Research On Fault Feature Extraction Of Charging Module Of DC Charging Pile

As the main charging facility of electric vehicle,DC charging pile can meet the demand of fast charging.In recent years,China has been vigorously developing DC charging piles.The supply and demand problems between the Charging Pile and Electric Vehicle have been improved,but it also faces the problem of frequent faults of DC charging piles.Among the many components of DC charging pile,the charging module is the most critical one with the highest failure probability.Moreover,the fault characteristics of charging module are not obvious and the causes of failure are difficult to identify,which leads to the time-consuming and laborious maintenance of charging module.Therefore,it is of practical and economic significance to study the fault process of charging module and extract effective fault feature vectors.In this thesis,a complete closed-loop simulation model of charging module is established.Based on the fault data of the simulation model,an effective fault eigenvector of charging module is extracted and fault diagnosis is performed.The main work is as follows:Firstly,the two-stage commutation electrical structure of the charging module is introduced,and VIENNA rectifier circuit is selected as its pre-stage structure according to the design purpose of the front and rear stages,which is responsible for the implementation of Active Power Factor Correction(APFC).The phase-shifting full-bridge converter circuit is the following structure and is responsible for electrical isolation and voltage conversion.The single-period control strategy and the average current control strategy were selected as the closed-loop control circuits of the two stages before and after the charging module,and the simulation was carried out on the MATLAB platform to verify the correctness of the simulation model of the charging module.Then open circuit simulation is performed for the easily faulty power device and electrolytic capacitor in the charging module.Based on the working principle and fault waveform of the charging module,the fault characteristics of the charging module are deeply analyzed.The three-phase input current is selected as the fault characteristic parameter of the charging module and fault signal is collected.And expounds the related theory,wavelet transform time-frequency analysis technology based on the charging module three phase input current under different fault state have the characteristics of different frequency components,using frequency analysis of wavelet packet energy spectrum method of three-phase input current of the charging module fault feature extraction,extraction of three-phase input current energy of each band with three-phase input current amplitude of the charging module of fault feature vector.Finally,the improved BP neural network method was used for fault diagnosis of the charging module.The designed fault feature vector samples were divided into training samples and test samples on the MATLAB platform to input the improved BP neural network for fault diagnosis of the charging module,to verify the validity of the extracted fault features and the correctness of the diagnosis method.