Statistical Process Control Based on Optimum Gages

Classical statistical process control (SPC) by attributes is based on counts of nonconformities. However, process quality has greatly improved with respect to past decades, and the vast majority of samples taken from high-quality processes do not exhibit defective units. Therefore, control charts by variables are the standard monitoring scheme employed. However, it is still possible to design an effective SPC scheme by attributes for such processes if the sample units are classified into categories such as 'large', 'normal', or 'small' according to limits that are different from the specification limits. Units classified as 'large' or 'small' will most likely still be conforming (within the specifications), but such a classification allows monitoring the process with attributes charts. In the case of dimensional quality characteristics, gages can be built for this purpose, making inspection quick and easy and reducing the risk of errors. We propose such a control chart, optimize it, compare its performance with the traditional (X) over bar and S charts and with another chart in the literature that is also based in classifying observations of continuous variables through gaging, and present a brief sensitivity analysis of its performance. The new chart is shown to be competitive with the use of (X) over bar -S charts, with the operational advantage of simpler, faster, and less costly inspection. Copyright (C) 2017 John Wiley & Sons, Ltd.