F1 cylinder head temperatures

[johnny comelately]

anyone hazard a guess at current F1 cylinder head temperatures?
as it relates to new efficiencies : quench and boundary layer thickness and materials and coatings.
different to cylinder temperatures? two different systems?

[63l8qrrfy6]

Probably peaks at about 350 C in the exhaust-exhaust bridge. Purely based on the fact that no Al alloy could possibly cope with anything more.

[hurril]

Probably peaks at about 350 C in the exhaust-exhaust bridge. Purely based on the fact that no Al alloy could possibly cope with anything more.

Are they not permitted more exotic surface materials?

[63l8qrrfy6]

Probably peaks at about 350 C in the exhaust-exhaust bridge. Purely based on the fact that no Al alloy could possibly cope with anything more.

Are they not permitted more exotic surface materials?

There is not much you can do to Al as it is a weak substrate for most coatings. Alfin bonds with high temperature irons are also poor and have lowish HTC - you would run much hotter with not a lot of extra strength (if any at all).

[hurril]

Probably peaks at about 350 C in the exhaust-exhaust bridge. Purely based on the fact that no Al alloy could possibly cope with anything more.

Are they not permitted more exotic surface materials?

There is not much you can do to Al as it is a weak substrate for most coatings. Alfin bonds with high temperature irons are also poor and have lowish HTC - you would run much hotter with not a lot of extra strength (if any at all).

Ahh I see. Thank you

[johnny comelately]

The thinking has been along the lines of:
to achieve higher efficiency they would have to have a higher fraction burn which means reducing the boundary layer thickness, which comes from quench.
And in order to do that they would have to run higher metal (and coolant) temperatures combined with coatings.
All the while avoiding detonation (hence the oil burn imo)

[gruntguru]

The simplest, most effective way to reduce surface temperature is to provide a "dry" (air only) boundary layer - should be do-able with DI.

[riff_raff]

Probably peaks at about 350 C in the exhaust-exhaust bridge. Purely based on the fact that no Al alloy could possibly cope with anything more.

Correct. Aluminum cylinder head material is limited by fatigue strength at temperature. The cylinder head surface at the exhaust bridge location will have fairly high temperatures due to limited heat transfer to the coolant flow. The combination of high temp and stress at this location will be a factor with fatigue life of the cylinder head structure.

[johnny comelately]

Ok, so any ideas of the quench temperature of stoichiometric petrol mixture at any given pressure?

[johnny comelately]

Anyway, the point being is that the subject is higher burn fraction as a factor in the higher efficiency F1 engines, maybe even 98% - 99%.
In order to do that the boundary layer has to be minuscule and that is possible with coatings.
Danger from detonation/pre-ignition is reduced dramatically with flame ignition producing very quick burn time and a more complete burn.

[gruntguru]

Anyway, the point being is that the subject is higher burn fraction as a factor in the higher efficiency F1 engines, maybe even 98% - 99%.
In order to do that the boundary layer has to be minuscule and that is possible with coatings.

Agree on the burn fraction - but that is burn-fraction of fuel - not air. There is a lot of excess air in current F1 engines - approx 100% so the boundary layer does not have to be minuscule provided there is no fuel in it.

[johnny comelately]

Anyway, the point being is that the subject is higher burn fraction as a factor in the higher efficiency F1 engines, maybe even 98% - 99%.
In order to do that the boundary layer has to be minuscule and that is possible with coatings.

Agree on the burn fraction - but that is burn-fraction of fuel - not air. There is a lot of excess air in current F1 engines - approx 100% so the boundary layer does not have to be minuscule provided there is no fuel in it.

I'm not quite with you there.
If 'burn fraction' is the wrong terminology i apologise for that.
My understanding it is the AF mix.
My premise is the amount of burn of the AF mix as it can contribute to the push.
This being done by a relatively complete burn contained in an environment that can handle it: coated chamber and piston (or carbon piston), flame ignited.
Essentially to get rid of quench induced boundary layer.
This is the fuels maximum contribution to the 50% thermal efficiency (which is such a complex total measurement)
One of the benefits is the extra heat available to drive the turbo and ancillary.

[Muniix]

The concept of TBC coated steel combustion surfaces with cooling system controlled by a 'wall heat transfer' simplified physics model running in a high performance auxillary processor running a real time physics engine. The temperature sensors measuring temperature in and out of the head and liner used to verify the removal of the correct amount of heat.

This way you get to maintain a more constant temperature to support more complete combustion with less cycle to cycle variance. Controlling a 48 volt electric coolant pump and fans.

A cylinder pressure sensor is needed for the 'wall heat transfer' function, having cylinder pressure provides really useful control and diagnosis of the combustion process. Initial engine calibration using a Kistler sensor is most likely needed, for production a virtual sensor maybe all that is needed but most likely a Sensata CPoS sensor providing data at 15,000 samples a second. One the wall heat transfer model has been calibrated with the Kistler sensor, using the AI with the CPoS sensor should work fine. Having been trained on the high resolution sensor data and what the CPoS return, it is a lot like facial recognition algorithms. Recognising combustion phasing indicators like;
SoC Start of Combustion
LPP Location of Peak Pressure
MHRR Maximum heat release rate
CA10 crank angle of 10% fuel burn
CA50
CA90
This way you can optimise the burn rate over the crank angle, minimising negative work before TDC and the pressure over the crank/rod angles to achieve maximum rotational torque or less losses in the cranktrain depending on engine mode selected. Running a powertrain model calculating the frictional losses in real time, like the Guzzomi model. Even a Kirin 970 SoC would have plenty of performance, thou a NXP a55/Mali G72 SoC would likely be a more traditional choice, if you really need lots of compute the AMD Ryzen mobile SoC with the latest Vega GPU core it has over a teraflop of compute for only 25 watts.

As engine control complexity increases it will need to go through some Deep Learning training the usual Artificial Intelligence. Languages such as Julia running on the more recent Arm cores that have GPU compute units, HSA architecture and Vulkan Compute API.

Julia was used to process the Astronomical survey data, 56 Petabytes to locate the binary neutron stars recently. They used 1.25 million threads on 650,000 cpu cores, that required one of the top 5 fastest computers on the planet. I was at CSIRO Lindfield just a few days ago, where they were involved with locating the source of the gravitational waves.