| dc.description.abstract | Inspection of parallelism between two surfaces (planes) requires the establishment of a datum feature (surface), and determination of the envelope of points containing the inspected feature parallel to that datum feature. Verifying parallelism thus requires the establishment of the datum feature, and determining the separating distance between the two parallel planes. The parallelism determination can change with sampling (number of data points) and fitting of points. The datum feature establishment is very important to the inspection, and is also dependent on the number of points used to define it. In this thesis, the datum feature is least squares fit from the data collected using coordinate metrology. The inspected feature is then enveloped by minimum separation planes that contain the maximum deviation between the points. The separating distance between the enveloping planes is calculated, and termed the Parallelism Tolerance.
Three levels each, of two sets of data representing the datum feature and inspected feature are collected, for 15 aluminum plates (3 parallelism geometries, 5 replicates). The independent factors are analyzed against the calculated parallelism values. Experimental analysis shows significant effect of sample size on the parallelism computed. As would be expected, the best parallelism values were obtained at the combination of the highest levels of sampling points for datum feature establishment and inspected feature verification. | en_US |
