Dynamic process monitoring based on a time-serial multi-block modeling approach. (May 2020)

• Record Type: Journal Article
• Authors: Wan, Xinchun; Tong, Chudong; Meng, Shengjun; Lan, Ting
• Is Part Of: Journal of process control. Volume 89(2020)
• Page Start: 22
• Page End: 29
• Publication Date: 2020-05
• Subjects: Dynamic process monitoring -- Multi-block modeling -- Time-serial auto-correlation; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2020.03.007
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: A novel time-serial multi-block modeling method is proposed for process monitoring. A unified framework of multi-block modeling and auto-correlation extraction is considered. Comparisons demonstrated the superiority and effectiveness of the TSMBM-based method. Abstract: A time-serial multi-block modeling (TSMBM) algorithm is proposed for dynamic process monitoring, which considers a unified framework of multi-block modeling and auto-correlation extraction. The proposed TSMBM-based method first constructs time-serial multi-blocks according to the sampling time nodes, the correlation between different blocks then serves as a good representation for the auto-correlated characteristic in the given data. With the utilization of multi-block projecting bases, three categories of auto-correlated variations can be captured in different block models. Furthermore, the Kalman filter is employed to generate dynamic noise and measurement noise inheriting little auto-correlation for online monitoring purposes. Finally, the effectiveness and superiority of the proposed method are validated through comparisons with other state-of-art dynamic process monitoring approaches.