Supercharger For Extracting Exhaust Gas.

[Rolls]

The whole idea is not worth the hassle IMO, because of 1 fundamental reason. You can not drop pressure below vacuum, i.e. there is no such thing as negative pressure. You're trying to increase pressure differential across the engine. In the best possible scenario there will be 1-0 pressure diff - atmosperic at the intake and 0 past the engine, and density of air trapped in chamber will be somewhere in between, i.e. below atmosperic.

Yeah it pretty much fails because of this.

[GTSBoy]

...vacuum in chamber prior to intake event wouldn't make cylinder filling good, because it's gas we're dealing with here, so when your engine opens its intake valves, gas just loses density, it won't "rush in".On n/a engines with some overlap, vacuum does not do the major part of cylinder filling. Gas momentum does. This momentum is created because some mass of gas gets moving - it moves because of pressure differential betw. chamber full of hot burnt gasses and atmosphere. You basically suggest to lower the past-valve pressure below atmospheric, make pressure differential greater and thus increase gas momenum. Good thinking, but it comes with the penalty. On a well-engineered exhaust manifolds exhaust events are arranged in such a way that when some cylinder opens its valves and commenses exhaust stroke, there is already vacuum in the manifold - it was created by exhaust stroke of the previous cylinder and consumes no engine power. Your idea is the same in that it lowers exh. manifold pressure too - but unlike "extractors" it uses engine power.

Well, that's not entirely true. A decent vacuum in the cylinder will in fact cause any air that is in the inlet tract to suddenly accelarate into the cylinder when the inlet valve opens. Gas doesn't just "lose density". Any such pressure waves move through a gas at the local sonic velocity. In the high pressure gas (assuming it's boosted) in the inlet tract the velocity of sound is pretty high, but in a good solid vacuum the sonic velocity is quite low. In fact the speed of sound becomes quite complicated as you get towards a hard vacuum. Anyway, ignoring that complication, if you have managed to reduce the static pressure in the cylinder to some value below what it would normally be in any other engine then you will increase the volumetric and mass flow rate of air into the cylinder - at least in the initial phase of cylinder filling. This would probably be true even if you lost some of the scavenging "momentum" effect of valve overlap.

Remember, ignoring the exact opening an closing times of the inlet and exhaust valves for a moment....the majority of the driving force for cylinder filling is in fact the piston falling down the bore. The main impact of the momentum of the inlet air flow continuing to fill the cylinder occurs after the piston has reached BDC and has started rising again for the compression stroke. Because the air is still flowing towards the cylinder, it continues to fill even against the rising pressure in the cylinder and will do so until either the inlet valve closes or the pressure gets too high and starts to push mixture back out of the inlet valve. Anyway, the momentum that helps to fill the cylinder only occurs because the piston dropping (plus whatever positive boost pressure is in the plenum) causes the flow to start happening. So any extra vacuum in the cylinder before you start will certainly help.

In an engine with overlap, the exhaust scavenging effect in the cylinder occurs at the tail end of the exhaust stroke and helps to start the inlet flow in the same way as the possible in cylinder vacuum that we're talking about. The question comes down to which would have the bigger effect. Remember that in a conventional engine the exhaust gases in the cylinder at that point are still at positive pressure compared to the exhaust port, and in a turbo engine it is quite possible that the exhaust port has quite a high pressure above atmospheric. If this were replaced with a residual gas that were at somewhat lower that atomospheric pressure then it would rather drastically change the way we think about charge contamination, timing vs detonation, etc etc etc. Lots of things going on. Lots of possibilities for it to be better. Just that it's unlikely that we could get a significant decrease to that sort of pressure on a turbo engine because we'd still have a reasonably high exhaust manifold pressure - certainly at least in the same order of magnitude as the boost level we're running.

By the way, there is no need to "suck the intake charge into the cylinder via the exhaust" because it significanlty reduces engine fuel efficiency, and SFC rises.

Except that this is exactly what you claim happens (and does in fact happen) with the momentum scavenging effect. It's just that its done with dynamic pressure instead of static pressure.

The whole idea is not worth the hassle IMO, because of 1 fundamental reason. You can not drop pressure below vacuum, i.e. there is no such thing as negative pressure.

Well, yes there is. If you define zero pressure to be at 1 atmosphere, as is the case with the commonly used gauge pressure system, then any pressure between that zero and a total vacuum is a negative pressure. Any time you read "negative pressure" in Rolls' posts, you just have to remember that that is what he means, not some imaginary pressure that is below absolute vacuum. It is merely a matter of definition.

You're trying to increase pressure differential across the engine. In the best possible scenario there will be 1-0 pressure diff - atmosperic at the intake and 0 past the engine, and density of air trapped in chamber will be somewhere in between, i.e. below atmosperic.

On the other hand you can make the same pressure differential by adding 1 bar of boost to intake side. Same diff, this time 2-1 though, and density of air again will be somewhere in between, but now between 2 and 1, above atmosperic. Bonus - more airmass - bigger bang! And another bonus - boost pressure does not have to be 1 bar - it can be 2, 3, sky is the limit. So with n/a engine you should just add a supercharger.

A quick word to the wise. The driving force may be the same between a 1bar to 0 bar pressure difference and a 2 bar to 1 bar pressure difference and a 3 bar to 2 bar pressure difference - but you will not get the same flow through a given nozzle size in all those cases. The downstream pressure has a significant influence on the total flow rate for a given pressure drop.

I'm not trying to disagree with you - just trying to clean up your terminology a bit.

[nisskid]

or possibly turn your turbo into an electrically driven motor, which free runs in normal driving, but you hit a button which adds power and eliminates lag, but then allows exhaust gases to take over the driving once they start overpowering the electric motor? bottom end of a supercharger, top end of the turbo without running a pulley off the engine to zap power. because it's only used in short periods, it would be renewable power after a bit of driving afterwards.

kinda like a twin charger i guess.

haha love threads like these.

[Legionnaire]

Well, that's not entirely true. A decent vacuum in the cylinder will in fact cause any air that is in the inlet tract to suddenly accelarate into the cylinder when the inlet valve opens. Gas doesn't just "lose density". Any such pressure waves move through a gas at the local sonic velocity. In the high pressure gas (assuming it's boosted) in the inlet tract the velocity of sound is pretty high, but in a good solid vacuum the sonic velocity is quite low. In fact the speed of sound becomes quite complicated as you get towards a hard vacuum. Anyway, ignoring that complication, if you have managed to reduce the static pressure in the cylinder to some value below what it would normally be in any other engine then you will increase the volumetric and mass flow rate of air into the cylinder - at least in the initial phase of cylinder filling. This would probably be true even if you lost some of the scavenging "momentum" effect of valve overlap.

Remember, ignoring the exact opening an closing times of the inlet and exhaust valves for a moment....the majority of the driving force for cylinder filling is in fact the piston falling down the bore. The main impact of the momentum of the inlet air flow continuing to fill the cylinder occurs after the piston has reached BDC and has started rising again for the compression stroke. Because the air is still flowing towards the cylinder, it continues to fill even against the rising pressure in the cylinder and will do so until either the inlet valve closes or the pressure gets too high and starts to push mixture back out of the inlet valve. Anyway, the momentum that helps to fill the cylinder only occurs because the piston dropping (plus whatever positive boost pressure is in the plenum) causes the flow to start happening. So any extra vacuum in the cylinder before you start will certainly help.

In an engine with overlap, the exhaust scavenging effect in the cylinder occurs at the tail end of the exhaust stroke and helps to start the inlet flow in the same way as the possible in cylinder vacuum that we're talking about. The question comes down to which would have the bigger effect. Remember that in a conventional engine the exhaust gases in the cylinder at that point are still at positive pressure compared to the exhaust port, and in a turbo engine it is quite possible that the exhaust port has quite a high pressure above atmospheric. If this were replaced with a residual gas that were at somewhat lower that atomospheric pressure then it would rather drastically change the way we think about charge contamination, timing vs detonation, etc etc etc. Lots of things going on. Lots of possibilities for it to be better. Just that it's unlikely that we could get a significant decrease to that sort of pressure on a turbo engine because we'd still have a reasonably high exhaust manifold pressure - certainly at least in the same order of magnitude as the boost level we're running.

Actually I have a lot to say about this, but it would be seriously OT here + I'm a bit lazy ATM and don't feel like typing everything I have to say.

I'm not saying anything about valve events. I agree with you, valve events and valve timing play major part in the equation. Pressure and vacuum waves that travel through intake and exhaust tracts depend on LOTS of parameters - temperature, RPM, port geometry, etc. Usually we have some of that parameters fixed - say, port and exhaust runner cross-section - that's why variable valve event systems, like vanos and variocam, are so beneficial. Ultimately of course we want infinitely variable valvetrain - with independently controllable opening and closing points and lifts. There are some very interesting experiments and studies about this, done by Audi and Fiat, I hope to see the results of those on production engines soon. For the time being though we have to deal with much less flexibility.

Generally, we have 3 problems with fixed cam profiles: exhaust reversion, intake reversion and overscavenging. With the second, we have to carefully choose cam geometry and some sort of VVT is desirable. The exhaust scavenger virtually eliminates the first problem, but makes the third one worse. Whether advantages of the former oveweight the drawbacks of the latter is to be tested.

The whole superscanvenger idea may benefit from relocation of the valves in the chamber and its reshaping.

An interesting thing to investigate would be power expenditure of the engine throughout non-power strokes for superscavenged vs. exhaust-tuned n/a engine.

Except that this is exactly what you claim happens (and does in fact happen) with the momentum scavenging effect. It's just that its done with dynamic pressure instead of static pressure.

It obviously happens with either system - it does not matter how exactly exhaust vacuum is created, it creates overscavenging and wastes fuel unless correct exhaust valve closing point is utilised. And that point is not the same for various loads and engine speeds.

Well, yes there is. If you define zero pressure to be at 1 atmosphere, as is the case with the commonly used gauge pressure system, then any pressure between that zero and a total vacuum is a negative pressure. Any time you read "negative pressure" in Rolls' posts, you just have to remember that that is what he means, not some imaginary pressure that is below absolute vacuum. It is merely a matter of definition.

And I never said he was using some imaginary negative pressure. If you set a zero to be at 1 atmosphere, then the minimum pressure will be at -1.

The point of that statement was that pressure can not be lower than vacuum, i.e. there exists a lower limit for pressure.

I was using absolute pressure in my post, hence no negative pressures.

A quick word to the wise. The driving force may be the same between a 1bar to 0 bar pressure difference and a 2 bar to 1 bar pressure difference and a 3 bar to 2 bar pressure difference - but you will not get the same flow through a given nozzle size in all those cases. The downstream pressure has a significant influence on the total flow rate for a given pressure drop.

Good to know, thank you. In what case the flow will be greater?

I'm not trying to disagree with you - just trying to clean up your terminology a bit.

I see that you're not disagreeing with me, admit my statements to be not 100% scientifically correct and appreciate your input with terminolody cleaninig. I was trying to express my ideas in a simple way, without maths and physics, as not everyone is into reading phormulas on an internet forum, and was trying to avoid over-complication of the tread with details - there are books, lots of them, and not thin books at that, that discribe different aspects of the subject.

[GTSBoy]

Good to know, thank you. In what case the flow will be greater?

Higher downstream pressure with same pressure drop gives greater volumetric (and mass) flow rate. This is because the downstream density is higher with higher downstream pressure, so the amount of mass that can flow through the nozzle at a given pressure drop (and hence velocity) will increase with density.

Just taken these numbers from a calculator spreadsheet that I use for designing nozzles for gas burners. So the flow rates are m3/h of natural gas (with the conditions for measuring the m3/h fixed to 1 atm pressure and 0°C, so that the volume flow rates quoted are effectively mass rates because the gas density is the same for all quote flow rates).

For a particular nozzle (78mm diameter with a given coefficiant of discharge of 0.8 - not that it matters) with the downstream pressure set to 50 kPa (abs) and upstream pressure at 150 kPa (abs), the gas flow rate is 4977 m3/h.

Same nozzle, downstream pressure 100 kPa (abs), upstream pressure 200 kPa (abs), gas flow rate is 6638 m3/h.

Same nozzle again, downstream pressure 200 kPa (abs), upstream pressure 300 kPa (abs), gas flow rate is 13275 m3/h.

Note that this is steady state flow, and that engines are of course a fairly dynamic flow situation with continuously changing downstream pressure and density throughout the intake stroke, so only take the broad outlines of this post forward.

Edited September 12, 2011 by GTSBoy

[Legionnaire]

Interesting bit of info nevertheless, thank you.