Experimental study of the use of a transfer function to find rail corrugation from axle-box accelerations

Abstract

This investigation uses a scale vehicle-track experimental facility to study the calculation of rail corrugation using vertical accelerations measured in the axle-box of rail vehicles and a transfer function (TF). The rail corrugated profile is machined in the rail heads of the scale track following a periodic function with four harmonics. Experiments are performed with a scale bogie-like vehicle at different forward velocities in the range inspection velocities. Two simple analytical forms of the TF are studied: the kinematic TF, that assumes that the axle box follows the rail profile, and the TF of a 2-dof model of the vehicle-track system. For the vehicle response analysis, this work proposes to normalize the measured acceleration with the square of the forward velocity of the vehicle, that is assumed to be approximately constant. This normalized acceleration reduces the effect of the forward velocity on the TF. Experimental results show that the kinematic TF can be used to measure the track corrugation for moderate forward velocities providing reasonable but not accurate results. The limitation of the kinematic TF is mainly due to free flights and wheel rail curvature incompatibility. The measured axle-box accelerations may include frequency peaks that are not excitation frequencies and can distort the rail profile measurement. Results show that linear elastic models like the assumed 2-dof model do not explain the appearance of these non-excitation peaks.