How do sails generate lift?

[topmast]

Measure the kinetic energy of every single molecule of air .Note its vectors relative to the sail fabric.Not only is it moving but it's vibrating (temperature) Add up all the forces on both sides of the sail. Pressure is the sum impact of molecules on the sail. Glancing blows(laminar flow) ,less pressure. Difference (speed of air passage changes vector of impact) is seen as lift .or force at any given point.

[Singularity]

Originally Posted by thinwater

I could start with the fact that the flow is never fully attached, so the air does not exactly follow the path drawn by the sail.

A truly massive amount of experimental work was done by NACA, and you can find it on-line. Engineers use wind tunnels in part because you can debate the math forever. Much more can be done now computers, calculating flows and forces based on finite elements, but there is always a measure of chaos and statistics in fluid flow....
---
My personal understanding? I always felt that the circulation, vacuum, and Bernoulli approaches were less than rigorous. Newton, on the other hand, is rock solid. Sails change both the momentum and energy of the wind. Some is transferred to the boat and some is dissipated. As a practical matter, the math is changing too fast on the water (don't forget the keel, rudder, and waves), so we go with observations and rules of thumb... and try to pass the other guy!

Agreed wholeheartedly.

I'd suggest that there's a fundamental problem conceptualizing lift that's quite analogous to the problem of conceptualizing centripetal force. In casual conversation no one tends to argue whether or not centrifugal force is "real" yet if someone brings up centripetal force...well at best you end up with a gentleman's agreement. When discussing lift people typically refer to the second law (maybe the macro scale that you can feel) but neglect the third (arguably the micro scale that you have to compute).

Here's a lecture from a rocket scientist on the matter of lift, taking it to the Euler level. In the first 10 minutes of the clip he outlines the inadequacy of the typical lift explanations of Bernoulli vs Newton. Insofar as men who grew up eating lead paint chips in their cereal put men on the moon with slide rules, this microscale lift stuff is sort of trivia.

[olewsaa]

The book "The Art and Science of Sails" by Tom Whidden and Michael Levitt has excellent chapters about this. In addition the rest of the books are equally informative. Perfect bedside reading. The references are also well worth reading.


Have fun,
Ole

[robinmay]

Originally Posted by Tayana42

If the air particles moving around the outside of a sail have to move faster to get around the curve (Bernoulli) why don’t the air particles moving around the inside of the curve move faster to follow the curve of the inside of the sail? And if the air traveling over the curve does begin to separate causing a partial vacuum and then accelerate causing lower pressure does not the air traveling along the inside of the sail not also create a partial vacuum (and turbulence behind the mast) negating the pressure difference?

I've been pondering the same thing and I'm so glad you mentioned this. Unsatisfied by the answer I found lower on this thread, I looked it up and found this: https://www.quora.com/What-are-the-physics-of-sails-I-know-about-the-Bernoulli-effect-and-airplane-wing-example-but-a-sail-does-not-look-like-an-airplane-wing-Air-flows-on-both-sides-of-the-thin-curved-sail-Is-there-some-other-principle

Which basically says (Quora doesn't allow copy paste) that both flows are indeed curved but the curved flow has lower pressure on the inside of the curve and higher on the outside of the curve. The inside of the curve is the one on the sail side in front of the sail and the outside of the curve is on the side of the sail on the back.

Mind = blown!

[Ken Fry]

Originally Posted by Tayana42

If an airplane with flat airfoil surfaces can fly both right side up and up side down given the proper angle of attack could not flat sails not also work on both tacks given the proper angle of attack? I’ve never heard anyone refer to Newton’s 2nd law to explain sails generating lift. Does that not apply to us?

Yes, the second law applies quite directly.

First, a sail and a wing work on exactly the same principles. The first works at rather low efficiency (for many reasons), the latter works at higher efficiency (with efficiency, as I am using it, meaning the ratio of lift to drag.) A really good airplane wing (as on a sailplane) operates at 60:1 L/D. An extremely good, brand new, high aspect ratio, well-trimmed soft sail might reach 8:1 L/D

Hang gliders fly with soft sails. C-class catamarans sail with rigid airfoils.

Your suggestion that Newton's second law comes into play is ever so much better and more reasonable than the entirely incorrect explanation that was still in use when I took high school physics many decades ago. If you google "equal transit time lift theory" (and so forth) you will find, among many other things, well-produced (albeit tedious) you tube videos by university professors explaining how lift is really created. Also Google circulation patterns in wings. Both unhelpful and the more helpful you tube videos and internet articles will mention the Coanda effect. However, simply giving a name to the effect that causes a fluid to cling to a curved face does not explain why that effect occurs.

Abbott and Doenhoff's "Theory of Wing Sections" is a good place to start. It is the classic areodynamics text, and contains loads of actual wind tunnel results for wings of various cross sectional shapes.

I fear that Scientific American is not so scientific these days.

Real aerobatic airplanes (the ones designed specifically for aerobatics) have wings that are not flat top and bottom, but they do have equal curvature top and bottom (called in aero terms, zero camber). In right-side- up cruise flight, such wings are not as efficient as a properly selected sized and designed cambered wing, but the differences are much subtler than most people assume.

Consider a balsa wood glider. Many of these have truly flat wings, and fly pretty well.

A 0012 wing section and a 2412 wing section are from the same family, and both are perfectly good wings. The first is symmetrical, and the second is cambered. Their thickness profiles are identical. Either will make an airplane fly and the former is a profile I used to make a sailboat go.

So your understanding of a wing as a streamlined air deflector is much better than the old discredited "equal transit time theory". That theory was proposed by an aerodynamicist who said something to the effect that "A plausible lie is better than an incomprehensible truth." (for explaning difficult subjects to the lay person)

Odd how many sailors think that cambering of a wing is a requirement for lift, while sailing with uncambered wings in their keels. (Both the keel and rudder for a J-24 and the vast majority of other sailboats boats that have any aspiration towards good performance, are shaped to profiles you can find in Abbot and Doenhoff, generally in the range 008 through 0012.

[Tayana42]

Ken, thanks for this very interesting post. Pondering the engineering of foil design and performance adds to my fascination with my boat and with sailing.

[thinwater]

Originally Posted by bison65

...The bottom line is that the jib generates a faster airflow along the main sail on one side causing low pressure and lift. Try sailing against the wind on a Sun Fish with no jib and see what happens.

And yet the C-class catamarans have evolved away from using a jib, so the function and action of the slot is not as he explained it. Also, the back of the mainsail of a masthead rig often has very little negative pressure sailing to windward, as evidenced by it blowing away from the slot if the jib is just a little too tight.

[olewsaa]

Lift boils down to two simple fundamentals, Viscosity and angle-of-attack. From there is just level of sophistication.

Ole

[Ken Fry]

Originally Posted by olewsaa

Lift boils down to two simple fundamentals, Viscosity and angle-of-attack. From there is just level of sophistication.

Ole

I try not to make a habit of correcting the posts of others, but what you write is "not even wrong" in the view of practicing aerodynamicists, and if people read this thread, and believed what you said, they would leave here less educated than when they came in. You have things largely backwards, in the sense that only at a high level of sophistication does viscosity become important.

You can buy your own inexpensive CFD program (such as PANDA, if it is still available) to prove to yourself that viscosity effects fall into the "esoteric" or sophisticated side of things, and can be safely ignored, especially at the level of crudeness of a sailboat. Less expensive CFD programs (like PANDA) are likely to be based on Euler equations rather than Navier-Stokes, and for the purposes of designing a boat or airplane in the early stages, ignoring viscosity effects works just fine. (For late stages in the design of boats like the new America's cup boats and in the comparatively early stages of airliner design, viscosity effects come into play.)

Using PANDA, you can plot the lift curve of an airfoil (such as a 2412, for which there is abundant wind tunnel and other practical data) and come very close to the curve you can find in Abbott and Doenhoff (in which numerous airfoil wind tunnel results are published). So the plot using inviscid flow is plenty close enough for many practical uses.

Had you written "density" (instead of viscosity) you would have been closer, because at least density shows up in the fundamental, practical lift and drag equations. L= 1/2 rho x V^2 x CL x Area. Rho is the mass density of the fluid in question.

In support of the OP's largely correct understanding of things, consider that Newton determined that F = MA. The nature of the lift formula above essentially says that a force (lift) equals a mass (Rho) times an acceleration (V^2). The coefficient of lift and the area scale that force.

(Bear in mind that acceleration is a change in velocity, and either a change in speed or direction qualifies. So, when an airfoil changes the direction of the airflow over it, an acceleration occurs.)

CL will vary with angle of attack, but also varies with wing profile, camber, flap deflection, slat deflection, and so forth.

If I recall, Steve Clark claimed a CL of nearly 4 for the wing on Cogito. Small airplane wings (for Cessnas, etc.) rarely exceed 2.0 (in landing configuration) , although large airliners can be higher than that, owing to very complicated flap and slat systems. A typical pretty good sail can have a CL of about 1.0 (plus or minus a lot, depending on sail shape, etc.) All these figures can be at somewhere around the same angle of attack. (And in simple aircraft, the maximum lift CL, seen when in landing configuration, [-- say 1.6] is obtained at a lower angle of attack [8 degrees] than in the climb-out condition, where the CL is lower [say 1.0] (10 degrees.)

As a gross generalization, one can say that a single surface wing (soft sail) will operate better at low Reynolds numbers, and a double surface wing will operate better at higher Reynolds numbers. But this is so oversimplified that it can make people come to unwarranted conclusions, such as the notion that at the low Reynolds numbers encountered in sailing, a rigid wing won't work well. The C class cats prove this to be an incorrect oversimplification. My own Windrocket sailed at about three times wind speed in perfect conditions (found about once per year); flat water, 8-9 knots true wind for 24 knots over the water. So, by designing the boat to be fast (in all other respects), the limitation of low Reynolds numbers are self-compensating, to a degree.

(At a boat show or two, people would ask me about putting a wing on a J-24 or an even slower boat. I'd tell them it makes little sense, and low Reynolds numbers has a lot to do with that -- every thing else on the boat has to support sailing at multiples of windspeed to make a rigid wing worth considering. However... my current build has a rigid wing, and it will be a pretty slow boat. So there are other advantages to wings: Walker was not altogether wrong. The line between slow boats and fast boats, as I think about them, might be drawn at the point that the boat either tacks downwind (fast boat) or gybes downwind (slow boat).

Wow! Have I got verbal diarrhea, or what?

[Ken Fry]

Originally Posted by arsenelupiga

Anyone has explanation ? Is boat creating wind or instruments somehow playing up? Boat speed drop is real.

Yes, the boatspeed is creating wind. In windsurfing, one becomes viscerally aware of this effect. Likewise, on an iceboat, it is unmistakable. In ten knots of wind and iceboat sails on its own 80 knot wind. The forces on the sail are generated by the apparent wind... so ice boats have huge mainsheet tackle advantage ratios, because the potential force can be huge (64 times the potential force at 80 knots as at 10 knots.

This can sound like a perpetual motion machine. But doing the math, and drawing the vector diagrams helps.

Or better, sail a windsurfer. When you just begin to move in 15 knots true wind, you cannot hook into the harness (the equivalent of hiking our in traditional sailing terms) Because the wind the sail sees is too light. But a few seconds later, the forces are many times higher (four times higher if the apparent wind is now 20 knots in the true 10 knot breeze) and you have to hook in to avoid having your arms abused.

A sailplane sails on its own wind, and can do just fine on a calm day.

A fun thought experiment: Imagine a sailboat crossing a river in a flat calm. The current is 4 knots. Why can the boat sail as if it is sailing on calm water in 4 knots of wind? In a dense fog on a large river, it would be impossible to tell the difference -- in other words you might swear you are sailing in a lake.

[Ken Fry]

Originally Posted by thinwater

And yet the C-class catamarans have evolved away from using a jib, so the function and action of the slot is not as he explained it.

Yes. And of course a Sunfish sails fairly well to windward, as does a Laser, or any other sailing dinghy. Cat-rigged boats sail on the same windward-leeward courses when racing that boats with other rigs (such as sloops) do.

[olewsaa]

Viscosity is essential, try taking it away (experiment is done by superfluid helium, which have zero viscosity, no lift, no drag). Review Arvel Gentry's papers. Or Wikipedia :
https://en.wikipedia.org/wiki/Superfluidity . Viscosity is an interesting property of any fluid, ranging from low level with noble gases to the more extreme normal glass (which is a liquid).

To understand lift and viscosity you could start with he very basic at a good University like MIT :
"Viscosity is essential in generating lift. The effects of viscosity lead to the formation of the starting vortex (see Figure 4), which, in turn is responsible for producing the proper conditions for lift."
https://web.mit.edu/2.972/www/report...l/airfoil.html

As viscosity is such an essential factor to understand flow, shear and flow of liquids. It's good to settle the fundamentals before embarking on any elaborate computer approximations. Not that there anything wrong with those .



Ole

[Don C L]

OK, I'd like to pose this to those more knowledgeable than I, is there a threshold of apparent wind speed where a jib, or something else creating a slot, becomes more disadvantageous than advantageous? And is this at all analogous to biplanes vs. monoplanes? Or is it analogous to the leading edge slats on STOL aircraft and jet airliners? And if this discussion has already appeared elsewhere feel free to redirect me.

I guess what I am getting at is: is it always advantageous to employ a jib, or employ a rig that depends on a jib, at the kinds of speeds a ballasted, non-planing, sailboat will be working in? My assumption had been no, but I am thinking now I was wrong.

[Ken Fry]

Originally Posted by olewsaa

Viscosity is essential, try taking it away (experiment is done by superfluid helium, which have zero viscosity, no lift, no drag).
Ole

Of course. But so is quantum mechanics.

However, a sailor contemplating a sloop rig vs a cat rig, or a soft sail vs a rigid wing need not bone up on either quantum mechanics or the viscosity of air vs that of other fluids. With a pencil and one small napkin, a wing designer can calculate the lift and drag of a given wing section operating at a particular angle of attack at a particular airspeed, knowing nothing more than the mass density of the fluid involved (which any designer knows by heart for air and water, -- whereas fewer can quote viscosity figures for each from memory.)

I provided the formula for lift (or drag with insignificant change) above, and nowhere in it is there a term for viscosity. (Although you could argue, obstinately, that the CL term is dependent upon Reynolds effect, and therefore on viscosity. But then I would argue, more obstinately, that the lift of many sections changes very little with change in Reynold's numbers of 3,000,000 vs 6,000,000, but change dramatically with a doubling of mass density.)

The same formula applies to the lift of a sail and the lift of a keel in water, but the differences in viscosity (air vs water) do not account for the difference in lift, whereas the differences is mass density do. The reason keels are very small and sails very large (despite, when close hauled, operating with similar lift forces) has to do with the large differences in mass density quite directly and with the much smaller
differences in viscosity only indirectly.

If you have a simple formula that a sailor can use to predict the lift of his sail based on the viscosity of air, then now would be the time to produce it.
Perhaps there are those here who would find that helpful.

In the back-of-an-envelope calculations that a designer does to predict performance of a wing or keel, mass density always shows up, and viscosity does not.

I agree that there is value in taking some fluid dynamics courses at MIT, but doubt that doing so will be practical for the members of this forum. I suspect that even this conversation is already too abstract for this audience -- it's too abstract for me, and I've built a couple rigid-wing-powered boats, so have a higher-than-average tolerance for such things.

Perhaps our different perspectives are like those of the mathematician and the engineer faced with making a series of halfway steps to the pretty girl on the other side of the room. The mathematician will never get there. The engineer gets close enough.