Copula-BN based risk assessment methodology of marine ship collisions

doi:10.16265/j.cnki.issn1003-3033.2023.09.1469

Abstract

Abstract:

In order to effectively prevent marine ship collisions, a Copula-BN-based dynamic risk assessment method for marine ship collisions was proposed. Gaussian Copula method was used to identify the edge distribution of each node, so that the complex relationships between risk factor and collision risk was more accurate and the Bayesian model inference results were more reasonable. Firstly, the data collected on ship collisions were combined with expert knowledge to identify and analyze the specific risk factors that led to ship collisions from the perspective of the human-ship-loop system. Gaussian Copula method was used to identify the edge distribution of each node and determine the optimal marginal distribution function of each node. Then BN structure was determined through expert experience and data structure analysis, thus completing the establishment of the Copula-BN model. Based on the correlation analysis, forward probability prediction analysis and backward fault diagnosis analysis, the key risk factors leading to the occurrence of ship collision were introduced based on the dependency between each risk factor and the target factor, and the whole risk system was dynamically analyzed through the state update of the nodes. The study shows that this method can identify the key risk factors leading to ship collision accidents and realize the dynamic assessment of ship collision risk.

Key words: marine ship collisions, Copula-Bayesian Networks(Copula-BN), correlation analysis, risk assessment, risk factors

LI Xinhong, FU Yaqian, LIU Yazhou, HAN Ziyue, ZHANG Renren. Copula-BN based risk assessment methodology of marine ship collisions[J]. China Safety Science Journal, 2023, 33(9): 204-213.

Figures/Tables 13

Tab.1

Common Copula functions

Copula函数	分布函数	$ρ / θ 取值范围$
Gaussian-Copula	$Φ ρ (Φ - 1 (u 1), Φ - 1 (u 2))$	[-1,1]
t-Copula	$T p, u [T μ - 1 (u), T μ - 1 (v)]$	[-1,1]
Gumbel-Copula	$e x p (- [(- l n u 1) θ + (- l n u 1) θ] 1 θ)$	0,1
Clayton-Copula	$m a x ([u 1 - θ + u 2 - θ - 1] 1 θ, 0)$	(0, $∞$ )
Frank-Copula	$1 θ l n (1 + e x p ((- θ u 1) - 1) e x p ((- θ u 2) - 1) e x p (- θ) - 1)$	(- $∞$ ,+ $∞$ )/{0}

Tab.1

Fig.1

Tab.2

Tab.3

Fig.2

Fig.3

Fig.4

Tab.4

Fig.5

Fig.6

Fig.7

Tab.5

Fig.8

References 16

[1]	CHEN Pengfei, HUANG Yamin, MOU Junmin, et al. Probabilistic risk analysis for ship-ship collision: state-of-the-art[J]. Safety Science, 2019, 117:108-122. doi: 10.1016/j.ssci.2019.04.014
[2]	KAPTAN M, UĞURLU Ö, WANG Jin. The effect of nonconformities encountered in the use of technology on the occurrence of collision, contact and grounding accidents[J]. Reliability Engineering & System Safety, 2021, 215:1-22.
[3]	信德海事网. 安联:2019年我国水上交通安全情况[EB/OL].(2020-03-12). https://www.xindemarinenews.com/topic/hangyunanquan/2020/0312/18926.html.
[4]	ZHENG Kai, JIANG Yi, ZHOU Shikun, et al. A comprehensive spatiotemporal metric for ship collision risk assessment[J]. Ocean Engineering, 2022, 265: DOI: 10.1016/J.OCEANENG.2022.112446.
[5]	郝勇, 吴志新, 傅春燕. 长江干线船舶碰撞原因的故障树分析:以武汉海事局辖区水域为例[J]. 中国安全生产科学技术, 2010, 6(5):171-176.
	HAO Yong, WU Zhixin, FU Chunyan. Fault tree analysis on collision between vessels in Changjing river:taking the water area of Wuhan Maritime Safety Administration for example[J]. Journal of Safety Science and Technoligy, 2010, 6(05):171-176.
[6]	司东森, 张英俊, 郎坤. 基于改进BN的集装箱船舶碰撞事故致因分析[J]. 中国安全科学学报, 2019, 29(10):31-37. doi: 10.16265/j.cnki.issn1003-3033.2019.10.006
	SI Dongsen, ZHANG Yingjun, LANG Kun. Causation analysis of container ship collision accidents based on improved BN[J]. China Safety Science Journal, 2019, 29(10):31-37. doi: 10.16265/j.cnki.issn1003-3033.2019.10.006
[7]	KHAN B, KHAN F, VEITCH B. A dynamic Bayesian network model for ship-ice collision risk in the Arctic waters[J]. Safety Science, 2020, 130:130-137.
[8]	AKHTAR M J, UTNE I B. Human fatigue's effect on the risk of maritime groundings: a Bayesian network modeling approach[J]. Safety Science, 2014, 62:427-440. doi: 10.1016/j.ssci.2013.10.002
[9]	MARTINS M R, MATURANA M C. Application of Bayesian belief networks to the human reliability analysis of an oil tanker operation focusing on collision accidents[J]. Reliability Engineering and System Safety, 2013, 110:110-121. doi: 10.1016/j.ress.2012.09.011
[10]	BAKSH A A, ABBASSI R, GARANIYA V, et al. Marine transportation risk assessment using Bayesian network: Application to Arctic waters[J]. Ocean Engineering, 2018, 159:422-436. doi: 10.1016/j.oceaneng.2018.04.024
[11]	GUO Chuanqi, KHAN F, IMTIAZ S. Copula-based Bayesian network model for process system risk assessment[J]. Process Safety and Environmental Protection, 2019, 123:317-326. doi: 10.1016/j.psep.2019.01.022
[12]	吴伋, 江福才, 姚厚杰, 等. 基于文本挖掘的内河船舶碰撞事故致因因素分析与风险预测[J]. 交通信息与安全, 2018, 36(3):8-18.
	WU Ji, JIANG Fucai, YAO Houjie, et al. An analysis and risk forecasting of inland ship collision based on text mining[J]. Journal of Transport Information and Safety, 2018, 36(3):8-18.
[13]	LAN He, MA Xiaoxiao, QIAO Weiliang, et al. Determining the critical risk factors for predicting the severity of ship collision accidents using a data-driven approach[J]. Reliability Engineering & System Safety, 2022, 230:230-244.
[14]	PAN Yue, OU Shenwei, ZHANG Limao, et al. Modeling risks in dependent systems: a Copula-Bayesian approach[J]. Reliability Engineering & System Safety, 2019, 188:416-431. doi: 10.1016/j.ress.2019.03.048
[15]	刘文黎, 吴贤国, 张文静, 等. 地铁健康监测PCBN模型的参数相关性分析[J]. 土木与环境工程学报:中英文, 2019, 41(2):45-52.
	LIU Wenli, WU Xianguo, ZHANG Wenjing, articele-title>Correlation analysis of parameters in PCBN model for metro health monitoring[J]. et al. Journal of Civil and Environmental Engineering: in English and Chinese, 2019, 41(2):45-52.
[16]	刘文, 刘文黎, 翟世鸿. 基于Copula相依模型的地铁结构安全可靠性分析[J]. 中国安全科学学报, 2019, 29(8):164-171. doi: 10.16265/j.cnki.issn1003-3033.2019.08.026
	LIU Wen, LIU Wenli, ZHAI Shihong. Metro structural safety and reliability analysis based on Copula dependency model[J]. China Safety Science Journal, 2019, 29(8):164-171. doi: 10.16265/j.cnki.issn1003-3033.2019.08.026

因素分类	具体节点内容
人为因素P	瞭望不当P₁、值守不当P₂、未及早采取避碰措施P₃、航速过快P₄、应急操作不当P₅、疲劳驾驶P₆、未履行相应义务P₇
船的因素S	导航设备故障或失灵S₁、船舶信号使用不当S₂、主机设备发生故障S₃
环境因素E	大风E₁、能见度不良E₂

风险因素	T	P₁	P₂	P₃	P₄	P₅	P₆
最优拟合函数	Normal	Weibull	Norma	Exponential	Gamma	Weibull	Norma
风险因素	P₇	S₁	S₂	S₃	E₁	E₂	—
最优拟合函数	Weibull	Gamma	Exponential	Norma	Gamma	Weibull	—

指标	P₁	P₂	P₃	P₄	P₅	P₆	P₇	S₁	S₂	S₃	E₁	E₂
各因素更新区域	[0.05,0.1]	[0.05,0.1]	[0,0.1]	[2,7]	[0.1,0.2]	[0,0.05]	[0.1,0.2]	[0.05,0.1]	[0.05,0.1]	[0,0.1]	[1,3]	[7,8]
T的均值	0.917	0.928	0.923	0.916	0.884	0.927	0.924	0.929	0.929	0.929	0.928	0.929
T的标准差	0.027	0.027	0.024	0.025	0.022	0.027	0.027	0.027	0.027	0.027	0.027	0.027

指标	P₁	P₂	P₃	P₄	P₅	P₆	P₇	S₁	S₂	S₃	E₁	E₂
原始均值	0.671	0.137	0.313	8.94	0.512	0.050	0.415	0.277	0.121	0.091	4.33	4.89
更新后均值	0.751	0.173	0.687	12.4	0.629	0.056	0.492	0.288	0.121	0.092	4.49	4.82
改变量/%	11.9	26.3	119.5	38.7	22.9	12.0	18.6	4.0	0	1.1	3.7	-1.4
原始标准差	0.254	0.158	0.4	3.38	0.109	0.020	0.22	0.171	0.073	0.062	1.77	1.22
更新后标准差	0.255	0.157	0.556	3.74	0.077	0.020	0.232	0.17	0.074	0.063	1.82	1.22
改变量/%	0.4	-0.6	39.0	10.7	-29.4	0	5.5	-0.6	1.4	1.6	2.8	0