Analytical forecasting of long-term railway track settlement

Railway tracks undergo plastic settlement when subject to repeated train loading. This occurs differentially along the track rather than in a uniform manner, and the profile is a key parameter when scheduling track maintenance operations. Therefore this paper presents a novel numerical approach to predict track irregularity evolution. The model combines empirical settlement laws with finite element theory, where the track-ground structure is modelled explicitly, and multi-body train-track interaction is considered. The stresses induced by rolling stock are solved using a hybrid frequency-wavenumber and time-space approach, considering non-linear track-soil material behaviour. It has several novelties: 1) after every load passage, the track profile is updated before applying the next load, meaning the train-track interaction is constantly evolving; 2) new empirical settlement laws are derived that account for evolving train-track forces and track profiles; 3) fully 3D stress fields in the track and ground are considered. First the model is described, before validating its prediction of track geometry evolution, captured from track recording vehicles. Next, it's used to show that modelling error is introduced if the geometry isn't updated frequently (e.g. after every load passage). Finally, a parametric study shows track subgrade material properties have a marked effect on track settlement.