Renewals scheduling for high-speed railway assets
Publication Date
July 31, 2026
Creator
Buttriss, Emily
Abstract
The increasing focus on reducing carbon emissions and easing traffic congestion means that the demand for rail services, particularly high-speed rail, grows ever higher. The cost of ensuring that the railway line remains in a safe and acceptable condition is reflected in the price a passenger pays for travel, hence an efficient asset management strategy should predict when various actions need to be scheduled so that appropriate funds can be provided when needed without overcharging for use of the service. One of the costliest actions is replacement, which requires an in-depth understanding of how the track degrades and what maintenance actions can be completed to improve its condition in order to discern when the useful life has been exhausted and a renewal should be completed. Modelling tools have been developed with aims to improve understanding of the life cycle of a track component, however a review of the surrounding literature revealed that the models rarely distinguished between different failures affecting the track, and the implemented renewal strategies were too simplistic.
This research aims to develop a methodology that can address these limitations by modelling the behaviour of four track assets: crossing panel, switch panel, plain line rail, and the ballast. The degradation process was studied through a reliability analysis completed on historical maintenance and condition databases of a high-speed line situated in the UK. The analysis fitted continuous probability distributions to time-to-failure datasets, to predict future degradation events. The mean time-to-failure and associated parameters of each distribution highlighted the most common failures affecting each track component and how the degradation rate of the fault was expected to change across the operational period. Railway standards were used to develop a Petri net model to represent the life cycle of the four components. Due to the complexities of asset management procedures, additional features such as place dependent transitions or inhibitor arcs were included within the structure. The methodology created sub-models for each of the inspection, degradation, maintenance, and renewal processes, with an individual degradation sub-model created for each failure type, and the ability to choose between a combination of five criteria for the implemented renewal strategy. The results from the previous reliability analysis were incorporated into the Petri net transitions, and Monte Carlo simulation was used to solve the models, to predict when a replacement is required. The models demonstrated their usefulness in investigating the renewal process, including the importance of each criterion for triggering when a replacement is scheduled or how varying the renewal threshold influences the expected lifetime.
The final stage of the research applied an artificial neural network to understand how the different features of the track can affect degradation and ultimately when a renewal is expected to be scheduled. A variety of training parameters, input variables, and data configurations were investigated but ultimately the models were unable to find a significant relationship between the characteristics of an asset unit and the degradation rate of a fault. A review into the behaviour of the studied railway line concluded that the track was too young and in too good a condition for a clear degradation trend to be established, causing the high prediction errors from the models.
Item Type
ethesis
Thesis Type
PhD
Supervisors
Andrews, John
Remenyte-prescott Rasa
Subjects (LC)
Associated Schools / Departments
Author URL(s)
UoN Repository URI
Except where otherwise noted, this item's license is described as
File(s)![Thumbnail Image]()
Name
Emily Buttriss thesis.pdf
Size
6.58 MB
Format
Adobe PDF
Checksum (MD5)
9f19d261b24edb0bb1df43a6466fdae2