by Nicolas P. D. Sawaya
figures by Brad Wierbowski
More renewable energy means more electricity storage
When the National Academy of Engineering ranked the twentieth century’s greatest engineering achievements, first on the list was “electrification,” beating out more obvious technologies like computers and spaceflight. If this choice seems banal, it is only because our electricity supply is so remarkably reliable that it never enters our minds.
In fact, the electrical “grid” – the intricate network of power plants, power lines, homes, industries, and power control infrastructure that sprawls across the country – operates something like a complex highway system, but with stricter constraints. The main challenge is that, at any instant, the electricity demand needs to match the supply. Utility companies have to plan for and stabilize any supply-demand mismatches on a huge range of time scales, from months down to seconds and milliseconds.
A new challenge is being introduced to the grid, as wind and solar begin to take more market share. Because the wind does not always blow and the sun does not always shine, it is becoming necessary to integrate electricity storage into the grid (Figure 1).
But this is not just an issue of, at will, supplying enough electricity to exactly match demand on scales of months, days, or seconds. You can think of renewable electricity as being a bit “bumpier,” in terms of milliseconds and seconds, than electricity from traditional power plants. As a result of these short power fluctuations, power infrastructure can be damaged. Electricity storage is necessary for smoothing out these short variations, in addition to the more obvious task of longer-term storage to save for cloudy days, solar-deprived winter, or a million people turning on their teapots at once.
All of the above: We need many storage technology options
It will be important for policymakers, utility companies, and scientists to keep in mind that an “all of the above” strategy is necessary for grid electricity storage, just as it has been for sourcing the electricity itself. Many countries, including the United States, produce their power from a broad mixture of nuclear, coal, natural gas, hydroelectric, wind, and solar power. Putting eggs in several baskets was a prescient plan, since it is difficult to predict how different technologies’ construction costs will improve, or how the price of a particular fossil fuel will change.
In the case of storage, a “smorgasbord” policy is similarly essential, largely because different technologies are cheaper for different purposes (Figure 2). You need to strike a cost balance between (a) how quickly you can release energy and (b) how much total energy you can store. You can think of this trade off with an analogy to a car. Paying for more energy output per second is like purchasing a more powerful car engine. On the other hand, paying for more total energy is like paying for a bigger gas tank. In electricity storage, some technologies have a more expensive “engine” than others, while other technologies are pricey if you want to increase the total stored energy.
Sometimes you need to be able to release lots of energy at once for just a few minutes but don’t necessarily need to do it for a long time. In this case, you care about the cost of the “engine” more than the cost of a larger “gas tank.” However, if you want to store weeks’ or months’ worth of energy, then your cost priorities are reversed.
Importantly, both energy release and energy storage considerations come into play when we talk about renewable energy. It turns out that wind power produces a choppier signal than solar power does, often necessitating the ability to release more electricity per second. On the other hand, solar is worse than wind on longer timescales, since virtually no solar is produced during a rainy week, and there is much less daylight in the winter. This means that total stored energy may be even more important for a more solar-heavy grid than a wind-heavy one.
In addition to cost considerations from energy per second output and total stored energy, there are three additional forces that require many storage options to be available. First is geography. Storing energy using a hydroelectric dam requires mountainous areas, for example, making it an option in only some locations. The second constraint is economical in nature. The availability of a broad set of cheap storage options provides a buffer to market fluctuations. For example, energy storage markets will be less affected by a cost spike in lithium-ion batteries if there are several alternative storage methods to choose from. Third, environmental consequences can affect the feasibility of building some types of storage: Building a hydroelectric dam for storing water may damage wildlife habitats, and certain battery types may be more prone than others to toxic chemical spill.
Though it is possible that, decades from now, just a few storage methods will outcompete all the others in cost, we don’t yet know which technologies will win that race. “We certainly need to be investigating many right now, because we’re very far from where we need to be in terms of low-cost energy storage,” says Jessika Trancik, a professor in energy studies at MIT. Trancik points out that although “it’s important to explore many options, the ones we end up with may be a handful, or may be more.”
Sampling the smorgasbord
Though various types of battery technologies are available, simply pushing water uphill remains the cheapest way to store massive amounts of total energy. This type of storage is called pumped hydroelectric power. When the energy is needed again, the water is run downhill through an electricity-generating turbine similar to the ones that produce electricity at hydroelectric plants like the Hoover dam. (A turbine is a large machine that produces power when you flow liquid or gas through it.)
However, geographical conditions must be met. In areas where there are no existing reservoirs or hills, pumped hydro is not a serious option. But in other regions, opportunity beckons: mountainous and hydro-friendly Norway is being touted as a possible battery of Europe.
Another promising option is compressed air storage. To store excess energy, air is pumped into a naturally occurring and sealed-off cavern, which increases the air pressure and density inside. Electricity is discharged when the pressurized air is released through a turbine.
Pumped hydro and compressed air are examples of technologies for which it is cheap to store lots of energy, since the energy is stored in natural reservoirs or caverns that already exist. Where geographically permissible, these are usually the cheapest methods of long-term and large-scale storage, like when you save lots of solar energy in preparation for the winter. However, it is expensive to build the small power plant that produces the electricity from the stored water or air, making these methods economically inefficient when you need to generate only minutes’ worth of electricity.
On the other extreme are mechanical flywheels, which can cheaply release a lot of energy at once. A flywheel is a large solid spinning disc. To store energy for later use, the flywheel is sped up using a large electric motor. When the energy is needed, an electric motor is essentially run in reverse, producing electricity from the spinning object. Because increasing the total amount of energy they can store (i.e., using a larger disc) is quite expensive, typical flywheel designs run out of energy after about 15 minutes. Still, flywheels will likely remain a cheap and reliable part of the storage smorgasbord for short-term functions.
Of course, electrochemical batteries like those used in your phone or laptop can be used for the grid as well. Batteries generally strike a balance between (a) cost for how much energy can be released per second and (b) cost for total energy stored, but don’t come with any geographical constraints. All batteries operate under the same principle: roughly speaking, positively charged particles, such as the chemical element lithium, move from one end of the battery to the other in order to store or release energy.
Large arrays of lithium-ion batteries, nearly identical to the ones used in laptops and electric cars, are already being used on the power grid. Though lithium-ion currently has a market advantage because it is already being massively produced for so many applications, other futuristic battery designs are poised to beat out lithium-ion in cost.
The chemicals and materials used to make these other battery types are cheaper. A promising replacement for lithium-ion batteries is the molten metal battery, which operates at hundreds of degrees Celsius. Because the molten metal does not degrade as quickly, the lifetime of these batteries is longer as well, which in turn drives down cost.
The flow battery is in play too, which cheaply stores chemical energy in liquid solutions in large tanks, before pumping it to an area that either generates electricity or stores energy, both through a chemical reaction. The advantage? Once the expensive components of the battery have been built, it is very cheap to increase the total amount of energy it can store; all you have to do is get bigger tanks and fill them with more of the appropriate chemicals. There is an especially large variety of materials and chemicals can be used in flow batteries, leaving a promising path to lower costs and improving performance.
The final essential storage category includes various forms of demand response, which can be roughly thought of as “virtual” storage (Figure 3). This is the notion that, instead of physical storage, we try to shift electricity use to different times of day. Simply put, utilities and legislators can get businesses and residences to avoid energy-intensive activities during peak hours. Ontario, for instance, charges different rates for off-peak and on-peak hours, which is then reflected in the customer’s bill. This is in fact cheaper than any storage method, as it requires no investment in real hardware.
We need strong policy to create a storage arsenal
When a utility company purchases electricity storage, it balances all the constraints mentioned above: costs for how fast energy can be released, costs for total stored energy, environmental considerations, geographical realities, and unpredictable market forces. Additionally, optimal storage implementation will depend on the fraction of electricity being produced from solar, wind, natural gas, nuclear, and coal. The pressing question is, how do we accelerate the development of electricity storage methods, to both bring down costs and to provide as many options as possible for tackling these constraints?
Trancik, the professor at MIT, appeals to diversity not only on the storage level, but also on the policy level. “We need to take a portfolio approach in policies as well,” she says. Generally there are three types of strategies. Taxing carbon dioxide emissions, funding very early-stage research at universities, and carefully designing utility regulation are all important strategies. The latter approach may be the most direct way to accelerate storage development in the short-term. California, Oregon, and Massachusetts have laws requiring their utility companies to purchase large amounts of storage, immediately creating an incentive for technological innovation that will help bring down costs. “These early markets can spur innovation and cost reductions. They’re really critical,” notes Trancik.
In a recent journal article, former energy secretary Stephen Chu and coauthors wrote that “the cost of decarbonizing the first 25% of the world economy is far less than the cost of decarbonizing the last 25%.” This last-quarter challenge, still decades away, is formidable largely because majority-renewable electricity is impossible without investing in energy storage. But we cannot rely solely upon lithium-ion batteries and storage dams, since they are unlikely to be the most cost-effective for every energy system and every region. Research funding and regulatory policy must be shaped to push for a wide variety of technologies to fill this need. Having an arsenal on hand is the best way to overcome the geographical constraints, energy production varieties, unexpected market shocks, and multiple functions that storage needs to provide.
Nicolas P. D. Sawaya is a PhD candidate in Chemical Physics at Harvard University.
For more information:
Energy storage: Tracking the technologies that will transform the power sector (Deloitte)
Value of storage technologies for wind and solar energy (Trancik Laboratory)