[CFD_Students_FR]

Hello everybody !

I have a problem when I’m trying to create symmetry zones in the Meshing of ANSYS v16.1.
I already read a lot of help about this topic, but I didn’t find out how to solve my problem.

The geometry that I want to make periodic is quite complicated, so I finally came back on a simple cube in order to better understand how to create these symmetry zones.
I have a fluid inlet and outlet which are facing each other. Meanwhile I want to get a symmetry on the two others pairs of faces (top and bottom, left and right), in order to create two periodic zones in Fluent.
I managed to create in the Mesh two symmetric parallel faces without problem by using some methods: by symmetry areas, cyclic regions, or by match control, etc…

My problem is actually that I don’t manage to create two pairs of symmetry zones which are in contact at an edge. Two cyclic regions are working together (I mean the mesh is created), but the pattern on each faces is inverted (because of the rotation around the Z axis of a cylindrical coordinate system). It means, that when I go to Fluent, the number of cells match but not the patterns!
Finally, I have always the same error message: “The mesh could not be generated because two faces adjacent to an edge have different match controls applied to them. To correct this, ensure that match control settings are the same for both adjacent faces.”
I think I tried every combinations of match control and symmetry areas, I didn’t manage to solve that!
If someone could help me it would be very great J

Thanks!

[harryosu]

Hi,

From my understanding, you are simulating whatever something that is periodic in both directions other than the inlet-outlet direction. What you can do is to create interfaces for these pair of surfaces in fluent and pair them to be "periodic". I think then the mesh generated in fluent will be adjusted thereafter to make them conformal automatically, a process similar to what you have been trying to do. That is to say, you don't need to do this symmetric or periodic or whatever meshing yourself in the meshing tool. Just go to fluent to finish this. Hopefully it helps.

Harry

[CFD_Students_FR]

Hello Harry,

Thank you for your answer!

What you propose is what I tried to do some days ago and yes it works.
I focused myself on trying to do a good meshing with match control, symmetries and so on, but in fact this solution is really easier.
Just, I think that it doesn't create a conformal meshing between the two interfaces, but Fluent tries to match every cell with a certain tolerance (so it is finally a non-conformal meshing).

[lisongying123]

Hi
I have the same problem, but I am doing it in static structure not in fluent, how to solve that problem:the mesh could not generated because two faces adjacent to an edge...
for a cube I want to set 3 directions of linear periodic symmetry, but it error all the time in meshing.

[CFD_Students_FR]

Hi,

I had the same problem that you describe and didn't find out the solution.

What I did (in Fluent, but it is maybe be the same kind of thing for another software) was to set each pair of surfaces as "interface". Then I set each pair or faces as "Periodic boundary conditions" and "Matching" and enter the distance in the right direction (as translation or rotational, depending you geometry).

In my case I checked that by looking at the mass flow: there is indeed mass flow conservation between each pair of faces. So the periodic conditions that I wanted works!

Hope it can help you (I forgot a bit it was for a student project last year!).

[vasava]

On a cube you can create periodic mesh by sweeping. If you use sweeping from top-to-bottom (in your case) it should automatically create a mesh that it periodic on four-sides and top-n-bottom.

[lisongying123]

thanks for the solution XD. I can understand your method in fluent, my research model is a rubber cube with iron particles in it, so I think I need analyze in static structure.
for that sweep method, this model is not a sweepable stuff because particles inside, but I test a normal cube it is true. so how can I do next?