[breakyerself]

Chernobyl design was never in use in the US, but nuclear went through a long period of near universal public opposition to its expansion because of the high profile disasters that it caused.

Now the cost of solar and storage are dropping at a rate I doubt nuclear is ever going to make a significant comeback. I'm not opposed to it, but I wonder if the economics will ever be favorable even with regulatory reform.

[ch4s3]

> Chernobyl design was never in use in the US

Commercially. Several early test reactors were essentially just graphite moderated piles not unlike Chernobyl, but they were abandoned for a reason.

[mikkupikku]

Graphite moderated reactors are broadly fine, the problem was with some technical specifics of that specific reactor design, and the operational culture that surrounded it. After Chernobyl, those flaws were corrected and operation of other RBMK reactors has continued to this very day, with no repeats.

[ch4s3]

That's good additional clarification, I only meant to point out that graphite moderated, water cooled reactors had existed in the US and UK.

[cyberax]

Graphite reactors are not a good idea because they inherently have, well, graphite. That can burn.

The worst possible case for water-moderated reactors is uncontained meltdown. And it's not _terribly_ horrible. You will get contamination with volatiles, mainly cesium. But there's not a lot of it in the reactor, so it'll affect only a small area around the plant. Some fuel might get initially mobilized by steam explosions, but again, only a fraction.

The worst case for a graphite reactor is an uncontained core fire. That can burn for weeks and spread a significant part of the fuel as particulates over large territories (Chernobyl).

Is it likely? Nope. But there are black swan events: earthquakes, mega-hurricanes, meteorite strikes, Godzilla attacks.

[WorldMaker]

Chernobyl may have done a lot to inflame cultural imagination of what could happen in the worst cases, but the US still had its own high profile disasters like Three Mile Island.

https://en.wikipedia.org/wiki/Three_Mile_Island_accident

[patmorgan23]

I would hesitate to call Three Mile Island a disaster, it was certainly a nuclear accident. A reactor was damaged, but no one was injured and an absolutely miniscule amount of radiation was released. The other units at the plant continued to operate until quite recently (and might actually be starting up again).

[LorenPechtel]

A "disaster" that killed nobody.

Likewise, an even bigger "disaster" at Fukushima--that killed nobody. (The deaths from the evacuation are not deaths from the incident--they wouldn't have died if they had stayed put.)

[wolvoleo]

Human deaths aren't the only measure of a disaster. The pollution was huge and even detected on the other side of the Pacific.

[LorenPechtel]

The pollution from Fukushima was very minor, blown all out of proportion by the reporting. It could be detected on the other side of the Pacific because the background levels are so low. But we can detect it far, far below the point of meaningful risk.

I don't know the numbers for Fukushima, but let's consider Three Mile Island. Same basic problem--some radioactive noble gas needed to be released to avoid trouble (and they actually released it rather than panic.) You are standing at the fence line, what do you do? Let's say you evacuate....hey, there's a street here. Cross it? Nope--it was more dangerous to walk across one ordinary street than to stay put.

[anonymars]

Now let's do coal and oil/gas

[wolvoleo]

No let's do renewables.

Nobody's arguing for more fossil. But nuclear isn't an option either.

[anonymars]

Are they so different?

There may be good reasons not to pursue nuclear (high complexity and upfront cost), but by the overall numbers I don't think pollution or death rate make that case

[LorenPechtel]

Yes, they are. They just don't know it.

Reality: arguing for renewables over nuclear actually ends up making more fossil fuel use as the renewables are not viable.

[danaris]

I guarantee you that "let's do" was "let's look at the pollution, including radioactive pollution, from".

Because coal mining releases more radiation into the world than nuclear power.

...But yes; with the efficiency and price of solar, wind, and batteries now, nuclear makes much less sense.

[anonymars]

Indeed. Banqiao Dam would like a word

And yet we still build hydroelectric dams

[mattmaroon]

It would. People are still building some natural gas plants even despite renewables being cheaper and nuclear is far cheaper over its lifecycle than that and, other than regulatory issues, is basically better in every way.

[prpl]

There will continue to be new gas plants as long as there are coal plants which will be converted, usually around the time a major overhaul would need to be taken anyway.

[9rx]

> nuclear is far cheaper over its lifecycle than that

That is the case for base load generation, where the plant can operate near 100% capacity all the time. But that isn't were gas is usually being deployed; it being used for reserve generation. The economics of nuclear isn't as favourable in that application as it costs more or less the same to run at partial generation, or even no generation, as it does when it is going full blast.

[mattmaroon]

Right, but in the context of data centers, it’s all about baseload anyway, right? If data centers become a big driver of energy use, there will be a lot lower fluctuation between peak and trough demand.

I can imagine a future in which every data center has a little baby nuclear plant built right next to it. Watts per acre may become a significant measurement of density. Solar’s environmental impact is of course dramatically overstated by its opponents, but it won’t be when we scale it up and have to start slashing forests for it.

[9rx]

If it were simply an option between nuclear or gas for that, nuclear would, generally, be the obvious choice. But it would be quite atypical to build a gas plant to provide base power. Typically they are being built to back up renewables.

Fair point that renewables may have a practical expansion limit, but for the time being are, by far, the cheapest option so a data centre is still going to prefer that source of power to the greatest extent possible, thereby leaving gas/nuclear only as reserve — of which nuclear has not proven to be cost effective at. Geothermal, hydro, etc. are hard to beat, but where you aren't sitting on the perfect environment, generally speaking, wind+solar+gas is about as good as it gets on a cost basis.

[mattmaroon]

Yeah, and I'm all for all of it. I just can see a future in which nuclear (through some combination of regulatory reform and new technology) ends up becoming cost-feasible and fossil fuels fade away.

[jerlam]

Nuclear might be better and cheaper over it's entire lifecycle; but given that the starting costs are so high, the time to build is so long, and the US has serious problems with cost overruns in public projects, as well as the fickleness of both government and public opinion, I don't expect another plant to be built.

[mattmaroon]

Well we were speaking of costs in a hypothetical future in which regulations are sane. I don’t expect that to happen either but if it did, the economics would work.

[LorenPechtel]

Renewables are only "cheaper" because the market forces major subsidies. The reality is the value of renewables is the fuel they save. They do not replace the generators or any of the other stuff.