[threeseed]

> people will use the losses the nuclear incurs to subsidize renewables as a proof that nuclear is uneconomical, and that renewables beat it handily

There is already plenty of proof around the world that renewables are cheaper than nuclear.

[Krssst]

Maybe cheaper as of today if we don't account for storage, but since buulding renewables use far more materials than nuclear, would fossil fuels which ensured cheap production and transport become lacking, or base materials extraction not being able to follow a rising demand, I am not sure it would still be the case.

[TheOtherHobbes]

If nuclear used fewer raw materials nuke plants wouldn't be huge - huge - capital projects, wouldn't take years to come online, and wouldn't have huge cleanup costs.

If the money spent on nukes had been spent on renewables and on developing storage we wouldn't have these problems.

This was predictable decades ago.

The reality is that nukes are a political solution to a political problem. It's nice that they sometimes generate energy for a while, but there is no sense in which they've ever been a rational economic choice.

[mathiasgredal]

Nuclear uses roughly an order of magnitude fewer ressources than wind, without factoring in storage: https://imgur.com/a/Kc2h21O.

Source: https://www.energy.gov/sites/prod/files/2017/03/f34/quadrenn... (page 390)

[Schroedingersat]

You can't use a report from 2015 citing data from 2010 as indicative of a technology that has dropped in cost 10-fold and nearly doubled capacity factor since then.

Additionally the overwhelming majority of that material is foundation and tower. Both of which can be reused by replacing the nacelle.

[mathiasgredal]

The Vestas V112-3MW from 2009 has a 42% capacity factor. I don’t know any modern wind turbines that has a capacity factor of 80%.

https://www.thewindpower.net/turbine_en_413_vestas_v112-3000...

https://en.m.wikipedia.org/wiki/Humber_Gateway_Wind_Farm

[Schroedingersat]

And this makes the point. A 2009 model turbine is a huge improvement on a retrospective on past history of plants that had been running for some time in 2010

https://www.vestas.com/content/dam/vestas-com/global/en/sust...

A V112 has 40% lower materials per TWh (or ~30% for the lower wind configuration) than your report without considering modular foundations which would drop it by another factor of 3 or being able to reuse the foundation and tower at least once which would drop it by another 2.

On the other end you're not considering Uranium mining or enrichment facilities or waste storage or that the metals are all recyclable immediately rather than half of them being LLW or SLIW.

[Manuel_D]

Nuclear plants use very little raw materials relative to the amount of power the produce. When built at scale, nuclear plants have been delivered at prices around $1-2 billion dollars per GW of capacity.

The cheapest form of carbon-free energy really depends on what the objective is: small reductions in a mostly fossil-fuel grid? Or total replacement of fossil fuels? Renewables are great for the former: you can throw up some solar panels or wind turbines and reduce a chunk of fossil fuels use. But once you try to start delivering significant portions of the energy grid through intermittent sources the surplus energy starts to get wasted, and the effectiveness drops.

[Schroedingersat]

Nuclear is roughly equal to wind on a modular foundation if you account for the fact that the tower and foundation outlast the nacelle. The "$2/GW" nuclear reactors were all built by state run agencies with opaque budgets and in France's, Japan's, and South Korea's cases have all proved wildly unreliable in addition to having opaque public subsidy on top of the very large visible subsidies in the supply chain and finance. If you think it's possible to match the prices China reports that megaprojects cost, I'd like to see any examples of projects in the global north with auditable accounting matching their figures in hydro, or highways, or rail, or ports or... basically anything.

In mediocre to good areas with something like the PEG racking system solar uses about the same raw material than nuclear already and it's almost all sand. By the time a new nuke came online this will be far less.

Both are recyclable. 12 hour storage adds negligible mass and can easily cover daily variation.

Intermittent power without storage can easily feed dispatchable loads like EV charging, chemical feedstock and heat production. These vastly exceed non-dispatchable electricity and can be used for virtual seasonal storage.

There are only a small handful of areas best served by nuclear, and most of them have hydro or nuclear already.

There's a narrow niche where nuclear is optimal:

Grid electricity between 50% and 80% penetration in the 50% of areas where hybrid CSP + e-fuel backup isn't better. This niche is rapidly shrinking and could easily be gone by the time one is built. More carbon can be removed faster and with fewer resources by throwing renewables at the other 10 or so TW of fossil fuels currently being burnt. Until those resources are committed, new nuclear just delays things.

[Manuel_D]

> Nuclear is roughly equal to wind on a modular foundation if you account for the fact that the tower and foundation outlast the nacelle.

Except intermittent sources also need storage. They also need long distance transmission lines to bring power from remote areas of generation to places of demand (whereas you can just place nuclear plants next to areas of demand).

This is a common pattern in renewables discussion: laser focus on generation and ignoring the fact that wind and solar have storage and transmission requirements that other energy sources don't have.

> The "$2/GW" nuclear reactors were all built by state run agencies with opaque budgets

Nope, do more research. These were all built in the US with public cost history.

https://en.m.wikipedia.org/wiki/Peach_Bottom_Nuclear_Generat...

https://en.m.wikipedia.org/wiki/Browns_Ferry_Nuclear_Plant

https://en.m.wikipedia.org/wiki/Byron_Nuclear_Generating_Sta...

https://en.m.wikipedia.org/wiki/McGuire_Nuclear_Station

https://en.m.wikipedia.org/wiki/Donald_C._Cook_Nuclear_Plant

https://en.m.wikipedia.org/wiki/Indian_Point_Energy_Center

https://en.m.wikipedia.org/wiki/Turkey_Point_Nuclear_Generat...

https://en.m.wikipedia.org/wiki/Arkansas_Nuclear_One

https://en.m.wikipedia.org/wiki/St._Lucie_Nuclear_Power_Plan...

https://en.m.wikipedia.org/wiki/North_Anna_Nuclear_Generatin...

> 12 hour storage adds negligible mass and can easily cover daily variation

12 hours of storage for the world is 30,000 GWh. This is 70-100 times the global battery production output. "Negligible mass" is going to have to see a hundredfold increase in som extraction industries. By comparison, nuclear already produces 20% of the US's electricity hand about a tenth of the world's electricity. A tenfold increase is much more manageable than a hundredfold increase.

> Intermittent power without storage can easily feed dispatchable loads like EV charging, chemical feedstock and heat production. These vastly exceed non-dispatchable electricity and can be used for virtual seasonal storage.

If you're going to tell chemical industries and metallurgy plants that they'll have to cease production for part of the year when renewables are producing lower than average output, then that has to be factored into your costs. If the price of steel and ammonia goes up because they can't run their plants as usual, then that cost is ultimately borne by consumers. You can't just use load shifting as part of the plan and ignore the costs of load shifting. "Virtual seasonal storage" amounts to "tell industries to shut off during winter". And no, heat production is not non-dispatchable unless you're okay with people freezing to death.

[Schroedingersat]

> Except intermittent sources also need storage. They also need long distance transmission lines to bring power from remote areas of generation to places of demand

You are correct. They also need storage like nuclear for non-dispatchable loads in areas without good hydro or CSP resource.

For the remainder your battery production figures are off by at least a factor of two. China delivered 280GWh in H1 2022 at the peak of a market crunch in an industry that is growing at 50% YoY. There's no compelling reason to think the 5TWh/yr of factories under construction won't be completed on time as the renewable industry has been consistently over-delivering for a decade.

Your scaling for nuclear is new capacity. Which is around 5GW/yr right now. It has to increase tenfold to match the last year of new renewable generation, or fifteenfold to match the new capacity weighted installation.

> (whereas you can just place nuclear plants next to areas of demand).

Incorrect. Seismic activity, ground, water, temperature, security and many other concerns limit siting severely.

> If the price of steel and ammonia goes up because they can't run their plants as usual, then that cost is ultimately borne by consumers. You can't just use load shifting as part of the plan and ignore the costs of load shifting. "Virtual seasonal storage" amounts to "tell industries to shut off during winter". And no, heat production is not non-dispatchable unless you're okay with people freezing to death.

Hydrogen or ammonia continue to exist after you make them. Simply overprovision your $300/kW electrolyser slightly and use chemical energy as your buffer. This has the added advantage of being an emergency or low CF backup at minimal extra cost.

[Schroedingersat]

> Nope, do more research. These were all built in the US with public cost history.

On top of not including the cost of finance, liabity, or upstream supply chain which was provided by state military projects. The first one you linked had repeated safety violations and maintenance issues and has no record if how much they cost to remedy. The quoted price leads here:

https://www.osti.gov/biblio/6259203

And is in nominal dollars not 1986 dollars. $763 million 1966 dollars is $7 billion, not $3 billion. Add in the cost of finance and you get $10-14 billion or around $7/watt for an unsafe, inefficient plant with corrupt management. And this was not a greenfield site, it already had work for unit 1.

Do more research.

Every pro nuclear claim turns out to be a lie when examined even with the slightest scrutiny. All of them.

[Schroedingersat]

Heating is trivially dispatchable over 24 hours. Put some sand, brick or water between what you want to heat and the heat source. This method has been used for thkusands of years.

Scale it up a bit and you can do it seasonally.

[mensetmanusman]

Not if you calculate in the cost of wanting to have a nuclear weapons and submarine fleet…

[CamperBob2]

They don't compete. Nuclear provides base load.

[Schroedingersat]

They both provide non-dispatchable power. Renewables have a slight edge at moderate penetration with no storage because you can turn them off whenever you want without incurring massive costs and solar output is biased towards peak time.

Then there's hybrid PV-CSP which is available in about half of the world and is dispatchable. I guess you're probably right in that nuclear doesn't compete because hybrid CSP is vastly cheaper even in FOAK form and dispatchable power is superior.

[HyperSane]

Nuclear is dispatchable.

[Schroedingersat]

No it isn't. Ramping is slow and can't be done beyond 20% very often or you destroy your fuel and control rods

Reducing output doesn't reduce costs, it increases them. This is the opposite of dispatchable.

If you can only pay for your reactor by coercing people into buying daytime electricity for 20c/kWh rather than buying a solar panel that will pay for itself in 3 years then it's not dispatchable.

[Manuel_D]

You don't need to alter the thermal output of the reactor to modulate a nuclear plant's electrical output. You can more aggressively cool the reactor to reduce the energy delivered to the turbine. This isn't often done since it's essentially deliberately reducing the efficiency of the plant.

> Ramping is slow and can't be done beyond 20% very often or you destroy your fuel and control rods

20% is all that's necessary to accommodate most load variations: https://www.eia.gov/todayinenergy/detail.php?id=42915

[Schroedingersat]

That is also slow.

Dispatchable generation ramps in tens of seconds and you don't pay $90-200/MWh for it when you're not using it.

[HyperSane]

Not when you take intermittency into account.