CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > ANSYS Meshing & Geometry

[ICEM] Hybrid mesh merging issues

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   March 26, 2014, 07:05
Default Hybrid mesh merging issues
  #1
Senior Member
 
Francisco
Join Date: Mar 2012
Location: Spain
Posts: 270
Rep Power: 6
Bollonga is on a distinguished road
Hi everybody,

I'm trying to make an hybrid mesh consisting on an external hex mesh and an internal tet mesh around two bodies (picture 1 and 2). There's a symmetry plane reducing the domain in half.

I make the hex mesh and the tet mesh separately (with the same node number in the common edges), I merge both meshes and try to use the edit mesh/Merge nodes/Merge meshes option, choosing the common surfaces as the merging interface. It gives me several error message:

Unhandled element type HEXA_8
0 vertices included
0 vertices removed
I am trying to include edge 13444 13442 but vertex 13444 isn't even in the region 1

And finally erases the surrounding hex elements (see picture 3).

Anybody can tell me which is the problem?
How can I make a conformal interface between the hex and the tet mesh?

And also, I've decided to continue with the non-conformal mesh, setting the interfaces as internal walls, but when I open it in Fluent I got the following warning:

Warning: Inappropriate zone type (interior) for one-sided face zone 2.
Changing to wall.

Then I cannot change the type back to internal.

Why do I get only one-sided faces? How can I fix that?

Any comment is highly appreciated! Thanks in advance.
Attached Images
File Type: jpg External_hex.jpg (46.0 KB, 24 views)
File Type: jpg Internal_tet.jpg (33.5 KB, 23 views)
File Type: jpg Erased_interface.jpg (98.5 KB, 24 views)
Bollonga is offline   Reply With Quote

Old   March 26, 2014, 12:46
Default
  #2
kad
Senior Member
 
Join Date: Feb 2013
Location: Germany
Posts: 171
Rep Power: 13
kad will become famous soon enoughkad will become famous soon enough
The interface should contain only quad_4 and tri_3 elements. I don't know why ICEM is finding hexa_8 there. So it could be a problem of wrong parts.

It is important that the interface in both domains have the same name e.g. MERGE. When merging the meshes you only select this one surface part. This part should not contain any other mesh or geometry.

As hex/tet merging only works well with nearly quadriliteral elements on the interface, there can also be a problem in the near wall region. In general element shape there is strongly non quadriliteral, because of the high resolution of the boundary layer. Btw. your geometry looks pretty doable for blocking.
kad is offline   Reply With Quote

Old   March 26, 2014, 18:17
Default
  #3
Senior Member
 
Francisco
Join Date: Mar 2012
Location: Spain
Posts: 270
Rep Power: 6
Bollonga is on a distinguished road
Quote:
Originally Posted by kad View Post
The interface should contain only quad_4 and tri_3 elements.
And so it is, only quad and tri.

Quote:
Originally Posted by kad View Post
It is important that the interface in both domains have the same name e.g. MERGE. When merging the meshes you only select this one surface part. This part should not contain any other mesh or geometry.
So it is, it is called internal and only contains the surface elements to merge.

Quote:
Originally Posted by kad View Post
As hex/tet merging only works well with nearly quadriliteral elements on the interface, there can also be a problem in the near wall region. In general element shape there is strongly non quadriliteral, because of the high resolution of the boundary layer. Btw. your geometry looks pretty doable for blocking.
What do you mean with quadrilateral? All quads are obviously cuadrilateral and tri cannot be quadrilateral. Did you mean cubical shape for the hexaedral elements?

I tried at first a blocking strategy but the two bodies are too close and I don't know how to solve that. How would you do it?

I attach the project, geometry, blocking, and both meshes in the link below. So you can check where the problem is.

https://dl.dropboxusercontent.com/u/...o_symmetry.zip

Thanks a lot!
Bollonga is offline   Reply With Quote

Old   March 27, 2014, 09:48
Default
  #4
kad
Senior Member
 
Join Date: Feb 2013
Location: Germany
Posts: 171
Rep Power: 13
kad will become famous soon enoughkad will become famous soon enough
Quote:
Originally Posted by Bollonga View Post
What do you mean with quadrilateral? All quads are obviously cuadrilateral and tri cannot be quadrilateral. Did you mean cubical shape for the hexaedral elements?

I tried at first a blocking strategy but the two bodies are too close and I don't know how to solve that. How would you do it?
Sorry, I meant equilateral (->aspect ratio near 1). This is pretty obvious when you think about how the process works. There are pyramids grown on the quad_4 elements of the hexa mesh and these are then connected to the tetra mesh. The quad_4 are somehow the baseplanes of the pyramids.If these elements are nowhere near equilateral the pyramid elements will become highly distorted or the process might even fail.

In your case I think there are two major problems. First your tet mesh contains prism. The merge tool can not handle these elements. No prism elements are allowed at the interface. There are some threads around here that say cou can do it with a workaround, but I don't know if this gives you satisfying results. The second point I have already stated above.

One obvious solution would be to make an all tet/prism mesh. You can also try the cutcell mesher in workbench, if you want to go for hexa. It should be able to produce a good mesh for this geometry.
kad is offline   Reply With Quote

Old   March 28, 2014, 09:17
Default
  #5
Senior Member
 
Francisco
Join Date: Mar 2012
Location: Spain
Posts: 270
Rep Power: 6
Bollonga is on a distinguished road
Quote:
Originally Posted by kad View Post
In your case I think there are two major problems. First your tet mesh contains prism. The merge tool can not handle these elements. No prism elements are allowed at the interface.
You were right, Kad. I erased the prism layer in the interface and I got the conformal surface mesh and the transition pyramids from tet to hex.

I wonder if using the cutcell method in ansys workbench it is possible to choose areas with different densities as in regular meshing method.

Thanks!
Bollonga is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Fluent converting 'wall' to 'interior' BC for hybrid mesh DarrenC FLUENT 6 November 6, 2014 06:00
[ICEM] Orthogonality/Skew issues in 3D unstructured mesh eddyy19g ANSYS Meshing & Geometry 3 February 13, 2014 10:36
3D Hybrid Mesh Errors DarrenC ANSYS Meshing & Geometry 11 August 5, 2013 06:42
snappyHexMesh won't work - zeros everywhere! sc298 OpenFOAM Native Meshers: snappyHexMesh and Others 2 March 27, 2011 21:11
[ICEM] hybrid mesh vmlxb6 ANSYS Meshing & Geometry 5 March 9, 2011 14:44


All times are GMT -4. The time now is 06:07.