Accounting for Specifications in Statistical Process Control

Abstract

Usually, statistical process control uses control charts for monitoring the evolution of a manufacturing process. Traditionally, control limits are computed without considering another aspect of statistical process control: the process capability (PC). This is essential information because it indicates the propensity of a manufacturing process of respecting the specifications (nominal and tolerances). The main problem when separating the PC from control charts is that the latter are not suitable for monitoring the manufacturing process evolution with respect to the specifications. Thus proposes a new approach for computing control limits: simultaneously accounts for the specifications and the PC, and leads to control charts useful not only for identifying a possible non‐random behaviour of the monitored process but also an erratic behaviour with respect to the specification limits. The main results are: (1) the control limits are no longer established using an interval of 3 standard deviation around a centre line (•); (2) the control limits are no longer restricted to be symmetric; and (3) there is no possibility for the control limits to stand outside the specification limits (which might be the case with the traditional approach).

Citation

Gélinas, R. (1994), "Accounting for Specifications in Statistical Process Control", International Journal of Quality & Reliability Management, Vol. 11 No. 6, pp. 57-64. https://doi.org/10.1108/02656719410064676