基于ACE-BN的通勤飞行事故/事件诱因分析

doi:10.16265/j.cnki.issn1003-3033.2023.02.1109

摘要/Abstract

摘要：

为降低我国通用航空短途运输运行安全风险,提出一种基于平均因果效应(ACE)和贝叶斯网络(BN)的事故/事件诱因分析方法。首先,分析102起美国通勤飞行事故/事件,共识别出 7类 19种诱因;然后,引入ACE公式确定节点的优先次序,使用K2算法构建BN结构,并采用最大期望(EM)算法进行网络参数学习,建立通勤飞行事故/事件诱因分析模型;最后,对各诱因的概率进行排序并分析诱因间的敏感性。结果表明:机组经验不足导致的事故/事件发生概率最高;机械设备情况容易受到雨雪等恶劣天气的影响;天气因素和监管因素对机组因素表现出较高的敏感性,其中,恶劣的天气会影响机翼等机械设备的正常运转,进而影响飞行员对于飞机的操作,技术和安全培训不到位会影响飞行员和管制人员的专业水平及经验技能。

关键词: 平均因果效应(ACE), 贝叶斯网络(BN), 通勤飞行, 事故/事件, 诱因分析

Abstract:

In order to reduce the operational safety risk of general aviation short-distance transportation in China, an accidents/incidents inducement analysis method based on ACE and BN was proposed. Firstly, 102 U.S.commuter flight accidents/incidents were analyzed, and a total of 7 categories and 19 causes were identified. Then, ACE formula was introduced to determine the priority of nodes, the K2 algorithm was used to construct the BN structure, and the maximum expectation (EM) algorithm was used to learn the network parameters, and the incentive analysis model of commuting flight accidents/events was established. Finally, the probability of each inducement was sorted and the sensitivity among inducements was analyzed. The results show that the probability of accidents/events caused by insufficient unit experience is the highest. The mechanical equipment is easily affected by bad weather such as rain and snow; weather factors and regulatory factors are highly sensitive to unit factors. Among them, bad weather will affect the normal operation of mechanical equipment such as wings, which in turn affects pilots' operation of aircraft. In adequate technical and safety training will affect the professional level and experience and skills of pilots and controllers.

Key words: average causal effect (ACE), Bayesian network (BN), commuter flight, accidents/incidents, incentive analysis

靳慧斌, 朱孟昌, 马明霞. 基于ACE-BN的通勤飞行事故/事件诱因分析[J]. 中国安全科学学报, 2023, 33(2): 96-102.

JIN Huibin, ZHU Mengchang, MA Mingxia. Cause analysis of commuter flight accidents/incidents based on ACE-BN[J]. China Safety Science Journal, 2023, 33(2): 96-102.

图/表 9

图1

图2

表1

通勤飞行事故/事件诱因分析指标体系

目标层	准则层	指标层
通勤飞行事故/事件诱因分析指标体系	机组原因	机组经验技能 $X 1$
		违规操作 $X 2$
		航前准备 $X 3$
		机组人员生理心理状态 $X 4$
	机务原因	飞机维护达标性 $X 5$
	机务原因	机型繁杂程度 $X 6$
	机械原因	发动机情况 $X 7$
		起落架情况 $X 8$
		机翼情况 $X 9$
		其他机械设备情况 $X 10$
	空管原因	空域繁忙程度 $X 11$
	空管原因	管制人员专业水平 $X 12$
	地面保障原因	特种设备使用情况 $X 13$
	地面保障原因	机场环境影响 $X 14$
	天气/意外原因	天气情况气象条件 $X 15$
	天气/意外原因	鸟类等野生动物影响 $X 16$
	监管原因	安全监督及管理 $X 17$
		规章标准制定执行情况 $X 18$
		技术和安全培训 $X 19$

表1

图3

表2

通勤飞行事故/事件诱因数据矩阵示例

编号	诱因分析指标
编号	$X 1$	$X 2$	…	$X 14$	$X 15$	…	$X 19$
$C 1$	0	0	0	0	1	0	0
$C 2$	1	1	0	1	1	0	0
$︙$	—	—	—	$︙$	—	—	—
$C 102$	0	0	…	0	0	…	0

表2

表3

图4

表4

图5

参考文献 19

[1]	中国航空运输协会通用航空分会. 中国通用航空短途运输行业发展报告[R], 2020.
[2]	MENG Guowang, YE Yangchun, WU Bo, et al. Risk assessment of shield tunnel construction in karst strata based on fuzzy analytic hierarchy process and cloud model[J]. Shock and Vibration, 2021, 2021:DOI: 10.1155/2021/7237136.
[3]	LIU Lingyun. Oil spill risk assessment of submarine oil pipeline based on fuzzy comprehensive evaluation and accounting[J]. Arabian Journal of Geosciences, 2021, 14(18): 1-18. doi: 10.1007/s12517-020-06304-8
[4]	余建星, 范海昭, 陈海成, 等. 海洋立管失效风险因素分析方法及应用[J]. 中国安全科学学报, 2021, 31(11):47-53. doi: 10.16265/j.cnki.issn 1003-3033.2021.11.007
	YU Jianxing, FAN Haizhao, CHEN Haicheng, et al. Analysis method and application of marine riser failure risk factors[J]. China Safety Science Journal, 2021, 31(11):47-53. doi: 10.16265/j.cnki.issn 1003-3033.2021.11.007
[5]	SHIN J, YOU I, SEO J T. Investment priority analysis of ICS information security resources in smart mobile IoT network environment using the analytic hierarchy process[J]. Mobile Information Systems, 2020, 2020(3):1-11.
[6]	ZHOU Di, ZHUANG Xiao, ZUO Hongfa. A hybrid deep neural network based on multi-time window convolutional bidirectional LSTM for civil aircraft APU hazard identification[J]. Chinese Journal of Aeronautics, 2022, 35(4):344-361. doi: 10.1016/j.cja.2021.03.031
[7]	HUANG Junbo, XU Zhen, YANG Feng, et al. Fire risk assessment and warning based on hierarchical density-based spatial clustering algorithm and grey relational analysis[J]. Mathematical Problems in Engineering, 2022, 2022: DOI: 10.1155/2022/7339312. doi: 10.1155/2022/7339312
[8]	RAO A H, MARAIS K. A state-based approach to modeling general aviation accidents[J]. Reliability Engineering and System Safety, 2020, 193: DOI: 10.1016/j.ress.2019.106670. doi: 10.1016/j.ress.2019.106670
[9]	王永刚, 孙瑶. 基于BN的FTA在通用航空风险评价中的应用[J]. 中国安全科学学报, 2010, 20(3): 19-23.
	WANG Yonggang, SUN Yao. Application of BN-based FTA in general aviation risk assessment[J]. China Safety Science Journal, 2010, 20(3): 19-23.
[10]	韩梅, 吴珊, 常青, 等. 基于事故树和模糊贝叶斯网络的铁路超限货物运输风险评估[J]. 铁道学报, 2021, 43(5): 9-17.
	HAN Mei, WU Shan, CHANG Qing, et al. Risk assessment of railway over-gauge cargo transportation based on fault tree and fuzzy Bayesian network[J]. Journal of the China Railway Society, 2021, 43(5): 9-17.
[11]	潘丹, 李永周, 罗帆, 等. 飞行区外来物入侵安全风险致因FTA-BN模型[J]. 中国安全科学学报, 2021, 31(6):7-13. doi: 10.16265/j.cnki.issn 1003-3033.2021.06.002
	PAN Dan, LI Yongzhou, LUO Fan, et al. FTA-BN model for the security risk of foreign object intrusion in the flight area[J]. China Safety Science Journal, 2021, 31(6):7-13. doi: 10.16265/j.cnki.issn 1003-3033.2021.06.002
[12]	王旭. 基于模糊贝叶斯网络的城市轨道交通运营安全风险评估研究[D]. 南昌: 华东交通大学, 2018.
	WANG Xu. Research on safety risk assessment of urban rail transit operation based on fuzzy Bayesian network[D]. Nanchang: East China Jiaotong University, 2018.
[13]	PEARL J, GLYMOUR M, NICHOLAS P J. Causal inference in statistics:a primer[M]. Hoboken: Wiley, 2016:636-642.
[14]	寇兴怡, 帅斌, 黄文成. 基于贝叶斯网络的高速动车组运营故障分析[J]. 中国安全生产科学技术, 2020, 16(4): 63-69.
	KOU Xingyi, SHUAI Bin, HUANG Wencheng. Operation failure analysis of high-speed EMU based on Bayesian network[J]. Journal of Safety Science and Technology, 2020, 16(4): 63-69.
[15]	周艳. 朴素贝叶斯分类器的研究与改进[D]. 厦门: 厦门大学, 2017.
	ZHOU Yan. Research and improvement of naive bayes classifier[D]. Xiamen: Xiamen University, 2017.
[16]	COOPER G F. A Bayesian method for the induction of probabilistic networks from data[J]. Machine Learning, 1992, 9(4): 309-347.
[17]	安宁, 滕越, 杨矫云, 等. 基于因果效应的贝叶斯网络结构学习方法[J]. 计算机应用研究, 2018, 35(12): 3609-3613.
	AN Ning, TENG Yue, YANG Jiaoyun, et al. A bayesian network structure learning method based on causal effects[J]. Computer Application Research, 2018, 35(12): 3609-3613.
[18]	Board Nation-Transportation-Safety. Aviation accident database &synopses[DB/OL]. [2022-03-10]. https://www.ntsb.gov/Pages/home.aspx
[19]	中国民用航空局航空安全办公室. 中国民航航空安全报告[R], 2019.

节点次序	事故/事件诱因	节点次序	事故/事件诱因
1	技术和安全培训X₁₉	4	机翼情况X₉
	天气情况气象条件X₁₅		发动机情况X₇
	鸟类等野生动物影响X₁₆		起落架情况X₈
	安全监督及管理X₁₇		其他机械设备情况X₁₀
	空域繁忙程度X₁₁	5	违规操作X₂
2	机组人员生理心理状态X₄	5	飞机维护达标性X₅
	航前准备X₃	6	机组人员经验技能X₁
	规章标准制定执行情况X₁₈		管制人员专业水平X₁₂
3	机型繁杂程度X₆		机场环境影响X₁₄
3	特种设备使用情况X₁₃	—	—