Artificial neural network-based fault detection and classification for photovoltaic system

This paper handles with the artificial neural network-based fault detection and classification method for a photovoltaic (PV) system. In fact, the diagnosis of the PV plants is extremely required in order to maintain the optimum performance. Therefore, in this paper, we used the artificial neural network for the fault classification. The studied system composed of a PV array with the implementation of Perturb and Observe (P&O) maximum power point tracking (MPPT) using boost converter. The simulation has been accomplished in MATLAB/SIMULINK software. Five different faults have been implemented in the PV system. We used the neural network fitting tool to build and train the network and evaluate its performance using the mean square error (MSE) and regression analysis. The proposed technique has proved effective in terms of results.