[oropolo]

> 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.

[_n_b_]

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.

[brmgb]

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.

[Animats]

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.

[rootusrootus]

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...

[Animats]

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.

[AniseAbyss]

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?

[pram]

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

[willvarfar]

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

[rkagerer]

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...

[jessaustin]

Please don't give them any ideas...

[Animats]

No.

[dylan604]

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?

[akiselev]

> 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...

[smorephism]

>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.

[kevinpet]

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

[jbay808]

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

[giarc]

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?

[lumost]

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.

[mikeyouse]

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.

[giantg2]

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.

[CogentHedgehog]

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.

[giantg2]

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.

[CogentHedgehog]

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.

[giantg2]

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.

[cratermoon]

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.

[DennisP]

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

[CogentHedgehog]

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.

[7952]

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.

[jessaustin]

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.

[echelon]

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?

[fulafel]

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).

[dylan604]

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.

[stjohnswarts]

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.

[manfredo]

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...

[CogentHedgehog]

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.

[godelski]

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

[stjohnswarts]

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

[manfredo]

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...

[CogentHedgehog]

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.

[sir_bearington]

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.

[CogentHedgehog]

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.

[bazooka_penguin]

Sounds like we just have to build even more plants