Jump to content


Photo
- - - - -

Mercedes-AMG achieves 50% thermal efficiency on 2016 F1 engine


  • Please log in to reply
111 replies to this topic

#51 Greg Locock

Greg Locock
  • Member

  • 5,656 posts
  • Joined: March 03

Posted 06 March 2016 - 08:28

Yes, but is your efficient 24cc petrol engine an exact scale model of the 24 litre one? No, it is not.

 

The first really clever thing the Victorian engineers did was to measure the indicator diagram of their engines. Have the model engineering boys done that? Have they measured their boiler efficiency? 



Advertisement

#52 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 06 March 2016 - 09:20

SI engines have an optimal cylinder size. TE falls either side. Mostly to do with friction as they get smaller and detonation (long flame front and long combustion time) as they get bigger.

 

Diesel engines get more efficient as cylinder size increases. Less friction, less heat loss (reducing surface area to volume).

 

I imagine reciprocating steam engines have the same issues as Diesels.



#53 Kelpiecross

Kelpiecross
  • Member

  • 1,458 posts
  • Joined: October 10

Posted 06 March 2016 - 10:08

http://www.modeleng....cles/hall01.pdf

I think the above demonstrates that model engineers are well aware of indicator diagrams etc. I suspect the unwritten underlying point of this article is attempting to explain why model efficiencies are so low.

Edited by Kelpiecross, 06 March 2016 - 10:13.


#54 J. Edlund

J. Edlund
  • Member

  • 1,323 posts
  • Joined: September 03

Posted 06 March 2016 - 23:26

I doubt the V8s were close to stoichiometric. Perhaps 0.95 in economy mode, 0.9 in power mode? (WOT of course)

 

AFR calc.

Mass Airflow (assume intercooling to ambient) = rho x Displacement/2 x RPM/60 x VE x PR = 1.2 x 1.6E-3 x 1/2 x 11,000/60 x 1.1 x 3.5 = 0.68 kg/s

Mass Fuel flow = 100/3600 = 0.0278 kg/s

AFR = 0.68/0.0278 = 24.4 = lambda 1.66

Higher CAT, lower MAP or lower trapped mass will make this richer but I don't think they are running less than 1.4. (I doubt there will be a lot of scavenge - they want to run the exhaust pressure as high as possible for turbine power)

 

Stratified charge is a no-brainer:

1. They are very lean - unlikely to be homogeneous.

2. No fuel near the walls where quenching kills combustion.

3. Shorter flame travel.

4. Cool end-gas.

 

The main objective with stratified combustion in a gasoline engine is reduced throttling losses at part load using very lean mixtures (which also gives reasonably low pre-cat NOx emissions to deal with). For a F1 or other racing engine part load efficiency is typically a minor issue. With a stratified charge, as engine load increase, the air excess decrease and the fuel injection duration increase and you end up with a larger "fuel cloud" than you would want. At high loads the mixture also tends to get too rich around the spark plug and the engine will begin to produce smoke. This means that as engine load increase, the fuel efficiency advantage of the stratified combustion will decrease and for a regular direct injected gasoline engine this fuel efficiency advantage will be totally gone once you reach a bmep of 5 bar or so, above that homogeneous lean is more fuel efficient.

 

At high loads the advantages of stratified combustion are small; due to the larger amount of air you have a slightly higher ratio of specific heats and slightly smaller heat losses to the cylinderwalls. But the leaner mixture also means you need more air and boost which means the compressor needs more power, which also means more heat needs to be removed by intercooling, and the lower turbine inlet temperature will also hurt turbine efficiency, Add to that the problems to get a good burn with a stratified charge at high load, and the potential problem with soot, this is not what I would call a no-brainer.

 

Power station turbines can be up towards 50% efficient - using every possible method of waste heat recovery.

 

 There has been much debate in model engineering circles as to why the model loco engines are so spectacularly inefficient   when they are pretty much exact scale replicas of a full size engine that may be approaching 10% efficient.     

 

 I have dabbled a little in experimental small steam engines  - mostly in the steam inlet region - using something like  a  petrol injector to meter the steam.  But without any great success.   Also - I am a little wary of home made boilers  - the Mythbusters episodes on exploding household water heaters (at about 350psi from memory)  are something of a warning.       

 

Power station steam turbines which have efficiencies close to 50% typically operate with superheated steam at 700 degC or so (300-350 bar), plus reheat, and have condensation cooling with seawater.



#55 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 07 March 2016 - 04:19

Thanks JE - interesting stuff.

 

 

For a F1 or other racing engine part load efficiency is typically a minor issue.

 

My comments considered full load operation only.

 

 

With a stratified charge, as engine load increase, the air excess decrease

Air excess can be whatever the engine team wish. The F1 engines can add as much air as they wish and fuel is limited by the regulations.

 

 

At high loads the mixture also tends to get too rich around the spark plug

Could not this be avoided with earlier injection, multi pulse injection, leaner AFR?

 

 

But the leaner mixture also means you need more air and boost which means the compressor needs more power, which also means more heat needs to be removed by intercooling, and the lower turbine inlet temperature will also hurt turbine efficiency

I think you meant turbine power not efficiency?

There is no doubt that intercooling hurts efficiency - especially in a compounded engine. I wonder about the extent of intercooling used (Honda deliberately limited intercooling in the RA168e). The engines are not airflow limited so improved charge air density is not an advantage. The remaining advantage (of intercooling) is detonation control (which is where I see the biggest benefit of stratified charge). This (limited intercooling) is supported by reports of 1000*C turbine inlet temp which is high for an engine running at lambda >1.3

 

At constant exhaust temperature, leaning through increased MAP does not result in a reduction in net power from the turbomachinery. In fact at 80% compressor and turbine efficiency or higher there is a net gain.

 

Reports from Mercedes and Ferrari are suggesting very rapid combustion. Mercedes video says 4 elephants on each piston which equates to about 400 bar peak cylinder pressure. If we assumed the peak cylinder pressure at 10 deg ATDC and CR =12:1, MAP = 3.5 bar, what would the CAT be? Would this also give EGT of 1000* C?


Edited by gruntguru, 07 March 2016 - 04:20.


#56 Kelpiecross

Kelpiecross
  • Member

  • 1,458 posts
  • Joined: October 10

Posted 07 March 2016 - 10:07


JE - 700C and over 5000psi - I would not have thought it possible. Would work well in a piston steam engine.

#57 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 07 March 2016 - 14:57

Lean stratified is certainly plausible. A quick calculation of expected airflow and limited fuel flow bear this out. It would also seem to be an absolute necessity to keep turbine inlet temperatures reasonable at a time when running rich is not an option. What tantalises me is how they do it. I want to see port/combustion chamber design; imagine injection strategy, etc. Achieving proper charge stratification - avoiding very rich zones that will cause smoke and lean zones that won't ignite at all - is very difficult, exponentially so at high RPM and load. I should know, that what my dissertation is on and I'm submitting SAE papers soon on one way to achieve this, albeit without F1 specifically in mind.



#58 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 07 March 2016 - 23:40

Perhaps droplet size has a lot to do with it. The smaller the droplets the more they tend to remain entrained and behave as a gas, maintaining the mixture distribution as delivered by the injector. (Of course turbulence after the injection event will still disrupt the stratification.) Apart from the 500 bar injection pressure, I the engines probably use heated fuel (up to 100* C) and high charge air temperature to increase vaporization (fuel vapor will obviously behave as a gas) and reduce droplet size even further.



#59 J. Edlund

J. Edlund
  • Member

  • 1,323 posts
  • Joined: September 03

Posted 28 March 2016 - 17:07


My comments considered full load operation only.

 

Yes, which means the main advantage with stratified charge, reduced pumping losses, is lost.

 


Air excess can be whatever the engine team wish. The F1 engines can add as much air as they wish and fuel is limited by the regulations.

 

Yes, but more air isn't free. It will cost power to compress and increase drag due to increased charge cooling.

 


Could not this be avoided with earlier injection, multi pulse injection, leaner AFR?

 

Earlier injection will make the mixture more homogeneous. The difference between a homogeneous lean and a stratified lean combustion is basically the point of injection, with the former being earlier.

 


I think you meant turbine power not efficiency?

 

I meant the thermal efficiency of the turbine as that is dependent on turbine inlet temperature. With increased air excess TIT will be lower but the energy available to the turbine will be the same (due to greater mass flow); turbine power will drop due to loss of thermal efficiency.

 


There is no doubt that intercooling hurts efficiency - especially in a compounded engine. I wonder about the extent of intercooling used (Honda deliberately limited intercooling in the RA168e). The engines are not airflow limited so improved charge air density is not an advantage. The remaining advantage (of intercooling) is detonation control (which is where I see the biggest benefit of stratified charge). This (limited intercooling) is supported by reports of 1000*C turbine inlet temp which is high for an engine running at lambda >1.3

 

Intercooling should increase cycle efficiency as it reduce compression work, its even used on some gas turbines for that purpose.

 

The racing fuel used by Honda had an initial boiling point of 100 degC which is quite high. I think they had some problems with evaporation, hence the limited charge cooling and fuel preheating.

 

The problem with a stratified charge as I see it, is that lambda 1.3 is very rich for a stratified combustion mode. Typically you can't go richer than about lambda 1.5 with stratified to avoid smoke, while homogeneous lean can't go leaner than roughly lambda 1.3 in order to avoid combustion instability.



Advertisement

#60 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 29 March 2016 - 01:07

 

Yes, which means the main advantage with stratified charge, reduced pumping losses, is lost.

So we move to the next advantage - ignitability of the richer core which also accelerates combustion of the lean main charge. Incidentally, this article suggests the possibility of Mercedes using a pre-chamber (which has a similar effect). https://translate.go...t-text=&act=url

 

 

Yes, but more air isn't free. It will cost power to compress and increase drag due to increased charge cooling.

Depending on isentropic efficiency of the turbomachinery, most of the extra compression work is recovered as increased power available to the turbine.

 

 

Earlier injection will make the mixture more homogeneous. The difference between a homogeneous lean and a stratified lean combustion is basically the point of injection, with the former being earlier.

My point exactly. Early pulse creates homogeneous lean @ perhaps 2.0 AFR and a smaller late pulse creates richer core at the spark plug.

 

 

I meant the thermal efficiency of the turbine as that is dependent on turbine inlet temperature. With increased air excess TIT will be lower but the energy available to the turbine will be the same (due to greater mass flow); turbine power will drop due to loss of thermal efficiency. . . .

 

. . . . Intercooling should increase cycle efficiency as it reduce compression work, its even used on some gas turbines for that purpose.

Notice the contradiction here? Low TIT? - No problem - reduce intercooling!

 

BTW TIT falls but mass flow increases and PR increases. I think you will find turbine power increases - even with intercooling.

 

Gas turbines are a very different kettle of fish because they are typically TIT limited. Intercooling reduces TIT and allows a higher PR which is the real reason for the increased TE. If PR is fixed, intercooling on its own will reduce TE.


Edited by gruntguru, 29 March 2016 - 04:11.


#61 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 29 March 2016 - 02:14

Hmmm this is getting very, very close to my PhD dissertation research...

 

So we move to the next advantage - ignitability of the richer core which also accelerates combustion of the lean main charge. Incidentally, this article suggests the possibility of a pre-chamber (which has a similar effect). https://translate.go...t-text=&act=url

 

 



#62 M100

M100
  • New Member

  • 9 posts
  • Joined: July 03

Posted 30 March 2016 - 09:04

Power station turbines can be up towards 50% efficient - using every possible method of waste heat recovery.

 

Around 46% using a supercritical steam cycle is about the best commercially deployed, but with district heating efficiencies can reach '91%'

 

(I'd take the latter figure with a large pinch of salt)

 

The current record for Combined Cycle Gas Turbines is 61.5%



#63 MatsNorway

MatsNorway
  • Member

  • 2,730 posts
  • Joined: December 09

Posted 30 March 2016 - 13:53

Is pre-chamber still used on diesel cars? im asking because mercedes was doing it earlier than the others. Its been years since i had a look at current gen diesel engines mind you..

Google says not anymore.

 

But i do believe classical Pre-chamber really does not work so well with high rpm.

 

Then again.. pre-chamber is just a extreme squish design.. a squish like design was a given in my mind to get the desired fuel mixture locally.


Edited by MatsNorway, 30 March 2016 - 14:09.


#64 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 30 March 2016 - 15:05

Recommended reading (not F1-specific) would be Heitland et al, SAE 981912 and Oppenheim, SAE 2002-01-0999.



#65 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 30 March 2016 - 23:02

 

Then again.. pre-chamber is just a extreme squish design.. a squish like design was a given in my mind to get the desired fuel mixture locally.

There are similarities. The big difference is that the pre-chamber design (especially along the lines of the Mahle Jet Ignition) projects high temperature gas (hot enough to ignite adjacent mixture) to all extremities of the combustion chamber.

 

Squish projects unburned mix towards the combustion - really just a turbulence enhancer.



#66 MatsNorway

MatsNorway
  • Member

  • 2,730 posts
  • Joined: December 09

Posted 31 March 2016 - 14:13

Forget squish and its intended and normal function. There will be no mixture to push into the area.

 

It would operate/function as a Pre-chamber, with a local mix that spreads out. Difference is that you do not have the chamber in the piston. In my mind you have to run something like this to get an ideal mix. And a light piston.

 

In other words; you inject "late" and ignite it just after.

You don't want the burning to go anywhere near the walls if you can help it. Thats just heat loss and extra cooling needed, you don't want turbulence either because that gives more heat loss due to the movement of gasses over the surfaces.

 

Its all about the injector creating a nice uniform mix. With a burn that goes evenly throught the mixture.

The initial burn chamber is probably ideally hemi like in design.

As the piston travels down the burning mixture heats up the non mixed air, that increases the pressure due to an overall increase in temp in the cylinder. Thats why they run high boost and high compression.

But the temps are overall low to minimise cooling needs.

 

Whatever ratio you guys think they need for maximum power from a fixed fuel, they can run "locally" if deemed most efficient. The exess air then will function to deny/minimise heat loss into the cyl walls etc.

 

Im also guessing a high airflow/low pressure is easier to take out energy from rather than a higher pressure, hotter mix. You want the air temp to drop to as near ambient temp as possible after the turbo.

 

Does it sound probable? does some of it make sense? I should have started to write sci-fi books right?


Edited by MatsNorway, 01 April 2016 - 14:20.


#67 Canuck

Canuck
  • Member

  • 2,072 posts
  • Joined: March 05

Posted 01 April 2016 - 18:47


Isn't the hemi chamber the exact opposite of reduced heat loss with its very high surface area?

#68 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 02 April 2016 - 00:22

Isn't the hemi chamber the exact opposite of reduced heat loss with its very high surface area?

A spherical chamber has the lowest surface to volume ratio and a hemisphere is not far behind.



#69 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 06 April 2016 - 23:21

 

I meant the thermal efficiency of the turbine as that is dependent on turbine inlet temperature. With increased air excess TIT will be lower but the energy available to the turbine will be the same (due to greater mass flow); turbine power will drop due to loss of thermal efficiency.

That is not usually considered "efficiency". Turbine efficiency is usually expressed as a percentage of isentropic efficiency. If you reduce the inlet temperature, that reduces the amount of heat energy available. A good turbine will still convert about 80% of isentropic which is the maximum possible.



#70 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 13 May 2016 - 15:21

Ferrari confirms lean-burn combustion process in F1

 

https://www.springer...rmel-1/10097594

 

Translation by me:

Ioannis Kitsopanidis has confirmed that the turbocharged 1.6-liter V6 petrol engine is used with a lean burn process in F1. But how the V6 ignites the lean mixture remains in the dark.
 
In a presentation at the 37th International Vienna Motor Symposium last April, Ioannis Kitsopanidis, head of engine development for Formula 1, confirmed that the turbocharged 1.6-liter V6 petrol engine designated 059/5 would be operated with a lean combustion process. A chart in his presentation "Evolutions in F1 Engine Technology: Pursuing Performance From Today's Power Unit Through Efficiency" shows that the currently homologated generation of engines operates unders a relative air-fuel ratio of lambda = 1.2 in the mixture formation. That a stratified charge is employed, is ruled out: The operating map window for the lean burn is known to be limited - at high speeds, the time between injection and ignition is not sufficient to form a stratified mixture cloud.
 
How the V6, with technical regulations limiting to a maximum speed of 15,000 / min and the boost pressure is theoretically unlimited - typically up to 3.5 bar (absolute) is reached - ignites the lean mixture, remains in the dark. Concerning the competitors Mercedes was already rumored that the engine of the AMG F1 W07 Hybrid is also operated with a lean burn process - and an evolved form of the scavenged prechamber is used. So far, this technique is mainly used for large-displacement gaseous-fueled engines. Whether and how material problems, caused at high mean effective pressures with the prechamber approach, could be solved in Formula 1 remains unclear. On request of MTZ, Kitsopanidis declined to comment on the ignition system employed. The thermal efficiency of Formula 1 engines according to the presentation figures up to 45 percent.


#71 J. Edlund

J. Edlund
  • Member

  • 1,323 posts
  • Joined: September 03

Posted 21 May 2016 - 23:10

A spherical chamber has the lowest surface to volume ratio and a hemisphere is not far behind.

 

Canuck is correct,a hemispherical combustion chamber have a very large surface area. That a hemisphere has a low surface to volume ratio doesn't help, for a given bore diameter a hemispherical chamber have a larger surface area than a flat chamber. To achieve a reasonable compression ratio with a hemispherical chamber you typically need a piston with a domed piston crown too, increasing the surface area even further.

 

That is not usually considered "efficiency". Turbine efficiency is usually expressed as a percentage of isentropic efficiency. If you reduce the inlet temperature, that reduces the amount of heat energy available. A good turbine will still convert about 80% of isentropic which is the maximum possible.

 

If TIT is lower but mass flow higher the amount of heat available will be the same, the isentropic efficiency of the turbine will obviously be the same but the total efficiency will be lower.

 

Ferrari confirms lean-burn combustion process in F1

 

https://www.springer...rmel-1/10097594

 

Translation by me:

Ioannis Kitsopanidis has confirmed that the turbocharged 1.6-liter V6 petrol engine is used with a lean burn process in F1. But how the V6 ignites the lean mixture remains in the dark.
 
In a presentation at the 37th International Vienna Motor Symposium last April, Ioannis Kitsopanidis, head of engine development for Formula 1, confirmed that the turbocharged 1.6-liter V6 petrol engine designated 059/5 would be operated with a lean combustion process. A chart in his presentation "Evolutions in F1 Engine Technology: Pursuing Performance From Today's Power Unit Through Efficiency" shows that the currently homologated generation of engines operates unders a relative air-fuel ratio of lambda = 1.2 in the mixture formation. That a stratified charge is employed, is ruled out: The operating map window for the lean burn is known to be limited - at high speeds, the time between injection and ignition is not sufficient to form a stratified mixture cloud.
 
How the V6, with technical regulations limiting to a maximum speed of 15,000 / min and the boost pressure is theoretically unlimited - typically up to 3.5 bar (absolute) is reached - ignites the lean mixture, remains in the dark. Concerning the competitors Mercedes was already rumored that the engine of the AMG F1 W07 Hybrid is also operated with a lean burn process - and an evolved form of the scavenged prechamber is used. So far, this technique is mainly used for large-displacement gaseous-fueled engines. Whether and how material problems, caused at high mean effective pressures with the prechamber approach, could be solved in Formula 1 remains unclear. On request of MTZ, Kitsopanidis declined to comment on the ignition system employed. The thermal efficiency of Formula 1 engines according to the presentation figures up to 45 percent.

 

 

Mahle have confirmed that they supply their jet ignition system to Ferrari, so that should answer the question how they ignite the mixture (even if there isn't any specific details how this system is designed for F1).

 

Lambda 1.2 is lower than I expected as some studies seem to suggest lambda 2 is possible with jet ignition. Assuming 3,5 bar absolute, 50 degC, 12,000 rpm and 110% VE the air flow should be around 2390 kg/h suggesting a lambda of 1.6 with 100 kg fuel per hour. However, for next year there is a new regulation that limit the geometric compression ratio to 18:1 so perhaps this difference is explained by the use of early or late inlet valve closure, which would reduce airflow and make such high high compression ratios possible.



#72 MatsNorway

MatsNorway
  • Member

  • 2,730 posts
  • Joined: December 09

Posted 22 May 2016 - 11:17

What is geometric compression ratio? and why are they putting on the brakes on the development?



#73 thegforcemaybewithyou

thegforcemaybewithyou
  • Member

  • 3,225 posts
  • Joined: April 12

Posted 22 May 2016 - 12:59

The max compression ratio of 18 seems very high for a turbocharged engine. Do you think it is possible that an Atkinson cycle (constant late inlet valve closing, no VVT) is already in use in F1?



#74 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 26 May 2016 - 15:17

What is geometric compression ratio? and why are they putting on the brakes on the development?

Geometric CR is merely the CR you generally know (swept volume + clearance volume) / clearance volume, to distinguish it from effective CR which takes valve timing events into account. I also question why the FIA would put a limit of the CR to 18. It would be great to see what the ingenuity of the F1 engine engineers could come up with to see how far one could go.

 

The max compression ratio of 18 seems very high for a turbocharged engine. Do you think it is possible that an Atkinson cycle (constant late inlet valve closing, no VVT) is already in use in F1?

Definitely Miller/Atkinson. My guess is that with the high RPMs late timing is the approach taken because early IVC with the desired valve lift for good volumetric efficiency would mean very high cam profile and valve lift accelerations. Late IVC can also better exploit ram effect and resonant tuning effects.



#75 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 27 May 2016 - 01:24

I doubt that Miller/Atkinson is being used. The intake airflow has to meet two requirements:

 1. Scavenge air to remove combustion products and cool hot components in the chamber.

 2. Meet the AFR requirement (remembering that fuel flow is regulated by rules)

 

I believe AFR is more critical to engine efficiency than CR. AFR dictates the charge mass required per cycle and CR would be increased to the maximum tolerated using that charge mass at the start of the compression process. (In addition, CR is a less effective means of increasing efficiency of a compound engine because the additional in-cylinder expansion has a penalty in reduced turbine work) 

 

Regulating this charge mass is best done by regulating boost pressure - reducing boost reduces compressor work leaving a higher percentage of turbine power available to the MGUH. Reducing charge with Miller/Atkinson valve timing increases the boost pressure required to obtain the required charge.

 

Late IVC also has an unfortunate side effect - VE increases with rpm. With a fixed fuel mass flow from 10,500 - 15,000 rpm these engines require a (somewhat?) reducing air charge per cycle in this range. This would be accomplished with a combination of falling boost pressure and VE.

 

EDIT. 18:1 is a CR limit rule. There is no suggestion that anyone is running CR that high, however the fact this rule is proposed suggests the rules folk are worried it might happen.


Edited by gruntguru, 27 May 2016 - 05:01.


#76 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 28 July 2016 - 04:43

 If TIT is lower but mass flow higher the amount of heat available will be the same, the isentropic efficiency of the turbine will obviously be the same but the total efficiency will be lower.

Leaner mix:

 - reduces TIT

 - increases mass flow

 - increases PR

 

The first will reduce turbine output. The other two will increase turbine output.

 

I did rough calculations for a simple steady flow turbine using constant Cp of 1. Turbine power increases with PR (almost linear up to PR=6 which is the highest I tried). Of course compressor power also goes up and surplus power (Pt - Pc) decreases slightly as you go from PR=3 to PR=4 and drops dramatically by the time you reach PR=6. That was assuming 80% efficiency for both turbine and compressor. At 90% the surplus power increases at PR=4 and lower at PR=6 so there is clearly an optimum PR for the fixed heat input (I assumed 480 kW which I think is a tad high but corresponds to exhaust temp of 1000C at PR=3).



#77 ray b

ray b
  • Member

  • 2,745 posts
  • Joined: January 01

Posted 12 August 2016 - 01:00

from Formula 1 | Yahoo Sport UK - The Incredible F1 Engine Innovations That Have Transformed Formula 1 ''It is quite remarkable that F1 has achieved so much more in two or three years than the entire automotive industry achieved in more than a century. There are, of course, myriad differences in the details of F1 and road car ICE engines but there’s one big secret that has only recently come to light, and it is transforming how F1 engines perform. One inefficiency engineers have really struggled to get a grip of is how the fuel mixture actually burns inside a cylinder. In your parents’ cars, and F1 cars of the past, a spark plug ignited that mixture around about the center of the cylinder. The spark didn’t get to much of the mix and the combustion was not efficient. Fuel injection improved matters but the same basic problem remained – one (or more) sparks still wasn’t causing the fuel to burn particularly efficiently. Enter the boffins at Mercedes. It is believed they’ve been working on technology known as turbulent jet ignition (TJI) and introduced it to their F1 cars when the sport switched to 1.6litre turbo engines in 2014. Here’s how TJI works. Instead of the spark plug being in the cylinder, it is instead housed in a much smaller ignition chamber. Fuel – around 97 per cent of the total for each cylinder cycle – is injected into the cylinder. The remaining 3 per cent (and this is the clever bit) is ignited in the little ignition chamber. This creates plasma which is injected through several small holes into the cylinder, igniting the main fuel mixture evenly and from the outside of the cylinder in, burning with far more efficiency. The process also swirls the plasma, promoting even better combustion. Instead of a spark in the cylinder igniting the fuel mix, you have several plasma jets doing the job – think of it as swapping a Zippo lighter for a handful of blow torches.'' __________________

#78 Greg Locock

Greg Locock
  • Member

  • 5,656 posts
  • Joined: March 03

Posted 12 August 2016 - 05:18

Ah, yes. Indirect injection diesel technology. No, he is absolutely right, nobody has ever heard of that. Especially not Harry Ricardo.



#79 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 12 August 2016 - 11:04

Or indeed Prof. Harry Watson from University of Melbourne, one of the originators of TJI.



Advertisement

#80 desmo

desmo
  • Tech Forum Host

  • 19,399 posts
  • Joined: January 00

Posted 15 August 2016 - 00:03

Ah, Mr. Harry Ricardo. There's an up and comer. Bleeding edge.



#81 ray b

ray b
  • Member

  • 2,745 posts
  • Joined: January 01

Posted 15 August 2016 - 17:16

http://arstechnica.c...ficiency-gains/

above link has a picture of the pre/spark/injector assembly
looks like it is an add on to the existing spark plug hole

Edited by ray b, 15 August 2016 - 17:19.


#82 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 15 August 2016 - 23:54

Nice.

One question remains. Everything published so far about Mahle TJI describes a system with 2 injectors - one in the pre-chamber delivering 2% - 3% of total fuel and another (usually described as a port injector but could easily be a direct injector) delivering the main, lean charge. The F1 rules only permit 1 injector per cylinder so it will be interesting to learn how they have implemented TJI without 2 injectors.



#83 Kelpiecross

Kelpiecross
  • Member

  • 1,458 posts
  • Joined: October 10

Posted 17 August 2016 - 05:36

http://arstechnica.c...ficiency-gains/

above link has a picture of the pre/spark/injector assembly
looks like it is an add on to the existing spark plug hole

Rather than compare this system to diesel indirect injection etc. surely it is a variation on stratified charge technology. It also appears that it could be used in normal road-going cars.

#84 ray b

ray b
  • Member

  • 2,745 posts
  • Joined: January 01

Posted 17 August 2016 - 14:47

&

Nice.
One question remains. Everything published so far about Mahle TJI describes a system with 2 injectors - one in the pre-chamber delivering 2% - 3% of total fuel and another (usually described as a port injector but could easily be a direct injector) delivering the main, lean charge. The F1 rules only permit 1 injector per cylinder so it will be interesting to learn how they have implemented TJI without 2 injectors.

&amp

I suspect the single injector in the pre-chamber is used for both
ether the main injector allows a bit of fuel to remain in the pre
or a second weak/small shot timed to the spark is used to get the 3% pre chamber load

but I just read the links so guessing

I do note I thought up a similar spark/injector DI use years ago
but droped it when a google patent search found bosch already had a patent on it
but the pre-chamber and resulting plasma jets is a new one
I wonder if the bosch patents are paid or not on this tec

Edited by ray b, 17 August 2016 - 14:50.


#85 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 17 August 2016 - 23:00

The general concept has been around for perhaps 100 years but the TJI refinement combined with the pent-roof chamber seems to be something of a breakthrough.

 

On the topic of "2 injectors per cylinder". TJI works well without the pre-chamber injector but doesn't enable the lean extension to 2.0+ that is possible with pre-chamber injection.



#86 Kelpiecross

Kelpiecross
  • Member

  • 1,458 posts
  • Joined: October 10

Posted 18 August 2016 - 05:43

Rather than compare this system to diesel indirect injection etc. surely it is a variation on stratified charge technology. It also appears that it could be used in normal road-going cars.

I mentioned the "stratified charge" approach as in this forum most of the speculation was that the F1 engine actually ran in certain modes by compression ignition. This stratified charge approach is very much the same as Honda's CVCC engine - down the point where the jets of flame emerge through tiny holes in a metal screen into the main combustion chamber. Or do some of the other F1 engines actually operate in a compression ignition manner?

#87 Wuzak

Wuzak
  • Member

  • 6,895 posts
  • Joined: September 00

Posted 24 October 2016 - 04:48

http://www.motorspor...barrier-829341/



#88 Wuzak

Wuzak
  • Member

  • 6,895 posts
  • Joined: September 00

Posted 26 October 2016 - 11:32

http://www.f1analisi...da.html?refresh

 

This claims that Mercedes has 975hp for qualifying and 940hp for the race, 960hp and 940hp for Ferrari and 940hp and 915hp for Renault.



#89 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 14 November 2016 - 17:49

Any remaining doubts that F1 is employing lean-burn and prechamber ignition?

 

Screenshot%202016-11-14%2012_11_42_zpsdi

From PMW Magazine Jan. 2017


Edited by TDIMeister, 14 November 2016 - 17:51.


#90 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 14 November 2016 - 23:47

That article still speaks of an auxiliary injector which contravenes the F1 rules of "one direct injector per cylinder". I am keen to learn what changes they have made to meet this regulation - or have they gained dispensation from the FIA for a "special case" injector?



#91 bigleagueslider

bigleagueslider
  • Member

  • 1,235 posts
  • Joined: March 11

Posted 15 November 2016 - 07:11

That article still speaks of an auxiliary injector which contravenes the F1 rules of "one direct injector per cylinder". I am keen to learn what changes they have made to meet this regulation - or have they gained dispensation from the FIA for a "special case" injector?

I suppose one could argue that the prechamber injector is "indirect". Diesel engines that have the injector located in a prechamber are normally referred to as indirect Injected (IDI), while those that have the injector located in an open combustion chamber are referred to as direct injected (DI).

 

Here's a relevant section of the 2016 F1 tech regs: "5.10.2 There may only be one direct injector per cylinder and no injectors are permitted upstream of the intake valves or downstream of the exhaust valves."



#92 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 16 November 2016 - 02:44

Yes that does look like a possible loophole. I am sure the original rule intent was "only one injector per cylinder" and "must be a direct injector".



#93 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 16 November 2016 - 02:51

That article still speaks of an auxiliary injector which contravenes the F1 rules of "one direct injector per cylinder". I am keen to learn what changes they have made to meet this regulation - or have they gained dispensation from the FIA for a "special case" injector?

Nowhere in that article do I read or interpret the use of an auxiliary injector, nor is there any need for one.  The article states, "Inside the jet igniter assembly is a small ignition chamber with a direct injector that provides a small amount of auxiliary fuel (<5% of the total system fuel)..."

 

This can easily be achieved by a single injector per cylinder, where fuel is introduced over multiple injections - an early one during the intake stroke that goes through the prechamber orifices (remember that there's 500 bar behind the injected fuel to ~MAP in the cylinder) and diffuses/entrains into the whole cylinder for good homogenisation, and one or more late injections before spark ignition that gets confined largely within the prechamber and in-close proximity thereof. 


Edited by TDIMeister, 16 November 2016 - 02:52.


#94 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 16 November 2016 - 08:42

TDIM. The system you are suggesting sounds unlikely. The main injection event would merely displace some of the gas remaining in the pre-chamber, leaving most of the fuel in the pre-chamber. The connection passages to the main chamber are quite narrow to generate high velocity and penetration to the chamber extremities. The result would be an excessively rich mixture in the pre-chamber. Even if it was ignitable, it would be impossible to control the pre-chamber mixture accurately enough under all operating conditions. Add to this the lack of auxiliary-charge precision - using the same injector to deliver full power main-charge and part load auxiliary-charge, a ratio of perhaps 500:1.



#95 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 16 November 2016 - 17:05

TDIM. The system you are suggesting sounds unlikely. The main injection event would merely displace some of the gas remaining in the pre-chamber, leaving most of the fuel in the pre-chamber. The connection passages to the main chamber are quite narrow to generate high velocity and penetration to the chamber extremities. The result would be an excessively rich mixture in the pre-chamber. Even if it was ignitable, it would be impossible to control the pre-chamber mixture accurately enough under all operating conditions. Add to this the lack of auxiliary-charge precision - using the same injector to deliver full power main-charge and part load auxiliary-charge, a ratio of perhaps 500:1.

None of your stated "challenges" have not be addressed by current established technology, a lot from the Diesel world. I'll attempt to do so in turn.

 

"The main injection event would merely displace some of the gas remaining in the pre-chamber, leaving most of the fuel in the pre-chamber."

I disagree and can draw from my own experience from my PhD research in a divided-chamber SI engine - I did schlieren imaging and CFD simulations that showed that a substantial amount of fuel injected into this chamber leaves it into the main cylinder. I used only 10 bar injection pressure, imagine what 500 bar would do. In usual DISI engine practise with or without a divided chamber, fuel injection would take place during the intake stroke to give sufficient time for fuel evaporation and formation of a well homogenised mixture in the main cylinder. I believe F1 does the same. My engine was a 2-stroke, so fuel injection occurred after exhaust port closure in a window from 90° to 85° BTDC. When cylinder pressure rises during the compression stroke, mass transfer of the gas mixture in the main cylinder - appreciably lean by design - go back into the divided chamber to equalise the pressures and reduce the equivalence ratio down to the range for most rapid and robust inflammation.

 

The connection passages to the main chamber are quite narrow to generate high velocity and penetration to the chamber extremities. 

Doesn't matter. There is a pressure gradient between the injected fuel and the prevailing pressures in the divided chamber and main cylinder, and mass transfer is driven by this gradient. High velocity, turbulent gas issuing out from the prechamber due to combustion-induced gas expansion is no different than going the other way back into it again due to differential pressure during the compression stroke, only a matter of degree.

 

The result would be an excessively rich mixture in the pre-chamber. 

Not necessarily. If well optimised, as I'm sure F1 engineers expend a lot of effort doing, you can control the lambda in the prechamber quite precisely. In my own research engine, the prechamber is, as expected and as you correctly stated at the outset, composed almost entirely of pure fuel in the moments right after the end of injection - lambda approaches zero, As the compression stroke proceeds and lean mixture returns back into the divided chamber from the main cylinder and remixes turbulently with the super-rich contents in the former, I can target my equivalence ratio to pretty much whatever I want through injection timing and strategy, limited on the lower end only by what the lambda of the mixture in the main cylinder is coming back in from the main cylinder.

 

Even if it was ignitable, it would be impossible to control the pre-chamber mixture accurately enough under all operating conditions. 

What makes it so impossible? Piezo injectors in Diesel engines are already delivering up to 7 separate injection events in a single cycle. Distinct injections can be controlled down to an order of 1°CA apart from the end of one to the start of the next. Granted, Diesels, run at much lower speeds than F1; let take one point at 3000 RPM vs. 15000 RPM -- assuming the same limiting response times, injections can be spaced in the order of just 5°CA apart in F1 engine speeds at max RPM.

 

Add to this the lack of auxiliary-charge precision - using the same injector to deliver full power main-charge and part load auxiliary-charge, a ratio of perhaps 500:1.

Addressed in my previous paragraphs, except I will add that injection quantities in Diesel piezo injectors have an incredible granularity of control that you may not be aware of. The pilot injection can go down to as little as 0.8 mg/injection, well less than the 5% of the total injected fuel quantity for the cycle (which can be anywhere from 20-50 mg/stroke in a typical Diesel engine at full load depending on per-cylinder output) quoted in the article I posted above. I understand what you're implying, but it is doubtful that the ratio is anywhere near 500:1. At its most simplistic, it's ~95:5 as per the article in a single cycle, but allowing for the range of operating loads, I get it. In Diesel engines, the oft-quoted injection pressure (e.g. 1800, 2000 bar, etc.) represents the peak. It doesn't operate at this pressure over the entire load and speed range. I don't suspect F1 engines run at the full limiting 500 bar from idle/overrun to max power either; maybe it does, but it doesn't matter and doesn't weaken the argument. Nevertheless, going back to Diesel engine practise, it is 100% reliant on fuel injection for load control, going from idle to upwards of 30 bar BMEP without any difficulty, and it has emissions regulations to worry about! The precision and control is only an encumbrance if you're stuck on the first point of the mass transfer interaction going on between the divided chamber and main cylinder. 


Edited by TDIMeister, 16 November 2016 - 17:58.


#96 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 16 November 2016 - 18:20

I might add that one potential way to get around the single direct injector per cylinder rule without the concerns that gg raised is simply to have an injector with two coaxial sets of spray cones and concentrically lifting injector needles. One spray cone goes into the main cylinder and the other is directed into the divided chamber. Both are actuated separately by concentric needles. I have seen drawings and patents of this but don't have the references at hand at the moment. 

 

However, Fig. 1 from here shows the idea.

ftp://78.38.77.30/cee/m.jafari/motor/ICE%20projects-89/0%20(29).pdf

 

Edit: I still believe my originally hypothesized scenario is the case because it is robust and is widely documented in the literature, if not specifically for F1 applications.


Edited by TDIMeister, 16 November 2016 - 18:26.


#97 Greg Locock

Greg Locock
  • Member

  • 5,656 posts
  • Joined: March 03

Posted 16 November 2016 - 21:42

Almost completely irrelevant war story-

 

When EFI was being introduced at Lotus they ran into a problem with idle stability because injectors big enough to handle the max power had a tolerance for fuel flow vs pulse width based on their max flow rate. So as they pushed the turbo towards the heady heights of power now available in a Camry the fuel flow at idle was too erratic. So two injectors were added further back in the manifold, meaning they could use smaller injectors, and only used 4 of them at idle. They could even have just used the two upstream ones, for better mixing, I suppose.


Edited by Greg Locock, 16 November 2016 - 21:43.


#98 gruntguru

gruntguru
  • Member

  • 6,733 posts
  • Joined: January 09

Posted 17 November 2016 - 03:51

The whole point of 500 bar injection pressure is atomization and droplet velocity. Injecting into a pre-chamber and accepting whatever fuel emerges through tiny holes into the main chamber is not the sort of precision that gets you beyond 50% TE. 

 

Same goes for the fuel quantity remaining in the pre-chamber and the resulting pre-chamber mix. This will be highly dependent on - not only the usual variations in operating point (load, speed, air temp, coolant temp, spark timing, boost, overall AFR etc) but also things like pre-chamber temperature which will be time and transient dependent. This makes accurate control and mapping of the auxiliary injection event highly problematic regardless of issues with injector resolution.

 

Regarding injector resolution - even if the dynamic range from largest main-injection-duration to smallest pilot-injection-duration is only 200:1 you still have the situation where a main-injection resolution of 1/10,000 becomes a pilot resolution of 1/50 which would of course be inadequate.

 

Regarding 500 bar injection "clearing" the pre-chamber (vs 10 bar). The pre-chamber does not see 500 bar. 500 bar is the pressure drop across the injector nozzle. The only effect in the pre-chamber is smaller droplets and higher velocities.

 

The volume of fuel (main-injection) is roughly 1/100th the volume of the pre-chamber. I'm not sure how much of that fuel is going to escape the pre-chamber through the fine orifices with no pressure differential other than the 1/100 of the pre-chamber that is displaced by the fuel spray. Consider also that only 10% of the main-injection needs to remain in the pre-chamber to result in stoichiometry. (Assuming lambda 2.0 for main charge)


Edited by gruntguru, 17 November 2016 - 03:52.


#99 MatsNorway

MatsNorway
  • Member

  • 2,730 posts
  • Joined: December 09

Posted 17 November 2016 - 12:53

Canuck is correct,a hemispherical combustion chamber have a very large surface area. That a hemisphere has a low surface to volume ratio doesn't help

 

Point is that it is better than flat top, flat piston. If a chamber shall contain 500cc of mixture you want it to be round to minimise surface area.

 

We "only" use "flat" top, "flat" piston to get the desired compression ratio. And the only reason they have to be flat top, flat piston is because bore/stroke ratio is more important with this type of engine and application.

 

A ship engine with more stroke and lower rpm runs more hemi like piston and head. I googled to prove my teories and sure enough:

 

http://www.marine-kn...ston-Rings1.jpg

http://www.bcmtourin...0009-jpg.59825/

 

Well i say only but there is also piston speed to consider, weight of the piston and the extra friction that causes, the flame front and all kinds of other variables. Hence modern piston designs.

 

To say it differently..

If i was only chasing Horsepower with unlimited fuel i would run much bigger bore like the old engines. Excess pressure would simply get dumped out so you could make more "peak" pressure to work with. Hence the rpm, overlap on the cams etc.

 

There is a sweetspot where bore/stroke ratio + rpm is optimal for a given fuel type and flow. If they open up the rules F1 engineers will find it..

They probably have some idea where that is. So they made the odd fuel flow ramping  and a locked Bore and stroke ratio to keep the show factor in the sport. "We can not simply be seen dieseling around Monaco can we"

 

Personally i still vote for allowing Straight 5 engines. Some would prob give it a go as it sounds gnarly (ferrari) and it has one less piston so it would be more efficient.

 

Offset for that engine is packaging.

 

Increasing efficiency has the added challenge of giving less reusable energy for the turbo KERS. If all the pressure drop happens in the engine the turbo is simply a back pressure generator. That will never happen ofc.


Edited by MatsNorway, 21 November 2016 - 07:58.


Advertisement

#100 TDIMeister

TDIMeister
  • Member

  • 286 posts
  • Joined: January 06

Posted 18 November 2016 - 18:16



The whole point of 500 bar injection pressure is atomization and droplet velocity. Injecting into a pre-chamber and accepting whatever fuel emerges through tiny holes into the main chamber is not the sort of precision that gets you beyond 50% TE. 

 

Same goes for the fuel quantity remaining in the pre-chamber and the resulting pre-chamber mix. This will be highly dependent on - not only the usual variations in operating point (load, speed, air temp, coolant temp, spark timing, boost, overall AFR etc) but also things like pre-chamber temperature which will be time and transient dependent. This makes accurate control and mapping of the auxiliary injection event highly problematic regardless of issues with injector resolution.

 

Regarding injector resolution - even if the dynamic range from largest main-injection-duration to smallest pilot-injection-duration is only 200:1 you still have the situation where a main-injection resolution of 1/10,000 becomes a pilot resolution of 1/50 which would of course be inadequate.

 

Regarding 500 bar injection "clearing" the pre-chamber (vs 10 bar). The pre-chamber does not see 500 bar. 500 bar is the pressure drop across the injector nozzle. The only effect in the pre-chamber is smaller droplets and higher velocities.

 

The volume of fuel (main-injection) is roughly 1/100th the volume of the pre-chamber. I'm not sure how much of that fuel is going to escape the pre-chamber through the fine orifices with no pressure differential other than the 1/100 of the pre-chamber that is displaced by the fuel spray. Consider also that only 10% of the main-injection needs to remain in the pre-chamber to result in stoichiometry. (Assuming lambda 2.0 for main charge)

I can understand your point and while we can go around in circles with speculation, it would be instructive to go right to the source of how TJI is implemented originally and go from there. With picture(s) at hand, we can move forward to imagine how it would be done in F1 whilst respecting the rulebook. So I'll let the picture first speak for itself...

 

Screenshot%202016-11-18%2012_11_10_zpstn

From Toulson et al (SAE 2010-01-2263) - A review of pre-chamber initiated jet ignition combustion systems.

 

In the roadcar engine implementation, there is a PFI injector in the intake port - not "apparently" allowed in the rules. However, given that the TJI shown is a threaded-in module that incorporates the spark plug, direct fuel injector and pre-chamber, it's not hard to imagine that the F1 guys have cleverly packaged a single fuel injector element that sprays a small amount of fuel into the prechamber and the rest into the main combustion chamber. Some variation of the coaxial/concentric spray hole idea I put forward before is quite conceivable.

 

Here is what I can show of my own research work:

http://bit.ly/2eOkXOV (The forum doesn't seem to allow me to embed a video directly)

 

What it doesn't show is the portion of the fuel that leaves the chamber from the channels at the bottom into the main cylinder and granted, a different fuel (hydrogen) directly injected is simulated. With gasoline, there is a much greater momentum of the liquid spray and additional processes of spray wall impingement and evaporation, but the principle holds. What ends up at the moment before ignition is nonetheless very homogeneous at a targeted equivalence ratio.


Edited by TDIMeister, 18 November 2016 - 18:23.