基于主动学习Kriging模型与序列重要抽样的随机-区间混合可靠性分析

针对随机-区间混合可靠性分析中复杂功能函数的高非线性和多设计点问题,本文提出了一种结合主动学习Kriging模型与序列重要抽样方法的混合可靠性分析方法.在序列重要采样方法中采用高斯混合分布作为提议分布进行逐级采样,逐步逼近最优重要抽样函数的采样样本;结合序列重要抽样方法的特点,提出了主动学习Kriging模型的两步学习...

A combined active learning Kriging model and sequential importance sampling for hybrid reliability analysis with random and interval variables

Aiming at the complicated performance functions with high nonlinearity or multiple design points in random-interval hybrid reliability analysis, a hybrid reliability analysis method combining active learning Kriging model and sequence importance sampling method is proposed in this paper. The sequential importance sampling method is employed to generate the approximate samples of the optimal importance sampling function gradually by using Gaussian mixture distribution as the proposed distribution. Then, combined with the sequential importance sampling method, a two-step active learning scheme of Kriging model is presented, which can improve the efficiency of the proposed method significantly while ensuring the accuracy. Finally, the proposed method is compared with some existing hybrid reliability analysis methods through several numerical examples, to verify the accuracy and efficiency of the proposed method.