CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   SU2 (https://www.cfd-online.com/Forums/su2/)
-   -   Interface boundary condition (https://www.cfd-online.com/Forums/su2/193277-interface-boundary-condition.html)

Yminjo September 21, 2017 00:11

Interface boundary condition
 
Hi, all.

As I know, the interface boundary condition in SU2 is just for postpocessing, and it does not affect any solution.
In my case, the solution diverges when I used the B.C..
Does anyone have the same problem with me?

Thank you.

talbring September 25, 2017 09:09

Hi Yeongmin,

you are right, the interface should not affect the solution. First of all make sure that you provide only one side of the interface in your mesh file, the other side is automatically constructed.

If that does not solve your problem, provide me the mesh file and config so that I can have a look at the problem.

Tim

Yminjo September 25, 2017 12:00

1 Attachment(s)
Quote:

Originally Posted by talbring (Post 665562)
Hi Yeongmin,

you are right, the interface should not affect the solution. First of all make sure that you provide only one side of the interface in your mesh file, the other side is automatically constructed.

If that does not solve your problem, provide me the mesh file and config so that I can have a look at the problem.

Tim


Dear Albring

Thank you for your reply so much.
I attached my mesh and configure files in the following URL because of its large size:

https://drive.google.com/file/d/0BxY...I2VWlLZE0/view

The zip file includes three mesh files with and without interface markers.

1. 2C_no_interface.su2 does not have any interface marker.
2. 2C_with_inter1_inter2.su2 has a pair of interface markers of inward 'inter1' and outward 'inter2'.
3. 2C_with_inter2_only.su2 has only the outward 'inter2' marker.
4. comp.cfg is the cofig file.

Also, the definition of markers is shown in the attached figure 'Markers.jpg'.

The mesh has very small domain size for testing the interface feature, but the computation converges well using '2C_no_interface.su2' file.
Because the three mesh files are exactly same excluding the marker definition, the computation using '2C_with_inter2_only.su2' must converge too,
However the computation diverges.

Could you please tell me what the current problem is?

Thank you for your effort in advance.

Sincerely,

Yeongmin

talbring September 26, 2017 05:01

Hi Yeongmin,

I think the problem is that the interface shares a point with the wall (airfoil). Can you try to increase the distance a little bit and try it again ?

Yminjo September 26, 2017 05:25

1 Attachment(s)
Quote:

Originally Posted by talbring (Post 665689)
Hi Yeongmin,

I think the problem is that the interface shares a point with the wall (airfoil). Can you try to increase the distance a little bit and try it again ?


Dear Albring

Thank you for your reply.
Actually, I think I found what the problem is.
The problem was not in SU2, but in the connectivity written in the mesh file.
For the interface boundary condition, I made the inner 'inter1' and outer 'inter2' boundaries using connectors in Pointwise.
I attached a figure to explain that situation: green connector for 'inter1' and blue connector for 'inter2'.
In my experience, I have to select a pair of inner and outer connectors in Pointwise whichever they are assigned as boundary conditions together or separately.

Anyway, I think that the way generates wrong connectivity information.
After generating a mesh file without the interface boundary in Pointwise, I made a new marker with connectivity manually in the mesh file.
Then, the solution converges well, and the interface feature works well.

Now, I'm trying to find a way to avoid the problem in Pointwise.
If you don't mind, could you please tell me your way to make the interface marker?

Thank you, again.

Sincerely,

Yeongmin


All times are GMT -4. The time now is 05:43.