How to get new nuclear built faster
samdumitriu.com
samdumitriu.com
How does that benefit the groups funding major candidates in the US? I don't know about the other countries but you can be confident that the amount of red tape involved in reactor construction in the US is not a bug but a feature to many of those responsible for it.
Nuclear power is generally stable and predictable which reduces potential peak prices, but it doesn't scale down well so having a lot of it during times of low load means off peak prices will fall. Expansion of nuclear power thus threatens the profit margins of a lot of powerful organizations.
---
See also: Cars. It is not technically possible to produce a single light passenger vehicle that can be legally registered without modification anywhere in the world.
The US's Federal Motor Vehicle Safety Standards and the UNECE Global Technical Regulations that apply in much of the rest of the world are incompatible in a few ways that mostly relate to lighting/reflectors and crash structures. As a result, every "world car" offered in all major markets in fact has at least two and usually three variants, typically UNECE LHD, UNECE RHD, and FMVSS LHD.
This wouldn't be particularly hard to solve by either updating one or the other to be compatible or by the US accepting UNECE LHD vehicles, but the existing automakers that sell around the world have no interest in making it easier for others to compete.
- On one important data transfer, from European module to American, somebody forgot convert units, and software decided to make full halt (to be honest, these things are extremely frequent, but mostly fixed after appear), and this time, unfortunately, this lead to mission loss.
But to be honest, civilian aviation is full of such issues, when appear some critical things, because people erroneous read measuring unit names.
For example, Gimli Glider case, when technicians erroneously read fuel amount in kilograms, but enter something other (their interface accepted centimeters, liters and kilograms and convert automatically), and as consequence, plane was carrying only 45% of it required fuel load. Fortunately, all alive and plane got minimal damage, but this is not always.
As usual, horrible ROI and massive cost overruns:
> During the construction of Vogtle's first two units, capital investment required jumped from an estimated $660 million to $8.87 billion.[7] ($18.3 billion in 2022 dollars[8])
> Two additional units utilizing Westinghouse AP1000 reactors were under construction since 2009, with Unit 3 being completed in July 2023.[9][10][11] This last report blames the latest increase in costs on the contractor not completing work as scheduled. Another complicating factor in the construction process is the bankruptcy of Westinghouse in 2017.[12] In 2018 costs were estimated to be about $25 billion.[13] By 2021 they were estimated to be over $28.5 billion.[14] In 2023 costs had increased to $34 billion, with work still to be completed on Vogtle 4.[15]
So why bother? Because base load? I mean damn, why not just toss in a couple Moss Landing type tesla Megapack installations instead?
At 2.2GW, even if u wanted to store power for 12h, thats 26GWh. Moss Landing is 2.8GWh and it cost $560 million. Ten of these is 5B, so an entire solar+megapack installation still ends up at half the cost of nuclear.
And you don’t need to store 12h of solar. And the rapid response of Megapacks is incredibly helpful for grid stability. And they can be spread out making the grid more resilient to natural disasters, attacks etc
No you don't need 12h of storage but 3 to 4 months to sustain the attrocious winter production where the highest load actually is.
And this is why the battery packs are uneconomical to smooth out renewables, it doesn't even get you 1% of the way with 5B.
Still cheaper, and getting cheaper. Don’t ignore wind power too.
> Still cheaper, and getting cheaper. Don’t ignore wind power too.
Wind is even worse than solar in this aspect, it can be down unpredictability and unlike solar, is closed to the maximum theoretical efficiency already anyways which means the only thing you can do is count on the economy of scale.
And before people tell me that it usually produces more in winter, yes it does, until it doesn't like the 3 weeks wind anomaly we had two years ago in the EU with close to no production.
I'm not predicting future tech of course, it's possible that "just build more" will work in the future, for now it doesn't though and from a large margin.
1. *Solar Insolation*: The average solar energy received per day in winter. 2. *Day Length*: Winter days are shorter, so there's less daylight for solar panels to operate. 3. *Panel Efficiency*: The efficiency of the solar panels in converting sunlight to electricity. 4. *Weather Conditions*: Overcast days are common in winter, which can reduce solar panel output.
Let's make some assumptions for a rough estimate:
- *Solar Insolation*: In Germany, the winter solar insolation is roughly 1.5 kWh/m²/day. - *Day Length*: We'll assume an average of 8 hours of effective daylight in winter. - *Panel Efficiency*: Modern solar panels are about 15-20% efficient. Let's use 18%. - *Weather Conditions*: We'll assume that overcast conditions reduce effective insolation by 25%.
Now, let's calculate:
1. The daily energy output (in kWh) = 1 GW * 8 hours * 18% efficiency * 75% weather factor. 2. Multiply this by the number of days in a month (30 days for simplicity).
A 1 GW solar power plant in Germany could produce approximately 32.4 million kWh (or 32.4 GWh) of energy during a typical 30-day winter month, considering the factors and assumptions outlined earlier.
Hit me up in 2040 when Poland’s first nuclear power plant finally opens a decade behind schedule snd costing the usual ~10x of the estimate.
>A 1 GW solar power plant in Germany could produce approximately 32.4 million kWh (or 32.4 GWh) of energy during a typical 30-day winter month, considering the factors and assumptions outlined earlier.
Okay, so this means on average 1GW of solar power should output ~1GWh of energy daily, right? So Germany's roughly 80GW of power would produce on average 80GWh of electricity? Well, let's look at _real_ data: https://app.electricitymaps.com/zone/DE
In the last 30 days there was exactly one day with over 80GWh of solar power produced: 9th January. And it was just 81GWh. Yet there was multiple days where actual electricity produced barely went over 10GWh... or not even 10GWh, like 21 December with 7GWh produced. Three days ago, it was 14,5GWh.
Overall solar power production in Germany in December was 795GWh, below 1/3 of your estimate.
Plenty of hydro too. Enough to push electricity prices below zero in may 2023.
Luckily youve also got a choice of nuclear reactors built by Russia or reactors where you get the privilege to fight with France over whose responsibility it is to pay for the billions of euros in cost overruns.
There will always be a local regulator, and the push to adopt common requirements (even enforced by the pertinent UN agency, the IAEA) is limited by the need to double-check (keeping many distinct eyes on the subject in order), its own tropisms and also will/need to justify itself by avoiding to simply copy-paste.
All this will probably prove to be quite an ordeal for any "let's build thanks to components sent by a factory" approach (SMRs...).
> One major driver of Hinkley Point C’s costs is that EDF were effectively forced to build a first-of-its-kind reactor design by the Office for Nuclear Regulation. To win regulatory approval in the UK, EDF had to make roughly 7,000 design changes from the reactor design they use in France. Not only do the design changes mean that 25% more concrete and 35% more steel is needed during construction, but also that many lessons EDF have learnt the hard way in France can’t be easily applied in the UK.
This is absurd, and explains a lot.
In any case, a lot has been learned from Flamanville (where there were poor welds that introduced years of delays) and Olkiluoto. The next batch of EPRs, probably of the EPR2 design which implements those hard and costly lessons, should be drastically quicker and cheaper to build. We'll see how Penly progresses.
https://www.nationalgrid.com/national-grid-ventures/intercon...
That's quite a difference between two first-world regulatory agencies.
https://www.edf.fr/groupe-edf/produire-une-energie-respectue...
That's the thing though. Even absurdly expensive as they currently are, the damage they don't cause through the emissions they will save over their lifetime is probably worth it. And considering all the lessons EDF have learned on EPRs, it's highly probable that the next generation will be cheaper and faster to build.
Assuming you could site a new nuclear plant somewhat closer to an interconnection point, it seems to me like even modern, over budget, uneconomical nuclear is not all that much more expensive than massive scale renewables. And the capacity factor can’t be beat.
If the west could build even a handful of additional plants after Flamanville, Vogtle and the handful of others, is there any reason to think that the next tranche of plants is just hopelessly uneconomical with renewables?
To me it seems like even if nuclear is twice as expensive than renewables plus transmission, that’s a reasonable price to pay for the base load generation capabilities and diversification of the overall energy mix.
But wait, there’s more! horrible ROI and massive cost overruns:
> During the construction of Vogtle's first two units, capital investment required jumped from an estimated $660 million to $8.87 billion.[7] ($18.3 billion in 2022 dollars[8])
> Two additional units utilizing Westinghouse AP1000 reactors were under construction since 2009, with Unit 3 being completed in July 2023.[9][10][11] This last report blames the latest increase in costs on the contractor not completing work as scheduled. Another complicating factor in the construction process is the bankruptcy of Westinghouse in 2017.[12] In 2018 costs were estimated to be about $25 billion.[13] By 2021 they were estimated to be over $28.5 billion.[14] In 2023 costs had increased to $34 billion, with work still to be completed on Vogtle 4.[15]
So why bother? Because base load? I mean damn, why not just toss in a couple Moss Landing type tesla megapack installations instead?
At 2.2GW, even if u wanted to store power for 12h, thats 26GWh. Moss Landing is 2.8GWh and it cost $560 million.
1. The whole point is Nuclear does not have to cost this much, thus your comparisons are irrelevant. China’s build costs are substantially lower so it can be done.
2. For a not-insignificant percentage of the global population that can consider mega projects of this scale neither wind nor solar are reliable enough during peak energy season - the winter - to use even massive battery packs like you have described.
And don't forget that a nuclear reactor can easily push 70+ years with some maintenance and refreshes, meanwhile solar panels and batteries are drastically less effective after ~15-20 years at most.
Who is going around investing in the overall energy mix, though? People with $10 billion to invest aren't thinking "how can I optimize the grid", they're thinking "how can I get the best return on my capital"; renewables offer substantially lower risk and quicker timelines than nuclear, so that's where the money seems likely to go. If we as a society want someone to build nuclear plants in the near term, for grid-stability reasons, we're probably going to have to subsidize the project.
There are lots of ways those costs can be allocated - from user fees in deregulated markets or from ratepayers in vertically integrated systems, and probably a million permutations in between - but those costs are there and someone’s gonna pay them.
A long-life highly available nuke plant at 3x the cost of nameplate renewable production doesn’t seem like much more than an immaterial cost difference at the end of the day.
You do the same as every other market, you make those companies pay for the externalities.
So they don't only pay for renewables but renewables + gas plants (and their environmental costs) + batteries + whatever else is necessary in a real grid.
Which is why we've built pumped hydro and implemented Time of Use pricing for many decades to help nuclear energy integrate well with the grid.
Where did you get this idea from?
I do think it was an environmental and diplomatic (as seen now in the Russian war) mistake.
[1] https://en.wikipedia.org/wiki/Hualong_One
[2] https://en.wikipedia.org/wiki/Karachi_Nuclear_Power_Complex
This is because different nature of issues.
In Pharma, main issue testing, which is basically very human-intensive task, at the moment machines are not involved. And mostly talking on drugs, we trying hard to achieve "invisible" drugs, which have extremely narrow appearance (as in space as in time), so ideally, drug make it's action extremely short time, after which no effects appear.
So, ideal drug, does not have any side effects at all.
In Tech, difference, it almost whole side effects.
1. Building starts with choose good place - using lot of sources to gather info, and sources see this and react.
2. Continued with negotiations, with interested parties (persons and entities), and on so large and sophisticated project, need to make some parties allies.
3. Building itself, involve lot of parties. Some parties making profits, other demand compensations.
4. All buildings have limited productive lifetime, after which appear decommission and in ideal case, annihilation of building and return land to some "first use" state, now most people consider wild nature state.
Imagine world, where you must negotiate with ALL people on your street usage of anti-flu drugs, including all possible side effects and all following effects, like attending of gym, etc :)
Moreover in South Korea some nifty lil'tricks may alleviate some of maintenance-related burden: https://en.wikipedia.org/wiki/South_Korean_nuclear_scandal
See FAA/Boeing.