BMW Oracle Racing's revolutionary solid wing sail
Gilles Martin-Raget BMW Oracle Racing
BMW Oracle Racing's revolutionary solid wing sail

That wing

We look at the nuts and bolts of the BMW Oracle Racing tri and its amazing solid wing sail rig

Wednesday February 3rd 2010, Author: James Boyd, Location: Spain

The 33rd America’s Cup may be a disaster from a legal standpoint, but it has created two of the most technologically advanced boats ever built and taken racing multihull design to a whole new inter-galactic level.

In the multihull world there has been a long standing debate over whether trimarans or catamarans are superior on the race course. Looking at the different circuits around the world over the last 30 years or so and there is conflicting data about this. In France, both the Formula 40 and ORMA 60 saw trimarans very much dominate. Round the world in the Jules Verne Trophy catamarans have proved superior - something that could change in the next couple of months. Catamarans also have the upper hand in the less well reported, but highly active, multihull scene on Lake Geneva, where they have dominated the Bol d’Or for more than 20 years.

And so with the two new America’s Cup multihulls we have two quasi-cat/tris. Instead of a centre hull, Alinghi 5 has her fore and aft spine system to provide fore and aft rig tension, similar to Ernesto Bertarelli’s Sebastien Schmidt/Jo Richards-designed triple Bol d’Or winning 41ft cat and the D35 one designs on Lake Geneva - a system first seen on Nick Keig’s Derek Kelsall-designed VSD offshore cat back in the early 1980s.

Meanwhile BMW Oracle Racing’s tri began life as a 90ft inshore version of an ORMA 60, but over her lengthy evolution her floats were replaced with ones a third longer – substantially longer than her centre hull – and the conventional ORMA 60-style rudder and daggerboard on her centre hull were removed. So in terms of foils she is now just like a catamaran – with the boards in her floats preventing leeway as well as providing vertical lift (something they achieve – unlike Alinghi 5 - by having a hydraulic canting mechanism attached to them – eat your heart out L’Hydroptere...)

When sailing, assuming helmsman James Spithill remains on the ball, her centre hull will be pretty much always clear of the water. So spine system or centre hull - they may be quite different in terms of hull shapes and structure, but the general configuration of the Alinghi/BMW Oracle platforms is now very similar.

It’s an academic point now, but one thing we hadn’t fully understood when there was the Supreme Court debate over BMW trimaran’s rudders forming part of her LOA, was not over the original transom-hung rudder on her centre hull (of which there are many examples of boats where the rudders are not included in the length measure) but to the rudders in her floats which are conventional spades, and located well aft of the centre hull transom and outside of the maximum 90x90ft load length/waterline permissible for their boat (for which there are no other examples...not that it would have made any difference to the court ruling as the rudders don’t exactly contribute to the load carrying of the boat.) In terms of measurement we understand that load waterline length is taken from her centre hull while beam load waterline is measured from the middle of her floats, both of which are in the water at standstill (this is best seen in our video - below). So how she floats at standstill is absolutely critical to her design if she is to measure.

Back to the cat-tri debate and BMW Oracle Racing’s chief design guru Mike Drummond ponders our question about if he had to start from scratch today - would he still have come up with a tri? “There are so many contradictory effects: The cat flies a hull at a very low heel angle. The trimaran has a better volume distribution in its floats. The trimaran tacks a bit better and is probably a bit better in the light - or they used to be - but both teams have gone for rigs of extreme height, so that has nullified a lot of the traditional advantage of a trimaran. So it is not clear cut. A catamaran has a bigger gap between its sails and the water. On a trimaran the hull tends to end plate the rig quite nicely. Then there is the whole structural equation – I certainly think a catamaran has attractions from simplicity, but the deeper you go into it, the more complex you realise these simple boats are. So I don’t know.”

Structurally the two boats are very different and it has been observed that the Alinghi cat is very stiff, while the BMW Oracle tri is less so. Drummond concurs with this: “From what I have seen it [Alinghi 5] looks like it has less deflection, but it is also narrower and lighter, so its loads are less. Imagine the two hulls are twisted, well we are wider so for the same rate of twist we have more. So we are not comparing apples with apples. We could have made a stiffer platform but we would have added weight.”

Drummond also reckons that Alinghi designed their boat for lighter conditions and this allowed them to trim safety factors when it came to the engineering of Alinghi 5.

So if the platforms of the two boats have more similarities than it might seem, the fundamental difference between the two is that the BMW Oracle Racing tri features a solid wing sail, while Alinghi 5 has the conventional racing multihull set-up of a towering rotating wingmast and softsails. On both rigs, rake can be altered and the mast can also be canted – this is nothing especially new as the ORMA 60s used to do this, only the degree is different. On the tri for example the rig can be canted by +/- 15 degrees whereas on the ORMA 60s it was typically +/-10 degrees.

According to Joseph Ozonne, who worked on the American team’s solid wing, they started design work on it in December 2008, but its build was only given the green light in March 2009 with a view to starting it in early May. So the design time was tight and the build-period even more so given that the mast was unveiled in San Diego in early November.

"We knew pretty well what we wanted, so it was pretty straightforward," says Ozonne of the design process. “You look at the span and the area, the configuration you want - the number of slots, sections, what sailing points you want to be [good] at and especially how you integrate the wing on the boat and for which conditions you want to be good in.”

A small group worked independently from the rest of the BMW Oracle Racing design team on the project led by Ozonne and CFD expert and hydro/aero dynamist Mario Caponnetto and his colleague Francis Hueber, with ex-Luna Rossa rig designer Scott Ferguson, Herve Devaux and Steven Robert on the structural and composite side, with Dimitri Despierres on the mechnical design and acting as the liaison between designers and builders while British ex-Formula 1 mechanical engineer Thiha Win working on the complex control mechanism for the wing. They had quite a task on, as Mike Drummond puts it: “We didn’t know exactly what the performance would be at the start, we didn’t know how we would handle it. We didn’t even know how to build it at the start.”

The result they came up with is a solid wing sail similar to those seen on recent C-Class catamarans. Think aeroplane wing pointing up rather than sideways and being ‘tackable’ – plane wings are the equivalent of a solid wing sail, but only working on one tack... Compared to the C-Class rigs, the one on the American tri is slightly dumbed down as the design remit between the two types of boat differs: With this America’s Cup anything goes, while in C-Class sail area is restricted. Thus in the C-Class the game is simply to design and build the most efficient sail plan possible. On the Cup multihulls the sail plan can be any size and so the BMW Oracle tri can fly gennikers downwind or in light conditions upwind. The USA wing is 68m tall, its chord (lateral thickness) ranges from 0.5 to 2m at its base while fore and aft it is 14m along its foot to 3m at the head. The published weight of the rig is 3,500kg which BMW Oracle maintain is lighter than their previous wingmast/softsail set-up.

Given the lack of any constraint the design team reckoned that there wasn’t the need to go for maximum efficiency (and thus added weight and complexity) so their wing has a single slot, while the wings on C-Class cats like Steve Clark’s Cogito, have two. (A physics recap: The slot, much like that between the main and a genoa, aims to increase lift of the sail plan by allowing wind to pass from the windward side to the leeward side of the mainsail/wing. This maximises the amount of flow that remains attached to the leeward side of the mainsail/wing – two slots typically extends the flow even further aft).

So the BMW Oracle Race wing comprises firstly a single piece forward (the front/primary/main ‘element’) extending up the full height of the rig. Like all wingmasts, this sits on a ball allowing the whole rig to rotate, cant, rake, etc. This main element is made up of a carbon fibre structural member in the middle with a Kevlar ‘nosecone’, while the trailing edge is built using a number of carbon fibre frames, covered in a thin, but robust, aircraft-grade plastic that is heatshrunk on to the frames.

Down the tip of the main element’s trailing edge are seven hinge points. These connect the trailing edge of the main element to the second/aft element. This second element is made up of another full height structural member, although much smaller than the main element’s, and hung off the back of this are the flaps extending vertically up the length of the rig. When this rig was originally stepped there were eight flaps but when the US team arrived in Valencia they increased the height of the main element to approaching 70m and added a ninth flap at the top - impressively both can be removed for racing in stronger conditions, although this is unlikely to happen given that racing is probably only going to take place in lower wind speeds in Valencia.

According to Mike Drummond the ninth flap was added when there was the possibility that they might have to sail in very light winds in RAK.

Impressively, camber (ie the athwartships angle relative to the rest of the wing) and twist can be induced in each of the nine flaps, via an intricate control mechanism inside the wing (the camber control rods run across what look like forward-pointing spreaders that are mounted at each of the hinge points. When there is zero camber up the rig, the slot is closed but increases when more camber is put into the flaps. (For a visual demonstration of this, see Steve Clark demonstrating the different rig functions on his C-Class cat, Cogito here)

Positioning of the wing relative to the wind is carried out by adjusting a combination of mast rotation and traveller (there is no main sheet). The shape vertically up the rig is altered via the camber and twist induced in each of the nine flaps. The clever part is that while each flap has individual twist and camber controls – with potentially more than 20 lines to operate this, twice that if its going to work on both tacks - in practise twist and camber across all the flaps are operated by just two lines thanks to a complex cascade inside the mast that allows a progressively changing profile to be maintained up the height of the rig.

Joseph Ozonne explains: “You have the camber control – the global camber. When you play with that you give the same amount of camber to all the flaps. If you change it you can have a deeper or flatter wing. Then there is the rotation – like the traveller. You rotate the wing. Then you have the twist – each individual flap can have a given angle so you open the top or close the top. That is controlled in a cascade. You have all lines coming from the top to this system [what we term the ‘gearbox’] and with this it is a simplification of the controls where you can play with all the flaps together.”

And in addition to this, as mentioned earlier, the whole deal must be tackable and it seems that to achieve this the wing has similar set-up to Cogito's rig with a kind of gear box located at the bottom hinge point. Hydraulics are used to help in the operation of the twist and camber however the traveller is on a 2:1 back to a winch. As Russell Coutts explains, the traveller loads generated by the solid wing rig are only around 3.5 tonnes, compared to more than 20 with a softsail set-up.

If this all seems impossibly complex then all it is trying to do is effectively mimic what your mainsail does, only because the wing has some chord to it, it is very much better at generating lift, while the flaps are like having battens that can be individually controlled. Most impressive is that this enormous rig is operated by just three main control lines – traveller, twist and camber – and on board is trimmed by just two crew – Dirk de Ridder and Joey Newton.

Compare this with the amount of cordage coming back to the cockpit from a conventional rig. As Mike Drummond emphasises: “It has remarkably few controls. Really the sophisticated thread layouts and batten stiffnesses and luff curves and mast bend and all those normal controls - they are all encapsulated in the design of it.”

Within the rig there are pressure sensors (behind the black circles) to monitor wind strength and according to Ozonne the trimmers have software that tells them the optimum settings for the rig, along with all the conventional target speeds, etc. “In the end it is still a question of feeling the boat – boat speed, and heel angle and all those kind of things. It is not a computer that drives the boat.”

As to the on the water experience with the wing compared to a soft sail/wingmast, Mike Drummond, who regularly sails on board says: “Everything is a lot simpler. It doesn’t flap when you go head to wind. When you are sailing inside a headsail it starts losing its drive, but when you sail inside the wing it is still producing lift, so you get more thrust for longer into the tack, you get less drag in the middle of the tack and it starts working for you immediately on the new tack, so there are some small advantages there. Also it is very repeatable. If you set it at 20 degrees today and you want to do that tomorrow, it is easy, it is not reliant on certain runner tension or the sail stretched or gotten old or strunk in the sun. So we should be able to sail the boat much more accurately.”

Drummond adds that the wing could allow them to point higher upwind, however with these extreme apparent wind generators the aim of the game is always VMG. “We can point higher with it, but the best VMG angle doesn’t significantly change [with the wing compared to softsails]. We hope we have saved a bit of drag on the rig and that has the same effect on the pointing ability as if we’d saved some drag in the rudders or the hull or the windage. If you can magically strip some drag off the yacht then there is a new best speed and a new best angle that it sails at. So even if we can sail higher it might not be the right mode to.”

On C-Class cats twist can also be induced up the height of the forward element too, but with the BMW rig the twist in this is only from what the structure allows. “We don’t play with that , for different reasons,” says Ozonne. “We are much faster because our apparent wind range is much smaller so we don’t need to open it and downwind in very light air we have the genniker, so we don’t need this extra feature.”

The result is phenomenal. The surface area of the wing is roughly 600sqm, while typically the upwind configuration with the team’s previous softsail set-up was 1200sqm and they achieve similar straight line speeds. So the efficiency of the solid wing over a softsail equivalent is roughly twice. Russell Coutts, an engineer as well as a sailor, makes the interesting point that one of the advantages with the wing’s structure is that it is supported in a number of places, whereas a conventional softsail you can only support it between the three corners.

As to the relative performance of Alinghi 5 and the USA tri, the feeling remains that the former could be the stronger in light winds and on the reaching legs on the second race, while the tri will be a weapon at the middle to upper end of the wind limits and upwind. “There is a certain logic to that and it is difficult to beat the laws of physics,” agrees Mike Drummond. “If we are wider and heavier, we just shouldn’t be as good in light airs and it is a question of whether light airs means, 5, 6 or 7 knots. I’m pretty sure it doesn’t mean 10.”

Both teams have been running simulations of their boat against the others, however Drummond states that the major limitation in this is knowing the weight of your opponents boat. In a previous article Franck Cammas (who was involved with the BMW Oracle Racing campaign) reckoned Alinghi 5 was around 13 tonnes, the BMW tri around 15. But this may no longer be the case.

Drummond says that the sailing team is now very confident is sailing their monster tri. This, he adds, is less from having had a year more than the Swiss to train but from respecting the loads the boat is designed to withstand and the boat being very well designed from the outset.

So can they throw it around? “Yes we can, within your limits. There are some manoeuvres that you just plain can’t do as quickly because of the 90 x 90ft platform and small rudders. The sailing crew – it is hard to keep up with the speed sometimes.”

Anyway, we shall know more when the boats, in theory at least, line up for the first time on Monday. The forecast for a 10am start is for 10-15 knots from the southwest, building through the afternoon (so it may be canned). If they do sail in those conditions, we know where we’ll be putting our money...

Latest Comments

Add a comment - Members log in

Latest news!

Back to top
    Back to top