A Study on Part-Number Selection for Gage R&R in Automated Measurement Systems: A Monte Carlo Simulation-based Analysis of Decision Risks

Purpose: The AIAG MSA manual presents Gage R&R study designs using multiple parts and repeated measurements, with 10 parts commonly used in its examples. However, securing sufficient parts is often impractical for automated measurement systems (Type 3), particularly in high-value manufacturing. This study investigates how reducing the number of parts while maintaining a constant total number of measurements affects MSA decision risk.Methods: Monte Carlo simulations were performed for automated measurement systems without operator effects. The total number of observations was fixed at 30 (number of parts x repetitions = 30), while the number of parts was varied from 10 to 5 and 3. Three Gage R&R levels (5%, 10%, and 15%) were simulated, and the pass rates of %Study Variation (%SV), %Tolerance (%TOL), number of distinct categories (ndc), and the final MSA decision were evaluated over 10,000 simulation runs.Results: Reducing the number of parts consistently decreased the overall MSA pass rate even when the total amount of measurement data remained constant. For acceptable measurement systems (5% Gage R&R), the pass rate decreased from 98% to 78%, indicating an increase in producer’s risk. Conversely, for unacceptable systems (15% Gage R&R), the false acceptance rate decreased from 1.0% to 0.2%, indicating a reduction in consumer’s risk.Conclusion: Reducing the number of parts does not simply increase MSA decision risk but redistributes it by increasing producer’s risk while reducing consumer’s risk, resulting in a more conservative decision. Therefore, the number of parts in Type 3 Gage R&R should be selected by considering the trade-off between practical constraints and decision risks rather than the total number of measurements alone.