#stirling engine

Fig. 6-8. Zwauer-Wankel diagram of a Displacer-free rotary Stirling engine.

A Stirling engine has a lower power-to-weight ratio than an internal combustion engine, but how bad is it? Well, the 1986 NASA project said they had gotten it to 3.35 kg/kW. (Has there been any progress since 1986 on Stirling engines? I don’t know. There have been some new developments with thermoacoustic Stirling engines and such, but I’m not sure they’d be any better than traditional Stirling engines for this application.) With a hybrid, the most sensible approach is to make the heavy Stirling engine as weak as it can be while sustaining a desired top cruising speed – bursts of acceleration are better handled by electrical motors with better power-to-weight ratios.

It’s commonly said that a car only needs maybe 20 hp (15 kW) to cruise at a highway speed of, say, 60 mph/100 kmh. That’s not very marketable, though. If we supposed that a top sustainable speed – the top speed at which you’re burning fuel, but not draining your batteries – should be more like 90 mph, then we might want 1.5^3 = 3.375 times that much power (this is on the pessimistic assumption that air resistance is basically all resistance, since it scales worst (cubically) with speed). Other situations besides speeding, like towing or going up a long-enough upward incline that it wouldn’t just be cleared on battery alone, might also be occasions for sustained power higher than the ~15 kW highway level.  So maybe 50 kW. A stirling engine capable of producing that might weigh 170 kg.

For our electric motors, we might want, say, 150 kW, for a marketably peppy response. The motor of a Tesla Model 3 (cited in Wikipedia on a list of power densities) has got 6.26 kW/kg.  That might require 30 kg. If we’re going for a serial hybrid rather than a parallel hybrid – with no direct connection between the Stirling engine and the drivetrain, only a connection that bypasses the battery to feed electrical power to the motors directly – then the power going to the tires is just from the electric motors, not the Stirling engine. This might simplify the engineering, though..

Then there are the batteries. If we wanted four minutes of battery power at near-maximum power  – making sure to compensate for any lag in the Stirling engine heating up – then we’d want ~100 kW * ~250 seconds = 25 MJ. This is also close to the minimum for an American tax-break eligible plug-in hybrid electric vehicle, which requires a battery of at least 7 kWh, or 25.2 MJ. If we got 0.4 MJ/kg then that’s another 65 kg. One question is whether the battery has enough power, not just enough energy. I am having a hard time finding sources on this. Wikipedia pages on power density seem too pessimistic given what we see out of electric vehicles like the Model S Plaid. (100 kWh battery, 1020 hp ~=760 kW, suggesting that a battery can discharge itself in 8 minutes at most, depending on whether the Plaid is basically designed with motors that are as powerful as possible given its battery.) We might need to double the size of the batteries to power the motors enough.

So we’re adding 265-340 kg to the vehicle so far. There will probably be a bit more – regenerative braking systems, power distributing linkages, etc. On the other hand, we’ve gotta remember that we’re also taking the internal combustion engine out.

Is 280 kg too much? I suppose a comparison might be plug-in hybrids, which are rather heavy on account of an ICE-electric hybrid energy system and a rather large battery. A Subaru Crosstrek conventional has a curb weight of 3298 lbs, while the PHEV weighs in at 3,717 lbs, a difference of 190 kg. The Kia Sportage has a curb weight of 3,373 lbs, while the PHEV version weighs 4211, a difference of 380 kg.

Perhaps with some sort of modular system could add versatility to the fueling types, with different modules for gasoline, compressed natural gas or hydrogen, waste vegetable oil (perhaps a preheater would be good for something that viscous). The Stirling engine might lose a little peak power on some fuels relative to gasoline, but for ordinary, not-too-fast, not-too-hilly cruising that might just mean it runs more and it doesn’t affect the performance too much – although in times of persistent high demand for power, like climbing up a hill, towing a trailer, or just speeding, you might notice the deficiency.

One possible issue is hydrogen embrittlement. If the Stirling engine has a working fluid other than hydrogen, it would be less efficient, while if it’s using materials that are resistant to hydrogen embrittlement, then that might make it less rugged and heavier. I think they used sufficiently embrittlement-resistant stainless steel and hydrogen as a working fluid in the 1986 NASA project, but one way to cheat for a short-term test of a relatively powerful, light Stirling engine would be to just use hydrogen as a working fluid and regular steel engine parts, and if they crack like a week later, that’s after your tests are done.

I don’t really know that much about the details of obstacles to alternative ideas for cars, though. I guess I’ve been a bit car-brained since I became a car owner almost a year ago. I suspect that when you first get a job as an engineer at Ford or Honda, the first thing in your inbox is a Powerpoint attachment titled “Here’s Why We’re Not Going to Consider Your Stupid Alternative Powertrain Ideas.” Maybe if someone could leak that to me that would give me some clarity.

I guess Dean Kamen tried to get some kind of Stirling-based car going, and there's some guy, Josh MacDowell, who turned a Ford F-150 into a Stirling-Electric hybrid? Given that there's no IP protection on Stirling engines, though, I think there must be more practical obstacles that I'm not considering.