4.02 Multivariate Statistical Process Control and Process Control, Using Latent Variables

Multivariate monitoring and control schemes based on latent variable methods are receiving increasing attention by industrial practitioners. Several companies have enthusiastically adopted these methods and have reported many success stories. Applications have been reported where multivariate statistical process control (MSPC), fault detection and diagnosis, is achieved by utilizing the latent variable space, for continuous and batch processes as well as for process transitions, for example, start ups and restarts. This work gives an overview of the latest developments in MSPC and its application for fault detection and isolation (FDI) in industrial processes. Recent applications of latent variable methods to process control as well as to image analysis for monitoring and feedback control are discussed.