[iso1631]

Not all batteries need to be mobile.

A 1 GWh grid scale battery takes up about 4 hectares at the moment. The UKs total energy use is about 2,000GWh a day.

It would need to use 240,000 hectares to store all energy requirements (eletric, transport, heating etc) for a whole month. Even in extreme cold conditions it would last a couple of weeks.

That would be about 1% of land use.

[epistasis]

Storing a month of energy use doesn't make any sense when it comes to renewable grids. Since you don't have to mine/extract and transport an inventory, far far less storage is required.

[iso1631]

Which just emphasises the point -- density and thus land use isn't really a major concern

You could store an entire years worth of energy (not just electricity) for a fairly dense country like the UK and still have 95% of the country left for other usage.

[epistasis]

Definitely! However I frequently encounter people on energy discussion forums that assume that we need 2-3 months of battery storage, perhaps because they see natural gas or other storage and assume that batteries need exactly the same thing, so I'm perhaps overly eager to respond to claims about month-long battery storage.

I had previously been hopeful that battery storage would allow more shared used of land, but the fire risks of batteries have been pretty severe. And since the land usage requirements are fairly low, there's no need to enhance the risk by putting batteries, say, in enclosed spaces of former natural gas generation facilities [1].

I'm hopeful that we'll see a lot more storage, say a shipping container's worth, at the end of distribution feeders, which helps suck up residential solar with minimal resources, but solving the problem of "who pays for the benefits for all" when the utility is incentivized to keep grid costs high means that nobody is knocking down doors to make that happen...

[1] https://www.utilitydive.com/news/moss-landing-fire-battery-s...

[iso1631]

You'd have to compare the fire risk of a high density battery vs a substation.

But the beauty of battery and solar/wind is that it can be far more distributed. This of course means far harder to take out -- a wildfire near a nuclear power plant could knock out 3GW of capacity, a wildfire near one of 50x 60MW solar plants would knock out 60MW of capacity.

Solar might not be the only solution in a country at a high latitude, but for the contiguous 48 states even in winter solar produces more than enough power per hectare that when coupled with an appropriately sized battery can generate the entire needs.

1 hectare in sourthern california generates about 3.5MWh a day in December. In North Dakota about 1.5MWh a day, battery size for that is a tiny amount of land.

To generate enough solar energy to power the entire US electric needs, on a bad winter day, the US would have to have 60,000 square km of solar+battery.

The US currently uses 150,000 sqkm just to grow corn ethanol

You'd need a decent HVDC connectors from the southern states to the more demanding northern ones.

But even factoring in current prices of wind, solar, battery, and grid, it would only cost somewhere in the $6T range. Over 30 years that's $200b a year, or about $80 per MWh

New nuclear is about $150 per MWh before decommissioning costs

[laxpri]

you mean fission . thats why we need fusion

[womble2]

A month is probably excessive, but there are large regions of the world where its not uncommon for both solar and wind power to be running at <5% for multiple weeks in a row.

[epistasis]

There's an engineering tradeoff between having excess generation capacity, for that seasonal minimum, and having more storage. The cost optimal decision on that tradeoff will be determined by the ratio of excess generation cost and storage cost for rarely-used storage.

Until that ratio falls by at least 3x to 5x in favor of batteries being cheaper than generation, extra generation is going to be the way that grids actually get built out. Batteries and generation are both falling in cost fairly quickly, but generation still has the overall edge in learning rate. Cost decreases won't bottom out for at least a decade, because there's been no slow down yet, so I wouldn't expect this ratio to change for a minimum of 20 years, which means that pretty much a full energy system interchange will have happened by the time that this price ration changes.

So there's at least a few assumptions about the current industry and it's future development that underlie my assertion that a month of storage makes no sense, but I'm confident enough that I'd place money on the bet, and there's very very few things I'd bet on.

Edit: one thing that would break my assumption is the industrial development of storage that's super cheap for once-per-year usage. Most storage now needs to be cycled about 300x per year to make economic sense. "Long duration" storage is actually better defined as "economical storage at few battery cycles per year". There's nothing like that in the research hopper, but that doesn't mean it couldn't appear tomorrow and be deployed within a decade. Something that only gets used once or twice a year has to be dirt cheap, even if you could get 10x or 20x normal electricity prices for it.

[PunchyHamster]

you don't want industry going down coz there has been bad weather for few days in a row. And other alternative is burning gas.

If we want to get out of fossils, we will have to store for longer. Or built a ton of atomics

[barbazoo]

Doesn't that depend on the sources of energy though? Pretty much constant short and mid term supply like geothermal, hydro, tides, etc vs unpredictable or variable in the short term like wind and solar.

As a society relying on solar, I'd want to have more of a buffer than one that relies on hydro.

[_carbyau_]

Even if it doesn't work for the UK. Australia could do this easily.

Proportionally less population, less snow covered territory and general coldness = easier requirements.

Land coverage? To say Australia has plenty of land is an understatement.

To amp up(excuse the pun) this mental exercise (don't mind the practicalities of building it!) a bit:

If efficient super-long distance power transmission was a thing, we could have batteries and solar for the whole world and locate it where few people would ever see it.

If placed into a single contiguous blob though, I imagine the solar panels would cause their own microclimate by sucking 20% of the suns energy out of many hectares/acres of land... so that may be a problem.

[thelastgallon]

Instead of batteries, why not green hydrogen or green ammonia? Ammonia is needed for fertilizer, storage/logistics are a solved problem. Emergency plants around that can use ammonia as fuel can solve the dunkelflaute problem?

Another option is natural gas peaker plant on a ship/barge. Have a fleet of these around that can dock at any port and supply electrictity. A peaker powership is essentially a mobile, marine-class version of an onshore peaking power plant.

What would be best long term if we have a few hundred - few thousand nuclear ships/subs that can go anywhere and supply power.

It can be the sharing economy for clean power at scale.

[epistasis]

Capital costs and round-trip efficiency are the primary reasons. Even at constant utilization, electrolyzers are expensive enough that it's hard to replace fossil-fuel generated hydrogen at the moment.

Once you 2x-10x the cost of electrolyzer capital by only using it rarely, more generation and throwing away the excess electricity often makes the most sense.

Nuclear ships are similarly super super expensive. The only reason we build them at all are for their unique and wonderful operational capabilities, as in not needing to surface or refuel. Using nuclear ships for power would so expensive that we may need to up our GDPs 10x before such wasteful use makes sense. (Though I'm hoping we do reach such future luxurious lifestyles!)

[bawolff]

> Another option is natural gas peaker plant on a ship/barge. Have a fleet of these around that can dock at any port and supply electrictity. A peaker powership is essentially a mobile, marine-class version of an onshore peaking power plant.

Why would you want this?