Photovoltaic Output Power Modeling: a Hybrid Approach

This study presents a physics-based model aimed at estimating the thermal-electrical performance of photovoltaic (PV) systems. The hybrid model integrates fundamental physical principles with empirical data, thereby improving reliability while reducing data requirements. The physics equations are derived from a thermoelectric dynamic model of the PV module, encompassing reflection losses, module temperature determination, PV module efficiency calculation, and estimation of systematic losses. In addition, Bayesian inference is employed to mitigate uncertainties associated with physical parameters values. Numerical simulations demonstrate the model's ability to accurately describe both module operating temperature and power output, accounting for the temperature effect on PV conversion efficiency. Error analysis underscores its superior accuracy compared to traditional physical modeling approach. This research contributes to advancing PV power modeling by offering a robust and reliable methodology combining physics-based and data-driven techniques.