sevenperforce

Whenever I think about next-generation nuclear reactor designs, I’m reminded of the Cecil Kelley criticality incident. There was an unfortunate excursion in a plutonium reprocessing lab that claimed the life of one of the technicians when an extraordinarily high level of dissolved plutonium was introduced to a vat in the lab. The operator turned on the vat to mix it, not realizing how high the plutonium levels were, and the plutonium-enriched layer of fluid at the top was pulled down into a vortex, creating a critical mass: One way to make an “inherently safe” reactor would be to use a liquid fuel solution that could have its criticality controlled by geometry through rotation, thus requiring a constant source of energy to maintain operation. Lose power, and the fluid spins down quickly and can no longer maintain critical geometry.

Efficiency loss due to physical separation hasn’t stopped molten salt reactors before; there are lead-cooled salt reactors, for example. And I would have to imagine that there would be a geometry solution to salt deposition, right? Helium is not known for being particularly dense….

Yep. This proposal quickly reduces to a pulsejet engine, which of course works very well for backyard projects or 1940s-era cruise missiles but is incapable of accelerating to spacecraft-useful speeds.

Is there anything you can use in place of graphite with the same design? That's one of the things we all love about molten salt reactors -- they are SUPPOSED to be melty, and if they get too hot they simply melt through the freeze plug and are dumped. It's also fairly low-pressure. One of the challenges in a molten salt reactor is that the salt is the coolant and so the salt has to be pumped around in a loop, which is very challenging to get right. Would it be possible to have a helium-cooled molten core reactor?

To start with, keep in mind that you want to conceptualize everything in terms of field interactions, not interaction between the object and the field. The electrically charged particle in a magnetic field experiences the Lorentz force because its electrostatic field is moving through a magnetic field and the motion of one field through the other produces a field interaction. a So, what does the magnetic field of a monopole look like? Unlike dipole electrostatic fields, magnetic dipoles do not arise from the aggregate of negative charges on one side and positive charges on the other, but rather arise from the intrinsic dipole magnetic moment of elementary particles. The field of a magnetic monopole would be a monopole magnetic moment, which is NOT merely the bisection of a dipole moment, but something else entirely which we haven't ever observed. But our best guess would be to analogize to the Coulomb and Lorentz interactions. This is the Lorentz equation: and this is the Lorentz equation with the monopole extension: So just as a moving electrical charge in a magnetic field experiences a force which is the cross product of its charge*velocity and the vector strength of the field, a moving magnetic monopole in an electric field experiences a force which is the cross product of its magnetic charge*velocity and the vector strength of the field. And so, by analogy, a monopole in a magnetic dipole field would act the same as an electrical charge in an electric dipole field. In other words, yes, your intuition is right.

It’s worse than that. As I pointed out upthread, the maximum detonation speed in a fuel-air mixture is 1,800 m/s, which means the greatest speed you can obtain using fuel-air detonations is 1,800 m/s. I completely reverse my prior stance. This is an excellent idea. Except you will want to use many small fuel-air bombs instead of large ones. By making the bombs smaller, you will be able to wrap the pressure plate around the explosion in a cone shape and thus make it more efficient. You will also want to have an extra hole in the center so that air can flow through to participate in the fuel-air bomb. You will need some sort of reverse-pressure-plate at the front to compress that air so that it comes in and mixes well with the fuel. Instead of dropping the bombs in one by one, you should make them into a liquid that can be sprayed continuously into the flow of compressed air so that it detonates continuously. Congratulations, you have invented the ramjet.

Is it like thermonuclear, where efficiency continues to go up no matter how big you get, or is there a power level where you basically max out your benefits? Modular-distributed makes a lot of sense for a lot of reasons, but I find myself wondering whether we could ever see the entire eastern seaboard running on one offshore megareactor.

I would like to get a better idea of the scaleability characteristics of nuclear. Is bigger better/more efficient? Is there a lower limit?

On suggestion by @CastleKSide, I decided to start a thread on nuclear reactor design, particularly with a focus on near-term next-generation reactors, but generally open. This will split the discussion on the SpaceX thread about startups working in modular reactors. Tagging possibly-interested parties @mikegarrison @KSK @JoeSchmuckatelli @RCgothic @CatastrophicFailure @Rakaydos @AHHans

Also, obligatory:

No, it will not work. The maximum detonation wave speed in an ideal thermobaric explosive is around 1.8 km/s. This means that once you are moving at 1.8 km/s, blowing up more fuel-air bombs behind you won’t do anything, because you’re moving away faster than the shockwave can reach you. A thermobaric Orion will not get you anywhere. There will NEVER be an occasion or situation where a "non-nuclear Orion" makes ANY sense. Never. Thermobaric explosives are effective because they are able to use atmospheric oxygen to support detonation, rather than depending on an oxidizer chemical premixed with the reducer chemical to create the explosive. Since oxidizer makes up over half the weight of an explosive, you can get more kaboom per kilogram (pow per pound in freedom units) if you can use atmospheric oxygen instead. But if you're trying to propel a vehicle and you want to use atmospheric oxygen, you use a jet engine. A pusher-plate system is WILDLY inefficient. The ONLY reason for a pusher-plate is if you cannot control the release of energy from your engine, which is ONLY applicable when using nukes. If you want to use atmospheric oxygen to help you get to orbit, try Skylon.

I suppose that would be necessary, yes. “The company began to create a budget microreactor that could fit in a regular shipping container. The megawatt model will be capable of powering up to 1,000 homes and is expected to use helium for cooling instead of water. The main places of use are remote settlements, zones of natural or other disasters, and military bases.” So presumably it would be limited to a locale with water access…not scaleable or suitable for residential solutions, for example. At least that’s what I would imagine. At 1 MW and a Braxton efficiency of ~65% (enabled by a high-temperature coolant), that’s 1.53 MW of thermal power that needs to be rejected. I ran the numbers (after typing that) and it’s not as bad as I thought. If your water heat sink loop raises water temperature by 20 degrees Celsius, you only need a flow of 1100 liters per minute.

Here's the typical schematic for a helium-cooled reactor. Cool, pressurized helium is pumped around the reactor and through the reactor core, gaining thermal energy. It is then expanded through a turbine, which operates a series of compressors. The helium is cooled via heat sink and passed through the compressors in cycles to return it to its compressed, cool state. The turbine also operates the generator. I wonder what kind of heat sink is used for cooling the helium. I'm also curious to know how cool the helium is supposed to be when it enters the reactor. Helium obviously can remain gaseous at very low temperatures; is there an advantage to getting the largest temperature swing possible? Would the system efficiency be improved by a secondary refrigerant loop?

Aye. There are radioactive isotopes of helium, but the only one that could be created in a reactor by neutron capture would be 5He, which decays back into 4He by neutron emission in about 6e-22 seconds.

It was not initiated by a helium leak; it was initiated by LOX crystals freezing in the composite overwrap layers, resulting in COPV failure. Of course once the COPV failed then the helium certainly "leaked" in a very sudden and dramatic sense. The first Falcon 9 failure was also a helium "leak" in the sense that a failed strut caused the COPV to break free and fall through the LOX and then fail on impact with the common bulkhead; the rupture of the COPV released the pressurized helium. AFAIK the latter. But that's mostly "I haven't heard anyone complaining about helium diffusing everywhere" and not "I know that helium is well behaved." Helium diffuses through solids just like hydrogen does -- in fact, I believe it may diffuse a little better. However, diffusion is not what makes hydrogen so nasty. Hydrogen embrittlement happens because individually-migrating hydrogen atoms meet up inside the tank/pipe wall and decide to bond with each other, essentially producing nanoscale bubbles inside the metal, rapidly degrading its material strength. It's like when Vision phases his hand into someone's chest and then unphases his hand while it's still inside.