A knowledge graph-based approach for exploring railway operational accidents. (March 2021)

• Record Type: Journal Article
• Authors: Liu, Jintao; Schmid, Felix; Li, Keping; Zheng, Wei
• Is Part Of: Reliability engineering & system safety. Volume 207(2021)
• Publication Date: 2021-03
• Subjects: Network analysis -- Topological analysis -- Knowledge graph -- Accident analysis -- Railway operational accident; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2020.107352
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: An accident investigation reports-based knowledge graph modeling method is developed. A knowledge graph-based topological analysis method is proposed for exploring railway operational accidents. An application to a real set of railway operational accidents is provided. Results show that the proposed approach can help railway operators to understand past accidents and to make accident prevention decisions. Abstract: Drawing lessons from past accidents is an essential way to improve the operational safety of railways. Various railway operational accidents and their related hazards constitute a causation network due to the interactions among the hazards. Some useful lessons can be captured from such a network. In this paper, a new knowledge graph-based approach to explore railway operational accidents is proposed, aiming to reveal the potential rules of accidents by depicting accidents and hazards in a heterogeneous network. This work serves as an extension and complement to classical homogeneous network-based accident analyses. Its originality is to apply the knowledge graph theory to railway operational accident analysis, by means of some topological indicators adapting to the heterogeneous structural features of knowledge graphs. To facilitate the construction of the accident knowledge graph, a modelling method is developed. The outcomes of the knowledge graph-based analysis provide railway operators with the decision-making basis for the investment of accident prevention efforts. An application on real railway operational accidents in the UK is presented. The results show the effectiveness of the proposed approach in terms of discovering the latent features of the corresponding railway operational accidents and assisting in formulating targeted preventive measures.