32 replies to this topic

[Ali_G]

Can anyone give me a drawing of a venturi tunnel and explain how it works. I know how ground effect works I just don't know how a venturi tunnel works. I have heard sugestions that the bottom of the car would be shaped like the bottom of an aerofoil. Is this true.

Niall

[desmo]

Here's the Lotus 79B showing how the undersides of the sidepods are indeed contoured like an inverted(compared to an airplane) wing. Same principle applies. The skirts kept airflow from spilling into the undercar area and diminishing the downforce.

[Ali_G]

Nice one desmo.

So in a way the car was one huge wing in a way.

Didn't they ban skirts in 1980 (note really sure).

Niall

[Rainer Nyberg]

A larger monochrome image of the same pic.



Btw, indycars are still using venturis to a lesser degree.

Rainer

[Ali_G]

Rainer: But they can't have skirts down the sides at all.

Niall

[MacFan]

Skirts aren't allowed in CART, but venturi tunnels still work without them. The skirts help prevent air at atmospheric pressure rushing in to the low pressure area under the car, but the air under the car will still be at lower than atmospheric pressure.

[Paolo]

Ehm, sorry guys,
but the Venturi tunnels principle is NOT to make the sidepods work as inverted wings.
It all has to do with the Venturi effect (here is where the name comes from...).
More or less, the thing works as follows :
1) the air enters the tunnel formed by the undersurface of the pods and the ground itself. The entrance has a relatively large area.
2) the air speeds through the tunnel, wich, due to the shape of the pds, has a decreasing section. Since the mass-flow (i.e. the quantity of air that must pass in a given time trough every section of the tunnel) must be constant, whrn the section becomes smaller the speed of the flow increase.
3) the pressure of the air under the tunnel decreases as the speed increases.
This has to do with bernoulli's law : the quantity
(pressure) + (1/2) * (air density) * (speed)^2
must remain constant from section to section.
So, if you increase speed, you decrease pressure.
4) the tunnel section gradually increases, and the pressure grows ; in the end, yhe air exits on the back of the car, at a pressure still lower thn atmospheric, since it's in the body wake.
One might wonder why not keep the tunnel small until the tail of the car, to have everywhere the minimum pressure.
Aerodynamics gives an answer (yes, I'm a tech...) but I don't want to embark in it there.
Suffices to say that you MUST gradually take the airflow to the exit pressure, or you don't get the downforce.
If you ar not gradual, by the way, the flow will separate and , again you don't get the downforce...

Since the pressure under the car is lower than atmospheric, the car is "sucked" on the ground.
By the way, this same principle applies also to modern "flat bottom and diffuser" ground effect cars ; the main difference with a Lotus 79 is that they lack the first "wide area" part of the tunnel.
The air flow behaves anyway as if a tunnel was present, so the air "vein" in front of the car restrictes itself while advncing towards th underbody and enters the trailing edge of the flat bottom at high speed already.
The difference with a '80 wing car is that, since the airflow must do this unaided by a properly shaped tunnels, it loses a lot of energy in the process, and becomes more easily separated.
This means you cannot have the same downforce, because you cannot speed the airflow as much.

[MrAerodynamicist]

Originally posted by Paolo
One might wonder why not keep the tunnel small until the tail of the car, to have everywhere the minimum pressure.
Aerodynamics gives an answer (yes, I'm a tech...) but I don't want to embark in it there.

Ah, go on....
a) challenge us
b) I should know why [I'm a 3rd year aerospace student] but I don't [well it is past midnight :)]

[Paolo]

Okay, I'll work on that. Gimme some time.
By the way, I' don't think of knowledge sharing as "challenge"

[desmo]

I knew as I described the underside of the venturi tunnel as an inverted wing that there were as many differences as similarities and it was a simplification. Of course having the undercar contoured such as on the Lotus is illegal now, but the raised trailing section of the underbody of the Lotus performed the same function as the diffuser in a modern F1 car, deccelerating the airflow in a smooth and orderly fashion.

[AndersF1]

Hmm, one remark about your description Paolo; remember that the inviscid, Euler, part of the Navier-Stokes Eqs is elliptic for subsonic flow, so the area of influence is both upstream and downstream, i.e. what happens downstream effects the flow upstream and vice versa. Another thing to remember is that you can't have a discontinuous pressure-jump in subsonic flow, density and velocity yes, but not pressure.

This means that you can start your calculations from the rear, at the exit of the venturi tunnel as the pressure here is close to athmospheric and then use conservation of mass(constant massflow) and Bernoulli to calculate the pressure. The other way around is much more difficult as you don't know the area of the "flowtube", or the unknown "wide area" as Paolo describes it. You can't get the area exact at the rear either, as the pressure is not exactly athmospheric, due to the wake, but it is easier.

BTW, as the pressure is close to athmospheric the flowspeed at the exit will also be close to the speed of the car, but have a different direction(upward). So, the flowspeed parallel to the ground is less after the car, than infront of, hence giving the car behind a reduced drag, as the local airspeed is less than the speed of the car. The change in vertical momentum of the air per timeunit is the same as the downforce achieved by the venturi-tunnels and the change in horizontal momentum per timeunit is the same as the drag.

The reason that you can't use a small area all the way, until the end, is for the same reason that you have maximum thickness of a wing close to the leading edge; you can't have a too large positive pressuregradient. Otherwise you will have flowseparation and hence, sudden increase of pressure before the end of the tunnel. The decrease in flowmomentum(lower flowspeed) is due to the difference in pressure over a "flowparticle". This is well balanced in the outer flow, but not close to the wall, in the boundary layer, where the flowspeed is much less, but the pressure is the same. So, if you have a very large pressuregradient the flow will change direction close to the wall and then you have a separated flow. Not good.

/Anders

[355 boy]

WAY past my elementary physics...;)

[desmo]

Anders, thank you for your post. I'd never thought about the downforce and drag being equal to the changes of the airflow's momenentums per time unit. That makes perfect sense!

[MrAerodynamicist]

Correct me if I'm wrong, but don't both the images above show cars based on inverted aerofoil designs? Whereas a simpler venturi only has the aft curve?

[Paolo]

The aerodynamic measuring device called Venturi (not to be confused with Venturi Tunnel used in F1) has usually a converging-diverging shape (i.e. both fore and aft curves).
In some situations, however, it only has one of the two curves. The shape depends on what you need to do with the venturi : accelerate a flow from supersonic to subsonic, decelerate it, stay subsonic but at different speeds etc. The different ambient pressures at the ends of venturi dictate the behaviour. Any good book in Gasdynamics will show you the math.

[Paolo]

Before start, thanks to Mr Aerodynamicist for hosting the image on his page , and to the others who gave advice about how to post .



Hallo everybody,
here comes what I believe to be a good qualitative explanation of ground effect in its various forms
It is a simple explanation, so I left away some technical accuracy for the sake of clarity.
Before you tell me, fellow engineers : I know about the Nav-Stoks, okay ?
Anyway, any correction will be greatly appreciated; I see there is a lot of skilled people on this board.

We shall assume some simplifications : the most important one is : flow is here considered two dimensional.
For non-techs, it means that the air flows neatly under the car from leading edge (the front part) of side pods to trailing edge (the rear part), without spilling from the sides, forming vortices, going sideways and so on.
Also, we will neglect the static pressure change beetween the front and rear of the car ; that is, we will ignore the effects of wake.
We will also ignore the effect of boundary layers (slower air near the surface of the car)
All of these simplificatons are not acceptable if you look for accurate results, but they do not affect the general behaviour of the phenomenon.

Here you see a picture : look at it with awe and respect, it took me an hour to draw the scheme…
We shall point out that here only the flow under the side pontoons is examined ; that ‘s why they are coloured yellow.
Yes, the smiley is also coloured yellow, but we will not investigate the airflow around him…

In this discussion I will use the “flow pipe” concept.
Imagine that air flows inside a pipe : when the pipe becomes smaller in section, the air accelerates , and where it becomes bigger in section , the air decelerates. Engineers say this is due to the conservation of mass flow applied to incompressible fluids ; anyway you can see the same effect when you give water to your garden : if you put a thumb in front of the water hose exit (diminishing the section) the water accelerates and the jet goes farther.
Because of Bernoulli’s law, when air accelerates its pressure decreases ; when it decelerates its pressure increases.
That’s why I put that colour scheme on the right of the picture : every color corresponds to a certain speed and pressure of the air flow, High speed corresponds to low pressures.

One fundamental notion : the air that exits the car must be at the same pressure that reigns on the car’s back so, at the end of its travel under the pontoons, the air must be at ambient pressure (actually wake pressure, but that’s not so important now, and the difference is not as big as you would expect). Why ? Believe that, O.K. ? It has to do with the fact that air cannot spontaneously increase its pressure in the free stream on the rear of the car ; it would separate first because of the infinite pressure gradient. Just raise your arm , close your eyes and say “I believe”; it’s faster, and in University it works better…
Also look at Anders’ post in this thread.

The Drawing :

In case A) we have a car without any kind of Venturi or rear diffuser ; more or less what you had in 1970, and what you have now in Formula Renault : a full flat bottom.
In this case , the airflow comes in and exits at the same pressure (I said we shall ignore the slight change due to wake, remember? ). This means that there is no change in the section of a given flow pipe : the air stays all the time at the same speed and pressure, (ambient pressure) from the start to the end of the pontoon.
That pressure, under the car, is the same as the one above it : no downforce.
(This is not completely true, because of the effects we didn’t consider, and some other reasons ; I will however discuss the matter this deep only when some team hires me … Mr. Hamidy, Mr. Newey, don’t let yourselves be starved of such knowledge…)

In case B) we have a modern car, with flat bottom and rear diffuser.
Let’s start from the rear : the air must exit at ambient pressure.
Following the flow upstream, from the back of the car to the front, we see that the flow pipe section gradually decreases, and the pressure decreases too. That’s the diffuser.
Under the flat bottom , the section of the pipe is costantly at a minimum: that is the lowest pressure area. Speed is at its maximum, here ; air travels much faster than the car.
Continuing upstream, we see that air in our hypotetical pipe cannot stay at that extremely low pressure : it must gradually reach the ambient pressure, so the pipe gradually expands, and the flow slows down to the car’s speed.
Of course, in reality everything happened the opposite way than we looked at it : the air comes towards the car at the car ‘s speed, then increases its speed (while decreasing pressure), travels under the flat bottom at high speed, slows down (and increases its pressure) through the diffuser and finally exits from the rear at ambient pressure.

In case C) we have a good old wing car from 1980.
It works more or less as a flat bottom with diffuser, with two important differences :
1) the front part is smoothly shaped to accompany the gradual decrease in the flow pipe’s section ; this saves a lot of internal energy in the flow. Also , the diffuser contour is smoother, since its lenght is increased; this saves more energy and (see later) imposes a smaller pressure gradient.
2) This internal energy can be used to overcome higher positive pressure gradients than these manageable in a “flat bottom+diffuser car”.
So, what is a positive pressure gradient? I know you know, Mr. Aerodynamicist, but some other my not…. 
The airflow is said to meet a positive pressure gradient when the pressure gradually grows as it advances.
The steepest the increase in pressure, the greater the gradient.
Positive pressure gradients have a nasty habit of making the airflow separate, i.e. not follow the accurately engineered contours of your car but going its own way.
When this happens, the diffuser part will not work properly : it will be exactly like having a smaller diffuser, with smaller flow pipe section restriction under the car and less downforce generated. If the flow separates at the start of the diffuser, you might not get any dowforce at all… or worse (see case C).
In the end, a properly designed Venturi pod can give more downforce than a flat bottomed car with rear diffuser, because it is possible to obtain a smaller pressure under the car and still not have flow separation when the pressure increases to ambient in the back.

Case D ) One might think : why have a diffuser on a wing car ? Let’s just decrease the section at the front, and then keep the airflow at maximum speed under the car without ever decelerating it ; so we get minimum pressure under the whole lenght of the side pods (or pontoons, anyway you like to call them…).
Sorry, doesn’t work.
Air must exit at the rear of the car at ambient pressure ; so you would have ambient pressure all the way under the car… except at the front. There, you have HIGHER than ambient pressure . The flow coming towards the car will decelerate to meet the increased section at the front ; so it will estabilish a high pressure area under the trailing edge of the pods and then gradually speed up… until it reachs the car’s speed (and ambient pressure) under the car. You got not downforce but LIFT. The pilot won’t be grateful…

Phew, it took over an hour to write this… hope it gives me back at least Warhol’s 15 minutes…

[mhferrari]

It is designed to force air on the aerofoil to force the car unto the road. This is why it affects the mechanical grip of the car.

[Matt Davis]

Hi
I am not an aero engineer, I prefer electonics.
I understand paolo's explanation, but,
You explanation works in a wind tunnel but when a car is moving on a track does the air underneath it get speeded up even though it was not moving in the first place(apart from any wind)?
I don't fully understand car D and the points you are making about it could you elaborate?
If you want something else intresting aero related have a look at my post's and repplys in the super-car toppic

[MrAerodynamicist]

"You explanation works in a wind tunnel but when a car is moving on a track does the air underneath it get speeded up even though it was not moving in the first place(apart from any wind)?"
Yes. Velocities are always relative, wind tunnels and moving objects are basically the same. [There are some difference, but this is not particually relevent at this level. The need for a rolling road wind tunnel for accurate testing cars is probabily the best example]. If the idea of the air "geting speeded up even though it was not moving in the first place" seems strange, then think of it this way - if the air didn't accelerate at all, then how would it get out of the way of any moving objects? Run your hand though some still water, and watch as high pressure occurs in front of your hand and water rushing to the lower pressure behind it.

[andy_bee]

if you have acess to F1-Racing december issue (with the drivers on the front-in the UK anyway) they have an excellent article by William Toet the chief aerodynamist at BAR and explains how the windtunnel works and has some very very good pictures

[Paolo]

Matt
case D) has never been used (to my knowledge) on real world cars, because it is an example of how you can generate lift (opposed to downforce) with a wrong contouring of the underside of pods.
Scheme D) was my answer to a discussion beetween Mr. A , Anders and me .
The question was :what happens if we use only the first half of Venturi ?

Regards

[Martin]

Paolo I thought your post was excellent, thank you.
A question: Is case D an explanation of why road cars go light at speed (I remember the original MR2 I once drove became quite "interesting" (ie terrifying) above about 150/90 kmh/mph) and if so why don't you see diffusers on ordinary road cars?
Cheers
Martin

[desmo]

Paolo's post was indeed a very interesting read. Being American, I wonder what the legal ramifications of a "safety" device that doesn't come into play until one is driving at well past the legal maximum speed in the country might be, although one could make the same argument of the spoilers and such that appear on tarted-up Sports versions of sedans. It would also take space away from where most cars store their spare tires!

[MrAerodynamicist]

Originally posted by Martin
Paolo I thought your post was excellent, thank you.
A question: Is case D an explanation of why road cars go light at speed (I remember the original MR2 I once drove became quite "interesting" (ie terrifying) above about 150/90 kmh/mph) and if so why don't you see diffusers on ordinary road cars?

I think the problem with most road cars is that they lack any real downforce devices - most wings on road cars tend to be fairly cosmetic, while ground effects are impratical - while the overall shapes tend to resemble wings and produce some level of lift.

[Paolo]

Martin,
I agree with Desmo : the main problem is that a diffuser on a road car takes space away from the luggage compartment.
Anyway some road legal sports car have it.
Another problem is that a Venturi has a damn non linear (and not quadratic, as , for example, wings) effect : its effectivenes changes a lot with ride height (i.e., with speed and bumps) , so it could be not easy for an ordinary driver to tackle it.
And Desmo's country is the place were Porsche was condemned in Court for not stating on the manual that their 911 Turbo was more difficult to drive than an ordinary sedan...

[Paolo]

Again about Martin's post :
I don't know the underside shape of MR2's front, but if it looks like case D), you bet you have a problem.
Anyway I doubt it can be worse thn a Triumph Spitfire .
And, for the series "Aero is not everything", Spitfire is the car I have loved most in my life. If only it had SOME reliability....

[Paolo]

Ahem,
I'm getting a little bit presentialist... third post today...
But, again about Martin's interesting post about front lift on road cars, I would like to point out that most road cars have (some) front lift , even if they are flat bottomed (they 're open bottomed, actually...)
This is because the flow separates at the start of the hood, because of the abrupt angle you find there.
The separation "bubble" that results causes lift.
Only cars with downward sloped front hoods (the front drive Lotus Elan comes to mind)totally avoid this problem

[AndersF1]

Paolo: I agree with the rest, a very good explanation about the venturi tunnel.
Actually, I think it's so good that I may "steal" your picture to explain this phenomena in a swedish F1 BB ( Forza motorsport bb's technical forum ), if it's OK with you (too lazy to draw my own picture )

As some of you seems to have some questions about case D, I'll give a little remark about this case. I don't know if it helps, or if it increases the confusion... At the front of the car it's OK to start with a narrow, small, area, because you will get a "flow pipe" in the free air anyway, for some conditions. It's not as effective as having a true venture-shaped flow pipe, but it works. This is not true at the rear, mainly because you would need a positive pressure gradient here, so the flow breaks down to a wake soon after it leaves the car. This means that you get a small lift, instead of downforce, even if you have the same front as case B and C, as long as you have a rear like case A and D.

If you look at the Koenigsegg in the supercars topic, you can see a road car that have venturi tunnels, or at least diffusers.

About regular cars having front lift instead of downforce; the same is true for the rear, on many cars. See for example the Audi TT, which had severe handling problems when it reached close to and above 200 km/h, before they redesigned it. They had many accidents with the TT on the autobahn in Germany, partly due to this. I think that even one of the Audi-bosses died in one of these accidents.

My field of research is aerodynamics, or CFD really, for military aircrafts, but I have some friends from University that now works in the automobile industry (SAAB and Volvo) and they say that a car is actually run more hours in the wind tunnel than an aircraft. So it is a little strange that they end upp with contructions that have lift instead of downforce. However, a car is full of compromises, luggage, space for your head, etc. Most of the wind tunnel hours are spent with practical problems related to ventilation, dirt on the side-windows, noice and so on.

[Paolo]

Anders,
of course you have permission to use my drawing.
Happy you appreciate it. Post the URL for us, after : I'm curious to see a Swedish explanation of my pic.
If any swedish gurl reads your forum , I'd be happy to give you also a personal pic : I'm the typical beautiful-clever-artistic-always happy-always singing Italian ... they'd love me, I know... really... please...

About wind tunnel study for road cars : a friend of mine has been working for some time in this field. He said me that he felt very frustrated : whatever he suggested, was rejected by the style department , even in front of major aerodynamic problems. Style department rules, that's all.

[kober]

Quite a lot offtopic:

Originally posted by AndersF1
See for example the Audi TT, which had severe handling problems when it reached close to and above 200 km/h, before they redesigned it. They had many accidents with the TT on the autobahn in Germany, partly due to this. I think that even one of the Audi-bosses died in one of these accidents.

I'm quite sure, that the TT case wasn't aero related. Though one of the solutions was adding a little wing between the taillights to get some downforce on the rear, the main point was tuning the dampers, springs and antirollbars towards more understeer. The Audi probably wanted to build fast, responsive and neutral car on the standard concern platform with front drive (also Seat Toledo, Audi A3, Golf and Skoda Octavia are using the same platform) and forgot, that this car will not be driven by experienced racers, but by businessmen and their wives. The modified TT is said to be very unresponsive and boring.

[Paolo]

Well, front wheel drive cars are boring anyway...

[Marco94]

From what I have heard, the TT case was definitly an aero problem! Because of the rounded shape, Von Karman vortices would appear at a speed of about 150 km/h. These vortices resulted in rear end instabilities, thus an unpleasant ride and possibly a crash. That's the reason for the little spoiler. It looks hiddious so Audi would not have used it if it was not necesarry.

Marco.

[kober]

Originally posted by Marco94
From what I have heard, the TT case was definitly an aero problem! Because of the rounded shape, Von Karman vortices would appear at a speed of about 150 km/h. These vortices resulted in rear end instabilities, thus an unpleasant ride and possibly a crash. That's the reason for the little spoiler. It looks hiddious so Audi would not have used it if it was not necesarry.

From what I have read, the problem appeared when throttle was lifted in fast curves. Every car has tendencies to oversteer then, but it was quite fast in TT. Really, never heard about any aero in it.