Abstract:Dimensionality reduction techniques are widely used in data driven process monitoring methods for extracting key process features from process data. Dimensionality reduction generally leads to information loss, and therefore may reduce the process monitoring performance. However, choosing an appropriate data projection matrix that can minimize the effect of dimensionality reduction on process monitoring performance is often challenging. In this paper, we introduce a method to construct a variable correlation-based sparse projection matrix (VCBSPM) for reducing the dimension of process data and for building the process monitoring model. The VCBSPM is constructed on the basis of variable correlations, with each column of VCBSPM corresponding to a variable group consisting of highly correlated variables. The VCBSPM has two advantages: (1) it implements dimensionality reduction only for highly correlated variables, and therefore the negative effect of dimensionality reduction on process monitoring performance is significantly reduced; (2) the sparsity of VCBSPM not only improves its interpretability, but also enables it to eliminate redundant interferences between variables and to reveal meaningful variable connections. These advantages make the VCBSPM-based monitoring model well suited for fault detection and diagnosis. In addition, to utilize meaningful variable connections revealed by the VCBSPM to improve the fault diagnosis capability, hierarchical contribution plots consisting of group-wise and group variable-wise contribution plots are developed for fault diagnosis. The hierarchical contribution plots can identify both the faulty groups corresponding to actual control loops or physical links in the process and the faulty variables responsible for the fault. The implementation, effectiveness, and key features of the proposed methods are illustrated by an industrial case study.

An Error-in-variable Projection to Latent Structure Framework for Monitoring Technical Systems with Orthogonal Signal Components

Process Monitoring with Sparse Bayesian Model for Industrial Methanol Distillation

Direct projection to latent variable space for fault detection

Dynamic Processes Modeling and Monitoring Based on a Novel Dynamic Latent Variable Model.

Robust Mixture Bayesian Latent Variable Regression with Structural Sparsity and Application to Inferential Sensing of Quality Variables

Model Order Selection In Errors-In-Variables System Identification

Improved Projection to Latent Structures for Quality-Relevant Process Monitoring

Structured Principal Component Analysis Model With Variable Correlation Constraint

Dynamic latent variable regression for inferential sensor modeling and monitoring

Total projection to latent structures for process monitoring

An Extended Closed-loop Subspace Identification Method for Error-in-variables Systems

Deep Learning of Latent Variable Models for Industrial Process Monitoring

Autoregressive Dynamic Latent Variable Models for Process Monitoring

Similarity Based Robust Probability Latent Variable Regression Model and Its Kernel Extension for Process Monitoring

A Variable-Correlation-Based Sparse Modeling Method for Industrial Process Monitoring

Nonlinear Projection to Latent Structures Method and Its Applications

Process Monitoring for Covariance Matrixes with Latent Structures

Quality-Relevant Fault Monitoring Based on Efficient Projection to Latent Structures with Application to Hot Strip Mill Process

Quality-Relevant Fault Monitoring Based on Locally Linear Embedding Orthogonal Projection to Latent Structure

Robust Dynamic Latent Variable Model for Dynamic Process Monitoring with Missing Data in Cyber-Physical Systems

Data-Driven Process Monitoring Based on Modified Orthogonal Projections to Latent Structures.