Hacker News Reader

Comment by acidburnNSA | Hacker News Reader

Nuclear reactor developer here. I personally have not really gotten on the small reactor bandwagon. I can see them as useful for trying out new tech but the way to really reduce costs is to pick a known and proven tech, serialize it, and increase power. I think SMRs are a response to boondoggle construction projects, but if and when a smr project goes well there will be lots of pressure to make bigger versions of the reactor.

For decarbonising rapidly we should currently be building lots of ABWR and APR-1400 and Hualong One reactors (big, modern, proven).

My take on economics of nuclear is here https://whatisnuclear.com/economics.html

View on HN

> I think SMRs are a response to boondoggle construction projects, but if and when a smr project goes well there will be lots of pressure to make bigger versions of the reactor.

I think there's some value in that right? Proving designs, rebuilding public trust, and building up all of the economic infrastructure again will be really useful for scaling up.

I'm definitely going to read your article, seems fascinating from the intro.

What about France's reactors? I hear that they standardized the design and are/were able to roll them out effeciently.

i believe that was the idea with the westinghouse design.

but it went sideways, and their recent projects had huge problems and overruns, and westinghouse ended up going bankrupt a few years ago..

Yup they're one of the best/only examples of large country-scale rapid deep decarbonization of electricity. ~50 standardized reactors in ~15 years and done.

I guess it depends on what the definition of a SMR is. A Terrestrial Energy is a 400-500MWe.

I think those kinds of reactors are just much smaller then your PWRs but its not exactly SMRs.

We should focus on GenIV designs.

However I would not be opposed to bulding traditional designs but it must be proven and have people who have done it before available.

> I think SMRs are a response to boondoggle construction projects...

This +10 (if I could).

Consider orbital-class rockets - another very advanced technology, and far less forgiving than nuclear power reactors. How much does it cost to launch (say) 10 tons of satellite on a rocket from plucky upstart SpaceX? Vs. how much did it cost back when only a few bloated "cost plus plus" mega-corporations were providing orbital rockets?

Orbital rockets are less forgiving than reactors in that they operate closer to engineering limits even when they work correctly. They are more difficult to make reliable. But they have a far more forgiving-to-neighbors failure mode. When a SpaceX test vehicle explodes only SpaceX is harmed. With reactors you can't start by testing a design that spontaneously disassembles itself and iterate toward a reliable one like the organizations that build rockets have done.

Well, true. But since reactors run nowhere nears engineering limits, they need "build, explode, iterate" R&D about as much as they need testing at 5g, or in vacuum. (Yes, I know that destructive testing of many reactor components & sub-systems is quite reasonable.)

Reply on news.ycombinator.com