CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM Running, Solving & CFD (
-   -   Strange boundaries for k and epsilon (

henning December 16, 2008 09:55

Hi, I have a pipe looking l

I have a pipe looking like this:

and its extruded along the flow direction (z-axis).

When I run simpleFoam (3000 Steps) I get unrealistic values for k(at least 184 though u = 3.725 m/s). Also near the boundaries they look strange. k should go back to 0 and epsilon should rise constantly.

I made a simpler case with a 2D channel and a similar mesh. Here the results look quite good.

Any idea what i am doing wrong?

B.c.s for k and epsilon are zeroGradient (similar to the tutorial cases in simpleFoam / turbFoam) and just for the internal field I gave values. I also tried values on inlet but then I get an error message saying I need zeroGradient at inlet patch.

Thanks for your help


pierre December 16, 2008 11:24

what turbulence model do you u
what turbulence model do you use?

Have you had a look at boundaryFoam and its tutorial? Even channelOodles. That might help.


juanduque December 16, 2008 15:10

Why don't you try a different
Why don't you try a different mesh configuration, check O-Grid configuration, that might help.

henning December 17, 2008 06:07

I am using the classical k-eps
I am using the classical k-epsilon model.

I took a look on boundaryFoam. It uses zeroGradient b.c. for both, k and epsilon as well. channelOodles didn't help me that much since it is LES and i am not so familar with LES.

And changing the mesh ... well, I don't like to use O-Grid for this solution since I would prefer a mesh that can be made finer at the boundaries easyly.

Maybe any other ideas?

Anyhow, thank you very very much for your help so far.


henning January 7, 2009 03:54

Well, found my problem. Did pa
Well, found my problem. Did patch instead of wall in blockMeshDict. It's always the simple things that make life so interesting ;)

All times are GMT -4. The time now is 16:55.