• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报

• 制造科学与技术 • 上一篇    下一篇

一种新的矩阵公差模型约束条件建立和优化方法

陈华1,  林观生2,  唐涛2,  杨嘉2,  金隼1   

  1. 1. 上海交通大学上海市复杂薄板结构数字化制造重点实验室  上海  200240;
    2.上汽通用五菱汽车股份有限公司  柳州  545007
  • 收稿日期:2015-01-13 修回日期:2015-08-13 出版日期:2016-01-05 发布日期:2016-01-05
  • 通讯作者: 金隼,男,1973年出生,博士,研究员,博士研究生导师。主要研究方向为数字化装配质量控制。 E-mail:jinsun@sjtu.edu.cn
  • 作者简介:陈华,男,1981年出生,博士研究生。主要研究方向为三维公差分析技术。 E-mail:chenhua@sjtu.edu.cn
  • 基金资助:
    国家自然科学基金(51121063,51175340)和国家科技重大专项(2014ZX04002011-3)资助项目

A New Approach of Constraints Establishment and Optimization for Matrix Tolerance Model

CHEN Hua1,  LIN Guansheng2,  TANG Tao2,  YANG Jia2,  JIN Sun1   

  1. 1. Shanghai Key Laboratory of Digital Manufacture for Thin-walled Structures, Shanghai Jiao Tong University, Shanghai 200240;
    2. SAIC-GM Wuling Automobile Co., Ltd., Liuzhou 545007
  • Received:2015-01-13 Revised:2015-08-13 Online:2016-01-05 Published:2016-01-05

摘要: 传统的矩阵公差模型依据特征顶点建立约束条件,不同位置和数量的顶点会导致计算精度和效率的不同。同时,传统算法难以应对存在大量非线性约束条件的优化问题。为此将一般性特征变动约束方程引入到矩阵公差模型中,通过蒙特卡洛仿真方法将符合约束条件的随机数代入目标函数进行优化计算,提高了模型的计算效率和可靠性,降低了优化难度。对比分析一个简单的装配案例,说明其有效性和实用性。

关键词: 公差, 矩阵模型, 蒙特卡洛, 优化, 约束

Abstract: The constraints of traditional matrix tolerance model are constructed by virtue of vertexes of features. Differences of locations and numbers of these vertexes may result in different computational accuracy and efficiency. Meanwhile, traditional algorithm of optimization is incapable of dealing with the situations where a large number of nonlinear inequalities exist. Therefore, the general variations and constraints of features are brought into the matrix model. Monte Carlo simulation is used to take these random numbers satisfying the constraints into computation process. This approach enhances the reliability and efficiency of the matrix model, and reduces its difficulty of the optimization. Comparative results of a simple assemble show the practicability and effectiveness of this approach.

Key words: constraints, matrix model, Monte Carlo, optimization, tolerance

中图分类号: