基于神经网络的机械零部件可靠性稳健设计  

Reliability-Based Robust Design for Mechanical Components with Neural Network

作　　者：张义民[1] 高娓[1] 贺向东[1] 宋相强[1] 黄贤振[1] 

机构地区：[1]东北大学,沈阳110004

出　　处：《应用力学学报》2009年第1期172-175,共4页Chinese Journal of Applied Mechanics

基　　金：国家高技术研究发展计划(2007AA04Z442);国家自然科学基金重点项目(50535010);高等学校创新团队项目

摘　　要：将可靠性优化设计理论、可靠性灵敏度技术和可靠性稳健设计方法相结合,利用神经网络的非线性映射功能,模拟得到随机设计变量与机械零部件结构响应之间的关系,以训练好的神经网络模型为基础,结合随机摄动技术,进行结构可靠性稳健设计。从而解决了工程实际中很难给出极限状态函数的显性表达式问题,并有效地克服了目前随机有限元法和响应面法存在的缺陷。Combining the theory of reliability-based optimization,reliability-based sensitivity and reliability-based robust design,the explicit expression between the design parameters and the system reliability is given correctly by NN with the approximation concepts inherent in reliability analysis.Following the trained NN,the probabilistic perturbation method is chosen to obtain the reliability-based robust design information.Therefore,the explicit expression of the limit state function which gets very difficult to be found in engineering,is proposed,and the limitations in finite element method and response surface method are made up.

关 键 词：可靠性优化设计 可靠性稳健设计 神经网络 随机摄动技术 

分 类 号：TH122[机械工程—机械设计及理论]