Jump to content


Photo
- - - - -

Water instead of wind tunnels?


  • Please log in to reply
20 replies to this topic

#1 Singularity

Singularity
  • Member

  • 848 posts
  • Joined: March 21

Posted 17 September 2021 - 18:12

Do anyone know if any F1 team has experimented with water tunnels? Seems to me that something like this could give some interesting observations:




Advertisement

#2 AustinF1

AustinF1
  • Member

  • 20,502 posts
  • Joined: November 10

Posted 17 September 2021 - 18:18

That's very cool.



#3 Rodaknee

Rodaknee
  • Member

  • 2,178 posts
  • Joined: June 19

Posted 17 September 2021 - 18:52

Spot the difference

 



#4 Beamer

Beamer
  • Member

  • 3,384 posts
  • Joined: September 00

Posted 17 September 2021 - 19:00

In general comparable (as in visualizing flow and turbulance) but in detail very very different. Water does not compress. Air does. Water has higher density and changes direction slower, et cetera. So for f1 of no use.

#5 Singularity

Singularity
  • Member

  • 848 posts
  • Joined: March 21

Posted 17 September 2021 - 19:09

In general comparable (as in visualizing flow and turbulance) but in detail very very different. Water does not compress. Air does. Water has higher density and changes direction slower, et cetera. So for f1 of no use.

Sparkling water maybe, that'll compress? Champagne? ;)



#6 PayasYouRace

PayasYouRace
  • Racing Sims Forum Host

  • 45,984 posts
  • Joined: January 10

Posted 17 September 2021 - 19:13

Water tunnels are good for visualising airflow, which is great for a teaching tool, but they are not very good at actually replicating the forces involved which is the important thing. That NASA tunnel is clearly a teaching tool.

 

Aerodynamic testing depends entirely on what's called the Reynolds Number. This number is a ratio of the inertial and viscous forces experienced by a fluid, and water is very much more viscous than air. So water is not very useful, other than producing pretty patterns with dyes. Unless you test at the correct Reynolds Number, you won't get the correct turbulent structures forming.

 

If you look at the video Rodaknee posted you can see a much more useful tool for F1 aerodynamicists, and that's the pressure sensor which is sweeping through the car's wake in the video.

 

Turbulence is also something that CFD is getting better and better at simulating.

 

 

In general comparable (as in visualizing flow and turbulance) but in detail very very different. Water does not compress. Air does. Water has higher density and changes direction slower, et cetera. So for f1 of no use.

 

At the speeds F1 cars moved at, the compressibility of air is negligible. Compressibility only becomes important above Mach 0.3, and an F1 car will reach a maximum of about Mach 0.25. At the speeds an F1 car runs at, air is considered incompressible.



#7 PlatenGlass

PlatenGlass
  • Member

  • 4,653 posts
  • Joined: June 14

Posted 17 September 2021 - 19:37

I dunno. Maybe Spa could have gone ahead if they'd done this.

#8 ARTGP

ARTGP
  • Member

  • 29,421 posts
  • Joined: March 19

Posted 17 September 2021 - 19:58

Not to mention, submerging a windtunnel model has its own host of difficulties. Making an F1 wind tunnel model water tight, while also maintaining the ability to quickly change components is damn near impossible. There are onboard electronics that you don't want to get wet. There is also the issue of long term corrosion of anything that is metal.  


Edited by ARTGP, 17 September 2021 - 20:00.


#9 Rodaknee

Rodaknee
  • Member

  • 2,178 posts
  • Joined: June 19

Posted 17 September 2021 - 20:01


If you look at the video Rodaknee posted you can see a much more useful tool for F1 aerodynamicists, and that's the pressure sensor which is sweeping through the car's wake in the video.

 

Other than that, cars are tested on a rolling road.  First used by John Barnard, I believe.



#10 Izzyeviel

Izzyeviel
  • Member

  • 3,172 posts
  • Joined: March 16

Posted 17 September 2021 - 20:05

Sparkling water maybe, that'll compress? Champagne?  ;)

 I believe Ferrari tried this in the late 2000's... apparently it didn't work out, someone kept drinking it... :lol:



#11 PayasYouRace

PayasYouRace
  • Racing Sims Forum Host

  • 45,984 posts
  • Joined: January 10

Posted 17 September 2021 - 20:19

Other than that, cars are tested on a rolling road.  First used by John Barnard, I believe.

 

The rolling road is a key part of a racing car wind tunnel, because it provides for the correct relationship between the air and the track. This is key for modelling ground effect. For those that may not realise it, ground effect is always a factor in any racing car design, not just when venturi tunnels are used. Air travelling over a surface (or vice versa) creates a boundary layer. Modelling the correct relationship between the car and the track allows for that boundary layer to be simulated correctly.

 

Though it wouldn't be that difficult to have a rolling road in a water tank. It just wouldn't be that useful overall.

 

Not to mention, submerging a windtunnel model has its own host of difficulties. Making an F1 wind tunnel model water tight, while also maintaining the ability to quickly change components is damn near impossible. There are onboard electronics that you don't want to get wet. There is also the issue of long term corrosion of anything that is metal.  

 

You wouldn't need to make the model water tight. You'd just need to make things water proof. Most wind tunnel models wouldn't have a long enough life to worry about corrosion either. Still, it's all academic as you wouldn't use a water tunnel in the first place.



#12 r4mses

r4mses
  • Member

  • 2,348 posts
  • Joined: April 09

Posted 17 September 2021 - 22:18

Newey maybe - when building his boat.



#13 Beamer

Beamer
  • Member

  • 3,384 posts
  • Joined: September 00

Posted 18 September 2021 - 17:18

Water tunnels are good for visualising airflow, which is great for a teaching tool, but they are not very good at actually replicating the forces involved which is the important thing. That NASA tunnel is clearly a teaching tool.

Aerodynamic testing depends entirely on what's called the Reynolds Number. This number is a ratio of the inertial and viscous forces experienced by a fluid, and water is very much more viscous than air. So water is not very useful, other than producing pretty patterns with dyes. Unless you test at the correct Reynolds Number, you won't get the correct turbulent structures forming.

If you look at the video Rodaknee posted you can see a much more useful tool for F1 aerodynamicists, and that's the pressure sensor which is sweeping through the car's wake in the video.

Turbulence is also something that CFD is getting better and better at simulating.



At the speeds F1 cars moved at, the compressibility of air is negligible. Compressibility only becomes important above Mach 0.3, and an F1 car will reach a maximum of about Mach 0.25. At the speeds an F1 car runs at, air is considered incompressible.


Thx, learned something!

#14 Joblo

Joblo
  • New Member

  • 2 posts
  • Joined: August 05

Posted 28 October 2023 - 07:21

-Minardi was approched by sweden SSPA (Statens Skepps Provnings Anstalt) Maritime Center in October 1986. It took almost 1 year to study the methodologies for this type of test. So from September to October 1987, a 1:1 M186 was towed at the bottom of a 80m x 30m water basin at a speed of 3m/second. (1988 February 02 Rombo #5).

In La Stampa (1989 July 18) Giancarlo Minardi mention: "We had found an extraordinary technical solution, testing a 1:1 scale car in a special pool, for aerodynamic flows. But it cost 5-6 million lira per hour and we had to abandon it."

 

-1988 Zakspeed 881 was tested in a water tunnel using a 1:4 scale model as said by designer David Kelly in 1988 Autosprint #11 (March 15): "We are assured of this by the data from our penetration values are markedly improved. We are assured of this by the data from our tests done in the wind tunnel in Aachen, Germany, where we also water tunnel tests with a 1:4 scale model." (original: Ce lo assicurano i dati dei nostri valori d penetrazione sono nettamente migliorati. Ce lo assicurano i dati dei nostri test fatti nella galleria del vento di Aachen, in Germania, dove abbiamo anche postato avanti prove nella galleria ad acqua con un modello in scala 1:4.)

 

-In the 2000's, the Honda F1 team tested a full-sized rear wing in a water towing tank ( https://www.highpowe...g-wing-airflows )


Edited by Joblo, 29 October 2023 - 18:02.


#15 Grippy

Grippy
  • Member

  • 421 posts
  • Joined: June 18

Posted 28 October 2023 - 10:58

Slightly related;

Have any teams considered 'AeroSHARK' film? It has recently been added to Lufthansa planes as a fuel saving measure, Based on the dendrites in shark skin it cuts drag on a Boeing 777 by 1% with equivalent fuel savings.

 

Main problem I visualise is dirt, second, the cars might not be fast enough to take advantage of it.

 

https://cleantechhub.../aeroshark.html



#16 GlenWatkins

GlenWatkins
  • Member

  • 2,084 posts
  • Joined: March 20

Posted 28 October 2023 - 11:25

Water tunnels are good for visualising airflow, which is great for a teaching tool, but they are not very good at actually replicating the forces involved which is the important thing. That NASA tunnel is clearly a teaching tool.

 

Aerodynamic testing depends entirely on what's called the Reynolds Number. This number is a ratio of the inertial and viscous forces experienced by a fluid, and water is very much more viscous than air. So water is not very useful, other than producing pretty patterns with dyes. Unless you test at the correct Reynolds Number, you won't get the correct turbulent structures forming.

 

If you look at the video Rodaknee posted you can see a much more useful tool for F1 aerodynamicists, and that's the pressure sensor which is sweeping through the car's wake in the video.

 

Turbulence is also something that CFD is getting better and better at simulating.

 

 

 

At the speeds F1 cars moved at, the compressibility of air is negligible. Compressibility only becomes important above Mach 0.3, and an F1 car will reach a maximum of about Mach 0.25. At the speeds an F1 car runs at, air is considered incompressible.

 

Interesting research paper on this subject that I ran across on F1Technical:  https://www.research...heel_racing_car

 

 

5.0 CONCLUSION

A numerical investigation into the influence of compressibility effects around a simplified open-wheel racing car was completed. Results demonstrated that for high-lift aerodynamic designs operating in close ground proximity the effects of compressibility are significant at speeds well below the Mach 0·3 threshold normally applied. Incompressible simulations are generally unsuitable even at Mach 0·15 and below, with pockets of local velocities exceeding three times the freestream value and creating exaggerated low-density, low-pressure flow. Notably, changes to the interactions between components were the key contributors to some of the most significant observed points of difference. Components operating in the most extreme ground effect were observed to be most affected. For the front wing, incompressible simulations underestimated the extent of flow acceleration and therefore maximum suction. The negative lift produced by the floor and diffuser was found to be the most influenced component; the consideration of compressible flow revealed a more detrimental interaction with the rear wheel and resulted in markedly less predicted negative lift when compared to the incompressible prediction even at Mach 0·0882 (30ms–1), but most notably at Mach 0·2646 where discrepancies of over 20% were observed. The effects towards the behaviour of the prominent vortices off the front and rear wings were small relative to the overall pressure distributions acting on components. Smaller vortices surrounding the rear wheel and diffuser, and general wake velocity deficits were found to be more affected due to compressibility. The car simulated here has a significantly less aerodynamic efficiency compared to modern vehicles, so the estimates presented here are expected to be conservative – actual discrepancies may therefore be greater than those evaluated in this study. The incompressible assumption is clearly not suitable for racing car aerodynamics if absolute accuracy is a key goal, despite the additional computational expense required to conduct compressible simulations. Compressible simulations become particularly important at Mach numbers of 0·15 and above, and may be one of several factors in poor correlation between wind-tunnel results and incompressible CFD simulations. Observed trends are non-linear and complex due to the close interaction between components. It is clear that a simple compressible correction would not be sufficient. Broadly speaking, Keogh, Doig & Diasinos Flow compressibility effects around an open-wheel racing car the results reinforce that any CFD being compared to physical experiments, should reproduce those experiments as closely as is possible and practical in terms of both the geometry and the flow physics.


Edited by GlenWatkins, 28 October 2023 - 11:28.


#17 1player

1player
  • Member

  • 1,235 posts
  • Joined: March 21

Posted 28 October 2023 - 11:36

I don't know anything about fluid dynamics, but I just got nerd sniped by this topic, thanks! :rotfl:

 

I reckon that converting simulations made with an incompressible fluid (water) to behaviour with a compressible fluid (air) must be such a pain that it is best to simulate with air directly. If they could easily translate (i.e. calculate) the behaviour in water to air, they wouldn't even need to do any real-life simulation in the first place.

 

That's my reductio ad absurdum of the problem. I have a mate that works with this kind of stuff every day (for industrial heat exchangers, not motorsports) and I wonder what is his opinion on this.



#18 wheadon1985

wheadon1985
  • Member

  • 124 posts
  • Joined: January 16

Posted 27 February 2024 - 20:35

Having listened to a FFS podcast with Eddie Jordan and David Coulthard, Eddie said Jordan couldn't afford wind tunnel time so had Gary Anderson design the car in a water tunnel as it was so much cheaper in cost. 

 

Maybe this is why it has such beautiful curves?! 



#19 Jazza

Jazza
  • Member

  • 1,805 posts
  • Joined: November 99

Posted 28 February 2024 - 05:35

Pretty sure Benetton in 98 and 99 used one. I remember seeing some pictures. I think they were even put in upside down with the car stuck to the top of the tunnel.

I doubt that it was solely a water tunnel that they used though, as I think it was more of a research link up with Oxford. (Although, with their performances in 99 and 2000 the car very well might have been made in a water tunnel…)

Edited by Jazza, 28 February 2024 - 05:41.


Advertisement

#20 kumo7

kumo7
  • Member

  • 7,169 posts
  • Joined: May 15

Posted 28 February 2024 - 12:49

Got some pressure issues indeed. ;-)
But the slow movement of the flow give me much needed time to rethink what it means, and that might be an interesting property. It is like a slow motion.

#21 KWSN - DSM

KWSN - DSM
  • Member

  • 36,090 posts
  • Joined: January 03

Posted 28 February 2024 - 13:56

https://arstechnica....-test-facility/