F1technical.net

Blackout wrote:I have a question. How do you call that 'technique' when you smooth the angles/add a radius to an angle to improve the airflow
(red in this rough illustration)
http://www.servimg.com/image_preview.ph ... u=14795526

Has someone a better illustration ?

Chamfer works too,

Hi,

I started with some simulations in Ansys, but I have some questions about it.

Is here someone who does CFD analyses for a job who wants to help me with my university project?
And a more urgent question: If you simulate forces on a 2D airfoil analyses. You print them, what is it exactly that you can use as the lift or drag? (you get a bunch of numbers and I don't know which one is usefull.)

Hope someone can help me.

A good place to find answers to this kind of qusetions is cfdonline.com

Not much, older model, but...

[youtube]http://www.youtube.com/watch?v=P3ulqzRR9UA[/youtube]

Another CFD simulation from Sauber

[youtube]http://www.youtube.com/watch?v=-Iyh-S-LGN8[/youtube]

Does anybody has a CATIA part (assembly) of a F1 car? I need one so I can make a graphical design for my final year thesis. It would be great if I can present a F1 car with the logo's of all my sponsors.

I hope someone can help me.

[youtube]http://www.youtube.com/watch?v=pREgFVRXeDI[/youtube]

Nando wrote:is CFD so advanced that it calculates individual molecules, or groups of molecules or does it work by other principles?

Where do they draw the limit?

Well... most on here will talk about and use solvers based on Navier-Stokes, which indeed assumes continuous fluid properties over a reasonably small space. For N-S, the individual molecules are never modelled or considered. Instead only fluid properties are considered.



There are others which use Lattice-Boltzmann... which track particles, not molecules (in this context, a particle could be considered a group of molecules, but of non-fixed size) and according to the powerpoint, they are great. But if I had a pound for every time powerpoint folks says their software will do A, B and C... and it doesn't.... I'd be a millionaire. Basically, LBM is still a work-in-progress, but it does show promise.

Thanks for this Kilcoo, I was only familiar with Navier-Stokers method of solving it, which is most probably the one used by almost everybody, including the CFDs on F1 teams departments

TDH wrote:Does anybody has a CATIA part (assembly) of a F1 car? I need one so I can make a graphical design for my final year thesis. It would be great if I can present a F1 car with the logo's of all my sponsors.

I hope someone can help me.

I am afraid this will be difficult. There are some guys that have modeled imaginary F1 cars in the Engineering projects section that might do this, but I really do not know of any existing model available.

kilcoo316 wrote:

Nando wrote:is CFD so advanced that it calculates individual molecules, or groups of molecules or does it work by other principles?

Where do they draw the limit?

Well... most on here will talk about and use solvers based on Navier-Stokes, which indeed assumes continuous fluid properties over a reasonably small space. For N-S, the individual molecules are never modelled or considered. Instead only fluid properties are considered.

LMB does show promise, but can be quite demanding in terms of resources ( I have seen models implementing LMB and requiring significantly more memory for the discrete probability dsitribution algorithms than we normally would have in classical/typical models).

There is however one thing that needs to be taken into consideration: it is perfectly suited for solving NS equations, but it does require some effort in order code algorithms that solve anything beyond NS.

There are other disadvantages (conditional stability, transient scheme, lack of flexibility), but it is gaining support (at least, I read/hear more clients/researchers probing methods based on LMB)

Just_a_fan wrote: We'll probably never need to carry out molecular-CFD on things like cars because the RANS/LES/DNS models will be "close enough". This is just as well as we'll probably never be able to do molecular-CFD on cars...the numbers are just stupidly huge!

To say never is in my view is shortsighted. If the physics are understood then the only thing stopping such a simulation is computing power, and to say never is to put a limit on the computing power achievable, and I don't believe that will happen. Whether it is ever useful to run such an in depth simulation is cause for speculation, but it is unlikely as you say. But maybe in 100 or 200 years time, a researcher somewhere with a large amount of computing resources available will decide to run such a simulation purely for shits and giggles.

I'd be hugely disappointed if we still have cars, or anything remotely like them, in 100 or 200 years. As for computing power, you're looking at having the equivalent of all of the world's supercomputers, squared, available for "shits and giggles". Sure, it might happen but only if there is nothing more interesting than studying an historic irrelevance - which is what F1 will be by then.

kosioBG wrote:All this CFD is so interesting, but so scary... i don't know where to start(i have no experience whatsoever) And i can only make 3d models with google sketchup, would that be ok or would i have to learn cad?

You need to know a lot about classically solved aerodynamics before you just model something up In CAD and push the CFD button.

You must have some hand calculated (ok, ok or spread sheet calculated) estimate to ensure the CFD result is reasonable. I've seen analysts at work (usually young ones) refuse to do a hand calc and get led far astray from the correct answer because of something as simple as entering fluid properties incorrectly. (After I did a hand calc on a kid's suspect CFD result, the kid's boss actually found a numerical entry off by a factor of 10...oops)

Yes, F1 airflow is very complex over the whole car, but you need to make some simplifications to get a order of magnitude estimate of the answer. Ideally, you'd also have some tunnel data to back up the CFD result.

You absolutely cannot just push the CFD button and trust the answer.

Just_a_fan wrote:I'd be hugely disappointed if we still have cars, or anything remotely like them, in 100 or 200 years. As for computing power, you're looking at having the equivalent of all of the world's supercomputers, squared, available for "shits and giggles". Sure, it might happen but only if there is nothing more interesting than studying an historic irrelevance - which is what F1 will be by then.

If you go back 10 years, the FLOPS record was 36 Terraflops, which cost millions to install. Today you can get that sort of power with 35 $450 graphics cards. A super computer from the early 90s is as powerful as a quad core laptop processor today. I don't see it as inconceivable that in 200 years that a single person will have access to such processing power.

Maybe in a fluid mechanics class 200 years from now, students will examine case studies of F1 cars as early examples of CFD.