ARB Incentive Program Puts 650 Hybrid Trucks on California Roads
Zero Motorcycles Awarded Additional Grant Funds to Support Advanced Electric Powertrain Development Program

Lund Team Shows 57% Thermodynamic Efficiency in a Gasoline-Fueled Heavy-Duty Diesel Engine Using PPC

Gross indicated efficiency (%) of Scania heavy-duty diesel running on gasoline using PPC. source: Bengt Johansson. Click to enlarge.

Researchers at Lund University in Sweden have shown a thermodynamic efficiency of 56% in a gasoline-fueled single-cylinder light-duty engine and 57% in a gasoline-fueled single-cylinder heavy-duty engine using the Partially Premixed Combustion (PPC) concept. Under higher load, they achieved 52-55% thermodynamic efficiency with 99.8% combustion efficiency and engine-out NOx below US10/Euro6 levels in the heavy-duty engine.

Prof. Bengt Johansson presented the results at the US Department of Energy’s 16th Directions in Engine-Efficiency and Emissions Research (DEER) Conference in Detroit.

Although HCCI (Homogeneous Charge Compression Ignition) was developed as a means to achieve ultra-low NOx and soot simultaneously with higher efficiencies, the viability of HCCI applications appears limited to lower load operations. Among the challenges facing HCCI combustion are acoustic noise, lack of direct control of the combustion and too diluted air-fuel mixture requirement.

Partially Premixed Combustion (PPC) was developed as a means to increase efficiency in the order of 50% or above, achieve low NOx and soot, and run the whole load range. PPC is a mix of the classic diesel combustion process and HCCI using a fully pre-mixed charge. If you consider standard diesel combustion as “black” and HCCI as “white”, Johansson said, PPC is some varying shade of gray.

PPC uses a rather early injection to create a premixed fraction of the fuel mixture, and a late injection to obtain stratification. By changing the ratio of these two injections, it is possible to tune the burn rate.

More fuel in the first injection means a faster, more HCCI-like combustion; more fuel in the second injection results in combustion more like regular diesel diffusion controlled combustion.

The Lund team worked with a Saab variable compression ratio engine, a GM L850 world engine, and the Scania heavy-duty diesel engine. As a baseline, they found that running the GM engine in spark ignition (SI) mode at a compression ratio of 9.5:1 (standard) at low load resulted in a thermodynamic efficiency of about 30%. Boosting the compression ratio to 18:1 resulted in thermodynamic efficiencies in the range of 30-40%. Switching combustion mode from normal flame propagation to HCCI resulted in roughly 50% thermodynamic efficiency across the three engine platforms, Johansson said.

We can also see all three engines would have about the same thermodynamic efficiency. There is no major difference...We can also see that HCCI kind of stops at about 6 bar BMEP. We did 20 bar HCCI, but I would not recommend it...There is actually a [combustion efficiency] penalty going from SI to HCCI. People saying that HCCI is an efficient combustion process, I wouldn’t really say so because you lose 10% of the fuel.

—Bengt Johansson

When using a diesel (compression ignition) engine running with diesel fuel, the load region in which the engine can run in PPC mode is limited to 5-6 bar gross IMEP, noted Vittorio Manente from Lund in his PhD thesis on the subject. Increasing the upper load PPC boundary in a diesel engine could thus either require a piston with much lower compression ratio has in conjunction with “an intolerable amount of EGR” to keep the start of combustion and the end of injection separated; or a fuel that is more resistant to auto ignition needs to be used: e.g., low-cetane diesel or gasoline.

The Lund team used regular US gasoline in their engines.




Their diagram says "gross indicated efficiency" so probably not the efficiency at the output...


I must remind you again that 57% is an indicated efficiency. If we presume that the mechanical efficiency is - at maximum - slightly better than 90%, we would give us a brake efficiency of 51-53% (e.g. 0.92*0.57=0.52). We must compare apples and apples. It is still very good results, though! Optimistic? Why? Researchers have to publish their work in peer reviewed journals and similar. It is not absolutely a proof of that everything is correct but it is still better than most of the data out here on the Internet.

Simple theoretical calculations do not provide that much insight about efficiency. The simplifications are too crude. The difference between the ideal and the real engine is great enough to indicate this fact. However, a theoretical engine does care how fuel is injected, or more correctly, how combustion proceeds. This is a way to distinguish between, e.g. otto, seiliger and diesel processes. In the past, the seiliger process (somewhere in between otto and diesel) was representative of a “real” diesel engine. Today, the “real” diesel engines have more and more approached the “ideal” diesel process. In contrast, I suppose that the PPC engine is best represented by a seiliger process. This indicates higher maximum cylinder pressure than a diesel process. The “real” limit is cylinder pressure, not the geometric compression ratio, as some might think. Thus, the compression ratio might have to be reduced for a PPC engine compared to a conventional diesel engine, or else, the engine structure must be developed to cope with higher pressure levels. Somehow, there is a misunderstanding that PPC/PCCI/HCCI engines would always be lighter than diesel engines. Maybe so, if this type of combustion is limited to low loads. If we strive to operate also at high load and at high compression ratios to obtain high efficiency, this engine could actually become heavier than a conventional diesel engine. Thus, the most obvious starting point would be to use a conventional diesel engine block, as the authors did in this case.

For those who care about particle emissions, DPF is a solution. Beyond doubt, this engine would have to use a DPF to achieve similar particle emissions as a diesel engine. With best type of DPF, we could get a particle level lower than in ambient air, which should fulfill EU regulations and give a much greater margin to all US regulations that are not as strict as EU regulations.

US gasoline can be obtained from several suppliers of test fuels in Europe, or else, it can be imported from the USA. Nothing magic… By the way, the specifications for those test fuels are much tighter than commercial fuels at the filling station. This is why researchers often use them. Another approach is to analyze the commercial fuel used. The problem here is that the fuel will change from batch to batch in a way that tightly specified test fuels do not. I am sure Johansson et al. are familiar with these issues.


From time to time OPEC TELLS they will cut their output, but there is a big difference between their intentions and what they really do.
Anyway, of the 86 million barrels produced today, only a relatively small (and decreasing) part is produced by OPEC members. they TOLD to decrease their production by 1.5 million barrels. that's not even 2% of global production, and eventually, they didn't cut anything.
What we need to do is cut our demand, simply by not using petrol anymore for heating and electricity generation, and switching to alternatives in other fields. The relatively small amounts of liquid fuels we will probably still need for decades to come, can easily be made from other sources.


FYI, "gross indicated" output probably means the power output plus the zero-fuel engine drag at the same speed (this is one of the tricks Detroit used to inflate horsepower ratings, publishing gross rather than net). The net efficiency at the crankshaft will be considerably lower.


If you were a "real" engineer, you should know the denotations in the business. There is a clear definition of this in the literature. If you really insist, I could explain it to you but not this time… I have already shown you how to re-calculate indicated efficiency into brake efficiency, using my assumption of a mechanical efficiency. It should be clear enough. Indicated efficiency is obtained by using the measured cylinder pressure trace. Researchers usually report these data when they, for example, have a single-cylinder research engine that has much higher friction than the corresponding multi-cylinder engine. It is also possible to compensate for higher friction to correct the numbers to better represent those of a multi-cylinder engine. However, using a measured value as they have done here is actually more straightforward. There are no dirty tricks involved here. You just need to know what they are talking about.


HCCI / PCCI would be very difficult to use over a normal operating range of an engine, but should be much simpler to use as a range extender. However it could be the case that a less efficient but much lighter engine could be a better choice


The author refers to PPC, not HCCI/PCCI, although one might discuss the difference between PPC and PCCI. If you look at the diagram you will see that they have been able to run on pretty high load. At even higher load, one idea is to use more conventional diesel combustion. If you could not run the engine at similar load as the diesel counterpart, it would be like anti-downsizing, so it is essential to achieve this target. Nevertheless, an engine as this one could be of interest as range extender, since this could simplify the engine control.

Roger Pham

Thanks, Peter, for the clarification.
It is perhaps the more rapid combustion of the PPC process that raises the indicated efficiency to 57% vs. the indicated efficiency of ~50% for traditional turbodiesel. At a leaner premixed fuel:air ratio, the compression ratio can be raised to higher than 10:1 for higher efficiency than traditional SI Otto-cycle engine, but the detonation risk will still be there, especially on an older engine with in-cylinder deposits. Perhaps a knock sensor can be used to lower the premixed fuel:air ratio whenever engine knocks is detected.

Professor Rolf Reitz has been researching along the same concept, using premixed gasoline by a port injector and trigger the ignition with an incylinder diesel injector. More reliable ignition is achieved this way, due to the higher ignitability of diesel fuel. The disadvantage is that two fuel tanks and two fuel injectors will be needed. But, Prof. Reitz must have his reason for doing so, since by the Lund's method, if lower octane gasoline is used, you would risk detonation and hence limit compression ratio, but if higher octane gasoline is used to raise compression ratio, then you would have ignition problem.

Brian P

Regarding the peak efficiency numbers: The standard efficiency calculations take no account of the Atkinson cycle. The expansion ratio is assumed to be the same as the compression ratio for both the ideal Diesel and Otto cycles. Using the Atkinson cycle extracts some useful work over and above that. Granted, we don't know if the research engine was using that, but it's not difficult to imagine that it did.

The research engine overcomes detonation of gasoline by injecting it late enough in the compression stroke that the fuel does not reach detonation conditions. It does self-ignite (that's the whole idea) but the idea is that the fuel is not evenly distributed in the cylinder, so it burns progressively rather than detonating. It is mechanically a combination of a gasoline engine and a diesel engine. Many production direct-injection gasoline engines are also doing this to fight detonation, although not even close to the same extent that this research engine does.

One would imagine that the combustion process is very fast - closer to the ideal Otto cycle (constant-volume). For a given compression ratio, the idealized Otto cycle has a higher efficiency than the idealized diesel cycle. I get 69% for an idealized Otto cycle at 18:1 compression and that doesn't account for the possibility of some Atkinson operation.

If you were a "real" engineer, you should know the denotations in the business.
I don't work in the research labs any more, so I don't rub elbows with the people who do this for a living and I've forgotten some of the jargon. Of course, if YOU were a real engineer you'd have a hyperlink to the definition of the term instead of just posting insults. I was just too hurried to look it up.
There are no dirty tricks involved here. You just need to know what they are talking about.
What's dirty about (a) noting HISTORIC practices of Detroit, with which YOU have no experience, and (b) noting that the efficiency at the consumer's crankshaft won't be as good as the lab figure?

Knowledge about what indicated efficiency is very elemental for an engineer. Therefore, I doubt your experience from this business. You can find the mentioned definition of in any text book on engines, so hyperlinks are not necessary this time.

The real problem is that many people who read the comments on this site might believe your misinformation. The results from this research group are among the most substantial findings in this field for the last decade. It is very important to note this fact. From the debate on this site, one might get the impression that it some kind of scam instead. The truth is the opposite. The DEER conference is for devoted to participants from the USA and the presentations are mostly from R&D projects funded by US DOE. In some exceptional cases, speakers are invited from abroad the USA. This presentation was one of these. This alone, should imply that it was a very important contribution at the DEER conference.

Bengt Johansson is not from Detroit. He is a professor at the Lund University in Sweden. By the way, I know a lot of people in Detroit. The majority of them are very decent people.


Brian P
Of course you can to an idealized (theoretical) calculation of an atkinson cycle. In fact, both atkinson-diesel and atkinson-otto and a cycle anywhere in between (atkinson-seiliger) can be used in idealized calculations. The test engine did not use any kind of Atkinson cycle.

Yes, the combustion is fast for this engine; it is probably best represented by a seiliger cycle (steep pressure rise and then a fairly constant pressure level).

Geometric compression ratio is not a real practical limit, albeit tolerances for the manufacturing process. It is possible to make an engine with a compression ratio of 50:1, if you like (of course it is not practical). The real limit is cylinder pressure and, of course, the problem of knock in the otto engine case. If you calculate the idealized (theoretical) efficiency for otto and diesel cycles, you will find that the diesel cycle is more efficient at any level of cylinder pressure. This simple fact is often overlooked in text books but the more in-depth ones can provide more thorough information on this topic. However, it is nothing new, since Rudolf Diesel himself discovered this fact.



I think I may have figured out the mystery of the efficiency percentages.

There are two ways to measure thermal efficiency: indicated thermal efficiency and brake thermal efficiency.

In general, we are more interested in brake thermal efficiency because that is efficiency at the driveshaft. In the graph, they state they are measuring gross indicated efficiency. I found the original paper from Lund University here:

In the paper, on page 23, you will see the "brake efficiency" is close to 49% not 57%. The "gross indicated effiency" is 57%.

Also, Roger, I am not an engineer, but the theoretical maximum efficiency for an Otto Cycle engine is 60% at 10:1 compression according to here:

Now, 49% is very good and is the best I've ever seen for a gasoline powered engine.

The question is will the production version of EcoMotors OPOC engine be the same or possibly higher?



Roger Pham

The 60% theoretical efficiency of the Otto cycle engine at CR = 10 is only a very gross approximation and over-simplification. To calculate a more accurate number, you must realize that when compress air adiabatically, you'll get a molar heat capacity ratio (Cp/Cv) of 1.4 for diatomic molecules, for both N2 and O2. However, in the adiabatic expansion phase, the production of combustion now contains ~21% of triatomic molecules, being H2O and CO2, having molar heat capacity ratio of only 1.3. If you would recalculate using this information, you would see that the 56% theoretical efficiency for Otto-cycle engine at a CR=10 is quite a generous estimation.


Ralph & Co.
Now, I have also had the time to go through the thesis. This is older than what was presented at DEER but most of it is surely representative of what Johansson presented anyway.

Since the authors have used a single cylinder engine, measuring the indicated efficiency (net or gross) is the best option but they have also calculated the brake efficiency for a hypothetic multi-cylinder engine would. To do this, one has to know the mechanical efficiency and pumping losses. The authors themselves calculated the brake efficiency to 50% (not 49%). In my previous simple calculation, I assumed a mechanical efficiency of 92% and no pressure drop over the engine (i.e. a gas exchange efficiency of ~100%). With these numbers you get 52% brake efficiency as I stated earlier. In their measurements of mechanical efficiency on a multi-cylinder engine of the same type, they have obtained a slightly higher mechanical efficiency. Thus, I was somewhat pessimistic. Usually, the pressure difference over the engine at the load and speed they tested is slightly positive or close to zero. This is of course dependent on the turbocharger efficiency and matching at that test point. Now I can see in the thesis that the authors have been more pessimistic than that. They have anticipated n-e-g-a-t-i-v-e pressure drop, i.e. gas exchange efficiency lower than 100%. Thus the turbocharger efficiency would be significantly below 50%. It is known from the literature that it can be higher, i.e. some 55% or even higher. If we use a mechanical efficiency of 93% and zero pressure drop over the engine, i.e. 100% gas exchange efficiency, we get a brake efficiency of 53%. Still, the assumption on gas exchange is rather pessimistic. If we use a higher value, close to 55% brake efficiency is in fact realistic. Turbocharging systems that can achieve this are already in production. What I refer to specifically is two-stage turbocharging with intercooler and aftercooler. Such turbocharging was not used on this particular multi-cylinder engine. (Turbocompounding is another way to improve efficiency.)

All-in-all, I would say that the assumptions the authors used to estimate brake efficiency actually have been on the pessimistic side. This also makes the achievements even more remarkable! Well, of course you can find inventors and development companies that claim much more, but it is mostly not substantiated by any reliable test data. Nor can you find such results in peer-reviewed publications. DOE has target in their R&D programs to achieve 55% efficiency and in fact, engine manufacturers are on the way to achieve this. Development in many areas, not only combustion but also engine friction, heat and pumping losses, auxiliaries, mild hybridization and heat recovery seems to be necessary to meet the target. Thus, the target of 60% in this project actually may not be impossible. It would, of course, also require attention in many other areas, as indicated above.


OPEC reducing output would be the best thing for EV's. Expensive gas is not enough, it's SHORTAGES that are needed. When people can't get fuel, that's when EV's will look great.


I agree, I do not want to see shortages, but they do focus attention. In the 1970s you could get gasoline at 70 cents per gallon, if you could get it. People were in lines at filling stations with odd and even licence plate last digits matching the date on the calender.

This put quite a ripple in the force. People did not know what was coming next and along with "stagflation" provided a backdrop for more uncertainty. I want us to get ahead of the curve and solve the problem before it occurs. It is a high probability with a ceiling on supply and ever increasing demand, crunch time will occur.

Stan Peterson

What OPEC could attempt to do in 1986 and what it could attempt in 2010, 25 years later is quite different. The aging OPEC cartel, is a lot weaker and more fractionated than then. It controls a lot less of the total liquid fossil market, even as that market decays.

Plus there is a tremendous amount of Oil not currently on the market but that could return soon. Iraq used to be the second or third largest producer before the wars. No, It has not returned to anywhere near that production, but it is starting to come back with the outbreak of relative "Peace".

Iran has seen its output plunge as the low maintenance of its Oil fields for a decade of loose punishments have still had their effects. Chavez and his cronies have killed the Venezuela production from under-investment and mis-management endemic to all leftist control. Canada's production from the world's largest Oil reserves, in its Oil sands continues to climb. Brazil is ready to be a world player with its deep sea Oil, as would the US if the anti-Science clique ever to get out of the way.

As for so-called "Peak Oil", remember if you ever even learned, that this was originaly about refining capacity for anything other than sweet, low viscosity crude, rather than oil per se. The mis-labeled "Peak Oil thesis" was that there would never be the Refineries to handle the high viscosity and sour crudes, becoming dominant. Well, virtually all the world's Refineries have already been upgraded, so the original Peak Oil thesis has been proven completely fallacious. I urge you to read the original Peak Oil thesis papers, rather than the warped interpretations from the Oil Spot, propagandists.

Will there be a genuine "Peak Oil" versus "Peak Refinery" crisis? Sure, in about half a millenia, in 2510 or so. Presumably we won't care then.

The comments to this entry are closed.