Adaptive Radial Based Sampling for Reliability and Reliability Sensitivity Analysis Involving Truncated Correlated Normal Variables

摘要

The reliability and reliability sensitivity (RS) models are presented for the engineering problem involving truncated correlated normal variables (CNV),and in the case an adaptive radial based sampling is used to analyze the reliability and the RS.In the presented models,the truncated CNV is transformed to general CNV,and the value domains of the truncated CNV are treated as multiple failure modes,then the reliability and the RS with the truncated CNV are transformed to the general cases,on which an efficient radial based sampling is used to analyze the transformed reliability and RS.An adaptive strategy is employed to search for the optimal radial in the sampling,by which the robustness of the method is improved.After the model concepts and the detailed implementation are given,several examples are presented to demonstrate the feasibility of the model and the efficiency of the solutions.