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...

 

[The Lurker Beneath]

Because we don't need seasonal storage. We need better transmission. Season storage is a concept designed around a grid that uses fossil fuels that can be readily stored. Baseload power is a concept designed around thermal power plants that cannot be rapidly cycled. Both of these are strategies meant to force new technology to adapt to the reality of the existing grid, to the (often intentional) benefit of those existing energy sources for which the existing grid was designed.

You only need weeks of stored energy if you intent to source all your energy locally. We already don't do that, but we distribute that energy in physical form before it's converted to electrical. Yes, building out terawatts of long distance interconnects is a very costly undertaking, but so was building our existing fuel distribution networks, and so would be hundreds of TWh of storage.

Humans didn't invent the seasons, and half of biology in temperate latitudes can be linked to seasonal storage of energy. You can't just magik it away.

 

[The Geeman]

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.

There are four pumped-storage hydroelectric power stations in the UK providing 2.8 GW of installed electrical generating capacity, and contributing up to 4,075 GWh of peak demand electricity annually. Not using sea water but fresh water.

 

[wagnerrp]

Humans didn't invent the seasons, and half of biology in temperate latitudes can be linked to seasonal storage of energy. You can't just magik it away.

Biology migrates. It solves the seasonal power issue by moving elsewhere. We can do the opposite, solving the seasonal issue by moving the power instead of us.

 

[C.M. Allen]

LiFePO4 batteries have cheap ass inputs. It is just a membrane, plus graphite and the electrolyte. The price is mostly determined by Capex for the factory and labour. The material inputs are a tiny fraction of cost.

Yes, and the labor and factory conditions required for manufacturing lithium batteries are significantly higher than it is for some chemistries that are stable at room temperature, non-flammable, and don't require clean rooms to produce.

But you STILL need investors to pay for those factories. And right now, the biggest driver of batteries sales is EVs and consumer electronics that all use lithium, so that's what attracts investors. Grid-scale energy storage is a nascent industry that's fighting against much bigger interests that aren't aligned with what's best for grid-scale energy storage.

 

[paulfdietz]

My understanding was it was more a physical requirement than a chemical/electrical one. While the sodium melts at 98°C, below 300°C, it's just too "thick", and doesn't make good contact with the electrolyte. That does ultimately result in poor ion conduction, because you can't conduct across a void.

I think it's that the conductivity of the beta-alumina to sodium ions is too low below a certain temperature, although this is being improved.

https://chemistry-europe.onlinelibrary.wiley.com/doi/full/10.1002/batt.202100131

 

[TreeCatKnight]

Your claim is that hydrogen cannot be stored and distributed in large volume, but you're wrong, because it is being stored and distributed in large
volume. It has for the better part of a century. You reject that data as it's "grey hydrogen", but that's wrong, because regardless of the source, it is still evidence of hydrogen being stored and distributed in large volume. Then you go on to repeat your claim that we're not actually doing it, because... doing it with grey hydrogen doesn't count?

Don't get me wrong. I don't believe hydrogen will be used heavily for power generation, except perhaps out of convenience at facilities that are already producing/using it for other purposes. My reasons are because the RTE is only half that of even low performance systems currently in use, and the justification (large energy storage) doesn't make sense unless you're insisting that we work within the power grid as it currently exists. The problem of slow leakage is basically a non-issue.

For a data point, maximum permissible leak rates for hydrogen tanks in mobile applications is 1scc/hr/L. For a 10L bottle at 700bar, that's around 4.4M scc total volume, or decades before the tank is empty. To be clear, the reason that hydrogen leaks are a problem is because leaks in an enclosed room can produce explosive mixtures near the ceiling.

OK. I'm excitable. I like to co-mingle too many replies to too many people, so I can see the confusion.

Let me be more succinct. My replies are exclusively in the context of power generation, as it all started with my reply to Ddopson way way above.

My claim is not that hydrogen cannot be stored long term. My claim is that storing it long term is not a desirable thing. It's an expensive thing that you avoid if you can.

It also seems that we're talking past each other regarding the source of the hydrogen: in the context of reducing atmospheric carbon (as related to the articles subject of energy storage, a prime complement to renewable energy) we have to adjust our previous knowledge to account for green and not grey/blue/etc options. This drastically changes the math (and required facilities. I'm clearly not an expert, but electrolyzing H2 VS steam reforming has to have crazy different requirements, and some of that likely means transportation in useful quantities.)

It seems that another poster from above has some useful input that helps my case as well:

Is it being store and distributed in large volumes though? The main used for industrial H2 is ammonia to make fertiliser and AIUI the H2 is steam reformed from methane and then immediately turned into ammonia in the same facility, so very short pipes and zero storage. Storing H2 or moving H2 for other purposes is comparatively rare. I’ve searched for the sizes of existing H2 storage facilities and not found a total, as for salt domes and H2, it seems to be all pilot projects and nothing serious.

Surely there does exist long term storage, but my understanding is that it's just not super common, especially for this application.

In any case, I will also respond to Ddopson here because it is relevant:

Please do enlighten the class. What's the easier solution you are hand-waiving about?

Solar/wind and battery storage (the very subject of this fine article) already exist, are far cheaper, and don't rely on a non existent pipeline of green hydrogen to make it not add CO2.

I don't consider this hand-waiving.

Edit: added comment about electrolysis VS steam reforming facilities.

 

[DDopson]

Humans didn't invent the seasons, and half of biology in temperate latitudes can be linked to seasonal storage of energy. You can't just magik it away.

As seductive as that intuition may be, it's not a good description of the actual needs of our power grid.

In biology, most seasonal storage is about food storage due to crops not growing in winter. Grains are harvested in Autumn and used to sustain the family all winter long.

The power grid looks nothing like that.

Using real-world 2011 weather data for the continental US, with wind and solar locations sampled in proportion to the square of the capacity factor at that location (ie, prefer locations that generate more power), here's what the month-by-month wind and solar output looks like:

A few observations:

• Wind varies more than solar, with 1.93X between min vs max for wind, versus 1.68X for solar.
• Solar generates year round. Sure, it generates more in summer, less in winter, but except at high latitudes, wintertime generation is still very substantial.
• Total variation doesn't seem that bad (on this graph) - the combined total varies by only 1.2X from min to max, and the weakest month isn't even in the winter; it's a month where wind output was unusually low.

Looking at the monthly data (which is hiding some key details), "seasonality" doesn't really feel like the right description, at least not in the US, where the inclusion of the southern states make the solar curve pretty mild. Go to Norway, and seasonality is a much bigger problem. Still, there's an enormous fraction of humanity that lives in latitudes that are at least as favorable to solar as these US numbers.

It's when we look at the daily data that something interesting jumps out..

Solar has a moderate seasonal curve (higher in summer; lower in winter), with only a small amount of daily variation. I can tell you from my house's output curve that individual panels can drop significantly when a cloud passes over, but when you average over a broad geographical region, these panel-by-panel variations largely average out, making solar production very predictable.

Wind has large day to day, week to week, and month to month variability, and to the extent that there's any seasonal correlation at all, it's different than solar's seasonal correlation.

Comparing the avg-to-min, it's 2.2 for Solar, 2.9 for Wind, and 1.7 for the sum. This means that if we have diurnal storage, but no shifting between days, we need to over-provision +70% more generation capacity.

Thus in a grid that has long-term storage, while there's some seasonal trending, most of the back-and-forth is driven by managing that wind generation variability, with the storage capacity being used an average of 3.36X per year, not the 1X that you'd predict for a naive "seasonal" strategy:

This is one of the "conclusions" that I summarized in my post near the top of page 2.

 

[THT]

As part of eliminating net metering entirely, they instituted a "solar+storage" rate for residential systems. There are various tweaks available that provide greater or lesser discounts depending on how much of your battery you'll turn over to utility management. If you have enough battery to run off-grid during evening peak periods, you can get a significant discount on power used at other times. But reading between the various lines these rates are for residences that meet most of their needs from solar - essentially, running the battery as storage for local net metering, not as net power producers, and any surplus power delivered to the grid is paid for at avoided cost (around $0.05/kwh). They apparently figure that if you put in enough solar+storage to cover (annually) up to 110% of your recent total usage, you'll still pay for some power for a month or 2 in midwinter, and perhaps a bit in late summer/fall (a/c is needed well into October these days), much like I do now with Net Metering 1.0.

That's essentially how residential distributed generation is now set up. If you want to generate more power than that, you're treated as a commercial co-generation site with much more complicated requirements for interconnection and utility management, and a lower, variable payment rate at true wholesale (seldom more than $0.04/kwh, and sometimes zero). Plus, your system needs to be fully dispatchable including remote shutdown. This is the category I would put the system you described into, unless it's a large rural property designed for full off-grid operation and larger electrical loads than the average suburban or urban site.

The system I described was for a home that uses about 16.5 MWH per year. I didn't think it was an unusual number at all. It's the usage for your basic 2500 to 3500 ft2 house in Texas. That's not an average house, but 2500 to 3500 ft2 houses are pretty common. Electricity usage would be even more if the house had a pool. Those folks really have to pay for having the pleasure of a pool, $600, $700, $800 per month during the summer.

Yes, once you are generating a lot more than you use, it gets complicated. Obviously the grid will not let you make money on it. A lot of the power providers here just limit how much output power you can have. I haven't heard about remote shut down capability for a solar+storage system before though. The signaling is from cellular or WiFi I assume? What happens during a grid outage?

 

[OrvGull]

Nobody mentioned a turnkey UPS, but you can buy a 12V 100Ah LFP battery for ~$150 which is cheaper per Wh, and much cheaper per usable Wh.

I feel like you might be low on that. When I replaced the 80 Ah flooded lead-acid battery in my camper with a 100 Ah LiFePO4 battery, it cost me about $900. Granted that was a pre-packaged unit with built in BMS, but I doubt the BMS was $750 by itself.

 

[hpsgrad]

I feel like you might be low on that. When I replaced the 80 Ah flooded lead-acid battery in my camper with a 100 Ah LiFePo4 battery, it cost me about $900. Granted that was a pre-packaged unit with built in BMS, but I doubt the BMS was $750 by itself.

The whole conversation was weird. @mmiller7 posted up unit characteristics that don’t appear to make sense (I can’t find any current APC UPSs that are rack mountable, take a 24V battery and are compatible with a remote battery unit; maybe that’s my overlooking something? Maybe APC no longer makes an equivalent unit? Etc).

I also couldn’t find anything less than $2k MSRP with a power curve that supported more than an hour of runtime at ~200W. Smt3000rm2u

At the same time, UPS costs vary wildly based on factors other than raw battery capacity. Offline versus line-interactive, stepped approximation of a sine versus pure sine output, monitoring/management capabilities, and more make a big difference.

If someone wanted to chase this down, we would probably want to get some clarification about the detailed specs/model is being used for a reference, and work from there.

 

[OrvGull]

The whole conversation was weird. @mmiller7 posted up unit characteristics that don’t appear to make sense (I can’t find any current APC UPSs that are rack mountable, take a 24V battery and are compatible with a remote battery unit; maybe that’s my overlooking something? Maybe APC no longer makes an equivalent unit? Etc).

It's been a long time since I was doing that myself as opposed to leaving it up to a data center with its own Immortal Power Supply. I do remember having a 3000 VA rack-mount unit, but I think it had a 48 volt battery in the form of two 24 volt packs. By the time you need enough capacity that you're using four SLA batteries anyway, it starts to make sense to go up in voltage. The biggest problem I remember is if a pack failed and swelled up it became impossible to pull it out of the case to replace it. We generally only got about four years out of a pack but I think that was partly because the cooling in that part of the data center wasn't great. SLA battery life is pretty directly related to temperature.

 

[afidel]

I feel like you might be low on that. When I replaced the 80 Ah flooded lead-acid battery in my camper with a 100 Ah LiFePO4 battery, it cost me about $900. Granted that was a pre-packaged unit with built in BMS, but I doubt the BMS was $750 by itself.

Nope, they're now in the $130 to $230 range.

Here's one for $130:
https://a.co/d/49mSGdC
I bought this one for $199, it's heated with Bluetooth and and the ability to talk canbus:

Roypow 12V 100Ah Lithium Battery - Bluetooth and Self Heating https://a.co/d/0CQI94H

 

[wagnerrp]

Looking at the monthly data (which is hiding some key details), "seasonality" doesn't really feel like the right description, at least not in the US, where the inclusion of the southern states make the solar curve pretty mild. Go to Norway, and seasonality is a much bigger problem. Still, there's an enormous fraction of humanity that lives in latitudes that are at least as favorable to solar as these US numbers.

Regions can have seasons and downturns, but someone somewhere will have some sort of power. It's just a question of how far you have to go, and then building out the transmission infrastructure to allow you to do so.

 

[wagnerrp]

(I can’t find any current APC UPSs that are rack mountable, take a 24V battery and are compatible with a remote battery unit; maybe that’s my overlooking something? Maybe APC no longer makes an equivalent unit? Etc).

I've never used the external battery, but I know I've bought SmartUPS 2200 units in the past with a heavy gauge plug on the back for chaining to additional battery units.

 

[DDopson]

Solar/wind and battery storage (the very subject of this fine article) already exist, are far cheaper, and don't rely on a non existent pipeline of green hydrogen to make it not add CO2.

I don't consider this hand-waiving.

You need to read my original post more carefully. It considers that option and finds that it costs 130% instead of 110% of the polluting solution's cost.

Wind+Solar+Batteries are a fantastically cost-efficient way to produce the first ~85% of electricity, and even in a world without carbon taxes or other policy incentives, the lowest cost solution produces the vast majority of all electricity from renewables. That's why renewables deployments are scaling exponentially. It's just plain cheaper.

But then you reach a point where further renewables deployments generate most of their power on days where you've already satisfied 100% of demand, leading to curtailment. And the time gap between this curtailed power and the next day on which you need it is too long for battery storage to be economical. In a world without carbon taxes, the most economical solution is to pollute by running a gas turbine.

But what if you don't want to pollute at all? Is a zero-emissions grid feasible?

For context, one of the evergreen "discussion topics" is whether renewables, with their obviously variable power output, are capable of providing a stable, reliable power grid, or whether it's all just a greenwashing scam that can't possibly scale to more than a few percent of our power and there will always be a coal power plant running under the hood. In large part, the persistence of this topic is driven by "common-sense" observations that the sun doesn't shine at night, that clouds can cover an individual solar panel, and that the wind people perceive at ground level varies wildly, with long periods of little movement punctuated by occasional gusts. People often don't realize that the wind at a turbine's height is nothing like what they experience at ground level and that cloud shading effects can be almost totally averaged away by summing across a large number of panels.

Unfortunately, renewables variability is a topic that can only be properly addressed statistically. The post that you originally responded to summarizes my conclusions from a dozen hours of data analysis using real-world weather data to estimate Wind and Solar production hour-by-hour and understand the impact that variability has on the lowest-cost solution for providing reliable electricity under various assumption-sets. I'm using a model that's simplified in several ways, including using a constant demand curve, but the conclusions it reaches are broadly consistent with much more complex grid simulations that run on finer temporal and spacial granularity including transmission modeling. It's infinitely more accurate than anyone's subjective hand-waiving about renewables variability, which is something that we get a lot of in the media, in politics, and in these discussion forums.

When you say "Solar/wind and battery storage already exist" you are effectively proposing the over-provisioning solution where we continue deploying more solar, wind, and batteries until we satisfy all demand, along with curtailing a non-trivial fraction of that extra generation capacity. That costs 130% of the cheapest solution with 15% pollution, which isn't exactly unaffordable, but it's three times the cost premium of a solution that uses some over-provisioning of solar/wind/batteries, but then addresses the last 3.3% of residual unmet demand by burning expensive green hydrogen that was (inefficiently) produced during times of excess power. When including green hydrogen as energy storage, zero-emissions is possible for only 110% the cost of the polluting solution.

The hydrogen cost assumptions aren't particularly rosy, and not only would the scenario's green hydrogen be significantly more expensive than grey hydrogen produced from methane, it would be much more expensive than burning methane in a gas turbine. But it's cheaper than building wind and solar resources that are >90% curtailed. It's cheaper than doubling the number of batteries, which the over-provisioning scenario does in addition to deploying heavily curtailed renewables. It's cheaper than carbon sequestration. It's cheaper than nuclear.

That conclusion surprised me, so I played with the model extensively to see if it was sensitive to any of the cost assumptions, and under all plausible assumption sets, including aggressively pessimizing hydrogen cost and being beyond aggressively optimistic about battery and solar prices, zero-emissions was always cheaper when the last X% of electricity was solved by burning expensive green hydrogen. Changing the assumptions just meant that X would be 1% instead of 3%.

I don't expect to see green hydrogen get deployed anytime soon for the simple reason that it's currently cheaper to pollute. Until either the policy or cost assumptions change, our grid will continue moving in the direction of the 85/15 lowest cost solution. Putting a price on pollution would be smart, and efficient, but we seem pretty far from that being politically feasible.

 

[DDopson]

Regions can have seasons and downturns, but someone somewhere will have some sort of power. It's just a question of how far you have to go, and then building out the transmission infrastructure to allow you to do so.

The data I'm using is a composite of the entire continental US, so if anything, it's probably too optimistic, effectively similar to assuming we have unlimited power transmission between states. The more serious simulations try to model the transmission infrastructure bottlenecks, which will chop it up so that the Northern states see a stronger seasonal effect on their solar than is modeled in this dataset.

I'm sure we trade at least some power with Canada or Mexico, but that's not going to be significant enough to change the conclusions.

 

[numerobis]

I feel like you might be low on that. When I replaced the 80 Ah flooded lead-acid battery in my camper with a 100 Ah LiFePO4 battery, it cost me about $900. Granted that was a pre-packaged unit with built in BMS, but I doubt the BMS was $750 by itself.

The inverter is the expensive component beyond the battery. But the price is crashing; $900 must have been about two years ago? I paid that much CAD last year, and it’s cheaper today.

 

[Asbjorn Johansen]

This is way to deep in the comments for Ars staff to notice but, I'd love to see a survey of transmission technology. There is a significant coverage of generation and storage, but I don't think transmission gets the detailed coverage. What is the cost of transmission per mile? What is the power loss? Without knowing that, you can't really understand the best mix for generation and storage.

 

[afidel]

This is way to deep in the comments for Ars staff to notice but, I'd love to see a survey of transmission technology. There is a significant coverage of generation and storage, but I don't think transmission gets the detailed coverage. What is the cost of transmission per mile? What is the power loss? Without knowing that, you can't really understand the best mix for generation and storage.

Here's AEP's numbers for cost per mile in 2008 dollars:
------------------- ------------------
765 kV Single Circuit $2.6 – 4.0 Million
500 kV Single Circuit $2.3 - 3.5 Million
345 kV Double Circuit $1.5 - 2.5 Million
345 kV Single Circuit $1.1 – 2.0 Million

For losses:
LINE LOSSES - MW/100 MILES
Resistive Corona* Total
765 kV LINE @1000 MW LOAD --------- ------- -----------
Original 4-conductor (“Rail”) bundle 4.4 6.4 10.8 (1.1%)
Newer 4-conductor (“Dipper”) bundle 3.3 3.7 7.0 (0.7%)
Current 6-conductor (“Tern”) bundle 3.4 2.3 5.7 (0.6%)
Planned 6-trapezoidal cond. (“Kettle”) bundle 3.1 2.3 5.4 (0.5%)
500 kV LINE @1000 MW LOAD
Typical 2-conductor bundle 11.0 1.6 12.6 (1.3%)
345 kV LINE @1000 MW LOAD
Typical 2-conductor bundle 41.9 0.6 42.5 (4.2%)

HVDC line losses are ~3.5% per 1,000km (0.56% per 100 miles) but they have higher losses at the station and much higher station costs.

 

[Walker On Earth]

In any case, I will also respond to Ddopson here because it is relevant:



Solar/wind and battery storage (the very subject of this fine article) already exist, are far cheaper, and don't rely on a non existent pipeline of green hydrogen to make it not add CO2.

I don't consider this hand-waiving.

Edit: added comment about electrolysis VS steam reforming facilities.

 

[DDopson]

Here's AEP's numbers for cost per mile in 2008 dollars:
------------------- ------------------
765 kV Single Circuit $2.6 – 4.0 Million
500 kV Single Circuit $2.3 - 3.5 Million
345 kV Double Circuit $1.5 - 2.5 Million
345 kV Single Circuit $1.1 – 2.0 Million

For losses:
LINE LOSSES - MW/100 MILES
Resistive Corona* Total
765 kV LINE @1000 MW LOAD --------- ------- -----------
Original 4-conductor (“Rail”) bundle 4.4 6.4 10.8 (1.1%)
Newer 4-conductor (“Dipper”) bundle 3.3 3.7 7.0 (0.7%)
Current 6-conductor (“Tern”) bundle 3.4 2.3 5.7 (0.6%)
Planned 6-trapezoidal cond. (“Kettle”) bundle 3.1 2.3 5.4 (0.5%)
500 kV LINE @1000 MW LOAD
Typical 2-conductor bundle 11.0 1.6 12.6 (1.3%)
345 kV LINE @1000 MW LOAD
Typical 2-conductor bundle 41.9 0.6 42.5 (4.2%)

HVDC line losses are ~3.5% per 1,000km (0.56% per 100 miles) but they have higher losses at the station and much higher station costs.

Great data! The other big reason to use HVDC is if you need to run it a significant distance underwater where the reactive losses will be greater. Or if you are interconnecting two AC grids that aren't phase-synchronized.

HVDC is far more interesting because the utility can actually decide which direction to pump the power, versus AC power just "flows downhill". And because of all the sexy sexy semi-conductors in the switching stations.

I suspect that falling battery prices are going to kill my dream of East-West cross-oceanic HVDC links that deliver solar power all night long. It's feasible -- not really any harder than running fiber optics -- but will be increasingly less competitive as batteries get ever cheaper.

There's not a whole lot to AC distribution beyond using transformers to step up the voltage. Arguably, I think the mechanical aspects of developing lighter stronger cables is more interesting than anything that's happening from an electrical perspective.

 

[TreeCatKnight]

You need to read my original post more carefully. It considers that option and finds that it costs 130% instead of 110% of the polluting solution's cost.

Wind+Solar+Batteries are a fantastically cost-efficient way to produce the first ~85% of electricity, and even in a world without carbon taxes or other policy incentives, the lowest cost solution produces the vast majority of all electricity from renewables. That's why renewables deployments are scaling exponentially. It's just plain cheaper.

But then you reach a point where further renewables deployments generate most of their power on days where you've already satisfied 100% of demand, leading to curtailment. And the time gap between this curtailed power and the next day on which you need it is too long for battery storage to be economical. In a world without carbon taxes, the most economical solution is to pollute by running a gas turbine.

But what if you don't want to pollute at all? Is a zero-emissions grid feasible?

For context, one of the evergreen "discussion topics" is whether renewables, with their obviously variable power output, are capable of providing a stable, reliable power grid, or whether it's all just a greenwashing scam that can't possibly scale to more than a few percent of our power and there will always be a coal power plant running under the hood. In large part, the persistence of this topic is driven by "common-sense" observations that the sun doesn't shine at night, that clouds can cover an individual solar panel, and that the wind people perceive at ground level varies wildly, with long periods of little movement punctuated by occasional gusts. People often don't realize that the wind at a turbine's height is nothing like what they experience at ground level and that cloud shading effects can be almost totally averaged away by summing across a large number of panels.

Unfortunately, renewables variability is a topic that can only be properly addressed statistically. The post that you originally responded to summarizes my conclusions from a dozen hours of data analysis using real-world weather data to estimate Wind and Solar production hour-by-hour and understand the impact that variability has on the lowest-cost solution for providing reliable electricity under various assumption-sets. I'm using a model that's simplified in several ways, including using a constant demand curve, but the conclusions it reaches are broadly consistent with much more complex grid simulations that run on finer temporal and spacial granularity including transmission modeling. It's infinitely more accurate than anyone's subjective hand-waiving about renewables variability, which is something that we get a lot of in the media, in politics, and in these discussion forums.

When you say "Solar/wind and battery storage already exist" you are effectively proposing the over-provisioning solution where we continue deploying more solar, wind, and batteries until we satisfy all demand, along with curtailing a non-trivial fraction of that extra generation capacity. That costs 130% of the cheapest solution with 15% pollution, which isn't exactly unaffordable, but it's three times the cost premium of a solution that uses some over-provisioning of solar/wind/batteries, but then addresses the last 3.3% of residual unmet demand by burning expensive green hydrogen that was (inefficiently) produced during times of excess power. When including green hydrogen as energy storage, zero-emissions is possible for only 110% the cost of the polluting solution.

The hydrogen cost assumptions aren't particularly rosy, and not only would the scenario's green hydrogen be significantly more expensive than grey hydrogen produced from methane, it would be much more expensive than burning methane in a gas turbine. But it's cheaper than building wind and solar resources that are >90% curtailed. It's cheaper than doubling the number of batteries, which the over-provisioning scenario does in addition to deploying heavily curtailed renewables. It's cheaper than carbon sequestration. It's cheaper than nuclear.

That conclusion surprised me, so I played with the model extensively to see if it was sensitive to any of the cost assumptions, and under all plausible assumption sets, including aggressively pessimizing hydrogen cost and being beyond aggressively optimistic about battery and solar prices, zero-emissions was always cheaper when the last X% of electricity was solved by burning expensive green hydrogen. Changing the assumptions just meant that X would be 1% instead of 3%.

I don't expect to see green hydrogen get deployed anytime soon for the simple reason that it's currently cheaper to pollute. Until either the policy or cost assumptions change, our grid will continue moving in the direction of the 85/15 lowest cost solution. Putting a price on pollution would be smart, and efficient, but we seem pretty far from that being politically feasible.

It's page 7 on an only sort of related side conversation in this article's comment thread.

Let's call it done and go get a beer?

 

[Walker On Earth]

I saw your earlier history with this poster. A few points:

a) I saw nothing 'straw maning', as Ddopson put it. And just making the accusation doesn't fly; specific instances must be noted along with reasons why these objections are a strawman.

b) The burden of proof is on the one making the claim, not on anyone else. Saying that you've got to show why hydrogen storage in infeasible is, at best, showing a profound misunderstanding of how this works. At best.

c) Twiddling the knobs on a model of costs isn't proof of the underlying feasibility of the technology in question. This one is bonkers insane. How anyone could think this is beyond crazy talk. If anyone truly believes this, I've got a a little model for the various costs of nuclear fusion, want to buy some shares I happen to have handy? I trust my point is made. The fact that it had to be made is deeply depressing, considering that posters here fancy themselves more tech-savvy than most, God help us all.

There are several more points I could have made, but I want these three, especially the last one to sink in.

 

[Walker On Earth]

I don't see much, if any, discussion on the technological hurdles to be overcome. These are -- or should be -- well known. For those who can't be arsed to research the subject, here's the hasenpfeffer lady giving a capsule summary. For anyone pushing alternative energy who denounces the awl companies as being pure evil while at the same time being on about the wonders of hydrogen, well, let me say they seem awfully gullible. Seeing how as a lot of that hydrogen hype is being pushed those self same companies and is widely denounced as yet another attempt at greenwashing.

 

[DDopson]

I saw your earlier history with this poster. A few points:

a) I saw nothing 'straw maning', as Ddopson put it. And just making the accusation doesn't fly; specific instances must be noted along with reasons why these objections are a strawman.

b) The burden of proof is on the one making the claim, not on anyone else. Saying that you've got to show why hydrogen storage in infeasible is, at best, showing a profound misunderstanding of how this works. At best.

c) Twiddling the knobs on a model of costs isn't proof of the underlying feasibility of the technology in question. This one is bonkers insane. How anyone could think this is beyond crazy talk. If anyone truly believes this, I've got a a little model for the various costs of nuclear fusion, want to buy some shares I happen to have handy? I trust my point is made. The fact that it had to be made is deeply depressing, considering that posters here fancy themselves more tech-savvy than most, God help us all.

There are several more points I could have made, but I want these three, especially the last one to sink in.

I'm not sure which planet you are returning from, but I never used the phrase "straw maning [sic]". Please don't put words into my mouth.

On your second point, are you really trying to argue that hydrogen storage is infeasible? Do you believe that the moon landings were faked? How does NASA fuel their rockets if it's infeasible to store hydrogen? How does the chemicals industry process 100 Mt / yr of grey hydrogen if it can't be stored?

Hydrogen has been well-understood for decades. Your comparison to fusion is facile. It's like saying that you disbelieve my price for sand grading equipment, without even asking what that price was, so you're going to make up a fake price for a semiconductor fab to produce beyond state-of-the-art atom-width transistors.

Regarding the "underlying feasibility of the technology in question", electrolysis isn't some mythical research technology; it's merely expensive, not as expensive as building a nuclear plant, but more expensive than producing grey hydrogen from natural gas. This year, global production of green hydrogen will exceed 1 million tons. Which is small potatoes compared to the 100 million of grey, but it's easily enough scale that the cost of the equipment isn't some mythical unknown on the level of "can fusion be made to work".

My analysis regarding pathways to a zero-emissions grid did NOT rely on green hydrogen getting cheaper and in fact, I modeled scenarios where electrolysis costs were pessimistically an order-of-magnitude more expensive than today's current price point. That's what made the result particularly interesting. Pick your own favorite capex/kW number for the electrolysis equipment, and as long as you are within an order of magnitude of the true value, my analysis won't change.

You clearly didn't even understand my post, so in this case, I will say it: you are "straw man[n]ing" my arguments. Badly.

 

[SpikeTheHobbitMage]

Yabut here's the thing: Even if whatever study being quoted is accurate, those figures are irrelevant. Irrelevant that is, if you're also claiming that APG is an existential threat that must be dealt with immediately, and with massive amounts of resources. So no, you can't use LCOS figures and expect to make anything sensible out of them.

APG isn't really relevant. Renewables are replacing other generation means because they are cheaper, which means they are only going to become more prevalent over time. Grid storage is a natural outgrowth because renewables benefit from time-shifting between generation and use.

 

[SpikeTheHobbitMage]

Would it be possible to convert abundant green energy not in hydrogen, but in methane? Use carbon-capture to create the methane, and after burning sequester the CO2. The process is more complex, but methane is a lot easier to handle and more energy dense.

Possible, but not as energy efficient. Converting hydrogen and carbon dioxide into methane and oxygen wastes a fair bit of energy as heat.

 

[SpikeTheHobbitMage]

That's a big bummer. So what's the next alternative? Go closed cycle and recapture and reuse the carbon out of your own flue gas?

Reduction of iron ore to iron via electrochemistry, no reducing agent needed. It can be done, just not economically yet, also issues around very high temperature electrodes needed.

Hot salt electrolysis using molten calcium chloride as the electrolyte is a solved problem. That's how all of the world's aluminum and titanium is produced. The only reason we don't use it for iron already is that carbon monoxide from fossil fuels is cheaper than electricity.

 

[hpsgrad]

I saw your earlier history with this poster. A few points:

a) I saw nothing 'straw maning', as Ddopson put it. And just making the accusation doesn't fly; specific instances must be noted along with reasons why these objections are a strawman.

b) The burden of proof is on the one making the claim, not on anyone else. Saying that you've got to show why hydrogen storage in infeasible is, at best, showing a profound misunderstanding of how this works. At best.

c) Twiddling the knobs on a model of costs isn't proof of the underlying feasibility of the technology in question. This one is bonkers insane. How anyone could think this is beyond crazy talk. If anyone truly believes this, I've got a a little model for the various costs of nuclear fusion, want to buy some shares I happen to have handy? I trust my point is made. The fact that it had to be made is deeply depressing, considering that posters here fancy themselves more tech-savvy than most, God help us all.

There are several more points I could have made, but I want these three, especially the last one to sink in.

i have a hard time deciding if it’s funnier that you think you have something to teach people, or that you think the above might be an example of such teaching rather than some boorish virtue signaling.

Either way the comic value is vastly outweighed by the arrogance and vapidity.

My offer from earlier stands. We could engage in something productive and quantitative, if you actually wanted.

ETA a link to the other thread where my offer was made: https://meincmagazine.com/civis/threa...rst-half-of-2024.1502569/page-6#post-43114734

 

[pepoluan]

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.

Though I agree with the numbers (I voted you up btw ), there's another thing: Political will.

If, say, regulation prohibits one from using Lithium batteries in large scale -- because seriously, they are potential bombs -- then these Lithium-battery-alternatives will be an answer.

Also if they're more robust / less maintenance they may be attractive even if the cost of storage is not less than Lithium-batteries.

 

[ALittleTeapot]

These batteries do not require the critical minerals that lithium-ion batteries need, which are sometimes from parts of the world that have unsafe labor practices or geopolitical risks. The minerals for these zinc-bromine batteries are affordable and easy to obtain.

Zinc - China produces more than four times as much as any other country, with only a couple of others being even within an order of magnitude. The country with the largest reserves is Iran. Bromine - mainly produced by Israel and Jordan from the Red Sea.

Yep, definitely no geopolitical risks or questionable labour practices involved anywhere.

 

[David Woodward]

Any idea what the actual chemistry of these are? They don’t disclose much more than it being aqueous, and mildly alkaline in the case of GridStar.

Neither company is open about the chemical details, trade secrets I guess. The best details I can find are:
Allegro - a microemulsion of a hydrophobic oil, with a surfactant to assist the emulsification. There’s a paper
“Electrochemical Double Layer Capacitor using a Microemulsion Electrolyte”
Fraser R Hughson, Rohan Borah, Thomas Nann
which is also available as a preprint at arxiv. That’s for work on supercapacitor electrolyte rather than flow battery, and of course they might have moved on from those particular components.

Lockheed Martin bought Sun Catalytix, a MIT spinoff, for the GridStar IP and talent. “Earth abundant” and cheap, non toxic is about all they mention. Old articles about Sun Catalytix mention “synthetic metal-ligand compounds”. There’s a Sun Catalytix patent
https://patents.google.com/patent/US8753761B2/enwhich gives examples of possible components, but not the specifics.

 

[The Lurker Beneath]

As seductive as that intuition may be, it's not a good description of the actual needs of our power grid.

In biology, most seasonal storage is about food storage due to crops not growing in winter. Grains are harvested in Autumn and used to sustain the family all winter long.

The power grid looks nothing like that.

Using real-world 2011 weather data for the continental US, with wind and solar locations sampled in proportion to the square of the capacity factor at that location (ie, prefer locations that generate more power), here's what the month-by-month wind and solar output looks like:

View attachment 90938

A few observations:

• Wind varies more than solar, with 1.93X between min vs max for wind, versus 1.68X for solar.
• Solar generates year round. Sure, it generates more in summer, less in winter, but except at high latitudes, wintertime generation is still very substantial.
• Total variation doesn't seem that bad (on this graph) - the combined total varies by only 1.2X from min to max, and the weakest month isn't even in the winter; it's a month where wind output was unusually low.

Looking at the monthly data (which is hiding some key details), "seasonality" doesn't really feel like the right description, at least not in the US, where the inclusion of the southern states make the solar curve pretty mild. Go to Norway, and seasonality is a much bigger problem. Still, there's an enormous fraction of humanity that lives in latitudes that are at least as favorable to solar as these US numbers.

It's when we look at the daily data that something interesting jumps out..

View attachment 90961
View attachment 90962

Solar has a moderate seasonal curve (higher in summer; lower in winter), with only a small amount of daily variation. I can tell you from my house's output curve that individual panels can drop significantly when a cloud passes over, but when you average over a broad geographical region, these panel-by-panel variations largely average out, making solar production very predictable.

Wind has large day to day, week to week, and month to month variability, and to the extent that there's any seasonal correlation at all, it's different than solar's seasonal correlation.

Comparing the avg-to-min, it's 2.2 for Solar, 2.9 for Wind, and 1.7 for the sum. This means that if we have diurnal storage, but no shifting between days, we need to over-provision +70% more generation capacity.

Thus in a grid that has long-term storage, while there's some seasonal trending, most of the back-and-forth is driven by managing that wind generation variability, with the storage capacity being used an average of 3.36X per year, not the 1X that you'd predict for a naive "seasonal" strategy:

View attachment 90963

This is one of the "conclusions" that I summarized in my post near the top of page 2.

Well, that's interesting, though it probably isn't going to help much in Northern Europe where the solar changes are large and we need heat in winter but little or nothing in the way of AC in summer. I guess if it works out then a system with a lot of renewables will tend to move naturally towards such an arrangement.

 

[bjn]

Well, that's interesting, though it probably isn't going to help much in Northern Europe where the solar changes are large and we need heat in winter but little or nothing in the way of AC in summer. I guess if it works out then a system with a lot of renewables will tend to move naturally towards such an arrangement.

Wind resources are much more abundant in winter at higher latitudes, making up for the shortage of summer photons. Yes, you can get intervals of high pressure in winter where you have little wind, but that not 3 months of the year, a week or two now and then.

 

[Kasper Wuyts]

Well, that's interesting, though it probably isn't going to help much in Northern Europe where the solar changes are large and we need heat in winter but little or nothing in the way of AC in summer. I guess if it works out then a system with a lot of renewables will tend to move naturally towards such an arrangement.

As mentioned, northern Europe has the bad luck of being an energy poor-region, but has some luck in that wind and solar complement eachother quite nicely season-wise.
In Belgium, our TSO basically divides the European 'storage' problem for a 100% renewable grid in 4 different problems:

1) Seasonal imbalance: In a 2050 winter in Europe, demand peaks, because heat demand in winter is larger than cooling demand in summer. We solve this by building out solar and wind in the correct proportions.

2) Day-to day storage, aka solving the 'duck curve problem: Solar peaks at noon, yet demand is higher in the evening and mornings. To compensate for this, batteries and thermal inertia of buildings are the solution. In our very well insulated home, shutting of the heat pump causes a very marginal drop in temperature in the hours that follow. Batteries, be they grid or behind the meter, will do wonders here.

3) Weather fluctuations. While sun and wind complement eachother seasonwise, this does not mean that this will occur day to day or even week to week. However, law of large numbers means this effect gets flattened out when looking at the entirety of Europe. So interconnections are the main solution to this problem.

4) dunkelflaute events; It might occur that over a period of some weeks in winter, both wind and solar produce much lower than average in the entirety of Europe. This is quite rare (2 weeks out of the year), but we need to be able to cope with it nonetheless. The economics are a lot different for this type of storage: We look towards energy sources with low capital costs per stored kwh (batteries are out of the question), but we allow much bigger operating cost per kwh. Because it's used so little, we don't see this as an urgent problem to solve and are currently quite agnostic about what will turn out the be the most economical option in 2050. Natural gas is currently the perfect solution, and this will probably remain the very last thing we do with natural gas up until 2050. Cheap to store, cheap to burn. Many say this is the place where hydrogen and green molecules in general will shine. The absolutely dreadful efficiency of producing green hydrogen is less of an issue because of how little we need it. Similarly, synthetic fuels, biofuels, waste fuels might be a solution here, although we will probably need those for aviation.

The irony about it all is that so much (online) discussion happens related to point number 4, which is an absolutely tiny part of the energy transition. Even if it turns out the tech is not there in 2050 to tackle point number 4 in a carbon neutral way, it will mean very little for the climate to just have that one point be something that's solved at a later date. (well, the tech is definitely here already, but I mean rolling it out at scale in time)

 

[numerobis]

Well, that's interesting, though it probably isn't going to help much in Northern Europe where the solar changes are large and we need heat in winter but little or nothing in the way of AC in summer. I guess if it works out then a system with a lot of renewables will tend to move naturally towards such an arrangement.

It will come as news to the French and increasingly to the Germans and Scandinavians that there’s no need for AC in summer. Before too long, London will also need much more widespread AC.

But in any case, if there’s a need for a lot of power to keep Northern Europe going, then grid connections to the med are a simple answer. It’s not that far, just 1-2,000 km, and you get to a very different energy regime.

 

[bjn]

Another point, given that the winter demand for energy in Europe is significantly due to heating, investment in insulating buildings will drop that demand and make coping with dunkleflaute events much easier. Conservation efforts can also be a great return on investment if done correctly, in the form of avoided expenditure. It also insulates yourself from price volatility and unlike investing in stocks, the ‘return’ is tax free to boot.

 

[mmiller7]

How “high voltage” are we talking. Telco (and networking) gear is all 48V, and all the stuff I use is 24V. I did once use a 600V UPS, but that was a very special application to feed an inverter for an induction motor.

When I was looking into it some years ago I'd come across some commercial switches and server gear for sale that if I recall were around 200-300 VDC input.

Its been a while since I've looked at it though and I haven't heard much since. Could be it was abandoned in favor of cheaper modern UPS gear and inverters have become better and cheaper over time (our APC UPSs at work run battery packs around 220V DC and supply 208V AC 30A single-phase to our racks)

 

[bjn]

It will come as news to the French and increasingly to the Germans and Scandinavians that there’s no need for AC in summer. Before too long, London will also need much more widespread AC.

But in any case, if there’s a need for a lot of power to keep Northern Europe going, then grid connections to the med are a simple answer. It’s not that far, just 1-2,000 km, and you get to a very different energy regime.

Don’t forget the Norwegians, they have huge amounts of hydro they export like you Canadians do. International interconnects are being built all the time, the idea being they use North Sea wind power from the UK/Netherlands/Belgium when the wind blows, and we use their hydro when nothing else is going.

https://en.m.wikipedia.org/wiki/North_Sea_Link

 

[paulfdietz]

Well, that's interesting, though it probably isn't going to help much in Northern Europe where the solar changes are large and we need heat in winter but little or nothing in the way of AC in summer. I guess if it works out then a system with a lot of renewables will tend to move naturally towards such an arrangement.

DDopson probably wasn't running his optimization exercises on Europe, but if did he'd see hydrogen having a larger role.

Constraining the problem to Germany, and using the 2011 weather data / 2030 cost predictions, including hydrogen cuts the cost nearly in half, a larger effect that he was seeing.

I haven't done Europe as a single entity, which might see less hydrogen due to statistical averaging of wind output.

EDIT: I selected most of Europe out to Poland and Hungary (excluding the UK, Norway, Sweden, and Finland). Hydrogen still saves, although not quite that much (53.6 euro/MWh vs. 79.1 euro/MWh). This is consistent with my expectations.