Lift

[skyking]

On the Jaguar picture above, the center 6 tufts are not looking so great downstream of the shark fin radio antenna.
That area does not look attached to my eye. Either side of it, the two rows are looking good and I think greatly influenced by flow wrapping around from the sides. In the center, the window slope is determined not by aero but by styling. It is truly too steep to stay attached.

[JulianEdgar]

Originally Posted by skyking

On the Jaguar picture above, the center 6 tufts are not looking so great downstream of the shark fin radio antenna.
That area does not look attached to my eye. Either side of it, the two rows are looking good and I think greatly influenced by flow wrapping around from the sides. In the center, the window slope is determined not by aero but by styling. It is truly too steep to stay attached.

As I've previously stated, there's a small separation bubble in the lower, middle part of the rear glass of the Jaguar. Fairly common - and interestingly, not shown in Jaguar's CFD.

There's attached flow either side and on the trunk lid.

This sort of little separation bubble is usually easily fixed with some vortex generators - it's quite minor.

Another pic:

[freebeard]

This sort of little separation bubble is usually easily fixed with some vortex generators - it's quite minor.

So, not a spoiler then?

[aerohead]

Originally Posted by JulianEdgar

Yes, you're right - my mistake. Otherwise, my point remains - flow wrapping around those upper curves create lift.



That diagram is for cars (Type 31 Volkswagen fastback, squareback and notchback) that are 59 years old!

Flow patterns on modern car shapes (eg notchbacks) are quite different. This is a primary cause for your confusion - car shapes (and their flow patterns) have changed, but you act as if they have not.

So let's consider 'flow-wrapping'.
1) lift is a function of pressure distribution.
2) pressure distribution is a function of velocity
3) velocity is a function of streamlines
4) streamlines are a function of body shape
-------------------------------------------------------------------------------------
* the lowest pressure on the upper body of a contemporary automobile is at the suction peak, just ahead of the windshield header. The streamlines are closest together, the air is at maximum velocity, and pressure at it's minimum.
* if the body cross-section contracts rearwards of the roof apex, then by default, as the flow travels further aft, it gets slower, and its static pressure increases.
* the further the flow gets away from the suction peak, the slower, and higher pressure it becomes.
-------------------------------------------------------------------------------------
A) if the flow can remain attached all the way to the rear, it will be at it's highest pressure, compared to the forward stagnation point, dependent upon aft-body length ( 'Verjungungsverhaltnis' as per 'Kamm's' research )
B) adding a spoiler would only force the flow into a faster, lower pressure regime. A spoiler's deflection, or wing's negative lift 'will ' develop downforce, but at a drag penalty.
--------------------------------------------------------------------------------------
* 'Downwash' does not qualify technically as 'attached' flow. Downwash is an artifact of separation-induced attached longitudinal vortices, which generate the highest drag. The small wake, which is also an artifact of flow separation is not the same as a small wake generated from a progressively diminishing body cross-section which produces only a moderate pressure increase, incapable of triggering separation of the boundary layer.
-------------------------------------------------------------------------------------
* modern notchbacks demonstrate architectures that are all over the map. There are Hyundai Accents, Kia Rios, Nissan Sentras, and the like which don't improve on some notchbacks of 1986.
* Hucho addressed in 1986 all the caveats of windshield- to- roof transition, roof camber, aft-body length percentage, onset flow off the roof, boot height, and especially C-Pillar shape, which would govern notchback performance.
-------------------------------------------------------------------------------------
1) I could not find a single instance in Hucho's 2nd-Edition where a spoiler wasn't in separated flow.
2) In his section on high-performance sports cars he recommended zero lift, or if you were going to introduce negative lift, then provide rear bias.
3) Schenkel's rear spoiler introduced front lift where none had existed.
4) beyond 17.8 mm height, Schenkel's spoiler increased drag, at the expense of downforce.
------------------------------------------------------------------------------------
* Kamm's research predicted Cd 0.303 for the 1964 Porsche 911 with a 'template' roofline, vs Cd 0.40.
* Kamm's research predicted Cd 0.29 for the VW New Beetle with a 'template' roofline, vs Cd 0.39.
* a Volkswagen wind tunnel photograph of the VW Beetle, which appears in the book, ' the new VW Beetle', by Jonathan Wood, page 28, shows smoke-flow imaging, with the boundary layer coming off the car on a 19-degree angle, exactly as predicted by the 'fast template.'
------------------------------------------------------------------------------------
Longitudinal centerline pressures on the upper body of a vehicle are meaningless without the underbody pressure profile, especially if there's an active diffuser.
--------------------------------------------------------------------------------------
Any information derived from tuft testing is extremely problematic. They clearly misrepresent the flow on the Beetle, New Beetle, and 911.

[aerohead]

Originally Posted by JulianEdgar

I wish you wouldn't keep saying this - saying it repeatedly does not make it true. The Insight is a good example:



Are you still trying to suggest that the low pressures over the roof and hatch are caused by separated flow?

If you have the color photo from Honda, of the Insight undergoing smoke flow testing in their wind tunnel, you'll be able to discern compromised flow over the hatch.
Turbulence cannot conduct pressure. Even burble-point is problematic. Pressure over the entire hatch area takes on the pressure of the separation line if it exists. The closer the separation is to the suction peak, the lower the pressure over all the roof behind it.
Aside from that, do you see how the negative pressure is trending downwards as it's further removed from the suction peak near the windshield?
Should the aft-body continue as an extended structure, you'd see an inflexion point, where the pressure goes neutral, then positive.
Hucho said 'neutral' lift was fine for passenger cars. Even for sports cars ( with caveats ).
As a BEV, the Insight would have a perfect belly pan and diffuser, which could easily address rear lift. The 'template' has a perfect belly and diffuser.
I own this car. I've never experienced objectionable driving behavior.
A neighbor's father in law, from which I bought the car, obtained 3-Wheeler's Insight boat-tail and 80-mpg HWY is not uncommon for him. And he reports no adverse driving behavior.

[aerohead]

Originally Posted by MeteorGray

Quote: "Flow patterns on modern car shapes (eg notchbacks) are quite different. This is a primary cause for your confusion - car shapes (and their flow patterns) have changed, but you act as if they have not."

Aerohead, do you acknowledge this, or do you refute it?

It's hard to argue points without a basis of common principles.

I have a hard time following all this, but it's interesting nevertheless.

I've addressed it in #34-permalink below.
Hucho addressed all this in his 2nd-Edition. There's nothing new that's transpired since 1986.

[aerohead]

Originally Posted by JulianEdgar

If you strip away the obscure language, it's actually easy to understand.

Old shape cars (eg notchbacks) had flow that separated at the end of the roof. So, my old W123 Mercedes shows flow separation from the end of the roof:



Note the tufts on the rear window and boot / trunk lid show separation (they're whirling around, not lined up in flat rows.) Basically, the wake extends to the height of the roof.

Conversely, modern cars have attached flow across the boot / trunk:



Note the tufts on the rear window and boot / trunk lid show attached flow (they're not whirling around but instead are lined up in flat rows.)

As far as I can ascertain, Aerohead believes the flow pattern shown on the Jaguar isn't real, in part apparently because the rear contours of the Jaguar dip more sharply than his Template. He believes that the flow is really separating far forward. (In fact, he has said at least once that it is separating at the windscreen header rail, believe it or not.)

Now to lift. If we look at the old Mercedes, we can see the boot / trunk lid is in separated flow, so in effect it is within the wake. The wake is a low pressure area and this low pressure is bearing on the boot / trunk lid, so causing lift. So in the old cars that Aerohead likes citing, in fact separated flow was causing lift.

But this theory is obviously quite invalid for modern car shapes (there's almost no separation), so where is the lift coming from? It is coming from the airflow wrapping around (ie attached to) the upper curves, so generating low pressures.

Here those low pressures can be seen on Jaguar CFD (hotter colour = lower pressures):



Not the low pressures all across that curved roof and part way down the rear window. The blue areas show that in fact Jaguar claim positive pressures on the boot lid (achieved by the rear spoiler).

My Insight shows it well - note the measured lift across the upper curves. The Insight has attached flow on all upper surfaces. (Length of arrows shows magnitude of force.)




So:

• the shape of the car determines the airflow pattern
• car shapes have changed a lot in the past 60 years
• theories of airflow need to take into account what is happening on current shape cars

Unfortunately, Aerohead has built entire theories on faulty and / or outdated premises, and then extrapolated them to the point where a great deal that he writes is completely wrong. (Not everything, but a lot.)

He then denigrates the real experts. Who? Well, he has made adverse comments here on a number of aerodynamicists, including the head of Porsche aero, the head of Jaguar aero, an F1 aerodynamicist, a former Tesla aerodynamcist - and so on.

So he doesn't read current aero literature, and doesn't want to learn from real experts.

That's not a good combination for giving people advice!

If you'd actually read Hucho's book you'd be stuck with the same conclusions.
Aeronautical engineer Larry Mauro doesn't know anything about automotive aerodynamics, although his friend Peter Brock certainly does.
My college chum Glenn got his Master's degree in aeronautical engineering from NASA, but it doesn't qualify him for automotive aerodynamics.
None of the car companies your associates worked for have generated novel low drag cars. The only reason Tesla gets a pass is because it doesn't have the cooling system and belly mutilations of a conventional ICE vehicle, otherwise, they're no better than a 1989 Opel Calibra.
Other than the popular literature, or what can be gleaned from online, my income precludes me from staying abreast of contemporary professional developments, however, the fundamentals don't change.
In 1986 Hucho thought the future would produce a Cd 0.20 car. In 1987, the Cd 0.19 Impact debuted. Hell of a ride! The future moved fast! I recommend we catch up to the past.

[aerohead]

Originally Posted by freebeard

The Type III Notchback will always be 'new' to me.



The [rear] trunk lid is remarkably short. I'm just holding my hands in the air but it seems like it was less than three feet. Not much to lift on.

Since I've got my albums open, here's my favorite Type III. I knew the guy, he used to deliver eggs by the dozen to me in it.

The fact is that, it had higher rear lift than the fastback.

[aerohead]

Originally Posted by skyking

On the Jaguar picture above, the center 6 tufts are not looking so great downstream of the shark fin radio antenna.
That area does not look attached to my eye. Either side of it, the two rows are looking good and I think greatly influenced by flow wrapping around from the sides. In the center, the window slope is determined not by aero but by styling. It is truly too steep to stay attached.

The 'template' comparison suggests a compromised boundary layer, possible attached longitudinal vortices, and vortex-induced downwash, exactly like, but not as severe as the Porsche 911/ 912.
If so, the area over the rear roof would carry the low pressure of the separation line upstream, closer to the suction peak, at a lower static pressure, explaining all the observed phenomena.
Only a smoke flow survey would reveal the actual air flow. I remain steadfast in my rejection of tufts as accurate indicators of flow.

[JulianEdgar]

Originally Posted by aerohead

So let's consider 'flow-wrapping'.
1) lift is a function of pressure distribution.
2) pressure distribution is a function of velocity
3) velocity is a function of streamlines
4) streamlines are a function of body shape

Yes, all correct.

-------------------------------------------------------------------------------------
* the lowest pressure on the upper body of a contemporary automobile is at the suction peak, just ahead of the windshield header. The streamlines are closest together, the air is at maximum velocity, and pressure at it's minimum.

Incorrect. This is the case only if the curve at the top of the header rail is the sharpest having attached flow. On the old cars you like to quote - probably yes. On modern shapes, often no.

* if the body cross-section contracts rearwards of the roof apex, then by default, as the flow travels further aft, it gets slower, and its static pressure increases.
* the further the flow gets away from the suction peak, the slower, and higher pressure it becomes.

Correct.

-------------------------------------------------------------------------------------
A) if the flow can remain attached all the way to the rear, it will be at it's highest pressure, compared to the forward stagnation point, dependent upon aft-body length ( 'Verjungungsverhaltnis' as per 'Kamm's' research )

Correct.

B) adding a spoiler would only force the flow into a faster, lower pressure regime. A spoiler's deflection, or wing's negative lift 'will ' develop downforce, but at a drag penalty.

Incorrect with regard to a spoiler. Spoilers can decrease drag and lift. If you don't believe me, take it up with Dr Wolf, head of Porsche aero. I can dig out the direct quote from him if you want.

Correct with regard to a wing.

--------------------------------------------------------------------------------------
* 'Downwash' does not qualify technically as 'attached' flow. Downwash is an artifact of separation-induced attached longitudinal vortices, which generate the highest drag. The small wake, which is also an artifact of flow separation is not the same as a small wake generated from a progressively diminishing body cross-section which produces only a moderate pressure increase, incapable of triggering separation of the boundary layer.

Incorrect. Attached flow is attached flow, and separated flow is separated flow. You have changed the definition of attached flow to suit yourself.

-------------------------------------------------------------------------------------
* modern notchbacks demonstrate architectures that are all over the map. There are Hyundai Accents, Kia Rios, Nissan Sentras, and the like which don't improve on some notchbacks of 1986.
* Hucho addressed in 1986 all the caveats of windshield- to- roof transition, roof camber, aft-body length percentage, onset flow off the roof, boot height, and especially C-Pillar shape, which would govern notchback performance.
-------------------------------------------------------------------------------------

Incorrect. Hucho does not address car shapes of the 2000s, especially notchbacks. None of his notchbacks look like today's cars.

1) I could not find a single instance in Hucho's 2nd-Edition where a spoiler wasn't in separated flow.

Page 173, 272 of 2nd edition. The 5th edition has an extensive section on rear spoilers, including showing via streamlines those in attached flow (page 311). Your understanding of how rear spoilers work on cars of the last 30 years is completely wrong.

2) In his section on high-performance sports cars he recommended zero lift, or if you were going to introduce negative lift, then provide rear bias.

Correct

3) Schenkel's rear spoiler introduced front lift where none had existed.
4) beyond 17.8 mm height, Schenkel's spoiler increased drag, at the expense of downforce.

This refers to really old shape cars, no?

------------------------------------------------------------------------------------
Longitudinal centerline pressures on the upper body of a vehicle are meaningless without the underbody pressure profile, especially if there's an active diffuser.
--------------------------------------------------------------------------------------

Partly correct, but overstated. They're not 'meaningless', they're just not the full story.

Any information derived from tuft testing is extremely problematic. They clearly misrepresent the flow on the Beetle, New Beetle, and 911.

When real world tuft tests don't match your theories, I bet you find them 'extremely problematic'!