code_saturne User's Forum

Hello,

I use the Code_Saturne version 2.0.1, and I want to use a .med mesh. I made a basic cube to see what it happens with a simple mesh. I obtain some warning in the listpre whatsoever with the next configuration :
- Support du format de fichiers CGNS 2.5.5
- Support du format de fichiers MED 2.3.6 (HDF5 1.8.6)
- Lecture de fichiers compressés ('.gz') avec Zlib 1.2.3,
or with the next configuration :
- Support du format de fichiers CGNS 2.5.3
- Support du format de fichiers MED 2.3.6 (HDF5 1.6.5)
- Lecture de fichiers compressés ('.gz') avec Zlib 1.2.3

It indicates :
"Le maillage MED comporte 180 éléments de type seg2 qui sont ignorés par le Préprocesseur.", and "9450 faces sur un total de 10800 ne sont pas coloriées et n'appartiennent Ã　 aucun groupe ...On attribue Ã　 ces faces une famille par défaut."

I want to know if it's serious and if it's a problem to the calculation (knowing I test the same mesh with the Code_Saturne version 1.3 and in this case, and I have no problem), or if it's just an indication.

I join the cube_test_box2004.listpre (listpre file).

Furthermore, can you post me a listpre, please, from EDF come from a .med mesh to the computer specialist installs the same libraries' versions of you and to reassure us.

Many thanks,

Best regards,

Pauline

Hello,

The messages are just additional information, not any indication of a problem (if I remeber correctly, 1.3 warned about defining a default family only for boundary faces, while 2.0 does it for interior faces too). And if you do have any missing face definitions leading for errors in the boundary conditions, 2.0 will also output an error.ensight (orr error.med, or error.cgns) directory (or file) allowing you to visualize where things are missing, so if you have any problems, you will have better diagnostics.

MED 2.3.6 with HDF5 1.8.6 should be fine, so do not worry.

Best regards,

  Yvan