From liquid air to supercapacitors, energy storage is poised for a breakthrough | Hacker News Reader

From liquid air to supercapacitors, energy storage is poised for a breakthrough

theguardian.com

50 pointsjsingleton68 comments

I've spent the last couple years working on a very interesting project using ultra-capacitors for energy storage. We are storing 33kWh of energy in ultra capacitors, charging and discharging with a very short period (seconds) for peak-demand shaving. It really is an exciting time for energy storage technology. I think we are going to see a lot of cool innovation in the next few years, even just finding new uses for existing technologies.

What sort of volume/weight dimensions are required for 33kWh of ultra capacitors?

They are quite large, but in the particular application weight and volume are not a huge concern. All sitting side by side they are about 20 meters long, 3 meters high, and a meter deep. Probably weighs on the order of 15 tonnes.

Wow, how much capacitance is that?

16.3 megafarads, assuming typical cell voltage of 2.7v, and total energy of 33kWh:

http://www.wolframalpha.com/input/?i=33kWh%2F%282.7v%29**2

I don't know much about this field, but how does charging/discharging on the order of seconds help shave peak demand?

Think of a sinusoidal demand for energy, with a period of a few tens of seconds. By charging when demand is low and discharging when demand is high, peak load is lowered.

Sounds like a logical extension to power factor correction. Large consumers have long been installing capacitors to compensate for inductive loads. Although this new tech is over a longer time frame.

To put this in context for others, although domestic customers are normally only billed for real power this is not the case for industrial consumers. Reactive power is significant as they are billed not just for energy used but also for peak power drawn.

A good overview, although missing Vanadium Redox Flow batteries (liquid battery), which is one of my favorites.

Redox Flow batteries can be pumped into storage tanks and distributed (theoretically anyway). In contrast, the entire set off Lithium Ion battery needs to be connected to the grid to take advantage of its storage, you can "pump" Redox Flow batteries into storage tanks to increase capacity without increasing power.

For day/night cycle storage (or longer), I'd expect Redox Flow to be more cost efficient than Lithium Ion, despite LiIon having numerous advantages (density / power / etc. etc.). Mainly because you can save on "Power Output" and instead build larger and larger tanks to store the charged electrolyte.

LiIons get more and more expensive ($$$ per KW-hr storage) the "bigger" they get. Redox Flow batteries achieve scalability (cheaper $$$ per KW-hr the bigger they get). I do realize I'm imprecisely using the word "bigger", but I hope yall will forgive me on that :-)

My personal favorites:

1. Pumped Hydro -- Old technology, but a favorite of mine. Very simple design, just take advantage of lakes that already exist in your vicinity and use the water in the lake as a gravity storage device.

2. Flywheels -- Store energy by spinning a heavy object. Spin-spin-spin... spinning and gyroscopes are cool.

3. Vanadium Redox Flow Battery -- Liquid batteries, how cool is that?

> (...) you can "pump" Redox Flow batteries into storage tanks to increase capacity without increasing power.

Why wouldn't you keep them at maximum capacity all the time?

Cost efficiency.

If your neighborhood only needs 50MW of power, but you want 1000 MWhr of storage (~20 hours at full power), a Redox flow battery would be able to do that. Or if you were building a Hospital (Lets say 5MW of power) and need to spec out 3-days of power (ie: 360 MWhrs of storage on 5MW Power), you just build a tank that can store a lot of the electrolyte.

I guess... the best way of putting it is that Lithium Ion batteries are like AAA batteries, while a Redox Flow battery could be configured as a larger D battery. D Batteries don't give any more power or voltage than AAA batteries, but are more efficient for energy storage.

It is very easy to imagine how to build a 3000MW-hr 50MW Redox Flow battery (just build a huge storage tank). It is almost impossible to imagine the same with Lithium Ion (I guess you can build massive Li-Ion cells, but note that the Tesla is powered by tiny 3.7V 18650 cells wired up together. So it really hasn't been done before)

Day/Night cycle storage is a low power (Megawatts) high-storage (Megawatt-hours) use case. Lithium Ion batteries are more useful for high-power / medium-storage situations (like frequency balancing, or maybe a couple hours-long storage).

Flywheels are exactly the opposite. They are high-power low-storage devices. Flywheels would be very useful for smoothing out power when birds / clouds come up in front of solar panels (for example). Its easy to build a 300MW flywheel with almost no capacity (only a few minutes of storage). Flywheels could be used when a power-source goes offline suddenly, while a few minutes need to be spent spinning up a new power-source. (ex: Main grid goes down, you need 30 seconds before your diesel generator fully turns on. The flywheel will keep very high-power for the next 2 minutes while things switch over). Although supercapacitors are also a promising technology for this problem.

Different storage technologies for different purposes.

They would be charged. You're confusing energy and power, by increasing the reservoir sizes you increase energy storage capacity but not the speed at which the total stored energy can change (power).

Pumped storage is great if you have the geography for it. In a lot of places (such as the UK) it's all already in use.

Yes. Some places, like the island of 'el hierro', even are disconnected of the main grid:

http://www.npr.org/sections/parallels/2014/09/17/349223674/t...

I posted this yesterday but today there's a story about "Welsh home installs UK's first Tesla Powerwall storage battery". Good news.

http://www.theguardian.com/environment/2016/feb/05/welsh-hom...

This project to store energy in trains also looks interesting.

http://www.aresnorthamerica.com

Energy storage is such a no-brainer. It makes solar/wind viable for both base load and peak load generation. The numbers required are very measurable and predictable. It doesn't require magical, exotic technology (as the elevated reservoir and compressed air mechanisms show). There's always going to be some thermal losses in transformation, but they're going to be pretty similar across technologies.

Another nice thing about solar + storage is that it can scale very small, very inexpensively. Imagine putting a set of solar panels or a wind turbine plus some storage in a village in Africa. It may not be worth wiring it to an expensive grid, but just getting electricity there can make so many other differences.

Beyond that, the ability to small-scale and decouple reduces the power and influence of large corporations and the government-industrial complexes that wreak so much havoc. You don't need the drama and abuse of the petroleum or nuclear industries anymore.

Scalable? Compressed air and reservoirs are about the opposite of energy-dense and require a lot of space per unit energy.

I'm optimistic about the role that energy storage can play in converting green tech to base-load capability, but it seems like it requires some technological advances unless you want to devote huge territories to becoming low-density batteries.

I'm not talking about physical size. I'm talking about cost and complexity. An air tank with a compressor pump at one end and a turbine at the other is simple, reliable, well-understood, off-the-shelf technology. How much energy can you store in an air tank that would fit on the back of a truck? Hook it up to another truck full of solar panels, and you have a local electric grid, capable of powering at least a few buildings, at very reasonable cost and minimal sophistication required for operation and maintenance.

Scalability covers spatial scalability, and so necessarily involves talking about physical size relative to energy stored in this context.

Is that actually important, though? Does it mean inexpensive small-scale energy caching in off-grid areas is impossible? Would a truck-carried air tank (or bank of them) hooked to solar-powered compressors not actually work, or be cost-prohibitive?

Size doesn't actually matter here.

But it does. The Wikipedia numbers [1] give a mass density for compressed air storage of 40 kJ/kg and volume density of 8 kJ/L. That's 0.011 kWh/kJ or 0.0014 kWh/L. So a 10,000 L tank gets you only 14 kWh, or ~$1.50 worth of energy.

[1] https://en.wikipedia.org/wiki/Compressed_air_energy_storage#...

This stuff is always "poised for breakthrough". But we are still stuck on LiOn batteries, which have been around for what, 20 years? Wake me up when we have something that doubles the capacity of LiOn batteries in a similar form-factor and safety rating, propelling Teslas to 600km+ ranges. That would be a real "breakthrough".

Could somebody enlighten me on the economics of energy storage?

The local retail energy prices are approximately 10c/kWh year round, rain or shine. The price is fixed, unless you buy a contract based on the spot market.

All the energy storage products I've heard about cost more or about the same per kWh just to store the energy. So even if the energy to charge the energy storage was free, it still would not make economic sense to store energy if you have grid power.

Am I missing something or is consumer energy storage only for off-grid applications?

Solar/wind are inconsistent, and their output doesn't map to consumption patterns. So a totally solar/wind grid would require massive overbuilding in order to provide sufficient power for peaks. By caching the surplus generated when supply exceeds demand, we can significantly reduce the overall supply footprint, by using the generated supply much more efficiently.

But why bother storing the solar/wind excess, if storage costs exceed traditional generation costs?

For a purely economical perspective, it does not make sense.

Because in the long run, storage costs will go down, and traditional generation costs will go up. At some point, there's a crossing where storage becomes cheaper. Much of that has to do with economies of scale and driving technological improvement with demand.

Beyond that, there are externalized costs to continuing a system that depends on fossil fuels and complex/expensive nuclear plants. In a macro world, governments and public utilities should be taking the externalized costs into account.

So the answer to my question is that the economics of energy storage do not work, but may at some point of the future?

I get both the technical and ecological argument, my point is that the economics to drive the change aren't here (yet).

We already do large scale grid storage.

https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit...

Flywheels are also used for short term demand spikes. https://en.wikipedia.org/wiki/Flywheel_energy_storage

PS: This is one of those problems that seems a lot more important than it actually is. The current state of the art is fairly cheap and reasonably efficient. Going from 90% to 99% efficiency really does not change much and is a far harder problem.

Part of the problem is that natural gas and diesel peaking plants have historically been cheap enough that there hasn't been a demand for alternatives.

Hydro only works if you have geography on your side. Flywheels don't have the capacity to do utility scale demand shifting.

So, excluding hydro, am I right in that the economics of energy storage do not work (yet)?

You can shift electricity 1,000 miles each way before losses stack up. So, geography is not really an issue.

ED: Currently, ~4,300 mi is about the limit on economical power transmission. But, that's a long trip.

So, can we build enough hydro to make a difference and how much will it cost to beef up the grid to handle all the power going into and out of pumped hydro?

Depends on what you want to do, but overall not much. Standard lines hung on towers run from $1 million to $4 million a mile, depending on terrain and other factors.

Though, we already have an electric grid and you can do a lot by simply shifting where each area's supply is coming from. AKA town A get's power from town B's power plant town B, get's power from town C's power plant etc.

Also, a lot of this infrastructure is a ~20+ year problem. So, a 50 billion dollar bill spread over 20 years spread over the entire US is ~8.50$ per year per person.

I want to replace peaker plants with storage. Somehow I don't think the economics work out. Do you disagree?

Even if we rebuild the whole continental US power grid (that's a lot more than $50 billion) to cope with the demands of using storage, it is unclear to me if we can build enough hydro to replace peaker plants. Apparently all other storage technologies other than hydro are cost-ineffective and/or do not have enough scale to solve the problem. Do you know if there is enough geography to go around for all the hydro we need or if there really are any cost-efficient alternatives?

You don't need to replace the grid, just add more capacity.

The economics for ~100% up-time only work out when utility's are regulated to act that way. That means spare capacity and backups for your backups. As such I feel a few Peaker plants are actually a good idea. They should not be turned on every day and should be able to pick up the stack when unexpected things happen.

In terms of hydro, we can build them to store ~12 hours worth of US electricity demand. Past that the geography still works just fine, but excess production becomes cheaper than excess storage.

PS: Don't forget we already use pumped hydro. It's just a maximization problem where cost of excess capacity vs. extra storage.

Batteries are becoming cost-competitive as well.

And as I said earlier, the economics don't need to work now in order to justify pilot programs and research, due to the long timelines involved in radically changing how the grid works. The assumption that the economics will work in the foreseeable future is enough.

> Batteries are becoming cost-competitive as well.

Can you quantify that?

As far as I can tell the Powerwall is about as cost-competitive as battery storage gets and that's still far more expensive than just using grid power.

Consider the time frame for this, as well. Radical alteration of the power grid will take at least a decade, probably multiple decades. Weaning it off coal, gas, and diesel is an enormous undertaking, even if it's technically feasible. Pilot programs now, even if they aren't immediately cost-effective, greatly reduce the risks of future change, when the cost equation is different. They'll shorten the timeline, reducing the long-term costs.

> All the energy storage products I've heard about cost more or about the same per kWh just to store the energy. So even if the energy to charge the energy storage was free, it still would not make economic sense to store energy if you have grid power.

1) They are re-usable. Power generation [e.g. Solar Panels] to refill them is frequently inconsistent enough that you will need to go 12+ hours without power. Cycle through enough time and paying a multiple of grid power for energy storage that lasts years is worth it.

2) If I lose grid power [happens a couple times a year] I need backup power.

EDIT:

Iirc, I'm pretty sure Powerwall and such is ~$.30/kwh with generation included and I know I eat ~$.30/kwh for a certain level of usage.

https://www.sce.com/wps/portal/home/residential/rates/Standa...

http://www.forbes.com/sites/christopherhelman/2015/05/01/why...

$.15/kwh during the day and $.30/kwh at night is actually comparable depending on usage.

If you don't want to do generation, you can arbitage rates at $.15/kwh storage + $.14/kwh charge from grid:

https://www.sce.com/wps/portal/home/residential/rates/Time-O...

No energy storage solution is infinitely re-usable. Thus there is a price per kWh per cycle.

Consumer energy storage solutions have a higher kWh price than traditional energy generation.

So why store energy as a consumer?

EDIT:

So energy storage only makes sense in a market where power is very expensive (~0.30c/kWh) and even then you might never recoup the investment required to store the energy.

Not all places are like this. Some places have variable time-of-use rates with high premiums at peak times and large discounts at times of extreme low demand. For example, rates with Southern California Edison can be as high as 45 cents/kWh and as low as 11 cents/kWh. If you can charge up a storage system at times of low demand and then use it at times of high demand, you could save a lot of money.

The bigger use is probably industrial applications, though. Commercial users can get hit with huge surcharges if they draw a lot of power, for example, so users which only draw huge amounts of power briefly could save a lot of money by using storage to smooth that out. The utilities themselves can end up paying extremely high rates for extra capacity at peak time, so if they can use storage to eliminate that need it can save them a lot of money.

> If you can charge up a storage system at times of low demand and then use it at times of high demand, you could save a lot of money.

In a sufficiently advanced energy market, you could actually make money.

> Not all places are like this.

Is there a reason why all places couldn't be like this? Wouldn't a well run and properly regulated power grid solve these problems?

Demand inherently varies at different times of day and different times of the year. Given variable demand, variable pricing is expected. Fixed prices are basically a leftover from when technology didn't support sophisticated metering for time-of-use pricing, and sticks around because people prefer simplicity in pricing.

For the sake of the argument, let's accept the premise that variable demand requires variable pricing.

Is there a technical reason price differentials have to be so large in California? In an efficient market, even with variable prices, would the differences be so extreme and would peak demand prices still exceed the cost of storage?

I'm pretty sure California is an example of a badly run system, and price differentials wouldn't be so big if it were run better.

In an efficient market, peak prices wouldn't exceed the cost of storage pretty much by definition, because the utility would use storage to even out demand as long as it's cost effective.

But I assume you mean whether it would happen without storage. That depends a lot on the underlying assumptions, I'd say. Does an efficient market mean one where you can build coal plants anywhere you feel like and pay only the cost of extraction, but not the cost of pollution? Or does it mean one where you pay the full cost of fossil fuels, rather than the weak and inconsistent controls we currently have?

I think you'll almost always benefit from some storage. You'll get diminishing returns as you increase generating capacity to match peak load, so storage would become the cheaper option at some point (assuming you're not shedding loads for extreme peaks).

Consider that no other industry operates on such a sharp just-in-time way, where supply and demand have to be matched on a second-by-second basis. Whether it's fruit or car parts or passenger airplanes, there's always some storage to allow slack in the system. The electrical grid historically hasn't been this way just because the technology wasn't there, but it's finally catching up.

A well-regulated and well-run utility does make consumer battery storage irrelevant.

Though, interestingly you could be taken as saying that flat rate electricity means well-regulated and well-run and often time-of-use rates lead to greater efficiency.

> The local retail energy prices are approximately 10c/kWh year round, rain or shine. The price is fixed, unless you buy a contract based on the spot market.

Chances are, you live close to the Bath County Pumped Storage Station.

https://en.wikipedia.org/wiki/Bath_County_Pumped_Storage_Sta...

Its because of these big infrastructure projects that your region has cheap energy all year (and all day) round. In areas like California, big infrastructure projects have been shuttered, and they have 41c / kWhr prices in the day, and 10c at night.

Other locations in the US have their own energy storage public-infrastructure projects that stabilize the grid. But some states (Hawaii or California) have... issues.

So energy storage is a technical solution to market failure?

Energy Storage allows energy arbitrage.

If a big utility company buys 10c "night time" energy and sells it at 20cs during the day, then they make money, and help fix the 40c daytime prices across California.

And unlike the decentralized "Powerwall" solutions, the poor will benefit from the big utility company performing energy arbitrage. (A poor family cannot afford a Tesla Powerwall to perform energy arbitrage for themselves. They rely on a larger, market actor to perform it on their behalf).

Basically, you don't notice it in your region because a big utility company is already doing this, which stabilizes prices for your region. The default state of the world however... is that electricity is cheap at night and expensive during the day. The various energy-storage projects in America exist to smooth out the day/night discrepancies.

I get the arbitration angle, but doesn't it purely depend on there being a market failure, i.e. failure to match production to demand and/or grid failure to transfer power between demand imbalance regions?

Wouldn't the better long term solution be to fix the market failure?

That's not how it works.

When a cloud covers up a solar panel, the region needs to grab energy from somewhere. When the wind stops blowing, the Wind Turbines require energy to keep them spinning (yes, during times of low-wind... wind power USES energy, not produces it).

It is very rare to find a stable source of energy. Nuclear and Coal are extremely stable, but the politics of those energy sources are poison to many communities.

------------

Note that the US uses twice as much energy during noon than it does at midnight. The only way this discrepency is solved right now is:

1. Energy storage -- store the energy generated at night to help out during the day. Nuclear and Coal power plants make just as much energy at night as they do during the day, so "storing" the excess from night will help out.

2. "Peaker" energy plants (Natural Gas can start and stop very quickly). But obviously, this is inefficient and wasteful.

Solar Panels sorta help from this regards, because they generate more electricity during the day rather than at night. But all Solar Panels will do in the long-term is shift the problem to 4:00pm when the Sun begins to set but electricity usage remains very high.

This isn't the case of "Kentucky is using more energy right now, pull from some Tennessee". This is the entire freaking US, simultaneously, uses double the electricity at one point in time, and then stops using it at another point in time. (A valid "load balancing" solution is to have night-shifts at factories btw. Not everything needs to be solved through technical innovations)

Finally, we have the issue of over-powering the grid. If the grid has too much power applied to it, it damages equipment. Energy storage can eat-up excess capacity on the grid safely and efficiently (and then sell it back at another time).

Energy storage is going on constantly under the modern grid. The US just needs more of it... we need more of it cheaper and in bigger amounts than ever before. Especially if we're going to start using unreliable sources of energy like Wind or Solar.

Given that it's never noon/midnight across the entire country, do you mean that at each point independently, usage is double at local noon vs. local midnight, or do you mean that there exists a single point in time where usage is double another point across the whole country?

If the former, if we had a better CONUS-wide grid, that could distribute power East/West more efficiently, how much would that help?

Each point independently, although the ratio can be as low as 1.5x during winter (summer air conditioning is a big problem).

East / West power distribution won't help too much, because there's only a 4-hour difference. The specific load curves of New England (which is only 1.5x load peak vs bottom at the moment... more typical of winter conditions) can be found here: http://www.iso-ne.com/

So, even in a prefect world, it all comes down to whether it's cheaper/better to overgenerate and store or to run peaker plants?

It seems storage has a long way to go as long as gas peaker plants are around 7c/kWh.

Energy storage is beating peaker plants today... under specific situations.

Note that gas peaker plants only run 10% of the time on the average (basically, they only work in the summer when our grid needs the most power, and then they don't make any money for the rest of the year).

From an economics / business point of view, having your power plants sit around only working 10% of the time is completely uneconomical. You need to be compensated severely to make it worth your while.

Or to put it another way: my energy storage company that gets 1c off energy arbitrage for all 12-months 100% of the year will make more money than your peaker plant that only gets 7c for 10% of the year.

And good luck convincing California to build a peaker plant. It is more politically sound to push for green-energy batteries. You'll get more approval in more locations in the US than gas (even if Natural gas is all-American and burns cleanly, its still a fossil fuel which is increasingly becoming a poison word in politics)

Your storage system must be able to store enough energy to cover the cumulative output of the peaker plant in order for it to be able to replace the peaker plant.

That's a lot of storage. And you need to have enough overgeneration to be able to reliable fill it up daily. This sounds very expensive. Could it really work from an economic perspective?

It already does work.

https://www.dom.com/corporate/what-we-do/electricity/generat...

This worked in 1970. Why wouldn't it work today? Energy storage solutions are deployed around the nation, and world, for this very reason.

The main advantage is that energy storage today is more efficient, with new technologies coming out.

> The main advantage is that energy storage today is more efficient, with new technologies coming out.

But not cheaper or anywhere close to the scale of pumped hydro?

Hydro requires favorable geography and/or rebuilding the grid to cope with the extremely large flows required to keep up with storage.

So, energy storage works, but is not economically feasible other than as research projects outside of hydro?

Looks like peaker plants are here to stay for a very long time.

In addition to the comments above about solar/wind inconsistency, it is also a lot cheaper to run baseload power plants constantly rather than to ramp peaker plants up and down. So, if they can store the baseload power and then tap it later the total cost of electricity will come down. Plus, storage makes it easier to level the grid to make output match demand.

Is storing power cheaper than using peaker plants? That is, is storing power at scales required to replace peaker plants economically viable?

Contextually relevant question: How much would it cost to do a major revamp of the North American electrical grid? (For various values of "major revamp" of course.) I found this article that mentions a $2 Trillion figure from Rocky Mountain Institute, but I take that with a big grain of salt. [1]

Since the patents around Tesla's Supercharger are open, what about the idea of giving the US highway commission a few hundreds of millions to double the number of Supercharger stations? (This would create more demand while not improving the infrastructure, but I would benefit personally.)

    [1] - http://www.renewableenergyworld.com/news/2015/04/us-power-grids-2-trillion-upgrade-needs-european-efficiency.html

Superchargers only work with Teslas, and Tesla is doing fine when it comes to doubling the number of installations on their own (there are currently 270, and they installed 113 of those last year), so having the government come in and install them doesn't seem like a good use of money.

Further, what's the relevance of Supercharger stations to electrical storage? Maybe I'm missing something but that seems like an irrelevant tangent here.

These buzz articles always focus on the sexy energy storage methods: lithium ion, liquid batteries, electrolysis, supercapacitors, etc. I've seen very little evidence that these can compete on either $/kWh, or efficiency, with pumped hydro (where available), thermal (Isentropic Energy Ltd [1]), or kinetic.

[1]: https://www.greentechmedia.com/articles/read/Isentropics-Pum...

Correct me if I'm wrong, but none of the options you listed are portable.

The article is all about storage for the electrical grid, to shave peak demand and allow more wind/solar use, so portability isn't a requirement.

Right. In the context of grid storage of renewable energy, that doesn't matter.

You need a li-ion or similar in your phone and your laptop (and maybe your car), but that doesn't make it the best tech for flattening out a peaky solar/wind grid.

"It doesn’t always rain when you need water, so we have reservoirs - but we don’t have the same system for electricity,” says Jill Cainey, director of the UK’s"

Actually... we do. It is called pumped hydroelectric storage [1], and it is essentially just a reservoir, a pump, and a turbine.

[1] http://energystorage.org/energy-storage/technologies/pumped-...

Reply on news.ycombinator.com