基于二参数寻优设计点的混合结构可靠性分析算法

A mixed reliability analysis algorithm based on two parameters is proposed for the mixed reliability problem of mechanical structure with random variables and interval variables. Interval variables make the reliability analysis problem become a double-loop optimization problem. In order to reduce the influence of double-loop optimization model on computational efficiency, it is decoupled into a high-efficiency sequence iterative model for probability and interval analysis. In order to optimize design point, two adjusting parameters are introduced to control the search step length and the search direction, respectively, in probability analysis, which ensures the convergence stability and the search efficiency. The interval analysis problem is transformed into a more solvable quadratic programming problem, and the gradient projection method is used to search for the interval optimum point. The analyzed results of two cases show that the relative error of maximun failure probability calculated by the proposed algorithm is within 5%, and the number of function calls can reach convergence within 1000 times compared with Monte Carlo sampling. When the performance function has a high degree of nonlinearity, the proposed algorithm also has high calculation accuracy, efficiency and convergence stability.