CapSizer |
December 2, 2009 04:55 |
Schemes to improve pressure drag?
I'm analysing a series of 3D streamlined shapes with simpleFoam, using meshes created with snappyHexmesh. As far as I can make out, the meshes are OK, as they produce drag figures in Fluent that quite closely match what I was able to achieve with a very high quality block-structured hexahedral mesh. The Fluent results make sense, with pressure drag making up about 10% of the total drag. Re-number based on length is about 10 million, and the finess ratio is around 7. In Fluent I know that least squares interpolation combined with second order momentum differencing is required to get the accuracy. However, in simpleFoam I'm getting pressure drag that is about 4 times too high. FWIW, I'm using the SST turbulence model with wall functions, but changing the turbulence model does not make much difference here. I suspect that although I'm using higher order schemes, there is something important that I am missing. Any suggestions, please?
|