200 Years Of The Stirling Engine

In the early years of the nineteenth century, steam engines were at work in a variety of practical uses. However, they were still imperfect in many ways. One particular problem were the boilers, that had a tendency to explode, causing injuries and fatalities. Reverend Dr. Robert Stirling, a Scottish clergyman, was concerned about the death toll from exploding boilers. Based on previous work by George Cayley (known for his pioneering work on aeronautics), Stirling filed his patent for a safer engine in 1816. That makes this year the bicentenary of this engine. The Stirling engine has the highest theoretical efficiency of any thermal engine. It is also a relatively simple machine. Unlike other types of engines, there are no valves, and that makes the mechanical design much simpler.

Principles of a Stirling Engine

In a Stirling engine, as with any other heat engine, there are two zones at different temperature, engineand the working fluid is moved between them to extract work. In the animation at the right, the cold zone is in blue and contains the power piston. The hot zone is in red and contains another piston called the displacer. Heat is added to the hot zone and the gas expands, pushing the power piston to produce work. When the power piston is about to reach the end of its stroke, the displacer moves to the left, moving the gas from the hot zone to the cold one. The gas is cooled and contracts, and the power piston moves to the left.  Then the cycle repeats.

alpha_stirling-1 There are two basic configurations for a Stirling engine, one is the beta type that we discussed above; the other is the alpha type, shown at the left. The principle of operation is the same. In the alpha engine, there are two power pistons that join at the same point in the crankshaft, and the gas is shuttled between the two cylinders. One key component and the basis of the 1816 patent is the regenerator, basically a heat exchanger that is placed between the hot and cold zones. The regenerator retains heat that would otherwise be lost to the environment, increasing the thermal efficiency. There are many variations in design that use the Stirling cycle, such as:

Advertising for an early Stirling fan, circa 1900.

Waiting for Material Science to Catch Up

Stirling continued to make refinements to the design, with the aid of his brother James (an engineer). Their engines were produced and sold as an alternative to the steam engine. But in order to achieve maximum power and efficiency, the Stirling engine needs to operate at high temperatures. The materials available in those early years simply were not up to the task, resulting in frequent failures. At the same time, the steam engine evolved and became safer.

Stirlings also were considerably larger than a steam engine of the same power. Because of that, they took a back seat to other competitors, mainly steam and the new internal combustion engine. Stirling engines were confined to low power applications, such as ventilators, water pumping and providing air for church organs. But eventually, even in this applications the technology was replaced by the electric motor. The development of steel solved many of the engine failures, but it was too late. In 1876, Stirling wrote:

…These imperfections have been in great measure removed by time and
especially by the genius of the distinguished Bessemer. If Bessemer iron or steel had been known thirty five or forty years ago there is scarce a doubt that the air engine would have been a great success… It remains for some skilled and ambitious mechanist in a future age to repeat it under favorable circumstances and with complete success….

Stirling Generator and Cryocooler

Further development of the engine halted, until the mid-1930s, thanks to the Phillips Research Laboratory in Eindhoven. In order to boost the sales of its tube radios, Phillips wanted a small, quiet and thermally powered electric generator for remote areas. In 1951, they succeeded in developing a Stirling generator capable of generating 200 watts of electrical power, but the price was not competitive.

Phillips Stirling generator.
Phillips Stirling generator.

By that time, transistor radios made their debut. With minimum power requirements, they could be powered with batteries, and the Stirling generator was no longer needed. However, the research by Phillips produced a wealth of new knowledge for further development of the Stirling engine. In fact, they developed in parallel a cryocooler, which is considered the only commercially successful application of Stirling engines. The engine, working in reverse, produces a temperature difference when mechanical work is applied to the shaft. Temperatures as low as -200 C can be obtained, enough to liquefy air.

Stirling in Automotive

There were efforts to develop an automotive version of the Stirling engine. In 1986, the MOD II project produced an engine capable of 38.5% thermal efficiency (modern automotive engines are in the 20-25% range). The engine was tested in a Chevrolet Celebrity notchback, and the urban gas mileage was increased from 26 to 33 mpg. One disadvantage was that the startup time was of around 30 seconds. However, Ford developed a version with internal heaters that could start the engine in a few seconds. Lack of interest by automakers and cutbacks in research funding prevented the continued development of an automotive Stirling.

Stirling radioisotope generator.
Stirling radioisotope generator.

Finding Its Niche

Besides the cryocooler, the Stirling engine is in use in some other niche applications. The free piston Stirling engine (FPSE), uses an alternating magnetic field to drive the piston and generate a temperature difference. Thanks to the magnetic drive, there is no need for seals or gaskets, and the unit can be completely sealed. Because of its high reliability, it is being used by NASA to cool instrumentation in satellites.

Possible future uses include desalination units, such as the slingshot, invented by Dean Kamen, that uses water vapor distillation to purify water. It can use any form of fuel to power its Stirling engine. In solar power, Stirling engines are being tested by placing them in the focus of parabolic mirrors and driving solarstirlingengineelectric generators. With this setup, the efficiency is better than with non-concentrated photovoltaics.

They have also been proposed to replace steam turbines in nuclear power plants, eliminating the need for water in the system. Another nuclear application is a Stirling engine that uses nuclear fuel as the source of heat, providing years of power for use in space exploration.

In recent years an interesting idea for home use has gained momentum: Micro combined heat and power, which is a technology that generates heat and electricity from the same energy source in individual homes. It is an engine-driven generator interconnected to the grid that runs on fossil fuel, the generator produces electricity and the waste heat goes to a heat exchanger which provides hot water.

If you want to experiment with Stirling engines, we have good news for you, as it is one of the simplest engines that you can build. After 200 years of history, it is not clear if we will ever see the Stirling take a major role in the engine world, or if it will stay in niche applications. We certainly hope for the first.

56 thoughts on “200 Years Of The Stirling Engine

  1. >” In 1986, the MOD II project produced an engine capable of 38.5% thermal efficiency (modern automotive engines are in the 20-25% range)”

    30-40% range actually. The article is confusing engine efficiency for whole drivetrain efficiency, where the latter is usually quoted by Stirling engine advocates, disingenuously comparing just the plain stirling engine to automobiles in general – or quoting completely out of date information.

    The Atkins cycle engine in a Toyota Prius for example achieves 31% efficiency over a wide range of loads. However, the whole drivetrain knocks off 15% in other losses, so the tank-to-wheel efficiency is around 23% – within the range often quoted.

    A Stirling engine in place of a car engine is subject to the same losses and non-optimal use, and in reality does not outperform a modern gasoline engine, far less a modern diesel engine. Small turbocharged diesels are pushing 40% and very large diesels surpass 50% efficiency, but again the drivetrain losses and other inefficiencies apply.

    1. The “wide range of loads” thing is what’s special about the atkinson cycle, typically, normal IC engine you get the 30% or better efficiency only at about 80% load at 80% of torque peak, and the rest of the time it can be as sucky as 10%. That’s all getting changed up because torque peak is determined by valve timing and cam, and variable timing in modern motors stretches out that “island” in the BSFC map so engine can be more efficient over a bit of a larger range.

      This is where heads get bent though, most motors are going to be more efficient in terms of work done, screaming up a hill at about 3000 RPM, vs chugging along in high gear at 1500 RPM on highway. That would be because you’re doing triple the work for maybe only double the fuel vs hwy cruise. From this we observe that really, this means your motor is too big and if sized so it was screaming along at 3000 rpm and at 80% of it’s capacity going down the highway, then it would be close to it’s ideal minimum fuel consumption per HP. But, it wouldn’t accelerate to that speed very fast and change the load or try to go up a steep hill and it’s either incapable or inefficient again. Ideally, a well designed displacement on demand system could put running part of motor into ~80% load state at highway cruise and give great efficiency. Consumer considerations and testing methods conspire against giving us a “perfect” version of this however. Due to the speed changes on EPA highway cycle tests it never makes more than an mpg of difference, to be able to show it’s worth it, and due to customers not liking to hear/feel the engine cutting cylinders in and out, it’s use is minimalised even in models that do use it. (Kinda like how they’re dumbing down CVTs to give the feel of gear changes now, and in the past how they slushed out auto transmissions to make them “buttery smooth” when mechanically, the best thing to do was slam frictions together as rapidly as possible and not make it try to be in two gears at once but gradually slipping between them.)

      1. Yes, but the point is that “modern” engines have evolved in general to the point that they are now at or above 30% efficient – everything from modified engine cycles to variable compression ratios, lean combustion etc.

        And no engine is optimized for “screaming at 3000 RPM” – they’re optimized for that 1500-2000 RPM range at cruising speed. At high RPM the gasses don’t have enough time to expand, so while you may be making more power the efficiency drops – at the other end at low RPM the heat escapes through conduction via the cylinder walls and that’s not good either. Going too fast increases friction losses as well.

          1. Nah, american consumers regard 3000 RPM as “screaming” it was tongue in cheek a bit. They’ll routinely break their trucks rather than shift down towing uphill.

        1. Gases not having enough expansion time isn’t reallllly a factor until 5 or 6 thousand RPM and depends on shit like stroke and rod ratio, strokers may be “there” at 5000 rpm, but undersquare short strokers may be way up near 10k before it happens.

        2. And yes they are optimised for about 3000, non-VVT motors, otherwise they’ve got no power band, bring torque peak down to 2000 and all it does is break the tires loose too easy trying to set off, and run out of breath by 4000. 3000 allows for reasonably low idle speed, reasonable pickup and useful operating range of 1000-6000 ish.

          Though VVT versions of those motors will manage to stretch that peak from 2000 to 5000 ish, and traction control and torque limiting nannies tame the low end tweakies.

    2. Something I have always wondered, could you use an electric motor and generator to replace the drivetrain? One motor for each of the two drive wheels will also remove the diff. It is my understanding that the electric motors they use have the torque from zero revs to not need gears.

      1. There are a bunch of companies/people trying to make in-wheel motors work. It’s not an easy problem – there’s not exactly a ton of space right behind a wheel, and the motors are obviously unsprung, so to a motor, it’s basically living in hell.

        1. A series hybrid means that the only engine that drives the wheels is an electric motor, and the gas engine just acts as a generator to supply the electric power. That’s not exactly what the poster was saying – those cars can still have a drivetrain because the engine is likely located under the hood, so you need some way to deliver the power to the wheels.

          Motors at the wheels obviously wouldn’t need much of a drivetrain at all, just power, but it moves the complication in the design to the motor.

      2. This is what diesel engine trains have done since since as early as 1925.

        Also, my Honda Accord has a CVT… more losses in the drivetrain but you maintain at the peak of the torque curve for longer leading to overall higher efficiency. I get about 35MPG on the highway to down around 21 as a minimum if I drive with a lead foot.

        …. Also building a Blazer with a 353 Detroit Diesel in it… which should have a much higher smiles per gallon efficiency rating. Especially running between 1500-2800 rpm :D (max RPM shuould be around 3400).

        1. My Toyota Corolla has a regular manual transmission and gets 35-38 MPG on the highway and the worst I’ve got in the city was around 27 MPG.

          Dunno what’s so special about the Honda CVT – just feels and sounds like driving a rubber band.

          1. And my bmw 530d gets 40 mpg around town and a sniff over 50 on a run, the girlfriends diesel golf is almost impossible to get below 55mpg amd regularly hits 60 on her commute (1.6 diesel) never understood why diesels in normal cars havnt took off in the states the fuel ecomomy over petrol counterparts is staggering, furthermore why the hell dont these hybrid cars have diesels lol i bet a prius with a 1.5 turbodiesel would make its own fuel.

          2. Mind, UK MPG is different to US MPG. 55 UK = 45 US

            Plus, there’s 10% more energy per gallon in diesel, so that 55 MPG-UK with diesel is equivalent to 41 MPG-US with gasoline which is just a regular economy car these days.

          3. I mostly agree… however, it is my daily driver and I drive a lot and in traffic so a manual isn’t as practical. The manual transmission achieves higher efficiency through less wasted power in the drive line itself, though it won’t transmit power at optimal torque consistently. Your corolla assuming it is a recent year model also has about 50 fewer horsepower, admittedly you probably don’t need the extra power since less of it is being wasted on the transmission. The CVT is quite boring but you get used to it… I just put some music on the radio and drive (I won’t be doing that in the 353 Blazer ha!)

            And I will point out that I do get the rated 28Mpg in city traffic when I’m not bashing on it.

      3. oops, hit the report comment by mistake…
        Large mining trucks…in the range of 700,000-1,000,000+lbs GVW use diesel electric hybrid drivetrains. There is a lot less mechanical complication in this arrangement, but there is a planetary gearset on each wheel motor.

        1. I was just thinking about this. (had to read R. G. LeTourneau’s autobiography to help someone study for a class. my understanding is the guy basically revolutionized heavy mining equipment and the individual electric motors for each wheel was one of his innovations.) could you replace the diesel in one of these with a sterling engine and some sort of capacitor/battery to compensate for the need to change speeds quickly, ie: have the engine running at constant speed while charging the battery/capacitor bank which then powers the motors at whatever speed they’d need to be running at? if nothing else, it would give you more flexibility in fuel options…

  2. >” One disadvantage was that the startup time was of around 30 seconds”

    That is only half the story. The main problem is the very poor throttle response of a stirling engine, because there’s a significant heat mass involved in the actual burner, which needs to get hot first before the actual engine can start drawing heat out of it. There’s essentially two throttles – one that controls the stirling engine itself and another that controls the burner that provides the heat. Before you can stomp on the throttle, you must turn the heater up and then accelerate, and likewise when you take your foot off the pedal, the burner is still going – there’s a massive lag in throttle response.

    That was also the reason for the poor practical efficiency of the engine – the mismatch between heat generated and heat used meant that the engine was repeatedly dumping the excess or maintaining too much heat production to better respond to a future throttle event. The transient response was far worse than with internal combustion engines.

    The 30 second startup time was just to the point where the car would move. Full power was not available until minutes later.

    1. Besides, the typical way to throttle a stirling engine is to change the amount of working fluid in the engine (pressure) and/or introduce more dead volume in the system which takes some of the gas inside out of action and reduces the amount of energy passed through per cycle, thereby throttling the output power.

      Both have the effect of dropping the engine efficiency, which makes the practical operation in a car difficult – the maximum power output can’t be much greater than the average power demand or else you’re continuosly running the engine far below its optimum, and so you can forget 100+ HP stirling engines since most cars cruise at around 25 HP.

      1. These issues could be mitigated by using a series-hybrid system with some local storage (probably ultracapacitors, low energy density requirement and high cycle count requirement). Have a control system that runs the stirling engine to maintain your average energy demand, with the caps and electric drive motors being used to allow good throttle response. However, these extra steps in the fuel -> wheels process will add some inefficiency, so in the end it’s probably no better than, say, a modern diesel engine, or hybrid, or gasoline engine with start-stop.

        1. Yes, but once you have a hybrid system then it’s no longer necesary to have a piston engine. A microturbine with heat recovery can reach the same efficiency with far fewer parts and a greatly simplified control scheme.

    2. Well said, all the relevant points I one post, so forgot any ideas for cars and trucks, a modern stirling has a similar weight and power output as a modern diesel engine but with shit acceleration, so we need a use for them that doesn’t need acceleration, such as, excavators, locomotives, grass cutting machines, generators te list goes on, so where are they?
      maybe fuel is still too cheap, beer and coke cost more!

      1. One major hurdle is power density. They do not have the same power to weight ratio, although it is possible with a caveat

        which is maintenance. To get high power, you need very high working fluid pressure, in the hundreds of bars, and the working fluid is a clean gas like nitrogen or hydrogen, sometimes CO2. You can’t just pull the engine apart in a garage or in the field, fix it, and expect it to work.

        1. One big issue with the power density is the heat exchanger.

          In an ICE the waste heat goes out the exhaust, but in a Stirling engine the waste heat has to be expelled by a radiator, and the efficiency of the engine depends on how well you can keep it cooled, so for high power output you also need a massive radiator and/or a big blower to push air through – especially if the engine is stationary. That pretty much kills the power density.

    1. Interesting article. But as silent as they are, it still should be possible to locate with a magnetic anomaly detector (such as on a P3 Orion) or even conventional sonar if on the move.

  3. I like the concentrated solar generator idea. Wonder if this could be done on a small scale for power generation on an RV to charge the battery bank when shore power is unavailable. Would be a hell of a lot quieter than an ICE generator.

  4. One of the casualties of Stirling marginalization was Coleman’s Stirling Power Cooler made in the early 2000s These were capable of near-cryogenic cooling while only drawing a few amps at 12V but the price tag (~$300) doomed them. There are similar models built for specialty operations that are currently in the $1200+ range but again the market is small when compared with thermoelectric/vapor compression units that are cheaper to buy, but are energy pigs.

    With some luck we may see Stirling fridges in our houses since refrigerators are big power consumers but again they’ll be expensive.


      1. I have a couple FLIR thermal camera units that have cryo coolers built in. Tiny little things, the modules were intended to go into their Recon III binoculars. In 6 minutes they will drop the sensor down to 77K.

  5. First, they are strling, not strling, that is a very common error.

    Second they are not really easy to build, at least in models. They generate very little power, so they cannot tolerate much drag, but they also cannot tolerate any leaks around the displacer piston rod and the power piston. I have built one that will run on the heat of your hand and a couple that run on flames. Model Internal combustion and steam engines are easier to build.

    A book was recently published with detailed drawings of the two early models Stiling built and donated to Universities in Scotland. I hope to build one of those someday.

  6. A small group of us in the UK are experimenting with Stirling engines for powering boats. We meet up every August to pootle up and down the River Thames. It’s an ideal application for Stirling Engines: they’re low noise, low vibration, low fumes (burning LPG), constant load (speed control of the boat is by two rudders used as a brake) and the river provides a practically limitless supply of cooling water. The low power is not a huge problem as the speed limit is 5mph anyway. If you’re interested, have a look at: https://www.youtube.com/watch?v=0QN8mXED0S8&t=3s

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.