2 ms·

- "For example, if we need a gigawatt of energy (10,000 people at 100 kW each) and space reactors weigh 150 T/MW, we’ll need to salami 150,000 T of reactors bet

- "For example, if we need a gigawatt of energy (10,000 people at 100 kW each) and space reactors weigh 150 T/MW, we’ll need to salami 150,000 T of reactors between 1500 Starship flights..." This is not a good estimate because the scaling laws you should be looking at are very strongly sub-linear. It's a major error to take a 10 kW design and flatly multiply it by 100,000x. This[0] is what mass-optimized, gigawatt-scale space nuclear reactors look like. They don't need 1,500 starships; they would fit inside one—they were designed to fit in one, because they are rocket propulsion engines. [0] https://en.wikipedia.org/wiki/Project_Rover https://en.wikipedia.org/wiki/Project_Rover ([late edit]: Anticipating the strongest criticism: these weren't electricity-generating reactors, true—but those subsystems can be highly miniaturized as well. The power density of gas turbines is incredible. A single Starship has 2 gigawatts of turbopump shaft power, distributed between its 30-something engine fuel pumps).

ceejayoz 2y ago
> This is what mass-optimized, gigawatt-scale space nuclear reactors look like. A nuclear rocket and a nuclear power plant are vastly different things, for the same reason a jet engine and an oil power plant look quite different. The reactor itself is a spatially small part of everything - the little doughnut in the middle. https://commons.wikimedia.org/wiki/File:EPR_1_EPR_2_perso.png https://commons.wikimedia.org/wiki/File:EPR_1_EPR_2_perso.pn...
- perihelions 2y ago
  I'm aware of that. But your commercial power plant example is also far from optimized for the requirements of a space power plant. ((edit): Your jet engine analogy is apt: a turbojet is a highly miniaturized version of the same type of turbine used in gas power plants. It'd be quite a mistake to look at a power plant turbine and analogize from that that jet airplanes are impossible, because, well, just look how huge those turbines are!)
  - ceejayoz 2y ago
    An in-space power plant isn't the proposal, though. It would be on Mars, and have similar requirements for containment, cooling, maintenance, etc. We are unlikely to want to spew radioactive hydrogen all over the landscape.
    - perihelions 2y ago
      It's a fascinating topic, isn't it? The hydrogen exhaust in nuclear thermal rockets isn't actually radioactive (well, in practice it was, because the rockets they actually built had a tendency to fall apart and eject bits of their nuclear fuel in the exhaust, but on paper that shouldn't happen). But the power-plant versions are closed gas loops anyway—you're driving the hydrogen or helium through a turbine, it's a sealed system. I don't think you'd choose to build any serious containment building on Mars. The baseline risk of death on Mars is extremely high; nuclear accident risks don't appreciably move the needle. My understanding is what you'd actually do is put a biological shield—maybe a pile of Martian rock—on the line-of-sight between the reactor and the astronaut base, and, simply, never walk that way. Cooling is substantially easier on Mars than in space, since there's an atmosphere that functions as a heat sink. The space version (i.e., in the nuclear electric propulsion context) has a more difficult challenge, with radiation as the sole heat dissipation mechanism. You're right that a space reactor would have to be designed for zero maintenance, else it'd be a non-starter.
  - PaulHoule 2y ago
    In theory you can build a low capital cost and compact nuclear reactor that operates at high temperature with a closed-cycle gas turbine powerset https://ntrs.nasa.gov/api/citations/20140016755/downloads/20140016755.pdf https://ntrs.nasa.gov/api/citations/20140016755/downloads/20... It could be a molten salt reactor or a liquid metal fast breeder reactor or a high temperature gas cooled reactor (carbide fuel in prismatic or pebble form) It could be highly competitive with solar as a carbon free energy source for terrestrial use but it is not a bird in the hand. (I think the robots in Gundam use a power plant like that which is why Zakus blow up when you hit them) I imagine the reactor, powerset and all, would be packed up into one Starship load but would have some civil works (cooling system) assembled on site. In another comment I point out Casey is accounting for 1 Starship launch of powerplant for 1 Starship load of colonists so this could scale OK from that point of view. Presumably they use a closed fuel cycle which is more like https://www.youtube.com/watch?v=KEfhx5ovuYk https://www.youtube.com/watch?v=KEfhx5ovuYk than Sellafield.
gamblor956 2y ago
Westinghouse recently announced a mini-nuclear reactor capable of powering several hundred homes (https://info.westinghousenuclear.com/news/first-canadian-evinci-microreactor-targeted-for-saskatchewan https://info.westinghousenuclear.com/news/first-canadian-evi...). The reactor itself is the size of small apartment and would fit aboard a spaceship (the now defunct Space Shuttle) and is light enough that a large rocket (like the Saturn) could carry it into space. The reactor plant would be constructed around the reactor at its final destination (in this particular case, Saskatchewan Canada, but for purposes of this discussion, Mars). As a side benefit of the design, it uses the heat of the reactor to provide heating, which a Mars colony would need a lot of.

ceejayoz 2y ago

> This is what mass-optimized, gigawatt-scale space nuclear reactors look like. A nuclear rocket and a nuclear power plant are vastly different things, for the same reason a jet engine and an oil power plant look quite different. The reactor itself is a spatially small part of everything - the little doughnut in the middle. https://commons.wikimedia.org/wiki/File:EPR_1_EPR_2_perso.png https://commons.wikimedia.org/wiki/File:EPR_1_EPR_2_perso.pn...

perihelions 2y ago
I'm aware of that. But your commercial power plant example is also far from optimized for the requirements of a space power plant. ((edit): Your jet engine analogy is apt: a turbojet is a highly miniaturized version of the same type of turbine used in gas power plants. It'd be quite a mistake to look at a power plant turbine and analogize from that that jet airplanes are impossible, because, well, just look how huge those turbines are!)
- ceejayoz 2y ago
  An in-space power plant isn't the proposal, though. It would be on Mars, and have similar requirements for containment, cooling, maintenance, etc. We are unlikely to want to spew radioactive hydrogen all over the landscape.
  - perihelions 2y ago
    It's a fascinating topic, isn't it? The hydrogen exhaust in nuclear thermal rockets isn't actually radioactive (well, in practice it was, because the rockets they actually built had a tendency to fall apart and eject bits of their nuclear fuel in the exhaust, but on paper that shouldn't happen). But the power-plant versions are closed gas loops anyway—you're driving the hydrogen or helium through a turbine, it's a sealed system. I don't think you'd choose to build any serious containment building on Mars. The baseline risk of death on Mars is extremely high; nuclear accident risks don't appreciably move the needle. My understanding is what you'd actually do is put a biological shield—maybe a pile of Martian rock—on the line-of-sight between the reactor and the astronaut base, and, simply, never walk that way. Cooling is substantially easier on Mars than in space, since there's an atmosphere that functions as a heat sink. The space version (i.e., in the nuclear electric propulsion context) has a more difficult challenge, with radiation as the sole heat dissipation mechanism. You're right that a space reactor would have to be designed for zero maintenance, else it'd be a non-starter.
- PaulHoule 2y ago
  In theory you can build a low capital cost and compact nuclear reactor that operates at high temperature with a closed-cycle gas turbine powerset https://ntrs.nasa.gov/api/citations/20140016755/downloads/20140016755.pdf https://ntrs.nasa.gov/api/citations/20140016755/downloads/20... It could be a molten salt reactor or a liquid metal fast breeder reactor or a high temperature gas cooled reactor (carbide fuel in prismatic or pebble form) It could be highly competitive with solar as a carbon free energy source for terrestrial use but it is not a bird in the hand. (I think the robots in Gundam use a power plant like that which is why Zakus blow up when you hit them) I imagine the reactor, powerset and all, would be packed up into one Starship load but would have some civil works (cooling system) assembled on site. In another comment I point out Casey is accounting for 1 Starship launch of powerplant for 1 Starship load of colonists so this could scale OK from that point of view. Presumably they use a closed fuel cycle which is more like https://www.youtube.com/watch?v=KEfhx5ovuYk https://www.youtube.com/watch?v=KEfhx5ovuYk than Sellafield.

gamblor956 2y ago

Westinghouse recently announced a mini-nuclear reactor capable of powering several hundred homes (https://info.westinghousenuclear.com/news/first-canadian-evinci-microreactor-targeted-for-saskatchewan https://info.westinghousenuclear.com/news/first-canadian-evi...). The reactor itself is the size of small apartment and would fit aboard a spaceship (the now defunct Space Shuttle) and is light enough that a large rocket (like the Saturn) could carry it into space. The reactor plant would be constructed around the reactor at its final destination (in this particular case, Saskatchewan Canada, but for purposes of this discussion, Mars). As a side benefit of the design, it uses the heat of the reactor to provide heating, which a Mars colony would need a lot of.