McClaren F1, driven by IT

McLaren F1, or rather West McLaren Mercedes: the very name signifies extreme endeavour in what is without doubt the highest tech, most expensive and glamorous sport on earth. The figures speak for themselves: last year’s McLaren MP4-16, powered by a 3 litre 10 valve Mercedes engine, develops an awesome 850hp, launching the car from a standing start to 185mph in just 3.6 seconds. The machine pulls three of four g as it eats up the course – and it’s capable of slamming back to rest, with brakes burning, in just 3.7 seconds. Every race weekend McLaren gets through 10 or more engines at $300,000 apiece, and they’re never raced again. Each machine is hand built to fit its driver – names that are legends among the 360 million in 200 countries that view every Grand Prix. And exotic materials range form carbon fibre, to honeycomb aluminium, titanium, kevlar and gold leaf. Even the individually designed steering wheels cost $60,000! The design, engineering and manufacturing effort involved is… well, prodigious is too small a word. Yet the team, while rightly proud of its heritage and sheer engineering achievements, notes that in some respects its just like any other manufacturer – except that time to market is only ever that left till the next race. Every hour counts, and there are lessons here. Indeed some of them McLaren is harnessing for itself. The company is currently putting the finishing touches to its $200,000 Mercedes SLR McLaren, to be manufactured at its huge new Paragon facility near Woking when it opens in August. And with volumes of 500 per year, it’s having to attend to familiar planning and scheduling and production management issues. But back to Formula One and design – and it’s all powered by technology partner Sun Microsystems hardware. With every season presenting new challenges, new race regulations and 90% of the car to redesign in just six months, the pressure is on to cut design cycles to the bone. So McLaren now uses two giant Sun TCF (technical compute farm) systems to handle all its solid modelling, FEA (finite element analysis) and CFD (computational fluid dynamics). It runs the Catia CAD/CAM/CAE suite for totally concurrent engineering, analysis and design for manufacture (74 seats across everything), with Enovia VPM for product data management (PDM) and change control. The system serves some 30 Sun Blade 1000 workstations, each with the UltraSparc III 64bit processor, and allows design all the way up from the component level, through large assembly in-context design to simulations of the full car in real time. This is totally virtual reality design and build at its best. TCF itself is a rack system integrating Sun’s enterprise servers, compute engines and storage arrays, plus networking kit and resource management software – it’s massive distributed computing. Kevin Colburn, McLaren’s CFD team leader, says: “We’re on a par with the whole of BAe in terms of processing power here.” And it’s used for absolutely everything. On CFD, for example, the TCF runs all the airflow simulations, enabling the team to refine designs fast, well before having anything physically built for wind tunnel testing. Colburn: “Historically, we looked at 10 or 15 wing configurations before selecting the best... By optimising designs on the TCF we cut this and use the tunnel much more effectively.” And it means hugely faster design cycles and much less physical build and testing costs. As for the CAD, McLaren doesn’t pretend the switch to 3D, and all the power it provides, was easy: Scott Bain, senior design engineer, makes the point that, like everyone else, most F1 designers came from a 2D background. But clearly it was well worth the learning curve. Bain: “There are a host of benefits from working with solid geometry. You can assemble bits of the car together and actually see a meaningful representation of the assembly, instead of just bits and pieces of partial geometry or incomplete models. You can work out the centre of gravity of complex assemblies very accurately and very quickly.” Single integrated view Then once the designers have created solids for all the major components, like the chassis, they can offset from them for all other internal details. On other components, like the gearbox main case, after the designers have modelled them as solids, they can change them very quickly if features need to be added as the development progresses. And, importantly, McLaren design engineers can transfer the solid data easily and directly to analysis and manufacturing. Bain adds: “The data integrity of Catia is so rigorous it forces people to model properly and not to cut corners and get away with the bare minimum in an effort to save time… Now we see the car evolve months before it ever goes near a track. We can spot errors immediately, or double check something that doesn’t look right. If we weren’t all working from a single master model that just wouldn’t be possible.” Clearly, when McLaren says it’s used for everything, it means it. Indeed, the team says it generates some 4,500 drawings to produce the car’s 10,800 components in its massively shortened design cycle – equivalent to 22 telephone directories. And the pace of design continues right through the season. At last year’s Austrian Grand Prix, for example, the car featured a specially redesigned front wing based on data resulting from simulations run on the TCF just the previous Sunday. Only by having everything on the system is it possible to change components in this way specifically for race circuit characteristics. And it doesn’t stop there. There’s also the track side computing. Sun systems go to all races in the form of the Battlestation – racks of ruggedised Sun servers, storage and workstations feeding data to McLaren race engineers and analysts at track side and via ISDN to Woking, as well as to the ‘pit-wall’ system for the race strategists. As the car speeds round the circuit, real time telemetred data from its 120 sensors monitoring some 190 parameters (from wheel speed, to ignition timing, temperatures and component performance) is fired back by radio to the Battlestation. Sun’s Java technology is used to sift and disseminate the data in appropriate formats to the race users, enabling them to deal with events as the race unfolds. McLaren makes the point: Java code never fails to deliver exactly the data needed when it’s needed, and it integrates across all platform types. No one is saying it’s not extreme. But there are aspects of this we should all be noting.