CFD - Computational Fluid Dynamics, Motorsport, Formula 1

[flyboy2160]

smokes and thor may be confusing the "free" simple floworks tool that comes with solidworks and the very capable flow simulation add-on. if your comments are indeed directed at flowsim 2012, you guys are 100% in error. and i had wind tunnel and in-flight optical distortion data to prove that.

[Th0R]

I just wanted to mention that neither autodesk nor solidWorks should be used for simulation exhaust gas flow.

I actually do not have any knowledge about autodesk cfd functionallity...

image

acutally the post before mine was aiming at solidworks. And to be honest i still doubt that autodesk cfd is that much better, i would like to see proof for that.

And yes i was referrring to the cfd tools supplied with the professional version. But they are not shipped with the standard version. If there are any addons for cfd simulations it would be nice to refer to them with their name and not with 'solidworks'.

And still my argument stands that it doenst make sense to simulate exhaust gas flow with any software as a semi - professional, if you have no data to back up your simulation. Even the formula one teams didnt get their designs right on the first try....

[slimjim8201]

acutally the post before mine was aiming at solidworks. And to be honest i still doubt that autodesk cfd is that much better, i would like to see proof for that.

Product names can be confusing these days with so many tools on the market. When Autodesk purchased Blue Ridge Numerics last year, they chose to rename the product "Autodesk Simulation CFD". The former name was CFdesign.

Simulation CFD is standalone product and while it integrates with Autodesk Inventor (and Fusion, Pro/E, SolidWorks, SolidEdge, UG NX, Spaceclaim, and all major cad packages) it is a unique product. Autodesk has over 170 unique products today.

Part of the confusion is that SolidWorks includes a bare-bones, limited functionality CFD tool with their CAD package. This is NOT to be confused with the full-blown "SolidWorks Flow Simulation". Gotta love generic names like "Flow Simulation" and "Simulation CFD"

[Th0R]

your right, i see my mistake now. But i was assuming that most people here dont have the very expensive commercial cfd tools. That's why i understood solidworks cfd as the one which is shiped with the professional edition.

Is there an overview of all the cfd - software products available somewhere, that would be nice to have.

[Cam]

Good read for CFD and the Diffuser from 2011.

LARGE EDDY SIMULATION OF THE FLOW AROUND A DIFFUSER-EQUIPPED BLUFF BODY IN GROUND EFFECT
Unsteady Large Eddy Simulation (LES) is carried out for the flow around a bluff body equipped with an underbody rear diffuser in close proximity to the ground, representing an auto- motive diffuser. The goal is to demonstrate the ability of LES to model underbody vortical flow features at experimental Reynolds numbers (1.01 × 106 based on model height and incoming veloc- ity).

ftp://202.38.89.18/incoming/ASME/data/p ... -62673.pdf

[flyboy2160]

Good read for CFD and the Diffuser from 2011.....

Why did they use a bluff body for the test? To match some previous wind tunnel model? Why not have a knife edge at the top of the diffuser?

It seems to me that the bluff body flow - the antithesis of what the F1 cars are trying to do - would interfere with the most accurate evaluation of the diffuser flow. But I've never run diffuser CFD...

[N12ck]

Good read for CFD and the Diffuser from 2011.....

Why did they use a bluff body for the test? To match some previous wind tunnel model? Why not have a knife edge at the top of the diffuser?

It seems to me that the bluff body flow - the antithesis of what the F1 cars are trying to do - would interfere with the most accurate evaluation of the diffuser flow. But I've never run diffuser CFD...

the problem with diffuser CFD is that you need all the bodywork ahead (front wing +wheels etc) to get an effective result, which is why the only proper way to run CFD is either on wing profiles, or on a full car to be effective,

[flyboy2160]

..

the problem with diffuser CFD is that you need all the bodywork ahead (front wing +wheels etc) to get an effective result, which is why the only proper way to run CFD is either on wing profiles, or on a full car to be effective,

i complelely disagree.

to get actual car downforce numbers, you may need everything ahead of the diffuser - and maybe even in a wind tunnel. but many times in engineering studies it is necessary to make simplifications (for cost or time constraints) to factors that are either too difficult to study or, if simplified, allow a reasonable study other factors.

you don't need a full up f1 car body to study certain design parameters of diffusers. there are previous studies - referred to on this forum - that appear valid and that use a simplified body ahead of the diffuser.

i stick to my assertion that a flat topped body with a knife edge upper diffuser would give more representative results than the bluff body used here. not actual car downforce numbers, but enough to make some assumptions about what the diffuser should look like as a guide for more detailed studies.

[N12ck]

..

the problem with diffuser CFD is that you need all the bodywork ahead (front wing +wheels etc) to get an effective result, which is why the only proper way to run CFD is either on wing profiles, or on a full car to be effective,

i complelely disagree.

to get actual car downforce numbers, you may need everything ahead of the diffuser - and maybe even in a wind tunnel. but many times in engineering studies it is necessary to make simplifications (for cost or time constraints) to factors that are either too difficult to study or, if simplified, allow a reasonable study other factors.

you don't need a full up f1 car body to study certain design parameters of diffusers. there are previous studies - referred to on this forum - that appear valid and that use a simplified body ahead of the diffuser.

i stick to my assertion that a flat topped body with a knife edge upper diffuser would give more representative results than the bluff body used here. not actual car downforce numbers, but enough to make some assumptions about what the diffuser should look like as a guide for more detailed studies.

If you want to develop a diffuser accurately you need to have the bodywork aswell, you can make a generic diffuser which is optimized as a basic diffuser using your methods as discussed, but to optimize a diffuser for a particular car you do need the bodywork ahead of the diffuser, to know fully how the air interacts with the diffuser

[Jersey Tom]

Preventing flow separation at the tail end of the diffuser seems like it would be a key thing. By itself, you can't run very aggressive divergence angles because of this reason (similar for divergent nozzle design if I recall correctly).

I'd be more curious about the interaction of the rear wing on keeping flow attached.

[Just_a_fan]

Good read for CFD and the Diffuser from 2011.....

Why did they use a bluff body for the test? To match some previous wind tunnel model? Why not have a knife edge at the top of the diffuser?

It seems to me that the bluff body flow - the antithesis of what the F1 cars are trying to do - would interfere with the most accurate evaluation of the diffuser flow. But I've never run diffuser CFD...

From the abstract and a skim read of the paper:

• They used a bluff body with diffuser because it represents an automotive diffuser [note: not an F1 diffuser specifically, an automotive diffuser i.e. it's a generic vehicle shape].
  
  They did the work to try to see if using LES would give insight in to how vortices form and develop in such a diffuser [note that they're using LES rather than the more usual RANS here - because LES deals with eddies (vortices are eddies). In general terms RANS is cheaper, computationally, than LES. But as computational power becomes cheaper then using LES on complicated models becomes more realisitic. So they are doing research on it - they even state that they're not aware of an LES approach having been tried on a bluff-bodied diffuser; it didn't show up in the research they carried out although other LES modelling of diffusers had been carried out.].
  
  The idea would be to allow the development of better vortex control [because if you better understand how/why they form then you have more chance of dealing with them].
  
  Having compared RANS and LES simulations, they compared the results to an identical bluff-bodied wind tunnel model and saw "a good level of agreement".

That's why the reference was given...

[stez90]

check out this new opensource GUI for OpenFoam. Look promising and user-friendly! http://engys.com/products/helyx-os

[godlameroso]

I hope to see that cars for next year are designed from the ground up to be the best at what they do most. I'm sure by now most teams have extensive data on what exactly a car does most during a specific race and the telemetry is so ultra analyzed that they even understand what parts of a track give them the most trouble. Perhaps it would be wise to design a car around exactly the sort of turns one experiences on the calendar the most frequently, then one would be able to take into account where exactly the car needs yaw stability, where it needs acceleration stability, or braking stability. Perhaps teams already do this, albeit perhaps not employing it as a conscious strategy.

[techF1LES]

Actual screens from Williams' design office

[shelly]

Thes pictures are a great find. We can see the floor suction peaks at the front and on the kink line, and the low pressure footprint of the bargeboard vortex protruding from the front suction peak.

There are a lot of other interesting details, such as the shape of the tyre contact patch.

Should help a lot in discussing underbody aero