Film formation and friction in grease lubricated rolling-sliding contacts

Abstract

A vast majority of rolling element bearings are lubricated with grease. Despite this, the mechanisms of grease lubrication are still poorly understood, as a result of which bearing performance is commonly predicted using lubrication theories developed for lubricating oils. Under certain operating conditions this approach can lead to erroneous predictions of lubricant film thickness in bearings and potentially, to the selection of a wrong grease for a given application. The poor understanding of grease lubrication derives in large part from the complex composition of greases which are visco-elastic semi-solids, consisting of base oil, thickener and suitable lubricating additives. Systematic, lab-based studies are further hindered by the fact that it is difficult to produce small batches of a specific grease with properties that are representative of those in large, commercially made quantities. The research presented in this thesis studies the mechanisms of grease lubrication by employing custom-made greases whose composition is systematically varied so that the effects of the thickener and the base oil can be isolated. Grease film thickness and friction in rolling-sliding, non-conformal contacts were measured over a range of entrainment speeds, temperatures and surface roughnesses. Thus, results cover a wide range of λ ratios including very low values where any differences between grease and base oil may be expected to be most evident. Studies were conducted on both lab-based, single-contact rigs and real rolling bearings and the two sets of results related to each other. To assess the effect of grease composition, custom greases were made with two thickeners, lithium and di-urea and two base oil viscosities. All tested greases built thicker films and had lower friction then their corresponding base oils at low ‘nominal’ (that based on oil properties alone) specific film thickness values (λ ratios) i.e. low speeds, high temperatures and high roughness. Here, grease friction and film thickness were determined by the thickener type alone, the base oil viscosity and test temperature had practically no effect. At higher nominal λ ratios, the grease film thickness and friction were almost identical to those of the corresponding base oils and were therefore strongly influenced by oil viscosity and temperature with no effect of thickener type. The transition from the thickener-dominated to oil-dominated region for a given grease was seen to occur at a set value of film thickness. Furthermore, grease film thickness with urea greases in particular was shown to increase with time at a fixed entrainment speed and to persist under load after rubbing stops. Based on these observations it is here proposed that the mechanism of grease film formation is analogous to that of colloidal dispersions, namely low speed grease films are built by the mechanical entrapment and deposition of thickener fibres; such films are destroyed as speed, and hence the shear rate, increases so that the normal EHL oil films are built at high speeds. The transition from the thickener-dominated to oil-dominated region occurs when the prevailing film thickness at the imposed conditions exceeds the characteristic dimension of thickener fibres for a specific grease, rather than a fixed ‘transition speed’. It was shown that grease friction does not follow the standard Stribeck type behaviour when plotted against the nominal λ ratio, in line with other studies. However, grease friction does follow Stribeck type behaviour when plotted against the ‘true’ λ ratio calculated from the actual measurements of the grease film thickness. Finally, it was shown that within the range of test conditions employed here, single-contact measurements can be used to ascertain the relative trends in behaviour of different greases in real bearings and that, with application of suitable analyses, it appears possible to estimate the actual absolute values of the sliding component of bearing friction torque using friction measurements made in simple ball-on-disc tests.