[gerritgroot]

Hi,

I ran SU2_PBC on a periodic grid (not the grid from the example), but SU2_PBC.exe breaks down when it checks the periodic boundary conditions.

MyDirectory>SU2_PBC addperiodic.cfg
....
-------------------- Setting the periodic boundaries --------------------
Setting the periodic boundary conditions.

And then it breaks down and asks me if I want to send a message to Microsoft (I said yes :-))

Apparently, there's something wrong with the grid, but visualizing it, I can't find anything. I don't have any grid generator, nor a tool to check grids (I got this grid from a third party)

I tried to add the periodic boundary conditions with the following cfg file

MESH_FILENAME=InputGrid.su2
MESH_FORMAT= SU2
MARKER_PERIODIC= (Periodic,PeriodicShadow,0.,0.,0.,0.,0.,-120.,0.,0.,0.,PeriodicShadow,Periodic,0.,0.,0.,0., 0.,120.,0.,0.,0.)
MESH_OUT_FILENAME=OutputGrid.su2
OUTPUT_FORMAT= TECPLOT

Is there some grid checking tool to see what's wrong with the *.su2 grid?
SU_PBC just breaks down, it doesn't tell me which element is wrong.

[economon]

Hi,

One thing you can quickly try to check the mesh is to simply run it through SU2_CFD instead for only a single iteration (set the periodic markers to be farfield BCs, and use EXT_ITER= 1). If this does not crash, it will write a solution file that can be opened in Tecplot, for example, which will allow you to study the mesh. SU2_CFD may also give you better warning or error information if there indeed is an issue.

Hope this helps!
Tom

[gerritgroot]

Yes, there seems to be something wrong with the grid.

SU2_CFD breaks down when...

"Searching for closest normal neighbor on the surface."

[gerritgroot]

Hi,

Couldn't find anything strange in the grid.

Could it be that quad elements are not allowed on the periodic boundaries?

[gerritgroot]

Hi,

Couldn't find anything strange in the grid.

Could it be that quad elements are not allowed on the periodic boundaries?

[economon]

Hmmm, well since it isn't running in SU2_CFD either, I don't think it is related to SU2_PBC.

Are you able to visualize your mesh in a different program before exporting it to SU2 format? Is it possible that there are duplicate nodes or unsupported unstructured elements (other than tri, quad, tet, hex, pyramid, prism)? Do you have a way to convert it to a different format that you can visualize?

[gerritgroot]

I visualised it already, there is nothing stange to be seen.
I'll check it again, it only has tetras and wedges.
What should I check it for?
- Duplicated nodes?
- Unused nodes?
- Bad node references?

[gerritgroot]

Hi,

I obtained the same grid saved in cgns format, it runs in fluent, but it breaks down in SU2_PBC.

[economon]

Have you tried converting the CGNS mesh to SU2 from within SU2? Is there an error written to the console when trying to read the grid in from CGNS format?