Aerodynamics - why wings create lift - current vs historical discussions

[cjl]

Had a read through the "arguments " here re air flow and wing lift. Not being an expert on such in depth matters, but I would like to query the "downwards" force explanation versus the pressure differential( if that how I am understanding the issues. If it was simple matter of "downward" force why would a layer of ice (specially on leading edges)on wings reduce/effect wing lift to the extent of causing aircraft to lose lift and crash(refer to "Air crash investigations)?.

To be clear, wings do direct air downwards. You can entirely explain the lift from a wing through downwash. However, the air deflection actually mostly occurs above the wing, not below it, and the upper surface is more critical to both maintaining lift and keeping drag low. If you have ice, the flow is prone to separating from the upper surface, and this both slightly (or dramatically, depending on how far past stall you are) reduces lift and substantially increases drag.

Happened on a number of occasions Why would it be nesessary to deice wings? All one would need would be increase the "angle of attack" to force more air downwards.

It's worth noting that even if the airplane is able to generate enough lift through extra angle of attack, the drag rise is also a huge problem unless you have enough engine power to deal with the extra drag. A lightly stalled wing is actually making pretty similar lift to one just before stall, but it'll have many times the drag. This can lead to a vicious cycle too, since as the aircraft slows down due to the excessive drag, the required Cl climbs, and therefore the wing stalls more deeply.

Finally, just as a fun bit of extra trivia, many airfoils do produce peak lift at around 45 degrees, or at least many thinner, lower-camber airfoils do. They do stall at 10 or 12 degrees, but the lift climbs back up at 30 or 40 degrees and matches or exceeds the lift performance at 10 degrees. Of course, the drag is much, much higher, so overall efficiency is terrible, but purely from a lift standpoint, it's better than you might think. Here's an example lift polar for a NACA 0012: https://i.stack.imgur.com/jpvEl.jpg.

I also note that excessive "angle of attack" will cause loss of lift.

Yes, and this is because as I mentioned above, most of the flow redirection happens above the wing rather than below it. When you reach an angle such that the flow over the top is unable to continue to follow the curvature, the flow separates and the flow over the top stops being directed downwards, losing you much of your lift.

I would expect there to be some element of both involved since aircraft use flaps when taking off or landing and these seem to direct airflow directly(as per the hand out the car window example)

Again, it's not either-or. The bernoulli relation holds everywhere around a wing in incompressible flow (below about mach 0.3), and even in compressible flow, it holds with some minor modifications (as long as you stay subsonic). If you cover a wing in pressure transducers, you'll see that 100% of the lift can be explained through differences in pressure around the wing.

In addition, you can measure downwash in the wake of the wing. If you have a good way to measure downwash quantity and velocity, you can also find that 100% of the wing's lift is explained through the amount of air it is pumping downwards as it travels through the air.

Flaps don't counteract this either. They provide a way to enlarge areas of low pressure above the wing and dramatically increase cl_max, which also increases downwash. Even a hand out a window probably makes a better lift coefficient than you might think, and even in the case of a flat plate, most of that lift does come from the suction side of the airfoil rather than the pressure side.

 

[russ_watters]

The common (and entirely wrong) "bernoulli' description of lift goes as follows:

1) The top of a wing is longer than the bottom of the wing due to the airfoil shape
2) The air going around the wing will all arrive at the back simultaneously (often with some handwaving about "continuity" or something)
3) The air going over the top must travel farther, therefore it must travel faster
4) Therefore, by the Bernoulli relation, the pressure is lower above the wing, creating lift

This makes a nice "just-so" explanation of lift, but it's actually entirely wrong. Specifically, step 2 is where it completely falls apart. There's no reason to expect the flow around both sides to arrive at the back of the wing simultaneously, and in fact, it usually doesn't.

Fair enough. For my part, I don't like saying something is "entirely wrong", when it contains multiple points and one at least (the last one) looks right as stated. I don't generally subscribe to the idea that one a line of logic fails everything else associated with it (or at least after it) must be wrong. In this case it gives the false impression that Bernoulli is of no help when describing lift. That's exactly the problem that led to the creation of the thread.

 

[sophiecentaur]

Fair enough. For my part, I don't like saying something is "entirely wrong",

Point two is the weak link, surely. Where can you define the effective trailing edge of the wing? This actual point only needs to be slightly below the physical trailing edge and things are more believable. Is the argument that laminar flow has to be occurring?

 

[cjl]

Point 2 is entirely wrong, regardless of where you define the "effective" trailing edge. The air passing above the wing will arrive at the back entirely misaligned with air that passed beneath. It also doesn't matter whether you're talking about a laminar or turbulent boundary layer - the air outside the boundary layer will still arrive misaligned.

 

[A.T.]

The version of the "downward force" explanation that I get from your post sounds like the "air bounces off wing" explanation that everyone agrees is incorrect.

It is correct that all the momentum transferred by the wing to the air (and vice versa) is accounted for by the direct collisions between the wing and air molecules.

Whether this qualifies as an explanation is not a matter of correct or incorrect, but of the expectations one has of an explanation.

 

[sophiecentaur]

Point 2 is entirely wrong, regardless of where you define the "effective" trailing edge. The air passing above the wing will arrive at the back entirely misaligned with air that passed beneath. It also doesn't matter whether you're talking about a laminar or turbulent boundary layer - the air outside the boundary layer will still arrive misaligned.

I an believe that 2 is just wrong but it's certainly what the elementary diagrams seem to tell us. I guess the diagrams all imply that there should be continuity of 'something' across the boundary between upper and lower airstreams. But there's nothing to say that the neighbouring regions of air where the streams come together should be the same adjacent regions when the streams come together. So there is no requirement for speed over the top to be higher than underneath - in the steady state - just different path lengths.

What would need to be continuous? Presumably the velocity profile across the boundary. But the vertical component of velocity would be greater over an aerofoil because the air has to move further up and down as the wing passes through it. That would cause a lower pressure on the top.

Explanations all seem to use the Wind Tunnel frame but a flying wing is going through stationary air and it must be pushing air forward causing drag and downward causing lift. The air way out in front and well to the rear is stationary. (I know it's just a choice of reference frame but some frames make more immediate sense than others, at times.

 

[profbuxton]

if a wing just pushes air forward(drag) and downward for lift, why not just use a flat wing,no aero foil needed. Note that wings have different aero foil shapes depending on application and speed. Also it appears to me that the shape of the leading edge in fairly critical for generating lift.
If a wing just relied on thrusting air downward surely it wouldn't matter what shape the leading edge was as long it didnt increase drag by excessive thickness.

 

[cjl]

I an believe that 2 is just wrong but it's certainly what the elementary diagrams seem to tell us. I guess the diagrams all imply that there should be continuity of 'something' across the boundary between upper and lower airstreams. But there's nothing to say that the neighbouring regions of air where the streams come together should be the same adjacent regions when the streams come together. So there is no requirement for speed over the top to be higher than underneath - in the steady state - just different path lengths.

This is one area where it can be misleading to rely too much on basic diagrams. You're right that there's no requirement for them to come together at the same spot, and a more accurate diagram would show that the air over the upper surface actually outruns the air beneath, despite the longer distance (so an "equal transit time" calculation would severely underpredict the lift). You can see this effect in this video, especially around 40-60 seconds.

What would need to be continuous? Presumably the velocity profile across the boundary. But the vertical component of velocity would be greater over an aerofoil because the air has to move further up and down as the wing passes through it. That would cause a lower pressure on the top.

There's actually not even a requirement that velocity be continuous, though in practice it will be because the pressure does have to match, and outside of the viscous dissipation region, there's no mechanism for a mismatch of velocity that wouldn't also result in a mismatch of pressure. However, behind a supersonic airfoil, you can get a so-called "slip line", where the velocity does have a mismatch off the top vs bottom surface of the airfoil. All that is actually required is that the flow coming off the back has to be parallel (flow off the top and bottom surface can't diverge or converge), and there can't be a vertical pressure gradient, so pressure just behind the trailing edge has to be the same both in flow from the bottom and top.

Explanations all seem to use the Wind Tunnel frame but a flying wing is going through stationary air and it must be pushing air forward causing drag and downward causing lift. The air way out in front and well to the rear is stationary. (I know it's just a choice of reference frame but some frames make more immediate sense than others, at times.

Physically, there's no difference between the two frames of course. The misalignment behind the wing in the wind tunnel frame corresponds to the wing actually pushing some of the air traveling over the suction side in the opposite direction as the flight direction, while dragging some air on the pressure side with it. So, after the airplane passes, some of the air that traveled above the wing is displaced slightly backwards and some of the air below the wing is displaced slightly forwards, even though you're right that they both end up stationary.

 

[cjl]

if a wing just pushes air forward(drag) and downward for lift, why not just use a flat wing,no aero foil needed. Note that wings have different aero foil shapes depending on application and speed. Also it appears to me that the shape of the leading edge in fairly critical for generating lift.
If a wing just relied on thrusting air downward surely it wouldn't matter what shape the leading edge was as long it didnt increase drag by excessive thickness.

Because a cambered, nonzero-thickness wing is better at redirecting flow downwards with less drag. As I've been saying, most of this redirection actually happens above the wing, so the way the flow behaves on the suction side is very important. A flat plate doesn't do a good job creating a nice large low pressure region on the suction side to redirect flow downwards, and it also has a bit of a tendency to stall at a significantly lower angle of attack than an airfoil, both of which make it a poor choice.

 

[russ_watters]

It's worth noting that even if the airplane is able to generate enough lift through extra angle of attack, the drag rise is also a huge problem unless you have enough engine power to deal with the extra drag. A lightly stalled wing is actually making pretty similar lift to one just before stall, but it'll have many times the drag. This can lead to a vicious cycle too, since as the aircraft slows down due to the excessive drag, the required Cl climbs, and therefore the wing stalls more deeply.

Another angle ( ) to the vicious cycle is that when the plane that is flying level stalls, it starts descending fairly rapidly, and that causes the relative wind to rotate down, further increasing the angle of attack and deepening the stall, unless the pitch-angle is rapidly lowered. Stalls can be pretty violent, and that cascade/cycle happens rapidly.

 

[russ_watters]

Point two is the weak link, surely. Where can you define the effective trailing edge of the wing? This actual point only needs to be slightly below the physical trailing edge and things are more believable. Is the argument that laminar flow has to be occurring?

I'm not sure I understand what you mean. The trailing edge is always in exactly the same place and is the most easily identifiable point in the flow. The whole point of making the trailing edge of the wing sharp is to enforce that that's the trailing stagnation point.

This is unlike the leading-edge stagnation point, which moves up and down as the pitch changes, and isn't even aligned with the freestream direction of flow (it's below it).

 

[russ_watters]

Because a cambered, nonzero-thickness wing is better at redirecting flow downwards with less drag. As I've been saying, most of this redirection actually happens above the wing, so the way the flow behaves on the suction side is very important. A flat plate doesn't do a good job creating a nice large low pressure region on the suction side to redirect flow downwards, and it also has a bit of a tendency to stall at a significantly lower angle of attack than an airfoil, both of which make it a poor choice.

Other side of the...coin... a sail (zero thickness, large camber) doesn't do a good job of creating a good high pressure region under the wing. I think it was you who pointed out that that was one of the first things the Wright brothers realized in studying their kites and testing airfoils in a wind tunnel.

 

[cjl]

That wasn't me, but it does provide an interesting case study. I would suspect the reason it doesn't work as well as a blunt nose is largely because at high angles of attack, the leading stagnation point actually moves around underneath the airfoil, so you get flow wrapping around the nose of the airfoil from beneath. If you have a large leading edge radius, that allows the flow to stay attached while this happens, allowing for better lift production at high angles of attack. With a cambered but very thin wing, the stagnation point really can't migrate beneath the wing, since the flow will just separate at the thin leading edge, and that causes it to be ineffective at creating high lift at high angles of attack.

(if it's unclear what I'm talking about, here's a diagram)

 

[Lnewqban]

I always thought de-icing was to remove weight from the airplane. I have been wrong about many things, maybe this is another?

 

[rcgldr]

leading stagnation point - separation of flow

Regardless of the airfoil shape, the separation point is below the leading edge of an airfoil making lift because the lower pressure above draws the air from in front and below the leading edge backwards and upwards to flow over the surface.

icing

Icing tends to fill in the reduced pressure zone on the upper surface, effectively changing the airfoil into one that produces less lift at the same AOA, requiring more AOA and associated drag to produce the same lift. In an extreme case, the wing may reach a stall state before it produces sufficient lift.

flat or thin wings

These work fine at low Reynolds numbers (small wing chord and slow speed). A small balsa glider with about a 30 inch wingspan glides reasonably well with barely any camber.

http://www.ericbrasseur.org/glider2.html

a more accurate diagram would show that the air over the upper surface actually outruns the air beneath, despite the longer distance

The whole point of making the trailing edge of the wing sharp is to enforce that that's the trailing stagnation point.

There are cases where the longer distance is on the bottom, and where the trailing edge isn't sharp. For example, the M2-F2 (glider) lifting body prototype re-entry vehicle. The trailing edge is not sharp because that is where the rocket engines were place in the later M2-F3.

 

[russ_watters]

There are cases where the longer distance is on the bottom, and where the trailing edge isn't sharp. For example, the M2-F2 (glider) lifting body prototype re-entry vehicle. The trailing edge is not sharp because that is where the rocket engines were place in the later M2-F3.

View attachment 259020

You've posted that claim and that image several times before, without evidence that your claim is true -- the picture isn't evidence because you can't see the flow. I'm not inclined to accept your claim, without evidence that a lifting body works any different from a normal wing. Specifically, I bet the stagnation point is far under the chin and the airflow does indeed take a longer path over the top than over the bottom. The video of it landing at very high AOA compared to the F-104 flying next to it implies that should be true.

Further, even if true I think it is unhelpful to use an outlier when trying to teach the basics. It's a great way to create wrong impressions of what is typical or best. I'm sure a brick generates lift at a small positive angle of attack, but we shouldn't be generalizing that or listing it as a noteworthy exception.

 

[cjl]

Regardless of the airfoil shape, the separation point is below the leading edge of an airfoil making lift because the lower pressure above draws the air from in front and below the leading edge backwards and upwards to flow over the surface.

It's not clear to me that this is always true, and at low Cl (such as in cruise), the stagnation point is effectively at the leading edge on pretty much any airfoil. Certainly any reasonable airfoil operating at higher Cl though will have the stagnation point well below the leading edge.

Icing tends to fill in the reduced pressure zone on the upper surface, effectively changing the airfoil into one that produces less lift at the same AOA, requiring more AOA and associated drag to produce the same lift. In an extreme case, the wing may reach a stall state before it produces sufficient lift.

The increased surface roughness, and associated higher drag and earlier stall angle is a much larger problem than gross airfoil shape changes. In general, adding to the top surface of the airfoil won't decrease the achievable Cl_max (unless you get to some fairly large changes there) - that tends to be more dependent on leading edge radius instead (sharply curved leading edges lead to lower Cl_max).

These work fine at low Reynolds numbers (small wing chord and slow speed). A small balsa glider with about a 30 inch wingspan glides reasonably well with barely any camber.

http://www.ericbrasseur.org/glider2.html

It's not just the low reynolds number helping him there, it's also the low angle of attack and required Cl. You're right that flat airfoils do perfectly fine as you go down to low reynolds numbers though (or more accurately, "traditional" airfoils do really badly - you'd be hard pressed to get a flat plate above a L/D of 10 or so, while 100:1 is pretty readily achievable with a more normal airfoil, but only at Re>10^5 or so)

There are cases where the longer distance is on the bottom, and where the trailing edge isn't sharp. For example, the M2-F2 (glider) lifting body prototype re-entry vehicle. The trailing edge is not sharp because that is where the rocket engines were place in the later M2-F3.

View attachment 259020

For the purposes of my statement, a flatback and a sharp trailing edge both achieve the same thing: they prevent flow from wrapping around and effectively enforce a rear stagnation point. There are some additional interesting effects when it comes to flatback airfoils that make them a really useful choice for very high thickness airfoils, especially when you also consider structural properties of your airfoil section on top of just the aerodynamic effects.

Also, it's not necessarily useful to use your M2-F3 example as an airfoil at all. It doesn't behave like an airfoil, because the flow isn't even close to 2D. I would bet that a large part of the lift on that is actually related to vortices around the sides onto the upper surface, similar to a delta wing at high AoA, and there are so many 3d effects impacting the flow that I don't even know that you could meaningfully talk about "path lengths" for the flow except right on the centerline.

 

[A.T.]

...the airflow does indeed take a longer path over the top than over the bottom.

Is this key to generating lift?

 

[FactChecker]

EDIT: Although the statements in this post are reasonable, I think it is safe to say that having a longer airflow over the top of the wing is essential to obtaining lift.

Is this key to generating lift?

There are multiple things adding to the total lift and they are all inter-related. It is hard to know how much lift to credit to each cause. I would not call any single cause "key". The total lift of a wing has a very messy explanation when you look critically at it. That is why CFD calculations are used.
The only exception is Newton's Third Law, where the lift of the wing is equal and opposite to the total downward force on the air. But that is a bottom-line number with messy intermediate details that again require CFD. It is almost a platitude.

 

[russ_watters]

Is this key to generating lift?

It's probably not a coincidence.

 

[rcgldr]

Regardless of the airfoil shape, the separation point is below the leading edge of an airfoil making lift because the lower pressure above draws the air from in front and below the leading edge backwards and upwards to flow over the surface.

It's not clear to me that this is always true, and at low Cl (such as in cruise), the stagnation point is effectively at the leading edge on pretty much any airfoil. Certainly any reasonable airfoil operating at higher Cl though will have the stagnation point well below the leading edge.

I was referring to the separation of flow point, which is well ahead of the stagnation point and below if the wing is generating lift.

m2-f2

Here's a better image, in this case the M2-F3. The AOA is high when landing, but this was a re-entry prototype that reached a max speed of mach 1.6 during tests. It does need a high AOA for landing, but at speeds around 400 knots, AOA and lift to drag ratio would be reasonable, probably similar to the Space Shuttle. In the case of the Space Shuttle, lift to drag is 1:1 at hypersonic speeds, 2:1 at supersonic speeds, 4.5:1 at sub-sonic speeds. These are somewhat low since the reentry to landing time is already 20 minutes, and reentry to landing distance is 5,000 miles. Two big turns in the shape of an S are done to scrub off speed and stay within reach of the landing zone.

 

[cjl]

I was referring to the separation of flow point, which is well ahead of the stagnation point and below if the wing is generating lift.

That isn't a meaningful distinction. The streamline that contains the stagnation point is also (by definition) the "flow separation" line. You're right that (at least for incompressible flow) this streamline will always be upward sloping for a 2D flowfield around an object making lift, so some air that starts below the wing will end up traveling above it.

I would kind of expect there to also always be some upwash in front of something like your M2-F3 example during flight in the incompressible regime, despite the much more complicated 3d flowfield. I just don't want to definitively make a statement that you couldn't have some weird 3d effects that would allow for lift generation without leading upwash. I can't envision how that would work off the top of my head though.

Also, just as an interesting counterexample (though this is way outside of the discussion above), in supersonic flight, the incoming flow will just split exactly where the leading edge is. Anything that starts above the wing stays above, and anything that starts below stays below. This is of course intuitively obvious though, because disturbances can't propagate forwards in a supersonic flow.

 

[russ_watters]

There are cases where the longer distance is on the bottom...

You haven't responded to my request for substantiation with regard to the lifting body, but here's evidence against, using the common example of an inverted airfoil:

As you can see, the stagnation point in both cases is underneath the "chin" thus the air flowing over the top does indeed follow a longer path/gets deflected more than the air flowing over the bottom. In both cases the flow just above the stagnation streamline curves up a lot as it approaches the wing, to flow over the top. And flow under doesn't have as far to bend down to flow over the bottom surface.

...with deference to @cjl 's point that since the shape is very non-uniform spanwise, the flow is going to be very 3 dimensional for a lifting body.

 

[rcgldr]

I was referring to the separation of flow point, which is well ahead of the stagnation point and below if the wing is generating lift.

That isn't a meaningful distinction. The streamline that contains the stagnation point is also (by definition) the "flow separation" line. You're right that (at least for incompressible flow) this streamline will always be upward sloping for a 2D flowfield around an object making lift, so some air that starts below the wing will end up traveling above it.

I should find a better term, since separation of flow normally refers to the detachment of flow over the upper surface of a wing. As for a description, the free stream flow well in front of a wing is horizontal, but the flow approaches the wing, the flow initially curves upwards, but then a bit closer to the leading edge, the streamlines separate into flows that will go over the wing and flows that go below the wing. Both of these transition points in the flow occur ahead of and below the leading edge.

As for the M2-F2 or M2-F3 at sub-sonic but fairly high speed (400 knots), the ratio of ambient pressure / pressure above the lifting body is probably less than the ratio of pressure just below the lifting body / ambient pressure, in otherwords, the pressure below deviates more from ambient than pressure above. This corresponds to a reduced lift to drag ratio (probably around the Space Shuttles 4.5 to 1), but as posted early, the Shuttle takes 20 minutes and 5000 miles from re-entry to landing, so a better lift to drag ratio isn't a goal here. One issue is that both the M2-F2 / M2-F3 / Shuttle have high landing speeds and need long runways to land.

 

[Aeronautic Freek]

Summary:: What is the real cause of lift, said to be true by current aerodynamics

My son and i were discussing aerodynamics and he brought up a paper from https://phys.org/news/2012-01-wings.html It seems that the latest discussions seem to completely discount the differential velocity of air flow as a cause of differential pressure, but point to a differential pressure (pressure gradient) caused by the force acting on the air being accelerated to cause the differential pressure. in other words, some of these papers and i don't know if they are outliers, are insinuating that it is the differential pressure causing the speed variance above and below an airfoil, and not the other way around.
thoughts??

Doug Mclean write one book where he only explain all these tons of false theories and missconceptions in aerodynamics.
Title of book is :Understading Aerodynamics : Arguing from Real Physics..
equal time theory
"bernulli" explanation
"Newtonian" explanation
coanda examples
wingtip vortices cause induced drag
etc
etc
etc
etc
etc
etceven in Aderson books ,there are some missconceptions..(water jet at spoon-test is not example of Coanda effect...)

 

[sophiecentaur]

all these tons of false theorys

OMG. So they're all wrong? Which is the right alternative?

PS I just watched the video. If our local experts in Maths can OK it then it seems that his explanation is fair enough. 'No downward Momentum' was a bit of a surprise for me.

Question is whether the pressure at the ground is high enough to measure for a plane at normal height.

 

[Aeronautic Freek]

OMG. So they're all wrong? Which is the right alternative?

PS I just watched the video. If our local experts in Maths can OK it then it seems that his explanation is fair enough. 'No downward Momentum' was a bit of a surprise for me.

Question is whether the pressure at the ground is high enough to measure for a plane at normal height.

Did you ask Math experts?

 

[sophiecentaur]

Did you ask Math experts?

I meant Maths Experts local to PF. Some PF members are no slouch when it comes to Maths.

 

[A.T.]

'No downward Momentum' was a bit of a surprise for me.

He says that "there is no downward momentum accumulated in the atmosphere" (at 45:10). This should not be surprising. How should the atmosphere as a whole accumulate downward momentum, with the ground in the way?

 

[sophiecentaur]

He says that "there is no downward momentum accumulated in the atmosphere" (at 45:10). This should not be surprising. How should the atmosphere as a whole accumulate downward momentum, with the ground in the way?

Ahh. Like a hamster on a hamster wheel when the wheel is accelerating. So it’s angular momentum?

 

[A.T.]

Ahh. Like a hamster on a hamster wheel when the wheel is accelerating. So it’s angular momentum?

No idea what you mean here.

 

[sophiecentaur]

No idea what you mean here.

The hamster imparts no linear momentum to the wheel but friction helps transfer linear momentum to the ground which allows the hamster to climb up a bit (angular momentum). Short term the hamster can derive ‘lift’ as it accelerates up the slope.
I can see an analogy.

 

[A.T.]

The hamster imparts no linear momentum to the wheel but friction helps transfer linear momentum to the ground which allows the hamster to climb up a bit (angular momentum). Short term the hamster can derive ‘lift’ as it accelerates up the slope.
I can see an analogy.

But a plane in level flight creates symmetrical counter rotating vertices, so their net angular momentum is zero.

 

[sophiecentaur]

But a plane in level flight creates symmetrical counter rotating vertices, so their net angular momentum is zero.

How is that relevant? The hamster could use two counter rotating wheels and my model would work. Also, is there no friction involved in the atmosphere?

 

[A.T.]

The hamster could use two counter rotating wheels and my model would work.

Then there would be no net angular momentum, so what was your point in bringing angular momentum up?