Hacker News new | past | comments | ask | show | jobs | submit login

> Imagine a world where economies of scale bring nuclear power down to the cost of fuel and maintenance.

Economies of scale were applied from the inception of commercial nuclear power, yet the promise of electricity too cheap to meter has never materialized. Quite the opposite. In fact, nuclear has always been the most expensive method of generating electricity, and the anti-nukes don't enter into it. Even if every individual on the planet was pro-nuke, it would still be too expensive. If it was otherwise, nothing whatsoever could prevent investors from coming out of the woodwork to fulfill your dream of nuclear power plants everywhere. Make nuclear energy economical and you can have all the nuclear power plants you want, as well as being absurdly wealthy. But when you fail, try to avoid blaming anything other than nuclear energy itself.




None of this is true in Japan, you might want to read up about the cost scaling that’s happened there.

The short version is, you need to have enough generations of reactor building to allow later projects to benefit from previous learnings. Because of various outside effects (this is a euphemism for the anti-nuke lobby) that type of iterative improvement and workforce skilling didn’t happen in other countries.


The capital costs of nuclear are so high the current model of marketing electricity doesn’t work.

The same will be true of fusion. Set a target level of service and make electric generation a government service. The government is better at capital projects anyway and if energy were a taxpayer funded service it would transform society in many ways.


> The capital costs of nuclear are so high the current model of marketing electricity doesn’t work.

It doesn't work because nuclear is the only power where all the externalities are accounted for. The average coal plant kills more people with radiation released from burning coal every year than all nuclear accidents combined.

But because it's nicely diffuse no one cares.


Coal is dying in the United States for economic reasons, anyway.


Yes, because natural gas is cheaper.

Amazing.

When people said we'd be using renewables I somehow didn't think dinosaur farts would be the largest slice of the pie.


Sorry, what radiation is released from coal? Do you mean gas emissions?



Thanks

The link you provided indicates the level of radiation is negligible, so I'm not sure what your point is.


It's higher (IIRC, orders of magnitude higher) than the amount of radiation released by an operating nuclear plant.


And that number includes all nuclear accidents as well iirc


Can you name any capital projects in the last 30 years where the US government did a great job of a) controlling costs, b) meeting the stated design goal, c) on time?


I’m very happy with the massive road construction and maintenance in my area.


I have a family member who was an engineer in the Federal Highway Administration for many years. The amount of waste and excess cost on these projects is really hard to exagerate.

Comparative costs to other OECD countries for similar infrastructure - there really isn't one, as the US is way out of line in terms of cost and quality.

My family member communicated with counterparts in other countries, impressed with what they were doing. When asked what the research was that they based their projects on, they all said it was the US funded research (which the US doesn't take advantage of).


The roads I’m talking about aren’t federal highways (“I-*”).

I’m sure the interstate highway system is the best the 1950s has to offer.


If you’re not talking about federal roads then I don’t think your comment is relevant to a discussion of US Government projects.


A few that come to mind:

- Alaskan Way Viaduct

- Tappan Zee bridge replacement

- Billions in highway projects.


I might not hold up the new Tappan Zee Bridge as a beacon of meeting design goals.

https://www.timesunion.com/news/article/mario-cuomo-bridge-s...


Maybe you should stop electing people whose campaign talk is that US agencies are a big mess?

I mean, these people basically gain popularity when government is shown to be inefficient, but at the same time, they're the ones tasked with running the government.


[flagged]


Fukishima was still probably cleaner per kWh than most coal plants.


You shouldn't leave it there, because that accident has nothing to do with cost scaling.

Why don't you leave the thinking for the ones better at it.


> You shouldn't leave it there, because that accident has nothing to do with cost scaling.

The accident, that single accident on its own, probably doubled the cost of nuclear power to Japan. It is hard to tell because the costs have not all been paid and are mounting still.

This is the Achilles heal of nuclear power, that the boosters want us to ignore: The risk of catastrophic failure. One reactor can be looked after carefully (not economically - others have pointed out here the eye watering costs) and safely. But lots of reactors? No. We will keep have catastrophic failures until we stop.


> But lots of reactors? No. We will keep have catastrophic failures until we stop.

The military have fielded many reactors for a considerable amount of time. We've got a bit over fifty years of fielding nuclear reactors in submarines. Where are all the inevitable disasters, if that is the case?


> The military have fielded many reactors for a considerable amount of time. > We've got a bit over fifty years of fielding nuclear reactors in submarines. Where are all the inevitable disasters, if that is the case?

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

Three from the USA navy. It is a lie that the nuclear reactors on submarines have been safe. Two have been lost at sea (one was sunk in harbour)

Nuclear reactors on ships generally, submarines in particular, are a terrible idea. As evidenced by the fact that the US Navy has sunk three, two catestrophically. If the US Navy cannot do it, who can?


2 disasters in 50 years is not exactly catostrophic.

We've had over 120 shipping disasters in that same timeframe, resulting in many thousands greater loss of life, and greater environmental disasters.

https://en.m.wikipedia.org/wiki/List_of_maritime_disasters_i...

Should global shipping be abolished?


Two disasters that leave nuclear reactors irrecoverably deep is indeed a catastrophe.

If you care at all for the future.

But the present trumps the future...


The number of oil spills that happened in that same time frame, that we're still trying to recover from, exceeds that by a great number. The environmental damage caused by those isn't purely theoretical.

Compare that, to the Exxon Valdez, and the Deepwater Horizon.


I'd like you to demonstrate what damage that's causing.


Why don't you keep reading the article on the list submarines that you referenced, all the way down to the 4th paragraph that talks about causes of sinking.

1 of 9 (Soviet K-37) was scuttled with no loss of life due to reactor issues spanning years. 1 had an undetermined cause of sinking with theories having to do with the usual submarine issues (fire, torpedo accent, ballast/pressure issues), not its reactor. All others were lost for reasons having nothing to do with nuclear reactors.


Huh? Only one of those subs had issues with its nuclear reactor, but there was no "disaster": it was scuttled in a controlled manner (though unfortunately in a location that the IAEA disagreed with).

All the other incidents listed had nothing to do with their nuclear reactors; they would have all sunk just the same had they been powered by diesel fuel.

I'm not sure what you mean by the US Navy sinking two "catastrophically": none of the deaths listed on the page you link have anything to do with the fact that nuclear power was used.

However, you probably should have linked this section of the overall article on nuclear submarines:

https://en.wikipedia.org/wiki/Nuclear_submarine#Reactor_acci...

It's weird that the section starts off with "Some of the most serious nuclear and radiation accidents by death toll in the world have involved nuclear submarine mishaps", as the following list doesn't seem to support that, only cataloguing 28 deaths and 162 cases that probably included lifelong health problems. (These numbers are not great, but still not as bad as for land-based reactor incidents, and tiny compared to deaths and health problems caused by burning fossil fuels.) They do claim "substantial radioactivity released" in several of the incidents, but don't quantify the effects, or what the effects would have been if it had been a land-based reactor (so maybe very bad!).

I hesitate to mention this (as I don't want to sound dismissive), but it might be notable to consider that all of these incidents occurred on Soviet subs, and that there hasn't been an incident since 1989. The navies of the US, UK, France, China, India, and post-Soviet Russia have never (at least as far as we know) had a reactor incident with their nuclear-powered subs.


What are the issues with sunk nuclear submarines?

I tend to think the bottom of the ocean is a reasonably good place for the nuclear reactors in those subs.


I don’t think you actually believe that the bottom of the ocean is a good place to keep nuclear waste.

You would probably agree with me—and most other people—that a nuclear power plant, the was caught throwing their waste into the ocean, would be an environmental disaster. The US navy doing the same thing is also an environmental disaster.


Presumably the reactors in submarines are commanded to scram when the sub is in trouble.

So the reactor is a high corrosion resistant stainless vessel in non-critical state at the bottom of many tens (hundreds?) of metres of ocean.

All told, that doesn't seem an entirely unreasonable place for it. Even if it was critical, in the active state before the sub sank, it'd probably just heat water for a decade or too before it eventually cools down.

Even if it suffered a meltdown, that's not ideal, but I'm not gonna lose any sleep over it.


Do you really think that?

We are dealing now with barrels of DDT dropped in deep water off the Californian coast


Genuine question: is this a nuclear power failing, or a business failing?

According to the wiki article, several issues from the inception of the plant through the failure were ignored, all of which either made the failure worse or would have arguably prevented it. My understanding is that the plant itself was already supposed to have been shut down, but duty was extended; I'm not sure if this was to move towards break-even on cost, or just to prevent a power generation gap, or make more money. Note that the plant in question is one that had been shut down previously because TEPCO had been caught falsifying inspection and repair records.

IMO (as a total non-expert) nuclear is probably safe, it's the people running it that screw it up for everyone. Maybe that's not a meaningful distinction? I'd sure like to have a discussion about nuclear that doesn't conflate failures of leadership with failures of technology, though.


you better leave it there because any other word would show how bad of a point it is.


> In fact, nuclear has always been the most expensive method of generating electricity

Just wonder if we're pricing in those externalities from coal, like hundreds of thousands of deaths per year due to air pollution and the fact that it is completely fucking up the atmosphere, leading to a global catastrophe. Because...we're not.


Including air pollution externalities, nuclear is definately cheaper.

Some of the costs/risks with nuclear are a bit black swannish though which complicates matters.

Energy efficiency, pollution controls and carbon taxes are the real magical tech we've not exploited.

Because cheaper in the long run means more expensive right now, which is a vulnerable political soft spot that psycopaths can easily exploit.


Great that the alternative is renewables then!


Nuclear is uneconomical because safe, economical nuclear is illegal. That's it. If everyone were pro-nuclear, those laws wouldn't exist (or would exist in a sane form), and nuclear would be cheap.


Nuclear never achieved economies of scale because of the anti nuclear crowd. The cumulative number of nuclear power plants built is a drop in the bucket compared to coal and gas plants.


I'm not sure how the anti-nuclear crowd stopped that. There was a general mismanagement and misestimstion of how much nuclear was needed. And then there was a massive amount of construction mismanagement.

Even in France, which has a very supportive population, and is usually heralded as a nuclear success story, saw increasing costs rather than decreasing costs as they build more of the same reactor model.

Construction productivity doesn't increase like manufacturing productivity does. Nuclear's failures are, as far as I can tell, entirely the responsibility of those tasked with building it. And it's quite possible that nuclear only makes sense at a certain level of economic development, with the right level of technology, but not too much technological development such that manufacturing has completely eclipsed construction.

Go to any investor that would build nuclear, and they won't cite public opposition as the reason not to build, they will cite the construction risk.

And that's why this SMR design is being tried rather than large reactors.


Hail Mary Reactors are being tried because they're nuclear's last chance.


> Even in France, which has a very supportive population, and is usually heralded as a nuclear success story, saw increasing costs rather than decreasing costs as they build more of the same reactor model.

I'd be curious to learn where you got your data for this, because I have never been able to find such data. Best I could get was average cost of a plant in a "pallier".

What I can provide is building duration, which definitely goes down as more of the same model are built, but there's no guarantee that building duration and cost are equal here.


I spent years asking people who claim that France's cost were low for the data behind it to no avail, but only finally stumbled upon somebody linking to this paper in a different context:

https://www.sciencedirect.com/science/article/abs/pii/S03014...

There's another paper somewhere in my bookmarks that shows this for reactors of the same model built in the US, and maybe also France, but I'm having trouble locating it right now...

As with any single publication, it can't be taken as the revealed truth, but it's the best I have at hand.

Edit: how could I have forgotten about the infamous Loveringe, Yip, Nordgard paper! This is impressive because it does the analysis across many countries, and finds a few with cost decreases, but a dominant trend towards more expensive construction:

https://reader.elsevier.com/reader/sd/pii/S0301421516300106?...

My personal, untested, hypothesis is that these variations have mostly to do with the general cost of labor in a country, and in particular their level of economic advancement, which establishes a floor on the value of an hour of a person's labor. Specifically, my guess is that construction is only feasible for nuclear when a country is in a magic sweet spot where labor is cheap, and technological advancement is moderate, like 1950s-70s US level. But not so advanced that labor has gotten so expensive that skilled labor like welders have more productive uses of their time.

If my theory is correct, future costs for South Korea should rise from where they were before stopping their program (in a flurry of corruption scandals, I would note. Nobody brings up South Korea as an example of successful construction anymore, instead having to resort to suggesting Rosatom should embark on a massive building spree across the globe.)


The abstract looks interesting, but I can't find it in open access. If you have access to it, what would be interesting to look at is cost evolution within a batch ("pallier" in french). Cost of newer batches compared to older batches isn't really relevant, as France progressively moved from 900MW reactors to 1450MW ones. Newer batches were also notably smaller, leaving less room for design improvement or cost optimization.


Do you have source for nuclear energy is “too expensive”?


Nuclear needs subsidies due to a long history of large cost overruns and other issues.

https://energypost.eu/how-profitable-is-an-investment-in-nuc...

And France, with many Nuclear plants, has recently had issues keeping the plants up. Many of the most problematic plants are newer ones.

https://www.ans.org/news/article-3939/frances-energy-woes-wo...

The reality is economies of scale have kicked in for solar and wind energy in a way they never can for nuclear. We are at the point where it makes sense, at times, to overprovision renewals to ensure enough supply.

The issue with renewables is storage, of course. But that problem looks to be more solvable than cost effective nuclear, a problem which we have not solved in over 50 years. One can say if we were only smarter we could make nuclear more cost effective, which is probably true. But we built a nuclear power plant where a tsunami occurred in the past, only to find it occurred again, so we aren't that smart. The issue with nuclear is everything has to be right for it to be cost effective and safe, and nuclear is too complex for humans to consistently do this.

Safe nuclear power that is also cost effective is not a problem we have solved.


SMRs are an opportunity for economies of scale to kick in for nuclear. If you can start building reactors in factories, and interate on designs, then you can see economies of scale kick in.*

That said, I am not super bullish on nuclear, and agree with your position. Solar and wind are just getting too cheap for nuclear to get the investment needed to catch-up. If we see someone crack the code on cheap storage of electiciy, solar and wind will runaway with the prize. Not coincidentally, storage is has vastly more investment of money and brains on looking for ways to scale vs nuclear.

*Safety concerns with nuclear will make iterative design difficult to pull off!


Nuclear needs subsidies

So do oil and coal [0]:

> Globally, fossil fuel subsidies were $5.9 trillion in 2020 or about 6.8 percent of GDP, and are expected to rise to 7.4 percent of GDP in 2025

[0]: https://www.imf.org/en/Publications/WP/Issues/2021/09/23/Sti...


Never say never. And about solvable storage: right now it is not solved. So it makes sense to go nuclear (again).


There was scarcely any financial incentive to address the grid scale energy storage problem until like two or three years ago. There are a half dozen startups working on it now.

My vote is to rapidly install renewables while we have the natural gas backstop, and give the storage startups a few years to make it happen. Energy Dome, Hydrostor, and Form Energy all look promising.

And if new nuclear turns out to be less expensive than expected, great.


> There are a half dozen startups working on it now.

Many more than that.


What are even the yardsticks by which nuclear storage success is even measured? /gen


It’s pretty easy to google comparative costs.

It’s really hard to invest billions of capital dollars that won’t produce any return for a decade while you can realize immediate return on solar with costs that drop every year.

Accountants are the worst enemy of nuclear, not activists.


If they were comparable that might be true, but they are not interchangeable.


The most expensive energy is wind and solar after a few weeks of low wind in a cloudy winter.


Fortunately, that very rarely happens, so backing it up with gas turbine plants (burning hydrogen, ultimately) is cheaper than dealing with the 365/24/7 high cost of nuclear.


Nice, i would have guessed that duplicating the production infrastructure and the added loss of hydrogen transformation would be more problematic.

Do you have any actual numbers at hand?

"...is cheaper..." is a bit vague.


I'll point to the simulation site I always link to: https://model.energy/

This uses real weather data for various countries and optimize solar + wind + batteries + hydrogen to provide a constant output at minimum cost. The results are interesting. You can tweak the cost assumptions.


This is an article about nuclear plants starting to be produced at scale.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: