Multivariate statistical process control applied to a high speed industrial film production facility

This paper addresses the development and implementation of Multivariate Statistical Process Control to a high speed film production facility. The plant data logging facilities provide a wealth of process data relating to each of the key process units. The data is highly correlated and difficult to monitor using traditional univariate statistical process control methods. Multivariate Statistical Process Control, and in particular the technique of Principal Component Analysis, provides a route to the detection of important process events. Two contrasting case studies are used to highlight the power and versatility of the methodology.