Learning Forum

[FEMfan]

Hello, 

I'm trying to simulate a geometry that's periodic in not one, but two directions. I'm using the transient thermal module. As a minimal example, let's look at this 10 mm cube:

I want all the heat flux that goes out of one side go into the opposite side, as if the cubes were arranged in an infinite XZ grid.

As you can see, I have applied linear periodic BCs in the X and Z directions, each for the corresponding pair of opposing faces. I have chosen the "high" regions so that they have the larger X or Z value. The "linear shift" is equal to 10 mm.

However, when I generate the mesh, I get this 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."

Next thing i tried is to set up Match controls for each pair of adjacent faces like this:

Which returns the error message 

The meshing algorithm cannot find matching topology. Please verify the topology and the position as well as the orientation of any associated coordinate systems. 

Can anyone help me with this? I'm sure it can be done, but I cannot figure it out! Thanks for reading.

[Rohith Patchigolla]

Hi,

Currently, one can have Linear Periodic symmetry only in one direction. 

This is documented in ANSYS help at this link.

"The Linear Periodic boundary condition is used to simulate models with translational symmetry, where the structure is assumed to repeat itself in one particular direction to infinity. This feature supports only a single direction for the entire model (more than one direction is not supported) for structural analyses and thermal/thermal-electric analyses."

Best regards,

Rohith

[JavierAguilar]

Hello,

I have the same problem, I cannot match surfaces in the x and z directions simultaneously since there is an overlapping region (a straight line in the cube of @FEMfan above) in which the two different matching criteria coexists. Somebody knows how to solve the problem? I would like to try to match the areas without including the nodes at the borders, but I don't know how to do that.

Since I need to match the surfaces in more than one direction, I am meshing without introducing the Linear Periodic symmetry.

Greetings,

Javier.