The numbers come from the PERRINN 2017 open-source F1 CFD dataset
(windtunnel_data/perrinn_cfd_data.csv): sCz = 3.25 m²
and sCx = 1.16 m² at 40 mm front / 50 mm rear ride
height. Those coefficients already include the reference area, so force is
½ ρ v² sC — no second multiplication by frontal
area. Setup changes scale each component: wing load moves roughly linearly with angle while its
induced drag grows with the square, and the floor follows a ground-effect curve that peaks near
24 mm and collapses below it, which is the porpoising cliff the grid rediscovered in 2022.
The picture is a 2D potential-flow approximation: doublets give the body its
thickness, Rankine vortices carry the circulation, and every element is mirrored about the ground
plane so y = 0 stays a streamline. Circulation is divided between the front wing,
floor, diffuser and rear wing using the same split as the force model, so the flow and the readouts
always agree. Pressure colouring is Bernoulli:
Cp = 1 − (V/V∞)².
Streamline magnitudes are scaled for legibility — treat the flow as qualitative and the forces
as quantitative.
Run a lap is a quasi-steady lap-time model over the real pole lap's
centreline: grip-limited speed at every point from the downforce your setup produces,
a forward pass limited by the friction circle and by power against drag, and a
backward pass for braking. Track curvature comes from the FastF1 position channel,
smoothed with a Savitzky–Golay fit; DRS zones come from the pole lap's DRS
channel where it exists and from the two longest full-throttle straights where it
does not. Tyre grip and power were fitted so the fastest setup the model can find
lands on the real pole time at all three circuits — the residuals are in the
exported data. It is a first-order model: it has no tyre temperature, no wind, no
kerbs and no driver.
Grip assumes a peak slick friction coefficient of 1.8 on the combined weight and
downforce; Vmax solves drag power against roughly 470 kW at the
wheels, calibrated so a Monza DRS setup tops out around 355 km/h. Both are first-order estimates, not a lap simulation.
