Relation between the geometry of rail welds and the dynamic wheel - rail response: Numerical simulations for measured welds

The primary mechanisms playing a role in the dynamic wheel-rail response to rail weld irregularities in a ballasted track are pointed out. The concept of P 1 and P 2 forces for metallurgical rail welds, introduced in a first paper [1] concerning the present research, is further elaborated. The dynamic wheel-rail response is simulated for a number of geometrical rail weld measurements. Results show a good correlation between the gradient of the rail weld geometry and the maximum dynamic wheel-rail contact forces, whereas the correlation with vertical peak deviations is shown to be very poor. Therefore, an assessment method based on the gradient (introducing a speed-dependent quality index [1]) is more consistent than a method based on vertical tolerances. An approximate formula is presented to calculate the maximum dynamic wheel-rail contact force as a function of the train velocity and the maximum gradient of the weld geometry, in analogy to Jenkins' formulae for calculating P 1 and P 2 forces at dipped rail joints.