Will SpaceX ever go nuclear propulsion?

[ArbitraryConstant]

MSRs have zero risk of loss of coolant accidents, because the fuel and the coolant are mixed together.

If the fuel and the coolant are the same thing, loss of coolant is loss of fuel. In practice, even ambient pressure systems leak, this has the added disadvantage that the fluid will self-heat wherever it ends up. Safety systems can hope to manage this, but let's not boast of certainty, the probability is not exactly 1. If there was ever a time nuclear could get away with that, it's been over for a while now.

Even if light water technology was a mistake, we now have a great deal of operational experience on existing designs, and safety features come from real accidents. The limited time on molten salt designs means there's inevitably going to be unknown unknowns.

It's not just a safety issue, on Mars it will impact the energy balance of the colony. Reactors must last as long as expected and require the amount of maintenance that's expected.

Those swelling issues mostly don't apply to MSRs.

Not sure that's true. I would expect the entire primary loop to be exposed to delayed neutrons. If it's a two fluid design there's obviously going to be significant neutron flux through the barrier. Not that this would be a deal breaker, but I am suspicious of claims that something doesn't apply when I'm pretty sure it at least needs proper engineering to address.

The big problem with solar isn't just having enough panels. Its energy storage. Like on earth, it's not always day in Mars. You must store energy overnight.

I am aware. Here on Earth batteries are already used for demand control and frequency regulation, it looks like they'll be competitive with natural gas for peaking plants in the next few years. We're probably less than ten years from the point where solar+battery on premises is competitive with utility rates.

The rapidity of improvement of these technologies leads to counterintuitive results, and trades are significantly impacted over only a few years. Given a 20+ year time horizon for large scale Mars activity, the outcome is entirely up in the air.

[Robotbeat]

Radioisotope power supplies actually trade very well compared to fission from a specific power standpoint. Standard RPS/RTGs get 6-10W/kg compared to about 3-4 W/kg for the fission reactor on NASA's 5.0 Mars DRA. At least for NASA, a combo RPS and solar power system seems like a pretty good fit, considering the huge advances in both photovoltaics and lithium sulfur batteries over the last decade or two (which an give nearly 10W/kg, effectively).

...at least at the scale of 20-50kWe NASA typically is thinking for surface missions. At larger scales, nuclear trades a bit better and RPS definitely trades worse since Pu-238 is so scarce.

[ArbitraryConstant]

3. Nuclear propulsion scores in the availability of large quantities of power, allowing rapid transit times under some designs, but loses on maturity, complexity and bureaucracy - especially the latter.

The problem for nuclear propulsion is that the specific power isn't really that good, for a short trip like Mars the aggregate acceleration never has time to accrue a large advantage. If upmass from Earth is cheap, as SpaceX wants it to be, an abundant chemical scheme wins. Same goes for solar electric, the deta-v is insane (the Dawn mission has more delta-v capability than SSTO) but it doesn't win to Mars with that huge initial chemical push available.

[JasonAW3]

Unless a ready and easily accessible and easily refinible source of Nuclear materials can be found, namely in an astroid or possibly on the moon, the cost of boosting that much mass into orbit, as well as the perceived hazards, will make nuclear powered space craft for all practical purposes, impossible.

When it takes over 20 years just to get all the paperwork done just to BUILD a Nuclear Power Plant, getting clearence to boost such a device into orbit would, at best, seem next to impossible.

[ArbitraryConstant]

Unless a ready and easily accessible and easily refinible source of Nuclear materials can be found, namely in an astroid or possibly on the moon, the cost of boosting that much mass into orbit, as well as the perceived hazards, will make nuclear powered space craft for all practical purposes, impossible.

I don't think that's so much of a problem - before a reactor goes critical Uranium is not dangerous. It's actually less of a radioactivity hazard than RTGs. If we don't want this stuff getting out, then we need to close every coal power plant first.

For mining materials locally, a real thorium breeder or a uranium fast reactor would, thermodynamically at least, be favorable basically anywhere. On Earth it would be worth it to extract the materials from sea water or granite. The energy density is in the gigawatt decades per cubic meter. I'm not convinced it'd work, but this is not one of my concerns.

When it takes over 20 years just to get all the paperwork done just to BUILD a Nuclear Power Plant, getting clearence to boost such a device into orbit would, at best, seem next to impossible.

You make sure it's not coming back before you switch it on.

Such things have been proposed, for example JIMO would have used it for high power ion propulsion and experiments around Jupiter. I'm disappointed JIMO never got properly funded, total mission delta-v would have been ~38 km/s. o_O

[nadreck]

Taking from the concept of a high power to weight ratio reactor like the molten salt ones, and bringing the conversation back to propulsion, how about a molten salt reactor that powers a low thrust high ISP monatomic hydrogen engine that, at highest thrust (still to low for anything except a space only stage), lowest ISP (where 100% of the power generated is used to produce monatomic hydrogen and maybe even a little lox is added to increase thrust) without the lox, ISP would be about 800 to 900, however you could get a much higher ISP if you took a couple of percent of the power to make monatomic hydrogen and feed it into an MHD drive using the rest of the available power. There you could theoretically get performance at or beyond what VASIMR provides.

The advantage in this process though in a Mars colonization effort only a couple of the reactors need to be left behind to provide electric power while the rest continue to ply the space between Mars and Earth (or later be adapted to cruise the 'roids, Trojans, Jovian moons).

hmm thinking about it further, I doubt adding oxygen would yield significant improvements in thrust over injecting a little H₂ (possibly to cool the chamber) and bringing the reaction temp down to about 3000K from 4000K, and the oxygen would probably slow the exhaust velocity down much more.

[Robotbeat]

You're missing that hydrogen is very low density, the Achilles heel of NTR.

[Vultur]

And then you need to account for energy storage that could be mostly avoided with two reliable nuclear reactors.

True, but that is a technology advancing really fast; in 15-20 years (ie by the time we are in a position to have a Mars colony, even optimistically) I don't think it will be a problem.

Recycling Water, Oxygen, CO2, human waste. Hydroponic food growth (with around the clock lighting).

Why use around-the-clock lighting? Mars has an Earthlike day/night cycle; the ability to grow crops with natural sunlight (saving energy) is one of its big advantages over the Moon.

I believe around-the-clock lighting even limits what crops you can grow (as some use day length to determine when to set fruit etc.)

The plants will also provide the oxygen/CO2 cycling and some parts of the waste recycling.

Plus placing solar panels exposed to the elements in Mars could risk damage from the sandstorms you mentioned.

Maybe, but Spirit and Opportunity have I think demonstrated that Mars is a much more benign environment for solar power than was previously expected.

Very long-lasting dust storms can happen, but they are not "zero solar power" by quite a stretch. You just need much more area... but thin films are light.

Kirk's argument is just as important about solar too. If a large enough share of solar panels fail for any reason, you die sooner or later (lack of oxygen, lack of drinking water, lack of food).

That's true, but if we're talking about a large area composed of separate thin-film modules, what would make them all (or mostly) fail at once?

[nadreck]

You're missing that hydrogen is very low density, the Achilles heel of NTR.

At the high end of ISP that doesn't matter, at the low end, if we are talking about a 1:5 to 1:10 TWR on the engine/reactor, it still doesn't matter, it will be a big tank but will not need to be structurally strong. Making H₂ space storable becomes the biggest problem. One other possibility is to use liquid methane in this engine. Break it into C H H H H and let it all recombine as it will might still end up as a pretty high ISP engine in high(er) thrust mode, and in MHD mode it might end up with higher thrust though somewhat lower ISP. 

[SoulWager]

"ever" is a long time, but I don't see anyone flying a nuclear thermal rocket in the foreseeable future. Electric propulsion kind of removes the need for it.

I think nuclear power will be limited to RTGs for unmanned vehicles, and maybe reactors for surface colonies, where solar isn't reliable enough to be suitable for life support.  For example, dust storms on mars, or long nights on the moon.

[Robotbeat]

You're missing that hydrogen is very low density, the Achilles heel of NTR.

At the high end of ISP that doesn't matter, at the low end, if we are talking about a 1:5 to 1:10 TWR on the engine/reactor, it still doesn't matter, it will be a big tank but will not need to be structurally strong. Making H₂ space storable becomes the biggest problem. One other possibility is to use liquid methane in this engine. Break it into C H H H H and let it all recombine as it will might still end up as a pretty high ISP engine in high(er) thrust mode, and in MHD mode it might end up with higher thrust though somewhat lower ISP.

Yeah, it DOES matter. You end up with a far bigger and more expensive stage. And it DOES of course need to be structurally strong since it's a pressure vessel, whose dry mass is proportional more to volume than it is to propellant mass. It eats up a lot of NTR's advantage, and certainly makes it a lot more expensive. Also, it makes ISRU a lot harder since you need to process several times more water for same delta-v.

[RanulfC]

Even if SpaceX were really gungho about nuclear (which they aren't), the immediate issue would be testing. While KIWI and NERVA proved the concept in the 1960s, they were not very reliable and would require extensive testing to reach a point where you would even consider using them on a crewed vehicle. Plus, there is literally nowhere on Earth today where you would be allowed to do open-air tests of a nuclear rocket, as was done at Jackass Flats. The testing would have to be done in a closed facility, which would be regulated even more than a commercial power reactor. Which is say, by the time the paperwork is complete, you'll be dead.

And that's not even considering that are far better options. SpaceX's favored approach is very big chemical rockets and ISRU, which has no regulatory issues. Likewise for solar electric, which completely proven technology (i.e. Dawn). If you really, really want to use fission, nuclear electric is always available, and requires considerably less testing (especially if using a solid state reactor, like Los Alamos has recently been pushing).

There is very little future for nuclear thermal rockets.

Late but a point here is that NERVA was to the point that a flight ready system was tested and ready. The designs were highly reliable and about the only testing needed was functional testing of a modern build. While facilitites are lacking there's been a good amount of continual study on what would be needed so we have a pretty firm idea on how much it would cost and what it would take to safetly test a nuclear thermal rocket if we wished to.

Having said that I don't see it being something SpaceX would do. Surface nuclear is possible but they don't seem to be interested in the legal and political hassels of trying to obtain NTRs.

Randy

[nadreck]

You're missing that hydrogen is very low density, the Achilles heel of NTR.

At the high end of ISP that doesn't matter, at the low end, if we are talking about a 1:5 to 1:10 TWR on the engine/reactor, it still doesn't matter, it will be a big tank but will not need to be structurally strong. Making H₂ space storable becomes the biggest problem. One other possibility is to use liquid methane in this engine. Break it into C H H H H and let it all recombine as it will might still end up as a pretty high ISP engine in high(er) thrust mode, and in MHD mode it might end up with higher thrust though somewhat lower ISP.

Yeah, it DOES matter. You end up with a far bigger and more expensive stage. And it DOES of course need to be structurally strong since it's a pressure vessel, whose dry mass is proportional more to volume than it is to propellant mass. It eats up a lot of NTR's advantage, and certainly makes it a lot more expensive. Also, it makes ISRU a lot harder since you need to process several times more water for same delta-v.

In space, as long as you resolve the boil off issues, it is no more a pressure vessel than a water tank. I am not talking NTR I am talking much lower TWR here, 5000 ISP at low thrust 10 newtons per ton(total WAG based on adding weight and inefficiency to VASIMR specs), and maybe 900 at high thrust 500 newtons per ton (total WAG based on last WAG x 50 for proportion of H₂ turned into monatomic H and the extra H₂ added for cooling/extra reaction mass. Note per ton of engine and reactor.

[RanulfC]

]I've often found that whenever someone ends a statement with the word "period," they inevitably are glossing over a bunch of important issues.

For instance, the Isp is double, but you also explode the dry mass because you need much larger tanks (tank mass is proportional to volume) and NTR engines are much heavier than the otherwise-equivalent chemical rocket engine. Also, NTR is much more expensive and harder to reuse. And (a minor note), you need a LOT more hydrogen (oxygen you get included if you're doing electrolysis and nearly-free from Earth's atmosphere), which takes more energy to generate or water to mine (on the Moon or Mars or whathaveyou). But really, the increased difficulty of reuse in my mind makes NTR not worth it at all.

This is always glossed over... NASA architectures always show disposable NTR stages. Who (besides NASA with Apollo-funding-on-steroids) could possibly afford to throw away nuclear reactors like that? Chemical stages, if you do docking etc with them, aren't terribly difficult to reuse in principle, so they're a far more cost-effective solution. Not only are they much cheaper to develop, but they're surely going to be far cheaper to build per unit and almost certainly much easier to reuse plus their propellant costs (if that becomes significant) are much less and the overall SIZE of the stage will be much smaller with chemical (because liquid hydrogen is basically the least dense liquid).

Also, while I think Solar Electric Propulsion is awesome, don't become enamored with VASIMR. There are a lot of other electric propulsion solutions out there that are less complicated and even potentially higher performing, not to mention more mature and proven.

Couple of points if I may
"Reuse" of NTRs isn't an issue really IF you start from that standpoint. Using NASA is a very bad example as they tend to not consider reuse very often as a standard. I'll be blunt and point out that according to NASA use for the most part you simply dump most of your equipment the second its jobs done as a STANDARD no matter the basic architecture. If you're only planning a few "missions" (which is the standard government model) you normally don't worry about reuse or reusablity. You're also missing a point on the propellant as an NTR CAN use LOX and most proposals recently have included the use of a LOX "afterburner" as standard along with longer-life and refurbishable NTRs. And it's not like any of the processed byproducts fo the propellant manufacturing are going to "waste" either.

The basic argument against NTRs boils down to expense (mostly due to regulatory and political issues) and complexity over more basic systems and truth be told we don't NEED nuclear for anything inside the asteroid belt really. It's a nice-to-have though and does in fact lend itself to reuabilty and high operations tempo IF you can avoid the afore mentioned regulatory and political pitfalls.

SpaceX seems to me to be going the full chemical route for propulsion with either solar or surface nuclear for other needs but myself I see a combination of various propulsion means coming into play as time goes on.

Randy

[ncb1397]

If you want propellant density in your nuclear propulsion scheme, I would suggest not using hydrogen or oxygen. Some of the heavier noble gases liquified would offer very good propellant density and ISRU would be very simple(Argon is 2% of Mars atmosphere). Having compounds that don't react with anything has a component of added safety to it and may aid in high levels of re-use. Some of the high efficiency heat-exchanger technology developed for skylon could aid in heating the working fluid.

argon liquid density: 1.3954 g/cm^3
argon boiling point: 87 K

Noble gases also are used in electric propulsion and so you could have one tank and two engines. One high efficiency NEP engine and a lower efficiency but higher thrust NTR scheme on the same vehicle. Attitude control would be through electric thrusters.

[sheltonjr]

Here is my unrealistic wildly optimistic timeline that could possibly happen. 

2016-2019  Prototype MSR developed built and tested. My favorite is Terrestrial Energy, FH DV2 flight to Mars

2020-2024 Small factory made MSRs are fielded in the Canada Oil Sands and perform perfectly

2024-2030 Factory made MSRs are fielded around the world solving world energy and producing millions of gallons of fresh water from waste heat

2027 First SpaceX MCT arrives on Mars with ISRU systems to refill tanks.

2029 Light weight 5 MWe Space MSR Design work starts. Cargo MCT arrives at Mars, refuels and returns.

2031 First Manned MCT arrives at Mars using Solar and Methane/Lox for power requirements

2033 3 MCTs arrives at Mars with 1 cold inert MSR for power. Buried and fissile fuel is added to take it critical.

2035 NASA uses a space version of the MSR for a proper JIMO mission with Bigelow Modules, MCT and NEP to visit the moons of Jupiter with probes to investigate under the ice on Europa.

I guess my point is that even for SpaceX, Mars is still 15 years away and with the recent resurrection of MSR technology that hopefully reason will overcome the FUD of nuclear power and will benefit the Earth and our prospects for exploring the solar system and becoming a multi-planetary civilization.

If the above schedule were to slide to the right 10-15 years, I would still consider that great progress. By 2040 I will be 74 years old and hope to see all this happen.

[Robotbeat]

Nuclear thermal rockets that don't use hydrogen have the low Isp of chemical rockets combined with the operational complexity and weight of a heavy rector and shielding.

Keep the reactor on the ground to produce propellant. Just because you COULD use a nuclear reactor doesn't make it a good idea. A lot of space enthusiasts treat nuclear power as space-magic, a wand to wave that makes everything magically feasible. If you actually look at the realistic engineering involved, nuclear thermal (and solar thermal for that matter) simply doesn't trade very well nowadays. In fact, I'd argue that the closer we get to routine, inexpensive space launch, the LESS sense nuclear thermal makes.

[RanulfC]

Nuclear thermal rockets that don't use hydrogen have the low Isp of chemical rockets combined with the operational complexity and weight of a heavy rector and shielding.

Not at all really. Ammonia as an NTR propellant still beats straight chemical for ISP, but your fixation on "bulk" is really a non-issue given any sort of general lowered standard of access prices. At the price-point being imagined its simple enough to ship water anywhere in the solar system if needs be and crack it on-site. The main advantage of methalox is the abilty to pretty cheaply haul it around rather than depending on "just" ISRU sources rather than trying to maintain large quantities of LH2/LOX which is arguably pretty easy once you figure enough mass for any decent solar/nuclear powered cryo-cooling ability.

Keep the reactor on the ground to produce propellant. Just because you COULD use a nuclear reactor doesn't make it a good idea. A lot of space enthusiasts treat nuclear power as space-magic, a wand to wave that makes everything magically feasible. If you actually look at the realistic engineering involved, nuclear thermal (and solar thermal for that matter) simply doesn't trade very well nowadays. In fact, I'd argue that the closer we get to routine, inexpensive space launch, the LESS sense nuclear thermal makes.

Abudant chemical as an architecture requires a lot of "on-site" power production as a basis. NTRs are not a cure-all by any means and few people who actually look into them treat them as such though "new-comers" tend to start out overly enthusiastic they learn better quickly. Still NTR has its place, I still don't see it having one at SpaceX planning table but we really don't know what Musk is planning and are specualting on very limited data

Randy

[Nindalf]

Nuclear thermal rockets that don't use hydrogen have the low Isp of chemical rockets combined with the operational complexity and weight of a heavy rector and shielding.

Keep the reactor on the ground to produce propellant. Just because you COULD use a nuclear reactor doesn't make it a good idea. A lot of space enthusiasts treat nuclear power as space-magic, a wand to wave that makes everything magically feasible. If you actually look at the realistic engineering involved, nuclear thermal (and solar thermal for that matter) simply doesn't trade very well nowadays. In fact, I'd argue that the closer we get to routine, inexpensive space launch, the LESS sense nuclear thermal makes.

This guy's stuff is pretty interesting:
http://neofuel.com/index_neofuel.html

He makes a good point that if abundant, easily-extracted water is available in space, for instance on one of Mars's moons, it may be more operationally efficient to simply use water for the propellant in a steam NTR (gives an Isp around 190-200s), rather than attempt any chemical processing.  He talks about 50,000 ton payloads from Deimos to Earth capture orbit, using a reusable engine that would be launchable on a BFR, and a lunar architecture that would put ~15,000 tons of water per year into low lunar orbit with just two small nuclear reactors (one melter/distiller, operating in lunar south pole craters, and one NTR for the shuttle).

No matter how cheap and routine launch gets, I think it's obviously preferable if we find an efficient way to get our propellant without lifting it from the Earth's surface.  Here on Earth, we conveniently have natural gas and oxygen just sitting around to be picked up.  Nuclear propulsion gives us the opportunity to use substances just sitting around to be picked up in space.

The native-material NTR system is a lot simpler than one where you first generate power, then use it to produce chemical propellants, then use a chemical rocket to burn the propellants, particularly when you're talking about handling water vs. cryogenic fluids.

[nadreck]

Nuclear thermal rockets that don't use hydrogen have the low Isp of chemical rockets combined with the operational complexity and weight of a heavy rector and shielding.

Keep the reactor on the ground to produce propellant. Just because you COULD use a nuclear reactor doesn't make it a good idea. A lot of space enthusiasts treat nuclear power as space-magic, a wand to wave that makes everything magically feasible. If you actually look at the realistic engineering involved, nuclear thermal (and solar thermal for that matter) simply doesn't trade very well nowadays. In fact, I'd argue that the closer we get to routine, inexpensive space launch, the LESS sense nuclear thermal makes.

This guy's stuff is pretty interesting:
http://neofuel.com/index_neofuel.html

He makes a good point that if abundant, easily-extracted water is available in space, for instance on one of Mars's moons, it may be more operationally efficient to simply use water for the propellant in a steam NTR (gives an Isp around 190-200s), rather than attempt any chemical processing.  He talks about 50,000 ton payloads from Deimos to Earth capture orbit, using a reusable engine that would be launchable on a BFR, and a lunar architecture that would put ~15,000 tons of water per year into low lunar orbit with just two small nuclear reactors (one melter/distiller, operating in lunar south pole craters, and one NTR for the shuttle).

No matter how cheap and routine launch gets, I think it's obviously preferable if we find an efficient way to get our propellant without lifting it from the Earth's surface.  Here on Earth, we conveniently have natural gas and oxygen just sitting around to be picked up.  Nuclear propulsion gives us the opportunity to use substances just sitting around to be picked up in space.

The native-material NTR system is a lot simpler than one where you first generate power, then use it to produce chemical propellants, then use a chemical rocket to burn the propellants, particularly when you're talking about handling water vs. cryogenic fluids.

That has steam at 200ISP with an 800K engine, you get just about as good from hydrogen peroxide, if you increase the temperature to 2000K (very doable with NERVA style NTR) you get 400 to 500.  I do believe the argument could be made that if a mirror coated 1 gigaton ice moon from Jupiter, Jovian Trojans, Saturn could be brought to say EM L2 or L3 would revolutionize the economics of exploitation of the inner system and asteroids.