[ariorus]

Hi,
I need to split my mesh in order to apply boundary conditions at some internal zero-thickness surfaces.

I used this approach:

1)I exported my mesh (from gambit) applying interior boundary at a zero-thickness surface named "thinsurf".
In this way I get a faceset named thinsurf.

2) modified boundary file by adding two empty boundaries with zero faces, called "thinsurf1" and "thinsurf2".

3) executed "splitMesh . mycase thinsurf thinsurf1 thinsurf2".

I tried using a very 2d simple mesh and it works ok, but using 3d meshes I have a lot of problems.

The output of splitMesh is the following:
[...]
Splitter
{
type attachDetach;
faceZoneName membraneFaces;
masterPatchName thinface1;
slavePatchName thinface2;
triggerTimes 1(0);
manualTrigger off;
active on;
}

Adding topology modifiers
--> FOAM Warning :
From function bool primitiveMesh::checkCellsZipUp(const bool, labelHashSet*) const
in file meshes/primitiveMesh/primitiveMeshCheck.C at line 1471
40 open cells found. Please use the mesh zip-up tool.
Failed 1 mesh topology checks.
Failed some mesh checks.
[...]


And the output of "checkMesh . mycase" gives me a lot of messages like this:


bool primitiveMesh::checkCellsZipUp(const bool, labelHashSet*) const :
Cell 6324 has got 4 unmatched edges:
4
(
(1702 1654)
(1702 551)
(551 192)
(192 1654)
)


--> FOAM Warning :
From function bool primitiveMesh::checkCellsZipUp(const bool, labelHashSet*) const
in file meshes/primitiveMesh/primitiveMeshCheck.C at line 1471
40 open cells found. Please use the mesh zip-up tool.
Writing 40 cells with over used edges to set zipUpCells


I think the problem is in the mesh but if I run checkMesh before using splitMesh everything seems ok. I don't know what to do...

Thank you for any suggestions.

[ariorus]

So far I succeeded if my thin surface is surrounded by layers of prisms.

Simple tetrahedra cells give errors. Probably the error is due to theaedra having sharp points on the surface...

[kati]

I had a similar problem with lots of unmatched edges and open cells after splitMesh. I then read the original mesh into Fluent (changed the interior to wall before that) and wrote it out as a case file, and ran fluentMeshToFoam on the case file. The resulting foam mesh is ok.

I also tested a very simple 3D case with internal boundary condition, and that was ok for splitMesh. The original case was about 2 million cells and complicated structure, no tetras, but some prisms. I don't know what exactly is the problem with the larger mesh. If someone is interested in developing splitMesh further, I can test it more or try to find the places and mesh structure where it goes wrong.

[ronaldo]

Hi,

please just tell me how to run splitMesh and CellSet?
What about the makeCellSets.setSet file.

Thank you in advance!

[purnp2]

Hello,
I am reviving this old post in hope of understanding the complete picture of how splitMesh works.

I have written my original post here:
boundary conditions for p, U, omega, k and nut for an "internal wall" of a mesh

My question, in the context of the current thread is: what boundary condtions should I use for the newly generated walls (thinsurf1 and thinsurf2).

1. For example, in 0/p.boundaryField

thinsurf1
{
     type    <?????>
}
thinsurf2
{
     type     <?????>
}

2. In 0.polyMesh

thinsurf1
{
       type          <?????>
       inGroup     1(<?????>)
       nFaces       283
       startFace   466642
}
thinthinsurf2
{
       type          <?????>
       inGroup     1(<?????>)
       nFaces       283
       startFace   466925
}

What should I assinge in places of <?????> ?