Right now, all 10 Formula 1 teams are hard at work configuring their responses to the reworked aerodynamic regulations coming into force for the 2019 season. The outline of these regulations was known for some time, but the specific details that would allow teams to accurately model the cars was deliberately left until quite late – in order to minimise the head start the bigger teams would gain over the smaller ones simply by dint of their greater resources. Although wind tunnel time and CFD capacity are both limited by regulation, the number of resources – such as people and computing – let loose upon analysing the results varies greatly between the teams.
At this relatively late stage of the process, the aerodynamicists and the designers following behind them will largely have configured the car. The hard points – gearbox length, axle placement, monocoque shape, cooling capacity – will long ago have been decided, narrowing down the area in which the aero departments will be continuing to strive for more performance by manipulating the body surfaces.
How they go about this stage has evolved over the years and continues to do so, and it’s now about so much more than simply conjuring the best possible lift:drag ratio (ie maximum downforce for minimum drag).
There are basically two methods currently being used between the various teams. But both are looking to find the best compromise between the peak lift:drag ratio on the one hand and, on the other, the best combination of useable, driveable, downforce across the full range of operating conditions.
A given lift:drag number combined with a given engine performance will in theory give you a predicted lap time, but that’s of limited value if the driver cannot access the car’s full theoretical aerodynamic performance because its traits make it impossible to drive on its limit.
The current most commonly used method in aero simulation is the ‘weighted number’ technique. The model of the car will be tested in the full range of ride heights front and rear, yaw angles, steering angles and roll. This will be the car’s basic aero map. The aerodynamicists can then manipulate that map by prioritising the parts of it deemed most critical to lap time. So it may be that the ultimate lift:drag number is produced by a certain configuration, but that number falls away way too fast at a certain angle of steering because of the way the steered wheels are interrupting the flow to the barge boards and down the side of the car. In this example, a lower than ultimate lift:drag number might have to be accepted in order to give the required downforce consistency. The downforce levels at the various dynamic states of the car will thus be massaged into a whole. This is represented numerically by giving appropriate weightings of importance to each of the lift:drag numbers at the various dynamic states of the car. These are then combined to give an overall weighted aero number and any progress is measured by this number.
Whatever car spec has given the best weighted number by the time the crucial parts of the actual car build has to begin (ready for winter testing) pretty much defines the initial car.
Assessing which dynamic states carry which weighting involves some subjectivity on the part of the aerodynamicist. How well they work with the vehicle dynamicists is crucial in this. But there is a development of the technique relying less on number weighting and more on manipulating the shape of slices taken through computer-generated images of the aero map, with the objective being to achieve a nicely progressive shape change from one dynamic state to the next.
Whether this is more subjective or less than the weighted numbers method is a point of some debate. An experienced aerodynamicist with a great feel for his craft might tell you that the subjective ‘shape’ method actually more accurately represents reality because the weighting numbers can never incorporate all the nuances. As Colin Chapman once remarked: “Air is funny stuff.”
But regardless of whether the weighting is numeric or shape-based, in at least one of the top teams – and probably in three of them – the process is further refined by using the driver-in-loop simulator to give a simulated lap time for a given aero map. So you can put your actual driver in the simulated model using different aero maps and in this way finesse the fidelity of your weighting – so you actively improve your methodology as you go along.
But it’s still all simulation at this point. The reality of the track could still hold a nasty surprise. That’s what progress in driver-in-loop simulation is seeking to minimise. But it’s an ever-evolving technology itself – and the budgets are virtually limitless. A standard contemporary driver-in-loop simulator can be yours for about 4.5 million euros – and it will typically take a couple of years to get it working properly. But if you want it to do tricks such as incorporating with your aero maps, that’s a new research and development programme in itself – and there is no predicting how expensive that may be.
All the above is by way of illustrating how there is still black art among the science and how incredibly intricate it all is.
But to get to play at the margins of where the science meets the black art, to attempt to become nature’s master, requires gazillions of pounds of budget and investment. If most of those gazillions were removed, the teams would just find other ways to do the best with what they had and to us outside observers the cars would still race each other and be the fastest road-racing cars on earth. It isn’t only air that’s funny stuff – the same is true of money.
Since he began covering Grand Prix racing in 2000, Mark Hughes has forged a reputation as the finest Formula 1 analyst of his generation