Atkinson/Miller/West engine cycle(s)...

[wwest]

What's wrong with this..??

An engine that runs in the highly fuel efficient Atkinson cycle when low torque is required for simply cruising along at constant speed but then transitions into miller cycle when high torque is required, say for acceleration.

The "key" would be a variable intake valve closing delay. Have a smallish DFI engine with a static compression ratio of 12:1 but an expansion ratio of 15-16:1 during the power stroke.

Then use a variable speed positive displacement SuperCharger to boost engine output when acceleration is required. The throttle plate/valve could be eliminated.

As boost rises the intake valve delay would be increased to allow for the dynamic rise in CR due to SC boost.

What do yawl think..??

 

[BrianPetersen]

The current base model Honda Civic engine uses its VTEC system and its variable-valve-timing system to largely implement paragraphs #2 and #3, albeit not with quite as extreme a compression ratio, because the engine still has to be capable of not detonating itself to death when running at full load (on the "power" cam profile that gives normal inlet valve timing).

The BMW Valvetronic system also implements paragraphs #2 and #3 but in a fully variable manner, still subject to the compression ratio limitations.

Fiat also has a system called Multiair that implements #2 and #3, which looks like it will see production in the next couple of years.

All of these systems run at normal cam timing at full load so that the engine can develop normal torque output. No supercharger. The need to run normal cam timing at full load limits the maximum possible compression ratio, and the proposed use of supercharging would limit this even further.

The Toyota and Ford hybrids use the Atkinson cycle with a raised compression ratio and just accept the reduced power output - the hybrid system is used to make up for it.

Delaying intake valve closing (to reduce the amount of air going into the cylinder to reduce dynamic CR) isn't constructive if you are simultaneously trying to increase the amount of air by supercharging. You might as well not delay intake valve closing and not supercharge as much ... same outcome (and less power demand to run the supercharger).

I do not know of any current Atkinson-cycle implementations that also use direct-injection. It could certainly be done, but I suppose every additional system adds more cost and has diminishing returns, and it gets to a point where (at least currently) it isn't worthwhile. Emission regulations have dictated that direct-injection spark-ignition engines run at stoichiometric anyway, which defeats much of the promised efficiency improvement.

If the "supercharging" is replaced with "turbocharging" then some engine downsizing could be done. VW/Audi are doing this in combination with direct-injection, and it looks like BMW may be heading down that path also. These engines require the use of premium high-octane fuel.

 

[TDIMeister]

The slider-crank mechanism does not allow for a large dissimilarity of the upward- and downward piston motions that would result in different "static" compression- and expansion ratios, respectively.

You my be able to get a half point increase in actual static ratios by offsetting the cylinders relative to the crank centerline as the Toyota Prius engine does, but the rest has to come from valve timing.

First of all, there's lots of confusion and a lack of convention in literature as to the definitions of the Miller- and Atkinson cycles. Some define the distinction between the two as applying to force-induction and naturally-aspirated engines, respectively, with no distinction of late- or early- intake valve closure; others say Miller = late IVC resulting in push-back of some charge back in to the inlet manifold, while Atkinson = early IVC before BDC resulting in a partial expansion of the charge before the compression stroke takes place as usual. I tend to prefer and use the latter definition.

For low engine speeds and part-load, I would use early IVC and at higher speeds and high-load, use late IVC with charge boosting, thereby also taking advantage of gas dynamic effects to improve volumetric efficiency or at least offset the loss that results from charge push-back into the intake manifold that is associated with LIVC.

The static geometric compression ratio would remain some high value like around 10.5:1 (in a forced induction engine). More is feasible in a forced-aspirated GDI engine, but this limits the attainable knock-limited BMEP and requires more full-load AFR enrichening that negates the fuel economy benefits of GDI. This is precisely the subject I am currently undertaking active research: 2-stage turbocharged GDI engine that achieves a constant 26 bar BMEP from 1500-5500 RPM...

 

[ivymike]

wwest, I must be missing something - when you say "static compression ratio," you're referring to the quantity calculated by taking the ratio of BDC volume over TDC volume? If yes, then how could a conventional piston engine have static compression ratio of 12:1 but an expansion ratio of 15-16:1 during the power stroke? Would static compression ratio not be the largest compression ratio, with various valve timing schemes resulting in reduced values of effective compression ratios?

 

[wwest]

The Mazda Millennia S used the modern day Atkinson cycle technique, delayed intake valve closing, to increase efficiency vs power and then added an SC, Miller cycle, to help make up the difference.

The Mazda used a positive displacement SC downstream of the throttle plate and ALWAYS provided boost when the throttle was open/cracked.

The same basic idea.

In order to turbocharge an engine its base efficiency, non-boost operation, must be sacrificed. The use of an variable speed SC allows efficient or BOOSTED operation throughout the operational engine range.

Using the Prius e/CVT technique if only atmospheric pressure is required the SC would simply idle along with very little engine drive or electric drive needed.

But yes, the "static" compression ratio by strict definition would need to be ~15-16:1 with DIVC bringing it down to 12:1.

Basically, except for DFI, we would have an engine equivalent to that in the Prius for simply cruising along and transition to one equivalent to the Madza Millenia S when power production became the mode.

 

[BrianPetersen]

Why would turbocharging require off-boost efficiency to be sacrificed any further than with supercharging?

The piston-and-valves-and-cams piece of the engine only sees "intake manifold pressure" in either case, it doesn't care how it was generated, and the turbocharger requires less parasitic loss from the crankshaft in order to operate it.

It's true that most traditional turbocharged production applications have been optimized for power rather than economy, but VW's TSI engines, Ford's Ecoboost engines, and GM's (hopefully) upcoming 1.4 litre turbo engine in the new Cruze are the other way 'round, and use turbocharging to allow the engine to be downsized.

 

[wwest]

"why would turbocharging..."

Because turbochaging require one hell of a lot of WASTE energy entering the exhaust manifold and that does not happen at cruise speed engine torque levels.

Once BOOST "arrives" the cylinder charge will rise accordingly, so prior to boost arriving the CR must be inordinately low, lower than in a even a non-boosted engine.

Using an HSD e/CVT SuperCharging can provide boost, or not, all the way from idle RPM up to the rev limiter.

TurboCharging only works when there is enough energy left in the exhaust stack to move the turbine. Use the Atkinson cycle, or Miller, and more of the energy of combustion is used up in pushing the piston downward with NOTHING left over for spinning a turbo.

Ford is done nothing other than hiding the "wizard" behind the curtain. The DFI CR advantage could be put to greater use.

 

[TDIMeister]

Current practise of typical intake valve closing timing well past BDC in the order of 40-70° already reduces the effective compression ratio well below that of the geometric CR. In literature it has been found that regardless of geometric compression ratio, a large scatterband of naturally-aspirated, pump gas-fuelled engines (non-DI, no otherwise unconventional combustion chamber- or piston bowl shapes or design measures) converge to a full-load knock-limited effective CR of around 9.5:1 - 10:1. Forced-inducted Miller-cycle HD-Diesel engines close the intake valves over 100° ABDC.

Current practise of typical EVO timings are also in the order of 40-70° BBDC. You could open the exhaust valve very near BDC, but whatever you might gain in indicated efficiency will be lost in increased pumping work during the exhaust stroke and increased residual gas mass fraction for the next cycle hindering combustion.

 

[wwest]

The basic idea is to VARY the delay of the intake valve closing so as to provide the HIGH FUEL EFFICIENCY of the Atkinson cycle and then transition into the Miller cycle mode when POWER output is required.

Starting with a mechanical CR of ~15-16:1 but in Atkinson mode using delayed valve closing to make the "effective" CR about 12:1 (DFI). Then as boost rises the valve closing delay would be extended as needed to prevent knock/ping.

 

[TDIMeister]

I hear you. Already being done in research engines (here at my workplace) using electromagnetic valvetrains). Nothing all that new.

 

[SomptingGuy]

Rumour has it that if your valve control is flexible enough, you can flip between 2-stroke and 4-stroke operation too. Fancy that!

- Steve

 

[TDIMeister]

Yup, being done at Ricardo (a competitor) among other places.

 

[wwest]

Yeah, SuperCharging a 2-stroke, now that would be real breakthrough engineering.

 

[BigVlad]

I have been following this thread with great interest. About 20 years ago I was involved in the building and testing of an A-Series BMC engine with a mechanical or geometrical CR of about 18:1. The manifold pressure was limited by various means to about 14 inches of Hg so that the pressure in the combustion chamber was about that of a engine with a CR of 9:1 or so - that is; not high enough to cause detonation. The whole idea was that the combustion gases were expanded much more than normal thus raising the thermal efficiency. (I don't think I had even heard the words "Atkinson Cycle" at the time). To cut a long story short, it all worked very much as predicted. The exhaust manifold temp. was markedly lower and the exhaust note much quieter (and other effects) showing that much more energy had been extracted from the combustion gases.
But to get back to Wwest's proposition: - you can't even run anywhere near full atmospheric manifold pressure with an Atkinson engine let alone above atmospheric as with supercharging. Really I think this is what Ivymike pointed out, and there is no way around it.
I think Miller cycle engines with their superchargers have confused the issue - I don't think this is true supercharging in the normal sense. The pressure in the combustion chamber after compression is the critical factor no matter what type of cycle is being used or how you arrive this point.
So no - I don't think you can change from Atkinson Cycle to Miller Cycle.

 

[SomptingGuy]

You have 3 choices:

1) Hand waving (very popular).
2) Hardware (expensive).
3) Simulation.

- Steve

 

[wwest]

"..I don't think you can change..."

But, why not...??

4 cylinder Atkinson cycle engines exist, work, and are in current production and use.

Miller cycle engines exist, work, and are in current use.

DFI engines exist, work, and are in current production and use.

Engines with VVT and VVT-i exist, work, and are in current production and use.

Apparently switching a 4 cylinder engine between the two, Atkinson vs Miller, might involve something as simple as an blower type SC with a clutch.

Aftermarket addition of an SC or Turbo to an otherwise off-the-shelf, standard, production engine is a fairly common event.

Obviously one could optimize, increase the level of boost possible, with the aftermarket addition of an SC or blower via reduction of the engine's CR. But that would result, also obviously, in a lower FE "off-boost".

All I'm suggesting is an explansion of teh use of VVT to accomplish a dynamic change in the CR as/when the SC begins producing boost.

 

[patprimmer]

The question has already been answered. What part of the answer don't you get.

The part about supercharging a high compression engine causes detonation.

The part that reducing compression by late intake valve closing reduces VE and power.

The part that by correcting that with supercharging only gets back the VE that you deliberately lost by late valve closing, but still suffers the parasitic losses of of driving a supercharger.

The part that no matter how you achieve the cylinder pressure, you can only tolerate a certain amount before it detonates.



Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[BigVlad]

Wwest: I think the main problem is that the Miller Cycle doesn't make what would appear to be commonsense. Apparently if the charge is compressed by a combination of supercharger and piston it takes less energy than if the charge is compressed by the piston alone. And this is the only real source of efficiency gain with the Miller Cycle. To me it is not obvious why this should be so but apparently it is true. Commonsense would suggest that to get a certain final charge pressure all methods would take the same amount of energy - that is; it would be "independent of path" (as they say in the classics).
I think a lot of magazine writers etc. have confused the issue by giving the false impression that the Miller Cycle allows a high mechanical CR (and expansion ratio) to be used by having LIVC to to reduce charge pressure and then regain all that has been lost by supercharging. This is really not what the Miller Cycle is.
I think the most surprising thing is that Ralph Miller realized that the combination of blower and piston took less energy than piston alone to get the final pressure.

 

[wwest]

There are two parts to the issue. First of all the rate of travel of the piston is not linear. As the crank leaves the BDC position the amount of angular motion is quite high but the level of piston motion is relatively low. The DIVC technique removes the requirement for any level of compression during this early part of the compression stroke.

The second part of the equation has to do with the fact that the air compressed by the SC can be COOLED post-compression.

So the charge reaching the cylinder will not be as hot as it would be were it compressed solely by the piston travel.

 

[BrianPetersen]

I do not recall whether the Mazda engine actually had an intercooler.

I think the idea with the Miller cycle is that during light to moderate load operation (95+% of most people's driving) the supercharger does not operate (and does not cause any meaningful power losses) so the engine can operate with reduced throttling losses. At full load, the supercharger makes up for the volumetric-efficiency disadvantage (it is NOT a high-pressure-ratio supercharger if I remember right - the power output of that engine was nothing special) to allow the engine to have a normal power output. But, by stuffing extra air into the engine it's still subject to limitations imposed by detonation, so the compression ratio can't be raised above normal. At full load, I would expect the power demand to run the supercharger to make the whole package slightly LESS efficient than a conventional unsupercharged engine. But, most people's driving spends so little time at full load that it really doesn't matter too much to the actual fuel consumption in normal driving.

The Civic engine (and the BMW Valvetronic) reaches a normal power output by varying the intake valve timing and lift but again the compression ratio has to remain "normal" because of detonation limitations at full load.

The Prius engine and the Atkinson version of the Ford 2.5 engine (used in Escape and Fusion hybrids) have delayed intake closing with a raised compression ratio and no supercharging and they deal with the reduced power output of the engine by using the hybrid system to make up for it.

Higher pressure supercharging whether "super" or "turbo" can be used by using very good intercooling without sacrificing the compression ratio too much.

Regarding not having enough exhaust energy left to run a turbo ... fat chance. In a normal application if the engine is to have anywhere near an acceptable power output, you won't be able to have enough difference between the effective compression and expansion ratios to use up ALL of the extra energy during the power stroke. Diesel engines already use up more of the energy to drive the piston than gasoline engines do, and automotive applications are universally turbocharged and intercooled (not mechanically supercharged) nowadays. It requires different calibration of the turbine side than a gasoline-application turbo does, that's all.