What causes the airflow separation from the airplane wing, that triggers a stall?

To be more precise, what is the detailed explanation of the physics for why flow separation occurs at certain speeds & angles of attack?

  • 8
    $\begingroup$ welcome to the site! To the downvoter: why the downvote? This is a perfectly valid question. If you think it requires improvement then how about helping a new user instead of alienating them? $\endgroup$ – Notts90 is off to codidact.org Jul 21 '19 at 10:37
  • 5
    $\begingroup$ @Notts90 100% agreed, too much anonymous downvoting to valid questions. $\endgroup$ – Koyovis Jul 21 '19 at 12:08
  • 1
    $\begingroup$ Just curious, but can one see all the up and downvotes? Because it appears there is only the net result showing. I personally always explain a downvote. Not sure in this case, but guessing any downvotes are because this is such a rudimentary concept. $\endgroup$ – Michael Hall Jul 21 '19 at 16:02
  • 1
    $\begingroup$ @MichaelHall Users with 1,000 rep or above (technically the "established user" privilege) can click on the net vote count to see the breakdown in up/downvotes. $\endgroup$ – a CVn Jul 21 '19 at 19:18
  • $\begingroup$ I think it is a more interesting question to ponder on why the smooth flow doesn't separate with an angle of attack. Why does it "turn the corner" and hang on at all? The magic of fluid flow... $\endgroup$ – MikeY Jul 22 '19 at 1:48

enter image description here

Air flowing underneath the wing at an angle is pushed downwards, regardless of the shape of the lower surface: high pressure has few practical limits.

Air flowing over the upper surface cannot suddenly change direction though because it is driven by an under-pressure gradient. It can follow a slowly curving surface to the limit of the pressure differential between ambient and suction, but too steep a curve and the airflow cannot follow and has to separate: a stall occurs.

enter image description here

Picture source

  • 10
    $\begingroup$ To the OP, if you are still unsure of this concept from the progression of figures above, take it to the extreme: Imagine a wing at 90 degrees AOA, the air simply cannot turn the corners fast enough to follow the curvature of the wing so there would be massive turbulence on the almost flat back side. Somewhere in between 0 and 90 this turbulence begins to have a critical effect on the efficiency of an airfoil. $\endgroup$ – Michael Hall Jul 21 '19 at 15:51
  • 1
    $\begingroup$ @Koyovis thanks for your gas dynamics link above. Clearly illustrated is the 0<M<1 case as compared to stationary. $\endgroup$ – Robert DiGiovanni Jul 22 '19 at 0:37
  • $\begingroup$ @RobertDiGiovanni the second link of @AEhere’s comment, the one pointing to the physics site, explains it best. $\endgroup$ – Koyovis Jul 22 '19 at 21:16
  • $\begingroup$ @Koyovis this activity has lead to 2 positive conclusions: 1. Oh, so that's where the Volume factor went in your P formula (It's held constant in incompressible modeling) 2. hydrofoils may have similar properties to airfoils at low speeds. Thanks! $\endgroup$ – Robert DiGiovanni Jul 22 '19 at 21:36
  • $\begingroup$ @RobertDiGiovanni No worries. Supersonic flow has much in common with water flow close to the surface: the water volume can be easily “ compressed” there by moving up in the air without much resistance. $\endgroup$ – Koyovis Jul 22 '19 at 21:51

The flow separation commences at the boundary layer, due to adverse pressure gradients (from Wiki):

adverse pressure gradient in a boundary layer

You can find a basic mathematical explanation in the linked article:

The streamwise momentum equation inside the boundary layer is approximately stated as

$$ {\partial u \over \partial s} = -{1 \over \rho}{dp \over ds} + {\nu} {\partial^2 u \over \partial y^2}$$ where $s,y$ are streamwise and normal coordinates. An adverse pressure gradient is when $dp/ds > 0$, which then can be seen to cause the velocity u to decrease along s and possibly go to zero if the adverse pressure gradient is strong enough.

In simpler terms, this means that the boundary layer on the upper surface of the wing is progressively slowed as it travels down the chord, until it cannot push against the higher pressure downstream. The air out of the layer can still advance, because it has higher momentum, but the bottom of the layer is forced to invert its direction, detaching from the surface.

As to where does this adverse pressure come from, it is because the accelerated air over the wing is at a lower pressure than the rest of the free airflow, so the air at the trailing edge pushes against the air over the wing (but cannot overcome its momentum).

Note that the flow is not guaranteed to fully separate at the point described above. This process can also force a laminar-turbulent transition, and a turbulent boundary layer can reattach.

  • $\begingroup$ But you forgot to add the link. And please cite the source also. Looks like MIT graphics. $\endgroup$ – Phil Sweet Jul 22 '19 at 20:27
  • $\begingroup$ @phil There is only one link in the article, but I´ll relink lower in the text too. The image is from wikipedia, maybe the MIT ripped their graphics from there? :) $\endgroup$ – AEhere supports Monica Jul 22 '19 at 20:49
  • $\begingroup$ Sorry, I didn't see it the first time. I spend most of my time on another SE that is seriously fussy about citing references. That sort of link doesn't pass muster and I tend not to notice them. $\endgroup$ – Phil Sweet Jul 22 '19 at 21:20
  • $\begingroup$ @PhilSweet no problem, proper sourcing is important, though I have been known to be lax on SE myself. It's a nice contrast to the multi-page reference lists on each 1-page technical note from work. $\endgroup$ – AEhere supports Monica Jul 22 '19 at 21:52

Jan's comment and AEhere's explanation are essentially the correct answer, but let me rephrase it in plainer terms of energy, without explicit math.

As flow is deflected downward by the wing, its inertia resists being redirected. The wing is sucking it down, and it exerts a reaction force on the wing--this is lift. In the process, an area of low pressure is created at the top of the curve, where the wing and the airflow are tugging at each other.

The airflow around the upper surface of the wing (never mind the bottom side, because it contributes only a small part of the lift at useful angles of attack) first accelerates, moving from higher to lower pressure and thus converting its pressure surplus into speed (pressure energy into kinetic energy to be exact).

Later along the curve, the flow passes the point of minimum pressure, and starts moving against adverse pressure gradient, using up its kinetic energy to compress itself and move up the pressure "hill".

But, in the boundary layer right next to the wing, some of the energy has been lost to skin friction. Thus there is a deficit of energy in the flow, and it is unable to recompress all the way; at some point its store of kinetic energy runs out, and it stops. It separates from the wing and is swept along a different path, towards lower pressure above and behind the wing, gradually regaining energy from the air surrounding the wake.

Meanwhile a bubble of whirling air is attached to the tail part of the wing surface, where the flow was unable to reach. The bubble, being attached to the wing, generates no lift in that "dead" area. Therefore wing lift starts reducing significantly once separation begins (strictly speaking, once it grows beyond a certain small size). Furthermore, conditions become highly unstable once separation starts, as the affected area changes a lot with even small fluctuations.

And why does separation only happens once you exceed some angle, and not before? Because the incoming airflow has a reserve of kinetic energy to begin with, which is at first sufficient to make it past the adverse pressure all the way to the end of the airfoil. It's when the losses and pressure gradients grow more severe with increasing angle, that separation eventually occurs.

  • 3
    $\begingroup$ +1 Good answer, welcome to the site. $\endgroup$ – Koyovis Jul 22 '19 at 11:43

"Airflow separating from the wing" is actually a very simplistic explanation that should not be interpereted literally. Air is a compressible gas, and until you get near near supersonic flight, it is best to understand it that way.

Lower (underneath) wing lift increases linearly with AOA all the way up to 45 degrees, but becomes hugely draggy.

Upper wing lift (from the airfoil shape) is much less draggy, and is what you lose when the wing "stalls".

Change in direction of airflow actually begins ahead of the wing and is vital to its lift creation. What happens is when AOA gets too high, the air flow over the top becomes turbulent and disorganized, losing lift efficiency, while drag continues to rise.

The solution is to lower the AOA. Watching the many available wind and smoke tunnel films may be greatly helpful.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ – Federico Jul 23 '19 at 5:18

The answer it's about the DECELERATION of the airflow over the wings, due to the Increased AOA above the stall margins, then separation takes place from the wing roots and closer to the trailing edges and moves across the wings and towards leading edge and spread outward until reaching the wing tip (this is done by aeronautical engineers to preserve the ailerons efficiency little time after the wing started to stall) , at any speed . There are many videos on Internet that have wings with " taft's" on top on them for visualization of turbulence. At this highest level of the AOA (17-30 degrees depending on the wing profile) the first layer of the airflow is deflected from the wings surface because it is not energized enough to follow the upper side and unstuck from it. So the airflow is not accelerated on top of the wing 's surfaces .

  • 1
    $\begingroup$ The question is about the physics for why flow separation occurs at certain speeds & angles of attack? This answer does not attempt to explain the fluid-physics involved. $\endgroup$ – ymb1 Sep 15 '19 at 16:01
  • 3
    $\begingroup$ "Airflow it will first become from laminar turbulent, then separation takes place" this is not correct, since laminar separation is a thing, however unusual on well-designed airfoils. $\endgroup$ – AEhere supports Monica Sep 15 '19 at 19:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.