Daimler and Mitsui invest in Mobility House for commercialization of energy management services with EVs
Hitachi Solutions C2X middleware platform now supports Japanese ITS standards

Researchers developing free-piston linear generator for exhaust waste heat recovery

Researchers in China have developed a novel free-piston linear generator (FPLG) to recover exhaust waste heat efficiently from a vehicle engine. The FPLG can be used in a small-scale organic Rankine cycle (ORC) system and can directly convert the thermodynamic energy of working fluid into electricity.

In a paper published in the journal Applied Thermal Engineering, the team from Beijing University of Technology, Collaborative Innovation Center of Electric Vehicles in Beijing, and Datong North Tianli Turbocharging Technology Co., Ltd. reports that the energy conversion efficiency of the FPLG can reach up to 45.82% with an intake pressure of 2.6 bar.

… the internal combustion engine is the power system of most vehicles. Therefore, it is necessary to recover exhaust energy of vehicle engine, and many strategies and methods have been adopted to reduce the energy waste and improve the fuel utilization efficiency.

In recent years, more and more researches focus on the Organic Rankine Cycle (ORC) system, which is an effective technical solution and a promising means for industrialization of energy savings among all the ways to recover waste heat. As a key component of the ORC system, the performance of the expander has a significant impact on the efficiency of the ORC system.

… In this study, a novel free piston expander coupling with a linear generator integrated unit, which can be used in a small-scale ORC system to recover exhaust waste heat from vehicle engine, is presented.

—Tian et al.

The FPLG in the study consists of a dual-opposed piston type free piston expander with a linear generator placed in the middle of two cylinders and a variety of auxiliary components. The piston is connected with linear generator mover by a connecting rod. The piston can move freely between its top dead center and bottom dead center. As the piston is not restricted by the crankshaft, its motion is influenced by in-cylinder gas pressure, electromagnetic force and mechanical friction. The researchers used a compressed air test rig for the experiments.

Prototype FPLG. Click to enlarge.

The two cylinders are alternately in the intake-expansion stroke and exhaust stroke. The compressed air flows alternately into each cylinder to drive the piston assembly back and forth, and the linear generator converts parts of the kinetic energy of the piston assembly into electricity.

Because the piston is not restricted by the crankshaft, top dead center and bottom dead center may be varied during the working process.

Among their findings:

  • The piston displacement profile is similar to a sinusoidal wave; the piston amplitude increases significantly with the increase of the intake pressure.

  • An increase in the operation frequency brings about a decrease in both piston amplitude and motion symmetry.

  • The peak piston velocity and power output show a nearly linear relation with the intake pressure, indicating that the power output is sensitive to the piston velocity. When the operation frequency is 1.0Hz and the external load resistance is 20Ω, the maximum peak velocity is about 0.69m/s and the highest power output 96W is obtained.

  • The conversion efficiency of the FPLG increases with the increase of the intake pressure.

  • When the operation frequency is 2.0Hz, energy conversion efficiency of the FPLG can reach up to 45.82% with an intake pressure of 2.6 bar.

Having determined the potential energy conversion efficiency of the system, the researchers will next investigate the effects of higher pressure and operation frequency on the FPLG and test FPLG in a small-scale ORC system.


  • Yaming Tian, Hongguang Zhang, Gaosheng Li, Xiaochen Hou, Fei Yu, Fubin Yang, Yuxin Yang, Yi Liu (2017) “Experimental study on free piston linear generator (FPLG) used for waste heat recovery of vehicle engine,” Applied Thermal Engineering, Volume 127, Pages 184-193 doi: 10.1016/j.applthermaleng.2017.08.031



96 watts.  When they said small, they weren't kidding.  But the unit proper is anything but small for a mere 96 watts.

The thermal efficiency is no doubt going to change with the properties of the working fluid.  γ for air is 1.4, but for most organic vapors I suspect it will be much less.  This requires a higher expansion ratio to achieve the same temperature drop in the vapor, with all that implies for heat transfer to the cylinder walls.  Of course, they could just use air and be able to discard the condenser.


Demystified this sits between the intake and post turbo exhaust if I am reading correctly.
Not sourcing the energy directly from heat but rather by the high / low pressure induction and expansion sides.
Frequency independent of rpm.


More likely pre-turbo


Post-catalytic converter.  You have to keep the gas hot to light off the cat, and putting a big heat exchanger upstream of a turbo would not only lose a lot of the gas volume but also dissipate the impulse energy.

I can see the boiler being incorporated into the muffler.

Brent Jatko

I wish there was a diagram of what they're trying to do.

Labeled photos of laboratory Stirling engines don't exactly give me the idea that this is coming anytime soon enough to matter.


It's not complicated.  They have a two-piston expander there.  The servomotors apparently operate the valves.

This would appear to be ideal for a sleeve-valve setup to radically decrease the flow losses over poppet valves.

Liviu Giurca

There are simpler solutions which integrate the expander in the volume of the engine itself. The internal heat recovery can be made in a compact and simple structure. Please see www.hybrid-engine-hope.com


This design is to complex to bring down to reasonable price point to even entice consumers to purchase vehicles with said system. Simplicity would be a rotary design using tested and tried existing designs like a scroll expander (1) or engineair (2) which I'll openly concede the second option won't be cost effective till his patents expire. Also, his motor design flaw is that it's using compressed air tanks which NOVEC 7200 or Vertrel SDG could be more ideal for waste heat recovery on EV's or traditional ICE's.

(1) https://www.youtube.com/watch?v=ZQDshN_5VkU - AirSquared
(2) http://www.engineair.com.au

Nobody wanted to pay Honda royalties to use their VTEC variable valve timing and now that their patents are expired every manufacturer has them. Nobody wanted to use Toyota's prius hybrid design but government forced them by threatening to bar sales in California if they didn't meet X hybrid targets. Just as Ford didn't want to pay the designer of the intermittent wiper motor which they even made a movie about it named, "Flash of Genius."


It seems as most of you are assessing this invention as if it would be almost production-ready unit. It is just a laboratory set-up, which, at most, qualifies as a proof-of-concept. For example, you cannot make a compact unit if you run at maximum at peak velocity of 0.69 m/s. A free-piston unit should achieve (mean) piston speed in the range of (or above!) modern light and heavy-duty engines, i.e. at least >10 m/s. Modern steam engine concepts also aim at similar piston speeds. Operating at higher pressures also decrease the size considerably. Note the comment: “…the researchers will next investigate the effects of higher pressure and operation frequency on the FPLG…”. Furthermore, note that steam engines are, per definition, 2-strokes, which gives double the number of power strokes compared to a 4-stroke design (in this case, without the drawbacks that plague 2-stroke ICEs). A power cylinder of a steam engine does virtually no compression work in the cylinder; only expansion, which also increase power density. With all positive factors combined, there is potential for very high power density, which also modern steam engines prove by being much more compact that a corresponding 4-stroke gasoline or diesel engine. Whether they will succeed in developing the concept to a commercial unit is another story.


Organic vapors generally have high MW and thus low jet velocities, which makes small turbines attractive.  I'd bet that you could get 96 watts out of a rotor the size of your thumbnail.


Organic fluids are not well-suited for high temperatures and pressures, so in case a piston expander is used, water or CO2 would be better options. With water, dimensions could be very small as can be seen in the two references below. At 250 bar, 450°C and 6 000 rpm, you get 1 200 kW per liter of displacement (fig. 8 in the first reference). That is sufficiently small for most applications and certainly much smaller than any conventional ICE today.




You also get extremely high jet speeds with steam, and thus rotational speeds.  There's a reason steam turbines have so many stages compared to gas turbines; it extracts the energy in smaller steps, with lower jet speeds.

The comments to this entry are closed.