Simulation techniques of non-gaussian random loadings in structural reliability analysis

The only currently available universal method for accurately solving problems in stochastic mechanics is Monte Carlo simulation. In situations where the system response is a function of loadings that deviate significantly from Gaussian, techniques for the accurate simulation of these loadings must be available in order to apply Monte Carlo simulation. This paper presents a novel approach for the simulation of non-Gaussian true-random processes. Firstly, a new algorithm for generating non-Gaussian and quasi-random loadings by inverse fast Fourier transform (IFFT) and second phase modulation (SPM) is proposed. Secondly, the effects of time domain randomization (TDR) on the kurtosis and skewness of quasi-random processes are studied, thus the analytical formulas about the relationships between input and output kurtosis and skewness values are obtainded. Finally, a general numerical procedure for the simulation of non-Gaussian true-random processes with specified PSD, skewness and kurtosis based on IFFT, SPM, TDR and iterative correction has been developed. Several simulation examples demonstrate the efficiency and accuracy of the proposed simulation technique. This toot will aid in the dynamic response and fatigue reliability analysis of structural systems exposed to non-normal loadings.