Study On Health Status Assessment And Failure Distribution Estimation For Traction Inverter Of Urban Rail Vehicle

The urban rail vehicle is an indispensable part of urban rail transit system.Electrical system is an important subsystem of urban rail vehicle.Traction inverter is core component of electrical system.Obviously,traction inverter is the key to safe and reliable operation of urban rail vehicle.Health status assessment theory mainly studies safety,and failure distribution estimation theory mainly studies reliability.Therefore,it is necessary to carry out health status assessment and failure distribution estimation for traction inverter to find its weak links as early as possible then to guide operation and maintenance of urban rail vehicle.The content of the thesis includes two parts:(1)Research on health status assessment for urban rail vehicle traction inverter based on cross entropy and SVM theory:In this section,the basic principles of cross entropy and SVM theory are introduced first briefly.The composition of urban rail vehicle traction inverter is summarized.The index system construction is complete in conjunction with relevant specifications.Simultaneously,five health status levels are established and data processing is complete.Then,the cross entropy theory is used to combine subjective method,that is analytic hierarchy process method,objective method,that is entropy weight method and data method,that is gray correlation degree method to solve the combined weights of each index.Afterwards,combined with operating data and an SVM assessment model which corresponds operating data to each status level of traction inverter,health status assessment for urban rail vehicle traction inverter is completed.(2)Research on failure distribution estimation for urban rail vehicle traction inverter based on cross entropy and Markov theory:In this section,based on a brief introduction of the basic principles of cross entropy and Markov theory,according to the basic working principles of urban rail vehicle traction inverter,combined with Markov state transition process,the transition probability matrix of inverter's operating states are obtained,and the initial expression of the state transition probability matrix is set.Then,by introducing the failure rate parameter and other related parameters,the initial parameter values are set.The cross entropy theory is used to fuse three sets of conventional different values of the key elements m and n to determine the combined model architecture.The failure condition within 1 hour and 20 years are analyzed mainly.Based on the above work,the failure distribution estimation for urban rail vehicle traction inverter is finally completed.The research results show that the health status assessment for urban rail vehicle traction inverter based on cross entropy and SVM theory solves the problem of traction inverteroperating status assessment,that is,safety analysis.While combining cross entropy with Markov theory,the failure distribution estimation for urban rail vehicle traction inverter has effectively completed the estimation of traction inverter failure condition,that is,the reliability research.