How do F1 Power Units Know When To Harvest?

[peanutaxis]

I don't see the drivers pressing buttons, so......how do the engines know when to harvest and charge batteries down the straight? Surely it can't be speed because that would compromise you if you are ever slow down the straight for traffic or something. So .... it must be GPS?

[George-Jung]

Probably they can program it in the software? So it knows due to throttle and brake in put where the car is on the circuit?

[Jolle]

Probably they can program it in the software? So it knows due to throttle and brake in put where the car is on the circuit?

and GPS....

[Frafer]

it's the kind of topic that's a bit too complicated in order to be meaningful assessed here. Let's say there are some software that handle the PU in every aspects, evaluating tons of variables, but for the sake of simplicity we'd say that the K charges the battery during braking, and so does the H (sort of); and that energy is deployed somehow (directly or not) to the crank in the first part of every straight; at the moments before braking a lot of things can happen, depending on mapping.
But still, we are miles behind a serious understanding of how things actually work

(of course gps position is one variable, like throttle pedal position, brake pedal position, mapping, energy managment settings, fuel managment ecc ecc)

[rekoildale]

I have heard or read that there ar no rear brakes they use the generator to slow the rear of the car and that was the initial problems with corner balance in the beginning of this era. But as for harvesting i think it would be all software as the parameters needed are all captured in microseconds. The rear harvest light occasionally has flashed on the straight and the cars nearly stop and they said at the time it was a malfunction of the software.

Sent from my SM-N910G using Tapatalk

[Sevach]

They can program it according to speed, which gear the car is in, and in some cases distance based(not "wasting" power between the 2 Lesmos for example).

Normally you get small doses of power in the lower gears, full in the middle gears and it goes into harvesting as you approach top speed(race mode we are talking here).

[wuzak]

I have heard or read that there ar no rear brakes they use the generator to slow the rear of the car and that was the initial problems with corner balance in the beginning of this era. But as for harvesting i think it would be all software as the parameters needed are all captured in microseconds. The rear harvest light occasionally has flashed on the straight and the cars nearly stop and they said at the time it was a malfunction of the software.

Sent from my SM-N910G using Tapatalk

As the rear braking energy recovery is restricted to 120kW it would be impractical to do without hydraulic brakes.

The rear brakes are smaller than they were up to 2013 because of the electronic assistance, and they fail quickly if the MGUK fails.

[J.A.W.]

Here O.P., this is all you need to know, subject-wise..

https://www.youtube.com/watch?v=QCRAUKWKKlo

[gruntguru]

At the most basic level the control algorithms could be made fairly simple:

MGUH. The engine management system demands a compressor speed setting based on engine speed (and throttle).
Under heavy engine load the MGUH will have to brake the turbocharger (harvest energy) to maintain that speed.
Under light engine load, the MGUH will have to motor the turbocharger (draw energy from the ES).

MGUK. The engine management computer adds power to the MGUK according to throttle setting starting at say 95% throttle and increasing to max (120 kW) at 100% throttle.
The engine management computer subtracts power (harvests) from the MGUK according to brake pedal position starting at 0% throttle and increasing to max (120 kW) at say 5% brake pedal position.

It is also likely that regenerative braking is to some extent linked to throttle position so that the lowest few percent of throttle operation actually controls harvest from the MGUK producing an effect similar to engine braking.

The above are very simple algorithms/tuning-tools - simpler than those already used for engine mapping.

EDIT. Not suggesting for a minute that the actual algorithms used are anywhere near this simple.

[henry]

I think there is also a mode in which the torque demand at low engine speed/partial throttle is met by running the ICE against the MGU-K and so charging the ES.

Given that the ICE will be more efficient at higher torque output the overall PU efficiency will be higher if the round trip losses, MGU-K to ES -and back to MGU-K, are lower than the difference in ICE efficiencies. This calculation would also need to include the MGU-H which might use less charge or contribute depending on the operating point.

However. The goal is not PU efficiency but lap time reduction. So in some circumstances there is probably a case to be made to burn additional fuel at times when low power is demanded, in the twisty bits, in order to have longer MGU-K deployment on the faster bits. This would mean that the charging strategy would be circuit dependant, relying on simulations to establish a State of Charge versus position in lap, providing a target for the ES algorithms. Whether this is done or how they might establish position in lap I don't know.

I recall seeing a report from Andy Cowell that the Mercedes ES algorithms are adaptive so that if something changes, say wind direction or tyre performance, the deployment will change over the course of a few laps. It was suggested that Hamilton's problem in Baku would have righted itself over the course of the race.

I would hazard a guess that even within the teams few people actually know concepts behind the strategies let alone the detail of how they are implemented.

[PlatinumZealot]

It is set by algorith then manually programmed. Usually the human programming is needed because real life running is always different from simulated.

honda for example still spend considerable timing manually tweaking the ERS strategy for each race.

The brakes is also the same thing. note how Haas needs like two weekends to fine tune the brakes.

The systems are so transtransient that the computer just doesnt know the big picture.. Even if you use multiple cascaded control loops the computer still does not know the big picture. You have to feed it the "feed forward" part and it will do the rest.

[wuzak]

How do F1 Power Units Know When To Harvest?

Black magic, obviously.

[henry]

Further to my suggestion above that one of the things that might be controlled is state of charge(SOC)versus track position. I have been thinking about what might be involved in relation to Andy Cowell's comment about wind direction and the comments from the Honda managers that the ERS strategy needs to be tuned to the car.

On wind direction I think this can be inferred from the air speed (pitot) and the road speed, (wheel speed or ground laser). You would probably want to use most deployment in the tailwind direction, so knowing this would allow the SOC to be modified on the fly to prioritise downwind straights, if the wind direction or strength changed.

On car parameters if you know grip level, a combination of tyre, aero and suspension, you can estimate how early on a straight you can start deployment and when you might expect to be able to charge. This will be dependant on aero configuration , tyres fitted, track evolution, rubbering in, and the aforementioned wind direction. Once again this would modify the SOC vs track position map, if there is more grip discharge earlier and de-rate (switch off MGU-K) earlier.

I would expect there to be monitoring of fuel use vs track position also. Again with a target map and a bunch of measured parameters. Outputs could be dash readings, lift and coast beeps, shift point settings and probably others.

Each strategy set on the steering wheel will probably have different maps, and they may or may not, cross populate.

Of course it may be nothing like this at all. And I still have no idea about the answer to the OP's original question, how does it know where it is on the track.

[MrPotatoHead]

A combination of Driver Mode Selection which then selects software / time / location / parameter driven harvesting.
But it is clear that the driver is often selecting the mode.

90% of the settings will be determined on the simulators before the team gets to the track.

[Craigy]

I see this as two questions really;

• How does the car charge (driver control or via mapping?), and
• How does a team determine the best place on the track to charge, or to deploy energy?

The former is fairly straightforward - there are predefined maps on the steering wheel for when to charge the car. The driver can select between these. At the extremes are "don't charge at all" (for the quali lap run) and ("charge everything to the max") which is most likely going to be used on the out lap before the quali run to the get the battery to a point where it has >4MJ of charge. (I say >4MJ because the exit from the last corner prior to the quali lap happening is not counted in the 4MJ, so actually the battery packs will have a higher capacity than 4).
There's also a separate button on the wheel for energy deployment (which is in addition to a mapped deployment which would occur regardless).
Combining the predefined maps for charging and deployment with the override for deployment puts the driver into fairly decent control of the system.

On the subject of "how do we map this in the first place" - there are lots of methods, but what's important is that there is always going to be a compromise between charging, deployment, car weight, racing line, tyre pressure, aero setup (lift/drag compromise), suspension setup, and racing position in terms of where the vehicle is faster compared to the competition on each portion of each lap. There are many other factors, but lets just look at these.

Why do I state all these things?

• Firstly - the cars are generally going to charge under braking. In this state, the car has to shed energy anyway - so it makes sense to do this here, and there are limited circumstances where one would want to not do this - related to better control of braking over bumps or heating the rear tyres, or something like that, where you only want the mechanical braking to work.
• The car will generally deploy energy once no longer traction limited out of corners. If there's an energy budget (the amount of energy collected on a lap) then there will be a decision on which corners to deploy on, and for how long.
• The car will generally deploy energy to spin up the ERS-H/compressor (this varies by manufacturer and year - Honda want to stay constant RPM, Ferrari used to do that but not in 2017, Mercedes have been and remain variable). In quali, it's likely that all the manufacturers do some level of ERS-H spinup this way.
• The car weight at the start of the race is a function of the fuel level. The cars are allowed up to 105kg, but generally every car will start below this level, for performance. The lightest cars will be kindest on their tyres and will be fastest at the start of the race (instead of weighing perhaps 35kg more than the competition and running a richer fuel mode for more power, they run lighter on fuel and consequently take less life from the tyre at the cost of less power from the leaner fuel mix).
• The racing line can be either longer or shorter, and tighter or more of a straight line. Each of these has a fuel economy consequence - the straighter line is obviously going to use less fuel because of a shorter distance, however a wider corner entry and exit may use less fuel than a tighter one due to lower tyre scrub, so there's a compromise to be made there too, and compromises on the drag/downforce and chassis settings can also affect this.
• Tyre pressures generally are as low as possible - however it's quite feasible that some cars run more than the minimum. Higher tyre pressures = lower rolling resistance, so they have an impact on fuel economy. Generally this will always come second to cornering and braking performance though.
• aero drag (=downforce) levels are going to vary and this is one of the biggest determinants of fuel economy. This is an aero and car concept philosophy thing - Mercedes run much higher downforce+drag than williams, for the same basic PU, for example. Williams are lighter, and have higher end-of-straight speeds, Mercedes are heavier (fuel load), and have more performance in corners and under braking. Mercedes have fewer problems keeping tyre temperature because of this. Williams makes for a better overtaking machine at end-of-straight because of their lower drag so they are choosing to make their car faster in certain places (straights) at the cost of cornering/braking performance.
• When in the race to use the performance. F1 races are obviously to the finish line (distance - 305km except Monaco, or 2 hour time limit, which is often only a factor in Singapore). In fact they often are a race to the track position at the last stop for each car (because overtakes on track are very difficult) - so the teams don't run the same on each lap. The first lap and each lap around the pitstops are the most important for track position so it's certainly the case that these will be run under less fuel economy pressure than the others. Equally, after the last stop, each of the cars will generally be in more of a fuel-saving mode than before that stop (because they want to get to that stop with the best track position, they will use more fuel on average before than after the last stops).
• Track elevation changes - >1MJ delta per lap at certain tracks (eg. Spa).

There are umpteen other parts to this, all of them have a knock-on consequence elsewhere.
For some people this would perhaps lead to an attitude of "well, there's no single answer to this" and a series of guesses/estimates would be made, but actually there is also a logic tool that can be applied in terms of simulation of all the factors.
It's a branch of maths called "optimal control" problems, and is essentially a statistical evaluation of multiple parameters (including all those above) needs to be more heavily weighted in order to achieve a certain goal (in this case, generally that's going to be elapsed time to get to the last pitstop while maintaining enough performance after that to not be overtaken).

There's a professor in Oxford who did this for Ferrari with the results starting in the 2014 year - You can find a link to a video of him talking about it here https://www.mathworks.com/videos/optima ... 96763.html

All of the top teams use this sort of technique to determine not only how to map the energy systems of the car to each track of the season, but also to determine where to get the most bang for their bucks in terms of developing the car through the season and what to do next season. It's how they decide how big of a cooling capacity to design into the car, how much drag they can live with, etc. It's not guesswork, it's based on cold logic. It's also how Merc end up with a car that can't run in traffic without overheating - it's optimised around being in cold air at the front of the pack.

I suspect this is a fuller answer than OP expected.