This paper aims to investigate uncertainties in railway vehicle suspension components and the implement of uncertainty quantification methods in railway vehicle dynamics. The sampling-based method represented by Latin Hypercube Sampling (LHS) and generalized polynomial chaos approaches including the stochastic Galerkin and Collocation methods (SGM and SCM) are employed to analyze the propagation of uncertainties from the parameters input in a vehicle-track mathematical model to the results of running dynamics. In order to illustrate the performance qualities of SGM, SCM and LHS, a stochastic wheel model with uncertainties of the stiffness and damping is firstly formulated to study the vertical displacement of wheel. Numerical results show that SCM, which can be easily implemented by means of the existing deterministic model, has explicit advantages over SGM and LHS in terms of the efficiency and accuracy. Furthermore, a simplified stochastic bogie model with three random suspension parameters is also established by means of SCM and LHS to analyze the critical speed, which is affected obviously by the parametric uncertainties. Finally, a stochastic vertical vehicle-track coupled model with parametric uncertainties is built comprehensively on the basis of SCM, by which the impact behavior of wheel-rail interaction under a rail defect is investigated and the dynamic response of vehicles under the track irregularity is explored in terms of the Sperling index. It concludes that the uncertainties of parameters have a significant influence on P2 force and Sperling index from the view of the running quality.
Application and Comparison of Uncertainty Quantification Methods for Railway Vehicle Dynamics with Random Mechanical Parameters
