Rolling contact fatigue behaviour of rails: Wedge model predictions in T-Gamma world

Abstract

In this study, T-Gamma and Wedge models have been compared with each other for the prediction of surface-initiated rolling contact fatigue cracks on rail surfaces. Both models are able to account for different observed rolling contact fatigue-wear regimes in tracks, but with very different physical backgrounds. The T-Gamma model uses empirically determined damage functions by introducing a relationship between the wear number (T-Gamma) and the rolling contact fatigue damage increment. Different rolling contact fatigue-wear regimes are considered in this empirical approach based on the idea that initiated cracks get partially or fully removed by the wear mechanism, not accounting for the full complexity of the occurring tribological phenomena. The Wedge model represents a physical approach, where contact stresses and its impact on plastic deformations and related material anisotropy are considered. Thus, the prediction of different rolling contact fatigue-wear regimes is based on these physical relationships, where plastic shear deformations in the near-surface layer play a key role. For comparison, the wheel–rail contact data from stochastic multibody dynamics simulations of a metro vehicle with conventional bogie technology running in three curve radii have been used. While the T-Gamma model always predicts the same rolling contact fatigue damage increment for a given T-Gamma value, the Wedge model shows a scattering of the predicted rolling contact fatigue damage increments when plotting them over T-Gamma because of the explicit consideration of contact stresses. Thus, each scenario consisting, for example, of certain vehicles, curve radius, wheel–rail profile combination, friction conditions, rail material, etc. needs its own damage function in the T-Gamma world. This should be kept in mind when applying the standard T-Gamma model to scenarios which differ significantly from the scenario it has been parameterised for.