Finding design point base on a quasi-importance sampling method in structural reliability analysis

Calculation of reliability index is still a significant challenge in reliability problems. The Monte Carlo simulation technique is beneficial for calculating failure probability, but given the need to generate many random samples, it has a high computational cost. Additionally, it cannot calculate the design point, while it can be important to determine the design point in large-scale practical problems.This paper develops an algorithm for calculating the design point using the modified importance-sampling method. In the proposed method, the standard deviation (SD) of sampling density-function is improved in each step. This enables the proposed algorithm to calculate the reliability index and design point using a small number of random samples. The examples are presented in this paper indicate the high efficiency of the proposed algorithm to calculate the reliability index and design point, even for nonlinear problems with correlated random variables.

Automated summary

This paper presents an algorithm for calculating the design point using the modified importance-sampling method. By improving the standard deviation of the sampling density function, the proposed algorithm is able to calculate the reliability index and design point using a small number of random samples, even for nonlinear problems with correlated random variables.