Model-plant mismatch detection for cross-directional processes. (November 2021)

• Record Type: Journal Article
• Authors: Lu, Qiugang; Forbes, Michael G.; Loewen, Philip D.; Backström, Johan U.; Dumont, Guy A.; Gopaluni, R. Bhushan
• Is Part Of: ISA transactions. Volume 117(2021)
• Page Start: 150
• Page End: 159
• Publication Date: 2021-11
• Subjects: Routine closed-loop identification -- Support vector machine -- High-order ARX -- Cross-directional processes -- Process monitoring; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2021.01.051
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: This paper presents a two-component framework to detect model-plant mismatch (MPM) in cross-directional (CD) processes on paper machines under model-predictive control. First, routine operating data is used for system identification in closed loop; second, a one-class support vector machine (SVM) is trained to predict MPM. The iterative identification method alternates between identifying the finite impulse response coefficients of the spatial and temporal models. It converges, and the parameter estimates are asymptotically consistent. Coefficient estimates drawn from normal operation are used to train a one-class SVM, which then detects model-plant mismatch in subsequent routine operation. This approach applies to routine operating data without requiring external excitations. It can also distinguish mismatches in the process model from changes in the noise model. Examples of CD processes on paper machines are provided to verify the effectiveness of both components. Highlights: Routine operating data can be used for detecting model-plant mismatch (MPM). Spatial and temporal models can be identified in closed-loop for CD processes. One-class SVM can detect the MPM in spatial and temporal models of CD processes. Noise changes can be separated from MPM by the one-class SVM. Paper machine control verifies the effectiveness of the proposed scheme.