Foiler Design

Main foils about same surface and aspect ratio.
In a couple of days we will have some pictures and a video on
www.moth.it
Marco.

 

Hi I am on page eight now..
It seems confusing to make such a qualitative difference between the two foil systems, uni and bifoiler. The canard does the same as the front "bifoil", just to a different degree. same goes for the uni "main foil" and the bi rudder.
Since noone so far (page eight) is experimenting with an auto-actuated rear foil, (I don't see why they would?), I'd concentrate on elongating the "wheelbase" to decrease the induced drag from pitching the boat (would also let the foils act within a smaller range of load, and decrease the effect of the rig on the helm), and see where it got me. Like, counter-rotating rudders fore and aft, wand actuated front foil for instance? I'm sure I will come to it during the night on one of these pages. Pretty good reading

You are leaning these foilers quite hard into the wind, would it be feasible to let the rig to leeward untill upright?

Tspeer said:
"A three-surface arrangement is a good idea. It lets each surface specialize in doing one task well - lifting, pitch stabilization and trim, heave stabilization and trim."

Since all those four tasks really are just lifting, I don't see why you need three surfaces?

And:
"Absolutely. I rather like the three-surface approach as a way to optimize the whole configuration without having to compromise any of the foils too much."

Can you elaborate how they would become compromised?

And: "I was talking about the static stability of a canard and main wing, with no aft surface and with the weight fixed in position. Say the two surfaces have the same aspect ratio. If the canard is trimmed at a higher lift coefficient than the main foil, then a change of, say, one degree in angle of attack will result in a smaller % change in the lift on the canard than it will on the main foil. This will result in a negative pitching moment, which is stabilizing."

I think you are thinking about planes too much. This canard in question *follows the surface*. It must be made so that only a very rare occation will force it under its designed depth.. the relative wing loading doesn't have anything to do with this. The canard is ment to track the surface either by having enough AOA and being static (planing) - you could make it as big as you wanted, it would not affect the stability. or it is actively guided by a sensor, in which case the same applies.

And: "The lift from all the foils should result in a uniform downwash left behind in the wake. If that is produced by more lift on the middle of the main foil and some negative lift on the aft foil, that's OK."

but if there wasnt a wing pulling the opposite way the other wings could be smaller.

 

Downwind Only Foiling

Seems to me that foils ,particularly the bi-foil arrangement, could be applied to a wider range of monohulls if the idea was to only foil downwind. Foil sizes could be reduced possibly making it easier to retract the main foil on a keel boat by having it simply pivot into the keel bulb in a similar way as the kFOIL does. Or perhaps the foil could have an adjustable angle of attack and be used upwind on a canting keel to add lateral resistance(as per Andy Dovell's wings) and then, with the keel locked, offwind as a hydrofoil.I'm thinking this might require on deck ballast as well.
Rereading the Australian Sailing article about the pioneering work David Lugg did with his I14 started me thinking about it since the design road he chose in 1999 was to optimize his 14 foils for downwind work. This resulted in a mainfoil with an area of about .8 sq.ft. as compared to around 1.04 sq. ft. for a bi-foil
Moth--this on a boat weighing over twice what the Moth weighs.
So you'd round the weather mark, pop the spin and set the foil. Seems like an interesting avenue of foiler design to look into more closely.

 

Just for the record re the development of David Lugg’s int 14 foiler
David Lugg writes (july 2006) “Our initial aim was never to fully fly at all. It was only after the centreplate foil angle adjustment failed that we realised the power of the main foil and then decided to add a rudder foil. I intended using a fully rotating rudder foil. I recalled Alans had been involved in aircraft design and anticipated that he would be able to assist us. Alan insisted on the flap arrangement on the rudder foil, which I think was a very sensible choice for us at the time”.


At the time David sort my assistance he was unaware of the real stability issues. Other foilers were surface piercing with inherent height stability coming as a result of foil area reducing as a function of height. David had considered the International 14 rules and concluded that only fully submerged foils (span less than that of the hull beam) would be acceptable. While aware of the wand control option, David and I reasoned that autopilot control would not be accepted by racing regulators. While I do not agree with the related decision allowing auto control of foils on Moths, history shows we should not have prejudged the decision. But so be it!

I knew that fully submerged non articulated foils would not be sufficiently dynamically stable for the fourteen to stay “up” any longer than seconds. Hence some form of man in the loop control was necessary. We chose a conventional aircraft control with pitch attitude control via an elevator on the rudder foil. The mechanics of this mechanism were developed jointly by David and Brett Burville ahead of John Illet using the same system on a Moth. And the real crunch; for man in the loop control the CofG needs to be slightly ahead of the combined centre of lift of the foil set and the Cof G must be sufficiently aft to balance the large nose down pitching moment generated by the forward thrust of the sails when you take moments about the centre of lift of the front foil. The result is a large rudder foil.

At the time my data on foil performance was scratchy to say the least ( not so today) and we did not believe foiling to windward would be possible in normal racing winds (mistake 2). Hence we chose a main foil area suitable for downwind foiling and small enough not to greatly reduce the upwind non-foiling performance

alans http://www.highspeedsailing.com

 

Probably this has been posted before, even by Tom Speer himself. But just in case, here it is:
http://www.basiliscus.com/CSYSpaper.pdf#search="rolf eliasson paper stability"

 

thought so, read the top, it is written by Thomas E Speer himself

 

FWIW, I think this is a good example of why it's important to look at a foiler design numerically before committing to building it. It would have been a very disappointing perfomer had the design been build as shown.

Basically, the results I had at that time say that the takeoff speed is too low and the foils are too large. The ladder foils aren't efficient because the lift distribution is way different from what most people assume. It's not until the craft is flying high on the foils that the drag becomes competitive with the narrow displacement hull alone. So the performance would generally be much better with the foils up than down!

It needs less foil area and a cleaner design. Here's my current conceptual configuration, but I haven't analyzed it for performance or loads, yet.

 

That was my impression of the early design.

Just curious, what is the state of your plans to build a foiler? Are you still thinking about it? If you have put the idea on a back burner is there any particular reason?

 

Still thinking about it, but not actively working on it. I still don't have a design that meets the three principal design objectives: cruise, fly, and be affordable. Especially the last one. It's almost certain that I won't build a whole boat from scratch - I'm just not saving the money fast enough to make that happen in the next 10 years, and by then I'll be too old to do it. If I build something, it will probably mean buying a used boat and designing foils for it.

But as much as anything, I just don't have the time - too busy at work, and the project I'm doing now is scratching the itch to design something from a clean sheet of paper.

 

http://www.highspeedsailing.com/Sailing/Moth/Moth configuration.htm

What is the reason for that?

edit: is boogie's website down, or just moved somewhere? did he stop making fins?

edit: alans and wardi, on the above site a surface piercing inverted V moth arrangement is shown. is it intended to be flapped? would it not be hard to balance in roll? is it in addition to other foils?

 

He's trying to maintain the same 3D lift curve slope, and thus the same pitch "stiffness" (change in lift per change in angle of attack) from the foil. This would leave the stability characteristics unchanged.

2D section characteristics:
alpha_2D = angle of attack of the two-dimensional section (deg)
Cl = 2D lift coefficient
ae = 2D lift curve slope (per deg)
Cl = 0 at alpha_2D = alpha0
Cl = ae * (alpha_2D - alpha0)

A foil with a finite span flies in a downwash of its own making, due to the wake it sheds. The shorter the span the stronger this downwash is, because the foil has to accelerate a smaller mass of water downward at a greater speed to maintain the same lift.

If you reference the angle of attack to the freestream flow direction, this appears to be a loss in lift for the 3D foil at the same angle of attack compared to the 2D section characteristics. The downwash angle of attack for an optimally loaded foil operating deep beneath the surface is constant along the span and equal to

alpha_downwash = CL / (pi * AR) (radians)

(The downwash is doubled for a foil operating at the surface.)

The foil has to operate at a higher angle of attack to make up for this, so

alpha = alpha_2D + 57.3 CL / (pi * AR) (deg)

The foil is tipped up at an incidence of alpha, but locally the flow is coming at it at alpha_2D because the local flow is actually moving downward, not coming straight on. It's really a choice in force bookkeeping to call this effect a loss of lift rather than a reduction in angle of attack. It could be legitimately done either way, but booking it as a lift loss is the conventional approach.

CL = 3D lift coefficient
a = 3D lift curve slope (per deg)
CL = a * (alpha - alpha0)

AR = aspect ratio
f = correction factor for taper, etc. (f is > 0.98 for all but diamond or delta planforms, so is usually taken as 1)

a = f * ae / [1 + 57.3 * ae / (pi * AR)] (a, ae per degree)

a = f * ae / [1 + ae / (pi * AR)] (a, ae per radian)

Since the theoretical value of ae is 2*pi per radian, this results in the 3D lift curve being approximately

a = f * ae / (1 + 2 / AR) (a, ae per degree or per radian)

qbar = dynamic pressure (1/2 density * velocity^2)
Lift = CL * area * qbar
Lift = a * (alpha - alpha0) * area * qbar
Lift = f * ae / (1 + 2 / AR) * (alpha - alpha0) * area * qbar

Lift = weight, so the lift has to be kept the same between the two candidate foil configurations. If the trim angle and stability characteristics are to be kept the same, then you want to produce the same lift at the same angle of attack. So you want to maintain the following ratio, which is formed by moving all the stuff that doesn't change to the left-hand side:

Lift / [f * ae * (alpha - alpha0) *qbar] = area / (1 + 2 / AR)

Multiplying top and bottom by aspect ratio, yields:

area / (1 + 2 / AR) = AR * area / (AR + 2)

If you want to keep the stability the same at the surface, you'd want to keep [AR * area / (AR + 4)] constant. So you're not going to be able to match the stability both deeply submerged and when running at shallow depths when you change the foil span. The wider foil is not going to suffer as much lift loss at the surface, and this can affect the heave stiffness of the foil. This can be compensated for by adjusting the "gain" - the gearing between the wand and the flap. The large-span foil will need more flap deflection per degree of wand movement than the shorter foil to maintain the same height control response.

 

Ncrit

Tom,
I saw on sailinganarchy (in a post back in 2004 when you were an anarchist ) that you used Ncrit = 3 in xfoil. But I also read here http://www.esotec.co.nz/condorkeel/Html/update_F.html that Ncrit = 1 should be used.

What is the correct number to use? What's the effect of using to high / to low? (for a blue water cruiser)

Thanks!
Mikey

 

p.s. The implication of this strategy is a reduction in foil area as the span is increased.

Baseline foil planform:
b1 = baseline span
S1 = baseline area
AR1 = baseline aspect ratio
AR1 = b1^2 / S1

New foil planform:
b2 = new span
S2 = new area
AR2 = new aspect ratio
AR2 = b2^2 / S2

The area ratio of the new foil to the baseline foil is:

S2/S1 = (b2/b1)^2 / {(b2/b1)^2 + 2*AR1*[(b2/b1)^2-1]}

If the span is increased, b2>b1, the terms in the denominator are all positive, and S2/S1 < 1.

Basically, the longer foil is more efficient and can be loaded more heavily while still maintaining the same stability characteristics (pitch stiffness).

The increase in span reduces the induced drag - especially important when running near the surface and at low speeds. The reduction in area reduces the parasite drag, which is especially important at high speeds. The limit to increasing the span is reached when the foil becomes too flexible or the area is so small that the foil stalls out at low speeds.

 

Ncrit is used in the boundary layer calculation. Thie idea is disturbances in the boundary layer start off small, and can be either damped out or amplified, depending on the boundary layer characteristics. When the flow is accelerating, the disturbances tend to be damped, and they are amplified when the flow is decelerating (positive pressure gradient).

If the oncoming flow is smooth, the disturbances in the flow are small, and they can be amplified a great deal before they are big enough to cause the boundary layer to be turbulent. If the oncoming flow is fairly turbulent to begin with, then less amplification can be tolerated before the disturbances get to be too large for the boundary layer to remain laminar.

If the surface is very smooth, there aren't as many new disturbances being introduced into the flow. If the surface is rough, the roughness itself introduces disturbances that are then amplified. So transition occurs earlier with a rough surface.

XFOIL allows you to change ncrit to model the freestream turbulence level and the surface roughness. Transition is assumed to occur when the amplitude of a disturbance in the laminar boundary layer grows to e^ncrit times its original amplitude. The other way to think about this is the initial disturbance is e^-ncrit times the critical disturbance level at transition.

There's no "right" value for ncrit. It depends on what kind of situation you have and how conservative you want to be. There are lots of these kinds of tuning parameters in engineering. If you have a lot of data, like an AC campaign would collect, then you'd fiddle with ncrit to best match the data for your application. Otherwise, you pick what you think is a reasonable value, and make some runs with higher and lower choices to see how sensitive the results are about your nominal choice. I think you'll find there's not a lot of sensitivity at the lower values of ncrit for most sections.

The default value for ncrit is 9, meaning the initial disturbances are exp(-9) times the critical level. This is quite small (0.01%), and represents very smooth air and a very smooth surface. It's about the best you can hope for, and XFOIL results using ncrit=9 do a good job of matching experimental data collected in modern, high-quality, low-noise laminar flow wind tunnels. The wind tunnels used for the original NACA airfoil data were not very good by today's standards, so a smaller value ncrit would be appropriate when matching those data.

When I use ncrit=3, I'm saying I'm assuming the real-life initial disturbances are 400 times bigger than in the best wind tunnels. I don't have any data to justify this - it's advice I've gotten from the net, too - from Dr. Drela on the XFOIL list, IIRC.

If you use ncrit = 1 instead of ncrit = 3, you're saying the initial disturbances are another 7 times greater. Or maybe you're saying the initial disturbances are the same size, but your surface is rougher - not a bad assumption if you're talking about a blue-water cruiser keel compared to a high-performance hydrofoil.

 

Thanks Tom, good explanations as always

I was thinking that a blue water cruiser will be "in the groove", keeping the keel clean would not be my top priority... and the keel should still be able to do what it has to do - also in rough weather and when not sailed to its optimum speed. Ncrit = 1 sounds realistic for those conditions.

I plan a rather large keel area and would therefore prefer a rather thin foil (I get ~60% ballast with a NACA 0008 lead keel ). I don't want to use NACA 000X and is looking for checking Eppler (E1161-24 and 28 and 1160- 28 and 32 series). Problem is I can't find any information on the net.

Do you know of any Eppler generators or where I can find coordinates?

Thanks
Mikey