Hello!

I'm currently near the end of my Ph.D. dissertation, and one of my committee members says that he'd like to see some 3D results as well as the 2D results that I've already created. But all the flow solvers I have now -- which are mostly things I've written myself -- are 2D (and a bit slow), so I'm looking around for a quick way to do some 3D calculations.

Essentially, what I need is this: a 3D box discretized with a regular cartesian grid of cubic grid cells, containing incompressible laminar flow with standard uniform-velocity inflow conditions on one and uniform-pressure outflow conditions on the other, and some form of symmetry conditions on the sides.

Furthermore (and this, of course, is the key piece), I have a distributed body force to apply to this flow, which amounts to a datafile containing a force to be applied to each grid cell in the mesh.

I'm wanting to have a steady-state solution for this flow, but a time-accurate solution that's been run to steady state would be ok. I expect the average mesh size will be something on the order of 32x32x128, and I'm hoping to get results on a desktop PC in a reasonable timeframe.

The main constraint, for me, is that ideally I'd like to be able to go from a state of downloading the 3D flow solver to running my calculations with no more than about two weeks of work. Thus, I need a program which is relatively easy to figure out how to use, and in which it will not be too difficult to implement this distributed body force field.

Would you recommend OpenFOAM for this purpose? What will be the hardest parts of doing this with OpenFOAM, and how difficult do you expect them to be? Are there other CFD programs that might be better for what I need?

Thanks very much!

[mattijs]

Almost all public domain or commercial solvers should be able to do your problem.

re. OpenFOAM:
- you should be able to do this in two weeks.
- hardest parts: inserting your body force in the equations

Main problem is the lack of documentation re. your specific problem. But you should be able to find enough on this site or on the wiki.

- generate a mesh: blockMesh

- reading file with coordinates:
IFstream str("coord.xyz");
pointField cellCentres(str);

- finding cell at point: search this site for 'findCell'.

- adding to e.g. simpleFoam. Look in the source code and add to the equation for U.

Thanks for the advice!

It does look like this wouldn't be too hard to figure out, indeed. (If nothing else, I could just look at the answers to the most recent questions -- seems I'm not the only person wanting to solve a flow with a body force!)