Why energy storage is a solvable problem

by John Quiggin on January 12, 2022

Most discussion of energy storage that I’ve seen has focused on batteries, with occasional mentions of pumped hydro. But in the last week, I’ve seen announcements of big investments in quite different technologies. Goldman Sachs just put $250 million ($US, I think) into a firm that claims to worked out the bugs that have prevented the use of compressed air storage until now

And several companies are working on gravity storage (raising and lowering massive blocks) to store and release energy

Underlying these points is a crucial fact in physics/engineering: Any reversible physical process is an energy storage technology.

That’s why concerns about the variability of wind and solar power will come to nothing in the end

Energy storage technologies differ in lots of dimensions, including efficiency (the proportion of stored energy that is released), speed of discharge, and the time for which energy can be stored without being lost. Conversely, an energy supply system has demands that vary on all the correspond dimensions. In particular, the storage time dimension can vary from milliseconds (the instant power needed to maintain system stability against shocks) to months (storage between seasons).



Brett 01.12.22 at 1:53 am

That Hydrostor set-up sounds pretty awesome. The upfront costs are probably a bit high, but if they can offer super-cheap storage after that then it would be worth it.

Basically any storage has to be capable of being deployed and produced en masse and over a wide area, and probably down around the $20-40/kW-hr price-point. Otherwise, Lithium-Ion batteries might steal their lunch – especially if they can get the latter down to below $100/kW-hr.

I don’t know where they’re drawing the estimate from, but the US EIA already has “solar photovoltaic plus storage” below $2000/kW in overnight construction costs. Cheaper than even the cheapest nuclear power plant in South Korea, although still pricy compared to natural gas. But getting closer.


Jim Harrison 01.12.22 at 3:31 am

Energy storage is certainly feasible. The problem lies in making it cost effective. That has turned out to be a neat trick. I worked as a technical writer and editor for power company researchers for a very long time. One of the first jobs I worked on back in ’88 or so was a project on compressed air energy storage (CAES). Didn’t work. Over the next 30 years, I copyedited papers on other storage schemes: gravity storage, molten salt, hydrogen, capacitors, fly wheels, many kinds of batteries. The most alarming idea was a plan to build an enormous tank of superconducting helium in which currents could be made to flow indefinitely. As I recall, analysis of what would happen in the event of an accident put an end to that one. My point is, energy storage is not a new goal. Before the utilities were worried about global warming, they were already obsessed with load leveling, the holy grail of the power industry. If the storage problem were easy to solve, it would have been solved years ago because serious research beginning Jimmy Carter’s day. Which is why I’m a hair skeptical about the umpteenth announcement a breakthrough. I’m not completely cynical, but I expect that progress in energy storage will continue be slow and incremental.


Ronald 01.12.22 at 4:33 am

For Australia, only hours of energy storage are required to meet electricity demand. Rather than go to the expense of building days or months of energy storage, it will be a lot cheaper to just to add more solar generating capacity. Wind capacity is also an option. This means we’ll have so much solar capacity the wholesale price of electricity will often be zero in the middle of the day, but this is a good thing. If anyone thinks high electricity prices are a good, I am quite happy to personally supply you with electricity that costs a fortune.

Another thing that gets overlooked is we are going to end up with a lot of energy storage on wheels in the form of electric vehicles. This applies whether we stick to private cars or end up with robot taxis. Charging road transport is something that can be done when renewable output is high and electricity prices low. This storage can also supply energy to the grid when it’s required. The wholesale price of electricity briefly hit $13 per kilowatt-hour in NSW this morning. I’m sure if EV owners were able, many would be happy to supply some energy from their vehicles for that amount.


James 01.12.22 at 5:01 am

The problem stems from the difference between kW and kWh. kWh much harder problem.


Kevin Cox 01.12.22 at 5:38 am

Estimates vary, but the Australian Household consumes about 30% of its energy heating water, and many have ways of storing hot water.
If a house has a swimming pool, about 10% of its energy pumps water through the pool. Most of the time, it doesn’t matter when we pump the water.
To take advantage of building heat and changing the time of energy use, we need regulations to allow 5-minute metering in the homes to match the 5-minute wholesale electricity market.
Such a system will also make more efficient use of batteries and other storage technologies.
It is an example of a relatively small change in regulations that allows markets to do their magic.
If we apply the same principle of capital flows to individuals over their lifetime, we can make adjustments and remove the need to charge rent for the use of money. A good example is the original HECS system.

Capital storage as ownership of assets is a costly way to store capital. We can learn from electricity storage and find cheap ways to store capital. For example, rather than giving a return on the use of the money, we give investors lower-priced goods and services for the use of their capital.


MFB 01.12.22 at 7:15 am

How much time have we got in which to develop viable energy storage systems, without which renewables are all but useless for the electricity grid? We certainly don’t have another three decades.


John Quiggin 01.12.22 at 7:32 am

MFB @6 We need to get it done by 2030, and large-scale installation didn’t start until about 2020, so 10 years. Mostly it’s not a problem of developing new technologies but installing enough of technologies that are already well developed. The big need is to shift solar a few hours from afternoon into early evening, which can be done with batteries – most new solar plants have a battery component. The technologies mentioned above are for longer periods.

Jim H. It’s helpful to distinguish between things that can’t be done with current knowledge (useful nuclear fusion) and things that can be done, but need more engineering effort than we have put in so far. Some theoretically feasible energy storage technologies will fall into the first category. GS are betting that compressed air falls into the second. Even if this particular bet fails, the point of the post is that there’s an essential boundless set of alternatives that can be tried.


Ronald 01.12.22 at 11:14 am

MFB, it’s a myth that renewables require energy storage to provide useful amount of energy to a grid. South Australia’s generation was over 40% wind and solar before any storage was built. Coal power had already been eliminated at that point. At the moment energy storage is still minimal and wind and solar are 62% of the state’s generation. So a lot can be done without storage. But, fortunately, the cost of battery storage has fallen so we don’t have to do without it.


LE 01.12.22 at 2:49 pm

Reversible process in a thermodynamic sense? Like you state later in the post, efficiency is still important. There are no real life reversible processes in the thermodynamics sense. For any unit of energy we store in as physical system we will only recover s fraction. I guess I don’t understand the point of the bold sentence.


Another Lurker 01.12.22 at 2:52 pm

Any reversible physical process is an energy storage technology.

You might wanna reword this, since the second law of thermodynamics tells us that no real physical process is reversible. In order to achieve reversibility you’d have to go infinitely slow and that’s certainly not a good feature of something used to tap energy storage from the grid.


Omega Centauri 01.12.22 at 5:52 pm

Jim @2 and Ronald @3.
Jim; During the period you were observing, storage was seen by utilities as an incremental improvement in operating costs. Now its becoming an essential, so the incentive is now orders of magnitude greater.

Ronald is right, that a system that is cost balanced, won’t have the sort of storage scale that the naysayers are always saying is impossible. With very cheap wind and solar, large solar/wind overbuild reduces the net cost of the system. Even on cloudy days, solar still delivers something, and say 3 times the solar capacity means decent supply even on cloudy days -that greatly reduces the draw on storage, as well as allowing rapid recharge when conditions improve. Its really a systems cost optimization problem. And not mentioned here, long distance transmission is another important way to reduce storage needs.


Tim Worstall 01.12.22 at 6:33 pm

I have long insisted – for what little that’s worth from an entirely non-engineering type like myself – that it will be hydrogen. If solar is cheap enough (and we’re seeing bids in Abu Dhabi at 1.8 cents etc) then electrolysing water works.

Long term storage of hydrogen is a problem, as is long distance transport. But that can be solved through known chemistry. If you’ve got cheap H2 then ammonia, or methane, or even jet fuel etc, can be made.

What you keep as H2 can be run through fuel cells like those Bloom boxes (I’ve worked on supplying the weird metals those use) and that’ll be just fine for local shifting of energy usage from generation by a few hours, maybe days. That you reform up to complex hydrocarbons can be used just like those products are now. It will – OK, I think it will – almost certainly be cheaper to go this route than it will to try and entirely change jet planes. That will work for longer term storage and or transport of that “battery” which is the electrolysed hydrogen.

Sure, any of this has significant losses during the process. But if your starting point is cheap enough – really low solar prices – then we can indeed end up with storable and moveable energy just like we have today. I have a feeling that some won’t like the likely outcome, which is that some of it will be stored in pretty much the same fuels we use today even if we have just created them by sucking the necessary C out of the air to start with.


Rapier 01.12.22 at 8:10 pm

Any mechanical energy storage system suffers large inefficiencies. In other words it takes a lot of energy to store energy. That is on strictly a thermodynamic basis. It is also a money problem and this solution like every energy solution being touted out there assumes if enough money is created we can fix the energy/warming problems.

The energy problem won’t be solved with more and more and more money. It’s a beautiful myth and seductive myth and I get why people sign on to it.


John Quiggin 01.13.22 at 3:24 am

A couple of commenters have expressed concerns about whether the statement in bold is consistent with the Laws of Thermodynamics. So, I’ll spell it out in more detail

  1. In ordinary language, lots of physical processes are reversible. For example, we can pump water from a reservoir up to the top of a hill, then run it down again.

  2. The Second Law of Thermodynamics tells us that any such process involves an increase in entropy. In this sense, no process is reversible. So, the only non-trivial meaning of “reversible” is the ordinary language meaning.

  3. As far as energy is concerned, the relevant implication is that the amount of energy (with given entropy) that we can get out of any reversible (ordinary language sense) process is less than the amount we have to put in. Again in ordinary language, the process is not perfectly efficient

  4. But the (economic and social) value of energy varies over time, on lots of different scales. As long as the value ratio outweighs the efficiency loss, energy storage produces a net increase in value.

  5. So, to restate, any reversible process is an energy storage technology


bad Jim 01.13.22 at 4:35 am

Here’s a fun new storage system:

The Ocean Batteryis a new bladder-based energy storage system for offshore wind farms, which garnered a “Best of Innovation” award at CES 2022.”


Alan White 01.13.22 at 5:21 am

Since solar energy made everything on this planet including us possible via planetary and biological evolution, aren’t all storage systems ultimately solar in origin if not immediately so? All I’m saying is that if not for the sun, then nothing in our galactic vicinity could go against entropy, even human-made nuclear power not to mention the obvious fossil fuels and such. Gotta give good old sol all the credit whatever the subsystem of storage.


notGoodenough 01.13.22 at 11:12 am

Just to throw my £0.02 out there quickly (lack of time precludes too much analysis atm.), my speculation (and this is considerable conjecture on my part) is that there are three areas of storage which will be increasingly important: electrical (e.g. batteries, supercapacitors, and supercapatteries), green hydrogen, and thermal.

Now this is strong generalisation (a lot is dependent on the transmission and generation that you have), but assuming there is a “sensible mix” (e.g. of VRE/nuclear/hydro/geothermal/etc.), then the main issues seem to me to be increasing efficiency, load levelling, and transportation. Batteries and supercapacitors can deal with electricity directly, and can offer very rapid charge/discharge and ca. hours storage. Green hydrogen (assuming electrolysers become sufficiently cheap and viable) will be very flexible, and can offer long term storage (as well as, potentially, key transport solutions). Thermal storage could be important, as it offers a way to recapture energy (making systems more efficient) as well as additional storage options (I do sometimes wonder if thermal storage will see a bump when the geophysicists/geologists/etc. from oil companies start putting their knowledge to different uses). I think pumped hydro will also be important and that there will be niches for other technologies too – there are some more speculative things out there (I for one would be very interested to see at scale demonstration data of Star’s HERO catalyst, for example), after all. But this would be my guess for now (again, very speculative).

The biggest problems I see are less technological (though certainly problems there still exist), but more socio-political – namely a) lack of investment (not just more money, though that might help, but also in messaging and concrete political action and commitments) and b) objections to deployment (it would be a bit unfairly dismissive to attribute this soley to NIMBYism, but certainly that would seem to be part of it). I certainly could be wrong (a lot is outside my area of expertise and research focus), but this is just my general impression.


oldster 01.13.22 at 11:55 am

I remember reading some years ago about a plan to use electric motors to drag heavy train cars up a hill, and then use their descent to power generators.

Seemed elegant and simple. Less likely that a cement block will escape from a rail car than that compressed air will escape from a chamber. And less invasive to aquifers. True, it requires some grade, but not much.
I wonder whether anyone is still pursuing this route.
Alan White — I am all in favor of reviving the old sun- worshipping religions. As long as they don’t require human sacrifice on Dec. 21.


David J. Littleboy 01.13.22 at 12:46 pm

As someone who TA’d freshman chemistry once, JQ’s on solid ground.

There’s always going to be some loss, of course; nothing’s free. Get over it.

But the devil’s in the details. How much loss, scalability, how much energy can you actually get out. I’m personally pessimistic, but could be wrong. For example, the last time I checked, H2 was hard to store and handle: known methods all leaked something fierce. But that was 5 or 10 years ago (there was an overview section in Science), and things change. The other day, Elon Musk was badmouthing H2 for cars. That means he’s worried about it eating his lunch, so I’m more hopeful for H2 than I was a couple of months ago. And I rode an H2 fuel cell powered bus here (Tokyo) when I went on an errand somewhat off my usual beaten track the other day.


Rapier 01.13.22 at 2:29 pm

All these energy storage, conversion, and production ideas are classic neoliberalism. Proposing to fix problems caused by the market, with the market.

Let’s unpack “Goldman Sachs just put $250 million….” This implies it was GS’s money. Well no. They found some marks to sell the investment to. It’s green, it’s cool, what’s not to love? All we need is more money to invest and build stuff. Big stuff. And get it they hope, or some other schemes.

That large chunks of what’s called Liberalism, is in the thrall of ‘markets are going to produce a world of plenty for most’, after 40 years of proving that markets wreck everything, is a sight to behold.

I suppose in 20 years, with modest incremental increases in energy efficiency with CO2 emissions falling, after hundreds of trillions of dollars are invested in the physical systems, that an OK for the average Joe and Jane citizen would be possible. The thing is the barbarians are at the gate and they want blood.


reason 01.13.22 at 9:57 pm

Seems to me this is a hard issue to generalise about. It is quite a different problem in Australia (where you need to store energy for 16 hours of so) compared to Russia or Scandinavia where you need to store it for 6 months. By my concern – is how quickly these solutions lose efficiency and what resource constraints you might get when you try to scale them up.


Omega Centauri 01.14.22 at 1:40 am

Green hydrogen isn’t a slam dunk. Some major investments are planned, especially by the EU, but its uncertain if the economics will support its usage on a large scale. Better early targets are as chemical feedstocks, including as a replacement for carbon in ironmaking. Roundtrip efficiency of
energy to H2 to energy is low. So far green H2 is several times the price of dirty H2 (from natural gas). It remains to be seen if the price can be brought down into the competitive range. One thing being pushed by fossil interests is methane to H2, with CO2 capture and storage. Its questionable whether this would be good for the environment.


Tom Slee 01.14.22 at 3:14 am

Some forms of storage have been around for ages. On a rainy afternoon about fifty years ago my parents took me around Dinorwig Power Station in North Wales and I was right impressed. It pumps water up to a reservoir at off-peak times and lets it out again during peak times. It seems odd that storage is still seen as a problem.


Seekonk 01.14.22 at 5:27 am

Using nuclear energy is reckless because of its toxicity, susceptibility to sabotage/accident, and the problem of storing the ‘spent’ fuel.

Burning carbon-based fuel is even worse because it is producing dangerous climate change and will very possibly result in a cascading environmental catastrophe.

So, we need to keep moving on the supply side by developing energy storage along with green/renewable solar, wind, and hydro-tide sources.

And it’s way past time to mention the unmentionable: manage our energy supply by reducing demand. Adopt a population-control policy. I suggest that we pay a stipend to women for each year that they defer having their first child.


KT2 01.14.22 at 6:19 am

+1 Rapier at 20.
Reading some of the background and ‘funded’ Hydrostor projects, ROI was between 100%, and -20% when replacing / augmenting mining production.

Initial tech & IP funded by Canadian government. Profit to capital. It seems the race to ‘unmeterable’ will, as always, disproportionately effect the poor. Technological improvements seem easy as compared with fixng inequality.

I love the idea of Hydrostor, big bladders etc, yet I still can’t see how inequality will be ameliorated. Unless governments fund scale up and production and retain returns.


Tim Worstall 01.14.22 at 6:41 am

Just a thought about this: “Even on cloudy days, solar still delivers something, and say 3 times the solar capacity”

Yes, I get the point, but has the cost of this been considered? All – to any useful level of accuracy – of the cost of solar is in the build out. So, if we’ve got to have 3 x capacity then our electricity must be costing 3 x. If it’s £50 a MWh or whatever the unit is then a system with 3 x capacity must be costing us £150 per unit. Which might well make solar at 3x capacity the desired technological solution.


Omega Centauri 01.14.22 at 2:41 pm

If storage is at the same site as solar gen, then some parts of the solar infrastructure don’t need to be overbuilt. These include inverters, and the tie-in to the grid. I’ve heard of solar farms being built
with 50% more panels than needed for full design output, which means significant curtailment under ideal conditions, but 50% better output during mediocre conditions. This is being done to optimize profits for the farm, rather than provide better grid services. Change the incentive structure and this overbuild ration should increase. Panels are cheap, and getting cheaper, so the raw cost may not be great.


Dsquared 01.14.22 at 5:25 pm

I think the important point here is also that if energy is cheap enough, even ludicrously inefficient storage technologies are economically viable. If we get anywhere near full exploitation of solar and wind, we can tolerate technologies which chuck away three quarters of the energy put into them and still have electricity that is too cheap to be worth metering


Tim Worstall 01.15.22 at 7:59 am

Green hydrogen isn’t a slam dunk.”

I agree. I was putting forward merely an opinion as to where I think it might end up. I’m also with you @27, and with D2 @28. If that initial ‘leccie production is cheap enough then many of these varied storage options become viable. “If”. Of them I think that H2 solves more problems than near any of the others therefore it’s the one we’ll finally end up with.

I will admit to a bias. I’ve worked on (do not now, so this is an emotional bias perhaps, not a current financial one) the weirder rare earths for many years, scandium particularly. And solid oxide fuels cells, which would be a vital part of the H2 cycle, preferentially use scandium (Bloom and their boxes is the world’s major consumer). Even to the point that an Sc/Y doped SOFC would be a “better” car powering fuel cell than the more usual PEM type.

I’ve long mused – and no more than that, technically minded I’m not – that a viable system could/would be constructed around solar and SOFCs. Solar, as prices have shown this past few decades, are subject to something akin to Moore’s Law at least concerning their price, perovskites might well push that out another decade or more. There’s nothing inherently expensive about silicon after all (yes, I have visited factories that make the ingots, one of the problems overcome was to make those ingots “less” pure than those used for computer chips, another was to upgrade the size of them, there being a cost per kg and also a cost per ingot, larger kg ingots being cheaper overall). The same is true of SOFCs, they’re also subject to something akin to Moore’s. Early Bloom boxes had the rare earths hand painted onto them – replacing that with thin film deposition would likely work and while I’m a decade out of date not done so far I think. For an SOFC really is just zirconia (cheap as, umm, chips) with your choice of rare earths and some circuitry traced upon it.

So, solar for ‘leccie, when not being used run that through an SOFC and collect the H2. When dispatchable power required, run the H2 back through the SOFC (or another one) to get ‘leccie.

Yes, H2 storage is a problem, so is kablooie, and yet petrol and natural gas also have similar problems. If, and it is if, solar panels/cells and SOFCs become cheap enough then that is a viable system. Would almost certainly be a highly decentralised system given H2 transport problems but so what, it is viable. Well, OK, maybe viable.

The longer term storage and transport problems of H2 might well be solved through chimstry. Once you’ve got cheap H2 then making say, jet fuel, isn’t difficult, in fact it’s well known. That might be cheaper than trying to change the globe’s air fleet. But that’s becoming more speculative….

Comments on this entry are closed.