Abstract:Data-driven process monitoring based on latent variable models are widely employed in industry. This article proposes a novel monitoring framework for latent variable models using hierarchical feature extraction, Bayesian inference, and weighting strategy. We first establish a deep structure to implement hierarchical latent variables extraction, the extracted features are used to construct diverse monitoring statistics. Then, we utilize Bayesian inference and proper weighting strategy to fuse various useful information. In line with the different characteristics of principal component analysis (PCA) and independent component analysis (ICA), we construct a deep PCA-ICA model for process monitoring according to the proposed framework. The deep PCA-ICA model performs hierarchical feature extraction, which can simultaneously extract deep Gaussian information and deep non-Gaussian information. The features extracted by different layers are then transformed to posterior probabilities through Bayesian inference. After that, different posterior probabilities are combined through appropriate weighting strategy to build new probabilistic statistic, which can give more synthetic monitoring results. Moreover, the Bayesian inference and weighting strategy are further used to integrate the advantages of different models by transforming various probabilistic statistics into an overall monitoring index, which can comprehensively indicate the process status. The Tennessee Eastman process is used to validate the superiority of the proposed model over the existing methods. Besides, the extracted features are further analyzed to show the effectiveness and benefits of the deep hierarchical feature extraction structure.

Extracting Latent Dynamics from Process Data for Quality Prediction and Performance Assessment Via Slow Feature Regression

Long-short Memory with Feature Representation by Latent-variable Model for Industrial Soft Sensor

Dynamic Probabilistic Latent Variable Model for Process Data Modeling and Regression Application

An Efficient Latent Variable Optimization Approach with Stochastic Constraints for Complex Industrial Process

Quality-Relevant Process Monitoring with Concurrent Locality-Preserving Dynamic Latent Variable Method

Locally weighted slow feature regression for nonlinear dynamic soft sensor modeling and its application to an industrial hydrocracking process

Probabilistic Slow Feature Analysis‐based Representation Learning from Massive Process Data for Soft Sensor Modeling

Sparse Slow Feature Analysis for Enhanced Control Monitoring and Fault Isolation

Probabilistic Slow Feature Regression For Dynamic Soft Sensing

Quality-relevant Dynamic Process Monitoring Based on Dynamic Total Slow Feature Regression Model

Assessment of process operating performance with supervised probabilistic slow feature analysis

Semi-Supervised Dynamic Latent Variable Regression for Prediction and Quality-Relevant Fault Monitoring

Locally Weighted Prediction Methods for Latent Factor Analysis with Supervised and Semisupervised Process Data

Concurrent Monitoring of Operating Condition Deviations and Process Dynamics Anomalies with Slow Feature Analysis

Fast Locally Weighted PLS Modeling for Large-Scale Industrial Processes

Deep Learning of Latent Variable Models for Industrial Process Monitoring

Monitoring and prediction of big process data with deep latent variable models and parallel computing

Weighted Linear Dynamic System for Feature Representation and Soft Sensor Application in Nonlinear Dynamic Industrial Processes.

Cooperative Deep Dynamic Feature Extraction and Variable Time-Delay Estimation for Industrial Quality Prediction

Enhancing Dynamic Soft Sensors Based on DPLS: A Temporal Smoothness Regularization Approach

Probabilistic latent variable regression model for process-quality monitoring