[STutexas]

Originally Posted by bigphil

Hi Bertrand,

Unfortunately I have not come across any other software to convert vtk/abaqus to OpenFOAM.
You could write a program to convert them but this may be a lot of effort.

Another possible option is to use a free meshing program that tet meshes your geometry, then import to OpenFOAM and then convert this geometry into polyhedral using polyDualMesh.

Philip

Hi Philip

I was reading this post and I have a similar problem of converting CT images to OpenFoam meshes. I can generate a gmsh file but I cannot define patches for the boundaries. Please let me know if you finally found any solution to this problem.

Thank you.

Saurabh

[bigphil]

Originally Posted by STutexas

Hi Philip

I was reading this post and I have a similar problem of converting CT images to OpenFoam meshes. I can generate a gmsh file but I cannot define patches for the boundaries. Please let me know if you finally found any solution to this problem.

Thank you.

Saurabh

Hi Saurabh,

Currently the solution we use is:

• generate castellated surface STL from CT images using 3Dslicer;
• smooth STL using volume preserving smoothing in meshlab;
• open smoothed STL in ANSYS ICEM CFD meshing program: cut surface geometry into desired boundary patches;
• create volume mesh using the STL and cfMesh OR ANSYS ICEM CFD.

As you can see we are trying to use open-source software but we still need ANSYS ICEM CFD to cut the STL into patches.

Philip

[STutexas]

Hi Philip

Thank you for replying so quickly and for your advice. Could you let me know how big is the mesh file that you use or if you have ever had an issue with memory for any of the mesh files.

Thanks again for the help.

Saurabh

[bigphil]

Originally Posted by STutexas

Hi Philip

Thank you for replying so quickly and for your advice. Could you let me know how big is the mesh file that you use or if you have ever had an issue with memory for any of the mesh files.

Thanks again for the help.

Saurabh

Hi Saurabh,

We have generated final "OpenFOAM" meshes of about 5 million cells with no problem, where the STL would be less than 100 MB; so I suppose we have not tried very big problems yet.
Depending on the final mesh density you desire, it can also be convenient to "decimate" (i.e. reduce the number of triangles) of the STL using meshlab too, after smoothing; but you should make sure that the STL resolution is still greater than the volume mesh resolution that you want.

Philip

[STutexas]

Hi Philip

Thanks again for replying so quickly and for all your advice. It was very insightful. I was wondering if you ever looked into the use of Iso2mesh toolbox in Matlab and if you used it for meshing of 3d CT images. Thanks again for the advice.

Saurabh

[bigphil]

Originally Posted by STutexas

Hi Philip

Thanks again for replying so quickly and for all your advice. It was very insightful. I was wondering if you ever looked into the use of Iso2mesh toolbox in Matlab and if you used it for meshing of 3d CT images. Thanks again for the advice.

Saurabh

Hi Saurabh,

No I have not tried the Iso2mesh toolbox, though it looks interesting from a quick look at the website.

As much as we would like it to be, our current procedure still requires quite a bit of manual effort and is potentially subjective. I might have a look at Iso2mesh in the future.

Best,
Philip

[STutexas]

Hi Philip

I have been able to create a pore scale mesh using gmsh and Iso2mesh software. The mesh quality is not very good and I still can't run simulations but the process is still very simple and repeatable if you have CT image slices. Can you advice me on how to improve the mesh. Can OpenFOAM do it for me or do I have to go back to the originial image and improve my mesh. This is the error that I get when I run simulation.

Thanks again for the help.

#0 Foam::error:rintStack(Foam::Ostream&) at ??:?
#1 Foam::sigFpe::sigHandler(int) at ??:?
#2 ? in "/lib/x86_64-linux-gnu/libc.so.6"
#3 Foam::fv::gaussGrad<double>::gradf(Foam::Geometric Field<double, Foam::fvsPatchField, Foam::surfaceMesh> const&, Foam::word const&) at ??:?
#4 Foam::fv::gaussGrad<double>::calcGrad(Foam::Geomet ricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::word const&) const at ??:?
#5 Foam::fv::gradScheme<double>::grad(Foam::Geometric Field<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::word const&) const at ??:?

[bigphil]

Originally Posted by STutexas

Hi Philip

I have been able to create a pore scale mesh using gmsh and Iso2mesh software. The mesh quality is not very good and I still can't run simulations but the process is still very simple and repeatable if you have CT image slices. Can you advice me on how to improve the mesh. Can OpenFOAM do it for me or do I have to go back to the originial image and improve my mesh. This is the error that I get when I run simulation.

Thanks again for the help.

#0 Foam::error:rintStack(Foam::Ostream&) at ??:?
#1 Foam::sigFpe::sigHandler(int) at ??:?
#2 ? in "/lib/x86_64-linux-gnu/libc.so.6"
#3 Foam::fv::gaussGrad<double>::gradf(Foam::Geometric Field<double, Foam::fvsPatchField, Foam::surfaceMesh> const&, Foam::word const&) at ??:?
#4 Foam::fv::gaussGrad<double>::calcGrad(Foam::Geomet ricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::word const&) const at ??:?
#5 Foam::fv::gradScheme<double>::grad(Foam::Geometric Field<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::word const&) const at ??:?

Hi Saurabh,

In my experience, OpenFOAM does not have tools to improve a mesh (though I think one of the foam extend dynamic tet meshers can be used to smooth a tetrahedral mesh).
The best option is probably to create a better mesh in the software that created it.

Alternatively, depending on the solver you are using, a more appropriate choice of schemes and boundary conditions might give you a result (search around the forum for similar problems).

Philip