Study identifies reasons for soaring nuclear plant cost overruns in the U.S. | Hacker News Reader

Study identifies reasons for soaring nuclear plant cost overruns in the U.S.

news.mit.edu

240 pointsconsumer451302 comments

I think a pervasive theme in US infrastructure (and why it's expensive) is that we've gotten out of practice of building. I think it's the deeper "why" to this issue.

We no longer have a broad reservoir/cadre/pool of practitioners of engineers, construction crews, experts, designers (except for a very concentrated few) who do this week in, week out. Every job now seems custom. Only Parsons or whoever has the ability to do it, and even they have to scrounge around for the project manager.

Other countries are in their building phase and have thousands of engineers who are practiced in standardizing designs and squeezing out inefficiencies. They have graduating classes full of highway engineers. For gods sake, they even have graduating classes full of engineers specializing in the tooling for highway / railroad engineering machinery.

Our building phase was post-war construction-crazy economic expansion era. Now we're in the maintenance phase, where big new projects are the exception. Not to denigrate the great importance of doing great maintenance, but in practice you can't keep (or attract) a large pool of well paid expert engineers going on that level of activity. They go where the projects go.

And as a result, as the report points out, when everything is "for the first time again" and custom because you do it once per decade, the particular factors of a given site are allowed to dictate what the system design should be. (Maybe also not insignificant: the environmental considerations allowed to creep in, just to point that out, as grateful as I am for the protection that we give these matters). Rather than (for better or worse, probably a bit worse) in other countries, they're stamping out highways, railroad, power plants by the week, and the environment is made to fit the construction.

It's choices we made, over time, that caused us to atrophy our capabilities here. It could be brought back with investment.

The gov should just pay people $15 an hour to study welding. It's the best way of;

i)getting people trained, ii)having a kind-of minimum wage accessible to many people.

Why would the gov. need to pay people? if there is a demand for welding people in the market, their pays would signal people to study for it isn't it?

They get payed for the work, not the training (with some exceptions). Not being payed for training is a barrier for people who are not in a good financial situation.

I wonder why banks don’t offer loans for job training programs. Seems like it might work. Or maybe they do?

You can get student loans to go to welding school - yes.

If there are lot of potential people with willingness to train and a market to pay for trained people a market for training them with deferred loans etc. would be there?

People change plans now to need fewer welders (like deciding not to build things) - because there are too few welders - which results in fewer welders in the future. Becomes a self-fulfilling prophecy. Government action can help break that cycle.

This sort of thing has been theorized under the name of the Job Guarantee.

https://en.wikipedia.org/wiki/Job_guarantee

Tons of welding jobs already exist and already pay a minimum of approximately $15/hr. Unless you meant welding engineers when you said 'study', like some of my compatriots building automated robotic weld cells, who get paid 4x that much.

And there's plenty of training courses at the local caterer center and community college for both those jobs at reasonable rates. Local manufacturers often pay hourly+tuition for their best fabricators or line workers who want a promotion to the welder position to go through one of those programs.

Welding seems to be this fruit that people reach for when they think of Trades.

There's lots of welding schools churning out 'graduates' throughout the year. Welders are almost always the first guys laid off in shops.

The 'learn to code' thing thrown around would be better.

I wish it would not be the narrative. I think in university and where else appropriate, it should be taught, like a friend of mine who works in finance said, everyone will need this as a skill eventually for complex work

The whole narrative of producing more 'software developers' in general though? No thank you. I see enough people coming off of less than stellar bootcamps (there are, in fact, good ones, though I'm finding this less true over time) who just see dollar signs who have no interest in this as a craft. I don't want to encourage that, at all.

In fact, I want certification for the title of Software Engineer, no different than having to be certified like civil engineers do. Uncle Bob is right about this one.

What's special about welding? You also need carpenters, plumbers, electricians, civil engineers, etc for any major infrastructure undertaking, having 50,000 unemployed welders doesn't really make any sense.

The issue is mostly political and not financial or knowledge problems. How you accomplish that is government spending, and there's a certain subsection of the population that has been convinced that government spending is universally bad.

There are vested interests waging a war against the US building infrastructure because its ideologically against the narrative developed.

"Starve the beast" has turned any kind of fiscal spending into a huge political battle. It used to be, funding basic things was just a given. Now it's a fight over every little dollar.

The factions against this kind of government spending also happen to be the same ones who blow up the debt and deficit on tax cuts for the rich when they're in power, then bang on the table about how it's the most important issue when they're out of power.

In approximately 61 days, conservatives in the US will be in an absolute panic about the financial state of the country.

The boomercons in my circle have given up on fiscal responsibility and have now just accepted the fact that we're going to default at some point, so there's no point in fretting about it. It'll be interesting to see if the Congress critters follow suit.

The US debt is denominated in a currency that it is the sovereign issuer of.

The US has full ability to pay such debts, at the cost of inflation.

Large deficits would be fine if the spending were for investments with a high ROI, such as infrastructure or nuclear power.

Indeed, going further into debt for an investment with great ROI makes sense. Regardless of the debt level, a good investment is a good investment. And a bad investment is a bad investment.

Alas, most spending of the US federal government isn't an investment of any type.

> Alas, most spending of the US federal government isn't an investment of any type.

Unless you consider being the sovereign issuer of the world's reserve currency, and the military, civil government, legal system, and property rights behind that to be an investment of some type.

The majority of spending goes towards pensions, healthcare, education, etc.

Education could be an investment but per student spending has tripled over 50 years, yet it would be hard to argue that there is a high marginal return on investment from the extra spending.

Same with healthcare. Plenty have noted that the US spends more per capita but has worse outcomes. Seems like there's simultaneously cheaper and more effective approaches out there.

Social security is just a transfer payment, but not sustainably structured.

> The US has full ability to pay such debts, at the cost of inflation.

That point could reasonably be interpreted as rather than taking in a dollar and paying back 50 cents, the US has the power to take in an hour of labour and pay back a half hour. And claim that it fulfilled the terms of the agreement.

That isn't an improvement. It is nothing more than muddying the waters enough it makes it hard to point out that the situation is no different from a default. Which I suppose is why they do it.

There is an amount of stuff in the world at this moment. The US either gives back more stuff than it borrowed, or less. The idea that it can give back less stuff but that is OK because it is a government is a foolish idea. Making it hard to understand who is getting what doesn't magically mean everyone is happy, it just means we only find out who is paying with benefit of hindsight.

> The idea that it can give back less stuff but that is OK because it is a government is a foolish idea.

Anyone can pay back less real value if they have terms where the rate of interest is lower than the real rate of inflation.

Creditors of a government can mitigate the risk of that happening due to the policies of the debtor government by setting repayment terms other than in fiat currency of the debtor.

They often choose not to; heck, in the case of the US, it's happened that they choose to lend despite negative nominal rates, even before inflation.

Capitalists?

Republicans of course.

> tax cuts for the rich

You mean those who actually pay taxes? The top 10% pays 70% of all federal tax.

The most current tax cuts also lowered the brackets in almost all cases, as well as raising the standard deduction. These changes reduced taxes for almost every American.

> In approximately 61 days, conservatives in the US will be in an absolute panic about the financial state of the country.

As opposed to the last 4 years where democrats in the US were in an absolute panic about everything.

The truth is that somehow we are conditioned to believe that only income deserves taxation. We have under-taxed wealth so much that there needs to be some equalization.

That was a great trick that the wealthy pulled, convincing us to jealously despise well-paid workers.

It comes in the form of a wealth tax. Instead of sitting on their cash/capital piles they have to do something with it or lose it.

But they own 95% of the wealth so shouldn't they be paying 95% of the federal taxes?

It's actually roughly 70% for wealth as well, so I guess it's fair eh?

> Deutsche Bank’s Torsten Sløk says that the distribution of household wealth in America has become even more disproportionate over the past decade, with the richest 10% of U.S. households representing 70% of all U.S. wealth in 2018, compared with 60% in 1989, according to a recent study by researchers at the Federal Reserve. [1]

[1] https://www.federalreserve.gov/econres/feds/files/2019017pap...

Yes, it's all a game. The minute they're in power, it's spend spend spend. The minute they're out of power, suddenly the deficit is a huge deal. How many actually care about the deficit is questionable, but it's clearly very low.

> Now it's a fight over every little dollar.

Unless a bank is in trouble or someone wants to fight a war.

> certain subsection of the population that has been convinced that government spending is universally bad.

unless that spending is on them. That subsection of the population LOVES welfare when they get it and it's called something else. You could just call this something other than welfare.

Please explain the $578 million dollar elementary school in Los Angeles, California in 2010

https://www.huffpost.com/entry/robert-f-kennedy-communit_n_6...

Excess on all sides

It's an outcome of polarization, imo. One side says government spending is evil so don't do any, so the other says government spending is an inherent good unto itself, so damn the costs.

Your link describes it as a "K-12 complex". Maybe you think they should have thrown in preschool and PhDs too for that price, but there's no need to exaggerate.

Google says it's actually six different K-12 schools crammed into one mega-complex.

The article explains why. Perhaps you should read it.

> "Starve the beast" has turned any kind of fiscal spending into a huge political battle.

Strangely though "Defund the police" goes the other way, politically. /ducks

There's no equivalence. The rationale behind each of these two positions is completely unrelated.

> and there's a certain subsection of the population that has been convinced that government spending is universally bad.

Oh of course! It's the Republicans' fault, how predictable. If only we had a Democrat utopia like California at a national level, we'd have nuclear plants everywhere and everything would be perfect.

Californians don't want nuclear plants in California! Wyoming or Utah might be the perfect distance away, and prevailing winds are also ideal...

Considering their abysmal failure in California, yes, we don't want them.

I don’t disagree with you generally.

Luckily for the nuclear industry, the engineering capability still exists in the US to design a whole plant, and a lot of the manufacturing base still exists. Big US nuclear players have had a base load of domestic upgrades and international new plant jobs that have kept these capabilities alive.

But for construction, we really are awfully thin on specialized trades. Welders, especially, are _always_ in short supply.

  Welders, especially, are _always_ in short supply.

I've heard this time and again over the years, but it doesn't make sense to me. If welders made > 300k/year, surely they wouldn't be in short supply anymore. Why isn't there simply a wage/price signal to compensate for this? It's not like medicine that requires 6-8 years of school, and years more of practice to fill certain positions (where additional constraints are the schools and hospital residencies limited capacity). For welders, labor supply should be fairly easy and quick to remedy.

I think there's an underestimated resistance to changing the labor rate. If a company quotes a typical job at 80 man-hours with $15/hr welders, and a year later you want to hire welders and can only find them at $20, they're more likely to complain there's a welder shortage and decline jobs because they don't have enough welders to complete them than increase their labor rate. Plus, if you increase the rate for a new guy, and your $12/hr hire from a decade ago hears the new hire is making $20, he's going to be pissed, so you'd have to increase costs across the board or suppress discussion of salaries.

This explanation doesn't really hold water. Some company is going to eventually win the contract by bidding with $20/hr welders right? It's not like the client is just going to give up on their project.

No, I think you do see clients give up. There are certainly more infrastructure maintenance projects in the US than governments are willing/able to pay for. I’m not sure the market can bear much higher costs for construction.

There was a rather famous story about Aceribo being abandoned due to lack of upkeep and high cost of repairs.

> This explanation doesn't really hold water.

During the 2010-2019 US recovery wage stickiness was noticed, even with everyone complaining about not being able to find qualified workers. Companies were very reticent about increasing wages, often preferring to go with signing bonuses.

It is suggested that this reticence/stickiness was because after 2008-2009 GFC many places were stuck with high wages (from the previous boom), and getting people to cut their salaries was almost impossible.

* https://www.investopedia.com/terms/s/sticky-wage-theory.asp

* https://en.wikipedia.org/wiki/Nominal_rigidity

I see this with those profiles of factory owners complaining about Those Dang Kids not wanting to work, while omitting the fact that they offer too little relative to other options. A lot of people running places like this don't want to accept they're also in competition with retail now.

People look at the small difference between pay at the factory or mine (dangerous) and pay at Walmart (not quite as dangerous + employee discounts) and make the rational choice.

It takes many years to become a skilled welder, and nuclear is the hardest class of welding to get certified for.

A few months at the local community college can get you to the point where you can do the most basic welding jobs, but these low skill jobs are the ones that either left the US or got automated, and the few such jobs that remain are mostly done by mediocre welders who should have retired years ago. If you want to do the high skilled welding that pays well, you need to essentially support yourself for years as you develop those skills on your own - and most people who have the means to do so have better opportunities in other fields.

Realistically, employers need to eat the cost of training new people if they want the pipeline to open back up again, but it's more economical to poach a good welder after some other sucker puts their resources into them, so you get a classic tragedy of the commons.

It takes literal years of daily practice for a welder to get good enough to do the high quality specialty materials pipe fitting welds that would make up so much of nuclear plant construction. I am not exaggerating. A good welder/fitter team can work pretty much wherever they want, they are in incredibly high demand. There aren't more of them because not that many people have the baseline talent to be able to achieve the ability, regardless of how much practice they put in.

Source: used to weld at the NASSCO shipyard in San Diego.

Could you expand on the "baseline talent"? This isn't the sort of hand-eye coordination that would be required to hit a major-league fastball, is it? (Somehow those employers are able to fulfill their skilled-labor needs.) How would an individual welder determine whether she has the talent that would make those years of practice worthwhile? If that is not possible, why don't interested organizations set up some testing programs?

Hand eye coordination isn't really the issue.

When welding, you put an incredible amount of heat into the material. This causes it to move in real time. A welder needs to predict how it will move ant actively account for this so that the final piece winds up in the right condition. The tolerances depend on the application - for furniture who cares if it's out a bit, for a coolant line to a reactor you need it to be damn near perfect.

Then there's heat penetration - adding heat to the material changes its material properties. Go too slow and steel becomes excessively weak, go too fast and you don't form a strong bond. Again, big differences between "probably won't break if you sit on it" and "this will kill a dozen people if it fails"

Finally, there's just plain old margin for error. If you're welding together consumer products, it doesn't matter if you scrap one every now and then. However if you're trying to join together two multi-million dollar custom fabrications, you need someone who's going to walk up to this and get it on the first try every time.

I'd compare it less to being a major league baseball player and more to being a professional musician - you can't determine that someone could never do it, but some people are certainly born with a knack for it and you can't simply churn out more on a whim.

Thanks for the detailed and insightful answer! It seems that welding is what takes over when engineering can't go any further.

> Welders, especially, are _always_ in short supply.

In short supply for the wages offered. Bid enough out there, and the positions will fill. Bid consistently such that precarity is avoided and bid even more beyond that, and the bench will deepen. Short supply is entirely addressable by buyers.

I think these arguments assume welding is a low-skill job. https://news.ycombinator.com/item?id=25166696 seems to explain why it isn't.

There are definitely some industries where the cost to get to the level where you make good money is too high, especially when you aren't sure when you start if you have the required talent to make a great living.

There will always be a shortage of great NBA centres - for reasons outside labour market functioning.

Welding is a huge problem space, and I see a lot of welding shops talk about how they're constantly lowballed by buyers who can't tell the difference between hot-glue-gun-for-big-box-store-steel-blanks work and the-device-to-certify-your-weld-costs-more-than-your-house work.

There is only ever a shortage of sufficient compensation to clear the market for a desired level of skills availability.

Welders are in short supply is just CEO talk for "we don't want to pay them more".

The only hope we have is to hire China to build plants for us at this point. The USA is forgetting how to do everything and becoming more antiscience every year.

This is exactly the reason I think Richard Nixon’s 1973 plan for “1000 nuclear power plants by the year 2000” would have been so impactful for us. It’s hard to get good at only building a handful of something.

e: I misremembered; it was actually 1000 by 1980

A thousand! We’d be so rich in carbon free energy we’d be exporting it to Canada and Mexico and still have hundreds of GW left over to, I dunno, produce hydrogen 24/7 for other uses.

We could decarbonize our grid today with only something in the vicinity of ~200 new plants. This is entirely doable. France showed how it’s done in the 80s. It’s sad we can’t start now.

Unfortunately that was a very different regulatory era for nuclear development than we have today, so who knows if it would be feasible. Watergate happened not long after that declaration, and one of Gerald Ford’s first major acts as replacement President was signing the Energy Reorganization Act of 1974, to split the Atomic Energy Commission (responsible for both civilian and military nuclear development) into the Nuclear Regulatory Commission (responsible for regulating civilian nuclear power) and the Energy Research and Development Administration (responsible for supporting military nuclear development).

To be fair those are completely different things. Commercial needs a profit and uses low enriched material. Military is not concerned with profit and uses highly enriched material.

I am not as familiar with the commercial side but I can say the military model is safe as all get out and produces less waste since you only need to refuel ever few decades or so.

Also NRC is very much still involved with Military applications, which at this point just involves moving expensive war ships around.

> produces less waste

But do nuclear reactors really produce any waste, or only “spent” fuel that’s recyclable?

Recycling used fuel still doesn't make much sense. It's very expensive to recycle and new nuclear fuel is cheap. Storing the waste isn't technically difficult either. The total quantity of waste is tiny, but politically it's very difficult to deal with.

Due to some logistics complexities recycling fuel doesn't make as much economic sense as it does for places like France. Though if we did have 1000 plants it would make a lot of sense haha.

But besides that there is always waste. The key part is understanding what level of waste. There's a fair amount of low level waste (low radiation levels). In fact, this is like 90% of total waste (even more if you count my volume). But this type of waste is not the kind people are typically concerned with because it isn't radioactive for long nor is it producing dangerous levels. These types are not really recyclable though (concrete, steel beams, etc).

I do suggest reading up on our own AcidburnNSA's post about waste[0]

[0] https://whatisnuclear.com/waste.html

To godelski point yes there is low level stuff that really isn't a concern except to jandrese's point it's political. If I walk out of a reactor room and my gloves from my contamination suit pop too high they get put in a bag that is marked nuclear waste.

Granted the too high mark could still be perfectly fine to throw into a landfill and not impact anything, but nobody wants that in their neighborhood.

As to high level stuff which is what you are asking about, the spent stuff is really hot. To recycle, it would need to cool down first. Commercial plants place that in cooling pools, which is kind of bad if you have a Tsunami hit your plant (Japan), our something else.

So as to the waste element think I have 2% enriched object that is 100 cubic meters. The same in a highly enriched say 90% would only be about 2 cubic meters. This means my pools would be smaller, my overall size is smaller, and I can contain the smaller pool easier preventing the release.

On the politics side the idea of highly enriched is dangerous because it could be used for other purposes, however when those decisions were made I don't think anyone every imagined the concept of a dirty bomb in which enrichment is really going to matter one bit.

France delivered 50% of their nuclear cores (the 900MW model) between 1980 and 1985. It isn’t very good because they will all EOL at the same time. I hope Nixon’s plan would have been to streamline and spread their production, which then becomes extremely efficient, a bit like our P’4 in this graph:

https://commons.wikimedia.org/wiki/File:Chrono-parc-nucleair...

I'll give the benefit of the doubt to the parent post here, and add, that in the construction industry in California over the last thirty+ years, the pressure on middle-tier contractors combined with human nature, made a blatant game of taking away the pay from anyone who could not defend themselves in a transaction. That sometimes is the workers, sometimes the suppliers of materials, sometimes the contract holder.. Theft escalated and wages stayed stagnant or actually fell. "Illegal" labor is ordinary and expected.

I would welcome some case studies from the commentors here, because "reasons" given in comments are not explaining what I have seen with my own eyes.

> "Illegal" labor is ordinary and expected.

I agree with your entire comment, but I want to emphasize the part about illegal labor.

There are entire industries built on the back of illegal labor, and I think many otherwise informed Americans are not aware of this.

We have some fairly substantial mismatches between our labor market and our immigration policies.

Before someone says “just don’t use illegal labor”, please realize that this means that you will be out of business due to a lack of competitiveness since literally everyone else in your line of business does it.

Personally, I have mixed feelings about this. On the one hand, maybe it’s just better off staying like it is — the market is clearing. On the other hand, I kind of wish proper enforcement of the laws would happen. This would force these companies to pay wages that legal workers would take. This, of course, would require them to raise their prices to the end user, but it would give legal laborers more... of something... definitely money, definitely rights, and maybe a bigger piece of the pie.

I don’t work in any of these fields, but I know people who do. I would love to hear from HNers who have first hand experience.

> many otherwise informed Americans

It also has to do with conversation venue. Think how many HN commenters work for companies like Uber/Lyft, just for one example.

There are entire industries built on the back of illegal labor, and I think many otherwise informed Americans are not aware of this.

Indeed. The game though, is a big reason for hiring people illegally is to keep costs down. Because illegal workers area afraid to complain and will work for less.

Now, when the average moron finds that there are lots of illegals in industry X, or in the country as a whole, they get up in arms about the illegals, support a crackdown but the crackdown doesn't remove the illegals but makes them more afraid, reinforcing the whole system. So the typical ideology of this country makes accomplices of people to the racket when sun light is (selectively) shed on the situation.

It's funny that China runs their on the same legal/illegal division of labor but in China the "illegals" are internal to China, people from the countryside who don't have official permission to be in the city.

To point out a particular example... One can hire various masons at various prices who would be capable of laying a flat brick or CMU wall. As soon as you require skilled masonry, of real natural stone or in walls or columns with interesting shapes and orientations? Now you're hiring Mexicans or Central Americans. At any price.

One can imagine the several decades of law enforcement that would be required to change this situation, but it is an open question whether it would be a good thing overall.

Theoretically, these businesses are already priced to maximize profit. So it could be that increased wages increases the cost to the consumer. It could also mean lower profits for the business if they've already priced at the max their customers are willing to pay.

Uncracked illegal labor is just a tax incentive. Tax incentivizing construction is not crazy.

Could you say a bit more about the game being played here? No need to spill any secrets, but I'd love some "for example" items to understand it better.

Maybe CA is different, so YMMV, but im NY the game is unions, and filling the site with so many workers, due to union regulations, that they can't even work from each other. So there is a lot of waiting, which of course is billable. Now on top of all this the workers might be working illegally. And who actually compares on-site events and crew numbers/sizes with invoices? Well, if you don't want trouble, you don't.

> I think a pervasive theme in US infrastructure (and why it's expensive) is that we've gotten out of practice of building. I think it's the deeper "why" to this issue.

When it comes to nuclear, that's a problem the world over. Out of practice of designing as well. Look no further than the mess that is the EPR. On its face it seems like a boondoggle, but it makes a lot more sense when you look back at the history of nuclear construction: from 1971 to 1993, France was continuously building plants, first the 34 900MW CP-series (1971 to 1988) then the 20 1300MW P-series (1978 to 1993). The CP-series took 5-6 years to build and the P-series a bit more (6-7). From the mid-70s to the mid 80s, there were a dozen plans being built at any time, and they were being completed.

Then came the N4, that thing had serious teething issues, the first model started construction in 1985 but took 12 years to complete. By the time the 4th was completed, it was down to 7 years. But the 4th was the last N4 to be built. It's unclear to me whether the N4 slowdown was intended all along or whether the teething issues made them take a look at construction and pump the brakes, but in the space of 5 years the country went from building 12 plants at a time to building 5, then 4, then 3, then 2, then 1.

N4 at least overlapped some with the tail end of the CP-series and the middle of the P-series, but the first EPR construction was started in 2007, 5 years after the last N4 (Civaux 2) was completed. I can't even imagine the loss in institutional knowledge over that span.

> When it comes to nuclear, that's a problem the world over.

I think there's a perception problem.

Do people readily go to school to study nuclear engineering?

I don't think it carries a stigma per se, but it would not be that sexy, and it would be harder to get a job and keep it subject to political whim on a 4-year-cycle.

(A more practical/guaranteed job would probably be a petroleum engineer, especially wrt fracking)

A lot of people are going to be discouraged simply because it sounds and is a difficult subject to study. And not everyone with enthusiasm is going to have the grades.

> Do people readily go to school to study nuclear engineering?

Yes, people do study nuclear engineering. I was considering that path myself in university, but ended up going more towards science and doing several years of research in nuclear physics labs instead.

I'm very glad I didn't end up doing nuclear engineering. The nuclear industry stagnated and failed to address cost problems and bad project delivery (cost overruns, delays, etc).

And then between 2010 and 2020 renewable energy scaled up and became incredibly cheap, in fact the cheapest source of energy in most countries: https://about.bnef.com/blog/scale-up-of-solar-and-wind-puts-...

Between 2010 to 2019 wind energy become 70% cheaper and solar became 89% cheaper: https://www.lazard.com/media/451082/lcoe-8.png

Battery costs have dropped 75% over the last 6 years: https://www.greentechmedia.com/articles/read/report-levelize...

> I think there's a perception problem.

Somewhat, but the nuclear industry did a massive PR campaign to address that.

Blaming perception or "green activists" or "fear of scawy radiashun!" is an easy straw-man for them, but the real obstacles are more practical and harder to address: the economics of nuclear are not good. This is compounded by a problematic history of delivery problems.

> > Do people readily go to school to study nuclear engineering?

> Yes, people do study nuclear engineering.

Note I said "readily" not "really". I see you made the calcluation when you made your decision.

It's a shame really because I believe nuclear power has great promise in benefitting society.

Too bad the moonshot level investment in nuclear went towards military ends (france in 80's excepted). And too bad funding is dependent on political whim.

Although solar and wind is working, I worry that might be a "save ourselves rich" strategy.

When you look at this table:

  material              energy density mj/kg
  Uranium               22,394,000,000
  Diesel fuel           38.6
  Lithium-ion battery  2.63

I think it would be nice to think about the benefits of growth on the high end not just savings on the low end. I believe making energy available in ever-increasing plentiful quantities at low (to zero) cost would really change the world.

Nah, the nuclear industry got a TON of R&D funding too, and continues to. Plus I don't think you realize just how much governments subsidize and assist nuclear power. $8-10 billion per reactor is more money than companies can easily plunk down for an investment that takes decades to pay off. They rely on government loans, subsidies, and loan guarantees to raise capital and help fund the projects.

Loan guarantees are a particularly common example -- for example, last year the US govt provided $3.7 BILLION in loan guarantees for the troubled Vogtle reactor builds in Georgia. The way loan guarantees work is that the government agrees to pay the debt owed if the company goes bankrupt or cannot complete the project; this is surprisingly common actually, since the same project had already driven Westignhouse to file bankruptcy.

Solar and wind aren't likely to make a handful of people extremely rich -- the profit margins on the components are pretty low due to the high degree of competition pushing them down. Building solar and wind farms to sell their power does produce a solid and reliable financial return on investment, but it's still less than 10% (even though the energy return on investment is good).

> When you look at this table [energy density]

For power generation, energy density is largely irrelevant.

Also these energy densities are apples-to-oranges comparisons. You can't burn uranium without a fairly large, heavy reactor. There are no "nuclear cars." Similarly diesel requires an engine although it's smaller. Batteries aren't a power source on their own, they're storage, and you're not "burning" them up.

Maybe the answer is (a) Green New Deal. Basically just rebuild the whole infrastructure - no tinkering around the edges, just replace the lot.

Start with engineering courses, bringing on new talent. Spin up hundreds of solar farms, battery factories etc

My question is are there real innovations to be made - do we need to rely on concrete and steel ? Are new homes possible in flat packs?

I would rather packet it into small parcels (10-100M) that explore the phase space of possible than think a quick government report can define infrastructure for a century

On top of which we should heed the strong town lesson - only build what we need, and cost in the car externalities in our cities.

Just thinking out loud ...

In 50 years do we come to the same point in the road with different infrastructure?

That would still be a better outcome than if in 50 years we came to the same point in the road with World War 2 infrastructure.

One of the problems with New Deal-like projects is what do you do with new engineering and construction talent once the projects are complete? China is pushing OBOR onto developing countries to make use of its excess construction capacity, with limited success. I don't think that's feasible with high-cost US labor.

Export it, if you have such a deal you will become a leading research and development center for it, it can become a feedback loop.

You will still need engineers for improving the technology, you will be on the bleeding edge. You will still need maintenance crews, the constructors are the ones that know the most about what they built, why not train them to maintain it as well?

By this time, there will be a demand for the development of space resources.

I'm sorry to disagree with this.

First - 'Green New Deal' is a slogan, not anything material that helps fix underlying operational problems.

Second - we don't need to 'rebuild' most things, we need to fix, maintain.

Third - this is not an issue with 'green' or 'innovation' in the technical sense - it's about the efficiency of the bureaucracy.

Fourth - 'All the wasted money' is going into the pockets of seriously powerful vested interests - they are not interested in reform.

Are you ready for the political cost of taking on the Unions? Esp. public sector unions? They may not have their fingers in the pot, but they still likely a massive barrier.

What about exposing the fact that a ton of workers in infrastructure are in the country illegally? Who's going to 'clamp down' on that and pay the political price?

What about the fact that all these firms are massive political donors to the powers that be? How do we solve that one?

And fighting entities: SF is a tiny city. So is Palo Alto. So is Oakland. The 'Bay Area' needs coordination and nobody really wants to give up their little fiefdoms of power. So there are a lot of problems like that.

The issues are complex and a lot of it has to do with vested interests and power, not 'technology' or anything like that.

>Third - this is not an issue with 'green' or 'innovation' in the technical sense - it's about the efficiency of the bureaucracy.

The entire discussion has been how this isn't the major driving factor of the cost of building. So why bring this up at all?

> Are you ready for the political cost of taking on the Unions? Esp. public sector unions? They may not have their fingers in the pot, but they still likely a massive barrier.

> What about exposing the fact that a ton of workers in infrastructure are in the country illegally? Who's going to 'clamp down' on that and pay the political price?

Unions are going to clamp down and pay the political price of squashing illegal labor. That's their whole purpose of existing! Anti-union sentiment in America reflects significant cognitive dissonance on this issue.

Public unions are a different matter. They're entrenched largely because private sector unions have disappeared. With greater private sector unionization public employees wouldn't dominate union lobbying as much as they do, and politicians wouldn't feel as obliged to protect them if there existed a larger pool of well paying, stable private sector jobs for the working class.

I think another important factor is that to the extent we have people with lots of daily experience, they're not in charge. We've shifted from work-focused to plan-focused behavior. To the extent that we don't even know there's another option.

A talk that really opened my eyes was Mary Poppendieck's "Tyranny of the Plan": https://chrisgagne.com/1255/mary-poppendiecks-the-tyranny-of...

It turns out the Empire State Building was created on time and under budget, and against a very hard deadline. They did not have a complete plan when they started; they were designing the top floors as they were building the bottom ones.

In software, we've become familiar with how over-planning creates a lot of waste. But I was still surprised how broadly that lesson applies.

Commercial skyscrapers get built on time and on budget all the time (in not-California-land, at least).

They sign big leases with big tenants before they can get financing, and the building needs to be open when the lease says it will.

Strangely enough, pretty much the last thing they do is to design the foundation. Once they know the building is a go, they drill and take core samples at dozens of places around the site and use that information to design a foundation that will hold the building as designed. This is the when things start to cost large amounts of real money, so they don’t do it until they have to.

The concrete mat that ties all the caissons together is a single pour that can involve hundreds of concrete trucks and take a day or more to do, all of it orchestrated to ensure every truck arrives exactly on schedule.

All this is to say that there is still an awful lot of skill in the construction industry.

I certainly agree there's a lot of skill; my point is more that it's often not running the show. As to on-time/on-budget rates, I'd love to see some data. Significant cost overruns and delays certainly still happen. E.g. this non-CA building was at least 20% over and a year late: https://www.inquirer.com/philly/business/comcast/comcast-tow...

And I'll note that the Empire State Building was not a standard office building of the day, but the tallest building in the world. (A title it held for 40 years.) People building today have an 80-year advantage, plus the rise of computers. You'd think we'd be able to do at least as well with every building, but we don't. That's in stark contrast to building safety, for example, where we're very good about learning lessons about keeping buildings from collapsing.

"We no longer have a broad reservoir/cadre/pool of practitioners of engineers, construction crews, experts, designers (except for a very concentrated few) who do this week in, week out. "

No, we definitely have those people.

Thinks are more complicated across the board, safety practices are higher, contrary to public opinion wages are higher in this sector and I think there's a soft kind of pervasive corruption going on at the government and contractor layer, where there's just padding, waste, overbidding, back-scratching etc..

Montreal had some serious problems with corruption and was able to make progress on it, states and municipalities need to do the same.

I would vote for the first person that wanted to make these things more efficient and transparent.

"[M]ore efficient and transparent" are at odds with each other. Transparency requires more process and scrutiny and inspection; more of these things increases project risk because bidders need to invest more money in getting the bid, and that increased risk is amortized across successful projects. It also has the effect of favouring entrenched interests and existing players because 1) they know the process, and 2) have the capital to endure longer lead time in sales.

>They go where the projects go.

So where do they go now? Are American trained Mech/Civil/Industrial/etc Engineers really going overseas in large numbers? We all know the cost of schooling is now exorbitant in the US and engineering salaries (certainly for software but I'd guess other disciplines as well) are generally higher here than anywhere else in the world. Where could they even go to make enough to pay off their student loans?

I feel critical of approaches that make generalization using "we" when talking about a serious of actions by some very specific institutions.

The US has a complex state and a number of private corporations with a relationship to the state and each other. Who are you talking about in particular?

Edit: I should add. It's not just the problem of having trouble building things but companies that get a lot of money out of failure and those companies symbiotic attachment to the government - consider the X billion dollar California rail planing process. Whoever did that didn't even have to create more than minimally to still leave with truckloads of cash. And they'll be on call for next time. Even that is just an example.

I have an extremely half-baked idea of allowing mutual colonies-- landswaps between countries with special local laws and mobility. This would result in specialized supply chains and labor groupings that would enable, say, groups of Dutch or German brick/stone road builders to operate in America and show us how it's done.

> building subsequent plants based on an existing design actually costs more, not less, than building the initial plant.

Unless building a nuke plant is a common thing where the crew that built a plant in Georgia can then go build one just like it in Alabama then one in Tennessee, then in Ohio, etc, then I don't see how building two plants with identical layouts can leverage economies of scale. For example: Wolf Creek NGS in Kansas and Calloway NGS in Missouri were apparently the first two plants in the US to be built using the same blueprints. If two more plants were built today with those same blueprints but by different construction crews who don't have any of the tribal knowledge from the construction of Wolf Creek and Calloway then what you have are four essentially bespoke plants that just happen to use the same blueprints.

But if subsequent plants are built with same drawings, you don't have a design engineering firm having to create that work from scratch. Should that alone not save money?

Not when those original plans and designs are 30-60 years old.

We have newer reactor designs, materials, and engineering knowledge today that could significantly reduce cost just in material savings or construction time.using something like FAST reactor, is significantly different that the older tech and could save 90% on fuel and waste.

The reactors being built today are Gen III/Gen III+ designs: AP-1000, the EPR, VVER-1200/ AES-2006, APR1400, etc. These include all the design and engineering refinements we've been able to cram in. They're a far cry from 30-60 year old tech.

Unfortunately most of the new reactors built in the US and Europe have run massively over time and over budget, despite new technology. Vogtle 3&4 in Georgia drove Westinghouse bankrupt. These were modern AP-1000 models. Flamanville 3 in France (an EPR) is running nearly triple its cost estimate and the delivery time ballooned to 15 years. Olkiluoto in Finland (the first EPR) went massively over time and budget as well. These reactors were specifically designed to be more cost-effective and promised much lower prices, but failed to deliver.

The problem in the nuclear industry isn't the technology itself, but the fact that they consistently fail to deliver projects within their allotted time and budget. Unfortunately this shows no signs of changing, and renewable energy industry looks poised to completely out-compete them in the energy market.

I say this all as someone that used to have high hopes for nuclear tech, after working in nuclear physics research all throughout university.

They may be gen3 and gen3+ and newer designs, but they are still just working on the old regular fission model. I'm talking about using breeder or FAST reactors, which are completely different in the underlying physics. So far I believe there haven't been any of those built commercially but have been in testing for 60 years without issue.

There's a reason countries don't build breeder reactors. Fuel costs are a tiny fraction of the costs for a nuclear powerplant: less than 10%. Breeders save a bit of money on fuel in exchange for a higher capital cost (cost of construction). For reactors, capital costs are a huge factor because reactors are extremely expensive already ($8-10Bn per reactor in the US/Europe). Increasing that further more than balances the savings on fuel.

Thus, breeders generally end up being more expensive than a conventional BWR or PWR.

Here I should mention that I spent some time in nuclear physics research. There's a lot of misinformation floating around about nuclear energy. Most of the "miracle solutions" don't live up to their promises (especially thorium tech and breeders). If they did, we'd already be using them -- nuclear engineers are not fools, and most of these reactor concepts have been kicked around for literally decades.

One other point: the physics behind breeders and conventional slow-neutron reactors isn't fundamentally different. Both neutron capture ("breeding") and fission ("burning") reactions happen in both, the ratios in a breeder are just optimized to favor the first process more. In fact in conventional light water reactors, around a third of the energy released comes from fissile isotope bred from fertile isotopes such as U-238.

With the newer breeders (FAST), the main difference is safety. You don't need the same control mechanisms because they automatically cool down due to the physics of them so that they don't melt down. I'd imagine that FAST reactors eliminating the need for ever increasing safety tech and complexity of it for conventional reactors would be a cost savings (or at least make it close). Not to mention the clean up costs when you compare to a conventional reactor that could melt down, even if it's rare.

I'm not saying that the engineers are idiots. But there are some engineers (supported by government or corporate funds) still building new prototypes and testing new designs, such as FAST. Especially in the US, a driving reason that new designs aren't used is that there have been few built I'm recent decades - partially due to lower cost alternatives and also due to public opinion.

I think people -might- be more on board if they knew the nuke plant sitting a couple of miles away wasn't going to go critical/poison their local acquifer and was provably safe beyond a doubt even to skeptics. I think that's what it will take, otherwise solar will have to be our savior.

Passive safety is what you describe. That's a requirement for reactors to be classified as Gen III, so all of the models listed above have some variant of that.

> eliminating the need for ever increasing safety tech and complexity of it for conventional reactors would be a cost savings

Passive safety features are useful, but they don't end up replacing active features (you still need to control the reactor during normal use). At best they might allow for reducing the redundancy level on a critical system -- which might save a bit bit of money, although not much.

Better safety is always a great feature in general, but it's not close to making breeders cost competitive on its own.

> Not to mention the clean up costs when you compare to a conventional reactor that could melt down, even if it's rare.

Actual meltdowns are exceedingly rare (and catastrophically expensive + devastating), so you don't really factor them into the cost equation for a normal reactor.

> Especially in the US, a driving reason that new designs aren't used is that there have been few built I'm recent decades - partially due to lower cost alternatives and also due to public opinion

Mostly cost tbh -- nuclear energy has been historically somewhat unpopular (especially after major accidents) but there's a lot of industrial projects that continue anyway despite being unpopular (oil pipelines etc). The financials for nuclear reactors are not great (it's a big, financially high-risk investment that takes decades to really pay off), so there's less incentive.

Depends so much on siting too.. a lot of assumptions on foundations and water handling are baked into plans and even if the reactor room is the same size, you’re going to need an engineering firm comfortable with signing off on a nuke plant run the math based on the new soil composition and 500-year storm rainfalls, etc.

Like any engineering activity it'll depend on how much knowledge of the design the engineering team has. If it's not the same engineering and regulatory teams - then there will be a lot of risk in certifying the plant, little reward, and I'd expect every design decision to be expensively questioned.

Not to mention the hundreds of small decisions related to "X part supplier went out of business. Y part is similar with a slightly different alloy and mechanical properties, is it a suitable replacement? how can we verify this?"

The original design teams would have auxiliary artifacts that were vetted, and tribal knowledge to help quickly answer these questions. Subsequent design teams will not. The timescales between plant construction exceeds most engineer's memory at 5-10 years.

Except that when a safety problem is found with a plant design, you can't just build more from the same blueprint, you have to both refit the older plants with the mitigation and redesign subsequent plants with the fix, so now you're building a new design.

That's the reason for small modular reactors, built in factories.

It's not looking so hot for SMRs right now though. NuScale, one of the most promising SMR companies, is losing backers for its first big reactor build. This is a result of cost under-estimates and delays:

https://www.sciencemag.org/news/2020/11/several-us-utilities...

> announced that completion of the project would be delayed by 3 years to 2030. It also estimates the cost would climb from $4.2 billion to $6.1 billion.

Unfortunately it has become a pattern now for the nuclear industry to promise that the next tech will suddenly make nuclear energy cheap and fast to build. They consistently have failed to deliver on their promises -- the fiasco of the AP1000 reactor build at Vogtle in Georgia is a recent example.

It seems like the problem of the nuclear industry is the industry, not the technology itself. They're too used to relying on fat taxpayer subsidies, and are not well equipped to compete against other energy sources on the free market.

The crew is only a small part of the price. The first plant has a lot of costs which shouldn't impact as much the next ones: r&d which gets amortized on more and more units, sourcing parts and certifying providers, eventual blueprint adjustments at building, certification and safety assessments.

This is also true of aircraft carriers. The head of Newport News Shipbuilding and Drydock once told Congress that if they'd order two at once, instead of one at a time, he'd throw in a third carrier for free.

Well, did they take him up on the offer? lol

No.

Congress, collectively, is not often referred to as "smart". If I'm a senator and a guy sitting in front of me makes that offer, I take him up on it and hold his feet to the fire. How do you explain that to your constituents? I was offered free, and I said no. Re-elect Me!!!

To be fair, the building of the actual carrier is just part of the price of a carrier. What's the cost to also fit that carrier out with all of the various aircraft required to make the carrier worth having? Didn't we see an example of this when the NSA donated Hubble equivalent satellites to NASA, but NASA had to politely say no since they had no budget to operate them?

>What's the cost to also fit that carrier out with all of the various aircraft required to make the carrier worth having?

Right idea! The answer is "hardly anything, relative to crewing it for its useful lifetime", but you're barking up the correct tree.

DOD budget is about $350K per employee. A typical E5 (sergeant or PO2 in the navy) makes $3K/mo, or 1/10 of that.

You could always sell the bare-bones carrier to an ally, the way Russia sold their extra carrier to China.

> Didn't we see an example of this when the NSA donated Hubble equivalent satellites to NASA, but NASA had to politely say no since they had no budget to operate them?

It was the NRO not the NSA and NASA did accept the donation. They announced the mission schedule for one of the satellites in 2016 [1]

[1] https://spaceflightnow.com/2016/02/18/nasa-moves-forward-wit...

Details https://news.usni.org/2018/04/16/newport-news-save-1-6-billi...

I don't see any "throw a third carrier in for free" mentioned in that article.

The link says the savings from Newport would be $1.6B.

It also quotes the Navy acquisitions chief pointing out "about a third of the cost of a carrier comes from government-furnished equipment that the Navy would contract for separately" and that he previously said the overall savings of ordering two at once would be around $2.5B after considering similar efficiencies of "scale" from those vendors.

Considering the cost[1] of these Ford-class carriers is around $13B this represents a savings of about 10% overall, at least according to my crude, back-of-the-napkin math [ 2.5B/(2x13B) or 1.6B/(2/3x2x13B) ].

So it might be closer to "buy 10, get one free".

[1] https://en.m.wikipedia.org/wiki/Gerald_R._Ford-class_aircraf...

Please don't give them any ideas...

Ha easy thing to say. And when he can't deliver on the promise and goes under who has to sweep in to save all those jobs?

The Navy has ordered two at a time, and they definitely didn't get a third one free. They did save about 15%, though.

https://news.usni.org/2019/01/31/navy-awards-2-carrier-contr...

That comment wasn't about this round, it was from the middle days of the Nimitz class, which was produced one at a time for almost 35 years, with occasional periods of no production.

I’ve been involved with large scale nuclear construction projects of different designs in multiple different countries and different regulatory schemes. I have a hard time squaring what I’ve seen on the ground with the summary of this study.

To respond to your exact point, big parts of the nuclear-qualified workforce in most places really do move from large job to large job (assuming relatively static demand), so you do get that transfer of knowledge on sequential jobs.

Also, don’t forget the design and manufacturing that happens off-site, which can be significant and often is a huge driver of risk and cost. (10 MW vertical motors! Big safety-related pumps! Mega-sized forged components! Specialized custom fabrications!) The back office engineers and subcontractors and factory floor people who make them remain pretty static as long as there is work to be done. By the Nth unit, there is know-how and known problems are worked out; some of this is translatable into drawing updates and schedule resequencing but really a lot of it is expertise that stays in peoples’ heads. Once everybody is demobed and scatters, that is all lost.

I'm not familiar with electric power infrastructure. Could we build the plants in middle America and transmit the power long distance?

I assume we'd lose power during the transformer steps, but would it preclude building the plants away from people and groundwater reservoirs?

Why do you assume "middle" America has no groundwater reservoirs to be concerned about? Also, sure, the population may be sparse in case of accident, but there goes your food supply. Also, "middle" America has another common name used to describe it: Tornado Alley. Weather plays a large part of site decisions.

To the first question: Yes. See eg https://en.wikipedia.org/wiki/List_of_HVDC_projects

It's not as attractive for managing nuclear accident risk as you might think, as nuclear is already struggling with cost competitivenss even without HVDC lines, and risk reduction per buck of other safety design features are better.

But it's brilliant for eg hydro (for obvious reasons), wind power (because 2000 km away it'll be windy when you have local lulls and vice versa) and solar (2 time zones worth of distance balance out production/consumption peaks nicely).

No you can't, there are too many losses in power lines over that kind of distances. Ideally they are in 0 to 100 miles of where the power is used.

Sounds like we just have to build even more plants

I have no experience in nuclear but have worked on infrastructure projects that have restarted. I think people have a new set of biases that are not present in the first project. They are optimistic which is a risky state of mind.

Also...

* Often the starting point is a poorly organized mess that needs to be unpicked and understood.

* People conduct "reviews" of existing work that consumes budget and leads to new issues being found.

* You are continuing an existing chain of communication, but with a loss of continuity. The people you need to ask can't remember, or are not available to ask.

* Overly optimistic budgets lead to corner cutting and mistakes.

That's why you don't build two plants you build 20 or 30. France's nuclear project built 34 and 20 plants of 900 MWe and 1300 MWe classes of plants respectively which make up the bulk of its reactor fleet: https://en.wikipedia.org/wiki/Nuclear_power_in_France#Techni...

The Messmer plan (France's big reactor build) was enacted 40-50 years ago. The energy market looked vastly different then. It would not work in 2020. In fact, France is looking to REDUCE their dependence on nuclear energy now. Source: https://www.reuters.com/article/france-electricity-solarpowe...

> France aims to rapidly develop renewable wind, solar and biomass capacity to curb its dependence on atomic power, reducing its share in its power mix to 50 percent by 2035, from 75 percent today.

Plans they continue to delay, because there's no feasible alternative to geographically independent and dispatchable energy save for fossil fuels: https://world-nuclear-news.org/Articles/French-bill-delays-n...

You realize that World Nuclear News is not an unbiased source, right?

The UK is doing fine with a renewables-heavy powergrid, as are Spain and Portugal. 40% of their electricity comes from this, and the share is rising steadily.

The above source is quoting a bill written by the French government:

> However, the bill also calls for "realistic goals to transform our energy model by increasing the timeframe for reducing nuclear power to 50% by 2035 instead of 2025, which would have required the construction of new gas-fired plants, and would have involved an increase in our greenhouse gas emissions."

I'm sure World Nuclear News is more favorable to nuclear power, but I don't see how that is relevant. Do you believe that this article fabricated the contents of the French bill?

This comes off more as a cheap jab than a substantive claim of bias.

Nothing in there actually refutes my point that France wants to REDUCE its dependence on nuclear energy overall. France is the poster-child for a nuclear-focused powergrid. If they want to move away from nuclear, that suggests it is not working out as well as people claim.

Legislation gets written and rewritten as timelines get tweaked. That's a reality. The core goal is unchanged.

The point is, they make lip service towards moving away from nuclear but don't actually make any strides towards doing so. The claim that France wants to reduce its share of nuclear would hold more water once they actually start doing so. Talk is cheap, actions are not.

They closed FESSENHEIM-1 and FESSENHEIM-2 this year

It seems like you're implying "nuclear = bad because France is moving away from it" which is not what France is doing.

> which is not what France is doing.

They literally are, and have formally said so. They're not shutting down all their reactors right away or anything. But for a country that took pride in betting big on nuclear, officially planning to go from 70% nuclear to 50% is a pretty huge drop. They've already shut down the first two reactors at Fessenheim as part of this.

Yes, they're kicking the can down the road slightly at last news. That's not surprising since many of the reactors may be possible to life-extend to 50 years, and renewable energy costs are dropping quickly. Delaying a few years saves some money. But that does not change the fact that many of the reactors will not be replaced when they hit end-of-life.

>> you're implying "nuclear = bad because France is moving away from it" which is not what France is doing.

> They literally are, and have formally said so

Citation needed.

As I stated in another reply to you, reducing dependence on nuclear has a lot of advantages unrelated to "nuclear is bad" or "nuclear is not clean" or etc. Like I said before, no matter what your energy source is, 75% of your power being produced from a singular source is not a good idea. You may be familiar with the old saying "don't put all your eggs in one basket." This is the same reason we don't invest in a single stock. Dropping to 50% is far from getting rid of nuclear and it is an absurd claim.

> But that does not change the fact that many of the reactors will not be replaced when they hit end-of-life.

Many, but many are planed to be replaced. You're focusing too much on the reduction aspect. Reduction is not elimination. France has no plans on eliminating nuclear in the foreseeable future. The only way this would happen is if ITER was a major success and small versions could reliably be developed and deployed. I'm not counting on that.

> Citation needed.

I already provided a citation for this - https://www.reuters.com/article/france-electricity-solarpowe...)

Quote

>> France aims to rapidly develop renewable wind, solar and biomass capacity to curb its dependence on atomic power, reducing its share in its power mix to 50 percent by 2035, from 75 percent today.

A quick google will turn up dozens of other sources discussing that if you like.

> reducing dependence on nuclear has a lot of advantages unrelated to "nuclear is bad" or "nuclear is not clean" or etc

When did I even say that? Please have the decency NOT to convert my educated and reasoned points into a straw-man argument. I spent years working in nuclear physics. I'm not making some idiotic "nuclear reactors will turn us into glowing mutants" argument here.

The real issues with nuclear are -- and always have been -- simple, practical problems of cost, time to construct, and the usual problems of cost overruns and delays that you see with constructing a very complicated system. The Flamanville EPR has been a fiasco, with costs triple its original estimate and a timeline that ballooned to 15 years.

In comparison, the competition from renewable energy has heated up rapidly. Economies of scale are rapidly driving down the costs renewables and batteries (and the underlying technology is improving rapidly). Between 2010 to 2019 wind energy become 70% cheaper and solar became 89% cheaper: https://www.lazard.com/media/451082/lcoe-8.png

Battery costs have dropped 75% over the last 6 years: https://www.greentechmedia.com/articles/read/report-levelize...

Nuclear energy in 2020 is often quite expensive compared to renewable energy, and the comparison is only getting worse over time.

> France has no plans on eliminating nuclear in the foreseeable future.

In 2020, Energy Minister Élisabeth Borne announced the government would not decide on the construction of any new reactors until Flamanville 3 starts operation after 2022. Citation: https://www.neimagazine.com/news/newsfrance-to-decide-on-new...

"France has undertaken to cut the nuclear share in its electricity mix from around 75% to 50% by 2035 while increasing the share of renewable solar, wind and biomass."

“Whether we are looking at 100% renewables or a percentage of new nuclear, we want to consider all the elements, including technical, economic,” she noted. “On such important subjects, we must make rational, reasoned decisions, and that is the objective of the various studies that have been launched.”

To put that all together: France is cutting their use of nuclear, and waiting to decide whether or not to replace reactors that are fast approaching end-of-life. They're going to look and see how the energy market shakes out.

>>>> you're implying "nuclear = bad because France is moving away from it" which is not what France is doing.

>>> They literally are, and have formally said so

>> Citation needed.

>> reducing dependence on nuclear has a lot of advantages unrelated to "nuclear is bad" or "nuclear is not clean" or etc

> When did I even say that?

Wow this conversation is a mess.

This "France is looking to reduce their dependence on nuclear energy" statement can get dicey because of different contexts that it is used. So I apologize if I am responding to a context you aren't implying. It is often used as "look, even France doesn't want nuclear!" which frankly just isn't true.

We have to look at France and how much carbon it produces. Right now it is one of the lowest producers in Europe. Let's look at the electricity map[0]. Sweden, Norway, and France are leagues ahead of others in terms of carbon emissions. France's plan is first to replace existing natural gas, biomass, and coal with renewable resources. The second part of the equation is that their reactors are reaching EOL, so do you build more or replace them? If you pay attention to energy trends solar and wind (something France has an abundance of) is getting much cheaper and battery storage is getting cheaper (France doesn't have to bet as much on battery storage since they can over produce and sell excess energy, which they currently do a fair amount of). So if you're going to take bets this is still a good bet. A big part of a good and stable power grid is by having a diversification of energy resources. 75% of your energy being dependent upon one resource is not a good idea. No matter the resource. Even 50% is high, but acceptable. They aren't planning from going away from reactors, there's even one under development. But you also want to hedge your bets. If any of these factors (solar, wind, battery storage, smart grids, ITER, etc) don't pay off, then they need to maintain their nuclear grid. It would take a large revolution in energy development for France to be able to still produce so little carbon and provide its citizens with a modernized (electrified) country.

Also consider that France doesn't have good access to hydro like Norway and Sweden do so its options for clean energy are nuclear, solar, and wind (lots!). They should, and are planning on, using a diversification of these. Nuclear provides a strong baseload and the others supplement. You may notice that this is a key argument made by many proponents of nuclear. Anyone that says the grid should be entirely nuclear is an armchair scientist who understands very little about nuclear or the climate. But the same is true for those that think we can solve the issue with just solar and wind.

So if you're saying France is turning away from nuclear, then this is adding desires into a plan that does not express or concern itself with those desires. A big part of this decision is about diversification and increasing energy independence (just like recycling fuel is a big part of their energy independence, which they power a whopping 17% of their grid with recycled nuclear alone).

[0] https://www.electricitymap.org/zone/FR

If it worked then it could work now. It's only cowardly politics that stop newer, safer reactors from being built.

The Callaway plant cost ratepayers so much that a state ballot proposition outlawing that accounting trick passed with 66% of the vote. The last time Ameren tried to remove that protection was 2015. Even they've noticed how cheap renewables are now.

I guess I have to read the full paper, because reading the link does nothing to clear up what is going on.

From reading the link I get, change orders cost money (duh!) and that there are a lot of change orders, particularly to older designs due to "changing regulation". Cause presumably before starting construction the older design has been mitigated to deal with newer regulations.

Which from my own biases, seems to mean that courts/etc are stepping in and modifying the regulations during construction and that costs a sh*tload. Otherwise the builders are negligent, or simply that it took 30 years to get approval for the plans, but now its a catch 22 because they are outdated.

All large systems are, in a sense, obsolete as soon as construction begins. The longer construction takes, the more out-of-date the finished product. And then there are the decades of operation.

Consider, for example, that one year into construction of a new power plant a completely unrelated investigation into fires concludes with the finding that a certain kind of electrical wire, when installed incorrectly, can short and cause a fire. The power plant has already installed miles of the exact same wire. The wire manufacturer has a new product that prevents incorrect installation, but the plant designs call for the old wire. What to do?

* Pause construction while they adjust the plant designs with the new wiring? * Continue construction, but inspect the existing wiring, and reinstall if necessary? * Leave the wiring in place, but install a fire-remediation solution for the already-installed wiring?

All large systems have to deal with these kinds of changes. (It’s interesting to look at the changes in a “class” of ships over them as they build each one, for example.)

Is there a field of study/practice that deals with such changes to large scale system? It does sound like a very useful thing to systematically study, if possible.

But those problems exist for large coal/etc plants as well, so some of that is baked into the equation and doesn't do much to explain why nuke plants end up paying 10x as much for a plant that is only slightly more complex.

My take: one significant issue is that nuclear power is the only large scale electricity source with meaningful, consistently enforced regulations.

There's no equivalent of the NRC setting welding standards for the steam systems of coal plants and demanding that control systems are guarded against wiring fires. General safety and environmental regulators are often deliberately underfunded by business-friendly legislatures, and even violations that they catch are rarely considered severe enough to shut down a site until remediation. And of course nuclear reactors are (rightfully!) expected to contain their waste products, while combustion plants still get a free pass to dump CO2 into the atmosphere.

The result is that very few industrial sectors in the US have experience receiving meticulous government oversight that they cannot ignore. Aerospace [1], pharmaceutical manufacturing, and nuclear power are about the only ones I can think of. If contractors don't have institutional knowledge of actually working on nuclear projects (which are already more complex than a big coal or gas project), they may not be prepared for regulators that actually check the work and demand it adhere to documented standards. Just about every other project can let marginal welds or inconsistent paperwork slide with "eh, you tried." The really expensive part about nuclear projects isn't doing the work. It's redoing the work that you failed to do up to standards the first time.

But you absolutely do not want to make the NRC more lax to bring it in line with the barely-there standards the rest of industry is used to ignoring. You don't want the nuclear equivalent of the West Fertilizer Company -- inspected twice in 28 years, minor violations ignored until the plant exploded [2].

[1] You might even put an asterisk on "aerospace" considering the 737 MAX debacle.

[2] https://en.wikipedia.org/wiki/West_Fertilizer_Company_explos...

Exactly, and when a reactor takes 10 years to build it's almost guaranteed you'll discover some issues after construction begins.

The article highlights that...

> Among the surprising findings was that contrary to expectations building subsequent plants based on an existing design actually costs more, not less, than building the initial plant.

Given your example wouldn't that mean replacing the wiring of all previous plants of the same design? I suppose this is one way cost overruns get compounded - an issue discovered during the current build would likely apply to the previous builds. (any actual examples of this?)

> building subsequent plants based on an existing design actually costs more, not less, than building the initial plant.

> if more components of the plant, or even the entire plant, could be built offsite under controlled factory conditions, such extra costs could be substantially cut.

Wait what? France builds with relative similarity and it's cheaper. The problem is with US regulation, politics and CYA hand-wringing, imo.

The French industry isn’t doing too well either; all the EPRs are also late and over budget and they took over Areva to bail it out.

not true

Taishan units are up and running

Open doesn't mean it was on time or on budget.

https://archive.is/MT9fA#selection-2255.0-2337.266

> CGN Power, the listed unit of state-owned China General Nuclear Power, China’s largest nuclear power projects developer, has announced a 17 per cent capital increase at its Taishan nuclear project, raising concern cost overruns and commissioning delays will continue.

> The company will inject 2.94 billion yuan into its 51 per cent-held unit Taishan Nuclear Power Joint Venture, which will see the unit’s total registered capital to 28.6 billion yuan from 24.4 billion yuan, the company said in a filing to Hong Kong’s bourse late on Tuesday. Provincial government-controlled Guangdong Yudean, the province’s largest power producer, owns 19 per cent stake, and CGN’s French technology partner EDF holds 30 per cent stake.

> CGN said the capital injection serves “to enhance its financing ability so as to meet its fund requirements for engineering construction.”

> The project, which CGN said on track to become the world’s first third-generation nuclear reactor built on the European pressurised reactors (EPR) design to achieve commercial operation, was originally expected to have its first generating unit come online in December 2013 followed by the second in October 2014.

> The timetable of the two units subsequently slipped to the first and second half of this year, and further to the first and second half of next year after “comprehensive evaluation on subsequent engineering construction plan and relevant risks,” said CGN a year ago.

Sure

but we're talking about 20% overrun, which I could call a good upfront guess

Flamanville, France:

The project was planned to involve around €3.3 billion of capital expenditure from EDF, but latest cost estimates (from 2019) are at €12.4 billion. Pierre Moscovici, president of the Court of Audit, gave a statement on 9 July 2020 concerning the release of the report on the delay costs of the Flamanville 3. The report of the Court of Audit reveals that the cost of Flamanville 3 could reach €19.1 billion when taking in account the additional charges due to the delay in construction.

https://en.wikipedia.org/wiki/EPR_(nuclear_reactor)#Flamanvi...

This one is a bit different though, it's the first of a kind of a new design (so count a lot of R&D there) and the original schedule was very unrealistic, not sure why anybody agreed on a timescale so short for a new design in the first place.

France hasn't built any reactors of the old design in what decades?

Worth noting that Unit 3 at Olkiluoto in Finland was started before Flamanville and it hit massive cost overruns and delays as well. It was another EPR -- it's not like Flamanville was the very first build of this design, they had another to learn with.

You should also add that China started to build two EPRs after France and Finnland and they are both already operating commercially.

They were both also late and over budget though. Less so than the two European units, but still.

https://archive.is/MT9fA#selection-2255.0-2337.266

> CGN said the capital injection serves “to enhance its financing ability so as to meet its fund requirements for engineering construction.”

China's nuclear industry and the way the government interacts with it is just... very different from pretty much any other country. This is true of some of their other sectors. Yes, they can do some things incredibly quickly by more or less ramming them through at top speed using a combination of government and state-affiliated corporations. But the ways they achieve those results at speed are not necessarily transferable to other countries.

I also get a bit nervous about how fast they've scaled their nuclear sector and pumped out plants. It reminds me a bit of the big historical push the USSR did into nuclear tech -- which culminated in the Chernobyl incident because they overlooked issues they should not have in order to keep things moving fast.

A lot of technology and engineering lessons have been learned since those days... and so far I haven't heard anything specifically concerning about China's nuclear reactor fleet. But there's always going to be a little doubt in the back of my mind.

Flamanville is one of the first EPRs built by France. The bulk of its nuclear fleet were built on serial production of the same design. 34 900 MWe reactors and 20 1300 MWe reactors make up the bulk of its generation: https://en.wikipedia.org/wiki/Nuclear_power_in_France#Techni...

Unit 3 at Olkiluoto in Finland was started before Flamanville and it hit massive cost overruns and delays as well. It was another EPR -- it's not like Flamanville was the very first build of this design, they had another to learn with.

Yes, and two EPRs up and running in Taishan, China, that were started after the two you mentioned.

don't forget the NIMBYs

I don’t know how anyone can look at the state of corporate America and not be a NIMBY when it comes to a new nuclear plant... which company exists right now that you would trust to run something correctly for the next 50 years and resist the pressure to cut costs as close to the bone as possible..

Why is this downvoted? I wouldn't trust a nuclear power plant run by a private company.

The Navy. Semi-serious; they’ve been operating nuclear reactors safely for decades.

I was skeptical, but it's true https://www.politifact.com/factchecks/2008/jun/09/john-mccai...

I trust airlines with my life without even thinking about it. Sometimes we even make fun of people who are afraid of flying.

Incentives make a big difference here. Ignoring everything else, airlines don't want to lose their hundred million dollar machine in an accident. And the person most directly responsible for the safety of the flight -- the pilot -- is sitting right there in the machine with the rest of us. It's a good arrangement with a long track record.

I could imagine corporate-run nuclear plants that have the same incentives as airlines. I could also imagine bad ones, run by Homer Simpson. It probably makes more sense to judge case by case.

Airlines in less developed countries are certainly not as safe as they are in the West.

Airlines are also different from power utilities, in that it's very easy to change airlines and not very easy to change grid operators. After Fukushima it's not like consumers could switch to a power utility other than TEPCO.

All good points!

Yeah, I had PG&E in mind too when I said "How could you trust anyone in corporate America" -- look at their record even before their equipment burned down 10 million acres in California with the San Bruno explosion since they decided to ignore their own engineers' recommendations on maintenance intervals.

There's a story (probably highly exagerated) I heard about a nuclear power plant being built in china a while back. During construction, flaws were found and it was discovered that inspectors had been bribed to look the other way, etc. The gov ordered that all plant construction executives and crew bosses and their families had to live next to the power plant. Lots more problems were suddenly found but soon it was built and operating safely.

I think people have a lot of suspicion and disdain for corporate ownership of infrastructure. People abhor the windmills of BigEnerCorp that "disrupt" their view, but the same sort of folks in the next town over love their energy cooperative and the ROI of their town-owned windmills or solar farm.

Counter-example: the Boeing 737 MAX

"I could imagine corporate-run nuclear plants that have the same incentives as airlines."

So how exactly do you enforce that corporate executives and their families must live on site for decades?

recently, boeing appears to be doing all it can to break the trust it has built up over decades.

Exactly. Watts Barr in Tennessee just got a big fine from the NRC for safety violations, after trying to lie about an incident to the NRC and firing a whistleblower who reported safety issues.

https://news.yahoo.com/federal-utility-fined-900k-nuclear-21...

> In a notice dated Nov. 6, regulators noted a “substantial safety culture issue” at Watts Bar at the time of the incident. They also found that “TVA senior management and staff failed to communicate with candor, clarity, and integrity during several interactions with the NRC during the course of the inspection and investigation.”

> The incident wasn't recorded in the plant's logbook and managers later misled NRC investigators about what had happened. The shift manager told investigators he wasn't truthful with them at first because he feared that whatever he said would be relayed back to management.

> last year, TVA was ordered to rehire and pay thousands of dollars in back pay and compensation to a whistleblower who raised concerns about nuclear safety.

And then there was the nuclear corruption scandal in South Korea: https://www.technologyreview.com/2019/04/22/136020/how-greed...

And let's not forget when a nuclear company literally bribed Ohio politicians with $60M to get a $1Bn+ bailout:

https://www.cleveland.com/open/2020/07/fbi-agents-deployed-t...

The NIMBY factor constrains where new projects get permits. It doesn't explain why permitted projects that were welcome in local communities still blew out their construction schedules and budgets. The AP1000 fiascoes at VC Summer and Vogtle weren't caused by NIMBY. Nor did NIMBY cause the problems with new European builds of EPRs.

"The problem is with US regulation, politics and CYA hand-wringing"

The article specifically addresses your cynicism about "regulation and politics", only they call it "safety regulations".

They found it not to be the deciding factor. Lack of construction outside the US would seem to support that conclusion or, alternatively, to accept those factors as unchanging, sort-of like the laws of nature, and to blaming democracy for all ills of society.

The principal author, Philip Eash-Gates, presented his master's thesis last year: Modeling barriers to cost change in solar and nuclear energy technologies [1]

[1]: https://dspace.mit.edu/handle/1721.1/122160

Even if they can reduce costs a bit, the economics are still not in favor of nuclear energy, and renewable energy has become incredibly cheap. Between 2010 to 2019 wind energy become 70% cheaper and solar became 89% cheaper[1] -- and they're still getting cheaper. We are now in a situation where we can build 3x as much renewables for the same price as nuclear[2] - nuclear has a serious cost problem. I'm very skeptical of any claimed nuclear energy cost reductions because time and again the industry has promised lower costs and failed to deliver; for example the new AP1000 and EPR reactors in the US and Europe have all vastly overrun their budgets and run years behind schedule.

1. https://www.lazard.com/media/451082/lcoe-8.png 2. https://www.lazard.com/media/451081/lcoe-2.png

I am very hopeful about renewable but I am also aware of their intrinsic weakness: They can't be provide base power (unless we build battery farms able to cover our needs).

Renewables are exciting. Jumping too fast into it like Germany did (and is now polluting as much as 8X France with 450g CO2 / kWh) is much less exciting.

I wish there was a bit more expert involvement in the way we choose our energy policies (and much less tribalism and populism).

source: https://www.electricitymap.org/zone/DE

Wind is a pretty solid source of baseload actually, as long as you build enough. The supplementary materials from Caldeira's Geophysical Constraints paper (usually used to argue AGAINST renewable energy) show that with 50/50 wind/solar mixes (figure S4) you can achieve:

* 1x capacity, 0 storage: 74% of electricity demand

* 1.5x capacity, 0 storage: 86% of electricity demand

* 1x capacity, 12h storage: 90% of electricity demand

* 1.5x capacity, 12h storage: 99.6% of electricity demand

Citation: http://www.rsc.org/suppdata/c7/ee/c7ee03029k/c7ee03029k1.pdf

> ElectricityMap

Nobody who follows the energy sector closely thinks ElectricityMap has any credibility for country-to-country comparisons. The datasets are extremely fragmentary and have huge yawning gaps with no data available, which should be the first red flag for anybody citing it. It might be useful for trends within a given country, but not the way you're citing it.

Also: accounts popping up out of the woodwork to argue passionately for an out-of-favor technology reeks of a dying industry trying to revive itself with public relations. 3/4 of those accounts seem to cite ElectricityMap, oddly enough...

The reason renewables are the cheapest source of energy today is at least in part because Germany made a big investment early on.

Ensuring that there is demand allows investment to be made in production, which allows competition and learning, and acceleration of the decrease in cost.

Batteries are halving in cost every handful of years because there is a guaranteed market that allows construction of more production facilities.

Batteries are currently being deployed very cost-effective for replacement of gas peaker plants and for frequency regulation. They are also being deployed on the grid as "non-wires-alternatives" to transmission upgrades. As we transition to a carbon-free grid, they will find even more use.

A huge fraction of planned solar farms in the US have storage built in, because of the efficiency of reusing the same DC to AC inverters, and because for quite a while now, panels have been cheap enough that some panel generation capacity is thrown away in order to get more output at other times of the day.

It is a mistake to think of batteries as needing some sort of technological leap to serve our needs. If we had to, we could deploy them at current prices and build a renewable, carbon-free grid more cheaply than we could with nuclear. But we are only installing them as necessary as we replace aging infrastructure, rather than shutting down existing infrastructure that hasn't worn down.

As the cost of storage+solar drops below the fuel+operations expense of natural gas, we will start shutting down natural gas plants before their natural end of life, resulting in wasted costs. I have a feeling that any naturals gas turbine installed today will be considered a boondoggle within 5-10 years.

The future is now, when it comes to storage, we just haven't had time for reporting to catch up.

I appreciate that someone has "done the math" on the soft costs for nuclear plants. One of the arguments against nuclear has always been "but it's so expensive" and yet the engineering of the plant itself is not particularly expensive relative to similar capacity coal or gas plants[1]. When we were discussing the "Ultra Safe Nuclear" (www.usnc.com) [2] I commented [3] that expense was more soft costs than engineering from my own experience of looking into the cost of nuclear. This paper just does the math and makes it more explicit.

[1] Yes, that is multiple plants to create the equivalent amount of capacity.

[2] Ultra Safe Nuclear -- https://news.ycombinator.com/item?id=24505727

[3] https://news.ycombinator.com/item?id=24511751

I feel like large multi-year construction costs have a tendency to increase significantly (outside of the normal increases in year-of-expenditure costs), and I wonder if this is because construction firms know something needs to be built and will behave accordingly.

This is in contrast to smaller projects, especially during recessions, where construction firms may be more likely to need the work (lower bids) and also don't have the expectation that the project will have to be completed (they can wait it out to extract more money from a contract).

Not quite this scale, but talking to people on large projects, I have been told that to bid what it's actually going to cost simply guarantees someone else gets the job. To stay in business, knowing how much to underbid is a crucial piece of unwritten knowledge.

I'm sure it's hard to shift these norms. As the buyer, it's hard to credibly threaten a hard limit. Everyone knows that, in 10 years time with an 80% done plant, and a construction company in bankruptcy proceedings, of course more money will be found.

Yes, nuclear power projects bid below what they know it will cost and then pull a shocked-pikachu face when the final bill ends up several times the original "cost estimate".

Look at Flamanville in France: budget triple the "estimate", 15 years to construct. France has more nuclear energy per-capita than any other country on Earth, so it's not like they don't know how to "do" nuclear.

Right, but it's hopeless to pose this as a moral question, and hope that businessmen & politicians will suddenly start behaving with boy-scout honor.

Those who decline to play by the rules we have ended up with, don't get the contract & go out of business, or don't get elected. It seems like an incentive problem. And a possibly unsolvable one, for huge mega-projects like this -- constructed over decades, dangerous & high-tech, and completely useless until 100% completed.

Wind energy seems much more healthy in this regard. You buy off-the-shelf units, from a factory. If you don't like the first 10 they install, you don't buy the other 90, and still get 10% the power. It's much closer to buying cars, or pencils, which our society is pretty good at optimizing. Every year we improve the process, and these steps add up.

Yeah, and there's not enough viable vendors in the nuclear sector to be able to just switch if someone does not play by the rules.

You are correct that wind has a lot more flexibility in this area than nuclear. It is also much easier to scale a project up by just adding more turbines. In the case of older, smaller turbines they can be replaced with more modern turbines that yield more power (called re-powering).

In general wind power and solar have much better economies of scale too -- the more you build, the cheaper it gets. Nuclear doesn't tend to follow that pattern -- although there are attempts to create it via small modular reactors (SMRS). The jury is out if that will succeed or not (early signs are not great though).

Yes. I wonder whether a push for something like small modular reactors, starting in say 1973, might have got us a real solution. Seems a bit late now, but would love to be proved wrong on that.

Small reactors were part of the original 1st gen nuclear reactors. The world went to bigger reactors in the 60s-70s because it ended up being much cheaper than building lots of small ones. Reactor fabrication is complex, and it's easier to just build them bigger.

From the 90s to the mid-2000s there was probably a window for development in this area. But the nuclear sector stagnated as a whole, and then fracking made natural gas cheap enough that people wouldn't consider it.

The jury is out on SMRs still. They're promising a lot, and the tech and engineering has improved. But plunging costs of renewables mean the window of opportunity for nuclear tech is probably mostly closed in the West. This article from a highly respected energy sector analysis group is worth a read: https://about.bnef.com/blog/scale-up-of-solar-and-wind-puts-...

India and China are the big markets for nuclear right now -- their energy demands continue to grow, and they're building reactors as part of an all-of-the-above model. France MIGHT be a potential buyer of SMRs (their reactor fleet is getting close to end-of-life), but their nuclear companies have put a lot of push behind the EPR design.

>Is there a field of study/practice that deals with such changes to large scale system? It does sound like a very useful thing to systematically study, if possible.

"Analysis points to ways engineering strategies could be reimagined to minimize delays and other unanticipated expenses."

(from comments below).

The two quoted authors in the press release are a prof in energy studies and a nuclear engineer.

What I notice is that MIT does not have a Systems Engineering undergraduate department, but numerous specialties.

In a large design/build project such as a nuclear power plant, the systems engineering group (there must be exactly one) keeps track of the performance and function of each of the subsystems (civil works like containment, basements, buildings; electrical; controls; HVAC; the nuclear bits; and so on). In addition, it is the system engineering group that responds (or directs responses) to a change order request, and the overall impact on expense, on functionality/reliability/safety, and schedule.

It seems these are responsibilities that are not identified here with a known role; instead the authors reinvent system engineering for themselves.

There is a lot of SE work done in numerous industries (Elon Musk is what I would call a systems engineer, based on how he identifies things to do and how he gets them done. His degree is not in SE though).

A short list of US universities offer SE or related disciplines; in California/New York/Illinois. MIT is not one: it has a research group not a degree-granting program.

References:

https://www.incose.org/ The International Council of System Engineering. Has published a handbook in numerous editions over the years.

https://www.nasa.gov/connect/ebooks/nasa-systems-engineering... First published 1995, based on other reports published in the 1960s.

MILSTD 499, 1970s, very dated by today's standards

Nuclear power plants are insanely complex. It’s just a massive water boiler. The problem is that all that steam and water is radioactive. So you have to design in all these reinforcements to ensure that the steam and liquids do not leak out.

It’s easy to say that your design is airtight, but the reality is far different, with the laws of entropy taking over after decades of operation. This is when your parts starts to fail. Seals starts to leak. Things just wear out and fall apart.

This is why the thing is so complex.

Then, you have to deal with the radioactive waste byproduct which lasts for 10,000+ years. How do you even deal with this, when no government in human history has lasted longer than a few hundred years.

As much as nuclear is a necessary evil to maintain baseload, since solar and wind doesn’t always function, I think we need a new energy paradigm.

I’m thinking we need to go with solar space power generation, and beam the energy down via microwave or laser. Maybe laser is a good candidate. Beam the laser down, and collect it with photovoltaics. And the laser can run 24/7, even at night.

Asking a friend: Maybe someone can run the efficiency numbers on this. But even if it’s less efficient than nuclear, this is probably far safer than nuclear ever will be.

Does water itself become activated? Or the problem is that something leaks into the water even during normal operating? Plus everything is designed to maintain containment, which makes the coolant loops a lot more complicated, right?

Putting a giant laser in space doesn't seems _that_ safe.

The proposed space based energy things are very low efficiency. After all, you can collect enough solar power on Earth already, the problem is efficient distribution. (And if you put a big laser/maser in space that still would not be able to provide baseload if it's not always in line of sight.)

And launch costs are prohibitively expensive right now. First we'd need to bootstrap a cislunar economy (eg mine and manufacture stuff in space, so we don't need to pay the launch costs from deep within Earth's gravity well): https://www.youtube.com/watch?v=gOr-Gd58zu8

Where is the Elon Musk of Nuclear Reactors? Where are the robotic assembly lines that crank out highly modular, efficient, miniature "assemble-yourself" nuclear power plants that any construction team could snap together and seat into a preformed foundation in 48 hours? Why doesn't this exist yet? I'm talking on the scale of 1 reactor to 1 neighborhood or 1 large business. Perhaps geothermal cooling.

NIMBYism, safety, and non-proliferation concerns.

> non-proliferation concerns

We're already entering into runaway proliferation.

Pakistan and India are our allies, but I don't see them keeping a stranglehold on the tech. Iran will have them. The Saudis will probably have them soon.

If a dysfunctional nation like North Korea can have nukes, what's stopping the rest of the world?

We're not even preventing ICBM capability from spreading!

What gives? This is way more of a nightmare scenario than climate change and nobody is batting an eye.

> This is way more of a nightmare scenario than climate change and nobody is batting an eye.

there is a probability that an irrational leader orders a nuclear attack and somebody executes it instead of deserting.

climate change is a certainty at this point and the only question is how bad will it be. for people living by the sea and in the parts of the tropics displacement in a single lifetime is next to certain.

The tech is 75 years old. One suspects a sufficiently-funded team of recent PhDs could build the entire infrastructure from scratch in less than a decade. The methods we've attempted to stem proliferation haven't worked. Trying them again, except harder this time, won't work either. We all know this, but few of us have the imagination to consider other options.

Thorium msr reactors almost solves the last two problems...

It does exist. But it uses nuclear fusion instead of fission. There's a big nuclear fusion plant at the center of our universe. It's called the "sun". It spreads out its energy for free. There are highly modular energy collectors you can buy that make use of that energy and they're getting cheaper at amazing rates.

You still need base load, ie. an option that can displace natgas in the way that natgas displaced coal.

There's a reason why even though solar is now cheaper than coal it didn't displace it like natgas has.

The reason is more about capex, the financial necessity of aging out existing plants and the time it takes to shift from one dominant source of power generation to another.

It wouldn't have been possible to shift to nuclear power overnight either.

Since natgas is 100% dispatchable and is 35% of power generation in the US it's not like nuclear's 20% is really contributing that much stability to the grid right now and natgas can be gradually dialed down as solar/wind/variability tech is dialed up.

> the financial necessity of aging out existing plants and the time it takes to shift from one dominant source of power generation to another.

Yes... but we just undertook the capex to do this with coal -> natgas in the last decade or so. So the barrier is not just the financial cost of overhauling.

and in the last 5 years we went from about 5% to 18% renewables.

The renewables industry in the US wasn't given any special dispensation so it had to compete pretty much without subsidy with natgas and coal which it could only really start to do about 5 years ago (it's cost competitive with coal only this year).

In fact, arguably the US subsidies oil and gas more than renewables.

Nuclear would require massive subsidies to grow beyond 20%. Renewables won't.

The green industry is subsidized in the US at both the state and federal levels.

Some. Dirt cheap solar panels from China also had tariffs slapped on them by Obama which offset much of that and killed the solar boom a few years ago.

The oil and gas industry is also subsidized at federal and state level, of course.

Last I checked we still use a lot of coal.

Solar hasn’t replaced natural gas, because they are entirely different things. It’s exceptionally easy to transport natural gas, it has a ridiculously good energy density and burns well, the logistics of natural gas require no technology in the lines that run between natural gas devices are a half pound of pressure meaning any idiot can hookup/install/modify.

There is a case to be made for hydrogen displacing natural gas, but in my opinion not until we have small local fusion reactors, and in that event you’re probably looking a battery anyway. Hydrogen has great potential for aviation, but home consumer use natural gas will be king for some time.

https://www.nytimes.com/interactive/2020/10/28/climate/how-e...

Rapidly declining, actually.

> they are entirely different things.

Uh, sure - but many of the reasons you're giving seem to be missing that I'm talking about for electricity generation, whereas you seem to be imagining in home usage?

Natural gas power plants can have 1000 psi or more in their lines.

> it has a ridiculously good energy density and burns well

Yes, good energy density - that does sound like something that might be important.

This is an irrelevant distraction. There's no reason we can't do both. And, since life on Earth is powered by the sun, everything is fusion power if you zoom out far enough.

> There's no reason we can't do both

Sure there is -- one option is a terrifically expensive and dangerous dead end.

We need to focus on figuring out the logistics for the capturing, storage, and distribution of our friendly neighborhood fusion reactor. If you look at the increasing energy density of battery technology over time[1], there is no practical application for fossil fuels or nuclear in the long run. There are startups[2][3] that are far closer to the "holy grail" of battery technology than any similar plan for nuclear based power.

The economics are already so clear that every major energy provider is moving to renewables. There is (correctly) no appetite for massive nuclear power plants in anyone's backyard. However, there is already a huge market for putting solar on top of your roof and a battery in the garage.

Tesla has proven the massive savings available to any sizeable government with the use of a truly smart grid.[4] Once the logistics are in place to recycle high performance electric vehicle batteries into infrastructure storage where charging times are less important, it's game over for non-renewables for everything but industrial applications. And there are already solutions for that.[5][6]

[1] https://www.researchgate.net/figure/History-of-development-o...

[2] https://venturebeat.com/2020/01/28/tony-fadell-and-mitsui-ki...

[3] https://venturebeat.com/2018/08/16/sila-nanotechnologies-rai...

[4] https://www.startupdaily.net/2020/03/south-australias-tesla-...

[5] https://phys.org/news/2019-03-renewable-energy-solution-indu...

[6] https://www.geekwire.com/2019/company-backed-bill-gates-clai...

Wish in one hand, * in the other and see which fills up faster.

I've been hearing similar stories for the last 30 years.

In that time the energy mix feeding my house has gone roughtly from 50% coal:40% nuke:10% NG, to 30% nuke:30% NG:30% coal: 10% wind. For an increase in total CO2 production, despite a slightly better CO2/MW rating due to moving from coal to NG.

No one AFAIK, has managed to get their "green" energy solutions down to the levels france was in the 1980's, despite the increases in cost that come with not only having to build energy production, but 2-3x the production in storage.

The day its actually cheaper to put solar on your average house with a tesla power wall will be the day there is a shitton of home equity loans being taken out to do so. Or for that matter people knocking on everyones door offering to do it for "free" to sell the power back to the grid.

Right now most of the green energy solutions are "inexpensive" because they are being supported by natural gas generators. Meaning they are cleaner than standalone gas, but still a carbon source.

This is... really off base for 2020. Solar and wind in the US are incredibly cheap sources of energy. Literally the cheapest options: https://www.lazard.com/media/451445/grphx_lcoe-02-02.jpg

Residential solar is expensive because you don't have the same economy of scale you see in utility-scale deployments. Also something like 1/3 of the cost is "soft costs" such as marketing and the crazy red tape associated with permitting.

The latest academic modelling shows that we can meet 70-80% of energy demand from renewable energy alone, even without storage. All it requires is building a modest excess of capacity and a 50/50 wind/solar mix. When it comes time to add storage, battery storage costs have been plummeting and already dropped 75% over the last 6 years: https://www.greentechmedia.com/articles/read/report-levelize...

But if its truly cheaper than all the worlds CO2 production problems are solved!!

Energy is sufficiently deregulated in enough of the US that they will sell to the areas that aren't, just like France was doing with all its "old dirty" nukes, to the countries around it building "green".

I can't tell from your graph, but I saw an actual cost estimate a couple months ago that points out what has been true for the past 10 years of "wind is cheaper" metrics.

Which is that its not, because its not continuous (or controllable) power delivery. Nor does it account for the fact that it also needs to be overbuilt if its going to supply an energy storage system either. Nor does it account for the energy storage costs.

So, yes in absolute KW produced its cheaper, but that does little but create an oversupply problem. Which is why in places like TX the power costs frequently go to zero when the wind is blowing and spike at other times. Making cheap power when you don't need it doesn't help. What TX needs is lots of power at 3PM (when in theory solar would be useful, but the existing smaller plants aren't making money either).

The net result in TX has been lots of wind install, but even more gas install. Because the gas plants are actually making money. If big battery plants are economically workable then we would also see a lot of companies arbitraging the free wind energy into $ when the price spikes but we don't see that either.

So its not a simple/sure bet like is being claimed.

(for those that don't know, TX is one of the largest green energy systems in the world https://en.wikipedia.org/wiki/Wind_power_in_Texas)

It's not solved yet, because energy is not deregulated enough, and entrenched financial interests won't go down without a legislative fight.

For example, in Ohio several state legislators were purchased, and passed regulation claiming to "save" nuclear but what it really did was bail out coal and prevent the cheapest source of energy from competing in the market. The most surprising is that this corruption is actually resulting in prosecutions, and the top regulator has now resigned too:

https://www.bloomberg.com/news/articles/2020-11-20/firstener...

Funny you should bring up Texas, it actually is installing massive batteries, with 17GW in the pipeline last time I heard. And natural gas is dwindling to nothing, getting replaced with solar. ERCOT is one of the very very very few places where cheapest cost can actually win, and it's where we are going to see natural gas die first because of that.

I live in tx, and have a bit of knowledge about this.

Gas plants are still being constructed, and there are more in planning. There is 0 indication its going anywhere anytime soon.

Gas plants are replacing coal plants: https://www.power-eng.com/2019/04/11/eia-gas-fired-combined-...

They're generally cheaper to run and better at load-following.

Coal has been on the way out for a decade or so, long before renewable energy started to be a big player in the US.

Right, but its the NG that enables the wind because of said load following. That combined with the fact that they are super cheap to build, and the US has a huge glut in NG due to fracking and regulation on how much we can sell internationally.

But, that isn't the point. The point is that even if we go to a ~70% wind model, we will _STILL_BE_WORSE_OFF_THAN_FRANCE_WAS_40_YEARS_AGO_.

The "Green energy" movement there is _INCREASING_ their CO2 production.

https://www.worldometers.info/co2-emissions/france-co2-emiss...

I don't get why this is so hard to understand for all those down voters. Worse, at the current rates, we won't get there for decades.

Click the link and look at the graph in my parent comment. Gas started rising in 2002, when wind and solar had an invisibly small presence in the US market.

It sure sounds like you were arguing that renewable energy had a time-travel effect causing the construction of gas before the renewables were added?

People are downvoting because what you're saying makes no sense.

You realize that the natural gas plants sit mostly idle when renewables are producing enough power, right? And the more renewable capacity we build, the more often that happens?

> Worse, at the current rates, we won't get there for decades.

So, about the same timespan it took to execute the Messmer plan in France, and at a fraction of the cost? (The Messmer plan was France's big nuclear buildout from the 70s through the 90s.)

On a side note, even if we started building reactors today they probably wouldn't be operational for a decade or more (including planning time).

> The point is that even if we go to a ~70% wind model, we will _STILL_BE_WORSE_OFF_THAN_FRANCE_WAS_40_YEARS_AGO_.

No.[1]

Denmark has about 48% of their energy from wind, and their per capita carbon emissions are at the same level as in 1960.[2] Ireland with 33% is at the same level as in 1980.[3] Portugal with 27% now has the same emissions as 1990.[4]

There is a trend here.

> I don't get why this is so hard to understand for all those down voters.

You are ignoring data that doesn't agree with your hypothesis.

[1] https://www.statista.com/statistics/217804/wind-energy-penet...

[2] https://data.worldbank.org/indicator/EN.ATM.CO2E.PC?location...

[3] https://data.worldbank.org/indicator/EN.ATM.CO2E.PC?location...

[4] https://data.worldbank.org/indicator/EN.ATM.CO2E.PC?location...

Since 2019/early 2020, the data disagrees with your assessment. Gas plants are leaving the interconnection queue, and massive amounts of solar and wind are coming in:

https://rmi.org/clean-energy-is-canceling-gas-plants/

The market is responding to the inflection point, where renewables are clearly the lowest cost energy.

The switch is happening right now!

> But if its truly cheaper than all the worlds CO2 production problems are solved!!

I mean, the International Panel on Climate Change certainly thinks renewable energy is a core part of solving carbon emissions. Their special Report on 1.5C AKA SR15 (https://www.ipcc.ch/sr15/chapter/spm/) says:

> In 1.5°C pathways with no or limited overshoot, renewables are projected to supply 70–85% (interquartile range) of electricity in 2050 (high confidence).

For the 3 scenarios where we achieve needed emissions reductions, renewables are 48-60% of electricity generation in 2030, and 63-77% in 2050.

> Energy is sufficiently deregulated in enough of the US that they will sell to the areas that aren't

In the US in 2020 the majority of new generating capacity being added is from solar or wind: https://www.eia.gov/todayinenergy/detail.php?id=42495# -- and if you do the math for capacity factors (around 40% for wind, 25%ish for solar, 60% is for natural gas) then you'll find that solar and wind capacity generates more electricity than the gas.

If the US grid operators and utilities are building all this renewable energy capacity, perhaps they know something...??

> in places like TX the power costs frequently go to zero when the wind is blowing and spike at other times.

Isn't that what a free market is supposed to do -- respond to supply and demand? Last I checked, we don't say that the stock market is broken because it goes up and down.

> that does little but create an oversupply problem

Are you saying free excess power is a BAD thing? I can think of a TON of ways to take advantage of a temporary oversupply; capturing it in storage is only one of them.

> If big battery plants are economically workable then we would also see a lot of companies arbitraging the free wind energy into $ when the price spikes but we don't see that either

Hold your horses -- they're literally starting to do this in Australia. Batteries were pretty expensive up until a couple years ago: https://arstechnica.com/science/2020/10/us-grid-battery-cost...

Now we're seeing a race to install batteries. Energy arbitrage is only one of the possible income streams -- grid services such as frequency regulation are an even bigger source of funds. The "Big Battery" in Australia has already paid for itself after just a couple year and they've already increased capacity by 50% and are installing a second one in Victoria.

> balancing with gas

In the US, gas capacity is mostly replacing dirtier, more expensive coal powerplants. I don't see a problem with using spare gas capacity to help balance the grid while storage gets ramped up -- the renewable generation is directly replacing fossil fuels except when they need an extra boost.

> In that time the energy mix feeding my house

That's a pretty small data point. Check out the worldwide trend.[1]

> No one AFAIK, has managed to get their "green" energy solutions down to the levels france was in the 1980's

The share of nuclear power generation in France peaked around 2005.[2] They had to temporarily shut down four plants in 2018 to avoid meltdowns[3] due to heatwaves which will become more frequent and they are cancelling nuclear projects in favor of renewable projects.

Renewable energy surpassed nuclear for the EU as a whole around 2014[4]. And the first half of this year, renewables provided more energy than fossil fuels:

"It’s official: in the first half of 2020, and for the first time, Europe generated more electricity from renewable sources than from fossil fuels. Not only that, but electricity is proving cheaper in countries that have more renewables.

From January to June, wind, solar, hydro and bioenergy generated 40% of the electricity across the EU’s 27 member states, while fossil fuels generated 34%. In the United States, by way of contrast, fossil fuels generated more than 62% of electricity last year, while renewables accounted for less than 18%."[5]

> The day its actually cheaper to put solar on your average house with a tesla power wall will be the day there is a shitton of home equity loans being taken out to do so.

"Per the US residential solar finance update: H2 2019, a bit over half of homeowners who upgrade their homes with solar panels get a loan, a sixth of the population pay cash, and the last third rent their roof out to a third party.

Within those numbers were $1.35 billion of residential solar loan ABS (asset backed security) 144A public issuances in 2019, doubling 2018’s $680 million."[6]

[1] https://ourworldindata.org/grapher/modern-renewable-energy-c...

[2] https://en.wikipedia.org/wiki/Energy_in_France

[3] https://www.independent.co.uk/news/world/europe/france-nucle...

[4] https://en.wikipedia.org/wiki/Renewable_energy_in_the_Europe...

[5] https://www.forbes.com/sites/davidrvetter/2020/07/23/europea...

[6] https://commercialsolarguy.com/2020/07/01/residential-solar-...

You can lie with press releases all day long. The pacific NW had more "renewable" energy production than carbon sources for decades, its carbon footprint per MW is about 3x that of France.

Shifing to a pacific NW model is economically something that is happening on its own, and has been for the 30 years I linked above. But not because we are building "Green" because all those intermittent sources are 30%+ backed by NG generators which can be spun up/down on demand and over the past 15 are super cheap to install. And NG has been super cheap due to fracking and the fact that the US can't sell it internationally.

But its not about percentages, its about how much CO2 is being produced.

https://www.nwcouncil.org/news/northwest-carbon-footprint-lo...

Today I learned that Wikipedia and OurWorldInData are "press releases"...

Yes, we get it that you have a wholly irrational dislike of renewable energy and are willing to grasp at straws to argue against it.

> because all those intermittent sources are 30%+ backed by NG generators which can be spun up/down on demand and over the past 15 are super cheap to install

I'm trying to follow what you're arguing here and it makes no sense. You're arguing that the Northwest is simultaneously using almost all renewables and using tons of fossil fuels...? It can't be both.

We'll ignore the fact that according to your link, wind+solar is less than 10% of the electricity generation there, and there's no real evidence of a big investment in either, just the pre-existing hydro power.

No, I just don't like people who pretend that wind/solar are actually going to solve the CO2 problem by pointing at some future technology or pricing model that is going to save us while stubbornly refusing to accept a 70 year old technology that is actually carbon free.

That is what got us were we are today. Because back in the 1970/80's people argued the exact same thing (no nukes, solar will save us). And what happened? We got more coal plants because they were way cheaper except in the few places that actually built nukes. The same argument happened in the 1990s/2000s and what did we get? A few percent of wind/solar, and a shift to NG due to fracking.

_TODAY_ if you want a competitive solution you build somewhere in the ballpark of 20-70% wind/solar and back it with NG.

Where will that put us in over the next 40 years?

Its going to create even more C02 because power utilization is going up. All that is going to do is slow the rate of increase, which is exactly what the current charts are showing. The rate of wind/solar rollout is barely exceeding the demand curve and in a lot of places its regressing, particularly in places where ancient nuke plants are being replaced by "green" technologies.

Yes, sure wind+NG is better than NG or coal, but its a shit solution. If batteries actually get cheap then we can do 2x wind+battery, but that isn't here today. What we have today are nuke plants, just like we did 40 years ago. If back then instead of waiting for the future to save us we were more pragmatic the ice caps wouldn't be melting.

My personal opinion is that in 40 years what is going to solve this problem is a giant war, because wishing for the future to save us hasn't worked yet.

You're trying to argue that the solar market in 2020 looks the same as it did in the 1970s? Seriously? When the cost of solar PV has dropped 70% just in the last decade alone...?

https://www.lazard.com/media/451447/grphx_lcoe-09-09.jpg

I can't even find LCOE figures that far back, but the cost of solar modules in 1975 was just over $100 a watt. Today, a solar panel can cost as little as $0.50/W

Citation: https://news.energysage.com/the-history-and-invention-of-sol...

That's literally a TWO HUNDREDFOLD decrease in price, and modern panels last longer. This is like arguing that computers are useless in 2020 because in the 70s they were not very powerful.

> If batteries actually get cheap then we can do 2x wind+battery, but that isn't here today

Battery storage costs have already dropped 75% over the last 6 years, and we're pretty close to making that possible: https://www.greentechmedia.com/articles/read/report-levelize...

In the field, people bidding on energy projects are doing mixed solar+storage and wind+storage at prices comparable to nuclear or lower. Today. With 2020 tech and 2020 pricing, not 2025, not 2030. Granted, these aren't including a lot of storage right now (1-4 hours generally) but as battery prices continue to drop that will increase.

Given how costs are dropping as the technologies scale up, in 2025 people won't even think twice before choosing renewables+storage over nuclear, because it'll be a no-brainer.

> What we have today are nuke plants, just like we did 40 years ago.

Yes, that's the problem. Nuke plants plants have advanced technologically in the last 40 years, but in terms of cost they're actually more expensive because we found more failure modes (and need to prevent them).

And this is disappointing because I worked in nuclear physics for a few years and really wanted to believe that nuclear energy was going to save us... and it catastrophically failed to deliver.

> Sure there is -- one option is a terrifically expensive and dangerous dead end.

You say this with such certainty... but all of your citations are about renewable and solar. What is your background that gives you the confidence to make such a statement when billions are being spent on nuclear power research by, what I presume you would consider to be, idiots?

I'll stop there because ad hominem arguments are insufficient anyway. The only argument that really matters here is, capitalism will decide the path forward, and not you (fortunately, imo). And the way it does that is that people, none of whom can predict the future with perfect clarity, make intelligent bets about what technologies will produce the most value. As the research and development pans out, the winners become evident, and either way, consumers win.

There's no need to try to strangle the nuclear baby in the crib. If it doesn't work, let it die in development. But, really, how can you make such a confident assertion about the prospects of nuclear? It boggles the mind.

Hopefully HN is still a place where people who are interested in technology can make predictions. I make an effort never to call anyone names, but to focus on the data. I don't think people who are excited about new nuclear technology are idiots. Obviously, nuclear physicists are smart people. But no one, no matter how intelligent, is immune to hubris.

> capitalism will decide the path forward, and not you (fortunately, imo)

Capitalism is the poster child of hubris, and ultimately a fallible set of human ideas -- but that's a different discussion. However, if that is your metric, then the argument is already over.[1]

> how can you make such a confident assertion about the prospects of nuclear? It boggles the mind.

"But even some proponents of nuclear power doubt the program will spur construction of new commercial reactors as long as natural gas and renewable energy remain relatively cheap. 'New builds can’t compete with renewables,' says Robert Rosner, a physicist at the University of Chicago. 'Certainly not now.'

...

"Ramana [U. of B.C. physicist] questions whether the U.S. nuclear industry can be saved. Although issues of dealing with waste and the public’s apprehension about radioactivity remain, the biggest issue confronting the nuclear industry is the high capital cost of new reactors, which can be $7 billion or more. In deregulated markets, utility companies cannot afford such capital expenses, which is why cheaper renewables may ultimately replace nuclear energy, he says. 'This is a sunset industry,' he says, 'and the sooner you recognize that the better.'[2]

There are competing arguments in the same article, but the argument is between nuclear designs that only exist on paper and wind/solar technology that has already been built and is already providing electricity. If you want to put a shovel in the ground today and have renewable energy in a few months, it's possible with existing solar and wind at prices cheaper than any proposed idea for nuclear.

Of course nuclear physicists want billions of dollars for R&D. Of course they believe they can create the one true design that doesn't have any of the fundamental problems of nuclear technology. But according to those people, even their best efforts won't produce any results for years and years.

To put it into startup language, do you want to build a service using an architecture, language, and set of libraries that have been tested and used for years, or do you want to wait a decade and then start building on the latest holy grail platform that no one has successfully used in production?

[1] https://en.wikipedia.org/wiki/Cost_of_electricity_by_source

[2] https://www.sciencemag.org/news/2020/05/us-department-energy...

Small scale reactors, built in an agile fashion, seems like a significant departure from the types of projects you describe. I agree with making predictions, but I think they should be limited and testable predictions. If you say nuclear is a dying industry, I'm going to say you're painting with too broad a brush.

I don't disagree that nuclear appears to be on the wrong track, but I've learned that people often conflate doing something wrong with that thing being impossible to do right.

Sure, except the Earth rotates and we can only collect said energy for half the day. It's also fluctuates in the amount of energy collected seasonally (tilt of Earth's axis), and most energy demand happens to occur in northern latitudes where these fluctuations are more extreme. Our ability to store energy remains orders of magnitude smaller than what is required to actually use this as a primary source of energy.

This doesn't provide a path to a decarbonized energy sector. It's a good way of mitigating emissions while a real decarbonization solution is implemented. Right now nuclear is the only known way of doing that, save for geographically dependent sources like hydroelectricity and geothermal power.

Multiple countries generate the majority of their electricity from nuclear power. No country generates more than 15% of their energy from solar power, and only 3 countries (Malta, Yemen, Honduras) generate more than 10%.

How many Falcon 9s did SpaceX blow up before they got a successful launch? Are you willing to have that many meltdowns in order to get your "highly modular, efficient, miniature 'assemble-yourself' nuclear power plants"? Even if you are, how are you going to convince people that it's worth it? What are you going to say when the press tars your project with the headline, "A Fukushima In Your Backyard?"

Yeah people suck at critical reasoning.

Falcon 9 had been an amazing success, which kind of defeats your point. Of course you need to perfect the technology (or close to it) before you deploy it. People are easily defeated by repetition. To paraphrase, if you have one reactor and it fails, it's a news story. If you have one million reactors and one fails, it's a statistic.

Yes. Build the first ones in a desert. Do chaos testing, stress testing and over-engineer each component and the whole system so that any teenager can install and operate it. I'm not kidding. Make it "teen certified".

> Build the first ones in a desert

Leaving aside the fact the deserts are essential biomes on their own and shouldn't be treated as garbage dumps, didn't Fukishima and Chernobyl teach us that putting nuclear power plants "way over there away from anything" doesn't really matter in the case of a serious accident?

Well, we've already got a tract of desert that's been heavily irradiated: https://en.wikipedia.org/wiki/Nevada_Test_Site

We've also got an ocean full of plastic and has hotspots where nuclear waste was dumped[1]. That's no reason to go around saying, "well, we've already trashed it, might as well let it continue to be trashed"

1. https://pubs.usgs.gov/circ/c1198/chapters/207-217_RadWaste.p...

> "way over there away from anything" doesn't really matter in the case of a serious accident?

The current level of burning coal equates to roughly 50 nuclear meltdowns a year because coal contains thorium. So in terms of "putting it anywhere," an in-place switch from coal to fission in a populated area is actually safer.

In addition, the article identifies outdated designs as the primary reason for the cost of nuclear plant rollout. Guess what? Modern plant designs are passively safe, particle physics (not electronics, nor mechanics, nor humans) prevents criticality events.

I'm afraid that the "coal is more radioactive" argument isn't going to get the traction that it did in 20 years ago. There are too many caveats and assumptions made in what the level of radioactivity in coal ash means to meaningfully compare it to fissionable radionuclides.

As measured by GHG emissions, nuclear power is certainly better for the climate. Whether or not that equates to "safer for human activity"? The devil is in the details.

> There are too many caveats and assumptions made in what the level of radioactivity in coal ash means to meaningfully compare it to fissionable radionuclides.

If the reactor is passively safe, the amortized level of radioactivity is zero, especially if the waste is "burned" in secondary reactors (such as TWR).

"if" and "amortized" are just different forms for caveats and assumptions. The level of radioactivity amortized over how long? Long enough time scales and the level of radioactivity is 0 for everything in the entire universe.

Way to needlessly misrepresent the argument.

Amortized over the duration where we are typically seeing meltdowns in gen1 reactors. 10-50 years?

> "if"

The article talks about modern designs. It is an aspiration, not an assumption. The central argument is that adopting modern reactor designs eliminates meltdown concerns (in addition to reducing costs as the article suggests). That isn't an assumption, that is a suggestion.

And presumably they would melt down a few plants in the Nevada desert nuclear sites before launching... there’s an obvious difference between testing and deployment

The first Falcon 9 launch was a success -- maybe you're thinking of the older Falcon 1 rockets that exploded during launch?

NuScale power just got approval a month or two ago, and is working on their small (50MW reactors)

NuScale just lost a bunch of their backers for their project after announcing several years of delays and cost overruns.

https://www.sciencemag.org/news/2020/11/several-us-utilities...

> announced that completion of the project would be delayed by 3 years to 2030. It also estimates the cost would climb from $4.2 billion to $6.1 billion.

Unfortunately this is a very old story with the nuclear industry: they consistently over-promise and under-deliver. The closer projects get to completion, the more costs grow and the further the completion date gets pushed out. This has been true of most of the recent builds in the US and Europe.

The tech is fine, but there seems to be something wrong with the industry.

And they can't find customers.

Their target price, of $55/MWh, would probably be acceptable if delivered today.

But when you look at hybrid solar/battery projects, we are already getting close to $55/MWh for fully dispatchable energy, that doesn't have to be run 24x7 like nuclear but actually matches the load curve of the grid.

Seriously, there are battery projects being deployed now, attached to solar, with a storage cost of $80/MWh, and energy cost of $20/MWh. Since overnight demand is small compared to daytime, only about half of the energy needs to be stored. That averages out to $60/MWh, today. In 5 years it will be something like $40-$50. In 10 years, when NuScale deploys, the cost of solar+storage will be $30-40/MWh. And it might even get cheaper.

NuScale may be useful for polar locations without much wind, but I don't foresee them having a huge addressable market.

Sure there are reasons this is a bad idea, but there are certainly reasons it's a good one. It seems like building large plants the way we typically do backloads tons of risk. And it's not really working.

How does the risk of nuclear meltdown scale down with size? Both in terms of likelihood and how catastrophic?

The actual Elon Musk is building large battery farms, which have ROI measured in single-digit years.

In Hollywood terms, solar/wind/battery plants are now bankable. I don't know who in their right mind would underwrite a nuclear plant now.

I suppose it would be useful in a space ship, or in sea vessels smaller than current nuclear subs.

> Where is the Elon Musk of Nuclear Reactors?

They are doing fusion.

Don't get me wrong, I have been pro-fission for a very long time (as it could get us out of this fossil fuel pit faster than any other option, even if we plan to replace it in the future). It has a very bad name because of bad green peace pseudo-science, fusion has better PR.

Costs go up because they get paid.

Once enough money is sunk, no amount more can kill the project. (Rare exception, Californian "hi-speed" rail! But that was politics.) Thing is, none of the money is wasted, as such: somebody gets every single dollar. The higher the costs, the more they get. So long as that somebody spreads enough of it around, everybody actually involved (the paying public not among them) benefits.

That model applies to any big-enough multi-year construction project, whether it's a nuke plant, a subway line, or a fleet of bombers. So, it's nothing to do with nukes, as such, except that for nukes there are always enough pigeons convinced that the next one will be cheap, for sure! And around we go.

I can't see any other reference to this but there is a whole field of study on this. See e.g. https://www.amazon.com.au/Cost-Disease-Computers-Cheaper-Hea... or google "cost disease".

One thing I have noticed is the incredible increase in the cost of regulatory compliance over the years.

It can take years to get the dozens to hundreds of approvals needed. Vast armies of people are needed to document and monitor compliance.

At a large bank where I worked, I analyzed the previous $600 million in IT projects. $350 million of that $600m was on compliance projects.

> At a large bank where I worked, I analyzed the previous $600 million in IT projects. $350 million of that $600m was on compliance projects.

We live in a complex world. "Traffic laws," to keep people from inadvertently interfering with — or harming — each other, are necessarily more complex than they were, say, 200 years ago.

http://rationallyspeakingpodcast.org/show/rs-236-alex-tabarr...

the current state of the argument, pretty interesting.

Sadly sci hub fails on the mentioned study :

https://www.cell.com/joule/fulltext/S2542-4351(20)30458-X

"Analysis points to ways engineering strategies could be reimagined to minimize delays and other unanticipated expenses."

Ah, okay, reimagined. Phew.

Let me guess. They need to be more Agile.

Does anyone have a PDF for https://www.sciencedirect.com/science/article/abs/pii/S25424...? Sci-hub has the wrong paper for this DOI.

NuScale's small modular reactor (SMR) seems like a good solution to this problem.

https://www.energy.gov/ne/nuclear-reactor-technologies/small...

Has this study included corruption as one of the key points of raising costs?

What about the possibility that their is no bad consequences for cost over runs? If fact, from a corporate view it is highly beneficial.

Show me the reward structure and I will show you the outcome.

I'm surprised by these findings but they do make sense. Seems to me like it's basically the giant construction project version of design for manufacturing.

The biggest reason for cost overruns in the US is that the companies that are doing the work make more money from cost overruns. The way nuclear plant financing works in the US is that taxpayers and ratepayers ultimately pay for the cost no matter what the cost is and how much the overruns are. The cost of the plant with all the overruns usually shows up on the electric bills of the consumers, so the more overruns the more one can charge the consumers. This overcharging may not happen fully intentionally, it may be happening as kind of a long bureaucratic slide into convenience but it is happening.

Now you may think there is an easy solution for this. Funding for these plants should be changed, so that private parties pay for them and they get their money by selling the electricity produced by the plants. I.e., they should be funded the way any private business is supposed to be funded in a capitalist economy. Sounds like a good idea at first, but there is a catch. If that is the rule, nobody in their right mind will ever build a nuclear plant ever again. Gas, wind and solar are all so much cheaper now. And yes, even considering back ups for when the wind is not blowing, wind is much much cheaper than nuclear.

So the obvious solution is just not to build nuclear. Even if you do not consider the real risk to human life, it is a an incredibly expensive extravagance. If you do consider the risk to human life it is absolute madness.

You are right about wind and solar being able to bid the lowest rate to the grid.

Newer reactor designs such as FAST reactors do not pose a threat to human safety.

> Gas, wind and solar are all so much cheaper now.

But cannot power our collective needs (especially industrial), unlike nuclear, which can (at least temporarily).

> an incredibly expensive extravagance

So is entire cities disappearing under growing oceans because nuclear makes us feel bad. And if we want nuclear to be cheaper, then we should collectively raise the cost of fossil fuels until gas is $20/gallon. Then we'll see how expensive nuclear is.

The article is fairly sparse in details and could be summarized by:

>so if more components of the plant, or even the entire plant, could be built offsite under controlled factory conditions, such extra costs could be substantially cut

From the paper (locked behind paywall) there is also a mention of rising labor costs.

I wonder if this has more to do with the size of reactors, which went up from 600-800MWhe up to the 1500MWhe range, i.e. if containment buildings (safety systems etc) at those power levels are just a massive dis-economy of scale.

How can you build today something, with any kind of electronic parts, and excpect a life from 50 years?

Overbuild it to be capable of taking many times what it is expected to encounter just like for something mechanical. It being electronic isn't the cause in itself and can make things significantly better from fewer moving parts.

You have drawings that show how it works and a supply of spare parts. When it stops working people can figure out why using the drawings and then fix or replace the broken or worn out components.

No I guess thats not so easy. Every new part in a design for a nuclear power station has to be certified. So just use another almost same chip is not so easy, because you have to make o lot of tests, so the new part is allowed. Thats why a lot of things for nuclear power station are so extreme expensive.

The components we are installing have chips in them, but we don’t worry about what chip because the manufacturer of the product keeps it working the same way. The modicon quantum Plc uses the pentium 166 chip and was released in 1994 and I can still buy it today from Schneider although they encourage new installations to use a newer product. I am sure I will still be encountering old quantum Plc in the field when I retire, the platform will have a 50 year lifespan.

Even when a line of controllers is end of life the manufacturers provide an upgrade path to their newer product line.

31$USD for a fucking pdf?

Please. We have known why this is the case (in the West) for decades.

http://www.phyast.pitt.edu/~blc/book/chapter9.html

This was certainly true in 1990 when it was published.

Sadly, the facts don't seem to have changed that much.

It switched from coal in 1990 to natgas and now renewables taking market share. Nuclear is still not cost competitive without lavish subsidy injections though, no. Never was, probably never will be.

Reply on news.ycombinator.com