When does nuclear thermal propulsion outperform chemical propulsion?

[mark_bose]

TL;DR Summary

Is nuclear thermal propulsion worth to be developed? advantages and disadvantages with respect old good chemical rockets.

Dear aerospace guys,

From time to time, I think about nuclear thermal propulsion. As a nuclear enthusiast, I'd love to see nuclear reactors in space.

Regarding Nuclear Thermal Propulsion (NTP), I understand that it theoretically offers a higher specific impulse compared to chemical propulsion (a bit more than double ?), but with lower thrust. I also know that, because of this, missions like reaching Mars would either take less time or use less propellant. However, I'm not an expert in the field, and I wonder whether it really is advantageous to use NTP, especially considering the following points:

1. Although NTP was developed in the past (e.g., NERVA project), building space nuclear reactors is still quite complex and expensive (low TRL).
2. If the amount of propellant is the problem, with in-orbit refueling, chemical propulsion might still be more competitive (or not?).
3. Public opinion and regulations around nuclear technology can be problematic.

That being said, I believe NTP should be reserved for missions that are truly impossible to accomplish with other propulsion technologies, rather than just being slightly more advantageous. I'd love to understand more about this, so I'm curious if anyone knows of any missions that cannot be accomplished with chemical propulsion (now or in the near future) or if I'm completely off the mark here.

 

[russ_watters]

Specific impulse is indeed the answer:
https://en.m.wikipedia.org/wiki/Specific_impulse

 

[Flyboy]

The specific impulse, especially when using LH2, is hard to beat, and for missions with high dV requirements, such as planetary exploration, the increased performance and lower propellant mass requirements make it worth looking into.

The costs and drawbacks, though, are why we haven’t seen one fly yet. The large propellant tank volumes, the radiation hazards if you leave the shadow shielded area shortly after a burn, the significantly higher cost to manufacture the stage… I don’t know if the public is quite ready to shove the nuclear power boogeyman into the trashcan, even (or especially) for the space program.

 

[Tom.G]

Rocket launches, and landings, do fail on occassion.

You are then looking at massive contamination that is not compatible with life, and that you may or may not be able to clean up.

Even if used only as a power source in satellites, they will eventually come down and burn up in the atmosphere - hope it isn't where you live.

 

[Flyboy]

Rocket launches, and landings, do fail on occassion.

You are then looking at massive contamination that is not compatible with life, and that you may or may not be able to clean up.

Even if used only as a power source in satellites, they will eventually come down and burn up in the atmosphere - hope it isn't where you live.

Wasn’t there some Soviet RORSAT with a nuclear reactor that came down in Canada? Can’t recall the exact details off the top of my head…

 

[mfb]

Rocket launches, and landings, do fail on occassion.

You are then looking at massive contamination that is not compatible with life, and that you may or may not be able to clean up.

Even if used only as a power source in satellites, they will eventually come down and burn up in the atmosphere - hope it isn't where you live.

Nuclear reactors are essentially inert before you turn them on. It's just a chunk of uranium. There is no application for nuclear thermal propulsion in low Earth orbit so they will not deorbit over time.

In-orbit refueling hasn't been done on a large scale yet either, but Starship should change that soon.

If you need the highest I_sp then ion thrusters (powered by solar or nuclear) win by a large margin. NTP only wins by thrust, i.e. if there is some time constraint on the mission. It's not an accident that crewed missions to Mars are discussed the most. A crew needs significant electric power as well, so having the reactor already on board is nice.
I don't see anything competing with Starship in the near future, however. By the time an NTP ship might fly, Starship refueling (enough to reach Mars quickly with chemical propulsion) should already be routine.

 

[mark_bose]

Thanks for your replies, I still have some doubts

Specific impulse is indeed the answer:
https://en.m.wikipedia.org/wiki/Specific_impulse

If that was the only reason then why don't we go directly with nucelar electric instead of complicate things with NTP? Isp is much larger.

The specific impulse, especially when using LH2, is hard to beat, and for missions with high dV requirements, such as planetary exploration, the increased performance and lower propellant mass requirements make it worth looking into.

The costs and drawbacks, though, are why we haven’t seen one fly yet. The large propellant tank volumes, the radiation hazards if you leave the shadow shielded area shortly after a burn, the significantly higher cost to manufacture the stage… I don’t know if the public is quite ready to shove the nuclear power boogeyman into the trashcan, even (or especially) for the space program.

I'm not really an expert on this, but I imagine that provided we have sufficient propellant with chemical propulsion, we could accelerate enough to reach the required ΔV, couldn't we? In that case, wouldn't chemical propulsion combined with in-orbit refueling be easier than nuclear? Is there any mission that would require so much propellant that it would be physically impossible even with in-orbit refueling?

Rocket launches, and landings, do fail on occassion.

You are then looking at massive contamination that is not compatible with life, and that you may or may not be able to clean up.

Even if used only as a power source in satellites, they will eventually come down and burn up in the atmosphere - hope it isn't where you live.

Yeah, that was one of my concerns. However, as far as I know, radioisotope thermal generators have been launched in the past, and despite being smaller than nuclear reactors, their radiotoxicity is much higher (at least if the reactor has not operated before the accident). But it is true that RTGs are smaller, and people might be less concerned.

 

[Astronuc]

If that was the only reason then why don't we go directly with nuclear electric instead of complicate things with NTP? Isp is much larger.

Mass. In NTP, one simply pumps H₂ through the core then out through a nozzle. In nuclear electric systems, there is the core, power conversion system (including circulation system, and possibly heat exchanger, turbine, generator, power converter, battery of propulsion devices, AND a massive radiator). One has to trade off thermodynamic efficiency where one wants the lowest possible temperature for high Carnot effiiciency against the radiator temperature, which one has to maximize to minimize radiator area and mass. Using cryogenic propellant as a heat sink in a hybrid system is one possibility for reducibg the size of the radiator, but that requires an added challege of propellant mass flow (and storage) with required thrust requirements (and ability to turn on/off propulsion) - as well as technical challenges related to propellant/structure interactions.

but I imagine that provided we have sufficient propellant with chemical propulsion, we could accelerate enough to reach the required ΔV, couldn't we? In that case, wouldn't chemical propulsion combined with in-orbit refueling be easier than nuclear?

On-orbit refueling requires one launch propellant form earth surface to orbit, which is quite expensive, although the cost/kg has come down since 4 decades ago. It is not practical, especially if one is sending a nuclear powered spacecraft to Mars or further out in the solar system.

Is there any mission that would require so much propellant that it would be physically impossible even with in-orbit refueling?

That depends on the payload and mission parameters. How big a payload, and how quickly one wants to arrive (days vs weeks vs months vs years vs decades).

The initial motivation behind NTP was propulsion of ballistic missiles designed to carry the first generation of thermonuclear warheads, which were massive, and the chemical rockets of the 1950s just weren't capable. Things changed payloads got smaller and chemical rockets more powerful.

There is no application for nuclear thermal propulsion in low Earth orbit so they will not deorbit over time.

This was a topic of discussion during a seminar on VLEO (vs LEO) applications. I have to look into the regulatory details, but it seems there is an 'altitude limit' below which nuclear is not considered due to safety/reliability concerns, and chemical works well.

Wasn’t there some Soviet RORSAT with a nuclear reactor that came down in Canada? Can’t recall the exact details off the top of my head…

Kosmos 954, part of the Russian RORSAT program.
https://en.wikipedia.org/wiki/Kosmos_954

 

[mark_bose]

Mass. In NTP, one simply pumps H₂ through the core then out through a nozzle. In nuclear electric systems, there is the core, power conversion system (including circulation system, and possibly heat exchanger, turbine, generator, power converter, battery of propulsion devices, AND a massive radiator).

It's true that, if we compare NTP and NEP reactor systems (at the same power), the latter would be heavier for the reason you mentioned. However, Isp in NEP is much higher than in NTP, shouldn't this allow to use less propellant and compensate for the heavier power system?
Moreover, as far as I know, NEP uses reactors with relatively low power. For instance, the SP-100 (2.4MWt, 100 kWe) weight is estimated around 4500 kg including shielding, energy conversion system and radiator, while the reactor only weights something like 650 kg. For comparison, in this publication are estimated the masses of NERVA-like reactors that turned out to be 2-4 times higher (1300-2600 kg). It's true that It's not really a fair comparison since thermal power in NERVA is probably much higher than in SP-100, but that impacts also the mass of the shielding, that probably would be havier compared to the one used for NEP.

So, wrapping up:
-The Isp alone does not explain the advantage of NTP over chemical, otherwise one could direcly go NEP.
-It seems to me that it is debatable the absolute need of NTP for mars mission. Space-x claims to reach mars in 6 months using in-orbit refueling, similar to what can be achieved with NTP.
-Maybe NTP has not competitors only beyond mars. In that case, I suppose that In-(earth)orbit refueling is not sufficient for chemical and would be necessary to make several stops around some lunar/martian refueling station. But I still have to hear a convincing argument for that.

 

[Flyboy]

NTP strikes a good balance between the light weight and raw thrust of a chemical propulsion engine, and the high efficiency but low thrust of electric propulsion.

The issue is that you have to trade one kind of performance for another. A big kick in the rear to make an inefficient but quick burn, or a long, gentle, but efficient burn.

 

[Astronuc]

Moreover, as far as I know, NEP uses reactors with relatively low power. For instance, the SP-100 (2.4MWt, 100 kWe) weight is estimated around 4500 kg including shielding, energy conversion system and radiator, while the reactor only weights something like 650 kg. For comparison, in this publication are estimated the masses of NERVA-like reactors that turned out to be 2-4 times higher (1300-2600 kg). It's true that It's not really a fair comparison since thermal power in NERVA is probably much higher than in SP-100, but that impacts also the mass of the shielding, that probably would be heavier compared to the one used for NEP.

40 years ago, I did a core design for a 300 MWt compact fast reactor (for an NEP system) using (U,Pu)N, with ²³⁹Pu content of 51%-45% radially zoned to flatten the peak power. The cladding was a W-Mo-Re alloy; reflectors were a YBe alloy. It's been a long time since I looked at the design, but I believe the mass was on the order of 10 tonne. The core was roughly a right circular cylinder (hexagonal geometry) with height of 1 m and diameter on the order of 1 m. One could use DUN blankets to provide some shielding. The primary coolant was Li, while the secondary coolant loop used K to drive a K-Rankine cycle.

I've learned a lot since then, and I would do some things differently now. The technology and understanding of materials has greatly improved, but on the other hand, not much progress has been made since SP-100.

A good metric is the kWt/kg (the higher the better) or conversely kg/kWt (the lower the better).

A Wikipedia article is reasonably accurate.
https://en.wikipedia.org/wiki/NERVA#Reactor_and_engine_test_summary