A Novel Hybrid Method Based on Fireworks Algorithm and Artificial Neural Network for Photovoltaic System Fault Diagnosis

In the last years, the need for robust photovoltaic systems has been escalated, which has been involved the high accuracy and the robustness of the diagnostic systems, whether at a normal run or when facing unexpected events. Therefore, the diagnostic systems must stick strictly to the security- or safety-criteria of the system components and humans, while strengthening their ability to improve the efficiency of the produced energy. Recently, the research has been focused mainly on the development of an intelligent diagnostic system characterized by precision, stability, and speed. In this article, a new diagnostic system is based on the Artificial Neural Network (ANN) and the Evolutionary Fireworks Algorithm (FWA) has been proposed. The method proposed in this article combines the prediction of two different algorithms to obtain more satisfactory accuracy, the objective of hybridization is to optimize the artificial neural network in terms of precision and convergence. The state of the photovoltaic field is determined by the diagnostic model, based on the voltage and current delivered by the field in real-time to make the diagnostic assignment uncomplicated and more reasonable. Our novel hybrid model FWA-ANN has defeated the PSO-ANN model in terms of high accuracy and reduced time of convergence by less than a half, which has proved that our proposed method is a promising model for Neural Network optimization for the PV field compared to the previous works. The results in this study prove that the hybrid method achieves an accuracy of 99.98% in 241 iterations compared to the ANN model which reaches only 99.94% in 682 and the PSO-ANN model which achieves 99.95% accuracy in 564 iterations.