Ferrari Power Unit Hardware & Software

[henry]

During coasting (i.e release the gas pedal before braking), it is perfectly feasible to harvest from the K.

SSP 682.5 (Self Sustain Plus, ICE + MGU-H + ES)
SS 625 (Self Sustain, ICE + MGU-H)

When you say MGU-H here by what is it driven in each case?

I was using a short-hand. The electrical power to the crank is always provided by the MGU-K. That power either comes directly from the MGU-H or indirectly from the ES.

When the MGU-H is providing power to the MGU-K the MGU-H is driven by the exhaust turbine, which also drives the compressor.

[chaoticflounder]

Control theory studies suggest that these modes are used “bang-bang”, that is they don’t use partial power versions of them.

* I have omitted Honda’s extra harvest mode. And there is another theoretical “extra power” for which I have seen no evidence of implementation.

Would you be able to link any of these control theory studies?

Are these the ones that have been linked previously in this thread?

Thanks,

C

[MtthsMlw]

SSP 682.5 (Self Sustain Plus, ICE + MGU-H + ES)
SS 625 (Self Sustain, ICE + MGU-H)

When you say MGU-H here by what is it driven in each case?

I was using a short-hand. The electrical power to the crank is always provided by the MGU-K. That power either comes directly from the MGU-H or indirectly from the ES.

When the MGU-H is providing power to the MGU-K the MGU-H is driven by the exhaust turbine, which also drives the compressor.

Thanks.
So the H is driving the Compressor with energy from the ES in e-boost mode? I guess it is also sometimes used partially with the wastegates not fully opened to reduce back pressure a bit.
Not taking the K - > H route in to consideration here.

[restless]

Comparison between Canada and Sochi
Gaining only at high speeds:
https://www.f1analisitecnica.com/wp-con ... 24x358.jpg

Gaining all the time..
You can see the end of straight drop off quite nicely here. On the pit straight and back straight (right before S2)
https://www.f1analisitecnica.com/wp-con ... 24x466.jpg
via f1analisitecnica.com

What I see is how Ferrari accelerates faster.
That means more power to the wheels.
They gain through whole straights. Harvesting gains is fine, but doesn't explain these 20-30-40hp more they have - except better ICE combustion

[henry]

Control theory studies suggest that these modes are used “bang-bang”, that is they don’t use partial power versions of them.

* I have omitted Honda’s extra harvest mode. And there is another theoretical “extra power” for which I have seen no evidence of implementation.

Would you be able to link any of these control theory studies?

Are these the ones that have been linked previously in this thread?

Thanks,

C

The first reference I have used is Prof. Limebeer’s early paper from 2013. http://vdol.mae.ufl.edu/JournalPublicat ... covery.pdf There’s a reference to bang-bang on page 20.

The second is as posted earlier in the thread Salazar’s paper of 2018. https://www.research-collection.ethz.ch ... sAllowed=y

Interestingly both studies involved Ferrari.

[subcritical71]

Control theory studies suggest that these modes are used “bang-bang”, that is they don’t use partial power versions of them.

* I have omitted Honda’s extra harvest mode. And there is another theoretical “extra power” for which I have seen no evidence of implementation.

Would you be able to link any of these control theory studies?

Are these the ones that have been linked previously in this thread?

Thanks,

C

The first reference I have used is Prof. Limebeer’s early paper from 2013. http://vdol.mae.ufl.edu/JournalPublicat ... covery.pdf There’s a reference to bang-bang on page 20.

The second is as posted earlier in the thread Salazar’s paper of 2018. https://www.research-collection.ethz.ch ... sAllowed=y

Interestingly both studies involved Ferrari.

I know a sliding concept, vs bang-bang, would add another dimension (and processing power) to the equation, but are they really that far off from being able to implement?

[henry]

Would you be able to link any of these control theory studies?

Are these the ones that have been linked previously in this thread?

Thanks,

C

The first reference I have used is Prof. Limebeer’s early paper from 2013. http://vdol.mae.ufl.edu/JournalPublicat ... covery.pdf There’s a reference to bang-bang on page 20.

The second is as posted earlier in the thread Salazar’s paper of 2018. https://www.research-collection.ethz.ch ... sAllowed=y

Interestingly both studies involved Ferrari.

I know a sliding concept, vs bang-bang, would add another dimension (and processing power) to the equation, but are they really that far off from being able to implement?

Can I ask what you mean by a sliding concept?

I think the logic is that since you want to apply the maximum power at the lowest speeds and to achieve that you apply the highest discrete power you have until you can’t and then switch to the next. I’m guessing there will be ramp functions for drivability and tyre preservation but probably quite short.

[63l8qrrfy6]

Interestingly both studies involved Ferrari.

Interestingly Giacomo has been employed by Mercedes for quite a few years now )

[subcritical71]

The first reference I have used is Prof. Limebeer’s early paper from 2013. http://vdol.mae.ufl.edu/JournalPublicat ... covery.pdf There’s a reference to bang-bang on page 20.

The second is as posted earlier in the thread Salazar’s paper of 2018. https://www.research-collection.ethz.ch ... sAllowed=y

Interestingly both studies involved Ferrari.

I know a sliding concept, vs bang-bang, would add another dimension (and processing power) to the equation, but are they really that far off from being able to implement?

Can I ask what you mean by a sliding concept?

I think the logic is that since you want to apply the maximum power at the lowest speeds and to achieve that you apply the highest discrete power you have until you can’t and then switch to the next. I’m guessing there will be ramp functions for drivability and tyre preservation but probably quite short.

I have the idea in my head but can't seem to articulate it well so I'll just give you what I've seen instead. If you have a maximum power mode and maximum efficiency mode, depending on conditions you can slide between the two instead of a binary on-off, one or the other (what I interpret the bang-bang as). So, if your at a max power mode and max power is no longer necessary the ECU would slide to the efficiency strategy mode with no driver interaction. And likewise when power demand is max again. It would essentially take out some of the required strat mode changes we see on the cars today. Another way to think about it, if an ES for example is reaching a critical temperature while in maximum power mode you can slide to the next 'lower' mode automatically with no driver interaction necessary to maintain as much power as possible. Think of the sliding as a linear progression between any two modes vs an on-off scenario.

[PlatinumZealot]

You guys are overthinking!

It CANNOT be an E-boost strategy that give Ferrari the advantage. This is simply too easy for Mercedes to sniff out and copy - even if to simulate it in one sector.

What is something that Mercedes cannot copy easily or simulate within the rules?

The power from the ICE.

Ferrari simply have a more powerful ICE. Yeah you know its there - but can you copy it within the bounds of the regulated fuel limit? Not easy to do. Even with E-boost.

As we had discussed before, a more powerful ICE is benefit upon benefit. More turbine power, more MGUH power, less compressor boost needed, Less MGUK needed. The ICE is the core of any big power boost. Be it fuel, lubricant or new pistons, new combustion method etc... The ICE hardware is still the great un-locker.

[henry]

I know a sliding concept, vs bang-bang, would add another dimension (and processing power) to the equation, but are they really that far off from being able to implement?

Can I ask what you mean by a sliding concept?

I think the logic is that since you want to apply the maximum power at the lowest speeds and to achieve that you apply the highest discrete power you have until you can’t and then switch to the next. I’m guessing there will be ramp functions for drivability and tyre preservation but probably quite short.

I have the idea in my head but can't seem to articulate it well so I'll just give you what I've seen instead. If you have a maximum power mode and maximum efficiency mode, depending on conditions you can slide between the two instead of a binary on-off, one or the other (what I interpret the bang-bang as). So, if your at a max power mode and max power is no longer necessary the ECU would slide to the efficiency strategy mode with no driver interaction. And likewise when power demand is max again. It would essentially take out some of the required strat mode changes we see on the cars today. Another way to think about it, if an ES for example is reaching a critical temperature while in maximum power mode you can slide to the next 'lower' mode automatically with no driver interaction necessary to maintain as much power as possible. Think of the sliding as a linear progression between any two modes vs an on-off scenario.

I think I see what you mean. I think the concept of Max power or max efficiency maps more on to Strat Modes. In my thinking a Strat Mode is made up of Power Modes. It is the switching between these modes that is Bang-bang. The Strat Modes are adaptive, they change in response to changing conditions. How they would respond to, say ES temperature I don’t know. I’m pretty sure that they respond to, say, wind direction, by applying more power downwind than upwind. I think they do this by changing the the mix of durations of the individual discrete Power Modes. The PhD paper you posted is about how you would design a real time optimiser to handle this. Or at least that’s how I read things. It considers only some of the potential Power Modes, I think there are about 10.

Like you I struggle to articulate the operation of these PUs and I still don’t have a good mental model of the finer gradations of Strat Modes.

[henry]

You guys are overthinking!

It CANNOT be an E-boost strategy that give Ferrari the advantage. This is simply too easy for Mercedes to sniff out and copy - even if to simulate it in one sector.

What is something that Mercedes cannot copy easily or simulate within the rules?

The power from the ICE.

Ferrari simply have a more powerful ICE. Yeah you know its there - but can you copy it within the bounds of the regulated fuel limit? Not easy to do. Even with E-boost.

As we had discussed before, a more powerful ICE is benefit upon benefit. More turbine power, more MGUH power, less compressor boost needed, Less MGUK needed. The ICE is the core of any big power boost. Be it fuel, lubricant or new pistons, new combustion method etc... The ICE hardware is still the great un-locker.

It may be as simple as better combustion. However, I believe Ferrari deploy a different charging strategy from Mercedes and that places higher demands on the ES and CE. it’s a difference and maybe Mercedes don’t know how they do that. I very much doubt that it is just one thing that makes the overall difference.

I don’t understand what you mean by “Less MGU-K needed”, or why they would need less compressor boost. Care to explain?

[subcritical71]

You guys are overthinking!

It CANNOT be an E-boost strategy that give Ferrari the advantage. This is simply too easy for Mercedes to sniff out and copy - even if to simulate it in one sector.

What is something that Mercedes cannot copy easily or simulate within the rules?

The power from the ICE.

Ferrari simply have a more powerful ICE. Yeah you know its there - but can you copy it within the bounds of the regulated fuel limit? Not easy to do. Even with E-boost.

As we had discussed before, a more powerful ICE is benefit upon benefit. More turbine power, more MGUH power, less compressor boost needed, Less MGUK needed. The ICE is the core of any big power boost. Be it fuel, lubricant or new pistons, new combustion method etc... The ICE hardware is still the great un-locker.

Agree 100%. My take is more what can be possible. Probably a bit off topic.

[chaoticflounder]

Do we know if any teams use the turbine as an engine brake during deceleration? (i.e. MGU-H to ES)

Opposite essentially of the vaccuum that would get pulled in a typical otto cycle engine when you take your foot off the gas and the throttle plate closes, bringing the vehicle into noticeable deceleration.

I understand they can use the MGU-K to decelerate at a rate of 120 kW. But has anyone taken into account also using the MGU-H to capture this energy through compressing combustion chamber products on the exhaust stroke of the engine (compressing it against the turbine)?

Looking at some rough numbers for it, it looks like it's possible.

https://www.grc.nasa.gov/WWW/K-12/airpl ... trbth.html
https://www.mercedes-amg-hpp.com/formul ... ne-facts/#

Using this formula for Turbine Work / mass:

W/m = Cp*(Tin - Tout)
Power = (mass flow rate)*(W/m) = (mass flow rate)*Cp*(Tin - Tout)

Cp = 1.150 [kJ/(kg*K)] approximate value for dry air at around 500 C, roughly equal to Cp of combustion products
Tin = 1000 C = approx 1270 K
Tout = 500 C = approx 770 K
mass flow = 100 [kg/h] * 17 (Lean Air/Fuel Ratio + 1) * 1/3600 [h/s] = 0.472 [kg/s]

Power = 0.472 [kg/s] * 1.150 [kJ/(kg*K)] * (1270 [K] - 770 [K])
Power = 271.5 [kW]

Now, a couple things:

Obviously all of this power can't be devoted to the MGU-H, they likely won't run full fuel rate on deceleration, and there will be a stall limit to the turbine (variable geometry guide vanes would help but I'm not sure if it's legal).

But there looks to be some potential there, but it seems too easy for the rest of the teams to miss something like this.

C

[Phil]

I always thought the theories surrounding whatever smart thing Ferrari is doing is related to energy as opposed to it being ICU, is because there is a clear pattern when it is being used and when it isnt.

E.g. during a hot lap in qualifying. At race starts (first lap). Restarts. It points to it being confined for one lap only. Simply a better ICU would IMO not show this pattern.