Fault Diagnosis of Electric Vehicle Charging Station Based on Empirical Wavelet Transform and Entropy

Abstract

Considering the influence of non-Gaussian noise in electric vehicle charging stations on converter fault signals, in order to detect fault types accurately, a method of fault detection of inverter was proposed, which combines empirical wavelet transform(EWT) and cyclic entropy(CCE) First of all, the model of electric vehicle charging station and analysis of the open circuit fault of IGBT components are given. Secondly, the improved EWT method is used to decompose the original signal, and the modes obtained from decomposition were screened. Once more, considering the problem of the degradation of EWT s decomposition performance of non-Gaussian noise, the reconstructed signal was mapped to the cyclic entropy spectrum by cyclic entropy, and the fault feature values of various types of open-circuit faults were calculated. Finally, the simulation experiment results show that this fault diagnosis method is suitable for inverter fault detection in a non-Gaussian noise environment, which has good accuracy and practicability.