Study on Reliability of Traction System of Subway Train

Subway vehicles are gradually becoming the main components of urban public transportation systems due to their large capacity,comfort,low energy consumption,greenness,and convenience.They have also become the preferred means of transportation for citizens.At present,China attaches great importance to the urban rail transit industry,and the scientific research level of subway vehicles has also been greatly improved.However,the sudden failure of subway vehicles will still cause large-scale traffic congestion,among which the failure of the vehicle’s electrical system accounts for a large proportion.Especially for the traction system of subway vehicles,its reliability directly affects the passenger comfort and safety of train operation,and improving the reliability of the traction system and the efficiency of maintenance work will be a focus of future subway development.This article takes the traction system of an AC vehicle of a subway line for a subway line as the research object,analyzes the main components of the traction system,common failure forms and possible causes,calculates the reliability parameters of each subsystem module of the traction system,and uses the hierarchy The analysis method determines the comprehensive weight of each subsystem module to the entire system.Based on this,the Markov reward process is used to evaluate the reliability of the traction system and its subsystem modules.It is given that the traction system and each subsystem module are different.The curve of the availability change over time under the attenuation coefficient;the historical fault data of the traction system is collected according to the common fault forms of each subsystem module,and the reliability prediction models of the traction system are established by using BP neural network and gray neural network,and compared The prediction accuracy of different algorithms finally determined that the gray neural network could better realize the prediction of the reliability trend of the traction system.The main tasks are as follows:(1)In-depth analysis of the working principle,structural composition,common failure forms and possible causes of various components of the traction system of subway trains,collecting historical fault data of the traction system of a subway line train,and determining the traction system and Reliability characteristic parameters of each subsystem module.(2)Through the analytic hierarchy process of the traction system,the reliability evaluation index of the traction system is determined,and the comprehensive weight of each subsystem module to the entire system is calculated by using the analytic hierarchy process.Based on learning to use the Markov process to evaluate the reliability of the system,the reward coefficient and attenuation coefficient are added.The comprehensive weight calculated by the analytic hierarchy process is used as the reward coefficient,and a reasonable attenuation coefficient is determined to establish a Markov-based reward process.The reliability evaluation model of the traction system,and based on the reliability evaluation results,a reasonable reference optimal maintenance cycle is given.(3)Based on the calculated reliability characteristic parameters of the traction system,the BP neural network and gray neural network algorithms were used to predict the reliability trend of the traction system,and the prediction accuracy of the two was compared.Finally,the gray neural network was determined as the system.Feasibility of reliability trend prediction.By predicting the trend of traction system reliability changes,maintenance strategies can be reasonably adjusted based on changes in reliability indicators,which is conducive to improving the maintenance plan of train traction systems and improving train performance