Analysis of Solar Photovoltaic Panel Integrated with Ground Heat Exchanger for Thermal Management

Photovoltaic panels (PV) constitute the most commonly adopted solar-to-electricity conver- sion technique worldwide to alleviate the world electric energy demand. However, climate in the deserts is representing a challenging barrier for solar PV to be adopted. Heat and dust both have a negative impact on the performance, durability and electrical efficiency of the PV. The problem is associated with high temperature during the day and condensation formation after midnight. The cause of the problem and the resulting consequences are summarized in the chart shown in Figure 1.

Commercial solar cells convert solar irradiation into electric power at a relatively low effi- ciency, around 20%. PV panel efficiency decreases due to high temperature by 0.4%/K, as noticed by Dorobantu et al. [1]. Because of this, researchers proposed many ideas to cool the PV panels during the day. PV integrated with a cooling method is called PV/T, referring to photovoltaic thermal combination.

Many PV cooling methods have been proposed and investigated. PV/T solar systems are beneficial devices, which enable production of electricity and heat simultaneously. Continu- ous water film flowing over the frontal surface of the PV module has been studied and analysed by many researchers. Dorobanţu et al. [1] presented their work on frontal water film cooling utilizing PV power plant installed at the University Politehnica of Bucharest. They used water film at 24°C flowing at 2 lit/min, which reduced the temperature on the back of the PV from 48°C to 35.5°C. The temperature difference between back-side and front-side of the panel remains the same at about 7–8°C. Experimental studies have been performed to compare the performance of a PV system combined with a cooling system consisting of a thin film of water running on the top surface of the panel. These studies contain the effects of the nominal power of array and system head on the operation of PV system by cooling the PV cells with a thin film of water (Krauter [2], Moharram et al. [3]).

Another proposed and widely investigated PV/T cooling method is by rear surface cooling. Bahaidarah et al. [4] evaluated the performance of a PV/T module by backside cooling and reported yielding a temperature reduction of 20%, which resulted in an efficiency increase of the PV panel by 9%. Hongbing et al. [5] investigated active cooling by attaching heat pipes beneath the PV panel, building up a PV-heat pipe (HP) combined system. They concluded that thermal efficiency decreased in the daytime; a turning point occurred in the afternoon and there was a minimum electrical efficiency in the daytime. Jakhar et al. [6] studied the rear side cooling by heat exchange with earth water and Chiasson et al. [7] by heat exchange with shallow water pond.

Another cooling method is by water spray on the frontal surface of the PV. A study by Odeh et al. [8] achieved cooling of the PV panel by water dripping on the upper surface of the PV panel and obtained an increase of about 15% in the system output at peak radiation conditions. A different investigation of front PV surface cooling was performed by Hosseini et al. [9], using a spray water cooling technique. Results showed that the cell power increased due to the spraying of water.

While the cooling of the PV module is needed only during day time, heating is needed in the night time to avoid condensation of air humidity on PV panels. This prevents the formation of mud in the presence of dusty winds, which increases maintenance costs and shortens the lifespan of the PV panels [10]. While Hegazy [11] investigated the effect of dust accumulation on flat plate solar collector, Kordzadeh [12] investigated the effects of nominal power of array and system head on the water pumping by PV set with the surface covered by a film of water and Al-Kayiem [13] investigated the condensation formation effect on the performance of the solar collector of solar chimney power plants. Hegazy [11] found that mud layers on the pan- els’ surface reduce the transmissivity of the glass layer and thus also PV efficiency and Al-Kayiem et al. [13] found that the condensation alter the solar chimney functioning as the early hours after sunrise are consumed for evaporation of the condensate instead of heating the absorber. However, condensation film is encouraging mud formation and frequent cleaning is leading to scratching the surface, which in the medium term also leads to a drastic decline in the efficiency of solar systems, like PV, flat palate solar collectors and solar chimneys.

Using the earth as a heat source or thermal buffer in a heating or cooling system is an attractive application from a thermodynamic point of view. Underground water cooling sys- tems are potentially more efficient than conventional cooling water-to-air systems. Ground horizontal cooling loops are preferred in many situations due to a consistent temperature since the ground temperature is not different within a specific depth where the surface heat flux does not play any role ([14], [15]).

The ground heat exchanger (GHEX) is a concept of heat exchange between soil and circu- lating thermal medium, usually fluid flowing in pipe. At a certain depth in the ground, the soil temperature is almost constant throughout the day, unless there is harsh weather change. In tropical areas, the soil temperature below a certain depth is almost constant over the year (Bansal et al. [15]). Chel et al. [17] suggested that the soil can be used as a thermal sink for the cooling of PV panels. Various studies adopted the GHEX concept as earth-water heat exchangers and earth-air heat exchangers for air conditioning purpose, where water and air are the cooling media, respectively [14, 16, 18, 19]. The GHEX pipes are buried in the ground at a particular depth, and the inlet of the tubes carries the hot water, which transfers the ther- mal energy from the high-temperature fluid to the soil, thus resulting in a decrease in outlet temperature [10–18].

In conclusion, most of the investigations and methods used in the PV/T approach are deal- ing with cooling of the PV during the day. No work is reported on mitigating the condensation during the night. The present article presents a CFD numerical procedure and the experimen- tal validation approach achieved by site measurements. The article presents the results of the new novel method to mitigate both, high temperature during the day and condense formation during the night by integration of the PV/T panel with GHEX. The numerical method is developed by modeling of the experimental PV/T- GHEX in ANSYS FLUENT. Presented results are focusing on the validation of the CFD method, analysis of the GHEX and PV/T thermal conditions at various weather conditions.

A developed experimental setup was designed and implemented under humid and warm cli- mate conditions at the Universiti Teknologi PETRONAS (UTP) - Malaysia. The location of the site is at a latitude of 4.39°N and longitude of 100.9°E.

The experimental setup is shown schematically and physically in Figure 2. It comprises five primary components: solar PV panel, modified panel to a PV/T by installation of water pocket in the back, water circulation system with a pump, ground heat exchanger and set of measuring instrumentation system. Water as cooling/heating medium is circulating between the backside water pocket and the GHEX by pump running at controlled discharge for prese- lected flow rate. The PV and PV/T are subjected to the same weather conditions and are subjected to measurements and observations, simultaneously.

The GHEX is made of copper tube coil embedded in the ground. The tube has inner diam- eter of 0.015 m, thickness of 0.003 m and a total length of 22.0 m. The heat exchanger is installed at 0.8 m below the ground surface. This depth has been decided based on tempera- ture distribution measurement of the ground in the experiment site. The measurements have been repeated for five days and the mean temperature values are considered for results anal- ysis. At 0.8-m-depth and below, the temperature is almost constant.

During the day, warm water flowing through the PV backside has been circulated through the GHEX, thus transferring heat to the soil and cooling the water. This process increases the soil temperature and store heat to the ground. However, during the night, when water flows through the ground, the storage heat increases the water temperature and heat the PV panel when circulating through the backside panel. The heat storage capacity of the ground is dependent on the thermal properties of the soil. Thermal conductivity of dry soil is measured and found to be about 0.2 W/m2·K.

The numerical analysis of the present work included a horizontal GHEX embedded under- ground. The simulation considered a block of the ground containing the GHEX, as shown in Figure 3. The selected ground block has 2-m-width × 3-m-length × 2-m-height. Those ground domain dimensions are large enough compared to the GHEX geometries for assuming adiabatic boundaries. The heat exchanger has similar geometries as in the exper- imental setup. The simulation has been conducted at three different depths, Z, of the GHEX installation: 0.5, 0.8 and 1.5 m. The computational model was created to simulate the experimental case and impose extended boundary conditions to gain more insight on the effectiveness of the proposed idea of integrating the PV/T with underground embedded coil, GHEX.

The simulation was achieved by solving the conservation equations of mass, momentum and energy under the assumption of steady, compressible and viscid pipe flow. The equations of conservation of mass, momentum and energy are:

Continuity

Momentum

Energy

where $\vec{V}$ is the velocity vector, u is the specific internal energy, q is the heat flux (= - k ÑT ), p is pressure, τ is the shear tensor, $\dot{S}$ is a heat source (from/to ground) and $\tau: \nabla V$ is the change in internal energy due to viscous dissipation.

The copper pipe loop and soil domain mesh is generated using the ANSYS meshing tool. In both domains, a hex-dominant method for reducing the number of cells is used. The cells in the copper pipe loops domain were kept uniform with a cell size of 0.2 cm, which gave enough accuracy of the pipe temperature distribution. The internal water flow domain uses cell inflation on the interior of the pipe wall to resolve the near wall boundary layers within the fluid. Cell inflation meshing is recommended for modelling wall-bounded flow effects due to the turbulent boundary layer low gradient near the internal wall. It ensures that the grid cell close to the wall has a height smaller than the laminar sublayer, not in the algorithm region of the turbulent boundary layer (Lim et al. [20]). The discretization of these domains provides additional information about heat transfer and fluid flow characteristics within the system. In order to save computation time and since the flow characteristics are relatively simple, a high resolution of these features was not considered. The mesh was concentrated inside the soil pile. The size of the individual mesh elements varied greatly because of the large variation in domain sizes. The size variation of the domains provides challenges in keeping the cell count of the mesh low. The model residual convergence criteria were set to 1×10-3. The CFD model mesh with the soil domain, the top and bottom of the soil pile and the copper pipe loops is displayed in Figure 3.

The simulation is performed numerically using control volume-based solver in the commer- cial software, ANSYS FLUENT. Reynolds-Averaged Navier–Stokes equations have been solved with a semi-implicit method for pressure-linked equations utilizing SIMPLE algo- rithm. The realizable k-ε turbulence model with enhanced wall treatment and thermal effect option has been adopted. The k-ε model is appropriate for wall-bounded and internal flows with small pressure difference across the flow passage (Lim et al. [20]). The k-ε model satis- fies certain mathematical constraints on the Reynolds stresses and is consistent with the physics of turbulent flows. Robustness, computational economy and reasonable accuracy for a wide range of turbulent flows justify the k-ε model popularity in engineering applications and heat transfer with fluid flow simulations (Ruzicka [21]).

The main parameters of the computational procedure are listed in Table 1. The numerical simulation has been carried out under the assumption of steady, incompressible and viscus flow conditions. The ambient temperature, solar irradiance and inlet water temperature are taken from the experimental data and input to the simulation. The computational domain is discretised adopting 3D-structured meshing criteria. The heat exchanger was simulated for different operat- ing conditions following the experimental measurements conditions at various times of the day.

For the simulation, the weather data acquired at the solar research site in UTP – Malaysia have been used. These considered the interactions of the soil with solar irradiation at the ground surface. Also, wind speed and ambient air temperature during the measurement time have been utilized as input parameters to the simulation. The applied boundary conditions used in the simulations are given in Table 2.

The values of the main thermodynamic properties of the materials included in the simula- tion are shown in Table 3.

The numerical simulation has been carried out at controlled water flow rate of 1.6 LPM. The experimental conditions during the measurement are essential to be identified prior to com- mencing the discussion of the results.

The experimental measurements were carried out first on the ground thermal conditions and then on the PV and PV/T-GHEX system. The measurements have been repeated for many days and an average over 5 days is selected for presentation.

The ground temperature at various depths is essential for making a correct decision on the depth at which the tube coil should be buried. Figure 4 shows the transient measured ground temperature at 0.5 and 0.8 m depth. The temperature results are almost identical in the early morning hours. Divergence begins after 11.30 AM in the morning until it reaches its maxi- mum value at around 4.30 PM. Two conclusions could be drawn from those measurements. First, it is clear that the soil temperature is almost constant at 0.8 m depth with slight increase in the afternoon, as the solar radiation with high irradiance and high incidence. Second, the temperature of ground at 0.5 m depth is not suitable for the installation of the GHEX as it is relatively high and the heat removal from the hot water will be inefficient.

Figure 5 shows the measured temperatures on the surfaces of the PV, PV/T panels and the solar irradiation in October month.

The thermal control system managed to reduce the PV/T panel temperature by about 8 K over the hot and high solar irradiance period during the day. This temperature reduction is enough to keep the PV panel temperature near the normal operating condition of about 40–45°C. During the condensation time (3:30 AM to 7:30 AM), the thermal control system managed to keep the PV surface temperature above the dew temperature by about 2–4°C. By this thermal control system, the PV efficiency can be increased by 10% and nocturnal con- densation can be avoided.

The numerical procedure developed in this research is aimed to simulate the operational per- formance of GHEX that is hypothesised to enhance the performance of a PV/T system.

Prior to presenting and discussing the numerical results, the computational procedure is val- idated. The validation has been achieved through comparison between the numerical prediction and experimental measurements. As a numerical analysis targeting ground thermal simulation and GHEX simulation, the validation procedure is based on both. Figure 6 repre- sents validation by comparing the experimental measurements and simulation prediction of dry soil ground temperatures at 0.8 m. The goodness of the simulation is fairly clear in terms of the trend and the deviation.

However, the simulation is over-predicting the ground temperature during the daytime and under-predicting the ground temperature after the midnight. The maximum relative error is around 1.9%, which indicates that the simulation is valid to predict the ground temperature. Comparison between various cases of measured and simulated ground temperatures is presented in Table 4. The results in the table demonstrate that the soil conditions, wet or dry, highly influence the temperature profile in the ground. Both, experimental and numerical results show that the dry soil temperature is higher than the wet soil temperature, under the same weather conditions.

Experimental water temperature at outlet of the GHEX as measured experimentally and that predicted numerically by the simulation are compared by referring to Figure 7. The short-term behaviour of the fluid outlet temperature was compared as obtained experimen- tally and numerically during every 1 h at a depth of 0.8 m. The maximum and mean differences between the simulation and experimental results are about 3.2% and 2.2%, respectively. So, it is reasonable to say that the outlet fluid temperature shows reasonable agreement with the experimental data. The numerical model and experimental data matched well in this time- frame, whereas a maximum deviation of 0.6 K was observed at 5 PM.

Figure 8 shows the temperature profiles in the soil pile at two different buried GHEX depths, 0.5 m and 1.0 m, at 12 PM, October 11. The temperature profiles show that under realistic weather conditions, the soil temperature varies minimally at 1.0-m-depth, whereas tempera- ture is almost constant below a depth of around 1.5 m. This is the ideal location for the ground heat exchanger as it was not affected much by the ambient weather conditions. At 1.5-m-depth and below, it is possible to obtain the maximum temperature between circulatory water inside the GHEX and the surroundings.

Figure 9 displays the heat dissipation in the buried GHEX at 0.8-m-depth, at 12:00 PM. Hot water flows in the GHEX at around 313 K and leaves at around 308 K. The GHEX managed to reduce the water temperature by around 5 K. This demonstrates the correctness of the hypothesis and the proposed technique for reducing the PV surface temperature.

Figure 10 shows the water outlet temperature from the GHEX at various times when the GHEX is located at depths of 0.8 m, 0.5 m and 1.5 m in dry soil. With larger depths of the copper pipe loop, the overall heat transfer area towards the colder soil is also growing. Besides, the water temperature was less affected by the thermal condition of the soil surface for a larger burial depth of the pipe. Therefore, the water outlet temperature decreases with increasing depth.

The PV-GHEX system has been run for different cases; in every case, one of the weather data is changed and other data are fixed.

Figure 11 shows the effect of wind velocity on the PV/T-GHEX panel surface temperature. The results show that when the wind velocity changed from 2 to 12 m/s, PV glass temperature changed from 304.6 K to 301.7 K during the night and from 292.6 K to 292.2 K during the day. For every 4 m/s rise in wind velocity, the PV glass temperature is reduced by 1.2 K during daytime. For every 4 m/s rise in wind velocity, the glass temperature is reduced by 0.3 K during the night-time. This is because the air flow will increase the convective heat transfer loses, which will decrease the glass surface temperature.

Figure 12 presents the effect of ambient air temperature, $T_{\text {air }}$ on the system surface tempera- ture. For air temperature changing from 8 °C to 48 °C, the night glass surface temperature, $T_{\text {glass }}$, changes from 288.7 K to 304.8 K and the day glass temperature changes from 301.4 K to 317.6 K. For every 10 K rise in $T_{\text {air }}$, $T_{\text {glass }}$ increases by 4 K. This increment in the PV glass surface temperature is due to the glass exchanging heat with the surroundings through con- vection as well as radiating heat from and to the atmosphere through radiation.

Figure 13 shows the PV-GHEX system surface temperature for one day. The dew point tem- perature is changed from 1°C to 15°C and other weather data were fixed. The PV/T glass temperature changes from 296.3 K to 295 K during the condensation time over the 1°C to 15o C increase in the dew point temperature. While the glass temperature is reduced by 2.5 K during the daytime. For every 5 K rise in dew temperature, the glass temperature increases by 0.4 K and 2.5 K during the night time and daytime, respectively.

It is demonstrated that installation of water pocket at the back of the PV and connected to ground heat exchanger would enhance the PV/T performance. Experimental and numerical assessment proved the functionality of the proposed method to mitigate the PV problems of high temperature in the day and condensation formation in the night. Results show that the GHEX cools the water in the daytime by around 8 K, leading to improved PV daily efficiency by about 9.0%. Experimental measurements revealed that the GHEX has contributed to warming up the PV/T in the night, resulting in elimination of the moisture condensation on the PV surface. On other hand, ground temperature profile measurement and simulation show that the wet soil has lower temperature than the dry soil.