AERODYNAMIC MECHANISMS ON A GROUND-MOUNTED SOLAR PANEL AT DIFFERENT WIND DIRECTIONS

: Aerodynamics mechanisms on a solar panel were studied using Computational Fluid Dynamics methodology at different wind directions. The wind velocity was chosen as 10 m/s that corresponding turbulent flow. The inclination angle of the panel was fixed as 25°, while the wind directions were varied as 180°, 135°, 45°, and 0°. Governing equations were solved by utilizing a finite volume method with a realizable k-ε turbulence model and standard wall functions. The results showed that the recirculation area occurred for the straight wind directions, but it was not observed for the oblique wind directions. The highest pressure coefficients occurred at the leading edges of the solar panel and they reduced to the trailing edge for all wind directions. The maximum drag and uplift coefficient was obtained at the wind direction of 0 0 and 180 0 , respectively. dolaşım bölgesinin düz rüzgar yönlerinde meydana geldiğini, ancak eğik rüzgar yönleri için gözlemlenmediğini göstermiştir. En yüksek basınç katsayıları panelin ön kenarlarında meydana gelmiş ve daha sonra tüm rüzgar yönleri için arka kenara doğru azalmıştır. 0 0 ve 180 0 rüzgar yönleri, maksimum


INTRODUCTION
Since energy demands are growing day by day, alternative energy research increases along with that. Solar energy is a reliable source, but it is a challenge to make cost-efficient and effective solar panels in comparison to traditional energy resources. Therefore, there have been many investigations on optimization studies of solar panels recently, such as Aly and Bitsuamlak (2013) carried out wind tunnel study for geometric scale effect on solar panels mounted on a low-rise building. They showed that the size of the model was not changed with the mean pressure loads. Kopp et al. (2012) examined the solar panel arrays that mounted on the roof and ground with the different inclination angles of the panel in a wind tunnel. They found that turbulence generated by the panels and pressure equalization were two main mechanisms causing the aerodynamic loads. Bitsuamlak et al. (2010) performed a study to investigate the wind angle effects on ground-mounted solar panels. The numerical results showed that the pressure coefficients were similar patterns, while the magnitudes of the pressure coefficients were underestimated to compare to full-scale measurements. Jubayer and Hangan (2016) performed simulations to determine the wind loads for different wind directions on the solar panels. They reported that the highest values of the C D and C L were determined on the first row for all four wind directions, while the maximum overturning moments occurred for the 45° and 135° wind directions. Shademan et al. (2014) performed CFD simulations to investigate the unsteady wind loads of the panels with the effects of ground clearance. They found that the bigger clearance caused higher wind loads and shedding frequencies, and stronger vortex shedding. CFD analyses were performed to analyze the wind loads of solar panel arrays at different wind angles of attack and panel inclination angles by Shaderman and Hangan (2010), who found that the simulations identified the corner panels as the critical ones, and the entire structure A numerical investigation was performed for wind loads on a solar panel at the various wind directions of 180°, 135°, 45°, and 0°, in this study.

METHOD AND APPROACH
In this section, numerical modeling setup, the panel geometry, boundary conditions, and computational domain dimensions were explained. The dimension of the panel with a 25° inclination angle was in length of 1 m, in width of 0.7 m, and in thickness of 0.025 m. The back support height of the panel was 1.3 m, and the height of the front supports was dimensioned according to the angle of panel inclination. The panel geometry and computational domain were modeled using the SOLID-WORKS design software package. The dimensions of the computational domain were 21H in length, 11H in width, and 5H in height, and the distance of the inlet area to supports was 7H, as given in Figure 1. The incompressible and unsteady continuity and momentum equations can be written as follows: The term of Reynolds stress ̅̅̅̅̅̅ is described by the following equations: where represents turbulent viscosity. The simulations were carried out with the realizable k-ε model. k is turbulence kinetic energy, and ε is dissipation rate. The three-dimensional realizable k-ε model is widely applied in CFD, especially to analyze complex flows with large strain rates such as recirculation, rotation, and separation. Eq. (4) and (5) defined the transport equations as, where, [ ] √ ( ) P k and P b is the turbulence kinetic energy generation because of the velocity gradients and buoyancy, respectively, which are both estimated as the standard k- model. The rate of the turbulence fluctuating to the entire dissipation rate is defined as . The values of and are 1.44, 1.9, 1.0, and 1.2, respectively. A standard wall function is employed with near-wall treatment.
The mesh was consisting of 1.3 million cells generated with ANSYS meshing, as shown in Figure 2. In the inlet domain, 10 m/s wind speed of and 5% turbulence intensity was chosen. The bottom surface of the domain, support structures, and panel was modeled as a no-slip rough wall, while the sides and top of the domain were considered as no-slip smooth walls. Zero gradients of pressure were applied at the outlet of the domain.

a.
b. Figure 2: The mesh of the computational domain; a. Isometric view b. Side view

RESULTS AND DISCUSSION
The simulations were conducted on a ground-mounted panel for 10 m/s wind speed at 25° panel inclination angle and four various wind directions. The flow fields of the panel were shown in Figure 3 a-d for the wind directions of 180°, 135°, 45°, and 0°, respectively. The leading edges of the panel for all wind directions were exposed to the relatively low speed of the wind as it's the stagnation point of the panel. The velocity around the leading edge was between 5-6 m/s and 8-9 m/s in straight (180° and 0°) and oblique (135° and 45°) wind directions, respectively. A recirculation area occurred at the straight wind directions, but it was not observed for the 135° and 45° wind directions. The recirculation area was obtained on the lower surface of the panel for the wind direction of 0°, while it occurred at the top surface of the panel for the wind direction of 180°. Only one vortex was observed for the straight wind directions. The flow accelerated as it moved through the trailing edge of the panel for straight wind directions. However, the effect of acceleration was observed on lower and upper surface for 180° and 0° wind direction, respectively. Even though velocity distributions were well proportioned in oblique wind directions, maximum velocities could be seen right after leading edges. For 45° wind direction, it occurred beneath the lower surface of the panel, while it was above the top surface of the panel for the wind direction of 135°. So, the flow behavior for straight and oblique wind directions was similar due to the symmetric distribution. The pressure coefficient distributions on the panel for the wind directions of 180° and 0° are given in Figures 4 and 5, respectively. It was observed that as flow accelerated, Cp values were tending to decrease at both wind directions. The highest pressure coefficient occurred as 0.98 at the leading edge of the upper surface for the wind direction of 0°, where the flow separation occurred. The Cp values decreased monotonically from the leading to trailing edge. The lower surfaces exposed to the wind at 180 0 wind direction, as opposed to 0 0 wind directions. For the wind direction of 180°, the highest pressure coefficient determined as 0.94 at the leading edge of the lower surface. The maximum Cp in absolute values for straight winds was found very close, as expected.

a.
b. c. Figure 4: Pressure coefficient values for 180° wind direction on  a. upper surface, b. lower surface, and c. sides of the panel  a. b. c.

CONCLUSIONS
This study was carried out to specify the effects of the wind direction on a solar panel numerically. The wind velocity and panel inclination angle were chosen as 10 m/s and 25° respectively, while the wind directions were varied as 180°, 135°, 45°, and 0°.
It was found that the recirculation areas occurred only at the straight wind directions, which were seen on the top and bottom panel surface for the 180° and 0 0 wind directions, respectively. The maximum velocities were found to be where the minimum pressure coefficients occurred for all wind directions.
For all wind directions, the highest pressure coefficients obtained at the leading edge of the solar panel, and the values dropped monotonically to the trailing edge. The symmetric Cp distributions were obtained for the wind directions of 180° and 0°, while the distributions were asymmetrical for 135° and 45° wind directions.
The highest drag coefficient was obtained for 0° wind direction, while the smallest value found for 135° wind direction. The negative C L values were occurred for 45 0 and 0 0 wind directions, while the positive lift values were determined for the wind directions of 135 0 and 180 0 . The maximum negative and positive lift was obtained for 0 0 and 180 0 wind directions, respectively. The values of the pressure, lift, and drag coefficients were higher for straight wind directions than oblique wind directions.