Control F1 car energy recovery system

[SBR9]

Very interesting work going on at the University of Oxford. I believe some of the work is being done in collaboration with Ferrari. http://control.eng.ox.ac.uk/Vehicular_Optimal_Control and http://ch.mathworks.com/videos/optimal- ... 96763.html also nice summery of the some of the work: http://mccabism.blogspot.co.nz/2015/04/ ... raris.html

[chuckdanny]

iNdeed thanks!

[gruntguru]

Thanks.

[Cold Fussion]

In the MathWorks video he showed the model using drawing energy from the ERS and the kinetic to the ERS (around the 21 minute mark) but it wasn't clear (to me at least) if the model allowed for the mgu-h to power the mgu-k directly.. Does anyone know if it does or not? Modeling that would complicate the energy usage quite a bit.

[gruntguru]

I am sure it would be in the model since the rules allow it and it is more efficient for the car to use the MGUH energy than to store it. How much would that complicate the modelling?

[R_Redding]

but it wasn't clear (to me at least) if the model allowed for the mgu-h to power the mgu-k directly..

A Big thanks to Gordon McCabe for presenting this information.



As Mr McCabe points out ... the information used in the papers and video is circa late 2013.

The newest of the publications (IEEE-CSM-14-0038.pdf)... the data published in the IEE Controls magazine in 2015 states that..
"It is not possible to identify easily the contributions being made by the MGU-H, but they will be reflected in discrepancies between the ESs state of charge and the MGU-K usage." (pg54 top left)

It seems that the control scheme presented was an early approach mainly utilising the MGU-K , ES store and turbo wastegate ,the MGU-H used primarily for ICE boost..and may not have realised the importance of a bigger compressor/turbo (Marmorini is quoted as saying aero packaging was priority over ICE size for 2014) and a greater importance and reliance on MGU-H energy as Mercedes did presumably.

Rob

[Wayne DR]

I am sure it would be in the model since the rules allow it and it is more efficient for the car to use the MGUH energy than to store it. How much would that complicate the modelling?

I think I remember reading somewhere that was the downfall of the Ferrari PU in 2014, they could not transfer power directly from the MGU-H to the MGU-K without it passing through the Energy Store. The whole presentation seems to totally ignore the nearly constant "Sustained" power that can be harvested from the MGU-H.

The numbers don't seem real. I would guess at peak around 80-90kW to run the compressor not 40kw, and the turbine would nominally put out say 160-180kW, leaving 80-100kW for the MGU-K... However the 20kW boost from fully opening the waste gate is probably real, if not slightly overstated!

[gruntguru]

Yes 40kW compressor power would equate to about 2.4 bar MAP. Early calculations may have been based on that sort of number and an AFR closer to stoichiometric.

[Wayne DR]

It seems that the control scheme presented was an early approach mainly utilising the MGU-K , ES store and turbo wastegate ,the MGU-H used primarily for ICE boost..and may not have realised the importance of a bigger compressor/turbo (Marmorini is quoted as saying aero packaging was priority over ICE size for 2014) and a greater importance and reliance on MGU-H energy as Mercedes did presumably.

I must say it took me a while to get my head around the potential of the Brayton Cycle power output from the compressor/turbo/MGU-H, which can be fed directly to the MGU-K constantly and without limit. Mercedes certainly capitalised on this. (I wonder how much was Andy Cowell and how much was Aldo Costa, if it was more the latter, Ferrari should rue letting him go!)

That is a step change from the previous KERS systems (automatic and continuous deployment) and possibly not immediately obvious.

Ferrari knew they were sacrificing power by reducing the compressor/turbo size, they just underestimated how much. To be beaten by the "Bibitari" (soft drink makers) is somewhat acceptable (as they have Adrian Newey), whereas to be beaten by another car/engine manufacturer with outright power and straight line speed is totally unacceptable (and very embarrassing)...

If Marmorini missed the importance of the Brayton Cycle and the potential of the unlimited MGU-H to MGU-K power link, this could explain his departure...

[Wayne DR]

Yes 40kW compressor power would equate to about 2.4 bar MAP. Early calculations may have been based on that sort of number and an AFR closer to stoichiometric.

Yes, that sounds about right for 2.4 bar (for stoich conditions), but leaner conditions with higher airflows (an AFR of say 1.4) takes boost pressure over 3.4 bar and would require much more power (around the 80kW). This is a significant design flaw, as a 45% airflow increase could require a full redesign of the cylinder heads, intake and exhaust systems.

If this was presented in 2013, surely Ferrari were considering leaner running by then?

[gruntguru]

This is a significant design flaw, as a 45% airflow increase could require a full redesign of the cylinder heads, intake and exhaust systems.

Not really. Increasing MAP increases mass airflow while volumetric airflow stays about the same. Flow passages do not need to be increased unless velocity (proportional to volume airflow) increases.

[Wayne DR]

Not really. Increasing MAP increases mass airflow while volumetric airflow stays about the same. Flow passages do not need to be increased unless velocity (proportional to volume airflow) increases.

You are right (as usual...)

I have only worked on NA engines, where restrictions that drops MAP, impact airflow and therefore performance. There is 45% more air, but it is (roughly) 45% is denser.

[DaveW]

Thanks people. A fascinating story.

On reflection, however, I am surprised that Professor Limebeer, et al. were allowed to publish late in 2014. Perhaps it coincided with Ferrari's dramatic turn over of staff in 2014 and with the sudden gain in performance of the vehicle in 2015, or am I being cynical... Maybe the reported installation of a full scale power train dyno was significant, too.

[gruntguru]

I haven't read anything he published on the matter, but the presentation was pretty general and certainly wouldn't have contained any surprises for the opposing teams.

[autogyro]

DW stated
Maybe the reported installation of a full scale power train dyno was significant, too.

Surely the need for a full power train dyno facility has been recognised by all the teams and power train developers and many years ago.
If not then I am not so far behind on development ideas as I thought I was.
I am also very surprised that Ferrari lacked the technology.
It was something I thought about over ten years ago.
Could this be because of a miss placed faith in remote power train development by the engine suppliers/builders only and or a cost issue.

I demonstrated such a full power train dyno for Gordon Murray at McLaren in the 90s.
It was a mobile hub mounted dyno for use with anything up to the fully assembled car at the circuit.

McLaren did not purchase one at the time.
Going by what DW has said of Ferrari's new acquisition, perhaps McLaren should get in touch with me again as it is perhaps something they are now in need of

If I am right then McLaren stand little chance of catching Merc and Ferrari now they both have such a lead on development in house.
Obviously Renault has dropped the ball somewhere, probably on compressor size and H energy control.
Honda could get up to speed with enough cash but they would have to bring a complete power train development factory to Woking to be certain.
Things change and hybrid systems require a lot more hands on work at least that is my opinion.
And all this with Bernie moaning about cuts in money seems a shame Max is no longer about, I dont think Todt would know a dyno from a bowl of flowers.