The numbers that reveal F1's aerodynamic efficiency improvements

Feedback is always welcome but when one of our loyal readers, Ben Halford, asked me to write about the historical development of aerodynamic efficiency, I hadn’t appreciated what a rabbit hole
I was about to disappear into.

Firstly we need to establish what we mean by aerodynamic efficiency and how we express it. In this context efficiency would be expressed as the total downforce divided by the total drag. Here comes the first problem. While it’s been possible (albeit difficult) to pull together some numbers for downforce, it’s been much harder to get drag figures – so for the most part we’ll focus on downforce here. 

Again we need to define how we will express downforce. It could be expressed as a force but this would then depend on speed. To get round this engineers use a ‘non-dimensional’ number which is generally known as the lift coefficient, or Cl, downforce just being negative lift. This should mean one car could be directly compared with another. Unfortunately this isn’t that easy for two reasons.

Firstly, to arrive at a lift coefficient one needs to equate it to a reference area which, in vehicles, is normally the frontal area. Now the frontal area of a Formula 1 car is just under 1.5 square metres so many teams will use 1.5 as the reference area. Others are more pedantic and will use a more accurate number – say 1.47 – while others still will say it’s only a notional thing, so let’s use one square metre. This means that if we quoted our lift coefficient as, say, 4.5 based on 1.5 square metres it would be 6.75 if based on one square metre.

It gets even worse, though. We like to express the lift coefficient as a single number but of course that number changes depending on the ride height of the car, the steering angle, and the angle of the wind amongst other things.

Aerodynamicists get around this by quoting a weighted number with the weighting they apply to each individual condition based on its effect on lap time. So, for example, the drag is much more important when the steering is straight and the ride heights are low – in other words at the end of a straight – than it is when the steering is turned and the ride heights are high, the conditions you would find in a slow corner and where the downforce is more important.

The Lotus 72 is the earliest F1 car with a lift coefficient and demonstrates how poorly it compares to modern day cars

Photo by: Motorsport Images

It’s also true that wind tunnel testing, which is where these numbers come from, has got significantly more precise since some of the earlier numbers were determined. So too have the methodologies for giving useful weighted values.

So, with these caveats, what does history tell us? Many people think the skirted ground-effect cars of the 1970s had enormous downforce. I’m sorry to disappoint you but they didn’t by modern standards.

The earliest numbers I can find are for a Lotus 72 which was scanned and analysed in CFD. This, as a non-ground effect car with a very high aspect ratio wing, was dreadfully inefficient with a Cl of 0.38 and a Cd (drag) of 0.74. The efficiency was therefore only around 0.5.

PLUS: The troubled story of F1's greatest racing car

The earliest numbers I can find for a skirted car are for the 1981 Ensign which has been analysed in CFD by Bolton University under Professor Willem Toet, a very experienced aerodynamicist who spent many years in Formula 1. This shows a downforce coefficient of 2.76 with drag at 1.13, so efficiency has jumped to 2.44. Of course this wasn’t a particularly successful car and the dominant Williams FW07 was undoubtably significantly better than this.

The new generation of cars for 2022 introduced much more ground effect and, while there was a small initial drop in performance it’s rapidly approaching 2021 values again

In 1983 ground effect was outlawed and numbers dropped significantly. In fact it took until 1987 before Cl exceeded 2.0 again. However this was also the time that aerodynamic research began to become more scientific and progress was rapid: Cl peaked in 1994 at over 3.5.

The regulation changes mandated after Imola ’94 brought an immediate drop and at the start of 1995 Cl was back to around 2.1. Progress was rapid from this point but in 1998 the cars became narrower which, together with other restrictions, pegged them back to around 2.25.

By this time Computational Fluid Dynamics (CFD) was vastly improving understanding and by the end of 2004 Cl had reached 3.2. A reduction in diffuser size for 2005 lost around 20 points, bringing it back to 3.0, but then progress was steady at around 10 to 15 points a year until the extreme regulations for 2009 were introduced which should have dropped downforce significantly. The infamous ‘double diffuser’ negated this and actually, by the end of 2009, Cl remained above 3.0.

The move to narrow body cars for 1998 pegged back the CI around 2.25, but this increased as CFD understanding improved

Photo by: Motorsport Images

Continuing improvements in wind tunnel test techniques and further gains in CFD together with the exploitation of blown diffusers yielded a steady 15-point yearly improvement. By 2013, Cl had reached around 4.6. However, the raft of new rules for 2014 wrought a drop back to around 3.7.

The next big change was the wider cars for 2017. These regulations allowed for more aerodynamic development freedom, which saw extreme gains of 60 points in the first year and improvements of around 20 points a year with the 2021 cars reaching Cl values approaching 5.5.

The new generation of cars for 2022 introduced much more ground effect and, while there was a small initial drop in performance it’s rapidly approaching 2021 values again.

Of course these numbers need to be weighed with caution. Techniques have changed – it wasn’t until the 1990s that wheel lift was even measured – and of course the weighting applied to the plethora of results by different teams can have a profound effect. When I was at Williams an aerodynamicist joined us from a rival team and told us some numbers his previous team were heading towards for 2014. I knew we wouldn’t achieve such numbers and was very concerned.

Ultimately we beat his former team by a mile in the constructors’ race so numbers are just numbers. Performance is a different subject.

Only Red Bull's data will truly reveal how strong the RB19 is - while performance is a different challenge entirely

Photo by: Jake Grant / Motorsport Images