Study On The Condition Monitoring And Fault Diagnosis Method Of High-speed Train Damper

As a demonstration line of China's high-speed railway,Jing Zhang railway requires the EMU to be "intelligent".One of the important contents is the fault monitoring and health management of the EMU.First of all,the fault diagnosis of the key components in the EMU is necessary to guarantee the safe operation of the train.Secondly,after a long-time running,the performance parameters of the key components of the vehicle will change,which will affect the performance of the vehicle.It is an important part of the health management of the high-speed train to master the evolution of the performance parameters of the key components.How to estimate the performance parameters and how to diagnose the fault of the key components by vibration signal is an urgent task to be solved.The high-speed train is a high degree of freedom of multi-body system,therefore,the effective use of the vibration signals collected on vehicle to establish an effective and efficient model,choosing a suitable method for the railway system to estimate the performance parameters and to diagnose the fault of key components,finally proposing a observation method with minimal amounts of observed quantity is the purpose of this study.The main work of the thesis is proceeded from the following aspects:(1)Through the analysis of dynamics model of vertical half car body system and the transverse half car body system,the paper chooses the vertical acceleration at front of the car body and the vertical acceleration at front of the frame as observed quantities based on sensors installed on test trains and vehicles on service,then establishes a state space model and a observation model of vertical system.For transverse system,the paper selects the lateral acceleration at front of the car body,the lateral acceleration at front of the frame,and the lateral acceleration of primary wheelset as the observed quantities,and then establishes the state space model and a observation mode.From the two aspects of state space model and observation model,the paper effectively reduces the dimension of the model,greatly reduces the computation and improves the computational efficiency.(2)In order to solve the damping parameter estimation of the EMU in health management,the paper estimates the damping of the primary vertical damper,secondary vertical damper,secondary lateral damper,and the yaw damper.The method to estimate the damping parameter of the damper off line is determined by analyzing and contrasting the performance of the two filters in depth.(3)In allusion to the problem of model error caused by ignoring the nodding and shaking movement in the half car body model.The paper corrects the vertical acceleration and lateral acceleration of car body by analyzing the error between calculated car body acceleration and real value.Then the relative error for estimation is reduced from 30% to below 10%.(4)Fault diagnosis and separation are performed using IMM Kalman filter and classical Kalman filter for the online fault diagnosis of damper.The classical Kalman filter needs to analyze the residual components to determine the location of the fault,in allusion to this issue,the paper provided a improved Kalman filter,which can directly determine the location of the fault.(5)In order to improve the reliability of vehicle and reduce the sensors installed on vehicles,the paper analyzes the influence of damping on the vibration acceleration of vehicle components,and determines the new observation quantities,providing theoretical basis for the improvement of sensor installation and improvement of the state monitoring and fault diagnosis of damper.