In this particular case, a power plant with these nuscale reactors would probably have survived a Fukushima type accident. One effect of being small is that the reactors are designed to be passively cooled after shutdown, using convection instead of pumped flow. So there is no need for emergency diesel generators to keep the coolant pumps running.
> So as long as the reactor vessel is intact it absorbs 100% of the neutrons? Going out as well as coming in?
Neutron diffusion is a statistical process, so yeah, a small amount of neutrons will escape through the reactor wall. But not enough that the neutron flux from one reactor would influence the next one in any measurable way.
> Because in a failure you’ll have only one pressure vessel between the two cores. If everything else goes right.
Sure, it's safer for reactors to tolerate flooding, if indeed these reactors actually do tolerate flooding. Definitely that would be a better design in flood/tsunami-prone areas. No design (from history, especially no reactor design) is perfect. A set of smaller reactors that were not co-located would be more tolerant of site-specific vulnerabilities in their design.
> ... if indeed these reactors actually do tolerate flooding
The reactor modules are partially immersed in a pond, the ultimate heat sink. Cooling is passive, i.e., no cooling pumps, and does not require electrical power.
>A set of smaller reactors that were not co-located would be more tolerant of site-specific vulnerabilities in their design
That depends on what risks you are trying to mitigate. IF you are talking about site specific natural disasters and freak occurrences damaging a site. Then spreading them out increases risk of failure.
If you are worried about power loss after a site destroys a site, then colocation is worse. Of the two, damage to a site is the more pressing concern.
The odds that a concentrated site has an event are much higher with colocation. But the changes of having an event at multiple locations are higher if they're spread out.
Some things that I believe would matter for colocation would be:
- chance of cascading failures
- economies of scale/safety in numbers (1 large team vs many small)
- plant-to-home efficiency (n fewer reactors due to smaller transmission losses)
