Reliability Evaluation of Semirigid Steel Frames Using Advanced Analysis

Abstract

In this paper, an efficient method for determining the reliability of semirigid steel frames is presented. An advanced analysis using the beam-column approach and the zero-length semirigid model is used to predict structural nonlinear inelastic behaviors and reduce computational effort. A new probability method, named as LPSS-EIS, is developed by integrating the Latinized partially stratified sampling (LPSS) and an efficient importance sampling (EIS). The robustness of the proposed method is demonstrated through two space semirigid steel frames subjected to strength and serviceability load combinations. Several popular reliability analysis methods, such as the Monte Carlo simulation, LPSS, and subset simulation, are used for comparison. The numerical results prove that LPSS-EIS predicts the structural failure probability accurately, saves the computational time, and works well for both very small and large values of structural failure probability. The effect of semirigid connection properties on structural reliability is also carried out.