Will SpaceX ever go nuclear propulsion?

[Robotbeat]

The power-beaming part is ten times harder than the solar array part. You could blanket 2 miles around a Mars colony in solar cells and Lithium Sulfur batteries for the difficulty of setting up power beaming.

[Vultur]

I didn't say SEP has no uses. But abundant chemical gives you a lot of delta-v for those high-delta-v missions.

That depends on what you consider a high delta-v mission. For example, going to Mercury and back(or any mission with a similar delta-v) with chemical is going to be really impractical even with abundant fuel in Earth orbit, while a similar mission with SEP is very feasible.

Yeah, that's the kind of thing I meant. Mars is within what's feasible with lots of chemical propellant. Mercury or reasonable manned trip times to the outer planets, not so much.

[Robotbeat]

Out planets are feasible with chemical in combination with aerobraking... But you're going to need nuclear (or ENORMOUS solar) power anyway, and orbital mechanics already guarantees long trip times even with high delta-v, and electric propulsion plus long trip times is a very good match since (from a certain perspective) you have a long time to generate electricity so your total energy that you're dumping into the propellant is proportional to the length of the trip... so that your delta-v can be roughly proportional to the square root of the trip time...

[clongton]

Re: Outer solar system
Almost certainly nuclear will be the electrical  power source.
Propulsively, I would suspect that the reactor will be dual purposed;
(1) Nuclear Thermal for power maneuvers such as departure, course correction and braking into orbit;
(2) Nuclear Electric for cruise mode.
I hesitate to speculate beyond that broad brush.

[Robotbeat]

Developing an NTR cost about the same as developing SLS, from what I've read in budget estimates. Hard to justify that. Use NEP, easier to test and better performance anyway.

[RanulfC]

The power-beaming part is ten times harder than the solar array part. You could blanket 2 miles around a Mars colony in solar cells and Lithium Sulfur batteries for the difficulty of setting up power beaming.

What "difficulty" would that be? Microwave you're putting up a simple rectenna. (And you never worry about it being "night" or any dust storm) Laser you're using those about at third as many cells and batteries. And you have the ability to spread the power among more than "just" the colony.

Of course this assumes you're not using the SEP as a cargo vehicle constantly shuttling between Earth and Mars which is the better solution in the first place. Again, your trying to hard to INSIST on nothing but abundent chemcial when its quite clear you don't NEED to do so.

While SpaceX would seem to be planning on focusing on abundent chemical with methalox it doesn't follow that that will be the ONLY option. (Though for SpaceX I suspect it will be, they probably won't be the only "players" at the point where colonization becomes a serious option)

Randy

[Robotbeat]

"Simple" rectenna would need to be just as big as the required solar array and wouldn't be any easier to deploy. Also wouldn't be simple.

[RanulfC]

Developing an NTR cost about the same as developing SLS, from what I've read in budget estimates. Hard to justify that. Use NEP, easier to test and better performance anyway.

NEP has most of the same issue as NTR and the additional problem that we have even LESS experiance/knowledge of building in space straight power reactors. Its actually easier to build a tri-modal NTR using what we already know than trying to develop a space power reactor from scratch.

And like all EP systems, NEP suffers with needing a higher T/W system for best efficency at point such as clongton pointed out.
Combining it all with the same "reactor" makes much more sense than anything else.

Randy

[Nilof]

Outer planets have high high transfer delta v's, but really low capture delta v's and Icy moons have a lot of ISRU potential. The short periods of Icy moons means you can let gravity assists do all the work, so for capture, landing on moons, and transferring between them you're likely best off with ISRU + chemical. Designing a nuclear reactor for use on a solid surface is easier.

SEP can do transfer orbits to the outer solar system just fine, and it is IMHO the most attractive option for the asteroid belt since transfers are longer than Mars and there is no local gravity well or atmosphere to brake, but there's still enough sunlight to work with(and indeed, it was selected for Dawn).

For an example of a transfer orbit to Jupiter with SEP, here's a very nice powerpoint slide that describes what happened to the CAT which had a kickstarter a while ago. It has an example of a 16 month fast Jupiter fly-by and of a 3 year Jupiter transfer for an SEP-equipped cubesat.

[go4mars]

"Simple" rectenna would need to be just as big as the required solar array and wouldn't be any easier to deploy. Also wouldn't be simple.

Not advocating here, but I will point out that orbital microwave (or other broadcast type) power sources are more versatile.  Can re-allocate by changing the orbit, and surplus power could go toward melting polar ice caps, degassing operations, smelting, trenching, or spice harvesters.  If there's ever political unrest, repurpose them elsewhere, perhaps elsewhere in the solar system.  Like Shell's floating LNG ships, mobile assets are a good idea for a variety of (unforeseen) reasons. 

The "General fusion high ISP space propulsion design" could be the orbital power source for microwave (or other) transmission.  Then your versatility is even greater.

I think a combination of surface nuclear and solar, is most likely.  Of course, these could be mobile too (with perhaps greater difficulty).

[Robotbeat]

Microwave isn't very portable because of need for huge, custom array needed on the ground. Solar+batteries (or fission) is much better.

[Nilof]

Imho, if you're willing to develop a big infrastructure, MW power beaming for powering an electric rocket makes little sense since the propulsive efficiency is low and the power beaming falls off with distance which limits the total delta-v that can be attained. Non-rocket options such as tether systems or magbeams make a lot more sense.

[Zed_Noir]

Reading the tea leafs. If the SpaceX CTO decides if some sort of nuclear power propulsion is needed, development will follow. After all the SpaceX R&D is what you called the Indiana Jones method. 

[john smith 19]

Hmm.

You're talking specifically propulsion?

That's tough. Keep in mind NTR for Nerva (according to my Sutton 4th Ed) is a T/W of 7.7:1 and an Isp of about 900secs.

That's a very poor T/W ratio by rocket engine standards (people think 40:1 is bad for reusuable "super performance" HTP engines in the 50's).

And 900secs is about 2x what LH2/LO2 can do.

But the cost to re-activate the nuclear programme, the enrichment (most designs are like 80% U235+).

And AFAIK NASA viewed them as 1 shot systems, 1 flight and then dispose of them. 

I'll note that SpaceX seem to be pretty good at raiding the libraries for historical data

TBH my instinct is if SpaceX were to go nuclear they'd look at the fission fragment rocket. The initial designs were all solid with carbon fibre "spokes" rotating through a slotted moderator forming the "throat" of an EM nozzle expelling most of the fragments (moving at 1-2% of the speed of light) out the back.

This was in 1988. The impressive points are
a)Isp of 100 000secs. Several times better than an Ion thruster.
b)Most of the system stays fairly cold as the spokes only get seriously reactive inside the moderator. The original concept was looking at some kind of Deuterated wax for the moderator.
c)Intense neutron emission only takes place within the moderator slot, simplifying the shielding a lot.
d)Thrust level in the 10s of Newtons. Not much by liquid rocket standards but huge by ion thruster standards.

The original team were looking at delivering a 6000Kg payload to Alpha Centuri in about 100 yrs with an Americurium fuel of <1.5Kg.

There is a new design based on using the fuel as a dust but the spokes-on-wheel design was viewed as viable in the late 80's. As both the thermal conductivity of certain grades of carbon fibre is 10x greater (along the fibre) and metal plating of highly electropositive metals should now be possible with "ionic liquids" this technology could be much simpler to test, if one had cheap(ish) access to space.

People have also looked at a kind of "after burner concept, like TAN, of adding chemical that is heated by the flux and increases the thrust (a lot) while lowering the Isp (a lot). Since we're looking at thrust, not Isp LH2 is no longer the automatic choice and something a bit cheaper (LN2?) might do just as well.

[Zed_Noir]

Hmm.

You're talking specifically propulsion?
...

@john smith 19 if your query is for me. Nuclear as in the use of some sort of nuclear reactor in a propulsion system. Was thinking of Nuclear electric ion drive among others.

[john smith 19]

Hmm.

You're talking specifically propulsion?
...

@john smith 19 if your query is for me. Nuclear as in the use of some sort of nuclear reactor in a propulsion system. Was thinking of Nuclear electric ion drive among others.

Apologies if I missed this but has anyone mentioned the Los Alamos work to link an actual reactor to a Stirling cycle engine.  The "DUFF" project? The goal is a system in the 1-10Kw electrical range

www.world-nuclear-news.org/C-Nuclear_and_Stirling_engines_spur_space_exploration_271112a.html

Most amazingly this project was pulled together in six months,which given the H&S environment around nuclear systems is quite amazing. 

Such a system would have several benefits.
a)Re introduce the concept of reactors in space, absent in the US since the late 60's.
b)Power levels can be throttled unlike an RTG who's power is gradually declining.
c)Can be implemented with various fuels, unlike RTG's, which rely exclusively on Pu241.

Such systems can be very useful for missions where there is a long coast period with fairly minimum power needs. It might also be the first unit for trialling ISRU systems.

In principle a developed system could have many uses across the solar system.

There are various challenges but one are that could make a big  difference is improved radiator technology allowing lighter systems to dump more heat into space, critical as even with the Stirling we're looking at less than 30% efficiency (although about 4x that of thermoelectric systems).

[EDIT I missed that their long term plan is a range of power reactors at the 1,10, 200 and 2000 Kw levels, however just getting the first level up and running would be a major achievement]

[Robotbeat]

Hmm.

You're talking specifically propulsion?

That's tough. Keep in mind NTR for Nerva (according to my Sutton 4th Ed) is a T/W of 7.7:1 and an Isp of about 900secs.

That's a very poor T/W ratio by rocket engine standards (people think 40:1 is bad for reusuable "super performance" HTP engines in the 50's).

And 900secs is about 2x what LH2/LO2 can do.

But the cost to re-activate the nuclear programme, the enrichment (most designs are like 80% U235+).

And AFAIK NASA viewed them as 1 shot systems, 1 flight and then dispose of them. 

I'll note that SpaceX seem to be pretty good at raiding the libraries for historical data

TBH my instinct is if SpaceX were to go nuclear they'd look at the fission fragment rocket. The initial designs were all solid with carbon fibre "spokes" rotating through a slotted moderator forming the "throat" of an EM nozzle expelling most of the fragments (moving at 1-2% of the speed of light) out the back.

This was in 1988. The impressive points are
a)Isp of 100 000secs. Several times better than an Ion thruster.
b)Most of the system stays fairly cold as the spokes only get seriously reactive inside the moderator. The original concept was looking at some kind of Deuterated wax for the moderator.
c)Intense neutron emission only takes place within the moderator slot, simplifying the shielding a lot.
d)Thrust level in the 10s of Newtons. Not much by liquid rocket standards but huge by ion thruster standards.

The original team were looking at delivering a 6000Kg payload to Alpha Centuri in about 100 yrs with an Americurium fuel of <1.5Kg.

There is a new design based on using the fuel as a dust but the spokes-on-wheel design was viewed as viable in the late 80's. As both the thermal conductivity of certain grades of carbon fibre is 10x greater (along the fibre) and metal plating of highly electropositive metals should now be possible with "ionic liquids" this technology could be much simpler to test, if one had cheap(ish) access to space.

People have also looked at a kind of "after burner concept, like TAN, of adding chemical that is heated by the flux and increases the thrust (a lot) while lowering the Isp (a lot). Since we're looking at thrust, not Isp LH2 is no longer the automatic choice and something a bit cheaper (LN2?) might do just as well.

I like fission fragment, but I really don't think it makes sense for anything closer than, say, the Kuiper belt. It just isn't needed. Too high of Isp for shorter missions (non optimal). It would be great for secondary propulsion on an interstellar flight, though (primary as "beamed near-relativistic dust" for acceleration and mag-sail braking against the interstellar medium for deceleration).

[john smith 19]

I like fission fragment, but I really don't think it makes sense for anything closer than, say, the Kuiper belt. It just isn't needed. Too high of Isp for shorter missions (non optimal). It would be great for secondary propulsion on an interstellar flight, though (primary as "beamed near-relativistic dust" for acceleration and mag-sail braking against the interstellar medium for deceleration).

As I noted there is an "afterburner" concept that trades thrust for Isp. With so much Isp that implies a pretty big increase in thrust.

However this thread topic is about what SpaceX would do.

I think this is an option for them. It's got a huge  potential and it's had limited investigation. SpaceX's progress hints at significant computing power which is useful for running the sort of intensive simulation software needed to refine such designs.

Setting up and running a Mars colony is going to need a serious ongoing logistics effort. Obviously good planning will reduce that effort but it seems clear some kind of "staged" architecture of getting stuff to LEO, then getting it to Mars, then to the surface is the way to go. The "middle bit" between Earth and Mars is the question.

[Robotbeat]

But getting cargo from Earth vicinity to Mars vicinity doesn't take much delta-v. I don't think fission fragment makes ANY sense for that trip. NTR makes more sense than fission fragment for that sort of thing.

A big, powerful ISRU capability on the surface of Mars, tens or later hundreds of Megawatts, could enable lots of flights to bring stuff sent on an Earth-to-Mars ballistic trajectory down to the surface without expensive hardware needing to make the trip. For cargo, you could just keep the rocket stages on each side, near Earth to send stuff to Mars (then do a quick burn to return immediately back to Earth for refueling and to send another package on a ballistic trajectory) and at Mars to capture the packages. A rocket stage or lander with a very good mass fraction could do this quite efficiently, making very good use of the same hardware used hundreds or even thousands of times.

That is one clever way to get a lot of reuse out of MCT and BFR, but it'd only work really for cargo because of safety and the long trip times.

[john smith 19]

But getting cargo from Earth vicinity to Mars vicinity doesn't take much delta-v. I don't think fission fragment makes ANY sense for that trip. NTR makes more sense than fission fragment for that sort of thing.

That is one clever way to get a lot of reuse out of MCT and BFR, but it'd only work really for cargo because of safety and the long trip times.

The question was SpaceX and nuclear propulsion.

Surface nuclear (and the work at LANL has plans up to 2MWe IIRC) is another topic.

The problem is that existing nuclear reactor options for propulsion are complex and a lot of the initial work would be "archeological engineering." The cycle of an FFR is relatively benign in terms of temperatures and neutron generation which should simplify the shielding, fueling and moderating.
The thing is NTR has very poor T/W, although some very speculative designs were talking about 40:1 even if that were possible I'm not sure the Isp of about 2x that of LO2/LH2 is worth that kind of money.

But an Isp of 100 000 secs? That's about 3x the Isp of an ion thruster and a thrust in 10s of Newtons (has anyone actually built a 1N electric thruster yet?)

Using a sub escape velocity upper stage to push payloads to Mars is a neat trick I first saw in the REL "Troy" study for a Skylon based Mars expedition. IMHO it's one I think should be in the tool box of any   exploration architecture programme manager for wherever you're going and with whatever engine.