基于语义网络的航空安全主动报告信息分析方法研究

发布时间：2018-04-20 14:51

本文选题：民航安全 + 主动报告　；参考：《中国民航大学》2017年硕士论文

【摘要】：航空安全信息是辨识安全隐患、控制安全风险、改善航空安全的重要基础。除事故、事故征候信息以外,来自从业人员的主动报告信息也是完善航空系统运行的重要信息来源。深入挖掘主动报告信息中的隐患,梳理报告中可能诱发严重后果的安全风险,建立同一类事件原因与可能衍生后果之间的语义网络关系,可以为进一步的安全风险分析提供基于数据的分析基础。本文以美国航空安全报告系统(ASRS)发布的近4年200份飞行机组疲劳报告信息为样本,采用语义网络分析法,研究报告中飞行员疲劳发生的时间、所处飞行阶段、可能的原因、引发的后果,按照报告原文关键信息提取、疲劳语义网络框架构建、各模块节点确定和节点关系确定的流程,建立飞行疲劳事件语义网络图;利用灰色关联分析方法计算原因节点与疲劳之间、原因节点与后果节点之间的灰色加权关联度大小排序,进一步分析不同层次节点之间相关性;利用Bow-Tie模型对飞行疲劳语义网络内容进行针对原因的预防控制措施及针对后果的缓解措施补充分析,最终建立能描述疲劳事件原因与可能后果的语义网络图。以2015年报告序号为1255485的疲劳事件为例进行实证分析结果表明:参照飞行疲劳语义网络图对疲劳事件过程进行分析,一是可以便捷地得出飞行员疲劳事件包含的疲劳原因、后果等信息,使飞行员疲劳事件演化过程的信息脉络简明、清晰;二是可以根据分析出的事件原因,利用节点关联度分析结果,分析可能引发不同后果的安全风险,并参照建立的飞行疲劳Bow-tie模型给出控制措施建议;三是可以将新的原因和后果信息补充到现有语义网络图中,使其内容不断丰富和完善,进而为科学控制疲劳风险提供支持和参考。综上,语义网络分析法能够很好地梳理和分析知识及之间存在的联系,除进行飞行疲劳报告原因及后果的分析之外,也为主动报告中同一类事件的分析提供了参考。但由于主动报告无固定字段、顺序等要求的特点,语义网络目前在主动报告的分析中仍然处于人工分析的阶段,在以后的分析中可增加分析工具的使用,提高分析效率。
[Abstract]:Aviation safety information is an important foundation to identify hidden dangers, control safety risks and improve aviation safety. In addition to accident and accident information, active reporting information from employees is also an important source of information to improve the operation of aviation system. The hidden dangers of active reporting information are deeply excavated, the security risks that may induce serious consequences in the report are combed, and the semantic network relationship between the causes of the same kind of events and the possible derivative consequences is established. It can provide data-based analysis basis for further security risk analysis. Based on 200 flight crew fatigue report information released by ASRSs, this paper uses semantic network analysis method to study the time, flight stage and possible causes of pilot fatigue in the report. According to the process of extracting the key information, constructing the fatigue semantic network framework, determining the node of each module and determining the relationship between nodes, the semantic network chart of flight fatigue event is established. The grey relational analysis method is used to calculate the grey weighted correlation degree between cause node and fatigue node, cause node and consequence node, and further analyze the correlation between different levels of nodes. The Bow-Tie model is used to analyze the prevention and control measures for the causes and the mitigation measures for the consequences of the flight fatigue semantic network. Finally, a semantic network diagram which can describe the causes and possible consequences of the fatigue events is established. Taking the fatigue event reported as 1255485 in 2015 as an example, the results show that the fatigue event process can be analyzed by referring to the flight fatigue semantic network diagram. First, the fatigue causes contained in the pilot fatigue event can be obtained conveniently. Such information as consequences makes the information of the evolution process of pilot fatigue events concise and clear. Second, the safety risks that may lead to different consequences can be analyzed by using the results of nodal correlation analysis according to the causes of the events analyzed. According to the established flight fatigue Bow-tie model, the control measures are suggested. Thirdly, the new cause and consequence information can be added to the existing semantic network diagram, which can enrich and improve the content of the model. Furthermore, it provides support and reference for scientific control of fatigue risk. In summary, the semantic network analysis method can sort out and analyze the knowledge and the relationship between them. Besides the analysis of the causes and consequences of the flight fatigue report, it also provides a reference for the analysis of the same kind of events in the active reporting. However, the semantic network is still in the stage of manual analysis in the analysis of active reports because of the characteristics of no fixed fields and order of active reports. In the later analysis, the use of analysis tools can be increased and the efficiency of analysis can be improved.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V328

【参考文献】