Meshing advice for complex domains
 

For automatic meshing of complex domains, the octree hexagonal meshers seem to cope best. Harpoon is an example of such a mesher, but it's quite costly. Luckily, snappyHexMesh is also that kind of a utility. I only experimented with it for a day or so, but I think it will do exactly what you want from it. It doesn't feature a GUI but that is less of an issue that it may seem at first. You can specify refinement regions, but I doubt you will really require this as the mesh refinement adapts automatically to curvatures, and the boundary layer creation is a special feature of this program, creating prismatic elements there.

I would guess that looking into snappyHexMesh is almost certainly your best path.

Thanks for the advice. When I
 

Thanks for the advice. When I upgrade to a newer OS I will give OF-1.5 a try.

Hello, we have developed an op
 

Hello, we have developed an open-source tool for unstructured (tetra/prism) meshing -- see post in the Meshing/Mesh conversion -> Open source section.

If you want to give it a try, I'd be happy to help.

Hi Oliver, What is the name
 

Hi Oliver,

What is the name of the open-source tool you are talking about?

Dan

It is called engrid -- I just
 

It is called engrid -- I just announced it yesterday: see here

Oliver, We just developed a
 

Oliver,

We just developed a variational approach to generating prisms on complex biological geometries, lungs and hearts so far. We are in the proccess of publishing it now. Do you have a manuscript that describes the method you use? We would very much like to compare to see if there are advantages/disadvantages of one or the other. It will all eventually be released in opensource.

Dan

Dan, I am in the process of
 

Dan,

I am in the process of writing it. If you want you can send me an email at ogloth (at) engits dot com and I can send you something as soon as it is available.

In a few words: Prisms are created on all surface triangles of a certain boundary condition. For now, their quadrangular faces have to be at the domain border (if not connected to another prism of the layer). After that, nodes are smoothed by minimising a local error measure (basically a bit of a hacked function with weighting coefficients for various geometrical issues). Only one layer of prisms is smoothed and afterwards it is divided into sub-layers. The smoothing process is a bit slow at the moment, but the resulting grids seem to be quite good -- the main issue being the remaining tets, but that can be tackled as well.

Oliver

Hi Martin, did you have a l
 

Hi Martin,

did you have a look at salome?
http://www.salome-platform.org/

there are tutorials at CAELINUX
showing how to use salome 4 OpenFOAM
http://www.caelinux.org/wiki/index.php/Doc:CAETutorials#Salome_.26_OpenFOAM_tuto rial:_3D_CFD_analysis_of_a_Y-shaped_pipe_.28CAELinux_2008_.2F_CAELinux_2007.29

http://www.caelinux.org/wiki/index.php/Doc:CAETutorials#Salome_.26_OpenFOAM_tuto rial:_3D_CFD_analysis_of_a_Y-shaped_pipe_.28CAELinux_2007.29

http://www.caelinux.org/wiki/downloa...7/PipeGeom.htm

Salome uses NETgen gmsh and other mesher
and if you want to do meshrefine with blockmesh
> salome will mesh your Step or Iges Data with groups
http://openfoamwiki.net/index.php/IdeasUnvToFoam
ideasUnvToFoam will do the converting

hope i helped u

elvis

Thanks for your replies everyb
 

Thanks for your replies everybody. I'll try each of these meshers out.

Hi guys, How can I convert
 

Hi guys,

How can I convert pointwise meshes into openFoam format?
I think it would be possible with multiple mesh conversion via starCD, CFX, Gambit ..., but I don't know how?

Thanks in advance,
Mir