A failure probability assessment method for train derailments in railway yards based on IFFTA and NGBN

Abstract

Derailment is one of the main hazards during train passes through railway turnouts (RTs) in classification yards. The complexity of the train-turnout system (TTS) and unfavorable operating conditions frequently cause freight wagons to derail at RTs. Secondary damages such as hazardous material spillage and train collisions can result in loss of life and property. Therefore, the primary goal is to assess the derailment risk and identify the root causes when trains pass through RTs in classification yards. To address this problem, this paper proposes a failure probability assessment approach that integrates intuitionistic fuzzy fault tree analysis (IFFTA) and Noisy or gate Bayesian network (NGBN) for quantifying the derailment risk at RTs. This method can handle the fact that the available information on the components of the TTS is imprecise, incomplete, and vague. The proposed methodology was tested through data analysis at Taiyuan North classification yard in China. The results demonstrate that the method can efficiently evaluate the derailment risk and identify key risk factors. To reduce the derailment risk at RTs and prevent secondary damage and injuries, measures such as optimizing turnout alignment, controlling impact between wagons, lubricating the rails, and regularly inspecting the turnout geometries can be implemented. By developing a risk-based model, this study connects theory with practice and provides insights that can help railway authorities better understand the impact of poor TTS conditions on train safety in classification yards.