Creep curve measurement to support wear and adhesion modelling, using a continuously variable creep twin disc machine

Predictive modelling of wear and adhesion at rolling–sliding contacts such as a railway rail and wheel depends on understanding the relationship between slip and shear force at the contact surface, i.e., the creep verses force curve. This paper describes a new approach to creep curve measurement using a twin disc machine running with a continuous programmed variation of creep, enabling an entire creep curve to be defined in a single experiment. The work focuses on very low levels of creep, ranging from zero to 1%, and shows clear correlation between the creep curve gradient and the full slip friction coefficient for dry and lubricated contacts. Comparison of data generated using the new approach with that generated using multiple tests each at a single creep level shows good agreement. Comparison is also made between the twin disc data and results for full size three dimensional rail–wheel contacts to examine how two and three dimensional contact adhesion data are related. The data generated has application in wear and rolling contact fatigue modelling, but the original motivation for the research was generation of creep curves to support prediction of low adhesion conditions at the rail–wheel interface based upon monitored running conditions prior to brake application. The range of contact conditions investigated includes those experienced in service and during driver training, with the correlation found between creep curve gradient (measurable prior to braking) and full slip friction coefficient (not measurable until brakes are applied) representing a key finding. ⺠Creep curve measurement using a twin disc machine with a continuous programmed variation of creep. ⺠Findings applicable to wear, rolling contact fatigue or adhesion in rolling–sliding contacts. ⺠Entire creep characteristic can be defined in a single experiment. ⺠Data generated for rail–wheel contact conditions at very low creep, ranging from zero to 1%. ⺠Identification of low adhesion through correlation of creep and friction coefficients.