A Segment-Based Model for Estimating the Service Life of Railway Rails

Mastering rail degradation is of great importance for railway infrastructure managers estimating the service and residual life of rails, making rational decisions about maintenance and repair to ensure the safety of trains. Rails have different degradation rules at different locations, because they are susceptible to heterogeneity factors such as the geographical environment, and transportation organization. In this paper, the authors propose a novel segment-based model, in which a continuous railway line is divided into several adjacent 1-kilometer-long segments, and the number of severely defective rails per kilometer is applied to quantify the segment condition. Employing the theory of Markov’s stochastic process, the segment deterioration law is evaluated. Parameters of the proposed model are estimated via a maximum log-likelihood method. The proposed model is verified with the measurement data obtained by rail flaw detector cars. The authors' evaluation demonstrates that the proposed model outperforms the traditional method based on accumulated gross tonnages.