CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Native Meshers: blockMesh (https://www.cfd-online.com/Forums/openfoam-meshing-blockmesh/)
-   -   Is blockMesh generally considered a good quality tool (https://www.cfd-online.com/Forums/openfoam-meshing-blockmesh/61842-blockmesh-generally-considered-good-quality-tool.html)

york November 15, 2006 13:44

Hello OpenFOAM users, I am
 
Hello OpenFOAM users,

I am a graduate student researching the motion of a buoy in the ocean. I am trying to decide which CFD software to use. I have already posted twice on these forums about other questions, and appreciate everyone's helpful responses.

My question now is this: do you generally consider blockMesh a good quality mesh generation tool? I understand that it does not have a GUI (someone actually hinted that making one would be a good idea). Other than that, is it good? Does it do what you need? Keep in mind, I do not know much about mesh generation yet, but I have to make a decision about which software to go with, so I want to hear what some people who are very good at this (i.e. all of you) think about blockMesh.

Thank you for your time,

hjasak November 15, 2006 13:50

The answer depends very much o
 
The answer depends very much on your geometry. For all Mickey Mouse cases (read: simple geometry or of cases of academic interest) that can be decomposed into a few blocks with or without curved edges, blockMesh does the job pretty well. I would not consider using anything else for e.g. vortex shedding behind a cylinder mesh, most 2-D meshes etc.

However, if you want to mesh an engine coolant jacket or a Formula 1 car, blockMesh is nowhere near good enough. This is why I have put so much effort into mesh import and interfacing with other CFD packages or meshing tools.

Automatic meshing in OpenFOAM is of great interest but still not ready for wide exposure or public domain (my support and admiration goes to the guy who is doing it - he knows why) ;-)

Hope this was useful,

Hrv

york November 15, 2006 14:10

Hrv, Yes, thank you very mu
 
Hrv,

Yes, thank you very much, it was quite useful.

If you would suffer more questions...what do you mean by automatic meshing? Do you refer to the moving mesh solver? Or is this a part of mesh generation that I just do not understand yet?

Do you think that blockMesh would be good enough for meshing a buoy floating in the ocean? The buoy would be basically a donut or cylinder, with a thinner cylindrical spar through the center (like a doughnut on a stick). The ocean would be, however you mesh the surface of water.

Thank you very much again,

vinz November 15, 2006 14:23

Hi, I'm working on OpenFOAM
 
Hi,

I'm working on OpenFOAM to see whether or not, this could be a good CFD tool. I must admit that the solvers look pretty good for the cases that I computed so far (tutorials, flat plate, cylinder....).
However, I'm not using blockmesh. Indeed, I made a good mesh it for the cylinder case. It took me a whole day to do it. At the same time, my collegue did a similar mesh in a few minutes with the software GridPro. And, as a real mesher, the mesh looks really great. I'm computing vortex shedding on both meshes and I'll probably post my results whenever possible.
In my opinion, blockmesh can be used for really simple meshes (as you said Hrvoje), but use another mesher as soon as the mesh is a bit more complicated. Otherwise, you'll use a lot of time.
Hope my comment will help you,

Vincent

hjasak November 15, 2006 14:42

Hi York, Doing cylinders an
 
Hi York,

Doing cylinders and donuts will be just fine. As a test, try drawing a piece of your domain and split it into blocks. Anything up to 20-30 blocks is easy (the largest mesh in blockMesh had something like 150 blocks but that was pretty hard to manage). Incidentally, I find blockMesh pretty good wneh you need to control mesh resolution and build high-quality meshes (say for LES).

In automatic meshing you provide the STL surface description of your domain and the rest happens automatically (with some interaction on mesh size, grading etc).

Hrv


All times are GMT -4. The time now is 07:40.