The Mystery of Fly Flight: Debunking Bernoulli's Law Explanation

[feynmann]

Why does fly fly? The theory based on Bernoulli's law fail to explain it since the wings of fly are flat, not curved like the wings of birds

"An air parcel going over the curved top of the wing has to travel a longer distance, but it has to arrive at the trailing edge at the same time, hence it has to travel faster, and Bernoulli's law says that pressure decreases as speed increases."

 

[mgb_phys]

Why does fly fly? The theory based on Bernoulli's law fail to explain it since the wings of fly are flat, not curved like the wings of birds

Only soaring birds wings are curved - if you flap the wings everything is different.

"An air parcel going over the curved top of the wing has to travel a longer distance,

True but irrelevent

but it has to arrive at the trailing edge at the same time,

Who says ?
This is a common misconception about how aeroplane wings work - do a search on this site.

 

[chroot]

Airplanes do not fly (entirely) due to explanation given by Bernoulli's law, though it does contribute to the lift created by some wings. Bernoulli's law as a sole explanation of lift is one of the most enduring popular myths in physics.

Aerobatic airplanes have wings with symmetrical cross-sections. Most airplanes with asymmetrical wings are also capable of flying inverted. Airplane wings generate lift mainly due to their angle of attack -- they push air down, and the reaction force pushes them up. Simple as pie.

- Warren

 

[Lok]

Flat wings or curved ones, it does not matter if u want to fly. Birds and flies alike push the air downwards creating lift, so the wing shape can be almost whatever you want it to be, even in nice sparkly pretty colors ( butterflies ).

But flies cannot glide, that means if they hold their wings still they drop. Birds have curved wings so they can glide.

 

[zoobyshoe]

Actually, I read about 20 years ago that insect wings are a bit stretchy and acquire camber on the downstroke. They're not flat during flight. They can, in fact, operate like hanglider wings. They are reoriented on the upstroke to cut through the air with the least amount of resistance and reposition themselves for the next downstroke.

What's really cool is the inertial guidance systems of flies. Also, they can walk upside down on the ceiling by virtue of the hairs covering their "feet" which gain purchase in the microscopic irregularities of the surface.

 

[chroot]

The hair on my feet, on the other hand, isn't really all that useful to me.

- Warren

 

[Andy Resnick]

Why does fly fly? The theory based on Bernoulli's law fail to explain it since the wings of fly are flat, not curved like the wings of birds

"An air parcel going over the curved top of the wing has to travel a longer distance, but it has to arrive at the trailing edge at the same time, hence it has to travel faster, and Bernoulli's law says that pressure decreases as speed increases."

Insect flight is only partially understood. YOu may be interested in looking through some of the referenced work here:

http://en.wikipedia.org/wiki/Insect_flight

and specifically this article and references:

http://www.cs.washington.edu/homes/diorio/MURI2003/Publications/sane_review.pdf

 

[zoobyshoe]

Insect flight is only partially understood.

I know some drosophila who would disagree with this.

 

[zoobyshoe]

The hair on my feet, on the other hand, isn't really all that useful to me.

That's just so "inside the box".

 

[rcgldr]

Bernoulli's "law" bascially states that when no work is done during an exchange between pressure and speed, the total pressure, the sum of static pressure and dynamic pressure (dynamic pressure is relative to speed ²) is constant.

At the air + wing interface, work is done, so Bernoulli's law is violated. Away from the wing, where no mechanical interaction takes place (direct deflection from the bottom surface or void "creation" from the top surface), Bernoulli's law applies because no work is being done once away from the immediate vicinty of the wing.

In the case of insects, and humming birds, the flapping of wings that rotate (sometimes via flexing) in coordination with the flapping, generates a downforce on the air coexisting with the accelerated air generating an upforce on the wings. A dragon fly has relatavely large enough wings that it can glide well, but a bumble bee, fly, or humming bird require constant flapping.

The pattern of the flapping is different. Humming birds use a figure 8 pattern and dynamically change this pattern to control movment. Butterflies "clap" their wings at the top of the flap stroke.

As mentioned in the Wiki article for smaller insects, the Reynolds number is smaller, and to the insect the air is "thicker" because of viscosity versus their size and weight (weight will ultimately be related to the required generated air stream speed to sustain flight).

One aspect of the constant flapping of wings that is interesting is minimizing the amount of power required to sustain flight. The muscles involved are very elastic, like springs, and only require a relatively small amount of power to keep flapping. In many insects, the muscles are attached to an elastic membrane that is then attached to the wings as mentioned in the wiki article.

 

[Andy Resnick]

I know some drosophila who would disagree with this.

Do you talk to them as well?

 

[zoobyshoe]

Do you talk to them as well?

No. I lurk at a fruit fly forum.

 

[vibjwb]

Jeff is right fly’s wings rotate. One way to think about it is that a fly is swimming through the air.

 

[Cyrus]

I didn't read all the posts here, but I will give a very basic overview of what happens in low Reynolds number insect flight.

First of all, the statement that flies have 'flat plate' wings is flat out wrong. Micro Air Vehicles (MAVs) of a size somewhat close to that of flies (most slightly larger - drag fly size), do have chamber to them.

The motion of a flapping wing MAV is also quite complex. There is an upstroke and backstroke. Each stroke takes advantage of different things. On the up stroke the wing goes through the air rather cleanly. It then reaches the forward position and puts the trailing edge down. This causes a ton of drag. Then the backstroke begins and the drag caused during the transition actually helps provide more lift on the return stroke. Its a non-intuitive effect.

My advisor is a leading expert on insect flight and navigation. The information above is from a talk given here by a guy from a lab up in Princeton.

We have small wind tunnels with high speed cameras that track fly flapping motion.

[1] http://www.terp.umd.edu/4.0/engineering/

 

[Cyrus]

Airplanes do not fly (entirely) due to explanation given by Bernoulli's law, though it does contribute to the lift created by some wings. Bernoulli's law as a sole explanation of lift is one of the most enduring popular myths in physics.

Aerobatic airplanes have wings with symmetrical cross-sections. Most airplanes with asymmetrical wings are also capable of flying inverted. Airplane wings generate lift mainly due to their angle of attack -- they push air down, and the reaction force pushes them up. Simple as pie.

- Warren

and that force is caused by a pressure distribution. Sorry, Bernoulli just got you.

 

[rcgldr]

and that force is caused by a pressure distribution. Sorry, Bernoulli just got you.

Bernoulli's equation describes a relationship between static pressure and speed^2 (component of dynamic pressure). There are pressure distributions on a wing that do not correspond to the speed of the air as described by Bernoulli's law, because work is done. For example, in the vicinity of the the upper leading edge of a wing , you have a significant component of centripetal acceleration of air, that corresponds to a reduction in pressure with no change in speed.

 

[parsec]

There's a good series called nature tech on discovery. I remember seeing a group at caltech that does flow visualisation on a large scale model fly in oil. I can't seem to find the video on youtube though.

Their consensus seemed to be that flies generate most of their lift from a leading edge vortex that stays attached to the wing throughout the fly's wing flapping movement.

 

[Cyrus]

Bernoulli's equation describes a relationship between static pressure and speed^2 (component of dynamic pressure). There are pressure distributions on a wing that do not correspond to the speed of the air as described by Bernoulli's law, because work is done. For example, in the vicinity of the the upper leading edge of a wing , you have a significant component of centripetal acceleration of air, that corresponds to a reduction in pressure with no change in speed.

Its not because work is done, it's because viscous forces are prevalent. Bernoullis law is just a special case of the Navier-Stokes Equations. The Navier-Stokes Equations always hold true, everywhere.

Work being done would be an effect of a helicopter rotor or propeller blade when you are using momentum theory.

(PS, I was giving warren a hard time about what he said for the forces on a wing. He's right, I was nitpicking).

 

[rcgldr]

Its not because work is done, it's because viscous forces are prevalent. Bernoullis law is just a special case of the Navier-Stokes Equations. The Navier-Stokes Equations always hold true, everywhere. Work being done would be an effect of a helicopter rotor or propeller blade when you are using momentum theory. (PS, I was giving warren a hard time about what he said for the forces on a wing. He's right, I was nitpicking).

I was being equally nitpicky about Bernoulli. Although the Navier-Stokes Equations hold in the real world (not sure how complicated turbulence makes this), classical Bernoulli doesn't because of the assumption that total energy is constant, or that total energy along a stream line is constant. The mass flow is constant, but the energy is clearly changed in the case of a propeller, rotor, or turbine, and although the amount of energy change is less with a wing, it's still there.

The "work being done" issue is mentioned in this Nasa article on propellers, and exit velocity. Not covered is what is happening at the outer edges of the decreasing diameter funnel of the main air stream.

"But at the exit, the velocity is greater than free stream because the propeller does work on the airflow. We can apply Bernoulli'sequation to the air in front of the propeller and to the air behind the propeller. But we cannot apply Bernoulli's equation across the propeller disk because the work performed by the engine (by the propeller) violates an assumption used to derive the equation."

http://www.grc.nasa.gov/WWW/K-12/airplane/propanl.html

 

[Cyrus]

I was being equally nitpicky about Bernoulli. Although the Navier-Stokes Equations hold in the real world (not sure how complicated turbulence makes this), classical Bernoulli doesn't because of the assumption that total energy is constant, or that total energy along a stream line is constant. The mass flow is constant, but the energy is clearly changed the case of a propeller, rotor, or turbine, and although the amount of energy changed involved is less with a wing, it's still there.

That is exactly correct. The NS equations are true - period. It is the NS equations that CFD solves.

The "work being done" issue is mentioned in this Nasa article on propellers, and exit velocity. Not covered is what is happening at the outer edges of the decreasing diameter funnel of the main air stream.

Thats called 'wake contraction'.

" But at the exit, the velocity is greater than free stream because the propeller does work on the airflow. We can apply Bernoulli'sequation to the air in front of the propeller and to the air behind the propeller. But we cannot apply Bernoulli's equation across the propeller disk because the work performed by the engine (by the propeller) [/i] violates an assumption used to derive the equation.[/i]"

http://www.grc.nasa.gov/WWW/K-12/airplane/propanl.html

That is called, 'momentum theory'.

 

[feynmann]

I didn't read all the posts here, but I will give a very basic overview of what happens in low Reynolds number insect flight.

First of all, the statement that flies have 'flat plate' wings is flat out wrong. Micro Air Vehicles (MAVs) of a size somewhat close to that of flies (most slightly larger - drag fly size), do have chamber to them.

[1] http://www.terp.umd.edu/4.0/engineering/

The question I have is this. Can fly fly even if the wings are flat? If yes, then it proves that the camber of the wing is not the reason that fly can fly. It certainly helps to have camber

 

[Cyrus]

The question I have is this. Can fly fly even if the wings are flat? If yes, then it proves that the camber of the wing is not the reason that fly can fly. It certainly helps to have camber

I guess in theory, but the aerodynamics here are very low Reynolds number - which means things are very sensitive to even small amounts of chamber, surface roughness, etc. So to say 'can they'...I don't know. MAYBE? Perhaps you could get a mechanical fly to, I don't know how the power curves look like on flat vs chambered for a fly. Its possible that the flat plate wings require more power than a fly can provide.

 

[getitright]

Airplanes do not fly (entirely) due to explanation given by Bernoulli's law, though it does contribute to the lift created by some wings. Bernoulli's law as a sole explanation of lift is one of the most enduring popular myths in physics.

Aerobatic airplanes have wings with symmetrical cross-sections. Most airplanes with asymmetrical wings are also capable of flying inverted. Airplane wings generate lift mainly due to their angle of attack -- they push air down, and the reaction force pushes them up. Simple as pie.

- Warren

I have to call BS on this one! No offense meant but the increased air rushing over the top of the wing creates a localize low pressure causing lift of the wing defeating the wind drag under the lower part of the wing at a slightly increased pressure.

I'm in Houston, where are all you astronauts and flight jocks to back me up with a better explanation?

 

[Cyrus]

I have to call BS on this one! No offense meant but the increased air rushing over the top of the wing creates a localize low pressure causing lift of the wing defeating the wind drag under the lower part of the wing at a slightly increased pressure.

I'm in Houston, where are all you astronauts and flight jocks to back me up with a better explanation?

I already gave an explanation. There is a pressure difference between the top and bottom wing. Forum Side note: why do you we have to have this discussion about how an airplane wing flies for the millionth time around here. Someone should just put a sticky that stays STOP ASKING about Bernoulli and wings. Good god, a thousand and one threads on this gets old fast. The OP didn't even ask about Bernoulli, so why are we even talking about it?

 

[rcgldr]

increased air rushing over the top of the wing creates a localize low pressure causing lift of the wing defeating the wind drag under the lower part of the wing at a slightly increased pressure. I'm in Houston.

Perhaps you could explain how your theory (air rushing over the top of the wing) applies to these pre-shuttle prototypes?

M2-F2 glider:

M2-F3 rocket powered version (max speed mach 1.6):

 

[getitright]

I already gave an explanation. There is a pressure difference between the top and bottom wing.


Forum Side note: why do you we have to have this discussion about how an airplane wing flies for the millionth time around here. Someone should just put a sticky that stays STOP ASKING about Bernoulli and wings. Good god, a thousand and one threads on this gets old fast. The OP didn't even ask about Bernoulli, so why are we even talking about it?

I get your idea, life is too short, huh. Some folks on this site can tell folks about the deflection of the wing of a fly on the upward motion as oppossed to the downward motion but you didn't mention that. Perhaps no one mentioned that the wing when going up is at a 75degree angle or that when it goes down it is at a 22 degree angle in stable flight. But then again maybe one of the experts you have access to has that high resolution slow motion video of the flight of a fly. Give them a call. I can give you a number if you need it.

 

[Cyrus]

I get your idea, life is too short, huh. Some folks on this site can tell folks about the deflection of the wing of a fly on the upward motion as oppossed to the downward motion but you didn't mention that. Perhaps no one mentioned that the wing when going up is at a 75degree angle or that when it goes down it is at a 22 degree angle in stable flight. But then again maybe one of the experts you have access to has that high resolution slow motion video of the flight of a fly. Give them a call. I can give you a number if you need it.

My advisor isn't about to distribute video of that on the web for a physics forum.

 

[getitright]

Perhaps you could explain how your theory (air rushing over the top of the wing) applies to these pre-shuttle prototypes?

M2-F2 glider:

M2-F3 rocket powered version (max speed mach 1.6):

These designs were specifically meant for high altitude and/or vertical travel at supersonic speeds and still allow a controlled descent.

 

[rcgldr]

m2-f2, m2-f3

These designs were specifically meant for high altitude and/or vertical travel at supersonic speeds and still allow a controlled descent.

And yet they still produce adequate lift at low sub-sonic speeds, and with the "hump" on the bottom of the wing instead of the top. Compare the angle off attack of the F104 (essentially a jet powered missle with tiny wings) to the M2-F2. The M2-F2 picture is a bit deceiving since the photo is angled a bit.

The M2-F2 and M2-F3 are good examples to "disprove" equal transit theory as the cause of lift, and I just find them interesting, as they are fairly unique.

I rotated the picture 2 degrees right to make the ground appear level, still the upper surface of the m2-f2 is nearly horizontal, but it could be in a "flare" since it's landing, but the wheels are still up, so even though it's fairly low, it's got some gliding time left to deploy the landing gear.

Also the top surface isn't completely flat, it tapers at the tail, but the main point is shape of the lower surface, and the fact that the bottom surface is the "longer path".

 

[Cyrus]

There is a two day seminar of flapping wing flight on campus tomorrow and the day after that I will be attending. I'll report anything interesting. Here is the schedule

7:30 - 8:00 ***** Registration ($20), Coffee and Donuts ******
8:00 – 8:30 Mechanics Center Introduction: Chopra
8:30 – 10:00 Task-1 : Aeromechanics: Humbert
1.1 Fundamental bio-inspired principles of flapping flight physics - - Humbert/Dickinson
1.2 Dual-plane particle image flow diagnostics of flapping-wing unsteady aerodynamics - - - Leishman
1.3 DNS/LES/RANS analysis for rotary- and flapping-wing-based MAVs- - - Baeder/Yamleev
10:00 – 10:15 - - - Coffee Break
1.4 Flight dynamics simulation modeling of MAVs - - - Celi
1.5 Aeromechanics of revolutionary cyclocopter and flapping rotors - - - Chopra/Benedict
1.6 Bio-inspired flexture-based wings and airframes - - - Dickinson/Humbert
1.7 Avian-based wing morphing for agile flight - - - Hubbard

12:15 – 1:15 - - - - Lunch Break - - - - -

1:15 – 3:15 Task-2 : Ambulation: Full
2.1 Bio-inspired dynamic modeling and simulation with parameters for ground contact model - - Full /Goldman
2.2 Bio-inspired principles of appendage and actuator design - - - Full/Fearing/Wood
2.3 Ambulatory design of body and appendages - - - Full/Fearing
2.4 Bio-inspired crawling, running, climbing robots - - - Fearing/Full/Wood
3:15 – 3:30 - - - Coffee Break - - -
3:30 – 5:00 Task-3: Hybrid Aeromechanics/Ambulation: Fearing
3.1 Thrust augmented entomopter: a revolutionary hover-capable high-speed MAV - - - Chopra/Wereley
3.2 Bio-inspired hybrid aerial and terrestrial locomotion - - - Fearing/Full/Wood/Humbert
3.3 MBMAC: Multi-body Microsystem Analysis Code for rotary, flap and ground - - Masarati/Goldman/Chopra
5:00 Demonstrations & Reception (Kim Engineering Building Rotunda)

Location: Kim Engineering Building
8:30 – 9:00 - - - Coffee & Donuts - - -
9:00 – 11:30 Task-4: Multifunctional, Actuation and Propulsion: Wood
4.1 High performance microactuators - - - Smela/Fearing/Wood
4.2 Smart composite-based rapid fabrication of micromechanical and micromechatronic structures - - Wood/Fearing
4.3 Ultra-light multifunctional composite structures based on electrospun fabric - - - Shivakumar/Lingaiah
4.4 Chemical energy conversion system - - - Cadou/Jackson
4.5 Distributed propulsion system for power efficiency - - - Fearing/Full/Wood
11:30 – 12:00 ARL and Government Meeting
12:00 – 1:00 - - - Lunch - - -
1:00 – 1:30 TMG Meeting and Hot Wash

 

[Cyrus]

And yet they still produce adequate lift at low sub-sonic speeds, and with the "hump" on the bottom of the wing instead of the top. Compare the angle off attack of the F104 (essentially a jet powered missle with tiny wings) to the M2-F2. The M2-F2 picture is a bit deceiving since the photo is angled a bit.

The M2-F2 and M2-F3 are good examples to "disprove" equal transit theory as the cause of lift, and I just find them interesting, as they are fairly unique.

I rotated the picture 2 degrees right to make the ground appear level, still the upper surface of the m2-f2 is nearly horizontal, but it could be in a "flare" since it's landing, but the wheels are still up, so even though it's fairly low, it's got some gliding time left to deploy the landing gear.

Also the top surface isn't completely flat, it tapers at the tail, but the main point is shape of the lower surface, and the fact that the bottom surface is the "longer path".

You have to be careful, just by looking at a picture gives you no indication of the AoA of those two aircraft. All you're seeing in this photo is the flight path angle - not the AoA.

Despite this, it is a very interesting aircraft design! It looks like a delta wing with the pilot all the way forward for stability in pitch.

 

[FredGarvin]

That is exactly correct. The NS equations are true - period. It is the NS equations that CFD solves.

Approximates. N-S equations have very few situations that can be solved exactly.

That is called, 'momentum theory'.

AKA "Actuator Disc Theory."

 

[Cyrus]

Approximates. N-S equations have very few situations that can be solved exactly.

AKA "Actuator Disc Theory."

N-S doesn't have a closed form solution, implied in my post was the necessity to use something like CFD to solve them. That does not make the N-S equation an approximation. The solution is an approximation to the N-S equations, but the Equations are not an approximation.

*The only real "approximation" is that the gas particles follow a continuum.

As for Actuator Disc Theory, tom-a-to, to-ma-to.

 

[Cyrus]

I left half way into the talks because I have work to do. But what I did see was pretty interesting. A guy from CalTech had optical sectioning images of a fly. Basically, they use special infrared laser beams to scan the fly and you can see all the internal structure of the fly just like a 3d MRI. The fly has two muscles that cause the complex flapping motion. One main muscle always powers the flies wings, while these two muscles adjust the tension, and thus the equivalent spring constant to change the flapping properties. Pretty non-intuitive. There were also a video of a honey bee inside a wind tunnel given a wind gust disturbance, and a high speed strobe video of a fly with a piece of string tethered onto its back to keep it stationary.

 

[lividz]

Look this was shown on discovery...
actually the wings of the fly create whirl of wind[sorry for that..I can't find the exact word now...hope u understand what I want to convey]...on its lower back side...this is the main cause of lift