A two-stage Subset Simulation-based approach for calculating the reliability of inelastic structural systems subjected to Gaussian random excitations

Abstract

This paper presents a methodology for calculating the reliability of inelastic structural systems subjected to Gaussian random excitations. The method adopts a two-stage approach, involving separate calculation of the failure probabilities associated with linear elastic and inelastic structural response. The method exploits the fact that the calculation of failure probabilities associated with a linear problem can be performed extremely efficiently, using minimal computational effort compared to the effort required for solving a corresponding nonlinear problem. The calculation of failure probability associated with inelastic response is performed using a modified Subset Simulation procedure where the first step involves the direct simulation of samples in the inelastic domain rather than standard Monte Carlo simulations as in Standard Subset Simulation. It is demonstrated with a numerical example that the proposed two-stage approach offers significant computational savings over the Standard Subset Simulation approach.