Will SpaceX use NASA Kilopower Nuclear Reactors?

[JamesH65]

I think nuclear is just one thing too many for Spacex at this time.

They need to develop so much stuff, in so many fields, that anything they can do without is a gain.

The question should really be: What can Spacex do without and still achieve its goal?

It's really a question of cost. They can always spin off an independent lab to do the work, but it all costs money. It's not like Musk himself needs to run everything. Better to get others to do the leg work.

[john smith 19]

Basically need a much smaller version of these SMR designs that Nuscale is working on, with a higher enriched fuel to be able to last longer between refuelling. The weight they list seems wrong though as you can't transport a 700 ton item by truck. 70 tons seems more realistic. 50 megawatts of electric power and 160 megawatts of thermal power would be quite useful.

http://www.nuscalepower.com/our-technology/technology-overview

According to this presentation

https://www.iaea.org/NuclearPower/Downloadable/Meetings/2013/2013-09-02-09-04-TM-NPTD/20_usa_colbert_nuscale.pdf

The standard reactor core will weigh 264 tonnes.

When one is actually built.

Kilopower has been built and is in test right now.  That makes quite a difference.

[lamontagne]

I think nuclear is just one thing too many for Spacex at this time.

They need to develop so much stuff, in so many fields, that anything they can do without is a gain.

The question should really be: What can Spacex do without and still achieve its goal?

It's really a question of cost. They can always spin off an independent lab to do the work, but it all costs money. It's not like Musk himself needs to run everything. Better to get others to do the leg work.

Financing is the key.  Not cost but getting people to invest.  The best market for small nuclear reactors may be Mars, but this market needs to exist before the product is developed!

For the moment, SpaceX needs to limit areas requiring investment.  As Spacex succeeds (one hopes) other will start putting money into the field.
History shows that in fact once something gets going there is over investment, as everyone tries to jump onto the bandwagon.  A typival example being rail in the nineteenth century.  What SpaceX is trying to do in survive the initial phase.  So they need to focus on the bare minimum requirements, in the expectation that once they succeed, others will follow.

A typical example, the Cseries frosm Bombardier.  Once it was shown to be viable, then Airbus snapped it up (anf Boeing tried to destroy it).   But they didn't risk anything before.  SpaceX wants to succeed with enough funds left over to continue existing.  A tough proposition.

[nacnud]

Could you make solar panels on Mars? Built a factory inside a BFR and land it as one piece. You'd need to refine the raw materials into semi conductors and conductors and then make into panels.

You would bring the harder to refine elements from Earth but if you could make silicon and aluminium that might be enough to be worth the bother.

[Patchouli]

Could you make solar panels on Mars? Built a factory inside a BFR and land it as one piece. You'd need to refine the raw materials into semi conductors and conductors and then make into panels.

You would bring the harder to refine elements from Earth but if you could make silicon and aluminium that might be enough to be worth the bother.

You'll have to send several hundred maybe even thousands of tons of hardware before you can manufacture silicon solar cells from Martian materials.
Building a solar concentrator and using a turbine based system would be a better option.
Even building nuclear power plant may be easier than setting up a semi conductor plant as most of the heaviest stuff is just steel and concrete which is pretty low tech stuff that could be manufactured using fairly simple hardware without an extensive chemical industrial infrastructure.

[AncientU]

Basically need a much smaller version of these SMR designs that Nuscale is working on, with a higher enriched fuel to be able to last longer between refuelling. The weight they list seems wrong though as you can't transport a 700 ton item by truck. 70 tons seems more realistic. 50 megawatts of electric power and 160 megawatts of thermal power would be quite useful.

http://www.nuscalepower.com/our-technology/technology-overview

Interesting.  Proposed same concept (except for 100MW electric per unit) in 1985-6 to GE Nuclear.  The economies-of-scale legion had a fit.  10 units to produce the GigaWatt electric, one being refueled, one in standby.  Leveled the utility workforce and made pre-constructed, modular units available to start producing power in 18-24 months instead of five years.  Full automated operation -- no TMI-Peach Bottom-like operators.

Half or quarter scale is just right for Mars.

[KelvinZero]

One interesting thing about nuclear power: It could make initial ISRU pretty trivial. You could land right on or near the exposed ice at Louth crater. Just trundle a nuclear heated probe onto the ice. Something a lot easier than the proposed europa landers could melt down 30 meters and begin creating a lake of water. It could be tiny compared to BFS cargo. A couple of tons. At that depth you get radiation shielding and robust earth pressure for free.

Would it just get too cold there? Maybe. However the crater is at a latitude of about 70°, so I don't think you would face permanent night, and one ridge of the crater should be well placed to catch the sun every day.

Another option could be to initially just land the BFS in the long summer and bring it home before the long winter. Permanent infrastructure could be buried in the ice, which is a good heat insulator.

I think we will find much better locations right near the equator, but if anyone is too worried about the unknowns, here is a known. You can see the ice right there in photos.

Maybe this could just be an initial location. You land your heavy prospecting equipment there in the summer and bring the BFS home, but then this equipment all moves in a caravan toward the equator, prospecting as it goes. This could all be done robotically because the ISRU is so simple.

[john smith 19]

You'll have to send several hundred maybe even thousands of tons of hardware before you can manufacture silicon solar cells from Martian materials.

IIRC Silicon is one of the 10 most common elements in the Universe. The question is how much capacity do you want to build out and how fast do  you want to do it.

Building a solar concentrator and using a turbine based system would be a better option.

Certainly more scalable, although prone to dust storms. Also much simpler supply chain that can (in principle) be built on Mars.

Even building nuclear power plant may be easier than setting up a semi conductor plant as most of the heaviest stuff is just steel and concrete which is pretty low tech stuff that could be manufactured using fairly simple hardware without an extensive chemical industrial infrastructure.

If you have a large supply of fossil fuel to use as fuel.

Which Mars does not.

It's only when you'd don't have it that you realize just how big an enabler of large engineering projects a large, ready source of energy is to making them happen. For example heating Silicon to its melting point (about 1450c) is every energy intensive but its heat of fusion is huge so getting it that 1 degree over the line to liquid Silicon doubles the energy bill at least. That's why you need a very big solar array already to make more of them on orbit for SPS, or anywhere else for that matter.

[john smith 19]

Interesting.  Proposed same concept (except for 100MW electric per unit) in 1985-6 to GE Nuclear.  The economies-of-scale legion had a fit.  10 units to produce the GigaWatt electric, one being refueled, one in standby.  Leveled the utility workforce and made pre-constructed, modular units available to start producing power in 18-24 months instead of five years.  Full automated operation -- no TMI-Peach Bottom-like operators.

Half or quarter scale is just right for Mars.

Do you know anyone who's constructing units of such size anywhere in the world?

Right now AFAIK the closest fit to a  Mars power plant would be the naval ship reactors, or the few designs developed for nuclear civilian vessels in the 1950's and 1960's by (IIRC) the US, Germany and Japan.

AFAIK most Western naval reactors are PWR's but they are (were?) sealed for life with suitable fuel loads and fissionable poisons to level the power output. They also have tended to run with HEU but I think that's changing.

On that basis if any navy is operating a refuellable LEU PWR for its fleet then that would be the best (near term) design you could get if SX could get it.

Otherwise AFAIK Kilopower is at the Nevada test site right now and going through tests. 

People seem to obsess about how it's only 10Kw, but you don't have to send just one, do you?
Incidentally Kilopwer is (in principle) moveable by a team of astronauts on a trolley (provided it's been shut down for about a week), no crane required. It has also been designed to be started by remote control. The biggest variable is the deployment of the cooling radiator, which will vary with the environment, but could be a fixed design, no moving parts needed.

I think we will find much better locations right near the equator, but if anyone is too worried about the unknowns, here is a known. You can see the ice right there in photos.

At this point there are enough unknowns that eliminating even one of them is pretty attractive. You'd have proved out ISR recovery and ISRU so their TRL's would have gone to 9.

In principle ISRU is a simple idea but it's the implementation (especially how to handle faults and keep working) that makes designing it such a PITA. Life gets very tough when you can't just send a guy out to fix it (because "the guy" is about 140 million miles from the hardware    ).

Maybe this could just be an initial location. You land your heavy prospecting equipment there in the summer and bring the BFS home, but then this equipment all moves in a caravan toward the equator, prospecting as it goes. This could all be done robotically because the ISRU is so simple.

Well simpler than having to drill through a surface layer to get to the ice that's true. Do we really know if that's just a liquid, or could it be more like permafrost, with lots of solids to filter out first?

I don't think there's any doubt that the first pair of BFS's to Mars will not be crewed.

[JamesH65]

You'll have to send several hundred maybe even thousands of tons of hardware before you can manufacture silicon solar cells from Martian materials.

IIRC Silicon is one of the 10 most common elements in the Universe. The question is how much capacity do you want to build out and how fast do  you want to do it.

I think availability of silicon* is fairly irrelevant to the cost of making solar panels. It's the process that costs, not the raw materials. http://www.madehow.com/Volume-1/Solar-Cell.html

* Although silicon of the required grade for chip manufacture is fairly thin on the ground on Earth.

[speedevil]

It's only when you'd don't have it that you realize just how big an enabler of large engineering projects a large, ready source of energy is to making them happen. For example heating Silicon to its melting point (about 1450c) is every energy intensive but its heat of fusion is huge so getting it that 1 degree over the line to liquid Silicon doubles the energy bill at least. That's why you need a very big solar array already to make more of them on orbit for SPS, or anywhere else for that matter.

I note that launch to Mars surface costs $130/kg or so. (BFR, 2016)

Picking the first more-or-less suitable cell off alibaba gives me this 3.5W cell for $1.7.

It produces 3.5W, and weighs 5g.

For $130 or so, you can transport the $360 worth of cells to Mars, where they will produce about 350W peak, and perhaps 100W average in a good location.

If you are claiming that it's worth it making cells on Mars, you are also implicitly assuming that it is considerably cheaper to make cells on Mars than on earth.

Which seems a rather extravagant claim.

The cells picked were on the first page of the listings for Solar Cell, the first bare cell. Somewhat thinner cells are coming onto the market, and thin film cells may be considerably lighter.

[jpo234]

* Although silicon of the required grade for chip manufacture is fairly thin on the ground on Earth.

I'm not sure what you mean here. Silicon with the "nine nines" purity required for chip making always requires multiple refinement steps...

[KelvinZero]

For $130 or so, you can transport the $360 worth of cells to Mars, where they will produce about 350W peak, and perhaps 100W average in a good location.

If you are claiming that it's worth it making cells on Mars, you are also implicitly assuming that it is considerably cheaper to make cells on Mars than on earth.

Which seems a rather extravagant claim.

You can't just do a mars dollar to earth dollar conversion though. For example, consider the extreme case where mars is 100% self sufficient and yet has $0 products to sell to earth. Mars would then be entirely capable of building it's own solar cells but entirely incapable of buying anything from earth.

[speedevil]

You can't just do a mars dollar to earth dollar conversion though. For example, consider the extreme case where mars is 100% self sufficient and yet has $0 products to sell to earth. Mars would then be entirely capable of building it's own solar cells but entirely incapable of buying anything from earth.

No, you can't.
However, even widening out the question a lot from 'Will SpaceX use NASA Kilopower Nuclear Reactors' to include possible alternatives from earth (note that the Kilopower reactor is from earth)  doesn't get you to questions of a local economy till a very, very long time in the future.

[WindnWar]

Interesting.  Proposed same concept (except for 100MW electric per unit) in 1985-6 to GE Nuclear.  The economies-of-scale legion had a fit.  10 units to produce the GigaWatt electric, one being refueled, one in standby.  Leveled the utility workforce and made pre-constructed, modular units available to start producing power in 18-24 months instead of five years.  Full automated operation -- no TMI-Peach Bottom-like operators.

Half or quarter scale is just right for Mars.

Do you know anyone who's constructing units of such size anywhere in the world?

Right now AFAIK the closest fit to a  Mars power plant would be the naval ship reactors, or the few designs developed for nuclear civilian vessels in the 1950's and 1960's by (IIRC) the US, Germany and Japan.

AFAIK most Western naval reactors are PWR's but they are (were?) sealed for life with suitable fuel loads and fissionable poisons to level the power output. They also have tended to run with HEU but I think that's changing.

On that basis if any navy is operating a refuellable LEU PWR for its fleet then that would be the best (near term) design you could get if SX could get it.

Otherwise AFAIK Kilopower is at the Nevada test site right now and going through tests. 

People seem to obsess about how it's only 10Kw, but you don't have to send just one, do you?
Incidentally Kilopwer is (in principle) moveable by a team of astronauts on a trolley (provided it's been shut down for about a week), no crane required. It has also been designed to be started by remote control. The biggest variable is the deployment of the cooling radiator, which will vary with the environment, but could be a fixed design, no moving parts needed.

I think we will find much better locations right near the equator, but if anyone is too worried about the unknowns, here is a known. You can see the ice right there in photos.

At this point there are enough unknowns that eliminating even one of them is pretty attractive. You'd have proved out ISR recovery and ISRU so their TRL's would have gone to 9.

In principle ISRU is a simple idea but it's the implementation (especially how to handle faults and keep working) that makes designing it such a PITA. Life gets very tough when you can't just send a guy out to fix it (because "the guy" is about 140 million miles from the hardware    ).

Maybe this could just be an initial location. You land your heavy prospecting equipment there in the summer and bring the BFS home, but then this equipment all moves in a caravan toward the equator, prospecting as it goes. This could all be done robotically because the ISRU is so simple.

Well simpler than having to drill through a surface layer to get to the ice that's true. Do we really know if that's just a liquid, or could it be more like permafrost, with lots of solids to filter out first?

I don't think there's any doubt that the first pair of BFS's to Mars will not be crewed.

US Naval reactors are all HEU reactors in order to have the power density and time between refuelling required for service, some of the older Los Angeles class subs had core lifespans of 10-15 years but refuelled cores used a newer design and both them and the newer Virginia class are all 30 plus year lifespans. They can be refuelled but it's a pretty massive operation. I grew up not far from Mare Island and knew a lot of folks that worked at the base doing refueling ops back in the day. Nuclear non proliferation agreements require commercial reactors to be 14% or less enrichment. Current US Naval reactors are 90% or more enrichment. Not sure what the lifespan would be using lower enriched fuel though it is possible to use it in them.

I would think it would require a pretty major waiver to be able to use and launch such a reactor, but that's not my specialty...

[Kenp51d]

First let admit to a relatively low knowledge level here, so this may not be the brightest question.
The constraints and risk can be fairly easy to emagine with luanching a fueled reactor.
How feasable would it be to luanch an unfueled reactor, and then fuel it in orbit (roboticly perhaps, yeah I know that even on Earth fuel is remotely handled).
Fuel sent up on another launch with extra safe guards, maybe?

Ken


Sent from my XT1565 using Tapatalk

[JamesH65]

* Although silicon of the required grade for chip manufacture is fairly thin on the ground on Earth.

I'm not sure what you mean here. Silicon with the "nine nines" purity required for chip making always requires multiple refinement steps...

You need very high silicon content sand, with very few impurities, which is not beach sand. Locations on Earth where the raw materials are a high enough quality are not common. I'm sure if you wanted to, you could make a process that removed the impurities and could therefor use 'beach' sand, but we don't do that on Earth. Of course, if you find a pure quartzite deposit on Mars, you could use that.

[john smith 19]

First let admit to a relatively low knowledge level here, so this may not be the brightest question.
The constraints and risk can be fairly easy to emagine with luanching a fueled reactor.
How feasable would it be to luanch an unfueled reactor, and then fuel it in orbit (roboticly perhaps, yeah I know that even on Earth fuel is remotely handled).
Fuel sent up on another launch with extra safe guards, maybe?

Key issues with reactors are crashes are a) Dispersion of nuclear material b)Reactor going critical.
 
Your idea means you now have to have two perfect launches and a rendezvous and a docking in space.

a) Conventional reactors use powered Uranium Oxide as the fuel. It's pretty safe but it's a ceramic, so if it gets crushed it'll generate powder, and there's a lot of it. Kilopower uses quite big lumps of solid metal. They can burn but the big lumps mean they can absorb a lot of heat before doing so.

b) Sea water (or water logged sand, which is the other test condition) is an effective moderator and a lot of conventional designs have "voids" where the conventional coolant could be forced out of and replaced by this better moderator, forcing the reactor critical.  Kilopower does not have a single coolant flow it uses multiple sealed heat pipes. Individual sealed fluid loops. To push moderator into the reactor one (or more) of these would have to rupture, along with the removal of the control rods that help prevent criticality to begin with.

Until startup reactors are fairly safe. The fuel is an Alpha emitter which requires minimal shielding. Essentially it's like the engine block of a truck. It'll kill you if it falls on you. Once it starts up (and the truck it's attached to starts moving) things change a lot.

That's when they become neutron emitters and dangerous to get close to.

Which is why AFAIK no space reactor has ever  launched while operating. They are always started up in space.  Kilopower (by design) is also meant to be approachable within 1 week of shut down, which is AFAIK a first for space nuclear reactors and may not be possible for nuclear thermal rocket IE direct thrust generation rather than electrical power applications.

[TrevorMonty]

What are mass estimates for operational 1KW and 10KW versions?

Would they stay mounted to lander that delivers them to surface? Lander may need  wheels to move to final location.

[Kenp51d]

Thanks John.
Good info. I have a much better understanding now.

Ken

Sent from my V10 using Tapatalk