Simultaneous Operating Temperature and Output Power Prediction Method for Photovoltaic Modules

Accurate cell temperature and output power prediction are vital for the optimal design and operation of photovoltaic (PV) systems. However, capturing the accurate relationships between cell temperature/circuit parameters and weather conditions is still a challenging task. In this study, a universal radial basis function neural network based hybrid modeling approach is proposed to model the cell temperature and circuit parameters. A simultaneous optimization model with l1 norm penalty is established and a separate parameter estimation strategy is proposed to handle the high computational parameter estimation procedure. The effectiveness of the proposed hybrid modeling approach is validated based on four practical experimental datasets of both commercial and laboratory PV plants. It is thus indicated that the proposed modeling approach could provide a promising potential solution framework for the accurate output power prediction under different PV types and relatively wide weather conditions.