coupling meshes at their interface

[Kevin]

Hi guys!
We have a mesh which contains two parts. The first part has 50 elements in z direction while the second part has 10 elements in that direction this is done to save the number of elements in the region where the flow is supposed to be inviscid and not turbulent. Also the length is equal in z direction and the number of elements in y direction is equal for both parts. The mesh is created by ICEM CFD. We want to couple these two parts at their interface. I've done this in ANSYS FLUENT by setting up interface boundary faces names as "interface1"&"interface2" and coupling them in interface settings and it was done very straightforward by a few number of clicks! I'm gonna check the ANSYS FLUENT accuracy and performance with code saturne because a friend told me that code saturne is much powerful and more accurate in turbulence than ASNSY FLUENT. How can I do this in code saturne? I couldn't find any options about mesh interface.
Thanks!

[Yvan Fournier]

Hello,

If I understand correctly, what you want is to join the non-conforming meshes into a single mesh. You can check the mesh joining options in the "Preprocessing" portion of the GUI for this. You also simply need to select the boundaries that should be joined.

Best regards,

Yvan

[Kevin]

Hello
Thank you much! I think it's now much easier in code saturne than ANSYS FLUENT It's working now.
Two more question:
1. Can I set up two different models for these two parts? For example Laminar + LES?
2. Code saturne is about ten times faster than ANSYS FLUENT while the solver precision is 1E-5 in code saturne and the residuals are 1E-5 in ANSYS FLUENT! I set the same solvers, models and other conditions in both softwares and equal number of cores are used. I think its wierd! Is there something special in code saturne? Is the solver precision the same as ANSYS FLUENT's residuals?

[Yvan Fournier]

Hello,

If you want to use different models (or computing options) in different meshes, the approach is quite different: in that case you must not join the meshes, but define 2 coupled cases (creating both cases at once with "code_saturne create", define a setup for each, and activate a coupling (boundary conditions-only, explicit as regards time-stepping) for both these cases (also defining the coupled zones). This is not possible using the GUI, but can be activated using the cs_user_coupling.c user functions (check the examples).

This was actually designed for RANS/LES coupling (with or without overlap), but is much more rarely used, less tested and may be incomplete. Also, as the coupling is by boundary conditions only, it is a quite weak coupling, so the incompressibility condition is not well enforced between domains. If you want to experiment with it, you are welcome, but in this case you should run small, easy to interpret test cases first to make sure the behavior fits your needs. We may be trying to improve some aspects of that coupling in the future so feedback is welcome also.

Best regards,

Yvan

[Kevin]

Hi admin. i think i'm not very expert in codesaturne to test this coupling! by the way thank you.