基于聚类分析的白车身零件工艺模板提取

Process template acquirement for body in white based on clustering analysis

为提高白车身工艺规划的效率,提出了基于聚类分析的白车身工艺模板提取方法。按照白车身装配的典型工艺,将其分为装配定位工序、连接工序和检验工序三类。采用专家打分法计算装配定位工序间的距离。对连接工序进行四位数字编码,在此基础上应用曼哈坦距离计算工序间的距离。为了提高距离计算的正确性和科学性,采用二值属性间距离公式求解检验工序间的距离。采用层次聚类法对白车身工艺聚类,用均方差法从聚类结果中提取了每个簇的工艺模板。通过实例验证了该方法能有效提取白车身零件的工艺模板。

To improve the efficiency of process planning for Body In White （BIW）, a methodology of process templatefor BIW parts based on clustering analysis was proposed. The process was classified into assembling position working procedure, joint working procedure and inspection working procedure according to the typical assembling sequence for BIW. The distance between assembling positions was calculated by expert scoring method. Joint procedures were coded by 4 digits and the distance was measured by Manhattan distance. To improve the efficiency andaccuracy of calculation, a formula solving the binary attribute distance was applied to calculate the inspection distance. Process clusters were eventually merged by the agglomerative hierarchical clustering method. Process template with the target of minimum variance between process distances was extracted from every cluster. A case studywas used to show the effectiveness of this methodology.