CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   problem with interfaces (http://www.cfd-online.com/Forums/star-ccm/102988-problem-interfaces.html)

sieginc. June 7, 2012 18:20

problem with interfaces
 
I'm trying to simulate the convective cooling of a heat generating block in duct. I want the block to be defined as a solid, but I am having trouble creating the interfaces. I split the geometry by part curves, and then create another part based on the surfaces of the block. Then I send the two parts to regions, and define one as the air and one as the solid.

When I try to mesh I get an error that says one of the edges is shared by an odd number (3) of faces. At first I thought this meant I had somehow defined two faces on top of each other, and then when the cube meets the base of the duct that would result in edges being shared by three faces. However, there has to be the same surface defined in each region (one in air and one in solid) so you can select them both and define it as an interface, correct?

The system looks something like this:
http://img840.imageshack.us/img840/1...istometric.jpg

Henry Arrigo June 7, 2012 18:54

Hi
I think you should use separate meshes for each region instead of one mesh for both regions.

sieginc. June 7, 2012 23:29

Quote:

Originally Posted by Henry Arrigo (Post 365357)
Hi
I think you should use separate meshes for each region instead of one mesh for both regions.

Unfortunately I don't think this will work. I've tried this in the past, and I've also read that you must have one conformal mesh in the whole geometry.

abdul099 June 8, 2012 03:12

You're both wrong.
First, you can create a conformal mesh, using one mesh continuum. It works when your geometry preparation is done the right way. That's everyday business.

Second, you don't need a conformal mesh. It also works with a non-conformal mesh, but it will smear the solution a little at the interface. Conformal meshes are recommended, but not essential.

sieginc.: Can you have a look on the surface repair? Do you have error free parts when you check them independently? Do both of your parts share the same block surface? Did you create the interfaces prior to meshing?

sieginc. June 8, 2012 15:22

Quote:

Originally Posted by abdul099 (Post 365384)
You're both wrong.
First, you can create a conformal mesh, using one mesh continuum. It works when your geometry preparation is done the right way. That's everyday business.

Second, you don't need a conformal mesh. It also works with a non-conformal mesh, but it will smear the solution a little at the interface. Conformal meshes are recommended, but not essential.

sieginc.: Can you have a look on the surface repair? Do you have error free parts when you check them independently? Do both of your parts share the same block surface? Did you create the interfaces prior to meshing?


I will check the surface repair. The parts share the same block surface. I'm assuming this is the problem? However, doesn't the same surface have to show up in each region so I can select both to form an interface?

rwryne June 8, 2012 16:17

For people new to interfaces, definitely recommend "Split by topology". It gets everything set up for you!

abdul099 June 10, 2012 14:35

Quote:

Originally Posted by sieginc. (Post 365502)
... However, doesn't the same surface have to show up in each region so I can select both to form an interface?

Yes, you need the same surface in every part, so when you create regions from parts, the interface faces will show up as boundaries.
I also agree to rwryne, split by topology would be easier.


All times are GMT -4. The time now is 10:41.