CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   Problems with cyclic boundaries in faMeshDefinition and surfactantFoam (http://www.cfd-online.com/Forums/openfoam/93088-problems-cyclic-boundaries-fameshdefinition-surfactantfoam.html)

safre October 4, 2011 17:43

Problems with cyclic boundaries in faMeshDefinition and surfactantFoam
 
I would like to use a solver similar to surfactantFoam but the with the finite volume solution to update the finite area velocity field. This works fine but I have problems with setting up the faMeshDefinition file and the surfactant concentration file Cs in the 0 directory when I use cyclic boundary conditions. The makeFaMesh is doing fine and creates the faBoundary file as it should but when I start the solver and the solver reads the Cs file I get.

"Floating point exception".

The setup works for other boundary conditions but not the cyclic one. Does anyone has experienced the same or what is wrong?

Plaes find my faMeshDefinition and mys Cs file below.

faMeshDefinition:

polyMeshPatches 1( surface );

boundary
{
frontAndBack
{
type cyclic;
ownerPolyPatch surface;
neighbourPolyPatch frontAndBack;
}

side
{
type cyclic;
ownerPolyPatch surface;
neighbourPolyPatch side;
}
}

Cs:
dimensions [1 -2 0 0 0 0 0];

internalField uniform 0;

referenceLevel 0;

boundaryField
{
frontAndBack
{
type cyclic;
}

side
{
type cyclic;
}
}

Best regards, Sam

ngj December 12, 2011 05:05

Hi Sam

I have had problems with finite area and cyclic boundary conditions, however, I have never had problems with initialising the field. The problem I encountered is only a concern if your fields on FAM are either vector or tensor fields.

Well, I had a more careful look at your faMeshDefinition-file, and your frontAndBack is set to cyclic, however, frontAndBack is typically empty, so does it make sense?

/ Niels

safre December 12, 2011 08:34

Hello Niels,

thank you for your answer. It might be misleading of me to call it frontAndBack but it is a three dimensional domain so yes it make sense to have them cyclic as well and it works well in the fvm solution.

You said that you had problems - did you make it work in the end? Was it a similar problem with a coupling between a fvm and a fam solutions? I solve for a velocity field and a scalar concentration field in the fam solver.

Best regards, Sam

ngj December 12, 2011 08:56

Hi Sam

The problem is/was that the transformations of vector/tensor fields across the cyclic boundary conditions are wrong! I have made a full bug-report here with a suggested solution:

http://sourceforge.net/apps/mantisbt...view.php?id=87

Nevertheless, I do not know if it has been integrated into the git, as I have not been using the finite area method for quite some time.

/ Niels


All times are GMT -4. The time now is 22:11.