A study into the relationship between surface roughness characteristics and sliding friction in nominally dry contacts

This thesis investigates the influence of surface topography on sliding friction in nominally dry, rough contacts of hard steels. A Bowden-Leben type rig was extensively rebuilt to allow for detailed studies of frictional behaviour on asperity level. Bearing steel components were manufactured with realistic, machined, rough surfaces to cover a wide range of spatial and height characteristics and tested in terms of their frictional and wear performance in pure sliding under different conditions of elastic-plastic contact. In parallel, numerical models for rough surface contacts were used to predict the real area of contact and pressure distributions. An attempt was then made to relate the observed frictional trends to these contact parameters. Finally, the most suitable roughness functional parameters, calculated from the height and wavelength content of the tested surfaces, in terms of their correlation to the measured friction were identified. The work is fundamental in nature but the practical application that motivated this study is the performance of contact between the inner ring of a rolling bearing and the shafts, where friction has an important effect on mounting/dismounting forces and the potential for fretting damage and creep of the inner ring. Such damage can lead to expensive maintenance and even critical bearing failures. One way to reduce the risk for relative motion between the components, and hence prevent fretting and creep, is to control the coefficient of friction by tailored surface topography of the bearing ring, hence the link to the current study which attempts to relate characteristics of machined surface roughness to sliding friction. Surface topography and its influence on coefficient of friction has been widely studied and argued in the past, including classical works of Bowden and Tabor in the 1950s, the plasticity work of Hirst and Hollander and Archard’s asperity-on-asperity representation of a rough surface contact. However, most available literature on this topic is based on Gaussian surfaces, considers purely plastic or elastic contacts and uses parameter descriptions limited to a single roughness scale with lack of complete description of the surface characterisations. This study employs a much wider range of surfaces than was done in the past, with very different surface roughness structures, and attempts to use rough contact simulations to better understand the observed trends. The experimental results show that the friction coefficient is not constant for a given pair of contacting materials and set contact conditions, but varies with the roughness structure of the surfaces. The observed trends were analysed and compared to basic laws of friction. It was shown that the existence of the proportionality relationship between friction force and real area of contact, and hence between the friction force and the applied load, is strongly influenced by the level of plasticity in the contact. The level of plasticity is in turn influenced by the roughness structure and the hardness of the surfaces. When extensive plasticity occurs in the contact, the friction coefficient is nearly constant regardless of roughness and the friction force is directly proportional to the predicted real area of contact. However, this relationship breaks down for a relatively wide range of surfaces under conditions of predominantly elastic contact. Poor correlation was observed between friction and simple roughness parameters, such as Rq. A few new functional roughness characterisations were linked to contact area, but often perform less well for friction. In other words, using measured surface characteristics it is easier to predict contact area than it is to predict friction. It was also found that different test conditions require different roughness aspects to describe friction. It was seen that wear and friction are intricately related to the multi-scale nature of surface topography including; asperity shape, magnitude and wavelength content. New surface characterisations were developed that better link the different aspects of surfaces to friction. Generally two or more parameters are required for this functional roughness characterisation to correlate well with measured friction.Open Acces