Wednesday, June 29, 2011

F1 Practice Session Experiments - What's Going On?

You may or may not know that this week is National Motorsport Week in the UK (I wonder why they didn't mention it at the Isle of Wight Speedway last night???), Over on GoMotorsport the UK-based F1 teams have been taking it in turns to run a daily competition, and I think I may have picked up one of the spots on the Virgin Racing factory tour (not sure if I replied to the email invite quickly enough)...:-)

SAM_1383

Anyway, on the off chance that I do have a place on the tour, I thought I'd do a bit of prepping around some questions I might ask... The obvious thing that comes to mind relates to the role of High Performance Computing and the use of CFD in the design of the Virgin car, as well some of the reasons why CFD might not be able to deliver all that wind tunnel testing does...

The visit is a factory tour, though, so I guess something about materials and maybe the use of 3D printing in prototyping may be appropriate?

The extent to which on-track captured telemetry data is used to replay stresses and strains on various bits of the car in various test rigs is something else I'd like to ask about (I assume this does go on...), as well as the role of the strategist over the race weekend...

But something else that's been niggling at me lately, and that I may get a chance to ask about (if I've got the place, and if the opportunity arises), is this: what goes on during practice, how is that information used to plan out race strategy, and is there anything in the publicly available practice timing data that the armchair strategist can draw on to tryi to figure out what might happen during the race itself?

To ground the question, here's a snapshot of Car 25's fast laptime data over the practice sessions from the 2011 European Grand Prix weekend:



The symbols denote stint and the colours practice session.

To pick apart the possible experiments I think there are several different sorts of regime that may be worth clarifying in terms of what we might guess is going on.

Here's the second practice session, and in particular the latter two stints:



So my questions would be these:

- does the first of those example stints provide us with a tyre model (for which tyre? Can we identify tyres from the practice timing sheets?
- in the second example stint, is there an experiment going on that accounts for the +0.6s delta in the fourth and seventh laptimes (diamond symbol)?

Looking at data from the (other) Spanish Grand Prix, we see similar signatures:



In the third stint, the crosses show a couple of "but one" tours that are a delta of 1 second or so off the pace - what's going on here? Is it the same sort of test? (Maybe not using KERS or DRS on the lap?)

And for the last couple of stints in this session:



do we again get information here that the armchair strategist can put into a crude tyre model?

As ever, any and all comments appreciated:-)

PS Bah - a gremlin in my config files means the fuel corrected laptimes are off - i was using settings for Monaco:-(

3 comments:

  1. The old days of practice being used to explore the track and settings and find an optimal setup are largely gone. This year's car is run on a static jig to simulate the bumps and loads on the tyres and similarly most of the basic setup is done in advance by simulation and prediction.

    Practice then becomes a series of exercises to fine tune and calibrate things like absolute grip levels (tarmac changes from year to year as do tyres etc), fuel loads, optimal engine mappings for each corner to minimise tyre wear, aero tweaks for this year's weather etc.

    I wonder to what degree they might also experiment with different lines through some corners, just to see whether the new car works better where last year's car was unsettled or compromised by some aspect (eg this year's car controls the wheels better over bumps, so can take a different line into a corner without locking a wheel, and thereby makes a faster line achievable) - usually all you'd learn is that the old fast line is the new fast line, but it could produce big gains if you find the odd such item.

    ReplyDelete
  2. @Tim I appreciate that, but the questions still arise:

    1) there appear to be some experiments going on (e.g. the fast-slow-fast-slow-fast pattern) - so what tests do occur that have a signature at the admittedly crude laptime level?

    2) can the amateur strategist pick up enough from practice session data to make a crude tyre model to plug into a strategy prediction formula of the sort hinted at but not described in "How Do You Calculate An Optimal Pit Stop Strategy for a 2011 F1 Race?" [ http://f1datajunkie.blogspot.com/2011/06/how-do-you-calculate-optimal-pit-stop.html ]

    ReplyDelete
  3. Hi Tony

    Remember this isn't just an engineering exercise, the drivers ability (and possibility) to be consistent needs to be factored here.

    Therefore I think you're missing a little context to these figures, to draw meaningful conclusion.

    Certainly I would expected experiments to be going on. However you need to be able to match your lap time numbers with a) car configuration, b) driver consistency, c) track conditions.

    Maybe for the next GP you could try and track this and then cross-correlate each of those to lap time?

    It might be a challenge but then I am sure any secrets will pop out.

    Good luck!

    ReplyDelete

There seem to be a few issues with posting comments. I think you need to preview your comment before you can submit it... Any problems, send me a message on twitter: @psychemedia