关键词: 结构可靠性, 二次插值抽样法, 设计点, 调整参数, 区间变量

Abstract: Aiming at the structural reliability problem in the presence of random and interval variables,an efficient sequence iterative reliability analysis algorithm was proposed.The High Dimensional Model Representation method (HDMR) was used to decouple the reliability model of probability-interval mixture into a single-layer reliability model with random and interval variable separation.Based on the Quadratic Interpolation Sampling method (QIS),unit functions that represented random and interval variables were fitted.In probability analysis,two adjustment parameters were introduced to control the search direction and search step size of optimizing design point respectively.In interval analysis,the projection gradient method was used to calculate interval optimal point,and the failure probability was solved by sequence iteration algorithm.Compared with the Monte Carlo Sampling method (MCS),the relative error of the algorithm was less than 5%,which verified the accuracy and efficiency.In addition,when the degree of nonlinearity of limit state function was high,the algorithm could also ensure the convergence of the calculation results.

Key words: structural reliability, quadratic interpolation sampling, design point, adjustment parameters, interval variables