An overview of multivariate statistical process control in continuous and batch process performance monitoring

Univariate SPC systems effectively only detect, or provide early warning of, off-specification production, process disturbances and process malfunctions related to individual quality measurement sources. Consequently, they provide little information about the interactions between the variables which are so important in complex processes such as those now found in the process and manufacturing industries. These limitations can be addressed through the application of Multivariate Statistical Process Control (MSPC). The bases of MSPC are the projection techniques of Principal Components Analysis (PCA) and Projection to Latent Structures (PLS). Through the application of PCA or PLS, the process can be defined in terms of a much reduced set of latent variables (a linear combination of the original variables), which reflect the true dimensionality of the process. This paper presents an overview of multivariate statistical process control for continuous and batch processes. The power of the methodology is demonstrated by application to two industrial processes.