Foil performance discussion

I beg to disagree. Our differences are largely semantic though. You claim: "isn't being reduced as fast" should be: 'isn't being reduced as much'. 'Fast' would relate to changes in the rate of acceleration. With this one word change, your claim is correct: but only when expressed relative to the at rest attack angle (absolute angle). My claim was relative to the current attack angle. Vector addition of the forward motion and upward speed determines the instantaneous attack angle. So, if the upward speed reduces, ceteris paribus, the attack angle becomes less negative (increases).

Yes, you are correct. I'm sorry if my analysis seemed critical. I guess I was focusing on ways the foil could perform more stably. The application itself is stable, but, moving toward full foiling, which is my interest, would I believe, require more.

I'll post some of my xfoil analysis of proa sections today if I get a chance.

Sorry about the delay in replying. Work and boat building make for long days and my father had a heart attack in the intervening time. He's doing alright now.

 

In this analysis I'm comparing a thin, fine entry, Speer inspired foil with the Eppler 817. Reynold's number is a bit high for small craft at 4m. I'll run a lower speced analysis if requested. I'll also expand the discussion when I have time, but for the moment, a few salient points:

First, compared to Rob's 12% section, these are only 8.85%, with 3% camber. Rob, your ogival section has rounded tip what is the radius? I request this in order to calculate the camber. My suspicions are that it is very high.

Secondly, comparing the 2nd and 3rd attachments, note that for zero alpha, the suction peak occurs at 50% chord. At 2.45 degrees the suction peak is flattened. from this, proa sections will perform better at an angle of attack that flattens Cp and raises Cl/Cd. Your 12% section I suspect is generating a Cp graph peaked at 50%, with much higher suctions than this board.

Thirdly, ogival with radius didn't perform as well as the section I trialed here. The suction peak couldn't be flattened.

I'll add some more info later.

 

Hi all,

I'm new to this forum. I'm the kite part of Rob's proa experiments and have been doing some analysis of the hydrofoil setup. I'm flat out with work at present but will make a couple of quick comments {OK, it's got longer than intended...but if you saw the work I should be doing.... }. At this stage I've only been considering the large scale geometry of the setup, not, for example, the section shape - that will obviously need further analysis.

As indicated earlier, the geometry is such that vertical lift is determined by the lateral load from the kite. That is, leeway adjusts such that the horizontal component of lift balances the lateral kite load (ignoring the lateral resistance generated by the hulls). For a given foil angle this determines the total lift vector (normal to the board) and hence determines the vertical lift. As also noted earlier, as the windward hull starts to fly, the board angle (cant/dihedral) changes. As the horizontal lift does not change (ie it is balancing the same lateral load from the kite), the vertical lift increases. So, before considering end span or near surface effects, this would be quite unstable - it would keep rising until leeway was such as to cause stall. I think I was even more surprised than Rob to see quite benign stable performance when we first flew it. So I conclude that end span and near surface effects are significant.

3kN is an upper limit on any loads that the type of kites we're using have generated. (Having said that, we're experimenting with quite different kites which we hope will have a much higher upper limit and provide several other benefits.) I suspect we may have seen loads of up to 2kN and would generally be sailing with less than that. Using Doug's numbers from post #7 (45° elevation, 15° azimuth from beam, foil angle 45°), 2kN would generate twice the vertical load he calculated using a 100 kg (1kN) kite load. So say 1.37kN (ie 137kg weight - forgive me, I prefer to work in N than kg weight and use g=10 for this). Although this is still a bit under Rob's advised mass that is being lifted (180kg), we're really talking about torques here rather than vertical lifting. Consider torques about the lee hull and bear in mind that
a) the centre of lift of the foil is a fair way to windward of the hull and
b) the estimated weight is a at least slightly to leeward of the windward hull, and
c) the horizontal component of the lift vector is contributing some lifting torque.

So somewhat more than 1.37kN force on a longer moment arm than the 1.8kN (180kg weight) weight seems about right. I have done the numbers accurately and it all makes sense. With a few different parameters I estimate a lift coefficient at take off speed (about 4m/s) of 0.18 (span 1.4m, chord 0.3m, foil angle 60° to vertical). At "cruising" height and speed (7m/s, span 0.5m, foil angle 45°) this is only slightly higher at 0.2. These are quite rough estimates but seem to be in the right ball park.

The configuration we have is very attractive for its simplicity and, dare I say, elegance. A single underwater appendage to provide all of flying (ok, it's only one hull!), steering (with the aid of the kite) and lateral resistance has considerable attraction. However, we're considering various ways to improve the downwind/reaching performance and mitigate the heave instability that would arise from high leeway at higher flying angles. One solution is to add a rudder. Although I've a lot more thinking to do on this, the basic aim (apart from steering, which would be a bonus), is to control leeway (and hence lift on the main foil). By raising or lowering the rudder you would effectively be trimming the main foil by adjusting leeway.

When beating the rudder has the effect that as the hull rises, the proportion of lateral lift borne by the rudder would increase. (Although the rudder angle would change with the rotation of the leeward hull, near which it is mounted, this is a small effect as the vertical projection would be the cosine of the heel angle. So I discount this effect in this discussion.) As the assumption is that total lateral lift is constant, lateral lift from the main foil decreases as it flies. Although the vertical proportion of the lift from the main foil still increases with flying height, dynamic analysis reveals this arrangement to be heave stable. Moreover, for given sailing conditions, the stable height is set by the amount of rudder immersed. The rudder taking more of the lateral load clearly also implies better upwind angles.

For strong winds the rudder would be fully immersed whereas for lighter winds a sweet spot may have relatively little in the water - and experience shows that, at least in some conditions with the current foil, no rudder is necessary.

For downwind sailing where we don't fly at deep angles currently because there is insufficient lateral load, the rudder oppose the upwind lift from the main foil. That is, currently we would have negative leeway when going downwind as this is the condition that matches the low lateral kite load. The rudder would mitigate this effect, hence retaining sufficient lift from the main foil to fly....at least that's the plan. There is a clear cost here in that drags from the opposing lifting surfaces would add.

Several other foil configurations have been considered. L, J, inverted L (effectively what we have but with a vertical cap like l'Hydroptere), both to leeward and windward. At present I'm leaning towards the straight canted foil to windward with a rudder for relative simplicity and ease of trimming. But I'm very interested in any other thoughts. (Sorry this has got to be quite so long.)

 

Done it again

 

This would indeed be the case if the lee hull flew as with SailRocket. Given the lateral resistance of the slender lee hull, the foil kite forces can go out of balance. You hint at this later when discussing negative leeway. So even when reaching, leeway might be a little negative but virtually undetectable, because the foil is producing excess lift. (Edit: this might appear at odds with my final remark, but the previous sentence must be read in the context of the foil having produced extra lift, generates negative leeway. From ensuing discussion this reduces the attack angle, reducing and modulating the foil lift).

Leeway, if sufficient, would lead to a stall. As I observed earlier, the attack angle will increase with leeway. Given that the foil is fixed, i.e. not rotated between shunts to increase attack angle, the angle starts at zero. Refer to attachments 2 and 3 in my previous post. At zero alpha the pressure distribution is peaked at 50%. At an attack angle 2.45 degrees, the pressure distribution is flatter. The foil is more efficient. The pressure peak has increased by less than 10%. Referring to attachment 5, the two upper right graphs show that the Cl and Cl/Cd roughly double as alpha increases from zero to 2.45. In the case of a thicker foil (12% as opposed to 8.85), the efficient angle is likely to increase to 3 or 4 degrees.

So, attack angle stall is not likely to be the means by which the foil loses lift. The lift increases with leeway. The principal heave modulator is reduced lift from reduced foil immersion area. You get a double whammy effect as the craft heels and foil immersion reduces. The leeway resistance diminishes faster than the lift.

I can't speak with certainty concerning your foil as I only know that it is ogival with flat under side, 12% chord, with leading edge radius. Without knowing the radius, I can't calculate the camber. However if it is a true ogival 6% camber (6% due to flat bottom - 12% camber if thin plate section - concave underside), Cp crit at 50% is likely to be in the vicinity of -1. Refer to the critical pressure coefficient graph in the link below. Cp of -1 will cavitate at 27 knots. I understand you aren't there yet, but this indicates why the foil is ventilating at 50%. The suction is high, close to cavitation. The solution is firstly a much less cambered foil, and secondly, 1 to 2 degrees of attack in order to make the foil more efficient (flatter pressure).

http://www.mh-aerotools.de/airfoils/hydrofoils.htm

So, with all the preceding in mind, as speed increases, the lift provided by the foil increases (not necessarily coupled to the kite pull), modulated by reduced surface area, aspect ratio and increased ventilation. Leeway is also likely to shift from negative to positive as speed rises. Is failure to fly downwind entirely due to the lack of sideways kite pull? I am inclined doubt it. Perhaps the kite doesn't have the power to shift the apparent wind far enough forward to increase speed to take off? Certainly negative leeway will result in a negative attack angle and less lift. The solution downwind might be to decrease the vertical angle of the foil. The rudder idea makes sense, but I'd still consider adjusting tilt as you don't need the drag of opposing foils to modulate directional stability until you are fully foiling.

David

 

I'm keen to see you get this working. I'm particularly interested in these symmetrical Tom Speer inspired foils for my proa. Stienar's and the Luca Antara 20m proas will require this type of foil. Tom has expressed his doubts re the actual performance of these largely untested foils, which is why I'm exploring the idea of a finer entry foil.

 

David : I've been assuming that rob meant "top" (the low pressure side), but misspelled it "tip." If so, that makes the camber 6%, which is awfully large. It possibly would need negative angles of attack to get its usual operating CL's. So, I agree that your 3% camber section is probably a lot better.

 

The rounded tip refers to the wing tip - ie at the end of the span. The LE/TE is very sharp with no measurable radius.

 

David,

Thanks also for the earlier foil analysis output and comments re the flatter suction curve for 2.45° alpha. The higher efficiency at moderate alpha suggests we should pursue the idea of being able to trim the board (other than with leeway). We have the beginning of a plan in that regard.

I accept that the lee hull is providing a fair bit of lateral resistance. We manage surprisingly well upwind without the foil. So it's functioning in a similar manner to my description of the rudder's leeway moderating effect. However, the long goal is to have the lee hull planing (notwithstanding the narrow "stern") and very lightly loaded (with displacement lessened by the kite, judicious loading and moving the foil further to lee) - so, although I'd like to get away without a rudder, it may yet be helpful.

Although measuring small negative (or positive) leeway is difficult, I think the negative leeway is the main thing stopping us flying downwind. We've had speeds downwind, even with all the drag from the foil and it's not-at-all-fair case, which are rather greater than our foiling speeds (take off) upwind. Varying cant angle should help this - eg 30° from vertical rather than the 60° we have. But that then requires quite a different case, introduces higher cantilever loads and produces quite a shallow inclination at the water surface (eg 20° if the vessel is heeling 10°). I remain unsure how important surface penetration angle is in terms of ventilation.

A better downwind solution would be to go fast enough to have the apparent sufficiently close to the beam to retain the lateral load that gets us flying upwind. That is, we want to only sail upwind relative to apparent. We'll be working on this next time we go out. It may mean that our downwind angles aren't as deep as they could be - but VMG may be better.

 

As far as foil configurations go, I've consider many for the windward hull. I referred to an option I'm strongly considering in an earlier post:

If the T foil was in a pivoting case, it could be raised and lowered to set the desired level of immersion. The tilt would balance leeway. Having a foil tip pierce the surface in order to modulate leeway I anticipate would result in far less ventilation at shallow angles. I know this adds complexity, but a T foil is undoubtedly simpler to accurately build than a curved foil. Additionally, the cantilever loading I expect would be reduced in all conditions.
The angled board is simple and shows promise on most points of sail in most conditions. Perhaps the T foil, or some other idea, will prove simpler than adding a dagger / rudder to leeward?

 

5 foil comparison

In this post, I've compared 5 foils in order to explain my thoughts on leading edge shape and camber.

Tom Speer has observed:
"The blunt trailing edge sections analyzed in this paper barely met the assumptions behind XFOIL's methodology. Therefore, the results should be treated with some caution until confirmed by experiment." http://www.basiliscus.com/ProaSections/Paper/ProaSections.htm

Empirical evidence is scarce to say the least. The only references I have found are from Kevin O'Neil and Rick Willoughby:

Kevin O'Neil
http://proafile.com/multihull-boats/discussion-forum/viewthread/429/
"Laurent’s implementation of the Speer section is great. It never flutters, and the boat goes to windward really well. I left it vertical and sailed down onto near a beam reach several times to see how it acted at higher speeds, and it never fluttered. I think the fastest I got it to was eight to ten knots of boatspeed."

Rick Willoughby
https://au.groups.yahoo.com/neo/groups/harryproa/conversations/messages/10467
"The 18m proa I sail on has a bi-directional cambered swinging dagger board. It is 2m deep and chord is 500mm. It is fibreglass over an aluminium former. The nominal section is a NACA 6516-15 mirrored - edges are nominally 4mm radius. There is no discernible vortex shedding. The thickness is based on strength requirements and cavitation is not a consideration upwind in this boat. The dagger board improved windward VMG by 10% to 15%. Pointing ability by more than 10 degrees. The board has a limited working range. Once the boat is more than 60 degrees to the apparent wind the board starts pulling the boat to windward. That slows the boat compared with just using the rudders. This is a brief clip of the board looking down through the deck slot during testing before the lateral brace was faired:
https://www.dropbox.com/s/cjaexyps6wqgu2i/100_2307.MOV?dl=0"

I am not sure of the top speeds of the 18m proa but believe it has done 15 knots. With this foil? I'm not sure. Given that it is not clear that the blunt trailing edge will perform well at higher speeds, I have focused upon designing a lifting foil with a much finer entry. The challenge, as I saw it, was to develop a shape that had high Cl / Cd within a sufficiently wide Cp min bucket.

The end result is a thinner foil than a conventional lifting foil like the Eppler 817. The necessarily finer leading edge would entail more top camber and consequently, increased suction, if it were to be as thick as the Eppler. The alternative is increased camber on the bottom surface as with the Speer 3 series proa boards, but I found that this solution had lower Cl /Cd.

Referring to the large graph in the first attachment, the Y41z 885 and the P30885 mod have the widest operating range to Cp min of -0.75: wider than the two Speer series 3 boards. The P30208 has the lowest suction value, being thinner, less cambered and blunter. Refer to the second and third attachments to view the foil and leading edge shapes. The a5, thinner and finer than the Y41, has a narrower Cp min range. This is due to the Cp peak at the fine entry, suggesting diminishing returns to sharpening the entry. Yet this quote from Tom Speer suggests that it might be possible to get away with this:

"If Ncrit really is lower for hydrofoils, this leads to earlier transition in the bubble and less succeptibility to laminar separation than for airfoils. You might be able to use that to your advantage by allowing your leading edge to be a bit more peaky than you'd otherwise be able to get away with."
http://www.boatdesign.net/forums/boat-design/xfoil-hydrofoils-2909.html

Referring to attachment 1 bottom left, Cl / Cd is mostly significantly higher for the finer entry boards. The intermediate blunt entry of the P30885 mod foil is almost as high as the finer entry boards and marginally outperforms the Y41z on Cp min. This suggests the board is a viable option if the blunter edge performs ok in practice. I suggest that these results hint at blunt trailing edge drag and drag from the blunter, more cambered underside of the Speer boards.

The P30885 requires a disclosure: the upper surface is that of the P30208 and the lower comes from the Y41z. The combination has then been scaled to 8.85% to match the Y41z.

With the limited empirical evidence, I am cautiously optimistic that the Y41z can perform. Whether it will perform up to its cavitation speed limit in the vicinity of 37 knots is certainly unknown. The entry is finer than the 0.8% radius on the 18m proa, so with a little fairing, sharper entries could be trialled and discarded.

The thin foil will potentially place engineering constrains on lifting foil aspect ratios. This is another reason why I lean toward a T foil for this application: reduced and largely balanced spans. Rotating the foil to leeward can also unload the stem of the T (other than when surface piercing), with the span of the T oriented to balance lift and leeway. A solution for the simple angled foil is to reduce the thickness and camber along the foil. The thicker more heavily cambered upper section is better suited to resisting the cantilever load and will be out of the water at higher speeds.

The final attachment shows the P41z performance over a wide range of Re.

 

Thanks for the thoughts on the canted T, David. Having not been familiar with the examples you gave I didn't initially follow the suggestion. Now having seen the Sunnucks Vampire I see what you mean. Interestingly my latest thoughts were not so different from your suggestion: the angled foil starting from a beam, midships or closer to the lee hull, and having a support strut from the beam near the windwrd hull and meeting the foil near the water surface at cruising height - the strut would be the stem of a canted T if the foil didn't extend as far as the beam.

Both mounts on the beam would rotate. This would allow
a kicking up in case of a collision;
b AoA to be trimmed; and
c retraction in light air.

You mention the possibility of having the T foil in a pivoting case - I think you're referring to pivoting about a roll axis. This would of course be nice as well - but I think we'll stick with something simpler (ie one axis rotation rather than a universal joint) in the first instance. You've also suggested that "Having a foil tip pierce the surface in order to modulate leeway I anticipate would result in far less ventilation at shallow angle". I don't have much feel for the importance of surface penetration angle for ventilation. I'd thought steeper would be better....but perhaps if the fence (stem of T) can be at the right point, that's less important.

With the board on the beam, we'd no doubt have 2 boards. Mounted on the beams also means they can be further to leeward. Positioning could be such as to optimise the amount of vertical lift. Mounted under the hull, the moment arm about the lee hull is necessarily quite long (and increases as the hull flies and the more distal parts of the foil provide the lift). This means that more weight is transferred to the lee hull and the goal of a minimal, though positive!, displacement lee hull is not achieved. If the foils are moved closer to the lee hull, the (net) displacement of the lee hull can be optimised. (With the kite also providing vertical lift, one has to be careful not to get into negative displacement mode! I don't think that would end well.)

Being able to trim the board could remove the need for the rudder to modulate leeway. That is, one could rely on leeway resistance from the leeward hull. For example, when reaching with low lateral loads, the board might be trimmmed for higher lift so that the lee hull would oppose negative leeway from the board and thereby maintain an adequate AoA to provide sufficient lift to fly.

Thanks for the section analysis. Unfortunately I'm struggling with a busy work period presently (just for another 2 weeks). If I start considering sections now I'll get carried away and the work won't get done - but I'll look forward to looking at this soon.

Trevor

 

Do you worry about the winch between your legs?!

 

Rob, amazing cool video and boat.
Looks like a great setup and geometry.
It is made so that it will sideslip when overpowered.

Surface piercing V or half V planforms are well covered in one of the papers at Tspeers site IIRC. What you want is to have a taper so that you keep some aspect ratio when you lift off.

Now comes the ubiquous suggestions of how to complicate your simple machinery.
Board inclined like yours will lift out and it will lose span, increasing both its induced drag and that of the hulls taking over the sideforce.
If you had used a J-foil, with the top part rather horizontal-ish, and the bottom rather vertical-ish, you would quickly be lifted out by the horizontal part, but then not be lifted further. Your vertical foil area would be intact. The effective AR would be much like the same foil if straightened out, same idea as upturned wingtips on span-limited airplanes.

I think as long as we are discussing sharp edged foils ventilation is going to occur, and then it is probably just as well to have a stable ventilation. The monofoil foil had an L/D of 9 IIRC, and that was a quite low AR foil. Proa-ized version of that idea depicted below.

None of these sections that we are discussing are creating very much sideforce until you have leeway. Leeway in these hull shapes is not super. So the next step in complication is to pivot the j fore-aft, on a vertical axis. Of course with the way you navigate, it would have to combine with a kickup. Sketch shows AoA adjustment but not cant and rake (kickup) restraints. As usual, all sorts of geometric aspects are a bit off - just a quick illustration of the idea.

I am not quite sure how you reel your kite in, this is the reel I use: Kiteline on the big spool, driven by the fat continuous purple rope. The block near it is a one way block and then there is the cleat. The red bunghy keeps the continuous rope in tension. I find it OK to use and it was easy to make.

 

Also, I recall us discussing my project, and you asked me 'why don't you put the kite on the small hull?'

I had reasons for not doing that, not pertinent to your boat as far as I can see.
So now I ask the same to you, why don't you put the kite on the small hull?

I think it'll provide a better geometry all in all: The leeward hull is going to lift off at some point, and with your setup, you now have a reduced footprint fore-aft, increasing pitch up and stalling your board.


EDIT: Adding a configuration which I tested in a model behind a motor boat. It worked fine until the speed made the middle foil bend (or rather straighten out) and it released from the water. Fixed foils, on long ww hull. The model didn't actually have hulls but the lee 'substitute for a hull' were kept above the water by the tow line.

EDIT:

Then consider lowering the kite attachment points. 800mm seems awful high. As low as possible is a good estimate of the desired placement I think. The 2nd attachment illustrates how the lee hull can be hammocking between the hydro center of lift and the kite.