Statistical Process Control

Statistical Process Control is a method of trending and analyzing both on line continuous and off line discretely sampled data for a process.  Metrics are established determining the performance of the system considering the variability that is inherent in the system.  The information presented for decision-making is delivered with the known errors so that the quality of the data can be appropriately applied to the nature of the decision being made on that data.  A common goal is to have all production specifications achieve and maintain a cpk of 1.33.  All measured control variables should only be reported from validated test methods or field instruments by qualified operators and analysts. 

All measured control variables shall be refereed by a laboratory that has greater validated precision than the plant claim and has accuracy tied to an NIST standard.  The referee validation of the plants measurements for each measured variable shall occur at a frequency no less than once per year.  No unit should be permitted to pass on a quality problem downstream for others to fix. 

The Western Electric Run Rules should be applied to each units measured control variables so that corrective action is taken prior to the control variable going out of specification, however action is not taken in response to the natural variability in the system.