Forbidding windtunnels in F1

[Just_a_fan]

CFD will be as good as a wind tunnel when it can model the movement of molecules around the car. Until then it is nothing more than a rough approximation. Even a sub-full size wind tunnel is an approximation but it is a lot closer than CFD is. CFD helps you to decide what to wind tunnel test.

One day we will be able to do DNS CFD at will and at reasonable cost. That day is a long way off, however, at least in terms of the lives of most F1 teams. Yes, I know that computers get more powerful at an astonishing rate but here's something discussing how CFD trails that increase quite markedly at this level.

Therefore, doubling the Reynolds number
from a currently attainable value would increase the computational cost (i.e. CPU time,
memory) by roughly a factor of 11! Assuming that the current trend of doubling computing
power every 18 months continues and numerical algorithms scale perfectly for even larger
grid counts, this implies that for this flow the Reynolds number can be doubled only every
five to six years. Although the exact Reynolds-number scaling of the cost will vary from flow
to flow, and will not always be as stringent as for homogeneous isotropic turbulence,‡ the
general point still holds that we are currently far away, and will remain so for the foreseeable
future, from being able to perform DNS for Reynolds numbers typical of full-scale engineering
and especially geophysical applications.

from: http://eprints.soton.ac.uk/66182/1/A_primer_on_DNS.pdf

As for cost, the current top supercomputer is in China (they're all basically state-owned at this level), cost $390million and consumes 24MW including cooling requirements (info from Wiki). We're a long way off before this level of performance (which is really needed to replace a wind tunnel fully) is available and usable by a racing car company.

[dren]

Teams will spend all of the money they have wrapped up in wind tunnels towards their CFD and more if needed. They will be in a constant cycle of upgrading their CFD computers and programs. Sure, CFD may not be expensive to operate, but it will be hugely expensive to continuosly update to obtain an advantage.

But as Wirth has proven, CFD only with the current technology doesn't quite work.

[wesley123]

Welcome to the era where everything that isn't regulated by a computer is "dinosaur technology".

Does the product get over-complicated by having lots of software and computer calculations in it?
- > No - > Add computers, preferably lots, lots of it.
- > Yes - > Perfect, but don't forget to upgrade yearly!


CFD is a great tool, as it doesn't require you to build the parts anymore, which would save time.

On the other hand, CFD is a computer program written by human beings, what it has coded is things known to mankind. Thus, it can't possibly calculate everything correctly, it can only calculate what is known.

A windtunnel on the other hand is real life, and although computers have come a long way, it can't possibly replace real life.

[Facts Only]

I'm actually coming round to this idea. There are a lot of people saying that CFD doesn't work as well as a wind tunnel and currently it doesn't but the great thing about F1 is that the competition and speed of development forces teams to make things work.
Secondly wind tunnels are very resource heavy especially in terms of electricity and resin which are not going down in price, whereas computing power is always getting cheaper so yes they will spend the money they save on supercomputers but the prices will always be reducing.
Also the Cap Ex requirements of CFD are a lot lower than a wind tunnel aiding new teams.
Lastly the teams that already have wind tunnels could probably make some income from them.

Anyway I doubt it will happen soon but maybe during a major shake up after 2020

[Facts Only]

On the other hand, CFD is a computer program written by human beings, what it has coded is things known to mankind. Thus, it can't possibly calculate everything correctly, it can only calculate what is known.

A windtunnel on the other hand is real life, and although computers have come a long way, it can't possibly replace real life.

A wind tunnel isn't real life either, its a simulation of real life just like CFD. Remember how much pain Ferrari have just gone through to try to correlate their wind tunnel data to on track results.

At the moment wind tunnels are only closer to real life than CFD because so much effort has been poured into correct correlation of wind tunnels. If all that effort was poured into CFD IT would quickly catch up.

[Jersey Tom]

-Environmental it would sense make since wind tunnels consume huge amounts of electricity.
-He called the wind tunnel "dinosaur technnology", claiming that if F1 wants to continue to push the boundaries, it should do so in the field of CFD.

Ludicrous on both counts, but especially the second. I don't know whether the right word is patronizing or insulting or what.

Taking test measurements is dinosaur technology? I thought that was kind of a key, fundamental part of engineering. Y'know, that whole scientific method thing based around improving hypotheses and mental models based on physical experiments. You can't push CFD or any CAE without measurement, validation, verification.

It's a farce. That's probably the best word I'm looking for. F1 does not push all the boundaries. It is bounded. It becomes increasingly bounded every year - banning wind tunnels is certainly another bound.

To be fair - that's all well and good. F1 is a sport. Can't argue that. Sports have rules. Sports are for fun competition and entertainment. So why does F1 struggle with its identity so much? Why pretend it's something it isn't? All the "green" malarkey and gimmicky tech while forbidding real cutting edge stuff.. it's embarrassing.

[Cold Fussion]

Banning wind tunnels on the grounds of environmentalism is the most ridiculous argument I've heard outside of F1 teams forum section. If we're so keen on environmentalism, why are persisting with freighting 40 + cars and thousands of people around the world 20 times year for the inane activity of driving from point a to point a the fastest? On the topic of environmentalism, does CFD really save any power at all to get to the accuracy of a wind tunnel? The fan in the Sauber tunnel uses 3000 kW, everything all together we can probably say it uses 5000 - 10000 kW while running. The Titan 2 super computer uses 8000 kW, so is a 27 Petaflop super computer enough to completely replace a wind tunnel? Once again Jersey Tom brings some much needed common sense to these forums.

[Ciro Pabón]

Although the wind tunnel vs computer is an interesting dilemma, let's not forget that wind tunnels are not "real life".

Real life is when the car is racing.

Even a wind tunnel is a simulation.

In real life there are things like wind, ground, engine, other cars and whatnot (sunspots, passing invisible unicorns, pit ghosts of past racers, involuntary and voluntary telekinesis of spectators, underground UFOs waiting for invasion command interred under the main straight foundations and their huge magnetic fields: you know, the usual engineering real life influences).

About F1 being a sport and Jersey Tom bringing common sense to the forum, well... what about the little cats they use in wind tunnels?

Doesn't Tom feel their pain when a wind tunnel is started? What is wrong with us, people?

[turbof1]

-Environmental it would sense make since wind tunnels consume huge amounts of electricity.
-He called the wind tunnel "dinosaur technnology", claiming that if F1 wants to continue to push the boundaries, it should do so in the field of CFD.

Ludicrous on both counts, but especially the second. I don't know whether the right word is patronizing or insulting or what.

Taking test measurements is dinosaur technology? I thought that was kind of a key, fundamental part of engineering. Y'know, that whole scientific method thing based around improving hypotheses and mental models based on physical experiments. You can't push CFD or any CAE without measurement, validation, verification.

It's a farce. That's probably the best word I'm looking for. F1 does not push all the boundaries. It is bounded. It becomes increasingly bounded every year - banning wind tunnels is certainly another bound.

To be fair - that's all well and good. F1 is a sport. Can't argue that. Sports have rules. Sports are for fun competition and entertainment. So why does F1 struggle with its identity so much? Why pretend it's something it isn't? All the "green" malarkey and gimmicky tech while forbidding real cutting edge stuff.. it's embarrassing.

Except that validation and verification happens on the track, not (ultimately) in the windtunnel. Don't ask me, ask Ferrari.

Banning windtunnels is of course another bound, but if you are going to take that way there a lot of other things you should worry about first, beginning with the hugely restrictive technical rule book.

Lastly, removing the windtunnel would certainly speed up CFD development. Yes, coming with a huge price too, but mostly in R&D, with base costs remaining quite a chunk lower then setting up a windtunnel.

Speaking of bounds, the one thing you curiously did not wanted to adress was the cost. Having a windtunnel or hiring a windtunnel means a huge base cost to begin with, demotivating new teams to enter. That's a big boundary on its own. Sure, CFD is quite a significant cost too, but nowhere near what a windtunnel would ever cost.

Besides that, it's like Facts Only said: today cfd doesn't work as well as a windtunnel, but we all know it has the potentional to outdevelop windtunnel within a decade or so. Yes indeed, windtunnels will eventually become redundant tools anyway.

Also, I'll be the first to underline the hypocrity in Fernley's comment that it'd be environment-friendly.
Again however, we aren't looking at making a product green, we are looking at branding a product green, just like producers of sun panels do (even though the polution by production of sun panels takes 20 years use of said panels to compensate) or hybrid cars (China has somewhere a huge toxic wasteland due production of the batteries). It's marketing, selling your product. Nobody even cares to check if the product truly is green.

[Just_a_fan]

Besides that, it's like Facts Only said: today cfd doesn't work as well as a windtunnel, but we all know it has the potentional to outdevelop windtunnel within a decade or so. Yes indeed, windtunnels will eventually become redundant tools anyway.

A decade? No way. And over that decade, how much money will be spent, how many computers built and replaced? The cost will be in the billions. Per team.

F1's idea of cutting costs is to ban something that is mature and replace it with something that will cost a fortune to develop over a few years just to get to the point they are today. The whole idea is just political posturing by teams that haven't done a good enough job. I'm sad that people on a technical forum like this have fallen for it.

[turbof1]

Besides that, it's like Facts Only said: today cfd doesn't work as well as a windtunnel, but we all know it has the potentional to outdevelop windtunnel within a decade or so. Yes indeed, windtunnels will eventually become redundant tools anyway.

A decade? No way. And over that decade, how much money will be spent, how many computers built and replaced? The cost will be in the billions. Per team.

F1's idea of cutting costs is to ban something that is mature and replace it with something that will cost a fortune to develop over a few years just to get to the point they are today. The whole idea is just political posturing by teams that haven't done a good enough job. I'm sad that people on a technical forum like this have fallen for it.

That's quite harsh; we all know something has to change anyhow. I'm by no means fans of knee-jerk reactions, but there's always the possibility to gradually limit the windtunnel spread over several years (like a decade).

I also think you are quite overreacting . a few hundred millions over a decade, I would have agreed with that, but billions? That's blowing things quite up.

Teams will always have their own motivations to do so. Just because they abuse sound arguments, does not mean those arguments are not sound.

[Andres125sx]

IMO, after watching how FIA ban techonolgies wich any team could use (active wings or suspensions for example) wich would make F1 cars way way faster, banning something wich is way more expensive and no every team can use (windtunnels) doesn´t look weird at all

This would equal teams perfomance and would also act as a cost cap.

Also, as some of you have already pointed out, they´re not real life experience, they use 1:2 scale models, they can´t simulate lateral wind, they can´t simulate kerbs, they can´t simulate front car slipstreams, they can´t simulate wings flex at bumps/kerbs.... they can´t simulate so many things they shouldn´t be considered foolproof.

CFD eventually will be able to simulate anything you want, with a lot of development to do yet, but it´s a tool even the smallest team can use. At a different scale obviously, but even so there would be more parity than currently with some teams using huge windtunnels while others only can simulate on basic CFD

So cost control, more parity between teams, way easier for new teams to join F1, CFD would be road relevant as this develpment will be used on productions cars.... I can only see advantages, and not one single disadvantage. If you´re worried about perfomance just remove some restriction, for example active wings, and that will make F1 cars way faster at a fraction of the cost of windtunnels

[Jersey Tom]

Except that validation and verification happens on the track, not (ultimately) in the windtunnel. Don't ask me, ask Ferrari.

[...]

Lastly, removing the windtunnel would certainly speed up CFD development.

[...]

Yes indeed, windtunnels will eventually become redundant tools anyway.

Not entirely true. Is the end goal to make the car go faster at the track? Sure. But the track has a long list of negatives associated with it. Much more uncontrolled environment, harder to measure small changes, expensive, infrequent, etc. If your CFD model predicts some small winglet or widget is going to give a small incremental increase - say a few pounds of downforce - you might be able to resolve that at the tunnel but it would be silly to do so at track.

To me, as an engineer and performance analyst, controlled physical test benches are never redundant or obsolete - though roles can certainly change. It is the trap of the naive or newbie analyst to think that they're going to make "the ultimate" computer model that captures everything. It's an exercise in futility. It's like trying to replace every tool in a mechanic's or machinist's toolbox with one mega Swiss army knife. The hand file has been around a long time, and it hasn't been made obsolete by CNC machinery and 3d printers.

As for advancing CFD itself, my thought is that it's silly to think that F1 is seriously going to advance the world of CAE on its own. CFD, vehicle dynamics, whatever. F1 is too small in scope and a drop in the bucket when it comes to people and budget. When you spend time in both the motorsport industry and the consumer industry, it is shocking just how small (and often behind the curve) the racing business is compared to the rest of the world.

Lastly, the cost bit - I just don't see how it will fundamentally change things. If a team gets sponsor money, it will be spent.

[turbof1]

Except that validation and verification happens on the track, not (ultimately) in the windtunnel. Don't ask me, ask Ferrari.

[...]

Lastly, removing the windtunnel would certainly speed up CFD development.

[...]

Yes indeed, windtunnels will eventually become redundant tools anyway.

Not entirely true. Is the end goal to make the car go faster at the track? Sure. But the track has a long list of negatives associated with it. Much more uncontrolled environment, harder to measure small changes, expensive, infrequent, etc. If your CFD model predicts some small winglet or widget is going to give a small incremental increase - say a few pounds of downforce - you might be able to resolve that at the tunnel but it would be silly to do so at track.

To me, as an engineer and performance analyst, controlled physical test benches are never redundant or obsolete - though roles can certainly change. It is the trap of the naive or newbie analyst to think that they're going to make "the ultimate" computer model that captures everything. It's an exercise in futility. It's like trying to replace every tool in a mechanic's or machinist's toolbox with one mega Swiss army knife. The hand file has been around a long time, and it hasn't been made obsolete by CNC machinery and 3d printers.

As for advancing CFD itself, my thought is that it's silly to think that F1 is seriously going to advance the world of CAE on its own. CFD, vehicle dynamics, whatever. F1 is too small in scope and a drop in the bucket when it comes to people and budget. When you spend time in both the motorsport industry and the consumer industry, it is shocking just how small (and often behind the curve) the racing business is compared to the rest of the world.

Windtunnel is just a simulation like cfd is. The only difference is that one is mechanical, the other digital. Yes of course, windtunnel today is an important tool to verify your findings in cfd (cfd has issues still). However, if you ban windtunnels, compettion will make sure cfd will grow towards a better simulating tool. It's the nature of the sport. What you are leaving out, purposely or not, is the incentive. Teams nowadays don't have the incentive to accelerate CFD/CAE development due windtunnel giving better results (which I fully underline that that is the case in the present). Take away the windtunnel and there will be a competitive incentive, and it will stop being just a drop in a bucket.

Lastly, the cost bit - I just don't see how it will fundamentally change things. If a team gets sponsor money, it will be spent.

The issue is that teams have, emphasis on have, to spend more then what their revenue is. If it was simple the case that the teams just spent what they have and after that wait for the next round of revenue, this discussion might not even take place. As an engineer and performance analyst, you should know that budgets more often then not are not nearly sufficient, no matter where you work. It'll always end up with overspending. Not a big issue on the short run, but on the long run this philosophy always ends up in tears.

Fine, you can always tell those teams to bugger off, only we have just 10 teams left. One way or another, either there has to be more income, or there have to be less expenditures in order to atleast safeguard 10 teams on the grid. I hope we can atleast agree on that, aside the discussion of banning windtunnels on the long run.

Let's also make one thing clear: banning windtunnels will NOT make cars slower. It will probably decrease the rate of development, but unless new technical restrictions are introduced it will not harm performance.

[Moxie]

Banning any technology for the purpose of reducing costs does not work. Teams will simply spend the money on the next most promising technology to gain a millisecond of advantage. Even if you ban the whole damned car the teams will strap $200 M rollerskates to the driver's feet. The only way to control costs is through economic disincentive to spend. Reduce the financial reward, and the financial risk will drop accordingly.