[cauch]

Demand following for nuclear is possible (after all, if you produce 10X but the demand suddenly drops to 7X, what you can always do is to "dump" 3X worth of steam instead of injecting it in the turbine), but because the cost of nuclear is mainly upfront, it is not cost efficient at all.

If it costs 10X dollars upfront to build a nuclear central that can produce 10X energy, then if you run it at 100%, it will cost 1 dollar per 1 unit of energy. If you follow the demand, you will not produce 10X, but let's say to illustrate maybe 5X, and it will cost 2 dollars per 1 unit of energy.

You are right about storage as a way to help with demand following, but if you build enough storage capacity, then you basically have solved "for free" a big part of the problem linked to the intermittence of renewables. In this case, you have the choice between building an expensive nuclear central and a distributed cheaper renewable generation.

I'm not saying it demonstrate renewables are better, but that it is true that nuclear is not the obvious winner it looks like before we look into the practical details.

[panick21_]

No, nuclear storage needs to be optimized for 1-2h peaks, if you build a renewable system you need, much, much more. And you have much more localized peaks and valleys depending on weather and such.

So basically, you can put some battery next to every nuclear plant and otherwise use the same grid.

For renewable you need a much more complex grid with much more battery.

[cauch]

Why do you say "optimized for 1-2h peaks"? The typical electricity demand has a trough during the night for few hours, then a first peak in the morning, than a second less-deep trough in the middle of the day (sometimes with a bump around lunchtime), than a second bigger peak in the afternoon for 3-4 hours. And of course this varies with season, day of week, regions, ...

Not sure why you are saying that renewable you need "much more battery": the overlap of generation means that you already have a "baseline" of generation even with renewables (sure, I know about dunkelflaute, but they are as frequent as unexpected shutdown of nuclear site), and therefore in both case, the game is mainly to "move the peaks around" which requires about the same capacity.

Not sure why you are saying the renewable led to a much more complex grid either. Sure, with a naive simplified grid, nuclear works well. But in practice, the modern grid is complex, and adding more nuclear does not really reduce the complexity.

Also, nuclear or not, having EV or heat-pump will be needed for decarbonisation, and therefore the demand becomes even more complex. With EV and heat-pump, local solar+battery is also a smarter choice. So it means that some storage will need to be built on the consumer site directly, even with nuclear.

I'm not saying that in some situation at the end of the computation, nuclear is not the best option, but it is not at all as simple as having a clear winner. Also, the reality is that you need to work with the uncertainties, so it is not like one solution has a score of 75 and the other has a score of 70, so the first is the obvious choice, it is more like one solution has a score of 75 +- 15 and the other 70 +- 5 (or even asymmetric errors), so you cannot directly conclude the first solution is the best. I think the conversation would be way more healthy if we could just avoid over-simplify into a pro-nuclear vs. pro-renewable partisan battle.

(also not sure about "you can put some battery next to every nuclear plant and otherwise use the same grid", why is this not true for renewable too? Just compute the average production of the site, and put storage that will charge when the site produces more than the average and discharge when the site produces less, and you get the same situation as the nuclear site (they may still have period of no generation, the same way nuclear sites have unexpected shutdown). Especially that with a renewable site, the cost is lower so the site owner can invest more in storage and manage it themselves: storage is part of the black box, the grid does not need to know, stay the same and no complexity is added)

[SoftTalker]

Renewable (solar for example) needs enough battery to cover the 12-18 hours a day where the sun isn't shining or is at too low an angle to effectively capture by solar panels. This is far more than covering a few hours of load peak. Solar might actually be a decent peaking power source, as it tends to peak at the same time as demand. But it's not a good base load source; it needs far too much battery buffer.

[cauch]

My point is that to follow the demand, you need ~12 hours of capacity: the trough is during the night around 1-4 am, the peak is mainly in the afternoon around 5pm. So if you have a battery for nuclear, you will charge it during the night and discharge it ~12 hours later.

Not sure about your 12-18 hours, it looks like you want to use the energy during the night, while it is the trough and does not require energy.