Nuclear: Not a Faustian bargain, a near-perfect providential gift (2008) [pdf]
gordianknotbook.com
gordianknotbook.com
All complex technologies are really best understood as socio-technical systems. They depend on people and machines, data and economics, and, in the case of nuclear, especially fickle things like geopolitics and warfare. It's all about unknown unknowns, and not the engineering kind.
If it weren't for the unknown unknowns we are going to face in the 21st century from climate change alone (and not just the first order effects, but the second and third order effects on economics and geopolitics, which are basically unknowable), nuclear might make sense. But we have cheaper (solar PV + storage is cheaper, soon others will be too) and safer alternatives.
The easy test is this: if a country has a currency crisis, a coup, two back to back 1000-year storms, another pandemic, or any number of other major events, and the plant gets damaged and nobody shows up to work for a month, is everything going to be fine? Will someone break in and steal something dangerous? Will something dangerous leak out? With modern renewables the cost is low, the tail risk is basically zero, and the power is compatible with hard times.
It is not, not at the scale needed to reliably replace fossil fuels. This just isn't true.
In the Nordics, there is no PV production for five months of the year. Two months have marginal production rates.
Zero times any large number is still zero.
Nuclear does not enjoy the benefit of scalability because it is complex, highly regulated, and the risks associated with failure are large.
Chernobyl was an awful incident no doubt, but the USCEAR report shows that it killed about 4000 people over the full course of time. Meanwhile, the US reliance on coal currently kills 25,000 people per year. The next-worst nuclear accident of all time, Fukushima, killed either 0 or 1 person. The next, Three Mile Island killed exactly zero.
We've surely lost millions of lives globally over the last 50 years thanks to sticking to coal when we could have just moved huge quantities of power generation over to nuclear.
In terms of cost, the reality is that a renewable + storage grid is still more expensive that nuclear power is today, let alone has been for the last 50 years.
[1] https://ourworldindata.org/grapher/death-rates-from-energy-p...
[2] https://www.unscear.org/unscear/en/areas-of-work/chernobyl.h...
[3] https://www.forbes.com/sites/jamesconca/2016/07/01/uranium-s...
This isn't true.
Leveled, unsubsidized cost ($/MWh) of:
Utility scale solar + storage: $46 - $102
Onshore wind + storage: $42 - $114
Nuclear: $141 - $221
Source: https://www.lazard.com/research-insights/2023-levelized-cost... (have to download the linked PDF: https://www.lazard.com/media/ruwg1jol/lazards-lcoeplus-april...)
This is not to say solar and wind + storage are less economical than nuclear. But LCOE is not a good measure of real costs.
The over provisioning could be considered more of a feature than a bug, offering new economic periods where the marginal energy cost of production go negative at times.
??? Like 4 days of downtime per year?
I've had about 1.5 hours of outage this year, which was an outlier (much more outage than normal).
3 9's is typical if you have above-ground utilities. If you have underground utilities, you may be doing far better than this...
Over-provisioning has a cost. If it is required to achieve characteristics required to do a like-for-like comparison, it should be accounted for.
The problem with these Nuclear conversations is that people are pushing their preferred option, always without coming to some sort of agreement on what the requirements are for the power generation option. Without agreeing on what needs to be achieved, everyone talks past each other, arguing with straw-man solutions.
Long distance transmission (1000km+) is surprisingly affordable and widely deployed in eg China: https://en.wikipedia.org/wiki/Ultra-high-voltage_electricity...
But note that even with significant over provisioning solar and wind is still cheaper than the cheapest nuclear costs if you over-provision in the low cost, high generation areas (another argument for transmission)
People from both "sides" always make claims that make their preferred solution look more favorable. The non-nuclear people point at measures like LCOE which are misleading. I'm sure pro-nuclear people are doing the same thing.
Personally, I do not have a side, I just want real numbers for replacing coal/gas and getting a similar or better level of grid stability.
> But note that even with significant over provisioning solar and wind is still cheaper than the cheapest nuclear costs if you over-provision in the low cost, high generation areas (another argument for transmission)
Can you provide some sort of evidence for this claim? How much does a UHV line cost per KM? How much will it take to build out? How much storage is required? How much over-provisioning is required?
You can over provision wind (without storage) 5x or 3x with storage and still be cheaper than nuclear.
> How much does a UHV line cost per KM? How much will it take to build out?
Electrical transmission is historically $41.50/MWh/1000 miles in the US and build costs are $1502/mile.
So transmission from Kansas (high wind zone) to California would be around $1.7B.\ to build out.
Given the average cost of a nuclear power station is at last $19B[2] it's pretty easy to see how economically far ahead renewables are.
> How much storage is required? How much over-provisioning is required?
I can't find any useful modelling to on this which is pretty astonishing really.
The closest I found was [3] which is German specific, where they find around 8 potential days of every 10 years are below the low wind power events for power from on land wind farms within Germany (so no offshore and no transmission) using 40 years of data.
The distribution of these events is more problematic: there was a 4 day lull in 1985, or 10 days with a broader definition of "lull".
[1] Table 1 and Table 5 https://www.sciencedirect.com/science/article/pii/S258900422...
[2] https://thebulletin.org/2019/06/why-nuclear-power-plants-cos...
[3] https://iopscience.iop.org/article/10.1088/1748-9326/ab91e9
https://iopscience.iop.org/article/10.1088/1748-9326/ab70bd
https://www.sciencedirect.com/science/article/abs/pii/S03605...
It’s just not feasible currently.
The further north you go, the less solar power you have. In the Nordics, you get no solar output for months.
You also have windstill days in the winter.
No amount of overbuilding is going to cure that.
Fortunately wind power and solar power are often seasonally reversed in many parts of the world (ie, lots of wind in winter, more solar in summer), which helps a lot too. See for example wind levels in the North Sea[1] where there is the most wind in January and December, and least in July. Perfect for Nordics!
(And of course with pumped storage it is perfectly feasible to power for days or weeks off storage anyway)
[1] https://weatherspark.com/y/25440/Average-Weather-in-North-Se... (look for "Wind")
It would be, let's say, less than ideal if Europe were dependent on North Africa for their energy security.
In practical terms this would mean that large countries would feel the need to establish military control over critical energy production areas.
Energy security is no laughing matter in the Nordics. In the winter, without power, people die.
Even if the political and operational risks are acceptable, building long distance transmission is no simple matter. Everybody knows that more transmission lines are needed between Norway, Sweden and Finland, but regardless no real progress has been made. Heck, Sweden can't even manage to connect their north and south regions!
The vagaries of weather and the requirement for redundancy would require significant over-investment in generation capacity and long distance transmission networks.
Looking at averages in weather and wind is also very myopic. Power generation is instantaneous, so you have to look at sustainable base load. But the question is, can wind sustain any kind of base load since it is so variable?
On Saturday wind production was down to 53 MW in Finland, less than 2% of peak production. And that wasn't even the lowest this month. What are you going to do, build 100x more wind power? And even if you do, you are only producing less than 20% of total power required to power the country, so it's more like you need 500x more wind power.
Since you cannot do base load with wind power, nuclear is off the table, what are you going to do? Hydro power?
You cannot run off hydro power for days or weeks since the generation capacity just isn't there. For the Nordics, Denmark has no hydro, Finland only has about 20% of generation capacity and Sweden has about 45%. Norway is the only country that can run off hydro and Iceland is a very far away.
This means you are back to peaker plants. Say bye-bye to zero carbon emissions.
So all this brings us back to the questions, are renewables really a realistic and cost effective option compared to nuclear?
Stability of output is a very important (and valuable) characteristic of a power source, and it is completely ignored in the LCOE calculation.
One major thing going for nuclear is that is has a stable and consistent output. It's great that solar and wind beat all options on a $$ when you ignore time effects. But if stability is important (and many argue it is), then solar and wind rely on masses of storage, overprovisioning, and extensive upgrades to the transmission network, then let's bring those into the discussion and do a fair comparison.
Oh. A private financial think tank that grossly underestimate the cost of storage when solar/wind investment actually can yield 10% ROI in some country.
I am so surprised :rolling-eyes:
But if you choose not to believe them there are numerous other sources that give the similar numbers.
https://www.irena.org/Energy-Transition/Technology/Energy-st...
https://www.energycouncil.com.au/analysis/big-battery-bonanz... (2021)
Which of course, like common sense would say (irony), would stay perfectly flat or decreasing when it already struggle to match the supply for EV which is two order of magnitude inferior to what would be needed.
Which make me back to my point: Yes, this is oriented garbage done by people that has financial interest in what they are selling.
There's likely to be a lithium oversupply: https://www.cnbc.com/2023/03/07/bank-of-america-sees-lithium...
There's no particular shortage of Lithium in the ground - it's just matching projects to demand. I'm not familiar with vanadium.
> oriented garbage done by people that has financial interest in what they are selling
They sell consulting services.
Lazard has mostly ignored those costs when it tries to come up with a LCOE for intermittent renewables. A real world grid operator can't ignore those costs. To take a real world example, Germany requires dramatic subsidies to build out its renewable energy portfolio. These subsidies occur at every level of electricity grid. There is no simple way to untangle the costs and come up with a single number that can be compared to non-intermittent sources of energy. The share of biomass in the German grid is strongly suggestive of the fact that renewable energy isn't cost competitive without significant subsidies.
There is no free lunch, decarbonizing energy generation requires more expensive sources of energy.
Nuclear replaces the first step, heating the water. The other steps are the same.
Pumped hydro does better than batteries, but building new dams causes massive ecological damage, and in some parts of the world fresh water is a rather scarce resource. The heated water from a nuclear power plant arguably does less environmental damage overall than a big new dam does.
2. How do you think other thermal powerplants (gas, coal, ...) work? They, too, need to heat water and dispose of it afterwards, so that's definitely not nuclear-specific.
But maybe these thousands of French river fish come from a very rare species or something.
I mean, I’m legitimately asking. Was it a big ecological problem? Thousands of fish doesn’t sound like a ton of fish.
Don't dismiss Chernoble as if it's over. Let's not also forget Fukushima. And hey, how about 3 mile island?
Nuclear seems to have potential to be a phenomenal gift, but the problem is giving it to humans.
The basic upshot is that nuclear power was not seen as economically viable if operators had to buy insurance to cover actual risk. The federal government (meaning taxpayers) pay everything above the $15 billion mark. Fukushima's cleanup cost is still unknown, but estimates went from $13 billion to $96 billion to $187 billion. If nuclear power is as safe as you say, then they shouldn't need these kinds of subsidies.
Going further, between American business culture and American regulatory culture, I have real concerns about our ability to safely and economically operate nuclear plants. Is a nuclear plant really compatible with a reward system that prioritizes short-term cash extraction? Where it isn't, can we really count on regulators to keep things safe without drastically inflating costs? Looking at the recent spate of bank failures leaves me with a lot of questions there.
Insuring long tail risk is really, really hard to do and is likely to be very overpriced in the market.
Regardless, I think nuclear power should pay its own way. I don't want to be on the hook for hundreds of billions in cleanup cost because some CEO got greedy.
In general, we do not require industries to insure for limitless, pie-in-the-sky risk. Sometimes the state is left holding the bag. Requiring infinite insurance would price every industry out of existence.
And the notion that Fukushima is just some sort of "pie-in-the-sky" scenario is bonkers. It is a thing that actually happened. An actual nuclear plant caused actual hundreds of billions of dollars in costs. If the best nuclear power advocates can come up with is, "Well that shit just happens, and we're sure not going to pay for it," then I think the state is better of spending its money on solar panels, wind, and other things where the downside risk is not sized like a major hurricane.
What you're talking about -- requiring insurance for outsized losses -- is not something we require of most industries. A chemical plant (say, making feedstocks for solar panels) can contaminate a water table and cause tens of billions of dollars in losses, but we attempt to address this risk with front-end regulation rather than requiring insurance for uninsurable losses.
Nuclear power was given a special subsidy because nobody wanted to run the risk of running a nuclear power plant if they actually had to be on the hook for it.
I never said we should require insurance for nuclear power, although that's not a bad idea. Plenty of states do it for cars, after all, so you could make a case that it should apply to bigger things. I am just saying that nuclear power should be priced fairly, which would include removing the special liability cap for nuclear plants.
Then there's Centralia, PA, which was turned into a virtual ghost town by a coal mine fire. https://en.m.wikipedia.org/wiki/Centralia_mine_fire
And we're not even considering the effects of global warming here. Sure, nuclear has risks, but our civilization needs energy and nuclear has proven to be one of the least risky ways to produce it.
It's complex because it's hard to tell if any of us are making apples-to-apples comparisons in these discussions (broadly). And...
I hope we figure it out before our current energy generation makes an entire planet uninhabitable.
And that is with less than 400 plants globally. If we 10x these then we should expect far more frequent accidents.
And let’s not get started on the topic of the nuclear waste.
https://www.statista.com/statistics/273002/the-biggest-nucle...
It's pretty ironic how this article talks about how new designs are unsafe and yet commentators here are saying how much safer new designs are:
> The Europeans at the WNA Symposium were talking proudly of how their latest reactor design, with its core catcher and superior leak-tight containment, is safer than the current plants’ designs. This is wrong in concept. Adding provisions to solve a nonproblem merely provides additional potential paths to failure.
There are still a few reactors of the same design as Chernobyl, generating power today.
These reactors have not exploded. Simply because they follow strict procedure.
Fukushima was a consequence of continuous cost cutting by management. It was fully preventable, and construction of improvements even started! They were cancelled due to "cost".
Onagawa Nuclear Powerplant survived the same earthquake and tsunami, despite being closer to the epicenter.
An analogy: All planes are dangerous because a human can push a red button, or cut costs and crash the airplane.
Nuclear Power has the same issue stemming from human idiocy and incompetence. This can be fixed, if only politics and MBAs got out of the way.
But does anyone foresee a modern America with a regulatory or financial environment that will allow for controversial projects that take decades to pay off? I don’t.
Onward to solar and wind. You can install it on your own house, and it doesn’t make the nightly news if somebody falls off their roof and dies while installing solar.
[1] https://www.nationalobserver.com/2022/12/22/news/japan-goes-...
I don’t know the answer but probably more than 150 I’d wager.
How much land is going to be rendered sub-arable over the next hundred years from climate change?
Let that sink in, after a nuclear disaster (and two MAJOR natural disasters), the most devastating impact of living in the area was having to live in underfunded temporary housing and the stigma of being viewed as "contaminated".
This is something every society needs a solution for. It's not just about nuclear, natural disasters happen all the time and the best system we have in place is "leave". We can do so much better. This is another instance where improving temporary public housing isn't "just for homeless people", it's for everyone.
> The year after the 2011 disaster, the Japanese government estimated that 573 people had died indirectly as a result of the physical and mental stress of evacuation.17 Since then, more rigorous assessments of increased mortality have been done, and this figure was revised to 2,313 deaths in September 2020.
Source: https://ourworldindata.org/what-was-the-death-toll-from-cher...
Don't get me wrong, I love nuclear power, but reporting numbers inaccurately/partially serves nobody.
EDIT: I am not sure what justifies the 3 instant downvotes, I am literally reporting accurate numbers in the face of someone minimizing them.
> No one died directly from the disaster. However, 40 to 50 people were injured as a result of physical injury from the blast, or radiation burns.
Since then one worker has died of lung cancer, and that is likely attributable. So zero or one death for the second worst civilian nuclear power incident in history.
It's very likely that area would have had to be evacuated one way or another, no? I think it's far more fair to attribute those 2313 deaths to the tsunami and earthquake.
If it was a coal plant, I don't think anyone would have been attributing 'stress deaths' to coal. Ditto any other kind of industrial or chemical plant that may or may not have been on that land.
The issue seems to have been a failure to properly safely evacuate people from an industrial zone.
It still is an exceptional incident and it is a ridiculously small death toll compared to the "normal" operation of coal plants yearly death toll... but it isn't a reason to minimize the numbers. As for the 1 direct death, I was referring to this part of my source:
> In 2018, the Japanese government reported that one worker has since died from lung cancer as a result of radiation exposure from the event.
Going by the official statement, I let them do the judging, they don't really have an interest in over-reporting as far as I know.
Do you have data for that? Or as compared to some other chemical or industrial plant?
Again, I think this just has to do with a failure to have an adequate evacuation plan, not the specific land use.
I don't think I have ever heard of such large zones for other industrial disasters, the estimatated maps I've seen for Bhopal, one of the if not the worst industrial chemical disaster ever, had a radius for the gas spread of about 7km (and it killed a lot more people really quickly). It is of course annecdotal, but each disaster has its own circumstances: wind, water/soil contamination, etc. that impose different measures to take... but in general nuclear disasters have a much wider radius for the fallout, that's why the security/processes for it are way more demanding and generally successfully applied.
But the main point was that we can't discount these risks and their results (deaths) when a disaster happened, as rare as they are (thankfully). They are part of it, and counting these deaths is just honest, as I said other sources of energy in the course of their "normal" operation kill way more people and destroy our environment in the process. It's a no-brainer to choose nuclear power over them.
Meanwhile, coal kills tens of thousands every year from pollution, windmills kill engineers and maintenance workers, and shoddy solar installs and battery banks set fire to buildings.
Nuclear is far from perfect, but Fukushima is a terrible example of the dangers of nuclear power.
And it seems every single technology proposed by nuclear advocates is rapidly deprecated. Notably replacement designs are almost all completely experimental.
Even this old, mismanaged plant had zero attributable deaths after what was basically the worst case scenario.
You should also look up Onagawa Nuclear Power Plant. It was closer to the epicenter than Fukushima No1, and was hit with a larger tsunami.
It survived because the plant director fought management, enforced strict safety regimens and performed emergency drills every 3 months.
Nuclear Power works, as evidenced by nuclear power plants and nuclear powered vessels across the planet. They are not exploding at random.
The problem is exclusively human.
For profit corporations cannot be trusted with nuclear power.
Lowest common denominators in governments cannot be trusted with nuclear power. (Japan, a prime example. I know because I live here.)
So who do we trust? A robust system of private and public organizations.
Will never happen here. All the nuclear authority directors have been idiots, purchased, or claimed the position through party politics.
That is the fundamental problem with Nuclear. If not managed properly it has the potential to be very unsafe.
Our world is too unstable and growing more so. I have doubts about our long term ability to safely manage reactors in a capitalist society.
As for the second Fukushima. Look at France. They are IMO one of the best examples of well regulated and managed nuclear power.
You will find mentions of maintenance recently, but this I because COVID delayed scheduled maintenance. (No one wanted the older specialized engineers catching COVID)
I guess the physics of power loss make it uneconomical?
Leveraging an alternative model about eight years ago, this plasma generates its own magnetic fields, avoiding the multi-billion magnet infrastructure required by Tokomak. This plasma stays "lit" almost indefinitely with less than 280 watts of power input.
The video is raw video footage, not CGI.
Climate change has already done much more damage to Europe than Chernobyl.
Fukushima No1 was a carbon copy of a US nuclear power plant that was intended to withstand earthquakes... And tornadoes.
Tsunamis were not factored into the design.
TEPCO wanted to minimize cost. So instead of redesigning the reactor, they just used the same blueprints and built a giant sea wall to block tsunamis.
They then refused to upgrade the sea wall. This is despite multiple earthquake forecasts by public and private authorities.
They even started a sea wall upgrade project 1 year before the earthquake.
It was cancelled due to "excessive costs".