Hacker News Reader

Comment by photochemsyn | Hacker News Reader

'handwave the unsolved-at-scale' ? All you have to do is look at Australia:

https://www.pv-magazine.com/2021/08/10/worlds-largest-grid-f...

> Battery rollout: "AGL Energy has committed to build 850 MW of battery-based assets by 2024. To that end, Torrens Island marks the first project to be constructed at the site of a fossil-fuel power plant. It won’t be the last, however, as big batteries are planned for New South Wale’s Liddell coal plant and Victoria’s Loy Yang facility. In January, AGL Energy revealed that it had secured both Wärtsilä and Fluence under non-exclusive framework agreements to supply up to 1 GW of large-scale battery storage."

So, would it make more economic sense for this 1GW battery-storage system to be replaced by a nuclear power plant? Australia after all is one of the places with the most high-grade uranium ore, being among the top three exporters.

Here's an example of more reliable analysis:

> "To illustrate this point, the 2,430 MW Vogtle nuclear plant could be expected to generate 21 million MWh per year. That is enough to power about 1.75 million residential households. Meanwhile, a hypothetical 3,500 MW solar power plant would be able to produce just under 6 million MWh of electricity per year. This number is enough to power only 500,000 homes, which is considerably less than nuclear power. For solar to produce as much electricity as is generated by a nuclear power plant, it would require about 13,000 MW of utility-scale solar capacity, which about four times as much as built in the existing plants. However, the cost to build this 13,000 MW facility would be $12.1 billion, which is still just 50% of the cost of the $25 billion Vogtle nuclear plant."

https://www.solarfeeds.com/mag/solar-power-vs-nuclear-power/...

View on HN

French electrical power is 70% nuclear and has an installed capacity of 135 GW of power today. That comes out to 94.5 GW of nuclear power in France alone. Is there any country in the world that can supply hundreds of gigawatts to the grid from batteries today? If not, then the problem of grid power storage is unsolved at scale compared to nuclear.

I don't think that's a common use of "unsolved at scale". We know how to do it, it just doesn't make economic sense right now. I mean, suppose you have hundreds of GW of storage of some form installed. You could just claim it's still not "solved at scale" because the global energy consumption is even higher.

The global supply chains for lithium batteries are already struggling to keep up just with current demand. Where are we going to get enough cobalt? Hopefully not DR Congo.

Who says Lithium batteries are the best technology for grid scale storage? (Besides Elon, I guess?)

There are multiple kinds of grid storage. Australia, for example, wants these batteries to provide fast-acting frequency stabilization... in addition to stabilizing the output of various wind and solar facilities.

>Where are we going to get enough cobalt?

Nobody would use this kind of lithium ion battery for grid storage. Instead use Lithium Iron Phosphate (LiFePo4) batteries- no cobalt needed. LiFePo4 batteries are absolutely good enough right now for grid storage.

And any sort of lithium battery is the most expensive storage.

Utilities are not generally in a habit of reflexively choosing the most expensive alternative, even if some have plumped for nukes in places where ratepayers have no say.

Batteries will be a negligible fraction of utility storage, so that is moot.

Why? What will the rest be?

Kinetic (water gravity, flywheel) or thermal (molten salt baths which can be supplemented with solar thermal).

None of these things exist at a large scale except for pumped hydro / water gravity, and that requires very convenient topography and geology to work. Generally, any non-chemical storage system is overpriced (too inefficient) or a maintenance disaster.

Do you think those technologies will be scaled such that batteries are a "negligible fraction" of utility storage, or are you just giving examples of possibilities?

Numerous technologies will be scaled. Those mentioned adjacent are just very easily understood examples. Utilities will have criteria other than "easily understood", cost chief among them.

Flywheels and molten salt both have pretty gnarly failure modes too. Not Chernobyl bad, but still no fun.

Underground flywheels are perfectly safe, and more easily kept evacuated. But they won't be used much at utility scale, because they are expensive per Wh stored. Maximum storage is dictated by tensile strength of the wheel material, and each wheel needs its own motor / generator.

A negligible fraction of storage will be in flywheels.

Pumped hydro is not kinetic.

Not much storage exists because it has not been built yet. It has not been built yet because we haven't the renewable generating capacity to charge it from yet. Thus, we are spending on renewable generating capacity first, and lately factories that will build storage systems.

Battery storage is the most expensive variety. Round-trip efficiency is not very important anymore. I.e., other concerns dominate now. Anyway, gravity storage efficiency has always been fine.

I know there are options like pumped water, compressed air, spinning disc. No idea how feasible they are at scale, but I can imagine a world where physical storage like that could be lower maintenance & cost in the long-term compared to chemical batteries.

Even if less efficient, you can get enormous potential storage from pumped water with relatively simple machines compared to massive battery farms.

I'm interested in those alternative technologies, but I suspect they won't scale as well as batteries. Pumped hydro has been around a long time but it hasn't scaled because it is just very geographically specific.

Pumped hydro is far, far less geographically specific than always assumed.

This is partly because it needs no watershed, and partly because transmission lines mean it need not be local.

Anyway pumped hydro is only one of many equally viable alternatives. All will be used. Eventually which ones are cheapest will become clear.

I think my assumption with pumped hydro is that there must be something holding it back because it's been mature for a long time, so if it were as useful and scalable as you say, I think it would have been scaled a lot more by now.

The overwhelming majority of utility-scale storage today is pumped hydro. It stores and releases many GWh every day. You don't hear much about it because it just works.

Yes but is still a small amount of energy relative to total demand and there don't seem to be serious efforts to scale it up massively. I don't claim to know why there aren't. Maybe there should be. It just makes me suspicious when I see a mature technology that still sees limited deployment, that there is a good reason for that.

Batteries have the very large disadvantage that cost is per Wh stored. Flow batteries can reduce this, as capacity becomes a matter of tankage and electrolyte.

The sweet spot for storage is near-zero opex, very low per-Wh stored capex, and cheap capex per-W in and, particularly, out. Thus, weight for watt-hours, and simple motor / generator for watts in and out.

Synthetic fuel has the appeal that the stored energy is readily tanked and transported, and very saleable; enough so to overcome low round-trip efficiency, expensive electrolysers for W-in, and complicated combined-cycle turbines for W-out.

I know a fellow who is running a startup that moves a (very) heavy weight up a hill, and going back down it provides power. He says it is something like 90% efficient at energy recovery, whereas pumped water is far less due to friction in the pipes.

He also said the inertia of a falling weight is enormous, and would help stabilize the grid.

Okay, but if his startup isn't named Sisyphus Energy, I don't know what he's doing with his life.

This is a genius idea, especially because it could be done anywhere with a hill, which makes it a lot more geographically accessible (since most places have hills)

It’s actually not a good idea at all. When you do the math ( or rather the physics) gravity batteries need so much mass and height to compete with other technologies that it’s just not worth it.

Except if water, pipes and pumps are used, as water is heavy, and pipes and pumps can transport a lot of it very efficiently and cheaply. That’s called pumped hydro storage and is currently the #1 grid scale storage in the world.

Trying to do the same with solid materials can sound like a genius idea but it’s really not.

Well, his company is building a pilot plant in Brazil. I suppose we'll find out if he's a genius or not.

> efficiently

Are you sure about that? He said the losses due to pipe friction were considerable.

Losses turn out not to matter much when opex is near zero and generation capex is very low. You just build a little more top-line generating capacity.

70% round trop efficiency. And it’s from actual operational data, not from an optimistic spreadsheet calculation.

70% is plenty, nowadays.

If the railroad is already built, and disused, that drives down cost.

Cost is what matters most.

You need a very, very large mass, and then rails and rolling stock to carry it.

That is why vertical motion in a disused mineshaft is attractive. A single 10,000t weight hanging from a winch has just one moving part, the winch, and cable all in tension, with no preparation of the shaft sides; everything is concentrated at the top. The mineshaft provides isolation from environmental disruption.

You get much the same effect with a float attached to a cable down to a pulley augered into the sea floor; thence to an onshore winch. The cable and pulley are exposed to sea (or lake) water, but you get the advantage that the winch may be shared among as many cable reels as you like, with only simple clutches. A 20m diameter float displaces 4000 tons of water, and the sea floor may be much more than a km deep.

(In principle you could share a winch drive among lots of mineshafts, but they are rarely close enough together. You could bore more, but that drives up cost. The world is absolutely perforated with disused mineshafts.)

I hadn't heard this sea float / winch idea before! Any links you could dig up for me? Anyone operationalizing this yet?

There was an "Undecided with Matt Ferrell" on YT some weeks back that mentioned a dumb version of it somebody at a university promoted.

That one had the winch and motor / generator at the sea floor, and a big rack of floats. It was, of course, too expensive. This is a common feature of energy storage ideas: people seem to love to promote the most stupidly expensive version of a good idea, and then give up when it is shot down for that.

Certainly, any scheme where you put complicated electromechanical stuff on the sea floor, it will be expensive. And likewise, where the expensive part is only usable with one "unit" of storage. Or anyway it just costs too much or needs perfect weather or something. (E.g., Energy Vault (NRGV) is stupid and their investors will lose their shirts.)

But there is usually a simple variation that lacks these problems.

10,000t weight? If it’s made up of rock with a density of 3000kg/m3, it’s 3333 cubic meters, so about 15x15x15 m3. Basically a large building. I’m not sure what kind of shaft can handle it. And even if you can pull it of, 10000t over 1000 meters is only 27 MWh of storage. 1 pumped hydro plant can store hundred of GWh!

(1) Rock is not very dense; pig iron density is 3x rock's. (2) A cube is not the only shape, for a weight. (3) Total capacity is not the only parameter of interest, for utility storage. (4) There is not only one mineshaft. (5) No utility is restricted to using only one kind of storage; different kinds favor different parameters, and we may expect utilities to use different kinds at the same time to maximize value and minimize cost.

Assuming utilities will do the stupid, expensive thing leads to failed analyses.

Rock is cheap. The shape is irrelevant. For grid scale, total capacity is what makes it grid scale. No amount of parameter tuning can change the fact that stored energy is at best mass * g * height. Adding more huge mineshafts is irrelevant. You can always add more units, it doesn’t make it more efficient.

Sounds like an Elon Musk solution. Make it bigger, taller, make more of it, and somehow it’ll bend basic physics facts to become the ground breaking disruptor technology of the next century.

Efficiency doesn't matter nearly so much as is usually assumed. What matters is cost. Exactly, cost per Wh stored, and per W in and per W out.

Mineshaft storage is very, very efficient, but has an upper limit on energy per mineshaft determined by its depth and how much weight fits in it. The cost of the winch apparatus is amortized over that amount. More mineshafts means more total storage.

Pumped hydro has very, very low cost per Wh stored, but kinda high per W in and out. But thousands of pumped hydro systems were already built, almost, so seem free; they just needed pumps added.

Synthetic H2 and NH3 are very, very cheap per Wh stored. W out uses existing turbines, so seems cheap. W in depends on your electrolysers, which are improving fast.

In batteries, cost per Wh stored has been large, but is different in different chemistries. In a better chemistry it is lower. Flow batteries have lower cost per Wh stored, for any liquid chemistry, but higher per W in or out; and need plumbing and pumps.

Details matter. Numbers matter. Dumb choices make dumb designs. Dumb designs can be improved on.

Drawing conclusions based on dumb choices is just dumb.

Musk is dumb, but just barely smart enough to hire smart people and, sometimes, let them be smart. So, Hyperloop was just dumb, no matter who works on it. Starship is kinda smart, depending on what you can think of to do with it. Mars colonization is dumb. Starlink could possibly be smart, depending on a lot of details.

A wholly new capability, like radically cheaper mass to orbit, can often enable smart ideas. Super-expensive mass to orbit was good enough for weather satellites, planetary probes, and geosynch COMSATs.

The downside of pumped water, of course, is you need a hydroelectric dam and there are not many sites suitable for that. Those dams also come with significant environmental disruption.

The weight system only requires part of a hillside, has a pretty small environmental footprint, and there are lots and lots of hills.

Not true. You can pump water to an existing hydroelectric dam, but you can also just pump it to a pond on a hilltop. No environmental disruption. There are lots of hills.

The rail system needs maintained rolling stock and maintained rails. The mineshaft method is overwhelmingly cheaper and more reliable, because all that must be kept working is an electric winch. There are lots and lots of disused mineshafts.

Sorry, you can't make hydropower without constructing a dam, turbines, and completely disrupting the stream. This is environmentally destructive, and due to the scale needed, there are few suitable sites.

There are indeed many unused mineshafts, but a shaft would be only suitable for a small weight, maybe enough for a single house. This does not scale.

Pumped hydro doesn't require a stream, since it isn't harvesting energy from a preexisting flow of water.

How are you going to create a large enough pond?

Maybe read up? When you just make shit up as you go along, you say dumb things.

Asking a question is not making things up.

This involves moving earth in places with suitable topography. The economics depends on both the vertical offset between the two reservoirs, and how much earth has to be moved per unit of reservoir volume.

Algorithms have been created to use GIS data to automatically find potential PHES sites. There are lots of candidates. Australia, for example, has 300x more potential that it would need for a 100% renewable grid.

http://re100.eng.anu.edu.au/global/

Often there is already a depression you can use as is. Or deepen, and use the excavation for a dike.

If not, maybe there is a box canyon you can close the end of with a dam.

In places, they have flattened a convex hilltop and used the material to make a dike.

Usually you don't need anywhere near as much storage volume as a typical hydroelectric dam reservoir holds.

It’s actually not a good idea at all. When you do the math ( or rather the physics) gravity batteries need so much mass and height to compete with other technologies that it’s just not worth it.

Now if you use water, pipes and pumps, instead of solid materials, now we’re talking. It’s called pumped hydro storage and is the number one grid scale storage in the world.

Hint: pumped hydro is, exactly, a gravity battery. A gravity "flow battery", if you like.

It is #1, just now, because a lot already existed. They cost a great deal to build, back when. New ones would cost as much, but a small variation is much cheaper, not needing a dam or a watershed.

A small variation on a bad idea can be a very good idea.

There is a number of battery chemistries more cheap and even efficient than Lithium-based, but with lower energy density. High energy density is important for handheld devices, for flying devices, even for cars. But stationary energy storage can be much heavier and bulkier per kWh stored, as long as it's cheaper, and can sustain more than a few thousand recharge cycles.

Lithium batteries have the momentum and are dropping in price because of the EV cars revolution happening right now. When solar revolution will happen i utility scale, the utility-type batteries that will emerge may be seriously different.

Since you are 100% certain that such technology will emerge in an economical viable state to solve the storage problem for Europe, could you please provide a date so we can draft a EU law that forbids burning of fossil fuel in the European market by that date.

If you can do it by this winter it would solve inflation, enable a complete sanction packet against Russian oil and gas, solve several political conflicts within EU in regard to the Ukraine war, eliminate the food shortage problem and significant reduce the cost of farming, make a large dent in reducing global warming, improve GDP, and boost prosperity in general.

Alternative chemisitries are many, and are actively developed; here's a 2021 overview: https://www.mining.com/non-lithium-battery-chemistries-to-gr...

What I can't predict is when some of them will become competitive enough to matter, and thus used in some large infrastructure projects.

If we had completely free, 100% efficient grid scale storage of unlimited capacity today we couldn't stop burning fossil fuels because we don't have enough renewable generation capacity installed to fill any form of storage.

If the storage solution is close to free then the capacity doesn't matter since any new investment would be guarantied profit. Wind, when the weather is optimal, is several time cheaper than fossil fuels. Everyone would rush to just install new renewable capacity.

Just this week in Europe, during the summer vacations and when energy usage is at the lowest point during the year, the energy market price still jumps by 3000% when the wind go from high to low.

The advantage of lithium batteries is energy/weight and energy/volume. With grid storage, neither weight nor volume is much of an issue.

A couple years back, "60 Minutes" ran a segment of somebody developing grid batteries with seawater and dirt. They didn't store much energy per weight and volume, but who cares for grid batteries. I wonder what happened to this technology.

At scale means with mining requirements as well.

With current and projected reserves we can manage intra-day requirements for the whole world. This is the we mine it all option.

But we need intra-year storage capacity.

Nuclear is clearly the easier and only option (and the most environment-friendly to me, considering the mining requirements)

Mining of what? Hydrogen as storage medium for example doesn't require all that much mining. Are you assuming storage will be in the form of Lithium batteries?

I'm assuming currently proven tech of batteries. And we need to act now for climate change, with current tech.

Electricity to H2 to electricity has horrendous efficiency (about 20% at most IIRC). We'd be much better off CO2-speaking using it to replace coal in steel production, and replace gas in fertilizer production. Both of which are proven tech (but not globally deployed, as coal and gas are cheap)

Power2Gas2Power has low round trip efficiency (say 40%), but as only a fairly small fraction of the energy flow goes through it that doesn't matter that much at the end of the day. If it did, that just means heavy industries would migrate to less seasonally variable places.

Inefficiency just requires extra capex, not opex. And the capex is capex is still falling at an exponential rate. So, no.

Where are you getting that 20% number from?

https://www.irena.org/-/media/Files/IRENA/Agency/Publication...

https://www.sciencedirect.com/topics/engineering/round-trip-...

The round trip efficiency is 30% or more and rising with better electrolysis and reverse electrolysis being developed.

Personally, I expect something made with iron or sodium to take over.

For hydrogen, we have to solve the horrible cycle efficiency of its storage. It's not so bad for use as a fuel (e.g. for airplanes), but as static storage hydrogen is currently a very bad choice.

Cycle efficiency is trivially resolved by feeding in more zero-opex power.

But hydrogen efficiency is improving as fast as all renewable tech.

I agree that hydrogen efficiency doesn't matter here, but the capital expenditure to build a giant electrolyzer with 5% utilization doesn't really make sense. The cost of the electrolyzer is the things that needs to greatly fall if hydrogen is going to be used for long-term storage.

Electrolyzers driven by otherwise curtailed renewables would have a capacity factor much higher than 10%. If I go to https://model.energy/ and solve for the optimal solution for the US for 2011 weather and 2030 cost assumptions, the capacity factor of the electrolyzers is 43%.

That is more of a what-if. If there was more wind and solar built out there would be enough curtailed energy to run electrolyzers at 43% capacity factor. I am saying today it doesn't make sense because there isn't a lot of over-provisioning taking place. There should be but what investors are going to jump first with a curtailed energy electrolyzer when curtailed energy is low.

It's a demolition of your nonsensical 5% number (which is also what-if).

In the state with the most solar built out in the US the curtailment is 5% https://www.eia.gov/todayinenergy/detail.php?id=49276.

Which is of course meaningless, since curtailment increases with the amount of installed wind/solar. What matters is how much curtailment there would be in an optimal all-renewable energy system.

I tell you what: show me a cost optimized renewable energy system in which the electrolyzers have a capacity factor of 5%.

Demand for hydrogen for all uses will drive elecrolyser use to as close to 100% as available energy permits. Ultimately, other cheap storage will be used to drive them.

Hydrogen has important industrial uses. Currently we get it from natural gas. That has to change in the future.

Hydrogen produced using NG will be deeply undercut by electrolysed hydrogen. Extracting NG will become unsustainably expensive as zero-opex synthesis surges.

Well nuclear proponents handwave away the massively higher costs, the still unsolved long term storage problem (it's been >70 years of nuclear power and there still does not exist a long term storage facility in the world), and insurance for the case of a disaster.

> it's been >70 years of nuclear power and there still does not exist a long term storage facility in the world

It's bizarre how anti-nuclear activists point at problems caused by anti-nuclear activism and ask, like Eric Andre, "how could nuclear power do this?"

Nukes' problems are not caused by activists. They have nowhere near that much influence. Nukes' problems are inherent.

Biggest, by a large margin, is capex which simply prices it completely out of the market.

Even a superficial glance would tell you that this doesn't make sense. We have had >70 years of peaceful nuclear power. A significant anti nuclear movement has existed for maybe 40 years and it has clearly not been influencial until maybe the last 20 years. So why has the nuclear power industry not solved the issues in the first 30 years? Moreover, why has France which hasn't had a significant antinuclear movement ever solved any of the issues?

The storage problem has yet to be solved for the same reason why any nation has yet to solve the issue of pollution from fossil fuels. It cost money and any increase in energy costs is detrimental to growth.

The insurance issue is similar identical to other energy sources. Why isn't hydro power insured in the case of accidental flooding? If they were it wouldn't be possible to run them at current energy prices. We don't see fossil fuel plants being insured against increased air pollution, nor do they have to pay if people around them get sick. If a hydrodam causes a flood, then there will be a flood. Hopefully those people downstream has their own insurance.

No.

The storage problem has not been solved because we do not have the technology. There is no evidence that the technology os possible.

The long term waste (not much in quantity, but it does mount up. A few million litres at the moment I believe ??) has to be stored for periods of time on the order of hundreds of thousands of years.

There is no way we can write a warning sign that we know can be read for that long.

We cannot build a structure that will last that long.

There are no places on this planet that we can be sure are stable for that long.

It is making future generations pay for our current consumption.

>It is making future generations pay for our current consumption.

Okay, then in places where wind/solar/storage aren't up to snuff we'll just keep burning coal.

The US had a solution at Yucca Mountain, except the solution has been legally and politically sabotaged by activists.

A significant anti nuclear movement has existed for maybe 40 years and it has clearly not been influencial until maybe the last 20 years.

That anti-nuclear movement was huge in the 1970s, and was successful in many states and countries at creating major roadblocks to new power plants. Their success in California back then is big reason why the last plants are going offline now, and there haven't been any new plants since Diablo Canyon, which was started in the 1960s. To say that anti-nuclear groups weren't influential until the 21st century is absurd.

> So why has the nuclear power industry not solved the issues

The technical issues are solved.

The political issue is that the "green lobby" will not accept any nuclear waste storage solution.

> A significant anti nuclear movement has existed for maybe 40 years and it has clearly not been influencial until maybe the last 20 years.

O_o There's been a significant and influential anti-nuclear movement for the last 60 years...

I think the only one very valid criticism for nuclear is long term storage problem. But also I think it's sunk cost. It's a big problem to have 10000t of nuclear waste from zero, but adding another 15000t from 10000t can be ignored, isn't it?

>the still unsolved long term storage problem (it's been >70 years of nuclear power and there still does not exist a long term storage facility in the world)

Storage has /been solved/ for decades.

High Level Waste is only a tiny fraction of total waste. As it stands, all of the HLW created by all the nuclear plants in the world could be kept in a football field sized area. The vast majority of the rest of the generated waste will decay to to harmlessness during the lifetime of a plant.

HLW regulations around transport are so high in most countries that the standards for transport casks has made them practically indestructible. They can literally be hit by trains and no be breached. Solutions like those developed by Deep Isolation also effectively solve any long term storage problem for the HLW.

Here [1] is a picture of the temporary storage at the Yankee nuclear power plant temporary storage (sorry I could not find better pictures from above to give more accurate size). There are ~90 commercial nuclear reactor in the US. I can immediately see that this would not fit into a football field sized area. And this is for a single country and ignores waste that was already transported away. And no things will not decay to harmlessness during the lifetime of a plant. They decay to a level where the containers are not very warm anymore so they can be more easily transported. Also low level waste is not harmless either, and there's going to be a lot of it when decommissioning a plant.

Sorry here's the link: https://vtdigger.org/wp-content/uploads/2017/06/Aerial-from-...

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

Operation will start next year.

Not just hundreds of Gigawatts, but also tens of thousands of Gigawatt-hours. Unless you're willing to accept regular black-outs, you need to build enough capacity to account for the occasional long, sustained period of cloudy days with low wind. The cost of that extra necessary-but-rarely-used capacity needs to be factored in to any comparison of wind/solar vs other energy sources.

It might make more sense to have either fossil-fuels as backups for such times, or for HVDC lines to connect geographically disparate regions (it's always sunny or windy somewhere).

As we're seeing from Russia and Europe, depending on someone else means you can find yourself being threatened or actually being cut off.

There will be plenty of tropical suppliers of synthetic anhydrous ammonia competing, if transmission capacity or availability falls short.

People might even keep a stock on hand.

Why would you need 100% storage supply. Also note if France would go 100% nuclear they would need either significant overprovisioning or storage as well. Nuclear is not an load following source. Even less so if you want to operate economically.

> Why would you need 100% storage supply.

So you don't have power outages on windless nights.

You generally use peaker plants or backup fossil fuel or hydro for that use case, at this point in time batteries are used to supply power for a couple of hours at best. You need that backup power anyway for when the nuclear plant is out of commission for some reason.

Regardless of how you're generating your power, you're going to need to overprovision your generation capacity slightly to account for downtime. You only need storage if all of your generation capacity is offline at the same time, which is inevitable with solar power (nighttime), possible with gas (pipeline failure?), but less inherently likely with nuclear.

> backup fossil fuel

In practice this means fossil fuels become the baseload power source.

I sincerely believe that HVDC is a game changer here, it allows you to move energy around at a fraction of the losses that were considered normal before allowing for overcapacity in one place to be used elsewhere eliminating a major requirement for storage, on that scale there is no lack of combined sun/wind availability.

Anyway transmission losses are immaterial when top-line generating capacity is so cheap.

China is building solar in Chile, at their antipodes, and a transmission line back to China with only 50% loss, for solar energy through the night.

> In practice this means fossil fuels become the baseload power source.

False. In the immediate future, synthetic anhydrous ammonia will displace it. Extracting NG will be too expensive to compete, because the only cost of ammonia will be capex. The more you make, the less it costs.

Your argument does not follow at all. I have the impression that you don't know what baseload is. Baseload is the minimum amount of demand over a given time (it is not a measure of supply at all).

The previous poster talked about using fossil (or any other means) to supply the mismatch between the supply and demand at the times where such a mismatch might exist. This is completely unrelated to baseload, it could in fact be larger than baseload. Note that such a mismatch would also exist for nuclear (dependent on overprovisioning). However it would in general be only needed of short periods of time.

It could be larger than baseload, but it's unlikely to be any less than baseload.

They are completely unrelated and yes in general they would be much smaller than baseload. Why wouldn't it? What the GP described is essentially what gas peakers do.

> yes in general they would be much smaller than baseload. Why wouldn't it?

Because if your baseload is your minimum power consumption (which generally occurs at night) and you have a windless night (therefore your renewables are generating 0 power), your "gas peakers" will have to provide the entire baseload amount of power.

> In practice this means fossil fuels become the baseload power source.

As the name baseload power implies, it would have to be providing power all the time and that's not what is needed.

That is correct. The term I see by my government is "reserve energy". In practice this mean a country subsidize a fossil fuel plant to keep engines running when demand is low, and then the plant get paid a second time if and when the plant is running.

It is one of the main reason why investing in fossil fuel energy when countries turn to wind power is fairly good choice. The plant has lower costs in terms of fuels, when it burn fuel it does so when the market price is at the highest possible, and it get paid regardless from the government. The government are also locked into continuing this scheme as the alternative is to not have the reserve energy when needed which makes voters very angry.

Burning fossil fuels when the price is lowest and without any government subsidies is just something people with too much fossil fuel can afford to do.

> Nuclear is not an load following source

In France it is, and has been successfully for more than 33 years [1]

[1] https://hal.archives-ouvertes.fr/hal-01977209/

Thank you for providing a trustable source.

Ok I didn't know that they implemented load following in France. However, note thatbbecause nuclear power cost is almost completely capex dominated so doing this essentially increases your power cost by the factor that you use for load balancing. So out of economic reasons you hardly ever want to not run a nuclear plant at 100%.

Seems like grid scale storage could "solve" this, in precisely the way it does for renewables (ie. Treat nuclear as a sun that shines 24/7 and build storage to allow you to generate a bit above the average and follow the actual usage curve with storage).

That is, assuming that grid scale storage is in fact viable.

Yup, but didn't we already say that renewables are cheaper per kWh? So then with any storage renewables should still be the cheap bulk generation.

You said that, that doesn't mean it is true. And it probably isn't true north of 50 degrees of latitute where lots of the people who need energy live.

France's nuclear fleet has an all-in cost of 7c/kWh, which is roughly the same as the cost of solar or wind power in France, but without the need for storage or smart grids that doubles the cost of solar/wind. (For all this remember that France has about a third of the insolation of California or Arizona per year, or maybe half at best in some places.)

For all this remember that France has about a third of the insolation of California or Arizona per year, or maybe half at best in some places.

I don't think that is correct. This is a table of average insolation for different cities around the world derived from NASA's Surface meteorology and Solar Energy data set: http://stalix.com/isolation.pdf

Los Angeles gets 5.4 kWh/m^2/day. San Francisco gets 4.9. Lyon, Paris, and Toulouse get 3.7, 3.3, and 3.7 respectively.

I dunno if this is what gp meant to say but didn't convey properly, but the average per year is pretty irrelevant once you leave the subtropics. You need enough power for the lowest level of light part of the year, which is also the highest total energy use part of the year (much of that is achieved through individual home heating with burning fuel, but electrification is necessary to have any hope of even stalling climate trends).

France installed nuclear capacity is 61370 MWe, I suspect you looked at MWth (thermal output, not electrical output). Source:

https://fr.wikipedia.org/wiki/Industrie_nucl%C3%A9aire_en_Fr...

Currently the capacity factor for french nuclear reactors is less than 50% because of various technical issues and maintenance schedule slippage.

It will go further down because of a very hot summer and regulations on maximum released water temperatures.

At a minimum, the EV demand will solve this. 150 wh/kg Sodium Ion (good enough for likely 300 mile cars) is coming next year, and 230 wh/kg LFP is coming later this year (350-450 mile cars). Grid storage should be a core use case of Sodium Ion.

Lithium Sulfur looks VERY PROMISING, and other schemes and grid-specific chemistries will be developed.

Within the timeframe of any new nuclear construction, you cannot predict the cost targets with respect to any of these industries/technologies: they all are in prime economies of scale and technological improvement curves. I don't think they will improve as well as the last decade where solar dropped double-digits per year, but we'll see.

Nuclear keeps trying to say the scale isn't solvable with wind/solar/battery. That is a completely unproven/unsubstantiated claim. It's basically FUD. The materials being targeted for near-term alt energy will be massively scalable.

> So, would it make more economic sense for this 1GW battery-storage system to be replaced by a nuclear power plant? Australia after all is one of the places with the most high-grade uranium ore, being among the top three exporters.

As much as I detest the pro nuclear shilling, this isn't a valid comparison. 250MWh of storage is only useful for very short term energy arbitrage and isn't a replacement for always-on power.

The real answer is a mix of solar, solar-thermal, wind, storage, and combined cycle plants for the holes (ideally eventually powered by plant-based and solar-cracked hydrogen/methane/etc.). This mix costs a bit more than nuclear does now, but doesn't have the huge massive downsides of total dependency on and subservience to a handful of companies operating out of the five or so countries that are allowed to produce fuel (as well as all the existential risk stuff, concentration of risk to mistakes/attacks, horrible un-accounted externalities and supply chain fragility). It will also likely cost less to build it at a time such that it would still be completed before a nuclear plant which was comissioned now.

Of course that's all moot because the balance *right now* is running existing fossil fuel infrastructure more vs. building solar and not storing the energy. And in this equation, the cost of solar is a tiny fraction of nuclear.

Your final comparison there is an excellent example of hand-waving away the problem of intermittency. This is something that advocates of renewables regularly want to dismiss because the discussion is a difficult one.

Wind and solar are cheap now because their fuel cost is zero and the cost of grid reliability is borne by us - consumers. In a better world, the cost of balancing the grid at night should be borne by a solar PV generator because it caused the problem.

Even if you match the energy output of a solar and nuclear plant, the component of time is ignored. On a grid powered by solar, where will energy come from at night? How many batteries will you construct? The scale of these batteries is again something that’s dismissed by claims of continental-scale grids but has that been achieved anywhere? All the batteries in the world combined can power à giga watt scale grid for hours. That’s right, hours.

Nuclear is a solution in the here and now. It buys us time to sort through these problems while we work towards a better future.

Nuclear has the same issue of intermittency. You have to either overprovisioning if you use nuclear or use battery storage. Nuclear is much more expensive than wind/solar so if you follow an overprovisioning strategy the best is to build wind/solar capacity, because it buys you more. Moreover, the larger and more integrated grids become (this is happening already for economic reasons in Europe) the less this is an issue.

Also why do you only mention solar and not wind? Maybe because eit would make your argument less convincing? The wind still blows at night, so you do not need battery storage to provide all the demand during the night.

Why are the same simplistic arguments being brought forward again and again. They have all been debunked.

> The wind still blows at night, so you do not need battery storage to provide all the demand during the night.

Is it enough wind to meet all the demand at night? And what happens when it stops blowing as much as expected for a long period of time[0]? You still need baseline power generation for quite some time to come until those questions that, well, keep getting hand waved away are solved.

0: https://fortune.com/2021/09/16/the-u-k-went-all-in-on-wind-p...

There are many studies which showed how much overprovisioning you need to go to 100% renewable with different size grids. None of these are outrageous, but it's also nothing we currently need to think about. We can easily scale up renewable capacity to 70% without significant overprovisioning. So we really just need to build lots more renewables at the moment.

If you don't have renewables for a while, and transmission line availability flags, you import anhydrous ammonia from the tropics. So, you need only minimal over-provisioning and local storage.

But then you presumably need to build these extra power plants that can use that fuel?

We have combined-cycle turbines already. More will be built wherever they will be needed.

A combined cycle powerplant is about 10% of the cost, per unit power output, of a nuclear power plant. So we could completely back up a renewable grid with CC plants and nuclear could still be much more expensive.

And, steam turbines cost more to keep working than CC turbines. But both are very mature, meaning costs won't change.

> The wind still blows at night, so you do not need battery storage to provide all the demand during the night.

The wind doesn't blow as strongly at night. It's much calmer.

Yes and demand is also much lower what is your point? Apart from the fact that your statement is highly dependent on location (it is much less true for sea installations for example).

> You have to either overprovisioning if you use nuclear or use battery storage.

Whereas with solar and wind, you need to do both, because you need the overprovisioned capacity to recharge the batteries.

No with a reasonably large grid you only need to overprovison no need for battery.

But this is really a question of economics, the mix of how much storage you want to include is really a question of cost and the specific grid. One should also say that pretty much every grid already has some level of overprovisioning, because otherwise you run into problems when some exceptional events hit.

In the end why would you pay for more expensive nuclear if you can get more capacity for the same price with wind/solar. So it is cheaper to build the overprovisioning with that. Moreover if the current trends for storage prices continue then things become cheaper still.

Grid scale storage need two things to be viable. Capacity and duration. The battery solutions being rolled right now are those that operate on the scale of hours, being charged each day and then discharge each night. When you can reliable assert that you can fully charge the storage each day, and reliable know you can sell the full charge each night at the price point when the market peaks, it becomes fairly trivial to determine when revenue will exceed that of investments.

Solar + battery is great in those locations, and basically no one is aruging using nuclear when solar and battery is viable for 365 days of the year. Australia has a fairly large desert where such assumption would be true. Finland, Sweden, Germany, France, UK and most of Europe do not have reliable 365 days of sun each year. There aren't many investors spending their money to build solar + storage in northern Finland.

>The battery solutions being rolled right now are those that operate on the scale of hours

The state where the mega batteries are installed uses about ~32 GWh total a day. An 850 MWh doesn't even come close to covering their needs. Covering ~2.6% of their load is equivalent to keeping power on for about 40 minutes.

>Solar + battery is great in those locations, and basically no one is aruging using nuclear when solar and battery is viable for 365 days of the year. Australia has a fairly large desert where such assumption would be true.

Deserts actually make for the worst place to install solar. Dust / sand covering panels substantially lowers their efficiency (to near zero), and keeping them clean involves a lot of water (Can you guess where it's hardest to get water? Yep, a desert). And lastly, because the deserts are so far away from everyone else, there is significant transmission loss.

850MW doesn't tell you how much energy it can store -- if you need 850MW to power your grid and your battery can supply 850MWh, where do you get power for the other 11 hours until the sun comes out?

For some reason, these journalists tend to drop the "h" portion of the rated capacity of energy projects. The battery capacity (I believe this will actually be a group of stand-alone installations) will be 850MWh, which is about 2.6% of that Australian state's daily consumption.

> the cost to build this 13,000 MW facility would be $12.1 billion, which is still just 50% of the cost of the $25 billion Vogtle nuclear plant.

Yes this is generally my argument. If you take the money you would spend on nuclear and spend it instead on solar, wind, gridscale storage, as well as EV and home battery rebates, you end up with a much more robust distributed energy production grid with very low maintenance costs. I think the idea of including rebates for home batteries and EVs is an important one, because it means that when major parts of the power line system are knocked out for some reason, you still have a lot of distributed storage.

You can also begin energy production with a solar and battery roll out MUCH sooner than new nuclear construction. From a climate change perspective, producing clean energy sooner is a big deal.

And we should not overlook that nuclear energy is a single target for terrorist attack and requires a powerful state to manage security for the plant and its supply chain, while batteries and renewables require no such thing. From a libertarian/anarchist perspective, we can have a weaker state and still have good energy production with renewables.

EDIT: I am also realizing that handing out a bunch of rebates to the voters for new stuff (rooftop solar, home batteries, new EVs) would be WAY more politically popular than "lets pay now for a big monolithic nuclear power plant that takes 20 years to build." If you want a plan that actually works, keeping the voters happy seems like an important component.

But if you just price in the cost of the rebates, aren’t you overlooking a large chunk of the cost, which is borne by homeowners? That’s still a cost, even if it’s not being directly paid by the government.

So you see a doubled cost for nuclear.

But you need to add that Renewable solution only last 20 years vs 60 for nuclear. That can be seen like triple the cost you factored in already.

And also, your renewable energy source won’t work 100% of time and not at 100% of the capacity. Which also implies that you need to build multiple time what you planned to really get that planned capacity.

I'm not sure what the service life of a wind turbine is, but 20 years is about right for a photovoltaic solar panel. The power output begins to fall off drastically at that point.

It's about 15-20 years for wind turbines, depending on the circumstances.

LCOE is usually used to more easily compare the investment cost. While this leads to more resource usage... Because of economies of scale, wind turbines from 20 years into the future will be less expensive, more efficient and overall better.

> So you see a doubled cost for nuclear.

I don’t know what you mean. I proposed spending the same total amount.

> And also, your renewable energy source won’t work 100% of time and not at 100% of the capacity. Which also implies that you need to build multiple time what you planned to really get that planned capacity.

This was already factored in to the calculation I was agreeing with. They said that to get sufficient capacity in renewables considering reduced availability, it still costs 1/2 what nuclear costs.

I will copy and paste the relevant section from that comment for clarity:

>> "To illustrate this point, the 2,430 MW Vogtle nuclear plant could be expected to generate 21 million MWh per year. That is enough to power about 1.75 million residential households. Meanwhile, a hypothetical 3,500 MW solar power plant would be able to produce just under 6 million MWh of electricity per year. This number is enough to power only 500,000 homes, which is considerably less than nuclear power. For solar to produce as much electricity as is generated by a nuclear power plant, it would require about 13,000 MW of utility-scale solar capacity, which about four times as much as built in the existing plants. However, the cost to build this 13,000 MW facility would be $12.1 billion, which is still just 50% of the cost of the $25 billion Vogtle nuclear plant."

> https://www.solarfeeds.com/mag/solar-power-vs-nuclear-power/...

Batteries are not a feasable seasonal storage. They can store for a couple of hours, no more.

For seasonal storage you need a hydrogen economy. H2 is very very expensive and has never been rolled out at scale. Not a single large gas turbine to burn pure H2 has ever been build.

In the middle of a crisis it's better to go for proven and economic technologies than trying to invent something completely new and untestet.

Reply on news.ycombinator.com