And at any rate, the article claims that

> The small size and large surface area-to-volume ratio of NuScale’s reactor core, that sits below ground in a super seismic-resistant heat sink, allows natural processes to cool it indefinitely in the case of complete power blackout.

Of course, the surface that allows efficient cooling here is the same surface that allows neutrons to escape, so my hunch is that it has poor neutron economy. (And of course, heat escaping through it will not turn turbines.)

> 2) refueling of this reactor does not require the nuclear plant to shut down.

That is very nice when you're trying to breed plutonium, the natural uranium fuel assemblies only have to irradiated a few weeks and then you would already need to shut the reactor down. When you can switch during operation, than there's less downtime.

So in total I guess it's a pretty inefficient reactor that's perhaps a nice addition to your weapons program.

Disclaimer: I'm of course one of the anti-nuclear types HN always tells me are only anti-nuclear because we don't understand these things.

So in total I guess it's a pretty inefficient reactor that's perhaps a nice addition to your weapons program.

This is incorrect. The NuScale reactor is not an online-refueling design like the CANDU (which is indeed easily adapted for breeding high grade plutonium simply by adjusting irradation time). It is an offline-refueling design like every other operating PWR. See this NuScale presentation about refueling operations:

https://www.nrc.gov/docs/ML1515/ML15159A311.pdf

The reactor has to be completely shut down for more than a week during refueling (slide 13). The plant can stay online because a plant contains a minimum of 4 reactor modules:

https://www.nuscalepower.com/about-us/faq

So by staggering refueling times, the plant can continuously generate at least 75% of rated output.

The problem that I was trying to show is that people try to answer a question "can a chernobyl(-like) event ever happen again?" which one simply _can't answer_. Not in a "science" way, since to answer this question you need to predict the existence (not probability!) of a literal _punctual_ event which depends on a gazillion factors outside your control (i.e. this is not simply computing a MTBF).