中国机械工程 ›› 2014, Vol. 25 ›› Issue (2): 246-250,272.

• 智能制造 • 上一篇    下一篇

基于多目标非耦合优化策略的可靠性稳健优化设计

张瑞军1,2;邱继伟1;王晓伟1,2;董明晓1,2   

  1. 1.山东建筑大学,济南,250101[]
    2.山东省高校机械工程创新技术重点实验室(山东建筑大学),济南,250101
  • 出版日期:2014-01-25 发布日期:2014-02-19
  • 基金资助:
    国家自然科学基金资助项目(51075244);山东省教育厅计划项目(J12LB60)

Reliability-based Robust Optimization Design Based on Strategy of Multi-objective Optimization Uncoupling or Decoupling

Zhang Ruijun1,2;Qiu Jiwei1;Wang Xiaowei1,2;Dong Mingxiao1,2   

  1. 1.Shandong Jianzhu University,Jinan,250101
    2.Key Laboratory of Mechanical Engineering & Innovation Technology in Universities of Shandong(Shandong Jianzhu University),Jinan,250101
  • Online:2014-01-25 Published:2014-02-19
  • Supported by:
    National Natural Science Foundation of China(No. 51075244)

摘要:

以实现多目标问题的可靠性稳健优化设计为目标,通过对目标函数和约束条件进行灵敏度分析,生成目标函数和约束函数的灵敏度附加项,建立了基于灵敏度附加目标函数的可靠性稳健优化设计模型;基于独立公理,使用正交试验和方差分析技术确定设计变量对各个设计目标的影响程度,将设计参数按无耦合或准耦合设计形式分组,把多目标优化问题单目标化,避免多个设计目标之间的反复权衡;结合增广乘子法,应用MATLAB的优化和符号工具箱来实现钳形盘式制动器的可靠性稳健优化设计。算例表明,提出的稳健设计方法具有较高的精度和可靠度。

关键词: 多目标优化, 灵敏度设计, 可靠性稳健设计, 公理设计

Abstract:

Targeting at realizing reliability-based robust optimization design of a multi-objective optimization, the additional items of sensitivity were generated via analysing of objective functions and constraints. And a model of reliability robust-optimization design was proposed based on the additional items of sensitivity. The impact degree of design variables on design objectives was determined by using the orthogonal experiment and variance analysis that were based on the axiomatic design. Design parameters were classified into groups by the uncoupled or decoupled optimization design that can make the optimization problems of multiple objective into single objective and can avoid the trade-off between design objectives. With the augmented multiplier method,reliability-based robust optimization design of caliper disc braker can be proceeded by using the software of symbolic toolbox and optimization toolbox of MATLAB. The numerical example shows that the proposed approach has higher precision and reliability.

Key words: multi-objective optimization, sensitivity design, reliability robust design, axiomatic design

中图分类号: