[dctoedt]

Nuclear power is expensive in no small part because of the safeguards needed to try to avert catastrophic accidents. Humans are fallible, and our best intentions can be subverted by inadequate training; fatigue; inattention; laziness; or what we used to call "a loss-of-brain accident." As a result, we can f[oul] up at any stage of design, construction, operation, or maintenance of a nuclear reactor.

(Neither Three Mile Island [0] nor Chernobyl [1] would have been so disastrous had it not been for cascading sequences of human error.)

Expecting nominal performance by people or machinery is ... unwise; as Admiral Rickover famously said, "you get what you INspect, not what you EXpect."

All that adds to costs.

Source: Former Navy nuclear engineering officer, qualified as [chief] engineer aboard the eight-reactor aircraft carrier USS Enterprise.

[0] https://www.energy.gov/ne/articles/5-facts-know-about-three-...

[1] https://www.iaea.org/newscenter/focus/chernobyl/faqs

[pdonis]

Newer reactor designs with entirely passive safety features remove the need for operators to take particular actions in the event of a problem, so "loss of brain accidents" are no longer possible. (Whether those designs can support the requirements for military use, which are quite different from those for civilian power generation, is a different question.)

That said, civilian nuclear power, at least in the US, was never operated with the same attention to detail and the same intolerance for f[oul]ups as Navy nuclear power; the kinds of mistakes that operators made at TMI (don't even start about Chernobyl, that's a whole other level of insanity) would have gotten Navy nuclear trainees kicked out of the program long before they were allowed to do anything with an actual reactor. (When I was an Engineering Duty Officer working at Norfolk Naval Shipyard, I saw a reactor officer get fired for an administrative error that probably would not even have been on the radar in a civilian plant.) So that can't be a significant part of the explanation of why civilian nuclear power is so costly in the US.

The high cost of civilian nuclear power in the US has always been primarily due to politics: things like unreasonable waste storage requirements imposed by the government (you don't need to store waste for 10,000 years if you reprocess it, like every other nuclear energy using country does) and endless lawsuits delaying plant construction being allowed to proceed even though they were based on no valid technical data whatever.

[dctoedt]

> Newer reactor designs with entirely passive safety features remove the need for operators to take particular actions in the event of a problem, so "loss of brain accidents" are no longer possible.

That would be excellent — although color me doubtful; people always seem to find new and innovative ways to f[oul] up ....

[pdonis]

> people always seem to find new and innovative ways to f[oul] up

That's why the passive safety features I referred to don't depend on people at all; they depend on the laws of physics, which are certainly more reliable than people. :-)

[bluejekyll]

I assume people are still building [those things](edit: these new reactors), and I also bet they require some regular amount of maintenance.

[pdonis]

> I assume people are still building those things

If you mean older reactor designs without the passive safety features I mentioned, I don't think the US is building much of anything.

> and I also bet they require some regular amount of maintenance.

Nuclear reactors of course require regular maintenance, like any large industrial plant, but those costs alone are not a significant impact. To the extent there are increased maintenance costs for nuclear, they are more than offset by lower fuel costs.

Operating costs in general for nuclear reactors are a significantly smaller fraction of total life cycle costs than for fossil fuels. The largest ongoing cost for a nuclear plant is usually the amortization of the initial capital investment, which is larger for nuclear because the plant has extra requirements like secondary containment and radiation shielding that aren't present in fossil fuel plants.

[bluejekyll]

>> I assume people are still building those things

> If you mean older reactor designs without the passive safety features I mentioned, I don't think the US is building much of anything.

No, I mean the new ones you mention, people will make mistakes building the new ones, including the failsafes, regardless of the physics.

> Nuclear reactors of course require regular maintenance, like any large industrial plant, but those costs alone are not a significant impact.

It's not the cost I'm responding to, but more the idea that human error will be entirely removed. People will be responsible for that maintenance, they will screw it up.

[Arnt]

Digressing, perhaps… do you know anything about why the passive designs weren't built? There are something like 500 nuclear reactors, of which approximately zero use entirely passive safety features. Not 250, not even 50.

[pdonis]

> do you know anything about why the passive designs weren't built?

Because they're too recent. By the time these designs were developed, the nuclear industry in the US was basically dead, so there has been no real opportunity for them to be commercially deployed.

[Arnt]

I seem to recall reading about them in the earlyish nineties. Specifically a fine article about a Swedish gravity-based design, which used electric power to keep things running and many, many failures should result in things being pulled down by gravity and failing safely.

[tuatoru]

Strong safeguards are also needed in iron smelters and steel foundries, in ammonia synthesis, and in making aluminum. In the mining of coal and extraction of oil and natural gas.

Yet all these materials are cheap.

[philipkglass]

The LD50 for arsenic in rats is 15 milligrams/kg: http://whs.rocklinusd.org/documents/Science/Lethal_Dose_Tabl...

The LD50 for strontium 90 in hamsters (90 day survival) is 2 millicuries per kilogram: http://www.rrjournal.org/doi/abs/10.2307/3573895

Given strontium 90's specific activity of 142 curies/gram (https://ntrs.nasa.gov/citations/19680020487), in mass terms that's 14 micrograms/kg for the LD50. Gram for gram, strontium 90 is about 1000 times as acutely deadly as arsenic, or 3 times as deadly as the chemical warfare agent sarin. A commercial power reactor of 1000 MWe output can have an inventory of tens of kilograms of strontium 90 in the core [1] along with even more acutely dangerous shorter-lived fission products.

Nuclear reactors are safe because of careful defense-in-depth in their engineering and operation. They need deeper, more stringent safety systems than steel plants or ammonia plants because they contain substances much more toxic than those found in steel plants or ammonia plants. You also see extreme safety practices in facilities that handle non-radioactive poisons, if the poisons are potent enough:

"Inside Fort Botox"

https://www.bloomberg.com/news/features/2017-10-26/inside-fo...

[1] https://www.osti.gov/servlets/purl/6822946 See table 2.2 "Fission product inventories"

[etchasketch]

This is an excellent explanation and it captures why the regulation around nuclear energy is so strict.

[chickenchicken]

Thanks for this insight and explanation.

[lnsru]

Add semiconductor foundries. They have tons of nasty materials there. Enough to make inhabitants of the next bigger city not alive anymore.

[pfdietz]

In all those cases if something goes wrong workers can be sent in afterwards to rebuild things. If something goes seriously wrong in a nuclear power plant the owner is out one nuclear power plant.

[marcosdumay]

> Yet all these materials are cheap.

Yes, in large part because the safeguards are not nearly as strong. (And a lot of the rest is because the safeguards impede research on how to make nuclear cheaper.)

[tuatoru]

Arguable, but I won't. The strength of the safeguards is not the problem.

The surface problem is that the rules keep changing. That's part of the expense: repeated redesign, over and over.

The underlying problem that causes that, is that western societies do not want nuclear power. More: they actively dislike it.

It's all academic now anyway, at least as far as stationary electric power generation is concerned. PV power and storage is far cheaper, and the growth of the gap is accelerating.

[ElijahLynn]

Yeah, I don't know how to convince myself that it will ever be truly safe.

It isn't if it fails, it is when it fails.

Seems highly risky compared to simpler alternatives.

[grecy]

> Former Navy nuclear engineering officer, qualified as [chief] engineer aboard the eight-reactor aircraft carrier USS Enterprise.

I imagine there's tons of stuff you're not allowed to talk about, but have you ever written down any stories in a book or blog or anything like that? I would love to learn more about your career, what you learned, saw, etc. etc. Thanks!

[dctoedt]

> I imagine there's tons of stuff you're not allowed to talk about

Yep.

> but have you ever written down any stories in a book or blog or anything like that?

Nope.

But thanks for asking — I appreciate the interest!

[grecy]

haha, thanks.

Please do know that if you ever decide to write a book, or a blog, or a YT series, or honestly anything at all I'm very sure you'd have a huge audience!

[xaduha]

I wonder what you think of these

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

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

[pfdietz]

Not competitive.