SFEM-based seismic risk analysis of nonlinear structures using sequential RSM

Abstract

A finite element-based reliability analysis approach is proposed to estimate the risk of structures under short duration dynamic loadings including seismic loading in the time domain. The proposed approach is parallel to the deterministic finite element method, except that it can incorporate the information on the uncertainty in the essential variables present in the problem under consideration. In the algorithm, the incorporation of uncertainties of all the seismic loading and resistance-related parameters is successfully accomplished by using the concepts of response surface method, the first-order reliability method, and an iterative linear interpolation scheme. Therefore, the rational integration of these methods with a finite element formulation leads to the stochastic finite element-based algorithm. It is capable of capable of capturing any special features that can be handled by the finite element method, making it a robust reliability assessment technique. One of its distinguishing features is that actual earthquake loading time histories can be used to excite structures so that the realistic loading conditions can be simulated. Furthermore, it has the potential to estimate the risk associated with both the serviceability and the strength limit states. The algorithm has been extensively verified using the Monte Carlo simulation technique. Although two examples are given for the steel frame building structure in this paper, applicability of the proposed approach can be easily extended to other types of structures.