Degradation and performance specification of Melbourne tram tracks

The increase in passengers and tramcars will wear down existing rail structures faster. This is forcing the rail infrastructure asset owners to incorporate asset management strategies to reduce total operating cost of maintenance whilst improving safety and performance. Analyzing track geometry defects is critical to plan a proactive maintenance strategy in short and long term. Repairing and maintaining the appropriate number, types and combination of geo-defects can effectively reduce the capital cost of maintenance operations. The main objective of this research is to explore maintenance needs regarding tram track geometry defects based on formulating the following two sets of data driven models: (1) non-parametric track deterioration models that are used to compare and capture the deterioration process of different track types, i.e. straight track, curved track, H-crossing and crossover. (2) Semi-parametric track deterioration models to assess probability of maintenance needs in the most critical segments as a function of explanatory variables such as: track surface, rail profile, rail type, rail support, maintenance, and traffic condition. In this paper, the proposed models are applied to geometry defects from inspection runs occurring seven times, from December 2009 to November 2013. In the first step, among Melbourne tram network routes, three routes are selected as case study. Non-parametric deterioration models shows curve type tracks and H-crossing have the highest probability of failure. This analysis shows that curved tracks are the most critical segments of the network. In the next step, the Cox regression procedure is used for modeling the time to exceeding the failure limits for curved tracks. The covariates that are used to estimate the probability of failure consists of curve radius, rail support (concrete/sleeper), rail type (Grooved or T-shapes), rail profile, track surface (asphalt/concrete), and traffic conditions.