Markov Chain Monte Carlo-Based Bayesian Identification of Interlayer Stiffness of Ballastless Track

Determining the stiffness of elastic components of ballastless track is critical for assessing the vibration and noise of urban rail transit. Therefore, this study combines Markov Chain Monte Carlo (MCMC)-based Bayesian method with the dynamic track model to identify the stiffness of rail pads and damping mats. First, the effectiveness and anti-noise performance of the framework are verified by identifying the interlayer stiffness of a three-story frame. Second, a finite element model for the dynamic response of the ballastless track is constructed to conduct a sensitivity analysis of the interlayer stiffness on the vertical acceleration of the ballastless track and to investigate the influence of the number of response points and unknown parameters on the identification results of the stiffness of rail pads and damping mats. Finally, the stiffness of the elastic components of the ballastless track is determined through the MCMC-based Bayesian method in a field impact test of the ballastless track. The results demonstrated that the method can identify the unknown stiffness of the elastic components of the ballastless track. The dynamic response of the rail is sensitive to both fastener stiffness and damping mat stiffness. The vertical acceleration of the track slab is more sensitive to the stiffness of damping mats than to that of fasteners. The parameter identification can benefit from the dynamic response close to the impact point, and the identification accuracy depends more on the location of response points than on the number of response points and unknown parameters.