Experimental Investigation of the Parapet Effect on the Wind Load of Roof-Mounted Solar Arrays

Abstract

A wind tunnel test was conducted to investigate the effects of parapets on the aerodynamic wind loads of roof-mounted solar arrays. The distribution of the mean wind pressure coefficient and the extreme wind pressure coefficient in the solar arrays were discussed in detail, and the results were compared with some national standards. Results show that the mean and extreme values of the wind pressure coefficient are larger at oblique wind angles. The presence of the parapet can reduce the mean and extreme area-averaged net pressure coefficients in some areas of the solar arrays, and both the most critical positive and negative shape factors reduce at the corner and outer rows while less change occurs at the inner zones. The experimental shape factors are much smaller compared to the values calculated by the standards. The guidelines for the design of wind pressure on roof-mounted solar arrays are very conservative, and more research is needed to make the standards applicable and complete.

1. Introduction

The rapid development of the photovoltaic industry will inevitably bring about some corresponding safety issues. Due to the light weight of the photovoltaic support and the solar arrays, the main controlled load for the calculation of the internal force of the solar arrays is the wind load [1]. Due to the high installation height of photovoltaic panels and the change in the wind field after the wind passes through the building, compared with large-scale solar arrays on flat ground, solar arrays installed on the roof are more affected by wind loads, and if an accident occurs, it will bring direct economic loss and is very likely to cause harm to the personal safety of nearby pedestrians.

Solar arrays are often installed on flat ground or on building roofs. There have been many studies on the wind load of solar panels on flat ground. For solar arrays, determining the wind load of a single solar panel is the basis for determining the wind load of the entire solar array. Velicu et al. [2] conducted a wind tunnel test study on the wind load of a single vertical solar panel, and the results show that the force coefficient is higher for low wind velocity, is decreasing and tending to stabilize for high wind velocity, and is relatively close and has higher values for high tilt angles. Similar wind tunnel tests were also conducted by Pfahl et al. [3], but they studied the wind load of solar panels with different aspect ratios, and a significant impact of the aspect ratio was measured. On this basis, some scholars have studied the wind load effect of photovoltaic arrays. Abiola-Ogedengbe et al. [4] conducted a wind tunnel test and found that for 0° and 180°, the pressure distribution on the surface of the solar module on the ground is symmetrical around the center line, but it is asymmetrical in other wind directions, and the average pressure values of the solar array are higher under downwind conditions than under open terrain winds. For some solar arrays, not only wind load but also wind and structure interactions need to be considered. Taylor and Browne [5] used two models, a pressure model at a relatively small scale and a sectional model at a larger scale, to propose a hybrid method and compared it to wind loads estimated using a dynamic amplification approach, and the results show that at high wind speeds, the self-excited forces become significant.

There is a significant difference in the wind load effect of solar panels located on the roof and on the ground due to the presence of buildings [6]. Kopp et al. [7,8,9] studied solar panels with different tilt angles through wind tunnel tests, and the results showed that the turbulence generated by larger tilt angles increases the net wind load, while for low tilt angles, the pressure balance dominates. When the tilt angle is less than 10°, the pressure coefficient has an approximately linear increase as the tilt angle increases, and for arrays with a tilt angle of 10° or greater, the wind load is relatively stable. Cao et al. [10] studied the wind load characteristics of roof arrays through wind tunnel tests and found that the negative pressure coefficient in the case of a single array is much larger than that in the case of multiple arrays. The tilt angle and the distance between arrays will increase the load pressure coefficient, but the influence of building height and parapet height on the negative pressure coefficient is not obvious. Shademan et al. [11] studied the influence of the lateral spacing between the sub-solar panels, the distance from the ground, and the wind direction on the wind load of the entire solar arrays through the three-dimensional Reynolds-averaged Navier–Stokes simulation (RANS), and found that wind load reaches the maximum value at the wind directions of 0° and 180°, and the results also show that the variation of the spacing between the panels changed the flow structure in the wake region. Warsido et al. [12] studied the effect of solar array spacing on the wind load coefficient through wind tunnel tests. The results showed that the wind load coefficient increased with the increase in distance between the rows of solar panels and decreased as the distance between the array and the roof skirt increased. Blocken et al. [13] and Stathopoulos et al. [14] studied the wind load of photovoltaic arrays on flat roofs of different heights and found that when the wind direction angle is 135°, the suction force of the rear photovoltaic panels is greater than that of the front photovoltaic panels, and the height of the building has a significant effect on the wind load. The influence of the total load of photovoltaic panels is relatively small. Estephan et al. [15,16] used the partial turbulence simulation (PTS) method to conduct wind tunnel tests on photovoltaic panels installed at different positions on the roofs of low-rise buildings. The peak static pressure coefficient obtained from the test is consistent with ASCE 7–16 [17]. The results show that the values for ASCE 7–16 are significantly lower than the crest coefficients in the current study. Peng et al. [18] studied the influence of length, inclination angle, position, spacing, and parapet height on the wind load of photovoltaic panels through wind tunnel tests. The results show that the PV panel position is a key factor for the wind load on PV panels, while the parapet greatly reduces the negative pressure peaks on the arrayed PV panels, and the PV panels on low-rise buildings are more susceptible to airflow reattachment compared with high-rise buildings.

Although there are many studies on the wind load characteristics of roof-mounted solar arrays, there are not many studies on the influence of parapets on the wind loads of photovoltaic panels, and the current research is not deep enough. To this end, this paper conducts pressure tests on two different models of roof-mounted solar arrays, with and without parapets, and analyzes their pressure coefficient distribution, extreme wind pressure coefficient, angle of appearance, and area-averaged pressure coefficients. The influence of the parapet was revealed, and finally the results were compared with the standards of various countries to provide a reference for the design.

2. Experimental Setups

2.1. Wind Tunnel Facility and Wind Profiles

The experiments were carried out at the ZD-1 Wind Tunnel Laboratory of Zhejiang University, a return-flow boundary layer wind tunnel of 18 m in length, 4.0 m in width, and 3.0 m in height, which is located in Hangzhou, China The test wind speed can range from 0 m/s to 55 m/s in this tunnel, and the non-uniformity turbulence intensity of the wind velocity in the test region is less than 1%. The aerodynamic wind pressure acting on the solar panels was measured by the Zoc33 digital modules, which are produced by the Scanivalve Company, Washington, DC, USA.

The experiments were carried out under simulated wind flow over open terrain, where the mean wind profile followed the power law with a power exponent of α = 0.12 (GB 50009-2012). The height of 1 m above the ground of the wind tunnel was chosen as the reference height for the measurement of the wind profile. The measured mean wind velocity and turbulence intensity profile as well as the target open terrain profiles from GB 50009-2012 are displayed in Figure 1, where $U$ and $U_r$ are the mean wind speed and reference mean wind speed, respectively, $I_u$ is the turbulence intensity, and z is the height above the ground of the wind tunnel. The simulated profiles agreed well with their theoretical values. The reference mean wind velocity ($U_r$) and turbulence intensity ($I_u$) at the reference height were 10.65 m/s and 5.83%, respectively. As the geometry scale of approaching flow was 1:50 and the time scale was 1:12.2 in the wind tunnel test, the velocity scale of the approaching flow was 1:4.1.

2.2. Pressure Model Configurations

The layout of roof-mounted solar arrays is shown in Figure 2. The flat-roofed low-rise building was 51.0 m in length, 19.0 m in width, and 25.0 m in height. Five rows of solar panels were set on the roof. The dimensions of the research object for this experiment were determined after investigating the dimensions of several industrial plants where solar panels were installed with the design institute, including the arrangement of solar panels and the height of the parapet, which are representative. Each row consisted of 22 modules, and each module had two units (see Figure 3a,b). The dimensions of each unit were 2.094 m in length and 1.038 m in width. With a gap of 0.02 m, the dimensions of a typical module were 2.094 m in length and 2.096 m in width, and its weight was 24.5 kg. Thus, the length of each row was 46.488 m since the rows were set with gaps of 0.02 m between modules. The solar arrays were symmetrically mounted on the roof with a row spacing of 1 m and a setback of 2.26 m from the edge of the roof in the longer and shorter directions of the building. The tilt angle (α) of the module was 10°, as shown in Figure 3c. The tilt direction of the solar panels is facing the direction of the wind with an angle of 0°. The module was supported by two columns, with the lower end being 0.5 m off the rooftop. Parapets (see Figure 2) are also set on the roof of the building, and a parapet is made of plate. To investigate their effects on the aerodynamic wind loads of solar arrays, low-rise buildings with parapet heights of 0 m and 1.2 m were modeled in this study. The solar arrays would be occluded when the parapet height was 1.2 m since the upper end of the model was less than 0.9 m from the rooftop.

Given the size of the wind tunnel test section and the constraints of blockage less than 5%, a geometric scale of 1:50 was used to make the test model. The aerodynamic wind pressures on solar arrays were measured in the wind tunnel test. The layout of pressure taps marked by ⚪ is shown in Figure 3a. A total of 496 pressure taps were drilled on the panels: half on the upper surface and the other half on the lower surface. Up to 24 taps were drilled on the module at the end of an array. Since the building and solar arrays were symmetrical, the pressure taps were only drilled on the half zone of the arrays. Then the wind pressure on the other half can be obtained by symmetry.

Pressure signals were acquired at a sampling frequency of 312.5 Hz over a period of 90 s. Thus, over 28,000 data points were recorded for each measured point. Measurements were taken for 24 incident wind angles at 15° intervals for the full 360° azimuths, where 0° was perpendicular to the wide face acting in the shorter direction of the building and 90° was perpendicular to the narrow face acting in the longer direction, as defined in Figure 2. As shown in Figure 3c, at a wind direction of 0°, the upper surface of the panel faces the approaching wind flow. Figure 4 shows the model arrays mounted on the roof of a low-rise building at a wind direction of 0° in the wind tunnel.

3. Data Processing

3.1. Pressure Coefficients

The wind pressure coefficients on the upper and lower surfaces of solar panels, denoted as $C_{pu}$ and $C_{pl}$, respectively, are defined as follows:

$$C_{pu}(i)=\frac{P_u(i)-P_{\infty }}{0.5\rho U_H^2}\begin{array}{cc}& (i=1,2,\dots ,248)\end{array}$$

(1)

$$C_{pl}(i)=\frac{P_l(i)-P_{\infty }}{0.5\rho U_H^2}\begin{array}{cc}& (i=1,2,\dots ,248)\end{array}$$

(2)

where $P_u(i)$ and $P_l(i)$ are the wind pressures on the upper and lower surfaces, respectively (Pa), $P_{\infty }$ is the ambient atmospheric pressure (Pa), $\rho$ is the air density (kg/m³), $U_H$ is the mean wind speed at the roof height, and $i$ is the number of pressure taps. The net pressure coefficient $C_{pn}$, defined as positive downward in Figure 3c, is the difference of the local pressure coefficients on the upper and lower surfaces:

$$C_{pn}(i)=C_{pu}(i)-C_{pl}(i)\begin{array}{cc}& (i=1,2,\dots ,248)\end{array}$$

(3)

Then the area-averaged net pressure coefficients $C_{An}$ on a module are defined as:

$$C_{An}={\displaystyle \sum _{i=1}^n{A}_i{C}_{pn}(i)}/{\displaystyle \sum _{i=1}^n{A}_i}$$

(4)

where n is the number of taps on a unit or module, $A_i$ is the tributary area of the pressure tap $i$. As shown in Figure 3a, n can be 1, 2, 4, or 6 for units and 2, 4, 8, and 12 for modules.

In the structure design of a solar panel, the largest and smallest peak pressure coefficients are typically considered by structural designers. To obtain peak pressure coefficients (including area averages) with a more statistically stable quantity, the Lieblein BLUE formulation (1974) was applied in this study. This involves two steps: first, divide the recorded time series into 10 equal segments and take the peaks from each of the segments; second, estimate the peak pressure coefficients by Type I extreme value distribution.

3.2. Shape Factors

In the Chinese standard (GB 50009-2012) shape factor is used to calculate the wind pressure acting on an object with a certain shape [19]. It can be obtained by the mean area-averaged net pressure coefficient ($C_{An,mean}$) and wind pressure height coefficient as follows:

$${\mu }_s=C_{An,mean}{\mu }_H/{\mu }_{zp}$$

(5)

where ${\mu }_s$ is the shape factor, ${\mu }_H$ and ${\mu }_{zp}$ are the wind pressure height coefficient at the mean heights of the roof (H = 25.0 m) and the solar panels ($z_p$ = 25.7 m), respectively, and ${\mu }_H/{\mu }_{zp}={\left(H/z_p\right)}^{2\alpha }=0.9934$.

In the Chinese standard NB/T 10115-2018 [20], when the tilt angle of a module is less than 15°, the shape factors are 0.8 and −0.95 for downward and uplifted wind loads, respectively. However, it is not distinguished in the NB/T 10115-2018 whether modules are mounted on the ground or on the roof, constraining the accuracy and practicality of its specified shape factors. In the Japanese standard JIS-C-8955 (2018) [21], when the tilt angle of a flat-roof mounted module is less than 10°, the shape factors are 0.75 and 0.60 for the downward wind load on exposed and non-exposed modules, respectively, and −0.6 for the uplift wind load on all modules. In the ASCE standard ASCE 7–22 (2022) [22], the net pressure coefficient, not the shape factor, is used for the determination of wind pressure for rooftop modules. However, the shape factor can be calculated from the net pressure coefficient as follows:

$${\mu }_s=K_d\left(G{C}_{rn}\right)$$

(6)

$$\left(G{C}_{rn}\right)={\gamma }_p{\gamma }_c{\gamma }_E{\left(G{C}_{rn}\right)}_{\mathrm{nom}}$$

(7)

where $K_d$$=0.85$ is the wind directionality factor, $\left(G{C}_{rn}\right)$ and ${\left(G{C}_{rn}\right)}_{\mathrm{nom}}$ are the net pressure coefficient and nominal net pressure coefficient, ${\gamma }_p$ is the parapet-related parameter, ${\gamma }_c$ the module width-related parameter, and ${\gamma }_E$ the module exposure-related parameter. The detailed information about the above parameters is described in Section 29.4-3 of ASCE 7–22.

4. Results and Discussions

4.1. Local Pressure Coefficients

4.1.1. Unfavorable Wind Direction

Some contour plots of wind pressure coefficients on the net surface of solar arrays at typical wind directions are presented in Figure 5. It can be seen from the figure that the maximum positive wind pressure coefficient and the maximum negative wind pressure coefficient of the solar arrays at the wind angles of 45° and 135° are larger than those at the wind angles of 0° and 90°. It can be further speculated that in some areas of the solar arrays, the extreme values of the wind pressure coefficient are larger at oblique wind angles.

The largest and smallest mean net pressure coefficients $C_{pn,mean}$ among all taps at each row with parapet height $h_{pt}$ = 0 m are shown in Figure 6. In general, solar panels are subjected to suction forces and pressure forces with comparable magnitudes when the tilt angle is 10°, almost parallel to the roof, which is different from the previous studies of Kopp et al. (2012) [7] and Peng et al. (2022) [18]. They found that when tilt angles are between 20° and 30°, the solar panels are mainly subjected to suction forces. So, it is reasonable to infer that the tilt angle has a great influence on the net pressure feature of the solar panel. The largest and smallest most critical $C_{pn,mean}$ occurred at oblique wind directions as conical vortices over rooftops and were generated by building corners. The most critical largest $C_{pn,mean}$ occurred with $\theta$ = 45°, while the most critical smallest $C_{pn,mean}$ appeared with $\theta$ = 135°, which is consistent with the literature of Kopp et al. (2012) [7], Cao et al. (2013) [10], and Peng et al. (2022) [18]. Rows 1# and 5# are the head-on rows at $\theta$ = 45° and 135°, reducing the most critical $C_{pn,mean}$. At $\theta$ = 45°, the most critical largest $C_{pn,mean}$ equaling 1.34 at row 1# is close to that at row 3#, and approximately 2.9 times as large as those at rows 2#, 4#, and 5#; at $\theta$ = 135°, the most critical smallest $C_{pn,mean}$ equaling −1.53 at row 5# is approximately 1.2, 2, and 4 times as large as those at rows 1# and 3#, 4#, and 2#, respectively. For non-exposed panels, the largest and smallest $C_{pn,mean}$ at row 3# are larger in magnitude than those at rows 2# and 4#; this is because taps at row 3# are closer to the edge of panels (see Figure 3a), facing more wind flow (Peng et al. (2022) [18]).

The largest positive and negative peak net pressure coefficients $C_{pn,peak}$ among all taps at rows 1#–5# with $h_{pt}$ = 0 m are shown in Figure 7. As the panels were almost parallel to the roof, the largest positive and negative $C_{pn,peak}$ are also comparable in magnitude. The unfavorable wind directions are from 30° to 75° for positive peaks, with the most critical peak of 2.81 at row 1#, and from 120° to 165° for negative peaks, with the most critical peak of 3.12 at row 5#. Similar to mean net pressure coefficients, the most critical peaks occur at the head-on rows.

Figure 8 and Figure 9 show the instantaneous point pressure coefficient distribution on the net surface, upper surface, and lower surface at the instant the largest positive peak net pressure coefficient appeared among taps on the entire solar array for a wind direction of 0° and 45°. It can be seen that the location where the largest pressure coefficient occurs on the three surfaces changes for the two wind angles, as does their magnitude. This indicates that the wind angle has a strong influence on the distribution of peak wind pressure on the solar arrays.

4.1.2. Effect of Parapet Height

The largest mean, smallest mean, largest positive peak, and smallest negative peak coefficients among all taps with wind directions are shown in Figure 10. In most wind directions, the parapet has little effect on pressure coefficients, as the picked values are relatively comparable. However, at oblique wind directions, the parapet has a significant impact on pressure coefficients. At $\theta =45°$, the largest mean and positive peak pressure coefficients are 0.83 and 1.69 with parapet, respectively, about 31% and 37% lower than the values without parapet. At $\theta =135°$, the smallest mean and negative peak pressure coefficients are −0.64 and −18.2, respectively, about 42% and 58% lower than the values without parapet. To further understand the effect of the parapet, the locations of the picked pressure coefficients are investigated and shown in Figure 11. As seen, the largest and smallest values occur at the top left and right corners without parapet, indicating that they are significantly influenced by the conical vortices generated by the building corners. However, when the parapet height is 1.2 m, the largest and smallest values occur in the inner zone, whereas the occlusion effect of the parapet reduces the pressure coefficients in the outer zone and has little impact on the inner zone. Wang et al. [23] also reached a similar conclusion, but there are some discrepancies at 180°. The study in this paper found that under certain wind direction angles, the negative peak coefficients of solar panels with parapet may be larger than those without parapet, but this phenomenon did not occur at 180°. This may be related to various factors, such as the tilt angle of the solar panel and the height of the building.

4.2. Area-Averaged Net Pressure Coefficients

The area-averaged net pressure coefficients of each row are plotted in Figure 12, with each row consisting of two PV panels, the front and the back, with black representing the back row and red representing the front. According to the results at the displayed wind angles, there are significant fluctuations in the values of the area-averaged net pressure coefficients at the ends of the rows compared to the other positions, particularly in the first and second rows at 45° and in the fourth and fifth rows at 135°. At 0° and 90° wind directions, the area-averaged net pressure coefficients do not vary very much from row to row at each position. In contrast, the difference between the wind pressure coefficients of the rows is relatively large at 45° and 135° wind directions, which is further illustrated by the fact that the wind pressure coefficient of the row on the windward side increases significantly under the effect of oblique wind directions. Comparing the results of the front and back rows in the same row, it can be seen that the area-averaged net pressure coefficients for the front and back rows show good consistency with the change in position.

The largest and smallest mean area-averaged net pressure coefficients and the largest positive and negative peak area-averaged net pressure coefficients on modules at rows 1#–5# were picked out, and their variation with wind direction is shown in Figure 13, Figure 14, Figure 15 and Figure 16, which include the cases of $h_{pt}$ = 0 m and $h_{pt}$ = 1.2 m. In the case of $h_{pt}$ = 0 m, the largest mean area-averaged net pressure coefficients in each row occur at a wind direction of 30–60°, and the smallest mean area-averaged net pressure coefficients occur at a wind direction of 135–165°. The largest positive and negative peak area-averaged net pressure coefficients occur in similar locations, but near the wind direction of 180°, the largest positive peak area-averaged net pressure coefficients also appear extreme values, but the largest mean area-averaged net pressure coefficients at the same wind angle are approaching 0, indicating greater local turbulence at these wind directions. A similar conclusion to those mentioned above can be obtained for the case of $h_{pt}$ = 1.2 m, except that the wind direction at which certain extremes occur changes slightly. In addition, for the mean and peak area-averaged net pressure coefficients of row 5#, the results of the case of $h_{pt}$ = 1.2 m are smaller than those of the case of $h_{pt}$ = 0 m. This indicates that the presence of the parapet can reduce the mean and extreme area-averaged net pressure coefficients in some areas of the solar arrays.

The most critical positive and negative peak area-averaged net pressure coefficients on modules and the associated wind directions for $h_{pt}$ = 0 m and $h_{pt}$ = 1.5 m are given in Figure 17 and Figure 18. The most critical positive peak area-average net pressure coefficients of all modules are mostly between 0.5 and ~1.0, and the most critical negative peak area-average net pressure coefficients of all modules are mostly between −0.5 and ~−1.0 for the two conditions. Compared to the condition without a parapet, the most critical (a) positive and (b) negative peak area-averaged net pressure coefficients on row 5# are reduced. Due to the existence of the parapet, the wind directions where the most critical positive peak area-averaged net pressure coefficients occur are more between 30 and −75°, and the area where the most critical negative peak area-averaged net pressure coefficients occur at 0° is greatly reduced.

4.3. Shape Factors

The module shape factors are calculated by the module’s largest and smallest mean pressure coefficients. The most critical positive and negative shape factors and their associated wind directions for parapet heights of 0 m and 1.2 m are shown in Figure 19 and Figure 20. When parapet height is 0 m, the most critical positive shape factors are larger at the ends of rows from 1# to ~3# and in the middle of rows from 3# to ~5#, with a maximum of 0.51. The shape factors in the middle of row 1# can be as small as 0.05. Consistent with the module pressure coefficients, the most critical shape factors occur at oblique wind directions from 30° to 75°. The most critical negative shape factors are larger at the ends of rows from 1# to ~5# and in the middle of rows from 1# to ~2#, with a maximum of −0.53. However, the critical negatives still occur when the lower module faces the direction of the wind flow (from 120° to 150°). When the parapet height is 1.2 m, both the most critical positive and negative shape factors reduce at the corner and outer rows, where the modules are directly impacted by the occlusion effect of the parapet, while less change occurs at the inner zones. The means of the most critical positive and negative shape factors with $h_{pt}$ = 1.2 m are 0.196 and −0.190, respectively, which are quite close in magnitude but are about 30% lower than the values (0.261 and −0.256, respectively) with $h_{pt}$ = 0 m. Nevertheless, the wind directions associated with the critical values are not changed by the parapet.

Shape factors of solar panels in The Chinese standard NB/T 10115-2018, Japanese standard JIS-C-8955 (2018) and American standard ASCE 7–22 are shown in Figure 21, Figure 22, Figure 23 and Figure 24. For the Chinese standard NB/T 10115-2018, the positive shape factors are uniform with a value of 0.80. In addition, so are the negative shape factors, with a value of −0.95. The positive and negative shape factors determined by the Japanese standard JIS-C-8955 (2018) are 0.75 and 0.60 at the exposed and unexposed zones, respectively, and uniformly −0.60. The positive shape factors determined by the American standard ASCE 7–22 are slightly smaller in magnitude than the negative shape factors. The positive shape factors with $h_{pt}$ = 0 m and 1.2 m are uniformly 0.55 and 0.58, respectively, while the negative shape factors are as large as −0.83 and −0.88 at the exposed zones. The means of positive and negative shape factors determined by the Chinese, Japanese, and American standards are 0.80, 0.64, and 0.55; $h_{pt}$ = 0 m), 0.58 ($h_{pt}$ = 1.2 m), and −0.95, −0.60, −0.69 ($h_{pt}$ = 0 m) and −0.73 ($h_{pt}$ = 1.2 m). Compared to the shape factors calculated by the standards, the experimental shape factors are much smaller. The experimental positive and negative shape factors of $h_{pt}$ = 1.2 m are only 24.4%, 30.5%, 33.5%, 20.0%, 31.7%, and 26.3% of the standard values. This indicates that the guidelines for the design wind pressure on roof-mounted solar arrays are very conservative when the tilt angle of the module is 10°, and more research is needed to make the standards applicable and complete. Compared to the research results of Wang et al. [24], the experimental values are smaller than the standard values when the subordinate area is relatively small. However, Wang et al. [24] also compared with the American standard ASCE 7–22 for large subordinate areas, and the result was the opposite.

5. Conclusions

A wind tunnel test was conducted to investigate the effects of parapets on the aerodynamic wind loads of roof-mounted solar arrays. The distribution of the mean wind pressure coefficient and the extreme wind pressure coefficient in the solar arrays were discussed in detail, and the results were compared with some national standards. Results show that the mean and extreme values of the wind pressure coefficient are larger at oblique wind angles. The most critical largest $C_{pn,mean}$ occurred at $\theta$ = 45°, while the most critical smallest $C_{pn,mean}$ appeared at $\theta$ = 135°. The unfavorable wind directions are from 30° to 75° for positive peaks and from 120° to 165° for negative peaks. The most critical positive peak area-average net pressure coefficients of all modules are mostly between 0.5 and ~1.0, and the most critical negative peak area-average net pressure coefficients of all modules are mostly between −0.5 and ~−1.0 for the conditions with and without parapet.

The presence of the parapet can reduce the mean and extreme area-averaged net pressure coefficients in some areas of the solar arrays, and both the most critical positive and negative shape factors reduce at the corner and outer rows, where the modules are directly impacted by the occlusion effect of the parapet, while less change occurs at the inner zones.

Compared to the shape factors calculated by the standards, the experimental shape factors are much smaller. The guidelines for the design of wind pressure on roof-mounted solar arrays are very conservative, and more research is needed to make the standards applicable and complete.

The wind load characteristics of solar panels are not only related to the height of the parapet but, more importantly, the tilt angle of the parapet, which will be studied in later research.