Syrthes 3.4

[nsellami]

Hi there,
I am a new user of both code saturne (2.1.4) and syrthes (3.4) with the purpose of using the CHT feature available with these codes.

I am trying to firstly run separatly a solid and liquid domains with each of these codes before using the coupling part. I have done alright with code_saturne but I have trouble converting a MED mesh with the version 3.4 script: convert2syrthes (Message: the mesh file is a 3.0.4 med file while the script is a 2.6 version of MED).

I tried to compile it independently with a more recent med version but that did not work (I saw on the forum that there were a bug to be fixed in the version 4).
I then used the syrthes version 4 script which allowed me to sucessfully convert the .med to .syr mesh file but now it is the solver syrthes 3.4 that does not work.

I am not sure what is the problem so far (message: Erreur de segmentation ) , but I wanted to know if the Mesh.syr file created by convert2syrthes4 is read and compatible with the syrthes 3.4 solver.

Thanks in advance for any assistane!

Nadia

[Yvan Fournier]

Hello,

The .syr file format has changed between Syrthes versions 3.4 and 4.0, so the error
you obtain is not surprising.

Code_Saturne 2.2 (compatible with Syrthes 4) should be released before the end of this month, so I recommend to wait a few more days (if you have the GNU autotools, flex, bison, pdflatex, and recode packages installed, you may even try downloading it from the Subversion mirror on this web site, using "svn checkout http://code-saturne.org/svn/saturne/branches/Version2_2 saturne-2.2" followed by "cd saturne-2.2 && ./sbin/bootstrap", but waiting a few days is probably simpler.

In the meantime, you may try exporting the MED file to ".unv" format with SALOME, and running trying that with Syrthes 3.4 (note that Syrthes 3.4 requires quadratic elements, while Syrthes 4 requires linear elements).

Best regards,

Yvan

[nsellami]

Hi There,
Thanks you for your quick reply! we really appreciate your assistance and contributions with CS/syrthes!

I basically tried your second solution as I am unable to get the version 2.2 from the mirror website to work at all.
I have a quadratic unv mesh file but I get this message when I run syrthes 3.4:

%% ERREUR LECID1 : UTILISATION D'ELEMENTS NON CONFORMES

I actually put the .unv file directly as the mesh file to be read in the syrthes.env file. Is that correct ?
Thanks again-
Nadia

[Yvan Fournier]

Hello,

Yes your use is correct. As the I-deas format allows several variant element types, it is possible that the variants supported by SALOME and Syrthes 3.4 are not the same.

In any case, moving to Code_Saturne 2.2 and SYRTHES 4 would be a better option, so if we do not finally release Code_Saturne 2.2 this week, I will send you a pre-release version.

Best regards,

Yvan

[nsellami]

Dear Yvan,
I will be waiting for the latest version as advised and I hope then to get the coupling with syrthes4 working finally!!
Have a great day-
Nadia

[nsellami]

Dear Yvan,
Do you think that the newest version will happen today?
Best Regards,
Nadia

[Yvan Fournier]

Hello,

I just sent you a pre-release of version 2.2 by our large message service.

Note that the user examples for specific physics are still a work in progress, but if you are using only the standard solver, this should not make a difference.

Best regards,

Yvan

[nsellami]

Dear Yvan,
Thanks for sending the latest version of CS.
I tried to re-compile the code so that it does find all the local paths for libraries and also for syrthes, but whenever I run the code_saturne create for coupled case with syrthes just to create the proper folder it does give me this error:

/home/nsellami/saturne/code_saturne-2.2.0/lib/python2.6/site-packages/code_saturne/cs_create.py", line 327, in create_syrthes3_cases
shutil.copytree(src_ref_syr, data_ref_syr)
File "/usr/lib/python2.6/shutil.py", line 136, in copytree
names = os.listdir(src)
OSError: [Errno 2] No such file or directory: '/home/nsellami/saturne/syrthes4.0.0-beta/arch/Linux_x86_64/data'

Meaning that it is expecting to create a syrthes 3 type of case yet I am using a version 4.
I am not sure if during the compilation or creation of the coupled case, I have to tell that it is a version 4 or that my compilation is wrong even though I tried short runs on independent cases and it seemed fine-
Thanks for your help in advance and sorry for taking more of your time!
Have a good day!
Nadia

[Martin Isack]

Hello,

is it possible that you compiled Code_Saturne with the “--with-syrthes=…” option? This configure option is for SYRTHES 3 only. For SYRTHES 4 you have to ignore this option. Instead you have to edit the file …/etc/code_saturne.cfg in the directory where you installed Code_Saturne. Just put the path to SYRTHES install directory in the appropriate line.
i. e.:

Code: Select all

syrthes = /path/to/syrthes4.0.0-beta/arch/Linux_x86_64

Regards

Martin

[nsellami]

Hello there,
I followed your advise since I was indeed compiling using the --with_syrthes option and now it seems to compile together. however, when I run the coupled case, I have this error in the liquid listing file:

fvm_to_med.c:4740: Erreur fatale.

MED doit avoir une connectivité.
Maillage : "SYRTHES face centers"
Gestionnaire : "results"

It sounds that my mesh is wrong even though separatly, both codes run on those same meshes. Is there a requirement for interfacial mesh- I thought it does not need to be a conformal mesh?

Thanks for any hint to help me resolve this issue!
Take care,
Nadia