Study of Structural Reliability Estimation Based on Importance Sampling Method-- Construction of Optimal Importance Sampling Density part 2

publisherThis paper is concerned with the improvement method of the previous report. The estimation method of the structural failure probability with multi-variables based on the conditional expectation procedure is proposed. The proposed method is composed of two stage procedures. In the first stage, The region of each variable is divided analytically into a specified number of finit element meshes and the optimal marginal probability density function generating samples is constructed getting a rough estimate of the failure probabibility, probability function and conditional probability. In the second stage, Failure probability estimates by using the the optimal importance sampling method based on the conditional expectation Numerical examples for the evaluation of structural failure probabilities are presented to show the validity of the proposed method in saving compuatation efforts and in getting low cov compared with a regular importance sampling method and conditional method