Aero analysis for F1 2022 based on CFD result

[Fluido]

Thank you very much, it's soooo encouraging !
Doing analyze is very interesting and good way to kill time, feeling closer to the amazing sport.

Is it more work doing geometry or CFD?
Can you post specifications of your PC/ work station that you used for this?

[Latios]

Thank you very much, it's soooo encouraging !
Doing analyze is very interesting and good way to kill time, feeling closer to the amazing sport.

Is it more work doing geometry or CFD?
Can you post specifications of your PC/ work station that you used for this?

CAD causes more, since it includes: read regulaction, create regulation's control volume/surface/limits, then draw the geo.
PC:
for CAD, just normal PC, 9700K+2080S+64GB RAM.
for CFD: an year-2016 old work station, 32 cores (each's speed is much lower than 9700K) and 64GB RAM

[Fluido]

for CFD: an year-2016 old work station, 32 cores (each's speed is much lower than 9700K) and 64GB RAM

Which CPU is that ,i7-9700K has 8 cores?

[Latios]

for CFD: an year-2016 old work station, 32 cores (each's speed is much lower than 9700K) and 64GB RAM

Which CPU is that ,i7-9700K has 8 cores?

The PC for CAD is using i7-9700K, has 8 cores.
The PC for CFD is using Xeon E5-2667 v4, 3.2GHz

[Fluido]

The PC for CAD is using i7-9700K, has 8 cores.
The PC for CFD is using Xeon E5-2667 v4, 3.2GHz

2x xeon gold 6134 (8cores, base 3.2Ghz, turbo 3.7Ghz,cache 24,75MB)
or
2x xeon gold 6142 (16cores ,base 2.6Ghz, turbo 3.7GHz, 22MB cache)

Which one is better for CFD?

[Latios]

The PC for CAD is using i7-9700K, has 8 cores.
The PC for CFD is using Xeon E5-2667 v4, 3.2GHz

2x xeon gold 6134 (8cores, base 3.2Ghz, turbo 3.7Ghz,cache 24,75MB)
or
2x xeon gold 6142 (16cores ,base 2.6Ghz, turbo 3.7GHz, 22MB cache)

Which one is better for CFD?

Not sure... You can use a simple CFD task to test.
Memory is also critical, it must be enough for your mesh.

[Fluido]

is better for CFD?


Not sure... You can use a simple CFD task to test.

You mean which one solve first?

[Latios]

is better for CFD?


Not sure... You can use a simple CFD task to test.

You mean which one solve first?

Yes

[Fluido]

is better for CFD?


Not sure... You can use a simple CFD task to test.

You mean which one solve first?

Yes

GHz vs cores

[Latios]

You mean which one solve first?

Yes

GHz vs cores

I mean you need to test.
With cores increases, the calculation speed is not increased linearly.

[Fluido]

Yes

GHz vs cores

I mean you need to test.
With cores increases, the calculation speed is not increased linearly.

when turbo starts?
If both has same turbo ghz doesnt that mean this with more cores is faster?

[vorticism]

...

What CAD program did you use for the model? Parametric, polygonal?

Might be fun to see:

-Model run in reverse flow, maybe 200kph, to see how much lifting capacity it has if spun 180*
-Various positive rake angles to see what angle of attack would be needed to make the car go airborne during an impact
-duplicate the model and test following distances (drafting)
-Mercedes sidepods effects
-Mclaren bib/t-tray effects (vs. what looks like RB18 version in your existing model)
-put a non-flat road surface to simulate riding the curbs (gap from floor to road increases), Mick Schumacher crash comes to mind

-omit the front wheel fairings/wings (do they actually do much?)
-Red Bull stacked beam wing effects
-Test a Monaco spec with the biggest wings you can fit in the legality volumes
-Test a large version of the Ferrari sidepods (essentially make the largest sidepods that can fit in the legality volumes)

[jjn9128]

What CAD program did you use for the model? Parametric, polygonal?

Might be fun to see:

-Model run in reverse flow, maybe 200kph, to see how much lifting capacity it has if spun 180*
-Various positive rake angles to see what angle of attack would be needed to make the car go airborne during an impact
-duplicate the model and test following distances (drafting)
-Mercedes sidepods effects
-Mclaren bib/t-tray effects (vs. what looks like RB18 version in your existing model)
-put a non-flat road surface to simulate riding the curbs (gap from floor to road increases), Mick Schumacher crash comes to mind

-omit the front wheel fairings/wings (do they actually do much?)
-Red Bull stacked beam wing effects
-Test a Monaco spec with the biggest wings you can fit in the legality volumes
-Test a large version of the Ferrari sidepods (essentially make the largest sidepods that can fit in the legality volumes)

Definitely these 3 would definitely be interesting. A lot of "copy team A" you can see trends but not really their flowfield - plus Vanja already did some of that - to me the big picture things are more interesting.

[Stu]

What CAD program did you use for the model? Parametric, polygonal?

Might be fun to see:

-Model run in reverse flow, maybe 200kph, to see how much lifting capacity it has if spun 180*
-Various positive rake angles to see what angle of attack would be needed to make the car go airborne during an impact
-duplicate the model and test following distances (drafting)
-Mercedes sidepods effects
-Mclaren bib/t-tray effects (vs. what looks like RB18 version in your existing model)
-put a non-flat road surface to simulate riding the curbs (gap from floor to road increases), Mick Schumacher crash comes to mind

-omit the front wheel fairings/wings (do they actually do much?)
-Red Bull stacked beam wing effects
-Test a Monaco spec with the biggest wings you can fit in the legality volumes
-Test a large version of the Ferrari sidepods (essentially make the largest sidepods that can fit in the legality volumes)

Definitely these 3 would definitely be interesting. A lot of "copy team A" you can see trends but not really their flowfield - plus Vanja already did some of that - to me the big picture things are more interesting.

By “big picture” are you talking about how the aero works as a whole, how different parts interact or where the bigger gains are coming from?

[jjn9128]

By “big picture” are you talking about how the aero works as a whole, how different parts interact or where the bigger gains are coming from?

As opposed to the minutiae of design variations between teams. Because that's so dependent on getting the whole system working.