Preventing abnormal combustion

[WhiteBlue]

The amount of air in the cylinder is larger with DI, because with port injection the fuel displaces some of the air before it is compressed in the combustion chamber.

I don't think that's the reason. Port injection causes a lot of fuel to stick to the walls instead of going into the mixture. DI wastes no fuel for such purposes. With good DI you my have air excess but no fuel excess. That makes a difference in terms of fuel efficiency.

[autogyro]

You are correct WB but so am I.
He asked what were the differences.

Taking the boost up on the turbo compensates for the air displaced by the fuel and wastes energy.
The fuel sticking to the port walls messes up the mixture and wastes fuel.
Neither matches good DI.

[pgfpro]

Good topic!!!

Looking at our new 2014 F1 engines at a fuel limit of 100 Kg/hr at 10500 rpm.

10500 rpm / 2 = 5250 cycles per min.
5250 cycles per min / 60 = 87.5 cycles per-second.

Fuel delivery
Our new 2014 F1 engine will use 100 Kg/hr or 100 kg / 60 min = 1.6667 kg/min or 1.6667 kg / 60 second = 0.0278 kg/sec.
0.0278 kg/sec / .001 = 27.77 grams a second
27.77 grams per second / 87.5 cycles per-second = 0.3175 grams of fuel per total engine cycle
0.3175 / 6 cylinders 0.0529 grams per cylinder

So imagine what 0.0529 grams of fuel looks like... a very small droplet. Then imagine what the space this droplet takes up in the intake track from the intake valves to T/B

[Edis]

If you accept that "knock" implies spontaneous detonation type combustion ahead of the flame front in an SI engine, then it is not so much time dependent. Instead, it relies more on flame speed, heat transfer, air/fuel mixture, and charge temperatures.

"Knock is the name given to the noise which is transmitted through the engine structure when essentially spontaneous ignition of a portion of the end-gas - the fuel, air, residual gas, mixture ahead of the propagating flame - occurs"
John B. Heywood

End gas temperature and time (along with fuel properties) are the main factors that determine if knock will occur. The higher the end gas temperature, and the longer the time available until the flame front arrives, the more likely is knock to occur.

The peak temperature of the endgas trapped in the quench area is almost entirely determined by heat transfer. The thickness of the endgas volume in the quench area at TDC in a high CR engine is only about 1mm. And it's precisely because this endgas layer is so thin and has a very high rate of heat transfer into the piston crown and cylinder head deck that it inhibits detonation. In fact, this one reason most production auto engines don't use much quench area. The endgas mixture in the quench area never gets hot enough to ignite, so it results in large amounts of unburned HC emissions.

Since the time available for heat transfer is very short, it's impact on the end gas temperature is not that significant.

Here is a plot of the unburned zone temperature, using the GT-Suite simulation software:

As you can see the temperature of the unburned zone is nearly 900 kelvin in this case - well above the self ignition temperature of gasoline - and you can see that the temperature rises rapidly after ignition have occurred just before TDC. You can ignore the rapid jump to 1500 K when the exhaust valve opens.

OK, so technically air/fuel mixtures don't self ignite "instantly". The combustion process requires the correct combination of fuel/oxygen association and (heat) energy input to initiate. But in reality, the constant volume combustion process of HCCI is virtually instantaneous, and it's because of the very rapid heat release rate it produces that it is so thermally efficient. In theory, the HHCI process can be used with any fuel or stoichiometric ratio. But the best efficiency results are achieved with lean mixtures and a detonation resistant fuel. The very rapid constant volume combustion of HCCI also results in much lower peak combustion temperatures, which virtually eliminates any formation of NOx compounds.

You need to separate "combustion process" from "ignition delay". Yes, the HCCI combustion process is very fast due to a lack of a flame and flame propagation, that is why very lean mixtures are used - they slow the combustion process down. But a fast combustion does not mean the fuel will ignite instantly, and with a HCCI engine there is plenty of time to self ignite, usually the whole compression stroke.

HCCI engines are in practice limited to lean mixtures, at richer mixtures they begin to suffer from a type of knocking. The fast burn at richer mixtures will also result in high cylinder pressures/temperatures, and this will increase the heat losses to the cylinder walls. However, since there are few temperature inhomogeneities NOx production is low.

As for combustion in DI CI diesel engines, the ignition delay is dependent upon many factors. The combustion process in a DI diesel engine is no different than that of any other engine. The combustion process will only initiate when there is the proper combination fuel/oxygen association and (heat) energy input. One of the biggest factors in DI diesel ignition delay is the mean droplet size of the fuel spray (Sauter). If you look at the combustion process at the micro level, you'll note that combustion in a DI diesel only occurs at the surface of each fuel droplet, where there is the correct ratio of oxygen and fuel. Plus, the greater mass of large fuel droplets means they take longer to heat up and evaporate. This is critical because only fuel vapor will combust, and not liquid fuel. Thus a well dispersed spray of numerous, very tiny fuel droplets will result in less ignition delay, and will produce faster combustion.

The combustion process of a diesel is different than in a SI or HCCI engine. In a SI engine we have a flame propagation, in the diesel we have a diffusion flame. In the HCCI engine we have no flame at all, but a slow chemical oxidation.

Ignition delay in a diesel is mostly a function of temperature (pressure) and fuel cetane. The following graph shows ignition delay in milliseconds against cylinder pressure at start of combustion. As we can see the ignition delay increase with a reduced pressure (temperature). The data are from measurements on a six cylinder commercial diesel engine.



When you've got a short ignition delay, less fuel will have evaporated before the fuel starts to burn.

The amount of air in the cylinder is larger with DI, because with port injection the fuel displaces some of the air before it is compressed in the combustion chamber.

I don't think that's the reason. Port injection causes a lot of fuel to stick to the walls instead of going into the mixture. DI wastes no fuel for such purposes. With good DI you my have air excess but no fuel excess. That makes a difference in terms of fuel efficiency.

Autogyro is indeed correct, and that particular effect is described already in NACA papers from the 1940'ties. With port injection some of the air is replaced with fuel vapor, causing a small loss in volumetric efficiency. The cooling offered by the fuel is not enough to compensate for the air displaced by fuel vapor.

The fuel that sticks to the walls in the intake ports is not "wasted" in any way, but it needs to be considered in the control of the engine. This is because fuel sticks to the walls when the pressure is increased, during a load increase for instance, then it leaves the walls when the pressure is decreased, like during a load decrease. This is why the air fuel mixture is enriched during a rapid load increase, and made leaner if the load on the engine rapidly decreases.

[riff_raff]

"Knock is the name given to the noise which is transmitted through the engine structure when essentially spontaneous ignition of a portion of the end-gas - the fuel, air, residual gas, mixture ahead of the propagating flame - occurs"
John B. Heywood

I like Dr. Heywood, and I have one of his books. But to be specific, the "knocking" noise is the block/cylinder head structure's acoustic response to the impact on their surfaces of high pressure, high velocity waves propagating through the combustion gas as a result of spontaneous combustion. Much of the temperature rise within the end gas results from the compression heating created ahead of the main flame front. The pressure wave situation is compounded when the main flame front and the pressure wave front from spontaneous ignition in the end gas strike each other.

Ignition delay in a diesel is mostly a function of temperature (pressure) and fuel cetane. The following graph shows ignition delay in milliseconds against cylinder pressure at start of combustion. As we can see the ignition delay increase with a reduced pressure (temperature). The data are from measurements on a six cylinder commercial diesel engine.

Actually, the injection spray characteristics can have a significant impact on DI diesel ignition delay. A well dispersed spray of fine droplets will produce less ignition delay than a confined spray of very large droplets.

[Tommy Cookers]

the F1 Ferrari engine tests show the combustion to be slow (relative to the ideal at 18000 rpm) pre tdc (and post tdc)
the ignition advance required (around 60-70 deg btdc) supports that view
in the pre tdc phase unnecessary (or at least non-ideal) work is being done against the piston
(riff raff, I know some is unavoidable, but it looks excessive here, a compromise in the cause of maximum power via maximum rpm)
the speeding of this early combustion would help by allowing less ignition advance
new DI allows very late, rapid, injection timing (helping CR), but that is rather wasted with extreme ignition advance

2014 (and current?) F1 bans any ignition method other than the usual (even having a list of banned potential injection technologies)
but .......
the latest DI can inject in less than 1 mSec

EDIT ................ RING SYSTEM ........ it's not CI, not SI, not HCCI
early combustion of the fuel/air mixture would be more rapid if combustion was started by controlled and contained detonation
combustion then proceeding in the normal way (of an otto cycle engine)
this has been done by using timed injection of small amounts of a pro-detonation fluid to initiate combustion by detonation
(the spark ignition equipment was not fitted to these engines),
as the 'Ring System', it was tested on a wide scale in WW2 aircraft engines by a country that already used DI in all such engines
the 'Ring System' was shown to improve BTE by 12-15% at 50-90% power (and seemed able to run 2.5x leaner than stoichiometric)
(it was considered for service use when spark plug problems had become severe, with high TEL and other problems)
the detonant fluid (ether ?) quantity was varied from 10% down to 2% (of fuel quantity) at high power
(oxygenates eg ethers are F1 mandated as a fuel constituent)

has this sort of thing been proposed or tried in modern times ?
(is it on the F1 banned list ?)

[riff_raff]

....new DI allows very late, rapid, injection timing (helping CR), but that is rather wasted with extreme ignition advance.......this was done experimentally on a wide scale with WW2 aircraft engines by a country that used DI in all such engines......it was shown to improve BTE by 12-15% at 50-90% power (and seemed able to run 2.5x leaner than stoichiometric)

TC-

The benefit of late GDI injection is improved fuel latent heat effect. When the fuel is injected later in the compression stroke, the temperature of the compressed intake air charge is higher, and the temp drop produced by the evaporation of the injected fuel is greater.

As for the GDI used by Daimler-Benz or BMW recip engines during WWII, it was used primarily to minimize mixture variations from cylinder-to-cylinder, and fuel starvation during high-g maneuvers. Best power and SFC are usually at F/A mixtures close to stoichiometric.

[olefud]

The benefit of late GDI injection is improved fuel latent heat effect. When the fuel is injected later in the compression stroke, the temperature of the compressed intake air charge is higher, and the temp drop produced by the evaporation of the injected fuel is greater.

Is this a greater ultimate temp drop or just a larger late stage drop? It would seem that solely regarding latent heat, PI would provide a cool temp earlier -promoting better flow density- with the same ultimate general temp, though the earlier temp drop could allow time for reheating prior to combustion.

[riff_raff]

Is this a greater ultimate temp drop or just a larger late stage drop? It would seem that solely regarding latent heat, PI would provide a cool temp earlier -promoting better flow density- with the same ultimate general temp, though the earlier temp drop could allow time for reheating prior to combustion.

Injecting the fuel late in the compression stroke, when the charge air temp is greater, results in more heat being absorbed via the fuel latent heat of evaporation. Thus, the charge temp near TDC is lower and less likely to detonate.

With PI, you might achieve a max temp drop of 30degF in the intake air from the fuel latent heat. With GDI later in the compression stroke, you could likely achieve 2 or 3 times that amount of temp drop in the intake air charge.

[olefud]

Is this a greater ultimate temp drop or just a larger late stage drop? It would seem that solely regarding latent heat, PI would provide a cool temp earlier -promoting better flow density- with the same ultimate general temp, though the earlier temp drop could allow time for reheating prior to combustion.

Injecting the fuel late in the compression stroke, when the charge air temp is greater, results in more heat being absorbed via the fuel latent heat of evaporation. Thus, the charge temp near TDC is lower and less likely to detonate.

With PI, you might achieve a max temp drop of 30degF in the intake air from the fuel latent heat. With GDI later in the compression stroke, you could likely achieve 2 or 3 times that amount of temp drop in the intake air charge.

I’m not questioning empirical results; but there’ a bit of an anomaly in the energy balance. The same amount of fuel provided PI or DI should provide the same latent heat cooling if both evaporate to the same degree. Both should be very near adiabatic – PI a bit less so. The PI fuel might see somewhat lower pre-TDC temps since it would have already cooled the charge somewhat, but the temp at TDC should be more than adequate to complete the evaporation.
Or the DI air could be cooled 30 degrees more than the PI charge since it arrives hotter yet end up at about the same temp.
A small point but sometimes that’s where the good stuff lives.

[WhiteBlue]

.. the DI air could be cooled 30 degrees more than the PI charge since it arrives hotter yet end up at about the same temp. A small point but sometimes that’s where the good stuff lives.

It appears that is the explanation if you have the liberty of freely selecting the charge air cooling that you want to apply.

For the 2014 F1 engines Renault predict a hot debate between the engine engineers and the chassis engineers over the amount of inter cooling. The engine guys will want to have maximum cooling capacity and the chassis guys will try to keep the size of the radiators down to help with packaging and aerodynamics. At least that was the story when I read the article in race car engineering.

[Tommy Cookers]

For the 2014 F1 engines Renault predict a hot debate between the engine engineers and the chassis engineers over the amount of inter cooling. The engine guys will want to have maximum cooling capacity and the chassis guys will try to keep the size of the radiators down to help with packaging and aerodynamics.

that is a hot debate because the need for charge cooling is greatly increased because the rules have banned any fuelling into the supercharger
have banned what has always been the overwhelming choice of designers
who believed that no amount of cooling after supercharging allowed as good a delivery from the supercharger as evaporative cooling in the supercharging process ??

[WhiteBlue]

that is a hot debate because the need for charge cooling is greatly increased because the rules have banned any fuelling into the supercharger
have banned what has always been the overwhelming choice of designers
who believed that no amount of cooling after supercharging allowed as good a delivery from the supercharger as evaporative cooling in the supercharging process ??

Do you realize that you are following a circular argument? It was already established that latent cooling from fuel evaporation works best when you inject directly into the cylinder. You can sensibly only inject as much fuel as the engine needs for optimum combustion. So any additional fuel to cool the charge air would be surplus to requirements. To use surplus fuel to cool the charge air appears foolish to me.

[riff_raff]

Do you realize that you are following a circular argument? It was already established that latent cooling from fuel evaporation works best when you inject directly into the cylinder. You can sensibly only inject as much fuel as the engine needs for optimum combustion. So any additional fuel to cool the charge air would be surplus to requirements. To use surplus fuel to cool the charge air appears foolish to me.

WB- This may or may not be true. Whether or not injecting fuel ahead of a compressor stage can be more beneficial than direct cylinder injection would depend upon the particulars of the supercharging system. For example, if you are using very high boost levels and multiple stages of supercharging, then there may be benefit to injecting some fuel ahead of the compressors. Using fuel latent heat to cool the air charge instead of air-air heat exchangers can result in less heat rejection. Plus, the compressors will give better performance at high PRs with cooler/denser airflow at the inlet.

[WhiteBlue]

WB- This may or may not be true. Whether or not injecting fuel ahead of a compressor stage can be more beneficial than direct cylinder injection would depend upon the particulars of the supercharging system. ..

@riff_raff
my exchange with TC was specific to the 2014 F1 engines because he referred to FiA policies and regulations. So I think for that particular case my remarks are still correct.

Also if we consider intake injection then we would loose the benefits of direct injection on the combustion efficiency. The potential power losses would most likely exceed any gains from the spraying. In the end in a fuel flow regulated formula the over all fuel efficiency of the engine determines the power output. I would be very sceptical that fuel spraying into the air intake can match or outperform the given 2014 engine design.