Hi!

I have a mesh which is basically quad but refines using tri elements in regular intervals. The problem is 2D and has a periodic bc at the left and right, moving wall at the top and fixed wall at the bottom. The bottom wall is not flat. If I run this with only quad at high resolution I see a vortex forming as expected. Now keeping the number of elements constant at the bottom wall and applying the "coarsening" openfoam does not capture the vortex and the flow is strangely laminar, even at high velocities. Is there something wrong with the mesh, sort of channelling the flow even though there is no evidence of this in the pressure field? Does openfoam struggle with the combination of quad/tri ?

paraFoam does produce an error message with this mesh but seems to work.

A very coarse and basic (but in principle the same) mesh can be found on:

http://openfoam:openfoam@217.46.165.161/openfoam

if asked for a user/password it is always openfoam (both, password and userid)

I checked the mesh as good as I could, normal of faces, volume of cells using Gauss and Stokes etc but could not find anything wrong with it ....

It is generated automatically so if the coarse one is okay the fine one I use will be in principle the same.

Tried QUICK as well as the standard linear solver but no luck so far.

Will try to run it through fluent in the meantime, we'll see what it says.

Any input more than welcome and appreciated!

Kind regards,

Markus

Just "learned" how to post attachements. Sorry for cluttering up your inboxes. So here come a few files, a pdf showing the mesh structure and all the mesh files like points, faces etc. I ran checkMesh in the meantime and it checkes out. But I can still not make much sense out of the different results I get from openFoam, nor the error message from paraFOAM about missing vertices in the tuples.

Thanks for anybody investing some time into having a quick look at the files!

Cheers,

Markus.

well, aparently some learning to do still .... IE closed the window on me, suppose that is why I usually avoid MS.




mesh.pdf
points.zip
neighbour
faces.zip
owner
boundary