A Fast Maximum Power Point Estimator for PV Panels Installed at the Rooftop of a Vehicle

This paper introduces a maximum power point tracking (MPPT) method based on an artificial neural network (ANN). The method uses measured voltage and power samples as inputs to predict the voltage adjustment required to achieve the global maximum power point (GMPP). This approach eliminates the need for irradiance and temperature sensors, making it a more compact and cost-effective solution for vehicle rooftop PV arrays (VRPVAs). Additionally, a perturb-and-observe procedure is incorporated to refine the ANN’s predictions for more accurate GMPP tracking. The proposed method has been experimentally validated under various partial shading conditions, including three static and two dynamic scenarios derived from Taipei streets. Comparative analysis with several existing MPPT methods, including two designed specifically for VRPVAs, demonstrates that the proposed method achieves superior efficiency and identifies the GMPP with remarkable speed, effectively substantiating its advantages. The dynamic tracking accuracy can be improved by more than 6.5% compared to some of the existing ANN, other conventional and hybrid methods.