Page 103 of 160
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 04:30
by wogx
Those swooshy RW's - are these aero-effective in their current (showcar) form? Is there a chance, that we will see more "flat" surfaces there during 22' season?
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 14:28
by jjn9128
wogx wrote: ↑02 Jan 2022, 04:30
Those swooshy RW's - are these aero-effective in their current (showcar) form? Is there a chance, that we will see more "flat" surfaces there during 22' season?
The show cars are based on FOMs working models so they're effective - teams will produce more downforce though.
As for the spoony rear wing... the tip has to come quite high to hide some surface for the endplate.
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 17:27
by godlameroso
BlueCheetah66 wrote: ↑01 Jan 2022, 19:43
For the noses next year, are teams still allowed to have a sharp change in the width of the nose, similar to what Mercedes have had for a while or has that been closed off in these regulations?
You can have an 'eagle' beak type nose. What they can't do is put any sort of winglets on the nose because they prescribe no concave surfaces.
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 17:31
by godlameroso
jjn9128 wrote: ↑01 Jan 2022, 21:59
aleks_ader wrote: ↑01 Jan 2022, 20:51
Hoffman900 wrote: ↑01 Jan 2022, 20:49
Who did the design of the FOM cars?
There seems to be at least some thought put into them.
Tombazis aero group with Symonds.
Tombazis is FIA. Symonds group is FOM, Jason Somerville led the aero design group under Symonds.
The FOM technical Hierarchy is:
1) Brawn
2) Symonds
3) Somerville (aero), Wilson (vehicle performance)
4) 3/4 aeros and surfacers + a couple of suspension and gearbox engineers
FOM created the car(s) and hand that off to Tombazis and the FIA who then drafted the rules with the teams. There was some crossover with the FIA and FOM but that's how it worked.
And all of them are out. Curious.
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 19:18
by Zynerji
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 21:52
by Stu
Zynerji wrote: ↑02 Jan 2022, 19:18
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
They are doing it to an extent (and that is controlled by the wind tunnel/CFD usage regulations. It is only a matter of time before limits are put on general data usage in other areas (simulation, FEA - also very effective used iteratively - and during race weekends)
Re: 2022 Aero Thread
Posted: 02 Jan 2022, 23:41
by godlameroso
Stu wrote: ↑02 Jan 2022, 21:52
Zynerji wrote: ↑02 Jan 2022, 19:18
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
They are doing it to an extent (and that is controlled by the wind tunnel/CFD usage regulations. It is only a matter of time before limits are put on general data usage in other areas (simulation, FEA - also very effective used iteratively - and during race weekends)
They already had the right idea back in 2018's outwashing front wing. The move to curved diffuser from the boxy designs of 2017-2018 vintage shows that geometry works. IIRC Haas was the first team to try the curved diffuser.
To quote a user here..."ellipses everywhere".
Re: 2022 Aero Thread
Posted: 03 Jan 2022, 00:58
by Zynerji
Stu wrote: ↑02 Jan 2022, 21:52
Zynerji wrote: ↑02 Jan 2022, 19:18
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
They are doing it to an extent (and that is controlled by the wind tunnel/CFD usage regulations. It is only a matter of time before limits are put on general data usage in other areas (simulation, FEA - also very effective used iteratively - and during race weekends)
CFD is controlled by TFLOPS. That can be sidestepped by analog co-processors that are much, much faster at the Navier Stokes computations.
https://penntoday.upenn.edu/news/penn-e ... -equations
That with a ML algorithm that iterates continuously after being trained on Petabytes of previous data can churn out some stunning things for the current limitations on computer hardware.
Re: 2022 Aero Thread
Posted: 03 Jan 2022, 14:05
by mzso
Zynerji wrote: ↑02 Jan 2022, 19:18
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
I think what you're thinking of are called evolutionary algorithms. But they certainly don't explore "everything", because that's impossible, it would be practically infinite.
Re: 2022 Aero Thread
Posted: 03 Jan 2022, 19:58
by Zynerji
mzso wrote: ↑03 Jan 2022, 14:05
Zynerji wrote: ↑02 Jan 2022, 19:18
I'm surprised that we haven't talked about the real elephant in the aerodynamic design room...
Machine learning, iterative design software that "explores" everything legal within the minimum and maximum dimensions.
I wish the teams shared more about this stuff.
I think what you're thinking of are called evolutionary algorithms. But they certainly don't explore "everything", because that's impossible, it would be practically infinite.
By definition. Having a minimum and a maximum sets boundaries that are finite.
Just because the number is big does not mean that it goes on forever.
With enough computation, just reducing the possible upgrade paths is worth its weight in gold.
Re: 2022 Aero Thread
Posted: 03 Jan 2022, 20:31
by BorisTheBlade
Zynerji wrote: ↑03 Jan 2022, 19:58
By definition. Having a minimum and a maximum sets boundaries that are finite.
By which definition? You do realise that even between the borders of 0 and 1 there is an infinite amount of values in-between?
Re: 2022 Aero Thread
Posted: 04 Jan 2022, 01:09
by Zynerji
BorisTheBlade wrote: ↑03 Jan 2022, 20:31
Zynerji wrote: ↑03 Jan 2022, 19:58
By definition. Having a minimum and a maximum sets boundaries that are finite.
By which definition? You do realise that even between the borders of 0 and 1 there is an infinite amount of values in-between?
Changing the ruler doesn't mean the object changes with your description. It still is limited by real constraints. Digital to Analog nit-picking aside.
Re: 2022 Aero Thread
Posted: 04 Jan 2022, 01:59
by godlameroso
More importantly, the tunnels aren't straight, there will be compression waves when airflow encounters a concave surface, and there will be expansion waves when air travels on a convex surface. Compression waves slow down air and increase drag, while expansion waves speed up air and slightly decrease drag locally. The geometry to optimize these phenomena will require many studies and iterations. I don't think machine learning will be that valuable until the meat sacks get some learnin' done.
Yes there will be sonic flow and choking effects, exploiting that will be key to getting the most performance from the floor.
A venturi tunnel speeds up air, if the car is going 33 meters per second, the local speed at the venturi tunnel can be as high as 100 m/s 50% increase in speed, and you're somewhere in the sonic regime ballpark, ie choked flow. The only way to increase mass flow at this point is to increase the pressure of the inlet, or air temperature, as ambient temperature affects the Mach number of air. If you achieve choked flow, the pressure downstream will drop like a neutron hammer. The airflow is literally choked at the inlet, like sand grains passing through an hourglass. Such a sharp pressure reduction will summon any surrounding air to fill that "void".
Re: 2022 Aero Thread
Posted: 04 Jan 2022, 04:13
by Hoffman900
godlameroso wrote: ↑04 Jan 2022, 01:59
More importantly, the tunnels aren't straight, there will be compression waves when airflow encounters a concave surface, and there will be expansion waves when air travels on a convex surface. Compression waves slow down air and increase drag, while expansion waves speed up air and slightly decrease drag locally. The geometry to optimize these phenomena will require many studies and iterations. I don't think machine learning will be that valuable until the meat sacks get some learnin' done.
Yes there will be sonic flow and choking effects, exploiting that will be key to getting the most performance from the floor.
A venturi tunnel speeds up air, if the car is going 33 meters per second, the local speed at the venturi tunnel can be as high as 100 m/s 50% increase in speed, and you're somewhere in the sonic regime ballpark, ie choked flow. The only way to increase mass flow at this point is to increase the pressure of the inlet, or air temperature, as ambient temperature affects the Mach number of air. If you achieve choked flow, the pressure downstream will drop like a neutron hammer. The airflow is literally choked at the inlet, like sand grains passing through an hourglass. Such a sharp pressure reduction will summon any surrounding air to fill that "void".
Do you have any data to suggest that the venturi's can come
anywhere near choke flow?
Re: 2022 Aero Thread
Posted: 04 Jan 2022, 04:45
by wogx
Would the incorporation of e.g. heat pipes in the top surface of Venturi tunnels have any positive effect on the air flow? Would it be legal?