基于数论网格法(NT-net)的二维多工位装配成功率计算方法与实例  

Evaluation Method and Case of Assembly yield in Two-dimension Multi-station Assembly Process Based on Number-Theoretical Net

作　　者：文泽军[1] 朱正强[1] 张小平[1] 赵延明[1] 余以道[1] 

机构地区：[1]湖南科技大学机械设备健康维护湖南省重点实验室,湖南湘潭411201

出　　处：《机械设计与研究》2012年第5期94-97,共4页Machine Design And Research

基　　金：国家自然科学基金(51075141);国家"十二五"科技支撑计划(2012BAF02B01)资助项目;湖南省产学研结合技术创新工程计划资助项目(2010XK6066);湖南省自然科学省市联合基金重点项目(11JJ8005);湖南省高校科技创新团队支持计划资助项目

摘　　要：提出一种基于数论网格法(NT-net)的二维多工位装配成功率计算方法。首先分析了数论网格法的偏差,描述了数论网格法产生glp集的原理。然后应用数论网格法对夹具定位偏差进行采样,将其代入二维多工位装配尺寸偏差传递状态空间模型中,计算得到二维多工位装配输出偏差。将计算结果和零件上测点允许偏差进行比较,统计出合格样本数,再将其除以总样本数,计算得到二维多工位装配成功率。最后,以车身地板装配为实例,在构建了车身地板二维三工位装配偏差传递模型基础上,求解其偏差传递矩阵,再运用数论网格法计算其装配成功率,并采用蒙特卡洛模拟法进行了验证和分析,结果表明了该方法的有效性。这为产品二维多工位装配成功率的预测提供了一种新的途径。A calculating method of assembly yield in two-dimension multi-station assembly process is developed based on Number-Theoretical Net（NT-net）.The discrepancy of NT-net is analyzed,and the principle of generating good lattice point（glp） based on NT-net method is introduced.Then,taking fixture locating variations sampled by NT-net method for input vectors,the samples are substituted into state space model of dimension variation propagation in multi-station assembly process to get output vectors.The statistics for qualified samples are accomplished,after comparing output vectors with the variations of measuring points on component.Assembly yield in two-dimension multi-station assembly process is gained when qualified sample divided by total sample.Finally,a real case in automotive body floor assembly is given as an example to calculate the assembly yield in two-dimension three-station assembly processes.The result is validated by using Monte carlo simulation.It provides a new way to predict assembly yield in two-dimension multi-station assembly processes.

关 键 词：数论网格法 多工位装配 装配成功率 蒙特卡洛模拟法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]