A model-based approach for attribute gauge analysis

A new method is proposed for the attribute gauge study. A mathematical model based on the Gauge Performance Curve (GPC) is introduced and the connection between the model parameters and gauge characteristics is discussed. The ability of the whole measurement system (the process and the gauge together) is characterized by two key probabilities: the probability of a rejected part being good and the probability of an accepted part being bad. These probabilities can be calculated either with the proposed mathematical model of GPC (GPC-method) or by using relative frequencies (AIAG-type method). The performance of the two approaches has been compared, and the proposed GPC-method is proven to be statistically favourable.