Technology Quarterly | Grid-scale storage

Smooth operators

Matching output to demand is hard with wind and solar power. The answer is to store surplus juice on the grid until it is needed

ON OCTOBER 28th a battery factory opened in Concord, North Carolina. That was good for an area which has seen dark economic times, but the event made few headlines. Perhaps it should have made more, though, for this factory’s owner, Alevo, a Swiss company, is not in the business of manufacturing cells for torches, mobile phones or even laptop computers. Rather, it is making batteries that can store serious amounts of electricity—megawatt-hours of it. And it plans to sell them to power-grid operators.

To start with, the new batteries will be used to smooth the consequences of irregular demand through the day by absorbing electricity during troughs and regurgitating it during peaks. If that pans out, it will eliminate the need for gas-powered “peaker” stations which fire up quickly when needed, but are expensive to run. It would also allow non-peaker stations to operate more efficiently. Alevo reckons that if a grid as big as America’s Western interconnection (which supplies the west of the United States and Canada) were to use 18GW-worth of its batteries the grid could save $12 billion a year. Though the company has no North American contract yet, it does have an agreement to deploy its batteries in Guangdong, China.

Smoothing the operation of existing grids, however, may be only the beginning. In the longer run, optimists believe, batteries like these, or some equivalent technology, are the key to dealing with the problem not just of irregular demand, but of irregular supply. As the unit cost of solar and wind energy drops ever closer to that of power from fossil fuels, the fact that the wind does not always blow and the sun does not always shine becomes more and more irksome. It is not just the great power-gap that is night which matters. As the chart below shows, even during the day—and even in deserts—the amount of sunlight can vary from minute to minute. And the wind, of course, is equally fickle.

Cheap grid-scale storage would overcome these irregularities. Renewables could then compete on cost alone. And there are many ideas for how to make this happen. Some, such as Alevo’s, are ready to be sold. Others work in laboratories but have yet to be scaled up for use in the real world. Others still are little more than twinkles of varying plausibility in their inventors’ eyes. But if even one of them is up to the task, then renewable energy may, at last, be able to stand on its own, rather than having to be subsidised and regulated into existence.

At the moment, grid-scale storage is dominated by pumped hydro. According to the Electric Power Research Institute, an American think-tank, 140GW-worth of this is installed around the world, with a capacity of 1.4TWhr. Pumped storage requires friendly geography. You need two reservoirs separated by a good gap of altitude. But it is then just a matter of linking them with pipes and using turbines that, if turned by falling water, generate electricity, but, when fed electricity, turn the other way to pump that water whence it came. Send it uphill when power is cheap, and let it flow down when there are spikes in demand, and you have a nice little business.

Not everywhere, though, has compliant hills and valleys. And pumped storage takes a long time, and a lot of money, to build. Technologies that start small, but can be scaled up as needed, are often a better answer.

Batteries now included

The immediate future of grid-scale storage, then, probably lies with real batteries rather than topographical ones. At least, Alevo thinks so. At full capacity, the firm’s factory in Concord should be able to turn out 16.2GWhr-worth of them a year. And Alevo is not alone. Tesla is building an even bigger factory near Reno, Nevada (see article) to make batteries for its electric cars and for local and grid storage.

Several stations that use batteries to regulate the output of wind farms have already been built, or are under construction. In Sendai, Japan, Toshiba is creating one based on lithium-ion batteries. This should open in 2015. It will have a maximum power of 40MW, and will be able to run at that rate for half an hour. The Notrees Battery Storage Project, which opened in Texas in 2013, uses lead-acid batteries—sophisticated versions of the type found in petrol and diesel cars. It has a maximum power of 36MW and could run for 40 minutes at full tilt. Another Japanese project, of 34MW, in Rokkasho, uses sodium-sulphur batteries. And one in Alaska, of 27MW, uses nickel-cadmium ones.

As that list suggests, many types of grid-scale battery technology are available. Alevo uses electrodes made of lithium iron phosphate and graphite. These are connected by an inorganic sulphur-based electrolyte, a combination, the firm claims, that is particularly propitious because cycling between charged and discharged states produces only a 1°C change in the battery’s temperature. This should eliminate the risk of overheating, to which some sorts of lithium-based cells are prone.

There are types of battery that actually require high temperatures to work. In sodium-sulphur cells of the sort deployed at Rokkasho both of those elements need to be liquid, meaning the battery has to be maintained at a temperature of 300-350°C. And an approach being developed by Donald Sadoway of the Massachusetts Institute of Technology would use two sorts of liquid metal, separated by a liquid electrolyte. The clever thing about this design is that, by picking a dense metal such as a mixture of antimony and lead, a light one such as lithium, and an electrolyte whose density falls between the two, the three substances will float as separate layers in a container, rather as oil separates from vinegar in a salad dressing.

Despite their superficial differences, one thing all these batteries have in common is that the energy they contain is stored chemically within their electrodes. This has a consequence, at least for those with solid electrodes. The constant change in the electrodes’ composition as they are charged and discharged gradually wears them out. This limited lifespan is one reason using batteries for grid-scale storage is still pricey. Indeed, Alevo’s claim that its batteries can undergo more than 40,000 cycles of charging and discharging without noticeable loss of function is an important part of its sales pitch.

An alternative approach, known as a flow battery, does not suffer from this difficulty. A flow battery’s energy is stored in its electrolytes (of which there are two, separated by a membrane), rather than its electrodes (see illustration 1). Not only does that stop the electrodes wearing out, it also means that there is no upper limit, based on the sizes of those electrodes, on how much energy such a battery can store. Its capacity depends instead on the size of the tanks used to hold the electrolytes.

Flow batteries are a much less developed technology than standard batteries, but they are beginning to become commercially available. Many of those on sale at the moment (by firms such as Gildemeister of Germany and UET of Washington state) use vanadium-based electrolytes. Vanadium is a good material because its multiple ionic states mean it can be used to store energy without having to involve other reagents, and thus complicate the design.

Unfortunately, vanadium is expensive. But systems that use cheaper materials are being developed. Several firms are trying zinc and bromine in electrolytes and others iron and chromium. Ideas still in the lab include flow batteries based on cheap organic compounds called anthraquinones. If these prove robust enough to commercialise, they will be strong competitors in the grid-scale storage market. But they will not be alone. For batteries are not the only route to the destination.

Pumped up

If the engineers at Gravity Power in Goleta, California, get their way, even pumped storage is in line for a makeover. Their approach, it should be said from the outset, is one of the most twinkly of the twinkling eyes in the field. Even if it ultimately fails it shows the originality of thought that is being brought to bear on the problem.

Instead of two large reservoirs at different altitudes on a hillside, Gravity Power proposes two water-filled cylindrical shafts—one wider than the other—dug into the ground (see illustration 2). The shafts will be linked top and bottom to form a circuit, with a combined pump-turbine, similar to the ones used in conventional pumped storage, in the upper link. The wider shaft will contain a huge cylinder, made either of the rock the shaft is cut through or of concrete, to act as a piston.

When the pump-turbine is opened, the piston sinks, driving water around the circuit and through the turbine, generating power. Spin the device the other way using electricity, and the reversed water flow pushes the piston up again.

How much energy this arrangement can store depends on how deep the shafts go. And that is where it gets tricky, for some serious civil engineering will be needed if the idea is to work. Gravity Power proposes the shafts descend hundreds of metres. This will require large thicknesses of suitable rock—in practice this will probably be limestone, which is soft enough to cut into—so deployment will be limited not so much by geography as geology. And making a good seal between piston and shaft will hardly be trivial. So it will be expensive. A unit 700 metres deep, with a main shaft 26 metres across and a return shaft (or penstock) of about a tenth of that, would cost $170m. It would, though, be able to store about 200MWhr of energy, with an output of 50MW. Building one that size is years away, but the firm hopes to start work in 2015 on a demonstration plant near Penzberg, in Germany, with a depth of 140 metres, a capacity of 500 kWhr and an output of 1MW.

Nor is Gravity Power’s approach the only one to rely on underground spaces and friendly geology. Another is to fill a subterranean cavern with compressed air. For that, the cavern needs to be hermetically sealed and this means using an underground salt dome that has been hollowed out by solution mining (ie, the salt has been extracted with hot water).

Given such a cavern, compressed-air storage is a bit like classical pumped storage, except with a gas, rather than a liquid. Air is pumped into the cavern, increasing its pressure, and then let out to drive a turbine. But there is a catch: gases heat up when compressed and cool when they expand. For compressed-air storage to work, therefore, the air released from the cavern has to be heated (usually by burning natural gas), otherwise it would freeze the turbine. That makes compressed-air storage inefficient—one reason there are only two grid-scale examples of it in the world (one in Germany, the other in Alabama).

This would change if the heat of compression could be captured, stored and recycled. And that is the goal of LightSail Energy, a firm based in Berkeley, California. LightSail has developed a small, but still grid-scale, compressed-air system that sprays water into the compression chamber, to cool the air as its volume shrinks. The air is then stored in a set of tanks with a total volume of 42,000 litres, and the water, with its heat load, is put into two tanks that have, in total, about a quarter of the volume of the air tanks.

At the moment, this device can store 700kWhr of energy, but that should rise to 1.1MWhr when (as is the plan) it is pressurised to 300 atmospheres instead of the current 200. That is a fraction more than one of Alevo’s battery packs, which store 1MWhr. For comparison, the Alabama salt dome can store 2.9GWhr.

If heat is to be stored at scale some inventors would prefer to simplify the process, get rid of the compressed air, and concentrate on sequestering the heat itself. Isentropic, a company in Fareham, Britain, plans to employ the compression and expansion of a gas (in this case, argon) to create heat and cold respectively in two large containers of gravel—one of the cheapest solid heat-storage media imaginable. Once again, a pump-turbine is involved. It does the compression and expansion when electricity is abundant, and when it is scarce the gas flow, and thus the heat flow and therefore the whole process, is reversed.

Nor are these ideas the end of the list. Several firms, from giants such as ABB of Zurich, to minnows such as Berkeley Energy Sciences, a neighbour of LightSail, are pushing giant flywheels as at least part of the answer. Another suggestion—for filling in the shortest irregularities in supply, those lasting a few seconds or minutes such as are caused by the passage of a cloud in front of the sun—is to use supercapacitors, which store electricity as an actual electric charge, rather than converting it into chemical or physical potential energy of a non-electric form. At the other end of the scale as far as the size of the gap in supply is concerned, namely the nocturnal hours when solar energy cannot operate, several research groups are trying to use molten salts (usually sodium and potassium nitrates) to store heat gathered during the day and then, at night, raise steam for generators with it.

The world would no doubt be a better place if the externalities imposed by fossil fuels were properly accounted for in the price of electricity

And there is one further idea around that, though it relies on new storage technology being developed, does not rely on that technology being developed specifically for grid-scale storage. This is to use the fleet of electric cars that its proposers hope will take over from ones driven by internal-combustion engines over the course of the next couple of decades.

In the imaginations of such people, the batteries of these cars (which would, when idle, be attached to the grid in order to charge them), could be employed as a giant storage network, to be plundered with the car owners’ permission at times of peak demand. It is an intriguing thought—but the overlap between those times and the times cars are most likely to be on the road might scupper it in practice. As might the answer to the question about how ubiquitous electric cars will actually become. For that will depend on the future success and affordability of batteries.

The path from startup to success is littered with corpses, and an awful lot of business models depend for their putative profit on what is, according to your point of view, either a subsidy or a factoring in of the economic externalities (in the form of climate change) imposed by fossil fuels. In particular, Germany’s Energiewende and California’s Renewable Energy Programme have, by requiring a large fraction of those jurisdictions’ electricity to be renewable, helped fuel the boom.

Your bill, sir

The world would no doubt be a better place if the externalities imposed by fossil fuels were properly accounted for in the price of electricity. But that is a hard sell, not least because of disagreements about those externalities’ true size. In the meantime, it is better if grid-scale storage can be rolled out without taxpayer support.

That is the main reason for watching the example of Alevo. It says it can make money even in unsubsidised grids, because it has been ruthless about reducing manufacturing costs and simplifying the technology as far as possible.

This is a businesslike approach. If it works, and others prove able to mimic it, then the cost of running a grid, and thus the price of electricity, will fall. That alone will be a good thing. But success will change the very nature of such a grid, enabling it to absorb more wind and solar power even if this is a consequence unintended by the grid owners. How much more is yet unknown, for fossil fuels (particularly natural gas) are getting cheaper too. But renewables will no longer be fighting the battle with one hand tied behind their back.

This article appeared in the Technology Quarterly section of the print edition under the headline "Smooth operators"

Sheikhs v shale

From the December 6th 2014 edition

Discover stories from this section and more in the list of contents

Explore the edition