Understanding Wing Technology

Twist Control on Wing Mast

Hi All,
I've been studing the Stephen Clark Open Wing and Patient Lady control dwgs and it has finally dawned on me how it works. Isn't it great when the Penny Drops!! I note that Doug has been contemplating it as well. Stephens words are a bit complicated so I thought I'd simplify it here. There are slave deltas on the flaps and master deltas at the boom. Whatever the master does the slave does. Say there are 3 master deltas stacked up at the boom so we have 3 slaves and each slave is at the top of each flap. A slot is down the centre of each master delta and a pin connects the three master deltas. Depending on where the pin is in the slot alters the relationship of the deltas. This is called proportional control. So the geometry of the slot (width) and position of the pin along the slot dictates the difference (proportion) in angle of each delta. If the pin is pulled right back (slot is same width as pin) then all flaps are straight if the pin is slide to half slot then the difference between the top and bottom delta may be say 10degs and if the pin is at the bottom of the slot the differance maybe 20degs (widest slot place). I hope this helps. The yellow colour of the text makes it difficult. In this way the control lines are pulled taunt and the deltas do the work, plus all flaps are controlled by one sheet and it looks after itself when it tacks/gybes. Nice. Quite clever really I wonder who sorted this? David Hubbard? Cheers Peter S

 

Wing Sail Twist Control

Hi All,
I've modelled a single delta and will now move onto a 3 flap design. The top master delta is being moved and the bottom slave delta mimicks the master. The master is moved 25mm. The strings are 3mm dyneema. Notice for the deltas to be syncronous the strings have to cross, if they don't cross the slave moves in the opposite direction to the master. Cheers PeterS

 

Wing Sail Twist Control

Hi,
Heres the full 3 flap model. Its not quite the Patient Lady master delta arrangment. I want to make the system with one master delta and multiple slave deltas. This model shows that the concept can work. There is a slight nonlinearity due to low string angles. In the real wing this will be fixed using a quadrant connection vs a point connection. I'll call them strings because in the FEA system the elements used are called strings. The master "delta bank" at the boom is moved 90mm and the top flap moves 90mm. Each flap down moves proportionally. Onto a real one now. Regards Peter S

 

Peter, that is really neat! Thanks for posting it.

 

Wing Sail Twist Control

Hi All - The beginning of the real moth wing model. I made a Moth wing 14m2 and it has an elliptical profile with panel 1 hingeline 60% of the cord. Main features are 8m luff, panel 1 foot(chord) 1300mm long. I used a NACA0012 for the section. I made the skin 0.5mm mylar. I made the head rib and foot rib 10mm carbon for the present. I then rotated the head 20deg (fixed foot) and had a look. Oh yes... the skin is "moulded" mylar so its the correct shape. Obviously the skin wrinkled, so I looked at where the wrinkles were worse and put in a fibreglass batten. 20x5mm. Then rerun and had to put in another batten near the head. The leach also needed a batten and this is dia10mm fibreglass all the way. I think the mylar is too thick so I'll have a look for the correct film thickness used on these things. Learnings - its clear the panel width has to be quite small or the mylar wrinkles "along" the wing, so the next model will have lots of battens say 500mm spacing which then can be deleted from the model, which is easier then adding them to the model. Once Panel 1 is sorted then panel 2(flaps) can be added, then add correct control system. Currently I'm forcing the deflection using node control. By the way the wing needs 35kg (applied at leach on head rib) to twist it 20deg. This is the structural load not the aero load. The assumption that if the wing is stiff enough to support its own weight and deflect as required then it will support the aero loads. I'll apply aero loads once the wing is stable and twists nicely. Probably a week before I can get back to this. Ideally I want to twist the P1 and P2 and have no flaps and be free standing Hmmm.... But what I might do is look at Stephen Clarks "open wing" and model that...Cheers Peter

 

Hi Peter,

You are full trottle, very impressive. I didn't understood your questions about twist control, but I noticed you get it now.

Very quickly, if you want to have your own source, in order to have the good loads for your FEA software, here is what I can suggest.

Just consider the boat (ie:A-cat) sailing winward with just enought wind for full trapeze(max righting moment) and not yet overpowered. (around 9knts true wind I guess)

1-You can calculate the righting moment, (3400N/meter)
2-you can calculate the distance between centerboard center of effort and sail center of effort (5meters)
3-you have the boat speed (8 knts)
4-you can calculate the apparent wind with true wind at 45° with around 10° for AoA
5-You can calculate the "implied 3D-Lift Coef) of the sail.
6-You assume a smart crew who achieves a perfect elliptical lift distribution across the span
7-You can approximate the load distribution along the span.

I did it along time ago to have a clear picture of drag and other forces, and as fara s I remenber, I used a 22° angle between the longitunal axis of the boat and the sail chord at the sail center of effort.

From that I got around 670 N lateral force, around 170N driving force.
with implied 3D lift coef around 1,2.

Then if you assume the boat is at equilibrium at constant speed, the driving force equal the total drag.

Also with the implicit lift coef, you can get an idea of the induced drag, just using the formulas.

You need a few assumption for wing section drag & friction drag for the sail.

I used the DTMB serie 64 for the wave resistance and a basic coef for the hull friction drag, including centerboard and rudder total drag.

I was surprised to see that aero drag was around 50%+ and more than 60%of it was the induced drag.

Of course these figures are a bit fuzzy in my memory, and I used so much approximations that it could look a bit blasphematory for professional boat designers. But to process it on the paper step by steps prooved to be very helpfull for me. That is just my experience.

Cheers Mate,

As I noticed you plug fiberglass battens here or there, Please find below some specs for widely available carbon pultruded tubes sections.

Inner Diameter -Outer Diameter- I (mm)^4- I/V(mm)^3-Weight(g/Meter)
3 mm 1 mm 3.93 2.62 9.7
5 3.5 23.31 9.33 15.5
5.5 4 32.35 4.76 17.3
6 4 45.22 15.07 20.2
8 6 137.44 34.36 34
10 8 289.8 58 43.6
12 10 527 87.83 53.3
14 12 867.9 124 63
16 14 1331.25 166.40 72.71
18 16 1936 215.11 82.41
20 18 2701 270 92
21 19 3149 300 97
20 22 3645 331 102
24 20 8432 702 213
26 22 10932 841 232
30 26 17329 1155 270
34 30 25836 1519 310

 

"peaked up" head

===========
Peter, have you done any analysis at any point that suggests an advantage for a more or less rectangular planform with the back end of the square top peaked up-say around 45 degrees. Makes the leach longer than the luff. I read about a couple of French guys that had done a study showing that this was an advantage but I've misplaced the link.

 

Wing Sail Twist Control

Hi Doug - Thats an aero question and you'll need to get some words from an aero person. The sydney 18 footers and some sailboard rigs do this as well. I think the main reason is a compliant top for gust control. Plus there are drag reduction reasons, plus in some rules it may be free sail area. I usually am given shapes by the aero/hydro guys and asked to make them work structurally. Doug have your questions from the early thread been answered?

Hi Erwan - Thanks for the stuff on CF battens. I've been designing composite and various structures for 20 years so I'm familiar with the materials. In this case I think we need more compliant materials. CF is quite stiff and its going to be a mission to twist this thing comfortably. Glass is 1/3 the stiffness of CF so takes 1/3 the energy to do it.

The answer will lie in the following. The front panel (P1)will need to be segmented like the rear flaps if "rigid" panels are used. The flaps create slip planes in the structure. We can slip the front panel but aerodynamicaly this is poor. I think we may be able to slip it along the trailing edge and create the foil as an "open" section. I beams and channels have low torsional stiffness but good bending stiffness. This is sort of what we want. The second problem is that we want the zero twist plane (shear centre) to be about the spar axis. If we have elliptical profiles or tapered profiles in P1 the shear centre moves around as we go along the panel. (the shear centre is the point at which the section wants to have zero twist, in an isotropic closed section this is the geometric centroid) If we dont have a planar shear center then the action will be quite sticky as it twists around the spar. The bigger the panel aspect ratio (chord/thickness) the worse this gets, maybe a 3 panel design is better? We shall see. More modelling and thought balloons needed. Cheers Peter S

 

==============
I think so Peter- need to spend a little time making sure.

 

Wing Sail Twist Control

Hi All - Heres the current direction for Panel 1 twist. The Leading edge and trailing edge are flat sheets of fibreglass. CF would be fine too. They are currently 50x5mm rectangular. The ribs are supported where the spar would go through and the spar goes from bottom to top. The load at the "bearing" spot is very low due to this twist which is good. It only takes 1kgf to twist it at the top rib leach. The trailing edge remains quite straight so I can put the same arrangment onto it and try for a "flapless" Panel 2. The model twists 30deg which should be enough. Its the moth wing so is 8m long. The next step is to skin it and see how that goes. Then add Panel 2. Cheers Peter S

 

Thank you for all this, Peter-it is very helpful.

 

Wing Sail Twist Control

Weekend Roundup - Discovery... The wing structure for the twisting Moth Panel #1 as designed prior performs quite well by itself. But once a 0.5mm mylar cover is placed on it, it goes awol. The structure itself requires 2kgf at 800mm arm to twist it 26degs. Once the film is on it requires 80kg to twist it 26degs plus the film wrinkles quite easily at very low loads (<1kg) Pretensioning the mylar makes it worse, as the film wants to turn into a hyperbolic surface not the intended surface shape. (The Clystar film is best used at the flat rear part of the wing not across the front area) This requires some thought. What comes to mind is that we need a permanently stretchy film (latex, gold spandex?) vs a hard film so that we can stretch it more than is required for the aero loads. In this way it stays in tension when the panel warps. The skin "buckles" in a classic shear buckle (see AVI) which means the distance across the diagonals of the panels is getting shorter. Its about 3-4mm shorter under load than not. So all you twisted wing guys out there we need solutions for the film. Cheers Peter S

The AVI shows the first skin buckle, there are 100's after this. It occurs at a load of ~1kg.

 

Hi Peter,

Very impressive you are moving forward full steam, congratulations.

It is a bit difficult for mme to keep up wth you enthusiasm, be indulgent, I do my best.

Fiisrt of all, please find a link for Moth sail measurement and wing info:

http://wiki.mothosphere.com/index.php?title=Wings#Measurement_of_Area

As I am not sure to understand in depth your FEA simulations, ,I would like to sum up the main advise of the most experienced wing builder ; Mr Steve Clark.

He mentioned that the point is to have a LIGHT and RELIABLE wing with good command-lines. Wing section is a second order parameter.

To be light, I guess you should better consider in your FEA model, a composite skin only for the first 25% or 30% of the wing section.

You can consider also a conservative assumption for the spar position on the wing section between 7.5% and 10%.

And in this case ( 25% Leading edge / spar @ 10%) your sheer centroÏd is not so far from the spar axis ??????

and may be it is not necessary to slice the leading edge anymore ?

Also, I guess that the AC45 is not a good benchmark for small boats, they have to package their boat in containers. For A-cat and Moth, I will favor a flap split that mimics the front element split, usually 2 panels (mast foot- hound)+(hound-mast tip) .

Now coming back to the first element, considering the appropriate technology of the second part of the wing section (between 25%chord and the structural trailing edge), as long as your wing section shape has a minimum convexity, the heat shrinked film like Clysar on feather light ribs is probably the lightest solution.

After that, you can add or not, on the first element an hinged flap, or use Mylar for a morphing trailing edge if you can control it. In this case, the structural trailing edge is not at the same place.

Now with regards to the wrinkles, I guess that the Clysar is double face-adhesive tapped on the structure before heat-shrinked, and I guess it does not need to be double-face adhesive on the ribs, only along spars vertical axis.

If it is the case, your diagonal will be different.It will be a straight line from the end of the leading edge nose at the bottom to the tip corner of the trailing edge at the top.

To address porosity on the compoosite leading edge, it could be clever to have the Clysar wrapped around it, as a result you will save 2 lenghts of double-face tap adhesive which are heavy, and some epoxy filler as well.

Hope iit can help to modelize light wings

Cheers Mate

Erwan

 

Peter,

Just a short message to provide you with a rough idea of the aerodynamic loads on a A-Cat sail surface
Please find below a proxy of aero loads meter by meter along the span, more or less mimicking an elliptical lift distribution.
From the top to the bottom

1st meter 25 N
2nd 52 N
3rd 73 N
4th 86 N
5th 94 N
6th 103 N
7th 103 N
8th 90 N
9th 70 N

For instance, the corresponding aera for the 6th & 7th panels can be assumed @ 2 sq meters

And for the Top panel between 0.75 and 1 sq meter.

Cheers Mate

Erwan

 

Wing Sail Twist Control

Hi Erwan - Thanks for the loads, but at this level of design really need a proper pressure contour map of the wing surface. The loads you have here are good for designing the spar but to try and use this to design the overall structure we need the 3D distribution. 0.5mm Clysar has an areal wt of 70g/m2 so is really light. Its difficult at present to make a lighter wing using anything else. Simulations - the models simulate the actual structure. They use plate and beam elements that represent the real thing. My models so far have been to investigate twisting. My aim is to have a free standing wing, with a front panel twisting and a rear panel twisting, no flaps. My assumption is that if the structure can twist and do the various mechanical things it has to do then it will support the local aero loads. The spar will support the global aero loads. I have designed and built a 22m free standing mast for a 60ft catamaran and also designed a 45m free standing mast that was not built so the spar is the 'easy bit".

There are two paths to tread - 1) stressed skin and 2)space frame with skin. Currently most builders are going the space frame approach. The D front section could be considered a stressed skin and the rear bit a spaceframe so hybrids are out there. As my panel needs to twist there is a conflict between the required aero profile and the structural shear centre. If we use an elliptical profile it has to be tweeked so that an axis of torsional zero(or low) stiffness exists. This is where the spar needs to be placed. We want the wing to twist about an axis so the "bearings" are lightly loaded and the twist is even. The D front section is a bit stiff and heavy for my twisting wing at present but looks like I'll have to visit it as the film doesn't form to the required shape at the leading edge. I can model the shrinking process (membrane preload) if needed. But from a structural perspective a shrunk film or a non shrunk film performs the same. Its the shape that matters so if the model is built in the "flying shape or tight shape" thats all you need to do for our FEA purposes. By the way the composite leading edge should not be porous. Moth Wings - may not need to be twisted as they are short and the gain may not be worth the effort, but the aero guys will have to figure this out. Its clear in the C=Class that its effective as twistable wings win. My current investigations have lead to the following * I need to look at the front D section for suitable materials, my current thought is moulded resinated peel ply. Peel ply is woven polyester or Mylar or I could vacuum form film * the twisting is possible * slips may need to be incorporated into the structure to keep the skins tight or a dynamicaly loaded skin may need to be looked at like hang glider techology. Similiar to soft sails, tensioned battens and leach lines with cuban fibre skins* there is lots of scope yet to build better wings! Cheers Peter S