Research On Railway Vehicle Track Spectrum Inversion And Application Based On Adaptive Kalman Filter

Track irregularity is the direct factor of wheel-rail dynamic response,the main excitation source of vibration of railway vehicle system and the important index of track system maintenance.Therefore,obtaining track condition is particularly important for vehicle system design and the maintenance of operation vehicle and track system.In this thesis,combining the current track detection and estimation techniques,an adaptive Kalman filter(AKF)was proposed to realize the track spectrum inversion.The main works are as follows:(1)The related theories of track spectrum inversion technology based on Kalman filter were introduced in detail.Firstly,according to the linear system theory,the state space description of dynamic system was studied,the time discretization method of state space equation and the judgment basis of system observability were derived.Secondly,based on the principle of orthogonal projection and the recursive estimation criterion of linear minimum mean square error,the Kalman filtering idea was explained,and the Kalman filtering algorithm formula was deduced.Finally,the Kalman filter parameter optimization and filter divergence were improved by adaptive algorithm;(2)The inversion method of tract spectrum was developed by using AKF.The results indicated that: compared with the KF,the AKF had faster convergence and higher accuracy,which well solved the optimization problem of filtering parameters.At the same time,with the vehicle system state response as the observation,wheel-rail excitation such as track irregularity and vehicle polygon were estimated by adaptive Kalman filter track spectrum inversion method.And the correlation coefficients between the inversion results and the real values were all above 0.6,which showed a strong correlation.While the inversion results presented amplitude differences in high frequency bands(> 70 Hz);(3)The application of track spectrum inversion method in on-site test was studied.The results indicated that: The inversion estimation of track spectrum can be realized based on the acceleration signal measured in on-site test.The track spectrum inversion results of Guangzhou Metro Line 3 were correlated with the measured irregularity,and the inversion results reflect the track characteristic excitations such as track corrugation and rail joint.In the wavelength range of 1～20 m,the inversion spectral density value of Chengdu-Guiyang highspeed railway was significantly higher than that of Wuhan-Guangzhou spectrum and AAR5.In the wavelength range of less than 0.7 m,the inversion spectral density value was significantly lower than that of Wuhan-Guangzhou spectrum and AAR5,which showed the difference of track excitation of each line.(4)The application of track spectrum in vibration fatigue design of vehicle system was studied preliminarily.The results indicated that: based on the rigid-flexible coupling dynamic model considering the flexibility of the car body,the dynamic weak position of the car body can be identified,such as the connection weld between the traction beam and the sleeper beam,the corner of the door and the window.Taking the inverse spectrum of Chengdu-Guiyang highspeed railway as the track excitation,the actual service life prediction of the car body structure was more accurate than that of the Wuhan-Guangzhou spectrum excitation.The vibration fatigue mileage life of the weld area between the traction beam and the sleeper beam was 6.63 million km,which should to be paid more attention in the real operation.