A laboratory study of rail–wheel interaction monitoring using acoustic emission: Effect of rolling conditions with and without lateral rattling

Abstract

This paper presents part of an extended laboratory study on the potential to apply the acoustic emission (AE) technique for in-situ rail–wheel interaction monitoring using rail-mounted sensors. The essential monitoring principle is that the intensity of the rail–wheel contact will affect the intensity of the AE that is generated during rolling. The current paper is confined to situations of ‘normal’ rolling, investigating the effects of wheel load, speed and lateral rattling of the generated AE, although the wider study includes the effects of rail and wheel irregularities on the AE generated. Rail–wheel contact was simulated on a scaled test rig consisting of a single wheel rolled round a circular track using a motor mounted at the centre of the track attached to a driving arm. AE was recorded at a fixed point on the track while the wheel was rolled around the track at a range of speeds with varying axle load. Wheel slip was found to be insignificant and tests using a simulated source confirmed that the energy recorded with a source at a given position on the track was repeatable. In one set of experiments, measures were taken to eliminate lateral rattling of the wheel and eccentricity of the wheel arc relative to the track arc. In a second set of experiments, some lateral float of the wheel along the axle was allowed and a limited amount of eccentricity was built into the track. The energy per wheel rotation for normal rolling was found to increase in a linear fashion with increasing axle load and centrifugal force, as would be expected due to the increasing rolling resistance with axle load and increasing lateral force with angular speed. The frictional instability associated with the lateral rattling was clearly detectible at low axle loads and lower speeds, and both natural and eccentricity (flange rubbing) rail abnormalities were detectible in the AE record. It is concluded that, with appropriate calibration, contact stresses between wheel and rail on straight or curved track can be monitored using track-mounted sensors so that cumulative contact stress can potentially be monitored.