Spatial and temporal analysis of wind effects on PV modules: Consequences for electrical power evaluation

•Surface temperature and performance of PV cells are affected by wind.•Spatial variations in temperature occur within PV modules.•Power output and hereby energy yield PV modules affected by spatial variations in temperature.•Spatial temperature variations need to be considered for accurate energy yield evaluation. Currently, most of the state-of-the-art energy yield prediction and evaluation models reported in the literature assume a uniform photovoltaic (PV) module temperature and air flow. However, during realistic wind flow conditions, local temperature variations occur over PV modules due to forced convection. This study aims to investigate how spatial temperature differences over the surface of a PV module that are caused by wind affect the convection of heat over the module, and how this influences the energy production by the module. A PV module composed of 2 rows of 9 solar cells each was placed horizontally in a wind tunnel. It has then been tested at free-stream wind velocities of 1, 2, 3 and 5m/s. Surface temperatures and ambient air temperatures have been measured along the module and vertical temperature gradients have been determined. The electrical characteristics of each individual cell in the PV module have also been measured. Substantial differences in performance between the solar cells caused by local variations in module temperature have been observed. These data have been used to calculate the heat transfer coefficient along the module. To illustrate the importance of including, in energy yield evaluation models, the temperature variations that occur between the cells of a same module, we have constructed an energy yield evaluation model for a standard 60-cell PV module illuminated at 1000W/m2 and calculated the power output for (1) a uniform heat transfer coefficient over the module, and (2) a realistic distribution of the heat transfer coefficient based on the wind tunnel experiments. Differences up to 4.3% have been observed, indicating that local variations in temperature within a PV module need to be taken into account for accurate energy yield evaluations. Using this improved modelling, the absolute errors remaining in the E-yield estimations can be reduced.