Reducing lift on different body shape cars

[JulianEdgar]

Originally Posted by AeroMcAeroFace

Surely aerohead, it would be better to argue from the point of drag, downwash induced attached flow will be almost always higher drag than if the flow were "real" attached flow. Or reattached flow is almost always higher drag than if it was attached flow.

Arguing that the flow isn't attached and the tufts are lying and every expert is wrong seems difficult to believe. Saying that you have a different definition of attached flow seems far more reasonable to me.

You don't even need a different definition of attached flow.

Attached flow: airflow is guided by shape of the car, and tufts line up on body in direction of flow

Separated flow: airflow is no longer guided by the shape of the car, and tufts whirl around and point in any old direction

Whether airflow is attached or separated is governed by a range of factors, including boundary layer thickness (thicker = more easily separates), rapidity of angle change (quicker downwards angle change = more easily separates), surface finish (eg steps up or down = more easily separates) and vortex behaviour (downwash can cause attachment that otherwise wouldn't occur).

If we then want to talk about pressure drag, we can talk about the drag induced by attached flow on inclined rear surfaces (because this flow will also be causing lift, you can call it induced drag, or simply separate the force vector into its lift and drag components), and we can talk about drag caused by the low pressures of the action of the vortices that are causing the downwash.

Unfortunately, Aerohead says stuff like this:

Downwash is not attached flow. It is symptomatic of flow separation. You'll get the highest drag, vortex-drag, plus a small, but very low pressure wake, and overall higher drag. To say otherwise is delusional. It may look like a duck, but it doesn't quack like a duck.

...which is full of confusion and, yes in his words, rather delusional.

For those with my big book, on Page 129 there is direct quote from the head of Jaguar aero that makes it clear that attached flow caused by vortices is still attached flow.

[AeroMcAeroFace]

Downwash is not attached flow. It is symptomatic of flow separation. You'll get the highest drag, vortex-drag, plus a small, but very low pressure wake, and overall higher drag. To say otherwise is delusional. It may look like a duck, but it doesn't quack like a duck.

"Attached flow caused by vortices is still attached flow." I can't believe this is even contested or debated, the only place I have ever seen it said that it isn't is in aerohead posts.

Out of curiosity I searched ecomodder for attached flow and found a quote, basically the same debate

Originally Posted by aerohead

Disagreement is healthy and I always learn.

Later saying...

With respect to the 30-degree fastback,flow will be attached along the centerline only,with attached longitudinal vortices on both sides.

So it seems in 2011, on a fastback the flow will be attached along the centreline, but also with attached vortices on both sides. Sounds pretty attached to me.

The same thread also discusses this chapter, https://books.google.co.uk/books?id=XEGZIZ5zhy4C&lpg=PP1&pg=PA115&redir_esc=y #v=onepage&q&f=false

aerohead said in that thread: "Once the tangent angle on the roofline exceeds 22-degrees(as measured against the horizon)the flow cannot remain attached." This picture (and many many others) says different (page 128)

[Vman455]

Originally Posted by AeroMcAeroFace

aerohead said in that thread: "Once the tangent angle on the roofline exceeds 22-degrees(as measured against the horizon)the flow cannot remain attached."

This 22-degree angle comes from Mair's work, as has been correctly stated on this site. But here's the thing: Mair was investigating flow over very long cylindrical bodies (14 times as long as their diameter) and very short, tapered tails (half as long as the diameter) to try and understand the mechanisms of flow attachment and the effect of taper on drag; these are bodies that look more like ICBMs than cars. He also reported on other researchers' work with the same types of cylindrical bodies who found drag minima at different trailing edge angles, as well as minima in side forces in yawed flow at different taper angles, and there are modern articles reporting on work with similar cylindrical bodies that find drag minima at differing taper angles depending on things like the transition between body and tail and other upstream features. 22-degrees was never presented as a hard-and-fast, "beyond this angle airflow detaches" rule because it isn't (except on this site, of course). And every aerodynamic textbook of which I'm aware gives a range of angles for rear taper within which minimum drag can be found. I'm not at home and don't have my books in front of me, but I believe Hucho even uses the word "broad" to describe the range of acceptable backlight angles for low drag, and Barnard writes about the effective angle between backlight upper edge and trunklid trailing edge, saying it should generally be between 10- and 30-degrees for lowest drag--a very wide range!

Contrary to what I believed in the past, the ideal angle for any car has to be found through testing. Depending on the body shape and details, an angle faster or slower than 22-degrees may be allowable or necessary to maintain flow attachment in the widest range of conditions. If you find yourself with a 40- or 50-degree angle, then there's a good chance the attached flow is due to vortex downwash, but the resulting high drag will show up in proper testing (that's how this phenomenon was discovered in the first place).

[JulianEdgar]

Originally Posted by Vman455

This 22-degree angle comes from Mair's work, as has been correctly stated on this site. But here's the thing: Mair was investigating flow over very long cylindrical bodies (14 times as long as their diameter) and very short, tapered tails (half as long as the diameter) to try and understand the mechanisms of flow attachment and the effect of taper on drag; these are bodies that look more like ICBMs than cars. He also reported on other researchers' work with the same types of cylindrical bodies who found drag minima at different trailing edge angles, as well as minima in side forces in yawed flow at different taper angles, and there are modern articles reporting on work with similar cylindrical bodies that find drag minima at differing taper angles depending on things like the transition between body and tail and other upstream features. 22-degrees was never presented as a hard-and-fast, "beyond this angle airflow detaches" rule because it isn't (except on this site, of course). And every aerodynamic textbook of which I'm aware gives a range of angles for rear taper within which minimum drag can be found. I'm not at home and don't have my books in front of me, but I believe Hucho even uses the word "broad" to describe the range of acceptable backlight angles for low drag, and Barnard writes about the effective angle between backlight upper edge and trunklid trailing edge, saying it should generally be between 10- and 30-degrees for lowest drag--a very wide range!

Contrary to what I believed in the past, the ideal angle for any car has to be found through testing. Depending on the body shape and details, an angle faster or slower than 22-degrees may be allowable or necessary to maintain flow attachment in the widest range of conditions. If you find yourself with a 40- or 50-degree angle, then there's a good chance the attached flow is due to vortex downwash, but the resulting high drag will show up in proper testing (that's how this phenomenon was discovered in the first place).

A well-supported summary. To put it another way:

The use of a template to guide rear extensions, judge the 'purity' of shapes, guide the positioning of rear spoilers on notchbacks, and assess whether a car will have attached or separated flows - they're all just rubbish.

I'm (again) putting it like that because I've noticed that lots of people here can't see the cognitive dissonance of agreeing with two completely contradictory ideas.

At least as a group we now seem to be moving away from invalid 'rules of thumb' to acknowledging that the real-world complexity of car aero modification requires testing, not guessing.

[aerohead]

Originally Posted by Cd

How does downwash appear different from the surrounding tufts ?
Do tufts in a downwash remain steady, or do they flutter in any way different from areas of the car that are not in downwash ?

What clues are there that certain tufts are in downwash ?

1) 'Appearance' is the main talking point.
2) You'd be hard-pressed to make the distinction without smoke.
3) A fatal trap!
--------------------------------------------------------------------------------------
Here's a reverse construction of the phenomena:
A) The tufts are held in place within a downwash.
B) The downwash is induced by attached, longitudinal vortices.
C) The attached longitudinal vortices are symptomatic of flow separation.
D) The flow separation is symptomatic of exceeding the threshold magnitude of adverse pressure gradient, responsible for triggering separation of the aft-body flow.
E) The trigger is provided by too radical a constriction of body cross-section and the accompanying pressure spike, intolerable to the TBL.
F) The body shape itself is the 'trigger.'

[aerohead]

Originally Posted by AeroMcAeroFace

Surely aerohead, it would be better to argue from the point of drag, downwash induced attached flow will be almost always higher drag than if the flow were "real" attached flow. Or reattached flow is almost always higher drag than if it was attached flow.

Arguing that the flow isn't attached and the tufts are lying and every expert is wrong seems difficult to believe. Saying that you have a different definition of attached flow seems far more reasonable to me.

* Exactly! Which Hucho succeeds with.
* Vortex-drag is of the highest drag.
* A too-radical profile may produce a small wake due to downwash, however, the wake itself is of very low pressure, creating a low base pressure, high pressure drag, and high overall drag, in addition to the long-lived, drag-producing vortices themselves, which cannot possibly offer any constructive pressure regain.
-------------------------------------------------------------------------------------
If the flow were actually 'attached' the TBL would be intact. The local streamline just outside the TBL would be imparting higher and higher pressure, the further rearwards on the body, all the way to the trailing edges.
If separation occurs, the pressure of that above the separation line, is transmitted throughout the entire region of separated flow. All the way into the wake unless there's a physical boundary to prevent it.
This is part of the reason a Mitsubishi Mirage 'G' model is Cd 0.27, whereas, a Porsche Macan is Cd 0.37. The mutilation of the Porsche roofline is robbing the ability of pressure regain over the segment cut away from the roofline. ( Exactly what Kamm studied at his FKFS).
There's a reason why the professional language differentiates between ' attached' and 'downwash' flow. They are completely different animals.

[aerohead]

Originally Posted by JulianEdgar

You don't even need a different definition of attached flow.

Attached flow: airflow is guided by shape of the car, and tufts line up on body in direction of flow

Separated flow: airflow is no longer guided by the shape of the car, and tufts whirl around and point in any old direction

Whether airflow is attached or separated is governed by a range of factors, including boundary layer thickness (thicker = more easily separates), rapidity of angle change (quicker downwards angle change = more easily separates), surface finish (eg steps up or down = more easily separates) and vortex behaviour (downwash can cause attachment that otherwise wouldn't occur).

If we then want to talk about pressure drag, we can talk about the drag induced by attached flow on inclined rear surfaces (because this flow will also be causing lift, you can call it induced drag, or simply separate the force vector into its lift and drag components), and we can talk about drag caused by the low pressures of the action of the vortices that are causing the downwash.

Unfortunately, Aerohead says stuff like this:

Downwash is not attached flow. It is symptomatic of flow separation. You'll get the highest drag, vortex-drag, plus a small, but very low pressure wake, and overall higher drag. To say otherwise is delusional. It may look like a duck, but it doesn't quack like a duck.

...which is full of confusion and, yes in his words, rather delusional.

For those with my big book, on Page 129 there is direct quote from the head of Jaguar aero that makes it clear that attached flow caused by vortices is still attached flow.

Would you like to qualify that assertion?

[aerohead]

Originally Posted by Vman455

This 22-degree angle comes from Mair's work, as has been correctly stated on this site. But here's the thing: Mair was investigating flow over very long cylindrical bodies (14 times as long as their diameter) and very short, tapered tails (half as long as the diameter) to try and understand the mechanisms of flow attachment and the effect of taper on drag; these are bodies that look more like ICBMs than cars. He also reported on other researchers' work with the same types of cylindrical bodies who found drag minima at different trailing edge angles, as well as minima in side forces in yawed flow at different taper angles, and there are modern articles reporting on work with similar cylindrical bodies that find drag minima at differing taper angles depending on things like the transition between body and tail and other upstream features. 22-degrees was never presented as a hard-and-fast, "beyond this angle airflow detaches" rule because it isn't (except on this site, of course). And every aerodynamic textbook of which I'm aware gives a range of angles for rear taper within which minimum drag can be found. I'm not at home and don't have my books in front of me, but I believe Hucho even uses the word "broad" to describe the range of acceptable backlight angles for low drag, and Barnard writes about the effective angle between backlight upper edge and trunklid trailing edge, saying it should generally be between 10- and 30-degrees for lowest drag--a very wide range!

Contrary to what I believed in the past, the ideal angle for any car has to be found through testing. Depending on the body shape and details, an angle faster or slower than 22-degrees may be allowable or necessary to maintain flow attachment in the widest range of conditions. If you find yourself with a 40- or 50-degree angle, then there's a good chance the attached flow is due to vortex downwash, but the resulting high drag will show up in proper testing (that's how this phenomenon was discovered in the first place).

1) Mair's model consisted of an elliptical nose, of 1,3-D in length, A cylindrical main body of 3-D in length, and the boat-tailed section of 1.9-D, which creates a total length of 6.2-diameters.
2) All values reflect zero viscosity effects, only pressure drag. Nothing to do with boundary layer thickness.
3) The take-away for aero modders is the 'lead- in' profile, leading to the final downslope angle. It's a known quantity.
4) Without the boat-tail we're looking at Cd 0.204 for Mair's model with skin friction.
5) With the boat-tail, Cd 0.066.
6) This 'lead-in' profile basically exists on all PGA Regulation golf balls.
7) 'Compound-curvature' roof modifications can entail 800-man-hours of construction time. Some form of Go NoGo might save many hundreds of hours of fabrication time. This would make something like Mair's profile quite
valuable. It may not be perfect, but short of a few hundreds of thousands of dollars for laboratory R & D, it might be handy for some with a more meager budget.
8) This was the original premise for sharing it. And perhaps why Hucho and others have presented it.

[JulianEdgar]

Originally Posted by aerohead

Would you like to qualify that assertion?

Which one?

[aerohead]

Originally Posted by JulianEdgar

A well-supported summary. To put it another way:

The use of a template to guide rear extensions, judge the 'purity' of shapes, guide the positioning of rear spoilers on notchbacks, and assess whether a car will have attached or separated flows - they're all just rubbish.

I'm (again) putting it like that because I've noticed that lots of people here can't see the cognitive dissonance of agreeing with two completely contradictory ideas.

At least as a group we now seem to be moving away from invalid 'rules of thumb' to acknowledging that the real-world complexity of car aero modification requires testing, not guessing.

Are we to expect that your second-edition will have your rules-of-thumb edited out?