Comment by illiarian | Hacker News Reader

Comment by illiarian | Hacker News Reader

Parent Full thread

> Demand isn't flat and nuclear can't scale quickly enough to meet the daily peaks and troughs.

Of course it can. https://www.oecd-nea.org/upload/docs/application/pdf/2021-12...

--- start quote ---

Modern nuclear plans with light water reactors are designed to have strong manoeuvring capabilities. Nuclear power plants in France and in Germany operate in load-following mode, i.e. they participate in the primary and secondary frequency control, and some units follow a variable load programme with one or two large power changes per day.

... according to the current version of the European Utilities Requirements (EUR) the NPP must at least be capable of daily load cycling operation between 50% and 100 % of its rated power P. with a rate of change of electric output of 3-5% of P, per minute.

Most of the modern designs implement even higher manoeuvrability capabilities, with the possibility of planned and unplanned load-following in a wide power range and with ramps of 5% P, per minute.

(Where P is rated power)

--- end quote ---

So you're overbuilding nuclear, and then throwing the extra power away when you don't need it with (at best) no savings on fuel or running costs? That doesn't sound cheap.

Wouldn't it make sense to take some of that power and make hydrogen or something that can be stored for later peaks? Then maybe you can close the nuclear plant you only turn on 1% of the time and fill that peak with hydrogen instead and save a lot of money.

Maybe offer cheap rates at night so EV users fill up on cheap nuclear power when it's available. Or store heat in water tanks or brick. Or use some of that power to pump water uphill like nuclear plants have been doing for half a century already.

You could do like France and schedule refueling during the summer when the demand is lower and taking a plant offline has less impact.

Lots of boring sensible things you can do to match nuclear supply to demand.

But of course if you admit these things exist, it kind of puts a big hole in your argument, so best to pretend they don't happen and are only needed for renewables.

> So you're overbuilding nuclear

So are complaining about changing demand or about overbuilding?

> Wouldn't it make sense to take some of that power and make hydrogen or something that can be stored for later peaks?

Yes, you could probably do that if any if those solutions scaled in any meaningful shape or form.

> But of course if you admit these things exist, it kind of puts a big hole in your argument

My argument was that your claim that "demand isn't flat and nuclear can't scale quickly enough to meet the daily peaks and troughs" is a lie.

None of the weird tangent you went off on makes a hole in this argument: nuclear power can scale quickly and is literally used to do that right now.

The fact that it isn't used for something that you want it used for is completely orthogonal to whether it can be scaled or not. For example, no one stores nuclear power in hydrogen or whatever because it's literally not needed because nuclear power can be scaled quickly on demand.

> My argument was that your claim that "demand isn't flat and nuclear can't scale quickly enough to meet the daily peaks and troughs" is a lie.

That was someone else's argument, and the second part is kind of irrelevant, but it's also true.

The speed that Nuclear can ramp, even in the best case is not enough to match demand spikes. It doesn't ramp as fast as gas plants and gas plants don't ramp as fast as batteries.

So even if you ramp nuclear, at an increased cost, you still need gas or batteries or hydro for fast ramps. So you might as well use the power you ramp down to fill batteries, make syngas or hydrogen or fill pumped storage.

Just like you'd do with excess renewables.

> The speed that Nuclear can ramp, even in the best case is not enough to match demand spikes.

Because Nuclear isn't 100% of all power output. Of course it can't ramp up from, say 20% to 100%.

And yet, it does ramp up quickly.

> It doesn't ramp as fast as gas plants and gas plants don't ramp as fast as batteries.

5% of rated power per minute is plenty fast. Not as fast as gas plants, true. And for batteries... There's this funny little things of: there are literally no batteries at scale required by demand spikes. And won't be in any near future.

Meanwhile if you followed the links, you could see the graph showing how nuclear power plants go from ~650 MW to ~1400MW to meat demand (which is mostly predictable day to day).

How many batteries you need to have extra ~800 MW per plant every day? And how many renewable sources to provide that excess?

(Edit: largest energy storage seems to be Ouarzazate Solar Power Station with 510 MW of storage, thermal storage. Everything else is lower. To keep up with a single nuclear plant in the report you'd need two of those)

> Just like you'd do with excess renewables.

Except, for "excess renewables" you need to both overbuild them significantly more, and provide significantly more batteries because besides fast load following (which they can't do, and nuclear can) their base load is literally zero (and it's never zero with nuclear unless you take the plant completely offline).

Spikes occur at various timescales. There are seasonal spikes (due to heating/AC), daily spikes (more power in morning and evening), and spikes on the millisecond to minute level due to random chance. Nuclear can't deal with the short spikes, and while it can deal with the long spikes, it's not efficient to do so since roughly 100% of the costs of nuclear are initial construction and personnel so running at reduced power levels doesn't save any cost.

Batteries already exist that handle demand spikes on the shorter timescales (e.g. in Australia). and hydro is currently the best for medium timescale storage (that said there's a bunch of R&D currently into better grid batteries since they have very different performance requirements than portable batteries since weight and volume don't matter, just cost and cycles).

Nuclear is pretty much the worst tech to overbuild because it's already expensive and it's price per kwh goes way up if you use it infrequently.

> Spikes occur at various timescales. There are seasonal spikes (due to heating/AC), daily spikes (more power in morning and evening), and spikes on the millisecond to minute level due to random chance.

The first two are easily handled by nuclear (which you'd know if you read the report).

Literally nothing can handle "millisecond spikes due to random chance".

Renewables like solar and wind can't really handle any of the three because their power generation is intermittent. So you have to overbuild both them and their storage.

> Batteries already exist that handle demand spikes on the shorter timescales (e.g. in Australia).

Of course they don't exist. Not anywhere near the scale required.

> and hydro is currently the best for medium timescale storage

Ah yes, hydro. That is so easily available and easy to build just about anywhere, and on the required scale.

> Nuclear is pretty much the worst tech to overbuild because it's already expensive and it's price per kwh goes way up if you use it infrequently.

Of course it's the best tech to "overbuild" (because you don't need to overbuild much).

Once again:

- it's the highest energy density we know

- it's stable power generation (not intermittent like most renewables)

- it requires significantly less storage (because it can scale up quickly: most renewables can't even handle daily power spikes)

- it requires significantly less area to build

- the costs for renewables somehow never include the costs for the required overbuilding and the costs for required storage.

Again: in the report you can see just one of the power plants easily scale from 650MW to 1400 MW during the day. And that's for power plants built 45 years ago. What are your renewable + battery solutions capable of that, and their cost?

> The first two are easily handled by nuclear

I never said or implied otherwise.

> Literally nothing can handle "millisecond spikes due to random chance".

Batteries can (see https://energytransition.org/2015/06/batteries-stabilize-the...)

> Of course they don't exist.

Wrong. https://www.cnbc.com/2021/12/08/australia-switches-on-victor...

> That is so easily available and easy to build just about anywhere, and on the required scale.

I know, right! It's so cool! https://www.energy.gov/eere/water/hydropower-program

> because you don't need to overbuild much).

Based on variation of demand, you need to overbuild any source by at least 50% https://en.wikipedia.org/wiki/Electricity_sector_of_the_Unit...

> it's the highest energy density we know

Doesn't matter at all. We don't pay by land used, we pay by total cost.

> it's stable power generation

True. That's why it makes sense to have for baseload (but is stupid to overbuild)

> it requires significantly less storage

Not really since no one would ever build a purely nuclear grid because it would cost way too much.

> it requires significantly less area to build

Again. Irrelevant

> the costs for renewables somehow never include the costs for the required overbuilding and the costs for required storage.

That's because the amount of overbuilding and storage needed depend strongly on a bunch of factors and aren't that useful.

Using https://www.lazard.com/research-insights/levelized-cost-of-e... as a source for LCOE numbers by power plant type (there may be a better one). Solar has LCOE of (very roughly) $70/MWH while nuclear has LCOE of roughly $175 if you ignore nuclear's decommissioning and operational costs (which are both significant and would almost certainly push the figure up to over $200).

This yields a pure nuclear grid cost of approximately $85B. Alternatively, you can overbuild solar and wind by 3x and have a couple days of storage (which is doable with hydro), and you come out ahead.

Reply on news.ycombinator.com