Simple periodicity problem
 

Hi guys,

I'm new in CFD.

I've been working for a few weeks on a model of combustor, drawing it with UG, and now I'm trying to mesh it with ICEM.
I keep having a problem with the periodical options (I have cut a 60° part to use the symetry of the device), so I decided to cut a small part of my combustor and try to make the periodical thing work on it.
Here is what the small part looks like : inlet, outlet, 2 surfaces for the symetry and a top wall.

http://www.freeimagehosting.net/uploads/28a97103ba.jpg

After completing the meshing, when checking the mesh, I keep having this message in ICEM :

maximum distance from periodic 5.96046e-008
between vertices 409 and 409; Fixed

everytime I try to check it.
When I try to import the file in Fluent and do a check grid, this is what I get :

Zone 37: average rotation angle (deg) = 6.168 (0.242 to 13.324)
stored zone rotation angle (deg) = 0.991
stored axis , (0.000000e+000, 0.000000e+000, 1.000000e+000)
stored origin, (0.000000e+000, 0.000000e+000, 0.000000e+000)

instead of the 60° I wanted.
The periodic conditions seems not to be respected at all.
I've been trying several things, but still cannot make the periodicity work.
Do you guys have an idea where the problem come from ?

I sent the .tin and .uns files of my project here : http://rapidshare.com/files/400310758/cut.rar.html
Hope it will be enough to make someone able to check what is going wrong...

Mesh is fine.
 

I think you just have a significant digits problem...

Your axis endpoints have a lot of trailing digits and exceed the precision of the mesher. The mesher then finds the problem and fixes it (notes you were 5.9 e-8 off periodic)... This is not a significant distance off (6 orders of magnitude smaller than your mesh size.) Perhaps if you centered your geometry on 0 0 0 or at least on a number where you had a more absolute number... But I don't think precision is stopping you.
Attachment 3824


I am guessing you didn't setup your bocos correctly?
Attachment 3825

Or perhaps the problem is that your one periodic surface just has surface, while the other has curves, points and surface (shells, lines and nodes). I am not sure how sensitive Fluent is to this.

Check with the Fluent Forum for tips about setting up periodic problems...

I have the same problem. I have tried evrything and it does not work