Hacker News Reader

Comment by Maximus9000 | Hacker News Reader

300 MW is still pretty big, isn't it? Why don't they build nuclear reactors that can be made on a factory assembly line, shipped on the back of truck, and installed right where it's needed in a town?

I assume the reactors in a nuclear sub are sort of like what I'm talking about.

View on HN

This is also being done. Micro-reactors are generally sub-20MWe and are intended to be portable. The biggest name is micro-reactors is probably Project PELE[1]. PELE is a DoD project to develop a trailer mounted micro-reactors for remote deployment.

PELE will be tested at the Idaho National Lab[2] (DoE's primary nuclear energy research lab). Speaking of INL, they are also building the MARVEL[3] reactor. MARVEL is another test reactor of the micro-reactor type. MARVEL is being funded by DOE-NE (if I recall) and is designed to be be a 100kW.

[1] https://www.cto.mil/pele_eis/ [2] https://gain.inl.gov/GAINEPRINEI_MicroreactorProgramVirtualW... [3] https://www.energy.gov/ne/articles/new-marvel-project-aims-s...

Aren't micro reactors meant to be like one-use batteries? That's the info I got from listening to the Omega Tau podcast about this.

One use of 2-3 years. You have two of them working at the same time and when it’s time to refuel you have a new one waiting and old gets shipped back to factory for refueling.

The expensive part of nuclear power plants is all the electricity generation part (turbines/generators/piping/transformers/etc) that is pretty much the same as for coal power plants. You don’t swap out any of that.

What are the chances the old one will be left on the site because it's not financially viable to ship it back and recycle it?

They are the size of a shipping container, not prohibitively expensive to move. The customers buying these things have no ability to recycle/refuel or dispose of the spent fuel. I presume any purchase agreement comes with a commitment to deal with it at end of life.

I guess if the supplier goes out of business, you might have trouble, but I'd expect these companies are putting things in place for that eventuality e.g. insurance.

They are one-use batteries in the sense that they are sealed and not intended to be refueled at the place of deployment, if ever. They can last for a decade without refueling though, so it may be fine.

There are economies of scale, in the sense that (very roughly) material costs are a rounding error and scaling up the reactor doesn't make the design more complex it just makes the parts larger. The Gen III+ reactor designs all went big as a cost-efficiency measure, not that it worked out particularly well. The major caveats being that very large pressure vessels required specialized forges to be constructed and that very large concrete pours are more difficult than simply large ones.

The "build the reactor in a factory" idea could still hold some merit, though there's not much reason to make it smaller once it's small enough to fit on a barge or a wide-load trailer. I'm not sure that making even smaller reactors than that to make more of them for economies of scale makes sense. It's hard to imagine that 2x 150MW reactors could ever be cheaper than 1x 300MW reactor. The economies of scale efficiency benefits tend to be very sub-linear.

Yes. We could basically say that there are two different markets for small nuclear, which will probably settle at different levels.

The first is just stepping back a bit from the GW scale plants to try to reap some economies of scale in terms of making the projects and processes less bespoke, even though the raw amount of material consumed might be higher.

The other would be uh... I guess "niche" applications? This is where scaling down even more (sub 50MWe?) happens. This could be some DoD related stuff, or trying to power remote communities (which aren't already grid connected). I'm sure there might be a handful of interesting industrial applications (I know the Tar Sands in Canada is interested in SMRs to power their steam extraction processes).

But the time to get your roi is likely to go down significantly. Which means overall costs could still go down and you can gradually add more capacity.

But I suppose the biggest advantage is easier access to funding. Do you really want to invest 8 billion for an all or nothing project?

The economics just don't work out. Small reactors still end up being enormously expensive, but without the huge power output of large ones. There is a reason why all power plants use gigawatt-scale reactors.

And big ones require government subsidies, tax breaks, and federal insurance since they are uninsurable on their own. Externalities are critical in nuclear and it still doesn't make money for the operators.

If these are modular and come on a truck, does that mean that basically Westinghouse will finance its own inventory, and that for the customer there's just a down payment until it's up and running ? This would change the structure of financing AFAIK.

It's really unfortunate that we fail to price in negative externalities for other energy sources.

Are the externalities for wind and solar, the fastest growing electricity sources, significant when compared to other electricity sources?

No, but the externalities of fossil fuels, the worst energy sources, are very significant. Yet these get subsidized in many countries and Germany is even moving to coal. I'm comparison between nuclear and fossil it definitely matters.

Germany just passed a law to exit coal in 2038. It's called Kohleverstromungsbeendigungsgesetz. [1]

Last month, Germany also closed down the last nuclear power plant [2]

[1] https://de.wikipedia.org/wiki/Kohleverstromungsbeendigungsge...

[2] https://de.wikipedia.org/wiki/Atomausstieg#2023:_Abschaltung...

So how are they going to heat their homes?

They plan to have 10GW of hydrogen electrolyzers by 2030, among other storage investments. Assuming they meet that and build 10GW worth of dispatchable storage annually starting in 2028 they should be able to shut down their current ~40GW of coal and ~40GW of gas by 2038.

Without nuclear that would probably work the same as their current arrangement is working now. Germany and Denmark currently have the most expensive electricity on the EU.

That's one of the reasons why we had to shut down the most expensive and inflexible source of energy.

Nuclear is the most expensive way you can think of to generate power, but the most expensive power plant unfortunately dictates the price on the energy market in the EU ("Merit-Order") [1]

[1] https://www.tengelmann-energie.com/merit-order/

Yet somehow France manages to stay below the EU mean with 70% nuclear in the mix.

The merit order article you have linked is typical lobbist brainwash as to why your country needs to depend on gas when the sun doesn't shine and the wind doesn't blow.

Simple: Renewables. In 2030, 80% of the generated power should come from renewables [1].

Starting from 2025, each year, 7,8GW of wind and 22GW solar power will be added into the mix.

For reference, the nuclear plants we just shut down produced 4GW in total.

Germany is exporting excess energy each years to France and other neighbours. 2002 was the last year we had to import energy. [2]

[1] https://www.bmwk.de/Redaktion/DE/Dossier/erneuerbare-energie...

[2] https://de.statista.com/statistik/daten/studie/153533/umfrag...

The externalities for wind and solar is the reserve energy that currently consumer expects governments (in EU) to supply and pay for. It is the most expensive form of externalities in the grid, it is one of the dirtiest, and since the government is paying for it it is also the most subsidized.

This socialization of costs is sadly mostly unknown to the consumer. Governments could try moving the costs closer to the consumer by putting the costs onto the power company, forcing them to either cut off consumer when demand exceed supply or build enough storage to cover periods of suboptimal weather conditions.

> Why don't they build nuclear reactors that can be made on a factory assembly line,

Rolls Royce is proposing just that: https://www.rolls-royce-smr.com/why-rolls-royce-smr

Westinghouse has made nuclear submarine reactors.

The UK small modular reactor project is being run by Rolls-Royce who make the equivalent submarine reactors.

Both SMR projects seem to be about 10x the power output of a submarine reactor. I'm guessing they have to use a very different level of fuel enrichment than is in subs.

Nuscale is building 6x77MWe reactors in Idaho.. https://www.nuscalepower.com/en/projects

Literally like you suggested, small and able to ship on a truck, with no 'active' cooling needed. Can also take offline one of the 6 at a time for refuelling, instead of shutting down whole plant.

At a Vogtle like cost of $120/MWh. The project is looking incredibly shaky.

https://www.wired.com/story/the-dream-of-mini-nuclear-plants...

You're saying $120/MWh like it's a bad thing. The current average price for electricity in the US is $166/MWh as of March 2023 [1]. In sunny San Diego it's $475/MWh.

[1] https://www.bls.gov/regions/midwest/data/averageenergyprices...

You're confusing the cost charged to a consumer vs. the wholesale cost of electricity. $120/MWh is completely uncompetitive as a wholesale cost. Not even the cost of storage for solar/wind would bring their cost up to that.

Yeah, wind and solar are estimated to be $42.8 and $48.8 per MWh respectively in 2023 [0].

[0]: https://en.wikipedia.org/wiki/Cost_of_electricity_by_source#...

I'm a big fan of wind and solar, but we should compare like for like. AFAICT the numbers quoted here are for wind and solar without storage.

We need to compare solar and wind to these SMRs with the desired output characteristics. How acceptable is dropping the output for 10 hours each night, and lose 50% of your max output during long periods? If that's ok for your application, Solar without storage is probably the perfect solution. If you 99.99 of your time to be at a certain output level, all year round, solar is going to require overbuilding, and a weeks worth of storage (which changes the costs per MWh considerably).

Likewise, if you need a lot of heat for your application, solar is not going to be as good as an SMR.

The best technology is most likely going to be different depending on the use case. SMR would operate in a decently sized niche in our current world. What is not clear to me is how big this niche is, and how quickly it will shrink as gridescale storage tech becomes more economical.

Levelized cost already takes into account the capacity factor of the sources. On a per-peak watt basis, PV is 10x cheaper than nuclear.

We can look at the cost of providing steady power (synthetic baseload) from renewables + storage. This is the most favorable comparison for nuclear, since any varying demand enables storage to serve double duty (smoothing variations in both supply and demand). For reasonable cost assumptions, particularly for when any nuclear plant started today could come online, renewables beat nuclear. It's important to use more than just batteries for storage; batteries are not well suited to long term storage.

https://model.energy/

Don't misunderstand: I am a huge proponent of solar/wind + storage and grid interconnectivity. But you have completely talked past my point.

> Levelized cost already takes into account the capacity factor of the sources. On a per-peak watt basis, PV is 10x cheaper than nuclear.

You sound like you know what you are talking about, so I presume that you know that one of the limitations of Levelized cost is that it doesn't control for "time effects" which was the entire thrust of my comment.

> We can look at the cost of providing steady power (synthetic baseload) from renewables + storage

Yes, that's what I suggested is a fair comparison.

> any nuclear plant started today could come online, renewables beat nuclear. It's important to use more than just batteries for storage; batteries are not well suited to long term storage.

There are a lot of unstated assumptions/conditions here. Do you have any references I can read?

Let's not get ahead of ourselves, they're not building it yet, they are hoping and planning and continually increasing costs as it gets closer to time to build:

https://ieefa.org/resources/eye-popping-new-cost-estimates-r...

And we will see what cost increase happen once they start putting shovels in the ground. That's where nuclear really starts tk have its big prices increases, usually.

The Rolls-Royce SMR is pitched as exactly that, based on the submarine plants they already build.

No takers yet as far as I know.

At some point the pressure vessel gets small enough that you don’t need a supersized hydraulic press to make it

https://world-nuclear.org/information-library/nuclear-fuel-c...

GE’s first BWRX300 is expected to be made in Canada which is not on the list of countries with huge presses and might help with manufacturing throughput.

> Why don't they build nuclear reactors that can be made on a factory assembly line, shipped on the back of truck, and installed right where it's needed in a town?

It's very difficult to get a permit.

Are you under the impression that nobody has had this idea before?

Couldn’t you also suspend them over a deep concrete lined pool of water and just drop them if there are any major problems?

Sure until it boils away all the pool water and melts through the concrete and creates an ecological disaster

Many of the newer reactor designs tout passive control mechanisms as their major selling points. We should be wary of claims of "idiot proof" designs, as we can always seem to be able to find a bigger idiot, but the risk of runaway reactions is far less with these newer designs.

Reply on news.ycombinator.com