Foiler Design

An aircraft wing with a reflex profile does not require a tailplane stabilizer and is often used for "all-wing" planes. It is similar to that shown in the paper linked in Paul's post. It would be a more compact amd "marina-friendly" solution than a control surface mounted on an extension from the wing. I suspect such "all-wing" designs were less efficient as far as average lift per unit area although I have never verified that. Nonetheless it is another solution to the problem. Of course the tialplane is a rather antiquated solution to the stability problem in aircraft and technically not required for a wing-sail on a boat where the mainsheet -or perhaps several as in the junk rig - can theoretically do the job, given adequate sensing, computational and actuator power. We seem to be off topic, however.

 

To aerodynamically trim a tailless wing, the section must be reflexed enough to produce a positive pitching moment, assuming that the pivot axis is ahead of the aerodynamic center so as to be statically stable. This will generally result in a large reduction in maximum lift, and requires more wing area to produce the same lift. The increase in wetted area of the wing is greater than the wetted area of the tail that will trim the wing at the same lift.

Tailless wings also have low damping. This is a real problem when the atmosphere is gusty. The wing will be excited by the gust and oscillate. The oscillations produce lift to one side and then the other, and increase the drag. This can lead to some pretty scary behavior at the mooring. It is even possible for an oscillating wing to pitchpole backwards.

I hardly think tails are antiquated at all. In fact, I think they would be an advance in softsail technology. A ballestron rig combined with a tail could be self-trimming, tracking changes in the apparent wind automatically and alleviating gust loads without any input from the crew. And doing it all with no power required. It would be the ultimate cruising rig, and possibly effective in short-handed racing as well because of its ability to maintain optimum trim.

If you want to go with an actively stabilized system, which requires considerable power BTW, you could move the pivot axis back so the wing is statically unstable. That would require an increasing positive flap deflection to trim, resulting in high camber at high lift when you need it. Other than power and complexity, a drawback of such a rig would be total loss of control if it experienced aerodynamic stall. So it would have to be operated with a considerable margin away from stall onset. This, again, limits the lift one can get from a given area and makes tail control attractive.

I've experienced what the behavior of such a system is like. When some friends and I built a landyacht with a rigid wing, we initially rigged it up with a canard for control. We found that we could control the wing with the canard at low angles of attack. But the canard could easily drive the wing to angles from which it could not recover. If the wing stalled, the rig would swing broadside to the wind in an uncontrollable manner. An actively controlled wing without a tail could experience similar behavior.

Studies of aircraft configurations have consistently shown that when aerodynamic trim is taken into account, the minimum drag and maximum lift is obtained with an aft tail of moderate size, with a small but positive static margin. Here is one example by Ilan Kroo:

 

I was over at the Boeing Museum of flight the other day and wound up spending the whole time there staring at the different types of wing warping technology on display. Would that be a compromise worth considering? Sailors do it already to some extent-

Paul

 

Wing warping is important. A major factor in the success of Cogito and USA 17 was their ability to vary the aerodynamic twist.

The way to reduce the height of the center of effort with minimum impact to the induced drag is to vary the deflection of the air linearly along the span. This is approximated pretty well by linear twist. So when the maximum heeling moment has been reached, the lift can be increased by sheeting in and twisting off. There will be an induced drag penalty for doing this, but the increased lift offsets the parasite drag and improves the overall lift/drag ratio.

Cogito and the Canadian C-class cats twisted the entire wing. USA 17 achieved the same effect by only twisting the flap. Thin airfoil theory says deflecting a 50% chord flap is 80% as effective as rotating the whole section, and this is not far off for a slotted flap that maintains attached flow over the whole section.

Steve Clark said Cogito's victory over Yellow Pages was due in large part to their ability to twist the wing. So warping of some sort has been a big part of the competitive wing rigs of the last decade and a half.

 

Would the effectiveness of a two part segemented wing lie somewhere between a fully twisted wing and a twisted flap wing, esp. figuring wind shear? Or would the exposed edges of each segment create more drag than the advantages of rotating the top section would give you, even if the top section was above the layer of most shear?

Paul

 

It's not the exposed edges that are the problem, it's the step change in lift across the joint. This results in shedding a vortex into the wake, which increases induced drag.

Vorticity is shed continuously along the span. It's not just shed at the tips, which is part of the problem with the "high pressure on one side flows around the tip to the low pressure on the other side" explanation for the tip vortices. It's true that most of the vorticity is shed at the tips, but not all of it, and a lot of vorticity can be shed elsewhere. The strength of the vortices shed along the span is proportional to the slope in the spanwise lift distribution. So where you have a sudden change in the lift, meaning a steep slope to the spanwise lift distribution, you get strong vortices shed into the wake.

You can see this in action if you watch airliners on approach to landing in humid conditions. Water vapor condenses in the low pressure of the cores of the trailing vortices. Vortices can be seen trailing from the ends of the flaps as well as from the wing tips. In fact, the vortices from the flaps are typically more visible than the vortices at the wing tips, and often only the flap vortices can be seen. There is, of course, a big drop-off in the lift between the flap and the outboard part of the wing, and the strength of the vortex shed at the end of the flap can be calculated from that drop-off.

The split panels of the HarborWing rig would be effective in reducing the loads due to gusts and shear. However, they are not the least draggy way of adapting the wing to shear. They have the advantage that they can react quickly and passively, so it's a practical solution rather than the most efficient solution. For best performance, you'd want the lift distribution to vary smoothly, and twist does that. But twist generally has to be applied by explicit control, and is usually a slow process.

That's not to say that twist and discrete panels are mutually exclusive. For example, the flaps in each panel could be twisted so as to smooth out the variation in lift across the panel joint. Although they are separate, the panels can be operated so they are lined up and act like a single continuous wing. This might be the steady state trimmed position of the panels. The panels would still react individually to rapid gusts, and there would be a vortex shed at the joint at that time. But the vortex wouldn't be as strong and it would go away when the panels came back into alignment as the gust died. So there would be a modest drag penalty for the dynamic variation in the wind but not for the mean wind shear. The cost of the increased efficiency would be the added hardware and complexity of having flap actuators at both the top and bottom to twist the flap, instead of a single actuator per flap.

It all depends on the mission and how the dynamic response stacks up against the need for performance. HarborWing are getting a lot of experience in the real-world aspects of operating a rigid wing rig at sea. It's not surprising their rig is evolving to emphasize those aspects, while C-class rigs have evolved in the direction of performance.

 

DSS on skinny hulls/multihulls

Has anyone run into an application of a horizontal retractable lifting hydrofoil to a long skinny hull/ama? It would be used instead of curved foils when just
lifting force is desired. I'm considering the use of such a foil on a 13' ama (as part of an 18' tri) with the possible use of a very small "blister" on the very bottom of the ama to get the foil down as low as possible. Foil area would be about 1.4 sq. ft-slightly bigger than a Moth foil.
1) would you think a small "blister" on the bottom of the hull would result in significant drag?
2) foil would pivot like a centerboard and/or pivot in the center emerging from both sides of the "blister"/hull bottom.

 

What I think, is that the wave drag is probably dominated by the rate of change of cross-sectional area. In other words, the shape of the cross-section can be allowed to change in weird ways. If you put a blister, I´d make it long and fair it well to the hull. But I don´t see how you would retract it in this way?
Do you mean, swing it backwards into the blister, like a centerboard? Then there is a looong gap in the blister, right? maybe not so good. Is it for full flying? What approximate angle is it to horizontal? What sort of boat? There is a way to hinge a short cord board so that it is pulled up like a cross between a daggerboard and centerboard. That way the gap is shortened vs daggerboard, but collision kickup is still possible. I forgot the link where it was illustrated.

 

======================
Thanks for the reply,Sigurd. I envision it swinging like a centerboard OR pivoting: one side goes forward the other side aft to retract. The centerboard version was designed by Hugh Wellbourn for a large "Wally" monohull. I still have concerns about the foil being deep enough(hence the possible blister) but the advantage over curved foils of nearly vertical lift(approx. half the area of a curved foil?) might be worth investigating for the amas of trimarans. The axis of the foil could easily be accessed on deck with a line/drum system that would allow movement of the foil(s) from the cockpit of a trimaran. Sealing the slot properly would be a challenge but there are lots of centerboard boats to research for a good system.

 

Bend: how much is too much

Wow - no response in a few days. Did this thread finally die?

____________________________________________________________

OK, so Bill dragged my non-flapped foil down the tank, and it has quite a bit less form drag than a flapped foil - perhaps 10% overall T-foil drag reduction at similar speed, load and vertical strut immersion ( www.mothchronicles.blogspot.com )

So...

I went and modeled the bend using some old MIT lecture notes I found online, integrating the moment spanwise, assuming load scales with the local chord for an elliptical planform, to get deflection, knowing Young's modulus for my carbon, which I got by making a test beam and bending it (or having my friend Kirk bend it).

I have a spreadsheet that will calculate the total T-foil drag given foil area, span and section (Cd variation with Cl is included from the XFOIL curve). So I pushed the span up and the area down until the takeoff drag was the same as my current foil But it is a VERY thin foil, in absolute terms.

So how much bend at the tips under load is too much? I'm at 3.8" on a 46" span. I was thinking to build in the prebend, so it sails with a little anhedral under load. But conventional wisdom holds stiff is better. So are bendy wings slow? if so, I need to order some high modulus carbon, or go for more planform area.

Thanks,
K

 

From a static, hydrodynamic point of view, I don't think the bend will make much of a difference.

Dynamically it might be a different story, especially if it twists as it bends. If the lift is changing dynamically, that will shed starting vortices into the wake and increase the drag. Depending on how it twists, the spanwise lift distribution could be affected and increase the induced drag.

Aircraft wings are designed so they wash out as they bend. This reduces the load on the tip, shifting the load inboard and stabilizing the bending. If the wing twisted so as to increase the angle of attack at the tip when it bent, that would lead to divergence when the speed was high enough, and require a lot more structure to stiffen the wing against it.

You could do the same thing with a foil. If the unidirectional fibers were swept forward somewhat instead of running straight down the axis of the foil, it will twist as it bends. How much twist would be useful (if any) will depend on the particular configuration, but it might be worth making an experimental foil or two to see the difference.

 

Thanks for your thoughts Dr. Speer.

I modeled a truly elliptical planform, because that is what my current foil is. Most of the bending occurs where the moment is greatest, which is at the root. The chord will actually be increased there for an intersection fairing, which will reduce the total deflection, but also control twist I would think. There is no sweep at all.

I will go a bit more conservative on the area to increase absolute thickness, switch to high modulus prepreg, beef up the root and see what happens. That's the nice thing about Moths - the landing is generally pretty soft when things go wrong. Worst case scenario is an unstable (or slow) foil, but it doesn't seem like there are any ways to absolutely control deflection or twist, so the real world becomes the test tank. Maybe I could put some of those optical sensors in the layup to get an idea of what is really happening down there under load...

Regards,
Karl

 

Foiler Design-stability paper

Here is a paper doing an analysis of the Moth Hydrofoil performance by stability and force balance criteria:

 

Twisting can be modelled using a hydro-elastic FEA analysis. If the pressure distribution can be calculated and Karl can provide the foil geometry I can do such an analysis. If you have some idea of the washout required I can trial a few laminates to see if the required twist is acheivable. This is also used on wind turbine blades to depower them in high winds. Peter Schwarzel

 

The usual problem with timber is strength. A very good grade of timber will have a flexural strength of Sf<100MPa. Whereas a good carbon laminate will have a Sf=>1400MPa. A good infused glass laminate will be +1000MPa and this can be done in your garage. The second criteria is stiffness. The timber will have a E=15Gpa but the Glass will be 31GPa and the std carbon 100GPa. Since a foil has to be made thin and stiff the carbon is the usual winner. High strength aluminium could be a contender with E=70GPa and strength ~300MPa. But making it hollow (to get to target weight) would be a good trick. If you did it in 7075 alloy then you can get this to 600MPa. Early wind surfer masts were of this alloy and they were hugely strong. In one of my early mast jobs I had to straighten wind surfing masts and I had to inform the owner that they had a 50/50 chance of me breaking it vs straightening it. Plus 7000 series corrodes really fast. The answer is to get familiar with infusion technology and then you can build glass or carbon laminates of autoclave quality at home. Another contender is kevlar or PBO if you can get the fabric in a suitable form. The usual problem with these types of fibres is the poor compressive properties. But this is because the fibres are less dense then the resin so they float in the resin and don't consolidate well. This produces poor hand laminates even when wet bagged. But if infused so the dry stack is consolidated before the resin is introduced the fibres are well packed and do great in compression. Our infused glass laminates test stronger and stiffer in compression than tension! This is in some ways due to how std tests are done but at least we dont have to knock down the compressive values so far as usual. It would be good to know if the foils are failing in direct compression or by buckling and of course what the laminate properties (or ply properties) are. Peter Schwarzel