Efficient Method for Simulating Nonstationary and Non-Gaussian Stochastic Ground Motion with Known Time-Varying Statistical Moments

The identification of a suitable transformation model and the calculation of correlation coefficients constitute essential steps in the simulation a nonstationary and non-Gaussian stochastic process. For stochastic ground motion, its nonstationary and non-Gaussian properties are well-known. In this paper, an easy-to-implement simulation method of a nonstationary and non-Gaussian stochastic process with known time-varying statistical moments is proposed. The approximate transformation model between the nonstationary and non-Gaussian stochastic process and the underlying Gaussian process, which contains the wider range of applicability and the availability of explicit expressions for determining the distribution parameters, is determined. The time-varying correlation coefficient of the underlying Gaussian process is calculated through the proposed interpolation method. Furthermore, the effectiveness of the proposed simulation method of the nonstationary and non-Gaussian stochastic ground motion is verified by numerical examples.