The duck in the room – the end of baseload | Hacker News Reader

The duck in the room – the end of baseload

jeromeaparis.substack.com

38 pointsguerby117 comments

Not a friend of nuclear, but I think OP's first point (duck curve) and second point (no place for baseload) contradict each other somewhat.

Or more specifically, why couldn't you convert any baseload plant into a "flexible supply" plant by adding storage?

You can interpret the post in two ways: Either, solar capacity will grow to such an extent that eventually it will be enough to cover all demand: Meaning, during summer days, there will be enough supply to cover the immediate demand and fill up storage enough to cover all evenings, nights and winter days. If that were the case then there wouldn't be the need for any power sources except solar and storage and we'd have basically solved energy. Hooray!

However, that seems pretty unrealistic. The second interpretation of the article is that solar will grow (and will be able to contribute to storage to some extent) but will not be able to fully cover the dark hours - hence the duck curve. For those times, other energy sources are still needed.

However, if there is still a need for other energy sources, then why couldn't baseload plants cover that need? Yes, it's only for a few hours per day (in summer), but then the baseload plant is needed the rest of the time to fill up storage - as solar, by initial assumption, would not be sufficient to do that.

> Or more specifically, why couldn't you convert any baseload plant into a "flexible supply" plant by adding storage?

Cost. New built nuclear costs $0.12 - 0.2/kWh [1], already completely uncompetitive. Now add storage on top of that.

There are ideas with for example salt based heat storage, but they all boil down to added cost on something uncompetitive if it ran at 100% all year around.

[1]: https://www.lazard.com/research-insights/levelized-cost-of-e...

Would that still hold if we got serious about phasing out gas-powered plants as well? (As we will probably have to at some time due to co2 emissions)

I could argue that nuclear power isn’t that expensive when I can buy nuclear power at retail for 9 cents per kWh, but that isn’t really an interesting point.

I’m more interested in learning where all the storage that’s going to be needed to decarbonize is going to come from.

According to Woodmac’s most recent grid storage update, there was only 13.4 GWh of grid storage in the US 2022.

Projected growth for the next five years is 10-17 GWh per year with a declining growth trend.

> I could argue that nuclear power isn’t that expensive when I can buy nuclear power at retail for 9 cents per kWh, but that isn’t really an interesting point.

Paid-off nuclear plants are in the same range as new built renewables in cost. They don't magically appear out of thin air though, someone has to pay to build them.

A brand new nuclear power plant was just taken into service here in Finland.

As a result electricity prices dropped about 5 cents per kWh across the board.

However, as noted before, this isn’t really all that interesting. The storage part is the big interesting trillion dollar question.

Of course, if someone puts up the money and eats the loss you get cheap electricity.

The problem is, someone has to eat the loss. People proposing nuclear power generally have no answer regarding who this "someone" is rather than generic handwaving about "the grid!"

> However, as noted before, this isn’t really all that interesting. The storage part is the big interesting trillion dollar question.

For the Nordics this is already solved with the existing hydropower. Sweden as of this week stores 19 TWh in their reservoirs. This is the reason no storage gets built in the Nordics, it is impossible to compete with scheduling the release of water more in line with renewables rather than daily consumption cycles.

> Of course, if someone puts up the money and eats the loss you get cheap electricity.

The power companies paid for the nukes and the French paid for botching the job.

> The problem is, someone has to eat the loss. People proposing nuclear power generally have no answer regarding who this "someone" is rather than generic handwaving about "the grid!"

The end user always pays in the end.

Around here people will rather pay for working power rather than freeze to death.

> For the Nordics this is already solved with the existing hydropower.

Um, no. Even if you aren’t from around here these things are easily looked up on Wikipedia.

Norway is the only Nordic country with enough hydro, but even they, like Sweden, are lacking in north-south transmission capacity.

Iceland is too far away.

Denmark has no hydro.

Sweden only has enough for 50%.

Finland only has about 20%.

To further point out the obvious, Finland has to import more than a gigawatt during peak power usage, even when using all generation capacity in the country!

> Sweden as of this week stores 19 TWh in their reservoirs. This is the reason no storage gets built in the Nordics

Um, no. Again.

The reason more hydro isn’t being built is because there is nowhere left to put more off it!

Plus, in Sweden’s case, lack of north-south transmission capacity. Which nobody is building.

So, tell me again, why isn’t more grid storage being built out in the Nordics?

French taxpayers are footing the real cost:

https://www.world-nuclear-news.org/C-Olkiluoto-3-EPR-parties...

> "In the event that the supplier consortium companies fail to complete the OL3 EPR project by the end of 2019, they will pay a penalty to TVO for such delay in an amount which will depend on the actual time of completion of the OL3 EPR project and may not exceed EUR400 million,"

Note the article is from 2018

> The latest schedule sees grid connection taking place in December, with the start of regular electricity production in May next year.

Whoops.

Again, the nuclear part isn’t the interesting bit. It’s the storage part that is interesting and keeps being ignored in these comments.

Yes, the French had to pay penalties and cost overruns. Even so that only doubles the system costs from 4 cents to 8 cents.

I think what you are missing from this discussion is that both renewables and nuclear require dispatchable power to fill the gaps. If storage is "impossible" then a nuclear grid is impossible unless you overbuild it leading to costs multiples higher than Ukrainian war gas crisis costs.

The problem is that renewables and nuclear are economically incompatible, like the article goes into depths about. Renewables easily win this battle as the cost for new built renewables are in the same range as operations and maintenance for paid off nuclear plants.

Generally, the research available see no issues building 100% renewable grids, so I do not know why you keep hampering on about it being impossible?

https://ieeexplore.ieee.org/document/9837910

>The result is a holistic vision of the transition towards a net-negative greenhouse gas emissions economy that can limit global warming to 1.5°C

The research in question ignores the fact that we've already passed that level.

>However, a clear deficit and research gap exist, since a detailed description of the industry sector, i.e. separated major industries such as cement, iron and steel, chemicals, aluminum, pulp and paper, etc., is lacking in almost all cases. Therefore, a full defossilization of the non-energy feedstock demand of the industry sector has not been modelled in global 100% RE analysis.

The research in question conveniently says "we don't really know how much industries use, but surely it can't be that bad right ? Anyways, we're not taking it into account"

>The industry sector is described in detail in Pursiheimo et al. [145], though the authors admit that TIMES, the model used, was not capable of applying full power-to-X functionality for the industry sector, thus fossil hydrocarbon inputs to the industry sector were still required by the model. Similarly, Teske et al. [125] and Luderer et al. [146] mention that the chemical industry is still fully based on fossil fuels.

The research in question confirms that the whole industrial sector still requires massive amounts of hydrocarbons.

Overall, all these full renewable grids papers assume one critical thing: a complete change of all of our infrastructure, to add storage, better grid organisation, and overall a complete rework of... everything. While that's a lofty goal, looking at the reality where we can't even phase out simple things makes it quite unbelievable. Everything is possible if you ignore reality.

> I think what you are missing from this discussion is that both renewables and nuclear require dispatchable power to fill the gaps.

Well, obviously. Either you overbuild generation capacity or you use dispatchable generation. This is 101 stuff.

Why would you assume I would be overlooking such an obvious thing?

> If storage is "impossible" then a nuclear grid is impossible unless you overbuild it leading to costs multiples higher than Ukrainian war gas crisis costs.

That’s an interesting comparison. How did you come up with the cost comparison?

A nuclear grid isn’t “impossible”. We already have a nuclear grid. What we don’t have is a 100% nuclear grid.

> The problem is that renewables and nuclear are economically incompatible

Yes, because of the rules imposed on the electrical market.

Begin by changing the rules so that they stop disincentivizing running nuclear at 100% all the time. Variable renewable sources should not be must take. Nuclear should be paid first and then, if there are any takers, variable renewables.

Non-variable renewables should be free to eat nuclear’s lunch.

> Renewables easily win this battle as the cost for new built renewables are in the same range as operations and maintenance for paid off nuclear plants.

Only because must take rules and renewables using other generation to compensate for their variability.

> Generally, the research available see no issues building 100% renewable grids, so I do not know why you keep hampering on about it being impossible?

Please stop putting words in my mouth. I have never said or implied that 100% renewable grids are impossible. Clearly they are from a theoretical perspective and small grids have been built using 100% renewables.

What I am specifically asking is, how are nation size grids going to be decarbonized, i.e. turned into 100% renewable grids in practice and what timescales are required?

As far as I can tell we either need a boatload of new nuclear, storage and/or megaprojects to build out trans/intercontinental transmission networks.

Storage projections don’t look optimistic.

Trans/intercontinental megaprojects are hard and come with fun spices such as geopolitical risks and massive failure modes.

Nukes are expensive and take a long time to build, but at least we know how to do it.

If we don’t have a plan and know how to build TWh grid storage faster than new nukes, then we should start mass producing nukes right now. At least we’ll have the nukes built eventually.

No one has to take power from renewable sources. They want to because it's the cheapest source of power at the time.

Obviously nuclear power is economically viable if you use price-fixing to make it economically viable, but that is a huge divergence from the way our economy normally operates.

> No one has to take power from renewable sources.

The EU disagrees with you. A certain percentage of power generation is mandated to come from renewables and nuclear is specifically excluded. IIRC there are no means for power companies to refuse feed-in of renewables.

> They want to because it's the cheapest source of power at the time.

The cheapness is mainly due to the artificial construction of the pricing mechanism.

Variable renewable generation should be priced differently from non-variable fossile free generation.

> Obviously nuclear power is economically viable if you use price-fixing to make it economically viable, but that is a huge divergence from the way our economy normally operates.

It's not a matter of price fixing, but rather of a more rational cost allocation.

Variable renewables are free riders on other generation methods.

Obviously nuclear power is economically viable if you use price-fixing to make it economically viable, but that is a huge divergence from the way our economy normally operates.

> Variable renewable generation should be priced differently from non-variable fossile free generation.

Some of the largest costs in grid infrastructure comes from the requirements when nuclear reactors trip.

You're amortising the cost over a time period that is in direct contradiction to reality, at an EAF that is in direct contradiction to reality at a cost of capital that doesn't reflect the likelihood of project failure or early closure. On top of excluding all the indirect subsidies.

€11bn for 1.6GW on a European nuke that will run about 70% of the time and statistically will close in around 30 years while costing $30/MWh for O&M is around 16c/kwh

Since nuclear fuel cost is very low and maintenance/lifetime is time based nuclear kWh price will be scaled by inverse of load factor.

So if for example 66% of the time solar+wind+hydro+misc provides enough and nuclear is not needed, nuclear kWh will be triple the price (estimated with historical load factors).

Already at play in Finland a few weeks after the nuclear power plant has been turned on it had to scale back its production to "not loose money":

https://oilprice.com/Alternative-Energy/Nuclear-Power/Finnis...

What will happen in a decade with more and more solar and wind?

Hopefully the antiquated must take rules for variable renewables will be replaced with must take rules for non-fossile deterministic power generation.

It's a cost of selling a kwh on an existing market with relatively low renewables penetration and supported by fossils. It's not a true aggregate cost of an energy system dominated by, say, solar.

Imagine we turn of all fossil generators tomorrow, forever. Solar panels will still be cheap, but instantaneous energy price in California will tend towards infinity every night, making solar-based generation an extremely expensive solution. That expense is not captured in Lazard's numbers.

Neither do they capture the cost of transmission lines and storage necessary to eliminate blackouts in that scenario. People talk a lot about the time it takes to build nuclear, but transmission lines aren't much faster, and unlike reactors they don't generate on their own.

Lazard has been tracking the cost of storage for a while now (and cost of hydrogen production).

They have a whole section when they compare renewables with battery storage against other "near firm" options.

It's still LCOE, which is by definition focused on generating assets. It doesn't track societal costs or costs offloaded to other entities, like grids or households, that can be attributed to a particular generating technology becoming dominant.

Things like sky high electricity prices during that blackout in Texas a few years ago don't really reflect in LCOE.

That particular firming calculation on page 8 (if that's the one you have in mind) is a bit misleading, they use average capacity instead of historical capacity factors [1], which doesn't make any sense in places with seasonal variability. You need waaay less storage to "firm" an energy source that is active 25% of every day vs one active only during the summer. Time distribution matters, and they do acknowledge it in small print in a footnote, suggesting that they only mean firming the "duck's head" in the duck curve [2]. Which only captures a tiny part of a system that would need to store a whole night (or season's) worth of energy.

[1]: "ELCC is an indicator of the reliability contribution of different resources to the electricity grid. The ELCC of a generation resource is based on its contribution to meeting peak electricity demand. For example, a 1 MW wind resource with a 15% ELCC provides 0.15 MW of capacity contribution and would need to be supplemented with 0.85 MW of additional firm capacity in order to represent the addition of 1 MW of firm system capacity."

[2] "For PV + Storage cases, the effective ELCC value is represented. CAISO and PJM assess ELCC values separately for the PV and storage components of a system. Storage ELCC value is provided only for the capacity that can be charged directly by the accompanying resource up to the energy required for a 4 hour discharge during peak load. Any capacity available in excess of the 4 hour maximum discharge is attributed to the system at the solar ELCC."

I don't really agree with this, though I come to the same conclusion via a different route.

Baseload as an abstract concept still makes perfect sense.

It is exactly equivalent to a flat reduction in demand. Which is something that would be worth paying for.

How much it is worth is the big question. I'd suggest it is worth less than the costs of burning coal or gas as baseload and that therefore those should be used only when necessary and be on standby the rest of the time.

I'd also suggest building new nuclear makes little sense as there's little chance of it being economically worthwhile.

But, current nuclear that's not got other problems isn't any big drama to integrate with increasing renewables. And I don't think we're quite at the point yet where new build renewables are cheaper than nuclear running costs. If anything it just gets you faster to the point where you have a daily bump of excess solar that can be matched with batteries (grid support or in vehicles).

Final shout-out for east-west orientated static solar, which can be used to make the solar curve a bit more of a square wave. They produce less but at times of day when it'll be worth more, once the standard peak is saturated.

Baseload only exists because we let it - partly we created it to soak up the outputs of previous generations of nuclear!

BTW, agreed on E- and W- facing PV widening generation and thus being more useful.

There are plenty of 24h demands. Office/Street lighting, industry, etc.

Baseload is just a product of certain demand.

lighting is by definition not a 24h demand, why are you running your street lights during the day?

It is absolutely a 24h demand for cases like warehouses, hangars, factories: these cannot rely on daylight alone, and account for most of the energy consumption related to lighting.

Idle question, following on from your last para:

Why aren't sun-tracking solar panels in use at scale? Economically infeasible I assume, but is that because of increased hardware costs, increased maintentance costs, or just because they won't add much extra output?

> Economically infeasible I assume, but is that because of increased hardware costs

Economically infeasible because of decreased hardware costs. Panels are so cheap compared to tracking systems.

(Quick envelope calculation: Amazon offered me a 400w panel for £200. Amortize that to about 6h of effective daytime use, that's 21600 seconds or 8kWh/day. 25 year warranty lifetime; derate that by, say, 20% to allow for degradation. 20 years is about 7300 days. That's about 58,400kWh. UK electric costs are all over the place at the moment, but at £0.30 that means .. you're buying £17,520 of electricity for £200.)

NOT INCLUDING BALANCE OF SYSTEM COSTS, solar panels are really cheap.

So if you want to generate more energy, "buy more land" is cheaper than "add tracking"? Seems like a solid answer, thanks.

More (cheap desert) land, more roofspace, more air above parking lots. It's also a matter of reliability, cheap moving parts = downtime.

Tracking does not significantly increase the amount of available energy out of a plot of land. It just reduces the amount of PV panels you need to efficiently utilize it. If you tile land with horizontal PV, all photons that hit that land hit them.

And once PV is cheap enough, the expense of tracking (and especially the maintenance expense of tracking!) to deploy a little less silicon does not make any sense.

https://www.canarymedia.com/articles/solar/a-100mw-solar-far...

I think a while back solar panels got cheaper than the curved mirrors used in molten solar based solar thermal plants.

'Cheap lower efficiency solar cells' became uneconomic when the non silicon cost of the panels became too high a proportion of the total cost. Because half the power per panel at 2/3rd the cost per panel doesn't pencil out.

Your calculation is way off. Getting 8kWh a day from a single 400Wp panel would already require 20 hours of operation at theoretical peak power, so under real-life conditions you can only expect a fraction of that per day.

This doesn't invalidate your point though - even if you're one order of magnitude off, your £200 solar panel will buy you about ten times its price worth of electricity.

Ugh, I did 1kWh = 1000 watt-seconds, which should be 3600 watt-seconds, so the calculations are off by a factor of four.

£0.30 per kWh is also optimistic over the lifetime of the panel.

Around these parts you get the spot price for any solar you generate. Prices can, and do, go negative.

In general prices tend to be low when the sun is shining and/or the wind is blowing.

So, you might even end up paying for the pleasure of providing solar power :)

I have 22 solar panels rated at 370w, I live in Phoenix Arizona where it’s normally cloud free and sunny. I have them facing southeast and southwest. My most productive panel produces about 2.5 KWh a day.

Not that I disagree with the conclusion, but you’d get no where near 8KWh a day from a 400W panel.

Trackers are used at scale but it depends where. If a solar plant isn't intended to maximize output because peak prices can be negative, then there isn't much point. However at locations where something other than price is a driving factor, trackers are common enough at grid-scale facilities.

The energy market systems are somewhat distorting perceptions of power in general and some reliability planners are worried that as older dispatchable generation goes offline for good, rotating blackouts may be in the future for many high demand areas.

The tracking hardware is expensive both in terms of initial outlay and maintenance, relative to the increase in output (which is significant: ~30% or so), when compared with the cost of just adding more panel area. Tracking is useful if you are constrained by the area of panels you have or the area you have to put the panels, but if you are constrained by neither and just optimising for price they lose to just building more fixed panels (in fact, with the rate at which panels are getting cheaper there's even a bit of a push towards not even bothering with building the mounts to orient the panels in the best fixed orientation and just putting them flat on the ground).

Baseload as an abstract concept makes sense.

But the concept of a specific type of power plant that specializes in satisfying the baseload is going away.

The idea that you provide the baseload with dirt cheap but inflexible nuclear and supplement the peaks with expensive but flexible things like coal made perfect sense, while nuclear was cheap.

Now nuclear isn't the cheapest thing around anymore. Which means that while solar is available the economically rational decision is to buy solar.

And this hurts the profitability of nuclear enormously. Now it can't run 24/7, because during the day it has competition. And since nuclear is almost all capital costs, any time it's not providing power is time it's not paying off those enormous loans. And renewables being very variable results in an environment nuclear isn't very well suited for either.

> Now nuclear isn't the cheapest thing around anymore. Which means that while solar is available the economically rational decision is to buy solar.

You are emphasizing one factor/capability: (economic) efficiency. That may not be the only factor worth considering. One such other factor could be reliability, and it may be worth paying more to gain that factor/capability.

Or would you also say: "Well, the load can be handled by a single server (or router), so a second (active-active or active-passive) system isn't worth paying for to get HA."?

In a capitalist system of electricity markets, things which are valuable have to be turned into prices to make them happen.

(France's big nuclear build out was extremely socialist planned economy, only later privatized as EDF)

EDF is a state company owned by French state. https://en.wikipedia.org/wiki/%C3%89lectricit%C3%A9_de_Franc...

For ones too lazy to open the Wikipedia page, it was state owned since its inception.

Founded 1946; 77 years ago

Founder French state under the direction of Provisional Government Minister for Industrial Production Marcel Paul

Nationalized last week due to the hole they dug for themselves using nuclear power.

https://www.reuters.com/business/energy/french-utility-giant...

> In a capitalist system of electricity markets, things which are valuable have to be turned into prices to make them happen.

Some things are mandated regardless of what they cost (reliability regulations, or lack thereof). Just ask Texas:

* https://en.wikipedia.org/wiki/2021_Texas_power_crisis

* https://en.wikipedia.org/wiki/2011_Groundhog_Day_blizzard#Te...

Another factor is the fact that they are working at night. And Europe is further to the north than California, nights has a tendency to get much longer in winter (up to 16 hours of night in Central Europe), so cheap energy from solar panels is instantly nullified by lack of sunshine during winter, when the energy is needed the most.

>(up to 16 hours of night in Central Europe)

Poland gets on average 1 hour of actual sunshine in December. I'll take nuclear power unless somebody really has reasonable strategy for getting 3-7 days of energy storage.

You need an energy storage solution for nuclear too, unless you want to heavily overbuild, which raises it from too expensive to stupid expensive.

And once you have that storage, a mix of renewables is cheaper.

> You need an energy storage solution for nuclear too, unless you want to heavily overbuild, which raises it from too expensive to stupid expensive.

Why do you energy storage for nuclear? As someone who lives in Ontario, Canada, where 8-9 GW of nuclear power chugs along continuously:

* https://www.ieso.ca/en/Power-Data

I'm not sure what needs to be stored. We just bring online hydro and gas as needed (along with having fluctuating wind).

Hydro and gas both involve storage (of water and methane), and both can be, and are, used with renewables and/or nuclear to fill gaps between supply and demand.

> You need an energy storage solution for nuclear too, unless you want to heavily overbuild, which raises it from too expensive to stupid expensive.

No, nuclear lowers baseline you need. If 50% of your energy is produced by nuclear, you just need to build the rest like your country is 50% smaller.

Reliability to most people is a property of the system, not of the power plant.

Meaning, I don't care whether the power I'm getting got made by a power plant that can keep going 24/7, or sometimes gets supplemented by a battery, or there's some constant delicate balancing act going on between 5000 distributed sources.

So as a consumer I don't really care that nuclear is stable and solar isn't, the output of the grid is all that matters.

Yes, I think it's important to say both that the idea of splitting up the demand curve into baseload and variable demand is a reasonable way to look at things when given a certain mix of generation options, but also that it is not some fundamental reality about how the grid needs to operate in the same way that supply and demand always matching up is.

Storage is the functional replacement for baseload. Lack of storage was the reason for baseload generation in the first place

Japan’s vast hydro storage was actually built to solve the issue of too much nuclear baseload generation at night (to be discharged during the day). Later as Japan added a lot of Photovoltaic after Fukushima, the same hydro storage was doubly useful

https://app.electricitymaps.com/zone/DK-DK2

is good for seeing details for each country (or part of country). I'm not affiliated to the site.

There is a lot of different duck patters. For some countries Wind Power has the same duck pattern as solar, for others it's more even.

It is also interesting to see how much electricity is imported/exported during the day to other countries. West European countries are definitely not isolated islands.

I the near future (20 years) with enough EVs we will have enough storage capacity (All cars in Germany EVs -> 2 days of electrical power storage capacity).

Everyone here currently adds local solar panels (which dropped 50% in the last year). Battery packs being no option 4 years ago are now offered for 1kwh household solar panels and dropped in price too (but not 50% yet). We will get one next year to move generated energy into consumption times in the evening (we have no car).

(This is all private consumption of course not industrial power usage)

Industrial power usage should anyway be designed as a function of electricity cost. Old-school always-on use is just not cost efficient (and thus not future proof).

There are certain use cases which kind of have to be always-on. Big freezers. Aluminium smelters.

Venezuela's terribly administered grid shut down power to its aluminium industry. This cannot survive without power to keep the smelters warm; two hours was enough to destroy the industry. https://www.alcircle.com/news/blackout-in-venezuela-brings-d...

Interestingly both of the use cases you mentioned are quite amenable to demand response. Big freezers can be set to a few degrees cooler and be switched off during peak electricity demand. In [1] the researchers even call them "ideal candidates for implementing demand response strategies". For aluminium smelters, you are correct in that they can't be turned off entirely (the cryolite will solidify) but it is often possible to reduce energy demand significantly by stopping the electrolysis reaction but still heating the mixture to keep it molten. Then later, when peak demand has passed you can switch on the electrolysis process again.

[1] https://www.osti.gov/servlets/purl/1237687

To me it reads like the author just handwaves away the problems of solar, the exercise is left to the reader. I also find it contradictory that if we find efficient ways to store/use surplus (solar)energy, why would that make baseload unnecessary instead of profitable? It makes no sense.

> why would that make baseload unnecessary instead of profitable

It is an argument hinging on cost. The cost for new built renewables are in the range of what existing paid off nuclear and fossil plants cost in operations and maintenance.

Try building new ones in that environment.

To a layperson like me, it is very hard to find the truth what is the true TCO when it comes to power plants, numbers depend on who you ask. E.g factoring in environmental impacts of manufacturing/shipping/fuel/installation, subsidies, how well does calculations generalizes across climates/countries/geography etc. I don't think there is a one size fits all solution, also, having diversity in the energy mix is valuable in itself.

https://www.lazard.com/research-insights/levelized-cost-of-e...

https://model.energy/

Use that to play with the numbers at different locations.

> It is an argument hinging on cost.

Well, if you ignore reliability.

To decarbonize without rolling blackouts you need nuclear or storage.

For nuclear this inflexibility comes from pure economics. It is economic suicide to build a new plant and operate it at 100%, now try operating it at less than 50% on average.

Generally, the research available see no issues building 100% renewable grids, so I do not know why you keep hampering on about it being impossible?

https://ieeexplore.ieee.org/document/9837910

> I think what you are missing from this discussion is that both renewables and nuclear require dispatchable power to fill the gaps.

Well, obviously. Either you overbuild generation capacity or you use dispatchable generation. This is 101 stuff.

Why would you assume I would be overlooking such an obvious thing?

> If storage is "impossible" then a nuclear grid is impossible unless you overbuild it leading to costs multiples higher than Ukrainian war gas crisis costs.

That’s an interesting comparison. How did you come up with the cost comparison?

A nuclear grid isn’t “impossible”. We already have a nuclear grid. What we don’t have is a 100% nuclear grid.

> The problem is that renewables and nuclear are economically incompatible

Yes, because of the rules imposed on the electrical market.

Non-variable renewables should be free to eat nuclear’s lunch.

> Renewables easily win this battle as the cost for new built renewables are in the same range as operations and maintenance for paid off nuclear plants.

Only because must take rules and renewables using other generation to compensate for their variability.

> For nuclear this inflexibility comes from pure economics. It is economic suicide to build a new plant and operate it at 100%, now try operating it at less than 50% on average.

Again, this is only due to outdated energy market rules. If

> Generally, the research available see no issues building 100% renewable grids, so I do not know why you keep hampering on about it being impossible?

What I am specifically asking is, how are nation size grids going to be decarbonized, i.e. turned into 100% renewable grids in practice and what timescales are required?

As far as I can tell we either need a boatload of new nuclear, storage and/or megaprojects to build out trans/intercontinental transmission networks.

Storage projections don’t look optimistic.

Trans/intercontinental megaprojects are hard and come with fun spices such as geopolitical risks and massive failure modes.

Nukes are expensive and take a long time to build, but at least we know how to do it.

If we don’t have a plan and know how to build TWh grid storage faster than new nukes, then we should start mass producing nukes right now. At least we’ll have the nukes built eventually.

Unfortunately this is article fails to talk about grid stability, frequency control, inertia, and other important operational factors that keep transmission grids running.

The grid the author is imagining would not survive more than an hour.

The grid that the author describes already IS running from time to time, eg low demand days in sunny parts of Australia.

Renewables and storage can provide the aux services needed, and already are on many grids.

> Renewables and storage can provide the aux services needed, and already are on many grids.

Which grids? I wasn't aware of grids powered entirely by renewables with grid scale storage as the backup, that would be pretty cool if true.

There is an island or two in UK that does that, involves signals in homes to disable various kinds of devices to drop usage whenever capacity falters.

They weren't talking about supplying energy with batteries.

The initial use case for small batteries on grids are providing the grid stability, frequency control, (synthetic) inertia and other important operational factors that the original comment thinks are being ignored.

> Unfortunately this is article fails to talk about grid stability, frequency control, inertia, and other important operational factors that keep transmission grids running.

This 'value stacking', which the immediate parent called 'aux services' makes them much more valuable to the grid than just their delivery of power would suggest.

Though we're now approaching a point in the mass deployment where the batteries have flooded that market on many grids and there's only actual energy storage left for them to do (and maybe avoid some transmission bottlenecks).

Fast Frequency Response is something that can be provided by all inverter-driven systems, but isn't. Once a strategy is agreed it can be retrofitted at very reasonable expense.

What would be the price to have it retrofitted onto the home solar I had installed last year? Unless my inverter firmware can be updated remotely that doesn't seem like a reasonable expense at all.

I don't think it would ever be retrofitted at the individual level, but there are plenty of bigger installations.

Also, it seems the main issues are shifted to the baseload/nuclear, instead of incentivizing or asking the flexible power sources like wind to switch on/off as per needed.

The root cause is that we build wind and solar on wrong incentives. If they get support of x per unit produced at any time. They are still making money as long as market prices is higher than -x...

Most UK systems are paid on a "contract for difference" basis. This is an option-like agreement where they get paid a subsidy when power is cheap .. but pay back when power is expensive.

The impact of that is that high UK natgas and power prices resulted in renewables with "negative subsidy" paying money back to the taxpayer. https://www.current-news.co.uk/cfds-set-to-pay-back-10-5bn-a...

(Heck, I myself am part of this: I'm paid a 15p subsidy for electricity generated by my panels. This was a good deal when electricity was 10p. Electricity is now 30p.)

short-term grid stability (on the order of seconds to hours) is essentially a non-issue with a renewable grid that has even the fraction of the storage needed to make up for longer term variability in supply (days/weeks). Inverter systems connected to batteries are already the best and most cost effective means of providing frequency control to a grid (they can simulate an incredible amount of rotor inertia): this is one of the ways in which Tesla's grid-scale battery in Australia is making lots of money (and eating the lunch of gas turbine operators who previously provided this). You can also add this capability to wind and solar generation, it's just not generally worth the extra effort (it's easier to coordinate a few larger providers somewhat specialised to the task than distribute it to everything on the grid).

The author both says that we will find a solution to excess energy (eg. Hydrogen generation), and at the same time says we don't have space for baseload. It feels a bit of cognitive dissonance.

I still feel that these articles massively underestimate the amount of energy required in the future, when all fossils (eg. air travel, transportation, industry) will migrate to fully electric. Unless we want to pave every surface with solar, baseload will still be needed...

> The author both says that we will find a solution to excess energy (eg. Hydrogen generation), and at the same time says we don't have space for baseload. It feels a bit of cognitive dissonance.

The point is that excess energy from solar especially is essentially free, while excess energy from nuclear or other traditional baseload generators is very far from free.

> I still feel that these articles massively underestimate the amount of energy required in the future

That's certainly an aspect. Here in Norway it increasingly looks like we'll end up with a lack of electricity by 2030 or so[1]. However production is just one aspect, another problem is distribution. Even if we tap into all that wind up north in Norway, for example, most of the consumption is in the south, and there's a huge lack of transmission capacity between the north and south today, with no current plans of building more.

[1]: A big factor here is if we want to use our precious "clean" hydro to electrify the oil platforms or not. They need a lot of energy to operate.

> A big factor here is if we want to use our precious "clean" hydro to electrify the oil platforms or not.

Perhaps a much less comfortable question for Norway is when to retire them entirely and leave the oil under the sea.

Maybe Norway need to build the Shetland interconnector as well as the 1400MW one that's just opened? Shetland is where all the never-built tidal schemes are always planned for.

> Perhaps a much less comfortable question for Norway is when to retire them entirely and leave the oil under the sea.

Oh absolutely. About as easy as stopping a heroin addiction though, I expect. Not only the direct income from them, but about 10% of the non-govt workforce is directly tied to the oil extraction or in supporting roles.

I don't see how flexible power demand (e.g. hydrogen generation) helps baseload power. If you always have a valuable way to use excess power, that just makes variable generation like solar even more valuable and makes it even harder for nuclear to compete.

In the YouTube video in a sibling comment the rebuttal was that most chemical processes are continuous processes and require constant power, i.e. they become new baseload.

Hydrogen generation does not have a flexible power demand, because as you the electrolysis ramps up and down efficiently goes to zero.

I think it should be mandatory to include the actual demand, as well as the net load on duck curve graphs.

And while I'm wishing, showing it for each season would be nice too.

It would clear up a lot of unnecessary drama caused mostly by poor information design.

edit, Wikipedia does okay with it's two graphics:

https://en.wikipedia.org/wiki/Duck_curve

Why do that, when you can just present the surplus/deficit coldest-sunniest day in late spring and pretend it is an essential aspect of all renewable generation?

I wonder if servers could be located close to solar or even attached to individual panels. So that during times of excess you sell cheap computing resources. This could reduce demand on the electricity grid at peak times.

When my off-grid system is producing well my Internet router is automatically powered from it, I run more tasks[0] on my off-grid server, charge my phone off-grid, and sometimes run my laptop off-grid. It can be done, but is decidely non-trivial. That's what we have a grid for, with variable per-MWh pricing.

[0] https://www.earth.org.uk/note-on-site-technicals-46.html#202...

buying the server (capex) is probably the biggest pricing factor

since dc cooling has become very efficient

also since 99% people are wake up during day time, actually you'd not have "excess" capacity, you'd have more stuff running

This is indeed the heart of the matter: humans are diurnal.

If we stopped trying to live to an entirely artificial rigid timetable (with some ineffectual fiddling at the edges with daylight saving time) that ignores seasons, weather, etc, solar would be an even better fit to demand.

… only in places where heating is not the lion’s share of energy demand. Solar is a great fit near the equator or in sunny California where cooling is needed more than heating.

It is not great for latitudes north of 45 (much of Europe) where energy needs due to heating increase at night and in winter when solar is not running at capacity…

I'm not in the cloud computing business, rather, the utility business. But at a glance it seems quite similar: you invest in capital and make your money a few cents a minute over a relatively long period. Costs are relatively granular and so you get more for more usage.

With that said, don't you want steady, consistent usage selling a server? I'd guess for exactly the same reason as a power plant, nobody wants to own a server that cranks up 2 or 3 hours a day... unless they absolutely HAVE to, to provide their service.

I don't mind owning a server (or a power plant for that matter) that only operates a few hours per day, as long as the price I get paid for those hours is high "enough" compared to the risk on the investment.

The problem is, of course, that at the prices needed for that most consumers will no longer be interested in buying my computing services or power.

Many things are possible, but I don't think this particular idea is very promising from an economic standpoint.

First, you reduce the need for extra power cables but in exchange you now need internet cables going to all the solar panels.

Second, for any capital good you need to take into account the utilization rate during its lifetime. Imagine you have a server costing 10k with an expected lifespan of 5 years. This means it needs to earn about 2k per year. Depending on how much power it uses, it may well be that 50% utilization with only cheap electricity is still less profitable than 100% utilization with cheap electricity 50% of the time and expensive electricity 50% of the time.

Third, physical servers need maintenance every once in a while. It will be mcuh cheaper to organize that if all the servers are in a few central locations than if they are spread out. (this was originally one of the main selling points of electricity in the first place! You could finally build your factory where the people and the resources were, rather than having to go where the power was)

Finally there are other options to consider which solve the same problem at potentially a lower cost, like building out the grid and introducing more demand response in the industrial base.

> I wonder if servers could be located close to solar or even attached to individual panels. So that during times of excess you sell cheap computing resources. This could reduce demand on the electricity grid at peak times.

That doesn't really make a lot of sense. Servers are too expensive to keep off except in "times of excess".

That being said, it's a long established practice to co-locate data centers with cheap power.

I sort of agree with the author's hypotheses, but reach diametrically different conclusions.

Price goes to zero at highest solar production within the day, ergo nuclear will suffer from this. Counterpoint: if electricity is sold at spot price, wouldn't it be solar mostly affected by this? If at its best hours price is zero or negative, this means mostly zero profits for solar certainly. Meanwhile nuclear still has nights to get some of the investment back?

The awkward thing is: mostly nuclear works. France is doing ok, thank you, has best CO2-impact, with flexible production tech though. But renewables suffer from the presence of nuclear.

While mostly renewables suffer from very high storage and flexibility requirements. The author mostly handwaves those away: "I have no doubt that entrepreneurs will find zillions of ways to use energy available". Sure free energy is nice, but that's a problem for solar investors surely (or the taxpayer).

But in the end what do we want? Renewables? Or the expectation that renewables reduce CO2? Zero-CO2 can be done with flexible nuclear.

----