[Gaetano03]

Hello everyone,
I have a bit annoying issue while exporting su2 format from gmsh. The problem is that it includes even construction points that don't belong to the mesh in the su2 file. So when I execute SU2 I get a Segmentation fault. Am I missing something or is it really not possible to use construction points that don't belong to the mesh when exporting the su2 format?


Thanks for any help


Gaetano

[TKatt]

Hi,
please try creating Physical Surfaces for 2D meshes (or Volumes in 3D) for all your inner domain and then export the mesh with the -saveall flag (from command line) or the equivalent option from the GUI.

That did the trick for me as I ran into that problem with construction points as well. See https://github.com/su2code/TestCases...ration/0__mesh for the way I did it with .geo script files. In the build_mesh.py you'll find the command how I build the mesh. You don't need to care about all the stuff as it is a multizone mesh.

I hope that helps, cheers,

Tobi

[Gaetano03]

Hello Tobi,
many thanks, I defined physical and volume groups and I don't have construction points listed in the su2 mesh anymore. But I still have an issue. During Geometry Preprocessing I got the following error:
"The surface element (i,j) doesn't have an associated volume element"


I have no idea how I can manage it. I attach my .geo and the .log file with the error in dwing.zip


Thank you very much again for your help


Gaetano

[TKatt]

Ok, could you provide the command with which you build the mesh or the steps in the GUI to reproduce your mesh. Secondly a .cfg which shows the error, doesn't need to be the original one if IP is an issue. These two are necessary to reproduce your error! Edit: and your .su2 mesh as well if it is not too big.

Some general Advice: Start with easy geometries which has the rough characteristics of the big mesh and few cells where you have fast turnaround times and can still see structures and even maybe see bad cells. This is beneficial for you because often you can debug your problem there and directly use it for the bigger problem. You are also more likely to get quick help together with the 2 missing pieces above as other people can test without big effort.

I loaded your .geo in gmsh and simply pressed Mesh->3D (as not further specified) which took some 7min on my local machine. The resulting mesh (without further checking) didn't look too good in the farfield. Did you already used that mesh before with another solver to know that it is otherwise correct?

I am optimistic we can get your mesh running

Cheers, Tobi

[Gaetano03]

Hello,
Thank you again for the help and the advice. I used the Frontal Algorithm to generate the mesh, because the default Delaunay worked too bad for the far field when you have to interface with prisms elements. Actually Gmsh pops up a warning message when I try to use Delaunay, advicing to use Frontal scheme. Then I just pressed 3 on the keyboard. The rest is all defined in the .geo file (attractor).

Honestly, I didn't test the mesh with other solvers.
What do you mean by .cfg with the error? Isn't it the .log file I have already provided?

By the way I am afraid I can't upload the su2 mesh


Thanks again,
Gaetano

[TKatt]

Hey,
With .cfg I mean, that the error occurs during the program execution "SU2_CFD <filename.cfg>" ... I would like to have <filename.cfg> to be sure that I can reproduce your error. Otherwise I would need to create/remix my own .cfg for the case which, again, costs time.

An idea towards the error message "The surface element (0, 0) doesn't have an associated volume element" -> Can you assure/test whether you have a matching mesh at the interface near the Wing with a structured mesh to the farfield with an unstructured? To me the error sounds like a non-matching mesh connection might be the problem here.

Cheers Tobi

[Gaetano03]

hello,
Here is the config file. Anyway, the first number in the error message ( 0 ) isn't it associated with the first marker in su2 mesh file, that is Farfield?


Cheers,
Gaetano