Quote:
Originally Posted by CaptainDoggles
That's an interesting take on direct injection, Viper.
I'd always assumed DI was first introduced to more reliably control mixture, eliminate pre-ignition and get a stratified charge in CI systems.
|
DI was indeed initially introduced to improve mixture control and avoid the various problems associated with carburettors.
In general, a single carburettor isn't likely to give good mixture distribution to a multi-cylinder naturally aspirated piston engine, because evaporation isn't completed before the charge reaches the intake manifold, and so you get a fairly complex multi-phase flow.
However, if you've got a supercharger between the carburettor and the intake manifold, things get a lot better because the supercharger homogenises the mixture and also increases its static temperature. This means that much more, if not all of, the fuel evaporates; diffusion is then very helpful in further homogenising the mixture. Furthermore, the flow is likely to warm the induction manifold sufficiently that fuel doesn't condense upon contact with it.
This means that supercharged engines which put the fuel into the airflow (whether via a carburettor or some kind of injection system) upstream of the supercharger will tend to deliver a pretty consistent FAR to all of their cylinders. This removes one of the main motivations for direct injection.
However, it must be stressed that this is something of a special case; take the supercharger away and FAR will vary considerably from cylinder to cylinder, whereas with DI you can set FAR quite accurately.
As an aside, WWII vintage technology would just meter the fuel into the cylinders, which is an open-loop approach. You'd still get FAR variations from cylinder to cylinder because although the fuel mass injected would be the same for all cylinders, the airflow would not.
A modern car engine would use an oxygen sensor to tune the fuel mass injected into the cylinder so as to maintain stoichiometry throughout the operating range of the engine; this is vital to the operation of 3-way catalysts. However, this sort of closed-loop approach requires computers, and it is primarily driven by emissions legislation rather than engine performance (power, SFC) considerations.
Without such constraints you'd run the engine leaner to improve SFC, or richer to improve power.
Fuel injection is very useful for CI engines because injection timing can be controlled in order to control the timing of the combustion event; it also allows the pressure profile of the combustion event to be controlled.
Limiting the peak cylinder pressure allows you to make the engine lighter.
With modern engines you can also just stop fuelling cylinders in order to reduce power. This is useful because the turn-down ratio of the injectors is limited if you want good atomisation; poor atomisation leads to reduced combustion efficiency and increased emissions (especially CO and UHC). The alternative is to use multiple injectors per cylinder, but that's a pain.
(If you really really want to then you can build a CI engine with a carburettor, but it's hard work, and it tends to be difficult to control combustion in a satisfactory manner, which hurts thermal efficiency and will tend to cause vibration due to considerable cycle-to-cycle variations in engine behaviour. You'll also find that the smoke limit is set by the richest cylinder, and smoke is a factor then this inevitably limits output.)
Anyway, DI is great for mixture control, but as with all aspects of engine design, it's an option within a tradespace, rather than an unmitigated upside.
If you're mostly interested in operating at a fixed design point then the advantages of DI may easily be outweighed by other options, especially if you're supercharging heavily. OTOH, if you're making a small engine for an economical passenger car today, it's very hard to beat a turbocharged CI engine with DI.
So I'm not suggesting that carburettors are magic; what I'm saying is that DI isn't magic either. Which is better depends upon your priorities, and the job you're trying to make your engine do.
In the very specific example of R-4360C, the primary goal was to get rid of the supercharger so that more useful work could be extracted from the exhaust via the turbocharger. The removal of the supercharger then drove the design towards DI. But it's not reasonable to say that DI is thermodynamically superior; the advantage comes from the improved utilisation of exhaust enthalpy, and DI is just a tool which allows this to be done. Indeed, had it been possible to inject fuel upstream of the turbosupercharger's impeller, superior performance would have been attained.
So DI is analogous to a bridge in this case; the economic benefit comes from the traffic which the bridge carries, rather than the bridge itself, and life would have been easier & cheaper if there had been no need to build a bridge in the first place.
But I must stress again that I'm talking about big aero-engines with high degrees of supercharge here; very different conclusions would be reached if the engine in question was designed to power a car, or even a significantly smaller aeroplane.