CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   Looking for a pimpleFoam tutorial using Salome (and hardware recommendations?) (http://www.cfd-online.com/Forums/openfoam/76488-looking-pimplefoam-tutorial-using-salome-hardware-recommendations.html)

madact May 26, 2010 23:30

Looking for a pimpleFoam tutorial using Salome (and hardware recommendations?)
 
I have working installs of Salome and openFOAM, and want to get straight into some simulation - I have read through the 2D tutorials, but am still somewhat at a loss (one might say 'impatient'...) to go from this to a 3D vehicle simulation scenario.

The problem to which I want to apply this is human-powered vehicle design (for racing): the target is for a cruise speed of at least 16m/s at 140W (including drivetrain and rolling losses) for a vehicle approximately 3m long and 1m wide. I have been told "just find a nice low-drag foil section, apply the cube-of-sqrt method to convert to a rotated 3D body, and concentrate on precise construction and sources of parasitic drag" which sounds like good advice, but I think there may be more to it than that :D.

Really, I just want some help setting up a 'virtual rolling-road wind tunnel' that can evaluate a a couple of cross/head/tail-wind scenarios, find a good procedure for meshing, simulation and viewing, and get on with low-drag body design. Simulation of lateral aerodynamic forces on the vehicle when turning would be helpful, too, as the events I plan on entering are run on tracks that resemble criterium courses more than velodromes...

From what I have read, I believe the most appropriate openFOAM solver for this problem will be pimpleFOAM with a DES or DDES model (not sure of the relationship between the openFOAM solvers and the type of model here)...

As a test case to get up and running, I have modelled the Ahmed body as a Salome solid for 25 and 35 degree cases, and would like to find a walkthrough I could follow for setting up a simulation scenario, meshing the model / simulation, simulating and viewing / analysing the results. It would be very helpful if someone could could either point me to an applicable tutorial and let me know the caveats, or have the patience to walk me through the process.

I have a limited knowledge of fluid dynamics - I'm an avid model aircraft builder, I have used 2-D panel method design in XFoil with good results, and I know the basic concepts, but I'm not an expert, and have no experience with FEA/CFD. I do develop spatial data analysis software as my day job, though, so I'm not clueless about mesh and boundary structure representations - just the software and data formats you guys use to do them ;)

Some system & software information:
- Salome: 5.1.3
- OpenFOAM: 1.6
- OS: Ubuntu 10.1 Karmic 32-bit
- Hardware: Core2 Quad 3GHz / 12MB cache (Q9650), 4GB RAM, NVidia 9800GT 512MB

P.S. Does anyone have any hardware recommendations? I have a limited budget, so I can't build a cluster :(, but I am curious whether going to 64-bit Linux and/or upgrading to 8GB RAM would help with simulation times or size capability in openFOAM? This is the one potential upgrade well within my budget...

madact May 27, 2010 01:24

Some references
 
1 Attachment(s)
FYI, the following two presentations describe pretty much the kind of result I want to replicate in my initial testing / validation, except that the first uses a RANS / simpleFOAM model and commercial mesher, and the second uses Blender / Calculix (which look like a decent option, too).

Both more-or-less skip everything between mesh generation and viewing the results (apart from providing some relevant simulation parameters). Neither go into enough detail to really follow step-by-step...

https://online.tu-graz.ac.at/tug_onl...cumentNr=81599

http://www.imac.unavarra.es/Cursos/A...thOpenFOAM.pdf

I've also attached iges models for the Ahmed 35 degree body, and a 'wind tunnel' volume 10m x 10m x 60m (20m upstream and 40m downstream) which I'm trying to simulate.

The compound body meshes well enough using the procedure demonstrated here:
http://openfoamwiki.net/index.php/IdeasUnvToFoam
but I'm pretty sure this is not the kind of mesh I need to be generating...


All times are GMT -4. The time now is 16:23.