基于概率支持向量机的可靠性分析与设计方法

发布时间：2018-05-07 03:27

本文选题：结构可靠性分析 + 优化设计　；参考：《应用力学学报》2017年01期

【摘要】：结构可靠性设计是现代结构设计中能够考虑不确定性因素影响的重要方法之一。针对结构可靠性设计中较为复杂的隐式多概率约束条件问题,本文提出了一种有效的代理模型方法。该方法基于多输入多输出支持向量机方法,并引入分类错误概念,可以同时对多个概率约束条件进行建模,同时对概率约束条件进行了相对保守的结构可靠性分析,得出非零失效概率的两者差值在1%~10%。将该可靠性评估方法嵌入可靠性设计优化框架内,形成了偏于安全的结构可靠性设计方法。文中还提出了一种新偏导数计算方法,用于可靠性优化设计。最后将所提方法应用于减速器可靠性分析和优化设计中,验证了其计算精度和效率。
[Abstract]:Structural reliability design is one of the important methods to consider the influence of uncertainty factors in modern structural design. In this paper, an efficient agent model method is proposed to solve the complex implicit multi-probability constraint problem in structural reliability design. This method is based on multi-input and multi-output support vector machine and introduces the concept of classification error. It can model multiple probabilistic constraints at the same time and make a relatively conservative structural reliability analysis of probabilistic constraints. The difference between non-zero failure probability and non-zero failure probability is 1 / 10. The reliability evaluation method is embedded in the framework of reliability design optimization, and a structural reliability design method is formed. A new method for calculating partial derivative is also presented, which is used in reliability optimization design. Finally, the proposed method is applied to the reliability analysis and optimization design of the reducer, and its calculation accuracy and efficiency are verified.
【作者单位】：南京航空航天大学飞行器先进设计技术国防重点实验室;
【基金】：国家自然科学基金民航联合研究基金(U1533109) 中央高校基本科研业务费(NS2015007)
【分类号】：TB114.3;TP181

【参考文献】