Hacker News Reader

Comment by throwaway2037 | Hacker News Reader

throwaway2037

> Nuclear. Start-up of the fourth reactor (1.1 GW) at Georgia’s Vogtle nuclear power plant

I checked the Wiki page: https://en.wikipedia.org/wiki/Vogtle_Electric_Generating_Pla...

> 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]

Those numbers are simply mind blowing. Why are these still built? How much solar/wind/battery capacity can 34B USD buy!?

I feel the same about the two big nuke projects in the UK (Hinkley Point C & Sizewell C). It makes no sense compared to solar/wind/battery.

View on HN

> How much solar/wind/battery capacity can 34B USD buy!?

According to How Big Things Get Done, at least 30B worth. Nuclear plants have some of the worst cost and time overruns for large projects, wind and solar some of the lowest (beaten only by roads).

And his theory as to why this happens, is that wind, solar, and roads are self-similar. Building the first 10% directly informs you on how to build the next 10%. Everything you learn along the way can be rolled back into the project in a feedback loop.

What you learn shingling a roof or building a nuclear reactor doesn't help this project all that much. It helps the next. And how many nuclear reactors does one worker build in their lifetime? 2? 4? That's not a lot of opportunity for process improvement.

This is why if nuclear is to have a future it would be via Small Modular Reactors. But even that appears to be too expensive.

Indeed. I mean, I'm not sure the smallness is a key factor, but the modularity for sure and maybe the size limitation comes with that.

I'm not much of a fan of nuclear (pun not intended?) and would like to see solar + wind (+ other renewables) cover most of the load. But if enough resources were pooled to design an SMR, with extensive safety studies, lifetime planning, involvement of multiple parties (e.g. scientific and engineering teams from different world states) - I don't think I could really oppose, especially given the global warming situation.

Of course - there's "ideal nuclear" and there's real existing nuclear, which is quite different and often very problematic.

Matt Ferrell posted a video about vertical bifacial solar (https://www.youtube.com/watch?v=LqizLQDi9BM) last week that has me hopeful for more demand-aligned solar output.

They absorb more indirect light, but the main advantage is that hot panels drop in efficiency, and standing them up improves heat dissipation substantially.

One of the bits that stood out to me is that it has an output curve that partially lines up with The Duck, because you get a period of low-angle exposure near sunset. I suspect you get quite a drop at high noon, but that's fine as long as not all solar panels in the grid are vertical.

I've been anti-nuclear for a long time, and SMR is the least unpalatable choice.

Not without caveats though. Humans are shit at statistics. Their sense of the aggregate danger of putting 100 'safe' items in close proximity often underestimates the odds by half, and in some cases by an order of magnitude.

You might get lucky and lose two personal hard drives in your entire life, but the IT guy managing 100 drives is dealing with failed drives 'all the time'. Backblaze reports on something like 10,000 drives and they are claiming something on the order of 200-300 failures. And I think that's per quarter, not per year.

What do you do with an array of reactors where one right in the middle did not fail cleanly?

> And how many nuclear reactors does one worker build in their lifetime? 2? 4? That's not a lot of opportunity for process improvement.

France built 58 over 20 years, that's how you make it economical. India says they want to do one a year IIRC, which is probably good enough too.

French nuclear power is still not particularly economical and they didn’t build as many plants as originally planned because of it, nor do they plan to build a sufficient number of new plants to sustain current capacity over the next decades. France justifies the expense with their nuclear bomb program that benefits from a strong civilian nuclear industry.

French nuclear power has a levelized cost of about $70/MWh - which is about the same cost as the levelized cost of solar or wind in France, without most the headaches of intermittency. (Remember the average French person lives further north than the average Canadian...)

Exactly, and don't forget that nuclear costs in France are being driven high by wind (and solar, but less so), because it cannot run at its optimal power because it has to reduce produced power when there's wind.

Downvote all you want, when you artificially decrease capacity factor from the average 90% to 65% because you have to accommodate for wind, in an industry when fixed costs are the vast majority of costs, you mathematically raise the cost per MWh…

> “Remember the average French person lives further north than the average Canadian...”

That doesn’t mean much. Regardless of latitude, the French climate is much milder than most of Canada.

In fact there’s plenty of places in North America that get much, much colder than equivalent latitudes in Europe.

Heck, Chicago is the same latitude as Barcelona(!!), but there can be 20°C or more difference in temperature during winter.

What are you saying? That Earth's rotation axis is, what, tilted? Outragious! Herasy!

Wait, Chicago is colder than Barcelona because of the Earth's tilt?!

I don't know!!!?!!!? You think I just had a streak of genius?

No. Since the earth spins they both get the same solar energy.

The way ocean currents work help Barcelona more than Chicago. Chicago is warmer than its western suburbs because the suburbs are farther from the lake. (there are many other confounding factors so we cannot measure this)

That's true, because of the gulf stream, but the only thing that matters from a solar panel point of view is the angle to the sun. In fact you might get slightly more out of a Canadian solar panel than a French one at the same latitude because the Canadian one will be colder!

Indeed. But in either case, solar doesn't help too much on cold winter days at higher latitudes. Especially if we want to decarbonise heating, leading to increased grid demand on the coldest days.

> "French nuclear power has a levelized cost of about $70/MWh"

That may be true for the existing nuclear fleet, but the cost of new-build nuclear is significantly higher. Flamanville 3, which is expected to finally come online in mid-2024 after 17 years of construction, was estimated at €125/MWh in 2022.

This is in large part because they are not building a bit over 2 reactors per year anymore.

France did not experience a significant fall in construction costs when they were building lots of reactors in the 1970s. There's not much hope of getting that €125/MWh below €100/MWh, and honestly that €125/MWh is likely a sever underestimate to start with.

Don't know about France but TVO in Finland is way under that. Their earlier reactors were built in the 70s.

They have been hitting 15 to 20€/MWh for decades now. Or at least this is the price they have been selling to their owners without going bankrupt for decades now. Target with OL3 in the mix is 40€/MWh so OL3 alone is probably in the 50€/MWh to 60€/MWh range. (and they got a really good deal with OL3)

https://www.tvo.fi/material/sites/tvo/pdft/rm89bf734/TVO_-_C...

Paid off reactors can get by on O&M only costs. The trouble is getting through that initial period of paying off the loan and interest. Also, older reactors were far simpler and reliable to construct. (Though "simpler" does not mean simple, they are still extremely complex beasts of machines.)

Finland was smart enough, and France dumb enough, to sign a fixed-cost contract for OL3. This ended up in complete disaster for the French companies doing the building, with the French government buying up the failed builder.

So you can thank French taxpayers for a €70-€100/MWh subsidy for the new clean energy in Finland.

What you're saying is just true of EPR, but not revelant at all in comparison to what happened when the nuclear plants where mass produced. As the parent said, nuclear has been very cheap for 3 decades.

One big difference in the costs is that now days, new reactors are required to fund their own eventual decommissioning and long-term waste handling costs from operating income.

That wasn’t always the case! In the UK, taxpayers have been left with enormous liabilities for managing and cleaning up old nuclear sites:

https://www.theguardian.com/environment/2022/sep/23/uk-nucle...

Finland deserves credit for actually building a long-term waste storage repository, which helps solve one of the biggest ongoing issues/costs with nuclear decommissioning.

The operators have been paying into a fund since the 80s which is meant for decommissioning and spent fuel storage.

Basically it is a legal requirement (some small fraction of a cent for every kWh produiced)

This is really the only sane way to do it.

Also we got lucky with the ground under Olkiluoto being good spot for nuclear waste storage so not much NIMBY stuff for that as it is already the site for the biggest nuclear power plant in the country. It is also small town so a huge % of the population there work at the plant or its sub contractors.

> which is about the same cost as the levelized cost of solar or wind in France

As of 2024. Remember that solar and wind are still on a downward price trajectory and the reduced investment in Nuclear has meant that the prices to up.

So if you have a figure for example of $70/MWh for PV and wind in mind from a couple of years ago, then this figure might be slightly outdated now and more slightly outdated tomorrow. PV might go under $20/MWh at the end of the decade for instance. Nobody will be able to compete with that.

> Nobody will be able to compete with that

It depends on what it will cost to level out times of no sun.

>Nobody will be able to compete with that.

Coal does. According to this IEA table, India and China achieved under $10/MWH

https://www.iea.org/data-and-statistics/data-tools/levelised...

The coal numbers food those countries are $70-100/MWh. Its only the capital cost that get to $10/MWh.

> PV might go under $20/MWh at the end of the decade

It's never clear to me when these numbers are cited, clearly you must be talking about PV + battery right? Otherwise it would not be a fair comparison.

> PV might go under $20/MWh at the end of the decade for instance. Nobody will be able to compete with that.

Except places where the sun doesn't shine much long periods of time (far north/south).

> Remember the average French person lives further north than the average Canadian

The opposite is true, I think.

Did you try to find data? Or just reporting on your intuition about this thing that someone else thought might be surprising and turned out to be surprising?

It seems like half of Canada's populations lives south of 45.5°N. [1]

Someone else calculated the mean population latitude of France to be just over 47°N [2]

Mean and median aren't the same, but in this case, the measures are far enough apart that it does seem like the average (however reasonably defined) French resident lives north of the average (same defined) Canadian resident.

[1] https://skeptics.stackexchange.com/questions/50122/do-50-of-...

[2] https://freakonometrics.hypotheses.org/1125

I wonder how economical it is if you factor externalities.

https://phys.org/news/2021-03-trillions-hidden-energy-extern...

What's interesting is that they mentioned the externalities for geothermal are very very low.

Wikipedia writes: "As of early September 2022, 32 of France's 56 nuclear reactors were shut down due to maintenance or technical problems" and "In 2022, Europe's driest summer in 500 years had serious consequences for power plant cooling systems, as the drought reduced the amount of river water available for cooling."

What has this to do with building? Summer gets you cheap PV battery electric energy anyhow. Lack of cooling towers etc. gives you trouble....

French reactors suffered a lack of suffciently cool water, what little water they had was too hot. Cooling towers numbers actually are sufficient.

There is still plenty of water, the problem is that you have to limit heating the river. With a cooling tower can cool with or without evaporating water depending on the design but always without heating the river. But it is cheaper to directly dump the heat into the water.

And you cannot empty the river neither, especially during a drought, can you?

My hunch is evaporative cooling uses not very much water. You could also fill a small dam when the water flow is higher. All means higher cost, of course.

It generally helps to no use hunches when it comes to hard technical questions.

Ok, the hard technical answer is that a phase change from liquid to gas eats a lot more energy than just heating the liquid however many degrees. :-)

My back of the envelope calculations say it's about 10 times less water to evaporate.

There are detailed diagrams and formulas to calculate that. Engineering, not back of the envelope highschool physics.

But guess what, people designing an building power plants know this. And whatever is built is the best compromise possible at the time. Backnof envelope calculus in 2023 or not.

Dry cooling tower exists but they are more expensive. Natural cooling with the evaporation eventually runs into the problem due to built-up of salts which forces you to replace non-evaporated cooling water with fresh river water and lets you dump the heated water into the river (which is limiting factor due to environmental concerns). The less water is flowing and the higher the temperature of the river are the sooner this point is reached

Related and interesting thing - China appears to be building a trial reactor in the Gobi desert [0] because they think they have air cooling sorted out.

Or at least close enough. I haven't checked if it is technically in the desert.

[0] https://en.wikipedia.org/wiki/TMSR-LF1

Even after the summer heat the reactor availability in France had trouble inching above 2/3 for a number of reasons. One of them being age of the installations and related unplanned outages.

> "As of early September 2022, 32 of France's 56 nuclear reactors were shut down"

This is cherry picking. Look at the 10 years, or even lifetime, average availability. It's 90 or 95%. The reason for this bad number is because of delayed maintenance due to COVID.

> "serious consequences for power plant cooling systems, as the drought reduced the amount of river water available for cooling."

The reduction of power output of French nuclear was something like 0.30%. You read that right. So I would call "serious consequences" a blatant lie.

> "Look at the 10 years, or even lifetime, average availability. It's 90 or 95%."

Average load factor for nuclear reactors globally hovers around 80% according to IAEA data [1]. In France the average is actually lower than this due to load following: by design, many French reactors don't always operate at full capacity because there isn't enough demand at off-peak times.

Few, if any, reactors reach 95%: planned outages for refuelling and maintenance takes up more than 5% of their time.

[1] https://pris.iaea.org/PRIS/WorldStatistics/WorldTrendinAvera...

What proportion of wind turbines were off-line for maintenance over a similar period?

According to this article, the staff for the regulator in Georgia (the only entity in the decision making process looking out for the customers, ratepayers) were banned from even considering the cost of renewables:

> Plant Vogtle proceeded with no cost-cap, no consumer protections, and Public Service Commission (PSC) staff were prohibited from conducting analysis comparing the costs of nuclear to clean energy alternatives.

https://www.powermag.com/blog/plant-vogtle-not-a-star-but-a-...

At the other pair of reactors started at the same time and with the same design in Soith Carolina, executives are in jail for lying about the project. The SC reactors were also abandoned.

Nuclear in the US is so expensive that even starting a new reactor requires considerable corruption.

>Nuclear in the US is so expensive that even starting a new reactor requires considerable corruption.

Best line I've come across in a while. I'm gonna use it. Is it yours?

Just seeing your comment 10 days after it was posted, but I have many fond memories of your work, and it would be an honor for Asa Dotzler to use one of my lines.

> And how many nuclear reactors does one worker build in their lifetime? 2? 4? That's not a lot of opportunity for process improvement.

And reactors are inherently too big and complicated. A wind farm, solar farm, or a road is intrinsically much simpler.

And reactors will suffer from decision makers sitting around a table and changing designs for the next one because engineers have thought up new failure modes of the last design and nobody wants to be the person who says "no" and then has a failure. Along with the natural tendency of US management to increase the spending of their departments in order to increase the size of their own personal kingdom. You'll never manage to stamp out reactor after reactor all of the same design.

What specifically is so expensive about the reactors? Having played the simulator[0] and now being an expert in running a nuclear power plant, the idea is "simple". Seemingly not much more complex than any other steam/heat power plant, so how do the costs run so high? Are the tolerances on everything turned to 11 to minimize chance of errors? Triple-checking all work?

As a point of comparison, the Large Hadron collider had a $9 billion budget, and that required a 27 kilometer circle to be dug.

[0] https://news.ycombinator.com/item?id=38515435

I worked for a Campbell, CA nuclear engineering partial employee-owned co-op formed by mostly ex-General Electric engineers (Stanford and Berkeley alums) ultimately acquired by Curtiss-Wright Corp. (And yes, computers were named after Simpsons' characters.)

Costs are related to the assurance of safety through careful documentation and processes beginning with regulation and insurance underwriting of design engineering, construction, operations, and maintenance. This sort of thing only makes sense to do so at large scale to make it economical. Scale also has its costs. Safety isn't something that can be Boeing'ed because the NRC can and will shut the whole thing down if it were to be found unsafe.

> "I feel the same about the two big nuke projects in the UK (Hinkley Point C & Sizewell C). It makes no sense compared to solar/wind/battery."

Renewables are far cheaper, and can be built much more quickly. But there is an argument for (some) nuclear around energy security and diversification of supply. Batteries are great but can't really provide long-term storage (ie: multiple days/weeks). Do we want to rely solely on natural gas as a backup during the cold, calm, winter weather patterns which in some years can persist for weeks? Last year's energy crisis suggests maybe not. Or can we build enough interconnections and rely on imports from our neighbours?

Also worth considering that Hinkley Point C & Sizewell C are really only replacing existing nuclear plants that are shutting down. Even if both are built, the UK will still have only around half the nuclear capacity in the 2030s that we had in the 1990s.

> Those numbers are simply mind blowing. Why are these still built?

I'm guessing because there was no step in the process where it seemed like a better idea to stop construction and have spent whatever costs so far to get nothing of value, rather than accept the cost increase and end up with high availability, somewhat dispatchable generation.

There's not a lot of new projects in nuclear, because construction costs are high and subject to cost overruns as projects get delayed, and delays seem inevitable. That's why there's all the talk of modular nuclear and what not. If it was feasible to build these plants on time and on budget, the generation parameters are good --- doesn't use much fuel, tends to be available outside of scheduled maintenance, can modulate generation to follow demand, if the economics make sense (as-is, most of the plant expenses are fixed cost, so you may as well generate as much as you can to amortize the cost over more kWh)

According to the NREL (National Renewable Energy Laboratory) in 2022 it cost approximately $1.06 per watt to install utility scale solar. So for the cost of the 1.1 GW Vogtle 4 reactor, utilities could have installed approximately 32 GW of utility scale solar. Obviously this is an extremely simplified and naive calculation that doesn't account for cost or availability of land, transmission infrastructure, battery or other energy storage, etc.

That number does include the cost of land for the solar, which is fairly insignificant.

For the cost of batteries, NREL published this review of reviews, and lands at ~480/kWh of capacity:

https://www.nrel.gov/docs/fy23osti/85332.pdf

So split the $35B (there's a recent $3B that hasn't made it wikipedia, as I understand it) half and half, and you get 17GW of solar, and 36GWh of storage.

As far as translating this into per-kWh costs, most estimates I have seen put Vogtle at $0.17-$0.18/kWh. The equivalent for solar is $0.04/kWh. To charge a battery with that same solar, and then deliver it later, it's $0.13/kWh, when doing the napkin math with those NREL numbers up there.

Lifetime of panels and batteries vs modern nuclear plant? I guess also do the nuclear numbers use the real liability and disposal costs, or subsidized ones?

The cost per kWh numbers take into account lifetimes and nuclear waste disposal

What about Proce Anderson liability cap? Potential trillion dollar incidents reduced to $10 billion cap.

PV panels will produce for at least 25-30 years with limited degradation in output. At that time, consider what state of the art will be, and that you can repower an existing PV plant trivially; you de-energize segments, manual labor replaces panels, and you re-energize. Old panels will get shipped for recycling (shredded and materials sorted for reuse).

Nuclear is the most clean energy that can be generated on-demand, solar and wind can only generate energy depending on the external conditions.

Batteries until today are not a viable solution.

When talking about the electrical grid you have to be able to generate energy in any amount whenever you want.

Providing "base load" is often touted as an advantage of nuclear power plants (NPP) here on HN. The reality is actually the opposite. As the International Atomic Energy Agency says[1]:

"Any unexpected sudden disconnect of the NPP from an otherwise stable electric grid could trigger a severe imbalance between power generation and consumption causing a sudden reduction in grid frequency and voltage. This could even cascade into the collapse of the grid if additional power sources are not connected to the grid in time."

Basically NPPs are designed to SCRAM for all sorts of reasons, then the sudden loss of multiple GW really ruins the grid managers' day. The first paragraphs of [1] make it clear that a large, stable, grid is a pre-requisite for NPPs not a result of NPPs.

[1] https://www.iaea.org/sites/default/files/gc/gc53inf-3-att5_e...

> When talking about the electrical grid you have to be able to generate energy in any amount whenever you want.

Not exactly. It is possible to manage the demand side to some degree.

For example, Octopus Energy in the UK has a "Intelligent Octopus Go" contract which offers much cheaper night rates, in exchange for giving up control over when and at what rate your EV charges. You just tell them what battery percentage you need by what time in the morning. They plan the charging within this constraint and get paid by the grid operator to balance the grid.

Another example are dynamically priced contracts where the prices vary hour by hour based on the day-ahead market prices. I have such a contract and I charge my EV only during the cheapest hours when other demand is low. Sometimes I postpone charging for a day or more because I have sufficient charge for my needs and I expect lower prices later, e.g. based on weather or upcoming weekend.

> When talking about the electrical grid you have to be able to generate energy in any amount whenever you want.

I think there is an error thinking this way.

It is like saying "everyone must own a pickup truck, because they must be able to move any household appliance or furniture item at any time."

"nobody can drive an electric vehicle because they must be able to drive to the next state at any time."

Honestly, you can have a very reliable electrical grid that is runs on solar/wind/batteries for most of the time.

Batteries enhance the grid, and if there was a period of no wind or no sun, it is quite easy to spin up gas turbines for a day or a month.

It's not quite the same when failing to meet peak demand means pipes freeze or people die of heat exhaustion. There are areas in the country that are over a hundred degrees at night. It's going to take a lot of solar to cool all those homes.

Could we all live a different way, communally instead of in our own big boxes? It's physically possible but socially impossible. The truth is we'd rather burn the planet to the ground and we will. That's our nature, little use fighting it.

> to be able to generate energy in any amount whenever you want.

This isn't true. Demand is predictable to some degree on multiple timescales.

But if it was true, why would that help nuclear? It is most suited to the role of producing a set amount of power continuously.

Why are so many people fans of this mythical nuclear power they've invented inside their own head? (or had fed to them)?

It all just seems like lazy, second-hand, anti-renewable propaganda.

Network stability, night and bad weather hedging is a problem though with full solar/wind, isn't it? We still need a mix until tech is better to solve these problems.

It gets really wild when trying to calculate the break-even point :)

Regulation plays a significant role in cost overruns. Also we (should) expect the nuclear plants to work for at least a century while the panels and wind turbines last what, 5-15 years assuming no major storms damage them?

Check your data, newish solar is expected to have less then 20% degradation for 30+ years, afterwards its still usable.. but an exchange might be financially beneficial.

Windturbines are rated for 20+ and can be extended.. current generation are usally changed as there are bigger models available, with higher profit margins

So still 1/3 and 1/5 expected lifetime for nuclear

Wich nuclear reactor has a planned lifetime of 100+ years without expensive rebuild?

That seems very far fetched

Design lifetimes for nuclear reactors are typically 30 to 40 years. Solar panels are often guaranteed to produce about 80% of their capacity after 20 years (and often exceed that) IIRC

> How much solar/wind/battery capacity can 34B USD buy!?

I'm guessing you'd run into space constraints pretty quickly. For one large project a nuclear reactor is incredibly compact compared to most alternatives except maybe hydro.

I don't think space is a concern in North America.

If space isn't a concern, then transmission is.

Also not really a problem in North America, nor is running more transmission lines as capacity and demand grow

There are some very long transmission lines:

https://www.statista.com/statistics/1305820/longest-power-tr...

You don't have to generate all of the power in one place, you know...

Reply on news.ycombinator.com