Periodic Boundary divergence

[Anonymous]

Hi I am trying to solve a problem using Periodic Boundary conditions to reduce the model. 

Attach you can find a sketch of the model I want to simulate same with the Boundary Conditions of the model

I have tryed to simulate the model using k-epsilon method for turbulence, and also Eddy Viscosity (because is suppose to be more stable)

Each time the solver reach divergence, and I can't figure why. What an I doing wrong?

I have used a low mesh Size (aprox 30 000 elements) But I have use other similar model at higher mesh sizes up to 900 000 elements and the solver diverge anyway

Any help will be apreciate

P.S. In the picture it is not shown but the the opposite surfaces with Periodic BC, have also Periodic BC's with side 2.

[apolo_vanderberg]

Romar,

   While you mention that you've tested a couple mesh sizes, I would say that the issue would be your mesh quality. 30 000 elements is very small and may not be sufficient for capturing the detail in the center section.

We would want to make sure that we have a decent mesh in the horizontal direction.

Can you post an image of a Plane showing the Shaded Mesh on the plane lengthwise from inlet to outlet?

Or post your support sharefile to look at?

Thanks

Apolo

[Anonymous]

Hi attach you can find the simulation support file, and also I have made a new simulation in which I have 1 000 000 fluid nodes. I will attach also a picture of a cut plane ( there are some confetty in the picture I don't know from where). You can see in the picture I have a maximum velocity which is much higher that should be taking into account that the inlet velocity BC is 2 m/s

[apolo_vanderberg · Accepted solution]

Romar,

From the image you posted and the sharefile I would say that your mesh is more then adequate. I was trying to ensure you had 3-4 elements between the edge of the interior baffle and the periodic BC.

To get this to run go in to Solution Control (on the Solve dialog), click Advection, and select either ADV2 or ADV5.

Both of these will be tighter on mass balance and will avoid creating the divergence taht you were seeing.

Apolo

[Anonymous]

apolo, with ADV5 it worked. 

Thanks