Photovoltaic Leakage Current Detection Method Based on Variational Mode Decomposition and Random Forest Algorithm

Aiming at the difficulty of accurate detection of leakage faults in photovoltaic low-voltage distribution area, a new method of photovoltaic leakage current detection based on variational mode decomposition and random forest algorithm is proposed. Firstly, the mechanism of leakage current generation of photovoltaic electrical equipment is deeply analyzed. Secondly, the signal is decomposed by variational mode decomposition to reduce the influence of high frequency signal on diagnostic accuracy. Then, in order to solve the problems of high complexity and easy overfitting when the number of continuous attribute values is too large, the improved C4.5 decision tree and classification regression tree algorithms are proposed. Finally, compared with support vector machine algorithm and traditional random forest algorithm, the accuracy of the proposed method can reach 98.89%, which effectively verifies the accuracy and feasibility of the proposed method.