I am in the process of validating OpenFOAM for couple of standard benchmark cases for turbulent flows with heat transfer. I already have converged solution by Fluent and wants to see how OpenFOAM compares. As a first case, I took validation for "backward facing step" and converted fluent mesh to OpenFOAM mesh. While running this case in OpenFOAM with "empty" BC, flow development was not accurate. But results looks much better if I use "slip wall" BC. As it is recommended to use "empty" BC for 2D cases, why do I see such a discrepancy? Two dofferent 3D Fluent grids were used, with one/three elements in the "Z" direction. I have also looked at the example case "simpleFoam", available in the OpenFOAM "tutorials" directory.

[henry]

"empty" BC is only correct for 2D cases, that is cases which are 1 cell thick with the front and back planes parallel. Cases which are more than one cell think or non-parallel front and back planes are not 2D because they allow flow in the third direction and are hence 3D. For such cases you must define all BCs and the fron and back planes may be either symmetry, cyclic or walls depending on how you want to constrain the flow in the 3rd direction.

Thanks Henry for the replay. As I mentioned, I tried running two cases, one with one cell thick and another with three cells thick. I faced the same problem. It works well with "slip" conditions. I faced convergence problem even with "symmetry" BC. Any thought.

Regards

Sorry for creating a confusion. Earlier we created a 3D grid with one cell in the "Z" direction in Gambit and tried to run it in the OpenFOAM. Basically I was not aware that it is possible to convert a 2D Fluent grid to Foam grid. This time we tried and it works with "empty" BC. Thanks Henry for your suggestions.

Now I want to apply a profile for the velocity and turbulence quantities at the inlet. How do I do that?

Regards