CFD

[grimreaper]

Hi all, this is my first post here - very interesting site.

The aerodynamic efficiency of a wing depends on the amount of downforce vs drag generated. Is it possible to get these numbers i.e. for downforce and drag out of a CFD application. I dont expect that they will have a button labeled "downforce" but a proxy for it : maybe pressure exerted on the wing?

thx
Grim

[akbar21881]

be it downforce, be it drag, both of these are pressure difference. if i'm not mistaken, cfd shows the airflow and pressure associated with it. its the same as wing lift calculation I think. get the pressure or velocity and of course drag and lift coefficient then you will straightaway get lift(or downforce) and drag.

[Monstrobolaxa]

Nothing to be added here...!

[Becker4]

so sorry, CFD means??

[Steven]

Computed fluid dynamics... With this are meant some programs that all the teams use to simulate airflow and pressure around their vehicle.
It's nothing more (or less if you like) than a big computer program, which requires kind of a powerful computer

[Becker4]

thanks, wonderful site BTW. very informative and comprenensive

[Monstrobolaxa]

According to my sources the most used CFD program in F1 is called:

Fluent 6.0

[Steven]

Catia is also something like that isn't it?
Although this might be the rendering tool... I'm not sure here

[grimreaper]

I managed to drag out the drag force (pun intended) from the CFD code I'm using - CFDesign. I need this to calculate the CoD.

Now all I need is to find someone who'll trust me with the aerodynamics of their (race ?) car...

cya
Grim

[PaulC]

From a converged FLUENT solution, you can pretty much ask for the vector force magnitudes such as downforce and drag. Drag Coefficient and other parameters of interest (i.e. Y* turbulence) are also readily available. As you can tell from the BAR website, the constructors are demonstrating a lot of effort and capacity in using CFD. I think that it is only recently that the lowered cost of computing capability expanded the use of CFD beyond simplifed 3D models and time independent solutions (i.e. steady-state). I imagine that time dependent phenomena such as turning and running into and out of a slipstream can greatly affect induced forces and it is within the current technology state to be able to simulate (wouldn't I love to try )

Check out this link to an article about how FLUENT was used in the design of the 'toblerone panels' just under the intake of the F2003...
http://www.cfdreview.com/article.pl?sid ... ode=nested

BTW- great site, I love the analysis of car designs after each race as in the F1 Racing magazine I have started reading. I'm a new fan of F1, but I've been a user of the FLUENT code for a few years. (Fortunately my work pays for the user license and hardware, one license costs around 10k USD per year and most users have several in order to run 'parallel' processes at once)

I'll try to answer CFD-type questions as best I can, but be forewarned that I am not a full-time 'power-user' in auto/aero aerodynamics. Rather, I occasionally use it as an analysis tool for our process equipment (toner manufacturing). Cheers!

[grimreaper]

Hi PaulC

Thanks for the tips. I read that article you mentioned regarding the toblerone panel - very interesting. I'm trying to get access to Fluent as it seems to be the code of choice for racing teams. I'm not a fulltime aero guy but I love to experiment.

As regards transient analysis it seems that Fluent 6.1 does support dynamic mesh modeling - very useful for understanding what happens during overtaking and other manouevers.

[KeithYoung]

Paul, it sounds like you know CFD rather well. if you could, would you email me at keith.a.young@comcast.net, I would like to discuss CFD and Formula technology in general. I look forward to hearing from you. regards... Keith

[PaulC]

Hey, glad to see the interest in CFD. I really like how aerodynamics plays such an important role in F1 and the relative freedom the constuctors have to exploit it (unlike stock car racing in the US)

For anyone interested in the steps of Computational Fluid Dynamics from a FLUENT user's perspective, here it is:

1) Geometry: Since you are looking at fluid flow, you basically take a CAD model and subtract out the solids to create the fluid geometry. The commercial CFD companies make it a priority to be able to interact with the popular CAD programs (the future wav is the all-in-one simulation of both mechanical and fluid interactions)

2) Mesh: In this step, the fluid domain is subdivided into small cells. The smaller (and more numerous) leads to better resolution but comes at the cost of high computational resources. This is a constant a balancing act for all CFD users. In the early days I heard this step used to take a lot of time and effort, but the advent of automatic meshing tools and non-conformal prismatic mesh elements have made this easier.

3) Boundary Conditions: So now we are prescribing the constraints and starting conditions to our multi-variable problem. Things like airflow direction and magnitude; frictional coefficients for surfaces; moving objects; etc. Turbulence can also be defines in a myriad of ways. In general, the more you can input - the better quality (and efficiency) of the solution.

4) Iterate: So CFD is basically a huge solver. Given the boundary conditions, it iterates to minimize the 'residual error' in mass, momentum, and energy balance over the entire fluid domain. If the planets and stars are aligned correctly, this results in a flowfied solution that is representative of the reality.

5) Analysis: The extraction of the required data, Which usually leads to more questions and 'what-ifs' and back to another simulation!

This is very simplified, but represents the basic steps by CFD users. FLUENT offers front ends that automate the process of high-volume type wprk such as chemical reactors, electronic board cooling, fan-blades, and some others.

Hope this helps,
Paul C
Sodus, NY USA

[Guest]

I have very little fluent experience but i have a few videos i made with it, if anyone is interested email me at keith.a.young@comcast.net . On one condition though, dont ask what pressures or velocities are involved and such because i dont know . also to note is these are of simple cylinders and stuff, not formula cars or anything like that yet.

[CFDruss]

May I add, Fluent is a very much dying CFD code, they have lost alot of contracts to probably the leader of CFD code, Star-CD, which is part of the adapco group. Most F1 team are now using Star-CD. Another package that is becoming very popular is CFX (which has now been took over by Ansys (the well know finite element analysis group!!!))

I currently use both Star-CD and CFX 5.6, and I would rate both as eqaully as good, although CFX struggle a little with supersonice multiphase flows.