Research On Fault Diagnosis For High-speed Train Motor System Based On Multi-model Estimation

Because of the wide operating range and complex operating environment of high-speed train in China,there is a high demand for safety and reliability of the system,and various kinds of faults should be diagnosed and detected in time.This work is supported by the National Natural Science Foundation of China,carries on the fault diagnosis research for the high-speed train traction motor system.The traction motor,traction inverter,rectifier and other components constitute the traction system of the high-speed train,which provides power for the train.In this paper,the fault diagnosis algorithms for the rotor broken bar fault of the traction motor and the IGBT component fault of the traction inverter are studied.The working principle of the high-speed train traction system is analyzed,and according to the working mechanism of the traction motor and the traction inverter,the healthy system models are established.Aiming at the broken rotor bar fault of traction motor,the model is simplified.The fault models of the system with different number of broken bars and IGBT fault in different position of inverter are established.When the inverter works,the switch state of the inverter is switching constantly,so the model parameters are changing.Based on this situation,the interactive multi-model estimation algorithm is improved to track the system state.The concepts of multi-model estimation and interactive multi-model are introduced,and the fault diagnosis method is proposed.Based on the model set of broken bar fault and the nonlinear characteristic of traction motor,a nonlinear filter bank is designed to obtain residual sets.An interactive multi-model diagnosis algorithm is used to estimate the state of the motor,and the system working model can be determined.Combined with the fault model set,the development degree of the broken bar fault can be diagnosed.Considering the situation that the output current sensor of the inverter in the high-speed train is a redundancy design composed of the same two groups.The distributed filter based on synchronous data fusion is designed to estimate the state of the inverter system under different working models.The inverter state is diagnosed by the interactive multi-model algorithm with filter state estimation and residuals.The fault IGBT components can be located with the fault model set.At last,the feasibility and validity of the fault diagnosis method are verified by simulation,the improvement and further research are prospected.