Intelligent shading fault detection in a PV system with MPPT control using neural network technique

Photovoltaic (PV) power generation systems know widespread in the power generation world due to their production efficiency of clean energy. This system is exposed to several faults and errors during the production process, which reduces the quality and quantity of the produced energy, among the most common defects is partial shading. This paper proposes a simplified method for fault detection based on the generation of residual signals sensitive to these faults. For this detection, we have developed a model of the healthy photovoltaic system based on an artificial neural network (ANN). The output of this model is compared to the PV generator controlled by maximum power point tracking (MPPT) to form a residue used to feed a mechanism dedicated to fault detection. For the detection mechanism, an ANN was used as a fault classifier. The proposed method makes it possible to determine the percentage of partial shading, even in the presence of climate change. The results have been verified and validated using MATLAB/Simulink.