BMW Oracle Wing

The issue is the wind. It's easy for the wing to get out of control when raising or lowering it. It can't weathervane, then, and it's not well restrained by the rigging.

There's a maximum wind strength for doing the maneuver safely. So you have to anticipate the weather and if high winds are coming, get the wing down before it's too late. Otherwise, you have no choice but to ride it out.

And once the wing is down, what do you do with it? It took the coordinated efforts of two cranes to lift the wing off the boat and set it on dollies to be rolled inside a tent. That's a lot of waterfront real estate compared to the footprint of a sail bag.

Paul Bieker, a Seattle naval architect was the guy that engineered the logistics and rigging to raise and lower the wing, and to move it from shore to boat and back. Herve Devaux was also involved with the logistics design.

Because the wind is not constant and fluctuates in direction. Google Katzmayr effect.

A wing has to be "flown" 100% of the time. When USA 17 was moored, a crew was aboard her 24/7, ready to sail if she should break free of her mooring. Magnus Clark, experienced wing sailor from Toronto and holder of the C-class Championship Trophy, was the night skipper. At least two people were on the boat at all times, and slept on the nets. More crew were sleeping close by on shore, on tap for immediate assistance if needed.

USA 17 did not hang back from her mooring, even when the wing was free to rotate. She was typically surging forward, like a racehorse wanting to be let free to run. It was not unusual for her to be surging back and forth 10 m. In high winds (and it experienced over 30 kt), they had RIB with a towline to her stern, pulling back on the boat all night long. They couldn't anchor from the stern, because it had to be allowed to swing as the wind changed direction. There were some pretty hairy nights out on that mooring.

 

I was out of line in putting Doug down - and apologise - i just dont think that this technology is appropriate to this boat. Again i am sorry and should have put it differently/better!!!!!

 

Alex,

As for the portion of your remark toward me... I thank you for your concern. By people who know me, I am not known for doing rash things... many of my friends would say I am too conservative. However, I am known for asking questions all over creation, not for direct applicability to a project, but just to fill my natural curiosity on the order of things. I have asked the question of my post #14 many times to people I respect. So far, all have replied with "you can't"... when pressed with why... "they can't". I am that precocious two year old that keeps saying WHY? Tom, apparently, is the first that is able and willing to take the time. And Doug has kindly not patronized me by saying, "you can't".

I remain aptly named,
Inquisitor

 

Tom,

Thank you for your posts... I have gone through your posts several times and will many more times. I find them mind boggling fascinating. Sorry... they open up more questions than they close for me. I respect your time and I'll try to digest more before I start blurting out "WHY".

First thing - I see that I did miss-phrase my question. "Just what is the issue of taking a wing down?" I more accurately meant, "Just what is the issue of needing to take a wing down?" I have Googled the Katzmayr effect and I will dig deeper, but the first MIT paper opens up with stating this effect lowers the drag of the wing. This would be encouraging my premise. However...

I think I see in your text and between the lines that my premise is only valid if I assume the foil weather vanes. And that, you're saying in the real world case the sudden changes in wind direction can not overcome the inertia of the wing quickly enough. That, for a moment, there is a significant AOA creating a major lifting force. That this transient force will be far more than the bare pole case. (in this example... 1000 lbs). Am I understanding correctly?

 

Yes. I can't speak to 1000 lbs, but the dynamic loads can be different from what you'd expect from static conditions.

The Katzmayr effect can produce forward thrust from a stationary wing. When the wind is coming from one side, the lift vector is tilted forward - this is how a boat sails, after all. Then when the wind shifts to the other side, the lift switches direction but is still tilted forward. So the average longitudinal force in an oscillating wind can be forward not aft - resulting in apparently negative drag!

This is different from oscillating the wing through the same angles of attack with a steady wind. Since lift is defined to be perpendicular to the wind, when oscillating the wing the lift always points to the side and the total force points aft, so the oscillating wing produces more drag than the stationary wing.

There's inertia to the wing, so it doesn't respond instantly to a change in wind direction. There will be a phase difference between the wing and an oscillating wind. Whether there is increased drag or even forward drive depends on that phase difference. The wing and the wind can get 180 degrees out of phase, so the wing is doing exactly the wrong thing! When the wind goes to one side, the trailing edge may already be going in that direction, with the wing stopping and swinging the other way just as the wind reverses direction, too.

The turbulence in the wind has a lot of different physical scales and varies over a wide range of frequencies. The wing can respond to gust frequencies that are lower than the wing's natural frequency, but to turbulence that is substantially higher than the wing's natural frequency the wing is effectively stationary.

And gusts can cause a significant heeling moment, too. It's not just the drag that is of concern.

The wing can't react to turbulence that has a scale smaller than the wing, so as you make the wing bigger, there is more turbulence that it can't feather out. Even if these scales don't produce a net force, they can produce moments. Think of a horizontal vortex 150 ft in diameter hitting the wing on USA 17. There may be counteracting forces at the foot and head, but there will still be significant heeling moment.

Plus, the mean wind isn't uniform all along the span, either. When there is shear (which is pretty much all the time), the wing may be feathered at one height but lifting at another. That's why Harborwing have split their wing into upper and lower halves that independently feather to the wind. The variation over each panel is a lot less, allowing the panel to do a better job of feathering itself.

With the wind doing one thing, the wing doing another, and the boat doing something else, it can get pretty wild on a gusty night!

 

Tom, I've no experience with full wings but know about fairly large sized wing masts in high winds. To reduce the driving moments of shifting and shearing gusts on the wing mast -and here I learned the hard way because earlier I had underhung rudders - so that in oscillating, or just continued strong wind from a generally fixed direction, the boat would sail like a lunatic while at anchor, in fact once sailed itself over while on a heavy mooring off Herne Bay - but what reduces this problem hugely is to have lift out rudders .... and then the platform just skids about without racing to the end of its tether and cartwheeling. Of course the daggerboards are up too. If they are contemplating wing rigs on a monohull I smell disaster because the keel plus rudder will send the boat into frenetic sailing while moored too. Maybe they could cant keel to the surface ... oops sudden wind change, holy **** our latest AC boat has sunk itself.
They couldn't lift BMW-O's rudders, could they? What a bloody nightmare.

 

Well, if you were actually interested in designing/building a wing sail, the best course of action would be to use a real CFD program and not mitchlet.

 

No, that's why they had a crew on board 24/7. Magnus Clark definitely had a huge responsibility as the night skipper. And Rossi, his companion from Toronto, that was on the night crew with him.

Guys like Mark "Tugboat" Turner and Tim Smythe justifiably get a lot of publicity for heading up the boat building team, but guys like Wolfgang Chamberlin and Scotty that had lead roles on the shore team also deserve a lot of credit. I've heard it took 42 people to get the boat sailing, and it's hard to appreciate the competence and professionalism of the whole BOR logistical effort until you've seen them in action.

 

Michlet doesn't have anything to do with wing sails, per se. It calculates wave drag of slender hulls.

Actually you'd be surprised at how much simple tools were used for design work on this project. The wing was analyzed using the highly sophisticated Navier Stokes codes of Applied Fluid Technologies. However, the actual shaping of the wing was done with spreadsheets and XFOIL. Mario Caponnetto had a lifting line spreadsheet that included an optimizer so he could compute the optimum planform shape. The slotted wing sections were analyzed using MSES, but the individual element sections themselves were shaped using XFOIL because it was so much easier to make specific changes using XFOIL's inverse method. So the actual shape of the wing was determined using the simplest of fluid dynamic tools.

The sophisticated CFD was primarily used to determine how to operate the wing. The predictions from the simple 2D methods were confirmed by the 3D Navier Stokes calculations, but there wasn't time to do a lot of design iteration based on the NS codes. The tables that were loaded into the simulation to predict performance, and used to determine target flap and wing trim angles onboard, were computed using CFD. But even then, the 2D results gave a lot of insight as to why the wing ought to be operated the way it was.

Navier Stokes codes are just too expensive to be used for preliminary design, and they only tell you what a given shape will do. For design, you really want codes that can tell you what the shape should be. That's why I really like inverse methods. They come up with shapes that I'd never be able to create using intuition and trial-and-error optimization.

No matter how sophisticated your analysis, once you've built it, it is what it is. The boat doesn't care whether you used CFD to analyze it or not. It will perform the same, regardless. What the CFD will tell you is what performance to expect. And why. On a big program with lots at risk, that's very important to know before committing to laying up carbon. On a small program, it's cheaper just to build the thing and go sailing.

If I were doing a C-class cat design, I'd definitely use Michlet as my workhorse for designing the hulls. XFOIL, Javafoil, and a lifting line spreadsheet would do most of the work in wing design. Spreadsheets would be used to do weight and balance accounting. Affordable CAD programs would provide hydrostatics, and the VPP could be done in a spreadsheet. The big uncertainty in the wing design would be maximum lift, so if I had some budget for CFD, that's where I'd spend it.

 

Thank you Tom for taking the time (lots of it) to explain the BMO wing

My knowledge of wings and rigs is obviously much more limited than yours. Most of the theory I gleaned from chats with Tony Marchaj at his house over 30 years ago

One thing he told me was that although a wing sail was the most efficient a small high aspect jib and big full battened mainsail was nearly as good - and a lot more practical. But only in stronger winds. In lighter winds then the jib should be replaced with a big masthead drifter.

Tom is right that rigid foils left up and tethered can fall down. I used to live opposite the Walker Wingsails yard (in fact you could see my house in their adverts). One night when a boat was ashore the rig was blown over. (They kept quiet about it)

Richard Woods of Woods Designs

www.sailingcatamarans.com

 

True. The Mast3 rig on USA 17 would have been competitive, too. The soft sail rig can be faster in higher winds because it can be reefed.

For a given hull, the optimum rig tends to have a similar aspect ratio as it is sized for different wind speeds. It needs to have enough area to power up, and the taller the rig the less drag it has. But it also comes up against heeling moment constraints from the hull. For the same heeling moment, the rig can be made taller for less drag, but the area has to be reduced to stay within the same heeling moment, which reduces the power. So there's an optimum height and area combination for that wind speed - an optimum aspect ratio.

In higher winds, the induced drag for a given lift (set by the heeling moment) goes down. This means the optimum rig can be shorter, because it's not as important to go tall for efficiency. With a shorter rig, more force can be generated for the same heeling moment. The optimum area is also less because more force is produced per unit area in high winds. These two trends tend to shrink the rig in proportion to each other, maintaining a similar optimum aspect ratio.

This is nicely achieved in practice by reefing, which the rigid wing cannot do. The rigid wing must operate at low lift coeffcients in high winds, where it has excessive parasite drag. It depends on twist to lower the center of effort, which makes the upper portions operate at even lower lift coefficients.

The multihull concept for the 34th AC says wings are allowed. It doesn't say wings are required. Given the large wind range for racing, and the high probability of being toward the upper end of the range if the racing is held in San Francisco, I think you may see a lot of development of rotating wingmast rigs if the multihull concept is selected.

A twisting wingmast would be an interesting approach - like Krazy Koyote. It is possible to design a wingmast that has attached flow on both sides - all the way back to the sail track - over a useful range of angles of attack. But, especially with the bending of the flow from a jib, it's hard to stay within that range over the whole span. So a wingmast that can be twisted would allow attached flow over the entire mast. That could make a soft sail rig pretty darn competitive with a rigid wing.

 

Wing vs Sail forces efficiency

I've been working on a spread sheet that predicts heel and drive forces based on the simple wing sections using Cl and Cd from a simple NACA 0012 foil http://www.cyberiad.net/foildata.htm. All it does is a trade study on the relative angle between the boat and apparent wind and the AOA of the foil relative to the apparent wind. It calculates the heeling and drive forces in the boat's coordinate system.

I haven't found much public information on the Heel to Drive force for soft sails. However, I did find this on John Shuttleworth's site. http://www.john-shuttleworth.com/Images/Fig10-D50-upwind-loads.jpg I'm assuming its a relatively high performance sail (probably laminate) as its for an all out racer where money seems to be of secondary concern. In it, it shows at a 24 degree apparent wind angle the Heel to Drive force is 5.9. With the spread sheet, I'm finding at a 24 degree apparent wind angle, AOA = 12 degrees (using the NACA 0012) the Heel to Drive ratio is 2.3.

My question: Does it seem reasonable that a simple airfoil (one without an advanced flap such as the BMO wing) would still be 2.5 times better than the best laminate soft sail?

 

It could very well be reasonable.

What is the ratio between the driving forces for sails vs NACA0012?

Usually it seems that a simple aerofoil will not develop enough lift at a reasonable area. So you end up needing a multi section aerofoil like the BMW one or a mainsail and a jib. Or so it seems.

Also the cl and cd for the aerofoil is probably for a 2D section whereas the numbers from the Shuttleworth site is probably 3D and including drag from the rig.

 

Ahhhh! You are correct!

I don't know if the forces on Shuttleworth's image are measured, real world or theoretical from some sail prediction program. My NACA0012 are 2D. I understand that aspect ratio and tip vorticies reduce the affective wing lift to drag ratio. However, I was under the impression that if I keep the plan view of the wing similar to an elipse (WW2 Spitfire wing) that it approaches the 2D case.

To answer your question... I was strictly dealing with ratios, so I hadn't sized the wing yet.

Thanks, I'll have to read up some more on the 3D case!

 

Are you considering the induced drag? You'll need to take into consideration the planform shape, camber and twist.

And what is your definition of "better"? Maximum lift? Maximum L/D?