Open AccessResearch on Fault Diagnosis of DC Charging Pile Power Device Based on Wavelet Packet and Elman Neural Network
Author(s) -
Jiajia Wang,
Xingying Chen,
Ji Li
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/486/1/012086
Subject(s) - fault (geology) , artificial neural network , rectifier (neural networks) , wavelet , network packet , signal (programming language) , computer science , wavelet packet decomposition , power (physics) , generalization , control theory (sociology) , electronic engineering , engineering , pattern recognition (psychology) , artificial intelligence , wavelet transform , recurrent neural network , mathematics , computer network , mathematical analysis , physics , stochastic neural network , control (management) , quantum mechanics , seismology , programming language , geology
In order to improve the fault diagnosis accuracy of DC charging pile power devices, a fault diagnosis method based on wavelet packet analysis (WPA) and Elman neural network is proposed in this paper. This method sampled the output voltage signal of DC bus in fault state, decomposed the three-layer db10 wavelet packet and reconstructed the single branch, then calculated the characteristic energy spectrum of the fault signal using the signal in the frequency band, and identified it by Elman neural network. In order to test the diagnostic ability of the model, the PWM rectifier model of DC charging pile is used as an example to simulate and compare with the diagnostic results of standard BP neural network. The simulation results show that the fault diagnosis method based on WPA and Elman neural network has faster diagnosis speed, higher accuracy and stronger generalization ability.