Dimensional Measurement Uncertainty and Inspection Planning

9622692库尔菲斯现代加工设施可以生产复杂而精确的几何形状的零件，但仍需要高精度的坐标测量机(CMM)来验证这些几何形状。坐标测量机测量零件表面上的点的笛卡尔坐标，这些点使用极小最大或最小二乘等标准与标称零件几何形状相匹配。这些方法产生相互矛盾或不一致的结果，因为分析没有解决测量不确定性问题。这项研究将开发一种方法，以最大化信息含量的方式，将多变量可信区间、某个特征的点测量的最佳位置和点的数量联系起来。一个与工业界合作的实验计划将测试这些在处理CMM数据方面的进步。这项研究对制造业的生产率和质量都有影响。例如，该方法的一个应用是基于时间或成本约束来指定要在CMM上收集的点的数量，然后为测量的特征生成采样计划和置信度区域的估计。或者，如果需要指定的准确度，则该方法可以提供用于进行测量的抽样计划，并指定达到指定的置信度区域所需的点数。

9622692 Kurfess Modern machining facilities can produce parts with complex and precise geometry, but high precision coordinate measuring machines (CMMs) are still needed to validate these geometries. The CMMs measure Cartesian coordinates of points on the part surface, and these points are fit to nominal part geometries using criteria like mini-max or least squares. These methods yield conflicting or inconsistent results because measurement uncertainty has not been addressed by the analysis. This research will develop methods that relate multivariate confidence intervals, the optimal location of point measurements for a feature, and the number of points in a way that maximizes information content. An experimental program, in cooperation with industry, will test these advances in processing CMM data. This research impacts both productivity and quality in manufacture. For example, one application of this methodology is to specify the number of points to be collected on the CMM based on time or cost constraints, and then generate a sampling plan and an estimate of the confidence regions for the measured features. Alternatively, if a specified degree of accuracy is needed, then the methodology can provide a sampling plan for making measurements and specify the number of points needed to achieve a the specified confidence region.