面向城市轨道交通列车运维安全的事理图谱构建研究  

Research on the Construction of Event Evolutionary Graph for Urban Rail Transit Train Operation and Maintenance Safety

作　　者：樊茜琪 左建勇 龚明 贾波 李政江 FAN Qianqi;ZUO Jianyong;GONG Ming;JIA Bo;LI Zhengjiang(College of Transportation Engineering,Tongji University,201804,Shanghai,China;CRRC Academy,100160,Beijing,China;Shanghai Railway Certification(Group)Co.,Ltd.,201804,Shanghai,China)

机构地区：[1]同济大学交通学院,上海201804 [2]中车工业研究院有限公司,北京100160 [3]上海轨道交通检测认证(集团)有限公司,上海201804

出　　处：《城市轨道交通研究》2025年第4期32-37,共6页Urban Mass Transit

基　　金：国家重点研发计划项目(2020YFB1600704)。

摘　　要：[目的]城市轨道交通运维系统庞大而复杂,使用传统数据准确分析运维风险具有一定难度,将大数据分析纳入风险管理是一种趋势。通过构建城市轨道交通列车运维安全事理图谱,为城市轨道交通安全管理和风险防控提供新的视角和工具,从而为提升城市轨道交通系统的安全性提供理论基础和方法支持。[方法]聚焦于列车运维安全事理图谱的构建,对构建事理图谱的3个关键步骤进行了优化。事件抽取环节构建了城市轨道交通专业词典并设计了触发词规则;事件关系抽取环节结合了语义依存与顺序关系;事理对齐环节采用优化的聚类算法来定义主要风险因素。通过对传统知识图谱方法、通用事理图谱方法以及本文方法的对比评价,验证本文方法在事理图谱构建中的准确性和有效性。最后通过Gephi软件实现了事理图谱用数据存储和可视化。[结果及结论]传统的基于数理统计的故障分析和依赖人工经验的风险分析在预见性、全面性和及时性等方面存在一定的局限性;展示了事理图谱在运维风险管理中从理论到应用的完整路径,并实现了数据分析与可视化表示的综合方法论。随着方法论的不断完善和实证研究的深入,这套城市轨道交通领域事理图谱抽取范式将能进一步促进城市轨道交通安全管理水平的提升。[Objective] As the urban rail transit operation and maintenance system is large and complex,it is somewhat difficult to accurately analyze the operation and maintenance risks using traditional data.Incorporating big data analysis into risk management is a development trend.By constructing an event evolutionary graph for urban rail train operation and maintenance safety,new perspectives and tools are provided for urban rail transit safety management,risk prevention and control,thereby providing a theoretical basis and methodological support for improving the safety of urban rail transit systems.[Method] Focusing on the construction of event evolutionary graph for urban rail train operation and maintenance safety,three key steps for constructing the event evolutionary graph are optimized.In the event extraction stage,a professional dictionary of urban rail is constructed and trigger word rules are designed;in the event relationship extraction stage,semantic dependency and sequential relationship are combined;in the event evolutionary alignment stage,an optimized clustering algorithm is used to define the main risk factors.Through comparative evaluation of traditional knowledge graph methods,general event evolutionary graph methods and the above proposed method,accuracy and effectiveness of the last one in the construction of event evolutionary graph are verified.Finally,the data storage and visualization of the event evolutionary graph are realized through Gephi software.[Result & Conclusion] Traditional fault analysis based on mathematical statistics and risk analysis relying on manual experience have certain limitations in terms of predictability,comprehensiveness,and timeliness.The complete path from theory to application of event evolutionary graph in operation and maintenance risk management is presented,and a comprehensive methodology for data analysis and visual representation is realized.With the continuous improvement of methodology and the deepening of empirical research,it is believed that this paradigm

关 键 词：城市轨道交通 列车运维安全 事理图谱 风险分析 

分 类 号：U298[交通运输工程—交通运输规划与管理]