结构可靠性分析的自适应子集模拟方法

Adaptive subset simulation for structural reliability analysis

结构可靠性分析需要精确高效的失效概率计算方法。为解决高维非线性可靠性分析问题中的失效概率计算问题,本文提出了两种以失效概率估计精度为停机控制参数的自适应子集模拟方法。理论分析和数值算例表明:(1)两种自适应子集模拟方法能根据失效概率的估计精度要求自适应调整样本量;(2)考虑样本量优化的自适应子集模拟方法能进一步减少总样本量,提高计算效率。本文所提方法为研究者对结构进行精确高效的可靠性分析提供了一条可行途径。

Structural reliability analysis requires accurate and efficient failure probability evaluation. Two Adaptive Subset Simulation（ASS） methods were developed in this paper to deal with the high dimensional nonlinear structural reliability problems. Stopping criteria of the proposed methods were based on the estimated coefficient of variation, which was more reasonable. The sample size of the proposed ASS methods could be adjusted adaptively according to the estimated coefficient of variation. The ASS method with optimal sample size was more efficient. The accuracy and efficiency of the two ASS methods were verified by a numerical example.