Windy city

Thedailysail visits BMW Oracle Racing's wind tunnel and talks to its proprietor Dave le Pelley

We may not be the brightest star in the firmament, but one aspect of the America's Cup that has recently been making an impression on the grey matter is just at how much higher a technological plane teams within this sphere operate compared to any other part of our sport. Quite simply within Cup teams there is more money and more resource to make sense and quantify the black art that is every aspect of sail boat design. From time to time results from this even leak out, filtering through to the less exclusive areas of our sport.

One man intensely involved in the science into what makes boats tick is Englishman David Le Pelley, who now considers Auckland home but spends a considerable proportion of his time in Valencia when he runs the wind tunnel for BMW Oracle Racing. This facility is based in an anonymous looking unit a 15 minute drive away from the glamorous Cup bases around Valencia's Darsena marina to which we are whisked in some £60,000 worth of brand new 5 Series.

So what is the point of a wind tunnel? When it comes to R&D, be it into sails or aspects of hull shape or deck gear, etc etc these items can obviously be tested full size but this is both costly, takes time and is generally inefficient. Thus much testing takes place in 'model form' - generally within the virtual world in instances where conditions can be simulated accurately in a computer and in instances where they can't physical devices such as wind tunnels and towing tanks get brought into play. A good example of this is sails: Upwind sails, where the flow is attached and things behave according to fairly straightforward (and programable) rules of physics, is carried out numerically in the computer whereas R&D into downwind sails, where the flow characteristics are a whole lot more complicated, are carried out in the real world, built to scale and then put through their paces in the wind tunnel. Thus minute changes can be made with relatively ease and minimal cost to a model sail compared to its full scale equivalent, this is then tested in le Pelley's tunnel and if it performs well then the odds are very good that it will work full scale.

A feature of the BMW Oracle Racing wind tunnel is that it can create a 'twisted flow', mimicing the often dramatic vertical wind sheer profiles found in different venues. It is believed that one aspect of Team New Zealand's technological success in the Cup during the 1990s came through their sail development work in the University of Auckland's wind tunnel where they were able to simulate the wind sheer profiles for the winds off San Diego, and subsequently Auckland through this process.

Le Pelley gives some insight into wind shear: "Wind has a velocity profile, so the wind at the surface of the water is going at 0 knots and as you go off the surface of the water the wind increases in a predictable pattern depending on a lot of different features. And since the boat is going at a constant speed through this vertical velocity profile, theoretically the apparent wind angle that the boat sails on the water is dead ahead – because there is no wind there. And as the wind velocity increases and the boat is moving at a constant speed, the angle of the wind varies." As the picture at the top of this article illustrates the wind at hull level is considerably more headed than it is further aloft.

The BMW Oracle Racing wind tunnel is a development of the University of Auckland, unsurprisingly as David le Pelley is also still responsible for running that facility, where he says these days due to his contract with Larry Ellison's team he gets chucked out of his own facility in Auckland whenever Emirates Team New Zealand are in there using it. "It is based on the same original design," he says of his tunnel in Spain. "This is slightly smaller. The Auckland one has progressed and it has got larger, we have done other things with this one and it has diversed away."



Working from the back forward, the BMW Oracle Racing wind tunnel is driven by two 45kW fans, each driving a giant 3m diameter blade. Le Pelley describes his tunnel as 'slow speed' with most testing carried out at around 5 knots. His tunnel in Valencia can run at up to 12 knots and he says it could blow harder except that they have only a short space to play with and the priority was to create a very high quality flow.

Thus the first step is to remove the turbulence and vorticity from the fans by forcing the wind through a honeycomb and then a fine steel mesh, a type of wind sieve. The subsequent clean air flow is then ready to be directed through a series of what are effectively vertical Venetian blinds. Horizontal steel rods connect these blinds at intervals up and down their height and are used to adjust the angle of the blinds over the course of their length, thus creating the 'twist' in the wind flow to suit a particular venue. Typically there is considerable twist down low. "For downwind sailing in a decent breeze it goes straight through at a full scale height of 10m and it is about +/- 10 degrees top to bottom - so 20 degrees of twist between the top and the bottom of the mast," says Le Pelley. "And that is quite significant. Especially in Valencia we are finding there are some quite strange twist and velocity profiles because when we get the big delay we are finding we are starting in a sea breeze, it does quite strange things to the wind in how it sets up, so we are looking hard at that."

Obviously research has been carried out into the complex wind shear behaviour that occurs off Valencia as being able to predict this is fundamental to Le Pelley's work.



The front end of the wind tunnel is the stage, the vertical Venetian blinds forming a floor to ceiling backdrop. Centre stage is a turntable on which a 2m tall 1:15 scale model of an America's Cup class yacht resides (when we visit a Version 4 boat has tactfully been placed in residence not to give away any secrets about the team's new Version 5 weapon) with cameras mounted on the turnable looking up at the sails from the bow, stern and beam.



How this model is located far from straightforward. The turntable allows the hull of the yacht to be rotated accurately so different wind angles can be presented to it, but other than rotating the model the turntable otherwise has no physical interaction with the model itself. To achieve this the model sits in a shallow basin of water - not evident from the photos thanks to the black turntable. The boat is attached at three points (the bow and stern quarters) by metal prongs and these in turn attach to sensors located below the turntable exactly measuring the movement of the boat. Measurement is now accurate to 0.85 Newtons (for those rusty on their physics, one Newton is the amount of force required to move 1kg at one metre per second). This accuracy, le Pelley says, is an improvement over his other set-up back in Auckland. "This is totally specialised for what we are testing here. Other wind tunnels have to do cars or environmental testing - we’re not interested in that, so anything we don’t need - we just get rid of it."

The model can be made to heel for the few occasions upwind sails are tested in the tunnel. This is made simpler as America's Cup class boats are typically sailed upwind at a particular angle of heel - 25-30 degrees. Another advantage of Cup boats is that thanks to their 20 or so tonne bulb on an all-up displacement of 24 tonnes they are generally sailed downwind near upright. The heel set-up for the models at the University of Auckland wind tunnel is a lot more complex as there they also test Volvo Ocean Race boats where heel angle is much more dynamic and affected further by positioning of movable ballast, etc.

The model is fitted with six mini motorised winches to trim the sails much like the full scale version except these are operated by a remote control. Loads in each of the mini-winches is accurately monitored and this data obviously logged, but also presented in a smart visual display, all the software to do this written by Le Pelley. The amount of line that goes in and out from each winch is also logged allowing for settings to be repeated.

"What we are looking at is the trim the driving force, the thrust force of the boats," le Pelley says. "Typically in a test session we might have three or four sails of a particular type, then often they have very very small differences. Normally in the wind tunnel we are looking at the smallest type of difference we can see, because already the sails are so refined with the America’s Cup. So you are trimming very carefully sitting here for quite a long time trying to get the maximum drive force from the sails."

Once the optimum trim is achieved the three cameras whirr into operation and the data is logged for around 90 seconds to achieve a reliable average result. A wealth of data is recorded, drive force, side force and vertical force being the important results. "Typically we are measuring about 20 Newtons of force with is about 2kg. And the accurate of the balance is to 5g so we can get a very fine resolution on the forces," says Le Pelley. The physical data of the set-up is also recorded such as the turntable angle (TWA), the wind twist profile, wind speed in several places, the temperature of the tunnel and atmospheric pressure,etc.

The turntable is then spun a few degrees, the optimum trim set-up and a fresh set of data recorded. Each sail they test in perhaps two different wind shear profiles and at six different wind angles, estimates Le Pelley. Then the data is examined and occasionally a test is repeated for example if the trim appears not to have been as good as it could have been: all in all quite a time consuming process.

While in the past model sails were built in strip cloth, today they are much more sophisticated. "We are having to move to construction like the full scale sails just to get the right tolerances and accuracies and get the sail to fly correctly," says Le Pelley. "The panel layout is the same as the existing sails and it gets plotted by a laser, so it is quite a careful process. There are about 30 panels in a typical model sail. So even the model sails now are taking a significant amount of time to build." At BMW Oracle Racing, the model sails are built down at the sail loft within their compound alongside the full scale sails.

One limitation is the material used in the construction of the miniature sails. At BMW Oracle Racing they use the lightest spinnaker cloth to take the load but unfortunately this is unable to replicate the stretch characteristics of the full-sized sail. "We are not applying [relatively] the same load to it in the wind tunnel so we don’t get the stretch. So we have to cheat a bit by lowering the head slightly so it doesn’t fly quite in the wind tunnel as it would at full scale. So the cloth will only wrinkle in a certain way and the wrinkles tend to be a bit bigger in the wind tunnel," says Le Pelley.

It is tempting to imagine that the loads recorded in the wind tunnel can be scaled up to full size, but Le Pelley says that is not how the data is used - results are viewed comparitively. "We are looking at the relative differencies in the wind tunnel. You can go further down the line and for velocity prediction programs you can start taking wind tunnel data to use in those, but really the America’s Cup has been going long enough that people know sensible sized numbers for things like velocity prediction. So really here we are just see what sail is fastest in the wind tunnel." Impressively he estimates that around 95% of the time sails that come out faster in the wind tunnel are also faster scaled up. "It is the other 5% where the skill comes in of the sail designers and the people analysing the sails, who say it has gone faster in the tunnel but we know in real life it might not be and we’ll have to take a bit of caution.

"Nothing is black and white even when it seems like science. It is all still a bit of an art to get the right answers out of a facility such as this. If you could scale it perfectly and it was a perfect world then we would have already designed the perfect sail and we wouldn’t need a wind tunnel and we wouldn’t be struggling for three years to try and make our sails faster and faster. But the trouble is it won’t be a complete replica of real life but we are striving to make it as close as we can."

In terms of practice, le Pelley and his team at the wind tunnel obviously work closely with sail designers within the team like JB Braun, but also the large number of talented sail makers within the sailing team itself such as Robbie Naismith, Zachary Hurst, Mark Bradford and Paul Westlake who can use the tunnel to try out new ideas.

"Earlier this year we were looking at trends and which way to go with certain sails as opposed to later on in the year when it will be a case of refining the sails for building for next year for winning the Cup," says Le Pelley. "The main advantage of the wind tunnel is that you can try out crazy ideas, take a pair of scissors to a sail and see ‘what if?’ Everyone down at the base has bright ideas and some of them work and some don’t but they can come down here and do things quickly and cheaply to see if that work is worth pursuing.

"It is also a very good step for the designers to be able to see the sails flying. There are computer programs that will try to predict how the moulded shape will look when it is flying but it is difficult to use. So to be able to build the sail and come in here and walk around it and take a pinch in the sail or see what a take up does, it is certainly far easier than building a new sail."

Even an experienced hand like Le Pelley remains impressed by the ability of the sail experts within the team to see the most minute differences between sails. "The sail designers have been designing these sails for year and while the sails are pretty similar from 50-60ft away they can tell me the difference between two sails that I can’t see. It is quite amazing even on the model sails, they can pick up the trends because they have been living these sails."

And this is just one aspect of the R&D work, going into one of the 12 teams competing for the America's Cup...

More photos on page 2...