Microsoft Word - 31-2787_s_ETASR_V9_N3_pp4235-4242 Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4235 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study Energy Design and Optimization of a Greenhouse A Heating, Cooling and Lighting Study Anouar Belkadi UR-LAPER, Faculty of Sciences of Tunis, University of Tunis El Manar Tunis, Tunisia Dhafer Mezghani UR-LAPER, Faculty of Sciences of Tunis, University of Tunis El Manar Tunis, Tunisia Abdelkader Mami UR-LAPER, Faculty of Sciences of Tunis, University of Tunis El Manar Tunis, Tunisia Abstract—This paper presents a new approach to properly optimize the energy consumption in a greenhouse. An improved intermediate modeling to establish the energy balance in a greenhouse within a higher precision was adopted. While the classical model focuses on the efficient cooling and heating demands and neglects the profound impact of lighting parameters, it was shown that these three necessary components are interdependent, and they should be taken into account all together to comfortably reach optimal crop production and energy consumption. This study’s contribution is the classical model’s improvement and the demonstration of the fact that the heat released by the luminaries and the energy used by this equipment has fundamental consequences on the energy balance as well as the preferred choice of the possible shape of the greenhouse and its adequate cover. Keywords-cooling; energy; greenhouse design; heating; lighting I. INTRODUCTION Facing a constant increase of world population, an efficient food production system is ought to aim accurately for cost minimization, to allow a sufficient supply and adequately meet the consumers’ demand. Being one of the most energy demanding sectors in modern agriculture, the greenhouse industry is in definite need of energy consumption reduction, while maintaining an adequate microclimate for cultivating crops, thus guaranteeing minimal production expenses and an optimal crop quality. Because of the considerable importance of thermal modeling of a greenhouse typically used for energy simulation, to optimize energy-efficient designs and to verify the economic feasibility of greenhouse production, studies have been conducted to develop the greenhouse thermal models and try to correctly describe the specific process of heat and mass transfer in greenhouses [1-3]. Greenhouse thermal models are generally classified into three specific groups: static, dynamic and intermediate. Static models [4-6] determine the productive capacities of heating and cooling equipment and are based on approximate heat gains and losses, depending on periodic collected meteorological data. On the other hand, intermediate models [7-14] consider the factor of sunlight, while complex dynamic models [9, 15-18] are principally based on the integration of numerous greenhouse components. Despite the relative accuracy of these models, the estimation of heating energy requirements over long periods or for a greenhouse with different configurations requires very complex modifications and does not mention the effect of evapotranspiration of plants and control systems of the environment on the greenhouse’s energy balance. In order to obtain higher yield and better production quality in the greenhouse, this study will start from the equations used in the energy balance of previous models as references. Our contribution is to develop an updated version of these models. II. METHODOLOGY A. Greenhouses Studied Shapes Four forms of greenhouses studied (elliptic, even span, uneven span and rectangular), as shown in Figure 1. Fig. 1. Analyzed greenhouse forms. The analysis of these greenhouses is carried out for the same greenhouse volume of 4420m 3 and the same dimension of the base surface (W=34m, L=40 m). B. Covering Materials of Studied Greenhouses Five different covering materials of greenhouses were studied. The characteristics of these materials are shown in Table I [9, 16, 20, 21]. C. Heat Balance Inside the Greenhouse To fix the climatic problem inside the greenhouse, several factors must be inevitably retained during the establishment of Corresponding author: Anouar Belkadi (belkadi.anouar@gmail.com) Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4236 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study the energy balance. The energy balance of the greenhouse is precisely the heat transfer and the mass exchange between the greenhouse and the external environment. This multidimensional quantity can be expressed according to (1): - = L Gs Q Q Q (1) In this energy equation, positive QS means that the heat loss is greater than the heat gain, which would imply a need for warming up the greenhouse, whereas in case this value is negative, it would indicate the heat gain is greater than the heat loss so there is necessarily a need, to cool it. We could classify the various parameters intervening in the development of the energy balance equation in two categories: • First category defines the structure of the greenhouse by its geometric shape, the diverse materials which constitute its shell, mainly in terms of materials selected for its cover, and the temperature, light and moisture to retain inside the greenhouse. • Second category: it is, essentially, related to the external environmental conditions of the greenhouse like solar radiation level, wind speed, external temperature, humidity degree, altitude, latitude and the orientation of the greenhouse. TABLE I. COVERING MATERIALS CHARACTERISTICS PF GL LEG PC Acr Index M1 M2 M3 M4 M5 Lc 0.000152 0,003 0,0032 0,01 0,006 Solar transmission 0,87 0,905 0,78 0,9 0,9 Long-wave transmission 50 3 <3 <3 <5 Kc 0,33 0,76 0,76 0,17 0,2 ACH 0,85 1,1 1,1 1,1 1,1 D. Traditional Model Design Our approach is essentially based on the results of previous intermediate models that took into account the different components that act on the energy balance. These theoretical models have paved the way for our contribution. In the traditional model the heat gain QG calculation depends only on the solar radiation as shown in (2): τ= sG r f IQ A (2) While the precise determination of the heat loss was generally based on the calculation of two fundamental parameters, the heat transfer caused by the phenomenon of conduction and convection Qcd-cv and the heat transfer due to the infiltration Qinf. The convection and conduction heat transfer is calculated by (3): ( ) G in outcd cv Q U A T T− = − (3) where U is the overall heat transfer coefficient and which was taken as general constant [1]. The heat transfer due to the Infiltration Qinf is estimated by (4): ( ) CH a a G in outinf A C VQ T Tρ= − (4) Figure 2 represents the energy demand of heating and cooling for the elliptic form equipped by polyethylene covering material applied in the climatic and geographical conditions of the Tunisian latitude, and by keeping the desired indoor temperature of the cultivated invariably to 20°C, according to the previous equations. Fig. 2. Energy demand of the heating and cooling using traditional approaches Note that despite the same temperature of some months (April and November, 17°C), (May and October, 21°C), we do not have the same behavior in terms of heating-cooling and this is expected since a sunny day with a low temperature does not mean we will have a need for heating and vice versa (a hot day with low sun does not mean that we will need ventilation). In order to represent the energy demand of all shapes and all covering materials studied in this paper, we replicated the same process for all the chosen shapes and covering materials, the results are presented in Table II. TABLE II. TRADITIONAL MODEL AVG DAILY HEATING AND COOLING APPLIED TO TUNISIA [KWHM -2 ] Even span Uneven span Elliptic Rectangular M1 4,472 4,483 4,799 4,549 M2 3,334 3,339 3,513 3,369 M3 4,940 4,956 5,387 5,055 M4 3,666 3,673 3,658 3,711 M5 4,102 4,112 4,396 4,179 We can deduct through Table II that if using the traditional model without any changes in the energy balance, the most appropriate shape is the even span equipped with M2, while the worst is the elliptic form equipped with M3. According to our analysis of the different computation models of the energy balance, the traditional model remains an ineffective one, despite it took into consideration the majority of the parameters intervening in the equation of the energy balance, since it ignored the impact of other parameters that must be included such as: • The global heat transfer coefficient U Has been defined as a constant whereas it should be dynamic since it depends on the internal and external Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4237 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study conditions of the greenhouse such as the properties of the covering material, the air velocity outside and inside the greenhouse, and the ambient and desired temperature. Therefore, the variable U cannot be considered as a constant when establishing the energy equilibrium equation of a greenhouse. • Heat loss due to long-wave radiation Qlw In general, thermal radiation emitted from inside the greenhouse can be partly absorbed by the cover, reflected in the greenhouse or transmitted to the outside of the greenhouse. Depending on the degree of transparency of the greenhouse cover, the exchange of long-wave radiation by the transparent surface will cause a significant effect on the heat loss so a considerable amount of heat will be transmitted to the outside. • Soil heat loss Qf and the heat due to the perimeter Qp The heat loss engendered from the greenhouse's floor is principally due to the conduction of the soil and the heat transfer of the perimeter. This quantity exerts an effect on the estimate of the equation of the energy balance that we should not neglect in order to increase the accuracy of the results of the energy balance equation. • Heat loss due to evaporation Qevap The evapotranspiration remains a crucially essential component that we should not neglect while establishing the equation of the energy balance because it represents a significant amount of heat loss. This quantity is produced from the plant leaves and the ground surface. • The absence of the lighting effect Qli Qli is the sufficient amount of specific heat typically emitted by the lighting system. Previous studies and researches have overlooked the energy weight of lighting in the choice of the shape and cover of the greenhouse and its impact on the energy balance equation, while our model in its analysis considers the predicted effects of lighting on the model. E. Presentation of BGHM Model Our developed model, named BGHM, is typically based on the cultivation of tomatoes inside greenhouses in the climatic and geographical conditions of the Tunisian latitude band. We took into account the neglected parameters of the traditional model: Qlw, Qf, Qp, Qevap, and the heat gains from lighting Qli. We have also recalculated the overall heat loss coefficient U which was taken static during the study of the traditional model [1], and which is in fact dynamic. We claim by our approach to highlight the direct impact of the specific shape of the greenhouse and its adequate coverage of the energy needed by acting on various components that have been severely neglected in the successful establishment of the energy balance equation: mainly cooling, heating and lighting. To perform this, we will update the traditional model by injecting the neglected parameters Qlw, Qf, Qp, Qevap with and without the heat gain from lighting Qli, as this factor has been neglected in the majority of studies who worked on the energy balance equation, in order to emphasize on the importance of the effect of the lighting component to retain an efficient system in terms of energy. 1) BGHM Program Design In order to develop our model, we have created the program presented in Figure 3. Fig. 3. Flowchart of the program 2) Study of BGHM Model without Lighting Impact We will therefore resume the traditional model and remodel it while injecting in the energy balance equation the neglected parameters Qlw, Qf, Qp, Qevap and the improvement of the Qcd-cv by adopting a dynamic heat transfer Ud. The equation Qcd-cv then becomes as follows: ( ) d G in ov tc ud c U A T TQ − = − (5) where: 1 1 1 1 0 [ ]idU h Lc Kc h − − − −= + + (6) and h0 and hi can be estimated as shown in (7)-(8): 0 2.8 1.2 V h W= + (7) 1/3 1/2 1.52[ ] 5.2 ( ) i in out CH G h T T A S L= − + (8) The heat transfer through the greenhouse floor Qf is: 1 ( ) s f if n s K d A TQ T −= − (9) The heat transfer through the greenhouse perimeter Qp, is estimated by (10): ( ) p G n oup i t F P T TQ = − (10) The long wave radiation heat loss Qlw is calculated as: 0 (1 ) ( ) l G in s yw k Q h A T Tτ= − − (11) Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4238 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study where, Tsky, is calculated by: 1.5 0.552( 273.15) -273,15 sky out T T= + (12) Finally, the heat transfer engendered by the process of evapotranspiration Qevap, is calculated as shown in (13)-(14): evap T V Q M L= (13) 1 ( ) ( ) T p M A wps wi Ra Rsρ −= − + (14) where Ap could be determined from the plant’s leaf area index, while Wps and Wi could be calculated as: 1 100.6219 1. )5( 32 ps ws ws W p p −= − (15) 1 1010 ..6 322 519 ( ) i w w W p p −= − (16) By assuming a negligible temperature difference between Tin and Tplant, we could deduce the actual vapor pressure from the relative humidity and the saturated vapor pressure as shown in (17)-(18): 33 1 6 2 1.10 . 2 3.10 4 .10 . 9 3 5. 10 . 6 . ln( ) 1 exp 1000 −− − + − + − − +    =        −  C a C C a C a C a C a P ws (17) where a=Tin +273.15, C1=-5.80002, C2=1.3915, C3=48.64024, C4=41.764768, C5=14.45209310, and C6=6.5459673. W ws h P P R= (18) In order to calculate Ra and Rs we could apply (19) and (20): 0.2 0.8 220 a f i R L V −= (19) 1 200 (1 (exp(0.05( 50))) ) sr Rs Iτ −= + − (20) We can apply the above theoretical development on the various shapes and covers of the greenhouse. Figure 4 accurately represents the need for heating and cooling of the elliptic form equipped by polyethylene covering material during the different months of the year. To conclude, we will precisely calculate the energy demand in specific terms of heating and cooling using the BGHM model without lighting potential effect to correctly determine the most efficient form and greenhouse cover (Table III). Based on the results of Table III, it is clear that the most efficient form is the elliptic if the greenhouse is adequately covered by the polyethylene film M3, while if we utilized the traditional models we would have shown that this form was the worst energy-efficient and therefore the most economically expensive. TABLE III. AVERAGE DAILY HEATING AND COOLING OF OUR MODEL WITHOUT LIGHTING IMPACT [KWH M-²] Even Span Uneven Span Elliptic Rectangular M1 3,677 3,683 3,669 3,723 M2 3,996 4,005 3,983 4,210 M3 3,462 3,469 3,452 3,514 M4 3,651 3,657 3,643 3,696 M5 3,554 3,562 3,543 3,612 Fig. 4. Estimated heating and cooling load of the elliptic shape 3) Study of BGHM Model with Lighting Impact We will calculate the artificial lighting requirements to determine the hourly lighting energy demand [W] which will allow us to deduce the heating effect of the lighting system. Starting from the fact that any lighting device emits heat as well as light, from our model we will check the impact of these two parameters on the choice of the form and the covering material which must be used. To do this we will, at first, present the need of illumination of a tomato plant (Table IV). Then we will determine the daily light to calculate the additional energy effort that meets the optimal need of the cultivation of our plant. TABLE IV. TOMATO PLANT NEEDS [23-27]. Parameters Value Daily photoperiod 16 Hours Required Illumination(1) 8475 Lux 1Lux=0.0177µmol×m -2 s -1 a) Lighting System Design • Daylight calculation To design the daylight system, we started with the 3D modeling of each shape of the selected greenhouse, then, we defined the covering material properties, the type of the reference sky (clear sky, overall sky, and covered sky), the hours and the days considered for the simulation to create the daylight scenes. Daylight from 07:00 am to 10:00 pm was simulated, for the 1st, 8th, 15th and 22th of each month. This process has been applied for all types and covering materials of the greenhouse. An example of the daylight simulation of the elliptic form equipped with M3 has been shown in Figure 5 (a) and (b) for August and December respectively. In order to compare the daylight of a different shape, we calculated the average daylight illumination of each greenhouse’s type applied for M3 covering material. The results are shown in Figure 6. By referring to Figure 6 we could deduce that the rectangular shape of the greenhouse is apparently the best in terms of daylight and that the elliptic shape remains the least efficient. In order to study the impact of coating materials on daytime lighting, we have simulated, through DIALUX EVO, all types of coating materials for a reference date and time 15/06/2018 at 1:00 pm), because the lighting depends not only Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4239 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study on the transparency but also on light reflection and absorption, and the height, size and shape of the greenhouse. The results are shown in Table V. The gain/loss illumination factor to all greenhouse types with combined effects of covering material and applied greenhouse types, is shown in Figure 7. (a) (b) Fig. 5. Daylight simulation results in (a) August, (b) December. Fig. 6. Daylight illumination of all greenhouse types. TABLE V. EFFECT OF THE COVERING MATERIALS Covering material Illumination simulated on (15-06-2018 at 1 PM) [KLux] Gain/loss compared to reference M3 [%] M1 13.185 4.5% M2 11.705 -7.6% M3 12.596 Taken as reference M4 13.182 4.4% M5 13182 4.4% Fig. 7. Average annual daylight illumination of all greenhouse types. Figure 7 shows the daylight illumination of all greenhouse types. It is clearly noticed that the rectangular greenhouse has the best value in terms of Lux level, while M3 and M4 have the lower lighting values. There is not a big difference in the rest of the material. • Lighting energy demand In order to design the required artificial lighting, we have applied the process described in Figure 8 to meet the target Lux level for the tomato plant which is 8475Lux [22]. Fig. 8. Lighting design process Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4240 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study After all calculations were done, we used 1400 luminaries of 60w suspended at 1.8m from the working plan (floor) and within an efficiency of 160lm·w -1 and a CCT of 4000K. According to the simulation, our target was achieved within an illumination of 8478 vs 8475 (target) and a consumption of about 61w·m -2 . An example of false color rendering of the simulation is shown in Figure 9, which proves that illumination is around our target light level and that the light distribution remains uniform throughout the entire greenhouse area. Fig. 9. A sample false color rendering of a lighting configuration, and light distribution map of the greenhouse To reduce the lighting energy, we have implemented a daylight system control consisting of a light sensor connected to a raspberry pi3 in order to transmit the value of the light in real time to calculate the difference between the target value and the sensor value in the PWM signal, to dim the luminaries in order to obtain a uniform illumination equal to the desired value at any time t from 7:00 am to 10:00 pm when the daily PPFD is lower than the target value of the tomato plant. Once the daily PPFD is reached, the lighting system will be automatically switched off. To estimate the average of the compensated illumination of all greenhouse types equipped with M3 we have to subtract the illumination of the total installed luminaries (when all luminaries are ON) from the daylight. Finally, we calculated the compensated need of illumination, to deduct the annual power consumption of lighting for all greenhouse types and covering as shown in Figure 10. We note that in Figure 10 the need for energy compensation in lighting remains high compared to the other greenhouse forms, especially in the case of the rectangular form. Despite the satisfactory results so far, in terms of the choice of shape and coverage and their impact on the efficiency of the lighting system, we will look further into our analysis to see if there is an impact of the heat emitted by the luminaries and, if applicable, its effect on the energy equilibrium equation and on the response of the equation to the variation of heating and cooling components. This amount of energy allowed us not only to estimate the energy requirement of the lighting, but also to estimate heat loss from the lighting equipment as mentioned in (21): ' " li l Q W k k= (21) Fig. 10. Lighting energy demand After all calculations are done, the estimated heat gains and losses are represented in Figures 11-12. Figure 11 shows the monthly evolution of energy captured by solar radiation Qsr , the evolution of the global greenhouse gain system which encompasses Qsr, and the heat effect gained from the lighting system. We also note that the difference between the heat gains due to the solar energy and the global heat gains are approximately identical during the hot period which is expected since the need of additional lighting in these months decreases because they have more daylight. Fig. 11. Monthly evolution of heat gain Fig. 12. Monthly evolution of heat loss Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4241 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study Figure 12 represents the evolution of the heat loss caused by conduction convection Qcd-cv, infiltration Qinf, perimeter heat transfer Qp and soil Qf, heat transfer due to long wave radiation Qlw, and the evapotranspiration Qevap. It can be seen from Figure 12 that the quantities of heat Qp, Qinf, Qcd-cv follow a double sign, which is expected since these quantities depend on both internal and external temperatures of the greenhouse. They are positive when the external temperature is lower than the internal temperature and vice versa. In order to compare the state of the different models in terms of energy demand (heating and cooling), the models are presented in Figure 13. Fig. 13. Monthly evolution of energy demand According to Figure 13 the energy demand estimate by the traditional model remains inaccurate, whereas with the improvement of this model by the injection of the neglected parameters, there is a clear improvement. We also note that the energy gap in cooling and heating is around 49.87% compared to the traditional model. Our developed model, even if the direct results at this elevated level are satisfactory, shows better results and allows demonstrating the considerable weight of the heat effect of the lighting on sufficient accuracy of the energy demand level in heating and cooling. The potential impact of lighting equipment’s electrical consumption for the various possible shapes and coverage materials is shown in Table VI. TABLE VI. TOTAL ENERGY DEMAND* [KWH M-²] Even Span Uneven Span Elliptic Rectangular M1 4,022 4,010 4,084 4,028 M2 4,205 4,196 4,262 4,237 M3 3,763 3,751 3,824 3,775 M4 3,996 3,984 4,058 4,001 M5 3,829 3,819 3,888 3,848 * Heating, cooling, and lighting According to Table VI, our developed model points that the uneven span is the most efficient form. Our model adequately supported the model parameters conveniently overlooked by the traditional model and the impact of the lighting system and its effects. III. CONCLUSION This research showed that the energy balance can exclusively reveal the optimal values of energy consumption if all the principal determinants that cause a direct effect on the fundamental process of greenhouse cultivation, are taken into account. From this study, the following conclusions are derived: • The need for energy consumption has been optimized by adopting the uneven span form. • In our energy equation, it is possible to confirm that the uneven span form could improve the efficiency of our energy balance more significantly by using appropriately polyethylene M3 as greenhouse cover. • The development of the heat release effect of the lighting system enabled a considerable gain in electric energy, especially during the cold period, since it enabled a residual heat to be amply provided. Therefore, a tangible net cost reduction efficient energy ordinarily required by the greenhouse tomato cultivation was observed. • The proposed model exhibited a substantial gain, reaching 49.87%, of the energy consumption of cooling and heating compared to the reference “traditional” model. NOMENCLATURE AG Greenhouse area [m²] Af Greenhouse floor area [m²] Ap Plant area [m²] Acr Acrylic ACH Number of air changes per hour [m 3 s -1 ] CCT Correlated color temperature, [ K] Ca Specific heat of air [J. kg-1 K -1 ] D Depth of constant soil temperature [m] Fp Perimeter heat loss factor, [w.m -1 . K -1 ] G Glass H Height of the greenhouse [m] h0 Outside convective heat transfer coefficients hi Ohe inside convective heat transfer coefficients Isr Solar radiation on the horizontal surface [w.m -2 ] K’ Lighting allowance factor K” Heat conversion factor Ks Thermal conductivity of soil [w m -1 K -1 ] Kc Thermal conductivity [w m -1 K -1 ] L Length of the greenhouse [m] Lf Characteristic length plant leaves [m] Lv Latent heat of water vaporization, [J kg -1 ] Lmax Maximum illumination inside the greenhouse [lux] LEG Low E glass LC Covering material thickness [m] MT Moisture transfer rate, [kg s -1 ] PF Polyethylene film PC polycarbonate PG Perimeter of the greenhouse [m] Pw Partial pressure of the water vapor, [kPa] Pws Partial pressure at saturation, [kPa] Qcd-cv Conduction and convection heat transfer [w] Qinf Infiltration heat transfer [w] Qf Floor heat transfer [w] Qp Perimeter heat transfer [w] Qlw Long-wave radiation heat transfer [w] Qevap Evapotranspiration heat transfer [w] Qs Total required heating and cooling power [w] Qsr Solar radiation heat gain [w] QLi lighting heat gain [w] QG Total heat gain inside the greenhouse [w] Engineering, Technology & Applied Science Research Vol. 9, No. 3, 2019, 4235-4242 4242 www.etasr.com Belkadi et al.: Energy Design and Optimization of a Greenhouse: A Heating, Cooling and Lighting Study QL Total heat losses from the greenhouse [w] Ra Aerodynamic resistance [s m -1 ] Rs Stomatal resistance, [s m -1 ] Rh Relative humidity SG Section of the greenhouse [m] T in Inside air temperature [°C] Tout Average outdoor air temperature [°C] Ts Greenhouse soil temperature [°C] Tsky Sky temperature [°C] Tplant Plant temperature [°C] U Static overall heat transfer coefficient Ud Dynamic overall heat transfer coefficient Vi Indoor airspeed airspeed, m s -1 VG Volume of the greenhouse [m] W Width of the greenhouse [m] Wl Lighting energy consumption [w] WV Wind velocity [m s -1 ] Wi Humidity ratio of air at indoor temperature Wps Saturated humidity ratio of air at the plant temperature ρα Density of internal air [kg m -3 ] τ Transmissivity of the greenhouse cover REFERENCES [1] S. D’Arpa, I. Petrosillo, V. F. Uricchio, G. Zurlini, G. Colangelo, G. Starace, D. E. Bruno, “Heating requirements in greenhouses farming in south of italy: evaluation of ground source heat pump utilization”, Energy Efficiency, Vol. 9, No. 5, 2015 [2] X. Qin, X. Shen, H. Sun, Q. Guo, “A quasi-dynamic model and corresponding calculation method for integrated energy system with electricity and heat”, Energy Procedia, Vol. 158, pp. 6413-6418, 2019 [3] K. A. Joudi, A. A. Farhan, “A dynamic model and an experimental study for the internal air and soil temperatures in an innovative greenhouse”, Energy Conversion and Management, Vol. 91, pp. 76–82, 2015 [4] O. Jolliet, L. Danloy, J. B. Gay, G. L. Munday, A. Reist, “Horticern: an improved static model for predicting the energy consumption of a greenhouse”, Agricultural and Forest Meteorology, Vol. 55, No. 3-4, pp. 265-294, 1991 [5] J. P. Chiapale, C. Kittas, O. De Villele, “Estimation regionale des besoins de chauffage des serres”, Acta Horticulturae, Vol. 115, pp. 493- 503, 1981 [6] L. G. Morris, “Some aspects of the control of plant environment”, Journal of Agricultural Engineering Research, Vol. 1, pp. 156–66, 1956 [7] J. J. J. Breuer, T. H. Short, “Greenhouse energy demands comparisons for the Netherlands and Ohio”, Acta Horticulturae, Vol. 174, pp. 145– 154, 1985 [8] M. Tunc, J. E. S. Venart, K. Sollows, “Bivalent (hybrid) heat pumpoil heating systems for greenhouses”, Journal of Heat Recover Systems, Vol. 5, No. 6, pp. 483–491, 1985 [9] J. M. Hill, Dynamic Modeling of Tree Growth and Energy Use in a Nursery Greenhouse Using Matlab and Simulink, Cornell University, 2006 [10] E. Fitz-Rodriguez, C. Kubota, G. A. Giacomelli, M. E. Tignor, S. B. Wilson, M. McMahon, “Dynamic modeling and simulation of greenhouse environments under several scenarios: a web based application”, Computers and Electronics in Agriculture, Vol. 70, No. 1, pp. 105–116, 2010 [11] P. Chandra, L. D. Albright, “Analytical determination of the effect on greenhouse heating requirements of using night curtains”, American Society of Agricultural and Biological Engineers, Vol. 23, No. 4, pp. 994–1000, 1981 [12] K. V. Garzoli, J. Blackwell, “An analysis of the nocturnal heat loss from a double skin plastic greenhouse”, Journal of Agricultural Engineering Research, Vol. 36, No. 2, pp. 75–86, 1987 [13] K. V. Garzoli, J. Blackwell, “An analysis of the nocturnal heat loss from a single skin plastic greenhouse”, Journal of Agricultural Engineering Research, Vol. 26, No. 3, pp. 203–214, 1981 [14] R. D. Singh, G. N. Tiwari, “Energy conservation in the greenhouse system: A steady state analysis”, Energy, Vol. 35, No. 6, pp. 2367–2373, 2010 [15] V. P. Sethi, “On the selection of shape and orientation of a greenhouse: thermal modeling and experimental validation”, Solar Energy, Vol. 83, No. 1, pp. 21–38, 2009 [16] F. Xu, S. Li, C. Ma, S. Zhao, J. Han, Y. Liu, B. Hu, S. Wang, “Thermal environment of chinese solar greenhouses: analysis and simulation”, American Society of Agricultural and Biological Engineers, Vol. 29, No. 6, pp. 991–997, 2013 [17] R. B. Ali, E. Aridhi, M. Abbes, A. Mami, “Fuzzy Logic Controller of Temperature and Humidity Inside an Agricultural Greenhouse”, 7th International Renewable Energy Congress, Hammamet, Tunisia, March 22-24, 2017 [18] G. Singh, P. P. Singh, P. P. S. Lubana, K. G. Singh, “Formulation and validation of a mathematical model of the microclimate of a greenhouse”, Renewable Energy, Vol. 31, No. 10, pp. 1541–1560, 2006 [19] http://www.holiday-weather.com/tunis/averages/ [20] S. Sanford, Reducing Greenhouse Energy Consumption-An Overview, University of Wisconsin–Extension, 2011 [21] G. Papadakis, D. Briassoulis, G. S. Mugnozza, G. Vox, P. Feuilloley, J. A. Stoffers, “Review Paper (SE—Structures and Environment): radiometric and thermal properties of, and testing methods for, greenhouse covering materials, Journal of Agricultural Engineering Research, Vol. 77, No. 1, pp. 7–38, 2000 [22] L. D. Albright, Environment Control for Animals and Plants, American Society of Agricultural Engineers, 1990 [23] G. N. Tiwari, M. Din, N. S. L. Srivastava, D. Jain, M. S. Sodha, “Evaluation of solar fraction (Fn) for north wall of a controlled environment greenhouse: an experimental validation”, International Journal of Energy Research, Vol. 26, No. 3, pp. 203–215, 2002 [24] M. Dorais, “The Use of Supplemental Lighting for Vegetable Crop Production: Light Intensity, Crop Response, Nutrition, Crop Management, Cultural Practices”, Canadian Greenhouse Conference, October 9, 2003 [25] C. A. Campiotti, G. Morosinotto, G. Puglisi, E. Schettini, G. Vox, “Performance evaluation of a solar cooling plant applied for greenhouse thermal control”, Agriculture and Agricultural Science Procedia, Vol. 8, pp. 664– 669, 2016 [26] Y. Noorollahi, P. Bigdelou, F. Pourfayaz, H. Yousefi, “Numerical modeling and economic analysis of a ground source heat pump for supplying energy for a greenhouse in Alborz province, Iran”, Journal of Cleaner Production, Vol. 131, pp. 145-154, 2016 [27] D. A. Damers, M. Dorais, C. H. Wien, A. Gosselin, “Effects of supplemental light duration on greenhouse tomato (lycopersicon esculentum mill.) plants and fruit yields”, Scientia Horticulturae, Vol. 74, No. 4, pp. 295-306, 1998