The Road to the 50% Thermally Efficient F1 Internal Combustion Engine

[Tommy Cookers]

A big part of modern engine design is about building in knock resistance by having very big focus on tumble strength and preservation beyond IVC.

the (hybrid F1) fuel quality seems conspicuously poor
ie MON in the lower 80s

as I forecast imo ...
they have cherry-picked from thousands of gasoline ingredients - the FIA rules can't prevent this
and sacrificed potential for unusually high ON to secure unusually high mass-specific heat value
(so this isn't entirely an efficiency formula)

remember ON is tested using (iso)octane as a reference fuel (some teachers imagine gasoline to be this)
F1 fuel is c.102 RON but only c.84 MON
iso-octane is 100 RON and 100 MON

isn't F1 injection pulsed according to real-time cylinder pressure rise ?
this must help with low ON fuels


yes the whole ICE thing is a surprise ....
such low heat to engine coolant without unprecedented leaning (though high heat to charge cooling)
and maybe in future a longer stroke/lower rpm ? - even 2 valves ???

[Hoffman900]

A 2 valve engine is always valve-area limited and that is even worst on a Miller Cycle engine. Furthermore, a 2 valve engine while potentially having a better chamber are not tumble designs (only a hemi is) and swirl isn’t preferred at all, tumble is the preferred motion in the cylinder.

But going back to my first point, you need all the valve-time-area you can get before IVC on the Miller Cycle and a 4 valve is always going to have that over a 2 valve.

F1, DTM, and most European series in general have used lower octane pump type fuels in racing for decades now. This is a big difference that must be considered when looking at amateur racing engines; like a 289ci Ford prepped for Goodwood / FIA series may have like 11-11.5:1 geometric compression where in the historic racing in States, with all the different fuels available $$, might be 13-14:1 geometric compression

[johnny comelately]

One of the sophisticated factors is the spark assisted HCCI, how is this fuel key for it being predictable and under control.
Has anyone calculated the cylinder pressures before ignition (careful not to use the term pre-ignition)

[Hoffman900]

One of the sophisticated factors is the spark assisted HCCI, how is this fuel key for it being predictable and under control.
Has anyone calculated the cylinder pressures before ignition (careful not to use the term pre-ignition)

Probably very! I believe Honda had to have their fuel reformulated to make the spark assisted HCCI stable and consistent. Otherwise it was an accidental anomaly (as has been written up elsewhere).

I’m sure someone has modeled it, looks like the FIA had to partner with a university to get some of that information because I am presuming the OEM’s won’t share it.

[johnny comelately]

One of the sophisticated factors is the spark assisted HCCI, how is this fuel key for it being predictable and under control.
Has anyone calculated the cylinder pressures before ignition (careful not to use the term pre-ignition)

Probably very! I believe Honda had to have their fuel reformulated to make the spark assisted HCCI stable and consistent. Otherwise it was an accidental anomaly (as has been written up elsewhere).

I’m sure someone has modeled it, looks like the FIA had to partner with a university to get some of that information because I am presuming the OEM’s won’t share it.

I meant: how (does this work)?

[Tommy Cookers]

F1, DTM, and most European series in general have used lower octane pump type fuels in racing for decades now.....

F1 has now and since 2014 unlimited ON
ok I wouldn't expect super hign ON fuel in a c. 20000rpm engine ie F1 pre 2014
for current F1 ie c. 10000rpm I did expect a MON in the high 90s not the low 80s
high ON is of course useless if stoichiometric-mass heat energy or mass-specific heat energy is uncompetitive

(I expect fancy US 'pump gas' race fuel includes quite a lot of xylene/metaxylene
at Goodwood you can get Avgas - ok Eu(ropean) Avgas isn't obliged to be the same formulation as US Avgas
- which maybe has xylene and maybe tests at c.106/130)

the presentation implies unease with the current 'cherry-picked' F1 fuel
ie an efficiency formula should involve carbon-efficiency count at source

[Hoffman900]

Everyone looks at octane, but I think the most telling as oxygen content as well as reid vapor pressure. Really, you want to run as little octane as you can get away with and it’s a solution to not having knock resistance built into the engine through good fast combustion.

But yeah, give me a fuel that has more oxygen and volatilizes better any day, but like most of us, we live in a world where we’re making up for lack of airflow with chemical energy. F1 engines are the opposite, they have an abundance of airflow and not enough fuel.

[johnny comelately]

This goes some way as an insight of combustion, an explanation as to the state of simulation development in 2016.

[johnny comelately]

The subject of combustion strategy as it applies to the current F1 engines is a complex matter, for me anyway.
The role of HCCI has been discussed elewhere but not so much in this context.
Previous to Pat Symonds expose, I viewed HCCI as a lost cause because of it originally being too hard to reliably control,ie, unpredictable.
With the introduction of spark assist some of the concerns have been alleviated.
Of course, these statements would look so amateurish to those in the know, but I shall battle on in the hope that those more knowledgable can contribute to the learning.
I get the impression that when the flame ignition is used as the initiator, because of the enormous pressures involved a very quick, high pressure rise occurs. As evidenced by the 4 elephant statement by Andy Cowell I think. and the roughly 500 BHP per litre.
On this point has anyone worked out the BSFC, maybe around 130 g/kwH ??
Pat mentioned the 50% MFB by 8 degrees ATDC, most engines reach peak pressure aroung 20 to 27 degrees ATDC, he doesnt mention the peak angle (damn him )
But it is some indication of the speed of burn.
Which brings me to the point of deflagration or controlled detonation.
Is HCCI controlled detonation, with the pressures causing combustion chamber wide?
Is there a controlled detonation in the prechamber facilitated by a Shchelkin Spiral for instance?
This would produce an enormous jet flame, but unlikely because of the rich stratification in that zone.
It seems that no one is sure if there is an affective, significant transition phase between deflagration and detonation , but if there is it makes some of this possible.

[gruntguru]

One of the sophisticated factors is the spark assisted HCCI, how is this fuel key for it being predictable and under control.
Has anyone calculated the cylinder pressures before ignition (careful not to use the term pre-ignition)

Charge temperature before ignition is the key determinant for pre-ignition and detonation.

For 18:1 compression, charge temperature at TDC is approximately 3 times the temperature at BDC so if we assume 50*C after intercooling (323*K) the TDC temperature is 969 K or 696 C. It is likely that the BDC temperature is somewhat cooler due to expansion after IVC (Miller cycle with EIVC). We can estimate this effect from the mass airflow suggested in the lecture (Lambda 1.4 = 34.3 kg/min). This airflow suggests a VE at 10,500 rpm of 3.4 (340%) whereas 5.5 bar MAP plus wave tuning would normally produce a VE around 600% with perfect intercooling, or 540% with intercooling to 30* above ambient.

So assuming cylinder pressure and temperature at IVC of 5.5 bar and 323*K at IVC, pressure of 3.4 bar at BDC, and no heat transfer between those points, there will be a temperature decrease of about 40*. This results in a decrease of 120* in the TDC temperature for 849*K (576*C). Working backwards gives an effective in-cylinder CR of 12.7:1

To answer your original question - pressure at TDC. Assuming 3.4 bar at BDC, 18:1 CR gives a TDC pressure of 183 bar.

If the entire heat of combustion is added at TDC (8.7 kJ/cycle added to 0.00408 kg air/cycle) the temperature will increase to 2,625 K and the pressure to 566 Bar.

Happy for anyone to check my calculations (I haven't) but the answers seem to be in the ball park.

EDIT 10/05/2022. This video from MB HPP https://youtu.be/Cx4HMSNx_MY says peak combustion temperature is 2760*C ( 3033*K) so BDC charge temperature must be considerably higher than I calculated.

[gruntguru]

The subject of combustion strategy as it applies to the current F1 engines is a complex matter, for me anyway.
The role of HCCI has been discussed elewhere but not so much in this context.
Previous to Pat Symonds expose, I viewed HCCI as a lost cause because of it originally being too hard to reliably control,ie, unpredictable.
With the introduction of spark assist some of the concerns have been alleviated.
Of course, these statements would look so amateurish to those in the know, but I shall battle on in the hope that those more knowledgable can contribute to the learning.
I get the impression that when the flame ignition is used as the initiator, because of the enormous pressures involved a very quick, high pressure rise occurs. As evidenced by the 4 elephant statement by Andy Cowell I think. and the roughly 500 BHP per litre.
On this point has anyone worked out the BSFC, maybe around 130 g/kwH ??
Pat mentioned the 50% MFB by 8 degrees ATDC, most engines reach peak pressure aroung 20 to 27 degrees ATDC, he doesnt mention the peak angle (damn him )
But it is some indication of the speed of burn.
Which brings me to the point of deflagration or controlled detonation.
Is HCCI controlled detonation, with the pressures causing combustion chamber wide?
Is there a controlled detonation in the prechamber facilitated by a Shchelkin Spiral for instance?
This would produce an enormous jet flame, but unlikely because of the rich stratification in that zone.
It seems that no one is sure if there is an affective, significant transition phase between deflagration and detonation , but if there is it makes some of this possible.

A few observations - some of them my own interpretations so not necessarily facts.
1. Four elephants on the piston crown is equivalent to 400 bar pressure. (5T elephant)

2. Normal combustion progresses as a "flame front" - a spherical surface moving outwards from the spark. Inside the sphere is burnt gas. Outside the sphere is unburnt charge. At the "front" itself unburnt gas is heated to its ignition point by conduction and radiation from adjacent burning molecules. Forced convection (e.g. turbulence) can speed up the progress of the flame front. Because the flame front travels much slower than the speed of sound (Mach 1), the pressure in the entire chamber increases at about the same rate.

3. TJI is flame-front ignition but is initiated almost simultaneously throughout the chamber by the high speed jets of combustion products coming from the pre-chamber. Combustion in the pre-chamber is also normal flame-front combustion.

4. Detonation ignition starts from some initial point and progresses at Mach 1 (much faster than the adiabatic flame speed of a normal flame front). This occurs because the remaining charge has been heated (by radiation from the flame) almost to the auto-ignition temperature and is so sensitive that it will ignite with a small pressure increase. The pressure increase at the combustion front can only spread to the remaining unburnt charge at the same rate (Mach 1) as the detonation front - so the pressure at the detonation front builds very rapidly to a level much higher than the rest of the chamber. This is a shock wave and is powerful enough to damage metal parts.

5. HCCI is simultaneous auto-ignition at a very large number of sites spread evenly throughout the chamber. Each one of these sites becomes a tiny flame kernel that spreads out until it bumps into the next kernel and runs out of air-fuel mix. Each kernel spreads via a flame front or even a detonation front. Because the kernel size and travel length is so small, the shock wave(s) produced are small and cancel each other to some extent anyway. High load HCCI has been around for many years in "diesel" (not really diesel) model engines - controlled using adjustable AFR and CR (and fuel with a low auto ignition temperature). Similarly HCCI is being achieved in F1 but only at a narrow range of operating points.

6. The fuel flow limit is 100 kg/hr and LHV is 44 MJ/kg so heat available is 4400 MJ/hr = 1,222 kW of heat. Symonds says ( ) 52% TE so engine output is 636 kW. BSFC is 100,000/636 = 157.2 g/kW.hr (Someone needs to add the F1 engines to the table at the bottom of the Wikipedia BSFC page https://en.wikipedia.org/wiki/Brake-spe ... onsumption)

[johnny comelately]

Thank you for this instructive information.
It highlights the part temperatures plays, something I havent considered enough apart from rudimentary factors.
These engines are more akin to the racing diesels of Audi etc than traditional petrol engines.

[vorticism]

Regarding TJI in this application, would it be correct to say:

(1) the prechamber a:f ratio is equal to or leaner than the main charge
(2) a greater % of combustion products remain in the prechamber after the exhaust stroke compared to the cylinder

[johnny comelately]

Regarding TJI in this application, would it be correct to say:

(1) the prechamber a:f ratio is equal to or leaner than the main charge
(2) a greater % of combustion products remain in the prechamber after the exhaust stroke compared to the cylinder

1. Richer
2. moot point, either way?

[vorticism]

1. Richer

What mechanism do you propose to ensure this? GIven the regulatory set.