Some important issues addressed by the project NGT are directly related to the wheel/rail interface: it is the origin of rolling noise, the major noise source below 200km/h, wear and the associated maintenance costs are generated there and running safety as well relies on the wheel/rail forces, which therefore are the relevant quantities for the homologation of a railway vehicle.
These problems are the essential motivation of our approach to extend and to improve the modelling of the wheel/rail interface by consideration in particular the structural dynamics of the wheel or wheel-sets and the rail or track system, respectively.
Exemplary wheel-set bending mode that corresponds to 83Hz eigenfrequency
To become descriptive, this means that e.g. the bending modes of a wheel-set such as shown in the figure above or the deformation behaviour of the rail visualized in the figure below are introduced instead of assuming both contact partners to be rigid as it is the state-of-the-art today.
Structure of the track modeland 2 exemplary deformation modes of the rails
We could show that the so-called hunting motion, a dangerous running state of high-speed-trains, may occur at lower velocities, if the structural elasticity of wheel and rail are taken into account, i.e. neglecting the elasticity leads to a smaller reserve of safety related to the running behaviour.
We propose to use the so-called Arbitrary-Eulerian-Lagrange (ALE) description to describe the deformation of the rotating wheels. This approach inter alia facilitates the modelling of rotor-like structures with non-rotating contact forces, as theyare present at the wheel/rail interface, but also e.g. at the brake-disc/pad- or the lathe/workpiece interface.
On contrary to other ALE-approaches inspired by fluid dynamics, we exploit the specific properties of axially symmetric or axially cyclic structures, so that the resulting formulation exposes a high computational efficiency, which is a requirement for the industrial application of the method.
In addition to the consideration of the structural dynamics the contact modelling itself also has to be refined in order to reasonably adjust the different model components. Due the geometrical shape of wheel and rail, the wheel/rail interface has a non-elliptic contact patch, see figure below, which may be solved using a fast iterative algorithm.
Normal pressure distribution at the contact patch between wheel an rail.