Freight-train derailments represents a significant hazard to human safety and the economy. However, few studies have explored the contributory and interactive characteristics of risk factors influencing the economic costs of derailments across various track classes. We proposed a novel Choquet integral-based multivariate polynomial regression (CI-MPR) model to predict the economic cost via the states of four risk factors—number of derailed cars (DC), derailment speed (DS), loading factor (LO), and residual train length (RL). The results indicated that the CI-MPR model attained better predictions than the Choquet integral-based multiple linear regression (CI-MLR) model. The Shapley value and interaction indices were used to elucidate the contributory and interactive characteristics of these risk factors. Across Track Classes 1 to 5, DC emerged as the most influential risk factor, contributing 34.22%, 45.81%, 51.82%, 63.94%, and 64.74%, respectively. As for interactive characteristics, DC prevalently interacted with and in a superadditive way across all track classes. Notably, the pair (DC, LO) demonstrated critical superadditive interactions in Track Classes 1, 2, and 3, while appearing as additive and subadditive in Track Classes 4 and 5. The findings suggest that managing operational speeds and enhancing the stability of leading cars could significantly reduce economic losses. Moreover, controlling operational speed may prevent interactions with LO, potentially mitigating the adverse effects of the superadditive interaction between DS and LO in higher track classes. This study provides insights into identifying critical risk factors for freight-train derailment and devising track-class-specific mitigation strategies to reduce derailment severity and enhance operational safety.
A Novel Choquet Integral-Based Regression Approach for Analyzing Freight-Train Derailment Severity
Transportation Research Record: Journal of the Transportation Research Board
2025-04-04
Article (Journal)
Electronic Resource
English
Using Machine Learning Methods for Modeling Freight Train Derailment Severity
Transportation Research Record | 2022
|Freight-train derailment rates for railroad safety and risk analysis
Online Contents | 2017
|Study on Critical Speed of Freight Train Derailment on Bridges
Tema Archive | 2013
|Study on Critical Speed of Freight Train Derailment on Bridges
Trans Tech Publications | 2013
|