RESEARCH NOTE COMPUTER SIMULATION OF THE PERFORMANCES OF SOLAR PONDS IN IRAN G.H. Toutounchi Department of Physic. Ferdowsi University Mashhad, Iran Abstract The Rabl-Nielsen formulas are used to make predictions for solar pond operations forfour locations situated in different solar insolation zones in Ian. The inputs to the computer program for {determining the temperature variation a the bottom of the-pond are the total insolation (corrected for cloudiness), ambient temperature, insolating layer thickness, convecting layer thickness and load. The ‘mean, the amplitude ofthe varying part, and the phase of the temperature atthe bottom isthe output of| the program. The computer simulations show that solar ponds may be promising as solar collectors for space heating in most parts of Iran. hese sith llie tet Sab OS olla Wee wtetiyt gle dgiges Slee OAS Ea 9s ess te ane Aa 5939 LI aslo ry gb ledges Sth ie a) al lame gl ol jee inn ebigs ESE Jae ald ge 9 gla agg sg abe Ls 9 lye ls ade ds pga ealipe gla ge abSLae LS Jy ciate a Ste ape aS hl Slender gl the Astin oJ paid ald ls Bgl ULEEL RLS cee tine iso Eat pepe Supa ys gig Mote INTRODUCTION Salt gradient solar ponds as a means of collecting and storing solar energy have been known for years (1) Normally a solar pond (Figure 1) consists of three water layers: a relatively thin homogeneus and convective fresh water surface layer; a non-convecting layer in the middle, ‘a meter or more thick, in which there are gradients in salinity (increasing downwards) and temperature (hot on the bottom, cold on the top); and a lower homogeneous- high salinity convecting layer, also a meter or so thick, which serves as a storage layer for solar radiation. Both directand diffused solar radiation penetrate throughout the pond. Even though very little radiation of solar spectrum wavelength longer than0.7 micrometer penetratesthrough as much as one meter of water, the short and medium wavelengths are absorbed throughout the pond and on the black bottom. The thermal radiation from the bottom in the Jong wavelengths is strongly absorbed in the lower cconvectine layer, where the heat i trapped since the upper layers non-convectingand water isa very poor conductor. Inthis study, the Rabl-Nielson (RN.) formulas (2] are Journal of Engineering, Islamic Republic of Iran ge E IS, gal ghia applied to the conditions in four locations with different insolation rates in tran to make predictions for solar ponds Siete convective rch war Noncomvecve yer CComectve yer Blam Figure 1, ayout a the solar pond Vol. 5, Nos. 384, November 1992 - 141‘TABLE 1. Mean Monthly Total Daily Solar Radiation Incident ‘upon a Horizontal Surface, Corrected for Cloudiness, (Wr - day), adopted from Ret (3. Giy_Rasht_ Mashhad Tehran Shiraz March 108164 188 219 April = 162-226 238 268 My = 195-276 282 301 June 204294 201 300 July 180285 276 286 August 140 27 27 264 Sebtember 105178 185 27 October 729 126 167 November 50 82 90 127 December 45 8 a 123 Januwy $7 95 107 147 February 76 122 97 189 operations, The computer program, developed for aN. two layer (upper non-convecting and lower convecting) solar pond, determines the temperature on the bottom of the pond as a function of time with solar, insolation, ambient temperature, convecting layer and non-convec ing layer thickness and load, as parameters. ‘TABLE 2, Location, Insolation and, Climate. Tocatfon, latte, —Tnsolavon Air emperaare and Cost HF TR & A Wine) CO)" rad) Rash,3732,0559 G18 79.01 184 94 0493 Matthad 3627.08 17980 11040 133 114 0453 Tehran, 3568,0886 18347 10888 168 127 0488 Shira, 2953,0869 21796 S005 175 108 0498 ‘THEORY Inhisrecent study ofthe solar energy for Iran, Samimit3} calculates the mean monthly total daily solar radiation, corrected for cloudiness, forannumber of cities. According, to his caleulations the total (direct plus diffused) solar energy flux, H(®), incident upon horizontal surface can be described by H@)=[5+0.15+03 (1-s)]1@)cos(®) (1 where 8,s and I(8) are the solar altitude angle, the ratio of ‘the number of sunny hours tothe total sunny hours avail ‘TABLE 3, Pond Parameters for an Average Year for Four Cities in Iran Location Insulating layer Convecing layer Heat load = Averagetemp = Tempamplitute Phase 1 ke s a a 8 100 (em) 100. (em) “0° (Win) B45 CC) 173 CC) 1:10 (rad) 100 125 128 300 96 138 100 2325 458 292 125 200 2d 143 300 162 133 100 100 0 1206 242, 10s 200 114 Lae Matha 300 BS 134 100) "38.96 608 Ba na 200 aud 130 300 240 140 100 100 “0 26.1 243 06 200 vs 12s 300 135 13s 100 16.69 65. 436 m2 200 as 13 400 243 tat 100) 100 0 Maa 202 1.06 : 200 146 Las Shear 300 U2 13s “4359 186 42 uu 2 n2 132 wx 241 142 142 - Vol. 5, Nos. 384, November 1992 Journal of Engineering, Islamic Republic of iranable, and the direct energy flux, respectively, In this equation itis simply assumed that the diffused solar fluxis| 10 percent for sunny days and 30 percent for cloudy days ofthe direct solar energy flux. The calculations of the mean ‘monthly total daily solarradiation isthenmade by empolying Equation (1) and summing over days in each month He 5 H(0)/ (Number of days in each month) (2) ‘Samimi's calculations of the mean monthly total daily insolation agree very well with the measured values for Tehran and are assumed to do s0 for other cities (3). His results for four locations: Rasht in low radiation zone (less than 350 cal/on?-day). Mashhad in intermediate radiation zone (350 to 390 calm? - day), Tehran in high radiation zone (390 to 430 cal/em?- day), and Shiraz in very high radiation zone (more than 430 cal/em® - day) are listed in Table 1. Rabl and Nielsen consider the absorption of radiation asitpasses through the water ofa several meters deep solar pond and derive equations for the resulting temperature range ofthe pond during a year round operation. They take into account the heat storage inthe convective layer in the bottom and in the ground underneath the pond and restrict their attention tothe steady state solutions ofthe equations, for along run performance. With this model, the incident slobal (solar) flux per unit area (insolation) H(t) and the surface temeperature T() (taking it to be equal to the ambient temperature) both have a constant and sinusoidal term and are described by H(= H+ Hicos wt @ and TW =Tt Tacos(wt-&) @ TEMPERATURE OF THE POND IN MASHHAD Figure 2. Temperature variation ofthe bottom of « pond in Mashhad with (A) and without (0) beat load together with the ambient temperature variation (Q). Journal of Engineering, Islamic Republic of Iran ‘where t=0 on 21 June and the common angular frequency w-corresponds t0 a period of 1 year. In these equations, H andi are the average insolation and the varying pat andT,,T,,and@, are the mean ambient temperature, the varytig “impute, ana ine phase 1ag win respect 10 insolation phase, respectively. Then, for a given rate of heat extraction from the bottom of the pond, the steady state temperature of convective layer is calculated from a radiation transition function to be also sinusoidal of the form T= T + Hicos (wt -8) oO) where T, T and 8 are the mean temperature, the varying amplitade, and the phase lag_with respect wo insolation phase. COMPUTER SIMULATION OF A SOLAR POND IN FOUR CITIES ‘The RNN. formulas in reference [2] are used to obtain the temperature variation at the bottom ofa solar pond located inthe four cities of Rasht, Mashhad, Tehran, and Shira.in ‘Iran, All notations are identical with the additional symbol S, defined to be the average power per square cm (load) being withdrawn from the pond, Similar simulations have been done for the southern part of Iran by Akbarzadeh and ‘Ahmadi (4 and for London by Bryant and Colbeck [5]. ‘The values of sotar insolation for each city in Table_| are fitted to Equation (3) to obtain the values of H , and H. ‘These values are listed in Table 2. The ambient air em- perature data supplied by the Tranian Weather Institute (1951-1976 inclusive) are fitted to Equation (4) and the determined values of T, T, and &are also listed in Table 2. The goal_of this simulation isto calculate the mean ‘temperature T, the amplitude T, and the phase lag 8 from TEMPERATURE OF THE POND IN SHIRAZ wa eset gE Figure 3. Comparison of the results of the present analysis (curve P) with the previous similar one(curve A-A)forasolar pond operating in Shiraz, Vol. §, Nos. 384, November 1992 - 143,Equation (5) for the aformentioned four cities. For the transmission coefficient of the radiation into the pond a conservative value of 0.83 is taken. The values of the effective cos (r= angle of reflection) are listed in Table 2. Itis assumed thatthe pond has a sufficiently large area so that edge effect can be neglected and itis dug into soit (the thermal conductivity k=0.0096 Wiem-c, the skin depth for seasonal variations 6, = (2k,/'w):, where k, = the ground diffusivity). The same physical parametersasinReference [2] are used, “RESULTS The results for insulating layer thickness of 1s= 100 cm, various choices of convecting layer thickness, 1c, and heat load, S, are summerizedin Table 3. In this table, in addition to $=0, other values of S are chosen as reasonable heat Toads to heat a 200 mr? house all year round. Basedona valueof 5.1310" //C-day (2700 BiwF-day) toheat ahouse with 200m*floor space, the average indoor temperature of 20°C, and the pond efficiency of 20 per- cent, the solar pond areas required to provide the loads are calculated to be Rasht: 117 n?, Mashhad: 107 n, Tehran: 52 ni, and Shiraz: 34 mé, corresponding to 1679, 2372, 1168, and 912 centigrade-degree-days per year, respec- tively. These estimates are valid for lange community ponds and obviously the performance of a finite pond would be different. More solar pond areas are required for individual house ponds in which the temperature at the bottom would be lower as the temperature near the outer edges is lower. The rato of the average temperature of a finite pond to the temperature of an infinite pond depends onthe radius in case of circular ponds, which could be as, small as 0.8 (2). As is evident froth Table 3, taking Mashhad as an example, the temperature at the bottom of solar pond with ‘one meter of non-convecting and one meter of convecting, layer thickness and an annual average load of 35.96 W/m is given by ‘T= 60.77 + 43.13 cos (2nt/12-1.12) © ‘The pond's temperature reaches a maximum of 104°C late in July and a minimum of 18°C late in January. ‘The temperature variation of the bottom of this pond versus time with and without load together with the ambi- 144- Vol. 5, Nos. 3&4, November 1992 ent temperature variation is plotted in Figure 2. It is ‘observed thatin the case of no load the temperature reaches a maximum of 145°C which exceeds the boiling point of ‘water at normal pressure. Since this disturbs the salt ‘gradient in the non-convecting layer, heat should be ex- tracted from the pond at an optimum rate to prevent boiling. A fixed percent (20%) of annual mean insolation removed, which i in fact the efficiency of the solar pond, keepsthe temperature of the pond reasonably below boiling point, as shown in the case with load. ‘Tocompare the results ofthe present simulation forthe ‘operation ofa solar pond in Shiraz withthe only previous similar simulation, the temperature variation of thebottom of the pond is plotted in Figure 3. Curve A-A shows the results of Reference (4) for a pond with one meter of non- cconvecting and zero convecting layer thickness and curve represents the analysis fora pond with one meter of non- ‘convecting, and one meter of convecting-layer thickness ‘operating inthe same area, Both curves are plotted for 20 percent of heat removal. It is observed that the two analyses do not compare properly. This may be due to the fact that making the convecting layer thicker reduces the seasonal variation and increases the phase lag without affecting the average temperature, as evident from Table 3. {As the presented computer simulation of the perfor- ‘mance of solar ponds located in different solar insolations shows, it appears that solar ponds are promising as solar collectors for residential heating in most parts of Iran. ACKNOWLEDGEMENTS ‘The author would like to thank his colleagues at the Physics Department of New Mexico University for their hospitality during the course of his sabbatical year. In particular he is grateful to professor Bryant who made valuable contributions to the success of this work. REFERENCES 1.H. Tabor and H. Weinberger. "The Solar Energy Handbook’, ‘chapter 10, McCraw Hil, (1981). 2.A.Rable and CE. Nielsen, Solar Energy 17, 11975). 3.1. Samimi, Iranian Journal of Physics 3, 21985), 4.4. Akbarzadch ndG. Ahmadi Solar Energy24,2, 43(1980) 5. H.C. Bryant and lan Colbeck, Solar Energy 19, 321(1977. Journal of Engineering, Islamic Republic of ran