Modelling of wheel/rail interaction considering roughness and discrete irregularities

Licentiatavhandling, 2008

The interaction between wheel and rail is the predominant source of noise emission from railway operations in a wide range of conventional speeds. On the one hand, this wheel/rail noise concerns rolling noise and impact noise caused by the vertical interaction excited by roughness and discrete irregularities of the wheel/rail running surfaces, respectively. On the other hand, it concerns squeal noise generated by the tangential interaction. While a well-established frequency-domain model is available for the prediction of rolling noise, likewise successful models do not yet exist for the prediction of impact noise and squeal noise. The overall aim of this thesis is to develop a model for the vertical wheel/rail interaction induced by roughness and discrete irregularities. The inclusion of tangential interaction is a subject of future work. In order to include the non-linearities in the contact zone, the interaction model presented in this thesis is formulated in the time domain. Wheel and track models are represented by Green’s functions, which leads to a computationally efficient formulation and allows inclusion of detailed non-Hertzian contact models. The first contact model considered is a two-dimensional (2D) model consisting of a bedding of independent springs. This model uses a simplified wheel and rail geometry and takes into account one line of wheel/rail roughness in the rolling direction. The second contact model is a three-dimensional (3D) model based on an influence-function method for the elastic half-space. This model considers the real three-dimensional wheel and rail geometry and includes the roughness in several parallel lines. In the thesis, the interaction model using both the 2D and the 3D contact models is applied to simulate the wheel/rail interaction caused by parametric excitation on a discretely supported rail and by wheel/rail roughness. The results indicate that the application of the 3D contact model is preferable when the degree of correlation between roughness across the width of the contact is low, although more simulations should be carried out before drawing a final conclusion. The interaction model using the 2D contact model is applied to simulate impact forces caused by wheel flats and shows encouraging agreement with field measurements.