[mdawson25]

Hello,

I am currently analyzing the flow over a cylinder using a geometry and mesh created in Gmsh. It is a cubic volume, four faces serving as farfield boundary conditions with the two that intersect the cylinder being symmetry. The surface of the cylinder is marked using the Navier-Stokes (no-slip) constant heat flux marker.

While running SU2, I see the following appear during the initial stages:

---Geometry Preprocessing---
Setting point connectivity.
Renumbering points (Reverse Cuthill McKee Ordering).
Recomputing point connectivity.
Setting element connectivity.
Checking the numerical grid orientation.
Not defined orientation change
Not defined orientation change
Not defined orientation change
Not defined orientation change
(this continues for several more lines)

Once the Residual Evolution Summary appears I see non-physical points in the solution and non-physical states in the upwind reconstruction. This persists until SU2 tries to create an output file and then SU2 crashes.

I believe this "Not defined orientation change" error is leading to the other errors but even after some googling I haven't located a solution. Is there something I need to set in my SU2 config file or is this an issue with the mesh I have created in Gmsh then converted to an .SU2 file?


Best Regards,

Matt

[fpalacios]

Thanks for using SU2,

Unfortunately I have found similar problems in the past with gmesh. Could you please try to generate a grid with only tetrahedra

Best,
Francisco

[mdawson25]

Francisco,

I have gone back and edited my gmesh file to output only tetrahedral elements. After running SU2_CFD, I no longer get the orientation errors or the non-physical locations in the solution! In turn, the solution file is able to be output and no crash occurs.


Thanks for the help,

Matt

[AlbertoPi]

Originally Posted by fpalacios

Thanks for using SU2,

Unfortunately I have found similar problems in the past with gmesh. Could you please try to generate a grid with only tetrahedra

Best,
Francisco

Hi
I'm meshing a 3D wing and I have the same error.
I have a prismatic boundary layer with 4 million elements. I have 7 million elements in total and with my RAM (32Gb) I can not have more than 8 million elements to run SU2_CFD.
If I do not use "Recombine" in the "Extrude" of the boundary layer to get only tetrahedra, I have too many elements.
Do you know any other solution for this?

Thank you.

[AlbertoPi]

Originally Posted by AlbertoPi

Hi
I'm meshing a 3D wing and I have the same error.
I have a prismatic boundary layer with 4 million elements. I have 7 million elements in total and with my RAM (32Gb) I can not have more than 8 million elements to run SU2_CFD.
If I do not use "Recombine" in the "Extrude" of the boundary layer to get only tetrahedra, I have too many elements.
Do you know any other solution for this?

Thank you.

Hi,
Finally I found the way to do it. Getting prims and tetrahedras but no pyramids at final mesh. It looks like the pyramids were the problem.

Now the message "Not defined orientation change" doesn't appear, but non-physical points still appering.

Regards.