Guidelines for Hemispherical Feature Inspection Using Coordinate Measuring Machines

Abstract

To increase efficiency and effectiveness of form inspection using probe-typecoordinate measuring machines (CMMs) on hemispherical feature, a data collection problemconsisting of sample sizes, sample point locations, and sampling sequences was investigatedin this work. The accuracy and probe path length of sphericity measurement were studiedcollectively to reduce measuring time. Three types of sampling strategies and six differentsample sizes were addressed in conjunction with the determination of an optimal sample pointsequence. A geometrical procedure for three dimensional arc length calculation was proposedto find distances between the sampled points along the curved surface. The concept of thetraveling salesman problem (TSP) was then applied next. A relatively new discreteoptimization algorithm, ant colony system (ACS), was easily adopted to find an optimal pathsequence due to its appealing properties for solving TSP problems, such as near-optimalsolutions, consistent results, and ability to handle large sample sizes. The obtained sequencewas then employed by a CMM to actually measure the manufactured hemispherical surface.Influence of accuracy and path length was combined by using a priority coefficient to selectan effective and efficient sampling strategy. Preliminary observations revealed that theHammersley sampling method was very attractive in most test cases.