ABSTRACT
Railway tracks experience different settlement over time due to repeated train passages. Typically, when the standard deviation of this vertical profile exceeds a threshold limit, maintenance (often tamping) is required to correct the profile. Therefore, when planning future maintenance and attempting to optimise track life-cycle cost, it is important to know when future maintenance will be required. To predict this, this paper presents a novel numerical approach to calculate track irregularity evolution. It is particularly useful for scenarios where it is difficult to extrapolate historical standard deviation changes into the future, such as new track constructions or increased freight tonnage. The model combines empirical settlement laws with finite elements for the track and ground. Perfectly matched layers are used to absorb outgoing wave energy, and a multi-body rolling-stock model is used for train-track interaction. The model is solved using a combination of frequency-wavenumber and time-space domains, to give a highly efficient solution approach. This optimised solution procedure allows for the track profile to be updated after every axle passage, before applying the next load, meaning the train-track interaction is constantly evolving like on a real track. This is highly novel considering the models ability to also calculate 3D stress fields. After showing model development, it is used to perform a parametric study that shows track subgrade material properties have a marked effect on track differential settlement.
