This action might not be possible to undo. Are you sure you want to continue?
Renewable Energy 33 (2008) 383–387 www.elsevier.com/locate/renene
Validation of CFD simulation for ﬂat plate solar energy collector
Mohamed Selmi, Mohammed J. Al-KhawajaÃ, Abdulhamid Maraﬁa
Department of Mechanical Engineering, University of Qatar, P.O. Box 2713, Doha, State of Qatar, Qatar Received 30 December 2006; accepted 3 February 2007 Available online 26 April 2007
Abstract The problem of ﬂat plate solar energy collector with water ﬂow is simulated and analyzed using computational ﬂuid dynamics (CFD) software. The considered case includes the CFD modeling of solar irradiation and the modes of mixed convection and radiation heat transfer between tube surface, glass cover, side walls, and insulating base of the collector as well as the mixed convective heat transfer in the circulating water inside the tube and conduction between the base and tube material. The collector performance, after obtaining 3-D temperature distribution over the volume of the body of the collector, was studied with and without circulating water ﬂow. An experimental model was built and experiments were performed to validate the CFD model. The outlet temperature of water is compared with experimental results and there is a good agreement. r 2007 Elsevier Ltd. All rights reserved.
Keywords: Solar energy; Solar radiation; Solar collector; CFD simulation
1. Introduction The solar energy collection as a renewable energy topic has been the primary interests of many engineers and researchers for the last two centuries due to its wide applications such as domestic water heating systems. Today, solar water heating systems are being used for single family houses, apartment buildings, schools, car washes, hospitals, restaurants, agricultural farms and different industries. Solar water heating can reduce domestic water heating costs by as much as 70%. Owners of these buildings have found that solar water heating systems are cost-effective in meeting their hot water needs all over the year. A more intensive attention was given to this topic from 1970s of the last century, particularly, when the worldwide crisis of 1973 has taken place. Since then, the efﬁciency of solar heating systems and collectors has improved. The efﬁciencies can be attributed to the use of low iron, tempered glass for glazing (low-iron glass allows the transmission of more solar energy than conventional glass), improved insulation, and the development of durable selective coatings. Thus due to its importance, a
ÃCorresponding author. Tel.: +974 4852109; fax: +974 4670421.
ﬂat plate solar collector with and without cover glass was analyzed using computational ﬂuid dynamics (CFD) software and simulated without water ﬂow in Refs. [1,2]. Here, the water ﬂow is added and the problem is examined both computationally using CFD software as well as experimentally and the results are compared to each other for validation purposes. The CFD package is from Computational Fluid Dynamics Research Corporation (CFDRC) and has been successfully used to simulate water currents and heat transfer inside a water-intake lagoon along a coastal zone near a power plant in Doha, Qatar by Raouf and Selmi . The software proved to be powerful and ﬂexible in modeling a wide range of practical problems. More on the CFDRC package is found in Ref. . The CFD analysis of the ﬂow and heat transfer in ﬂat plate solar collectors is computationally quite difﬁcult and the number of research works on this subject is quite low. We refer the reader to the paper by Morrision et al.  for further details. 2. Experimental model A simple model for the solar collector was built (See Fig. 1), which is used for implementing the required experiment, that can be helpful in gaining real outputs that
E-mail address: firstname.lastname@example.org (M.J. Al-Khawaja). 0960-1481/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.renene.2007.02.003
and outside of the solar collector. W/m2 residuals temperature. Both plate and pipe are ﬁxed at a certain level inside the wooden box. The inner walls of the box. in order to connect the water inlet and outlet ﬁttings easily. i. and a calibrated pot. The project model is oriented to face the sun rays normally. (4) pipe temperature. Selmi et al. kg/m3 are compared to those simulated by the CFD package described later in this work. 2). / Renewable Energy 33 (2008) 383–387 Nomenclature h n p q R T Tin heat transfer coefﬁcient.ARTICLE IN PRESS 384 M. in accordance to the proposed geometries (see Fig. (2) water outlet temperature. (5) ambient temperature. To gain simulated results using CFD software. The combination of these geometries forms the module body. but without the status Fig. aluminum plate has a density of 2770 kg/m3 and conductivity of 177 W/1C m. where module volumes. 3. The box contains an aluminum plate used as an absorber. m/s v velocity component in y-direction. boundaries. The readings are automatically monitored and plotted. and (7) water ﬂow rate. 1C and K Tout. The water ﬂow is controlled by means of a manual control valve. while the outside box is painted with white color. ambient. (6) solar radiation intensity. The following measurements are required: (1) water inlet temperature. and grids are created. while all other parameters are kept the same. 1C and K water inlet temperature. The procedure starts by drawing the module ﬁrst. The project model consists of a wooden box of 1. Model geometry. 1. There are two types of experiments which are implemented by the built up project model: The ﬁrst experiment is performed by still water. In addition. The procedure requires setting the boundary and volume conditions of the simulated module. W/m2 1C number of iterations pressure.e. exp experimental water outlet temperature. m/s Greek letters r ﬂuid density. through the thermocouples. CFD simulation The CFD simulations are done by the CFD package from the CFDRC. beside others for other materials are given in Table 1. a simulation procedure has to be followed. Due to the high insulation efﬁciency and easy forming. m/s w velocity component in z-direction. ﬁxed to it from the top a copper pipe of 1-in diameter. The pipe extends slightly on both sides outside the housing frame. the housing frame was made of wood.5 m long. A number of thermocouples are attached to the absorber plate. and tilted from time to time to keep tracing the sun position for most of the experiment time. CFD-Geom. The model pipe is connected to water source through manual control valve at one end. and a polyurethane insulation from the 2 bottom. absorber plate. and the second is performed by providing a controlled ﬂow of water. Assuming that the collector is a simple ﬂat plate solar collector. . and a drain hose at the other. This stage is done with the CFDRC geometric modeling module. and pipe are covered by black mate paint. to measure the collector inside. upon complete set up of all components and steady-state conditions are achieved. no water ﬂow. 1C and K u velocity component in x-direction. by means of a set of measuring devices connected to the absorber plate and pipe. since it is painted with black mate paint. boundaries should have both convection and radiation heat transfer mechanisms. Aluminum plate is set to have both emissivity and absorptivity of 1. N/m2 solar heat ﬂux. sim simulation water outlet temperature. Measurements starts. All these properties. interfaces. it has a density of 8800 kg/m3 and conductivity of 401 W/1C m. covered with a transparent glass. except for the face and bottom surfaces. thus the maximum energy can be gathered. pipe at some selected points. water inlet and outlet temperatures. (3) absorber plate temperature. The copper pipe has the same radiation characteristics to be an ideal emissive and absorptive surface. 166 mm wide and 70 mm high. 1C and K Tout.
039 0 — 385 Fig. The air block can be selected and its properties are set up or entered into the solving module. they have both convection and radiation heat transfer with the ambient temperature equals to the corresponding experimental one. for the inlet and outlet . The next stage is converting the bone geometries to a graphical user-interface module which can be used in the CFDRC solving module. 3. this can be done by selecting the boundary or volume which will cause a setting window to appear on the screen that has the setting cells of each property or condition. and properties.569 — — Al plate 2770 875 177 0 1 Cu pipe 8800 420 401 0 1 Insulation block 72 1 0. All boundaries except the top glass. the air speciﬁc heat and thermal conductivity are set to be 1007 J/kg 1C and 0. and radiation. The next step in the simulation modeling process is to set up the boundaries.ARTICLE IN PRESS M. The cover glass is also exchanging energy by convection with the outside environment. deﬁnition of boundaries and volumes. Selmi et al. A complete simulation process is performed by using the experimental results of 7 December 2004 (see Fig. 3 for total solar radiation). the conditions are considered to be without heat transfer and radiation effects. and bottom of the model have an emissivity of unity as one of the radiation options. Furthermore. respectively. / Renewable Energy 33 (2008) 383–387 Table 1 Volume condition setting for properties of blocks Air block Density (kg/m3) Speciﬁc heat (J/kg 1C) Thermal conductivity (W/m 1C) Transmissivity Emissivity See Eq. in Doha city. ð2Þ For the module side walls and bottom walls. Drawn module in CFD-Geom. This simulation is to be modeled in a manner that it will have the same initial conditions and parameters as the experiments. 2. volume conditions. a varied density which is a function of temperature is used for air as rair ¼ 8:147 À 0:0682 þ 2:688 Â 10À4 T 2 À 5:387 Â 10À7 T 3 þ 5:299 Â 10À10 T 4 À 2:0282 Â 10À13 T 5 ð1Þ and a transmissivity of unity is set to be the block radiation property as the air is assumed to be transparent to radiation coming from the sun rays and hitting the cover plate or bouncing off the cover plate. The same was done for the aluminum plate. CFD-ACE. to end up with a comparison chart between the experimental and simulated results. Finally. so the emissivity of the boundaries is set to be zero and these walls are considered adiabatic. face. all these properties are shown in Table 1.0263 W/m 1C. The volume blocks can also be set up by selecting the intended block in the model. The simulation modeling can be described starting from selecting the problem type as ﬂow. and exchanging radiative energy with the environment or the neighboring walls enclosing the air block. and water blocks. In addition. but with different properties and conditions. Also. Again. Measured total solar radiation on 7 December 2004. (1) 1007 0. (2) 4230 0. The water density is estimated by the function rwater ¼ 197:105 þ 7:14299T À 0:0200628T 2 þ 1:711949787 Â 10À5 T 3 . copper pipe. insulation.0263 1 — Water block See Eq. the model dimensions are taken to be exactly the same as those of experimental model. heat transfer. Fig.
the properties of collector body material. 6). 4). Temperature distribution simulation (in K) of the collector inlet cross section for 1. at a certain time (11:45 AM). Residuals. n.36 g/s mass ﬂow rate at 9:55 AM of 7 December 2004. heat transfer. Fig. e. which represent the atmosphere properties. the residual magnitudes become small.ARTICLE IN PRESS 386 M. and it is less than that for zero-ﬂow rate Fig. and above properties into the solving module CFD-ACE software. except that it has a ﬂow rate of water equal to 1. Selmi et al. . The module initial conditions were applied similarly as the actual experimental ones.. then for each case the results are analyzed and outlet water temperatures are plotted and compared to the experimental ones for zero-ﬂow rate and varied ﬂow rate. Temperature distribution simulation (in K) at the collector outlet cross section for 1. ﬂow rate.0170484 m/s. Temperature distribution (in K) for zero-ﬂow rate simulation at 11:45 AM of 7 December 2004. After introducing the experimental results. 4. Fig. CFD-View. From one iteration to the next. the results in terms of temperature.g. 6. boundary conditions. pressure. 5. / Renewable Energy 33 (2008) 383–387 walls. R. The residuals are shown as curves of residual magnitude versus number of iterations (see Fig. Fig. As it is shown in Fig.36 g/s mass ﬂow rate at 11:45 AM of 7 December 2004. The ﬁrst module simulation was for the collector with zero-ﬂow rate. The next module simulation is for a random but reasonable ﬂow rate. pressure at the grid points change and the difference of the values between the current solution and previous one are called residuals. which is 345 K (see Fig. and radiation iteratively starting from zero solution or previous one after inserting the experimental parameters in it with the aid of a graphical user interface. versus number of iterations. heat transfer rates and so on can be visualized or calculated by the CFDRC viewing module. velocity.361573 g/s. the average speed of water inside the tube has been set to a magnitude of 0. and the heated water temperature inside the heat-carrying pipe is almost 361 K. The CFDRC solution module CFD-ACE solves the equations of ﬂuid ﬂow. This simulation has the same initial conditions of the zero-ﬂow rate module. the simulated results such as temperature. and the ﬂuids involved in the process. the maximum registered temperature was for the absorber plate. 7. 370 K. The maximum temperature is registered for the absorber plate. and achieving the conversion of the steady-state solutions. 5. When the results are converged. The parameters are managed through the boundary and volume conditions.
a comparison between the experimental and simulated results have to be done. Groenhout NK. while keeping other parameters ﬁxed. December 5–8. 7 at 9:55 AM. an inlet temperature distribution simulation of the collector cross section is shown in Fig. Maraﬁ A. To accomplish that. Huntsville. These disturbances or inefﬁciencies could be associated with the experimental measurements. Optimal design of advanced solar water heaters. the simulated temperature curve has the same behavior as that experimental one and they are close to each other. it shows that the experiment has numerous disturbances. The water inlet temperature is reported to be 309. Although there are some small discrepancies due to some experimental imperfectness matters. As shown in Fig. Selmi M. such as geometries. CFD simulation of ﬂuid ﬂow and heat transfer in a ﬂat plate solar collector. Also inlet temperature is shown. Fig. we still have a good conﬁdence in the CFD simulation program that can be used in the future for more complex solar collector problem. curves are plotted to indicate the inlet and experimental and References  Selmi M. Behnia M. insulation. Boulder. 2000. 2004. 1999. In: Proceedings of IMEC2004 international mechanical engineering conference. August 28–September 3. CO. material. 4. Kuwait. The registered water temperature at the outlet is 327. which affect the behavior of the experimental temperature curves. Cook M.4. USA: CFD Research Corporation. It seems the temperature difference between the water inlet and outlet is almost 9 1C over all times. As an interesting result. 5.  Morrision GL. In addition.6 K which is much lower than the absorber plate temperature. A comparison chart is plotted to indicate the experimental and simulated temperatures versus time. measuring aids. Selmi et al.  Selmi M.38 K (see Fig. Comparison of results In order to draw up a conclusion. and recommend a decision in regards of the project model. In: Proceedings of the 2004 world renewable energy congress. Performance prediction of a ﬂat plate solar collector using CFD software. Al-Khawaja M. Comparison between experimental and simulated outlet water temperatures. which is less than the water temperature in the previous simulation. the solar radiation ﬂux for the simulation module is varied. Mills DR.  Raouf MIN. / Renewable Energy 33 (2008) 383–387 387 simulated outlet temperatures versus time. Conclusion There is a good agreement between the experimental and simulated results for outlet water temperatures. 2004. to match the experimental readings. Al-Khawaja M. Maraﬁ A. module.6:175–93. as the collector is a hand made model and it could have obvious imperfectness issues. such as leakage. Also. 8.  CFD-ACE(U) users manual. etc. Murdoch University. optimize a performance. 350 K. 8. In: Proceedings of the sixth Australasian natural convection workshop. It was found that the temperature is lower in the case for water ﬂow than that one without ﬂow due to the forced convection heat transfer. J Mar Environ Eng 2001. This assures that our built-up solar collector is a fairly effective model. . 6). version 6. Modeling of temperature variations and water currents for coastal zones near power generations and water desalination plants.ARTICLE IN PRESS M. AL. This is due to the effect of the heat transfer while water is ﬂowing inside the pipe.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.