In a normal internal combustion engine, speed is controlled by using a throttle valve on the air intake. This valve reduces the amount of fuel/air mixture that is drawn into the engine on each stroke, and thus reduces the amount of power produced. Unfortunately, this valve can waste a very large amount of energy. When cruising at highway speeds--the most common state for most cars--this valve can waste several horsepower or more. Given that cruising doesn't require much power to begin with, this waste represents a significant portion of the power the engine is producing, and thus will impose a severe penalty on highway fuel economy.

Suppose that the throttle valve setting is such that the engine is operating at 10 inches (Hg) of vacuum. This means that the intake pressure will be 2/3 atmosphere. The engine will thus spend about 30 joules per cycle per liter of displacement drawing a vacuum, of which it will recover about 5 joules on the upstroke (since the crank case pressure will be higher than the cylinder pressure until the piston is 1/3 of the way up). At 3,000 RPM, a 3 liter engine drawing 10 inches of vacuum will waste 2.5hp. A larger engine would waste even more.

To avoid this waste, I would propose eliminating the throttle valve(*) and using computer-controlled intake valves to throttle the engine instead. If the computer valves were wide open until the piston was 2/3 of the way down and then closed, the amount of air admitted to the cylinder would be the same as if the valve was open for the whole stroke but a throttle restricted intake pressure to 2/3 atmosphere. Instead of having to operate against a 1/3 atmosphere pressure differential for the entire stroke, however, the piston would only have to draw vacuum during the last portion of its travel. This would take 5 joules of energy per cycle instead of 30, and this 5 joules of energy would be recovered on the following compression stroke.

(*) Actually, the throttle valve should probably be kept as a means of providing engine braking. If someone lets up on the gas for the purpose of slowing down the car, the engine controller should slow the car down. While it would be more "energy efficient" to let the car coast, it would be contrary to the intention of the driver. Additionally, some automotive systems may want to use engine vacuum; either a throttle valve would have to be used to provide vacuum for them, or some other power source would have to be used.
-- supercat, Apr 29 2004

First and Second Law Analyses of a Naturally-Aspirated, Miller Cycle, SI Engine with Late Intake Valve Closure http://me.engin.umi...ations/P1998_02.htm
I'm not the only guy to have thought of this idea (though I wasn't aware of the other when I thought of it) [supercat, Oct 05 2004]

BMW Valvetronic engine http://www.bmwworld...ogy/valvetronic.htm
[squigbobble, Aug 29 2005]

So, tell me how it is that the throttle body/airflow valve is wasteful?
-- bristolz, Apr 29 2004


+! It’s a digital throttle instead of the old analogue. A switching regulator is more efficient than a DC regulator.
-- Fussass, Apr 29 2004


The flow of air from a region of higher pressure to a region of lower pressure will release energy proportional to the volume of air flowing and the difference in pressure between the two regions. A throttle limits the flow of air by using air friction to convert the released energy into heat. This energy is thus wasted--gone--kaput. Actually, it's worse than wasted since engines run more efficiently with cooler intake air than with warmer air.

If you are travelling along at moderate speed and release the gas (and, in the case of a manual-transmission car, downshift) you will notice that the vehicle slows down much more quickly when in gear than when in neutral. When you do this, the kinetic energy of the car gets converted into heat in the throttle.
-- supercat, Apr 29 2004


Upon further consideration, it appears that rather than closing the intake valve early to throttle the engine, it might be more practical to close it late. Although this would cause some energy to be wasted pushing air back out through the intake valve (where, on a vehicle with a multiple of four cylinders, it would immediately flow into the next cylinder starting its intake cycle) it would probably make the intake valve easier to operate since it would only have to resist flow in one direction. Although this would in some regards resemble a Miller cycle engine, I'm not aware of anyone using variable valve timing on a Miller cycle engine as an alternative to throttling.
-- supercat, Apr 29 2004


I hate it when I answer my own question, but once I decided to look at late-closing rather than early-closing, I was eventually able to google a paper on the subject. Turns out there is a factor I hadn't considered with late closing: the fact that some residual burned fuel from the previous cycle will be present in the cylinder during the intake stroke, and in a Miller-cycle engine (delayed closing) will get pushed into the fuel intake where its heat is a bad thing.

Perhaps what's needed is a small supplemental cylinder which would face the main cylinder and whose pistol would go up and down once every two revolutions of the engine so that between the compression and power strokes there would be some volume in the cylinder above the piston but between the exhaust and intake strokes there would be nearly none.
-- supercat, Apr 29 2004


supercat i seem to remember from reading articles on the miller cycle or atleast Mazda's implementation of it that a supercharger was used to provide pressure allowing the intake valve to remain open into the compression cycle.

also as i read your idea the vacuum is above the cylinder shirley this will aid the intake as a vacume above an expanding volume should result in an inflow of air.
-- engineer1, Apr 29 2004


If the throttle is fully closed, how is energy lost across it? I thought engine braking was due to friction in the engine.
Take away the fuel, and an engine is an air pump. A similar example would be a centrifugal fan with inlet damper: the least power (and work done) is when the intake damper is fully closed. That's why big fans are started with the inlet damper closed. When the fan rotates in a partial vacuum, it isn't churning so much air, so friction is less.
But the basic valve throttle idea is workable, I think, and iVTEC engines from Honda come close. I think they close off some cylinders during cruising. So +
-- Ling, Apr 29 2004


engineer1: An engine operating with late intake valve closure will generate less power per unit displacement than one with "normal" valve closure. An engine which operated that way all the time and didn't have a supercharger would simply be anemic for its size. My idea would be to have a naturally-aspirated engine which would operate on a conventional cycle when maximum power was needed, and on a Miller cycle when it wasn't. According to an SAE paper I found this mortning, someone did some simulation tests and found significant efficiency gains.
-- supercat, Apr 29 2004


Ling: Take away the fuel and close the throttle, and the engine will be an air pump which takes air from the lower-pressure intake manifold and pumps it into the higher-pressure atmosphere.

When the engine draws a particularly deep vacuum, the amount of gas being drawn in will decrease, consequently reducing the energy required to draw that vacuum; holding a near perfect vacuum (in the absense of leaks) would require little more energy than holding none. At moderate vacuum levels, however, the pump has to do a lot of work.
-- supercat, Apr 29 2004


The idea that all of this energy would be wasted in the throttle valve is absurd. If that were true, the valve would get white hot and melt. Actually, the loss of efficiency is due to the reduction in compression ratio, which you can’t avoid by using this scheme.
-- ldischler, Apr 29 2004


Forget about closing the intake valve early or late. The way to do this is control both intake and exhaust valves via computer as you indicate. However, the computer would always select valve timings (both opening and closing) which optimize fuel consumption for any desired power output. The computer would also optimize fuel delivery and ignition timing. BTW, there are a few research groups working on this around the world.
-- kcamer, Apr 29 2004


ldischler- Throttle valve pumping losses are real and have been recognized for a long time. The heat is transfered to the intake air. Since the lowered pressure downstream of the throttle valve tends to lower the temerature, one usually wouldn't notice the extra energy unless one was doing careful enthalpy measurements.
-- kcamer, Apr 29 2004


//The idea that all of this energy would be wasted in the throttle valve is absurd. If that were true, the valve would get white hot and melt. Actually, the loss of efficiency is due to the reduction in compression ratio, which you can’t avoid by using this scheme.//

The valve does indeed waste large amounts of energy, but the airflow past it is sufficient to keep it cool.

Reduced compression ratio does cause a further reduction in engine efficiency which is not overcome by this approach, but IIRC according to the SAE paper I found this morning, a small 2.0L engine operating at WOT but using delayed valve closing to reduce output will be about 15% more efficient than one which whose output is reduced by throttling. Not a humungous gain in efficiency, but nothing to sneeze at either.
-- supercat, Apr 29 2004


Arghh! I don't even know where to begin with this one... it's just non-sense.
-- zigness, Apr 29 2004


I hesitate to become involved in this, that said, here we go: Increasing the complexity of a system to reduce perceived inefficiency almost never works. Assuming you could take a VANOS or VTEC type system and adapt it, the gains would completely outweighed by the cost of the system and the inherent unreliability of complex non fixed systems at high RPM, and the assumption that the all powerful, all knowing ECU can handle the increased load without going bonkers. It seems like it would be much easier to use a guillotine type throttle body and a turbocharger and keep valve timeing out of the picture completely.
-- austere_apathy, Apr 29 2004


zigness: It's not nonsense. After posting the idea and looking through about 100 sites on Google, I finally found one that actually discusses the same concept I'd thought of and seems to view it favorably.

austere_apathy: Some vehicles already use throttle-by-wire and computer-operated valves. Using valve timing when possible instead of throttling to control engine power would require merely enhancing the controller software. To be sure, it would make that software a bit more complicated, but probably not outrageously so.
-- supercat, Apr 29 2004


half: //Semi-relevant interjection: A typical diesel engine has no air throttle. Fuel metering controls engine output.//

Gasoline works well only in a fairly narrow range of fuel/air mixtures, but over a fairly wide range of pressures. The usable pressure range for diesel engines is much smaller, and so they have to vary the fuel/air mixture. Engine braking in a car using the throttle is effective because the engine acts as a vacuum pump. Diesels don't draw vacuum by design, and thus some other form of engine braking is required.
-- supercat, Apr 29 2004


I doubt this has been baked with valve timing alone, but by combining variable valve timing and lift it can be done... by BMW.
-- squigbobble, Aug 29 2005


yeah, definitely baked. what you describe is exactly how the valvetronic BMW system works. the problem is that you still need a butterfly valve throttle body as means of a fail safe, so you still get pumping losses from the turbulance it makes, even at WOT. furthermore it takes a very fast computer to do calculations that quickly to controll the valve actuation, but that technology is quickly getting better. note: i said "controll" the valve actuations, not necessarily POWER the valve actuation. for that you would likely need a complex cam system (which should be avoided if at all possible), or you could do it with pnumatic valves (F1 style) or electromagnetic valves (which again, would run into problems with the computers response speed). Very good idea though, may be practicall on expensive european cars that are sold to blokes that never intend on personally opening the bonnet (which is a big market).
-- auricom_mech, Aug 02 2006


auricom_mech,

F1 pneumatic valves are not pneumatically actuated valves but pneumatic 'springs'. They are still essentially passive devices with only perhaps the spring rate dynamically controlled. The cam is still there.
-- djmips, Apr 07 2007



random, halfbakery