Hacker News Reader

Comment by KennyBlanken | Hacker News Reader

Exactly. People who shout about how wind/solar are not reliable and need storage / fossil fuel backup and then spout about nuclear have no idea how grid-scale energy generation works.

Nuclear takes hours to days to respond to load change, so there has to be much more reactive power available. And nuclear plants need to be able to draw megawatts of power from the grid any time they're not generating power. You can't "cold boot" a nuclear power plant without a massive external power supply and it takes days to weeks to do so.

It's so bizarre hearing people advocating for nuclear talking about storage like it's some barely-understood, dangerous, hideously expensive sci-fi pipe dream. Grid-scale energy storage is almost 100 years old and battery based systems have been actively deployed for many years at all levels of the grid, right down to residential.

View on HN

The entire United States has something like 3 seconds of energy storage, total.

Just getting that to an hour means a three order of magnitude change in the amount of connected grid storage. Good luck finding that much pumped hydro! The easy spots are taken already!

An hour is not enough, of course. That might handle some small variations in wind only. Of course, one diurnal solar cycle means you need to store power for 8-12 hours. That is another order of magnitude. So in order to do this, we need to increase our battery capacity by four orders of magnitude. Multiple days? Weeks?

Now consider that the demand curve will become even more extreme, as climate change increases demand for electric heating and cooling. And consider that you will also need to account for electrifying the cars on the road (because building out storage like this is a multiple decades project), which will place additional capacity demands on the grid.

The thing with storage is not that it's difficult, it's that it's big. Like the scale is you need to nationalize your extraction industries and turn your economy into a wartime economy whose singular purpose is producing batteries big. It's not a Tesla or Google or even a DOE scale problem.

Well, 3 seconds if you turn everything off, but it is not what we are talking about here, especially if the question is peak shifting. The majority of peak shifting is done at local level, with solution like, for example, domestic wall battery like Tesla or electric-vehicle-to-grid. Still not ideal, but I don't think your depiction of the situation is very fair if it presents it in a totally unrealistic way.

I also don't understand your argument about solar cycle: solar is producing close to peak hours, while nights are non-peak hours. So, in practice, it's not like you need to store the electricity locally for more than a couple of hours. I guess you are talking of a situation where the grid is 100% solar, but that's obviously a ridiculous scenario that no one has never pushed for.

I agree with the fact that the demand curve will become more extreme, and it is exactly the reason of my intervention here: this is a complicated situation, and it does not matter if the grid is mainly nuclear or mainly renewable, it is still problematic.

I agree that the storage problem is tricky and that the scale is a challenge. But this challenge still exist in full nuclear grid, unless we end up with a lot of idle nuclear plant running only few hours per day.

The United States added more battery energy storage to the grid in 2021 than the United States had at the beginning of 2021. The situation here is one of rapid transition.

> Nuclear takes hours to days to respond to load change

Days? Seriously? You know that is not true, don't you? The French nuclear power plants are an existence proof of that.

Nuclear does need to be supplemented with something else with low construction/fixed costs for handling the rare, extreme peaks. That will most often be some sort of gas-powered plants. This has nothing to do with nuclear powers "inability" to "respond to load change" and everything to do with the cost of building enough nuclear power capacity that will almost never be used.

On the other hand, we are still very, very far from being able to do really large-scale battery storage. We don't have the manufacturing capacity for the enormous number of battery cells needed and even if we did, the cost would be prohibitive. This will of course change for the better in the future but that is how things are for now.

I thought French nuclear were only doing base load, the load following part being done by other sources.

I believe the French facilities have the capacity of doing load following, but are not doing it because it is less efficient (and therefore more expensive). So I would also not say "days", but it is also true that one cannot directly take a situation where nuclear has, so far, being used only for base load and pretend that load following will work exactly the same way.

I'm not sure why you say "rare, extreme peaks": the current demand is non-flat every day, with a difference between the trough and peak that is typically ~50% of the peak.

I agree that large scale battery storage is challenging. Maybe the solution would be to have some storage and some "wasted efficiency" by using nuclear plants in load following and turning them in stand-by every day during off-peak hours.

Correct, French nuclear is base load, but it still does a non-trivial amount of load following to e.g. absorb peaks. Check out [0] for some cool live graphs!

[0] https://www.rte-france.com/en/eco2mix/power-generation-energ...

Nuclear plants can respond to changes in reactive power demand in ~100ms. It is a function of the voltage regulator and not the prime mover. Perhaps you didn’t mean reactive?

Reply on news.ycombinator.com