Some knowledge we gain is negative. We've learned nuclear wasn't as promising as had been claimed.
More generally, in engineering most technologies must fail. That's because there are multiple technologies that can solve any given problem, and only one can come out on top. The number of niches < the number of different solutions.
If you think just wait, we'll learn more and nuclear wins, why doesn't that same argument apply to the technologies nuclear is competing against?
You've got it backwards. We've built nuclear plants. We've encountered unknown difficulties and overcome them. Our cost estimates for nuclear power are historical estimates. We did built nuclear at scale before, and it did cost less.
By comparison, the cost estimates for building things like massive amounts of energy storage are cost projections. These costs represent what we think these projects will cost. But these projections, by definition, don't account for unforseen engineering challenges. Historical costs do.
Comparing the two is apples to oranges. Historical costs include the cost of engineering challenges, cost projections for new technologies do not.
Yes, we've built nuclear power plants. And we've discovered they are complex and expensive, and we've discovered they take forever to design, and that very often the construction process fails and the investment is lost.
Energy storage technologies have the property that they can rapidly iterate new designs, on a smaller scale than nuclear. They are inside nuclear's decision loop. The mammals are outbreeding the dinosaurs (yes, I know that metaphor is unfair to the actual animals).
There's a reason business people don't swallow the kind of BS you sling on nuclear, and it isn't because they're Greenpeace members.
But the point is, we have built nuclear at scale. And built it at scale. France has over 70% of its electricity coming from nuclear, Belgium over 50%. The US is at ~10%. We've encountered the emergent issues, and fixed them. Cost estimates on nuclear are historical estimates. They're not projections.
We haven't built energy storage at anywhere close to the required levels of scale. We have 4 minutes worth of energy storage. Most of that is from hydro electric storage. 1% and 3% are from battery and thermal storage, respectively [1].
What will be the cost of scaling this up to the hours or days required to make intermittent sources viable? We can only offer cost projections: what we think massive grid-scale projects will cost. But cost projections do not include the costs of unknown engineering challenges. Remember, early nuclear proponents claimed it would be an order of magnitude cheaper than fossil fuels and had the studies to prove it. But real-world experience demonstrated otherwise. What makes us think that increasing our amount of energy storage by 2 or 3 orders of magnitude won't encounter similar unforeseen costs?
We have built nuclear at scale, and yes we have discovered engineering challenges that increase cost. But we haven't built energy storage at scale, and we haven't encountered it's unforeseen engineering challenges.
This post offers an excellent analogy on this topic [2]. With nuclear, we've walked this path, and know firsthand the difficulties it takes. With new technologies we're looking at the map. Those are two fundamentally different measurements of cost.
And the response to that is that (1) that golden age of nuclear was far less successful than you depict, and (2) attempts to replicate it have gone disastrously wrong.
Cost overruns and failures have been the norm for nuclear. In France, we don't actually know the cost of that initial roll out. In the US, nuclear construction routinely overran budgets, often by large factors. When PURPA subjected US power markets to more market discipline nuclear construction evaporated. This should tell you that it was never a low cost source.
Today, even the construction that occurred then is not possible. The nuclear supply chain is atrophied, especially without nuclear weapons programs to provide support. Attempts to repeat the past glories have gone even more disastrously wrong than they did then.
The best people know not to go into nuclear now, so the expectation of its failure is self-reinforcing. Nuclear programs at universities are closing for lack of interest. Hell, even steam turbines (!) risk becoming a marginalized (and perhaps, eventually, forgotten) technology.
> What will be the cost of scaling this up to the hours or days required to make intermittent sources viable?
It will almost certainly be much cheaper than building nuclear reactors, and by a very large margin. But if it's not, no big deal. We could build reactors later if by some miracle it turns out they're needed. I think the best you can expect now is a dribble of seed money to keep the nuclear industry from disappearing entirely, just to keep some level of expertise around as an insurance policy. But if and when renewables prove themselves to a point even you cannot deny, even that could dry up.
Cost overruns have been the norm, but nuclear power generation at scale during the 1970s was still at most half as expensive as the Votgle and Flamanville projects that you like to cherry-pick.
The atrophy of the nuclear supply chain is precisely why nuclear plants are cheaper when built at scale. When you have a steady demand for reactor pressure vessels, high pressure heat exchangers, and other specialized machinery for nuclear plants its cheaper on a per-unit basis.
Also, I'm not sure why you think weapons programs have anything to do with nuclear power generation's supply chain. The supply chain issues pertain to the construction of pressure vessels, pumps, and pipes that need to withstand the more difficult environment inside a nuclear reactor. The only intersection that nuclear weapons manufacturing has with nuclear power is uranium enrichment. But the cost of fuel is negligible in nuclear power.
> It will almost certainly be much cheaper than building nuclear reactors, and by a very large margin.
Right, just like how early nuclear advocates said it would almost certainly be cheaper than fossil fuels. We don't know this. This isn't a historical cost estimate, it's cost speculation. And cost speculation fundamentally cannot account for unknown challenges.
> But if it's not, no big deal. We could build reactors later if by some miracle it turns out they're needed.
Trillions of dollars wasted is "no big deal"? Not to mention, exacerbating global warming because we delayed nuclearization of our energy sector pursuing energy storage? These are pretty big deals. Decarbonization of the energy sector is going to be one of the biggest endeavors that countries conduct. It's no time to bet on unproven technologies. We know nuclear is expensive. But we know that it works, and we have experience building it at scale. We're walking in the footsteps of what we've done before, not bushwhacking through unknown engineering challenges.
> Votgle and Flamanville projects that you like to cherry-pick.
Yes, I have cherry picked 100% of the current era new nuclear construction projects in France and the US. How selective of me! /s
I suppose I should also have included all the projects that were abandoned before ground was even broken?
> Also, I'm not sure why you think weapons programs have anything to do with nuclear power generation's supply chain. The supply chain issues pertain to the construction of pressure vessels, pumps, and pipes that need to withstand the more difficult environment inside a nuclear reactor.
It includes such things as instrumentation, enrichment, fuel fabrication, and (most importantly) manpower and expertise.
> Trillions of dollars wasted is "no big deal"?
Investments are made on expected payoff. The expectation that renewables will do well is high. The off chance that they do not, oh well! Nuclear could be a backup in that case. There is no need to pump trillions into nuclear now to restore your putative golden age.
(I'd actually put carbon capture over nuclear as the backup plan. Nuclear would be the backup to the backup.)
> Yes, I have cherry picked 100% of the current era new nuclear construction projects in France and the US. How selective of me! /s
Yup, precisely. You've conveniently ignored successful projects elsewhere, like the Hanul nuclear power plant and the Taishan nuclear plant. Also, you're incorrect that Votgle is the only nuclear project in the US, Turkey Point is installing two more AP1000 reactors.
> It includes such things as instrumentation, enrichment, fuel fabrication, and (most importantly) manpower and expertise.
Enrichment is not a substantial portion of nuclear power's cost. Almost all of it is in construction and fabrication of the plant. I'm also not sure why you think building a nuclear warhead translates into running a nuclear plant. A nuclear weapons designer wants a very violent runaway nuclear reaction. This is what a nuclear plant designer wants to avoid.
Nuclear power and nuclear weapons don't use the same fuel. That's not quite true with recent plants: a few of the newer nuclear plants can mix in some weapons-grade plutonium - but the mix is still mostly uranium. This more about dismantling plutonium, as it's the only way we know how to get ride of weapons-grade fissile material - well, besides detonating the nuclear weapon.
> Investments are made on expected payoff. The expectation that renewables will do well is high. The off chance that they do not, oh well! Nuclear could be a backup in that case. There is no need to pump trillions into nuclear now to restore your putative golden age.
You continue to miss the point. With renewables and storage, we're talking in terms of hypothetical costs. What we think it will cost. Projected costs are cheap, because by definition they exclude the costs of unforeseen issues. Historical costs do include the costs of unforeseen issues that emerged in implementation.
Nuclear costs are historical costs. We know what nuclear costs, because we've built it before. And we've built it at scale.
The cost of proposed storage technologies are hypothetical costs. Nobody has built a 1TWh storage facility, let alone a 10 TWh storage facility. We can make educated guesses at what this will cost, but they're still guesses. We'll only know the real cost after it's been built. And projected costs are almost always optimistic.
Comparing historical costs to projected costs is comparing apples to oranges.
> (I'd actually put carbon capture over nuclear as the backup plan. Nuclear would be the backup to the backup.)
You seem to be particularly determined to avoid the one solution that is actually proven to work.
> Nuclear costs are historical costs. We know what nuclear costs, because we've built it before. And we've built it at scale.
Do we really know the real costs? How many countries outside of Canada has designed and built reactor cores without a nuclear weapons programme? How do you untangle the huge amount of money poured into nuclear weapons programmes from nuclear energy? Do you account for education/training of nuclear engineers and scientists? I know what you'll say - that we must account for it with the alternatives too. But those educations are more general purpose, so it's not directly comparable. You don't need to educate lots of "wind power scientists" to continue to scale wind power, and their education wouldn't be wasted if it turned out that wind power reached a dead end.
The costs of future nuclear power is absolutely hypothetical costs. You simply can't just magically reproduce the exact same conditions as the mid/late 1900s. At the very least, I'm sure that most countries will demand reactors that are fundamentally more safe than the reactors built during the golden area of nuclear power. So the reactor technology will not be the same. You might call it irrational, and I'd be inclined to agree. Most reactors of that era was incredibly safe. But you can't wish away human irrationality. Maybe you can educate people, but not without massive resources poured into education and public relations. Add that to the indirect costs.
We don't have strong government countries in the west anymore, governments that can just push through huge multi-decade bets like that. And I don't see any chance of that changing. So the most likely path for a nuclear renaissance is through small modular reactors, since you might find applications with shorter return on investments, and then gradually scale it up. But that's a a fundamentally different model from the golden area of nuclear power, so historical lessons about costs do not apply.
Germany, Japan, Sweden, Korea, Ukraine, Belgium, Spain, Finland, and Argentina form a non-exhaustive list of countries that have built reactors without weapons. The reactor designs are often licensed, but reactors are constructed on site.
I'm not sure why you seem to think that nuclear weapons are connected with nuclear power. These are very different technologies. You don't need control rods, pressure vessels, and heat exchangers to build nuclear weapons. You don't need precise blasting plugs and explosive lenses to build nuclear power plants. The only shared technology is nuclear enrichment, but enrichment infrastructure already exists and countries can buy enriched fuel from other countries.
Taishan in China. China is basically the single remaining bright spot in nuclear, anywhere in the world. But to treat it as such, you have to believe the financial information coming from a communist government. I am not as trusting as you are on that.
> Also, you're incorrect that Votgle is the only nuclear project in the US, Turkey Point is installing two more AP1000 reactors.
"Nuclear units Florida Power & Light was to build at Turkey Point have been delayed indefinitely as it’s too expensive to compete with natural gas, FPL says."
> You continue to miss the point. With renewables and storage, we're talking in terms of hypothetical costs.
Yes, all future costs are uncertain. So any plans involve risk, and all investments are gambles. The market has judged that the risk favors renewables and storage, not nuclear.
> Nuclear costs are historical costs. We know what nuclear costs, because we've built it before. And we've built it at scale.
They are costs from another time. Unless you have a time machine, those costs are just of historical interest. What matters today is what it costs to build today. And even back then, the costs became unacceptable. There's a reason the first nuclear expansion stopped, and it's not regulation.
> You seem to be particularly determined to avoid the one solution that is actually proven to work.
You have linked to a solar + storage + transmission solution, not needing any nuclear. Are you sure that's what you were wanting to link to? I'm fine with that solution. Note that the "three weeks of storage" there would be better done with hydrogen than with batteries.
And we also have examples of France's Messmer plan. And the US nuclear boom during the 1970s. But I'm sure you're prepared to hand-wave away these examples, too.
Countries have built nuclear power beyond 80% of total energy consumption in the case of France. Over 50% in the case of Belgium, 40% in the case of Sweden, and 22% in the case of Korea. Nobody matches matches France with renewable energy. The closest is Denmark with 42% wind power generation. Germany is the only large country with substantial renewable generation (other than hydroelectric or geothermal power), generating 25% of its electricity from wind with a further 9% from solar, but that still adds up to less than 40%. You insist that the market favors renewables, but the share of electricity generation demonstrates otherwise.
Those costs are indeed of historical interest. In particular, when we analyze the cost history we see a clear pattern that serial construction of nuclear plants is cheaper than one-off construction. Production chains for nuclear equipment can leverage economies of scale. We have historical precedence for this relationship between number of plants built and individual plant cost.
You're right: nuclear power wasn't killed by regulation. It was never cost competitive with fossil fuels, and interest in climate change wasn't very strong back then. And when the pace of building nuclear plants diminished, individual plant costs increased.
> You have linked to a solar + storage + transmission solution, not needing any nuclear. Are you sure that's what you were wanting to link to? I'm fine with that solution. Note that the "three weeks of storage" there would be better done with hydrogen than with batteries.
And how much hydrogen electrical storage do we currently have? Again, this solution involves technologies that we have no experience working with at scale. I'm sure these cost estimates are very competitive - just like how analysis predicted nuclear would be cheaper than fossil fuels before we had experience building nuclear plants at scale. These cost estimates exclude the cost of overcoming emergent engineering challenges.
Which is a more reliable prediction?
* Predicting cost of a project based on what it cost to build the projects of the same scale in the past.
* Predicting the cost of a project based purely on educated guesses, or based on small-scale experiments.
More generally, in engineering most technologies must fail. That's because there are multiple technologies that can solve any given problem, and only one can come out on top. The number of niches < the number of different solutions.
If you think just wait, we'll learn more and nuclear wins, why doesn't that same argument apply to the technologies nuclear is competing against?