Exploring elastohydrodynamic lubrication using finite-volume computational modelling techniques

Abstract

Elastohydrodynamic lubrication is a regime of lubrication that occurs in lubricated rolling-sliding non-conformal contacts, such as those in rolling bearings and gears. Under EHL, the solid surfaces are completely separated by a lubricant film that is built up through a hydrodynamic action and the fluid pressures are sufficiency high to cause significant elastic deformation of the solids. EHL is typically modelled using numerical solutions to some form of Reynolds equation. These methods can be used to simulate EHL to predict variables such as friction, film thickness and temperature rise. The Reynolds equation is derived from the general fluid flow equations by assuming certain gradients are negligible. In contrast, Computational Fluid Dynamics (CFD) methods can solve the general fluid flow and heat transfer equations fully. Advantages include modelling the full convective heat transfer effects at the contact entrance, extension of the simulation domain further than Reynolds models and truer simulated multiphase flow. Comparisons between the simulation methods are very limited, with only a few papers that compare CFD solutions to a Reynolds models under conditions that show solution differences. This project is aimed at investigating CFD simulation of EHL, chiefly in situations where the assumptions of Reynolds derived models of EHL are tentative or incorrect by comparing Reynolds derived EHL models to a CFD model parametrically. In developing and using the CFD model to gain new insight into EHL the CFD model solutions were compared with experimental results for film thickness profile at varying entrainment speeds and slide-roll ratios (SRRs). The CFD model was also used to study the thermal effects in EHL in relation to Archards temperature rise equation inside the contact, as well as the thermal effects outside of the contact. Thermal solution comparisons between a thermal modified Reynolds model and the CFD model at varying SRR (0–1.8) and dimensionless entrainment (U = 1 × 10−10–5 × 10−10) were carried out and finally parametric comparisons between CFD predicted and literature film thickness reduction factors were made. Results show that at moderate conditions the solutions obtained using these two modelling strategies are nearly identical in terms of film thickness (Thermal loading parameter below 1); however, at high viscosity, high entrainment and high SRR, CFD solutions differ in predicted film thickness and/or friction from Reynolds-based models due to terms neglected in the Reynolds equation derivation. In general CFD results predict higher temperature rise due to thermal convective effects at the contact inlet.