[york]

Hello all,

I am a graduate student studying the motion of buoys in the ocean. I am learning about wave theory and CFD, trying to get an understanding of things. I am trying to decide on a CFD software to use for my research. Of course, like I said, I am new to CFD; I have to pick my software now, though. I know, it doesn't make total sense, but that is how it is.

I have been trying to learn enough about what is important that I can make a fairly educated decision. Sometimes, it is hard to interpret the information. I have searched the OpenFOAM web page and the forums, but there are still some things that confuse me. I know that some of these things could very well be discussed on the board already, and I may have read them and just not understood. I hope someone may be patient with me and help me to understand.

1) It seems that OpenFOAM handles unstructured grids. Does blockMesh create unstructured grids, then?

2) Does it solve implicitly, explicitly, or both?

3) Does it use adaptive meshing? It describes the mesh thusly: "a cell is simply represented as a list of faces and a face as a list of vertices," but I am not sure how to interpret that.

4) Does it use a boundary layer around structures? I have been told this helps a lot with accuracy.

5) How does it do with fluids interacting with solid objects? Does anyone know of someone who has used OpenFOAM to model an ocean buoy?

6) Would its discretization be considered finite volume method? Or some clever hybrid?

7) It seems that mesh creation can be a major part of the work in a CFD simulation. What are your thoughts on the OpenFOAM mesh creation tools?

Like I said earlier, I know this will probably annoy some people, but if anyone is feeling generous, I would sure appreciate your replies. And, as I mentioned, if anyone knows of work someone has done modeling ocean buoys in OpenFOAM, that would be nice, too.


Thanks much for all help,

[dmoroian]

Hi Charles,
Although I'm not an experienced OpenFOAM user I will try to answer to your questions so you can have a small image of this library.

1. Yes blockMesh creates grids that can be used by OpenFOAM, and they are unstructured.
2. If you refer to the time discretisation, then in the documentation it says that it can solve both implicitly and explicitly, but the options I know for time discretisations are only implicit. There are other operators (divergence, gradient, laplacian) that are available in both versions: implicit and explicit.
3. As far as I know, it can handle adaptive meshing.
4. Yes it can solve inside a boundary layer but, I think, here you refere at a certain feature of Gambit called boundary layer, which is nothing else but a fine discretisation of the domain in zones with high variations of the solved variables.
5. Well it can handle this types of problems too (Fluid Structure Interaction), just not in parallel, yet.
6. Yes it solves partial differential equations using finite volumes (which is clever enough in my opinion)
7. This is somehow a drawback in OpenFOAM: it lacks good mesh generators. The only tool that I know is blockMesh which can generate only block meshes of hexahedral (not very sure about this) cells. The point is, that for complex meshes, you need an appropriate preprocessor. There are quite a lot and for most of them, OpenFOAM has good translators.

I hope you will find this useful,
Dragos

[alberto]

Hello,
for fluid-structure interaction there is a technology demo on the OpenFOAM wiki:

http://www.cdlact.unileoben.ac.at/non_cdl/OpenFOAMWiki/index.php/Contrib_icoStru ctFoam

Regards,
Alberto

[york]

Thank you all for your responses. I will definitely look into that demo once I get OpenFOAM going.

I also found the training info at :

http://www.mfix.org/mwiki/index.php/Main_Page

which is quite helpful.

I still appreciate the input, insights, or experience of anyone else who wishes to post.


Thanks very much