A Canadian company has gotten funding to move forward on a molten-salt reactor. I think a lot of sensible people are realizing that if carbon really is a problem, nuclear is the solution, despite the insanity of people like Naomi Oreskes.
But my question is: Would this be a useful tech for space, either for electric power generation or propulsion? The company could do a spin off called Extraterrestrial Energy.
Off the top of my head:
1) You do need to remove some fission products from the salt periodically, and I think they have to helium-sparge the Xe-135 in some of the transient modes. That either limits the design life of the space version or forces you to get awfully creative with the chemical engineering.
2) Not sure how much the TE design relies on convection for circulating the salt. The reactor is thin and tall, which leads me to believe that convection is a non-trivial effect.
3) The whole freeze plug idea, for both fail-safety and servicing, isn’t going to work. Probably not an issue with a short enough design life.
4) IIRC, the neutron economy for MSRs isn’t that great, so they need graphite moderation in the core. I wonder if graphite moderators are tough enough to survive launch.
5) I suspect you have to go 2-loop with an MSR, because the primary coolant was critical fuel just moments before it hits the heat exchanger. Not the sort of thing you want in close proximity to your avionics.
6) On the plus side, you can be as conservative as you want on launching MSR fuel, to achieve whatever safety profile is necessary. You could even launch parcels of frozen salt in cubesats and pour ’em into the reactor on-orbit. That’s hard to do with solid fuel assemblies.
Rand, are you playing around with board settings?
You mean the blog? No, haven’t touched anything. I do need to do a WordPress update, though.
Graphite moderated thermal reactor cores are necessarily fairly large. Aren’t all space reactors fast reactors?
” Would this be a useful tech for space, either for electric power generation or propulsion?”
Better, it would be an excellent demo of the usefulness of space resources. You see, one of the bugaboos still afflicting regulatory thoughts about MSRs is corrosion-caused microcracks in the plumbing, even with corrosion resistant alloys. In the freefall vacuum of space strong, lightweight carbon-in-carbon composites can be made and used for plumbing, with something to shield it from the fluoride salts. Now, it just so happens that CC asteroids contain *lots* more Iridium than most PGM ores here on Earth do. True, Iridium is heavy, but a thin coat of it to stop corrosion wouldn’t be. Remember that graphite gets *stronger* until the temperature hits about 2000ºK. With a nice Iridium coating. Use supercritical CO2 as the working fluid. Use dynamic radiators. that combo would give us a lightweight nuclear power unit, that also tells the world just how much space resources can do for them!
Transatomic believes they can use a zirconium hydride moderator instead of graphite in a molten salt reactor:
http://www.transatomicpower.com/wp-content/uploads/2015/04/transatomic-white-paper.pdf
“TheRadicalModerate
January 13, 2016 at 1:43 PM
Off the top of my head:
…
1) You do need to remove some fission products from the salt periodically, and I think they have to helium-sparge the Xe-135 in some of the transient modes. That either limits the design life of the space version or forces you to get awfully creative with the chemical engineering.”
Why use He? All you need is a liquid (molten salt) to gas phase interface. The Xe will partition to the gas phase. It’s simpler with gravity, but gravity is not required. This is one of the charms of the molten salt design. Removing Xe is comparatively trivial compared to anything where fuel is in the solid phase. And if you do need He for more rapid purging, don’t forget that the reactor will be making He. In fact, it will make a lot more He than Xe. There’s going to be a lot of alphas with all those reactions + decaying nuclei.
“3) The whole freeze plug idea, for both fail-safety and servicing, isn’t going to work. Probably not an issue with a short enough design life.”
How so? It’s how they shut down the Oak Ridge reactor routinely for many years (6, iirc). Turn off the cooling fan for the plug, the plug melts, the liquid drains into a holding tank, salt solidifies. No pressurization, and the tank can be designed to handle the heat from the decay of fission products. In the morning, melt the salt, pump it into the reactor, re-establish the plug and you’re in business again. Thermally stable. In this case, gravity is a plus, but I’m guessing a spinning design could do the same thing. The nice thing is that if you lose power, the reactor automatically safely shuts down. Ultimate in fail-safe.
1) I had the wrong sparge gas, and the wrong purpose for the sparging. Here’s the relevant Wikipedia blurb (so it’s gotta be true!):
As for Xe-135 removal, I’m skeptical that it works without convection. I assume that you have to wait for bubbles to consolidate and then get buoyed up to some “surface” before they’ll separate from the melt. No gravity, no buoyancy.
Any sort of closed-loop management of sparge gases is going to be dicey.
2) The freeze plug is gravity-fed–that’s why it’s fail-safe. You can certainly substitute a battery- or APU-powered pump, and a long-duration power loss is likely to be lethal in space anyway.
Which brings up an interesting question: Since you obviously launch the reactor cold, with solid salt, how much energy does it take to create the melt? That energy has to come from batteries or from a fueled generator of some sort.
RTG?
The odds of my dropping a few decimal points here are extremely high, so check me:
LiF melting point is 845°C, heat capacity 2.4 J/gK (that’s actually for FLiBe in liquid phase, but it’s what I could find), and heat of fusion of about 990 KJ/kg. If I SWAG the heat capacity to be the same in liquid and solid and assume that we have to heat from 0°C solid, it looks like you need about 3 MJ/kg to get it melted. Let’s say we need 500 kg of salt/fuel for a reactor (a SWAG) and we get about a 5% heat transfer efficiency. (That’s a big SWAG. I’m assuming electric heating and hefty losses to thermal radiation, especially as the whole thing gets hotter. Anybody have insulation coefficients and passable differential equations skills?) Then we need 30,000 MJ for startup.
Assume we have to start the reactor within 2 weeks of launch (another SWAG). Then we need about a 25 kW RTG. Modern RTGs (much smaller ones) run with We/kg ratios of about 5. So you’re looking at about 5 tonnes of RTG to melt the salt. That’s not so good.
On the other hand, if your entire RTG contains nothing but Pu-238 and you can heat the salt directly (preventing, of course the core neutrons from smacking into the Pu-238 after startup, which would probably be unpleasant), insulating the whole shebang with unobtanium to limit the heat loss to 50%, then you only need about 5 kg of Pu-238. So we only have a factor of 1000 difference. But I don’t think we’re talking about about anything we’d call an RTG in the latter case.
Hmm. What’s the specific impulse of the molten fluid?
You can’t use the salt as reaction mass–it has the fissile stuff dissolved in it. If you’re going to be that profligate (and dirty), you’d do better with a dusty plasma nuke.
If you wanted to use a MSR to power a nuclear rocket, I expect it could operate at higher temperature than a conventional solid core reactor. But reactor power to mass may not be as good.
“As for Xe-135 removal, I’m skeptical that it works without convection. I assume that you have to wait for bubbles to consolidate and then get buoyed up to some “surface” before they’ll separate from the melt. No gravity, no buoyancy.”
Well, I’m just remembering thing gleaned from some Energy from Thorium presentations, but I believe Xe is handled by continuously pumping the fluoride salts (U fuel and Th breeder), and there’s a loop where there’s a low pressure liquid/gas interface, which continuously pumps out the Xe (i.e., they don’t rely on convection). Should be fairly simple with gravity, but in zero-g, maybe a central vacuum pump in a centrifuge that has liquid fluoride salt flowing around the rim would do the trick. IIRC, one of the charms of fluoride salts is their stability and that they don’t corrode the vessel.
To get the xenon out of solution, you have to make it boil out.
The nifty thing about low pressure is that it will cause dissolved gases to boil. When a bubble forms in a liquid, it’s essentially because the partial pressure of a few dissolved atoms is big enough to overcome the pressure of the liquid in which they’re dissolved. That pressure is the sum of the total pressure of the gas “above” the gas/liquid interface, plus the hydrostatic pressure of the liquid at a particular “depth”. After that, they’re buoyant, and they make their way “up” to the big gas/liquid interface.
I’m clueless how this works in microgravity, where you can form bubbles like crazy (lots of vacuum, no hydrostatic pressure), but how you get them buoyed up to separate from the liquid is beyond me. Run them through a centrifugal pump?
Suffice it to say that the chemical engineering in a microgravity MSR is going to be wildly different than one sitting on the ground. Boiling won’t work. Sparging won’t work.
Maybe the answer is to spin the whole friggin’ reactor on a tether. Hmm…
OK, now I’m falling down the rabbit hole on this, because I suddenly don’t understand how tank pressurization works in zero-g. If you pressurize with helium under acceleration, that works fine, but how do you do restarts on-orbit? Doesn’t the helium and the LH or LOX just mix together into some weird colloid? What’s the secret sauce here?
You typically do a settling burn with cold gas. The ACES stage, with its Integrated Vehicle Fluids, will have gaseous LOX/LH2 thrusters. It doesn’t need helium, because it does autogenous pressurization.
Ah. Always wondered what those were for. Thanks!
I want to express my disagreement with your assertion that nuclear is “the solution”. Nuclear’s economic prospects do not look good. The big problem is nuclear’s lousy history of cost reduction w. experience (basically, it hasn’t seen any). Apologists have blamed this on regulation, but the effect is pretty much universal, even in countries that greatly supported nuclear (like France).
The much ballyhooed “nuclear renaissance” in the US fizzled out when the cost estimates proved wildly optimistic. The handful of reactors that got started here are already experiencing large cost overruns. Small modular reactors (molten salt or otherwise) have been presented as a savior, but their economics doesn’t look better, as they involve more material and worse economies of scale.
In contrast to this, renewable costs HAVE been dropping, with proven experience curves. This means that by the time a new nuclear reactor type can be demonstrated and tested, years from now, renewable costs will likely be even lower than they are now (and they are already getting pretty low).
“In contrast to this, renewable costs HAVE been dropping, with proven experience curves.”
Renewables could be free and still be uneconomical. That is what you people simply don’t get.
With the exception of Space Solar and Hydro, you can’t base load it or dispatch it. Absent those two, you must be able to store it and the storage issue is Godzilla in the room, a problem so gigantic that there is simply no solution on the horizon as far as the eye can see.
“a problem so gigantic that there is simply no solution on the horizon as far as the eye can see.”
This is simply wrong, there are actually many solutions on the horizon (or nearer). If you project forward the cost of grid-connected energy storage using the learning curves those technologies have demonstrated, they also will become quite cheap by the time substantial penetration has occurred. The real question is not whether cheap storage will be available, but rather which of the technologies will come out ahead.
There are growing markets for storage going forward, even before those longer time prices are reached, and these markets will drive the technologies down the learning curves.
Yeah, I’ll believe the storage “solutions” when I see them.
Li-ion batteries have come down dramatically in cost, with a learning rate between 15 and 21% per doubling of cumulative production. That rate is sufficient to make Li-ion batteries cheap enough for diurnal storage if their production has expanded to the scale needed for that globally.
But there are other technologies (flow batteries, zinc-air batteries, others) that may be even less expensive. Eos Energy System’s Zynth product is being advertised as going for $160/kWh with a 5,000 cycle lifetime, for example, and to be less expensive than pumped hydro.
http://www.eosenergystorage.com/products/
BTW, you might want to look up what Elon Musk has to say about space solar power, if you think diurnal variation is going to be an unsolvable problem.
So basically you’re saying that we’ve got a race between one currently uneconomical technology (Gen III+ nukes) and another currently uneconomical technology (high-scale grid storage systems). And we’ve got people working hard to get their technology over the finish line first.
That sounds like a good thing to me. As long as we don’t intervene to help or hinder one or the other, two evolving technologies are better than one, especially when either of them could hit a show-stopper problem.
TRM: Gen III+ nukes are currently uneconomical in all uses. Batteries have numerous lesser markets in which they are profitable that are driving their development. They are also much easier to develop than nukes. So it’s not really an equal comparison.
“To get the xenon out of solution, you have to make it boil out.
The nifty thing about low pressure is that it will cause dissolved gases to boil. When a bubble forms in a liquid, it’s essentially because the partial pressure of a few dissolved atoms is big enough to overcome the pressure of the liquid in which they’re dissolved. That pressure is the sum of the total pressure of the gas “above” the gas/liquid interface, plus the hydrostatic pressure of the liquid at a particular “depth”. After that, they’re buoyant, and they make their way “up” to the big gas/liquid interface.”
Sorry, no. Xe dissolves in water . It comes out of solution from water when the temperature rises without forming bubbles. All it takes is for solution/exsolution to take place at the liquid/gas interface. This is a slow process if and only if the transport mechanism of Xe to the interface is dominated by diffusion. If, however, you use advection via a stirring/pumping system, this can be quite rapid and efficient, especially if you increase the surface are to volume ratio (e.g. O2, CO2 exchange in your lungs). Think of it this way: divers live because they exsolve N2 from their blood stream WITHOUT forming bubbles. Sure it’s slower, but it still takes place reasonably quickly.
Even advection isn’t going to work unless you can get the entire volume of fluid pumped to some kind of liquid/gas interface, and that’s not easy. To get lots of fluid to turn over at that interface, you have to apply some force to it, which in space means getting it to hit some solid barrier. That’s not a liquid/gas interface. Furthermore, you need to get the Xe-135 out of solution pretty quickly, or you might as well let it decay to Cs-135 or +n1 to Xe-136 and have done with it.
It occurs to me that we may be arguing about the wrong thing. Xe-135 poisoning is mostly a transient problem as you increase power. Reducing that transient is important for a dispatchable grid nuke (which is why TE brings the issue up) but I’m not sure how important it is in a space application. You have to be able to dump the full heat load no matter what, so maybe you just run the thing at constant (max) power all the time.
Even if you can dodge the xenon issue, you’ve still got lots of other things you need to cleanse from the salt. Those are also going to require wildly different mechanisms than they would in a gravity field.