Applying real measured surface roughness to EHL contacts

On-going work in the CPDE Unit, led by Prof Martin Berzins, is into elastohydrodynamic lubrication of real rough surfaces. This work is funded by Shell Global Solutions and is relevant to accurate friction modelling of engine performance.

The following measured surface roughness patern has been applied to EHL contacts.

The data of resolution 256x256 data was used at the mesh points on a grid 7 mesh (ie 257x257 with the last on each row being an averaged value of the first and the last). I then produced coarser meshes by simly taking the corresponding mesh points, and finer meshes by linear interpolation.

The only part of the provided data used is the bottom half of the top picture. This is then reflected around the central axis.

The solver was then run on grid 9 (1025x1025 points).

Pressure, inflow from left. Geometry, inflow from left.

Solution time for 30 V-cycles on four processors, grid 9: 20 minutes. Residuals at 10e-9, still dropping off steadily towards round off.


When the roughness is oriented the otherway, ie such that the ghrooves are perpendicular to the flow direction the pictures look like...:

Pressure, inflow from left. Geometry, inflow from left.


Movies

Jpeg Movie here

Mpg Movie here

Introduction - Variable Timestepping - Mesh Adaptation - PSEs - Parallelism - The Grid