Under floor flow & diffusers

[Paul]

Don't think of the aerofoil as a diffuser, think of it as a floor with a diffuser.

[henra]

The diffuser does not make a low pressure area. Just before the mouth of the diffuser is the lowest pressure area, the diffuser then slows the air down, increasing air pressure back to ambient as it exits the diffuser.

When assuming constant mass flow (no Spillage) this slowing down to ambient means that at the entry of the diffuser the air needs to move faster than at the Exit. When air Speed at the Exit is ambient this means that at the mouth the air flow is faster than ambient and thus Bernoulli tells us that there will be lower static pressure than above the car and thus Downforce exerted.
Thus a working diffuser does create low pressure at the mouth. It is its basic working principle.

[shelly]

You're talking a different thing. Diffuser have clear definition. Diffuser without the floor would itself be floor .
If you look at pressure distribution on this picture, you see that the lowest pressures are at the zone which is closest to the ground:
http://www.mh-aerotools.de/airfoils/images/wig_5.jpg
Now, I am not saying diffuser does not provide downforce. However, most of its effect is in enhancing the flow upstream of it.

PS, which is why it is possible to have to big a diffuser -- if the feeding flow is inadequate.

That 'inverted' aerofoil is not behaving the same as a F1 diffuser. Look at the streamlines ahead of the aerofoil - it is drawing air from way above the leading edge into the venturi gap that the aerofoil makes with the ground. In a F1 diffuser that air is not available so it does not behave as shown by this diagram.

I disagree with that. A F1 floor, in spite of all the regulation forcing flat sections and steps and so on, still behaves like an inverted aerofoil. The front of the sidepod is the leading edge (hence the rules on the 50mm radius at the edge and the importance of undercut), then it has to be flat up to the diffuser kink line (aka kick).

If we had a flow viz of the stream line before the sidepod we would see something similar to timbo's image, with squeezed streamlines

There are 2 strong low proessure peaks under a floor -diffuser: one at the leading edge, the other at the diffuser kink line. Then thera are 3d effects (side sealing, vortex acceleration) and trapped vortices (i.e. vortex tubes running beneath the floor giving downforce.

[chuckdanny]

Here is an example from sauber



It seems that it is a double diffuser year model with those throat in front of the extractor

[tok-tokkie]

You're talking a different thing. Diffuser have clear definition. Diffuser without the floor would itself be floor .
If you look at pressure distribution on this picture, you see that the lowest pressures are at the zone which is closest to the ground:
http://www.mh-aerotools.de/airfoils/images/wig_5.jpg
Now, I am not saying diffuser does not provide downforce. However, most of its effect is in enhancing the flow upstream of it.

PS, which is why it is possible to have to big a diffuser -- if the feeding flow is inadequate.

That 'inverted' aerofoil is not behaving the same as a F1 diffuser. Look at the streamlines ahead of the aerofoil - it is drawing air from way above the leading edge into the venturi gap that the aerofoil makes with the ground. In a F1 diffuser that air is not available so it does not behave as shown by this diagram.

I disagree with that. A F1 floor, in spite of all the regulation forcing flat sections and steps and so on, still behaves like an inverted aerofoil. The front of the sidepod is the leading edge (hence the rules on the 50mm radius at the edge and the importance of undercut), then it has to be flat up to the diffuser kink line (aka kick).

If we had a flow viz of the stream line before the sidepod we would see something similar to timbo's image, with squeezed streamlines

There are 2 strong low proessure peaks under a floor -diffuser: one at the leading edge, the other at the diffuser kink line. Then thera are 3d effects (side sealing, vortex acceleration) and trapped vortices (i.e. vortex tubes running beneath the floor giving downforce.

Thanks for that. So the leading edge is way forwards and we have a much extended low pressure area but still the drawing in of streamlines - even though the leading edge is sharp rather than radiused as on an aerofoil.

There was an earlier nice graph of the pressure distrubution under the floor showing the big drop in pressure at the start of the diffuser. It was in this external link: http://www.mulsannescorner.com/diffuser.htm

[SiLo]

Here is an example from sauber

http://oi61.tinypic.com/34o4n46.jpg

It seems that it is a double diffuser year model with those throat in front of the extractor

I didn't realise the front of the floor generate that much downforce as well. Or at least, that much extraction.

[chuckdanny]

Here is an example from sauber

http://oi61.tinypic.com/34o4n46.jpg

It seems that it is a double diffuser year model with those throat in front of the extractor

I didn't realise the front of the floor generate that much downforce as well. Or at least, that much extraction.

Yes!
And as i tried to express, the pattern seen suggest the structure of the flow :


That is, it is slightly divergent at the inlet due to the sidewash to shield this inlet from the turbulent wake of front tire but at the back it seems that it act as a big convergent thanks to the presence of the rear tires.
But the front wing accelerate the flow also because it is shaped (section reduction between road and wing) and the front of the floor is shaped also plus the skid plank, it's not flat. And of course front tire create a convergent structure.
The document posted by tok-tokkie on mulsanncorner say that regardless, a simple plate with rake would create the same pattern. Now in which quantity we don't know.

[SiLo]

Here is an example from sauber

http://oi61.tinypic.com/34o4n46.jpg

It seems that it is a double diffuser year model with those throat in front of the extractor

I didn't realise the front of the floor generate that much downforce as well. Or at least, that much extraction.

Yes!
And as i tried to express, the pattern seen suggest the structure of the flow :
http://oi59.tinypic.com/sq5h8n.jpg

That is, it is slightly divergent at the inlet due to the sidewash to shield this inlet from the turbulent wake of front tire but at the back it seems that it act as a big convergent thanks to the presence of the rear tires.
But the front wing accelerate the flow also because it is shaped (section reduction between road and wing) and the front of the floor is shaped also plus the skid plank, it's not flat. And of course front tire create a convergent structure.
The document posted by tok-tokkie on mulsanncorner say that regardless, a simple plate with rake would create the same pattern. Now in which quantity we don't know.

You might find that the double diffuser cars generate a much lower pressure area a the front of the floor simply because they were extracting more air from underneath the car. I would imagine that area would be of a higher pressure now, but most likely the same across the board. In fact, the front floor should theoretically work along the same level as the diffuser. The more air the diffuser can extract the less air underneath the entire body of the car.

[olefud]

The slipstream is accelerated and displaced by the vehicle bodywork. It is desirable to decelerate (raise the pressure of) the slipstream before it exits the rear of the vehicle to avoid a low pressure area at the rear of the vehicle that would increase drag. By optimizing the slipstream velocity at the throat of the diffuser the pressure can be lowered to develop downforce. There’s no real conflict between these two diffuser functions. Design engineers no doubt optimize the downforce aspect as it produces minimum drag compared to a wing. The latter also produces the turbulence that bedevils following cars and the upward displacement of the slip stream.

[mandark]

Wow lots of misinformation put forward by some, in authorative tone as well.

Interesting pic shelly. Those low pressure areas (at inlet and kink) are much larger (longer in x) than you'd expect in a 2d view of an undertray, but it makes sense when you bear in mind the changing section area in y.

One interesting thought..the air under the floor is propelled backwards relative to the road despite a 200mph car shooting forwards over it?

[shelly]

From the CFD thread of this forum, some months ago:

Actual screens from Williams' design office

leading edge suction peak clearly visible, along with vortex low pressure strake at the front (the blue narrow zone protruding from the front).
Also noticeable double diffuser effct and diffuser kink line suction peak

[olefud]

One interesting thought..the air under the floor is propelled backwards relative to the road despite a 200mph car shooting forwards over it?

At least according to Bernoulli. But then it’s at essentially zero velocity before the car comes along.
Also, Bernoulli address a free, incompressible, inviscid fluid which is hardy the case in the throat of the diffuser. With the static ground boundary layer and the dynamic diffuser throat boundary layer, there is probably a spectrum of velocity realities.

[PlatinumZealot]

Well, first it's not new, the also called bat wing was there with the W05 while not at every grand prix.
They are many theories about the use of vortices. Some claim it seal the underfloor, it improves the speed carried through the bottom of the sidepod by the stream under the nose and maybe burst right in the coke bottle shape area behind the sidepod at the very top speed where maybe the flow is detaching creating drag.

For me, it "obviously" can't seal the under floor. It's rotating the wrong way to do that. This would actually draw air out from under the floor, not force it to stay under there.

Hey Moose check out this direction of rotation.
We know the Main Vortex that seals the floor is from the curly bits on the corners of the leading edge of the floor, I don't know if the y250 helps with that or do something else, but look at the direction of curl of these Y250 vortex. It is the same direction as the bat wing. It is outwardly rotating relative to the edge of the floor. What do you think of this?

[CBeck113]

One interesting thought..the air under the floor is propelled backwards relative to the road despite a 200mph car shooting forwards over it?

At least according to Bernoulli. But then it’s at essentially zero velocity before the car comes along.
Also, Bernoulli address a free, incompressible, inviscid fluid which is hardy the case in the throat of the diffuser. With the static ground boundary layer and the dynamic diffuser throat boundary layer, there is probably a spectrum of velocity realities.

That is not fully correct - as I said in an earlier post in this discussion, the calculation is done with the car stationary, and the air moving under the floor, while the reality is that the air is actually being pulled in the direction of the car's movement, with boundry layers on both the ground and on the floor. This change in POV is done to simplify the calculation - it is much easier to calculate air movement over the floor as a closed system than to try and calculate what happens to a volume of air when a car passes through it. The risk is that the simplification causes the engineers to miss important details, and the simulated data does not depict reality (anyone remember the Ferrari aero discussions with the poor coorelation?)

[MadMatt]

From the CFD thread of this forum, some months ago:

Actual screens from Williams' design office

http://a.yfrog.com/img850/8106/hcjw.png

leading edge suction peak clearly visible, along with vortex low pressure strake at the front (the blue narrow zone protruding from the front).
Also noticeable double diffuser effct and diffuser kink line suction peak

Yes you really see how powerful the management of the flow upstream of the floor is. Very impressive! All down to the front wing?