|
[Sponsors] |
March 29, 2007, 06:58 |
GGI and 1:1 Domain Interface connections
|
#1 |
Guest
Posts: n/a
|
I have a 3D geometry composed by a domain and two subdomains.
I have solved a problem using a 2D extruded mesh from CFX-MESH. With tetrahedral elements. I have defined the Domain Interface connection as 1:1 and it runs perfect. Then I tried to solve it using a 2D extruded mesh from ICEM CFD. With hexahedral elements. I can't use a 1:1 interface connection, so the problem doesn't converge. Why can't I use 1:1 connection here? Can I change some parameter? |
|
March 30, 2007, 07:33 |
Re: GGI and 1:1 Domain Interface connections
|
#2 |
Guest
Posts: n/a
|
If I do a tetrahedral mesh with CFX-MESH I can use 1:1 connection; but if I do the same mesh with ICEM CFD I canīt use 1:1 (only GGI connection). Why???
|
|
April 1, 2007, 18:20 |
Re: GGI and 1:1 Domain Interface connections
|
#3 |
Guest
Posts: n/a
|
Hi,
I suspect the meshes are not lining up on either side of the interface. The difference might only be small. There is a tolerance for determining whether the mesh on either side of a periodic boundary, you could try increasing it if your meshes are almost but not quite matching. On a side issue: GGI interfaces for most flows are quite stable and should not cause failure to converge. I suspect there is a problem with your simulation setup. Glenn Horrocks |
|
April 2, 2007, 05:37 |
Re: GGI and 1:1 Domain Interface connections
|
#4 |
Guest
Posts: n/a
|
I have increased a lot the mesh tolerance but I still have the problem (I canīt use a 1:1 connection). Perhaps there is an option in ICEM CFD to change the tolerance.
|
|
|
|