(First written in august 2001, and updated for the Bahrain Grand Prix of 2004)
SET-UP: Setting up a car for a grand prix is a tricky business - no two circuits are the same. But there is a certain amount of predictability about the process.
Engineers will have worked on computer modelling back at base - with the circuit's co-ordinates plotted in - analysing the theoretical impact of set-up changes. The test team will also have put in mileage at a circuit, possessing corners with similar characteristics to the one about to be visited.
There are a number of 'set-up' variables that engineers have to work on. Then, of course, there's that universal variable known as the weather, with a car's handling affected by changes in temperature and wind direction, as well as whether it's wet or dry.
Once at the circuit, cars go out for two, one-hour sessions separated by an hour on the Friday in which teams work towards a 'balanced' set-up (where the car is moving efficiently around the track) from which they can develop the often very different set-ups that will work in qualifying and the race.
Computer simulation will have reduced the car's 'aerodynamic' settings from a potential 20, to just three or four. Allied to a car's aerodynamics are its ride height and suspension requirements. In a perfect world, engineers would run their cars as close to the track as the rule book allows, but circuits like Brazil's Interlagos are bumpy and running the cars low would lead to them 'bottoming' (where the underside of the car actually hits the track's surface) and the drivers losing control, so they have to be set up high.
Settings will have been tested first on a six-post testing rig - a device that simulates how the car runs on the track - at the team's base, leaving the team to work through five or six set-ups. The engineers also choose which of three or four brake materials is best suited to the characteristics of the circuit. Circuits such as Monza that ask for heavy braking require thicker discs and pads that offer the driver less feel.
Finally, engineers have to consider what gearing to use, looking for the gears that will make the most of the engine's power curve while not costing them top speed. A circuit like Monaco doesn't have a straight long enough to make top speed a factor, so gearing will be fitted for maximum acceleration.
Set-up planning for a new circuit though is somewhat trickier. For the Bahrain Grand Prix, the first at the circuit, preparation began many months ago, as Renault's Executive Director of Engineering, Pat Symonds, explains: "Logically, the earlier one can start the process of informing the lap simulation, then the better prepared one is; however, this is not always easy."
The starting point for a lap time simulation usually involves obtaining an accurate circuit map from which the trajectory of the car along the racing line can be described mathematically in three dimensions. Once the circuit configuration has been finalised, detailed maps are issued by the FIA, and work can begin. The boundaries of the tarmac are digitised, and fed into the first stage of the simulation programme, which uses mathematical 'cost' functions to determine the ideal racing line - this is what a skilled driver does intuitively.
"Once the racing line has been established, a car model with an 'average' set-up is introduced. From this, a variety of wing settings and gear ratios can be evaluated to get the basis of the set-up. Once this has been done, refinements are made to weight distribution and suspension settings in order to minimise the virtual lap time," says Symonds. "Subsequently, the team can begin to look at energy requirements from the tyres, to help choose the appropriate compound, and at this stage brake energy requirements, and to some extent brake cooling requirements, can be calculated."
However scientific this may sound, though - as indeed it is - a number of 'imponderables' can lead to errors in the simulation, that often cannot be corrected until the circuit has been seen and, indeed, until the car has first run on it. The first of these are the kerbs: Where the simulation is only able to assume a driver will use the limits of the tarmac, in reality, they use kerbs wherever they can. If they are relatively smooth and low, they will be used to shortcut what had previously been the 'ideal' racing line.
"Much more problematic, though, is the variation in grip levels," says Symonds.
Of the circuits raced on in 2003, grip levels varied by as much as 15 per cent, and if one includes the early sessions at Monaco in the equation, this figure rises to 24 per cent. Naturally, working 5,000 km from the circuit, the teams can do nothing but assume an average level of grip and work from there until further information is available.
"For Bahrain, however, we know the grip level to be similar to that at Silverstone, which is just one per cent below the average," says Symonds.
However, to put this into perspective, a three per cent variation in grip level on an average circuit can bring a change in lap time of around one second and just to make things worse, this parameter varies continually, even during the same day.
"In order to insure against this, the team conducts numerous simulations at different grip levels in order to have a bank of data at its disposal in the event of changes, so that the appropriate car set-up can be decided upon as quickly as possible," says Symonds. "Once that has been completed, and the car is running, we then begin running more simulations for race strategy. In fact, before making our decision on Saturday, over one million race scenarios will have run through the team's computers!"