Williams FW43B

[jjn9128]

Hey that's me! Disclaimer: the meshes used for those are pretty poor (65M cells, no volumetric refinement). I just needed something eyecatching for now, so I went with images that don't make too many claims. Expect real results in ~6 weeks.

Can I just congratulate you on using coefficients

[ScrewCaptain27]

Looks like all the wheel internals are missing

They are far too roughly modelled to be any use.

That's just like, your opinion, man.

https://imgur.com/lLOOF3o

(disclaimer, garbage mesh)

Did you remodel them? They looked pretty poor when I tried extracting them into Blender and also into Solidworks.

By the way, looking forward to the full article! Will you also mention what rig setup you are using?

[jjn9128]

Did you remodel them? They looked pretty poor when I tried extracting them into Blender and also into Solidworks.

By the way, looking forward to the full article! Will you also mention what rig setup you are using?

My opinion was the front drums looked good but the rears looked like crap. Brake discs and calipers looked good though. The rear drums winglets are responsible for a fair bit of downforce and helping the diffuser so them being so low quality can't be helping.

What I can't figure out is how Variante and DP_CFD are working around the cooling ducts - which as far as I can tell terminate inside the body. Sealing it off and using an outlet?

[DP_CFD]

65M cells is more than what Toyota use for their Le Mans cars.
It might not be an optimized mesh, but it's definitely not poor.

https://i.imgur.com/m2KOqax.png
(14M cells here. And Williams wasn't kind enough to provide internal aero, but that doesn't have such an impact for most uses)

It's basically just a decent surface mesh with 14 prisms. I'm told 200M is a good number to aim for, although that's probably including internals. Did you cap off the wheels to avoid the bad geometry, or are you cooking up some MRFs?

Did you remodel them? They looked pretty poor when I tried extracting them into Blender and also into Solidworks.

By the way, looking forward to the full article! Will you also mention what rig setup you are using?

The front ones were fine, I just had to remodel a surface named "structure" that was completely garbage. The rear ones are fairly rough as you know, but definitely still usable. Right now I'm just doing simulation setup on my CFD workstation, AMD EPYC 7551P (32 cores) + 128GB DDR4, but I'm planning to make use of something better for the actual runs.

My opinion was the front drums looked good but the rears looked like crap. Brake discs and calipers looked good though. The rear drums winglets are responsible for a fair bit of downforce and helping the diffuser so them being so low quality can't be helping.

What I can't figure out is how Variante and DP_CFD are working around the cooling ducts - which as far as I can tell terminate inside the body. Sealing it off and using an outlet?

I'm partially considering remodeling those winglets on the rear cake tins, although they're performing better than expected in CFD so far.

For the cooling ducts, exactly that. I know someone with F1 experience giving me some numbers to use, although they might not translate perfectly to the 2020 Mercedes PU.

[variante]

What I can't figure out is how Variante and DP_CFD are working around the cooling ducts - which as far as I can tell terminate inside the body. Sealing it off and using an outlet?

At the moment, i'm simply not working on them. They're ducts that lead nowhere. And that's part of the reason why i haven't made any claim concerning the true performance of the car, so that no casual is tempted to use those numbers as a reference.

Why haven't i bothered with them? Because they play a very secondary role. The main flow structures of the car are still very much there.

But if someone is willing to spend time modelling some internal geometries and applying some boundary conditions, he can probably dare to make such claim (especially with aforementioned 65M cells, which my computer can't handle).
It still won't be the official super optimized billion cells simulation, but it won't be too far off either.

[Owen.C93]

Yeah I would at most just let them bleed off through a small outlet so air isn't spilling out of the sidepod inlets. Maybe an 75% restriction to simulate the radiators.

[ScrewCaptain27]

Yeah I would at most just let them bleed off through a small outlet so air isn't spilling out of the sidepod inlets. Maybe an 75% restriction to simulate the radiators.

Good idea on the restriction, best thing would be to let them bleed into the hot air exits at the back.

[Racingf1]

It's still possible to get the raw geometry of the leaked car from the app?

If you, you could help me to tell me where do I need to look for it? I'm really looking forward try run some CFD stuff with Star and learn from what a current F1 car does and compare it with a very simple design that I did a year ago during the first lockdown period in Spain.

Thanks in advance to all of you!

[Owen.C93]

It's still possible to get the raw geometry of the leaked car from the app?

If you, you could help me to tell me where do I need to look for it? I'm really looking forward try run some CFD stuff with Star and learn from what a current F1 car does and compare it with a very simple design that I did a year ago during the first lockdown period in Spain.

Thanks in advance to all of you!

It's on page 1. https://t.co/akFaAkKBux?amp=1

[ScienceTechnology]

Guys a friendly word of warning as I have now seen several people posting CFD analysis of the Williams CAD files on their own LinkedIn profiles...

For those of you with the goal of an F1 career, posting these simulations under your own name could seriously jeopardise your chances of ever working for an F1 team. The teams take IP breaches extremely seriously and will not risk the potential consequences of employing someone who has published analysis of these CAD files on a public forum!

If you have no interest working in F1 then by all means post away, it really is great insight for everyone else!

[DP_CFD]

Guys a friendly word of warning as I have now seen several people posting CFD analysis of the Williams CAD files on their own LinkedIn profiles...

For those of you with the goal of an F1 career, posting these simulations under your own name could seriously jeopardise your chances of ever working for an F1 team. The teams take IP breaches extremely seriously and will not risk the potential consequences of employing someone who has published analysis of these CAD files on a public forum!

If you have no interest working in F1 then by all means post away, it really is great insight for everyone else!

Noticed you made an account just to say this, are you speaking as someone on the inside?

For me, I'm viewing this as the chance to show to the teams that I know what I'm doing. Being outside of the typical candidate pool means I'm a bit easier to dismiss, so I want to be hard to ignore.

As for 'damage' to Williams, my take is that it's ultimately inconsequential other than a bit of embarrassment, given that:

• The models are incomplete

• The models have sections of poor quality

• This is the last year of the aero regulations

• The car is quite similar to previous year's, which teams will have had all season to copy

• Any community CFD will be highly questionable, we don't have a team of CFD Aerodynamicists and CFD Methodologists to help us out

It's stuff like this I plan to make clear, that my results are for the most part not to be trusted. Hell, I might even ask Williams for more reasons it's wrong!

On principle, I understand what you're saying about teams exercising caution. Had this been a perfect model, I would likely keep the CFD to myself, and perhaps send it to Williams with a resume. I'm only sharing results because I don't think I'm doing any technical damage.

[jjn9128]

Yeah, not sure on the legal ramifications there. It wasn't stolen or hacked, it was simple to find in a badly designed app, the onus is on Williams to make sure the CAD is secure, IMO if they basically gave it away through stupidity it's on them.

That said I never downloaded the app or looked at the T&Cs.

[DP_CFD]

Yeah, not sure on the legal ramifications there. It wasn't stolen or hacked, it was simple to find in a badly designed app, the onus is on Williams to make sure the CAD is secure, IMO if they basically gave it away through stupidity it's on them.

That said I never downloaded the app or looked at the T&Cs.

Regardless of legality, I think the concern that ScienceTechnology brought up is that people are 'participating' in the leak. Whether or this will cause technical damage isn't clear cut, so while I personally think I'm in the clear, there's a chance that a team might want to avoid anybody involved in such grey areas.

[PlatinumZealot]

Guys a friendly word of warning as I have now seen several people posting CFD analysis of the Williams CAD files on their own LinkedIn profiles...

For those of you with the goal of an F1 career, posting these simulations under your own name could seriously jeopardise your chances of ever working for an F1 team. The teams take IP breaches extremely seriously and will not risk the potential consequences of employing someone who has published analysis of these CAD files on a public forum!

If you have no interest working in F1 then by all means post away, it really is great insight for everyone else!

So how will they reliably prove that the CFD was from their file? Ha!

CFD away my fellow geeks!

[Hoffman900]

While fun to observe, I struggle with how much I can learn from community CFD projects. There are major shortcomings in computing power, and given that the teams themselves struggle with correlation issues between the wind tunnel, track, and CFD, with all the data at their disposal, I’d question how accurate someone at home can get.

It is useful from a 1000’ / macro view of what the car is doing, but I wouldn’t read too much into the details of the data.

Applying to a job, I wouldn’t show any F1 team my home analysis. Only show what you have to, any more, and you’re opening yourself up to more critique. Once inside, you might see you would have been pretty far off in some aspects of your modeling.

YMMV.