A Data-Driven Calibration of the FEM Temperature Model with Wind Direction Input

This paper proposes an enhanced method for photovoltaic (PV) temperature (TPV) modeling using a datadriven approach. By modifying the methodology of Herteleer et al. [1], we aim to make the FEM temperature model easier to calibrate and more accurate. The calibration of Faiman’s model [2] coefficients is optimized using modern computational techniques, minimizing the error between measured and estimated temperatures. A novel component is introduced: the inclusion of wind direction as a variable influencing the cooling effect on PV systems. Our results demonstrate an improvement in model accuracy, reducing the mean absolute error (MAE) by 0.29°C compared to previous methods. This advancement provides a more refined tool for estimating TPV in real-world conditions.