Abstract:Most chemical processes are time-varying and nonlinear due to load changes, product demand transitions or other causes, causing unsatisfactory performance of conventional non-parametric multivariate statistical fault detection method. In this work, a probabilistic multivariate state estimation fault detection method based on multi-output Gaussian process autoregression is first developed to overcome the linear limitation and noise information ignorance of multivariate state estimation technology-based fault detection methods. In multi-output Gaussian process regression framework, the Leave-One-Sample-Out strategy is developed to construct the dataset by leaving each historical sample from the memory matrix out as one training output in turn while regarding the corresponding remaining historical samples as one training input. Then, we design the multi-output Gaussian process autoregression model to estimate the current state of the system. Utilizing the estimation covariance information, a probabilistic index based on multivariate Kullback-Leibler divergence is developed for fault detection. Moreover, a multi-scale adaptive fault detection enhancement based on discrete multi-objective Archimedes optimization and just-in-time learning is further proposed to handle the time-varying problem. In this scheme, the memory matrix is batch-adaptively updated at monthly intervals based on discrete multi-objective Archimedes optimization to handle the slow change of system characteristics caused by equipment degradation. On the other hand, to cope with the rapid change of system characteristics caused by operation transition, the just-in-time learning is employed to select samples from the memory matrix that best match to the current system state for sample-adaptive updating. Finally, two industrial cases are used to verify the performance of the proposed method.

Functional State-Space Model for Multi-Channel Autoregressive Profiles with Application in Advanced Manufacturing

Autoregressive Dynamic Latent Variable Models for Process Monitoring

Functional Graphical Models for Manufacturing Process Modeling

Weakly Correlated Profile Monitoring Based on Sparse Multi-Channel Functional Principal Component Analysis

In-profile Monitoring for Cluster-Correlated Data in Advanced Manufacturing System

Product Quality Improvement Method In Manufacturing Process Based On Kernel Optimisation Algorithm

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

Online monitoring of profiles via RVC model with an application to industrial busbar

Multiple Profiles Sensor-Based Monitoring and Anomaly Detection

Monitoring Heterogeneous Multivariate Profiles Based on Heterogeneous Graphical Model.

A quality-driven stability analysis framework based on state fluctuation space model for manufacturing process

Robust Supervised Probabilistic Factor Analysis and Its Application to Industrial Soft Sensor Modeling.

Multimode Process Monitoring Based on Switching Autoregressive Dynamic Latent Variable Model

Multirate Dynamic Process Monitoring Based on Multirate Linear Gaussian State-Space Model.

Multi-rate Principal Component Regression Model for Soft Sensor Application in Industrial Processes.

Linear Modeling and Control of Dimensional Variation in Multi-Station Manufacturing System

Operational risk modeling based on operational data fusion for multi-state manufacturing systems

Multi-subspace Factor Analysis Integrated with Support Vector Data Description for Multimode Process Monitoring.

Multi-scale adaptive multivariate state estimation fault detection enhancement for time-varying industrial system based on multi-output Gaussian process autoregression

Employing machine learning techniques in monitoring autocorrelated profiles

An ensemble framework based on multivariate statistical analysis for process monitoring