Visiting the most expensive nuclear station | Hacker News Reader

Visiting the most expensive nuclear station

samdumitriu.com

50 pointsdukeyukey51 comments

> To produce as much electricity with solar as Hinkley Point C would use a plot of land almost fifty times bigger than Hyde Park.

Hyde Park is about 1.4 km2, so that would be 70 km2, which even in a dense country like the UK is not that much. It seems like a no-brainer to go for solar instead.

Even China, who nuclear advocates point to as “getting it right” look like they realize this now .. https://johnmenadue.com/chinas-quiet-energy-revolution-the-s...

(need storage too though that's more a cost problem than a space problem)

> Nuclear, of course, is the safest form of power generation there is.

It's odd that one person would make a claim, then link to data that shows it not to be true.

Even odder, I've seen this done repeatedly with the same claim. Is it intentional? Some kind of collective mental blindspot?

It's funny. I suppose Rank Order is not that useful when the top two ranks are very similar. The linked data is Nuclear 0.03 deaths / Solar 0.02 deaths.

But we shouldn't use the rank order claim then haha. In fact, there's not that much utility to it consider 1,2,3 are in one class and then 4 is two orders of magnitude higher, and then 7 and 8 are one more magnitude higher.

Nuclear actually is safe by pure incident metrics, even when counting higher casualty estimates for Chernobyl.

The safety, of course, is not intrinsic: nuclear material is obviously very dangerous, especially in a running reactor.

The safety record stems from the considerable regulation around building and operating these reactors, and the fact that the reactor has so little external surface area once running:

* A very small fuel acquisition operation in comparison to fossil fuels.

* Likewise, no externally released pollution outside of accidents, which is rare.

* sites chosen for construction are picked for their stability, and are heavily engineered, meaning you also don’t have the installation worksite deaths which run up numbers for wind and solar.

But again, this is only realized if the operational safety onsite is maintained. That said, it’s not the only dangerous power generation site: dammed hydroelectric can be a considerable danger depending on what is downstream.

In general, I think nuclear would be very popular if natural gas and solar were not available; however, the costs to keep it safe are too high for it to be economical compared to those two sources.

Generally I consider nuclear to be an incredible potential, hugely capable.

But civilizationally I don't see us as being able to keep doing the right thing with it. If windmills mess up and come crashing down in 20 years, there's negative impact, but locally and short term. If nuclear waste of contaminated decommissioned sites have a particularly bad day in 300 years, it could cause massive long term widescale problems for potentially centuries.

I want this to be something that governments get behind and do, want it to be a high priority that we put beyond the whims of the market. I want better realistic views of pricing in & maintaining the very very very long lived negative externalities. I want intense research in what makes good long term sense, what's sustainable.

Nuclear has so much potential, is so compelling to me. But I do not see a species organized or driven enough to meet with the very long lived complexities and challenges, do not see the appetite to do the job extremely well. We are very safe about it, but ultimately our scope is short cheap reactors, not doing nuclear in a big lasting fashion, at scale to justify figuring out systematically.

Breeder reactors remain this fancy expensive thing we once did, but don't do again. Cancellation of Integral Fast Reactor & failure to make a PRISM derivative is a really sad failure to mature; here we had a much cheaper safer proloferation-safe way to care for the whole nuclear lifecycle, and we never could muster the try, to see how we might do better. What few reactors that are getting built tend towards unambitious fuel-inefficient simple designs, that saddle us with long term problems.

Edit: -2? What nonsense. Say something! I spent the effort to lay out a view & case, have some decency, downvote-squad; contribute back.

I think your down votes are because people are tired of rebutting the same old anti-nuclear arguments.

"Civilizationally" The evidence is nuclear has remained safer than alternatives well over half a century even when we have failed organizationally to do the right things (e.g. Chernobyl, Fukushima). IMO let us move on and use technologies that might prevent civilizational collapse rather than avoid them and make such a thing more likely. (Although it's unlikely under any scenario.)

"Proliferation" as a product of civilian nuclear power has been studied and discussed for its entire history and has been disproven. There's no link. In general having civilian nuclear power allows more oversight by international bodies about what you're doing, whereas regimes pursuing nuclear weapons tend to pursue them in secret and using infrastructure fit for the purpose of producing weapons materials.

"Fuel efficiency" simply isn't important when the fuel is so abundant and so cheap. We can afford to worry about that in future if we ever wind up building enough nuclear power it becomes a problem. If anything this is a good reason to stop freaking out about "nuclear waste" i.e. mildly used and 95% reusable fuel and leave that where it's been sitting perfectly safe for decades, above ground.

If someone had the time they could mine every nuclear thread on Hacker News and pull out all the common tropes and rebut them someplace in a similar vein to Skeptical Science's list for Climate Change (https://skepticalscience.com/argument.php). @acidburnNSA's https://whatisnuclear.com/ might be the closest thing. But then nobody would read it, and the problem would continue.

I remember reading a list of industrial disasters, and Dam breaks always had absurdly high death tolls. I remember reading one from China that had several hundred thousand IIRC.

I couldn't find the exact Wikipedia article, but this one is still a pretty interesting read.

https://www.wikipedia.org/wiki/List_of_accidents_and_disaste...

Edit: I think it was this one: https://en.m.wikipedia.org/wiki/1975_Banqiao_Dam_failure

Though, of course, the exact numbers are contested.

Attributing casualties from a dam break as energy-related is a stretch. The vast majority of dams are built for flood control. Energy production just helps pay the bills. It's not the 'purpose' of the dam.

Article reads as overly dismissive of safety measures as well. Seems to be implying that measures to prevent construction workplace accidents are driving the cost up, for example.

Well it doesn't need to imply it, those measures do drive up the cost without debate. It's fine that they do though as we value people's lives more than that.

The thing linked to doesn't show it not to be true. It gives point estimates of 0.03 deaths per TWh for nuclear and 0.02 deaths per TWh for solar (and higher figures for everything else) and says "the uncertainties around these values mean they are likely to overlap".

So it indeed doesn't show that "nuclear is the safest" is true, and if forced to guess on the basis of the numbers there you'd do best to guess that actually it's second-safest after solar, but it also doesn't show that "nuclear is the safest" is false" and a more accurate description would be "wind, nuclear and solar are all comparable to one another, and all much much much safer than any fossil fuels, there probably isn't that much difference between them, and we don't really have the data to tell what order they should go in".

(Some slightly more concrete numbers: total electricity generation of the US is about 4000 TWh, so those figures suggest an average of about 120 deaths per year if that were all nuclear and about 90 deaths per year if it were all solar. For comparison, according to https://www.statista.com/chart/6024/causes-of-death-in-the-u... these are on the same order of magnitude as "deaths by lawnmower" and "deaths by autoerotic asphyxiation".)

It would probably have been better to say something like "the safest, along with solar and wind power". But everything we know is perfectly compatible with nuclear in fact being the safest, or the second safest, or the third safest.

There’s a lot of dudes on a lot of roofs for solar. A lot of electrical work done, maybe by electricians, probably by mildly upskilled roofers.

I would be highly skeptical that solar is safer than nuclear. There’s just so much poorly regulated low skill low accountability construction work going on keeping it all running.

Maybe the figures are for large-scale power-grid solar, or something?

(I share your suspicion, though. And the figures being used for nuclear seem too high to me; e.g., no one is building nuclear power facilities as dangerous as Chernobyl any more, so far as I know.)

You are right, th link shows that nuclear is the safest right after solar, if you count dead/Terawatt-hour. 0.03 for nuclear and 0.02 for solar.

Solar is still very small (and intermittent) when it comes to production, for instance, less than 2% for US (source: https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/U...). This single new UK nuclear power plant alone would account for 7% of UK production according to the article.

The article basically lists all of the problems with the common law system. I tend to agree that it would be cheaper to build a large civil project under another system of laws but unless you are planning a revolution you have to take the political and administrative conditions into account.

> more than four times more expensive on a pound-for-megawatt basis

oh, the units people use today :)

Pounds sterling, Britain’s currency.

I know, thus the smiley :)

How does the joke go?

"Americans will measure with anything to avoid learning the metric system"

(I know this is about a British power plant but the tongue in cheek remark still works)

American units are actually pegged to metric standards, and quite a few people use SI units here and there.

I don't see cm overtaking inches anytime soon though.

Just the other day, I opened up my fancontrol-gui and saw my CPU temperatures all in the 200⁰F range, I almost vomited in disgust, then thought this is how Europeans must feel. Who uses Fahrenheit for CPU temperature monitoring (legitimate question)?

Fahrenheit is still better for daily weather though.

Chances are the actual thermal mgmt chip is doing everything in Celsius, and your GUI layer is thoughtfully converting it to Fahrenheit to match your locale settings.

You could probably override your local info and get it directly in Celsius if you wanted.

You are correct of course, https://github.com/Maldela/fancontrol-gui/commit/1667dc46651...

However, I don't want my locale to reflect all units this way.

I might try overriding the locale for this single process, however I it was just easier for me to compile a patched version of the code that was available on the Arch User Repository already.

Though I have no idea why the git version defaults to Celsius.

31k page environmental document. My God, why even bother with nuclear when you can slap out coal power plants instead of dealing with that regulation mess.

Coal runs out?

To an unskilled observer, it sure must appear that in most countries, doing literally anything on a scale greater than a moderately-sized industrial hall is bound to be shut down by an unspoken alliance of bureaucrats, tree-huggers, and various public-private partnership agencies that want to get in on the money.

I live in Portishead (mentioned in this article), we frequently see great barges sail by with massive blocks of concrete destined for HPC. Quite a spectacle to behold!

One thing I found out recently found out about nuclear that really shuts down home "great" it is as a fix for climate change is how limited fuel is for it.

Some numbers

Nuclear currently uses about 60,000 tonnes per year of uranium [1] Nuclear is about 10% of electricity, 4% of energy as a whole [2] There is about 8,000,000 tonnes of uranium reserves world wide [3]

For a 100% of current electricity demand by nuclear that's 13 years of fuel,

For 100% of energy (e.g. gas heaters replaced by electric powered by nuclear) that's 5 years of fuel.

Doesn't look so great when you do the math.

[1] https://www.iaea.org/newscenter/news/iaea-symposium-examines...

[2] https://ourworldindata.org/nuclear-energy

[3] https://en.wikipedia.org/wiki/List_of_countries_by_uranium_r...

It’s BS. There are far greater uranium reserves, even with minimal exploration, and most fuel can actually continue being used if desired.

The equipment for recycling used-up fuel rods is the same for creating nuclear weapon payloads. Governments struggle stomaching proliferation risk in the name of fuel efficiency.

If there was an actual shortage, that problem would be resolved,

They have no need to do so, because actual uranium reserves are huge. We haven’t been attempting to find or open new ones for a long time.

That's mostly just fear mongering because mitigating that risk with organization and a little bit of planning is trivial. Just put the reactors and the fuel reprocessing in different sites under control of different organizations. To produce plutonium usable for nuclear weapons you need to pull the fuel out of the reactor way ahead of schedule for power generation. The longer you keep the fuel in the reactor, the more Pu-240 you produce alongside the Pu-239 and you can't make nuclear weapons with that.

Weapons grade plutonium requires very low concentrations of Pu-240 and that requires running the reactor specifically for this. If the reactor and fuel reprocessing are done by different organizations, then neither can make nuclear weapons without the cooperation of the other.

Managing risk with organization structure is a technique well known to virtually all governments, from western democracies to the worst dictatorships, so this isn't breaking any new ground.

Those are proven reserves, not all the uranium in the crust. It’s just the confirmed high value locations that people have bothered to prospect.

At over 3 ppm there’s at least 4 billion tons of uranium in seawater alone.

Fun fact: the estimated cost of extracting uranium from seawater is $1000/kg [1] (the link shows a lower number, but I'm making it round and higher to account for inflation), which is about 7 times higher than the current market price. If we were to use this, it would increase the price of nuclear generated electricity by ¢0.83/kWh. For comparison the average price of electricity in the US in February 2024 (last month published) was ¢ 16.1 / kWh.

Here's the math, for those curious: modern nuclear power plants produce about 50 GWd per ton of uranium fuel. It takes about 10 tons of mined uranium to produce one reactor grade ton (because of the enrichment). So, that's about 5 GWd per ton, or 120000 MWh/T, or 120000 kWh/kg. If one kg of Uranium is $1000, you get 120000 kWh out of that, which comes at $1/120 = ¢0.83 per kWh.

[1] https://inis.iaea.org/collection/NCLCollectionStore/_Public/...

[2] https://www.eia.gov/electricity/monthly/epm_table_grapher.ph...

As everyone mentioned there is no much prospection, and we are only scratching the surface.

And this is only counting with the old nuclear power plant designs. The current Gen3 and coming Gen4 do not need nearly as much as the previous generations. They squeeze more juice out. (source: https://www.amacad.org/sites/default/files/academy/pdfs/nucl...)

We might unfortunately (because of the chaos that it will create) ran out of oil before uranium.

Fossil fuel crunch will eventually happen, and it would not be pretty, unless we electrify most of our economies (source: https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/U...) and replace all the fossil fuel based production... This is the national security issue that every country should put as a priority.

The need to electrify the economy is why nuclear is so bad, it simply can't compete on cost even.

This article is about a power plant that cost £46Billion and nameplate of 3260 MW

Assuming 100% load factor for the plant (looks like 70-80% is more common but I'll be generous) that's 28,557,600MWh per year. Or a cost of £1610 per MWh per year.

Taking just one of the latest wind farm in UK South Kyle Wind Farm, Cost £38Million with a nameplate of 240MW.

Assuming 10% load factor (30% is common but I'll be pessimistic for this case).

That's 210,240MWh per year,(2400.1 24 * 365) that's £180 per MWh per year, (Life span differences of wind(30 year expected) vs nuclear(40-60) could increase the cost of the wind by up to double if you took the worst case but I've already given a 3x disadvantage on load factor) Even with the deck stacked in nuclear's favour it's 10x more expensive than wind, you will simply never migrate a factory using thermal gas with the cost of electricity made by nuclear.

Edit: also Flamanville 3 in france costs are better but still so much worse than wind, 13billion Euro(~11Billion GBP) for 1600MW nameplate, comes out for 713gbp per MWh per year, still 3x worse than wind.

There is some premium to be paid to choose when you get the electricity and when you decide it is a convenient time to shut down for maintenance.

If the alternative is having no lights and no industry operating the extra cost of dispatchable power pales in comparison to the losses due to blackouts.

The need to electrify the economy has more to do with peak fossil fuel or level of CO2 in atmosphere, than nuclear.

Comparing nuclear and wind is a bit apples and oranges. One is intermetent and one is not. The lifespans are different too, as nuclear usually last 60 to 80 years, when wind farm is 30 years. As mentioned in the article the cost of nuclear is artificially inflated, like for redundant safety measures, and specifically in UK by the financing used (most of the cost come from interest).

We need to build cheaper and more nuclear power plants, and more wind farms too.

Reply on news.ycombinator.com