SimScale CAE Forum

Periodic BC produces error


#1

Hey guys,

I am trying to simulate a compressor and I am having trouble with the periodic BC.

I am using a converted fluent mesh and started with a very basic setup (incompressible, inlet with slow velocities, outlet, Walls). This setup works fine. Since the mesh includes a wedge with one single blade, I want to apply a periodic BC to end up with a full cylinder. To do that, I apply the periodic BC separately to each corresponding patch pair. The Patch pairs seem to be the same size. If I now run the simulation, an error is raised before the first iteration. Can somebody have a look at the error?

https://www.simscale.com/projects/jbergmeier/eckardt_compressor/


#2

Hi @jbergmeier,

I’m not too familiar with this type of simulation case, but from looking at your setup for the compressible flow one, it just isn’t possible. There is no outlet for flow to exit out of and you will hit a continuity error straight away.

Can you elaborate on why does your outlet in the compressible flow have to be a pressure inlet instead of outlet? I’m guessing it has to do with your need for a periodic BC. If I’m not wrong there is a BC for cyclic which should give you what you want and you have assign both inlet and outlet as cyclic.

Let me know how it goes.

Cheers.

Regards,
Barry


#3

The setup for the compressible flow was just a test to get familiar with the setup. To avoid further confuision I now removed it. The error is beeing raised in the incompressible setup. And please don’t be confused by the patch_names, which I was provided along with the mesh. The “inlet” patch is the only inlet, while the other patches containg “inlet” in their name are just close to that area.


#4

Hi @jbergmeier,

Understood. I am however, woefully unfamiliar with this type of simulation case. Upon going through some sources, it seems implementation is not that simple. I will link the resources below that may help you. They go into detail about OpenFOAM and may be slightly different than what you see in SimScale, but SimScale is using OpenFOAM so the implementation should be identical.

From what I’ve read, it seems like implementation requires the use of AMI and coupled with the periodic BC. Along with that, you will need to define a mean velocity through the mesh and a pressure to reference from.

Otherwise, I will call on @jousefm & @1318980 who may have some idea on how to do this.

Cheers.

Regards,
Barry


#5

Hi @Get_Barried @jbergmeier, what you describe Barry in terms of the periodic BC is what is done behind the scenes, however:

https://www.simscale.com/docs/content/simulation/model/boundaryConditionTypes/OF_new_cyclic.html

The two faces have to be an almost perfect similarity in terms of shape and area, and the meshes have to be sufficiently self-similar also, not exactly the same but approximately.

That’s all I can say on the periodic part :slight_smile:

Best,
Darren


#6

This might actually be the problem. I will check if they are identical and let you know. Thanks!