论文部分内容阅读
将均值1阶Esscher近似可靠性分析方法、灵敏度分析技术及稳健设计方法相结合,提出了一种新的可靠性稳健设计方法,即基于均值1阶Esscher近似可靠性的稳健优化设计。可靠性稳健设计方法将质量和灵敏度作为目标,可靠度作为约束,最终归结为一个多目标优化问题。均值1阶Esscher近似可靠性方法适用于基本随机变量是正态或非正态、连续或独立的失效概率的计算,但要求这些随机变量相互独立,有矩母函数,并且极限状态函数具有显示表达式。因此,提出的可靠性稳健设计方法对于含非正态变量的稳健优化设计问题具有较高的计算精度。
Based on the first order Esscher approximate reliability analysis method, the sensitivity analysis technique and the robust design method, a new robust robust design method is proposed, which is a robust optimization design based on the average first order Esscher approximation reliability. Robust design approach to reliability and quality as the goal, reliability as a constraint, and ultimately boils down to a multi-objective optimization problem. The mean-order Esscher approximation reliability method is suitable for the calculation of the failure probability that the basic random variables are normal or non-normal, continuous or independent, but these random variables are required to be independent of each other, have the moment mother function, and the limit state function has the display expression formula. Therefore, the proposed robust robust design method has higher computational accuracy for robust optimization design problems with non-normal variables.