To solve this problem, a model is required to predict the degradation of tram tracks. In this chapter, this will be done by analysing and trending the tram track attributes/ variables over a period of time. Afterwards, the correlation between those variables and track deterioration will be identified and a model will be developed to predict the track deterioration/degradation based on tram track variables. The results of this chapter can be used to identify the maintenance and replacement activities and prepare a priority list for reparation of track rails. This will assist in minimising the maintenance costs and preventing unnecessary maintenance actions and therefore, saving time (Jovanovic, 2004; Larsson, 2004; Budai, Huisman, & Dekker, 2005; Guler, Jovanovic, & Evren, 2011; Caetano & Teixeira, 2013; Peng & Ouyang, 2013). Moreover, by preventing unnecessary maintenance actions, disturbance/ interruption of traffic will be reduced and the delay time experienced by private car drivers and tram passengers will be decreased (Shafahi & Hakhamaneshi, 2009; Liu, Xu, & Wang, 2010; Sadeghi, 2010; Sadeghi & Askarinejad, 2010; Peng & Ouyang, 2012; Wu, Flintsch, Ferreira, & Picado-santos, 2012; Zakeri & Shahriari, 2012; Yaghini et al., 2013; Tang, Boyles, & Jiang, 2015). This chapter aims to gain deep insights into the tram track (light rail) maintenance dataset, examine the relationship between various variables and understand how different factors might impact degradation of tram tracks for the first time in the literature. This chapter also develops a time dependent Artificial Neural Networks (ANN) model to predict degradation of tram tracks for the first time in the literature as a function of various external factors over time.
This chapter begins by explaining the data used in this study. Then, the variables influencing the degradation of tram tracks will be identified. It is followed by presenting an ANN model to predict the degradation of tram tracks. The final section summarises the insight from this study and identifies recommendations and directions for future research.