Atkinson Cycle engines on non-hybrid vehicles

Status
Not open for further replies.
Joined
Jun 27, 2017
Messages
2,014
Location
OK
Just wanted opinions on this question I've been having about the Atkinson Cycle engine. While I know a bit about it and it's definitely not new technology, I had never really heard of it used very often until manufacturers started putting it in their hybrids. It seemed like a perfect fit, an engine that suffers from a lack of low end power and a battery to help make up for that. Now, however, I'm seeing it put into a couple of other applications, including non-hybrid vehicles. For example, the Elantra in my sig has an Atkinson Cycle engine, as well as 2016+ Tacomas.

What I'm wondering is if there's any reason to think that these engines may not experience the same longevity as the standard, Otto cycle engine? Having never seen an Atkinson Cycle engine outside of a hybrid, I really have no idea what to expect here. I mean, there are plenty of 200k+ mile Prius engines out there which are Atkinson Cycle, but the life of a hybrid engine is a pretty easy one.
 
If I'm remembering correctly, non-hybrid Atkinson engines use their VVT systems to simulate Atkinson cycle and can switch to the more powerful Otto cycle on demand.
 
That's an interesting point Vette. I hadn't heard that, but that would make sense.
 
Originally Posted By: VetteElite
If I'm remembering correctly, non-hybrid Atkinson engines use their VVT systems to simulate Atkinson cycle and can switch to the more powerful Otto cycle on demand.


I'm not sure about the Elantra, but I know this is the case with the Tacoma.
 
Originally Posted By: VetteElite
If I'm remembering correctly, non-hybrid Atkinson engines use their VVT systems to simulate Atkinson cycle and can switch to the more powerful Otto cycle on demand.

Yup, SkyActiv from Mazda can adjust intake and exhaust valves to achieve Otto or Atkinson. Miller cycle engines, aka Mazda Millennia, add supercharging to make up for the lack of low end breathing.
 
Originally Posted By: glock19
Originally Posted By: VetteElite
If I'm remembering correctly, non-hybrid Atkinson engines use their VVT systems to simulate Atkinson cycle and can switch to the more powerful Otto cycle on demand.


I'm not sure about the Elantra, but I know this is the case with the Tacoma.

Here's what I've found on the Elantra:

"For 2017, the new Elantra receives two all-new powertrains designed for improved fuel efficiency and everyday drivability performance. The standard engine available on the base SE and Limited trim is a 2.0-liter Nu MPI Atkinson four-cylinder engine producing a peak 147 horsepower at 6,200 rpm and 132 lb. ft. of torque at 4,500 rpm (estimated). This Atkinson cycle-type engine reduces pumping loss by delaying the close timing of the intake valve in the compression stroke, therefore maximizing the expansion ratio – it is the only Atkinson cycle engine to be combined with multi-port injection in the compact class. This greater expansion ratio is made more efficient by allowing additional energy to be produced. Furthermore, this engine features a high compression ratio of 12.5.

Other enhancements include intermediate valve cam phasing that increases the operational range of the intake valves and helps to reduce pumping loss."

Link
 
Thought the real McCoy (origin Elijah) had a variable stroke - not VVT ... some (good) modern day technology around but does the name hold true?
 
I did a google search for Atkinson Cycle engine . Read several articles & know a tiny bit more than I did about a Atkinson Cycle engine .

OK , the spark is 100% under control of the computer .

Fuel is 100% under control of the computer .

In an automatic transmission , the gearing is 100% under the control of the computer .

Valves , through the cam phaser , are under partial computer control . If you had the valves actuated by electric / electronic solenoids , you would have them under 100% control of the computer .

You could really get arrive with the software . If you could implememt AI , yhe software / computer could learn to tune the engine for economy , low emissions , torque , max horse power . Or , maybe things I have not thought of .

Wonder is anyone is researching this idea ?

Best of luck to you , :)
 
Originally Posted By: VetteElite
If I'm remembering correctly, non-hybrid Atkinson engines use their VVT systems to simulate Atkinson cycle and can switch to the more powerful Otto cycle on demand.


The R18 engine in 2006-2015 Civic's have a simulated Atkinson cycle. On the economy cam profile the intake valve stays open longer and lets part of the air/fuel mixture back out of the combustion chamber. On the "power" cam the intake valves close sooner. This also reduces pumping loses.
 
Last edited:
Originally Posted By: skyactiv
Mazda uses what they call the Mazda Miller cycle for obtaining an Atkinson cycle on the Skyactiv engines.

Its not a Miller Cycle on the Skyactiv engines, unless they have a supercharger, which I don't think they do. Miller Cycles require superchargers by definition.

Nobody runs Atkinson Cycle using the bizarre crankshaft that the original patent in 1887 used. Its always done with late intake valve closing.
180px-Atkinson_Engine_with_Intake.gif


As for the original question, the only durability issue I can think of is having the air-fuel mixture washing ('burping back') over the intake valves may cause extra intake valve stem deposits. Potentially you would think.

One more note: The high compression ratios in Atkinson Cycle engines are a little decieving, since late intake valve closing reduces the effective compression stage.
What is nice is that a good VVT system can come and go in and out of Atkinson as conditions dictate.
 
Last edited:
Originally Posted By: WyrTwister
You could really get arrive with the software . If you could implememt AI , yhe software / computer could learn to tune the engine for economy , low emissions , torque , max horse power . Or , maybe things I have not thought of . Wonder is anyone is researching this idea ?

Yes engine makers have been using Artificial Intelligence (AI) algorithms to optimize engine operation. Much of this is done off-line, in the lab or field testing, where the engine maps are adjusted to reach optimality goals (performance, MPG, emissions, you know...). Some trimming and adjustments are allowed as the engine is used in real-time, although the algorithms are very limited in authority so they don't diverge in new unknown directions. Actually the 'Integral' (I) part of traditional PID control systems is a trimming adjustment used for the last 30 years in engine control systems, a basic 'learning' adjustment.
 
I think one reason that dedicated Atkinson Cycle engines are only used in hybrids is that their specific output (i.e. horsepower per engine displacement) is low. In a hybrid this is fine because the electric motor makes up the difference, but it would require an unacceptably large engine in a non-hybrid.
 
Originally Posted By: NateDN10
I think one reason that dedicated Atkinson Cycle engines are only used in hybrids is that their specific output (i.e. horsepower per engine displacement) is low. In a hybrid this is fine because the electric motor makes up the difference, but it would require an unacceptably large engine in a non-hybrid.
I'm not sure what you mean by "dedicated" Atkinson when VVT allows it to come and go on the fly, and all operate that way. It may be true hybrid engines do Atkinson more, as a percentage of total running time. All are part-time Atkinson anyway. When we say 'Atkinson', we simply mean those moments when the intake valve closes late, thats all. Nothing more complicated than that. Any engine with a VVT can claim it if it does late valve closing!!
 
AI has been the most abused word in the last 5 years. The supposedly official definition is to find match using linear algebra (to get the closest distance between 2 points) that is not yet determined and "guess" which one of the multiple choice is the answer.

Fuel map is pre determined by trial and error, so technically it is a calibration, and then looking up the calibration at run time. It is technically a brute force pre-calculated approach, not "intelligent", but better if you can test everything out beforehand than let the computer guess by trial and error when you are driving.
 
Last edited:
Originally Posted By: WyrTwister
I did a google search for Atkinson Cycle engine . Read several articles & know a tiny bit more than I did about a Atkinson Cycle engine .

OK , the spark is 100% under control of the computer .

Fuel is 100% under control of the computer .

In an automatic transmission , the gearing is 100% under the control of the computer .

Valves , through the cam phaser , are under partial computer control . If you had the valves actuated by electric / electronic solenoids , you would have them under 100% control of the computer .

You could really get arrive with the software . If you could implememt AI , yhe software / computer could learn to tune the engine for economy , low emissions , torque , max horse power . Or , maybe things I have not thought of .

Wonder is anyone is researching this idea ?

Best of luck to you , :)


Many people and companies have been researching fully active valvetrains for decades. But nobody seems to have reached the point where the efficiency benefit of the fancy valve system overcomes the additional parasitic loss. If they had, such systems would be on the market now. I interviewed for a job at Sturman Engine Systems back in the year 2000. They were working on a fully variable hydraulic valve actuation system, and wanted me to design a V8 engine to demonstrate the technology. I became less excited about the technology when they showed me a room full of destroyed test rigs where electrical glitches had caused the ECU to open valves at the wrong time, destroying themselves against intruding pistons. A "stupid" mechanical camshaft doesn't have that problem. Camshafts are really pretty elegant devices; most of the energy they take from the engine when opening valves is returned as the springs close the valves. In a heavy duty diesel running at full power, it only takes ~2hp to run the camshaft.
 
Originally Posted By: A_Harman
Many people and companies have been researching fully active valvetrains for decades. But nobody seems to have reached the point where the efficiency benefit of the fancy valve system overcomes the additional parasitic loss. If they had, such systems would be on the market now. I interviewed for a job at Sturman Engine Systems back in the year 2000. They were working on a fully variable hydraulic valve actuation system, and wanted me to design a V8 engine to demonstrate the technology. I became less excited about the technology when they showed me a room full of destroyed test rigs where electrical glitches had caused the ECU to open valves at the wrong time, destroying themselves against intruding pistons. A "stupid" mechanical camshaft doesn't have that problem. Camshafts are really pretty elegant devices; most of the energy they take from the engine when opening valves is returned as the springs close the valves. In a heavy duty diesel running at full power, it only takes ~2hp to run the camshaft.


I'm not seeing how the parisitic power loss of a Sturman-like system was the problem. And then you make a different point that the commands to lift and close aren't reliable enough, which is a different issue than parasitic losses. I've always thought reliability was the problem with electrically actuated valves, not energy efficiency of the actuation itself. Remember a spring takes more energy to compress than no spring at all. Energy return does happen when it decompresses, although that initial expenditure of power is not there if the spring is missing. Also, cams have friction, which is a loss.

Originally Posted By: PandaBear
AI has been the most abused word in the last 5 years. The supposedly official definition is to find match using linear algebra (to get the closest distance between 2 points) that is not yet determined and "guess" which one of the multiple choice is the answer.
Not true at all. Thats not the "official" definition of AI. AI is a very broad term which can mean any of about a thousand different algorithms and a many custom heuristics, not just linear algebra. AI isn't an abused term, but it is very broad and can mean any adaptive method is being used.

Originally Posted By: PandaBear
Fuel map is pre determined by trial and error, so technically it is a calibration, and then looking up the calibration at run time. It is technically a brute force pre-calculated approach, not "intelligent", but better if you can test everything out beforehand than let the computer guess by trial and error when you are driving.
Good point. A big problem with many AI algorithms is that we can't guarantee convergence to a good solution in all situations. Some AI algorithms work better than others, and success of any algorithm depends on what the application is.
 
Originally Posted By: oil_film_movies
Originally Posted By: A_Harman
Many people and companies have been researching fully active valvetrains for decades. But nobody seems to have reached the point where the efficiency benefit of the fancy valve system overcomes the additional parasitic loss. If they had, such systems would be on the market now. I interviewed for a job at Sturman Engine Systems back in the year 2000. They were working on a fully variable hydraulic valve actuation system, and wanted me to design a V8 engine to demonstrate the technology. I became less excited about the technology when they showed me a room full of destroyed test rigs where electrical glitches had caused the ECU to open valves at the wrong time, destroying themselves against intruding pistons. A "stupid" mechanical camshaft doesn't have that problem. Camshafts are really pretty elegant devices; most of the energy they take from the engine when opening valves is returned as the springs close the valves. In a heavy duty diesel running at full power, it only takes ~2hp to run the camshaft.


I'm not seeing how the parisitic power loss of a Sturman-like system was the problem. And then you make a different point that the commands to lift and close aren't reliable enough, which is a different issue than parasitic losses. I've always thought reliability was the problem with electrically actuated valves, not energy efficiency of the actuation itself. Remember a spring takes more energy to compress than no spring at all. Energy return does happen when it decompresses, although that initial expenditure of power is not there if the spring is missing. Also, cams have friction, which is a loss.


And the story continues...

Then a few years later, I was an advanced valve train system engineer working at Eaton Corp. Round about 2005, they bought the rights to Lotus Engineering's Active Valve Train hydraulic fully variable system, and set to work trying to commercialize it at their tech center in Farmington Hills. Right from the start, they identified reducing parasitic loss of the hydraulic pump as a major enabler to making AVT viable for use in mass-production applications. Eaton is a good company to attempt this work, as they own Vickers Hydraulics, and would presumably be able to design an efficient variable displacement pump. Now it's 2017, and Eaton has not gotten AVT into production.

So there you have it, I have talked to engineers at two major developers of the technology, and they both said that parasitic loss of the hydraulic system stands in the way of major applications. From the OEM's standpoint, why go through all the trouble of putting the expensive new system into production if it doesn't provide a net improvement in efficiency?

Yes, cams have friction. So do hydraulic pumps, only they are very inefficient in terms of mechanical power input to hydraulic power output. And then there is also the flow losses and resultant heat generation of the hydraulic fluid in the system. And the additional design complexity of having a separate hydraulic fluid circuit in the engine, with the potential for leaks. Or if the system is designed to use engine oil as the hydraulic fluid, there is always the risk of thick-oil lovers running 20w50 in it instead of 0w20!
 
All of this is interesting .

I had in mind a system where electric solenoids operated the valves . W/o hydraulics .

100% electrical / electronic .

Go not know if that is workable ?

Thanks for the information , :)
 
Status
Not open for further replies.
Back
Top