Research On Load Identification Of Axle Box Bearing Of High-speed Train

With China’s continuous improvement of high-speed train’s speed,increasing operating mileage,it is essential to ensure the safe and stable operation of trains.Axle box bearings are key components in the transmission system of high-speed train,bearing failures will directly affect safe operation.Accurately obtaining the load of the axle box bearing under the operating of the train can provide simulation and experimental data for studying the failure evolution and remaining life of the axle box bearing,it can even carry out condition monitoring and fault diagnosis according to the bearing load.However,due to the special assembly form of the axle box bearing,it is impossible to directly measure the axle box bearing through the test,so an indirect method—load identification is needed,and the external load on the axle box bearing is reversely calculated by measuring the vibration acceleration of the axle box.The main work of this thesis is as follows:(1)Combined with the Newmark-β integration method,the system motion differential equation is transformed into the state space equation form,and the load identification forward equation is established.On this basis,the ill-posed problems in load identification are analyzed,and use the truncated singular value decomposition(TSVD)and Tikhonov regularization to deal with the ill-conditioned characteristics of inverse problems,and established a load identification method based on state space equation.By single degree of freedom system and multiple degrees of freedom systems(MIMO)numerical calculating,compare the anti-noise performance of different methods,the truncated singular value decomposition(TSVD)performs better than the Tikhonov regularization on anti-noise performance.(2)Based on regularization method,determine the objective function of the error equation,combine the state space equation and dynamic programming to discretize the continuous timedomain process,and divide the optimal decision of the whole process into multiple subdecisions problem.Based on the principle of optimal Bellman,establish a recursive formula for optimal identification load,optimize the solution processing,computational efficiency,while introducing wavelet noise reduction method effectively improves the anti-noise performance of the method.By numerically calculating the load identification of single degree of freedom system and multi degrees of freedom system(MIMO),the load identification method based on dynamic programming has achieved a good identification effect and is suitable for the solution of complex systems with multiple degrees of freedom.(3)Combined with the vehicle system dynamics and the load identification method based on dynamic programming,the load identification of the high-speed train’s axle box bearing was carried out.Considering the limitation of measuring data,the axle box vibration acceleration data and related loads which is used to verify the load identification results are obtained through SIMPACK simulation calculation.On the basis of vehicle system dynamics model,the load identification model of rail vehicle system is established by dynamic programming and regularization method.The axle box vibration acceleration is used as the input of the load identification model to calculate the vertical and lateral loads of the wheel and rail.Then,according to the force analysis of the wheelset and the load of the bearing,the load of axle box bearing is calculated.Comparing the identification results with the simulation results,the vertical load correlation coefficient of the wheel and rail is greater than 0.88,which has a very strong correlation;the horizontal load correlation coefficient of the wheel and rail is between0.6-0.8,which has a strong correlation;the correlation coefficient of lateral and vertical load of the axle box bearing is between 0.4-0.6,which has a medium correlation.