Home  
Site Partners: SpotterGuides Veloce Books  
Related Sites: Your Link Here  

Go Back   TenTenths Motorsport Forum > Racing Talk > Racing Technology

Reply
 
Thread Tools Display Modes
Old 7 Jun 2006, 10:50 (Ref:1629148)   #1
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Diffusers

Hi,
I'm managing to get myself a bit confused with how diffusers work and why they are benificial. I understand that, to have what is called 'ground effects', you want to narrow the gap between the road and the underbody to increase the velocity of the air and reduce pressure. What I don't quite understand is why you then need to have a diffuser at the rear of the car for pressure recovery and why you don't just want the underbody-road gap to narrow, without getting wider again?

Is it becuase the boundry layer will get to large and choke the flow?

Is the main function of a diffuser more to channel the flow exiting the car upward to get downforce, rather than for pressure recovery?

Would having the narow gap continue all the way to the back of the car have a lower mass flow than with a diffuser?

I'd really love someone to clear this up for me.
Thanks
browney is offline  
Quote
Old 7 Jun 2006, 11:24 (Ref:1629163)   #2
glenn22481
Rookie
 
Join Date: Dec 2004
Guernsey
Posts: 30
glenn22481 should be qualifying in the top 10 on the grid
have a look at http://scarbsf1.com/diffuser.html

it doesnt go into the physics side of a diffuser but will simlify the understanding of what they do and how they work. Also worth noting is that rules and regs will effect overall design.
glenn22481 is offline  
Quote
Old 7 Jun 2006, 13:40 (Ref:1629248)   #3
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Thanks for the link,

I'm still not 100% clear on it though. I keep thinking of pipe flow where the pressure will be lower in a constriction but higer in a wider part. That would make pressure higher in a diffuser, which would give lift not downforce from the diffuser.
browney is offline  
Quote
Old 7 Jun 2006, 21:11 (Ref:1629526)   #4
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
The diffuser part allows the flow to slow back down and the pressure to gradually head back towards the freestream value rather than having a kind of jet of high-speed air coming out from under the car. The loss of energy from such a jet mixing with near-stationary is much, much larger and means that the effective resistance of the underbody channel is very high.

As i gather you already know, the underbody airflow is very similar to flow through a normal pipe. Since the pressure at the front of the car is effectively fixed and the base pressure (general level of static pressure in the 'void' like region immediately behind the vehicle) normally only varies over a fairly narrow range, the pressure difference across the 'pipe' is pretty much fixed. This means the amount of airflow is essentially controlled by the resistance of the pipe. Low resistance means more flow. More flow for a fixed cross-sectional area means higher velocity. Higher velocity means lower pressure. Lower pressure leads to more downforce.

For maximum effectiveness you can try and reduce the pressure in the base area by careful location of a wing. This low pressure will increase the difference across the underfloor 'pipe' and either give you more flow for a given geometry or allow you to angle the diffuser more steeply by reducing the metephorical hill the air has to climb on it's way back to freestream static pressure. A steeper diffuser means you can have a longer flat section to the floor and/or exaggerates the suction peak you get at the very apex of the diffuser, where it joins onto the flat bit of the floor. Both mean more downforce, partly by increasing the size of the region exposed to the lowest pressure and partly because you're turning the flow upwards around a larger angle.

I'm sorry if any of this is teaching you to suck eggs (always thought that's a strange expression), but i hope some of it helps.

Cheers

Rob
Locost47 is offline  
Quote
Old 7 Jun 2006, 21:27 (Ref:1629539)   #5
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
Sorry, just read the scarbsf1 link. It's not right. A diffuser IS named appropriately, does NOT actively scavenge airflow and does NOT reduce boundary layer thickness. That last point is the exact opposite of what actually happens - the adverse pressure gradient in the diffuser causes the boundary layers to thicken and is ultimately the limiting factor on how steep you can go. They've made a good effort but the information is misleading.
Locost47 is offline  
Quote
Old 8 Jun 2006, 01:00 (Ref:1629631)   #6
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
I have no idea what 'suck eggs' means!!!

O.k, That's helping. It seems like your saying that having a diffuser helps because it allows the faster air of the underbody to mix more easily with the slower freestream.

So if you have a diffuser the velocity in the thinest road to underbody region of the underbody will be higher than for a car where the same road to underbody distance is maintained all the way to the back of the car?

What causes the pressure to be lower at the start of the diffuser, where the body goes from flat to turning upward? I get confused because when thinking about pipe flow, like a venturi tube, as you increase thickness of the pipe the pressure increases, seemingly the oposite to what happends at the start of a diffuser.
browney is offline  
Quote
Old 8 Jun 2006, 08:38 (Ref:1629777)   #7
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
Never heard the expression "teaching your grandmother to suck eggs"? Actually that's probably for the best as it doesn't make a lot of sense. It's supposed to mean telling someone something they already know but what's that got to do with sucking eggs? Strange.

Anyway, yes, the velocity under the car will ganerally be greater for a car with diffuser than for one without. For real cars there can be several other factors which strongly affect this as well but the principle is still correct.

What causes there to be a peak in the low pressure at the apex of the floor/diffuser intersection is the way the air turns to follow the surface of the car, which is now angled upwards. I should clarify that the peak suction occurs on the car side of the 'pipe' and *not* across the whole width, though it may extend quite a way across. What happens isn't straighforward to explain - but i'll try so please bear with me but tell me if it's not clear or doesn't make sense.

The air is compelled to stick to the car body by virtue of its (low but very important) viscosity. Also it is still subject to Newton's laws so to change it's direction there must be a force acting on it to change direction. If the corner is not too sharp, the viscous effect between car-air is enough to pull around the air immediately adjacent to the surface but the air-air viscosity isn't enough away from the surface and so something else has to do it.

A pressure gradient is required within the air perpendicular to the diffuser apex to push/pull it around the corner. Pressure and speed are intrinsically linked via an often mis-qouted relationship derived by a chap called Bernoulli. Bernoulli's (complete) formula shows that a fluid can 'borrow' energy from one form of pressure and lend it to another. In this way speed and pressure are inversely proportional to each other (with a couple of caveats which i won't go into here).

The required pressure gradient thus forms via a speed gradient within the air. The air near to the surface but not immediately adjacent has to turn through the greatest angle so it needs the greatest pressure force acting on it, hence the speed there must be the greatest. Further out the turn radius is larger so the change in pressure doesn't have to be so big and so the speed change is smaller.

You get the same effect wherever air flows around a convex surface but where the curvature is not so severe that viscosty isn't enough to provide the force necessary to turn the air immediately adjacent to the surface around that tight a radius. Hence you get faster airflow and lower pressures around smoothly curving surfaces where the air has to change direction.

This is the same as what happens around wings which are essentially posh turning vanes. It happens at the front of the bonnet of your car, around the sides of the front bumper, over the top of the windscreen, around the A-pillars (though often rain gutters mess this up) - bascially wherever the air can still make the turn around the corner. Where it can't, it will head off in something vaguely resembling a straight line.

With a race car diffuser they often make the flloor/diffuser junction as sharp as they can get away with so that the air has to make the tightest turn and you get the lowest pressure at that point. This is great for peak downforce but can make the car very sensitive to small changes in pitch, heave (ride height) yaw or the presence of another car up ahead. Often you have to take a less aggressive approach and use a larger radius (still tiny in real terms) so that the air will still be able to make the turn under a wider range of upstream conditions.

Is that any clearer of have i made things worse?
Locost47 is offline  
Quote
Old 8 Jun 2006, 11:01 (Ref:1629877)   #8
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Yep, Your explanation makes sense. The diffuser acts like a wing, where a pressure has to act to accelerate the air around the curvature. It makes sense that you don't want the diffuser too steep, otherwise the boundry layer will seperate effectivly stalling the diffuser. I guess another way to think about it, is with conservation of momentum/newton's 3rd too. The air is accelerated upwards by the diffuser so the diffuser is being pushed downwards.

Are you able to tell me then, why is it then that an increase in cross sectional area causes the fluid to slow down (like in the diffuser on the exit side of a venturi), seemingly the opposite to this. I was thinking about it using the equation 'q = v1 A1 = v2 A2' Where the area was increasing so the air was slowing.

You won't confuse me (well hopefully not!) with getting complex, I'm 3/4 of the way through an automotive engineering degree, so I've done 2 years of fluid mechanics. It's funny, I've looked through my fluids text book and a text book on road vehicle aerodynamics I borrowed but the best explanation I get is on an internet forum!

Last edited by browney; 8 Jun 2006 at 11:10.
browney is offline  
Quote
Old 8 Jun 2006, 11:31 (Ref:1629892)   #9
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
The increase in cross-secional area does cause the fluid to slow down 'as a whole', in order to satisfy the conservation of mass, but the portion of it near to the diffuser apex goes quicker. If you integrate the speed distribution across the gap between the ground and the car then you still get the usual V1A1=V2A2 relationship overall.
Locost47 is offline  
Quote
Old 8 Jun 2006, 12:03 (Ref:1629914)   #10
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Ahh, O.K thanks! Now I get it!!!

Does the acceleration at the apex give a net downforce on the car if you look at the diffuser as a whole or is the effect there to help speed up the flow under the body of as a whole.

Thank you so much for explaining this to me. I asked my Aero lecturer but his explanation wasn't clear. He spoke more about the diffuser getting the boundry layer away from the ground so as not to choke the flow.

Last edited by browney; 8 Jun 2006 at 12:05.
browney is offline  
Quote
Old 8 Jun 2006, 16:31 (Ref:1630073)   #11
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
Both! The peak suction at the diffuser apex is an additional downforce contribution, but even with a really large radius there you will still have more massflow through the underbody and hence the high speed/low pressure situation.

On race cars where either the regulations aren't too strict or where other considerations reduce the importance of diffusers you sometimes see a quite gradual blending of the flat part of the floor into the diffuser. One of the highest-downforce car's i've ever tested was actually a road car but with diffuser channels running almost the entire length of the car, from a narrow 'throat' between the front wheels to twin 500mm square tunnel outlets at the rear. The only thing i've tested that had more downforce was a full-on ground effect F1 car with completely sealing side-skirts, and even that wasn't much better.

Not sure what your lecturer means. You can sometimes get a separation from the ground just behind a car with a diffuser. This is down to a boundary layer forming on the ground as well as the car because the flow under the car is faster than the road speed of the vehicle. This b.l is then still subject to the influences of pressure gradients and a combination of the upsweep from the diffuser and the return to freestream pressure from the low pressure under the car causes it to separate. Other than that, i've never known a diffuser be 'choked' by viscous effects unless it's at something daft like 2-3mm ride height. Not a realistic condition for anything with any suspension travel/compliance, though someone did once tell me that they run their F3 car at a ride height of -3mm, so the underfloor plank would be ground away down to the limit of the regulations and so get their car's underbody lower than was really intended by the rules. Interesting...
Locost47 is offline  
Quote
Old 8 Jun 2006, 23:06 (Ref:1630282)   #12
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Thanks for answering my questions. I'm sure you have better things to do!

My lecturer was saying that if you just had a flat underbody, low to the ground and no diffuser, the boundry layer could get thick enough to start to choke the flow. Thus reducing downforce compared to having a diffuser.

Am I right in this flollowing image, showing the pressure difference between a flat underbody and having a diffuser?


Can I ask, what do you do that means you understand this so well. Are you an engineer or scientist?

Last edited by browney; 8 Jun 2006 at 23:14.
browney is offline  
Quote
Old 9 Jun 2006, 07:01 (Ref:1630390)   #13
greenamex2
Veteran
 
Join Date: Aug 2002
England
Hertfordshire
Posts: 1,686
greenamex2 should be qualifying in the top 10 on the grid
Really simple question to which I would appreciate a really simple answer, being a really simple person.

Does the diffuser on an Elise/Exige work?

Thanks.
greenamex2 is offline  
Quote
Old 9 Jun 2006, 08:46 (Ref:1630445)   #14
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
Firstly to answer Dennis - yes, the Lotus underbodies do work. The top aero guy there really does know what he's doing. Conversely, the Ferrari 360 & 550 have underbodies which are pretty but don't work. The 360's is too steep and the 550's is scuppered by poor management of the engine cooling airflow right at the front. It can be no coincidence that the Challenge Stradale 'race-spec' version of the 360 has an extra element inside the standard diffuser which makes it a whole lot shallower...


Back to browney - yes those pressure distributions are about right, though the pressure recovery from the entrance will be more rapid than that so there will be a peak there, similar to the one at the apex of the diffuser, before the gradual decay. Be aware though that once past that initial zone, the main reason the pressure increases is because of inflow from the side edges of the car. If you seal those edges (with sharper / deeper sills or, even better, full side skirts) then the pressure stays very low all the way to the back - hence massive downforce! The pressure rise is contrary to the boundary layer development on the car/floor which acts to constrict the flow further and so increase the speed and reduce the pressure, though in practice that's not really an issue.

The boundary layer thickness only becomes a problem at very, very low ride heights. You're probably only looking at maybe 20-30mm boundary layer thickness as a maximum, right at the back of the upper surfaces of the car but underneath the sizable acceleration of the flow thins them out much more, to maybe 5mm? A thinner one will form on the floor as well due to the mismatch between accelerated flow beneath the car and the car's travelling speed. Overall i'm confident that you won't get a viscous limiting of the flow under the body until you're down below maybe 10mm ride height (assuming you still design the inlet and diffuser sections appropriately).

If the underbody isn't smooth, then that distance increases, but then you're significantly reducing the chances that anything other than the shallowest of diffusers will work anyway.

To answer your last question, i'm an aerodynamics consultant by trade, though i tend to get dragged into various other areas of automotive development as well, if only as an extra pair of hands. I'm lucky in that i get to work on all different types of vehicle from trucks to F1 cars. The variety stops things getting boring and keeps you on your toes, as you are expected to know your way around each vehicle as if it was your whole life. Takes a lot of work but i wouldn't have it any other way.
Locost47 is offline  
Quote
Old 9 Jun 2006, 09:07 (Ref:1630457)   #15
greenamex2
Veteran
 
Join Date: Aug 2002
England
Hertfordshire
Posts: 1,686
greenamex2 should be qualifying in the top 10 on the grid
Ta for that. Sounds like as well as the planned tacking on of a Lotus Elise diffuser I also need to concetrate on smoothing the underside of the car AND blocking all the holes/'transmission tunnel' ahead of my planned wind tunnel session.

Oh, and get an engine.

Out of interest how big a protusion into the underneath of a car does it take to affect the workings of a diffuser? Are we talking rivett heads, bolts head or one inch box section?
greenamex2 is offline  
Quote
Old 9 Jun 2006, 10:08 (Ref:1630505)   #16
browney
Racer
 
Join Date: Aug 2003
Australia
Adelaide, South Australia
Posts: 316
browney should be qualifying in the top 10 on the grid
Denis Bassom
I guess it would depend on the front spoiler as well. A front 'lip' can effectivly shelter some of the protrusions at the bottom of the engine bay also reducing underbody drag. I guess it is possible for flow to seperate in the underbody if you have protrusions, and that would reduce the effect of the diffuser as well as increase drag.
Although, as I have demonstrated, I'm not an expert.

Locost47
I guess that another advantage of a diffuser is that by raising the pressure at the back and 'boat tailing' the profile of the rear, you will propably decrease drag too. Does anyone ever use a 'nozzle' shape on the front of the car to try and get faster underbody flow, or will this increase front end lift too much?

Do you know where I would be able to get some info on diffusers? I've read Hucho's book on the 'Aerodynamics of road vehicles' but our library doesn't seem to have much more and I can't find any journals that I have full text access for on the subject. It seems like getting answers on bluff body aero is really difficult (unless you hang aroudn on ten-tenths ;-) ).I'd love to find out more about the difference between mixing the air at the rear of the car at high speed vs. bringing back down to closer to the free stream.

I hope you love your job, I dream of doing stuff like that!

Last edited by browney; 9 Jun 2006 at 10:16.
browney is offline  
Quote
Old 9 Jun 2006, 10:46 (Ref:1630533)   #17
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
Denis (with one 'n' now, sorry about that)
As Browney says, things like nuts & rivets won't matter so much, within reason, but it's normally big stuff like chassis members or suspension gubbins which scupper the effectiveness of a diffuser but i would also add poorly planned cooling system outflows and the disturbances from the wheels.

Ideally the cooling airflow should be completely ducted to the radiator/intercooler etc and then compeltely ducted out of the car, although some provision may have to be made for direct cooling of engine/gearbox or specifically sensitive components. Regardless, this cooling flow must not be allowed to leave via the underside of the car if the diffuser is to work well.

For the wheels it would be nice to have endplates which extend forwards to the front of the rear wheelarches in order to try and stop the peak suction at the apex of the diffuser drawing in air through and around the rear wheels. This will make an unholy mess, though will be handy for brake cooling.

Also important is to have a smooth-ish transition from the floor to the diffuser. A lot of people, including specialist car companies who should know better, manufacture it out of two flat pieces, with a join exactly at the apex of the diffuser. Unless this panel match is perfect you will get a gap, or even worse a step, in the floor which can knacker the diffuser. Best to have it made from a single flattened v-shape piece with the seam 100mm or more upstream, so that the disturbance from any mismatch has some time to sort itself out before it reaches the critical bit at the start of the diffuser.

browney
Yes a rounded front end will increase front lift but it doesn't have to be much of a radius at all since air is more forgiving when it first arrives at the car and hasn't had any of its energy stripped from it through viscosity and turbulence. Also you benefit becaus this further reduces the energy losses through the underbody and so increase massflow etc and you do get some of that benefit at the front of the floor. The 360 Challenge Stradale is quite a good example of this but you can spot it on the newer breed of Le mans cars too, which all tend to have a raised and sometimes radiussed centre section to the front splitter.

For more information you could try:

Competition Car Aerodynamics
Simon McBeath
Haynes Publishing 2006
ISBN 1 84425 230 2

or

Race Car Aerodynamics: Designing for Speed
Joseph Katz
Robert Bentley Publishers, 1996
ISBN: 0 8376 0142 8

Simon's book is more modern and has lots of cool pictures (done by our competitors at Advantage CFD, grrrr!) but has a misleading and technically incorrect explanation of how wings work, although unfortunately this is the version which is often written in textbooks. We're actually working with Simon on his next set of articles for Racecar Engineering magazine so we're putting him straight. Apart from that though his book is excellent. Both books are available from places like Amazon or Waterstones etc.

Cheers

Rob
Locost47 is offline  
Quote
Old 13 Jun 2006, 12:45 (Ref:1633457)   #18
zac510
Veteran
 
zac510's Avatar
 
Join Date: Nov 2005
Posts: 2,714
zac510 should be qualifying in the top 10 on the grid
Interesting read. I've recently purchased CCA but have only covered the first couple of chapters.

I'd like to hear thoughts on the use of strakes in diffusers. The current GT (endurance) trend seems to be more, more, more! Sometimes over half a dozen strakes that run all the way to the trailing edge.
zac510 is offline  
Quote
Old 13 Jun 2006, 13:47 (Ref:1633514)   #19
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
I've not done any real parametric tests with the number of strakes in a diffuser so i can't say what the optimum is likely to be. I know the primary function they serve is to reduce the amount of cross-flow in the diffuser and try to keep it more like the 2-dimensional idealised situation.

The more strakes you have the more '2D' the flow should be. I presume there would be a point though where the amount of effective diffuser area you're losing due to the unpleasant boundary layer goings-on in all the 90-degree corners betwwen strake & diffuser surface outweighs the other gain.

Also there are occasions where cross-flows can be beneficial, however, such as when you want to run a wing/diffuser steeper than you could normally get away with. By allowing some flow around the side edges (making it more '3D') you can re-energise the tired boundary layers and keep the flow attached for longer. There is a fairly hefty drag penalty for this but, if done cleverly, the overall balance between downforce/drag can still work in your favour.

If you're being really cunning you can deliberately manipulate the cross-flow to generate strong vortices whose extremely low pressure core acts on a conveniently located nearby surface and pulls the car down. A bit like how delta wings work on, though there are other things going on as well. It's also now a commonly used trick on the top flaps of F1 car rear wings.
Locost47 is offline  
Quote
Old 14 Jun 2006, 12:10 (Ref:1634249)   #20
zac510
Veteran
 
zac510's Avatar
 
Join Date: Nov 2005
Posts: 2,714
zac510 should be qualifying in the top 10 on the grid
Thanks, that's some more food for thought
zac510 is offline  
Quote
Old 14 Jun 2006, 13:12 (Ref:1634303)   #21
Redracer77
Veteran
 
Join Date: Oct 2003
Location:
Didsbury/Chorley
Posts: 3,446
Redracer77 should be qualifying in the top 10 on the grid
an interesting read, that for all the input. Now I know why my F4 was rubbish when the diffuser fell off!!
Redracer77 is offline  
Quote
Old 31 Jul 2006, 15:26 (Ref:1669015)   #22
zac510
Veteran
 
zac510's Avatar
 
Join Date: Nov 2005
Posts: 2,714
zac510 should be qualifying in the top 10 on the grid
I was reminded of this thread by this picture, which a forum-friend took during a visit to the Ferrari Gallery at Maranello.




However I think the 'pressure paint' is merely airbrushed on to represent distribution.
zac510 is offline  
Quote
Old 1 Aug 2006, 08:10 (Ref:1669661)   #23
Locost47
Racer
 
Join Date: Apr 2006
England
Posts: 185
Locost47 should be qualifying in the top 10 on the grid
LOL! Yes that's a highly stylised interpretation of some CFD results as painted onto a model. It's what you would *want* it to look like...

I was disappointed to see that pressure sensitive paint only works in monochrome. You have to shine a fancy light at the model whilst in the wind tunnel and then measure the light intensity reflected back from it. The paint absorbs light in different amounts depending on the pressure exterted upon it so you end up with a kind of black & white version of those pictures above, only with real results!
Locost47 is offline  
Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Diffusers on Grand-Am DSPs Dauntless North American Racing 11 3 Jan 2003 20:25


All times are GMT. The time now is 11:26.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Original Website Copyright © 1998-2003 Craig Antil. All Rights Reserved.
Ten-Tenths Motorsport Forums Copyright © 2004-2021 Royalridge Computing. All Rights Reserved.
Ten-Tenths Motorsport Forums Copyright © 2021-2022 Grant MacDonald. All Rights Reserved.