装配成功率可控、面向成本优化的尺寸链容差设计  

Research on A Tolerance Design Method of Dimensional Chain for Controllable Assembly and Cost Reduction

作　　者：杨宇龙 韩彪 洪巾钧 杨洪 刘衎 潘絮微 魏冕 YANG Yulong;HAN Biao;HONG Jinjun;YANG Hong;LIU Kan;PAN Xuwei;WEI Mian(Information Digitization Room,Lingyun Tech Group Co.,Ltd,Wuhan 430040,China;1th Mechanical and Electrical Workshop,Lingyun Tech Group Co.,Ltd,Dangyang Hubei 444100,China;Spare Parts Division,Lingyun Tech Group Co.,Ltd,Dangyang Hubei 444100,China)

机构地区：[1]凌云科技集团有限责任公司信息数字化室,武汉430040 [2]凌云科技集团有限责任公司机电一车间,湖北当阳444100 [3]凌云科技集团有限责任公司零备件事业部,湖北当阳444100

出　　处：《机械设计与研究》2024年第4期25-30,共6页Machine Design And Research

摘　　要：采用蒙特卡罗法求解、遗传算法优化来进行尺寸链容差设计,设计目标是在装配成功率真实可控的前提条件下尽量优化机械加工的成本。通过深入分析,装配成功率经成本优化后会极限接近合格线,这样会导致两个问题:(1)经重复多次采样,装配成功率的值分布在及格线上下,其中大量不合格;(2)遗传算法每代最优个体无法长久保存,导致优化效率低下。文中采用二次采样蒙特卡罗法将装配成功率的采样波动纳入考虑,有效解决了上述两个问题,得到的结果更符合设计目标的要求。This article presents a tolerance design method for dimensional chains based on the Monte Carlo method and the genetic algorithm to achieve controllable assembly and cost reduction.It is found that the assembly success rate will approach the qualified limit to the extreme after cost reduction,which can lead to two problems:(1)after repeated sampling,the value distribution of the assembly success rate is around the qualified limit,with a large number of unqualified values;(2)the optimal individual of each generation in the genetic algorithm cannot be saved for a long time,resulting in low optimization efficiency.This article adopts the second sampling Monte Carlo method to take into account the sampling fluctuation of assembly success rate,effectively solving the above two problems.

关 键 词：容差设计 遗传算法 蒙特卡罗法 尺寸链 装配 成本优化 

分 类 号：TG801[金属学及工艺—公差测量技术]