recognition of boundary conditions by code Saturn from a STAR-CCM+ mesh

[john.poubelle]

Hi,

I would like to use Code SATURNE to compute a STAR-CCM+ generated tetrahedral mesh.

Code SATURNE accepts the mesh and does not complain about the quality. I transferred the mesh as a .case file.

I did not find how to make the link between the surfaces defined in STAR-CCM+ and Code SATURNE. in Salome, I can see the diffrent surfaces have a different, discrete, value of vtkCompositeIndex, but how to use this in Code SATURNE ?


Thanks for your answers

[Yvan Fournier]

Hello,

Which format did you use ? .ccm or CGNS ?

In any case, the boundary zones should appear as "groups" whose names you can use to define zone selection criteria for boundary conditions (i.e. for example, if you have a group named "inlet", using "inlet" as a selection criterion for the matching inlet should do).

Regards,

Yvan Fournier

[john.poubelle]

Hi and thank you for your answer

I was using .case (ensight). I also tried CGNS but I did get this message :


Error in preprocessor execution
===============================

../../../preprocessor/pre-post/ecs_pre_cgns.c:255: Fatal error.

CGNS error:
CGNS file 0 is not open

This may be due to a CGNS library supporting only ADF (legacy)
or HDF5 (future default) and the file not matching. In this case,
running the adf2hdf of hdf2adf converters may solve the issue.



The preprocessor does not propose groups : just group "1" which is the main fluid domain.

In SCCM+, i used the "export mesh & solution data" option. Is there Something else required by Code_SATURN ?

Thanks again

[john.poubelle]

One last thing : I am running on Windows 10 computers. I do not know how to run hdf2adf on Windows. Is there a way ?

Thanks again

[Yvan Fournier]

Hello,

We did not yet generate and a 5.0.8 build for Windows, so if you are using 5.0.7, the support for polyhedron elements in CGNS (used by Star-CCM+) is missing. We will probably release 5.0.9 soon, so I'll try to remember to generate a build for Windows, but this build is fragile due to recent versions of some optional libraries not being compatible with the build method (see this thread for details: viewtopic.php?f=4&t=2034).

And the Widows build does not contain the CCM format support either (as its licence is proprietary,, we ensure Code_Saturne can use it when available, but do not distribute it ourselves).

So using the EnSight format as you did is probably the simplest solution, but I am not sure how it behaves in this type of file. Could you post your "preprocessor.log" file (or EnSight mesh, if it is small enough and can be shared).

Regards,

Yvan

[john.poubelle]

Thanks for your reply this is great.

Attached is the preprocessing.log.

Thanks again for your consideration.

[Yvan Fournier]

Hello,

Unfortunately, looking at the preprocessor.log, groups are not translated using the EnSight output. You can still post a smal example of an EnSight output, in case some other option combinations can be used (I can check more with the mesh).

So CGNS would be a better option. If you can post your mesh (or a simpler one) using CGNS output, I can test it on Linux first and test it on the next Windows builds (though can't guarantee it will work). If you have several options for CGNS output, could you try and post meshes exported with different option settings, to see if some have better compatibility than others ?

Regards,

Yvan

[john.poubelle]

OK, I will see what i can do.

Thanks again

[john.poubelle]

OK, problem : my mesh is 32 Mo => over max. size.

I will try generating a new simple mesh using basic output options. I do not see much of those… it is essentially "export/mesh/CGNS => GO".

Thanks for trying this.

[john.poubelle]

K.

I just tried several options in SCCM+ : add vertices data => no change.

I generated this very simple case : a cube with 12 tetras. It has 6 different boundary surfaces, that i cannot see in code_saturne.

Attached are :

- .case file. i had to change the extension into .log to upload it.
- .geo file. i had to change the extebsion into.log to upload it
- preprocessor.log

I am trying to get around this by using SALOME to set the boundary conditions. Problem => to generate groups I have to click or to put Inside a box all the cells of each side. is there a faster way (i.e / select from one starting point & angle like in Gmsh for instance ? Sorry but I am kind of beginner with Salome).

Thanks