Project Description

The current preventive maintenance for railway track is based on inspection and on an assessment of deterioration to calculate the maintenance requirements. As a consequence of the predictions, and to restore geometric quality of the track, maintenance machines and respective crews are sent to locations to undertake maintenance tasks. It is known that these tasks are very costly because they involve specialized crews and heavy equipment that, usually, are working overnight. Also, the means of intervention require expensive travelling times from their bases and the track is occupied while maintenance machines are travelling and maintaining the track, resulting in alterations for the train traffic. However, these specific track maintenance topics were not explored extensively in previous research, especially in what concerns a detailed tactical and operational analysis of the steps required by the machines to access the intervention sites. In this sense, the current work proposes a method to plan and schedule heavy maintenance interventions during night periods, where both travel costs and number of workdays on track are minimized, and at the same time ensuring that an adequate track quality is taken into account. With the proposed models, it was possible to evaluate key elements that contribute to the high cost of the heavy maintenance in track geometry correction.

The starting point of the current work was to establish an optimal formulation of the planning/scheduling problem, resulting in a complex structure which involved a considerable number of constraints. Two heuristic models were then created that validated and enlarged the scope of analysis of the optimal formulation. However, the reduced limits in terms of temporal horizon (one month) of the optimal and heuristic models, did not allow one to draw relevant conclusions in the medium and long term. Thus, a decision rules model was introduced, which was different from the previous models in the sense that no optimisation was used, but was able to provide solutions by following a set of rules created, using as a base the resolution method of the previous models. In this way, planning/scheduling solutions for annual periods were reached and, considering wider temporal scenarios, it was possible to observe the influence of the various factors that impact on the maintenance costs.

It was concluded that the problem of heavy maintenance planning for the task of the correction of track geometry and as applied to the real context of the Portuguese Northern Line, can be reduced to determining the frequency of full traverses of the line for maintenance, for a given temporal horizon. Following the decision rules model and adopting different work intensities and minimum maintenance cycles, significant reductions in the unit maintenance cost were obtained. Finally, through a sensitivity analysis, it was possible to understand the relevance of various maintenance parameters. From usual values and by making individual changes to those parameters, for instance, the number of days worked by week, the hours available during the night-time track possession, the performance of the machines while maintaining the track or travelling speed and in the availability of more sidings, it was possible to observe that the changes can contribute individually to savings of up to 12% in total annual maintenance costs.

Research Team

University of Lisbon

• Rui Santos
• Paulo Teixeira (supervisor)

CITTA

• António Pais Antunes (supervisor)

Financial Support

• FCT

Stage of Progress

• Finished in 2015