Application of Back-Propagation Neural Network in Multiple Peak Photovoltaic MPPT

Abstract

In a photovoltaic (PV) system that consists of multiple series-connected PV modules with bypass diode, there could be multiple peaks in the P-V curve of the PV system when the irradiance on PV modules become non-uniform, which results in conventional MPPT methods' failure in tracking the global maximum power point (GMPP). In view of this problem, we propose a novel GMPP tracking method based on back-propagation neural network (BPNN). The BPNN takes the irradiance on each PV module as input variables. After identification by the BPNN, the GMPP voltage is obtained, which acts as reference voltage to the DC-DC converter circuit to keep the PV system operating at GMPP. Simulation results showed that the proposed method has good adaptability and high precision.