Grid-scale batteries: They’re not just lithium

When size and weight don't matter, lots of other battery chemistries can work.

See full article...

 

[paulfdietz]

Pumping water up the hill then collecting power from it when it runs back down is standard pumped-hydro. It's a great setup if you have enough elevation difference over a short distance and plenty of water available. Both of those are limiting factors.

The water needed isn't that much; it can be obtained from wells if nothing else. For example:

https://www.whitepinepumpedstorage.com/
5000 acre-feet to fill, 600 acre-feet year to top off. Capacity 8 GWh.

The average golf course in Southern Nevada consumes 725 acre-feet of water per year.

 

[avhn]

I'm not an expert in this area... but what about just pumping a lot of seawater up a mountain, and then slowly releasing it through a turbine generator whenever you need more electricity? There's no loss of stored energy over time like with chemical batteries.

Side benefit: if you happen to have extra energy, you can use the water pressure to push that seawater through reverse-osmosis filters and get freshwater, which is scarce in certain parts of the country, like California.

Is anyone doing something like this already?

There are several problems with doing this at large scale. OTTOMH:

1. It only works in areas with mountainous coasts.
2. You need a gigantic reservoir at the top of the mountain. Pumping a cubic meter (1,000 kg) of water up a moderate 1km mountain stores a measly 2.7 kWh.
3. If your reservoir leaks or fails, you need to be prepared to contain or direct it somehow, or else deal with an ecological disaster from releasing a lake of seawater into an inland ecosystem. Even slow leaks are potentially dangerous, salting the earth/groundwater with your pumped brine.

 

[zaco]

I'm not an expert in this area... but what about just pumping a lot of seawater up a mountain, and then slowly releasing it through a turbine generator whenever you need more electricity? There's no loss of stored energy over time like with chemical batteries.

Side benefit: if you happen to have extra energy, you can use the water pressure to push that seawater through reverse-osmosis filters and get freshwater, which is scarce in certain parts of the country, like California.

Is anyone doing something like this already?

Pumped hydro energy storage is like 90% of the current global energy storage market. But its drawbacks are that it takes up a lot of space and needs mountains. California already has a few significant mountain reservoirs, but it needs those for municipalities and agriculture.

 

[johnz]

Is any of this interesting in a home use environment? What I'm imagining is a battery service, where every house has a pad somewhere close by, and battery cubes get delivered when the old one wears out. Quick connect cables, a machine to move the cubes, and recycling at nearby(ish) depot.

Maybe not today, but lots of houses are getting solar installations. Pretty up the cubes so they don't look too industrial, and it'll become one of those things that's /just there/, like omnipresent HVAC appliances on the sides of houses.

 

[afidel]

Is any of this interesting in a home use environment? What I'm imagining is a battery service, where every house has a pad somewhere close by, and battery cubes get delivered when the old one wears out. Quick connect cables, a machine to move the cubes, and recycling at nearby(ish) depot.

Maybe not today, but lots of houses are getting solar installations. Pretty up the cubes so they don't look too industrial, and it'll become one of those things that's /just there/, like omnipresent HVAC appliances on the sides of houses.

Hmm, interesting concept but standardization is hard. The other major problem is cost, I wouldn't mind a pad mount battery that basically took the place of a standby generator but even without the power electronics a battery that could run my very modest house for a few days would be ~$10k or twice what a natural gas generator that can run it effectively indefinitely would cost.

 

[DDopson]

This depends on the storage use case. For very long term storage with few charge/discharge cycles, the "cost of inefficiency" becomes unimportant compared to the capex of energy storage capacity. On that, hydrogen can be far superior to batteries.

There was an illuminating discussion of this recently in another thread.

https://meincmagazine.com/civis/threa...rst-half-of-2024.1502569/page-7#post-43119624

Yes, in that thread we played with a grid simulation, https://model.energy/, which takes various assumptions as inputs, including a specific year worth of real-world weather data for your choice of country so that the tool can estimate the timing of wind and solar generation with 3hr granularity, and then the tool solves for the lowest-cost solution to satisfy a constant 100 MW demand 24/7/365 using a mix of wind + solar + batteries + hydrogen, and by setting a flag, I can allow the tool to use fossil fuels and / or nuclear. One limitation is that the tool isn't modeling a real-world demand curve, so it probably deploys less solar than the real-world optimum, but it's still a highly informative way to quantify the trade-offs, and is much better than random hand-waiving about the impact of renewables variability. It also changed my mind on a few things.

Some of the conclusions from that exercise:

* Lowest-cost = 85% renewables: Even when allowing fossil fuel gas turbines w/o a carbon tax, the lowest-cost solution for the US 2030 scenario w/ 2011 weather data obtains 85% of electricity from renewables+storage and only 15% of electricity from 65 MW of gas peakers (per 100 MW of constant demand) running 23.2% of the hours in the year.

* Hydrogen competes with gas peakers and is non-viable without a carbon tax - When fossil fuels were allowed w/o a carbon tax, the solution included zero hydrogen infrastructure. With a carbon tax of $50 / ton, the optimal solution included a small amount of hydrogen infrastructure and better than halved the gas peaker duty cycles such that gas was only 5% of all generation (the number of gas turbines decreasing by only a third, to 44 MW). With the carbon tax but disallowing hydrogen, gas peakers were 7.7% of generation, so allowing hydrogen in the solution was responsible for a material decrease.

* Over-generation isn't prohibitively expensive, but it's more expensive than hydrogen storage - compared to the lowest-cost solution where 15% of electricity is still coming from fossil fuel gas turbines, sufficiently over-provisioning wind+solar+batteries to satisfy all demand costs +30% more (huge increase in battery capex), but when including hydrogen storage in the solution, the cost premium for zero emissions drops to less than +10%. What really blew my mind, because I'm used to thinking of hydrogen as a boondoggle, is that hydrogen continued to have a significant role in the lowest-cost zero-emissions solution even when I strongly pessimized the hydrogen cost assumptions, including a 100X increase in hydrogen storage cost (to match some numbers for aboveground steel tanks), a 10X increase in electrolyzer cost, or a 10X increase in hydrogen turbine cost.

* Seasonal storage is mis-understood - yes, the hydrogen discharge occurred more in the winter than in the summer, but the main driver of hydrogen discharge was managing multi-day fluctuations in wind power output. I figured this out by opening the hour-by-hour simulation data in Excel. Even during the winter, there were periods where excess wind power enabled the electrolyzers to add hydrogen to the storage reservoir. Even during the summer/spring/fall, there were periods where hydrogen discharge would cover for lulls in wind generation. Hydrogen and wind power were best friends. Solar was best friends with batteries. On net there was a trend of shifting power from summer (when solar was at a max, and demand was unrealistically constant) to winter, but the storage was used 3.36X per year, not the 1X you'd expect for pure seasonal shifting with only one charging period and one discharging period.

Overall, that exercise gave me a much more nuanced understanding of the potential role for long-term energy storage solutions. Which is why I was so bummed to see a total lack of numbers in this article, as I'd love to plug some cost assumptions into the model and what role these technologies might have in the model's solutions.

 

[C.M. Allen]

Hmm, interesting concept but standardization is hard. The other major problem is cost, I wouldn't mind a pad mount battery that basically took the place of a standby generator but even without the power electronics a battery that could run my very modest house for a few days would be ~$10k or twice what a natural gas generator that can run it effectively indefinitely would cost.

1) Cost depends on the storage tech and how mature it is. Non-lithium alternatives aren't fully mature or at mass-production scale, so their cost is not reflective of what it could easily be in ~10 years.

2) Electricity storage is an energy abstraction and can come from all manner of sources, including by the homeowner with roof solar or small wind turbines, etc, which is 'free' energy.

3) A NG generator is not 'indefinitely' -- it needs a feed line to a NG supplier or a truck to refill it, and NG costs are heavily subsidized and not all reflective of its true market cost without those subsidies. Even worse, as NG demand goes down in response to more renewable energy, the FIXED operational costs of NG extraction, transportation, and refinement will have to be recouped by inflating that price further. This is a death spiral, and precisely what the people selling NG are trying to stop or delay as long as possible.

 

[James McP]

Personally, I like the iron flow batteries that are being used for grid scale storage. Their primary ingredients are abundant (water, salt and iron), they have very long lifespans (25+ years) with near zero degradation, they're not explosive/flammable, and they have a very wide operating temperature range. The one thing they lack, energy density, is not really an issue for grid scale storage. I've never really seen a great critique on their shortcomings, but from what I've read they seem ideal.

Their power is low relative to their energy. Iirc, it takes 100hrs to discharge. If you have 100MWhr of storage, you can only pull 1MW. If you had 100MWhr of Li-Ion, you could pull up to 100MW.

The math on high-power/high-cost batteries like Li-ion vs low-power/low-cost batteries will come down to use case and available land. It is probably a good thing for solar and wind farms where the land is a sunk cost and the batteries would guarantee a cost effective level of base load over days.

It's probably not a good idea for residential or commercial use, outside of some rare cases like maybe a hospital that would need 4 days of power but couldn't have a generator/fuel tank for some reason.

 

[DDopson]

Internet is full of data about most of these technologies, they are not exactly top secret.

When I search for "Zinc-bromine cost per mwhr", "Sulfur-sodium batteries cost per mwhr", or "Redox-flow cost per mwhr" I get a bunch of different prices, mostly somewhere between $300 to $500 per kWhr, which is several times more expensive than vanilla Lithium-ion battery packs.

This article is based on conversations with several start-ups, and I'd like to think that those start-ups have their own aspirational price targets that are hopefully a bit rosier than whatever 2020 DoD reports I can pull up on these technologies. Surely, their pitch isn't "pay 3X more per kWhr because it's a flow battery".

And yeah, that information doesn't seem to be public. I dug around on several of their websites and couldn't find any price numbers, just glossy non-quantitative pitch materials.

Price is the single most important aspect of a new battery technology. We already have batteries that are excellent on all other performance metrics - round-trip efficiency, durability, discharge rate, etc. What we desperately need is something that's cheaper per MWhr of storage capacity. Something to sit between Li-ion and the hypothetical role hydrogen storage could play. The entire point of flow batteries is to be cheaper per unit of energy storage, so price needs to be part of the pitch deck. Otherwise it's just a story.

If these technologies can't beat Lithium-ion on $/MWhr of storage capacity, then it won't matter how sexy their engineering story is. A flow battery that isn't cheaper per unit of energy storage has no role in our grid.

 

[DDopson]

See: https://en.wikipedia.org/wiki/Mediterranean-Dead_Sea_Canal. That's a fairly straightforward hydroelectric scheme, not a battery, and not based on any kind of pumping. Once-through.

One could conceive of something similar between, say, the Gulf of California and the Salton Sea, but the difference in elevation is less and the collateral damage (loss of agriculture and other things - like the Coachella festival - in the region as Lake Coahuilla redevelops); and there's speculation that the southern San Andreas would come unstuck if the Lake reappeared providing water to lubricate the fault.

Pumping water up the hill then collecting power from it when it runs back down is standard pumped-hydro. It's a great setup if you have enough elevation difference over a short distance and plenty of water available. Both of those are limiting factors.

It's also worth noting that even traditional hydro can look surprisingly similar to a battery when there's flexibility to choose when to run the turbines. Sometimes the downstream watershed has minimum flow requirements, and the turbines have a maximum power rating, so there are some limits on this, but given the scale at which hydro is deployed, the effect is quite significant.

On the CAISO grid, the difference between hydro's daytime minimum and hydro's maximum during the evening net-demand peak is roughly 3.8 GW.

 

[xoe]

Are some of these small enough to fit in a big truck? Because I wanna know how long an electrical cargo truck would last running hooked to one of these.

Now I'm imagining AUS style road trains hauling loads with one of these in the line, eco friendly trucking at it's best.

 

[kltye]

Indeed - too many people seem to forget there's other technology that has really good attributes.

One I've often seen come up is people wanting Lithium battery powered UPSs instead of "old obsolete lead acid" for home network gear, followed by complaining about the high cost and how many portable power stations that can be used as an EPS unit require manual cycling to keep the BMS in calibration since Lithium has such a flat discharge curve the calibration drifts if it doesn't hit the ends.

Except for something that will sit its entire life in a single place without moving, heavy lead batteries don't really matter, have a much better discharge curve for identifying SOC, don't really care about sitting at 100% for months or years on end, and are WAY cheaper.

My problem with "old obsolete" SLA is its high internal resistance, and how you really shouldn't run it to near 0% SoC, even for deep-cycle batteries. High internal resistance means you can't easily put a prolonged high load on it (voltage dips too quickly at lower SoC), and a reduced depth-of-discharge rate means you're actually only able to access 60-70% of the energy available in lead acid.

With LFP batteries being ridiculously cheap these days - and I would say almost on par with SLA per Wh, since you can actually use more of the energy stored - there's no good reason to stay with SLA for indoor use. (Not being able to charge below freezing point is a significant drawback for some applications, I'll concede that)

LFPs can sit at 100% for a prolonged time without degradation (it's not that your SLAs in UPSes are staying good forever either), and a good BMS with a shunt and proper calibration shouldn't have any issues figuring out pack SoC.

 

[StikyPad]

While weight might not matter (much), size probably does matter to a larger degree, because the batteries have to be stored somewhere, and larger spaces are inherently more expensive to build and maintain. It's the same reason a lithium ion battery still makes more sense than a lead-acid battery for solar energy storage at home, even though a garage doesn't really care how much the batteries weigh (again, to an extent), but you probably do care about how much of your garage has to be dedicated to batteries.

 

[numerobis]

Thank you for this. We don't usually hear much about the non-Lithium Ion battery tech currently in use.

“In use” is doing a lot of heavy lifting there. These aren’t in widespread deployment.

The Chinese battery industry has basically crushed everything in its path when it comes to alternatives to lithium ion. Technologies that seemed promising even just a handful of years ago have been faced with lithium ion battery prices falling by more than an order of magnitude.

 

[Properjob70]

“In use” is doing a lot of heavy lifting there. These aren’t in widespread deployment.

The Chinese battery industry has basically crushed everything in its path when it comes to alternatives to lithium ion. Technologies that seemed promising even just a handful of years ago have been faced with lithium ion battery prices falling by more than an order of magnitude.

I guess that, like wind & solar a couple of decades ago, governments will have to give incentives for currently price-uncompetitive technologies to gain a foothold in the market. Wind/solar only become price-competitive when produced at scale, so the howls of outrage from fossil fuel interests lasted several election cycles. It's difficult to pick winners 20 years from now, but the pitfalls of going all-in on scaling up lithium technologies as a default are at least apparent when making other choices to sponsor

 

[afidel]

While weight might not matter (much), size probably does matter to a larger degree, because the batteries have to be stored somewhere, and larger spaces are inherently more expensive to build and maintain. It's the same reason a lithium ion battery still makes more sense than a lead-acid battery for solar energy storage at home, even though a garage doesn't really care how much the batteries weigh (again, to an extent), but you probably do care about how much of your garage has to be dedicated to batteries.

Not really, to give a sense of perspective Iron flow batteries might only achieve 30Wh/L, a small fraction of the 210Wh/L of LFP but a standard olympic swimming pool is about 4M Liters so in a relatively small structure you can store 120MWh. The site of a small coal power plant near me that was recently cleared was 62 acres, you could fit 200 pools on that site or 24GWh of Iron flow battery. The power plant was only 250MW when it was operating meaning in the same footprint you can store about 100 hours of what the old plant could output. So land use is not going to be the limiting cost factor at grid scale.

Edit
I missed I decimal place, it's 24GWh, not 2.4, makes it even less of an issue

 

[leonwid]

Yes, in that thread we played with a grid simulation, https://model.energy/, which takes various assumptions as inputs, including a specific year worth of real-world weather data for your choice of country so that the tool can estimate the timing of wind and solar generation with 3hr granularity, and then the tool solves for the lowest-cost solution to satisfy a constant 100 MW demand 24/7/365 using a mix of wind + solar + batteries + hydrogen, and by setting a flag, I can allow the tool to use fossil fuels and / or nuclear. One limitation is that the tool isn't modeling a real-world demand curve, so it probably deploys less solar than the real-world optimum, but it's still a highly informative way to quantify the trade-offs, and is much better than random hand-waiving about the impact of renewables variability. It also changed my mind on a few things.

Some of the conclusions from that exercise:

* Lowest-cost = 85% renewables: Even when allowing fossil fuel gas turbines w/o a carbon tax, the lowest-cost solution for the US 2030 scenario w/ 2011 weather data obtains 85% of electricity from renewables+storage and only 15% of electricity from 65 MW of gas peakers (per 100 MW of constant demand) running 23.2% of the hours in the year.

* Hydrogen competes with gas peakers and is non-viable without a carbon tax - When fossil fuels were allowed w/o a carbon tax, the solution included zero hydrogen infrastructure. With a carbon tax of $50 / ton, the optimal solution included a small amount of hydrogen infrastructure and better than halved the gas peaker duty cycles such that gas was only 5% of all generation (the number of gas turbines decreasing by only a third, to 44 MW). With the carbon tax but disallowing hydrogen, gas peakers were 7.7% of generation, so allowing hydrogen in the solution was responsible for a material decrease.

* Over-generation isn't prohibitively expensive, but it's more expensive than hydrogen storage - compared to the lowest-cost solution where 15% of electricity is still coming from fossil fuel gas turbines, sufficiently over-provisioning wind+solar+batteries to satisfy all demand costs +30% more (huge increase in battery capex), but when including hydrogen storage in the solution, the cost premium for zero emissions drops to less than +10%. What really blew my mind, because I'm used to thinking of hydrogen as a boondoggle, is that hydrogen continued to have a significant role in the lowest-cost zero-emissions solution even when I strongly pessimized the hydrogen cost assumptions, including a 100X increase in hydrogen storage cost (to match some numbers for aboveground steel tanks), a 10X increase in electrolyzer cost, or a 10X increase in hydrogen turbine cost.

* Seasonal storage is mis-understood - yes, the hydrogen discharge occurred more in the winter than in the summer, but the main driver of hydrogen discharge was managing multi-day fluctuations in wind power output. I figured this out by opening the hour-by-hour simulation data in Excel. Even during the winter, there were periods where excess wind power enabled the electrolyzers to add hydrogen to the storage reservoir. Even during the summer/spring/fall, there were periods where hydrogen discharge would cover for lulls in wind generation. Hydrogen and wind power were best friends. Solar was best friends with batteries. On net there was a trend of shifting power from summer (when solar was at a max, and demand was unrealistically constant) to winter, but the storage was used 3.36X per year, not the 1X you'd expect for pure seasonal shifting with only one charging period and one discharging period.

Overall, that exercise gave me a much more nuanced understanding of the potential role for long-term energy storage solutions. Which is why I was so bummed to see a total lack of numbers in this article, as I'd love to plug some cost assumptions into the model and what role these technologies might have in the model's solutions.

Thanks for sharing this in-depth analysis. Even without demand modeling this is already very informative and it certainly reset my opinion on hydrogen.

 

[alexvoda]

Indeed - too many people seem to forget there's other technology that has really good attributes.

One I've often seen come up is people wanting Lithium battery powered UPSs instead of "old obsolete lead acid" for home network gear, followed by complaining about the high cost and how many portable power stations that can be used as an EPS unit require manual cycling to keep the BMS in calibration since Lithium has such a flat discharge curve the calibration drifts if it doesn't hit the ends.

Except for something that will sit its entire life in a single place without moving, heavy lead batteries don't really matter, have a much better discharge curve for identifying SOC, don't really care about sitting at 100% for months or years on end, and are WAY cheaper.

Rather than a Li-Ion UPS, what I would like is a integrated PSU + UPS to avoid going from AC->DC->AC->DC.

 

[Ultor]

These technologies are all very interesting, but what I don't see on any of them is numbers for $/MW and $/MWhr. I'd like to think that even in the research phase, there would be an eye towards the eventual at-scale cost targets.

Yeah, this article was a bit light on the numbers. Also, regarding the second battery described: provide “ power for five to 10 hours” is not a helpful statement. It could probably power my laptop for weeks, but so what? We need its capacity and maximum discharge rate, etc.

 

[sorten]

So, engineering then?

 

[Malmesbury]

I'm not an expert in this area... but what about just pumping a lot of seawater up a mountain, and then slowly releasing it through a turbine generator whenever you need more electricity? There's no loss of stored energy over time like with chemical batteries.

Side benefit: if you happen to have extra energy, you can use the water pressure to push that seawater through reverse-osmosis filters and get freshwater, which is scarce in certain parts of the country, like California.

Is anyone doing something like this already?

That’s called pumped storage

https://en.wikipedia.org/wiki/Dinorwig_Power_Station
Good luck finding a mountain that you will be allowed, these days, to hack a great big hole into the top of.

I’ve calculated that Li-Ion batteries could do the same thing for about the same price.

The other benefit the batteries have is that they scale to zero. They can start with a single ISO container - which holds 3-4MWh.

In the U.K., that means no planning permission required. Just park it on your land and hook it up. A chap I know is getting storage, on that basis, for his solar farm - time shifting to create 24/7 power production.

 

[SiberX]

Indeed - too many people seem to forget there's other technology that has really good attributes.

One I've often seen come up is people wanting Lithium battery powered UPSs instead of "old obsolete lead acid" for home network gear, followed by complaining about the high cost and how many portable power stations that can be used as an EPS unit require manual cycling to keep the BMS in calibration since Lithium has such a flat discharge curve the calibration drifts if it doesn't hit the ends.

Except for something that will sit its entire life in a single place without moving, heavy lead batteries don't really matter, have a much better discharge curve for identifying SOC, don't really care about sitting at 100% for months or years on end, and are WAY cheaper.

Lead-acid batteries also have much worse lifespans (3-10 years) than lithium ion (5-20 years), especially LFP. When you factor in the cycle counts and replacement intervals, lithium batteries can come in pretty competitive with lead-acid over long periods and save the grief and downtime of replacing on as frequent an interval.

LFP is denser than lead-acid, can be cycled far more deeply without damage, and has better round-trip efficiency. You can fit the same effective capacity in a smaller space (and it weighs less) so if you expect your installation to last more than 5 years or so, it's very much worth considering the more expensive (up front) lithium options.

 

[Scuba-Man-1970]

I was musing about how much trouble I could get into with my Electrical Safety Authority (ESA [Ontario, Canada]) if I decided to get one of those hot Sulphur rigs plopped onto our property. Right beside our off-grid home (under construction). If money wasn't a issue for me.

Apropo of nothing: here's an image of our system (work in progress):

... strictly fyi eh. The Li-Ion rack will be installed behind that BBQ owned by the electricians!

 

[ChrisRed]

Yes, in that thread we played with a grid simulation, https://model.energy/, which takes various assumptions as inputs, including a specific year worth of real-world weather data for your choice of country so that the tool can estimate the timing of wind and solar generation with 3hr granularity, and then the tool solves for the lowest-cost solution to satisfy a constant 100 MW demand 24/7/365 using a mix of wind + solar + batteries + hydrogen, and by setting a flag, I can allow the tool to use fossil fuels and / or nuclear. One limitation is that the tool isn't modeling a real-world demand curve, so it probably deploys less solar than the real-world optimum, but it's still a highly informative way to quantify the trade-offs, and is much better than random hand-waiving about the impact of renewables variability. It also changed my mind on a few things.

Some of the conclusions from that exercise:

* Lowest-cost = 85% renewables: Even when allowing fossil fuel gas turbines w/o a carbon tax, the lowest-cost solution for the US 2030 scenario w/ 2011 weather data obtains 85% of electricity from renewables+storage and only 15% of electricity from 65 MW of gas peakers (per 100 MW of constant demand) running 23.2% of the hours in the year.

* Hydrogen competes with gas peakers and is non-viable without a carbon tax - When fossil fuels were allowed w/o a carbon tax, the solution included zero hydrogen infrastructure. With a carbon tax of $50 / ton, the optimal solution included a small amount of hydrogen infrastructure and better than halved the gas peaker duty cycles such that gas was only 5% of all generation (the number of gas turbines decreasing by only a third, to 44 MW). With the carbon tax but disallowing hydrogen, gas peakers were 7.7% of generation, so allowing hydrogen in the solution was responsible for a material decrease.

* Over-generation isn't prohibitively expensive, but it's more expensive than hydrogen storage - compared to the lowest-cost solution where 15% of electricity is still coming from fossil fuel gas turbines, sufficiently over-provisioning wind+solar+batteries to satisfy all demand costs +30% more (huge increase in battery capex), but when including hydrogen storage in the solution, the cost premium for zero emissions drops to less than +10%. What really blew my mind, because I'm used to thinking of hydrogen as a boondoggle, is that hydrogen continued to have a significant role in the lowest-cost zero-emissions solution even when I strongly pessimized the hydrogen cost assumptions, including a 100X increase in hydrogen storage cost (to match some numbers for aboveground steel tanks), a 10X increase in electrolyzer cost, or a 10X increase in hydrogen turbine cost.

* Seasonal storage is mis-understood - yes, the hydrogen discharge occurred more in the winter than in the summer, but the main driver of hydrogen discharge was managing multi-day fluctuations in wind power output. I figured this out by opening the hour-by-hour simulation data in Excel. Even during the winter, there were periods where excess wind power enabled the electrolyzers to add hydrogen to the storage reservoir. Even during the summer/spring/fall, there were periods where hydrogen discharge would cover for lulls in wind generation. Hydrogen and wind power were best friends. Solar was best friends with batteries. On net there was a trend of shifting power from summer (when solar was at a max, and demand was unrealistically constant) to winter, but the storage was used 3.36X per year, not the 1X you'd expect for pure seasonal shifting with only one charging period and one discharging period.

Overall, that exercise gave me a much more nuanced understanding of the potential role for long-term energy storage solutions. Which is why I was so bummed to see a total lack of numbers in this article, as I'd love to plug some cost assumptions into the model and what role these technologies might have in the model's solutions.

Thank you for this, here in the UK and likely everywhere, there are lots of very loud voices shouting about wind not blowing or sun not shining. I wish they'd just listen to experts for once without whining. Renewables are the essential majority component of a modern energy grid and anyone arguing otherwise is a fool or paid to say so. Debate should be about the lowest cost way to get peaker-equivalent generation, not the rest.

 

[bjn]

Other chemistries are interesting and can hopefully fill other niches more cost effectively than lithium chemistries. However, given that currently most grid storage batteries have lithium iron phosphate chemistries, the repeated assertion in the article like the one below is just coal hugger FUD.

This means that they do not require any materials that come from conflict countries that have unsafe working conditions.

The main exporter of lithium is Australia, a country not noted for forcing African children down lithium mines, and while the country’s politics can be quite feisty, it’s not exactly a conflict country. Oz is pretty big on iron exports as well, and phosphate is widely available and its usage in batteries is a rounding error as compared to what is used as fertiliser.

 

[UpNorth2]

I'm not an expert in this area... but what about just pumping a lot of seawater up a mountain, and then slowly releasing it through a turbine generator whenever you need more electricity? There's no loss of stored energy over time like with chemical batteries.

Side benefit: if you happen to have extra energy, you can use the water pressure to push that seawater through reverse-osmosis filters and get freshwater, which is scarce in certain parts of the country, like California.

Is anyone doing something like this already?

Hydro power is the golden standard for long term energy storage. You do not want to do this with salt water, as it affects the soil. The big issue is that most places that are suitable for hydro power are already in use.

 

[MilkyBarKid]

Pumped hydro energy storage is like 90% of the current global energy storage market. But its drawbacks are that it takes up a lot of space and needs mountains. California already has a few significant mountain reservoirs, but it needs those for municipalities and agriculture.

It doesn’t necessarily need mountains or space; caves will work. https://arena.gov.au/blog/compressed-air-to-secure-power-supply-for-broken-hill/

This pilot will use two pipes into a cave. One pipe pumps in air, the other pipe has a water reservoir on top of it. Rather than pushing water through the turbine, air is pushed through the turbine with the pressure/force and cooling provided by the water.

Australian is like California only drier and flatter, and it has a range of pumped storage systems in development.

 

[bjn]

Hydro power is the golden standard for long term energy storage. You do not want to do this with salt water, as it affects the soil. The big issue is that most places that are suitable for hydro power are already in use.

Pumped hydro is different to trad hydro. Trad hydro you damn a river and flow water through turbines once to generate power. The locations you can do that is limitted by the geography and having a sufficiently large river to make it worth while. Pumped hydro recycles the same water betweeen two reservoirs, you only need to replenish losses due to evaporation and leakage. This opens alot more locations for pumped hydro. For example, you can't damn Loch Ness for trad hydro, but you can use it as the bottom reservoir for a pumped hydro scheme.

 

[InIgnem]

So I have the same musing a lot of you do - we used about 16.5MW of power in the last year at my home, and so I would have loved to see more details about how many MWH/etc. these things could provide. As somebody else mentioned above:

Is any of this interesting in a home use environment? What I'm imagining is a battery service, where every house has a pad somewhere close by, and battery cubes get delivered when the old one wears out. Quick connect cables, a machine to move the cubes, and recycling at nearby(ish) depot.

Maybe not today, but lots of houses are getting solar installations. Pretty up the cubes so they don't look too industrial, and it'll become one of those things that's /just there/, like omnipresent HVAC appliances on the sides of houses.

Yes, you can currently get generators that run on LNG/etc. around here for the occasional power outage, but imagine if every home had a "power bank" that could power their house for upto 6-12 months during a power outage. And even in communities with smaller lots, a "small" one of these providing upto a 1-3 month supply? Especially with a wide solar initiative, it seems like we could then really focus on optimizing our power grid and lowering costs. (Not that the privatized monopolistic power companies in the US would allow it, but....)


All that aside, it's cool to watch how the power storage industry is moving forward esp. with advances in car and solar technology. Slow but steady progress!

 

[Vertigre]

These technologies are all very interesting, but what I don't see on any of them is numbers for $/MW and $/MWhr. I'd like to think that even in the research phase, there would be an eye towards the eventual at-scale cost targets.

Those numbers won't be super interesting here, as they would change vastly with larger scale.

Maybe the best way to get a number like this would be to just add up the raw materials cost, since that's relatively unchanging and can't easily be optimized away like the assembly costs can. I suspect all three of these are quite low from that stand point, not including any rare or expensive ingredients.

 

[rochefort]

There are several problems with doing this at large scale. OTTOMH:
<snip>
3. If your reservoir leaks or fails, you need to be prepared to contain or direct it somehow, or else deal with an ecological disaster from releasing a lake of seawater into an inland ecosystem. Even slow leaks are potentially dangerous, salting the earth/groundwater with your pumped brine.

Look, if salting the earth was a problem, surely you'd hear the Carthaginians complaining about it, right?

 

[GreenEggsAndCrack]

Personally, I like the iron flow batteries that are being used for grid scale storage. Their primary ingredients are abundant (water, salt and iron), they have very long lifespans (25+ years) with near zero degradation, they're not explosive/flammable, and they have a very wide operating temperature range. The one thing they lack, energy density, is not really an issue for grid scale storage. I've never really seen a great critique on their shortcomings, but from what I've read they seem ideal.

Don't iron chemistry batteries have high self-discharge rates? Maybe this isn't an issue for grid storage, since they'll be charged and drained typically within 12 hours or so.

A while ago we looked at nickel-iron batteries for an off-grid solar system, but apart from the high cost, the high self-discharge rate was an issue. Expecting the system to run several days into a cloudy period seemed a little sketchy.

 

[mcswell]

"In the stack, zinc is plated onto a carbon surface during the charging process. It is then dissolved into the liquid during the discharging process," When I was a kid (before most of you were born), we had carbon-zinc batteries. Terrible, always running out of juice when you most needed them, and of course non-rechargeable. I'm glad to hear someone has fixed that technology.

Oh, and get off my lawn.

 

[mcswell]

Look, if salting the earth was a problem, surely you'd hear the Carthaginians complaining about it, right?

Carthago delenda est. Ceterum autem censeo, Moskva delenda est.

 

[DovePig]

The Chinese battery industry has basically crushed everything in its path when it comes to alternatives to lithium ion. Technologies that seemed promising even just a handful of years ago have been faced with lithium ion battery prices falling by more than an order of magnitude.

Northvolt already produces sodium‑ion batteries at Skellefteå (Sweden), don't they? In addition to their other lithium‑ion gigafactories there. 160 Wh/kg, IIRC, with big plans to scale up production (of course, who knows if that will work out). And they aren't just your small startup, they are already one of the biggest European li‑ion battery producers, I think.

But you are still right, most of sodium‑ion batteries is made in China.

 

[numerobis]

While weight might not matter (much), size probably does matter to a larger degree, because the batteries have to be stored somewhere, and larger spaces are inherently more expensive to build and maintain. It's the same reason a lithium ion battery still makes more sense than a lead-acid battery for solar energy storage at home, even though a garage doesn't really care how much the batteries weigh (again, to an extent), but you probably do care about how much of your garage has to be dedicated to batteries.

Far simpler reason: it’s a lot cheaper to use lithium ion batteries than lead acid for that. This started being true about ten years ago, and is an order of magnitude more true today. You can’t just look at the cost to buy the cells, you also need to keep track of how long they’ll last.

Cheap enough for grid use was hit back in 2018 or so for lithium, whereas it’s never been true for lead acid.

 

[numerobis]

Northvolt already produces sodium‑ion batteries at Skellefteå (Sweden), don't they? In addition to their other lithium‑ion gigafactories there. 160 Wh/kg, IIRC, with big plans to scale up production (of course, who knows if that will work out). And they aren't just your small startup, they are already one of the biggest European li‑ion battery producers, I think.

But you are still right, most of sodium‑ion batteries is made in China.

NaS batteries might actually work out; you can tell because the Chinese have started producing them. Unlike all the systems mentioned in this story.

 

[raschumacher]

Meanwhile the state of Texas has a $5 billion (next year, $10 billion) fund providing loans at 3% for building new fossil-fueled plants. The fund explicitly excludes batteries and renewables.

 

[sartalon]

Indeed - too many people seem to forget there's other technology that has really good attributes.

One I've often seen come up is people wanting Lithium battery powered UPSs instead of "old obsolete lead acid" for home network gear, followed by complaining about the high cost and how many portable power stations that can be used as an EPS unit require manual cycling to keep the BMS in calibration since Lithium has such a flat discharge curve the calibration drifts if it doesn't hit the ends.

Except for something that will sit its entire life in a single place without moving, heavy lead batteries don't really matter, have a much better discharge curve for identifying SOC, don't really care about sitting at 100% for months or years on end, and are WAY cheaper.

What is really appealing to me about redox flow over lead-acid and lithium, is its charge/discharge life and overall shelf life. You are looking at a battery that just needs mild upkeep every 10 years and last for 30, with very little loss of capacity. But I'm also expecting it to take up a big chunk of space.

 

[DovePig]

NaS batteries might actually work out; you can tell because the Chinese have started producing them. Unlike all the systems mentioned in this story.

Northvolt's isn't NaS, and the newer Chinese ones aren't either, IIRC. Na‑ion, instead. Most of the current production and research seems to be focused on Prussian blue analogues as the cathode, with carbon anodes. I don't think there is any sulphur in the electrolyte either. There was some research in China on MoS₂ anodes, though.