Harman and Tsinghua University establish joint research lab with focus on connected vehicles
Plug Power and Hyundai Hysco sign MOU to form fuel cell JV in Asia

Toyota Central R&D developing free-piston engine linear generator; envisioning multi-FPEG units for electric drive vehicles

Toyota’s FPEG features a hollow step-shaped piston, combustion chamber and gas spring chamber. Click to enlarge.

A team at Toyota Central R&D Labs Inc. is developing a prototype 10 kW Free Piston Engine Linear Generator (FPEG) featuring a thin and compact build, high efficiency and high fuel flexibility. Toyota envisions that a pair of such units (20 kW) would enable B/C-segment electric drive vehicles to cruise at 120 km/h (75 mph). The team presented two papers on the state of their work at the recent SAE 2014 World Congress in Detroit.

The FPEG consists of a two-stroke combustion chamber, a linear generator and a gas spring chamber. The piston is moved by the combustion gas, while magnets attached to the piston move within a linear coil, thereby converting kinetic energy to electrical energy. The main structural feature of the Toyota FPEG is a hollow circular step-shaped piston, which Toyota calls “W-shape”. The smaller-diameter side of the piston constitutes a combustion chamber, and the larger-diameter side constitutes a gas spring chamber.

Cartoon of the W-shape piston. Click to enlarge.

FPEGs are attractive for a number of reasons, the Toyota researchers note, including thermal efficiency, low friction, and low vibration. Two basic design approaches have emerged: the first is a structure with two opposed combustion chambers; the second, a structure with one combustion chamber and one gas spring chamber. In the latter approach, the gas spring chamber is responsible for returning the piston for the subsequent combustion event. This second configuration is the one selected by Toyota for further investigation by both numerical simulation and experimentation.

The Toyota FPEG is based on a double piston system; at one end is the combustion chamber, and at the other, the adjustable gas spring chamber. Burned gas is scavenged out through exhaust valves mounted in the cylinder head of the combustion chamber; fresh air is brought in through the scavenging port at the side wall of the cylinder liner.

A portion of the kinetic energy of the piston is stored in the gas spring, and extracted on the return stroke to the combustion chamber side. A magnetic “mover” is mounted at the outer periphery of the piston; the mover and the stator coils together comprise the linear generator component of the FPEG.

Animation of Toyota’s FPEG. Click image and video will launch in large separate window. Video played in embedded player on Toyota site is available here.

The “W-shape” piston design offers several advantages, Toyota says:

  • The larger cross-sectional area of the gas spring chamber leads to lower compression temperature of the gas spring chamber and consequently decreased heat loss.

  • The piston has a hollow structure and moves along a column stay, which in turn enables the construction of a cooling oil passage within the stay. The key technologies to deliver stable continuous operation of an FPEG are lubricating, cooling, and control logic. (The Toyota team’s second paper deals exclusively with the control system.)

  • The inner periphery of the hollow piston also serves as a sliding surface on the column stay, enabling a steady small clearance between the magnets and coil for improved generating efficiency.

  • The magnet is set far from the piston top, preventing magnet degaussing by heating.

The researchers developed a one-dimensional cycle simulation to investigate the performance of the proposed structure, and used it to assess spark ignition combustion (SI) and premixed charged compression ignition combustion (PCCI). They achieved output power of 10 kW with both SI and PCCI combustion cases; the PCCI combustion case realized 42% thermal efficiency.

They then constructed an FPEG prototype with a uni-flow scavenging type, two-stroke SI combustion system as an experimental study. They used ceramic-coated piston rings and cylinder liner they developed in order to ensure the smooth sliding of the piston even under insufficient lubrication. Poppet valves seated in a water-cooled cylinder head were actuated by hydraulic valve trains to control exhaust valve timing. Direct injection reduced unburned hydrocarbon emissions exhausted through the scavenging process.

A pressure regulating valve in the gas spring chamber enabled a variable gas mass, thereby varying the stiffness of the gas springs—one of the variables to shift the FPEG to different operating points.

The linear generator was a permanently excited synchronous machine consisting of the stationary coil, the mover (based on neodymium-iron-boron magnets) attached to the piston, and iron-cored stator. The poer electronics drive the machine as both a motor and a generator.

The researchers designed the prototype control system to ensure that the compression ratio was kept to the values which enable stable combustions—i.e., the generating load coefficient is variable, not constant. The coefficient is determined by a feedback control method based on the postion and velocity of the piston.

As there is no crank mechanism, the piston position in an FPEG is not defined with crank angle. However, knowing the piston position is critical not only to timings (fuel injection, ignition, opening/closing exhaust valves), but also to mode selections of driving or generating. To determine piston position, the Toyota researchers count plural-lines grooves engraved on the side surface of the piston body with gap sensors fixed on the inner wall of the cylinder block. (The detailed method of detecting and controlling piston position is the subject of the second paper.)

The generator control logic must meet the following requirements, according to the researchers:

  • Assuming a multi-unit vehicle application, the multiple FPEGs would cancel out vibration through a horizontally opposed layout; the frequency and phase of the piston oscillation should be controllable.

  • TDC and BDC need to be precisely controlled for stabilizing two-stroke combustion.

  • After knocking or misfire, the oscillation must continue robustly.

The prototype FPEG with W-shape piston and two-stroke SI combustion system achieved stable operation for more than 4 hours without any cooling and lubricating problems.

The experimental analysis also showed that the precise control of ignition position is essential for stable operation of the FPEG.

In future work, the research team plans to improve the power generation of the system and to perform a quantitative analysis of the efficiency.


  • Kosaka, H., Akita, T., Moriya, K., Goto, S. et al. (2014) “Development of Free Piston Engine Linear Generator System Part 1 - Investigation of Fundamental Characteristics,” SAE Technical Paper 2014-01-1203 doi: 10.4271/2014-01-1203

  • Goto, S., Moriya, K., Kosaka, H., Akita, T. et al. (2014) “Development of Free Piston Engine Linear Generator System Part 2 - Investigation of Control System for Generator,” SAE Technical Paper 2014-01-1193 doi: 10.4271/2014-01-1193


Steve Reynolds

This looks like an excellent alternative to use in vehicles that Toyota is designing to use fuel cells. Then a version of essentially the same vehicle can be sold where hydrogen is not readily available.


Wow. Haven't seen one of these since a Stirling cycle Solar Dynamic concept in the 80's at NASA Lewis. How the world rediscovers innovation... not a jibe, just an observation. (Probably fueled by anguish over cable news talking heads today announcing the "incredible" vertical landing of the Grasshopper, recalling our DC-X doing the same thing about 20 years ago.) What's old is new again.

Back in the day, a fuel-burning version of the FPE was always daunted by thermal problems with the electromagnetics (assuming you actually had the piston control problems solved). Here's hoping Toyota has it licked (ditto NVH challenges that are surely not trivial). Here's hoping they have all this stuff licked. It would be a great addition to the solution portfolio.


I can't believe they are still pouring money into range extender technology. Within a few years, electric cars will have enough range and fast charging capabilities to turn the plugin hybrid into a temporary quirk of history. Maybe there is a longer lasting niche market for plugin hybrids, but a niche market doesn't justify spending a lot of R&D money.



Consider the possibility that Toyota has more insight into the medium-term prospects of electric-only cars.
Toyota sells cars in all markets, so they realize that solutions that work well in infrastructure-rich locations with mild weather may not sell elsewhere.

Witness the Prius. It's undoubtedly a success, but the majority of its worldwide sales are in two markets: Japan and California.


It looks like a great idea to me - a 20Kw range extender.
if you had one, you could reduce the battery range of BEVs to your daily requirements and use the extra weight and space for the range extender.

I suppose it all comes down to noise and vibrations and neat packaging - you don't want to lose your trunk to fix one of these.

Also, it had better be quieter than the BMW I3 range extender, which, by all accounts is very nosy.


"Free Piston...The experimental analysis also showed that the precise control of ignition position is essential for stable operation of the FPEG.."

There is part of the problem, you horizontally oppose to cancel vibration, but the free piston design is hard to synchronize both.


SJC: It's not that hard to synchronize them since the linear generator is also a linear motor. Of course when you exercise a fairly rigid control on the piston position, you'll need some decent sized capacitors to smooth the output power.

In terms of internal combustion machines, I think this is a very exciting development. It basically is to crank-tied piston machines what fuel injection is to carburetors. It allows fine grained variable control of piston position, compression ratio, etc. And it also appears to have the benefit of being extremely compact.


Many years ago I looked at this problem. The hydraulics lab happened to have had a free piston engine dating from 1964. One thing for sure, that device certainly was noisy when running. And just like hydrogen fuel cells - as someone noted here - it will be tomorrows solution. ALWAYS.

But it forces you to appreciate rotary motion all the more.

However regarding rotating pulsating energy systems, the current practice is to use a crankshaft flywheel and a damper disk to smooth pulsations to a 3-phase generator which of course would prefer a constant torque input delivery.

Noting that flywheels and damper disks are large and not inexpensive there must be some way to eliminate both these middlemen !

The mechanically "stiff" flywheel is not sympathetic to pulsating power and despite the use of a damper disk it is the cause of subsequent engine vibration.

An idea I suggest is to electronically switch the rectifier tank circuit of the generator so that it would efficiently absorb the energy into an LC network and provide more of a dampening effect to the crankshaft. Basically you are understanding and accepting that the adiabatic expansion that forces the piston down the cylinder bore must be received with a counterbalancing electromagnetic force. A mechanical phasor would be required initially in the lab - the printing industry has used double planetaries for colour registration for many years - to align the generator magnetic poles on the rotor with the point of maximum thrust from the crankshaft.

Consider that a single cylinder engine will provide one mechanical power pulse every 720 degrees.
Consider also that a two pole single phase generator will absorb one mechanical power pulse every 180 degrees.
Clearly one match would be to use a 4:1 reducer gear between the engine and the generator.
Or one could take a four cylinder engine which would provide one mechanical power pulse every 180 degrees and therefore directly connect it to the generator.


"The generator control logic must meet the following requirements, according to the researchers:

Assuming a multi-unit vehicle application, the multiple FPEGs would cancel out vibration through a horizontally opposed layout; the frequency and phase of the piston oscillation should be controllable.

TDC and BDC need to be precisely controlled for stabilizing two-stroke combustion.

After knocking or misfire, the oscillation must continue robustly."

Pretty much what I said.
In the engineering world it is referred to as TBD..To Be Determined.

Roger Pham

This is great as a stationary genset, being very simple and compact, running at a fixed 3,600 rpm for 60Hz AC or 3,000 rpm for 50Hz AC. It can run on NG or even H2, or a mixture thereof, thereby can provide co-generation of power and heat for the present, on NG, or for the future, on H2 from local H2 piping network. For co-generation purpose of heat and power, the somewhat lower thermal efficiency than a PEM-FC is not a disadvantage, while the ability to run on both NG and H2 is a major advantage. The crankless design offers perhaps higher durability and lower vibration and higher efficiency than a slider-crank mechanism.

However, for automotive use, the ability to contribute rotational torque directly to the wheels to augment the power of the motor is very important. The ICE in a PHEV can double the total power output of the vehicle, thereby increase sale and higher price potential of the vehicle. Therefore, I won't expect this free-piston engine in a vehicle at anytime to come.



I agree with you that is going to be the strong trend, but there's still cold places that basically need a PHEV too, and also people who must park on the street without easy access to overnight topping off of batteries.

@Roger Pham

Very good point. BYD's Qin offers a very very compelling package with the no compromise performance by doing exactly that.



Toyota Central R&D Labs has invented a latest technology which will be used in electric vehicle for increasing efficiency & performance. The tool is named as Free Piston Engine Linear Generator. The tool consists of two-stroke combustion chamber which will convert the kinetic energy to electric energy. Now a day the demand of electric cars is increasing. The car makers are focusing on development in that particular area. Hence the FPEG will be a latest invented technology for electric cars.

John Burns

This free piston generator looks to be cheaper to run than using grid power electricity to charge batteries and run an EV car. If so, which looks the case, then the onboard battery set will be just a buffer. This may increase emissions in cities, which is not desirable. Incorporate supercapacitors along with batteries for kinetic energy brake reclaim and acceleration, emissions will be very low indeed. It may get to the point it may be worth running your can in the driveway and plugging it into your house to heat a thermal store cylinder of water to use to heat the house and use for hot water.

Henry Gibson

Free piston Stirling designs from Infinia now Qnergy might be interesting, but have more heat transfer problems. ..HG..

The comments to this entry are closed.