Winglets on sails?

[champ0815]

Is there a reason for not using winglets on sails?
On planes they are used to increase the efficiency of the wings (http://en.wikipedia.org/wiki/Wingtip_device), why shouldn't they be used on sailboats?

[yipster]

Sail plan from the bit harder to find boatwiki http://boatdesign.net/wiki/Main_Page
read under Sail Plan - Needs additional contributions arvelgentry on how the jib acts as LEX and search various threads (and check my gallery) for winglet
another great thread, sail airodynamics sail aerodynamics

[champ0815]

Wow, that's something to read...
Thanks a lot - must have missed it in my short survey about the topic...

[tspeer]

The use of winglets hinges on the question, "Does a winglet reduce the induced drag due to lift more than it adds in parasite drag?" If you are adding surface area the size of the winglet, the least induced drag is obtained when that area is used to extend the span of the lifting surface itself.

If the span is limited for some reason, say, in the case of a Standard Class sailplane or the need to fit airplanes next to each other at terminal gates, then a winglet makes sense. Winglets make sense on keels because the depth is limited by grounding considerations.

But for a sail rig, there are rarely any a priori limits on the height, so it's better to just make the rig taller. An exception might be if you are designing to get under the bridges on the Intracoastal Waterway, etc.

Other factors include not being able to reef a winglet, getting a design that works for both tacks and at a number of angles of heel, and weight aloft.

[champ0815]

Originally Posted by tspeer

Other factors include not being able to reef a winglet, getting a design that works for both tacks and at a number of angles of heel, and weight aloft.

Well, greater height of a sail is a problem since the heeling momentum is increased, or am i wrong?
As for the reefing, i thought about a square top sail with the whatsoever winglet attached to the top batten, so it would reef with the sail...

[yipster]

thanks again for the points Tom, very interesting subject nevertheles http://aero.stanford.edu/Reports/VKI_nonplanar_Kroo.pdf
more reading from under the search button on winglets http://www.boatdesign.net/forums/sea...archid=1251173

[champ0815]

Nevertheless, most of the posts about winglets are about the underwater side of the boat. There is no thread dedicated to the aerodynamic effects on the upper (and lower) end of the sail. (at least, I haven't found one)
I just wanted to start a discussion about the potential increase in efficiency of a sail (same area, same aspect ratio) when the flow around the upper and lower edge of the sail are somehow diminished.

[quicksail]

I have to agree with Tspeer on this. As we all know design is all about compromise. Yes you can make the "sail" more effiecient by adding winglets but at what cost. Will the suface drag be increase? What about the added weight aloft and the reduction in righting moment? This consequences might hinder the overall performance of the boat.

I to have been interested to see if winglets can be used. Especially now with large square topped mainsails. But how big would they need to be? It is hard enough to support the square tope mainsails now. How much added weight would winglets be? Lots of questions that would need to be answered. Sounds like a good project for a student. Could be fun.

[champ0815]

Originally Posted by quicksail

Will the suface drag be increase?

Sure it would, but it would depend on the material and the size...

Originally Posted by quicksail

What about the added weight aloft and the reduction in righting moment?

Can you explain the reduction in righting moment? I thought it would result in a reduction of heeling moment... . Or do you mean reduction in righting following additional weight in the mast top?

Originally Posted by quicksail

But how big would they need to be?

Judging from the size of winglets on planes, the area would not be to so large. But maybe it's a question of velocity...

Originally Posted by quicksail

It is hard enough to support the square tope mainsails now. How much added weight would winglets be? Lots of questions that would need to be answered. Sounds like a good project for a student. Could be fun.

What about endplates on the lower edge of the sail? Maybe some structure on the boom or only a tiny gap between boom and the roof of the cabin on larger boats (each tack and each jibe cleans your roof top window or your solar panels )? There, the added weight (if any) would be closer to the COG of the boat.

[TTS]

Does anyone have a design for sail winglets. It might be interesting to try on an A-Cat.

Tom, contact Magnus("High on Carbon"); he and Fredo tried them on a C CLass cat. Find his contact info here(click on "High on Carbon") :
http://boatdesign.net/forums/showthr...t=18998&page=2

[miloman]

The first thing to do in order to increase the efficiency of a sail is to change its shape. Triangles are poor foils, rectangles, trapezoids, and elipses all generate much more lift for their drag than triangles. You've probably already noticed that there are no triangular keels. You can see on many modern rigs that mainsails are growing more eliptical or rectangular. Fat headed rigs, and rigs with fully battened mains and huge roaches are examples of this trend.

The reasons that modern sails tend to be triangular is that they are a natural shape to fit on our tall masts with standing backstays. The gaff main is at least in theory not a terrible shape for a sail. Don't let anyone tell you that the gaff disappeared because it was less efficient than the marconi sail. The gaff rig disappeared because the International and Universal rules only measured the amount of sail area a boat could set, and this doesn't take account for the shorter gaff rigs ability to carry more sail. Steve Dashew's Beowulf shows the potential of shorter more efficiently shaped sails. The masts on Beowulf are stubby and short but they carry full-battened mains with a practically eliptical, she is quite fast. Eric Sponberg also has some information about sail shape on his website.

Basically changing the shape is a much quicker way to increase the efficiency of a sail than adding winglets.

[champ0815]

Well, I completely agree with you! Winglets on triangular sails wouldn't make sense, since there is no top edge around which the air can stream from high to low pressure side (although there should be some tip vortice?).
Anyhow, I have to repeat my question:
For a given efficient sail shape, preferably rectangular at the top, and a given area, would there be any advantage to add winglets in order to increase the efficiency?
Or the other way round, can winglets increase the efficiency in a way to allow for the reduction of mast height with equal performance to a non-winglet sail? I think, this is an interesting question, since a lower rig results in lower heeling momentums and therefore in faster and safer sailing for both mono- and multihulls.

[oldsailor7]

IMHO an attempt to put winglets on the top of a sail is impractical for the reasons already stated. There is also the fact that a pitching motion of the boat is going to be continually altering the AOA of the winglet,thus reducing its efficiency.

However any sail configuration has to be related to the hull of the boat it is going to be driving.

It would be an incongruity to put a high aspect ratio square top fully battened sail on a Drascombe Lugger for instance. A short masted lug rig is more effective, despite its poorer L/D ratio, because it is more suited to the limited hull speed of the Lugger.

On the other hand the tall higher efficiency sails are eminently suitable on boats such as an A Class Catamaran, or any boat that has a fine and slippery hull shape and low windage.

[tspeer]

Originally Posted by champ0815

Well, I completely agree with you! Winglets on triangular sails wouldn't make sense, since there is no top edge around which the air can stream from high to low pressure side (although there should be some tip vortice?).

This is a widely repeated explanation for the induced drag of a lifting surface, but it is wildly inaccurate and misleading. Induced drag is due to the fact that the lifting surface has a finite span, but air escaping around the tip has very little to do with it.

Induced drag is the result of deflecting a finite amount of air through a finite angle, imparting a sideways velocity to it. It's not just the air at the tip that counts, it's the entire wake that matters. In effect, a boat sails in a header of its own making. This self-generated header results in the lift vector being tilted backwards compared to the free-stream velocity upstream, and this looks like a drag component.

The trailing vortices are the result of the rest of the air getting out of the way of the slice of deflected air, and flowing in behind to replace it. They are exactly the same as the vortices you see at the edge of a canoe paddle as you draw it through the water. With the paddle, it's clear that the water is flowing outboard from the center of the paddle, circulating around and coming in behind the paddle.

The same thing happens to the air in the wake behind a sail. The deflected air on the windward side of the wake is shoved to windward and towards the top and bottom of the wake. The deflected air has to be replaced by air to leeward, and at the edges of the wake the air is flowing around and coming back in.

The planform shape, along with the camber and twist, determine how the lift is distributed along the span. The lift at a given spanwise station is proportional to how much the air is deflected at that station.

The amount of vorticity (swirling) shed into the wake is proportional to how much the lift is varying along the span. At the tip, the lift has to go to zero. So there's a lot of vorticity shed there as the lift goes from being substantial to being nothing. A triangular planform ramps the lift down over a longer distance, so the vorticity is more spread out. But it's still there.

For a given efficient sail shape, preferably rectangular at the top, and a given area, would there be any advantage to add winglets in order to increase the efficiency?

The minimum induced drag is obtained when the induced velocity (the self-generated header) is uniform along the span. This is true, even if the freestream is not uniform, or what the planform shape is, or if the surface is experiencing interference from some source such as the water's surface.

If you have a non-planar lifting surface, like a sail rig with winglets, the minimum induced drag is obtained when the induced velocity is uniform along each panel and proportional to the cosine of the dihedral angle of that panel. In other words, if the sail is perpendicular to the apparent wind and the winglet is perpendicular to the sail, then because the sail has a zero dihedral angle and the winglet has a 90 degree dihedral angle, the induced velocity (called the "downwash" in aeronautical contexts) will be uniform along the sail and zero along the span of the winglet - the winglet will be providing just enough lift to cancel out the induced velocity that would be swirling inboard to leeward of the head (or foot) of the sail.

When you calculate the optimum planform shape with and without the winglet, you'll find the lift distribution does not go to zero at the ends of the sail with a winglet, as it would have to do without the winglet, and the planform shape will be fuller toward the junction between sail and winglet.

Or the other way round, can winglets increase the efficiency in a way to allow for the reduction of mast height with equal performance to a non-winglet sail? I think, this is an interesting question, since a lower rig results in lower heeling momentums and therefore in faster and safer sailing for both mono- and multihulls.

There is, indeed, and interesting tradeoff here. If you design for the minimum induced drag for a given heeling moment, then instead of a uniform induced velocity distribution, the optimum is a linearly varying spanwise induced velocity distribution. It is stronger at the foot and weaker at the head. Here is what the comparison looks like for single planar surfaces:



The solid dot corresponds to a half-ellipse planform sealed to the water's surface. The span, gap, and induced drag are taken as a ratio to this case to show the sensitivity of different design variations. Every point on this graph represents a different design, optimized for the particular combination of span, gap, and induced velocity distribution, while producing the same lift.

The dashed lines correspond to the uniform induced velocity distribution, with variations in span and the gap between the foot and the water. As the span is made longer, the induced drag drops significantly - inversely proportional to the square of the span. The gap also has a significant effect, increasing the induced drag. As you'd expect, the center of effort is higher if you design for a greater span or larger gap.

The solid lines correspond to the case where the induced velocity varies from a maximum at the foot to zero at the head. This shifts the lift downward, reducing the height of the center of effort. For a given span, the induced drag is greater for the tapered induced velocity distribution than for the uniform velocity distribution. However, for the same height of the center of effort - same heeling moment - you can make the mast taller with the tapered velocity distribution and get less induced drag. So it really depends on what your basis for comparison is - fixed span, or fixed height of the center of effort.

The dash-dot line shows limiting case where the chord goes to zero at the head and the induced velocity actually reverses sign near the head. The upper trailing vortex is effectively being shed some distance below the head.

A winglet changes this tradeoff in detail, but not in degree. A winglet changes the effective span because you can always get the same induced drag of a rig with a winglet by making the rig taller. So the effectiveness of a winglet on the drag can be expressed as a multiplier to the span. The winglet will have a lower center of effort but the same drag as a single surface whose physical span is the same as the winglet's effective span.

You can add the winglet span and winglet dihedral angle to the trades shown by the figure above. If the sail span is kept the same, increasing the winglet span (for, say, a 90 degree winglet) will decrease the induced drag but increase the height of the center of effort somewhat because of the fuller lift distribution of the sail and the fact that the downward-directed lift of the winglet displaced to leeward adds to the heeling moment. So the trend lines will have a similar shape, but different slopes.

However, there's another factor for the winglet. If you keep the sail area the same and add winglet area, then the total wetted surface is increasing. This brings with it a parasite drag penalty. There may also be a rating penalty if the winglet area is judged to count as sail area. On this basis, you might want to ask the question as to what produces the minimum induced drag, constrained by both total area and heeling moment. That would keep the parasite drag the same.

Even if the winglet area doesn't count, so the basic sail area can be held constant, the increase in effective span you get with a winglet requires more winglet area than the increase in sail area required to extend the tip for the matching physical span. In other words, when you optimize a winglet's dihedral angle to minimize the parasite drag penalty, it turns out that the optimum winglet dihedral angle is zero - the winglet is a straight extension of the sail.

It would be an interesting tradeoff to constrain both the height of the center of effort (heeling moment) and induced drag, and see how the parasite drag changes with winglet dihedral angle. The span and induced velocity distribution would be changing to meet the constraints.