Page 1 of 1

Coupling Code saturne 7 syrthes 5

Posted: Wed Sep 22, 2021 7:28 am
by vlado285
Hello,

I am working on coupling the code saturne 7 with syrthes 5. The sequential and parallel calculation without coupling run on both without problem. But if I want to run the coupling calculation I get some issue. It seems to be something wrong with MPI but the both were compiled with the same mpicompiler.
I attached the resu files from simulation.

Re: Coupling Code saturne 7 syrthes 5

Posted: Wed Sep 22, 2021 1:27 pm
by Yvan Fournier
Hello,

This thread belongs in the "Syrthes related questions" or "Installation issues" section, not here. So I'll move it soon (you can move it yourself I believe, which is even better).

This seems to be a crash on the Syrthes side. So probably an installation issue, but hard to analyze just from the stack trace. If you can compile a debug build of Syrthes, we could run this under a debugger to pinpoint the cause of the crash, which would help. You have a "specific_debug" option in the Syrthes initialization.

In any case, there have been some runs combining code_saturne 7 and Syrthes 5, so thre is no fundamental incompatibility, but probably a build issue or a bug.

Otherwise, unless you need some Syrthes-specific features (advanced radiative model options, with a different feature set than those of code_saturne), you can use code_saturne's built-in "internal coupling".

Best regards,

Yvan

Re: Coupling Code saturne 7 syrthes 5

Posted: Wed Sep 22, 2021 2:46 pm
by vlado285
Hello Yvan,

If I run the internal fluid-thermal coupling in code saturne. Is it possible to export the solid domain in med file and then use it for simulation in code aster as thermal load(temperature field)?

Re: Coupling Code saturne 7 syrthes 5

Posted: Thu Sep 23, 2021 1:57 am
by Yvan Fournier
Hello,

Yes, absolutely. You simply need to define a postprocessing mesh whose selection criteria matches only the solid part, define a writer in MED format, and associate the 2 (all in the graphical interface).

The output will have cell-centered values instead of vertex-based values, but code_aster should be OK with that. If not, you can always use SALOME's PARAVIS module to import the MED file, apply ParaView's cell data to point data filter, and re-export it to MED format. Actually, this can even be done quite easily direclty in code_saturne, with user-defined functions (using a new "cs_cell_to_vertex" function).

Best regards,

Yvan

Re: Coupling Code saturne 7 syrthes 5

Posted: Thu Sep 23, 2021 7:27 am
by vlado285
Hello,

thank you very much. It's very good news for me.
I would have one question more.
My goal is to simulate the more domains together. 2 domain as fluid and one as solid. I attached a picture.
Solid is as pipe. Then in pipe flow the hot water and outside of pipe flow the cold air.
Is it possible simulate something like that in code saturne or it is required the coupling with syrthes?

Re: Coupling Code saturne 7 syrthes 5

Posted: Thu Sep 23, 2021 8:43 am
by Yvan Fournier
Hello,

Yes, this is possible with internal coupling. You can define physical properties by zone, so the air/water aspect should not be an issue.

Best regards,

Yvan

Re: Coupling Code saturne 7 syrthes 5

Posted: Fri Sep 24, 2021 1:53 pm
by vlado285
Hello Yvan,
I tried to make the simulation as I sent you before.
The simulation ran but I have no possibilities to set the specified properties for various domain (air and water).
For solid domain it is possible to set them but for next other no.

Re: Coupling Code saturne 7 syrthes 5

Posted: Fri Sep 24, 2021 3:15 pm
by Yvan Fournier
Hello,

Did you check the "physical properties" box for the matching zones in the general "Volume zones" settings ?

Regards,

Yvan

Re: Coupling Code saturne 7 syrthes 5

Posted: Fri Sep 24, 2021 6:11 pm
by vlado285
Hello,
Yes I checked them for all domain.

Re: Coupling Code saturne 7 syrthes 5

Posted: Sat Sep 25, 2021 2:50 pm
by Yvan Fournier
Hello,

Checking your case, you need to set the physical property type to "user law" for the "all cells" section before you can define the user law for the other zones.

This is related to the inner workings of the code (where the property needs to be defined "per cell" and not globally), but should be better automated, and is not very intuitive....

But in any case, doing this should work fine.

Best regards,

Yvan