Multiscale tribological modelling of the mixed lubrication regime

Microscale topography covers every surface and has a substantial
impact on overall tribological performance. The separation of scales between
the macroscale contacting surfaces and the microscale roughness renders
accurate modelling difficult. Contemporary literature is dominated by the
efficient flow factors methods that cannot make accurate microscale predictions,
and the accurate but computationally expensive deterministic approach. By
using the Heterogenous Multiscale Methods (HMM), the speed of the flow
factors methods can be achieved with the accuracy of a deterministic approach.
Experimental measurements of asperity height and frequency are
applied across a model of a tilted pad bearing and used to generate more
representative microscale models across the surface. Three otherwise identical
models are created with different, normalised microscale topographies. Clearly
indicating that the morphology of the microscale topography has a substantial
impact on the overall performance of lubricated contacts. The HMM models are
at least sixty times less computationally expensive than a deterministic
approach for modelling varying surface topography on tilted pad bearings.
Mixed lubrication is modelled within the HMM for the first time, enabling
the generation of a Stribeck curve that spans the mixed and hydrodynamic
lubrication regimes. The developed framework models local deformation,
cavitation, piezoviscosity, and the effect of pressure on density. This framework
is then validated against experimental data from literature. A normalised root
mean squared error of 0.023 is calculated between model predications and
experimental data from literature.
Laser Surface Texturing (LST) is used to apply a microscale topography
and control friction. In this work, the impact of LST is explored on a point
contact problem. Experimental data is contrasted with data from numerical
models. A normalised root mean square error of 0.0241 is achieved for the
numerical predictions of experimental outputs with LST. Through this
comparison, the efficacy of the HMM for determining the effect that microscale
topography has on macroscale performance is evidenced.