subcritical gas core reactor idea

[Asteroza]

A terrible idea occurred to me recently, that was half-baked enough to merit opinions. To make it make some sense, here is the rough train of thought that brought forth this concept:

The chinese recently announced more information about their ADANES project, which is an accelerator driven subcritical reactor concept using a few new tricks. Notably the accelerator portion has a new trick, using microscopic SiC coated tungsten grains as the particle accelerator's neutron spallation target. This gives the higher heat handling of a solid target, with the pumpability of a liquid, the aim being to handle beam trip transient heat shocks.

The thermal shock for some reason reminded me of a Popular Mechanics article a ways back (june 1998) for a nuclear reciprocating engine, using fissile embedded pistons with a steam moderator/working fluid, referenced as NPTRE. Interesting in it's own right, but having a piston stuck at top dead center would suck.

https://www.osti.gov/scitech/biblio/760128-nJ9K8C/webviewable

Which led down a rabbit hole to pulsed gas core reactor configurations where the gas is mixed in with the moderator

http://ufdcimages.uflib.ufl.edu/UF/00/08/41/76/00001/UF00084716.pdf
http://www.gera-e.com/technology/papers.html

The GERA paper is interesting because they propose a Wankel rotary with external "spark" neutron source, which could easily be a packet switched particle accelerator having particle bunches routed to different spallation targets on demand. That, and switching to a LiquidPiston style inverse wankel layout may be advantageous...


So this finally lead to the weak spark of an idea. These are effectively using reciprocating pumps to achieve their compression, and the fission reaction mainly occurs at peak compression, implying a subcritical situation otherwise. Are there any other compressor configurations that can achieve near instantaneous compression states, and preferably rotary based?

Well, a ramjet configuration comes to mind, notably something like a rotary ramjet, with a shared open combustor zone like

https://web.archive.org/web/20160319113542/http://www.camusdesign.ca/FR/Recherche/Ramjet.html

The peak compression for a ramjet inlet generally is at the inlet throat, creating a normal shock. This normal shock zone would be the region where criticality in the fissile gas and moderator occur. The peak heating occurs there, and residual heating from fissioning occurs past that in the subsonic "combustor" zone of a ramjet, finally wrapped up by ramjet exit nozzle (supersonic expander) providing shaft power and driving the compressor ram inlet.

Assuming you could get a closed cycle brayton engine setup to work with an appropriate heat exchanger cooling the reactor gas, how sane is this? Will it still need an external "spark" neutron source near the normal shock zone of the ramjet as the gas is too subcritical otherwise? This setup favors direct shaft power, but how usable would that be?

[Robotbeat]

I like your thinking.

[JasonAW3]

Too bad that we can't develop an internal combustion engine that could pressurize a Tritium-Deuterium mix to the point of spontaneous fusion combustion.

[Asteroza]

Too bad that we can't develop an internal combustion engine that could pressurize a Tritium-Deuterium mix to the point of spontaneous fusion combustion.

Wait, wasn't focus fusion or a similar fusion startup doing a spherical piston compression setup (effectively cylinderless?)

Current research into rotary ramjets is probably strongest with RamGen, which who got sucked into Dresser-Rand for their CO2 compressor using a ram inlet only.

Continuing my bad idea, it primarily is shaft power in output (which means collocating a generator in a neutron rich environment which is difficult at best, nevermind what the shaft cracking will look like under neutron embrittlement), but since it uses a supersonic expander nozzle it has a high speed exit flow. If desired reactor output is electricity, I wonder if some kind of MHD setup is possible? The gas mixture for example might be fissile gas and H2 moderator, but MHD usable conditions may not exist downstream of the expander due to cooling (necessitating seeding?).

[JasonAW3]

Too bad that we can't develop an internal combustion engine that could pressurize a Tritium-Deuterium mix to the point of spontaneous fusion combustion.

Wait, wasn't focus fusion or a similar fusion startup doing a spherical piston compression setup (effectively cylinderless?)

Current research into rotary ramjets is probably strongest with RamGen, which who got sucked into Dresser-Rand for their CO2 compressor using a ram inlet only.

Continuing my bad idea, it primarily is shaft power in output (which means collocating a generator in a neutron rich environment which is difficult at best, nevermind what the shaft cracking will look like under neutron embrittlement), but since it uses a supersonic expander nozzle it has a high speed exit flow. If desired reactor output is electricity, I wonder if some kind of MHD setup is possible? The gas mixture for example might be fissile gas and H2 moderator, but MHD usable conditions may not exist downstream of the expander due to cooling (necessitating seeding?).

Why not use the expander to do double duty as part of the MHD generator system?   The placement is good, so, unless there is some technical reason that I don't know about, it should be doable.

[ChrisWilson68]

The GERA paper is interesting because they propose a Wankel rotary with external "spark" neutron source, which could easily be a packet switched particle accelerator having particle bunches routed to different spallation targets on demand. That, and switching to a LiquidPiston style inverse wankel layout may be advantageous...

I don't know enough to judge most of this proposal, but I have a minor nit to pick: calling a particle accelerator "packet switched" is an abuse of the term.  The whole point of using the term "packet switched" is that it means that each packet contains within it address information and the switching is done based on looking into each packet and routing based on the address information in the packet.  For the term "packet switched" to make sense here, the particles themselves would have to carry address information which would be used to determine which target they were routed to.  Clearly, that would be a terrible idea, and it's clearly not what was meant here.

If you have the choice of where to route things coming from a sideband channel rather than from the things being routed themselves, it's not packet switched.

[Asteroza]

Why not use the expander to do double duty as part of the MHD generator system?   The placement is good, so, unless there is some technical reason that I don't know about, it should be doable.

The expander could be a good MHD duct (since it would likely be itself a duct) assuming you could get a good mag field in place and the appropriate gas conductivity conditions. The problem here being if the fissile/moderator gas mixture is at near plasma conditions, the gas gets flung against the outer wall of the rotor/duct, which could lead to bad erosion. Which implies things like a SiC or some other ceramic coated outer wall for the duct, and duct walls (which are radially oriented) being the conductive/magnetic surface (which could be done if the duct walls are under light tension due to the outer wall/rotor outer rim being compressive).

I don't know enough to judge most of this proposal, but I have a minor nit to pick: calling a particle accelerator "packet switched" is an abuse of the term.  The whole point of using the term "packet switched" is that it means that each packet contains within it address information and the switching is done based on looking into each packet and routing based on the address information in the packet.  For the term "packet switched" to make sense here, the particles themselves would have to carry address information which would be used to determine which target they were routed to.  Clearly, that would be a terrible idea, and it's clearly not what was meant here.

If you have the choice of where to route things coming from a sideband channel rather than from the things being routed themselves, it's not packet switched.

Fully guilty of this as protons in a particle accelerator are not self describing. As an analogy though, the movement of discrete groups of particles in a particle accelerator (this grouping appears to be called bunches) which are path switched by particle accelerator switching gear may be better understood as a network by some people. Though calling it a distributor cap analogue may have been more precise, in the context of hitting separate neutron spallation targets for each cylinder/rotor in a piston setup in an appropriate order relative to piston timing.


I also had a problem point thought. What happens if the expander overexpands the exhaust? Would you have mach diamonds? Would mach diamond shocks be sufficient to cause criticality in the gas stream? Does that mean you have some sort of delayed fission afterburner effect within mach diamond shocks (gives new meaning to the term reheat...)?

[Hanelyp]

A quibble about the reaction rate: the rate of change in reaction rate is a function of criticality.  Fissile gas being compressed will start fissioning with an accelerating rate as it is compressed past criticality, and reach a peak when it  passes criticality going the other direction.  So the peak reaction rate would not be at the compression shock but shortly downstream.

[ChrisWilson68]

I don't know enough to judge most of this proposal, but I have a minor nit to pick: calling a particle accelerator "packet switched" is an abuse of the term.  The whole point of using the term "packet switched" is that it means that each packet contains within it address information and the switching is done based on looking into each packet and routing based on the address information in the packet.  For the term "packet switched" to make sense here, the particles themselves would have to carry address information which would be used to determine which target they were routed to.  Clearly, that would be a terrible idea, and it's clearly not what was meant here.

If you have the choice of where to route things coming from a sideband channel rather than from the things being routed themselves, it's not packet switched.

Fully guilty of this as protons in a particle accelerator are not self describing. As an analogy though, the movement of discrete groups of particles in a particle accelerator (this grouping appears to be called bunches) which are path switched by particle accelerator switching gear may be better understood as a network by some people. Though calling it a distributor cap analogue may have been more precise, in the context of hitting separate neutron spallation targets for each cylinder/rotor in a piston setup in an appropriate order relative to piston timing.

Yeah, a distributor cap is a good analogue.  In data networking, this is called TDM (time domain multiplexing).  Each channel is allocated a repeating time slice and all the data in that time slice goes to a particular destination.  It works really well when the bandwidth requirements for each channel are very regular.  It doesn't work so well when the bandwidth requirements are bursty.

[Asteroza]

A quibble about the reaction rate: the rate of change in reaction rate is a function of criticality.  Fissile gas being compressed will start fissioning with an accelerating rate as it is compressed past criticality, and reach a peak when it  passes criticality going the other direction.  So the peak reaction rate would not be at the compression shock but shortly downstream.

Which in some ways is the desired outcome, because you want your max heating to occur in the subsonic "combustor" zone of a ramjet anyways. At least, in trying to implement a ramjet style setup using an unconventional (and depending on implementation, not entirely internal) heat/pressure source. The conventional combustion equivalent is an air/fuel premix ingesting ramjet (carbureted?), with inlet normal shock ignition, full combustion in the subsonic zone, then expansion/acceleration in an exit nozzle.