FRESH
Hacker News
Hacker News Hacker Newshttps://en.wikipedia.org/wiki/Sodium-cooled_fast_reactor
[0] https://www.nrc.gov/reactors/new-reactors/advanced/who-were-...
Todays U.S. meeting "Roundtable on Ratepayer Protection Pledge" with the U.S. President himself leading that meeting garnished commitments from Big Tech as it relates to energy. In time Big Tech Energy divisions will be thing and some citizens will be paying their utilities bill to them.
In a thermal reactor, reactivity is maintained by a carefully designed lattice of fuel elements and moderator. Disrupt this lattice and reactivity goes down. Thermal neutrons are also highly absorbed by certain neutron poisons with resonances that enable neutron capture at low energy; these can be added to shut down any potential reaction.
Fast reactors aren't like that. If fuel rearranges (for example, by melting and flowing into coolant channels) reactivity can increase. A fasts reactor will have ~100 times the "bare core" critical mass of fissionable material in it, so there's plenty of room for serious rearrangement to bring fission material into a prompt fast supercritical configuration.
That by itself could give you an explosion. But if the explosion then compresses some other part of the system beyond supercriticality, one could get an even more serious explosion. The possibility with something with a yield in the kiloton range can't easily be ruled out. This would be far worse than Chernobyl.
The fast reactor concepts I've seen deal with this by saying "our design can't ever melt down". Color me skeptical on that, and defense in depth says you don't believe such claims when failure could be so catastrophic. Even if regulators can be convinced (or be made to say they are convinced), the first experience that indicates the assumption wasn't true will lead to all reactors of that design being permanently shut down. This would be a serious financial risk to anyone thinking of building them.
If I were dead set on a fast reactor I'd look at something like a fast MSR (chloride salt) where such rearrangement could be ruled out.