A method to enhance the quality and reliability of inspection data for precision manufacturing

Dimensional inspections are commonly used to scrutinize the quality of manufactured products against established standards and specifications. In order to effectively conduct a dimensional inspection in a computer-integrated manufacturing (CIM) environment, first one needs to derive a sampling strategy that can be used to specify a set of measuring points. The designed sampling strategy must lead to an accurate sampling while minimizing the sampling time and cost. To further enhance the quality and reliability of any inspection process, potential measurement errors need to be addressed for all data acquisition activities. One of the prominent sources for measurement errors is due to the imperfection of a measuring device and the compound effect of its imperfection with geometric characteristics of a measured feature. Accordingly, a method is needed to identify and decouple the effect of compound errors. If this can be done, then the collected data can be properly adjusted and a more meaningful analysis can result.; In this research, first a feature-based sampling strategy using the Hammersley sequence is developed for the various geometric features. Then, an analytical method is presented to discuss the issue of error identification and reduction in machine calibration and dimension measurement. The developed method is applicable for various artifacts with both elementary and complex geometric features. The measurement method and error classification for both the machine-related error and workpiece geometry-related error are first discussed. Generalized models are then derived to identify and decouple the compound effect of both types of measurement errors. Analytical models are demonstrated on 2-D, 3-D and free-form geometrical features. As the compound effect is decoupled, the accuracy and reliability of the measured data can be significantly improved. The derived models are validated for common measuring devices and measurement procedures that are adopted by industry. Finally, the effectiveness and the limitation of the derived sampling strategy, the measurement error models, and the measurement error reduction algorithms are also discussed.