Towards active statistical process control

Statistical process control (SPC) research has focused on charting techniques which are employed for process monitoring. Unfortunately, little attention has been paid to the importance of bringing the process in control automatically via these charting techniques. Drawing upon concepts from automatic process control (APC) it is possible to devise schemes whereby the process is monitored and automatically controlled via SPC procedures. This paper describes the formulation of such a method, where partial correlation analysis is used to determine the variables which have to be monitored and manipulated as well as the corresponding control laws. The capabilities of the proposed SPC methodology are demonstrated by application to a simulated reaction process.