Most larger liquid-fuelled rocket engines1 use fuel and oxidiser turbopumps driven by the hot, high-pressure gasses produced by burning said fuel and oxidiser:

  • The gas-generator cycle burns some of the fuel and oxidiser in its namesake gas generator to produce hot, high-pressure gasses; these gasses are then used to drive the engine’s turbopumps, before being dumped overboard.
  • The staged-combustion cycle is similar to the gas-generator cycle, but the gasses from the gas generator (now frequently known instead as a preburner), after being used to drive the turbopumps, are fed back into the main combustion chamber rather than being dumped overboard.
  • The combustion-chamber tapoff cycle drives the turbopumps with gasses from the main combustion chamber, rather than using a separate gas generator/preburner.

The fuel turbopumps are fairly easy to deal with; they are almost2 always driven by fuel-rich combustion gasses, which are friendly to engine plumbing and turbomachinery and don’t pose a risk of reacting with the fuel if they leak past the pump seals and come into contact with said fuel.

The oxidiser turbopumps, however, are a very different beast. When producing hot, high-pressure combustion gasses to drive an oxidiser turbopump, there are basically three options, each with its own problems:

  1. The gas generator can run at the stoichiometric mixture ratio (just enough oxidiser to completely burn the fuel provided, with no extra of either reactant), which has a habit of melting the pump’s turbine wheels.
  2. The gas generator can run rich (more fuel and less oxidiser than stoichiometric), which requires really good sealing to prevent the combustion gasses from leaking past the pump seals, coming into contact with the oxidiser, and exploding (or vice versa); this generally necessitates the use of two sets of seals, with the space between the two filled with a nonreactive gas3 at positive pressure relative to both the combustion gasses and the oxidiser.
  3. The gas generator can run lean (less fuel and more oxidiser than stoichiometric), which produces large amounts of superheated oxidiser-rich gasses, which are hideously-difficult to deal with, due to their tendency to eat engine plumbing and turbomachinery.

Physically separating the oxidiser pump from the turbomachinery powering it would allow the turbopump to be driven by docile fuel-rich gasses without needing a complex gas-purged seal system; one way of doing this would be to use a turboelectric-drive system, with the combustion-gas-driven turbine driving an electrical generator and the resulting electricity being used to power an electric motor driving the oxidiser turbopump. The electrical transmission between the turbine and the pump would add some mass and slightly reduce the pump’s efficiency (though not by much - well-designed electric motors and generators can have conversion efficiencies well north of 90%), but would eliminate the need for a complicated and heavy gas-purge system or for difficult-to-engineer superheated-oxidiser-containment-and-transport equipment.

Why do no liquid-fuelled rocket engines, to the best of my knowledge, use turboelectric-drive oxidiser turbopumps?

1: Smaller liquid-fuelled engines tend to use the pressure-fed or expander cycles, which are very simple but scale poorly, while the newer electric-pump-fed cycle requires heavy battery banks to drive its turbopumps.

2: A very few engines use oxidiser-rich combustion gasses to drive their fuel turbopumps; these are almost exclusively those that already use oxidiser-rich gasses for the oxidiser turbopumps. This arrangement is extremely uncommon, as it combines the disadvantages of oxidiser-turbopump drive methods 2 and 3 without the benefits of either.

3: Generally helium, which is extremely light and almost completely inert (although also extremely expensive).

  • 2
    $\begingroup$ "eliminate the need for a complicated and heavy gas-purge system" by adding a complicated and heavy generator and motor? $\endgroup$ – Organic Marble Sep 14 at 3:24
  • 1
    $\begingroup$ Just because the gas seal is easier to handle than the electric motors. The cooling of electric motor is just as hard to handle, if not more and certainly much more heavy. Maybe hrdrolic coupling is slightly better? $\endgroup$ – user3528438 Sep 14 at 4:51
  • 1
    $\begingroup$ @OrganicMarble electric motors are much simpler, although maybe heavier, and for the cooling part, you could just syphon off a small amount of the fuel/oxidizer that already passing through the pump anyway. $\endgroup$ – Reuben Farley-Hall Sep 14 at 5:19
  • 1
    $\begingroup$ I certainly wouldn't call oxidizer-rich staged combustion "extremely uncommon", given that it powers most of the Soviet/Russian launch vehicles (essentially everything except for Soyuz) and the Chinese space program (Long March 5-7) and is becoming more popular among US launch vehicles as well: Atlas III/V on the RD-170 family, Antares using NK-33s, Vulcan and New Glenn using BE-4. $\endgroup$ – TooTea Sep 14 at 12:20
  • $\begingroup$ @ReubenFarley-Hall I admire your optimism that a generator/motor "delivering as much horsepower as 28 locomotives" and operating in a savage thermal environment is simple. Admire, but do not share. $\endgroup$ – Organic Marble Sep 14 at 13:48

Working off the Falcon pump power of 7500 kW and power to weight of 10 kW/kg (electric aircraft engine) and assuming generator weighs the same as the motor, that gives a best case motor/generator set for a merlin engine of 1500kg, on a existing engine mass of 750kg so certainly not lighter than the seals and associated plumbing.

Some possible cases where it might make sense is beyond earth Orbit, where things like the ability to start the engine on electric power, fine speed control, possible cross connect for redundancy and ability to just run the generator for electricity might conceivably make up for that mass penalty but these start getting complicated to calculate.

| improve this answer | |
  • $\begingroup$ Do we have some additional sources for the 7500kW (18 kW/kN) number? Electron's Rutherford engine is cited with 6 times less power 74 kW (3 kW/kN). $\endgroup$ – asdfex Sep 14 at 12:44
  • $\begingroup$ @asdfex I'm wondering if that is the combined power of all 9 engines, making 2kW/kN and a 83kg motor generator set per 750kg Merlin, and somewhere around 7.4kg of electric motor on a Rutherford, all of which seem more reasonable. $\endgroup$ – GremlinWranger Sep 14 at 13:35
  • $\begingroup$ Yes. Just one of the 4 turbopumps in the one SSME "delivered as much horsepower as 28 locomotives". (alternatewars.com/BBOW/Space_Engines/…) Might be kind of a big electric motor. $\endgroup$ – Organic Marble Sep 14 at 13:46
  • 1
    $\begingroup$ This SSME link gives us another point at 25 kW/kN for the fuel pump only. In total about 3 times more than the Merlin and a factor 18 above the Rutherford. We have to keep in mind that it's not flow, but flow times pressure that determines the power, but the same is true for thrust... $\endgroup$ – asdfex Sep 14 at 14:12
  • $\begingroup$ For J-2X I found 2x 10 kW/kN. I guess our main problem is that one lists power on the driving shaft and others list the total thermodynamic power. There's a factor of 4 between them due to efficiency. But this also invalidates the comparisons to locomotives $\endgroup$ – asdfex Sep 14 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.