[rhee]

Hi,

I have a question about converting the GGI inteface into the AMI for the overlapped interface patches. I am in the process of exploring "Arbitrary Meshing Interface (AMI)" available in OpenFOAM-2.0 or newer. I already went through three examples under tutorials/.../pimpleDyMFoam, all of them starting from scratch using either blockMesh or snappyHexMesh. But in my case, I created two separate meshes - a body mesh with a circular disc carved out and a blade mesh inside the disc. Then I merged the two meshes using mergeMeshes tool. I already tested out my merged mesh with body/blade interface using the GGI in OpenFOAM-1.6-ext. I would like to get some help in setting up my case for the AMI.

Here are my settings used for the GGI interace:

1. ++++++++++++
in constant/polymesh/boundary file,
body interface
{
type ggi;
nFaces 18272;
startFace 29142236;
shadowPatch blade-interface;
zone body-interface-zone;
bridgeOverlap no;
}

blade-interface
{
type ggi;
nFaces 66880;
startFace 29075356;
shadowPatch body-interface;
zone blade-interface-zone;
bridgeOverlap no;
}

2. ++++++++++++

in constant/dynamicMeshDict,

dynamicFvMesh mixerGgiFvMesh;
mixerGgiFvMeshCoeffs
{
type cylindrical;
origin (0.98 0 0);
axis (-1 0 0);
direction (0 0 1);
}

rpm 500;

attachDetach false;
rotatingRegionMarkerPoint (0.98 0 0);

slider
{
moving (blade-interface);
static (body-interface);
}

}

3. +++++++++++++++++++++

in setBatch file,

faceSet blade-interface-zone new patchToFace blade-interface
faceSet body-interface-zone new patchToFace body-interface

4. +++++++++++++++

run setSet -batch setBatch

5. +++++++++++++

run setsToZones -noFlipMap

[hz283]

Hi Bong,

I had the similar problem to the one mentioned in your thread. For my case, I divided the computational domain into several bodies, and generate the meshes body by body using ICEM and then merge them together. There will be interface between neighboring bodies and on these interfaces the mesh nodes do not coincide. These interface are stationary and the meshes in the computational domain do not move during the simulations. These interfaces are equivalent to the interior in Fluent. For this situation, can I use AMI ?

best
H

Originally Posted by rhee

Hi,

I have a question about converting the GGI inteface into the AMI for the overlapped interface patches. I am in the process of exploring "Arbitrary Meshing Interface (AMI)" available in OpenFOAM-2.0 or newer. I already went through three examples under tutorials/.../pimpleDyMFoam, all of them starting from scratch using either blockMesh or snappyHexMesh. But in my case, I created two separate meshes - a body mesh with a circular disc carved out and a blade mesh inside the disc. Then I merged the two meshes using mergeMeshes tool. I already tested out my merged mesh with body/blade interface using the GGI in OpenFOAM-1.6-ext. I would like to get some help in setting up my case for the AMI.

Here are my settings used for the GGI interace:

1. ++++++++++++
in constant/polymesh/boundary file,
body interface
{
type ggi;
nFaces 18272;
startFace 29142236;
shadowPatch blade-interface;
zone body-interface-zone;
bridgeOverlap no;
}

blade-interface
{
type ggi;
nFaces 66880;
startFace 29075356;
shadowPatch body-interface;
zone blade-interface-zone;
bridgeOverlap no;
}

2. ++++++++++++

in constant/dynamicMeshDict,

dynamicFvMesh mixerGgiFvMesh;
mixerGgiFvMeshCoeffs
{
type cylindrical;
origin (0.98 0 0);
axis (-1 0 0);
direction (0 0 1);
}

rpm 500;

attachDetach false;
rotatingRegionMarkerPoint (0.98 0 0);

slider
{
moving (blade-interface);
static (body-interface);
}

}

3. +++++++++++++++++++++

in setBatch file,

faceSet blade-interface-zone new patchToFace blade-interface
faceSet body-interface-zone new patchToFace body-interface

4. +++++++++++++++

run setSet -batch setBatch

5. +++++++++++++

run setsToZones -noFlipMap