Dept. of Energy signs agreements to develop small nuclear generators | Hacker News Reader

Dept. of Energy signs agreements to develop small nuclear generators

arstechnica.com

58 points3lit3H4ck3r9 comments

Excellent news, if it pans out. Here are the reactors they're working on:

Hyperion: A 25 Megawatt-electric (MWe) lead-bismuth-cooled fast reactor. Operates at a high temperature, which makes it potentially useful as a source of industrial process heat, e.g. the Haber-Bosch process for synthesizing ammonia. Has a nice negative temperature coefficient of reactivity: if it gets too hot, the reaction slows down.

Westinghouse SMR: Essentially a much-scaled-down version of Westinghouse's AP1000. Pretty conventional. At 200 MWe, it's the largest of the three.

NuScale: Another conservative pressurized light water reactor. Uses shortened versions of standard fuel rods, so the fuel supply chain should be pretty simple. Smaller than the Westinghouse SMR, at 45 MWe.

All three are small enough that they can be manufactured at factories and shipped out by boat, rail, or a particularly big truck. Their pressure vessels are light enough that they don't need to wait in line for heavy forgings from the handful of companies capable of making pressure vessels for the gigawatt behemoths. I hope that the NRC doesn't kill them.

I was really hoping to see a thorium[1] reactor in this list. Apparently, several government labs are working on a small, portable thorium reactor that would be great for this type of thing. The SSTAR[2] (small, sealed, transportable, autonomous reactor) system will produce 10 to 100 megawatts electric and can be safely transported on ship or by a heavy-haul transport truck. Perfect!

1:http://en.wikipedia.org/wiki/Thorium

2:http://www.thorium.tv/en/thorium_reactor/thorium_reactor_1.p...

I read a paper about it recently, and it states while everything is good, it will still takes years of research and development, trial and testing before it moves into deployment stage.

So unless someone with many BILLIONS decide to tackle the problem, the time it takes Thorium to get traction making it not really an alternative ATM.

About a year ago China committed a billion dollars to developing liquid thorium reactors. Australia and Czechoslovakia have a partnership with about $300 million committed so far.

Sorenson and others have estimated that a billion or two is about what we'd need, and about ten years with a strong effort.

There are a lot of variants, some more difficult than others. Instead of going for the LFTR right off, we could start with the DMSR. It's a non-breeding molten salt reactor using uranium, but about five times more efficiently than LWRs, excellent proliferation resistance, and all the safety advantages of LFTR.

I think you've misread your article. SSTAR is a not a thorium reactor; it is a liquid-metal cooled fast reactor. It's a sealed "nuclear battery" type reactor which lasts extremely long without refueling (15-30 years). It uses medium-enriched uranium, ~20% U-235 (on the border of what regulators consider "highly enriched"), far higher than commercial LWRs (~3-5%). It also burns a large amount of plutonium which it creates in situ, transmuting non-fissile U-238 into fissile Pu-239.

(Actually it looks kind of similar to one of the reactors in the DoE agreement, the one from Hyperion (developed at Los Alamos). Both are small, liquid-lead cooled fast reactors designed to never be refueled in their lifetime.)

[1] http://www.gen-4.org/Technology/systems/lfr.htm

[2] https://smr.inl.gov/Document.ashx?path=DOCS%2fSMR+technologi...

Here's a nice, succinct overview excerpted from [2]:

"The secure transportable autonomous reactor (SSTAR) fast neutron spectrum lead-cooled reactor is a concept designed to achieve the major desired attributes for the worldwide deployment market described above [5]. The SSTAR reactors are ‘right sized’ for initially small but fast growing electric grids; they provide energy security for nations not wanting the expense of an indigenous fuel cycle and waste repository infrastructure but willing to accept the guarantee of services from regional fuel cycle centers by virtue of a long (15- to 30-year) refueling interval [6]. The SSTAR initial fissile inventory is relatively large; nevertheless, the one-time initial fissile loading is substantially less than the lifetime 235U consumption of a LWR for same the energy delivery."

"Once loaded, SSTARs are fissile self-sufficient as they operate with a conversion ratio of about 1.0. As such, they provide an alternative approach to actinide management in which these nuclear materials are securely ‘stored’ in long core lifetime power reactors instead of being transmuted in advanced recycle reactors."

"The current design concept for the SSTAR, under development in the US is a 20 MWe natural circulation pool-type reactor with a small shippable reactor vessel. Specific features of the lead coolant, transuranic nitride fuel, fast spectrum core, and small size have been incorporated to achieve proliferation resistance, fissile self-sufficiency, autonomous load following, simplicity of operation, reliability, transportability, as well as a high degree of passive safety. Conversion of the core thermal power into electricity at a high plant efficiency of 44% is accomplished by utilizing a supercritical carbon dioxide Brayton cycle power converter."

It wouldn't surprise me to see these or similar small nuclear generators powering datacenters in the near future. I've always thought turning old nuclear submarines into small datacenters would be an interesting way to recycle them.

Afaik, U.S. nuclear submarines in particular are designed to run on very highly enriched Uranium (>90% U-235), which makes them both security-sensitive and expensive to fuel.

I hope we do it, AND, I hope we closely regulate them. Nuclear power is dirty. So is coal, oil, clear-cutting forests, and turning corn into fuel.

Not a big fan of the collapse thing, so little nuke plants seem to be the lesser evil.

Whatever happened to the talk about "pebble beds" -- little fuel pellets in ceramic marbles that can't pack close enough to go critical?

> Nuclear power is dirty. So is coal, oil, clear-cutting forests, and turning corn into fuel.

They are not the same amount of dirty, not by a long shot. The regulatory burden on nuclear power is psychotically out of proportion to its actual safety risk and environmental impact, and gives an unfair advantage to energy sources like coal and methane which are much dirtier and more dangerous. Beyond a certain point, safety regulations on nuclear power make us less safe. We're well past that point.

(Regarding pebble beds: China has an experimental pebble bed plant under construction. Each module consists of two 105 MWe reactors sharing the same turbomachinery for generating electricity. It looks like the plan is to get the manufacturing infrastructure working, get some operational experience actually running the things, and then start building more and possibly exporting them commercially to other countries.)

Reply on news.ycombinator.com