An efficient time-dependent reliability method

Abstract

Abstract Reliability-based methods have been widely utilized for assessing the time-dependent performance of various structural components and systems subjected to random process loadings. The time-dependent reliability is regarded as the probability that the structural random response process does not exceed the specified failure threshold within the forecast time period. In this context, Poisson outcrossing rate methods using the first-order reliability method are widely employed. However, these methods are inaccurate for low boundary reliability problems with dependent outcrossing events. In this paper, a new time-dependent reliability method is presented. Instead of analyzing outcrossing rates, the time-dependent reliability is formulated as a large-scale series system consisting of time-independent response functions obtained by discretizing time-dependent continuous response functions within the forecast time period. Efficient methods developed recently are adopted to address the large system dilemma with respect to design point evaluation and multi-normal integral. The accuracy and efficiency of the proposed method are examined on four numerical examples.