摘要
Conventional reliability-based design optimization （RBDO） requires to use the most probable point （MPP） method for a probabilistic analysis of the reliability constraints. A new approach is presented, called as the minimum error point （MEP） method or the MEP based method, for reliability-based design optimization, whose idea is to minimize the error produced by approximating performance functions. The MEP based method uses the first order Taylor＇s expansion at MEP instead of MPP. Examples demonstrate that the MEP based design optimization can ensure product reliability at the required level, which is very imperative for many important engineering systems. The MEP based reliability design optimization method is feasible and is considered as an alternative for solving reliability design optimization problems. The MEP based method is more robust than the commonly used MPP based method for some irregular performance functions.
Conventional reliability-based design optimization （RBDO） requires to use the most probable point （MPP） method for a probabilistic analysis of the reliability constraints. A new approach is presented, called as the minimum error point （MEP） method or the MEP based method, for reliability-based design optimization, whose idea is to minimize the error produced by approximating performance functions. The MEP based method uses the first order Taylor＇s expansion at MEP instead of MPP. Examples demonstrate that the MEP based design optimization can ensure product reliability at the required level, which is very imperative for many important engineering systems. The MEP based reliability design optimization method is feasible and is considered as an alternative for solving reliability design optimization problems. The MEP based method is more robust than the commonly used MPP based method for some irregular performance functions.
基金
This project is supported by National Natural Science Foundation of China(No.50575072)Outstanding Youth Fund of Hunan Education Department, China (No.04B007).
作者简介
LIU Deshun was born in 1962 in China. He is currently a professor in Hunan Provincial Key Laboratory of Health Maintenance for Mechanical Equipment, Hunan University of Science and Technology, China. He recelved his PhD degree from Central Southern University, China, in 1996. His research interests include probabilistic design, multidiscipline robust design optimization, etc. Till now he has published morn than 100 journal papers and conference papers. Tel： ＋86-732-8290380; E-mail： liudeshun@,hnust.edu.cnYUE Wenhui is currently an associate professor in Hunan Provincial Key Laboratory of Health Maintenance for Mechanical Equipment, Hunan University of Science and Technology, China. And he is a doctoral candidate in China University of Mining and Technology, China. His research interests include multidisclpline robust design optimization, advanced manufacturing technology, etc. Tel： ＋86-732-8290288; E-mail： jwc 1999178@ 126.comZHU Pingyu was born in 1971, associate professor, post-PhD. She received her PhD from Central South University, China, in 2003. Her research interests include measuring theory and technology for mechanical engineering. She has been working in this field for many years and has published more than 20 papers in magazines and international conferences at home and abroad. Tel： ＋86-732-8290090; E-mail： hnkjdxzpy@ 163.comDU Xiaoping is currently an assistant professor in Department of Mechanical and Aerospace Engineering, University of Missouri-Rolla, Rolla, USA. His research interests include multidiscipline robust ＆sign optimization, engineering probabilistic design, etc. Tel： ＋1-573-341-7249; E-mail： dux@umr.edu
