Periodic boundary conditions: joining is non-conforming

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Post Reply
Boone11
Posts: 20
Joined: Fri Jan 13, 2023 10:47 am

Periodic boundary conditions: joining is non-conforming

Post by Boone11 »

Hi all,

I am performing a computation of flow through a downward facing step with periodic lateral boundaries. I use a mesh composed of both structured cells in some regions and unstructured mesh elsewhere. I need to use the joining face option in order to include the non-structured area in the domain (otherwise it is considered as a wall). I have to avoid non-conformal mesh because of some user functions.

The mesh is conformal but cells at the front periodic boundary are different from those at the rear wall. I think that this is the reason for the following message in the listing file (see attached file listing.txt):
Periodicity type: translation
...
Global number of intersections detected: 1068880
Vertex-Vertex intersections: 995952
Other intersections: 72928

Joining operation is non-conforming.
Am I right? What are the solutions to keep mesh conformal while performing computation with periodic boundary conditions?

I selected all the surfaces of the non-strucutured area for the joining face and I declared the entire front and rear surfaces as periodic conditions (including the surfaces of the non structured areas). Is that the good way of doing it?

The setup file and an insight of the used mesh are attached (note that to simplify the reading, the cell size in the non structured area has been increased, the cells of the front and rear periodic conditions are very close when I use more refined mesh, but they are not exactly the same...).

Thanks a lot for your help!
Attachments
setup.txt
(26.28 KiB) Downloaded 52 times
listing.txt
(19.38 KiB) Downloaded 46 times
mesh_insight.png
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: Periodic boundary conditions: joining is non-conforming

Post by Yvan Fournier »

Hello,

If the meshes are almost but not quite perfectly aligned due to precision issues, you can try to increase the periodic joining's tolerance. Increasing it too much will degrade quality and possibly lead to a crash, so go by small steps.

Depending on how you build the mesh, you may have more or less control. If you build in a bottom-up manner (channel sides first), you may be able to apply the mesh from one side to that of the other before meshing the volume.

Best regards,

Yvan
Boone11
Posts: 20
Joined: Fri Jan 13, 2023 10:47 am

Re: Periodic boundary conditions: joining is non-conforming

Post by Boone11 »

Thanks for your answer Yvan. I forced GMSH to produce identical cell pattern at both periodic frontiers. One problem solved ! ;)
Post Reply