Nuclear Thermal Rocket (NTR) propulsion

[Astronuc]

NASA, DARPA Will Test Nuclear Engine for Future Mars Missions - But then that's been stated periodically in the past.
https://www.nasa.gov/press-release/nasa-darpa-will-test-nuclear-engine-for-future-mars-missions

NASA and DOE are working another commercial design effort to advance higher temperature fission fuels and reactor designs as part of a nuclear thermal propulsion engine. These design efforts are still under development to support a longer-range goal for increased engine performance and will not be used for the DRACO engine.

Obviously we won't be sending a manned mission to Mars by 2020 or 2024, and likely not before 2030, if ever.
https://www.physicsforums.com/threads/manned-mars-mission-to-mars-before-2020.243659/

See some history here
https://www.physicsforums.com/threads/the-nuclear-power-thread.9091/post-6879441

 

[Andrew Mason]

Interesting post and links. I was a bit surprised to see that a nuclear powered rocket engine was not only conceived as an idea but was actually developed by NASA over a 20 year period ending in 1973 (Project NERVA). As I understand the NERVA Wikipedia article, NERVA worked well and was fully developed. The only thing it lacked was the political nerve of Congress to continue funding it. It is being considered again as an engine for interplanetary travel.

The main advantage of the nuclear rocket is the ability to provide a greater specific impulse than chemical rockets. This means they can propel a spacecraft to much greater speeds than chemical rockets with the same amount of working mass. If the specific impulse is twice that of conventional rockets, the momentum gain per kg of working mass is more than double and the travel time would be less than half that of conventional rockets.

One problem that is not really explained, is how the core of the reactor of the rocket engine on a spacecraft would kept cool. Liquid hydrogen acted as both coolant and working mass. At some point, the craft would either run out of hydrogen or it would be too warm. I can see liquid hydrogen cooling the reactor as it passes through the rocket and is expelled at high speed. But when the liquid hydrogen ran out, there would be residual heat from the radioactive core that would still need to be cooled, I would think.

AM

 

[Filip Larsen]

If the specific impulse is twice that of conventional rockets, the momentum gain per kg of working mass is more than double and the travel time would be less than half that of conventional rockets.

This doesn't sound right. With double the ejection speed you get double delta-V for same mass ratioes, but if you want to go Mars you still have to coast in an (almost) Hohmann transfer orbit between Earth and Mars for a more or less a fixed amount of time. I am not familiar with the details of NERVA, but I would think that a more efficient engine would primarily be used to allow for a higher payload mass ratio.

 

[gmax137]

But when the liquid hydrogen ran out, there would be residual heat from the radioactive core that would still need to be cooled, I would think.

Interesting question! I skimmed thru the wiki article, and it seems the longest test run was 1199 MW for 3620 seconds. Short reactor operating time will result in a rapidly decreasing residual heat. So maybe some kind of radiator system would be adequate to dump the heat during a mission?

I don't know what the operational plans were (how long the engine would run during a mission).

 

[Andrew Mason]

This doesn't sound right. With double the ejection speed you get double delta-V for same mass ratioes, but if you want to go Mars you still have to coast in an (almost) Hohmann transfer orbit between Earth and Mars for a more or less a fixed amount of time. I am not familiar with the details of NERVA, but I would think that a more efficient engine would primarily be used to allow for a higher payload mass ratio.

I think the idea is that the spacecraft would not have to travel as far and would use a non-Hohmann transfer orbit. This would require a higher Δv to enter the non-Hohmann orbit and a greater negative Δv to enter Martian orbit but the total distance is much less. See: https://orbital-mechanics.space/orbital-maneuvers/non-hohmann-transfers.html

It is not as efficient as the Hohmann transfer but with human beings as part of the payload, the shorter time would be worth it.

AM

 

[Filip Larsen]

It is not as efficient as the Hohmann transfer but with human beings as part of the payload, the shorter time would be worth it.

True. It's just that I have never seen a serious near-future mission design that do not go for near-minimum transfer energy (delta-V) requirement which, as I understand it, is simply because not doing so makes the design more likely to fail in practice (e.g. more narrow launch windows, more narrow cargo selection reducing redundancy, etc).

Of course, with double the specific impulse it doesn't have to be a complete either-or choice between low transfer time and energy, a mix of the two is probably also quite feasible. But, still I don't see the business case⁽¹⁾ for low transfer time for, say, Earth-to-Mars missions considering the crew will spend years training for the mission anyway and you would have to sacrifice a lot of payload "just to save" a month or two. To put some number on it, for a final-to-initial mass ratio in range 0.5 to 0.1, the relative loss in final mass capacity if "converting" a doubling of the specific impulse fully into saving transfer time would be in the range 40% to 216% of the initial mass.

(1) I am not aware of the current understanding of the radiation hazard on a manned Earth-to-Mars mission, but radiation hazard could potentially be an argument for reducing crew transfer time if it somehow can be shown that the extra mass you need to safe-guard crew on a longer coast phase can more than be saved by going faster.

 

[Andrew Mason]

True. It's just that I have never seen a serious near-future mission design that do not go for near-minimum transfer energy (delta-V) requirement which, as I understand it, is simply because not doing so makes the design more likely to fail in practice (e.g. more narrow launch windows, more narrow cargo selection reducing redundancy, etc).

NASA has a web page calculator for such trajectories. The first calculation for an Earth-Mars trajectory using a maximum Δv of 4.6 km/s for conventional chemical rockets and these constraints to minimize duration:

resulted in trajectories taking from 304 to 336 days:

The second calculation for an Earth-Mars trajectory doubling the maximum Δv to 9.2 km/s for nuclear NERVA type-rocket and these constraints:

resulted in trajectories taking from 144 to 176 days:

(1) I am not aware of the current understanding of the radiation hazard on a manned Earth-to-Mars mission, but radiation hazard could potentially be an argument for reducing crew transfer time if it somehow can be shown that the extra mass you need to safe-guard crew on a longer coast phase can more than be saved by going faster.

Here is an interesting Nature paper on the potential health hazards of long duration Mars space travel: https://www.nature.com/articles/s41526-020-00124-6

AM

 

[Filip Larsen]

The second calculation for an Earth-Mars trajectory doubling the maximum Δv to 9.2 km/s

The trajectories exists for sure, and the duration can get even shorter if waiting for the right launch window is allowed, that is, if the high-energy transfers you listed that costs around 45 km2/s2 can be done as energy optimal transfer with nearly same duration at a different launch date or, in other words, by picking a different launch date you could get even lower transfer times with a high-energy transfer.

Also, with aerocapture (i.e. using fly-by instead of rendezvous in the tool) the number of low-duration transfer opportunities in the next couple of decades seems increases a lot (although some of them might have too high arrival speed for aerocapture), so that might also be an option, at least to the extend aerocapture can be considered a realistic capture option for a crewed vehicle.

By the way, regarding opportunities there is also a nice collection of Earth-Mars transfer opportunities listed in https://ntrs.nasa.gov/api/citations/20100037210/downloads/20100037210.pdf that can be used to ball-park energies and transfer times.

But all that said it would still be interesting to see a serious mission design that actually planned with a short duration crew transfer (it may very well exists, I am just not aware of any). I assume such design would have to be a multi-vehicle/phase mission where equipment needed for exploration on Mars are placed in orbit or on the Mars surface before arrival of the crew, as this is likely to be a flexible (albeit expensive) mission design template even without access to NERVA or similar effective propulsion, and this is (as I understand it) also the current template used by NASA and others for manned Mars mission designs.

 

[sophiecentaur]

The only thing it lacked was the political nerve of Congress to continue funding it. It is being considered again as an engine for interplanetary travel.

I find that delay strangely comforting. You have to hope they will wait a suitable time before starting practical experiments. Maybe after they can experiment on a robotic space station, a long way away.

One problem that is not really explained, is how the core of the reactor of the rocket engine on a spacecraft would kept cool.

What you need is a reactor that operates at 'red heat' temperatures which would shift a reasonable amount of heat energy (many kW/m²). Looks pretty inconvenient to engineer.

 

[Andrew Mason]

The trajectories exists for sure, and the duration can get even shorter if waiting for the right launch window is allowed, that is, if the high-energy transfers you listed that costs around 45 km2/s2 can be done as energy optimal transfer with nearly same duration at a different launch date or, in other words, by picking a different launch date you could get even lower transfer times with a high-energy transfer.

Also, with aerocapture (i.e. using fly-by instead of rendezvous in the tool) the number of low-duration transfer opportunities in the next couple of decades seems increases a lot (although some of them might have too high arrival speed for aerocapture), so that might also be an option, at least to the extend aerocapture can be considered a realistic capture option for a crewed vehicle.

By the way, regarding opportunities there is also a nice collection of Earth-Mars transfer opportunities listed in https://ntrs.nasa.gov/api/citations/20100037210/downloads/20100037210.pdf that can be used to ball-park energies and transfer times.

Using the link you provided, I was able to find the quickest trip to Mars which has a launch date of May 19, 2033 and only takes 96 days:

There is a slightly longer trip - lasting 112 days - leaving the same day but with a smaller total Δv:

The higher eccentricity of Mars' orbit and the almost 2° inclination of its orbital plane relative to the Earth's orbital plane make the timing of the launch particularly important.

But all that said it would still be interesting to see a serious mission design that actually planned with a short duration crew transfer (it may very well exists, I am just not aware of any). I assume such design would have to be a multi-vehicle/phase mission where equipment needed for exploration on Mars are placed in orbit or on the Mars surface before arrival of the crew, as this is likely to be a flexible (albeit expensive) mission design template even without access to NERVA or similar effective propulsion, and this is (as I understand it) also the current template used by NASA and others for manned Mars mission designs.

Assuming there is enough water on Mars, they could produce their working mass or fuel for the return travel. With a nuclear engine, it would be a matter of producing liquid hydrogen on Mars in order to get back to Earth - probably set up an automated system for production before the first manned mission. (I am assuming there would be preceding unmanned missions to set up the infrastructure needed to support the manned mission, including fuel/working mass production for the return voyage to Earth). I noticed that for the return trip, assuming a total round trip duration of 2 years or less does not leave a great deal of time on Mars to make fuel - most missions allow only 30 days on Mars using these parameters.

AM

 

[Filip Larsen]

Eric Berger at Arstechnica just had a piece relevant to this thread:
https://arstechnica.com/space/2023/...-nuclear-powered-rocket-engine-in-four-years/

 

[Flyboy]

One problem that is not really explained, is how the core of the reactor of the rocket engine on a spacecraft would kept cool. Liquid hydrogen acted as both coolant and working mass. At some point, the craft would either run out of hydrogen or it would be too warm. I can see liquid hydrogen cooling the reactor as it passes through the rocket and is expelled at high speed. But when the liquid hydrogen ran out, there would be residual heat from the radioactive core that would still need to be cooled, I would think.

AM

Interesting question! I skimmed thru the wiki article, and it seems the longest test run was 1199 MW for 3620 seconds. Short reactor operating time will result in a rapidly decreasing residual heat. So maybe some kind of radiator system would be adequate to dump the heat during a mission?

I don't know what the operational plans were (how long the engine would run during a mission).

What you need is a reactor that operates at 'red heat' temperatures which would shift a reasonable amount of heat energy (many kW/m²). Looks pretty inconvenient to engineer.

To answer these questions all at once, if memory serves the burn would also include a 30 minute cooldown phase where they put the reactor subcritical and flow a reduced amount of LH2 into the engine, It would then take away any residual heat. I'll have to dig around for the exact reference I stumbled across that said that, though.

 

[gmax137]

take away any residual heat

In nuclear engineering, "residual heat" refers to the heat released by the fission products. It decreases exponentially once the reactor is subcritical, but it never really goes to zero. So the question as I understood it, is how the design removes this energy in the long term.

I'll have to dig around for the exact reference I stumbled across that said that, though.

If you do turn up the reference, please come back and post!

EDIT: I will have to come back and clean this up. There is residual heat and decay heat; they are not exactly the same thing.

 

[gmax137]

https://en.wikipedia.org/wiki/Decay_heat provides a pretty good discussion of decay heat.

"Residual heat" usually refers to the total energy release, which is the sum of the decay heat and the heat from the fissions which occur after shutdown. Fission continues with the reactor subcritical, due to the delayed neutrons. See https://en.wikipedia.org/wiki/Delayed_neutron

 

[Flyboy]

In nuclear engineering, "residual heat" refers to the heat released by the fission products. It decreases exponentially once the reactor is subcritical, but it never really goes to zero. So the question as I understood it, is how the design removes this energy in the long term.

Ah, fair point. I know there's been suggestions/proposals in the past to use Brayton cycle generators to tap that decay heat to provide electrical power to supplement or replace photovoltaic systems. They're called Bimodal NTRs, and NASA has looked at them somewhat seriously in the past.

As for the NERVA design... I think they planned to just straight up quench the core with LH2 and then called it good. For ground tests, they used liquid nitrogen for a long time to get it down to temperatures where they could safely disassemble it and study it after firing.

 

[Flyboy]

If you do turn up the reference, please come back and post!

Found it! There's an old report from Aerojet Nuclear Systems Company about the specs for the classic NERVA program, and it doesn't list any specific times, just a lot of "TBD sec". So, I don't know where the duration came from that I listed. I'll keep looking.

It's an interesting read, though. Check it out.

 

[Peter034]

This doesn't sound right. With double the ejection speed you get double delta-V for same mass ratioes, but if you want to go Mars you still have to coast in an (almost) Hohmann transfer orbit between Earth and Mars for a more or less a fixed amount of time. I am not familiar with the details of NERVA, but I would think that a more efficient engine would primarily be used to allow for a higher payload mass ratio.

He must have been talking about a smaller mass ratio for the more than double part of it.

 

[Filip Larsen]

He must have been talking about a smaller mass ratio for the more than double part of it.

Not quite sure what you mean by this, but for me the discussion for this bit was about if a significant improvement in ejection speed realistically would be used to reduce transit time or increase the payload (or a mix of the two), where I was arguing that mission designs almost always seem to go towards highest possible payload.

 

[Peter034]

Not quite sure what you mean by this, but for me the discussion for this bit was about if a significant improvement in ejection speed realistically would be used to reduce transit time or increase the payload (or a mix of the two), where I was arguing that mission designs almost always seem to go towards highest possible payload.

Yes. I would think so too. I was reinterpreting Mason. At first reading I saw it just like you did. It didn't sound right as you put it. I would have said just the same thing. When I went back to try and imagine how he came to imagine that you get more than twice the momentum gain per unit working mass it seemed that he must have imagined a mission that would tend to require a large mass ratio in the case of some old fashioned rocket technology, and then, if I am getting this right, he would get is more than twice the delta velocity per unit working mass by reducing the mass ratio.

The second reading told me that he probably knows that you would not get more than twice the momentum gain with the same mass ratio.