Process alarm propagation path analysis method based on time delay estimation

doi:10.16265/j.cnki.issn1003-3033.2022.11.2321

Abstract

Abstract:

In order to avoid the problem of associated alarm flooding due to the large number of process equipment and the increase of process alarm configuration in complex large-scale systems, a process alarm propagation path analysis method based on time-delay analysis was proposed for the existing causal analysis methods with strong subjectivity and uncertainty factors and lack of validity test of time-delay relationship between variables. Based on K nearest neighbor impuation(KNNI) method, the time delay between variables was estimated to determine the direction of process risk propagation. By calculating the Sorgenfrei similarity coefficient, the alarm correlation degree between the two variables was determined, and the process alarm propagation path diagram was established. This method was applied to a central heating system in a school. The results show that the proposed method can accurately identify two associated alarm propagation paths in the process, namely X₁ (outlet temperature of No.102 heat exchanger) → X₂ (hot water supply temperature in school) → X₃ (hot water supply temperature of No.3 building in Tercero area) and X₁ → X₂ → X₄ (hot water supply temperature of canteen in Tercero area). The obtained time-delay estimation result are in line with the actual situation of the process. According to the model, the time delay of X₂→X₃ (5 min) is much smaller than the time delay of X₂→X₄ (24 min), but the degree of alarm correlation is similar. Based on the difference of the time delay between the two paths, if X₁ has a high alarm, it can prompt the operator to take prior measures (such as reducing hot water flow) to prevent X₃ from having high alarm in a short time, so as to prevent alarm and and ensure the process safety.

Key words: time delay estimation, alarm propagation path, analysis method, alarm flooding, associated alarm

CAI Shuang, WAN Aying, XU Qingqing. Process alarm propagation path analysis method based on time delay estimation[J]. China Safety Science Journal, 2022, 32(11): 126-133.

Figures/Tables 11

Fig.1

Fig.2

Tab.1

Fig.3

Fig.4

Fig.5

Tab.2

Time delay estimation results between variables

变量i	变量j	$d^i, j$	$ρ (d^i, j)$
1	2	2	0.83
1	3	7	0.81
1	4	26	0.80
2	3	5	0.80
2	4	24	0.79
3	4	19	0.79

Tab.2

Fig.6

Tab.3

Alarm correlation coefficient between variables

变量i	变量j	$d^i, j$	$N 1$	$N 2$	c( $d^$ )	ρ_sor( $d^i, j$ )
1	2	2	69	68	53	0.60
1	3	7	69	61	55	0.72
1	4	26	69	63	53	0.65
2	3	5	68	61	52	0.65
2	4	24	68	63	54	0.68
3	4	19	61	63	54	0.76

Tab.3

Fig.7

Fig.8

References 15

[1]	STOLAR A, FRIEDL A. Process safety for sustainable applications[J]. International Journal of Reliability Quality and Safety Engineering, 2021, 28(5):DOI:10.1142/S0218539321500339. doi: 10.1142/S0218539321500339
[2]	胡瑾秋, 张来斌, 伊岩, 等. 非正常工况下化工过程设备故障实时关联预警研究[J]. 中国安全科学学报, 2016, 26(9):140-145.
	HU Jinqiu, ZHANG Laibin, YI Yan, et al. Real-time fault associated early warning for petrochemical process equipment under abnormal condition[J]. China Safety Science Journal, 2016, 26(9):140-145.
[3]	WANG Jiandong, YANG Fan, CHEN Tongwen, et al. An overview of industrial alarm systems:main causes for alarm overloading, research status, and open problems[J]. Automation Science and Engineering, 2016, 13(2):1045-1061.
[4]	BAUER M, THORNHILL N F. A practical method for identifying the propagation path of plant-wide disturbances[J]. Journal of Process Control, 2008, 18(7):707-719. doi: 10.1016/j.jprocont.2007.11.007
[5]	BRESSLER S L, SETH A K. Wiener-Granger causality:a well established methodology[J]. Neuroimage, 2011, 58(2):323-329. doi: 10.1016/j.neuroimage.2010.02.059
[6]	DUAN Ping, YANG Fan, CHEN Tongwen, et al. Detection of direct causality based on process data[C]. The 2012 American Control Conference (ACC). MontrÉal:IEEE 2012: 3522-3527.
[7]	蔡爽, 胡瑾秋, 张来斌. 基于传递熵与KELM的炼油化工过程风险传播路径分析方法[J]. 中国安全生产科学技术, 2018, 14(3):19-26.
	CAI Shuang, HU Jinqiu, ZHANG Laibin. Analysis method of risk propagation paths in refining petrochemical process based on transfer entropy and KELM[J]. Journal of Safety Science and Technology, 2018, 14(3):19-26.
[8]	李威君. 风险动态评估理论与方法研究及其在天然气站场的应用[D]. 北京: 中国石油大学(北京), 2017.
	LI Weijun. Risk dynamic evaluation theory and method research and their application to gas transmission station[D]. Beijing: China University of Petroleum(Beijing), 2017.
[9]	PENG Di, GU Xiangbai, XU Yuan, et al. Integrating probabilistic signed digraph and reliability analysis for alarm signal optimization in chemical plant[J]. Journal of Loss Prevention in the Process Industries, 2015, 33:279-288. doi: 10.1016/j.jlp.2015.01.005
[10]	WANG Jia, LI Hongguang, HUANG Jingwen, et al. Association rules mining based analysis of consequential alarm sequences in chemical processes[J]. Journal of Loss Prevention in the Process Industries, 2016, 41:178-185. doi: 10.1016/j.jlp.2016.03.022
[11]	YANG Fan, SHAH S L, XIAO Deyun, et al. Improved correlation analysis and visualization of industrial alarm data[J]. ISA Transactions, 2012, 51(4):499-506. doi: 10.1016/j.isatra.2012.03.005 pmid: 22503464
[12]	GAO Huihui, LIU Feifei, ZHU Qunxiong. A correlation consistency based multivariate alarm thresholds optimization approach[J]. ISA Transactions, 2016, 65:37-43. doi: S0019-0578(16)30334-2 pmid: 27665140
[13]	BENATIA M A, LOUIS A, BAUDRY D. Alarm correlation to improve industrial fault management[J]. IFAC-PapersOnLine, 2020, 53(2):10485-10 492. doi: 10.1016/j.ifacol.2021.04.044
[14]	KACED R, KOUADRI A, BAICHE K, et al. Multivariate nuisance alarm management in chemical processes[J]. Journal of Loss Prevention in the Process Industries, 2021, 72:DOI:10.1016/j.jlp.2021.104548. doi: 10.1016/j.jlp.2021.104548
[15]	STOCKMANN M, HABER R, SCHMITZ U. Source identification of plant-wide faults based on k nearest neighbor time delay estimation[J]. Journal of Process Control, 2012, 22(3):583-598. doi: 10.1016/j.jprocont.2012.01.005

变量	位号	描述
X₁	TT8227	102号热交换器出口温度
X₂	TT8238	校内热水供给温度
X₃	TT8271	Tercero区域3号楼热水供给温度
X₄	TT8281	Tercero区域食堂热水供给温度