Re: Physics of Flight

[John Denker <jsd@MONMOUTH.COM>]

At 05:32 PM 8/16/99 -0400, Michael Edmiston wrote:
> In the summary of his third chapter he says: "A wing is very effective
> at changing the speed of the air. The air above is speeded up; the air
> below is slowed down. Each air parcel gets a temporary change in speed
> and a permanent offset in position."
>
> The problem I am having is... why?  Why is a wing very effective at
> changing the speed of the air.

If you find the wing amazing, well, so do I.  I've spent thousands of hours
studying wings of various kinds, and I still find it amazing how well they
work.  It literally sends shivers of amazement up my spine to think about it.

As Hans Bethe said in another context, there are two types of magic.  The
humdrum sort of magic is fascinating when/because you don't know how it's
done.  Then there's the really special sort of magic, where even if you
know exactly how it's done, it's still amazing.

> As an object moves through a fluid, surely the atoms/molecules ahead of
> the object

Indeed, remarkably *far* ahead.

> are being displaced to some other location; likewise the
> atoms/molecules must "fill-in" behind the object or else we're left
> with a vacuum behind the object.

OK.  There's more to the story than that, but that's not wrong.

> For the moment, let's ignore what happens in front of the object
> because it is behind the object that gives me the most problems.  I
> would say that the air going over the top of the wing is speeded up
> because it is being sucked into the void created behind the moving and
> tilted wing.

Yes, the air accelerates into regions of less pressure and decelerates into
regions of higher pressure.

But no, the low-pressure region is not "behind" the wing.  The point of
lowest pressure is rather far foward on the top side of a typical wing.
The point immediately aft of the wing is a point of high pressure.

The analytical program you are suggesting -- finding the pressure
distribution and then deriving the velocities from that -- is perfectly
sound in principle, but it is not very convenient.  There are no easy
principles to allow you to find the pressure distribution, without first
knowing the velocities.

In contrast, there are relatively feasible ways of determining the
velocities, and then determining the pressures therefrom.  Specifically,
the flow is a superposition of the fore-to-aft flow (due to the ordinary
airspeed) plus the circulatory flow.  The fore-to-aft flow is easy to
visualize.  A vortex is less familiar, but you can easily stir up an
example and look at it.  If you are the least bit interested in fluid
dynamics, it's well worth spending an hour or two doing this.

> Then the question becomes: in what manner does it fill
> this void?  If it turbulently swirls into this region it doesn't have
> any obvious net momentum and it is difficult for me to imagine lift.
> In addition, it seems that drag is high if "air friction" changes from
> bv^1 to bv^(more than one) when turbulence sets in.

1) There's no "void".
2) Turbulence is not required for producing lift.

> On the other hand, if it is laminar flow and the air follows the upper
> surface of the wing, then it will have momentum directed along the
> angle of attack (i.e. momentum typically directed downward).  To me
> that implies lift and also low air friction.

Laminar flow is nice and elegant.  But the flow is never perfectly laminar,
and sometimes a turbulent boundary layer works better than something more
nearly laminar.  But having lots of turbulence would not be good.

> I am not sure how to understand turbulent flow versus separated flow
> (words Denker distinguishes).

The distinction is important.  Can you formulate a more specific question?

> John's computer generated wind tunnel
> drawings, and real wind tunnel & smoke photos I have seen, certainly
> look like laminar flow to me, and they look like turbulent flow once
> the wing has stalled.

I doubt the photos you've seen can show what's going on in the boundary layer.

>  Is this all wrong?  I can imagine some degree of
> turbulence that somewhat follows the wing shape.

That's the right idea.

> And I can imagine
> some degree of separation without gross turbulence.

Also correct.  It is common for the flow to detach at point "A" and then
re-attach at point "B" a few inches downstream.  The air between the
detached flow and the wing need not be grossly turbulent.  It just hangs
around, flows weakly in the wrong direction, et cetera.

>  But, in general, I
> sure have the impression that a major stall has both gross turbulence
> and separation behind/above the wing.

That's true for a "major" stall (by which I assume you mean flight at
angles of attack well beyond the critical angle of attack).

On the other hand, remember:
  Separation is the cause of stalling,
  just as water is the cause of drowning
  --- but you can get quite wet before you drown.
  A wing in normal flight always has *some* separation,
  and the amount of separation grows gradually as the
  angle of attack grows toward, through, and beyond the
  critical angle of attack.

Nothing qualitative happens to the airflow pattern at the critical angle of
attack.  The coefficient of lift one degree above critical is essentially
the same as one degree below critical.  Lift does not suddenly go to zero
at the critical angle of attack.  (Now, certain stability derivatives *do*
go throuh zero at or near the critical angle of attack, and this has a big
effect on the airplane's handling characteristics, but that's a different
story entirely.)

OK? --- jsd