基于QAR数据的飞行员个体超限风险精细化评价模型

doi:10.16265/j.cnki.issn1003-3033.2024.08.1824

摘要/Abstract

摘要：

为定量评价飞行员个体超限风险,提出一种基于快速存取记录器(QAR)数据和飞行品质监控(FOQA)的飞行员个体超限风险精细化评价模型。首先,根据事故统计结果、国际民航组织 (ICAO)及FOQA基站划分的核心风险类别,选取其中3类风险的FOQA监控项目作为评价指标,计算飞行员个体单项核心风险值;然后,运用熵权逼近理想解排序法(TOPSIS)测算各项核心风险值所占权重,提出飞行员个体超限风险精细化评价模型;最后,将模型运用于实际飞行风险量化评价,通过采集中国民航(CAAC)FOQA基站中共9 317条多源融合数据,量化飞行员个体超限风险,得到飞行员个体超限风险量化值排序,结合K-means聚类算法划分飞行员个体超限风险等级。结果表明:该模型可量化排序1 693名飞行员个体超限风险,并将飞行员超限风险等级划分为高风险、中风险和低风险3类。

关键词: 快速存取记录器(QAR)数据, 个体超限风险, 精细化评价, 飞行品质监控(FOQA), 熵权逼近理想解排序(TOPSIS)法, K-means聚类

Abstract:

In order to achieve a quantitative evaluation of individual pilot's exceedance risk, a refined evaluation model for pilot's individual exceedance risk was established based on QAR data and the flight operations quality assurance(FOQA) monitoring items. Firstly, according to accident statistics, International Civil Aviation Organization (ICAO) and the core risks divided by the FOQA station, the FOQA monitoring items associated with the three types of core risks were selected as the evaluation indexes, and the risk value for each core risk of the individual pilots was calculated. In the next step, the weight of each core risk value was calculated by entropy weighted TOPSIS. Then, the refined evaluation model for the pilot's individual exceedance risk was established. Finally, the model was applied to the quantitative evaluation of actual flight risk. By collecting 9 317 pieces of multi-source fusion data from the FOQA station of Civil Aviation Administration of China (CAAC), the individual exceedance risk of the pilots were quantified, the ranking of individual pilots' exceedance risk was obtained, and the pilot's individual exceedance risk levels were also divided with the use of K-means clustering algorithm. The results show that the model can quantify and rank 1 693 individual pilots' exceedance risk, and divide the pilot's exceedance risk into three types, including high risk, medium risk and low risk.

Key words: quick access recorder (QAR) data, individual exceedance risk, refined evaluation, flight operations quality assurance (FOQA), entropy-weighted technique for order preference by similarity to an ideal solution(TOPSIS), K-means clustering

中图分类号:

X949

汪磊, 安佳宁, 赵新斌, 俞力玲. 基于QAR数据的飞行员个体超限风险精细化评价模型[J]. 中国安全科学学报, 2024, 34(8): 35-42.

WANG Lei, AN Jianing, ZHAO Xinbin, YU Liling. Refined evaluation model for pilot's individual exceedance risk based on QAR data[J]. China Safety Science Journal, 2024, 34(8): 35-42.

图/表 11

表1

表2

表3

表4

图1

图2

表5

图3

表6

表7

图4

参考文献 22

[1]	International Civil Aviation Organization. Safety report 2020[R], 2020.
[2]	中国民航局. 中国民航航空安全报告, 2022[R], 2022.
[3]	International Civil Aviation Organization. Safety report 2022[R], 2022.
[4]	Airbus SAS. A statistical analysis of commercial aviation accidents 1958-2022[R], 2022.
[5]	汪磊, 梁妍. 全球民航事故调查数据统计与分析[J]. 综合运输, 2021, 43(3):7-12.
	WANG Lei, LIANG Yan. Statistical analysis on global civil aviation accident investigation data[J]. China Transportation Outlook, 2021, 43(3): 7-12.
[6]	汪磊, 孙瑞山, 吴昌旭, 等. 基于飞行QAR数据的重着陆风险定量评价模型[J]. 中国安全科学学报, 2014, 24(2):88-92.
	WANG Lei, SUN Ruishan, WU Changxu, et al. A flight QAR based model for hard landing risk quantitative evaluation[J]. China Safety Science Journal, 2014, 24(2): 88-92.
[7]	孙瑞山, 杨绎煊, 汪磊. QAR数据在飞行安全评价中的应用[J]. 中国安全科学学报, 2015, 25(7):87-92.
	SUN Ruishan, YANG Yixuan, WANG Lei. Study on flight safety evaluation based on QAR data[J]. China Safety Science Journal, 2015, 25(7): 87-92.
[8]	LU Binbin, SUN Huabo, ZHANG Xiaoyue, et al. Exploring spatial patterns of flight turbulence incidents in China with QAR data[C]. 2019 IEEE 1^st International Conference on Civil Aviation Safety and Information Technology (ICCASIT), 2019: 570-573.
[9]	LI Xu, SHANG Jiaxing, ZHENG Linjiang, et al. Curvecluster+: curve clustering for hard landing pattern recognition and risk evaluation based on flight data[C]. IEEE Transactions on Intelligent Transportation Systems, 2022: 12 811-12 821.
[10]	汪磊, 孙景陆, 王文超, 等. 基于QAR数据的着陆超限风险贝叶斯网络分析模型[J]. 安全与环境学报, 2023, 23(1):26-34.
	WANG Lei, SUN Jinglu, WANG Wenchao, et al. Bayesian network analysis model on landing exceedance risk based on flight QAR data[J]. Journal of Safety and Environment, 2023, 23(1): 26-34.
[11]	康宗伟. 基于QAR数据和深度学习的民航客机冲出跑道事件预警研究[D]. 重庆: 重庆大学, 2022.
	KANG Zongwei. Research on early warning of airliner runway overrun events based on QAR data and deep learning[D]. Chongqing: Chongqing University, 2022.
[12]	陈农田, 满永政, 李俊辉. 基于QAR数据的民机高高原进近着陆风险评估方法[J]. 北京航空航天大学学报, 2024, 50(1):77-85.
	CHEN Nongtian, MAN Yongzheng, LI Junhui. Risk assessment method for civil aircraft approach and landing at high plateau based on QAR data[J]. Journal of Beijing University of Aeronautics and Astronautics, 2024, 50(1):77-85.
[13]	WANG Lei, GAO Shan, HONG Ruiyuan, et al. Effects of age and flight exposure on flight safety performance: evidence from a large cross-sectional pilot sample[J]. Safety Science, 2023, 165: DOI: 10.1016/j.ssci.2023.106199.
[14]	WANG Lei, GAO Shan, TAN Wei, et al. Pilot's mental workload variation when taking a risk in flight scenario: a study based on flight simulator experiments[J]. International Journal of Occupational Safety and Ergonomics, 2023, 29: 1-35.
[15]	占欣. 基于QAR数据的冲/偏出跑道风险评估研究[D]. 天津: 中国民航大学, 2020.
	ZHAN Xin. Research on risk evaluation of runway excursion based on QAR data[D]. Tianjin: Civil Aviation University of China, 2020.
[16]	Loss of control in-flight accident analysis report[R/OL].[2019-10-11]. https://www.iata.org/contentassets/b6eb2adc248c484192101edd1ed36015/loc-i_2019.pdf.
[17]	Guidance material and best practices for the implementation of upset prevention and recovery training[Z].[2023-11-12]. https://www.iata.org/contentassets/b6eb2adc248c484192101edd1ed36015/gmbp_uprt.pdf.
[18]	杜红兵, 王雪莉. 基于贝叶斯网络的可控飞行撞地事故原因分析方法[J]. 安全与环境学报, 2009, 9(5):136-139.
	DU Hongbing, WANG Xueli. Cause analysis of the controlled flight bumping into terrain based on the Bayesian network[J]. Journal of Safety and Environment, 2009, 9(5): 136-139.
[19]	方方, 赵新斌, 张日权, 等. 一种航空风险评价方法、装置及计算机设备: 中国, CN202010399853.2[P]. 2023-06-13.
[20]	赵新斌, 张日权, 王浩锋, 等. 一种运行风险量化方法、运行风险评价方法及装置: 中国,CN201811238049.5[P]. 2023-05-23.
[21]	任亮, 张海涛, 魏明珠, 等. 基于熵权TOPSIS模型的智慧城市发展水平评价研究[J]. 情报理论与实践, 2019, 42(7):113-118,125.
	REN Liang, ZHANG Haitao, WEI Mingzhu, et al. Research on the evaluation of development level of smart city based on entropy TOPSIS model[J]. Expert Systems with Applications, 2019, 42(7): 113-118, 125.
[22]	International Air Transport Association. Interactive safety report[R/OL].[2023-12-31]. https://www.iata.org/en/publications/safety-report/interactive-safety-report/.

时间	核心风险类别	飞行员	风险值
2020-11-22	冲偏出跑道	F₁	0.004 980 916
2021-03-14	可控飞行撞地	F₂	0.001 598 222 3
2021-03-22	冲偏出跑道	F₃	0.004 465 648 9
2021-06-20	冲偏出跑道	F₄	0.000 590 422 2
2021-09-5	空中失控	F₅	0.002 491 506 2

F	RE风险值	LOC风险值	CFIT风险值
飞行员F₆	5.200 982 523	0	46.012 269 938
飞行员 F₇	6.204 539 982	0.226 500 563	0.154 041 108
飞行员 F₈	1.204 283 986	3.233 034 572	1.090 901 236
飞行员 F₉	19.947 326 070	1.424 487 463	0
飞行员 F₁₀	10.276 362 606	36.363 636 363	25.144 733 431

指标	发生频率	信息熵值	信息效用值	权重
RE	0.846	0.935	0.065	0.084 74
LOC	0.032	0.634	0.366	0.474 46
CFIT	0.122	0.660	0.340	0.440 80

飞行员个体数	相关系数
Top 100%	0.965
Top 50%	0.926
Top 10%	0.751

聚类类别	频数	占比/%	均值±标准差	聚类中心
类别1	1 490	88.009	0.02±0.017	0.019 735 623 2
类别2	176	10.396	0.119±0.041	0.336 348 159 5
类别3	27	1.595	0.336±0.077	0.118 726 863 1