Traction prediction in rolling/sliding EHL contacts with reference fluids

Abstract

Machine elements such as rolling bearings and gears transmit forces and permit relative motion in concentrated contacts, whereby elastohydrodynamic lubrication (EHL) plays a major role in surface protection. The friction/traction in a rolling/sliding EHL contact is hard to predict due to non-Newtonian rheology and concomitant thermal effects. In the last decade, much effort has been made to study the EHL traction using reference fluids. However, considerable discrepancies still exist between predictions and measurements.

This work continues the effort to predict the EHL traction with model fluids (mainly with squalane) and investigates the influence factors that lead to the differences between simulations and experiments. An EHL model has been developed for traction prediction accounting for non-Newtonian and thermal effects by embedding fluid models of thermo-physical-rheological properties (such as viscosity, thermal conductivity, shear thinning, and limiting shear stress) supported by independent high-pressure measurements. On the experimental aspects, traction curves have been measured on two traction machines with different contact geometries, i.e. twin-disc and ball-on-disc.

Three factors have been found to complicate the accuracy of the EHL traction prediction, namely solid body temperature effect, thermal conductivity of solids and lubricants, and scale/geometrical effects. Firstly, the bulk temperature of specimens exceeds the supplied oil temperature during traction measurements. Bulk temperature has been adopted as thermal boundary condition in thermal EHL simulations and it has found that the solid body temperature reduces oil viscosity and then reduces the film thickness and traction. Secondly, solid and lubricant thermal conductivity affects heat conduction and the maximum temperature rise, as well as the traction. For lubricants, thermal conductivity doubles its value at about 1 GPa; for solids, recent measurements have shown that thermal conductivity of through-hardened 52100 bearing steel should be around 21 W/mK rather than the widely cited 46 W/mK in literature. The effects of both solid and liquid thermal conductivity are analyzed. Thirdly, the traction curves measured from the two traction rigs for the same fluid are different at comparable operating conditions, i.e. at the same pressure, speed and supplied oil temperature. This phenomenon has been studied through thermal EHL analysis and it shows that the reason lies in the difference in the reduced radius of curvature which affects the film thickness and the heat distribution of the two traction machines. For a thicker EHL film thickness, shear is mainly localized in the middle film due to a temperature-viscosity gradient across the film. The scale effect studied here belongs to thermal effects.

For practical traction predictions, a simplified traction calculation method for highly loaded rolling/sliding EHL contacts has been developed without solving the Reynolds equation and the surface deformation equation. Using a bilinear limiting shear stress model extracted from traction experiments, the discrepancy between measurements and numerical simulations is smaller than 15% over a wide range of operating conditions (e.g. velocity, pressure, and slide-to-roll ratio), which is mainly caused by the solid body temperature effect.