Research on Fault Diagnosis of Photovoltaic Array Based on ACA-RBF Neural Network Model

Abstract

photovoltaic (PV) array is always outdoors, which is affected by the environment and its own aging, resulting in frequent failures. If the PV array fault is not detected and processed in time, it will not only reduce the efficiency of the photovoltaic power generation system at the source, but also bring serious security risks to the system. Traditional fault diagnosis methods have some problems, such as high cost, high dependence on equipment, weak universality, and difficult mathematical modeling. How to reduce the time cost and detection cost of photovoltaic array fault diagnosis and increase the accuracy rate has become an urgent problem to be solved. In this paper, a photovoltaic array fault diagnosis model based on improved Ant Colony Algorithm (ACA) optimized Radial Basis Function Neural Network (RBFNN) is proposed. The central parameters of RBF neural network are optimized by the improved ant colony algorithm, and the optimized parameters are used to establish a model to effectively diagnose photovoltaic array faults. In practice, the probability of multiple faults is still relatively small, and it is difficult to obtain enough fault samples as training data. Therefore, the data set generated by the simulation model is used to train the model with considering the occurrence of four kinds of photovoltaic array faults and their combination. Experimental results show that the proposed fault diagnosis model can effectively distinguish typical photovoltaic array faults with high fault diagnosis accuracy. Compared with BP neural network and support vector machine model, the proposed fault diagnosis model has a faster convergence rate and higher accuracy.