Aerodynamic development uses several methods, none of which is definitive on its own.
Wind tunnels
Scale models tested under controlled conditions.
Which is expensive and heavily regulated in some series.
Computational fluid dynamics
Simulating airflow numerically.
Which is cheaper and depends entirely on model validity.
Track testing
Flow visualisation and pressure measurement on the actual car.
Which is the reference the others are validated against.
Correlation problems
Results disagreeing between methods.
Which has cost teams entire seasons.
Why no single method is trusted
Wind tunnels test scale models in idealised airflow, computation tests a mathematical model, and the track tests the real thing under uncontrolled conditions.
Which means each has a systematic error the others do not share.
Teams run all three and treat disagreement between them as the signal that something is wrong somewhere.
Flow visualisation
Paint, smoke and arrays of sensors on the track car.
Which shows where air actually goes.
Regulated development
Limits on tunnel hours and computing resource.
Which several series impose, sometimes weighted by championship position.
Scale effects
Model testing not reproducing full-size behaviour exactly.
Where gains now come from
Small refinements rather than large discoveries.
How development actually proceeds
Ideas generated computationally, promising ones tested in the tunnel, and the best of those taken to the track for correlation.
Which filters a large number of concepts down to a small number of parts that get made.
The proportion of ideas that survive to the car is small, which is why development resource restrictions bite hard.
Rake and ride height
Attitude of the car affecting floor performance.
Which is tested across a range rather than at one setting.
Aero maps
Downforce and drag as functions of ride height, yaw and steering.
Which feed the simulator.
Tunnel constraints
Model scale, wind speed and hours restricted by regulation.
Upgrades
Parts introduced through a season with measured expectations.
What a development cycle looks like
A concept is modelled computationally, screened, refined, tested at scale in the tunnel, then manufactured and validated on track.
Which takes weeks and is running continuously on many parallel ideas.
Most concepts are discarded at the computational stage, which is why computing resource restrictions have real competitive effect.
Testing restrictions in practice
Allocations of tunnel occupancy and computational hours.
Which some series weight against championship position to level competition.
Sensor arrays on track
Rakes of pressure probes measuring real flow.
Which is used in specific test sessions.
Yaw and cornering
Aerodynamic behaviour when the car is not pointing straight.
Which is harder to test and increasingly important.
Diminishing returns
Mature regulations producing progressively smaller gains.
Why correlation failures are so damaging
If the tunnel and the track disagree, a team cannot tell which of its development directions are real.
Which means months of work and manufactured parts may be worthless, and worse, the team does not know which months.
Several well-funded teams have lost entire seasons to exactly this, and recovering requires rebuilding confidence in the tools from the ground up.
Model construction
Scale models built to very tight tolerances.
Which are themselves expensive precision engineering.
Rolling roads
Moving belts reproducing relative motion between car and ground.
Which is essential for floor work.
Data processing
Enormous computational output requiring automated analysis.
Road car applications
The same tools used for efficiency and stability in production vehicles.
A note on sources and figures
Technical detail in motorsport is unusually well documented in some areas and closely guarded in others. Regulations, safety standards and championship structures are published openly by governing bodies. Setup data, aerodynamic figures and strategy models are competitive assets and are not.
What circulates publicly about the guarded material comes from team personnel speaking in general terms, from technical journalists with paddock access, and from the small amount that emerges through regulation disputes. It is generally directionally right and rarely precise, and anything quoted as an exact figure should be treated with some caution.
Where to look for more
Governing body technical and sporting regulations are freely available and are the authoritative source on what is and is not permitted. Specialist technical journalism, engineering society publications and books by former engineers cover the underlying principles properly. Team media output is informative and is promotional material rather than documentation.
Why any of this matters to a spectator
Motorsport is more interesting when you can see what is actually being decided. A driver lifting on a straight, a team pitting a lap earlier than expected, a car running a visibly different wing setting from its team mate: each of those is a choice with reasoning behind it.
Broadcast coverage has become much better at explaining this than it once was, and there is a limit to what fits between corners. Knowing the underlying mechanisms fills the rest in, and it turns a procession into something considerably more absorbing.