[crostolo]

Hi all,
I have a problem on wich I'm smashing my head for a month. I'm trying to mesh the fluid field around a wing.

I use unstructured mesh. I have create a "wind gallery" of a third of a tube (so with an angle of 120°). Then I mesh the surface and create --> periodic --> rotate for the other face.

When I import the cae file into fluent he generates many errors due to incosistent rotation of the periodic domain, in particular when I try to transform it into polyhedra mesh.
The only strange thing I have noticed is that I can't project all the mesh surfaces oh my gemoetry surfaces (in particular the boundary node appears not to be well projected).

Is there someone could help me? Thanks

edit: I add some more info. I have made my geometry on SolidWorks with rotational extrusion for the domain. Here are some imagines, of the domain (a coarse one (external) a fine one and then inside it the wing) and the problem I met. All the faces are not 100% projected and only if I remove the thick from "interior only" i can project 100% my faces.
Then when I try to do ne periodic rotate face it is not 100% projected :/

[crostolo]

Here are the imagines

[dgarlisch]

I strongly suggest you contact Pointwise support for help on this problem.

If support is not available, you will need to provide more information that just a few images. Attach the pw file along with the exact steps you use to repeat the bug.

It is better if you can repeat the issue in a simple pw file.

[tcarrigan]

That looks very similar to the HAWT case we did a couple years ago. Take a look at this video, it may answer your questions:

http://www.youtube.com/watch?v=KOvWlgBEzo4

[crostolo]

Originally Posted by tcarrigan

That looks very similar to the HAWT case we did a couple years ago. Take a look at this video, it may answer your questions:

http://www.youtube.com/watch?v=KOvWlgBEzo4

Thanks a lot for the video, very useful.
Seems not to resolve my problem with periodic btw.

Seems to be a problem on the code when I import mesh into fluent if I use periodic "feature" of pointwise, so I have resolved my problem simply making mesh of an half of my domain and the mirroring along y axis and it works

[dgarlisch]

Try exporting the grid from Pointwise with some other type of boundary condition on the periodic domains (wall?).

Then inside Fluent, change the BCs back to periodic.

Good luck.

[Shubham_SD]

Hey, guys, I have a similar problem as discussed above. On importing the domains in fluent, the interfaces lose periodicity.

Warning: Number of nodes read does not match number referenced . Resetting counter to match the number referenced.

Error at Node 2: Couldn't intersect threads 14 and 15.


This is likely due to incorrect specification of periodic
offset values. Please check the offset values specified
and recreate the interface.

Error: Couldn't intersect threads 14 and 15.

Sometimes I saw that the rotational periodic shows 168 degrees (fluent, default) instead of 180 deg.

I also tried importing the periodic zones in wall and other BCs, but that didn't work

Neither did the TUI commands for making periodics in Fluent.

Would be grateful for your help!

Thanks
Shubham

[dgarlisch]

I assume this is rotational periodic?

I am not a Fluent user, but I have seen problems with symmetry exports if the symmetry domains are NOT truly flat (within a tolerance).

Are the periodic domains database constrained to a plane entity?

If not, try this:

• Create a database plane at the symmetry location.
• Project the periodic domains and their connectors to the plane.

This should force all points to be coplanar to a very tight tolerance.

You can use Examine, Database associativity... to verify that all appropriate grid points are constrained to the plane.

It is possible this operation might break the periodic connection between the domains. I have not tried it.

Good luck.

[Shubham_SD]

Thanks, Garlisch. It is indeed a rotational periodic.

I tried it your way (All domains projected on respective planes) but that doesn't solve the problem.

Also, when the rotor domain rotates, it creates -ve volume elements.

I tried it with 2 basic cases geometries:
(a). without database import (+unstructured mesh) and
(b). imported database (& structured mesh) (domains projected on planes).

The basic geometry being a cylinder rotating inside another, although I did not use periodic condition here.

In case (a), the sliding mesh method works quite great, while in (b). it shows the -ve volume error.
Can't figure out what the problem is !

While this works fine if the mesh is not made in pointwise.

Thanks
Shubham

[Shubham_SD]

Hope these images would clarify the problem.

(Axis of rotation is y-axis)