Estimation of gauge corner and flange root degradation from rail, wheel and track geometries

A methodology to predict gauge corner and flange root degradation in terms of rolling contact fatigue (RCF) and wear from measured rail, wheel and track geometries by so-called meta-models is presented. The methodology sets out from a parametrisation of the gauge corner and the flange root geometries. Then statistics of measured rail and wheel profiles are employed in a experimental design methodology to create scenarios, which are employed in multibody simulations from where degradation indices are evaluated. Finally, classification and regression analysis is used to derive decision boundaries and meta-models. These predict whether contact will occur at the gauge corner and (when this is the case) quantify RCF and wear impacts. The derived meta-models are used to rank measured profiles based on how detrimental they are. Robustness is assessed through comparisons between two vehicle types. The ability to rank measured profiles without the need for computationally expensive simulations (including the need for additional presumptions that such simulations require) makes the study valuable e.g. for maintenance planning.