基于支持向量机回归的机械零件可靠性分析  

Reliability Analysis of Mechanical Parts Based on Support Vector Regression

作　　者：程跃[1] 郑严[2] 

机构地区：[1]成都大学工业制造学院,四川成都610106 [2]西南交通大学机械工程研究所,四川成都610031

出　　处：《成都大学学报（自然科学版）》2012年第1期73-76,共4页Journal of Chengdu University（Natural Science Edition）

摘　　要：针对机械零件中隐式极限状态方程的可靠性分析,提出了一种基于支持向量机回归的机械零件可靠性分析方法.将支持向量机回归作为隐式极限状态方程的重构工具,用训练后的支持向量机模型替代隐式极限状态方程,结合改进的一次二阶矩法,给出了基于支持向量机回归的机械零件可靠性分析流程,并用2个算例验证了该方法的可行性、正确性和计算精度.结果表明,该方法能够正确有效地解决机械零件可靠性分析问题,具有一定的工程实用价值.For reliability analysis of implicit limit state function in mechanical parts, a method for reliability analysis of mechanical parts was proposed on the basis of support vector regression. Support vector regres- sion was used as a tool for reconstruction of implicit limit state function, and Support Vector Machine model after training was used to replace implicit limit state function. Combining the advanced first order second moment method, a computation flow of reliability analysis for mechanical parts based on support vector re- gression was given. Finally, two examples of testing were used to verify the feasibility, correctness and accu- racy of this method. The results show that this method can solve the reliability analysis problem of mechani- cal parts rightly and efficiently, which has certain practical engineering value.

关 键 词：支持向量机回归 机械零件 可靠性分析 隐式极限状态方程 

分 类 号：TB114.3[理学—概率论与数理统计]