Railway Track Geometry Defect Modeling: Deterioration, Derailment Risk, and Optimal Repair

Analyzing track geometry defects is critical for safe and effective railway transportation. Repairing the right number and type of track geo-defects can appropriately reduce the probability of derailments. Additionally, prioritized track geometry repair work reduces dynamic vehicle and track interaction, thus reducing the stress state of the railroad. In this paper, the authors propose an analytical framework for making optimal geo-defect repair decisions by minimizing total expected costs, which include potential derailment costs and repair costs. This paper's major contribution lies in formulating and integrating the following three data-driven models: 1). A track deterioration model to study the degradation of Class II geo-defects; 2). A survival model to assess the derailment risk as a function of the track condition; 3). An optimization model under uncertainty for track repair decisions. In real-world examples, compared with heuristic strategies in practice, our proposed models can reduce 20% of the total composite cost on average, and potentially even more for long track sections.