An accelerated alignment method for analyzing time sequences of industrial alarm floods. (September 2017)

• Record Type: Journal Article
• Authors: Guo, Cen; Hu, Wenkai; Lai, Shiqi; Yang, Fan; Chen, Tongwen
• Is Part Of: Journal of process control. Volume 57(2017)
• Page Start: 102
• Page End: 115
• Publication Date: 2017-09
• Subjects: Industrial alarm systems -- Alarm floods -- Sequence alignment; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2017.06.019
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: An accelerated method is proposed to generate alarm flood sequence alignments. The method is robust, computationally efficient, and capable of exploring alarm time information. A plant dataset is used to demonstrate the method through specific alignment cases. Abstract: In industrial processes, analyzing and predicting process faults are quite important, which could help operators to take timely and effective responses to ensure process safety and prevent further losses, especially during alarm floods. Various fault analysis methods have been proposed so far, among which the alarm flood sequence alignment (AFSA) methods, unlike other traditional model-based or statistical methods, provide fault inference from the perspective of alarm sequence similarity assessment. A new AFSA method, the match-based accelerated alignment (MAA) is proposed to generate insightful and informative alarm sequence alignments. MAA focuses mainly on alarm match analysis and outperforms other methods in terms of robustness towards nuisance alarms and improved computational efficiency. More importantly, the alarm time information is well considered and explored in MAA, rendering its alignment results capable of revealing the real similarity of alarm floods. Numerical examples and a real chemical plant case are studied to demonstrate the effectiveness and efficiency of the proposed MAA method.