AP08_5.vp 1 Introduction The availability of solar radiation data and the relevant meteorological parameters are important to solar engineers and architects in order to give an accurate estimate of the available solar energy resource. Solar radiation data is always a necessary basis for the design of any solar energy conversion device and for a feasibility study of the possible use of solar energy [1]. The sunshine duration at a given site depends on the topography of the site and the prevailing meteorological conditions, such as the clearness of the sky and the height above sea level, water vapor content, air temperature, pres- sure, humidity, wind direction and force, etc. [2]. The first attempt at estimating global solar radiation was the well-known empirical relation between global solar radia- tion under clear sky conditions and bright sunshine duration, given by Angstrom, see [3]. Theoretical and empirical models have been postulated to compute the components of the inso- lation [4–13]. Some of these models are theoretical, dealing with the solution of the radiative transfer equation, while others are simply regression models. In this paper o give an accurate estimate of the available solar energy resource.sitesvaries an empirical sunshine-based model is applied to match observed values of the global solar radiation of selected geographical sites in Egypt. This paper is simply a continuation of several previous studies conducted at different sites in Egypt to define the degree of empirical in- solation models in the examined area [14–16]. 2 Methodology The original form, proposed by Angstrom, expressed the correlation between clear sky global solar radiation and sun- shine duration as follows: G G a b S Nc� �{ ( )}. (1) The linear relation correlating G G0 and S N is given in [4] as: G G a b S N0 � �{ ( )}. (2) Where G0 is the extraterrestrial solar radiation on a hori- zontal surface in kW/m2, given by: G I ESC0 0 24 180 � � � � � � �� � � � � � �� � �cos cos sin sin sin . (3) Where E0 is the correction factor of the Earth’s orbit and � is the sunrise/sunset hour angle given by: E dn 0 1 0 033 2 365 � � � � � � �. cos � , (4) � � �� �cos (tan tan ).1 (5) The declination angle of the sun � is given in degrees according to [17] as: � � � � � ( . . cos . sin . cos 0 006918 0 399912 0 070257 0 006758 2 � � � � � � � � � 0 000907 2 0 002697 3 0 00148 3 180 . sin . cos . sin ) , � (6) where � is the day angle in radiance, it is represented by: � � �2 1 365 �( )dn . (7) t-statistics [18] is applied as an indicator to select a coeffi- cient of the empirical best method that gives the smallest per- centage error, in the estimated G-values, t n � � � � � � � � ( )( ) ( ) ( ) 1 2 2 2 1 2MBE RM E MBES , (8) where (RMSE) is the root mean square error and (MBE) is the mean bias error to assess the performance of the relativity model, and n is the number of data pairs. The absolute percentage error of the estimated values of the global solar radiation at each site may be calculated from the following equation: � % � � � G G G m es m 100. (9) 3 Observational data Observations of the total solar radiation were carried out using a Pyronometer with sensitivity 9 �V/Wm2. The Pyrono- meter was originally introduced by Kimball and Hobbs in 1923. The detector is a differential thermopile with the hot- -junction receiver blackened and the cold-junction receiver whitened. 48 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 48 No. 5/2008 An Empirical Method for Estimating Global Solar Radiation over Egypt S. A. Khalil, A. M. Fathy Global solar radiation has been estimated on the basis of measurements of sunshine duration for different selected sites in Egypt; (Marsa-Matruh, Cairo, Aswan, Al-Kharga, Abu-Simble and Halaib-Shalatin). The regression coefficients (a) and (b) of Angstrom type correlation are calculated for the selected sites. The values of the regression coefficients are found to vary from 0.219–0.611 and 0.107–0.576, respectively. These values have been calculated by three different approaches. The estimated values of the global solar radiation are compared with the measured values. Although the (a) and (b) values differ from one site to another; the summation (a+b) is almost the same for the selected sites. The difference between the estimated and measured values of the global solar radiation at the various sites varies from 4 % to 12 %. Keywords: Regression coefficient, sunshine duration, global solar radiation. The calibration of the Pyronometer readings was carried out by the Egyptian Meteorological Authority, and the de- fined errors of the observations range from 5 % to 8 %.The data in this paper was obtained from the Meteorological Au- thority of Egypt. The data was gathered at six selected sites in Egypt: Marsa-Matruh (lat. 310°33´N & long. lat. 270°35´E), Abu-Simble (lat. 220°34´N & long. lat. 310°63´E), Cairo (lat. 300°05´N & long. lat. 320°17´E), Aswan (lat. 230°58´N & long. lat. 320°47´E), Al-Kharga (lat. 250°27´N & long. lat. 300°34´E) and Halaib-Shalatin (lat. 230°30´N & long. lat. 340°30´E). The data was gathered at the Marsa-Matruh, Cairo and Aswan sites during the period (1991–1993) and at the Abu- -Simble, Al-Kharga and Halaib-Shalatin sites during the period (1992–1995). 4 Computations Substituting the global solar radiation data in equation (2), we obtained the values of the regression coefficients con- stant a and b for the examined sites by least square method. The a and b values were determined for different meth- ods. In the first method, we substituted by the daily data values, while in the second method, the constants were calcu- lated using the monthly data values of G G0 and S N for each month according to the available data. © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 49 Acta Polytechnica Vol. 48 No. 5/2008 Methods Marsa-Matruh Cairo Aswan Abu-Simble Al-Kharga Halaib-Shalatin (1) a 0.351 0.449 0.219 0.241 0.311 0.191 b 0.406 0.281 0.553 0.502 0.429 0.582 a+b 0.757 0.730 0.772 0.743 0.740 0.773 (2) a 0.249 0.461 0.446 0.472 0.291 0.490 b 0.576 0.259 0.297 0.281 0.439 0.223 a+b 0.825 0.720 0.743 0.753 0.730 0.713 (3) a 0.338 0.568 0.596 0.611 0.609 0.593 b 0.425 0.215 0.174 0.107 0.137 0.098 a+b 0.763 0.783 0.770 0.718 0.746 0.691 Table 1: Values of Angstrom coefficients given by different methods at the selected sites Month G0 S/N Gm Method 1 Method 2 Method 3 Ges � % Ges � % Ges � % Jan. 6.63 0.911 4.64 4.95 6.6 4.70 1.3 4.76 2.6 Feb. 7.75 0.939 5.61 6.15 9.5 5.93 5.7 6.10 8.7 Mar. 9.19 0.825 6.51 6.09 6.5 6.36 2.3 5.91 9.3 Apr. 10.61 0.841 7.54 7.91 4.9 7.47 0.9 7.64 1.4 May 11.34 0.826 7.84 7.64 2.5 7.72 1.5 7.45 5.0 Jun. 11.04 0.853 8.20 8.42 2.7 8.23 0.4 7.82 4.6 Jul. 10.54 0.947 7.76 8.12 4.8 7.92 2.1 7.76 0.1 Aug. 10.32 0.924 7.53 7.33 2.7 7.62 1.3 7.44 1.3 Sep. 10.07 0.886 6.99 6.80 2.6 7.17 2.6 6.99 0.04 Oct. 8.58 0.850 6.19 6.45 4.2 6.59 6.6 6.50 5.1 Nov. 7.56 0.892 5.32 5.63 5.8 5.51 3.7 5.21 2.1 Dec. 6.42 0.814 4.55 4.36 3.2 4.64 2.0 4.47 1.8 MBE 0.193 0.136 0.072 RMSE 0.783 0.491 0.463 T 0.625 0.511 0.439 Table 2: Comparison between measured Gm and estimated Ges global solar radiation values (kW/m 2) at Abu-Simble given by different methods In the third case, the constants were calculated using the monthly mean daily values of the global solar radiation. The different values of the constants determined using the three methods for the given locations are listed in Table 1. These constants were in turn used to recalculate the estimated values of the global solar radiation at the selected sites. A comparison of the measured and estimated global solar radi- ation values is shown in Tables 2–7. The physical significance of regression coefficients a and b is that a represents the case of overall atmospheric transmis- sion for overcast sky conditions, i.e. S N � 0, while b is the rate of increase of G G0 with S N. The summation (a b� ) denotes the overall atmospheric transmission under clear sky condi- tions or the clearness index. 5 Results and discussion For Cairo, the a-values given by the three methods are generally higher than the b-values (see Table 1). In the coastal site at Marsa-Matruh an opposite behavior is noted, i.e. the b-values are higher than the a-values. Aswan and Abu-Simble have the same behaviors, i.e. the b-values are higher than the a-values for method one, while the reverse results are pro- vided by the second and the third methods. At the Al-Kharga and Halaib-Shalatin sites, the results reveal fluctuating behav- ior of these parameters. However, the summations (a b� ) are almost the same at the examined sites, except for the results of the third method at Halaib-Shalatin, where the summation is slightly lower. In fact, under full clear sky conditions, i.e. S N� , accord- ing to Eq. (2), we find that, the values of G G0 are equal to the value of a at the examined location. The applied empirical methods give estimated values of global solar radiation nearly coinciding with the measured values at the various selected sites, where the errors range from 4 % to 12 % (see Tables 2–7). According to Eq. (9), we obtained the values of (� %) for the different methods at all the selected locations, where we compare the values of (� %) for the different methods at all the sites. The smallest values are considered the best method, but we have to compare their MBE, RMSE and t-test values. Thus the method which gives the smallest values for the t-test is considered as the best method for estimating the global solar radiation for different selected sites with an acceptable error. In fact, it is difficult to select one empirical method that explains the time fluctuations of the observed global solar radiation values at various sites. For example, methods 1 and 2 are more applicable at Aswan, Al-Kharga and Halaib- -Shalatin, but method 2 is best for estimating the global solar radiation at these sites. At Cairo and Abu-Simble, methods 2 and 3 are more applicable throughout the various seasons, but method 3 is best for estimating at Abu-Simble, whereas method 2 is best at Cairo. At Marsa-Matruh, method 1 and 3 seem to the best for representing the trend of the measured global solar radiation values throughout the seasons. Method 1 is generally best at Marsa-Matruh. 6 Conclusion The results of this paper clearly indicate the primary importance of developing empirical approaches for formulat- ing the global solar radiation field reaching the Earth at various geographical sites in Egypt. Method two provides 50 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 48 No. 5/2008 Month G0 S/N Gm Method 1 Method 2 Method 3 Ges � % Ges � % Ges � % Jan. 6.16 0.598 3.40 3.29 3.1 3.13 7.9 3.26 4.2 Feb. 6.64 0.647 3.92 3.67 6.2 3.75 4.3 3.83 2.3 Mar. 8.24 0.689 5.26 5.07 3.5 4.99 5.1 5.07 3.6 Apr. 10.04 0.771 6.20 5.96 3.8 6.37 2.7 6.08 1.8 May 11.06 0.815 7.33 7.59 3.5 7.47 2.0 7.45 1.7 Jun. 11.51 0.859 7.99 8.11 1.5 7.77 2.7 8.17 2.3 Jul. 11.28 0.883 7.76 7.84 1.1 7.81 0.6 7.95 2.5 Aug. 10.60 0.809 7.26 7.17 1.1 7.45 2.6 7.36 1.8 Sep. 9.99 0.731 6.18 6.57 6.3 6.05 2.1 6.37 3.1 Oct. 8.09 0.702 5.26 4.97 5.3 5.07 3.5 5.37 2.1 Nov. 6.15 0.693 3.94 4.34 10.2 4.11 4.5 4.07 3.6 Dec. 5.16 0.645 3.30 3.44 4.0 3.49 5.6 3.54 7.1 MBE 0.185 0.122 0.489 RMSE 0.649 0.571 0.631 T 0.797 0.460 1.227 Table 3: Comparison between measured (Gm) and estimated (Ges) global solar radiation values (kW/m 2) at Cairo given by different methods. © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 51 Acta Polytechnica Vol. 48 No. 5/2008 Month G0 S/N Gm Method1 Method2 Method3 Ges � % Ges � % Ges � % Jan. 6.64 0.863 4.79 4.75 0.8 4.99 4.1 4.90 2.3 Feb. 7.93 0.905 5.86 5.79 1.2 5.97 1.9 5.71 2.7 Mar. 9.47 0.846 6.62 6.66 0.6 6.71 1.3 6.79 2.5 Apr. 10.91 0.849 7.98 8.12 1.8 8.19 2.6 7.95 0.4 May 11.34 0.872 8.16 8.07 1.1 8.02 1.7 8.29 1.5 Jun. 11.66 0.909 8.29 8.16 1.6 8.04 2.9 7.92 4.5 Jul. 11.60 0.913 7.92 7.81 1.4 7.80 1.5 7.76 2.1 Aug. 10.79 0.932 7.76 7.87 1.5 7.98 2.8 7.37 4.9 Sep. 10.32 0.895 7.19 7.29 1.3 7.38 2.6 6.98 2.9 Oct. 9.33 0.879 6.49 6.42 1.1 6.35 2.2 6.30 2.9 Nov. 7.88 0.845 5.87 5.80 1.2 5.97 1.7 5.77 1.6 Dec. 6.59 0.870 4.96 4.76 4.2 5.09 2.7 5.04 1.6 MBE 0.283 0.149 0.262 RMSE 0.641 0.627 0.539 T 0.539 0.391 0.473 Table 5: Comparison between measured (Gm) and estimated (Ges) global solar radiation values (kW/m 2) at Al-Kharga given by different methods Month G0 S/N Gm Method 1 Method 2 Method 3 Ges � % Ges � % Ges � % Jan. 6.52 0.875 4.70 4.51 4.1 4.79 1.9 4.75 1.1 Feb. 7.77 0.911 5.78 5.68 1.8 5.92 2.4 5.87 1.6 Mar. 9.39 0.825 6.58 6.36 3.4 6.28 4.6 6.49 1.4 Apr. 10.81 0.859 7.87 7.75 1.6 7.61 3.4 7.44 5.6 May 11.31 0.813 8.03 7.82 2.6 7.90 1.6 7.82 2.6 Jun. 11.63 0.881 8.25 8.01 3.0 7.84 5.0 8.10 1.9 Jul. 11.43 0.925 7.90 7.76 1.8 7.54 4.6 7.63 3.4 Aug. 10.66 0.961 7.70 7.62 0.9 7.59 1.4 7.73 0.4 Sep. 10.21 0.905 7.20 7.06 1.9 7.30 1.4 7.26 0.8 Oct. 9.00 0.863 6.50 6.38 2.0 6.62 1.8 6.59 1.3 Nov. 7.81 0.815 5.59 5.79 3.5 5.76 2.9 5.86 4.8 Dec. 6.44 0.85 4.77 4.86 1.8 4.70 1.5 4.96 3.9 MBE 0.171 0.69 0.059 RMSE 0.711 0.532 0.496 T 0.692 0.486 0.436 Table 4: Comparison between measured (Gm) and estimated (Ges) global solar radiation values (kW/m 2) at Aswan given by different methods 52 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 48 No. 5/2008 Month G0 S/N Gm Method1 Method2 Method3 Ges � % Ges � % Ges � % Jan. 6.82 0.882 4.83 4.95 2.6 4.69 2.8 4.60 4.8 Feb. 7.97 0.895 5.99 6.22 3.9 5.81 2.9 5.78 3.4 Mar. 9.67 0.932 6.66 6.47 2.9 6.74 1.1 6.37 4.4 Apr. 10.91 0.865 8.18 7.97 2.5 8.09 1.1 7.99 2.8 May 11.36 0.876 8.26 8.20 0.6 8.14 1.3 7.94 3.8 Jun. 11.80 0.942 8.43 8.68 2.9 8.24 2.3 8.04 4.7 Jul. 11.56 0.963 8.20 8.09 1.4 7.85 4.2 7.92 3.9 Aug. 10.99 0.951 7.90 7.74 2.0 7.66 2.9 7.54 4.5 Sep. 10.38 0.918 7.40 7.26 1.8 7.54 1.9 7.49 1.3 Oct. 9.51 0.885 6.89 6.66 3.3 6.61 4.2 6.75 2.1 Nov. 8.09 0.839 5.99 5.79 3.9 6.09 1.5 6.22 3.8 Dec. 6.74 0.874 5.14 5.27 2.5 5.34 4.0 5.24 1.9 MBE 0.231 0.181 0.211 RMSE 0.682 0.575 0.425 T 0.573 0.431 0.511 Table 7: Comparison between measured (Gm) and estimated (Ges) global solar radiation values (kW/m2) at Halaib-Shalatin given by different methods Month G0 S/N Gm Method1 Method2 Method3 Ges � % Ges � % Ges � % Jan. 5.45 0.672 2.81 2.85 1.4 3.02 7.4 3.18 13.3 Feb. 6.66 0.692 3.86 3.97 3.0 3.68 4.6 4.02 4.1 Mar. 8.47 0.715 4.97 4.80 3.5 4.89 1.6 4.67 6.1 Apr. 9.78 0.759 6.42 6.61 3.0 6.74 4.9 6.65 3.6 May 10.51 0.771 7.20 7.45 3.1 7.03 2.3 7.26 0.8 Jun. 10.91 0.815 7.87 8.21 4.3 7.75 1.5 8.04 2.2 Jul. 10.60 0.849 7.94 8.28 4.3 7.82 1.5 7.99 0.6 Aug. 10.39 0.807 7.26 7.54 3.7 7.35 1.1 7.45 2.6 Sep. 9.76 0.782 6.34 6.48 2.2 6.37 0.5 6.56 2.5 Oct. 8.11 0.731 5.14 5.32 3.6 5.38 4.8 5.49 6.8 Nov. 6.66 0.671 3.92 4.18 6.7 4.14 5.6 4.07 3.9 Dec. 5.24 0.619 3.13 3.31 5.6 3.40 8.4 3.42 9.2 MBE 0.086 0.137 0.479 RMSE 0.763 0.849 1.063 T 0.211 0.473 1.395 Table 6: Comparison between measured (Gm) and estimated (Ges) global solar radiation values (kW/m 2) at Marsa-Matruh given by different methods good agreement at the Cairo, Halaib-Shalatin and Al-Kharga sites, while at Abu-Simble and Aswan sites the third method is considered best. For Marsa-Matruh, the first method is considered best. Other topographic, climatological and envi- ronmental parameters should be inserted into the adopted empirical formula to increase the accuracy of the estimated values of the observed quantities. The dependence of coef- ficients a and b in the Angstrom formula should be tested as a function of the prevailing environmental conditions at the tested sites. List of symbols Gc clear sky global solar radiation in kW/m 2, S sunshine duration in hours, N length of the daylight in hours, G0 extraterrestrial solar radiation on a horizontal sur- face in kW/m2, E0 correction factor of the Earth’s orbit, � sunrise/sunset hour angle, dn day number in the year (1 365� �dn ), � declination of the sun, � latitude of the station, � day angle in radiance, RMSE root mean square error, MBE mean bias error, n number of data pairs, � % absolute percentage error, Gm measured values of global solar radiation, Ges estimated values of global solar radiation. References [1] Gopinathau, K. K.: Solar Sky Radiation Estimation Techniques, Solar Energy, Vol. 49 (1992), No. 1, p. 9–11. [2] Ready, S. S.: An Empirical Method for the Estimation of the Total Solar Radiation, Solar Energy, Vol. 24 (1971), p. 13. [3] Angstrom, A.: Solar and Terrestrial Radiation, Roy. Met. Soc., Vol. 50 (1924), p. 121–127. [4] Prescott, J. A.: Evaporation from a Water Surface in Re- lation to Solar Radiation, Trans. R. Soc. S. Austr., Vol. 64 (1940), p. 114–118. [5] Hourmitz, B.: Insolation in Relation to Cloudiness and Cloud Density, J. Met., Vol. 2 (1945), p. 154–156. [6] Daneshyar, M.: Solar Radiation Statistics for Iran, Solar Energy, Vol. 21 (1978), p. 345–349. [7] Davies, J., Abdel-Wahab, M., Mekay, D.: Estimating So- lar Irradiance on Horizontal Surface, Int. J. Sol. Energy, Vol. 2 (1984), p. 405. [8] Abdel-Wahab, M.: Simple Model for Estimation of Glo- bal Solar Radiation, Solar and Wind Technology, Vol. 2 (1985), No. 1, p. 69–71. [9] Srivastava, S. K., Sinoh, O. P., Pandy, G. N.: Estimation of Global Solar Radiation in Uttar Pradesh (India) and Comparison of Some Existing Correlations, Solar En- ergy, Vol. 51 (1993), p. 27–90. [10] Abdel-Wahab, M.: New Approach to Estimate Angstrom Coefficient, Solar Energy, Vol. 51 (1993), p. 241–245. [11] Bodescu, V.: Verification of Some Very Simple Clear and Cloudy Sky Models to Evaluate Global Solar Radiation, Irradiance, Solar Energy, Vol. 61 (1997), p. 251–264. [12] Hamid, R. H.: Formulation of the Global Solar Radia- tion Using Sunshine Duration over Egypt, Journal Astro- nomical Society of Egypt, Vol. 11 (2003), p. 39–52. [13] Beheary, M. M.: Using the Global Solar Radiation to Estimate the Spectroscopic Structure of the Normal Inci- dent Solar Radiation at Selected Sites in Egypt, Al-Azhar Bull. Sci. Vol. 15 (2004), No. 2, p. 93–106. [14] Kamal, M. A., Shalaby, S. A., Mostafa, S. S.: Solar Radia- tion over Egypt; Comparison of Predicted and Mea- sured Meteorological Data, Solar Energy, Vol. 50 (1993), p. 463–470. [15] Ibrahim, S. M. A.: Predicted and Measured Global So- lar Radiation in Egypt, Solar Energy, Vol 35 (1985), p. 185–190. [16] Tadros, M. T. Y.: Uses of Sunshine Duration to Estimate the Global Solar Radiation over Eight Meteorological Stations in Egypt, Renewable Energy, Vol. 21 (2000), p. 231–240. [17] Spencer, J. W.: Fourier Series Representation of the Position of the Sun, Search, Vol. 2 (1975), No. 5, p. 165–172. [18] Iqbal, M: An Introduction to Solar Radiation, Edited by Ac- ademic Press, 1983. Doc. Ahmed Mohamed Fathy Phone: +2-011-2689337 Fax: +2-02-25548020 e-mail: amfathy2003@yahoo.co.uk National Research Institute of Astronomy and Geophysics Solar and Space Department Helwan, Egypt. Doc. Samy Khalil Abd El Mordy Fax: +2-02-25548020 e-mail: Samyakalil@yahoo.com National Research Institute of Astronomy and Geophysics Helwan, Egypt. © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 53 Acta Polytechnica Vol. 48 No. 5/2008