CFD for Thesis project

[gildasd]

Hello,

This project is going to contain some electronics (sensors, data loging etc) but right now I'm checking the premises:

A big part of the project is comparing mathematical and CFD data to real wind conditions.
I'll be doing a 3 month data collection campaign to get some statistically relevant data - It's a bit more complicated than this brief explanation, but it should give you the basic context:

I need to do "basic" CFD on a 1m by 1m by 1m square with right angles, the bottom side is neglected (on the "ground" - or deck of a ship in this case).
The results that I need to obtain:
 - Lift.
 - Drag.
 - Resulting effort in kilos (or Nm).
 - At different angles, 5° to 45° in 5° increments should do it for a symmetrical shape.

I've been trying out Flowsquare V4 http://flowsquare.com/ and it's great for visualisation, but:
 - No real numerical data.
 - Only 2D.
 - A right pain in the posterior to set up flow angles.
 - Sloooow.

I only have a PC with Windows 7 64 bit and that's it. All my funds will go in the second phase on the project.
What free CFD programs out there do you recommend?

Thanks!
G.
(And please don't propose to do for me, I need the ability to rerun stuff at the behest of my thesis overlord.)

[awallin]

OpenFOAM is the free and open-source alternative. Steep learning curve with lots of different programs used along the tool-chain (meshing, solving, post-processing, etc).
If your university has Comsol it's very much easier to start with but may not provide the advanced features and detailed control that other packages dedicated to CFD are. Ansys is good I hear but very expensive.

You'll want all the RAM you can get, 32 Gb minimum, or preferably a CPU/mb combination that supports 64 Gb.

If you are investigating situations with turbulent flow then most of the black art of this kind of simulation is selecting the appropriate approximate turbulence model that is valid for your situations.
http://en.wikipedia.org/wiki/Computational_fluid_dynamics#Turbulence_models
The big boys do simulate turbulent flow also directly (requires even more CPUs and RAM!)
http://en.wikipedia.org/wiki/Direct_numerical_simulation

[gildasd]

Thanks,

I have 16Gb of fast ram, I can't afford any more.
My University is a Maritime institute, so we have a CFD suite but only for incompressible fluids, I don't remember it's name, but getting the "airflow" module cost can be calculated in Corvettes Z06's - that's a dead end.
And I can't use my soft on the CFD farm because it's not connected to any network and simply sticking a USB stick in a terminal can get you ejected from the campus !!!!
It has to run on my desktop.

I used to use the airflow module in Solidworks to evaluate convection inside museum cases and it was pretty simple, but my key has long run out!

I've checked out OpenFOAM, it seems a bit of an overkill (grams is more than sufficient for a 1m3 shape in my case) and I don't have a month to get learn the UI for some non central data...


Edit: Found a flaw in Flowsquare; No Karman eddies if the shape is symmetrical and composed of flat sides...