A novel hybrid image processing-based reconfiguration with RBF neural network MPPT approach for improving global maximum power and effective tracking of PV system

Abstract

The solar photovoltaic (PV) array output is reduced significantly by the fre quently occurring inevitable partial shading conditions. In consequence, the array exhibits multiple peaks in its characteristics that cause the conventional maximum power point tracking (MPPT) algorithms to get stuck at the local maximum. So, to track the global maximum power (GMP) among the multiple peaks, a novel radial basis function (RBF)-based neural network approach has been proposed for predicting the optimal GMP. Additionally, a novel and intel ligent encryption-based ruler transform (RT) reconfiguration approach is pro posed to disperse the shading effect enhancing the GMP and mitigating the multiple peaks. The effectiveness of the proposed RBF-MPPT and novel RT reconfiguration strategies has been tested and analyzed for a 5 7 PV array under distinct dynamic, uniform, and nonuniform shading conditions. The results of the proposed RBF have been compared with the conventional incre mental conductance (INC) algorithm before and after reconfiguration of the PV array. Further, the ease of GMP tracking by a simple conventional INC due to the reduction of peaks after the array reconfiguration under shading condi tions has been demonstrated and discussed in detail. After reconfiguration, the GMPis enhanced by 37.35%, 31.41%, 30.86%, 21.46%, 13.69%, and 8.88%, using the proposed RBF for the considered five shading conditions. The steady-state oscillations are also considerably mitigated by employing the proposed reconfiguration and RBF strategies.