Development of a numerical tool for the computation of rail wheel – brake shoe thermo-mechanical interaction based on a 2D plane finite element model

During tread braking operations, a friction heat flux is generated at the wheel-shoe contact interface, which can increase the wheel surface temperature. The thermal flux flowing into the wheel is not uniform, as it depends on the distribution of the contact pressure. At the same time, the remaining portion of the wheel is cooled down by natural convection, radiation and rail chill effect, i.e., the cooling due to contact with the fresh rail. Therefore, the wheel is prone to thermo-mechanical stresses and strains, which can eventually damage the wheel surface, due to shelling and spalling. Furthermore, in special conditions, a microstructural transformation of the wheel steel can occur, and brittle martensite can be generated. Hence, it is essential to predict the wheel-shoe thermal behaviour in the frame of predictive maintenance as well as to increase the braking performances of new shoe materials. The present work deals with the description and preliminary validation of a new numerical tool for the simulation of the wheel-shoe thermomechanical behaviour, which was developed with the aim to find the best compromise between a detailed modelling of the phenomenon and computational efficiency. The tool includes a Matlab routine for the solution of the equations describing the dynamics of a braked railway wheel and two plane finite element (FE) modules, implemented in ANSYS Mechanical APDL. The first FE module performs a static structural analysis to calculate the distribution of the normal and tangential pressures at the contact interface, while the second FE module performs a transient thermal analysis to compute the evolution of the wheel temperature during a braking operation. To reduce the computational needs, the mechanical and thermal problems are decoupled and only a limited portion of the wheel is modelled in the thermal module, by superimposing adiabatic boundary conditions at the lateral edges. The present paper describes the modelling strategy adopted and the implementation of the FE modules in ANSYS, also giving light to the preliminary results obtained with the new code. The code validation shows a good agreement of the calculated temperature with results available in the literature. For drag braking operations, it is shown that the 1Bg configuration is more thermally harmful with respect to the 2Bg arrangement, and that at constant braking power an increase in the running speed reduces the wheel temperature thanks to air convection.