An Attempt Towards FE-Modelling of Fracture Propagation in Railway Wheels

Abstract: The demand for higher velocities and heavier axle loads for freight trains leads to higher forces on the railway wheels which in turn lead to an increase in stresses on and below the surface of the wheel-rail contact. By time, this induces wear on the wheels which consequently lead to higher maintenance costs and in some cases accidents. The ability to predict the evolution of wheel profiles due to uniform wear has been demonstrated with a rather accurate precision in most operational conditions. These wear models are based on wear coefficients and since they are not usually valid for real operational conditions, the models are generally calibrated against real-life scenarios in order to adjust the coefficients from test conditions to real-life lubrication conditions. This engineering approach can be useful in prediction of wear in systems where the materials and contact conditions do not vary. However, when addressing material development focused on reducing specific damage modes, the approach is of limited use because the obtained wear coefficients are not directly related to material properties. Therefore, attempts towards developing physical fracture propagation models that relates to the contact conditions and material properties have been made. The purpose has been to retrieve vital information about where a fracture initiates and how it propagates. In the long run, it is of great interest to be able to attain information about how a material particle is removed from the contact surface. Studies for this type of model was done in the 70’s and 80’s mainly with pin-disk experiments but has not been utilized in the specific field of wheel-rail contact. The thesis is part of the FR8RAIL project arranged by the European rail initiative Shift2Rail. Literature studies have been the basis for the thesis in order to gain vital insights into fracture mechanics and other related fields. The physical fracture propagation models have been constructed in the FE software Abaqus with the implementation of the XFEM. For the 2D model, the fracture initiates at the top of the implanted inclusion when the friction coefficient is  and propagates upwards a few elements. For , the fracture initiates at the right surface boundary where the pressure distribution and traction is applied. The fracture propagation angle increases relative to the surface as the friction coefficient value is increased. The fracture for the 3D model extends broader compared to the 2D model at the top of the inclusion in the case of . The fracture initiates at the same surface location as for the 2D model for . The fracture propagation is however non-existent due to convergence problems. The FE-models constructed are initial steps towards analysing the fracture propagation and closely related phenomena for a railway freight wheel in detail. At the end of the thesis, the simplified models give mainly information about the fracture initiation, propagation and its patterns. From this first phase, further adjustments and improvements can take place in order to eliminate the margins of error. In the long run, fully integrated models with further implementations such as detailed microstructure for the contact conditions, plastic behaviour for the material, and complete three-dimensional models can finally be employed.