Research On Reliability Evaluation Methods Of Emu System

The reliability of the EMUs has become an important subject of concern in the rail industry.EMUs have sophisticated operating systems and are running in a complex environment,so the research on the reliability evaluation methods of EMUs is playing a significant role in mastering the regular patterns of EMU’s malfunctions,optimizing EMU’s maintenance procedures,improving the reliability of EMUs,and guarding the safe operations of the rail system.In this paper,the EMU and its subsystems are taken as the research object.Based on the data of EMU’s malfunctions in the daily operations,it applies the relevant theories and technologies such as statistics,reliability theory,optimization theory,machine learning and neural network to analyze and evaluate the operational reliability of the EMUs.The relevant research contents and main conclusions are as follows:(1)Based on the analysis of the basic structure of EMU body and end connection system,bogie system,air supply and braking system,high-pressure traction system and network control system,the fault tree model of each system is established,and the method of minimum cut set is used to determine the weak links of each system.(2)Based on the Weibull distribution model,the cumulative failure probability function,reliability,mean distance between failures,failure rate function and other related reliability indexes of each subsystem of the EMU are calculated,and the reliability of the whole EMU system is evaluated.Using the optimization theory,the Adam optimization algorithm is applied to the parameter estimation of Weibull distribution.The experimental results show that the distribution model obtained by Adam optimization algorithm can improve the accuracy of fitting results.(3)Based on machine learning and neural network algorithm,the average failure rate of EMU is predicted.According to the existing calculation method of the average failure rate of EMU,the average failure rate of EMU is calculated by taking 10000 km as the time interval.On this basis,the EMD-LSTM model is established to predict the average failure rate of EMU,and the future 10 average failure rate values are predicted.The prediction results show that the model can achieve a certain accuracy for the prediction of the average failure rate,and can effectively predict the trend of the average failure rate.(4)Based on the requirements of EMU operational reliability analysis and system design principles,the operational reliability analysis system of EMU system is designed and developed by using Python language and Flask framework.The system realizes the functions of EMU fault data management,reliability quantitative index calculation,EMU average failure rate prediction and so on.The system is easy to operate and improves the efficiency of EMU Operation Reliability Analysis and the practicability of EMU operation reliability analysis.