A couple of additions and caveats.

Friend of mine at ORNL has been talking up compressed air energy storage (CAES). I'm somewhat dubious due to the Boyles law heating/cooling problem (gasses heat when compressed, cool when expanded, and the resulting heat flows affect both efficiency and equipment function -- gas turbines don't spin well when iced solid).

Pumped-hydro is the most efficient storage solution bar none, but is painfully restricted by siting opportunities. Being able to use oceans as a lower baisin might address this, though salt-water engineering is a significant concern (corrosion and other issues).

Interesting to see molten salt as at large scale that pencils out a viable for even, say, a two week total energy replacement solution for the US. You'd need tankage (insulated) roughly corresponding to current Oaklahoma oil transshipment storage facilities.

Electricity-to-fuel is another option, with Sabatier and Fischer-Tropsch cycles both proven. Carbon source from seawater appears somewhat promising, though my concern is that there's been 50 years of research without a large-scale demonstration project, making me suspect deeper issues.

I actually am a fan of load-shedding, and suspect that the concept of dispatchable load as opposed to supply may be part of the future way of thinking.

Coal => gas is a no-brainer, though also not a total solution, by a long shot.

Geothermal is highly underrated, especially if currently protected areas are opened for consideration. The Yellowstone caldera could possibly supply up to 15% of total US electrical demand.

There also are ideas to use the ocean as the higher basin. See https://trustyetverify.wordpress.com/2014/12/11/energy-from-..., https://www.technologyreview.com/s/510806/a-manmade-island-t...

I haven't seen anything about it recently, but the Navy was/is funding a demonstration plant in Hawaii. I really hope it works out as I'm skeptical that the developing world will shift to electric vehicles anytime soon.

Her Ted Talk[1] and their website[2] explains it much better than I do.

[1] https://www.youtube.com/watch?v=hZiaTV6uvFQ [2] http://www.lightsail.com/

When sizing an industrial air system a rule of thumb is to budget 7-8hp of compressor for 1 hp of air motor. That is 12% efficiency, or only worth it if you had no other source of power. I'll give them the benefit of the doubt and assume a tightly integrated system you could do 3x better, but it still doesn't come close to batteries.

If someone has plug-plug numbers that are better, let me know.

My understanding isn't particularly deep, and I'd like to run numbers on storage, volume, pressure, etc. But I suspect this'll have to be big enough that constructing cans really isn't viable.

If it was more efficient, than we would see air powered cars (not the scam that surfaces every other year).

Industry only uses it to power things because you can have simple & compact actuators and high rpm motors.

Pumped hydro's remarkably efficient, fast-responding (minutes), but extremely limited in scale. See Tom "Do the Math" Murphy, but I think you'd need ~1,500 or was it ~15,000 Hoover-dam sized project for US storage capacity.

Batteries are fairly efficient, but respond slowly (you can neither charge nor discharge them rapidly), and at scale virtually all have massive substrate / materials shortages -- lead and lithium particularly. Iron-based batteries should be abundant, and liquid metal or molten-salt batteries seem to be the most abundant. Storage densities aren't great, but if you're building stationary facilities, that's not a concern. The 600C - 800C temperatures may be though. Having a neighbourhood melt-out wouldn't be pretty.

Biomass might work for standby thermal generation.

Thermal salt storage also pencils out. But any thermal storage system, that is, hot stuff you use to boil a working fluid that runs through a gas turbine, suffers from Carnot efficiency losses, about 35-45 max efficiency. If in a desert area, you've got problems of cooling your working fluid without venting too much of it (or the coolant). Not insoluble problems, but an issue.

Heat also doesn't store indefinitely though it should be good for hours to a few days.

Fuel synthesis has the advantage of scaling fairly arbitrarily large, particularly with liquid hydrocarbons. Put them in liquid-proof tanks and they'll stay there. Storage stability is proven to 100s of millions of years, so there's that. They're also transportable and can be utilised on-site. You lose 50% in hydrogen electrolysis, plus 65% on Carnot, for a net return of about 17%, but if you've sufficient peak surplus, that's viable.

CAES has the hot/cold problem, but there's a lot of air, and there's a lot of underground reservoir. Again I'm not sold that it pencils out, but it's possible.

Other options include direct banking of heat for buildings and/or industrial processes -- these use a lot of energy, and direct application and storage avoids electric or electric-to-fuel conversion losses.

It'll be interesting.

Anythhing exposed to fluid isn't. Groundwater flws, tanks in air.

I'm not sure about possible ignition explosion risk from residual methaane or hyrogen, say.

Molten salt is efficient because of the phase change, but also much more difficult and expensive to handle than a big (well enormous) bunch of hot rocks. Last thing I heard, he's currently trying to get demonstration plant up and running - it sounded like they'd already done some experiments at Siemens before he left.

Anyway, nowhere near ready, but just to show that there are still low-tech options out there being examined.

Another low-tech option is biomass, either wood chips/pills or other stuff processed into methane.

Yeah, it's not as cheap as coal, but with good grid connections we need less storage than people seem to think, at least according to the research papers I've seen.

Some numbers: 100% renewable energy with a capacity of 340% would require 8% storage just for Germany, 3.8% for Europe (due to larger area which means more diverse weather). Also 8% for Europe using no solar and more wind instead.

The 340% capacity is needed to get to the 100% coverage since there is more downtime compared to old energy which had 180% in 2007.

If we go to 390% capacity (which corresponds to 170 GW additional production as calculated for Germany) we only need 0.9% storage.

Depending on the relative pricing of additional capacity and storage, the price for building the whole infrastructure can go down by half. (The upper limit given in numbers is of the same order of magnitude as Europe's GDP in 2009, 12·10^12 €)

[0] (German) http://onlinelibrary.wiley.com/doi/10.1002/piuz.201201301/ab...

Much better than water - higher energy capacity means smaller units and higher performing insulation making losses much lower.

Heat loss is a bit problem in stored water DHW scenarios and will get even more of a problem if we begin building houses properly (which surely has to happen at some point).

1) You can move them around, as in an electric car. That's useful if you want to do mobile stuff like driving. Or if you just flat-out want to move the facility somewhere it will be more useful.

2) They can be efficiently decentralized, which benefits reliability, and decreases transmission loss.

But anyway, the solution is to use any and all technologies, right? Batteries make sense some places, gravity storage elsewhere.

While an interesting theory, I can't think of a single scenario where this would be economically justified.

> 2) They can be efficiently decentralized, which benefits reliability, and decreases transmission loss.

Transmission losses are about 6%, so not really any economic gain.

http://www.rmi.org/electricity_battery_value

I have no idea if that's realistic. Does someone know what the weight/distance to power output would be limited at...

Play with this and you'll immediately see it doesn't seem useful. e.g. storing a 50-tonne weight at 10m gives you less than 1.5 kWh of energy.

Flywheels that don't become fly-apart wheels is one problem.

Containment of fly-apart wheels another. Buried concrete pipes are usually what you're looking at. You need mass to contain these things.

The per-kg energy densities just aren't all that high. If you want reactive load to soak up very short-term demand / supply fluctuations, you can get that. And you've got crazy charge-discharge cycles.

Bearings are mostly solvable with magnets, as I understand. But you still lose a given percentage of storage over short time -- ~1% per hour or so, within a few powers of 2. Enough that long-term storage isn't viable.

And the thing that really screws you over is that the Earth moves. Geometric precession of your flywheel is something you've got to deal with.

There's a place, but I don't think it's large-scale, long-term grid storage.

There is a modern effort to use the technology in a massively scaled up fashion by Heindl Energy[1]. Their concept uses pumped water to raise a rock weight 100 meters or more in diameter.

[0] https://en.wikipedia.org/wiki/Hydraulic_accumulator [1] http://www.heindl-energy.com/hydraulic-rock-storage/overview...