Improving the capacity and performance assessment of railway nodes on the French network

Abstract

Capacity and performance (evaluated through delay propagation) analysis methods have mostly focused on railway line track sections, but less attention has been given to nodes. Still, a few analytical methods for the capacity and performance assessment of the switch areas between station platform tracks and line tracks can be classified upon their reliance on a timetable (“timetable-based” methods) or not (“timetable-free” methods). The relation between capacity utilisation and performance has rarely been tested for railway nodes, and critical capacity utilisation thresholds remain to be investigated. The comparison of timetable-based and timetable-free methods also needs to be conducted. Filling these knowledge gaps will help the French infrastructure manager SNCF Réseau improve its analyses of nodes capacity utilisation and performance in the long-term planning stages. This paper investigates a small set of timetable-based and timetable-free methods either taken and adapted from the literature, such as the Potthoff and UIC 406 methods, or developed for the need of this research, such as adaptations of the UIC method and a method developed from SNCF Réseau’s previous works. The methods are applied on a case study on the French network, first evaluating their indicators’ magnitude and trends with artificial traffic data, and then comparing their outputs to real data. It is found that the Potthoff method and a timetable-free UIC-adapted method for capacity utilisation evaluation and an SNCF-adapted method for delay propagation provided results that are relevant in terms of magnitude and trends for long-term assessment. The timetable-based UIC 406 method for node capacity utilisation assessment can be used to study specific timetables. No satisfactory timetable-based delay propagation method was found in this paper. The timetable-free methods are further used to study the capacity utilisation – delay propagation relationship, which takes the form of an exponential function. Attempts to determine capacity utilisation thresholds are also conducted. It is recommended to perform further research with extended traffic data on different node layouts to consolidate these preliminary findings before applying them in real studies.