Reblog - Willem Toet - Power - More is Always Better, Right?

This is reblogged from Willem's LinkedIn account with his permission: https://www.linkedin.com/pulse/what-did-i-do-formula-1-willem-toet?trk=mp-reader-card

For UK followers of the blog the following opportunities exist for anyone interested in Willem's work...

2 Dec  18:30 to 20:30 plus Oxford - Free public lecture - F1 Performance, Design and (maily) Aerodynamics see toet.eventbrite.co.uk   There will be some entertaining stories and time for questions.  Free, book early.  This is the one to aim for if you work as it starts at 6:30 pm for 7 pm talk start point.   Organised by the IMechE.  Some refreshments available from 6:30 pm.   Questions and discussion due to finish at 20:30 but I'm happy to discuss any questions you may have for a bit longer.  I will bring additional material so we have the potential to illustrate answers to questions.

4 Dec - 13:00 - 17:00 approx.  Southampton University (Building 45 room 0045 which should be on the ground floor and is a large lecture theater - see site map here https://www.southampton.ac.uk/assets/sharepoint/groupsite/Administration/SitePublisher-document-store/Documents/About/visit/highfield_accessible_routes.pdf).   Guests welcome and free.   First lecture is similar to the one on the 2nd - Formula 1 performance, design, & aerodynamics.  This does not require specialist knowledge.   Would be suitable for higher school pupils, motorsport enthusiasts, engineering students and engineers.  After the first lecture we then focus more on the use of CFD to develop a race car (aimed at university students but anyone using CFD may find it interesting) and how new aerodynamic testing restrictions (in the FiA regulations) are changing the approach F1 teams are taking to aerodynamic development.   There will also be discussions with the Formula Student team which are probably not open to all. - See more at: http://www.somersf1.co.uk/2015/11/mexico-city-track-actually-is-about.html#sthash.nbelG4xp.dpuf

In F1 today differences between powertrains have led to unwise outbursts from some team bosses.  But how do teams know exactly how much others have?

The most powerful race car I ever made a contribution to was the 1986 Benetton BMW.   In qualifying trim it had somewhere between 1200 and 1500 horsepower.  A figure of 1350hp was often reported, but in the team we didn’t really know how much, exactly, because it did vary quite a bit (by engine and with engine use).   I remember (team driver) Teo Fabi describing a qualifying lap (with a special qualy engine) being like pulling the trigger (car accelerated like a bullet out of a gun) every time you went full throttle and doing the best you could to keep the car under control.  Then, where you ended up in qualifying depended to a large extent on where on your hot lap the engine would start to lose power - or blow up.  Most failures were literally explosive.  I also remember failures of the intercooler header tanks and cores, which were already quite thick aluminium, because of the combination of boost levels with high input charge air temperature (into the intercooler).

1986 Austrian GP start – Berger and Fabi leading after qualifying 1st (Fabi) and 2nd in the Benetton BMW’s.  Both qualifying engines held on for the whole lap!  In the race Fabi’s engine let go just as he took the lead, and Berger lost time in a long pitstop to resolve battery issue so didn’t score any points.  Image thanks to StatsF1.  According to Gerhard (Berger) and Teo (Fabi) they were fun cars to drive in qualy mode.   I believe it!

For today’s Formula 1 cars adding 1% more power (let’s call this 8 horsepower) the car will go about 0.14 seconds faster (varies track to track and with ambient conditions).   That can be several positions on the grid, as well as meaning a little more straight line speed.  The present status of F1 engines is highlighting differences between power units clearly, with really quite remarkable outbursts from the bosses of the teams that are behind.  The regulatory system of using up tokens for modifications (you can’t just design a new engine) makes it harder to catch up if you’re behind.  The token system also makes it a little harder for those who are ahead to run away even further.

In Formula 1 today the contribution from differences in powertrains to lap times is over a second a lap.  Chassis differences are greater than that.  The truth is that with the previous generation of engines there were also differences.  For example, when the Brawn team had to buy engines for 2009, they changed to a more powerful unit which did their championship hopes no harm at all.   Most reports of that year focus on the remarkable performance of the chassis/aerodynamic side, but the engine also played a significant part.   On the other hand, when BMW withdrew from F1 at the end of 2009, the Sauber team took a small step in the other direction with their new power unit.

How do you REALLY know how much difference there is?  Teams very often try to measure power via the gearbox input shaft (engine output) and / or via driveshaft torque sensors.   This always leads to arguments but also does not directly tell you what the opposition have.   At what is now the Mercedes team, from back when it was British American Racing (BAR), students were taken on for a gap year in the middle of their university studies.   Many were asked, during part of their time (typically around GP weekends), to do competitor comparisons.

How is this done?   Sound, vision and post-processing with software.  Let’s start with sound.  Analyse the onboard video for engine sound.  From the sound, reverse engineer the rpm.  (Four-stroke) engines fire every 2nd revolution and, depending on the crank arrangement, usually fire individually (one can make “big bang” engines where several cylinders fire together but you can “hear” if that has been done).  With every firing there is a shock wave.  Add the waves together and you have the tone of the engine at a particular engine speed.   A V12 will have a higher pitched sound at the same rpm than a V8.  The concept is simple once you have the idea to measure it.  So how does this help you?  Back when the teams started to do this, one of the first things they learned was how many gears a rival team was using.   F1 gear changes have always been fast but you can also listen to exactly how fast.   You can also work out rpm changes between gears, which pretty much tells you gear ratios.  The teams have access to sector times and sector top speeds.   As soon as you have one speed point and gear ratios, you have a calculation of speed at any time you have the continuous source of sound.
Briefly I’ll mention tyre growth here – with some types of tyres the number of revolutions per kilometre changes with speed.  Tyre “growth” sometimes happens with crossply tyres in particular.

These are small with F1 tyres and also well understood (everyone has the same sort of tyres nowadays).  Also for road car tyres it has become pretty insignificant.  In addition long lasting wheelspin can prove challenging to allow for.  If you spin the wheels the engine sound will tell you that you’re going faster than you really are.  There are ways to detect and allow for that too.  There is actually some “wheelspin” all the time when accelerating but that is getting too detailed.
So now you have speed against time for your competitors.   If you do the same analysis for your own car, you can compare.   You can also compare your actual measured data against data derived from sound analysis – important for correlation and confidence.  I’m sure you’ve worked out that comparing to real data from your own car is the place to start.   Since you know speed against time, you can calculate distance covered at any point in time.

Of course there is much more that can be done with competitor comparisons.   Just think how much more information you have from video.  Fortunately the cameras used in F1 have all been fixed-lens cameras (same zoom level, all cars).   Not great for TV (and I don’t know why they have stuck with the fixed lenses for so long), but useful for engineers.  This means that every pixel in the image comes from the first visible object along a fixed vector direction from the lens of the camera.  So, if you know what you are looking at (e.g. the GPS coordinates of every visual marker in the individual images of the video – you want three coordinates to triangulate), you can work out not only the position of the car but also the attitude it has relative to the track.  All you need to do is reverse engineer the coordinates of every painted line, advertising hoarding, grandstand seat, bridge etc. and you can work out the line the car is taking but also how quickly things change – such as rate of change of yaw (speed of direction change) and stability.  The pictures can, of course, provide velocity against time, to further refine what you have from sound.  In particular brake lockups make gathering speed against time data difficult with just sound.

So what is done with this information?   The first thing is to do a “compare elapsed time” or a “delta time” plot.   A plot showing how much more (or less) time than a competitor you have taken to arrive at different points on the track.

Competitor Comparison between Panis and Coulthard from 2000.   Image thanks to BAT.  Horizontal axis is the distance from the start finish line, the vertical axis shows speed (one car) and the delta time between the drivers.  In this example the loss is almost continuous (Panis slower), but it is particularly evident in two braking areas (that’s where the orange data line drops more steeply).

Most teams have driver in the loop simulators but all have purely software-based lap time simulations (probably more useful for this analysis, and that was all that BAR had access to at the time).   By tweaking parameters in the software simulation you can try to match what a competitor must have in terms of mechanical grip, aerodynamics, powertrain characteristics, weight (some teams run more weight on Fridays than others).   At BAR I’d get a report after each event that told me how good or bad the car was for such parameters.  It can be useful in deciding where to deploy resources for the future.   On the other hand it can mean that everyone points fingers.   Drag and horsepower are possible to distinguish but not perfectly and back at BAR we were always told the car had too much drag.  Later we learned, when the team changed engines, that there was a significant power deficit.  This information is useful, provided you have enough resources to spend on development.  In hindsight I’d say that BAR spent too many resources in areas that made too little difference (they had lovely trucks and motor homes at the track for instance), and not enough in the most important area they were able to directly influence.  Of course I have to admit to being biased – sorry but nobody’s perfect.

There is lots more I’d like to say about power and how it influences performance but that will have to wait until another post.  And of course more power is not always better, either politically speaking, or in a car.

UPDATE - I've inserted two received comments here because they have engineering information in them.
Ian Wright  (Chief Engineer - Driving Simulator Laboratory at Pratt & Miller Engineering)                Another very good article. As the person who actually did the competitor comparison analysis in the early days at BAR and then led their development when Honda got involved, I can say that it is pretty much spot on. The more advanced competitor comparison toolset (after Willem left BAR / Honda) was used to accurately predict the reasons for small differences in competitor performance (aero and engine).

Greg Davies   (Manufacturer of Machined Components).                            
Great post, it reminds me of Jonathon Palmers technical contributions to the BBC's F1 coverage of the mid 90's. I was always fascinated by how much detail goes into everyday events that are normally taken for granted. It was just this type of article that got me interested in engineering and keeps me interested to this day. On the subject of power, Red Bulls championships were all won with engines that were less powerful than there rivals. They even won on the traditional power tracks like Spa and Monza. Regarding the current turbo era, whilst I like the technology I don't like the way that its been forced into the sport. Turbo's 1st time around came into the sport on merit. During the development era this led to some great competition between the competing technologies. Would the 1st turbo era have been held in such high regard if it had never had the chance to overhaul the then status quo?

This is reblogged from Willem's LinkedIn account with his permission: https://www.linkedin.com/pulse/power-more-always-better-right-willem-toet?trk=prof-post

For UK followers of the blog the following opportunities exist for anyone interested in Willem's work...

2 Dec  18:30 to 20:30 plus Oxford - Free public lecture - F1 Performance, Design and (maily) Aerodynamics see toet.eventbrite.co.uk   There will be some entertaining stories and time for questions.  Free, book early.  This is the one to aim for if you work as it starts at 6:30 pm for 7 pm talk start point.   Organised by the IMechE.  Some refreshments available from 6:30 pm.   Questions and discussion due to finish at 20:30 but I'm happy to discuss any questions you may have for a bit longer.  I will bring additional material so we have the potential to illustrate answers to questions.

4 Dec - 13:00 - 17:00 approx.  Southampton University (Building 45 room 0045 which should be on the ground floor and is a large lecture theater - see site map here https://www.southampton.ac.uk/assets/sharepoint/groupsite/Administration/SitePublisher-document-store/Documents/About/visit/highfield_accessible_routes.pdf).   Guests welcome and free.   First lecture is similar to the one on the 2nd - Formula 1 performance, design, & aerodynamics.  This does not require specialist knowledge.   Would be suitable for higher school pupils, motorsport enthusiasts, engineering students and engineers.  After the first lecture we then focus more on the use of CFD to develop a race car (aimed at university students but anyone using CFD may find it interesting) and how new aerodynamic testing restrictions (in the FiA regulations) are changing the approach F1 teams are taking to aerodynamic development.   There will also be discussions with the Formula Student team which are probably not open to all.