Fault Diagnosis And Remaining-Life Prediction Of Traction Converter For High Power Electric Locomotive

At present,the high-speed train technology in China is in the high speed development stage.The accurate fault diagnosis and remaining-life prediction of the traction drive system can ensure the stable operation of the train,which has always been the focus of attention and research by the scholars.This paper studies the fault diagnosis and remaining-life prediction of the converter based on the project "Converter Health Management Basic Platform Research ".The specific research contents are as follows:(1)Research on transient fault diagnosis of converter.Taking the HXD1 C locomotive as the background,the converter structure and the operating principle of the inverter are introduced first.The direct torque control model of the inverter is built through the simulation software.Secondly,the open circuit fault of the inverter is selected as the main research object,and the output current waveform of the inverter is obtained by simulation of different kinds of open circuit faults.Wavelet analysis is used to decompose the waveform,and extract the energy and waveform proportional coefficients as the feature vectors of fault diagnosis.(2)A fault diagnosis system based on BP neural network and SOM neural network is studied.First,the BP neural network algorithm and the SOM neural network algorithm are introduced,and the fault category is coded as the required output of the BP network.At the same time,the structure of the BP neural network and the SOM neural network are designed respectively.Then,the effectiveness of the two kinds of diagnostic methods is verified by simulation,and the two kinds of methods are compared and analyzed.(3)Based on the IGBT switch experiment,the remaining-life prediction of converter is studied.Firstly,the dynamic characteristics and monitoring parameters of IGBT are studied.Secondly,88 groups of IGBT switch experiments were carried out,and the conducting voltage drop of Collector-Emitter was chosen as the monitoring parameter in this paper,and the data obtained were analyzed statistically.Finally,based on the grey model,a IGBT state degradation model is proposed,and the accuracy of the model is verified.