On inverse identification method of bearing load based on railway vehicle acceleration

The bearings, as key parts of the drive system of high-speed trains, directly affect the operation safety of high-speed train once the fault occurs. Accurately obtaining the dynamic load of the bearing during the high-speed train operation can provide real boundary input conditions for the early design, fatigue life and service performance of the bearing, which has important engineering practical value. In this paper, 18-DOF vertical and 21-DOF lateral vehicle dynamics models were established considering Maxwell shock absorber model and axle box rotary arm, and then the state-space equation of the vehicle system was determined. Aninverse method for identifying bearing load from vibration acceleration of high-speed train vehicle was proposed by using the Kalman filter algorithm. Finally, the vehicle dynamics model was established by using the SIMPACK® software program, and the method was verified by applying typical random excitation of track irregularity and periodic high-frequency excitation of wheel polygon, respectively. The results show that the variation trend and amplitude of the vertical bearing load identified by the inverse method were consistent with simulation results in time-frequency domain. Under the random excitation of track irregularity, the vertical correlation coefficient of each axle box bearing is greater than 0.7, and the lateral correlation coefficient is greater than 0.6. Under high-frequency excitation of polygon wheel, the vertical correlation coefficient of bearing is 0.97, and the lateral correlation coefficient is 0.96. Under the combined action of track irregularity and wheel polygon excitation, the vertical correlation coefficient of bearing is 0.79 and the lateral correlation coefficient is 0.77. To sum up, theinverse method for identifying bearing load takes into account the movement of the axle box bearing itself based on the traditional vehicle model. It has high inversion accuracy and line adaptability, and can obtain accurate bearing load to provide data basis for studying the remaining service life of the bearings and their structural design and optimization. © 2023, Central South University Press. All rights reserved.