Towards a digital twin for prediction of rail damage evolution in railway curves
Licentiate thesis, 2023

Railway maintenance incurs significant expenses, and reducing the costs while maintaining safety levels and functionality is of great industrial and societal interest. Accurately predicting rail damage is crucial for cost-effective maintenance and extending rail life. This thesis focuses on developing a digital twin that can predict the rail damage evolution in a curved track under operational traffic, considering plastic deformation, wear, and surface Rolling Contact Fatigue (RCF) by accurate and fast simulations.

The first part of the thesis contains results from long-term simulations of rail damage for a given traffic scenario. The adopted simulation methodology is calibrated and validated using field tests. The numerical methodology in Paper A employs a framework that incorporates multiple steps that are applied iteratively to predict rail head degradation. The different steps are (1) the dynamic vehicle-track interaction for a given load sequence, (2) elastic-plastic wheel-rail contact, (3) accumulative rail damage due to cyclic plasticity, wear, and surface RCF, and (4) rail profile update from the accumulated wear and plasticity. Thereafter, the simulation steps can be repeated. It is demonstrated how the numerical model can be calibrated to match field measurements of the averaged geometry changes of the rail cross-section and predict surface RCF crack initiation. To allow for better predictions of the local wear distribution on the rail head, a model sensitivity study is performed in Paper B to determine the most significant model parameters for vehicle and track. Also, the effect of different contact modeling approaches is investigated. By including a freight vehicle, using different measured wheel profile samples in the load sequence, and considering a different contact modeling approach, promising predictions of the wear distribution are obtained.

Furthermore, a reduced-order model based on Proper Generalized Decomposition (PGD) is developed in Paper C for accurately simulating the deformation in the rail under varying contact loads. It enables cost-efficient simulations required for the digital twin framework. As a first step towards replacing computationally demanding nonlinear finite element simulations for the plastic deformation evaluation, the current version of the PGD model involves a domain decomposition of a three-dimensional elastic rail head combined with a parameterized discrete load to account for different load scenarios in an automated fashion. It is shown that the three-dimensional PGD model can be computed with two-dimensional complexity and can generate the parametrized displacement field for various loading scenarios.

Finite Element Modelling (FEM)

model reduction


Rolling Contact Fatigue (RCF)

Proper Generalized Decomposition (PGD)


dynamic vehicle-track interaction

digital twin


Virtual Development Laboratory Chalmers Tvärgata 4C
Opponent: Professor Mathias Wallin, Lund University, (Online presentation password: 508398)


Caroline Ansin

Chalmers, Industrial and Materials Science, Material and Computational Mechanics

Simulation and Field Measurements of the Long-term Rail Surface Damage due to Plasticity, Wear and Surface Rolling Contact Fatigue Cracks in a Curve

CM 2022 - 12th International Conference on Contact Mechanics and Wear of Rail/Wheel Systems, Conference Proceedings,; (2022)p. 591-601

Paper in proceeding

C. Ansin and B. Pålsson. "Influence of model parameters on the predicted rail profile wear distribution in a curve"

C. Ansin, F. Larsson and R. Larsson. "Fast Simulation of 3D Elastic Wheel-Rail Contact Using Proper Generalized Decomposition"

Driving Forces

Sustainable development

Areas of Advance


Materials Science

Subject Categories

Applied Mechanics

Other Materials Engineering



Virtual Development Laboratory Chalmers Tvärgata 4C


Opponent: Professor Mathias Wallin, Lund University, (Online presentation password: 508398)

More information


5/9/2023 1