Causality-Network-Based Critical Hazard Identification for Railway Accident Prevention: Complex Network-Based Model...

Area of Science:

• Railway Safety Engineering
• Risk Management
• Systems Engineering

Background:

• Railway accidents pose significant safety risks.
• Effective hazard identification is crucial for prevention.
• Existing models may not fully capture hazard interactions.

Purpose of the Study:

• To develop a novel method for critical hazard identification in railway accident prevention.
• To model the complex interactions between hazards and accidents.
• To identify key factors for effective accident mitigation.

Main Methods:

• Proposed a new accident causation network model.
• Developed a method for measuring node distance to align path lengths.
• Introduced a critical hazard identification model based on controllability analysis.
• Evaluated five critical hazard identification methods.

Main Results:

• Transformed the most-likely causation path problem into a shortest causation path problem.
• The weighted directed accident causality network considering length demonstrated superior performance.
• Integer programming-based critical hazard identification combined with the proposed network showed the best results for accident prevention.

Conclusions:

• The proposed accident causation network and controllability analysis provide an effective framework for hazard identification.
• The integrated approach offers a robust solution for railway accident prevention.
• Prioritizing identified critical hazards can significantly enhance railway safety.