect/1309584629869809664 https://www.x-mol.com/paper/journal/12462?r_detail=1309584629869809664 1. introduction 2. experimental work 2.1. frp bars 2.3. flexural test procedure 3. test results and discussions 3.1. summary of results 3.2. failure modes of tested bars 3.3 effect of temperature level and bar diameter (type g1 bars) 3.4. effect of frp type 4. conclusion conflict of interest funding acknowledgment references the flexural strength was derived using the following equations (hibbler r.c. 2018): where, d is the diameter of the gfrp bar (mm) ,𝑓-𝑏.= , 8𝐹𝐿𝜋 ,𝑑-3.. (a6) appendix b: eqn. (2): flexure modulus, (eb) the calculations of the flexural modulus were derived using the double integration method given in (hibbler r.c. 2018). copy of copy of comp020604.qxd the journal of engineering research vol. 3, no. 1 (2006) 43-54 1. introduction the advanced society is strongly supported by electronics. with the progress of society, much more aid of electronics in the wide area including hard environment ("hard electronics"), such as high-temperature, high power, high irradiation, and high frequency, is needed. the term of "hard electronics" was incepted in japan in 1991. it means a new kind of electronics needed in hard circumstances utilizing hard materials, which are essentially tough themselves and very hard to be handled. the hard materials involve wide band gap semiconductors and new kind of oxides (ma et al. 2002; prelas et al. 2002; _______________________________________ *corresponding author's e-mail: touatif@squ.edu.om spitsyn et al. 2002). hard electronics research continued to develop throughout the 80s with increasing emphasis on high-temperature operation, i.e. high-temperature electronics (htes) (patrick et al. 1997). the energy crisis precipitated by the middle east saw a clear focus on improving petroleum recovery from oil wells by improved down hole monitoring, necessitating increased temperature coverage by the electronic systems. there are many equally potential applications with increasing commercial opportunities afforded by programs of more electric aircraft and more electric vehicles (kodjak, 1993 and 1996). a main theme also has been in distributed control systems for improved reliability and increased fuel economy, either directly by improved combustion control or indihigh-temperature electronics: status and future prospects in the 21st century f. touati*, f. mnif and a. lawati department of electrical and computer engineering, sultan qaboos university, p. o. box 33, al-khoud, muscat, 123, sultanate of oman received 2 june 2004; accepted 18 september 2005 abstract: this paper reviews the state of current electronics and states the drive toward high-temperature electronics. the problems specific to high-temperature effects on conventional electronics and prospects of alternative technologies like silicon-on-insulator, silicon carbide, and diamond are discussed. improving petroleum recovery from oil wells with hightemperature coverage of downhole electronics, making combustion processes more efficient utilizing embedded electronics, programs for more electric aircraft and vehicles necessitating distributed control systems, and environmental protection issues stress the need to use and develop high-temperature electronics. this makes high-temperature electronics a key-enabling technology in the 21st century. actual applications using high-temperature electronics are discussed in some details. also information and guidelines are included about supporting electronics needed to make a complete high-temperature system. the technology has been making major advancements and is expected to account for 20% of the electronics market by 2010. however, many technical challenges have to be solved. keywords: high temperature, hard electronics, downhole monitoring, electric aircrafts and vehicles, efficient use of energy, distributed control systems .øjô°û©©dgh …oé◊g ¿ô≤dg ‘ á«∏ñ≤à°ùÿÿg äé©bƒàdgh ‹é◊g ™°vƒdg :á«dé©dg iqgô◊g äé«fhîµdg »àjƒj .±@≈«jgƒdg .´ h ∞«æe .± , ::áá°°uuóónnäé«fhîµdód á≤agôÿg πcé°ûÿg åëñdg gòg ¢ûbéæj .á«dé©dg iqgô◊g äé«fhîµdg ¤g ∫é≤àf’g ò°s í°vƒjh é«dém á∏ª©à°ùÿg äé«fhîµdód á©lgôe åëñdg gòg øª° àj á°séÿg iqhô° dg äé≤«ñ£àdg øe ~j~©dg í°vƒj .¢séÿgh ¿ƒµ«∏°ùdg ~«hqéc ,∫ré©dg ≈∏y ¿ƒµ«∏°ùdg äé«æ≤j øe é¡d á©bƒàÿg πfg~ñdgh iqgô◊g äélqo ´éøjqg aéækg áj~«∏≤àdg i~e ioéjr ∫ón øe ôø◊g πngo áñbgôÿg iõ¡lg ôjƒ£àh §øædg qéhbg øe ∫hîñdg ¢uóîà°sg ôjƒ£j :äé≤«ñ£àdg √òg ºgcg øeh .á«dé©dg iqgô◊g äé«fhîµdg ôjƒ£jh ωg~îà°s’ ìcg ~ªà©j π≤f §fé°sh hg ägôféw êéàf’ iqƒ£àÿg èegèdg ,iô°ûàæÿg äé«fhîµdg iõ¡lg ωg~îà°séh iaéøc ìcg ¥gîm’g ägƒ£n π©l ,é¡jé«fhîµdg πñb øe √飨ÿg iqgô◊g ¿ô≤dg é«lƒ∏æµj ‘ ~yé°ùe éméàøe á«dé©dg iqgô◊g äé«fhîµdg π©éj gòg πc .áä«ñdg ájéªëh á≤∏©àÿg äé≤«ñ£àdg gòngh ,iô°ûàæe iô£«°s iõ¡lg ¤g áléëh ¿ƒµj aéhô¡µdg ≈∏y πlg øe äé«fhîµd’g ºy~d ägoé°tq’gh äéeƒ∏©ÿg é° jg åëñdg øª° àj éªc .åëñdg ‘ é¡à°ûbéæe ” á«dé©dg iqgô◊g äé«fhîµd’ á«∏©ødg äé≤«ñ£àdg øe ~j~©dg .øjô°û©dgh …oé◊g ~j~©dg ¿éa gòg ™e .2010 ωéy ™e äé«fhîµd’g ¥ƒ°s øe %20 »£¨j ¿cg ™bƒàÿgh é«lƒ∏æµàdg √òg ≈∏y cgôw òñc ω~≤j ∑éæg .á«dé©dg iqgô◊g äé«fhîµd’ πeéc ωé¶f aéæh .πm ¤g áléëh á«æ≤àdg äéj~ëàdg øe ::áá««mmééààøøÿÿgg ääggooôôøøÿÿggiõ¡mg ,é«lƒdƒæµà∏d céøc’g ωg~îà°s’g ,aéhô¡µdg ≈∏y ìcg ~ªà©j π≤f §fé°shh ägôféw ,ôø◊g πngo áñbgôÿg ,áñ∏°üdg äé«fhîµd’g ,á«dé©dg iqé◊g äé«fhîµdg .iô°ûàæÿg iô£«°ùdg 44 the journal of engineering research vol. 3, no. 1 (2006) 43-54 rectly by reduced weight by reducing the requirement for ancillary equipment. there are still many problems to overcome for the full adoption of hte systems in "everyday" electronics. high-temperature designs need components rated to operate in the hard environment in which they will be used. this paper reviews the state and future prospects of htes as a technology increasingly affecting strategic sectors in the 21st century. some details of actual applications using htes have been discussed. information has also been included about components and materials needed to make a complete high-temperature system. 2. high-temperature effects on device properties 2.1 general tendency as temperature increases in a semiconductor, there are challenging issues to resolve (lall et al. 1997): intrinsic carrier density increases. junction leakage current increases exponentially (p-n junctions, bjts, and fets), which presents the biggest problem. device parameters vary (forward bias voltage, threshold voltage, forward current, carrier mobility, input impedance, transconductance, susceptibility to latchup). electromigration of interconnection traces increases. chemical reactivity of ohmic contacts increases. dielectric breakdown strengths decrease. tec mismatches stress the die mechanically. the most important physical properties for a good hightemperature semiconductor material are a wide bandgap and high thermal conductivity. bandgap, because it determines the amount of leakage current across the junction. thermal conductivity, because it determines the semiconductor's ability to dissipate heat to the ambient environment. the above effects show that the temperature is a major limitation for electronic devices. the upper temperature limit for conventional electronics is 100oc, with some special military components being operated at a maximum of 125oc as shown in fig. 1. possible way to cope with a high-temperature environment is to use active cooling systems and to place electronics some distance away in a moderate environment. however, large, heavy, and complex cooling systems and the cabling and interconnections they require are not at all advantageous from the viewpoint of accuracy, reliability, size and cost. the development of htes is, therefore, eagerly sought after. 2.2 bipolar technology typically, amplifiers must have low offset, high impedance, and high gain to minimize errors in a control loop. unfortunately, bipolar junction transistors, which are typically used for these reasons, do not work well at high temperatures because of: forward current gain increases with temperature at about 30%/100ºc. forward voltage decreases at about 200 mv/100ºc. reverse leakage of the base/collector junction with open emitter doubles every 8ºc. figure 1. temperature limits for electronic applications *eor = enhanced oil recovery oc 45 the journal of engineering research vol. 3, no. 1 (2006) 43-54 resistivity of silicon (si), which is temperature dependent. a loss of linearity that occurs in the saturation region with temperature increase. increasing leakage current, which is often the dominant effect, eventually causes loss of rectification and isolation. bipolar circuits start to fail above 150ºc. in the bipolar transistor characteristics, not only is the collector current dramatically reduced, but there is significant distortion and loss of linearity at high temperatures. also, the output resistance is not constant. 2.3 cmos technology at high temperatures, cmos devices also experience reduced performance. unfortunately, most devices are not tested above 150ºc. characterization testing of a typical mosfet device in the range of 100ºc-200ºc indicates that: breakdown voltage increases with temperature at about 8%/100ºc. on-resistance increases with temperature at about 70%/100ºc. leakage current increases rapidly with temperature, changing four orders of magnitude/100ºc, doubling every 8 ºc. gate-source threshold voltage decreases with temperature, about 24%/100ºc. conventional cmos and mosfet devices usually operate to 160ºc. with dielectric isolation, this limit can be extended to about 200ºc, and when designed for high temperature, to about 300ºc. 3. why are high-temperature electronics needed? the industries most interested in designing products using htes today can be broken along field lines as follows: 3.1 oil and gas well logging the energy crises of the 70s backed by a second energy crisis precipitated by the middle east nations emphasis on the need for a clear focus on improving petroleum recovery from oil wells by improved downhole monitoring, necessitated increased temperature coverage by downhole instruments. this requires sensors and data acquisition systems used to log the environment around drilling heads in wells, measuring reservoir potential and environmental conditions, and geothermal implementation of 'permanent sensors' intended to allow long life; up to 5 years operation with no failures. this will also help discovering new markets in deeper (i.e. hotter) wells operating at 400oc and above. this has presented electronics with a challenging environment. locating an instrument for pressure or flow measurement at the end of 3 miles of wire poses problems for electronics including temperatures of 185°c (extended periods) to 250°c (shorter periods). geothermal wells push continuous operating temperatures to 300°c, with survival to 350°c required (normann, 1998). traditional limits of several hours to 100 hours are being pushed out to several thousand hours, and in the case of permanent gages, to several years of continuous operation. there are three typical types of instruments which are discussed below. 3.1.1 logging tools logging tools are deployed for gathering snapshot information revealing the quality and potential production of the well. these can be wire-connected (wireline tools, receiving power and communications from the surface) or steel cable connected (memory tools, which are self-powered and contain memory which is retrieved with the tool). 3.1.2 measurement while drilling tools these are used for directional drilling information such as hole inclination, hole azimuth, and the tool face direction. in addition, other measurements are now performed including resistivity, natural gamma, and neutral density. downhole weight on bit, torque, vibration levels and accelerations are sometimes also incorporated into the measurement. this environment is extremely rugged requiring endurance of high vibrations and shock and thermal cycling every several days as the bit is pulled up and changed. because of the high cost of drilling operations ($20,000 to $100,000 per day) these instruments need to be highly reliable. in some cases, the drilling operation is allowed to continue without operating instrumentation. 3.1.3 permanent gauges permanent gauges are installed in producing wells and monitored at the surface periodically over several years. temperature, pressure, flow and different chemical sensors are deployed. these devices must have good longterm stability without susceptibility to premature wearout mechanisms such as electromigration (discussed later). 3.2 military and space the mechanisms of failure of electronic components are being looked at afresh. until quite recently the us military handbook 217 provided design guidelines linking failure rate exponentially to steady state temperature which led to overly large cooling systems adding weight, cost and complexity to aircraft. from the engine manufacturers' viewpoint, transducers with high-temperature capability would enable more accurate assessment of parameters in hard environments. this would facilitate the extended use of data buses to replace dedicated wiring from engine probes to remote electronic processing circuits. the impact would give considerable weight and reliability. total weight of connectors and wiring in the harness is around 60-65 kg in current large civil aircraft. in the aerospace industry, htes could potentially 46 the journal of engineering research vol. 3, no. 1 (2006) 43-54 enable improved system designs resulting in lighter, faster, cheaper, and more fuel-efficient aircraft. most severe environment is around the engine. controlling electronics operate in an ambient of up to 250oc and cooled by routing and interfacing engine fuel (at less than 105oc) with the heat sinks. in two specialist applications, electronics already operate above 100oc: in electronic braking systems for large aircraft, uv flame detection for turbine engine combustor con trol. htes are also required for the implementation of multiplexed systems for gas turbine engine control, which would enable complex cooling and connection requirements to be eliminated. studies done at the us air force research laboratory show that better regulation of engine thrust can result in doubling of the thrust/weight ratio, 40% fuel burn reduction, at the same time providing a 35% decrease in engine life cycle costs (lewis, 1998). to achieve these goals, control logic must be designed to use a broad array of hundreds of sensors/actuators operating over a wide operating temperature range to control thrust while protecting against aerodynamic, thermal, or material strength limitations. the engine case is a significant source of radiant heat, approaching 560°c at high altitudes. conventional uncooled electronics cannot be used here because of thermally induced wear factors leading to degradation and premature failure. although fuel cooling has been very effective at improving the electronics environment, this approach favors centralized control architectures as well as the added weight and space required for the cooling pumps, heat exchanger, and piping. also, deep space exploration of planets like neptune requires electronics which are able to operate below -220°c on the surface of triton or pluto. electronics for venus 2006-2007 missions need to operate at above 470°c and it seems that suitable htes may not be ready in time except for "grab-and-go" or limited life" operations. 3.3 automotive there is a rising need for more electric aircraft and more electric vehicles which are more reliable, faster, simpler, lighter, and less polluting. the automotive sector presents, by far, the largest potential market for htes. however, automotive applications will prove to be the most difficult for hte to penetrate, due to: extremely pricing pressures, very demanding requirements for long term reliability, the long cycles for new designs. five automotive applications will provide the main drivers for htes and high-temperature power devices: engine control, transmission control, abs (antilock braking systems), active suspensions, sensors for wheel speed and map (manifold absolute pressure (systems). htes and sensors are becoming increasingly necessary for automotive control applications such as engine control units (ecus), in-cylinder pressure sensing and camshaft position sensing because of international requirements to rapidly reduce engine emissions. reducing emissions depends on more precise and local control of valve timing, which in turn depends on primary, not inferential, sensing of cylinder pressure and camshaft position. improveing the precision of these measurements requires the sensors and signal conditioning electronics to be mounted close to, or in some cases inside, the engine itself. ecus, typically mounted on the firewall of the engine compartment, can also contribute to reduce emissions by being mounted on the engine itself. the shorter control loops to/from the ecu lead to more precise control. 3.4 efficient energy use and environment protection the global energy demand is growing at a rate of 3 to 4% a year. accordingly, the world energy consumption will double around 2050. a survey of future carbon dioxide emission predicts environmental crisis such as: increase in atmospheric temperature, rise of sea surface level, so this will be a challenge for mankind to avoid the crisis in the 21st century. three basic technological measures to take are as follows, . energy system change from fossil fuel to non-fossil fuel use, . high efficiency energy utilization, . control of carbon dioxide emission and collection. as to the energy system change, some report says that today's ratio of fossil (80% of the 11 terawatts) to non-fossil energy (20%) utilization have to be changed into the ratio of 40% to 60% of the 30 terawatt total until middle of the 21st century to stabilize carbon dioxide concentration in the air (matin, 1998). most of non-fossil energy will be electricity, supplied by atomic energy and reproducible natural energy such as solar energy, water power, wind power and terrestrial heat. in japan, about 40% of the total input energy has been already supplied by electricity. the term high-efficiency energy utilization is two fold. first, it will be more and more important for mankind to develop high-efficiency electrical energy utilization social system in this century, because electrical energy demand will rapidly grow in the world. second, improvement of combustion control system is necessary to attain efficient energy use and minimal pollution. as to 47 the journal of engineering research vol. 3, no. 1 (2006) 43-54 the first, expected electric power leading energy technologies for this century are: power generation: atomic, new systems to find out power transmission: power system linkage power storage: pumped storage, hydraulic generation power conversion: ev, high speed railway system (inverters) where power electronics will be essential for the new technologies. the corresponding power devices must, therefore, be of: low power consumption, higher switching speed, higher temperature operation. as to the second, in japan, in 1994 the ministry of international trade and industry (miti) began the project called "research and development on combustion control for energy conservation". the project had a term of 6 years. miti has delegated management of the research and development to the ion engineering research institute corporation through the new energy and industrial technology development organization (nedo). the objectives of this r&d are to establish technologies for realizing sic-based devices applicable to the advanced combustion control system as shown in fig. 2. more specifically, to achieve the advanced new system, r&d is focused on sic devices operable in a high-temperature and oxidizing environment. 3.5 communication sector the communication sector is also very much concerned by hard electronics devices. the communication systems in 1970's to 1980's were characterized by: high-power land/satellite base stations and home receivers for broadcasting, analog transceiver for wireless telecommunications and cable network communication. from the beginning of 1990's, new systems characterized by: very small-sized personal mobile telephones (cellular), personal computer network (internet). have been implemented. keywords in these new systems: digital, high speed, mobile and personal. in the beginning of this century, 50 cc size and 50 g weight cellular handset (can be wristband type) with a high rf performance has appeared. also, for this century various high-frequency communications are projected. very small base stations are strongly demanded in order to install them on rooftops of buildings. hard electronics will contribute a great deal: 3.6 other applications in addition, markets needing htes are forming in the areas of integrated motor controls, industrial process controls, building fire safety controls, environmental test equipment, and other areas. a number of recent requirement changes have pushed the upper operating temperature range limit higher, as shown in table 1. figure 2. schematic new combustion control system using sic-based control devices requirement effect on product explore new, deeper oil wells hotter, i.e., equipment exposed to higher temperatures downhole drilling cost increased durable and more reliable equipment required improve aircraft performa nce distributed engine controls optimize fuel consumption lighter controls and monitors remove cooling systems hotter operat ing conditions with no airflow or heat sink 48 the journal of engineering research vol. 3, no. 1 (2006) 43-54 standard, commercial off the shelf (cots) bulk si processed semiconductors, however, are not able to handle these increased demands. they are designed for a service range of -55°c to +70°c for commercial, +85°c for industrial or +125°c for military temperature range. use of these components beyond their intended range is not specified or recommended by their manufacturers. the discussion above on the various strategic sectors emphasizes the need for devices having higher temperature ratings. in fact, the 2002 five-year update to the international technology roadmap for semiconductors (itrs) did not reflect the immediate need for higher operating temperatures for complex integrated circuits or asics; but did recognize increasing temperature requirements for power and linear devices as shown in table 1 (johnson, 2004). nevertheless, in the 2003 edition of the itrs, the maximum junction temperatures identified for hard-environment complex integrated circuits was raised to 150°c through 2018 (johnson, 2004). the ambient operating temperature extreme for hard-environment complex integrated circuits was defined as 40°c to 125°c through 2009, increasing to 40°c to 150°c for 2010 and beyond. 4. status and prospects of high-temperature electronics 4.1 high-temperature semiconductor technologies generally, the main part of electronics is composed of semiconductors. ordinary semiconductors si, germanium (ge), and gallium arsenide (gaas), have rather small energy gaps as shown in table 2. due to the limitation of material properties caused by the bandgap, they cannot withstand severe environment. so, we need wide bandgap semiconductors such as silicon carbide (sic), gallium nitride, diamond, and oxides. the same reason has been behind the interest shown on si over ge (esi, gap > ege, gap) fifty years ago and the technological development of the si industry so far. however, as table 3 shows, si and silicon on insulator (soi) are the only technologies with good high-temperature characteristics that is mature enough for production use. soi technology was originally developed to meet aerospace needs for instrumentation, data acquisition and distributed control. the temperature and reliability requirements for the aerospace applications supported the 2001 2002 2003 2004 2005 2006 2007 complex ics -40°c to +125°c -40°c to +125°c -40°c to +125°c -40°c to +125°c -40°c to +125°c -40°c to +125°c -40°c to +125°c power/linear -40°c to +150°c -40°c to +150°c -40°c to +180°c -40°c to +180°c -40°c to +200°c -40°c to +200°c -40°c to +200°c table 1. international technology roadmap for semiconductors ambient operating temperatures for hard environments (automotive) element bandgap e g (ev) electron mobility (cm/s) breakdown voltage (v/cm) si 1.12 1x107 2x104 ge 0.67 5x107 7x103 gaas 1.43 2x107 3x104 sic 2.2 3.00 2x107 25x105 diamond 4.8 1x107 8x10 5 table 2. the main physical parameters for various semiconductors technology maturity temperature source products si cmos (bulk) production -50ºc to 150ºc multiple sensor i/f, logic, memory, power high-temp. tolerant si (soi) production -50ºc to 300ºc honeywell sensor i/f, logic, memory, power e/d gallium arsenide (gaas) production -50ºc to 150ºc vitesse logic, memory complimentary gaas development -50ºc to 350ºc honeywell logic, power sic development -50ºc to 60 0ºc cree, siemens, abb very few (discrete) diamond early development -50ºc to 1000ºc research labs none table 3. high-temperature tolerant semiconductor technologies (hiten, london, england) 49 the journal of engineering research vol. 3, no. 1 (2006) 43-54 development of reliable integrated circuits to operate for long periods of time at high temperatures. with the fullydepleted soi mos technology, si has very good thermal properties, theoretically reaching temperatures of 400°c, and low power si devices have been widely demonstrated to operate at 300°c (brusius, 1994; lemnios, 1995; naefe et al. 1998; pinardi, 2003; shahidi et al. 1993). since by far, most applications operate under 300°c, si-based technologies dominate the field now and will for many years to come. another major selling point for si is that the manufacturing technology is already in place for high volume and low cost products. alternatively, silicon carbides demonstrated ability to function under extreme high-temperature, high-power, and high-radiation conditions. therefore, sic is expected to enable significant enhancements to a far-ranging variety of applications and systems. specific applications include high frequency, high power radar systems, jet ignition systems, turbine engine monitoring and control, spacecraft power conditioning, nuclear power instrumentation, in-cylinder engine sensors, and numerous others. however, improvements in crystal growth and device fabrication processes are needed before sic-based devices and circuits can be scaled-up and reliably incorporated into electronic systems (eriksson, 2002; ziegler et al. 2004). in summary, temperature ranges for high-temperature electronics can be broken down into 3 ranges as shown in table 4. market usage is also shown for relative importance. 4.2 htes devices and systems in the past few years research and development in digital, analog, and power devices for htes have been very active. the ultimate goal is to integrate on-chip application-specific integrated circuits (asics) as solutions for cost, reliability, space/weight, and power issues. in the following sections, we are going to highlight recent applications. 4.2.1 digital devices table 5 shows, mainly well-documented high-digital devices available in soi and sic technology. these devices demonstrated good operations at temperatures up to 300°c. as can be seen, soi technology is dominant. although it is finding wide applications in power systems, it started also in the digital world and a lot of work is being put to produce sic-based high-temperature digital systems. 4.2.2 analog devices high-temperature analog devices such as operational amplifiers (ht1104), analog-to-digital (a/d) converters, analog multiplexers and switches, 5v voltage references all operating at 250°c and above, and many others are commercially available (not restrictively from honeywell, ibm, and cree). a well-documented application of the opamp ht1104 concerns the control of combustion process in engines, as shown in fig. 3, by either simple flame indicators which have a binary output corresponding to the occurrence of flame or not or more complex light spectrum sensors which have a complete analog output (sims, 1996). 4.2.3 power devices far-ranging development and applications of power devices have been reported. recently, for example, in the automotive sector, the impact of sic power electronic devices for hybrid electric vehicles, traction drives, dcdc converters, pwm inverters, and igbts with field stop temperature range percent of market recommend for lo-power recommend for hi-power up to 200ºc 90.2% soi soi 200 to 300ºc 8.6% soi sic above 300ºc 1.2% sic sic table 4. recommended technologies for given temperature ranges (hiten, london, england part temp. range features reference/source soi ibm microprocessor to 300 ºc >500mhz, 32b, 6.5m devices ibm microelectronics division. soi microprocessor ht83c51 to 300 ºc 20mhz, 8k rom, 256 bytes ram, 3 -16b timer/counters honeywell ht83c51 soi sram ht6256 to 300 ºc 200mhz, 32kx8 static ram honeywell ht6256 soi cmos sram to 250 ºc 64kb (fallet et al., 1996) soi gate array ht 2007, 2080, 2400 to 300 ºc sea of 5 k to 200k transistor gates honeywell ht 2007, 2080, 2400 sic cmos ics to 300 ºc p-well process, 5v power supply (ryu et al., 1997, 1998) table 5. samples of available high-temperature digital devices 50 the journal of engineering research vol. 3, no. 1 (2006) 43-54 technology was studied (tolbert et al. 2002 and 2003; ozpineci et al. 2002; hasanuzzaman et al. 2004; spitz et al. 1998; ishii et al. 2001; matsumoto et al. 2001; bauer et al. 2001). honeywell has reported the availability of power positive and negative linear regulators (0.5a, ±5v, ±10v, ±15v) operating well above 200ºc. 4.2.4 high-temperature asics vlsi technologies today offer the potential for integrating mixed signal circuits as asics in high-temperature applications. this solution opens up the possibility of obtaining microelectronics circuits at a lower cost and with better reliability compared to discrete solutions. furthermore, the scale of the technology leads to reduced power supply voltages, which in turn reduces the field levels, the high-temperature effects and the power consumption. as a result, low power supply voltage and high-temperature operation can be used together effectively in the design of high-temperature mixed signal asics (vermesan, 1997). examples of such implementations are described below. ssi/lsi components for hybrid-mounted: typical system control circuits involve analog and digital signal conditioning and processing, memories and peripherals are implemented (fallet et al. 1996). such components are already available for hybrid mounting and operation above 175°c. examples of circuits designed for high-temperature operation are shown in fig. 4. control electronics module for future all-electric aircraft (aea): an interesting example is the actuator control electronics (ace) module, shown in fig. 5. it is targeted for completely distributed control in future aea figure 3. block diagram of engine combustion control circuit figure 4. typical circuit architecture for system monitoring applications figure 5. targeted actuator control electronics module for future aea applications se n so r s a c t u a t o r s 51 the journal of engineering research vol. 3, no. 1 (2006) 43-54 and operation up to 200oc for the wings, 315oc for the engines (bottner et al. 1991). ace modules will require high-speed data bus interface, large memories, powerful microprocessor, precision analog functions as well as power drivers (brown et al. 1994). the development of such systems is still years ahead but a few circuits already aim at this direction. downhole monitoring: the market for electronics in instrumentation and control systems for downhole applications demands increased performance for gauges both in terms of duration and temperature. many companies now require instrumentation equipment that can tolerate up to 200°c up to two months. there is no doubt as to the future market requirements even higher temperatures for longer periods. figure 6 shows architecture for a wire line permanent gage which utilizes a custom gate array and asic to interface pulse width modulated pressure sensors and temperature sensors, and a single bi-directional communication line (normann et al. 1998). 4.3 support electronics & materials used in making high-temperature circuits 4.3.1 passive & discrete devices for passive components, packaging durability at high temperatures is of prime importance. very often, joining, sealing, and package ruggedness are the main issues affecting the reliability of the components, whereas the basic property materials are very robust. there is a strong desire to use surface mount technology (smt) components. more passive high-temperature smt components are now available, making the decision to go to smt easier (naefe et al. 1998). resistors: are the largest utilized components in electronic design. therefore, failure modes, shifts, and thermal drifts all can have very detrimental effects on circuits designed for high-temperature operations. honeywell, heraeus-cermalloy, and huntington electric have for example, manufactured wire wound, thick-film and thinfilm resistors for power applications. these resistors withstand temperatures above 200°c (naefe et al. 1998; grzybowski et al. 1998). to stay below their material temperature limits, these resistors should be derated significantly. capacitors: the strong temperature dependency of the dielectric constant makes capacitances change significantly with increasing temperature. high-temperature capacitors usually have low capacitance due to mechanical breakage caused by thermal packaging stresses. that is why capacitors will need to be paralleled to get high values. electrolytic capacitors do not operate above 150°c due to dielectric breakdown. values of several tens of microfarads are typically wet slug tantalum devices. solid tantalum capacitors are available in values of several microfarads. high-temperature ceramic capacitors tend to be very stable for long periods of time exceeding 5000 hours in some tests, however, their values are usually smaller than 0.1 microfarad. teflon electrostatic capacitors are being developed by custom electronics, which have very low tc of capacitance from room to 200°c, though the parts are usually small valued. some published temperature test results demonstrate that these capacitors show gradual loss of functionality and initial aging shift in capacitance e.g. matsuo solid tantalum and novacap ceramic capacitors (naefe et al. 1998; grzybowski et al. 1998). electromigration: probably the most significant limiting aspect of an electronic component's lifetime at high temperature is the ability of the metal films to carry currents for extended periods of time. over years, or even months, with sufficient current density present, metal figure 6. downhole tool architecture 52 the journal of engineering research vol. 3, no. 1 (2006) 43-54 atoms tend to migrate in an uncontrollable dissipation manner (electromigration). lifetimes can, thus, be improved by reducing current density. manufacturers of high-temperature-tolerant electronics eliminate this issue by using layout rules that allow at most 1/10 the current density of conventional bulk si design. ohmic contact and packaging: high temperatures cause increased chemical reactivity of metal to si junctions. metals with high electrical conductivity present poor interfaces and thus do not form stable ohmic contacts. to overcome this problem, intervening layers of materials (e.g. tiw), called diffusion barriers, that are chemically neutral to the interconnect metal and the semiconductor contact, are being used. a significant amount of study is devoted to researching ohmic contacts in sic, an issue that has slowed this technology from entering volume production (cole et al. 2002; lee et al. 2003; madsen, 2001; saxena et al. 1999; touati and saji, 1997). it is also true that packaging and interconnection problems present some of the biggest challenges for electronics at temperatures above 250ºc, often being more of a barrier to successful operation than the intrinsic material issues. when htes are exposed to high temperatures for extended periods of time, the materials used for substrates, die attach, wire bonding, and metal interconnects all have to be chosen carefully and should have good hightemperature properties. some of the important properties for htes packaging and interconnections include: materials used should have matching tecs. interfacial materials used should be chemically inert. interconnect metals should not be mixed if possible, to avoid intermetallics. substrates should have high dielectric strength at high temperature. 4.4 design techniques for high-temperature applications design for high-temperature applications cannot dispense with circuit techniques to solve problems due to high-temperature operation. by combining techniques such as process technology, physical design, and layout techniques cost effective and reliable high-temperature devices and asics can be designed. above 200°c, the soi and epi-cmos combined with conservative layout rules (e.g. guarding and interdigitized structures) and techniques such as low power supply voltage, operating the devices into the zero-temperature coefficient (ztc) biasing point and scaling of transistor aspect ratio offer cost effective solutions for implementing reliable high-temperature systems (fallet et al. 1994; shoucair et al. 1984; vermesan, 1997; wu and brown, 1991). in the aerospace market today, the soi technology coupled with the above high-temperature design techniques are being used. there are currently several major projects ongoing which are in various stages of development, testing, and demonstration. the reliability testing of components has been ongoing for several years. data collection of reliability at the component level has indicated the ability to operate within specification for 5 years or more at 225°c. these results have included thermal cycling, package mechanical testing as well as accelerated temperature and voltage testing (hunter, 2004; lewis, 1998; naefe, 1998). in the industrial control and instrumentation arena, high-temperature circuit techniques are used to overcome the inherent drift and non-linearity of sensors, conditioning blocks (e.g. instrumentation amplifiers, a/d converters), and actuators due to temperature effects (brusius, 1994; bryzek, 1993). in low-temperature applications inherent drift and non-linearity in the sensor due to temperature effects are generally neglected. high-temperature approaches such as sensor's scaling (normann and livesay, 1998), removing of thermal drift of instrumentation amplifiers (paine instruments), and overcoming thermal effects when controlling actuators using pulse-width modulation (pwm) signals (honeywell) were demonstrated. consequently, all "screw" it up and let industry and market guide the way. 5. market overview the world market for hte components reached $148 million in 1995. industry analysts estimate that this will reach $1.5 billion by 2010 as shown in table 6. the market is currently quite a bit small; about $1billion, because of component scarcity, component cost, the lack of semiconductor fabrication facilities capable of producing these devices, and immature packaging and materials technology. it is highly expected that key drivers like well logging equipment, aerospace, and automotive industries, plus several niche applications will boost the hte market. by temperature range, operation up to 200oc satisfies most requirements currently. as the technology matures, 1995 2000 2010 actual market by application sector ($ million) well logging equipment 62 93 313 aerospace 6 38 221 automotive 73 274 848 other 7 31 118 total 148 436 1500 by temperature range, as % of total up to 200 oc 92 88 80 200-300oc 8 11 12 >300oc 0 2 8 by semiconductor technology, as % of total si/sio 88 86 83 gaas 12 13 9 wbs 0 1 8 table 6. the world market for high-temperature electronics 53 the journal of engineering research vol. 3, no. 1 (2006) 43-54 applications at high temperatures will be enabled and are expected to account for 20% of the market by 2010. sibased technologies dominate applications up to 200250oc. gaas and wide bandgap semiconductors (sic, diamond and group iii nitrides) will play critical roles beyond 250oc, accounting for 17% of the market by 2010. 6. conclusions status and future prospects of high-temperature electronics have been reviewed. high-temperature electronics are a key enabling technology for many applications in a wide range of industrial sectors. therefore, they have been recognized as a strategically potential technology. development of high-temperature electronics is being driven mainly by vertically integrated system organizations in strategic sectors such as oil and gas logging, space, automotive, and communications along with small specialist companies developing new semiconductor technologies. these electronics are based on soi and wide band-gap semiconductors like silicon carbide, diamond, and oxides by virtue of their superior physical properties compared to those of conventional technologies (si, gaas). modern systems require distributed electronics to achieve adequate performance and accuracy, but conventional components cannot handle high temperatures because of severe derating and thermal wearout. the major market opportunities will arise from incorporating high-temperature electronics in subsystems/systems, as opposed to high-volume component markets. recently, the technology has witnessed major advancements toward systematic implementations. however, many technical challenges are still pending and need to be surmounted. acknowledgments the authors wish to thank sultan qaboos university (squ) for providing the facilities and resources. references bauer o., auerbach j.g., porst f., roth a., ruething r. and schilling, h., 2001, "6.5 kv-modules using igbts with field stop technology," ieee proceedings of the 13th int. symposium on power semiconductor devices and ics, pp. 121-124, osaka, japan. bottner, t., 1991, "high-temperature electronics for automobile, aeronautical, geothermal, nuclear, and space," proceedings of the 1st int. high-temperature electronics conference, pp. 77-82, albuquerque, nm. brown, d. m., 1994, "high-temperature silicon carbide planar ic technology and first monolithic sic operation amplifier ic," proceedings of the 2nd int. high temperature electronics conference, pp. xi17xi22, charlotte, nc. brusius, p., 1994, "reliable high-temperature soi process," proceedings of the 2nd int. high-temperature electronics conference, pp. 15-19, charlotte, nc. bryzek, j., 1993, "the evolution of smart sensor and transducer design," sensors, pp. 12-22. cole, m.w., joshi, p. c., hubbard, c., demaree, j. d. and ervin, m., 2002, "thermal stability and performance reliability of pt/ti/wsi/ni ohmic contacts to n-sic for high-temperature and pulsed power device applications," j. of applied physics, vol. 91(6), pp. 3864-3868. eriksson, j., 2002, silicon carbide microwave devices,. publisher: chalmers university of technology, sweden. fallet t., 1996, "a high-temperature voltage regulator chip for downhole applications," proceedings of the 3rd int. high-temperature electronics conference, pp. 3-8, charlotte, nc. fallet, t., gakkestad, j. and forre, j., 1994, "application specific integrated circuits for high temperature oil well applications," proceedings of the 2nd int. high-temperature electronics conference, pp. vii9vii14, charlotte, nc. flandre, d. and colinge, j. p., 1994, "high-temperature characteristics of cmos devices and circuits on silicon-on-insulator (soi) substrates," proceedings of ix sbmicro 94, pp. 777-786, rio de jeneiro. grzybowski, r., 1998, "long term behavior of passive components for high-temperature applications-an update," proceedings of the high-temperature electronic materials, devices and sensors conference, pp. 172-179, albuquerque, nm. hasanuzzaman, m., islam, s. k. and tolbert, l. m, 2004, "effects of temperature variation (300°-600°k) in mosfet modeling in 6h silicon carbide," solid state electronics, vol. 48(1), pp. 125-132. hunter, g. w., neudeck, p. g., xu, j., lucko, d., trunek, a., artale, m., lampard, p., androjna, d., makel, d., ward, b. and liu, c.c., 2004, "development of sicbased gas sensors for aerospace applications," materials research society symposium proceedings, vol. 815, pp. 287-297, san francisco, ca, usa. ishii, s., yasuoka, k. and ibuka, k., 2001, "pulsed power application assisted by power semiconductor devices," ieee proceedings of the 13th int. symposium on power semiconductor devices and ics, pp. 11-14, osaka, japan. johnson, r. w., evans, j. l., jacobsen p., thompson, j. r. and christopher, m., 2004, "the changing automotive environment: high-temperature electronics," ieee transactions on electronics packaging manufacturing, vol. 27(3), pp. 164-176. kodjak, d., 1993, "emissions, quantifying the air quality impact of ev recharging," green car j., october issue, pp. 116-120. kodjak, d., 1996, "evs: clean today, cleaner tomorrow," technology review, august issue, pp. 66-67. 54 the journal of engineering research vol. 3, no. 1 (2006) 43-54 lall., p., pecht, m. and hakim, e., 1997, influence of temperature on microelectronics and system reliability: a physics of failure approach. crc press, new york. lee, b. t., shin, j. y., kim, s. h., kim, j. h., han, s. y. and lee, j. l., 2003, "investigation of ti/al and tin/al thin films as the stable ohmic contact for ptype 4h-sic," j. of electronic materials, vol. 32(6), pp. 501-509. lemnios, j., 1995, "manufacturing technology challenges for low power electronics," proceedings of vlsi technology symposium, pp. 5-8, kyoto, japan. lewis, t., 1998, "military aircraft turbine engine electronics and requirements," proceedings of the 4th int. high-temperature electronics conference, pp. 1-23, albuquerque, nm. ma, y., wang, w. l., liao, k. j. and kong, c. y., 2002, "study on sensitivity of nano-grain zno gas sensors," j. of wide bandgap materials, vol. 10, pp. 113-120. madsen, l. d., 2001, "formation of ohmic contacts to ti-sic and their impact on devices," j. of electronic materials, vol. 30(10), pp. 1353-1361. matin, i. h., 1998, "experts urge all-out push for nonfossil energy," nature, november 19 issue. matsumoto, s., hiraoka s. and sakai y., 2001, "a highefficiency 5-ghz-band soi power mosfet having a self-aligned drain offset structure," ieee proceedings of the 13th int. symposium on power semiconductor devices and ics, pp. 99-102, osaka, japan. naefe, j., johnson, w. and grzybowski, r., 1998, "hightemperature storage and thermal shock studies of passive component attach materials," proceedings of the fourth int. high-temperature electronics conference, pp. 68-78, albuquerque, nm. normann, r. and livesay, b., 1998, "geothermal hightemperature instrumentation applications," proceedings of the 4th int. high-temperature electronics conference, pp. 21-31, albuquerque, nm. ozpineci, b., tolbert, l. m., islam, s. k. and hasanuzzaman, m., 2002, "system impact of sic power devices," int. j. of high speed electronics and systems, vol. 12(2), pp. 439-448. patrick, f., mccluskey, grzybowski, r. and podlesak, t., 1997, high-temperature electronicsm. crc press, inc., new york. pinardi, k., 2003,"modelling of high power soi vertical dmos transistors and flip-chip packages". doctoral dissertation, publisher: chalmers university of technology, sweden. prelas, m. a., ghosh, t. k., loyalka, s. k. and tompson, r. v., 2002, "hydrogen storage in diamond films," j. of wide bandgap materials, vol. 10, pp. 99-111. ryu, s., kornegay, k. t., cooper, j. a. and melloch, m. r., 1997, "6h-sic cmos digital ics operating on a 5v power supply," ieee device research conf., pp. 38-39, fort collins, co, usa. ryu, s., kornegay, k.t., cooper, j. a. and melloch, m. r., 1998, "digital cmos ics in 6h-sic operating on a 5v power supply," ieee trans. on electron devices, vol. 45(1), pp. 45-53. saxena, v., nong, j. s. and steckl, a. j., 1999, "highvoltage niand pt-sic schottky diodes utilizing metal field plate termination," ieee transactions on electron devices, vol. 46, no. 3, pp. 456-464. shahidi, g. g., 1993, "soi for a 1-volt cmos technology and applications to a 512-kb sram with 3.5 ns access time," proceedings of iedm, pp. 813-816, washington dc, usa. shoucair, f. s., hwang, w. and jam, p., 1984, "electrical characteristics of large scale integration (lsi) mosfets at very high temperatures, part i: theory," microelectronics reliability, vol. 24(3), pp. 465-485. shoucair, f. s., hwang, w. and jam, p., 1984, "electrical characteristics of large scale integration (lsi) mosfets at very high temperatures, part ii: experiment," microelectronics reliability, vol. 24(3), pp. 487-510. spitsyn, a. b., tompson, r.v., prelas, m. a. and. ghosh, t. k., 2002, "charge carriers removal from 4hsic using field enhanced by optical activation diffusion method," j. of wide bandgap materials, vol. 10, pp. 89-98. spitz, j., melloch, m. r., cooper, j. a. and capano, m. a., 1998, "high-voltage (2.6 kv) lateral dmosfets in 4h-sic," ieee electron device letters, vol. 19, pp. 100-107. sims, p. e., 1996, "high-temperature electronics for distributed control systems," proceedings of the 3rd int. high-temperature electronics conference, pp. x9-15, albuquerque, nm. tolbert, l. m., ozpineci, b., islam, s. k. and peng, f. z., 2003, "impact of sic power electronic devices for hybrid electric vehicles," j. of passenger cars: electronic and electrical systems, pp. 765-771. touati, f. and saji, m., 1999, "electrical properties and interface chemistry in ti/3c-sic contact system," special issue of the ieee transactions on electron devices on silicon carbide, vol. 46(3), pp. 186-189. vermesan, o., 1997, "asic design for high-temperature applications," proceedings of the 2nd int. hightemperature electronics conference, invited paper, pp. 1-7, charlotte, nc. ziegler, k. j., lyon., d. m., holmes, j. d., erts, d., polyakov, b., svensson, k., olin, h. and olsson, e., 2004, "nanomechanical non-volatile memory element," applied physics letters, vol. 84(20), pp. 4074-4076. wu, k. and brown, r. b., 1991, "high-temperature design rules," proceedings of the 1st int. high temperature electronics conference, pp. 267-272, albuquerque, nm. template for electronic submission to acs journals the journal of engineering research (tjer), vol. 17, no. 2, (2020) 100-111 *corresponding author’s e-mail: razzaqul@squ.edu.om doi:10.24200/tjer.vol17iss2pp100-111 analysis of loss reduction techniques for low voltage distribution network a. al-badi1, r. ahshan1, *, s. al-hinai1, a. moosa2, r. shah2, m. al hasani2, and s. khan2 1 department of electrical & computer engineering, college of engineering, sultan qaboos university, muscat, oman 2 muscat electricity distribution company, muscat, oman abstract: energy losses in a typical distribution system can be in a range of 6 to 10%, and it depends on the system characteristics, installed equipment, and operating strategies of the distribution network. losses reduction during peak periods needs special attention since the losses in the system and the cost of the losses are the highest during this peak. distribution system owner (dso) always strives to reduce power losses in the distribution network that eventually leads to energy saving and cost reduction. this paper presents the model of a selected 33 kv, 11 kv and lv network of a representative primary substation, which is a part of the muscat electricity distribution company (medc) network. in order to quantify the losses in various components, the numerical simulation is carried out using the etap software package. the technical losses, power factors, and voltage profiles are quantified and analyzed. this paper also investigates on the optimal conductor and cable selection for 11kv lines, capitalization values for transformer losses to alleviate system losses and hence the system operational cost. the method of determining optimal conductor and cable size for an 11kv distribution network is presented, where the cost of losses for various conductors with their extra construction or material cost are compared. it also presents the detailed model of calculating capitalization values for distribution transformer losses and sample calculation of the capitalization values. utilizing these capitalization values, the transformer buyer can calculate the total life cycle cost of the distribution transformers and select the most economical one. keywords: distribution network; loss reduction method; system modelling; optimal conductor selection; capitalization values. المنخفض الجھد تحلیل تقنیات الحد من خسائر الطاقة لشبكة توزیع خان أ. موسى، ر. شاه، م. الحسنى، س.، س. الھنائى، *ع. البادى ، ر. احسان ٪، وتعتمد على خصائص النظام والمعدات المركبة 10إلى 6یمكن أن تتراوح خسائر الطاقة في نظم التوزیع من : الملخص الخسائر في النظم واستراتیجیات التشغیل لشبكة التوزیع. یحتاج تخفیض الخسائر خالل فترات الذروة إلى اھتمام خاص حیث أن وتكلفة الخسائر ھي األعلى خالل ھذه الذروة. یسعى مالك نظام التوزیع دائًما لتقلیل فقد الطاقة في شبكة التوزیع مما یؤدي في من v 415كیلوفولت و 11كیلوفولت و 33النھایة إلى توفیر الطاقة وخفض التكلفة. تقدم ھذه الورقة نموذج شبكة مختارة بجھد رعیة وھي جزء من شبكة شركة مسقط لتوزیع الكھرباء. من أجل تحدید الخسائر في المكونات المختلفة ، یتم إجراء محطة ف یتم تحدید وتحلیل الخسائر التقنیة وعوامل الطاقة ومالمح الجھد. تبحث ھذه . etapالمحاكاة العددیة باستخدام حزمة برنامج خسائر المحوالت للتخفیف من لقیم الرأسمالیّة الكیلو فولت ، و 11كابالت لخطوط الورقة أیًضا في الموصل األمثل واختیار ال كیلو فولت ، 11خسائر النظام وبالتالي التكلفة التشغیلیة للنظام. یتم تقدیم طریقة تحدید الموصل األمثل وحجم الكابل لشبكة توزیع أو المواد اإلضافیة. یتم عرض النموذج التفصیلي الحتساب قیم حیث تتم مقارنة تكلفة الخسائر لمختلف الموصالت بتكلفة البناء ، ویتم توفیر حساب عینة منھا. باستخدام قیم الرأسمالیّة ھذه، یمكن لمشتري المحوالت حساب التوزیع الرأسمالیّة لخسائر محوالت .التكلفة اإلجمالیة لدورة الحیاة لمحوالت التوزیع واختیار األكثر اقتصاداً .؛ قیم الرأسمالیّةاألمثل ؛ اختیار الموصلنمذجة النظام كة توزیع؛ طریقة الحد من الخسارة؛شب :ةالمفتاحی الكلمات analysis of loss reduction techniques for low voltage distribution network 101 nomenclature i current flowing through the cable/conductor in ampere n number of months in a year p total peak losses in three-phase lines in kw h number of hours in a year r resistance of the cable in ohm/km x no-load losses capitalization value in omr/w y load losses capitalization value in omr/w. tη transmission efficiency kiθ y-bus elements angle in radian ,k iδ δ voltage angles in radian iv voltage magnitude at the ith bus in per unit kv voltage magnitude at the kth bus in per unit kiy the magnitude of y-bus elements in per unit dso distribution system owner lf load factor medc muscat electricity distribution company accc annual carrying charge on extra construction asc cost of annual loss savings in omr dcc annual demand cost in omr/kw ec levelized cost of energy in omr/kwh ecc annual energy cost of peak losses omr/kw kwhc cost of unit energy in omr/kwh ,net asc net annual savings in omr scc system capacity cost in omr tacc total annual cost for peak losses in omr /tac kwc total annual cost for per kw peak losses in omr ,tac psc total annual loss cost of a particular size of the conductor in omr ,tac ssc total annual loss cost of the smallest size conductor in omr wecc wholesale electricity purchase cost per month in omr/kw fi increasing factor fl loss factor plp uniform annual peak load factor kp real power at the kth bus ,loss kip real power losses between the k-i branches kq reactive power at the k th bus ,loss kiq reactive power losses between k-i branches fr peak responsibility factor fcr levelized annual fixed charge rate in omr ,loss kis power losses between the k-i branches 1. introduction with the continuous growth in size and complexity of the power distribution network, reduction in losses can return significant savings for the distribution system owner (dso). losses reduction strategy provides other advantages, such as relief in system capacity, and the possibility of extending capital expenses for improving and expanding the system for the dso (emmanuel m et al. 2017). to meet load demand, distribution companies need to alleviate the system losses and enhance the quality of power supply to achieve social and economic development. losses associated with the distribution network categorize as technical and non-technical losses. in distribution networks, the fixed losses are in a range of 1/4 to 1/3 of the technical losses (inan h et al. 2014). such fixed losses occur due to corona, leakage current, dielectric losses, no-load losses, and current flow through measurement and control elements and realize as heat and noise. the variable losses represent 2/3 and 3/4 of the distribution losses and depend on the magnitude of the current flow (medc 2018). such variable losses include the copper losses in the distribution lines and transformers. an increase in load demand causes an increase in current flow through the lines and the distribution transformers windings, and hence an increase in losses. such losses are significant in distribution networks because of the involvement of a large number of distribution transformers. in addition, system unbalances due to unbalanced load at the consumer side increases losses in the transformer. a small amount of copper losses also contributes to the system due to the harmonic currents. however, the high-frequency harmonic voltage has a greater contribution to the core losses of the transformer (al-badi ah et al. 2011, daut i et al. 2013). depending upon only the initial cost of the transformers is not an economical choice in buying the efficient distribution transformers. the losses in distribution transformers, especially the load loss occurs based on the load pattern, which is variable during its operational lifetime (al-badi ah et al. 2011). therefore, the transformer buyer needs to evaluate the no-load and load losses capitalization values for their requisition transformers over the transformer lifespan. the capitalization values of the transformer depend on system capacity cost, levelized cost of the energy, load profile and the economic consideration (wijayapala wdas et al. 2016). the detailed model of calculating capitalization values for distribution transformer losses are presented, and the model is applied to capitalization values for medc distribution transformers. further, the distribution system losses can be alleviated by increasing the conductor size (zhu z et al. 2016). however, increasing the conductor size without engineering sense may increase the cost and losses. therefore, an economic a. al-badi, r. ahshan, s. al-hinai, a. moosa, r. shah, m. al hasani, and s. khan 102 optimal conductor size needs to be selected. in this study, the differences in losses are not compared with the total material and labour costs for building/rebuilding the line; instead, the differences in the cost of losses for various conductors are compared with only construction costs above those required to build the line with the smallest suitable conductor. the saving resulted from loss reduction does not only reflect on the financial aspect of the saved energy but also reflects on releasing the system capacity that can reduce the requirement of system development and lessen the deteriorating of system components. in the usa, the average losses in transmission and distribution systems are around 7.5%, whereas, the average losses in the distribution system only are about 6% (inan h et al. 2014). the total losses in a distribution system for one of the distribution companies in oman reached to 6.92% in 2018 (medc (2018)), which is very close to the reported percentage in the usa. the energy loss in the main interconnected system (mis) in oman reached 1.43% in 2018 (oetc 2019), which is well within international norms. in order to quantify the technical losses in medc distribution network, the model of a selected 33 kv, 11 kv and lv network of a representative primary substation is developed in this work. the distribution network components (transformer, cable, and load) model parameters are obtained from the concerned company using the system single-line diagrams (sld). the transformer parameters include kva rating, impedance ratio, and rated voltage. the cables are modelled using the cable type, length and cross-section area. the loads are modelled based on the collected data that is distributed in each feeder equally to all loads. etap software is used to implement and simulate the network model. this introduction is followed by presenting the methods of minimizing losses in section 2. section 3 presents the capitalization of losses for distribution transformers. section 4 discusses the optimal conductor size selection for 11kv and lv networks. modelling of a selected 3kv, 11kv and an lv network representative primary substation is presented in section 5. section 6 summarizes the main conclusions of this study. 2. losses reduction methods several ways are available to alleviate losses in the distribution network (al-sarmi, s et al. 2019). however, some mechanisms require additional equipment to be installed in the system that can increase the financial burden for companies. several devices such as fluorescent lamps, distribution transformers at no-loads, induction motors with light or no loads in the distribution networks can lead to poor power factor in the system. power factor improvement using capacitors is an effective method, which helps in reducing distribution system losses and maximizing the revenue. power factor improvement can result in a reduction in the phase angle difference between the voltage and the current. the greater part of the loads in the distribution system is the inductive type, which needs reactive power to work. installing a capacitor bank in parallel with the loads provides them with the necessary reactive power that lowers the phase angle difference between the voltage and current. installing a capacitor bank can reduce the upstream current flow through the distribution lines, thus, reduces losses in the system. another benefit of installing a capacitor bank is reducing voltage drop during heavy load periods. such a capacitor bank installation requires determining the location of the capacitor bank placement, along with their types and proper sizes (samineni s et al. 2010). switching optimization is also known as reconfiguration. it is a way of relocating the switching devices that already exist or introduce new devices in the appropriate location depending upon the size of loads, and the length and size of conductors. an effective method, switching optimization, helps in reducing technical losses in the distribution networks and improves its security. compared to the method of reconductoring or new installation of feeders, switching optimization has been found as one of the most cost-effective methods to reduce the technical losses. although the switching optimization method needs several new devices, the low-cost switching devices makes this method cost-effective compared to the reconductoring or new feeder installation method (phetlamphanh v et al. 2012). the selection of an appropriate conductor size can reduce the losses. the losses in the conductors depend on the connection quality at each end of the conductor, conductor size relative to the amount of current it carries, and the conductor operating voltage level. the line loss is inversely proportional to the conductor size and directly proportional to the square of the current that passes through the conductor. a smaller size of conductor can result in higher i2r losses and a higher voltage drop that causes a loss of credit for the dso. the suggested practice is to ensure that the conductor is capable of delivering the peak demand of the consumers at the standard voltages. in other words, the voltage drop has to maintain within the standard range (aburn g, and hough m 2015). the shorter length distribution network can reduce distribution losses as studied in (sadati smb et al. 2012). current density and heuristic index-based approximated optimal solution for conductor size selection for radial distribution system has been presented in (wang z. et al. 2000). the main losses in distribution sub-stations are transformer losses. two types of transformer losses: one is core (no-load) losses that are typically 25% to 30% of the total distribution losses and independent of the load (al-sarmi, s et al. 2019). this loss varies with the transformer size and the materials that are used to manufacture the transformer. the other one is analysis of loss reduction techniques for low voltage distribution network 103 the copper losses that mainly depends on the magnitude of current passes through the windings of the transformer, and it dissipates as heat. with the increase in loads, the material behaves like more resistive and hence increases line losses. for better management of distribution transformers, it is recommended to de-energize the transformer one or more times at low-load periods, which can help to reduce excessive core losses. similarly, the distribution transformers need to switch them on at high-demand periods to reduce excessive copper losses. furthermore, the dso needs to identify customers having premises connected to oversized (or undersized) line transformers so that the dso can optimize the transformers’ sizes. an undersized transformer serving a particular load can operate with high losses. the voltage upgrade is the changeover of lines and substations to the higher voltages. this method has developed practically to meet load growth or transmission requirements. it has several advantages in addition to the practical aspect; the economic advantages may be a benefit, especially if some of the used equipment can be used again with minimal modifications (panek j, elahi h 1989). this method plays a very important role in alleviating losses in case of the distribution network is overloaded. the best method in load balancing is to utilize the current duration curve, which can be developed for all three phases by the distribution system planner. accordingly, if the loads in each phase of the distribution line are re-distributed, the losses can be minimized (al-sarmi, s et al. 2019). a tap-changing transformer allows adjusting the voltage level by altering turns the number of the transformer winding using a tap changer. in order to achieve a controllable voltage level, taps are normally adjusted on the hightension side of the transformer. off-load and on-load types of tap changing transformer are commonly utilized in the distribution network (al-sarmi, s et al. 2019). with demand-side management, the dso can reduce the overall system load, especially during peak periods, by turning off particular types of load or catering some stimulus for customers. customer motivation to use smart and high efficient motors, refrigerators, and lighting systems can reduce the overall load (al badi ah et al. 2020). advanced metering infrastructure (ami) allows automated and bi-directional interaction between the smart utility meter with its ip address and the distribution company. the purpose of an ami architecture is to update periodically about the real-time data related to power consumption to both the distribution companies and the consumers. it helps the customers to schedule their appliances to operate at the time of the best price and hence can reduce their cost of energy consumption (al badi ah et al. 2020). 3. research methodology some of the loss reduction methods unfolded in section 2 have been used in the medc network. such methods are capacitor banks for power factor improvement, transformer size and location, and tapchanging transformers. recently, the ami method is currently under installation in the network. however, there is an opportunity to reduce the distribution network losses by selecting economic optimal conductors and cables and account for the cost of losses for the transformer load and no-load losses during the time of buying a new transformer. the following subsections present the detailed models to calculate losses in a representative distribution network, models for calculating optimal conductor and cable selection, and models for determining no-load and load losses capitalization values of distribution transformers. 3.1 system modelling and losses quantification a load flow study is conducted to quantify the losses in a representative distribution network. the load flow model uses actual data for load, lines, transformers, and short circuit capacity, which are given by the medc. the load flow model solves the following power balance equations to determine voltages at the different buses, real and reactive power flows through the lines (saadat h 2011). ( ) 1 cos n k k i ki k ki i i p v v y δ θ δ = = − −∑ (1) ( ) 1 sin n k k i ki k ki i i q v v y δ θ δ = = − −∑ (2) the non-linear power balance eqns. (1) and (2) can be solved by different iterative methods such as the newton-raphson, the gauss-seidel, and the fastdecoupled methods. in this study, the newtonraphson method available in the etap software package is applied to solve the load flow equations. after solving the load flow problem, the system losses are computed by calculating the losses in any branch k-i using the following equations (albadi m et al. 2017). , , ,loss ki loss ki loss ki ki iks p jq s s= + = + (3) * * and ki k ki ik i iks v i s v i= = (4) 3.2 optimal cable and conductor selection increasing the cross-sectional area of the conductors/cables can reduce the energy losses in the cable; however, the large size conductors cannot be a choice for the dso. this subsection presents a simple method of selecting the most economical a. al-badi, r. ahshan, s. al-hinai, a. moosa, r. shah, m. al hasani, and s. khan 104 conductor/cable sizes among the available sizes given by medc (booth et al. 1988). the most economical conductor/cable size is determined by comparing the cost of losses for various conductors/cables with their extra construction/material cost considering peak load conditions. the total peak power loss for three phases are calculated using the following equation, 23p i r= (5) the annual demand cost per kw of peak losses is calculated as dc wecc c n= × (6) the annual energy cost per kw of peak losses is calculated as ec kwh fc c l h= × × (7) fl is the loss factor and can be determined as ( ) ( )fl a lf b lf= × + × (8) where lf is the load factor, 0.2a = , and 1 .b a= − lf is defined as the ratio between the average load and the peak load for a given load pattern. the total annual loss cost for per kw peak losses is determined as / ( )tac kw ec dcc c c= + (9) the total annual loss cost for total peak losses is determined using the following equation ( )tac ec dcc c c p= + × (10) the cost of annual loss savings for a cable/conductor is calculated as , ,( )as tac ps tac ssc c c= − (11) the net annual savings for a cable/conductor is determined as , ( )net as as accc c c= − (12) the maximum net annual savings of a cable/conductor gives an indication about the most economical cable/conductor for a given annual peak load. 3.3 capitalization of losses for distribution transformers the dsos are always in the process of buying and installing distribution transformers because of the continuous expansion of the electrical distribution networks. transformer economics is heavily connected to the pricing of the energy losses, e.g. load and no-load losses that occur during the transformer operation. the cost of such losses is important to calculate over the lifetime of the distribution transformers (szwander w (1945)). the transformers with low initial cost may increase the cost of such losses over its lifetime, and vice versa. therefore, the transformer-purchasing group needs to determine the cost of losses for the duration of the transformer lifespan in evaluating the most economical distribution transformer. no-load losses capitalization value refers here as the value of one unit power loss in a distribution transformer under the no-load condition for the transformer lifetime. no-load losses capitalization value depends on the system capacity cost, and cost for generating, transmitting and distributing energy. no-load losses capitalization value is computed using the following equation (al-badi ah et al. 2011, charalambous ca et al. 2013, and wijayapala wdas et al. 2016). 1000 sc e t fc f c hc x r iη   + =        (13) load losses capitalization value refers here as the value of one unit power losses in a distribution transformer under load condition for the transformer lifetime. the load pattern, growth of the load, and nature of the load profile are the main reasons for varying load losses capitalization value. load losses capitalization value depends on the system capacity cost, the cost for generating, transmitting and distributing energy, yearly loss factor, peak responsibility factor, yearly peak load, and transformer fixed charge rate. load losses capitalization value can be determined as (al-badi ah et al. 2011, charalambous ca et al. 2013, and wijayapala wdas et al. 2016). 2 2 1000 sc f e f pl t fc f c r hc l y p r iη   + = ×       (14) analysis of loss reduction techniques for low voltage distribution network 105 the uniform annual peak load that can be determined based on ieee loss evaluation guide as given in (c57.120-2017 standard). pl f lf p l = (15) the increasing factor is calculated as (wijayapala wdas et al. 2016) purchase cost + overhead + taxes purchase costf i = (16) 4. results and discussion 4.1 system modelling and losses quantification this study model a selected 33 kv, 11 kv and lv network of a representative primary substation of medc network. the network consists of three 20mva, 33/11kv transformers, 15 feeders and three 5mvar capacitor banks. five feeders, namely fdr4, fdr5, fdr6, fdr7 and fdr8, are modelled in details up to the low voltage (415v) level. the components model such as transformers, lines, capacitors banks, loads and grid available in the etap software package are utilized to model the selected network. the parameters such as kva rating, impedance ratio, rated voltage for the transformers, type, length and cross-section for the cable, type and amount of load, short circuit capacity for the grid are obtained from the medc. figure 1 shows the etap model for the selected 33 kv, 11 kv and lv network, while figure 2 demonstrates the detailed model for one of the feeders fdr6. the power flow problem is solved for the developed model at different load conditions. such load conditions are maximum load, 80% of maximum load, 70% of maximum load, and minimum load. the load current drawn by each feeder in the developed model is comparable with the field-recorded load current for the corresponding feeder during peak load. the closeness of both current values verifies the accuracy of the developed model. the load flow solution provides voltage profiles for each bus, line flows and losses occur in the network components. figure 3 shows the voltage profile for the selected buses for different loading conditions. it reveals that voltages for all selected buses are within the recommended limit ( ± 6%) except one bus bus162fdr4, which has under-voltage at the maximum loading and 80% of the maximum loading. it also reveals that buses such as bus r4/31 in fdr5 and bus 167 in fdr7 maintain the minimum level of the required voltage during maximum loading. the total kw and kvar losses for distribution networks at different load conditions are depicted in figure 4. an increase in the load results in an increase in both the real power losses and reactive power requirements for the system. this result demonstrates that the system faces more losses during the peak demand, especially during the summertime. table 1 shows real and reactive power losses for different feeders. it is important to mention that real power losses are significant in the cables/lines because the current medc network has fatter cables than required. the use of the optimal size of the conductor can reduce these losses. the transformers also cause real power losses; however, the reactive power losses due to the transformer is significant as table 1 revealed. these losses influence the transformer’s total owning cost as the losses vary based on loading. since the total owning cost has two major components such as the costs of load and no-load losses, therefore, the determination of capitalization values of losses at the time of buying a new transformer can reduce such losses during the transformer lifespan. 4.2 optimal cable and conductor selection for 11kv networks 4.2.1 economic optimal cable selection table 2 shows a summary of the detailed calculation for economical cable selection based on peak load conditions—the study conducted for cable sizes, as mentioned in the first column of table 2. the cable resistances are obtained from the datasheet found in the oman cable website for 3c xple copper cable. the annual peak load for the cable is assumed to be 120 ampere, which can be adjusted by the distribution company as they required. the total annual loss cost for the total annual peak losses is calculated using eqn. (10). the results are tabulated in table 2. the total annual loss cost per kw peak loss is calculated using eqns. (5)-(9). the results have been presented in table 3. the extra cost compared to the smallest size (70 sq.mm) of the cables is obtained from the medc. the annual carrying charge on extra construction cost is assumed to be 12% as found in (booth et al. 1988). the cost of annual loss savings of each cable and the net annual savings for all cables have been obtained using eqns. (11) and (12), respectively. all these calculated results are tabulated in table 2. figure 5 shows the net annual savings versus the size of the cables. it reveals that 185 sq.mm is an optimal economic choice for carrying annual peak load current 120 ampere and at a 0.5 load factor. it also indicates that the variation of the load factor, while keeping the peak load current constant, does not affect the cable size variation. in other words, the cable size remains the same as the optimal one; however, there a change in the amount of net annual savings is observed. on the other hand, it reveals that the optimal economic size of the cable can be changed if the peak load current is far from the expected peak load current. a. al-badi, r. ahshan, s. al-hinai, a. moosa, r. shah, m. al hasani, and s. khan 106 figure 1. components modelling in etap for the selected 33 kv, 11 kv and lv network. analysis of loss reduction techniques for low voltage distribution network 107 figure 2. feeder 6 (fdr6) detailed model that shows each low voltage (11kv/415v) distribution transformer and load. a. al-badi, r. ahshan, s. al-hinai, a. moosa, r. shah, m. al hasani, and s. khan 108 figure. 4. total losses (kw and kvar) in the distribution network for different loading conditions. table 3. annual demand cost and energy cost detail. wholesale electricity purchase cost/kw-month (omr) 17.7 annual demand cost/kw (omr) 212.4 load factor 0.5 loss factor 0.3 cost/kwh (omr) 0.015 annual energy cost/kw (omr) 39.42 total annual loss cost/kw(omr) 251.8 figure. 3. voltage profiles for the selected buses at different loading conditions. table 1. real and reactive power loses in distribution transformers and cables for different feeders. kw losses (fdr kvar losses (fdr4) kw losses (fdr5) kvar losses (fdr5) kw losses (fdr6) kvar losses (fdr6) kw losses (fdr7) kvar losses (fdr7) kw losses (fdr8) kvar losses (fdr8) cables/line 159.7 -28.3 81.8 -19.4 42.5 -60.8 81.8 -56.2 46.6 -39.6 transformers 17.0 76.2 17.2 73.4 11.4 59.8 11.3 58.2 8.5 44.8 total losses 176.8 47.9 99.2 53.9 53.9 1.0 93.1 2.1 55.5 5.2 table 2. cost of loss and net annual savings for various sizes of cables compared to the smallest size. cable sizes (sq. mm) cable resistance (ohm/km) total peak loss for three phases (kw) total annual loss cost (omr) extra construction costs in compared to the smallest size conductor (omr) annual carrying charge on extra construction cost (omr) annual loss savings compared to the smallest size conductor (omr) net annual savings (omr) 70 0.3420 14.77 3720.49 0.00 0.00 0.00 0.00 120 0.1960 8.47 2132.21 4100.00 492.00 1588.28 1096.28 150 0.1590 6.87 1729.70 6757.00 810.84 1990.79 1179.95 185 0.1280 5.53 1392.46 8698.00 1043.76 2328.03 1284.27 240 0.0982 4.24 1068.28 15099.00 1811.88 2652.21 840.33 300 0.0794 3.43 863.76 18300.00 2196.00 2856.73 660.73 400 0.0636 2.75 691.88 23100.00 2772.00 3028.61 256.61 analysis of loss reduction techniques for low voltage distribution network 109 figure. 5. annual net savings variation for different sizes of cables for annual peak load current 120 a, 0.5 load factor, where the optimal size is 185 sq. mm. figure. 6. annual net savings variation for different sizes of overhead conductors for annual peak load current 120 a, 0.5 load factor, where the optimal conductor size is 200 sq. mm. 4.2.2 economic optimal overhead conductor selection the detailed calculation is carried out using eqns. (5) – (12) as described for the economic optimal cable selection, and the detailed results are presented in table 4. table 4 shows a summary of the detailed calculation for economic optimal conductor selection based on peak load conditions. the study conducted for conductor sizes is mentioned in the first column of table 4. the conductor resistances are obtained from the datasheet found in the oman cable website for acsr conductors. the annual peak load for the conductors is assumed to be 120 amperes, which can be adjusted by the distribution company as they required. figure 6 shows the net annual savings versus the size of the conductors. it is found that 200 sq.mm is an optimal economic choice for carrying annual peak load current 120 ampere with a 0.5 load factor. 4.3 capitalization of losses for distribution transformers 4.3.1 capitalization value for no-load loss the capitalization value for no-load loss (x) is determined using eqn. (13). considering the distribution transformers as fixed assets (wijayapala wdas et al. 2016), the fixed charge rate is taken by combining the opportunity cost or minimum acceptable return as 12.75% and the book depreciation as 4%. since there is no income tax in the country, thus it is taken as zero; however, local property tax and insurance is considered as 0.1 %. thus, the fixed charge rate is found at 16.85%. since the overall system loss for medc is 6.92%, thus the system efficiency is considered as 93.08%. equation 16 is used to calculate the increasing factor considering there is no overhead fee and taxes, and the factor is found 1. the system capacity is taken approximately as 40 omr/kw/year as per (crt 2017). the average cost of energy for the year 2018 is considered to calculate the levelized cost over the transformer lifetime (capability statement 2018). the discount rate and the inflation rate are considered as 7.5% and 6.425%, respectively. the cost recovery factor is utilized to calculate the annual levelized cost of energy. the levelized annual cost of energy is calculated as 0.0296 omr/kwh. thus, the no-load loss capitalization value was found 2.0467 omr/w, as shown in table 5. the calculated value in this study is higher than the value given in the oman electricity standard (oes) as 0.8 omr/w. such a discrepancy may arise due to the differences in the load pattern, the system capacity cost, and the other parameters that were assumed in this study. 4.3.2 load loss capitalization value the load loss capitalization value (y) is determined using eqn. (14). the loss factor is determined using the load factor of the transformer used in the distribution network. in order to determine the transformer load factor, it is essential to know the transformer load pattern even for a short period like one day during system peak load. with the use of the load profile of a distribution transformer, the load factor is determined by the ratio between the average load and the peak load. the load factor was calculated as 47.3% for a 2mva transformer using the load profile provided by the medc. with this load factor, the loss factor was calculated as 0.2749. the peak responsibility factor was determined based on the ieee loss guide (c57.120-2017) and was found as 0.9375. the uniform annual peak load is calculated using eqn. (15) as 0.9047. all calculated parameters are substituted to the eqn. (14) to calculate the capitalization value of load loss, and it is found as 0.5959 omr/w, as shown in table 5. the calculated value in this study is higher than the value given in the oes (0.3 omr/w). such a discrepancy may arise due to the differences in the load pattern and transformer type (urban or rural), the system capacity cost, and the other parameters that are assumed in this study. 0 200 400 600 800 1000 1200 1400 50 100 150 200 250 300 350 400 a n n u a l n e t sa v in g s (o m r ) cable sizes (sq. mm) 0 500 1000 1500 2000 2500 50 100 150 200 250 300 350 400 a nn ua l ne t sa vi ng s (o m r ) conductor sizes (sq. mm) a. al-badi, r. ahshan, s. al-hinai, a. moosa, r. shah, m. al hasani, and s. khan 110 table 5. no-load and load loss capitalization values of a 2 mva distribution transformer. capitalization values omr/w no-load loss 2.0467 load loss 0.5959 5. conclusion distribution companies consider the power loss in the distribution lines as a serious problem because of the energy and money wasted. this paper has presented different methods to minimize such losses. distribution system losses cannot be phased out; however, it can be decreased by appropriate planning of systems to assure that power loss remain within an acceptable range. this paper reveals the following conclusion and recommendation for reducing losses in the distribution system and hence increasing the system efficiency: • losses in the distribution system can be quantified by modelling and simulating the network. such a kind of study reveals the components or subsystems that contribute to higher losses in the system. through modelling and simulation of a representative distribution network, it was found that transformers and cables/lines contribute higher losses among all other components in the system. • determining an economic optimal conductor/ cable from the available sizes by comparing the cost of losses for various conductors with their extra construction/material cost can reduce the losses in the cables/lines. the study is carried out based on peak load conditions instead of average load conditions because the losses change with the square of the current. it is shown that 185-sq.mm cable is an optimal economic choice for carrying an annual peak load current 120 ampere with a load factor 0.5. at the same time, 200-sq.mm, an overhead conductor is found as an optimal economic choice for carrying annual peak load current 120 ampere with a load factor 0.5. • capitalization values of the losses have a significant influence in evaluating the most economical distribution transformer. the utility can assess the capitalization value of the no-load and load losses using the measured load profile of various distribution transformers considering their lifetime. the utility can provide such values to the transformer manufacturing company and ask to design the requisition transformers accordingly. it was found in this paper that both the no-load and load loss capitalization values are higher than the values recommended by the oman electrical standards (oes). such differences may arise due to the differences in the variables such as the load pattern, load factor, transformer type (urban or rural), and the cost parameters, such as the system capacity cost, unit cost of energy, discount and inflation rates considered in the current study and the study done before. this study suggests recording the actual load profile of a particular distribution transformer for a period of time to determine its loss of capitalization values. • regular inspection and maintenance of distribution equipment such as isolators, connections, transformer and transformer bushings, lt switches and dropout fuses are required. optimal location and appropriate size of the distribution transformers is an important factor. this study suggests placing the distribution transformers closer to the load centre and maintaining the minimum number of transformers, if possible. • the feeders can directly supply heavy loads such as large buildings and industrial loads. load balancing and appropriate load management can be considered for further reduction of the losses. the integration of digital, tamper-proof meters can help to reduce non-technical losses. in addition, the dso can operate their system at a higher power factor. these suggestions can be further investigated for the network presented in this research. table 4. cost of loss and net annual savings for various sizes of conductors compared to the smallest size. conductor sizes (sq. mm) conductor resistance (ohm/km) total peak loss for three phases (kw) total annual loss cost (omr) extra construction costs in compared to the smallest size conductor (omr) annual carrying charge on extra construction cost (omr) annual loss savings compared to the smallest size conductor (omr) net annual savings (omr) 70 0.4156 17.95 4521.16 0.00 0.00 0.00 0.00 100 0.2885 12.46 3138.48 1681.70 201.80 1382.67 1180.87 150 0.193 8.34 2099.57 4737.10 568.45 2421.58 1853.13 200 0.1439 6.22 1565.43 6649.10 797.89 2955.72 2157.83 400 0.0712 3.08 774.56 20510.00 2461.20 3746.60 1285.40 analysis of loss reduction techniques for low voltage distribution network 111 conflict of interest the author declares no conflicts of interest. funding muscat electricity distribution company provided financial support to perform this research work. acknowledgment the first three authors would like to acknowledge squ and muscat electricity distribution company for the financial support received to perform this research work. references albadi m, soliman h, awlad thani m, al-alawi a, al-ismaili s, al-nabhani a, baalawi h (2017), optimal allocation of pv systems to minimize losses in distribution networks using ga and pso: masirah island case study. j. electrical systems, 13(4): 678-688. al-sarmi s, al-ansari h, al-badi a, ahshan r, alhinai s, al-hadi hilal (2019), “designing a system to minimize power losses in distribution network”, in proc. 15th international conference for gcc cigre, muscat, sultanate of oman, pp. 432-440. aburn g, hough m (2015), implementing epa’s clean power plan: a menu of options. technical report, national association of clean air agencies (nacaa). al-badi ah, elmoudi a, metwally i, al-wahaibi a, al-ajmi h, al bulushi m (2011), “losses reduction in distribution transformers”, in proc. of the international multiconference of engineers and computer scientists, hong kong, pp. 1-5. al-badi ah, ahshan r, hosseinzadeh n. ghorbani r, hossain e (2020), survey of smart grid concepts and technological demonstrations worldwide emphasizing on the oman perspective. appl. syst. innov. 3, 5. booth and associates, inc. (1988), distribution system loss evaluation manual. technical report, the university of michigan, usa. charalambous ca, milidonis a, lazari a, nikolaidis ai (2013), loss evaluation and total ownership cost of power transformers—part i: a comprehensive method. ieee transactions on power delivery, 28(3):1872-1880. c57.120-2017 ieee guide for loss evaluation of distribution and power transformers and reactors. crt (2017), cost reflective tariff (crt) rates as per (2017), -muscat electricity distribution company, oman. capability statement (2018), distribution system capability statement, muscat electricity distribution company, 2019 – 2021. daut i, hasan s, taib s (2013), magnetizing current, harmonic content and power factor as the indicators of transformer core saturation. journal of clean energy technologies, 1(4): 304-307. emmanuel m, rayudu r, welch i (2017), grid capacity released analysis and incremental addition computation for distribution system planning. electric power system research, 152: 105-121. inan h, batson j, scheibe m (2014), systems loss reduction. leidos, techadvantage. medc (2018), muscat electricity distribution company, annual report, muscat, oman. oetc 2019, five-year annual transmission capability statement (2019-2023). technical report, oman electricity transmission company. phetlamphanh v, premrudeepreechacharn s, ngamsanroaj k (2012), "technical losses reduction of electrical distribution system in vientiane capital," in proc. international conference on renewable energy research and applications (icrera), nagasaki, pp. 1-6. panek j, elahi h (1989), "substation voltage upgrading," in ieee transactions on power delivery, vol. 4, no. 3, pp. 1715-1724. samineni s, labuschagne c, pope j (2010), "principles of shunt capacitor bank application and protection," in proc. of 63rd annual conference for protective relay engineers, college station, tx, usa, pp. 1-14. saadat h (2011), power system analysis. psa publishing llc. szwander w (1945), the valuation and capitalization of transformer losses. journal of the institution of electrical engineers part ii: power engineering, 92(26): 125-134. sadati s. m. b., yazdani-asrami m, takami m. m. m. (2012), decreasing length of distribution network for loss reduction: technical and economical analysis. research journal of applied sciences engineering and technology 4(7): 813-818. wijayapala wdas, gamage srk, bandara hmslg (2016), determination of capitalization values for no load loss and load loss in distribution transformers. journal of the institution of engineers, srilanka, 49(3): 11-20. wang, z., h. liu, d.c. yu, x. wang, h. song (2000), a practical approach to the conductor size selection in planning radial distribution systems. ieee transaction on power delivery, 15(1): 350353. zhu z, lu s, gao b, yi t, chen b (2016), life cycle cost analysis of three types of power lines in 10 kv distribution network. inventions, 1, 20. 1. introduction 2. losses reduction methods 3. research methodology 3.3 capitalization of losses for distribution transformers 4. results and discussion 4.1 system modelling and losses quantification 4.2 optimal cable and conductor selection for 11kv networks 4.2.1 economic optimal cable selection 4.3 capitalization of losses for distribution transformers 4.3.1 capitalization value for no-load loss 4.3.2 load loss capitalization value 5. conclusion the journal of engineering research (tjer), vol. 17, no. 2, (2020) 135-141 *corresponding author’s e-mail: smmukherjee3@gmail.com doi:10.24200/tjer.vol17iss2pp135-141 development of spinel magnesium aluminate by solution combustion route using thiourea and urea as fuel srinath ranjan ghosh 1 , soumya mukherjee 2, *, and sathi banerjee 1 1 department of metallurgical & materials engineering, jadavpur university, india 2, * department of metallurgical engineering, kazi nazrul university, india abstract: in the present article spinel phase magnesium aluminate was synthesized using magnesium nitrate, alnitrate precursors in 1:2 molar ratio using urea and thiourea as fuel and reducing agent. nitrate salts were mixed in a stoichiometric ratio in distilled water with three different molar ratios of two different fuels: urea and thiourea. the temperature of crystallization was obtained after thermal analysis followed by annealing at a fixed temperature, fixed soaking period for urea as fuel while variable temperature and soaking period were required for thiourea as fuel. ftir analyses were carried out of the samples to verify the m-o co-ordinations for the phase formation. prominent octahedral m-o stretching was noted at about 609 cm-1 while that of al-mg-o stretching was noted at about 1100 cm-1 for thiourea based samples. using urea as fuel al-o stretching was noted at about 539cm-1 while that of al-mg-o vibration was noted at about 677cm-1. morphological features of the synthesized samples were observed by sem. agglomeration was noted for both urea and thiourea as fuel having irregular polygon shape. using thiourea as fuel, a bit of porous structure was noted while for urea as fuel negligible porosity was noted. keywords: magnesium aluminate; thermal analysis; phase analysis; m-o co-ordinations; morphology باستخدام عن طریق حرق المحلولالومینات المغنیسیوم تحضیر معدن السبینیل ذو یا كوقودیا والیوریوالثیور ، وساثي بنرجي*سریناث جوش، سومیا مخرجي تصنیع ألومینات المغنیسیوم في طور اإلسبنیل باستخدام نترات المغنیسیوم وسالئف النترات خالل ھذه الدراسة تم الملخص: بنسب متكافئة في الماء اختزال. تم خلط أمالح النترات ریا كوقود وعواملیوكال من الیوریا والثیو باستخدام یةموالر 2: 1بنسبة اسفرت الدراسة بعد التحلیل الحراري والتلدین للیوریا . لنوعي الوقود المذكورین اعالهالمقطر مع ثالث نسب موالریة مختلفة ة ریا وجود درجة حرارة وفتریوالثیووقود بینما تطلب استخدام ثابتتین، ریا إلى أن فترة نقع الیوریا ودرجة حرارة تبلوره یووالثیو ثماني السطوح m-oلتشكیل الطور. لوحظ تمدد m-oوتم إجراء التحلیل الطیفي للعینات للتأكد من تنسیقات ،نقع متغیرتین ریا. وفي حالة یوللعینات القائمة على الثیو 1-سم 1100عند حوالي al-mg-oبینما لوحظ امتداد 1-سم 609البارز عند حوالي سم 677عند حوالي al-mg-oبینما لوحظ استخدام اھتزاز 1-سم 539عند حوالي al-o استخدام الیوریا كوقود لوحظ تمدد من نتائج التجارب أن لوحظ. تمت مالحظة السمات المورفولوجیة للعینات المركبة بواسطة المجھر المسحي األلكتروني. كما 1كان للیوریا بینما تؤخذ بعین اإلعتبار ریا بنیة مسامیةیولثیول أنكما ،ریا لھ شكل مضلع غیر منتظمیوتكتل كل من الیوریا والثیو .بنیة أقل مسامیة .وقود، ریایوثیو، یوریا، تحلیل حراري، الومینات المغنیسیوم الكلمات المفتاحیة: development of spinel magnesium aluminate by solution combustion route using thiourea and urea as fuel 136 1. introduction magnesium aluminate has lots of industrial applications due to its high melting point (2135°c), high-temperature mechanical strength, chemical inertness, thermal shock resistance, chemical resistance, and low thermal expansion coefficient. (c păcurariu et al. 2007; s.r. ghosh et al. 2018; ali saberi et al. 2008; p v marakkar kutty et al. 2013) the presence of such properties makes the compound suitable as high-temperature refractory material for both cement rotary kiln, and steel ladle. in the initial decades, chromite-based spinel is extensively used for steel making, glass tank regenerators, rotary kiln and even for copper industry. however, environmental concern regarding cr+6 species from the chromite spinel is a hindrance to its applications. (chandrima ghosh et al. 2015) the above compound has spinel structure represented by the formula ab2o4 where a is the divalent ion, b is the trivalent ion and o represents the anion. divalent mg and trivalent al ions occupy tetrahedral (1/8 of available) and octahedral sites (1/2 of available) respectively. (shiva salem 2015) the particular spinel ceramic also has excellent optical properties and can be used for transparent ceramics. the optical property enhances its eligibility as a prominent candidate material for transparent armour and visible-infrared windows. it is also utilized for ceramic paints and for making catalyst support, membranes, dye absorbent and as sensors. for the successful application of spinel aluminate as dye absorbent, catalyst, sensors, spinel fabricated needs to possess high purity, controlled particle size, high surface area and uniform pore size distribution. (narges habibi et al. 2017) spinel magnesium aluminate also acts as a potential candidate for humidity sensors. in recent researches, it is noted that spinel magnesium aluminate as one potential candidate for photocatalyst to decompose reactive red methylene used as a dye for industrial operations. (mostafa y nassar et al. 2014) purity, particle size, chemical homogeneity, and reactivity of spinel magnesium aluminate are influenced by the synthesis route. (ali saberi et al. 2008) the material is found to be synthesized by various methods in addition to solid-state reaction like sol-gel (debsikdar j.c 1985; naskar m.k. et al. 2005) precipitation (li j-g et al. 2000), aerosol method (yang n et al. 1992), co-precipitation (guo j et al. 2004), combustion synthesis, freeze-drying, decomposition of an organometallic complex in super critical fluids, hydrothermal route, plasma spray decomposition of powders, (bickmore r clint et al. 1996; bratton r.j. 1969; barj m. et al. 1992; pommier c. et al. 1990; yang ning et al. 1992; varnier olivier et al. 1994) microwave-assisted combustion route (torkian leila et al. 2011), polymerized complex method (lee p.y et al. 2006; du xuelian et al. 2014), mechanochemical route (domanski d. et al. 2004), self-propagating high-temperature synthesis (gorshov v. a. et al. 2017) and others. in the present article, spinel magnesium aluminate is synthesized using urea and thiourea as fuel and reducing agent for three molar ratios along with nitrate precursors. the annealing temperature is confirmed after carrying thermal analysis of mixed solutions followed by phase analysis, bonding analysis and morphological studies. 2. experimental methods ar grade of magnesium nitrate, aluminium nitrate, urea and thiourea were used as precursors for synthesizing spinel. the stoichiometric ratio of magnesium nitrate, aluminium nitrate were taken in 1:2 molar ratios in distilled water for stirring. the solution undergoes stirring by magnetic beads for sufficient time. after proper mixing of nitrates urea, thiourea was added in 1.25, 1.50 and 1.75 molar ratio with respect to nitrate salts and again undergoes stirring for about 1 hour to ensure homogeneity. the resultant solution undergoes drying in an oven (heater) at about 80°c for about 3-4 hours to obtain a dry gel-like mass. the gel was put into dta-tga analyzer (diamond pyris, perkin elmer) in presence of nitrogen atmosphere at the heating rate of 10°c/min for determining the thermal characteristics of the sample. crystallization temperature for spinelization was obtained and after that, the gel was put for annealing at 700°c for 5 hours in the case of urea as fuel. the annealing temperature was 700°c for 5 hours, 800°c for 4.5 hours and 900°c for 4.5 hours respectively for thiourea as fuel. phase analysis was carried by xrd (rigaku, ultima iii) having cu kα wavelength of 1.54å, 40kv with a scan range of 10-80° having a scan rate of 5°/minute. bonding analysis was carried by ftir (ir prestige-21, shimadzu) to determine m-o coordination of the required phase after preparing pellet samples with kbr. morphological structures were analyzed by sem (jeol, jax 840a) using carbon conducting tape to stick the powder sample for better resolution and to avoid electrostatic charging. 3. results and discussions dta-tga curve exhibits initial weight loss in the temperature range between 100-200°c. from the curve in fig. 1 two prominent endothermic peaks at about 100°c and 200°c are noted. first one suggests the removal of physically absorbed water and 2nd one suggests removal of structural water. a large exothermic peak is observed due to burning or oxidation of organic compounds which is mainly mainly a combustion of carbonaceous and sulphur 0 100 200 300 400 500 600 700 800 900 -40 -20 0 20 40 60 temperature o c he at fl ow (m w ) 4 5 6 7 8 9 10 11 12 weight (m g) figure 1. dta-tga curve of magnesium nitrate: alnitrate precursors in 1:2 molar ratio with thiourea as fuel and reducing agent having a heating rate of 15°c/min. srinath ranjan ghosh , soumya mukherjee, and sathi banerjee 137 elements due to thiourea corresponding to 2nd prominent weight loss. a minor endothermic peak is observed around 350°c due to the initial formation of γ-alumina. (macêdo maria laponeide fernandes et al. 2007) the second endothermic peak with minor weight change is noted due to possible dissociation of nitrate precursor salts at about 450°c (macêdo maria laponeide fernandes et al. 2007). a broad weak exothermic peak ranging from 500°c to 700°c is noted which corresponds to the initiation of the crystallization process and records the third weight loss. crystallization of gel from precursors always involves weight loss and undergoes exothermic reaction. weight loss occurs possibly because of the presence of the multicomponent compound and their interactions. the stability of weight loss indicates full onset of the crystallization process. figures 2-7 represent xrd spectra of the synthesized sample using urea and thiourea respectively as fuel. three different molar ratios of 1.25, 1.50 and 1.75 are considered for each fuel respectively. it has been noted that for spinel magnesium aluminate synthesized by urea annealing temperature and soaking period are kept fixed. for such conditions, spinel has been noted for all molar ratios of urea as fuel, reducing agent. peak intensity is optimum for all cases. slight amorphous nature of spectra is noted for all cases. lower urea fuel ratio of 1.25 exhibits much better crystalline nature of spectra than higher fuel ratio. it may be possible due to higher reduction tendency and calorific value of fuel with more molar ratio which could also possibly accelerate the reaction and shorten the transformation period to induce the required phase. in contrast with thiourea as fuel, the reducing agent annealing temperature is increased. annealing temperature and soaking period carried are carried as follows 700°c for 5 hours, 800°c for 4.5 hours and 900°c for 4.5 hours, respectively. the crystallinity of peaks got increased with temperature for thiourea but in the present research the focus is on lower temperature spinelization hence for both fuels 700°c for 5 hours is taken as reference. for both cases, successful spinel phase formation is noted. for urea, spinel formation is verified by comparing the xrd spectra with jcpds card file #01-077-1193, #01-075-1800, #01-082-2424 and #01-077-0437 respectively. for thiourea based fuel spinel is verified by comparing the xrd spectra with jcpds card file #01-086-0085, #01-075-1800, #01-082-2424 and #01-077-0437 respectively. using urea as fuel all peaks are indexed as spinel phase. no intermediates or other phases are noted. major planes of growth for spinel is noted along (311), (220), (111), (400), (511) and (440) planes. a similar trend is also noted for thiourea as fuel and reducing agent. for thiourea based fuel no intermediates or presence of other phases are noted except for only spinel phase. crystallite size is calculated by scherrers formula t=0.9λ/βcosθ where t is the crystallite size, λ is the wavelength, cu kα = 1.54å, β is the full width of half mean, θ is the angle corresponding to the spectra. for 1.25, 1.75 molar ratio of urea, crystallite size is found to be about 36.52nm, while for 1.50 molar ratio it is noted to be about 48.69nm. in the case of thiourea, crystallite size is calculated and noted to be about 42.96nm, 48.69nm and 73.02 nm respectively. such variation in crystallite size is noted since annealing temperature varies. in the present context, the focus is on lower temperature synthesis henceforth, further analyses will be carried for samples synthesized at 700°c for 5 hours for both urea, thiourea as fuel where full phase development is noted from xrd analysis. figures 8-9 represent ftir spectra of spinel synthesized using urea and thiourea as fuel. for both cases, scanning range is within 450-4500cm-1 while major m-o coordinations are noted within 1000cm-1. it has been noted for the urea-based synthesis of spinel, m-o coordinations are noted mostly within 1000cm-1. al-o stretching is noted at about 539cm-1 while mg-o-al vibration is observed at 677cm-1 approximately. the ftir analysis is noted to be in correspondence with research findings by (s.r. ghosh et al. 2018; mukherjee 2020). for thiourea, m-o coordinations are noted for al-o stretching, al-mg-o stretching at about 609cm-1 and 1100cm-1 respectively. spectral peaks at about 1655cm-1, 2356cm-1 and 3311cm-1 are noted for h-o-h stretching, ch3-ch2 vibration, and o-h bonding vibration. the presence of h-o-h stretching, o-h vibration is possibly due to some physically absorbed moisture on the surface of the sample since ftir analysis is carried in a normal atmosphere without any purging of gas. figures 10 and 11 represent morphology obtained by sem analysis of spinel after annealing at 700°c for 5 hours using thiourea and urea of 1.25 molar ratios as fuel. for both fuels, agglomeration tendency is observed with the irregular polygonal shape of the agglomerated chunk. figs. 10 a and b represent morphology for thiourea based spinel samples with minor porosity on the surface and negligible interconnected porosity. figure 10 b exhibits agglomerated chunk for spinel by thiourea as fuel having irregular step around the periphery of agglomerate with some convex fracture at some portion of the chunk morphology. figures 10 a-d execute individual particulates to be spherical or bean shape to irregular polygonal shape for thiourea based fuel for spinel formation. figure 10 d represents a bit flaky structure for some portion. figure 11 b represents an agglomerate chunk with a polygonal shape having sharp edges while the size of the agglomerated chunk is noted to be about 4 μm while individual particulates are about 0.2μm. figures 11 ac execute dense compact formation of spinel using urea as fuel in compare to a bit porous structure noted using thiourea as fuel. individual particulates are spherical with agglomerate having dimensions close to 2.5μm to 5μm. individual particulates are noted to be about 0.3μm to 0.4μm in range. development of spinel magnesium aluminate by solution combustion route using thiourea and urea as fuel 138 10 20 30 40 50 60 70 80 100 200 300 400 in ten sit y c ps angle 2θ (urea)1.25700 oc 5hrs [111] [220] [311] [400] [511] [440] figure 2. xrd spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with urea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. 10 20 30 40 50 60 70 80 50 100 150 200 250 300 350 in te ns ity c ps angle 2θ (urea)1.5700 oc 5hrs [111] [220] [311] [400] [511] [440] [533] figure 3. xrd spectra of magnesium nitrate: alnitrate precursors in 1:2 molar ratio with urea as fuel and reducing agent having 1.50 molar ratio after annealing at 700°c for 5 hours. 10 20 30 40 50 60 70 80 50 100 150 200 250 300 350 in te ns ity c ps angle 2θ (urea)1.75700 oc 5hrs [111] [220] [311] [400] [511] [440] [533] figure 4. xrd spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with urea as fuel and reducing agent having 1.50 molar ratio after annealing at 700°c for 5 hours. 0 10 20 30 40 50 60 70 80 90 0 100 200 300 400 500 in te ns ity cp s angle 2θ (thio)1.25700 oc 5hrs [111] [220] [311] [400] [511] [440] [533] figure 5. xrd spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with thiourea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. 10 20 30 40 50 60 70 80 100 200 300 400 in te ns ity cp s angle 2θ (thio)1.5800 oc 4.5 hrs [111] [220] [311] [400] [511] [440] [622] figure 6. xrd spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with thiourea as fuel and reducing agent having 1.25 molar ratio after annealing at 800°c for 4.5 hours. 10 20 30 40 50 60 70 80 200 400 600 800 in te ns ity c ps angle 2θ (thio)1.75900 oc 4.5 hrs [111] [220] [311] [400] [422] [511] [440] [533] figure 7. xrd spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with thiourea as fuel and reducing agent having 1.25 molar ratio after annealing at 900°c for 4.5 hours. srinath ranjan ghosh , soumya mukherjee, and sathi banerjee 139 1000 2000 3000 4000 0.0 0.1 0.2 0.3 0.4 0.5 ab so rb an ce (a .u ) wave number (cm-1) (urea)1,25 700 oc 5hrs539 677 1403 3435 figure 8. ftir spectra of magnesium nitrate: al-nitrate precursors in 1:2 molar ratio with urea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. 1000 2000 3000 4000 0.0 0.1 0.2 0.3 0.4 ab so rb an ce (a .u ) wave number (cm-1) (thio)1.25 700 oc 5hrs 1100609 1655 2356 3311 figure 9. ftir spectra of magnesium nitrate: al nitrate precursors in 1:2 molar ratio with thio urea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. figure 10. sem morphology of spinel magnesium aluminate using thiourea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. a d c b development of spinel magnesium aluminate by solution combustion route using thiourea and urea as fuel 140 figure 11. sem morphology of spinel magnesium aluminate using urea as fuel and reducing agent having 1.25 molar ratio after annealing at 700°c for 5 hours. conclusion spinel magnesium aluminate was prepared using thiourea and urea as a fuel and reducing agent. thermal analysis of precursors along with fuel indicate the onset of crystallization in the range of 500-700°c. the goal of low-temperature spinelization was achieved for both cases as the temperature was about 700°c with 5 hours soaking period and confirmed from xrd phase analysis. ftir analysis confirmed the m-o coordinations and it was noted that al-o coordination was about 539cm-1 while mgo-al vibration was observed at about 677cm-1. agglomerated mass with irregular polygon shape was noted as morphological features after using both urea and thiourea as fuel, reducing agent for synthesizing spinel magnesium aluminate. using thiourea as fuel, the dimension of the agglomerated chunk was about 4μm and for individual particulate, it was about 0.2μm. similarly, using urea as fuel agglomerate was in the range of 2.5 to 5μm and for individual particulate was about 0.3 to 0.4μm. thiourea induces a slight porosity in structure while negligible dense mass was noted using urea as fuel. conflict of interest the authors declare that there are no conflicts of interest as regards this article. funding there is no funding from any external agency to carry out the experiment acknowledgment the authors would like to thank the department of metallurgical & material engineering for providing technical support to carry out dsc-tga analysis, xrd, ftir and morphology studies by sem. references: barj m, bocquet j.f, chhor k, pommier c, (1992), submicronic mgal2o4 powder synthesis in supercritical ethanol. journal of materials science 27: 2187-2192. bickmore r clint, waldner f. kurt, treadwell r. david, laine m. richard, (1996), ultrafine spinel powders by flame pyrolysis of a magnesium aluminium double hydroxide. journal of the american ceramic society 79(5): 1419-1423. bratton r.j, (1969), coprecipitates yielding mgal2o4 powders. american ceramic society bulletin 83: 759-762. domanski daniela, urretavizcaya guillermina, castro j. facundo, gennari c. fabiana, (2004), mechanochemical synthesis of magnesium aluminate spinel powder at room temperature. journal of the american ceramic society 87(11): 2020-2024. du xuelian, liu yaqiang, li liqiang, chen wencong, (2014), synthesis of mgal2o4 spinel nanoparticles via polymer-gel and isolation –medium-assisted calcination. journal of materials research 29(24): 2921-2927. ghosh chandrima, ghosh arup, halder manas kamal (2015), studies on densification, mechanical, micro-structural and structure properties relationship of magnesium aluminate spinel a b c srinath ranjan ghosh , soumya mukherjee, and sathi banerjee 141 refractory aggregates prepared from indian magnesite. materials characterization 99: 84-91. ghosh s.r, mukherjee s, banerjee s (2018), solution combustion synthesis of alumina spinel and its characterization. interceram 67: 34-41. gorshkov v.a, miloserdov p.a, yukhvid v. i, sachkova n.v, kovalev i.d, (2017), preparation of magnesium aluminate spinel by self propagating high temperature synthesis metallurgy methods. inorganic materials 53: 1046-1052. habibi narges, wang yuan, arandiyan hamidreza, rezaei mehran (2017), low temperature synthesis of mesoporous nanocrystalline magnesium aluminate (mgal2o4) spinel with high surface area using a novel modified sol-gel method. advanced powder technology 28: 1249-1257. lee p.y, suematsu h, yano t, yatsui k, (2006), synthesis and characterization of nanocrystalline mgal2o4 spinel by polymerized complex method. journal of nanoparticle research 8: 911-917. macêdo maria laponeide fernandes, bertran celso aparecido, osawa carla cristiane, (2007), kinetics of the γ → α-alumina phase transformation by quantitative x-ray diffraction. journal of materials science 42: 2830–2836. https://doi.org/10.1007/s10 853-006-1364-1 marakkar kutty p v, dasgupta subrata (2013), low temperature synthesis of nanocrystalline magnesium aluminate spinel by a soft chemical method. ceramics international 39(7): 7891-7894. mukherjee soumya, (2020), evolution of spinel magnesium aluminate by combustion route using glycine as fuel and its characterization. international journal of emerging trends in science and technology 6(1): 24-27. nassar y. mostafa, ahmed s. ibrahim, samir ihab (2014), a novel synthetic route for magnesium aluminate (mgal2o4) nanoparticles using sol-gel auto combustion method and their photocatalytic properties. spectrochimica acta part a: molecular and biomolecular spectroscopy 131: 329-334. păcurariu c, lazău i, ecsedi z, lazău r, barvinschi p, mărginean g, (2007), new synthesis methods of mgal2o4 spinel. journal of the european ceramic society 27: 707-710. pommier c, chhor k, bocquet j.f, barj m, (1990), reactions in supercritical fluids, a new route for oxide ceramic powder elaboration, synthesis of the spinel mgal2o4. materials research bulletin 25(2): 213-221. saberi ali, golestani-fard farhad, sarpoolaky hosein, willert porada monika, gerdes thorsten, simon reinhard (2008), chemical synthesis of nanocrystalline magnesium aluminate spinel via nitrate-citrate combustion route. journal of alloys and compounds 462: 142-146. salem shiva (2015) application of autoignition technique for synthesis of magnesium aluminate in nano scale: influence of starting solution ph on physico-chemical characteristics of particles. materials chemistry and physics 155(1): 59-66. torkian leila, amini m mostafa, bahrami zohreh, (2011), synthesis of nano crystalline mgal2o4 powder by microwave assisted combustion. journal of inorganic materials 26(5): 550-554. varnier olivier, hovnanian nadine, larbot andré, bergez pierre, cot louis, charpin jean, (1994), sol-gel synthesis of magnesium aluminum spinel from a heterometallic alkoxide. materials research bulletin 29(5): 479-488. yang ning, chang li, (1992), structural inhomogenity and crystallization behavior of aerosol reacted mgal2o4 powders. materials letters 15(1-2): 84-88. figures 8-9 represent ftir spectra of spinel synthesized using urea and thiourea as fuel. for both cases, scanning range is within 450-4500cm-1 while major m-o coordinations are noted within 1000cm-1. it has been noted for the urea-based synthesis of ... conflict of interest funding acknowledgment references:   *corresponding author’s e-mail: mabualtayef@iugaza.edu.ps doi:10.24200/tjer.vol17iss1pp11-23  the journal of engineering research (tjer), vol. 17, no. 1, (2020) 11-23 applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip mazen abuatayef a, *, abedelmajeed kahail b, hassan al-najjar, c and thaer abushbak d a, * environmental engineering department, the islamic university of gaza, gaza, palestine b ctg-united nations office for project services, ctg-unops, gaza, palestine c civil engineering department, institute of natural sciences, sakarya university, sakarya, turkey d central laboratory for water and soil, ministry of agriculture, gaza, palestine abstract: reverse osmosis (ro) technology shows common popularity in the field of water treatment as an advanced stage to eliminate the residual biogenic elements and dissolved impurities after the traditional treatment processes. this article highlights the applicability of using ro membrane technology as a post-treatment stage to treat the discharged effluent from the gaza wastewater treatment plant. the designed experimental model reveals optimal removal efficiency between 92 and 100% for a number of physical, chemical and biological pollutants. the ro membrane unit demonstrates significant removal efficiency compared to the sand filter where the ro removal efficiency for bod, tss, tds, fecal coliform and no3 were 100, 97.5, 92, 100 and 100%, respectively. the quality of reclaimed wastewater was idealistic where the contents of bod, fecal coliform and no3 in the permeate were nil, and the concentrations of tds and tss were 20 and 296 ppm, respectively. practically, the results confirm that the wastewater with the reclaimed quality could be used for agricultural activities with no degree of restriction according to fao’s guidelines water quality for irrigation. according to the palestinian standard (ps), the quality of reclaimed wastewater is high, class (a), and it could be used without restrictions to irrigate many crops and for the purposes of groundwater replenishment. related to energy estimation and cost analysis, the numerical model and the market analysis study demonstrate the energy of 1.23 kwh and total cost, i.e. fixed and energy costs, of 0.58 usd to produce 1 m3 of reclaimed wastewater using the ro membrane in the gaza strip over a projected lifespan of 5 years. keywords: gaza strip; wastewater; reclamation; reverse osmosis; post-treatment; agriculture. التناضح العكسي الستصالحمدى قابلية تطبيق تقنية المياه العادمة في قطاع غزة مازن أبو اللطيف * ، عبدالمجيد كحيل، حسن النجار و ثائر أبو شباك في ضوء التطور السريع في تكنولوجيا معالجة المياه، أصبح توجه استصالح المياه العادمة خياراً واعداً لتخفيف الملخص: زيادة االستدامة في نظام إدارة المياه في المناطق التي تعاني عجز حقيقي في مصادر المياه. تكنولوجيا التناضح ل أزمة المياه و في العكسي تشكل التقنية األكثر رواجاً في مجال معالجة المياه و إزالة العوالق و الملوثات البيولوجية و غيرها. نسلط الضوء ضح العكسي لمعالجة المياه العادمة الخارجة من محطة غزة للمياه العادمة و ذلك على استخدام أغشية التناهذا المقال لخزان الجوفي حسب المعايير الفلسطينية للمياه. أظهر النموذج ا للوصول إلى مياه قابلة لالستخدام في مجال الزراعة و حقن الكيميائية ، % في إزالة الملوثات الفيزيائية 100إلى 92المصمم باستخدام أغشية التناضح العكسي كفاءة تتراوح بين ، 100 ، 92، 97.5، 100كالتالي bod ،tss ،tds ،fecal coliform ،3noلـ والبيولوجية. وكانت كفاءة إزالة حسب المعايير ) aبدرجة ( تصنيفها . عملياً يمكن االستنتاج أن جودة المياه المعالجة بهذه التقنية يمكن على التوالي % 100 تعني أنه يمكن استخدامها بدون محددات في الزراعة والري وحقن الخزان الجوفي. من ناحية أخرى وهذا ، الفلسطينية 0.58كيلو وات لكل متر مكعب و تبلغ تكلفة معالجة المتر المكعب 1.23زمة يومياً هي الال اظهرت النتائج أن كمية الطاقة سنوات للمشروع. 5دوالر أمريكي حسابًا على فترة عمر .زراعة ؛ المعالجة ما بعد ؛ تناضح عكسي ؛ استصالح ؛ مياه عادمة ؛ قطاع غزة :ةالكلمات المفتاحي applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 12  1. introduction the water scarcity crisis has raised as one of the most pressing problems on the sustainability of life due to the effect of population growth, climate change, water abuse, pollution of water resources, improper management of water and other factors (mehta, 2007; karagiannis and soldatos, 2008; vörösmarty et al. 2010; jacobson et al. 2011 ang et al. 2015). the water crisis threatens about four out of every five persons of the human communities and more than two-thirds of the environmental habitats (jacobson et al. 2010; vörösmarty et al. 2010). in an effort to alleviate the water shortage, new water sources have been identified and supplied through desalination of seawater or brackish water, and by reclaiming wastewater (zhang and he, 2013). however, the utilizing of treated wastewater in agriculture and in the replenishment of water resources (i.e., surface and groundwater) is less expensive than using the desalinated water and it is more affordable for most crops (fao 2006; zhang and he, 2013; jaramillo and restrepo 2017). moreover, the worth of wastewater comes from being a promising constituent in promoting smart sustainable development especially in arid and semi-arid areas and many counties worldwide. the merit of wastewater as an efficient alternative water source is derived from the fact that about 99% of its ingredients are water and only less than 1% is solid wastes and hence the waste components reduction of wastewater to acceptable levels by means of water recycling could be worthy environmentally and economically to the sustainability of water resource management. nowadays, the rapid development in membrane manufacturing technologies, such as micro, ultra, nanofiltration, and reverse osmosis (ro) membrane technologies plays a significant role in enhancing the feasibility of water and wastewater reclamation by adding a quaternary-post-treatment process to remove the remaining pollutants, e.g. pyrogens, color, submicron colloidal matter, bacteria, and viruses, after the tertiary treatment (metcalf and eddy, 2013). in comparison to other membrane technologies, the use of ro technology is the most prestigious in the purification of wastewater due to the low energy consumption and the high rate of pollutant and contaminant removal. edwards and schubert (1974) and fang and chian (1976) reported that the ro membrane technology could eliminate 51% up to 99% of various pesticides. al jlil and sajid (2014) found that the ro water treatment process could remove all the cations and anions from wastewater or seawater especially removing the monovalent ions such as chloride (cl-) with a removing efficiency of 94.4%. suzuki and minami (1991) revealed through several experiments that ro membranes could remove up to 99%, 90% and 99.9% of total dissolved solids (tds), total organic carbon (toc) and fecal coliform, respectively. abid et al. (2012) addressed the dye removal efficiency of more than 97% from industrial wastewater by membrane technologies of ro. practically, agriculture shows a huge demand for water as it is considered the biggest consumer of water in volume among all other water sectors. the possibility of using reclaimed wastewater forms an attractive water source to farmers for several reasons such as its reliability, its low or zero cost and its valuable nutrient content, which increases crop production without adding artificial fertilizers (u.s. epa, 2004; jiménez, 2005; qadir et al. 2010). in specific, the typical domestic wastewater is a rice source of the nitrogen, phosphorus, potassium, micronutrients and organic matters that are naturally required for agricultural harvesting (fao 1992). however, the main challenge that faces the sustainability of any wastewater reclamation activity is compromised by public health assurances. the wastewater is most likely to contain pathogenic organisms similar to those in the original human excreta such wastes: bacteria, viruses, protozoa, and helminth (fao 1992). depending on the harsh effects of sunlight and desiccation, almost all excreted pathogens can survive in soil and on crop surfaces for a sufficient length of time ranges between two days and many months which is a sufficient to indicate that public health risk and infections by diarrhea, typhoid, salmonella type and other virus diseases could be conjugated (who 1989). moreover, the accumulation of heavy metals the wastewater demonstrates a major potential risk to consumers and farmers (jaradat, 2016). gumbo et al. (2010) found significant public health risks for using wastewater for agricultural irrigation. jan et al. (2010) detected heavy metals concentrations in a number of tested food crop samples irrigated with wastewater. generally, the long term using of partially treated wastewater in irrigation can degrade soil and cause public health risks through the transmission of enteric diseases and heavy metal accumulation (jamil et al. 2010). obviously, the different water activities and uses in terms of public health and environmental protections are principally governed by specific guidelines and frameworks that are set and adopted by the world health organization (who) or by the national legislation of countries. in palestine, the relevant policies, legislation, and regulation related to water quality are set and adopted by the palestinian water authority (pwa) with direct negotiations and discussion with the related ministries such as the ministry of health, ministry of agriculture and the environmental quality affairs (pwl, 2002). generally, the wastewater treatment passes through a tertiary treatment process which contains an ultraviolet (uv) disinfection stage to eliminate the pathogenic organisms from the treated wastewater. unfortunately, the technique of uv is unavailable in most of the regions that suffer from a lack of facilities and energy resources. the wastewater is a valuable source of water to balance the water deficit in the gaza strip, therefore, this study demonstrates the applicability of using the ro membrane technology as a post-treatment stage instead of the uv stage to eliminate the remained nutrients and pathogenic organisms from the reclaimed wastewater in order to be efficient in term of quality to be applicable in different water sectors according to specified palestinian water standards and guidelines. moreover, not only in the gaza strip, the applicability of ro membrane technology as a posttreatment stage could be a potential unique solution for different countries in the middle east and north africa mazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak 13  which suffer from poor facilities and energy resources to overcome the need to tertiary treatment stage of uv which needs a well-established preparations and facilities. 2. environmental standards and regulations investigating the efficiency of any water treatment systems is performed in terms of the produced quality of water and its coinciding with the adopted guidelines and standards. water quality, which is as significant and important as quantity, has been overlooked for decades in terms of legislation, investment and public awareness. water and sanitation infrastructure, water and sanitation policies, good governance and practice, proper legislation, regulations, and standards are key issues to safeguard humans and their surrounding environment. various legislations and regulations were designed worldwide to regulate water resource protection, water supply, sanitation, protection of the environment, and prevention of pollution. standards have been adopted for different purposes of water uses and reuse. monitoring systems have been put in place to ensure responding to these regulations and standards. the availability of adopted water quality standards and regulatory aspects in forms of national or international standards is important to control and restrict the arbitrary use or disposal of wastewater. there are no unified criteria for wastewater reuse in the world and each country proposes its own best practices criteria based on different environmental, economic, social and political conditions. globally, many organizations and agencies like the world health organization (who), food and agriculture organization (fao) and environmental protection agencies (epas) suggest guidance standards for wastewater reuse for different purposes. in the 1970s, the who started developing guidelines, shown in table 1, for wastewater reuse in agricultural irrigation to ensure human health and environmental integrity. these guidelines are based on the consensus view that the actual risk associated with irrigation with treated wastewater is much lower than previously thought and that earlier standards and guidelines for effluent quality (who, 2006). traditionally, irrigation water is grouped into various quality classes in order to guide the user to the potential advantages as well as problems associated with its use and to achieve optimum crop production. the water quality classifications are only indicative guidelines and their application has to be adjusted to conditions that prevail in the field. this is so because the conditions of water use in irrigation are very complex and difficult to predict (who, 2006). the suitability of water for irrigation greatly depends on the climatic conditions, physical and chemical properties of the soil, the salt tolerance of the crop grown and the management practices. thus, the classification of water for irrigation will always be general in nature and applicable under average use conditions. many schemes of classification for irrigation water have been proposed. fao (1985) classified irrigation water into three groups based on salinity, sodicity, toxicity, and miscellaneous hazards. these general water quality classification guidelines help to identify potential crop production problems associated with the use of conventional water sources. the guidelines are equally applicable to evaluate wastewaters for irrigation purposes in terms of their chemical constituents, such as dissolved salts, relative sodium content and toxic ions. depending on the sodium ion concentration relative to the concentrations of calcium and magnesium ions (as indicated by sar) and the total salt concentration, the potential content of sodium ions in the irrigation water affects the infiltration rate and the soil permeability. obviously, it is clear to indicate that for a given sar value, an increase in total salt concentration is likely to increase soil permeability and, for a given total salt concentration, an increase in sar will decrease soil permeability. this illustrates the fact that soil permeability (including infiltration rate and surface crusting) hazards caused by sodium in irrigation water cannot be predicted independently of the dissolved salt content of the irrigation water or that of the surface layer of the soil (fao 1985). in palestine, the palestinian water laws clearly declare the responsibility of the pwa, in consultation with the ministry of health, ministry of agriculture and environmental quality authority, to establish guidelines and framework for monitoring the water-related practices, protecting the water resources and ensuring the safety of palestinian public health (pwl 2002, 2014). the palestinian standard (ps 2003) imposes a set of mandatory technical instructions for using treated wastewater in agricultural irrigation to maintain the safety of humans, animal health and the crops as well as to sustain environmental constituents. the ps prohibits the usage of treated wastewater for watering livestock and poultry, irrigation of all types of vegetables, groundwater recharge through direct injection and fish farming. based on the quality of treated wastewater, the ps categorizes the effluents into four classes based on the water quality as shown in table 2. the ps states that “the irrigation using reclaimed wastewater should be stopped before three weeks of the yield time for the fruit crops and before two weeks for field and feeding crops” to take into consideration the sensitivity of some crops to the properties and the elements of treated wastewater. the ps permits the use of treated wastewater, without any restrictions on the irrigation activities, for harvesting the yields of cotton, luffa, brooms and other crops that are not in direct contact with the public. however, concerning the crops that are in direct contact with the public through touch or eating, the ps imposes intensive preventive restrictions and protective procedures for using the treated wastewater in irrigation. the ps adopts eleven irrigation methods that enable the farming activities from using the reclaimed wastewater for irrigating specific types of crops. each of these methods is classified as a two-way or a one-way restriction method (ps, 2003). one-way restriction irrigation methods:  a distance ≥ 25 cm above the ground between the dripping point and the crop or fruits.  a distance ≥ 50 cm between the level of sprinklers and the fruits.  a ground plastic cover between the treated wastewater and the fruits.  crops or fruits with a single crust or an uneaten shell. applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 14  table 1. who microbiological quality guidelines for wastewater use in agriculture (who, 2006). category reuse conditions exposed group intestinal nematodes (arithmetic mean no. of eggs per liter) fecal coliforms (geometric mean no. per 100 ml) treatment to meet the micro-biological guidelines a irrigation of crops probably to be consumed uncooked, sports fields, public parks farmers, consumers, public ≤ 1 ≤ 1000 a series of stabilization ponds designed to attain microbiological quality indicated, or equivalent treatment b irrigation of cereal crops, industrial crops, fodder crops, pasture and trees farmers ≤ 1 no standard recommended retention in stabilization ponds for 8–10 days or equivalent helminths and fecal coliform elimination c localized irrigation of crops in category b if exposure to workers and the public does not occur none not applicable not applicable pretreatment as essential by irrigation technology but not less than primary sedimentation table 2. classification of wastewater quality (ps, 2003). parameter unit class (a) class (b) class (c) class (d) quality type high good medium low bod5 mg/l 20 20 40 60 tss mg/l 30 30 50 90 fecal coliform cfu /100 ml 200 1,000 1,000 1,000 cod mg/l 50 50 100 150 do mg/l ˃ 1 ˃ 1 ˃ 1 ˃ 1 tds mg/l 1,200 1,500 1,500 1,500 ph 6.00 to 9.00 6.00 to 9.00 6.00 to 9.00 6.00 to 9.00 fat, oil & grease mg/l 5.00 5.00 5.00 5.00 phenol mg/l 0.002 0.002 0.002 0.002 methylene blue active substance (mbas) mg/l 15 15 15 25 no3‐n mg/l 20 20 30 40 nh4‐n mg/l 5 5 10 15 total‐n mg/l 30 30 45 60 cl mg/l 400 400 400 400 so4 mg/l 300 300 300 300 na mg/l 200 200 200 200 mg mg/l 60 60 60 60 ca mg/l 300 300 300 300 sar 5.83 5.83 5.83 5.83 po4‐p mg/l 30 30 30 30 al; fe mg/l 5 5 5 5 as; cr mg/l 0.1 0.1 0.1 0.1 cu, mn, ni, pb mg/l 0.2 0.2 0.2 0.2 se mg/l 0.02 0.02 0.02 0.02 cd mg/l 0.01 0.01 0.01 0.01 zn mg/l 2 2 2 2 cn; co mg/l 0.05 0.05 0.05 0.05 hg mg/l 0.001 0.001 0.001 0.001 b mg/l 0.7 0.7 0.7 0.7 e. coli cfu /100 ml 100 1,000 1,000 1,000 nematodes eggs/l ≤ 1 ≤ 1 ≤ 1 ≤ 1  crops or fruits eaten only cooked.  sand filter detain the wastewater for at least 15 days or water ponds that contain less than 10% of treated wastewater.  disinfection by chlorine so that the residual chlorine is not less than 0.5 mg/l and the contact time is not less than an hour half or by using any other sterilization method. two-way restriction:  a distance ≥ 50 cm above the ground between the dripping point and the crop or fruits. irrigation methods:  underground drip irrigation. the ps requires specific standard considerations to irrigate crops by the reclaimed wastewater. the ps restricts the irrigation activities for gardens, playgrounds, and parks to be by a reclaimed wastewater quality of class (a). however, the processes of groundwater replenishment by infiltration and the offshore disposal of about 500 m into the sea should be by wastewater quality of not less than class (c). the standards permit the irrigation of crops of seeds production, dry feeds, forest trees, industrial corps mazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak 15  (a) (b) figure 1 (a). geographical location of the gaza strip, and (b) positions of wastewater treatment plants in the gaza strip. and grains using low quality of treated wastewater. the ps allow using treated wastewater with class quality of (a), or other quality with suitable one-way or two-way restriction irrigation methods to irrigate the crops of artichoke, eaten corns, citrus fruits, uneaten shell fruits, falling leaves trees, tropical crops, grapes, cactus, dates, olives, and flowers (ps, 2003). in the area of the gaza strip, four types of crops are cultivated and harvested: (a) field crops: wheat, potatoes, and onions; (b) vegetables: tomatoes, cucumbers, eggplants, squash, green beans, and paprika; (c) fruit trees: citrus fruits, guava, dates, and almonds; and (d) olives (arij, 2015). generally, drip irrigation remains the domain irrigation system and widely used for agricultural production by farmers in the gaza strip (arij, 2015). 3. water issues of the study area the gaza strip (shown in fig. 1(a)) is a stretch of area that extends on the coast of the eastern mediterranean with about 42 km long, between 6 and 12 km wide, and an area of 365 km2. the gaza strip is recognized as one of the most densely populated areas in the world. the estimations of the palestinian central bureau of statistics (pcbs) for the year 2018 revealed that the population of the gaza strip was more than two million inhabitants (pcbs, 2018). concerning the current situation of the water sector, the region of the gaza strip is characterized by many parties as an area in a continuous humanitarian crisis due to the lack of both quantity and quality of water resources for different uses (pwa, 2011, 2013, 2014). the underlying coastal aquifer is the only water resource that is exploited in the gaza strip for the domestic, industrial and agricultural uses. the status of the coastal aquifer shows permanent deterioration on both quantity and quality. the annual replenishment volume of the aquifer is between 55 and 60 million cubic meters while the palestinian abstraction from the aquifer in the gaza strip is quantified by about 180 million cubic meters per annum. the agricultural sector consumes the largest portion of nearly 50% of the entire pumped groundwater water from the aquifer comparing to the other remaining domestic and industrial sectors that tighter use the second half of the extracted groundwater. accordingly, the records of water balance show a severe and continuous deficit of about more than 120 million cubic meters per annum (pwa, 2013). in terms of quality, the present rate of deterioration demonstrates that over 95% of the coastal aquifer is contaminated with unacceptably high levels of either nitrate (no3) or chloride (cl), posing significant health risks to gaza’s residents (pwa, 2014). the sustainability of coastal aquifer became a crucial and sensitive issue for the palestinian water authority (pwa) and other palestinian water-related organizations and agencies. hence, in order to alleviate the huge pressing on the coastal aquifer, the pwa pays more attention to using new alternative and other nonconventional water sources. i. purchased water: the water sector in the gaza strip is supplied by 10 million cubic meters per year. this quantity is bought from israeli water utility (mekorot). in addition, under the negotiation’s agreement of oslo ii, there is a commitment to increase the supplied quantity by 10 million cubic meters (pwa, 2014). ii. desalination of seawater: desalination of seawater became a strategic option for the pwa to overcome the shortage in the water supply. hence, the pwa provided a strategic plan for water supply along with the short and long terms. nowadays, short-term low-volume (stlv) seawater desalination plants, the overall capacity of 13 million cubic meters per annum, are built to enhance the quality and quantity supply of water in the gaza strip. from a long-term perspective, a plan to construct a regional strategic large-scale applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 16  seawater desalination was adopted. the first phase of the regional seawater is intended to provide water sector by 55 million cubic meters yearly while the second phase will lift the plant’s capacity to 120 million cubic meters per annum (pwa, 2015). iii. treated wastewater reuse: the reclamation of wastewater draws the attention of pwa as a potential source that could be used to maintain the coastal aquifer sustainable. exploiting the reclaimed wastewater can achieve the sustainability of groundwater by 50%. practically, the pwa starts studying the willingness of using reclaimed wastewater for irrigating agriculture (pwa, 2013). on average, the water services suppliers estimate the total amount of produced wastewater in the gaza strip by approximately 50 million cubic meters per annum (arij 2015; pcbs, 2018). currently, there are four small scale and one large scale operating wastewater treatment plants, geographically demonstrated in fig. 1(b), in the gaza strip. the four small scale wastewater treatment plants of gaza, wadi gaza, khanyunis and rafah have the capacity to treat partially, and not in an acceptable level of quality, about 38.7 million cubic meters of the collected wastewaters in the gaza strip. this quantity which forms a ratio of about 75-80% of the total generated municipal effluents in the gaza strip is directly disposed into the marine ecosystem without enough proper treatment and without any mitigation measures. as well, the operating large scale of north gaza (ngest) wastewater treatment plant treats about 12.8 million cubic meters of wastewater which is utilized to recover the coastal aquifer through a number of wastewater infiltration basins. the characteristic properties of raw wastewater quality for different governorates in the gaza strip, as depicted in table 3, show that the effluent of wastewater has high organic matter, i.e. bod5 of 724 mg/l, compared with the average bod5 levels in the developing countries which range from 200-300 mg/l (polprasert, 2007). these high pollutant contents affect adversely the treatment efficiency of the in-operation wastewater treatment plant. the high growth in the population of the gaza strip increases the load on the capacity and affects the initial quality design parameters that have been used to plan the current wastewater treatment plants which affect the treatment efficiency. therefore, the sanitation services in the gaza strip are in curial crisis and the existing wastewater treatment plants work under overloaded flow rates and so some wastewater is being treated and the large amount is returned raw and pumped into the sea. based on the design parameters of bod, cod, and tss, illustrated in table 4, the designed efficiency of the small-scale wastewater treatment plants is about 75% (cmwu, 2012). however, nowadays, and due to over rates of wastewater and the continuous deterioration in the treatment efficiency of the small-scale wastewater treatment plants, the pwa suggested closing these plants and constructing instead three large scale wastewater treatment plants on the eastern border of the gaza strip (pwa, 2014). to enhance the status of public health and sanitation services, the pwa plans to replace the current over-loaded wastewater treatment facilities by three new large-scale high-efficiency wastewater treatment plants located as shown in fig. 1(b) in the north, central, and south of the gaza strip. recently, as shown in the provided timeline plan in table 5, the wastewater treatment plant of north gaza entered the service to replace the old plant at beit lahia. also, the gaza central and south wastewater treatment plants are under construction and the decision-makers confirm that they will be ready to operate by 2022. the efficiency of the three planned large-scale wastewater treatment plants will be very acceptable, where, the effluent characteristics are between 10 to 20 mg/l for bod5, 15-20 mg/l for suspended solids, 10 to 15 mg/l for total nitrogen, less than 1 eggs per liter from helminths and less than 200 cfu per 100 ml for fecal coliforms (pwa, 2011). the quality of effluent indicates that these strategic plants were designed to provide treated wastewater of class (a). according to the monitoring records of the pwa, the volume of treated wastewater that was used in 2012 is 1 million cubic meters which represent only 3% of the available partially treated wastewater. the pwa put a strategic plan, shown in table 6, extends to the year 2032 to enhance the water management system in the gaza strip (pwa, 2013). according to this plan, by 2022, in the short-term strategy plan, it is expected to lift the exploited treated wastewater to more than 25% of the available partially treated wastewater. however, by 2032, the expectations of the long-term strategy plan address that about 25 million cubic meters per annum and nearly 75 million cubic meters of treated wastewater are going to be used for agriculture and aquifer recharging, respectively. in order to alleviate the deterioration in the coastal aquifer, the long-term plan aims to decrease the overall groundwater abstraction in the gaza strip from the present rate of 180 million cubic meters per annum to about 70 million cubic meters per annum in 2032. as shown in table 6, the long-term strategic plan aims to increase the sustainability of the coastal aquifer by using the reclaimed wastewater as a new water source for agricultural purposes. the plan expects that the total anticipated demands of water for agriculture will be reduced as a result of urbanization and population growth, hence, the projected potential irrigated lands will be about 90,000 dunums by 2032. 4. aterials and methods this article highlights the performance of ro membrane technology as a post-treatment stage in enhancing the treatment efficiency to improve the quality of wastewater reclamation. the performance of this technology was investigated by setting up an experimental model and by testing and analyzing, in the laboratory, the quality of wastewater before and after the ro treatment. the experimental physical model, shown in fig. 2, consists of: (1) collecting tanks: high-density polyethylene (hdpe) feeder, permeate and concentrate tanks each of cubic meter volume were exploited to collect the raw and the reclaimed wastewater before and after the treatment process; (2) sand filter: a composite of sand and well-graded sizes of gravels range from 2.37 and 9.5 mm was prepared to configure a cylindrical sand filter with a length of 50 cm and mazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak 17  table 3. characteristics of wastewater in the gaza strip (cmwu, 2012). parameter unit characteristics of wastewater according to area north area gaza rafah average biochemical oxygen demand (bod5) (mg/l) 728 667 777 724 chemical oxygen demand (cod) (mg/l) 1385 1306 1399 1,363 suspended solids (ss) (mg/l) 663 617 540 607 ss/bod ratio 0.9 0.95 0.69 0.85 bod/cod ratio 0.526 0.51 0.56 0.532 table 4. efficiency of wastewater treatment in the gaza strip (cmwu, 2012). wastewater plant biochemical oxygen demand (bod5) chemical oxygen demand (cod) suspended solids (ss) influent effluent removal influent effluent removal influent effluent removal mg/l mg/l % mg/l mg/l % mg/l mg/l % gaza 500 105 79 1020 220 78 550 110 80 rafah 560 120 81 1160 255 78 550 122 79 khanyunis 520 155 70 1090 322 70 580 141 76 beit lahia 440 133 70 980 250 74 480 222 71 average 505 128 75 1,063 262 75 540 149 77 table 5. characteristics of large-scale wastewater treatment plant in gaza strip (pwa, 2011). facility flow (m3/day) current status future status final destination north gaza 35,000 north gaza wastewater treatment plant; opened in 2018 to replace bait lahia treatment plant which is now out of service northern gaza emergency sewage treatment (ngest) will be upgraded to a capacity of 70,000 m3/d infiltration basins central gaza 65,000 gaza wastewater treatment plant; established in 1979; now the plant is overloaded with a capacity of 65,000 m3/day central gaza treatment facility with a capacity of 65,000 m3/d to replace the current plant; it will be operated in 2020 seawater 14,000 middle gaza wastewater treatment plant; established in 2014 with a capacity of 12,000 m3/day south gaza 14,000 khan yunis wastewater treatment plant; consists of three lagoons were built in to collect and treat wastewater during 2003 – 2009 south gaza treatment plant with a capacity of 26,000 m3/d as phase i operated in 2019 and 44,000 m3/d operated in 2022 to replace the current treatment plant infiltration basins table 6. potential reuse of treated wastewater (pwa, 2013). a diameter of 30 cm to eliminate the content of large suspended solids before the ro unit; (3) microfiltration (mf) and ultrafiltration (uf) unit: are cellulose acetate semipermeable membranes that remove too large solids before the ro unit, the microfiltration membranes range from 0.1 to 10 μm, and ultrafiltration membranes range from 0.1 to 0.01 μm; and (4) ro unit: consists of a semipermeable membrane thin polyamide layer (< 200 nm) deposited on top of a polyethersulfone or polysulfide porous layer (about 50 microns) on top of a nonwoven fabric support sheet. these semipermeable membranes have the ability to remove ions, molecules and larger particles from water by applying pressure to overcome the osmotic pressure. regarding, quality control and quality assurance (qc/qa), many experimental trials were executed during february and april of 2018 to take into consideration the seasonal change in the characteristics of wastewater that reaches the wastewater treatment plants in the gaza strip. for each experimental trial, approximately 500 liters of wastewater sample was collected from the effluent of gaza wastewater treatment plant and transported to the laboratory of the islamic university of gaza using barrels. the barrels were preserved into an icebox to inhibit the metabolic derangements of the wastewater organic content. in the laboratory, the collected wastewater samples from the gaza items years 2012 2022 2027 2032 quantity of reclaimed wastewater 33.2 59.3 75.8 99.9 irrigation portion (million m3/year) 1 14.8 19 25 aquifer recharge (million m3/year) 32.2 44.5 56.9 75 portion of groundwater for irrigation (million m3/year) 86 59 45.5 32 potion of constructed dams for irrigation (million m3/year) 0 5 7.5 10 total available quantity for irrigation (million m3/year) 87 78.8 72 67 irrigation needs (million m3/year) 74.1 74.1 74.1 74.1 irrigable land (dunum) 133,000 123,000 118,000 113,000 potential irrigated land (dunum) 117,403 106,383 97,112 90,401 percent of irrigable land (dunum) 88.3% 86.5% 82.3% 80% applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 18  wastewater treatment plant were poured into the feeder tank. the wastewater was passed through the sand filter, which composes of graded-size gravel, the flow of wastewater inside the sand filter is governed by the force of gravity. the general design criteria for the slow sand filter recommend a filtration rate between 0.1 and 0.2 m per hour through the filter layers (kawamura 2000). however, the filtration treatment through the designed sand filter demonstrated a slow sand filtration, hence, the water head reached about 20 cm above the sand filter layers with a filtration rate of 0.5 liters per minute. the passed wastewater through the sand filter is entered into the three stages of mf/uf membrane unit with 5and 1-micron pores diameter which was selected as a second pretreatment unit after sand filter to protect the ro device from fouling and to reduce suspended solids. after sufficient pretreatment, the water was pumped by a main pump of 130 psi (8.844 bar) through the ro unit with a flow of the filter was 1.8 liters per minute, the recovery rate was 50% from the fed water. in order to enhance the recovery efficiency of the system, a repass cycle was inserted into the system, where, a small repass pump of 4.23 bar was operated figure 2. scheme of the experimental physical model. figure 3. model configuration of winflows. mazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak 19  to repump the concentrate through the ro unit. the operation of the repass cycle was designed to operate when the concentrate in the concentrate tank reaches a volume of 250 liters or more. finally, the treated water i.e., permeate water, was collected into the permeate tank (5) which is the treatment product that was tested in the lab. for each experimental trial, four samples were taken in every stage of filtration in each of the two experimental trials for testing and analysis purposes. the samples were put in polythene bottles that were pre-washed with acid and distilled water and then were dried. the first sample was taken from the feeder tank and before sand filter, the second was taken after the sand filter, and finally, the last two samples were taken from ro concentrate and permeate tanks, respectively. the four gathered samples were preserved at 4°c in an icebox and brought to the testing department of the laboratory for testing and results declaration. the quality of the treated wastewater in every phase of the treatment process was tested in order to examine the efficiency of each component in the experimental model in terms of biochemical oxygen demand (bod), total suspended solids (tss), total dissolved solids (tds), hydrogen ion concentration (ph), electrical conductivity (ec), fecal coliform (fc) and nitrate (no3-n). the described standard methods for the examination of water and wastewater by the american public health association, american water works association, and the water environmental federation (2017) have been employed to examine the water quality parameters. the content of the organic matters has been tested in terms of the 5-day bod test according to the procedure of the 5210 b. the described standard method in the procedure of 2540 d and 2520 b were applied to measure the tds and the ec in the reclaimed wastewater samples, respectively. furthermore, the acidity of the treated wastewater in term of ph was determined according to the procedure of 4500-h+. concerning the pathogenic content, the fc concentration was detected according to the procedure of the 9211 b seven-hour fecal coliform test. else, the concentration of no3 in the wastewater was determined using the nitrate electrode method according to the procedure of 4500-no3 d practically, testing the efficiency of membrane technology as a post-treatment in wastewater reclamation from the view of quality is not sufficient to judge on the whole efficiency of this technology. principally, the main challenge governs the development and the popularity of membrane technology is the challenge of energy. thus, it is critical to estimate the feasibility of this system from the views of cost and energy consumption. it is difficult to address the cost and energy consumption through the experimental physical model. hence, with the rapid increase in computer power, it is not surprising to note the highly dependent on numerical simulation as a prediction tool in the predesign and design stages. in this article, the simulation model of win-flows ro system design and simulation software was used to estimate numerically the cost and energy consumption of wastewater reclamation through the designed experimental model. to ensure better simulation results, the software was run several times and a cartridge-filter-pre-treatment unit was added to the simulation system to provide a better presentation of the reality. the win-flows has some distinctive features like three pass systems, permeate split and recycle, anti-scalant dosing, energy recovery devices and ability to combine stages that make its ability higher than other similar software. the best presentation windflows model for the experimental model of this study, shown in fig. 3, was designed to treat 1,000 m3 per hour of tertiary treated wastewater with a tds of 3,800, ph of 7.8 at a temperature of 16°c. after a series of trials and errors, the final design of the model system consisted of cartridge filter as pretreatment and two stages of ro elements, 90 pressure vessels, and 65 pressure vessels, every vessel contains 7 elements of duraslick anti-fouling membranes that are designed especially for wastewater treatment. the selected membrane model was dsl ro8040 to maintain the continuity of membrane efficiency for 5 years. in the numerical model, the water is pumped through the system using two pumps. in the first treatment stage, the main pump, which has a pressure of about 18 bar draws the feedwater by a capacity of 505 m3 per hour and pumps it into the ro membrane system. however, the rejected water from the first stage is mixed with a flow of feedwater and repumped again into the membrane system using a repass pump with a pressure of 4.23 bar and a capacity of 495 m3 per hour. the numerical model demonstrates that the system gives a recovery ratio of 50% which is acceptable and similar to the recovery ratio of the experimental model. 5. results discussion and cost estimation reusing of reclaimed wastewater is expected to overcome the deficit in the water management system as a new proper water source. the advance development in membrane technology supports the progress in reusing wastewater; however, the energy challenges could be the main obstacle to make these technologies practical. the applicability of ro membrane technology as a post-treatment unit to recycle wastewater was investigated based on the quality of the reclaimed wastewater and on the energy efficiency of the whole system. the quality of reclaimed wastewater was addressed by the experimental examination for the efficiency of using ro membrane technology in treating wastewater. the experimental trials and laboratory testing were performed several times to ensure better qc/qa and to present precise results and decisions. table 7 illustrates the average removal efficiencies of impurities and pollutants through the ro experimental model in terms of different physical, chemical and biological parameters for all runs. the quality of collected wastewater from gaza wastewater treatment shows higher concentrations than designed quality effluent by about 138%, and excess concentrations above the average designed concentrations for three operating wastewater treatment plant in the gaza strip by 95%. the tested samples provide practical evidence that the treatment plants in the gaza strip operate with an unacceptable quality level. thus, this article demonstrates efficiency investigation for the applicability of using membrane technology to reclaim wastewater in a region suffers from inefficient wastewater treatment plants. the applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 20 depicted experimental model results demonstrate that the treated wastewater using ro membrane technology achieves acceptable water quality levels for the agricultural and groundwater replenishment activities. the quality of the treated effluent wastewater through the designed experimental model was examined in terms of the physical, i.e. bod5, tss and tds, the biological, i.e. fc and no3, and the acidity or basicity index of ph. in the light of bod5, the whole system shows a high overall efficiency of 100% for organic contaminants rejection, and the efficiency of the sand filter and ro unit were 59% and 100%, respectively. the system units and the whole system reveal lower removal efficiency for tss and tds than other parameters. the capacity of the system in eliminating tss was higher than that of tds by about 5%. the sand filter contribution in the whole eliminating of tss and tds is insignificant and the efficiency of sand filter in removing tds is almost 0%. the ro unit forms the main removing part by the efficiency of 97.5 and 92% for tss and tds rejection, respectively. the contents of fc in the tested wastewater samples were about within 1300 cfu per 100 ml. the ability of sand filters to eliminate the fc from the samples was about 95% and the ro unit shows full ability to remove the remaining fc after the primary treatment by a sand filter. the nitrification process takes a significant place in the experimental system; hence, the sand filter trends to increase the concentration of no3 in reclaimed water as a result of conversion from nh4 into no3 through the nitrification process. the no3 is a rich nutrient to plant growth and it is considered as a valuable element to the agricultural activities. however, the ro unit has a high ability to remove no3 and in all the whole system was able to fully remove all no3 from the permeate. related to the ph, the nitrification process trends to covert the state of water into more alkalinity, hence, the results show that the ph of the treated wastewater rises from 7.7 to 8.7 after treatment so for the purposes of agricultural activities the treated wastewater must be adjusted to be within a ph value of 6.5 to 8.4 by adding a point entry injection system to pump an acid solution of acetic acid (white vinegar), citric acid or alum to reduce the basicity of treated water. the energy and cost estimations are two linked tasks, the specific energy consumption (sec) for the ro system is commonly calculated using over simple analyses depending on the average operation tasks for the specific plant. the results of the built numerical model through the winflows ro system design and simulation software indicate that the sec of the whole experimental system is about 1.23 kwh per one cubic meter of permeate. however, the local membrane technology market indicates to set up a membrane-based treatment facility with a capacity of one million cubic meters of permeate per year over a five years lifespan is approximately 6 million united states dollars (usd) as shown in table 8. thus, the fixed cost estimation addresses that the loaded fixed costs on the unit volume of treated wastewater are about 0.4 usd per cubic meter of permeate. in the gaza strip, the cost of the municipal energy is approximated quantified by 0.15 usd for 1 kwh. hence, the total cost to reclaim one cubic meter of wastewater is about 0.6 usd. reviewing the cost and energy consumption estimations for similar ro wastewater treatment plants worldwide could introduce better indication about the expected cost and energy to produce one cubic meter of reclaimed wastewater. facing all the challenges table 7. experimental results of ro wastewater reclamation. bod5 (mg/l) tss (mg/l) tds (mg/l) ph fc (cfu /100 ml) no3 (mg/l) raw wastewater 250 1,070 3,800 7.7 1,250 0.4 after sand filter 103 837 3,800 7.8 67 15 concentrate 230 1,700 7,300 7.8 10 2 permeate nil 20 296 8.7 nil nil sand filter efficiency (%) 59 22 0 increases ph 95 increases no3 ro-unit efficiency (%) 100 97.5 92 increases ph 100 100 system efficiency (%) 100 98 92 increases ph 100 100 table 8. estimated fixed cost for applying the model. series item unit no. unit price (usd) total price (usd) 1 8-inch low fouling duaslick membrane no. 1,085 1,900 2,061,500 2 8-inch vessels (7 elements per vessel) no. 155 4,000 620,000 3 cartridge filter no. 8 3,000 24,000 4 primary pump no. 10 5,000 50,000 5 high pressure pump no. 10 15,000 150,000 6 pressure exchange no. 2 50,000 100,000 7 dossing pump with tanks no. 2 10,000 20,000 8 pressure exchange no. 2 50,000 100,000 9 dual media filters no. 4 7,000 28,000 10 backwash pumps no. 4 3,000 12,000 11 flow meters no. 4 1,000 4,000 12 skids no. 1 30,000 30,000 13 plc no. 1 200,000 200,000 14 fittings and connections no. 1 100,000 100,000 15 h2so4 along five years kg. 45,000 2 90,000 16 hanger and other plant structure no. 1 750,000 750,000 17 effluent and permeate tanks 4000 m3 no. 2 800,000 1,600,000 18 total price usd 5,939,500 19 amount of permeate along five years m3 15,000,000 20 total cost per 1 m3 of permeate usd 0.394 21  and constrains relating to using ro in the reclamation of wastewater as new technology, the main two largest membrane-based treatment plants in the world, sulaibiya wastewater treatment plant in kuwait and orange country wastewater treatment plants in the united states (usa) produce daily hundreds of thousands cubic meters of reclaimed wastewater with stability and continuance. sulaibiya wastewater treatment plant was opened in 2004; the plant is the biggest membrane-based water treatment plant. currently, the capacity of sulaibiya wastewater treatment plant is about 375,000 m³ per day and it is planned to expand it to reach a capacity of 600,000 m³ per day and it is anticipated to cover 26% of kuwait’s gross water needs, and is expected to reduce the demand from 140 to 25 million cubic meters per annum (pearce, 2008). the orange country plant was opened in 2004, the facility treats 320,000 m3 daily of wastewater effluent using ro membrane technology. the plant is going to be expanded to a capacity of 590,000 m3 per day (water technology, 2018). generally, the estimations reveal that the conventional treatment of wastewater by the ro membrane technology, the costs about 85 to 90 cents per cubic meter of permeate (pearce, 2008; water technology, 2018). on the other hand, the requirement of energy to purify water depends on the quality of raw water and the applied treatment technique. the energy consumption for the purification of natural sources such as surface water and groundwater using membrane filtration is about 0.1 kwh per cubic meter. however, the energy consumption for treating surface water using both conventional treatment and uf/mf membrane is estimated by about 0.25 to 0.35 kwh per cubic meter. in general, the energy of water treatment could be estimated based on the tds of the raw water, for tds of less than 3,000 ppm about kwh of energy is needed to purify one cubic meter however for tds between 3,000 and 11,000 ppm nearly 1.7 kwh of power is required to treat a cubic meter of water. regarding, wastewater reclamation the energy requirement to recycle wastewater using dual membrane filtration of mf/uf and ro needs from 0.8 to 1.2 kwh per cubic meter. however, the use of a membrane bioreactor followed by ro requires energy between 1.2 and 1.5 kwh per cubic meter. in the countries of the mediterranean sea, the ro seawater desalination technique consumes about 2.3 to 4 kwh per cubic meter of freshwater (pearce, 2008). 6. conclusion reclamation of wastewater is a promising source to meet the increasing water demands in the arid and semi-arid regions. the agricultural sectors consume more than 50% of the water budget and providing treated wastewater with acceptable quality levels for irrigation can increase the sustainability of the water system and the fertility of agricultural lands. the rapid advance in the membrane technology draws the attention toward using ro membrane technology as a post-treatment process for wastewater treatment. the pollutants removal efficiency of ro technology, as well as the low cost of operation and maintenance, makes this technology dominant in water treatment. this study experimentally concludes that the romembrane technology system has the ability to remove more than 90% of solids pollutants and the efficiency of eliminating the other bacterial and ionic pollutants could reach 100%. the study estimates the energy consumption for wastewater treatment using ro by around 1.23 kwh per cubic meter of permeate which is far less than the consumption of other alternative solutions brackish water and seawater desalination that consume about 1.7 kwh and up to 4 kwh per cubic meter, respectively. this technology provides a sustainable treatment process for countries suffering from a lack of electrical sources since the experimental model of this study estimates that the fixed cost to treat one cubic meter of permeate costs approximately 40 cents, compared to other comparison studies, the described model of this study could save more than 50% of the fixed costs to treat one cubic of wastewater. therefore, the findings of this study clarify that the romembrane technology system as a post-treatment stage in the wastewater treatment plants of the gaza strip could improve the wastewater quality to meet ps guidelines and produce a reclaimed wastewater of class (a) that could be used without restrictions to irrigate gardens, playgrounds, parks, groundwater replenishment and for the offshore disposal at about 500 m into the sea. moreover, the quality of the reclaimed wastewater could be used for planting crops of seeds production, dry feeds, forest trees, industrial crops and grains, crops of artichoke, eaten corns, citrus fruits, uneaten shell fruits, falling leaves trees, tropical crops, grapes, cactus, dates, olives, and flowers. regarding the existing drip irrigation system in the gaza strip which is widely common, this study recommends the use of the reclaimed wastewater using ro membrane technology to irrigate the crops of citrus, guava, dates, almonds, and olives under a two-way restriction measures by lifting the irrigation system by at least 0.5 meters from the ground. in conclusion, the applicability of the ro treatment system could be a significant component to enhance the strategic water development plan of the pwa to maintain the sustainability of the coastal aquifer. conflict of interest the authors declare no conflicts of interest. funding no funding was received for this project. references abid mf, zablouk ma, abid-alameer am (2012), experimental study of dye removal from industrial wastewater by membrane technologies of reverse osmosis and nanofiltration. iranian j environ health sci eng 9(1): 1-17. al jlil s, sajid m (2014), study of wastewater treatment plant. applied numerical mathematics and scientific computation, proceedings of the 2nd international mazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak applicability of using reverse osmosis membrane technology for wastewater reclamation in the gaza strip 22  conference on applied mathematics and computational methods:132-140. alhumoud jm, al-humaidi h, al-ghusain in, alhumoud am (2010), cost/benefit evaluation of sulaibiya wastewater treatment plant in kuwait. international business & economics research journal 9(2): 23-32. american public health, american water works association, water environment federation (2017), standard methods for the examination of water and wasewater 23rd edition. baird rb, eaton ad, rice ew (edss). publication office, american public health association, washington, dc. ang wl, mohammad aw, hilal n, leo cp (2015), review on the applicability of integrated/hybrid membrane processes in water treatment. desalination 363 (1): 2-18. arij (2015), status of the environment in the state of palestine. applied research institute – jerusalem. cmwu (2012), environmental and social impact assessment (esia) and environmental and social management plan (esmp) for gaza water supply and sewage systems improvement project (wsssip) phase 1 and additional financing (af). coastal municipalities water utility. edwards vh, schubert pf (1974), removal of 2,4-d and other persistent organic molecules from water supplies by reverse osmosis. j am waterworks assoc 12: 610616. fang h, chian e (1976), optimization of ns‐100 membrane for reverse osmosis. applied polymer science 20(2): 303-314. fao (1985), the state of food and agriculture. food and agriculture organization. fao (1992), wastewater treatment and use in agriculture. food and agriculture organization. fao (2006), water desalination for agricultural applications. proceedings of the fao expert consultation on water desalination for agricultural applications, food and agriculture organization. fao-aquastat (2018), global information system on water and agriculture. http://www.fao.org/nr/water /aquastat/main/ index.stm (accessed in june 2018). gumbo jr, malaka em, odiyo jo, nare l (2010), the health implications of wastewater reuse in vegetable: a case study from malamulele, south africa. international journal of environmental health research 20(3): 201-211. jacobson ks, drew dm, he z (2011), efficient salt removal in a continuously operated up-flow microbial desalination cell with an air cathode. bioresour technol 102 (1): 376–380. jamil m, zia ms, qasim m (2010), contamination of agro-ecosystem and human health hazards from wastewater used for irrigation. journal of chemical society of pakistan 32(3): 370-378. jan fa, ishaq m, khan s, ihsanullah i, ahmad i, shakirullah m (2010), a comparative study of human health risks via consumption of food crops grown on wastewater irrigated soil (peshawar) and relatively clean water irrigated soil (lower dir). j. hazard mater 179: 612-621. jaradat d. (2016), reality and challenges of water quality in palestine: focus on regulations and monitoring of wastewater treatment and reclaimed water use. master thesis. birzeit university, palestine jaramillo mf, restrepo i (2017), wastewater reuse in agriculture: a review about its limitations and benefits. sustainability 9(10): 1-19. jiménez b (2005), treatment technology and standards for agricultural wastewater reuse: a case study in mexico. irrigation and drainage 54 (1): 23-33. karagiannis ic, soldatos pg (2008), water desalination cost literature: review and assessment. desalination 223 (1-3): 448–456. kawamura s. (2000), integrated design and operation of water treatment facilities. 2nd edition, canada, john wiley & sons, inc. mehta l (2007), whose scarcity? whose property? the case of water in western india. land use policy 24 (4): 654–663. metcalf and eddy (2013), wastewater engineering treatment and reuse. 4th edition. mcgraw-hill: usa. middle east and north africa. irrigation and drainage systems 24(1-2): 37–51. pcbs (2018), water and wastewater in palestine. iop publishing physics web. http://www.pcbs.gov.ps accessed 25 june 2018. pearce gk (2008), uf/mf pre-treatment to ro in seawater and wastewater reuse applications: a comparison of energy costs. desalination, 222(1-3): 66-73. polprasert c. (2007), organic waste recycling technology and management. 3rd edition, london, uk, iwa publishing. ps (2003), the palestinian standards for treated wastewater. palestinian standard. pwa (2011), the comparative study of options for an additional supply of water for the gaza strip (cso-g). the pwa (2013), status report of water resources in the occupied state of palestine-2012. palestinian water authority, ramallah, palestine. pwa (2014), gaza water resources status report, 2013/2014. palestinian water authority, gaza, palestine. pwa (2015), gaza strip: desalination facility project: necessity, politics and energy. palestinian water authority, gaza, palestine. pwl (2002). palestinian water law, decree no 3: 2002. pwl (2014) palestinian water law, decree no 14: 2014. qadir m, bahri a, sato t, al-karadsheh e (2010), wastewater production, treatment, and irrigation in the production, treatment, and irrigation in the middle east and north africa. irrigation and drainage systems 24(1-2): 37–51. suzuki y, minami t (1991), technological development of a wastewater reclamation process for recreational reuse: an approach to advanced wastewater treatment featuring reverse osmosis membrane. water science and technology 23: 16-29. u.s. epa (2004), guidelines for water reuse. epa/625/rmazen abualtayef, abedelmajeed kahail, hassan al-najjar, and thaer abushbak 23  04/108. u.s. environmental protection agency, washington. updated final report, palestinian water authority, gaza, palestine. vörösmarty cj, mcintyre pb, gessner mo, dudgeon d, prusevich a, green p, glidden s, bunn se, sullivan ca, reidy liermann c, davies pm (2010), global threats to human water security and river biodiversity. nature 467: 555–561 water technology (2018), groundwater replenishment system (gwrs), orange county, california. https://www.water-technology.net/projects retrieved 15 july 2018. who (1989), health guidelines for the use of wastewater in agriculture and aquaculture. the report of the scientific group on health aspects of the use of treated wastewater for agriculture and aquaculture. the world health organization, geneva. who (2006), a compendium of standards for wastewater reuse in the eastern mediterranean region. the world health organization, regional office for the eastern mediterranean, regional centre for environmental health activities, ceha. zhang b, he z (2013), improving water desalination by hydraulically coupling an osmotic microbial fuel cell with amicrobial desalination cell. journal of membrane science 441: 18–24. chen060214.qxd the journal of engineering research vol. 5, no.1 (2008) 47-54 1. introduction activated carbon is used in many applications due to its adsorption capacity and large internal surface area. these applications are partly facilitated by the fact that the carbon precursor or the preparation conditions could be chosen to achieve a unique set of desirable properties (luehrs, et al. 1998; wartelle, et al. 2000 and byrne, et al. 1997). both naturally occurring and synthetic carbonaceous solid have been used (gergova, et al. 1997; jagtoyen, et al. 1992; benaddi, et al. 2005; yalcm, et al. 2000 and nagano, et al. 2000). though new synthetic precursors such as polymeric fibers and phenolic-resin have been reported to produce high surface area products (cai, et al. 2004), agricultural wastes are increasingly being investigated as precursors (wartlle, et al. 2004; byrne, et al. 1997; gergova, et al. 1996; yalcm, 2000; liou, 2004 and carvalho, et al. 2003). wood is the most favored source of activated carbon production (35%) followed by coal (28%), lignite (14%), coconut shell (10%), peat (10%) and others (3%) (byrne, et al. 1977). _____________________________________ *corresponding author’s e-mail: baba@squ.edu.om other lignocellulosic materials are also used which include nutshell (wartelle, et al. 2000), apricot stones (gergova, et al. 1996), rice husks (yalcm, et al. 2000), cork waste (carvalho, et al. 2003) and dates pits (caturla, et al. 1991; molina-sabio, et al. 1995; wigmans, et al. 1989; pelekani, et al. 2000 and daifullah, et al. 2003). the activating agents for preparing activated carbons can be used before or after carbonizing the raw materials. some of the activating chemicals used in this process are boric acid, phosphoric acid, nitric acid and zinc chloride (girgis, et al. 2002 and el-hendawy, 2006). once carbonization is achieved, the agents most widely used to activate the char are: steam, carbon dioxide, air or their combinations. for example, when phosphoric acid is used as an activation chemical, the acid react with the lignocellulosics thermal decomposition products. this leads to bond weakening and formation of cross-linked structure. this reduces the release of volatile materials, restricts tar formation and loss of porosity (teng, et al. 1998). in addition, the effect of the chemical agent leads to lower activation temperature and higher carbon yields (el-akkad, et al. 19978). study of date palm stem as raw material in preparation of activated carbon o. houache1, r. al-maamari2, b. al-rashidi2 and b. jibril*2 1oman polypropylene llc, sohar industrial port complex, p.o. box 277, pc 322, falaj, al-qabail, sohar, oman *2department of petroleum and chemical engineering, sultan qaboos university, p.o. box 33, al khod, pc 123, muscat, oman received 14 february 2006; accepted 10 march 2007 abstract: activated carbon adsorbent was prepared using omani date palm tree stem as a precursor. precursor samples were subjected to thermal treatment (at 400, 500 and 600 oc) before or after impregnation with either h3po4 (85 wt %) or koh (3 wt %). the activated carbon obtained was characterized by bet (surface area and porosity), gas pycnometry (true density) and sem (texture). sample subjected to carbonization, without chemical activation, exhibited low surface areas ~ 1.0 m2/g at 400 and 500 oc and 124 m2/g at 600 oc. further treatment of such samples with either the acid or the base did not show improvement in surface area or other properties. impregnations of the precursor with acid before carbonization significantly improved the surface area to as high as 1,100 m2/g at a carbonization temperature of 500 oc. thus, activated carbon with a moderate surface area could be produced from date palm stem using low carbonization temperature. keywords: activated carbon, chemical activation, bet surface area, porosity, sem, date palm stem §°ûæÿg ¿ƒhôµdg ò° – ‘ ωén ioéªc πîædg ´hòl ωg~îà°sg á°sgqo ¢tgƒm ôªy1…ôª©ÿg ~°tgq ,2…~°tgôdg q~h ,22*πjèl éhéh , áá°°uuóóÿÿgg,500 ,400 iqgôm äélqo ≈∏y ájqgô◊g á÷驪∏d â©° ncg πîædg ´hòl äéæ«y ,§°ûæÿg ¿ƒhôµdg êéàfe’ ωén ioéªc »∏ùg πîædg ´hòl ωg~îà°sg ” á°sgq~dg √òg ‘ : á«ë£°ùdg ámé°ùÿg ¢sé«b ≥jôw øy §°ûæÿg ¿ƒhôµdg ioƒl º««≤j ” .…~yéb hg »° ªm ∫ƒ∏fi ‘ π«ªëàdg h ôª¨dg ∫ón øe »f髪«µdg §«°ûæàdg ~©h hcg πñb ájƒäe álqo 600 h â©° ncg »àdg iônc’g äéæ«©déh áfqé≤e á«ë£°ùdg ámé°ùÿg å«m øe i~«l èféàf â£ycg ájqgô◊g á÷é©ÿg πñb ¢ ª◊éh é¡à÷é©e ” »àdg äéæ«©dg .á«æñdg h áaéãµdg h á«eé°ùÿg h .ájqgô◊g á÷é©ÿg ~©h é«f髪«c é¡£«°ûæj ” »àdg ∂∏j hg ≥ñ°ùe »f髪«c §«°ûæj ¿h~h ájqgô◊g á÷驪∏d áá««mmééààøøÿÿgg ääggooôôøøÿÿgg.πîædg ´hòl ,êhîµd’g ô¡ûéh í°ùÿg ,á«eé°ùÿg ,á«ë£°ùdg ámé°ùÿg ,»f髪«µdg §«°ûæàdg ,§°ûæÿg ¿ƒhôµdg : 48 the journal of engineering research vol. 5, no.1 (2008) 47-54 based on the foregoing, we found it of interest to study different lignocellulosics materials as sources of activated carbon. here, we report phosphoric acid and potassium hydroxide treatment of an omani date palm stem for preparing activated carbon. oman and neighboring areas have millions of date-palm-trees. every year tons of date palm tree trunk and stem are disposed off. this gives a free or low-cost, abundantly supplied precursor. as stated above, many workers have reported different aspects of production of activated carbon from date pits, but to our knowledge, there are no reports on the use of date-palmtree stem for such endeavor. the aim of this work was to prepare and characterize activated carbons from date palm stems using chemical activation method which will contribute to further understanding the effects of activating chemical on the precursor's surface and other properties of the final products. 2. experimental setup activated carbon was prepared and characterized at the petroleum and chemical engineering laboratory, sultan qaboos university, oman. 2.1 sample preparation date palm stems were obtained from nizwa region of oman. they consist of a major outer part having softmesh and channel-like structure and inner hard part having longitudinal fibrous structure, as shown in fig. 1. samples were prepared by cutting slices of 1 cm thick from different parts of the stem before drying them in a muffle heater at 105 oc for 24 hrs. longer drying time did not show observable change in the samples weight. the dried samples were then used as precursors to obtain char and activated carbon. before or after activation (impregnation) with either h3po4 (85 wt %) or potassium hydroxide koh (3 wt %) the weighed samples were carbonized in a muffle furnace under nitrogen atmosphere. the chemical activation was accomplished by having 30g of each dried sample immersed in 100g of activating agents [either h3po4 (85 wt%) or koh (3 wt%)] at room temperature then increasing the heating temperature to 85 oc at a heating rate of 15 oc/min. the sample was exposed to this temperature for 3 hrs before cooling to ambient temperature and left at rest for 16 hrs. these treated samples were then dried at 105oc. in all cases, carbonization was done under flow of nitrogen (250 ml/min). it took place by heating the samples at 100 oc for 20 min before raising the temperature in steps of 50 oc up to the desired temperature (400, 500 or 600 oc). the final temperature was held constant for two hours before cooling the sample to ambient temperature. density and surface analysis took place after grounding and sieving the samples to sizes between 212 and 250 microns. in the test, , the samples are represented based on the activation procedure as follows: s-untreated = date palm stem; s-c = carbonized date palm stem; s-c-h3po4 = carbonization of date palm stem followed by acid activation; s-c-koh = carbonization of date palm stem followed by base activation; s-h3po4-c = acid activation followed by carbonization of date palm stem and s-kohc = base activation followed by carbonization of date palm stem. 2.2 characterization of samples samples of the activated carbon were characterized by physical adsorptions of n2 at 77.4 k, using a quantachrome autosorb 1c analyzer. the properties obtained are: (i) bet surface area (sbet, m2/g), (ii) total pore volume from nitrogen held as liquid at p/po = 0.95 (vp, cm3/g) and (iii) average pore diameter. prior to the analysis, samples were cleaned by heating them, under flow of helium, at 300 oc for 3 hrs. densities of the samples were measured by displacement of helium using a quantachrome ultrapycnometer 1000. the textures of samples of the products were observed using sem scanning electron microscope (jeol jsm 840a). it was equipped with an energy dispersive x-ray analyzer that allowed semi-quantitative analysis of solids. 3. results and discussion the precursor was subjected to different carbonization times and activations and the decreases in sample weight were recorded. the char yield of date palm stem was compared with those obtained from other cellulose materials such as almond shell and coconut shell, in fig. 2. it could be observed that the char yielded from date palm stem is substantially higher than that from other lignocelluloic sources (al-rashidi 2004). carbonization without chemical impregnation of the precursor yielded char that exhibited surface areas of about 1.0 m2/g at 400 and 500 oc. the area increased to 124 m2/g at 600 oc. the yields were improved by impregnations of the precursor with either h3po4 or koh before carbonization. the surface area also increased. the base treatment increased the area figure 1. sample of the date palm stem 49 the journal of engineering research vol. 5, no.1 (2008) 47-54 by one order of magnitude at 400 or 500 oc, but led to an inferior product at 600 oc, as shown in table 1. on the other hand, acid treatment improved the surface areas at all temperatures: 690.9, 1101.0 and 632.2 m2/g at 400, 500 and 600 oc respectively. this suggests that by chemical treatment of date palm stem, the chemistry of the carbonization was changed. due to dehydration, different macromolecular networks were formed which resulted in promotion of cross linking of structure and suppression of tar formation during the carbonization (attia, 1997). furthermore, the increase and decrease in surface areas with the chemical treatment, indicate the complex and interacting effects of different changes in the sample. due to different degrees of interaction, some volatiles components may react and link to carbon structure, or other compounds could be formed thereby changing the product porosity and the extent of exposed surface area. the characteristic of the product was further explored by measuring the true densities. figure 3 shows increase in densities with increase in carbonization temperature. when the char formed was subjected to post-treatment with either acid or base, different behaviors were observed. for the char produced at 400 oc, impregnation with koh exhibited a little decrease in density. however, at higher temperatures, there was an increase of the density by 21.2 and 5.8% at 500 and 600 oc respectively. on the other hand, samples treated with acid, show a decrease in the densities; 33.5, 33.3 and 8.8% at 400, 500 and 600 oc respectively. in all cases, the samples were treated to adjust the ph of the water in which the char was suspended. depending on the surface morphology of the char, the koh may be binded with surface functional groups (shih, et al. 2005). such interactions would lead to corsslinking reactions with some of the volatiles thereby retaining them in the solid matrix. this would result into increase in the net mass of the sample, thereby increasing the density. addition of h3po4 on the other hand, may lead to dehydrogenation and release of gases, thereby decreasing the net mass and density of the char. impregnation of the precursor prior to carbonization lead to partly physical diffusion of the chemical to the internal structures of the samples and partly chemical reaction with compound in the structure. carbonization resulted into chars of different characteristics. their differences decrease with an increase in the temperature (fig. 3). all samples appear to converge to constant similar densities. this implies that the final products obtained may have similar true densities (shih, et al. 2005). the sem micrographs of (a) date palm stem and chars produced at (b) 400, (c) 500 and (d) 600 oc are shown in fig. 4. the micrograph of the precursor exhibited a dense 0 10 20 30 40 50 400 500 600 carbonization temperature, oc c ha r y ie ld (% ) date stem coconut shell regenerated cellulose figure 2. comparison of char yields for different precursors at different carbonization temperature temperature ( oc) sample 400 500 600 s-c 0.0936 0.7700 119.0 s-c-h3po4 1.2600 0.3610 31.50 s-h3po4-c 691.0 1100.0 632.0 (sbet, m2/g) s-koh-c 27.80 29.40 94.70 s-c 539.0 295.0 28.20 s-c-h3po4 104.0 67.00 41.40 s-h3po4-c 36.90 41.60 37.30 avg. pore diameter (å) s-koh-c 45.90 50.70 36.00 s-c 0.0012 6 0.00568 0.0839 s-c-h3po4 0.0033 0.00061 0.0326 s-h3po4-c 0.6370 1.1500 0.5890 (vp, cc/g) s-koh-c 0.0320 0.0372 0.0852 table 1. surface area, acverage pore diameter, and total pore volume of samples prepared at different heat treatment 1 1.5 2 2.5 350 450 550 650 carbonization temperature, oc d en si ty , ( cm 3 / g) s-c s-h3po4-c s-koh-c s-c-h3po4 s-c-koh figure 3. density of different samples after carbonization 50 the journal of engineering research vol. 5, no.1 (2008) 47-54 structure and smooth surface. the chars show irregular and flaky texture. interconnecting large pores that increase with the temperature could be observed. this suggests the increase in the release of the volatile compounds from the samples, in line with the increase in the true density. the surface textures were changed when the chars were subjected to h3po4 and koh (fig. 5). the texture became denser (higher for acid treated fig. 5(a-c), than base-treated samples fig. 5(d-f)) and exhibited pores of different sizes and shapes. this indicates the chemical effects on the samples or physical washing of organic matter from the surface. the base-treated samples exhibited grey particles covering the surface. acid and base impregnation before carbonization resulted into more observable porous surface (fig. 6 a-c (acid) and d-f (base)). again, pores of different sizes and shapes could be observed. the higher surface area of samples (fig. 6b) than that of (fig. 6c) could be related to the fact that heating above 500 oc destroyes the tiny pores structure and therefore reduces the available surface area. the smaller pore perimeters of samples in fig. 6d-f than that of fig. 6a-c suggest that the chemical washing done by the acid is superior to the one done by the base which might have led to higher surface area. the acid-impregnated samples exhibited surface area an order of magnitude higher than the base treated ones. the base treatment resulted into interaction that produced porous surface, with shallow pores, as indicated by the sem images. surface areas and porosities for the treated carbons are summarized in table 1. the surface area increases and the average pore diameter decreases with temperature. carbonized date palm stems are characterized by poorly developed porous structures of low surface area (0 119 m2/g), low pore volumes (0.00126 0.0839 cm3/g) and medium to large average pore diameter (28.2 539.0 å). subsequent acid treatment of chars reduced porosity development with a decreased surface area (0.36-31.5 m2/g), decrease in the pore volume (0.0006-0.0326 cc/g) and consequent increase in the mean pore dimension (41.4-104 å) compared to carbonization. perhaps inorganic materials started to react at lower temperature but on heat treatment to higher temperature when the pores were being reorganized; inorganic materials might have filled or blocked some portion of the macropores (lua, et al. 2004). activation with phosphoric acid and then carbonization at 500 oc led to the best porous product with surface area of 1100 m2/g, internal porosity of 1.15 cm3/g and a mean pore dimension of 41.6 å. it can be observed a b c d figure 4. sem micrographs of date palm stem chars produced at different temperatures: (a) precursor (as is), (b) 400oc (c) 500oc, (d) 600oc 51 the journal of engineering research vol. 5, no.1 (2008) 47-54 a b c d e f figure 5. sem micrographs of chars produced at different temperature and subjected for either h3po4 or koh: (a) s-400-acid, (b) s-500-acid, (c) s-600-acid (d) s-400 base, (e) s-500-base, (f) s-600-base 52 the journal of engineering research vol. 5, no.1 (2008) 47-54 a b c d e f figure 6. sem micrographs of chars produced by subjecting the precursor to either h3po4 or koh and then carbonized at different temperature: (a) s-acid-400, (b) s-acid-500, (c) s-acid-600, (d) s-base-400, (e) s-base-500, (f) s-base-600 53 the journal of engineering research vol. 5, no.1 (2008) 47-54 that activation with phosphoric acid and then carbonization showed superiority under all treatments. the improvement in surface characteristics when comparing phosphoric acid activation to carbonization were very highly significant; also phosphoric acid activation resulted in a highly significant improvement over potassium hydroxide activation. the results showed that the date palm stem chars consist mainly of macropores and mesopores at low treatment temperature. this result may be further discussed based on the fact that the proportion of a particular porosity generated in the active carbon depends on the preparation conditions, natural and chemical composition of the precursor (plante, et al. 1988). the conversion of cellulose material to solid carbon yields porous carbon. the porosity depends on the original anatomy of the precursor. the activation treatment given; either by heat treatment or by chemicals results in the formation of surface oxygen complexes. this also results in the increase in the aromaticity of carbon surface due to cross-linking and polymerization of organic molecules, removal of volatiles and opening of blocked pores as the heat treatment temperature is raised (kidena, et al. 2004). during the process of activation, the spaces between the elementary crystallites become cleared of various carbonaceous compounds. this complex and different degrees of physical and chemical changes in the sample led to similarly irregular surface area, porosity and texture of the samples. 4. conclusions porous activated carbon adsorbent has been prepared from date palm trees. high carbon yield was observed. the products exhibited dominance of meso and macropores, especially at low carbonization temperature (400 500oc). generally, phosphoric acid treated precursor yielded higher porosity. a surface area of 1100 m2/g was obtained from the precursor impregnated with h3po4 and carbonized at 500 oc. the study has shown that date palm trees stem could be used as raw material for the preparation of activated carbon with good surface properties. more investigations are necessary to give further insight into the nature of the surface functional groups and other characteristics of the product. acknowledgments the financial assistant of sultan qaboos university is gratefully acknowledged. we would like to thank mr. saif almaamari of the college of science, sultan qaboos university, for the sem analysis. references al-rashdi, b., 2004, preparation of activated carbon from date palm stem,” final year student project, department of petroleum & chemical engineering, sultan qaboos university, oman. attia, a.a., 1997, "some surface properties of activated carbons prepared by gasification with different gases," adsorpt. sci. technol. vol. 15, pp. 707-715. benaddi, h., bandosz, t.j., jagiello, j., schwarz, j.a., rouzaud, j.n., legras, d. and béguin, f., 2005, "surface functionality and porosity of activated carbons obtained from chemical activation of wood," carbon, vol. 38, pp. 669-674. byrne, e. and nagle, d.c., 1997, "carbonization of wood for advanced materials applications," carbon, vol. 35, pp. 259-266. cai, q., huang, z.h., kang, f. and yang, j.b., 2004, "preparation of activated carbon microspheres from phenolic-resin by supercritical water activation," carbon, vol. 42, pp. 775-783. carvalho, a.p., cardoso, b., pires, j. and brotas de carvalho, m., 2003, "preparation of activated carbons from cork waste by chemical activation with koh," carbon, vol. 41, pp. 2873-2876. caturla, f., molina sabio, m. and rodríguez-reinoso, f., 1991, "preparation of activated carbon by chemical activation with zncl2 ," carbon, vol. 29, pp. 9991007. daifullah, a.a.m. and girgis, b.s., 2003, “impact of surface characteristics of activated carbon on adsorption of btex,” physicochem eng. asp. vol. 214, pp. 181-193. gergova, k. and eser, s., 1996, "effects of activation method on the pore structure of activated carbons from apricot stones," carbon, vol. 34, pp. 879-888. girgis, b.s. and el-hendawy, a.a., 2002, "porosity development in activated carbons obtained from date pits under chemical activation with phosphoric acid," micro. meso. mater. vol. 52, pp. 105-117. jagtoyen, m., thwaites, m., stencel, j., mcenaney, b. and derbyshire, f., 1992, "adsorbent carbon synthesis from coals by phosphoric acid activation," carbon, vol. 30, pp. 1089-1096. kidena, k., matsumoto, m., katsuyama, s., murata, s. and nomura, m., 2004, “development of aromatic ring size in bituminous coals during heat treatment in the plastic temperature range,” fuel proc. techn. vol. 85, pp. 827-835. liou, t.h., 2004, "evolution of chemistry and morphology during the carbonization and combustion of rice husk," carbon, vol. 42, pp. 785794. lua, a.c. and yang, t., 2004, "effects of vacuum pyrolysis conditions on the characteristics of activated carbons derived from pistachio-nut shells," j. coll. interface sci. vol. 276, pp. 364-372. luehrs, d.c., rogers, t.n. and godbole, k.a., 1998 "partition coefficients of organic solutes between water and activated carbon," encyclopedia of environmental analysis and remediation, meyers ra, (ed): john wiley & sons. pp 3404-3415. molina-sabio, m., rodríguez-reinoso, f., caturla, f. and sellés, m.j., 1995, "porosity in granular carbons activated with phosphoric acid," carbon, vol. 33, pp. 1105-1113. 54 the journal of engineering research vol. 5, no.1 (2008) 47-54 nagano, s., tamon, h., adzumi, t., nakagawa, k. and suzuki, t., 2000, "activated carbon from municipal waste," carbon, vol. 38, pp. 915-920. pelekani, c. and snoeyink, v.l., 2000, "competitive adsorption between atrazine and methylene blue on activated carbon: the importance of pore size distribution", carbon, vol. 38, pp. 1423-1436. plante, p., roy, c. and chornet, e., 1988, "co2 gasification of wood charcoals derived from vacuum and atmospheric pyrolysis," can. j. chem. eng. vol. 66, pp. 307-312. shih, c.c. and chang, j.r., 2005, "genesis and growth of platinum subnano-particles on activated-carbon characterized by x-ray absorption spectroscopy: effects of preparation conditions," mater. chem. phy, vol. 92, pp. 89-97. teng, h., sheng yeh and li-y hsu, 1998, “preparation of activated carbon from bituminous coal with phosphoric acid activation,” carbon, pp. 1387-1395. wartelle, l., marshall, w., toles, c. and johns, m., 2000, "comparison of nutshell granular activated carbons to commercial adsorbents for the purgeand-trap gas chromatographic analysis of volatile organic compounds," j. chrom. a, vol. 879, pp. 169-175. wigmans, t., 1989, "industrial aspects of production and use of activated carbons," carbon, vol. 27, pp. 1322. yalçιn, n. and sevinç, v., 2000, "studies of the surface area and porosity of activated carbons prepared from rice husks," carbon, vol. 38, pp. 1943-1945.