Uncertain Dynamic Process Monitoring Using Moving Window Interval PCA

Imen Hamrouni; Hajer Lahdhiri; Okba Taouali; Kais Bouzrara; IEEE

doi:10.1109/SSD52085.2021.9429483

Back

Conference proceeding

Uncertain Dynamic Process Monitoring Using Moving Window Interval PCA

Imen Hamrouni, Hajer Lahdhiri, Okba Taouali, Kais Bouzrara and IEEE

2021 18th International Multi-Conference on Systems, Signals & Devices (SSD), pp.580-588

22/03/2021

DOI: https://doi.org/10.1109/SSD52085.2021.9429483

Abstract

Actuators

Analytical models

Data models

Fault detection

Fault detection and Isolation

Heuristic algorithms

Interval Data

interval moving window

Measurement uncertainty

Monitoring statistics

Principal Component Analysis

Process monitoring

Tennessee Eastman process (TEP)

Uncertainties

In this paper, we developed a new online approach for uncertain dynamic process using Moving Window Principal Component Analysis (MWPCA) for interval-valued data. The proposed approach is built based on the principal component analysis (CIPCA); thus deriving new Moving Window complete information PCA (MWCIPCA) approach. The proposed technique has been used to fault detection on a static system and Tennessee Eastman process (TEP) data. The performance of the proposed online MWCIPCA was compared with previously approach such CIPCA methods.

Metrics

1 Record Views

Details

Title: Uncertain Dynamic Process Monitoring Using Moving Window Interval PCA
Creators - without role: Imen Hamrouni - University of Sousse
Hajer Lahdhiri - University of Monastir
Okba Taouali - University of Tabuk
Kais Bouzrara - University of Monastir
IEEE
Publication Details: 2021 18th International Multi-Conference on Systems, Signals & Devices (SSD), pp.580-588
Publisher: IEEE
Identifiers: 9933815008331
Academic Unit: University of Tabuk
Language: English
Resource Type: Conference proceeding