Mercedes's Andy Cowell: We exploit HCCI (at least SPCCI)

Mercedes's Andy Cowell is quoted in this article in Motorsport Magazine saying:

“We use a conventional spark to start our combustion. Does the pressure increase that then occurs as a consequence of combustion cause spontaneous combustion, as opposed to flame front combustion? The answer is ‘yes’. Do we welcome that? Sometimes it’s welcome, sometimes unwelcome combustion? I think that is the challenge of taking combustion to another level.”

 

 

The full context is here: 

 

 

So what makes the power units so complex? There are too many individual factors to list them all, but a few do stand out. The first is the efforts taken to refine the combustion process; the key to making power is efficiency with these engines and that means running exceptionally lean. The leaner the engine runs, the greater the issues with knock (where the fuel/air mix starts to combust ahead of the flame front, leading to unstable combustion and expensive noises). This means precise control is needed over combustion conditions which is where one of the more public innovations comes to light, pre-chamber ignition.

 

First revealed as being part of Ferrari’s package, this relies on a small chamber within the main combustion chamber which is filled with rich mixture and then ignited. The burning mixture is expelled from the chamber via nozzles to the main combustion chamber, igniting the much leaner mixture within. The result is an (almost) simultaneous ignition of all of the mixture, negating the problems of pre-ignition found with a standard spark and allowing the engine to run very lean.

 

Making these systems work is complicated, particularly within the constraints of the rules. There are also times when some pre-ignition is desired, effectively controlled knock, as Andy Cowell once alluded to. “We use a conventional spark to start our combustion. Does the pressure increase that then occurs as a consequence of combustion cause spontaneous combustion, as opposed to flame front combustion? The answer is ‘yes’. Do we welcome that? Sometimes it’s welcome, sometimes unwelcome combustion? I think that is the challenge of taking combustion to another level.”

Source: https://www.motorspo...y-of-f1s-hybrid

 

 

SPCCI (Spark Plug-Controlled Compression Ignition) is essentially what Mazda is doing with its Skyactiv'X.

https://www2.mazda.c...ion/technology/

 

Discuss!

Weren't Lotus pushing this back in the 90s?

Mercedes's Andy Cowell is quoted in this article in Motorsport Magazine saying:

 

The full context is here: 

 

Source: https://www.motorspo...y-of-f1s-hybrid

 

 

SPCCI (Spark Plug-Controlled Compression Ignition) is essentially what Mazda is doing with its Skyactiv'X.

https://www2.mazda.c...ion/technology/

 

Discuss!

read a university paper on hcci combustion...one of them was proposing an idea to condition the cylinder with fuel at 80% hcci potential so it wont combust on its own then upon ignition it increase cylidner pressure and temperature then hcci would occur. i think its more like mazda's but this paper is ages ago...ill post it here if i manage to find it again

Needs to be temperature not pressure - and probably radiant heating not compressive heating. For compressive heating the combustion will propagate in a wave at sonic velocity. This will generate a shock wave akin to detonation. True HCCI occurs simultaneously at many points throughout the chamber.

As much and for as long as this has been discussed in the context of F1, I'd be amazed if it weren't universal now on the grid. Ferrari. for one, have been quite explicit about using it.

Dieseling a Petrol engine is hard, especially when you only get max fuel flow at 10k and up so i hope they get their heads out of their asses and implement a flat fuel flow at some point. Then again if you leave bore and stroke as is they might avoid it initially for reliabilitys sake.

the irony about using hcci in an f1 engine is that hcci charge window is very limited before it burns on its own. high pressure for targeted performance results such as high boost or those high compression ratios and high compression pumping speeds with high rp/m they are running can easily trigger pre-mature hcci charge. on a normal production test engine, controlling volatility was one of their problems as engine operating levels rises, even a naturally aspirated test engines for production cars cannot operate on hcci when more engine load and rp/m is applied

You can actually see evidence of the HCCI burn on the crown of a modern F1 piston- https://twitter.com/...397030445129729

What you can see is witness marks from the TJI Jets.

Are these mutually exclusive?

Off-centre to the bore?

Are these mutually exclusive?

TJI is an accelerated ignition technique. The combustion regime that follows is a conventional flame front (or fronts).

HCCI is a different combustion regime where unburned charge experiences widespread spontaneous ignition at multiple locations without contact with either a flame front or a pressure wave. (pressure wave ignition is typically associated with detonation).

The video from Honda shows true HCCI occurring independent of the TJI ignited combustion.

 

https://streamable.com/e/kc9qv0

I titled the tread to call the process HCCI based on Andy Cowell's quoted statement in the first post. Based on that statement, HCCI would be accurate, but it is more likely precise that the process to which Cowell alludes is more like SPCCI (Spark Plug-Controlled Compression Ignition), which is what Mazda calls it for its Skyactiv-X, or SACI (Spark Assisted Compression Ignition). These processes can operate over a much wider load/speed range are much less sensitive to reaction kinetics and prevailing conditions than pure HCCI.

the irony about using hcci in an f1 engine is that hcci charge window is very limited before it burns on its own. high pressure for targeted performance results such as high boost or those high compression ratios and high compression pumping speeds with high rp/m they are running can easily trigger pre-mature hcci charge. . . .

A significant difference here is the presence of pressure sensing on all cylinders of an F1 engine. Analysing pressure rise for every firing allows the system to determine the combustion regime for every firing and adjust input parameters to maintain HCCI without excessive detonation. Which parameters are controlled is debateable.

 

Essentially it is charge temperature at TDC that will determine whether HCCI will occur. I don't believe the spark is initiating HCCI so changes to TJI, like spark timing and pre-chamber AFR will not be sufficient to create HCCI. Perhaps main chamber AFR or perhaps even end-gas AFR (controllable with injection event timing???) could be used to adjust charge temp at TDC.

 

Obviously MAP and MAT will have a major effect but the time constant for adjusting these is very long and I don't think they can be controlled on a per-cylinder basis.

I titled the tread to call the process HCCI based on Andy Cowell's quoted statement in the first post. Based on that statement, HCCI would be accurate, but it is more likely precise that the process to which Cowell alludes is more like SPCCI (Spark Plug-Controlled Compression Ignition), which is what Mazda calls it for its Skyactiv-X, or SACI (Spark Assisted Compression Ignition). These processes can operate over a much wider load/speed range are much less sensitive to reaction kinetics and prevailing conditions than pure HCCI.

If you look at the Honda video, the HCCI events at the perimeter of the chamber begin before the jets appear at the pre-chamber orifices. I can't see how the spark can be initiating the HCCI - unless a pressure wave from the pre-chamber is able to reach the main chamber perimeter before the jets even emerge - which I doubt very much.

 

(Look at 0:22 in the video)

Edited by gruntguru, 16 May 2022 - 23:33.

There is zero context in those videos, except I see Honda and F1 mentioned. Pure implementations of HCCI doesn't even need direct injection of any sort and most are in fact port injected - again, more context needed.

HCCI doesn't need a spark at all to get ignition started, at least the video seems to show that happening. That combustion appears to occur first at the periphery rather than at the centre where the spark plug or prechamber might be is also fully explained by the reaction kinetics. Ignition occurs first where the conditions are most amenable to autoignition, namely having the required conditions of temperature, pressure, air-fuel ratio, and presence of reactive species. The outer edges of the combustion chamber is usually the coolest, so that tends to work against autoignition, but there can be locally richer zones due to piston ring crevice volumes, and the presence of engine oil on the cylinder bore and ring pack also has been shown to have several points lower effective octane rating than the fuel itself and therefore be a site of preferential autoignition.

It bears noting that once you have any significant stratification by design or otherwise, you no longer have HCCI by definition ("H" denoting "homogeneous"), but rather gives rise to another four-letter acronym - SCCI (Stratified Charge Compression Ignition). Although there may be a small degree of local enrichening stratification due to the crevice volume, I don't believe we have SCCI formally happening in the video. Many of the in-cylinder imaging I have seen of canonical HCCI also has ignition starting at the periphery precisely due to the aforementioned crevice volume enrichment and presence of more autoignition-prone engine lube oilm

There are a few fascinating papers that show that engine oil can have a big impact on knocking tendency, and that has a contribution to HCCI, too. IIRC, ester-based engine oils were least negatively impactful as far as knocking is concerned.

Edited by TDIMeister, 17 May 2022 - 00:10.

Also note how with the conventional engine, after combustion is done in the main chamber, what is very likely unburned fuel issues from the pre-chamber orifii.  This is very like a diesel nozzle where residual fuel in the sac is boiled and/or blown out after the combustion event resulting in HC exhaust emissions.

IIRC, most modern Diesels use VCO (valve covers orifice) injectors, precisely to address HC emissions as you stated. With holes under 150 microns in diameter, there is a vanishingly small volume in the holes themselves to contribute to HC/soot issues, but the sac volume is no longer in significant play as far as emissions are concerned.

Your point, however, of having some trapped unburned mixture in the prechamber potentially leading to higher HC emissions certainly has merit, which is also why thoroughly scavenging the prechamber is an area of active R&D with CFD, and why it is preferred to keep prechamber volumes very small compared to the total swept and clearance volumes (~3-5%). The small volume also reduces heat loss; the higher heat loss is why prechamber Diesels suffered a ~15-20% BSFC penalty compared to direct injected ones.

Edited by TDIMeister, 17 May 2022 - 00:24.

A possible explanation for HCCI (or SCCI) at the perimeter is higher charge temperature. Although the cylinder walls are cooler than the compressed charge, the fresh, uncompressed charge is much cooler than all the chamber surfaces, especially piston crown and exhaust valves. The uncompressed charge is tumbling and absorbing heat at these surfaces. Rapid compression then results in a higher temp towards the perimeter (although cooler at the extreme boundary due to quenching).

 

F1 engines are all (anecdotally) running HCCI - relying on lube oil contamination would surely be excessively unreliable.

 

Crevices would be more severely quenched - charge would be cooler there.

As it happens, the trend with modern light duty diesels went from VCO > mini-sac > micro-sac because sac-type nozzle give better spray atomization and thus lower Pm emissions hence fewer trap regeneration events.  Depending on required fuel flow, sac volumes can fall between 0.06 mm^3 and 0.2 mm^3.  If that volume gets boiled out after every injection, particularly given the wide use of multiple injections, the HC emissions can add up.  Fortunately, oxidation catalysts are quite effective.  VCO nozzles may make a come-back if we get the chance to use low carbon smokeless fuels like DME in the future.

"Omega bowl" F1 piston from "c.2019". https://twitter.com/...836447845339137 Doesn't look remotely SI does it? 

"Omega bowl" F1 piston from "c.2019". https://twitter.com/...836447845339137 Doesn't look remotely SI does it? 

Very cool find!