On-line Batch Process Monitoring Using SVM-Based MPLS

Abstract

This paper integrates support vector machines (SVM) into multiway partial least squares (MPLS) resulting in a nonlinear MPLS model and the model is developed for on-line fault detection in batch processes. Process data from normal historical batches are used to develop the MPLS model, and a series of single-input-single-output SVM networks are adopted to approximate nonlinear inner relationship between input and output variables. In addition, the application of a time-lagged window technique not only makes the complementarities of unmeasured data of the monitored batch unnecessary, but also significantly reduces the computation and storage requirements in comparison with the traditional MPLS. The proposed approach is able to used for real-time process monitoring and fault diagnosis via T2-chart, SPE-chart, and contribution plot for a fed-batch penicillin production.