A reduced polynomial chaos expansion model for stochastic analysis of a moving load on beam system with non-Gaussian parameters

Abstract

Stochastic analysis of a load moving on a beam is conducted in which the Spectral Stochastic Finite Element Method (SSFEM) is adopted to simulation the uncertainties in the system parameters and excitation forces. Since the dimension of Polynomial Chaos Expansion (PCE) of the beam responses will exponentially grow with the number of K-L components of the system and excitation uncertainties, this limits the application of the SSFEM. A reduced PCE model is proposed in this paper to improve the computational efficiency. The non-Gaussian random variables in the Karhunen-Loéve Expansion (KLE) of the non-Gaussian system parameters are assumed “uncoupled”. Numerical simulations show that the computational effort can significantly be reduced while accurate predictions on the response statistics can still be achieved. Studies on the effect of different level of randomness in the system parameters and excitation forces show that the proposed method has good performance even with a high level of uncertainty.