How to Lose F1 Races with a Wind Tunnel and CFD

I recently came across a blog post "How to lose races with CFD" [source: Another Fine Mesh] related to the CFD-only strategy pursued by the Marussia Formula 1 Team (formerly Virgin Racing) during the 2010 and 2011 F1 seasons. It seems such an easy conclusion to make, right? During its first 2 F1 seasons the Marussia team garnered a lot of publicity thanks to their CFD-only (no wind tunnel testing) strategy and during that period they didn't win a race, therefore CFD must lose races, right? Not so fast.

The fact is during the 2010 and 2011 F1 seasons only 3 teams out of 12 won any races - those teams being Red Bull, McLaren, and Ferarri. So 75% (9/12), the vast majority, of F1 teams consistently lost races and of those teams 89% (8/9) used a combination of a wind tunnel  (WT) and CFD to develop their cars. Therefore, correcting for the faulty logic of the headline writers, we could say that WT + CFD loses F1 races - the correlation is much higher at 89% for WT + CFD, compared to 11% (1/9) for CFD only.

I think the headline writers are really trying to highlight the fact that Marussia didn't only lose races (plenty of teams do that), but the team's performance was particularly poor in that it hadn't scored any constructors points since entering F1 in 2010. But this was also true for 2 other teams, Team Lotus and HRT. It's interesting to note that Lotus and HRT both entered F1 in 2010 at the same time as Marussia. Turning to statistics again, we arrive at a 67% (2/3) correlation between WT + CFD to score zero constructors points compared with a 33% (1/3) correlation with CFD only. Still not a favorable correlation between CFD only and poor race results.

I think the correlation between results and WT + CFD compared to the correlation between results and CFD-only is dwarfed by the correlation between results and team funding. I don't think the Marussia CFD-only strategy can be judged in isolation from the fact that Marussia, along with Lotus and HRT, were all poorly funded  teams in F1 terms. This means they can't attract the best aerodynamicists. Until a top (i.e., well funded) team like Red Bull, McLaren, or Ferarri with proven race winning aerodynamicists tries a CFD-only strategy then I don't think we can pass judgment on whether CFD only loses races. However, this is unlikely to happen any time soon, in the same way no F1 team would consider a WT-only strategy, because WT + CFD together are such a formidable combination.

In my opinion, the Marussia CFD-only experiment proves nothing other than under-funded teams struggle in F1.

The great thing about racing is that the results will speak for themselves. It will be interesting to watch Marussia in the new 2012 F1 season and see if there is a marked improvement in their results now they have adopted a WT + CFD strategy while still appearing to suffer from restricted funding.

Comments

Marussia found that their

Marussia found that their cars lacked performance in high-speed corners
=>Their CFD / CFD is currently not adequate to simulate F1 cars in in high-speed corners

So said the team

"CFD is currently not adequate to simulate F1 cars in in high-speed corners"

So said the team's new technical consultant Pat Symonds - do you believe it? So with a little tweak we will see a competitive team in 2012? Great news for sponsors then. Could there be a conflict of interest in what teams say (to attract/retain sponsors) and what is actually happening?

If high-speed corners were the problem and assuming low speed corners were fine then we should have seen markedly better results at Monaco, but didn't.

It's easier to make a scapegoat of CFD and the newly departed CFD team, than admit the reality of funding in F1.

Here are a couple of

Here are a couple of questions which probably show my ignorance. Who supplies the aerodynamicst? From your blog it sounds like the team does and not the analysis house. And, who interprets the results, determines uncertainties in the modeling of the physics, and determines the margins? Is it the analysis house/CFDer or the team? It sounds, at least to me now, that running CFD and determining physics+consequences is uncoupled.

Aerodynamicist is decoupled from CFD

Typically F1 teams perform all aerodynamic design and CFD in house. Whether the aerodynamicists in Marussia's case were employed directly or through Wirth I don't know.

Often in an F1 team the aerodynamicist is decoupled from the CFD and the wind tunnel. I.e., the aerodynamicist doesn't perform the simulation or test. However, they will likely work very closely with the CFD and wind tunnel personnel to determine "physics + consequences".

Sorry, I don't mean to

Sorry, I don't mean to belabor this. But I come from a background where aerodynamics and CFD are highly coupled (a small aeronautical consulting company where the aerodynamicist is the CFDer). So, this is the way I think. Therefore, in the case at hand, it is not likely that the CFD modeling was bad, i.e. it fell out of expectations of a knowledgeable aerodynamicist. I assume the results are what they are. However, there is a higher probability that aerodynamicist was not (not sure of a nice way to say this) up to the task.

CFD at Marussia

Marussia's technical consultant, Pat Symonds talks to the BBC about IT, including CFD, at "Formula 1's IT crowd: Software engineers power Marussia".