Wear at the wheel/rail interface when sanding is used to increase adhesion

Sanding is used in train operation to improve adhesion at the wheel/rail interface during both braking and traction. An experimental study has been carried out to determine the effect of sanding on the wear of wheel and rail materials. Static tests were performed using actual wheel and rail sections. Dynamic tests were carried out with and without sand on a twin disc machine, where wheel and rail steel discs are loaded together and driven under controlled conditions of rolling and sliding. Sand was fed into the disc contact through a standard compressed air sanding valve. In both static and dynamic tests, sand caused severe surface damage. During the dynamic tests, because of the application of sand, wear increased by factors between 2 and 10. The wheel steel wear rates showed the largest increases. Wear in wet conditions was higher than that in dry conditions because the wet discs entrained a larger amount of sand through the contact that otherwise was ejected when the discs were run dry. The mechanisms of sanding wear have been investigated. Severe plastic flow as well as a high material removal rate and surface corrugation takes place. The rail discs are scored by sand particles that embed and stick in the softer wheel disc. As well as wear by abrasion, the discs were subject to a rapid fatigue process and large chunks of material fractured from the surface. A simple abrasive wear model has been developed to predict wear of rail material caused by sand in the wheel/rail contact, which shows good correlation with test results. There are a number of idealizations inherent in the test simulations that lead to increased severity over the actual wheel/rail contact. These include the amount of sand entering the contacts in both types of test and the disc geometry and motion in the dynamic tests. While the twin disc specimens have been scaled down, the sand is as used on the railway network, which may lead to surface damage appearing more severe than it would be in an actual wheel/rail contact. Results are therefore only to be taken as a guide to what happens in the full size wheel/rail interface.