Published Papers

Volume 31 - Issue 4 (Award Issue)
Shortlisted Papers

AI-Driven railway regulator inspection planning system: enhancing railway safety inspection prioritisation and incident management

Coe M Y Chiu
Pages: 1-7Published: 20 Dec 2024
DOI: 10.33430/V31N4THIE-2024-0012
Cite thisHide

Chiu MY, AI-Driven railway regulator inspection planning system: enhancing railway safety inspection prioritisation and incident management, HKIE Transactions, Vol. 31, No. 4 (Award Issue), Article THIE-2024-0012, 2024, 10.33430/V31N4THIE-2024-0012

 Copy

Abstract:

This paper presents an innovative approach, with an application named AI-Driven Railway Regulator Inspection Planning System (the System), to enhance the safety inspection prioritisation and incident management of railways from regulator's perspective. The conventional approach to inspection planning relies on predetermined schedules and the manual selection of inspection items, which is time consuming and not conducive to objectively identifying high-risk assets and optimising resources. The system utilises Natural Language Processing (NLP) and Convolutional Neural Network (CNN) algorithms to quickly extract the semantic meaning of incident and maintenance deferral details, analyse information from various data sources, and derive the relationship among the incidents, maintenance deferrals, and inspection items. Finally, the System assigns weights based on the severity level to generate the normalised risk score of assets. The System also collects and analyses incident data to identify patterns and facilitate trend analysis, with a user-friendly search interface for the efficient retrieval of past cases, assisting in incident analysis. The System was successfully deployed in Electrical and Mechanical Services Department (EMSD), the Government of the Hong Kong Special Administrative Region (HKSAR) of the People’s Republic of China, demonstrating benefits such as an enhanced understanding of incident relationship, streamlined inspection planning, and proactive safety improvement. The System offers valuable guidance for wider application in railway systems, fostering improved safety, operational efficiency, and resource optimisation.

Keywords:

Railways; smart planning; risk-based prioritisation; artificial intelligence; data analytics; incident analysis; natural language processing; machine learning

Reference List:

  1. Alawad H and Kaewunruen S (2023). Unsupervised machine learning for managing safety accidents in railway stations. IEEE Access.
  2. Alawad H, Kaewunruen S and An M (2020). A deep learning approach towards railway safety risk assessment. IEEE Access, 8, pp. 102811-102832.
  3. Aven T (2010). Misconceptions of risk. Reliability Engineering & System Safety, 95(1), pp. 1-8.
  4. Commission Directive 2009/149/EC of the European Parliament and of the Council of 27 November 2009.
  5. Conneau A, Khandelwal K, Chaudhary NGV, Wenzek G, Guzman F, Grave E, Ott M, Zettlemoyer L and Stoyanov V (2020). Unsupervised Cross-lingual Representation Learning at Scale.
  6. International Civil Aviation Organization (ICAO). (2013). Safety Management Manual (SMM).
  7. Khan F and Abbasi T (2015). Risk-based inspection planning methodology for railways. Safety Science, 76, pp. 107-114.
  8. Lai KH and Cheng TCE (2016). A review of risk assessment methodologies in the railway industry. Journal of Safety Research, 57, pp. 39-48.
  9. Leitner B (2017). A General Model for Railway Systems Risk Assessment with the Use of Railway Accident Scenarios Analysis. Procedia Engineering, 187(2017), pp. 150-159. Elsevier
  10. Liu C and Yang S (2022). Using text mining to establish knowledge graph from accident/incident reports in risk assessment. Expert Systems with Applications, 207, pp. 117991.
  11. Podofillini L, Zio E and Vatn J (2006). Risk-informed optimisation of railway tracks inspection and maintenance procedures. Reliability Engineering & System Safety, Elsevier, 91(1), pp. 20-35.
  12. Raffel C, Shazeer N, Roberts A, Lee K, Narang S, Matena M, Zhou Y, Li W and Liu PJ (2020). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer.
  13. Railway Safety Directive 2004/49/EC of the European Parliament and of the Council of 29 April 2004.
  14. Seo M, Kembhavi A, Farhadi A and Hajishirzi H (2017). Bi-Directional Attention Flow for Machine Comprehension.
  15. Syeda KN, Shirazi SN, Naqvi SAA and Parkinson HJ (2017). Exploiting natural language processing for analysing railway incident reports. IGI Glob. Publ., pp. 1-18.
  16. Vladimir DV, Maria AB and Olga VI (2016). An Approach to Interactive Information Processing for Situation Awareness About Incidents in Railway Infrastructure Management System. Part of the Advances in Intelligent Systems and Computing book series (AISC), pp. 451.
  17. Wang Y, Zhu C, Guo Q and Ye Y (2023). A domain semantics-enhanced relation extraction model for identifying the railway safety risk. Complex & Intelligent Systems, pp. 1-15.
  18. Yamada I, Asai A, Shindo H, Takeda H and Matsumoto Y (2010). LUKE: Deep Contextualized Entity Representations with Entity-aware Selfattention.
  19. Zhou P and Ranjitkar P (2018). A review of risk assessment methods for railway infrastructure. Journal of Rail Transport Planning & Management, 8, pp. 1-14.
  20. Zhou PG (2011). Quantitative Approach in Regulating Railway Safety.
>> more<< less