The second sentence on that page: "This page is a grand tour of reactor development programs from 1945 to about 1970, also known as the nuclear heyday." So yeah, no wonder Oklo is not included.
As for bipartisan nuclear support in the US, IIRC Clinton shut down the IFR (one of the most promising advanced reactor + reprocessing concepts) in the early 90'ies, largely due to anti-nuclear sentiments.
For anyone reading this I also suggest acidburn's article on waste[0] and economics[1]. (I always recommend searching for his comments in HN nuclear discussions)
Also, always good to see you and congrats on a post of yours getting to the front page.
I loved this episode. It resonated with all I learned in writing the history page. There's almost no type of reactor that hasn't been tried, and so hoping that some new kind of reactor is the answer to all problems in nuclear is a bit misguided.
> There's almost no type of reactor that hasn't been tried, and so hoping that some new kind of reactor is the answer to all problems in nuclear is a bit misguided.
Reading the history and economics pages, and searching for hopeful paths forward, I thought of the NASA-SpaceX partnership. Where it wasn't that the component technologies had never been tried, or weren't known, but that company and systems optimized for atypical objectives might find an order of magnitude to pick up, and maybe soon two.
OP here - thanks for writing it! Really enjoyed it!! Curious if you wrote anything more specific about future of nuclear power, and if you have any readings to recommend.
Thanks for posting it. It's always a dream to get on here.
The follow-up piece to this one is the Economics page, and this section in particular talks about what's likely necessary to have a good future in nuclear [1] (it's a collection of conclusions from expert groups on the topic, mostly focused on economics).
Your work is a gift to the rest of us: excellent scholarship on a crucial topic. Thank you.
The "economics" essay underplays two important issues and elides a third. First, learning curves. You are reluctant to let go of "economies of scale", for which you seem to place, ahem, a thumb on the scale against "economies of mass production". What you miss is the dynamic nature of learning. If you produce two reactors of size X/2 rather than 1 reactor of size X, you get twice as much learning, and that learning compounds. C.f. wind turbines.
Second, nuclear has many more serious competitors than it used to. Geothermal and various storage technologies produce market-friendly energy at gigawatt scale, and capital is being deployed to improve them as fast as it can be absorbed.
Third (and related): all economies are political economies, and nuclear has the worst politics of any technology. To their enormous credit, nuclear advocates are finally taking this issue seriously, but they have so much work to do.
Unfortunately, these three issues feed into each other. The happy case—the one I'm rooting for—is that a 5-10 megawatt technology gets so much capital behind it that costs are brought down to "vaguely competitive with offshore wind", and that it becomes an important option with gigawatts deployed annually starting around 2030. Unfortunately, by that point we're going to have GWh-storage paired with solar that is either 1/10th that cost or twice as efficient as currently, or both. The path for nuclear is narrow and treacherous.
After completing this, I thought geez I should just make this a book. It's a relatively fringe subject, but hey maybe it'd be a good use of a sabbatical or something.
There is a lot of investor interest in advanced nuclear reactors because of climate change, so it could be useful to make this info more broadly available.
It's a pity we seem to be losing knowledge from time to time. Like Antikythera mechanism, representing technology that was developed and lost for 1,500 years.
The little bits of information - that are currently scattered across hundreds of impenetrable reports - might still be interesting to a regular person, and who knows, inspire someone to pick up the new research.
I have a 100-line Python script that I wrote that runs hourly via cron. It scours my apache access log and counts up the number of times each referrer has shown up in the past hour. If it's above a threshold, it emails me with the referrer names in order. So I just get email that basically say: "Hey you're on hacker news" or "Hey check this reddit thread" or whatever.
What criteria do you use to determine "best"? Safety, construction cost, fuel efficiency, capacity factor?
Pretty much all modern designs are much safer than designs of Fukushima and Chernobyl reactors developed in 60s.
One of the most recent designs (classified as generation 3+) with operational reactors is VVER-1200, operational reactors: Leningrad-2, Novovoronezh-2 and soon Astravets (Belarus). Next iteration VVER-TOI is under construction in Kursk-2 and Akkuyu (Turkey). Both of those designs easily satisfy post-Fukushima safety requirements.
Practically, it's the PWR- Pressurised Water Reactor. I say practically as there are always new developments, like Thorium-Salt reactors that made the headlines a while back, but these are the reactors that are producing the majority of nuclear electricity and haven't had a single major accident. They contain less fissile material in the reactor core, they are purely Boron moderated (so Chernobyl 2 couldn't happen) and there is no risk of contamination as the pressurised water in the reactor core cannot mix with the water turned into steam for the turbine.
They do have a few downsides, like frequent fuel rod swaps and boric acid buildup, but in the real world this is the safest working reactor we have today.
> Practically, it's the PWR- Pressurised Water Reactor. I say practically as there are always new developments, like Thorium-Salt reactors that made the headlines a while back, but these are the reactors that are producing the majority of nuclear electricity and haven't had a single major accident.
Three Mile Island was a PWR. Of course, it wasn't a major accident in the sense that no significant radioactivity escaped the plant, but it was a major economic hit for the operator, and gave an unfortunate boost to various anti-nuclear groups.
I think in practice, BWR's are considered about equally safe as PWR's. Yes, Fukushima was BWR's, but it's uncertain if same vintage PWR's with equivalent safety equipment would have fared any better.
> They contain less fissile material in the reactor core
Compared to what? A fast reactor of equivalent size, sure, but this isn't really the thing holding fast reactors back..
> they are purely Boron moderated
No, they are water moderated. Boron, which is an efficient neutron adsorber, is used to control the reactivity during the burnup. That is, with fresh fuel rods the boron concentration in the water is higher to compensate for the higher reactivity of the fuel.
There is ongoing research to design LWR's without boron, in order to avoid the corrosion issues and the equipment needed for controlling boron concentration. One way is to have rods filled with DU, which are slowly pulled out (and the volume replaced by water) as burnup proceeds. This approach has its own downsides, but, we'll see.
FTA: "Note: This is written largely from the US perspective. Developments in other countries are not well covered here. Also, the chains of events are difficult to classify so the time linearity of the following is not perfect."
Perhaps you could write a similar page about European and Chinese development of nuclear power. Who knows, you might end up on the front page of HN as well!
I would love to add more about this story in other countries as well. I just ran out of time and steam when writing the US perspective. Give me more time to recharge.
One thing that's nice about the US side for me is that we have vast libraries of technical reports available for me to peruse. It would be impossible for me personally to do such a deep dive in Russia or France or China or any other non-English-publishing place simply because I don't speak the language.
The development of electricity generation from nuclear has deep roots in europe: USSR in 1954, Uk in 1957.
AGR design got deployed at Torness. (I protested FWIW, a position I might vary on now with hindsight)
The chinese work on pebble-bed interests me hugely. small modular reactor stuff is fascinating.
The UK and France (aside from Russia) operate their own submarine reactors. I think its odd they didn't move this up the foodchain to surface ships the way the Americans did.
Both the UK and USSR had significant accidents, which informed choices around future reactor design.
The role of Klaus Fuchs in the emergence of the east german nuclear power industry would be worth a write up all of its own.
The extent of French dependency on nuclear power is amazing. And the German retreat from nuclear power, likewise the scandi experience. China and India, I think will be making a huge move off coal but its less clear this will be to nuclear. Japan is obviously in a complicated situation, perhaps not unlike France's dependencies.
According to this chart [0] the installed capacity of fossil fuel plants is 4TW. Lets cut it down to 2TW because we don't know how much of that infrastructure is actually utilized. Let's say I'm sold on nuclear and want 2TW of installed nuclear plant capacity. Each power plant should be able to generate 700MW. This means we have to build 2857 new power plants. Half of them need to be done by 2050. The other half should be in construction by 2050 but are allowed to take 10-15 years to be built. Surely economies of scale should be able to reduce nuclear plant costs to around 2 billion USD per plant [1] but that would still cost 5714 billion USD.
Open questions:
Which nuclear technology is available today that can be deployed 1400 times today as soon as today without any massive safety flaws? Theoretical reactor designs may sound cool but nobody is going to wait for them.
Who is going to pay for the prototypes and first dozen plants? How are you going to convince the rest of the planet to install your nuclear plants?
Is this ambitious timeline even possible? Can you train enough people and build an extensively optimized nuclear supply chain that can scale thousands of reactors?
What happens once the "building" rush is over? Will the supply chain die again and be hard to start back up?
Can every country be trusted to operate their own nuclear plants? Will they openly report flaws and safety problems in the power plant and take those reports seriously by turning off defective or unsafe plants until they restored the safety of the plant? Fukushima was easily preventable but there was no accountability and no incentive to prevent an accident. Nobody in charge felt responsible to actually add adequate safety measures.
Can we mine enough fuel to power these plants? Will mining activity have to increase to meet the new demand? Can that mining be powered by nuclear energy? Is mining for nuclear fuel worse than coal?
Is waste disposal feasible today in every country? Can this cost be reduced through economies of scale or is it fixed per unit of waste?
None of these things are impossible but they require time, money and trust.
People who don't care about climate change are not going to give you money for nuclear or renewables.
Time isn't exactly on our side. We are at a point where we have figured out lots of alternatives to fossil fuels, it is just a matter of sticking with one choice and actually doing it. Hesitation will cost a lot of time so it's now or never.
We are also short on trust. The UK doesn't like the EU. Other continents are not unified at all and it is common for countries to be at war in the middle east and Africa. Why would they want a power plant that is a juicy target for terrorism?
Because of these concerns I'm not going to hope for a nuclear future but I'll let myself be surprised if things turn out better than I predict.
Of course it is. Pick a design that can be built right now, which means the Chinese or Russians have to do it, because everyone else (Siemens, Areva, Westinghouse, ...) seem to have forgotten how to build theirs. One reactor produces 600-1400MW, so we're talking roughly 1500-3000 reactors. You set a timeline of about 40 years, or 15000 days. This means, construction of a reactor has to start every 5-10 days. Construction of a nuclear plant takes about 4 years or 1500 days, unless treehuggers are allowed to interfere with frivolous lawsuits. This means, 150-300 reactors will be under construction concurrently. Worldwide, that doesn't sound too bad.
Let's build one reactor per week.
> What happens once the "building" rush is over?
More building to replace aging plants. At the end of your 40 year timeline, the plants built at its beginning are (nominally) at the end of their life. (Not really, but then again, 40 years into the future, 2TW won't be enough.)
Let's build one reactor per week, forever.
> that would still cost 5714 billion USD
So what? Those 2TW of generating capacity are still going to be built, if only to replace broken down old plants. Capital cost for power plants varies a bit by technology, but 2000G$ are going to be spent on power plants no matter what; if the future is solar, it's going to be a lot more than 6000G$.
> Can we mine enough fuel to power these plants? Will mining activity have to increase to meet the new demand? Can that mining be powered by nuclear energy? Is mining for nuclear fuel worse than coal?
Yes, yes, mostly, no. If we actually embark on a massive nuclear program like this, we will run out of fuel fairly quickly (about 100 years). One generation of once-through burner reactors can be fueled easily, maybe two. After that, the Uranium may need to come from low grade sources, which is annoying, or breeders have to be used.
> Is waste disposal feasible today in every country? Can this cost be reduced?
Yes, and why bother? Waste disposal is more a political problem than a technical one. It isn't actually expensive. Arguably, the "waste" isn't waste and should be recycled instead of being disposed of.
> Why would they want a power plant that is a juicy target for terrorism?
Bullshit.
This is called "drive by slander". Why do you do this when you (otherwise) appear to like nuclear power as a solution to climate change?
1. for the first time since the 1970's, support for nuclear power is non-partisan in the US - both parties have endorsed it.
2. a good addition to the list would be mentioning a YC company, Oklo (https://www.businesswire.com/news/home/20200615005643/en/Okl...)