Rail Degradation Predication: Melbourne Tram System Case Study

Nowadays, tram as an accessible and convenient mode of public transport is implemented and used in different cities. Due to high frequency of accelerations and decelerations along their routes and sharing the route with other vehicles, the rate of degradation of tram tracks (light rail) is different from the degradation rate of train tracks (heavy rail). In this paper, track gauge deviation as an indicator of irregularities on the rail-wheel contact surface is used for developing the track degradation model. Data set used in this study includes the curve sections of Melbourne tram network and divided into repaired and unrepaired segments. For data analysis more than 13 km of curved tracks are examined. Annual tonnage, previous gauge deviation and track structural properties like track surface, rail support, rail profile and gauge deviation are considered as the influencing variables on the future gauge deviation. Two different models including a regression model and an Artificial Neural Networks (ANN) model have been developed for predicting tram track gauge deviation. According to the results, the performances of the regression models are not very different from the ANN models. The determination coefficients of the selected models are approximately 0.8 and higher.