CFD - Computational Fluid Dynamics, Motorsport, Formula 1

[Alonso Fan]

some of them have ''not found'

[andylaurence]

At the risk of stating the obvious, it might be worth trying to find them. Here's what my settings show and I've been able to successfully produce meaningful results:

Code: Select all

Utility tools: 
Tool gnuplot: "C:/Program Files (x86)/gnuplot/bin/gnuplot.exe" Edit tool location Clear 
Tool gmsh: "C:/Program Files (x86)/Gmsh/gmsh-2.5.0-Windows/gmsh.exe" Edit tool location Clear 
Tool Paraview: "C:/Program Files (x86)/ParaView 3.14.1/bin/paraview.exe" Edit tool location Clear 
Tool EnSight: "C:/Program Files (x86)/CEI/bin/ensight100.bat" Edit tool location Clear 
Tool curl: "C:/PROGRA~2/Google/GOOGLE~1/Plugins/Khamsin/Third Party/curl.exe" Edit tool location Clear 
OpenFoam library and solvers: 
Tool openFoam-settings: "C:/Program Files (x86)/blueCFD-SingleCore-2.1/OpenFOAM-2.1/etc/batchrc.bat" Edit tool location Clear 
Tool openFoam-mpisettings: *** Not Found *** 
Tool groovyBC: *** Not Found *** 
Dolfyn solvers: 
Tool Dolfyn: "C:/Program Files (x86)/Dolfyn/dolfyn.exe" Edit tool location Clear 
Tool gmsh2dolfyn: "C:/Program Files (x86)/Dolfyn/gmsh2dolfyn.exe" Edit tool location Clear 
Tool DolfynPreprocessor: "C:/Program Files (x86)/Dolfyn/preprocessor.exe" Edit tool location Clear 

[Alonso Fan]

ok, thanks, i will try it

[julien.decharentenay]

Alonso, How are you going with setting up Khamsin's third parties. Let me know if there is anything I can help with.

On similar note, I have recently put online a web portal https://aerodynamicondemand.appspot.com to submit CFD analysis with just car geometry, speed and yaw. The up side is that it does not require the painful installation process of Khamsin's third parties (nor do you need to keep your computer running and some post-processing is done for you). The downside is that it is not free...

[KeithYoung]

Hey guys,

I am preparing to write articles on Bluff Bodies (or Blunt Bodies). The goal will be to compare changes in things such as Diffuser Length, Diffuser Angle, Rake, various arrangements such as Skirts, and an invention of mine that I was told had some promise by one of the top experts in the field.

The computer I have all my Open Source CFD installed on only has 2GB of ram.

I was wondering if anyone is comfortable with getting consistent results, and is willing to run the simulations on the models I send them?

You need to be able to mesh reliably with many cells in the Boundary Layer.

I need you to be able to find the lift and drag on the Body (I don't know how to find lift and drag in Paraview), and export an Ensight file for me to be able to use with Paraview.

Those interested can reply here, PM me, or preferably contact me via email which is listed on my website. Be ready to answer questions about things such as residuals, Reynolds Number, and solvers etc.

The results of this study will be posted on my website (posted in my signature), as well as in these Forums.

You will be given credit for the CFD analysis, and any other help you provide.

I anticipate choosing the person for the project in a week, around 9-8-2013.

Applications are being accepted now.

One last thing. Who has access to a wind tunnel at their University, as well as the ability to make a simple model with static ports? When I was encouraged by this expert to pursue the research on my invention, the school gave me wind tunnel time but didn't really have the ability to make the model I needed. I intend to validate the CFD with Wind Tunnel Tests if possible.

Recruiting Page

Thanks,

[KeithYoung]

The work already began days ago. Sorry if you found this late and didn't get a chance to apply.

I expect I can finish up this first article shortly.

Sorry to those of you that applied and weren't chosen.

[LogicPro]

Hi all. I am trying to learn the basics of CFD analysis for external flows (I'm using Autodesk CFD and following wiki.autodesk tutorials)...and I was wondering if there's any "standard" dimensions to create the external air volume in order to get meaningful results. I couldn't find any information about dimensions... just looking at some pictures I figured out that the air volume may be around 5 times the vehicle wheelbase on the x axis and 4 times the vehicle height on the z axis, but I don't know if there is some sort of "rule".
And does the air volume depend on the air speed or on other parameters?

[KeithYoung]

Most typically what I see people worry about is the "blockage".

Googling Wind Tunnel Blockage got me a ton of results.

I try to aim for 10% blockage or less. If you have a computer with low ram you may find yourself constantly battling with these issues like I do.

[LogicPro]

Most typically what I see people worry about is the "blockage".

Googling Wind Tunnel Blockage got me a ton of results.

I try to aim for 10% blockage or less. If you have a computer with low ram you may find yourself constantly battling with these issues like I do.

Ok, thanks for the hint: so I will select the cross-sectional area of the air volume so that the projected area of the car is around 10% of it. But what about the length of the air volume? Is it enough to have 5 or 6 times the car wheelbase (with the car positioned towards the inlet... for instance at 1/3 of the whole length of the air volume) ?
I know I may just do some experiments with different size...but I have no experience on CFD and I am just trying to get some "guidelines" to obtain reasonable results.

[KeithYoung]

The bigger the better. In my case due to mesh limitations I often only have enough ram for 3-4 carlengths behind the car.

One issue is you want tight mesh behind the car or from my experience it seems any vortex or disturbed flow dies down very quickly in a course mesh.

My advice is to just start meshing, and see what it does to your ram then adjust accordingly.

[MadMatt]

From my personal experience, blockage ratio should be lower than 5%, and flow domain should extend 2 car lengths in front and 5 behind the car. As Keith said, problem then is RAM. Try to use control volumes in specific regions to overcome this but if you want sufficient precision and quality, you should have at least 8GB of RAM

[Just_a_fan]

Extra RAM is a worthwhile performance upgrade and not too expensive either. I'm running with 8GB on my "old" quad core Q6600 which was put together to run some fire-related CFD stuff a couple of years ago.

[LogicPro]

Thanks all; I'll follow your advices!

[julien.decharentenay]

A good test case if you are just starting is the Ahmed body - a very simplified car body. It has been quite well documented and researched. You can download some experimental results from http://cfd.mace.manchester.ac.uk/cgi-bi ... &1&0&0&0&0 [You need to register (free) to download the results].

If you are trying the above, my advice is to initially use a rear end slant angle of less than 25 degree (i.e before flow separation). Once you build experience with a slant angle of less than 25 degree, try higher slant angle.

[techF1LES]

Not CFD, but still interesting... vortex formed behind the Red Bull's RWEP trailing edge as a result of the lift-induced drag. Gif is too large for embedding [9.56MB], so click here or on the picture below to see full animation of the vortex build-up: