Hacker News Reader

Comment by _ph_ | Hacker News Reader

The huge problem with nuclear vs. natural gas is, that nuclear power plants have only a limited range, in which the power output can be regulated and they are not too quick in that either. So you need something else you can power switch quickly to maintain grid integrity. Natural gas is great for that, as the plants are very quick in response and can be regulated between 0 and 100% and back.

Nuclear is great to replace coal, but you still need the same amount of e.g. storage to complement changing net load and production (renewables).

View on HN

Nuclear fuel is cheap enough that it could be economical to simply [0] discard the unneeded production.

[0] It’s not that simple, but dump loads exist, as do mechanisms to intentionally waste power inside a plant without producing electricity. For example: https://www.thermal-engineering.org/what-is-turbine-bypass-s...

The other capital and operational costs stay the same even if you just discard the production you don't need, so just focusing on fuel costs is not very helpful to determine if it's economical to do that.

Financial profitability is not a requirement for infrastructure projects. If it is the best way to supply power in the most sustainable way, it should be subsidized by governments to make it happen. If you want to pretend it balances on paper add eco taxes elsewhere, but again thats not necessary.

It is not that easy, we have privatised the energy markets so there are rules that need to be followed. Ignoring cost is not something governments can or should do, there are many more factors that need to be taken into account and this combination is what should be optimized for instead of just one factor. The most sustainable way to supply power is to reduce the need to supply it in the first place, insulation and other energy saving measures are usually subsidized already.

My point is that, once you’ve paid the capital cost for, say, 1GW of nuclear, the marginal cost of energy produced is low and you can adjust the production on fairly short timescales. I’m not saying you should, but the calculation is different from a natural gas or coal plant in which the marginal cost of energy is much higher.

The OpEx is coal < nuclear < gas [0]. This is for regular nuclear plants, SMRs will have higher OpEx. Not sure how they'll fare against gas exactly. The SMR handbook has some estimations I believe, but I don't recall the numbers.

[0] https://www.e-education.psu.edu/eme801/node/530

Sounds like a case for Bitcoin ASICs if I've ever heard one . . .

> Nuclear fuel is cheap enough that it could be economical to simply [0] discard the unneeded production.

Just run wind turbines backwards, easy!

This is the actual plan for renewables, because they are just cheap period (though their fuel is pretty darn cheap too). Demand response, simply dumping the excess, storing excess as heat etc.

It's not a good idea for nuclear, because even nuclear that runs all the time can't compete on cost with renewables and any deviation from that schedule destroys the economics conpletely.

A small amount that runs all the time is probably a good idea in places far from the equator, but even then the plummeting costs of renewables opens a few other doors, like shipping green hydrogen around.

Right, any dollar diverted from renewables + storage to anything else, whether nukes or gas, brings climate disaster ever nearer.

24 GW of renewables in my country are not delivering even 1 GW when demanded currently.

That's not demand response. That's failure.

Can machines for extracting CO2 from the atmosphere have how hard they are running be ramped up and down quickly? Or do they just have a rate they work at, where changing how much one is doing would require changing how much machine there is / which ones are active ?

Machines for removing CO2 from the air need consume no power at all. (E.g., a big fabric corral that waves slop into, connected to a fabric tube 10 m across that opens 1000m below.)

Machines to synthesize H2 and NH3 could be ramped up and down, but in practice will run 24/7 except when price peaks and stockpiles are burned for power instead.

Small modular reactors may improve on that. Molten salt reactors in particular are supposed to be able to adjust to demand within a minute or so.

Another possibility, for any sort of high-temperature reactor, is to use molten salt as energy storage. Heat the storage salt directly from the reactor, and put the turbine after that, so you don't have conversion losses before storage. Terrapower is working on this with their Natrium project.

I've seen non-photovoltaic solar collectors that heat molten salt and then store it for off-peak use. I wonder if they could be hybridized with small modular MSRs.

Those are no longer cost competitive with PV and batteries.

But a lot of that is the cost of concentrated solar.

On the PV side, batteries are not known for being especially cheap. Lazard puts the wholesale cost of PV plus four hours storage at 8 to 14 cents/kWh, as of 2020: https://www.lazard.com/perspective/levelized-cost-of-energy-...

And four hours storage is just what you need to shift peak daytime sun to peak evening demand. You need more to get through a windless night. To cover reduced supply on cloudy winter days you need some combination of overcapacity and long-duration storage, which typically involves some kind of turbine.

Batteries are the most expensive storage. Storage cost is plummeting even faster than solar or wind ever did.

>Molten salt reactors in particular are supposed to be able to adjust to demand within a minute or so.

Nobody is making these commercially and the SMR companies out there are making standard pressurized water reactors afaik.

A bunch of companies are developing molten salt reactors commercially, including Terrestrial Energy, Moltex, Elysium, Seaborg, Thorcon, and Terrapower.

And if a fraction of the funding that was dumped on frivolous "renewables" in the last 20 years (Solara? Ha - tip of the iceberg for waste!) was instead spent on new reactor designs we wouldn't even be having this discussion.

Which is probably why nuclear continues to be demonized. Follow the money...

The last plant built in the US was the Watts Bar Unit 2, estimated cost around 6.1B USD for ~1.1 GW of output and construction took over 40 years. Average cost for Solar is something like 2-3$/watt. Follow the money indeed.

Solara failed purely because cost for the competition dropped well below theirs. It was an honest business failure, like many others; and ultimately good news for the rest of us. They burned through $1/2 billion. Nukes always burn much more; US subsidies are more, every year, for equal value in the end.

Harping on Solara only calls attention to nukes' overwhelmingly worse performance. Nukes are "demonized" because they cost overwhelming more than alternatives.

Commercially usually means they are actually outputting things.

Lftrs have been hyped since like 2010 and I don't think any new molten salt reactor has made any net power on the grid since then. From what I could tell of the companies you listed not a single one even has a demo reactor and the earliest ones are slated for years from now.

So you wouldn't define pre-release tech startups as commercial? Ok. They have investors who hope for profits, call them what you like.

None of these projects are LFTRs. They're generally uranium-fueled, which retains the safety and cost advantages of MSRs without the radioactive chemical engineering of LFTRs.

Terrestrial Energy, to take one example, has been doing well with regulators in Canada, and plans to operate a 390MW plant in 2028: https://www.powermag.com/terrestrial-energy-launches-390-mw-...

So yes, years away, but not that many years.

Well forgive the pessimism but these claims feel awfully like "Fusion is just 30 years away" which people have been saying for 50 years.

That's handled by grid-scale storage: Batteries, pumped storage, compressed air.

Renewable energy has the same need.

BTW, when I was a teenager I got to go deep inside of a mountain to tour a pumped storage facility built to pair with some early 1970s nuclear power plants. The nuclear plants are long-decommissioned, but the pumped storage is happily running with renewables.

The general preferred choice for energy that can quickly turn on and off (in terms of weather changes that cause low production in wind/solar) is hydro. As long you don't deplete the water reserves it can be turn on and off fairly fast, through there can be considerations downstream. Nuclear can help in keeping those reserves high by reducing demand and letting hydro deal with the peaks.

Using natural gas is non-viable. It funds war as has been fairly obvious lately, and it puts a dependence between those that has the natural resource and those that don't. The climate can also not continue to take the continuous damage that burning/leaking natural gas causes.

The choices that we have is either nuclear, renewables, and batteries. The natural gas plants need to be replaced and so one can either choose nuclear, renewables or batteries.

It’s not really that huge of a problem to have more energy than needed. It just takes additional engineering to put the excess to use. Having a national grid it can be moved around. Or used for energy intensive uses that can be useful elsewhere. Kind of like the flare gas oil rigs burn when those gasses aren’t economic. I’d like to see us have that excess power pull useful gasses out of the atmosphere or sea water, for instance. Done right it could act as a hydrogen source for hydrogen fuel cell vehicles, for instance. Or power a national water grid. Tons of uses for more power than needed.

It's not a huge engineering problem. But it is a huge problem, lower the rates when you have excess power and we will have more uses for the energy. But if the rates are low the revenue is low.

(This would not be so significant a problem if the capital costs weren't so high)

This is not true. Modern nuclear plants can load follow.

The reason they usually don't is because fuel is such a small portion of the cost of operating a nuclear plant. It's comparatively more efficient to run nuclear at 100% and idle fossil plants. But nuclear can absolutely load follow if there's a reason to, such as load sharing with renewables.

You will find further information on this web site: https://en.m.wikipedia.org/wiki/Load-following_power_plant

If they load follow, their cost per joule produced becomes even worse than it already is.

Anyway there will be no need to load follow, because load producing H2 and NH3, and pumping hydro and compressed air, can follow price, instead.

Only in a very limited way. As we move forward, we are quickly reaching the point where for more and more time, renewables can carry the whole grid load alone. Nuclear power plants would have to be switched off in those times. And while they can follow the load over a large part of their power output, they can't be swithed off and on again. So as the transition to renewables progresses, nuclear power plants are less and less a good counterpart. Gas plants however are. They are very quick and switching them off is possible and saves precious fuel. As the transition progresses, they should use less and less fuel as they are used more rareley and the fuel could be switched to renewable sources (bio, power2gas). The pre-war plans of the current German government included extending the gas infrastructure but with a clear ability for new installations to be also able to use H2 as it becomes available.

Seems like a perfect pairing for direct air carbon capture systems, which are primarily energy limited.

No, you just build extra for the peaks and use the low-demand period to drive capture of carbon.

I think the real problem there is carbon capture is very new and often still on the drawing board. But definitely one way to get carbon capture working is to start with whatever we’ve got.

Just need Bitcoin miner’s next door to regulate output

Reply on news.ycombinator.com