车身小样本检测数据多异常值判别方法EI北大核心CSCD

针对车身检测数据样本量较小、某时域范围或某类型的测量数据呈现非正态分布的特点,提出了一种基于距离测度的车身检测数据多异常值排序检出算法,并推导了区分由生产过程突变造成的异常值和由测量粗大误差造成的异常值的判断方法。以某车型的间隙与面差检测数据为例,验证了本方法的有效性。     

Study on Dimension Chain Generation for Auto-body Tolerance Analysis

Introduction The construction of assembly function is oneof the keys to evaluate assembly dimensional quali-ty in the process of computer-aided product processconcurrent design. The assembly function mustrepresent the relationship among parts being as-sem…