Reliability Estimation with Time-Variant Loads and Resistances

Abstract

This paper presents an efficient simulation-based technique to estimate the system reliability of structures that experience variation in loads and resistance over time. To arrive at a realistic reliability estimate of such structures, random process description of load variation and consideration of resistance degradation and periodic repair must be included. An adaptive importance sampling methodology for estimating the time-dependent system reliability of brittle structures is presented in this paper. In the proposed method, a simple enumeration scheme is used to identify one complete failure sequence of the structural system. This defines the initial failure domain to start the simulation. As the simulation progresses, the initial failure domain is refined adaptively to include other significant failure sequences and to accurately estimate the failure probability of the system. The proposed method removes the idealized simplifying assumptions of earlier analytical methods, facilitates easy addition of the reliability analysis module to structural analysis, and provides a fast and accurate estimate of the failure probability. Numerical examples presented to illustrate the application of the proposed method demonstrate its efficiency and promise.