Fault diagnosis in industrial process manufacturing using MSPC

Abstract

The basis of multivariate statistical process control (MSPC) are the multivariate projection techniques of principal components analysis (PCA) and projection to latent structures. This paper focuses upon PCA. The philosophy behind these techniques is to reduce the dimensionality of the problem by forming a new set of latent variables. If the variables are highly correlated, then the process can be defined in terms of a reduced set of latent variables, which are a linear combination of the original variables. Principal component analysis is an analysis tool which reduces the dimensionality of a single data matrix. The principal components generated from the analysis form the cornerstone of the multivariate statistical process control charts.